Comparative Genome Analysis Reveals Differences in Biocontrol Efficacy According to Each Individual Isolate Belonging to Rhizospheric Fluorescent Pseudomonads

Biocontrol fluorescent pseudomonads produce a number of antibiotic organic compounds, including 2,4-diacetylphloroglucinol, pyoluteorin, pyrrolnitrin, and phenazine. We previously classified rhizospheric fluorescent pseudomonads harboring antibiotic biosynthetic gene clusters into 10 operational taxonomic units (OTUs). In the present study, we report the complete genome sequences of selected strains from these OTUs. The genetic diversity of antibiotic biosynthetic gene clusters and their surrounding sequences correlated with the OTU classification. In comparisons of the biocontrol activity and distribution of antibiotic biosynthetic gene clusters, we found that the pyrrolnitrin biosynthetic gene cluster more effectively controlled the growth of Rhizoctonia solani.

Pseudomonas is one of the most diversified bacterial genera and is widespread in multiple environments (Peix et al., 2009). Fluorescent pseudomonads are a specific group of the genus Pseudomonas that produce an extracellular, water-soluble, yellow-green pigment that fluoresces under UV irradiation (Bultreys et al., 2003). Several fluorescent pseudomonads produce antibiotics, exhibit biocontrol activities, and protect host plants (Höfte and Altier, 2010). These biocontrol fluorescent pseudomonads produce a number of antibiotic organic compounds, such as 2,4diacetylphloroglucinol (PHL), pyoluteorin (PLT), pyrrolnitrin (PRN), and phenazine (PHZ) (Gross and Loper, 2009), which exhibit effective biocontrol activities against phytopathogens within their respective antimicrobial spectra Stipanovic, 1979, 1980;He et al., 2004). In contrast, the diversity of fluorescent pseudomonads and the distribution of antibiotic biosynthesis genes have not yet been clearly elucidated. We previously isolated more than 3,000 strains of fluorescent pseudomonads from plant roots in Japan and screened for the presence of antibioticsynthesizing genes (Someya et al., 2020). A cluster analysis based on 16S rRNA sequences revealed that fluorescent pseudomonads may be classified into 10 operational taxonomic units (OTUs). When the presence of antibiotic bio-synthesis genes was assessed via PCR with specific primers, OTU HLR comprised three antibiotic biosynthesis genes (PHL, PLT, and PRN), OTU RZ contained two antibiotic biosynthesis genes (PRN and PHZ), and OTU H1, H2, H3, H4, H5, H6, H7, and H8 comprised only one antibiotic biosynthesis gene (PHL); however, when the biocontrol efficacy of each OTU was examined using a cabbage-Rhizoctonia solani pathosystem, the OTU classification and possession of antibiotic biosynthesis genes did not show a clear relationship with biocontrol efficacy (Someya et al., 2020). Based on these findings, we hypothesized that the genetic diversity of antibiotic biosynthetic gene clusters in the genome of fluorescent pseudomonads may be closely related to the production of antibiotics and biocontrol efficacy. In the present study, we report the complete genome sequences of 13 strains selected from the aforementioned 10 OTUs, clarify the distribution of antibiotic biosynthetic gene clusters in their genome, and reveal the relationships among the diversity of antibiotic biosynthetic gene clusters, production of antibiotic compounds, and biocontrol efficacy.
using Canu version 1.9 (Koren et al., 2017). Circular contigs were subjected to the prediction and annotation of genes using DFAST pipeline 1.2.11 (Tanizawa et al., 2018). Coding sequences (CDSs) were predicted using Prodigal 2.6.3 (Hyatt et al., 2010). Genes coding for tRNA and rRNA were discovered using Aragorn 1.2.38 (Laslett and Canback, 2004) and Barrnap 0.8 (https://github.com/ tseemann/barrnap), respectively. A homology search was conducted using BLAST software (Altschul et al., 1990) in SequenceServer 1.0.14 (Priyam et al., 2019). The phylogenetic tree based on whole genome alignments was constructed using REALPHY 1.12 (Bertels et al., 2014). The general features and DDBJ/ENA/GenBank accession numbers of the complete genome sequences of the Pseudomonas strains used in the present study are described in Table 1.
We assessed the presence of antibiotic biosynthetic gene clusters in the complete genome of 13 strains. A phylogenetic tree based on the complete sequences of the chromosome of Pseudomonas strains and the presence of an antibiotic biosynthesis gene cluster is shown in Fig. S1. The phl gene cluster for PHL biosynthesis, which consists of nine genes, was present in all complete genome sequences, except for Tre132 from OTU RZ. The phl gene cluster was widely distributed in Pseudomonas protegens, Pseudomonas saponiphila, Pseudomonas fluorescens, and Pseudomonas brassicacearum, which comprised OTU HLR, H1, H2, H3, and H4 (Fig. S1). We compared the sequences of the phl gene cluster and its surrounding regions in the complete genome sequences of these strains (Fig. 1). The sequences surrounding the phl gene cluster from OTU HLR and H1 were almost identical. The upstream sequences of the phl gene cluster of OTUs H2, H3, and H4 were very similar, and the downstream sequences were closely related to the phylogenetic relationship described in Fig. S1. The sequence of the phl gene cluster and its surrounding region in Cab53 from OTU H5 and its related strain Pseudomonas chlororaphis UFB2 indicated high similarity; however, the complete genome sequence of UFB2 markedly differed from that of P. chlororaphis subsp. aurantiaca type strain DSM 19603 T , which did not possess the phl gene cluster (Fig. 1). These results showed that UFB2 and Cab53 may be more accurately classified as other Pseudomonas species and not P. chlororaphis. The complete genome sequences of Pseudomonas baetica a390 T , Pseudomonas agarici LMG 2112 T , and Pseudomonas alcaligenes NBRC 14159 T , which were phylogenetically related to Seg1 from OTU H6, Ost2 from OTU H7, and Pc102 from OTU H8, respectively, did not contain the phl gene cluster. The downstream sequences of the phl gene cluster of Ost2 were only detected in the genome of P. chlororaphis PCL1606, which did not possess the phl gene cluster. Furthermore, the surrounding sequence of the phl gene cluster in Seg1 and PC102 was not found in any sequences deposited in the DDBJ/ENA/GenBank databases. Based on these results, the phl gene cluster did not appear to be originally equipped in OTU H6, H7, and H8, and was presumably acquired due to the introduction of another large DNA molecule into their genome by horizontal gene transfer.
To evaluate the relationship between the diversity of the phl gene cluster and production of PHL, we assessed PHL production by 31 strains from the phl gene clusterpositive OTUs described in a previous study (Someya et al., 2020) (Table S1). To extract PHL from the culture supernatant, bacterial strains were inoculated into 4 mL of nutrient broth (Becton, Dickinson and Company) containing 1% (w/v) glucose (NBGlu) medium. After an incubation for 18 h, bacterial cultures were inoculated into 4 mL of fresh NBGlu medium (1% inoculum) and incubated for a further 16 h. Thereafter, culture supernatants (200 μL) were mixed with an equal volume of acetonitrile. Samples (20 μL) were chromatographed on an HPLC system (Jasco) with a UV/VIS detector set at 270 nm using a Mightysil RP-18GP column (250×4.6 mm, particle diameter of 5 μm; Kanto Kagaku). Samples were eluted isocratically with water:acetonitrile:acetic acid (50:50:0.1 [v/v/v]) at 2 mL min -1 . The amount of PHL was estimated by comparing the peak areas for a given retention time with respect to the PHL solution of a known concentration. The results of the PHL quantification are shown in Fig. S2. The nine strains of OTU H2 showed high production levels of PHL, while six strains of OTUs H4, H5, and H6 showed moderate production levels of PHL. In contrast, the other 15 strains belonging to OTU HLR, H1, H3, H7, and H8 exhibited low production levels of PHL. Although the complete genome sequence and surrounding sequence of the phl gene cluster correlated among OTUs H2, H3, and H4 ( Fig. 1 and S1), a marked difference was observed in PHL production levels among these three OTUs. Differences in the regulatory system for the expression of the phl gene cluster was assumed to have affected PHL production in the three OTUs.
The prn gene cluster for PRN biosynthesis, which comprises four genes, was only present in the complete genome sequences of OTU HLR and RZ, whereas Eqa60, belonging to OTU HLZ, did not possess this gene cluster. We compared the prn gene cluster and its surrounding sequence in the complete genome sequence of the OTU HLR strains, Cab57, Boi14, and Eqa60 and P. protegens type strain CHA0 T (Fig. 2A), and found that the sequence of the prn gene cluster and its surrounding sequences, CHA0, Cab57, and Boi14, were similar. In the genome sequence of Eqa60, a 16-kbp region containing the prn gene cluster was deleted, and a 6-kbp sequence containing a putative transposon gene was inserted (Fig. 2A). The prn gene cluster was presumable eliminated due to the insertion of a transposon in the genome of Eqa60. The plt gene cluster, which consists of 17 genes, was only present in the complete genome sequence of OTU HLR. The sequence of the plt gene cluster and its surrounding sequence in CHA0, Cab57, and Eqa60 were well conserved (Fig. 2B). In contrast, Boi14 belonging to OTU HLR did not possess the plt gene cluster in its genome, and a 36-kbp region containing the plt gene cluster, which was conserved in other OTU HLR strains, was completely deleted in the genome of Boi14 (Fig. 2B). The phz gene cluster for PHZ biosynthesis, which consists of 10 genes, was only present in the complete genome sequence OTU RZ. The sequence of the phz gene cluster was similar between Tre132 and P. chlororaphis subsp. aurantiaca type strain DSM 19603 T (data not shown).
We previously assessed the biocontrol activities of the 13 complete genome-sequenced strains examined in the present study using a cabbage-R. solani pathosystem (Someya et al., 2020). The findings obtained demonstrated that most strains of OTU HLR, RZ, and H7 exhibited higher biocontrol activities (Table 1). Although PHL is an antibiotic produced by OTU HLR and H7 strains, it was produced in limited quantities by OTU HLR and H7 (Fig. S2). In contrast, although OTU H2 and H5 strains showed higher production levels of PHL, they did not exhibit any significant biocontrol activity in this system. PHL plays a pivotal role in the biological control of diseases caused by R. solani (He et al., 2004). The present results suggest that not all strains harboring the phl gene cluster effectively produce PHL on the plant surface; however, the activation of the phl gene cluster may be necessary to enhance the biocontrol activity of the strains that affect PHL production. The OTUs HLR and RZ, which exhibited higher biocontrol activities, share the prn gene cluster. Moreover, PRN was shown to exert inhibitory effects against diseases caused by R. solani (Cartwright et al., 1995). Most strains belonging to OTU HLR exhibited high biocontrol activities, with Boi14 showing the highest biocontrol activity among all strains used in this study (Someya et al., 2020). In contrast, one of the HLR strains, Eqa60, did not exhibit any biocontrol activity (Table 1). The comparative genome analysis performed in this study revealed that Eqa60 and Boi14 lacked the prn and plt gene clusters, respectively, in their genomes. These results suggest that the production of PRN rather than PLT is important for biocontrol activity against diseases caused by R. solani. In a future study, a mutation analysis of each cluster will reveal their functional relevance. As an exception, Ost2 from OTU H7, which did not possess the prn gene cluster, exhibited slight biocontrol activity, indicating that unknown genes involved in biocontrol activity are encoded in the genome of Ost2.
In conclusion, we herein demonstrated a relationship between the distribution of antibiotic biosynthetic gene clusters and their biocontrol efficacy through the complete genome sequencing of rhizospheric fluorescent pseudomonads. The present results suggest that PRN is the most effective antibiotic in the control of diseases caused by R. solani. Regarding other antibiotics, particularly PHL, the expression of antibiotic biosynthesis gene clusters may not be effectively performed on the surface of plants because some strains harboring the phl gene cluster did not exhibit biocontrol activity using the cabbage-R. solani pathosystem, but showed higher production levels of PHL under our experimental conditions. A previous study reported that the efficacy of biological control differed between controlled and commercial field conditions, and was less effective or completely ineffective when introduced under the latter conditions (Guetsky et al., 2001). To develop more effective biocontrol agents, further studies need to investigate not only on the presence of antibiotic biosynthesis gene clusters in the genome, but also attempt to activate and stabilize the expression of these gene clusters under uncontrolled field conditions.