Characterization of denitrifying granular sludge with and without the addition of external carbon source

Bioresource Technology 124 (2012) 413–420

Contents lists available at SciVerse ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Characterization of denitrifying granular sludge with and without the addition of external carbon source Beni Lew a,e,⇑, Peter Stief b, Michael Beliavski c, Aviad Ashkenazi c, Olivera Svitlica b, Abid Khan d, Sheldon Tarre c, Dirk de Beer b, Michal Green c a

The Volcani Center, Institute of Agriculture Engineering, P.O. Box 6, Bet Dagan 50250, Israel Max Planck Institute for Marine Microbiology, Celsiusstr. 1, D-28359 Bremen, Germany Faculty of Civil and Environmental Engineering, Technion, Haifa 32000, Israel d Department of Civil Engineering, IIT Roorkee, NH 58, Uttrakhand 247667, India e Department of Civil Engineering, Ariel University Center of Judea and Samaria, Ariel 40700, Israel b c

h i g h l i g h t s " The denitrification rate was five times lower without external electron donor. " The nitrite removal rate was similar with or without external electron donor. " Nitrate without external electron donor led to a slight drop in microprofiles pH. " Nitrite without external electron donor led to an increase in microprofiles pH. " Large increases in pH were observed when acetate was used with nitrate or nitrite.

a r t i c l e

i n f o

Article history: Received 10 May 2012 Received in revised form 10 August 2012 Accepted 12 August 2012 Available online 19 August 2012 Keywords: USB reactor Denitrification Granular sludge Acetate concentration Microsensor profile

a b s t r a c t In this study granular sludge taken from a denitrifying upflow sludge reactor was characterized. Denitrification rates were determined in batch tests with and without external carbon source addition and pH microprofiles of the granules were studied. The microbial community structure was also determined. The results showed that denitrification without carbon source addition occurs; however, the process rate was lower than with external carbon source. This suggests that bacteria use dead biomass and extracellular material in the granular sludge as a carbon source when readily available substrate has been exhausted and nitrate is still present. Microprofiles showed a slight pH decrease for denitrification without external carbon source addition, and an increase in pH when using nitrite as the electron acceptor. Microprofiles using acetate as the carbon source for denitrification showed a significant increase in pH. Clone sequences obtained were close to the species Vitellibacter sp., Denitromonas indolicum str. and Denitromonas aromaticaus sp. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Contamination of groundwater with nitrate is a common phenomenon as a result of fertilizer application and disposal of animal waste. High nitrate concentrations in drinking water, and their subsequent conversion to nitrite and/or N-nitrosamines in the body, may contribute to adverse health effects. Consequently, the European Union has set a maximum concentration of nitrate in drinking water of 50 mg/L and nitrite of 0.1 mg/L (EU, 1998). According to the water framework directive given by the European

⇑ Corresponding author. Address: Agricultural Research Organization, P.O. Box 6, Bet Dagan 50250, Israel. Tel.: +972 39683453; fax: +972 34604704. E-mail address: [email protected] (B. Lew). 0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.08.049

Union, sustained upward trends of groundwater contaminants caused by human activities must be reversed by 2015. There are several options available to solve the problem of high nitrate concentrations in groundwater including improving farming practices, implementation of aquifer protection zones, or blending of affected sources with low-nitrate water supplies. However, these options are often not available within legislative constraints, and further problems may arise related to logistics and/ or cost, thus, the chosen method of water treatment is often the only practical option left for dealing with contaminated aquifers. To remove excess nitrate from groundwater supplies, a range of methods can be used including phytoremediation, ion exchange, distillation and reverse osmosis; however, these methods can be impractical and/or too expensive (Soares, 2000; Van der Hoek et al., 1987).

414

B. Lew et al. / Bioresource Technology 124 (2012) 413–420

Biological denitrification describes the stepwise conversion of nitrate (NO 3 ) to nitrogen gas (N2), Eqs. (1)–(3), and it is commonly used for the removal of nitrates from drinking water.

NO3 þ 2e þ 2Hþ ! NO2 þ H2 O NO2



ð1Þ 

þ

þ 2e þ 2H ! 0:5N2 O þ 0:5H2 O þ OH

0:5N2 O þ e þ Hþ ! 0:5N2 þ 0:5H2 O

ð2Þ ð3Þ

Summarizing the above equations for nitrate reduction to nitrogen gas reveals that five electrons are required and for every mole of nitrate reduced 1 mole of hydroxyl ion is released (Eq. (4)).

NO3 þ 5e þ 5Hþ ! 0:5N2 þ 2H2 O þ OH

ð4Þ

The different reaction steps can be carried out by different microorganisms, such as: Achromobacter, Acinetobacter, Agrobacterium, Alcaligenes, Arthrobacter, Bacillus, Chromobacterium, Corynebacterium, Flavobacterium, Hypomicrobium, Moraxella, Neisseria, Paracoccus, Propionibacterium, Pseudomonas, Rhizobium, Rhodopseudomonas, Spirillum, Vibrio, Thiobacillus denitrificans and Thiomicrospira denitrificans (Gerardi, 2006). Heterotrophic denitrifying bacteria require an organic carbon source for respiration and growth. A wide variety of organic compounds has been used as an external carbon source, such as methanol, ethanol, glucose, acetate, aspartate, or formic acid; as well as different industrial wastes including molasses, whey, distillery stillage and sulfite waste liquor (Cuervo-Lopez et al., 1999; Bernet et al., 1996). However, most of the published research regarding drinking water denitrification involves the use of methanol, ethanol and acetate. The metabolic stoichiometric relationships describing denitrification using methanol, ethanol and acetate are given in Eqs. (5)–(7), respectively (McCarty et al., 1969; Green et al., 1994; Bandpi et al., 1999).

NO3 þ 1:08CH3 OH ! 0:065C 5 H7 O2 N þ 0:47N2 þ 1:44H2 O 

þ 0:76CO2 þ OH

NO3 þ 0:59C2 H5 OH ! 0:09C5 H7 O2 N þ 0:45N2 þ 0:95H2 O þ 0:73CO2 þ OH

ð5Þ ð6Þ

NO3 þ 0:84CH3 COO ! 0:08C5 H7 O2 N þ 0:46N2 þ 0:06H2 O þ 1:3CO2 þ 1:84OH

ð7Þ

As observed in Eqs. (5)–(7) the metabolic stoichiometric equation of denitrification shows a higher production of hydroxyl ions (alkalinity forming product) than carbon dioxide (acidity forming product), causing an increase in pH for these carbon compounds (Santos et al., 2004; Bernet et al., 1996; Istok et al., 2007). Besides external carbon addition, excess sludge disposal can also increase water treatment costs. Granular sludge in upflow sludge blanket (USB) reactors is made up of living and dead cells entrapped in a biofilm matrix. We hypothesize that dead cells in granular biomass can provide a valuable additional carbon source to denitrification, thereby reducing external carbon costs and minimizing sludge production. The catabolic equation of denitrification with dead cells is given in Eq. (8).

NO3 þ 0:25C5 H7 O2 N ! 1:25CO2 þ 0:5N2 þ 0:25NH3 þ OH

denitrifying granules from a USB reactor were studied in batch tests with and without the addition of external carbon to determine substrate conversion and kinetics. pH and NOx microprofiles were measured inside the granules were determined to confirm the results observed in the batch tests. The micro-organisms present in the system were also determined.

ð8Þ

The metabolic stoichiometric equation of denitrification using dead cells shows a production of 1.0 mole of hydroxyl ion (alkalinity forming product) to 1.25 mole of carbon dioxide (and thus acidity formation is higher than accompanying alkalinity leading to an expected decrease in pH. Denitrification with different electron donors has been largely studied and well documented, but only little is known about denitrification fuelled with complex organic carbon compounds from dead cells. The objective of this work is to study the kinetic rates and the effects of denitrification on the water chemistry with and without the addition of external carbon source. To this end,

2. Methods A set of experiments was conducted to determine denitrification fuelled by organic carbon from dead cells in granular sludge from a continuous flow USB denitrifying reactor operated at different retention times of 5 h to 20 min, at around 25 °C for 12 months. The reactor was fed with a simulative brine consisting of tap water at pH 6.5, spiked with NaNO3 (125 mg N-NO3/L), KH2PO4 (5 mg/L), NaCl (7.5 g/L), CaCl22H2O (0.75 g/L), MgCl26H2O (0.75 g/L) and acetic acid as the carbon source. Acetate was added to the feed solution in an amount enough to ensure near zero concentrations of nitrate and acetate in the reactor effluent, 0.86 mole of acetic acid per mole of nitrate. The different retention times applied to the reactor operation led to different influent loading rates, from 2 to 10 g N/day. However, the denitrification efficiency was always higher than 95% for all loads studied. The granules utilized for the experiments were collected from the USB reactor after 10 months of operation, when the sludge was comprised of stable granules with diameter of up to 2.5 mm, organic matter fraction (volatile suspended solids to total suspended solids) of 84.7 ± 3.1% (n = 38) and a low SVI (Sludge Volume Index) value of 18.6 ± 5.5 ml/g (n = 38). 2.1. Denitrification kinetics in batch tests Four batch experiments were carried out in parallel in 500 mL gas-tight bottles with different acetate:nitrate ratio (without acetate (WO), lower than stoichiometric (Low), higher than stoichiometric (High) and at stoichiometric (Equal) concentration according to Eq. (7)), at 25 °C. Acetate was the only external electron donor added to the experiments. Initial acetate and nitrate concentrations of each experiment are shown in Table 1. Twenty milliliters of freshly settled granular sludge from the USB reactor was added to each batch experiment (60 mL for the zero acetate addition). Alkalinity and pH values were measured at the beginning and at the end of the experiments. Alkalinity was measured using the Gran titration technique (Gran, 1952) with H2 CO3 , NHþ 4 and  þ H2 PO 4 as reference species (denoted Alkalinity (H2 CO3 , NH4 , H2 PO )) and taking as well the acetate concentration in the liquid 4 phase into account. Nitrate, nitrite and gas pressure were measured continuously. Nitrate and nitrite were measured using a Metrohm compact ion chromatography unit with a 150 mm column and retention time of 16 min. Nitrogen gas formation was calculated based on the increase in gas pressure in the head space and taking into account gas dissolution in the liquid phase, according to Henry constant. Moreover, to overcome CO2 formation from carbon degradation, which adds to the gas pressure in the head space, NaOH pellets were used in the gas phase to absorb the CO2 produced.

Table 1 Initial acetate and nitrate concentrations in different batch tests. Experiment

Exp. Name

Acetate (mg/L)

Nitrate (mg/L as N)

Without acetate addition Low acetate: nitrate ratio Stoichiometric acetate: nitrate ratio High acetate: nitrate ratio

WO Low Equal High

– 42.9 42.9 42.9

18.09 29.11 21.26 12.30

B. Lew et al. / Bioresource Technology 124 (2012) 413–420

CO2 removal from the system (absorption to NaOH) is necessary in order to calculate the nitrogen gas formation, however, it affects the pH in the system (CO2 removal causes an increase in pH values). pH and alkalinity changes were used for indirect determination of denitrification. For this purpose, duplicate bottles for each batch test were run without NaOH absorption in the gas phase and pH was measured at the beginning and end of the experiments. 2.2. Microprofiles in the biofilm To show denitrification inside the granules, microprofiles of pH and NOx concentrations were measured in the granular biofilms at four different conditions: 2.8 mM nitrate with and without acetate addition (2.4 mM), and 1.5 mM nitrite with and without acetate addition (0.5 mM). Acetate concentration addition was stoichiometric. Individual granules were incubated in flow-cells (Stief and Eller, 2006) and continuously supplied with deoxygenated 1% saline with varying nitrate/nitrite concentrations and stoichiometric amounts of acetate added from a 10-L reservoir. Microsensors for pH (de Beer et al., 1997) and NOx (Larsen et al., 1996) were constructed at MPI Bremen and calibrated before and after profiling. One microsensor at a time (pH or NOx) was attached to the clamp of a micromanipulator that was mounted on a vertical motor sledge (Polerecky et al., 2007). The microsensor tip was positioned exactly at the surface of a granule using a micromanipulator and a dissection microscope. The custom-made programs l-Profiler, DAQ-server, and LINPOS-server were used for automated profile measurements (L. Polerecky, MPI Bremen, http://www.microsenwiki.net). Each microprofile started 500 lm above the granules and continued into the granule to a depth of 1000–2000 lm. Data points were recorded every 25–100 lm, depending on the sensor used. Microprofiles were repeated in 5–10 granules per treatment. All measurements were conducted at 21 °C. 2.3. Microbial community structure of the sludge The community structure was determined using a denaturing gradient gel electrophoresis (DGGE) method, cloning, sequencing and phylogenetic analysis based on 16S rRNA gene sequences. Biomass samples were taken for DNA and RNA extraction in a number of repetitions. Samples were taken after the bioreactor reached steady state (complete nitrate reduction and not less than 10 weeks (70 days) from startup of the experiment) and after 120 days of operation. All the equipment used for sampling was sterilized for 8 h at 180 °C. Samples were frozen at 80 °C until DNA and RNA extraction. 2.3.1. Simultaneous DNA and RNA extraction DNA and RNA were extracted using the phenol/chloroform method according to Hurt et al. (2001) and Ludemann et al. (2000). The RNA sample was treated with DNase (RQ1 RNase-free DNase, Promega, USA) according to the manufacturer’s instructions. Complementary DNA (cDNA) was synthesized from RNA using a commercial kit (Verso, Abgene, UK) according to the manufacturer’s instructions. 2.3.2. PCR and DGGE primers, equipment PCR was carried out in a reaction volume of 25 ll for cloning and 50 ll for DGGE reactions. The reaction volume contained GoTaq Green master mix (including 200 lM dNTP, 1.5 mM MgCl2; Promega, USA), primers (Table 2) and 0.5 ll sample DNA extract as template. The reaction was set in a Biometra thermocycler (Germany). In the PCR reaction designated for sequencing, the

415

same reagents were used except for a fluorescence loading buffer that interrupts the sequencing reaction. Nested PCR was carried out according to the thermocycling parameters described by Sanchez-Melsio et al. (2009). Separation of DNA molecules and PCR products according to size was performed by agarose gel electrophoresis. The agarose concentration was determined according to the fragment size: 1% for DNA and 2% for RNA. PCR products separation was made by loading into 6% polyacrylamide gel with denaturing gradient of urea and formamide between 30% and 63%. The gel was cast in Decode™ system (BioRad, Singapore) between two glass plates. Electrophoresis was performed in buffer 1⁄TAE (biological industry, Israel) with a voltage of 80–100 volt at a constant temperature of 60 °C for 17 h. Following the electrophoresis, the gel was suspended in ethidium bromide solution for half an hour, exposed to UV light and photographed (BioDoc it™ system, UVP, USA). Subsequently the bands were excised for further reamplification and sequencing. The size of the fragments were estimated by comparison to DNA Ladder Lambda/HindIII digested (Fermentas, Lithuania) or 100 bp DNA ladder (Norgen, Canada). The cloning of PCR products was performed using Escherichia coli strain DH5-a with the genotype:recA1, endA1, thi-1, hsdR17, sup44, gryA96, (NalR), relA1 (Hanahan, 1983). Ligation of DNA fragments into a vector was made with the commercial kit pGEMÒ-T easy vector systems (Promega, USA) according to the manufacturer’s instructions. 2.3.3. Sequencing PCR for sequencing was performed by Macrogen (Korea). According to the DGGE analysis, selected clones representing the bioreactor sample were amplified with M13F/R primers (colonies with pGEMÒ-T easy vector) and pJET1.2R/F primers (colonies with pJET 1.2 vector). 2.3.4. Phylogenetic analysis Acquired 16S rRNA gene sequences were compared with the available database sequences via Basic Local Alignment Search Tool (BLAST) search to determine their phylogenetic position (GreenGenes database, http://greengenes.lbl.gov November 2008, supplemented with additional sequences from GenBank). Further analysis was carried out with the ARB software package (Ludwig et al., 2004) with the neighbor joining and maximum parsimony algorithms. Sequences with identity higher than 99% homology were considered the same operational taxonomic unit (OTU). 3. Results and discussion 3.1. Denitrification kinetics in batch tests In all of the experiments, an increase in alkalinity was observed (results not shown) as expected when denitrification occurs. Moreover, the measured and the calculated alkalinity increase (calculated alkalinity based on nitrate consumption using Eq. (7)) were similar (±2% difference) for each experiment. Changes in nitrate, nitrite and nitrogen gas concentration with time in each experiment are shown in Fig. 1. Moreover, for each experiment the theoretical nitrogen gas formation at a given time was calculated based on the initial nitrate concentration minus the nitrate and nitrite concentration at the given time. An almost constant nitrate consumption rate was observed in all experiments (Fig. 1) and calculated according to a zero-order model to be 2.97, 17.88, 17.04 and 16.20 mg NO 3 -N/g VSS/h for the without acetate (WO), lower than stoichiometric (Low), stoichiometric (Equal) and higher than stoichiometric (High) experiments, respectively. The zero-order consumption rate of

416

B. Lew et al. / Bioresource Technology 124 (2012) 413–420

Table 2 Primers used for PCR reaction (⁄forward primer,

++

⁄⁄

reverse primer, +added to eub341F primer for DGGE application).

References

Anneal temp.

Target bacteria

Sequence++

Primer

Muyzer et al. (1993) Muyzer et al. (1995) Rich et al. (2003)

57 °C 57 °C 57 °C

50 -CCTACGGGAGGCAGCAG-30 50 -CGTCAATTCMTTTRAGTTT-30 50 -CGC TGT TCN TCG ACA GYC AG-30

eub341F⁄ univ907R⁄⁄ nosZ-F-1181⁄

Rich et al. (2003)

57 °C

50 -AT GTG CAK NGC RTG GCA GAA-30

pGEMÒ T easy vector system, Promega pGEMÒ T easy vector system, Promega CloneJETTM, Fermemtas CloneJETTM, Fermemtas Muyzer et al. (1993)

50 °C

All bacteria All bacteria Denitrifying bacteria Denitrifying bacteria pGEMÒ-T vector

50 -GTTTTCCCGTCACG-30

nosZ-R1880⁄⁄ M13-F (-40)

50 °C

pGEMÒ-T vector

50 -CAGGAAACAGCTATGAC-30

M13-R⁄⁄

60 °C 60 °C

Clone JET vector Clone JET vector

50 -GACTCACTATAGGGAGAGCGGC-30 50 -GAACATCGATTTTCCATGGCAG-30 50 -CGC CCG GGG CGC GCC CCG GGC GGG GCG GGG GCA CGG GGG G-30

pJET1.2F⁄ pJET1.2R⁄⁄ GC Clamp+

⁄

R = A + G, M = A + C, K = G + T, Y = C + T, N = A + C + G + T.

a

b

c

d

 Fig. 1. Measured nitrate (NO 3 ), nitrite (NO2 ), nitrogen gas (N2), and theoretical nitrogen gas (Theo) concentrations shown as function of time in batch tests, (a) without acetate addition, (b) low, (c) equal, and (d) high initial acetate:nitrate ratio.

nitrate in the WO experiments is around five times lower than in all experiments with acetate addition. This is attributed to the different electron donors in the experiments, i.e., dead cells only in the WO vs. acetate and dead cells in all other experiments. The organic compounds comprising dead cells are of higher complexity and require hydrolysis as opposed to a readily degradable substrate such as acetate. Another explanation would be that a residual amount of acetate from the reactor was available in the WO experiment; however, this assumption was eliminated based on the fact that low amounts of acetate from the reactor were also available in the Low experiment, which showed a similar nitrate consumption rate as the other experiments with acetate addition.

The calculated nitrate degradation rate for the experiments with acetate addition (Low, High and Equal) were very similar and comparable to the rate observed by Henze et al. (1994), indicating that the amount of acetate added to all experiments was used first for complete nitrate reduction to nitrite. Moreover, the rates observed in this study were higher than the ones observed by Isaacs and Henze (1995), for the same conditions, 3 mg NO 3N/g VSS/h. The calculated nitrate degradation rate for the WO experiment was similar to the ones observed for denitrification using hydrolyzed/fermented sludge as electron donor (Moustafa, 2004; Isaacs and Henze, 1995). The nitrite concentration increased in all experiments as long as nitrate was available. When all of the nitrate was converted into

B. Lew et al. / Bioresource Technology 124 (2012) 413–420

nitrite, a constant nitrite consumption was observed in all the experiments and calculated according to a zero-order model to be 3.26, 3.77, 4.66 and 4.77 mg NO 2 -N/h/g VSS, for the WO, Low, Equal and High experiments, respectively. The maximum nitrite reduction rates for the experiments with acetate addition were lower than the ones observed for nitrate reduction and explain the accumulation of nitrite in the experiments. However, for the WO experiment the nitrite consumption rate (after depletion of nitrate) was slightly higher than the nitrate consumption rate. In this case, the nitrite accumulation can be explained by the fact that at the beginning of the experiments nitrate being at the maximum concentration was converted into nitrite at a high rate (zero order); however, nitrite was at a minimum concentration and its consumption proceeded at a lower rate (probably first order). An increase in the zero-order consumption rate of nitrite was observed with an increasing acetate:nitrate ratio, indicating that the electron donor concentration affects the nitrite consumption rate. The concentration of nitrogen gas increased with time in all the experiments. However, the measured amount of nitrogen gas (based on head space pressure) was lower than the theoretical amount of nitrogen gas produced in all the experiments. This difference is probably caused by nitrous oxide (N2O) formation and accumulation. A pH change was observed from the beginning to the end of all experiments. For the experiment without acetate addition (WO) a slight drop in pH was observed, from 7.0 to 6.8. However, all other experiments showed a typical increase in pH, from 7.2 to 7.8 in the Low experiment and from 7.2 to 8.4 in the Equal and High experiments. The observed increase in pH in the experiments with acetate addition and the decrease in the WO experiment are in accordance to the stoichiometry of Eqs. (7 and 8). The lower pH increase observed in the Low experiment vs. in the High and Equal experiments can be explained by the fact that in the Low experiments both acetate and dead cells were used as electron donor; whereas it is very likely that only acetate was used as electron donor for nitrate reduction in the High and Equal experiments.

417

Fig. 2. pH microprofiles in granular biofilms using incubations with nitrate and acetate (grey diamonds), nitrate only (open diamonds; the error bars are so small that they are within the symbol), and nitrite only (closed diamonds).

3.2. Microprofiles in the biofilm pH microprofiles of granular biofilm incubated with nitrate and acetate, nitrate only, and nitrite only are seen in Fig. 2. When acetate was used as the electron donor, a strong increase in pH from 7 to 9 was observed in the denitrifying biofilm as expected. This is due to the much greater alkalinity relative to acidity produced during denitrification using acetate (Eq. (7)). Only a slight pH decrease was observed with biofilm depth for the experiment with nitrate only, from pH 7 to 6.8. This conforms to the smaller increase in acidity relative to alkalinity suggested by the catabolic equation for dead cells (Eq. (8)). However, the slight decrease in pH could also be the result of a small production of CO2 in the granules and no denitrification. It was possible to indirectly show using pH microsensors that the granules do indeed denitrify deep in the biofilm by performing incubations using nitrite instead of nitrate. Nitrite accepts two electrons less than nitrate for complete reduction to N2 leading to less acidity production, 0.75 mole CO2 to 1.0 mole hydroxyl ion in the case of dead cells (Eq. (9)). The pH will thus increase in the denitrifying biofilm using nitrite only in contrast to decrease using nitrate only. NO2 þ 0:15C5 H7 O2 N þ 0:2H2 O ! 0:5N2 þ 0:75CO2 þ 0:15NH3 þ 1:0OH

ð9Þ In incubations with nitrite only, microprofiles did indeed increase in pH with depth as opposed to incubations with nitrate only (Fig. 2). The pH increase in the incubation with nitrite alone,

Fig. 3. DGGE gel of 16S rRNA from the denitrifying USB reactor, (a) PCR products with two repeats (two wells), and (b) gene amplified from DNA (D) and cDNA (synthesized from RNA) extracted from granular biofilms (R).

from pH 7.4 to 8.4, was less pronounced than with acetate addition (results not shown). The lower increase observed for the

418

B. Lew et al. / Bioresource Technology 124 (2012) 413–420

incubation without acetate addition can be explained by fact that the metabolism of dead biomass adds a higher amount of acidity to the system as compared to acetate. This higher acidity production is responsible for the lower pH increase. Jin et al. (2012) also showed very high pH increase in a USB reactor for nitrite denitrification using methanol as the electron donor. Direct proof of granule denitrification by microsensors was shown by NOx microprofiling (results not shown). The NOx concentration decreased with biofilm depth with or without the addition of acetate. The decrease was more pronounced in the experiments with acetate addition and, as previously explained, organic matter from dead cells used as electron donor is probably less easily degradable in comparison to acetate, leading to a lower degradation rate. The results of the batch tests and microsensors profiles show that active bacteria in denitrifying granular sludge can use dead biomass as a carbon source for denitrification. However, batch conditions are significantly different than continuously fed mixed

reactor conditions that typically have low nitrate concentrations. At low nitrate concentrations, only the outer most layers of the biofilm are active with much less access to dead matter found deeper in the granule. In addition, the dosing of external carbon in continuous denitrifying reactor operation by trial and error to attain near zero effluent nitrate concentrations probably takes into account dead cells as part of the overall electron donor demand. While the results suggest that sequential batch reactors inherently benefit more from dead biomass as an additional carbon source, effective use of dead cells in continuous reactor systems could involve the alternate dosing of deficient and sufficient amounts of external carbon source. 3.3. Microbial community structure of the sludge PCR and DGGE analyses performed on replicate samples resulted in the detection of two dominant bands, 18 and 19 (Fig. 3a). These clone sequences are close to the sequence of

Fig. 4. Phylogenetic trees of the bacterial 16S rRNA gene sequences from the USB denitrifying bioreactor. The trees were constructed using Maximum likelihood algorithm. The clones are underlined and the clone name contains the serial clone number and the matching band from Fig. 3a.

B. Lew et al. / Bioresource Technology 124 (2012) 413–420

Vitellibacter sp. from Flavobacteriaceae family, Bacteroidetes phylum (band 18) and Denitromonas indolicum str. and Denitromonas aromaticaus sp. from Rhodocyclaceae family, beta Proteobacteria phylum (band 19). The genus Denitromonas found in this study has been previously found by Etchebehere et al. (2003) in a study to identify the genus responsible for granulation in a denitrifying USB system fed with acetate. However, in the same study, an USB denitrifying system with pure Denitrimonas culture could not granulate. Probably, an interaction between different microorganisms is necessary for granulation, for example, one species can produce the extracellular polymers responsible for another species, adhesion in the growing biofilm. The clone sequences found in the reactor suggest microbial heterogeneity in the denitrifying granular sludge. For both samples taken from the bioreactor, a replicate was performed in an independent and parallel fashion in order to extract and amplify DNA fragments. In Fig. 3a, it can be seen that the replicates are identical except for individual bands that exhibit different band intensity. For example, band 23 (not sequenced) is displayed at higher intensity in the right well than the left well. This difference can be caused by a slightly different composition of microbial populations in different sludge aggregates. DGGE of DNA and cDNA from the bioreactor were conducted in order to determine the active population in the bioreactors (Fig. 3b). cDNA was synthesized from RNA. The main band structure is very much alike between the DNA and cDNA from the same extraction as most of the bands displayed in the DNA well are also displayed in the cDNA well, i.e. most of the populations found in the bioreactors are active. This finding matches the assumption that the selection rate in the bioreactors is high enough to eliminate bacteria that are not adapted to the reactor conditions. As a result of constant wash out of sludge from the bioreactor, these populations eventually are diluted from the bioreactor. In contrast, active populations with a removal rate smaller than their growing rate will remain in the reactor. Only a few bands displayed with a greater relative intensity in the DNA well do not appear in the cDNA well, for example band 21. Probably, these populations exist in large numbers but are not active. Moreover, a few bands were dominant in the cDNA but not in the DNA well, for example band 3 (Methyloversatilis universalis str.). These bands may represent microbial populations with a very high activity, but account for a relatively small portion in the total bioreactor’s bacterial population. Most of the bands which appeared on the gel were sent to sequencing with several repetitions. The sequences were aligned and compared to sequences from the GenBank data base. A phylogenetic tree was constructed (Fig. 4) to show the relationships between the microorganisms detected in this system and microorganisms in other denitrifying reactors. The observed microorganisms in this system were from the Proteobacteria (alpha, beta and gamma Proteobacteria) and Bacteroidetes phylum. In Fig. 4, two clones were affiliated with the alpha subdivision of Proteobacteria and clustered with the Rhodobacter group, clones from band 20 and 21 from Fig. 3a. The clone from band 20 is closely related to Paracoccus sp. str., suggesting that it may belong to this species. Members of the genus Paracoccus are frequently isolated denitrifiers (Mateju et al., 1992; Karanasion et al., 2010). Moreover, the clone from band 21 is related to Rhodovolum sp. and Rhodobacter sp., which have been observed in a denitrifying reactor treating landfill leachate (Etchebehere et al., 2002). Paracoccus sp. and Rhodobacter sp. are well documented denitrifying microorganisms that use acetate as carbon source (Osaka et al., 2006; Ginige et al., 2005). In the gamma subdivision, the clone from band 13 is closely related to Methylophaga sp., which, together with Paracoccus sp. (band 21) have been observed to function at high rate in denitrify-

419

ing reactors with 4% NaCl concentration treating ethanol (Osaka et al., 2008). In the beta subdivision, the clone from band 22 is closely related to Alcaligenes sp., which has also been observed in several denitrifying reactors and also is a well known denitrifying microorganism (Mateju et al., 1992; Karanasion et al., 2010; Etchebehere et al., 2002). The clone from band 3, M. universalis sp. is a well known denitrifying microorganism generally observed in reactors using methanol or ethanol. Moreover, physiological and metabolic responses by M. universalis sp. point to a well-defined strategy to overcome carbon limitation for surviving in engineered or natural denitrifying environments (Baytshtok et al., 2009). 4. Conclusions Based on the results of this investigation, it can be concluded that active bacteria in granular sludge can use dead biomass as a carbon source for denitrification. However, the denitrification rate is slower than with external carbon source, such as acetate, and a slight drop in pH is expected, opposite to the common observation of pH increase in denitrification processes. Microprofiles showed a slight pH decrease in the biofilm using nitrate only. In the case of nitrite without external carbon source, an increase in pH was observed in the microprofiles, although lower than the increase observed when using an external carbon source. Acknowledgements The authors wish to thank the Stephen and Nancy Grand Water Research Institute and the BMBF-MOST German-Israeli Water Technology Research Fund (Grant No. WT0703) for their support. References Bandpi, M.A., Elliott, J., Memeny-Mazdek, A., 1999. Denitrification of groundwater using acetic acid as a carbon source. Water Science and Technology 40 (2), 53– 59. Baytshtok, V., Lu, H., Park, H., Kim, S., Yu, R., Chandran, K., 2009. Impact of varying electron donors on the molecular microbial ecology and biokinetics of methylotrophic denitrifying bacteria. Biotechnology and Bioengineering 102 (6), 1527–1536. de Beer, D., Schramm, A., Santegoeds, C.M., Kühl, M., 1997. A nitrite microsensor for profiling environmental biofilms. Applied Environmental Microbiology 63, 973– 977. Bernet, N., Habouzit, F., Moletta, R., 1996. Use of an industrial effluent as a carbon source for denitrification of a high-strength wastewater. Applied Microbiology Biotechnology 46, 92–97. Cuervo-Lopez, F.M., Martinez, F., Gutierrez-Rojas, M., Noyola, R.A., Gomez, J., 1999. Effect of nitrogen loading rate and carbon source on denitrification and sludge settleability in upflow anaerobic sludge blanket (UASB) reactors. Water Science and Technology 40 (8), 123–130. Etchebehere, C., Errazquin, M.I., Dabert, P., Muxi, L., 2002. Community analysis of a denitrifying reactor treating landfill leachate. Microbiology Ecology 40, 97–106. Etchebehere, C., Cabezas, A., Dabert, P., Muxí, L., 2003. Evolution of the bacterial community during granules formation in denitrifying reactors followed by molecular, culture-independent techniques. Water Science and Technology 48 (6), 75–79. EU, 1998. Council Directive 98/83/EC of 3 November 1998 on the Quality of Water Intended for human consumption. Official Journal of the European Community L330, pp. 0032–0053. Gerardi, M.H., 2006. Wastewater Bacteria. Wiley-Interscience, Hoboken, New Jersey. Gran, G., 1952. Determination of the equivalence point in potentiometric titrations, Part II. Analyst 77, 661–671. Green, M., Loewenthal, R.E., Schnitzer, M., Tarre, S., 1994. Denitrification of drinking water - A bioenergetic evaluation. Water SA 20 (3), 223–230. Ginige, M.P., Keller, J., Blackall, L.L., 2005. Investigation of an acetate-fed denitrifying microbial community by stable isotope probing, full-cycle rRNA analysis, and fluorescent in situ hybridization-microautoradiography. Applied Environmental Microbiology 71, 8683–8691. Hanahan, D., 1983. Studies on transformation of Escherichia coli with plasmids. Journal of Molecular Biology 166, 557–580. Henze, M., Kristensen, G.H., Strube, R., 1994. Rate-capacity characterization of waste-water for nutrient removal processes. Water Science and Technology 29 (7), 101–107.

420

B. Lew et al. / Bioresource Technology 124 (2012) 413–420

Hurt, R.A., Qiu, X.Y., Wu, L.Y., Roh, Y., Palumbo, A.V., Tiedje, J.M., Zhou, J.H., 2001. Simultaneous recovery of RNA and DNA from soils and sediments. Applied Environmental Microbiology 67, 4495–4503. Isaacs, S.H., Henze, M., 1995. Controlled carbon source addition to an alternating nitrification denitrification waste-water treatment process including biological p-removal. Water Research 29 (1), 77–89. Istok, J.D., Park, M.M., Peacock, A.D., Oostrom, M., Wietsma, T.W., 2007. An experimental investigation of nitrogen gas produced during denitrification. Ground Water 45 (4), 461–467. Jin, X., Wang, F., Liu, G., Liu, Y., 2012. Characteristics of denitrifying granular sludge grown on nitrite medium in an upflow sludge blanket (USB) reactor. Water Science and Technology 65 (8), 1420–1427. Karanasion, K.A., Vasiliadou, I.A., Pavlou, S., Vayenas, D.V., 2010. Hydrogenotrophic denitrification of potable water: a review. Journal of Hazardous Material 180, 20–37. Larsen, L.H., Revsbech, N.P., Binnerup, S.J., 1996. A microsensor for nitrate based on immobilized denitrifying bacteria. Applied Environmental Microbiology 62, 1248–1251. Ludemann, H., Arth, I., Liesack, W., 2000. Spatial changes in the bacterial community structure along a vertical oxygen gradient in flooded paddy soil cores. Applied Environmental Microbiology 66, 754–762. Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar Buchner, A., Lai, T., Steppi, S., Jobb, G., Förster, W., Brettske, I., Gerber, S., Ginhart, A.W., Gross, O., Grumann, S., Hermann, S., Jost, R., König, A., Liss, T., Lüßmann, R., May, M., Nonhoff, B., Reichel, B., Strehlow, R., Stamatakis, A., Stuckmann, N., Vilbig, A., Lenke, M., Ludwig, T., Bode, A., Schleifer, K.H., 2004. ARB: a software environment for sequence data. Nucleic Acids Research 32 (4), 1363–1371. McCarty, P.L., Beck, L., St. Amant, P.P., 1969. Biological denitrification of agricultural wastewaters by addition of organic materials. In: Proceedings of the Twenty Fourth Waste Conference, vol. 135. Purdue University Extension Service, pp. 1271–1285. Mateju, V., Cizinska, S., Krejci, J., Janoch, T., 1992. Biological water denitrification- A review. Enzyme Microbial Technology 14, 170–183. Moustafa, M.A.E., 2004. Improving the denitrification potential in biological wastewater treatment by dosing carbon from sludge hydrolysis. Alexandria Engineering Journal 43 (2), 143–151.

Muyzer, G., Dewaal, E.C., Uitterlinden, A.G., 1993. Profiling of complex microbial – populations by denaturing gradient gel-electrophoresis analysis of polymerase chain reaction amplifies genes-coding for 16S ribosomal-RNA. Applied Environmental Microbiology 59, 695–700. Muyzer, G., Ramsing, N.B., 1995. Molecular methods to study the organization of microbial communities. Water Science and Technology 32, 1–9. Osaka, T., Yoshie, S., Tsuneda, S., Hirata, A., 2006. Identification of acetate or methanol assimilating bacteria under nitrate reducing conditions by stable isotope probing. Microbial Ecology 52, 253–266. Osaka, T., Shirotani, K., Yoshie, S., Tsuneda, S., 2008. Effects of carbon source on denitrification efficiency and microbial community structure in a saline wastewater treatment process. Water Research 42, 3709–3718. Polerecky, L., Bachar, A., Schoon, R., Grinstein, M., Jørgensen, B.B., de Beer, D., Jonkers, H.M., 2007. Contribution of Chloroflexus respiration to oxygen cycling in a hypersaline microbial mat from Lake Chiprana, Spain. Environmental Microbiology 9, 2007–2024. Rich, J.J., Heichen, R.S., Bottomley, P.J., Cromack, K., Myrold, D., 2003. Community composition and functioning of denitrifying bacteria from adjacent meadow and forest soils. Applied Environmental Microbiology 69, 5974–5982. Sanchez-Melsio, A., Ciliz, J., Balaguer, M.D., Colprim, J., Vila, X., 2009. Development of batch-culture enrichment coupled to molecular detection for screening of natural and man-made environments in search of anammox bacteria for Nremoval bioreactors systems. Chemosphere 75, 169–179. Santos, S.G., Zaiat, M., Varesche, M.B., Foresti, E., 2004. Comparison of methanol, ethanol, and methane as electron donors for denitrification. Environmental Engineering Science 21 (3), 313–320. Soares, M.I.M., 2000. Biological denitrification of groundwater. Water, Air and Soil Pollution 123, 183–193. Stief, P., Eller, G., 2006. The gut microenvironment of sediment-dwelling Chironomus plumosus larvae as characterised with O2, pH and redox microsensors. Journal of Comparative Physiology B 176, 673–683. Van der Hoek, J.P., Latour, P.J.M., Klapwijk, A., 1987. Denitrification with methanol in the presence of high salt concentrations and at high pH levels. Applied Microbiology and Biotechnology 27, 199–205.