Evaluation of sequence-based tools to gather more insight into the positioning of rhizogenic agrobacteria within the Agrobacterium tumefaciens species complex

Abstract

Rhizogenic Agrobacterium, the causative agent of hairy root disease (HRD), is known for its high phenotypic and genetic diversity. The taxonomy of rhizogenic agrobacteria has undergone several changes in the past and is still somewhat controversial. While the classification of Agrobacterium strains was initially mainly based on phenotypic properties and the symptoms they induced on plants, more and more genetic information has been used along the years to infer Agrobacterium taxonomy. This has led to the definition of the so-called Agrobacterium tumefaciens species complex (Atsc), which comprises several genomospecies. Interestingly, the rhizogenic Agrobacterium strains are found in several of these genomospecies. Nevertheless, even up until today Agrobacterium strains, and in particular rhizogenic agrobacteria, are prone to misclassification and considerable confusion in literature. In this study, we evaluated different phylogenetic analysis approaches for their use to improve Agrobacterium taxonomy and tried to gain more insight in the classification of strains into this complex genus, with a particular focus on rhizogenic agrobacteria. The genome sequence analysis of 580 assemblies, comprising Agrobacterium, Allorhizobium and Rhizobium strains demonstrated that phylogenies based on single marker genes, such as the commonly used 16S rRNA and recA gene, do not provide sufficient resolution for proper delineation of the different genomospecies within the Atsc. Our results revealed that (in silico) multi-locus sequences analysis (MLSA) in combination with average nucleotide identity (ANIb) at a 94.0% threshold delineates genomospecies accurately and efficiently. Additionally, this latter approach permitted the identification of two new candidate genomospecies.

Introduction

The genus Agrobacterium is part of the Alphaproteobacteria class, more specifically the Rhizobiaceae family [1], and consists of rod-shaped, non-spore forming, gram-negative bacteria [2]. Agrobacterium is a very diverse genus comprising strains that can live in different environments, occurring mainly in soil and rhizosphere and even being capable of surviving under oligotrophic conditions such as pure water [3]. The diversity of this genus is also exemplified by the diverse effects it exerts on plants, with several Agrobacterium spp. being beneficial or non-pathogenic, while other species induce severe plant diseases, such as crown gall disease or hairy roots disease (HRD) on a wide range of host plants, often resulting in a considerable decrease in yield and associated economic losses [4–7].

Based on symptoms induced after infection, this genus was initially divided in the following three species: (i) Agrobacterium tumefaciens, causing tumours in a wide range of host plants; (ii) Agrobacterium rhizogenes, causing hairy roots in host plants; and (iii) Agrobacterium radiobacter, being non-pathogenic strains [9–11]. At that time, pathogenicity was considered the single trait to assign strains to specific Agrobacterium species [8–12]. Later on, three new species were added to the genus: Agrobacterium larrymoorei, which is capable of infecting figs (Ficus spp.) [13, 14]; Agrobacterium rubi, capable of infecting berries (Rubus spp.) [15, 16]; and Agrobacterium vitis, causing woody galls on grapevine (Vitis spp.) [17]. However, it was observed later on that pathogenicity was determined by the type of plasmid carried by the pathogenic strain, i.e., the tumor-inducing plasmid (pTi) for tumorigenic strains or the root-inducing plasmid (pRi) for rhizogenic strains [18, 19].

Subsequently, Agrobacterium species were classified into three biovars, mainly based on biochemical and serological tests [20, 21]. It should be noted that in the case of Agrobacterium, the term “biovar” does not have the usual meaning of a specific phenotypic form within a species, but rather corresponds to biological species within the genus [22]. Because the delineation in biovars is based mainly on the results of biochemical tests, rhizogenic, as well as tumorigenic and non-pathogenic strains, could be assigned to more than one biovar.

With the dawn of the genomic era, Young et al. [12] proposed to merge the Agrobacterium genus with that of Rhizobium based on 16S rDNA (rss) analysis, since both formed a monophyletic clade, along with Allorhizobium [12]. However, this suggestion was challenged by Farrand et al. [23] who claimed that only biovar 2 strains belong to the genus Rhizobium, giving birth to the name of Rhizobium rhizogenes. The epithet “rhizogenes” did not imply pathogenic properties [1]. Eventually, biovar 3 strains were reclassified from A. vitis into Allorhizobium vitis [17, 24]. All these changes resulted in the fact that strains initially classified into three different biovars of the same genus, ended up being classified into three different genera: biovar 1 strains were classified as A. tumefaciens, biovar 2 strains as R. rhizogenes, and biovar 3 strains as All. vitis [25, 26]. This separation has been further confirmed by the absence of a linear chromid in R. rhizogenes and All. vitis strains, which is considered a critical taxonomic characteristic for most Agrobacterium strains, except for A. divergens [27–29].

Considering the complexity of classification of bacterial strains causing HRD, as well as the corresponding genera described above, there is clearly a need for tools to unequivocally assign strains to dedicated genera and species. Deeper taxonomic analyses, including the analysis of specific marker genes such as genes encoding the 16S rRNA and a DNA recombinase (recA), revealed that former Agrobacterium biovar 1 is not a homogeneous species but a complex of several genomic species, also called genomovars or genomospecies [30]. For this reason, it has been proposed to rename Agrobacterium biovar 1 to ‘A. tumefaciens species complex’ (Atsc), until these genomospecies are unequivocally named and delineated [30]. To date, the Atsc contains 15 genomospecies (S1 Table), referred to as G1 to G9, G11, G13 to G15, G19, and G20 [31, 32]. Until now, three genomospecies have not received yet binomial name (G5, G6, and G13) [1, 32, 33], while for 12 genomospecies either a valid (formal) name has been assigned, or a species name has been suggested (S1 Table).

However, neither 16S rRNA nor recA marker gene phylogeny provide the resolution required to unequivocally classify strains into Agrobacterium genomospecies. To further scrutinize taxonomy in such complex genera, multi-locus sequence analysis (MLSA), comparative genome hybridization (CGH), or core-genome phylogeny have been used [30, 34–37]. More and more, overall genome relatedness indices (OGRI) that use genome-wide comparisons are commonly used to delineate prokaryotic species, such as ANI using BLASTN (ANIb) [26, 38–41].

In conclusion, even though recently a set of guidelines have been proposed for the description of new genera and species of rhizobia and agrobacteria [42], they have not been applied on a large collection of Agrobacterium strains to confirm their accuracy. Moreover, the taxonomic position of rhizogenic strains has also been poorly investigated with only a few studies available. The main objectives of this study were therefore to: (i) gain more insight into the taxonomic classification of Agrobacterium strains; (ii) evaluate and compare the ability of different DNA-sequence based tools for the delineation of (genomo)species, in particular within the AtSC; and (iii) examine the positioning of rhizogenic agrobacteria within the Rhizobiaceae family. To achieve this, we compared 580 publicly available genome sequences by means of different genetic analyses, including 16S rRNA and recA marker gene phylogeny, MLSA (including the housekeeping genes 16S rRNA, atpD encoding for the ATP synthase B subunit, gyrB encoding for a DNA gyrase subunit, recA, and rpoB encoding for an RNA polymerase subunit), WGS phylogeny, and OGRI values, such as ANIb, AAI, and nonmetric multidimensional scaling (NMDS) analysis based on and codon usage. Additionally, to increase our understanding of phylogenetic relationship of rhizogenic Agrobacterium strains, we included a selection of 24 (in-house) sequenced strains that were isolated from hydroponic crops showing HRD symptoms.

Materials and methods

Genome sequencing and collection of publicly available genome sequences

A total of 580 genomes were retrieved, including 579 quality-checked (based on NCBI metadata annotation for contaminated, unverified, or fragmented) genome sequences were downloaded from the NCBI database in September 2023 using the get_assemblies v.0.10.0 package (Davis, 2020), and an additional genome retrieved rom MGMI database [43], including 413 strains classified as Agrobacterium, 8 as Allorhizobium (ex A. vitis), and 159 as Rhizobium (S2 Table). Additionally, the genome assembly of strain Agrobacterium sp. R-31762, identified as the type strain for A. divergens, with accession code Ga0496675 was manually retrieved from the Integrated Microbial Genomes & Microbiomes database (IMG/M). The main focus was on strains previously classified as Agrobacterium and/or identified as rhizogenic strains (classified as Agrobacterium spp. Or R. rhizogenes strains). The binomial names associated with each assembly, as retrieved in the NCBI database, were respected to facilitate retrieving info on these strains. However, it has to be noted that the species names of the genomes uploaded in NCBI are not always correct anymore due to the changes in Agrobacterium taxonomy. To avoid confusion the up-to-date and correct species names are also added in S2 Table. In addition, the genome sequence of type strains of several Rhizobium and Agrobacterium strains were included to validate the confidence of the phylogenetic trees (S2 Table).

This collection of genome assemblies included a representative set of 24 rhizogenic Agrobacterium strains previously isolated from HRD-infested greenhouses from different plant hosts (tomato, cucumber and melon) and countries across Europe, the United States, Canada and Japan. [44, 45] (S2 Table). For details regarding sequencing and assembly, we refer to Kim et al. [45]. Additionally, three Ensifer adhaerens assemblies (Rhizobiaceae, Hyphomicrobiales, Alphaproteobacteria) were used as an outgroup for the downstream analyses (S2 Table). As a result, the complete set of genome assemblies used in downstream analysis was composed of 583 strains.

Construction of phylogenetic trees based on 16S rRNA gene, recA gene, multilocus sequence analysis (MLSA), and whole genome sequences

Two phylogenetic trees were constructed based either on the single marker gene 16S rRNA (1,515 bp) or recA (1,092 bp). In addition, an MLSA-based tree was constructed based on a concatemer of the 16S rRNA gene and four housekeeping genes, i.e., atpD (1,485 bp), gyrB (2,442 bp), recA (1,092 bp), and rpoB (4,149 bp), yielding a final concatemer of 10,683 bp. These trees were built using the autoMLSA2 workflow developed by Davis et al. [46]. Briefly, the DNA sequence of the individual marker genes of reference strain Agrobacterium NCPP2659 were downloaded from the NCBI GenBank database and used as query to retrieve homologous genes from the genome assemblies of the other 581 strains used in this study (S3 Table). For the individual marker genes as well as for the concatemer, the sequences were aligned using MAFFT v.7 [47] with default parameters, and a phylogenetic tree was generated using IQTREE v.2.2 [48–50]. The -program parameter was set to “BLASTN”. When more than one copy of the query gene 16S rRNA was found, only the highest scoring BLASTN hit was retained. The phylogenetic tree was constructed applying the Maximum Likelihood (ML) method and was based on the most optimal model found and a bootstrap of 1,000 replicates.

Whole genome-based phylogeny was constructed using the Phylophlan v.3.0 package [51]. First, a database of UniRef90 core proteins for Agrobacterium was generated by setting the ‘g’ parameter into s__Agrobacterium_tumefaciens, which resulted in 1,524 markers. Next, a configuration file was generated indicating that the mapping of the database markers to the input genomes should be conducted using DIAMOND [52], the multiple sequence alignment should be constructed with MAFFT, that sequences should be trimmed using TrimAl, and that the trees should be first generated using FastTree and then enhanced using RaxML. Subsequently, the phylogeny was built by indicating the parameter–‘diversity’ as “low” since the genomes analysed were of closely related species. The resulting phylogenetic tree was recalculated using 1,000 bootstrap replicates. The resulting phylogeny was constructed using a ML method and the CAT General Time Reversible (GTRCAT) approximation was used as nucleotide substitution model. The obtained phylogenetic trees were annotated using the iTOL platform [53]. If available, genomospecies assignments were adopted from Costechareyre et al. [30] and Bosmans et al. [44].

Four phylogenetic signal indices (Blomberg’s K, Pagel’s l, Moran’s I and Abouheif’s C_mean) were calculated in order to determine if variation in the genomic guanine-cytosine (GC) content among agrobacterial strains was linked to their phylogenetic affiliation. These phylogenetic signal analyses used as inputs: (i) the ML MLSA tree trimmed to include only Agrobacterium strains as well as the Ensifer outgroup; (ii) a table including all the GC content percentages, obtained from the assembly entries at the NCBI (S4 Table). The phylosig() function of the R package “picante” v.1.8.2 was used to calculate Blomberg’s K and Pagel’s l [54], while the R packages phylobase v.0.8.10 [55] and abouheif.moran() function of the R package “adephylo” v.1.1–13 were used to calculate Moran’s I and Abouheif’s C_mean (method = “Abouheif” and “oriAbouheif”, respectively) [56], all considering 1000 simulations.

Genome relatedness indices calculation

All genomes were compared in a pairwise manner to evaluate their phylogenetic relationship based on two OGRIs, ANIb and average amino acid identity (AAI). ANIb values were calculated using the average_nucleotide_identity.py script from pyani v.0.2.10 [57]. The threshold applied initially to delineate species was 95% based on López-Guerrero et al. [58]. This threshold was further fine-tuned in this study using the bactaxR package by identifying medoid genomes, i.e., genomes presenting the smallest average dissimilarity among themselves, using different thresholds [59]. AAI values were calculated using CompareM v.0.1.1 [60] under default settings using the aai_wf command, which uses protein coding sequences (CDS) predicted by Prodigal [61]. Then an all-vs.-all reciprocal sequence similarity search was conducted by DIAMOND [81] and finally pairwise AAI values based on the orthologous fraction of the two genomes was obtained. The threshold applied to delineate species in the AAI analysis was 95% based on Fan et al. [62]. Resulting distance matrices were visualized with heatmaps generated with the R package pheatmap v.1.0.12 in the RStudio v.3.3.0 platform [63–65].

Analysis of codon usage patterns

Possible differences in the codon usage patterns among ANIb groups were visualized by principal component analysis (PCA), as implemented by the prcomp() function of the ‘stats’ R library [63]. The input data for the PCA was a data frame of the normalized relative synonymous codon usage (NRSCU) values calculated for the genomes included in this study using the Dynamic Codon Biaser webserver (http://www.cbdb.info/; [66]. PCA loading plots and scores plots were created using the ‘ggplot2’ R library [67].

Results

Phylogenetic analysis of single marker genes

To have a comprehensive understanding of the phylogenetic relationships within the Agrobacterium genus, and the position of rhizogenic agrobacteria in the Rhizobiaceae family in particular, we analysed the genome sequence of 583 strains, including 24 in-house sequenced rhizogenic Agrobacterium strains. In a first approach, the phylogenetic relationships between strains, and more in particular, the association with previous genomospecies assignment was assessed based on 1,515 bp of the 16S rRNA gene (S1 Fig). The best substitution model for the 16S rRNA tree was Transition 3 model with empirical base frequencies and invariable site plus FreeRate with three categories model (TIM3+F+I+R3), based on Akaike information criterion (AIC) and Bayesian information criterion (BIC) scores of 17094.858 and 17158.736, respectively.

The 16S rRNA-based phylogeny showed that almost all strains clustered in three well-supported clades (>70% bootstrap values, 94.2% sequence similarity): (i) the R1 clade containing all the Rhizobium strains; (ii) a homogenous AV clade containing all the All. vitis (former A. vitis) strains; and (iii) the Agrobacterium clade containing all bona fide Agrobacterium strains. When zooming in on the R-clade, a subclade (R2) containing very closely related strains (>80% bootstrap values, 98.2% similarity) could be observed. This subclade contained all strains classified as R. rhizogenes (former A. rhizogenes or Agrobacterium biovar 2), as well as four Agrobacterium sp. strains. The R2 subclade also included R. radiobacter K84^T, and the type strain of R. rhizogenes (i.e. strain NBRC 13257^T = LMG150^T). In the Agrobacterium clade, seven well-supported subclades were identified (>80% bootstrap values, 96.8% similarity). The first cluster (Agr1_{16S rRNA}) included strains that are part of the so-called Agrobacterium rubi-larrymoorei subclade, namely A. vaccinii, A. rubi, A. skierniewicense, A. rosae, and A. bohemicum strains as well as their respective type strains. The other clusters (Agr2-7_{16S rRNA}) contained strains of different genomospecies (G1, G2, G3, G4, G5, G7, G9, G14, G19, and G20) that seemed to be interspersed throughout the clusters, except for Agr5_{16S rRNA} that was composed mainly of strains classified as A. larrymoorei. Noteworthy, the so-called Clade 3 of the Agrobacterium genus containing A. divergens strains took a basal and distinctive position among the different Agrobacterium clades delineated. Additionally, rhizogenic agrobacteria strains were observed in both the R. rhizogenes cluster as well as eight Agrobacterium (genomo)species including G2, G3, G7, G9, G19, G20, Agrobacterium cucumeris, and Agrobacterium leguminum.

Next, a phylogenetic tree was constructed based on a 1,092-bp alignment of the recA gene (S2 Fig). The best substitution model for the recA tree was Transition 3 model with empirical base frequencies and invariable site plus FreeRate with five categories model (TIM3+F+I+R5), based on AIC and BIC scores of 71080.772 and 71160.704. The recA-based phylogenetic tree presented two strongly supported clades (>80% bootstrap values, 91.5% similarity): (i) a clade containing most Rhizobium strains and (ii) the polyphyletic clade containing the All. vitis and Agrobacterium strains. Similar to the 16S rRNA tree, the Rhizobium clade could be divided into two subclades (R1 and R2), with subclade R2, mainly consisting of rhizogenic Rhizobium strains, including the type strain R. rhizogenes NBRC 13257^T (= LMG150^T) and R. radiobacter K84^T. The Allorhizobium/Agrobacterium main clade could be further sub-divided into the All. vitis subclade (including the type strain All. vitis NCPPB 3554^T) and the Agrobacterium subclade. The Agrobacterium subclade presented three clusters, with the so-called Agrobacterium Clade 3 containing A. divergens strains in a basal position, then the A. rubi-larrymoorei cluster with nine groups, and the Atsc cluster with 17 groups (≥84% similarity; >80% bootstrap values). The Atsc subclade was mainly composed of monophyletic groups consisting either of distinct genomospecies (based on the recA phylogeny reported by Costechareyre et al. [30]). However, strains identified as “A. deltaense" and A. salinitolerans were dispersed both into two polyphyletic groups, Atsc7_recA and Atsc9_recA, with Atsc9_recA containing the type strain “A. deltaense" YIC 4121^T, and Atsc14-15_recA, with Atsc14_recA containing the type strain A. salinitolerans YIC 5082^T, respectively. Similarly, A. pusense strains were split in two polyphyletic groups (Atsc16-17_recA), with Atsc17_recA including the type strain LMG 25623^T (= NRCPB10^T) and Atsc16_recA being composed of strains isolated from Pigeon Pea roots in India [68]. Additionally, Atsc11_recA and Atsc8_recA contained strains with no previously assigned genomospecies affiliation nor any species names validly published or pending for approval nomenclature. Overall, the genomospecies delineation presented by the recA phylogeny indicated that this marker does not provide sufficient resolution to unequivocally classify all Agrobacterium strains in dedicated monophyletic genomospecies. Rhizogenic agrobacteria strains were observed in the same Agrobacterium (genomo)species as in the 16S rRNA phylogeny.

Phylogenetic analysis using multiple marker genes using MLSA- and WGS-based trees

To overcome possible bias resulting from the use of single marker genes, an MLSA was conducted based on a concatemer sequence (10,683 bp in size) of five house-keeping genes (Fig 1). The best substitution models for the MLSA tree were Transition 3 model with empirical base frequencies and invariable site plus FreeRate with four categories model (TIM3+F+I+R) for 16S rRNA, Transition model 3 with empirical base frequencies and invariable sites plus FreeRate with six categories model (TIM3+F+I+R6) for atpD and rpoB, general time reversible model with unequal rates and unequal base frequencies and empirical base frequencies with invariable sites plus FreeRate with five categories model (GTR+F+I+R5) for gyrB, and Transition model 3 with unequal and empirical base frequencies with invariable sites plus FreeRate with five categories model (TIM3+F+I+R5) for recA. These models were selected based on BIC scores of 19104.979, 291897.533, 197574.819, and 69398.569 and AIC scores of 19030.454, 291778.075, 197470.409, and 69318.637 for 16S rRNA, atpD and rpoB, gyrB, and recA, respectively.

[Figure omitted. See PDF.]

A ML phylogeny of 583 strains including 413 Agrobacterium, 8 Allorhizobium, and 159 Rhizobium strains is shown. The phylogeny was built from the concatenated alignment (10,683 bp) of five chromosomal genes (16S rRNA, atpD, gyrB, recA, and rpoB) using IQTREE. Bootstrap support value ranging from 70% to 100% based on 1,000 repetitions is represented by thicker branches. Leaves have been highlighted according to genus affiliation. Leaves have also been annotated in the outer rings according to (from inside to outside) original genomospecies affiliation, MLSA subclades delineation, and ANIb subgroups (using a 94.0% identity threshold). Type strains are highlighted in light-blue background and bold font (see detailed annotation in online version https://itol.embl.de/shared/pme_bim). The scale bar represents the number of expected substitutions per site under the best-fitting model for each gene (TIM3+F+I+R for 16S rRNA, TIM3+F+I+R6 for atpD and rpoB, GTR+F+I+R5 for gyrB, and TIM3+F+I+R5 for recA; see details in Materials and Methods). Strains that did not cluster with any other are annotated as STN for singleton. E. adhaerens strains were used as outgroup.

In general, the MLSA phylogeny showed a similar structure as the recA tree, and presented two strongly supported clades (>80% bootstrap values, ≥91% similarity): (i) the Rhizobium clade, which could be subdivided in a subclade R2 containing closely related (mostly R. rhizogenes) strains and another subclade, consisting of more distantly related Rhizobium strains; and (ii) the clade containing all the All. Vitis and Agrobacterium strains. Similar to the recA phylogeny, the All. vitis/Agrobacterium clade could be further subdivided into the All. Vitis subclade (including the type strain NCPPB 3354^T), the Agrobacterium subclade, and a set of closely related recently described genera including Heterorhizobium, Paenirhizobium, Peteryoungia, Affinirhizobium, Neorhizobium, and Alirhizobium [69]. Additionally, rhizogenic agrobacteria strains were observed in the same clusters as in both the 16S rRNA and recA phylogenies.

When looking deeper into the Agrobacterium clade, eight clusters were part of the A. rubi-larrymoorei subclade (Agr1-8_MLSA) and contained bona fide species, while 14 clusters were grouped together in the Atsc subclade and included mainly known GSs, bona fide validly published species names or species names awaiting for nomenclature. Similarly, strains identified as A. pusense, A. salinitolerans, and "A. deltaense" also clustered monophyletically. Noteworthy, strains classified as A. leguminum clustered together with "A. deltaense" strains. Additionally, clades Atsc1 and Atsc4 contained strains with no previous GS affiliation clustered apart. Noteworthy, the Agrobacterium Clade 3 containing A. divergens strains was not included in this phylogeny due to insufficient homology with the selected query genes.

To resolve the ambiguities between the recA and MLSA trees, a WGS phylogeny was constructed. The phylogeny reconstruction was based on a concatemer of 852,736 bp that presented 851,130 informative sites. The likelihood value of the final tree obtained was –74728904.543924. The WGS phylogeny presented three strongly supported clades (>80% bootstrap values, 92% similarity): (i) the Rhizobium clade; (ii) the Allorhizobium clade; and (iii) the Agrobacterium clade. Similar to the MLSA phylogeny, Heterorhizobium, Paenirhizobium, and Peteryoungia clustered closely with Allorhizobium. The WGS phylogeny showed an easily distinguishable pattern in the Agrobacterium clade and displayed nine and 15 groups in the A. rubi-larrymoorei and the Atsc subclades, respectively, as well as the Agrobacterium Clade 3 composed of A. divergens strains in a basal position (Fig 2). Most of the defined subclades coincided with genomospecies assignment, validly published species or species waiting for name validation. Similar to the recA and MLSA phylogenies, strains identified as "A. deltaense" and A. leguminum clustered together in Atsc14_WGS. Noteworthy groups Agr9_WGS, Atsc7_WGS, Atsc10_WGS, and Atsc12_WGS contained strains with no previous GS assignation. Moreover, rhizogenic agrobacteria species were observed in the same Agrobacterium (genomo)species as in 16S rRNA, recA, and MLSA phylogenies.

[Figure omitted. See PDF.]

The tree was obtained using the Maximum likelihood (ML) method. The phylogeny was reconstructed initially with FastTRee and then enhanced with RAxML applying the ML method and a CAT General Time Reversible (GTRCAT) approximation. The phylogenetic analysis was based on a concatenated alignment of the most discriminative amino acid positions of 1,524 markers conserved among all 583 sequenced genomes (413 Agrobacterium, 8 Allorhizobium, and 159 Rhizobium and three E. adhaerens strains used as outgroup) identified with Phylophlan 2. Bootstrap support values ranging from 70% to 100% based on 1,000 repetitions is represented by thicker branches. Leaves have been highlighted according to genus affiliation. Type strains are highlighted in a light-blue background and bold font (see detailed annotation in online version https://itol.embl.de/shared/pme_bim). Leaves have also been annotated in the outer rings according to (from inside to outside) genomospecies affiliation, WGS subclades delineation, and ANIb subgroups. The scale bar represents the number of expected substitutions per site.

Assessment of phylogenetic signals based on GC-content, presence/absence of Agrobacterium genes, and codon usage bias analysis

The phylogenetic signal analysis involving GC content of all the Agrobacterium strains from the MLSA phylogeny yielded statistically significant results in all four metrics, more specifically Blomberg’s K = 2.6·10⁻⁶ (p-value = 0.001), Pagel’s Lambda = 0.976 (p-value = 2.73·10⁻²⁰²), Moran’s I = 0.81 (p-value ≤ 0.001) and Abouheif’s C_mean = 0.849 (p-value ≤ 0.001), meaning that the distribution of GC values across the MLSA phylogeny was not random but, instead, depended on the phylogenetic relatedness between strains (i.e., close relatives displayed more similar values among themselves than with unrelated strains).

Finally, it was also assessed whether the codon usage of Agrobacterium strains could be correlated to genomospecies classification (S4 Fig). Inspection of the results revealed a wide variation of the PCA scores within most ANIb groups, which resulted in the presence of a number of outliers and precluded a clear separation of most groups. Therefore, it was concluded that codon usage patterns might not be a good criterion for taxonomic delimitation.

Use of overall genome relatedness indices for species delineation in the Agrobacterium tumefaciens species complex

In order to better determine the species boundaries within the Agrobacterium genus, an ANIb was conducted as a complementary method since it is considered the golden standard for prokaryotic species definition [41]. The strains examined in this study yielded 162,409 pairwise comparisons showed pairwise distance values ranging from 68.06%% to 99.99% (S5 Table). At first glance and in contrast with the recA, MLSA, and WGS phylogenies, the ANIb-based hierarchical tree suggested that All. vitis and Rhizobium strains seemed to be more closely related to each other and more distant from Agrobacterium (S4 Fig). Remarkably, Agrobacterium species composing the A. rubi-larrymoorei clade of the three phylogenies (recA, MLSA, WGS) reconstructed, clustered within the Rhizobium clade in the ANIb-based tree.

Several studies have indicated that an ANIb cut-off value of 94–96% similarity should be applied to determine a species threshold among prokaryotes [39, 58, 70, 71]. To establish a more accurate cut-off value for Agrobacterium species delineation, we analysed the ANIb-values of all strains and observed a clear gap in the distribution of ANIb values between 93.7% and 94.2% (S3 Fig). For this reason, we have set the cut-off to delineate distinct subclades at an ANIb of 94.0%.

Using the optimized ANIb cut-off for species delineation, the Agrobacterium clade presented 25 groups (AG1-25) and 7 singletons, with pairwise distance values ranging from 76.58% to 99.99%. Annotating the ANIb results on the recA, MLSA, and WGS phylogenies indicated that MLSA and WGS analysis resulted in a more accurate delineation in species compared to recA phylogeny. In the WGS phylogeny, ANIb annotation helped to better delineate clades, mainly merge and form monophyletic groups. For example, ANIb allowed: (i) merging Agr3_WGS strains with neighbouring singletons and form a single group of A. vaccinii strains; (ii) merging all neighbouring A. rubi strains into a single group Agr5_WGS; (iii) merging all neighbouring A. pusense strains into a single group Atsc11_WGS; (iv) merging all neighbouring A. salinitoelrans strains into a single group Atsc13_WGS; and (v) merging all neighbouring A. radiobacter strains into a single group Atsc15_WGS. Strains in the All. vitis clade showed an identity range of 77.25%-99.99% and revealed three subclades and three singletons at a 94% identity threshold (Fig 3). The groups as defined based on ANIb analysis corresponded perfectly with the MLSA and WGS phylogeny.

[Figure omitted. See PDF.]

Pairwise ANIb values are shown of 412 Agrobacterium, 8 Allorhizobium, and 159 Rhizobium genomes. A total of 25 ANIb groups (AG1-25) were observed in the Agrobacterium genus. In addition 7 strains formed singletons, i.e. did not show an average nucleotide identity above 94% with any other strain. Pairwise comparison values ≥ 94.0% are highlighted in dark blue(corresponding to species level) and those between 80.0% and 94.0% are highlighted in light blue (corresponding to genus delineation). The hierarchical tree, located on the left side of the figure, was constructed based on the pairwise comparison values using a “complete” clustering algorithm. The hierarchical tree has been annotated with ANIb groups based on pairwise comparison values of ≥ 94.0% among strains.

In addition, pairwise AAI values were calculated to evaluate the capacity of this OGRI to delineate species in the Atsc. AAI analysis assessed the homology (evaluating both sequence similarity and alignment length) of shared CDSs. The strains examined in this study showed pairwise AAI distance values ranging from 67.3% to 99.99% (S6 Table). Applying a cut-off of 80% similarity for genus delineation as advised by Liu et al. [72], 13 distinct clusters could be identified: (i) a homogenous cluster containing all All. vitis strains; (ii) a cluster containing strains of the A. rubi-larrymoorei clade of the three phylogenies reconstructed; (iii) three remaining Agrobacterium clusters, one of them composed of A. divergens strains; and (iv) nine Rhizobium clusters, one of them being composed exclusively of strains identified as R. rhizogenes. By applying a threshold identity value of 95% to delineate species, as recommended by Fan et al. [62], a total of 27 clusters (9 in the Rhizobium clade, 2 in the Allorhizobium clade, and 16 in the Atsc clade) and 55 singletons were identified (S6 Table). Remarkably, one singleton identified as Agrobacterium sp. a22-2 and two clusters with strains identified as Agrobacterium (Agrobacterium sp. MA01 and RAC06, and A. albertimagni AOL15^T and Agrobacterium sp. SCN 61–19) were observed closely related to other Rhizobium singletons.

Discussion

The main objective of this study was to further elucidate and delineate (genomo)species in the Agrobacterium genus and to have better insight into the taxonomic positioning of rhizogenic strains, i.e. strains that induce HRD. To achieve this, a large set of more than 400 publicly available Agrobacterium genomes originally identified as members of genus Agrobacterium were analysed using several analysis tools that are frequently used to infer phylogenetic relationships, which also allowed evaluating their usefulness for species delineation within Agrobacterium. Because we were particularly interested in the taxonomic positioning of rhizogenic strains (i.e., strains causing HRD), we also included a set of 63 R. rhizogenes strains and 24 in-house sequenced rhizogenic Agrobacterium strains (previously classified as A. rhizogenes). In order to position rhizogenic Agrobacterium and Rhizobium strains in a broader phylogenetic context, we also included 18 All. vitis (former A. vitis) strains, and 96 Rhizobium type strains in this study. The phylogenetic relationship among Agrobacterium, Allorhizobium, and Rhizobium species was evaluated by generating four phylogenetic trees: two based on single marker genes (16S rRNA or recA gene), one comprising concatemers of the 16S rRNA gene and four housekeeping genes (rpoB, recA, gyrB, and atpD), and one using WGS data. Additionally, the pairwise ANI and AAI were calculated for all strains, and among Agrobacterium strains it was assessed whether presence/absence of genes, codon usage, or GC-percentage contain a phylogenetic signal. The results obtained have confirmed previously reported findings, but also provided new insights in the positioning of rhizogenic agrobacteria within the taxonomy of the Atsc. This study also allowed to evaluate the accuracy of different analyses to infer taxonomy within the genus Agrobacterium.

Single-marker phylogeny does not provide sufficient taxonomic resolution in the Agrobacterium genus

The sequence of the 16S rRNA gene was used to infer taxonomic classification since it is a well-known and valuable marker for bacterial taxonomy [73]. This marker is commonly included in polyphasic taxonomy approaches for bacteria due to the availability of universal primers and inherent high copy number in bacterial genomes [74]. In our study, the 16S rRNA-based phylogeny proved useful to distinguish strains at the genus level, and presented three large clades corresponding to Rhizobium, Agrobacterium, and Allorhizobium. Interestingly, a few strains identified as Rhizobium did not cluster in any of these large clades, indicating that these strains do not belong to genus Rhizobium, refuting their original classification. These strains have been recently moved into new genera including, (i) R. subbaraonis JC85^T and R. rhizolycopersici DBTS2^T to Mycoplana; (ii) R. album NS-104^T to Metarhizobium (also considered at basal position in Rhizobium phylogeny); (iii) R. wenxinae DSM 100734^T and CGMMCC 1.15279 as well as Agrobacterium sp. Ap1 to Alirhizobium; and (iv) All. pseudoryzae DSM 19479^T and R. rhizoryzae DSM 29514^T to Affinirhizobium. Nevertheless, it is a well-known fact that 16S rRNA does not provide sufficient resolution for species delineation in these complex genera [25, 75, 76].

In general, protein-encoding genes are considered more informative for taxonomic classification [25, 77]. Aiming for a clearer delineation of Agrobacterium (genomo)species, a phylogenetic analysis was conducted using the housekeeping gene recA. This gene encodes a DNA recombinase and was initially used for assigning strains in the Atsc to a specific genomospecies [25, 30]. Our recA phylogeny was mainly congruent with that presented by Costechareyre et al. [30], who defined thirteen distinct clades, nine of which were identified as Agrobacterium genomospecies G1-9 and four that were identified as A. larrymoorei, A. rubi, A. rosae, and A. vitis (later renamed to All. vitis). Even though the use of recA was initially used to differentiate between genomospecies, our results showed that recA was not always sufficient to make a clear delineation between genomospecies. For instance, in our study the recA-based phylogenetic tree did not generate monophyletic groups for some (genomo)species including A. vaccinii, A. rubi (G13), A. salinitolerans (G9), A. pusense (G2) and "A. deltaense" (G7). For that reason, we conducted an (in silico) MLSA that concatenated the 16S rRNA gene and four housekeeping genes (rpoB, recA, gyrB, and atpD). The application of multiple genes to resolve taxonomic issues has been previously documented to overcome the conflicting branching patterns of individual genes and has even been recommended for the genus Agrobacterium [34]. Although the MLSA tree was mainly congruent with the recA tree, it was clear that MLSA provided a higher taxonomic resolution. For instance, the MLSA tree clustered together all strains identified as A. pusense (G2) and those identified as “A. deltaense” (G7), unlike the recA phylogeny in which they were split in two clusters. Interestingly, strains identified as A. leguminum (including type strain A. leguminum MOPV5^T) were part of this G7-containing MLSA cluster.

Considering that the MLSA tree topology as well as genomospecies delineation was largely supported by both the WGS tree and ANIb analysis, it can be concluded that an MLSA-based tree outperforms phylogenetic trees based on either 16S rRNA or the recA genes, which are clearly not sufficient for accurate taxonomic positioning of strains within the Agrobacterium genus.

A combination of phylogenetic analyses provides support to “new candidate genomospecies” and allows identification of misclassified strains

The MLSA tree identified strains and subclades that clustered separately from previously reported genomospecies, suggesting the identification of new and previously unreported genomospecies. We provided support for this hypothesis by using a combination of ANIb analysis and WGS phylogeny. Our results indicated that, in the Agrobacterium clade of the WGS tree, strains Agrobacterium sp. DE0009, A. tumefaciens UBA5877, A. tumefaciens UBA6714, A. tumefaciens 6N2, and A. tumefaciens KCJK1736 were clustered in subclade Atsc10_WGS. This clustering was supported by ANIb, since all the strains in question were part of the ANI-group AG8 at the cut-off (94.0%) defined in this study. For these reasons, we suggest renaming it as new candidate genomospecies G21.

Similarly, strains in both Atsc4_MLSA and Atsc12_WGS group also clustered together in the same ANIb AG3 group, indicating that these strains belong to the same species. These strains were isolated from hospital intensive care unit sinks in the United States of America in 2023. As a result, we suggest considering these strains as new candidate genomospecies G22.

Moreover, a cluster composed exclusively of MAGs strains was observed in the A. rubi-larrymoorei. The ANIb results also pointed out that these strains are part of the same AG16 group. Based on these findings, we suggest considering these strains as new candidate genomospecies G23, supported by Minute 10 of the 2022 meeting of the Subcommittee on the Taxonomy of Rhizobia and Agrobacteria, wherein the SeqCode was endorsed [78]). Additionally, strains A. tumefaciens BIN17 and Agrobacterium sp. SOY 23 clustered apart in the recA (Atsc11_recA) and WGS (Atsc7_WGS) phylogenies as well as in the ANIb AG13 group. These results suggests that these strains could be also part of a new candidate genomospecies G24. So far, it seems Agrobacterium BIN17 and SOY23 are non-pathogenic species since they were isolated from a lignocellulose-decomposing bacterial consortium from soil associated with dry sugarcane straw and soybean rhizosphere soil [79]. Additionally, strain ST15.13.057 is hypothesized to be a new genomospecies since it did not cluster with any other strain in any of the phylogenies reconstructed nor in the ANIb analysis. This strain was isolated from an HRD-infested tomato greenhouse. Interestingly, compared to several other rhizogenic Agrobacterium strains, ST15.13.057 seems to be a highly aggressive strain resulting in severe symptoms of excessive root growth (Kim et al., unpublished data). It is also characterized by a comparatively large genome size (S1 Table). Nevertheless, more related strains are needed in order to propose ST15.13.057 as a new genomospecies according to the rules provided by the International Committee on Systematics of Prokaryotes (ICSP) Subcommittee on the Taxonomy of Rhizobia and Agrobacteria [80].

Although several complementary analyses (including MLSA- and WGS-based phylogeny, and ANIb analysis, confirm our hypothesis that strains mentioned above are members of new genomospecies, additional analyses (including fatty acid profiles, cell wall composition, and exopolysaccharides, as well as morphological, biochemical, and enzymological characterization) are required to confirm that there is sufficient difference among the strains of these clades to formally describe them as new species within the genus Agrobacterium [81].

Finally, our analysis also provides support for recently described new species. Indeed, the MLSA- and WGS-based phylogeny showed that strains previously assigned to G3 were clearly divided in two separate subclades. While strains of subclade Atsc8_MLSA or Atsc6_WGS were assigned to A. tomkonis (G3), strains of the other subclade were reclassified into A. cucumeris [82]. This was also supported by the ANIb analysis that showed <94.0% identity among strains of both subclades.

In addition to identifying candidate new genomospecies, MLSA analysis enabled a more accurate classification into genomospecies than recA analysis. For instance, while recA phylogeny positioned some outside the A. pusense (G2) clade (Atsc16_recA), the MLSA phylogeny positioned these strains within the G2-containing subclade (Atsc2_MLSA). Classification of these strains as G2 was also supported by ANIb analysis and WGS phylogeny. A similar situation was observed with “A. deltaense” (G7) strains that were separated into two clusters in the recA phylogeny (Atsc7_recA and Atsc9_recA) but were merged in the MLSA phylogeny (Atsc5_MLSA).

Within the Agrobacterium clade, a set of strains seemed to be misclassified. Such was the case for two A. larrymoorei strains, SORGH AS 1126 and SORGH AS974, that actually clustered apart in both the MLSA and WGS phylogenies and showed <94% nucleotide identity (namely 78.3–78.6%) with other A. larrymoorei strains (including the type strain). Likewise, strains previously identified as A. leguminum, including the type strain A. leguminum MOPV5^T, clustered together with “A. deltaense” (G7) strains in both the MLSA and the WGS phylogenies. This suggests that A. leguminum and A. deltaense strains both belong to the same species (previously assigned as G7), which is also supported with ANIb values >95% between A. deltaense and A. leguminum strains. This apparent synonymy has also been previously suggested by Naranjo and colleagues [29]. Similarly, other strain identifiers would need to be adjusted, including R. rhizogenes strains B 4.1 and SBV 302 82 into A. tumefaciens B 4.1 and A. burrii SBV 302 78 2, and R. oryzihabitans M15 should be A. oryzihabitans (G13) [69]. Likewise, all ST in-house strains need to be reclassified according to their respective Agrobacterium cluster. Also, outside the Agrobacterium clade, our study identified strains that seemed to be misclassified. For instance, A. albertimagni AOL51^T, which clustered closely with the Agrobacterium clade in the 16S rRNA-based phylogeny (98.4%), clustered apart from the Agrobacterium clade in the recA (73% similarity) and MLSA-based (76% similarity) phylogenies. In a previous study, the name Rhizobium aggregatum complex was proposed for the clade containing strain A. albertimagni AOL51^T [34]. However, in our study both the MLSA and WGS phylogeny and ANIb analysis showed that A. albertimagni AOL51^T is more closely related to Allorhizobium, suggesting that “R. aggregatum complex” is not the best choice. This is in agreement with Kuzmanović et al. [83], who proposed the name “Peteryoungia” for strains formerly identified as “R. aggregatum complex”, and also included A. albertimagni AOL51^T.

In the Rhizobium clade, R. rhizogenes TPD 7009 should be reclassified as Agrobacterium sp., R. rhizogenes Y79 and FIT62 should be reclassified as Rhizobium sp. and Agrobacterium strains 13–626, SHOUNA 12C, BETTINA 12B, and ICMP 7243 in the Rhizobium rhizogenes group should be reclassified as R. rhizogenes. Additionally, a set of strains identified as Rhizobium, Allorhizobium, or Agrobacterium clustered in three groups next to the Agrobacterium clade. These results are in agreement with [69], who suggested the transfer of these strains to the genera Affinirhizobium, Alirhizobium, and Neorhziobium. Similarly, strains identified as Rhizobium or Agrobacterium clustered also apart in three groups already suggested by Ma et al. [69] as Peteryoungia, Paenirhizobium, and Heterorhizobium.

Rhizogenic strains are dispersed over at least two genera and at least nine (genomo)species

Another objective in this study was to shed more insight into the taxonomy of strains causing HRD. Interestingly, while HRD was originally associated with pathogenic R. rhizogenes (former biovar 2) strains, in recent studies it was observed that on (hydroponically grown) Cucurbitaceae and Solanaceae plants the causative agent is generally Agrobacterium biovar 1 harbouring a root-inducing plasmid [5, 6, 44, 84].

In previous research, Bosmans et al. [44] demonstrated that rhizogenic agrobacteria isolated from infested hydroponic greenhouses showed a remarkable phenotypic and genetic diversity. This genetic diversity is also seen here, since the rhizogenic strains included in this study are scattered over at least seven Agrobacterium (genomo)species (G2, G3, G7, G9, G19, G20, A. cucumeris, and A. leguminum) [44, 85]. Strain ST15.13.057, which was clearly separated from the other genomospecies, could be the first strain to be assigned to new genomospecies G25. This particular strain was isolated from tomato plants grown in a hydroponic greenhouse located in Belgium in 2014 [44], and showed high symptom severity compared to most other rhizogenic Agrobacterium strains [Kim et al. in prep]. This strain clustered closest to G8 strains, but only with 92.4–92.9% identity, supporting the idea that this strain could be a member of a new genomospecies. Nevertheless, more strains are needed to define the limits of this candidate new genomospecies and more biochemical analysis are required to confirm it as a new genomospecies [80]. Additionally, a G1 strain (ST15.13/045) and a G4 strain (KACC 21759) for which no genome sequences were available at the moment, were also reported previously to induce HRD [44, 85]. Although rhizogenic Agrobacterium strains are assigned to several genomospecies, most isolated Agrobacterium strains inducing HRD belong to G9 and to the bona fide species A. cucumeris. Nevertheless, so far no rhizogenic strains classified as G6 nor G13 were observed in our study, not ruling out that future isolations may detect rhizogenic strains in these genomospecies as well.

In contrast to rhizogenic Agrobacterium strains that were scattered over diverse genomospecies, the rhizogenic Rhizobium strains (previous A. rhizogenes) clustered closely together, including the type strain R. rhizogenes NBRC 13257^T (= LMG150^T). Interestingly, the R. rhizogenes subclade also contained the avirulent Rhizobium radiobacter K84 strain (former A. radiobacter K84), which is known for its antagonistic activity towards A. tumefaciens causing crown gall disease in field crops [86]. All these results highlight the fact that rhizogenic agrobacteria are polyphyletic, being located in at least two genera in the Rhizobiaceae family.

Finally, the rhizogenic strain R. rhizogenes YR147 and OV677, clustered as singletons closely to the R. rhizogenes subclade in all the phylogenies reconstructed and shared 90–91% identity according to ANIb with other R. rhizogenes strains, suggesting these strains needs a different specific epithet.

Combination of MLSA and ANIb analysis is the best practice for robust classification into genomospecies in the Agrobacterium tumefaciens species complex

Compared to the recA tree, the MLSA and WGS phylogenies reflected better the genomospecies delineation provided by the ANIb groups, suggesting that the recA phylogeny, which has been initially used for genomospecies classification is not robust enough. The use of ANI has been suggested as a tool for resolving ambiguous phylogenies, which has become possible thanks to the enormous evolution in sequencing technologies [87]. ANIb has already been successfully used for accurate species delineation in complex genera, such as Pseudomonas, Arcobacter, Stenotrophomonas, and Burkholderia [88–91]. As a result, we applied ANIb in this study in order to shed more insight in the taxonomy of Agrobacterium, aiming to conciliate or resolve contradictory results in the different phylogenetic trees. For species delineation, the cut-off value of 94–96% for ANI has been recommended for prokaryotic organisms [92, 93]. More specifically, López-Guerrero et al. [58] recommended 95% similarity as a threshold for the delineation of species in the Rhizobiaceae family. As a result, a number of studies have applied the same cut-off value for the delineation of species in the Rhizobiaceae family [32, 77, 94, 95]. However, using a considerably large set of strains in this study (>400), we were able to fine-tune the cut-off to 94.0%. This allowed the identification of 24 distinct (genomo)species within the Agrobacterium genus. ANIb analysis also enabled delineation of three distinct species that clustered together in the MLSA- and WGS-based phylogenetic tree, namely A. rubi, A. bohemicum, and A. rosae, showing the usefulness of using this approach in a complementary manner with phylogenetic analysis based on MLSA or WGS. However, the ANIb analysis cannot be used as a stand-alone tool to infer phylogenetic relationships. For instance, the ANIb-based hierarchical tree, showed that, in contrast to MLSA and WGS phylogeny, that the A. rubi-larrymoorei clade clustered within the Rhizobium clade. This different clustering can be partially explained by the fact that pairwise distances values resulting from the ANIb analysis were clustered based on the specific algorithm known as “complete linkage” [96]. Moreover, ANI analysis does not apply an evolutionary model like phylogenetic analysis, and it is not considered an appropriate method for estimating genome relatedness among more divergent genomes, i.e., effective delineation of relationships at genus and family levels, for which AAI or phylogenomic treeing approaches are considered more appropriate [97].

In addition to calculating pairwise ANI distances, AAI distances were also calculated since it has been suggested that a 95% identity threshold could be applied for prokaryotic species delineation [98]. The AAI analysis did not provide a high-resolution species level delineation as compared to ANIb. Indeed, some of the separate ANIb groups were merged in the same AAI group when a cut-off of 95% was applied for AAI analysis, indicating that AAI was not able to delineate closely related (genomo)species within the Atsc. This is in agreement with the recommendation of using AAI comparisons for more distant relationships, e.g., for genus delineation [99].

Additionally, it was shown that GC content was strongly associated with the phylogenetic relationship present among the strains as observed in the MLSA tree. In other words, the more phylogenetically close Agrobacterium strains are, the more similar GC content values they display, again confirming the robustness of the MLSA analysis.

Altogether, we showed that using five genes in the MLSA, allowed a secure phylogenetic placement of AtSC strains. Nevertheless, in some cases the MLSA phylogeny could be improved by the use of ANIb.

The application of WGS-based phylogeny or phylogenomics is expected to yield more accurate phylogenies in contrast to other approaches using one or a limited set of genes [99]. This is not surprising since both methods use whole genomes, with more informative power and providing a better resolution for taxa delineation. Nevertheless, in our study, the subclades topology observed using WGS was largely similar to that observed in the MLSA phylogeny. The major downside of WGS phylogeny, however, seems to be the computing time that is needed. On desktop computer with 256 GB of storage memory, 16 GB of RAM memory, and 8 CPUs available, the WGS phylogeny analysis (i.e., the ML phylogeny reconstruction) for 200strains can take approximately 15 days, with the bootstrap value calculation taking a similar time, for a total of 30 days of continuous analysis. However, if computing power is moderate (e.g.: 4 GB RAM and 120 GB storage capacity), a combination of (in silico) MLSA and ANIb would be a fast (approx. four hours) and cost-effective method to unequivocally classify strains into genomospecies in the AtSC. This suggestion is in line with that of Kämpfer and Glaeser [100], who claim that MLSA represents an intermediate analysis between single marker and genome wide-based approaches to resolve phylogenetic resolution at the species level.

Conclusion

Based on the comparison of different tools to infer phylogenetic relationships, we propose conducting an MLSA in combination with ANIb analysis for accurate taxonomic resolution among AtSC strains, which requires only limited computational power. This approach allowed us to reclassify several strains in the Atsc cluster with high confidence. In addition, it allowed us to identify new candidate genomospecies. And finally, we obtained a comprehensive view of rhizogenic agrobacteria and showed that these strains are found in at least eight different (genomo)species and two genera.

Supporting information

S1 Fig. Phylogenetic tree based on the 16S rRNA universal marker gene of Agrobacterium, Allorhizobium, and Rhizobium.

A maximum likelihood phylogeny using as input a collection of 413 Agrobacterium, 8 Allorhizobium, and 159 Rhizobium strains is shown. The phylogeny was built from the alignment the 16S rRNA gene (1,495 bp) using IQTREE. Bootstrap support value >70% based on 1,000 repetitions is represented by branches in bold. Leaves have been highlighted according to the defined 16S rRNA subclades. The colors in the outer rings indicate the 16S subclade or the genomospecies assigned to the Agrobacterium strains. Type strains are indicated in bold (see detailed annotation in online version https://itol.embl.de/shared/pme_bim). The scale bar represents the number of expected substitutions per site under the best-fitting model for the marker gene used (TN+F+I+G4 for 16S rRNA). Three Ensifer adhaerens strains (including type strain E. adhaerens Casida A^T) were used as outgroup.

https://doi.org/10.1371/journal.pone.0302954.s001

(EPS)

S2 Fig. Phylogenetic tree based on the recA marker gene of Agrobacterium, Allorhizobium, and Rhizobium.

A maximum likelihood phylogeny using as input a set of 413 Agrobacterium, 8 Allorhizobium, and 159 Rhizobium strains is shown. The phylogeny was built from the alignment the recA gene (1,089 bp) using IQTREE. Bootstrap support value >70% based on 1,000 repetitions is represented by branches in bold. Leaves have been highlighted according to genus affiliation. The outer rings show the ANIb subgroups assigned to the strais (STN: singleton), the defined MLSA and recA subclades as well as the genomospecies assigned to the Agrobacterium strains. Leaves highlighted in bold represent type strains (see detailed annotation in online version https://itol.embl.de/shared/pme_bim). The scale bar represents the number of expected substitutions per site under the best-fitting model for the marker gene used (TIM3+F+I+G4 for recA). Strain that did not cluster with any other are named as STN for singleton. Three Ensifer adhaerens strains (including type strain E. adhaerens Casida A^T) were used as outgroup.

https://doi.org/10.1371/journal.pone.0302954.s002

(EPS)

S3 Fig. Histogram of pairwise average nucleotide identity using BLAST (ANIb) values.

Pairwise ANIb values are shown for 413 Agrobacterium, 8 Allorhizobium, and 159 Rhizobium genomes. The pyani 0.2.10 package was used to calculate all pairwise ANIb values. The dashed red line represents this study’s optimized ANIb threshold for the delineation of Agrobacterium species at 94.0%.

https://doi.org/10.1371/journal.pone.0302954.s003

(EPS)

S4 Fig. Principal component analysis (PCA) of normalized relative synonymous codon usage (NRSCU) obtained for different Atsc genomospecies.

(A) Scores plots showing the distribution of the Atsc genomes analyzed in this study according to the first two components (which explained 79.7% of the total variance) obtained by principal component analysis (PCA) of the normalized relative synonymous codon usage (NRSCU) values. Colors of points and the 68% data concentration ellipses denote different ANIb groups. (B) PCA loading plot, where each dot represents the loadings on the first two principal components for one factor (NRSCU for a particular codon).

https://doi.org/10.1371/journal.pone.0302954.s004

(TIFF)

S1 Table. List of species originally classified as Agrobacterium, of which some were re-classified later (adapted from Flores-Félix et al., 2020).

https://doi.org/10.1371/journal.pone.0302954.s005

(XLSX)

S2 Table. List of genome assemblies downloaded from the NCBI database and sequenced in this study.

https://doi.org/10.1371/journal.pone.0302954.s006

(XLSX)

S3 Table. List of markers used in this study and retrieved from the GCF_001649535 assembly for strain K599, also known as NCPPB2659.

https://doi.org/10.1371/journal.pone.0302954.s007

(XLSX)

S4 Table. GC content of selected strains from the Agrobacterium tumefaciens species complex.

https://doi.org/10.1371/journal.pone.0302954.s008

(XLSX)

S5 Table. Matrix presenting the pairwise values resulting from average nucleotide identity analysis using BLAST (ANIb).

The analysis was conducted for the 580 Agrobacterium, Allorhizobium, and Rhizobium quality-checked assemblies in this study. STN represents singletons and highlighted in orange, while 22 in-house sequenced rhizogenic Agrobacterium strains are highlighted in green.

https://doi.org/10.1371/journal.pone.0302954.s009

(XLSX)

S6 Table. Matrix presenting the pairwise values resulting from average amino acid identity (AAI) analysis.

The analysis was conducted among the 580 Agrobacterium, Allorhizobium, and Rhizobium quality-checked assemblies in this study.

https://doi.org/10.1371/journal.pone.0302954.s010

(XLSX)

Acknowledgments

We would like to thank Professor Sam Crauwels for sharing ideas, discussing, and giving advice on genome assembly methods, and Mrs. Alexa de Knijf for assistance with figures.

References

1. 1. Flores-Félix JD, Menéndez E, Peix A, García-Fraile P, Velázquez E. History and current taxonomic status of genus Agrobacterium. Systematic and Applied Microbiology. 2020;43:126046.

* View Article

* Google Scholar

2. 2. Young JM, Kerr A, Sawada H. Agrobacterium. Bergey’s Manual of Systematics of Archaea and Bacteria. 2015;1–15.

* View Article

* Google Scholar

3. 3. Dessaux Y, Faure D. Niche Construction and Exploitation by Agrobacterium: How to Survive and Face Competition in Soil and Plant Habitats. In: Gelvin S, editor. Agrobacterium Biology, from basic science to biotechnology. Springer; 2018. p. 55–86.

4. 4. Chihaoui S-A, Trabelsi D, Jdey A, Mhadhbi H, Mhamdi R. Inoculation of Phaseolus vulgaris with the nodule-endophyte Agrobacterium sp. 10C2 affects richness and structure of rhizosphere bacterial communities and enhances nodulation and growth. Archives of Microbiology. 2015;197:805–13.

* View Article

* Google Scholar

5. 5. Bosmans L, Moerkens R, Wittemans L, De Mot R, Rediers H, Lievens B. Rhizogenic agrobacteria in hydroponic crops: epidemics, diagnostics and control. Plant Pathology. 2017;66:1043–53.

* View Article

* Google Scholar

6. 6. Vargas P, Bosmans L, Van Kerckhove S, Van Calenberge B, Raaijmakers JM, Lievens B, et al. Optimizing Biocontrol Activity of Paenibacillus xylanexedens for Management of Hairy Root Disease in Tomato Grown in Hydroponic Greenhouses. Agronomy. 2021;11:817.

* View Article

* Google Scholar

7. 7. Puławska J. Crown gall on stone fruit trees. Acta Horticulturae. 2016;37–42.

* View Article

* Google Scholar

8. 8. Smith EF, Townsend CO. A PLANT-TUMOR OF BACTERIAL ORIGIN. Science. 1907;25:671–3.

* View Article

* Google Scholar

9. 9. Riker AJ, Banfield WM, Wright H, Keitt GW, Sagen HE. Studies on infectious hairy root of nursery Apple trees. Journal of Agricultural Research. 1930;41:507–40.

* View Article

* Google Scholar

10. 10. Beijerinck M, Van Delden A. Über die Assimilation des freien Stickstoffs durch Bakterien. Zentralblatt für Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene. 1902;2:3–43.

* View Article

* Google Scholar

11. 11. Conn HJ. Validity of the Genus Alcaligenes. Journal of Bacteriology. 1942;44:353–60.

* View Article

* Google Scholar

12. 12. Young JM, Kuykendall LD, Martínez-Romero E, Kerr A, Sawada H. A revision of Rhizobium Frank 1889, with an emended description of the genus, and the inclusion of all species of Agrobacterium Conn 1942 and Allorhizobium undicola de Lajudie et al. 1998 as new combinations: Rhizobium radiobacter, R. rhizogenes, R. rubi, R. undicola and R. vitis. International Journal of Systematic and Evolutionary Microbiology. 2001;51:89–103.

* View Article

* Google Scholar

13. 13. Bouzar H, Chilton WS, Nesme X, Dessaux Y, Vaudequin V, Petit A, et al. A new Agrobacterium strain isolated from aerial tumors on Ficus benjamina L. Applied and Environmental Microbiology. 1995;61:65–73.

* View Article

* Google Scholar

14. 14. Bouzar H, Jones JB. Agrobacterium larrymoorei sp. nov., a pathogen isolated from aerial tumours of Ficus benjamina. International Journal of Systematic and Evolutionary Microbiology. 2001;51:1023–6.

* View Article

* Google Scholar

15. 15. Hildebrand EM. Cane gall of brambles caused by Phytomonas n. sp. J Agric Res. 1940;61:685–96.

* View Article

* Google Scholar

16. 16. Starr MP, Weiss JE. Growth of phytopathogenic bacteria in a synthetic asparagin medium. Phytopathology. 1943;33:314–8.

* View Article

* Google Scholar

17. 17. Ophel K, Kerr A. Agrobacterium vitis sp. nov. for Strains of Agrobacterium biovar 3 from Grapevines. International Journal of Systematic Bacteriology. 1990;40:236–41.

* View Article

* Google Scholar

18. 18. Kerr A. Transfer of Virulence between Isolates of Agrobacterium. Nature. 1969;223:1175–6.

* View Article

* Google Scholar

19. 19. Genetello C, Van Larebeke N, Holsters M, De Picker A, Van Montagu M, Schell J. Ti plasmids of Agrobacterium as conjugative plasmids. Nature. 1977;265:561–3.

* View Article

* Google Scholar

20. 20. Gelvin SB. Agrobacterium-Mediated Plant Transformation: the Biology behind the “Gene-Jockeying” Tool. Microbiology and Molecular Biology Reviews. 2003;67:16–37.

* View Article

* Google Scholar

21. 21. Keane P, Kerr A, New P. Crown Gall of Stone Fruit II. Identification and Nomenclature of Agrobacterium Isolates. Australian Journal of Biological Sciences. 1970;23:585.

* View Article

* Google Scholar

22. 22. Lindström K, Young JPW. International Committee on Systematics of Prokaryotes Subcommittee on the taxonomy of Agrobacterium and Rhizobium. International Journal of Systematic and Evolutionary Microbiology. 2011;61:3089–93.

* View Article

* Google Scholar

23. 23. Farrand SK, van Berkum PB, Oger P. Agrobacterium is a definable genus of the family Rhizobiaceae. International Journal of Systematic and Evolutionary Microbiology. 2003;53:1681–7.

* View Article

* Google Scholar

24. 24. Mousavi SA, Österman J, Wahlberg N, Nesme X, Lavire C, Vial L, et al. Phylogeny of the Rhizobium–Allorhizobium–Agrobacterium clade supports the delineation of Neorhizobium gen. nov. Systematic and Applied Microbiology. 2014;37:208–15.

* View Article

* Google Scholar

25. 25. Gan HMi Savka MA. One more decade of Agrobacterium taxonomy. In: Gelvin S, editor. Agrobacterium Biology, from Basic Sicence to biotechnology. Switzerland: Springer; 2018. p. 2–14.

26. 26. Ormeño-Orrillo E, Servín-Garcidueñas LE, Rogel MA, González V, Peralta H, Mora J, et al. Taxonomy of rhizobia and agrobacteria from the Rhizobiaceae family in light of genomics. Systematic and Applied Microbiology. 2015;38:287–91.

* View Article

* Google Scholar

27. 27. Ramírez-Bahena MH, Vial L, Lassalle F, Diel B, Chapulliot D, Daubin V, et al. Single acquisition of protelomerase gave rise to speciation of a large and diverse clade within the Agrobacterium/Rhizobium supercluster characterized by the presence of a linear chromid. Molecular Phylogenetics and Evolution. 2014;73:202–7.

* View Article

* Google Scholar

28. 28. Kuzmanović N, Biondi E, Overmann J, Puławska J, Verbarg S, Smalla K, et al. Revisiting the taxonomy of Allorhizobium vitis (i.e. Agrobacterium vitis) using genomics—emended description of All. vitis sensu stricto and description of Allorhizobium ampelinum sp. nov. 2020.

* View Article

* Google Scholar

29. 29. Naranjo HD, Liesbeth Lebbe, Cnockaert M, Lassalle F, Chin Chin Too, Willems A Phylogenomics reveals insights into the functional evolution of the genus Agrobacterium and enables the description of Agrobacterium divergens sp. nov. Systematic and applied microbiology (Print). 2023;46:126420–0.

* View Article

* Google Scholar

30. 30. Costechareyre D, Rhouma A, Lavire C, Portier P, Chapulliot D, Bertolla F, et al. Rapid and Efficient Identification of Agrobacterium Species by recA Allele Analysis. Microbial Ecology. 2010;60:862–72.

* View Article

* Google Scholar

31. 31. Parte AC, Sardà Carbasse J, Meier-Kolthoff JP, Reimer LC, Göker M. List of Prokaryotic names with Standing in Nomenclature (LPSN) moves to the DSMZ. International Journal of Systematic and Evolutionary Microbiology. 2020;70:5607–12.

* View Article

* Google Scholar

32. 32. Singh NK, Lavire C, Nesme J, Vial L, Nesme X, Mason CE, et al. Comparative Genomics of Novel Agrobacterium G3 Strains Isolated From the International Space Station and Description of Agrobacterium tomkonis sp. nov. Frontiers in Microbiology. 2021;12.

* View Article

* Google Scholar

33. 33. Mafakheri H, Taghavi SM, Puławska J, de Lajudie P, Lassalle F, Osdaghi E. Two Novel Genomospecies in the Agrobacterium tumefaciens Species Complex Associated with Rose Crown Gall. Phytopathology®. 2019;109:1859–68.

* View Article

* Google Scholar

34. 34. Mousavi SA, Willems A, Nesme X, de Lajudie P, Lindström K. Revised phylogeny of Rhizobiaceae: Proposal of the delineation of Pararhizobium gen. nov., and 13 new species combinations. Systematic and Applied Microbiology. 2015;38:84–90.

* View Article

* Google Scholar

35. 35. Lassalle F, Planel R, Penel S, Chapulliot D, Barbe V, Dubost A, et al. Ancestral Genome Estimation Reveals the History of Ecological Diversification in Agrobacterium. Genome Biology and Evolution. 2017;9:3413–31.

* View Article

* Google Scholar

36. 36. Lassalle F, Campillo T, Vial L, Baude J, Costechareyre D, Chapulliot D, et al. Genomic Species Are Ecological Species as Revealed by Comparative Genomics in Agrobacterium tumefaciens. Genome Biology and Evolution. 2011;3:762–81.

* View Article

* Google Scholar

37. 37. Portier P, Fischer-Le Saux M, Mougel C, Lerondelle C, Chapulliot D, Thioulouse J, et al. Identification of Genomic Species in Agrobacterium Biovar 1 by AFLP Genomic Markers. Applied and Environmental Microbiology. 2006;72:7123–31.

* View Article

* Google Scholar

38. 38. Chun J, Oren A, Ventosa A, Christensen H, Arahal DR, da Costa MS, et al. Proposed minimal standards for the use of genome data for the taxonomy of prokaryotes. International Journal of Systematic and Evolutionary Microbiology. 2018;68:461–6.

* View Article

* Google Scholar

39. 39. Hayashi Sant’Anna F, Bach E, Porto RZ, Guella F, Hayashi Sant’Anna E, Passaglia LMP. Genomic metrics made easy: what to do and where to go in the new era of bacterial taxonomy. Critical Reviews in Microbiology. 2019;45:182–200.

* View Article

* Google Scholar

40. 40. Zhou Y, Zheng J, Wu Y, Zhang W, Jin J. A completeness-independent method for pre-selection of closely related genomes for species delineation in prokaryotes. BMC genomics. 2020;21.

* View Article

* Google Scholar

41. 41. Konstantinidis KT, Ramette A, Tiedje JM. Toward a More Robust Assessment of Intraspecies Diversity, Using Fewer Genetic Markers. Applied and Environmental Microbiology. 2006;72:7286–93.

* View Article

* Google Scholar

42. 42. de Lajudie PM, Andrews M, Ardley J, Eardly B, Jumas-Bilak E, Kuzmanović N, et al. Minimal standards for the description of new genera and species of rhizobia and agrobacteria. International Journal of Systematic and Evolutionary Microbiology. 2019;69:1852–63.

* View Article

* Google Scholar

43. 43. Davis E. get_assemblies v0.10.0. Available at https://github.com/davised/get_assemblies. 2020.

44. 44. Bosmans L, Álvarez-Pérez S, Moerkens R, Wittemans L, Van Calenberge B, Kerckhove SV, et al. Assessment of the genetic and phenotypic diversity among rhizogenic Agrobacterium biovar 1 strains infecting solanaceous and cucurbit crops. FEMS Microbiology Ecology. 2015;91: fiv081.

* View Article

* Google Scholar

45. 45. Kim N, Vargas P, Fortuna K, Jeroen Wagemans, Rediers H. Draft Genome Sequences of 27 Rhizogenic Agrobacterium Biovar 1 Strains, the Causative Agent of Hairy Root Disease. 2023;12.

* View Article

* Google Scholar

46. 46. Davis EW II, Weisberg AJ, Tabima JF, Grunwald NJ, Chang JH. Gall-ID: tools for genotyping gall-causing phytopathogenic bacteria. PeerJ. 2016;4: e2222.

* View Article

* Google Scholar

47. 47. Katoh K, Standley DM. MAFFT Multiple Sequence Alignment Software Version 7: Improvements in Performance and Usability. Molecular Biology and Evolution. 2013;30:772–80.

* View Article

* Google Scholar

48. 48. Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, et al. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Molecular Biology and Evolution. 2020;37:1530–4.

* View Article

* Google Scholar

49. 49. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. UFBoot2: Improving the Ultrafast Bootstrap Approximation. Molecular Biology and Evolution. 2017;35:518–22.

* View Article

* Google Scholar

50. 50. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nature Methods. 2017;14:587–9.

* View Article

* Google Scholar

51. 51. Asnicar F, Thomas AM, Beghini F, Mengoni C, Manara S, Manghi P, et al. Precise phylogenetic analysis of microbial isolates and genomes from metagenomes using PhyloPhlAn 3.0. Nature Communications. 2020;11.

* View Article

* Google Scholar

52. 52. Buchfink B, Xie C, Huson DH. Fast and sensitive protein alignment using DIAMOND. Nature Methods. 2014;12:59–60.

* View Article

* Google Scholar

53. 53. Letunic I, Bork P. Interactive Tree Of Life (iTOL) v4: recent updates and new developments. Nucleic Acids Research. 2019;47: W256–9.

* View Article

* Google Scholar

54. 54. Kembel SW, Cowan PD, Helmus MR, Cornwell WK, Morlon H, Ackerly DD, et al. Picante: R tools for integrating phylogenies and ecology. Bioinformatics. 2010;26:1463–4.

* View Article

* Google Scholar

55. 55. R Hackathon, Bolker B, Butler M, Cowan P, de Vienne D, Eddelbuettel D, et al. Package “phylobase” v. 0.8.10. https://github.com/fmichonneau/phylobase. 2022

56. 56. Jombart T, Balloux F, Dray S. adephylo: new tools for investigating the phylogenetic signal in biological traits. Bioinformatics. 2010;26:1907–9.

* View Article

* Google Scholar

57. 57. Pritchard L, Glover RH, Humphris S, Elphinstone JG, Toth IK. Genomics and taxonomy in diagnostics for food security: soft-rotting enterobacterial plant pathogens. Analytical Methods. 2016;8:12–24.

* View Article

* Google Scholar

58. 58. López-Guerrero MG, Ormeño-Orrillo E, Velázquez E, Rogel MA, Acosta JL, Gónzalez V, et al. Rhizobium etli taxonomy revised with novel genomic data and analyses. Systematic and Applied Microbiology. 2012;35:353–8.

* View Article

* Google Scholar

59. 59. Carroll LM, Wiedmann M, Kovac J. Proposal of a Taxonomic Nomenclature for the Bacillus cereus Group Which Reconciles Genomic Definitions of Bacterial Species with Clinical and Industrial Phenotypes. mBio. 2020;11.

* View Article

* Google Scholar

60. 60. Parks D. CompareM: A toolbox for comparative genomics. https://github.com/dparks1134/CompareM. 2014.

61. 61. Hyatt D, Chen G-L, LoCascio PF, Land ML, Larimer FW, Hauser LJ. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics. 2010;11.

* View Article

* Google Scholar

62. 62. Fan B, Blom J, Klenk H-P, Borriss R. Bacillus amyloliquefaciens, Bacillus velezensis, and Bacillus siamensis Form an “Operational Group B. amyloliquefaciens” within the B. subtilis Species Complex. Frontiers in Microbiology. 2017;8.

* View Article

* Google Scholar

63. 63. R Core Team. R: A language and environment for statistical. https://www.r-project.org/. 2022

64. 64. RStudio Team. RStudio: Integrated Development for R. RStudio, PBC, Boston, MA. 2020.

65. 65. Kolde R. pheatmap: Pretty Heatmaps. R package version 1.0.12. 2019.

* View Article

* Google Scholar

66. 66. Dehlinger B, Jurss J, Lychuk K, Putonti C. The Dynamic Codon Biaser: calculating prokaryotic codon usage biases. Microbial Genomics. 2021;7.

* View Article

* Google Scholar

67. 67. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. ISBN 978-3-319-24277-4, https://ggplot2.tidyverse.org. 2016.

68. 68. Chalasani D, Basu A, Pullabhotla SVSRN, Jorrin B, Neal AL, Poole PS, et al. Poor Competitiveness of Bradyrhizobium in Pigeon Pea Root Colonization in Indian Soils. mBio. 2021;12: e0042321.

* View Article

* Google Scholar

69. 69. Ma T, Xue H, Piao C, Jiang N, Li Y. Phylogenomic reappraisal of the family Rhizobiaceae at the genus and species levels, including the description of Ectorhizobium quercum gen. nov., sp. nov. Frontiers in microbiology. 2023;14.

* View Article

* Google Scholar

70. 70. Ciufo S, Kannan S, Sharma S, Badretdin A, Clark K, Turner S, et al. Using average nucleotide identity to improve taxonomic assignments in prokaryotic genomes at the NCBI. International Journal of Systematic and Evolutionary Microbiology. 2018;68:2386–92.

* View Article

* Google Scholar

71. 71. Chun J, Rainey FA. Integrating genomics into the taxonomy and systematics of the Bacteria and Archaea. International Journal of Systematic and Evolutionary Microbiology. 2014;64:316–24.

* View Article

* Google Scholar

72. 72. Liu A, Zhang Y-J, Cheng P, Peng Y-J, Blom J, Xue Q-J. Whole genome analysis calls for a taxonomic rearrangement of the genus Colwellia. Antonie van Leeuwenhoek. 2020;113:919–31.

* View Article

* Google Scholar

73. 73. Rajendhran J, Gunasekaran P. Microbial phylogeny and diversity: Small subunit ribosomal RNA sequence analysis and beyond. Microbiological Research. 2011;166:99–110.

* View Article

* Google Scholar

74. 74. Woo PCY, Lau SKP, Teng JLL, Tse H, Yuen K-Y. Then and now: use of 16S rDNA gene sequencing for bacterial identification and discovery of novel bacteria in clinical microbiology laboratories. Clinical Microbiology and Infection. 2008;14:908–34.

* View Article

* Google Scholar

75. 75. Tong W, Li X, Huo Y, Zhang L, Cao Y, Wang E, et al. Genomic insight into the taxonomy of Rhizobium genospecies that nodulate Phaseolus vulgaris. Systematic and Applied Microbiology. 2018;41:300–10.

* View Article

* Google Scholar

76. 76. Eardly BD, Nour SM, van Berkum P, Selander RK. Rhizobial 16S rRNA and dnaK Genes: Mosaicism and the Uncertain Phylogenetic Placement of Rhizobium galegae. Applied and Environmental Microbiology. 2005;71:1328–35.

* View Article

* Google Scholar

77. 77. Young JPW, Moeskjær S, Afonin A, Rahi P, Maluk M, James EK, et al. Defining the Rhizobium leguminosarum Species Complex. Genes. 2021;12:111.

* View Article

* Google Scholar

78. 78. Seyed Abdollah Mousavi Young PW. International Committee on Systematics of Prokaryotes Subcommittee on the Taxonomy of Rhizobia and Agrobacteria. Minutes of the closed annual meeting: videoconference on 11 October 2022 followed by online discussion until 31 December 2022. International Journal of Systematic and Evolutionary Microbiology. 2023;73.

* View Article

* Google Scholar

79. 79. Weiss B, Carolina A, Tavares M, Danillo Oliveira Alvarenga, ylro VS, Carareto LM, et al. Unraveling a Lignocellulose-Decomposing Bacterial Consortium from Soil Associated with Dry Sugarcane Straw by Genomic-Centered Metagenomics. Microorganisms. 2021;9:995–5.

* View Article

* Google Scholar

80. 80. de Lajudie PM, Young JPW. International Committee on Systematics of Prokaryotes Subcommittee on the taxonomy of rhizobia and agrobacteria Minutes of the closed meeting, Granada, 4 September 2017. International Journal of Systematic and Evolutionary Microbiology. 2018;68:3363–8.

* View Article

* Google Scholar

81. 81. Varghese NJ, Mukherjee S, Ivanova N, Konstantinidis KT, Mavrommatis K, Kyrpides NC, et al. Microbial species delineation using whole genome sequences. Nucleic Acids Research. 2015;43:6761–71.

* View Article

* Google Scholar

82. 82. Warabieda M, Kuzmanović N, Trzciński P, Puławska J. Agrobacterium cucumeris sp. nov. isolated from crazy roots on cucumber (Cucumis sativus). Systematic and Applied Microbiology. 2023;46:126402–2.

* View Article

* Google Scholar

83. 83. Kuzmanović N, Fagorzi C, Mengoni A, Lassalle F, diCenzo GC. Taxonomy of Rhizobiaceae revisited: proposal of a new framework for genus delimitation. 2021.

* View Article

* Google Scholar

84. 84. Vargas P, Van Kerkckhove S, Van Calenberge B, Bosmans L, Lievens B, Rediers H. First Report of Hairy Root Disease, Caused by Rhizogenic Agrobacterium Biovar 1, in Hydroponic Bell Pepper Crop in South Korea. Plant Disease. 2019;104:968.

* View Article

* Google Scholar

85. 85. Han I, Park K, Lee H, Park B, Lee Y, KIM J. First report of root mat disease in a hydroponic tomato production system caused by rhizogenic Agrobacterium biovar 1 in South Korea. Plant Disease. 2020. https://doi.org/10.1094/pdis-10-20-2179-pdn.

* View Article

* Google Scholar

86. 86. Penyalver R, López MM. Cocolonization of the Rhizosphere by PathogenicAgrobacterium Strains and Nonpathogenic Strains K84 and K1026, Used for Crown Gall Biocontrol. Applied and Environmental Microbiology. 1999;65:1936–40.

* View Article

* Google Scholar

87. 87. Richter M, Rosselló-Móra R, Oliver Glöckner F, Peplies J. JSpeciesWS: a web server for prokaryotic species circumscription based on pairwise genome comparison. Bioinformatics. 2015;32:929–31.

* View Article

* Google Scholar

88. 88. Girard L, Lood C, Höfte M, Vandamme P, Rokni-Zadeh H, van Noort V, et al. The Ever-Expanding Pseudomonas Genus: Description of 43 New Species and Partition of the Pseudomonas putida Group. Microorganisms. 2021;9:1766.

* View Article

* Google Scholar

89. 89. Tran PN, Savka MA, Gan HM. In-silico Taxonomic Classification of 373 Genomes Reveals Species Misidentification and New Genospecies within the Genus Pseudomonas. Frontiers in Microbiology. 2017;8.

* View Article

* Google Scholar

90. 90. Vinuesa P, Ochoa-Sánchez LE, Contreras-Moreira B. GET_PHYLOMARKERS, a Software Package to Select Optimal Orthologous Clusters for Phylogenomics and Inferring Pan-Genome Phylogenies, Used for a Critical Geno-Taxonomic Revision of the Genus Stenotrophomonas. Frontiers in Microbiology. 2018;9.

* View Article

* Google Scholar

91. 91. Jin Y, Zhou J, Zhou J, Hu M, Zhang Q, Kong N, et al. Genome-based classification of Burkholderia cepacia complex provides new insight into its taxonomic status. Biology Direct. 2020;15.

* View Article

* Google Scholar

92. 92. Konstantinidis KT, Tiedje JM. Towards a Genome-Based Taxonomy for Prokaryotes. Journal of Bacteriology. 2005;187:6258–64.

* View Article

* Google Scholar

93. 93. Richter M, Rosselló-Móra R. Shifting the genomic gold standard for the prokaryotic species definition. Proceedings of the National Academy of Sciences. 2009;106:19126–31.

* View Article

* Google Scholar

94. 94. Chou L, Lin Y-C, Haryono M, Santos MNM, Cho S-T, Weisberg AJ, et al. Modular evolution of secretion systems and virulence plasmids in a bacterial species complex. 2021. https://doi.org/10.1101/2021.05.20.444927.

* View Article

* Google Scholar

95. 95. Volpiano CG, Sant’Anna FH, Ambrosini A, de São José JFB, Beneduzi A, Whitman WB, et al. Genomic Metrics Applied to Rhizobiales (Hyphomicrobiales): Species Reclassification, Identification of Unauthentic Genomes and False Type Strains. Frontiers in Microbiology. 2021;12.

* View Article

* Google Scholar

96. 96. Garrido-Sanz D, Meier-Kolthoff JP, Göker M, Martín M, Rivilla R, Redondo-Nieto M. Genomic and Genetic Diversity within the Pseudomonas fluorescens Complex. PLOS ONE. 2016;11: e0150183.

* View Article

* Google Scholar

97. 97. Konstantinidis K, Ruiz-Perez C, Gerhardt K, Rodriguez-R L, Jain C, Tiedje J, et al. FastAAI: Efficient Estimation of Genome Average Amino Acid Identity and Phylum-level relationships using Tetramers of Universal Proteins. 2022. https://doi.org/10.21203/rs.3.rs-1459378/v1.

* View Article

* Google Scholar

98. 98. Thompson CC, Chimetto L, Edwards RA, Swings J, Stackebrandt E, Thompson FL. Microbial genomic taxonomy. BMC Genomics. 2013;14:913.

* View Article

* Google Scholar

99. 99. Kim D, Park S, Chun J. Introducing EzAAI: a pipeline for high throughput calculations of prokaryotic average amino acid identity. Journal of Microbiology. 2021;59:476–80.

* View Article

* Google Scholar

100. 100. Kämpfer P, Glaeser SP. Prokaryotic taxonomy in the sequencing era—the polyphasic approach revisited. Environmental Microbiology. 2012;14:291–317.

* View Article

* Google Scholar

Citation: Vargas Ribera PR, Kim N, Venbrux M, Álvarez-Pérez S, Rediers H (2024) Evaluation of sequence-based tools to gather more insight into the positioning of rhizogenic agrobacteria within the Agrobacterium tumefaciens species complex. PLoS ONE 19(11): e0302954. https://doi.org/10.1371/journal.pone.0302954

About the Authors:

Pablo Roberto Vargas Ribera

Roles: Conceptualization, Formal analysis, Investigation, Methodology, Writing – original draft

Affiliations: Sustainable Plant Protection, Centre de Cabrils, IRTA–Institute of Agrifood Research and Technology, Cabrils, Spain, CMPG Laboratory for Process Microbial Ecology and Bioinspirational Management (PME & BIM), Department Microbial and Molecular Systems (M2S), KU Leuven, Leuven, Belgium

ORICD: https://orcid.org/0000-0003-0435-284X

Nuri Kim

Roles: Formal analysis, Writing – review & editing

Affiliations: CMPG Laboratory for Process Microbial Ecology and Bioinspirational Management (PME & BIM), Department Microbial and Molecular Systems (M2S), KU Leuven, Leuven, Belgium, Leuven Plant Institute (LPI), Leuven, Belgium

Marc Venbrux

Roles: Formal analysis, Writing – review & editing

Affiliations: CMPG Laboratory for Process Microbial Ecology and Bioinspirational Management (PME & BIM), Department Microbial and Molecular Systems (M2S), KU Leuven, Leuven, Belgium, Leuven Plant Institute (LPI), Leuven, Belgium

Sergio Álvarez-Pérez

Roles: Formal analysis, Investigation, Methodology, Writing – review & editing

Affiliation: Animal Health Department, Faculty of Veterinary Medicine, Complutense University of Madrid, Madrid, Spain

ORICD: https://orcid.org/0000-0002-6587-8995

Hans Rediers

Roles: Funding acquisition, Supervision, Validation, Writing – original draft, Writing – review & editing

E-mail: [email protected]

Affiliations: CMPG Laboratory for Process Microbial Ecology and Bioinspirational Management (PME & BIM), Department Microbial and Molecular Systems (M2S), KU Leuven, Leuven, Belgium, Leuven Plant Institute (LPI), Leuven, Belgium

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

[/RAW_REF_TEXT]

References

1. Flores-Félix JD, Menéndez E, Peix A, García-Fraile P, Velázquez E. History and current taxonomic status of genus Agrobacterium. Systematic and Applied Microbiology. 2020;43:126046.

2. Young JM, Kerr A, Sawada H. Agrobacterium. Bergey’s Manual of Systematics of Archaea and Bacteria. 2015;1–15.

3. Dessaux Y, Faure D. Niche Construction and Exploitation by Agrobacterium: How to Survive and Face Competition in Soil and Plant Habitats. In: Gelvin S, editor. Agrobacterium Biology, from basic science to biotechnology. Springer; 2018. p. 55–86.

4. Chihaoui S-A, Trabelsi D, Jdey A, Mhadhbi H, Mhamdi R. Inoculation of Phaseolus vulgaris with the nodule-endophyte Agrobacterium sp. 10C2 affects richness and structure of rhizosphere bacterial communities and enhances nodulation and growth. Archives of Microbiology. 2015;197:805–13.

5. Bosmans L, Moerkens R, Wittemans L, De Mot R, Rediers H, Lievens B. Rhizogenic agrobacteria in hydroponic crops: epidemics, diagnostics and control. Plant Pathology. 2017;66:1043–53.

6. Vargas P, Bosmans L, Van Kerckhove S, Van Calenberge B, Raaijmakers JM, Lievens B, et al. Optimizing Biocontrol Activity of Paenibacillus xylanexedens for Management of Hairy Root Disease in Tomato Grown in Hydroponic Greenhouses. Agronomy. 2021;11:817.

7. Puławska J. Crown gall on stone fruit trees. Acta Horticulturae. 2016;37–42.

8. Smith EF, Townsend CO. A PLANT-TUMOR OF BACTERIAL ORIGIN. Science. 1907;25:671–3.

9. Riker AJ, Banfield WM, Wright H, Keitt GW, Sagen HE. Studies on infectious hairy root of nursery Apple trees. Journal of Agricultural Research. 1930;41:507–40.

10. Beijerinck M, Van Delden A. Über die Assimilation des freien Stickstoffs durch Bakterien. Zentralblatt für Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene. 1902;2:3–43.

11. Conn HJ. Validity of the Genus Alcaligenes. Journal of Bacteriology. 1942;44:353–60.

12. Young JM, Kuykendall LD, Martínez-Romero E, Kerr A, Sawada H. A revision of Rhizobium Frank 1889, with an emended description of the genus, and the inclusion of all species of Agrobacterium Conn 1942 and Allorhizobium undicola de Lajudie et al. 1998 as new combinations: Rhizobium radiobacter, R. rhizogenes, R. rubi, R. undicola and R. vitis. International Journal of Systematic and Evolutionary Microbiology. 2001;51:89–103.

13. Bouzar H, Chilton WS, Nesme X, Dessaux Y, Vaudequin V, Petit A, et al. A new Agrobacterium strain isolated from aerial tumors on Ficus benjamina L. Applied and Environmental Microbiology. 1995;61:65–73.

14. Bouzar H, Jones JB. Agrobacterium larrymoorei sp. nov., a pathogen isolated from aerial tumours of Ficus benjamina. International Journal of Systematic and Evolutionary Microbiology. 2001;51:1023–6.

15. Hildebrand EM. Cane gall of brambles caused by Phytomonas n. sp. J Agric Res. 1940;61:685–96.

16. Starr MP, Weiss JE. Growth of phytopathogenic bacteria in a synthetic asparagin medium. Phytopathology. 1943;33:314–8.

17. Ophel K, Kerr A. Agrobacterium vitis sp. nov. for Strains of Agrobacterium biovar 3 from Grapevines. International Journal of Systematic Bacteriology. 1990;40:236–41.

18. Kerr A. Transfer of Virulence between Isolates of Agrobacterium. Nature. 1969;223:1175–6.

19. Genetello C, Van Larebeke N, Holsters M, De Picker A, Van Montagu M, Schell J. Ti plasmids of Agrobacterium as conjugative plasmids. Nature. 1977;265:561–3.

20. Gelvin SB. Agrobacterium-Mediated Plant Transformation: the Biology behind the “Gene-Jockeying” Tool. Microbiology and Molecular Biology Reviews. 2003;67:16–37.

21. Keane P, Kerr A, New P. Crown Gall of Stone Fruit II. Identification and Nomenclature of Agrobacterium Isolates. Australian Journal of Biological Sciences. 1970;23:585.

22. Lindström K, Young JPW. International Committee on Systematics of Prokaryotes Subcommittee on the taxonomy of Agrobacterium and Rhizobium. International Journal of Systematic and Evolutionary Microbiology. 2011;61:3089–93.

23. Farrand SK, van Berkum PB, Oger P. Agrobacterium is a definable genus of the family Rhizobiaceae. International Journal of Systematic and Evolutionary Microbiology. 2003;53:1681–7.

24. Mousavi SA, Österman J, Wahlberg N, Nesme X, Lavire C, Vial L, et al. Phylogeny of the Rhizobium–Allorhizobium–Agrobacterium clade supports the delineation of Neorhizobium gen. nov. Systematic and Applied Microbiology. 2014;37:208–15.

25. Gan HMi Savka MA. One more decade of Agrobacterium taxonomy. In: Gelvin S, editor. Agrobacterium Biology, from Basic Sicence to biotechnology. Switzerland: Springer; 2018. p. 2–14.

26. Ormeño-Orrillo E, Servín-Garcidueñas LE, Rogel MA, González V, Peralta H, Mora J, et al. Taxonomy of rhizobia and agrobacteria from the Rhizobiaceae family in light of genomics. Systematic and Applied Microbiology. 2015;38:287–91.

27. Ramírez-Bahena MH, Vial L, Lassalle F, Diel B, Chapulliot D, Daubin V, et al. Single acquisition of protelomerase gave rise to speciation of a large and diverse clade within the Agrobacterium/Rhizobium supercluster characterized by the presence of a linear chromid. Molecular Phylogenetics and Evolution. 2014;73:202–7.

28. Kuzmanović N, Biondi E, Overmann J, Puławska J, Verbarg S, Smalla K, et al. Revisiting the taxonomy of Allorhizobium vitis (i.e. Agrobacterium vitis) using genomics—emended description of All. vitis sensu stricto and description of Allorhizobium ampelinum sp. nov. 2020.

29. Naranjo HD, Liesbeth Lebbe, Cnockaert M, Lassalle F, Chin Chin Too, Willems A Phylogenomics reveals insights into the functional evolution of the genus Agrobacterium and enables the description of Agrobacterium divergens sp. nov. Systematic and applied microbiology (Print). 2023;46:126420–0.

30. Costechareyre D, Rhouma A, Lavire C, Portier P, Chapulliot D, Bertolla F, et al. Rapid and Efficient Identification of Agrobacterium Species by recA Allele Analysis. Microbial Ecology. 2010;60:862–72.

31. Parte AC, Sardà Carbasse J, Meier-Kolthoff JP, Reimer LC, Göker M. List of Prokaryotic names with Standing in Nomenclature (LPSN) moves to the DSMZ. International Journal of Systematic and Evolutionary Microbiology. 2020;70:5607–12.

32. Singh NK, Lavire C, Nesme J, Vial L, Nesme X, Mason CE, et al. Comparative Genomics of Novel Agrobacterium G3 Strains Isolated From the International Space Station and Description of Agrobacterium tomkonis sp. nov. Frontiers in Microbiology. 2021;12.

33. Mafakheri H, Taghavi SM, Puławska J, de Lajudie P, Lassalle F, Osdaghi E. Two Novel Genomospecies in the Agrobacterium tumefaciens Species Complex Associated with Rose Crown Gall. Phytopathology®. 2019;109:1859–68.

34. Mousavi SA, Willems A, Nesme X, de Lajudie P, Lindström K. Revised phylogeny of Rhizobiaceae: Proposal of the delineation of Pararhizobium gen. nov., and 13 new species combinations. Systematic and Applied Microbiology. 2015;38:84–90.

35. Lassalle F, Planel R, Penel S, Chapulliot D, Barbe V, Dubost A, et al. Ancestral Genome Estimation Reveals the History of Ecological Diversification in Agrobacterium. Genome Biology and Evolution. 2017;9:3413–31.

36. Lassalle F, Campillo T, Vial L, Baude J, Costechareyre D, Chapulliot D, et al. Genomic Species Are Ecological Species as Revealed by Comparative Genomics in Agrobacterium tumefaciens. Genome Biology and Evolution. 2011;3:762–81.

37. Portier P, Fischer-Le Saux M, Mougel C, Lerondelle C, Chapulliot D, Thioulouse J, et al. Identification of Genomic Species in Agrobacterium Biovar 1 by AFLP Genomic Markers. Applied and Environmental Microbiology. 2006;72:7123–31.

38. Chun J, Oren A, Ventosa A, Christensen H, Arahal DR, da Costa MS, et al. Proposed minimal standards for the use of genome data for the taxonomy of prokaryotes. International Journal of Systematic and Evolutionary Microbiology. 2018;68:461–6.

39. Hayashi Sant’Anna F, Bach E, Porto RZ, Guella F, Hayashi Sant’Anna E, Passaglia LMP. Genomic metrics made easy: what to do and where to go in the new era of bacterial taxonomy. Critical Reviews in Microbiology. 2019;45:182–200.

40. Zhou Y, Zheng J, Wu Y, Zhang W, Jin J. A completeness-independent method for pre-selection of closely related genomes for species delineation in prokaryotes. BMC genomics. 2020;21.

41. Konstantinidis KT, Ramette A, Tiedje JM. Toward a More Robust Assessment of Intraspecies Diversity, Using Fewer Genetic Markers. Applied and Environmental Microbiology. 2006;72:7286–93.

42. de Lajudie PM, Andrews M, Ardley J, Eardly B, Jumas-Bilak E, Kuzmanović N, et al. Minimal standards for the description of new genera and species of rhizobia and agrobacteria. International Journal of Systematic and Evolutionary Microbiology. 2019;69:1852–63.

43. Davis E. get_assemblies v0.10.0. Available at https://github.com/davised/get_assemblies. 2020.

44. Bosmans L, Álvarez-Pérez S, Moerkens R, Wittemans L, Van Calenberge B, Kerckhove SV, et al. Assessment of the genetic and phenotypic diversity among rhizogenic Agrobacterium biovar 1 strains infecting solanaceous and cucurbit crops. FEMS Microbiology Ecology. 2015;91: fiv081.

45. Kim N, Vargas P, Fortuna K, Jeroen Wagemans, Rediers H. Draft Genome Sequences of 27 Rhizogenic Agrobacterium Biovar 1 Strains, the Causative Agent of Hairy Root Disease. 2023;12.

46. Davis EW II, Weisberg AJ, Tabima JF, Grunwald NJ, Chang JH. Gall-ID: tools for genotyping gall-causing phytopathogenic bacteria. PeerJ. 2016;4: e2222.

47. Katoh K, Standley DM. MAFFT Multiple Sequence Alignment Software Version 7: Improvements in Performance and Usability. Molecular Biology and Evolution. 2013;30:772–80.

48. Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, et al. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Molecular Biology and Evolution. 2020;37:1530–4.

49. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. UFBoot2: Improving the Ultrafast Bootstrap Approximation. Molecular Biology and Evolution. 2017;35:518–22.

50. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nature Methods. 2017;14:587–9.

51. Asnicar F, Thomas AM, Beghini F, Mengoni C, Manara S, Manghi P, et al. Precise phylogenetic analysis of microbial isolates and genomes from metagenomes using PhyloPhlAn 3.0. Nature Communications. 2020;11.

52. Buchfink B, Xie C, Huson DH. Fast and sensitive protein alignment using DIAMOND. Nature Methods. 2014;12:59–60.

53. Letunic I, Bork P. Interactive Tree Of Life (iTOL) v4: recent updates and new developments. Nucleic Acids Research. 2019;47: W256–9.

54. Kembel SW, Cowan PD, Helmus MR, Cornwell WK, Morlon H, Ackerly DD, et al. Picante: R tools for integrating phylogenies and ecology. Bioinformatics. 2010;26:1463–4.

55. R Hackathon, Bolker B, Butler M, Cowan P, de Vienne D, Eddelbuettel D, et al. Package “phylobase” v. 0.8.10. https://github.com/fmichonneau/phylobase. 2022

56. Jombart T, Balloux F, Dray S. adephylo: new tools for investigating the phylogenetic signal in biological traits. Bioinformatics. 2010;26:1907–9.

57. Pritchard L, Glover RH, Humphris S, Elphinstone JG, Toth IK. Genomics and taxonomy in diagnostics for food security: soft-rotting enterobacterial plant pathogens. Analytical Methods. 2016;8:12–24.

58. López-Guerrero MG, Ormeño-Orrillo E, Velázquez E, Rogel MA, Acosta JL, Gónzalez V, et al. Rhizobium etli taxonomy revised with novel genomic data and analyses. Systematic and Applied Microbiology. 2012;35:353–8.

59. Carroll LM, Wiedmann M, Kovac J. Proposal of a Taxonomic Nomenclature for the Bacillus cereus Group Which Reconciles Genomic Definitions of Bacterial Species with Clinical and Industrial Phenotypes. mBio. 2020;11.

60. Parks D. CompareM: A toolbox for comparative genomics. https://github.com/dparks1134/CompareM. 2014.

61. Hyatt D, Chen G-L, LoCascio PF, Land ML, Larimer FW, Hauser LJ. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics. 2010;11.

62. Fan B, Blom J, Klenk H-P, Borriss R. Bacillus amyloliquefaciens, Bacillus velezensis, and Bacillus siamensis Form an “Operational Group B. amyloliquefaciens” within the B. subtilis Species Complex. Frontiers in Microbiology. 2017;8.

63. R Core Team. R: A language and environment for statistical. https://www.r-project.org/. 2022

64. RStudio Team. RStudio: Integrated Development for R. RStudio, PBC, Boston, MA. 2020.

65. Kolde R. pheatmap: Pretty Heatmaps. R package version 1.0.12. 2019.

66. Dehlinger B, Jurss J, Lychuk K, Putonti C. The Dynamic Codon Biaser: calculating prokaryotic codon usage biases. Microbial Genomics. 2021;7.

67. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. ISBN 978-3-319-24277-4, https://ggplot2.tidyverse.org. 2016.

68. Chalasani D, Basu A, Pullabhotla SVSRN, Jorrin B, Neal AL, Poole PS, et al. Poor Competitiveness of Bradyrhizobium in Pigeon Pea Root Colonization in Indian Soils. mBio. 2021;12: e0042321.

69. Ma T, Xue H, Piao C, Jiang N, Li Y. Phylogenomic reappraisal of the family Rhizobiaceae at the genus and species levels, including the description of Ectorhizobium quercum gen. nov., sp. nov. Frontiers in microbiology. 2023;14.

70. Ciufo S, Kannan S, Sharma S, Badretdin A, Clark K, Turner S, et al. Using average nucleotide identity to improve taxonomic assignments in prokaryotic genomes at the NCBI. International Journal of Systematic and Evolutionary Microbiology. 2018;68:2386–92.

71. Chun J, Rainey FA. Integrating genomics into the taxonomy and systematics of the Bacteria and Archaea. International Journal of Systematic and Evolutionary Microbiology. 2014;64:316–24.

72. Liu A, Zhang Y-J, Cheng P, Peng Y-J, Blom J, Xue Q-J. Whole genome analysis calls for a taxonomic rearrangement of the genus Colwellia. Antonie van Leeuwenhoek. 2020;113:919–31.

73. Rajendhran J, Gunasekaran P. Microbial phylogeny and diversity: Small subunit ribosomal RNA sequence analysis and beyond. Microbiological Research. 2011;166:99–110.

74. Woo PCY, Lau SKP, Teng JLL, Tse H, Yuen K-Y. Then and now: use of 16S rDNA gene sequencing for bacterial identification and discovery of novel bacteria in clinical microbiology laboratories. Clinical Microbiology and Infection. 2008;14:908–34.

75. Tong W, Li X, Huo Y, Zhang L, Cao Y, Wang E, et al. Genomic insight into the taxonomy of Rhizobium genospecies that nodulate Phaseolus vulgaris. Systematic and Applied Microbiology. 2018;41:300–10.

76. Eardly BD, Nour SM, van Berkum P, Selander RK. Rhizobial 16S rRNA and dnaK Genes: Mosaicism and the Uncertain Phylogenetic Placement of Rhizobium galegae. Applied and Environmental Microbiology. 2005;71:1328–35.

77. Young JPW, Moeskjær S, Afonin A, Rahi P, Maluk M, James EK, et al. Defining the Rhizobium leguminosarum Species Complex. Genes. 2021;12:111.

78. Seyed Abdollah Mousavi Young PW. International Committee on Systematics of Prokaryotes Subcommittee on the Taxonomy of Rhizobia and Agrobacteria. Minutes of the closed annual meeting: videoconference on 11 October 2022 followed by online discussion until 31 December 2022. International Journal of Systematic and Evolutionary Microbiology. 2023;73.

79. Weiss B, Carolina A, Tavares M, Danillo Oliveira Alvarenga, ylro VS, Carareto LM, et al. Unraveling a Lignocellulose-Decomposing Bacterial Consortium from Soil Associated with Dry Sugarcane Straw by Genomic-Centered Metagenomics. Microorganisms. 2021;9:995–5.

80. de Lajudie PM, Young JPW. International Committee on Systematics of Prokaryotes Subcommittee on the taxonomy of rhizobia and agrobacteria Minutes of the closed meeting, Granada, 4 September 2017. International Journal of Systematic and Evolutionary Microbiology. 2018;68:3363–8.

81. Varghese NJ, Mukherjee S, Ivanova N, Konstantinidis KT, Mavrommatis K, Kyrpides NC, et al. Microbial species delineation using whole genome sequences. Nucleic Acids Research. 2015;43:6761–71.

82. Warabieda M, Kuzmanović N, Trzciński P, Puławska J. Agrobacterium cucumeris sp. nov. isolated from crazy roots on cucumber (Cucumis sativus). Systematic and Applied Microbiology. 2023;46:126402–2.

83. Kuzmanović N, Fagorzi C, Mengoni A, Lassalle F, diCenzo GC. Taxonomy of Rhizobiaceae revisited: proposal of a new framework for genus delimitation. 2021.

84. Vargas P, Van Kerkckhove S, Van Calenberge B, Bosmans L, Lievens B, Rediers H. First Report of Hairy Root Disease, Caused by Rhizogenic Agrobacterium Biovar 1, in Hydroponic Bell Pepper Crop in South Korea. Plant Disease. 2019;104:968.

85. Han I, Park K, Lee H, Park B, Lee Y, KIM J. First report of root mat disease in a hydroponic tomato production system caused by rhizogenic Agrobacterium biovar 1 in South Korea. Plant Disease. 2020. https://doi.org/10.1094/pdis-10-20-2179-pdn.

86. Penyalver R, López MM. Cocolonization of the Rhizosphere by PathogenicAgrobacterium Strains and Nonpathogenic Strains K84 and K1026, Used for Crown Gall Biocontrol. Applied and Environmental Microbiology. 1999;65:1936–40.

87. Richter M, Rosselló-Móra R, Oliver Glöckner F, Peplies J. JSpeciesWS: a web server for prokaryotic species circumscription based on pairwise genome comparison. Bioinformatics. 2015;32:929–31.

88. Girard L, Lood C, Höfte M, Vandamme P, Rokni-Zadeh H, van Noort V, et al. The Ever-Expanding Pseudomonas Genus: Description of 43 New Species and Partition of the Pseudomonas putida Group. Microorganisms. 2021;9:1766.

89. Tran PN, Savka MA, Gan HM. In-silico Taxonomic Classification of 373 Genomes Reveals Species Misidentification and New Genospecies within the Genus Pseudomonas. Frontiers in Microbiology. 2017;8.

90. Vinuesa P, Ochoa-Sánchez LE, Contreras-Moreira B. GET_PHYLOMARKERS, a Software Package to Select Optimal Orthologous Clusters for Phylogenomics and Inferring Pan-Genome Phylogenies, Used for a Critical Geno-Taxonomic Revision of the Genus Stenotrophomonas. Frontiers in Microbiology. 2018;9.

91. Jin Y, Zhou J, Zhou J, Hu M, Zhang Q, Kong N, et al. Genome-based classification of Burkholderia cepacia complex provides new insight into its taxonomic status. Biology Direct. 2020;15.

92. Konstantinidis KT, Tiedje JM. Towards a Genome-Based Taxonomy for Prokaryotes. Journal of Bacteriology. 2005;187:6258–64.

93. Richter M, Rosselló-Móra R. Shifting the genomic gold standard for the prokaryotic species definition. Proceedings of the National Academy of Sciences. 2009;106:19126–31.

94. Chou L, Lin Y-C, Haryono M, Santos MNM, Cho S-T, Weisberg AJ, et al. Modular evolution of secretion systems and virulence plasmids in a bacterial species complex. 2021. https://doi.org/10.1101/2021.05.20.444927.

95. Volpiano CG, Sant’Anna FH, Ambrosini A, de São José JFB, Beneduzi A, Whitman WB, et al. Genomic Metrics Applied to Rhizobiales (Hyphomicrobiales): Species Reclassification, Identification of Unauthentic Genomes and False Type Strains. Frontiers in Microbiology. 2021;12.

96. Garrido-Sanz D, Meier-Kolthoff JP, Göker M, Martín M, Rivilla R, Redondo-Nieto M. Genomic and Genetic Diversity within the Pseudomonas fluorescens Complex. PLOS ONE. 2016;11: e0150183.

97. Konstantinidis K, Ruiz-Perez C, Gerhardt K, Rodriguez-R L, Jain C, Tiedje J, et al. FastAAI: Efficient Estimation of Genome Average Amino Acid Identity and Phylum-level relationships using Tetramers of Universal Proteins. 2022. https://doi.org/10.21203/rs.3.rs-1459378/v1.

98. Thompson CC, Chimetto L, Edwards RA, Swings J, Stackebrandt E, Thompson FL. Microbial genomic taxonomy. BMC Genomics. 2013;14:913.

99. Kim D, Park S, Chun J. Introducing EzAAI: a pipeline for high throughput calculations of prokaryotic average amino acid identity. Journal of Microbiology. 2021;59:476–80.

100. Kämpfer P, Glaeser SP. Prokaryotic taxonomy in the sequencing era—the polyphasic approach revisited. Environmental Microbiology. 2012;14:291–317.