Abstract: At present, there is a risk of cross-infection caused by blood collection in the commonly used serological detection methods of African swine fever virus (ASFV) antibodies, and the serological detection sensitivity is affected by the relatively delayed serological conversion of serum antibodies. Therefore, it is important to establish a non-invasive sampling serological method that can realize early diagnosis and monitoring of ASFV infection. In this study, pFastbac1-P30-His recombinant expression vector was constructed by artificial synthesis. The recombinant ASFV P30 protein (SUMO-P30) was expressed by baculovirus-insect cell eukaryotic expression system, purified by Ni column affinity and used as the coating antigen. After a series of conditions optimization, an indirect ELISA method of ASFV antibody targeting the mucosal sIgA antibody in porcine oral fluid was established. The results showed that the eukaryotic recombinant P30 protein (SUMO-P30) was about 47 ku, and Western blot analysis showed that it had good reactivity; The optimal reaction condition of ELISA was determined after optimization: the coating amount was 1.0 μg·mL^－1, 5% skimmed milk was the best sample dilution, the best mixing ratio with the oral liquid to be tested was 2∶8, the sample reaction time was 120 min, the best dilution of rat anti-pig IgA-HRP was 1∶5 000, the reaction time was 60 min, and the best color development time was 15 min. The antibody titer of ASFV infected oral fluid detected by this method could reach 1∶32, and there was no cross reaction with CSFV, PRV, PRRSV mucosal antibody positive oral fluid, which indicated its good sensitivity and specificity. By using this method and the ASFV commercialized ELISA serum antibody detection kit, the oral fluid and serum samples of the same pig infected with ASFV virulent or artificially attenuated strains were detected at different time points. The mucosal sIgA antibody in the oral fluid showed a significant increase in S/P values 3-5 days after infection, while serum antibodies were not detected to be positive during the monitoring period. Oral fluid and serum samples from challenged or indirectly infected piglet with natural variant ASFV strains were detected at different time points post infection. The positive coincidence rate of this method with the commercial ASFV antibody detection kit was 100%, the negative coincidence rate was 37.5%, and the total coincidence rate was 61.5%. In conclusion, this study has established an ELSIA detection method of ASFV mucosal antibody with oral fluid as sample without blood collection, which can realize non-invasive early diagnosis of ASFV infection. This method provides new technical support and supplement for ASFV monitoring, prevention and control.

Key words: African swine fever, sIgA antibody, ELISA, early diagnosis