Prolixibacter bellariivorans gen. nov., sp. nov., a sugar-fermenting, psychrotolerant anaerobe of the phylum Bacteroidetes, isolated from a marine-sediment fuel cell

Electricity can be produced by sediment fuel cells placed in anoxic sediments: these cells contain natural populations of micro-organisms that transfer electrons to current-harvesting electrodes (Bond et al., 2002; Holmes et al., 2004b; Reimers et al., 2001; Tender et al., 2002). These electrons then flow to the cathode in the overlying aerobic water, where they react with oxygen. Several organisms that are able to oxidize organic acids, such as acetate, to CO₂ and utilize an electrode as the sole electron acceptor have already been isolated (Bond et al., 2002; Bond & Lovley, 2003, 2005; Chaudhuri & Lovley, 2003; Holmes et al., 2004d). However, there has been very little emphasis on the role that fermentative organisms might play within the microbial community of the sediment fuel cell.

Molecular analyses of marine-sediment fuel cells have led to the suggestion that fermentative organisms account for approximately 10 % of the microbial community present on the surfaces of current-harvesting anodes (Bond et al., 2002; Holmes et al., 2004b; Tender et al., 2002). In addition, micro-organisms with 16S rRNA gene sequences most similar to those of fermentative micro-organisms have been detected on anodes from microbial fuel cells powered by other fuel sources, such as organic wastewater (Kim et al., 2004), activated sludge (Lee et al., 2003) and oligotrophic river water (Phung et al., 2004). In fact, when wastewater from a corn-processing plant was provided as the electron donor in a microbial fuel cell powered by activated sludge, up to 19 % of the 16S rRNA gene sequences detected were most similar to known fermentative micro-organisms from the phylum Bacteroidetes (Kim et al., 2004).

Although fermentative micro-organisms are important members of the electrode community, they are unlikely to be directly involved in current production by the microbial fuel cell. Several micro-organisms with fermentative types of metabolism have been able to grow and produce current in microbial fuel cells; however, electron transfer to the electrode performed by these organisms was very inefficient (Park et al., 2001; Pham et al., 2003). For example, when Clostridium butyricum was grown in an anodic chamber with glucose as the electron donor, only 0.04 % of the electrons available in glucose were transferred to the electrode surface (Park et al., 2001) and less than 0.02 % of the electrons available from yeast extract (calculated from the chemical oxygen demand, 250 mg l⁻¹) were transferred to an electrode by Aeromonas hydrophila (Pham et al., 2003). In addition, only ∼25 % of the electrons available from the incomplete oxidation of pyruvate, lactate and propionate were transferred to the electrode surface by Desulfobulbus propionicus, via a mixed fermentative metabolism (Holmes et al., 2004a).

Dissimilatory metal-reducing micro-organisms, on the other hand, are able to transfer 80–95 % of the electrons available from the complete oxidation of an electron donor to a current-harvesting electrode (Bond & Lovley, 2003; Chaudhuri & Lovley, 2003; Holmes et al., 2004d). On the basis of these findings, it is likely that fermentative micro-organisms that can convert complex organic matter in anaerobic sediments to fermentation products such as acetate are involved in the initial steps of power production by the microbial fuel cell. These side-products are then completely oxidized by dissimilatory metal-reducing micro-organisms that are able to transfer electrons directly to the current-harvesting anode and produce a current.

Information concerning the role of fermentative micro-organisms on electrodes will help in the optimization of microbial fuel cells capable of harvesting electricity from waste biomass produced by agricultural, municipal and industrial sources in which energy is primarily stored in the form of carbohydrates. Here we report the characterization of a novel strain of bacteria, strain F2^T, isolated from the surface of an electricity-harvesting electrode incubated in marine sediments, that is likely to be a member of the fermentative niche within the microbial community on such electrodes.

Strain F2^T was isolated from a marine-sediment fuel cell that was constructed in the laboratory with sediments collected from Boston Harbour (MA, USA) near the World's End peninsula, at a water depth of 5 m, as described previously (Bond et al., 2002; Holmes et al., 2004b). After incubation and energy harvesting at 15 °C for 6 months, the anode was pulled from the sediment, washed with sterile anaerobic marine medium lacking electron donors or acceptors and containing Na₂S (1.0 mM) as a reducing agent. The anode surface was then scraped with a sterile razor blade into a slightly modified version of APW medium (Coates et al., 1995), referred to as A1 medium (Holmes et al., 2004d). Serial dilutions in A1 medium containing poorly crystalline Fe(III) oxide (100 mM) provided as the electron acceptor and benzoate (0.5 mM) as the electron donor yielded an enrichment culture comprising three organisms: a Geopsychrobacter species, a Fusibacter species and strain F2^T. In order to separate this tri-culture and isolate strain F2^T, eight 10-fold serial dilutions, with glucose (5 mM) provided as the fermentation substrate, were performed at 22 °C. The highest dilutions that were able to grow were diluted to extinction three more times and were then transferred to solidified medium (2 % agar) with glucose (5 mM). Isolated colonies were then selected and grown in A1 medium supplemented with glucose (5 mM).

For morphological and phenotypic characterization, strain F2^T was cultivated under anoxic conditions (N₂/CO₂; 80 : 20, v/v) in A1 medium with galactose (6 mM) as the fermentation substrate. Growth on an electrode was tested as previously described (Bond et al., 2002; Bond & Lovley, 2003) and utilization of the various electron donors and acceptors was only considered positive after four transfers (5 % inoculum). The end-products of glucose fermentation were analysed with a high-pressure liquid chromatograph (series 1100; Hewlett Packard) on a Fast Acid Analysis column (Bio-Rad) with an eluant of 5 mM H₂SO₄ and absorbance detection at 210 nm. Strain F2^T exhibited a mixed acid fermentative metabolism when grown on glucose; the end products included acetate, succinate and propionate.

Temperature, pH and salinity ranges were determined as described previously (Nevin et al., 2005). For salinity characterization tests, media contained NaCl without any divalent salts. There was no requirement for seawater. The API 20NE test (bioMérieux) was used according to the manufacturer's directions, with cells grown anaerobically with galactose (5–6 mM) as the fermentation substrate. Strain F2^T was positive for hydrolysis of aesculin and gelatin, indole production, β-galactosidase activity and arabinose assimilation and was negative for reduction of nitrate and nitrite, glucose fermentation, arginine dihydrolase, urease and the assimilation of glucose, mannose, mannitol, N-acetylglucosamine, maltose, potassium gluconate, capric acid, adipic acid, malic acid, trisodium citrate and phenylacetic acid. Growth on urease agar, MacConkey agar and DNase agar, acid production from sucrose and glucose, indole production, capnophilic metabolism and gliding motility were assessed using standard tests. Strain F2^T was found to be negative for growth on urease, DNase agar and MacConkey agar, and negative for acid production from sucrose and glucose, capnophilic metabolism and gliding motility. The strain was sensitive to penicillin. Pigment production was determined by using two methods: for flexirubin, the KOH method (Reichenbach et al., 1974) was used; for carotenoids, an acetone extraction method was used to produce an absorption spectrum (Denger et al., 2002). For cytochrome determinations, whole-cell suspensions were analysed using dithionite-reduced and air-oxidized spectra (350–700 nm) by using a UV2401-PC dual-beam spectrophotometer (Shimadzu). No peaks were observed in the dithionite-reduced minus air-oxidized difference spectrum of strain F2^T, a result that is consistent with the absence of c-type cytochromes. The results from these tests are shown, in comparison with those for genera within the phylum Bacteroidetes, in Table 1⇓. The physiology of strain F2^T was most similar to that of Alkaliflexus imshenetskii.

Cell morphology was observed under a Nikon Eclipse E600 microscope. Electron microscopic observations were performed in the microscopy facility at the University of Massachusetts, Amherst (Fig. 1⇓).

The phospholipid fatty acid (PLFA) composition (Pinkart et al., 2002), the double-bond geometry and the positions of monounsaturated PLFAs were determined by using previously described techniques (Dunkelblum et al., 1985). Seventeen major PLFAs and 15 trace PLFAs were observed in strain F2^T (see Supplementary Table S1 available in IJSEM Online). The major PLFAs were as follows: a-15 : 0 (23.29 %), 15 : 0 (20.73 %), i-15 : 0 (13.65 %), 17 : 0 (8.78 %), 16 : 0 (5.63 %), i-16 : 0 (4.51 %), 17 : 1ω8 (3.40 %), i-17 : 1ω8a (3.16 %), a-17 : 0 (2.83 %), i-14 : 0 (2.45 %), i-16 : 1ω7 (2.40 %), a-17 : 1ω8 (1.80 %), i-17 : 0 (1.49 %), a-15 : 1ω8 (1.03 %), 15 : 1ω6 (0.78 %), i-17 : 1ω8b (0.75 %) and 16 : 1ω7c (0.72 %). PFLA analysis showed that 35.1 % of the fatty acids in strain F2^T were normal saturated, 48.2 % were terminally branched saturated, 9.14 % were branched monounsaturated and 4.90 % were monounsaturated. The fatty acid composition of strain F2^T was similar to those of other species within the phylum Bacteroidetes (Alkaliflexus imshenetskii, Marinilabilia salmonicolor and Anaerophaga thermohalophila) (Table 2⇓).

Analysis of the DNA G+C content (Cashion et al., 1997) was performed by the Identification Service of the Deutsche Sammlung von Mikrooganismen and Zellkulturen (Braunschweig, Germany). Genomic DNA for the 16S rRNA gene and fusA, the gene that encodes elongation factor G, sequence analysis was extracted using the FastDNA spin kit (BIO 101) according to the manufacturer's instructions. The 16S rRNA gene sequence for strain F2^T was amplified with primers 8 forward (Lane et al., 1985; Lane, 1991) and 1492 reverse (Achenbach & Woese, 1995; Amann et al., 1990) as described previously (Holmes et al., 2004c). A gene fragment from the fusA gene in strain F2^T was amplified with primers fusAf and fusAr (Berchet et al., 2000) as described previously (Holmes et al., 2004c). The 16S rRNA and fusA gene fragments were compared with the GenBank nucleotide and protein databases using the blastn and blastx algorithms (Altschul et al., 1990). Nucleotide and amino acid sequences were manually aligned and hypervariable regions were masked in the Genetic Computer Group (GCG) sequence editor (Wisconsin Package, version 10). Aligned sequences were imported into paup 4.0b 4a (Swofford, 1998) in which the phylogenetic trees were inferred. The branching order was determined and compared with those obtained using maximum-parsimony, maximum-likelihood and distance-based algorithms. Similarity matrices were generated using align (Pearson, 1990).

Phylogenetic analysis of 16S rRNA and fusA gene sequences indicated that strain F2^T clusters within the phylum Bacteroidetes in the Bacteria (Fig. 2⇓). Strain F2^T is unique in that the most similar 16S rRNA gene sequence from an organism that is available in pure culture (Alkaliflexus imshenetskii) has a similarity of only 87.5 % (see Supplementary Table S2 available in IJSEM Online). The amino acid sequence of elongation factor G from strain F2^T was also only 71 % similar to that of another species within the Bacteroidetes, Porphyromonas gingivalis.

Although organisms phylogenetically related to strain F2^T are not available in pure culture, the 16S rRNA gene sequence from an uncultured organism detected in 2-bromophenol-dehalogenating and phenol-degrading consortia (Knight et al., 1999) was 97.2 % similar to that of strain F2^T. These results are interesting, as sediments previously collected from Boston Harbour were contaminated with polycyclic aromatic hydrocarbons (Hayes & Lovley, 2002; Hayes et al., 1999; McGroddy & Farrington, 1995; McGroddy et al., 1996; Rothermich et al., 2002; Rudnick & Chen, 1998; Wang et al., 2001). Other uncultured 16S rRNA gene sequences closely related to that of strain F2^T have been detected in permanently cold marine environments, such as sediments from the Arctic Ocean (Ravenschlag et al., 1999), the Japan Trench land slope (Li et al., 1999) and Antarctic sea ice (Brown & Bowman, 2001). Again, this is not very surprising, as strain F2^T is capable of growth at temperatures as low as 4 °C. On the basis of physiological, biochemical and phylogenetic data, strain F2^T represents a novel genus and species within the phylum Bacteroidetes, for which the name Prolixibacter bellariivorans gen. nov., sp. nov. is proposed.