Microbial Community Structure and Activity in a Compartmentalized, Anaerobic Bioreactor Largus T. Angenent, Dandan Zheng, Shihwu Sung, Lutgarde Raskin
ABSTRACT: The objective of this study was to evaluate staging and its effects on reactor performance in a compartmentalized bioreactor, designated the anaerobic migrating blanket reactor (AMBR). The AMBR was operated by reversing the flow several times per day, which allowed for substantial biomass migration without biomass accumulation in the final compartment. During reactor startup, the structures of the microbial communities in the five compartments were quite similar despite substantial differences in substrate types and concentrations in the different compartments. During the rest of the operational period, biomass migration was reduced by changing operating conditions and, as a result, a larger difference in the structures of the microbial communities developed for the different compartments (biomass staging). For example, after changing operating conditions, rRNA levels for the acetate-utilizing methanogen Methanosaeta concilii were approximately 35 and 10% of the total rRNA in the middle and outside compartments, respectively (before changing operating conditions these levels were approximately 20 and 12% of the total rRNA, respectively). Promoting larger differences in the structures of the microbial communities for the different compartments did not improve reactor performance as lower levels of M. concilii in the outside compartments hindered acetate removal and compromised effluent quality. Water Environ. Res., 74, 450 (2002). KEYWORDS: anaerobic wastewater treatment, anaerobic migrating blanket reactor, staging, methanogens, oligonucleotide probes, 16S rRNA.
Introduction During the past three decades, numerous full-scale anaerobic systems have been successfully used for the treatment of industrial and domestic wastewater. The upflow anaerobic sludge blanket (UASB) reactor, a single-vessel treatment system, has been the most frequently used high-rate anaerobic reactor worldwide. Despite the popularity of the UASB reactor, the development of simple and affordable alternative wastewater treatment technologies is still desirable for small and medium industries (Hulshoff Pol et al., 1997). Such alternatives include staged anaerobic treatment systems. In staged configurations, which consist of two or more reactors or compartments in a series, hydrolysis and fermentation take place mostly in the first stage and acetogenesis and methanogenesis occur primarily in the final stage(s). Hence, the different reactors or compartments receive different substrate types and concentrations, which results in the development of different microbial communities because substrate type and concentration ultimately determine the biomass composition (Fox and Pohland, 1994; van Lier et al., 1996; van Lier et al., 1997). In this manuscript, the development of different microbial community structures in the different compartments of staged systems is referred to 450
as biomass staging. Biomass staging can be prevented by continuous biomass recycling (Duran and Speece, 1998) or by combining the biomass of the different compartments on a regular basis. As a result, biomass staging will not develop, but substrate staging will still be observed. Substrate staging can also be accomplished in compartmentalized reactor systems through internal recycling of biomass (Angenent and Sung, 2001) and in one-reactor batch-fed systems, which exhibit substrate staging over time rather than over space (Sung and Dague, 1995). Several studies have found staged reactors to be very stable (Angenent et al., in press; Bachman et al., 1985; Duran and Speece, 1998; Guiot et al., 1995; van Lier, 1995). One reason for the improved stability of staged versus one-reactor systems is the maintenance of low hydrogen levels in the final stages, which promotes propionate oxidation during shock-load conditions (Harper and Pohland, 1986; Stams, 1994; Thauer et al., 1977). In addition, a shift toward the production of more desirable intermediates (butyrate rather than propionate) in the initial stage because of lower pH levels and higher substrate concentrations during shock-load conditions further explains the improved stability observed in staged systems (McCarty and Mosey, 1991). The objective of the research presented herein was to study staging in a five-compartment anaerobic migrating blanket reactor (AMBR). To study the effects of different modes of staging on AMBR performance and stability, the substrate staging was promoted and biomass staging was prevented during the initial period of operation, and subsequently the development of biomass staging was allowed. Oligonucleotide probe hybridizations were used to study the development of biomass staging and to visualize the spatial arrangement of methanogens in granules. Materials and Methods Configuration and Operating Conditions of the Anaerobic Migrating Blanket Reactor. The AMBR consisted of a rectangular, Plexiglas reactor (inside dimensions: length ⫽ 65 cm, height ⫽ 33 cm, width ⫽ 13 cm) with an active volume of 20 L, which was divided into five compartments of 4 L each (Figure 1). Holes (2.54 cm in diameter) were placed in the walls separating the compartments at a height of 17.5 cm from the bottom, while the liquid level was at a height of 23.6 cm. Sampling ports were placed 3 cm from the bottom of each compartment. The five compartments were mixed gently with mixers (model 5vb, EMI, Inc., Clinton, Connecticut), which were operated simultaneously for 10 seconds every 10 minutes at 100 rpm. The use of paddles (four vertical bars, width ⫽ 1.25 cm and length ⫽ 10 cm, were mounted on two horizontal bars, width ⫽ 1.25 cm and length ⫽ 9 cm; Water Environment Research, Volume 74, Number 5
Angenent et al.
Figure 1—Schematic diagram of AMBR. The top and bottom horizontal lines in each compartment indicate the height of the liquid and the height of the biomass, respectively.
Figure 1) further ensured gentle, but thorough mixing. At a rotational speed of 100 rpm, these paddles produced a root-meansquare-velocity gradient, G, of 232/s in a 4-L compartment as determined by a rotating torque meter (Bex-O-Meter, model 38, The Bex Company, San Fransisco) described by Sajjad and Cleasby (1995). The headspaces of the compartments were separated and biogas was collected in separate systems. A biogas collection system consisted of an observation bottle, a gas sampling port, and a wet-test gas meter (type 1 L, Schlumberger Industries, Dordrecht, The Netherlands). Baffles were placed in front of the effluent ports to prevent floating granules from washing out with the effluent. The baffles were closed at the top to separate the headspaces of the outside compartments from the air (Figure 1). This configuration allowed the effluent to flow out of the reactor by gravity without losing biogas from the headspaces. Two automatic ball valves with an internal diameter of 2.54 cm were used to open and close effluent ports (True Blue Electric Actuator, model EBV-6, Plasto-matic Valves, Inc., Cedar Grove, New Jersey). Programmable timers (ChronTrol Corporation, San Diego) were used to control reactor operation. The flow over the horizontal plane of the reactor was reversed four times a day. After feeding the initial compartment for 4 hours, the second, adjacent compartment was fed for 2 hours before the flow was reversed. The temperature of the AMBR was maintained at 35 ⫾ 1 °C by circulating warm water through a reactor jacket (heating recirculator, model 210, Polyscience, Niles, Illinois). A concentrated stock solution consisted of sucrose, nutrients, bicarbonate, yeast extract, and trace elements (the trace element solution was modified from Zehnder et al. [1980] and van Lier [1995]). The concentrated stock solution consisted of 96 g/L sucrose, 55 g/L sodium bicarbonate (NaHCO3), 0.3 g/L yeast extract, 10 g/L ammonium chloride (NH4Cl), 2.0 g/L dibasic potassium phosphate (K2HPO4), 1.7 g/L monobasic sodium phosphate (NaH2PO4•H2O), 1.0 g/L ferrous chloride (FeCl2•4H2O), 0.2 g/L cobaltous chloride (CoCl2•6H2O), 0.1 g/L ethylenediaminetetraacetic acid (EDTA), 0.05 g/L manganese chloride (MnCl2•4H2O), 0.02 g/L resazurin (C12H7O4N), 0.0142 g/L nickel chloride (NiCl2•6H2O), 0.0123 g/L sodium selenite (Na2SeO3), 0.009 g/L aluminum chloride (AlCl3•6H2O), 0.005 g/L boric acid (H3BO3), 0.005 g/L zinc chloride (ZnCl2), 0.005 g/L heptaammonium molybdate ((NH4)6Mo7O24•4H2O), 0.0038 g/L cupric chloride (CuCl2•2H2O), and 0.1 mL/L hydrochloric acid (HCl) (37.7% solution). The stock solution was kept at 4 °C and was mixed to September/October 2002
keep all components in solution. The low temperature, combined with the high sugar and salt concentrations (the water activity was estimated to be approximately 0.97 [Scott, 1957]), prevented microbial activity in the stock solution. Makeup water (tap water from Ames, Iowa), preheated to a temperature of 35 ⫾ 1 °C, was added to the stock solution before the solution was fed to the reactor. This water contributed additional nutrients such as calcium, magnesium, and sulfate. The AMBR was inoculated with granules from a UASB reactor operated by the Heileman brewery in La Crosse, Wisconsin. The granules were added to obtain a volatile suspended solids (VSS) concentration of 30 g/L in the reactor and were crushed by mixing the reactor contents at a high rotational speed (⬎500 rpm; G ⬎ 2500/s) for 24 hours. After addition of the inoculum, a 24-hour rest period was implemented to eliminate oxygen. Chemical Analyses. Effluent samples, samples from individual compartments, and biogas samples were taken at the midpoint of the time interval between two reversals of flow (or at a specified time during the flow cycle). Liquid samples were filtered through 0.45-m pore-size filters and stored at 4 °C until analysis. For volatile fatty acid (VFA) analyses, samples were acidified with hydrochloric acid. Biogas production was monitored with five gas meters. Concentrations of nitrogen, methane, and carbon dioxide in biogas samples and in biogas collected during activity tests were measured using gas chromatography (model 350, Gow-Mac Instruments, Co., Bridgewater, New Jersey) with a thermal conductivity detector (column: 1.7 m ⫻ 3 mm stainless steel Poropack Q [Gow-Mac] 149/177 m diameter openings [80/100 mesh]; carrier gas: helium). The hydrogen concentration in the biogas was measured with a reduction gas analyzer (model RGA3, Trace Analytical, Menlo Park, California) with a mercuric oxide assembly bed and mercury detector and with nitrogen as the carrier gas. Individual VFA levels were measured by ion chromatography (model DX-500, Dionex, Sunnyvale, California) with a CD 20 conductivity detector and anion micromembrane suppresser (column: Ion Pac ICE-As1; eluant: 0.8 to 1.0 mM heptafluorobutyric acid). Total alkalinity, total chemical oxygen demand (TCOD), soluble chemical oxygen demand (SCOD), and concentrations of total VFAs, total suspended solids, and VSS were performed according to Standard Methods (APHA, 1995). Standard methane production rate (SMPR), methane-based chemical oxygen demand (MCOD) removal efficiency, TCOD removal efficiency, SCOD removal efficiencies of the system and individual compartments, and the free energy changes of propionate conversion (⌬Gpropionate) were calculated according to Angenent et al. (in press). Biomass Characteristics. Biomass samples were taken from separate compartments of the AMBR and were used immediately for VSS measurement and activity tests. The acetate methanogenic activity (AMA) and the propionate acetogenic activity (PAA) were assessed using the “headspace method” according to specific methanogenic activity tests described by Rinzema et al. (1988) in 250-mL serum bottles at 35 °C. The PAA was estimated by measuring the methane production rate after adding propionate as the only substrate. The sucrose acidogenic activity (SAA) test was based on the specific acidogenic test described for glucose by Vanderhaegen et al. (1992), in which the production of acid was determined by measuring the equivalents of base added with a syringe pump (kdScientific, New Hope, Pennsylvania) to maintain a pH of 5.8. Instead of glucose, sucrose was used as the substrate at an initial concentration of 2 g/L. A 150-mL respirometer batch 451
Angenent et al.
Table 1—Oligonucleotide probes used, including their target groups, characteristics, and references. Probea S-*-Univ-1392-a-A-18 S-D-Bact-0338-a-A-18 S-D-Arch-0915-a-A-20 M-O-Mmic-1200-a-A-21
Most organisms Most Bacteria Most Archaea Methanomicrobiales
S-F-Mbac-0310-a-A-22
Methanobacteriaceae
S-F-Mcoc-1109-a-A-20
Methanococcaceae
S-F-Synm-0700-a-A-23
Syntrophomonadaceae
S-G-Dsbb-0660-a-A-20 S-*-Synb-0838-a-A-21
Desulfobulbus spp. Syntrophobacter spp.
S-G-Msar-0821-a-A-24
Methanosarcina spp.
S-S-M.con-0381-a-A-22
Methanosaeta concilii
a b
Relevant characteristicsb
Target group
Most methanogens in this order use H2–CO2 and formate Most methanogens in this family use H2–CO2, some use H2–CO2 and formate Most methanogens in this family use H2–CO2 and formate Members of this family are saturated fatty acid-oxidizing bacteria, responsible for butyrate degradation in the AMBR Propionate-degrading sulfate reducing bacteria Syntrophic propionate oxidizing bacteria Methanogens in this genus use acetate and other substrates (H2–CO2, methanol, and methylamines), and exhibit high half-saturation constants and maximum specific growth rates for acetate This methanogen can only use acetate, and exhibits low half-saturation constants and maximum specific growth rates for acetate
Zheng et al. (1996) Amann et al. (1990) Stahl et al. (1991) Raskin and co-workers (1994b) Raskin and co-workers (1994b) Raskin and co-workers (1994b) Hansen et al. (1999)
Devereux et al. (1992) Zheng and Raskin (unpublished data) Raskin and co-workers (1994b)
Zheng et al. (2000)
Probe nomenclature according to Alm et al. (1996). According to Raskin and co-workers (1994a).
vessel (Ellis et al., 1996) was modified to create an anaerobic headspace flushed with nitrogen gas at a low flow and the oxygen probe was replaced with a pH probe. The temperature was maintained at 35 ⫾ 1 °C by circulating warm water through the jacket of the batch vessel. The contents were stirred vigorously to break up granules and prevent diffusion limitations. Samples from different compartments were combined for granularsize measurements (on one occasion, the individual samples were analyzed separately). To study the change in granular size over time, n the arithmetic mean diameter (兺i⫽1 diameteri/n) and area-weighted n n mean diameter ([(兺i⫽1 (diameteri)3])/([兺i⫽1 (diameteri)2]) were calculated with automated image analysis. Biomass samples were mixed and diluted to obtain a mixture of clearly visible and nonoverlapping biomass particles. Next, 1.75 mL of this diluted sample was added to a counting chamber, which was prepared by cementing two, 3-mm thick glass slides together with a 2.5-cm diameter hole in the top slide. The counting chamber was closed with a cover slip to eliminate air bubbles. The automated image analysis setup consisted of a blackand-white video camera (Dage-MTI series 68, Michigan City, Indiana), a microscope (Olympus SZH, Melville, New York), and a personal computer with Quartz PCI Imaging software (Quartz Imaging Corporation, Vancouver, British Columbia, Canada). Some manual editing of the image was necessary to separate adjacent granules. Particles smaller than 0.1 mm were not included in calculations of size distribution (Grotenhuis et al., 1991). rRNA-Targeted Hybridizations. Biomass samples were divided into two 2.2-mL screw-cap microcentrifuge tubes, the tubes were centrifuged at 350 ⫻ g at 4 °C for 2 minutes, and the supernatant was removed from both tubes. The biomass in one of the two tubes was crushed with an RNase-free glass rod, frozen immediately in an ethanol dry ice bath, and stored at – 80 °C until RNA isolation using a low-pH hot-phenol extraction method (Raskin et al., 1995; Stahl et al., 1988). The biomass in the second 452
Reference
tube was prepared for fluorescence in situ hybridization (FISH) by overnight fixation with 4% paraformaldehyde as previously described (de los Reyes et al., 1997). The granules were then embedded in paraplast medium (Sigma, St. Louis) (Harmsen et al., 1996) and sectioned (thickness of 5 m) at the Histological Laboratory of Diagnostic Medicine (College of Veterinary Medicine, University of Illinois at Urbana-Champaign, Urbana). The ribbons were stretched in 50 °C water and transferred to microscope slides (positively charged slides with saline; Cell-Line Associates, Inc., Newfield, New Jersey). The slides were dried at 42 °C overnight and deparaffinated as described by Harmsen et al. (1996). Membrane hybridizations with 32P-labeled oligonucleotide probes (Table 1) were conducted as previously described (Raskin and co-workers, 1994a). The hybridization signals were quantified using an Instant Imager (Packard Instruments, Meriden, Connecticut) and the abundance of each target group was expressed as a percentage of the total 16S rRNA determined using a universal probe, S-*-Univ-1390-a-A-18 (Zheng et al., 1996). Fluorescence in situ hybridization with a tetramethylrhodamine isothiocyanate (TRITC)-labeled Archaea probe and a fluorescein isothiocyanate (FITC)-labeled Bacteria probe were conducted at 46 °C in a moisture chamber (Amann et al., 1990); washes were performed at 48 °C (Zheng and Raskin, 2000). The hybridization and wash buffers contained no formamide. Micrographs were observed using an epifluorescence microscope (Axioskop; Carl Zeiss, Jena, Germany) equipped with filter sets 41001 (for FITC) and 41002 (for TRITC) (Chroma Technology Corporation, Brattleboro, Vermont). A liquid-cooled charge-coupled device camera (Photometrics Ltd., Tucson, Arizona) and IPLab Spectrum software (Signal Analytics, Vienna, Virginia) were used to capture the image, which was then exported to Adobe Photoshop 3.0 (Adobe, Seattle). Statistical Analysis. Standard deviations for membrane hybridization data were determined by error propagation from triplicate Water Environment Research, Volume 74, Number 5
Angenent et al.
applications of the same rRNA sample to a membrane, and hence do not represent the variability associated with analysis of biomass samples obtained independently from the reactor. A cluster analysis using the unweighted arithmetic average was performed with the AMA and Methanosaeta concilii data to detect discontinuities in staging patterns in the reactor over time. The 2-method (weighted expression D15 in Legendre and Legendre [1998]) was used to calculate the distances of the patterns without considering the distance of averages. The distances were converted to similarities by dividing the distance by the maximum distance and equating the similarity to one minus the normalized distance. All calculations were performed with “The R Package v. 4.0”, which was downloaded from http://www.fas.umontreal.ca/BIOL/legendre/ on November 1, 2000. In addition, a Mantel test was performed between AMA and M. concilii staging patterns over time to test for the association of these patterns (Legendre and Legendre, 1998). The multinormality of the data was verified to test for the legitimacy of using the Mantel test (Dutilleul et al., 2000). Results and Discussion Experimental Conditions. The objectives of this study were to determine the effect of biomass staging on reactor performance and stability and to evaluate potential advantages and disadvantages of substrate staging compared with biomass staging. The AMBR has been shown to be able to treat a nonacidified carbohydrate-rich wastewater at a high loading rate while maintaining stable performance (Angenent and Sung, 2001; Angenent et al., in press). The AMBR in the current study was operated at a high volumetric loading rate (VLR) of up to 42 g of chemical oxygen demand (COD)/L䡠d using a sucrose-based synthetic wastewater (with a mean concentration of 6.8 g COD/L [standard error ⫽ 0.9; n ⫽ 74]), while limiting the addition of alkalinity (0.55 g NaHCO3/g COD). With this substrate choice, low pH values were anticipated in the initial compartments of the AMBR because of acidification. Given the high VLR and low alkalinity addition, it was necessary to reverse the flow four times a day to prevent the pH in the initial compartments from dropping lower than 6 for extended periods of time. Four cycles per day resulted in a cycle period of 6 hours, during which time one of the outside compartments was fed first for 4 hours. Then, the second, adjacent compartment was fed for 2 hours to prevent a breakthrough of substrate after the flow was reversed. Subsequently, the same process was repeated, but the flow was reversed. Because the reactor consisted of five compartments, the middle (third) compartment was never fed. To facilitate startup, granules from a UASB reactor used to treat preacidified brewery wastewater (consisting mostly of VFAs) were added to the AMBR and were crushed to prevent flotation of granules when contacted with the new substrate (sucrose) at the high food/microorganism ratio (F/M) in the initial compartment. The granules exhibited a relatively high AMA of 1.25 g COD/g VSS䡠d (standard deviation ⫽ 0.03; n ⫽ 3). At the start of operation, the hydraulic retention time (HRT) was set at 13 hours with a VLR of 7 g COD/L䡠d. Subsequently, the HRT was decreased gradually, while the substrate concentration was kept constant, resulting in a gradual increase of the VLR. The operation of the AMBR was divided into five periods. During the initial period, the VLR was increased gradually without biomass wasting (promotion of substrate staging) (period 1, days 0 to 81). During the second period, biomass was wasted to promote biomass staging, which led to unstable conditions and made it necessary to September/October 2002
lower the VLR (period 2, days 82 to 93). For the next 20 days, feeding was terminated (period 3, days 94 to 113). Feeding was resumed on day 114 at a lower VLR (period 4, days 114 to 130). Finally, the reactor was deliberately overloaded to promote biomass staging until the end of the operating period, while biomass was wasted to maintain biomass levels at 35 g VSS/L (period 5, days 131 to 154). Reactor Performance. Period 1: Increasing Volumetric Loading Rate (Days 0 to 81). During the initial period of operation, the VLR was increased from 7 to 42 g COD/L䡠d (Figure 2a). Despite this considerable increase in loading rate, total VFA concentrations in the effluent (samples were taken at the midpoint between flow reversals) remained lower than 250 mg/L as acetate (Figure 2d). This resulted in high SCOD removal efficiencies (average SCOD removal was 98.5% for five data points between a 12- and 8-hour HRT [standard error ⫽ 0.9]) (Figure 2d). At a VLR of 42 g COD/L䡠d (HRT ⫽ 4 hours), the biogas production was greater than 500 L/d and the SMPR was approximately 11.7 L/L䡠d (Figure 2e). The AMBR was able to accumulate biomass up to a concentration of 50 g VSS/L (Figure 2b). Retention of biomass at these conditions was possible because of the formation of a granular blanket, which was formed within 2 weeks after startup (the granular size increased substantially during period 1 [Figure 2f]). The pH and individual VFA concentrations of the reactor contents and methane levels in the biogas for the five compartments were measured several times during the operating period to evaluate the performance of the AMBR in detail (Table 2). On days 15 and 25, when the VLRs were 10 and 14 g COD/L䡠d, respectively, acetate and propionate were present at substantial levels in the first two compartments, but were low (or even undetectable) in the last three compartments (data not shown). This resulted in a low total VFA concentration in the effluent (i.e., last compartment) of 69 and 206 mg/L on days 15 and 25, respectively (Figure 2d). On day 81, when the VLR had been increased to approximately 42 g COD/L䡠d, VFA levels had increased somewhat in all compartments resulting in a total VFA concentration in the effluent of 236 mg/L (Table 2 and Figure 2d). To evaluate biomass characteristics during period 1, activity tests were performed with biomass samples taken from each compartment. On day 34, the average AMA for the five compartments was 2.01 g COD/g VSS䡠d (standard error ⫽ 0.08; n ⫽ 9) and the AMA values were similar for all five compartments (Figure 3). On day 66, the average AMA was lower compared with the average AMA on day 34, but differences in AMA values among the compartments were still small, suggesting that the biomass community had similar characteristics throughout the AMBR (Figure 3). Because of the high biomass level in the reactor (VSS ⫽ 48 g/L), the top of the granular blanket reached the holes connecting the compartments, which resulted in significant biomass migration through the reactor. In addition, flow reversal contributed to the fairly uniform biomass characteristics throughout the reactor. To promote biomass staging, biomass migration was limited by reducing the height of the granular blanket through biomass wasting initiated on day 82. Period 2: Initial Biomass Wasting (Days 82 to 93). Biomass wasting resulted in a decrease in the VSS concentration (Figure 2b) and an increase in the F/M ratios of the initial compartment and the overall reactor (Figure 2c). As a result, reactor performance deteriorated as evidenced by a decrease in COD removal efficiencies (Figure 2d) and an increase in VFA levels throughout the reactor and in the effluent (Table 2 and Figure 2d). Notably, the 453
Angenent et al.
Figure 2—Operational conditions and reactor performance: (a) volumetric loading rate (\ \) and HRT (E); (b) VSS level in reactor (MLVSS) (■) and VSS in effluent (}); (c) F/M ratio of the initial compartment (■) and of the overall reactor volume (}); (d) TCOD (}), SCOD (■), MCOD () removal efficiencies, and VFA concentrations in the effluent (䊐); (e) biogas production (\ \) and SMPR (); and (f) area-weighted mean diameter of granules of overall reactor volume (■), arithmetic mean diameter of granules of overall reactor volume (}), for day 147, the area-weighted mean diameter of granules in the middle (top) and one of the outside (bottom) compartments (䊐), and arithmetic mean diameter of granules in the middle (top) and one of the outside (bottom) compartments ({). propionate and butyrate concentrations increased substantially (Table 2), which indicated that unstable conditions were imminent. In addition, the high acetate concentrations in the final compartment (Table 2) resulted in more biogas production, which increased the turbulence in this compartment and caused more biomass loss through the effluent (Figure 2b). Although conditions corresponding to this slight overload apparently induced biomass staging (as suggested by the higher AMA in the middle compartment compared with the outside compartments on day 90; Figure 3), the chain of events triggered by biomass wasting on day 82 led to unstable operating conditions. To prevent the AMBR from becoming unstable, the VLR was decreased to 36 g COD/L䡠d on day 91. As a result, the total VFA concentration in the effluent decreased to 400 mg/L as acetate on day 93 (Figure 2d). Period 3: Starvation (Days 94 to 113). From day 94 to day 113, feeding and mixing were stopped and the biomass was kept at 25 °C in the reactor. Performance parameters were not measured during this starvation period. Starvation deteriorated the granular structure, as shown by a decrease in the mean diameters of the granules (Figure 2f). It is likely that the granular structure in the AMBR deteriorated because of digestion of fermenters, which presumably caused the decrease in the VSS levels (Figure 2b). 454
Indeed, fermenters had reached high levels in the biomass as suggested by an SAA between 2 and 3 meq. H⫹/g VSS䡠d on day 66, which is close to the SAA of 3.6 meq. H⫹/g VSS䡠d reported for granular biomass in UASB reactors fed fresh vinasse (Vanderhaegen et al., 1992). Period 4: Resume Feeding (Days 114 to 130). When feeding was resumed on day 114 at a VLR of 31 g COD/L䡠d, a sharp increase in biogas production was observed (Figure 2e), the COD removal efficiencies were higher than before the starvation period (Figure 2d), and the VFA concentrations in the individual compartments were lower (Table 2). To promote biomass staging and to monitor reactor performance during conditions of stress, the VLR was increased slightly on day 131. Period 5: Overload Conditions (Days 131 to 154). By increasing the VLR to 35 g COD/L䡠d while wasting biomass from all compartments to keep the VSS concentration at approximately 35 g/L, the F/M ratio in the initial compartment was maintained at approximately 4.5 g COD/g VSS䡠d (Figure 2c). The F/M ratio for the initial compartment was higher than 4.0 g COD/g VSS䡠d because this value corresponded to the upper limit for optimal performance before day 82 (Figures 2c and 2d). After the increase in VLR, biogas production first increased and then decreased Water Environment Research, Volume 74, Number 5
Angenent et al.
Table 2—Characteristics of the liquid and gas phases of compartments of the AMBR over the operational time period. Samples were taken at the midpoint between reversals of flow (i.e., at the time the initial compartment had been fed for 3 hours). Compartment Operational time (d) 81, Period 1
91, Period 2
123, Period 4
142, Period 5
147, Period 5
Parametera
Initial
Second
Middle
Fourth
Final
Formate Acetate Propionate Butyrate pH Methane Formate Acetate Propionate T-Butyrate pH Methane
0 435 696 88 6.0 35 0 689 1709 205 5.9 25
0 231 362 0 6.6 61 0 384 1076 150 6.4 51
0 99 88 13 6.8 65 0 225 485 444 6.7 62
0 94 108 0 6.8 61 0 143 276 382 6.7 59
0 81 76 0 6.7 60 0 96 193 319 6.7 58
Formate Acetate Propionate Butyrate pH Methane
18 491 449 351 6.0 32
0 321 298 128 6.6 55
0 197 161 34 6.7 61
0 120 107 0 6.7 57
0 89 100 0 6.7 55
Formate Acetate Propionate Butyrate pH Hydrogen ⌬Gpropionate Methane
27.0 644 520 290 6.0 5.5 ⫻ 10⫺2 16 22
0 676 558 217 6.2 1.4 ⫻ 10⫺4 ⫺30 39
0 439 332 84 6.6 8.0 ⫻ 10⫺5 ⫺35 54
0 274 280 26 6.7 3.0 ⫻ 10⫺5 ⫺43 56
0 171 178 0 6.6 5.0 ⫻ 10⫺5 ⫺39 53
Formate Acetate Propionate Butyrate pH SCOD level SCOD removed SCOD efficiency Hydrogen ⌬Gpropionate Methane
12.6 725 466 280 5.8 3075 17.0 58.7 5.0 ⫻ 10⫺2 16 26
10.3 455 305 183 6.3 1650 5.6 19.2 3.0 ⫻ 10⫺4 ⫺24 54
0 328 270 0 6.5 975 2.6 9.1 8.0 ⫻ 10⫺5 ⫺35 60
0 364 271 0 6.5 960 0.1 0.2 6.0 ⫻ 10⫺5 ⫺37 53
0 394 307 0 6.4 1395 0 0 3.0 ⫻ 10⫺5 ⫺43 50
NOTE—atm ⫻ 101.3 ⫽ kPa. a Hydrogen (partial pressure) and methane were measured in separate samples and are expressed as atm and %, respectively; VFA and SCOD levels in the liquid phase are given in mg/L; SCOD removed is expressed as g/h; SCOD efficiency is the SCOD removal efficiency for the individual compartments expressed as a percentage of the total influent COD, hence the summation of the individual SCOD efficiencies gives the total SCOD removal efficiency of the system. The free-energy change for propionate conversion (⌬Gpropionate) is given in kJ/mol.
slowly for the rest of the operational period. Moreover, removal efficiencies decreased and the VFA levels in the effluent increased (Figure 2d). The AMBR was overloaded as anticipated and its performance deteriorated. Nevertheless, most of the influent COD (58.7%) was removed in the initial compartment (Table 2). Using information from Table 2, it was estimated that the free energy change for propionate conversion was ⫹16.3 and ⫹16.1 kJ/mol in the initial compartment for days 142 and 147, respectively. Hence, September/October 2002
it is unlikely that propionate was degraded in the initial compartment. To evaluate the performance during this period in more detail, individual VFA and SCOD concentrations, pH of the reactor contents, and hydrogen and methane levels in the biogas of the five compartments were measured every hour for six hours on day 147 (Table 3). Despite overload conditions, these operating parameters changed rapidly over the horizontal plane of the AMBR. The 455
Angenent et al.
Figure 3—Acetate methanogenic activity (AMA) (■), propionate methanogenic activity (PMA) (}), and sucrose acidogenic activity (SAA) (Œ) of the biomass for individual compartments for the operational time period. acetate concentration, for example, started at a minimum of 289.9 mg/L in the initial compartment (t ⫽ 0 h) and reached 797.6 mg/L after feeding for 4 hours. Simultaneously, the effluent quality 456
improved (the SCOD in the final compartment decreased from 3375 mg/L to 1320 mg/L). The pH gradients also reversed over the horizontal plane of the reactor. A minimum pH of 5.7 in the initial compartment at t ⫽ 4 h and a maximum pH of 6.5 in the middle compartment at t ⫽ 1, 2, and 3 h and in compartment 4 at t ⫽ 3 h were measured. These low pH levels indicated conditions of poor performance. Nevertheless, hydrogen was depleted quickly in the initial compartment, as the hydrogen content decreased from 0.05 atm to 1.1 ⫻ 10⫺4 atm (5 kPa to 0.011 kPa) within 2 hours after feeding to this compartment was terminated (Table 3). Some depletion of hydrogen can be attributed to venting hydrogen with biogas. However, microbial hydrogen consumption likely was the primary reason for depletion because the initial compartments were still producing large amounts of methane. Hourly methane production fluctuated in the outside compartments depending on feeding conditions (on day 147 the methane production at t ⫽ 4 h in the initial compartment was 3 times higher than in the final compartment and the biogas production was 6 times higher). Biomass Staging. Granular Structure. Granules were mostly dark gray in color and increased in diameter during the initial 65 days of operation (Figure 2f). At approximately day 66, small, light-colored (white-yellow) granules appeared in the reactor. These granules were first observed in the outside compartments, but were later distributed throughout the reactor because of biomass migration. Their appearance resulted in a decrease in the average arithmetic mean diameter of the granules (Figure 2f). Meanwhile, large, dark-colored (grayish-brown) granules increased in size and were located mostly in the middle compartment. Hence, the average area-weighted diameter increased to 2.1 mm because large granules have a greater effect on the areaweighted mean diameter compared with the arithmetic mean diameter. Although different types of granules were observed in all compartments, some degree of stratification of different granular morphologies and sizes remained visible, which suggests the initiation of biomass staging during the final part of period 1 (days 66 to 81). Figure 4a (inside back cover) shows an epifluorescence image of a large, dark-colored granule sampled on day 66. The granule consisted of a layered structure of Bacteria and Archaea (the Archaea were located mainly in the core). These layered granules also have been found in several UASB reactors fed a carbohydrate substrate (Guiot et al., 1992; Rocheleau et al., 1999; Sekiguchi et al., 1999). In contrast, the epifluorescence image of small, lightcolored granules also sampled on day 66 shows that Bacteria and Archaea were clustered in a colony-like fashion and that some Archaea were located at the surface of these granules (Figure 4b, inside back cover). Biomass wasting on day 82 resulted in an increase in the levels of small, light-colored granules and a visual stratification of biomass in the compartments. After starvation (period 3), most large granules had disappeared and samples collected on day 123 showed a granular structure similar to the one depicted in Figure 4b or a layered structure with a mixture of Bacteria and Archaea within close vicinity (Figure 4c, inside back cover). Activity Tests. Differences in AMA between compartments were small during period 1, as demonstrated by the flat shape of the AMA curves (Figure 3, days 34 and 66). The initiation of biomass staging suggested by visible changes in granular structure at approximately day 66 was confirmed by AMA experiments with biomass sampled on day 90 (Figure 3). Moreover, by statistically comparing the AMA data from different compartments on days 0, Water Environment Research, Volume 74, Number 5
Angenent et al.
Table 3—Volatile fatty acid, SCOD, and pH levels of the liquid contents and methane and hydrogen levels of headspaces for the compartments of the AMBR during one flow cycle on day 147 (VLR ⴝ 35 g COD/L䡠d); samples were taken every hour during one 6-hour cycle. At t ⴝ 0 h, the initial compartment had been fed for 1 minute. At t ⴝ 4 h, the initial compartment had been fed for 4 hours, and the flow was redirected to the second compartment. At t ⴝ 6 h, the end of the first cycle was reached (second compartment had been fed for 2 hours). Data for t ⴝ 3 h are given in Table 2 (day 147; period 5). Time of cycle (h) 0
1
2
4
5
6
Compartment Parametera
Initial
Second
Middle
Fourth
Final
Formate Acetate Propionate Butyrate SCOD level pH Methane Formate Acetate Propionate Butyrate SCOD level pH Methane Formate Acetate Propionate Butyrate SCOD level pH Methane Formate Acetate Propionate Butyrate SCOD level pH Methane Formate Acetate Propionate Butyrate SCOD level pH Methane Formate Acetate Propionate Butyrate SCOD level pH Hydrogen Methane
0 290 243 0 1065 6.5 55 15 519 321 151 2475 6.0 27 11.3 698 432 259 2925 5.9 26 14 798 585 296.2 3075 5.7 27 0 823 529 211 2775 5.8 42 0 701 530 100 1950 5.8 1.1 ⫻ 10⫺4 43
0 433 305 0 1515 6.5 56 0 346 241 0 1050 6.4 53 0 420 298 70 2100 6.4 53 0 557 411 120 1650 6.1 53 3 542 387 168 1200 6.0 34 4.5 585 420 196 2100 5.8 2.7 ⫻ 10⫺2 25
0 628 383 140 2100 6.3 45 0 350 268 0 975 6.5 55 0 339 248 0 825 6.5 59 0 378 289 0 1050 6.4 62 0 411 293 0 1500 6.3 59 0 419 310 80 1050 6.2 3.0 ⫻ 10⫺4 47
30 681 410 278 3225 5.9 24 0 568 377 98 2175 6.3 41 0 424 290 0 975 6.4 49 0 349 270 0 1050 6.4 56 0 342 284 0 1080 6.3 58 0 393 281 0 1305 6.3 6.0 ⫻ 10⫺5 56
0 928 515 234 3375 6.0 40 0 665 406 111 1800 6.2 46 0 467 323 0 1515 6.4 49 0 393 310 0 1320 6.3 53 0 371 289 0 1200 6.2 54 0 335 289 0 1230 6.2 2.0 ⫻ 10⫺5 54
NOTE—atm ⫻ 101.3 ⫽ kPa. a Methane and hydrogen were measured in separated samples and are expressed as % and atm, respectively; VFA and SCOD are given in mg/L.
34, 66, 90, 123, 135, and 141 using a cluster analysis, a spatial discontinuity was found between days 66 and 90. In other words, changes in biomass staging based on methanogenic activity tests occurred between days 66 and 90. This coincided with the start of biomass wasting and reduction in biomass migration. On day 90, biomass from the middle compartment exhibited an AMA of 1.75 g COD/g VSS䡠d (n ⫽ 2), while biomass from compartments September/October 2002
1 and 5 had AMA values of 1.27 (n ⫽ 2) and 1.09 g COD/g VSS䡠d (n ⫽ 2), respectively (Figure 3). The AMA levels on day 123 were similar to those on day 90. In addition, a similar pattern was observed for PAA levels (0.27 [n ⫽ 2] and 0.18 g COD/g VSS䡠d [n ⫽ 4] for the middle and outside compartments, respectively) and an inverted staged pattern was detected for SAA values (2.21 [n ⫽ 2] and 3.64 meq. H⫹/g VSS䡠d [n ⫽ 4] for the middle and 457
Angenent et al.
Figure 5—Relative levels of 16S rRNA in each compartment over the operational time period obtained with: (a) Archaea probe (䊐), M. concilii probe ({), and Methanosarcina spp. probe (E); (b) Methanobacteriaceae probe (䊐), Methanomicrobiales probe ({), and Methanococcaceae probe (E); and (c) Desulfobulbus spp. probe (䊐), Syntrophomonadaceae probe ({), and Syntrophobacter spp. probe (E). outside compartments, respectively). As anticipated, the highest SAA was found in the outside compartments, in which most of the acidogenesis was taking place and conditions were less suitable for acetate-utilizing methanogens and acetogens resulting in low AMA and PAA levels. Membrane Hybridizations. On days 66, 90, 123, 135, and 141, biomass samples from all five compartments were obtained to characterize the microbial community structure. A limited amount of biomass staging was observed on day 66 using quantitative membrane hybridization data obtained with an archaeal probe and probes for aceticlastic methanogens (M. concilii and Methanosar458
cina spp.) and hydrogenotrophic methanogens (Methanomicrobiales, Methanobacteriaceae, and Methanococcaceae) (Figures 5a and 5b). This is in disagreement with results obtained from AMA tests, which suggested that biomass staging was initiated between days 66 and 90, but is consistent with visual observations of the aforementioned granular characteristics. For days 90, 123, 135, and 141, the biomass staging pattern based on membrane hybridization data became more pronounced (Figures 5a and 5b), which was consistent with results from activity tests (Figure 3). For example, the relative levels of M. concilii rRNA were approximately 35% in the middle compartment compared with approxiWater Environment Research, Volume 74, Number 5
Angenent et al.
mately 10% in the outside compartments on days 123 and 135. Relative signals of archaeal rRNA were higher than M. concilii rRNA in all compartments and probe nesting showed that relative rRNA levels of the family Methanobacteriaceae and the order Methanomicrobiales approximately made up the difference. Comparing the graphs within Figure 5b shows that relative signals of Methanobacteriaceae rRNA increased substantially after feeding was resumed following a period of starvation (period 4), and remained high during overload conditions (period 5). These results suggest that Methanobacteriaceae are more competitive at high hydrogen partial pressures. Despite a slight discrepancy between the hybridization outputs and AMA data for day 66, a significant association was shown between the M. concilii staging pattern (Figure 5a) and the AMA results (Figure 3) (Mantel’s r ⫽ 0.51; P ⬍ 0.1 because of similarities between these two data sets for days 90, 123, 135, and 141). However, a stronger association was anticipated (and thus a higher Mantel’s r) because almost all aceticlastic methanogens present in the AMBR were M. concilii. This may indicate that the AMA test is not a very sensitive method to evaluate population abundance. Relative rRNA levels of syntrophic bacteria were determined with probes for Syntrophomonadaceae and Syntrophobacter spp. In addition, a Desulfobulbus spp. probe was used to determine the relative abundance of propionate-degrading sulfate-reducing bacteria. The concentration of sulfate in the tap water used to dilute the feedstock solution was approximately 100 mg/L (information from city of Ames, Iowa). The hybridization results obtained with these three probes did not reveal a clear staging pattern, although some staging was apparent for Syntrophomonadaceae and Syntrophobacter spp. (Figure 5c). During period 5, which was characterized by overload conditions (days 135 and 141), relative levels of Syntrophobacter spp. rRNA were lower than those during the earlier periods of operation (days 66, 90, and 123), while relative signals of Syntrophomonadaceae and Desulfobulbus spp. were similar for all sampling points. This could indicate that Syntrophobacter spp. are more susceptible to stressed conditions such as low pH and high VFA levels. This is consistent with the presence of some degree of staging for Syntrophobacter spp. (Figure 5c) in that their rRNA levels were generally lowest in the outside compartments, which exhibited the lowest pH values and highest VFA concentrations (Tables 2 and 3). The biomass staging pattern became more pronounced in the AMBR with time, particularly during the period of overload conditions (Figures 5a and 5b). On day 141, the relative rRNA levels of methanogens in the outside compartments had decreased substantially compared with those on the previous sampling days. This decrease in relative concentrations may be the result of an increase in the absolute levels of fermenters resulting from the availability of more substrate during overload conditions, a decrease in the absolute levels of methanogens resulting from adverse conditions (e.g., lower pH), or a combination of both. On day 147, the acetate concentration in the final compartment was slightly higher than the acetate concentration in the fourth compartment, whereas the reverse was observed for the other sampling points (Table 2). These data indicate that the aceticlastic methanogenic activity and the levels of M. concilii in the outside compartments had dropped. Thus, during periods of overload, excessive biomass staging can be detrimental to reactor performance. To ensure effluent quality (low levels of VFAs), the levels of aceticlastic methanogens should be maintained at a sufficiently high level in all compartSeptember/October 2002
ments. This can be accomplished by increased biomass migration or biomass recycling while maintaining some degree of biomass staging. During overload conditions, relative rRNA levels of methanogens in the middle compartments remained high (Figures 5a and 5b). By maintaining high levels of methanogens in part of the AMBR, the reactor has the potential to recover quickly after severe overload conditions. In contrast, research with a staged, unidirectional reactor (upflow staged sludge bed reactor) showed that granules with a high acidogenic activity grown in the first compartment were distributed throughout the reactor, triggering the deterioration of reactor performances after 8 months of operation (van Lier, 1995). In other words, the fast-growing fermenters had diluted methanogens and acetogens, which decreased the AMA and PAA of the biomass in the final compartments. Feast and Famine Conditions for the AMBR. Besides biomass quantity and quality control and maintenance of methanogenic conditions in all compartments, reversing the flow in the AMBR results in feast and famine conditions (alternating high and low substrate levels). In single-vessel anaerobic systems, such as the anaerobic sequencing batch reactor, feast and famine conditions are synonymous with substrate staging (Sung and Dague, 1995). For the compartmentalized AMBR, feast and famine conditions occurred in combination with biomass staging, although some of the granular biomass was cycled through the system because of biomass migration. Therefore, the biomass in the AMBR was temporarily exposed to very high concentrations of substrate and intermediates (VFA and hydrogen). This resulted in high relative rRNA signals for syntrophic propionate-oxidizing bacteria (SPOB) and hydrogenotrophic methanogens. The averages of the sums of the relative 16S rRNA levels were 3.5% for Syntrophomonadaceae, Syntrophobacter spp., and Desulfobulbus spp. (SPOB), and 7.4% for Methanobacteriaceae, Methanomicrobiales, and Methanococcaceae (hydrogenotrophic methanogens). In individual compartments, these levels were as high as 6 and 13%, respectively. Instead, single-vessel anaerobic reactors that were operated under stable conditions (and, hence, low VFA and hydrogen levels) consisted of biomass with considerably lower rRNA levels for these organisms (Griffin et al., 1998; McMahon et al., 2001). Moreover, McMahon et al. (2001) observed that digesters with a history of unstable operation reduced propionate faster (and, thus, stabilized sooner) compared with digesters with a stable history because of the higher rRNA levels of SPOB and hydrogenotrophic methanogens in the digesters with the unstable history. This observation is in accordance with Harper and Pohland (1986) and McCarty and Mosey (1991), who predicted inadequate removal of hydrogen and propionate during a sudden increase in VLR for previously stable anaerobic reactors. Methanosaeta spp. and Methanosarcina spp. The acetateutilizing methanogen M. concilii was abundant, while Methanosarcina spp. were virtually absent in all compartments of the AMBR throughout the operational period (Figure 5a). This result was not anticipated because acetate concentrations were always greater than 200 mg/L in the initial compartments (Table 2). It is generally accepted that M. concilii with a high affinity for acetate is dominant in systems with low acetate concentrations, while Methanosarcina spp. are more competitive in systems with high acetate concentrations (Huser et al., 1982; Jetten et al., 1992; Smith and Mah, 1978). Because methanogens were present on the surfaces of some granules (Figure 4b), they were exposed to the acetate levels in the bulk liquid, suggesting that mass-transfer 459
Angenent et al.
limitations were not the reason for the absence of Methanosarcina spp. Grotenhuis et al. (1992) linked the hydrophobic and filamentous character of M. concilii to a high adherence capacity. Therefore, a competitive advantage of M. concilii over Methanosarcina spp. for the AMBR may exist because of a better retention of M. concilii. This is especially important for the AMBR because high mechanical shear and biomass migration may have caused washout of the less hydrophobic and less adhering Methanosarcina spp. Conclusions The AMBR was able to achieve SCOD removal efficiencies greater than 97% at a VLR of 42 to 45 g COD/L䡠d when fed a nonacidified sucrose substrate solution. A combination of a compartmentalized design, mechanical mixing, and reversing the flow over the horizontal plane of the reactor resulted in this performance. Biomass staging was observed in the AMBR in that relative 16S rRNA signals for Archaea, such as M. concilii and Methanobacteriaceae, were higher in the middle compartment compared with the outside compartments. Long-term overload conditions resulted in too much biomass staging, and low levels of acetate-utilizing methanogens in the outside compartments diminished acetate degradation in the final compartment. Meanwhile, Archaea (methanogen) levels remained high in the middle compartment, which allowed a fast recovery after unstable conditions. Thus, it was demonstrated that recycling of biomass between compartments was important to maintain sufficient levels of acetate-utilizing methanogens in the outside compartments. In conclusion, the results indicated that it was very helpful to monitor microbial population levels (in particular, levels of aceticlastic methanogens) to help improve the operating performance of the AMBR. Future work should focus on the development of rapid, quantitative molecular tools that can be used to monitor microbial population dynamics in parallel with performance measures to help make rational decisions about operating conditions. Acknowledgments Credits. This research was supported by grants from the U.S. Department of Agriculture (contract number 91-34188-5943), Washington, D.C., through the Iowa Biotechnology Byproducts Consortium, Ames, Iowa, and from the University of Illinois at Urbana-Champaign, Urbana, through its Critical Research Initiative Program. The authors acknowledge Pu Yong, New Jersey Institute of Technology, Newark, for analytical assistance; Dominic Frigon, University of Illinois at Urbana-Champaign, for assistance with statistical analyses; Willy Verstraete, University of Gent, Belgium, for providing an acidogenic activity protocol; and Daniel F. Voytas, Iowa State University, Ames, for providing space in his – 80 °C freezer. Authors. At the time of this work, Largus T. Angenent and Shihwu Sung were graduate student and assistant professor, respectively, in the Department of Civil and Construction Engineering at Iowa State University, Ames, and Dandan Zheng and Lutgarde Raskin were graduate student and associate professor, respectively, in the Department of Civil and Environmental Engineering at the University of Illinois at Urbana-Champaign, Urbana. Largus T. Angenent is currently an assistant professor in the Environmental Engineering Science Program at Washington University, St. Louis, Missouri. Dandan Zheng is a principal microbiologist for Alpha Therapeutic Corporation, Los Angeles, California. Correspondence should be addressed to Lutgarde Raskin, Associate Professor, University of Illinois at Urbana-Champaign, 460
Civil and Environmental Engineering, 3221 Newmark Civil Engineering Laboratory, MC-250, Urbana, IL 61801; e-mail:
[email protected]. Submitted for publication February 11, 2002; revised manuscript submitted and accepted for publication June 20, 2002. The deadline to submit Discussions of this paper is January 15, 2003. References Alm, E.; Oerther, D.; Larsen, N.; Stahl, D.; Raskin, L. (1996) The Oligonucleotide Probe Database. Appl. Environ. Microbiol., 62, 3557. Amann, R. I.; Krumholz, L.; Stahl, D. A. (1990) Fluorescent-Oligonucleotide Probing of Whole Cells for Determinative, Phylogenetic, and Environmental Studies in Microbiology. J. Bacteriol., 172, 762. Angenent, L. T.; Abel, S.; Sung, S. (in press) Effect of an Organic Shock Load on the Stability of AMBR. J. Environ. Eng., in press. Angenent, L. T.; Sung, S. (2001) Development of Anaerobic Migrating Blanket Reactor (AMBR), a Novel Anaerobic Treatment System. Water Res., 35, 1739. American Public Health Association; American Water Works Association; Water Environment Federation (1995) Standard Methods for the Examination of Water and Wastewater, 19th ed.; Washington, D.C. Bachman, A.; Beard, V. L.; McCarthy, P. (1985) Performance Characteristics of the Anaerobic Baffled Reactor. Water Res., 19, 99. de los Reyes, F. L.; Ritter, W.; Raskin, L. (1997) Group-Specific Small Subunit rRNA Hybridization Probes to Characterize Filamentous Foaming in Activated Sludge System. Appl. Environ. Microbiol., 63, 1107. Devereux, R.; Kane, M. D.; Wilfrey, J.; Stahl, D. (1992) Genus- and Group-Specific Hybridization Probes for Determinative and Environmental Studies of Sulfate-Reducing Bacteria. Syst. Appl. Microbiol., 15, 601. Duran, M.; Speece, R. E. (1998) Staging of Anaerobic Processes for Reduction of Chronically High Concentrations of Propionic Acid. Water Environ. Res., 70, 241. Dutilleul, P.; Stockwell, J. D.; Frigon, D.; Legendre, P. (2000) The Mantel Test versus Pearson’s Correlation Analysis: Assessment of the Differences for Biological and Environmental Studies. J. Agric., Biol., Environ. Stud., 5, 131. Ellis, T. G.; Barbeau, D. S.; Smets, B. F.; Grady, C. P. L., Jr. (1996) Respirometric Technique for Determination of Extant Kinetic Parameters Describing Biodegradation. Water Environ. Res., 68, 917. Fox, P.; Pohland, F. G. (1994) Anaerobic Treatment Applications and Fundamentals: Substrate Specificity During Phase Separation. Water Environ. Res., 66, 716. Griffin, M. E.; McMahon, K. D.; Mackie, R. I.; Raskin, L. (1998) Methanogenic Population Dynamics During Start-Up of Anaerobic Digesters Treating Municipal Solid Waste and Biosolids. Biotechnol. Bioeng., 57, 342. Grotenhuis, J. T. C.; Kissel, J. C.; Plugge, C. M.; Stams, A. J. M.; Zehnder, A. J. B. (1991) Role of Substrate Concentration in Particle Size Distribution of Methanogenic Granular Sludge in UASB Reactors. Water Res., 25, 21. Grotenhuis, J. T. C.; Plugge, C. M.; Stams, A. J. M.; Zehnder, A. J. B. (1992) Hydrophobicities and Electrophoretic Mobilities of Anaerobic Bacterial Isolates from Methanogenic Granular Sludge. Appl. Environ. Microbiol., 58, 1054. Guiot, S. R.; Pauss, A.; Costerton, J. W. (1992) A Structured Model of the Anaerobic Granule Consortium. Water Sci. Technol., 25 (7), 1. Guiot, S. R.; Safi, B.; Frigon, J. C.; Mercier, P.; Mulligan, C.; Tremblay, R.; Samson, R. (1995) Performances of a Full-Scale Novel Multiplate Anaerobic Reactor Treating Cheese Whey Effluent. Biotechnol. Bioeng., 45, 398. Hansen, K. H.; Ahring, B. K.; Raskin, L. (1999) Quantification of Syntrophic Fatty Acid--Oxidizing Bacteria in a Mesophilic Biogas Reactor Water Environment Research, Volume 74, Number 5
Angenent et al. by Oligonucleotide Probe Hybridization. Appl. Environ. Microbiol., 65, 4767. Harmsen, H. J.; Kengen, H. M.; Akkermans, A. D.; Stams, A. J.; de Vos, W. M. (1996) Detection and Localization of Syntrophic PropionateOxidizing Bacteria in Granular Sludge by In Situ Hybridization Using 16S rRNA-Based Oligonucleotide Probes. Appl. Environ. Microbiol., 62, 1656. Harper, S. R.; Pohland, F. G. (1986) Recent Developments in Hydrogen Management During Anaerobic Biological Wastewater Treatment. Biotechnol. Bioeng., 28, 585. Hulshoff Pol, L.; Euler, H.; Eitner, A.; Grohganz, D. (1997) GTZ Sectorial Project “Promotion of Anaerobic Technology for the Treatment of Municipal and Industrial Sewage and Wastes”. In Proceedings of the 8th International Conference on Anaerobic Digestion, Sendai, Japan; International Association on Water Quality: London. Huser, A.; Wuhrmann, K.; Zehnder, A. (1982) Methanothrix soehngenii gen. nov. spec. nov., a New Acetotrophic Non-Hydrogen-Oxidizing Methane Bacterium. Arch. Microbiol., 132, 1. 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, 181. Legendre, P.; Legendre, L. (1998) Numerical Ecology; Elsevier Science B.V.: Amsterdam. McCarty, P. L.; Mosey, F. E. (1991) Modeling of Anaerobic Digestion Processes (A Discussion of Concepts). Water Sci. Technol., 24 (8), 17. McMahon, K. D.; 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, 1817. Raskin, L.; Poulsen, L. K.; Noguera, D. R.; Rittman, B. E.; Stahl, D. A. (1994a) Quantification of Methanogenic Groups in Anaerobic Biological Reactors by Oligonucleotide Probe Hybridization. Appl. Environ. Microbiol., 60, 1241. Raskin, L.; Stromley, J. M.; Rittmann, B. E.; Stahl, D. A. (1994b) GroupSpecific 16S rRNA Hybridization Probes to Describe Natural Communities of Methanogens. Appl. Environ. Microbiol., 60, 1232. Raskin, L.; Zheng, D.; Griffin, M. E.; Stroot, P. G.; Misra, P. (1995) Characterization of Microbial Communities in Anaerobic Bioreactors Using Molecular Probes. Antonie van Leeuwenhoek, 68, 297. Rinzema, A.; van Lier, J.; Lettinga, G. (1988) Sodium Inhibition of Acetoclastic Methanogens in Granular Sludge from a UASB Reactor. Enzyme Microb. Technol., 10, 24. Rocheleau, S.; Greer, C. W.; Lawrence, J. R.; Cantin, C.; Larame´ e, L.; Guiot, S. R. (1999) Differentiation of Methanosaeta concilii and Methanosarcina barkeri in Anaerobic Mesophilic Granular Sludge by Fluorescent In Situ Hybridization and Confocal Scanning Laser Microscopy. Appl. Environ. Microbiol., 65, 2222. Sajjad, M. W.; Cleasby, J. L. (1995) Effect of Impeller Geometry and Various Mixing Patterns on Flocculation Kinetics of Kaolin Clay Using Ferric salts. In Proceedings of the 1995 Annual Conference of the American Water Works Association, Denver, Colorado.
September/October 2002
Scott, W. J. (1957) Water Relations of Food Spoilage Microorganisms. Adv. Food Res., 7, 83. Sekiguchi, Y.; Kamagata, Y.; Nakamura, K.; Ohashi, A.; Harada, H. (1999) Flourescence In Situ Hybridization Using 16S rRNA-Targetted Oligonucleotides Reveals Localization of Methanogens and Selected Uncultured Bacteria in Mesophilic and Thermophilic Sludge Granules. Appl. Environ. Microbiol., 65, 1280. Smith, M. R.; Mah, R. A. (1978) Growth and Methanogenesis by Methanosarcina Strain 227 on Acetate and Methanol. Appl. Environ. Microbiol., 36, 870. Stahl, D. A.; Amann, R. I. (1991) Development and Application of Nucleic Acid Probes. In Nucleic Acids Techniques in Bacterial Systematics; Stackebrandt, E., Goodfellow, M., Eds.; Wiley & Sons: New York; p 205. Stahl, D. A.; Flesher, B.; Mansfield, H. R.; Montgomery, L. (1988) Use of Phylogenetically Based Hybridization Probes for Studies of Ruminal Microbial Ecology. Appl. Environ. Microbiol., 54, 1079. Stams, A. J. M. (1994) Metabolic Interactions Between Anaerobic Bacteria in Methanogenic Environments. Antonie van Leeuwenhoek, 66, 271. Sung, S.; Dague, R. R. (1995) Laboratory Studies on the Anaerobic Sequencing Batch Reactor. Water Environ. Res., 67, 294. Thauer, R. K.; Jungermann, K.; Decker, K. (1977) Energy Conservation in Chemotrophic Anaerobic Bacteria. Bacteriol. Rev., 41, 100. Vanderhaegen, B.; Ysebaert, E.; Favere, K.; van Wambeke, M.; Peeters, T.; Pa´ nic, V.; Vandenlangenbergh, V.; Verstraete, W. (1992) Acidogenesis in Relation to In-Reactor Granule Yield. Water Sci. Technol., 25 (7), 21. van Lier, J. B.; Groeneveld, N.; Lettinga, G. (1996) Development of Thermophilic Methanogenic Sludge in Compartmentalized Upflow Reactors. Biotechnol. Bioeng., 50, 115. van Lier, J. B.; Rebac, S.; Lens, P.; van Bijnen, F.; Oude Elferink, S. J. W. H.; Stams, A. J. M.; Lettinga, G. (1997) Anaerobic Treatment of Partly Acidified Wastewater in a Two-Stage Expanded Granular Sludge Bed (EGSB) System at 8 °C. Water Sci. Technol., 36 (6 –7), 317. van Lier, J. B. (1995) Thermophilic Anaerobic Wastewater Treatment; Temperature Aspects and Process Stability; Environmental Engineering, Wageningen Agricultural University: Wageningen, The Netherlands; p 181. Zehnder, A. J. B.; Huser, B. A.; Brock, T. D.; Wuhrmann, K. (1980) Characterization of an Acetate-Decarboxylating, Non-Hydrogen-Oxidizing Methane Bacterium. Arch. Microbiol., 124, 1. Zheng, D.; Alm, E. W.; Stahl, D. A.; Raskin, L. (1996) Characterization of Universal Small-Subunit rRNA Hybridization Probes for Quantitative Molecular Microbial Ecology Studies. Appl. Environ. Microbiol., 62, 4504. Zheng, D.; Raskin, L. (2000) Quantification of Methanosaeta Species in Anaerobic Bioreactors Using Genus- and Species-Specific Hybridization Probes. Microbial Ecol., 39, 246.
461
From Angenent et al.:
Figure 4 —Epifluorescence micrograph of a FISH experiment with a FITC (green)–labeled probe for Bacteria (S-D-Bact-0338-a-A-18) and a TRITC (red)–labeled probe for Archaea (S-D-Arch-0915-a-A-20) on 5 m-thick sectioned granules; superimposition of image fields obtained with the TRITC and FITC filter sets results in a predominantly yellow color when Bacteria and Archaea were in a close vicinity to each other. (a) Large, layered granule sampled from compartment 2 on day 66. The bar is equal to 100 m. (b) Small granule sampled from compartment 1 on day 66. The bar is equal to 50 m. (c) Granule sampled from compartment 2 on day 123. The bar is equal to 100 m.
From Biesterfeld and Figueroa:
Figure 3—Microscope FISH images: image A (red) is the autofluorescent area, image B (green) is the EUB338 stained area for the same microscope field, and image C (grey) is the Nso190 stained area. The Nso190 probe is labeled with Cy5 and is in the very far red of the spectrum, not visible to the human eye. Each color or channel is collected separately. Magnification is 400ⴛ. This sample is from a 36-day-old biofilm collected from sample port 2 of the NTF.
