**4. Materials and Methods**

## *4.1. Samples, RT-PCR, Phylogenetic Tree, and Virus Isolation*

Blood samples from pigs (30–90 days old) with suspected CSF were collected from 16 pig farms from June 2014 to December 2018. Samples were collected from farms in Xuan Truong-Nam Dinh, Hi Duong, Nghe An, Hung Yen, Bac Giang, My Xa-Nam Dinh, and Ninh Giang-Hai Duong (Table 1 and Figure 1). None of the pigs on the 16 pig farms were vaccinated. Total RNA was extracted from blood using a micro column-based QIAamp Viral RNA Mini kit (Qiagen, USA) to identify CSFV. The RT-PCR conditions and specific primers used to amplify the complete E2 gene have been reported previously [4]. The complete E2 gene sequences of CSFVs (including the Vietnam strains examined in this study) were obtained from the NCBI GenBank database and aligned using the CLUSTAL X alignment program. Next, a BEAST input file was generated using BEAUti within BEAST package v1.8.1 [20]. Rates of nucleotide substitution per site per year and the tMRCA were estimated using a Bayesian MCMC approach. The exponential clock and expansion growth population model in the BEAST program was used to obtain the best-fit evolutionary model and the MCC tree was visualized using Figtree 1.4 [21]. For virus isolation, CSFV-positive blood samples were inoculated to porcine kidney 15 (PK-15) cell (grown to 80% confluence in 6-well plates) using alpha-minimum essential medium (GIBCO Cat. No. 12571-063) with L-glutamine (GIBCO Cat. No. 25030-081), sodium pyruvate (GIBCO Cat. No. 11360-070), and antibiotic-antimycotic solution (GIBCO Cat. No. 15240-062). The six-well plate (PK-15 cells) were inoculated for 3–5 days at 37 ◦C and the virus was identified by immunochemical staining using the CSFV monoclonal antibody 3B6 (Median Diagnostics Co., South Korea).

#### *4.2. Animal Experiments*

After virus inoculation, pigs were monitored to confirm the pathogenicity of two isolated sub-genotypes (2.2 and 2.1c). ND20 strain (sub-genotype 2.2) and HY58 strain (sub-genotype 2.1c) were each inoculated into pigs (30 days old) via two simultaneous administration (2 mL orally and 2 mL intramuscularly) (105.0 TCID50/mL). GPE<sup>−</sup> strain, a CSFV vaccine strain used in Japan, was inoculated via the same route and dose as a control. Pigs were divided into group 1 (n = 2, GPE− strain), group 2 (n = 5, ND20 strain), and group 3 (n = 5, HY58 strain), and observed for 30 days. To detect CSFV RNA copies and to check for leukopenia (leucocyte counts below 9000/μL), blood samples were collected at 0, 3, 5, 7, 10, 14, and 21 dpi. During the course of the experiment, pigs were monitored daily for clinical signs (anorexia, diarrhea, dehydration, tremble, congestion, conjunctivitis, and hind leg paralysis). The clinical score was determined by ten parameters following previous procedure [22]. Briefly, ten parameters (liveliness, body tension, body shape, breathing, walking, skin, eyes, appetite, defecation, and leftovers in feeding trough) were graded according to the following scoring system: 0, normal; 1, slightly altered; 2, showing distinct clinical signs; and 3, showing severe CSF clinical signs. Virus strains were classified as highly virulent (total clinical score: >15), moderately virulent (5–15), low virulent (<5), or avirulent (0) [22]. Rectal temperature was checked at the time of blood collection. Pigs were necropsied and ten organs (lung, heart, liver, spleen, tonsil, lymph node, intestinal, kidney, bladder, and central nervous system) were examined to detect histopathologic lesions and the presence of CSFV antigens. Collected tissues were processed routinely for histopathologic examination and stained with hematoxylin and eosin (H&E). Mortality was calculated over the 30-day observation period.
