An Avirulent Ralstonia Solanacearum Strain Undergoes Phenotype Conversion from a Pathogenic Strain Under Natural Environment

An avirulent R. solanacearum strain named FJAT-1458 was isolated from living tomato vessel and it showed no toxicity to tomato, pepper and eggplant. Multilocus sequence analysis (MLSA) based on eight genes (egl, hrpB, mutS, pehA, recA, rpoA, rpoB and rpoC) and whole genome average nucleotide identity (ANI) analysis suggested that strain FJAT-1458 belong to phylotype I. Genome sequence of the strain FJAT-1458 revealed a circular chromosome and a circular megaplasmid with whole genome size of 6,059,899 bp and GC content of 66.78%. Functional annotation of FJAT-1458 showed a total of 5,442 genes, with 5,166 protein-encoding genes, 202 pseudogenes and 74 noncoding RNA genes. Among which, 3,938 protein-coding genes can be assigned to 23 COG families, and 1,521 of them had KEGG orthologs. Prophage prediction using PHASTER revealed 12 prophages, including 7 intact, 1 questionable and 4 incomplete prophages. Comparative genome analyses between GMI1000 and FJAT-1458 showed that most of the virulence factors were well conserved and only small portion of them were distinct between them. Two genes, including a methyltransferase and an ISL3 family transposase genes, were identified to be inserted immediately upstream (141 bp) of phcA gene, which assumed to be responsible for avirulence of strain FJAT-1458. It is suggested that strain FJAT-1458 was originated from a wild-type pathogenic strain through an accident phenotype conversion, which is like those when cultured under experimental conditions. Our study provides new insight into the evolution of virulence in R. solanacearum strain under natural environment.


Introduction
Ralstonia solanacearum is one of the most important bacterial plant pathogens, which causes lethal wilts on more than 200 plant species belonging to over 50 different botanical families over a broad geographical range [1]. The bacterium enters plant roots, invades the xylem vessels and spreads rapidly to aerial parts of the plant through the vascular system where it reproduces in large number within a few days, leading to the death of plant [2]. In addition to its lethality, R.
solanacearum is able to survive in soils, waters and under various abiotic stresses for a long period without losing the ability to wilt host plants [3], which makes it even harder to be controlled [4]. Several strategies were employed to control of diseases caused by R. solanacearum, including use healthy plant seeds [5], crop rotation for 2-5 years [6], chemical control [1,7]. An alternative strategy was to use biological control agent such as antagonistic bacteria or avirulent mutants of R. solanacearum. However, the promising results could only be obtained under controlled conditions while was not confirmed in field [7].
R. solanacearum have evolved an elegant and effective system to invade host and expand its plant host range. There are two major pathogenicity determinants exist in R. solanacearum, the type III secretion system (T3SS) and extracellular polysaccharide (EPS). T3SS injects the "effector proteins" into the plant cell cytosol to favour infection [8,9], and EPS is largely responsible for the vascular dysfunction that causes wilt symptoms in susceptible hosts and promotes rapid systemic colonization as well [10]. Besides, R. solanacearum produces plenty of other factors that are potentially involved in the infection, including type II secreted plant cell wall-degrading enzymes, motility or attachment appendages, aerotaxis transducers, cellulases and pectinases [11]. The pathogenesis process was controlled through a sophisticated regulatory circuit and the LysR family transcriptional regulator PhcA plays a central role which regulates directly and indirectly many of those virulence genes [12].
Interestingly, under certain growth conditions, some members of the R. solanacearum population spontaneously undergo phenotype conversion (PC) from a wild-type pathogenic to a nonpathogenic form when it was allowed to grown to a high concentration [13]. Several genetic studies have revealed that PC is often resulted from mutation in phcA gene [14][15][16][17]. However, the phenomenon of PC under natural environment has never been reported probably due to that only few avirulent wild-type R. solanacearum strains have been identified. In our previous study, a R. solanacearum strain named FJAT-1458 has been isolated from living tomato vessel [18].
The objective of this study was to determine if the FJAT-1458 is born with nonpathogenic or the result of PC from a wild-type pathogenic strain. The complete genome sequence of strain FJAT-1458 was determined with single molecular real-time sequencing (SMRT) biotechnology. The genome comparison between strain FJAT-1458 and other R. solanacearum strains was performed. Besides, to elucidate the reason why strain FJAT-1458 is avirulent against Solanaceae plants, the virulence factors of strain FJAT-1458 were compared with strain GMI1000.

Pathogenicity Analysis
The Pathogenicity of the R. solanacearum strain FJAT-1458 was determined with a leaf-cutting method by using four-to six-leaf stage plants of tissue culture seedling of tomato (Solanum lycopersicum L. var. goldstone No. 1). The second, third and fourth leaves below the terminal bud of each seedling were cut about 1 cm length of wounds by a scissor. They were soaked in a bacterial suspension containing approximately 1×10 7 cfu/ml for 20 minutes. The treated seedlings were incubated in a greenhouse at 30°C and 80% of relative humidity for 12 hours light/dark. SPA medium was used as negative control. Each treatment was replicated 10 times and the whole experiment was repeated twice. The disease incidence (wilt symptoms) of plants were monitored daily for 6 days.

Genomic DNA Preparation, Library Construction, Sequencing and Assembly
Bacterial cells were grown at 30°C overnight in SPA liquid medium. Genomic DNA was isolated and purified according to the manufacture's instruction (Pacific Biosciences, Menlo Park, CA, USA). A 20-kb single-molecule real-time (SMRT) bell library was prepared with the SMRTbell template prep kit version 1.0 reagents (Pacific Biosciences, Menlo Park, CA, USA). The library was sequenced on a PacBio RS II sequencing platform using the C4 sequencing chemistry and P6 polymerase with 1 SMRT cells. The raw reads were filtered and assembled de novo following the Hierarchical Genome Assembly Process (HGAP) version 3.0 [20]. The polished assemblies were examined for circularity based on the presence of overlapping sequences at both ends of the contigs. Location of the overlapping sequence were determined using MUMmer version 3.0 [21].
successively to build the single super alignment. The best model for the alignment was estimated by MEGA software version 7.0 [29]. The phylogenetic tree was then constructed using Maximum Likelihood (ML) method in MEGA software version 7.0 [29]. The Jones-Taylor-Thornton (JTT) model assuming a discrete Gamma distribution (+G) with five rate categories was used for construction of a ML tree and the tree topology was evaluated by bootstrap analysis (1,000 replicates). Pairwise average nucleotide identity among these 19 strains were calculated with ANIm module of standalone version of JSpecies v1.2.1 with default parameters [30].

Nucleotide Sequence Deposition
The whole genome shotgun project has been deposited at DDBJ/EMBJ/GenBank under the accession number CP016554 and CP016555 (BioProject ID PRJNA329182, and BioSample ID SAMN05392572).

Pathogenicity Test of Strain FJAT-1458
The pathogenicity of R. solanacearum strain FJAT-1458 was determined with a leaf-cutting method by using tomato seedling with 5-6 leaves. Our results showed that strain FJAT-1458 is not able to cause wilt symptoms. However, the tomato seedlings began wilting at 4 days and reached to 58.33% incidence rate at 6 days after inoculated with R. solanacearum strain GMI1000 (Figure 1). The pathogenicity of R. solanacearum strain FJAT-1458 against pepper and eggplant were also tested, and it is not able to cause wilt symptoms in either of these two plants as well. Our results suggested that strain FJAT-1458 is probably avirulent.

General Genomic Features and Genome Annotation of Strain FJAT-1458
Whole genome sequencing was carried out with single molecular real-time sequencing (SMRT) biotechnology on the PacBio RS II platform. The completed genome of R. solanacearum strain FJAT-1458 was 6.06 Mb with a GC content of 66.78%. It contained one circular chromosome (3.98 Mb with a GC content of 66.71%, Figure 2a) and one circular megaplasmid (2.08 Mb with a GC content of 66.92%, Figure 2b). A total of 5,166 protein-coding genes were predicted (chromosome and megaplasmid encoded 3,633 and 1,533 genes, respectively). The FJAT-1458 genome contained 12 rRNA, 58 tRNA and 4 other non-coding RNAs. Of the 5,166 CDSs, 3,938 protein-coding genes can be assigned to 23 COG families. In addition, a total of 1,521 protein-coding genes had KEGG orthologs. 548 (10.61%) of protein-coding genes were predicted as classical secretory proteins with SignalP and 1,165 (22.55%) protein-coding genes have more than one transmembrane helix (Table 1).

Prophages
A total of 7 intact, 1 questionable and 4 incomplete prophages were identified in the genome of R. solanacearum strain FJAT-1458 by PHASTER [23]. These prophages were designated as Prophage 1-12 (Supplementary Table A1). Among which, Prophage 1-10 were located on the chromosome and Prophage 11-12 were located on the mega-plasmid. Prophage size ranged between 5.17 kb and 58.9 kb in length and GC content ranged between 57.85% and 66.54%. The GC content of Prophage 1 (66.54%), Prophage 2 (66.19%), Prophage 5 (66.31%) and Prophage 6 (66.26%) were very close to the average GC content of the whole genome (66.78%), indicating that they might have been integrated into R. solanacearum genome long time ago.
Besides phage structural genes and IS elements, prophage carries large number of genes which provide new functions to the host (Supplementary Table A2). For example, we identified five DNA methyltransferase, three methylase, three endonuclease and one exonuclease genes from these prophages, suggesting their role in DNA restriction and modification. We also identified two type III effector proteins and one type VI secretion system tip protein VgrG, which might contribute to virulence. Besides, we discovered a HicBA toxin-antitoxin II system in prophage 5, which encode a stable HicA toxin and a labile HicB antitoxin. TA systems are reported to be strongly correlated with physiological processes such as gene regulation, growth arrest, survival and apoptosis [31][32][33][34].

Comparative Genome Analysis
Phylogenetic relationships of R. solanacearum strain FJAT-1458 with other strains of R. solanacearum were assessed by performing Multilocus Sequence Analysis (MLSA) using eight genes (egl, hrpB, mutS, pehA, recA, rpoA, rpoB and rpoC). The result showed that strain FJAT-1458 belonged to phylotype I and was closest to strain YC-45 and SEPPX05 ( Figure 3). Based on average nucleotide identity (ANI) values, strains of R. solanacearum were grouped into three different genomespecies (Supplementary Table A3). The first group consists of strains from phylotype I and III. The second group includes phylotype II strains and the last group comprises strains from phylotype IV. ANI value is considered as one of the most robust measurements of genomic relatedness between strains and an ANI thresholds range (95~96%) correspond to ≥ 70% DDH standard for species definition [35]. These ANI results indicate that these three genomespecies groups could be considered as separate species. Our results are consistent with previous reports [36,37].

Genes Involved in Virulence
In this study, we showed that strain FJAT-1458 is avirulent to tomato, pepper and eggplant. To explore if avirulence of FJAT-1458 is due to the absence of key virulence factors, we created an inventory of 35 genes involved in virulence from GMI1000 [38], including exopolysaccharide (EPS) biosynthetic genes, cell wall degrading enzyme (CWDE) genes, response genes to the host defenses, key virulence regulator genes, chemotaxis genes, and genes involved in motility. Protein sequences predicted from the genome of strain FJAT-1458 were then searched against the database built from these virulence factors. Our results showed that these genes were well conserved between strain FJAT-1458 and GMI1000 (Supplementary Table A4), and the amino acid sequence identities were more than 99%, except phcB and twitching motility gene pilA, whose identities were 86.02% and 93.49%, respectively.

T2SS
Type II Secretion System (T2SS) is one means by which Gram-negative pathogens secrete proteins into the extracellular milieu and/or host organisms [39]. In R. solanacearum strains, lots of proteins secreted in a Type-II-dependent manner which contribute to its virulence, and R. solanacearum strain with a defective type II secretion system (T2SSs) is weakly virulent [40]. Similar to strain GMI1000 [41], FJAT-1458 harbors three type II secretion systems (T2SS) (Figure 4). The first one is the orthodox system which contains 12 genes in the chromosome (from 355,322 to 367,621). This gene cluster is well conserved and shares a high sequence identity with strain GMI1000 (average amino acid sequence identity of 99.32%). The other two T2SSs are unorthodox systems. One includes seven core genes in the chromosome (ranging from 1,233,265 to 1,243,218) with four hypothetical genes inserted between gspE and gspD (Figure 4 b). The other possesses six core genes located in the mega-plasmid (from 1,835,267 to 1,844,143), with four hypothetical genes inserted between gspD and gspE (Figure 4 c). These two gene clusters are also conserved between strain GMI1000 and strain FJAT-1458 (average amino acid sequence identity of 98.68%).

T3SS
The type III secretion system (T3SS) is widely spread in gram-negative bacteria, and is responsible for delivering bacterial proteins, termed effectors, from the bacterial cytosol directly into the interior of host cells [42]. These translocated proteins facilitate bacterial pathogenesis by specifically interfering with host cell signal transduction and other cellular processes [43]. The hrp gene cluster of T3SS is the key virulence determinant in R. solanacearum. In strain FJAT-1458, it is located on the megaplasmid and spans 29.655 kb (from 785,966 to 815,620), composed of 30 genes. Most of genes were well conserved between strain FJAT-1458 and GMI1000 ( Figure 5), sharing amino acid sequence identity from 94.19% to 100.00%. One gene named popC was presumed to be pseudo due to internal stop codon.
A total of 68 T3es were identified in FJAT-1458 genome (Supplementary Table A5). Among which, 62 (91.18%) of them were also present in the strain GMI1000. Six T3es were present in strain FJAT-1458 but absent in strain GMI1000, while 12 T3es were absent in strain FJAT1458 but were present in strain GMI1000.

T4SS
The Type IV Secretion System (T4SS) plays diverse important roles in virulence and adaption. In strain GMI1000, the T4SS gene cluster is comprised of 17 genes (RSc2574-RSc2588, RSp0179, and RSp1521) [41]. No experimental evidence is yet available in support of a role for these genes as a "fitness island" or an "ecological island" [44]. These 17 T4SS genes were searched against the genome of strain FJAT-1458, and none of them were identified in this genome.

T6SS
The type VI secretion system (T6SS) is a complex and widespread gram-negative bacterial export pathway with the capacity to translocate protein effectors into a diversity of target cell types [45]. Previous study has showed that T6SS could contribute to pathogenicity, swimming motility and mediate biofilm formation [46]. In strain GMI1000, the T6SS locus spans an approximate 45.2-kb region in the mega-plasmid, comprised of 15 core genes and 16 additional genes. In strain FJAT-1458, the T6SS locus located in a 46 kb region in the mega-plasmid, containing 15 core genes and 19 additional genes. The core T6SS genes between these two strains were conserved, with average amino acid identity up to 99.11% ( Figure 6). The inserted ORFs between vgrGA2 and impA varied significantly (8 additional ORFs in strain GMI1000, while 11 additional ORFs in strain FJAT-1458). Among these 11 ORFs, four of them were identified as transposon, suggesting this region to be a hotspot for insertion.

A Fragment was Inserted Immediately Upstream of Phca Gene
Comparison of virulence factors between strain GMI1000 and strain FJAT-1458 revealed that most of them are well conserved between them, and only small portion of virulence factors are distinct. These distinct virulence factors might lead to weakened virulence instead of loss of virulence. Previous study revealed that the LysR family transcriptional regulator PhcA plays a central role which regulates directly and indirectly many of those virulence genes [12] and inactivation of phcA gene resulted in loss of virulence [14][15][16]. A previous genetic study carried out with strain AW1 showed that spontaneous PC can be attributed to insertions with phcA gene (2-bp, 200-bp or 1-kb insertions) [14]. Another independent research with three spontaneous PC mutation strains of ACH0158 also showed that inactivation of phcA gene (an insertion of ISRso4, a 132bp-deletion and a 2-bp insertion) results in loss of virulence [16]. Thus, we carefully examined the phcA gene and its flanking sequence. We didn't find any insertion sequence in phcA gene and it is well conserved between strain GMI1000 and strain FJAT-1458 (coverage 100.00%, identity 99.71%). However, we did find two genes, including a methyltransferase and an ISL3 family transposase genes, inserted immediately upstream (141 bp) of phcA gene (Figure 7). This insertion sequence might suppress the expression of phcA gene as it is just located in the promoter region of phcA gene. The low levels of functional PhcA could not be able to activate the transcription of its downstream genes associated with the virulence of the pathogen [47], which finally resulted in avirulence of FJAT-1458. Further experiments (such as transcriptome analysis and genetic compensation experiment) are still required to confirm.

Conclusions
An avirulent R. solanacearum strain named FJAT-1458 was isolated from living tomato vessel and it showed no toxicity to tomato, pepper and eggplant. Comparative genome analyses between GMI1000 and FJAT-1458 revealed that a fragment, containing a methyltransferase and an ISL3 family transposase genes, was inserted immediately upstream (141 bp) of phcA gene, which assumed to be responsible for avirulence of FJAT-1458. It is suggested that strain FJAT-1458 was originated from a wild-type pathogenic strain through an accident phenotype conversion, which is like those when cultured under experimental conditions. Our study provides new insight into the evolution of virulence in R. solanacearum strain under natural environment.

Author Contributions
CD, LB and ZH conceived and designed the experiments. CD, LY, CY, CJ performed the experiments. CD, CJ and WJ generated and analyzed the data. CD wrote the paper.  I  III  IV  IV  II  II  II  II  II  II  FJAT-1458  I Po82  IBSBF  1503  UW163  RS489   I  III  IV  IV  II  II  II  II  II  II  IBSBF1503  II