The amplification of microbial nucleic acids from serum was based on a modified version of the SISPA method (Allander et al. 2001; Ambrose and Clewley 2006; Jones et al. 2005). After filtration through a 0.2-μm filter and digestion with turbo DNase (Ambion), benzonase (Novagen) and RNase One (Promega), RNA was extracted from serum (140 μL) of patient HB29. RNA preparations were reverse transcribed into cDNA and second strand cDNA were synthesized. After purification, the DNA was ligated to a phosphorylated blunt adapter E19 (5′-AGCAATTCCGTTGCTGTCG-3′); and E12 (5′-P-GGCAACGACAGC-3′). The ligation product was PCR amplified and separated by agarose gel electrophoresis. Fragments of different size were isolated and cloned. A total of 576 cDNA clones were picked by SISPA from the serum of patient HB29 and sequenced. After trimming to remove sequences derived from the amplification primer, the data set was subjected to homology search with the GenBank databases of nucleic acids and proteins using BLASTN and BLASTX. Whereas the nucleotide sequence was essentially unrelated to other sequences in the existing GeneBank database (http://blast.ncbi.nlm.gov), the deduced amino acid sequence from 14 cDNA clones showed 20–30% homology to viral proteins, including RNA dependent RNA polymerase (RdRP), membrane polyprotein and nonstructural S protein, of known phleboviruses such as Rift Valley fever virus and Uukuniemi virus. Walking primers from both 5′ and 3′ directions were designed based on the sequence of gene fragments obtained from SISPA. The three genomic segments were assembled from a series of overlapping cDNA clones. The 5′ and 3′ termini of viral RNA segments were determined by rapid amplification of cDNA ends (RACE) using reagents purchased from Invitrogen. To eliminate the influence of cloning bias on genome sequencing, direct PCR re-sequencing of all three genomic segments were performed with newly designed primers. The partial sequences were obtained from the first isolated virus strain DBM, and the complete genomes of 11 isolates of SFTS virus were further sequenced (Yu et al. 2011). All isolates including strain DBM were closely related (Yu et al. 2011). Termini of three genomic segments of SFTS virus were found to be similar to counterparts in other phleboviruses.
Phylogenetic analyses were performed with the neighbor-joining method using the Poisson correction and complete deletion of gaps. Bootstrap values were estimated from 2000 replicates (95% confidence) with a random seed. Phylogenetic trees based on complete viral genomic sequence of L, M and S segments from strains (HB29, HN6, AN12, LN2, JS3 and SD4) from 6 provinces in China in comparison with other known phleboviruses showed that SFTS virus was related to prototypic viruses of Phlebovirus. Phylogenetic analysis using the deduced amino acid sequences of RdRp, glycoproteins (Gn and Gc), N and NSs proteins, the generated phylogenetic tree showed all SFTS virus isolates clustered together, but are almost equally distant from other two groups (Fisher et al. 2003), including the sandfly fever group (Rift Valley fever, Punta Toro, Toscana, Massila, sandfly fever Sicilian viruses) and Uukuniemi group. The comparison of amino acid sequence similarity provided further support to the separation of SFTS virus from other phleboviruses. SFTS virus was assigned to be a novel species at the genus of Phlebovirus in Bunyaviriade in 2014, and was resigned to be a novel species at the genus of Phlebovirus in the family of Phenuiviridae by International Committee on Taxonomy of Viruses (ICTV) in 2016.