wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 7

Avai lab le at www.sc iencedi rect .com

journa l homepage : www.e lsev ier . com/ loca te /wat res

Methanogenic community shift in anaerobic batch digesterstreating swine wastewater

Woong Kim a, Seungyong Lee a, Seung Gu Shin a, Changsoo Lee b, Kwanghyun Hwang a,Seokhwan Hwang a,*aSchool of Environmental Science and Engineering, Pohang University of Science and Technology, San 31, Hyoja-dong, Nam-gu,

Pohang, Kyungbuk 790-784, Republic of KoreabDivision of Environmental and Water Resources Engineering, School of Civil and Environmental Engineering,

Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

a r t i c l e i n f o

Article history:

Received 11 April 2010

Received in revised form

25 June 2010

Accepted 9 July 2010

Available online 16 July 2010

Keywords:

Microbial community change

Swine wastewater

Redundancy analysis (RDA)

Methanobacteriales (MBT)

Methanomicrobiales (MMB)

Methanosarcinales (MSL)

methanogenesis

Anaerobic digestion

* Corresponding author. Tel.: þ82 54 279 228E-mail address: shwang@postech.ac.kr (S

0043-1354/$ e see front matter ª 2010 Elsevdoi:10.1016/j.watres.2010.07.029

a b s t r a c t

Qualitative and quantitative molecular analysis techniques were used to determine asso-

ciations between differences in methanogenic microbial communities and the efficiency of

batch anaerobic digesters. Two bioreactors were initially seeded with anaerobic sludge

originating from a local municipal wastewater treatment plant and then supplemented

with swine wastewater. Differences were observed in the total amount of methane

produced in the two bioreactors (7.9 L/L, and 4.5 L/L, respectively). To explain these

differences, efforts were taken to characterize the microbial populations present using

a PCR-based DGGE analysis with methanogenic primer and probe sets. The groups Meth-

anomicrobiales (MMB), Methanobacteriales (MBT), and Methanosarcinales (MSL) were detected,

but Methanococcales (MCC) was not detected. Following this qualitative assay, real-time PCR

was used to investigate quantitative differences in the populations of these methanogenic

orders. MMB was found to be the dominant order present and its abundance patterns were

different in the two digesters. The population profiles of the other methanogenic groups

also differed. Through redundancy analysis, correlations between the concentrations of

the different microbes and chemical properties such as volatile fatty acids were calculated.

Correlations between MBT and MSL populations and chemical properties were found to be

consistent in both digesters, however, differences were observed in the correlations

between MMB and propionate. These results suggest that interactions between pop-

ulations of MMB and other methanogens affected the final methane yield, despite MMB

remaining the dominant group overall. The exact details of why changes in the MMB

community caused different profiles of methane production could not be ascertained.

However, this research provides evidence that microbial behavior is important for regu-

lating the performance of anaerobic processes.

ª 2010 Elsevier Ltd. All rights reserved.

2; fax: þ82 54 279 8299.. Hwang).ier Ltd. All rights reserved.

wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 7 4901

1. Introduction correspondence analysis ordination andmultiple regressions.

1.1. Anaerobic digestion

Anaerobic digestion is now being used widely due to dramatic

increase in the production of swine waste during the past

decade commensurate with increase in meat consumption

(Speece, 1996; Deng et al., 2007; Lee et al., 2008). Anaerobic

digestion can bedivided into acidogenesis andmethanogenesis

(Speece, 1996). These broadsteps involve a series of interactions

between twomicrobial groups: acidogens andmethanogens.

Acidogens break down organic compounds using hydrolysis

and fermentation. Hydrolysis involves the decomposition of

complex substrates such as proteins, carbohydrates, and lipids

into smaller units (i.e. amino acids, simple sugars, short chain

lipids). Fermentation follows hydrolysis and consists of the

transformation of these small organic compounds into

hydrogen or organic acids such as volatile fatty acids (VFAs).

The products of fermentation by acidogens can then be

metabolized by methanogenic bacteria to yield methane. Thus

in order for waste to be fully broken down in an anaerobic

digester, acidogens and methanogens must act in concert

(Fernandez et al., 2000). An understanding of microbial

community structure and dynamics is therefore a fundamental

requisite for developing anaerobic digestion technology.

Although field scale anaerobic digesters have been established

already using the limited information known about the micro-

bial ecosystems involved in the process (Fernandez et al., 2000;

Hori et al., 2006), further insights into microbial community

structure could allowmore efficient digesters to be developed.

1.2. Linking molecular characterization technologies tobioreactor performance

The development of culture-independent molecular charac-

terization technologies, such as those based on 16S rRNA gene

sequences, has greatly assisted the study of microbial commu-

nities. These techniques allow the identification of all microor-

ganisms involved in a bioprocess rather than the limited set of

organisms, for which a culture method has been established.

Previously, analysis of anaerobic digestion has primarily relied

upon qualitative or semi-quantitative methods such as clone

libraries, molecular fingerprinting and nucleic acid hybridiza-

tion. While these methods represent an advancement over

culture based characterization, they can only give a limited

insight into the quantitative population dynamics that are

important for theperformanceofanaerobicdigesters (Zumstein

et al., 2000). For this reason, quantitative characterization

methods such as real-time PCR offer great potential for better

understanding the processes involved. By using quantitative

characterization methods together with multivariate analysis,

the relative abundances of each sub-group of microorganisms

can be associated with their functional contribution to anaer-

obic digestion (Akarsubasi et al., 2005).

1.3. Redundancy analysis

Redundancy analysis (RDA) is a multivariate analysis method

based on an iterative process of reciprocal averaging/

It generates a direct gradient ordination that is related to two

sets of variables: the dependent species data and the inde-

pendent environmental data (Jan and Peter, 2003).

RDAhasbeenwidelyusedtodeterminecorrelationsbetween

environmental variablesandmicrobial communitydynamics in

various systems, such as in soils (Bossio and Scow, 1995; Bremer

et al., 2007), dunes (Comoret al., 2008), grassland (Kennedyet al.,

2004), groundwater (Imfeld et al., 2008) and biominerallization

system of groundwater (Michalsen et al., 2007; Imfeld et al.,

2008). Recently, it was reported that multivariate statistical

techniques have also been effectively applied in wastewater

treatment systems to study the environmental effects on

microbial community structure (Gilbride et al., 2006; Hori et al.,

2006; Bernall et al., 2008; Elissen et al., 2008) and the dynamics

of microbial population (Hori et al., 2006; Roy et al., 2009).

1.4. Aims

This study aimed to investigate microbial community

dynamics using RDA correlation analysis to develop an

understanding of how microbial community composition can

impact on chemical properties such as methanation in

anaerobic digesters.

2. Materials and methods

2.1. Reactor operating conditions

Two identical anaerobic continuously stirred tank reactors

(CSTRs), each with a working volume of 6 L and equippedwith

temperature controllers, were operated in batch mode for

methanogenesis. Anaerobic seed sludge from a local munic-

ipal wastewater treatment plant in Pohang, South Korea, was

mixedwith swinewastewater. Themixing ratio of seed sludge

was 1% of the total suspended solids (w/v). The batch-mode

methanogenic digesters were operated for 79 days to culture

a mixed population of methanogens. The operating temper-

ature was kept at 35� 0.2 �C by the temperature controller.

Preliminary experiments showed that pH was almost static

during the process, so pH was not actively controlled in the

batch-mode digesters.

2.2. DNA extraction

For DNA extraction, sampling was carried out from the

anaerobic digesters. One milliliter of the raw sample was

diluted by adding 9 ml of deionized and distilled water (DDW).

A sample of 200 ml was centrifuged and then diluted 1:2 in

DDW. Following the centrifugation, 100 ml of supernatant was

decanted. After addition of DDW, centrifugation and removal

of supernatant were repeated again, the pellet was re-sus-

pended in 100 ml of DDW. Total DNA in the suspension was

immediately extracted using an automated nucleic acid

extractor (Magtration System 6GC, PSS, Chiba, Japan). Purified

DNA was eluted with 100 ml of TriseHCl buffer (pH 8.0) and

stored at �20 �C for further analyses. DNA extraction was

performed in duplicate.

wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 74902

2.3. DGGE analysis

In order to obtain the DGGE profile, 16S rRNA gene primers

(Table 1) for Archaea or for specific orders of methanogens

were exploited to amplify the extracted DNA. To amplify the

whole archaeal 16S rDNA gene within the extracted DNA,

primers ARC787F (50-ATTAG ATACC CSBGT AGTCC-30), and

ARC1059R (50-GCCATGCACCWCCTCT-30) were used (Yu et al.,

2005). In addition, to amplify each methanogenic 16S rDNA

gene at order level, specific primers for the orders Meth-

anomicrobiales (MMB), Methanosarcinales (MSL), and Meth-

anobacteriales (MBT) were used. To obtain the amplified target

DNA, a touchdown PCR method was used with the following

conditions: initial denaturation at 94 �C for 10 min, followed

by 20 cycles of denaturation at 94 �C for 1 min; annealing at

a temperature that decreased by 0.5 �C every cycle from 65 �Cto the ‘touchdown’ at 55 �C, remaining at each temperature for

1 min; and chain extension at 72 �C for 1 min. This was fol-

lowed by 20 cycles of denaturation at 94 �C for 1 min,

annealing at 55 �C for 1 min, and extension at 72 �C for 1 min.

Thus, the PCR was performed in a total of 40 cycles. A final

extension step was performed at 72 �C for 3 min. The PCR

products were loaded onto 8% polyacrylamide gels containing

a range of different denaturant concentrations (100% dena-

turant was a mixture of 7 M urea and 40% [v/v] formamide).

The ARC PCR products were applied to a gradient of 40%e60%

denaturant. The denaturant gradients used to separate the

amplified MMB, MSL, and MBT PCR products were 30e60%,

35e65%, and 40e60%, respectively. Each DGGE was performed

for 7 h at 150 V in 1� TAE electrophoresis buffer with the

D-Code system (BioRad, Hercules, CA). Following electropho-

resis, the gel was stained with ethidium bromide solution for

20 min, rinsed for 20 min in deionized water (DW), and pho-

tographed under UV transillumination. The visible DGGE

bands in each DGGE profile were excised directly from the gels

with a sterile blade, mixed with 40 mL of DW, and incubated

overnight at 4 �C. Each band was then eluted into solution and

2 mL was used as the template in a reamplification reaction

using the specific target primers.

The PCR fragments were purified using a purification kit

(General biosystem, Seoul, Korea) and cloned in Escherichia coli

DH5 alpha using a commercial cloning vector (pGEM-T Easy

Vector, Promega, Mannheim, Germany) according to the

Table 1 e The characteristics of the primer and probe sets for

Name Functionb Target group Sequence (5

MBT857F F primer Methanobacteriales CGWAG GGAAG CTG

MBT929F TaqMan AGCAC CACAA CGCG

MBT1196R R primer TACCG TCGTC CACT

MMB282F F primer Methanomicrobiales ATCGR TACGG GTTG

MMB749F TaqMan TYCGA CAGTG AGGR

MMB832R R primer CACCT AACGC RCAT

MSL812F F primer Methanosarcinales GTAAA CGATR YTCG

MSL860F TaqMan AGGGA AGCCG TGAA

MSL1159R R primer GGTCC CCACA GWGT

a Lee et al. (2009), Yu et al. (2005).

b F primer, Taqman, and R primer indicate forward primer, Taqman pro

c The number in parentheses indicates culture collection number.

manufacturer’s instructions. Cloned plasmids were isolated

from randomly selected colonies using a commercial kit

(General biosystem, Seoul, Korea) and used as templates for

the DNA sequencing analysis. Sequencing was performed

with T7 primers using an automated sequencer (3730� DNA

Analyzer, PerkineElmer, CA, USA). The sequencing results

were compared with the reference sequences in the BLAST

program in the National Center for Biotechnology Information

(NCBI) database.

2.4. Real-time PCR analysis

Quantitative real-time PCR was carried out using the Light-

Cycler 1.2 system (Roche, Mannheim, Germany) with three

primer and probe sets for the quantification of methanogens

(Table 1). Most methanogens in anaerobic bioreactors can be

covered by these primers and probe sets (Yu et al., 2005).

Consequently, three methanogenic groups in order level,

whole Archaea, and bacteria were monitored using real-time

PCR with the corresponding primer and probe sets. Each

reaction mixture was made using the LightCycler FastStart

DNA master Hybridization Probes kit (Roche): 9.6 ml of PCR-

grade pure water, 2.4 ml of MgCl2 stock solution (final concen-

tration 4 mM), 1 ml of each primer (final concentration 500 nM),

2 ml of the TaqMan probe (final concentration 200 nM), 2 ml of

reaction mix 10� solution, and 2 ml of template DNA. Amplifi-

cationwasperformed in a two-step thermal cyclingprocedure:

predenaturation for 10 minat 94 �Cfollowedby40cyclesof 10 s

at 94 �C and 30 s at 60 �C (except 63 �C for MMB-set). All DNA

templates were analyzed in duplicate.

The standard curves for the primer and probe sets used

were constructed as previously described (Yu et al., 2006)

using the representative strains listed in Table 1. The target

rRNA gene sequences were amplified from each strain by PCR

with the corresponding primer sets (Table 1) and cloned into

pGEM-T vectors (Promega, Mannheim, Germany). For each

plasmid, a 10-fold serial dilution series ranging from 102 to

109 copies/ml was generated and amplified in triplicate using

real-time PCR with the corresponding primer and probe sets.

The threshold cycle (CT) values determined were plotted

against the logarithm of their initial copy concentrations. The

standard curves constructed using different strains for

a primer and probe set showed no significant differences in

real-time PCR used in this researcha.

0 / 30) Represemtative strainsc

TT AAGT Methanobacterium thermoautotrophicum (DSM1053)

T GGA Methanobrevibacter arboriphilicus (DSM 1536)

C CTT

T GGG Methanocorpusculum parvum (DSM 3823)

A CGAAA GCTG Methanomicrobium mobile (DSM 1539)

H GTTTA C Methanospirillum hungatei (DSM 864)

C TAGGT Methanosarcina barkeri (DSM 800)

G CGARC C

A CC

be, and reverse primer, respectively.

wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 7 4903

their slopes at a 1% a-level. Thus, the average values of slope

and intercept for each set were used to quantify methano-

genic rRNA genes.

2.5. Statistical analysis

Redundancy analysis (RDA) is a linear, direct gradient con-

strained ordination method by which response variables are

constrained to be linear combinations of explanatory vari-

ables. If preliminary analysis of the main species shows

a linear response in their abundance in relation to the envi-

ronmental variables used, the use of the RDA will offer

a greater percentage of the variance explained regarding

canonical corresponding analysis (CCA), which is more suit-

able when there is a unimodal response (Sanchez et al., 2008).

RDA was used to investigate the statistical significance of

factors. Constrained ordination explicitly puts two matrices

into relationships: one dependentmatrix and one explanatory

matrix. Both are implied at the stage of the ordination

(Albrecht et al., 2008). Factors are constrained tomaximize the

redundancy index, which is defined as the product of the

variance in the predictor variable explained by predictor

factor and the variance in the response variable explained by

the predictor factor. The sum of canonical eigenvalues in RDA

equals the amount of variance in the response variable

explained by the predictor variable (Michalsen et al., 2007).

DNA fingerprinting techniques, such as PCR-DGGE, may

furtherelucidatemethanogenic community structureandallow

assessment of the changing composition of their communities

in anaerobic digesters. For this purpose,multivariate ordination

methods can be used to highlight the possible environmental

descriptors governing the ordering of information, such as

microbial community structure variation. In addition, these

methods have the major advantage of condensing the infor-

mation on a simple scheme (Imfeld et al., 2008).

2.6. Source of wastewater and physicochemical analysis

Swine wastewater (500 L) was collected from a local pig farm

(the Sansugol pig farm) in Kyungju, South Korea, and pre-

screened through an 850 mm sieve. The concentration of total

solids was 30.9 g/L. The sample was mixed homogeneously.

Physicochemical parameters were periodically analyzed

throughout the operation of the reactors. Chemical properties

such as chemical oxygen demand (COD), soluble chemical

oxygen demand (SCOD), and protein concentrations as well as

physical properties including total solids (TS), volatile solids

(VS), total suspended solids (TSS) and volatile suspended

solids (VSS) were determined according to the procedures in

Standard Methods (APHAeAWWAeWEF, 2005).

A gas chromatograph (6890 plus, Agilent, Palo Alto, CA)

equipped with an Innowax capillary column and a flame ioni-

zation detector was used to determine the concentrations of

volatile fatty acids (VFAs).Heliumwasusedas the carrier gasat

a flow rate of 2.5 mL/min with a split ratio of 10:1. Another

identical gas chromatographwith a HP-5 capillary column and

a thermal conductivity detector was used to analyze gas

composition in the biogas. Heliumwas the carrier gas at a flow

rate of 8 mL/min with a split ratio of 70:1. All analyses were

duplicated, and the results quoted as mean values.

3. Results and discussion

3.1. Reactor performance

The initial total VFA concentrations (TVFAs) in both reactors

were 7.5 g/L. Of all these VFA products, acetate and propionate

were the major fermentation products, comprising 67.7% of

TVFAs. Acetate alone accounted for more than 50%, and

propionate covered approximately 17% of TVFAs. As shown in

Fig. 2, the velocity of acetate decomposition during the batch

processwas faster than that of propionatemost likely because

propionate removal is a rate-limiting step in anaerobic

digestion (Tatara et al., 2008).

After 79 days, the accumulated methane production

volumes in reactors M1 and M2 were 7.9 L/L, and 4.5 L/L,

respectively (Figs. 2(A) and 3(A)). These results show that

methanogenic boosting was successful in the batch process

since all of the decomposed VFAs were transformed into

methane gas (Fig. 2).

3.2. PCR-based DGGE profile

To investigate themicrobial structures of the processes during

the batch step, DGGE analysis was used to determine the

methanogenic groups present in each reaction period. The

domain Eukarya was not analyzed because it has been repor-

ted that they are present at very low concentrations (below

0.8%) in anaerobic processes treating swine wastewater, even

though they are thought to play a role in a variety of other

anaerobic environments (Griffin et al., 1998). DGGE banding

profiles (Fig. 1) showed changes in the composition of

methanogen communities during the batch process in M1 and

M2. To compare the distinct DGGE bands patterns with time,

16S rRNA genes amplified from the reactor solution were

loaded onto both gels. Although there were many discernible

and weak DGGE bands in the two DGGE profiles, all of the

species identified from DGGE analysis could be grouped into

three different orders: MBT, MMB, and, MSL. Methanococcales

(MCC) was not detected in our system, which was ascribed to

the requirement of high salt concentration for their growth

(0.3e9.4% (w/v) NaCl) (Boon and Castenholz, 2001).

To quantify the microbial populations in during the batch

process, three different specific primer and probe sets for the

orders present, one archaeal sets, were used for the investi-

gation of microbial community dynamics.

3.3. Microbial community dynamics and theirredundancy analysis

Changes in the 16S rRNA gene concentrations of the microbes

present in the digesters are shown in Figs. 2(B) and 3(B). In

addition, redundancy analyses of shifts in the quantitative

structure of methanogenic community and their correlation

between each order of Archaea, VFAs and accumulated

methane in M1 and M2 are exhibited in Fig 4. In order to clarify

changes inmethanogenic community structure, RDAwas used

because it is known to be the most generally effective ordina-

tion method for ecological community data (McCune and

Grace, 2002). The RDA plot from the analysis of methanogenic

Fig. 1 e Methanogenic DGGE banding profiles of the PCR products amplified with 16S rRNA gene primers. The template DNA

was extracted from each experimental system at steady state.

Fig. 2 e Chemical profiles along with methane production

(A) and changes in methanogenic 16S rRNA gene

concentration (B) in M1.

wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 74904

16S rRNA gene concentration profiles (Fig. 4) displays the

contributions to shifts in methanogenic community structure

as well as correlations between the abundances of methano-

genic groups and the concentration of VFAs or methane. For

the plot, the eigenvalues of M1 and M2 were 0.75, and 0.86,

respectively, meaning that summation of each variable’s

dispersion could provide an explanatory power of 74.9% and

86.3% for the respectivemodels. Finally, the p-values of the two

models were 0.07, and 0.02 showing that each model was

meaningful within 7% and 5% a-level, respectively.

The rRNA concentration profiles for reactors M1 and M2

showed that therewere significant variations inmethanogenic

populations with respect to performance data. Although the

physicochemical conditions were almost the same in both

digesters, the pattern of shifts in microbial community

dynamics was different. In M1, the initial gene copy concen-

tration of MMB was higher than those of MBT and MSL. The

concentration of MMB was maintained at a roughly constant

level of 4.9� 108; copies/mL during days 0e40 (the lag phase of

methanation). After day 40, MMB andMSL gene copy numbers

began to fluctuate, correspondingwith the point thatmethane

gas began to be produced. As acetate was degraded, the

concentration ofMMB16S rRNAgene copies began to increase,

reaching a peak of 1.0� 109; copies/mL (approximately a two-

fold increase from the initial value) at the final point.

In reactor M2, the initial gene copy concentration of MMB

was also higher than any other methanogenic group at order

level, and its dominance was maintained at 3.0e4.0

� 108 copies/mL until ca. day 40. At that point, there were fluc-

tuations inMMBlevelsprobablydue tocompetitionwithMSL.At

the end of the process, MMB was still the dominant order with

Fig. 4 e Redundancy analyses of shifts in the quantitative

structure of methanogenic community and their

correlation between species of Archaea, VFAs, and

accumulated methane in M1 digesters (A) and in M2

digesters (B).

Fig. 3 e Chemical profiles along with methane production

(A) and changes in methanogenic 16S rRNA gene

concentration (B) in M2.

wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 7 4905

6.7� 108 copies/mL. Although the initial concentration of MSL

was lower than that of MMB, MSL was able to proliferate along

with MMB from day 45 probably because MSL belongs to the

aceticlastic group, which can induce first methanation using

acetate (Lee et al., 2009). As shown in Fig 4(a) and (b), the

approximated correlation between aceticlastic methanogens

and acetate (HAc) was positive. However, when acetate was

almost used up, MSL could no longer compete with MMB, and

MMB group became dominant once more (Figs. 2 and 3). This

result was consistent with previous evidence that populations

of hydrogenotrophic methanogens such as MMB increase

gradually after the termination of aceticlasticmethanation (Lee

etal., 2009), andwith thenegativecorrelationbetweenMMBand

acetate (HAc) (Fig. 4(a) and (b)).

The initial gene concentrations for MBT were 1.3� 108 and

8.9� 107 copies/mL in M1 and M2. This group also began to

fluctuate soon aftermethane production phase began. Its final

concentrations were 7.8� 107 and 9.9� 107 copies/mL in M1

and M2.

In our trials, even though MBT also belongs to the class of

hydrogenotrophic bacteria, a group which is often present

during the methane production phase, its population was not

dominant compared to MMB. Furthermore, MBT’s profile in

RDA was comparatively due to the fact that in Fig. 4(a) and (b),

its correlation with acetate (HAc) and propionate (HPro) was

positive and that there was almost no correlation between

MBT and MSL.

Taken together, the results showed that in both reactors,

MSL had a positive correlation with acetate, and no correla-

tion with propionate. Similarly, MBT’s approximated correla-

tion patterns were roughly equivalent in both digesters. For

this reason, it is likely that the differences in the MMB

community structure between the two anaerobic digesters

were responsible for the differences in the total amount of

methane produced. According to Fig. 4, the major factor

determining the ability to produce methane in each digester

was the MMB arrow line in RDA. Although MMB was found to

be the dominant group in both of the reactors at the end of the

experiment, the correlation pattern between MMB and

propionate differed in the two reactors (Fig. 4(a) and (b)). The

wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 74906

positive association between MMB and propionate was

stronger in M1 than in M2. This could explain the higher

amount of methane produced in M1 and suggests that effec-

tive metabolism of propionate by MMB following the onset of

acetate consumption was a crucial factor controlling produc-

tion of methane in our systems.

To our knowledge, this may be the first report on the

correlationbetweenmethanogenic communitydynamicswith

major VFAs such as acetate and propionate. The study may

also be the first to use a quantitative approach to examine how

methanogenic communities evolve differently in digesters by

using real-time PCR together with a multivariate ordination

technique (RDA). The results could not explain precisely why

changes in theMMB community led to differences inmethane

production. However, it should be noted that relationships

between microbial groups can impact on their metabolism

patterns, and this is likely to be related to the amount of

methane generated. Furthermore, because microbial pop-

ulations that comprise less than 1% of the total target

community are generally not able to be identified using

molecular fingerprinting methods (Forney et al., 2004), it is

possible that a functionally important yet numerically negli-

gible population could have been missed in the diversity

analysis results. For this reason, further research applying

quantitativemolecular characterization techniques overmore

taxonomic levels could be undertaken to allow for a better

understanding of microbial behavior in anaerobic processes.

4. Conclusions

In this study, quantitative changes in methanogenic

community structure were related to changes in chemical

properties in two anaerobic digestion systems operated under

physicochemically similar conditions by using multivariate

analysis. The experiments can be summarized as follows.

(1) The qualitative structure of the methanogenic communi-

ties was assayed in two batch anaerobic digesters. The

methanogenic groups, MBT, MMB, and MSL, were identi-

fied at order level. MCC was not detected.

(2) Based on this, primer and probes sets designed to amplify

DNA from methanogens were used to perform a quantita-

tive PCR-based analysis of microbial population dynamics.

MMBwas found to be the dominant group and its variation

pattern was different in two digesters. The population

profiles of other methanogenic groups were also different.

(3) Through RDA, correlations between MBT and MSL pop-

ulations with chemical properties such as VFAs were

shown to be consistent in both digesters, whereas the

correlation between MMB and propionate was different.

(4) This suggests that despite MMB remaining the dominant

methanogenic group in both reactors its interactions with

other methanogens can affect final methane yields.

The experimental results show that more study should be

devoted to quantitative and multivariate analysis of meth-

anogenic populations to better understand the influence of

their metabolism and behavior on methanogenesis in anaer-

obic digesters.

Acknowledgements

This research was supported by the Korea Ministry of

Knowledge and Economy (MKE) as a Manpower Development

Program for Energy & Resources, and the Ministry of Envi-

ronment (MOE) as a Human Resource Development Project for

Waste to Energy.

r e f e r e n c e s

Akarsubasi, A.T., Ince, O., Kirdar, B., Oz, N.A., Orhon, D.,Curtis, T.P., Head, I.M., Ince, B.K., 2005. Effect of wastewatercomposition on archaeal population diversity. Water Res. 39,1576e1584.

Albrecht, R., Joffre, R., Gros, R., Petit, J.L., Terrom, G., Perissol, C.,2008. Efficiency of near-infrared reflectance spectroscopy toassess and predict the stage of transformation of organicmatter in the composting process. Bioresour. Technol. 99 (2),448e455.

APHA, AWWA, WEF, 2005. Standard methods for the examinationof water and wastewater. American Public Health Association.American Water Works Association and Water EnvironmentFederation, Washington DC, USA.

Bernall, C.B., Vazquez, G., Quintal, I.B., Bussy, A.L., 2008.Microalgal dynamics in batch reactors for municipalwastewater treatment containing dairy sewage water. WaterAir Soil Poll 190 (1-4), 259e270.

Boon, D.R., Castenholz, R.W. (Eds.), 2001. Bergey’s Manual ofSystematic Bacteriology: the Archaea and the DeeplyBranching and Phototrophic Bacteria, second ed. vol. 1Springer, New York.

Bossio, D.A., Scow, K.M., 1995. Impact of carbon and flooding onthe metabolic diversity of microbial communities in soils.Appl. Environ. Microbiol. 61 (11), 4043e4050.

Bremer, C.G., Braker, G., Matthies, D., Reuter, A., Engels, C.,Conrad, R., 2007. Impact of plant functional group, plantspecies, and sampling time on the composition of nirK-typedenitrifier communities in soil. Appl. Environ. Microbiol. 73 (21),6876e6884.

Comor, V., Orgeas, J., Ponell, P., Rolando, C., Delettre, Y., 2008.Impactof anthropogenic disturbancesonbeetle communities ofFrench Mediterranean coastal dunes. Biodivers. Conserv. 17 (8),1837e1852.

Deng, L., Caic, C., Chen, Z., 2007. The treatment of pig slurry bya full-scale Anaerobic-Adding Raw Wastewater-IntermittentAeration Process. Biosyst. Eng. 98 (3), 327e334.

Elissen, H., Edwin, T., Peeters, H.M., Buys, B.R., Klapwijk, A.,Rulkens, W., 2008. Population dynamics of free-swimmingAnnelida in four Dutch wastewater treatment plants in relationto process characteristics. Hydrobiology 605 (1), 131e142.

Fernandez, A.S., Hashsham, S.A., Dollhopf, S.L., Raskin, L.,Glagoleve, O., Dazzo, F.B., Hickey, R.F., Criddle, C.S., Tiedje, J.M.,2000. Flexible community structure correlates with stablecommunity fuction inmethanogenic bioreactor communitiesperturbated by glucose. Appl. Environ. Microbiol. 66, 4058e4067.

Forney, L.J., Zhou, X., Brown, C.J., 2004. Molecular microbialecology: land of the one-eyed king. Curr. Opin. Microbiol. 7,210e220.

Gilbride, K.A., Frigon, D., Cesnik, A., Gawat, J., Fulthorpe, R.R.,2006. Effect of chemical and physical parameters on a pulpmill biotreatment bacterial community. Water Res. 40 (4),775e787.

Griffin, M.E., McMahon, K.D., Mackie, R.I., Raskin, I., 1998.Methanogenic population dynamics during start-up of

wat e r r e s e a r c h 4 4 ( 2 0 1 0 ) 4 9 0 0e4 9 0 7 4907

anaerobic digesters treating municipal solid waste andbiosolids. Biotechnol. Bioeng. 57 (3), 342e355.

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. Appl. Environ. Microbiol. 72, 1623e1630.

Imfeld, G., Nijenhuis, I., Nikolausz, M., Zeiger, S., Paschke, H.,Drangmeister, J., Grossmann, J., Richnow, H., Weber, S., 2008.Assessment of in situ degradation of chlorinated ethenes andbacterial community structure in a complex contaminatedgroundwater system. Water Res. 42 (4e5), 871e882.

Jan, L., Peter, S., 2003. Multivariate Analysis of Ecological DataUsing CANOCO. Cambridge University Press, Cambridge.

Kennedy, N., Brodie, E., Connolly, J., Clipson, N., 2004. Impact oflime, nitrogen and plant species on bacterial communitystructure in grassland microcosms. Environ. Microbiol. 6 (10),1070e1080.

Lee, C., Kim, J., Hwang, K., Flaherty, V., Hwang, S., 2009.Quantitative analysis of methanogenic community dynamicsin three anaerobic batch digesters treating differentwastewaters. Water Res. 43, 157e165.

Lee, C., Kim, J., Shin, S.G., Hwang, S., 2008. Monitoring bacterialand archaeal community shifts in a mesophilic anaerobicbatch reactor treating a high-strength organic wastewater.FEMS Microbiol. Ecol. 65, 544e554.

McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities.MjM Software Design, USA, Glenden Beach, OR.

Michalsen, M.M., Peacock, A.D., Spain, A.M., Smithgal, A.N.,White, D.C., Sanchez-Rosario, Y., Krumholz, L.R., Istok, J.D.,2007. Changes in microbial community composition and

geochemistry during uranium and technetiumbioimmobilization. Appl. Environ. Microbiol. 73 (18), 5885e5896.

Roy, C.S., Talbot, G., Topp, E., Beaulieu, C., Palin, M.-F., Masse, D.I.,2009. Bacterial community dynamics in an anaerobic plug-flow type bioreactor treating swine manure. Water Res. 43 (1),21e32.

Sanchez, F., Serrano, A., Parra, S., Ballesteros, M., Cartes, J.E.,2008. Habitat characteristics as determinant of the structureand spatial distribution of epibenthic and demersalcommunities of Le Danois Bank (Cantabrian Sea, N. Spain).J. Marine Systems 72 (1e4), 64e86.

Speece, R.E., 1996. Anaerobic Biotechnology for IndustrialWastewates. Archae Press, Nashville, TN.

Tatara, M., Makiuchi, T., Ueno, Y., Goto, M., Sode, K., 2008.Methanogenesis from acetate and propionate by thermophilicdownflow anaerobic packed-bed reactor. Bioresour. Technol.99 (11), 4786e4795.

Yu, Y., Kim, J., Hwang, S., 2006. Use of real-time PCR for group-specific quantification of aceticlastic methanogens inanaerobic processes: population dynamics and communitystructures. Biotechnol. Bioeng. 93, 424e433.

Yu, Y., Lee, C., Kim, J., Hwang, S., 2005. Group-specific primer andprobe sets to delete methanogenic communities usingquantitative real-time polymerase chain reaction. Biotechnol.Bioeng. 89, 670e679.

Zumstein, E., Moletta, R., Godon, J.J., 2000. Examination of twoyears of community dynamics in an aerobic bioreactor usingfluorescence polymerase chain reaction (PCR) single-strandconformation polymorphism analysis. Environ. Microbiol. 2,69e78.