Bioprospecting for Exopolysaccharides from Deep-Sea Hydrothermal Vent Bacteria: Relationship between Bacterial Diversity and Chemical Diversity

Many bacteria biosynthesize structurally diverse exopolysaccharides (EPS) and excrete them into their surrounding environment. The EPS functional features have found many applications in industries such as cosmetics and pharmaceutics. In particular, some EPS produced by marine bacteria are composed of uronic acids, neutral sugars, and N-acetylhexosamines, and may also bear some functional sulfate groups. This suggests that they can share common structural features with glycosaminoglycans (GAG) like the two EPS (HE800 and GY785) originating from the deep sea. In an attempt to discover new EPS that may be promising candidates as GAG-mimetics, fifty-one marine bacterial strains originating from deep-sea hydrothermal vents were screened. The analysis of the EPS chemical structure in relation to bacterial species showed that Vibrio, Alteromonas, and Pseudoalteromonas strains were the main producers. Moreover, they produced EPS with distinct structural features, which might be useful for targeting marine bacteria that could possibly produce structurally GAG-mimetic EPS.


Introduction
Glycosaminoglycans (GAGs) are structurally diverse animal glycopolymers with key roles in cell physiology and in some pathologies. These polysaccharides are polyanionic macromolecules composed of alternating uronic acids and hexosamines or neutral sugars [1,2]. Upon biosynthesis, GAG backbones undergo a series of specific enzymatic reactions leading to a large variety of Nand O-sulfations and epimerization patterns; these molecules are also heterogeneous with respect to size [3].
Hyaluronan (HA) is a non-sulfated and high molecular weight polymer. It is used in osteoarthritis treatment, in ophthalmology, and in wound healing as well as in the cosmetic industry due to its visco-elastic properties and biological properties on the cartilage and skin [1,4]. In contrast to HA, other GAGs (heparan, dermatan, chondroitin, and keratan sulfate) are highly sulfated; this remains crucial in some biological activities such as in anti-coagulation, cell-proliferation, and inflammation, as well as in fighting viruses, prions, Alzheimer's and Parkinson's diseases [1]. Heparin is a leading drug against many diseases and has been used medically for over 70 years [5]. However, finding carbohydrates that may influence EPS production. The ZHPUF medium was buffered by a mixture of 100 mM HEPES and PIPES (Sigma Aldrich, St. Louis, MO, USA) and pH was adjusted to 7.5. Glucose, saccharose, or mannitol was sterilized separately and added at a final concentration of 30 g·L −1 . Incubation temperature was set to 25 • C or 37 • C (Table S1). Liquid cultures were agitated at 150 rpm.

EPS Production
Preliminary screening of EPS production was performed in 1.2 mL tubes containing 1 mL of culture. Samples were incubated at 30 • C under 150 rpm agitation during 48 h. Growth was confirmed by absorbance at 600 nm. Cultures were centrifuged at 15,000 rpm, 30 min, 4 • C (Allegra™ 25R, Beckman Coulter, Brea, CA, USA). A 500 µL sample of the supernatant was withdrawn and stored at −20 • C before analysis on agarose gel stained with Stains all (Sigma Aldrich, St. Louis, MO, USA), as previously described in Ref. [31].
EPS extraction and purification was carried out on 100 mL of culture in a 500 mL Erlenmeyer incubated at 30 • C, 150 rpm during 48 h. After incubation, each culture was centrifuged at 8000× g, 4 • C during 30 min (Avanti J-E Beckman Coulter, Villepinte, France). Supernatant was filtered under vacuum on 2.6 µm and 0.7 µm glass microfiber membranes (Whatman, Maidstone, UK), purified by tangential ultrafiltration with a membrane of 100 kDa cut-off (Millipore, VWR, Fontenay-sous-Bois, France) and freeze dried. Production yield was estimated by the weight of the recovered powder.

Chemical Analysis
EPS solutions were prepared at 2 mg·mL −1 in water under very soft overnight agitation. The protein, sugar, and sulfur contents were determined by the corresponding method as follows. Co-extracted proteins were analyzed using the bicinchoninic acid method (BCA-Kit, Sigma). Neutral sugars were assessed by the method of Dubois et al. described in Reference [32] using glucose as the standard. Uronic acids were determined using the method proposed by Blumenkrantz and Asboe-Hansen [33], modified by Filisetti-Cozzi and Carpita [34]. The neutral sugar amount was corrected by the method of Montreuil and Spik [35] to consider the interfering uronic acid sugars. Sulfur content, indicating the presence of sulfate groups, was semi-quantitatively estimated by Azure A (Sigma) as follows. The EPS solution (50 µL) was mixed with 1 mL Azure A at 10 µg/mL in water. Absorbance at 535 nm gave the sulfur content (% w/w) based on an in-lab EPS reference whose sulfur content was determined by elemental analysis (Institut de Chimie des Substances Naturelles, ICSN-CNRS, Gif-sur-Yvette, France). Monomer compositions were determined by gas chromatography (GC) as per Karmeling et al. [36], modified by Montreuil et al. [37].

Phylogenetic Analysis
16S rRNA genes were amplified by PCR (Polymerase Chain Reaction) on a whole cell extract prepared as follows. A 500 µL culture was centrifuged at 10,000 rpm, 4 • C. The bacterial cell pellet was washed with PBS (Phosphate Buffer Saline) and subsequently suspended in 40 µL of sterile water. Samples were stored at −20 • C until analysis.
PCR were carried out with GoTaq ColorlessMasterMix 2× (Promega) on ten-time diluted samples with 55 • C hybridization and 1.5 min elongation using appropriate primers (8F (5 -AGAGTTTGATCATGGCTCAG-3 ) and 1489R (5 -GTTACCTTGTTACGACTTCAC-3 )). All of the PCR products were run on a 1% agarose gel electrophoresis with Sybr safe (Invitrogen, Fisher Scientific, Illkirch, France) and purified or extracted from gel with the GeneJET Gel Extraction and DNA Cleanup Micro Kit (Thermofisher, Waltham, MA, USA) before sequencing (GATC Biotech, Konstanz, Germany). A consensus sequence was built using forward and reverse sequences with Bioedit software [38]. Strain identification was performed by nucleotide Blast (megablast program) on the NCBI nr/nt Nucleotide Collection [39].
A phylogenetic tree was established using sequenced 16S rRNA genes. Reference strain 16S rRNA sequences were retrieved from NCBI (Table S3). Sequences were aligned with ClustalW and phylogenetic analyses were conducted with MEGA6 software [40]. The evolutionary history was inferred using the Neighbor-Joining method [41], and the optimal tree shown. The percentage of replicate trees where the associated taxa clustered together in the bootstrap test (1000 replicates) were shown next to the branches [42]. The tree was drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Jukes-Cantor method [43] and were in the units of the number of base substitutions per site. The analysis involved 40 nucleotide sequences. All ambiguous positions were removed for each sequence pair. E. coli was chosen as the outgroup.

Production of Exopolysaccharide
In this work, 51 strains isolated from deep-sea hydrothermal vents were analyzed for their polysaccharide production ability (Table S1). They were selected after a preliminary screening on a solid medium for mucoid colony formation indicating the putative ability of the bacterial strain to biosynthesize EPS (data not shown).
Bacteria were grown on liquid ZPHUF medium supplemented with glucose, saccharose, or mannitol (30 g·L −1 ). Both Alteromonas infernus and Vibrio diabolicus, producing GY785 and HE800 EPS, respectively, were used as laboratory references [44,45]. EPS released outside the cell were visualized by agarose gel electrophoresis ( Figure 1). EPS were produced in larger amounts when glucose or saccharose was used. A basal production was also observed on the ZPHUF medium. The electrophoretic migration depends not only on molecular weight, but also on the charge of the EPS molecule that appeared relatively homogeneous in the four different media, suggesting that there were no major variations in molecular weight and global charge regardless of the carbon source used. As observed in Figure 1, the MA896B and HE799 strains produced two different EPS whose electrophoretic migration clearly differed.
widespread substrate for fermentation bioprocesses in industry and would allow an easier scale up of the production process.

Chemical Characteristics of the EPS
Recovery yield and chemical composition (monosaccharide, protein, and sulfur contents) of the EPS produced by the different strains were determined (Table S2). Yield varied from 0.1 g·L −1 for MS919 and reached the highest yield of 0.58 g·L −1 for HY766. In comparison, in 100 mL culture, EPS produced by V. diabolicus and A. infernus were recovered with similar yields of 0.35 g·L −1 for HE800 EPS and 0.55 g·L −1 for GY785 EPS. This is consistent with the production yield usually encountered for marine EPS, which is around 1 g·L −1 [48]. However, the recovery yields of EPS from these reference strains obtained in bioreactors were considerably higher: 2.5 g·L −1 for HE800 EPS [29] and 5.5 g·L −1 for GY785 EPS [30]. Indeed, growth parameters were well-controlled in bioreactors. In particular, pH decreased when glucose was consumed, usually reaching 5 and resulted with lower growth in flasks. Although culture medium was buffered in our study, pH slightly decreased to 6.
The presence of both uronic acids and neutral sugars was assessed in all of the EPS by colorimetric assays (Table S2). Monosaccharide composition (GC analysis) revealed the presence of glucuronic and galacturonic acids (GlcA and GalA), N-acetylglucosamine, and N-acetylgalactosamine (GlcNAc and GalNAc) as well as neutral sugars such as glucose (Glc), galactose (Gal), mannose (Man), fucose (Fuc), and rhamnose (Rha) (Figure 2, Table S2). Among the EPS, four strains were particularly rich in uronic acids (GalA and GlcA) and N-acetylhexosamines (GlcNAc and GalNAc): HE800 EPS produced by V. diabolicus, and EPS produced by MS1004, MA893 and HE799. Nine other strains were also rich in uronic acids (>5% w/w); however, they did not contain any N-acetylhexosamines. Instead, they were rich in neutral sugars such as Glc and Gal that were present in GY785 EPS produced by A. infernus and EPS from MS1000, GY768, and GY791. In addition, some EPS exhibited important amounts of Fuc (ST717, BI725, and HE801) and Rha (ST717, BI725, HY765, GY791, MS919, and MS970). The presence of Fuc has been shown to be very important for some biological activities [49]. Although considered as a rare sugar together with Rha [49], they have both already been described in EPS from Pseudoalteromonas and Polaribacter strains isolated from Antarctic sea water, and sea ice and ocean sediments [50,51] as well as in Vibrio parahaemolyticus EPS [52] and that of Enterobacter A47 [53].
Sulfur was found in the EPS from 21 strains in a range from 2.6-5.5% (w/w) ( Figure 2, Table S2). It should be noted here that the four EPS composed of uronic acids and N-acetylhexosamines did not bear any sulfate group. Two sulfated EPS did not contain any uronic acid or N-acetylhexosamine (BI731 and GY788) and could therefore resemble sulfated mannans already identified in seaweeds [54]. Pseudoalteromonas sp. strain SM20310 has already been shown to produce a EPS molecules revealed by stains with all coloration adapted for acidic polysaccharides were found in 42 strains including A. infernus and V. diabolicus, i.e., 79% of the screened strains. There was no EPS production in 11 strains; moreover, 11 of the 42 positive strains produced EPS in very low amounts and four of them produced low molecular weight (LMW) EPS (Table S1). It could not be excluded that growth and production optimization could allow a better production yield of high-molecular weight (HMW) EPS from all the strains used in this study, but this possibility was not further investigated in this screening step.
The 27 strains producing high amounts of EPS were selected to be grown in a 100 mL culture (Table S1) including A. infernus and V. diabolicus. From this culture volume, we expected a sufficient amount of EPS to conduct chemical analyses. Moreover, through better agitation, we expected a better oxygen transfer rate, which was important for the EPS production yield [46,47]. An EPS was detected in the ZPHUF supplemented with glucose for all the selected strains with a good yield; consequently, glucose was chosen as the growth substrate for further studies. Indeed, glucose is a widespread substrate for fermentation bioprocesses in industry and would allow an easier scale up of the production process.

Chemical Characteristics of the EPS
Recovery yield and chemical composition (monosaccharide, protein, and sulfur contents) of the EPS produced by the different strains were determined (Table S2). Yield varied from 0.1 g·L −1 for MS919 and reached the highest yield of 0.58 g·L −1 for HY766. In comparison, in 100 mL culture, EPS produced by V. diabolicus and A. infernus were recovered with similar yields of 0.35 g·L −1 for HE800 EPS and 0.55 g·L −1 for GY785 EPS. This is consistent with the production yield usually encountered for marine EPS, which is around 1 g·L −1 [48]. However, the recovery yields of EPS from these reference strains obtained in bioreactors were considerably higher: 2.5 g·L −1 for HE800 EPS [29] and 5.5 g·L −1 for GY785 EPS [30]. Indeed, growth parameters were well-controlled in bioreactors. In particular, pH decreased when glucose was consumed, usually reaching 5 and resulted with lower growth in flasks. Although culture medium was buffered in our study, pH slightly decreased to 6.
The presence of both uronic acids and neutral sugars was assessed in all of the EPS by colorimetric assays (Table S2). Monosaccharide composition (GC analysis) revealed the presence of glucuronic and galacturonic acids (GlcA and GalA), N-acetylglucosamine, and N-acetylgalactosamine (GlcNAc and GalNAc) as well as neutral sugars such as glucose (Glc), galactose (Gal), mannose (Man), fucose (Fuc), and rhamnose (Rha) (Figure 2, Table S2). Among the EPS, four strains were particularly rich in uronic acids (GalA and GlcA) and N-acetylhexosamines (GlcNAc and GalNAc): HE800 EPS produced by V. diabolicus, and EPS produced by MS1004, MA893 and HE799. Nine other strains were also rich in uronic acids (>5% w/w); however, they did not contain any N-acetylhexosamines. Instead, they were rich in neutral sugars such as Glc and Gal that were present in GY785 EPS produced by A. infernus and EPS from MS1000, GY768, and GY791. In addition, some EPS exhibited important amounts of Fuc (ST717, BI725, and HE801) and Rha (ST717, BI725, HY765, GY791, MS919, and MS970). The presence of Fuc has been shown to be very important for some biological activities [49]. Although considered as a rare sugar together with Rha [49], they have both already been described in EPS from Pseudoalteromonas and Polaribacter strains isolated from Antarctic sea water, and sea ice and ocean sediments [50,51] as well as in Vibrio parahaemolyticus EPS [52] and that of Enterobacter A47 [53]. non-sulfated mannan [12]. The protein content of the EPS varied from 1% (w/w) for HE799 to 30.7% (w/w) for BI731. Proteins were most likely co-extracted with the EPS from the supernatant and were thus considered as contaminants of the EPS preparation. Indeed, since the EPS had a negative charge due to uronic acid and sulfate groups, their ability to bind and trap several compounds was very high [1,13]. Glucose can also be co-extracted with the EPS when residual glucose is present in the culture broth, but has also been described as a component of several heteropolysaccharides [21,44]. The average molecular weight ranged from 1.4 for MS1004 to 9.2 × 10 6 g·mol −1 for MS919. Thus, the EPS produced by the selected marine bacteria were high molecular weight (Table S2). The polydispersity index was close to 1, which emphasized that the EPS produced were highly monodisperse. No typical monosaccharide was identified in the EPS produced by the MA889 strain, despite its high recovery yield of 0.55 g·L −1 and colorimetric assays showing uronic acid and neutral sugar presence ( Figure 2, Table S2). It was thought that the cleavage by methanolysis might be incomplete, as already described for certain glycosidic bonds [55][56][57]. In addition, some monosaccharides have been shown to be degraded depending on methanolysis conditions [58]. Sulfur was found in the EPS from 21 strains in a range from 2.6-5.5% (w/w) (Figure 2, Table S2). It should be noted here that the four EPS composed of uronic acids and N-acetylhexosamines did not bear any sulfate group. Two sulfated EPS did not contain any uronic acid or N-acetylhexosamine (BI731 and GY788) and could therefore resemble sulfated mannans already identified in seaweeds [54]. Pseudoalteromonas sp. strain SM20310 has already been shown to produce a non-sulfated mannan [12].
The protein content of the EPS varied from 1% (w/w) for HE799 to 30.7% (w/w) for BI731. Proteins were most likely co-extracted with the EPS from the supernatant and were thus considered as contaminants of the EPS preparation. Indeed, since the EPS had a negative charge due to uronic acid and sulfate groups, their ability to bind and trap several compounds was very high [1,13]. Glucose can also be co-extracted with the EPS when residual glucose is present in the culture broth, but has also been described as a component of several heteropolysaccharides [21,44].
The average molecular weight ranged from 1.4 for MS1004 to 9.2 × 10 6 g·mol −1 for MS919. Thus, the EPS produced by the selected marine bacteria were high molecular weight (Table S2). The polydispersity index was close to 1, which emphasized that the EPS produced were highly monodisperse.
No typical monosaccharide was identified in the EPS produced by the MA889 strain, despite its high recovery yield of 0.55 g·L −1 and colorimetric assays showing uronic acid and neutral sugar presence ( Figure 2, Table S2). It was thought that the cleavage by methanolysis might be incomplete, as already described for certain glycosidic bonds [55][56][57]. In addition, some monosaccharides have been shown to be degraded depending on methanolysis conditions [58].

Phylogenetic Analysis
New 16S rRNA sequences obtained in this study are reported in Supplementary File S4. The blast similarity search on the NCBI database for all 51 strains is reported in Table S1 where it was revealed that the initial strain list contained 28 Alteromonas sp. bacteria (15 were selected for in depth analysis), 12 Pseudoalteromonas sp. strains among which five were selected, and five strains of Vibrio sp. (all EPS were studied). In addition, the Leclercia, Psychrobacter, Shewanella, Staphylococcus, and Tenacibaculum species had only one representative in the collection. Among them, the only species having shown EPS production were Psychrobacter and Staphylococcus, but the EPS was too low and these strains were not studied further. This suggested that the main deep-sea bacteria producing EPS belonged to the gamma-proteobacteria Vibrio, Alteromonas, and Pseudoalteromonas species.
To infer the phylogenetic relationships of EPS-producing strains, a tree was constructed using the 16S rRNA genes. 16S rRNA gene sequences of 15 reference strains were retrieved from the NCBI (Table S3) including V. diabolicus and A. infernus. A phylogenetic tree was built with these 40 16S rRNA sequences (Figure 3). The three main phylogenetic groups were indicated. The analysis of EPS composition of each taxonomic group revealed that none of the EPS produced by Vibrio sp. was sulfated ( Figure 4). On the other hand, they contained higher amounts of uronic acids (GlcA and GalA) and, in contrast with Alteromonas sp. and Pseudoalteromonas sp. EPS, possessed N-acetylhexosamine monomers (GlcNAc and GalNAc). Glucose and protein amounts were higher in EPS from Alteromonas and Pseudoalteromonas strains, suggesting that if they were not components of the glycopolymer, they could be trapped by the sulfate negative charge (Figure 4). For each EPS, when present, the GlcA amount was higher than the GalA. Overall, the Vibrio EPS appeared to be poor in neutral sugars excluding N-acetylhexosamines. Fucose was detected in each taxonomic group, while mannose was absent from all Vibrio EPS, and rhamnose was mainly present in Pseudoalteromonas EPS (Figure 4). In conclusion, we observed that general traits of the EPS composition were specific to the bacterial genera.

Conclusions
In this study, 51 new bacterial strains isolated from deep-sea hydrothermal vents were investigated for their ability to produce anionic polysaccharides. Chemical and phylogenetic analyses indicated that the majority of the EPS produced were anionic in nature and the producing strains in the collection mainly belonged to Alteromonas, Pseudoalteromonas, and Vibrio strains. Therefore, the marine environment appears to be a promising source to isolate new polysaccharides sharing common features with animal GAGs. Some EPS (MA886, MA889) may have unusual chemical structures and further studies are necessary to investigate them.
In the search for polysaccharides putatively mimetic of GAG, such as HA and heparin, HMW polyanionic carbohydrates are considered of high interest. Our data suggested that the best sources Purified EPS from marine bacteria are usually heteropolysaccharides bearing functional groups such as hydroxyl, carboxyl (from uronic acid), and N-acetylamine groups. Some rare sugars such as fucose or rhamnose containing methyl groups may also be encountered [21]. In addition, inorganic (sulfate, phosphate) or organic (such as pyruvate, lactate, glycerol-1-phosphate (Gro1P)) acids are also frequently identified in EPS [11,20,21,[59][60][61]. The chemical composition seems to be species-dependent, even strain-specific [48]. Growth conditions (carbon and nitrogen sources, pH, temperature) may also have a great influence on the structural features of the EPS [62]. Our findings were consistent with the previous description of bacterial EPS.
So far, EPS production has been mainly found in three main genera (Vibrio, Alteromonas, Pseudoalteromonas) and bioprospecting continues as marine environment contains a high microbial diversity [67]. In other studies, several EPS producing marine bacterial strains have also been identified in the genera Halomonas and Enterobacter [11,68], in Pseudomonas and Paracoccus strains isolated from French Polynesia microbial mats [61] as well as in Cobetia marina [69].
Previous data have suggested that the EPS may be strain specific [48]; however, we demonstrated that at the level of the bacterial genus, EPS structural features exhibited general specific traits. This further suggests that similar biosynthetic gene clusters might be responsible for the production of such carbohydrates. Little variation in gene regulation or expression, as well as in encoded proteins could result in slight molecular structural variability of the EPS. It has already been shown that the syp gene cluster encoding the biosynthesis of EPS in V. fischeri and V. diabolicus could be widespread among Vibrionaceae [70].

Conclusions
In this study, 51 new bacterial strains isolated from deep-sea hydrothermal vents were investigated for their ability to produce anionic polysaccharides. Chemical and phylogenetic analyses indicated that the majority of the EPS produced were anionic in nature and the producing strains in the collection mainly belonged to Alteromonas, Pseudoalteromonas, and Vibrio strains. Therefore, the marine environment appears to be a promising source to isolate new polysaccharides sharing common features with animal GAGs. Some EPS (MA886, MA889) may have unusual chemical structures and further studies are necessary to investigate them.
In the search for polysaccharides putatively mimetic of GAG, such as HA and heparin, HMW polyanionic carbohydrates are considered of high interest. Our data suggested that the best sources for HA-like EPS were bacterial members of the Vibrio species. On the other hand, sulfated carbohydrates of lower molecular weight can be alternatives to heparin-like GAG, and Alteromonas and Pseudoalteromonas appeared to be good candidates for such high-value carbohydrates. After the selection of bacterial strains of interest in the production of EPS of a particular composition, optimization of the production can be carried out. In addition, carbohydrate specific features can be further targeted by in vitro modification using a range of physical, chemical, mechanical, and even enzymatic methods.
Supplementary Materials: The following are available online at www.mdpi.com/2076-2607/5/3/63/s1, Table S1: List of bacterial strains studied in this work. EPS results from the screening step are also indicated as well as the phylogenetic identification, Table S2: Chemical features of EPS. Colorimetric assays, GC results and molar mass. nd: not detected, Prot: protein, UA: uronic acids, D: polydispersity, EC: standard deviation, Yield is in g·L −1 , all values are in weight percentage (% w/w), Table S3: 16S rRNA gene sequences of reference strains retrieved from NCBI, File S4: 16S rRNA genes sequenced in this study.