Ochrobactrum quorumnocens sp. nov., a quorum quenching bacterium from the potato rhizosphere, and comparative genome analysis with related type strains

Ochrobactrum spp. are ubiquitous bacteria attracting growing attention as important members of microbiomes of plants and nematodes and as a source of enzymes for biotechnology. Strain Ochrobactrum sp. A44T was isolated from the rhizosphere of a field-grown potato in Gelderland, the Netherlands. The strain can interfere with quorum sensing (QS) of Gram-negative bacteria through inactivation of N-acyl homoserine lactones (AHLs) and protect plant tissue against soft rot pathogens, the virulence of which is governed by QS. Phylogenetic analysis based on 16S rRNA gene alone and concatenation of 16S rRNA gene and MLSA genes (groEL and gyrB) revealed that the closest relatives of A44T are O. grignonense OgA9aT, O. thiophenivorans DSM 7216T, O. pseudogrignonense CCUG 30717T, O. pituitosum CCUG 50899T, and O. rhizosphaerae PR17T. Genomes of all six type strains were sequenced, significantly expanding the possibility of genome-based analyses in Ochrobactrum spp. Average nucleotide identity (ANIb) and genome-to-genome distance (GGDC) values for A44T and the related strains were below the single species thresholds (95% and 70%, respectively), with the highest scores obtained for O. pituitosum CCUG 50899T (87.31%; 35.6%), O. rhizosphaerae PR17T (86.80%; 34.3%), and O. grignonense OgA9aT (86.30%; 33.6%). Distinction of A44T from the related type strains was supported by chemotaxonomic and biochemical analyses. Comparative genomics revealed that the core genome for the newly sequenced strains comprises 2731 genes, constituting 50–66% of each individual genome. Through phenotype-to-genotype study, we found that the non-motile strain O. thiophenivorans DSM 7216T lacks a cluster of genes related to flagella formation. Moreover, we explored the genetic background of distinct urease activity among the strains. Here, we propose to establish a novel species Ochrobactrum quorumnocens, with A44T as the type strain (= LMG 30544T = PCM 2957T).


Introduction
Ochrobactrum spp., together with the closely related Brucella, Agrobacterium, and Rhizobium genera, belong to the class of Alphaproteobacteria [1][2][3]. Although the genus is most often associated with O. anthropi [4] and O. intermedium [1], which cause opportunistic infections in humans, the bacteria from the Ochrobactrum spp. genus adapted to a variety of environmental niches and can be found in soil [5], wastewater [6], in association with plants [6,7], and animals [8,9]. The ability of Ochrobactrum spp. members to utilize xenobiotic compounds led to their exploration as potential bioremediation agents [10][11][12][13] or a source of enzymes for the biotech industry [14,15]. Ochrobactrum spp. are also of interest as plant beneficial bacteria [16,17]. The plant-derived strains, such as O. lupini LUP21 T [7] and O. cytisi ESC1 T [18], are able to nodulate roots to fix nitrogen, underlining their close association with the host plants. Currently, the Ochrobactrum genus comprises 18 species [19] (Fig 1A). However, scientific data concerning the majority of the strains is limited to information of taxonomic value.
A44 T is a potato rhizosphere isolate, obtained in 2002 from the experimental field in Bennekom, the Netherlands, and studied for its ability to inactivate N-acyl homoserine lactones (AHLs) [20] via AHL-acylase activity [21]. The AHLs are small molecules produced, secreted, and auto-sensed by many Gram-negative bacteria in a regulatory mechanism called quorum sensing (QS). In QS, the expression of target genes is modulated in the presence of a threshold concentration of signal molecules in the immediate environment of the cell [22,23]. QS is known to govern vital metabolic processes in the AHL-utilizing bacteria, including production of virulence factors and biofilm formation. This enables strains like A44 T to deplete the pool of AHLs and interfere with QS of other species of scientific and applicational interest [24,25]. It was shown that through inactivation of AHLs, A44 T is able to attenuate the QS-dependent virulence of Pectobacterium parmentieri SCC3193 [26] and P. carotovorum EC71 [27] and hence protect the plant tissue from soft rot caused by these pathogens [20].
In this study, we establish the taxonomic position of Ochrobactrum sp. strain A44 T using sequence-based, genome-scale calculations, supported by the analysis of chemotaxonomic markers and phenotypic assays. Based on the results, we propose delineation of a novel species Ochrobactrum quorumnocens sp. nov. with A44 T (= LMG 30544 T = PCM 2957 T ) as a type strain. Furthermore, we perform comparative genome analysis of A44 T and the related Ochrobactrum spp. and use the genomic data to link chosen differential phenotypes for the studied group to their genetic background. The only study involving genetic analysis of Ochrobactrum spp. concerns variation among isolates of Ochrobactrum intermedium and was presented by Aujoulat et al. [28] and Kulkarmi et al. [29]. To our knowledge, this is the first attempt of such analyses for other members of this genus.

16S rRNA gene analysis and MLSA
First, we performed phylogenetic analysis of A44 T and the type strains of all 18 validly published Ochrobactrum spp. (Fig 1A; S1A and S1B Fig). Based on 16S rRNA gene sequence analysis, the closest relatives of A44 T were O. thiophenivorans DSM 7216 T , isolated from wastewater in Germany [6], O. pseudogrignonense CCUG 30717 T , isolated from a blood sample in Sweden [30], O. grignonense OgA9a T , obtained from bulk soil in France [5,30], O. rhizosphaerae PR17 T , originating from the roots of potato grown in Austria [6], and O. pituitosum CCUG 50899 T , isolated from industrial environment in Sweden [31] (Fig 1A; S1 Table). It was already shown that the listed A44 T -related strains cluster on a separate branch of the phylogenetic tree of Ochrobactrum spp. [30]. To obtain higher phylogenetic resolution within this group, we employed Multilocus Sequence Analysis (MLSA) [32] based on the concatenated sequences of 16S rRNA, gyrB, and groEL genes, which suggest that the closest relative of A44 T is O. rhizosphaerae PR17 T (Fig 1B).

Genome sequencing and genome-based phylogeny
At the time of this study, genome sequences of only two out of eighteen Ochrobactrum sp. type strains were publicly available, namely O. anthropi ATCC 49188 T (CP000758.1-CP000763.1) and O. intermedium LMG 3301 T (ACQA00000000.1), thereby limiting the application of the whole-genome resolution tools in Ochrobactrum taxonomy, as well as comparative genomics. Hence, we obtained the genome sequences of A44 T and the five closely related type strains of the  [33]. Basic features of the obtained genomes are presented in S2 Table. The complete genome of A44 T consists of four replicons: the main chromosome (2585.393 kbp; CP022604.1), with classical replication system, a chromid (2008.185 kbp; CP022603.1) carrying a plasmid-type replication system and the genes essential for the basic metabolism [34], [35], and two plasmids pOqn1 (1032.012 kbp; CP022605.1) and pOqn2 (19.701 kbp; CP022602.1). Genomes comprising several replicons were also reported for O. anthropi ATCC 49188 T [36] and the non-type O. pituitosum strain AA2 [37]. Interestingly, pOqn2 from A44 T , as well as the pOAN03 from O. anthropi ATCC 49188 T [36], lacks the typical replication/partition plasmid characteristics. Presence of multiple replicons, including chromids and megaplasmids, is widespread among the members of Alphaproteobacteria [35].
Out of the total of 5272 open reading frames comprising the genome of A44 T , 3329 (57.3%) have assigned function, 1443 (24.9%) were considered conserved hypothetical, and the remaining were either of unknown function: 546 (9.4%) or unclassified: 480 (8.3%). Detailed role category breakdown for all ORFs of A44 T genome is given in the S4 Until June 2018, sixty six genome assemblies for strains identified as belonging to the Ochrobactrum genus were available in GenBank (NCBI). Twenty two of the sequenced strains were classified only to the genus level. Others were assigned to certain species, however, most of them without documented phylogenetic studies. We used JSpeciesWS (http://jspecies. ribohost.com/jspeciesws/) to calculate ANIb value between the genome of A44 T and all 65 assemblies in order to find other strains potentially representing O. quorumnocens sp. nov. No other representatives of O. quorumnocens were found based on this search. The highest ANIb score (87.54%) was obtained between A44 T and strain SJY1 (AZRT01000000), designated as O. rhizosphaerae (S3 Table). Interestingly, the A44 T is also very closely related (ANIb>87%) to a set of Ochrobactrum spp. isolates recently obtained from the nematode Caenorhabditis elegans dwelling on a rotten apple (S3 Table). Data obtained in this work can significantly facilitate assigning those isolates to the respective species. We have also found that the ANIb value for O. rhizosphaerae SJY1 and PR17 T , the O. rhizosphaerae type strain, is 86.10%, thus below the single species threshold. ANIb calculation for O. rhizosphaerae strain SJY1 and O. pituitosum CCUG 50899 T suggests that SJY1 should be classified as O. pituitosum (ANIb = 96.97%).

Comparative genome analysis
Obtaining genome sequences of A44 T and the 5 related type strains (O. pituitosum CCUG 50899 T , O. rhizosphaerae PR17 T , O. grignonense OgA9a T , O. pseudogrignonense CCUG 30717 T , and O. thiophenivorans DSM 7216 T ) enabled us to perform comparative genome analyses using EDGAR [39]. Sequences uniformly annotated with Prokka [40] were used as the input files. Core genome for the six newly sequenced strains consists of 2731 coding sequences, making up 50-66% of each individual genome and 27% of the pan-genome (10296 CDSs) (Fig  2; S5 Table). The set of unique genes harbored by each strain varies from 354 for O. rhizosphaerae PR17 T to 817 for O. pseudogrignonense CCUG 30717 T (8-15%) (S6-S11 Tables-supplementary online content for the amino acid sequences). The remaining genes (32-46%) are shared by a subset of strains. Whereas hypothetical proteins constitute 25% of the core genome, this fraction is significantly enriched within the singletons subsets, ranging from 59% in O. thiophenivorans DSM 7216 T to 80% in O. rhizosphaerae PR17 T . The fact that each type strain of the analyzed group contributed new genes to the pan-genome suggests that the pangenome of Ochrobactrum spp. is "open" (as defined by Guimarães et al. [41]). However, this hypothesis needs to be confirmed on a broader pool of Ochrobactrum spp. genomes, as their availability in public databases continuously increases. Amino acid sequences encoded by 2604 core genes established for A44 T , its 5 closest relatives, and O. anthropi ATCC 49188 T were used to construct a core genome-based phylogenetic tree. The analysis was performed with EDGAR [40]. According to this approach, grouping of A44 T , O. rhizosphaerae PR17 T , O. grignonense OgA9a T , CCUG 50899 T , O. pseudogrignonense CCUG 30717 T , and O. thiophenivorans DSM 7216 T was analogous to that obtained by MLSA, with O. rhizosphaerae PR17 T indicated as the closest relative of A44 T (Fig 3). Flower plot depicting the core genome (in the center) and strain-specific genes (in the petals) for the group of six analyzed Ochrobactrum spp. type strains. Accession numbers for the sequences can be found in Table 2. Comparative genome analysis was performed using EDGAR. Total number of coding sequences identified by Prokka annotation tool is given next to designations of the particular strains.

Biochemical and phenotypic traits
A set of phenotypic assays was performed for A44 T and the closely related strains: Comparison of biochemical traits, determined with GEN III MicroPlates (Biolog, USA), revealed twelve differences between A44 T and O. rhizosphaerae PR17 T , twelve between A44 T and O. grignonense OgA9a T , fifteen for A44 T and O. thiophenivorans DSM 7216 T , sixteen for A44 T and O. pituitosum CCUG 50899 T , and seventeen for A44 T and O. pseudogrignonense CCUG 30717 T (S14 Table). Moreover, with the API 20NE test (bioMérieux, France), we determined that A44 T is negative for the production of urease, unlike the urease-positive O. thiophenivorans DSM 7216 T , and is unable to reduce nitrates to nitrites, as opposed to O. grignonense OgA9a T and O. pseudogrignonense CCUG 30717 T . Within the Ochrobactrum genus, urease activity can vary between species and should not be considered as a criterion for genus identification [43]. In this work, strain A44 T , along with the type strains of O. pseudogrignonense CCUG 30717 T , O. rhizosphareae PR17 T , O. grignonense OgA9a T , and O. pituitosum CCUG 50899 T , was shown to be urease-negative on API strips and in the urea-indole medium. On the contrary, strain O. thiophenivorans DSM 7216 T , for which no previous data concerning urease activity had been available, gave a positive reaction in the urease assays (S3 Fig).
Motility of the strain, its closest relatives, and O. anthropi ATCC 49188 T , was assessed in glass tubes assay after 4 days incubation at 28˚C. Strain A44 T was found to be motile under all tested conditions. The same was observed for O. rhizosphaerae PR17 T and O. grignonense OgA9a T , whereas O. pseudogrignonense CCUG 30717 T was motile only in the presence of casamino acids, and O. thiophenivorans DSM 7216 T was immotile irrespective of the tested conditions. O. anthropi ATCC 49188 T was motile only in the presence of glucose (with and without casamino acids) but not in glycerol, unless supplemented with casamino acids (S15 Table).
We  Fig). When grown for 5 days in the LB medium containing varying concentrations of NaCl, all tested strains could grow in medium containing up to 4.5% NaCl (S16 Table). However, the growth of A44 T , O. pseudogrignonense CCUG 30717 T , O. thiophenivorans DSM 7216 T , and O. anthropi ATCC 49188 T was less affected by the increasing concentration of NaCl than it was in the case of other investigated strains. Only those four microorganisms showed any growth at 6% of NaCl, with O. anthropi being the most resistant to salinity. According to the results from GEN III Micro-Plates (Biolog, USA), none of the tested strains (viz. rhizosphaerae PR17 T ) could grow at pH 5, and all strains, apart from O. rhizosphaerae PR17 T , could grow at pH 6 (S14 Table).
Considering that A44 T was previously studied for its ability to inactivate the AHL signal molecules, we verified the potential of the related Ochrobactrum sp. type strains to inactivate AHLs. Interestingly, O. rhizosphaerae PR17 T was found to be unable to interfere with quorum sensing. Using a modified assay as described by Jafra and van der Wolf [44] (Fig 4).
In line with this finding, strain O. rhizosphaerae PR17 T showed a significantly lower potential to attenuate soft rot caused by P. parmentieri SCC3193 when compared to A44 T , as demonstrated in a potato tuber slices assay conducted according to Jafra et al. [20] (Fig 5).
List of the selected phenotypic traits which can be used for efficient discrimination between A44 T and the related strains is given in Table 2. Ochrobactrum quorumnocens sp. nov. and comparative genome analysis of Ochrobactrum spp. type strains

Determining the genetic background of distinct urease activity and motility among strains
It is often found in literature that the presented results of genome analyses or genome comparisons are limited to 'dry' numeric data sets. To go a small step beyond this scenario, we investigated the genetic background for two phenotypes with differentiating values for the studied group of Ochrobactrum spp. strains.
Urease-related genes. Ureases are enzymes which catalyze the hydrolysis of urea to ammonia and carbon dioxide. Bacterial ureases are complex nickel-dependent enzymes, the synthesis of which requires three structural genes, ureABC, encoding gamma, beta, and alpha subunits, respectively, and four accessory genes, ureDEFG, involved in the protein assembly. We investigated the presence and the relative position of the ureABCDEFG genes in the six newly-sequenced Ochrobactrum spp. genomes. We found that all analyzed strains, irrespective of being urease-positive or -negative, possessed homologues of genes from the ure operon. However, the organization and the genomic context of the urease clusters showed high diversity (Fig 6).
Among the urease-negative strains, O. quorumnocens A44 T , O. rhizosphaerae PR17 T , and O. grignonense OgA9a T had a single urease cluster (further referred to as type I) with a similar gene organization. In this type of cluster, apart from ureABCDEFG, also the ureJ gene, encoding transmembrane hydrogenase/urease accessory protein of unclear function, was identified [46]. O pituitosum CCUG 50899 T and O. pseudogrignonense CCUG 30717 T were found to have two urease clusters-one of type I and a second, distinct operon (type II). The type II cluster had a different organization of genes and contained an urea transporter. For the structural proteins UreA, UreB and UreC, the identity was higher for homologues derived from the same cluster type, although from different strains, than between homologues from different cluster types. For instance the identity of protein homologues for type I clusters from of A44 T and O. pseudogrignonense CCUG 30717 T was 98% for UreC and 95% for both UreA and UreB, in comparison to 70%, 71% and 66%, respectively, for the homologues of these proteins from clusters I and II of O. pseudogrignonense. intermedium M86, and Brucella abortus 2308 (Fig 5). O. anthropi ATCC 49188 T harbors a single cluster resembling that of type I, however settled in a different genetic context. In O. intermedium M86, single homologues of the ure genes could be found as well, as reported earlier by Kulkarni et al. [47]. We found that the ure operon in O. intermedium M86 is similar to that of type II from O. pituitosum CCUG 50899 T and O. pseudogrignonense CCUG 30717 T , yet deprived of the ureE and ureG accessory genes. The missing genes could be located in a remote region of the genome. Neither cluster type I nor II of the studied Ochrobactrum spp. strains resembled the two urease clusters present in B. abortus 2308 [48]. Difference between the urease type from O. intermedium M86 and that of Brucella sp. has been previously reported by [49].
O. thiophenivorans DSM 7216 T , the only urease positive strain of the close relatives of A44 T harbors a single urease cluster, however of an entirely different type than those observed for the other analyzed Ochrobactrum spp. (Fig 5). We found that the protein encoded by the ureA allele from O. thiophenivorans DSM 7216 T , while showing >58% identity to that of other Ochrobactrum spp. strains, shows high homology to that of a particular group of saline waterisolated alphaproteobacterial strains: Pseudoruegeria sp. SK021 (query coverage 91%, identity 91%), Martelella mediterranea (91%, 89%) and Nitratireductor sp. OM-1 (91%, 88%) [50]. Further analysis revealed that the nearly 7 kb fragment containing the urease cluster from O. thiophenivorans (nine genes, from ureA2 to the ammonium transporter encoding gene), shows  Ochrobactrum quorumnocens sp. nov. and comparative genome analysis of Ochrobactrum spp. type strains 91% query coverage and 78% identity at the nucleotide level with an analogous cluster from Pseudoruegeria sp. SK021 (MTBG01000044.1: 6099-12600), isolated from the North Sea sediment [51]. Thus, it is highly possible that the unique, in terms of Ochrobactrum spp., urease cluster was obtained by O. thiophenivorans via horizontal gene transfer. We may only hypothesize that O. quorumnocens A44 T and the related Ochrobactrum spp. are urease-negative despite harboring urease-encoding operons due to the point mutations, frameshifts, or deletions/insertions that have accumulated in the (ancestor) clusters. The activity of urease results in increase in pH due to production of ammonia. For human pathogens such as Helicobacter pylori, Yersinia enterocolitica, and Brucella sp. urease activity increases the survival rate in the acidic environment of the host's gastric tract [48,51,52]. Analogous function was proposed for ureases produced by clinical isolates of Ochrobactrum spp. which, like H. pylori, can also be found in urease-positive gastric biopsies [47,53]. Also, the non- The clusters were divided into two groups according to their origin from urease-negative (I) and urease-positive (II) strains, as determined in a diagnostic urease assay. ORF colors represent: structural genes (red), accessory genes (blue), transporter genes (green), nickel metabolism related genes (yellow), tRNA encoding regions (violet). Differentiating design patterns were used for the structural genes ureC, ureA, and ureB to indicate different/similar alleles. Matching background colors were used for clusters/cluster regions showing structural resemblance. ORF numbers represent the respective loci in the genomes of the analyzed strains. https://doi.org/10.1371/journal.pone.0210874.g006 Ochrobactrum quorumnocens sp. nov. and comparative genome analysis of Ochrobactrum spp. type strains pathogenic soil microorganisms including Rhizobium spp. [54] produce ureases. Mineralization of nitrogen from urea by the soil microbiota makes it more accessible to the plants and is essential for completing the nitrogen cycle. This includes both the naturally occurring urea and the urea-based fertilizers-the most widely used nitrogen fertilizers in agriculture (http:// faostat.fao.org). Gram-negative bacteria can obtain nitrogen both from organic compounds, such as amino acids, and from inorganic ammonia-the decomposition product of urea [55].
Flagella-related genes. O. thiophenivorans DSM 7216 T was shown to be non-motile, both in the original study [6] and under the conditions tested in this work. To investigate the genetic background of this phenotype, we searched the genomes of all six strains for genes related to flagella formation [56]. Our results showed that O. thiophenivorans DSM 7216 T lacks 12 out of 25 investigated genes, 11 of which were among the 24 genes present in every other strain ( Table 3).
The missing genes include fliC encoding flagellin, flgE and flgK encoding hook-associated proteins, and fliF, fliQ, and fliR encoding flagellar biosynthetic proteins. In the motile strains, 11 out of the 12 missing genes occur in a cluster, indicating that the loss of motility by O. thiophenivorans was due to a major deletion of genes essential for flagella formation. Ochrobactrum quorumnocens sp. nov. and comparative genome analysis of Ochrobactrum spp. type strains

Formal description of Ochrobactrum quorumnocens sp. nov.
Ochrobactrum quorumnocens sp. nov. (quo.rum.nocens L. neut. noun, quorum quorum; L. present participle nocens, from L. verb noceo incapacitating, referring to the fact that the type strain was initially studied for its ability to perform quorum quenching through inactivation of AHL signal molecules). Cells are Gram-negative, short non-spore-forming rods (1.4 μm in length and 0.9 μm in width), occurring singly and motile by a unipolar flagellum (S5 Fig), urease negative, oxidase and catalase positive, and non-hemolytic on Columbia blood agar. Colonies grown on LB agar are of round shape, smooth surface and edges, low-convex, shiny, and exhibit light beige color. After 5 days of growth at room temperature (22˚C) on Tryptic Soy Broth agar, the colonies may become light orange. Growth occurs in 0-4.5% (w/v) NaCl (optimum below 3%), at pH 5.5-10, and temperature range of 7-37˚C (optimum around 28˚C). The strain grows on McConkey agar but does not ferment lactose. The predominant cellular fatty acid is C 18:1 9 . In line with the new standards [57], biochemical traits for the formal description of the strain are given in tabular format (S17 Table).
The type strain A44 T (= LMG 30544 T = PCM 2957 T ) was isolated from the rhizosphere of a potato plant grown in a field in the Netherlands. Genome sequence of the type strain is available from DDBJ/EMBL/GenBank under the accession CP022602.1-CP022605.1. It comprises a main chromosome (2.59 Mbp), a second chromosome (chromid) (2.01 Mbp), and two plasmids (1.03 Mb and 0.02 Mbp). The average G+C content is 53.16%. A44 T is able to inactivate AHLs, which are the signal molecules of many Gram-negative bacteria. The formal proposal of the new species Ochrobactrum quorumnocens sp. nov. is given in S17 Table with the Taxonumber TA00464 (http://imedea.uib-csic.es/dprotologue/).

Conclusions
Ochrobactrum spp. receive growing interest as members of plant and nematode microbiomes and as potent tools in biotechnology. In this study, on the basis of sequence-based and phenotypic analyses, we propose to establish the QS-interfering strain A44 T as a type strain of a novel species: Ochrobactrum quorumnocens sp. nov. (Taxonumber TA00464). The strain possesses a multi-replicon genome, characteristic for this genus. Before this study, genomes of only 2 out of 18 Ochrobactrum spp. type strains were available. Obtaining draft genomes of 5 of A44 T -related type strains significantly contributes to the application of golden standard, genome-based approaches in taxonomy of this genus, especially in the light of an increasing number of genome assemblies for new strains, provisionally classified as Ochrobactrum sp. To our knowledge, this is the first study in which comparative genome analyses was performed for members of different Ochrobactrum species. Moreover, phenotype-to-genotype approach revealed the genetic background of the lack of motility in O. thiophenivorans DSM 7217 T and showed the diversity of urease cluster organization in the studied group of Ochrobactrum spp. strains. Kämpfer (Justus-Liebig-Universität Gießen, Germany). For routine propagation, all strains were cultured in Miller's Lysogeny Broth (LB) or on LB solidified with agar (Novagen/Merck, Germany). Ochrobactrum spp. strains and P. parmentieri SCC3191 were grown at 28˚C. Escherichia coli [pSB401] was grown at 37˚C, with the addition of tetracycline (15 μg�mL -1 ). E. coli [pSB401] was used as a lux-based biosensor producing bioluminescence in the presence of AHLs [45]. The ability of Ochrobactrum spp. strains to grow at 7, 10, 20, 28, 37, and 42˚C for a total of 7 days was assessed on LB agar plates, spot-inoculated with 2 μL of saline suspension of bacterial cells (turbidity adjusted to 0.3 McF). Additionally, the growth rate at 37˚C was assessed in the LB medium for a total of 25 h, in 3 biological replicates, 4 technical replicates each. Salt tolerance was determined in LB base (10 g�L -1 peptone, 5 g�L -1 yeast extract) supplemented with NaCl concentrations ranging from 0 to 9%, in 3 replicates. All growth-related experiments in LB were performed in 96-well format, with orbital shaking (120 rpm) and 200 μl of medium inoculated with 4 μl of cell suspension (0.3 McF) per well. EnVision Multilabel Reader (PerkinElmer, USA) was used to measure OD 600 of the cultures. The pH growth range of A44 T was determined using the PM10 MicroPlate (Biolog, USA), in two replicates.

16S rRNA gene analysis and MLSA
Phylogenetic analyses were performed with the use of MEGA 7 software [58]. Dendrograms for the near complete 16S rRNA gene sequences (1335 bp) of A44 T and the type strains of all 18 validly published Ochrobactrum spp. were obtained using the three clustering algorithms: maximum likelihood based on the Tamura 3-parameter model, neighbor joining, and maximum parsimony. All positions containing gaps and missing data were eliminated. There was a total of 1331 positions in the final dataset [59,60]. Multilocus Sequence Analysis (MLSA) [32] was performed for a concatenated set (3512 bp) of three genes: 16S rRNA gene (1335 bp), gyrB (1012 bp), and groEL (1165 bp).

Genome sequencing and genome-based phylogeny
Genome sequencing of Ochrobactrum sp. type strains, with the exception of O. pituitosum CCUG 50899 T , was performed at BaseClear B.V., the Netherlands. Draft genome of O. pituitosum CCUG 50899 T was obtained by a combined use of Illumina MiniSeq and Oxford Nanopore MinION platforms at IFB UG & MUG, Poland. Sequence of O. pituitosum CCUG 50899 T was automatically annotated using the NCBI Prokaryotic Genome Annotation Pipeline [61]. All other sequences were annotated using the IGS Annotation Engine [33] (Institute for Genome Sciences, University of Maryland School of Medicine, USA). The whole-genome average nucleotide identity (ANIb) was calculated with the use of JSpecies software with default settings [62]. Digital DNA-DNA hybridization was carried out with Genome-to-Genome Distance Calculator 2.0 (GGDC) using the recommended BLAST+ alignment with formula 2 (identities/HSP length) (http://ggcd.dsmz.de/distcalc2.php) [63]. The single species thresholds recommended ANIb and GGDC are 95% [62] and 70% [63], respectively. Analysis of the functional annotation was done using the Kyoto Encyclopedia of Genes and Genomes (KEGG) (https://www.genome.jp/kegg/) [38].

Comparative genomics
CDS (CoDing Sequences) count for O. pituitosum CCUG 50899 T was obtained using Prokka [40]. Other CDS counts were derived from the respective GenBank files. Comparative genome analysis was performed using EDGAR plaftorm (http://edgar.computational.bio) [39]. The core genome and the singletons for a set of 6 related Ochrobactrum spp. strains were calculated for Prokka-annotated genomes using EDGAR (http://edgar.computational.bio).

Determination of chemotaxonomic markers
Bacterial fatty acids were obtained from the whole bacterial cells grown overnight on the LB medium at 28˚C. The fatty acids were extracted according to the method described by Bligh and Dyer, 1959 [42] and subjected to gas chromatography and mass spectrometry (Shimadzu QP 2010 SE system). Identification of FAMEs was performed with respect to standards (Bacterial Acid Methyl Esters, Sigma-Aldrich, USA). For the whole-cell MALDI-TOF mass spectrometry profiles, bacteria were grown overnight on LB agar plates at 28˚C. The analysis was performed in ferulic acid (10 mg�mL -1 ) dissolved in 17% formic acid, 33% acetonitrile and 50% water as matrix. Protein mass fingerprints were obtained using the MALDI-TOF/TOF 5800 mass spectrometer (AB Sciex, Framingham, MA, USA), with detection in the linear middle mass (4000-20000 Da), positive ion mode for a total of 1000 laser shots by a 1 kHz Opti-Beam laser (YAG, 349 nm). Registered spectra were analyzed with Data Explorer software (AB Sciex).

Biochemical and phenotypic assays
Biochemical traits and enzymatic activities were determined using GEN III Microplates (Biolog, USA) and API 20NE tests (bioMérieux, France) according to the manufacturer's protocols, in two independent experiments. The urease activity of the tested strains was also verified, in two biological replicates, using the urease-indole diagnostic broth (BioMaxima, Poland). The results were collected after 24 and 48 hours of incubation at 28˚C. The Columbia blood agar (BTL, Poland) was used to assess hemolytic activity. The MacConkey agar from BTL (Poland) was used to determine the growth on this medium. Catalase activity was determined by assessing the generation of oxygen bubbles in the presence of 3% (v/v) hydrogen peroxide. Oxidase activity was tested using plastic strips with a paper zone saturated with N,Ndimethyl-p-phenylenediamine oxolate and α-naphthol (Sigma-Aldrich, USA). LB agar and Tryptic Soy Agar TSA (Thermo Fisher Scientific, USA) were used for determination of colony morphology. Motility was assessed in glass tubes filled with 5 mL of the M9 minimal medium [64] containing 0.3% agar supplemented with 0.4% glycerol or 0.4% glucose, and with or without the addition of 0.5% casamino acids. Media in tubes were stab inoculated and incubated at 28˚C for 4 days.

Cell micrography
Imaging of the A44 T cells was performed with the use of atomic force microscope (AFM). Bacteria were grown in the M9 minimal medium supplemented with 0.4% glucose, overnight on a microscopy glass cover slide placed in a Petri dish, with gentle shaking (60 rpm) at 28˚C. The slides were washed with water, samples were fixed with 2.5% glutaraldehyde (Sigma Aldrich) for 2 hours, washed again and dried in air. Cells were imaged in air using Bioscope Resolve (Bruker), in ScanAsyst (Peak Force Tapping) mode, with the application of ScanAsyst Air probe (f0 7.0 kHz, diameter <12 nm, k:0.4 N/m).

AHL inactivation assay
The ability of the tested Ochrobactrum sp. type strains to inactivate AHLs was determined in a modified assay described by Jafra and van der Wolf [44]. Briefly, bacterial cells from overnight cultures were harvested and re-suspended in the MOPS-buffered LB, pH 6.5. Turbidity was adjusted to 5 units in McFarland scale (McF) (~3.5×10 8 cfu�mL -1 ). 50 μl of cell suspension was added to 50 μl of 10 μM C6-HSL (Sigma-Aldrich, USA) in M63 0.12% agar in white/clear bottom 96-well plates [36]. The plates were incubated for 16 h at 28˚C, without shaking. The remaining AHLs were detected using E. coli [pSB401] biosensor [45], emitting light in the presence of C6-HSL.

Plant protection assay on potato tuber slices
Potato tuber slices assay was performed as previously described [20]. Ware (table) potato tubers of cultivar Gala were obtained in local grocery stores (Gdansk, Poland).

Genome mining for urease and flagella-related genes
BLAST search tools [65] available from NCBI were used to determine the presence of urease and flagella-related genes in the genomes of A44 T and the related Ochrobactrum spp. (Gen-Bank accession numbers for genomes given in Table 1). For O. pituitosum, annotation of clusters A (extracted from PYSY02000003.1) and B (extracted from PYSY02000002.1) was performed with RAST: (http://rast.nmpdr.org/). The CLC Main Workbench 7 (QIAGEN Bioinformatics) and PowerPoint 2016 (Microsoft) were applied to visualize the urease clusters.