Comparative Metagenomic Analysis of Microcosm Structures and Lignocellulolytic Enzyme Systems of Symbiotic Biomass-Degrading Consortia

Comparative metagenomic analysis of microcosm structures and lignocellulolytic enzyme systems of symbiotic biomass-degrading consortia

Sarunyou Wongwilaiwalin & Thanaporn Laothanachareon & Wuttichai Mhuantong & Sithichoke Tangphatsornruang & Lily Eurwilaichitr & Yasuo Igarashi & Verawat Champreda

Received: 2 October 2012 /Revised: 3 January 2013 /Accepted: 7 January 2013 # Springer-Verlag Berlin Heidelberg 2013

Abstract Decomposition of lignocelluloses by cooperative industrial pulp waste with CMCase, xylanase, and β- microbial actions is an essential process of carbon cycling in glucanase activities in the supernatant. Shotgun pyrosequenc- nature and provides a basis for biomass conversion to fuels ing of the BGC-1 metagenome indicated a markedly high and chemicals in biorefineries. In this study, structurally stable relative abundance of genes encoding for glycosyl hydrolases, symbiotic aero-tolerant lignocellulose-degrading microbial particularly for lignocellulytic enzymes in 26 families. The consortia were obtained from biodiversified microflora pres- enzyme system comprised a unique composition of main- ent in industrial sugarcane bagasse pile (BGC-1), cow rumen chain degrading and side-chain processing hydrolases, domi- fluid (CRC-1), and pulp mill activated sludge (ASC-1) by nated by GH2, 3, 5, 9, 10, and 43, reflecting adaptation of successive subcultivation on rice straw under facultative an- enzyme profiles to the specific substrate. Gene mapping oxic conditions. Tagged 16S rRNA gene pyrosequencing showed metabolic potential of BGC-1 for conversion of bio- revealed that all isolated consortia originated from highly mass sugars to various fermentation products of industrial diverse environmental microflora shared similar composite importance. The symbiotic consortium is a promising simpli- phylum profiles comprising mainly Firmicutes, reflecting fied model for study of multispecies mechanisms on consol- convergent adaptation of microcosm structures, however, with idated bioprocessing and a platform for discovering efficient substantial differences at refined genus level. BGC-1 compris- synergistic enzyme systems for biotechnological application. ing cellulolytic Clostridium and Acetanaerobacterium in stable coexistence with ligninolytic Ureibacillus showed the Keywords Biorefinery . Glycosyl hydrolase . highest capability on degradation of agricultural residues and Lignocellulose . Metagenome . Microbial consortium

S. Wongwilaiwalin : T. Laothanachareon : W. Mhuantong : L. Eurwilaichitr : V. Champreda (*) Introduction Enzyme Technology Laboratory, Bioresources Technology Unit, National Center for Genetic Engineering and Biotechnology, Lignocellulosic plant biomass is the most abundant organic 113 Thailand Science Park, Klong Luang, Pathum Thani 12120, Thailand carbon reservoir on earth, providing a renewable carbon e-mail: [email protected] source for the chemosphere and the prospective biorefinery industry. Development of carbon neutral biomass conver- S. Tangphatsornruang sion processes has attracted much public interest in recent Genome Institute, National Center for Genetic Engineering and Biotechnology, 113 Thailand Science Park, years as a promising alternative platform for production of Klong Luang, Pathum Thani 12120, Thailand fuels and chemicals, which could reduce the global dependence on petroleum (FitzPatrick et al. 2010). Extensive Y. Igarashi research has been conducted to increase our understanding Department of Biotechnology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, of the complex natural bioconversion of lignocelluloses, Tokyo 113-8657, Japan particularly on degradation of recalcitrant plant biomass,

Reproduced from Appl. Microbiol. Biotechnol. DOI 10.1007/s00253-013-4699-y.

163 which is an essential starting step in biorefinery. Insights Culture-independent direct 454 pyrosequencing of envi- into biomass decomposition by means of intricate biological ronmental metagenomes has previously been applied to processes not only provides us a fundamental basis on study diversity and metabolic capabilities of natural biomass essential roles of microbes in maintaining carbon balance decay microbiomes in various environments, e.g., tropical of the eco-system but is also of great interest for future peat swamp forest (Kanokratana et al. 2010), termite gut establishment of viable biorefineries for our sustainable (Warnecke et al. 2007; Nimchua et al. 2011), cow rumen society. (Brulc et al. 2009), and switch grass-adapted compost com- Lignocellulosic materials consist of mainly three different munity (Allgaier et al. 2010). These works illustrate the types of polymers, i.e., cellulose, hemicelluloses, and lignin, highly diversified nature of unexplored plant biomass- which are associated with each other (Fengel and Wegener degrading microbes and their enzymes awaiting discovery 1984). In nature, lignocelluloses are degraded by a complex for biotechnological uses. In this study, comparative analy- multistep process involving cooperative actions of various sis of structurally stable cellulolytic consortia isolated from microorganisms producing a variety of cellulolytic, hemicel- biodiversified microcosms in various environments active in lulolytic, and ligninolytic enzymes that act specifically and plant biomass deconstruction was reported using tagged 16S synergistically to degrade plant biomass. The concept on rRNA gene pyrosequencing, which demonstrated effects of utilization of simplified multispecies consortia, ranging from the enrichment process on shaping of the community struc- dual-species systems to complex microcosms on plant bio- tures. Lignocellulolytic enzyme systems of the consortium mass decomposition, is a promising strategy to achieve high were further explored using shotgun metagenome pyrose- degradation efficiency compared to using a single cellulolytic quencing, which showed the highly enriched nature of gene isolate (Lynd et al. 2002). A number of structurally stable sets encoding a distinct enzyme profile compared with pre- multispecies consortia with high cellulose-degrading activity viously reported plant biomass degrading microbiomes. The have been obtained by successive culture enrichments using work provides the first in-depth analysis of lignocellulolytic agricultural biomass as the sole carbon source under meso- enzyme systems and metabolic potential of a symbiotic and thermophilic conditions. These symbiotic consortia can biomass degrading consortium by metagenomic approaches. efficiently degrade various cellulosic materials, such as agro- This study gives valuable insights into the mechanism in industrial residues, filter paper, cotton, and pulp wastes consolidated lignocellulose degradation and conversion by a (Haruta et al. 2002; Guo et al. 2010; Wongwilaiwalin et al. symbiotic consortium and provides a promising gene min- 2010;Fengetal.2011;Wangetal.2011). A defined cellulose- ing platform for discovery of efficient synergistic enzyme degrading consortium was assembled from microbial isolates systems for biotechnological application. in a mixed culture microcosm (Kato et al. 2004), and study of this simplified microcosm revealed complex metabolic interdependences among the community members, which contrib- Materials and methods ute to the consortium stability (Kato et al. 2005, 2008). These symbiotic consortia typically comprise mainly anaerobic cel- Materials lulolytic bacterial taxa in stable coexistence with various noncellulolytic members, which play important roles in Sources of microflora for preparation of microbial consortia metabolite dependence and environment modification, lead- in this study were collected from (1) sugarcane bagasse ing to enhancement in biomass hydrolysis efficiency of the decomposed soil from an industrial bagasse collection site multispecies systems. Modification of lignocellulose degrad- at the Phu Kheio Bio-Energy factory (Chaiyapoom, ing consortia was first explored by introducing and acclima- Thailand), (2) cow rumen fluid from Holstein cows housed tion of a cellulolytic clostridial strain into the system, which at the Khon Kaen Agriculture and Training Center (Roi Et, showed a way for alteration of the consortium structures and Thailand), and (3) activated sludge from a local soda pulp functions (Narisawa et al. 2007). The conceptual design on mill (Ratchaburi, Thailand). Rice straw and other agricul- development of symbiotic consortia comprising either natural tural by-products used as substrates in enrichment and bio- or engineered members has been recently reported as a prom- mass degradation efficiency analysis were obtained locally. ising approach for the establishment of microbial systems for The agricultural by-products were pretreated with alkali consolidated bioprocessing (Zuroff and Curtis 2012). Thus, (10 % w/v NaOH with 3:1 liquid/solid ratio at 90 °C for these studies indicate potential of the symbiotic microbial 90 min) for partial delignification. The native rice straw consortia, not only as the highly efficient biomass degrading contained 41.90 % cellulose, 34.24 % hemicelluloses, and systems but also for further modification of interspecies met- 22.44 % lignin, while the alkali-treated rice straw comprised abolic flux in designated consortia aiming for direct conver- of 68.69 % cellulose, 19.41 % hemicellulose, and 4.45 % sion of biomass to fuels and chemicals of industrial lignin according to analysis by Technical Association of importance. Pulp and Paper Industry (TAPPI) methods (T203om-83,

164 T222om-83, and T223hm-84; TAPPI 1992). Industrial eu- Denaturing gradient gel electrophoresis of 16S rRNA gene calyptus pulp waste was obtained from a local soda pulp fragments mill and used without pretreatment. The filter paper used in this study was Whatman number 1 (Whatman, Maidstone, Total genomic DNA was extracted from microbial consortia Kent, UK). The cellulosic substrates were autoclaved twice using the benzyl chloride method (Zhu et al. 1993). The at 121 °C for 15 min before use. purified DNA was used as a template for amplification of the partial 16S rRNA gene fragment. The reactions contained Construction of lignocellulolytic microbial consortia 30 ng of template DNA in 1× PCR buffer with 10 μmol of 338GC-F forward primer, which was attached to a GC The lignocellulose degrading microbial consortia were clamp at the 5′-terminus (5′-cgcccgccgcgcgcggcgggcggg obtained from three different environmental microbial sour- gcgggggcacggggggactcctacgggaggca-3′,GCclampse- ces: (1) bagasse (BG-Env), (2) cow rumen fluid (CR-Env), quence is underlined) and 518R reverse primer (5′- and (3) activated sludge (AS-Env) and prepared as described attaccgcggctgctgg-3′) encompassing the V3 hypervariable by Haruta et al. (2002). One gram (or 1 ml) of sample was region using Taq DNA polymerase (Fermentas, Vilnius, inoculated into a 50-ml screw-cap disposable tube containing Lithunia) according to the manufacturer’sprotocol 30 ml of PCS medium (0.1 % yeast extract, 0.5 % peptone, (Wongwilaiwalin et al. 2010). Reactions were performed

0.5 % CaCO3, 0.5 % NaCl, and 1 % alkali-pretreated rice on a MyCycler thermal cycler (Bio-Rad Laboratories, straw, pH8.0) with a filter paper strip (0.3 g) as an indicator for Hercules, CA, USA). The temperature profile consisted of cellulase activity. The mixture was incubated at 50 °C with 94 °C for 3 min, followed by 30 cycles of denaturation at continuous shaking at 200 rpm until the filter paper had 94 °C for 1 min, annealing at 55 °C for 1 min, and extension disintegrated (approximately 3–4 days), after which 1 ml of at 72 °C for 2 min, followed by a final extension step at 72 °C the culture was then transferred into fresh medium. The pro- for 3 min. The profiles of amplified 16S rRNA gene fragments cedure was then repeated until a structurally stable cellulolytic were analyzed by denaturing gradient gel electrophoresis microbial community was obtained as indicated by denaturing (DGGE) on the DCode™ system (Bio-Rad Laboratories, gradient gel electrophoresis. The structurally stable microbial Hercules, CA, USA) according to Wongwilaiwalin et al. consortia originating from bagasse, cow rumen, and activated (2010). sludge were designated BGC-1, CRC-1, and ASC-1, respectively. The consortia were maintained in PCS medium con- Tagged 16S rRNA gene pyrosequencing taining 20 % glycerol with no cellulosic substrate at −80 °C for long-term storage. The structurally stable consortia were The purified metagenomic DNA was used as the template used for all subsequent experimental studies. for amplification of the partial 16S rRNA gene using universal bacterial primers 338F (5′-ACTCCTACGGGAG Enzyme activity analysis GCAGCAG-3′ ) and 786R (5′ -CTACCAGGGT ATCTAATC-3′) encompassing the three and four hypervari- The crude enzyme (culture supernatant) of BGC-1 was clari- able regions in prokaryotic 16S rRNA gene attached with fied by centrifugation at 7,500×g for 15 min and used for tagged barcode sequences (Meyer et al. 2008). Polymerase enzyme profile analysis. Lignocelllulose degrading activities chain reactions were performed using DyNAzyme EXT in the supernatant were analyzed based on the amount of DNA polymerase (Finnzyme, Espoo, Finland) on a liberated reducing sugars using the 3,5-dinitrosalisylic acid MyCycler thermocycler for 30 cycles of denaturation at method according to the method described in Rattanachomsri 94 °C for 1 min, annealing at 55 °C for 1 min, and extension et al. (2009). One-hundred-microliter reactions contained the at 72 °C for 1 min 45 s. The amplicons were purified using a appropriate dilution of enzyme in 50 mM sodium acetate QIAGEN PCR purification kit (QIAGEN, Hilden, buffer, pH5.0, and 1 % (w/v) of the corresponding substrate: Germany) and were quantified using a NanoDrop ND- carboxymethyl cellulose for CMCase activity, birchwood xy- 1000 spectrophotometer (ThermoScientific, Wilminton, lan for xylanase activity, and β-glucan for β-glucanase activ- DE, USA). The sequences were determined using pyrose- ity. The reaction was incubated at 50 °C for 60 min, and the quencing on a 454-Life Sciences GS-FLX Genome amount of reducing sugars was determined from the absor- Sequencer System (Roche, Branford, CT, USA) following bance measurements at 540 nm and interpolated from a stan- the manufacturer’s recommended protocols. The 16S rRNA dard curve of the corresponding sugars. Assay reactions were sequences were deposited in the NCBI Sequence Read performed in triplicates, and the mean was reported with Archive (SRA) with the accession number SRX122677. standard deviation ≤5 %. One international unit (IU) was The pyrosequencing dataset was classified into groups defined as the amount of enzyme which produced 1 μmol of based on the tagged sequence using a customized Python reducing sugar in 1 min. script (www.python.org). The read sequences were trimmed

165 by removing the tag and primer sequences using GS De Novo straw under facultative anoxic conditions. After several Assembler version 2.6 (Roche, Branford, CT, USA). Chimeric rounds of subcultivations, the most active consortium was sequences were identified and removed using UCHIME selected as a representative from each line based on their (Edgar et al. 2011). Sequences that were 200 bps or longer cellulose degrading capability. According to the DGGE in length were selected for further analysis. Taxonomic clas- patterns, the community structures of the isolated symbiotic sification was assigned using the RDP classifier tool (Wang et consortia tended to stabilize after 10–15 successive subcul- al. 2007) to determine approximate phylogeny with 95 % tivations with a typical incubation time of 3–4 days for each confidence threshold and further classified using BLASTN batch cycle (Fig. 1a). Overall, similar DGGE patterns in six against the NCBI 16S rRNA database for phylogenetic as- generations were observed with no significant different (p= signment at the genus level using an expected E value cutoff 0.736, 0.976, and 0.909 for BGC-1, CRC-1, and ASC-1, of 1e−06. Operational taxonomic units (OTUs) were deter- respectively), indicating stable coexistence of the composite mined at sequence similarity levels of 0.03, 0.05, and 0.15 by microbes in the consortia and hence stability of the overall the furthest-neighbor method of MOTHUR (Schloss et al. community structures. The DGGE profiles of consortia iso- 2009). The diversity indexes and nonparametric diversity lated from different environmental sources were markedly were calculated to measure and compare diversity among different, suggesting variation in composite microbes the dataset. Good’scoverage(G), an estimator for sampling among the consortia. The consortia were functionally stable completeness, was calculated as G=1−(n/N), where n is the according to their degradation efficiencies on the pretreated number of singleton phylotypes and N is the total number of substrates in each batch cycle. The microcosms were aero- sequences in the sample. tolerant and capable of growth under nonreduced conditions. They were tolerant to freezing at −80 °C in PCS Shotgun pyrosequencing medium containing 20 % glycerol with no change in microbial composition as shown by DGGE patterns of the revived The purified metagenomic DNA was processed for pyrose- communities (data not shown). This illustrated the high quencing on a 454-Life Sciences GS-FLX Genome Sequencer stability of the microbial communities and allowed long- System using the titanium platform following the manufac- term storage of the seed culture for further experimental turer’s protocols. The sequences obtained from pyrosequenc- studies and application. ing were assembled using Newbler version 2.6 (Roche, Branford, CA, USA). The metagenome sequence reads Comparative microbial community structures obtained are available in the NCBI Sequence Read Archive (www.ncbi.nlm.nih.gov/Trace/sra) with the accession number In order to investigate community structures of the micro- SRX122677. The assembled sequences were predicted for cosms, the purified metagenomic DNAs of the environmen- open reading frames using GeneMark (Lukashin and tal microflora and their corresponding consortia cultured for Borodovsky 1998). Translated ORFs≥30 amino acid residues 4 days were used as the templates for direct amplification of in length were searched for annotated proteins using BLASTP the partial 16S rRNA genes attached with specific tag against the nonredundant protein sequences (nr) using an sequences. The amplicon sequences of 440 bps were ana- expected E value cutoff at 1e−06. Functional gene sequences lyzed by pyrosequencing. The reads were filtered to remove were analyzed and mapped to Kyoto Encyclopedia of Genes short sequences (<200 bases) and grouped according to their and Genomes terminology using BLAST2GO (Conesa et al. attached tag. In total, filtered reads were obtained with a 2005). Glycosyl hydrolase putative sequences were identified range of average read length from 306.28 to 382.38 bases by searching the pyrosequencing dataset against a library of (Table 1). The number of reads obtained from different modules derived from all entries in the carbohydrate-active samples was in the range of 3,597 to 63,872 reads/sample. enzyme database (CAZy) (www.cazy.org) (Cantarel et al. The average %GC content was 53.85 % with little variation 2009). in %GC among different samples (SD=0.91). Identification of unique phylotypes and estimation of bacterial richness using MOTHUR revealed that the dataset Results represented 1,852–21,295 OTUs for microflora in environments and 798–2,518 OTUs for the stable consortia based Construction of stable lignocellulolytic microbial consortia on a >97 % identity cutoff. The higher Shannon diversity index was found for the environmental samples compared Structurally and functionally stable lignocellulose- with the stable consortia, indicating the higher diversity of degrading microcosms originating from different environ- microbes in the starting communities compared with their mental biomass-degrading microbiomes were obtained by corresponding symbiotic microcosms (Table 1)(p<0.0001 successive subcultivation of microflora on delignified rice at 95 % confidence). Good’s coverage showed that the

166 Fig. 1 Characterization of symbiotic microbial consortia. a Structural stability of composite microbes in microbial consortia. Total genomic DNA was extracted from the corresponding consortium incubated for 4 days and used as the template for 16S rRNA gene DGGE analysis. Generation numbers of the consortium samples were indicated. b Bacterial phyla profiles of original microflora and stable lignocellulolytic microbial communities. c Classification of composite bacterial genera in stable lignocellulolytic consortia

dataset covered >95 % of the underlying bacterial diversity This was different from the microflora in cow rumen dom- at the genus level for most samples, except for the highly inated by anaerobic Bacteroidetes and Firmicutes, particu- diversified BG-Env. A higher coverage was observed for the larlyinordersBacteroidales and Clostridiales and the stable consortia, i.e., 85.34–94.96 % at the species level microflora in activated sludge, which predominated by aer- compared with the environmental microflora, reflecting the obic bacterial taxa in phylum Proteobacteria, specifically in selective enrichment of specific bacterial taxa under the orders Rhodospirillales and Sphingobacteriales. Successive subcultivation conditions. subcultivation led to enrichment of specific bacterial groups, The taxonomic classification results using RDP classifier particularly in phyla Firmicutes and Bacteroidetes resulting revealed substantial differences in overall microbial com- in relatively similar overall bacterial profiles of all stable munity profiles between the environmental microflora and consortia although they originated from highly diversified symbiotic consortia (Fig. 1b). The community profiles at microflora from different sources. phylum levels were highly different among the three envi- Composite microbe analysis at deeper taxonomic levels by ronmental microbiomes (BG-Env, CR-Env, and AS-Env), BLASTN showed the coexistence and possible interdepen- while they were relatively similar in the structurally stable dency of specific groups of microbes in the corsortia when consortia after subcultivation (BGC-1, CRC-1, and ASC-1). compared with the highly diverse original microcosms The major taxa found in BG-Env contained diverse bacterial (Fig. 1c). Successive subcultivation led to enrichment of phyla with different oxygen requirements, predominated by orders Clostridiales, Bacillales, and Bacteroidiales in all Proteobacteria, Bacteroidetes, and Actinobacteria, compris- structurally stable consortia. Focusing at a more refined taxo- ing mainly orders Rhodospirillales and Pseudomonadales. nomic level, genera Clostridium, Acetanaerobacterium,

167 Table 1 Statistic analysis of biodiversity indexes of microbial BG-Env BGC-1 communities in environments Number of reads 63,862 19,705 and structurally stable consortia Statistical index 0.03 0.05 0.15 0.03 0.05 0.15 Richness 21,295 16,861 9,919 2,518 1,116 184 Chao 77,097 55,337 27,481 4,343 1,640 241 ACE 153,606 106,792 48,438 5,899 2,000 270 Shannon 8.4211 7.9085 6.6292 6.0381 5.1706 3.7437 Simpson 0.0035 0.0043 0.0117 0.0109 0.0162 0.038 Coverage 75.28 81.12 89.73 93.66 97.8 99.76 CR-Env CRC-1 Number of reads 45,202 3,597 Statistical index 0.03 0.05 0.15 0.03 0.05 0.15 Richness 8,585 5,389 1,368 798 477 107 Chao 17,957 9,294 1,743 1,863 960 170 ACE 27,288 12,479 1,763 2,629 1,276 194 Shannon 8.0971 7.387 5.8444 5.2041 4.5072 3.0049 Simpson 0.0009 0.0018 0.0061 0.0253 0.0382 0.1173 Coverage 83 91.04 98.72 85.34 92.42 98.92 AS-Env ASC-1 Number of reads 7,998 13,257 The needleman alignment meth- Statistical index 0.03 0.05 0.15 0.03 0.05 0.15 od was used to generate an Richness 1,852 1,180 440 1,411 742 160 alignment with a gap-opening Chao 4,030 2,038 543 2,528 1,195 203 − penalty of 1. Distance matrix ACE 6,143 2,816 565 3,458 1,578 267 was generated using the furthest method with a cutoff of 0.15. All Shannon 6.1448 5.5013 4.4867 5.348 4.5273 3.2397 statistical diversity indices and Simpson 0.0113 0.0163 0.0312 0.0163 0.0291 0.0632 estimators were performed under Coverage 86.1 92.62 98.42 94.6 97.55 99.64 10,000 iterations

Ruminococcus, Caloramator,andUreibacillus constituted the including alkali-pretreated rice straw, sugarcane bagasse, majority of the composite microbial population in BGC-1. and corn stover, and nonpretreated eucalyptus pulp waste, Substantial similarities were shared by CRC-1, but with a with 72.4±2.8 %, 69.7±2.2 %, 72.7±3.2 %, and 48.2± higher abundance of Ruminococcus and the presence of 3.2 % substrate decomposition, respectively, after incuba- Tepidimicrobium and Lutispora. ASC-1 contained the highest tion for 7 days under anoxic conditions. Native lignocel- abundance of Clostridium compared to the other consortia in lulosic plant biomass was also degraded, but with less coexistence with various bacterial genera with different oxy- efficiency than the pretreated substrates. A substantial gen requirements, e.g., Proteiniphilum and Azotobacter.The decrease in pH from 8 to 6.5 was found in the first overall community profiles thus suggest enrichment of spe- 2 days of filter paper degradation, and the pH then cific groups of cellulolytic bacteria together with some facul- slightly increased to 6.9 during the later phase after the tative non-cellulolytic anaerobes or microaerobes during the 7-day incubation period. A similar trend of pH profile subcultivation process, indicating the effects of selective sub- was also observed for degradation of agro-industrial substrate and environmental conditions on shaping the composite strates (data not shown). bacterial profiles of the stable consortia. Lignocellulolytic enzyme activity profiles of the most efficient consortium, BGC-1, were analyzed in the crude Biomass degrading efficiency of symbiosis microcosms extracellular fraction. Continual increases in plant biomass degrading enzyme activities were observed during the incu- The lignocellulolse-degrading capability of the stable con- bation time. The endo-glucanase (CMCase), xylanase, and sortia was assessed by determining the extent of biomass β-glucanase activities reached maximum levels of 1.31, degradation during a 1-week incubation period (Fig 2a). All 1.02, and 4.80 IU/mg, respectively, at the end of 10-day consortia were capable of decomposition of all cellulosic incubation period (Fig. 2b). This indicated the capability of substrates, but to different extents. BGC-1 showed the high- the consortia for production of key cellulolytic and hemi- est degradation efficiency on all cellulosic materials cellulolytic enzymes and their potential as a platform for

168 excluded from the analysis. The filtered dataset after removing repeats and low quality sequences contained 66,517 sequences with an average read length of 1,163.4 and 467.7 bases for the contigs and singletons, respectively, with the total reads covering 34.9 Mb. In total, 78,514 ORFs were identified from the filtered dataset. Annotation of functional genes in the metagenome was initially determined by BLASTP. From the filtered dataset, 73.70 % of the translated query sequences could be mapped against protein sequences in the nonredundant (nr) protein database. According to classification of the sequence origins by mapping to NCBI taxonomy, the majority of the organisms belonged to bacteria (98.55 %) while very few archaea (0.92 %) and eukaryote (0.31 %) related sequences were identified (Fig. 3). Firmicutes (50.1 %), Bacteroidetes (23.8 %), Proteobacteria (14.1 %), and Spirochaetes (5.1 %) constituted the most abundant phyla, in addition to other minor bacterial phyla including Actinobacteria, Synergistetes, Planctomycetes,andThermotogae. Focusing on the metabolic capability of the consortium on lignocellulose degradation, the pyrosequencing dataset was searched for homology to glycosyl hydrolase (GH) families targeting on lignocellulosic polysaccharides. Putative genes encoding plant polysaccharide-degrading enzymes were initially annotated based on retrieved sequences from CAZy with an E value cut-off<1e−40 (Table 2). Using these criteria, a total of 957 GH-related sequences in 69 families were identified in the metagenome accounting for 1.65 % of the total hit sequences. Of these sequences, 485 reads were closely related to lignocellulose degrading enzymes in 26 GH families, in addition to the other GHs Fig. 2 Biomass degradation efficiency and enzyme activity profiles of lignocellulolytic microbial consortia. a Degradation of lignocellulosic targeting noncellulosic substrates. The genes encoding for substrates by lignocellulolytic consortia constructed from different oligosaccharide-degrading enzymes represented the majori- environmental microflora. The consortia were incubated in PCS medi- ty (~62 %) of the annotated genes for lignocellulolytic um with different substrates at 50 °C, 200 rpm for 7 days with enzymes. Most were related to GH2 and GH3 families substrates in uninoculated medium as a control. b Enzyme activity β profiles of BGC-1. The reactions were performed in triplicate, and the comprising mostly -glucosidases and other enzymes hy- averages of the results were reported drolyzing β-linked dimers from hydrolysis of hemicelluloses. Endocellullases were dominated by GH5 and GH9, which are major classes of endo-glucanases attacking plant discovery of various lignocellulose-degrading enzymes for biomass. Hemicellulases of types GH10, 26, and 28 were biotechnological application. found, which included a diverse range of endo-acting enzymes attacking hemicelluloses and pectin, e.g., endo-β- Direct shotgun pyrosequencing analysis of BGC-1 1,4-xylanases, β-mannanases, and polygalacturonases, in metagenome addition to various debranching enzymes working cooperatively on hydrolysis of hemicelluloses. The consortium BGC-1 was further investigated for their Classification of the overall plant biomass degrading metabolic capability using direct metagenome shotgun enzymes based on their origins is shown in Fig. 4. pyrosequencing. Sequencing of isolated total community Annotation of the GH sequence taxonomic origins showed DNA resulted in 144,253 raw reads with an average read that the majority of endocellulases were closely related to length of 479.0 bases totaling 69.1 Mb of sequence infor- those from phyla Firmicutes and Bacteroidetes, particularly mation. A substantial fraction (54.9 %) of all reads could be from the genera Clostridium, Niastella,andTannerella. assembled to 5,399 contigs of varying lengths, while the rest Hemicellulases, including various endo-acting enzymes were singletons (42.34 %) and outliers (3.44 %), which were attacking the heterogenous hemicelluloses, e.g., xylan and

169 Fig. 3 Taxonomic classification of microbial population in BGC-1 based on origins of putative functional genes

mannan, were related to those from genera Clostridium, vary greatly among these environments, as Proteobacteria, Bacteroides,andPedobacter. Diverse downstream exo- Bacteroidetes,andActinobacteria predominated in the mixed acting oligosaccharide-degrading enzymes and debranching aerobic/anoxic microenvironment of bagasse decomposed soil, accessory enzymes are closely related to their homologues while anaerobic Bacteroidetes/Firmicutes predominated the from Bacteroidetes origin, particularly genera Bacteroides rumen microbiome, and aerobic Proteobacteria predominated and Tannerella and also from Firmicutes in genera the microcosm in activated sludge from a eucalyptus pulp mill. Clostridium, Paenibacillus, and Thermobacillus. This indi- The community profiles for each of the three consortia cates that bacterial genera belonging to phyla Firmicutes obtained from these diverse microcosms in our study share and Bacteroidetes were the major plant biomass-degrading similarity at the phylum level, comprising Clostridiales, enzyme producing microbes in the consortium, with a minor Bacteroidales, and Bacillales as the major taxa. The predom- contribution from Proteobacteria and Spirocheates.The inance of these three phyla suggests that convergent adaptation result thus indicated cooperation of various enzymes from is driven by the selective pressure applied during the isolation diverse microbial sources, particularly from the anaerobic procedure. This phenomenon is indicated further by the re- cellulolytic bacterial taxa on degradation of lignocellulosic spective decreases in biodiversity indexes of the consortia substrates. compared with the original microbiomes. Structures of the lignocellulolytic consortia grown under initial aerobic to later facultative anoxic conditions in this Discussion study are broadly similar to microbial consortia originating from bagasse compost (Wongwilaiwalin et al. 2010)and Lignocellulose degradation in nature is a complex microbial rice straw compost (Haruta et al. 2002) prepared under static process comprising of diverse eukaryotic and prokaryotic aerobic conditions, suggesting that the interplay of compos- microbes playing cooperative actions on hydrolysis of the ite microbes on lignocellulose degradation mechanisms are recalcitrant lignocellulose structures under aerobic and anoxic similar among lignocellulolytic consortia obtained from dif- conditions. Understanding of the interplay of microbes on ferent substrates and conditions. Anaerobic cellulolytic clos- lignocellulose degradation in nature is thus challenging. tridia and microaerophilic or facultative anaerobic bacteria Among various biomass degrading microbial systems, struc- are often simultaneously detected in different environments turally stable symbiotic lignocellulolytic consortia represent a where degradation of plant biomass occurs under anoxic promising model for study of microbe interactions and enzyme conditions (Mitchell 1998) and also in the previously synergism on plant biomass degradation by microbial commu- reported consortia. The balance of various types of symbi- nities. In this study, symbiotic consortia have been isolated otic relationships, e.g., on metabolite interdependences and from diversified plant biomass-degrading microflora inhabit- environmental controls among these bacteria is considered ing various environments with high potentials for lignocellu- essential for their stable coexistence in the community, losic decomposition. The natural microbiomes were found to which results in efficient biomass degradation (Kato et al.

170 Table 2 Comparative relative abundance of genes encoding for lignocellulose degrading enzyme homologs in different environmental metagenomes with high lignocellulose degrading potential

CAZy family Annotated enzyme homologs BGC-1 Termite Bovine Wallaby in BGC-1 Hindgut (Warnecke Rumen Foregut et al. 2007) (Brulc et al. 2009) (Pope et al. 2010)

Cellulases GH5 Endoglucanase (3.2.1.4) 22 56 7 10 Glucan β-1,3-glucosidase (3.2.1.58) Mannan endo-β-1,4-mannosidase (3.2.1.78) GH9 Endoglucanase (3.2.1.4) 25 9 6 0 β-Glucosidase (3.2.1.21) GH8 Endoglucanase (3.2.1.4) 4 5 4 1 GH44 Endoglucanase (3.2.1.4) 3 6 0 0 GH45 Endoglucanase (3.2.1.4) 0 4 0 0 Total 54 (5.6 %) 80 (11.4 %) 17 (1.8 %) 11 (2.0 %) Endohemicellulases GH10 Endo-1,4-β-xylanase (3.2.1.8) 22 46 7 11 GH11 Xylanase (3.2.1.8) 4 14 1 0 GH26 β-Mannanase (3.2.1.78) 7 15 5 5 GH28 Polygalacturonase (3.2.1.15) 22 6 5 2 GH53 Endo-β-1,4-galactanase (3.2.1.89) 3 12 17 9 Total 58 (6.1 %) 93 (13.2 %) 35 (3.7 %) 27 (4.8 %) Cell wall elongation GH16 Endo-1,3(4)-β-glucanase (3.2.1.6) 9 1 0 0 Endo-β-1,3-galactanase (3.2.1.–) GH17 Glucan 1,3-β-glucosidase (3.2.1.58) 3 0 0 1 GH74 Endoglucanase (3.2.1.4) 2 7 0 0 Total 14 (1.5 %) 8 (1.1 %) 0 (0.0 %) 1 (0.2 %) Debranching enzymes GH51 α-L-Arabinofuranosidase (3.2.1.55); 21 18 64 12 endoglucanase (3.2.1.4) GH54 α-L-Arabinofuranosidase (3.2.1.55) 0 0 1 0 β-Xylosidase (3.2.1.37) GH62 α-L-Arabinofuranosidase (3.2.1.55) 1 0 0 0 GH67 α-Glucuronidase (3.2.1.139) 8 10 0 5 GH78 α-L-Rhamnosidase (3.2.1.40) 29 0 34 25 Total 59 (6.2 %) 28 (4.0 %) 99 (10.3 %) 42 (7.5 %) Oligosaccharide-degrading enzymes GH1 β-Glucosidase (3.2.1.21); 7 22 10 61 β-Galactosidase (3.2.1.23) GH2 β-Galactosidase (3.2.1.23) 92 23 186 24 GH3 β-Glucosidase (3.2.1.21) 71 69 176 72 Xylan 1,4-β-xylosidase (EC 3.2.1.37) β-N-acetylhexosaminidase (3.2.1.52) Glucan 1,3-β-glucosidase (3.2.1.58) α-L-Arabinofuranosidase (3.2.1.55) GH29 α-L-Fucosidase (3.2.1.51) 16 0 28 2 GH35 β-Galactosidase (3.2.1.23) 3 3 12 3 GH38 α-Mannosidase (3.2.1.24) 7 11 17 3 GH39 β-Xylosidase (3.2.1.37) 11 3 2 1

171 Table 2 (continued)

CAZy family Annotated enzyme homologs BGC-1 Termite Bovine Wallaby in BGC-1 Hindgut (Warnecke Rumen Foregut et al. 2007) (Brulc et al. 2009) (Pope et al. 2010)

GH42 β-Galactosidase (3.2.1.23) 4 24 11 8 GH43 β-Xylosidase (3.2.1.37) 87 16 61 10 α-L-Arabinofuranosidase (3.2.1.55) GH52 β-Xylosidase (3.2.1.37) 2 3 0 0 Total sequences 300 (31.3 %) 174 (24.7 %) 503 (52.6 %) 184 (33.0 %) (% in total GHs)

2008). In the previously reported consortium originating and the Clostridium/Proteiniphilum-enriched ASC-1 con- from rice straw compost, noncellulolytic facultative anaero- sortia. We infer that specific microbial profile of BGC-1 bic bacteria have been proposed to play a major role in provides the basis for its high lignocellulose degrading oxygen consumption in the early stage, triggering oxygen capability compared with other consortia. Clostridium are depletion in the system and allowing the growth of cellulo- important plant biomass degraders in anoxic conditions, as lytic clostridia in the later anoxic condition (Kato et al. has been reported in several symbiotic microcosms prepared 2004, 2005). In addition, the noncellulolytic members also under meso- and thermophilic conditions (Haruta et al. contribute to cellulose degradation by controlling pH and 2002;Fengetal.2011; Wongwilaiwalin et al. 2010). consuming metabolites, which would otherwise deteriorate Acetanaerobacterium are obligate anaerobic bacteria closely the cellulolytic activity, e.g., by enzyme inhibition and me- related to Clostridium, which was previously isolated from a tabolite repression. pulp mill (Chen and Dong 2004), and in hydrogen fermen- Although the overall scheme of interaction among micro- tation reactors (Ren et al. 2007). Ureibacillus was previous- organisms may be similar, microbial profiles among the ly reported as a ligninolytic bacterium, which plays a key consortia in this study are substantially different from those role in detoxification of biomass hydrolysate during fermen- previously reported. The BGC-1 consortium, which showed tation (Okura et al. 2008). Caloramator boliviansis is a the highest biomass-degrading efficiency, comprised a thermoanaerobic in Clostridiaceae family, which could fer- unique combination of cellulolytic Clostridium in coexis- ment biomass composite sugars, e.g., cellobiose, xylose, and tence with Acetanaerobacterium, Ruminococcus, sugarcane bagasse hydrolysate to ethanol (Crespo et al. Caloramator,andUreibacillus. The profile of BGC-1 is 2012). The differences in microbe composition among con- markedly different from the Ruminococcus-lead CRC-1 sortia also imply differences in their lignocellulolytic

Fig. 4 Classification of putative genes encoding for lignocellulose degrading enzymes in BGC-1 metagenome based on closest taxonomic origins

172 enzyme systems and specificities as well as their biomass of genes encoding for plant biomass-degrading GHs in the degradation efficiency and fermentation capability. metagenome as a result of the prolonged subcultivation pro- Exploration on metabolic-interdependent interactions of cess under selective conditions. the composite microbes in the system is of interest for According to annotation of functional GHs based on the providing the basis for further control and modification of enzyme classification (Allgaier et al. 2010), the BGC-1 the consortium for biotechnological application. metagenome contains diverse genes encoding for enzymes The highly enriched natures of lignocellulolytic microbes targeting plant cell wall degradation. Of these genes, 5.6 % and enzymes together with their synergistic interactions are are putative cellulases, mostly belonging to the major endo- plausible mechanisms to explain BGC-1’s high lignocellu- glucanase families GH5 and GH9, which are >2.5 times lose degradation efficiency. The enriched consortium dis- higher than that in rumen and wallaby gut microbiomes, plays a very high relative abundance of putative GHs genes but only half of that in termite gut. Diverse putative hemi- (~1.65 % of the total gene), which is significantly higher cellulase genes in the BGC-1 consortium metagenome are than that of microcosms in other plant biomass degrading related to genes encoding endo- and exo-acting enzymes environments, e.g., termite gut symbionts (Li et al. 2009) targeting degradation of the heterogeneous hemicellulose and the switchgrass-adapted compost community (Allgaier structures, e.g., xylan, mannan, and galactan dominated by et al. 2010), and much higher than the average value for all GH10, GH26, and GH28, which are major hemicellulases in examined environmental metagenomes (0.64 %) at the same E plant biomass degradation (Brulc et al. 2009). Putative value cut-off at 1e−40 (Li et al. 2009)(Fig.5). This is in line genes encoding oligosaccharide-degrading enzymes have with the increased abundance of GHs in environments with been annotated including those in the GH3 and GH43 high turnover of complex plant biomass. Comparative analy- families, which comprise exo-acting cellulolytic and hemi- sis of pyrosequencing dataset (Table 2)alsoshowsahigh cellulolytic enzymes, including mainly β-glucosidases and relative abundance of genes encoding for lignocellulose- β-xylosidases. A high proportion of enzymes involved in degrading enzymes in the BGC-1 metagenome, which is side chain processing (debranching and cell wall elongation comparable to those from other enriched lignocellulose enzymes) have also been annotated consistent with those degrading niches with a high turnover of complex plant bio- found in the rumen microbiome. The major debranching mass, i.e., termite gut (Warnecke et al. 2007), cow rumen enzymes are α-arabinofuranosidases of the GH45 and (Brulc et al. 2009), and wallaby gut (Pope et al. 2010). GH51 families that cleave off arabinose side chains of Altogether, the results thus suggested that the increased met- glucoronoarabinoxylan, which is a major hemicellulose in abolic capability of BGC-1 is due to the increased abundance grass and other monocot plants. These enzymes have been

Fig. 5 Relative abundance of genes encoding for glycosyl hydrolase homologs in different microbiomes. The GH abundance was compared to datasets reported in Li et al. 2009 and Allgaier et al. 2010 at E value cut-off at 1e−40

173 reported to work cooperatively on degradation of lignocel- secreted forms and the highly active cellulosomal systems luloses (Alvira et al. 2011). The relative abundance of (Doi 2008). The presence of cellulosomes in the enzyme putative genes encoding major main-chain attacking GHs system of BGC-1 is suggested by the presence of dockerin and accessory lignocellulolytic GHs of BGC-1 is interme- encoding sequences in the metagenome (data not shown). diate between the termite hindgut microbiome, which shows Bacteroides has also been reported to produce a range of high endo-activity required for decomposition of hardwood hydrolytic enzymes, which are involved in processing of core fiber (Warnecke et al. 2007), and the rumen micro- complex plant polysaccharides such as cellulose, xylan, biome, which shows high metabolic potential for accessory arabinogalactan, and pectin in various environments, e.g., enzymes needed for degradation of the easily available side soil, animal gut, and anaerobic digesters (Weber et al. 2001). chains in forages (Brulc et al. 2009). This unique profile Synergistic enzyme action has been reported for many cel- might reflect the more exposed nature of plant polysacchar- lulolytic and hemicellulolytic enzymes, both for free ides in the delignified rice straw used as the cellulosic enzymes and cellulosomes from various microbial origins material during consortium adaptation, which thus led to (Doi 2008). Cellulosomes are complex enzyme systems the enrichment of endo-acting GHs with the intermediate highly efficient in lignocellulose degradation due to the abundance of side-chain processing enzymes characteristic synergistic action of catalytic cellulosomal components. for nonwood substrates. Most lignocellulose-degrading This is due to the proximity effect of various cellulases enzymes belong to Firmicutes and Bacteroidetes in several and hemicellulases arranged on the core scaffoldin, which genera. According to gene mapping (Fig. 6), Clostridium leads to their enhanced cooperative action on plant biomass and Bacteroides were the most dominant lignocellulolytic decomposition (Doi 2008). Synergism between metabolical- enzyme producers in the consortia. Clostridia are capable of ly active cellulolytic microbes and enzymes has also been producing various plant biomass-degrading enzymes in both proposed based on removal of hydrolysis inhibitory

Fig. 6 Gene mapping on lignocellulose degradation and sugar fermentation pathways. Taxonomic sources of the gene sets involved in the synthesis pathways of indicated fermentation products are shown in pile charts

174 products and surface phenomena involving adherent cellu- Alvira P, Negro MJ, Ballesteros M (2011) Effect of endoxylanase and lolytic microorganisms (Lu et al. 2006). Further character- α-L-arabinofuranosidase supplementation on the enzymatic hydrolysis of steam exploded wheat straw. Bioresour Technol ization of novel cellulolytic microbes and enzyme systems 102:4552–4558 from the established consortia with different reaction spe- Brulc JM, Antonopoulos DA, Miller MEB, Wilson MK, Yannarell AC, cificities and their synergistic mechanism are thus of great Dinsdale EA, Edwards RE, Frankh ED, Emersoni JB, Wacklini P, interest for biotechnological research. Coutinhoj PM, Henrissatj B, Nelsoni KE, White BA (2009) Gene- centric metagenomics of the fiber-adherent bovine rumen micro- Analysis of metabolic potential of BGC-1 using gene biome reveals forage specific glycoside hydrolases. Proc Nat mapping shows the capability of the consortium on process- Acad Sci USA 106:1946–1953 ing of the complex lignocellulosic biomass to composite Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard V, sugars, i.e., glucose and pentoses, and subsequently to fer- Henrissat B (2009) The Carbohydrate-Active EnZymes database (CAZy): an expert resource for glycogenomics. Nucleic Acids mentation products of industrial importance, e.g., ethanol, Res 37:D233–238 butanal, lactic acid, and hydrogen (Fig. 6). This can be Chen S, Dong X (2004) Acetanaerobacterium elongatum gen. nov., sp. achieved through the sequential sugar fermentation by an- nov., from paper mill waste water. Inter J Sys Evol Microbiol aerobic metabolic pathways downstream of the central gly- 54:2257–2262 Conesa A, Gotz S, Garcia-Gomez JM, Terol J, Talon M, Robles M colysis pathway mostly related to metabolism of Firmicutes, (2005) Blast2GO: a universal tool for annotation, visualization particularly by clostridia. Application of consortia on one- and analysis in functional genomics research. Bioinformatics step bioprocessing of biomass to ethanol and other valorized 21:3674–3676 fermentation products has been reported (He et al. 2011;Fu Crespo CF, Badshah M, Alvarez MT, Mattiasson B (2012) Ethanol production by continuous fermentation of D-(+)-cellobiose, D- and Holtzapple 2010). Understanding the relationships and (+)-xylose and sugarcane bagasse hydrolysate using the thermoa- metabolic flux among key microbes in the consortia for naerobe Caloramator boliviensis. Bioresour Technol 103:186–191 further “consortium engineering” is essential in paving the Doi RH (2008) Cellulases of mesophilic microorganisms cellulosome way for future development of consolidated bioprocessing and noncellulosome producers. Ann NY Acad Sci 1125:267–279 Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R (2011) for one-step conversion of plant biomass to valorized fer- UCHIME improves sensitivity and speed of chimera detection. mentation products based on the lignocellulolytic consortia Bioinformatics 27:2194–2200 as established in this study. Direct application of the con- FitzPatrick M, Champagne P, Cunningham MF, Whitney RA (2010) A sortia for accelerating degradation of fibrous materials in biorefinery processing perspective: treatment of lignocellulosic materials for the production of value-added products. Bioresour anaerobic digestion process for biogas production from ag- Technol 101:8915–8922 ricultural wastes (Zhang et al. 2011) is also promising. Fengel D, Wegener G (1984) Wood: chemistry, ultrastructure, reac- In conclusion, the pyrosequencing approach was employed tions. De Gruyter, Berlin, Germany to elucidate the structures and metabolic capability of struc- Feng Y, Yu Y, Wang X, Qu Y, Li D, He W, Kim BH (2011) Degradation of raw corn stover powder (RCSP) by an enriched microbial consortium turally stable symbiotic aero-tolerant lignocellulose-degrading and its community structure. Bioresour Technol 102:742–747 consortia. With the high-throughput metagenomic analysis, Fu Z, Holtzapple MT (2010) Consolidated bioprocessing of sugarcane our study reveals the effects of the enrichment process on bagasse and chicken manure to ammonium carboxylates by a shaping composite community structures of the symbiotic mixed culture of marine microorganisms. Bioresour Technol 101:2825–2836 consortia, its lignocellulolytic enzyme systems, and its meta- Guo P, Zhu W, Wang H, Lü Y, Wang X, Zheng D, Cui Z (2010) bolic potential. The consortium provides a promising model to Functional characteristics and diversity of a novel lignocelluloses study lignocellulose-degradation mechanism and serves as a degrading composite microbial system with high xylanase activ- platform for further exploration of multispecies lignocellulo- ity. J Microbiol Biotechnol 20:254–264 Haruta S, Cui Z, Huang Z, Li M, Ishii M, Igarashi Y (2002) Construction lytic enzyme systems for biotechnological application. of a stable microbial community with high cellulose-degradation ability. Appl Microbiol Biotechnol 59:529–534 Acknowledgments This project is under the JSPS-BIOTEC collab- He Q, Hemme CL, Jiang H, He Z, Zhou J (2011) Mechanisms of orative project supported by a research grant from the National Science enhanced cellulosic bioethanol fermentation by co-cultivation of and Technology Development Agency (P-11-00926). Proofreading and Clostridium and Thermoanaerobacter spp. Bioresour Technol comments by Dr. Philip Shaw and Dr. Varadom Charoensawan are 102:9586–9592 appreciated. Kanokratana P, Uengwetwanit T, Rattanachomsri U, Bunterngsook B, Nimchua T, Tangphatsornruang S, Plengvidhaya V, Champreda V, Eurwilaichitr L (2010) Metagenomic analysis on phylogenetics and metabolic potential of microbial community in primary trop- References ical peat swamp forest. Microb Ecol 61:518–528 Kato S, Haruta S, Cui ZJ, Ishii M, Igarashi Y (2004) Effective cellulose degradation by a mixed-culture system composed of a cellulolytic Allgaier M, Reddy A, Park JI, Ivanova N, D’haeseleer P, Lowry S, Clostridium and aerobic non-cellulolytic bacteria. FEMS Sapra R, Hazen TC, Simmons BA, Vander Gheynst JS, Microbiol Ecol 51:133–142 Hugenholtz P (2010) Targeted discovery of glycoside hydrolases Kato S, Haruta S, Cui ZJ, Ishii M, Igarashi Y (2005) Stable coexistence from a switchgrass-adapted compost community. PLoS One 5: of five bacterial strains as a cellulose-degrading community. Appl e8812 Environ Microbiol 71:7099–7106

175 Kato S, Haruta S, Cui ZJ, Ishii M, Igarashi Y (2008) Network relation- Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister ships of bacteria in a stable mixed culture. Microb Ecol 56:403– EB, Lesniewski RA, Oakley BB, Parks DH, Robinson CJ, Sahl 411 JW, Stres B, Thallinger GG, van Horn DJ, Weber CF (2009) Li L-L, McCorkle SR, Monchy S, Taghavi S, van der Lelie D (2009) Introducing mothur: open-source, platform-independent, Bioprospecting metagenomes: glycosyl hydrolases for converting community-supported software for describing and comparing mi- biomass. Biotechnol Biofuel 2:10 crobial communities. Appl Environ Microbiol 75:7537–7541 Lu Y, Zhang Y, Lynd LR (2006) Enzyme-microbe synergy during TAPPI (1992) TAPPI test method: determination of cellulose and cellulose hydrolysis by Clostridium thermocellum. Proc Nat pentosan contents of wood. Technical Association of Pulp and Acad Sci USA 103:16165–16169 Paper Industry (TAPPI) Press, Atlanta Lukashin AV, Borodovsky M (1998) GeneMark.hmm: new solutions Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naive Bayesian for gene finding. Nucleic Acids Res 26:1107–1115 classifier for rapid assignment of rRNA sequences into the new Lynd LR, Weimer PJ, Van Zyl WH, Pretorius IS (2002) Microbial bacterial taxonomy. Appl Environ Microbiol 73:5261–5267 cellulose utilization: fundamentals and biotechnology. Microbiol Wang W, Yan L, Cui Z, Gao Y, Wang Y, Jing R (2011) Characterization Mol Biol Rev 66:506–739 of a microbial consortium capable of degrading lignocellulose. Meyer M, Stenzel U, Hofreiter M (2008) Parallel tagged sequencing on Bioresour Technol 102:9321–9324 the 454 platform. Nat Protoc 3:267–278 Warnecke F, Luginbühl P, Ivanova N, Ghassemian M, Richardson TH, Mitchell WJ (1998) Physiology of carbohydrate to solvent conversion Stege JT, Cayouette M, McHardy AC, Djordjevic G, Aboushadi by clostridia. Adv Microb Physiol 39:31–130 N, Sorek R, Tringe SG, Podar M, Martin HG, Kunin V, Dalevi D, Narisawa N, Haruta S, Cui ZJ, Ishii M, Igarashi Y (2007) Effect of Madejska J, Kirton E, Platt D, Szeto E, Salamov A, Barry K, adding cellulolytic bacterium on stable cellulose-degrading mi- Mikhailova N, Kyrpides NC, Matson EG, Ottesen EA, Zhang X, crobial community. J Biosci Bioeng 104:432–434 Hernandez M, Murillo C, Acosta LG, Rigoutsos I, Tamayo G, Nimchua T, Thongaram T, Uengwetwanit T, Pongpattanakitshote S, Green BD, Chang C, Rubin EM, Mathur EJ, Robertson DE, Eurwilaichitr L (2011) Metagenomic analysis of novel Hugenholtz P, Leadbetter JR (2007) Metagenomic and functional lignocellulose-degrading enzymes from higher termite guts inhab- analysis of hindgut microbiota of a wood-feeding higher termite. iting microbes. J Microbiol Biotechnol 22:462–469 Nature 450:560–565 Okura N, Soneura M, Ninomiya K, Katakura Y, Shioya S (2008) Weber CF, Stubner S, Conrad R (2001) Bacterial populations coloniz- Biological detoxification of waste house wood hydrolysate using ing and degrading rice straw in anoxic paddy soil. Appl Environ Ureibacillus thermosphaericus for bioethanol production. J Microbiol 67:1318–1327 Biosci Bioeng 106:128–133 Wongwilaiwalin S, Rattanachomsri U, Laothanachareon T, Eurwilaichitr L, Pope PB, Denman SE, Jones M, Tringe SG, Barry K, Malfatti SA, Igarashi Y, Champreda V (2010) Analysis of a thermophilic lignocel- McHardy AC, Cheng JF, Hugenholtz P, McSweeney CS, lulose degrading microbial consortium and multi-species lignocellu- Morrison M (2010) Adaptation of herbivory by the Tammar lolytic enzyme system. Enzym Microb Technol 47:283–290 wallaby includes bacterial and glycoside hydrolase profiles dif- Zhang Q, He J, Tian M, Maoa Z, Tang L, Zhang J, Zhang H (2011) ferent from other herbivores. Proc Nat Acad Sci 107:14793– Enhancement of methane production from cassava residues by 14798 biological pretreatment using a constructed microbial consortium. Rattanachomsri U, Tanapongpipat S, Eurwilaichitr L, Champreda V Bioresour Technol 102:8899–8906 (2009) Simultaneous non-thermal saccharification of cassava pulp Zhu H, Qu F, Zhu LH (1993) Isolation of genomic DNAs from plants, by multi-enzyme activity and ethanol fermentation by Candida fungi and bacteria using benzyl chloride. Nucleic Acids Res tropicalis. J Biosci Bioeng 107:488–493 21:5278–5280 Ren N, Xing D, Rittmann BE, Zhao L, Xie T, Zhao X (2007) Microbial Zuroff TR, Curtis WR (2012) Developing symbiotic consortia for community structure of ethanol type fermentation in bio- lignocellulosic biofuel production. Appl Microbiol Biotechnol hydrogen production. Environ Microbiol 9:1112–1125 93:1423–1435

176