A pyrosequencing-based metagenomic study of methane-producing microbial community in solid-state biogas reactor.

1.1 MB·English

by  Li, An

Lietal.BiotechnologyforBiofuels2013,6:3
http://www.biotechnologyforbiofuels.com/content/6/1/3
RESEARCH Open Access
A pyrosequencing-based metagenomic study of
methane-producing microbial community in
solid-state biogas reactor
An Li1†, Ya’nan Chu2†, Xumin Wang2, Lufeng Ren2, Jun Yu2, Xiaoling Liu1, Jianbin Yan1, Lei Zhang1,
Shuangxiu Wu2* and Shizhong Li1*
Abstract
Background: Asolid-state anaerobic digestion method is used to produce biogas from various solid wastes in
China but the efficiency of methane production requires constant improvement.The diversity and abundanceof
relevant microorganisms play important roles in methanogenesis of biomass.The next-generation high-throughput
pyrosequencingplatform (Roche/454 GS FLX Titanium) provides a powerful tool for thediscovery ofnovel
microbes within the biogas-generatingmicrobialcommunities.
Results: Toimprovethe power of our metagenomic analysis, we first evaluated five different protocolsfor
extracting total DNA from biogas-producing mesophilic solid-state fermentation materials and then chose two
high-quality protocols for a full-scaleanalysis.The characterization of both sequencing reads and assembled contigs
revealed thatthe mostprevalent microbes of thefermentationmaterials are derived from Clostridiales (Firmicutes),
which contribute to degrading both protein and cellulose. Other important bacterialspecies for decomposing fat
and carbohydrateare Bacilli, Gammaproteobacteria, and Bacteroidetes(belonging to Firmicutes, Proteobacteria, and
Bacteroidetes, respectively).The dominantbacterial species are from six genera: Clostridium, Aminobacterium,
Psychrobacter, Anaerococcus,Syntrophomonas,and Bacteroides.Among them, abundant Psychrobacter species, which
produce low temperature-adaptive lipases, and Anaerococcus species, which have weak fermentation capabilities,
were identified for the first time inbiogas fermentation. Archaea, represented bygenera Methanosarcina,
Methanosaeta and Methanoculleus of Euryarchaeota,constitute onlya small fraction of theentire microbial
community. The most abundant archaealspeciesinclude Methanosarcina barkeri fusaro,Methanoculleus marisnigri
JR1,and Methanosaeta theromphila,and all are involved inboth acetotrophic and hydrogenotrophic
methanogenesis.
Conclusions: The identification of new bacterial genera and speciesinvolved in biogas production provides
insights into novel designs of solid-state fermentationunder mesophilic or low-temperature conditions.
Keywords: Solid-state fermentation, Biogas production, Pyrosequencing,Metagenomics, DNA extraction,
Anaerococcus,Psychrobacter
*Correspondence:[email protected];[email protected]
†Equalcontributors
2TheCASKeyLaboratoryofGenomeSciencesandInformation,Beijing
InstituteofGenomics,ChineseAcademyofSciences,No.1-7BeichenWest
Road,ChaoyangDistrict,Beijing100101,China
1InstituteofNuclearandNewEnergyTechnology,TsinghuaUniversity,
TsinghuaGarden,HaidianDistrict,Beijing100084,China
©2013Lietal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative
CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and
reproductioninanymedium,providedtheoriginalworkisproperlycited.
Lietal.BiotechnologyforBiofuels2013,6:3 Page2of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
Background wastewaters [9,10], thermophilic upflow anaerobic filter
Due to fossil fuel crisis, atmospheric pollution, and global reactor [11], fully- stirred reactor fed with fodder beet
warming, the development of renewable and clean energy silage [12], thermophilic municipal biogas plant [13], and
forms has become a critical task for the human society. two-phaseliquidbiogasreactoroperatedwithsilages[14],
The production of biogas through biomass fermentation, have been conducted. Thermophilic anaerobic municipal
regarded as an environment-friendly, clean, and renewable solid-waste digester [15] and packed-bed reactor for
resource, has been gaining more attention in many deve- degrading organic solid wastes of artificial garbage slurry
loped and developing countries [1,2]. In China, solid [16] were also studied. However, the methodology used
biomass wastes (SW), such as kitchen, livestock, and for these studies was based on constructing and sequen-
agricultural wastes (largely crop straws and stalks), cing 16S rDNA and mcrA clone libraries, and the choice
are produced at the multi-million ton level annually of PCR primers for amplifying sequence fragments of the
[3] and the untreated disposals of such wastes may target genes and other sequences typically creates biases,
lead to severe long-term environmental hazards and and it has been difficult to cover the entire complexity of
resource wasting. Therefore, the utilization of anaerobic microbial communities based on just the sequences from
fermentation to convert SW into biogas represents a a limited number of gene-specific clones. The next-
promising effort if it can be accomplished at an industrial generation sequencing technologies have overcome many
scale and in an economical way. In recent years, solid- of these problems, particularly the pyrosequencing
state anaerobic digestion (SS-AD) has been promoted in platform (such as the Roche/454 GS FLX sequencer) that
China because of its many advantages, including less generates longer readlengthsranging from 200to400bp
reactor-capacity demand, lower heating-energy need, and as compared to other platforms (such as the Illumina
no stirring-energy consumption, particularly as opposed Hiseq2000 system that generates 50–150 bp reads in its
to liquid-state anaerobic digestion [2]. However, the yield single-directional sequencing runs [17-20]) and creates
of methane, the major end-product of this process, has less bias in sequencinglibrary construction [21,22]. Based
not been sufficient for an industrial-scale promotion, let on this platform, a German group conducted the first
aloneeconomicalplausibility. metagenomic analysis on a complex system of biogas-
The biochemical process for anaerobic methane produc- producingplant [20],and developed relatedbioinformatic
tion is complex. The diversity and abundance of microbes methods[23,24].They further revealed that in addition
involved in the process certainly play a major role, which to the archaeal methanogen Methanoculleus species
are influenced by microbial community compositions, (which play a dominant role in methane production)
fermentation materials, climate variations, and designs of and abundant numbers of cellulolytic Clostridia
chambers,tonamejustafew.IntheinitialstepsofSS-AD, (which were important for the hydrolysis of cellulosic
hydrolytic Firmicutes reduce large macromolecules (inclu- plant biomass for acetogenesis) other methanogen
dingproteins,complexfats,andpolycarbohydrates)totheir taxa (including Streptococcus, Acetivibrio, Garciella,
building blocks (i.e., amino acids, long-chain fatty acids, Tissierella, and Gelria) are also detected but their
and monosugars) and other bacteria (including acidogens precise functional roles in methane formation remain
and acetogens) further degrade them into smaller inter- to be elucidated [17,25]. A similar study that used a
mediates (such as acetate, carbon dioxide, and hydrogen). SOLiD™ short-read DNA sequencing platform has
In the later steps, methanogens, which are mainly derived recentlyconfirmedtheimportanceofhydrogenmetabolism
fromArchaea, convertthesmaller substratesinto methane in biogas production [26]. Nevertheless, a metagenomic
through both aceticlastic and hydrogenotrophic pathways study on the SS-AD system based on deep-sampling and
[4-6].Therefore,athoroughunderstandingofcomposition, long-read sequencing supported by the next-generation
structure, and function of the microbial communities resi- sequencing platforms is of essence in moving the
ding in anaerobic reactors is crucial for developing novel field forward.
fermentationstrategiesandimprovingmethaneyieldofthe Aside from sequencing and bioinformatic analysis, DNA
existingbiogasreactorsaswellasideasfornoveldesigns. extraction and its quality yield from samples of complex
Although the biochemistry and enzymology ofmethano- materials(suchasliquidvs.solidandsourcevs.processing)
genesis for model organisms are well characterized [7], the also greatly affect results of metagenomic sequencing
structure and function of biogas-producing microbial [27,28]. DNA extraction efficiency and quality from biogas
communities have not been sufficiently explored, samples have also been compared to PCR-based analyses
particularly under different anaerobic fermentation [29], but a robust method, particularly for analyzing
conditions. In the past decade or so, investigations of samples from SW biogas fermentation materials and
different biogas-producing systems and waste treatment based on high-throughput shotgun pyrosequencing, has
conditions,includinganaerobicmesophilicsludgedigester yettobereported.Towardthisend,wefirstevaluatedfive
[8], mesophilic anaerobic chemostats fed with synthetic DNA extraction protocols (including four based on
Lietal.BiotechnologyforBiofuels2013,6:3 Page3of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
commercial kits and one derived from the classic Table1ComparisonofDNAyieldandpurityamongfive
chloroform-isoamylalcohol method) for samples collected DNAextractionprotocols
from a mesophilic SS-AD fermenter fed with SW. After Method ParametersofDNAquantityandpurify$
theT-RFLP evaluation (see Methods for details), we then DNAyield(ng/μl) A /A A /A
260 280 260 230
chose the two better protocols and prepared DNA
P 20.5 1.61 0.88
samples for our pyrosequencing-based metagenomic
F 153.6 1.81 0.33
study. Our results have led to novel insights into
E 160.5 1.88 1.89
microorganism composition, gene content, and metabolic
capacityoftheSWfermentation. EY 121.4 1.81 1.82
S ND ND ND
Results and discussion $DNAquantityandpurityweremeasuredfluorometricallybasedon
EvaluationofDNAextractionmethodsfor NanoDrop.Thepurityparameterconcerningcarbohydrate,phenoland
aromaticcompoundcontaminationswascalculatedbasedontheratioofthe
high-throughputpyrosequencing absorptionsat260nmto230nm(A260/A230).Tocheckforprotein
Biogasfermentationsamplesareextremelycomplexdueto contamination,theratiooftheabsorptionsat260nmto280nmwas
evaluated(A /A ).ND,notdeterminedduetotheimpurityofDNA
the presence of multiple organic compounds and diverse 260 280
solution.ThepurityofDNApreparationswasmeasuredbasedonA /A >
260 280
degradation products. In SS-AD samples, microorganisms 1.8andA /A >1.8.
260 230
bindstronglytosolidmaterialsandhavearatherheteroge-
neous distribution inside the samples. In order to find a which are critical for DNA yield and quality especially
better protocol for the isolation of high-quality DNA when field sampling is the only source [29,30]. According
preparationsforpyrosequencing,wesetouttoevaluatefive to our results, Protocol P (the Mo-Bio PowerSoil DNA
DNA extraction methods. Using electrophoresis assay for Isolation Kit) showed the lowest DNA yield, suggesting
checking quality and yield of genomic DNA extracts insufficient lysing despite the use of vigorous mechanical
(Figure1andTable1),wefoundthatProtocolsE,EY,and force (vortexing for 15 min), especially when compared
Fgaverisetothehighestyields,rangingfrom~160.5ng/μl with the corresponding steps in other related protocols
to ~121.4 ng/μl, while Protocol P produced the lowest (hand shaking for a few minutes or vortexing for 30 s).
yield, with ~20.5 ng/μl. However, Protocol F showed the Therefore, we realized that in addition to mechanical
highestdegreeofsmearing(Figure1)andbothProtocolsF forces, lysis reagents used for the protocols may also be
and S showed low purity based on A /A ratios crucialforpreparingbettercelllysis.
260 230
(Table1).TheDNAextractfromProtocolSappeareddark WefurtherevaluatedtheDNApreparationsfromallfive
yellowish and its quality could not be measured based on methods based on T-RFLP analysis. The Shannon-Weiner
spectrophotometry. The differences among the five index was used to indicate diversity and complexity, and
methods were primarily observed at the cell lysing steps, the Simpson index was used to measure abundance.
Bacteria and archaea were analyzed separately. The
results showed that Protocol E resulted in the highest
bacterial diversity (Shannon-Weiner index of 3.6) and
thehighestabundance(Simpsonindexof0.95;Table2),fol-
lowed by Protocols P and EY. Protocols E and EYshowed
higher archaeal enrichment than that of Protocols P, Fand
S.Wetherefore choseDNAextractsfromProtocolsE and
EY for pyrosequencing, which consistently lead to higher
yield,purerDNA,andhighmicrobialdiversity.
Table2DiversityindexesofT-RFLPanalysisfordifferent
protocols
Method Diversityindexforbacteria Diversityindexforarchaea
Shannon- Simpson Shannon- Simpson
WeinerIndex(H) Index(D) WeinerIndex(H) Index(D)
E 3.578163 0.945856 2.76696 0.88734
Figure1TheelectrophoresisresultsofDNApreparations F 3.293445 0.923434 1.245221 0.561454
basedonthefivemethodsfromabiogasreactorsample.M,
P 3.410967 0.944772 1.984324 0.762865
molecularmarkerTransplus2K.P,F,E,EY,Srefertothefive
protocols,respectively.ExceptforProtocolP,whichwasloadedwith S 3.116197 0.921699 1.634894 0.665427
5μlofundilutedDNAsolution,only1μlofundilutedDNAsolution
EY 3.313463 0.924693 2.670502 0.872466
wasloadedonthe0.8%agarosegelsfortherestoftheextracts.
Foreachprotocol,bacterialandarchaealdiversitieswereanalyzedseparately.
Lietal.BiotechnologyforBiofuels2013,6:3 Page4of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
Sequencingandmetagenomicassembly 100%ofthetotalreads.Ourresultssuggestthatthecurrent
Pyrosequencing of two DNA libraries (from Protocols E samplingdepthisnotyetclose to the natural status for
andEY),namely“BE”and“BEY”,wereperformedandthe bacteria but may be saturated for archaea.
data from the experiments were summarized in Table 3. Matching the sequencing reads from BE-1 and BEY to
ThefirstsequencingrunsofBE(namedasBE-1)andBEY sequences collected in NTand NR databases, we dissected
resulted in 266,781,751 bp sequences from 738,005 reads microbial community structure of the two libraries (exclu-
(an average read length of 362 bp) and 197,514,392 bp dingthereadswithno-hits;Figure2B),showingthatatthe
sequences from 551,339 reads (an average read length of domainlevelthereissignificantdifferencebetweenthetwo
358bp),respectively.Itisobviousthattherearemoredata libraries in the proportion of reads assigned to bacterial,
and higher microbial richness (Figure 2A in section 3.3) archaeal,viral,andeukaryoticsequences(Figure2B).Inthe
obtained from BE (BE-1) than from BEY. Therefore, the BE-1dataset,4.7%and90.9%ofthereadswereassignedto
BEsamplewassequencedtwiceagainasBE-2andBE-3. archaea and bacteria, but decreased to 3.0% and 71.2% for
Since the BE sample was sequenced three times, it those of BEY, respectively. In contrast, only 3.4% of the
yielded 647,369,218 bp sequences from 2,280,601 reads (in reads were assigned to eukaryotes and almost no viral
anaveragereadlengthof283bp).Theassemblyofthetotal sequence was detectable in BE-1, but eukaryotic and viral
readsgaveriseto118,433contigscontaining76,759,543bp, detectionsweresignificantlyincreasedto20.5%and9.3%in
which were accounted for approximately 12% of the total BEY, respectively. The taxonomic bias in the microbial
sequences measured in basepairs generated in this study. communitiesdetectedbetweenthedatasetsfromProtocols
Thenumberoflargecontigs(>500nt)was37,276(anN50 E and EY may reflect the thoroughness of sample pre-
of 1,712 bp), in which the largest contig contains washing with TENP buffer that may partially wash off
158,075bp.TheaverageGCcontentofthetotalreadsfrom bacteria known to be lightly adhered to solid matrix.
theBEsampleis46%(Additionalfile1:FigureS1). Examining the taxonomies built from mapped sin-
gle reads of libraries BE-1 and BEY (Additional file 1:
Comparisonofmicrobialcompositionsbetweensamples Figure S2), we observed that the dominant taxa at
BE-1andBEY the genus level for archaea and bacteria (such as
We used rarefaction analysis to assess species richness of Methanosarcina and Clostridium for the former and
thesystem.UsingMEGAN(ameta-genomicbioinformatic the latter, respectively) are comparable between the two
tool) and at the best resolved levels based on the NCBI libraries; but there were greater numbers of microbial taxa
taxonomy database and our sequence data, we analyzed observed in the BE (combining BE-1, BY-2, and BE-3) and
the microbial richness, based on sequence reads, between BEY libraries (Table 2 and Figure 2A). Therefore, Protocol
libraries BE-1 and BEY (Figure 2A) and revealed that the E in combination with the E.Z.N.A.TM Soil DNA Kit
number of taxonomic leaves or clades of BE-1 are all should be considered as the most suitable procedure for
higher than those of BEY, and the result indicated that theSWfermentationsamples.
BE-1 contains more microbial taxa than BEY, and indeed
BE-1 and BEYcontain 717 and 643 leaves for all assigned Microbialcompositionanalysisbasedonsequencing
taxa, respectively. Furthermore, the rarefaction curves of readsandassembledcontigs
both libraries in archaea appear close to saturation at 20% Weanalyzedthemicrobialcommunitycompositionofthe
ofthetotalreads,whereasthoseinbacteriaareincreasedto BE library using MEGAN and mapped individual reads
Table3Summaryofsequencingresults
BEY BE-1 BE-2 BE-3 BE(total)
Numberofreads 551,339 738,005 781,293 761,304 2,280,601
Numberofbases 197,514,392 266,781,751 218,761,141 161,838,822 647,369,218
Averagereadlength(bp) 358 362 280 212.58 283
Numberoflargecontig(>500nt) 11897 17,933 10,861 7,354 37,276
Numberofallcontig 24,052 33,750 20,281 17,860 118,433
Maxcontiglength 18,435 35,598 35,601 11,781 158,075
Numberofbasesinallcontigs 16,193,286 25,185,216 16,441,312 11,362,543 76,759,543
Assembledbases% 8.2% 9.44% 7.5% 7% 11.9%
N50(largecontigs)(bp) 1105 1204 1440 1231 1712
TheBEandBEYsequenceswereassembledseparatelybyusingGSdenovoassembler.
Lietal.BiotechnologyforBiofuels2013,6:3 Page5of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
A
800
700
600
s
e
v
a 500
e
l
f
r o 400 BE-1-total BEY-total
e
b
m
u 300 BE-1-Bac BEY-Bac
N
200 BE-1-Arc BEY-Arc
100
0
0 20% 40% 60% 80% 100%
Percentage of reads
B
100
80
s
d BE-1
a
e
r 60 BEY
f
o
t
n 40
e
c
r
e
P 20
0
Bacteria Archaea Eukaryota Viruses
Taxon
Figure2Rarefactioncurves(A)andtherelativeproportionsoftaxonomicclassificationofbacteria,archaea,eukaryota,andviruses(B)
fromtheBE-1andBEYlibraries.Fortherarefactioncurves,theanalysiswasperformedonthetotal(includingbacteria,archaea,eukaryota,
virusesandenvironmentalsequences),archaeal,andbacterialtaxafromthetwolibraries.
first. Figure 3 shows the statistics of the total assigned (39.0% of hit-reads), followed by Proteobacteria (17.3% of
reads and their annotations for the popular genera and hit-reads)and Bacteroidetes(7.0%of hit-reads), whichare
phyla. The reads assigned to the superkingdoms Bacteria responsible for biomass degradation and fermentation.
(~9.8%)andArchaea(~0.5%)wereaccountedforapproxi- The 4th most abundant taxon is Euryarchaeota (4.3% of
mately10.3%ofthetotalreads,whereas88.4%ofthetotal hit-reads), involved in methane synthesis and taking a
reads obtained have no hit in the present database, small fraction of the community. In addition, massive
indicating that there are still an immense amount of bacterial taxa are distributed in phyla Thermotogae
unknown/uncultured species in this complex anaerobic (2.4% of hit-reads), Actinobacteria (2.2% of hit reads),
biogas-producingsample(Figure3A). Chloroflexi (1.2% of hit-reads), Cyanobacteria (0.4% of
Based on single-read assignments (Figure 3B), the most hit-reads), Chlorobi (0.3% of hit-reads), and Fusobacteria
prevalentbacterialtaxaatthephylumlevelareFirmicutes (0.3%ofhit-reads),andtheresultagainindicatesthatthere
Lietal.BiotechnologyforBiofuels2013,6:3 Page6of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
A B
100000 12000
ds 80000 Number of reads Number of contigs 10000 gs
r of rea 60000 68000000 of conti
be r
m 40000 be
u 4000 m
N u
20000 N
2000
0 0
C
20,000 2,500
s
d s
of rea 1105,,000000 Number of reads Number of contigs 12,,500000 contig
mber 5,000 510,0000 er of
b
Nu 0 0 m
u
Clostridium Aminobacterium Psychrobacter Anaerococcus Bacteroides Syntrophomonas Pseudomonas Alkaliphilus Streptococcus rmoanaerobacter Enterococcus Lactobacillus Kosmotoga Enterobacter Bacillus Methanosarcina Methanoculleus Methanosaeta N
e
h
T
Bacteria Archaea
Figure3Thestatisticsforthereadsassigned(A),thenumberofreads/contigsforthe11mostprevalentphyla(B),andthenumberofreads/
contigsof15mostprevalentgenerainbacteriaandthe3mostprevalentgenerainarchaea(C)obtainedinthetaxonomicclassification
analysisusingBLASTN/BLASTXtoolsagainsttheGenBankNT/NRdatabasewithaE-valuecutoffof10-5basedontotalreads/contigs.
are more complex microbial components residing within Psychrobacter (7,823 reads, 4.9% of hit-reads), Anaerococ-
this system and that some of the components may reflect cus (6,544 reads, 2.7% of hit-reads), Bacteroides (5,655
their original environments rather than characteristic of reads,2.3%ofhit-reads),andSyntrophomonas(4698reads,
theSWfeedsingeneral. 1.9% of hit-reads). For archaeal species, the three most
At the class level, the prevalent reads are distributed prevalent genera are Methanosarcina (6,522 reads, 2.7%),
over Clostridia (66,455 reads, 27.2% of hit-reads), Bacilli Methanoculleus (1,102 reads, 0.5%), and Methanosaeta
(16,767reads, 6.9% of hit-reads), Gammaproteobacteria (750reads,0.3%),whichallbelongtoclassMethanomicro-
(19,085 reads, 7.8% of hit-reads), Bacteroidetes (11,765 biaofEuryarchaeota(Table4,Figure3C).
reads, 4.8% of hit-reads), and Methanomicrobia (9180 At the species level (Figure 4), although Clostridium is
reads, 3.8% of hit-reads), which belong to phyla Firmi- the predominant genus, the three most abundant bacte-
cutes, Proteobacteria, Bacteroidetes, and Euryarchaeota, rial species are Aminobacterium colombiense DSM 12261
respectively (Table4). (11,868 reads, 5.2%), Anaerococcus prevotii DSM 20548
At the genus level, there are 429 generaof bacterial and (6,507 reads, 2.9%), Syntrophomonas wolfei subsp. Wolfei
39 generaof archaeal origins. The six most prevalent gen- str. Goettingen (4,685 reads, 1.9%), while the dominant
era for bacteria are Clostridium (17,975 reads, 7.4% of hit- Clostridium species were identified as C. thermocellum
reads), Aminobacterium (11,870 reads, 5.2% of hit-reads), ATCC 27405 (4,204 reads, 1.7%), C. tetani E88 (1,378
Lietal.BiotechnologyforBiofuels2013,6:3 Page7of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
Table4The21mostprevalentgeneraoftaxonomicclassificationofsampleBEbasedonreadcounts
Phylum Class Order Family Genus Countsofreads
Firmicutes Clostridia Clostridiales Clostridiaceae Clostridium 17,975
Clostridia Clostridiales Syntrophomonadaceae Aminobacterium 11,870
Clostridia Clostridiales Peptostreptococcaceae Anaerococcus 6,544
Clostridia Clostridiales Syntrophomonadaceae Syntrophomonas 4,698
Clostridia Clostridiales Clostridiaceae Alkaliphilus 2,637
Bacilli Lactobacillales Enterococcaceae Enterococcus 1,948
Bacilli Lactobacillales Streptococcaceae Streptococcus 2,462
Clostridia Clostridiales Syntrophomonadaceae Thermanaerovibrio 1,123
Bacilli Bacillales Bacillaceae Bacillus 1,282
Clostridia Thermoanaerobacteriales Thermoanaerobacteriaceae Thermoanaerobacter 2,302
Bacilli Lactobacillales Lactobacillaceae Lactobacillus 1,907
Bacteroidetes Bacteroidetes Bacteroidales Bacteroidaceae Bacteroides 5,655
Bacteroidetes Bacteroidales Porphyromonadaceae Parabacteroides 1,295
Proteobacteria Gammaproteobacteria Pseudomonadales Moraxellaceae Psychrobacter 7,823
Gammaproteobacteria Pseudomonadales Pseudomonadaceae Pseudomonas 2,850
Gammaproteobacteria Enterobacteriales Enterobacteriaceae Enterobacter 1,339
Chloroflexi Chloroflexi Chloroflexales Chloroflexaceae Roseiflexus 968
Thermotogae Thermotogae Thermotogales Thermotogaceae Kosmotoga 1,738
Euryarchaeota Methanomicrobia Methanosarcinales Methanosarcinaceae Methanosarcina 6,522
Methanomicrobia Methanomirobiales Methanomicrobiaceae Methanoculleus 1,102
Methanomicrobia Methanosarcinales Methanosaetaceae Methanosaeta 750
#Readcountsmatchedtothe21mostabundantmicrobialgenera.TheresultswereobtainedfromthebestBLASTxhitsbysearchingthenucleotidesequence
databaseofNCBIGenBank.
reads, 0.6%),C. kluyveri(1,100reads.0.5%),and C.phyto- Since16SrDNAiswidelyusedfortaxonomicandphylo-
fermentans ISDg (607 reads, 0.3%). In archaea, the three genetic studies due to its highly conserved sequences in
most prevalent species are Methanosarcina barkeri str. bothbacteriaandarchaeaanditshypervariableregioncan
Fusaro(2,611reads,1.1%),MethanoculleusmarisnigriJR1 also be used for accurate taxonomic evaluation, we
(1,101 reads, 0.5%), Methanosaeta theromphila PT extracted793contigs(only0.7%oftotalcontigs)thatcon-
(724 reads, 0.3%). tain 16S rDNA sequences (an average length of 1,068 bp)
Meanwhile, we also analyzed microbial community for further analysis. When submitted to the RDP database
compositions based on the assembled contigs, with (with 80% confidence), approximately 68.6% and 1.3% of
BLASTN (version 2.2.13) against the NTand NR data- themwereclassifiedintobacteriaandarchaea,respectively.
bases with E-value cutoff of 10-5. Among a total of At the class level, the dominant taxa include Clostridia,
118,433 contigs, 26,332 of them (22.2%) were assigned to Anaerolineae, Synergistia, Methanomicrobia, Bacilli, and
356 taxa (genus), containing 330 bacterial genera and 26 Gammaproteobacteria (Additional file 1: Table S2), mostly
archaeal genera. Comparable to the taxonomic structure from Firmicutes, Proteobacteria and Euryarchaeota. It is
generated from the output ofBLAST based onreads, our noteworthy that the detection of classes Anaerolineae and
analysis showed that Firmicutes (32.2%), followed by Pro- Synergistia to be the dominant taxa based on 16S rDNA
teobacteria (14.1%), Bacteroidetes (3.8%), and Euryarch- sequences differs from those based on reads and contigs.
aeota (5.5%), are most dominant (Figure 3B, Additional Thereasonsaremorecomplex.Oneofthemmaybeinfor-
file1:TableS1).ThedominantclassesinbacteriaareClos- mation loss in short contigs and sequence assemblies of
tridia (8,249 contigs), Gammaproteobacteria (1,972 con- low-abundance species that are difficult to annotate based
tigs), Bacilli (484 contigs), Bacteroidetes (315 contigs), and on limited matches. Another may be due to the lower
thoseinarchaeawere mappedtoMethanomicrobia(1,279 matching rate for the 16S-associated contigs, where only
contigs). The 17 most prevalent genera and the 6 most approximately 7% of the contigs were classified at the
prevalent species are also consistent with the taxonomic genus level. Therefore, an even lower number of 16S
structure based on mapped reads (Figure 3C, Additional rDNA geneswas detected and assigned tothe profilewith
file1:TableS1). significantcertainty.
Lietal.BiotechnologyforBiofuels2013,6:3 Page8of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
Number of reads
0 2000 4000 6000 8000 10000 12000 14000
Aminobacterium colombiense DSM 12261 11868
Anaerococcus prevotii DSM 20548 6507
Syntrophomonas wolfei subsp. Wolfei str. Goettingen 4685
Clostridium thermocellum ATCC 27405 4204
Methanosarcina barkeri str. Fusaro 2608
Thermanaerovibrio acidaminovorans DSM 6589 1948
Kosmotoga olearia TBF 19.5.1 1738
Alkaliphilus metalliredigens QYMF 1469
Psychrobacter cryohalolentis K5 1431
Bacteroides thetaiotaomicron VPI-5482 1406
Bacteroides vulgatus ATCC 8482 1402
Clostridium tetani E88 1378
Parabacteroides distasonis ATCC 8503 1295
Enterobacter cloacae subsp. cloacae ATCC 13047 1156
Finegoldia magna ATCC 29328 1138
Methanoculleus marisnigri JR1 1101
Clostridium kluyveri 1100
Alkaliphilus oremlandii OhILAS 1046
Methanosaeta theromphila PT 724
Candidatus Azobacteroides pseudotrichonymphae… 654
Clostridium phytofermentans ISDq 607
Clostridium celluloyticum H10 595
Clostridium novyi NT 565
Psychrobacter sp.PRwf-1 524
Enterococcus faecalis V583 520
Alistipes shahii WAL 8301 491
Bacillus megaterium QM B1551 486
Lysinibacillus sphaericus C3-41 453
Clostridium beijerinckii NCIMB 8052 435
Psychrobacter arcticus 273-4 425
Roseiflexus castenholzii DSM 13941 424
Heliobacterium modesticaldum Ice1 407
Roseiflexus sp. RS-1 385
Prevotella ruminicola 23 366
Clostridium acetobutylicum ATCC 824 354
Syntrophobacter fumaroxidans MPOB 331
Herpetosiphon aurantiacus 328
Methanosarcina acetivorans 327
Pedobacter heparinus DSM 2366 326
Petrotoga mobilis SJ95 324
Desulfitobacterium hafniense Y51 323
Faecalibacterium prausnitzii SL3/3 320
Desulfomicrobium baculatum DSM 4028 298
Methanosarcina mazei 292
Streptococcus gallolyticus UCN34 292
Fusobacterium nucleatum subsp. Nucleatum ATCC… 290
Thermoanaerobacter tengcongensis MB4 283
Pirellula staleyi DSM 6068 282
Symbiobacterium thermophilum IAM 14863 249
Halothermothrix orenii H 168 242
Figure4StatisticsforthereadsassignedtomicrobialgenomesequencesusingBLASTN/BLASTXtoolsagainsttheGenBankNT/NR
databasewithanE-valuecutoffof10-5basedonthetotalreads.Thex-axisdenotesthenumberofreadsassignedtothe50mostprevalent
microbialstraingenomes.
Lietal.BiotechnologyforBiofuels2013,6:3 Page9of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
Novelanaerococcusandpsychrobacterandtheir thefermentationpowerisweak[33].However,theinvolve-
characteristics ment of abundant Anaerococcus species in anaerobic fer-
Weidentified abundant reads assigned to two novel bacte- mentation has never been reported. The function of
rialgenerainthefermentation,andamongthese,thegenus abundant Anaerococcus species in the fermentation
Anaerococcus of class Clostridia was identified, which was remains unknown. Its existence in the fermentation
representedbythesecondmostabundantbacterialspecies, may be contaminations from kitchen wastes or variability
A.prevotiiDSM20548(alsoasA.prevotiiPC1,Peptococcus ofthegenusitself;althoughgeneticallyidentical,thespecies
prevotii) (6,507 reads, 2.7%), an obligate anaerobic bacte- possess significant discrepancies and are subject to adapta-
rium. A. prevotii often presents in oral cavity, skin, tion under certain fermentation conditions. In this study,
vagina, gut [31], and deep-seated soft tissue, causing several important enzymes in the metabolic pathways for
abscessesoranaerobicinfectionsinhumans[32].Thusfar, methane synthesis were detected in association with
in Anaerococcus, only this strain has been completely Anaerococcus based on the results of the KEGG analysis
sequencedduetoitsclinicalsignificance,andthehit-reads (Figure 5, Additional file 1: Table S3), such as ackA
forAnaerococcuswereallassignedtothisstrain.Therefore, (acetate kinase, EC: 2.7.2.1), pta (phosphate acetyl-
at the species level, the taxonomic prediction should be transferase, EC: 2.3.1.8) and transporters/antiporters,
treated with caution. Our taxonomic assignment depends including NhaC and V-type ATPase subunit D (EC:
on the comparison of amino acid sequences deduced 3.6.3.14), indicating that Anaerococcus species likely
from reads encoding protein sequences of known taxo- contribute to the methane synthesis in the fermentation.
nomicorigin.Therefore, onlypreviouslysequencedspecies Therefore, it is necessary to isolate Anaerococcus species,
can be identified and there are possibilities that other to characterize their phylogenetic relationships, and to
Anaerococcus species do exist in the fermenter but study their biological and ecological functions in SS-AD
conformations from further studies are inevitable. andmethanogenesisfromSWinfuturestudies.
Accordingtotheliterature,mostspeciesinAnaerococcus The genus Psychrobacter (7,823 reads, 3.2%) of class
are capable of fermenting several carbohydrates, although Gammaproteobacteria(phylumProteobacteria)isprimarily
Figure5MethanogenesispathwaypredictedinourfermenterbasedontheKEGGanalysis.Theellipsesdenotethesubstancesinvolvedin
thereaction.Theboxesshowtheenzymesinvolvedinmethanogenesis.Theacetotrophic(green),thehydrogenotrophic(blue),andtheshared
pathways(grey)areallcolor-encodedandthecolordepthindicatestheamountofreadsassigned.Theenzymesthatarepotentiallyassociated
withAnaerococcusandPsychrobacterinthemethanogenesispathwayareshownintheyellowbox.
Lietal.BiotechnologyforBiofuels2013,6:3 Page10of17
http://www.biotechnologyforbiofuels.com/content/6/1/3
comprised of P. cryohalolentis (1,431 reads, 0.6%), P. 12261andtherelativespeciesofAminobacteriumareboth
arcticus (425 reads, 0.2%), and Psychrobacter sp. syntrophic, capable of anaerobic degradation of amino
PRwf-1 (524 reads, 0.2%). Most species of this genus acids, particularly without saccharides and consistently
can adapt to cold conditions, such as polar permafrost identified in anaerobic environment, such as sludge and
and ice, and are capable of reproducing at temperatures compost [46,9]. The abundance of such amino acid-
rangingfrom−10°Cto40°C[34].Thespeciescanevenbe metabolizing organisms indicates high protein content in
found in the relatively thermophilic environment of the theSWsamples.
digestion conditions. So far, Psychrobacter sp. have been S. wolfei subsp. Wolfei str. Goettingen (4,685 reads,
defined as aerobic and mesophilic bacteria. However, 1.9%) belongs to genus Syntrophomonas (Clostridia),
some researchers have shown that some of their strains often isolated from anaerobic environments, such as
were also able to grow in fermenting environments aquaticsedimentorsewagesludge,growingtogetherwith
[35-37] and anaerobic conditions, such as facultative methanogens (such as Methanospirillum hungatii) and
anaerobic bacteria [38]. The Psychrobacter species often other H -using and/or formate-using microorganisms
2
produce variable lipases (including phenylalanine dea- [47]. S. wolfei subsp. Wolfei str. Goettingen participates in
minase, alkaline phosphatase, esterase (C4), esterase anaerobic fatty acid degradation [48] through the degra-
lipase (C8), lipase (C14), leucine arylamidase, and dation of long-chain fatty acids into acetate and H [49]
2
lecithinase [39,40]) and play essential roles in fat decom- duetotheactivityofacyl-CoAdehydrogenase,CoAtrans-
position reactions. They have been isolated from the facial ferase, enoyl-CoA hydratase, and other related enzymes
and body tissues of animals, poultry carcass, fermented [48], and this species plays a significant role in fatty acid
seafood [38,41-43], and groundwater, but the isolation of decompositionandmethanogenesis.
Psychrobacter species has never been reported in biogas Thereadsfor genusBacteroides(5,655reads,2.1%)were
fermentation samples. In this study, several important mainly assigned to B. thetaiotaomicron VPI-5482 (1,406
enzymes in methane metabolism pathways associated with reads, 0.6%), B. vulgatus ATCC 8482 (1,402 reads, 0.6%),
thisgenuswerealsodetected basedontheKEGGanalyses and B. fragilis YCH46 (183 reads, 0.1%). Bacteroides often
(Figure 5, Additional file 1: Table S3), including ackA (ace- reside in human and animal intestines so that they exhibit
tate kinase, EC: 2.7.2.1), pta (phosphate acetyltransferase, symbiotic relationship with E. coli and other species.
EC: 2.3.1.8), acetyl-CoA synthetase (EC: 6.2.1.1) and trans- Bacteroides are involved in the fermentation of dietary
porters/antiporters (such as NhaA and NhaC), and the polysaccharides, utilization of nitrogenous substances, and
results indicate that these enzymes most likely participate biotransformationofbileacidsandothersteroidsinhuman
in fat hydrolysis in SW samples for methanogenesis, colon [50]. Most intestinal bacteria are saccharolytic
particularly under mesophilic conditions. and obtain carbon and energy through hydrolysis of
Fat hydrolysis is the primary reaction of lipases. More- carbohydrates.
over, lipases catalyze esterification, interesterification, aci- The dominant Clostridium species was identified as C.
dolysis, alcoholysis and aminolysis reactions in addition thermocellum (4,204 reads, 1.7%), which directly con-
to the hydrolytic activity on triglycerides [44]. Therefore, verts cellulosic substrate into ethanol with high effi-
cold-active lipases, largely distributed in psychrophilic ciency and is a good candidate for the degradation of
microorganisms and showing high catalytic activity at low cellulosic materials from plant biomasses [20,51-53].
temperatures, are added to detergents for cold washing, In addition to plants, some Clostridium species can
industrial food fermentation samples, environmental also be isolated from animal feces and cultured with
bioremediation (digesters, composting, oil or xenobiotic Methanobacterium thermoautotrophicum [51]. The
biology applications) and biotransformation processes [44]. tetanus-causing bacterium C. tetani (1,378 reads, 0.6%) is
Some cold-adaptive lipase genes from Psychrobacter sp. anobligateanaerobethatreliesonfermentation.Itcanbe
had been previously cloned and expressed [45]. Therefore, found inmanure-treatedsoil,animalfeces, and fermenta-
the abundance of Psychrobacter sp. in this fermenter tionsamplesfrombiogas-producingplant[20].C.kluyveri
demonstrates great potential for use in SW treatment for (1,100reads.0.5%)growsanaerobically,usingethanoland
methane production or in other bio-energy conversion acetate as sole energy sources to produce butyrate, cap-
processes based on fatty-rich substrates, particularly under roate,andH [54].C. kluyveriisoriginallyidentifiedfrom
2
lowtemperatureconditions,innorthernChina. canal mud [55] and C. phytofermentans (607 reads, 0.3%)
iswidely distributed in soil, capable of producingethanol,
Otherdominantbacterialspeciesandtheircharacteristics acetate, CO , and H through fermenting cellulose [56].
2 2
Themostfrequentlypredictedspeciesinthisfermenteris Therefore, in the initial steps of biomass digestion,
A.colombienseDSM12261(11,868reads,5.1%),primarily members of Clostridium produce a wide variety of
isolated from anaerobic sludge and belongs to genus extracellularenzymestodegradelargebiologicalmolecules
Aminobacterium (Clostridia) [46]. A. colombiense DSM (such as cellulose, xylans, proteins, and lipids into