Evolution and genetic diversity of atypical porcine pestivirus (APPV) from piglets with congenital tremor in Guangxi Province, Southern China

Abstract Atypical porcine pestivirus (APPV) was identified and associated with congenital tremor (CT) type A‐II in new born piglets and has been reported in many countries. In China, the first APPV identification in swine herds was reported in Guangdong province in 2016. To investigate the genetic characteristics of APPV in Guangxi province, 53 tissue samples from neonatal piglets with CT were collected and detected from October 2017 to May 2019. Five APPV strains which were named as GX04/2017, GX01‐2018, GX02‐2018, GX01‐2019 and GX02‐2019 were obtained. Sequence analysis revealed that all six APPV strains from Guangxi province, including five strains from this study and one from a previous report, shared 83.3%‐97.5% nucleotide identity of complete genome and 91.7%‐99.1% amino acid identity of the open reading frame (ORF), and shared 77.7%‐97.7% nucleotide identity of complete genome and 90.6%‐99.3% amino acid identity of ORF with reference strains. Phylogenetic analysis indicated that all APPV strains could be divided into three clades based on the complete genome, Npro, Erns and E2 gene sequences, respectively; and the APPV strains from Guangxi province distributed in two clades (clades I and II). No sign of recombination was observed from Guangxi strains. Evolution analysis performed on the complete genome of 58 APPV strains showed that America, Europe and Asia strains during 2006–2019 evolved at a mean rate of 1.37 × 10–4 substitutions/site/year, and the most recent common ancestor (tMRCA) of them was estimated as 1,700.5 years ago. The findings of this study indicated that there existed a high degree of genetic diversity of APPV from Guangxi province, Southern China, which provided important information on the epidemiological features and evolutionary relationships of APPV.

The clinical presentation of APPV-infected pigs was characterized by CT type A-II in piglets, whereas adult pigs might become persistent carriers and shedders (de Groof et al., 2016;Schwarz et al., 2017). High APPV loads were detected by qRT-PCR in semen, serum and different tissue samples of infected pigs (Gatto et al., 2018;Liu et al., 2019;Schwarz et al., 2017). Besides horizontal transmission through oronasal pathway, APPV could also be vertically transmitted by transplacental infection (Arruda et al., 2016;de Groof et al., 2016). It could occur as a sporadic disease affecting single litters or as an outbreak over several weeks affecting multiple litters, and the losses of 2.5 piglets per sow and a 10% drop in pig reproductive performance was occurred on a farm in Austria (Schwarz et al., 2017). One retrospective study showed that 41.8% (51/122) sampled pig farms and 16.3% (182/1115) porcine serum samples could be detected antibodies against APPV in Germany (Michelitsch et al., 2019). Due to the abovementioned complicated epidemiology and the huge loss of infected swine herds, APPV has attracted great attention in many countries.
APPV is a newly discovered virus. Two independent studies have been reported that CT was reproduced following experimental inoculation with serum or tissue-homogenate-pools containing APPV (Arruda et al., 2016;de Groof et al., 2016), and APPV has been proven to fulfil Mokili's Metagenomic Koch's Postulates, so APPV is associated with CT type A-II (Stenberg et al., 2020). Some studies on APPV molecular epidemiology have found that there existed high genetic variation among different strains (Choe et al., 2020;Schwarz et al., 2017;Williamson, 2017;Wu et al., 2018;Zhang et al., 2017;Zhou et al., 2019), with up to 21% genetic distance among the viruses (Postel et al., 2017). However, APPV strains available for biological, origin and evolution analysis are still scarce so far, and it is inadequate for studies on understanding APPV genetic diversity and evolutionary relationships. This study was intended to investigate the evolutionary relationships and genetic diversity of APPV strains from Guangxi province, Southern China. The results would provide valuable information for research on epidemiological and evolutionary characteristics of APPV in China.

| Collection of clinical samples
From October 2017 to May 2019, newborn piglets with clinical CT were reported from eighteen farms in Guangxi province. Piglets presented CT soon after birth and some of them died within a week.
Fifty-three samples, including brain, liver, spleen and lymph node from each dead or sick piglet, were collected and submitted to our laboratory for diagnostic investigation. Identification of CT-associated viral pathogens was detected by PCR or RT-PCR for APPV, classical swine fever virus (CSFV), porcine circovirus type 2 (PCV2), porcine pseudorabies virus (PRV) and Japanese encephalitis virus (JEV).

| Detection and sequence determination of APPV genome
The tissue samples were homogenized in phosphate-buffered saline solution (PBS, pH7.2) and used to determine APPV presence by RT-PCR. Viral RNA was extracted from the tissue superna- To determine the complete genome, RT-PCR was used with eight pairs of specific primers (Table 1)

| Evolution and phylogenetic analysis of APPV genome
To determine the genetic characteristics of APPV, evolution and phylogenetic analysis was focused on the complete genome, N pro , E rns and E2 genes of 58 APPV strains from different countries (including 6 strains from Guangxi province, China) available in GenBank on February 29, 2020 (Table 4). Nucleotide and amino acid identity analysis were calculated using MegAlign program of the DNAStar package (DNASTAR, USA). Phylogenetic trees were constructed by MEGA X (http://www.megas oftwa re.net/).

Recombination evaluation was analysed using Recombination
Detection Program 4 (RDP4) and then confirmed with SimPlot software (version 3.5.1). Bayesian inference analysis was performed using BEAST software package (version 1.10.4) (http:// beast.commu nity), and the times of the most recent common ancestor (tMRCA) and mean rate of molecular evolution were calculated by BEAST software.

| Determination of APPV genome from positive samples
All the tissue samples collected from 53 piglets with CT were detected by PCR/RT-PCR, and all of them were negative for CSFV, PCV2, PRV and JEV, whereas 41 of them were positive for APPV (41/53, 77.36%).
Five APPV-positive samples were selected randomly for amplification and sequencing, and finally, the complete genome of five APPV strains were obtained. The complete genomes were 11 534-11 565 nucleotides (nt) in full-length, with a 5′ UTR of 358-378 nt, followed by a single large ORF and a 3′ UTR of 268-279 nt. The ORF was 10 908 nt in length, which encoded a polyprotein of 3 635 amino acids (aa). The polyprotein was composed of 12 proteins, including four structural proteins (C, E rns , E1 and E2) and eight nonstructural proteins (N pro , P7, N2, NS3, NS4A, NS4B, NS5A and NS5B) ( Table 2)

| Phylogenetic analysis of APPV genome
Phylogenetic trees were constructed using the complete genome, N pro , E rns and E2 gene sequences of 58 APPV strains available in GenBank (Table 4) and all APPV strains could be divided into three clades. The phylogenetic trees based on N pro , E rns and E2 gene sequences showed quite similar topology with that of the complete genome ( Figure 1).

| Evolution analysis of APPV strains
Recombination Detection Program 4 (RDP4) and Simplot 3.5.1 were used to analyse the recombination events of 58 APPV strains from different countries available in GenBank (Table 4) Evolutionary estimation based on the complete genome, N pro , E rns and E2 gene sequences of 58 APPV strains from different countries was conducted by Bayesian analysis. The results indicated that APPV genomic sequences evolved at a mean rate of 1.37 × 10 -4 (95% highest probability density, HPD: 5.12 × 10 -6 -3.02 × 10 -4 ) substitutions/site/year (s/s/y), and the mean rates of molecular evolution of N pro , E rns and E2 gene sequences were 1.00 × 10 -4 (8.47 × 10 -5 -1.17 × 10 -4 ), 1.56 × 10 -4 (9.81 × 10 -5 -2.20 × 10 -4 ) and  detected in serum, thymus, peripheral lymphoid organs (spleen, tonsil, submaxillary lymph node and inguinal lymph node), nervous system (brain stem, brain and cerebellum), digestive system (duodenum) and semen (Gatto et al., 2018;Hause et al., 2015;Munoz-Gonzalez et al., 2017;Postel et al., 2017;Yuan et al., 2017), which indicated that this virus has widespread tissue tropism. However, it was very difficult for this virus to culture and identify. Many researchers have attempted to seek appropriate cell lines to culture and acquire a higher titre of APPV for identification and characterization, but unfortunately, no one has been successful until now because this virus could not replicate in the selected cell lines or the viral titre was too low (Cagatay et al., 2018;Gatto et al., 2018;de Groof et al., 2016;Hause et al., 2015;Postel et al., 2016;Schwarz et al., 2017). Until now, all the APPV sequences obtained were directly amplified, cloned and sequenced from positive serum and tissue samples (de Groof et al., 2016;Hause et al., 2015;Pan, Yan, et al., 2019;Schwarz et al., 2017;Shen et al., 2018;. Therefore, we also amplified and sequenced the APPV genome of all five Guangxi strains from clinical tissue samples and used to analyse their evolution and genetic diversity in this study.  (Beer et al., 2017;Postel et al., 2016). The genomic sequences of Austrian APPV shared 90% nucleotide identity with those from the United States and 92% nucleotide identity with those from Germany (Schwarz et al., 2017 To our knowledge, this is the first report on the estimation of the mean rate of molecular evolution and dates of the tMRCA about APPV. The results revealed that APPV complete genome, N pro , E rns and E2 gene evolved at an evolutionary rate of 1.37 × 10 -4 1.00 × 10 -4 , 1.56 × 10 -4 and 1.01 × 10 -4 s/s/y, respectively, and their tMRCA were estimated as 1,700.5, 1,495.5, 1,390.8 and 1823.0 years ago, respectively, which indicated that N pro , E rns and E2 genes had similar evolutionary rate and tMRCA with the complete genome. Some previous studies had been done on the evolution of other pestiviruses. A mean substitution rate of 1.4 × 10 -3 s/s/y was found across both bovine viral diarrhoea virus 1 (BVDV-1) and BVDV-2 and the tMRCA was estimated to be 1,498 years ago (Chernick & Meer, 2017); the mean evolutionary rate of the 5' UTR sequences of border disease virus (BDV) was 2.9 × 10 −3 s/s/y (Luzzago et al., 2016). The full-length E2 gene of CSFV from different countries had an evolutionary rate of 3.2 × 10 -4 s/s/y and the origin of CSFV was estimated to be the mid-1500s (Garrido Haro et al., 2018), whereas the whole CSFV genome evolved at a rate of 1.03 × 10 -4 s/s/y and the tMRCA appeared 2,770.2 years ago (Kwon et al., 2015). As for other highly variable RNA virus, such as porcine reproductive and respiratory syndrome virus (PRRSV) and swine influenza virus (SIV), some studies reported that the evolutionary rate of PRRSV genome ranged from 1.98 × 10 -3 to 3.29 × 10 -3 s/s/y (Song et al., 2010;Yoon et al., 2013) and that of SIV HA gene ranged from 1.03 × 10 -3 to 3.18 × 10 -3 s/s/y (Al Khatib et al., 2018;Wei et al., 2016). These results showed that the evolutionary rate of APPV was similar to that of CSFV, and lower than those of other RNA viruses. This study might provide a complementary reference for evolutionary information of APPV.