Biochemical Engineering Journal 91 (2014) 283289

Contents lists available at ScienceDirect

Biochemical Engineering Journal

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

Regular Article

Acute impact of tetracycline and erythromycin on the storage

mechanism of polyhydroxyalkanoates
Tugce Katipoglu-Yazan a, , Ilke Pala-Ozkok a , Emine Ubay-Cokgor a , Derin Orhon a,b
a
Faculty of Civil Engineering, Environmental Engineering Department, Istanbul Technical University, 34469 Maslak, Istanbul, Turkey
b
The Science Academy, 34353 Besiktas, Istanbul, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: The study investigated acute impact of tetracycline and erythromycin on substrate storage under aerobic
Received 21 March 2014 conditions. A ll and draw reactor fed with peptone mixture was maintained at steady-state at a sludge
Received in revised form 29 July 2014 age of 10 days; the acclimated biomass was used in a series of batch runs. The rst run served as control
Accepted 1 September 2014
reactor with organic substrate alone and the others were started with antibiotic doses of 50 mg/L and
Available online 6 September 2014
200 mg/L for assessing intracellular storage. Parallel batch reactors were also conducted for recording
oxygen uptake rate proles. Both antibiotics enhanced substrate storage, leading to higher levels of
Keywords:
polyhydroxyalkanoates incorporated into biomass, but they impaired its internal utilization for microbial
Substrate storage
Polyhydroxyalkanoates
growth. The observed decrease in oxygen consumption under the acute effect of antibiotics could partially
Kinetic parameters be related to substrate storage except for 50 mg/L of erythromycin dosing suggesting an additional
Dissolved oxygen substrate binding mechanism by antibiotics, leading to residual biodegradable substrate.
Peptone 2014 Elsevier B.V. All rights reserved.
Antibiotic

1. Introduction
Storage of intracellular biopolymers is now recognized as a sig-
nicant auxiliary process during substrate utilization by microbial
cultures. It was mainly observed under transient feeding condi-
tions, where sequential feast and famine phases trigger substrate
storage [1]. In essence, storage results from an imbalance between
removal of available substrate and microbial growth potential;
while substrate may be removed, limitations on the metabolic reac-
tions leading to growth may prevent consumption of all energy
and divert a fraction of the substrate for generating intracellu-
lar biopolymers [2]. Batch reactors are often selected as the most
suitable experimental tool for investigating different aspects of
substrate biodegradation, mainly because they offer accurate eval-
uation of transient responses and resulting concentration proles
of major parameters [3,4]. Dynamic conditions sustained in batch
reactors approximate pulse feeding and induce a physiological
adaptation for the microbial community, often leading to substrate
storage.
The storage mechanism is often studied with simple, readily
biodegradable substrates like acetate or glucose, generating poly-
hydroxybutyrate (PHB) and glycogen as storage products. Reported
results suggested that up to 70% of the simple substrate could be
Corresponding author. Tel.: +90 2122856542; fax: +90 2122853781.
E-mail address: katipoglut@itu.edu.tr (T. Katipoglu-Yazan).
converted to storage products under pulse feeding [5,6]. They also
indicated that the magnitude of storage was likely to exhibit signif-
icant variations depending on the nature of substrate, the feeding
regime and culture history i.e. sludge age of the microbial culture
[714]. Expected storage would be substantially lower in com-
plex substrate mixtures such as domestic sewage where only a
small percent of the organic matter consist of readily biodegrad-
able compounds favoring formation of intracellular biopolymers.
In fact, in a study using a peptone-meat extract mixture as the
organic carbon source with biodegradation characteristics similar
to domestic sewage, storage of only 30 mg COD/L of intracellular
biopolymers was observed, corresponding to 60% of the available
readily biodegradable COD at a sludge age of 8 days [15].
The storage mechanism is also important when studying inhi-
bition of substrate biodegradation under the impact of adverse
chemicals and xenobiotics. Traditionally, inhibition was inter-
preted by means of enzyme analogy affecting only microbial
growth [16,17], simply because corresponding substrate utiliza-
tion was only characterized by a single overall parameter such
as BOD or COD. After recognition and experimental assessment
of COD fractions with different biodegradation characteristics in
domestic sewage and industrial wastewaters [15], interpretation of
biodegradation involved all these fractions and related biochemi-
cal processes [18]. While multi-component models incorporated
all these processes, they initially disregarded substrate storage
[19,20]. Later, these models were modied to account for simulta-
neous substrate utilization for growth and storage [21,22]. Thanks

http://dx.doi.org/10.1016/j.bej.2014.09.002
1369-703X/ 2014 Elsevier B.V. All rights reserved.
284 T. Katipoglu-Yazan et al. / Biochemical Engineering Journal 91 (2014) 283289
Nomenclature
Cs initial biodegradable substrate (mg COD/L)
CSG fraction of substrate directly utilized for microbial
growth (mg COD/L)
CSGP fraction of the initial substrate potentially available
for microbial growth (mg COD/L)
CSR residual fraction of the initially available biodegrad-
able substrate (mg COD/L)
CST fraction of substrate diverted to storage (mg COD/L)
CSTR substrate COD equivalent of residual PHA entrapped
in biomass (mg COD/L)
CSTU fraction of substrate stored and internally utilized
(mg COD/L)
kSTO maximum storage rate (1/day)
KSTO half saturation coefcient for storage and internal
growth (mg COD/L)
OUR oxygen uptake rate (mg O2 /L h)
PHA polyhydroxyalkanoates
PHB polyhydroxybutyrate
PHV polyhydroxyvalerate
SO oxygen consumed (mg/L)
SS readily biodegradable substrate (mg COD/L)
SOT theoretical oxygen demand (mg/L)
XSTO stored PHA (mg COD/L)
XSTR residual PHA entrapped in biomass (mg COD/L)
YH yield coefcient (mg COD/mg COD)
YSTO storage yield (mg COD/mg COD)
3H2MV 3-hydroxy-2-methylvalerate
to these modeling tools, interpretation of inhibitory substances
was no longer limited to substrate utilization for microbial growth,
but it could also cover all related biochemical mechanisms such
as hydrolysis, endogenous respiration and substrate storage [14].
In this context, the storage mechanism is particularly important,
mainly because it is most likely to be affected under stress con-
ditions, i.e. diversion of substrate from growth to storage under
adverse impact and when disregarded it may signicantly distort
the kinetics of parallel biochemical processes, leading to misinter-
pretation of the inhibitory impact.
The previous works investigated the impact of selected antibi-
otics on either heterotrophic or autotrophic bacteria in terms of
general process kinetics. The main objective of this part of the study
was to evaluate the acute impact of two selected inhibitors on the
mechanism of substrate storage in a nitrifying microbial commu-
nity. Two major antibiotics, namely tetracycline and erythromycin
were selected as the inhibitor compounds, due to the fact that they
take place in our daily life, discharged into the wastewaters and to
the environment [23] and well studied for their inhibitory effects
[2426]. The acute impact was evaluated on the biodegradation
of peptone-meat extract mixture called thereafter peptone mix-
ture for simplicity under aerobic conditions. It was selected as
the complex organic carbon source because it is prescribed as the
standard substrate by ISO 8192 [27] in similar inhibition studies
[28] and it approximates the COD fractionation and biodegradation
characteristics of domestic sewage [29].
2. Materials and methods
2.1. Experimental rationale
The experiments were conducted as part of a comprehensive
study investigating the impact of antibiotics on the heterotrophic
and autotrophic communities in single sludge systems operated
Table 1
Operating characteristics of batch experiments [33].
Runs Antibiotic Biomass (mg VSS/L) S0 /X0 (mg T ( C)
COD/mg VSS)
Run 1 920 0.55 21
Run 2.1 50 mg TET/L 1005 0.50 21
Run 2.2 200 mg TET/L 995 0.51 24
Run 3.1 50 mg ERY/L 995 0.51 25
Run 3.2 200 mg ERY/L 1035 0.49 25
Peptone mixture having 506 mg COD/L was added in all runs.
under aerobic conditions. The rst part involved running a ll and
draw reactor with biomass taken from a domestic wastewater
treatment plant. The system was maintained at steady-state at a
sludge age of 10 days, a value compatible with the level selected for
biological treatment systems targeting combined organic carbon
removal and nitrication [30].
The acclimated biomass from the ll and draw reactor was used
for starting a series of batch runs, each run including two parallel
batch reactors, one for the assessment of intracellular biopolymers
and the other, for the respirometric analysis of oxygen uptake rate
(OUR) proles. The initial COD concentration of the peptone mix-
ture was adjusted to around 500 mg/L. The rst run served as the
control reactor; the other two runs included in a feed having ini-
tial dose of 50 mg/L and 200 mg/L of the antibiotic for its acute
impact, primarily on substrate storage. The highest antibiotic level
was selected mainly to approximate efuent discharges of phar-
maceutical industries and hospitals which were reported to reach
up to 100500 mg/L [23,31]. As an example, tylosin antibiotic was
detected in the range of 20200 mg/L in an antibiotic wastewater
[32]. The lowest concentration was selected based on the respiro-
metric studies previously conducted on three different antibiotics
indicating oxygen uptake rate reduction by more than 50% for the
same substrate [17]. Batch experiments were conducted separately
for tetracycline and erythromycin.
2.2. Batch experiments
Batch experiments were conducted in aerated 2 L cylindrical
benchtop reactors having 30 cm diameter. The dissolved oxygen
was supplied air supplying diffusers connected to airline. The oxy-
gen level in the reactors was maintained above 3.0 mg/L (33%
oxygen saturation) by mixing with mechanical impellers to pre-
vent any oxygen limitations on biochemical reactions. In order to
determine the effects of selected antibiotics on substrate storage
mechanism, related parameters such as COD and storage com-
pounds were monitored. Batch reactors were supplied with 50
and 200 mg/L of TET or ERY solutions in addition to peptone mix-
ture. Initial food to microorganisms ratio (S0 /X0 ) was adjusted
to 0.490.55 mg COD/mg VSS where macro (320 g/L K2 HPO4 , and
160 g/L KH2 PO4 ) and micro nutrients (15 g/L MgSO4 7H2 O, 0.5 g/L
FeSO4 7H2 O, 0.5 g/L, ZnSO4 7H2 O, 0.41 g/L MnSO4 H2 O, 2.65 g/L,
CaCl2 2H2 O g/L) were also fed to support microbial activity. The sys-
tem was fed once on a daily basis and continuously aerated to have
a hydraulic retention time of 1 day. Activated sludge was settled
for 1 h and efuent was decanted to 2 L. COD removal efciencies
and internal storage biopolymers, polyhydroxyalkanoates (PHAs)
including polyhydroxybutyrate (PHB), polyhydroxyvalerate (PHV)
and 3-hydroxy-2-methylvalerate (3H2MV) were monitored during
the experiments for the determination of biological activity [33].
The tests lasted for 24 h and pH remained at neutral levels for this
time period. Information related to batch experiments was given in
Table 1. Repeated experiments yielded the same proles of related
parameters.
 T. Katipoglu-Yazan et al. / Biochemical Engineering Journal 91 (2014) 283289
2.3. Measurements of storage products
Samples taken from mixed liquor were ltered through AP40
glass bers for suspended solids (SS) and volatile suspended solids
(VSS) analyses as described in Standard Methods [34]. For COD anal-
ysis, 0.45 m Millipore membrane lters were used. The ltrate
was preserved with H2 SO4 and analyzed according to ISO 6060 [35].
pH was monitored by Orion 520 A pH meter. The pH meter was con-
nected with an online dosing system for a bicarbonate-carbonate
buffer solution that will keep the pH at neutral ranges.
Mixed liquor was directly preserved with formaldehyde and
biological activity was prevented in PHAs samples. K-P buffer solu-
tion was used for washing and samples were freeze dried. Boiling
at 100 C resulted in extraction, hydrolysation, and esterication
of the samples [5]. The free acids were water extracted and the
remaining organic phase was measured by gas chromatography
(Agilent 6890 N). Benzoic acid was used as internal standard.
Samples were analyzed in duplicate for each parameter. Standard
deviations of COD were 3%.
2.4. Respirometric measurements
Previous works involving respirometric analysis indicated that
evaluation of OUR proles enabled determination of the effects of
inhibitors on substrate biodegradation kinetics in terms of model
coefcients [28]. These model coefcients exert variable sensitivity
at different regions of OUR prole that are used to identify the most
sensitive parameter [10,15,36].
Modeling exercise involved analyzing and calibrating the OUR
and PHA proles. Kinetic and stoichiometric parameters were
estimated by using SIMPLEX algorithm and AQUASIM simulation
program [37]. Parameter identiability procedure was also imple-
mented to determine best identiable parameter subsets. The
UNCSIM module proposed by Brun et al. [38] was used for the cal-
culation of sensitivity rankings of model parameters as previously
reported [29].
In this study computer controlled oxygen uptake rate (OUR)
measurements, with well proven reliability, were conducted with
TET and ERY having identical conditions with batch experiments.
Biomass seeding alone to obtain the initial endogenous OUR level
was supplied with peptone mixture and antibiotic solutions to
have desired S0 /X0 ratios. A continuous lab-scale online respirom-
eter (Ra-Combo) was used for OUR measurements (Applitek Co.,
Nazareth, Belgium). Oxygen depletion was measured with highly
developed dissolved oxygen probes located in closed vessels of the
respirometer [39]. Biological activity was monitored with addi-
tional COD sampling. 69 mg/L of dissolved oxygen level was
maintained in the reactor and pH was recorded during the exper-
iments. Respirometric data was analyzed considering pH and
temperature corrections.
3. Results and discussion
The nature and magnitude of substrate storage mainly depends
on substrate characteristics, the feeding regime and the culture his-
tory i.e. the sludge age of the microbial culture. As in this study,
utilization/biodegradation of organic substrate is now evaluated
by means of respirometric procedures in waste/wastewater [15];
they require the use of batch reactors for assessing the evolution of
the oxygen uptake rate (OUR) prole associated with the selected
experimental conditions. Accordingly, all auxiliary measurements
supporting OUR proles are also derived from parallel batch reac-
tors, which conveniently yield concentration transients of relevant
parameters. Batch reactors start with an initial substrate concen-
tration, acting as a pulse feeding, which then leads to a sequence of
285
Fig. 1. Storage proles in the control reactor for PHA and individual storage com-
ponents PHB, PHV and 3H2MV [33].
feast and famine conditions abundance and absence of substrate
favoring substrate storage [1].
The peptone mixture, selected as the source of organic matter in
the study is a complex substrate with a COD fractionation quite sim-
ilar to domestic sewage and most industrial wastewaters: previous
studies indicated that a small fraction less than 10% of its COD
content was of readily biodegradable nature and the rest included
two fractions of slowly biodegradable components with different
hydrolysis rates [15,40]. Therefore, the substrate fraction diverted
to the generation of intracellular biopolymers was generally lower
as compared with simple/readily biodegradable compounds like
acetate or glucose [9,41].
Substrate storage observed for the control reactor oper-
ated without antibiotic dosage was illustrated in Fig. 1:
It essentially involved polyhydroxyalkanoates (PHA), including
polyhydroxybutyrate (PHB), polyhydroxyvalerate (PHV) and 3-
hydroxy-2-methylvalerate (3H2MV) with no detectable glycogen
often associated with carbohydrates. As shown in this gure, the
observed PHA prole exhibited an initial peak, reecting PHA accu-
mulation due to diversion of a fraction of available substrate to
storage; then, it was characterized by a descending curve result-
ing from biochemical consumption of the stored PHA as internal
substrate for microbial growth.
The prole started with a slight initial PHA pool of 35 mg COD/L
incorporated into the biomass seed during the continuous opera-
tion of the ll and draw reactor. As shown in Fig. 1, PHA gradually
increased and reached to a plateau with the rst 2.5 h. A maximum
of 22 mg COD/L PHA was stored. Stored PHA was consumed as inter-
nal substrate and decreased down to the initial pool level within
the next 20 h. Although PHA data is not available after 6 h, the PHA
plateau indicated a clearly denable gradual decrease within next
18 h. PHA storage mainly consisted of PHB (18 mg COD/L) and a
slight PHV fraction (4 mg COD/L). The initial PHA pool consisted of
30 mg COD/L 3H2MV (Fig. 1). It is interesting to note that while
the initial pool essentially composed of 3H2MV (30 mg COD/L), no
detectable accumulation of 3H2MV was observed during the test.
The results indicated that the amount of substrate stored was rel-
atively minimal i.e. less than 5% with respect to the initial level
of 506 mg COD/L selected as the initial level of peptone mixture.
3.1. Impact of tetracycline
Tetracycline was observed to inict a signicant change on the
storage mechanism. As shown in Fig. 2, the modied shape of
the PHA prole signaled a different storage mechanism under the
inuence of the antibiotic. At 50 mg/L dose, the PHA level rapidly
increased to 60 mg COD/L and the increase was continued at a
slower pace to reach a nal plateau of around 90 mg COD/L within
286 T. Katipoglu-Yazan et al. / Biochemical Engineering Journal 91 (2014) 283289
Fig. 2. Storage proles at tetracycline dosing of (a) 50 mg/L and (b) 200 mg/L for
PHA and individual storage components PHB, PHV and 3H2MV.
the rst 4 h. Afterwards, the PHA prole exhibited a slight decrease
of only 26 mg COD/L at the end of the monitoring period of 18 h,
indicating that its utilization as an internal substrate was partially
blocked due to inhibitory action of tetracycline. The same trend was
also observed for the two major components, PHB and 3H2MV; no
appreciable generation could be measured for PHV (Fig. 2a). This
way, 50 mg COD/L of PHA was generated and incorporated into
biomass. Based on storage yield, YSTO of 0.80 mg PHA COD/mg COD,
it could be estimated that 64 mg COD/L, corresponding to 13% of the
peptone mixture was diverted to intracellular storage rather direct
utilization for microbial growth. This level is almost three times
higher than the one associated with the control reactor. Further-
more, at the end of the observation period, 25 mg COD/L of PHA,
equivalent to 32 mg COD/L of peptone mixture initially available
for biodegradation, was not utilized and remained entrapped in
the biomass.
Increase of the tetracycline dose to 200 mg/L reduced the total
PHA accumulation, which remained around 30 mg COD/L, only
slightly higher than the level in the control reactor. However, it fur-
ther decreased the internal utilization of the stored PHA down to
only 5 mg COD/L; at the end of the observation period, 25 mg COD/L
of PHA remained unutilized and still entrapped within the biomass,
as in the case of the lower TET dose. The higher TET dose also
appeared to affect the metabolism of individual storage compo-
nents: As plotted in Fig. 2b, 3H2MV was still the major component
around 30 mg COD/L and a slight generation of PHV was observed
at the expense of a lower PHB formation.
Fig. 3. Storage proles at erythromycin dosing of (a) 50 mg/L and (b) 200 mg/L [33]
for PHA and individual storage components PHB, PHV and 3H2MV.
3.2. Impact of erythromycin
The impact of erythromycin on the storage mechanism, while
equally signicant, was different as compared with tetracycline
in the sense that a much larger portion of the substrate ini-
tially available for microbial growth was diverted to storage: At
50 mg/L dose, the PHA prole exhibited the same rapid increase
to around 50 mg COD/L; however, the increase continued to
reach 119 mg COD/L after 5 h of monitoring. At this level PHA
was composed of 46 mg COD/L of PHB, 33 mg COD/L of 3H2MV
and 40 mg COD/L of PHV (Fig. 3a). Calculations yielded an over-
all PHA accumulation of around 80 mg COD/L, indicating that
100 mg COD/L of the peptone mixture was diverted to stor-
age under the effect of erythromycin. Similar to the observed
effect of tetracycline, erythromycin exerted a retardation effect of
internal utilization of storage products; the slower consumption
resulted in a residual PHA of 46 mg COD/L after 19 h of monitor-
ing, corresponding to 58 mg/L of substrate initially converted to
PHA.
When the erythromycin dose was increased to 200 mg/L, PHA
accumulation basically remained the same as 83 mg COD/L within
the rst 6.7 h. However, the higher dose further decreased the rate
of internal PHA utilization, which was limited to 21 mg COD/L of the
generated PHAs (Fig. 3b). The slower internal utilization increased
the residual PHA to 62 mg COD/L, i.e. 78 mg/L of initially available
substrate converted to PHA remained incorporated into biomass at
the end of the experiment.
 T. Katipoglu-Yazan et al. / Biochemical Engineering Journal 91 (2014) 283289 287

200 Table 2

180 OUR Model coefcients associated with storage assessed by model calibration [33].

160 O2=O2_growth+O2_storage+O2_growth on PHA Experimental runs kSTO (1/day) 

STO (1/day)
OUR (mg O2/L.h)

140 Control 1.30 0.33

120
100 Endogenous respiration level
80
60
40
O2
20
0
-1 1 3 5 7 9 11 13
Time (h)
O2_endogenous respiration
Fig. 4. Evaluation of OUR prole for microbial growth, substrate storage, and
endogenous respiration.
3.3. Theoretical concepts
These results may better be evaluated when supported by mod-
eling and the respirometric data collected together with substrate
storage tests. Respirometry is an extensively used method for the
experimental assessment of biodegradation characteristics of sub-
strates. Obtained oxygen uptake (OUR) prole is interpreted by
modeling to estimate biodegradable substrate COD fractions and
biodegradation kinetics. The area under OUR curve, above endoge-
nous decay level reects the amount of oxygen consumed for
available substrate [39].
In this study, mass balance on PHA storage and microbial
growth were based on overall oxygen utilization, O2 , exclud-
ing endogenous respiration. This term refers to the net amount of
oxygen utilized for growth and storage processes after subtracting
oxygen requirement for endogenous respiration (cells mainte-
nance/respiration processes) as clearly indicated in Fig. 4.
Some studies considered similar material balances of biomass
precursors like NADH2, ATP, AcCoA, substrate, and nutrients as
basis for modeling. However, they focused on production of storage
compounds mainly in pure cultures [42,43]. In this study however,
evaluation of oxygen uptake rate enabled estimation of oxygen con-
sumption for substrate storage and external substrate utilization in
mixed cultures.
Modeling of substrate storage involves two biochemical reac-
tions, namely the storage mechanism and internal use of storage
products for microbial growth. Generation of intracellular storage
products, YSTO is generally dened in terms of the following surface-
saturation type of a rate expression as given in Eq. (1) where SS is
readily biodegradable substrate [20]:
dXSTO SS
= YSTO kSTO XH
dt SS + KSTO
where kSTO is the maximum storage rate; KSTO half saturation coef-
cient for storage, and XH active heterotrophic biomass; a typical
Monod expression is usually adopted for the growth process on
stored products Eq. (2):
dXSTO 1 XSTO /XH
= 
STO XH
dt YH KSTO + XSTO /XH
where YH is the heterotrophic yield coefcient,  STO , the maximum
internal growth rate and KSTO , half saturation coefcient for inter-
nal growth. Calibration of the kinetic expressions dened above,
yielded the set of model coefcients outlined in Table 2, which
provided best ts with the experimental data on substrate stor-
age, as illustrated in Figs. 13. It should be noted that the model
evaluation also covered oxygen uptake rate proles and included
TET 50 2.20 0.20
TET 200 4.00 0.08
ERY 50 3.16 0.22
ERY 200 3.16 0.16
all processes related to organic carbon removal. Detailed analysis
of the modeling exercise is reported elsewhere [33]. Values of the
model coefcients conrm and provide numerical support for the
15 17 observed impact of the antibiotics on the storage mechanism.
Oxygen uptake rate (OUR) proles were also obtained from
parallel batch experiments. It is now well known that the OUR
proles provide two signicant information related to the course
of biochemical reactions taking place as well as to the stoichio-
metric balance between utilization of biodegradable substrate and
consumption of oxygen as the nal electron acceptor [40]: (i) bio-
chemical reactions related to substrate utilization are terminated
when the OUR level drops down to endogenous respiration level;
(ii) the area of the OUR curve above the endogenous respiration
level yields oxygen consumed as related to substrate actually uti-
lized in Eq. (3):
O2 = So = (1 YH )CS (3)
where So is the concentration of oxygen consumed; CS , the con-
centration of biodegradable substrate and YH , the yield coefcient.
When there is simultaneous storage, Eq. (3) has to be modied,
with the assumption that growth on internal substrate proceeds
with the same yield coefcient:
O2 = SO = (1 YSTO YH )CST + (1 YH )CSG (4)
where CST is substrate diverted to storage, CSG , substrate utilized
for microbial growth and YSTO , the storage yield. Eq. (4) basi-
cally applies to the respirometric results of the control reactor.
Since the tested antibiotics inhibited internal utilization of stored
biopolymers, leaving a residual PHA concentration at the end of the
observation period, related mass balance equation should account
for the COD fractions, namely COD stored and internally utilized,
CSTU , substrate COD equivalent to the residual PHA (XSTR ) entrapped
in the biomass, CSTR , and the fraction directly utilized for microbial
growth, CSG (Eq. (5)):
O2 = SO = (1 YSTO YH )CSTU + (1 YSTO )CSTR + (1 YH )CSG (5)
Observation of the OUR proles shows, as illustrated in Fig. 5, a
signicant decrease in the amount of oxygen consumed between
the control reactor and those operated with antibiotic dosing,
despite the fact that they were all started with the same substrate
(1) concentration which should theoretically be available for microbial
uptake. For the control reactor, the mass balance dened by means
of Eq. (4) is satised by a YH of 0.63 mg cell COD/mg COD, a value
quite in agreement with the default value of 0.64 mg cell COD/mg
COD suggested in ASMs [20] and 0.60 mg cell COD/mg COD adopted
for model calibration of the biodegradation of the same peptone
mixture [26,33,44]. This balance also conrms, as expected, com-
(2) plete utilization of the available substrate as plotted in Fig. 5a, also
including the observed substrate storage.
The same balance is also calculated for every experiment con-
ducted with antibiotic dosing, based on Eq. (5); This evaluation
outlined in Table 3 gives a clear indication that the impact of antibi-
otics disturbs the basic stoichiometry for substrate utilization and
identies a residual fraction of the initially available biodegradable
substrate, CSR which remains unutilized, despite the fact the corre-
sponding OUR proles suggest completion of related biochemical
288 T. Katipoglu-Yazan et al. / Biochemical Engineering Journal 91 (2014) 283289

Table 3
Stoichiometry and mass balance between oxygen consumed, stored PHA and substrate utilized for microbial growth.

Experimental runs Antibiotic Stored PHA, XSTO Residual PHA, XSTR Oxygen consumed, Theoretical oxygen Substrate utilized Residual
concentration (mg COD/L) (mg COD/L) SO (mg COD/L) demand, SOT (mg/L) for growth, CSG (mg biodegradable
(mg/L) COD/L) substrate, CSR (mg
COD/L)

Control 22 185 185 463

TET-50 50 50 25 153 175 354 59
TET-200 200 30 25 58 182 132 336
ERY-50 50 81 46 180 182 397 7
ERY-200 200 83 62 150 178 327 75

In all runs initial substrate concentration, CS , was 506 mg.

reactions. In the traditional inhibition concept, the observed COD
difference is explained in terms of uncompetitive inhibition, where a
fraction of the available substrate remains bound with the inhibitor
[17]. It should be noted that substrate binding should now be con-
sidered as a simplistic explanation covering any possible metabolic
impact impairing substrate utilization for growth. The results pre-
sented above also show that enhanced substrate storage is likely to
be one of the important factors inducing lower oxygen consump-
tion.
A clear explanation on the structure of Table 3 should rst be
provided, for properly interpreting the implications of related mass
balance: The rst part of the table reports observed experimental
results, namely stored PHA, XSTO , the residual/unutilized fraction
of the PHA at the end of the experiment, XSTR and the observed
oxygen consumption SO calculated from the OUR proles in par-
allel respirometric experiments. The rst two parameters enable
to calculate COD equivalents of stored PHA fractions, CSTU and CSTR
using the previously adopted value of 0.80 mg PHA COD/mg COD
for the storage yield, YSTO ; the fraction of the initial COD potentially
(a)
200
180 OUR_Total_Data
160
OUR (mg O2/L.h)
available for microbial growth, CSGP , may also be calculated the
same way from mass balance. Following this explanatory note, cal-
culations for the TET-200 experiment for example, would give that
a total amount of 38 mg/L of COD was diverted to storage and only
6 mg/L could be internally utilized (CSTU = 6 mg/L; CSTR = 32 mg/L);
as the initial COD level of the peptone mixture in the experiment
was 506 mg/L, that leaves a CSGP of 468 mg/L that may be poten-
tially utilized for microbial growth. Substituting this information
in Eq. (5) yields three signicant results, also listed in Table 3: (1)
the theoretical oxygen consumption, SOT based on CSTU , CSTR and
CSGP i.e. the expected oxygen consumption if all available substrate
were to be utilized for growth, aside from storage; for TET-200, SOT
was calculated as 182 mg/L, much higher than the observed value
of only 58 mg/L, showing that all available substrate was indeed not
utilized. (2) Substrate actually utilized for growth, CSG , calculated
by using the observed oxygen consumption in Eq. (5); as shown in
the table, the CSG level corresponding to 58 mg/L of O2 consump-
tion was 132 mg/L for the TET-200 run. (3) Residual biodegradable
substrate, CSR , calculated from mass balance (CSGP CSG ), again cal-
culated as 336 mg/L for the TET-200 experiment.
In this context, the stoichiometric evaluation presented in
Table 3 indicated that antibiotic dosing signicantly reduced the
theoretical oxygen consumption corresponding to full utilization
of the substrate fraction available for microbial growth, paral-
lel to storage. The level of residual biodegradable substrate was

140 much higher for tetracycline compared to erythromycin. In fact,

120 for 50 mg/L dose of erythromycin, the slight change in the level
100 of oxygen consumption could be explained solely on the basis of
80 enhanced substrate storage within limits of analytical precision
60 associated with oxygen and COD measurements.
40
20
0 4. Conclusions
-1 1 3 5 7 9 11 13 15 17
Time (h) The major outcome of the experimental results outlined above
was (i) a noticeable increase in the magnitude of the biodegradable
(b) substrate diverted to intracellular storage, and (ii) inhibition and
140
retardation of the utilization of the stored biopolymers for internal
OUR_Total_Data
120 growth so that a residual PHA was always observed at the end of the
experiment. The impact of erythromycin in diverting a higher frac-
OUR (mg O2/L.h)

100
tion of the available substrate toward internal storage was much
80 more pronounced as compared to tetracycline.
Tetracycline and erythromycin signicantly affected substrate
60 storage. While the magnitude of the PHA incorporated into biomass
40 was increased, its internal utilization for microbial growth was
severely impaired. Results suggested that OUR is an essential tool
20 for understanding the impact of antibiotics. Moreover, kinetic and
stoichiometric evaluations on inhibitory impacts should not be
0
limited to microbial growth alone, but they should cover all bio-
-1 1 3 5 7 9 11 13 15 17
chemical processes including substrate storage, where relevant.
Time (h)
The observed decrease in the level of oxygen consumed under the
Fig. 5. OUR proles obtained (a) in the control reactor (b) in the reactor started with acute effect of antibiotics could not be solely attributed to enhanced
200 mg/L tetracycline dose [45]. storage except in the experiment started with 50 mg/L of
 T. Katipoglu-Yazan et al. / Biochemical Engineering Journal 91 (2014) 283289
erythromycin suggesting an additional substrate binding mech-
anism by antibiotics.
Acknowledgments
This study was conducted as a part of the project Evaluation of
the Biodegradation Characteristics and Toxicity/Inhibition Effects
of Selected Antibiotics on Nitrication Systems and supported by
joint doctoral degree program of Turkish Academy of Sciences and
Scientic Research Fund of Istanbul Technical University.
References
[1] M.C.M. van Loosdrecht, M.A. Pot, J.J. Heijnen, Importance of bacterial storage
polymers in bioprocesses, Water Sci. Technol. 35 (1997) 4147.
[2] G.T. Daigger, C.P.L. Grady Jr., The dynamic of microbial growth on soluble sub-
strates: a unifying theory, Water Res. 16 (1982) 368382, Review paper.
[3] C.B. Youssef, A. Zepeda, J. Gomez, Modelling the combined inhibition and time-
dependent inactivation effects of toluene on ammonia and nitrite oxidation
using a nitrifying sludge, Biochem. Eng. J. 80 (2013) 3744.
[4] A. Chiavola, G. Farabegoli, F. Antonetti, Biological treatment of olive mill
wastewater in a sequencing batch reactor, Biochem. Eng. J. 85 (2014) 7178.
[5] J.J. Beun, F. Paletta, M.C.M. van Loosdrecht, J.J. Heijnen, Stoichiometry and kinet-
ics of poly--hydroxybutyrate metabolism in aerobic, slow growing, activated
sludge cultures, Biotechnol. Bioeng. 67 (2000) 379389.
[6] K. Dircks, J.J. Beun, M. van Loosdrecht, J.J. Heijnen, M. Henze, Glycogen
metabolism in aerobic mixed cultures, Biotechnol. Bioeng. 73 (2001) 8594.
[7] F. Carta, J.J. Beun, M.C.M. Van Loosdrecht, J.J. Heijnen, Simultaneous storage and
degradation of PHB and glycogen in activated sludge cultures, Water Res. 35
(2001) 26932701.
[8] A.S. Ciggin, S. Rossetti, M. Majone, D. Orhon, Effect of feeding and sludge age
on acclimated bacterial community and fate of slowly biodegradable substrate,
Bioresour. Technol. 102 (2011) 77947801.
[9] A.S. Ciggin, G. Insel, M. Majone, D. Orhon, Respirometric evaluation and mod-
elling of acetate utilization in sequencing batch reactor under pulse and
continuous feeding, Bioresour. Technol. 107 (2012) 6169.
[10] A.S. Ciggin, D. Orhon, D. Capitani, A. Miccheli, C. Puccetti, M. Majone, Aerobic
metabolism of mixed carbon sources in sequencing batch reactor under pulse
and continuous feeding, Bioresour. Technol. 129 (2013) 118126.
[11] D. Dionisi, M. Majone, V. Papa, M. Beccari, Biodegradable polymers from organic
acids by using activated sludge enriched by aerobic periodic feeding, Bio-
technol. Bioeng. 85 (2004) 569579.
[12] T. Katipoglu, E.U. Cokgor, G. Insel, D. Orhon, Biodegradation kinetics of 2,6-
dihydroxybenzoic acid and peptone mixture by acclimated microbial culture
at low sludge age, J. Environ. Sci. Health. A 45 (2010) 18851891.
[13] I. Pala-Ozkok, D. Orhon, Chronic effect of erythromycin on substrate biodegra-
dation kinetics of activated sludge, Biochem. Eng. J. 81 (2013) 2939.
[14] I. Pala-Ozkok, G. Kor-Bicakci, A. Ural, T. Katipoglu-Yazan, N. Yagc, E. Ubay-
Cokgor, D. Orhon, Acute impact of sulfamethoxazole on the utilization of simple
and complex substrates by fast growing microbial culture, J. Chem. Technol.
Biotechnol. (2013), http://dx.doi.org/10.1002/jctb.4165.
[15] D. Orhon, E.U. Cokgor, G. Insel, O. Karahan, T. Katipoglu, Validity of Monod kinet-
ics at different sludge ages peptone biodegradation under aerobic conditions,
Bioresour. Technol. 100 (2009) 56785686.
[16] A. Mulchandani, J.H.T. Luong, Microbial inhibition kinetics revisited, Enzyme
Microb. Technol. 11 (1989) 6673.
[17] I. Pala-Ozkok, T. Katipoglu-Yazan, E.U. Cokgor, G. Insel, I. Talinli, D. Orhon,
Respirometric assessment of substrate binding by antibiotics in peptone
biodegradation, J. Environ. Sci. Health A 43 (2011) 15881597.
[18] H. Spanjers, P. Vanrolleghem, Respirometry as a tool for rapid characterization
of wastewater and activated sludge, Water Sci. Technol. 31 (1995) 105114.
[19] F.G. Babuna, D. Orhon, E.U. Cokgor, G. Insel, B. Yaprakl, Modelling of activated
sludge for textile wastewaters, Water Sci. Technol. 38 (1998) 917.
[20] M. Henze, W. Gujer, T. Mino, M. van Loosdrecht, Activated sludge models ASM1,
ASM2, ASM2d and ASM3, in: Scientic and Technical Report 9, IWA Publishing,
London, 2000.
289
[21] O. Karahan-Gul, N. Artan, D. Orhon, M. Henze, M.C.M. van Loosdrecht, Experi-
mental assessment of bacterial storage yield, J. Environ. Eng.-ASCE 128 (2002)
10301035.
[22] C. Krishna, M.C.M. van Loosdrecht, Effect of temperature on storage polymers
and settleability of activated sludge, Water Res. 33 (1999) 23742383.
[23] K. Kmmerer, Drugs in the environment: emission of drugs, diagnostic aids and
disinfectants into wastewater by hospitals in relation to other sources, review,
Chemosphere 45 (2001) 957969.
[24] B. Halling-Srensen, G. Sengelv, J. Tjrnelund, Toxicity of tetracyclines
and tetracycline degradation products to environmentally relevant bacteria,
including selected tetracycline-resistant bacteria, Arch. Environ. Contam. Tox-
icol. 42 (2002) 263271.
[25] J.N. Louvet, H. Heluin, G. Attik, D. Dumas, O. Potier, M.N. Pons, Assessment of
erythromycin toxicity on activated sludge via batch experiments and micro-
scopic techniques (epiuorescence and CLSM), Process Biochem. 45 (2010)
17871794.
[26] I. Pala-Ozkok, E.U. Cokgor, Z.P. Cakar, D. Orhon, Acute impact of erythromy-
cin on substrate utilization by activated sludge: effect of sludge age, J. Chem.
Technol. Biotechnol. (2013), http://dx.doi.org/10.1002/jctb.4208.
[27] ISO 8192, Water Quality-Test for Inhibition of Oxygen Consumption by Acti-
vated Sludge, International Standards Organization, Switzerland, 2007.
[28] G. Insel, O. Karahan, S. Ozdemir, I. Pala, T. Katipoglu, E. Ubay-Cokgor, D. Orhon,
Unied basis for the respirometric evaluation of inhibition for activated sludge,
J. Environ. Sci. Health A 41 (2006) 17631780.
[29] E.U. Cokgor, G. Insel, T. Katipoglu, D. Orhon, Biodegradation kinetics of peptone
and 2,6-dihydroxybenzoic acid by acclimated dual microbial culture, Bioresour.
Technol. 102 (2011) 567575.
[30] D. Orhon, S. Sozen, E. Ubay, Assessment of nitrication-denitrication potential
of Istanbul domestic wastewaters, Water Sci. Technol. 30 (1994) 2130.
[31] D.G. Larsson, C. de Pedro, N. Paxeus, Efuent from drug manufactures con-
tains extremely high levels of pharmaceuticals, J. Hazard. Mater. 148 (2007)
751755.
[32] S. Chelliapan, T. Wilby, P. Sallis, Treatment of pharmaceutical wastewater con-
taining tylosin in an anaerobicaerobic reactor system, Water Pract. Technol.
5 (2010) 116, http://dx.doi.org/10.2166/WPT.2010.016.
[33] T. Katipoglu-Yazan, I. Pala-Ozkok, E.U. Cokgor, D. Orhon, Acute impact of eryth-
romycin and tetracycline on the kinetics of nitrication and organic carbon
removal in mixed microbial culture, Bioresour. Technol. 144 (2013) 410419.
[34] A.P.H.A., Standard Methods for the Examination of Water and Wastewater, 21st
ed., 2005, Washington, DC.
[35] ISO 6060, Water Quality-Determination of the Chemical Oxygen Demand,
International Standards Organization, Switzerland, 1986.
[36] G. Insel, D. Orhon, P.A. Vanrolleghem, Identication and modelling of aerobic
hydrolysis mechanism-application of optimal experimental design, J. Chem.
Technol. Biotechnol. 78 (2003) 437445.
[37] P. Reichert, J. Ruchti, W. Simon, AQUASIM 2.0, in: Swiss Federal Institute
for Environmental Science and Technology (EAWAG), 1998, Duebendorf,
Switzerland.
[38] R. Brun, P. Reichert, R.K. Knsch, Practical identiability analysis of large envi-
ronmental simulation models, Water Resour. Res. 37 (2001) 10151030.
[39] D. Orhon, F. Germirli Babuna, O. Karahan, Industrial Wastewater Treatment by
Activated Sludge, IWA Publishing, London, 2009.
[40] I. Pala-Ozkok, A. Rehman, N. Yagci, E. Ubay-Cokgor, D. Jonas, D. Orhon, Charac-
teristics of mixed microbial culture at different sludge ages: effect on variable
kinetics for substrate utilization, Bioresour. Technol. 126 (2012) 274282.
[41] G. Insel, G. Celikyilmaz, E. Ucisik Akkaya, K. Yesiladali, P. Cakar, C. Tamer-
ler, D. Orhon, Respirometric evaluation and modeling of glucose utilization
by Escherichia coli under aerobic and mesophilic cultivation conditions, Bio-
technol. Bioeng. 96 (2007) 94105.
[42] C. Chatzidoukas, G. Penloglou, C. Kiparissides, Development of a structured
dynamic model for the production of polyhydroxybutyrate (PHB) in Azohy-
dromonas lata cultures, Biochem. Eng. J. 71 (2013) 7280.
[43] J.M.L. Dias, L.S. Seram, P.C. Lemos, M.A.M. Reis, R. Oliveira, Mathematical
modelling of a mixed culture cultivation process for the production of poly-
hydroxybutyrate, Biotechnol. Bioeng. 92 (2005) 209222.
[44] T. Katipoglu-Yazan, E.U. Cokgor, G. Insel, D. Orhon, Is ammonication the rate
limiting step for nitrication kinetics? Bioresour. Technol. 114 (2012) 117125.
[45] T. Katipoglu-Yazan, Evaluation of the Biodegradation Characteristics and Tox-
icity/Inhibition Effects of Selected Antibiotics on Nitrication Systems (PhD
thesis), Istanbul Technical University, Turkey, 2013.