conditions in the SBR (feast famine regime) caused fast removal of most COD in the anoxic phase (in particular acetate, ethanol, glutamic acid and glucose were ...
D. Dionisi*, C.Levantesi**, V. Renzi*,V. Tandoi** and M. Majone* * Department of Chemistry, University of Rome “La Sapienza”, p.le A. Moro 5, 00185 Rome, Italy ** Water Research Institute, National Research Council, via Reno1, 00198 Rome, Italy Abstract An activated sludge was cultivated on a mixture of several soluble substrates (acetate, ethanol, glucose, glutamic acid, peptone, Tween 80, starch, yeast extract) in an anoxic/aerobic SBR. Highly dynamic conditions in the SBR (feast famine regime) caused fast removal of most COD in the anoxic phase (in particular acetate, ethanol, glutamic acid and glucose were totally removed) and relevant contribution of storage. In spite of that, filament abundance was always high, as is typical of bulking sludges. Filaments which developed in the reactor were characterized on a morphological basis and on the basis of their ability to grow and to store polyhydroxyalkanoates (PHAs). Three main filaments prevailed in the biocenosis, whose relative abundance was varyng with time: Nostocoida limicola II, (two different morphological types), Haliscomenobacter hydrossis and an unidentified one. It was found that maximum growth rate was higher for flocformers than for filaments on each of the tested substrates. Epifluorescence showed that storage ability was more widespread among flocformers than in the filaments. Only one type of Nostocoida limicola II was able to store PHAs. The obtained data show that aerobic growth on the little residual fraction of COD from the anoxic phase was enough to support high abundance of filamentous microorganisms. Keywords Bulking; denitrification; Haliscomenobacter hydrossis; Nostocoida limicola II; PHAs; storage
Introduction
In several plant configurations, the biomass grows under transient (unsteady) conditions, even though the overall process can be considered under steady-state conditions. Transient conditions are typical for configurations where a concentration gradient of the substrate is produced (like plug-flow configuration of the aeration tanks) or, to a greater extent, where biomass experiences alternately (vs. time or flow direction) high and low substrate concentrations, as in sequencing batch reactors, contact-stabilization processes or processes with a selector for bulking control. More importantly, transient conditions are also present in nutrient removal processes where the biomass experiences alternating anaerobic, anoxic or aerobic environments and a metabolic stress is added to the kinetic one (Daigger and Grady, 1982). It is widely accepted that under these dynamic conditions growth becomes unbalanced, i.e. substrate can be removed without a corresponding regular increase in every cellular component (with no or little synthesis of active biomass): in particular intracellular storage of the substrate (i.e. substrate uptake and related synthesis of internal polymers) can become important (Majone et al., 1999). Therefore, storage can play a dominant role in substrate uptake in nitrogen removal processes based on predenitrification, particularly when an external source of readily biodegradable COD (RBCOD) is added to increase the denitrification rate (Isaacs et al., 1994; Hasselbad and Hallin, 1996). Storage can be a key-factor in microbial competition and therefore in bulking control. As a matter of fact, highly dynamic processes are reported to be effective in avoiding proliferation of filaments, because floc-formers are reported to have in general higher storage capacity than filaments (Grau et al., 1982). Under anoxic conditions a metabolic pressure is added to the kinetic one, because most filamentous microorganisms are reported not to be
Water Science and Technology Vol 46 No 1–2 pp 337–344 © IWA Publishing 2002
PHA storage from several substrates by different morphological types in an anoxic/aerobic SBR
337
D. Dionisi et al.
able to denitrify. In spite of highly dynamic conditions and of the metabolic pressure, nitrogen removal processes often suffer for bulking problems (Wanner, 1993). Till now literature evidence of storage is mainly focused on acetate or glucose, under aerobic conditions, and each substrate being the only carbon source. On the contrary, wastewater is a complex mixture of several substrates (Henze et al., 1994) and in many plants the carbon source is firstly taken up by biomass under anoxic conditions. Therefore the aim of the present paper is to study the behaviour of activated sludge in a nitrogen removal SBR in the presence of a mixture of several substrates (low and high molecular weight), with main reference to the role of poly-hydroxyalkanoate (PHA) storage and its effect on population dynamics. Methods SBR operation
The SBR was operated with 4 cycle per day (6 h/cycle, first 2 h 30 min anoxic, remaining 3 h 30 min aerobic, feed 1 min starting 30 min after the beginning of the cycle, in order to assure anoxic conditions during the feed, Figure 1). The mixed liquor was withdrawn just before the new feed, without any settling phase; so hydraulic retention time and sludge age were the same (6 d) and well controlled independently from settling properties of the sludge (Table 1). The particular arrangement was also chosen in order to study population dynamics in the sludge as resulting only from biological competition for the substrate between flocformers and filaments, with no additional effect due to different settling properties. Due to high HRT, feed COD was quite high (3,800 mg/L), in order to apply the chosen organic load rate (OLR 633 mgCOD/Ld, Table 1). The carbon sources in the feed (Table 2) were several soluble compounds representing different classes of organic substances in a range of molecular-weights (MW): volatile fatty acids (acetate), sugars (glucose and starch), alcohols (ethanol), proteins (glutamic acids, peptone and yeast extract) and long chain fatty acids (Tween 80). Thiourea and nitrate were added in the feed to inhibit nitrification and to provide nitrate for denitrification, respectively. Mineral medium composition is reported in Majone et al. (2001). During the last period of SBR operation the feed pattern was changed: only 1 cycle per day was made (the first 8 hours were anoxic, the remaining period was aerobic). All the other operating parameters remained unchanged. During operation, total suspended solids (TSS), volatile suspended solids (VSS), sludge volume index (SVI) were measured three times per week by sampling the mixed liquor at the end of the aerobic phase. Once a week soluble COD, acetate, glutamic acid, ethanol and glucose were determined, by sampling the mixed liquor at the end of both anoxic and aerobic phase. The analytical procedure was reported elsewhere (Majone et al., 2001). Microbial identification
Once a week the SBR biomass was characterized by microscopic observation (bright field, phase contrast and epifluorescency), after Gram, Neisser and Nile Blue staining. Filament morphological identification and filament abundance were done according to Eikelboom and van Buijsen (1983) and Jenkins et al. (1993). FISH analysis was performed according to Amann et al. (1995), using the probe ALF1b (Manz et al., 1992) for detecting the α-subclass of Proteobacteria and the eubacterial probe EUB338 (Amann et al., 1990). The ALF1b and EUB338, labelled respectively at the 5’ end with CY3 and fluorescein, were purchased from MWG Biotech, Germany. A Gram post-staining was performed according to Jenkins et al. (1993). Batch tests 338
In order to investigate the role of each substrate as the single carbon source, two types of
Anoxic Nitrogen bubbling
phase (2h 30 or 8h 30)
Feed (30 min after phase start, 1 min long)
Aerobic phase (3h 30 or 15h 30) Aeration
Sludge and effluent withdrawal (2 min long)
Figure 1 Operating cycle of the SBR D. Dionisi et al.
Table 1 SBR operating parameters and average performance Operating parameter
Value
Hydraulic retention time and sludge age (d) Influent COD (mg/L) Organic load rate (CODmg/L d) Cycle length (min) Feed length (min) pH Temperature (ºC)
6 3,800 633 360 1 7.5 25
Average performance
Value
Biomass (mgVSS/l) Residual COD (mg/l) (anoxic phase) Residual COD (mg/l) (aerobic phase) Overall COD removed (%) Fraction (%) of the overall COD removal
1044 275 247 93 82 18
anoxic phase aerobic phase
Table 2 Composition of the SBR feed Carbon sources
Sodium acetate Glucose Ethanol L-glutamic acid Soluble starch Peptone Yeast extract Tween 80
mg/l
mgCOD/l
Electron acceptor
mg/l
782 225 115 247 1,031 731 240 120
600 240 240 240 1,200 800 240 240
Sodium nitrate
2,731
Nitrification inhibitor Thiourea
20
batch tests were performed. In order to evaluate the relative ability of the filamentous and flocforming biomass to grow under excess substrate condition, the first type of batch test (low biomass inoculum, approximately 1 mgVSS/l) was performed with each substrate, under anoxic (nitrate) and aerobic conditions. During this type of test, the mixed liquor was sampled and slides were taken at the end of the test for microscope observation. In order to investigate the role of storage, the second type of batch tests (high sludge inoculum, approximately 250 mgVSS/l) was performed with each of the four low-MW well defined substrates (acetate, ethanol, glutamic acid and glucose), in the presence of either nitrate, nitrite or oxygen. Storage was quantified during aerobic or anoxic (nitrate) tests, where the mixed liquor was regularly sampled for substrate, PHA, carbohydrates and ammonia analysis, according to the methodology reported elsewhere (Majone et al., 2001). Moreover, in order to investigate the ability of individual morphotypes to store PHAs the mixed liquor was sampled and slides were taken at the beginning and at the end of the test for epifluorescence microscope observation.
339
Results and discussion SBR operation
D. Dionisi et al.
The SBR was operated for a long time under a typical feast-famine regime (feed in 1 minute over a 6 hour cycle). In particular, the four low-MW substrates that were analytically determined (acetate, ethanol, glucose and glutamic acid) were completely removed during the first part of the anoxic phase. Also for other substrates, most removal was in the anoxic phase (Table 1). It is also noteworthy that in spite of high removal (93%), the residual COD was still quite high (247 mgCOD/l at the end of the anoxic phase): this residual COD was apparently “non biodegradable” under the chosen OLR and sludge age. During SBR operation, SVI was usually high, even though highly variable (Figure 2). Three filamentous morphotypes were prevailing in the biomass, whose relative abundance was varying with time (Figure 3): Nostocoida limicola II (two different morphological types), Haliscomenobacter hydrossis and an unidentified filament that did not correspond to any of the Eikelboom types. Morphological characterization of the main observed morphotypes is reported in Table 3. One type of Nostocoida limicola II (named type A) was Gram variable and Neisser positive, with filament length usually > 200 µm, while type B was both Gram and Neisser negative, with filament length usually < 100 µm. The unidentified filamentous bacterium was straight or smoothly curved, of variable length (100–300 µm), diameter 1 µm, composed of cells of variable shape (square, rectangular, barrel), Gram and Neisser negative and gave negative response to the S test. Besides filamentous organisms, many tetrads forming Gram positive bacteria were detected in the sludge, morphologically differing from the G bacteria isolated and characterized by Blackall et al. (1997); in fact Fish analysis showed that tetrad-forming bacteria did not belong to the α-subclass. A slight improvement was observed in the last period of the SBR run when the feed pattern was modified to 1 cycle/day: the SVI profile and the abundances of the various filaments showed a slight and slow tendency to decrease after the change of operating conditions. Storage and growth in batch tests
Growth during substrate removal was investigated starting from a very small inoculum (batch tests at 1 mgVSS/l) under aerobic and anoxic (nitrate) conditions with each of the eight substrates of the SBR feed as single carbon source. Once a significant amount of biomass had been formed, filaments or tetrad-forming bacteria were not detectable with any substrate tested. These data show that floc-forming bacteria have a higher maximum growth rate than filamentous bacteria with each one of the substrates in the feed, as usually reported in the literature (Wanner, 1993). Therefore, the growth of filaments was probably 1200
6
5
filament abundance
800 SVI (ml/g)
1 cycle per day
SVI
4
600
3
400
2
200
1
0
0 0
50
100
150
200
days after inoculum
340
Figure 2 SVI and overall filament abundance during SBR runs
250
300
350
filament abundance
4 cycles per day
1000
6 N ostocoida lim icola II (A and B) H aliscom enobacter hydrossis unidentified
4 cycles per day
1 cycle per day
4
3
D. Dionisi et al.
filament abundance
5
2
1
0 0
50
100
150
200
250
300
350
d ays a f t e r i n o c u l u m
Figure 3 Individual filament abundance during SBR runs Table 3 Filament morphology Morphology
Nostocoida limicola II type A Nostocoida limicola II type B Haliscomenobacter hydrossis Unidentified filament Tetrads forming
length > 200 µm, diameter 1.8–2 µm, Gram variable, Neisser +ve lenght < 100 µm, diameter 1.8–2 µm, Gram –ve, Neisser –ve Identical to Jenkins et al. (1993) variable lenght (100–300 µm), diameter 1 µm, Gram –ve, Neisser –ve, S test –ve Gram +ve, Neisser +ve
supported by aerobic utilization of residual slowly biodegradable substrate: indeed, even though all substrates were soluble, 28 mgCOD/l apparently behaved as slowly biodegradable, so remaining available in the aerobic phase at low concentrations. PHA formation during substrate removal was also investigated (batch tests at about 250 mgVSS/l) for the four low-MW well defined substrates (acetate, ethanol, glucose and glutamic acid, each one as the only carbon source). Storage phenomena (PHA or glycogen) exerted a significant role (both under anoxic and aerobic tests) for acetate, ethanol and glucose (Table 4). No significant PHA or glycogen storage was observed from glutamic acid, even though an intense fluorescence after Nile Blue staining was developed in the sludge. The presence of fluorescence could indicate storage products of lipidic nature other than PHA, as already hypothesized for pure cultures grown on acetate (Fixter et al., 1986; Tandoi et al., 1998). On the other hand, storage of glutamic acid in form of poly-aminoacid has been reported under anaerobic conditions (Satoh et al., 1998) and would be coherent with combined carbon and nitrogen balance obtained in the present study (Majone et al., 2001). However, further research is clearly needed to ascertain whether glutamic acid can be actually stored and eventually in which form. As reported in Figure 4 for a typical anoxic batch test with ethanol, the fluorescence microscopy after Nile Blue staining made it also possible to observe which morphological types of bacteria where able to store PHB. In general, fluorescence was present in the largest amount inside the flocs under all conditions (including in the presence of nitrite as the only electron acceptor). Among the five morphological types which were individuated (Table 5), only Nostocoida limicola II (type B) was able to store PHA (from acetate and ethanol) or other compounds of lipidic nature fluorescent under Nile Blue staining (from glutamic acid). The evidence of storage by Nostocoida limicola II (type B) was found in the presence of oxygen or nitrate while it was not in the presence of nitrite. Specific FISH analysis directed to the identification of the Nostocoida limicola II morphotype B has been
341
Table 4 Storage contribution to substrate removal and solid formation
storage on removed substrate ratio • PHA • Carbohydrate storage on solid formation ratio
Acetate
Ethanol
Glucose
Glutamic acid
0.30 0 0.46
0.34 0 0.48
0 0.73 0.98
0 0 0
D. Dionisi et al. Figure 4 Bright field and fluorescence photo of a Nile Blue stained slide taken during an anoxic batch test with ethanol (three out of the five described morphological types are visible) Table 5 Fluorescence for the different morphological types Acetate aerobic
Nostocoida limicola II type A Nostocoida limicola II type B Haliscomenobacter hydrossis Unidentified filament Tetrads forming Floc formers
342
Ethanol
anoxic
anoxic
with
aerobic
Glutamic acid
anoxic
anoxic
with
with
nitrate
nitrite
–
–
–
+
+
– – – +
– – – +
aerobic
Glucose
anoxic
anoxic
every
with
with
with
tested
nitrate
nitrite
nitrate
nitrite
condition
–
–
–
–
–
–
–
–
+
+
–
+
+
–
–
– – – +
– – – +
– – – +
– – – +
– – – +
– – – +
– – – +
– – – –
done. This filamentous bacterium, differently from Nostocoida limicola II recently isolated from conventional wastewater treatment plants (Actinomycetes subphylum, Blackall et al., 2000), belongs to the α-subclass of Proteobacteria. The results of FISH analysis are reported in Figure 5 (a, b).
a
b
D. Dionisi et al.
Figure 5 a) FISH analysis using eubacterial probe EUB338 b) FISH analysis using specific probe ALF1B for the α subclass of Proteobacteria
Conclusions
The presented data show that filamentous microorganisms can survive also in highly dynamic systems. Even though the system exerted a strong pressure, both kinetic and metabolic (the removed COD was mainly consumed in the anoxic phase), three different filaments (two types of Nostocoida limicola II, Haliscomenobacter hydrossis, and an unidentified one) grew at significant abundance (at typical values of bulking sludge). Neither storage response nor fast growth were probably key factors in filament survival under the studied conditions. As a matter of fact, while storage exerted an important role in substrate removal and storage ability was widespread among flocformers, only one of the described filamentous organisms was able to store PHA (α Nostocoida limicola II, also in anoxic conditions). None of the filaments could grow faster than floc-formers with any of the different substrates. The little amount of COD that was degraded aerobically was apparently enough to allow filament growth in the described system. In particular, the occurrence of Halicomenobacter hydrossis, which is usually considered an “oxic zone grower”, was confirming such evidence. When the feed pattern was modified from 4 to 1 cycle per day, the SVI and abundance of main filamentous morphotypes showed a slight and slow tendency to decrease. However, such an evidence has to be confirmed by more experimental work. Acknowledgements
The authors are pleased to acknowledge Dr. Simona Rossetti, for FISH analysis and for the critical discussion. References Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R. and Stahl, D.A. (1990). Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial popuklation. Appl. Environm. Microbiol., 56, 1919–1925. Amann, R.I., Ludwig, W. and Schleifer, K.H. (1995). Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol. Rev., 59, 143–169. Blackall, L.L., Rossetti, S., Christensson, C., Cunningham, M., Hartman, P., Hugenholtz, P. and Tandoi, V. (1997). The characterization and description of representatives of “G” bacteria from activated sludge plants. Lett. Appl. Microbiol., 25, 63–69. Blackall, L.L., Seviour, E.M., Bradford, D., Rossetti, S., Tandoi, V. and Seviour, R.J. (2000). “Candidatus Nostocoida limicola”, a filamentous bacterium from activated sludge. Int. J. Syst. Bacteriol., 47, 479–491. Daigger, G.T. and Grady, Jr, C.P.L. (1982). Review paper. The dynamic of microbial growth on soluble substrates. A unifying theory. Wat. Res., 16, 368–382. 343
D. Dionisi et al. 344
Eikelboom, D.H. and van Buijsen, H.J.J. (1983). Microscopic Sludge Investigation Manual, 2nd edn., The Netherlands: TNO, Delft. Fixter, L.M., Nagi, M.N., McCormack, J.G. and Fewson, C.A. (1986). Structure, distribution and function of wax esters in Acinetobacter calcoaceticus. J. Gen. Microb., 132, 3147–3157. Grau, P., Chudoba, J. and Dohànyos, J. (1982). Theory and practice of accumulation-regeneration approach to control of activated sludge filamentous bulking, in: Bulking of Activated Sludge. Preventive and Remedial Methods. Chambers, B. and Tomlinson, J. (eds.), Ellis Horward Publ., Chichester, pp. 111–127. Hasselbad, S.and Hallin, S. (1996). Intermittent dosage of ethanol in a pre-denitrifying activated sludge process. Wat. Sci. Tech., 34(1–2), 387–389. Henze, M., Kristensen, G.H. and Strube, R. (1994). Rate-capacity characterization of wastewater for nutrient removal processes. Wat. Sci. Tech., 29(7), 101–107. Isaacs, S.H., Henze, M., Søeberg, H. and Kümmel, M. (1994). External carbon source addition as a means to control an activated sludge nutrient removal process. Wat. Res., 28(3), 511–520. Jenkins, D., Richard, M.G. and Daigger, G.T. (1993). Manual on the Causes and Control of Activated Sludge Bulking and Foaming, 2nd edn. Lewis Publishers, Michigan. Majone, M., Dircks, K. and Beun, J.J (1999). Aerobic storage under dynamic conditions in activated sludge processes. The state of the art. Wat. Sci. Tech., 39(1), 61–73. Majone, M., Beccari, M., Dionisi, D., Levantesi, C. and Renzi, V. (2001). Role of storage phenomena on removal of different substrates during pre-denitrification. Wat. Sci. Tech. 43(3), 151–158. Manz, W., Amann, R.I., Ludwig, W., Wagner, M. and Schleifer, K.H. (1992). Phylogenetic oligodeoxynucleotide probes for the major subclasses of proteobacteria: problems and solutions. Syst. Appl. Microbiol., 15, 593–600. Satoh, H., Mino, T. and Matsuo, T. (1998). Anaerobic uptake of glutamate and aspartate by enhanced biological phosphorus removal activated sludge. Wat. Sci. Tech., 37(4–5), 579–582. Tandoi, V., Majone, M., May, J. and Ramadori, R. (1998). The behaviour of polyphosphate accumulating Acinetobacter isolates in an anaerobic-aerobic chemostat. Wat. Res., 32(10), 2903–2912. Wanner, J. (1993). Activated Sludge Bulking and Foaming Control. Technomic Publishing Co., Inc., Lancaster, USA.
