Comparative Genomic Analysis of Delftia tsuruhatensis MTQ3 and the Identification of Functional NRPS Genes for Siderophore Production

Plant growth-promoting rhizobacteria (PGPR) are a group of rhizosphere bacteria that promote plant growth. Delftia tsuruhatensis MTQ3 is a member of PGPR that produces siderophores. The draft genome sequence of MTQ3 has been reported. Here, we analyzed the genome sequence of MTQ3 and performed a comparative genome analysis of four sequenced Delftia strains, revealing genetic relationships among these strains. In addition, genes responsible for bacteriocin and nonribosomal peptide synthesis were detected in the genomes of each strain. To reveal the functions of NRPS genes in siderophore production in D. tsuruhatensis MTQ3, three NRPS genes were knocked out to obtain the three mutants MTQ3-Δ1941, MTQ3-Δ1945, and MTQ3-Δ1946, which were compared with the wild-type strain. In qualitative and quantitative analyses using CAS assay, the mutants failed to produce siderophores. Accordingly, the NRPS genes in MTQ3 were functionally related to siderophore production. These results clarify one mechanism by which plant growth is promoted in MTQ3 and have important applications in agricultural production.


Introduction
Plant growth-promoting rhizobacteria (PGPR) could promote plant growth by a variety of mechanisms, such as siderophore production [1], antibiotics secretion, phytohormone generation, and the induction of systemic resistance [2,3]. PGPR have an important value in agricultural production.
Iron is essential to the growth of plants, but soluble iron is often insufficient in soil. Siderophores, which are low-molecular-weight molecules with a high affinity for ferric iron [4] secreted by PGPR, can mitigate that limitation, to some extent. Many microorganisms, including bacteria and fungi, can secrete siderophores [5]. In general, siderophores can be separated into two types based on their biosynthetic mechanism. One type depends on the nonribosomal peptide synthetase (NRPS) pathway, and the other is NRPS-independent [6].
Delftia has the ability to biodegrade organic pollutants, such as aniline [7], phenolic compounds [8], 2,4-dichlorophenoxyacetic acid (2,4-d) [9], and acetochlor. Delftia tsuruhatensis MTQ3 (MTQ3 for short) has been isolated by our group from the rhizosphere of tobacco in Guizhou, China. As an environmentally friendly PGPR, MTQ3 exhibits the potential to produce siderophores. Although species in the genus Delftia have been described as PGPR [10], their ability to produce siderophores has not been verified. The draft genome sequence of MTQ3 was formerly reported [11], but the mechanism of siderophore production is unclear.
In this work, a comparative genomic analysis of MTQ3 and the related genome sequences of Delftia sp. Cs1-4, D. acidovorans SPH-1, and D. tsuruhatensis 391 was performed. We characterized the genetic differences between the four Delftia strains. According to the genome annotation of MTQ3, a gene cluster that included three NRPS genes on scaffold 2  TAC GGC TAC CTT GTT ACG ACTT was found. The NRPS modular organizations were predicted using the PKS/NRPS Analysis website [15]. We constructed NRPS gene knockout mutants to analyze gene function with respect to siderophore biosynthesis.  obtained from GenBank were used for a comparative genome analysis with MTQ3 (LCZH00000000). The clusters of orthologous groups (COG) functional categories were analyzed by exploring all predicted proteins in the COG database using BLASTP [21]. Nonribosomal peptide and polyketide synthesis gene clusters were recognized using anti-SMASH (http://antismash.secondarymetabolites.org/help.html) [22], and their structures were compared to those of other known clusters.

Materials and Methods
2.6. Gene Knockout. All molecular genetic procedures for the genes orf-1941, orf-1945, and orf-1946 were performed according to the methods described in [23].

Medium and Cultivation.
For plasmid construction, E. coli strains were cultured in Luria-Bertani (LB) medium with gentamycin (50 g/mL) or kanamycin (100 g/mL), as needed, at 37 ∘ C. D. tsuruhatensis MTQ3 and the mutants were grown in LB medium with rifampicin (10 g/mL). The 5% sucrose plus LB plates with kanamycin and rifampicin were used to screen the recombinant strains. A CAS-agar plate [24] was used to qualitatively detect siderophores. For the quantitative analysis of siderophores, sucrose-asparagine (SA) medium was necessary, which included (per liter) 20 g of sucrose, 2.0 g of l-asparagine, 0.5 g of K 2 HPO 4 , and 0.5 g of MgSO 4 ⋅7H 2 O [25].

Qualitative and Quantitative Analyses of Siderophores.
Single clones of strains MTQ3, MTQ3-Δ1941, MTQ3-Δ1945, and MTQ3-Δ1946 were cultivated in LB plates at 37 ∘ C overnight. Then, the bacterial lawn was inoculated on a CAS-agar plate for cultivation at 37 ∘ C for 2-3 days, and the development of a color ring was monitored. Single clones of MTQ3, MTQ3-Δ1941, MTQ3-Δ1945, and MTQ3-Δ1946 were inoculated into 10 mL of sucroseasparagine (SA) liquid medium and incubated at 180 rpm and 37 ∘ C for two days. SA liquid medium [25] contains 20 g L −1 sucrose, 2.0 g L −1 l-asparagine, 0.5 g L −1 K 2 HPO 4 , and 0.5 g L −1 MgSO 4 ⋅7H 2 O. The supernatants of liquid cultures ( ) were collected by centrifugation at 10000 rpm for 15 min and then mixed with the CAS assay solution (at a volume ratio of 1 : 1) [26]. Uninoculated SA liquid medium was also treated following the same protocol as a reference ( ). After a 1-hour reaction period, the absorbances of and were detected at 630 nm [27]. Siderophore units were estimated as [( − )/ ] × 100 = siderophore units (%) [24]. Units not exceeding 10 were regarded as negative and no color change was found in the mixture.

Phylogenetic Analyses.
A phylogenetic tree was constructed based on the 16S rRNA sequences of Delftia spp. [28] in MEGA5 (Figure 1). These results suggested that strain MTQ3 had high homology with D. tsuruhatensis.

Comparative Analysis with Strains Cs1-4, SPH-1, and 391.
The general features of the four genomes are summarized in Table 3. There was clear variation in genome size. Specifically, the genome size varied from 5.7 Mb in MTQ3 to 6.7 Mb in SPH-1. Chromosomal coding DNA sequences (CDS) varied between 4103 in strain 391 and 6040 in SPH-1. The (G + C) mol% of the species shared a mean value of 66.60% and no strain differed from the mean value by >0.3%. These results indicated a genetic relationship between the species, to some extent.
To compare these genomes, the orthologous and unique genes among the four genomes were analyzed (Figure 2) [21]. The orthologous genes are contained in all strains, while the unique genes are owned by only one strain. A total of 2540 orthologous genes were shared and represented 51.05%, 43.34%, 42.05%, and 61.91% of all genes in MTQ3, Cs1-4, SPH-1, and 391, respectively. MTQ3 shared 4470, 4414, and 2782 orthologous genes with Cs1-4, SPH-1, and strain 391,  respectively. Meanwhile, MTQ3 possessed the least unique genes and accounted for 5.14% of all genes in its genome, but strain 391 contained the most unique genes and accounted for 20.94% in its genome. That might indicate that the genes of MTQ3 presented more conserved core genome for D. tsuruhatensis. Based on the COG-based analysis, the genes of these four genomes showed some similarities with respect to the distributions of COG categories (Figure 3). For the four genomes, genes related to transcription, amino acid transport and metabolism, and lipid transport and metabolism (COG categories K, E, and I, resp.) were relatively abundant functional categories, in addition to R and S, which represent general predicted functions and unknown functions, respectively, and provide little information regarding protein function [29]. The genome of MTQ3 included a larger proportion of genes involved in carbohydrate transport and metabolism compared with the other three genomes.
Bacteriocins, which are antimicrobial peptides or proteins produced by bacteria, could enhance environmental adaption. Enzymes related to the synthesis of nonribosomal peptides (NRP) and polyketides (PK) are modular and composed of a series of domains including adenylation, thiolation, condensation, and esterification domains [29]. We compared the NRPS gene cluster between the four genomes of Delftia, which are summarized in Figure 4. Bacteriocin synthetic gene clusters are also listed in Figure 4(b). The query sequence refers to the sequence of MTQ3. These results showed that not all Delftia genomes have identical clusters of NRPS and bacteriocin synthesis genes. These differences

COG Description
Nucleotide transport and metabolism Function unknown Signal transduction mechanisms Cell motility Transcription Amino acid transport and metabolism Defense mechanisms Cytoskeleton Secondary metabolites biosynthesis, transport, and catabolism Cell wall/membrane/envelope biogenesis Energy production and conversion Replication, recombination, and repair RNA processing and modification Translation and ribosomal structure and biogenesis Posttranslational modification, protein turnover, and chaperones Extracellular structures Inorganic ion transport and metabolism Chromatin structure and dynamics Coenzyme transport and metabolism Cell cycle control, cell division, and chromosome partitioning General function prediction only Lipid transport and metabolism Carbohydrate transport and metabolism Intracellular trafficking, secretion, and vesicular transport Delftia acidovorans SPH-1 Delftia sp. Cs1-4 Delftia tsuruhatensis 391 Delftia tsuruhatensis MTQ3 may reflect adaptations of the strains to their specific environments.

The Knockout of NRPSs in MTQ3.
To reveal the function of the NRPS genes (Figure 4(a)), gene knockouts were performed. The amplified target fragments of three NRPSs were ligated into the pGEM-T easy vector between sites Spe I and Xho I, which were then religated to the suicide plasmid pJQ200SK using the same restriction endonucleases for cutting. The resulting suicide plasmids were then linearized by Pst I. Km fragments were cloned from the plasmid pUC4k by Pst I digestion and then ligated into the above linearized suicide plasmids. Finally, the resulting three plasmids pk-1941, pk-1945, and pk-1946 were transformed into E. coli DH5 [30]. With the help of plasmid pRK2013, triparental mating [31] was used to generate the recombinant strains MTQ3-Δ1941, MTQ3-Δ1945, and MTQ3-Δ1946 ( Figure 5). The three mutants were selected by 5% sucrose plus LB plates with kanamycin and rifampicin and further confirmed by PCR using primers for Km and sequencing.

Qualitative and Quantitative Analyses of Siderophores.
On the CAS-agar plates, we observed an orange halo around the colony of MTQ3 after 2-3 days of incubation, but no orange ring around the mutants ( Figure 6). The presence of the orange ring suggested that MTQ3 can produce siderophores to chelate iron in the medium, thus resulting in a color change of the medium surrounding the colony. The mutants lost the ability to produce siderophores.
The quantitative measurements of siderophores (Table 4) indicated that the siderophore units of the wild-type strain  some similarities with respect to COG categories, but the proportions are somehow different. MTQ3 contains a larger proportion of genes involved in carbohydrate transport and metabolism, which indicates its better potential for carbohydrate utilization. For secondary metabolite prediction using anti-SMASH, the gene clusters of NRPS and bacteriocin synthesis genes are not identical in the Delftia genomes. The different features of the four genomes may be explained by adaptations of strains to their specific environments.
As PGPR, MTQ3 exhibits the potential to produce siderophores. To determine the genes involved in siderophore production, three NRPS genes orf-1941, orf-1945, and orf-1946 of MTQ3 were figured out to test the function. They belong to one gene cluster, the genes of which show 85% similarity to the nonribosomal peptide metallophore, delftibactin [32,33]. Compared with the wild-type strain MTQ3, three obtained mutants lost their ability to produce siderophores. These results indicated that siderophores could be biosynthesized by the NRPS modular multienzymes in MTQ3. Siderophores could improve the absorption of iron by plants, therefore promoting growth [4]. The ability of siderophores produced by bacteria to repress phytopathogens could be of significant importance in agriculture. This may explain why D. tsuruhatensis MTQ3 could stimulate plant growth. To the best of our knowledge, this is the first study to verify the gene cluster for siderophore production in D. tsuruhatensis. Meanwhile, the PGPR of MTQ3 may have important applications in agriculture.