Microbial Communities in Activated Sludge Performing Enhanced Biological Phosphorus Removal in a Sequencing Batch Reactor Che Ok Jeon, Dae Sung Lee, Jong Moon Park*

Water Research 37 (2003) 2195–2205

Microbial communities in activated sludge performing enhanced biological phosphorus removal in a sequencing batch reactor Che Ok Jeon, Dae Sung Lee, Jong Moon Park*

Department of Chemical Engineering, School of Environmental Science and Engineering, Pohang University of Science and Technology, Biotechnology Lab, San 31, Hyoja-dong, Nam-gu, Pohang 790-784, Kyoungbuk, South Korea

Received 13 June 2002; received in revised form 25 October 2002; accepted 23 November 2002

Abstract

Microbial communities ofactivated sludge in an anaerobic/aerobic sequencing batch reactor (SBR) supplied with acetate as sole carbon source were analyzed to identify the microorganisms responsible for enhanced biological phosphorus removal. Various analytical methods were used such as electron microscopy, quinone, slot hybridization, and 16S rRNA gene sequencing analyses. Electron photomicrographs showed that coccus-shaped microorganisms of about 1 mm diameter dominated the microbial communities ofthe activated sludge in the SBR, which had been operated for more than 18 months. These microorganisms contained polyphosphate granules and glycogen inclusions, which suggests that they are a type ofphosphorus-accumulating organism. Quinones, slot hybridization, and 16S rRNA sequencing analyses showed that the members ofthe Proteobacteria beta subclass were the most abundant species and were afﬁliated with the Rhodocyclus-like group. Phylogenetic analysis revealed that the two dominating clones ofthe beta subclass were closely related to the Rhodocyclus-like group. It was concluded that the coccus-shaped organisms related to the Rhodocyclus-like group within the Proteobacteria beta subclass were the most dominant species believed responsible for biological phosphorus removal in SBR operation with acetate. r 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Microbial community; Enhanced biological phosphorus removal; Sequencing batch reactor

1. Introduction derived from indirect observations and theoretical considerations. Therefore, the identiﬁcation of the Activated sludge processes with cyclic changes of phylogenetic and taxonomic groups ofbacteria respon- anaerobic and aerobic conditions have been used for sible for phosphorus removal remains as homework for phosphate removal from wastewater. Several biological environmental scientists in order to understand the models have been proposed to explain how this EBPR mechanism and to control the EBPR processes. enhanced biological phosphorus removal (EBPR) is Since Fus and Chen [3] ﬁrst described Acinetobacter achieved [1,2]. These models have been established spp. as a microorganism playing an important role in primarily based on the results ofmixed-cultures in EBPR, most subsequent studies have focused on this activated sludge. Knowledge ofthe biochemical reac- bacterial genus [4–6]. However, the Acinetobacter spp. tions involved in the EBPR process has mostly been did not perform the key biochemical transformations observed in EBPR sludge [7,8]. In fact, the majority of *Corresponding author. Tel.: +82-54-279-2275; fax: +82- microorganisms in activated sludge are non-culturable 54-279-8299. and this caused the emphasis on them [9,10]. In recent E-mail address: [email protected] (J.M. Park). years, new attempts have been made to describe

0043-1354/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0043-1354(02)00587-0 2196 C.O. Jeon et al. / Water Research 37 (2003) 2195–2205 bacterial communities without direct cultivation [11–13]. Corp., USA). The mixed liquor suspended solid (MLSS) Immunofluorescence [14] and quinone-profiling studies and total phosphorus content in the sludge were [12] have indicated that the numbers of Acinetobacter analyzed as described by the American Public Health spp. are comparatively low in EBPR processes. Mean- Association (APHA) [18]. The polyhydroxyalkanoic while, quinone analyses have shown that members ofthe acids (PHA) and glycogen in the sludge and the acetic Proteobacteria beta subclasses and Actinobacteria group acid in the supernatant were analyzed according to the are abundant in the EBPR system [12]. It has been method ofJeon and Park [19]. shown in 16S rRNA clone library studies by the polymerase chain reaction (PCR) approach and rRNA in situ hybridization that members ofthe beta subclass 2.3. Electron microscopy of Proteobacteria are the major population in the EBPR system [8,15]. The results ofrRNA-targeted oligonucleo- Sludge samples for the electron microscopic studies tide probing demonstrated that the Actinobacteria group were collected from the reactor at the end of the aerobic consisted ofa high proportion ofthe microbial popula- stage. The sludge samples were fixed with 3% glutar- tion in the EBPR process [8,16]. However, according to aldehyde and 1% osmium tetroxide. For scanning the 16S rRNA clone library studies, this phylogenetic electron microscopy (SEM), the fixed samples were group was not the major population [11, 13]. dried with a critical point dryer using liquid carbon In this paper, polyphasic analytical approaches dioxide as the transition fluid. The dried samples were incorporating electron microscopy, quinone, slot hybri- sputter-coated with gold under vacuum and then dization, and 16S rRNA gene sequencing analyses were examined with a scanning electron microscope (S- used to analyze the communities ofactivated sludge and 2460N, Hitachi Corp., Japan). For transmission electron to characterize the microorganisms responsible for microscopy (TEM), the fixed samples were embedded in phosphate removal in an anaerobic/aerobic sequencing an epoxy resin and polymerized in an oven for 48 h at batch reactor (SBR) supplied with acetate as a sole 60C [5,20]. The hardened samples were trimmed with a carbon source. glass knife and sectioned to 100 nm with a diamond knife on an Ultramicrotome (MT-7000, RMC-EM Corp., USA). The sectioned samples were stained with 2. Materials and methods 0.2% uranyl acetate and lead citrate and then examined with a transmission electron microscope (H-7000, 2.1. Operation of SBR Hitachi Corp., Japan).

A cylindrical vessel with a 4-L working volume was used for the SBR; it was operated in a fill-and-draw mode 2.4. Quinone analysis with a cycle of8 h. Microbial inoculum was obtained from an activated sludge treatment plant at Pohang Quinones from sludge were extracted three times with University ofScience and Technology. Each cycle a chloroform-methanol mixture (2:1 [v/v]), evaporated consisted of15 min anaerobic fill, 2 h anaerobic react, under vacuum conditions and re-extracted three times 4 h 10 min aerobic react, 60 min settle, 30 min decant with n-hexane-water (1:1 [v/v]) according to the previous stage, and 5 min idle phase. Two liters ofclarified methods [21,22]. The extracts were purified with Sep- supernatant were withdrawn at the end ofthe settling Pak Plus Silica (Waters Corp., USA) [12,22]. Quinone phase. Seven hundred seventy mg ofsodium acetate per components were separated and identified with an liter were used as the sole carbon source. The preparation HPLC (Waters Corp., USA) equipped with a Waters ofthe synthetic wastewater has been described elsewhere 996 PAD (photodiode array detector), a 4.6 i.d. + 3 [17]. The amount of40 mg/L NH 4 -N and 15 mg/L PO4 - 250 mm ODS column (Alltech Corp., USA), and an P were loaded into the SBR. The mean cell residence time IBM PC with the Millenium program (Waters Corp., (sludge age) was controlled at about 10 days by USA) for data analysis [12,23]. Methanol-isopropyl withdrawing sludge from the reactor at the end of the ether (9:2, v/v) was used as an eluent at a flow rate of aerobic phase. During the anaerobic periods, the reactor 1 mL/min. Standard ubiquinones along with their was continuously stirred but kept in an anaerobic respective UV spectra and extinction coefficients were condition by gassing with nitrogen. The temperature used for the identification and measurement of the was controlled at 20C using a water circulation system. quinone homologues [22–24]. Ubiquinones (Qs) and menaquinones (MKs) with n isoprene units in their side 2.2. Chemical analyses chains are designated as Q-n and MK-n, respectively. Partially hydrogenated MKs are designated MK-n (Hx), The soluble phosphate and nitrate in the solution were where x indicates the number ofhydrogen atoms analyzed using a DX-120 ion chromatograph (Dionex saturating the side chain. C.O. Jeon et al. / Water Research 37 (2003) 2195–2205 2197

2.5. RNA extraction and slot hybridization (Techne Ltd, UK) for the rRNA fixation. The DNA oligonucleotide probes (signature probes) were 50-end The total RNA from the sludge was recovered with labeled with [g-32P] ATP using a DNA 50-end labeling modifications by Ultraspec RNA isolation kit (Biotecx system (Promega Corp., USA) to a specific activity of Lab., USA). Sludge samples from the reactor at the end 108–109 cpm/mg for each probe. Hybridization and ofanaerobic and aerobic stages were collected by dehybridization were conducted as described by pre- centrifugation (10,000 g, 3 min), and 10 ml ofUltraspec vious papers [9,26–28]. The total rRNA abundance was solution was added. The samples were resuspended and inferred by the use of a universal hybridization probe sonicated for 2 min (2 s On; 1 s Off; intensity 35) on ice (EUB) complementary to all characterized 16S rRNA of with ultrasonic generator (Hielscher Systems GmbH, eubacteria [25,29]. Hybridization signals were quantified Germany). Chloroform 2.4 mL was added to the with PhosphoImager and ImageQuant software (Mole- samples and the mixtures were vigorously vortexed. cular Dynamics, USA). The samples were placed on ice for 5 min and centrifuged at 15,000 rpm, 4 C for 15 min. Then, the 2.6. Extraction of total genomic DNA crude rRNA extracts were purified by the method ofLin and Stahl [25] and treated with RNase free DNAse. The Modification ofthe DNA extraction methods ofBond rRNA extracts were resuspended in TE buffer. et al. [11] and Lin and Stahl [25] was used to isolate the The rRNA-targeted DNA oligonucleotides were genomic DNA from the sludge. Sludge samples were designed with reference to several previous reports collected by centrifugation (10,000 g, 3 min) from the [8,9,26] and the Ribosomal Database Project [27]. SBR at the end ofthe aerobic stage and resuspended in Sequences, target positions, and bacterial specificities 1 mL of a TE buffer (10 mM Tris-HCl, 1 mM EDTA, ofthe signature probes are listed in Table 1. A slight pH 8.0). Lysozyme was added at a concentration of modification ofthe rRNA hybridization procedure of 2 mg/mL, and the suspended solution was incubated for Lin and Stahl [25] was used to quantify specific rRNA in 1 h at 37C. To the sample solution 0.3 g ofzirconium the sludge from the SBR. The nucleic acid samples beads (0.1 mm diameter), 50 mL of20% sodium dodecyl (about 400 ng) were applied to nylon membrane with a sulfate [wt/v], and 700 mL of TE buffer-equilibrated slot device (Schleicher & Schuell Corp., Germany) under phenol were added. Mechanical disruption was achieved slight vacuum. The membranes were air dried comple- using a Mini-beadbeater (Biospec Corp., USA) for 1 min tely and irradiated with an Ultraviolet crosslinker in conical 2.2-mL screw-cap polypropylene vials. The

Table 1 Oligonuleotide probes, their sequences, target positions and speciﬁcities

Probe Sequence (50-30) Target position Speciﬁcity Ref.

EUB GCTGCCTCCCGTAGGAGT 16S, 338-355 Eubacteria [26] ALF CGTTCGYTCTGAGCCAG 16S, 19-35 a-subclass proteobacteria, Several d- [26] subclass proteobacteria, Most spirochetes BET GCCTTCCCACTTCGTTT 23S, 1027-1043 b-subclass proteobacteria [26] GAM GCCTTCCCACATCGTTT 23S, 1027-1043 g-subclass proteobacteria [26] DEL CGGCGTCGCTGCGTCAGG 16S, 385-402 Most members of d subclass of [26] proteobacteria, few gram(+) bacteria CF TGGTCCGTGTCTCAGTAC 16S, 319-336 Cytophaga-ﬂavobacterium cluster of [26] CFB-phylum (most) CTE TTCCATCCCCCTCTGCCG 16S, 659-676 Rhodocyclus purpureus, Comamonas [26] testosteroni, Brachymonas denitriﬁcans, Leptothrix discophora HGC TATAGTTACCACCGCCGT 23S, 1901-1918 Gram (+) with high DNA G+C [26] Content ACA ATCCTCTCCCATACTCTA 16S, 652-669 Acinetobacter species [8] AER CTACTTTCCCGCTGCCGC 16S, 66-83 All hitherto sequenced Aeromonas spp. [9] Except A. schubertii (1 mismatch), all other available Sequences (16S and 23 S) had at least 2 Mismatches PSE GCTGGCCTAGCCTTC 23S, 1432-1446 Most true Pseudomonas spp. [26]

Y=C or T. 2198 C.O. Jeon et al. / Water Research 37 (2003) 2195–2205 effectiveness of the cell lysis procedure was confirmed by increased over the operation time and complete EBPR microscopic examination of the samples before and after was accomplished after about 180 days (Fig. 1). Profiles the lysis treatment. After mechanical disruption, the ofother important compounds such as acetic acid, PHA nucleic acid extracts were purified by phenol-chloro- (the sum of3-hydroxy butyric acid [3HB], 3-hydroxy form-isoamyl alcohol (100:24:1 [v/v/v]) and chloroform valeric acid [3HV], 3-hydroxy-2-methyl butyric acid [3H extraction. 2MB], and 3-hydroxy-2-methyl valeric acid [3H 2MV]) and glycogen are shown in Fig. 2. It was found that 2.7. Amplification, cloning, and sequencing of 16s rRNA acetate uptake took place concurrently with glycogen genes consumption and PHA production in the cell during the anaerobic stage. During the aerobic stage, PHA Amplification ofthe 16S rDNA was performedas decreased near to zero, but on the other hand, glycogen described in Beffa et al. [30]. PCR on 16S rRNA (B1500 increased, which coincided well with the model reported base-pair product) was performed by amplifying the 16S by Mino et al. [1]. The total phosphorus contents ofthe rDNA genes using the eubacterial primers 27fand sludge at the end ofanaerobic and aerobic phases were 1492r. The PCR reaction was conducted as previously about 4.9% and 7.3%, respectively. At this stage, MLSS described by Beffa et al. [30]. The amplified PCR (about 3200 mg/L) and the concentration profiles of fragments of 16S rDNA were legated into pCR2.1 phosphate and carbon compounds were relatively vector containing kanamycin and ampicillin resistance constant. genes and the constructed plasmids were introduced into E-coli INV-a F’ competent cells (Invitrogen, USA). All sequencing was completed with M13 reverse and forward sequencing primers on an ABI model 377 instrument (Applied Biosystems, USA). Partial 16S rRNA gene sequences ofthe 35 clones fromthe SBR reactor have the accession numbers AF245316– AF245348.

2.8. Phylogenetic analysis

The 35 partial sequences obtained were compared with available 16S rRNA gene sequences from GenBank using the BLAST program to determine their approx- imate phylogenetic afﬁliations. The similarity values between the 16S rRNA gene sequences were calculated using SIMILARITY MATRIX (version 1.1) in the Ribosomal Database Project (RDP) [28]. These sequences were also checked for their chimeric properties Fig. 1. Typical proﬁles ofsoluble orthophosphate-P ( K) and using CHIMERA CHECK from the RDP. The 16S nitrate-N (J) concentrations during anaerobic and aerobic rRNA gene sequences were aligned with representative reactor cycle stages. 16S rRNA gene sequences ofrelated taxa using CLUSTAL W [31]. A phylogenetic tree was constructed using the neighbor-joining method within the PHYLIP package [32]. The stability ofthe relationships was assessed based on a bootstrap analysis of1000 data sets using the programs SEQBOOT, DNADIST, NEIGH- BOR, and CONSENSE ofthe PHYLIP package.

3. Results

3.1. SBR performance

The SBR was continuously operated for more than 18 months with acetate as the sole carbon source. Phosphate release during the anaerobic period and Fig. 2. Typical proﬁles ofacetate-C ( K), glycogen-C (’) and phosphate uptake during the aerobic period gradually PHA-C (J) during anaerobic and aerobic reactor cycle stages. C.O. Jeon et al. / Water Research 37 (2003) 2195–2205 2199

3.2. Electron microscopic analysis of the sludge that double staining using uranyl acetate and lead citrate both highlights glycogen and PHA inclusions and makes The SEM image revealed the presence ofseveral polyphosphate granules appear black [20,33,34].TheTEM morphologically distinct bacteria in the inoculum, image ofthe sludge showed that the dominating coccus- including coccoid, rod, and long ﬁlamentous cells shaped microorganisms contained small black polypho- (Fig. 3a). However, the microbial communities in the sphate granules and large bright white inclusions (possibly SBR fed with acetate were simpliﬁed after extended glycogen) inside the cells (Fig. 4a). The TEM images also operation under the same conditions for more than 18 showed that small coccus-shaped (about 0.3 mmdiameter) months (Fig. 3b). At this time, coccus-shaped micro- and rod-shaped (about 0.3 mmwidthand1mmlength) organisms ofabout 1 mm diameter dominated the microorganisms, which had not been found by SEM microbial sludge in the SBR. The relationship between analysis, were present as a minority. As shown in Fig. 4b, morphological characteristics and phosphate removal these small microorganisms did not contain any polypho- was investigated by using TEM. It is a well-known fact sphate granules or glycogen inclusions.

Fig. 3. Electron scanning photomicrographs ( 10,000) of(a) inoculum, a mixture ofcoccus-, rod-, and ﬁlamentous-shaped bacteria and (b) SBR sludge, dominated by coccus-shaped microorganisms.

Fig. 4. Transmission electron microscopic observations of activated sludge from sequencing batch reactor fed with acetate, (a) Coccus- shaped microorganisms ofabout 1 mm in 15 diameter ( 15,000). (b) Coccus-shaped microorganisms ofabout 0.3 mm in diameter and rod-shaped microorganisms ofabout 0.3 mm in width and 1 mm in length ( 30,000). 2200 C.O. Jeon et al. / Water Research 37 (2003) 2195–2205

3.3. Quinone analysis MK-9 (H4) and MK-9 (H8), whereas in the SBR sludge, MK-7 and MK-8, not the long-chain homologs, made The quinone compositions ofthe sludge fromthe SBR up only about 8% and 3.5% ofthe total quinones, reactor and the inoculated seed were determined by an respectively. All these MKs had two absorption maxima HPLC equipped with PAD after partial purification of at 247 and 266 nm and were assigned as MKs. the quinone extract with Sep-Pak Plus Silica, which is known to enhance the accuracy and reliability ofa 3.4. Slot hybridization quinone analysis [12,22]. The UV spectra ofquinones and standard quinones were used for identification and Organism abundance can be estimated from the verification ofthe peaks. Fig. 5 shows the HPLC elution fractional contribution of its specific rRNA molecules profiles ofthe microbial quinones extracted fromthe to the total ribosomal RNA population [29]. Here, the SBR sludge and the inoculum. While the quinone abundance ofspecific species was expressed as percen- extracts ofthe inoculum had complicated HPLC tage oftheir specific rRNA in the total rRNA. There profiles, those ofthe sludge showed relatively simple were no significant differences in the abundance between elution patterns (Fig. 5). In the ubiquinones ofthe the anaerobic and aerobic stages (Fig. 6). rRNA inoculum, Q-9 and Q-10 appeared as two major peaks belonging to the beta subclass ofthe Proteobacteria and Q-8 was not detected, as shown in Fig 5a. A peak was most dominant in rRNA extracted from the reactor. was detected at a retention time similar to that ofQ-8 as The rRNAs matching CTE and HGC probes were next in Fig. 5a. However, co-injection with Q-8 standard and in abundance in the acetate-fed SBR. However, Acine- scanning by PAD both showed that the peak was not of tobacter ssp., Aeromonas ssp., and Pseudomonas ssp. quinone compounds. A few other peaks did not show constituted less than 10% ofthe total rRNA. typical absorption spectra ofQs or MKs by PAD either, suggesting that they were not quinones. On the other 3.5. 16S rRNA gene sequencing and phylogenetic analysis hand, in the SBR sludge, Q-8 was the predominant type ofquinone (about 70% ofthe total quinones) and Q-10 To obtain more detailed taxonomic information on was the second most common component (about 18% the dominant EBPR bacteria in the SBR operation, ofthe total quinones). However, Q-9 was not detected in cloning libraries ofthe 16S rRNA fromthe EBPR the SBR sludge. The spectral analysis showed that all sludge were constructed by PCR. Partial sequences of these ubiquinone components had an absorption max- the 16S rRNA gene (about 400 nucleotides in length) imum at 275 nm and were therefore ascertained as Qs. were determined and compared for 35 SBR clones. Similarly, the MK patterns in the inoculum were Those clones that did not have >97% similarities with complicated by MK-6, MK-7, MK-8, MK-8 (H4), any other clones were considered as different species.

Fig. 5. HPLC elution proﬁles ofmicrobial quinones frominoculum and EBPR sludge at 275 nm (a) inoculum; (b) EBPR sludge. C.O. Jeon et al. / Water Research 37 (2003) 2195–2205 2201

Fig. 6. Abundance of speciﬁc microbial rRNA inferred by nucleotide probe hybridization relative to total RNA extracted from sludge sampled from SBR fed with sodium acetate. The bar graph represents the average of multiple hybridization experiments.

Table 2 Phylogenetic position and frequency of clones in the SBR clone librarya

Phylum Closest group Similarity (%) No. ofclones

Beta subclass Proteobacteria Leptothrix 98 1 Matsuebacter 94 1 Propionivibrio 97 5 Rhodocyclus 96 10 Alpha subclass Proteobacteria Rhodobacter 97 1 Purple nonsulfur bacteria 95 3 Environmental samples 95 2 Gamma subclass Proteobacteria Ectothiorhodospira 93 2 Methylocaldum 90 2 Rhodanobacter 96 1 Xanthomonas 91 1 Planctomycete group Planctomyces 93 2 87 1 Flexibacter-Cytophaga-Bacteriodes Cytophaga 91 2

a A species was deﬁned as a group ofclones having >97% sequence similarity.

The bacterial ‘‘species’’ compositions ofthe SBR clones sequences was also analyzed using the neighbor joining are presented in Table 2. Most clones from the SBR method. The resulting phylogenetic tree indicated that sludge were afﬁliated with the beta subclass of the the two representative clones ofthe Proteobacteria beta Proteobacteria (17 of35 clones). The clones ofthe beta subclass were distantly related to Propionivibrio dicar- subclass were divided into two major groups, Propioni- boxylicus DSM 5885 and Rhodocyclus tenuis DSM 109 vibrio (5 of35 clones) and Rhodocyclus (10 of35 clones). with about 95% and 96% sequence similarities, respec- Representatives of these two different groups were tively (Fig. 7). The gamma and alpha subclass of further sequenced up to more than 1500 nucleotides in the Proteobacteria were equivalent to six ofthe 35 length and then compared with the sequences available SBR clones, respectively. The gamma subclass clones in GenBank. The phylogenetic afﬁliation of the were divided into Ectothiorhodospira, Methylocaldum, 2202 C.O. Jeon et al. / Water Research 37 (2003) 2195–2205

Fig. 7. Phylogenetic tree based on 16S rRNA gene sequences showing position ofdominating clones ofbeta subclass and some other taxa. Numbers in parentheses indicate the number ofclones having >97% sequence similarity. Scale bar represents 0.01 substitution per nucleotide position. Bootstrap probabilities are indicated at the branch points.

Rhodanobacter, and Xanthomonas groups. However, their biases. In this paper, a polyphasic approach clones representing Acinetobacter ssp., Aeromonas ssp. incorporating a quinone analysis, microscopic observa- and Pseudomonas ssp., which historically have been tions, slot hybridization, and PCR cloning was used to considered to dominate phosphate-removing sludge analyze activated sludge communities and to character- populations, were not present. ize the microorganisms responsible for EBPR in SBR. Electron microscopic analysis showed that the dominating microorganisms, which were coccus-shaped, 4. Discussion contained large bright inclusions and black polyphosphate granules, suggesting that these organisms were Non-culture dependent analytical methods ofmicro- concerned with EBPR during the acetate-fed SBR bial ecology, such as quinone analysis, slot hybridiza- operation (Fig. 3 and 4). The white inclusions were tion, and the PCR cloning approach, have been used to presumed to be glycogen granules because the sludge analyze the microbial communities ofactivated sludge to sample collected from the reactor at the end of the avoid culture limitations [11,12,29]. Molecular methods aerobic stage ofthe SBR cycle forthe TEM analysis had have become popular in wastewater microbiology and low PHA content (Fig. 2). TEM observation ofthe SBR microbial ecology. However, these approaches can sludge also showed that small coccus- and rod-shaped contain biases stemming from DNA extraction, PCR microorganisms, which did not contain any polypho- amplification, gene copy number, the contents of16S sphate granules or glycogen inclusions, were present as a rRNA, and cloning artifacts [11,35]. As a direct chemical minority. It was inferred that they existed inside the analytical method for the determination of in situ microbial floc; they were not observed by SEM analysis. bacterial populations in environmental samples, qui- The quinone profile showed that Q-8 was predomi- none analysis has been considered to have a higher nant among quinones in the SBR sludge. This suggests reproducibility and reliability. However, the information that Q-8-producing microorganisms (e.g., species be- on a microbial community from quinone profiling is also longing to the beta subclass ofthe Proteobacteria or limited because quinone contents ofbacterial cells are some members ofthe gamma subclass) were the most variable among taxa or in response to environmental abundant species, which coincides with previous findings stress. Since the application ofa single community ofother researchers ( Fig. 5) [23]. Slot hybridization and analyzing method could provide false information about 16S rRNA gene sequencing studies also showed that the the community structure, a combinational approach largest groups ofclones belonged to members ofthe using various analytical methods is necessary to offset Proteobacteria beta subclass (Fig. 6 and Table 2). C.O. Jeon et al. / Water Research 37 (2003) 2195–2205 2203

Within the beta subclass, most ofthe clones were the contents of16S rRNA in cells can be considerably assigned to the genus Propionivibrio or Rhodocyclus. different according to the microorganisms being studied According to Hippe et al. [36], the genus Propionivibrio or their particular activities [26]. However, the small is fairly closely related to members of the genus number ofclone sequencing (35 clones) could be a cause Rhodocyclus, and the genus Propionivibrio can be unified ofthese discrepancies. Members ofthe alpha subclass of with the Rhodocyclus-like group. The fact that the the Proteobacteria were one ofthe major populations, as rRNAs matching the CTE probe (the probe comple- revealed by the Q-10 content. Slot hybridization and 16S mentary to a region ofthe 16S rRNA specific for rRNA gene sequence approaches showed similar results. Rhodocyclus purpureus, Comamonas testosteroni, Bra- Although the results ofthe 16S rRNA gene sequen- chymonas denitrificans, and Leptothrix discophora) were cing and bacterial quinone analysis are somewhat relatively abundant in the activated sludge also sup- different due to several inherent biases, the use of ported the view that the Rhodocyclus-like group was the polyphasic approaches helped to solidify the conclusions major population (Fig. 6). Therefore, we were able to that the coccus-shaped organisms ofabout 1 mm conclude that the Rhodocyclus-like group within the beta diameter, which are related to the Rhodocyclus-like subclass of Proteobacteria was important for EBPR in group within the Proteobacteria beta subclass, are the the SBR process and possibly acted as a biological most dominant species and are likely to be responsible phosphorus remover. Other researchers have also for the biological phosphorus removal. Interestingly, arrived at similar conclusions [15,37–39]. Two represen- Acinetobacter spp., Aeromonas ssp., and Pseudomonas tatives ofthe Proteobacteria beta subclass were the most ssp., which have historically been thought to dominate distantly related to Propionivibrio dicarboxylicus DSM phosphate-remover, were only minor phylogenetic 5885 and Rhodocyclus tenuis DSM 109, with about 95% groups for EBPR in the SBR sludge. More definitive and 96% sequence similarity, respectively (Fig. 7). identification remains, as does identification ofphysio- On the other hand, there have been several previous logical characteristics ofthe phosphorus-accumulating reports claiming that members of Actinobacteria ac- organisms, for complete understanding of the EBPR counted for a large proportion of the microbial mechanisms. Only pure cultures ofthe phosphorus- population in the EBPR process, a fact determined by accumulating organisms in the EBPR process could fluorescent in situ hybridization [8,15] and quinone provide a basis for the complete understanding of the analysis [12,40]. However, our quinone analysis pro- characteristics ofphosphorus-accumulating organisms vided different results. The peaks of long isoprene and and their mechanisms ofEBPR. partially saturated chains (MK-8 (H4), MK-9 (H4), MK-9 (H8)), e.g., species associated with the Actino- bacteria, were not present in the MK profiles ofthe SBR References sludge. Only MK-7 and MK-8, e.g., bacteria in the Cytophaga-Flavobacterinm group and Plantomyces [1] Mino T, Arun V, Tsuzuki Y, Matsuo T. Effect of group, were the major MKs in the SBR sludge sample. phosphorus accumulation on acetate metabolism in the Furthermore, clones ofthe Actinobacteria were not also biological phosphorus removal process. In: Ramadori R, detected in the 16S rDNA cloning studies. These results editor. Biological phosphate removal from wastewaters. agreed with those ofother 16S rDNA cloning studies Oxford, UK: Pergamon Press, 1987. p. 27–38. [11,13], which indicates that the Actinobacteria do not [2] Wentzel MC, Lotter. LH, Ekama GA, Loewenthal RE, act as biological phosphorus removers in the SBR Marias GvR. Evaluation ofbiochemical models for biological excess phosphate removal. Water Sci Technol process. 1991;23:567–76. The low Q-9 content in the SBR sludge suggests that [3] Fuhs GW, Chen M. Microbiological basis ofphosphate Acinetobacter and Pseudomonas (fluorescent group) removal in the activated sludge process for the treatment of species, which are representatives ofthe gamma wastewater. Microb Ecol 1975;2:119–38. subclass, constitute a minor population. However, slot [4] Auling C, Pilz F, Busse H-J, Karrasch S, Streichan M, hybridization and 16S rRNA gene sequencing studies Schon G. Analysis ofthe polyphosphate-accumulating showed that the Proteobacteria gamma subclass was one microflora in phosphorus-eliminating, anaerobic–aerobic ofthe dominant populations. Slot hybridization also activated sludge systems by using diaminopropane as a showed that the rRNA contents of Acinetobacter ssp., biomarker for rapid estimation of Acinetobacter ssp. Appl Aeromonas ssp., and Pseudomonas ssp. were present with Environ Microbiol 1991;57:3585–92. [5] Bark K, Sponner A, Kampfer. P, Grund S, Dott W. low portions, however the Q-9 peak ofquinone and the Differences in polyphosphate accumulation and phosphate clone ofthe 16S rRNA genes were not detected ( Fig. 6, adsorption by Acinetobacter isolates from wastewater Table 2). The approaches used in this work can contain producing polyphosphate: AMP phosphotransferase. some biases [11,13,26,35]. Therefore, these discrepancies Water Res 1992;6:1379–88. were presumed to be due to several biases introduced by [6] Beacham AM, Seviour RJ, Lindrea KC, Livingston I. these methods. For example, it has been reported that Genospecies diversity of Acinetobacter isolates obtained 2204 C.O. Jeon et al. / Water Research 37 (2003) 2195–2205

from a biological nutrient removal pilot plant of a activated sludge. Appl Microbiol Biotechnol 1993;38: modified U. C. T. configuration. Water Res 1990;24:23–9. 734–7. [7] Jerkins D, Tandoi V. The applied microbiology of [21] Hu HY, Lim BR, Goto N, Fujie K. Analytical precision enhanced biological phosphate removal; accomplishments and repeatability ofrespiratory quinones forquantitative and needs. Water Res 1991;25:1471–8. study ofmicrobial community structure in environmental [8] Wagner M, Erhart R, Manz W, Amann R, Lemmer H, samples. J Microbial Methods 2001;47(1):17–24. Wedi D, Schleifer K-H. Development of an rRNA- [22] Hu HY, Fujie K, Urano K. Enhanced development ofa targeted oligonucleotide probe specific for the genus novel solid phase extraction method for the analysis of Acinetobacter and its application for in situ monitoring bacterial quinones in activated sludge with a higher in activated sludge. Appl Environ Microbiol 1994; reliability. J Biosci Bioeng 1999;87(3):378–82. 60(3):792–800. [23] Hiraishi A, Ueda Y, Ishihara J, Mori T. Comparative [9] Kampfer. P, Erhart R, Beimfohr C, Bohringer. J, Wagner lipoquinone analysis ofinfluent sewage and activated M, Amann R. Characterization ofbacterial communities sludge by high-performance liquid chromatography and from activated sludge: culture-dependent numerical identi- photodiode array detection. J Gen Appl Microbiol fication verse in situ identification using group- and genus- 1996;42:457–69. specific rRNA-targeted oligonucleotide probes. Microb [24] Kroger A. Determination ofcontents and rodox states of Ecol 1996;32(2):101–21. ubiquinone and menaquinone. Methods Enzymol [10] Wagner M, Amann R, Lemmer H, Schleifer K-H. Probing 1978;53:579–91. activated sludge with oligonucleotide specific for proteo- [25] Lin C, Stahl DA. Taxon-specific probes for the cellulolytic bacteria: inadequacy ofculture-dependent methods for genus Fibrobacter reveal abundant and novel equine- describing microbial community structure. Appl Environ associated populations. Appl Environ Microbiol 1995; Microbiol 1993;59(5):1520–5. 61(4):1348–51. [11] Bond PL, Hugenholtz P, Keller J, Blackall LL. Bacterial [26] Amann R, Ludwig IW, Schleifer K-H. Phylogentic community structures ofphosphate-removing and non- identification and in situ detection ofindividual microbial phosphate removing activated sludges from sequencing cells without cultivation. Microbiol Rev 1995;59(1): batch reactors. Appl Environ Microbiol 1995;61(5): 143–69. 1910–6. [27] Maidak BL, Cole JR, Parker CT, Garrity JGM, Larsen N, [12] Hiraishi A, Ueda Y, Ishihara J. Quinone profiling of Li B, Lilburn TG, McCaughey MJ, Olsen GJ, Overbeek R, bacterial communities in natural and synthetic activated Pramanik S, Schimidt TM, Tiedje JM, Woese CR. A new sewage for enhanced phosphate removal. Appl Environ version ofthe RDP (Ribosomal Database Project). Nucl Microbiol 1998;64(3):992–8. Acids Res 1999;27(1):171–3. [13] Snaidr J, Amann R, Huber I, Ludwig W, Schleifer K–H. [28] de los Reyes FL, Ritter W, Raskin L. Group-specific Phylogenetic analysis and in situ identification ofbacteria small-subunit rRNA hybridization probes to characterize in activated sludge. Appl Environ Microbiol 1997;63(7): filamentous foaming in activated sludge systems. Appl 2884–96. Environ Microbiol 1997;63(3):1107–17. [14] Cloete TE, Steyn PL. A combined membrane filter- [29] Stahl DA, Flesher B, Mansfield HR, Montgomery L. Use immunofluorescent technique for the in situ identification ofphylogenetically based hybridization probes forstudies and enumeration of Acinetobacter in activated sludge. ofruminal microbial ecology. Appl Environ Microbiol Water Res 1988;22:961–9. 1988;54:1079–84. [15] Bond PL, Erhart R, Wagner M, Keller J, Blackall LL. [30] Beffa T, Blanc M, Lyon P-F, Vogt G, Marchiani M, Identification ofsome ofthe major groups ofbacteria in Fischer JL, Aragno M. Isolation ofthermus strains from efficient and nonefficient biological phosphorus removal hot composts (60 to 80C). Appl Environ Microbiol activated sludge systems. Appl Environ Microbiol 1999; 1996;62(5):1723–7. 65(9):4077–84. [31] Thompson JD, Higgins DO, Gibson TJ. CLUSTAL W: [16] Manz, W, Wagner M, Amann R, Schleifer KH. In situ improving the sensitivity ofprogressive multiple sequence characterization ofthe microbial consortia active in two alignment through sequence weighting, position specific wastewater treatment plants. Water Res 1994;28:1715–23. gap penalties and weight metrix choice. Nucl Acids Res [17] Smolders GJF, van der Meij J, van Loosdrecht MCM, 1994;22(22):4673–80. Heijnen JJ. Model ofthe anaerobic metabolism ofthe [32] Saitou N, Nei M. The neighbor-joining method: a new biological phosphorus removal process: stoichiometry and method for reconstructing phylogenetic trees. Mol Biol ph influences. Biotechnol Bioeng 1994;43:461–70. Evol 1987;4(4):406–25. [18] American public health association. Standard methods for [33] Clark JE, Beegen H, Wood HG. Isolation ofintact chain the examination ofwater 15 and wastewater, 19th ed. ofpolyphosphate fromPropionibacterium shermanii Washington, DC: American Public Health Association, grown on glucose or lactate. J Bacteriol 1986;168(3): 1995. 1212–9. [19] Jeon CO, Park JM. Enhanced biological phosphorus [34] Liesack W, Weyland H, Stackebrandt E. Potential risks of removal in a sequencing batch reactor supplied with gene amplification by PCR as determined by 16S rDNA glucose as a sole carbon source. Water Res 2000; analysis ofa mixed culture ofstrict barophilic bacteria. 34(7):3470–80. Microb Ecol 1991;21:191–8. [20] Rees GN, Vasilidis G, May JW, Bayly RC. Production of [35] Hippe H, Hagelstein A, Kramer I, Swiderski J, poly-b-hydroxylate in Acinetobacter spp. isolated from Stackebrandt E. Phylogenetic analysis of Formivibrio C.O. Jeon et al. / Water Research 37 (2003) 2195–2205 2205

citricus, Propionivibrio dicarboxylicus, Anaerobiospirillum [38] Nielsen AT, Liu W-T, Filipe C, Grady L JR, Molin S, thomasii, Succinimonas amylolytica and Succinivibrio dex- Stahl DA. Identification ofa novel group ofbacteria in trinosolvens and proposal ofSuccinivibrionceae fam.nov. sludge from a deteriorated biological phosphorus removal Int J Syst Bacteriol 1999;49:779–82. reactor. Appl Environ Microbiol 1999;65(3):1251–8. [36] Jeon CO, Lee DS, Park JM. Morphological Characteristics [39] Crocetti GR, Hugenholtz P, Bond PL, Schuler A, ofmicrobial sludge performing enhanced biological Keller J, Jenkins D. Blackall LL. Identification of phosphorus removal in a sequencing batch reactor fed polyphosphate-accumulating organisms and design of with glucose as sole carbon source. Water Sci Technol 16S rRNA-directed probes for their detection and 2000;41(12):79–84. quantification. Appl Environ Microbiol 2000;66(3): [37] Hesselmann RPX, Werlen C, Hahn D, van der Meer JR, 1175–82. Zehnder AJB. Enrichment, phylogenetic analysis and [40] Liu WT, Linning K, Nakamura K, Mino T, Matsuo T, detection ofa bacterium that performsenhanced biological Forney LJ. Microbial community changes in biological phosphate removal in activated sludge. Syst. Appl. phosphate-removal systems on altering sludge phosphorus Microbiol. 1999;22(3):454–65. content. Microbiology 2000;146:1099–107.