Environmental Science and Pollution Research

, Volume 20, Issue 3, pp 1225–1238

Extent of intracellular storage in single and dual substrate systems under pulse feeding

Authors

    • Faculty of Civil Engineering, Environmental Engineering DepartmentIstanbul Technical University
  • Simona Rossetti
    • Instituto di Ricerca Sulle Acque C.N.R
  • Mauro Majone
    • Department of ChemistrySapienza University of Rome
  • Derin Orhon
    • Faculty of Civil Engineering, Environmental Engineering DepartmentIstanbul Technical University
    • Turkish Academy of Sciences
Review Article

DOI: 10.1007/s11356-012-1291-4

Cite this article as:
Ciggin, A.S., Rossetti, S., Majone, M. et al. Environ Sci Pollut Res (2013) 20: 1225. doi:10.1007/s11356-012-1291-4

Abstract

The study investigated the effect of acetate/starch mixture on the formation of storage biopolymers as compared with the storage mechanism in systems fed with these compounds as single substrates. Experiments involved two sequencing batch reactor sets operated at steady state, at sludge ages of 8 and 2 days, respectively. Each set included three different runs, one fed with acetate, the other with starch and the last one with the acetate/starch mixture. In single substrate systems with pulse feeding, starch was fully converted to glycogen, whereas 25 % of acetate was utilized for direct microbial growth at sludge age of 8 days, together with polyhydroxybutyric acid (PHB) storage. The lower sludge age slightly increased this fraction to 30 %. Feeding of acetate/starch mixture induced a significant increase in acetate utilization for direct microbial growth; acetate fraction converted to PHB dropped down to 58 and 50 % at sludge ages of 8 and 2 days respectively, while starch remained fully converted to glycogen for both operating conditions. Parallel microbiological analyses based on FISH methodology confirmed that the biomass acclimated to the substrate mixture sustained microbial fractions capable of performing both glycogen and PHB storage.

Keywords

AcetateStarchStorageWastewater treatmentFilamentous growthGlycogenPolyhydroxybutyric acid (PHB)

Introduction

Evaluation and control of substrate storage mechanism in mixed microbial cultures require recognition and full understanding of a number of significant issues, essentially including conditions favouring storage, nature and composition of substrate, and also culture history of the microbial community. Engineered systems, such as the activated sludge process, are mostly operated under dynamic conditions due to natural fluctuations in the feed flow, substrate composition, temperature, etc. Moreover, a few modifications of the activated sludge process, like intermittent aeration (Hanhan et al. 2002; 2011) and sequencing batch reactor (Wilderer et al. 2001; Tomei and Annesini 2008), are specifically designed to involve and benefit from transient conditions for improved process efficiency. Dynamic conditions induce a process of physiological adaptation for the microbial community, often leading to substrate storage (Daigger and Grady 1982; Gangurde et al. 2012). Therefore, the recent researches underline the importance of considering substrate storage as an integral component of substrate removal mechanism (Basak et al. 2012). Storage results from an imbalance between substrate removal and growth potential; although substrate is transported into the cell, the biosynthetic growth limitations may prevent consumption of all energy, diverting a fraction of the substrate to storage (van Loosdrecht et al. 1997). Substrate storage has been explained in different ways involving RNA limitation (Grady et al. 1996) or disturbance of the balance between catabolic and anabolic activities (Cortassa et al. 1995; Liao et al. 1996). From a practical standpoint, assessment of storage is quite important in order to evaluate the amount of organic carbon available for significant treatment processes such as denitrification or enhanced biological phosphorus removal (Henze et al. 1995; Mino et al. 1998; Yagci et al. 2011).

The physiological state of microbial cultures is also a function of culture history, which is mainly controlled by the sludge age sustained in biological systems. The sludge age (θX) defines culture history of the enriched microbial community in activated sludge systems (Frigon et al. 2006), and hence, it is expected to exert an impact on metabolic activities related to growth and storage. Studies on the storage mechanism with pure cultures at low θX—i.e. high growth rate—showed that storage was correlated with the difference between the substrate uptake rate and its utilization rate for growth (Van Aalst-van Leeuwen et al. 1997). When the culture was operated at a growth rate close to its maximum substrate uptake rate, storage was observed to be negligible; in this range, the lower the θX, the less substrate was converted into storage polymers (Van Loosdrecht and Heijnen 2002; Majone et al. 2007).

The nature and composition of the organic substrate also play a central role in the storage mechanism as well as the nitrogen and phosphate removal (Hu et al. 2012). The effect of dynamic conditions on substrate storage has been preferentially studied by means of single substrates, mostly acetate or glucose, mainly because the fate of these compounds and generated storage biopolymers—polyhydroxybutyric acid (PHB) and glycogen respectively—could be easily monitored and evaluated for scientific purposes (Beun et al. 2000; Dirks et al. 2001). Most studies conducted with single substrates were related to enhanced biological phosphorus removal focusing on PHB utilization under anaerobic conditions and its conversion to glycogen under subsequent aerobic conditions (Mino et al. 1995; Pereira et al. 1996). Many studies were also devoted to the assessment of the storage yield (Goel et al. 1998a; Martins et al. 2003). Fewer studies were focused on the fate of mixed substrates, but not with conclusive results: In a study carried out under aerobic conditions with a mixture of similar substrates—a mixture of acetic, lactic and propionic acids—a strong decrease in the removal rates of acetic and lactic acids was observed in the presence of another substrate, as compared with systems using each substrate as the only organic carbon source (Dionisi et al. 2002). In case of a mixture of different substrates, i.e. volatile fatty acids and carbohydrates, Carta et al. (2001) reported that no difference was depicted in the uptake rates of acetate and glucose compared to experiments with single substrates. However, in another study conducted with acetate/starch mixture, the individual removal rates of acetate and starch in the mixture were observed to be slightly lower than the ones associated with single substrate experiments (Karahan et al. 2008).

Experiments on intracellular storage using single substrates or mixtures of similar/simple substrates have been quite useful, mostly to clarify the scientific basis of the storage mechanism. However, the results obtained were not relevant for practical applications: Natural or engineered environments sustaining microorganisms do not often involve single substrates but mixtures of carbon and energy sources. In this respect, wastewaters include a number of organic fractions with different biodegradation characteristics (Henze 1992; Henze et al. 1995; Hocaoglu et al. 2011), and results of studies conducted on single substrates run the risk of not providing an accurate image for substrate utilization in real systems. In fact, exploring the utilization mechanisms of substrate mixtures with different biodegradation characteristics is of paramount practical importance for wastewaters, which incorporate a large spectrum of different organics difficult—if not impossible—to identify on an individual basis. In this respect, since chemical oxygen demand (COD) is commonly used as a parameter to reflect the amount of organic pollutants found in water (Zhan et al. 2010), the adoption of the COD as an overall organic substrate parameter has been a major milestone: It enabled to characterize two major groups of organics with different biodegradation characteristics, namely readily biodegradable COD and slowly biodegradable COD, which requires hydrolysis as a first step before microbial utilization (Ekama et al. 1986), aside from non-biodegradable COD fractions and residual microbial products (Germirli et al. 1991; Orhon et al. 1999). In fact, almost all wastewaters are well characterized in terms of different COD fractions; the major COD fraction in domestic sewage is now identified as slowly biodegradable accounting, for 65–75 % of its total COD, together with 10–15 % readily biodegradable COD and 10–15 % non-biodegradable COD (Henze 1992; Orhon et al. 1998; Tas et al. 2009). Therefore, the fate and utilization of substrate mixtures representing the composition of domestic sewage still require extensive investigation to be able to cope with the delicate organic carbon requirements of different microbial processes, such as denitrification, enhanced biological phosphorus removal (Mino et al. 1987).

In this context, the study primarily intended to generate scientific information on the extent of intracellular biopolymer generation from a substrate mixture including readily biodegradable and slowly biodegradable fractions as in most wastewaters: It investigated the effect of using this substrate mixture on the magnitude of observed intracellular storage as compared with the storage mechanism in systems fed with these compounds as single substrates. This information is of vital importance as it would define an accurate basis for a model-based evaluation of substrate utilization mechanism; it is equally significant for establishing an on-line strategy for optimal operation of wastewater systems. Since different storage polymers originate from different metabolites, the study was carried out with two different carbon sources: acetate, which is an important readily biodegradable component in most wastewaters and stored mainly as PHB, and starch, which requires hydrolysis prior to its utilization and stored as glycogen. Accordingly, the substrate mixture consisted of acetate and starch. The experiments were conducted using a sequencing batch reactor (SBR) system offering the most appropriate setup for the envisaged evaluation, and for this reason used in many similar studies (Yagci et al. 2006a; Hocaoglu et al. 2011).

Materials and methods

Experimental design

Three main factors were considered in designing the experimental system in the study: (1) culture history, i.e. the sludge age (θX) of the microbial community; (2) the nature of organic substrate and (3) the feeding regime. A pulse feeding system was selected for all experimental runs to provide optimal conditions for the generation of storage products from external substrate. It is now well known that storage becomes the major metabolic process for systems that can be operated with pulse feeding, which sustains a sequence of feast (presence of external substrate) and famine (absence of external substrate) conditions (Van Loosdrecht et al. 1997)

Culture history of the microbial community was defined in terms of the θX of biological reactors used in the study. Related research has indicated that θX is not only an operational parameter but also an important factor affecting the composition of the microbial community and process kinetics (Ciggin et al. 2012a). Two different θX values of 8 and 2 days have been selected for the experiments; the former (θX = 8 days) approximates the level that is commonly adopted for biological treatment systems operated for organic carbon removal. The second value (θX = 2 days) reflects high rate operation. This combination was also adopted in similar studies (Ciggin et al. 2011; Dionisi et al. 2002).

Experiments were conducted on three different organic carbon sources. Substrate differentiation and selection was made according to major developments in the last 30 years evaluating the organic carbon content of wastewaters in terms of COD fractions with different biodegradation characteristics. In this respect, acetate and starch were selected as single substrates. They are markedly different from one another because acetate is extensively utilized in similar studies as a readily biodegradable compound (Yagci et al. 2011; Dionisi et al. 2002), whereas starch is a complex substrate that first undergoes hydrolysis before microbial utilization (San Pedro et al. 1994; Karahan et al. 2006). The acetate/starch mixture served as the dual substrate, mainly for exploring the behaviour and fate of individual components as compared with single substrate systems.

The experiments involved two sets of SBR operation, one at a θX of 8 days and the other at a much lower θX of 2 days. Each set included three different runs, one with acetate, the other with starch and the last one with the acetate/starch mixture. Observations were conducted on three full cycles after each SBR operation reached steady state, yielding identical /repeating concentration profiles for all monitored parameters. The experiments required sequential operation of the SBR system with an initial start-up phase before steady state so that the microbial culture was fully acclimated to the selected organic carbon source. Dual substrate experiments were purposefully conducted with the same overall initial COD—i.e. 50 % reduced single substrate concentrations—to approximately sustain the same F/M ratio and OLR as in SBRs fed with single substrates. During each SBR operation, cyclic concentration profiles of the two components—i.e. acetate and starch—of the substrate mixture were individually monitored, together with the corresponding storage biopolymers—i.e. PHB and glycogen—to allow comparison with the results of single substrate experiments. The operational characteristics of the experimental system are summarized in Table 1, where each measurement reflects the average value of the 3 cycles monitored, with a variation of less than 5 %, compatible with the level analytical accuracy.
Table 1

Operating conditions of the SBR systems

Run no.

SBR 1

SBR 2

SBR 3

SBR 4

SBR 5

SBR 6

Carbon source

Acetate

Starch

Dual substrate

Acetate

Starch

Dual substrate

Design θX (days)

8

8

8

2

2

2

OLR (mg COD/L day)

1,200

1,216

1,212

1,440

1,602

1,608

Biomass (mg VSS/L)

1,479

2,000

2,380

845

1,300

1,410

F/M (mg COD/mg VSS)

0.81

0.61

0.51

1.70

1.23

1.14

Feast phase (min)

15

10

7

35

15

15

COD removal (%)

89

92

90

95

90

85

Experimental setup

The activated sludge inocula obtained from the ISKI Pasakoy Advanced Biological Wastewater Treatment Plant, in Istanbul, Turkey were acclimated to different carbon sources in laboratory-scale sequencing batch reactors (SBRs) at different θX of 8 and 2 days. The SBR operation was designed to run 6 cycles per day with 4-h cycles. Operation of each cycle basically included a 10 min of idle phase, 5 min of nutrient feeding phase, 165 min of aerobic reaction phase, 30 min of the settling and finally 30 min for effluent withdrawal. At the beginning of the reaction phase, the organic carbon source was fed within 1 min. The θX of the SBRs were maintained by discharging mixed liquor during the last 2 min of the reaction phase, with due consideration of the amount of the volatile suspended solid (VSS) which is lost with the supernatant during effluent discharge period. The total reactor volume was adjusted to 2.0 L during reaction phase by adding 1.0 L of carbon and nutrient sources to the 1.0 L of the stationary volume (V0) holding settled biomass.

The necessary micronutrients were supplied by means of the micronutrient solution with a composition of 120 g/L NH4Cl, 160 g/L KH2PO4, 320 g/L K2HPO4, 15 g/L MgSO4.7H2O, 0.5 g/L FeSO4.7H2O, 2 g/L CaCl2.7H2O, 0.5 g/L MnSO4.H2O and 0.5 g/L ZnSO4.7H2O. The level of the nutrient supply was adjusted to be in excess with respect to metabolic requirements of the biomass so that the organic substrate remains to be the rate-limiting parameter. The carbon sources were prepared as sodium acetate trihydrate (CH3COONa.3H2O) and soluble starch stock solutions in distilled water according to the desired organic loading rate (OLR). For the mixed substrate experiments, these carbon sources were approximately 50 % diluted in order to feed biomass with same OLR with single substrate experiments.

Analytical methods

The SBRs were controlled at pH 7.5 ± 0.5 at the 25 °C with the dissolved oxygen (DO) concentration higher than 2 mg/L during the reaction phase. DO, pH and temperature were monitored with an ADI 1010 bio-controller and BioXpert software. Samples taken for COD and acetate analyses were filtered through 0.45-μm PVDF syringe filters. COD samples were preserved with H2SO4 and acetate samples with 10 M H3PO4. The COD analysis was carried out in accordance with the ISO6060 (1986). Acetate was analysed by gas chromatograph (Agilent 6890N). VSS analyses were conducted by using the procedure defined in Standard Methods (1995).

The PHB content of biomass was determined by implementing extraction, hydrolization and esterification after the lyophilisation process (Ciggin et al. 2009). The extracted propyl esters were analysed by gas chromatograph (Agilent 6890N). Starch samples were subjected to the acidic hydrolysis to glucose by sulphuric acid, and then the glucose content of the samples was measured with a high-pressure liquid chromatograph. Glycogen samples were analysed similarly, in accordance with the method described by Smolders et al. (1994). When necessary, analytical results were converted into COD units by using the correspondent oxidation stoichiometry: 1.067 mg COD/mg acetate, 1.60 mg COD/mg PHB, 1.067 mg COD/mg glucose, 1.185 mg COD/mg glycogen and 1.42 mg COD/mg VSS.

FISH analysis

Fluorescence in situ hybridization (FISH) analyses were performed to detect dominant bacterial groups depending on the carbon source and/or sludge age after the observation of the steady-state conditions in each SBR. The details on oligonucleotide probes specific to main phyla are available at the probeBase (Loy et al. 2007). The properties of the group specific oligonucleotide probes used for FISH experiments were given in Table 2. All hybridizations with group-specific probes were carried out simultaneously with probe EUB338mix for the detection of most bacteria and with DAPI staining for quantifying the total number of cells. Probes were synthesized with FITC and Cy3 labels purchased from MWG AG Biotech (Germany). The filamentous bacteria were identified according the morphological peculiarities and phylogenetic affiliation estimated by FISH as described in Levantesi et al. (2004).
Table 2

Properties of the oligonucleotide probes

Probe name

Specificity

Formaldehyde concentration (%)

References

EUBmix (EUB338 I + II + III)

Most Bacteria

20–35

Amann et al. 1990;

Daims et al. 1999

ALF968

Alphaproteobacteria

20

Neef 1997

BET42a

Betaproteobacteria

35

Manz et al. 1992

GAM42a

Gammaproteobacteria

35

Manz et al. 1992

HGC69A

Actinobacteria

25

Roller et al. 1994

LGC354mix

Firmicutes

35

Meier et al. 1999

THAU832

Thauera spp.

35

Loy et al. 2005

AZA645

Most members of the Azoarcus

35

Hess et al. 1997

CF319a

Cluster most Flavobacteria, some Bacteroidetes, some Sphingobacteria

35

Manz et al. 1996

CFX1223

Phylum Chloroflexi (green nonsulfur bacteria)

35

Björnsson et al. 2002

Results and discussion

Single substrate experiments

Selection of single organic carbon sources allowed independent monitoring of respective storage products, as acetate was stored as PHB and starch was stored as glycogen. Also, acetate and starch could be measured along with the analytical assessment of the overall COD.

SBR operation at θx = 8 days

Major experimental results obtained at θx = 8 days are summarized in Table 3. As shown in this table, the initial concentration of both single substrates was adjusted to around 200 mg COD/L. For both substrates, pulse feeding sustained a storage pool that could be predicted by mass balance between the formation rate and internal utilization of the stored organics. At steady state, the level of PHB pool associated with acetate feeding was 128 mg COD/L, whereas starch induced a much higher glycogen pool of around 416 mg COD/L. Within each cycle, the pool exhibited a fluctuation—i.e. initial accumulation followed by gradual depletion—due to substrate diverted to storage.
Table 3

Experimental results of SBR operation a θX of 8 days

Run no.

SBR 1

SBR 2

SBR 3

Carbon source

Acetate

Starch

Acetate

Starch

Experimental θX (days)

7.78

8.04

8.05

8.05

Initial COD (mg COD/L)

200

203

100

102

Biomass (mg VSS/L)

1,479

2,000

2,380

2,380

Feast phase (min)

15

10

7

7

PHB pool (mgCOD/L)

199 ± 9

270 ± 11

Glycogen pool (mg COD/L)

416 ± 15

313 ± 19

Amount of stored PHB (mg COD/L)

128 ± 4

49 ± 2

Amount of stored glycogen (mg COD/L)

170 ± 7

85 ± 3

COD used for storage (%)

75

100

58

100

COD used for growth (%)

25

42

qS (mg COD/g COD h)

381

426

254

259

qP (mg COD/g COD h)

257

355

123

218

qP/−qS

0.68

0.83

0.48

0.84

In case of acetate feeding as a single carbon source, the amount of PHB storage in each cycle was increased with an average level of 128 mg COD/L, corresponding to 64 % to the initial substrate concentration in the pulse feeding. Often, PHB generation is also assessed using the ratio between the rate of storage formation, (qP), and the rate of substrate utilization, (−qS). In this study, the qP/−qS ratio, which is the ratio of storage biopolymers generated to the initial substrate available in the experimental setup, was computed as 0.68 mg COD/mg COD as given in Table 3. These values are in close agreement with similar results reported in the literature: Beccari et al. (1998) investigated the influence of storage on population dynamics using a CSTR operated at a θX of 3 days with intermittent acetate feeding, i.e. 2 min pulse feeding at the beginning of every cycle, and they monitored storage in batch reactors. Under selected operating conditions, the PHB/acetate ratio was found to be 0.69 mg COD/mg COD. Majone et al. (1996) similarly tested an intermittently fed CSTR system sustained at the same θX of 3 days, with aerobic/anoxic batch reactors where the PHB/acetate ratio was found as 0.70 mg COD/mg COD. The corresponding qP/−qS ratio was 0.68 mg COD/mg COD. Beun et al. (2000) investigated storage in an SBR system involving pulse feeding, and they reported a PHB/acetate ratio of 0.69 mg COD/mg COD at a θX of 10 days, which practically remained the same (0.70 mg COD/mg COD) when the θX was increased to 20 days. Carta et al. (2001) studied simultaneous utilization of acetate and glucose in an SBR system sustained at steady state at θX of 6.1 days, with pulse feeding within 10–13 min of the cycle; they were able to observe a slightly variable PHB/acetate ratio, which increased from 0.56 to 0.66 mg COD/mg COD within consecutive cycles at steady state. Dirks et al. (2001) run batch experiments with pulse addition to activated sludge taken from full-scale nutrient removal treatment plants. They found that the rate of substrate utilization and PHB formation was much slower in real wastewater treatment systems. The experiments yielded a PHB/acetate ratio of 0.69 mg COD/mg COD for the plant operated at θX of 4 days and 0.73 mg COD/mg COD for θX of 21 days. They concluded that 90 % of microbial growth occurred in the famine period, and the glycogen-corrected value for the observed heterotrophic yield for the feast phase was around YHN = 0.05–0.07 mg cell COD/mg COD. Beun et al. (2002) performed a similar SBR study with pulse feeding of acetate at θX of 4 days and found a PHB/acetate ratio of 0.70 mg COD/mg COD confirming previous results. They also conducted an excellent literature evaluation which led them to conclude that the PHB/acetate ratio is a constant value of 0.68 mg COD/mg COD (0.6 Cmol/Cmol) for aerobic systems, irrespective of the specific microbial growth rate.

Accurate evaluation of the results above, i.e. the level of acetate diverted to PHB formation, requires recognition of energetics and basic stoichiometry of PHB storage and the corresponding value of the storage yield coefficient, YSTO. This issue was first considered together with the key role of storage compounds like PHB and glycogen in the metabolism of microorganisms performing enhanced biological phosphorus (EBPR) removal (Mino et al. 1987). In EBPR systems, the storage mechanism occurs under anaerobic conditions. In this context, Mino et al. (1995) calculated the corresponding yield coefficient, YSA as 0.70 mg COD/mg COD from theoretical/modelling considerations. Later, Yagci et al. (2004) experimentally determined the YSA value as 0.77 mg PHACOD/mg COD acetate for an EBPR system sustaining both phosphorus-accumulating organisms (PAOs) and glycogen-accumulating organisms (GAOs). The same group suggested that YSA would vary within a narrow range of 0.72–0.78 mg COD/mg COD on the basis of a metabolic model that accounted for the activities of PAOs and GAOs (Yagci et al. 2006b).

Similarly, for the aerobic storage of acetate, Van Aalst van Leeuwen et al. (1997) suggested a YSTO value of 0.73 mg COD/mg COD for pure cultures of Paracoccus pantotrophus. Beccari et al. (1998) stated that around 20 % energy was needed for storage, which implied that YSTO should be taken as 0.80 mg COD/mg COD. Beun et al. (2000) assumed a value of 0.70 mg COD/mg COD (0.62 Cmol/Cmol) based on metabolic modelling considerations. The group used the carbon mole unit for the evaluations; this unit does not reflect the oxidation state of the organic compound evaluated. In the modelling of substrate biodegradation, chemical oxygen demand replaced this unit as it established an electron balance between substrate utilized, oxygen consumed and biomass generated (Ekama et al. 1986). Activated Sludge Model No.3 suggested a higher YSTO of 0.85 mg COD/mg COD as a default value for domestic sewage. Accurate understanding of the process stoichiometry was achieved with the introduction of respirometric analyses where the area under the oxygen uptake rate profile obtained from a batch experiment yields the oxygen equivalent of substrate utilized. Karahan-Gul et al. (2002a) developed an experimental oxygen uptake rate (OUR) method where YSTO was determined as 0.76 mg COD/mg COD for acetate. The same group also tested acetate, glucose, acetate/glucose mixture and acetate/domestic sewage mixture for the assessment of the corresponding YSTO coefficients; the study yielded a YSTO range of 0.75–0.82 mg COD/mg COD with an average value of 0.78 mg COD/mg COD for acetate (Karahan-Gul et al. 2002b). With a similar approach, YSTO for acetate was calculated as 0.76 mg COD/mg COD using the generated OUR profiles (Karahan et al. 2008). As respirometric evaluations relied on the assumption of full conversion of acetate into PHB, these values need to be slightly corrected to 0.80–0.84 mg COD/mg COD range to account for the minor acetate fraction which is directly utilized for microbial growth.

In the earlier part of the study, a YSTO value of 0.84 mg COD/mg COD was calculated for acetate, on the basis of experimental OUR profiles obtained under different operating conditions (Ciggin et al. 2012b), in compliance with the range suggested in the literature. This YSTO value enables to compute that around 75 % (150 mg COD/L) of the available acetate was diverted to PHB storage, and the remaining 25 % was utilized directly for microbial growth.

Starch feeding as a single organic source was designed to reflect, as previously mentioned, the fate of slowly biodegradable substrate. The work of Goel et al. (1998a) is significant for the understanding of the hydrolysis mechanism and the utilization of slowly biodegradable substrate: They promoted starch as a model compound for slowly biodegradable substrate and underlined the difficulty of separately evaluating hydrolysis and storage, based on OUR measurements. This difficulty is currently overcome, as in this study, because storage products can be directly assessed on an individual basis. The same group previously run an experiment with an anaerobic/aerobic SBR operated at a θX of 10 days under pulse feeding, which consisted of glucose, peptone, yeast extract and acetate mixture (Goel et al. 1998b). Only 36 % of the initial starch concentration in the batch reactor was converted to glycogen as the system was not acclimated to starch, a common problem for batch tests. Karahan et al. (2006) operated an SBR system at a θX of 5 days with pulse feeding within 3 min at the beginning of each cycle. Two different types of starch, namely NPS (36 % soluble) and SolS (97 % soluble), were used as the organic carbon source in the experiments. The results indicated that adsorption was the initial mechanism for starch removal followed by fast enzymatic hydrolysis within the flocs. Maltose was the predominant hydrolysis product, aside from glucose and fructose. The glycogen/substrate ratio was found as 0.89 mg COD/mg COD for maltose, 0.80 mg COD/mg COD for SolS and a slightly lower value of 0.75 mg COD/mg COD for NPS.

In this study, the average amount of glycogen stored in each cycle was 170 mg COD/L, representing around 84 % of the starch introduced in the SBR by pulse feeding. As previously mentioned, storage requires energy, consuming a fraction of the available substrate for this purpose, so that a storage yield, YSTO, is involved depending upon specific substrate and storage mechanism (Goel et al. 1998a). Karahan-Gul et al. (2002a) found a YSTO of 0.87 mg COD/mg COD, a value very close to the glycogen/maltose ratio of 0.89 previously mentioned. Later, Karahan et al. (2008) evaluated YSTO of starch as 0.91 mg COD/mg COD much like glucose and other simple sugars. The fraction of 84 % observed in this study is quite close to the reported YSTO range of 0.87–0.91 mg COD/mg COD, clearly indicating that glycogen storage may be considered as the major mechanism of starch removal with no appreciable utilization for direct microbial growth under the selected operating conditions.

SBR operation at θx = 2 days

SBR experiments at a much lower θX of 2 days were started with relatively higher initial substrate concentrations of 240 mg COD/L for acetate and 267 mg COD/L for starch, to partly compensate lower biomass levels sustained at the selected θX; therefore, the food to microorganism (F/M) ratios were almost double the values associated with experiments at higher θX, as indicated in Table 1. Yet, the extent of storage was observed to remain slightly lower for both substrates: Starch fraction converted to glycogen at the beginning of each cycle was measured as 80 %, still indicating almost full storage considering the corresponding storage yield, aside from a small fraction of around 10 % that may be directly utilized for microbial growth after hydrolysis. It should be noted that glycogen measurements after acidic hydrolysis do not allow being more precise because they do not differentiate stored glycogen from intracellular and extracellular/adsorbed carbohydrates.

The amount of PHB generated based on acetate utilization was 143 mg COD/L, yielding a PHB/acetate ratio of 0.59 mg COD/mg COD; the corresponding qP/−qS ratio was computed as 0.58 mg COD/mg COD, almost identical to the observed PHB/acetate ratio (Table 4). These values agree well with similar results in the literature indicating that growth rate of the microbial culture affects the level of storage: Van Aalst van Leeuwen et al. (1997) investigated the behaviour of pure cultures of P. pantotrophus using chemostat/batch systems. In the experiments, the θX was changed in the range of 0.80–3.20 days, and the corresponding PHB/acetate ratio was observed to change between 0.20 and 0.33 mg COD/mg COD, i.e. only 25 to 40 % of acetate was converted to PHB storage. Krishna and van Loosdrecht (1999) run SBR experiments at a θX of 2.5 days, at steady state for different temperatures, which obviously affected the specific growth rate of the microbial culture: While the PHB/acetate ratio was 0.68 mg COD/mg COD at 15 and 20 °C, it dropped to 0.56 mg COD/mg COD at 25 °C, 0.22 mg COD/mg COD at 30 °C and 0.13 mg COD/mg COD at 43 °C. Similarly, Beun et al. (2000) reported a PHB storage ratio of 0.46 mg COD/mg COD for an SBR system sustained at a θX of 3 days. In this context, the observed values in this study showed that the amount of acetate directly introduced to metabolic reactions for growth was slightly increased to 30 %, compared with 25 % at θx = 8 days.
Table 4

Experimental results of SBR operation at θX of 2 days

Run no.

SBR 4

SBR 5

SBR 6

Carbon source

Acetate

Starch

Acetate

Starch

Experimental θX (days)

2.09

2.07

2.10

2.10

Initial COD (mg COD/L)

240

267

128

140

Biomass (mg VSS/L)

845

1,300

1,410

1,410

Feast phase (min)

35

15

15

15

PHB pool (mg COD/L)

159 ± 6

221 ± 8

Glycogen pool (mg COD/L)

282 ± 11

262 ± 14

Amount of stored PHB (mg COD/L)

143 ± 7

56 ± 2

Amount of stored glycogen (mg COD/L)

213 ± 8

111 ± 7

COD used for storage (%)

70

100

50

100

COD used for growth (%)

30

50

qS (mg COD/g COD h)

343

579

256

280

qP (mg COD/g COD h)

199

460

116

224

qP/−qS

0.58

0.80

0.45

0.80

Cyclic concentration profiles of all significant parameters for both acetate and starch selected for two different θX are illustrated in Figs. 1 and 2. Aside from the results summarized above, they show that while single substrates (acetate and starch/glucose) were totally removed at the start of the cycle, there was always a residual COD level, which should be attributed to soluble microbial products’ generation, as experienced in many similar studies (Orhon et al. 1999; Ciggin et al. 2011).
https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1291-4/MediaObjects/11356_2012_1291_Fig1_HTML.gif
Fig. 1

Cyclic PHB, acetate and COD profiles for pulse feeding of acetate at aθx = 8 days and bθx = 2 days

https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1291-4/MediaObjects/11356_2012_1291_Fig2_HTML.gif
Fig. 2

Cyclic glycogen, glucose and COD profiles for pulse feeding of starch at aθx = 8 days and bθx = 2 days

Experiments with acetate/starch mixture

The experimental results obtained with the acetate/starch mixture are outlined in Tables 3 and 4 and plotted in Figs. 3 and 4. They showed that the removal and fate of starch as a substrate fraction in the dual substrate feeding remained exactly the same as in the SBRs fed with starch as a single substrate: The amount of glycogen accumulation at the beginning of the cycles was as high as 80–83 % of the initial starch concentration in the pulse feeding, and this implied again almost full storage at both θX, considering the corresponding storage yield for starch.
https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1291-4/MediaObjects/11356_2012_1291_Fig3_HTML.gif
Fig. 3

Cyclic concentration profiles for pulse feeding of acetate/starch mixture at θx = 8 days; a glycogen, glucose and COD profiles; b PHB and acetate profiles

https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1291-4/MediaObjects/11356_2012_1291_Fig4_HTML.gif
Fig. 4

Cyclic concentration profiles for pulse feeding of acetate/starch mixture at θx = 2 days; a glycogen, glucose and COD profiles; b PHB and acetate profiles

However, dual substrate feeding significantly affected acetate utilization: At θx of 8 days, the average cyclic PHB accumulation was measured as 49 mg COD/L; using the same YSTO value of 0.84 mg COD/mg COD, around 58 % of acetate fed at the beginning of the cycle was utilized for internal storage based on the observed PHB accumulation; this indicated that the remaining 42 % was directly utilized for microbial growth, a level significantly higher than the 25 % level associated with SBR operation using acetate as the sole organic carbon source. At θx of 2 days, the ratio of acetate in the substrate mixture directly utilized for microbial growth exhibited an additional increase to reach 50 %; this level should be compared to 30 % obtained with the single substrate SBRs using acetate.

As previously mentioned, the majority of similar studies were conducted with mixtures of simple substrates (Carta et al. 2001; Dionisi et al. 2002; Karahan-Gul et al. 2002b); the results obtained do not set a scientific analogy for wastewaters mainly because the selected simple compounds are all considered as a combined single entity within the readily biodegradable COD fraction. Karahan et al. (2008) tested the same acetate/starch mixture for the generation of storage biopolymers, without running parallel single substrate studies. The storage mechanism was also investigated using a peptone–yeast extract mixture as the organic carbon source, in a study conducted for testing the validity of Monod kinetics at different θX (Orhon et al. 2009). The peptone/yeast extract mixture served as the standard substrate in many studies, mainly because it approximates the biodegradation characteristics of domestic sewage with a 9 % readily biodegradable, 42 % readily hydrolysable and 47 % slowly hydrolysable COD fractions. The results yielded a limited PHB storage of around 30 mg COD/L corresponding to only 60 % of the available readily biodegradable COD at θX of 8 days; PHB storage was reduced by 50 % at a lower θX of 2 days. Despite its importance, generation of storage products has not been experimentally evaluated for domestic sewage. The concept of substrate storage was only introduced in the modelling of tannery wastewaters, with a volatile fatty acid content of 820 mg COD/L including an acetate fraction of 74 % (Dizdaroglu-Risvanoglu et al. 2007); model calibration yielded a storage yield (YSTO) value of 0.83 mg COD/mg COD, quite compatible with the results of this study, and a storage rate coefficient (kSTO) much higher than the specific heterotrophic growth rate (μH) at the two food to microorganism (F/M) ratios of 0.07 and 0.2 mg COD/mg VSS tested in the experiments.

Microbiological observations

The bacterial composition of SBRs acclimated under different operating conditions were estimated by FISH. The majority of bacteria were identified with oligonucleotide probes, which are specific to main phyla within the Bacteria domain. The main result gathered from FISH analyses was the observation of all microbial groups in dual substrate systems, which were detected as dominant species in SBRs fed with different carbon sources. This observation indicated that the consumption of acetate and starch as well as the storage of PHB and glycogen could be attributed to different microbial groups. The distribution of phyla in total bacteria is given in Fig. 5.
https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1291-4/MediaObjects/11356_2012_1291_Fig5_HTML.gif
Fig. 5

Distribution of the detected phylum in each SBR at aθx = 8 days and bθx = 2 days

As shown in Fig. 5, Betaproteobacteria, Alphaproteobacteria and Gammaproteobacteria were detected as main phyla in the SBRs fed with acetate, although the SBRs fed with starch were mainly dominated by the Actinobacteria. Significant differences in bacterial composition were observed in SBRs operated at different θx with acetate (SBR1 and SBR4), while almost same bacterial compositions were observed in SBRs fed with starch independently from the θx. The dominant phyla detected in SBRs fed with different carbon sources agree with the results of previous studies observing Betaproteobacteria as a main species responsible for PHB storage (Dionisi et al. 2005) and Actinobacteria as responsible for starch hydrolysis (Xia et al. 2008) as well as the removal of slowly degradable compounds (Lew et al. 2011). Figure 6 gives the FISH images, which provide a visual support for the dominance of both Betaproteobacteria and Actinobacteria in the SBR system fed with the acetate/starch mixture, each one representing the dominant phylum in similar SBR experiments conducted with acetate and starch as single carbon sources.
https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1291-4/MediaObjects/11356_2012_1291_Fig6_HTML.gif
Fig. 6

Dominant bacterial species detected by FISH: aBetaproteobacteria in acetate fed SBR, bActinobacteria in starch fed SBR, cBetaproteobacteria and dActinobacteria in SBR fed with acetate/starch mixture (Bar is 10 μm)

The main filamentous bacteria were investigated by microscopic analysis as a result of the bulking problems encountered in SBR4, where the filamentous bacteria were observed in all flocs, and their abundance was scored as “dominant” according to Jenkins et al. (2004). The microscopic analysis showed marked differences in the ratio of floc-formers/filamentous bacteria in systems fed with different carbon sources as illustrated in Fig. 7. The higher amount of filamentous bacteria observed in SBR fed with acetate at the θx of 2 days is in good accord with bulking problems usually observed in systems operated at lower θX, whereas no bulking problems were experienced at θx of 8 days. On the other hand, an opposite trend was observed in SBRs fed with starch. Bulking in SBR fed with starch at θX of 8 days can be explained by the fact that substrate adsorbed onto biomass becomes available through hydrolysis; substrate limitation due to the slower rate of hydrolysis would favour preferential growth of filamentous microorganisms. When substrate is taken up at low concentrations from the bulk solution, this would lead to gradients through the floc giving advantage to filamentous organisms which extend outside the floc. In contrast, the observation of less filamentous bacteria in the SBR unit fed with starch at the lower θX was explained with the uniformly distribution of the hydrolysis products inside the flocs (Martins et al. 2010).
https://static-content.springer.com/image/art%3A10.1007%2Fs11356-012-1291-4/MediaObjects/11356_2012_1291_Fig7_HTML.gif
Fig. 7

Filamentous bacteria in each SBR at aθx = 8 days and bθx = 2 days

Evaluation of results

The evaluation of the results presented in the previous sections provided clear indication that the nature and the extent of intracellular storage become specifically related to the components, i.e. readily biodegradable and slowly biodegradable fractions, in the substrate mixture. Therefore, they cannot be predicted by means of single substrate experiments. Supporting observations can be summarized as follows:

Dual substrate feeding significantly increased the level of the storage pool. At θx = 8 days, the total storage (PHB + glycogen) induced by the substrate mixture was 583 mg COD/L as compared to 199 mg COD/L of PHB and 416 mg COD/L of glycogen in single substrate systems started with the same COD level. The relative increase in both the PHB and glycogen pools may also be visualized by the corresponding storage pool/initial substrate ratios: The PHB pool/acetate ratio of 1.0 and the glycogen pool/starch ratio of 2.08 in SBRs fed with single substrates were increased to 2.7 and 3.06, respectively, for dual substrate feeding at θx of 8 days. Mass balance dictates that a pool develops on the basis of a difference between the formation and the utilization rates of the stored organics. Thus, the results indicated a wider rate difference or, more precisely, a slower internal utilization rate of storage biopolymers in the case of mixed substrate feeding. SBR operation at θx of 2 days yielded almost the same results as the glycogen pool was only reduced to 46 % and the PHB pool to 69 % of the levels sustained at SBR operation with single substrates, despite around two times decrease in acetate and starch concentrations. The results basically showed that the same process kinetics remained valid—i.e. the same lower internal utilization rates—to induce a compatible if not identical relative increase in the PHB and glycogen pools, regardless of the selected θX.

The magnitude of the cyclic generation/accumulation of storage products induced by dual substrate feeding was quite different compared to single substrate systems: The level of storage associated with the acetate/starch mixture was always higher than PHB formation from acetate feeding and lower than glycogen formation from starch feeding. In fact, at θX = 8 days, the observed total storage of 134 mg COD/L generated by the dual substrate mixture remained between 128 mg COD/L of PHB and 170 mg COD/L glycogen from the same initial levels of acetate and starch feedings. A similar result was also observed for θX = 2 days, as outlined in Table 4.

The composition of the overall storage was quite variable as a function of the components in the acetate/starch mixture. At θX = 8 days, while the dual substrate mixture included equal amounts of acetate (102 mg COD/L) and starch (100 mg COD/L), the accumulated glycogen was almost twice the amount of PHB; at θX = 2 days, the pulse feeding of 268 mg COD/L with similarly balanced acetate and starch fractions generated 111 mg COD/L of glycogen as compared to 56 mg COD/L of PHB. The observed difference between PHB and glycogen was sustained mainly because a much larger fraction of acetate was directly utilized for microbial growth in the presence of starch with relatively slower biodegradation characteristics. The results strongly suggest that the balance of substrate utilization between microbial growth and storage for a selected culture history mainly depends on metabolic availability, i.e. relative magnitude and biodegradation characteristics of the substrate fractions in the mixture. The metabolic balance preferentially selects the more available substrate fraction—i.e. acetate/the readily biodegradable fraction as in this study—for direct microbial growth, even under pulse feeding conditions, which inherently favour substrate storage. This balance is bound to be quite different and obviously distorted in single substrate systems.

Substrate requirements obviously vary the basis of the microbial community sustained SBR system under the selected operating conditions. The results of the study also indicated that the nature and composition of the microbial community exhibited significant changes for dual substrate feeding, while conserving biomass fractions responsible for PHB and glycogen storage. In this context, the effect of microbial community structure on substrate selection for intracellular storage certainly deserves additional research work.

Conclusions

The results in this study suggest that the dual substrate feeding induced utilization of a significantly higher fraction of acetate—the readily biodegradable component of the mixture—for direct microbial growth, while starch—the slowly biodegradable component of the mixture—remained almost fully converted to storage. The increase in direct utilization of acetate is compatible with the fact that corresponding metabolic reactions are adjusted to the rate of protein synthesis that can be sustained for microbial growth at the selected sludge age, and they require a relatively higher level of acetate as it only provides a fraction of the organic carbon available in the substrate mixture. This observation also explains the relatively higher storage pools—i.e. lower internal utilization rates induced by a dual substrate environment, in agreement with higher utilization of external substrate.

In this respect, the results also indicated that single substrate experiments bear the risk to mislead substrate utilization mechanism, when extrapolated for real systems—i.e. wastewaters—with a similar readily/slowly biodegradable substrate composition, as tested in the study. They showed that interaction between removal potential of individual substrates should be evaluated together if the wastewater treatment plant is to be designed for the treatment of more than one type of wastewater source, especially for the industrial wastewater treatment plants. The sludge age did not have a noticeable effect on acetate/starch utilization, either as single substrates or in the mixture, other than a slight increase in direct utilization of acetate for growth, again explainable by the requirements of the protein synthesis system under different growth conditions. FISH analyses confirmed the simultaneous utilization of different carbon sources as the microbial community sustained with the substrate mixture included dominant species observed in single substrate experiments. The results obtained strongly suggest that significant biochemical/enzymatic reactions be emphasized in the following studies to further clarify the metabolic aspects of dual substrate metabolism.

Acknowledgments

Research work of Aslı S. Ciggin, conducted as part of this study at the Sapienza University of Rome, was supported by the grant of The Scientific and Technological Research Council of Turkey (Tubitak).

Copyright information

© Springer-Verlag Berlin Heidelberg 2012