b i om a s s an d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 1

Avai lab le at www.sc iencedi rect .com

ht tp : / /www.e lsev ier . com/ loca te /b iombioe

The buffer composition impacts the hydrogen production andthe microbial community composition in non-axenic cultures

Gustavo Davila-Vazquez a,1, Antonio de Leon-Rodrıguez b, Felipe Alatriste-Mondragon a,Elıas Razo-Flores a,*aDivision de Ciencias Ambientales, Instituto Potosino de Investigacion Cientıfica y Tecnologica, Camino a la Presa San Jose 2055,

Lomas 4a seccion, C.P. 78216, San Luis Potosı, S.L.P, MexicobDivision de Biologıa Molecular, Instituto Potosino de Investigacion Cientıfica y Tecnologica, Camino a la Presa San Jose 2055,

Lomas 4a seccion, C.P. 78216, San Luis Potosı, S.L.P, Mexico

a r t i c l e i n f o

Article history:

Received 23 July 2010

Received in revised form

16 April 2011

Accepted 21 April 2011

Available online 17 May 2011

Keywords:

Biohydrogen

Buffer

Cheese whey

Clostridium

Dark fermentation

Lactose

* Corresponding author. Tel.: þ52 444 834 20E-mail addresses: gdv76@prodigy.net.mx

1 Present address: Unidad de Tecnologıa Am(CIATEJ), Av. Normalistas 800, Colinas de la0961-9534/$ e see front matter ª 2011 Elsevdoi:10.1016/j.biombioe.2011.04.046

a b s t r a c t

Hydrogen obtained from biomass via dark fermentation is considered a sustainable and

clean energy carrier. Batch fermentations with cheese whey powder were performed to

assess total hydrogen production (Hmax), volumetric hydrogen production rate (VHPR),

maximum lactose consumption (Smax), maximum lactose consumption rate (Rmax,S),

hydrogen molar yield (HMY) and the bacterial species present using two mineral

media formulation (A, B). The highest VHPR was 304.8 cm3 dm�3 h�1 and the HMY was

1.8 mol mol�1. Medium B yielded around twice the VHPR than the attained with medium A,

but HMY only had a slight increment with the use of medium B. The values reached for

Smax (17.3 g dm�3), Hmax (4.863 dm3) and Rmax,S (2.7 g dm�3 h�1) were also enhanced with

medium B. Results suggest that butyrate levels and lower pH are the reasons for dimin-

ished hydrogen production with medium A. The microbial communities were analyzed

using polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE). Only

one band was observed in the experiments with medium A, the sequence retrieved from

this band presented a closest relative match to the sequence from Citrobacter freundii JCM

(100% identity); whereas for medium B, three bands were detected. Sequences from these

bands presented high homology to sequences from Clostridium perfringens W11 (95% iden-

tity), uncultured Lachnospiraceae bacterium clone MS146A1 E12 (100% identity) and Enter-

obacter cloacae GH1 (100% identity). From the results obtained it is clear that the formulation

of culture media had a strong effect on hydrogen production, kinetics and also on the

microbial diversity.

ª 2011 Elsevier Ltd. All rights reserved.

1. Introduction biological processes commonly used to treat wastewaters are

A large effort to find both renewable and sustainable energy

sources is being undertaken around the world. Anaerobic

00x2026; fax: þ52 444 834(G. Davila-Vazquez), erazbiental, Centro de Inves

Normal, C.P. 44270, Guadier Ltd. All rights reserve

able to generate sustainable fuels such as methane or

hydrogen, and rather than aerobic processes, are considered

to be energy producer systems [1]. Special attention has been

2010.o@ipicyt.edu.mx (E. Razo-Flores).tigacion y Asistencia en Tecnologıa y Diseno del Estado de Jaliscoalajara, Jal. Mexico.d.

Table 1 eMineral medium composition A (modified from[19]) and B (modified from [12]).

Compound Medium A,carbonate-basedbuffer (mg dm�3)

Medium B,phosphate-basedbuffer (mg dm�3)

NH4HCO3 2500

KH2PO4 1250

MgSO4,7H2O 125

NaCl 12.5

Na2MoO4,2H2O 12.5 12.5

CaCl2,2H2O 12.5

MnSO4,7H2O 18.75 15

FeCl2 3.48

CoCl2,8H2O 3 3

NiCl2,6H2O 1.5

ZnCl2 0.75 75

NH4H2PO4 4500

Na2HPO4 11,867

K2HPO4 125

MgCl2,6H2O 100

FeSO4,6H2O 25

CuSO4,5H2O 5

b i om a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 1 3175

paid to biohydrogen production from organic wastes such as

wastewaters and organic residues or by-products [2e5].

Regarding biohydrogen production, it is known that the by-

products from the dairy industry represent a potential

source of substrates for energy generation. Among them,

cheese whey (CW), a lactose-rich by-product, which accounts

for the 85% of the total volume of processed milk [6] and

cheese whey powder (CWP) have been tested.

According to Eq. (1), the maximum theoretical hydrogen

molar yield (HMY) is 8 mol H2 per mol of lactose consumed,

when acetate is the main metabolite produced, whereas HMY

of 4 mol mol�1 is typically obtained when the butyric fermen-

tation occurs Eq. (2), [7]. However, the HMY based on lactose

could be lower than 4 mol mol�1 if other metabolites such as

lactic acid, ethanol, butanol or acetone are also formed [8].

C12H22O11 þ 5H2O/4CH3COOHþ 4CO2 þ 8H2 (1)

C12H22O11 þH2O/2CH3CH2CH2COOHþ 4CO2 þ 4H2 (2)

Biohydrogen generation by dark fermentation is highly

dependent on the process conditions such as temperature,

pH, mineral medium formulation, kind of inoculum, profile of

organic acids produced, type of substrate and concentration,

hydrogen partial pressure, and reactor configuration [9,10].

Particularly, pH is a key parameter in biological processes,

because it affects the enzyme activities, the metabolite

transport, the microbial community in non-axenic cultures,

among others [11,12]; therefore the production media formu-

lation must include buffering compounds to reduce pH vari-

ations during cultivations. There are reports about

biohydrogen production using CW, and in all cases the use of

a carbonated compound in the buffer mineral media such

as sodium or ammonium bicarbonate prevails [13e18],

presumably because carbonated-media were extensively and

successfully used in anaerobic digestion processes (meth-

anogenesis). Unlike methanogenic process where organic

acids are consumed, in fermentative hydrogen production

besidesH2 and CO2 production, there is an intrinsic generation

of organic acids mainly acetic, propionic, lactic and butyric.

These acids could be toxic to cells, but they also react with the

bicarbonate, generating additional dissolved CO2, and the

buffer capacity of the medium is diminished (Eq. (3)).

HCO�3 þHþ%H2CO3%H2Oþ CO2 (3)

As pH is an important parameter influencing the efficiency

and productivity of fermentative biohydrogen process, the use

of carbonate-buffered media is being reconsidered [11,12]. In

this work phosphate salts were considered as the alternative

buffering supplement due to their wide application even at

full-scale level processes such as in wastewater treatment

plants [11].

Thus, the aim of this study was to assess the effect of two

different mineral media composition on the important

parameters in fermentative hydrogen production such as

hydrogen yield, volumetric production and pH profiles due to

the change in metabolites’ composition. Besides, the effect on

the microbial communities was also analyzed using poly-

merase chain reaction-denaturing gradient gel electropho-

resis (PCR-DGGE).

2. Materials and methods

2.1. Biohydrogen production experiments

Cheese whey powder (CWP, Land O’Lakes Inc., Minnesota,

USA) was used as a source of lactose (770 g kg�1), and protein

(110 g kg�1). All chemicals were purchased as reagent grade

from SigmaeAldrich. Anaerobic granular sludge from a full-

scale up-flow anaerobic sludge blanket reactor (UASB) was

used as inoculum for biohydrogen production. The UASB

reactor treats wastewater from a candy factory in San Luis

Potosı, Mexico. The granular sludge was washed with tap

water and then boiled for 40 min to inactivate methanogenic

microorganisms. Batch experiments were conducted in

duplicate using 120 cm3 vials with a working volume of

80 cm3. Two different mineral media were used: medium A,

modified from [19] and medium B, modified from [12]. The

mineral media composition is shown in Table 1. CWP at

25 g dm�3 was used as substrate and the inoculum concen-

tration as volatile suspended solids (VSS) was 4.5 g dm�3. An

initial pH of 7.5 was selected and vials were set as previously

described [17]. All experiments were performed without pH

control in order to follow the pHdrop due to volatile fatty acids

(VFA) production with each medium.

One scale-up batch experiment for each medium was

carried out in a 3 dm3 (2.4 dm3 working volume) stirred

bioreactor equipped with an ADI 1030 system controller, ADI

1035 bioconsole and BioXpert 1.3 data-acquisition software

(Applikon Biotechnology, Schiedam, TheNetherlands). Mixing

at 4.16 Hz was performed with two Rushton turbines and pH

was monitored online using an autocleavable pH electrode

(AppliSens, Applikon, Schiedam, The Netherlands). Temper-

ature was kept at 37 �C using an electric jacket. Initial CWP

concentration and pH were 25 g dm�3 and 7.5, respectively.

Gas production and composition in the headspace were

measured periodically as described in analytical methods.

b i om a s s an d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 13176

2.2. Analytical methods

Gas production was measured using two different liquid-

displacement devices filled with acidified water (pH ¼ 2). For

the 120 cm3 vials, an inverted burette (250 cm3) modified to

have a gas-sampling port in the top was used, while in the

case of the 3 dm3 bioreactor, a manometer (SEV, Puebla,

Mexico) calibrated to periodically count a fixed volume of gas

was employed. Cumulative hydrogen production was calcu-

lated for each vial considering the headspace composition and

the volume of gas released at each time interval, using a mass

balance equation [4,17,20]. Standard condition for tempera-

ture and pressure were considered to report gas volumes (0 �Cand 101.325 kPa).

H2 and CO2 were quantified using a gas chromatograph

6890 N (Agilent Technologies,Waldbronn, Germany) equipped

with a thermal conductivity detector (Agilent Technologies,

Waldbronn, Germany). The column used was a Hayesep D

(Alltech, Deerfield, Illinois, USA). Temperatures of the injec-

tion port, oven and the detector were 250, 60 and 250 �C,respectively. Nitrogen was used as carrier gas with a flow-rate

of 12 cm3 min�1.

Liquid samples were withdrawn at the end of each experi-

ment (for the 120 cm3 vials) and periodically for the 3 dm3

reactor as follows: 10 cm3 of liquid samples were taken and

0.06 cm3 of HgCl2 (16 g dm�3) were added before centrifugation

at 6610g for 15 min to minimize microorganisms activity [21].

Both, remaining substrate andproducts, suchas formic, acetic,

propionic, and butyric acids (VFA) were analyzed in the filtrate

by capillary electrophoresis in the samerun [17].Analyteswere

quantified by comparison with high purity standards. For this

purpose a capillary electrophoresis system (Agilent 1600A,

Waldbronn, Germany) was used with a basic anion buffer

(Agilent, pH¼ 12.1) anda fused silica capillary column (Agilent,

id¼ 50 mm, L¼ 80.5 cm, effective length¼ 72 cm). Temperature

and voltage were 20 �C and �30 kV, respectively. The samples

were injected with a pressure of 30 kPa for 6 s. Detection was

carried out with indirect UV detection using a diode-array

detector. The signal wavelength was set at 350 nm with

a reference at 230nm.Abuffer flush for 4min at 1 barwas done

prior to each run. Solvents such as acetone, ethanol, propanol

andbutanolwereanalyzedby injectinga1mm3sample inagas

chromatograph 6890 N equipped with an auto-sampler 7863

(Agilent, Wilmington, USA) and a capillary column HP-

Innowax (30 m � 0.25 mm i.d. � 0.25 mm film thickness; Agi-

lent, Wilmington, USA). Helium was used as carrier gas at

a flow rate of 1.5 cm3min�1. Temperatures for the injector and

flame ionization detector (FID) were 220 and 250 �C, respec-tively. The solvents’ analyseswere performedwith a split ratio

of 1:0.1 anda temperatureprogramof 35 �C for 2min, increased

to 80 �C (10 �Cmin�1), andwasmaintained at this temperature

to a final time of 15 min. VSS were analyzed according to the

Standard Methods [22].

2.3. Data analysis

Once cumulative hydrogen production was calculated from

experimental data, amodified Gompertz equationwas used to

fit the kinetics of biohydrogen production and to obtain the

parameters Hmax, Rmax and l using KaleidaGraph ver. 4.0

(Synergy software). This equation has been widely used to

model gas production data [23e26]:

HðtÞ ¼ Hmax � exp�� exp

�2:71828 � Rmax

Hmaxðl� tÞ þ 1

��(4)

where H(t) in cm3 is the total amount of hydrogen produced at

culture time t (h); Hmax (cm3) is the maximal amount of

hydrogen produced; Rmax (cm3 h�1) is the maximumhydrogen

production rate; l (h) is the lag time before exponential

hydrogen production. HMY and VHPR were defined as

response variables; HMY was calculated from Hmax and

defined as mol H2 (mol consumed substrate)�1 whereas VHPR

was obtained from Rmax standardized to the working volume

(cm3 H2 dm�3 h�1).

For the experiments in the 3 dm3 reactor, a modeling of the

substrate consumption was performed according to a modi-

fied Gompertz model Eq. (5), [24]):

S0 � S ¼ Smax � exp�� exp

�2:71828 � Rmax;S

Smaxðl� tÞ þ 1

��(5)

where S0 is the initial lactose concentration (g dm�3); S is the

lactose concentration (g dm�3) at fermentation time t (h); Smax

is maximum lactose consumption (g dm�3); Rmax,S is the

maximum lactose consumption rate (g dm�3 h�1); l is the lag

time before exponential substrate consumption (h).

2.4. Microbial community analyses by 16S rRNA genesusing DGGE

2.4.1. DNA extractionTen-milliliter samples were withdrawn from the 120 cm3 vials

at the end of the exponential hydrogen production phase and

were stored at �20 �C in glycerol (150 dm3 m�3) until analysis.

DNA extraction protocols reported elsewhere [27,28] were

optimized to be used on granular sludge samples. In brief, the

samples were slowly thawed and 0.5 cm3 were taken and

centrifuged at 7000 g for 10 min. The pellet was washed two-

times with PBS buffer (10 mmol dm�3, pH 7.5), resuspended

in extraction buffer (10 mmol dm�3 Tris/HCl, pH 7.5,

50 mmol dm�3 EDTA, 0.5 mol dm�3 NaCl) and sonicated for

5 min in an ultrasonic processor (Sonics & Materials, New-

town, USA). Afterward, the mixture was incubated with

0.01 cm3 of RNase (20 mg cm�3, Invitrogen, Carlsbad, USA),

0.02 cm3 of lysozyme (20 mg cm�3, USB, Cleveland, Ohio, USA)

and 0.015 cm3 of proteinase K (20 mg cm�3, Invitrogen,

Germany) during 2 h at 37 �C and 5.83 Hz. 0.1 cm3 of SDS

(100 g kg�1) and 0.2 cm3 of sodium acetate (5 mol dm�3, pH 8)

were added to the mixture and incubated for 10 min at 60 �C.For the purification of the nucleic acids present in themixture,

one volume of chloroform-isoamylic alcohol (24:1) was added,

mixed and centrifuged during 10 min at 7000g. This purifica-

tion step was repeated two times with the liquid supernatant.

To precipitate nucleic acids, one volume of isopropanol

at �20 �C was added to the supernatant and incubated for 3 h

at �20 �C before centrifugation at 4 �C for 20 min at 7000g. The

pellet was first washed with absolute ethanol and centrifuged

at 7000g for 1 min, and then washed again with diluted

ethanol (700 cm3 dm�3) and centrifuged with the same

conditions. DNA pellet was dried at room temperature and

b i om a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 1 3177

resuspended in 0.05 cm3 of sterile deionized water. A DNA

integrity analysis was performed in 10 g dm3 agarose gels,

stained with ethidium bromide.

2.4.2. PCR amplificationAmplification of the hypervariable 3 region of the 16S rRNA

gene from the purified nucleic acids preparations was carried

out by PCR using Pfu polymerase (Biotools, Spain). The PCR

primers used were the forward primer C356F (50-CGCCCGCCGCGCCCCGCGCCCGTCCCGCCGCCCCCGCCCCCCTACGG-

GAGGCAGCAG-30) and the reverse primer 517R (50-ATTAC-CGCGGCTGCTGG-30) described elsewhere by Muyzer et al. [29]

for bacterial identification of complex microbial population.

The primer C356F contains the GC clamp shown with bold

letters led to carry out the DGGE. Reaction conditions were as

follows: initial DNA denaturation at 95 �C for 1min using 1.5 U

of Pfu polymerase (Biotools, Madrid, Spain), followed by 10

cycles of denaturation at 95 �C for 30 s and annealing from 65

to 60 �C for 30 s, lowering the temperature 0.5 �C each cycle,

and followed by an extension at 72 �C for 1min. In addition, 20

cycles at 95 �C for 30 s, 60 �C for 30 s and 72 �C for 1 min; with

a final extension at 72 �C for 7 min were performed. The PCR

product was loaded onto a 15 g dm3 agarose gel and stained

with ethidium bromide to assess the size, purity and

concentration of DNA.

2.4.3. DGGE analysisDGGE was performed with DCode Universal Mutation Detec-

tion System (Biorad, Hercules, California, USA). The PCR

products were loaded onto 100 g dm�3 polyacrylamide gels in

1 � TAE buffer (20 mmol dm�3 Tris-acetate, 10 mmol dm�3

sodium acetate, 0.5 mmol dm�3 EDTA, pH 7.4) with a dena-

turing gradient (urea-formamide) that ranged from 300 to

600 g dm�3. Electrophoresis was carried out at 60 �C at

a constant voltage of 39 V during 14 h. After electrophoresis

the gel was stained using SYBRª Safe for 30 min (Molecular

Probes Inc., Eugene, Oregon, USA) before being visualized on

a UV transilluminator (Biorad, Hercules, California, USA). The

dominant bands were excised from the gel, eluted in

10mmol dm�3 Tris-EDTA buffer (pH 7.5) overnight at 4 �C. The

Table 2 e Kinetic parameters and performance of batch experi(modified from [12]).

12

Medium Aa

Hmax (cm3) 73.1

VHPR (cm3 dm�3 h�1) 94.5

Smax (g dm�3) e

Rmax,S (g dm�3 h�1) e

HMY (mol mol�1) 1.1

Final pH 4.8

Acetic acid (mg dm�3) 1523

Propionic acid (mg dm�3) ND

Butyric acid (mg dm�3) 1945

Ethanol (mg dm�3) 1036

Lactose consumption (%) 63.7

ND ¼ Not detected.

a Means from replicated experiments (n ¼ 2).

eluted DNA was reamplified by PCR with the conditions

mentioned before (Section 2.4.2). The PCR products from

reamplification were sent to purification and sequencing to

Molecular Cloning Laboratories (MCLAB, San Francisco, Cal-

ifornia, USA). Sequence data were analyzed with Bioedit v

7.0.9 software (Ibis Bioscience, Carlsbad CA, USA) and

submitted to the non-redundant nucleotide data base at

GenBank using the BLAST program (http://www.ncbi.nlm.nih.

gov/blast/) and Ribosomal Database Project (http://rdp.cme.

msu.edu/index.jsp) for bacterial identification.

3. Results and discussion

3.1. Biohydrogen production in the 120 cm3 vials

Batch fermentations carried out in 120 cm3 serum vials were

conducted for biohydrogen production. After 44 h of fermen-

tation, there was a clear difference in biohydrogen production

experiments using the two media (Table 2). One considerable

difference in the experiments was the lactose consumption

rate with eachmedium; lactose was not detected at the end of

the experiments with medium B, while in contrast only 63.7%

of initial lactose was consumed using medium A. This

significant difference is reflected in the higher maximum

hydrogen production achieved (Hmax: 170.7 cm3) using the

medium B, which yielded around twice the volume produced

withmediumA (73.1 cm3). Also the production rate (Rmax) was

enhanced with the use of medium B (Table 2).

Due to the lost in buffer capacity for medium A from the

reaction with acids (Eq. (3)), there was a difference of one unit

in the final pH of the medium, being higher with the use of

medium B. Consequently, due to the highest hydrogen

production with medium B, total VFA concentration was

7.427 g dm�3, while the use of medium A resulted in a total

VFA concentration of 3.468 g dm�3. Ethanol was the only

solvent detected at around 2.4 g dm�3 with the use of medium

B, and near half the concentration withmediumA. In terms of

the profile of VFA, these results are similar to those obtained

by Yang et al. [18]. Using a carbonate-based buffer, mixed

ments with medium A (modified from [19]) and medium B

0 cm3 vials 3 dm3 fermenter

Medium Ba Medium A Medium B

170.7 3005 4863

163.8 179.8 304.8

e 12.1 17.3

e 1 2.7

1.7 1.5 1.8

5.8 4.9 5.4

4159 1097 2287

740 2174 3416

2528 1814 5048

2409 44 82

>99 65 >99

Fig. 1 e Biohydrogen production in the 3 dm3 fermenter.

Cumulative H2 production using medium A (�) and

medium B (C). The profile of online pH for medium A (>),

and B (A) is also shown.

b i om a s s an d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 13178

microflora and CWP as substrate the authors found mainly

acetic and butyric acids with propionic acid present at low

substrate-to-microorganism ratio. Moreover, the presence of

residual lactose was detected in their experiments with

uncontrolled pH that fell from 7.33 to as low as 4.48. One

difference is that the authors found amaximumVHPR of near

400 cm3 dm�3 h�1 while in this work we report a maximum of

94.5 cm3 dm�3 h�1.

To our knowledge, there is only one report regarding non-

continuous lactose or CW fermentation for biohydrogen

production using mixed cultures and phosphate-based

medium [7]. The authors used pure lactose, as well as

xylose, at thermophilic conditions at 55 �C in both controlled

and uncontrolled pH experiments in fed-batch operation for

each substrate. For lactose, they found the best results on pH

Fig. 2 e Lactose concentration (-), and volatile fatty acids produ

A (a) and medium B (b) in fermentations for biohydrogen produ

controlled at 5.3 with HMY reaching 3.45 mol mol�1 and VHPR

was 57 cm3 dm�3 h�1. In their uncontrolled pH experiments,

VHPR were below 9 cm3 dm�3 h�1 and lactose consumption

was 62% with an average HMY of 1.5 mol mol�1; the pH

dropped from an initial value of 6.0 to 4.2. In all cases the

initial lactose concentration was 2 g dm�3. The lactose

consumption and HMY obtained by the authors are similar to

the reported here for uncontrolled pH experiments with

carbonate-based medium in which pH fell from 7.5 to 4.8.

However, the VHPR we report for the mesophilic experiments

are at least 10 times higher than the best result reported by

Calli et al. [7] in thermophilic controlled pH experiments. One

reason for this is due to both higher initial substrate concen-

tration (25 g dm�3) and initial pH of 7.5 used in this work, since

previous findings showed that both factors had a significant

effect on hydrogen production from either pure lactose or

CWP [17].

3.2. Biohydrogen production in the bioreactorexperiments

To further explore the behavior of pH, and follow the substrate

consumption, metabolites production and hydrogen genera-

tion, single scale-up experiments were performed for each

medium in a fermenter with a working volume of 2.4 dm3.

As noticed in Fig. 1 total hydrogen production was

4.863 dm3 with medium B while 3 dm3 of H2 were produced

using the medium A. The VHPR was also increased with the

use of medium B (Table 2). The increment in the working

volume from vials to fermenter by a factor of 30, was reflected

in the increase of Hmax being the increment factor of 41 for the

mediumA and 28.5 for themedium B. However, the VHPRwas

nearly doubled from the vials to the fermenter for both media

(Table 2).

Considering the lactose content (770 g kg�1) in the CWP, it

was expected to detect 19.25 g lactose dm�3 as initial

concentration in the batch experiments, however amaximum

initial lactose concentration of around 16.5 g dm�3 was

measured (Fig. 2a). A possible reason for this could be the

ction (,: acetic;>: propionic; �: butyric acid) with medium

ction with the 3 dm3 fermenter.

Fig. 3 e DGGE profile of partial 16S rRNA genes amplified

from the duplicated batch experiments (120 cm3 vials) run

with medium A (A) and medium B (B). The excised and

sequenced bands (1e4) correspond to the closest relative

matches: 1: Citrobacter freundii JCM (100% identity,

accession number AB548824.1); 2: uncultured

Lachnospiraceae bacterium clone MS146A1_E12 (100%

identity, accession number EF706759.1); 3: Clostridium

perfringens W11 (95% identity, accession number

JF499889.1) and 4: Enterobacter cloacae GH1 (100% identity,

accession number JF261136.1).

b i om a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 1 3179

adsorption of lactose to anaerobic biomass and/or to protein

present in CWP, and since the methodology includes the

precipitation of the solids, only dissolved lactose could be

detected. Therefore, all calculations were performed using

dissolved lactose found in the media.

Fig. 2a shows that using medium A, lactose was not

completely consumed by the microorganisms, and a residual

concentrationof around6gdm�3wasobserved.Whereas,with

the use of medium B (Fig. 2b), lactose was rapidly metabolized

and was not detected in the culture medium after 18 h of

fermentation. This difference could be explained due to the

different rates at which pH dropped for each medium (Fig. 1),

but also due to the concentration of undissociated acids at

certain fermentation times, because someVFA, such as acetic,

butyric and propionic acids, are toxic to cells or inhibitory for

the hydrogen production metabolism [30,31]. Both features

(low pH and concentration of undissociated VFA) are con-

nected because for a lower pH there is a higher amount of

undisocciated acid form according to the Hender-

soneHasselbalch equation [30]. Therefore, for medium A the

consumption of only 1.1 g lactose from fermentation time

17.3 he19.3 h, where consumption stopped, could be related to

the concentration of undissociated acids such as butyric acid

between those fermentation times. At time 17.3 h (Fig. 2a) the

undissociated concentration of butyric acid (pH 5.16) was

583mgdm�3 (6.6mmol dm�3) andhad a slight decrease at time

19.3 h. In the case of medium B, the concentration of butyric

acid at time 15 h was only 21mg dm�3 (0.24 mmol dm�3) at pH

5.88, this concentration is far below the reported threshold for

butyrate toxicity to cells (2e30mmoldm�3). It suggests that the

low butyric acid concentration in time 15 h allowed the

bacterial cells to further metabolize the residual 6 g lactose

dm�3 in 3 h, because from time 18 h lactose was not detected

(Fig. 2b). This was reflected in a higher maximum lactose

consumption (Smax) for medium B of 17.3 g dm�3 calculated by

the fitting of Eq. (5), although the real figure was 16.3 g dm�3 as

shown as initial lactose concentration in Fig. 2b. The low

butyrate concentration with the use of medium B during the

first 15hof fermentationcouldbe the reasonof ahigher lactose

consumption rate (Rmax,S; Table 2) that was near three times

the calculated for the experiment with medium A. Besides

butyric acid concentration, the faster drop in pH formediumA

could have a negative effect in cells and also inhibit hydrogen

production metabolism. Fig. 1 shows that the beginning of

exponential hydrogen production with medium A occurs at

time 12.3 h, and the pH had already dropped to 5.45, while for

the same fermentation time the pH formediumBwas 6.75. It is

reported that the optimum pH for fermentative hydrogen

production for different substrates is from 5.5e6 [19,25], but

hydrogen production has also been reported for higher pH

[12,32]. However, pH below 5.0 has not been reported as

optimum because it is around the pH where the switch from

acidogenesis to alcohol production triggers [33]. Thus, the

lower pH at the beginning of the exponential phase using

medium A, could also explain the lower biohydrogen produc-

tion compared to medium B, because for the latter there was

a bigger period of time before reaching a harsh low pH. More-

over, for medium B, pH fell to 5.4 which still is near the

optimum reported range for hydrogen production, while for

medium A the pH fell to below 5. It is hypothesized that these

different rates of pH drop could also have an effect over the

microbial communities that better adapt for each pH change.

3.3. Microbial community analyses

The effect of the A and B culture media on the microbial

communities was studied through the PCR-DGGE technique in

the 120 cm3 vials. Fig. 3 shows that one clear band (1) was

present in the experiment with medium A, while three bands

(2, 3 and 4) were observed when medium B was used in batch

fermentations. It is important to remark that the results from

these duplicated analyses are shown in Fig. 3. For the exper-

iment with medium A the DGGE band profile was exactly the

same, while for experiment with medium B only one band (3)

was predominant in the duplicate (Fig. 3). The band detected

in the experiments with medium A, had a closest relative

match to Citrobacter freundii JCM (100% identity, accession

number AB548824.1); while for medium B, three bands were

observed (Table 3). Sequences from these bands presented

high homology to sequences from Clostridium perfringens W11

Table 3 e Microorganisms showing the highest percentage of identity in the output result from analysis in the non-redundant nucleotide database from NCBI using the BLAST program.

Band Medium Closest relative Accession number Percentage of identitya E-valueb

1 A Citrobacter freundii JCM AB548824.1 100% 1e-75

2 B Uncultured Lachnospiraceae bacterium

MS146A1 E12

EF706759 100% 3e-82

3 B Clostridium perfringens W11 JF499889.1 95% 7e-73

4 B Enterobacter cloacae GH1 JF261136.1 100% 2e-77

a % of the sequence that is invariant to the closest BLAST relative.

b The number of different alignments with scores equivalent to or better than S that are expected to occur in a database search by chance. The

lower the value, the more significant the score.

b i om a s s an d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 13180

(95% identity, accession number JF499889.1), uncultured

Lachnospiraceae bacterium clone MS146A1 E12 (100% identity,

accession number EF706759) and Enterobacter cloacae GH1

(100% identity, accession number JF261136.1) (Table 3).

Although species of Clostridium, Enterobacter and Citrobacter

genera are known as efficient hydrogen-producing microor-

ganisms [34e45], the results presented here suggest that,

under the culture conditions assessed, the selection of both

Clostridium and Enterobacter genera might be related to best

results obtainedwithmedium B. These findings are supported

by reports in which bacterial populations from a mixed

culture were pH- and substrate-dependent [46,47]. Lin and

Hung performed batch cultivations using cow dung sludge as

inoculum and found Klebsiella, Pseudomonas, Clostridium and

Streptococcus species at different initial pH (5.5e9) with xylose

and cellulose as substrates [46]. While Xiao et al., showed that

even with the same inocula, the use of different substrates

and also the change in pH during hydrogen production in

batch tests would lead to the development of different

microorganisms under a transient biohydrogen production

pH environment, with glucose and protein as substrates [47].

4. Conclusions

From the results showed in this work it is clear that the

utilization of two different mineral media formulation, with

CWP as substrate, had a strong effect on biohydrogen

production. Differences were observed regarding the micro-

organisms present in each culture condition, consequently,

deviations from metabolic hydrogen-production pathways

occurred. Moreover, the kinetics had also an effect on the

microbial community developed with two media and there-

fore these could be the reasons of higher hydrogen production

with the use of medium B. These findings are significant

because the improvement of the hydrogen production is

critical for the scaling-up of bioenergy production processes.

Acknowledgments

This work was supported by the Fondo Mixto San Luis Potosı -

ConsejoNacional deCiencia yTecnologıa, project FMSLP-2005-

C01-23. The authors acknowledge the technical assistance of

Erika Nahomy Marino-Marmolejo, Leandro G. Ordonez-Ace-

vedo, Dulce Partida Gutierrez and GuillermoVidriales Escobar.

The use of the infrastructure of the “Laboratorio Nacional de

Biotecnologıa Agrıcola, Medica y Ambiental (LANBAMA)” is

also acknowledged.

r e f e r e n c e s

[1] Calli B, Zhao J, Nijssen E, Vanbroekhoven K. Significance ofacetogenic H2 consumption in dark fermentation andeffectiveness of pH. Water Sci Technol 2008;57:809e14.

[2] Kapdan IK, Kargi F. Bio-hydrogen production from wastematerials. Enzym Microb Tech 2006;38:569e82.

[3] Venkata Mohan S. Harnessing of biohydrogen fromwastewater treatment using mixed fermentative consortia:process evaluation towards optimization. Int J HydrogenEnerg 2009;34:7460e74.

[4] Van Ginkel SW, Oh SE, Logan BE. Biohydrogen gas productionfrom food processing and domestic wastewaters. Int JHydrogen Energ 2005;30:1535e42.

[5] Ntaikou I, Antonopoulou G, Lyberatos G. Biohydrogenproduction from biomass and wastes via dark fermentation:a review. Waste Biomass Valor 2010;1:21e39.

[6] De Leon-Rodrıguez A, Rivera-Pastrana D, Medina-Rivero E,Flores-Flores JL, Estrada-Baltazar A, Ordonez-Acevedo LG,et al. Production of penicillin acylase by a recombinantEscherichia coli using cheese whey as substrate and inducer.Biomol Eng 2006;23:299e305.

[7] Calli B, Schoenmaekers K, Vanbroekhoven K, Diels L. Darkfermentative H2 production from xylose and lactose-Effectsof on-line pH control. Int J Hydrogen Energ 2008;33:522e30.

[8] Levin DB, Pitt L, Love M. Biohydrogen production: prospectsand limitations to practical application. Int J Hydrogen Energ2004;29:173e85.

[9] Davila-Vazquez G, Arriaga S, Alatriste-Mondragon F, deLeon-Rodrıguez A, Rosales-Colunga LM, Razo-Flores E.Fermentative biohydrogen production: trends andperspectives. Rev Environ Sci Biotechnol 2008;7:27e45.

[10] Hallenbeck PC. Fermentative hydrogen production:principles, progress and prognosis. Int J Hydrogen Energ2009;34:7379e89.

[11] Lin CY, Lay CH. Effects of carbonate and phosphateconcentrations on hydrogen production using anaerobicsewage sludgemicroflora. Int JHydrogenEnerg 2004;29:275e81.

[12] Wang C-H, Lin P-J, Chang J-S. Fermentative conversion ofsucrose and pineapple waste into hydrogen gas inphosphate-buffered culture seeded with municipal sewagesludge. Process Biochem 2006;41:1353e8.

[13] Ferchichi M, Crabbe E, Gil GH, Hintz W, Almadidy A.Influence of initial pH on hydrogen production from cheesewhey. J Biotechnol 2005;120:402e9.

b i om a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 3 1 7 4e3 1 8 1 3181

[14] Venetsaneas N, Antonopoulou G, Stamatelatou K,Kornaros M, Lyberatos G. Using cheese whey for hydrogenand methane generation in a two-stage continuous processwith alternative pH controlling approaches. BioresourTechnol 2009;100:3713e7.

[15] Castello E, Garcıa y Santos C, Iglesias T, Paolino G, Wenzel J,Borzacconi L, et al. Feasibility of biohydrogen productionfrom cheese whey using a UASB reactor: links betweenmicrobial community and reactor performance. Int JHydrogen Energ 2009;34:5674e82.

[16] Azbar N, Dokgoz FT, Keskin T, Kormaz KS, Syed HM.Continuous fermentative hydrogen production from cheesewhey wastewater under thermophilic anaerobic conditions.Int J Hydrogen Energ 2009;34:7441e7.

[17] Davila-Vazquez G, Alatriste-Mondragon F, de Leon-Rodriguez A, Razo-Flores E. Fermentative hydrogenproduction in batch experiments using lactose, cheese wheyand glucose: Influence of initial substrate concentration andpH. Int J Hydrogen Energ 2008;33:4989e97.

[18] Yang P, Zhang R, McGarvey JA, Benemann JR. Biohydrogenproduction from cheese processing wastewater by anaerobicfermentation using mixed microbial communities. Int JHydrogen Energ 2007;32:4761e71.

[19] Van Ginkel SW, Sung S, Lay JJ. Biohydrogen production asa function of pH and substrate concentration. Environ SciTechnol 2001;35:4726e30.

[20] Argun H, Kargi F, Kapdan IK, Oztekin R. Biohydrogenproduction by dark fermentation of wheat powder solution:effects of C/N and C/P ratio on hydrogen yield and formationrate. Int J Hydrogen Energ 2008;33:1813e9.

[21] Park W, Hyun SH, Oh SE, Logan BE, Kim IS. Removal ofheadspace CO2 increases biological hydrogen production.Environ Sci Technol 2005;39:4416e20.

[22] APHA, AWWA, WEF. Standard methods for the examinationof water and wastewater. 20th ed. Washington DC, USA:American Public Health Association (APHA), AmericanWaterWorks Association (AWWA), Water Environment Federation(WEF); 1998.

[23] Khanal SK, Chen WH, Li L, Sung SW. Biological hydrogenproduction: effects of pH and intermediate products. Int JHydrogen Energ 2004;29:1123e31.

[24] Mu Y, Yu H-Q, Wang G. A kinetic approach to anaerobichydrogen-producing process. Water Res 2007;41:1152e60.

[25] Lay JJ. Biohydrogen generation by mesophilic anaerobicfermentation of microcrystalline cellulose. BiotechnolBioeng 2001;74:280e7.

[26] Lin CY, Lay CH. A nutrient formulation for fermentativehydrogen production using anaerobic sewage sludgemicroflora. Int J Hydrogen Energ 2005;30:285e92.

[27] Sekiguchi Y, Kamagata Y, Syutsubo K, Ohashi A, Harada H,Nakamura K. Phylogenetic diversity of mesophilic andthermophilic granular sludges determined by 16S rRNA geneanalysis. Microbiology-Sgm 1998;144:2655e65.

[28] Wisotzkey JD, Jurtshuk Jr P, Fox GE. PCR amplification of 16SrDNA from lyophilized cell cultures facilitates studies inmolecular systematics. Curr Microbiol 1990;21:325e7.

[29] Muyzer G, de Waal EC, Uitterlinden AG. Profiling of complexmicrobial populations by denaturing gradient gelelectrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl Environ Microbiol1993;59:695e700.

[30] Van Ginkel S, Logan BE. Inhibition of biohydrogen productionby undissociated acetic and butyric acids. Environ SciTechnol 2005;39:9351e6.

[31] Jones DT, Woods DR. Acetone-butanol fermentationrevisited. Microbiol Rev 1986;50:484e524.

[32] Wang C-H, Lu W-B, Chang J-S. Feasibility study onfermentative conversion of raw and hydrolyzed starch tohydrogen using anaerobic mixed microflora. Int J HydrogenEnerg 2007;32:3849e59.

[33] Oh SE, Van Ginkel S, Logan BE. The relative effectiveness ofpH control and heat treatment for enhancing biohydrogengas production. Environ Sci Technol 2003;37:5186e90.

[34] Nath K, Kumar A, Das D. Effect of some environmentalparameters on fermentative hydrogen production byEnterobacter cloacae DM11. Can J Microbiol 2006;52:525e32.

[35] Kim S, Seol E, Mohan Raj S, Park S, Oh Y-K, Ryu DDY. Varioushydrogenases and formate-dependent hydrogen productionin Citrobacter amalonaticus Y19. Int J Hydrogen Energ 2008;33:1509e15.

[36] Oh Y-K, Kim H-J, Park S, Kim M-S, Ryu DDY. Metabolic-fluxanalysis of hydrogen production pathway in Citrobacteramalonaticus Y19. Int J Hydrogen Energ 2008;33:1471e82.

[37] Oh YK, Seol EH, Kim JR, Park S. Fermentative biohydrogenproduction by a new chemoheterotrophic bacteriumCitrobacter sp Y19. Int J Hydrogen Energ 2003;28:1353e9.

[38] Mandal B, Nath K, Das D. Improvement of biohydrogenproduction under decreased partial pressure of H2 byEnterobacter cloacae. Biotechnol Lett 2006;28:831e5.

[39] Shin J-H, Hyun Yoon J, Eun Kyoung A, Kim M-S, Jun Sim S,Park TH. Fermentative hydrogen production by the newlyisolated Enterobacter asburiae SNU-1. Int J Hydrogen Energ2007;32:192e9.

[40] Collet C, Adler N, Schwitzguebel JP, Peringer P. Hydrogenproduction by Clostridium thermolacticum during continuousfermentation of lactose. Int J Hydrogen Energ 2004;29:1479e85.

[41] Chen WM, Tseng ZJ, Lee KS, Chang JS. Fermentativehydrogen production with Clostridium butyricum CGS5isolated from anaerobic sewage sludge. Int J Hydrogen Energ2005;30:1063e70.

[42] Zhang HS, Bruns MA, Logan BE. Biological hydrogenproduction by Clostridium acetobutylicum in an unsaturatedflow reactor. Water Res 2006;40:728e34.

[43] Levin DB, Islam R, Cicek N, Sparling R. Hydrogen productionby Clostridium thermocellum 27405 from cellulosic biomasssubstrates. Int J Hydrogen Energ 2006;31:1496e503.

[44] Collet C, Gaudard O, Peringer P, Schwitzguebel J-P. Acetateproduction from lactose by Clostridium thermolacticum andhydrogen-scavenging microorganisms in continuouscultureeEffect of hydrogen partial pressure. J Biotechnol2005;118:328e38.

[45] Ferchichi M, Crabbe E, Hintz W, Gil G-H, Almadidy A.Influence of culture parameters on biological hydrogenproduction by Clostridium saccharoperbutylacetonicum ATCC27021. World J Microbiol Biotechnol 2005;21:855e62.

[46] Lin C-Y, Hung W-C. Enhancement of fermentative hydrogen/ethanol production from cellulose using mixed anaerobiccultures. Int J Hydrogen Energ 2008;33:3660e7.

[47] Xiao B, Han Y, Liu J. Evaluation of biohydrogen productionfrom glucose and protein at neutral initial pH. Int J HydrogenEnerg 2010;35:6152e60.