Ralstonia solanacearum is a soil-borne, Gram-negative bacterium that causes bacterial wilt disease in many plants, including tomato, potato, eggplant and tobacco [5, 18]. This bacterium has an unusually wide host range of over 200 plant species belonging to more than 50 families [5]. R. solanacearum is highly heterogeneous, and bacterial strains have been divided into five races on the basis of their host ranges, and six biovars based on their ability to utilize sugar alcohols and disaccharides [5]. Recently, a new classification system for R. solanacearum strains divided them into four phylogenetic groups that roughly correspond to geographic origin [2, 17]. Given the high genetic diversity and broad host range of this pathogen, the common agronomic strategies, such as using disease-resistant plant varieties, crop rotation, and chemical bactericides, are not fully effective to control this pathogen in various plants. Therefore, more research is required to develop innovative and integrated approaches for effective management of this disease.

Bacteriophage (phage) therapy, a method of using virulent phage to control bacterial diseases, was first introduced by Félix d’Herelle, who discovered phages in 1917 [1]. During that time, phage therapy was regarded as a promising method to control bacterial disease. However, this ambition was abandoned after the 1940s with the discovery of various antibiotics. Recently, the idea of using phage therapy to control bacterial diseases has been gaining attention around the world as a potential solution to the appearance of drug-resistant bacteria [3, 11, 14].

In this paper, a novel lytic phage that infects the pathogen of tobacco wilt disease, R. solanacearum, was isolated using the host strain TB15-14 from soil samples collected from long-term cultivated tobacco fields in Fujian Province, South China. The phage was observed and purified using a double agar overlay plaque assay [7]. After 4 rounds of purification, a lytic phage was obtained. The morphological characteristics of the phage were observed under a transmission electron microscope (JEM-1400, JEOL, Japan), with results showing that it had an icosahedral head with a diameter of 51 ± 4 nm and a short tail with a length of 13 ± 5 nm. This finding indicated that this phage belongs to the family Podoviridae, and it was subsequently named phage RSPI1.

The genome of phage RSPI1 was sequenced using the high-throughput sequencing platform IonTorrent PGM Sequencer (Thermo Fisher Scientific, USA) at the State Key Laboratory of Pathogen and Biosecurity, Beijing Institute of Microbiology and Epidemiology [19]. Reads less than 30 bp were filtered and low quality (Q value < 20) bases were trimmed using NGS QC Toolkit v2.3.3 [10]. After data filtering, all sequences were assembled using Roche Newbler version 2.9 (Roche Applied Science), and the genomic termini were determined according to a method described previously [20]. Genome annotations and open reading frames (ORFs) were identified using RAST at http://rast.nmpdr.org/ [8]. The putative functions of the ORFs were examined for homologs using the deltaBLAST program at NCBI website (http://blast.ncbi.nlm.nih.gov/).

The dsDNA sequence of phage RSPI1 had a length of 43,211 bp with a mean G + C content of 61.5%. The sequence was submitted to the NCBI website for a BLASTn search and results showed that it was most similar to the Podophages Bp-AMP1, Bp-AMP2, Bp-AMP3 and Bp-AMP4 (GenBank accession numbers for those phages are HG793132, HG796219, HG796220 and HG796221, respectively) that infect Burkholderia pseudomallei, with a query cover of 78–79%. The highest identity was 84% between these sequences. In addition to those four phages, RSPI1 also had a query cover of 10% and 78% identity within coverage to phage RS-PII-1 (GenBank accession number KY316062), which was isolated by our group from the host R. solanacearum TB15-15 [13], and had a query cover of 8% and 73% identity within coverage to phage RSB1 (GenBank accession number AB451219), also isolated from R. solanacearum [6].

In total, 48 ORFs were identified with lengths ranging from 132 bp to 5061 bp and with encoded amino acid (aa) polypeptides ranging from 44 to 1687 aa in size. There were no tRNA sequences within this genome. Of the 48 ORFs, 42 initiated with the start codon ATG, 2 with GTG, and 4 with TTG. A map of predicted ORFs (Fig. 1) was generated using CGView [12]. Among the 48 ORFs, 11 were identified as functional genes which encoded proteins such as a phage integrase, lysozyme, phage DNA packaging protein, putative transglycosylase, phage terminase, phage RNA polymerase, phage-associated DNA ligase, RNase H superfamily, phage-associated DNA polymerase I, replicative DNA helicase, and phage-associated DNA primase/helicase; 12 were identified as structural genes which encoded proteins such as phage tail fibers, phage capsid and scaffold, and phage collar. The remaining 25 ORFs encoded hypothetical proteins and were characterized as unknown genes. A BLASTp search of all ORF sequences showed that 45 ORFs shared highest identities with Burkholderia phages Bp-AMP1, Bp-AMP2 and Bp-AMP4. Exceptions included three ORFs (orf1, orf47 and orf48) that encoded a hypothetical protein with less than 50% identity to Synechococcus phage syn9 (GenBank accession number YP_717848) and two uncultured Mediterranean phage sequences (GenBank accession number BAQ94277 and BAR32162) obtained by metagenomics sequencing [9] (Table S1).

Fig. 1
figure 1

Genome map of the phage RSPI1. Patterns are divided into five circles: the outermost circle represents the full length of the genome; the second circle represents the forward reading frame; the third circle represents the reverse reading frame; while on the fourth circle, red and green represent GC content, wherein red represents less than the total genome’s average GC content while green is the opposite. The most inner circle of orange and sky-blue indicates a GC skew of G-C/G+C, wherein orange is less than 0 and sky-blue is greater than 0. The ORFs marked with blue, orange and yellow indicate genes encoding functional proteins, structural proteins and hypothetical proteins, respectively

Full size image

To reveal the relationship between phage RSPI1 and other members classified within the Podoviridae, two neighbor-joining phylogenetic trees were constructed using Molecular Evolutionary Genetic Analysis (MEGA) version 5.0 [15] based on the amino acid sequences of the phage DNA polymerase I (orf31) and the phage tail fiber protein (orf15). The tree constructed using phage DNA polymerase I sequences showed that phage RSPI1 and Burkholderia phage Bp-AMP1 formed a 100% bootstrap value cluster, and this cluster was closely related to Ralstonia phages RS-PII-1 and RSB1 (Fig. 2A). Similarly, the tree for the phage tail fiber protein indicated that phage RSPI1 formed a 100% bootstrap value cluster with Burkholderia phage Bp-AMP1, which was part of a larger cluster (bootstrap value of 59%) that included Ralstonia phages RSJ2 and RSJ5 (Fig. 2B). Our previously isolated Ralstonia phage RS-PII-1 formed a small cluster with Ralstonia phage RSB1, and this cluster was located below the position of RSPI1. Thus, irrespective of the results of the BLAST search and the two phylogenetic trees, our findings indicated that phage RSPI1 was more similar to Burkholderia pseudomallei phages than R. solanacearum phages. This result was not surprising, since Kawasaki et al. [6] had already demonstrated that the genome sequence of R. solanacearum phage RSB1 was more similar to phages of Xanthomonas oryzae and Pseudomonas aeruginosa than to phages of R. solanacearum. Recently, Thi et al. [16] also observed that sequences in the range of 12-25 kbp in the genome of R. solanacearum phage RS138 were similar to sequences in phages infecting Burkholderia and Pseudomonas. In addition, a previous paper stated that both Ralstonia and Burkholderia belong to the order Burkholderiales and may share common phages [4]. Our phylogenetics data showing close proximity of Burkholderia phages and Ralstonia phages (Figure 2) supports the conclusions of Fujiwara et al. [4]. Given the whole genome sequence of RSPI1 had a low query cover, low similarity to known Burkholderia phages, and also had a very low query cover (less than 10%) to known Ralstonia phages, especially between PSPI1 in this study and PS-PII-1 from previous reports [13], these findings show that phage PSPI1 is a novel lytic podophage that infects R. solanacearum.

Fig. 2
figure 2

Neighbor-joining phylogenetic trees for the amino acid sequences of (A) DNA polymerase I protein and (B) tail fiber protein showing the relationships between phage RSPI1 and other known podophages. The values at the nodes indicate the bootstrap support scores as calculated using 1000 replicates. Numbers in parentheses are the accession numbers of the amino acid sequences at the NCBI web site

Full size image