The complete genome of Banana streak GF virus Yunnan isolate infecting Cavendish Musa AAA group in China

Materials

Banana leaves showing streak symptoms were collected by permissions of a farmer (Shao-cheng Shen) from Yunnan, China in 2009. The banana belongs to the Cavendish Musa AAA group. Total DNA was extracted and stored at −80 °C. Previous study showed that the sample was infected with Banana streak virus Acuminata Yunnan (BSV-Acum), and its complete genome was obtained and analyzed in our laboratory (Zhuang et al., 2011). In order to further determine whether other BSV species or strains co-infect in the same plant, the complete genome amplification based on PCR method was conducted on the total DNA.

Cloning and sequencing of viral genome

According to the complete genome sequence of BSGFV deposited in NCBI (GenBank accession number NC_007002.1), seven specific pairs primers were designed by Primer Premier 5.0 (Table 1). The primers were synthesized by Invitrogen (Guangzhou, China) Trading Co., Ltd. The PCR reaction system (50 µL) contained EasyTaq® DNA Polymerase 1 µL (Beijing, China), 2.5 mM dNTPs 4 µL, 10 ×EasyTaq buffer 5 µL, 5 µM forward and reverse primers 4 µL, total DNA template 2 µL, and added ddH₂O to 50 µL. PCR reaction condition was pre-denaturation at 94 °C for 3 min; followed by 35 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 90 s; and a final extension step at 72 °C for 10 min. The 6 µL PCR product was analyzed by electrophoresis in 1.5% agarose gel, and the remaining PCR product was purified by an OMEGA gel recovery kit (Bio-Tek, USA). The purified PCR product of each DNA fragment was cloned into the pMD18-T vector (Takara, China), and then transformed into E. coli DH5a competent cells (2nd Lab, Shanghai, China). Three positive clones were randomly selected for Sanger sequencing at Invitrogen (Guangzhou, China).

Complete genome sequence assembly and analysis

The seven overlapping fragments were edited by ChromasPro software (Technelysium Pty. Ltd., Australia) and were used to assemble into the complete genome of BSGFV-YN by BioEdit software (Hall, 1999). The genome was further analyzed by Blastn and Blastx at NCBI website (http://blast.ncbi.nlm.nih.gov/Blast.cgi), and the possible species or strain of BSV was preliminarily identified. Based on the classification of ICTV (2018b release), 37 badnaviruses were downloaded from GenBank for further comparison and analysis (Table 2). The open reading frames (ORFs) of BSGFV-YN were predicted by ORF Finder online website (https://www.ncbi.nlm.nih.gov/orffinder/) and bioinformatic analysis. Subsequently, the sequence similarities of complete genomes, ORFI∼III and their putative proteins between BSGFV-YN with other 15 badnaviruses were analyzed by Sequence Identity Matrix Program in BioEdit software. The movement protein (MP), coat protein (CP), aspartic protease (AP), reverse transcriptase (RT), ribonuclease H (RNaseH) and two cysteine-rich domains in BSGFV-YN ORFIII were predicted according to the report by Li et al. (2018), and sequence similarities with other homologous proteins were aligned by the GeneDoc software. The repeated sequences in the intergenic regions between ORFIII and ORFI of BSGFV-YN and BSGFV goldfinger (NC_007002.1) were also analyzed by GeneDoc software.

Phylogenetic analysis

In order to further determine the homology and evolutionary relationships of BSGFV-YN with other badnaviruses, two phylogenetic trees based on the complete genomes and ORFIII polyproteins of BSGFV-YN and other 37 badnaviruses were constructed by using MEGA6.0 software (Table 2). The evolutionary history was inferred using the Neighbor-Joining method and its distance was computed using the p-distance method (Saitou & Nei, 1987). Meanwhile, to analyze the evolutionary relationship between BSGFV-YN and other BSV species in different banana hosts, a phylogenetic tree was also constructed based on the complete genome sequences.

Cloning and assembly of BSGFV-YN genome

To obtain the complete genome sequence of BSGFV-YN, 7 specific pair primers were used for segmental PCR amplification (Fig. 1A). The PCR results showed that each DNA fragment was specifically amplified and the length was consistent with the expected size. Bioinformatic analysis showed that the sequence of each fragment was highly similar to the BSGFV by blastn. The seven overlapping DNA fragments were assembled via head-to-tail method by BioEdit software, and a circular double-stranded DNA of 7,325 bp with 42% GC was finally obtained. The complete genome sequence of BSGFV-YN was deposited in GenBank under accession number MN296502.

Genomic analysis of BSGFV-YN

Sequence analysis showed that the characteristic of BSGFV-YN complete genome was similar to the typical badnavirus. In detail, the BSGFV-YN contains three open reading frames (ORFs) on the positive strand of viral genome. The first ORF (ORFI) was found at 484–1,011 bp while the second ORF (ORFII) was located at 1,008–1,346 bp, both of which were predicted to encode proteins with unknown function. However, the third ORF (ORFIII) located at 1,343-6,841 bp encoded a polyprotein with the largest in size (Fig. 1B). Further analysis revealed that the termination codon region of BSGFV-YN ORFI has a 4-base overlap (ATGA) with the start codon region of ORFII; similarly, the termination codon region of ORFII and the start codon region of ORFIII are also overlap of a 4-base, with ATGA as well.

Complete genome sequence analysis showed that BSGFV-YN was 98.14% sequence similarity with BSGFV Goldfinger, while it was 49.10–57.09% sequence similarity with other BSV species. The genome size of BSGFV is the smallest among all BSV species. In addition, the sequence similarity of BSGFV-YN with Canna yellow mottle associated virus (CaYMV), Kalanchoe top-spotting virus (KTSV), Pineapple bacilliform CO virus (PBCOV), or Sugarcane bacilliform MO virus (SBMOV) was 53.58%, 52.92%, 51.49% and 49.16%, respectively. Further, the BSGFV-YN ORFs (I∼III) were 99.43%, 99.11% and 98.91% nt sequence similarities with those of BSGFV Goldfinger, while they shared 99.62%, 98.82% and 99.02% at the amino acid levels. The sequence similarity between BSGFV-YN ORFs (I∼III) and other badnaviruses ORFs (I∼III) were even more lower (Table S1).

Badnavirus ORFIII, the largest ORF, encodes a ∼200 kDa polyprotein, which can be hydrolyzed by protease into several small proteins, such as MP, CP, AP, RT and RNase H. In this study, BSGFV-YN ORFIII polyprotein also contains MP, CP, AP, RT, RNaseH and two cysteine-rich zinc finger-like RNA-binding regions (Figs. 1B and 2). The domains of BSGFV-YN ORFIII are highly similar or identical to the corresponding domains of BSGFV Goldfinger. Further analysis showed that the RT domain was the most conservative with the similarity of 72.25∼84.29%, while the AP was the most variable, with the similarity of 39.39∼66.33% (Table S2).

The non-coding region of BSGFV-YN genome is 968 bp, which contains a specific binding site, the 12 bases (5′-TGG TAT CAG AGC-3′), of tRNA^met. This binding site is a common feature among all badnaviruses. The first nucleotide “T” is a highly conserved base required for replication and is set as the starting base of the sequence. The non-coding region also contains a TATA box domain (TATATA), which is located at 7,162–7,168 bp. In addition, the non-coding region has a 63 bp sequence that longer than BSGFV Goldfinger, and two kinds of repeating sequences were found in the 63 bp and adjacent region. A short repetition “Repeat S” and relative long repetition “Repeat L” have been found (Fig. 3).

Phylogenetic analysis

In order to analyze the evolutionary relationships between BSGFV-YN and 37 other badnaviruses, a phylogenetic tree based on their complete genomes was firstly constructed. The results showed that these badnaviruses can be divided into two distinct groups, Group I and Group II. BSGFV-YN and other BSV species were classified into the Group II. Furthermore, 11 BSVs can be further divided into two subgroups, BSV-I and BSV-II subgroups. BSGFV-YN was the highest homology with the BSGFV Goldfinger, but it was more close to KTSV than other BSV species (Fig. 4). Similar result has been reported in badnaviruses (Rumbou et al., 2018).

To further determine the phylogenetic relationship, another phylogenetic tree was constructed based on ORFIII polyproteins. The results confirmed that these badnaviruses can be divided into Group I and Group II. However, the Aglaonema bacilliform virus (ABV) group was clustered into Group I, which is different from the result of phylogenetic tree based on the complete genome. This phylogenetic tree also showed that BSGFV-YN had the highest homology with BSGFV Goldfinger (Fig. 5).

As shown in Fig. 6, phylogenetic tree was constructed by the eleven complete genomes of BSVs. The results showed that BSGFV-YN and other BSV species formed to two branches. In detail, BSUMV, BSUIV and BSULV formed to one branch, while BSGFV-YN, BSGFV Goldfinger, BSIMV, BSMYV, BSOLV, BSVNV, BSUAV, and BSV-Acum were clustered into another branch. The BSGFV-YN was the highest homology with BSGFV Goldfinger, which was consistent with results as Figs. 4 and 5A. It is interesting that BSGFV Goldfinger has been reported to infect Goldfinger (Musa AAAB), but the BSGFV-YN was cloned from another banana line (Musa AAA), indicating the BSGFV can infect different banana varieties.