Journal of Chemical Technology and Biotechnology J Chem Technol Biotechnol 81:1745–1752 (2006)
Effects of bioaugmentation
by an anaerobic lipolytic bacterium on
anaerobic digestion of lipid-rich waste
Dores G. Cirne,
1,2∗
Lovisa Bj
¨
ornsson,
1
Madalena Alves
2
and Bo Mattiasson
1
1
Department of Biotechnology, Center for Chemistry and Chemical Engineering, Lund University, Lund, Sweden
2
Centro de Engenharia Biologica, Universidade do Minho, PT-4710-057 Braga, Portugal
Abstract: The effect of bioaugmentation with an anaerobic lipolytic bacterial strain on the anaerobic digestion
of restaurant lipid-rich waste was studied in batch experiments with a model waste containing 10% lipids
(triolein) under two sets of experimental conditions: (A) methanogenic conditions, and (B) initially acidogenic
conditions in the presence of only the lipolytic strain biomass (4 days), followed by methanogenic conditions.
The bioaugmenting lipolytic strain, Clostridium lundense (DSM 17049
T
), was isolated from bovine rumen. The
highest lipolytic activity was detected at the beginning of the experiments. A higher methane production rate,
27.7 cm
3
CH
4(STP)
g
−1
VS
added
day
−1
(VS, volatile solids) was observed in experiment A with the presence of the
bioaugmenting lipolytic strain under methanogenic conditions. The highest initial oleate concentration, 99% of
the total oleate contained in the substrate, was observed in the experiments with the bio augmenting lipolytic strain
under treatment A conditions; the levels of palmitate a nd stearate were also higher until day 15, indicating that
the bioaugmentation strategy improved the hydrolysis o f the lipid fraction. In general, the results indicated that
degradation of the long chain fatty acids (LCFAs) controlled the digestion process.
2006 Society of Chemical Industry
Keywords: biological anaerobic treatment; bioaugmentation; inhibition; hydrolysis; lipid-rich waste; LCFA
INTRODUCTION
Lipids (characterized as oil, grease, fat and free long
chain fatty acids, LCFAs) can be a major organic
component in wastewater. Approximately 15–20%
of the total solids in sewage sludge consists of
lipids.
1
Lipids are preferentially found as triglycerides,
and oleate is one of the most abundant LCFAs.
2,3
Triglycerides can comprise up to 65% (w/w) of meat
industry waste and contribute to waste solids from the
food processing industry.
4
Large amounts of lipid-rich
waste are also generated in the edible oil processing
industry, in the dairy industry and in restaurant waste.
Lipids are attractive substrates for anaerobic digestion
due to the higher methane yield obtained compared
with proteins or carbohydrates.
5
However, anaerobic
microbial degradation of lipids is one of the least
investigated topics in this area.
There is much discussion in the literature as to which
step limits conversion of the substrate to biogas when
complex substrates such as lipids are degraded. Studies
have shown that the hydrolysis of lipids to glycerol and
LCFAs can be rapid, and that the main problem during
lipid digestion is the further degradation of LCFAs.
6,7
However, the substrate interface area available for
hydrolysis may be a limiting factor.
8
In the case
of slaughterhouse wastewater, where high amounts
of suspended solids are present, the liquefaction of
colloids adsorbed onto the biomass and the hydrolysis
of suspended solids entrapped within the biomass bed
were found to be the limiting steps in biodegradation.
9
Petruy and Lettinga
10
also found liquefaction to be
the rate limiting step during the digestion of milk fat.
Salminen and co-workers
11
reported that hydrolysis
limited the digestion of a poultry slaughterhouse
waste due to a high concentration of propionate,
which was the consequence of the presence of
LCFAs. The type of lipid also has an influence
on what step in the degradation process will be
limiting.
12
Lipid hydrolysis limitation has not been
studied as much as LCFAs degradation. Different
pretreatment methods have been investigated to
improve the digestion process, such as: (1) chemical
pretreatment of the waste by NaOH, Ca(OH)
2
or HCl;
(2) enzyme addition; and (3) biological pretreatment
by utilization of hydrolytic microorganisms, covering
a variety of wastes, such as activated sludge,
lipid-rich wastewaters, household solid waste and
slaughterhouse wastewater.
2,6
Each method, however,
has serious drawbacks. Masse and co-workers
6
do
not recommend pretreatment with an alkali because
itresultsinanincreaseinpHinthedigestion
process. Pretreatment by enzyme addition was shown
to have positive effects (increased free LCFAs
concentration), however, the process can be expensive
and thus not economically feasible. The possibility
of pretreatment with enzyme-producing aerobic
∗
Correspondence to: Dores G. Cirne, Department of Biotechnology, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, SE-221
00, Lund, Sweden
E-mail: Dores.Cirne@biotek.lu.se
(Received 12 F ebruary 2006; revised version received 31 March 2006; accepted 31 March 2006)
Published online 2 October 2006
; DOI: 10.1002/jctb.1597
2006 Society of Chemical Industry. J Chem Technol Biotechnol 0268– 2575/2006/$30.00
DG Cirne et al.
microorganisms has been demonstrated with a lipolytic
fungus
13
and with mixed bacterial cultures comprising
lipase, protease and amylase producers.
14
However,
when using aerobic microorganisms for pretreatment,
oxygen supply may be required, increasing the cost
of the process. While the addition of anaerobic
microorganisms has been investigated as a means of
improving xylanolytic, cellulolytic and hemicellulolytic
activities, it has, to the authors’ knowledge, not
previously been investigated using strains exhibiting
lipolytic activity.
15,16
In the present study, the effect of bioaugmentation
by an anaerobic lipolytic strain as a means of improving
hydrolysis and solubilization of lipids in the anaerobic
digestion process of restaurant lipid-rich waste was
studied using a model substrate. The bioaugmenting
lipolytic bacterium is a strict anaerobe isolated from
bovine rumen.
17
MATERIALS AND METHODS
Substrate
The effect of bioaugmentation on the anaerobic diges-
tion of lipid-rich waste was studied in batch experi-
ments using a well-defined substrate (Table 1). The
substrate composition was based on the composition
of the restaurant waste from the University of Minho,
located in Campus de Gualtar (Braga, Portugal). This
waste consisted of a one week basis sample from the
waste produced in the restaurant. The amount of each
component was based on the chemical oxygen demand
(COD); 10% was the contribution from lipids, 45%
from protein, and 45% from carbohydrate (30% from
starch and 15% from cellulose) (Table 1). Nutri-
ents with the following final composition were added
to ensure that no nutrient deficiency would occur
(mg dm
−3
):
18
NH
4
Cl, 1000; NaCl, 100; MgCl
2
·
6H
2
O, 100; CaCl
2
· 2H
2
O, 50; K
2
HPO
4
· 3H
2
O,
400; cysteine HCl, 500; FeCl
2
· 2H
2
O, 2; H
3
BO
3
,
0.050; ZnCl
2
0.050; CuCl
2
, 0.030; MnCl
2
· 4H
2
O,
0.050; (NH
4
)
6
Mo
7
O
24
· 4H
2
O, 0.050; AlCl
3
, 0.050;
CoCl
2
· 6H
2
O, 0.050; NiCl
2
, 0.050; EDTA, 0.500;
Na
2
SeO
3
· 5H
2
O, 0.100; biotin, 0.020; folic acid,
0.020; pyridoxine HCl, 0.100; riboflavin, 0.005; thi-
amine HCl, 0.005; cyanocobalamine, 0.001; nicotinic
Table 1. Composition of the substrate used in the experiments
Composition
Substrate
COD
(gg
−1
)
TS
(gg
−1
)
TKN
a
(gg
−1
)
Amount
(COD %)
Lipid – triolein (65%
purity)
2.67 1.01 0 10
Protein – whey protein
(80% purity)
1.25 0.94 0.12 45
Carbohydrate – soluble
starch
1.01 0.92 0 30
Carbohydrate – α-
cellulose
1.10 0.96 0 15
a
TKN – Total Kjeldahl nitrogen.
acid, 0.005; p-aminobenzoic acid, 0.005; lipoic acid,
0.005; and DL-panthothenic acid, 0.005. The pH
was adjusted to 7.0 with 5 mol dm
−3
NaOH. Bicar-
bonate was added at a concentration of 14 g dm
−3
to
provide buffering capacity. The final substrate total
solids (TS) content in the experiments was 6.2% and
the total COD was 0.8 g per bottle, corresponding to
33.3gdm
−3
.
Methanogenic inoculum
Sludge from an anaerobic digester treating municipal
sewage sludge and potato processing waste (TS 4.8%
and volatile solids, VS 3.1%) (Ellinge, Sweden)
wasusedasinoculumataVSratioof1.35
(substrate:inoculum).
Bioaugmenting strain
The bioaugmenting lipolytic strain (Clostridium lun-
dense, DSM 17 049
T
) was isolated from bovine rumen
fluid.
17
The cells were cultivated for 88 h in an anaer-
obic medium consisting of the basal salts described
by Markossian and co-workers,
19
0.0025% (w/v) reza-
surin, 120 mg dm
−3
L-cysteine hydrochloride, pep-
tone, yeast extract and glucose all at 0.5% (w/v); olive
oil 2% (v/v) was the lipid source. The culture broth
was washed once with oxygen-free distilled water to
remove most of the remaining substrate. The amount
of lipolytic strain biomass added corresponded to 1.3%
of the VS of the methanogenic inoculum added.
Experimental set-up
The effect of bioaugmentation on the overall
biomethanation process was studied under two kinds
of conditions: treatment (A), methanogenic condi-
tions where the lipolytic strain biomass was added on
day 0 together with the methanogenic inoculum; and
treatment (B), initially acidogenic conditions followed
by methanogenic conditions. In this case, only the
lipolytic strain biomass was added at day 0, and after
4 days the methanogenic inoculum was added. Control
experiments without substrate were used to evaluate
the contribution of the methanogenic inoculum alone
(blank 1) as well as together with active (blank 2)
and inactivated lipolytic strain biomass (autoclaved
for 40 min at 121
◦
C) (blank 3). Control experiments
were also performed with substrate and methanogenic
inoculum in the presence (control 1) and absence of
inactivated lipolytic strain biomass (control 2). The
tests were performed in 100 cm
3
serum bottles with
a liquid volume of 24 cm
3
kept in a nitrogen atmo-
sphere. The bottles were incubated at 37
◦
C with
agitation (stirring) at 150 rpm. Assays were run using
11 replicates and at liquid phase sampling, one vial
was randomly taken for analysis.
1,11
Three bottles were
used for gas phase studies during the experiment and
the liquid contents in these were analysed at the end of
the experiment. Liquid phase sampling was performed
on days 0, 1, 2, 3, 5, 9, 15, 22 and on the final day in
treatment A and on days 0, 1, 2, 4 (before and after
1746 J Chem Technol Biotechnol 81:1745– 1752 (2006)
DOI: 10.1002/jctb
Anaerobic digestion of lipid-rich waste – effects of bioaugmentation
the addition of the methanogenic inoculum), 5, 9, 12
and on the final day in treatment B.
Analysis
Biogas production was measured using a hand-
held pressure transducer capable of measuring a
pressure variation of 2 bar (0 to ± 202.6kPa) over
a device range of −200 to +200 mV, with a minimum
detectable variation of 0.005 bar, corresponding to
0.05 cm
3
of biogas in a 10 cm
3
headspace (Centrepoint
Electronics, Galway, Ireland).
20
Gas composition was
analysed using a Varian 3350 GC-TCD (Walnut
Creek, CA, USA) in accordance with the method of
Mshandete et al.
21
The values of methane production
were corrected for standard temperature and pressure
conditions (STP). For liquid phase sample analysis,
the whole content of a serum bottle was centrifuged
at 4000 rpm for 30 min. The supernatant was taken
for analysis of concentration of volatile fatty acids
(VFAs), lipase activity and soluble COD. The solids
were washed twice with 10 cm
3
of distilled water,
acidified with 5 mol dm
−3
ofHCltopH2,and
stored at −20
◦
C for analysis of biomass-associated
LCFAs. Aliquots of the biomass (ranging from 2 to
4cm
3
) were dried at 105
◦
C for 20 h. The total lipids
were extracted from the dried solids as described
by Bligh and Dyer.
22
Samples (1.5cm
3
) from the
chloroform extract were passed through silica-based
columns (Bond Elut
LRC-Si, 100 mg, Varian,
Middelburg, the Netherlands) for separation of the
different classes of lipids. The neutral lipid fraction
was eluted with 1.5cm
3
chloroform and the eluate
was evaporated under nitrogen. The LCFAs present
in the residue were then methylated by dissolving them
in a methylation agent (methanol containing 5% (v/v)
sulphuric acid). The reaction was allowed to proceed
for 2 h at 50
◦
C, and was stopped by adding 5 cm
3
of a 5% (w/v) NaCl solution. Finally, the methyl
esters were extracted twice with 5 cm
3
n-hexane.
23
The extracts were stored at −20
◦
C until analysis.
LCFA methyl esters were analysed using a Varian
3400 GC-FID as described by Lyberg et al.
24
Total solids, VS and COD (total and soluble)
were measured according to standard methods.
25
The COD of the substrate components was deter-
mined using suspensions of each component, which
were homogenized using a homogenizer Disp 25
(20 500 rpm; Inter Med, Roskilde Denmark). Volatile
fatty acid concentrations were measured using HPLC
in accordance with the method of Mshandete et al.
21
Lipase activity was measured in accordance with
the method of Winkler and Stuckmann (1979).
26
The assay, using p-nitrophenylpalmitate (0.30 g dm
−3
)
(Sigma St Louis, MO, USA) as substrate in S
¨
orensen
phosphate buffer (pH 8.0, 0.05 mol dm
−3
) contain-
ing isopropanol (10 cm
3
per 90 cm
−3
of buffer),
sodium deoxycholate (2.07 g dm
−3
) and gum ara-
bic (1.00 g dm
−3
), was carried out with incubation
at 37
◦
C for 15 min followed by measuring the
absorbance at 410 nm against a control with inacti-
vated enzyme. One enzyme IU is defined as 1
µmol
of p-nitrophenol enzymatically released from the sub-
strate per minute under the conditions of the assay.
Total Kjeldahl nitrogen was determined after diges-
tion of samples according to the manufacturer’s
instructions by colorimetric analysis using a FIAS-
tar 5000 analyser coupled to a 5027 sampler (Foss
Tecator AB H
¨
ogan
¨
as, Sweden). The pH was mea-
sured with a CG 842 pH meter (Schoot, Ger
¨
ate
GmbH, Hofheim, Germany) immediately after sam-
pling.
RESULTS
Methane production
A higher methane production rate was observed in the
presence of the bioaugmenting lipolytic strain biomass
in treatment A (Fig. 1 and Table 2). In treatment
B, no bioaugmentation effects were observed. For all
conditions investigated, the percentage of substrate
methanization was above 90%, i.e. the fraction of the
theoretical methane yield that was obtained exper-
imentally (Table 2). The methane yields obtained
were similar for test samples and controls for both
treatments (Table 2).
0
200
400
600
800
0 20406080
Methane production (cm
3
CH
4(STP)
g
-1
VS
added
)
0
200
400
600
800
0 20406080
Time (days)
Figure 1. Specific cumulative methane production; top – treatment
A, bottom – treatment B. The arrow indicates the addition of
methanogenic inoculum in treatment B; inocula methane production
not subtracted. -
°
- control 1; -- control 2; -
- test sample.
J Chem Technol Biotechnol 81:1745–1752 (2006) 1747
DOI: 10.1002/jctb
DG Cirne et al.
0
10
20
30
40
50
0 5 10 15 20 25
Soluble COD (g dm
-3
)
0
10
20
30
40
50
0 5 10 15 20 25
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
0 5 10 15 20 25
Time (days)
pH
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
0 5 10 15 20 25
Time (days)
Figure 2. Soluble COD (top) and pH (bottom) profiles up to day 22; left – treatment A, right – treatment B. Thearrowindicatestheadditionof
methanogenic inoculum in treatment B. -
°
- control 1; -- control 2; -
- test sample.
Table 2. Summary of the results of biomethanation experiments (mean values ± standard deviation are shown, n = 3)
Treatment A Treatment B
Parameter Control 1 Control 2 Test sample Control 1 Test sample
Maximum methane production rate (cm
3
CH
4(STP)
g
−1
VS
added
day
−1
)
21.6 ± 1.5
c
18.7 ± 1.1
d
27.7 ± 1.0
e
24.0 ± 3.5
f
25.2 ± 3.5
g
Methane yield
a
(cm
3
CH
4(STP)
g
−1
VS
added
) 447 ± 36 434 ± 103 444 ± 26 415 ± 5 444 ± 41
Percentage methanization
b
(%) 98 ± 895± 23 98 ± 691± 198± 9
a
Inocula, i.e. methanogenic biomass and lipolytic strain biomass methane production subtracted.
b
Calculated by comparison with the theoretical methane potential of the substrate after subtracting inocula’s methane production.
c
Determined between days 18 and 22.
d
Determined between days 16 and 26.
e
Determined between days 10 and 26.
f
Determined between days 18 and 28
g
Determined between days 18 and 24.
COD and pH
The initial soluble COD was similar for test samples
and controls (Fig. 2). In treatment A the soluble COD
observed was slightly higher when the active lipolytic
strain biomass was present. For both the test sample
and control 1 (containing inactivated lipolytic strain
biomass), higher concentrations of soluble COD were
observed than in control 2 (with only methanogenic
inoculum) with the exception of a deviation on day 2.
After day 15, the soluble COD decreased for the test
sample and controls under treatment A conditions. In
treatment B, no differences between the test sample
and control with inactivated lipolytic strain biomass
were observed. On day 4 the soluble COD decreased
due to a dilution effect caused by the addition of the
methanogenic inoculum. Thereafter, as a consequence
of microbial activity, the soluble COD concentration
increased. In treatment A the pH showed substantial
differences between test sample and controls, with
the pH always lower in the test sample than in
controls.
Biomass-associated LCFAs and lipolytic activity
The biomass-associated LCFAs were analysed only
in treatment A since adding the lipolytic strain
biomass produced significant effects. Higher initial
oleate concentration was detected in the experiment
in which the active lipolytic strain was added,
corresponding to 99% of the total oleate that could
be released by trioleate hydrolysis (Fig. 3). This
observation was in agreement with the lipolytic
activity data (data not shown). The initial lipolytic
1748 J Chem Technol Biotechnol 81:1745– 1752 (2006)
DOI: 10.1002/jctb
Anaerobic digestion of lipid-rich waste – effects of bioaugmentation
0
10
20
30
40
50
60
0 5 10 15 20 25
Oleate (mg g
-1
d.w.)
0
2
4
6
8
10
0 5 10 15 20 25
Stearate (mg g
-1
d.w.)
0
10
20
30
40
50
60
0 5 10 15 20 25
Time (days)
Palmitate (mg g
-1
d.w.)
0
2
4
6
8
10
0 5 10 15 20 25
Time (days)
Myristate (mg g
-1
d.w.)
Figure 3. LCFA profiles up to day 22 in treatment A. -
°
- control 1; -- control 2; -
- test sample.
activity (91 IU dm
−3
) was approximately 2.5 times
higher than in the controls. The highest activity was
observed at the beginning of the experiment. After
24 h of incubation the measurable lipolytic activity
decreased to 5 IU dm
−3
. In treatment B, after 24 h,
the lipolytic activity decreased to 57% of the initial
value. After the initial high concentration, the oleate
concentration decreased to low values, while stearate
and palmitate concentrations increased to values above
7.5 and 40 mg g
−1
d.w. (d.w., dry weight), respectively
(Fig. 3). The palmitate concentration in the test
sample remained higher than in the controls until
day 15, after which it decreased to 7 mg g
−1
d.w.
The stearate concentration was also higher in the test
sample until day 15 (6–8 mg g
−1
d.w.). No significant
differences were observed between the test sample and
controls for myristate and the concentrations were the
lowest observed, 2–4 mg g
−1
d.w.
VFAs
Acetate (6–8gdm
−3
) and propionate (2.5–4gdm
−3
)
were the most abundant acids for both treatment
conditions investigated (Fig. 4). The concentration of
acetate was higher in the test sample for treatment A.
After day 15, the concentration of acetate decreased
while a slight increase in propionate concentration
was observed. Other VFAs showed similar levels in
test sample and controls for both treatment conditions
investigated (Fig. 4). In treatment B, data obtained
after the addition of the methanogenic inoculum
(day 4) and up to day 12 are presented. The initial
hydrolysis of the lipid fraction of the substrate had
no significant effect on the VFAs produced (data
not shown). At the end of the experiments, the
concentration of VFAs was between 0.2gdm
−3
and
1gdm
−3
.
DISCUSSION
When the effect of bioaugmenting the anaerobic
digestion process with a specific microorganism is
positive, it can result in two effects: an enhancement
in methane yield and/or an increase in the methane
production rate. If an increase in methane yield
is observed, it results from increasing the ultimate
bioavailability of the substrate. Increased methane
production rate results from faster conversion of the
substrate(s) involved in the limiting conversion step.
Improved enzymatic hydrolysis should primarily help
to reduce the duration of the hydrolytic step. This
was the main effect observed in this study. The
addition of lipases and lipolytic microorganisms has
been successfully used to treat wastewaters containing
high levels of lipids under aerobic conditions.
2,12,13
The enzyme and/or the microorganisms are in many
cases added to the aeration step at wastewater
treatment plants. Regarding anaerobic digestion,
this treatment strategy, when applied, has been
done under aerobic conditions before digestion. The
results in this case are contradictory.
12,13
The major
drawback of using bioaugmentation under anaerobic
conditions is the accumulation of LCFAs, which
may inhibit the digestion process. In a previous
study on the influence of lipid concentrations on
hydrolysis and biomethanation of lipid-rich waste,
the addition of a commercial lipase improved lipid
hydrolysis.
33
In the same study, the inhibition of
J Chem Technol Biotechnol 81:1745–1752 (2006) 1749
DOI: 10.1002/jctb