- Email: [email protected]
Inhibition of an inflammatory response is mediated by a 38-kDa protein of cowpox virus
VIROLOGY
172, 262-273 (1989)
Inhibition of an Inflammatory Response Is Mediated by a 38-kDa Protein of Cowpox Virus GREGORY J . PALUMBO,*' 1 DAVID 1 . PICKUP,t TORGNY N . FREDRICKSON,t LAURENCE J . McINTYRE,§ AND R . MARK L . BULLER* *Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892 ; tDepartment of Microbiology and Immunology, Duke University Medical Center, Durham, North Carolina 27710; #Department of Pathobiology, University of Connecticut, Storrs, Connecticut 06268; and § Vector Laboratories, Inc ., Burlingame, California 94010 Received March 10, 1989, accepted May 3, 1989 The Brighton Red (BR) strain of cowpox virus induces a flat, bright red pock on the chorioallantoic membrane (CAM) of the 12-day-old chick embryo . In contrast, mutants with a deleted 38K gene (which is located 31 to 32 kb from the right-hand end of the virus genome) induced a raised, white, and opaque pock . During the first 24-hr p .i ., both CPV-BR and the 38K deletion mutants replicated similarly in the CAM of the chick embryo, as indicated by immunocytochemical detection of similar amounts of virus antigen . By 48 hr p.i ., the pocks induced by the mutant and CPV-BR are strikingly different . The pocks induced by the 38K deletion mutants were infiltrated by large numbers of heterophils and macrophages, which correlated with a reduction in the levels of virus antigen and virus infectivity . The CPV-BR pock had an absence of inflammatory cells and increased levels of virus antigen and infectivity . By 72 hr p .i., many of the pocks induced by the mutant were undergoing resolution of the virus infection, as indicated by further decrease of virus antigen and visible signs of healing, whereas CPV-BR pocks continued to be a site of active viral replication . These data are consistent with a model where this 38-kDa protein directly or indirectly inhibits the generation of chemotactic molecules which are elicited during virus replication in the CAM or, alternatively, blocks the interaction of these molecules with cells of the host inflammatory response . n 1999 Academic Press, Inc .
INTRODUCTION
several plasma proteins that are inhibitors of serine proteases ; however, the recombination of this gene alone into the genome of white-pock deletion mutants did not result in the complete restoration of the wildtype, red pock (Pickup et al., 1986) . This suggested that additional genes may contribute to the red pock phenotype . The objective of this study was to determine the 38K gene function in the pathobiology of virus infection of chick embryo CAM in general, and to pock morphology in particular .
Cowpox virus produces two distinct types of pocks when inoculated onto the chorioallantoic membrane (CAM) of the chick embryo : one is flat, well demarcated, and bright red, while the other is raised, opaque, and white in character (Downie, 1939 ; Downie and Haddock, 1952 ; van Tongeren, 1952) . White pocks were detected at a frequency of about 1%, and on repeated passage bred true (Downie and Haddock, 1952 ; van Tongeren, 1952 ; Fenner, 1958) . The viruses that produced these white pocks appeared to be less virulent than the wild-type virus in inoculations of rabbits, mice, guinea pigs, nonhuman primates, and cows (van Tongeren, 1952 ; Fenner, 1958) . Studies of the genomes of these white-pock variants determined that this pock phenotype was associated with the loss of at least 32 kbp of DNA from the righthand end of the genome (Archard et al., 1984 ; Pickup etal., 1984) . Further studies by Pickup and colleagues (1986) identified the location of at least one major gene (designated 38K) involved in the determination of the white-pock phenotype, This gene was located between 31 and 32 kbp from the right-hand end of the CPV-BR genome, and encoded a 38-kDa protein predicted to have amino acid sequence homology with
MATERIALS AND METHODS Viruses, cells, and eggs CPV strain Brighton Red (CPV-BR) was originally isolated by Downie (1939) . It was grown in, or plaqued on, BS-C-1 monolayers cultured in minimum essential medium (MEM, GIBCO Laboratories) containing 10% bovine calf serum (BCS) . Eleven-day-old embryonated hen's eggs were obtained from Truslow Farms, Inc . (Chestertown, MD) . The eggs were incubated at 3738 .5° in approximately 50% humidity without rocking, or exogenous CO 2 for 24 hr prior to inoculation . Virus was diluted in PBS (pH 7 .2) alone, and 0 .1 ml inoculated in quadruplicate onto the CAM of the chick embryo (Beveridge and Burnet, 1946 ; Hawkes, 1979) .
To whom requests for reprints should be addressed . 0042-6822/89 $3 .00 Copyright rd 1989 by Academic Press . m a . All rights of reproduction in any form reserved.
262
INHIBITION OF INFLAMMATION BY CPV
Enzymes and chemicals Restriction enzymes were supplied by Bethesda Research Laboratories, Inc . (Gaithersburg, MD), New England BioLabs, Inc . (Beverly, MA), or Boehringer-Mannheim Biochemicals (Indianapolis, IN) and used as specified by the manufacturers . The Klenow fragment of DNA polymerase I, T, DNA ligase, T, DNA kinase, 5bromo-4-chloro-3-indolyl-fl-D-galactopyranoside (XGal), and isopropyl-,6-D-thiogalactopyranoside (IPTG) were obtained from Bethesda Research Laboratories, Inc . Alkaline phosphatase from calf intestine was purchased from Boehringer-Mannheim Biochemicals . Virus infectivity assays Virus pocks were excised from the CAM as a 4-mm' piece of tissue and added to 1 ml of PBS containing 0 .1% BSA in 15-m1 conical tubes and stored at -20° . Virus was released from the tissue by grinding in a tissue grinder (Bellco, Inc ., Vineland, NJ) and dispersed by two 30-sec periods of sonication in a chilled-cup attachment to a Model W-385 sonicator (Heat SystemsUltrasonics, Inc ., Farmingdale, NY) . Virus infectivity was estimated as described previously (Buller and Wallace, 1985) . Virus plaques were visualized after 2-3 days by the addition of 0 .5 ml 0 .3% crystal violet formalin solution to each well (Nakano, 1979) . Virus infection of cells Cultures containing 1 x 10 6 BS-C-1 cells were infected with viruses at a multiplicity of infection (m .o .i .) of 0 .01 plaque-forming unit (PFU) in 0 .5 ml of MEM for 1 hr at 37° . The virus inocula were removed and the cultures were then washed twice with prewarmed MEM containing 10% BCS and the cells were fed with MEM containing 10% BCS . At the indicated time points, virus-infected cells were scraped into the culture supernatant, which was transferred to a 15-ml conical test tube, and stored at -20° . Prior to estimation of virus infectivity, the samples were frozen and thawed three times, followed by two 30-sec periods of sonication . DNA hybridization analyses of virus genomic DNA DNA was isolated from purified virus or virus-infected cells by SDS/protease K digestion followed by phenol extraction (Garon et al„ 1978) . The isolated DNA was cleaved with restriction enzymes, and the DNA fragments were resolved by electrophoresis on a 0 .6 or a 1 % agarose gel in a Tris-borate-EDTA buffer system . DNA fragments were transferred from the agarose gel to nitrocellulose membranes by a modification of the procedure of Southern (1975) . DNAs labeled by
26 3
the random primer method were hybridized to the immobilized DNA as described previously (Moss et al ., 1981 ; Feinberg and Vogelstein, 1983) . Autoradiographs were made by placing the washed nitrocellulose sheets in contact with X-ray film for 1-4 days at -70° . Construction of a deletion in the 38-kDa gene The 5 .2-kbp EcoRl restriction endonuclease fragment of CPV DNA containing the 38K gene and its flanking regions was ligated into EcoRl restriction endonuclease-digested and alkaline phosphatasetreated p1247 (a derivative of pUC19 lacking all restriction endonuclease enzyme sites in the polylinker region except EcoRl and HindIII) . Competent Escherichia coli JM105 cells were transformed, and ii-Gal colonies (white versus O-Gal' blue) were isolated by using the XGal-IPTG screening system . A miniprep from a selected colony (pGP1 247 .2) was digested with Pstl, and the large fragment was purified away from the 688-bp Pstl fragment (see Fig . 1A) . The large fragment was treated sequentially with T, DNA polymerase (under conditions optimal for the formation of flush-ended termini) and alkaline phosphatase . This dephosphorylated, blunt-ended vehicle was ligated with a 4 .6-kbp selection cassette . The cassette was constructed first by inserting the 3 .1-kbp Xbal/Smal fragment from pSC10, which contained the vaccinia virus late 11 K promotor coupled to the E. coli lacZ gene, into Hincll- and Xbal-digested pUC19 (Chakrabarti et al., 1985) . In addition, into the same plasmid was cloned an appropriately modified 1 .5-kbp HindllI and Xmal fragment from Tn5, that contained the entire gene encoding resistance to neomycin/kanamycin and which was under the control of the vaccinia virus early/late 7 .5 K promotor (Southern and Berg, 1982 ; Franke et al., 1985) . This cassette was further modified to allow excision by either Smal or Nod . The ligation reaction was used to transform competent E. coli JM105 cells, and blue colonies (Q-Gal') which grew in the presence of kanamycin were isolated . Transfection of plasmid and isolation of recombinant viruses A 35-mm-diameter petri dish of CV-1 cells was infected with virus at a m .o .i . of approximately 0 .05 PFU per cell followed by transfection with calcium phosphate-precipitated plasmid (Stow et al., 1978), and the resultant progeny virus was screened as described previously (Chakrabarti et al., 1985) . The frequency of recombination of the target sequence into the virus genome was between 0 .1 and 0 .01 % .
26 4
PALUMBO ET AL . A 38 kDa E
p
p1247
P
D
ECO 111
I T4 DNA Llgaaa
B TRANSFECTION
RECOMBINANT -
GENOTYPE
Intact ORF
Virus
Plasmid
CPV-Br
NA
-P'
NA
CPV-Br
pGPI .7
-p .
CPV-BR .D1
Deletion
+
Cassette
CPV-Br
pGPI .7
-p .
CPV-BR .D2
Deletion
+
Cassette
CPV-Br .D1
PGP1247.2-P'
CPV-BR.D1 .R2
Intact ORF
CPV-Br .D1
PGP1247.2-r
CPV-BR.D1 .R4
Intact ORF
FIG. 1 . Deletion of the DNA sequences within the 38K gene of CPV . (A) Methodology employed to construct two plasmids : one that contained the entire 38K ORF, and the other which replaced approximately 50% of the coding region of the 38K gene with a #-Galrrn5 neo selection markercassette . (B) Strategyfor the production of recombinants of CPV-BR which contained only a deletion in the 38K gene, and the subsequent reversion to an intact 38K ORF .
Histopathology CAMs were excised from the egg and placed into 10 ml of PBS in a 100-mm-diameter plastic dish . The CAMs were washed twice with PBS prior to fixation with Tellyesnickie's fixative (738 ml ethanol, 50 ml glacial acetic acid, 100 ml 37% formaldehyde solution to 262 ml with distilled water) . After 1 hr at room temperature or 8 hr at 4°, the fixative was aspirated from the plate and the CAMs were cut into 2 x 10-mm strips which were sectioned at 5 µm and stained with hematoxylin and eosin . Other sections were stained for fibrin (Humanson, 1972) . Immunocytochemistry A New Zealand White rabbit was bled and serum (preimmune) was prepared . The same rabbit was then injected intradermally at two sites with 0 .1 ml of 0 .01 M Tris, pH 8 .0, containing 2 X 10 9 PFU of gradient purified vaccinia virus strain WR (Joklik, 1962). Four weeks later, the rabbit was boosted by subcutaneous injec-
tions at 10 sites with 0 .1 ml of complete Freund's adjuvant containing 1 x 10 9 PFU of virus . Twelve weeks later, the rabbit was again boosted by subcutaneous injections at 10 sites with 0 .1 ml of incomplete Freund's adjuvant containing 1 x 10 9 PFU of virus . The rabbit was bled and serum prepared (anti-virus) 2 weeks after this last immunization . A horseradish peroxidase developing system was used to detect antigen-antibody complexes . All tissue sections were treated with 0 .3% hydrogen peroxide in water for 30 min to reduce the endogenous peroxidase activity prior to incubation with a 1 :30,000 dilution of preimmune or immune sera . The primary antibody binding was detected with a Vectastain Elite ABC kit with an avidin/biotin/peroxidase complex developing system, following the manufacturer's instructions (Vector Laboratories, Burlingame, CA) . Slides were counterstained using Gill's hematoxylin . The sensitivity of the assay permitted the detection of even a single antigen-positive cell at 24 hr p .i . in a tissue section of CAM (McIntyre, unpublished results) .
INHIBITION OF INFLAMMATION BY CPV
X
C
C
CK
CX
r 1 Kb
SP41
265
R, BS P K C K E . . . . . „ .
I
1 ITA
FIG . 2 . Structure of the 38K gene in CPV- BR, mutant, and revertant viruses . Digestion of viral DNA with the restriction endonucleases Pstl (A), Clal (Band D), and EcoRl (C) . Viral DNAs were isolated from BS-C-1 cells infected with CPV-BR, the two mutants CPV-BR .D1 or CPV-BR .D2 that
haves deletion of the 38K gene, or either of the two revertants CPV-BR .D1 .R2 and CPV-BR .D1 R4 that have an intact 38K ORF . A 32P-labeled plasmid containing the 5 .2-kbp EcoRl fragment (heavy black line (F)) of CPV-BR that contained the entire coding sequence of the 38K gene was used as the probe in (A), (B), and (C) . A 32P-labeled oligonucleotide specific for the Tn5 neo gene was used as the probe in (D) . (E) Restriction endonuclease map of the right-hand end of CPV-BR ; BamHl (B), Clal (C), EcoRl (E), Hindlll ( H), Kpnl (K), Pall (P), Sall (S), Xhol (X), serine protease inhibitor-1 (SPI-1), and inverted terminal repeat (ITR) .
RESULTS Construction of a deletion in the gene encoding the 38-kDa protein Two different plasmids were constructed for use in the production of recombinants of CPV-BR . The 5 .2kbp EcoRl fragment that encompassed the entire open reading frame (ORF) of the 38K gene was inserted into p1247 (Fig . 1 A ; Materials and Methods) . This plasmid, pGP1 247 .2, contained two Pstl restriction endonuclease sites which bound a 688-bp sequence (Pstl-R fragment) within the OFF of the 38K gene . Digestion of pGP1 247 .2 with the restriction endonuclease Pstl and the replacement of the Pstl-R fragment with the p-Gal/ Tn5 neo cassette resulted in the isolation of pGP1 .7 .
The DNA structure of pGP 1 .7 was confirmed by EcoRl, Pstl, and Clal restriction endonuclease digestions and dideoxy sequencing from the termini of the cassette into the gene (Palumbo, unpublished results ; Hattori and Sakaki, 1986) . The plasmid GP1 .7 was used to make two individual recombinants of the Brighton strain of CPV, CPVBR .D1, and CPV-BR .D2, which had deletions of approximately half of the 38K gene, and insertions of the /3-Gal/Tn5 neo cassette (Fig . 18) . Recombinant CPVBR .D1 and pGP1 247 .2 (intact 38K ORF) were used to construct and to isolate independently two revertant viruses, CPV-BR .D1 .R2 and CPV-BR .D1 .R4, which had regained an intact 38K ORF . The CPV-BR .D1 virus replicated in BS-C-1 cells in a
2 66
PALUMBO ET AL.
INHIBITION OF INFLAMMATION BY CPV
manner indistinguishable from CPV-BR . At 40 hr p .i ., the geometric mean plaque diameter of CPV-BR .D1 was 0 .62 X 1 .05 mm compared to the CPV-BR value of 0 .64 X 1 .04 mm (Student's t test, P > 0 .05 ; 20 of) . Similarly . BS-C-1 cell cultures, infected with CPV-BR and CPV-BR .D1 viruses and harvested at 10, 24, and 48 hr p .i ., gave total culture virus infectivity values at each time point which were not significantly different from one another . Southern blot analysis of the genomes of the recombinant viruses To confirm the predicted DNA genome structure of the isolated recombinant viruses, Southern blot DNA analyses were performed on total DNA from virus-infected cells (Fig . 2) . For each virus, restriction endonuclease-generated DNA fragments were separated by agarose gel electrophoresis, transferred to nitrocellulose, and hybridized to 32 P-labeled DNA . The hybridized DNAwas visualized through autoradiography . The sizes of the restriction endonuclease fragments generated for both CPV-BR and the recombinant viruses were compared to the sizes of the restriction fragments expected upon successful generation of the desired recombinant CPV . A 32 P-labeled 5 .2-kbp EcoRl fragment (Fig . 2E ; probe) detected the 688-bp Pstl-R fragment (indicated as 0 .7 kbp in Fig . 2A) present in CPV-BR and the revertant virus DNAs, but this fragment was absent in the DNAs from mutant viruses CPV-BR .D1 and CPVBR .D2 . Analysis of the Clal restriction endonucleasedigested viral DNA confirmed the insertion of the cassette into the 38K open reading frame (Fig . 2B) . Of the five Clal restriction endonuclease fragments (Fig . 2E)
267
detected in CPV-BR and revertant virus DNAs (Fig . 2B), three were also present in the mutant virus DNAs, as were two Clal fragments of 2 .4 and 4 .5 kbp . The latter fragment was created by the fusion of DNA sequences present in the S-Gal/Tn5 neo cassette (containing a single Clal site approximately 2 .2 kbp from the end of the /acZ gene) with the 2 .2-kbp DNA fragment (deleted 2 .4kbp WT fragment) from the right-hand side of the 38K gene (Fig . 2E) . The smaller 2 .4-kbp Clal fragment was formed by fusion of 218 nucleotides derived from the left-hand end of the 38K ORF and approximately 2 .2 kbp of sequence from the /3-Gal/Tn5 neo cassette . In addition, the same 5 .2-kbp EcoRl fragment that contained the intact 38K gene was present in CPV-BR and revertant virus DNAs, while new restriction fragments of predicted size were detected in the viruses with a deletion in the 38K gene (Fig . 2C ; f3-Gal/Tn5 neo cassette contains one EcoRl site 3 .1 kbp from the end of the /acZ gene) . Additional Southern blots in which recombinant virus DNAs were digested with Pstl, EcoRl, Clal, and Hindlll restriction endonucleases and probed with 22 P-labeled total CPV-BR DNA showed no additional changes in DNA structure of the deletion mutants (data not shown) . Deletion of the cassette from the revertant viruses was confirmed by using an oligonucleotide probe to a Tn5 neo DNA sequence which hybridized only to the 4 .5-kbp Clal restriction endonuclease fusion fragment (Fig . 2D) . Thus, we have constructed and isolated recombinant viruses, two of which showed DNA structure alterations consistent with only a deletion/insertion into the 38K gene sequence, and a further two recombinants that appeared by DNA structure analysis to be identical to the wildtype CPV-BR .
FIG . 3 . CPV-BR and CPV-BR .D1 pock morphology on chick embryo CAMs 96 hr after virus inoculation . The CAMs of 12-day-old chicken embryos were inoculated with viruses and harvested at 96 hr p .i . The CAMs were washed several times with PBS and photographed with backand side-lighting . (A) Discrete, bright red pocks induced by CPV-BR scattered along the vascular tree ; (B) the greater magnification of the pock indicated by the arrow in (A) shows the network of congested blood vessels constituting the bulk of the pock ; (C) white pocks induced by CPVBR .D1 were mostly solid while, but the ones indicated by arrowheads contain the red center commonly seen in earlier pocks induced by CPVBR .D1 ; (D) a greater magnification of the pock indicated by the arrow in (C) showing the raised white center and discrete border . The contrast with the red pock (B) is evident [line : 1 .4 cm (A and 10 .22 cm (B and D)] . Fio . 5. The distribution of virus-specific antigen within CPV-BR and CPV-BR .D1 pocks on chicken embryo CAMs at various times after infection . (A) CPV-BR-induced pock at 24 hr p .i . showing a dense concentration of virus antigen in the central area and ectodermal proliferation (arrows) . Note the lack of both hemorrhage and cellular infiltration within the mesoderm . (B) CPV-BR .D1-induced pock at 24 hr p .i . showing viral antigen and ectodermal proliferation (arrows) similar to the CPV-BR pock . At this time, the mesodermal vessel remains patent, and there is minimal adhesion of inflammatory cells to the endothelium . (C) CPV-BR pock at 48 hr p .i . with extensive vacuolahon (v) and the presence of large amounts of viral antigen not only within the expanded ectoderm, but also in fibroblasts in the mesoderm (arrowheads) . Note the general lack of extravasation of erythrocytes from engorged blood vessels (C, arrows) . This is shown in greater detail in a higher magnification of the boxed region in (C) which is presented in (E) . Note the intact endothelium of the blood vessels (arrowheads), although there are some erythrocytes present in the mesoderm (arrows) . (D) CPV-BR .D1 pock at 48 hr p .i . showing the extensive infiltration of the mesoderm with inflammatory cells which have isolated antigen containing areas (arrows) that probably represent scattered remnants of the original ectoderm . At lower right is the endoderm, which shows proliferative changes . A scab overlies the ectoderm constituting the original pock . (F) Enlargement of the area boxed in (D) . Antigen was visualized by immunological staining with a horseradish peroxidase developing system followed by hematoxylin counterstaining (see Materials and Methods) [line : 6 .3 µm (A and B) ; 13 .4 gm (C and O) ; 4 .2 g m (F) ; and 6 .3 pro (F)] .
268
PALUMBO ET AL .
Fic . 4 . Histochemical examination of CPV-BR and CPV-BR .D1 virus pocks from chick embryo CAMs . CAM of 12-day-old embryo at 721 (A) CPV-BR-induced pock . The ectodermal layer is thickened and is beginning to undergo lytic changes and vesicle formation (v) assoc with virus replication, as evidenced by numerous intracytoplasmic inclusion bodies (white arrows) . Note the clear line of demarcation maint between ectoderm and mesoderm (black arrows in opposition) . Most of the cells within the mesoderm are erythrocytes (arrowheads) . (B) induced with CPV-BR .D1 showing the massive inflammatory cell influx into the mesoderm, central erosion of the ectoderm, and scab form
INHIBITION OF INFLAMMATION BY CPV
A
PFU/Pock at 96Hr Post Infection (1X10 -5 ) 0
10
15
20
CPV-Br (wt) CPV-Br .DI .R2 CPV-Br .D1 .R4 CPV-Br .D1 CPV-Br .02
B
PFU/Pock (1x10 -4 ) 0 1 2 3 4 5 8 7 8 9
10
48 hr
72 hr
96 hr
FIG . 6 . Virus infectivity recovered from pocks isolated from CAMs of chicken embryos . (A) Comparison of virus recovered from pocks induced by infection of CAMs with CPV-BR, mutant, or revertant viruses . (B) Virus infectivity recovered at the indicated time points from pocks induced by either CPV-BR (.) or CPV-BR .D1 (D) . Individual pocks were excised from the CAM, and infectivity was quantitated as described under Materials and Methods . Arithmetic means and one SEM were calculated from a minimum of tour pocks from an individual egg . Previous studies have indicated that there was considerable variation in the virus infectivity yields among eggs inoculated with the same dose of virus (Butler et at, 1988) . In all cases, of the four replicate eggs inoculated with virus, the egg which gave virus pocks with the maximal virus infectivity was used to calculate the means reported .
Macroscopic character of the wild-type and 38K deletion mutant pocks To determine the effect of a deletion in the 38K gene on the CPV-BR red-pock phenotype, the CAMs of 12day-old chick embryos were infected with CPV-BR or
269
the mutant viruses . The CPV-BR pock appeared macroscopically by 24 hr as a slightly raised, translucent thickening of the CAM but, by 48 hr, blood vessels became dilated and congested, giving it a more pronounced red color. Pocks became larger so that, by 96 hr, they had reached a diameter of 2-3 mm (Fig . 3A) . The dilated and congested nature of the blood vessels was shown clearly in the higher magnification of the pock indicated in Fig . 3B . In certain experiments, by 96 hr p .i . some pocks formed hematocysts . The two revertant viruses, CPV-BR .D1 .R2 and CPV-BR .D1 .R4, were examined in a similar fashion and yielded a pock morphology which was indistinguishable from CPV-BR at all time points examined . The 48-hr pocks induced by CPV-BR .D1 differed from CPV-BR, as they were raised and opaque, with small red centers . By 72 hr p .i ., the pocks had reached maximal sizes of 2-3 mm, and the red centers were less distinct . At 96 hr p .i ., CPV-BR .D1 pocks appeared dull and had lost their sheen (Fig . 3C) . A higher magnification of a typical pock showed the center of the pock to be raised and surrounded by a white region within the CAM tissue (Fig . 3D) . Identical results were obtained with CPV-BR .D2 and a deletion mutant, CPVBR .DC1, which lacked the /3-Gal/Tn5 neo cassette . The development and character of CPV-BR .D1, CPVBR .D2, and CPV-BR .DC1 pocks were similar to that described for a naturally occurring white-pock variant, W2, isolated by Pickup and co-workers (1984) . Histology of CPV-BR, mutant, and revertant pocks The difference between the pocks of CPV-BR and the CPV-BR .D series of mutants was even more apparent on histological examination of CAM tissues at 72 hr p .i . (Fig . 4) . The CPV-BR pocks were characterized by extensive ectodermal proliferation at the advancing margin of the pock (Bulleretal ., 1988), while at thecenter, there was erosion of the ectoderm layer with resultant vesicle formation and an abundance of inclusion bodies (Fig . 4A) . A sharp line of demarcation remained between the ectoderm and the mesoderm . The blood vessels were dilated and congested with certain pocks showing minimal extravasation of blood cells into the mesoderm, but other pocks showed severe hemorrhage accompanied by fibrin deposition . A consistent finding was the total lack of an inflammatory response .
associated with this mutant . The scab is outlined by arrowheads . The endoderm remains relatively normal . (C) Higher magnification of the boxed area shown in (B) . Note the lack of well-demarcated inclusion bodies (possible inclusions are indicated by white arrows) in the relatively intact ectoderm . The junction between the ectoderm and mesoderm is indistinct . Most cells within the mesoderm appear to be macrophages (black arrow) ; within the scab, heterophils predominate . (D) A healed pock induced with CPV-BR .D1 with regrowth of the ectoderm (arrowheads) capped by a small scab . Macrophage infiltration of the dermis is reduced compared to early pocks . All sections were stained with hematoxylin and eosin [line : 4 .2 pro (A); 13 .4 µm (B) ; 4 .2 µm (C); and 13 .4 µm (D)] .
270
PALUMBO ET AL .
The two revertant viruses induced pocks with a morphology identical to that of CPV-BR . In contrast, the pocks induced by the mutant CPVBR .D1 were dominated by a massive influx of inflammatory cells composed of macrophages and heterophils which resulted in destruction of the integrity of the ectodermal layer and the formation of a scab (Fig . 4B) . A higher magnification of this region is presented in Fig . 4C . Within a thickened mesoderm, there was moderate proliferation of fibroblasts and accumulations of macrophages . Blood vessels within the mesoderm were moderately dilated and contained thrombi . As with the wild-type virus, fibrin deposits were detected in pocks associated with hemorrhage . As indicated in Fig . 4D, pocks by 72 hr p .i . were observed in an advance stage of ''healing," as indicated by the extensive regrowth of ectodermal cells and the presence of diminished numbers of inflammatory cells, of which macrophages were more prominent than heterophils . In some of these pocks, virus antigen could be seen sequestered to the outer surface of the regenerated ectoderm layer. CPV-B R . D2-infected pocks yielded findings identical to CPV-BR .D1 . These findings illustrated the fundamental difference between the CPV-BR and mutantvirus infections with regard to induction of the host inflammatory response . Distribution of virus antigen within tissue of the pock To determine if CPV-BR and CPV-BR .D1 viruses differed in their replication and spread within the ectoderm, and thereby caused dissimilar levels of cellular necrosis, virus pocks were analyzed at 24 and 48 hr p .i . for antigen distribution . Both CPV-BR and CPV-BR .D1 pocks at 24 hr p .i . appeared very similar with minimal perturbations of the ectodermal layer, no involvement of the mesoderm, and a lack of inflammatory cell infiltrates (Figs . 5A and 5B) . A similar distribution of antigen-positive cells was detected in both infections . Ectodermal hyperplasia was associated with the advancing margin of the pock . The similarity between the CPV-BR and CPV-BR .D1 24-hr pocks argued against tissue necrosis as the sole explanation for the differences in the virus-induced inflammatory responses observed at 72 hr p .i . in Fig . 4 . By 48 hr p .i ., clear differences in virus antigen staining were detected between CPV-BR and CPV-BR .D1 . In comparison to CPV-BR .D1, CPV-BR pocks showed more intense staining, indicative of the presence of more virus antigen within infected cells, and a greater number of antigen-positive cells (compare Fig . 5C with 5D) . The pock in Fig . 5C clearly depicted dilation and congestion of the blood vessels beneath and within the ectoderm, and the lack of inflammatory cells within the
mesoderm . A higher magnification of this region showed the endothelial cells lining the blood vessels to be intact with a minimal extravasation of erythrocytes into the mesoderm (Fig . 5E, see also Fig . 4A) . In CPVBR virus pocks, antigen-positive fibroblasts could be detected in the mesoderm at a great distance from the center of infection (Fig . 5C, arrowheads) . On occasion, endothelial cells lining the arterioles and venues also stained positively for virus antigen . In contrast, the distribution of antigen-positive cells in CPV-BR .D1 pocks appeared more restricted and was fragmented by the inflammatory infiltrate (Figs . 5D and 5F) . It is uncertain whether these antigen-positive cells are producing virus progeny or are phagocytes containing virus antigenic debris . As noted previously in Fig . 4 and seen again in Fig . 5D, the demarcation between the ectoderm and mesoderm was obliterated by the magnitude of the inflammatory influx into the pocks induced by the 38K deletion mutants . Virus infectivity recovered from pocks To determine if the observed reduction in detectable antigen in pocks of CPV-BR .D1 compared to CPV-BR was paralleled by similar reductions in virus progeny, individual pocks from CPV-BR- and CPV-BR .D1-infected CAMs were analyzed for virus infectivity . The average infectivity yields of pocks isolated at 96 hr p .i . from CAMs infected with CPV-BR, the two independently isolated mutants lacking an intact 38K gene, and two independently isolated revertants of CPV-BR .D1 are depicted in Fig . 6A . Both mutants showed reductions in average yields of virus recovered from pocks in comparison to CPV-BR (P < 0 .05), 2 while the two 38K gene revertants of CPV-BR .D1 gave pock infectivity yields similar to those for CPV-BR pocks . To investigate whether the differences between virus pock infectivity yields varied with time p .i ., CPV-BR-and CPV-BR .D1infected CAMs were isolated at 48, 72, and 96 hr p .i ., and individual virus pocks were analyzed forvirus infectivity yields . Less virus infectivity was recovered from pocks induced by the mutant, CPV-BR .D1, in comparison to CPV-BR at all times examined (Fig . 513) (P < 0 .05) . 2 These findings are consistent with the diminished virus antigen detected in mutant pocks . DISCUSSION Variants of CPV which are identified by a white pock morphology induced on the CAM occur regularly with2 Data for each virus was subjected to sing le-factor analysis of variance (ANOV) (Zar, 1974) . Upon rejection of the null hypothesis that the mean virus yields were equivalent by ANOV, Newman-Keuls multiple comparison analysis of means was performed (a = 0 .05) (Zar, 1974).
INHIBITION OF INFLAMMATION BY CPV
out obvious selection for a genetic determinant (Downie and Haddock, 1952 ; van Tongeren, 1952) . These white-pock variants are associated with DNA rearrangements which typically involved the replacement of up to 39 kbp of the right-hand end of the genome, with an inverted copy of up to 50 kbp from the left-hand end of the genome (Archard et al., 1984 ; Pickup et al., 1984) . This type of rearrangement results in the loss of at least 29 kbp of mainly unique sequences from the right-hand end of the virus genome, which makes it difficult to relate alterations in pock morphology with the loss of a specific gene . Using a battery of independently isolated white-pock variants and marker-rescue techniques, Pickup and colleagues showed that a 38K gene which mapped between 31 and 32 kbp from the right-hand end of the genome was a major component in the determination of the white-pock phenotype ; however, recombination of this 38K gene into the TK gene of white-pock variants did not result in pocks as red as CPV-BR wild type (Pickup et al., 1984, 1986) . These experiments suggested the presence of additional genes which influenced pock phenotype and were encoded in the DNA sequences of the right-hand end of the CPV-BR virus genome . A candidate is the CPV homolog to the serine protease inhibitor-1 (SPI-1) gene of vaccinia virus that is located in this region of both the vaccinia and CPV genomes (see Fig . 2E) and has significant homology at the amino acid level with the 38-kDa protein (Kotwal and Moss, 1989 ; Palumbo and Buller, unpublished data) . Further evidence for the presence of additional genes involved in CPV-BR-pock phenotype came from the work of Chernos at al . (1985), who found that selected recombinants between vaccinia and ectromelia viruses gave red pocks on the CAM, whereas the parental viruses gave only white pocks . Direct evidence is presented in this report that inactivation of the 38K gene alone is sufficient to change the wild-type virus red pock phenotype to white . The whitepock phenotype was specifically the result of a deletion in the 38K gene, and not a second mutation elsewhere in the genome, since recombination between a 38K deletion mutant and a plasmid containing an intact 38K gene resulted in the isolation of revertant viruses which were indistinguishable from wild-type CPV-BR, These experiments strongly suggested that the white-pock phenotype was either directly or indirectly the consequence of the loss of the 38-kDa protein and that this protein is normally required to produce the red pock phenotype, but it does not rule out the possibility that other genes are also involved . To address this last possibility, experiments are in progress to delete the SPI-1 gene of CPV-BR .
27 1
A number of the differences in the pocks that form on a CAM between CPV-BR and 38K gene deletion mutants observed in this study were similar to those previously reported using naturally occurring white-pock deletion mutants : (1) The CPV-BR gave a red pock which was characterized by hemorrhage at 72 to 96 hr p . i . ; (2) the raised, white, and opaque phenotype of the white-pock variants was correlated with an influx of inflammatory cells into the pock ; and (3) the pocks of the white-pock variants, on average, had less infectivity than comparable CPV-BR pocks (Downie, 1939 ; Downie and Haddock, 1952 ; van Tongeren, 1952 ; Fenner, 1958 ; Baxby, 1969) . This study extended our understanding of CPV-BR virus pock development in the CAM of the chick embryo, delineated more clearly the differences between CPV-BR and white-pock variant viruses, and suggested a possible role for the 38-kDa protein . Within the first 24 hr after inoculation of the CAM with virus, sites of viral infection could not be detected reliably by microscopic observations of stained tissue sections . The early events in virus infection of the CAM were, however, revealed using immunocytochemistry and were found to be similar for both viruses, with no detectable perturbations of either the ectoderm or mesoderm . In contrast, by 48 hr p .i ., there were gross differences in pock color and form . Microscopic observations described above indicate that the red color of the CPV-BR pock is best explained by dilation and congestion of the dense intra- and epiectodermal capillary beds in the CAM which provide the respiratory function of the chick embryo (Ramanoff, 1960) . This was clearly shown in Figs . 5C and 5E where the individual blood vessels underlying the ectoderm were dilated to such an extent as to appear as one continuous vessel . Since the mutant also showed congestion and dilation of blood vessels, the presence of a red color only in the center of the mutant pocks can be explained by the screening effect of inflammatory cells migrating into the mesoderm and ectoderm . The reduction of virus infectivity and in the number of antigen-positive cells in pocks induced by the deletion mutant by 48 hr p .i . in comparison to CPV-BR can be explained as a direct result of the inhibition of virus spread by infiltrating inflammatory cells . Thus, these results suggest that CPV-BR evades the inflammatory response-a nonspecific host defense mechanism-by production of the 38-kDa protein . The simplest hypothesis would predict that the 38kDa protein expressed in the wild type, but not the mutant virus infection, must hinder the generation/action of a chemotactic substances) from within the pock which is directly or indirectly responsible for the migra-
272
PALUMBO ET AL .
tion into the lesion of first heterophils and then macrophages . In light of the complicated nature of the inflammatory response, we can only speculate as to the target of the 38-kDa protein . One parameter that can be eliminated is any contribution by the immune system, since mature T and B cells are not yet present in the blood of the chick embryo at 12-15 days of incubation (Eerola et al., 1987 ; Chen and Cooper, 1987 ; Chen et al., 1986, 1988) . Further, in the absence of IgE, it is unlikely that basophils or tissue mast cells play a role . In addition, the classical complement pathway is inoperative, since anti-virus antibody is lacking, and previous investigators have shown that the 12- to 15-day-old chick embryos have total C and C1 levels of 1-4% of the adult chicken (Gabrielsen et al., 1973) . On the other hand, during inflammation three mediator-producing systems-the kinin, the coagulation, and the fibrinolyticas well as the catabolic pathways of arachidonic acid (Samuelsson et al., 1979) can become activated and these would be prime candidates as targets of the 38kDa protein . Computer-assisted searches of the protein database of the National Biomedical Research Foundation have shown that the 38-kDa protein has significant sequence similarities to several inhibitors of serine proteases . For example, it is similar to both human antithrombin III and placental plasminogen activator inhibitor (having optimized match scores of 366 and 479, respectively, out of a possible score of 1679) . The structural similarity between these inhibitors and the 38-kDa protein suggested that the viral protein might induce red-pock formation by inhibiting serine proteases involved in either the blood coagulation pathway or the normal processes of wound-containment and healing . The results of these studies on the development of pocks show that the primary effect of the 38kDa protein is to repress the massive infiltration of inflammatory cells which normally occurs within 48 hr of infection . At this stage of the infection, the redness of the pocks produced by the wild-type virus is the result of both the lack of infiltration by inflammatory cells and the dilation and congestion of the blood vessels, Only at later stages of infection (72-96 hr), when there is greater tissue damage, do pocks begin to show evidence of extensive hemorrhage . Accordingly, the characteristic hemorrhage in pocks produced by the wildtype virus may be the consequence of the lack of an appropriate inflammatory response, rather than a direct inhibition of the enzymes involved in the blood coagulation pathway. Such a mechanism is consistent with the observations that, under certain conditions, white-pock variants of CPV can generate red pocks
(Baxby, 1969) . Such atypical pocks may be formed in the embryo if it is incubated at suboptimal temperatures, or if it is infected with more than 100 pock-forming units of virus ; under these conditions, the embryo's normal inflammatory response may be impaired or overwhelmed . To date, placental plasminogen activator inhibitor has yielded the most significant level of relatedness to the 38-kDa protein . This protein inhibits the conversion of plasminogen to plasmin . Plasmin may generate chemotactic substances either by direct action on complement component C5 to produce a C5a-like molecule (Larsen and Henson, 1985) or by dissolving a fibrin clot (Belew et al., 1978) . C5a has been shown to be a potent chemoattractant for neutrophils (see review, Goldstein, 1988) . Other proteases, including trypsin and kallikrein, have also been implicated in the generation of chemotactic fragments from C5 (Larsen and Henson, 1983), so it is conceivable that serine proteases directly or indirectly released as a consequence of virus infection of ectodermal cells could directly act on C5 . The 38-kDa protein would specifically inhibit some step in this process . Proper evaluation of a C5amediated inflammatory response induced by direct action of proteases on C5 must await measurements of the levels of C5 in the chick embryo and in vitro studies . ACKNOWLEDGMENTS We thank Mr . C . Duarte and Dr . D . Ailing for expert technical and statistical assistance, respectively ; Drs . B . Moss, M . Boyle, S . Isaacs, G . Kotwal, and J . Sechler for helpful discussions ; and Ms . B . R. Marshall for accurate and efficient manuscript production .
REFERENCES ARcHARO, L . C ., MACKETn, M ., BARNES, D . E ., and DuMBELL, K . R . (1984) . The genome structure of cowpox virus white pock variants . J . Gen . Wool. 65, 875-886 . BA
INHIBITION OF INFLAMMATION BY CPV
sual screening of recombinant virus plaques . Mol . Cell. Biol. 5, 3403-3409 . CHEN, C .-L . H ., AcER, L . L ., GARTLAND, G . L ., and COOPER, M . D . (1986) . Identification of a T3/T cell receptor complex in chickens . J.Exp . Med. 164, 375 -380 . CHEN, C .-L . H ., CIHAK, J ., LoSCH, U ., and COOPER, M . D . (1988) . Differential expression of two T cell receptors, TcR1 and TcR2, on chicken lymphocytes . Fur. J. Immunol. 18,539-543 . CHEN, C .-L . H ., and COOP=ER, M . D . (1987) . Identification of cell surface molecules on chicken lymphocytes with monoclonal antibodies . In ''Avian Immunology: Basis and Practice" (A . Toivanen, and P . Toivanen, Eds .) . CRC Press, Boca Raton, FL . CHERNOS, V . I ., ANTOtaovA, T . P ., and SENKEVICH, T . G . (1985) . Recombinants between vaccinia and ectromelia viruses bearing the specific pathogenicity markers of both parents . J Gen . Viral. 66, 621626 . DOWNIE, A . W . (1939) . A study of the lesions produced experimentally by cowpox virus . I Pethol. 48, 361-379 . DOWNIE, A . W ., and HADDOCK, D . W . (1952) . A variant of cowpox virus .Lancet 1,1049-1050 . EEROLA, E ., VEROMAA, T ., and TOIVANEN, P . (1987) . Special features in the structural organization of the avian lymphoid s ystem . in ''Avian Immunology : Basis and Practice" (A . Toivanen, and P . Toivanen, Eds .) . CRC Press, Boca Raton, FL . DE MERS, A . P ., and VOGELSTEIN, B . (1983) . A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity . Anal. Biochem . 132, 6-13 . FENNER, F- (1958)- The biological characters of several strains of vacclnla, cowpox and rabbitpox viruses . Virology 5, 502-529 . FRANKE, C . A . . RICE, C. M ., STRAUSS, J . H ., and HRUSY, D . E . (1985) . Neomycin resistance as a dominant selectable marker for selec tion and isolation of vaccinia virus recombinants . Mol. Cell. Riot. 5, 1918-1924 . GARRIEESEN . A . E ., PICKERING, R . J ., LINNA, T . 1 ., and GOOD, R . A . (1973)- Haemolysis in chicken serum . II . Ontogenetic development . lmmunology25, 179-184 . GARON, C . F ., BARBOSA, E ., and Moss, B . (1978) . Visualization of an inverted terminal repetition in vaccinia virus DNA . Proc. Nell . Acad . Sci . USA 75,4863-4867 . GOLOSTEIN, I . M . (1988) . Complement : Biologically active products . In 'Inflammation : Basic Principles and Clinical Correlates'' (1 . I . Gallin, I . M . Goldstein, and R . Snyderman, Eds .) . Raven Press, New York .
273
ettsial and Chlamydial Infections' (E . H . Lennette, and N . J . Schmidt, Eds .) . American Public Health Association, Inc ., Washington, DC . HUVANSON, G . L . (1972) . ''Animal Tissue Techniques'' 3rd ed ., p . 266 . Freeman, San Francisco, CA . JoKLIK, W . K . (1962) . The preparation and characteristics of highly purified radioactively labeled poxviruses . 8iochim . Biophys . Acta 61,290-301 . KowtAL, G . J ., and Moss, B . (1989) . Vaccinia virus encodes two proteins that are structurally related to members of the plasma serine protease inhibitor superfamily .I Viral 63, 600-606 . LARSEN, G . L ., and HENSON, P . M . (1983) . Mediators of inflammation . Annu . Rev. Immunol. 1, 335-359Moss, B ., WINTERS, E ., and COOPER, l . A . (1981) . Deletion of a 9,000base-pair segment of the vaccinia virus genome that encodes nonessential polypeplides . J.. ViroL 40, 387-395 . NAKANO, 1. H . (1979) . Poxyvrruses . In "Diagnostic Procedures for Viral, Rickettsial and Chlamydial Infections," 5th ed . (E . H . Lennette, and N . J . Schmidt, Eds .), pp . 293 296 . American Public Health Association, Washington, DC . PICKUP, D . J ., INK, 8 . 5 ., Hu, W ., RAY, C- A-, and JOKUK, W . K . (1986) . Hemorrhage in lesions caused by cowpox virus is induced by a viral protein that is related to plasma protein inhibitors of serine proteases . Proc. Natl . Acad. Sci. USA 83, 7698-7702 . PICKUP, D . J ., INK, B . S ., PARSONS, B . L, Hu, W and JOKLIK, W . K .
(1984) . Spontaneous deletions and duplications of sequences in the genome of cowpox virus . Proc . Natl. Acad . Sci. USA 81, 68176821 . RoMANOEE, A . L . (1960) . ''The Avian Embryo . Structural and Func tional Development ." MacMillan Co ., New YorkSAMUELSSON, B ., HAMMARSTROM, S ., and BORGEAI, P . (1979) . Pathways of arachidonic acid metabolism . In ' Advances in Inflammation Research" (G- Weissman, et al., Eds-) . Raven Press, New York SOUTHERN, E . M . (1975) . Detection of specific sequences among DNA fragments separated by gel electrophoresis . J. Mol. Biel . 98, 503-517 . SOUTHERN, P . J ., and BERG, P . (1982) . Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter . J Mol. Appl. Genet . 1, 327-341 . Stow, N . D ., SunAK-SHARPE, J . H ., and WILKIE, N . M . (1978) . Physical mapping of herpes simplex virus type 1 mutations by marker rescue .l. Viral. 28, 182-192 .
HA1TOR ., M ., and SASAKI, Y. (1986) . Dideoxy sequencing method using denatured plasmid templates . Anal . Biochem . 152, 232-238 .
VAN ToNOEREN . H . A. E . (1952) . Spontaneous mutation of cowpox virus by means of eggpassage . Arch . Gesamte Virusforsch . 5, 3552 .
HAwKES, R . A. (1979) . General principles underlying laboratory diagnosis of viral infections . In "Diagnostic Procedures for: Viral, Rick-
ZAR, J . H . (1974) . "Blostatistical Analysis-" Prentice-Hall, Inc ., Englewood Cliffs, NJ .
172, 262-273 (1989)
Inhibition of an Inflammatory Response Is Mediated by a 38-kDa Protein of Cowpox Virus GREGORY J . PALUMBO,*' 1 DAVID 1 . PICKUP,t TORGNY N . FREDRICKSON,t LAURENCE J . McINTYRE,§ AND R . MARK L . BULLER* *Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892 ; tDepartment of Microbiology and Immunology, Duke University Medical Center, Durham, North Carolina 27710; #Department of Pathobiology, University of Connecticut, Storrs, Connecticut 06268; and § Vector Laboratories, Inc ., Burlingame, California 94010 Received March 10, 1989, accepted May 3, 1989 The Brighton Red (BR) strain of cowpox virus induces a flat, bright red pock on the chorioallantoic membrane (CAM) of the 12-day-old chick embryo . In contrast, mutants with a deleted 38K gene (which is located 31 to 32 kb from the right-hand end of the virus genome) induced a raised, white, and opaque pock . During the first 24-hr p .i ., both CPV-BR and the 38K deletion mutants replicated similarly in the CAM of the chick embryo, as indicated by immunocytochemical detection of similar amounts of virus antigen . By 48 hr p.i ., the pocks induced by the mutant and CPV-BR are strikingly different . The pocks induced by the 38K deletion mutants were infiltrated by large numbers of heterophils and macrophages, which correlated with a reduction in the levels of virus antigen and virus infectivity . The CPV-BR pock had an absence of inflammatory cells and increased levels of virus antigen and infectivity . By 72 hr p .i., many of the pocks induced by the mutant were undergoing resolution of the virus infection, as indicated by further decrease of virus antigen and visible signs of healing, whereas CPV-BR pocks continued to be a site of active viral replication . These data are consistent with a model where this 38-kDa protein directly or indirectly inhibits the generation of chemotactic molecules which are elicited during virus replication in the CAM or, alternatively, blocks the interaction of these molecules with cells of the host inflammatory response . n 1999 Academic Press, Inc .
INTRODUCTION
several plasma proteins that are inhibitors of serine proteases ; however, the recombination of this gene alone into the genome of white-pock deletion mutants did not result in the complete restoration of the wildtype, red pock (Pickup et al., 1986) . This suggested that additional genes may contribute to the red pock phenotype . The objective of this study was to determine the 38K gene function in the pathobiology of virus infection of chick embryo CAM in general, and to pock morphology in particular .
Cowpox virus produces two distinct types of pocks when inoculated onto the chorioallantoic membrane (CAM) of the chick embryo : one is flat, well demarcated, and bright red, while the other is raised, opaque, and white in character (Downie, 1939 ; Downie and Haddock, 1952 ; van Tongeren, 1952) . White pocks were detected at a frequency of about 1%, and on repeated passage bred true (Downie and Haddock, 1952 ; van Tongeren, 1952 ; Fenner, 1958) . The viruses that produced these white pocks appeared to be less virulent than the wild-type virus in inoculations of rabbits, mice, guinea pigs, nonhuman primates, and cows (van Tongeren, 1952 ; Fenner, 1958) . Studies of the genomes of these white-pock variants determined that this pock phenotype was associated with the loss of at least 32 kbp of DNA from the righthand end of the genome (Archard et al., 1984 ; Pickup etal., 1984) . Further studies by Pickup and colleagues (1986) identified the location of at least one major gene (designated 38K) involved in the determination of the white-pock phenotype, This gene was located between 31 and 32 kbp from the right-hand end of the CPV-BR genome, and encoded a 38-kDa protein predicted to have amino acid sequence homology with
MATERIALS AND METHODS Viruses, cells, and eggs CPV strain Brighton Red (CPV-BR) was originally isolated by Downie (1939) . It was grown in, or plaqued on, BS-C-1 monolayers cultured in minimum essential medium (MEM, GIBCO Laboratories) containing 10% bovine calf serum (BCS) . Eleven-day-old embryonated hen's eggs were obtained from Truslow Farms, Inc . (Chestertown, MD) . The eggs were incubated at 3738 .5° in approximately 50% humidity without rocking, or exogenous CO 2 for 24 hr prior to inoculation . Virus was diluted in PBS (pH 7 .2) alone, and 0 .1 ml inoculated in quadruplicate onto the CAM of the chick embryo (Beveridge and Burnet, 1946 ; Hawkes, 1979) .
To whom requests for reprints should be addressed . 0042-6822/89 $3 .00 Copyright rd 1989 by Academic Press . m a . All rights of reproduction in any form reserved.
262
INHIBITION OF INFLAMMATION BY CPV
Enzymes and chemicals Restriction enzymes were supplied by Bethesda Research Laboratories, Inc . (Gaithersburg, MD), New England BioLabs, Inc . (Beverly, MA), or Boehringer-Mannheim Biochemicals (Indianapolis, IN) and used as specified by the manufacturers . The Klenow fragment of DNA polymerase I, T, DNA ligase, T, DNA kinase, 5bromo-4-chloro-3-indolyl-fl-D-galactopyranoside (XGal), and isopropyl-,6-D-thiogalactopyranoside (IPTG) were obtained from Bethesda Research Laboratories, Inc . Alkaline phosphatase from calf intestine was purchased from Boehringer-Mannheim Biochemicals . Virus infectivity assays Virus pocks were excised from the CAM as a 4-mm' piece of tissue and added to 1 ml of PBS containing 0 .1% BSA in 15-m1 conical tubes and stored at -20° . Virus was released from the tissue by grinding in a tissue grinder (Bellco, Inc ., Vineland, NJ) and dispersed by two 30-sec periods of sonication in a chilled-cup attachment to a Model W-385 sonicator (Heat SystemsUltrasonics, Inc ., Farmingdale, NY) . Virus infectivity was estimated as described previously (Buller and Wallace, 1985) . Virus plaques were visualized after 2-3 days by the addition of 0 .5 ml 0 .3% crystal violet formalin solution to each well (Nakano, 1979) . Virus infection of cells Cultures containing 1 x 10 6 BS-C-1 cells were infected with viruses at a multiplicity of infection (m .o .i .) of 0 .01 plaque-forming unit (PFU) in 0 .5 ml of MEM for 1 hr at 37° . The virus inocula were removed and the cultures were then washed twice with prewarmed MEM containing 10% BCS and the cells were fed with MEM containing 10% BCS . At the indicated time points, virus-infected cells were scraped into the culture supernatant, which was transferred to a 15-ml conical test tube, and stored at -20° . Prior to estimation of virus infectivity, the samples were frozen and thawed three times, followed by two 30-sec periods of sonication . DNA hybridization analyses of virus genomic DNA DNA was isolated from purified virus or virus-infected cells by SDS/protease K digestion followed by phenol extraction (Garon et al„ 1978) . The isolated DNA was cleaved with restriction enzymes, and the DNA fragments were resolved by electrophoresis on a 0 .6 or a 1 % agarose gel in a Tris-borate-EDTA buffer system . DNA fragments were transferred from the agarose gel to nitrocellulose membranes by a modification of the procedure of Southern (1975) . DNAs labeled by
26 3
the random primer method were hybridized to the immobilized DNA as described previously (Moss et al ., 1981 ; Feinberg and Vogelstein, 1983) . Autoradiographs were made by placing the washed nitrocellulose sheets in contact with X-ray film for 1-4 days at -70° . Construction of a deletion in the 38-kDa gene The 5 .2-kbp EcoRl restriction endonuclease fragment of CPV DNA containing the 38K gene and its flanking regions was ligated into EcoRl restriction endonuclease-digested and alkaline phosphatasetreated p1247 (a derivative of pUC19 lacking all restriction endonuclease enzyme sites in the polylinker region except EcoRl and HindIII) . Competent Escherichia coli JM105 cells were transformed, and ii-Gal colonies (white versus O-Gal' blue) were isolated by using the XGal-IPTG screening system . A miniprep from a selected colony (pGP1 247 .2) was digested with Pstl, and the large fragment was purified away from the 688-bp Pstl fragment (see Fig . 1A) . The large fragment was treated sequentially with T, DNA polymerase (under conditions optimal for the formation of flush-ended termini) and alkaline phosphatase . This dephosphorylated, blunt-ended vehicle was ligated with a 4 .6-kbp selection cassette . The cassette was constructed first by inserting the 3 .1-kbp Xbal/Smal fragment from pSC10, which contained the vaccinia virus late 11 K promotor coupled to the E. coli lacZ gene, into Hincll- and Xbal-digested pUC19 (Chakrabarti et al., 1985) . In addition, into the same plasmid was cloned an appropriately modified 1 .5-kbp HindllI and Xmal fragment from Tn5, that contained the entire gene encoding resistance to neomycin/kanamycin and which was under the control of the vaccinia virus early/late 7 .5 K promotor (Southern and Berg, 1982 ; Franke et al., 1985) . This cassette was further modified to allow excision by either Smal or Nod . The ligation reaction was used to transform competent E. coli JM105 cells, and blue colonies (Q-Gal') which grew in the presence of kanamycin were isolated . Transfection of plasmid and isolation of recombinant viruses A 35-mm-diameter petri dish of CV-1 cells was infected with virus at a m .o .i . of approximately 0 .05 PFU per cell followed by transfection with calcium phosphate-precipitated plasmid (Stow et al., 1978), and the resultant progeny virus was screened as described previously (Chakrabarti et al., 1985) . The frequency of recombination of the target sequence into the virus genome was between 0 .1 and 0 .01 % .
26 4
PALUMBO ET AL . A 38 kDa E
p
p1247
P
D
ECO 111
I T4 DNA Llgaaa
B TRANSFECTION
RECOMBINANT -
GENOTYPE
Intact ORF
Virus
Plasmid
CPV-Br
NA
-P'
NA
CPV-Br
pGPI .7
-p .
CPV-BR .D1
Deletion
+
Cassette
CPV-Br
pGPI .7
-p .
CPV-BR .D2
Deletion
+
Cassette
CPV-Br .D1
PGP1247.2-P'
CPV-BR.D1 .R2
Intact ORF
CPV-Br .D1
PGP1247.2-r
CPV-BR.D1 .R4
Intact ORF
FIG. 1 . Deletion of the DNA sequences within the 38K gene of CPV . (A) Methodology employed to construct two plasmids : one that contained the entire 38K ORF, and the other which replaced approximately 50% of the coding region of the 38K gene with a #-Galrrn5 neo selection markercassette . (B) Strategyfor the production of recombinants of CPV-BR which contained only a deletion in the 38K gene, and the subsequent reversion to an intact 38K ORF .
Histopathology CAMs were excised from the egg and placed into 10 ml of PBS in a 100-mm-diameter plastic dish . The CAMs were washed twice with PBS prior to fixation with Tellyesnickie's fixative (738 ml ethanol, 50 ml glacial acetic acid, 100 ml 37% formaldehyde solution to 262 ml with distilled water) . After 1 hr at room temperature or 8 hr at 4°, the fixative was aspirated from the plate and the CAMs were cut into 2 x 10-mm strips which were sectioned at 5 µm and stained with hematoxylin and eosin . Other sections were stained for fibrin (Humanson, 1972) . Immunocytochemistry A New Zealand White rabbit was bled and serum (preimmune) was prepared . The same rabbit was then injected intradermally at two sites with 0 .1 ml of 0 .01 M Tris, pH 8 .0, containing 2 X 10 9 PFU of gradient purified vaccinia virus strain WR (Joklik, 1962). Four weeks later, the rabbit was boosted by subcutaneous injec-
tions at 10 sites with 0 .1 ml of complete Freund's adjuvant containing 1 x 10 9 PFU of virus . Twelve weeks later, the rabbit was again boosted by subcutaneous injections at 10 sites with 0 .1 ml of incomplete Freund's adjuvant containing 1 x 10 9 PFU of virus . The rabbit was bled and serum prepared (anti-virus) 2 weeks after this last immunization . A horseradish peroxidase developing system was used to detect antigen-antibody complexes . All tissue sections were treated with 0 .3% hydrogen peroxide in water for 30 min to reduce the endogenous peroxidase activity prior to incubation with a 1 :30,000 dilution of preimmune or immune sera . The primary antibody binding was detected with a Vectastain Elite ABC kit with an avidin/biotin/peroxidase complex developing system, following the manufacturer's instructions (Vector Laboratories, Burlingame, CA) . Slides were counterstained using Gill's hematoxylin . The sensitivity of the assay permitted the detection of even a single antigen-positive cell at 24 hr p .i . in a tissue section of CAM (McIntyre, unpublished results) .
INHIBITION OF INFLAMMATION BY CPV
X
C
C
CK
CX
r 1 Kb
SP41
265
R, BS P K C K E . . . . . „ .
I
1 ITA
FIG . 2 . Structure of the 38K gene in CPV- BR, mutant, and revertant viruses . Digestion of viral DNA with the restriction endonucleases Pstl (A), Clal (Band D), and EcoRl (C) . Viral DNAs were isolated from BS-C-1 cells infected with CPV-BR, the two mutants CPV-BR .D1 or CPV-BR .D2 that
haves deletion of the 38K gene, or either of the two revertants CPV-BR .D1 .R2 and CPV-BR .D1 R4 that have an intact 38K ORF . A 32P-labeled plasmid containing the 5 .2-kbp EcoRl fragment (heavy black line (F)) of CPV-BR that contained the entire coding sequence of the 38K gene was used as the probe in (A), (B), and (C) . A 32P-labeled oligonucleotide specific for the Tn5 neo gene was used as the probe in (D) . (E) Restriction endonuclease map of the right-hand end of CPV-BR ; BamHl (B), Clal (C), EcoRl (E), Hindlll ( H), Kpnl (K), Pall (P), Sall (S), Xhol (X), serine protease inhibitor-1 (SPI-1), and inverted terminal repeat (ITR) .
RESULTS Construction of a deletion in the gene encoding the 38-kDa protein Two different plasmids were constructed for use in the production of recombinants of CPV-BR . The 5 .2kbp EcoRl fragment that encompassed the entire open reading frame (ORF) of the 38K gene was inserted into p1247 (Fig . 1 A ; Materials and Methods) . This plasmid, pGP1 247 .2, contained two Pstl restriction endonuclease sites which bound a 688-bp sequence (Pstl-R fragment) within the OFF of the 38K gene . Digestion of pGP1 247 .2 with the restriction endonuclease Pstl and the replacement of the Pstl-R fragment with the p-Gal/ Tn5 neo cassette resulted in the isolation of pGP1 .7 .
The DNA structure of pGP 1 .7 was confirmed by EcoRl, Pstl, and Clal restriction endonuclease digestions and dideoxy sequencing from the termini of the cassette into the gene (Palumbo, unpublished results ; Hattori and Sakaki, 1986) . The plasmid GP1 .7 was used to make two individual recombinants of the Brighton strain of CPV, CPVBR .D1, and CPV-BR .D2, which had deletions of approximately half of the 38K gene, and insertions of the /3-Gal/Tn5 neo cassette (Fig . 18) . Recombinant CPVBR .D1 and pGP1 247 .2 (intact 38K ORF) were used to construct and to isolate independently two revertant viruses, CPV-BR .D1 .R2 and CPV-BR .D1 .R4, which had regained an intact 38K ORF . The CPV-BR .D1 virus replicated in BS-C-1 cells in a
2 66
PALUMBO ET AL.
INHIBITION OF INFLAMMATION BY CPV
manner indistinguishable from CPV-BR . At 40 hr p .i ., the geometric mean plaque diameter of CPV-BR .D1 was 0 .62 X 1 .05 mm compared to the CPV-BR value of 0 .64 X 1 .04 mm (Student's t test, P > 0 .05 ; 20 of) . Similarly . BS-C-1 cell cultures, infected with CPV-BR and CPV-BR .D1 viruses and harvested at 10, 24, and 48 hr p .i ., gave total culture virus infectivity values at each time point which were not significantly different from one another . Southern blot analysis of the genomes of the recombinant viruses To confirm the predicted DNA genome structure of the isolated recombinant viruses, Southern blot DNA analyses were performed on total DNA from virus-infected cells (Fig . 2) . For each virus, restriction endonuclease-generated DNA fragments were separated by agarose gel electrophoresis, transferred to nitrocellulose, and hybridized to 32 P-labeled DNA . The hybridized DNAwas visualized through autoradiography . The sizes of the restriction endonuclease fragments generated for both CPV-BR and the recombinant viruses were compared to the sizes of the restriction fragments expected upon successful generation of the desired recombinant CPV . A 32 P-labeled 5 .2-kbp EcoRl fragment (Fig . 2E ; probe) detected the 688-bp Pstl-R fragment (indicated as 0 .7 kbp in Fig . 2A) present in CPV-BR and the revertant virus DNAs, but this fragment was absent in the DNAs from mutant viruses CPV-BR .D1 and CPVBR .D2 . Analysis of the Clal restriction endonucleasedigested viral DNA confirmed the insertion of the cassette into the 38K open reading frame (Fig . 2B) . Of the five Clal restriction endonuclease fragments (Fig . 2E)
267
detected in CPV-BR and revertant virus DNAs (Fig . 2B), three were also present in the mutant virus DNAs, as were two Clal fragments of 2 .4 and 4 .5 kbp . The latter fragment was created by the fusion of DNA sequences present in the S-Gal/Tn5 neo cassette (containing a single Clal site approximately 2 .2 kbp from the end of the /acZ gene) with the 2 .2-kbp DNA fragment (deleted 2 .4kbp WT fragment) from the right-hand side of the 38K gene (Fig . 2E) . The smaller 2 .4-kbp Clal fragment was formed by fusion of 218 nucleotides derived from the left-hand end of the 38K ORF and approximately 2 .2 kbp of sequence from the /3-Gal/Tn5 neo cassette . In addition, the same 5 .2-kbp EcoRl fragment that contained the intact 38K gene was present in CPV-BR and revertant virus DNAs, while new restriction fragments of predicted size were detected in the viruses with a deletion in the 38K gene (Fig . 2C ; f3-Gal/Tn5 neo cassette contains one EcoRl site 3 .1 kbp from the end of the /acZ gene) . Additional Southern blots in which recombinant virus DNAs were digested with Pstl, EcoRl, Clal, and Hindlll restriction endonucleases and probed with 22 P-labeled total CPV-BR DNA showed no additional changes in DNA structure of the deletion mutants (data not shown) . Deletion of the cassette from the revertant viruses was confirmed by using an oligonucleotide probe to a Tn5 neo DNA sequence which hybridized only to the 4 .5-kbp Clal restriction endonuclease fusion fragment (Fig . 2D) . Thus, we have constructed and isolated recombinant viruses, two of which showed DNA structure alterations consistent with only a deletion/insertion into the 38K gene sequence, and a further two recombinants that appeared by DNA structure analysis to be identical to the wildtype CPV-BR .
FIG . 3 . CPV-BR and CPV-BR .D1 pock morphology on chick embryo CAMs 96 hr after virus inoculation . The CAMs of 12-day-old chicken embryos were inoculated with viruses and harvested at 96 hr p .i . The CAMs were washed several times with PBS and photographed with backand side-lighting . (A) Discrete, bright red pocks induced by CPV-BR scattered along the vascular tree ; (B) the greater magnification of the pock indicated by the arrow in (A) shows the network of congested blood vessels constituting the bulk of the pock ; (C) white pocks induced by CPVBR .D1 were mostly solid while, but the ones indicated by arrowheads contain the red center commonly seen in earlier pocks induced by CPVBR .D1 ; (D) a greater magnification of the pock indicated by the arrow in (C) showing the raised white center and discrete border . The contrast with the red pock (B) is evident [line : 1 .4 cm (A and 10 .22 cm (B and D)] . Fio . 5. The distribution of virus-specific antigen within CPV-BR and CPV-BR .D1 pocks on chicken embryo CAMs at various times after infection . (A) CPV-BR-induced pock at 24 hr p .i . showing a dense concentration of virus antigen in the central area and ectodermal proliferation (arrows) . Note the lack of both hemorrhage and cellular infiltration within the mesoderm . (B) CPV-BR .D1-induced pock at 24 hr p .i . showing viral antigen and ectodermal proliferation (arrows) similar to the CPV-BR pock . At this time, the mesodermal vessel remains patent, and there is minimal adhesion of inflammatory cells to the endothelium . (C) CPV-BR pock at 48 hr p .i . with extensive vacuolahon (v) and the presence of large amounts of viral antigen not only within the expanded ectoderm, but also in fibroblasts in the mesoderm (arrowheads) . Note the general lack of extravasation of erythrocytes from engorged blood vessels (C, arrows) . This is shown in greater detail in a higher magnification of the boxed region in (C) which is presented in (E) . Note the intact endothelium of the blood vessels (arrowheads), although there are some erythrocytes present in the mesoderm (arrows) . (D) CPV-BR .D1 pock at 48 hr p .i . showing the extensive infiltration of the mesoderm with inflammatory cells which have isolated antigen containing areas (arrows) that probably represent scattered remnants of the original ectoderm . At lower right is the endoderm, which shows proliferative changes . A scab overlies the ectoderm constituting the original pock . (F) Enlargement of the area boxed in (D) . Antigen was visualized by immunological staining with a horseradish peroxidase developing system followed by hematoxylin counterstaining (see Materials and Methods) [line : 6 .3 µm (A and B) ; 13 .4 gm (C and O) ; 4 .2 g m (F) ; and 6 .3 pro (F)] .
268
PALUMBO ET AL .
Fic . 4 . Histochemical examination of CPV-BR and CPV-BR .D1 virus pocks from chick embryo CAMs . CAM of 12-day-old embryo at 721 (A) CPV-BR-induced pock . The ectodermal layer is thickened and is beginning to undergo lytic changes and vesicle formation (v) assoc with virus replication, as evidenced by numerous intracytoplasmic inclusion bodies (white arrows) . Note the clear line of demarcation maint between ectoderm and mesoderm (black arrows in opposition) . Most of the cells within the mesoderm are erythrocytes (arrowheads) . (B) induced with CPV-BR .D1 showing the massive inflammatory cell influx into the mesoderm, central erosion of the ectoderm, and scab form
INHIBITION OF INFLAMMATION BY CPV
A
PFU/Pock at 96Hr Post Infection (1X10 -5 ) 0
10
15
20
CPV-Br (wt) CPV-Br .DI .R2 CPV-Br .D1 .R4 CPV-Br .D1 CPV-Br .02
B
PFU/Pock (1x10 -4 ) 0 1 2 3 4 5 8 7 8 9
10
48 hr
72 hr
96 hr
FIG . 6 . Virus infectivity recovered from pocks isolated from CAMs of chicken embryos . (A) Comparison of virus recovered from pocks induced by infection of CAMs with CPV-BR, mutant, or revertant viruses . (B) Virus infectivity recovered at the indicated time points from pocks induced by either CPV-BR (.) or CPV-BR .D1 (D) . Individual pocks were excised from the CAM, and infectivity was quantitated as described under Materials and Methods . Arithmetic means and one SEM were calculated from a minimum of tour pocks from an individual egg . Previous studies have indicated that there was considerable variation in the virus infectivity yields among eggs inoculated with the same dose of virus (Butler et at, 1988) . In all cases, of the four replicate eggs inoculated with virus, the egg which gave virus pocks with the maximal virus infectivity was used to calculate the means reported .
Macroscopic character of the wild-type and 38K deletion mutant pocks To determine the effect of a deletion in the 38K gene on the CPV-BR red-pock phenotype, the CAMs of 12day-old chick embryos were infected with CPV-BR or
269
the mutant viruses . The CPV-BR pock appeared macroscopically by 24 hr as a slightly raised, translucent thickening of the CAM but, by 48 hr, blood vessels became dilated and congested, giving it a more pronounced red color. Pocks became larger so that, by 96 hr, they had reached a diameter of 2-3 mm (Fig . 3A) . The dilated and congested nature of the blood vessels was shown clearly in the higher magnification of the pock indicated in Fig . 3B . In certain experiments, by 96 hr p .i . some pocks formed hematocysts . The two revertant viruses, CPV-BR .D1 .R2 and CPV-BR .D1 .R4, were examined in a similar fashion and yielded a pock morphology which was indistinguishable from CPV-BR at all time points examined . The 48-hr pocks induced by CPV-BR .D1 differed from CPV-BR, as they were raised and opaque, with small red centers . By 72 hr p .i ., the pocks had reached maximal sizes of 2-3 mm, and the red centers were less distinct . At 96 hr p .i ., CPV-BR .D1 pocks appeared dull and had lost their sheen (Fig . 3C) . A higher magnification of a typical pock showed the center of the pock to be raised and surrounded by a white region within the CAM tissue (Fig . 3D) . Identical results were obtained with CPV-BR .D2 and a deletion mutant, CPVBR .DC1, which lacked the /3-Gal/Tn5 neo cassette . The development and character of CPV-BR .D1, CPVBR .D2, and CPV-BR .DC1 pocks were similar to that described for a naturally occurring white-pock variant, W2, isolated by Pickup and co-workers (1984) . Histology of CPV-BR, mutant, and revertant pocks The difference between the pocks of CPV-BR and the CPV-BR .D series of mutants was even more apparent on histological examination of CAM tissues at 72 hr p .i . (Fig . 4) . The CPV-BR pocks were characterized by extensive ectodermal proliferation at the advancing margin of the pock (Bulleretal ., 1988), while at thecenter, there was erosion of the ectoderm layer with resultant vesicle formation and an abundance of inclusion bodies (Fig . 4A) . A sharp line of demarcation remained between the ectoderm and the mesoderm . The blood vessels were dilated and congested with certain pocks showing minimal extravasation of blood cells into the mesoderm, but other pocks showed severe hemorrhage accompanied by fibrin deposition . A consistent finding was the total lack of an inflammatory response .
associated with this mutant . The scab is outlined by arrowheads . The endoderm remains relatively normal . (C) Higher magnification of the boxed area shown in (B) . Note the lack of well-demarcated inclusion bodies (possible inclusions are indicated by white arrows) in the relatively intact ectoderm . The junction between the ectoderm and mesoderm is indistinct . Most cells within the mesoderm appear to be macrophages (black arrow) ; within the scab, heterophils predominate . (D) A healed pock induced with CPV-BR .D1 with regrowth of the ectoderm (arrowheads) capped by a small scab . Macrophage infiltration of the dermis is reduced compared to early pocks . All sections were stained with hematoxylin and eosin [line : 4 .2 pro (A); 13 .4 µm (B) ; 4 .2 µm (C); and 13 .4 µm (D)] .
270
PALUMBO ET AL .
The two revertant viruses induced pocks with a morphology identical to that of CPV-BR . In contrast, the pocks induced by the mutant CPVBR .D1 were dominated by a massive influx of inflammatory cells composed of macrophages and heterophils which resulted in destruction of the integrity of the ectodermal layer and the formation of a scab (Fig . 4B) . A higher magnification of this region is presented in Fig . 4C . Within a thickened mesoderm, there was moderate proliferation of fibroblasts and accumulations of macrophages . Blood vessels within the mesoderm were moderately dilated and contained thrombi . As with the wild-type virus, fibrin deposits were detected in pocks associated with hemorrhage . As indicated in Fig . 4D, pocks by 72 hr p .i . were observed in an advance stage of ''healing," as indicated by the extensive regrowth of ectodermal cells and the presence of diminished numbers of inflammatory cells, of which macrophages were more prominent than heterophils . In some of these pocks, virus antigen could be seen sequestered to the outer surface of the regenerated ectoderm layer. CPV-B R . D2-infected pocks yielded findings identical to CPV-BR .D1 . These findings illustrated the fundamental difference between the CPV-BR and mutantvirus infections with regard to induction of the host inflammatory response . Distribution of virus antigen within tissue of the pock To determine if CPV-BR and CPV-BR .D1 viruses differed in their replication and spread within the ectoderm, and thereby caused dissimilar levels of cellular necrosis, virus pocks were analyzed at 24 and 48 hr p .i . for antigen distribution . Both CPV-BR and CPV-BR .D1 pocks at 24 hr p .i . appeared very similar with minimal perturbations of the ectodermal layer, no involvement of the mesoderm, and a lack of inflammatory cell infiltrates (Figs . 5A and 5B) . A similar distribution of antigen-positive cells was detected in both infections . Ectodermal hyperplasia was associated with the advancing margin of the pock . The similarity between the CPV-BR and CPV-BR .D1 24-hr pocks argued against tissue necrosis as the sole explanation for the differences in the virus-induced inflammatory responses observed at 72 hr p .i . in Fig . 4 . By 48 hr p .i ., clear differences in virus antigen staining were detected between CPV-BR and CPV-BR .D1 . In comparison to CPV-BR .D1, CPV-BR pocks showed more intense staining, indicative of the presence of more virus antigen within infected cells, and a greater number of antigen-positive cells (compare Fig . 5C with 5D) . The pock in Fig . 5C clearly depicted dilation and congestion of the blood vessels beneath and within the ectoderm, and the lack of inflammatory cells within the
mesoderm . A higher magnification of this region showed the endothelial cells lining the blood vessels to be intact with a minimal extravasation of erythrocytes into the mesoderm (Fig . 5E, see also Fig . 4A) . In CPVBR virus pocks, antigen-positive fibroblasts could be detected in the mesoderm at a great distance from the center of infection (Fig . 5C, arrowheads) . On occasion, endothelial cells lining the arterioles and venues also stained positively for virus antigen . In contrast, the distribution of antigen-positive cells in CPV-BR .D1 pocks appeared more restricted and was fragmented by the inflammatory infiltrate (Figs . 5D and 5F) . It is uncertain whether these antigen-positive cells are producing virus progeny or are phagocytes containing virus antigenic debris . As noted previously in Fig . 4 and seen again in Fig . 5D, the demarcation between the ectoderm and mesoderm was obliterated by the magnitude of the inflammatory influx into the pocks induced by the 38K deletion mutants . Virus infectivity recovered from pocks To determine if the observed reduction in detectable antigen in pocks of CPV-BR .D1 compared to CPV-BR was paralleled by similar reductions in virus progeny, individual pocks from CPV-BR- and CPV-BR .D1-infected CAMs were analyzed for virus infectivity . The average infectivity yields of pocks isolated at 96 hr p .i . from CAMs infected with CPV-BR, the two independently isolated mutants lacking an intact 38K gene, and two independently isolated revertants of CPV-BR .D1 are depicted in Fig . 6A . Both mutants showed reductions in average yields of virus recovered from pocks in comparison to CPV-BR (P < 0 .05), 2 while the two 38K gene revertants of CPV-BR .D1 gave pock infectivity yields similar to those for CPV-BR pocks . To investigate whether the differences between virus pock infectivity yields varied with time p .i ., CPV-BR-and CPV-BR .D1infected CAMs were isolated at 48, 72, and 96 hr p .i ., and individual virus pocks were analyzed forvirus infectivity yields . Less virus infectivity was recovered from pocks induced by the mutant, CPV-BR .D1, in comparison to CPV-BR at all times examined (Fig . 513) (P < 0 .05) . 2 These findings are consistent with the diminished virus antigen detected in mutant pocks . DISCUSSION Variants of CPV which are identified by a white pock morphology induced on the CAM occur regularly with2 Data for each virus was subjected to sing le-factor analysis of variance (ANOV) (Zar, 1974) . Upon rejection of the null hypothesis that the mean virus yields were equivalent by ANOV, Newman-Keuls multiple comparison analysis of means was performed (a = 0 .05) (Zar, 1974).
INHIBITION OF INFLAMMATION BY CPV
out obvious selection for a genetic determinant (Downie and Haddock, 1952 ; van Tongeren, 1952) . These white-pock variants are associated with DNA rearrangements which typically involved the replacement of up to 39 kbp of the right-hand end of the genome, with an inverted copy of up to 50 kbp from the left-hand end of the genome (Archard et al., 1984 ; Pickup et al., 1984) . This type of rearrangement results in the loss of at least 29 kbp of mainly unique sequences from the right-hand end of the virus genome, which makes it difficult to relate alterations in pock morphology with the loss of a specific gene . Using a battery of independently isolated white-pock variants and marker-rescue techniques, Pickup and colleagues showed that a 38K gene which mapped between 31 and 32 kbp from the right-hand end of the genome was a major component in the determination of the white-pock phenotype ; however, recombination of this 38K gene into the TK gene of white-pock variants did not result in pocks as red as CPV-BR wild type (Pickup et al., 1984, 1986) . These experiments suggested the presence of additional genes which influenced pock phenotype and were encoded in the DNA sequences of the right-hand end of the CPV-BR virus genome . A candidate is the CPV homolog to the serine protease inhibitor-1 (SPI-1) gene of vaccinia virus that is located in this region of both the vaccinia and CPV genomes (see Fig . 2E) and has significant homology at the amino acid level with the 38-kDa protein (Kotwal and Moss, 1989 ; Palumbo and Buller, unpublished data) . Further evidence for the presence of additional genes involved in CPV-BR-pock phenotype came from the work of Chernos at al . (1985), who found that selected recombinants between vaccinia and ectromelia viruses gave red pocks on the CAM, whereas the parental viruses gave only white pocks . Direct evidence is presented in this report that inactivation of the 38K gene alone is sufficient to change the wild-type virus red pock phenotype to white . The whitepock phenotype was specifically the result of a deletion in the 38K gene, and not a second mutation elsewhere in the genome, since recombination between a 38K deletion mutant and a plasmid containing an intact 38K gene resulted in the isolation of revertant viruses which were indistinguishable from wild-type CPV-BR, These experiments strongly suggested that the white-pock phenotype was either directly or indirectly the consequence of the loss of the 38-kDa protein and that this protein is normally required to produce the red pock phenotype, but it does not rule out the possibility that other genes are also involved . To address this last possibility, experiments are in progress to delete the SPI-1 gene of CPV-BR .
27 1
A number of the differences in the pocks that form on a CAM between CPV-BR and 38K gene deletion mutants observed in this study were similar to those previously reported using naturally occurring white-pock deletion mutants : (1) The CPV-BR gave a red pock which was characterized by hemorrhage at 72 to 96 hr p . i . ; (2) the raised, white, and opaque phenotype of the white-pock variants was correlated with an influx of inflammatory cells into the pock ; and (3) the pocks of the white-pock variants, on average, had less infectivity than comparable CPV-BR pocks (Downie, 1939 ; Downie and Haddock, 1952 ; van Tongeren, 1952 ; Fenner, 1958 ; Baxby, 1969) . This study extended our understanding of CPV-BR virus pock development in the CAM of the chick embryo, delineated more clearly the differences between CPV-BR and white-pock variant viruses, and suggested a possible role for the 38-kDa protein . Within the first 24 hr after inoculation of the CAM with virus, sites of viral infection could not be detected reliably by microscopic observations of stained tissue sections . The early events in virus infection of the CAM were, however, revealed using immunocytochemistry and were found to be similar for both viruses, with no detectable perturbations of either the ectoderm or mesoderm . In contrast, by 48 hr p .i ., there were gross differences in pock color and form . Microscopic observations described above indicate that the red color of the CPV-BR pock is best explained by dilation and congestion of the dense intra- and epiectodermal capillary beds in the CAM which provide the respiratory function of the chick embryo (Ramanoff, 1960) . This was clearly shown in Figs . 5C and 5E where the individual blood vessels underlying the ectoderm were dilated to such an extent as to appear as one continuous vessel . Since the mutant also showed congestion and dilation of blood vessels, the presence of a red color only in the center of the mutant pocks can be explained by the screening effect of inflammatory cells migrating into the mesoderm and ectoderm . The reduction of virus infectivity and in the number of antigen-positive cells in pocks induced by the deletion mutant by 48 hr p .i . in comparison to CPV-BR can be explained as a direct result of the inhibition of virus spread by infiltrating inflammatory cells . Thus, these results suggest that CPV-BR evades the inflammatory response-a nonspecific host defense mechanism-by production of the 38-kDa protein . The simplest hypothesis would predict that the 38kDa protein expressed in the wild type, but not the mutant virus infection, must hinder the generation/action of a chemotactic substances) from within the pock which is directly or indirectly responsible for the migra-
272
PALUMBO ET AL .
tion into the lesion of first heterophils and then macrophages . In light of the complicated nature of the inflammatory response, we can only speculate as to the target of the 38-kDa protein . One parameter that can be eliminated is any contribution by the immune system, since mature T and B cells are not yet present in the blood of the chick embryo at 12-15 days of incubation (Eerola et al., 1987 ; Chen and Cooper, 1987 ; Chen et al., 1986, 1988) . Further, in the absence of IgE, it is unlikely that basophils or tissue mast cells play a role . In addition, the classical complement pathway is inoperative, since anti-virus antibody is lacking, and previous investigators have shown that the 12- to 15-day-old chick embryos have total C and C1 levels of 1-4% of the adult chicken (Gabrielsen et al., 1973) . On the other hand, during inflammation three mediator-producing systems-the kinin, the coagulation, and the fibrinolyticas well as the catabolic pathways of arachidonic acid (Samuelsson et al., 1979) can become activated and these would be prime candidates as targets of the 38kDa protein . Computer-assisted searches of the protein database of the National Biomedical Research Foundation have shown that the 38-kDa protein has significant sequence similarities to several inhibitors of serine proteases . For example, it is similar to both human antithrombin III and placental plasminogen activator inhibitor (having optimized match scores of 366 and 479, respectively, out of a possible score of 1679) . The structural similarity between these inhibitors and the 38-kDa protein suggested that the viral protein might induce red-pock formation by inhibiting serine proteases involved in either the blood coagulation pathway or the normal processes of wound-containment and healing . The results of these studies on the development of pocks show that the primary effect of the 38kDa protein is to repress the massive infiltration of inflammatory cells which normally occurs within 48 hr of infection . At this stage of the infection, the redness of the pocks produced by the wild-type virus is the result of both the lack of infiltration by inflammatory cells and the dilation and congestion of the blood vessels, Only at later stages of infection (72-96 hr), when there is greater tissue damage, do pocks begin to show evidence of extensive hemorrhage . Accordingly, the characteristic hemorrhage in pocks produced by the wildtype virus may be the consequence of the lack of an appropriate inflammatory response, rather than a direct inhibition of the enzymes involved in the blood coagulation pathway. Such a mechanism is consistent with the observations that, under certain conditions, white-pock variants of CPV can generate red pocks
(Baxby, 1969) . Such atypical pocks may be formed in the embryo if it is incubated at suboptimal temperatures, or if it is infected with more than 100 pock-forming units of virus ; under these conditions, the embryo's normal inflammatory response may be impaired or overwhelmed . To date, placental plasminogen activator inhibitor has yielded the most significant level of relatedness to the 38-kDa protein . This protein inhibits the conversion of plasminogen to plasmin . Plasmin may generate chemotactic substances either by direct action on complement component C5 to produce a C5a-like molecule (Larsen and Henson, 1985) or by dissolving a fibrin clot (Belew et al., 1978) . C5a has been shown to be a potent chemoattractant for neutrophils (see review, Goldstein, 1988) . Other proteases, including trypsin and kallikrein, have also been implicated in the generation of chemotactic fragments from C5 (Larsen and Henson, 1983), so it is conceivable that serine proteases directly or indirectly released as a consequence of virus infection of ectodermal cells could directly act on C5 . The 38-kDa protein would specifically inhibit some step in this process . Proper evaluation of a C5amediated inflammatory response induced by direct action of proteases on C5 must await measurements of the levels of C5 in the chick embryo and in vitro studies . ACKNOWLEDGMENTS We thank Mr . C . Duarte and Dr . D . Ailing for expert technical and statistical assistance, respectively ; Drs . B . Moss, M . Boyle, S . Isaacs, G . Kotwal, and J . Sechler for helpful discussions ; and Ms . B . R. Marshall for accurate and efficient manuscript production .
REFERENCES ARcHARO, L . C ., MACKETn, M ., BARNES, D . E ., and DuMBELL, K . R . (1984) . The genome structure of cowpox virus white pock variants . J . Gen . Wool. 65, 875-886 . BA
INHIBITION OF INFLAMMATION BY CPV
sual screening of recombinant virus plaques . Mol . Cell. Biol. 5, 3403-3409 . CHEN, C .-L . H ., AcER, L . L ., GARTLAND, G . L ., and COOPER, M . D . (1986) . Identification of a T3/T cell receptor complex in chickens . J.Exp . Med. 164, 375 -380 . CHEN, C .-L . H ., CIHAK, J ., LoSCH, U ., and COOPER, M . D . (1988) . Differential expression of two T cell receptors, TcR1 and TcR2, on chicken lymphocytes . Fur. J. Immunol. 18,539-543 . CHEN, C .-L . H ., and COOP=ER, M . D . (1987) . Identification of cell surface molecules on chicken lymphocytes with monoclonal antibodies . In ''Avian Immunology: Basis and Practice" (A . Toivanen, and P . Toivanen, Eds .) . CRC Press, Boca Raton, FL . CHERNOS, V . I ., ANTOtaovA, T . P ., and SENKEVICH, T . G . (1985) . Recombinants between vaccinia and ectromelia viruses bearing the specific pathogenicity markers of both parents . J Gen . Viral. 66, 621626 . DOWNIE, A . W . (1939) . A study of the lesions produced experimentally by cowpox virus . I Pethol. 48, 361-379 . DOWNIE, A . W ., and HADDOCK, D . W . (1952) . A variant of cowpox virus .Lancet 1,1049-1050 . EEROLA, E ., VEROMAA, T ., and TOIVANEN, P . (1987) . Special features in the structural organization of the avian lymphoid s ystem . in ''Avian Immunology : Basis and Practice" (A . Toivanen, and P . Toivanen, Eds .) . CRC Press, Boca Raton, FL . DE MERS, A . P ., and VOGELSTEIN, B . (1983) . A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity . Anal. Biochem . 132, 6-13 . FENNER, F- (1958)- The biological characters of several strains of vacclnla, cowpox and rabbitpox viruses . Virology 5, 502-529 . FRANKE, C . A . . RICE, C. M ., STRAUSS, J . H ., and HRUSY, D . E . (1985) . Neomycin resistance as a dominant selectable marker for selec tion and isolation of vaccinia virus recombinants . Mol. Cell. Riot. 5, 1918-1924 . GARRIEESEN . A . E ., PICKERING, R . J ., LINNA, T . 1 ., and GOOD, R . A . (1973)- Haemolysis in chicken serum . II . Ontogenetic development . lmmunology25, 179-184 . GARON, C . F ., BARBOSA, E ., and Moss, B . (1978) . Visualization of an inverted terminal repetition in vaccinia virus DNA . Proc. Nell . Acad . Sci . USA 75,4863-4867 . GOLOSTEIN, I . M . (1988) . Complement : Biologically active products . In 'Inflammation : Basic Principles and Clinical Correlates'' (1 . I . Gallin, I . M . Goldstein, and R . Snyderman, Eds .) . Raven Press, New York .
273
ettsial and Chlamydial Infections' (E . H . Lennette, and N . J . Schmidt, Eds .) . American Public Health Association, Inc ., Washington, DC . HUVANSON, G . L . (1972) . ''Animal Tissue Techniques'' 3rd ed ., p . 266 . Freeman, San Francisco, CA . JoKLIK, W . K . (1962) . The preparation and characteristics of highly purified radioactively labeled poxviruses . 8iochim . Biophys . Acta 61,290-301 . KowtAL, G . J ., and Moss, B . (1989) . Vaccinia virus encodes two proteins that are structurally related to members of the plasma serine protease inhibitor superfamily .I Viral 63, 600-606 . LARSEN, G . L ., and HENSON, P . M . (1983) . Mediators of inflammation . Annu . Rev. Immunol. 1, 335-359Moss, B ., WINTERS, E ., and COOPER, l . A . (1981) . Deletion of a 9,000base-pair segment of the vaccinia virus genome that encodes nonessential polypeplides . J.. ViroL 40, 387-395 . NAKANO, 1. H . (1979) . Poxyvrruses . In "Diagnostic Procedures for Viral, Rickettsial and Chlamydial Infections," 5th ed . (E . H . Lennette, and N . J . Schmidt, Eds .), pp . 293 296 . American Public Health Association, Washington, DC . PICKUP, D . J ., INK, 8 . 5 ., Hu, W ., RAY, C- A-, and JOKUK, W . K . (1986) . Hemorrhage in lesions caused by cowpox virus is induced by a viral protein that is related to plasma protein inhibitors of serine proteases . Proc. Natl . Acad. Sci. USA 83, 7698-7702 . PICKUP, D . J ., INK, B . S ., PARSONS, B . L, Hu, W and JOKLIK, W . K .
(1984) . Spontaneous deletions and duplications of sequences in the genome of cowpox virus . Proc . Natl. Acad . Sci. USA 81, 68176821 . RoMANOEE, A . L . (1960) . ''The Avian Embryo . Structural and Func tional Development ." MacMillan Co ., New YorkSAMUELSSON, B ., HAMMARSTROM, S ., and BORGEAI, P . (1979) . Pathways of arachidonic acid metabolism . In ' Advances in Inflammation Research" (G- Weissman, et al., Eds-) . Raven Press, New York SOUTHERN, E . M . (1975) . Detection of specific sequences among DNA fragments separated by gel electrophoresis . J. Mol. Biel . 98, 503-517 . SOUTHERN, P . J ., and BERG, P . (1982) . Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter . J Mol. Appl. Genet . 1, 327-341 . Stow, N . D ., SunAK-SHARPE, J . H ., and WILKIE, N . M . (1978) . Physical mapping of herpes simplex virus type 1 mutations by marker rescue .l. Viral. 28, 182-192 .
HA1TOR ., M ., and SASAKI, Y. (1986) . Dideoxy sequencing method using denatured plasmid templates . Anal . Biochem . 152, 232-238 .
VAN ToNOEREN . H . A. E . (1952) . Spontaneous mutation of cowpox virus by means of eggpassage . Arch . Gesamte Virusforsch . 5, 3552 .
HAwKES, R . A. (1979) . General principles underlying laboratory diagnosis of viral infections . In "Diagnostic Procedures for: Viral, Rick-
ZAR, J . H . (1974) . "Blostatistical Analysis-" Prentice-Hall, Inc ., Englewood Cliffs, NJ .