Phylogenomic reclassification of Cellulophaga species to Paracellulophaga gen. nov. and description of Allocellulophaga tsushimaensis gen. nov., sp. nov., a novel bacterium from coastal seawater of Tsushima Island, Japan
Yu Nakajima, Shu-Kuan Wong, Yuki Muramatsu, Marie Johanna Cuadra, Kei Zenimoto, Keigi Ou, Hiroshi Xavier Chiura, Keiji Nakamura, Yasuhiro Gotoh, Tetsuya Hayashi, Yasuyoshi Nakagawa, Susumu Yoshizawa
TL;DR
A new marine bacterium, Allocellulophaga tsushimaensis, is described, and the genus Cellulophaga is reclassified into two new genera.
Contribution
The paper introduces Allocellulophaga tsushimaensis and reclassifies four Cellulophaga species into a new genus, Paracellulophaga.
Findings
Strain TSM-5T is distinct from Cellulophaga and represents a new genus, Allocellulophaga.
Phylogenomic analysis suggests splitting Cellulophaga into two genera, with four species moved to Paracellulophaga.
The new species Allocellulophaga tsushimaensis is characterized by unique physiological and biochemical traits.
Abstract
A novel Gram-staining-negative, rod-shaped, aerobic bacterial strain, designated as TSM-5 T, was isolated from surface seawater of a marine inlet. This strain grew well at 10–30 °C, pH 5.0–10.0 and in the presence of 0–4% (w/v) NaCl. Strain TSM-5 T was non-flagellated and positive for catalase and oxidase activities, and the major respiratory quinone was menaquinone-6 (MK-6). The major fatty acids present in the strain were iso-C15:1 G, iso-C15:0, iso-C15:0 3-OH, iso-C17:0 3-OH, and summed feature 3 (comprising C16:1ω7c and/or C16:1ω6c). The major polar lipid of the strain was phosphatidylethanolamine. Phylogenetic analysis based on 16S rRNA gene sequences showed that strain TSM-5 T shared 92.6 to 94.0% sequence similarity with valid species of the genus Cellulophaga, and the phylogenetic branch of this strain divided the genus into two groups. Based on the average nucleotide identity…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4- —https://doi.org/10.13039/501100001691Japan Society for the Promotion of Science
- —The University of Tokyo
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGenomics and Phylogenetic Studies · Aquaculture disease management and microbiota · Bacteriophages and microbial interactions
Introduction
The family Flavobacteriaceae (Bernardet et al. 1996, 2002), the largest family within the phylum Bacteroidota, currently comprises 155 genera, as referenced in the LPSN (Parte 2018). Species within this family are characterised as non-spore-forming, rod-shaped bacteria, with most species being aerobic. The type genus, Flavobacterium, was named for its production of yellow-orange pigments, primarily flexirubin (Reichenbach et al. 1980) and carotenoids, with the latter often being the most abundant pigments in marine lineages. Flexirubin-type pigments have been used as chemotaxonomic markers for the phylum Bacteroidota, and a biosynthetic gene cluster (BGC) for flexirubin production containing the darA and darB genes was identified in Flavobacterium johnsoniae (McBride et al. 2009). However, within the family Flavobacteriaceae, not all members encode flexirubin-producing BGCs (Silva et al. 2023). In the genus Cellulophaga, Cellulophaga. lytica was described as the type species (Johansen et al. 1999), and seven species are currently valid. Among these, most of the type species do not exhibit flexirubin-positive reaction. In closely related genera, Sediminicola and Arenibacter are also not considered to produce it (Khan et al. 2006).
Here, we present the results of a polyphasic study describing a new strain isolated from the surface seawater that is closely related to genus Cellulophaga but shows a presence of flexirubin. Based on phylogenetic analysis using genomic information, we propose this as a novel genus, and also propose reclassifying the valid species of Cellulophaga into two genera.
Materials and methods
Isolation and culture conditions
Surface seawater samples were collected using a fishing boat from Hitakatsu Port (34° 39.0’ N 129° 29.2’ E) in northwestern Tsushima, Japan. A 100 mL aliquot of seawater was inoculated on 1/10 strength ZoBell medium [0.5 g peptone, 0.1 g yeast extract, 15 g agar in 1 L of 80% aged natural seawater (80% seawater + 20% distilled water)] and incubated aerobically at 20 °C for 5 days. An orange colony designated as strain TSM-5^ T^, was isolated and purified through repeated streaking.
Genome sequence and phylogenetic analysis
The genomic DNA of strain TSM-5^ T^ was extracted using the Wizard DNA purification kit (Promega, Madison, WI). A genomic DNA library for short-read sequencing was prepared using the QIAseq FX DNA Library Kit (Qiagen), and sequenced on the MiSeq platform using the MiSeq Reagent kit version 3 (Illumina) to generate paired-end sequence reads (301 bp × 2). In addition, long-read sequences were generated using a Nanopore MinION flowcell R9.4.1 (Oxford Nanopore Technologies (ONT)) with the Rapid Barcoding Kit (ONT) and base-called using Albacore v2.3.1 (ONT), and the short and long read sequences were hybrid-assembled using Unicycler v. 0.4.6 (Wick et al. 2017). The genome sequence was annotated using DFAST v. 1.2.18 (Tanizawa et al. 2018), and secondary metabolism biosynthetic gene cluster (BGC) was predicted with antiSMASH v.8 (Blin et al. 2025). The genome completeness and contamination ratio were evaluated by CheckM2 version 1.1.0 (Chklovski et al. 2023). The 16S rRNA gene was extracted from the genome of strain TSM-5^ T^ (accession number: LC864043), and a BLASTn search against the EzBioCloud and NCBI database was performed. For phylogenetic analysis, the 16S rRNA sequences of closely related type species belonging to family Flavobacteriaceae were collected and aligned using MAFFT v7.402 (Katoh and Standley 2013) with default settings. Phylogenetic and molecular evolutionary analyses were conducted using IQ-TREE v.2.2.6 (Minh et al. 2020) with “-m GTR + F + I –mtree -bb 1000” and with “-m GTR + F + I + G –mtree -bb 1000” options. Genomic classification with GTDB maker genes was performed using GTDB-Tk v. 2.4.0 (Chaumeil et al. 2022) with the release 226 dataset. Additionally, to generate phylogenetic tree of ribosomal protein, 45 ribosomal protein genes were extracted using proteinortho6 (Klemm et al. 2023). These concatenated sequences were aligned using MAFFT v7.402, and phylogenetic analysis was performed using IQ-TREE with “-m LG + F + G + C60 –mtree -bb 1000”. These phylogenetic trees were visualized with iTOL (Letunic and Bork 2007).
To evaluate the threshold of whole genome relatedness indices for species and genus classification, Average Nucleotide Identity (ANI), and Average Amino acid Identity (AAI) values were calculated using FastANI v.1.33 (Jain et al. 2018), FastAAI v.0.1.17 (Gerhardt et al. 2025) among strain TSM-5^ T^ and closely related genera such as Cellulophaga, Sediminicola, Arenibacter, Maribacter used in phylogenetic trees of 16S rRNA and ribosomal protein. Furthermore, to augment the comparison of overall genome relatedness indices, a digital DNA-DNA hybridization (dDDH) value was calculated with Genome-to-Genome Distance Calculator (GGDC) 3.0 by DSMZ (Meier-Kolthoff et al. 2022) among strain TSM-5^ T^, genera Cellulophaga and Sediminicola. The valid species genomes of family Flavobacteriaceae used in ANI/AAI/dDDH and invalid isolates/MAGs currently included in genus Cellulophaga are listed in Table S1. The ANI, AAI, and dDDH value matrix was then visualized by python scientific computing library, matplotlib v.3.7.1(Hunter 2007), scipy1.13.1(Virtanen et al. 2020), and seaborn 0.13.2 (https://seaborn.pydata.org/).
Physiology and chemotaxonomy
The optimal temperature range for growth was tested on 1/2 strength Marine Agar (1/2 MA) [18.7 g Difco Marine Broth (MB) 2216, 15 g agar supplemented with 1.0% NaCl (w/v) in 1L milli-Q water] at 4, 10, 15, 20, 25, 30, 35 and 40 °C. For the salt tolerance test, cells were incubated in different concentrations of NaCl ranging from 0 to 10% (at 1% intervals. w/v) and for pH tolerance (pH 4–10 at 1 pH unit intervals). Salt and pH tolerance were tested using modified 1/2 MA, which contained (per liter distilled water) 2.5 g peptone (Bacto Peptone, BD), 0.5 g yeast extract (Bacto yeast extract, BD), 0.05 g ferric citrate, 0.9 g CaCl_2_ 2H_2_O, 3.0 g MgCl_2_ 6H_2_O (adjusted to pH 7.5 for NaCl tolerance test) and 1% NaCl. The pH values of the modified 1/2 MA were adjusted with either HCl or NaOH before and after sterilizing the medium. Cell morphology and size were examined using cells grown on 1/2 strength Marine Broth Difco 2216 supplemented with 1% NaCl for 2 days at 25 °C. The cells were negatively stained with 2% uranyl acetate for 30 s and observed at 75,000 × magnification at an acceleration voltage of 80 kV using a JEM-1400EX transmission electron microscope (JEOL Inc. Japan) according to Børsheim et al. 1990 (Børsheim et al. 1990) with modifications(Chiura 1997).
Enzyme activities and other biochemical properties were determined using API 20E, API 20NE, API 50CH and API ZYM (bioMérieux). All suspension media in API test strips were supplemented with 3% (w/v) NaCl (final concentration). API 20E, API 20NE, API 50CH were recorded after 5 days and those of API ZYM were recorded after 2 days. Degradation tests for DNA [using DNase agar (Oxoid) and flooding the plates with 1 M HCl], starch, agar, and carboxymethylcellulose were performed by Bowman 2000 and evaluated after 4 days. The presence of flexirubin-type pigments was tested using 500 µL of 5 N KOH solution to the TSM-5^ T^ and C. lytica cells (Fautz and Reichenbach 1980), and further testing with 525 µL of 6 N HCl was conducted to determine whether the color change was reversed. Both strains cells were incubated in 1/2 MB aerobically at 25 °C for three days, and washed twice with 0.1 M NaCl after centrifugation (5,800 × g for 10 min at 4 ºC; MX-305). Catalase activity was determined by observing bubble formation in 3% hydrogen peroxide solution. Oxidase activity was tested using cytochrome oxidase paper (Nissui Pharmaceutical Co.). Growth under anaerobic condition was performed after incubation with AnaeroPack (Mitsubishi Gas Chemical Co.) on 1/2 MA for 2 weeks. Gram staining was performed according to instructions provided in the Gram Stain Kit (BD). Gliding motility was also determined using cells plated onto modified 1/2 soft Marine Agar (0.3–1.0% agar content) (Harshey 2003; Nan and Zusman 2011). Chemotaxonomic studies, except for fatty acid analysis, were conducted by cultivating strain TSM-5^ T^ in liquid MB medium at 25 °C for 3 days. Amino acids and their isomers in cell-wall hydrolysates, isoprenoid quinones.
Polar lipids were extracted using the method described by (Minnikin et al. 1979) and analysed by TLC with chloroform/methanol/water (65: 25: 4, by vol.) in the first direction and chloroform/acetic acid/methanol/water (80: 18: 12: 5, by vol.) in the second direction. Polar lipids were visualized by spraying the TLC plate with 5% molybdophosphoric acid. Dittmer and Lester reagent (phosphorus) (Dittmer and Lester 1964), ninhydrin (amino group), Schiff’s reagent (glycol), and anisaldehyde (sugar) were also used as specific spray reagents for polar lipids. For fatty acid methyl ester analysis, strain TSM-5^ T^ was cultivated on MB medium for 3 days at 25 °C. Cellular fatty acid methyl esters were analysed by Gas Chromatography (GC) (model 6890 N; Agilent Technologies) according to the standard protocol of the MIDI Sherlock Microbial Identification System with the TSBA6 database (version 6.2).
Results and discussion
16S rRNA gene phylogeny and phylogenomic analysis
The sequence similarity of 16S rRNA gene between strain TSM-5^ T^ and type strains of recognized species within the genus Cellulophaga ranged from 92.6% to 94.0%. The highest similarity was observed with Cellulophaga tyrosinoxydans (93.96%), while the similarity between TSM-5^ T^ and Cellulophaga lytica NBRC 14961^ T^, the type species of the genus Cellulophaga, was 92.65%. According to the recently proposed rank boundary analyses (Hackmann 2025), the sequence identity of strain TSM-5^ T^ suggests that it belongs to either the genus Cellulophaga or the family Flavobacteriaceae. Notably, strain TSM-5^ T^ formed a distinct branch that diverged from the genus Cellulophaga in the 16S rRNA gene phylogenetic tree (Fig. 1). While the tree topology of TSM-5^ T^, Cellulophaga, and Sediminicola varied depending on the evolutionary model used, genus Cellulophaga was consistently divided into two clades: one containing C. lytica and the other containing C. tyrosinoxydans.Fig. 1. Phylogenetic tree based on nearly complete 16S rRNA gene sequences showing relationships between strain TSM-5^ T^ and related members of the family Flavobacteriaceae with GTR + F + I (a) or GTR + F + I + G evolutionary optimization models. The tree was described with IQ-TREE (Minh et al. 2020). Bootstrap values based on 1,000 replications. Only a bootstrap value of > 80% was shown. The bar shows 0.1 substitutions per nucleotide position. Mesonia algae KCTC 12089^ T^ (AF53683) was used as an outgroup
The complete genome sequence of strain TSM-5^ T^ consists of a single scaffold (total length: 4,450,019 bp), with a G + C content of 32.0%. The completeness and contamination ratio were 99.99% and 0.04%, respectively. The genome contained 3,832 coding sequences, nine rRNA genes (three copies each of 5S, 16S, and 23S rRNA genes, which matched 100% identity), 49 tRNAs as annotated by DFAST. A phylogenomic tree constructed using GTDB-Tk also showed that the genus Cellulophaga was divided into two major clades, with strain TSM-5^ T^ occupying an intermediate branch (Fig. 2a). In this tree, the sister clade to genus Cellulophaga was the genus Arenibacter, while the branch consisting of genus Sediminicola (GTDB taxonomy: g_YIK13) was a further outgroup. Similarly, a phylogenetic tree of ribosomal proteins showed that C. lytica clade and the C. tyrosinoxydans clade were widely separated, with several genera positioned between them (Fig. 2b).Fig. 2. Phylogenetic tree based on the genomic sequences showing relationships among genus Cellulophaga, strain TSM-5^ T^ and closely related genera. (a) The tree was created with GTDB-Tk v. 2.4.0 (Chaumeil et al. 2022) using R226 dataset. The underline of each name indicates a valid name on ICNP. (b) Phylogenetic tree of concatenated 45 ribosomal proteins from strain TSM-5^ T^ and related valid species. Only a bootstrap value of > 80% was shown. The bar shows 0.05 (a) and 0.2 (b) substitutions per nucleotide position, respectively. Mesonia algae KCTC 12089^ T^ was used as an outgroup in the phylogenetic tree of ribosomal protein
ANI and AAI comparisons showed that strain TSM-5^ T^ had values ranging from 74.1–76.2% (ANI) and 62.0–73.2% (AAI), respectively (Fig. 3). Based on the thresholds and distributions of ANI and AAI values for the delineation of intraspecies or intragenus, TSM-5^ T^ may represent a novel species that could fall within the genus Cellulophaga(Kim et al. 2014; Konstantinidis et al. 2017). However, in closely related genera, genus-level boundaries were consistently supported by pairwise ANI/AAI values higher than those of surrounding lineages. For example, in the genus Arenibacter, the intergenus values were 75.3% (ANI) or 75.9% (AAI), whereas the highest intragenus values were 75.2% (ANI) or 73.0% (AAI). Similarly, for Sediminicola species (only two valid species genomes available), ANI and AAI with other genera were 74.0–75.2% (ANI) and 62.4–73.0% (AAI), respectively.Fig. 3ANI and AAI matrices of strain TSM-5 ^T^ and related Flavobacteriaceae species. Blue solid-line boxes indicate the ANI/AAI values within the same genus. Blue dashed-line boxes indicate current genus Cellulophaga
dDDH identity was also compared with Sediminicola species, which is the closest relative to genus Cellulophaga in the 16S rRNA phylogenetic tree as well as ANI and AAI identities. As the result, strain TSM-5^ T^ showed dDDH identities of only 15.1–16.1% with other Cellulophaga and Sediminicola species. Considering the dDDH values within the C. lytica clade, the C. tyrosinoxydans clade, and the genus Sediminicola, the possibility of strain TSM-5^ T^ being an independent lineage was supported (Fig. S1).
Based on these phylogenetic and phylogenomic results, physiological tests were primarily conducted using type species of Cellulophaga (C. lytica) and Sediminicola (S. luteus). Additional physiological data were also referenced from C. tyrosinoxydans, and Cellulophaga baltica (a type species first described within the C. tyrosinoxydans clade.).
Phenotypic characterization
Colonies of strain TSM-5^ T^ grown on 1/2 MA after 3 to 4 days at 25 °C were orange. The cells were non-flagellated rods (0.55 ± 0.17 µm wide and 3.26 ± 2.20 µm long; Supplementary Fig.S2), Gram-staining-negative, and gliding motility was not observed on 0.3–1.0% soft agar. Strain TSM-5^ T^ exhibited KOH positive reaction characterized by a change in the yellow bacterial cells to a brown color for detecting flexirubin-type pigment. The change was reversed to yellow by adding HCl (Fig. S3). Although flexirubin-type pigment was tried to be extracted in acetone or acetone: methanol (7:3) by modified methods in Siddaramappa et al. 2019 from the strain TSM-5^ T^ cells, no extract exhibiting an absorption maximum at approximately 450 nm was obtained. Anaerobic growth was not observed. Strain TSM-5^ T^ showed positive catalase activity, and the detailed results of the morphological, physiological, and biochemical tests are shown in the species description and in Table 1. Table 1. Differential phenotypic characteristics of strain. 1: TSM-5^ T^ (this study), strain 2: C. lytica NBRC 14961^ T^ (this study), strain 3: C. tyrosinoxydans [Paracellulophaga] tyrosinoxydans DSM 21164^ T^ (Kahng et al. 2009), strain 4: C. baltica [Paracellulophaga] NN015840 ^T^ (Johansen et al. 1999; Nedashkovskaya et al. 2004) and strain 5: S. luteus NBRC 100966^ T^ (this study). +, Positive reaction; -, Negative reaction; w, weak positive reaction. GL, Glycolipid; AL, Aminolipid; PE, phosphatidylethanolamine; L, undefined lipid. All strains are positive for assimilation of alkaline phosphatase, esterase (C4), esterase lipase (C8), lipase (C14), leucine arylamidase, valine arylamidase, cystine arylamidase, trypsin, α-chymotrypsin, acid phosphate, β-glucosidase, N-acetyl-β-glucosaminidase (API ZYM); NO_3_ reduction, arginine dihydrolase, urease, esculinase (API 20NE); arabinose (API 20E); utilization of α-Methyl-D-Mannoside, α-Methyl-D-Glucoside, N-Acetyl-Glucosamine, amygdalin, arbutin, esculin, salicin, cellobiose, maltose, lactose, melibiose, L-fucose (API 50CH). All strains were negative for β-glucuronidase, α-fucosidase (API ZYM); L-tryptophane, L-arabinose, mannose, mannitol, N-acetyl-glucosamine, maltose, potassium gluconate, capric acid, adipic acid, malic acid, trisodium citrate, phenylacetic acid, cytochrome oxidase, fermentation of glucose, gelatinase, assimilation of glucose, (API 20NE); arginine dihydrolase, lysine decarboxylase, ornithine decarboxylase, citrate utilization, H_2_S production, urease, tryptophane deaminase, indole production, mannitol, inositol, sorbitol, rhamnose, sucrose, melibiose, amygdalin (API 20E); utilization of glycerol, D-arabinose, adonit, 2-keto-gluconate (API 50CH)Characteristics12345pH range for growth (optimum)4–10 (7–8)5–10 (7–9)6.5–9 (7–8.5)7–86.0–8.5 (7.0–7.5)NaCl requirement (%) (optimum)0–4 (2)0–5 (0–2)1–7 (2–4)1–5 (2)0.5–10 (3.5)**Temperature range for growth (optimum)10–30 (15–30)5–35 (10–30)15–35 (25–30)2–30 (26–30)10–30 (20)**Flexirubin pigment + ––––Hydrolysis of Starch– + * + + + ** Agar– + – + + ** Cellulose + N.D + + – Catalase + –– + + ** Oxidase + + + –w DNase + N.DN.D + –Enzyme activity (API ZYM) α-galactosidase– + –N.DW α-glucosidase– + –N.D + β-galactosidase– + –N.D–API 20NE 4-nitrophenyl-β-galactosidase– + –N.D–API 20E β-galactosidase, Glucose– + N.DN.D– Acetoin production– + N.DN.D + Gelatinase + –N.DN.D–Acid production from (API 50CH) Erythrit,β-Methyl-D-Xyloside– + N.DN.D– L-Arabinose, L-Xylose– + N.D–– D-Xylose– + N.D–W Fructose, Dulcitol, Sorbose– + N.DN.D + Galactose, Glucose, Rhamnose– + N.D– + Mannose– + N.DN.D + Inositol, Sorbitolw + N.DN.D + Mannitolw + N.D + + Polar lipidL1 ~ 12, PE, GL1 ~ 2L1 ~ 2, AL, APN.DN.DN.D DNA G + C content (mol%)32.0§33.033.533.038.0 FlagellaNoNoNoNoNo QuinoneMK-6MK-6MK-6MK-6MK-6 Cell shapeRodRodRodRodRod^*^ Data from Johansen et al. (1999); (Park et al. 2012)^^ Data from Khan et al. (2006); (Khan et al. 2006; Hwang et al. 2015)^§^ Data from genome informationN.D.: Not determined
Strain TSM-5^ T^ can be differentiated from C. lytica NBRC 14961^ T^ by various phenotypic characteristics such as the inability to hydrolyze starch and agar, and the presence of enzyme activity for α-galactosidase, α-glucosidase, and β-galactosidase. In addition, TSM-5^ T^ was unable to produce acid from erythritol, L-Arabinose, L-Xylose, β Methyl-D-Xyloside, D-Xylose, galactose, glucose, fructose, mannose, sorbose, rhamnose, and dulcitol. On the other hand, TSM-5^ T^ exhibited positive gelatinase, DNase, and catalase activities, unlike C. lytica.
The major fatty acid of strain TSM-5^ T^ consisted of iso-C_15:0_, iso-C_15:1_ G, iso-C_15:0_ 3**-OH, iso-C_17:0_ 3-OH, and summed feature 3 (comprising C_16: 1_ω7c and/or C_16: 1_ω6c) (Table 2). In contrast, C. lytica possessed these fatty acids in addition to iso-C_16:0_ 3-OH whereas the major cellular fatty acids for C. tyrosinoxydans are C_15:0_, iso-C_15:0_, iso-C_15: 0_ G, iso-C_15: 0_ 3-OH, C_15:1_ω6c, iso-C_16: 0_ 3-OH, iso-C_17: 0_ 3-OH, and summed feature 3 (comprising C_16: 1_ω7c and/or iso-C_15: 0_ 2-OH)(Kahng et al. 2009). In addition, S. luteus differed from other species mainly in that it contained anteiso-C_15:0_, iso-C_15:1_, iso-C_16:0_, and iso-C_17:1_ ω9c (Khan et al. 2006). Thus, the composition of major and minor fatty acids of the three reference strains differed from those of strain TSM-5^ T^. The predominant isoprenoid quinone of strain TSM-5^ T^ was menaquinone-6 (MK-6), similar to C. lytica, C. tyrosinoxydans, and S. luteus. The polar lipid profile of strain TSM-5^ T^ consisted of two glycolipids (GL), one phosphatidylethanolamine (PE), and 12 undefined lipids (L1-12) (Supplementary Fig. S4). PE and GL were not part of the main polar lipids of C. lytica. Instead, the polar lipid profiles of C. lytica contained two unidentified lipids, one unidentified aminolipid and one unidentified aminophospholipid. The polar lipid profile of strain TSM-5^ T^ was therefore distinct from that of C. lytica (Table 1). Table 2. Cellular fatty acid compositions (in percentage) of strain. 1: TSM-5^ T^, 2: C. lytica NBRC 14961^ T^, strain 3: C. tyrosinoxydans DSM 21164^ T^ (Kahng et al. 2009), strain 4: C. baltica NN015840 ^T^ (Bowman 2000) and strain 5: S. luteus NBRC 100966^ T^ (Khan et al. 2006; Hwang et al. 2015). -, Not detected or less than 1%. Major fatty acid (> 5%) are shown in bold characterFatty acid12345C_10:0_––2.2––C_14:0_1.2––––iso-C_14:0_––1.0–1.0C_15:0_––15.212.2–iso-C_15:0_10.120.817.013.612.0anteiso-C_15:0_-3.8–2.611.0iso-C_15:1_ G14.915.09.5****––anteiso-C_15:1_ A1.3––––iso-C_15:1_––––11.0C_15:0_ 2-OHC_15:0_ 3-OHC_15:1_ 3-OHiso-C_15:0_ 3-OHC_15:1_ ω6ciso- C_15:1_ ω10cC_16:0_iso-C_16:0_C_16:0_ 3-OHiso-C_16:0_ 3-OHiso-C_16:0_ HC_16:1_ ω7ciso-C_16:1_iso- C_16:1_ ω6cC_17:0_ 3-OHiso-C_17:0_ 3-OHiso-C_17:1_ ω9ciso-C_17:1_ ω7cC_17:1_ ω6c1.43.2–5.13.3–2.5–4.33.7––––1.219.8**––1.7–––12.7––1.9–2.47.5–––––18.8–––**–1.55.3–5.0–1.11.82.29.01.6––––8.91.6–1.2–––6.12.39.83.4–1.17.3–16.9–1.114.0–5.2––––4.0–––5.05.0****–––3.0––12.09.0****––C_17:1_ ω8c2.0––––Summed feature 2^#^1.0––––**Summed feature 3^^**17.0**7.311.2–14.0Summed feature 9^†^4.03.1–**––^#^Summed feature 2 contained unknown fatty acid (possible to 12:0 aldehyde)^^Summed feature 3 contained C_16:1_ ω7c and/or C_16:1_ ω6c^†^Summed feature 9 contained C_17:1_ iso ω9c and/or C_16:0_ 10-methyl
Genomic features for physiological characteristics
According to the results of phenotypic tests, strain TSM-5^ T^ may contain flexirubin pigment. Therefore, we investigated the presence or absence of the flexirubin synthesis gene operon in Cellulophaga genomes, including species with validly published name, species without any validly or effectively published scientific name and MAGs. Genes were considered present if their BLASTP hits against the reference genomes of Flavobacterium johnsoniae UW101 (NZ_CP140160.1) and Chitinophaga pinensis DSM 2588^ T^ (NC_013132.1) had e-values below the threshold of 1e-3. The flexirubin biosynthesis pathway was analysed using gene annotations and operon configurations based on F. johnsoniae and C. pinensis as described by (Schöner et al. 2014). Strain TSM-5^ T^ possessed 21 out of the 25 related genes (Table 3). Interestingly, several Cellulophaga species with validly published names that do not produce flexirubin still retained many of the flexirubin biosynthesis genes. However, because C. baltica DSM 24729^ T^ possessed the same gene set as strain TSM-5^ T^, it was not possible to conclusively differentiate flexirubin pigment production based solely on homology to genes from F. johnsoniae and C. pinensis. This suggests two possibilities: (1) genes unique to strain TSM-5^ T^ may be essential for pigment production. In the family Flavobacteriaceae, several genus-specific gene clusters putatively involved in flexirubin biosynthesis have been suggested (Silva et al. 2023); or (2) some Cellulophaga species may indeed be capable of flexirubin pigment production. Table 3. Presence or absence of genes in the biosynthesis gene cluster of flexirubins. Flavobacterium johnsoniae UW101 and Chitinophaga pinensis DSM 2588^ T^ are reference genomes; each character and number represents same functional gene with Schöner et al. 2014(Schöner et al. 2014). A (9): Histidine ammonia-lyase; B (7): Short-chain dehydrogenase/reductase; C (6): Beta-ketoacyl synthase; E (4): Lipid A biosynthesis acyltransferase; I (93): beta-ketoacyl-[acyl-carrier-protein] synthase family protein; J: methyltransferase; K (10): NAD(P)/FAD-dependent oxidoreductase; N (88): beta-ketoacyl-[acyl-carrier-protein] synthase family protein; O (87): beta-ketoacyl synthase N-terminal-like domain-containing protein; P (86): polysaccharide deacetylase; Q (85): outer membrane lipoprotein carrier protein LolA; T (79): DUF2062 domain-containing protein; U (78): trifunctional MMPL family transporter/lysophospholipid acyltransferase/class I SAM-dependent methyltransferase; V (77): phytoene desaturase; W (76): Ntn hydrolase; Y (75): phenylacetyl-CoA ligase; A (darA): 3-oxoacyl-[acyl-carrier-protein] reductase; B (darB): beta-ketoacyl-ACP synthase III; G (98): hypothetical protein; H (97): BtrH N-terminal domain-containing protein; I (96): ABC transporter ATP-binding protein; J (95): ABC transporter permease; C: Acyl-carrier protein; D (88): beta-ketoacyl-[acyl-carrier-protein] synthase; E: M14 family metallopeptidase. The character “r” shown in TSM-5^ T^ and Cellulophaga genomes column indicate genes that matched with the reference genome. The character “c” indicates that while these genes did not match with the genes in the two reference genomes, but matched with the genes which were present in strain TSM-5^ T^ as same annotation genes. In the column of “Flexirubin”, flexirubin pigment-producing bacteria indicate “Y”, and “N” indicates the bacteria cannot produce it. “N.D.”: Not determinedValid speciesFlexirubinPosition of flexirubin pigment synthesis pathway genesFlavobacterium johnsoniaeY97649310888786857978777675darAdarB9897969588Chitinophaga pinensisYABCEIJKNOPQTUVWYABGHIJCDETSM-5^ T^YrrrcrrrrrccrrrrRRcrrrCellulophaga algicolaNrrrcrrrrrcrrrrRRrrrrrCellulophaga tyrosinoxydansNrrrcrrrrcrrrRRrcrrCellulophaga lyticaNrrrcrrrrcrrrrrrrcrrCellulophaga fucicolaNrrrcrrrrcrrrrrrrrrrCellulophaga balticaNrrrcrrrrrcrrrrrrrcrrNot valid isolatedCellulophaga sp. L1A9N.DrrrcrrrrrcrrrrrrrrrrrCellulophaga sp. HaHa_2_95N.DrrrcrrrrccrrrrrrrcrrCellulophaga sp. HaHaR_3_176N.DrrrcrrrrcrrrrrrrcrrCellulophaga baltica B2M06N.DrrcrrrrcrrrrrrrrrrCellulophaga sp. F20128N.DrrrcrrrrcrrrrrrcrrCellulophaga sp. 20_2_10N.DrrrcrrrrcrrrrrrcrrCellulophaga sp. Hel_1_12N.DrrrcrrrrcrrrrrrrrcrrCellulophaga omnivescoria W5CN.DrrrcrrrrcrrrrrrcrrCellulophaga sp. RHA19N.DrrrcrrrrcrrrrrrcrrMAGCellulophaga sp947496695N.DrrcrrrrcrrrrrrcrrCellulophaga sp030731855N.DrrrrrrrcrrrrrrrcrrCellulophaga sp947491225N.DrrrrrrrcrrrrrrrrcrrCellulophaga sp963974535N.DrrrrrrrcrrrrrrrrrrCellulophaga sp036478775N.DrrrcrrrrrcrrrrrrrrcrrCellulophaga sp947492955N.DrrcrrrrrcrrrrrrrcrrCellulophaga sp041738035N.Drrrcrrrrcrrrrrrrcrr
Secondary metabolism biosynthetic gene cluster (BGC) prediction using antiSMASH revealed that strain TSM-5^ T^ possessed terpene, T3PKS (Type III polyketide synthase) and NRPS (non-ribosomal peptide synthase) BGCs (Fig. 4a). Because flexirubin pigments are synthesized via the polyketide biosynthesis pathway and incorporate acyl groups derived from fatty acids, the presence of T3PKS domains in strain TSM-5^ T^ may be related to its positive flexirubin reaction.Fig. 4. Annotated biosynthetic gene cluster (BGC) on current genus Cellulophaga. (A) BGC of strain TSM-5^ T^ genomic flanking region detected by antiSMASH v.8 (Blin et al. 2025). (B) Annotated domains of strain TSM-5^ T^ and current genus Cellulophaga species genomes. T3PKS: Type III polyketide synthase (PKS); NRPS: non-ribosomal peptide synthase; RiPP-like: unspecified ribosomally synthesised and post-translationally modified peptide product (RiPP); NI-siderophore: NRPS-independent, IucA/IucC-like siderophore. Bashed-lines indicate boundaries of the three different proposed genera
Across the 22 Cellulophaga genomes analysed, several biosynthetic domains were preferentially located on one side of the phylogenomic tree based on GTDB R226, including in strain TSM-5^ T^. The clade containing C. lytica tends to have fewer BGC domains, whereas the clade containing C. baltica and C. tyrosinoxydans had more.
Taxonomic conclusion
Overall, based on phylogenetic analysis and whole genome relatedness indices, the current genus Cellulophaga needs to be reclassified into two distinct genera. One clade includes C. lytica, and the other includes C. baltica. The latter tended to have more BGCs than the former. Furthermore, the novel strain TSM-5^ T^ showed ANI/AAI/dDDH values of approximately 75–76%, 69–73%, and 15–16% with both clades, respectively, suggesting that this strain belongs to another genus based on intragenus thresholds within the family Flavobacteriaceae. In terms of genotypic, chemotaxonomic and phenotypic characterizations, strain TSM-5^ T^ exhibited similar characteristics to those of the genus Cellulophaga, especially C. tyrosinoxydans, in G + C content, quinone type, fatty acid composition, and several phenotypic reactions. However, this strain can be differentiated from other members of the genus Cellulophaga by its positive flexirubin pigment reaction, its distinct branching in the phylogenomic tree, and unique BGC profile. Strain TSM-5^ T^ is therefore considered to represent the type strain of a novel genus with the newly proposed name Allocellulophaga.
Emended description of Paracellulophaga Johansen (Johansen et al. 1999).
Paracellulophaga (Pa.ra.cel.lu.lo'pha.ga. Gr. pref. para, beside, alongside; N.L. neut. n. cellulosum, cellulose; Gr. inf. v. phageîn, to eat; N.L. fem. n. Paracellulophaga, cellulose eater alongside, referring to the endocellulase enzyme activity and close phylogenetic relationship to the genus Cellulophaga.)
Description of the genus Paracellulophaga is partially based on that given by(Johansen et al. 1999; Bowman 2000; Nedashkovskaya et al. 2004; Kahng et al. 2009).
Cells are rod-shaped, Gram-staining-negative, catalase-positive, and non-motile. Growth occurs at pH 7–8 and 2–30ºC under aerobic condition. The predominant respiratory quinone is MK-6. The major fatty acids are C_15:0_, iso-C_15:0_, iso-C_15:0_ 3-OH, iso- C_15:1_ ω10c, iso-C_16:0_ 3-OH, C_16:1_ ω7**c, iso-C_17:0_ 3-OH, and iso-C_17:1_ ω7c. The DNA G + C content is 33.0%. The type species is Paracellulophaga baltica.
Description of Allocellulophaga gen. nov.
Allocellulophaga (Al.lo.cel.lu.lo'pha.ga. Gr. pref. allos, other, different; N.L. neut. n. cellulosum, cellulose; Gr. inf. v. phageîn, to eat; N.L. fem. n. Allocellulophaga, another cellulose eater, referring to the cellulase activity and phylogenetic distinctiveness from the genus Cellulophaga.)
Cells are rod-shaped, Gram-staining-negative, catalase- and oxidase-positive, and non-motile. Growth occurs at pH 4–10, with 0–4% NaCl, and at 10–30ºC under aerobic condition. The predominant respiratory quinone is MK-6. The major fatty acids are iso-C_15:0_, iso-C_15:1_ G, iso-C_15:0_ 3-OH, iso-C_17:0_ 3-OH, and summed feature 3 (C_16:1_ω7c and/or C_16:1_ω6c). The DNA G + C content is 32.0%. The type species is Allocellulophaga tsushimaensis.
Description of Allocellulophaga tsushimaensis sp. nov
Allocellulophaga tsushimaensis (tsu.shi.ma.en’sis. N.L. fem. adj. tsushimaensis referring to Tsushima, an island in Nagasaki prefecture, Japan, where the species was first isolated).
This species exhibits the following properties in addition to those given in the genus description. Cells are approximately 0.55 µm wide and 3.26 µm long. Optimal growth occurs at 15–30 ℃ (growth range of 10–30 ℃), pH 7–8 (growth range of pH 4–10) and 2% NaCl (growth range 0–4%, w/v). Catalase- and oxidase-positive. Flexirubin reaction is positive. DNA and cellulose are hydrolyzed, while starch and agar are not. In the API ZYM gallery, positive results were observed for alkaline phosphatase, esterase, esterase lipase, valine arylamidase, trypsin, α-chymotrypsin, acid phosphate, naphthol-AS-BI-Phosphohydrolase, β-glucosidase, N-acetyl-β-glucosaminidase and α-mannosidase. Positive for gelatinase and arabinose utilization in the API 20E gallery. In the API 50CH gallery, acid is produced from inositol, mannitol, sorbitol, α-Methyl-D-Mannoside, α-Methyl-D-Glucoside, N-Acetyl-Glucosamine, amygdalin, arbutin, esculin, salicin, cellobiose, maltose, lactose, melibiose, melezitose, raffinose, gentiobiose, D-fucose, L-fucose, L-arabitol, gluconate, and 5-keto-gluconate. The major cellular fatty acids (> 5%) are iso-C_15:1_G, iso-C_15:0_, iso-C_15:0_ 3-OH, iso-C_17:0_ 3-OH, and summed feature 3 (comprising C_16: 1_ω7c and/or C_16: 1_ω6c). The G + C content of the genomic DNA is 32.0%. The type strain, TSM-5^ T^ (= KCTC 52508^ T^ = NBRC 112430^ T^), was isolated from Tsushima, Nagasaki prefecture, Japan.
Supplementary Information
Below is the link to the electronic supplementary material. Supplementary material 1 (PDF 20406 kb)Supplementary material 2 (XLSX 12 kb)