Use of Fluorescent Oligonucleotide Probes to Characterize Vertical Population Distributions of Nitrifying Bacteria in a Full-Scale Nitrifying Trickling Filter
Sidney Biesterfeld1,2, Linda Figueroa1*, Mark Hernandez3, and Phil Russell2
1 Division of Environment Science and Engineering, Colorado School of Mines, Golden, Colorado 80401
2 Littleton/Englewood Wastewater Treatment Plant, Englewood, Colorado 80110
3 Department of Civil, Environmental, and Architectural Engineering, University of Colorado, Boulder, Colorado 80309
Presented at the IAWQ Conference on
Microbial Ecology of Biofilms: Concepts, Tools and Applications
Lake Bluff, IL on October 8-10, 1998
Environmental Science and Engineering Division
Colorado School Mines
Golden, CO 80401
Phone: (303) 273-3491
FAX: (303) 273-3413
USE OF FLUORESCENT OLIGONUCLEOTIDE PROBES TO CHARACTERIZE VERTICAL POPULATION DISTRIBUTIONS OF NITRIFYING BACTERIA IN A FULL-SCALE
NITRIFYING TRICKLING FILTER
Sidney Biesterfeld1,2, Linda Figueroa1, Mark Hernandez3, and Phil Russell2
1Department of Environmental Science and Engineering, Colorado School of Mines,
Golden, Colorado 80401
2Littleton/Englewood Wastewater Treatment Plant, Englewood, Colorado 80110
3Department of Civil, Environmental, and Architectural Engineering, University of Colorado, Boulder,
Process control and optimization of nitrifying trickling filters are currently based on macroscale chemical parameters and bulk measures of biomass (e.g. volatile solids). A relatively new microbiological technique that can provide phenotype information from an intact biofilm is Fluorescent In Situ Hybridization (FISH). The application of FISH to characterize nitrifying populations in full-scale wastewater treatment systems is limited and recent (Wagner et al. 1995, Mobarry et al. 1996, Schramm et al. 1996). FISH can yield important information about bacterial populations that could be correlated with system performance. Quantification of the active population may be used for process modeling and optimization.
The main research objective is to investigate the use of FISH to characterize intact biofilm samples from various depths within a nitrifying trickling filter and to determine whether this information can be used to improve process modeling and optimization. The research hypothesis is that the vertical population distribution of nitrifying bacteria as determined by FISH will correlate to measured changes in the concentrations of ammonia, nitrate, and nitrite. Preliminary data indicates a good correlation between the quantified bacterial population at a given sample location and the amount of ammonia removed.
Ammonia; Biofilm; FISH; Nitrification; Oligonucleotide; Trickling filter
Nitrification is carried out by two unrelated groups of chemolithotrophic bacteria. The first group converts ammonia to nitrite, which is then converted to nitrate by the second group. The best known catalysts of these reactions are Nitrosomonas and Nitrobacter respectively, although other groups have also been identified (Belser 1979, Bergey’s 1989). Taken together, these bacteria can be used to ameliorate damage to the environment by reducing the ammonia content of wastewater prior to discharge. Effective ammonia removal is extremely important as ammonia is toxic to aquatic life and creates a large oxygen demand (USEPA 1993).
While bulk chemical methods have been used to determine rates of nitrification, very little is actually known about the number, placement, and activity of the bacteria themselves. This is partly because conventional enumeration techniques, such as MPN and selective plating, are not suited to these slow growing organisms (Brock et al. 1987).
A rapid and reliable method for enumerating these organisms would be invaluable since their number and activity are the limiting parameters for ammonia removal (Belser 1979). Additionally, nitrification can be difficult to maintain in engineered systems as Nitrosomonas and Nitrobacter are sensitive to changes in temperature and pH as well as a wide variety of organic compounds (USEPA 1993). For this study, a relatively new microbiology technique, Fluorescent In Situ Hybridization (FISH), has been used to quantify the ammonia and nitrite-oxidizing populations at various depths within a nitrifying tricking filter. Population data is then compared to ammonia removal data for the same time period.
In situ hybridization is the process of annealing a small fragment of DNA or RNA to a specific target strand of DNA or RNA in a morphologically preserved cell. Under the right conditions, this fragment or probe, will bind to a specific genetic sequence within that cell. Because all organisms contain some unique sequences while having others in common, probe specificity is freely adjustable. The target may be as specific as a single organism or as broad as all prokaryotes. Probes used for this study target the 16S subunit of ribosomal RNA. Ribosomal RNA is an ideal target molecule as a single cell may contain as many as 10,000 ribosomes – each containing a copy of the target sequence (Manz et al. 1994). To detect probes after hybridization, a reporter molecule must be attached. Fluorescent dyes are used because they are easily read and recorded using a fluorescent microscope. If the complementary sequence is not present, hybridization does not occur and the probe is washed from the cell.
To date, this technique has been used successfully to characterize sulfate-reducing biofilms (Amann et al. 1992, Poulsen et al. 1993), to study methanogenic bacteria (Raskin et al. 1995), and to investigate foam-causing bacteria in activated sludge (De Los Reyes et al. 1998). Further, oligonucleotide sequences corresponding to regions of the 16S subunits of various ammonia and nitrite-oxidizing bacteria have been published (Mobarry et al. 1996, Wagner et al. 1995). These probes have used to characterize nitrifying bacteria in wastewater treatment systems (Wagner et al. 1997, Wagner et al. 1996, Schramm et al. 1996, Urbain et al. 1997). Schramm et al. (1996) reported successful in situ hybridization for both ammonia and nitrite-oxidizers in nitrifying biofilm samples.
METHODS AND MATERIALS
Sampling location. The Littleton/Englewood Wastewater Treatment Plant, Englewood, Colorado, has three nitrifying trickling filters (NTFs) that are operated in parallel. These filters were designed with 24 ft of media depth, are 105 feet in diameter, have an average hydraulic loading rate of 1.23 gpm/sf, and cross-flow media with a specific surface area of 42 sf/cf. High ammonia centrate solution is returned to the NTFs during periods of low flow to minimize diurnal variations which could result in reduced performance. Nine capped sampling ports are spaced approximately 2.5 feet apart from top to bottom and extend approximately five feet into the filter.
Chemical analysis. Grab and composite samples of trickle filter influent and effluent were collected three times per week and analyzed for ammonia, pH, alkalinity, CBOD, and TSS. Grab samples were taken from the secondary clarifier effluent, trickle filter influent, each trickle filter sampling port, and trickle filter effluent once monthly. These monthly sampling events are referred to as profile sampling. All samples were analyzed by accepted USEPA methods or in accordance with Standard Methods for the Examination of Water and Wastewater (1995).
Biofilm samples. Biofilm samples were collected on glass microscope slides from sampling ports 2, 4, 6, and 8. Slides were held in place for 10 days with a section of PVC pipe extending five feet into the filter media. The PVC was notched at regular intervals to accommodate up to eight plastic slide holders. All slides were held perpendicular to flow to minimize disruptions of the forming biofilm and to ensure uniform growth. The removal of slides with biofilm samples was coordinated with the monthly profile sampling i.e. 10 days prior to a scheduled profile sampling, slides were placed into the filter.
Biofilm preparation. Individual slides were collected in sterile, disposable 50 ml centrifuge tubes and submerged in fresh, cold fixative solution (4% paraformaldehyde in PBS, pH 7.2) at 4oC for 6-8 hours (Amann et al. 1990a). Excess fixative was removed by immersing slides twice for three minutes in PBS followed by ethanol dehydration as previously described (Manz et al. 1992). Prepared slides that could not be hybridized immediately were stored in a dessicator at 4oC for up to two weeks (Amann et al. 1992a).
Reference cells. Cultures of Nitrosomonas eutropha, Nitrosomonas europaea, and Nitrobacter winogradsky were obtained from the American Type Tissue Collection (ATCC). Krummel and Harms 1992 media recipe (Bergey’s 1989) and ATCC culture media recipe 480 were used to maintain active cultures of Nitrosomonas and Nitrobacter winogradsky, respectively. Cultures were maintained in a shaker at 20oC. Nitrosomonas cultures were checked daily for pH. Adjustments were made with sterile 5% sodium carbonate solution as needed to maintain the pH between 7.2 and 7.8. For all cultures, substrate and waste product concentrations were checked monthly. Reference cultures were fixed as previously described (Amann et al. 1990b) and stored in a mixture of 50% PBS and 50% ethanol at –20oC indefinitely.
Oligonucleotide probes. Probe sequences for NEU23a, NIT2, NIT3, and CNIT3 were taken from Schramm et al. (1996) and Mobarry et al. (1996). NIT2 and NIT3 are specific for Nitrobacter and NEU23a is specific for halophilic and halotolerant members of the genus Nitrosomonas and Nitrosococcus mobilis. The competitor oligonucleotide CNIT3 was also used to enhance specificity.
Modified oligonucleotides were purchased directly from Gemini Biotech, located in The Woodlands, Texas. Reporter molecules attached were Cy2 for NEU23a and BODIBY-TMR for both NIT2 and NIT3. Purification was accomplished by reverse phase HPLC and size confirmation done by 20% poly-acrylamide gel electrophoresis.
In situ hybridization. Two slides from each sampling port were hybridized. Positive and negative controls of previously prepared aliquots of Nitrosomonas eutropha, Nitrosomonas europaea, and Nitrobacter winogradsky were spotted onto clean microscope slides, ethanol series dehydrated, and hybridized alongside the sample slides. Generally, the hybridization protocol described by Manz et al. (1993) was followed. Some modifications were necessary because the biofilm samples were collected on regular pre-cleaned microscope slides. Aliquots of 100 ul of hybridization solution (0.9 M NaCl, 0.01% Sodium dodecyl sulfate, 20 mM Tris-HCl [pH 8.0], 500 ng each probe) were deposited into disposable, RNAse free HybriSlips (American Histology Company). Prepared slides were lowered onto the HybriSlips, placed into airtight plastic containers, and incubated at 46oC for five hours. A 20-minute wash step was performed with probe free hybridization solution as previously described (Schramm et al. 1996). Slides were permanently mounted with ProLong (Molecular Probes) and dried completely before viewing.
Microscopy and image analysis. Biofilm samples were examined with a Leitz Orthopol microscope fitted for fluorescence with a 100 W high pressure mercury source. The primary filter set consisted of a narrow band pass exciter filter with an excitation maximum of 488 nm and a wide band pass barrier filter with a 525 nm cutoff. A 1.4 million pixel, cooled CCD color SPOT camera captured images for digital analysis. Exposure times were fixed at 5 seconds. Images were analyzed with Simple PCI version 1.0.
Ten fields from each sample slide were randomly selected for analysis for a total of twenty fields examined per sample port. Because biofilms do not grow uniformly, the differences in signal intensity between fields was notably variable. To minimize this influence, signal intensity readings for all twenty fields from each sample port were averaged to obtain the signal intensity per port.
RESULTS AND DISCUSSION
Profile data have been collected monthly since July of 1996, a total of twenty-three sampling events. Throughout this period, flows and influent ammonia concentrations to the filters have remained relatively constant. Influent ammonia concentrations averaged 26 +/- 5 mg/L over two years. To reduce the amount of variability in the data and because biofilm samples were only taken from every other sampling port, the filter was broken into four discrete segments. The first segment begins at the NTF influent and ends with sampling port 2. The second segment begins at port 2 and ends at port 4. The third and fourth segments run from ports 4 to 6 and ports 6 to 8. Segment one is 6.5 feet high and all other segments are 5.0 feet high. The relative standard deviations for the amount of ammonia removed in mg/L for each segment are between 27% and 77%. There is also a decline with depth in the amount of ammonia removed.
To date, three sets of biofilm data have been collected. The first set of slides was removed from the trickling filter after a 10-day incubation period on June 7, 1998 and hybridized ten days later. Ten fields on each slide were examined using a 400x objective and FITC filter set. When NIT2 and NIT3 probes, specific for Nitrobacter, were used, signals were observed in both the biofilm and the positive control suggesting that Nitrobacter are present. However, the observed signals faded rapidly making quantification difficult. Schramm et al. (1996) also reported the presence of Nitrobacter in an intact nitrifying biofilm sample when using both NIT2 and NIT3 tagged with Cy3. The observed fading may have resulted from the use of BODIPY-TMR, a lower intensity dye. Another reason for weak signal may be low cell density or ribosome content of the biofilm and control samples. Biofilm samples examined by Schramm et al. (1996) were taken from a nitrifying filter treating aquaculture waters that can contain ammonia concentrations five times higher than municipal wastewater. Additionally, Wagner et al. (1996) have suggested that other groups of nitrite-oxidizing bacteria probably exist and may account for a significant amount of observed nitrite oxidation in wastewater plants. Future work will utilize a brighter dye i.e. Cy3 that is less prone to fading and additional probes for other species of nitrite-oxidizers (e.g., nitrospira). Signal intensities given in Figures 3-5 are for probe NEU23a; specific for Nitrosomonas and Nitrosococcus.
Average field intensities for each segment were compared to the averaged aqueous phase grab sample profile data collected on May 27, 1998 and June 24, 1998. The nitrification rate data from the two profile events was averaged in this case because the biofilm samples were collected between profile events. The amount of ammonia removal observed correlates well with the average signal intensity observed per segment resulting in a correlation coefficient for this data set of 0.975.
A second set of slides was removed from the trickling filter on June 24, 1998 and hybridized four days later. Aqueous phase profile samples were collected at the same time. Ten fields on each slide were examined using a 1000x objective and FITC filter set for a total of twenty fields per filter segment. The correlation between the amount of ammonia removed and average signal intensity for the second data set is 0.951
A third set of slides and aqueous phase samples were collected from the trickling filter on September 24, 1998. The correlation between the amount of ammonia removed and average signal intensity for the third data set is 0.970.
Ammonia removal rates for the profile sampling events are between zero and 2.38 g N/m2/day which are consistent with apparent zero-order nitrification rates of 0.93 to 1.8 g N/m2/day reported by USEPA (1993). Removal rates appear to be dependent upon influent loading and filter segment examined. Suppression of removal rate has been observed when influent CBOD5 exceeds 10 mg/L and/or when the influent exceeds 10 mg/L TSS (Parker et al. 1986).
For the September 24th profile, influent TSS and CBOD5 concentrations, averaged for ten days prior to sampling, were 9.7 mg/L and 11 mg/L, respectively. As expected, the highest removal rate is in the first segment. For the June 24th profile event, influent TSS and CBOD5 concentrations, also averaged for ten days prior to sampling, were 19 mg/L and 12 mg/L, respectively. As predicted by Parker et al. (1986), ammonia removal was suppressed in the first segment dropping to 1.22 g/m2/day. Signal intensity data for Nitrosomonas and Nitrosococcus shows a similar trend implying that they are not just inhibited, but are unable to survive under higher CBOD5 and TSS loading.
Observed signal intensities for the June 24th and September 24th profile samplings are much higher than those observed for the May 27th sampling. This difference results from the May 27th slides being observed at 400x and the remaining sets at 1000x. This may be compensated for by multiplying the signal intensities at 400x by 2.5. After correction, the correlation coefficient for all data points combined is 0.826.
Nitrification rates and signal intensity are much lower in the bottom half of the filter for all profile samplings. This may reflect patchy biofilm development due to an inadequate food supply or because of alkalinity limitations. Groeneweg et al. (1994) showed that the amount of ammonia required at pH 6.0 and 20oC to meet the half saturation constant of ammonia-oxidizers was 6.16 mg/L as N. Below this concentration, uptake by ammonia-oxidizers could be suppressed. Over the last two years, the average 24-hour composite ammonia effluent concentration was 3.9 mg/L. For the May 27th profile sampling, ammonia concentrations drop below 6.16 mg/L at the beginning of segment three – corresponding to a significant decrease in the rate of removal down to 0.75 g/m2/day. For the June 24th event, ammonia concentrations don’t fall below the threshold until the middle of segment three. Consequently, the observed removal rate for segment three is slightly higher at 0.97 g/m2/day. On September 24th, the ammonia concentration drops below this threshold in the middle of segment two.
Alkalinity concentrations below 50 to 150 mg/L as CaCO3 are thought to reduce the maximum rate of nitrification (Siegrist and Gujer 1987, Szwerinski et al. 1986, Gujer and Boller, 1984). Alkalinity data averaged over the last two years of profile data drops below 150 mg/L as CaCO3 at the bottom of segment two and frequently fell below 100 mg/L at the filter effluent. Low alkalinity may be contributing to the decreased removal rates observed in segments three and four.
Well-formed Nitrosomonas/Nitrosococcus colonies were found throughout the filter, although smaller and less numerous in the bottom half. Colonies were round or elliptical and up to 60 um across. This characteristic shape has been observed previously in activated sludge flocs (Wagner et al. 1995. Wagner et. al. 1997, Mobarry et al. 1996). Frequently, several colonies were found grouped together. Figure 6 shows a Nitrosomonas eutropha control hybridized with NEU23a. Figure 6 shows a cluster of colonies from the second filter segment that were hybridized in situ. Individual bacteria can easily be distinguished inside the larger colonies.
Figure 6 Nitrosomonas eutropha control.
Figure 7 Ammonia-oxidizer colonies in situ.
Both Nitrobacter and Nitrosomonas/Nitrosococcus were found in NTF biofilm samples using FISH, a technique suited to identifying specific phenotypes within intact biofilm samples. A good correlation does appear to exist between Nitrosomonas and Nitrosococcus populations quantified by FISH and observed ammonia removal. However, other species of ammonia-oxidizers may also be present. Wagner et al. (1997) reported the dominant ammonia-oxidizer in the Kraftisried wastewater plant to be Nitrosococcus mobilis. In future work, additional probes will be utilized to identify other species of ammonia-oxidizers present.
As nitrification rates decrease through the filter, there is a corresponding decrease in signal intensity. It is important for process control to understand whether the decrease in signal intensity is due to an inhibition of bacteria that remain in place or if the numbers of bacteria present are actually changing. An inhibited system will recover more quickly from an upset than a system requires repopulating. The addition of a universal bacterial probe for future sampling events may provide some insight.
Please note that non-stringent hybridization conditions were utilized and may have resulted in some binding of the probe to organisms other than the targets. This initial work is intended to demonstrate the feasibility of using fluorescent in situ hybridization in a full-scale system. Additionally, the eventual goal is to develop a monitoring tool to be used by wastewater treatment plants with relative ease. For these reasons, it was decided that high stringency hybridizations were not necessary.
Nitrobacter populations could not be quantified with the protocol used. However, the near stoichiometric production of nitrate in the NTF indicates high nitrite oxidizing activity. Changing to a brighter dye that is less prone to fading may increase signal intensity and make quantification of Nitrobacter populations feasible. The presence of additional species of nitrite-oxidizers will be also investigated. Probes for Nitrospina, Nitrococcus, and Nitrospira are currently under development (Wagner et al. 1996) and will be applied as the sequences become available.
Signal intensities per segment agreed well with what was predicted given bulk chemical measurements for CBOD5, TSS, and alkalinity. Increased CBOD5 and TSS loads to the NTF depressed nitrification rates in the top segment of the filter. There was a corresponding decrease in signal intensity indicating that the nitrifying bacterial population had been severely inhibited or washed out of that filter segment. Low signal intensities in the bottom half of the NTF equated with low ammonia and alkalinity concentrations.
FISH allows for rapid assessments of the microbial population’s diversity and activity. Community structure may be important to how well an NTF recovers from inhibitory events such as high CBOD5 and TSS pulses. As a monitoring tool, FISH can provide supplemental information to bulk chemical measurements. Information obtained will be beneficial to process modeling, design, and control/operation of NTFs.
This work was supported by a grant from the National Science Foundation (BES-9753086) and the Littleton/Englewood Wastewater Treatment Plant. We thank Mark Hernandez at the University of Colorado, Boulder for providing the fluorescent microscope and image analysis system and Jodi Flax at Northwestern University, Illinois for her technical assistance.
Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R., and Stahl, D. (1990a).
Combination of 16S rRNA-Targeted Oligonucleotide Probes with Flow Cytometry for Analyzing Mixed Microbial Populations. Applied and Environmental Microbiology 56:1919-1925.
Amann, R.I., Stromley, R., Devereux, R, and Stahl, D. (1992). Molecular and Microscopic Identification
of Sulfate Reducing Bacteria in Multi-species Biofilms. Applied and Environmental Microbiology 58:614-623.
Amann, R.I., Krumholz, L., and Stahl, D. (1990). Fluorescent-oligonucleotide Probing of Whole Cells for
Determinative, Phylogenetic, and Environmental Studies in Microbiology. Journal of Bacteriology 172:762-770.
Belser, L. W. (1979). Population Ecology of Nitrifying Bacteria. Ann. Rev. Microbiol. 33:309-333Bergey’s Manual of Systematic Bacteriology: Volume 3 (1989) Edited by James T. Staley and John G.
Holt, Williams & Wilkins, Baltimore.
Brock, T. D. (1987). The Study of Microorganisms In Situ; Progress and Problems. Symp. Soc. Gen.
De los Reyes, M.,. De los Reyes F.L, Hernandez M., and Raskin, L. (1998). Identification and
Quantification of Gordona amarae Strains in Activated Sludge Systems Using Comparative rRNA Sequence Analysis and Phylogenic Hybridization Probes. Water Science and Technology 37:521-528
Gujer, W., and Boller, M. (1984) Operating Experience with Plastic Media Tertiary Trickling Filters for
Nitrification. Wat. Sci. Tech. 16:201-213
Groeneweg, J., Sellner, B., Tappe, W. (1994). Ammonia Oxidation in Nitrosomonas at NH3
Concentrations Near Km: Effects of pH and Temperature. Wat. Res. 28(12):2561-2566
Manz, W., Szewzyk, U., Ericsson, P., Amann, R., Schleifer, K-H., and Stenstrom, T-A. (1993). In Situ
Identification of Bacteria in Drinking Water and Adjoining Biofilms by Hybridization With 16S and 23S rRNA-Directed Fluorescent Oligonucleotide Probes. Applied and Environmental Microbiology 59:2293-2298.
Mobarry, B.K., Wagner, M., Urbain, V., Rittmann, B.E., and Stahl, D. (1996). Phylogenetic Probes for
Analyzing Abundance and Spatial Organization of Nitrifying Bacteria. Applied and Environmental Microbiology 62:2156-2162.
Parker, D., Lutz, M., Andersson, B., and Aspegren, H. (1995). Effect of Operating Variables on
Nitrification Rates in Trickling Filters. Wat. Env. Res. 67(7):1111-1118
Poulsen, L., Ballard, G., Stahl, D. (1993). Use of rRNA Fluorescence In Situ Hybridization for Measuring
the Activity of Single Cells in Young and Established Biofilms. Applied and Environmental Microbiology 62:1354-1360.
Raskin, L., Amann, R., Poulsen, L., Rittman, B., and Stahl, D. (1995). Use of Ribosomal RNA-Based
Molecular Probes for Characterization of Complex Microbial Communities in Anaerobic Biofilms. Water Science Technology 31:261-272.
Schramm, A., Larson, L.H., Revsbech, N.P., Ramsing, N.B., Amann, R., and Schleifer, K-H. (1996).
Structure and Function of a Nitrifying Biofilm as Determined by In Situ Hybridization and the Use of Microelectrodes. Applied and Environmental Microbiology 62:493-397.
Seigrist, H. and Gujer, W. (1987) Demonstration of Mass Transfer and pH Effects in a Nitrifying Biofilm.
Water Res. 21(12):1481-1487
Standard Methods for the Examination of Water and Wastewater (1995). 19th edn, American Public
Health Association/American Water Works Association/Water Environment Federation, Washington DC, USA.
Szwerinski, H., Arvin, E., and Herremoes, P. (1986) pH-decrease in Nitrifying Biofilms. Wat. Res.20(8):971-976
United States Environmental Protection Agency (1993). Manual: Nitrogen Control, Manual
EPA/625/R-93/010, Office of Research and Development, Office of Water, Washington DC, USA.
Urbain, V., Mobarry B., de Silva V., Stahl D.A., Rittman B.E., and Manem J. (1997). Integration of
Performance, Molecular Biology, and Modeling to Describe the Activated Sludge Process. Proceedings of 2nd International Conference on Microorganisms in Activated Sludge and Biofilm Processes.
Wagner, M., Rath, G., Amann, R., Koops, H-P, and Schleifer, K-H. (1995). In Situ Identification of
Ammonia-oxidizing Bacteria. System. Appl. Microbiology 18:251-264.
Wagner, M., Rath G., Koops H-P., Flood J., and Amann R.. (1996). In Situ Analysis of Nitrifying
Bacteria in Sewage Treatment Plants. Water Sci. Tech. 34:237-244.
Wagner, M., Noguera, D., Juretschko, S., Rath, G., Koops, H-P., and Schleifer, K-H. (1997). Combining
Fluorescent In Situ Hybridization (FISH) with Cultivation and Mathematical Modeling to Study Population Stucture and Function of Ammonia-oxidizing Bacteria in Activated Sludge. Proceedings of 2nd International Conference on Microorganisms in Activated Sludge and Biofilm Processes, Berkley, CA.