Introduction

The human gut harbours a large number of bacteria and other organisms that have been extensively studied and correlated with health and diseases (Clemente et al. 2012). Besides the fundamental roles that the human gut plays (metabolic functions, energy harvesting, immunity, etc.), it is a good target for the development of therapeutic strategies, especially when specific bacterial profiles can be drawn in certain diseases (Million et al. 2016; Tidjani Alou et al. 2017). With the advancement of culture-independent techniques, our knowledge has been enhanced in terms of human gut microbiota composition, although several drawbacks have emerged such as the accumulation of unidentified sequences, the inability to provide material for further investigations, failure to identify certain pathogenic species and depth bias (Greub 2012). Thus, culturomics was established with the purpose of culturing what was previously called ‘un-cultivable’ organisms, complementing metagenomics by providing identities to the operational taxonomic units (OTUs) and providing material for further investigations (Lagier et al. 2012a). This technique has succeeded in describing the human gut microbiota by isolating a significant number of new bacterial species and describing these via the ‘taxonogenomic’ approach (Fournier and Drancourt 2015; Lagier et al. 2016; Bilen et al. 2018). Herein, we describe using taxonogenomics, a new bacterial species, Dysgonomonas massiliensis sp. nov. represented by strain Marseille-P4356T (= CCUG 71356T = CSUR P4356T), isolated from a stool sample of a healthy Pygmy male. The genus Dysgonomonas was first isolated from a human gall bladder and members are known to be Gram-stain negative, coccobacillus-shaped (Hofstad et al. 2000) and associated with gastroenteritis in immunocompromised patients (Murray et al. 2013).

Materials and methods

Before starting this project, approval from the ethics committee under the number 09-022 from the Institut Fédératif de Recherche 48 was obtained along with a signed informed consent from the donor.

Strain isolation and identification

Stool samples were taken from a 39-year-old healthy Pygmy male according to the Nagoya protocol. The shipment of samples was performed using a C-Top Ae-Ana medium (Culture Top, Marseille, France) and stored at − 80 °C at the URMITE laboratory (Marseille, France). Diluted in phosphate buffer saline, the stool sample was incubated in anaerobic culture bottles (BD BACTEC®, Plus Anaerobic/F Media, Le Pont de Claix, France) supplemented with 5% (v/v) sheep blood and 5% (v/v) sterile-filtered cow rumen at 37 °C. The growth of colonies was evaluated by sub-culturing samples on 5% sheep blood–enriched Columbia agar (bioMérieux, Marcy l’Etoile, France) and incubating them at 37 °C under anaerobic conditions. The identification of colonies was performed as previously described using MALDI-TOF MS (Matrix assisted laser desorption ionization time of flight mass spectrometry; Seng et al. 2010; Lagier et al. 2012b). When MALDI-TOF MS failed to identify the studied organism, 16S rRNA gene sequencing was performed for further analysis as previously described (Drancourt et al. 2004). Sequences were optimised and assembled with CodonCode Aligner tool (http://www.codoncode.com) and a Blast analysis was performed using the NCBI database and BLAST engine (http://blast.ncbi.nlm.nih.gov.gate1.inist.fr/Blast.cgi). A 16S rRNA sequence divergence of more than 5% was required to classify isolates into a new genus and a similarity of less than 98.65% was required to propose a new species (Kim et al. 2014). The 16S rRNA gene sequence of newly isolated bacterial species and the accompanying MALDI-TOF mass spectrum were deposited in EMBL-EBI and UMRS databases, respectively.

Growth conditions, phenotypic, biochemical and antibiotic resistance analyses

Optimal growth conditions were evaluated by performing culture assays at different temperatures (25, 37, 45 and 55 °C), pH (6, 6.5, 7 and 8.5), oxygen environments (anaerobic (GENbag anaer, bioMérieux), aerobic and micro-aerophilic (GENbag Microaer, bioMérieux)) and NaCl concentrations (0, 5, 10, 50, 75 and 100 g L−1).

In addition, API strips (20A, 50CH and ZYM; bioMerieux, France) were used according to the manufacturer’s guidelines in order to determine the major enzymatic and biochemical activities of the unidentified organism. Spore-forming ability was evaluated by performing a culture assay using a bacterial suspension, previously exposed to a 20 min heat shock at 80 °C. A DM1000 photonic microscope (Leica Microsystems, Nanterre, France) was used as previously described for motility and Gram assessments and cell images were obtained as previously (Bilen et al. 2017).

Bile solubility test was performed by checking the turbidity after adding 0.5 mL of 2% sodium desoxycholate (Sigma Aldrich) to 0.5 mL of 1 McFarland bacterial suspension (in normal saline). The latter was incubated at 37 °C for 10 min before reading the results. A tube with 0.5 mL of 1 McFarland bacterial suspension was taken as a control with normal saline added instead 2% sodium desoxycholate.

The antibiotic resistance profile of strain Marseille-P4356T was determined by performing different E-tests with: vancomycin, imipenem, rifampicin, amoxicillin, benzylpenecilin, minocycline, colistin, amikacin, metranidazol, teicoplanin, ertapenem, daptomycin, ceftriaxone, doxycycline, fosfomycin and kanamycin (bioMérieux, France).GC/MS and fatty acid analysis of strain Marseille-P4356T was carried out using 50 mg of bacterial biomass per tube as previously described (Dione et al. 2016). A Clarus 500 chromatography system was used connected to a SQ8s mass spectrometer (Perkin Elmer, Courtaboeuf, France) in order to analyse short chain fatty acids, as previously described (Zhao et al. 2006; Togo et al. 2017). Acetic, propionic, butyric, isobutyric, valeric, isovaleric, caproic, isocaproic, enanthic and isoenanthic acids were used as reference standards (Sigma Aldrich, Lyon, France).

Genome sequencing and analyses

For genomic analyses, genomic DNA of strain Marseille-P4356T was first mechanically extracted using an acid-washed glass beads (G4649-500 g Sigma) treatment in a FastPrep BIO 101 instrument (Qbiogene, Strasbourg, France) at maximum speed (6.5) for 90 s. After 2 h of lysozyme incubation at 37 °C, an EZ1 biorobot (Qiagen) was used for DNA extraction with the EZ1 DNA tissues kit. The elution volume was 50 µl. Qubit assay was used for genomic DNA quantification with the high sensitivity kit (Life technologies, Carlsbad, CA, USA) to 71.6 ng/µl. Next, MiSeq Technology (Illumina Inc, San Diego, CA, USA) was used for genomic DNA sequencing by means of the Nextera mater pair strategy prep kit (Illumina). The studied organism, along with 11 other projects, was barcoded and mixed within the same run.1.5 µg of genomic DNA was used for library preparation using the Nextera mate pair Illumina guide. Tagmentation with a mate pair junction adapter was performed after fragmentation, the pattern of which was previously validated on an Agilent 2100 BioAnalyzer (Agilent Technologies Inc, Santa Clara, CA, USA) with a DNA 7500 labchip. Optimal DNA fragments were 8.933 kb and ranged from 1.5 kb up to 11 kb. 600 ng of the tagged fragmented DNA were circularised with no size selection and then mechanically sheared to small fragments with an optimal size at 721 bp on a Covaris device S2 in T6 tubes (Covaris, Woburn, MA, USA). The final library concentration was measured at 20.98 nmol/l and its profile was visualised on a High Sensitivity Bioanalyzer LabChip (Agilent Technologies Inc, Santa Clara, CA, USA). Normalization of the library at 2 nM was performed before pooling, followed by a denaturation and dilution step at 18 pM. Cluster generation and sequencing were set automatically for a single 39-h run in a 2 × 151-bp. Total information of 6 Gb was obtained from a 623 K/mm2 cluster density with a cluster passing quality control filters of 97.5% (11,904,000 passing filter paired reads). Within this run, the index representation for strain Marseille-P4356T was determined to be 7.32%. The 871,666 paired reads were trimmed and assembled as previously described (Lagier et al. 2012b). The Rast tool was used for genome annotation and analysis (Aziz et al. 2008; Overbeek et al. 2014; Brettin et al. 2015). Phages were detected using PHAST tool (Zhou et al. 2011), rRNA was detected with RNAmmer (Lagesen et al. 2007) and Artemis was used for genome circular representation (Carver et al. 2009).

For phylogenetic analyses, the NCBI nucleotide database (https://www.ncbi.nlm.nih.gov/nucleotide/) was use to retrieve 16S rRNA gene sequences of the phylogenetically closely related species with standing in nomenclature of strain Marseille-P4356T after performing a Blast search at https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastSearch. Alignment was done using the ClustalW tool in MEGA (Kumar et al. 2016), which was adapted for phylogenetic inference generation within the maximum-likelihood method with 500 bootstraps (Tamura and Nei 1993; Tateno et al. 1994).

To further confirm the novelty of strain Marseille-P4356T, OrthoANI and dDDH (Digital DNA:DNA hybridization) estimates were calculated between the genome of this strain and the genomes of its phylogenetically close species using OAT software (Lee et al. 2016) and Genome-to-Genome Distance Calculator 2.1 (http://ggdc.dsmz.de), respectively.

Results

Strain identification

MALDI-TOF MS failed to identify strain Marseille-P4356T due to the absence of a matching spectrum in the current database. Therefore 16S rRNA gene sequencing was performed and revealed that this strain exhibited 92.96% sequence similarity with Dysgonomonas gadei strain JCM 16698T (NR_113134.1) and 93.57% Dysgonomonas mossii strain JCM 16699T (NR_113135.1), currently the phylogenetically closest species with standing in nomenclature (Fig. 1; Supplementary Table 1). Even though the 16S rRNA gene sequence of strain Marseille-P4356T diverges by more than 5% with its phylogenetically close species, this strain falls between other Dysgonomonas species in the phylogenetic tree (Fig. 1). Accordingly, we classify strain Marseille-P4356T as a new bacterial species belonging to the genus Dysgonomonas. The 16S rRNA gene sequence of this strain has been deposited in GenBank with the accession number LT934441 and the mass spectrum generated by MALDI-TOF MS (Fig. 2) uploaded to the URMS database (http://www.mediterranee-infection.com/article.php?larub=280&titre=urms-database). The 16S rRNA gene similarities between strain Marseille-P4356T and closely related species are listed in Supplementary Table 1.

Fig. 1
figure 1

Phylogenetic tree representing the position of strain Marseille-P4356T relative to other closely related species. 16S rRNA sequences were recovered after a nucleotide Blast search of the Silva project “The All-Species Living Tree” database (LTPs121). Muscle was used for sequence alignment and Fast tree for approximately maximum likelihood sequence tree generation. Local values obtained with the Shimodaira–Hasegawa test are shown on the nodes

Full size image
Fig. 2
figure 2

Reference MALDI-TOF mass spectrum representing D. massiliensis strain Marseille-P4356T obtained after comparing 12 spectra

Full size image

Main characteristics of Strain Marseille-P4356T′

Cells of strain Marseille-P4356T were observed to be Gram-stain negative, catalase positive but oxidase negative cocci. After 48 h of growth under anaerobic conditions on COS medium + 5% (v/v) sheep blood (bioMerieux, France) at 37 °C, the strain was observed to form smooth grey colonies of 0.03–1 mm diameter. Under electron microscope, cells were found to have a mean diameter of 0.6 µm (Table 1). Strain Marseille-P4356T was found to be able to grow at temperatures between 25 and 37 °C under aerobic and anaerobic conditions, optimally at 37 °C under anaerobic conditions to the strain can grow at pH between 6 and 8.5 and NaCl concentration below to 50 g L−1. Strain Marseille-P4356T is non-motile, asporogenous and bile soluble.

Table 1 Differential characteristics of Dysgonomonas massiliensis (DM), Dysgonomonas gadei (DG), Dysgonomonas capnocytophagoides (DC), Dysgonomonas macrotermitis (DMM), Dysgonomonas hofstadii (DH), Dysgonomonas oryzarvi (DO), Dysgonomonas alginatilytica (DA), Dysgonomonas mossii (DMMM) and Dysgonomonas termitidis (DT).
Full size table

In antibiotic susceptibility tests, strain Marseille-P4356T exhibited the following minimal inhibitory concentrations: vancomycin (> 32 μg mL−1), rifampicin (2 μg mL−1), benzylpenecilin (0.032 μg mL−1), amoxicillin (0.023 μg mL−1), imipenem (0.094 μg mL−1), amikacin (> 256 μg mL−1), minocycline (16 μg mL−1), teicoplanin (3 μg mL−1), colistin (> 256 μg mL−1), daptomycin (> 256 μg mL−1), metronidazole (1.5 μg mL−1), ceftriaxone (0.38 μg mL−1), ertapenem (0.016 μg mL−1), fosfomycin (< 1024 μg mL−1), doxycycline (24 μg mL−1) and anamycin (4 μg mL−1) respectively. These indicate possible resistance mechanisms towards metronidazole, doxycycline, amikacin, colistin, fosfomycin and vancomycin. The closely related species D. gadei exhibited minimal inhibitory concentrations as follows: metronidazole (1.5 μg mL−1), doxycycline (0.19 μg mL−1), imipenem (0.5 μg mL−1), ceftriaxone (> 256 μg mL−1), vancomycin (48 μg mL−1), teicoplanin (> 256 μg mL−1), amoxicillin (6 μg mL−1), benzylpenicillin (32 μg mL−1) and azithromicin (16 μg mL−1), showing also a diverse resistance profile (Hofstad et al. 2000). To have an unidentified bacterial species with such a wide resistance profile should encourage the scientific-medical community to enhance their work on characterising the human microbiota and studying unknown commensal bacteria as they may have the capacity to become pathogenic (Isenberg 1988). Using API 50CH tests, positive reactions were observed with l-arabinose, d-xylose, d-galactose, d-glucose, d-fructose, d-mannose, d-mannitol, N-acetylglucosamine, amygdalin, arbutin, aesculin ferric citrate, salicin, d-maltose, d-lactose, starch, glycogen, gentiobiose, d-arabitol and potassium gluconate. For API ZYM tests, positive reactions were observed with esterase (C4), esterase lipase (C8), lipase (C14), valine arylamidase, α-chymotrypsin, naphthol-AS-BI-phosphohydrolase, α-galactosidase, α-glucosidase, β-glucosidase and N-acetyl-β-glucosaminidase. In API 20A tests, positive reactions were observed with glucose, mannitol, lactose, maltose, xylose, arabinose and mannose. A comparison of the general phenotypic and biochemical characteristics between this strain and close relatives is shown in Table 1 (Hofstad et al. 2000; Lawson et al. 2002, 2010; Kodama et al. 2012; Yang et al. 2014; Kita et al. 2015; Pramono et al. 2015). All Dysgonomonas species are facultatively anaerobic, Gram-stain negative, non-motile and produce acids from d-mannose, d-glucose, d-maltose and d-lactose. Only strain Marseille-P4356T is alkaline phosphatase-positive.

The most abundant fatty acid identified in strain Marseille-P4356T is anteiso-C15:0 (12-methyl-tetradecanoic acid). Several specific 3-hydroxy structures were also identified. Minor amounts of unsaturated and other saturated fatty acids were also detected. The closely related species, D. gadei contains iso-C14:0 (12-methyl-tridecanoic acid), anteiso-C15:0 (12-methyl-tetradecanoic acid), C16:0 (hexadecanoic acid) and iso-3OH C16:0 (3-hydroxy-14-methyl-Pentadecanoic acid) as the major fatty acids (Hofstad et al. 2000) which were also all detected in strain Marseille-P4356T (Table 2). These results are in accordance with the known fatty of acid profiles of these organisms, which are generally characterised by the presence of a large amounts of anteiso- and iso-methyl branched, straight- chain saturated and 3-hydroxy long-chain fatty acids (Wallace et al. 1989). Regarding short chain fatty acid fermentation products, after 72 h of culture in a hemoculture flask supplemented with blood, production of acetic (> 10 mM), propanoic (6.8 ± 0.3 mM), isobutanoic (0.5 ± 0.1 mM) and isopentanoic (1.1 ± 0.1 mM) acids was detected. Butanoic, pentanoic, hexanoic, isohexanoic and heptanoic acids were not detected.

Table 2 Fatty acids profile of strain Marseille-P4356
Full size table

Genomic analyses and comparison

The draft genome of strain Marseille-P4356T is 3,472,011 bp long with 37.3 mol% G+C content. It is composed of 26 contigs (26 scaffolds). 2994 genes were detected with 2907 coding DNA sequences. 51 genes (RNA), 7 rRNA (5, 1, 1 for 5S, 16S, 23S rRNA, respectively) and 42 tRNAs were detected. No CRISPRs repeats were found. 1395 proteins were annotated as hypothetical proteins. Using the PHAST tool, 4 prophages regions were identified, of which one region is intact, 3 regions are incomplete (not 100% compatible with phage sequences). The intact region was located in the region 3097681–3103890 with 43.95 mol% G+C content and shares a high number of proteins (70%) with Enterobacteria phage phiX174 sensu lato (NC_001422.1). The importance of tracking phage sequences is to check for horizontal gene transfer scenarios that might have occurred and led to specific virulence traits for the studied microorganism (Arber 2014; Penadés et al. 2015). Based on the RAST annotation, 41 proteins were correlated with virulence factors and defense, of which 28 were related to antibiotic and toxic compound resistance as follows: Copper homeostasis (2), Cobalt-Zinc-cadmium-resistance (4), vancomycin resistance (1), tripartite multidrug resistance system (3), streptothricin resistance (1), tetracycline resistance-ribosome protection type (4), fluoroquinolones resistance (4), beta-lactamase (1) and multidrug resistance pumps (8). No spore protection system was detected, which is in accordance with our phenotypic results. Moreover, no motility features were detected, which is also in accordance with our observations. A graphical representation of the draft genome of strain Marseille-P4356T is shown in Fig. 3. When compared to phylogenetically closely related species with standing in nomenclature, the draft genome sequence of strain Marseille-P4356T size is smaller than that of, Dysgonomonas macrotermitis strain DSM 27370, Dysgonomonas hofstadii MX 1040T, Dysgonomonas capnocytophagoides DSM 22835, Dysgonomonas gadei ATCC BAA-286T (D. gadei), Dysgonomonas mossii DSM 22836 and Paludibacter propionicigenes WB4 (4.7, 5.0, 4.4, 5.2, 3.9 and 3.7 Mb, respectively).

Fig. 3
figure 3

Circular representation of the D. massiliensis strain Marseille-P4356T draft genome. From the outer strand to the inner strand: Coding DNA sequences on the forward strand, coding DNA sequences on the reverse strand, rRNA, tRNA and GC plot and skew

Full size image

The G+C content of strain Marseille-P4356T is higher than that of D. macrotermitis, D. hofstadii, D. capnocytophagoides, D. gadei, D. mossii and P. propionicigenes (40, 45.3, 37.7, 39.6, 37.5 and 38.9%, respectively) (Supplementary Table 2).

The OrthoANI values (%) between strain Marseille-P4356T, P. propionicigenes, D. macrotermitis, D. hofstadii, D. capnocytophagoides, D. gadei and D. mossi were 66.54, 69.06, 69.67, 69.38, 69.76 and 69.77, respectively (Supplementary Table 3). These values are below the cutoff threshold (95–96%), determined to discriminate bacterial species (Lee et al. 2016). Additionally, dDDH estimates between strain Marseille-P4356T, P. propionicigenes, D. macrotermitis, D. hofstadii, D. capnocytophagoides, D. gadei and D. mossi were 18.4 [16.2–20.7], 21.3 [19–23.7%], 20 [17.8–22.4], 21.9 [19.6–24.3], 20 [17.8–22.4%] and 21.8 [19.5–24.2%], respectively (Supplementary Table 4). dDDH estimates of all species between each other were below 30% and below the considered threshold with intervals suggested to delimitate bacterial species (Meier-Kolthoff et al. 2013).

Phylogenetic, phenotypic and biochemical analyses demonstrated that strain Marseille-P4356T merits recognition as a novel member of the genus Dysgonomonas for which the name Dsygonomonas massiliensis is proposed. The type strain is Dysgonomonas massiliensis strain Marseille-P4356T (= CSUR P4356T = CCUG 71356T), from the human gut. The Digital Protologue database (Rosselló-Móra et al. 2017) TaxoNumber for strain Marseille-P4356T is TA00816.

Description of Dysgonomonas massiliensis sp. nov.

Dysgonomonas massiliensis (mas.si.li.en’sis. L. fem. adj., massiliensis, pertaining to Massilia, the ancient name of the city of Marseille, where this bacterium was characterised).

Cells are Gram-stain negative cocci, 0.6 μm in diameter, non-motile and asporogenous. Forms smooth grey colonies, 0.03–1 mm diameter. Catalase positive and oxidase negative. Grows optimally at 37 °C under anaerobic conditions and tolerates pH of 6–8.5 and NaCl concentration below 50 g L−1. The major biochemical test reactions are given in Table 1. The most abundant fatty acid is anteiso-C15:0.

The type strain is Marseille-P4356T (= CSUR P4356T = CCUG 71356T), which was isolated from the stool sample of a healthy 38-year-old pygmy male from the Democratic Republic of the Congo. The draft genome of strain Marseille-P4356T is 3.5 Mb long with 37.3 mol% of G+C content. The 16S rRNA gene and genome sequences of the type strain have been deposited in EMBL-EBI under accession numbers LT934443 and OEPV00000000, respectively.