ARTICLE IN PRESS

Bioresource Technology xxx (2005) xxxxxx

Comparison of long-term performances and nal microbial compositions of anaerobic reactors treating landll leachate
B. Calli
a b

a,*

, B. Mertoglu a, K. Roest b, B. Inanc

Department of Environmental Engineering, Marmara University, 34722 Goztepe, Istanbul, Turkey Laboratory of Microbiology, Wageningen University, H.v. Suchtelenweg 4, 6703 Wageningen, The Netherlands c National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba, Ibaraki 305-8506, Japan Received 16 March 2004; received in revised form 26 November 2004; accepted 22 March 2005

Abstract Laboratory scale anaerobic upow lter, sludge blanket and hybrid bed reactors were operated for 860 days in the treatment of high ammonia landll leachate. Organic loading was gradually increased from 1.3 to 23.5 kg COD/m3 day in the start-up period and then uctuated according to the COD concentration of raw leachate. To prevent free ammonia inhibition, inuent pH was reduced to 4.5 after Day 181 and consequently COD removal eciencies above 80% were achieved in all reactors. However, the anaerobic lter and hybrid bed reactor were generally found slightly more ecient and stable than the UASB reactor. In addition to conventional anaerobic reactor control parameters, the complementary techniques of denaturing gradient gel electrophoresis (DGGE), cloning and uorescent in situ hybridization (FISH) were used to identify and compare the microbial proles in the reactors at Day 830. Molecular analyses revealed that acetoclastic Methanosaeta species were prevalent in all reactors and conguration did not have an impact on microbial diversity in the long-term. 2005 Elsevier Ltd. All rights reserved.
Keywords: Ammonia; Anaerobic treatment; DGGE; FISH; Landll leachate; Methanogens

1. Introduction To start-up and operate an anaerobic reactor properly, the microbial diversity and activity of the biomass should be continuously monitored. In this manner, investigations should include the identication and quantication of microorganisms that lead to the formation and stability of biomass in the anaerobic reactor. Moreover, identication of microbial communities in the reactors is necessary for explanation of the observed dierences in the performances of reactors treating similar wastewaters. Although microbial compositions of anaerobic reactors are reported in the literature, there
*

Corresponding author. Tel.: +90 216 347 40 90; fax: +90 216 348 02 E-mail address: bcalli@eng.marmara.edu.tr (B. Calli).

93.

is no study reporting the inuence of reactor conguration on the composition of the microbial community. Insights into the diversity, structure and function of mixed microbial communities in anaerobic reactors are essential for advancing reactor performance and stability against inhibitory compounds. However, knowledge about microbial diversity and activity of the seed biomass is needed for a successful start-up, since, as a general application, seed biomass is taken from another anaerobic reactor which is not adapted to the new wastewater. Without doubt, in the last decade, rRNA/rDNAbased advanced molecular techniques have become the most signicant detection and identication methods in the determination of the diversity of complex microbial populations in anaerobic reactors. Sequence analysis of the rRNA or the rRNA gene (rDNA) has revealed

0960-8524/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2005.03.021

ARTICLE IN PRESS
2 B. Calli et al. / Bioresource Technology xxx (2005) xxxxxx

that 16S and 23S rRNA can be used as evolutionary biomarkers. The 16S and 23S rRNA contain both highly conserved and highly variable regions. At present, a large database of 16S rRNA sequences is available, which is expanding daily, and 16S rRNA analysis has become a signicant tool for the identication of new isolates (Raskin et al., 1994). Several rRNA/rDNAbased methods have been developed to identify and quantify microorganisms in anaerobic reactors (Sekiguchi et al., 1999; McMahon et al., 2001). These methods can even be used without cultivation of the microorganisms (Stams and Oude Elferink, 1997; Oude Elferink et al., 1998). In this study, in addition to conventional reactor control parameters, denaturing gradient gel electrophoresis (DGGE), cloning and uorescent in situ hybridization (FISH) techniques were used to compare the prevalent microbial populations in three dierent anaerobic reactors, of dierent types, treating the same leachate. On Day 830 of operation sludge samples were taken from the bottom sampling ports of each reactor. 16S rDNA isolated and amplied from these samples was analyzed by DGGE for bacterial and archaeal communities. For isolation and identication of dominant archaeal bands on DGGE gel, cloning and sequencing were applied in order. The results of DGGE and cloning were further conrmed with uorescent in situ hybridization (FISH).

2. Methods 2.1. Operation of anaerobic reactors Three dierent laboratory scale anaerobic reactors, namely upow anaerobic lter (AF), sludge blanket (UASB) and hybrid bed (lter over sludge blanket) reactors, were operated for 860 days in the treatment of high ammonia landll leachate. Congurations and dimen-

sions of the reactors are shown in Fig. 1. Details of the experimental set-up and more information about the reactors can be found in another paper (Inanc et al., 2000). Filter media used in the anaerobic upow lter and hybrid bed reactor were plastic helezonic rings. The specic surface area and void ratio of the lter media were 112 m2/m3 and 87%, respectively. A constant temperature room was used to maintain the temperature of the reactors at 35 3 C. The reactors were seeded with sludge taken from an anaerobic treatment plant in the bakers yeast industry. The leachate was brought from IstanbulKomurcuoda Sanitary Landll, once every week. Komurcuoda Landll leachate was characterized by high COD and ammonia concentrations ranging from 12,350 to 47,800 mg/l, and from 1500 to 2680 mg/l, respectively (Inanc et al., 2000). After being characterized, leachate was fed to the reactors without any pH adjustment from Day 1 to Day 181. From Day 181 to Day 860 inuent pH was adjusted to 4.5 to reduce the free ammonia concentration in the reactors. In the start-up and in some periods when the COD concentration in leachate rose suddenly, dechlorinated tap water was used for dilution of the leachate to adjust the organic loading. Phosphoric acid was added to the feed to provide a COD/N/P ratio of 500/7/1 due to phosphorus deciency of the leachate. To monitor the performances of the reactors, inuent and euent pH and alkalinity were measured on a daily basis, while COD and NH3-N were monitored three times a week. TSS and VSS were monitored once a week. All analyses were carried out in accordance with the Standard Methods (APHA, 1998). To compare the nal biomass concentration and distribution in the reactors, sludge samples were taken from all of the sampling ports at Day 830 and SS, VSS, pH and COD were measured. For identication of microbial communities in reactors, biomass samples were taken at Day 830 from the lowest sampling port of each reactor. Although bio-

Fig. 1. Experimental set-up and reactor congurations.

ARTICLE IN PRESS
B. Calli et al. / Bioresource Technology xxx (2005) xxxxxx 3

lm samples could not be taken from the anaerobic lter by using the sampling ports, it was assumed that biolms sloughed o from the lter medium were also present in the sample. Thus, both attached (biolm) and non-attached (oc) microorganisms in the anaerobic lter could be suciently represented by using this sample. 2.2. DNA isolation and PCR amplication The sludge samples were mechanically bead beaten using a vortex mixer modied as a bead beater. After the mechanical disruption, released DNA was isolated and puried with phenol/chloroform/isoamyl alcohol (25:24:1) and chloroform/isoamyl alcohol (24:1) extraction according to the procedure given in Calli et al. (2003). The V6V8 region of the 16S-rDNA genes, from the isolated DNA, was amplied using polymerase chain reaction (PCR) with two sets of PCR primers for DGGE analysis. The primer set specic to the Bacteria domain was GC968 for and UNI1401 rev (Nu bel et al., 1996). The other primer set specic to the Archaea domain was A-109 for (Grosskopf et al., 1998) and GC515 rev (Lane, 1991). For the amplication of the 16SrDNA gene for cloning, A-109 for (Grosskopf et al., 1998) and 1510 rev were used as the primer set specic to the Archaea domain. The PCR programs were performed in a Progene PCR thermocycler (Techne). In the PCR for total archaeal 16S-rDNA and for DGGE the following program was used: pre-denaturation (94 C, 5 min), 34 cycles of annealing (52 C, 40 s), elongation (68 C, 1 min), denaturation (94 C, 30 s), a nal annealing (52 C, 40 s) and post-elongation (68 C, 5 min). In the PCR for bacterial 16S-rDNA the following program was used: pre-denaturation (94 C, 5 min), 30 cycles of denaturation (94 C, 30 s), annealing (56 C, 20 s), elongation (68 C, 40 s) and post-elongation (68 C, 5 min). The reactions were subsequently cooled to 4 C. PCR products were examined on ethidium bromide-stained agarose gels, and were used for DGGE analysis and cloning. 2.3. Denaturing gradient gel electrophoresis (DGGE) and cloning DGGE of the PCR amplied bacterial and archaeal partial 16S rDNA was performed with the BioRad DCode Universal Mutation Detection System according to Nu bel et al. (1996). The polyacrylamide concentrations used in the analysis of DGGE were 8% and the denaturing gradients were 3550%. The denaturing gradient gel was electrophoresed at 85 V at 60 C for 16 h after a voltage of 200 V for 5 min. Silver-staining and development of the gels were performed as described elsewhere (Sanguinetti et al., 1994). The amplied archaeal 16S rDNA products were puried with a QIAquick PCR purication kit (Qiagen)

and cloned in E. coli JM109 by using the pGEM-T Easy vector system (Promega) with ampicillin selection and blue/white screening, according to the manufacturers manual. The inserts were screened by Restriction Fragment Length Polymorphism (RFLP) analysis with the enzyme MspI (fragments were compared by electrophoresis on a 2% (w/v) agarose gel with ethidium bromide staining) and by mobility comparison on DGGE. Plasmids of selected transformants were puried using the QIAprep spin miniprep kit (Qiagen). Sequencing analysis was carried out with the Sequenase sequencing kit (Amersham) using the sequencing primers Sp6, complementary to one adjacent sequence of the pGEM-T cloning site, and T7, complementary to the other adjacent sequence of the pGEM-T cloning site, labelled with IRD8000 (MWG-Biotech). The sequences were analyzed automatically on a LI-COR (Lincoln) DNA sequencer 4000L and corrected manually. A similarity search in the GenBank database, with the derived partial (app. 800 bp) 16S rDNA sequences from the clones, was performed by using the NCBI sequence search service, available on the internet [NCBI]. 2.4. Fluorescent in situ hybridization (FISH) The sludge samples were xed overnight in 4% paraformaldehyde-phosphate-buered saline at 4 C. Fixed cells were spotted on gelatin-coated [0.1% gelatin and 0.01% KCr(SO4)2] multiwell glass slides (10 wells/slide; 10 ll of sample/well) and allowed to dry in a sterile hood (Raskin et al., 1994). Hybridizations were performed at 46 C for 2 h with a hybridization buer (0.9 M NaCl, 20 mM Tris/HCl, pH 8.0, 0.01% SDS) containing each labeled probe (3 ng/ll for Cy3 and 5 ng/ll for FLUOS) (MWG Biotech). Formamide was added to the nal concentrations listed in Table 1 to ensure the optimal hybridization stringency. After hybridization, unbound oligonucleotides were removed by rinsing with washing buer containing the same components as the hybridization buer except for the probes. For detection of all DNA, 4,6-diamidino-2-phenylindole (DAPI) was added to the wash buer (100 ll of 0.1% DAPI). Slides were subsequently incubated at 48 C for 10 min with washing buer, rinsed briey with ddH2O and immediately air-dried. Vectashield (Vector Laboratories) was used to prevent photo bleaching. The slides were examined with an Olympus BX50 microscope.

3. Results and discussion In the start-up period, organic loading rate was gradually increased from 1.3 to 23.5 kg COD/m3 day and it then uctuated between 2.9 and 23.5 kg COD/m3 day according to the COD concentration of raw leachate (Fig. 2). Hydraulic retention time was reduced from

ARTICLE IN PRESS
4 B. Calli et al. / Bioresource Technology xxx (2005) xxxxxx

Table 1 Oligonucleotide probes (Raskin et al., 1994) Probe name EUB 338 ARC 915 MX 825 MS 821 Labeled with FLUOS Cy3 Cy3 FLUOS Formamide (%) 0 0 35 35 Sequence CTGCTGCCTCCCGTAGGAGT GTGCTCCCCCGCCAATTCCT TCGCACCGTGGCCGACACCTAGC CGCCATGCCTGACACCTAGCGAGC Target group Bacteria domain Archaea domain Methanosaeta Methanosarcina

Fig. 2. Organic loading rate and COD removal eciencies.

2.4 to 2.0 days after Day 181 and kept constant until the end of the study. During the start-up period, all reactors showed similar performances with COD removal eciencies between 75% and 95%, until free ammonia reached inhibitory levels (Fig. 2). Due to high inuent total ammonia concentrations of about 2500 mg/l and alkalinity production as a result of anaerobic COD removal, pH was generally above 7.7 in the reactors and exceeded 8.3 when not regulated (Fig. 3a and b). To cope with high pH levels and free ammonia inhibition, pH was adjusted to 4.5 with concentrated HCl in the inuent after Day 181 (Fig. 3b). Thus, pH and free ammonia gradually decreased from 8.3 to 7.5 and from 330 to 30 mg/l, respectively (Fig. 3b and c). Generally, free ammonia inhibition in anaerobic treat-

ment at mesophilic conditions has been reported in the range of 50150 mg/l (Koster, 1986). Therefore, soon after the drop in free ammonia levels, COD removal eciencies returned to 85%. In general, COD removal eciencies were usually above 80% in all of the reactors while the lowest eciencies were always associated with the UASB reactor (Fig. 2b). Likewise, Borzacconi et al. (1999) observed deterioration in the performances of a laboratory scale UASB reactor treating young landll leachate when total ammonia and pH exceeded 1500 mg/l and 8.5, respectively. By adjusting the inlet pH, they reported that ammonia inhibition was reverted and COD removal eciencies above 80% were obtained at 20 kg COD/ m3 day.

ARTICLE IN PRESS
B. Calli et al. / Bioresource Technology xxx (2005) xxxxxx 5

Fig. 3. Total and free ammonia concentrations and pH.

The higher resistance of the anaerobic lter and hybrid bed reactors to elevated free ammonia levels during the rst 180 days might have been due to limited diusion of free ammonia into the depths of biolm. Another reason might be accumulation of more methanogens or methanogens acclimated to elevated free ammonia levels in these biolm supported reactors than in the UASB reactor. Since the sludge samples for microbial identication were taken on Day 830, 650 days after the last signicant free ammonia inhibition (Day 175), it was dicult to comment on the details of the acclimation by using sludge samples not exposed to elevated free ammonia levels for a very long time. In another study, COD removal above 80% at 300 mg/ l free ammonia concentration in an upow anaerobic lter treating sh meal processing wastewater was similarly related to immobilization of biomass, either by attachment on supporting medium or by entrapment (Guerrero et al., 1997). To explore the eects of low inuent pH, which was reduced to prevent free ammonia inhibition, VSS, COD and pH proles were monitored along the height of the reactors at Day 830 (Fig. 4). The results obtained were very interesting, especially for the UASB reactor. pH increased from 4.5 to around 7 within the rst 20 cm between the inlet and the rst sampling port in

the anaerobic lter and hybrid bed reactor, while this was not observed in the UASB reactor until 50 cm (Fig. 4c). This pattern was similar for COD proles. For example, COD decreased from 16,000 mg/l to below 5000 mg/l within the rst 20 cm in the anaerobic lter and hybrid bed reactor, while in the UASB reactor it decreased to 5000 mg/l levels only after 50 cm from the inlet (Fig. 4b). The VSS proles in Fig. 4a showed that the reason for this dissimilarity was the dierence in the amounts of biomass accumulated in the reactors. While active biomass was immobilized or trapped intensively in the bottom sections of the anaerobic lter and hybrid bed reactor, it was more homogeneously distributed in the UASB reactor as suspended ocs. Since the active biomass concentrated in the bottom, almost all biodegradable COD was removed within the rst 20 cm of the anaerobic lter and hybrid bed reactor (Fig. 4b). Consequently, pH immediately rose up to 7.0 in this zone in these reactors (Fig. 4c). On the other hand, COD removal was conducted notably between 40 and 60 cm in the UASB reactor (Fig. 4b). In contrast to the dissimilarity in VSS, COD and pH proles, banding patterns of bacterial and archaeal communities obtained with DGGE analysis were found to be similar in all reactors at Day 830. This indicated that reactor conguration did not have a signicant eect on

ARTICLE IN PRESS
6 B. Calli et al. / Bioresource Technology xxx (2005) xxxxxx

Fig. 4. COD and pH proles along the height of the reactors.

microbial distribution (Fig. 5). For identication of dominant archaeal bands on DGGE gel, cloning and DNA sequencing were applied. Amplicons of the cloned inserts were compared with the DGGE prole of sludge samples by running them together on the same DGGE gel to identify the matching bands. After DNA sequencing of these bands and blast search it was revealed that acetoclastic Methanosaeta species were the prevalent methanogens in the archaeal community. The remaining portion of archaea in the reactors consisted mainly of hydrogenotrophic Methanoculleus-like species (Fig. 5). FISH results with archaea specic ARCH915 and Methanosaeta specic MX825 probes conrmed the dominance of Methanosaeta-related cells (Fig. 6). Since almost no signals were detected with the MS821 probe, it was revealed that Methanosarcina-like methanogens were not frequent in the reactors. Although Methanosaeta is known to be more sensitive to elevated ammonia levels than Methanosarcina (Angenent et al., 2002), suitable conditions for the prevalence of Methanosaetarelated species were provided by keeping the free ammo-

Fig. 6. Photomicrograph of anaerobic lter (AF) sludge hybridized with Cy3-labeled Arch915 probe. Arrows indicate dominant Methanosaeta-related methanogens.

Fig. 5. DGGE patterns for bacteria and archaea.

nia concentration below 30 mg/l in the reactors by lowering the inuent pH. Moreover, low acetate concentrations resulting from high COD removal within the rst 20 cm of the anaerobic lter and hybrid bed reactor promoted the presence of Methanosaeta-related species instead of Methanosarcina as the main acetate utilizing methanogens. In the literature, it is generally accepted that at low acetate concentrations long sheathed rod Methanosaeta species with high surface to volume ratio have a competitive advantage over coccus like Methanosarcina species (Jetten et al., 1992; Raskin et al., 1994). On the other hand, low COD removal was observed near the rst sampling port of the UASB reactor in spite of the presence of Methanosaeta. Probably these Methanosaeta cells were in ocs settled from the active sludge blanket, especially suspended around 40 cm above the inlet, and could not function in the bottom of the reactor because of unfavorable pH and acetic acid values. This also explains

ARTICLE IN PRESS
B. Calli et al. / Bioresource Technology xxx (2005) xxxxxx 7

why there was a retardation in COD removal and consequent pH rise along the UASB reactor (Fig. 4b and c).

References
APHA, AWWA, WEF, 1998. Standard Methods for Water and Wastewater Examination, twentieth edn. American Public Health Association, Washington, DC. Angenent, L.T., Sung, S., Raskin, L., 2002. Methanogenic population dynamics during startup of a full-scale anaerobic sequencing batch reactor treating swine waste. Water Res. 36, 46474654. Borzacconi, L., Lopez, I., Ohanian, M., Vinas, M., 1999. Anaerobic aerobic treatment of municipal solid waste leachate. Environ. Technol. 20, 211217. Calli, B., Mertoglu, B., Tas, N., Inanc, B., Yenigun, O., Ozturk, I., 2003. Investigation of variations in microbial diversity in anaerobic reactors treating landll leachate. Water Sci. Technol. 48 (4), 105 112. Grosskopf, R., Janssen, P.H., Liesack, W., 1998. Diversity and structure of the methanogenic community in anoxic rice paddy soil microcosms as examined by cultivation and direct 16S-rRNA gene sequence retrieval. Appl. Environ. Microbiol. 64 (3), 960969. Guerrero, L., Omil, F., Mendez, R., Lema, J.M., 1997. Treatment of saline wastewaters from sh meal factories in an anaerobic lter under extreme ammonia concentrations. Bioresource Technol. 61, 6978. Inanc, B., Calli, B., Saatci, A., 2000. Characterization and anaerobic treatment of the sanitary landll leachate in Istanbul. Water Sci. Technol. 41 (3), 223230. Koster, I.W., 1986. Characteristics of the pH-inuenced adaptation on methanogenic sludge to ammonia toxicity. J. Chem. Technol. Biotechnol. 36, 445455. Jetten, M.S.M, Stams, A.J.M., Zehnder, A.J.B., 1992. Methanogenesis from acetate: a comparison of the acetate metabolism in Methanothrix soehngenii and Methanosarcina spp. FEMS Microbiol. Rev. 88, 181198. Lane, D.J., 1991. 16S/23S rRNA sequencing. In: Stackebrandt, E., Goodfellow, M. (Eds.), Nucleic Acid Techniques in Bacterial Systematics. J. Wiley & Sons, Chichester, UK, pp. 115147. McMahon, K., Stroot, P.G., Mackie, R.I, Raskin, L., 2001. Anaerobic codigestion of municipal solid waste and biosolids under various mixing conditions-II: microbial population dynamics. Water Res. 35 (7), 18171827. Nu bel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A., Amann, R.I., Ludwig, W., Backhaus, H., 1996. Sequence heterogeneities of genes encoding 16S rRNAs in Paenibacillus polymyxa detected by temperature gradient gel electrophoresis. J. Bacteriol. 178 (19), 56365643. Oude Elferink, S.J.W.H., van Lis, R., Heilig, H.G.H.J., Akkermans, A.D.L., Stams, A.J.M., 1998. Detection and quantication of microorganisms in anaerobic bioreactors. Biodegradation 9 (3/4), 169177. Raskin, L., Poulsen, L.R., Noguera, D.R., Rittman, B.E., Stahl, D.A., 1994. Quantication of methanogenic groups in anaerobic biological reactors by oligonucleotide probe hybridization. Appl. Environ. Microbiol. 60 (4), 12411248. Sanguinetti, C.J., Dias Neto, E., Simpson, A.J., 1994. Rapid silver staining and recovery of PCR products separated on polyacrylamide gels. Biotechniques 17, 914921. Sekiguchi, Y., Kamagata, Y., Nakamura, K., Ohashi, A., Harada, H., 1999. Fluorescence in situ using 16S rRNA-targetted oligonucleotides reveals localization of methanogens and selected uncultured bacteria in mesophilic and thermophilic sludge granules. Appl. Environ. Microbiol. 65 (3), 12801288. Stams, J.M.A., Oude Elferink, S.J.H.W., 1997. Understanding and advancing wastewater treatment. Curr. Opin. Biotechnol. 8, 328 334.

4. Conclusions Results indicated that landll leachate containing total ammonia concentrations above 2000 mg/l can be treated successfully by using anaerobic reactors congured as upow lter, sludge blanket or hybrid bed reactors. By reducing the inuent pH to 4.5 to control the free ammonia levels in the reactors, COD removal eciencies above 80% were achieved without any significant loss in methanogenic activity. However, the lowest eciencies were always associated with the UASB reactor, particularly when free ammonia rose up to inhibitory levels. Dierent responses of the anaerobic lter and hybrid bed reactor to ammonia inhibition were related to accumulation of more biomass, acclimated to elevated free ammonia levels, either by attachment on supporting medium or by being entrapped. The dierences in VSS, COD and pH proles were consistent with the dissimilarities in biomass concentration and distribution in the reactors. However, as no sludge samples were taken during any period of ammonia inhibition, comment on the acclimation process is not possible. While biomass was trapped intensively in the bottom sections of the anaerobic lter and hybrid bed reactor, it was more homogeneously distributed as suspended ocs in the UASB reactor. On the other hand, the results of molecular analysis carried out to identify and compare the prevalent microorganisms in the reactors revealed that the reactor conguration did not have a signicant inuence on the diversity of microbial communities in the long term. After sequencing of dominant DGGE bands and visualizing the hybridization signals, acetoclastic Methanosaeta species known to be sensitive to ammonia inhibition were found to be prevalent in all reactors. Keeping the free ammonia concentration in the reactors below 30 mg/l by lowering the inuent pH provided suitable conditions for prevalence of Methanosaeta species. Besides, the retarded COD removal along the UASB reactor in spite of an intense Methanosaeta presence at the bottom of the reactor was related to inhibition of these methanogens because of low inuent pH and associated high unionized acetic acid concentrations. Acknowledgements The authors would like to acknowledge the nancial support of TUBITAK (Project No.: ICTAG A034, A036 and A051).