- Email: [email protected]
Microbial communities involved in electricity generation from sulfide oxidation in a microbial fuel cell
Biosensors and Bioelectronics 26 (2010) 470–476
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Microbial communities involved in electricity generation from sulfide oxidation in a microbial fuel cell Min Sun a , Zhong-Hua Tong a,∗ , Guo-Ping Sheng a , Yong-Zhen Chen a , Feng Zhang a , Zhe-Xuan Mu b , Hua-Lin Wang b , Raymond J. Zeng a , Xian-Wei Liu a , Han-Qing Yu a,∗ , Li Wei c , Fang Ma c a b c
Department of Chemistry, University of Science & Technology of China, 96 Jinzhai Road, Hefei 230026, China School of Chemical Engineering, Hefei University of Technology, Hefei 230092, China State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, China
a r t i c l e
i n f o
Article history: Received 22 May 2010 Received in revised form 7 July 2010 Accepted 20 July 2010 Available online 27 July 2010 Keywords: Biocatalysis Microbial community diversity Microbial fuel cells Sulfate-reducing bacteria Sulfide oxidation Sulfur-oxidizing bacteria
a b s t r a c t Simultaneous electricity generation and sulfide removal can be achieved in a microbial fuel cell (MFC). In electricity harvesting from sulfide oxidation in such an MFC, various microbial communities are involved. It is essential to elucidate the microbial communities and their roles in the sulfide conversion and electricity generation. In this work, an MFC was constructed to enrich a microbial consortium, which could harvest electricity from sulfide oxidation. Electrochemical analysis demonstrated that microbial catalysis was involved in electricity output in the sulfide-fed MFC. The anode-attached and planktonic communities could perform catalysis independently, and synergistic interactions occurred when the two communities worked together. A 16S rRNA clone library analysis was employed to characterize the microbial communities in the MFC. The anode-attached and planktonic communities shared similar richness and diversity, while the LIBSHUFF analysis revealed that the two community structures were significantly different. The exoelectrogenic, sulfur-oxidizing and sulfate-reducing bacteria were found in the MFC anodic chamber. The discovery of these bacteria was consistent with the community characteristics for electricity generation from sulfide oxidation. The exoelectrogenic bacteria were found both on the anode and in the solution. The sulfur-oxidizing bacteria were present in greater abundance on the anode than in the solution, while the sulfate-reducing bacteria preferably lived in the solution. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Microbial fuel cells (MFC) provide a new approach for electricity generation from biomass. By utilizing microbial metabolism an MFC produces an electrical current from the degradation of organic/inorganic matters. Different types of exoelectrogenic bacteria have been isolated from naturally colonized anodes. However, many of these strains generate low power densities when grown as pure cultures (Rabaey et al., 2004; Zuo et al., 2008). It is therefore proposed some synergistic interactions might exist among members of an exoelectrogenic community (Logan, 2009). Molecular techniques have been widely applied to examine the composition of microbial communities in MFCs inoculated with mixed cultures (Kim et al., 2007; Lee et al., 2003; Phung et al., 2004). The communities in MFCs display great bacterial diversity, depending on the inoculum source or electron donors. An acetatefed system inoculated with activated sludge exhibited a near-even distribution of ˛-, -, and ı-Proteobacteria (Lee et al., 2003). In
∗ Corresponding authors. Fax: +86 551 3601592. E-mail addresses: [email protected] (Z.-H. Tong), [email protected] (H.-Q. Yu). 0956-5663/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2010.07.074
an ethanol-fed two-chamber MFC, 83% of cloned 16S rRNA gene sequences were related to ˇ-Proteobacteria (Kim et al., 2007). A ˇ-Proteobacteria-dominated MFC community was reported with a river sediment inoculum and river water feeding, while in another MFC fed with glucose and glutamate the predominant clones were ˛-Proteobacteria (Phung et al., 2004). Bacteria in MFCs can survive either through being electrochemically active or through interactions with other bacteria (Logan, 2007). A synergistic relationship between photosynthetic bacteria and heterotrophic exoelectrogenic bacteria was proposed to exist in a self-sustained phototrophic MFC (He et al., 2009). Glucose conversion in MFCs was demonstrated to be a complex process, in which different bacterial populations were involved. Hydrogen and acetate produced by fermentative bacteria could be used by exoelectrogenic bacteria for electricity generation (Freguia et al., 2008). The synergistic mechanism also plays an important role in electron transfer process. Lactobacillus amylovorus and Enterococcus faecium could make use of phenazine-based mediators produced by Pseudomonas sp. to transfer electrons (Rabaey et al., 2005). Wastewaters usually contain not only organics but also inorganic matters, such as sulfide. Sulfide is a hazardous substance and needs to be removed from wastewater before discharged
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
into environment. The MFC system was found to be effective for simultaneous sulfide removal and electricity generation (Rabaey et al., 2006). Recently, we have demonstrated that microbial catalysis played an important role in the electricity generation from a sulfide-fed MFC (Sun et al., 2009). Although the microbes involved in the oxidation of carbon-based compounds have been extensively studied (Kim et al., 2007; Lee et al., 2003; Phung et al., 2004), information about the microbial communities in a sulfide-fed MFC is not available yet. Therefore, in order to further elucidate the microbes involved and their roles in such a process, in the present work we examined the microbial community diversity and its catalytic activity in a sulfide-fed MFC anode. The microbial communities on the anode and in the solution were characterized by the 16S rRNA clone library analysis. 2. Materials and methods
471
sets (Suzuki and Giovannoni, 1996) with a GeneAmp PCR System PCT-200 (MJ Research Co., USA). The PCR products amplified using a 27f/1492r primer pair were cloned into the pGEM-T vector system (Promega Co., USA) and transformed into competent Escherichia coli TOP10 competent cells (TianGen Biotech. Co. Ltd., China). The transformants were plated on Luria-Bertani agar medium containing 5 g mL−1 ampicillin and 5-bromo-4-chloro3-indolyl--galactopyrano-side (X-gal). Ampicillin-resistant and -galactosidase-negative white clones were selected and transferred to a liquid medium of the same composition. Plasmids were then extracted from the colonies using the Plasmid Mini Kit (Huashun Inc., China). PCR amplification was performed twice according to the program. The first PCR amplification was performed with T7 forward (5 -TAA TAC GAC TCA CTA TAG GG-3 )/SP6 reverse (5 -ATT TAG GTG ACA CTA TAG AAT-3 ) primers. The second PCR amplification was performed using the amplicons of the first PCR as a DNA template with 27f forward/1492f reverse primers.
2.1. MFC construction and operation 2.4. Phylogenetic analysis The single-chamber MFC with air cathode as described previously (Sun et al., 2008) was used in this study (Fig. S1). The anodic chamber was filled with 350 mL of medium consisted of (in 1 L of 50 mM phosphate buffer, pH 7.0): NH4 Cl, 310 mg; KCl, 130 mg; Na2 CO3 , 750 mg; CaCl2 , 50 mg; MgCl2 ·6H2 O, 100 mg; NaCl, 10 mg; FeCl2 , 25 mg; CoCl2 ·2H2 O, 5 mg; MnCl2 ·4H2 O, 5 mg; AlCl3 , 2.5 mg; (NH4 )6 Mo7 O24 , 15 mg; H3 BO3 , 5 mg; NiCl2 ·6H2 O, 0.5 mg; CuCl2 ·2H2 O, 3.5 mg; ZnCl2 , 5 mg. The MFC was inoculated with 10 mL of sludge collected from an anaerobic reactor fed with wastewater rich in sodium sulfide. Sodium sulfide was added to a final concentration of 2 mM after the anodic chamber was purged with N2 to remove dissolved O2 . The fed-batch mode was adopted to incubate the active microbial communities. The MFC, designated as Reactor A, was operated with a fixed external resistance of 1 k. Electricity was generated in the first cycle and sodium sulfide was amended when the current density was less than 20 mA m−2 . Two identical MFCs were operated, with one for electrochemical evaluation and another for microbial community analysis. An abiotic reactor with a sterile anode, designated as Reactor B, was operated as control under the same conditions as Reactor A. All experiments were conducted in duplicate at 25 ◦ C. 2.2. Electrochemical analysis Voltage across a 1 k resistor was continuously recorded with an electrochemical workstation (660C, CH Instruments Inc., USA) and used to calculate the circuit current. Current density was then normalized by the projected surface areas of the two sides of anode (45 cm2 ). After the MFC was operated for half a year, polarization curves were obtained by applying a linear potential decrease of 1 mV s−1 from the open circuit voltage to 0 mV. 2.3. 16S rRNA clone library construction After the MFC had been operated for half a year, microbial analysis of the anode-attached and planktonic communities was performed. The anode was removed from the anodic chamber and rinsed with sterile distilled water to remove debris and loosely attached bacteria. Anodes were then cut and fragmented using sterile scissors. Planktonic bacteria were collected by centrifugating 10 mL of anode solution at 12,000 × g for 15 min at 4 ◦ C. Genomic DNA was extracted using FastDNA® SPIN Kit for Soil (MP BIO Qbiogene Inc., USA) according to the manufacturer’s instructions. The DNA was amplified using the 27f (5 -AGAGTTTGATCCTGGCTCAG3 ) and 1492r (5 -GGTTACCTTGTTACGACTT-3 ) universal primer
The clones were sequenced and analyzed in the Genbank database (http://www.ncbi.nlm.nih.gov) and Ribosomal Database Project II (RDP II, http://rdp.cme.msu.edu). All sequences were checked for chimeras using the CHECK CHIMERA program (http://rdp8.cme.msu.edu/cgis/chimera.cgi?su=SSU). Nonchimeric sequences were aligned with Clustal X v.1.8 (Thompson et al., 1997) and manual adjustment was performed with the alignment-visualizing tool of MEGA4 (Kumar et al., 2004). Operational taxonomic units (OTUs) from clone libraries were defined with Mothur v.1.6.0 program at a cutoff value of 0.03 (http://schloss.micro.umass.edu/wiki/Main Page). Phylogenetic trees were constructed with MEGA4 by using the UPGMA method and bootstrapping with 1000 repetitions. Mothur v.1.6.0 program was used to generate richness (Chao1 and ACE) (Chao, 1984; Chao and Lee, 1992) and diversity estimators (Mills and Wassel, 1980) for individual libraries. Clone libraries were compared using LIBSHUFF based on Jukes-Cantor pairwise distance matrices (Schloss et al., 2004). Nonparametric estimators of the fraction and richness of OTUs shared between two communities were analyzed using Mothur v.1.6.0 program. Partial 16S rRNA gene sequences have been deposited in the Genbank database under accession numbers FJ347713-FJ7719, GQ472933–GQ472959, and GU247454–GU247471.
3. Results 3.1. Enrichment of microbial community in sulfide-fed MFC The potential of sulfide as a substrate to generate electricity was examined by monitoring current evolution during the culture enrichment in the sulfide-fed MFC (Fig. 1A). After inoculation, electricity was generated immediately after sulfide dose. In the 1st feed cycle, the current density jumped to 75 mA m−2 , and then rapidly dropped to zero after 25 h. In the subsequent three cycles, the current densities maximized at ca. 75 mA m−2 and stabilized above 17 mA m−2 after 25-h cultivation. In the 5th and 6th cycles, the current density reached the same maximum level (78 mA m−2 ) as before, but declined more slowly. It took 60 h for the current density to decrease to 20 mA m−2 . The total charges produced by the MFC increased by 4 times from 4.9 C in the 1st cycle to 21.4 C in the 6th cycle. The polarization and power–current curves of this sulfide-fed MFC are shown in Fig. 1B. The maximum power density of 13 mW m−2 was achieved with a current density of 96 mA m−2 .
472
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Fig. 1. (A) Current generation in the sulfide-fed MFC enrichment and (B) polarization curves and power–current curves of the MFC after operated for half a year.
3.2. Electricity generation under diverse microbial catalysis After the MFC had been operated for half a year, an active consortium built up in the anodic chamber (Sun et al., 2009). In order to examine the microbial catalysis in sulfide oxidation, the MFC (Reactor A) and an abiotic reactor (Reactor B) were operated concurrently for comparison (Fig. 2A). In Reactor A, the maximum current density of 80 mA m−2 was obtained and it maintained at a level over 20 mA m−2 for 75 h. In Reactor B, the current density peaked at 74 mA m−2 and then decreased rapidly to nearly zero within 25 h. During 50 h of operation, Reactor A produced more than 3 times of charges than Reactor B, with their values of 17.8 and 5.1 C, respectively. These results suggest that sulfide oxidation could produce a higher and more persistent current through microbial catalysis. When the current density in Reactor A was below 20 mA m−2 , sulfide became exhausted (Sun et al., 2008). The biofilm-attached anode in Reactor A was replaced with a new electrode, but the solution with planktonic bacteria was retained (designated as Reactor C). The anode with biofilm built up, which was removed from Reactor A, was put into a reactor with fresh medium (designated as Reactor D). Current evolution in Reactors C and D were recorded after dose of 2 mM sulfide (Fig. 2B). As shown in Fig. 2B, in Reactor C, an instantaneous current of 80 mA m−2 was observed after sulfide dose, and this current sustained above 20 mA m−2 after 50 h of operation. In Reactor D, the current maximized at 74 mA m−2 and it decreased to 20 mA m−2 after 30 h of operation. In the 50-h operation, Reactors C and D produced 10.9 and 11.5 C of charges, respectively, which were higher than those produced by Reactor B, but lower than those by Reactor A. After 50 h of operation, the anodes in Reactors C and D were exchanged. Current in Reactor C was immediately restored to 68 mA m−2 with the active anodeattached and planktonic communities, even though no sulfide was amended. By contrast, the undeveloped electrode and medium
Fig. 2. Electricity generation in the: (A) Reactors A and B and (B) Reactors C and D in a batch test. At the 50th h (dashed line), the anodes in Reactors C and D were exchanged.
produced no current (Reactor D). Because sulfide can be quickly oxidized to other sulfur compounds via electrochemical reaction (Sun et al., 2009), the restored current in Reactor C probably resulted from the oxidation of sulfur compounds formed from the sulfide oxidation before anode exchange. 3.3. Phylogenetic distribution of sequences In order to further elucidate the microbial community structure of the sulfide-fed MFC, two clone libraries of the 16S rRNA
Fig. 3. Microbial community distribution on the anode and in the medium in the sulfide-fed MFC.
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
gene were constructed for the anode-attached and planktonic communities from Reactor A. One hundred positive clones for each sample were randomly selected for sequencing. Near full-length 16S rRNA gene sequences of 84 and 86 clones were obtained for the anode-attached and the planktonic population, respectively. The populations were measured by OTU abundance and referred in the
473
Genbank for their closest cultivated strains (Tables S1 and S2). The anode-attached and planktonic clone libraries were divided into 21 and 24 OTUs clusters by using a 3% cutoff criterion, respectively. The anode-attached population could be assigned to four groups, comprising ˛-Proteobacteria, ˇ-Proteobacteria, -Proteobacteria, and Firmicutes. The dominant genera on the anode were Pseudomonas
Fig. 4. Phylogenetic tree of bacteria community in the sulfur-fed MFC derived from 16S rRNA gene clone libraries. The clones with labels “♦” and “” are from the anode and medium, respectively. Bootstrap confidence levels (replicate 1000 times) greater than 50% are indicated at the nodes. Bar indicates 20% divergence. Methanobacterium curvum was used as an outgroup archeon.
474
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Fig. 5. Comparison of diversity parameters between the anode-attached and planktonic communities. Error bars indicate confidence levels of 95%. Coverage estimate is calculated as [1−(n/N)]×100, where n is the number of singleton clones, and N is the total number of clones.
sp. (24 of 84 clones) and Acinetobacter sp. (20 of 84 clones), with 28.6% and 23.8% abundance, respectively. In addition to the bacteria phylum above, ı-Proteobacteria, Actinobacteria, Gemmatimonadetes, and Bacteriodetes were also found in the reactor solution. The most abundant planktonic genera were Comamonas sp. (24 of 86 clones) and Acinetobacter sp. (22 of 86 clones), accounting for 27.9% and 25.6% abundance, respectively. The microbial community distributions on the anode and in the solution were compared (Fig. 3), and a phylogenetic tree was constructed to find out the relationship among these clone sequences (Fig. 4). The planktonic community showed a higher diversity than the anode-attached community. -Proteobacteria was present in much abundance both on the anode and in the solution. ˛-Proteobacteria was the second dominant phylum on the anode, whereas only few of them were found in the solution. ˇProteobacteria was the predominant phylum in the solution, while on the anode they are present as minor phylum. ı-Proteobacteria, Acinobacteria, Gemmatimonadetes, and Bacteroidetes were only found in the solution. 3.4. Diversity estimation and clone library comparison The richness and diversity of anode-attached and planktonic communities were compared (Fig. 5). The coverage of the clone libraries was 0.884 for anode and 0.893 for solution, suggesting that all major bacterial phyla present in the reactor were represented in the samples. The Simpson index of diversity for the anode-attached community was 0.087, while for the planktonic community this value was slightly lower, 0.080. The Shannon index showed a slightly greater diversity of the planktonic community (2.73) than the anode-attached bacteria (2.62). However, the rich-
ness estimates and diversity indices based on the 95% confidence indicate no differences between the two libraries. A LIBSHUFF analysis (Schloss et al., 2004) of the coverage curves was employed to compare the difference between phanktonic (X) and anode-attached communities (Y) (Fig. S2). The CXY and CYX were 0.0106 and 0.0204, respectively, and the P values were less than 0.0001. Because both P values were less than 0.025 (critical P value), the anode-attached and planktonic communities could be considered to be significantly different. Nonparametric estimators show the distribution of OTUs and Chao1 between two libraries in a Venn diagram (Fig. 6). The number in the overlapping region was the fraction of shared richness. Approximately 28.6% of OTUs (10 of 35) and 16.9% of Chao 1 (10.3333 of 61) were shared by the two communities. 4. Discussion MFC systems have been proven to be effective for simultaneous sulfide removal and electricity generation. A maximum current of 11 mA and power generation of 37 mW L-net anodic compartment−1 were obtained in a square-type MFC upon addition of 0.1 g L−1 sulfide (Rabaey et al., 2006). Our previous study demonstrated that sulfide could be electrochemically oxidized to generate instantaneous current, with polysulfides, polythionates, thiosulfate, sulfate and elemental sulfur as the final products (Sun et al., 2009). However, such a current did not sustain, but declined very rapidly. A higher and more sustainable current was obtained by microbial catalysis. As shown in Fig. 2A, the electricity output was enhanced through the microbial enrichment, suggesting that an active bacterial consortium capable of electricity production from sulfide oxidation was established in the MFC.
Fig. 6. A Venn diagram shows the shared OTU and Chao1 memberships derived from the anode-attached and planktonic clone libraries at a distance of 0.03.
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Electrochemical analysis shows that both the anode-attached and planktonic bacteria were responsible for the power generation in the sulfide-fed MFC. As shown in Fig. 2, the reactors with either anode-attached bacteria or planktonic bacteria produced more power than the abiotic control. This observation implies that both anode-attached and planktonic bacteria could work independently. This is different from the MFCs fed with organic acids, in which the anode-attached consortium is the main microbes responsible for the electricity production (Liu et al., 2005). The anode-attached and planktonic bacteria exhibited a synergistic interaction when they worked together. The highest power output was observed in the reactor with both anode-attached and planktonic bacteria, whereas the absence of either of them resulted in a reduced power output. Sulfide was incompletely oxidized when the anode-attached or planktonic community worked independently. The sulfide oxidation products in the process above could be further oxidized to generate electricity when the two communities were put together. Similar synergistic interaction has also been found in other studies. Rabaey et al. (2004) reported that, in a glucose-fed MFC, both anode-attached and planktonic bacteria contributed to the power output, with planktonic bacteria producing soluble electron shuttles for anode-attached bacteria. A further study confirmed that phenazine secreted by Pseudomonas aeruginosa facilitated the electron transfer by other bacterial species and thus improved electricity production of the MFC (Rabaey et al., 2005). In our MFC, the communities on the anode and in the solution shared most of the species found in the MFC. Universal richness and diversity estimation reveal no difference between the two communities. However, LIBSHUFF analysis shows that they had significantly different structures. The difference was further confirmed by comparing the bacterial species. For example, the clones of Rhodobacter sp. and Agrobacterium tumefaciens were only found on the anode, while Desulforhabdus amnigen and Thauera sp. were just discovered in the solution. Comamonas sp. was the dominant genus in the solution, whereas on the anode only a few bacteria belonged to this genus. In contrast, Pseudomonas sp. was the dominant genus on the anode. Such community differences suggest different microbial processes on the anode and in the solution, and also imply the possible synergistic interaction between the two communities. The bacterial communities in the sulfide-fed MFC share some similarities with those in the MFCs fed with carbohydrates. In our sulfide-fed MFC, -Proteobacteria was present in the highest abundance in the anodic chamber, with Acinetobacter sp. as the most numerous one. In some carbohydrates-fed MFCs, -Proteobacteria was also found to be the dominant species (Choo et al., 2006; Park et al., 2008), or occupied a great percentage of the community (Back et al., 2004; Lee et al., 2003). Pseudomonas sp. and Clostridium sp., both present on the anode and in the solution in our sulfide-fed MFC, were ever discovered in the carbohydrates-fed MFCs as well (Back et al., 2004; Lee et al., 2003). Pure cultures belonging to the species mentioned above have been isolated and demonstrated to be electrochemically active. Obligatory anaerobic bacterium, i.e., Clostridium butyricum EG3, utilized glucose with an anode as electron acceptor (Park et al., 2001). P. aeruginosa isolated from a glucose-fed MFC was able to produce endogenous chemical mediator pyocyanin, which was assumed to be responsible for electron transfer (Rabaey et al., 2004). Pseudomonas sp. and Clostridium sp. found in our sulfide-fed MFC were presumed to be electrochemically active and participate in the electron transfer in the sulfide oxidation. Sulfur-oxidizing and sulfate-reducing bacteria are two groups of bacteria participating in the global sulfur cycle. The sulfur-oxidizing bacteria can oxidize hydrogen sulfide, sulfur, sulfite, thiosulfate, and various polythionates under alkaline, neutral or acidic condi-
475
tions. Pseudomonas sp. and Rhodobacter sp., which were found in the sulfide-fed MFC, have been reported as the genera of sulfuroxidizing bacteria (Friedrich et al., 2001). The sulfate-reducing bacteria are a specialized group of anaerobic microbes that use sulfate as electron acceptor for respiration (Cooney et al., 1996). The sulfate-reducing bacterium, Desulfobulbus sp., has been demonstrated to contribute to power output in the seafloor fuel cell (Ryckelynck et al., 2005; Tender et al., 2002), as they have ability to oxidize S0 to SO4 2− with a graphite anode as electron acceptor (Holmes et al., 2004). Desulforhabdus sp. detected in the anodic chamber was the representatives of the sulfate-reducing bacteria (Muyzer and Stams, 2008). In the sulfide-fed MFC, Rhodobacter sp. was found only on the anode, and Pseudomonas sp. was in greater abundance on the anode than in the solution. Desulforhabdus sp. was only present in the solution. The bacterial species above may play a role in the electricity production from the sulfide oxidation. The fact that the sulfur-oxidizing bacteria were mainly located on the anode, whereas the sulfate-reducing bacteria grew preferentially in the solution, implies that the two groups of bacteria may perform different catalytic tasks. 5. Conclusions In the present study, we demonstrated that electrical power could be generated in an MFC using sulfide as the substrate. Community analysis of the sulfide-fed MFC showed a great diversity of bacteria in the anodic chamber, including the exoelectrogenic bacteria and sulfur-related bacteria. The anode-attached and planktonic communities shared similar richness and diversity, while their structures were significantly different according to the LIBSHUFF analysis. Synergistic association between the anode-attached and planktonic bacteria was proposed to play an important role in the electricity generation from the sulfide oxidation process in the MFC. A further study is warranted to elucidate how the microbes perform catalysis at different stages of sulfide oxidation in such a sulfide-fed MFC. Acknowledgements The authors wish to thank the NSFC (50625825), the NSFC-JST (21021140001), and the CAS (KSCX2-YW-G-001, KZCX2-YWQN504 and KJCX2-YW-H21-01) for the partial support of this study. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2010.07.074. References Back, J.H., Kim, M.S., Cho, H., Chang, I.S., Lee, J., Kim, K.S., Kim, B.H., Kim, Y.I., Han, Y.S., 2004. FEMS Microbiol. Lett. 238, 65–70. Chao, A., 1984. Scand. J. Stat. 11, 265–270. Chao, A., Lee, S.M., 1992. J. Am. Stat. Assoc. 87, 210–217. Choo, Y.F., Lee, J., Chang, I.S., Kim, B.H., 2006. J. Microbiol. Biotechnol. 16, 1481–1484. Cooney, M.J., Roschi, E., Marison, I.W., Comninellis, C., von Stockar, U., 1996. Enzyme Microb. Technol. 18, 358–365. Freguia, S., Rabaey, K., Yuan, Z., Keller, J., 2008. Environ. Sci. Technol. 42, 7937–7943. Friedrich, C.G., Rother, D., Bardischewsky, F., Quentmeier, A., Fischer, J., 2001. Appl. Environ. Microbiol. 67, 2873–2882. He, D., Kan, J.J., Mansfeld, F., Angenent, L.T., Nealson, K.H., 2009. Environ. Sci. Technol. 43, 1648–1654. Holmes, D.E., Bond, D.R., Lovley, D.R., 2004. Appl. Environ. Microbiol. 70, 1234–1237. Kim, J.R., Jung, S.H., Regan, J.M., Logan, B.E., 2007. Bioresour. Technol. 98, 2568–2577. Kumar, S., Tamura, K., Nei, N., 2004. Brief Bioinform. 5, 150–163. Lee, J., Phung, N.T., Chang, I.S., Kim, B.H., Sung, H.C., 2003. FEMS Microbiol. Lett. 223, 185–191. Liu, H., Cheng, S.A., Logan, B.E., 2005. Environ. Sci. Technol. 39, 658–662. Logan, B.E., 2007. Microbial Fuel Cells. Wiley, New York, pp. 48, 127. Logan, B.E., 2009. Nat. Rev. Microbiol. 7, 375–381.
476
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Mills, A.L., Wassel, R.A., 1980. Appl. Environ. Microbiol. 40, 578–586. Muyzer, G., Stams, A.J., 2008. Nat. Rev. Microbiol. 6, 441–454. Park, H.S., Kim, B.H., Kim, H.S., Kim, H.J., Kim, G.T., Kim, M., Chang, I.S., Park, Y.K., Chang, H.I., 2001. Anaerobe 7, 297–306. Park, H.I., Sanchez, D., Cho, S.K., Yun, M., 2008. Environ. Sci. Technol. 42, 6243–6249. Phung, N.T., Lee, J., Kang, K.H., Chang, I.S., Gadd, G.M., Kim, B.H., 2004. FEMS Microbiol. Lett. 233, 77–82. Rabaey, K., Boon, N., Hofte, M., Verstraete, W., 2005. Environ. Sci. Technol. 39, 3401–3408. Rabaey, K., Boon, N., Siciliano, S.D., Verhaege, M., Verstraete, W., 2004. Appl. Environ. Microbiol. 70, 5373–5382. Rabaey, K., van de Sompel, K., Maignien, L., Boon, N., Aelterman, P., Clauwaert, P., de Schamphelaire, L., Pham, H.T., Vermeulen, J., Verhaege, M., Lens, P., Verstraete, W., 2006. Environ. Sci. Technol. 40, 5218–5224. Ryckelynck, N., Stecher III, H.A., Reimers, C.E., 2005. Biogeochemistry 76, 113–139.
Schloss, P.D., Larget, B.R., Handelsman, J., 2004. Appl. Environ. Microbiol. 70, 5485–5492. Sun, M., Mu, Z.X., Chen, Y.P., Sheng, G.P., Liu, X.W., Chen, Y.Z., Zhao, Y., Wang, H.L., Yu, H.Q., Wei, L., Ma, F., 2009. Environ. Sci. Technol. 43, 3372–3377. Sun, M., Sheng, G.P., Zhang, L., Liu, X.W., Xia, C.R., Mu, Z.X., Wang, H.L., Yu, H.Q., Qi, R., Yu, T., Yang, M., 2008. Environ. Sci. Technol. 42, 8095–8100. Suzuki, M.T., Giovannoni, S.J., 1996. Appl. Environ. Microbiol. 62, 625–630. Tender, L.M., Reimers, C.E., Stecher III, H.A., Holmes, D.E., Bond, D.R., Lowy, D.A., Pilobello, K., Fertig, S.J., Lovley, D.R., 2002. Nat. Biotechnol. 20, 821–825. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, G.M., 1997. Nucleic Acids Res. 24, 4876–4882. Zuo, Y., Xing, D., Regan, J.M., Logan, B.E., 2008. Appl. Environ. Microbiol. 74, 3130–3137.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Microbial communities involved in electricity generation from sulfide oxidation in a microbial fuel cell Min Sun a , Zhong-Hua Tong a,∗ , Guo-Ping Sheng a , Yong-Zhen Chen a , Feng Zhang a , Zhe-Xuan Mu b , Hua-Lin Wang b , Raymond J. Zeng a , Xian-Wei Liu a , Han-Qing Yu a,∗ , Li Wei c , Fang Ma c a b c
Department of Chemistry, University of Science & Technology of China, 96 Jinzhai Road, Hefei 230026, China School of Chemical Engineering, Hefei University of Technology, Hefei 230092, China State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology, Harbin 150090, China
a r t i c l e
i n f o
Article history: Received 22 May 2010 Received in revised form 7 July 2010 Accepted 20 July 2010 Available online 27 July 2010 Keywords: Biocatalysis Microbial community diversity Microbial fuel cells Sulfate-reducing bacteria Sulfide oxidation Sulfur-oxidizing bacteria
a b s t r a c t Simultaneous electricity generation and sulfide removal can be achieved in a microbial fuel cell (MFC). In electricity harvesting from sulfide oxidation in such an MFC, various microbial communities are involved. It is essential to elucidate the microbial communities and their roles in the sulfide conversion and electricity generation. In this work, an MFC was constructed to enrich a microbial consortium, which could harvest electricity from sulfide oxidation. Electrochemical analysis demonstrated that microbial catalysis was involved in electricity output in the sulfide-fed MFC. The anode-attached and planktonic communities could perform catalysis independently, and synergistic interactions occurred when the two communities worked together. A 16S rRNA clone library analysis was employed to characterize the microbial communities in the MFC. The anode-attached and planktonic communities shared similar richness and diversity, while the LIBSHUFF analysis revealed that the two community structures were significantly different. The exoelectrogenic, sulfur-oxidizing and sulfate-reducing bacteria were found in the MFC anodic chamber. The discovery of these bacteria was consistent with the community characteristics for electricity generation from sulfide oxidation. The exoelectrogenic bacteria were found both on the anode and in the solution. The sulfur-oxidizing bacteria were present in greater abundance on the anode than in the solution, while the sulfate-reducing bacteria preferably lived in the solution. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Microbial fuel cells (MFC) provide a new approach for electricity generation from biomass. By utilizing microbial metabolism an MFC produces an electrical current from the degradation of organic/inorganic matters. Different types of exoelectrogenic bacteria have been isolated from naturally colonized anodes. However, many of these strains generate low power densities when grown as pure cultures (Rabaey et al., 2004; Zuo et al., 2008). It is therefore proposed some synergistic interactions might exist among members of an exoelectrogenic community (Logan, 2009). Molecular techniques have been widely applied to examine the composition of microbial communities in MFCs inoculated with mixed cultures (Kim et al., 2007; Lee et al., 2003; Phung et al., 2004). The communities in MFCs display great bacterial diversity, depending on the inoculum source or electron donors. An acetatefed system inoculated with activated sludge exhibited a near-even distribution of ˛-, -, and ı-Proteobacteria (Lee et al., 2003). In
∗ Corresponding authors. Fax: +86 551 3601592. E-mail addresses: [email protected] (Z.-H. Tong), [email protected] (H.-Q. Yu). 0956-5663/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2010.07.074
an ethanol-fed two-chamber MFC, 83% of cloned 16S rRNA gene sequences were related to ˇ-Proteobacteria (Kim et al., 2007). A ˇ-Proteobacteria-dominated MFC community was reported with a river sediment inoculum and river water feeding, while in another MFC fed with glucose and glutamate the predominant clones were ˛-Proteobacteria (Phung et al., 2004). Bacteria in MFCs can survive either through being electrochemically active or through interactions with other bacteria (Logan, 2007). A synergistic relationship between photosynthetic bacteria and heterotrophic exoelectrogenic bacteria was proposed to exist in a self-sustained phototrophic MFC (He et al., 2009). Glucose conversion in MFCs was demonstrated to be a complex process, in which different bacterial populations were involved. Hydrogen and acetate produced by fermentative bacteria could be used by exoelectrogenic bacteria for electricity generation (Freguia et al., 2008). The synergistic mechanism also plays an important role in electron transfer process. Lactobacillus amylovorus and Enterococcus faecium could make use of phenazine-based mediators produced by Pseudomonas sp. to transfer electrons (Rabaey et al., 2005). Wastewaters usually contain not only organics but also inorganic matters, such as sulfide. Sulfide is a hazardous substance and needs to be removed from wastewater before discharged
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
into environment. The MFC system was found to be effective for simultaneous sulfide removal and electricity generation (Rabaey et al., 2006). Recently, we have demonstrated that microbial catalysis played an important role in the electricity generation from a sulfide-fed MFC (Sun et al., 2009). Although the microbes involved in the oxidation of carbon-based compounds have been extensively studied (Kim et al., 2007; Lee et al., 2003; Phung et al., 2004), information about the microbial communities in a sulfide-fed MFC is not available yet. Therefore, in order to further elucidate the microbes involved and their roles in such a process, in the present work we examined the microbial community diversity and its catalytic activity in a sulfide-fed MFC anode. The microbial communities on the anode and in the solution were characterized by the 16S rRNA clone library analysis. 2. Materials and methods
471
sets (Suzuki and Giovannoni, 1996) with a GeneAmp PCR System PCT-200 (MJ Research Co., USA). The PCR products amplified using a 27f/1492r primer pair were cloned into the pGEM-T vector system (Promega Co., USA) and transformed into competent Escherichia coli TOP10 competent cells (TianGen Biotech. Co. Ltd., China). The transformants were plated on Luria-Bertani agar medium containing 5 g mL−1 ampicillin and 5-bromo-4-chloro3-indolyl--galactopyrano-side (X-gal). Ampicillin-resistant and -galactosidase-negative white clones were selected and transferred to a liquid medium of the same composition. Plasmids were then extracted from the colonies using the Plasmid Mini Kit (Huashun Inc., China). PCR amplification was performed twice according to the program. The first PCR amplification was performed with T7 forward (5 -TAA TAC GAC TCA CTA TAG GG-3 )/SP6 reverse (5 -ATT TAG GTG ACA CTA TAG AAT-3 ) primers. The second PCR amplification was performed using the amplicons of the first PCR as a DNA template with 27f forward/1492f reverse primers.
2.1. MFC construction and operation 2.4. Phylogenetic analysis The single-chamber MFC with air cathode as described previously (Sun et al., 2008) was used in this study (Fig. S1). The anodic chamber was filled with 350 mL of medium consisted of (in 1 L of 50 mM phosphate buffer, pH 7.0): NH4 Cl, 310 mg; KCl, 130 mg; Na2 CO3 , 750 mg; CaCl2 , 50 mg; MgCl2 ·6H2 O, 100 mg; NaCl, 10 mg; FeCl2 , 25 mg; CoCl2 ·2H2 O, 5 mg; MnCl2 ·4H2 O, 5 mg; AlCl3 , 2.5 mg; (NH4 )6 Mo7 O24 , 15 mg; H3 BO3 , 5 mg; NiCl2 ·6H2 O, 0.5 mg; CuCl2 ·2H2 O, 3.5 mg; ZnCl2 , 5 mg. The MFC was inoculated with 10 mL of sludge collected from an anaerobic reactor fed with wastewater rich in sodium sulfide. Sodium sulfide was added to a final concentration of 2 mM after the anodic chamber was purged with N2 to remove dissolved O2 . The fed-batch mode was adopted to incubate the active microbial communities. The MFC, designated as Reactor A, was operated with a fixed external resistance of 1 k. Electricity was generated in the first cycle and sodium sulfide was amended when the current density was less than 20 mA m−2 . Two identical MFCs were operated, with one for electrochemical evaluation and another for microbial community analysis. An abiotic reactor with a sterile anode, designated as Reactor B, was operated as control under the same conditions as Reactor A. All experiments were conducted in duplicate at 25 ◦ C. 2.2. Electrochemical analysis Voltage across a 1 k resistor was continuously recorded with an electrochemical workstation (660C, CH Instruments Inc., USA) and used to calculate the circuit current. Current density was then normalized by the projected surface areas of the two sides of anode (45 cm2 ). After the MFC was operated for half a year, polarization curves were obtained by applying a linear potential decrease of 1 mV s−1 from the open circuit voltage to 0 mV. 2.3. 16S rRNA clone library construction After the MFC had been operated for half a year, microbial analysis of the anode-attached and planktonic communities was performed. The anode was removed from the anodic chamber and rinsed with sterile distilled water to remove debris and loosely attached bacteria. Anodes were then cut and fragmented using sterile scissors. Planktonic bacteria were collected by centrifugating 10 mL of anode solution at 12,000 × g for 15 min at 4 ◦ C. Genomic DNA was extracted using FastDNA® SPIN Kit for Soil (MP BIO Qbiogene Inc., USA) according to the manufacturer’s instructions. The DNA was amplified using the 27f (5 -AGAGTTTGATCCTGGCTCAG3 ) and 1492r (5 -GGTTACCTTGTTACGACTT-3 ) universal primer
The clones were sequenced and analyzed in the Genbank database (http://www.ncbi.nlm.nih.gov) and Ribosomal Database Project II (RDP II, http://rdp.cme.msu.edu). All sequences were checked for chimeras using the CHECK CHIMERA program (http://rdp8.cme.msu.edu/cgis/chimera.cgi?su=SSU). Nonchimeric sequences were aligned with Clustal X v.1.8 (Thompson et al., 1997) and manual adjustment was performed with the alignment-visualizing tool of MEGA4 (Kumar et al., 2004). Operational taxonomic units (OTUs) from clone libraries were defined with Mothur v.1.6.0 program at a cutoff value of 0.03 (http://schloss.micro.umass.edu/wiki/Main Page). Phylogenetic trees were constructed with MEGA4 by using the UPGMA method and bootstrapping with 1000 repetitions. Mothur v.1.6.0 program was used to generate richness (Chao1 and ACE) (Chao, 1984; Chao and Lee, 1992) and diversity estimators (Mills and Wassel, 1980) for individual libraries. Clone libraries were compared using LIBSHUFF based on Jukes-Cantor pairwise distance matrices (Schloss et al., 2004). Nonparametric estimators of the fraction and richness of OTUs shared between two communities were analyzed using Mothur v.1.6.0 program. Partial 16S rRNA gene sequences have been deposited in the Genbank database under accession numbers FJ347713-FJ7719, GQ472933–GQ472959, and GU247454–GU247471.
3. Results 3.1. Enrichment of microbial community in sulfide-fed MFC The potential of sulfide as a substrate to generate electricity was examined by monitoring current evolution during the culture enrichment in the sulfide-fed MFC (Fig. 1A). After inoculation, electricity was generated immediately after sulfide dose. In the 1st feed cycle, the current density jumped to 75 mA m−2 , and then rapidly dropped to zero after 25 h. In the subsequent three cycles, the current densities maximized at ca. 75 mA m−2 and stabilized above 17 mA m−2 after 25-h cultivation. In the 5th and 6th cycles, the current density reached the same maximum level (78 mA m−2 ) as before, but declined more slowly. It took 60 h for the current density to decrease to 20 mA m−2 . The total charges produced by the MFC increased by 4 times from 4.9 C in the 1st cycle to 21.4 C in the 6th cycle. The polarization and power–current curves of this sulfide-fed MFC are shown in Fig. 1B. The maximum power density of 13 mW m−2 was achieved with a current density of 96 mA m−2 .
472
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Fig. 1. (A) Current generation in the sulfide-fed MFC enrichment and (B) polarization curves and power–current curves of the MFC after operated for half a year.
3.2. Electricity generation under diverse microbial catalysis After the MFC had been operated for half a year, an active consortium built up in the anodic chamber (Sun et al., 2009). In order to examine the microbial catalysis in sulfide oxidation, the MFC (Reactor A) and an abiotic reactor (Reactor B) were operated concurrently for comparison (Fig. 2A). In Reactor A, the maximum current density of 80 mA m−2 was obtained and it maintained at a level over 20 mA m−2 for 75 h. In Reactor B, the current density peaked at 74 mA m−2 and then decreased rapidly to nearly zero within 25 h. During 50 h of operation, Reactor A produced more than 3 times of charges than Reactor B, with their values of 17.8 and 5.1 C, respectively. These results suggest that sulfide oxidation could produce a higher and more persistent current through microbial catalysis. When the current density in Reactor A was below 20 mA m−2 , sulfide became exhausted (Sun et al., 2008). The biofilm-attached anode in Reactor A was replaced with a new electrode, but the solution with planktonic bacteria was retained (designated as Reactor C). The anode with biofilm built up, which was removed from Reactor A, was put into a reactor with fresh medium (designated as Reactor D). Current evolution in Reactors C and D were recorded after dose of 2 mM sulfide (Fig. 2B). As shown in Fig. 2B, in Reactor C, an instantaneous current of 80 mA m−2 was observed after sulfide dose, and this current sustained above 20 mA m−2 after 50 h of operation. In Reactor D, the current maximized at 74 mA m−2 and it decreased to 20 mA m−2 after 30 h of operation. In the 50-h operation, Reactors C and D produced 10.9 and 11.5 C of charges, respectively, which were higher than those produced by Reactor B, but lower than those by Reactor A. After 50 h of operation, the anodes in Reactors C and D were exchanged. Current in Reactor C was immediately restored to 68 mA m−2 with the active anodeattached and planktonic communities, even though no sulfide was amended. By contrast, the undeveloped electrode and medium
Fig. 2. Electricity generation in the: (A) Reactors A and B and (B) Reactors C and D in a batch test. At the 50th h (dashed line), the anodes in Reactors C and D were exchanged.
produced no current (Reactor D). Because sulfide can be quickly oxidized to other sulfur compounds via electrochemical reaction (Sun et al., 2009), the restored current in Reactor C probably resulted from the oxidation of sulfur compounds formed from the sulfide oxidation before anode exchange. 3.3. Phylogenetic distribution of sequences In order to further elucidate the microbial community structure of the sulfide-fed MFC, two clone libraries of the 16S rRNA
Fig. 3. Microbial community distribution on the anode and in the medium in the sulfide-fed MFC.
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
gene were constructed for the anode-attached and planktonic communities from Reactor A. One hundred positive clones for each sample were randomly selected for sequencing. Near full-length 16S rRNA gene sequences of 84 and 86 clones were obtained for the anode-attached and the planktonic population, respectively. The populations were measured by OTU abundance and referred in the
473
Genbank for their closest cultivated strains (Tables S1 and S2). The anode-attached and planktonic clone libraries were divided into 21 and 24 OTUs clusters by using a 3% cutoff criterion, respectively. The anode-attached population could be assigned to four groups, comprising ˛-Proteobacteria, ˇ-Proteobacteria, -Proteobacteria, and Firmicutes. The dominant genera on the anode were Pseudomonas
Fig. 4. Phylogenetic tree of bacteria community in the sulfur-fed MFC derived from 16S rRNA gene clone libraries. The clones with labels “♦” and “” are from the anode and medium, respectively. Bootstrap confidence levels (replicate 1000 times) greater than 50% are indicated at the nodes. Bar indicates 20% divergence. Methanobacterium curvum was used as an outgroup archeon.
474
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Fig. 5. Comparison of diversity parameters between the anode-attached and planktonic communities. Error bars indicate confidence levels of 95%. Coverage estimate is calculated as [1−(n/N)]×100, where n is the number of singleton clones, and N is the total number of clones.
sp. (24 of 84 clones) and Acinetobacter sp. (20 of 84 clones), with 28.6% and 23.8% abundance, respectively. In addition to the bacteria phylum above, ı-Proteobacteria, Actinobacteria, Gemmatimonadetes, and Bacteriodetes were also found in the reactor solution. The most abundant planktonic genera were Comamonas sp. (24 of 86 clones) and Acinetobacter sp. (22 of 86 clones), accounting for 27.9% and 25.6% abundance, respectively. The microbial community distributions on the anode and in the solution were compared (Fig. 3), and a phylogenetic tree was constructed to find out the relationship among these clone sequences (Fig. 4). The planktonic community showed a higher diversity than the anode-attached community. -Proteobacteria was present in much abundance both on the anode and in the solution. ˛-Proteobacteria was the second dominant phylum on the anode, whereas only few of them were found in the solution. ˇProteobacteria was the predominant phylum in the solution, while on the anode they are present as minor phylum. ı-Proteobacteria, Acinobacteria, Gemmatimonadetes, and Bacteroidetes were only found in the solution. 3.4. Diversity estimation and clone library comparison The richness and diversity of anode-attached and planktonic communities were compared (Fig. 5). The coverage of the clone libraries was 0.884 for anode and 0.893 for solution, suggesting that all major bacterial phyla present in the reactor were represented in the samples. The Simpson index of diversity for the anode-attached community was 0.087, while for the planktonic community this value was slightly lower, 0.080. The Shannon index showed a slightly greater diversity of the planktonic community (2.73) than the anode-attached bacteria (2.62). However, the rich-
ness estimates and diversity indices based on the 95% confidence indicate no differences between the two libraries. A LIBSHUFF analysis (Schloss et al., 2004) of the coverage curves was employed to compare the difference between phanktonic (X) and anode-attached communities (Y) (Fig. S2). The CXY and CYX were 0.0106 and 0.0204, respectively, and the P values were less than 0.0001. Because both P values were less than 0.025 (critical P value), the anode-attached and planktonic communities could be considered to be significantly different. Nonparametric estimators show the distribution of OTUs and Chao1 between two libraries in a Venn diagram (Fig. 6). The number in the overlapping region was the fraction of shared richness. Approximately 28.6% of OTUs (10 of 35) and 16.9% of Chao 1 (10.3333 of 61) were shared by the two communities. 4. Discussion MFC systems have been proven to be effective for simultaneous sulfide removal and electricity generation. A maximum current of 11 mA and power generation of 37 mW L-net anodic compartment−1 were obtained in a square-type MFC upon addition of 0.1 g L−1 sulfide (Rabaey et al., 2006). Our previous study demonstrated that sulfide could be electrochemically oxidized to generate instantaneous current, with polysulfides, polythionates, thiosulfate, sulfate and elemental sulfur as the final products (Sun et al., 2009). However, such a current did not sustain, but declined very rapidly. A higher and more sustainable current was obtained by microbial catalysis. As shown in Fig. 2A, the electricity output was enhanced through the microbial enrichment, suggesting that an active bacterial consortium capable of electricity production from sulfide oxidation was established in the MFC.
Fig. 6. A Venn diagram shows the shared OTU and Chao1 memberships derived from the anode-attached and planktonic clone libraries at a distance of 0.03.
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Electrochemical analysis shows that both the anode-attached and planktonic bacteria were responsible for the power generation in the sulfide-fed MFC. As shown in Fig. 2, the reactors with either anode-attached bacteria or planktonic bacteria produced more power than the abiotic control. This observation implies that both anode-attached and planktonic bacteria could work independently. This is different from the MFCs fed with organic acids, in which the anode-attached consortium is the main microbes responsible for the electricity production (Liu et al., 2005). The anode-attached and planktonic bacteria exhibited a synergistic interaction when they worked together. The highest power output was observed in the reactor with both anode-attached and planktonic bacteria, whereas the absence of either of them resulted in a reduced power output. Sulfide was incompletely oxidized when the anode-attached or planktonic community worked independently. The sulfide oxidation products in the process above could be further oxidized to generate electricity when the two communities were put together. Similar synergistic interaction has also been found in other studies. Rabaey et al. (2004) reported that, in a glucose-fed MFC, both anode-attached and planktonic bacteria contributed to the power output, with planktonic bacteria producing soluble electron shuttles for anode-attached bacteria. A further study confirmed that phenazine secreted by Pseudomonas aeruginosa facilitated the electron transfer by other bacterial species and thus improved electricity production of the MFC (Rabaey et al., 2005). In our MFC, the communities on the anode and in the solution shared most of the species found in the MFC. Universal richness and diversity estimation reveal no difference between the two communities. However, LIBSHUFF analysis shows that they had significantly different structures. The difference was further confirmed by comparing the bacterial species. For example, the clones of Rhodobacter sp. and Agrobacterium tumefaciens were only found on the anode, while Desulforhabdus amnigen and Thauera sp. were just discovered in the solution. Comamonas sp. was the dominant genus in the solution, whereas on the anode only a few bacteria belonged to this genus. In contrast, Pseudomonas sp. was the dominant genus on the anode. Such community differences suggest different microbial processes on the anode and in the solution, and also imply the possible synergistic interaction between the two communities. The bacterial communities in the sulfide-fed MFC share some similarities with those in the MFCs fed with carbohydrates. In our sulfide-fed MFC, -Proteobacteria was present in the highest abundance in the anodic chamber, with Acinetobacter sp. as the most numerous one. In some carbohydrates-fed MFCs, -Proteobacteria was also found to be the dominant species (Choo et al., 2006; Park et al., 2008), or occupied a great percentage of the community (Back et al., 2004; Lee et al., 2003). Pseudomonas sp. and Clostridium sp., both present on the anode and in the solution in our sulfide-fed MFC, were ever discovered in the carbohydrates-fed MFCs as well (Back et al., 2004; Lee et al., 2003). Pure cultures belonging to the species mentioned above have been isolated and demonstrated to be electrochemically active. Obligatory anaerobic bacterium, i.e., Clostridium butyricum EG3, utilized glucose with an anode as electron acceptor (Park et al., 2001). P. aeruginosa isolated from a glucose-fed MFC was able to produce endogenous chemical mediator pyocyanin, which was assumed to be responsible for electron transfer (Rabaey et al., 2004). Pseudomonas sp. and Clostridium sp. found in our sulfide-fed MFC were presumed to be electrochemically active and participate in the electron transfer in the sulfide oxidation. Sulfur-oxidizing and sulfate-reducing bacteria are two groups of bacteria participating in the global sulfur cycle. The sulfur-oxidizing bacteria can oxidize hydrogen sulfide, sulfur, sulfite, thiosulfate, and various polythionates under alkaline, neutral or acidic condi-
475
tions. Pseudomonas sp. and Rhodobacter sp., which were found in the sulfide-fed MFC, have been reported as the genera of sulfuroxidizing bacteria (Friedrich et al., 2001). The sulfate-reducing bacteria are a specialized group of anaerobic microbes that use sulfate as electron acceptor for respiration (Cooney et al., 1996). The sulfate-reducing bacterium, Desulfobulbus sp., has been demonstrated to contribute to power output in the seafloor fuel cell (Ryckelynck et al., 2005; Tender et al., 2002), as they have ability to oxidize S0 to SO4 2− with a graphite anode as electron acceptor (Holmes et al., 2004). Desulforhabdus sp. detected in the anodic chamber was the representatives of the sulfate-reducing bacteria (Muyzer and Stams, 2008). In the sulfide-fed MFC, Rhodobacter sp. was found only on the anode, and Pseudomonas sp. was in greater abundance on the anode than in the solution. Desulforhabdus sp. was only present in the solution. The bacterial species above may play a role in the electricity production from the sulfide oxidation. The fact that the sulfur-oxidizing bacteria were mainly located on the anode, whereas the sulfate-reducing bacteria grew preferentially in the solution, implies that the two groups of bacteria may perform different catalytic tasks. 5. Conclusions In the present study, we demonstrated that electrical power could be generated in an MFC using sulfide as the substrate. Community analysis of the sulfide-fed MFC showed a great diversity of bacteria in the anodic chamber, including the exoelectrogenic bacteria and sulfur-related bacteria. The anode-attached and planktonic communities shared similar richness and diversity, while their structures were significantly different according to the LIBSHUFF analysis. Synergistic association between the anode-attached and planktonic bacteria was proposed to play an important role in the electricity generation from the sulfide oxidation process in the MFC. A further study is warranted to elucidate how the microbes perform catalysis at different stages of sulfide oxidation in such a sulfide-fed MFC. Acknowledgements The authors wish to thank the NSFC (50625825), the NSFC-JST (21021140001), and the CAS (KSCX2-YW-G-001, KZCX2-YWQN504 and KJCX2-YW-H21-01) for the partial support of this study. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2010.07.074. References Back, J.H., Kim, M.S., Cho, H., Chang, I.S., Lee, J., Kim, K.S., Kim, B.H., Kim, Y.I., Han, Y.S., 2004. FEMS Microbiol. Lett. 238, 65–70. Chao, A., 1984. Scand. J. Stat. 11, 265–270. Chao, A., Lee, S.M., 1992. J. Am. Stat. Assoc. 87, 210–217. Choo, Y.F., Lee, J., Chang, I.S., Kim, B.H., 2006. J. Microbiol. Biotechnol. 16, 1481–1484. Cooney, M.J., Roschi, E., Marison, I.W., Comninellis, C., von Stockar, U., 1996. Enzyme Microb. Technol. 18, 358–365. Freguia, S., Rabaey, K., Yuan, Z., Keller, J., 2008. Environ. Sci. Technol. 42, 7937–7943. Friedrich, C.G., Rother, D., Bardischewsky, F., Quentmeier, A., Fischer, J., 2001. Appl. Environ. Microbiol. 67, 2873–2882. He, D., Kan, J.J., Mansfeld, F., Angenent, L.T., Nealson, K.H., 2009. Environ. Sci. Technol. 43, 1648–1654. Holmes, D.E., Bond, D.R., Lovley, D.R., 2004. Appl. Environ. Microbiol. 70, 1234–1237. Kim, J.R., Jung, S.H., Regan, J.M., Logan, B.E., 2007. Bioresour. Technol. 98, 2568–2577. Kumar, S., Tamura, K., Nei, N., 2004. Brief Bioinform. 5, 150–163. Lee, J., Phung, N.T., Chang, I.S., Kim, B.H., Sung, H.C., 2003. FEMS Microbiol. Lett. 223, 185–191. Liu, H., Cheng, S.A., Logan, B.E., 2005. Environ. Sci. Technol. 39, 658–662. Logan, B.E., 2007. Microbial Fuel Cells. Wiley, New York, pp. 48, 127. Logan, B.E., 2009. Nat. Rev. Microbiol. 7, 375–381.
476
M. Sun et al. / Biosensors and Bioelectronics 26 (2010) 470–476
Mills, A.L., Wassel, R.A., 1980. Appl. Environ. Microbiol. 40, 578–586. Muyzer, G., Stams, A.J., 2008. Nat. Rev. Microbiol. 6, 441–454. Park, H.S., Kim, B.H., Kim, H.S., Kim, H.J., Kim, G.T., Kim, M., Chang, I.S., Park, Y.K., Chang, H.I., 2001. Anaerobe 7, 297–306. Park, H.I., Sanchez, D., Cho, S.K., Yun, M., 2008. Environ. Sci. Technol. 42, 6243–6249. Phung, N.T., Lee, J., Kang, K.H., Chang, I.S., Gadd, G.M., Kim, B.H., 2004. FEMS Microbiol. Lett. 233, 77–82. Rabaey, K., Boon, N., Hofte, M., Verstraete, W., 2005. Environ. Sci. Technol. 39, 3401–3408. Rabaey, K., Boon, N., Siciliano, S.D., Verhaege, M., Verstraete, W., 2004. Appl. Environ. Microbiol. 70, 5373–5382. Rabaey, K., van de Sompel, K., Maignien, L., Boon, N., Aelterman, P., Clauwaert, P., de Schamphelaire, L., Pham, H.T., Vermeulen, J., Verhaege, M., Lens, P., Verstraete, W., 2006. Environ. Sci. Technol. 40, 5218–5224. Ryckelynck, N., Stecher III, H.A., Reimers, C.E., 2005. Biogeochemistry 76, 113–139.
Schloss, P.D., Larget, B.R., Handelsman, J., 2004. Appl. Environ. Microbiol. 70, 5485–5492. Sun, M., Mu, Z.X., Chen, Y.P., Sheng, G.P., Liu, X.W., Chen, Y.Z., Zhao, Y., Wang, H.L., Yu, H.Q., Wei, L., Ma, F., 2009. Environ. Sci. Technol. 43, 3372–3377. Sun, M., Sheng, G.P., Zhang, L., Liu, X.W., Xia, C.R., Mu, Z.X., Wang, H.L., Yu, H.Q., Qi, R., Yu, T., Yang, M., 2008. Environ. Sci. Technol. 42, 8095–8100. Suzuki, M.T., Giovannoni, S.J., 1996. Appl. Environ. Microbiol. 62, 625–630. Tender, L.M., Reimers, C.E., Stecher III, H.A., Holmes, D.E., Bond, D.R., Lowy, D.A., Pilobello, K., Fertig, S.J., Lovley, D.R., 2002. Nat. Biotechnol. 20, 821–825. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, G.M., 1997. Nucleic Acids Res. 24, 4876–4882. Zuo, Y., Xing, D., Regan, J.M., Logan, B.E., 2008. Appl. Environ. Microbiol. 74, 3130–3137.