Genome sequencing and analysis of plant growth-promoting attributes from Leclercia adecarboxylata

Abstract Plant growth-promoting bacteria are ecological alternatives for fertilization, mainly for gramineous. Since plant x bacteria interaction is genotype and strain dependent, searching for new strains may contribute to the development of new biofertilizers. We aim to characterize plant growth-promoting capacity of Leclercia adecarboxylata strain Palotina, formerly isolated by our group in corn. A single isolated colony was taken and its genome was sequenced using Illumina technology. The whole genome was compared to other Leclercia adecarboxylata strains, and their biological and growth-promoting traits, such as P solubilization and auxin production, were tested. Following that, a 4.8 Mb genome of L. adecarboxylata strain Palotina was assembled and the functional annotation was carried out. This paper is the first to report the genes associated with plant growth promotion demonstrating in vitro indole acid production by this strain. These results project the endophyte as a potential biofertilizer for further commercial exploitation.

Plants and microorganisms naturally interact in the soil, forming a narrow and complex communication network. This network operates on biochemical to molecular signals, which can be altered according to the type of association (Souza et al., 2015). The promotion of direct growth occurs through the availability of nutrients, nitrogen, phosphate, as well as the production of plant regulators as auxins, cytokinins and amino acids. These regulators mainly promote central and lateral root growth, increasing the absorption surface, which in turn increases the root's nutrient and water uptake (Beneduzi et al., 2012;Jha and Saraf, 2015).
However, the promotion of indirect growth occurs by means of induced systemic resistance (ISR). Some biocontrol mechanisms of pathogens are antibiosis, parasitism, competition for nutrients, production of hydrogen cyanide, siderophores, including the ones involved in responses to abiotic stresses, such as drought, salinity, extreme temperatures (Moreira et al., 2016).
Although this organism has been reported globally in food, water and animals (Tamura et al., 1986;Anuradha, 2014), for instance in strawberry root (Laili et al., 2017), evidences of its efficiency as plant growth promoter bacteria is scarce. In this context, we sequenced the complete genome of Leclercia adecarboxylata strain Palotina carrying out a comparative analysis with genomes of 16 different strains. This study provides new insights into genetic determinants, and as such may clarify some reported metabolic abilities of the Palotina strain, offering basic information on genetic plant growth promotion that may be relevant for biotechnological interest.

DNA extraction and sequencing
Genomic DNA was extracted from the isolated strain following the protocol by Souza et al. (1991), using as template for a PCR reaction, Y1 and Y3 primers for amplification of the 16S rRNA gene (Cruz et al., 2001). Amplicons were enzymatically treated with ExoI/SA and the sequencing was performed on BigDye® Terminator v3.1 Cycle Sequencing in an ABI3500xL. The resulting sequences were assembled with CAP3 using BLASTn for comparison at NCBI. Snak et al. 2 The gDNA of Leclercia was quantified with Qubit, diluted and used for the construction of genomic DNA sequencing libraries using Illumina NexteraXT kit, according to the manufacturer's recommendations. The libraries were quantified and the quality was verified by means of Bioanalyzer. The libraries were diluted to 500 pM and pooled. This pool was quantified by qPCR using the Kapa Biosystems kit, and 17.5 pM of pooled libraries were sequenced in the Illumina MiSeq with 500V2 kit in paired-end, generating paired reads of 250 base pairs from DNA fragments.

Genome assembly, annotation and serotyping
Overall, 5,795,728 reads were generated, representing a 31-fold coverage for the strain Palotina. FastQC (www. bioinformatics.babraham.ac.uk/projects/fastqc/) was used to check the quality of the reads. SPAdes program (Bankevich et al., 2012), version 3.11.1 was used to reassemble the sequence dataset, which were deposited at NCBI site under the BioSample access number SAMN09791487. In order to identify putative coding sequences (CDS) and provide an initial automatic annotation, the genome sequences were submitted to the RAST server annotation pipeline (Aziz et al., 2008) and Artemis (Sanger Institute, Cambridge, UK) was used to curate annotations manually.
Also, three genes related to plant growth promotion (P metabolism and auxins biosynthesis) were selected and compared by BLASTn against all genomes. In phylogenetic analyses, a Neighbor-Joining tree (Saitou and Nei, 1987) was constructed with 98 genomes with NCBI Tree Viewer (version 1.17.5).

Biochemical characterization of L. adecarboxylata
At first, L. adecarboxylata strain Palotina was isolated in LB medium, growing well in DYGS (Dobereiner et al., 1995), following the isolation protocol by (Chaves et al., 2019). Visual assays determined bacteria morphology and colony color. Bacterial phosphate solubilization was detected in vitro by inoculation in NBRIP medium (Nautiyal et al., 2000). A bacterial colony was collected with a toothpick, and each ¼ plate of NBRIP medium plate was inoculated. A halo around the colonies was observed after 10 days in a culture incubated at 28 °C. The solubilization index (SI) was calculated as: SI=Diameter of Halo (mm) / Diameter of colony (mm) (Nautiyal, 1999). Indole-3-acetic acid (IAA) production by bacteria was based on the Glickmann and Dessaux (1995) protocol. Isolates were inoculated into glass vials (penicillin-type) containing 4 mL of medium with tryptophan (5.0 g.L -1 glucose, 0.025 g.L -1 yeast extract and 0.204 g.L -1 L-TRP) and no tryptophan supplementation (Sarwar and Kremer, 1995). Triplicate vials were incubated in a shaker cooled at 28 °C in the dark at 120 rpm. After the growth, which occurred in 48 h, 2 mL of the culture medium was centrifuged at 10000 g for 10 min at 4 °C. Next, 1 mL of the bacterial suspension supernatant was transferred to a 15 mL Falcon-type tube with the addition of 1 mL of Salkowski reagent. The standard curve was assayed for final concentrations of 0 to 0.03 mg mL -1 . Samples were left in the dark for 30 min and the AIA quantification was performed by spectrophotometer reading at 535 nm.
L. adecarboxylata was primarily screened and further grown on LB plate containing 1%, 2%, 5% and 10% NaCl separately for 48 h at 30 °C. In addition, the optimum pH was checked using LB liquid medium with different pH (4; 5; 5.5; 6; 6.5; 7; 7.5 and 8). The growth temperatures assessed were 25 and 37 o C. The presence of oxidase was tested using TEMED 1 % (N-N-dimetil-p-phenilenediamine) (Kovacks,1956). The presence of catalase was verified by the presence of bubbles when hydrogen peroxide was deposited in a colony (Yano et al., 1991). All assays were made in triplicate.
Blood agar plates (5 % (v/v) sheep blood) were used for biosafety test (Russell et al., 2006;Suleman et al., 2018). The hemolytic capacity was evaluated after 48 h from fresh culture of L. adecarboxylata streaked onto blood agar plates and incubated at 37 ± 2 o C.

Genome assembly, annotation and comparative genomics
After de novo assembly, the genome of Leclercia adecarboxylata strain Palotina was represented in 20 contigs, sized 4,801,735 bp, with GC content of 55.7%, 4.379 coding sequences and no plasmid were observed. The comparison showed differences in the genome of L. adecarboxylata strain Palotina ( Figure 1, Table 1). The size of L. adecarboxylata strains ranged from 4.72 to 5.76 Mb, their CG content between 55-56% CG content. Some contained plasmids (up to 7). The BRIG genomic analyses showed CRISPR system and mobile elements as phages were absent in some strains. One interesting data is that no Leclercia sp contained the indole acetamide hydrolase gene. However, group genes related to bacterial systems, such as several hypothetical proteins, Type I restriction and cobalt/cadmium/zinc RND efflux, were absent in all strains used in the comparison (Figure 1, Figures S1 to S3).
Of all identified coding sequences, Rast server classified 2597 genes (60%) in categories (Subsystems) and 1782 (40%) were grouped as not classified (Not in Subsystems). The categories with the highest number of genes were carbohydrates metabolism (613 genes), followed by amino acids and derivatives (470 genes), and protein metabolism with 302 genes (Figure 2). Dormancy, sporulation and secondary metabolism showed the lowest gene number (only 5).
BLASTp comparison revealed 14 genes related to plant growth promotion that showed high identity (> 97%) and high e-value (Table S2). The phylogenetic analysis of all the 98 genomes found at NCBI, belonging to "Leclercia", showed a higher similarity between L. adecarboxylata strain Palotina and USDA-ARS-USMARC-60222, isolated from calf nasopharynx, an indicative that these bacteria can be associated to agricultural area, unlike clinical strains (Figure 3).

Biochemical and molecular characterization of L. adecarboxylata
The biological and plant growth promotion traits are summarized in Table 2. L. adecarboxylata strain Palotina is a cream rod-shaped, non-spore-forming, motile, Gramnegative bacillus of family Enterobacteriaceae, having an optimum pH growing range between 5.0-8.0, 25-37 o C for growth temperature and a low salinity toleration (below 5 %).
The strain also presented oxidase and negative catalase response (Table 2). In addition, genes for chitinase production were found in the genome. Antifungal resistance was not tested in L. adecarboxylata strain Palotina by the inoculation with Aspergillus flavus.
Halo zone formation on blood agar medium was observed in vitro, which points to hemolysin gene expression, confirming the opportunistic pathogen trait. A lipase gene was annotated in the genome, demonstrates a potential use of this strain for biotechnological purposes. Moreover, we identified genes that can be related to the improvement of nutrient availability to plants (Tables 3 and 4), which is consistent with many plant growth promoting bacteria (PGPB). The genome of L adecarboxylata strain Palotina possesses genes encoding glucose dehydrogenase (gcd), the major enzyme responsible for the production of gluconic acid. Palotina strain showed a medium capacity of P solubilization (2 < PSI > 4) ( Table 2). UDP-glucose dehydrogenase gene was present in all 16 Leclercia genomes (Table 3).
Trp cluster (trpC, D and F), tryptophan-permease, tryptophan-synthase (a) and (b) genes, and indole acetamide hydrolase gene involved in tryptophan biosynthesis were found in the genome (Table 4). We observed an increase of 2.3-fold in IAA production when tryptophan was added to the culture medium (Table 2). When we compared all 16 Leclercia genomes, the phosphoribosyl anthranilate isomerase gene was found in all strains (Table 3).   (Saitou and Nei, 1987). The 1,000 resampling bootstrap values are shown.
Ammonia assimilation genes, among others, seem to be the main N metabolism pathway, confirmed by the presence of several genes, such as GS type I (Glutamine synthase); NADPH-GOGAT; Amt (ammonia transporter); NRI (protein regulator) and PKII (Table 4).

Discussion
Our findings indicate the complete absence of the RND protein family, which was reported as a group of bacterial transport proteins involved in cell division, nodulation and heavy metal resistance (Nies, 2003). Another gene sequence that appeared to be distinct among Leclercia genomes is the clustered regularly interspaced short palindromic repeats (CRISPR), which is related to the microbial immune system. It contains a family of proteins whose functional domains are related to polynucleotide-binding proteins, polymerases, nucleases, and helicases (Horvath and Barrangou, 2010;Ishino et al., 2018). This region was observed in only three of Leclercia strains, including strain Palotina, which shows a horizontal gene transfer promoting a genomic differentiation among strains (Portillo and Gonzalez, 2009). A Type I restriction system or Restriction modification system (R-M system) was absent in all compared genomes. R-M system has large pentameric proteins with separate restriction, methylation and DNA sequencerecognition subunits (Loenen et al., 2014), which grants to the host bacterium a selective advantage (Sitaraman, 2016).
Carbohydrate metabolism genes were present in L. adecarboxylata strain Palotina enabling this bacterium to grow in different media using different carbohydrate/energy sources, including root exudates and other organic compounds. Moreover, this strain would interact positively with plants under harsh soil conditions. Although our strain was able to carry out an alpha hemolysis, Muratoglu et al., (2009) and Anuradha (2014), who tested L. adecarboxylata Ld1 and human isolates respectively, found a negative response to blood hemolysis. The contrasting results could possibly be explained by the presence of the hemolysin gene set found in the genome of our strain.
Strain Palotina showed a P solubilization capacity, probably explained by the presence of the gcd gene. Glucose dehydrogenase is the key enzyme in the biosynthesis of gluconic acid in the direct oxidation pathway of glucose, responsible for P solubilization (Chen et al., 2016;Suleman et al., 2018). The amount of gluconic acid released would control the availability of soluble phosphates (De Werra et al., 2009). Also, UDP-glucose dehydrogenase found in all compared strains (Table 3) catalyzes an NAD + -dependent twofold oxidation of UDP-glucose to generate UDP-glucuronic acid (Chen et al., 2019). This acid is also a precursor to UDPxylose component of the cell wall polysaccharides in plants (Gibeaut and Carpita, 1994).
Genomic analyses identified the indole acetamide hydrolase gene, which explains the IAA production mainly by IAM pathway suggesting the tryptophan-dependent IAM pathway function in strain Palotina. The main pathway to IAA production in PGPB is via indole-3-pyruvic acid, dependent on L-tryptophan (Souza et al., 2015). Not all Leclercia adecarboxylata and no Leclercia sp. strains present the indole acetamide hydrolase gene, which suggests that this gene has been acquired. This fact explains the association between bacteria and corn plants.
In addition, we identified the sequence of phosphoribosyl anthranilate isomerase (PRAI) encoded by trpC (Table 4). This enzyme is responsible for the conversion of N-(5′-phosphoribosyl)-anthranilate (PRA) to 1-(o-carboxyphenylamino)-1-deoxyribulose 5-phosphate (CdRP), the fourth step in tryptophan biosynthesis (Thoma et al., 2000). Moreover, monoamine oxidase plays an important role in tryptamine biosynthesis, whose oxidative deamination of tryptamine to indole acetaldehyde is known to be the main course for IAA formation, despite the fact that the role of monoamine oxidase has not been completely characterized (Ueno et al., 2003). The presence of these genes suggests that the tryptophan-dependent IAM and TAM pathways function in L. adecarboxylata.
L adecarboxylata produced 2.6 µg.mL -1 of IAA (Table 2). Albeit the variable levels, Gupta et al., (2014) related IAA production of 1.2-2.5 ug.mL -1 to candidate PGPB strains isolated from coconut, cocoa and arecanut plants, while Moreira et al., (2016) found strains that could produce more than 80 µg.mL -1 of indolic compounds. We did not identify an acdS gene coding for ACC deaminase enzyme in our strain, which demonstrates the absence of this enzyme among PGP traits. However, Kang et al. (2019) suggested that the IAA and ACC deaminase helped tomato (Solanum lycopersicum) plants to tolerate salt stress, despite having found acdS gene in L. adecarboxylata strain MO1.
Ammonia assimilation, among others, seems to be the main N metabolization pathway from nitrate. In addition, this strain exhibits the genes for denitrification used as energy source. These genes indicated that L. adecarboxylata has an important role in soil N cycling system. The results agree with Muratoglu et al. (2009) who observed an absence of nitrogen fixation capacity as well as a presence of NO 2 metabolism in Ld1 strain. From these data, L. adecarboxylata can be used as a model for PGP bacteria exclusively by auxins production.
Peroxidases, catalases, superoxide dismutase, and glutathione transferases genes found at L. adecarboxylata genome could help plants to overcome oxidative stress. Also, heat and cold shock genes could support bacteria to survive during abiotic or biotic stress (Gupta et al., 2014), which enable bacteria to adapt to adverse growth conditions.
Another strategy to copy with abiotic stresses is the accumulation of compatible solutes, such as trehalose, proline and glycine betaine, among others, by some soil bacteria (Suarez et al., 2019). The strain Palotina genome contains trehalose-6-phosphate synthase involved in GDP-or UDPglucose conversion to trehalose (Avonce et al., 2006). Also, glycine betaine/proline betaine-binding periplasmic protein (ProX) is one of three genes from operon VWX involved in binding compatible solutes with high affinity and specificity (Schiefner et al., 2004).
We also found genes related to acetoin and 2,3 butanediol production, which are volatile compounds (VOCs) involved in plant growth bacteria/fungi interaction as acetolactate synthase large and small subunit (Yi et al., 2016;Fincheira and Quiroz, 2018). VOCs are synthesized by the condensation of two pyruvate molecules into acetolactate by acetolactate synthase, which forms acetoin by acetolactate decarboxylase decarboxylation. The reduction of acetoin by acetoin reductase results in 2,3-butanediol (Suarez et al., 2019).
The strain Palotina contains phzF encoding phenazine biosynthesis. Phenazines can modify the cellular redox state by electron transport, acting in the cell signaling regulating gene expression. By contributing to biofilm formation and architecture, it can enhance bacterial viability in the rhizosphere (Pierson and Pierson, 2010).
Genome sequencing of a strain might provide more abundant screening tools for the PGPB, which could be readily detected in genomes (Finkel et al., 2017). The authors mentioned that the presence of minimal Nif cluster and genes required for indole acetic acid production are potent markers, albeit at variable levels, for screening potential strains, making the process faster and less labor extensive. Scagliola et al. (2016) affirmed that a potential PGPB candidate must have the ability to solubilize phosphate and iron (siderophores) and IAA. The data pointed to a PGP strain candidate and further studies should be conducted to reveal the full genetic mechanisms of plant interaction.

Supplementary Material
The following online material is available for this article: Table S1 -List of genes and their function in L. adecarboxylata strain Palotina. Table S2 -List of genes related to plant growth promotion in L. adecarboxylata strain Palotina. Figure S1 -Genomic map of CRISPR system. Figure S2 -Genomic map of RND efflux system. Figure S3 -Genomic map of Type I restriction-modification system.

Associate Editor: Célia Maria de Almeida Soares
License information: This is an open-access article distributed under the terms of the Creative Commons Attribution License (type CC-BY), which permits unrestricted use, distribution and reproduction in any medium, provided the original article is properly cited.