Microbial Community Structures in Foaming and ...

Microbial Community Structures in Foaming and Nonfoaming Full-Scale Wastewater Treatment Plants Francis L. de los Reyes 111, Dagmar Rothauszky, Lutgarde Raskin

ABSTRACT: A survey of full-scale activated-sludge plants in Illinois revealed that filamentous foaming is a widespread problem in the state, and that the causes and consequences of foaming control strategies are not fuilly understood. To link microbial community structure to foam occurrence, microbial populations in eight foaming and nine nonfoaming full-scale activated-sludge systems were quantified using oligonucleotide hybridization probes targeting the ribosomal RNA (rRNA) of the mycolata; Gordonia spp., Gordonia amarae; "Candidatzis Microthrix

parvicella": the et-, f-, and y-subclasses of the Proteobacteria, and members of the Cytoplhaga-Flavobacteria. Parallel measurements of microbial population abundance using hybridization of extracted RNA and fluorescence in situ hybridization (FISH) showed that the levels of mycolata, particularly Gordonia spp., were higher in most foaming systems compared with nonfoaming systems. Fluorescence in situ hybridization and microscopy suggested the involvement of "Candidatus Microthrix parvicella" and Skernania pinifornis in foam formation in other plants. Finally, high numbers of "Candidatus Microthrix parvicella' were detected by FISH in foam and mixed liquor samples of one plant, whereas the corresponding levels of rRNA were low. This finding implies that inactive "Candidants Microthrix parvicella" cells (i.e., cells with low rRNA levels) can cause foaming. Water Environ. Res., 74, 437 (2002).

KEYWORDS: activated sludge, filamentous foaming, microbial community structure, oligonucleotide probes.

Introduction Understanding the link between microbial community structure and function in activated-sludge systems is crucial to improving system operation. For example, early studies on identifying microbial populations in activated sludge using microscopy and staining techniques were instrumental in describing filament "types" and correlating their occurrences to operating conditions (Eikelboom, 1975. 1977; Farquhar and Boyle, 1971: Jenkins et al., 1993). Other researchers used cultivation-based techniques to study microbial populations in activated sludge (Benedict and Carlson, 1971; Dias and Bhat, 1964). However, morphology- and staining-based identification methods have drawbacks that limit their use in structure-function studies. Bacteria have limited morphological diversity and misidentification of related species is common (Amann et al., 1998; Kampfer. 1997; Soddell and Seviour, 1990). Furthermore, the same organisms sometimes exhibit variable staining characteristics or undergo morphological changes (Kampfer, 1997; Soddell and Seviour, 1990). Cultivationbased studies have also been criticized for their failure to determine true community diversity and their inability to identify the important (in terms of abundance and activity) populations in September/October 2002

complex ecosystems (Amann et al., 1992; Brock, 1987). In contrast, molecular techniques such as approaches based on comparative sequence analysis of libosomal RNA (rRNA) allow community characterization without these limitations (Amann et al., 1995; Olsen and Woese, 1993; Raskin et al., 1997; Stahl and Amann, 1991). Numerous studies have used oligonucleotide hybridization probes targeting the rRNA of defined phylogenetic groups to identify and quantify specific microbial populations in activated sludge (Amann et al., 1998; Kampfer et al., 1996). However, despite the abundance of work in this area, relating activatedsludge microbial community structure to function or operating performance remains a significant challenge. Linking microbial population dynamics to activated-sludge op,eration is especially relevant when investigating specific operational problems such as filamentous foaming. The formation of stable, brown foam on the surfaces of aeration tanks, secondary clarifiers, and aerobic and anaerobic digesters has long been attributed to filamentous organisms, particularly Gordonia (formerly Nocardia) amarae. However, other members of the mycolic acidcontaining actinomycetes (mycolata) and another actinomycete, "CandidatusMicrothrix parvicella" (hereinafter referred to as M. parvicella), have been implicated in foam formation (Sodell and Seviour, 1990). Therefore. oligonucleotide hybridization probes have been designed to identify and quantify the mycolata, Gordonia, G. attarae (de los Reyes et al., 1997, 1998a), and M. parvicella (Erhart et al., 1997). de los Reyes and Raskin (2002) recently linked the levels of mycolata to the initiation and stability of activated-sludge foam using probes for the mycolata, Gordonia, and G. amarae. The levels of M. parvicella and other microbial populations were not quantified in this study. While the link between the presence of G. amnarae above a certain threshold and foaming has been established (Davenport et al., 2000; de los Reyes and Raskin, 2002), it is possible that other microbial groups also play a role in foam formation or stabilization. The present study was conducted to characterize microbial communities in full-scale activated-sludge systems to gain insight to the possible roles of other populations in foaming. The extent and severity of foaming problems in Illinois were evaluated by conducting a survey of activated-sludge plants using probes specific for the primary microbial groups in activated sludge (Amann et al., 1998; Manz et al., 1992) and M. parvicella to compare the community structure of foaming and nonfoaming systems. Because some of these probes had previously been characterized for fluorescence in situ hybridization (FISH) only, additional probe characterization studies were required. Furthermore, 437

de los Reyes et al. the probes were used to evaluate changes in community structure before and during a foaming episode in a plant with seasonal foaming problems.

Materials and Methods Survey of Activated-Sludge Plants. To assess the prevalence of filamentous foaming in Illinois, a survey of full-scale municipal wastewater treatment plants was conducted. Contact information was obtained from the Illinois Association of Wastewater Agencies (2001) and the U.S. Environmental Protection Agency Envirofacts Warehouse Database (2001). Surveys with questions about the occurrence, control, and possible causes of filamentous foaming were sent to 114 plants in Illinois. The survey asked the operators to assess the frequency of foam occurrence ("never" = 1 to 2; "occasionally" = 3 to 4; "frequently" = 5 to 6; and "continually" = 7 to 8) and to rate the effectiveness of control strategies ("very successful" = 1 to 2; "moderately successful" = 3 to 4; "not successful" = 5 to 6). Thirty-one facilities responded to the questionnaire and samples for microbial characterization were requested from 15 of these plants. Additional samples were obtained from a municipal plant outside of Iliinois (Olympia, Washington) and an industrial plant (designated "Plant P") with foaming problems. Tables 1 and 2 list the municipal plants from which samples were analyzed with some of the plant characteristics. Samples. Grab samples were taken from the activated-sludge mixed liquor, activated-sludge foam, and the anaerobic or aerobic digester sludge of the plants listed in Tables 1 and 2 and the industrial plant. Samples were collected in 50-mL tubes and sent overnight to the laboratory or prepared on-site as described in the following paragraph. Total and suspended solids concentrations were determined according to Standard Methods (APHA et al., 1992). In addition, 50-mL grab samples of activated-sludge mixed liquor and foam were collected from the Urbana-Champaign Sanitary District Northeast (UCSD-NE) wastewater treatment plant every 2 to 3 days from March to September 1999. For membrane hybridizations, 14-mL, samples were centrifuged at 2040 X g and cell pellets were stored at -80 "C until nucleic acid extraction. For FISH, 3-mL samples were immediately fixed on-site after sampling or in the laboratory after overnight shipment in 9-mL 4% (w/v) paraformaldehyde for 1 minute at room temperature (de los Reyes et al., 1997), and stored in phosphate buffered saline-ethanol (1:1 v/v) at -20 °C until hybridization. Organisms and Culture Conditions. Microthrix parvicella strain RN1 (Rosseti et al., 1997) was provided by Valter Tandoi from the Italian National Research Council (CNR), Water Research Institute (Rome, Italy) on agar plates. Liquid medium R2A (Reasoner and Geldreich, 1985) and R2A agar plates were inoculated with the culture and incubated at 18 'C. Cells were harvested by centrifugation at 2040 X g and stored for FISH after fixation (as described above) and for RNA and DNA extractions by storage at -80 "C. For probe specificity studies, a variety of nontarget organisms were obtained from the American Type Culture Collection (ATCC, Rockville, Maryland) and various culture collections from the University of Illinois, Urbana, and the University of California, Berkeley. Cells were grown as suggested by ATCC to mid-log phase, harvested by centrifugation, and stored for FISH after fixation or stored at -80 "C for RNA extraction. Nucleic Acid Preparation. The RNA was extracted from environmental samples and pure cultures of nontarget organisms using a low-pH, hot-phenol bead-beating method (Raskin et al., 438

1995; Stahl et al., 1988). Because of the low growth of M. parvicella, an adequate amount of biomass for RNA extraction was not obtained. Therefore, M. parvicella 16S rRNA was generated by in vitro transcription (McMahon et al., 1998). Cells of M. parvicella taken from the agar plates were lysed by sonication (FS15 ultrasonicator, Fisher Scientific, Pittsburgh, Pennsylvania) for 20 minutes at room temperature. The 16S ribosomal DNA (rDNA) gene was amplified by the polymerase chain reaction (PCR) using standard primers S-*-Bact-0011-a-S-17 and S-DBact-1492-a-A-21 (Kane et al., 1993). The PCR was performed in a model PTC-200 thermocycler (MJ Research, Inc., Watertown, Massachusetts) using 30 cycles of 1 minute at 92 °C, 1 minute at 55 °C, and 1 minute at 72 °C (final extension for 7 minutes), following an initial denaturing step of 5 minutes at 94 °C (de los Reyes et al., 1998a), and repeated to obtain a higher yield of the amplified sequence. The amplified rDNA was ligated into the PCR 2.1 vector and Escherichiacoli INVoaF' was transformed using the Invitrogen TA Cloning Kit Version B (Invitrogen Corporation, San Diego, California) according to the manufacturer's instructions. Initial screening of transformants was performed by restriction enzyme digestion (Sambrook et al., 1989). To verify the 16S rDNA sequence ofM. parvicella and to select a plasmid with an insert that would produce dense RNA (after transcription), plasmids with different digestion patterns were partially sequenced by the University of Illinois Biotechnology Center (Urbana) using the M13(-20) forward primer (Invitrogen Corp.). Plasmids that matched the 16S rRNA of M. parvicella (obtained from the Ribosomal Database Project (RDP) [Maidak et al., 1999]) were used for transcription. Plasmids were linearized using the restriction enzyme Sacd (5'-GAGCTC-3'), extracted with phenol-chloroform, precipitated with ethanol, and examined using agarose gel electrophoresis to verify that the plasmid was linearized and cut only once. In vitro transcription was performed with the AmpliScribe Transcription Kit (Epicentre Technologies, Madison, Wisconsin) (McMahon et al., 1998). The transcript was purified with RNase-free DNase (Epicentre Technologies), extracted with phenol-chloroform and chloroform, and precipitated with ammonium-acetate and absolute ethanol according to the manufacturer's instructions (Epicentre Technologies). The quality of the transcript was evaluated using polyacrylamide gel electrophoresis. Oligonucleotide Probes and Membrane Hybridization. The oligonucleotide probes used, their target groups, and posthybridization wash temperatures (Tm) are listed in Table 3. Probes synthesized by the University of Illinois Biotechnology Center were 5'-end labeled with [-y3 2-P] adenosine 5'-triphosphate (ATP), and membrane hybridizations were performed as previously described (de los Reyes et al., 1997; Raskin et al., 1994). In brief, denatured RNA samples and dilution series of RNA from appropriate pure cultures were applied in triplicate to Magna Charge nylon membranes (Micron Separations, Inc., Westborough, Massachusetts). These membranes were hybridized with universal and specific probes, and the hybridization responses were measured using an Instant lmager (Packard Instruments, Meriden, Connecticut). The results were expressed as the concentration of target rRNA over the total rRNA (de los Reyes et al., 1997; Raskin et al., 1994). Characterization of Oligonucleotide Probes. To detennine optimal posthybridization wash conditions for membrane hybridization for probe S-*-aProt-0019-a-A-17 (targeting the (xsubclass of the Proteobacteria), probe S-*-Cyf-0319-a-A-18 (targeting the CytoWater Environment Research, Volume 74, Number 5

de los Reyes et al. Table 1-List of wastewater treatment plants that experienced foaming at the time of sampling, design flowrates, activated-sludge configurations, and membrane hybridization results. Proteobacteria Mycolata

Gordonia

G. amarae

City and sample locationa

M. parvicella

Alpha

Beta

Gamma

CFb

Total 16S rRNA in sample (%)

Downer's Grove, Illinois (design flowrate = 3.78 x 103 m3/d, conventional activated-sludge configuration) AS mixed liquor AS foam AS sludge

2.0 4.4 8.5

0.9 1.8 6.2

0.8 0.4 1.5

O5