F.H. Liu and S.B. Wang have contributed equally to this study.
Zhihua Zhou, Key Laboratory of Synthetic Biology, Institute of Plant Physiology and Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 300 Fenglin Road, Shanghai, 200032, P.R. China. E-mail: email@example.com
Aims: To identify the bacterial and archaeal composition in a mesophilic biogas digester treating pig manure and to compare the consistency of two 16S rDNA-based methods to investigate the microbial structure.
Methods and results: Sixty-nine bacterial operational taxonomic units (OTU) and 25 archaeal OTU were identified by sequencing two 16S rDNA clone libraries. Most bacterial OTU were identified as phyla of Firmicutes (47·2% of total clones), Bacteroides (35·4%) and Spirochaetes (13·2%). Methanoculleus bourgensis (29·0%), Methanosarcina barkeri (27·4%) and Methanospirillum hungatei (10·8%) were the dominant methanogens. Only 9% of bacterial and 20% of archaeal OTU matched cultured isolates at a similarity index of ≥97%. About 78% of the dominant bacterial (with abundance >3%) and 83% of archaeal OTU were recovered from the denaturing gradient gel electrophoresis (DGGE) bands of V3 regions in 16S rDNAs.
Conclusions: In the digester, most bacterial and archaeal species were uncultured; bacteria belonging to Firmicutes, Bacteroides and Spirochaetes seem to take charge of cellulolysis, proteolysis, acidogenesis, sulfur-reducing and homoacetogenesis; the most methanogens were typical hydrogenotrophic or hydrogenotrophic/aceticlastic; DGGE profiles reflected the dominant microbiota.
Significance and Impact of the Study: This study gave a first insight of the overall microbial structure in a rural biogas digester and also indicated DGGE was useful in displaying its dominant microbiota.
As the biggest developing country in the world, China has a population of 1·3 billion. Among these people, 72·3% of them are living in the rural areas. Because of the rapid growth of industry and energy consumption in the city, energy shortage becomes a big problem [The Priority Programme for China’s Agenda 21; see the World Wide Web (WWW) site: http://www.acca21.org.cn/indexe8.html]. Furthermore, the demands of energy in the rural areas also increase daily. On the other hand, there is a great deal of waste biomass produced in the countryside, including a large amount of lignocellulose byproducts in agriculture and excrement of poultry and cattle, which creates problems for the rural environment. To improve the environmental and living conditions in the countryside and to create a sustainable development in rural economy (Gupta 2003), it is necessary to resolve the energy and sanitary problems in the rural areas.
Anaerobic fermentation of waste biomass not only generates biogas fuel for cooking, lighting and heating, but also reduces waste biomass, potentially providing a mutually beneficial situation for the environmental, social and cycling-economic development in the rural areas. The technology for constructing biogas digesters at different scales and for different applications for treating rural wastes is well established (Chynoweth et al. 1999), while the efficiency of biogas production needs to be improved urgently. The composition of the microbial community in a biogas digester directly determines its efficiency and biogas yield. The process of anaerobic conversion of different biomasses to methane usually includes four steps: hydrolysis, acidogensis, acetogenesis and methanogenesis, in which hydrolytic, fermentative bacteria, acetogens and methanogens play distinct roles, respectively (Pretty et al. 2002; Angenent et al. 2004).
Culture-independent approaches, mainly 16S rDNA-based methods, e.g. cloning and sequencing, D/TGGE (denaturing/temperature gradient gel electrophoresis) fingerprinting, FISH (fluorescence in situ hybridization), RFLP [restriction fragment length polymorphism, or ARDRA (amplified rDNA restriction analysis)] and T-RFLP (terminal-RFLP), are applied to elucidate the diversity and composition of microbial communities in anaerobic methanogenic digesters widely used for the treatment of municipal and industrial wastewater (Karakashev et al. 2005; Connaughton et al. 2006; Mladenovska et al. 2006; Cirne et al. 2007; Klocke et al. 2007; Sousa et al. 2007a; Ye et al. 2007). Connaughton et al. (2006) and Sousa et al. (2007a) reported changes over time of the bacterial and archeal population in an anaerobic digester, the former compared the biomass composition with the activity of the digester and the latter compared the composition of microbiota in the presence and absence of long chain fatty acids. Ye et al. (2007) reported changes in digesters operated at several pH. Several studies involved investigation of microbial compositions in anaerobic methanogenic digesters treating rural waste (Mladenovska et al. 2006; Cirne et al. 2007; Klocke et al. 2007). By TGGE and cloning library and sequencing, Mladenovska et al. (2003) found that the most dominant methanogens in lab-scale anaerobic digesters with cattle manure or a mixture of cattle manure with glycerol trioleate were phylogenetically related to Methanosarcina siciliae. The bacterial and archaeal composition identified by T-RFLP analysis of 16S rRNA genes were found to be identical in two thermophilic continuously stirred tank reactors (CSTR), treating nontreated manure and pretreated manure for 40 min at 140°C, respectively (Mladenovska et al. 2006). However, to our knowledge, no previous studies characterized the overall microbial communities in a biogas digester treating rural waste.
Compared with the sequencing analysis of 16S rDNA genes in clone libraries, DGGE profiles of the V3 region was simpler and less costly for analysing the structural variation between different microbial systems or the spatio-temporal dynamics of the same system (Muyzer et al. 1993). It could also be useful in tracing the variation of dominant microbial organism in a biogas digester. However, previous studies have indicated that DGGE profiles were not consistent with the results from the sequencing of clone libraries in different microbial systems (Krave et al. 2002; Freeman et al. 2008; Wakase et al. 2008). It is important to ensure whether the results from DGGE profiles are consistent with clone library analyses before they are applied in biogas digesters.
For the purpose of both basic research and biogas biotechnology, there is considerable interest in elucidating the microbial composition and metabolic diversity involved in biogas production, as well as setting up an applied and less costly method to trace the variation of microbial structure in anaerobic bio-digesters. In this study, the overall microbial communities in a mesophilic anaerobic biogas digester were investigated by analysing the diversity of 16S rDNA using DGGE and sequence analysis. Such an investigation for the composition of the microbial community in the biogas digesters would be the first step to elucidate the relationship between the efficiency of biogas production in the digesters and the structure and variation of the microbiota. The purpose of this study was twofold: (i) to characterize the microbial diversity of an anaerobic biogas digester using culture-independent methods and (ii) to compare the discriminatory power of DGGE vs 16S library screening.
Materials and methods
A biogas slurry sample (3 l) was collected in June 2005 from a biogas digester in Qianwei Village, Chongming County, Shanghai, China, which was built in the early of 1980s and kept running for about 25 years. The digester has a volume of 600 m3 for anaerobic fermentation and produces 150 m3 biogas per day. The main fermentation substrate is pig manure from a nearby piggery breeding 1500 pigs. The digester can treat 1800 tonnes of pig manure from the piggery every year (Hao and Liu 2006).
To remove extracellular DNA and soluble organic contaminants, the fresh slurry samples (50 ml) were washed thrice in five volumes of TENP buffer [50 mmol l−1 Tris-HCl, 20 mmol l−1 EDTA, 100 mmol l−1 NaCl, 0·01 g ml−1 polyvinylpyrrolidone (PVP), pH 10], twice in five volumes of sterile phosphate-buffered saline (PBS buffer, 137 mmol l−1 NaCl, 2·7 mmol l−1 KCl, 1·5 mmol l−1 KH2PO4, 8·1 mmol l−1 Na2HPO4 in distilled water, pH 7·4) and vortexed for 5 min and centrifuged for 10 min at 10 000g. The washed cell pellets were resuspended in a suitable volume of sterile PBS buffer containing glycerol at a final concentration of 20%, divided into 4-ml aliquots, and stored at −70°C until nucleic acid extraction.
After thawing and centrifugation of subsamples (1 ml), the supernatants were removed. The genomic DNA was then extracted using QIAamp DNA stool mini kit (Qiagen, Heidelberg, Germany) according to the manufacturer’s instruction with minor modification. The treated sample was carefully disrupted by Cell Disruptor Genie (Scientific Industries Inc., New York) in the 2·0-ml microcentrifuge tube after adding the extraction buffer. The purified DNA was quantified with a Biophotometer (Eppendorf), and stored at −20°C until use.
Establishment of bacterial and archaeal 16S rDNA libraries
Bacterial and archaeal clone libraries were generated from polymerase chain reaction (PCR)-amplified 16S rDNA using bacterial primers 27f (5′-GAG AGT TTG ATC CTG GCT CAG-3′) and 1495r (5′-CTA CGG CTA CCT TGT TAC GA-3′) (Bianciotto et al. 1996) and the archaeal primers 1Af (5′-TCY GKT TGA TCC YGS CRG AG-3′) and 1100Ar (5′-TGG GTC TCG CTC GTT G-3′) (Embley et al. 1992). Reaction mixtures (25 μl) contained 2·5 μl of 10 × PCR buffer (TaKaRa Inc., Dalian, China), 0·2 mmol l−1 each of deoxyribonucleotide triphosphates (dNTP), 0·5 μmol l−1 of each primer, 1 U of Ex Taq DNA polymerase (TaKaRa Inc.) and 5 ng of template DNA. PCR reactions were performed on a Flexigene thermal cycler (Techne Flexigene, Cambridge, UK). The PCR reaction for archaeal 16S rDNA was performed using the following programme: initial denaturation for 3 min at 94°C; 30 cycles of denaturation (1 min at 94°C), annealing (1 min at 55°C) and extension (2 min at 72°C) with a final extension of 72°C for 10 min. The optimized PCR amplification conditions for bacterial 16S rDNA were as follows: initial denaturation at 95°C for 1·5 min; 5 cycles of 95°C for 30 s, 60°C for 30 s, 72°C for 1·5 min; 5 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 1·5 min; 15 cycles of 95°C for 30 s, 50°C for 30 s, 72°C for 1·5 min and a final extension of 72°C for 10 min.
To minimize PCR artefacts, ‘reconditioning PCR’ was performed as described by Thompson et al. (2002) after the initial amplification of the bacterial and archaeal 16S rDNA as described before. The initial PCR-amplified reaction was diluted 10-fold in a fresh reaction mixture of the same composition and cycled thrice using this programme. ssDNA and heteroduplex DNA could be minimized by adding excess primer during the ‘reconditioning PCR’ (Zhang et al. 2005).
Cloning and sequencing
Exactly 100 μl of bacterial and archaeal 16S rDNA reconditioning PCR products were electrophoresed on 1·0% agarose and the bands of the correct size (c. >1·5 kb for bacteria, and 1·1 kb for archaea) were purified using 3S PCR Product Purification Kit V2·0 (Shenergy Biocolor Biological Science & Technology Co., China), respectively. Finally, the purified product was cloned into the pMD18-T plasmid vector (TaKaRa Inc.) following the manufacturer’s instructions. The ligated products were transformed into Escherichia coli TOP10 competent cell (Invitrogen) with ampicillin and blue/white screening, and positive clones were arrayed in 96-well plates and stored at −80°C for long-term storage. Plasmid inserts were checked by PCR amplification using the M13 PCR set. Exactly 310 bacterial and 192 archaeal positive insert-containing clones were randomly selected for gene sequencing. The template DNA was prepared from overnight cultures of selected clones using an alkaline miniprep kit (Qiagen), sequencing were performed on an ABI 3730 DNA sequencer with Big Dye terminator chemistry as specified by the manufacturer (Applied Biosystems).
16S rDNA V3 region amplification
The V3 regions of bacterial and archaeal 16S rDNA from the biogas slurry DNA extract were amplified by PCR and the amplified products were used for analysis by DGGE. The bacterial primer sets EubacVf (5′-CGC CCG CCG CGC GCG GCG GGC GGG GCG GGG GCA CGG GGG GCC TAC GGG AGG CAG CAG-3′) and Vr (5′-ATT ACC GCG GCT GCT GG-3′), archaeal primer sets PARCH340f (5′-CCC TAC GGG GYG CAS CAG-3′) and PARCH519r (5′-TTA CCG CGG CKG CTG-3′), GC clamp and PCR amplification used were the same as described by Muyzer et al. (1993) and Øvreås et al. (1997). Reconditioning PCR products (200 μl) were concentrated with two volumes of ethanol and finally dissolved in 20 μl of water for further DGGE analysis. PCR product was observed on agarose gel (2·0%) with 1 × TAE buffer (40 mmol l−1 Tris-HCl, 40 mmol l−1 acetate, 1·0 mmol l−1 EDTA) and ethidium bromide (0·5 μg ml−1) under ultraviolet (UV) light.
Denaturing gradient gel electrophoresis
DGGE of the PCR products was performed by the method described by Muyzer et al. (1993) with the DCode Universal Mutation Detection system (Bio-Rad Laboratories, Hercules, CA < USA). Denaturing gradient gel [1 mm thickness × 160 × 160 mm; 1 × TAE (40 mmol l−1 Tris base with 1·0 mmol l−1 EDTA and 20 mmol l−1 sodium acetate at pH 7·4), 8% acrylamide-bisacrylamide (37·5 : 1), and 25–60% (35–70% for archaea, based on our unpublished results of perpendicular DGGE) denaturant (100% = 7 mol l−1 urea with 40% formamide)] was poured with a gradient delivery system (model 475; Bio-Rad Laboratories). Electrophoresis was performed at a constant temperature of 60°C, first for 10 min at 25 V and then for 5 h at 200 V in 1 × TAE buffer. After the electrophoresis, the gels were stained with AgNO3 as described by the manufacturer.
Clone library construction of single DGGE bands
Each gel slice that contained an obvious DNA band was excised with a clean razor blade and placed in an 1·5-ml Eppendorf tube. The gel slice was crushed and incubated with 50 μl of TE buffer at 4°C overnight. The 3-μl supernatant was subjected to a second PCR under the same conditions as described before. The re-amplified PCR products were examined by DGGE to confirm that single bands were present at the same positions. The PCR products were then purified with Mini-DNA Rapid Purification Kit (BioDev, Beijing, China) and cloned into the pMD18-T Vector (TaKaRa Inc.) to construct the clone libraries. Five clones were picked from each library and sent to Invitrogen (Shanghai, China) for sequencing.
Sequences were edited manually to remove vector and ambiguous sequences at the ends by scanning of the individual chromatograms using Chromas software ver.2.23 (Technelysium, Shanghai, China). Chimeras were checked by the CHIMERA_CHECK programme (Cole et al. 2003) in Ribosomal Database Project (RDP) at first, and then were further firmed by Bellerophon programme on the Greengenes website (DeSantis et al. 2006). All reference sequences were obtained from the GenBank and RDP. Then all the sequences and their closest relatives were fitted into an alignment using the automated tools of the ARB software package (Ludwig et al. 2004). Aligned sequences were added to the ARB neighbour-joining tree (based on pairwise distances with Olsen correction) with the parsimony insertion tool as described by Ley et al. (2005). Sequences with internal regions of poor quality leading to alignment problems were excluded from further analysis. Dotur (Schloss and Handelsman 2005) was used to cluster sequences into operational taxonomic units (out) by % pairwise identity (%ID, using a furthest-neighbour algorithm and a precision of 0·03). The stability of tree branches was assessed by the bootstrap method using 1000 replicates.
Microbial community analysed by clone library-bacterial community
The 16S rDNA genes were amplified from the total DNA extracted from the biogas slurry sample with a bacterial primer set 27f/1495r and amplicons were ligated to pMD-18 T vector to construct a library. In total, 310 clones were randomly selected and sequenced. After eliminating low-quality (68 clones) sequences and chimeric sequences (53 clones), 189 sequences were used for the following analyses. The coverage of the library was 81·0%, indicating that the library was large enough for further analyses. The 189 sequences of 16S rDNA genes were classified into 69 OTU (Table 1). The abundances of all OTU in the library were less than 7%, where only nine OTU were more than 3%. The accession number, sources and described functions of their phylogenetically closest matched organisms are also listed in Table 1.
Table 1. Taxonomic relationship of bacterial 16S rDNA sequences from Chongming biogas digester compared (Blast) with public databases (RDP, Greengenes and NCBI)
% of total
Phylogenetically most closely related organism (accession no.)
Only 14 of the 69 OTU in the bacterial library were matched to the closest related known sequences deposited in NCBI and RDP at a similarity index of more than 97%, which was regarded as an experiential index for differentiating species (Table 1). Twenty-nine of the 69 OTU were matched to the closest related known sequences deposited in the databases at a similarity index of between 90% and 96%, while 26 of the 69 OTU were matched to the closest related known sequences at a similarity index of between 80% and 90%. Nearly 80% of the bacteria in this digester may be new, previously undescribed species. Even among the 14 OTU most closely related to known sequences, only 6 OTU matched with the cultured bacterial strains, which were Leuconostoc citreum, Clostridium quinii, Clostridium chartatabidum, Streptococcus alactolyticus, Lactobacillus reuteri and Clostridium sp. The remaining eight OTU were from uncultured bacteria. These results indicated that more than 91% bacteria in this digester were uncultured.
Within the 69 OTU, 41 were classified as Firmicutes, 16 as Bacteroides and 8 as Spirochaetes. In a phylogenetic tree (Fig. 1), 24 OTU of Firmicutes were clustered with the pure cultures belonging to Clostridia, the most of which belonged to the family Clostridiaceae. Three Firmicutes OTU were assigned to different families of Bacilli. All of the 16 Bacteroides OTU clustered together and divided into four subgroups, belonging to family Porphyromonadaceae, Rikenellaceae and two uncultured Bacteroidetes. In the eight Spirochaetes OTU, two clustered together with genus Treponema, five with genus Spirochaeta and the remaining one with uncultured Leptospiracheae. The remaining four sequences were grouped together with uncultured Fibrobacteres, Xanthomonas vasicola, uncultured planctomycete and uncultured Verrucomicrobia, respectively.
The phylogenetically most closely matched bacteria were mostly detected from the intestine of pig or other animals, waste-water treatment plant (sludge or biofilm), anaerobic reactor or digester or landfill leachate and compost, which were all related to anaerobic fermentation (Table 1). The matching micro-organisms of most OTU belonged to phylum Firmicutes, Bacteroides and Spirochaetes, which were fermentative acidogens. Several OTU from Firmicutes were cellulolytic and those mainly from Bacteroides were proteolytic. The most closely matched micro-organisms of three OTU from Firmicutes were homoacetogens and that of another one was the sulfur-reducing bacteria (Table 1).
Microbial community analysed by clone library-archaeal community
The 16S rDNA genes amplified from the total DNA extracted from biogas slurry sample with a specific archaeal primer set of 1Af/1100Ar were ligated into the pMD-18 T vector to construct a library. Exactly 186 clones were used for the following analyses and the length of the amplified genes was approximately 1·1 kbp. All of the sequences were classified into 25 OTU (Table 2). The coverage of the library was 94·1%. OTU AS01, AS02 and AS04 accounted for 27·4%, 29·0% and 10·8% of the sequenced clones in the archaeal library, respectively.
Table 2. Taxonomic relationship of archaeal 16S rDNA sequences from Chongming biogas digester compared (Blast) with public databases (RDP, Greengenes and NCBI)
% of total
Phylogenetically most closely related organism (accession no.)
Only 6 of the 25 OTU had matches to most closely related sequences in the databases at a similarity more than 97% (Table 2), among which 5 OTU were matched to cultured isolates. Eighteen of the twenty-five OTU were matched to the closest related known sequences at a similarity index of between 90% and 96%; the remainders were matched to the closest related known sequences at a similarity index of 89%. More than 80·0% of the archaeal OTU in this digester may be uncultured.
All of the archaeal OTU were classified into Methanomicrobia of phylum Euryarchaeota and assigned to two branches: Methanomicrobiales and Methanosarcinales (Fig. 2). In branch Methanomicrobiales, four OTU were clustered with two isolates from genus of Methanogenium, one OTU was grouped to genus Methanoculleus, five to genus Methanospirillum and the remaining seven were divided into four different sub-branches, which might represent new genus or new families. Within the Methanosarcinales branch, three OTU clustered with genus Methanosarcina, one OTU belonged to genus Methanosaeta and the remaining four (AS10, AS14, AS15 and AS16) were clustered together but separated from genus Methanosarcina. It suggested that OTU AS10, AS14, AS15 and AS16 might represent a new genus of Methanosarcinaceae. However, the three most abundant OTU, AS02 (29·0%), AS01 (27·4%) and AS04 (10·8%), comprising more than 67% clones in the library, matched pure cultures of Methanoculleus bourgensis, Methanosarcina barkeri and Methanospirillum hungatei at a similarity index of 97%, respectively. This indicates that the four methanogens were the dominant archaeal species in this biogas digester.
The phylogenetically assigned archaea OTU were divided into three functional groups. Most OTU, including the most and the third most abundant OTU, were hydrogenotrophic methanogens (Table 2). The most closely matched archaea to five OTU, including the second most abundant OTU, were hydrogenotrophic/aceticlastic methanogens, and one OTU was an aceticlastic methanogen.
Structure of the dominant microbiota in the digester analysed by DGGE fingerprinting and sequencing the DGGE bands
The structures of the dominant bacteria and archaea in the slurry sample were analysed by DGGE fingerprinting (Fig. 3). Twenty-four detected bands were found in the bacterial DGGE profile (CB1–CB24) and nine bands (CA1–CA9) in the archaeal profile (Fig. 3). These results demonstrated a higher diversity of bacteria than archaea in the digester.
All of the detected bands from bacterial and archaeal DGGE fingerprinting were excised, amplified and cloned. The mobility of the inserted fragments of three clones randomly selected from each clone library (33 libraries in total) were checked via DGGE and compared with the original DGGE pattern. Clones of the inserts that migrated to the same locations as the original bands in the DGGE profile were sequenced. In total, 34 different sequences from the bacterial libraries and 11 sequences from archaeal libraries were obtained (Tables 3 and 4). Multi-fragments were found in a single band in at least 12 of the 33 recovered DGGE bands.
Table 3. Bacterial 16S rDNA sequence similarities of excised bands that appear in Fig. 3
Phylogenetically most closely related organism (accession no.)
Corresponding OTU no.
Sm1, similarity between the sequences of denaturing gradient gel electrophoresis (DGGE) band and its phylogenetically closely related organism.
Sm2, similarity between the sequences of DGGE band and its corresponding operational taxonomic unit (OTU).
Nineteen of the bacterial sequences were assigned to Firmicutes, ten of the sequences were clustered to Bacteroidetes, three were grouped to Spirochaetes and one to Proteobacteria and Verrucomicrobiales, respectively (Table 3). Seven of the 34 bacterial sequences matched one isolate at a similarity index of ≥97%, the remainder matched isolates at a similarity index of <97% (Table 3), indicating that nearly 80% of the dominant bacteria in the biogas digester were unclassified.
In the DGGE profile, most of the fragments representing V3 regions of phylum Firmicutes were found to melt at high denaturant concentration areas, and that of phylum Bacteroidetes tended to melt at low denaturant concentration areas (Table 3 and Fig. 3).
Three archaeal sequences were assigned to genera Methanoculleus and Methanosarcina, respectively; one belonged to genus Methanospirillum, three grouped to order Methanomicrobiales and the remaining was grouped to order Methanosarcinales (Table 4). About half of the sequences were matched to uncultured taxa.
The consistency of dominant bacterial and archaeal composition revealed by clone library and DGGE profile analysis
Eighteen bacterial sequences representing 16 of the 24 recovered DGGE bands were also present in the clone library. On the other hand, seven of the nine OTU with abundance >3% were matched with the V3 sequences recovered from the DGGE bands (Table 3). This means that 78% dominant OTU (with abundance >3%) in the bacterial clone library could be also detected by DGGE profile analysis.
Ten archaeal sequences representing all of the nine recovered DGGE bands were in the clone library. Five of the dominant six archaeal OTU with abundance more than 3% were found to match with the V3 sequences recovered from DGGE bands. Exactly 83% of the dominant OTU (with abundance >3%) in the archaeal library were observed in the DGGE profile.
Cloning and sequencing the full length of 16S rDNA has been frequently applied to elucidate the exact composition of a microbial community. However, DGGE profiles of the V3 regions can be employed to reflect the dominant community structure and to analyse the structural variation between different systems or the dynamics of the same system (Muyzer et al. 1993). In this study, sequencing of V3 regions from recovered DGGE bands and 16S rDNA from cloning libraries were both used to analyse the microbial community in an anaerobic digester. Both analyses indicated that the community was mainly composed of phyla Firmicutes, Bacteroides and Spirochaetes for bacteria, and orders Methanomicrobiales and Methanosarcinales for archaea. The analyses also indicated that the diversity of bacteria was higher than that of achaera. Importantly, 78% of dominant OTU (with abundance >3%) in the bacterial library and 83% of dominant OTU in the archaeal library could be detected in the DGGE profile. These results indicate that the DGGE profile in this study clearly reflected the dominant composition of the microbial community in the Qianwei biogas digester. Additionally, the major components of this bacterial community could be separately located on the different areas in the DGGE profile, e.g. Firmicutes tended to appear at the high denaturant concentration area and Bacteroidetes at low denaturant concentration areas in the DGGE profile. The analysis of DGGE profiles may be useful in displaying the dominant microbial composition in biogas digesters in our further studies. However, more OTU were identified by sequencing 16S rDNA libraries, indicating that DGGE may underperform in elucidating the diversity or the exact composition of a complex microbial community.
Anaerobic digesters are widely used to treat different wastes, e.g. brewery and pulp industry wastewater containing different carbohydrates, long-chain fatty acids, volatile fatty acids, methanethiol, terephthalate (Mata-Alvarez et al. 2000; Yadvika et al. 2004). Many studies focus on analyses of microbial community in such anaerobic digesters by using different approaches (Cotta et al. 2003; Snell-Castro et al. 2005; Peu et al. 2006). In these studies, the microbial structures in these systems varied greatly owing to the difference of the substrates used. Firmicutes, Nitrospira and Deferribacteres were found to be the predominant bacteria and Methanosaeta concilii was the dominant methanogenic archaea in an anaerobic digester treating wastewater from a beer brewery (Diaz et al. 2006). In an anaerobic digester treating long-chain fatty acids, species of Syntrophomonadaceae and Syntrophobacteraceae families, which oxidize fatty acids, were the predominant bacteria (Sousa et al. 2007b). When degrading methanethiol from paper mill wastewater, methylotrophic methanogens Methanomethylovorans hollandica were enriched (de Bok et al. 2006). In a laboratory methanogenic digester amended with glucose, Spirochaetes-, eubacterium- and propionibacterium-like bacteria were found to be dominant (Fernandez et al. 1999, 2000). Pig manure, which was mainly comprised of undigested biomass and some fatty acids, such as acetic acid and propionic acid, was the sole substrate in Qianwei biogas digester analysed in this study. Firmicutes (47·2%), Bacteroides (35·4%) and Spirochaetes (13·2%) were found to be the three most abundant bacterial phyla in this study. Within phylum Firmicutes, class Clostridia was the most dominant of the bacterial community (45·5% of the clones).
Several other studies have shown that within the bacterial and archaeal community of a pig manure slurry and a manure storage pit, Eubacterium, Clostidium, Bacillus–Lactobacillus–Streptococcus subdivision, Mycoplasma and the Flexibacter–Cytophaga–Bacteroides were the main components of the bacterial communities (Snell-Castro et al. 2005; Peu et al. 2006). Hydrogenotrophic methanogens, such as Methanoculleus, Methanogenium and Methanobrevibacter, dominated the archaeal communities (Whitehead and Cotta 1999; Tang et al. 2004; Hori et al. 2006). This agrees with the results from our digester, suggesting most archaea in our digester might originate from pig manure.
Most phylogenetically closely matched bacteria to the OTU identified in the biogas digester were found in anaerobic environments, such as the guts of animal or insects, sediments, anaerobic digesters and faeces (Table 1). Many phylogenetically closest matched taxa were uncultured organisms or function-unidentified organisms. The metabolic functions of their related OTU in the biogas digester were unknown. Among the function-identified bacteria, most were acidogenic, producing H2, CO2, formate, acetate and other fatty acids as well as a small amount of ethanol from cellobiose, d-fructose, N-acetylglucosamine, d-glucose, maltose, mannose and saccharose (Vandamme et al. 1999); several OTU related to the Clostridium sp. might be cellulolytic and the other three ones were related to proteolytic, both of which might take charge of decomposing polymers in the pig manure to monomers; Only three OTU might be homoacetogenic. This seems to be consistent with the fact that most archaeal OTU or the most abundant OTU were hydrogenotrophic methanogens, and only one OTU at low frequency was identified as the aceticlastic methanogen (Table 2).
The phylum Euryarchaeota was the major methanogenic archaeal group in anaerobic fermentation environments. Methanoculleus bourgensis, Methanosarcina barkeri, Methanospirillum hungatei and Methanomicrobiales archaeon were the most abundant methanogenic species in our digester (Table 2). Each of them showed some specific characteristics in methanogenic metabolism. Methanoculleus bourgensis was reported to use H2–CO2 or formate as a substrate for growth and methanogenesis, and is a hydrogenotrophic methanogen (Blotevogel et al. 1992). Methanospirillum hungatei produces methane only from H2–CO2 or formate, but not from acetate or ethanol and methanol, being a strictly hydrogenotrophic methanogen (Ferry et al. 1974). Methanosarcina barkeri could be used in different substrates to produce methane, including H2–CO2, methanol, mono-, di- and trimethylamines, acetate and CO (Bryant and Boone 1987), and is a hydrogenotrophic or aceticlastic methanogen.
Chimeric sequences of the full length of 16S rDNA were usually found while analysing the compositions of complex microbial communities (Ashelford et al. 2006). Therefore, several programmes were designed to identify chimeric sequences (Cole et al. 2003; DeSantis et al. 2006). Ashelford et al. (2006) reported that the average chance to falsely identify a sequence as chimeric by the Bellerophon programme was 7·2%. Fifty-three 16S rDNA sequences from bacteria were identified to be putatively chimeric. To ensure most of them to be assigned rightly, the putative chimeric sequences were confirmed by two different programmes, and some of them were further confirmed by PCR amplification. However, it was still possible that some of the 53 sequences be assigned as chimeric by wrong, and thus might underestimate the bacterial diversity in the biogas digester.
High diversities of microbial composition and metabolism were found in this microbiota in this study. It provides a pool of functional micro-organisms involved in biomass transformation. However, most micro-organisms, more than 91% bacteria and 80% of archaea, in this digester were uncultured. Presently it is impossible to elucidate the total metabolic process in the biogas digester using only the analysis of microbial composition. The recent application of metagenomic techniques suggest they may provide a new approach to obtain function genes related to biomass transformation within the assemblage. A lab anaerobic fermentation system and its control system are necessary in elucidating the relationship between the biogas-producing efficiency and microbiota composition under variable substrate conditions, with the addition of an inhibitor or accelerant, and/or adding specific micro-organisms for bio-augmentation.
The authors are very grateful for the technical support from Dr Huirong Li at Key Laboratory for Polar Science of State Oceanic Administration, Polar Research Institute of China. They thank Dr Yongping Huang and Prof. Xiangxiong Zhu for their critical reading of the manuscript. They also thank Dr Hewson at the Department of Ocean Sciences, University of California Santa Cruz, for language polishing. This work was financially supported by Knowledge Innovation Key Program of the Chinese Academy of Sciences project (no. KSCX2-YW-G-022), the National High Technology Research and Development Key Program of China (863 Key Program, grant no. 2007AA021302) and the National Natural Science Foundation of China (grant no. 30700017).