Pseudoprevotella muciniphila gen. nov., sp. nov., a mucin-degrading bacterium attached to the bovine rumen epithelium

A Gram-negative, strictly anaerobic mucin-degrading bacterium, which we designated strain E39T, was isolated from the rumen epithelium of Korean cattle. The cells were non-motile and had a coccus morphology. Growth of strain E39T was observed at 30–45°C (optimum, 39°C), pH 6.5–8.5 (optimum, pH 7.5), and in the presence of 0.0–1.0% (w/v) NaCl (optimum, 0.0–0.5%). Strain E39T contained C16:0, C18:0, C18:1 ω9c, iso-C15:0, and anteiso-C15:0 as the major fatty acids. The major polar lipids were phosphatidylethanolamine, unidentified aminophospholipid, and unidentified lipids. The major respiratory isoprenoid quinones were MK-8 and MK-9. The major fermented end-products of mucin were acetate and succinate. The G+C content of the genomic DNA was 46.4 mol%. Strain E39T was most closely related to Alloprevotella rava 81/4-12T with an 87.3% 16S rRNA gene sequence similarity. On the basis of phenotypic, chemotaxonomic, and molecular properties, strain E39T represents a novel genus of the family Prevotellaceae; as such, the name Pseudoprevotella muciniphila gen. nov., sp. nov. is proposed. A functional annotation of the whole genome sequences of P. muciniphila E39T revealed that this bacterium has a putative mucin-degrading pathway and biosynthetic pathways of extracellular polymeric substances and virulence factors which enable bacteria to adhere to the epithelial cells and avoid the host’s immune responses.


Bacterial growth and genomic DNA extraction
Cells of strain E39 T were grown in 250 ml basal mucin medium at 39˚C for 24 h and harvested by centrifugation (10,000 rpm for 5 min). The genomic DNA of the cells was extracted, according to standard procedures, including phenol-chloroform extraction and ethanol precipitation [18].

Phenotypic and chemotaxonomic characterization
The cell morphology of cells grown on basal mucin agar medium at 39˚C for 3 days was investigated using phase-contrast microscopy and transmission electron microscopy (Talos L120C; FEI) at 120 kV. Gram staining was performed using a Sigma Gram staining kit following the manufacturer's protocol. The growth of strain E39 T , as determined from the optical density (OD) at a wavelength of 600 nm, was evaluated by culturing the cells in basal mucin medium, brain heart infusion (BHI) broth (BD), trypticase soy broth (TSB; BD), Columbia broth (Acumedia), and anaerobe basal broth (Oxoid) at 39˚C for 48 h. The optimum temperature, pH, and NaCl concentration for growth were determined by culturing the cells on basal mucin medium for 48 h at different temperatures (5-45˚C, at 5˚C intervals), pH (5.0-9.0 at 0.5 pH unit intervals), and NaCl concentrations (0.0-2.0% at 0.5% intervals). To determine the optimum pH, different pH buffers were used in the appropriate pH range (Na 2 HPO 4 -NaH 2 PO 4 buffer at pH 5.0-7.5; Tris-HCl buffer at pH 8.0-9.0) and the pH values were adjusted before and after autoclaving (121˚C, 15 min) [22]. Oxygen tolerance was investigated by measuring growth (OD at 600 nm) in the absence of a reducing agent (cysteine sulfide solution) or in the aerobic condition on basal mucin medium. A. tannerae ATCC 51259 T , A. rava 81/4-12 T , Paraprevotella clara YIT 11840 T and Prevotella melaninogenica ATCC 25845 T , which is the type species of the genus Prevotella, were used as reference strains to compare enzyme profiles and cellular fatty acid composition. The enzyme profiles were determined using an API Rapid ID 32A identification kit (bioMérieux) following the manufacturer's instructions. Analysis of cellular fatty acids was performed according to a standard MIDI protocol. All of the strains were cultivated in peptone-yeast extract-glucose (PYG) broth, except strain E39 T , which was cultivated in basal mucin medium. Cells were harvested at the late exponential phase and cellular fatty acids were extracted from the cells following four steps (saponification, methylation, extraction, and base wash). Fatty acid methyl esters were analyzed by gas chromatography (Hewlett Packard 6890) and identified using the RTSBA6 database of the Microbial Identification System (Sherlock ver. 6.0B) [23]. The polar lipid profiles were analyzed by thin-layer chromatography following the Minnikin et al. method [24]. The following reagents were used to detect different types of polar lipids: 10% ethanolic molybdophosphoric acid (for total lipids), ninhydrin (for aminolipids), Dittmer-Lester reagent (for phospholipids), and α-naphthol (for glycolipids). The isoprenoid quinones of strain E39 T , A. tannerae ATCC 51259 T , A. rava 81/4-12 T , P. clara YIT 11840 T , and P. melaninogenica ATCC 25845 T were extracted from their exponentially grown cells according to the procedure described by Jeon et al. [25] and analyzed at 40˚C using an Agilent infinity 1290 UHPLC equipped with a photodiode array detector (PAD) and an Agilent 6550 ifunnel Q-TOF MS (Agilent Technologies, USA). Briefly, 2 μl of quinone samples were injected into an Agilent Eclipse Plus C-18 column (2.1 mm × 100 mm, 2.1 μm) and eluted at 40˚C using water (A) and acetonitrile (B) containing 0.1% formic acid as a mobile phases with the following gradient: 0 min, 85% B; 30 min, 100% B; 40 min; and flow rate, 0.4 ml/min. Isoprenoid quinone peaks in the chromatograms were identified by their UV spectra generated by PAD and their molecular masses were assessed using Q-TOF MS. The mass spectrometry was performed under the following conditions: polarity, positive; gas temp, 250˚C; nebulizer, 35 psi; capillary, (+) 4,000 V; MS range, 100-1,500 m/z.

Metabolite analysis using 1 H NMR spectroscopy
Metabolic compounds including amino acids, monosaccharides, and organic acids in cultured broth of strain E39 T were analyzed using 1 H NMR spectroscopy, as described previously [26]. Briefly, the basal mucin medium (2.5 g hog gastric mucin per liter; no glucose), glucose medium (5 g glucose per liter; no hog gastric mucin), and mucin-glucose medium (2.5 g hog gastric mucin and 5 g glucose per liter) were prepared based on the basal mucin medium to investigate mucin and glucose utilization by strain E39 T and their fermentation products. Strain E39 T was cultured in 5 ml of each broth at 39˚C for 0, 9,18,27,36, and 54 h. The growth of the cells was monitored by measuring OD at 600 nm. The culture broths were centrifuged, filtered with a 0.45 μm syringe filter, and 0.3 ml of filtrate was mixed with 0.3 ml of 99.9% D 2 O (Sigma-Aldrich, USA) containing 5 mM sodium 2,2-dimethyl-2-silapentane-5-sulfonate (DSS, 97%; Sigma-Aldrich). The mixtures were transferred into NMR tubes and their 1 H NMR spectra were measured on Varian Inova 600-MHz NMR spectrometer (Varian, USA). Metabolic compounds were identified and quantified using the Chenomx NMR Suite program (ver. 6.1; Chenomx, Canada).

Genome sequencing and analysis
De novo genome sequencing was performed using a Pacific Biosciences (PacBio) RSII platform at Macrogen (Seoul, Korea; http://www.macrogen.com). A library was prepared using PacBio DNA Template Prep Kit 1.0. After sequencing, reads were trimmed to obtain high quality region and then assembled using RS hierarchical genome assembly process (HGAP ver. 3.0) [27]. The complete genome was annotated using a software tool Prokka with default parameter (ver. 1.12) [28].

16S rRNA gene and genome based phylogeny
Comparative analysis of the 16S rRNA gene sequences revealed that strain E39 T was closely related to the genera Alloprevotella, Paraprevotella, Prevotella, and Bacteroides. Alloprevotella rava 81/4-12 T , Paraprevotella clara YIT 11840 T , Paraprevotella xylaniphila YIT 11841 T , and Bacteroides gallinarum JCM 13658 T were most closely related to strain E39 T with 87.3%, 86.6%, 86.3%, and 85.9% 16S rRNA gene sequence similarities, respectively. The phylogenetic trees based on the ML algorithm and 92 bacterial core genes showed that strain E39 T was affiliated with the family Prevotellaceae and close to the genera Alloprevotella and Paraprevotella (Fig 1) [35][36][37]. Similarly, the phylogenetic trees based on the NJ and MP algorithms also showed that strain E39 T was affiliated with the family Prevotellaceae (S1 and S2 Figs).
However, all of the phylogenetic trees showed that strain E39 T formed a distinct phylogenic lineage from the other genera. These molecular and phylogenetic analyses suggest that strain E39 T represents a novel genus of the family Prevotellaceae.

Phenotypic and chemotaxonomic characterization
The transmission electron microscopic analyses showed that the cells of strain E39 T were coccus in morphology (680-820 nm in diameter), and lacking in flagella (Fig 2). In addition, filamentous structures were observed from the cell surface. Among the various types of media, including BHI broth, TSB, Columbia broth, and anaerobe basal broth, strain E39 T could only grow on basal mucin medium. Cells grew at temperatures between 30 and 45˚C, pH between 6.5 and 8.5, and in NaCl concentration between 0.0 and 1.0%. When the headspaces were filled with anaerobic gas, growth was observed in the absence of a reducing agent, but at a lower rate than in its presence. Cells did not grow in an aerobic atmosphere regardless of whether a reducing agent was present or not. These results demonstrate that strain E39 T prefers obligate anaerobic conditions but tolerates a trace amount of oxygen (S3 Fig).
In the API Rapid ID 32A panel, strain E39 T had positive activities of mucin-degrading enzymes including β-galactosidase, N-acetyl-β-glucosaminidase, and α-fucosidase [8]. Lack of urease activity suggest that strain E39 T may not attend to the digestion of urea, which is one of the roles of epimural bacteria. In particular, strain E39 T was distinguished from the reference taxa by a positive activity of arginine dihydrolase ( Table 1). The major cellular fatty acids

PLOS ONE
(> 5% of the total fatty acids) of strain E39 T were C 16:0 , C 18:0 , C 18:1 ω9c, iso-C 15:0 , and anteiso-C 15:0 . Among strain E39 T and the reference taxa of the family Prevotellaceae, iso-C 15:0 and anteiso-C 15:0 were commonly detected as major cellular fatty acids. Especially, strain E39 T had higher portion of C 16:0 and C 18:0 compared with the reference taxa ( Table 2). The major polar lipids of strain E39 T were phosphatidylethanolamine (PE), unidentified aminophospholipid (APL), and three unidentified lipids (L) (S4 Fig). Strain E39 T and the reference taxa of the family Prevotellaceae commonly contained PE and APL (except P. melaninogenica) ( Table 1). Menaquinone (MK)-8 and MK-9 were identified from strain E39 T as major respiratory quinones. However, MK-9 was identified from A. tanneare and A. rava, the most closely related strains, as the sole respiratory quinones, while MK-10 and 11 and MK-7 and 11 were identified from P. clara and P. melaninogenica, respectively (Table 1). It has been reported that members of the family Prevotellaceae contain MK of between 10 to 13 as isoprenoid numbers [37]. However, MK-9 was identified from strain E39 T and Alloprevotella members and MK-8 was identified from only strain E39 T (Table 1), which differentiated strain E39 T from other genera of the family Prevotellaceae.   [45]. All strains are positive for the following characteristics: activity � of β-galactosidase, alkaline phosphatase and alanine arylamidase. All strains are negative for the following characteristics: activity � of urease, β-glucosidase, proline arylamidase, phenylalanine arylamidase, leucine arylamidase, pyroglutamic acid arylamidase, tyrosine arylamidase, glycine arylamidase, histidine arylamidase and serine arylamidase, reduction of nitrates, and indole production.

PLOS ONE
In conclusion, 16S rRNA gene and genome based analysis clearly supported identification of strain E39 T as a novel genus of the family Prevotellaceae. On top of that, strain E39 T was distinguished from the reference taxa by several traits, including coccus shape, growth only on media containing mucin, a positive activity for arginine dihydrolase, high portion of saturated fatty acids (C 16:0 and C 18:0 ),andtypical polar lipids (unidentified lipids), and MK-8 and MK-9 as major respiratory quinones (Table 3). Taken together, strain E39 T is considered to represent

Metabolite changes during mucin and glucose fermentation
The growth of strain E39 T was observed in the basal mucin medium and mucin-glucose medium, but not in glucose medium (Fig 3A). Strain E39 T grew in the mucin-glucose medium better than in the basal mucin medium. Metabolites including free sugars, organic acids, and amino acids were analyzed by 1 H NMR spectroscopy. The concentration of acetate and succinate increased continuously in the both media ( Fig 3B). The concentration of mannose and Nacetylglucosamine which are sugar residues of mucin structure, increased in both media during the early fermentation (0-9 h), and then decreased after 9 h (S5 Fig). Alanine, glycine, and valine were detected as the major amino acids during the fermentation in the both media (S6  Fig). Strain E39 T was not able to grow in media without mucin, and the growth of strain E39 T was promoted by the addition of glucose. These results suggest that mucin is an essential growth nutrient for strain E39 T and it can utilize glucose as an energy source only under the presence of mucin. Furthermore, it was shown that strain E39 T might utilize mucin as an energy source during the early fermentation and produce acetate, succinate, and several amino acids as major fermented end-products.
(1) Mucin degradation and utilization. Mucin is host-derived glycoprotein composed of a protein backbone, lots of O-glycan, and a small number of N-glycan. N-acetlygalactosamine (GalNAc) is O-glycosylated to proline-threonine-serine (PTS) domain of a protein backbone and addition of galactose or N-acetlyglucosamine (GlcNAc) forms 8 types of O-glycan core structures. Addition of extended core (galactose and GlcNAc) and terminal residues (sialic acid and fucose) makes mucin structure more complex [39]. To utilize mucin as an energy source, bacteria need to have series of enzymes to degrade complex mucin structure. Several glycoside hydrolases (GHs) were known as enzymes involved in mucin degradation such as sialidases (GH33), α-fucosidases (GH29, GH95), exo-and endo-β-N-acetylglucosaminidases (GH84, GH85), β-galactosidases (GH2, GH20, GH42), α-N-acetylglucosaminidases (GH89), endo-β1,4-galactosidases (GH98), and α-N-acetylgalactosaminidases (GH101, GH129) [39,40]. In this study, the result of CAZyme annotation using dbCAN2 tool showed that P. muciniphila E39 T had 64 GHs, 5 carbohyrate-binding modules, 49 glycosyltransferases, 10 carbohydrate esterases, and no polysaccharide lyases (Table 4, S1 Dataset). Among 64 GHs, 28 GHs were putative enzymes involved in mucin degradation (Table 5). Compared to a Gram-negative mucin-degrading bacterium, B. thetaiotaomicron, P. muciniphila E39 T had fewer but similar kinds of mucin-degrading GHs. Bacteroides thetaiotaomicron has Sus-like systems for utilization of mucin glycan [41]. Glycan binding proteins on the cell surface bind polysaccharide and GHs partially degrade glycan. And then, two outer membrane proteins, homologs of SusD and SusC, import oligosaccharide into periplasm. After transportation of glycan from extracellular place to periplasm, glycan is degraded into monosaccharides by GHs and transferred to cytoplasm through inner membrane transporters [40,41]. We confirmed that P. muciniphila E39 T also had homologs of SusC and SusD through a BlastP search against the genome of P. muciniphila E39 T (S1 Dataset). Based on a KEGG pathway analysis and BlastP analysis, we constructed putative mucin degrading pathway of P. muciniphila E39 T . There were genes associated with metabolism of carbon sources including galactose, sialic acid (Neu5Ac), fucose, GlcNAc, and mannose on the results of KEGG pathway analysis (Fig 4). However, the proposed metabolic pathway was incomplete because of the absence of two genes (galactose 1-phosphate uridyltransferase, EC 2.7.7.12; N-acetylglucosamine kinase, EC 2.7.1.59) involved in galactose and GlcNAc metabolisms, respectively. In addition, P. muciniphila E39 T was negative in the mannose fermentation activity despite it harbored a mannose metabolic pathway. Further studies on mucin degrading pathway like monosaccharide transporting systems are needed to explain and understand mucin metabolic features of the P. muciniphila E39 T .
(2)Extracellular polymeric substances biosynthesis. Bacteria produce biofilms for various purposes. Biofilms are composed of extracellular polymeric substances including nucleic acids, lipids, proteins, and exopolysaccharides and this complex compounds have a wide range of roles in adhesion to other bacterial cells or host cells, protection from stresses such as antibiotic substances or harmful chemicals, and provision of structure for stratification against rapid environmental changes [42]. In the Wzx/Wzy-dependent pathway, one of exopolysaccharides biosynthesis pathways, glycosyltransferases assemble repeating units and a flippase (Wzx protein) translocates the units into the periplasmic place. The repeating units are elongated by a polymerase (Wzy protein) and transported across the outer membrane through a polysaccharide export protein [43]. P. muciniphila E39 T also produced branch-shaped extracellular structures (Fig 2) and these structures are predicted to contribute to adhesion to mucin or host cells. We identified that P. muciniphila E39 T had 1 putative extracellular polysaccharide biosynthesis locus (L_02166 -L_02195) through a BlastP search against the genome of P. muciniphila E39 T (S1 Dataset). There were 8 putative glycosyltransferases (L_02166, L_02169, L_02175, L_02179, L_02180, L_02181, L_02183, L_02184), 1 polymerase (L_02174), 1 flippase (L_02189), 1 serine acetyltransferase (L_02182), 1 N-acetyltransferase (L_02182), 1 aminotransferase (L_02188), 1 polysaccharide pyruvyl transferase (L_02191), and 1 polysaccharide export protein (L_02193).
(3)Virulence factors. Bacterial virulence factors allow bacteria to survive in the host, participating in adhesion, colonization, invasion, evasion or inhibition of immune responses, etc [44]. P. muciniphila E39 T had 17 putative virulence factors and they involved in adherence (KpsF, htpB, glf, hasB), antiphagocytosis (cps4J, cps4K, cps4L, hasB), immune evasion (tviC), stress reaction (clpC, clpP), O-antigen (galE, fcl), lipopolysaccharide (gmd), and metabolic adaptation (panD) (S1 Dataset). The virulence factors encoded in these genes may enable P. muciniphila E39 T to survive on the rumen wall, a place which host's defense mechanism is most active in the rumen, by attaching to epithelial cells and avoiding the host's immune responses. The absence of genes encoding exotoxin and involved in invasion suggests that P. muciniphila E39 T may have low pathogenicity. However, we cannot ignore the possibility that P. muciniphila E39 T is a potential pathogen because of its endotoxin (lipopolysaccharide) and mucinolytic ability. Further researches at molecular level or in vivo studies are required to determine the pathogenicity of P. muciniphila E39 T .

Conclusions
The genetic, physiological, and chemotaxonomic features support that strain E39 T represents a novel genus of the family Prevotellaceae. As such, the name Pseudoprevotella muciniphila gen. nov., sp. nov. is proposed. Pseudoprevotella muciniphila E39 T was isolated from the bovine rumen epithelium and this bacterium utilized mucin as a sole carbon source. The functional annotation of the complete genome of P. muciniphila E39 T supported that P. muciniphila E39 T possess a series of mucin degrading enzymes and putative mucin-degrading pathway. In addition, P. muciniphila E39 T is predicted to have putative metabolisms to synthesize extracellular polymeric substances and virulence factors for adhering to rumen epithelial cells and evading the host's immune responses. In short, this study contributes to discovery of a novel mucindegrading bacterium which has a potential ability to significantly affect host's physiology and Table 5. List of predicted mucin degrading enzymes in the complete genome of strain E39 T . its putative metabolic pathways which can assist to predict its function in epimural community.  Proposed mucin degrading and utilizing pathways of the P. muciniphila E39 T . These putative pathways were constructed based on CAZyme annotation, KEGG pathway analysis and BlastP analysis. Mucin is initially degraded into oligo-or monosaccharides by mucin degrading glycoside hydrolases (GHs) followed by transportation into periplasm by Sus-like outer membrane proteins. Additional degradation is occurred by periplasmic glycoside hydrolases and mucin-derived monosaccharides are imported through unidentified transporters and utilized as carbon sources. Metabolic pathways that are present in the P. muciniphila E39 T are depicted in gray, and metabolic pathways that are not present in the P. muciniphila E39 T are depicted in red. GlcNAc: N-acetlyglucosamine, GalNAc: N-acetylgalactosamine, Neu5Ac: sialic acid, ManNAc: Nacetylmannosamine.

PLOS ONE
are phosphatidylethanolamine (PE), unidentified aminophospholipid (APL), and three unidentified lipids. The major respiratory quinones are MK-8 and MK-9. The major fermented end-products of mucin are acetate and succinate. The genus is a member of the family Prevotellaceae of the phylum Bacteroidetes. The type species is Pseudoprevotella muciniphila.