67
Journal of Virological Methods, 28 (1990) 67-78
Elsevier VIRMET 1000
Evaluation of sensitivity of different antigen and DNA-hybridization methods in African swine fever virus detection M.J. Pastor and J.M. Escribano Depamnt
of Animal Health, National Institute for Agrarian Research, Madrid, Spain
(Accepted 5 December 1989)
ELISA, immunodot and DNA hybridization methods have been adapted to detect African swine fever virus (ASFV), and their sensitivities were compared using virus obtained from cell cultures. About 2.3 x l@ 50% hemadsorbing doses (HADso) of virus were detected with ELISA sandwich using an anti-ASFV IgG biotinylated followed by avidin-peroxidase. The immunodot technique showed similar sensitivity, detecting about 4.6 x l@ HADso of virus. ASFV-DNA was detected using radioactive DNA probes and molecular hybridization. The maximal viral detection capacity of this technique was about 1.8 x 103 HAD,,. The antigenic and DNA detection of ASFV during the infection of animals with virulent and attenuated viruses, was also studied. For this purpose, sera and red blood cells from several infected pigs were obtained at different days post-inoculation. The virus was detected at the third day after infection by the three methods. However, ASFV-DNA detection was more efficient than antigenic detection at nine days post-inoculation, when antigen detection failed, because immunocomplexes with circulating viruses were formed in the subacute infection. African swine fever virus; ELBA;
Immunodot;
DNA hybridization
Correspondence ro: J.M. Escribano, Department of Animal Health, National Institute for Agrarian Research, Embajadores 68, 28012 Madrid, Spain.
0166-0934/90/$03.50 0 1990 Ekvier
Science Publishers B.V. (Biomedical Division)
68
Introduction African swine fever virus (ASFV) is a DNA virus that infects domestic pigs. ASF is one of the most important problems of the pig industry in affected countries. The lack of a vaccine makes the diagnostic procedures the only methodology that can help to plan the complete eradication of the disease in these countries. African swine fever occurs in forms varying from peracute to chronic and inapparent. In the peracute and acute forms the virus is present in large amounts in the blood and tissues such as the spleen, kidneys, lymph nodes and lungs. About 17% of the samples from outbreaks, submitted for testing in Spanish laboratories, are examined for detection of virus (data not published). In addition, nonhemadsorbing (Pini et al., 1974; Vigario et al., 1974; SanchezBotija et al., 1977) or noncytopathogenic strains of ASFV may cause problems in detection by virus cultivation in pig buffy coat cultures, the only way, at present, to accomplish ASFV isolation (Sanchez Botija et al., 1977, Hess, 1981). At present, there are well characterized tests for animals that produce antibodies following subacute ASFV infections. In contrast, a few previous reports have described antigenic (Crowther et al., 1979; Wardley et al., 1980) and ASFV-DNA detection (Caballero et al., 1986, Tabarts, 1987). No comparison of these methodologies has been performed to evaluate the sensitivity of ASFV detection in vitro and in infected pigs. In this report we have adapted several ELISA and immunodot (ID) techniques, and their sensitivities were compared to the use of viral DNA probes, both in vitro and in infected pigs.
Materials and Methods Cell, viruses and sera For in vitro assays the Spain-70 (E& strain of ASFV was used after forty-six passages in a monkey stable (MS) cell line (Alcaraz et al., 1989). Four different virus isolates were used to test the stability of cloned viral DNA fragments used as DNA probes. The viruses were ho, ET5, Em and EM strains isolated in different years and geographical areas of Spain. For pig inoculation, two different viruses were used. The wild-type ET0 strain as virulent virus, and the E,, strain, after four passages in CV, cells (ATCC, CCL 70) (E,s-CVt4), was used as attenuated virus. The first virus produces an acute disease in all the infected pigs, and the second virus produces a subacute disease in a high percentage of infected pigs. The viral titres were obtained by the haemadsorption test (Malmquist et al., 1960), and expressed in 50% haemadsorbing dose (HADso). The antiserum for the ASFV antigenic detection assays was obtained from five pigs that had been immunized with the nonvirulent strain of ET0 ASFV obtained after 46 passages in MS cells (ET0 MS&. The pigs were challenged with the ho-
69
mologous virulent virus and bled after viraemia had subsided (Ruiz-Gonzalvo et al., 1986). Each antiserum, before being pooled, presented ELISA titres higher than 1:20000. The IgG fraction of the pool was obtained by precipitation with ammonium sulphate. Pig inoculation
Four NIH minipigs (Sachs et al., 1976) of SLA ap weighing about 20 kg each were divided into two groups and held in separate isolated boxes for the duration of the experiment. Two animals were infected by intramuscular infection of 10r HADSo of E,,, virus, and two with the same dose of E,5 CV,4 virus. Peripheral blood was collected prior to inoculation and on subsequent days until the pig died, or until 12 days after inoculation in one surviving pig, given the attenuated virus. The serum was separated from each sample and the coagulated mass was treated with 2 volumes of distilled water to lyse the red blood cells with the associated virus (Wardley et al, 1977; Wesley et al., 1984). The recovered supematant was called lysed red blood cells (LRBC) fraction. Semipurification
of viral particles
MS cells were infected with E,,, h4S4s virus at a multiplicity of infection of about 2. Supematant was collected 24 h after infection and clarified by centrifugation at 1000 x g for 15 min. The supematant was ultracentrifuged through a cushion of 20% (W/W) sucrose in 50 mM Tris-HCl, pH 8, at 100000 x g for 75 min. The viral sediment was titrated by inoculation in huffy coat cultures. Preparation of viral DNA
Viral sediments, sera (200 l.~l)or LRBC fraction (200 ~1) were made to 10 mM EDTA, 1% SDS and 500 &ml proteinase K and incubated for 30 min at 45°C. The samples were deproteinized with phenol:chloroform:isoamylalcohol(25:24:1) and viral DNA precipitated with 2.5 volumes of ethanol and resuspended in TE buffer. Viral DNA probes
Two fragments of ASFV-DNA from the E,a MSM strain were used in the construction of the recombinant plasmids. The fragments were obtained after digestion of viral DNA with Hind111 endonuclease, and they were inserted into pACYC177 plasmid (Chang et al., 1978) at Hind111 site located in the kanamycin resistant region, using the E. coli C600 strain as host bacterium. The two Hind111 fragments cloned correspond to fragments E and F of E,c MS& viral DNA, with 4.1 and 3.9 Kpb, naming the recombinant plasmids ~161 and ~105, respectively (Fig. 1). The 32P-labeled probes were prepared by nick translation following the instructions supplied with kit (Bethesda Research Laboratories, U.S.A.), and used
70
E F PACYC 177
Fig. 1. Recombinant plasmids ~161 and ~105 obtained with the E and F Hind111 fragments of ASFVDNA inserted in the kanamycin resistance region of pACYC 177 plasmid. The figure shows an EtBr staining gel of the ASFV-DNA compared with recombinant plasmids ~161 and ~105 digested with HindIII.
with a specific activity of 9 X 10’ cpm/pg. The concentration of the probes in the hybridization was 0.33 Fg per 2.5 ml of hybridization buffer. Hybridization
technique
Viral DNA was filtered through a nitrocellulose filter presoaked in 2 x SSC using a biodot vacuum manifold. The hybridization process was performed at 55°C in plastic bags as described (Tabares, 1987). Briefly, the filter was rinsed in 2 x SSC, dried and baked at 80°C for 2 h. Before hybridization, the filter was incubated at 55°C for 1 h in pre-hybridization buffer (30% formamide, 6 x SSC, 0.1% SDS and 2 mg/ml each of bovine serum albumin, Ficoll and polyvinylpyrrolidone). For hybridization the filters were incubated at 55°C for 20 h in plastic bags with the labeled probe in pre-hybridization buffer containing 25 pg/ml of salmon sperm DNA. After hybridization the filters were rinsed three times in 2 x SSC at room temperature and once in pre-hybridization buffer at 55°C for 15 min. The filters were finally washed in 2 x SSC and dried. The hybrids were visualized by autoradiography. ELISA
and immunodot
assays
A 140 dilution of the anti-ASFV IgG in phosphate buffered saline (PBS) was used to coat microtitre plate wells. After 12 h at 4°C the microtitre plates were
71
washed with PBS-0.05% Tween 20 and serial two fold dilutions in PBS of samples including semipurified ASFV, sera, or LRBC fraction were added. After 1 h of incubation the immunocomplexes were developed by two different procedures. In the first anti-ASFV IgG labeled with peroxidase was added at 1:lOO dilution for 1 h. For the second labeled biotin anti-ASFV-IgG (Kodama et al., 1987) at the same dilution was reacted with the samples for 1 h and after washing, avidin-peroxidase was added at a working dilution in PBS-Tween 20 for 1 h. In both procedures the microtitre plates were then washed and substrate, orthophenylene diamine (OPD, Sigma, St. Louis, U.S.A.), was added. The reaction was stopped with 3 N sulphuric acid after 5 min. The immunodot assay (ID) was performed on the same samples used in ELISA assay by filtering through a nitrocellulose filter, using a biodot vacuum manifold (BioRad Laboratories, California, U.S.A.). The filter was washed in PBS and dried, and the immune reactions carried out using the same methodologies employed above which involved the use of anti-ASFV IgG labeled with peroxidase, or labeled with biotin followed by avidin-peroxidase. The presence of immunocomplexes was developed with the 4-chloro-1-naphthol technique (Hawkes et al., 1982).
Sensitivityof antigenic ASFV detection assays Supernatants of infected MS cells were used to semipurify the virus. It was titrated by the haemadsorption technique and used for antigenic detection experiments. Two ELISA sandwich techniques, using an anti-ASFV IgG labeled either with peroxidase or biotin followed by avidin-peroxidase, were adapted to detect ASFV obtained from cell cultures. The biotin-avidin procedure shows more sensitivity, detecting about 2.3 x 16L HADw, of virus, while anti-ASFV IgG labeled with peroxidase procedure detected 9.3 X 16L HADSo of virus (Fig. 2). The ID technique, directly fixing the virus to the filter and using both procedures employed in ELISA, detected about 4.6 X ld HADSO with biotin followed by avidin peroxidase and 9.3 x l@ HADSo with anti-ASFV IgG labeled with peroxidase methodology. The intensity of the spots for the biotin-avidin method was higher than anti-ASFV IgG labeled with peroxidase (Fig. 3). Sensitivityof ASFV-DNA detection assays Dot hybridizations were carried out using two different recombinant plasmids, ~161 and ~105 prepared with the E and F Hind111 fragments of ASFV-DNA of the E,,, MS, strain. The stability of the viral DNA fragments used as probes was determined by hybridization with four wild-type virus strains (E,a, E7s, E,, and E,&. Both probes resulted in a similar autoradiograph signal (Fig. 4, panel 1).
72
2.0
5: t d 6
1.0
0.1 3x10'
9.3~102
5.8X10
HAD5O Fig. 2. Comparison of detection of semipurified ASFV from infected cell culture supernatants, by two different ELBA methods. Serial ASFV dilutions were first reacted with anti-ASFV IgG adsorbed to the ELISA plate, and then with the same IgG labeled with peroxidase (0) or with biotin followed by avidin-peroxidase (a). Normal pig IgG, labelled with peroxidase (A) or with biotin (V) were used as controls. Absorptton at 0D4s0 was measured 5 min after addition of substrate.
Fig. 3. Comparison of detection of semipurified ASFV from infected cell culture supernatants by two different immunodot procedures. Serial ASFV dilutions were adsorbed to the nitrocellulose filter and reacted with anti-ASFV IgG labeled with peroxidase (1) or with biotin followed by avidin-peroxidase (2). The reactivity of both procedures was measured by densitographic analysis after substrate addition.
73
Fig. 4. Dot hybridixations of ASFV-DNA using the ‘*P-recombinant plasmids ~161 and ~105. Hybridizations were carried out using a concentration of the probes in the hybridization of 0.33 ug per 2.5 ml of hybridization buffer, employing 3 x 10’ cpm in each assay. Panel 1. Hybridization of ~105 and ~161 plasmids on nitrocelhdose filter with the DNA from 3 X 10’ HADW of four different ASFV isolates E,,,, E15, Ee,r and EM (1-t) after 24 h of autoradiographic exposure. Panel 2. Fixation of serial 2-fold dilutions of ASF’V-DNA starting with the viral DNA from 3 X lo’ HADs,, to nitrocellulose (A), Zeta-Probe (B) and Hybond-N (C) filters. The hybridization with ~105 plasmid was autoradiographed for 72 h. Panel 3. Hybridization of ~105 (A) and ~161 (B) plasmids with the ASFV-DNA diluted as described for panel 2, on nitroceUulose titters autoradiographed at 24 h (1) and 72 h (2). Panel 4. Hybridization of a mixture of ~105 and ~161 plasmids with a final specitic activity of 9 x 10’ cpm/pg, with the ASFV-DNA diluted as described in panel 2 and autoradiographed for 24 h.
Different filters were selected to analyze the capacity for adsorption of viral DNA. After hybridization, higher viral DNA was detected on the nitrocellulose filter (BioRad Laboratories, 0.45 pm, California, U.S.A.) than on nylon filters, Zeta-Probe (BioRad Laboratories) and Hybond-N (Amersham, 0.45 pm, England) (Fig. 4, panel 2). The sensitivity of viral DNA detection was assayed using each recombinant plasmid separately. After 24 and 72 h of autoradiography exposure viral DNA detection was the same with both probes, detecting about 1.8 x 103 HADSO at 72 h, which corresponds to 0.30 pg of viral DNA. Twenty-four hours exposure resulted in detection of 3.6 X 103 HAD% (Fig. 4, panel 3). When ~161 and ~105 plasmids were used jointly in the hybridization, at the same specific activity, the sensitivity of the technique in viral DNA detection was not increased, when compared to either of the plasmids alone (Fig. 4, panel 4).
A
2.0
s
d
B
I.0
d
0.1 03456789101112
DAYS AFTER
03456769101112
INFECTION
Fig. 5. ASFV detection by ELISA sandwich using the biotin-avidin procedure at different days after inoculation in the LRBC fraction (A) and in sera (B) from four ASFV-inoculated animals. Two pigs were inoculated with lo3 HAD,, of virulent virus (V,A) and two with the same dose of attenuated virus (@,o).
Comparison
of viral antigens and viral-DNA detection in ASFV
infected animals
Blood samples from two ASFV-infected pigs with the virulent ET0 strain (pigs A and B) and two with attenuated E,, CV14 virus (pigs C and D) were used to compare ASFV detection. By ELISA sandwich, using the biotin-avidin procedure, the virus was detected both in serum and in the LRBC fraction in animals inoc-
Fig. 6. ASFV detection by immunodot assay using biotin-avidin procedure from four inoculated animals: Pigs A and B were inoculated with virulent virus and C and D with the attenuated virus.
75
ulated with virulent and attenuated ASFV. Antigen detection was performed in one pig (A) both in serum and LRBC fraction at day 3 after inoculation and at day 4 in the other inoculated pig with the same virus (B) (Fig. 5). In the animals inoculated with the attenuated virus, it was detected in the pig that subsequently died (C) at day 4 after inoculation in the LRBC fraction, and at day 6 in the serum of the same animal (Fig. 5). In the other surviving pig (D), low amounts of viral antigens were detected in serum and in LRBC between 5 to 7 days after inoculation (Fig. 5). By the immunodot technique using the biotin-avidin procedure only the sera of infected pigs could be tested because of the backgrounds obtained with haemoglobin contamination from red blood cell lysates. In the sera of pigs inoculated with the virulent ASFV, virus was detected at days 4 and 5 by this ID technique (Fig. 6, rows A,B). In the case of attenuated ASFV, the virus was detected in serum only at day 6 in one pig and from days 6 to 9 after infection in the other (Fig. 6, row C, D). By dot hybridization using the ~105 32P-probe, ASFV-DNA was detected after 72 h of autoradiography, in pigs inoculated with virulent virus from days 3 to 5 in the LRBC fractions and at days 4 and 5 in the sera (Fig. 7). In the animals inoculated with attenuated ASFV, the viral DNA was detected in the LRBC fraction from days 3 to 6 in one pig (C) and from days 6 to 12 in the other (D) (Fig. 7). In the sera of animals inoculated with attenuated ASFV, viral DNA was detected in one pig (C) at day 6 only and in the other (D) from day 7 to the end of day 12 (Fig. 7).
Fig. 7. ASFV detection by dot hybridization using the 3zP-recombinant plasmid ~105 in LRBC fraction (I) and sera (II) from four animals at different days after inoculation. Two pigs were inoculated with the virulent virus (A, B) and two with the attenuated virus (C, D).
76
The intensity of the spots both in immunodot and in dot hybridizations was consistently lower in samples from animals inoculated with attenuated ASFV than in samples from animals inoculated with virulent ASFV. Finally, the induction of ASFV antibodies was measured by ELISA in the four inoculated animals. Specific ASFV antibodies were detected only in the surviving pig, inoculated with the attenuated virus, from day 7 after inoculation; maximal antibody level was at day 9 in the first 12 days of viraemia (data not shown).
Discussion ASFV detection in blood samples has been described in this report both by antigen and DNA-hybridization methods and their sensitivities have been compared. Other authors have previously described antigen (Crowther et al., 1979, Wardley et al., 1980) and ASFV-DNA (Caballero et al., 1986, Tabares, 1987) detection. All of these authors agree that the haemadsorption technique (Malmquist et al., 1960) is more sensitive for ASFV detection than tests based on viral antigen or DNA detection. However, both for routine and rapid sample screening, and for non haemadsorbing or noncytopathogenic strains of ASFV it is necessary to employ other techniques. The RIA technique has a detection limit of 50-500 HAD,, of virus (Wardley et al., 1980). We have shown similar sensitivities using the biotin-avidin ELISA sandwich, with the ID technique being even less sensitive. Previous studies (Tabares, 1987) demonstrated that 20 pg of viral DNA was detected by dot molecular hybridization using a “*P-probe with a specific activity about 2 x lo* cpm/u.g. We have detected 0.30 pg of viral DNA using a probe with a specific activity of 9 x 10’ cpm/pg. This 60-fold difference between detection capacity could be explained by the ratio of infective virus titrated by the HADso test and physical viral particles contained in the sample and the ability of dot molecular hybridization to measure both infective and non-infective viral particles. Similar sensitivities were obtained by other authors when infective virus was measured instead of purified viral DNA (Caballero et al., 1986), although our technique required only half the time of autoradiograph exposure. Our use of two recombinant plasmids as probes did not increase the sensitivity of viral DNA detection as suggested by other authors (Tabares, 1987). Viral detection in blood samples was found both in antigen and DNA hybridization methods, sooner in the post infection period for animals inoculated with virulent virus than for animals inoculated with attenuated virus. This observation agrees with those of other authors (Coggins et al., 1968; Mebus et al., 1978; Schlafer et al., 1984) showing viral titres in the blood of animals infected with attenuated virus lower than those infected with virulent virus. Virus could not be detected in the surviving pig by antigen detection methods after nine days of infection either associated with the LRBC fraction or in serum. However, by dot hybridization, virus was detected at least between the sixth and twelfth days. This could be explained by the presence of immunocomplexes that inhibit.antigen detection. The failure of antigen detection after day 10 is coinci-
77
dent with the peak of ASFV-antibody titre. Other authors (Sanchez Botija et al., 1977) have previously described the inhibition by immunocomplexes of viral detection using direct immunofluorescence. We therefore conclude that for rapid ASFV detection in acute disease, both antigen and DNA hybridization methods can be used with similar efficiency. However, in subacute ASF the viral DNA detection will be much more useful to avoid the problem of false negative diagnosis due to the presence of viral immunocomplexes from about nine days after infection, when high antibody titres are produced. In the future, when enough ASFV DNA sequence is available to design the appropriate primers, the polymerase chain reaction (PCR) technique will surely provide the highest sensitivity for examination of subacute infections by DNA hybridization.
Acknowledgements We are grateful to Dr J.K. Lunney for her useful discussion, and to M. Sevilla for technical assistance. This work was supported by a grant from Instituto Nacional de Investigaciones Agrarias, Madrid, Spain. References Alcaraz, C., Pasamontes, B., Ruiz Gonzalvo, F. and Escribano, J.M. (1989) Virology 168, 406-408. Caballero, R.G. and Taba&, E. (1986) Arch. Virol. 87, 119-125. Coggins, L., Moulton, J.E. and Colgrove, G.S. (1968) Cornell Vet. 58, 525-540. Crowther, J.R., Wardley, R.C. and Wilkinson, P.J. (1979) J. Hyg. 83, 353-361. Hess, W.R. (1981) Adv. Vet. Sci. Comp. Med. 25, 39-69. Malmquist, W.A. and Hay, D. (1960) Am. J. Vet. Res. 21, 104-108. Mebus, C.A., Dardiri, A.H., Hamdy, F.M., Ferris, D.H., Hess, W.R. and Callis, J.J. (1978) Proc. US Anim. Health Assoc. 82,232~236. Pini, A. and Wagenaar, G. (1974) Vet. Rec. 94, 2. Ruiz Gonzalvo, F., Caballero, C., Martinez, J. and Camero, M.E. (1986) Am. J. Vet. Res. 47, 18511862. Sachs, D.H., Leight, G., Cone, J., Schwarz, S., Stuart, L. and Rosenberg, S. (1976) Transplantation 22,559-567. Sanchez-Botija, C., Ordas, A., Gonzalvo, F. and Solana, A. (1977) Commission of the European Communities EUR 5904 EN, 642-652. Schlafer, D., Mebus, C.A. and McVicar, J.W. (1984) Am. J. Vet. Res. 45, 1367-1372. Taban%, E. (1987) Arch Virol. 92,233-242. Vigario, J.D., Terrinha, A.M. and Moura-Nunes, J.F. (1974) Arch. Ges. Virusforsch. 45, 272-277. Wardley, R.C. and Wilkinson, P.J. (1977) Arch. Viral. 55, 327-334. Wardley, R.C. and Wilkinson, P.J. (1980) Vet. Microbial. 5, 169-176. Wesley, R.D., Quintero, J.C. and Mebus, C.A. (1984) Am. J. Vet. Res. 45, 1127-1131.
Journal of Virological Methods, 28 (1990) 67-78
Elsevier VIRMET 1000
Evaluation of sensitivity of different antigen and DNA-hybridization methods in African swine fever virus detection M.J. Pastor and J.M. Escribano Depamnt
of Animal Health, National Institute for Agrarian Research, Madrid, Spain
(Accepted 5 December 1989)
ELISA, immunodot and DNA hybridization methods have been adapted to detect African swine fever virus (ASFV), and their sensitivities were compared using virus obtained from cell cultures. About 2.3 x l@ 50% hemadsorbing doses (HADso) of virus were detected with ELISA sandwich using an anti-ASFV IgG biotinylated followed by avidin-peroxidase. The immunodot technique showed similar sensitivity, detecting about 4.6 x l@ HADso of virus. ASFV-DNA was detected using radioactive DNA probes and molecular hybridization. The maximal viral detection capacity of this technique was about 1.8 x 103 HAD,,. The antigenic and DNA detection of ASFV during the infection of animals with virulent and attenuated viruses, was also studied. For this purpose, sera and red blood cells from several infected pigs were obtained at different days post-inoculation. The virus was detected at the third day after infection by the three methods. However, ASFV-DNA detection was more efficient than antigenic detection at nine days post-inoculation, when antigen detection failed, because immunocomplexes with circulating viruses were formed in the subacute infection. African swine fever virus; ELBA;
Immunodot;
DNA hybridization
Correspondence ro: J.M. Escribano, Department of Animal Health, National Institute for Agrarian Research, Embajadores 68, 28012 Madrid, Spain.
0166-0934/90/$03.50 0 1990 Ekvier
Science Publishers B.V. (Biomedical Division)
68
Introduction African swine fever virus (ASFV) is a DNA virus that infects domestic pigs. ASF is one of the most important problems of the pig industry in affected countries. The lack of a vaccine makes the diagnostic procedures the only methodology that can help to plan the complete eradication of the disease in these countries. African swine fever occurs in forms varying from peracute to chronic and inapparent. In the peracute and acute forms the virus is present in large amounts in the blood and tissues such as the spleen, kidneys, lymph nodes and lungs. About 17% of the samples from outbreaks, submitted for testing in Spanish laboratories, are examined for detection of virus (data not published). In addition, nonhemadsorbing (Pini et al., 1974; Vigario et al., 1974; SanchezBotija et al., 1977) or noncytopathogenic strains of ASFV may cause problems in detection by virus cultivation in pig buffy coat cultures, the only way, at present, to accomplish ASFV isolation (Sanchez Botija et al., 1977, Hess, 1981). At present, there are well characterized tests for animals that produce antibodies following subacute ASFV infections. In contrast, a few previous reports have described antigenic (Crowther et al., 1979; Wardley et al., 1980) and ASFV-DNA detection (Caballero et al., 1986, Tabarts, 1987). No comparison of these methodologies has been performed to evaluate the sensitivity of ASFV detection in vitro and in infected pigs. In this report we have adapted several ELISA and immunodot (ID) techniques, and their sensitivities were compared to the use of viral DNA probes, both in vitro and in infected pigs.
Materials and Methods Cell, viruses and sera For in vitro assays the Spain-70 (E& strain of ASFV was used after forty-six passages in a monkey stable (MS) cell line (Alcaraz et al., 1989). Four different virus isolates were used to test the stability of cloned viral DNA fragments used as DNA probes. The viruses were ho, ET5, Em and EM strains isolated in different years and geographical areas of Spain. For pig inoculation, two different viruses were used. The wild-type ET0 strain as virulent virus, and the E,, strain, after four passages in CV, cells (ATCC, CCL 70) (E,s-CVt4), was used as attenuated virus. The first virus produces an acute disease in all the infected pigs, and the second virus produces a subacute disease in a high percentage of infected pigs. The viral titres were obtained by the haemadsorption test (Malmquist et al., 1960), and expressed in 50% haemadsorbing dose (HADso). The antiserum for the ASFV antigenic detection assays was obtained from five pigs that had been immunized with the nonvirulent strain of ET0 ASFV obtained after 46 passages in MS cells (ET0 MS&. The pigs were challenged with the ho-
69
mologous virulent virus and bled after viraemia had subsided (Ruiz-Gonzalvo et al., 1986). Each antiserum, before being pooled, presented ELISA titres higher than 1:20000. The IgG fraction of the pool was obtained by precipitation with ammonium sulphate. Pig inoculation
Four NIH minipigs (Sachs et al., 1976) of SLA ap weighing about 20 kg each were divided into two groups and held in separate isolated boxes for the duration of the experiment. Two animals were infected by intramuscular infection of 10r HADSo of E,,, virus, and two with the same dose of E,5 CV,4 virus. Peripheral blood was collected prior to inoculation and on subsequent days until the pig died, or until 12 days after inoculation in one surviving pig, given the attenuated virus. The serum was separated from each sample and the coagulated mass was treated with 2 volumes of distilled water to lyse the red blood cells with the associated virus (Wardley et al, 1977; Wesley et al., 1984). The recovered supematant was called lysed red blood cells (LRBC) fraction. Semipurification
of viral particles
MS cells were infected with E,,, h4S4s virus at a multiplicity of infection of about 2. Supematant was collected 24 h after infection and clarified by centrifugation at 1000 x g for 15 min. The supematant was ultracentrifuged through a cushion of 20% (W/W) sucrose in 50 mM Tris-HCl, pH 8, at 100000 x g for 75 min. The viral sediment was titrated by inoculation in huffy coat cultures. Preparation of viral DNA
Viral sediments, sera (200 l.~l)or LRBC fraction (200 ~1) were made to 10 mM EDTA, 1% SDS and 500 &ml proteinase K and incubated for 30 min at 45°C. The samples were deproteinized with phenol:chloroform:isoamylalcohol(25:24:1) and viral DNA precipitated with 2.5 volumes of ethanol and resuspended in TE buffer. Viral DNA probes
Two fragments of ASFV-DNA from the E,a MSM strain were used in the construction of the recombinant plasmids. The fragments were obtained after digestion of viral DNA with Hind111 endonuclease, and they were inserted into pACYC177 plasmid (Chang et al., 1978) at Hind111 site located in the kanamycin resistant region, using the E. coli C600 strain as host bacterium. The two Hind111 fragments cloned correspond to fragments E and F of E,c MS& viral DNA, with 4.1 and 3.9 Kpb, naming the recombinant plasmids ~161 and ~105, respectively (Fig. 1). The 32P-labeled probes were prepared by nick translation following the instructions supplied with kit (Bethesda Research Laboratories, U.S.A.), and used
70
E F PACYC 177
Fig. 1. Recombinant plasmids ~161 and ~105 obtained with the E and F Hind111 fragments of ASFVDNA inserted in the kanamycin resistance region of pACYC 177 plasmid. The figure shows an EtBr staining gel of the ASFV-DNA compared with recombinant plasmids ~161 and ~105 digested with HindIII.
with a specific activity of 9 X 10’ cpm/pg. The concentration of the probes in the hybridization was 0.33 Fg per 2.5 ml of hybridization buffer. Hybridization
technique
Viral DNA was filtered through a nitrocellulose filter presoaked in 2 x SSC using a biodot vacuum manifold. The hybridization process was performed at 55°C in plastic bags as described (Tabares, 1987). Briefly, the filter was rinsed in 2 x SSC, dried and baked at 80°C for 2 h. Before hybridization, the filter was incubated at 55°C for 1 h in pre-hybridization buffer (30% formamide, 6 x SSC, 0.1% SDS and 2 mg/ml each of bovine serum albumin, Ficoll and polyvinylpyrrolidone). For hybridization the filters were incubated at 55°C for 20 h in plastic bags with the labeled probe in pre-hybridization buffer containing 25 pg/ml of salmon sperm DNA. After hybridization the filters were rinsed three times in 2 x SSC at room temperature and once in pre-hybridization buffer at 55°C for 15 min. The filters were finally washed in 2 x SSC and dried. The hybrids were visualized by autoradiography. ELISA
and immunodot
assays
A 140 dilution of the anti-ASFV IgG in phosphate buffered saline (PBS) was used to coat microtitre plate wells. After 12 h at 4°C the microtitre plates were
71
washed with PBS-0.05% Tween 20 and serial two fold dilutions in PBS of samples including semipurified ASFV, sera, or LRBC fraction were added. After 1 h of incubation the immunocomplexes were developed by two different procedures. In the first anti-ASFV IgG labeled with peroxidase was added at 1:lOO dilution for 1 h. For the second labeled biotin anti-ASFV-IgG (Kodama et al., 1987) at the same dilution was reacted with the samples for 1 h and after washing, avidin-peroxidase was added at a working dilution in PBS-Tween 20 for 1 h. In both procedures the microtitre plates were then washed and substrate, orthophenylene diamine (OPD, Sigma, St. Louis, U.S.A.), was added. The reaction was stopped with 3 N sulphuric acid after 5 min. The immunodot assay (ID) was performed on the same samples used in ELISA assay by filtering through a nitrocellulose filter, using a biodot vacuum manifold (BioRad Laboratories, California, U.S.A.). The filter was washed in PBS and dried, and the immune reactions carried out using the same methodologies employed above which involved the use of anti-ASFV IgG labeled with peroxidase, or labeled with biotin followed by avidin-peroxidase. The presence of immunocomplexes was developed with the 4-chloro-1-naphthol technique (Hawkes et al., 1982).
Sensitivityof antigenic ASFV detection assays Supernatants of infected MS cells were used to semipurify the virus. It was titrated by the haemadsorption technique and used for antigenic detection experiments. Two ELISA sandwich techniques, using an anti-ASFV IgG labeled either with peroxidase or biotin followed by avidin-peroxidase, were adapted to detect ASFV obtained from cell cultures. The biotin-avidin procedure shows more sensitivity, detecting about 2.3 x 16L HADw, of virus, while anti-ASFV IgG labeled with peroxidase procedure detected 9.3 X 16L HADSo of virus (Fig. 2). The ID technique, directly fixing the virus to the filter and using both procedures employed in ELISA, detected about 4.6 X ld HADSO with biotin followed by avidin peroxidase and 9.3 x l@ HADSo with anti-ASFV IgG labeled with peroxidase methodology. The intensity of the spots for the biotin-avidin method was higher than anti-ASFV IgG labeled with peroxidase (Fig. 3). Sensitivityof ASFV-DNA detection assays Dot hybridizations were carried out using two different recombinant plasmids, ~161 and ~105 prepared with the E and F Hind111 fragments of ASFV-DNA of the E,,, MS, strain. The stability of the viral DNA fragments used as probes was determined by hybridization with four wild-type virus strains (E,a, E7s, E,, and E,&. Both probes resulted in a similar autoradiograph signal (Fig. 4, panel 1).
72
2.0
5: t d 6
1.0
0.1 3x10'
9.3~102
5.8X10
HAD5O Fig. 2. Comparison of detection of semipurified ASFV from infected cell culture supernatants, by two different ELBA methods. Serial ASFV dilutions were first reacted with anti-ASFV IgG adsorbed to the ELISA plate, and then with the same IgG labeled with peroxidase (0) or with biotin followed by avidin-peroxidase (a). Normal pig IgG, labelled with peroxidase (A) or with biotin (V) were used as controls. Absorptton at 0D4s0 was measured 5 min after addition of substrate.
Fig. 3. Comparison of detection of semipurified ASFV from infected cell culture supernatants by two different immunodot procedures. Serial ASFV dilutions were adsorbed to the nitrocellulose filter and reacted with anti-ASFV IgG labeled with peroxidase (1) or with biotin followed by avidin-peroxidase (2). The reactivity of both procedures was measured by densitographic analysis after substrate addition.
73
Fig. 4. Dot hybridixations of ASFV-DNA using the ‘*P-recombinant plasmids ~161 and ~105. Hybridizations were carried out using a concentration of the probes in the hybridization of 0.33 ug per 2.5 ml of hybridization buffer, employing 3 x 10’ cpm in each assay. Panel 1. Hybridization of ~105 and ~161 plasmids on nitrocelhdose filter with the DNA from 3 X 10’ HADW of four different ASFV isolates E,,,, E15, Ee,r and EM (1-t) after 24 h of autoradiographic exposure. Panel 2. Fixation of serial 2-fold dilutions of ASF’V-DNA starting with the viral DNA from 3 X lo’ HADs,, to nitrocellulose (A), Zeta-Probe (B) and Hybond-N (C) filters. The hybridization with ~105 plasmid was autoradiographed for 72 h. Panel 3. Hybridization of ~105 (A) and ~161 (B) plasmids with the ASFV-DNA diluted as described for panel 2, on nitroceUulose titters autoradiographed at 24 h (1) and 72 h (2). Panel 4. Hybridization of a mixture of ~105 and ~161 plasmids with a final specitic activity of 9 x 10’ cpm/pg, with the ASFV-DNA diluted as described in panel 2 and autoradiographed for 24 h.
Different filters were selected to analyze the capacity for adsorption of viral DNA. After hybridization, higher viral DNA was detected on the nitrocellulose filter (BioRad Laboratories, 0.45 pm, California, U.S.A.) than on nylon filters, Zeta-Probe (BioRad Laboratories) and Hybond-N (Amersham, 0.45 pm, England) (Fig. 4, panel 2). The sensitivity of viral DNA detection was assayed using each recombinant plasmid separately. After 24 and 72 h of autoradiography exposure viral DNA detection was the same with both probes, detecting about 1.8 x 103 HADSO at 72 h, which corresponds to 0.30 pg of viral DNA. Twenty-four hours exposure resulted in detection of 3.6 X 103 HAD% (Fig. 4, panel 3). When ~161 and ~105 plasmids were used jointly in the hybridization, at the same specific activity, the sensitivity of the technique in viral DNA detection was not increased, when compared to either of the plasmids alone (Fig. 4, panel 4).
A
2.0
s
d
B
I.0
d
0.1 03456789101112
DAYS AFTER
03456769101112
INFECTION
Fig. 5. ASFV detection by ELISA sandwich using the biotin-avidin procedure at different days after inoculation in the LRBC fraction (A) and in sera (B) from four ASFV-inoculated animals. Two pigs were inoculated with lo3 HAD,, of virulent virus (V,A) and two with the same dose of attenuated virus (@,o).
Comparison
of viral antigens and viral-DNA detection in ASFV
infected animals
Blood samples from two ASFV-infected pigs with the virulent ET0 strain (pigs A and B) and two with attenuated E,, CV14 virus (pigs C and D) were used to compare ASFV detection. By ELISA sandwich, using the biotin-avidin procedure, the virus was detected both in serum and in the LRBC fraction in animals inoc-
Fig. 6. ASFV detection by immunodot assay using biotin-avidin procedure from four inoculated animals: Pigs A and B were inoculated with virulent virus and C and D with the attenuated virus.
75
ulated with virulent and attenuated ASFV. Antigen detection was performed in one pig (A) both in serum and LRBC fraction at day 3 after inoculation and at day 4 in the other inoculated pig with the same virus (B) (Fig. 5). In the animals inoculated with the attenuated virus, it was detected in the pig that subsequently died (C) at day 4 after inoculation in the LRBC fraction, and at day 6 in the serum of the same animal (Fig. 5). In the other surviving pig (D), low amounts of viral antigens were detected in serum and in LRBC between 5 to 7 days after inoculation (Fig. 5). By the immunodot technique using the biotin-avidin procedure only the sera of infected pigs could be tested because of the backgrounds obtained with haemoglobin contamination from red blood cell lysates. In the sera of pigs inoculated with the virulent ASFV, virus was detected at days 4 and 5 by this ID technique (Fig. 6, rows A,B). In the case of attenuated ASFV, the virus was detected in serum only at day 6 in one pig and from days 6 to 9 after infection in the other (Fig. 6, row C, D). By dot hybridization using the ~105 32P-probe, ASFV-DNA was detected after 72 h of autoradiography, in pigs inoculated with virulent virus from days 3 to 5 in the LRBC fractions and at days 4 and 5 in the sera (Fig. 7). In the animals inoculated with attenuated ASFV, the viral DNA was detected in the LRBC fraction from days 3 to 6 in one pig (C) and from days 6 to 12 in the other (D) (Fig. 7). In the sera of animals inoculated with attenuated ASFV, viral DNA was detected in one pig (C) at day 6 only and in the other (D) from day 7 to the end of day 12 (Fig. 7).
Fig. 7. ASFV detection by dot hybridization using the 3zP-recombinant plasmid ~105 in LRBC fraction (I) and sera (II) from four animals at different days after inoculation. Two pigs were inoculated with the virulent virus (A, B) and two with the attenuated virus (C, D).
76
The intensity of the spots both in immunodot and in dot hybridizations was consistently lower in samples from animals inoculated with attenuated ASFV than in samples from animals inoculated with virulent ASFV. Finally, the induction of ASFV antibodies was measured by ELISA in the four inoculated animals. Specific ASFV antibodies were detected only in the surviving pig, inoculated with the attenuated virus, from day 7 after inoculation; maximal antibody level was at day 9 in the first 12 days of viraemia (data not shown).
Discussion ASFV detection in blood samples has been described in this report both by antigen and DNA-hybridization methods and their sensitivities have been compared. Other authors have previously described antigen (Crowther et al., 1979, Wardley et al., 1980) and ASFV-DNA (Caballero et al., 1986, Tabares, 1987) detection. All of these authors agree that the haemadsorption technique (Malmquist et al., 1960) is more sensitive for ASFV detection than tests based on viral antigen or DNA detection. However, both for routine and rapid sample screening, and for non haemadsorbing or noncytopathogenic strains of ASFV it is necessary to employ other techniques. The RIA technique has a detection limit of 50-500 HAD,, of virus (Wardley et al., 1980). We have shown similar sensitivities using the biotin-avidin ELISA sandwich, with the ID technique being even less sensitive. Previous studies (Tabares, 1987) demonstrated that 20 pg of viral DNA was detected by dot molecular hybridization using a “*P-probe with a specific activity about 2 x lo* cpm/u.g. We have detected 0.30 pg of viral DNA using a probe with a specific activity of 9 x 10’ cpm/pg. This 60-fold difference between detection capacity could be explained by the ratio of infective virus titrated by the HADso test and physical viral particles contained in the sample and the ability of dot molecular hybridization to measure both infective and non-infective viral particles. Similar sensitivities were obtained by other authors when infective virus was measured instead of purified viral DNA (Caballero et al., 1986), although our technique required only half the time of autoradiograph exposure. Our use of two recombinant plasmids as probes did not increase the sensitivity of viral DNA detection as suggested by other authors (Tabares, 1987). Viral detection in blood samples was found both in antigen and DNA hybridization methods, sooner in the post infection period for animals inoculated with virulent virus than for animals inoculated with attenuated virus. This observation agrees with those of other authors (Coggins et al., 1968; Mebus et al., 1978; Schlafer et al., 1984) showing viral titres in the blood of animals infected with attenuated virus lower than those infected with virulent virus. Virus could not be detected in the surviving pig by antigen detection methods after nine days of infection either associated with the LRBC fraction or in serum. However, by dot hybridization, virus was detected at least between the sixth and twelfth days. This could be explained by the presence of immunocomplexes that inhibit.antigen detection. The failure of antigen detection after day 10 is coinci-
77
dent with the peak of ASFV-antibody titre. Other authors (Sanchez Botija et al., 1977) have previously described the inhibition by immunocomplexes of viral detection using direct immunofluorescence. We therefore conclude that for rapid ASFV detection in acute disease, both antigen and DNA hybridization methods can be used with similar efficiency. However, in subacute ASF the viral DNA detection will be much more useful to avoid the problem of false negative diagnosis due to the presence of viral immunocomplexes from about nine days after infection, when high antibody titres are produced. In the future, when enough ASFV DNA sequence is available to design the appropriate primers, the polymerase chain reaction (PCR) technique will surely provide the highest sensitivity for examination of subacute infections by DNA hybridization.
Acknowledgements We are grateful to Dr J.K. Lunney for her useful discussion, and to M. Sevilla for technical assistance. This work was supported by a grant from Instituto Nacional de Investigaciones Agrarias, Madrid, Spain. References Alcaraz, C., Pasamontes, B., Ruiz Gonzalvo, F. and Escribano, J.M. (1989) Virology 168, 406-408. Caballero, R.G. and Taba&, E. (1986) Arch. Virol. 87, 119-125. Coggins, L., Moulton, J.E. and Colgrove, G.S. (1968) Cornell Vet. 58, 525-540. Crowther, J.R., Wardley, R.C. and Wilkinson, P.J. (1979) J. Hyg. 83, 353-361. Hess, W.R. (1981) Adv. Vet. Sci. Comp. Med. 25, 39-69. Malmquist, W.A. and Hay, D. (1960) Am. J. Vet. Res. 21, 104-108. Mebus, C.A., Dardiri, A.H., Hamdy, F.M., Ferris, D.H., Hess, W.R. and Callis, J.J. (1978) Proc. US Anim. Health Assoc. 82,232~236. Pini, A. and Wagenaar, G. (1974) Vet. Rec. 94, 2. Ruiz Gonzalvo, F., Caballero, C., Martinez, J. and Camero, M.E. (1986) Am. J. Vet. Res. 47, 18511862. Sachs, D.H., Leight, G., Cone, J., Schwarz, S., Stuart, L. and Rosenberg, S. (1976) Transplantation 22,559-567. Sanchez-Botija, C., Ordas, A., Gonzalvo, F. and Solana, A. (1977) Commission of the European Communities EUR 5904 EN, 642-652. Schlafer, D., Mebus, C.A. and McVicar, J.W. (1984) Am. J. Vet. Res. 45, 1367-1372. Taban%, E. (1987) Arch Virol. 92,233-242. Vigario, J.D., Terrinha, A.M. and Moura-Nunes, J.F. (1974) Arch. Ges. Virusforsch. 45, 272-277. Wardley, R.C. and Wilkinson, P.J. (1977) Arch. Viral. 55, 327-334. Wardley, R.C. and Wilkinson, P.J. (1980) Vet. Microbial. 5, 169-176. Wesley, R.D., Quintero, J.C. and Mebus, C.A. (1984) Am. J. Vet. Res. 45, 1127-1131.
