ORIGINAL PAPER

Effect of temperature and hydraulic retention time on volatilefatty acid production based on bacterial community structurein anaerobic acidogenesis using swine wastewater

Woong Kim • Seung Gu Shin • Juntaek Lim •

Seokhwan Hwang

Received: 26 December 2012 / Accepted: 15 January 2013 / Published online: 1 February 2013

� Springer-Verlag Berlin Heidelberg 2013

Abstract To investigate the effect of hydraulic retention

time (HRT) and temperature (T) on bacterial community

structure and volatile fatty acids (VFAs) production of an

acidogenic process, and VFA production and changes in

the bacterial community in three identical automated

anaerobic continuously-stirred tank reactors were analyzed

using response surface analysis (RSA) and nonmetric

multidimensional scaling (NMDS). For RSA, 11 trials were

conducted to find the combination of T and HRT under

which VFA production was greatest; VFA production was

affected more by HRT than by T. To identify the bacterial

community structure in each trial, DNA from each exper-

imental point of the RSA was analyzed using denaturating

gradient gel electrophoresis (DGGE), and eight bacteria

species were detected. NMDS was conducted on band

intensities obtained using DGGE, and bacterial community

structure was affected more by T than by HRT. Taken

together, these results suggest that VFA production during

acidogenesis was more dependent on the physicochemical

properties of acidogens, such as their specific growth rate

or contact time with of substrates, than on changes in the

microbial community.

Keywords Anaerobic digestion � Response surface

analysis � Nonmetric multidimensional scaling analysis �Swine wastewater

List of symbols

ANOVA Analysis of variance

COD Chemical oxygen demand

CSTR Continuously stirred tank reactors

DDW Deionized and distilled water

DGGE Denaturating gradient gel electrophoresis

HRT Hydraulic retention time

NMDS Nonmetric multidimensional scaling

OD Optical density

RSA Response surface analysis

RSM Response surface methodology

sCOD Soluble chemical oxygen demand

T Temperature

TS Total solids

TSS Total suspended solids

TVFA Total volatile fatty acids

UPGMA Unweighed pair group method with arithmetic

VFA Volatile fatty acid

VS Volatile solids

VSS Volatile suspended solids

Introduction

The swine industry is growing rapidly; the trend is toward

more-concentrated piggeries with herd numbers in the

W. Kim

Department of Chemical and Biomolecular Engineering,

KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon 305-701,

Republic of Korea

S. G. Shin

Infrastructure and Environment Division, School of Engineering,

University of Glasgow, Glasgow G12 8LT, UK

J. Lim

POSCO, Gangnam-gu, Seoul, Republic of Korea

S. Hwang (&)

School of Environmental Science and Engineering,

Pohang University of Science and Technology, San 31,

Hyoja-dong, Nam-gu, Pohang, Kyungbuk 790-784,

Republic of Korea

e-mail: shwang@postech.ac.kr

123

Bioprocess Biosyst Eng (2013) 36:791–798

DOI 10.1007/s00449-013-0905-7

thousands. These large herds produce large quantities of

wastewater, and treatment of this wastewater requires

reducing the large amount of residue and the high levels of

organic and inorganic nutrients [1].

Anaerobic digestion generates little residue and is a

cost-effective method of treating swine wastewater; in

addition, the methane produced can be used as an energy

source [2]. Other advantages of the anaerobic treatment of

swine wastewater include nutrient conservation and odor

reduction. Anaerobic digestion is a multi-stage biochemical

process in which acidogenic bacteria (acidogens) ferment

into intermediates (mostly VFAs), which are subsequently

reduced to methane and CO2 by methanogenic archaea

(methanogens) [3]. Anaerobic reactors can be classified

into two types: single-stage, in which acidogenesis and

methanogenesis are conducted in a single reactor and two-

stage, in which they are conducted in separated reactors

with acidogen byproducts being fed to methanogens. The

two-stage process is advantageous because conditions in

the two reactors can be optimized separately for the acid-

ogens and the methanogens, but the requirement for an

additional large reactor for the acidification of influent

organics may make the two-stage process more expensive

than the single-stage process.

Recently, it was reported that metabolism of VFAs by

bacterial community varied with the type of methanogenic

group. These methanogenic community structure and

dynamics was deeply related to the successful process

operation as a result of VFAs metabolism [4]. Also, the

effect of methanogenic community structure on pilot plant

of anaerobic process was also studied well [5]. According

to these studies, even though aceticlastic methanogens can

grow on VFA-like acetate, hydrogenotrophic methanogens

using H2/CO2 were considered as more meaningful archaea

of anaerobic digestion to VFAs metabolism and methane

production [6]. However, these previous studies have been

primarily focused on the methanogenic community in the

anaerobic system. Overall enhancement of anaerobic

digestion requires understanding of optimum growth con-

ditions and behavior of acidogens in a two-stage process,

because they are the primary producers of short-chain

organic acids which are major substrates for methanogens.

Although organic acids are important substrates in metha-

nogenesis, high concentrations of these acids can cause

anaerobic digestion systems to fail [7]. Furthermore,

acidogens include many microbial species, and little

information is available concerning their characteristics

and growth conditions.

Acidogenesis can be classified into two classes: meso-

philic acidogenesis operates at T & 40 �C and thermo-

philic acidogenesis operates at T [ 50 �C, and can provide

many benefits including increased rates of digestion and

hydrolysis, decreased digester volume, ease of liquid–solid

separation, and efficient production of VFAs [8]. In prac-

tice, thermophilic acidogenesis may be initiated with

mesophilic sludge inoculum because the thermophilic

process is less prevalent than the mesophilic process in

field-scale applications [9]. However, thermophilic anaer-

obic digesters generally harbor less microbial diversity

than do mesophilic digesters, with thermophilic bacteria as

dominant species. Because microbial diversity is closely

related to functional stability, the effects of high-T on the

bacterial community dynamics in the anaerobic digestion

process during acidogenesis using mesophilic sludge

inoculum [10] must be studied.

Despite the promise of two-stage digestion as an efficient

treatment of swine wastewater, little information is avail-

able about two-stage anaerobic processes that include an

initial thermophilic acidogenic stage optimized to maxi-

mize the rate of acetic acid production, which is related to

bacterial community structure and environmental factors

such as HRT and T. For this reason, our previous study [11]

evaluated the optimum HRT and T for thermophilic

acidogenic reactors to maximize the production of acetic

acid. The aim of the present study was to investigate how

environmental factors influence VFA production by affect-

ing bacterial community structure during acidogenesis. For

this purpose, RSA was conducted to determine the combi-

nation of HRT and T that maximized VFA production in the

acidogenic digesters, DGGE was conducted to characterize

microbial communities by recovering and sequencing

amplification products [12], and NMDS was conducted to

determine the correlation among HRT, T, bacterial com-

munity structure, and VFA production.

Materials and methods

Bioreactor operation

Three identical automated anaerobic CSTRs, each with a

working volume of 4 L and equipped with temperature

controllers, were operated in batch mode and continuous

mode sequentially for acidogenesis. Anaerobic seed sludge

from a municipal wastewater treatment plant in Pohang,

South Korea, was mixed (1 % w/v of total suspended

solids) with swine wastewater as a substrate for acidogens

and cultivated in batch mode in the three bioreactors for

36 h to enrich a mixed population of acidogens. For the

RSA, three trials were conducted using three identical

automated CSTRs simultaneously; a total of 11 combina-

tions of HRT and T were evaluated using four operations of

three CSTRs. When the production of TVFAs was maxi-

mized after 36 h (data not shown) after the batch mode

process, the continuous acidogenesis was initiated by

turning on the pump. TVFA concentrations were measured

792 Bioprocess Biosyst Eng (2013) 36:791–798

123

and subjected to RSA with HRT and T as independent

variables. To monitor the process, physical properties such

TS, VS, TSS, and VSS, and chemical properties such as

COD and sCOD were monitored; pH and the biogas pro-

duction were also measured (‘‘Denaturating gradient gel

electrophoresis analysis’’). In preliminary experiments, pH

was nearly constant during the process, so pH was not

usually controlled in the optimization study.

Central composite in cube design and selection

of variables

Microbial growth and enzyme activity have specific opti-

mum conditions with regard to many variables, including

pH, HRT, T, and the presence of inhibitors. For example,

the pH and dilution rate of the bioreactor affect the rate of

formation of acetic, propionic, and butyric acids [13].

RSM was applied to analyze and to optimize the factors

that affect TVFA production associated with simultaneous

changes in HRT and T during thermophilic acidogenesis.

A sequential procedure of collecting data, estimating

polynomials (Eq. 1), and checking the adequacy of the

model was used [14]:

ga ¼ c0 þXn

i¼1

aixi þXn

i¼1

aiix2i þ

X

i

X

j

aijxixj; i\j

ð1Þ

where ga is measured acetic acid concentration (mg/L), xk

is independent variable k (1 = HRT; 2 = T), co is the

regression constant and ak are regression coefficients of the

independent variable k. The least squares method was used

to estimate the parameters in the approximating polyno-

mials (Eq. 1). The orthogonal design, which consists of a

2 9 2 factorial design augmented by a center and 2 9 2

axial points was employed [15, 16].

Based on the population growth rate of anaerobes on

swine wastes in a CSTR, 0.5 B HRT B 2.5 d was chosen

to give sufficient residence time for acidogenic activity

[17], and 40 B T B 60 �C was chosen because pathogenic

bacteria can be inactivated or HRT can be shortened using

thermophilic temperature processes at [50 �C rather than

the mesophilic process [8]. Thus, the trial points of RSM

based on central composite selection of variables were

shown in Table 1.

DNA extraction

Extraction of DNA from anaerobic sludge and bioreactor

samples was performed as described elsewhere [18].

Appropriate dilutions were performed to obtained VSS

concentrations \1.5 g/L. Cells from 500 lL samples were

harvested by centrifuging at 14,000 rpm for 5 min, then

decanting the supernatant and resuspending the residual

pellet with 1 mL of DDW, centrifuging again in the same

manner to ensure maximal removal of residual medium.

The supernatant was carefully removed, and the pellet was

resuspended in 100 lL of DDW before DNA extraction.

A fully-automated nucleic acid extractor (Magtration

System 6GC, PSS Co., Chiba, Japan) employing magnetic

bead technology [19] was used to extract and purify

genomic DNA. The automated instrumentation extracts

DNA with a consistent efficiency and high purity by

eliminating manual preparation steps that may cause cross-

contamination [20]. Genomic DNA extracted from the

samples was stored at -20 �C until analysis.

Denaturating gradient gel electrophoresis analysis

For DGGE analysis, the DNA samples from an acidogen-

esis bioreactor were used. The V3 to V5 region of 16S

rRNA genes in the extracted DNA was amplified using

PCR with a set of universal bacterial primers, BAC 228F

with a 40-bp GC-clamp (50-CGCCC GCCGC GCGCG

GCGGG CGGGG CGGGG GCACG GGGG G-30) [21]

and BAC 805R. A touch-down PCR program was used for

all amplifications to minimize non-specific amplification

[22]. After an initial denaturation at 94 �C for 10 min, 20

cycles of touch-down PCR were performed (denaturation at

94 �C for 30 s, annealing for 30 s with a 0.5- �C/cycle

decrement at 65–55 �C and extension at 72 �C for 1 min),

followed by 15 cycles of regular PCR (94 �C for 30 s,

55 �C for 30 s, and 72 �C for 1 min and a final extension

step at 72 �C for 7 min).

A 20-lL sample of each PCR product was loaded onto

8 % acrylamide gel containing a linear gradient ranging

from 30 to 60 % denaturant (100 % denaturants is a

Table 1 Experimental design and observed concentration of total

volatile fatty acids production in the acidogenesis using swine waste-

water [6]

Trial Independent variables TVFAs (g/L)

Temperature

(�C)

HRT

(days)

Linear design 1 40 0.5 0.29

2 40 2.5 1.44

3 60 0.5 0.72

4 60 2.5 1.35

5a 50 1.5 1.70 ± 0.03

Quadratic design 6 40 1.5 1.55

7 60 1.5 1.72

8 50 0.5 0.71

9 50 2.5 1.77

a Center point was repeated by tree times

Bioprocess Biosyst Eng (2013) 36:791–798 793

123

mixture of 7-M urea and 40 % [vol/vol] formamide). The

DGGE was performed for 7 h at 150 V in 1x TAE elec-

trophoresis buffer in a D-code system (Bio-Rad, Inc.,

Hercules, CA, USA). Following electrophoresis, the gel

was stained in ethidium bromide solution for 20 min,

rinsed for 20 min in DDW, and photographed under UV

transillumination.

For further identification of representative DGGE bands

in individual samples, DGGE fragments were excised

directly from the gels with a sterile blade, mixed with

40 lL of DDW, and incubated overnight at 4 �C. A 2-lL

sample of each band elution solution was re-amplified with

no GC-clamp BAC 338F and BAC 805R. The amplified

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 manufacturer’s instructions. Cloned plasmids were

isolated from randomly selected colonies of the library

using a commercial kit (General biosystem, Seoul, Korea)

and used as templates for the DNA sequencing analysis.

After DNA sequencing with T7 primers, the results were

compared with reference sequences in GenBank to identify

phylogenetic affiliations.

Multivariate scaling of DGGE profiles

DGGE gel images were scanned and then processed using

LABWORKS software (version 3.0.3, UVP, Upland, CA,

USA), and their band intensities were scored by measuring

the absolute integrated optical density of each band using

the software. The Jaccard similarity produced was analyzed

using NMDS, which is a multivariate ordination method

that reformats complex data to construct a new set of

variables; it can find a non-parametric monotonic rela-

tionship between the dissimilarities in the item–item matrix

and the Euclidean distance between items, and the location

of each item in the low-dimensional space [23]. NMDS

condenses the band pattern data of one lane to one point in

a two-dimensional plane, so that highly similar data are

plotted close together [24]. Therefore, changes in com-

munity structure can be visualized by connecting consec-

utive data points. UPGMA clustering analysis was

conducted based on the Jaccard similarity coefficient. The

intensity of each band in each lane from DGGE analysis

was quantified using optical density analysis (LabWorks

software, version 4.0, UVP, UK); this band represents one

microbial species. To investigate the relationship between

microbial community structure in RSM trials and envi-

ronmental factors used as independent variables in RSM,

NMDS was conducted using PC-ORD software (version 5,

MjM Software, Wagga Wagga, NSW, Australia); the two

independent variables were HRT and T.

Wastewater and physicochemical analysis

Swine wastewater (500 L) of DGGE was collected from

the Sansugol pig farm in Kyungju, South Korea, pre-

screened through an 850-lm sieve, then mixed to homog-

enize it.

Physicochemical parameters were periodically analyzed

throughout the operation of the three reactors. Chemical

properties such as COD and SCOD, and physical properties

such as TS, VS, TSS, and VSS were determined according

to the procedures in Standard Methods [25]. A gas chro-

matograph (6890 plus, Agilent, Palo Alto, CA) equipped

with an Innowax capillary column and a flame ionization

detector was used to determine the concentrations of VFAs.

He was the carrier gas at a flow rate of 2.5 mL/min with a

split ratio of 10:1. He was 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 means.

Results and discussion

Reactor performance and response surface

methodology

For the RSA, three identical acidogenic bioreactors were

operated simultaneously as continuously fed systems after

1.5 d in batch mode to boost acidogens in them. COD

concentration of the substrate was maintained at 70 g/L

and steady state was assumed after 10 turnovers. Inde-

pendent variables were HRT and T, and the dependent

variable was TVFA, including acetate, propionate,

n-butyrate, isobutyrate, n-valerate, isovalerate, n-caproate,

and isocaproate. The operating condition at the center point

was 1.5 days HRT and 50 �C. Repeated observations at the

center were used to estimate the experimental error. A part

of the present investigation has been reported in a previous

study, which reported an experiment for RSA that revealed

the production pattern of given categories of TVFA during

acidogenesis using swine wastewater [11]. This statistical

model produced ANOVA results of Eq. (1). Two-dimen-

sional response surfaces of the quadratic model described

an estimated optimum point for generating TVFAs. In the

cited article, lack of fit was not meaningful but regression

analyzed was significant at the 1 % a level, which meant

that the quadratic model in RSA fit the response model.

From this, we have extended our investigation to determine

the significance of the independent variables. The interac-

tion between HRT and T was not significant at the 1 % alevel, whereas HRT had a significant effect on TVFA

production in the criteria given (‘‘Central composite in

cube design and selection of variables’’) at a given T

(Table 2). T affected TVFA production at the 5 % a level.

794 Bioprocess Biosyst Eng (2013) 36:791–798

123

However, mean squares and p values indicated that HRT

had a greater effect on TVFA production than did T.

DGGE analysis of bacterial community at each

experimental trial of RSA

For DGGE analysis, DNA was extracted to examine shifts

in bacterial community structure. Ten DGGE bands, des-

ignated B1–B10, were visualized (Fig. 1). The DGGE band

was not detected at the combination of HRT = 0.5 d and

T = 60 �C. The affiliations of the 16S rRNA gene

sequences were determined by comparing them to the

GenBank data base (Table 3).

B2 and B10 showed 99 % similarities to several Bacil-

lus halodurans strains (Table 3). This species is a ther-

motolerant b-galactosidase and xylase producer that can

use polysaccharide as an energy source in the presence of

Na? and NH3 at 7 B pH B 10.5 [26]. The presence of

B. halodurans in all trials was assumed to be a result of its

thermotolerance; its maximum apparent rate of population

increase exceeded those of the other bacteria detected.

B3 and B4 were closely related to swine effluent bac-

teria with 98 % similarity (Table 3). Although this species

is yet to be unclassified, it was closely associated with a

Pseudomonas sp., a kind of c-proteobacterium that is

generally found in thermophilic digesters [27].

B5 and B6 were closely associated with Pseudomonas

sp. BBTR25, which is also closely related to swine effluent

bacteria [28]. Pseudomonas sp. BBTR25 has the ability to

denitrify nitrite or nitrate [27]. Therefore, the ever-present

bands of swine effluent bacteria in all trials including swine

wastewater and anaerobic seed sludge indicate that deni-

trification might occur during acidogenesis only in the

presence of nitrite or nitrate.

B7 was closed related to Lutispora thermophila, a spore-

forming, rod-shaped bacterium that grows at moderately

thermophilic condition (55 B T B 58 �C); its optimal pH

is 7.5–8.0 [29]. That this species was not observed in T1,

T4, and T7 (T \ 55 �C) is consistent with its thermophi-

licity. However, B7 cannot be L. thermophila because its

similarity to band 7 was rather low (88 % homology).

L. thermophila can grow on peptone, tryptone, Casamino

acids, casein hydrolysate, methionine, threonine, trypto-

phan, cysteine, lysine, and serine, but has no known ability

to metabolize carbohydrates [29]. Actually, the protein

degradation efficiency based on the Total Kjeldahl Nitro-

gen measurement was higher in 50 and 60 �C trials than in

40 �C trials (data not shown); this result suggests that B7 is

likely to be related to L. thermophila.

B9 was closely related to Thermomonas koreensis, a

carbohydrate utilizer that produces acetic or propionic acid

and which grows in a comparatively wide range of tem-

perature from 18 to 50 �C, but prefers slightly thermophilic

conditions [30, 31]. This band became more intense as T

increased (Fig. 1). Therefore, due to its slight preference

for T [ 50 �C, T. koreensis in mesophilic sludge or swine

wastewater influent may gradually become more common

as T increases during acidogenesis.

B1 and B8 were uncultured species. With 95 % simi-

larity, B1 was broadly related to Sedimentibacter sp.,

which is a variable bacterial group that has not been spe-

cifically classified. With 89 % similarity, B8 was slightly

related to Gracilibacter thermotolerans, which ferments

carbohydrates to ethanol, acetate, or lactate [32]. That B8

was only detected at 40 �C (T1, T4, and T7 in Fig. 1) is

consistent with G. thermotolerans’ optimum T range of

42.5–46.5 �C [32].

Multivariate analysis of DGGE bands

The intensity of each band in each lane from DGGE

analysis was quantified using optical density analysis; each

band represents a microbial species. To investigate the

relationship between microbial community structure in

RSM trials and environmental factors used as independent

variables in RSM, NMDS was conducted; the independent

variables were HRT and T.

Table 2 ANOVA results of quadratic model for optimization of

acidogenesis using swine waste with respect to two independent

variables and their interactions [6]

Source Mean square DF P value

Temperature 0.0633 2 0.0318

HRT 1.1024 2 0.0000

Temperature versus HRT 0.0645 1 0.0403

Fig. 1 Bacterial DGGE profiles of the PCR products amplified with

16S rRNA gene primers at each trial of response surface analysis (SLinitial seed, SW swine wastewater as a substrate of acidogens), lane

labels show each trial for response surface analysis (T1, T4, and T7,

T = 40 �C; T2, T5, and T8, T = 50 �C; T3 and T6, T = 60 �C;

T1–3, HRT = 2.5 d; T4–6, HRT = 1.5 d; T7–8, HRT 0.5 d; Trial at

T = 60 �C and HRT = 0.5 d was not detected; Codes on the DGGE

gel identify the bands excised for sequencing)

Bioprocess Biosyst Eng (2013) 36:791–798 795

123

The NMDS map (Fig. 2) of bacterial communities was

obtained from DGGE profiles. In view of the correlation

between bacterial communities and environmental factors

HRT and T, microbial community structures at 50 and

60 �C were estimated to be similar to each other because

T2, T3, T5, and T6 were clustered in the same position on

the map. Although T8 represented 50 �C, the bacterial

community at this point was different from those in other

50 or 60 �C trials, possibly because band B10 was weaker

in T8 than in other lanes (Fig. 1). However, B. halodurans,

regarded as B10, was observed as B2, which has a pattern

similar to those in T2, T3, T5, T6, and so T8 could also be

clustered in the microbial community structure at 50 or

60 �C. Meanwhile, the bacterial community structure at

40 �C (T1, T4, and T7) was revealed to vary because these

points were the most widely dispersed in the map [33]; i.e.,

at 40 �C, microbial communities varied along with HRT

rather with T. The weaker intensities of B1 related to

Sedimentibacter sp., B2 and B10 closely associated

with B. halodurans, and the shorter HRT was at 40 �C.

Table 3 Bacterial identification of amplified 16S rRNA gene sequences excised from the DGGE gels shown in Fig. 1

Bacterial DGGE band

(based on the bacterial 16S rDNA)

Nearest species and taxon GenBank accession

number

Similarity

(%)

B1 Uncultured bacterium clone ATB-KM1254 DQ390276 97

Uncultured bacterium clone C23B EU219936 96

Sedimentibacter sp. JN18_V27_I EF059533 95

B2 Bacillus halodurans AB274919 99

Bacillus halodurans strain PPKS2 EU118675 99

Bacillus halodurans strain XJRML-1 EF466141 99

B10 Bacillus halodurans strain PPKS2 EU118675 99

Bacillus clausii strain XJU-2 AY960115 99

Bacillus sp. A-59 AB043856 99

B3, B4 Swine effluent bacterium CHNDP38 DQ337540 98

Pseudomonas sp. 98S1 EU370416 98

Pseudomonas sp. 91S1 EU370417 98

B5, B6 Pseudomonas sp. BBTR25 DQ337603 99

Uncultured rumen bacterium clone BRC56 EF436342 99

B7 Uncultured bacterium 30BF17 AB330617 88

Lutispora thermophila AB186360 88

Thermoactinomyces dichotomicus AF138733 88

B8 Uncultured bacterium clone P1fT EF551941 99

Uncultured bacterium clone A55_D21_H_B_G02 EF559064 97

Gracilibacter thermotolerans strain JW/YJL-S DQ117469 89

Uncultured Thermoanaerobacteriaceae bacterium AY684101 87

B9 Thermomonas haemolytica isolate TG15 AF508110 99

Thermomonas koreensis DQ154906 99

Thermomonas fusca AF508110 99

The numbers show the bands whose sequences were identified with the BLAST program in the National Center for Biotechnology Information

(NCBI) database

BA denotes the domain Bacteria

Fig. 2 NMDS map of bacterial communities analyzed from DGGE

profiles (filled circles, designated as T1–8, indicate trials of response

surface analysis in Fig. 1; open circle, designated as abbreviated

letters, represent microbial species shown in Table 3; Gt, Gracillib-acter thermotolerans; Se, Swine effluent bacterium; Ps, Pseudomonassp.; Un, uncultured bacterium; Bh, Bacillus halodurans; Th, Ther-momonas haemolytica)

796 Bioprocess Biosyst Eng (2013) 36:791–798

123

B. halodurans grows thermophilically in the presence of

Na? and NH3, and can use polysaccharide as an energy

source, catalyzing the hydrolysis with b-galactosidase [26];

therefore, the community structure of this species seemed

to be affected more by HRT at 40 �C than at 50 or 60 �C.

Also, Sedimentibacter sp. could be considered to be similar

to the case of B. halodurans.

In view of relationship between microbial groups and

environmental factors, intensities of B2 and B9 related to

B. halodurans and T. haemolytica, respectively, became

more intense as T increased (Fig. 1). T. haemolytica grows

at T & 50 �C and can ferment carbohydrates to acetate

and propionate [31]; consequently, the increase in band

intensities of these two bacterial species was assumed to be

due to their thermophilicity. In contrast, B5 and B6, des-

ignated as Pseudomonas sp. BBTR25, faded as T and HRT

increased (Fig. 1), because Pseudomonas sp. BBTR25 is

mesophilic, and its maximum growth rate was 0.2 h-1

which was faster than other acidogenic bacteria. Therefore,

its band intensity was greater at short-HRT than at long-

HRT because this species has a competitive advantage over

other microbial groups at low HRT [11, 34]. B8, designated

as uncultured bacterium clone P1fT, was only detected at

40 �C; thus this clone must be obligatorily mesophilic.

Taken together, the community structure of most bac-

teria detected in each trial of RSA was affected more by

change in T than by change in HRT. Therefore, we infer

that TVFA production during acidogenesis was affected by

the physicochemical properties of acidogens, such as their

specific growth rate or contact time with substrates, rather

than by shift of microbial diversity.

Conclusions

When three identical acidogenic digesters were operated in

continuous mode, RSA was conducted with two independent

variables, HRT and T, and one dependent variable, TVFA

production. According to ANOVA from response surface

methodology, TVFAs produced by acidogens in the acido-

genic digesters were affected more by HRT than by T.

However, based on qualitative analysis of microbial com-

munity structure using NMDS, the shift of bacterial com-

munity structure was affected more by T than by HRT. These

results will be useful in optimizing conditions for production

of VFAs during acidogenesis in anaerobic digesters.

Acknowledgments This work was supported by the Advanced

Biomass R&D Center (ABC) of Global Frontier Project funded by

the Ministry of Education, Science and Technology (ABC-2010-

0029728), and was also the New & Renewable Energy of the Korea

Institute of Energy Technology Evaluation and Planning(KETEP)

grant funded by the Korea government Ministry of Knowledge

Economy (Grant no. 20103020090050).

References

1. Beaudet R, Gagnon C, Bisaillon JG, Ishaque M (1990) Micro-

biological aspects of aerobic thermophilic treatment of swine

waste. Appl Environ Microbiol 56:971–976

2. Speece RE (1996) Anaerobic Biotechnology for Industrial

Wastewaters. Archae Press, Nashville

3. Ueno Y, Haruta S, Ishii M, Igarashi Y (2001) Changes in product

formation and bacterial community by dilution rate on carbohy-

drate fermentation by methanogenic microflora in continuous

flow stirred tank reactor. Appl Microbiol Biotechnol 57:65–73

4. Hwang K, Song M, Kim W, Kim N, Hwang S (2010) Effects of

prolonged starvation on methanogenic population dynamics in

anaerobic digestion of swine wastewater. Bioresource Technol

101(Suppl 1):S2–S6

5. Song M, Shin SG, Hwang S (2010) Methanogenic population

dynamics assessed by real-time quantitative PCR in sludge

granule in upflow anaerobic sludge blanket treating swine

wastewater. Bioresource Technol 101:S23–S28

6. Kim W, Lee S, Shin SG, Lee C, Hwang K, Hwang S (2010)

Methanogenic community shift in anaerobic batch digesters

treating swine wastewater. Water Res 44:4900–4907

7. Stronach SM, Rudd T, Lester JN (1986) Anaerobic digestion

processes in industrial wastewater treatment. Springer, New York

8. Aitken MD, Mullennix RW (1992) Another look at thermophilic

anaerobic digestion of wastewater sludge. Water Environ Res

64:915–919

9. Yilmaz T, Yuceer A, Basibuyuk M (2008) A comparison of the

performance of mesophilic and thermophilic anaerobic filters

treating papermill wastewater. Bioresource Technol 99:156–163

10. Curtis TP, Sloan WT (2004) Prokaryotic diversity and its limits:

microbial community structure in nature and implications for

microbial ecology. Curr Opin Microb 7:221–226

11. Kim W, Hwang K, Shin SG, Lee S, Hwang S (2010) Effect of

high temperature on bacterial community dynamics in anaerobic

acidogenesis using mesophilic sludge inoculum. Bioresource

Technol 101(Suppl 1):S17–S22

12. Curtis TP, Craine NG (1993) The comparison of the diversity of

activated sludge plants. Water Sci Technol 137:71–78

13. Hwang S, Lee Y, Yang K (2001) Maximization of acetic acid

production in partial acidogenesis of swine wastewater. Bio-

technol Bioeng 75:521–529

14. Yang K, Oh C, Hwang S (2004) Optimizing volatile fatty acid

production in partial acidogenesis of swine wastewater. Water Sci

Technol 50:169–176

15. Lee H, Song M, Yu Y, Hwang S (2003) Production of Ganoderma

lucidum mycelium using cheese whey as an alternative substrate:

response surface analysis and biokinetics. Biochem Eng J 15:93–99

16. Box GEP, Draper NR (1987) Empirical model-building and

response surfaces. Wiley, New York

17. Hwang SH, Hansen CL, Stevens DK (1992) Biokinetics of an

upflow anaerobic sludge blanket reactor treating whey permeate.

Bioresource Technol 41:223–230

18. Yu Y, Kim J, Hwang S (2006) Use of real-time PCR for group-

specific quantification of aceticlastic methanogens in anaerobic

processes: population dynamics and community structures. Bio-

technol Bioeng 93:424–433

19. Obata K, Segawa O, Yakabe M, Ishida Y, Kuroita T, Ikeda K,

Kawakami B, Kawamura Y, Yohda M, Matsunaga T, Tajima H

(2001) Development of a novel method for operating magnetic

particles, magtration technology, and its use for automating

nucleic acid purification. J Biosci Bioeng 91:500–503

20. Yu Y, Lee C, Hwang S (2005) Analysis of community structures

in anaerobic processes using a quantitative real-time PCR

method. Water Sci Technol 52:85–91

Bioprocess Biosyst Eng (2013) 36:791–798 797

123

21. Muyzer G, De Waal EC, Uitterlinden AG (1993) Profiling of

complex microbial populations by denaturing gradient gel elec-

trophoresis analysis of polymerase chain reaction-amplified genes

coding for 16S rRNA. Appl Environ Microbiol 59:695–700

22. Kowalchuk GA, de Bruijn FJ, Head IM, Antoon D, Akkermans L,

Van Elsas JD (2004) Molecular microbial ecology manual.

Academic Publishers, Dordrecht

23. Roy CS, Talbot G, Topp E, Beaulieu C, Palin MF, Masse DI

(2009) Bacterial community dynamics in an anaerobic plug-flow

type bioreactor treating swine manure. Water Res 43:21–32

24. Fromin N, Hamelin J, Tarnawski S, Roesti D, Jourdain-Miserez

K, Forestier N, Teyssier-Cuvelle S, Gillet F, Aragno M, Rossi P

(2002) Statistical analysis of denaturing gel electrophoresis

(DGE) fingerprinting patterns. Environ Microbiol 4:634–643

25. APHA-AWWA-WEF (2005) Standard Methods for the exami-

nation of water and wastewater, 21st Ed., American Public Health

Association, Washinton, DC

26. Takami H, Horikoshi K (1999) Reidentification of facultatively

alkaliphilic Bacillus sp. C-125 to Bacillus halodurans. Biosci

Biotech Biochem 63:943–945

27. Patil S, Kumar M, Ball A (2010) Microbial community dynamics

in anaerobic bioreactors and algal tanks treating piggery waste-

water. Appl Microbiol Biotechnol 87:353–363

28. Szekely A, Sipos R, Berta B, Vajna B, Hajdu C, Marialigeti K

(2009) DGGE and T-RFLP analysis of bacterial succession dur-

ing mushroom compost production and sequence-aided T-RFLP

profile of mature compost. Microb Ecol 57:522–533

29. Shiratori H, Ohiwa H, Ikeno H, Ayame S, Kataoka N, Miya A,

Beppu T, Ueda K (2008) Lutispora thermophila gen. nov., sp.

nov., a thermophilic, spore-forming bacterium isolated from a

thermophilic methanogenic bioreactor digesting municipal solid

wastes. Int J Syst Evol Microbiol 58:964–969

30. Alves MP, Rainey FA, Nobre MF, da Costa MS (2003) Ther-momonas hydrothermalis sp. nov., a new slightly thermophilic

gamma-proteobacterium isolated from a hot spring in central

Portugal. Syst Appl Microbiol 26:70–75

31. Busse HJ, Kampfer P, Moore ERB, Nuutinen J, Tsitko IV,

Denner EBM, Vauterin L, Valens M, Rossello-Mora R, Sal-

kinoja-Salonen MS (2002) Thermomonas haemolytica gen. nov.,

sp. nov., a c-proteobacterium from kaolin slurry. Int J Syst Evol

Micr 52:473–483

32. Lee YJ, Romanek CS, Mills GL, Davis RC, Whitman WB,

Wiegel J (2006) Gracilibacter thermotolerans gen. nov., sp. nov.,

an anaerobic, thermotolerant bacterium from a constructed wetland

receiving acid sulfate water. Int J Syst Evol Micr 56:2089–2093

33. Lee C, Kim J, Hwang K, O’Flaherty V, Hwang S (2009) Quan-

titative analysis of methanogenic community dynamics in three

anaerobic batch digesters treating different wastewaters. Water

Res 43:157–165

34. Gummadi S, Santhosh D (2010) Kinetics of growth and caffeine

demethylase production of Pseudomonas sp. in bioreactor. J Ind

Microbiol Biotechnol 37:901–908

798 Bioprocess Biosyst Eng (2013) 36:791–798

123