Identity, abundance and ecophysiology of filamentous Chloroflexi species present in activated sludge treatment plants

Abstract

Filamentous Chloroflexi species are often present in activated sludge wastewater treatment plants in relatively low numbers, although bulking incidences caused by Chloroflexi filaments have been observed. A new species-specific gene probe for FISH was designed and using phylum-, subdivision-, morphotype 1851- and species-specific gene probes, the abundance of Chloroflexi filaments were monitored in samples from 126 industrial wastewater treatment plants from five European countries. Chloroflexi filaments were present in 50% of the samples, although in low quantities. In most treatment plants the filaments could only be identified with phylum or subdivision probes, indicating the presence of great undescribed biodiversity. The ecophysiology of various Chloroflexi filaments was investigated by a suite of in situ methods. The experiments revealed that Chloroflexi constituted a specialized group of filamentous bacteria only active under aerobic conditions consuming primarily carbohydrates. Many exo-enzymes were excreted, e.g. chitinase, glucuronidase and galactosidase, suggesting growth on complex polysaccharides. The surface of Chloroflexi filaments appeared to be hydrophilic compared to other filaments present. These results are generally supported by physiological studies of two new isolates. Based on the results obtained in this study, the potential role of filamentous Chloroflexi species in activated sludge is discussed.

Chloroflexi, filamentous bacteria, activated sludge, ecophysiology

Introduction

Members of the phylum Chloroflexi, formerly known as the green nonsulphur bacteria, have primarily been associated with extreme habitats, e.g. microbial mats in hot springs (Boomer et al., 2002; Nubel et al., 2002) and hypersaline environments (Nubel et al., 2001), where they are known as (filamentous) anoxygenic photothophs (Hanada et al., 2002; Hanada & Pierson, 2002; Nubel et al., 2002). Filamentous members of the phylum Chloroflexi have also been found in activated sludge wastewater treatment plants (WTP), and they have occasionally been associated with bulking incidences (Beer et al., 2002; Bjornsson et al., 2002; Schade et al., 2002).

A common filamentous microorganism, Eikelboom's morphotype 1851, has been micromanipulated, cultured, and sequenced, which has revealed that this morphotype belongs to the phylum Chloroflexi (Beer et al., 2002). A species-specific gene probe for FISH demonstrated its presence in activated sludge systems. Bjornsson et al. (2002) developed a phylum-specific as well as two subdivision-specific gene probes for Chloroflexi based on five clones originating from a sequencing batch reactor as well as publicly available Chloroflexi sequences. Phylogenetic analysis of the sequences revealed four subdivisions in total, but target sites suitable for probes were only found for two (div. 1 and 3). In subdivision 1, mainly clones from environmental sources were located, e.g. from polluted habitats; isolates obtained from activated sludge were also found here (Juretschko et al., 2002; Yamada et al., 2005). In subdivision 3, most characterized isolates clustered, e.g. morphotype 1851 (Beer et al., 2002), Chloroflexi ssp., Oscillochloris ssp., Roseiflexus carstenholzii and Herpetosiphon ssp. The closest relative to morphotype 1851 was Roseiflexus carstenholzii [however, it only shares 84% 16S rRNA gene sequence similarity; Beer et al. (2002)]. Many Chloroflexi members, at least those found in activated sludge, resemble morphotype 1851 as defined by Eikelboom & van Buijsen (1983) and Jenkins et al. (1993). This morphotype is characterized as having a cell diameter of 0.5–0.8 μm, length of filaments >200 μm, with rectangular cell shape and weak Gram-positive and Neisser-negative stain. They possess an atypical Gram-negative cell wall (Beer et al., 2002), and staining results seem to some extent to depend on the dye used in Gram solution A (carbol gentian violet vs. crystal violet) (D.H. Eikelboom, personal communication). They are frequently observed with epiphytic bacteria, especially in municipal WTP, and are often found in bundles.

Little is known about the physiology of filamentous Chloroflexi in activated sludge. Only a few pure culture studies have been conducted; one on the morphotype 1851 identified as Chloroflexi (Kohno et al., 2002) and another on filamentous Chloroflexi isolated from mesophilic and thermophilic methanogenic sludge granules (Yamada et al., 2005). In the first study, five filamentous orange-pigmented strains were isolated, characterized as morphotype 1851, and named Kouleothix aurantiaca. Minor discrepancies were seen between pure culture observations and the characterization manuals based on in situ observations, relating to Gram staining, gliding motility on solid media and filament length. The isolates grew mainly on sugars (e.g. glucose, mannose, trehalose, and xylose) and pyruvate. Two of the five strains were able to reduce nitrate to nitrite, and all strains were able to ferment glucose (Kohno et al., 2002). Three strains were also isolated from thermophilic and mesophilic methanogenic sludge granules, and pure culture investigations showed that they were strict anaerobes and were specialized on carbohydrates (glucose, fructose and sucrose). However, no growth was observed if yeast extract was not added (Yamada et al., 2005). Recently, filamentous Chloroflexi were identified in a nitrifying biofilm by applying the phylum-specific gene probes (CFX1223 and GNSB941) (Kindaichi et al., 2004). Studies of the ecophysiology using FISH-micro-autoradiography (MAR) showed that they took up N-acetylglucosamine and a mixture of amino acids, but they were never observed to take up acetate.

Recent studies on the ecophysiology of filamentous microorganisms in activated sludge wastewater treatment plants (WWTP) suggest that it is impossible to make general statements regarding their physiology (Wagner et al., 2002). Properties such as substrate uptake capability and rates, substrate affinity, storage abilities, surface properties, and exo-enzyme activity depend on the species examined. Studies on Microthrix (Nielsen et al., 2002), Thiothrix (Nielsen et al., 2000), filamentous Alphaproteobacteria (Kragelund et al., 2006), Aquaspirillum-related filaments (Thomsen et al., 2006), Skermania (Eales et al., 2005) and TM7-related filaments (Thomsen et al., 2002) revealed that, generally, two types of physiological strategies are exhibited by these filamentous bacteria. Some filamentous bacteria are versatile in substrate utilization, appearing as general consumers of organic matter and exemplified by the filamentous Alphaproteobacteria (Kragelund et al., 2006). Others are very restricted and thus consumers of only few specific organic compounds such as lipids by Microthix (Andreasen & Nielsen, 2000; Nielsen et al., 2002). Some filamentous species are also able to take up substrates with electron acceptors other than oxygen and have a large storage capacity. Based on this, it appears that many filamentous bacteria possess an unusual physiology and ecology. If detailed knowledge about the ecophysiology of specific filamentous bacteria were combined with the characterization of WWTP process conditions, better and more efficient WWTP control strategies could be developed to prevent sludge bulking.

In this study, the identity, abundance, and ecophysiology of Chloroflexi species in industrial and municipal WWTP were investigated. For this purpose, a new species-specific gene probe was designed and the ecophysiology of filamentous Chloroflexi in industrial and municipal WWTPs was investigated by applying several in situ methods in combination with FISH. Also included are physiological characteristics of two new Chloroflexi isolates, which appear to be closely related to K. aurantiaca.

Materials and methods

Activated sludge

Activated sludge samples used for the survey were fixed directly at the WWTP in 50% ethanol or 3.6% formaldehyde to preserve both Gram-positive and Gram-negative bacteria. In total, 126 samples from different industrial WWTPs and five samples from municipal WWTPs were used to monitor the presence and abundance of filamentous Chloroflexi as well as other filamentous species. Samples from many different industrial plants were included [agro industry (6), brewery (2), chemical (27), dairy (10), fish (3), food (11), potato (12), pulp and paper (22), textile (5), tannery (4), other (13), domestic (4), and unknown (6)]. Samples were collected in Denmark, Italy, Poland, Germany and the Netherlands. Of the 126 different industrial plants, 68 had nitrification and 58 also had denitrification.

Ecophysiology experiments were carried out with activated sludge from industrial and municipal plants from the Netherlands (TNO17 and TNO25), Italy (CNR1) and Denmark (Egå and Skagen). Plant descriptions are presented in Table 1 for all plants except TNO25, for which data were not available. All five WWTPs had nitrification, three had denitrification and two had enhanced biological phosphorus removal. The temperature of the process tank varied, and industrial WTP were operated at a higher temperature than most municipal WTP. A selector [a small compartmentalized tank where raw influent is mixed with return sludge to control filaments, see e.g. Martins et al. (2004)] was present in TNO17.

1

Overview of wastewater treatment plants used for ecophysiological studies

	AAE	CNR1	Skagen	Egå	TNO17
Type of industry	Municipal	Tannery	Fish industry	Municipal	Paper
Nitrification	+	+	+	+	+
Denitrification	+	+	+	+	−
Phosphorus removal (Biological/Chemical)	B+Ch	Ch	B+Ch	B+Ch	Ch
Temperature of process tank (°C)	8–18	18–25	8–18	8–18	20
Sludge age (days)	25	15	25	25	30
Selector present*	−	−	−	−	+

*

A small compartmentalized tank with a short residence time where raw influent is mixed with return sludge.

The activated sludge was collected and sent to Aalborg (Denmark) by express mail overnight. In situ experiments were conducted immediately after arrival of the sample. Undiluted sludge was used for exo-enzyme and surface property experiments. In all other experiments the sludge was diluted to 1 gSS L⁻¹ with nitrate or nitrite-free supernatant from the activated sludge.

Isolation, phylogenetic analysis and probe design

Thee isolates of filamentous bacteria morphologically identified as morphotype 1851 (Strain EU25, Ver9Iso1 and Ver9Iso2) were obtained by micromanipulation from activated sludge samples originating from industrial WWTP treating pulp and paper waste and brewery waste. EU 25 was isolated on a simple medium (MSV+acetate, MSV+A) composed of MSV mineral base (Williams & Unz, 1989) acetate (0.5 g CH₃COO⁻ L⁻¹) as sole carbon source and Eikelboom vitamin solution (1% v/v Eikelboom, 1975). The rich R2A medium (Reasoner & Geldreich, 1985) was used to isolate Ver9Iso1 and Ver9Iso2.

PCR, purification of products and sequencing were performed as described in detail elsewhere (Levantesi et al., 2004). The sequences were edited using Sequencer DNA sequencing software (Gene Codes Inc., Ann Arbor, MI). Checks for chimeric sequences were conducted using the check_chimera program from Ribosomal Database Project (http://rdp.cme.msu.edu) and the program bellerophon (Hugenholtz & Huber, 2003). 16S rRNA gene sequences were compiled and aligned using the automatic nucleic acid aligner in the arb software package (Ludwig et al., 2004), and alignments were refined manually. Aligned sequences were used for calculation of trees by neighbour-joining, distance matrix, parsimony, and maximum likelihood approaches using default settings in the arb software. Oligonucleotide probes were designed using the probe design/match tools in the arb software package. To evaluate the formamide concentration for optimum stringency, the designed probe was analyzed on ethanol-fixed EU25 culture applying hybridization buffer containing 0–60% formamide (5% increments).

Identification and abundance

The filamentous bacteria present in industrial WWTPs were morphologically identified using the Eikelboom classification system and the filament index (FI) (Eikelboom, 2002). FI determines the population size of filamentous bacteria and ranges from 0 (no filaments) to 5 (very many). Furthermore, FISH was applied with 16S rRNA gene-targeted oligonucleotide probes targeting all Bacteria, the Chloroflexi phylum, subdivisions and specific species within Chloroflexi (Table 2). Further details on oligonucleotide probes are available at probeBase (Loy et al., 2003). Oligonucleotides were labelled with 5(6)-carboxyfluorescein-N-hydroxysuccinimide ester (FLUOS) or with the sulphoindocyanine dyes (Cy3 and Cy5) (Thermohybaid Interactive, Ulm, Germany). A confocal laser scanning microscope, CLSM (LSM 510, Carl Zeiss, Oberkochen, Germany) equipped with a UV laser (351 and 364 nm), an Ar ion laser (458 and 488 nm), and two HeNe lasers (543 and 633 nm) were used to record fluorescent signals from the gene probes.

2

Overview of specificity, sequences and hybridization conditions of oligonucleotide probes used

Probe name	Specificity	Probe sequence (5′–3′)	% FA*	Reference
EUB338	Most Bacteria	GCTGCCTCCCGTAGGAGT	0–60	Amann et al. (1990)
EUB338-II	Planctomycetales	GCAGCCACCCGTAGGTGT	0–35	Daims et al. (1999)
EUB338-III	Verrucomicrobiales	GCTGCCACC CGTAGGTGT	0–60	Daims et al. (1999)
CFXMIX(CFX1223+GNSB941)	Phylum Chloroflexi	CCATTGTAGCGTGTGTGTMG+AAACCACACGCTCCGCT	35	Bjornsson et al. (2002), Gich et al. (2001)
CFX109	Chloroflexi subdivision 3	CACGTGTTCCTCAGCCGT	30	Bjornsson et al. (2002)
CFX784	Chloroflexi subdivision 1a+b	ACCGGGGTCTCTAATCCC	35	Bjornsson et al. (2002)
Chl1851	Filamentous bacterium Eikelboom morphotype 1851	AATTCCACGAACCTCTGCCA	35	Beer et al. (2002)
EU25-1238	Kouleothix aurantiaca Isolate EU-25	CTGCGCATTGCCACCGACAT	35	This study

*

Formamide concentration in hybridization buffer (v/v).

The abundance of Chloroflexi species was determined in 126 industrial and five municipal WTP with different process designs. The above-mentioned gene probes were applied to estimate the filament abundance of the Chloroflexi phylum (probe CFX1223), subdivision 1 (probe CFX784) and 3 (probe CFX109) of Chloroflexi, morphotype 1851 (probe Chl1851) as well as the K. aurantiaca-related bacteria targeted by probe EU25-1238. Hybridizations with any of the above-mentioned Chloroflexi probes for the screening were done in combination with CFX_MIX (CFX1223+GNSB941).

In most cases, combinations of MAR-FISH, ELF-FISH, and MAC-FISH (see below) were used for the studies of ecophysiology, and hybridization was always done with species-specific gene probes combined with the EUB_MIX (EUB338+EUB338-II+EUB338-III) targeting all Bacteria. All the above-mentioned methods were examined by CLSM before hybridization with gene probes, except MAR-FISH. For MAR-FISH, hybridization with gene probes was performed before applying the photographic emulsion. Positions of interest were recorded by an automatic stage controller, and digital images of filaments were recorded. Fresh samples used for ELF and MAC were fixed in 4% paraformaldehyde for 1 h before FISH was conducted as described by Amann (1995) and the stage control enabled relocation of the microscopic field. Slight modification of MAC-FISH and ELF-FISH were necessary, for details see Kragelund et al. (2005).

Physiological characterization of isolates EU25 and Ver9Iso2

The aerobic growth of EU25 and Ver9Iso2 was analyzed on a range of carbon and nitrogen sources at 20°C. Ver9Iso1 was not further characterized. Tests were performed in triplicate, and positive growth was determined against a negative control without any carbon or nitrogen source. The inoculum for these experiments was grown in liquid MSV+A medium, and cultures were inoculated at 1% (v/v). MSV mineral base with Eikelboom vitamin solution 0.1% (v/v) was used for substrate utilization tests; carbon source concentration was 0.5 g L⁻¹ except alcohols and Tween 80 added at 1% (v/v). Acetate consumption during EU25 growth was ascertained by gas chromatography (Perkin Elmer 8500, stationary phase Carbopack B-DA 80-120 4% CW 20 M) measuring substrate concentration before inoculum and after 10 days when growth was clearly visible.

Growth of isolates on different nitrogen sources (ammonium, nitrate, nitrite, urea) was assessed by adding these compounds (at 0.011% w/v of nitrogen) to basal medium devoid of other nitrogen-containing compounds. For these tests, acetate was used as substrate (0.5 g L⁻¹), and Eikelboom vitamin solution 0.1% (v/v) was added to the media. Autotrophic growth was tested in liquid MSV mineral base with Eikelboom vitamin solution 0.1% (v/v) without any added organic carbon source and with the addition of NaHCO₃ solution (0.42 g L⁻¹) and NaS₂O₃·5H₂O (0.4 g L⁻¹). Denitrification tests were carried out in tubes containing liquid medium. R2A liquid medium was used for Ver9Iso2 and MSV+A medium for EU25 with the addition of 0.1% KNO₃. The vials contained a Durham tube that allowed visualization of any gas produced during incubation. Nitrite production was determined colorimetrically. The ability to grow under anaerobic conditions was analyzed using several different experimental strategies to obtain anaerobiosis; anaerobiosis in a serum bottle with N₂, anaerobiosis in oxoid anaerobic bags (with oxygen indicator), and anaerobiosis in an anaerobic chamber. The temperature growth range of strains EU25 and Ver9Iso2 was determined in liquid R2A medium incubated at 10, 15, 20, 25, 35 and 40°C.

Nucleotide accession numbers

Partial length 16S rRNA gene sequences (1127 bp) of EU25, Ver9Iso1 and Ver9Iso2 (1330 bp and 1252 bp, respectively) were deposited in Genbank under the accession numbers DQ232757, DQ812549 and DQ812550, respectively.

MAR and MAR-FISH

The micro-autoradiography experiments were performed using ³H-labelled and ¹⁴C-labelled organic compounds and ¹⁴C-labelled bicarbonate. Details of the procedure, which includes incubation, fixation, and hybridization with gene probes, addition of a radiosensitive emulsion, exposure, processing, and microscopical investigations, are given elsewhere (Andreasen & Nielsen, 1997; Lee et al., 1999; Nielsen et al., 2000). In brief, two series of studies were conducted. In the first series, various potential substrates were tested for uptake under aerobic conditions to see whether gene probe-defined representatives of Chloroflexi (probe-positive phylum filaments (CFX1223), probe-positive subdivision 3 filaments (CFX109), probe-positive Type 1851 filaments (Chl1851) or probe-positive K. aurantiaca-like filaments (EU25-1238) were specialized or general consumers of organic substrates. For this, a selection of substrates was chosen representing short and long chain fatty acids, sugars, alcohols, and amino acids. In the second series, potential use of electron acceptors other than O₂ was tested by studying uptake of the same positively tested organic substrates with nitrate or nitrite present as electron acceptor or under anaerobic conditions (no oxygen, nitrate, or nitrite present). In all experiments, 2 mL diluted activated sludge (1 gSS L⁻¹) was transferred to glass serum vials with a final substrate concentration of 2 mM, and the labelled fraction was 10 μCi per vial. Incubation time was 3 h (5 h for bicarbonate). In all anaerobic incubations with nitrate or nitrite as e-acceptor and strict anaerobic experiments, a preincubation step of 2 h was included with unlabelled organic substrate (2 mM, oleic acid 0.5 mM). In this way, only bacteria able to take up large amounts of the substrate under these conditions (for storage or growth) would be MAR-positive (Andreasen & Nielsen, 2000). After the preincubation period, labelled (10 μCi per vial) and unlabelled substrate was added to a final substrate concentration of c. 2 mM (oleic acid 0.5 mM). The samples were incubated for 3 h. All vials for anaerobic incubation with either nitrate or nitrite as e-acceptor and strict anaerobic incubations were closed with a gas-tight rubber stopper and flushed with ultrapure nitrogen gas before incubation. When thiosulphate or nitrate was added, a final concentration of 2 mM was used (nitrite only 0.5 mM). A minimum of 30 filaments of each gene probe-defined organism were investigated in each incubation to determine potential uptake. In most experiments, MAR-positive and MAR-negative filaments were assessed by comparing silver grains on top of filaments to the background level. As a control for chemography, sludge was pasteurized at 70°C for 10 min just before incubation under defined conditions. Light microscopy was used to detect silver grains from MAR. [³H]N-acetylglucosamine was purchased from American Radiolabelled Chemicals Inc. (Bio Nuclear AB, Sweden). Details of other radiochemicals used in this study can be found elsewhere (Kragelund et al., 2005).

The isolate EU25 was grown in R2A medium. It was washed thee times in liquid MSV media before MAR incubation to remove residual carbon substrate. MAR experiments were performed as described above. Additional experiments were designed to test aerobic the photoautotrophic or photoheterotrophic ability of EU25. For this purpose, labelled bicarbonate was used in combination with unlabelled substrates (acetate and thiosulphate), and incubations were carried out in both light and darkness. An anaerobic experiment was carried out to determine the potential uptake of labelled bicarbonate together with acetate and thiosulphate in the presence of light.

Enzyme-labelled fluorescence (ELF), ELF-FISH

The presence of exo-enzyme activity was determined using enzyme-labelled fluorescence (ELF^®-97, Molecular Probes, Eugene, OR), where substrates upon enzymatic cleavage form a fluorescent precipitate (excitation 345 nm/emission 530 nm) on the surfaces of bacteria or microcolonies within flocs (Van Ommen & Geesey, 1999; Nielsen et al., 2002). The enzymatic activities of chitinase, esterase, galactosidase, glucuronidase, lipase, and phosphatase activities were investigated. An optimized protocol can be found elsewhere (Kragelund et al., 2005, 2006).

MAC, MAC-FISH

Surface properties were investigated using microsphere adhesion to cells (MAC) where sulphate-modified microspheres with hydrophobic characteristics and a diameter of 0.02 μm were applied (Molecular Probes). Details of microspheres and protocol have previously been described (Nielsen et al., 2001; Kragelund et al., 2005).

Results

Phylogenetic analysis and gene probe design

Isolate EU25 and Ver9Iso2 shared 98% sequence similarity. EU25 clustered together with the Eikelboom morphotype 1851 sequence obtained by Beer et al. (2002) in subdivision 3, along with most of the other isolates of Chloroflexi sp. The isolate EU25 was closely related (99.1–99.9% 16S rRNA gene sequence similarity) to the published sequences for K. aurantiaca (GenBank AB079637-41). The sequence similarity of isolate EU 25 and Eikelboom Type 1851 (Beer et al., 2002) was 94.7%. Ver9Iso 1 and Ver9Iso2 shared 99% sequence similarity. The K. aurantiaca sequences and Ver9Iso2 shared 98%. Ver9Iso2 and Eikelboom Type 1851 (Beer et al., 2002) shared c. 93% sequence similarity.

The phylogenetic tree based on all publicly available 16S rRNA gene sequences of primarily activated sludge clones including isolate EU25 and Ver9Iso1 and 2 belonging to the Chloroflexi phylum is shown in Fig. 1. Subdivisions 1–4 defined by Bjornsson et al. (2002) are marked with digits. All sequences included are targeted by the phylum-specific probes (CFX_MIX), and the subdivision probes are also denoted with a digit [subdivision 1 (CFX784) and subdivision 3 (CFX109), respectively]. All sequences within subdivision 3 were a perfect match to CFX109. Within this subdivision, the probe Chl1851 designed for morphotype 1851 in 2002 by Beer et al. also targeted all K. aurantiaca strains as well as isolates EU25, Ver9Iso1 and Ver9Iso2. One 16S rRNA gene-targeted oligonucleotide probe (EU25-1238) in this project was designed in 2001 to target EU25 and was applied on all industrial samples collected in this study. The probe sequence for EU25-1238 was 5′-CTGCGCATTGCCACCGACAT-3′, and the optimal formamide concentration was determined as 35%. Isolates EU25, Ver9Iso1, Ver9iso2 and K. aurantiaca strains were a perfect match to the probe EU25-1238.

1

16S rRNA gene tree showing the phylogenetic affiliation of strain EU25, isolates Ver9Iso1 and Ver9Iso2 within the Chloroflexi phylum. The tree was calculated by the neighbour-joining method using ARB default settings. The scale bar corresponds to 10% estimated sequence divergence. Perfect-match organisms targeted by gene probes are illustrated with grey boxes. Acinetobacter was chosen as out-group (not shown). Numbers correspond to subdivisions defined by Bjornsson et al. (2002).

Morphology of Chloroflexi filaments

Morphological characteristics such as filament length, width, and cell shape were noted for filaments included in the ecophysiological study, and measurements were carried out on FISH-labelled filaments rather than on fresh samples. This was done to locate the filaments otherwise hidden inside floc material and, thus, cell diameter of the filaments might be slightly biased due to the fixation procedure and the FISH protocol. Only minor differences were observed between the different gene probe-defined Chloroflexi sp. Variable diameters were observed, ranging from 0.5 to 1.0 μm and variable Gram staining results. Moreover, epiphytic bacteria were absent from most of the Chloroflexi filaments in industrial samples, whereas heavy growth of epiphytic bacteria was noted in municipal WTP. All representatives had a rectangular cell shape, were relatively short and contained small PHA granules. Most Chloroflexi filaments targeted by any of the above-mentioned probes in the survey also hybridized with the probe for all Bacteria (EUB_MIX). However, only about half of the CFX109-positive populations in the WWTP used for ecophysiological studies showed a positive EUB_MIX signal, indicating that the EUB_MIX is unable to hybridize with all Chloroflexi species.

Abundance

The filament index of the 126 industrial samples screened exceeded 1.5 for 104 samples and, of these, 92% had FI greater than 2.5. Industrial WWTP often contained co-occurring filamentous populations and this was found in c. 75 of the WWTPs, 35% of which contained Chloroflexi with a filament index >1.5. Chloroflexi were identified in 63 of the WWTPs; of these, 52% contained Chloroflexi populations larger than FI>1.5. Only in 16 WWTPs were Chloroflexi present as FI>2.5. In general, no industrial waste particularly favoured the presence of filamentous Chloroflexi; they were detected in all types of industries as well as in all the municipal plants. Filamentous Chloroflexi were found in WWTP with carbon removal, nitrification and denitrification. In 32 of the Chloroflexi- containing WWTP samples it was not possible to identify the filament beyond phylum level (CFX 1223), indicating the presence of many yet unidentified Chloroflexi species. In seven samples probe CFX1223 positive filaments were higher than FI>1.5. Subdivision 3 affiliated Chloroflexi positive by probe CFX109 were responsible for the high filament index (FI>1.5) in four WWTPs, probe positive EU25-1238 targeted filaments in two WWTPs, probe-defined CFX784 filaments and probe Chl1851 positive filaments each in one plant. Approximately 50% of the probe EU25-1238 defined filaments did not hybridize with the probe Chl1851, although the sequence of EU25 should have a perfect match to probe Chl1851. This could indicate an undescribed phylogenetic diversity. The gene probe-defined filaments were often located in bundles within the floc material, rendering a morphological identification difficult. They were rarely found outside flocs, except if filament index exceeded 2, as was the case in most samples used for ecophysiological studies.

Different morphologies were targeted by the Chloroflexi probes in the FISH survey; CFX1223 targeted primarily typical morphotype 1851 as well as some without epiphytic bacteria. Also, some filaments were very thin (0.5 μm), and others with a diameter of 2.0 μm were targeted. A few showed similarities to a thick morphotype 0041, and others looked like morphotype Type 021N. CFX 109 targeted mainly thin filaments without attached growth. The probes Chl1851 and EU25-1238 both primarily targeted morphotype Type 1851, although some differences were seen with respect to epiphytic bacteria and Gram staining results. Apart from filamentous bacteria, single cells were targeted by the phylum-specific probes (CFX1223 and CFX_MIX) as well as the subdivision-specific probes (CFX784 and CFX109).

Pure culture physiology

Isolates Ver9Iso2 and EU25 grew preferentially on carbohydrates (glucose, fructose), organic acid (acetate, pyruvate) and yeast extracts; no alcohols were utilized (Table 3). Minor differences were observed between the two isolates, in particular in propionate usage. However, these differences might account for the longer incubation time of strain Ver9Iso2, where it was observed that a larger selection of substrates could be utilized by increasing the incubation time from 40 to 90 days. Uptake of acetate was validated for isolate EU25 by gas chromatography. Here, acetate concentration decreased to 24% of initial concentration after c. 10 days of incubation. None of the isolates was able to grow chemo-autotrophically. Neither isolate EU25 nor Ver9Iso2 was able to denitrify. The capability to grow anaerobically could not be verified due to replica failure. However, only scarce growth was observed whenever positive anaerobic growth replicas were obtained. The temperature growth range was identical for both isolates, where 15°C was the minimum temperature sustaining growth, and 35°C the maximum temperature.

3

Physiology of isolates EU25 and Ver9Iso2 in pure culture and Kouleothix aurantiaca data from Kohno et al. (2000)

Metabolism	EU25	Ver9Iso2	Kouleothix aurantiaca
Chemoautotrophic	−	−	ND
Denitrification	−	−	−
Reduction of NO3−-NO2−	−	−	+ (two strains)
Anaerobic growth	+−*	+−*	+ (glucose and fructose)
Substrate utilized for growth	+	+	+
Glucose	+	+	ND
Fructose	+	+	+
Lactose	ND	+	ND
Galactose	+	+	ND
Acetate	+	+	+
Pyruvate	−	+*	ND
Propionate	ND	+	ND
Lactate	−	ND	ND
Oxalacetate	−	−	ND
Citrate	+	+	ND
Yeast extract	ND	+	ND
Casaminoacids	−	−	ND
Ethanol	−	−	ND
Propanol	−	ND	ND
Butanol	−	ND	ND
Tween 80
Temperature growth range	15–35°C	15–35°C	25–30°C†	range

*

Growth after 90 days of incubation.

†

Optimum growth temperature. ND, not determined; +−, failure between replica; +, growth; −, no growth.

The pure culture EU25 was tested for uptake of several substrates (Table 4) using MAR, and only glucose and mannose were taken up; uptake of fatty acids, amino acids or ethanol was not observed. Minute traces of glucose were taken up under conditions where nitrate served as e-acceptor. No aerobic photoautotrophic or photochemotrophic behaviour was observed for EU25. Anaerobic incubation with labelled bicarbonate together with acetate and thiosulphate did not result in a positive MAR signal (data not shown).

4

Uptake of substrates by the different filamentous Chloroflexi species under aerobic conditions as investigated by MAR

	CFX1223		CFX109		Chl1851		EU25-1238
Probe WTP	CNR1	AAE	Egå	Skagen	TNO17	Skagen	TNO25	Pure culture EU25
Formate	−	−	−	−	−	−	−	−
Acetate	−	+	−	−	−	−	−	ND
Propionate	−	ND	−	−	ND	−	−	−
Butyrate	−	++	−	++	ND	−	+	−
Pyruvate	−	(+)	−	−	++	−	−	−
Oleic acid	−	−	−	−	−	−	−	−
Glucose	++	(+)	+/++	+/++*	++	++	+	++
Mannose	++	ND	−	−	+	−	−	+
Galactose	+	ND	−	−	−	−	−	−
Leucine	−	ND	−	−	−	(+)	−	−
Glycine	−	−	−	−	ND	−	−	−
Thymidine	ND	++	ND	ND	ND	ND	ND	ND
Ethanol	−	ND	−	−	ND	−	−	−
N-acetyl-glucosamine	ND	++	−	−	ND	++	ND	ND
Bicarbonate+thiosulphate	−	ND	−	−	ND	−	−	−

*

Not all filaments are positive.

ND, not determined; −, No silver grains (no substrate uptake); (+), some filaments slightly positive, some silver grains; +, few silver grains, but clearly positive; ++, positive, many silver grains.

Ecophysiology

Substrate assimilation profile

A number of WWTPs were selected for detailed studies of the ecophysiology of various probe-defined Chloroflexi species. Type 1851 (positive with Chl1851) was present in one treatment plant. A closely related species positive with the probe EU25-1238, but not probe Chl1851, was present in another two plants. Some filaments, not positive with the specific probes, but positive with the broader probe CFX109, were also studied as well as some filaments only positive with the phylum probe CFX1223. Uptake of various carbon substrates by the probe-defined Chloroflexi under aerobic in situ conditions is shown in Table 4. All substrates tested were taken up by some floc-forming bacteria and single cells during all incubations, serving as positive controls.

Type 1851 (positive with Chl1851 or EU25-1238) mainly consumed glucose and N-acetylglucosamine among the substrates tested. They did not consume a range of other substrates such as acetate, ethanol and amino acids. However, some filaments also took up butyrate and pyruvate; this was largely in agreement with the pure culture studies. Other Chloroflexi species positive with the broader probes (subdivision 3-targeted filaments and phylum-specific filaments) also all consumed glucose and, to a varying degree, the other substrates. Some filaments only positive with the phylum probe (CFX1223) showed a slightly broader uptake spectrum, which reflects the existence of one or more new Chloroflexi species in the plant. Uptake of substrates under denitrifying or anaerobic conditions was never observed for any Chloroflexi (data not shown).

Surface properties and exo-enzymatic activity

The distribution of hydrophobic and hydrophilic surfaces was investigated in the different sludges by MAC (Table 5). Sludge flocs from all industrial WWTPs showed parts of the flocs covered with microspheres and other parts without microspheres, thus acting as control. All gene probe-defined filaments tested hydrophilic, as no hydrophobic microspheres attached to their surface. Although probe EU25-1238-defined filaments from TNO25 appeared slightly less hydrophilic with few microspheres attached, other filamentous species present in the sludge were more hydrophobic. The pure culture EU25 was also characterized as hydrophilic.

5

Surface properties and exo-enzymatic activity of the Chloroflexi species determined by MAC and ELF, respectively

WTP	Gene probe defined filament	Surface properties (MAC)	Exo-enzyme activity (ELF)
CNR1	CFX1223	−	CHIT/GLU
AAE	CFX1223	ND	CHIT/GLU
TNO17	Chl1851	−	EST/GLU
TNO25	EU25-1238	(−)	EST/GAL
Pure culture EU25	EU25-1238	−	EST/GAL

ND, not determined; −,no hydrophobic beads attached; (−), few hydrophobic beads attached; CHIT, chitinase; EST, esterase; GAL, galactosidase; GLU, glucuronidase.

The presence of exo-enzyme activity in the sludge flocs and on the filament surfaces was determined by enzyme-labelled fluorescence assays (Table 5). Sludge flocs from all plants exhibited exo-enzyme activity for all enzymes tested, although some enzyme activity was low and some very high, e.g. lipase and esterase, respectively. Esterase activity was observed for all Type 1851 (Chl1851 or EU25-1238-positive filaments) and for the pure culture. Glucuronidase and galactosidase activity was also found for some Type 1851. The filaments that were positive only with the phylum probe CFX1223 exhibited chitinase and glucuronidase activity.

Discussion

This study presents a comprehensive investigation of filamentous Chloroflexi species present in both municipal and industrial WWTP. At present, very little is known about the physiology of phylum representatives of Chloroflexi present in wastewater systems, where only in situ data from an autotrophic nitrifying biofilm (Kindaichi et al., 2004) and two pure culture studies exist (Kohno et al., 2002; Yamada et al., 2005). This paper reports the first investigation from WWTP where information on identity, abundance and ecophysiology is combined.

Identity and abundance

Phylogenetic analysis of activated sludge clones belonging to the Chloroflexi phylum showed that these are found in subdivisions 1 and 3, as defined by Bjornsson et al. (2002). The isolate EU 25 and different K. aurantiaca strains sharing between 99.1 and 99.9% 16S rRNA gene similarity are located in subdivision 3. The isolated morphotype 1851 identified by Beer et al. (2002) and isolates EU 25 and Ver9Iso2 from this study most likely belong to two different species with 94.7% and 93% 16S rRNA gene similarity, but this should be confirmed by DNA : DNA hybridization.

Filamentous Chloroflexi were present in half of the 126 industrial WTP investigated in this comprehensive study and thus verify the observations by Beer et al. (2002) and Bjornsson et al. (2002) that Chloroflexi is a normal member of the activated sludge microbial community. In 33 WWTP samples, Chloroflexi filaments were present as FI>1.5, indicating potential bulking. In c. 12% of the WWTP samples, high amounts of Chloroflexi filaments were found (FI>2.5), identifying them as important filamentous bacteria involved in bulking incidences. We found that a large fraction of the samples had unidentified Chloroflexi species that only hybridized with the phylum probe, not with subdivision- or species-specific probes. The new probe designed in this study hybridized with almost the same sequences as probe Chl1851 (Beer et al., 2002). It was designed before Beer's probe was published and was therefore used in the survey. However, when Chl1851 subsequently was used on the same sludge samples, it appeared that there was not a complete overlap of the two probes on Chloroflexi filaments, indicating that the diversity is still poorly described in activated sludge. Probe Chl1851-positive filaments were identified in 22% of the Chloroflexi positive population. The probe EU25-1238 targeted an additional 16% and when the two probes were used in combination, 38% of the Chloroflexi could be identified. For this reason, the two probes are recommended to be applied together for identification of filamentous Chloroflexi in activated sludge communities. The Chloroflexi species targeted by probe EU25-1238 or probe Chl1851 was, however, abundant in only three plants (FI>1.5).

In 34 other WWTPs, filaments with the same morphological appearance, but only hybridizing with the subdivision probe (10 samples) or the phylum probe (26 samples), were responsible for a high filament index. This emphasizes that unidentified species are more common than these isolates in full-scale WWTPs. It also shows that phylum probes (CFX_MIX) should be applied for screening the presence of filamentous Chloroflexi species in WWTP. Almost all probe-defined Chloroflexi filaments also hybridized with EUB_MIX; however, only half of the CFX109-positive filaments detected in the samples used for ecophysiological studies gave a positive EUB_MIX signal, suggesting that some Chloroflexi species did not hybridize with the EUB_MIX probes. This needs to be resolved in future studies.

The morphology of most gene probe-defined Chloroflexi from the ecophysiological studies fell within the broadly defined morphotype 1851. Minor differences were seen, for example, in cell diameter and Gram staining but these characters also differ in the two identification manuals (Eikelboom, 2002; Jenkins et al., 2004). Morphotype 1851 filaments without epiphytic bacteria have been observed frequently in industrial WWTP, and in this study with gene probe-defined Chloroflexi, no epiphytic bacteria were observed in three of four industrial WTP subjected to ecophysiological characterization. Almost all Chloroflexi sp. examined here showed variability in Gram staining, which might reflect an unusual cell wall, as reported in Beer et al. (2002). This has also been observed for a close relative, Chloroflexi aurantiacus, which also stains Gram-negative, although the peptidoglycan composition has Gram-positive characteristics (Garrity & Holt, 2001).

In 26% of the WWTPs investigated, probe-defined Chloroflexi constituted a filament index higher than 1.5, which could have an impact on floc structure and WWTP operation. In only a few of the cases observed here (TNO17 and TNO25) were representatives of the Chloroflexi directly involved in bulking. Morphology-based surveys have also been reported, and morphotype 1851 only dominated in a few incidences. However, these results are difficult to compare as the FISH screening in our study showed that filaments with different morphology than that of morphotype 1851 were in some cases targeted by some of the probes applied.

Ecophysiological behaviour

The Chloroflexi sp. in the WWTP samples examined were only active in situ under aerobic conditions, which is different from many other filamentous bacteria in activated sludge (Nielsen et al., 2002, Thomsen et al., 2002, Kragelund et al., 2006). The substrates taken up were mainly sugars. Butyrate and a few short chained fatty acids were also used by some filamentous Chloroflexi, but not all. Butyrate is not often taken up by activated sludge bacteria (Kragelund and Nielsen, unpublished results). Acetate was taken up only by filaments in one WTP targeted by the Chloroflexi phylum probe. Whether some unknown Chloroflexi species can consume this compound, as was shown for the new isolates in the pure culture growth tests, or whether the observation is due to unspecific phylum probes is so far unknown. N-acetylglucosamine uptake for Chloroflexi was observed in half of the WWTPs tested. This monosaccharide is a component in lipopolysaccharides and peptidoglycan, constituting the cell wall of most bacteria (Barker & Stuckey, 1999). It is not a substrate commonly used by filamentous bacteria in activated sludge (Kragelund and Nielsen, unpublished results), but would be available continuously due to cell decay and subsequent release of N-acetylglucosamine units. Uptake of N-acetylglucosamine has also been observed under aerobic conditions for filamentous Chloroflexi present in a biofilm (Kindaichi et al., 2004). The expression of exo-enzymes also supports the degradation of polysaccharides, e.g. chitinase, galactosidase, and glucuronidase activity. All examined Chloroflexi filaments appeared hydrophilic, and they were often observed in large bundles inside sludge flocs and were not always visible using phase contrast microscopy.

Comparing the isolates EU25 and Ver9Iso2 with the strains of Kohno et al. (2002), many identical physiological traits were observed. We tested more substrates supporting the specialization on sugars and some short chain fatty acids. However, reduction of nitrate to nitrite observed by Kohno and coworkers (two of the five strains) was not seen for isolate EU25 or Ver9Iso2. The strains isolated by Kohno and coworkers were able to grow anaerobically on sugars (glucose and fructose), whereas the ability to grow anaerobically for strain EU25 and Ver9Iso2 was not confirmed in our study. A small uptake of glucose using MAR was observed under conditions where nitrate served as e-acceptor for isolate EU25, but growth could not be observed.

Comparing the results from the pure cultures with in situ studies it is clear that the pure culture shows greater versatility in physiology than the gene probe-defined filaments in corresponding activated sludge samples. For instance, fewer substrates could be assimilated, and no uptake under anaerobic conditions was recorded. This phenomenon is in agreement with other studies (Rossetti et al., 2005) showing that isolates tend to show their greater physiological potential in pure culture, whereas under in situ conditions they are more restricted in their physiological activity.

Possible role in sludge

Filamentous members of the Chloroflexi phylum are frequently observed in activated sludge and contribute to the overall filament index number. Most likely, the population sizes of Chloroflexi species in sludge have been underestimated by conventional microscopical identification due to their location inside sludge flocs. Furthermore, many as yet unidentified members are present in activated sludge samples which no gene probes target beyond phylum level. Interestingly, acetate was not used (except possibly in one case) by Chloroflexi filaments under in situ conditions, although it is one of the most common substrates present in activated sludge (Hvitved-Jacobsen et al., 1995). Chloroflexi filaments appear to be specialized in polysaccharide degradation as different monosaccharides were taken up, combined with exo-enzyme activity used for polysaccharide degradation. The location of the investigated filaments inside sludge flocs and preferential use of sugars suggest that they grow on colloids and particles from the incoming wastewater trapped in the surrounding exopolymeric matrix, on exopolymers produced by other microorganisms, and on decaying cells. Other filamentous members in activated sludge, e.g. members of Bacteroidetes (Kindaichi et al., 2004), have also been shown to use the same type of substrate, and competition between these species is therefore likely. The influence of morphotype 1851 on floc structure has been observed, causing bridging and even open structured sludge flocs if they become more abundant. However, this requires a large population size and is not frequently observed. All in all, filamentous members of Chloroflexi are commonly observed in municipal and industrial WTP, but are only occasionally involved in bulking or foaming incidences.

Acknowledgements

This study was funded by the EU program ‘Dynamics and composition of filamentous micro-organism communities in industrial water systems’ (DYNAFILM).

Author notes

Editor: Michael Wagner