Taxogenomics of the Genus Cyclobacterium: Cyclobacterium xiamenense and Cyclobacterium halophilum as Synonyms and Description of Cyclobacterium plantarum sp. nov.

The genus Cyclobacterium belongs to the phylum Bacteroidetes and includes eight species. Our study, based on the genomic parameters in silico DNA–DNA hybridization (GGDC), average nucleotide identity (OrthoANI), and average amino acid identity (AAI), confirmed that all current species of Cyclobacterium belong to this genus and constitute a coherent phylogenomic group, but with species forming two separate branches. In addition, the genome-based analyses revealed that Cyclobacterium xiamenense and Cyclobacterium halophilum are members of the same species. Besides, we carried out a taxonomic characterization of the new strain GBPx2T, isolated from the halophytic plant Salicornia sp. Analysis of its 16S rRNA gene sequence showed the highest sequence similarity (97.5%) to Cyclobacterium lianum HY9T. Percentages of GGDC and OrthoANI between strain GBPx2T and species of the genus Cyclobacterium were lower than the threshold value for species delineation. The DNA G+C content was 43.0 mol%. The polar lipids included phosphatidylethanolamine as well as one unidentified phospholipid and four unidentified lipids, and its major cellular fatty acids were iso-C15:0 and summed feature 3 (C16:1ω7c and/or iso-C15:0 2-OH). The only quinone present was menaquinone 7. Based on a combination of phenotypic, chemotaxonomic, and phylogenomic features, the GBPx2T strain represents a novel species of the genus Cyclobacterium, for which the name Cyclobacterium plantarum sp. nov. is proposed. The type strain of Cyclobacterium plantarum is GBPx2T (= IBRC-M 10634T = LMG 28551T).


Phylogenetic Analysis Based on 16S rRNA Gene Sequence Comparison
The partial 16S rRNA gene was amplified using the universal primer pairs 16F27 and 16R1488 [19]. The PCR products were visualized on 1% agarose gel. The forward and reverse sequences were assembled by using Chromas Pro 1.7.7 (Technelysium Pty Ltd., South Brisbane, Australia). The 16S rRNA gene sequence of strain GBPx2 T was obtained and used for BLAST searches in GenBank and phylogenetic analysis. The identification of phylogenetic neighbors and calculation of pairwise 16S rRNA gene sequence similarity were achieved using the EzBioCloud server (https://www.ezbiocloud. net/) [20] and the alignments were performed by CLUSTAL-X [21]. Evolutionary distances between aligned 16S rRNA gene sequences of strain GBPx2 T with the most closely related type strains were calculated using the Jukes-Cantor model, and phylogenetic trees were reconstructed by the neighbour-joining [22], minimum-evolution [23], and maximum-likelihood [24] methods using the MEGA version 6 program [25]. Bootstrap analysis was carried out to evaluate the tree topology by performing resampling 1000 times [26]. The GenBank/EMBL/DDBJ accession number for the 16S rRNA gene sequence of strain GBPx2 T is MG457806. The 16S rRNA gene sequences of the reference type strains used for the phylogenetic comparison were obtained from GenBank database and their accession numbers are shown in Figure 1.

Genome Assembly and Annotation
The novo assembly of the reads of the genome of strain GBPx2 T was performed using Spades 3.13.0 [27]. The quality of final contigs was assessed by bioinformatics tools CheckM v1.0.5 [28] and Quast v2.3 [29]. The genome sequence was annotated using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) [30]. The genome of strain GBPx2 T was deposited in GenBank/EMBL/DDBJ under the accession number JAANYN00000000.

Phylogenomic Comparative Analysis
For the phylogenomic comparative analysis we used genomes available from GenBank database. The characteristics of the genomes and their accession numbers of the type strains of species of the genus Cyclobacterium are shown in Table 1. The quality of these genome sequences was in accordance with the recommended minimal standards for the use of genome data for the taxonomy of prokaryotes [31].
To determine the core-genome, the Enveomics [32] tool was used. To identify clusters of orthologous genes (OGs), an all-versus-all BLAST search based on protein-coding gene annotated sequences of strain GBPx2 T and all type species of the genera included in the family Cyclobacteriaceae available in databases was carried out. Those OGs shared among all taxa and present in a single copy per genome were selected. They were aligned with MUSCLE v. 3.8.31 [33] and subsequently concatenated. A maximum-likelihood tree was constructed using FastTree v. 2.1.9 [34] with the JTT replacement matrix [35] under the CAT approximation (single rate for each site) with 20 rate categories. Local support values were estimated with the Shimodaira-Hasegawa test [36]. The genomic parameters of in silico DNA-DNA hybridization (GGDC), average nucleotide identity (OrthoANI), and average amino acid identity (AAI) among strain GBPx2 T , the type strains of species of the genus Cyclobacterium, and the type species of the family Cyclobacteriaceae available from databases were determined. GGDC was calculated by the bioinformatic tool Genome-to-Genome Distance Calculator (GGDC version 2.1) available from the Leibniz Institute DSMZ [37]. The OrthoANI was calculated with ChunLab's Orthologous Average Nucleotide Identity Tool (OAT) [38]. For the estimation of the AAI, the CompareM program (https://github.com/dparks1134/CompareM) was used.

Phenotypic Characterization
Cell morphology and motility were examined using an Olympus BX51 microscope equipped with phase-contrast optics with cells from exponentially growing cultures. Gram staining was performed by the Burke method [39]. Motility was determined by the wet-mount method [39]. Colony morphology was observed on MA agar medium under optimal growth conditions after incubation at 25 • C for two days. To determine the temperature and pH ranges for growth, broth cultures of MA medium were incubated at 0, 4, 10, 15, 20, 25-37 (at intervals of 1.0 • C), 40, and 45 • C and at pH 5-10 at intervals of 0.5 pH units; the buffers sodium acetate/acetic acid (pH 5.0-6.0), Tris/HCl (pH 6.5-8.5), and glycine/sodium hydroxide (pH 9.0-10.0) were added at a concentration of 50 mM. The requirements for NaCl for growth were determined in media containing 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.5, 10.0, 12.5, and 15.0% (w/v) NaCl. Liquid cultures were incubated on a shaking incubator at 150 rpm and growth rates were determined by monitoring the increase in the optical density (OD) at 600 nm (ThermoSpectronics Spectronic 20D+).
Catalase and oxidase tests, nitrate and nitrite reduction, hydrolysis of aesculin, and production of indole and H 2 S were carried out as recommended by Smibert and Krieg [40], using media with 5% (w/v) NaCl. Hydrolysis activity of Tween 20, 40, and 80 was detected as described by Gutiérrez and González [41]. Hydrolysis of gelatin, casein, tyrosine, and starch, and activity for urease and DNase were determined as described by Mata et al. [42]. The anaerobic growth of the strain was tested in the presence of nitrate by adding 0.1% (w/v) KNO 3 to the medium with 5% (w/v) NaCl in filled stoppered tubes in an anaerobic chamber [43]. Acid production from carbohydrates was tested in unbuffered medium and was determined by measuring the initial and final pH of the medium. The culture was considered positive for acid production if the pH decreased by at least 1 unit. Tests for the utilization of different compounds as the sole source of carbon and energy were performed as recommended by Ventosa et al. [44].

Chemotaxonomic Characterization
Cell biomass for fatty acids, isoprenoid quinones, and polar lipids analyses was obtained by cultivation on MA medium at pH 8 and 25 • C. Cells were harvested in the mid-exponential growth phase determined spectrophotometrically with an optical density at 600 nm (OD 600 ). The whole-cell fatty acids composition of strain GBPx2 T was determined according to the standard protocol of the Microbial Identification System (MIDI, Version 6.1; Identification Library TSBA40 4.1; Microbial ID). Extracts were analyzed using a Hewlett Packard model HP6890A gas chromatograph equipped with a flame-ionization detector as described by Kämpfer and Kroppenstedt [45]. Fatty acids peaks were identified using the TSBA40 database. The polar lipids and respiratory quinones of strain GBPx2 T were analyzed as described by Groth et al. [46].

Phylogenetic Analysis Based on 16S rRNA Gene Sequence Comparison
The 16S rRNA gene sequence comparative analysis of strain GBPx2 T (1438 nt) showed the highest similarity to Cyclobacterium lianum HY9 T , Cyclobacterium jeungdonense HMD3055 T , Cyclobacterium xiamenense KD51 T , and Cyclobacterium halophilum IBRC-M 10761 T with 97.5%, 96.7%, 96.2%, and 96.2% sequence similarity, respectively, and values lower than 92.3% with species of other genera, such as Belliella or Fontibacter. These percentages were obtained by the EzBioCloud tool and indicate that strain GBPx2 T is a member of the genus Cyclobacterium.
The 16S rRNA gene sequence phylogenetic analysis using the maximum-likelihood algorithm showed the position of the novel strain within the genus Cyclobacterium ( Figure 1). The phylogenetic position was also confirmed in trees generated using the minimum-evolution and neighbour-joining algorithms. This phylogenetic tree shows that the genus Cyclobacterium is not monophyletic; the species of this genus are grouped into two clearly differentiated branches supported with 100% values of bootstrap. On the one hand, C. xiamenense KD51 T , C. halophilum GASx41 T , C. jeungdonense HMD3055 T , C. lianum HY9 T , and the new isolate GBPx2 T appear grouped, and on the other hand C. marinum LMG 13164 T , C. qasimii M12-11B T , C. caenipelagi HD-17 T , and C. amurskyense KMM 6143 T are clustered. To determine the relationship between these two clusters, a phylogenomic comparative analysis Filled circles indicate nodes that were also obtained in trees based on minimum-evolution and maximum-likelihood algorithms. Bootstrap values (for 1000 replicates) over 70% are shown at the nodes. The sequence accession numbers are shown in parenthesis. Bar, 2% estimated sequence divergence. The sequence of Parapedobacter soli DCY14 T (EF151805) was used as outgroup.
This phylogenetic tree shows that the genus Cyclobacterium is not monophyletic; the species of this genus are grouped into two clearly differentiated branches supported with 100% values of bootstrap. On the one hand, C. xiamenense KD51 T , C. halophilum GASx41 T , C. jeungdonense HMD3055 T , C. lianum HY9 T , and the new isolate GBPx2 T appear grouped, and on the other hand C. marinum LMG 13164 T , C. qasimii M12-11B T , C. caenipelagi HD-17 T , and C. amurskyense KMM 6143 T are clustered. To determine the relationship between these two clusters, a phylogenomic comparative analysis between them and also with members of other genera of the family Cyclobacteriaceae was performed.

Phylogenomic Comparative Analysis
We carried out phylogenomic comparative analysis and obtained the core-genome tree, based on 1309 single-copy translated genes of strain GBPx2 T , the genomes available for the type strains of Cyclobacterium species (Table 1), and the genomes of all type species of the genera of the family Cyclobacteriaceae available in databases ( Figure 2). This analysis shows that strain GBPx2 T constitutes a taxon which is sufficiently different from the other species of Cyclobacterium so as to be considered as a new species. Further, as occurred in the phylogenetic tree based on the 16S rRNA, the species of the genus Cyclobacterium appeared grouped in two different branches. Finally, this phylogenomic tree showed a close phylogenetic relationship between Cyclobacterium xiamenense CGMCC 1.12432 T and Cyclobacterium halophilum IBRC-M 10761 T , two species that were described almost simultaneously in 2014 [5,11], and so they were not considered for a comparison between them. Besides, the genomes of these two species are only now available for comparison and the current comparative data show in this study revealed that both are members of the same species. Phylogenomic tree based on the core orthologous translated genes of strain GBPx2 T , type species of Cyclobacterium, and type species of the genera of the family Cyclobacteriaceae obtained from the genomes available in databases, based on the maximum-likelihood algorithm. This tree was obtained after the alignment of 1309 shared orthologous single-copy translated genes of these genomes. Bootstrap values higher than 70% are indicated at branch-points. Bar, 0.1 substitutions per amino acid position.

in silico DNA-DNA Hybridization (GGDC), ANI, and AAI Values
In order to confirm that strain GBPx2 T was indeed a new taxon and the relationship between C. xiamenense and C. halophilum and the two clusters of the genus Cyclobacterium, average nucleotide identity (OrthoANI), average amino acid identity (AAI), and in silico DNA-DNA hybridization (GGDC) for the strain GBPx2 T and members of the family Cyclobacteriaceae were calculated.
GGDC percentages above or equal to 70% indicate that the strains can be assigned to the same species, and values under 70% indicate that the strains belong to different species [47][48][49]. GGDC values were equal or lower than 35% between strain GBPx2 T and species of the genus Cyclobacterium ( Table 2), proving that this strain constitutes a new species. In addition, the GGDC value of 81.6% which was determined between C. xiamenense CGMCC 1.12432 T and C. halophilum IBRC-M 10761 T ,  Figure 2. Phylogenomic tree based on the core orthologous translated genes of strain GBPx2 T , type species of Cyclobacterium, and type species of the genera of the family Cyclobacteriaceae obtained from the genomes available in databases, based on the maximum-likelihood algorithm. This tree was obtained after the alignment of 1309 shared orthologous single-copy translated genes of these genomes. Bootstrap values higher than 70% are indicated at branch-points. Bar, 0.1 substitutions per amino acid position.

in silico DNA-DNA Hybridization (GGDC), ANI, and AAI Values
In order to confirm that strain GBPx2 T was indeed a new taxon and the relationship between C. xiamenense and C. halophilum and the two clusters of the genus Cyclobacterium, average nucleotide identity (OrthoANI), average amino acid identity (AAI), and in silico DNA-DNA hybridization (GGDC) for the strain GBPx2 T and members of the family Cyclobacteriaceae were calculated. GGDC percentages above or equal to 70% indicate that the strains can be assigned to the same species, and values under 70% indicate that the strains belong to different species [47][48][49]. GGDC values were equal or lower than 35% between strain GBPx2 T and species of the genus Cyclobacterium (Table 2), proving that this strain constitutes a new species. In addition, the GGDC value of 81.6% which was determined between C. xiamenense CGMCC 1.12432 T and C. halophilum IBRC-M 10761 T , which was higher than the threshold percentage of 70% for species delineation, shows that both species belong to the same taxon [31,37]. With respect to the GGDC values between the other members of this family, all were lower than 70%, showing that all of them can be considered different taxa at the species level.
OrthoANI percentages calculated between strain GBPx2 T and species of the genus Cyclobacterium ranged from 71.8% to 79.2% (Table 2), lower than the threshold value for species delineation (95%-96%) [31,38,49,50], showing that strain GBPx2 T belongs to a different species. Values between 67.2% and 69.9% with the type species of the other genera of the family Cyclobacteriaceae were obtained. Further, the OrthoANI value of 97.8% between C. xiamenense CGMCC 1.12432 T and C. halophilum IBRC-M 10761 T showed again that both species constituted a single taxon.
An alternative to GGDC and ANI for more distantly related genomes is the AAI. In this case, to confirm that strain GBPx2 T and all species of Cyclobacterium were well assigned to this genus, the AAI percentages between them were calculated. The AAI values between each other were in the range of 72.2%-97.9% (Table 3). These values were above the threshold considered for species of the same genus (65%) [50][51][52], so we can affirm that all species belong to the genus Cyclobacterium. It is remarkable to highlight that AAI values between C. marinum DSM 745 T , C. qasimii M12-11B T , and C. amurskyense KMM 6143 T were higher (83.5% to 87.5%) as compared to with other species of Cyclobacterium (72.2%-73.5%), and lower than 66.2% with respect to species of the rest of genera of the family Cyclobacteriaceae. Similar results were observed in the other group of species of the genus Cyclobacterium that appear grouped in the 16S rRNA phylogenetic tree ( Figure 1) and also in the core-genome tree (Figure 2). This group included C. xiamenense KD51 T , C. halophilum IBRC-M 10761 T , C. jeungdonense HMD3055 T , C. lianum CGMCC 1.6102 T , and the new isolate GBPx2 T . AAI values between them ranged from 77.2% to 73.5%. With respect to the other species of the genus Cyclobacterium the AAI ranged between 72.2% and 73.5%, and values ranged between 61.6% to 68.1% with regard to the rest of the genera of the family Cyclobacteriaceae. All these data showed that the percentages for species of Cyclobacterium were always higher than 65% and thus they are members of the same genus, although there was a higher similarity between the respective members of the two phylogroups. Therefore, we conclude that the genus Cyclobacterium is monophyletic within the family, but once differentiated, it is divided into two clearly separated groups, as observed previously in both the 16S rRNA and core-genome trees (Figures 1 and 2). On the other hand, the value of 97.9% confirms that C. xiamenense and C. halophilum are members of the same taxon, as was described by Konstantinidis et al. [51] who established the threshold AAI range of 95%-100% for strains of the same species.

Chemotaxonomic Characterization
The cellular fatty acid profile of strain GBPx2 T was characterized by the presence of iso-C 15:0 (26.3%), summed feature 3 (C 16:1 ω7c and/or iso-C 15:0 2-OH; 23.9%), iso-C 17:0 3-OH (12.5%), anteiso-C 15:0 (12.1%), and iso-C 17:1 ω9c (9.6%) as the major fatty acids. The fatty acid profile of the strain was similar to that of the other type strains of species of the genus Cyclobacterium (Table 5). However, the percentages of these fatty acids were different from those obtained for other phylogenetically related species. Table 5. Cellular fatty acid composition (%) of strain GBPx2 T and related species of the genus Cyclobacterium. Strains: 1, GBPx2 T ; 2, Cyclobacterium lianum IBRC-M 10422 T ; 3, Cyclobacterium jeungdonense IBRC-M 11102 T . All strains were grown under the same conditions (Marine agar medium, 25 • C, and 2 days of incubation). Fatty acids accounting for < 1% of the total content in the strains are omitted. Summed feature 3 comprised iso-C 15:0 2-OH and/or C 16:1 ω7c and summed feature 4 comprised anteiso-C 17:1 B and/or iso-C 17:1 I. The polar lipids determined for strain GBPx2 T were phosphatidylethanolamine (PE), one unidentified phospholipid (PL), and four unidentified lipids (Supplementary Figure S1). The polar lipids pattern is similar to that of other species in the genus Cyclobacterium, except for Cyclobacterium halophilum, which has phosphatidylcholine as the major polar lipid [11].

Fatty
Menaquinone 7 (MK-7) was the only respiratory quinone present in strain GBPx2 T , which was typically found in members of the genus Cyclobacterium [6].

Conclusions
On the basis of the results of the taxogenomic and polyphasic taxonomic analysis, it is concluded that strain GBPx2 T should be considered as a novel species of the genus Cyclobacterium, for which the name Cyclobacterium plantarum sp. nov. is proposed. We enclose below the taxonomic description of this new species. As a result of the genomic analysis we can conclude that the genus Cyclobacterium is a coherent genus within the family Cyclobacteriaceae, and that all species currently described are members of the genus, even considering that they constitute two separate phylogenomic clusters. On the other hand, this genome-based study shows that Cyclobacterium xiamenense and Cyclobacterium halophilum constitute a single species, having priority the name Cyclobacterium xiamenense according to the International Code of Nomenclature of Prokaryotes [53]. Thus, Cyclobacterium halophilum  Cyclobacterium plantarum (plan.ta'rum. L. gen. pl. n. plantarum, of plants). Cells are Gram-stain-negative, non-motile, and strictly aerobic curved ring-like or horseshoe-shaped rods with sizes of 0.3-0.5 µm in width and the outer diameter of rings is 0.9-1.9 µm when grown on marine medium under optimal conditions. Colonies are small, circular, convex with entire margins, translucent, and smooth, with a light pink pigmentation. The strain is moderately halophilic and slightly alkaliphilic, growing over a wide range of temperatures from 4 to 40 • C (optimal growth at 25 • C), pH 6.5-9.0 (optimally at pH 8.0) and at 3%-10% (w/v) NaCl (with best growth at 5% [w/v] NaCl). It is positive for catalase and oxidase. Nitrate and nitrite reduction are positive and gas is formed from nitrate. Indole is not produced from tryptophan and H 2 S production is negative. Aesculin and Tween 40 are hydrolyzed, whereas casein, DNA, gelatin, starch, Tween 20, Tween 80, tyrosine, and urea are not. Acid is not produced from D-arabinose, cellobiose, D-galactose, D-glucose, lactose, maltose, melezitose, melibiose, sucrose, raffinose, D-rhamnose, D-ribose, trehalose, or D-xylose. Methyl red and Voges-Proskauer tests are negative. D-arabinose, D-galactose, D-glucose, D-fructose, D-maltose, D-mannitol, D-mannose, D-melibiose, myo-inositol, cellobiose, sucrose, D-xylose, ribose, L-alanine, L-ornithine, L-proline, L-serine and L-threonine are utilized as sole source of carbon and energy but L-glutamic acid is not. Polar lipids are phosphatidylethanolamine, one unidentified phospholipid, and four unidentified lipids. The only isoprenoid quinone is MK-7 and the predominant fatty acids are iso-C 15:0 , summed feature 3 (C 16:1 ω7c and/or iso-C 15:0 2-OH), iso-C 17:0 3-OH, and anteiso-C 15:0 . DNA G+C content of DNA is 43.0 mol% (genome).
The GenBank/EMBL/DDBJ accession number for the 16S rRNA gene sequence and complete genome sequence of the type strain are MG457806 and JAANYN00000000, respectively.
The description is that of Chen et al. [14], with the following modification: Growth occurs at 1.0%-10% (w/v) NaCl. Hydrolysis of Tweens 20 and 60 and aesculin is variable. The major cellular fatty acids are those reported previously in the species description plus iso-C 15:0 2-OH and anteiso-C 15:0 2-OH. Polar lipids are phosphatidylethanolamine, phosphatidylcholine, and several unidentified lipids. The DNA G+C range is 48.4-48.5 mol% (genome).
The type strain is KD51 T (= CGMCC 1.12432 T = KCTC 32253 T ), isolated from aggregates of Chlorella autotrophica in Xiamen, China. The DNA G+C content of the type strain is 48.5 mol% (calculated from the genome sequence).