ww.sciencedirect.com

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 5

Available online at w

journal homepage: www.elsevier .com/locate/watres

Microbial community structure and dynamics intwo-stage vs single-stage thermophilic anaerobicdigestion of mixed swine slurry and market bio-waste

Giuseppe Merlino a, Aurora Rizzi a, Andrea Schievano b, Alberto Tenca b, Barbara Scaglia b,Roberto Oberti b, Fabrizio Adani b, Daniele Daffonchio a,*aDepartment of Food Environmental and Nutritional Sciences (DEFENS), University of Milan, Celoria 2, 20133 Milan, ItalybDepartment of Agricultural and Environmental Science (DiSAA), University of Milan, Celoria 2, 20133 Milan, Italy

a r t i c l e i n f o

Article history:

Received 5 August 2012

Received in revised form

27 November 2012

Accepted 4 January 2013

Available online 18 January 2013

Keywords:

Bacterial and archaeal

anaerobic consortia

Bio-hydrogen

Bio-methane

PCR-DGGE

Real-time PCR

* Corresponding author. Tel.: þ39 0250319117E-mail address: daniele.daffonchio@unim

0043-1354/$ e see front matter ª 2013 Elsevhttp://dx.doi.org/10.1016/j.watres.2013.01.007

a b s t r a c t

The microbial community of a thermophilic two-stage process was monitored during two-

months operation and compared to a conventional single-stage process. Qualitative and

quantitative microbial dynamics were analysed by Denaturing Gradient Gel Electro-

phoresis (DGGE) and real-time PCR techniques, respectively. The bacterial community was

dominated by heat-shock resistant, spore-forming clostridia in the two-stage process,

whereas a more diverse and dynamic community (Firmicutes, Bacteroidetes, Synergistes) was

observed in the single-stage process. A significant evolution of bacterial community

occurred over time in the acidogenic phase of the two-phase process with the selection of

few dominant species associated to stable hydrogen production. The archaeal community,

dominated by the acetoclastic Methanosarcinales in both methanogen reactors, showed

a significant diversity change in the single-stage process after a period of adaptation to the

feeding conditions, compared to a constant stability in the methanogenic reactor of the

two-stage process. The more diverse and dynamic bacterial and archaeal community of

single-stage process compared to the two-stage process accounted for the best degradation

activity, and consequently the best performance, in this reactor. The microbiological per-

spective proved a useful tool for a better understanding and comparison of anaerobic

digestion processes.

ª 2013 Elsevier Ltd. All rights reserved.

1. Introduction a revived and increased interest as an environmentally-

Anaerobic digestion (AD) process is an effective way to treat

organic waste producing energy in the form of biogas of high

calorific value (methane and hydrogen) (Angenent et al., 2004).

This technology has been successfully used to produce

methane since several decades, and recently its use has raised

; fax: þ39 0250319238.i.it (D. Daffonchio).ier Ltd. All rights reserved

friendly alternative to fossil fuel-derived energy.

AD is a complex biological process operated by different

functional groups of microorganisms that convert organic

matter to methane through three major steps (hydrolysis/

acidogenesis, acetogenesis and methanogenesis). AD is

commonly run in single-stage process, however recently

.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 51984

a two-stage design that splits the overall process in two

phases, operated in two reactors in series with production of

hydrogen and methane separately, has been established

(Demirel et al., 2010). In the first-stage reactor, bacteria fer-

ment organic compounds, generally carbohydrates, directly

to hydrogen, carbon dioxide, organic acids and alcohols

(Valdez-Vazquez and Poggi-Varaldo, 2009).

In the second-stage reactor, methanogens convert the re-

sidual energy contained in the high-volatile fatty acids (VFAs)

effluent to bio-methane.

Early studies of bio-hydrogen production were focused on

pure cultures of Clostridia and Enterobacteria fermenting simple

soluble substrates, like starch, glucose or sucrose. However,

pure carbohydrates are very expensive and the use of pure

culture system is problematic as they are prone to con-

tamination. In following studies, mixed cultures fermenting

low cost solid organic waste were proved to be effective, easily

maintained and controlled. In general, the choice of sub-

strates for bio-hydrogen production is based on some major

criteria such as availability, cost, biodegradability and

carbohydrate-content because of the high bio-hydrogen pro-

ducing potential. The use of combined wastewaters has led to

a new path for bio-hydrogen production (Guo et al., 2010). In

particular, co-digestion of nitrogen-rich livestock manure

mixed with carbohydrate-rich materials, has been considered

for bio-hydrogen production (Guo et al., 2010).

Hydrogen is an important energy transfer intermediate in

AD. In the acidogenic phase of a two-phase system, the key

issue is to enable the accumulation of hydrogen, typically

consumed very quickly by microorganisms. Three microbial

groups are key players in hydrogen turnover: H2-producing

fermenting bacteria (HPB), H2-consuming methanogens and

H2-consuming acetogens. In order to facilitate HPB, while

preventing H2-consuming microorganisms, pretreatments

and biokinetic control of parameters, such as, pH and hy-

draulic retention time (HRT) are used. Heat-treatment of the

inoculum, selecting for spore-forming bacteria, decreases

methanogen content. Maintaining a low in-reactor pH and

a high dilution rate prevent the growth of methanogens and

possibly other H2-consuming microorganisms (Valdez-

Vazquez and Poggi-Varaldo, 2009). However, the knowledge

on microbial community structure and dynamics in the

two-stage processes is still limited. Studies focusedmainly on

the microbial qualitative diversity (Jo et al., 2007; Xing et al.,

2008; Luo et al., 2011b) but not on the qualitative and quanti-

tative dynamics of the key functional microbial groups

occurring in the two-stages AD processes. As the microor-

ganisms with their biochemical reactions are the key players

of the process, investigating the complexity of the microbial

community and its dynamics is a prerequisite to understand

the AD process, control it and improving its efficiency.

The aim of this study was to characterize and compare,

qualitatively and quantitatively, the bacterial and archaeal

community of a two-stages and a conventional single-stage

processes, both fed with the same mixture of swine manure

and fruit and vegetable market wastes. In a previous study we

compared the energetic and chemical performances of

a two-stages and a single-stage AD processes run in lab-scale

thermophilic intermittent-continuous stirred tank reactors

(I-CSTR) (Schievano et al., 2012). We found comparable overall

energy recovery for the two processes, though some organic

matter was left undegraded in two-stage process indicating

partial inefficiency. Here we report the results of a study on

the structure and dynamics of microbial communities in the

two processes assessed by PCR-Denaturing Gradient Gel

Electrophoresis (DGGE). We complemented the study of the

microbial communities dynamics by evaluating the temporal

quantitative changes of themajor functional microbial groups

involved in the two processes by quantitative Real-Time PCR.

Our aim was to give a contribution to the knowledge of

microbiological aspects of two types of reactor processes

investigating, together with previous analytical data, micro-

bial signatures associated to the performance of two

processes.

2. Materials and methods

2.1. Bioreactor set up and operation

Three previously described (Schievano et al., 2012) anaerobic

completely stirred tank reactor (CSTR) were operated and

used as source of biomass samples. The two-stage process

consisted of a hydrogen-producing reactor (R1) with 2.3 l

working volume and a methane-producing reactor (R2) with

14.7 l working volume. The single-stage process was a reactor

with 14.7 l working volume (R3). R1 was inoculated with heat-

shocked (100 �C for 2 h) anaerobic seeding sludge from a full-

scale biogas plant treating household source-separated bio-

waste and agro-industrial by-products. The same sludge,

without heat-shock, was used as inoculum for both R2 and R3.

The feeding substrate, a mixture (4:1 w/w ratio) of swine

manure and fruits and vegetables market waste (a chemical

characterization presented in Schievano et al. (2012)) was

supplied intermittently to R1 and R2 by peristaltic pumps. The

operational hydraulic retention time (HRT) was 3, 22 and 25

days in R1, R2 and R3, respectively. Temperature was main-

tained at 55 � 2 �C, pH was measured in continuous and not

actively controlled. Qualitative and quantitative biogas anal-

ysis were performed automatically in each reactor by gas

flow-meters (Schievano et al., 2012). The two-stage hydrogen-

methane and the single-stage methane AD processes were

monitored for 2 and 1 month, respectively.

2.2. DNA extraction

Reactor samples were centrifuged (10,000 � g, 30 min, 4 �C),the resulting pellet washed twice with sterile water and cen-

trifuged again in the same conditions. Variable volumes

(2e3 ml) were used for centrifugation to obtain a final pellet of

100mg. The pellets were stored at�20 �C until DNA extraction

performed using the PowerSoil DNA Isolation kit (MoBio Lab-

oratories, Inc., Milan, Italy) according to the manufacturer’s

instructions. All DNA were extracted in duplicate.

2.3. PCR-DGGE analysis

Bacterial and archaeal 16S rRNA gene were amplified by PCR

using the primer sets GC-357-F/907-R and GC-ARC787-F/

ARC1059-R, respectively (Sass et al., 2001; Hwang et al., 2008).

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 5 1985

PCR reactions and thermal programs were performed as pre-

viously described (Merlino et al., 2012). PCR products (approx.

300 ng) were loaded onto 7% (w/v) polyacrylamide gels

(0.75 mm) containing a denaturant gradient of 30e70% or

40e60% for Bacteria and Archaea, respectively (100% denatur-

ant contained 7 M urea and 40% formamide). Electrophoresis

was run in 1 � TAE buffer using a D-Code electrophoresis

system (BioRad) at 90 V and 60 �C for 17 h. Gels were stained

with SYBR(R) Green I Nucleic A (Invitrogen) and documented

with the GelDoc 2000 apparatus (BioRad) by using the Di-

versity Database software (BioRad). Relevant DNA bands were

excised from the gels and eluted in 50 ml of TriseHCl 10 mM.

Five microliters of DNA was PCR reamplified and the obtained

sequences (Macrogen, Seoul, Korea) were compared against

the NCBI genomic database with the BLAST search alignment

tool. Sequence alignment and phylogenetic trees were carried

out using theMEGA software, version 5.0 (Tamura et al., 2011).

The trees were constructed using the Maximum Likelihood

algorithm and the Tamura Nei parameter correction and were

bootstrapped 2000 times.

DGGE gel profiles were analysed using the Quantity One

software (Biorad). Lane background was subtracted by the

“rolling disk” tool. Bands were detected automatically and

matched manually. DGGE-based molecular parameters,

namely dynamycs (Dy), richness (Rr) and community organi-

zation (Co), were calculated as previously described (Marzorati

et al., 2008). Briefly, Dy was calculated from the similarity

matrix (100 e similarity%); Rr was the total number of bands

multiplied by the percentage of denaturing gradient used; Co

was the percentage of Gini coefficient, a value describing the

degree of evenness within a community by measuring the

normalized area between a given Lorenz curve and the perfect

evenness line. The Co parameter informs on the functional

organization of the microbial community describing the spe-

cies abundance distribution within a microbial community in

terms of degrees of evenness (0e100). Low Co values represent

a highly even community, whereas high Co values are char-

acteristic of uneven communities. Average Co (Co 45e60)

values correspond to balanced community, characterized by

most functional stability and resilience. The Co coefficient

Table 1 e Real time PCR primer sets used in this study.

Target group Name Sequence

Bacteria Bac357-F CCTACGGGAGGCAGCAG

Bac907-R CCGTCAATTCCTTTGAGTTT

Hydrogen-producing

bacteria (HPB)

hydF1 GCCGACCTKACMATMATGGA

hydH ATRCARCCRCCSGGRCAGGCCAT

Acetogens fhs1 GTWTGGGCWAARGGYGGMGAA

FTHFS-r GTATTGDGTYTTRGCCATACA

Sulphate-reducing

bacteria (SRB)

Drs1þ-F ACSCACTGGAAGCACGGCGG

Dsr-R GTGGMRCCGTGCAKRTTGG

Archaea Arch 931-F AGGAATTGGCGGGGGAGCA

ArchM1100-R BGGGTCTCGCTCGTTRC

Methanosarcinales Msl812-F GTAAACGATRYTCGCTAGGT

Msl1159-R GGTCCCCACAGWGTACC

was calculated based on the Gini value, Dy was determined by

the moving window analysis (Marzorati et al., 2008).

2.4. Real-time PCR analysis

Quantitative PCR assays were performed using primer set

reported in Table 1. Considering that in anaerobic reactor

most Archaea are methanogens (Yu et al., 2005), an archaeal

PCR real time assay was used to estimate quantitatively

methanogens. PCR SYBR green reactions were prepared by

using the “Brilliant SYBR Green QPCR Master Mix” kit (Stra-

tagene) in 96-well plates on the I-Cycler (Biorad). The reaction

mix (25 ml) contained: 1 � Brilliant SYBR Green (2.5 mMMgCl2),

0.12 mM of each primers, and approx. 100 ng of extracted DNA.

In the case of primer set Msl812-F/Msl1159-R extra MgCl2 was

added to a final concentration of 4.0 mM. One real time assay

was carried out per extracted DNA. The thermal cycling pro-

gram consisted of 10 min at 95 �C, followed by 40 cycles of 30 s

at 95 �C, 1 min at X �C (X ¼ 58 �C for Bac357-F/Bac907-R, 49 �Cfor hyd-F1/hyd-R1, 55 �C for fhs1-F/THFS-R, 59 �C for Drs1þ-F/

Dsr-R, 64 �C for Arch 931-F/ArchM1100-R, 60 �C for Msl812-F/

Msl1159-R) and 1min at 72 �C. Finally, amelting curve analysis

was performed for verifying the specificity of PCR products.

The program was as follows: denaturation of 1 min at 95 �C,cooling of 1 min at 55 �C and then 95 �C again, at a rate of

þ0.5 �C per cycle. Cycle threshold (Ct) values were calculated

using the Biorad real-time software (version 3.0a) according to

the manufacturer’s instructions. Standard curves were gen-

erated by tenfold diluting the standard plasmids to obtain

a series of concentrations ranging from 10 to 108 copies of

plasmid DNA. The standard plasmids were constructed as

previously described (Merlino et al., 2012) by cloning frag-

ments obtained from PCR amplification of genomic DNA from

the bacterium Asaia for bacteria) or from total DNA (for

archaea) from an anaerobic batch digester (Table 1). Conver-

sion of 16S rRNA gene copy numbers to cell number was done

considering the average 16S rRNA gene copy numbers of

bacteria (4/cell) and methanogens (2.5 copies/cell) reported in

the Ribosomal RNA Database (rrnDB, Lee et al., 2009). In the

case of real-time PCR targeting functional genes, it was

Targetgene

Ampliconsize (bp)

Closest relative ofstandard fragment

(% similarity)

Reference

16S

rRNA

550-585 Asaia sp. AM404260

(100%)

Favia et al.,

2007

hydA 700 Uncultured bacterium

EU828435 (75%)

Xing et al.,

2008

GG fhs 250 Clostridium beijerinckii

CP000721 (76%)

Xu et al.,

2009

dsrA 221 Desulfobacterium

autotrophicum

CP001087 (98%)

Kondo et al.,

2004

16S

rRNA

169 Methanobrevibacter

sp. DQ402034 (98%)

Einen et al.,

2008

16S

rRNA

354 Methanosarcina mazeii

LM5 DQ987528 (98%)

Yu et al.,

2005

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 51986

assumed that copy number was equivalent to cell number

based on the premise that the majority of known bacteria in

the database have a single copy of the functional genes con-

sidered (Kondo et al., 2004; Xu et al., 2009).

3. Results

3.1. Operation performance of anaerobic bioreactors

The two-stage process was monitored over a period of two

months including the start-up period (days 0e9), a steady

Fig. 1 e Hydrogen/methane productions in R1 (A), R2 (B) and R3 (

start-up period and a 17-long days period after the steady state

state period (days 9e43) described in Schievano et al. (2012),

and a further 17-days long period where some imbalances

occurred (Fig. 1A and B), whereas the single-stage process was

monitored only along the steady state period (Fig. 1C). As

previously reported (Schievano et al., 2012), at the steady state

hydrogen production rate was of 1.5 Ndm3 H2/L d (45% [v/v]

content in biogas) in the acidogenic reactors (R1), whereas

methane production rates of 0.53 and 0.54 Ndm3 CH4/L d were

registered for the methanogenic reactors of the two- (R2) and

single-stage (R3) processes, respectively (68% and 54% [v/v]

content in biogas). In R1 methane was detected between day

40 and day 50 with percentages of 1e5% of total biogas. A

C) during the observed operational period. In R1 and R2 the

are indicated by shading.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 5 1987

serious biogas production failure occurred from days 49e51

both in R1 and R2, and was attributed to electric energy

shortage that determined a stop of the thermal control and

a drop of the reactor temperatures.

In R1 the major acidogenic byproducts were hexanoic acid,

acetate, butyrate and propionate. Total VFAs concentration

was approx. 3900 mg acetate/L and acetate accumulated at

more than 2500 mg/L. In R2 the concentration of acetate

decreased and remained at low level (<500 mg/L) indicating

a consuming activity by acetotrophic methanogens. Butyrate

was completely degraded suggesting the presence of active

syntrophic bacteria. After the steady-state, on day 59, an ini-

tial accumulation of VFA (1045 mg acetate/L), and con-

sequently a partial inhibition of the process, occurred

(1890 mg acetate/L at day 66). In R3 the total VFAs were

detected at concentrations about approx. ten times lower than

R2 (763e1056 mg acetate/L) with acetate as the major product

and both butyrate and propionate under the limit of detection.

3.2. Microbial community characterization of two-stageanaerobic process

3.2.1. Hydrogenogenic acidogenic processThe PCR-DGGE of Bacteria is showed in Fig. 2A and the phy-

logenetic positions of the identified sequences are indicated in

Fig. 3 (sequences affiliation in Table S1 of supplemental ma-

terial). Most bands were not closely related to known species.

All bands were assigned to the phylum of Firmicutes; nine fell

into the Clostridiales order and two into the Thermoanaer-

obacterales order. Band H1, strictly related to an uncultured

Fig. 2 e Bacterial (A, B, C) and archaeal (A0, B0, C0) DGGE profiles

extracted from samples obtained from R1 (A, A0), R2 (B, B0) andreactor (days), lane F indicate the DGGE profile of the feeding so

species from an hydrogen fermenter (Lee et al., 2010b), was

detected throughout the entire time-course of the process

and, according to its intensity, appeared as a predominant

microorganism. Several other bands (H2-H8), mainly detected

from day 9 to day 15, were also closely related (99e100%

similarity) to uncultured bacteria detected by the same au-

thors (Lee et al., 2010b). Bands H1, and H5-H8 were assigned to

the Ruminococcaceae family, and showed 94% similarity to

Clostridium sp. BS-1, a sludge isolate fermenting D-galactitol to

H2, acetate, butyrate and hexanoic acid (Jeon et al., 2010), and

93% similarity to Clostridium sp. strain Z6, isolated from paper

mill wastewater, and Clostridium sporosphaeroides, capable to

produce hydrogen (and acetate) from glutamate. Bands H2

and H3, assigned to unclassified Lachnospiraceae, were closely

related to an uncultured bacterium from an hydrogen digester

(Lee et al., 2010a) and, more distantly, to uncultured Clostri-

dium clones involved in the cellulosic and lignocellulosic

waste digestion (Shiratori et al., 2006). Bands H9 and H10,

showing a high intensity mostly after 36 day of operation,

grouped into the Clostridiaceae cluster I and fullymatchedwith

many uncultured bacteria associated to faeces and anaerobic

reactors. Band H7 matched (99.4%) with Clostridium cellulosi,

a thermophilic cellulolytic bacterium frequently detected in

H2-producing systems. Band H4, detected at days 11e15,

strictly matched (99.7%) with an uncultured Thermoanaer-

obacterium from a hydrogen reactor treating food waste

(Wang, 2008) and was also related (99.1%) to Thermoanaer-

obacterium thermosaccharolyticum, a thermophilic saccharolytic

microorganism involved in production of large hydrogen

amount (Ueno et al., 2001).

of the 16S rRNA gene PCR products amplified from DNA

R3 (C, C0). Lanes are labelled with the sampled time of the

urce. Dots and numbers indicate the bands sequenced.

Fig. 3 e Phylogenetic tree showing the phylogenetic relationships of bacterial 16S rRNA sequences affiliated to Firmicutes

phylum with reference sequences deposited at the GenBank database. Sequences from feeding source, R1, R2 and R3 are

indicated with the capital letters F, H, T and S, respectively. Un indicates an uncultered bacterium. Numbers at nodes

represent bootstrap values. The scale bar represents a sequence divergence of 5%.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 51988

The bacterial DGGE pattern was further characterized

using parameters independent from the DGGE run, namely,

microbial richness (Rr), dynamics of change (Dy) and com-

munity organization (Co) (Fig. 4). Rr was higher in the start-up

period (Rr-indices of 30e40), thereafter Rr decreased reaching

at the end of the sampling values (6e10) corresponding to low

range-weighted richness. The Dy values, were generally kept

high, indicating the adaption of the community during the

Fig. 4 e Microbial richness (Rr), dynamics (Dy), and community organization (Co) parameters from bacterial (A, B) and

archaeal (C, D) DGGE profiles of R1 (A), R2 (C) and R3 (B, D).

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 5 1989

process. Dy decreased after day 43 and was very low the last

day of sampling (rate of 8%). Variable values of Co were

observed during the start-up period, thereafter stabilizing to

values (on average approx. 40) representing a relatively mod-

erate organized community.

In order to account for the methane production recorded,

a PCR-DGGE analysis was carried out on Archaea (Fig. 2A0). Aphylogenetic tree of identified sequences is shown in Fig. 5

(sequences affiliation in Table S2 of supplemental material).

A dominant band (h2) affiliated (96.9%) to Methanosaeta was

found in the initial days of the process, but after day 36 was

replaced, at least in term of dominant intensity, by two bands

(bands h5 and h6) affiliated (98%) to Metanogenium sp.

Quantitative measurements of the bacterial abundance in

the acidogenic reactor are shown in Fig. 6. Bacteria were pre-

sent at high concentrations (108e109 bacteria per ml). Within

Bacteria, HPB represented only 0.05e1.4% of total Bacteria,

whereas acetogens were 0.6e6.2% of total Bacteria. Number of

HPB remained almost constant during the process (106 bac-

teria per ml), though a decrease was observed from day 36 to

day 50 with a ratio HPB to total Bacteria of about 0.05%. The

acetogens, with an abundance similar to that of HPB in the

first ten days of hydrogen production, thereafter increased of

one order magnitude higher than HPB. Sulfate-reducing bac-

teria (SRB) counted one order magnitude lower than HPB

(about 105 bacteria per ml). A slight increase was observed at

day 43 in correspondence of a declining trend of HPB and also

of an increase of total Bacteria. Archaea were at low titre after

heat shock treatment (104 bacteria/ml) and were of two-three

orders of magnitude lower than Bacteria. Within Archaea,

Methanosarcinales represented approx. 1% of total Archaea,

declining to 0.1% at day 64. In the influent, Methanosarcinales

were detected at low concentration too (7%).

3.2.2. Methanogenic processThe PCR-DGGE of Archaea is showed in Fig. 2B and the phylo-

genetic positions of the identified sequences are indicated in

Fig. 5 (for sequences affiliation see Table S2 of supplemental

material). The PCR-DGGE profiles showed always three

strongly intense bands (t1, t2, t3), closely related to each other

and to Methanosarcina mazeii (>98%). Bands t5 and t6, appear-

ing at day 60, were closely related (>99%) to the genus Meth-

anothermobacter which depends entirely on H2/CO2 as energy

and carbon sources (Schill et al., 1999). The statistical analysis

of DGGE profiles evidenced a stable, highly specialized (Rr < 4)

community. A notable change (37%) occurred only at day 60 in

correspondence to a partial accumulation of VFAs. The Co

values were around 40e50, usually reported for good perfor-

mance reactors (Carballa et al., 2011).

PCR-DGGE analysis carried out on Bacteria evidenced in R2,

like in R1, a bacterial community dominated by Firmicutes

(Fig. 2B0 and Table S2). In R2was found C. cellulosi (band T1) and

other microorganisms (bands T3, T7, T8) already identified in

R1 and assigned to unclassified Ruminococcaceae. The other

identified sequences could not be attributed to known species,

but were highly similar (>99%) to sequences from thermo-

philic reactors (Tang et al., 2011; Sasaki et al., 2011; Goberna

et al., 2009; Shiratori et al., 2006); in particular, bands T10

and T11 matched (99e100%) with the unknown DAD cluster 3

(Tang et al., 2011). Band T2 was affiliated to the Thermotogae

Fig. 5 e Phylogenetic tree showing the phylogenetic relationships of archaeal 16S rRNA sequences with reference sequences

deposited at the GenBank database. Sequences from feeding source, R1, R2 and R3 are indicated with the lower-case letters

f, h, t and s, respectively. Un indicates an uncultered bacterium. Numbers at nodes represent bootstrap values. The scale bar

represents a sequence divergence of 10%.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 51990

phylum, consisting of anaerobic hyperthermophilic bacteria

capable of using a great variety of carbohydrates, and gen-

erating hydrogen (Eriksen et al., 2010).

In R2 the abundance of the different populations remained

almost constant during theprocess (Fig. 6).Archaea and Bacteria

were present at rather similar concentrations. Methanogen

numberwas higher than in R1, around 107e108 bacteria perml,

a range typical of anaerobic reactors (Yu et al., 2005; Lee et al.,

2008). Methanosarcinales were the dominant methanogens rep-

resenting approx. 50% of the total Archaea, in agreement with

the PCR-DGGE data. Archaea and Methanosarcinales showed the

same trend during the operation. The number of Meth-

anosarcinales decreased in correspondence to the partial inhi-

bition of the process, in accordance with PCR-DGGE data.

SRB, potential competitors of methanogens, were four

order magnitude lower than Archaea (103 bacteria per ml),

while acetogens and the HPB remained relatively stable dur-

ing all the operation (average values of 2 � 107 and 5 � 106

bacteria per ml, respectively).

3.3. Microbial community characterization of single-stage anaerobic process

PCR-DGGE bands of Archaea (Fig. 2C0) belonged mainly to the

Methanosarcinales (Fig. 5 and Table S2). Bands s1, s2 and s3,

identical to those detected in R2, were associated to the same

sludge used for the start-up of the two processes. Bands s4 and

s6 were related (>97.7%) to Methanosarcina spp. Bands s9 and

s10 both matched, with 97.3% and 99.5% similarity respec-

tively, with Methanosarcina mazeii and Methanosarcina lacustris.

Band s5, which showed a strong intensity after 15 day, was

affiliated to Methanosaeta concilii. The faint bands s7 and s8

were affiliated (>98.0%) to Methanothermobacter. The archaeal

community structure, on the contrary of R2, changed over

time. After two weeks, the community was drastically shifted

(rate of 88%) and thereafter stabilized. The community was

richer than in R2 (Rr average value of 6.8). Co values were on

slightly higher than those of R2, indicating a moderately

organized community.

PCR-DGGE analysis of Bacteria indicated a dominance of

Firmicutes, Clostridia and Bacilli classes (Fig. 2C). The bands of

Bacilli (S1eS3) were replaced after day 15 by high intensity

bands related to Bacteroidetes (S4eS7). Bands S1eS3 were

highly similar (99.1e100%) to Bacillus infernus (96.8e97.8%), an

anaerobic species able to ferment glucose and utilize formate

and lactate for growth (Boone et al., 1995). Bands S4eS7 were

assigned to unclassified Porphyromonadaceae, bacteria capable

of producing VFA from carbohydrates or proteins (Ziganshin

et al., 2011). Considering the strong intensity of these bands,

probably these bacteria played an important role in

Fig. 6 e Concentrations of microorganisms in R1 (A), R2 (B) and R3 (C) during the observed period. Values are averages of two

measurements. HPB and SRB were not detected at day 0 (after heat-shock treatment).

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 5 1991

hydrolysis and acidogenesis. A microorganism stably detec-

ted throughout the process was Anaerobaculum (band S8), able

to ferment mainly peptides and organic acids to acetate,

hydrogen and CO2 (Menes and Muxı, 2002). Bands S9 and S10

belonged to the Clostridiaceae cluster I. Bands S11 and S12

clustered with unknown clones, from thermophilic reactors,

grouping in the cluster DAD 1 (Tang et al., 2011) and DAD3,

respectively. Band S13 and S14 were correlated to Thermace-

togenium (Hattori et al., 2000) and Tepidanaerobacter (Sekiguchi

et al., 2006), thermophilic syntrophic acetate-oxidizing bac-

teria capable to form methane in association with hydro-

genotrophic methanogens. Statistical analysis of bacterial

PCR-DGGE profiles indicated a very high diversity (average

Rr value of 65), higher than in R1. Dy had a very similar trend

to that observed for Archaea, with a notable community shift

at day 15 (rate 66%). Constant Co values of approx. 40 indi-

cated a moderately even community that remained stable

during the operations.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 51992

Quantitativemeasurements of Bacteria, acetogens, SRB and

HPB in reactor R3 showed values similar to those in R2 (Fig. 6).

On the contrary, both totalArchaea andMethanosarcinaleswere

estimated in lower numbers than in R2 (below 107 bacteria per

ml). The number of Methanosarcinales decreased with the

changing of the community structure, dropping to 32% of total

Archaea at day 15 and increasing again to 58% at day 29.

4. Discussion

In this study the dynamics of microbial community structure

in a two- and a single-stage AD reactors have been inves-

tigated and compared. As models were used I-CTSR reactors

operating with some equal working conditions (inocula from

the same methanogenic sludge, feeding source, temperature,

and non-controlled pH) and some operational parameters

(HRT, loading rate, heat-shock treatment of acidogenic reac-

tor) specifically designed for the two AD processes.

Overall, the study showed that the microbial community

structure and dynamic was different in twoAD processes both

for Bacteria and Archaea. Resistant spore-formers Firmicutes

selected by the heat-shock treatment dominated in R1 and R2,

whereas a more diverse community (Firmicutes, Bacteroidetes,

Synergistes) was found in R3. In R1 and R2, most of the bacteria

were related to the order Clostridiales, more specifically to the

Clostridium genus. Clostridium spp. are capable of ferment

cellulose and various carbohydrates mainly to acetate, buty-

rate and hydrogen (Valdez-Vazquez and Poggi-Varaldo, 2009)

and their prevalence in stable H2-producing systems has been

already documented (Jo et al., 2007). Hence, they can account

for the hydrogen production recorded in the acidogenic reac-

tor. The majority of identified species, however, were not

referable to known cultured species, with the exception of C.

cellulosi. Nevertheless, many of the identified microorganisms

were phylogenetically related to microorganisms from ther-

mophilic acidogenic anaerobic reactors fed with vegetable

kitchen waste and, more distantly, to Clostridium species

(Clostridium sp. BS-1, Clostridium sp. Z6) included into the

Ruminococcaceae cluster. The presence in the digester of spe-

cies with degrading ability similar to that of Clostridium sp. BS-

1 may account for the detection of high hexanoic acid in the

reactor. Bioavailable D-galactitol, a reduced form of D-gal-

actose, is in fact contained in many fruit and vegetable resi-

dues like those used as feeding source. On the other hand, the

dominance of uncultured bacteria affiliated to Clostridium sp.

Z6 has been previously reported in other hydrogen-producing

reactors (Chu et al., 2010; Lee et al., 2010a) which operated at

55 �C treating food waste without heat treatment of inoculum.

Hence, it is likely that in this study the selection and the

dominance of these clostridia has been favoured by a combi-

nation of various operational parameters (temperaure, feed-

ing source, reactor type, TS, pH and HRT) rather than by the

inoculum pre-treatment, in agreement with the findings of

Luo et al. (2011).

In R2, as well in R1, not only the species diversity was rel-

atively low, but a relatively dynamic bacterial community

simplified over time. A specialized community, however,

though highly functional, is more sensitive to changes since it

lacks alternative players when impaired by stresses. This may

explain the partially inefficient biodegradation observed in R2

as deduced by the chemical characterization of reactors ma-

terials. The high concentrations of VFAs, alcohols and other

intermediatemetabolites (amines, amino acids, phenols) in R2

(Schievano et al., 2012) probably exerted inhibiting effects on

many microorganisms, including methanogens, decreasing

the community diversity and its potential of adaptation. On

the contrary, in R3 was maintained a more diverse and dy-

namic community that probably has a richest network of

metabolic pathways explaining the most efficient degrading

activity observed in this digester. In particular, in R3 were

found microorganisms related to the Porphyromonadaceae

family and Anaerobaculum genus. These bacteria, capable of

fermenting peptides and amino acids, were possibly respon-

sible for the low levels of nitrogenous compounds detected in

R3 and found instead undegraded in R2. Bacteroidetes are more

efficient than Firmicutes in degrading plant polyphenols and

less sensitive to phenols (Rastmanesh, 2011 and references

therein). It is speculated that a phenol/polyphenols rich

feeding promoted the Bacteroidetes growth in R3, but left these

compounds not degraded in R2.

The archaeal community in two methanogenic reactors

was dominated, though at different levels, by the Meth-

anosarcinales (average value of 70% and 58% of total Archaea in

R2 and R3, respectively), suggesting that acetoclastic meth-

anogenesis was the major pathway of methane production in

both systems. In R2, Methanosarcinales were up to 90% of

methanogens andwere represented at the steady state only by

the genus Methanosarcina. Methanosarcina spp., prevailing at

high acetate concentration (Jetten et al., 1992), were previously

detected as dominant in other thermophilic methanogenic

reactors from two-stage processes (Chu et al., 2010; Luo et al.,

2011) and in general from digesters treating manure (Demirel

and Scherer, 2008). Hence, their abundance in R2 is sustained

by the high levels of acetate (after hexanoic acid) detected in

gas and liquidphasesof R1 (Schievanoet al., 2012). Particularly,

the identified Methanosarcina-like species were related to M.

mazei andM. siciliaewhich are able to utilize various substrates

(methanol, methylamines and also H2/CO2 in the case of M.

siciliae) other than acetate (Liu et al., 2009; Lee et al., 2010a).

In R3, though the high level of Methanosarcinales also pre-

sent, the archaeal community was more diverse and dynamic

as compared to R2. Methanosarcina, Methanosaeta and Meth-

anothermobacter specieswere simultaneously detected, though

at different density over the course of the operation. The

contribution of hydrogenotrophic methanogenesis in R3 was

highest than in R2 as confirmed by the detection in the latter

of thermophilic acetate-oxidizing bacteria (Thermacetogenium

and Tepidanaerobacter) capable of form methane in coopera-

tion with hydrogenotrophic methanogens (Hattori, 2008).

Such an archaeal community may have been advantageous to

the process performance, promoting an improved adaptation

potential.

Data of quantitative determinations of AD functional

groups indicated that their abundance remained rather con-

stant at the steady state despite some variations in biogas

production during the period.

In the two methanogenic reactors the abundance of dif-

ferent microbial groups were at the same order of magnitude,

except for higher methanogens in R2, probably enriched by

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 5 1993

the separation of acidogenic and methanogenic phases. In R1

the high abundance of acetogens suggested that theywere the

major competitors of HPB. Acetogens are capable to catalyse

the reductive synthesis of acetate from CO2 switching be-

tween heterotrophic and autotrophic metabolism depending

on substrate availability (Drake et al., 2002). Their contribution

to hydrogen consumption in bioreactors depends on several

chemical and physical factors (acetate concentrations,

hydrogen partial pressure, mass transfer phenomena be-

tween H2-producers and H2-consumers) and the history of the

inocula. The abundance of acetogens in R1 reactor ranged

from 1.4 � 108 to 9.3 � 108 FTHFS (formyltetrahydrofolate

synthetase) genes per gram dry weight, in accordance with Xu

et al. (2009) who reported 108-109 FTHFS gene copies per gram

dry weight in a sludge under H2/CO2 enrichment conditions.

Thus, it is likely that acetogens, even at low percentages

(Kraemer and Bagley, 2008), may have somehow contributed

to hydrogen consumption in R1, suggesting that heat-

treatment is not sufficient to control spore-forming H2-con-

sumers. In addition, it was observed that, after some days of

operation, acetogens prevailed over methanogens. This is in

agreement with previous findings indicating that generally

methanogenesis prevails over acetogenesis due to its more

favourable thermodynamics and affinity for H2 (Liu and

Whitman, 2008). However, acetogenesis can effectively out-

competemethanogenesis in certain conditions, like under low

pH and accumulation of H2 (Drake et al., 2002). The methane

detected in the acidogenic reactor, presumably generated by

Metanogenium species, was in biogas in low percentage and for

a limited operational time, confirming the efficacy of low pH

condition to inhibit methanogenesis.

Quantitative data allowed also to explain some failures in

biogas production of the two-stage process occurred after the

steady state, like the hydrogen production drop occurred from

days 48e50 accompanied by an almost one order magnitude

decrease of HPB.

Partial accumulation of VFAs, particularly acetate and

propionate (370 and 325 mg/L, respectively, at day 59) may

explain the one order magnitude decrease of methanogens

and the acetotrophic methanogen proportion (50%). This VFA

accumulation may also account for the appearance in the

methanogen population ofMethanothermobacter, less sensitive

than acetoclastic methanogens to increases in VFAs concen-

tration (Hori et al., 2006). All these microbiological evidences

support the non-optimal condition in general occurring in R2.

Overall, the higher diversity and dynamic of prokaryote

community, especially the fermentative bacterial one, in the

single stage process as compared to the two-stage process,

may account for the best degradation efficiency observed in

R3. The difference in bacterial community and performance

between the two AD processes is likely a consequence of

decoupling of acidogenesis from methanogenesis in the two

stage stystem and of the different configurations and opera-

tional parameters of the two systems. In the two-stage pro-

cess, the separation of fermentative and methanogenic

environments might have affected negatively syntrophic as-

sociations among microorganisms and probably reduced the

number of degradation pathways. Particularly, the enriched

simplified community established in R2 proved to be unable to

completely degrade many intermediate metabolites causing

inhibition effects determining a methane production lower

than the expected one. Further research is needed to evaluate

the possibility of optimization of two-stage process ensuring

a stable and efficient microbial community.

5. Conclusions

The study showed that qualitative and quantitative data on

microbial community provided valuable information on the

functionality of the AD processes. In particular, the data

allowed to: i) correlate the structure of microbial communities

to the functionality of the processes accounting for the partial

inefficiency of R2 and highest performance of R3; ii) inves-

tigate the relations (and their fluctuations) among the differ-

ent trophicmicrobial groups, and with the functionality of the

AD processes iii) evidence the key role of operational param-

eters and of reactors configuration in driving the dominant

microbial communities. Microbiological investigation pro-

vided a useful tool to get a better insight into the factors

determining the performance of the two processes.

Acknowledgements

This study was funded by the project “Produzione di bio-

idrogeno ed energia rinnovabile da residui agro-zootecnici e

AgrIdEn” by Regione Lombardia. Partial support comes from

the project “Miniaturizzazione e semplificazione di linee di

trasformazione per piccole produzioni agroalimentari e

impiego di energie rinnovabili e MIERI” by Ministero per le

Politiche Agricole Alimentari e Forestali.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.watres.2013.01.007.

r e f e r e n c e s

Angenent, L.T., Karim, K., Al-Dahhan, M.H., Wrenn, B.A.,Domıguez-Espinosa, R., 2004. Production of bioenergy andbiochemicals from industrial and agricultural wastewater.Trends in Biotechnology 22, 477e485.

Boone, D.R., Liu, Y., Zhao, Z.J., Balkwill, D.L., Drake, G.R.,Stevens, T.O., Aldrich, H.C., 1995. Bacillus infernus sp. nov., anFe (III)- and Mn (1V)-reducing anaerobe from the deepterrestrial subsurface. International Journal of Systematic andEvolutionary Bacteriology 45, 441e448.

Carballa, M., Smits, M., Etchebehere, C., Boon, N., Verstraete, W.,2011. Correlations between molecular and operationalparameters in continuous lab-scale anaerobic reactors.Applied Microbiology and Biotechnology 89, 303e314.

Chu, C.-F., Ebie, Y., Xu, K.-Q., Li, Y.-Y., Inamori, Y., 2010.Characterization of microbial community in the two-stageprocess for hydrogen and methane production from foodwaste. International Journal of Hydrogen Energy 35,8253e8261.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 51994

Demirel, B., Scherer, P., 2008. The roles of acetotrophic andhydrogenotrophic methanogens during anaerobic conversionof biomass to methane: a review. Reviews in EnvironmentalScience and Biotechnology 7, 173e190.

Demirel, B., Scherer, P., Yenigun, O., Onay, T.T., 2010. Productionof methane and hydrogen from biomass through conventionaland high-rate anaerobic digestion processes. Critical Reviewsin Environmental Science and Technology 40, 116e146.

Drake, H.L., Kusel, K., Matthies, C., 2002. Ecological consequencesof the phylogenetic and physiological diversities of acetogens.Antonie van Leeuwenhoek 81, 203e213.

Einen, J., Thorseth, I.H., Øvreas, L., 2008. Enumeration ofArchaea andBacteria in seafloor basalt using real-time quantitative PCR andfluorescencemicroscopy.FEMSMicrobiologyLetters282,182e187.

Eriksen, N.T., Riis, M.L., Holm, N.K., Iversen, N., 2010. H2 synthesisfrom pentoses and biomass in Thermotoga spp. BiotechnologyLetters 33, 293e300.

Favia, G., Ricci, I., Damiani, C., Raddadi, N., Crotti, E.,Marzorati, M., Rizzi, A., Urso, R., Brusetti, L., Borin, S., Mora, D.,Scuppa, P., Pasqualini, L., Clementi, E., Genchi, M., Corona, S.,Negri, I., Grandi, G., Alma, A., Kramer, L., Esposito, F., Bandi, C.,Sacchi, L., Daffonchio, D., 2007. Bacteria of the genus Asaiastably associate with Anopheles stephensi, an Asian malarialmosquito vector. PNAS 104, 9047e9051.

Goberna, M., Insam, H., Franke-Whittle, I.H., 2009. Effect ofbiowaste sludge maturation on the diversity of thermophilicbacteria and archaea in an anaerobic reactor. AppliedEnvironmental Microbiology 75, 2566e2572.

Guo, X.M., Trably, E., Latrille, E., Carrere, H., Steyer, J.-P., 2010.Hydrogen production from agricultural waste by darkfermentation: a review. International Journal of HydrogenEnergy 35, 10660e10673.

Hattori, S., 2008. Syntrophic acetate-oxidizing microbes inmethanogenic environments. Microbes and Environments 23,118e127.

Hattori, S., Kamagata, Y., Hanada, S., Shoun, H., 2000.Thermacetogenium phaeum gen. nov., sp. nov., a strictlyanaerobic, thermophilic, syntrophic acetate-oxidizingbacterium. International Journal of Systematic andEvolutionary Microbiology 50, 1601e1609.

Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y., 2006. Dynamictransition of a methanogenic population in response to theconcentration of volatile fatty acids in a thermophilicanaerobic digester. Applied and Environmental Microbiology72, 1623e1630.

Hwang, K., Shin, S.G., Kim, J., Hwang, S., 2008. Methanogenicprofiles by denaturing gradient gel electrophoresis usingorder-specific primers in anaerobic sludge digestion. Appliedand Environmental Microbiology 80, 269e276.

Jeon, B.S., Kim, B.C., Um, Y., Sang, B.I., 2010. Production ofhexanoic acid from D-galactitol by a newly isolated Clostridiumsp. BS-1. Applied Microbiology Biotechnology 88, 1161e1167.

Jetten, M.S.M., Stams, A.J.M., Zehnder, A.J.B., 1992.Methanogenesis from acetate: a comparison of acetatemethabolism in Methanothrix soehngenii and Methanosarcinaspp. FEMS Microbiology Letters 88, 181e198.

Jo, J.H., Jeon, C.O., Lee, D.S., Park, J.M., 2007. Process stability andmicrobial community structure in anaerobic hydrogen-producing microflora from food waste containing kimchi.Journal of Biotechnology 131, 300e308.

Kondo, R., Nedwell, D.B., Purdy, K.J., de Queiroz Silva, S., 2004.Detection and enumeration of sulphate-reducing bacteria inestuarine sediments. Geomicrobiology Journal 21, 145e157.

Kraemer, J.T., Bagley, D.M., 2008. Measurement of H2 consumptionand its role in continuous fermentative hydrogen production.Water Science and Technology 57, 681e685.

Lee, C., Kim, J., Shin, S.G., Hwang, S., 2008. Monitoring bacterialand archaeal community shifts in a mesophilic anaerobic

batch reactor treating a high-strength organic wastewater.FEMS Microbiology Ecology 65, 544e554.

Lee, C., Kim, J., Shin, S.G., O’Flaherty, V., Hwang, S., 2010a.Quantitative and qualitative transitions of methanogencommunity structure during the batch anaerobic digestion ofcheese-processing wastewater. Applied Microbiology andBiotechnology 87, 1963e1973.

Lee, Z.M.-P., Bussema III, C., Schmidt, T.M., 2009. rrnDB:documenting the number of rRNA and tRNA genes in bacteriaand archaea. Nucleic Acids Research 37, D489eD493. http://rrndb.mmg.msu.edu/search.php.

Lee, Z.-K., Li, S.-L., Kuo, P.-C., Chen, I.-C., Tien, Y.-M., Huang, Y.-J.,Chuang, C.-P., Wong, S.-C., Cheng, S.-S., 2010b. Thermophilicbio-energy process study on hydrogen fermentation withvegetable kitchen waste. International Journal of HydrogenEnergy 35, 13458e13466.

Liu, F.H., Wang, S.B., Zhang, J.S., Zhang, J., Yan, X., Zhou, H.K.,Zhao, G.P., Zhou, Z.H., 2009. The structure of the bacterial andarchaeal community in a biogas digester as revealed bydenaturinggradient gel electrophoresis and16S rDNAsequencinganalysis. Journal of Applied Microbiology 106, 952e966.

Liu, Y., Whitman, W.B., 2008. Metabolic, phylogenetic, andecological diversity of the methanogenic archaea. Annals ofthe New York Academy of Sciences 1125, 171e189.

Luo, G., Karakashev, D., Xie, L., Zhou, Q., Angelidaki, I., 2011. Long-term of inoculums pretreatment on fermentative hydrogenproduction by repeated batch cultivations: homoacetogenesisand methanogenesis as competitors to hydrogen production.Biotechnology and Bioengineering 108, 1816e1827.

Luo, G., Xie, L., Zhou, Q., Angelidaki, I., 2011b. Enhancement ofbioenergy production from organic wastes by two-stageanaerobic hydrogen and methane production process.Bioresource Technology 102, 8700e8706.

Marzorati, M., Wittebolle, L., Boon, N., Daffonchio, D.,Verstraete, W., 2008. How to get more out of molecularfingerprints: practical tools for microbial ecology.Environmental Microbiology 10, 1571e1581.

Menes, R.J., Muxı, L., 2002. Anaerobaculum mobile sp. nov., a novelanaerobic, moderately thermophilic, peptide-fermentingbacterium that uses crotonate as an electron acceptor, andemended description of the genus Anaerobaculum.International Journal of Systematic and EvolutionaryMicrobiology 52, 157e164.

Merlino, G., Rizzi, A., Villa, F., Sorlini, C., Brambilla, M.,Navarotto, P., Bertazzoni, B., Zagni, M., Araldi, F.,Daffonchio, D., 2012. Shifts of microbial community structureduring anaerobic digestion of agro-industrial energetic cropsand food industry byproducts. Journal of ChemicalTechnology and Biotechnology 9, 1302e1311.

Rastmamesh, R., 2011. High Polyphenol, Low Prebiotico Diet forWeight Loss Because of Intestinal Microbiota Interaction, 189,pp. 1e8.

Sasaki, D., Hori, T., Haruta, S., Ueno, Y., Ishii, M., Igarashi, Y.,2011. Methanogenic pathway and community structure ina thermophilic anaerobic digestion process of organic solidwaste. Journal of Bioscience and Bioengineering 111, 41e46.

Sass, A.M., Sass, H., Coolen, M.J., Cypionka, H., Overmann, J., 2001.Microbial communities in the chemocline of a hypersalinedeep-sea basin (Urania basin, Mediterranean Sea). Appliedand Environmental Microbiology 67, 5392e5402.

Schievano, A., Tenca, A., Scaglia, B., Merlino, G., Rizzi, A.,Daffonchio, D., Oberti, R., Adani, F., 2012. Two-stage vs. single-stage anaerobic digestion: comparison of energy productionand biodegradation efficiencies. Environmental Science andTechnology 46, 8502e8510.

Schill, N.A., Liu, J.-S., von Stockar, U., 1999. Thermodynamicanalysis of growth of Ethanobacterium thermoautotrophicum.Biotechnology and Bioengineering 64, 74e81.

wat e r r e s e a r c h 4 7 ( 2 0 1 3 ) 1 9 8 3e1 9 9 5 1995

Sekiguchi, Y., Imachi, H., Susilorukmi, A., Muramatsu, M.,Ohashi, A., Harada, H., Hanada, S., Kamagata, Y., 2006.Tepidanaerobacter syntrophicus gen. nov., sp. nov., an anaerobic,moderately thermophilic, syntrophic alcohol- and lactate-degrading bacterium isolated from thermophilic digestedsludges. International Journal of Systematic and EvolutionaryMicrobiology 56, 1621e1629.

Shiratori, H., Ikeno, H., Ayame, S., Kataoka, N., Miya, A.,Hosono, K., Beppu, T., Ueda, K., 2006. Isolation andcharacterization of a new Clostridium sp. that performseffective cellulosic waste digestion in a thermophilicmethanogenic bioreactor. Applied and EnvironmentalMicrobiology 72, 3702e3709.

Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M.,Kumar, S., 2011. MEGA5: molecular evolutionary geneticsanalysis using maximum likelihood, evolutionary distance,and maximum parsimony methods. Molecular Biology andEvolution 28, 2731e2739.

Tang, Y.-Q., Ji, P., Hayashi, J., Koike, Y., Wu, X.-L., Kida, K., 2011.Characteristic microbial community of a dry thermophilicmethanogenic digester: its long-term stability and changewith feeding. Applied Microbiology and Biotechnology 91,1447e1461.

Ueno, Y., Haruta, S., Ishii, M., Igarashi, Y., 2001. Characterizationof a microorganism isolated from the effluent of hydrogen

fermentation by microflora. Journal of Bioscience andBioengineering 92, 397e400.

Valdez-Vazquez, I., Poggi-Varaldo, H.M., 2009. Hydrogenproduction by fermentative consortia. Renewable andSustainable Energy Reviews 13, 1000e1013.

Wang, Y.H., 2008. Process study and hydrolysis mechanism studyof thermophilic hydrogen production with starch-richedkitchen waste fermentation. Phd thesis.

Xing, D., Ren, N., Rittmann, B., 2008. Genetic diversity ofhydrogen-producing bacteria in an acidophilic ethanol-H2-coproducing system, analyzed using the [Fe]-hydrogenasegene. Applied and Environmental Microbiology 74, 1232e1239.

Xu, K., Liu, H., Du, G., Chen, J., 2009. Real-time PCR assaystargeting formyltetrahydrofolate synthetase gene toenumerate acetogens in natural and engineeredenvironments. Anaerobe 15, 204e213.

Yu, Y., Lee, C., Kim, J., Hwang, S., 2005. Group-specific primer andprobe sets to detect methanogenic communities usingquantitative real-time Polymerase Chain Reaction.Biotechnology and Bioengineering 89, 670e679.

Ziganshin, A.M., Schmidt, T., Scholwin, F., Il’inskaya, O.N.,Harms, H., Kleinsteuber, S., 2011. Bacteria and archaeainvolved in anaerobic digestion of distillers grains withsolubles. Applied Microbiology and Biotechnology 89,2039e2052.