Journal of Microbiological Methods 78 (2009) 119–126

Contents lists available at ScienceDirect

Journal of Microbiological Methods

j ourna l homepage: www.e lsev ie r.com/ locate / jmicmeth

Optimization of three FISH procedures for in situ detection of anaerobic ammoniumoxidizing bacteria in biological wastewater treatment

Marko Pavlekovic a, Markus C. Schmid b,g, Nadja Schmider-Poignee a, Stefan Spring c, Martin Pilhofer a,Tobias Gaul d, Mark Fiandaca e, Frank E. Löffler f, Mike Jetten g, K.-H. Schleifer a, Natuschka M. Lee a,⁎a Department of Microbiology, Technische Universität München, D-85354 Freising, Germanyb Department of Microbial Ecology, Universität Wien, A-1090 Wien, Austriac DSMZ German Collection of Microorgansims and Cell Cultures, D-38124 Braunschweig, Germanyd Institute for Biophysics, Leibniz Universität Hannover, D-30419, Hannover, Germanye AdvanDx, Inc. Woburn, MA 01801, USAf School of Civil and Environmental Engineering and School of Biology, Georgia Institute of Technology, Atlanta, GA 30332-0512, USAg Department of Microbiology, RU Nijmegen, NL 6525 ED, The Netherlands

⁎ Corresponding author. Tel.: +49 8161 71 5444; fax:E-mail address: nlee@microbial-systems-ecology.de

0167-7012/$ – see front matter © 2009 Elsevier B.V. Aldoi:10.1016/j.mimet.2009.04.003

a b s t r a c t

a r t i c l e i n f o

Article history:Received 30 December 2008Received in revised form 5 April 2009Accepted 10 April 2009Available online 20 April 2009

Keywords:Anammox bacteriaNitrogen transforming bacteriaReductively dechlorinating bacteriaSulphate-reducing bacteriaFISHPNA FISHCARD-FISH

Fluorescence in situ hybridization (FISH) using fluorochrome-labeled DNA oligonucleotide probes has beensuccessfully applied for in situ detection of anaerobic ammonium oxidizing (anammox) bacteria. However,application of the standard FISH protocols to visualize anammox bacteria in biofilms from a laboratory-scalewastewater reactor produced only weak signals. Increased signal intensity was achieved either by modifyingthe standard FISH protocol, using peptide nucleic acid probes (PNA FISH), or applying horse radishperoxidase- (HRP-) labeled probes and subsequent catalyzed reporter deposition (CARD-FISH). Acomparative analysis using anammox biofilm samples and suspended anammox biomass from differentlaboratory wastewater bioreactors revealed that the modified standard FISH protocol and the PNA FISHprobes produced equally strong fluorescence signals on suspended biomass, but only weak signals wereobtained with the biofilm samples. The probe signal intensities in the biofilm samples could be enhanced byenzymatic pre-treatment of fixed cells, and by increasing the hybridization time of the PNA FISH protocol.CARD-FISH always produced up to four-fold stronger fluorescent signals but unspecific fluorescence signals,likely caused by endogenous peroxidases as reported in several previous studies, compromised the results.Interference of the development of fluorescence intensity with endogenous peroxidases was also observed incells of aerobic ammonium oxidizers like Nitrosomonas europea, and sulfate-reducers like Desulfobacterpostgatei. Interestingly, no interference was observed with other peroxidase-positive microorganisms,suggesting that CARD-FISH is not only compromised by the mere presence of peroxidases. Pre-treatment ofcells to inactivate peroxidase with HCl or autoclavation/pasteurization failed to inactive peroxidases, butH2O2 significantly reduced endogenous peroxidase activity. However, for optimal inactivation, different H2O2

concentrations and incubation time may be needed, depending on nature of sample and should thereforealways be individually determined for each study.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Fluorescence in situ hybridization (FISH) with specific oligonucleo-tide probes is a useful tool for cultivation independent in situidentification of target cells in environmental samples. The first FISHprotocol for prokaryotic organisms was based on fluorochrome-labeled oligonucleotide probes (Giovanonni et al., 1988; Amann et al.,1990). While this method proved to be useful for target cellvisualization in habitats such as activated sludge, it was of limitedvalue for detecting target cells in oceans, soil or even specific bacterial

+49 8161 71 5475.(N.M. Lee).

l rights reserved.

species with rigid cell walls such as Mycobacterium (Amann andFuchs, 2008; Lehtola et al., 2006; Wagner et al., 2003). To overcomethese limitations, different strategies have been used, focusing eitheron increasing the probe signal intensity, by for example polynucleo-tide FISH (Zwirglmaier, 2005) and catalyzed reporter deposition FISH(CARD-FISH, Pernthaler et al., 2002; Hoshino et al., 2008), orminimizing probe penetration problems and increasing hybridizationefficiencies by employing different probe chemistries, such as peptidenucleic acid FISH (PNA FISH, Perry-O'Keefe et al., 2001) and lockednucleic acid FISH (LNA FISH, Kubota et al., 2006).

Standard FISH protocols using fluorochrome-labeled oligonucleo-tide probes have been successfully applied for in situ detectionof anammox bacteria, a recently discovered group within the

120 M. Pavlekovic et al. / Journal of Microbiological Methods 78 (2009) 119–126

Planctomycetes that present a novel metabolic group involved in theanaerobic oxidation of ammonium to dinitrogen gas (see e.g., Kuenen,2008). However, it was recently reported that a specific CARD-FISHprotocol had to be employed for enhanced in situ detection of particleassociated anammox cells in oligotrophic environmental samples (e.g.Woebken et al., 2007). Here, we report on similar observations onfailure of standard FISH protocols to detect anammox bacteria in abiofilm sample from a wastewater laboratory reactor. To increase thepossibilities for in situ detection of anammox bacteria in this system,we modified the standard published FISH protocol and employed twoother FISH protocols (FISH using peptide nucleic acid probes (PNAFISH) and horse radish peroxidase (HRP) labeled probes andsubsequent catalyzed reporter deposition (CARD-FISH)) to demon-strate their applicability on anammox samples retrieved fromdifferent laboratory wastewater reactors designed for enhancedbiological nitrogen removal. Three of the samples originated fromlaboratory reactors with suspended biomass, consisting of aggregatedenrichment cultures of “Candidatus Kuenenia stuttgartiensis”, “Candi-datus Brocadia anammoxidans” and “Candidatus Anammoxoglobuspropionicus”, respectively, and one of the samples from a biofilmlaboratory reactor consisting of a mixture of “Candidatus Kueneniastuttgartiensis” and “Candidatus Brocadia anammoxidans”, nitrifyingbacteria and heterotrophs (T. Gaul, Leibniz Universität Hannover,unpublished results). The standard FISH protocol worked well on thesuspended biomass samples, but produced only weak probe signals inthe biofilm sample. Since several parameters may cause low probesignals when applying the standard FISH protocol, such as penetrationproblems of the probe through the cell wall boundary, or lowribosome content, we employed three ways to increase the probesignal intensities on all four different anammox samples: i) weinvestigated the effect of different enzymatic (lysozyme and protease)pre-treatment procedures of the exopolymer/cell walls prior to thehybridization when applying the standard FISH protocol; ii) weapplied and modified the PNA FISH protocol; and iii) we applied andmodified the CARD-FISH protocol with respect to optimal pre-treatment of cells and inactivation of endogenous peroxidases. Apartfrom application of these FISH protocols on the four anammoxsamples in this study, we included also 28 additional non-anammoxreference strains, representing genera from the four main metabolicgroups within the nitrogen cycle (aerobic and anaerobic ammoniumoxidation, diazotrophic bacteria, denitrification, and nitrite oxidation),two other obligate anaerobic groups (reductively dechlorinatingbacteria and sulphate reducing bacteria), some standard aerobiclaboratory strains, and different complex, anaerobic environmentalsamples like activated sludge and dechlorinating consortia.

2. Material and methods

2.1. Samples and reference strains

Samples containing high amounts of anammox bacteria wereobtained from a laboratory scale biofilm (moving bed) reactor,maintained in the Department of, Biophysics, Leibniz University,Hannover, Germany, and from suspended biomass enrichmentcultures containing granules of “Candidatus Kuenenia stuttgartiensis”,

Table 116S rRNA-targeted oligonucleotide probes used in this study.

Probe (full name) Target group Sequence (5′–3′)

EUB338 (S-D-Bact-0338-a-A-18) Domain Bacteria GCTGCCTCCCGTAGEUB338II (S-⁎-BactP-0338-a-A-18) Planctomycetales GCAGCCACCCGTAGEUB338III (S-⁎-BactV-0338-a-A-18) Verrucomicrobiales GCTGCCACCCGTAGBacUniv-1 Bacteria CTGCCTCCCGTAGGAMX368 (S-⁎-AMX-0368-a-A-18) All known Anammox-bacteria CCTTTCGGGCATTGNONSENSE Nonsense AGAGAGAGAGAGA

Escherichia coli 16S rRNA numbering, Brosius et al. (1981). Full name of 16S rRNA gene-targ

“Candidatus Brocadia anammoxidans” and “Candidatus Anammox-oglobus propionicus” from reactors maintained in the Department ofMicrobiology, Raboud University of Nijmegen, The Netherlands. Allsamples contained significant amounts of anammox bacteria, verifiedby activity measurements and 16S rRNA gene based studies. Furtherdetails about microbiology, operation and performance of thesereactors have been reported elsewhere (Kartal et al., 2007, 2008;Schmid et al., 2005; Strous et al., 2006). All samples were fixed with4% paraformaldehyde for 3 h as described by R. Amann (1995).Samples were stored in 50% ethanol-PBS (390 mM NaCl, 10 mMNaxPO4, pH 7.2–7.4) buffer at −20 °C. Escherichia coli K-12 cells wereused as a positive control during all different hybridization proce-dures. Additional 28 reference strains and environmental samples,including different pure cultures representing ammonium oxidizingbacteria, diazotrophic bacteria, denitrifying bacteria, nitrite-oxidizingbacteria, reductively dechlorinating bacteria, sulphate-reducing bac-teria, different aerobic laboratory isolates, dechlorinating consortiaand wastewater samples (summarized in Table 2) were used forCARD-FISH experiments.

2.2. Enzymatic pre-treatment of fixed samples prior to FISH usingfluorochrome labeled oligonucleotides (standard FISH)

Two different enzymatic pre-treatment protocols were tested,employing two different enzymes: a) lysozyme (Merck, Darmstadt,Germany) at concentrations of 0, 40,000, 400,000, 4,000,000 U/ml,b) protease from Streptomyces griseus (Sigma, Steinheim, Germany) atconcentrations of 0, 4.8, 48, 480 U/ml; dissolved in buffer containing100mMTris–HCl, 50mMEDTA, pH 8.0. Prior to the enzyme treatment,the fixed sample was homogenized for 20 s using an Ultra-Turrax(Janke & Kunkel, IKA-Labortechnik, Germany). Fixed homogenizedsamples (20 µl) were then washed by centrifugation with 1x PBS(390 mM NaCl, 10 mM NaxPO4, pH 7.2). After adding 20 µl of enzyme(lysozyme or protease), the sample was vortexed for 15 s andincubated for 20 min at room temperature and 37 °C. After washingwith 1x PBS, the cell pellet was resuspended in 96% ethanol to achievea final concentration of 50% ethanol for long-term storage.

2.3. Fluorescence in situ hybridization using fluorochrome labeledoligonucleotide probes (standard FISH)

Five different probes were used; two probes targeting anammoxbacteria at the phylum level, Planctomycetales (EUB338-II) and(AMX368); the general reference Bacteria probes EUB338-I, EUB338-III; and the negative control probeNONSENSE. The probe sequences andtarget sites for these probes are listed in Table 1. Oligonucleotides werelabeled either with 5(6)-carboxyfluorescein-N-hydroxysuccinimideester (FLUOS) or hydrophilic sulphoindocyanine dyes (Cy3, Cy5). Allprobeswere ordered fromThermoElectronGmbH, Ulm, Germany. All insitu hybridizations were performed as described by Amann (1995). Inbrief, 10 µl of the fixed, pre-treated and homogenized samples wereapplied on a teflon coated microscope slide containing ten wells(Marienfeld, Germany) and dried at 46 °C for 15 min, followed by asuccessive dehydration step with 50%, 80% and 96% ethanol for 3 mineach. 9 µl of hybridization buffer containing 0.9MNaCl, 20mMTris–HCl,

Target site Formamide concentration (%) Reference

GAGT 338–355 0–50 Amann et al. (1990)GTGT 338–355 0–50 Daims et al. (1999)GTGT 338–355 0–50 Daims et al. (1999)A 340–354 0 Perry-O'Keefe et al. (2001)CGAA 368–385 15 Schmid et al. (2005)GAGAG – 0 Loy et al. (2002)

eted oligonucleotide probe based on the nomenclature of Alm et al. (1996).

121M. Pavlekovic et al. / Journal of Microbiological Methods 78 (2009) 119–126

0.02%SDS, pH8.0with the optimal formamide concentration, as listed inTable 1, was mixed with 1 µl of labeled probe (30 ng/µl for Cy3 and Cy5labeled probes, 100 ng/µl for FLUOS labeled probes), and then pipettedon themicroscope slide. Hybridizationswere performed at 46 °C for 2 h.After this, the microscope slides were rinsed with prewarmed (48 °C)washing buffer, followed by 15 min immersion in washing buffer at48 °C. Simultaneous hybridization with probes requiring differentstringency was realized by a successive hybridization procedure(Wagner et al., 1994).

2.4. FISH using peptide nucleic acid probes (PNA FISH)

The probe BacUni-1 was applied as a general Bacteria probe,labeled with fluorescein (Table 1, AdvanDx, Inc. Woburn, USA).

Table 2Summary of peroxidase analyses and CARD-FISH experiments without HRP labelled probes

Cultures/samples Cultivation

Nitrifying bacteriaNitrosomonas europaea Nm50 Koops et alNitrosomonas eutropha Nm70 Koops et alNitrosomonas sp. III7 Utaker andNitrosospira sp. 40KI Utaker and

Nitrate reducing bacteriaAnaeromyxobacter dehalogenans 2CP-C ATCC BAA-259 R2A, 30 °CParacoccus denitrificans DSM 413T DSM mediuZoogloea ramigera ATCC 25935 ATCC Medi

Diazotrophic bacteriaAzotobacter vinelandii DSM366 DSM mediuKlebsiella oxytoca DSM7498 LB, 37 °C, a

Reductively dechlorinating bacteriaAnaeromyxobacter dehalogenans 2CP-C ATCC BAA-259 ATCC mediDehalobacter restrictus DSM 9455 DSM MediuDesulfitobacterium hafniense DSM 10664T DSM MediuGeobacter lovleyi SZ ATCC BAA-1151 DSM MediuSulfurospirillum halorespirans DSM 13726 DSM Mediu

Sulfate-reducing bacteriaDesulfobacter postgatei DSM2034T DSM mediuDesulfovibrio longus DSM6739T DSM Mediu

Other test bacteriaBacillus subtilis strain collection microbiology, TUM HD, 30 °C,Cytophaga johnsonae strain collection microbiology, TUM R2A, 30 °CEscherichia coli K12 LB, 37 °C, aMyxococcus xanthus DZW CYE, 30 °C,Pseudomonas fluorescens DSM 50090T LB, 30 °C, aNeisseria canis LMG8383T LB, 37 °C, aSaccharomyces cerevisiae, collection microbiology TUM DSM Mediu

Consortia with anammox bacteriaCandidatus Anammoxoglobus propionicus Kartal et alCandidatus Brocadia anammoxidans Kartal et alCandidatus Kuenenia stuttgartiensis Strous et aMoving bed biofilm reactor sample n.a.

Consortia with reductively dechlorinating bacteriaBioDechlor INOCULUM (Frank E. Loeffler, Georgia Tech) Daprato etDechlorinating consortium OW (Frank E. Loeffler, Georgia Tech) Amos et alTCE-contaminated aquifer sMilledgeville, GA, USA (Frank E. Loeffler, Georgia Tech) n.a.Chloro-organic contaminated aquifer sample Erding, Germany n.a.

Activated sludgeActivated sludge with nitrification and denitrification Lee et al. (2Activated sludge with denitrification Lee et al. (2

ATCC: American Type Culture Collection; DSMZ: Deutsche Sammlung von MikroorganismeHD: yeast dextrose agar; LB: Luria Bertani agar; CYE: casitone-yeast extract agar; R2A: Reas

a Detected with H2O2.b According to PeroxiBase (http://peroxibase.isb-sib.ch/).

Hybridization using PNA probes was performed in solution asdescribed by Perry-O'Keefe et al., 2001. In brief, 100 µl of fixed samplewas washed with 1x PBS, resuspended in 100 µl hybridization buffercontaining 25 mM Tris–HCl, pH 9.0, 100mMNaCl, 0.5% (w/v) SDS and100 nM labeled probe. Hybridizations were performed at 55 °C for30 min. After hybridization, the samples were resuspended in 500 µlof washing buffer containing 10 mM Tris, pH 9.0, 1 mM EDTA andincubated at 55 °C for 10 min. The washing step was repeated threetimes. After this, the sample was resuspended in 50 µl of washingsolution. Two modifications of this protocol were tested: a) extendinghybridization time, ranging from 30 min, to 2 and 15 h, respectively;b) hybridization with or without lysozym pre-treatment, as describedin Section 2.2. E. coli was used as a positive reference strain. Twonegative references were used: hybridization of anammox samples

in different anaerobic and aerobic pure cultures and complex environmental samples.

conditions Catalase/peroxidasea

Fluorescent CARD-FISHsignals withoutHRP-labelled probe

Peroxidasepresentb

. (1991) + + +

. (1991) + − +Nes (1998) + − n.a.Nes (1998) + − n.a.

, anaerobic − − +m 1, 30 °C, aerobic − − +um 13, 26 °C, aerobic + − n.a.

m 3, 30 °C, aerobic − − +erobic + − +

um 2282, 30 °C, anaerobic − − +m 723, 30 °C, anaerobic − − n.a.m 720, 30 °C, anaerobic − − +m 732, 30 °C, anaerobic + − +m 732, 30 °C, anaerobic − − n.a.

m 193, 37 °C, anaerobic + + n.a.m 63, 37 °C, anaerobic + − n.a.

aerobic + − +, aerobic + − +erobic + − +aerobic − − +erobic + − +erobic + − +m 186, 30 °C, aerobic ND ND ND

. (2007) + + n.a.

. (2008) + + n.a.l. (2006) + + +

+ + n.a.

al. (2007) + − n.a.. (2008) + − n.a.

+ − n.a.+ − n.a.

003) + No/weak signals n.a.003) + No/weak signals n.a.

n und Zellkulturen GmbH/German Collection of Microorganisms and Cell Cultures;oner and Geldreich, 1985; n.a.: not applicable; ND: not determined.

Fig. 1. Staining of fixed biomass frommoving bed reactor with fluoresceinisothiocyanat(A) and Concanavalin A (B) for visualization of proteins and sugar residues. Scale bar20 µm.

122 M. Pavlekovic et al. / Journal of Microbiological Methods 78 (2009) 119–126

without the BacUni-1 probe, and Saccharomyces cerevisiae as anegative reference strain for the probe BacUni-1.

2.5. FISH using horse radish peroxidase (HRP) labeled probes andsubsequent catalyzed reporter deposition (CARD-FISH)

Four different probes were used for CARD-FISH: EUB338-I,EUB338-II, EUB338-III and NON-EUB (Table 1). All probes werelabeled with horse-radish peroxidase, and ordered from ThermoElectron GmbH, Ulm, Germany. These probes were used according tothe CARD-FISH protocol described by Teira et al. (2004). In brief,100 µlof fixed sample was incubated in 500 µl eppendorf tubes withlysozyme (10mg/ml in 0.05 EDTA, 0.1 Tris–HCl [pH 8]) at 37 °C for 1 h.The sample was washed with distilled water by centrifugation andincubated at room temperature for 20 min with hydrochloric acid(0.01 M). After a two time washing step with sterile, distilled waterthe sample was dehydrated with absolute ethanol. Hybridization wasperformed at 35 °C for 15 h using 100 µl of hybridization buffer (0.9 MNaCl, 20 mM Tris–HCl [pH 7.5], 10% [w/v] dextran sulfate, 0.02% [w/v]sodium dodecyl sulphate,1% blocking reagent [BoehringerMannheim,Mannheim, Germany], and 55% [v/v] formamide) together with thehorse-radish peroxidase labeled probe (0.28 ng/μl). After hybridiza-tion, the hybridization buffer was removed by centrifugation, and thesample was incubated with 100 µl of prewarmed washing buffer(5 mM EDTA [pH 8], 20 mM Tris–HCl [pH 7.4 to 7.6], 0.01% [w/v]sodiumdodecyl sulphate,13mMNaCl) at 37 °C for 10min, followed byphosphate-buffered saline (PBS-T) (145 mM NaCl, 1.4 mM NaH2PO4,8 mM Na2HPO4 [pH 7.6]), containing 0.05% Triton X-100 (Sigma-Aldrich, Germany) at room temperature for 15min. After this,100 µl ofamplification buffer (10% [w/v] dextran sulfate, 2 M NaCl, 0.1% [w/v]blocking reagent, 0.0015% H2O2 in PBS) were added to the sample,mixed with 0.68 µl of tyramide-Alexa 488 (1 mg/ml), and thenincubated at 37 °C for 45 min. After amplification, the sample waswashed in PBS-T at room temperature for 15 min, followed by distilledwater and absolute ethanol. For some experiments, four differentattempts to inactivate endogenous peroxidases were made, prior tothe permeabilisation step: i) acid treatment using hydrochloric acid(0.01 M) for 40 and 60 min, respectively (modified from Teira et al.,2004); ii) addition of 30% H2O2 (Merck, Darmstadt, Germany) for 20,60 and 180 min, respectively; iii) heat treatment at 80 °C for 10 min;and iv) autoclaving of sample for 20 min. Four anammox samples and28 different reference strains, representing anaerobic (denitrifying,reductively dechlorinating and sulphate-reducing strains), aerobic(e.g. nitrifying strains) and ten different complex environmentalsamples like activated sludge and dechlorinating methanogenicconsortia were used as reference strains where appropriate (seefurther Table 2).

2.6. Staining of exopolymeric substances with fluoresceinisothiocyanat(FITC) and Concanavalin A (ConA)

For microscopic detection of proteins and amine-sugars inextrapolymeric substance (EPS), the fixed biofilm sample was stainedwith fluoresceinisothiocyanat (FITC, Molecular Probes, Schmid et al.,2003), and for sugar residues like glycoconjugates with ConcanavalinA (ConA, Molecular Probes), following the protocol of McSwain et al.,2005. For this, the sample was homogenized and mixed with 100 µlFITC (0.01%) or 100 µl ConA (100 µg/ml) in 0.6 ml eppendorf tubes for15 min at room temperature in the dark. After this, the sample waswashed with 100 µl phosphate-buffered saline (1× PBS).

2.7. Screening of peroxidases and catalases

For macroscopic detection of catalases/peroxidases a drop of 30%H2O2 (Merck, Darmstadt, Germany)was applied on a suspended smearof sample on a microscope slide. The formation of gas bubbles is taken

as an indication for the presence of active catalases, and to some extentof peroxidases (Gerhardt et al., 1994). For microscopic detection ofintracellular peroxidases, a modified staining protocol of Arroyo et al.(1999), based on 2′,7′-dichlorofluorescein diacetate (Sigma-Aldrich,Germany) was used. After washing the sample once with distilledwater, 2′,7′-dichlorofluorescein diacetate was added to a finalconcentration of 5 µl/l, with and without H2O2 (1 µl per 100 µlsample) and incubated at 37 °C for 20min. Green fluorescence is takenas an indication for presence of peroxidases. Screening for peroxidasegene coding regions in available genome sequenceswere performedbythe search tool of the PeroxiBase (Passardi et al., 2007) web site(http://peroxibase.isb-sib.ch/).

2.8. Microscopy and image analysis

For microscopy, teflon coated microscope slides containing tenwells (Marienfeld, Germany) were used. Samples that had beenhybridized in solution (PNA FISH, CARD FISH) were applied on thewells (10 µl) and dried at 46 °C for 15 min. After this, they wereembedded with 0.5% agarose in TAE buffer (4 M Tris, 2 M ice coldacetic acid, 0.2 M EDTA) and dried for 10 min at 40 °C. All hybridizedsamples were mounted with Citifluor AF1 (Citifluor Ltd, UK) prior tomicroscopic observation. For image acquisition, a Zeiss LSM 510 laserscanning confocal microscope (Carl Zeiss, Göttingen, Germany)equipped with two Helium/Neon-Lasers (543 nm and 633 nm), oneArgon-ion Laser (450–514 nm) and one UV-Laser (351–364 nm) wasused together with the Zeiss LSM software package V.2.01 SP2.Systematic evaluation of the signal intensities was done by recordingimages of independent visual fields (encompassing at least 100 cells),followed by digital image analysis using the daime software (Daimset al., 2005).

3. Results and discussion

3.1. Enzymatic pretreatment of cells for optimization of standard FISHwith fluorescently labeled oligonucleotide probes

Good probe signals were obtained with the EUB probe mix with allthree suspended anammox biomass samples and the reference E. colistrain, but not with the anammox biofilm sample, possibly due toprobe penetration problems through exopolymeric substances (EPS)surrounding the cells embedded in the biofilm, or rigid cell structures.

Fig. 2. Standard FISHwith probe EUB II on biofilm sample before (A) and after (B) enzymepre-treatment (400,000 U/mg of lysozyme). FISH images are always shown on the leftside, and the corresponding phase contrast image on the right side. Scale bar 20 µm.

123M. Pavlekovic et al. / Journal of Microbiological Methods 78 (2009) 119–126

The chemical nature of the EPS in the biofilm sample was exploredwith the fluorescent stains, FITC and ConA (Schmid et al., 2003;McSwain et al., 2005), in order to identify an appropriate strategy forreducing probe penetration problems. Although these stains havebeen used successfully in other studies (e.g. McSwain et al., 2005),these stains did not yield any useful information on the anammoxsamples in this study. FITC showed some indications on protein-likestructures, whereas ConA only produced fluorescent signals inparticles of unknown character distributed throughout the wholesample (Fig. 1). This suggests that further systematic investigations,employing for example a panel of different lectin probes, are neededto optimize the usefulness of such protocols. Since these so farpublished protocols could not present more precise information aboutthe nature of the matrix of EPS and cell wall complex, two differentenzymes, lysozyme and protease, were then chosen to increase theprobe permeability in general as these enzymes have often been usedfor this purpose (e.g. Meier et al., 1999). Pre-treatment of fixed biofilmwith both enzymes at different concentrations produced significantlyhigher probe signal intensities (Fig. 2). However, protease treatment

Fig. 3. PNA FISH with and without probe Bac-Uni 1 on biofilm sample after 30 min (A), ancorresponding phase contrast image on the right side. Scale bar 20 µm.

often resulted in a severe loss of biomass during subsequenthybridization steps. Thus, the final optimal enzymatic pre-treatmentprotocol was based on lysozyme, at concentrations ranging from40,000 and 400,000 U/ml, for 20 min incubation at 37 °C and pH 7.

3.2. Optimization of PNA FISH with fluorescently labeled PNA probes

PNA probes are generally considered superior to standardfluorescently labeled oligonucleotide probes (Perry-O'Keefe et al.,2001). However, when applying the BacUniv-1 probe on theanammox samples according to Perry-O'Keefe et al. (2001), strongprobe signals could be detected only in the suspended biomassanammox samples, but not in the anammox biofilm sample. Twomodifications were tested for optimization of the PNA FISH protocolon the biofilm sample: i) pre-treatment of fixed samples with twodifferent enzymes as described in Section 2.2; and ii) increase ofhybridization time, from 30 min to 2 h and 15 h, respectively. Ofthese, only prolonged hybridization (2 h or more) produced sig-nificantly enhanced probe signals in the biofilm sample that werecomparable to the results with oligonucleotide probes presented inSection 2.2 (Fig. 3). All suspended samples, as well as E. coli cells(positive control for all experiments), always showed equally strongprobe signals despite different modifications of the hybridizationprocedure. However, compared to the probe signals obtained from E.coli cells after the modifications of the PNA FISH protocols, the probesignals in the suspended samples were always slightly blurred(results not shown). Negative controls, using either autoclaved E.coli cells or fixed S. cerevisae, produced no probe signals with theBacUniv-1 probe.

3.3. CARD-FISH with horse-radish peroxidase (HRP) labeledoligonucleotide probes

When applying CARD-FISH to the samples listed in Table 2, signalswere up to four times stronger compared to signals obtained with thestandard FISH protocol (Figs. 4 and 7). In addition to this, a higherdiversity of cell morphotypes could be detected with CARD-FISH incomparison with standard FISH and PNA FISH (Fig. 4). However, allnegative controls (hybridizationwithout a HRP labeled probe (Fig. 5);and the NONSENSE HRP labeled probe (results not shown), respec-tively) showed equally strong signals in similar morphologicalstructures as the hybridization with HRP labeled probes, whereas

d 2 h (B) hybridization time. FISH images are always shown on the left side, and the

Fig. 4. CARD-FISHwith EUB probemix on biofilm sample. Comparison of different morphologies detectable with CARD-FISH (A) and Oligo FISH (B). FISH images are always shown onthe left side, and the corresponding phase contrast image on the right side. Scale bar 20 µm.

124 M. Pavlekovic et al. / Journal of Microbiological Methods 78 (2009) 119–126

the probe signals in E. coli cells were negative under all theseconditions (results not shown). Such unspecific CARD-FISH signalshave been attributed to endogenous peroxidases in a number ofprevious studies (e.g. Ishii et al., 2004; Teira et al., 2004; Woebken etal., 2007). The influence of endogenous peroxidase on the unspecificCARD-FISH signals in the anammox samples in this study was furtherconfirmed by the strong detection of catalases/peroxidases in thecatalase test, as described in Section 2.7. However, when applyingCARD-FISH on 28 other different catalase/peroxidase positive bacteria(aerobic ammonium oxidizing bacteria, diazotrophic bacteria, deni-trifying, nitrite oxidizing bacteria, reductively dechlorinating bacteria,and sulphate-reducing bacteria) pure cultures and ten differentcomplex, environmental samples (activated sludge, anaerobic aquifersamples and dechlorinating consortia), only a few strains (aerobicammonium oxidizing species (Nitrosomonas europaea and onesulfate-reducing strain (Desulfobacter postgatei) produced unspecific

Fig. 5. CARD-FISH after different pre-treatment approaches to eliminate unspecific signals onwas done without any horse-radish peroxidase labeled probe. A–C: Acid treatment at differeperoxidases with H2O2. FISH images are always shown on the left side, and the correspond“Candidatus Kuenenia stuttgartiensis”, and “Candidatus Anammoxoglobus propionicus”, aPavlekovic_et_al_2009a.html.

CARD-FISH signals (Table 2). This suggests that the attribution ofunspecific CARD FISH signals to endogenous peroxidases needs to befurther investigated to reveal the true cause of unspecific CARD-FISHsignals (see further Section 3.5).

3.4. Inactivation of endogenous peroxidases prior to CARD-FISH

Different protocols have been employed to inactivate endogenousperoxidases in cells, such as hydrochloric acid for 20 min; orincubation of cells with 0.15–3% H2O2 for 10–30 min (Teira et al.,2004; Ishii et al., 2004; Woebken et al., 2007). However, hydrochloricacid treatment did not inactivate the peroxidases in the anammoxsamples in this study, despite prolonged incubation periods to 40 and60 min (Figs. 6 and 7). Thermal treatment, such as pasteurization(80 °C for 10min) or autoclaving, was also not suitable for inactivationof the endogenous peroxidases. Pasteurization showed no effect,

anammox enrichment cultures of “Candidatus Brocadia anammoxidans”. Hybridizationnt time intervals (20, 40, 60 min). D: Pasteurization. E: Autoclavation. F: Denaturation ofing phase contrast image on the right side. Scale bar 20 µm. For similar image series ofnd anammox biofilm samples, see the weblink www.microbial-systems-ecology.de/

Fig. 6. Inactivation of unspecific fluorescence signals probably caused by endogenousperoxidase activity using different pre-treatment methods. After pre-treatment, CARD-FISH hybridization and amplification steps were performed as previously described,except that hybridization was performed without the addition of any horse-radishperoxidase labeled probe. For amplification, the tyramide dye Alexa Fluor 488(Molecular Probes) was used.

125M. Pavlekovic et al. / Journal of Microbiological Methods 78 (2009) 119–126

whereas autoclaving resulted in a somewhat weaker (but stilldetectable) probe signals following CARD-FISH without HRP labeledprobes (Figs. 6 and 7). The most optimal inactivation of endogenousperoxidases (including also other enzymes such as catalases) wasobtained with pre-treatment with H2O2, where the concentration andincubation time proved to be critical, and optimal conditions for theanammox samples in this varied from those previously applied onother samples as described by e.g. Ishii et al. (2004), and Woebkenet al. (2007). Successful inactivation of endogenous peroxidases wasonly obtained after incubation for maximum 20 min with 30% H2O2.Longer exposure of the sample to H2O2 deteriorated the consistencyand cellular structures of the sample (results not shown).

3.5. Screening of endogenous peroxidases/catalases

For screening of endogenous peroxidases, three differentapproaches were employed as described in Section 2.7. This wasapplied on 32 different aerobic and anaerobic pure cultures andcomplex, anaerobic environmental samples, as listed in Table 2. Forthe anammox samples, live, pasteurized as well as autoclaved cellswere used in addition to the catalase test. Live samples from thepositive control E. coli and the suspended anammox samples(“Candidatus Kuenenia stuttgartiensis”, “Candidatus Brocadia ana-mmoxidans” and “Candidatus Anammoxoglobus propionicus”) pro-

Fig. 7. Comparison of standard FISH and CARD-FISH probe signal intensities dependingon the incubation time with lysozyme. More than 4 times stronger signal intensities(based on evaluation of the probe signal intensities with the daime software (Daims etal., 2005) were measured for CARD-FISH in comparison to standard FISH. Prior to CARD-FISH the sample was incubated for 20 min with H2O2 for inactivation of peroxidases.Probes were labeled with FLUOS dye. w/o = without.

duced large amounts of gas bubbles after addition of H2O2 (results notshown). This reaction was somewhat weaker if the sample waspasteurized or autoclaved (exception autoclaved “Candidatus Brocadiaanammoxidans” where no gas bubbles could be observed). Theseobservations corresponded well with the false-positive probe signalswhen CARD-FISH without HRP labelled probes was applied on fixed,pasteurized or autoclaved samples, and thus indicate that endogenousperoxidases are responsible for false-positive probe signals inanammox bacteria. 2′,7′-dichlorofluorescein diacetate producedstrong signals for the positive control E. coli, and for the anammoxsamples green fluorescent signals could be obtained. However, thesignals obtained were faint and blurred and it is therefore question-able if this stain can be used as a significant indicator of peroxidases inanammox samples. Screening for peroxidase coding regions inavailable genome sequences revealed a large phylogenetic diversityof different peroxidase classes in every species investigated (Table 2).This raises the question if all, or, if a specific peroxidase class should beattributed for unspecific CARD-FISH signals. We speculate that theunspecific CARD-FISH signals could either be attributed to a specificperoxidase class characteristic for certain prokaryotic groups likeanammox bacteria, or to a minimum concentration level of perox-idases (the amount of peroxidases is particularly high in e.g. anammoxbacteria and nitrifying bacteria). However, more detailed researchabout the influence of the peroxidase biochemistry in differentprokaryotic groups on the CARD-FISH reaction is needed in order toevaluate this further.

4. Conclusions

➢ In contrast to previous FISH studies on anammox biomass inwastewater treatment samples, we found that standard FISHprotocols may not be applicable on dense biofilm samples. Forimproved hybridization result of dense biofilms, an enzymatic pre-treatment, prior to the hybridization may have to be implemented.

➢ In contrast to previous studies claiming the superiority of PNA FISHprobes over standard fluorescently labeled oligonucleotide FISHprobes, we found that PNA FISH may not be applicable on allsamples, in particular on dense biofilm samples. For our particularsamples, higher PNA FISH signal intensities (at least equal to theintensity of standard FISH) could be obtained by a prolongedhybridization time.

➢ The CARD-FISH protocol produced not only up to four timesstronger probe signals compared to the standard and the PNA FISHprotocol, but detected also a significantly higher diversity ofbacteria than the other FISH tools. The main drawback with CARD-FISH is however that unspecific probe signals due to endogenousperoxidase activity may be obtained, which demands, with respectto species or sample type, further individual optimization of

Table 3Summary of the limits and the possibilities of the three FISH protocols applied in thisstudy on anammox biomass from laboratory wastewater reactors.

Parameter OligonucleotideFISH

PNA FISH CARD-FISH

Enzymatic pre-treatment Partly necessary Partly necessary NecessaryDenaturation of disturbingenzymes

No No Yesa

Hybridization time ~2 h 30 min to 2 h ≤15 hTotal process time(after fixation)

~3 h 1 h to 3 h N15–18 h

Shelf life of probes NN1 year NN1 year 6 months–1 yearCosts Low High MediumSimultaneous combinationwith other probes

Possible Possible Only possible insuccessivehybridization steps

a Dependent on the endogenous peroxidase activity or concentration in certainspecies (see further Table 2).

126 M. Pavlekovic et al. / Journal of Microbiological Methods 78 (2009) 119–126

specific inactivation protocols by for example H2O2 treatment priorto hybridization. However, we also observed that most otheraerobic or anaerobic peroxidase positive strains (including e.g.certain groups in the nitrogen cycle, reductively dechlorinatingbacteria and sulfate-reducers) or complex environmental samplesdo not produce unspecific CARD-FISH signals. Since CARD-FISH isincreasingly used in microbial ecological studies, we suggest thatthe current attribution of endogenous peroxidase activity as themain contributor to unspecific CARD-FISH signals needs to beexpanded by more detailed knowledge about the peroxidasebiochemistry and other biochemical phenomena that may influ-ence CARD-FISH.

➢ In summary, all three FISH protocols applied in this study can besuitable for in situ detection of anammox bacteria in wastewatersamples, once individual optimization of the protocols have beendone with respect to the sample type. Each protocol presents itsindividual advantages as well as disadvantages (summarized inTable 3), depending on the user's needs and possibilities.

Acknowledgements

Research on anammox and diazotrophic bacteria was supported bythe DFG projects KU 602/6-1 and LU412/4-2. Research on reductivelydechlorinating strains was supported by the Helmholtz project Virtualinstitute for isotope biogeochemistry and redox gradients undercontract VH-VI-155 (to N. M. L). We thank Bernhard Fuchs and RudolfAmann, Max Planck Institute, Bremen, for valuable discussions ondifferent aspects of FISH technology, and Sabine Kunst, PotsdamUniversität, and Wolfgang Liebl, TUM, for general support.

References

Alm, E.W., Oerther, D.B., Larsen, N., Stahl, D.A., Raskin, L., 1996. The oligonucleotideprobe database. Appl. Environ. Microbiol. 62, 3557–3559.

Amann, R., 1995. In situ identification of micro-organisms by whole cell hybridizationwith rRNA-targeted nucleic acid probes. In: Akkermans, A.D.L., van Elsas, J.D., deBruijn, F.J. (Eds.), Molecular Microbial Ecology Manual. Kluwer Academic Publish-ers, Dordrecht, pp. 1–15. section 3.3.6.

Amann, R., Fuchs, B., 2008. Single-cell identification in microbial communities byimproved fluorescence in situ hybridization techniques. Nat. Rev. Microbiol. 6,339–348.

Amann, R., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R., Stahl, D.A., 1990.Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry foranalyzing mixed microbial populations. Appl. Environ. Microbiol. 56, 1919–1925.

Amos, B.A., Ritalahti, K.M., Cruz-Garcia, C., Padilla-Crespo, E., Löffler, F.E., 2008. Oxygeneffect on Dehalococcoides viability and biomarker quantification. Environ. Sci.Technol. 42, 5718–5726.

Arroyo, G., Sanz, P.D., Préstamo, G., 1999. Response to high-pressure, low-temperaturetreatment in vegetables: determination of survival rates of microbial populationsusing flow cytometry and detection of peroxidase activity using confocalmicroscopy J. Appl. Microbiol. 86, 544-544.

Brosius, J., Dull, T.J., Sleeter, D.D., Noller, H.F., 1981. Gene organization and primarystructure of a ribosomal RNA operon from Escherichia coli. J. Mol. Biol. 148, 107–127.

Daims, H., Bruhl, A., Amann, R., Schleifer, K.H., Wagner, M., 1999. The domain-specificprobe EUB338 is insufficient for the detection of all bacteria: development andevaluation of a more comprehensive probe set. Syst. Appl. Microbiol. 22, 434–444.

Daims, H., Lücker, S., Wagner, M., 2005. daime, a novel image analysis program formicrobial ecology and biofilm research. Environ. Microbiol. 8, 200–213.

Daprato, R.C., Löffler, F.E., Hughes, J.B., 2007. Comparative analysis of three tetrachlor-oethene to ethene halorespiring consortia suggests functional redundancy.Environ. Sci. Technol. 41, 2261–2269.

Gerhardt, P., Murray, R.G.E., Wood, W.A., Krieg, N.R., 1994. Methods for General andMolecular Bacteriology. ASM, Washington DC.

Giovanonni, S., DeLong, E., Olsen, G., Pace, N., 1988. Phylogenetic group-specificoligodeoxy-nucleotide probes for identification of single microbial cells. J. Bacteriol.170, 720–726.

Hoshino, T., Yilmaz, L.S., Noguera, D.R., Daims, H., Wagner, M., 2008. Quantification oftarget molecules needed to detect microorganisms by fluorescence in situhybridization (FISH) and catalyzed reporter deposition-FISH. Appl. Environ.Microbiol. 74, 5068–5077.

Ishii, K., Mußmann, M., MacGregor, B.J., Amann, R., 2004. An improved fluorescence insitu hybridization protocol for the identification of bacteria and archaea in marinesediments. FEMS Microbiol. Ecol. 50, 203–213.

Kartal, B., Rattray, J., van Niftrik, L., van de Vossenberg, J.L.C.M., Schmid, M.C., Webb, R.I.,et al., 2007. “Candidatus Anammoxoglobus propionicus” a new propionate oxidizingspecies of anaerobic ammonium oxidizing bacteria. Syst. Appl. Microbiol. 30,39–49.

Kartal, B., van Niftrik, L., Rattray, J., van der Vossenberg, J.L.C.M., Schmid, M.C., SinningheDamsté, J., et al., 2008. “Candidatus Brocadia fulgida”: an autofluorescent anaerobicammonium oxidizing bacterium. FEMS Microbiol. Ecol. 63, 46–55.

Koops, H.-P., Böttcher, B., Möller, U.C., Pommerening-Röser, A., Stehr, G., 1991.Classification of eight new species of ammonia-oxidizing bacteria: Nitrosomonascommunis sp. nov., Nitrosomonas ureae sp. nov., Nitrosomonas aestuarii sp. nov.,Nitrosomonas marina sp. nov., Nitrosomonas nitrosa sp. nov., Nitrosomonas eutrophasp. nov., Nitrosomonas oligotropha sp. nov. and Nitrosomonas halophila sp. nov.J. Gen. Microbiol. 137, 1689–1699.

Kubota, K., Ohashi, A., Imachi, H., Harada, H., 2006. Improved in situ hybridizationefficiency with locked-nucleic-acid-incorporated DNA probes. Appl. Environ.Microbiol. 72, 5311–5317.

Kuenen, J.G., 2008. Anammox bacteria: from discovery to application. Nat. Rev.Microbiol. 6, 320–326.

Lee, N.M., Nielsen, P.H., Aspegren, H., Henze, M., Schleifer, K.H., la Cour Jansen, J., 2003.Long-term population dynamics and in situ physiology of the microbial communityin wastewater treatment plants with enhanced biological phosphorus removal.Syst. Appl. Microbiol. 26, 211–227.

Lehtola, M.J., Torvinen, E., Miettinen, I.T., Keevil, C.W., 2006. Fluorescence in situhybridization using peptide nucleic acid probes for rapid detection of Mycobac-terium avium subsp. avium and Mycobacterium avium subsp. paratuberculosis inpotable-water biofilms. Appl. Environ. Microbiol. 72, 848–853.

Loy, A., Lehner, A., Lee, N., Adamczyk, A., Meier, H., Ernst, J., et al., 2002. Oligonucleotidemicroarray for 16S rRNA gene-based detection of all recognized lineages of sulfate-reducing prokaryotes in the environment. Appl. Environ. Microbiol. 68, 5064–5081.

McSwain, B.S., Irvine, R.L., Hausner, M., Wilderer, P.A., 2005. Composition anddistribution of extracellular polymeric substances in aerobic flocs and granularsludge. Appl. Environ. Microbiol. 71, 1051–1057.

Meier, H., Amann, R., Ludwig, W., Schleifer, K.-H., 1999. Specific oligonucleotide probesfor in situ detection of a major group of gram-positive bacteria with low DNA G+Ccontent. Syst. Appl. Microbiol. 22, 186–196.

Passardi, F., Theiler, G., Zamocky, M., Cosio, C., Rouhier, N., Teixera, F., 2007. PeroxiBase:the peroxidase database. Phytochemistry 68, 1605–1611.

Pernthaler, A., Pernthaler, J., Amann, R., 2002. Fluorescence in situ hybridization andcatalyzed reporter deposition for the identification of marine bacteria. Appl.Environ. Microbiol. 68, 3094–3101.

Perry-O'Keefe, H., Rigby, S., Oliveira, K., Sorensen, D., Stender, H., Coull, J., et al., 2001.Identification of indicator microorganisms using a standardized PNA FISH method.J. Microbiol. Methods 47, 281–292.

Reasoner, D.J., Geldreich, E.E., 1985. A newmedium for the enumeration and subcultureof bacteria from potable water. Appl. Environ. Microbiol. 49, 1–7.

Schmid, M., Thill, A., Purkhold, U., Walcher, M., Bottero, J.Y., Ginestet, P., et al., 2003.Characterization of activated sludge flocs by confocal laser scanning microscopyand image analysis. Water Res. 37, 2043–2052.

Schmid, M.C., Maas, B., Dapena, A., van de Pas-Schoonen, K., van de Vossenberg, J.,Kartal, B., vanNiftrik, L., Schmidt, I., Cirpus, I., Kuenen, J.G.,Wagner,M., Damste, J.S.S.,Kuypers, M., Revsbech, N.P., Mendez, R., Jetten, M.S.M., Strous, M., 2005. Biomarkersfor in situ detection of anaerobic ammonium-oxidizing (ANAMMOX) bacteria. Appl.Environ. Microbiol. 71, 1677–1684.

Strous, M., Pelletier, E., Mangenot, S., Rattei, T., Lehner, A., Taylor, M.W., et al., 2006.Deciphering the evolution and metabolism of an anammox bacterium from acommunity genome. Nature 440, 790–794.

Teira, E., Reinthaler, T., Pernthaler, A., Pernthlaer, J., Herndl, G.J., 2004. Combiningcatalyzed reporter deposition-fluorescence in situ hybridization and microautor-adiography to detect substrate utilization by bacteria and Archaea in the deepocean. Appl. Environ. Microbiol. 70, 4411–4414.

Utaker, J.B., Nes, I.F., 1998. A qualitative evaluation of the published oligonucleotidesspecific for the 16S rRNA gene sequences of the ammonia-oxidizing bacteria. Syst.Appl. Microbiol. 21, 72–88.

Wagner, M., Amann, R., Kämpfer, P., Assmus, B., Hartmann, A., Hutzler, P., 1994.Identification and in situ detection of gram-negative filamentous bacteria inactivated sludge. Syst. Appl. Microbiol. 17, 405–407.

Wagner, M., Horn, M., Loy, A., 2003. Fluorescence in situ hybridization for theidentification and characterization of prokaryotes. Curr. Opin. Microbiol. 6,302–309.

Woebken, D., Fuchs, B.M., Kuypers, M.M., Amann, R., 2007. Potential interactions ofparticle-associated anammox bacteria with bacterial and archaeal partners in theNamibian upwelling system. Appl. Environ. Microbiol. 73, 4648–4657.

Zwirglmaier, K., 2005. Fluorescence in situ hybridisation (FISH) — the next generation.FEMS Microbiol. Lett. 246, 151–158.