©2019 The Author(s)

This is an Open Access book distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying and redistribution for noncommercial purposes, provided the original work is properly cited and that any new works are made available on the same conditions (http://creativecommons.org/licenses/by/4.0/). This does not affect the rights licensed or assigned from any third party in this book.

This title was made available Open Access through a partnership with Knowledge Unlatched.

IWA Publishing would like to thank all of the libraries for pledging to support the transition of this title to Open Access through the KU Select 2018 program.

# **FISH Handbook for Biological Wastewater Treatment**

Identification and quantification of microorganisms in activated sludge and biofilms by FISH

Editor(s): Per Halkjær Nielsen, Holger Daims and Hilde Lemmer

Editor(s): Per Halkjær Nielsen,

by IWA Publishing user on 04 February 2019

Holger Daims and Hilde Lemmer

Identification and quantification of microorganisms in activated sludge and biofilms by FISH

**FISH Handbook for Biological Wastewater Treatment**

www.iwapublishing.com

epifluorescence microscopy.

reliable FISH identification and quantification.

present wastewater treatment plants.

readers to use FISH in the context of wastewater treatment.

The *FISH Handbook for Biological Wastewater Treatment* provides all the required information for the user to be able to identify and quantify important microorganisms in activated sludge and biofilms by using fluorescence in situ hybridization (FISH) and

It has for some years been clear that most microorganisms in biological wastewater systems cannot be reliably identified and quantified by conventional microscopy or by traditional culture-dependent methods such as plate counts. Therefore, molecular biological methods are vital and must be introduced instead of, or in addition to, conventional methods. At present, FISH is the most widely used and best tested of these methods. This handbook presents all relevant information from the literature and, based on the extensive experience of the authors, advice and recommendations are given for

The overall purpose of the book is to help scientists, consultants, students, and plant operators to get an overview of important microorganisms in biological wastewater treatment and to explain how FISH can be used for detecting and quantifying these microbes. A proper and reliable identification of dominant microorganisms is of great importance for research and new developments in the wastewater treatment industry, and it is important for optimization and troubleshooting of operational problems in

The book encompasses an overview of dominant microorganisms present in the wastewater treatment systems, which oligonucleotide probes (gene probes) to select for detection of these microbes by FISH, how to perform FISH (detailed protocols), how to quantify the microbes, and how to solve common problems of FISH. The book addresses several functional groups: nitrifiers, denitrifiers, polyphosphate-accumulating organisms, glycogen-accumulating organisms, bacteria involved in hydrolysis and fermentation, filamentous bacteria from bulking sludge, and scum-forming bacteria. A comprehensive collection of FISH-images showing dominant representatives of these groups helps

ISBN 13: 9781843392316 ISBN: 1843392316

## FISH Handbook for Biological Wastewater Treatment

## FISH Handbook for Biological Wastewater Treatment

Identification and quantification of microorganisms in activated sludge and biofilms by FISH

Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer

#### Published by IWA Publishing Alliance House 12 Caxton Street London SW1H 0QS, UK Telephone: þ44 (0)20 7654 5500 Fax: þ44 (0)20 7654 5555 Email: publications@iwap.co.uk Web: www.iwapublishing.com

First published 2009 *#* 2009 IWA Publishing

Cover images provided by Simon Jon Mcllroy, Jeppe Lund Nielsen, and Kilian Stoecker Cover design by www.designforpublishing.co.uk Typeset in India by Alden Prepress Services Private Limited.

Index provided by Alden Prepress Services Private Limited.

Printed by Page Bros Ltd, Norwich, UK.

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the UK Copyright, Designs and Patents Act (1998), no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior permission in writing of the publisher, or, in the case of photographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of licenses issued by the appropriate reproduction rights organization outside the UK. Enquiries concerning reproduction outside the terms stated here should be sent to IWA Publishing at the address printed above.

The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility or liability for errors or omissions that may be made.

#### Disclaimer

The information provided and the opinions given in this publication are not necessarily those of IWA and should not be acted upon without independent consideration and professional advice. IWA and the Author will not accept responsibility for any loss or damage suffered by any person acting or refraining from acting upon any material contained in this publication.

British Library Cataloguing in Publication Data A CIP catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress

ISBN 10: 1843392316 ISBN 13: 9781843392316




by IWA Publishing user on 04 February 2019

## List of contributors

## Editors:

## Professor Per Halkjær Nielsen

Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark Fax: 45 9814 1808 Phone: 45 9940 8503 Email: phn@bio.aau.dk

### Dr. Hilde Lemmer

Bavarian Environment Agency, Kaulbachstr. 37, D-80539 Munich, Germany Phone: þ49 89 2180 2783 Fax: þ49 89 2800838 Email: lemmer@oec.net

### Dr. Holger Daims

Department of Microbial Ecology Vienna Ecology Centre University of Vienna, Althanstrasse 14 A-1090 Vienna, Austria Phone: þ43 1 4277 54392 Fax: þ43 1 4277 54389 Email: daims@microbial-ecology.net

## Other contributors:

### Dr. Aviaja A. Hansen

Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark Email: aah@bio.aau.dk

### Dr. Caroline Kragelund

Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark Email: ckr@bio.aau.dk

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

### Dr. Frank Maixner

Department of Microbial Ecology, Vienna Ecology Centre University of Vienna, Althanstrasse 14, A-1090 Vienna, Austria Email: maixner@microbial-ecology.net

#### Simon Jon McIlroy

Biotechnology Research Centre, La Trobe University, Bendigo, Vic 3552, Australia Email: simon.j.mcilroy@gmail.com

#### Artur Tomasz Mielczarek

Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark Email: atm@bio.aau.dk

#### Dr. Elisabeth Mu¨ller

Institute of Water Quality Control, Technical University Munich, Am Coulombwall, D-85748 Garching, Germany Email: Lisa.L.Mueller@bv.tum.de

#### Hien Thi Thu Nguyen

Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark Email: hn@bio.aau.dk

#### Dr. Jeppe Lund Nielsen

Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Sohngaardsholmsvej 49, DK-9000 Aalborg, Denmark Email: jln@bio.aau.dk

#### Dr. Margit Schade

Bavarian Environment Agency, Kaulbachstr. 37, D-80539 Munich, Germany Email: margit.schade@lfu.bayern.de

### Dr. Markus C. Schmid

Department of Microbial Ecology, Vienna Ecology Centre University of Vienna, Althanstrasse 14, A-1090 Vienna, Austria Email: schmid@microbial-ecology.net

### Professor Robert J. Seviour

Biotechnology Research Centre, La Trobe University, Bendigo, Vic 3552, Australia Email: R.Seviour@latrobe.edu.au


*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

## Introduction

## Per Halkjær Nielsen, Holger Daims, and Hilde Lemmer

## 1.1 IDENTIFICATION OF MICROORGANISMS IN ACTIVATED SLUDGE AND BIOFILMS

Until very recently, culture-dependent methods such as plate count or Most-Probable Number (MPN) counting have widely been used for enumeration and detection of bacteria being relevant to biological wastewater treatment performance. In fact, such standard methods are in many cases still used for effluent quality control, particularly with respect to pathogens and various indicator organisms (e.g. APHA Standard Methods). However, today we know that these methods suffer from severe limitations as from all types of microbes in environmental samples (also pathogens), only a very small fraction is cultivable on media generally applied. Therefore, this approach is prone to lead to serious misinformation. Thus, we strongly advocate for a change to using culture-independent molecular methods in all sorts of microbiological investigations in wastewater treatment plants (WWTPs).

Among the cultivation-independent methods for detection, fluorescence in situ hybridization (FISH) with ribosomal RNA (rRNA)-targeted probes (gene probes) is a very powerful tool for identification of microorganisms in activated sludge and biofilm biocenoses from WWTPs. This method is described in detail in Chapters 7 and 8 of this book. Most known functional key microorganisms in wastewater systems can be reliably identified and quantified by this method.

Furthermore, other molecular methods exist, primarily PCR-based methods. Quantitative PCR (q-PCR) is now getting more commonly applied in environmental samples, but the method has several drawbacks compared to FISH. This is due to biases concerning nucleic acid extraction, the PCR reaction, and also the fact that PCR-based approaches do not quantify microbial cells, but measure copy numbers of marker

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

genes. DNA microarrays carrying rRNA-targeted probes, so-called ''phylochips'', have a great potential as high-throughput tools for the qualitative detection of hundreds or even thousands of different uncultured microbes in only one experiment. When combined with autoradiography, phylochips become ''isotope arrays'' useful to track functional traits of microbes such as nitrifying bacteria in WWTP (Adamczyk et al., 2003). However, to date no quantitative phylochip-based assay exists that would allow for observing shifts in the abundances of probe-target microbial populations. This limitation is due to technical problems, such as saturation effects during hybridization, with the microarrays which are very difficult to overcome.

In contrast to the other methods, by using FISH it is possible to observe the morphology and to quantify numbers of bacteria or the equivalent biovolume. Thus, in our opinion, FISH is for the time being the method of choice for detection and quantification of microorganisms in WWTP as detailed in Chapters 7 and 8.

A special case is the identification of filamentous bacteria. These have primarily been identified based on their morphology and simple staining techniques using light microscopy since their first comprehensive description by Eikelboom (Eikelboom, 1975). Several manuals have since been published (e.g. Eikelboom 2000; Jenkins et al., 2004), all of them being based on his original work. However, today it is clear that although some filamentous bacteria can be fairly reliably identified in this way, the majority can not. As described in Chapter 5, we strongly recommend to also apply FISH for the identification of filamentous microorganisms, after having accomplished a preliminary morphological identification using the manuals.

## 1.2 THE MICROBIOLOGY OF BIOLOGICAL WASTEWATER TREATMENT

Biological treatment of municipal and industrial wastewater worldwide is primarily carried out by the activated sludge (AS) process. New technologies are being developed such as biofilm reactors, membrane bioreactors, sequencing batch reactors, etc., but they basically all derive from the AS process. The common purpose of all these technologies is the use of microorganisms to remove carbon (C), nitrogen (N), phosphorus (P), micropollutants and pathogens.

New interesting more sustainable solutions are appearing. They include for example recovery of nutrients (e.g. P) from wastewater, or conversion of organic waste components to usable, valuable compounds such as bioplastics (polyhydroxyalkanoates, PHA). Conversion of organic waste to energy by methane production during anaerobic digestion has been utilized for decades and these processes are further being developed together with other energy yielding processes such as microbial fuel cells.

Management of these complex microbial systems (or 'microbial resource management' for the new sustainable solutions, Verstraete et al., 2007) relies on a fundamental knowledge about the microbial populations being involved and about the factors that regulate their activity. A reliable identification of the microorganisms involved is fundamental and with the today's toolbox of various culture-independent methods is possible with a high sensitivity and precision. Not only the identity, but also knowledge about their ecophysiology, ecology, and population dynamics is essential. The present methodological approaches range from single cell microbiology (e.g. microautoradiography and FISH-Raman microspectroscopy; Huang et al., 2007), expression of specific functional genes, systems biology (genomics, transcriptomics, and proteomics) to lab-scale reactors and full-scale studies of chemical transformations. In this way we are gaining a rapidly increasing understanding of key microorganisms being involved in many processes and how to affect their presence and activity. However, there is still

#### Introduction 3

much to learn about full-scale systems since most studies so far have been carried out in lab-scale and pilot-scale reactors.

Several functional groups of bacteria being involved in the most common treatment processes are now fairly well identified and described. It is primarily bacteria involved in nitrification and to some extent those involved in denitrification, many bacteria involved in the enhanced biological P-removal (EBPR), and most bacteria causing settling problems (bulking) or foam/scum formation. In each functional group, for example the nitrifiers, a limited number of phylogenetic lineages (510) is encountered in nitrifying plants in general with only a few dominant populations (3–5) being present in a particular plant within the majority of full-scale plants. We should try to avoid the term ''species'', because a concise species definition is lacking in microbial ecology. Lineages, strains or ecotypes might be equally important for WWTP functioning as ''species''. The exact microbial community composition in a particular plant depends on wastewater composition, process design, and plant operation, see below. However, in common for most functional groups, the controlling factors determining the community composition is still poorly understood.

Does it matter which bacteria are present in each functional group in a certain treatment plant? This question can in some cases be answered with a clear 'yes', in others 'perhaps' or 'we don't know'. For the filamentous bacteria it is a clear 'yes'. Certain ecotypes cause severe settling properties, others are (in low number) important for sludge flocs as a backbone and thus for the floc structure. A proper identification is essential for the selection of efficient control measures towards the unwanted filamentous bacteria causing settling problems like bulking or foam.

It is more uncertain how important the knowledge about the exact community composition is for the nitrification performance in a certain activated sludge treatment plant, for example. Based on in situ observations it has been suggested that the presence of several lineages of ammonia and nitrite oxidizers ensures a more robust and stable system (e.g. Daims et al., 2001c) compared to the presence of only a single lineage from each functional group. But it is less documented what the exact community composition of a certain functional group means for plant stability and operation. Future studies, all based on a reliable identification of the microbial populations, for example by FISH, will show. A related question is whether strain-level microbial diversity influences process stability in WWT. For instance, very closely related nitrite oxidizers, which are slightly different on the genome level, may co-exist in the same plant. Due to a high sequence similarity (or even identity) of their 16S rRNA, the diversity of these strains would easily be overlooked by current approaches using methods such as FISH. However, the genomic differences may result in a biologically significant variety of phenotypes, which respond differently to events such as changes in wastewater composition or bacteriophage attack. With the latest molecular methods such as environmental genomics and deep DNA sequencing at hand, future research will elucidate the importance of such microbial microdiversity from an applied perspective.

Only a few comprehensive studies have been published describing the total community composition from full-scale wastewater treatment plants. They deal with activated sludge plants, such as industrial plants for C-removal or N-removal (Juretschko et al., 2002) or a plant with biological N- and P-removal from a mixture of domestic and industrial wastewater (Kong et al., 2007). However, several studies of specific populations have been carried out in various full-scale plants, referring solely to nitrifiers, denitrifiers or filamentous bacteria, for example. These studies are briefly described in the specific chapters. A summary of the most commonly observed species and genera encountered in WWTPs is shown in Table 1.1.


Table 1.1. Commonly reported microorganisms in wastewater treatment systems.

## 1.3 FACTORS OF IMPORTANCE FOR THE GROWTH OF MICROORGANISMS

The species composition in treatment plants depends on wastewater composition, process design, and plant operation. It is obvious that certain functional groups are dominant only if specific processes are included in the plant's process design (e.g. N-removal or EBPR). As mentioned above, the controlling factors

#### Introduction 5

determining the species composition are still poorly understood for many species, so during studies of the microbiology in WWTPs it is important to observe and register the potential factors that may be decisive for the presence of the different species. An overview of such factors is given in Table 1.2 and more advice on these is given elsewhere (e.g. Wilderer et al., 2002).


Table 1.2. Overview of important factors determining the microbial population structure in WWTPs.

Most important for activated sludge plants is whether solely C-removal is included in the plant design – meaning that only aerobic tanks are present (besides clarifier), or whether denitrification/EBPR are also included, meaning that anoxic/anaerobic tanks are present in addition to the aerobic ones. The selective pressure due to anoxic/anaerobic tanks substantially changes the population structure. Sludge age (mean cell residence time), which is determined by the sludge loading, is also extremely important. A low sludge

age (55–10 days) may not allow nitrification to occur due to the low growth rates of the nitrifiers, whereas a high sludge age (420–30 days) is important to obtain full N- and P-removal in temperate climates. Treatment plants running at very high temperatures (440*–*C), such as those treating special industrial wastewater, often select for unusual microbial communities.

The composition of the incoming wastewater is another decisive factor for bacterial growth. Industrial wastewater is often less complex than municipal wastewater, meaning fewer microorganisms may dominate in the treatment plants. Furthermore, the soluble fraction is often higher. Industrial wastewater may also not include important nutrients, such as P or N or other micronutrients. Domestic wastewater is usually more complex with a high fraction of particulates and a more balanced ratio of organics and nutrients. Other important factors are temperature, salinity, presence of toxic substances, and pH value. Likewise, the incoming microorganisms may affect the population structure in the treatment plant.

The operation of the plant may also affect the population structure. It is closely interrelated to the process design of the plant. The exact operation of anoxic/anaerobic mean cell residence time, oxygen concentration in aerobic tanks, addition of chemicals, carbon sources, and many other factors may affect the population composition.

The importance of the technology platform being applied to carry out a specific process, such as nitrification, is also poorly understood. Do we get the same nitrifiers in a full-scale plant based on activated sludge with different process configurations (e.g. sequencing batch reactors, continuously stirred reactors or plug-flow reactors), biofilm reactors (e.g. upflow or downflow biofilters, airlift reactors, granules reactors), or membrane bioreactors treating the same wastewater? Few studies have investigated this in detail, but the general impression is that we are often dealing with the same species/groups, although perhaps with slightly different strains/ecotypes (see also above). More studies based on a reliable identification of the microbial populations are needed.

## 1.4 THE USE OF THIS FISH HANDBOOK

This handbook contains a detailed description of the FISH protocol for identification and quantification of various bacteria typically encountered in biological wastewater treatment. The bacteria included cover several functional groups: nitrifiers, denitrifiers, polyphosphate-accumulating organisms (PAOs), glycogen-accumulating organisms (GAOs), filamentous bacteria involved in bulking or foaming, and some others. They can be found in aerobic or aerobic/anoxic/anaerobic treatment systems based on activated sludge or biofilms applied in various technologies. We have not included bacteria present in digesters (anaerobic digestion), fuel cells or bacteria carrying out more rarely encountered treatment processes such as treatment of S-containing waste (for S<sup>o</sup> production) or removal of specific pollutants (e.g. from polluted sites). Some of these groups may be included in future editions of this book.

The handbook does not cover protozoa as the molecular methods are still not ready for a proper detection of these.

Detailed information about the ecophysiology, ecology or management of the different bacteria is outside the scope of this handbook. Such information can be found in the specific literature, manuals, and books (see the different chapters) or in the new IWA book on ''Microbial Ecology of Activated Sludge'' (Eds. Seviour and Nielsen, 2009).

For each functional group an overview of the identities of the bacteria and their abundances is presented. Extensive tables describe the gene probes to be applied for the detection of the microbes, and phylogenetic trees show the coverage of the various probes. Furthermore, we have included a description of the typical morphologies targeted by the probes and many of these are documented by FISH images in the color figure section (Chapter 9). Based on our own experience we recommend the most suitable

#### Introduction 7

probes. Most probes are selected on the following criteria: they were designed based on published fulllength sequences, the probes are published, and details can be viewed in probeBase (www.microbialecology.net/probebase). Other probes are briefly mentioned in the text. They might be relevant in special cases.

New or improved gene probes for detection of relevant microbes are continually being developed, so we plan ongoing revisions of this book, the first in 1–2 years. Information about new editions and other relevant updates can be found on the web page of the IWA specialist groups (www.iwahq.org). This handbook has been developed in conjunction with the IWA specialist group on 'Activated Sludge Population Dynamics'.

Errors, suggestions for inclusion of other probes, important experiences or other comments are more than welcome as they can be beneficially included in future editions. Please e-mail Per Halkjær Nielsen (phn@bio.aau.dk).

## 2

## The nitrifying microbes: Ammonia oxidizers, nitrite oxidizers, and anaerobic ammonium oxidizers

## Holger Daims, Frank Maixner, and Markus C. Schmid

## 2.1 INTRODUCTION

Nitrification is a key process of nitrogen removal in most municipal and industrial wastewater treatment plants (WWTPs). The term ''nitrification'' refers to the aerobic, sequential oxidation of ammonia to nitrite and of nitrite to nitrate. These two steps are catalyzed by specialized chemolithoautotrophic prokaryotes: ammonia-oxidizing bacteria (AOB), ammonia-oxidizing archaea (AOA), and nitrite-oxidizing bacteria (NOB). To date, no organism is known to be able to catalyze both autotrophic nitrification steps.

Anaerobic ammonium oxidation (the ''anammox'' process) was described as a shortcut in the biogeochemical N-cycle, where ammonium is oxidized, nitrite is reduced, and N2 gas is formed (van de Graaf et al., 1995). As anammox is catalyzed by strictly anaerobic and autotrophic bacteria, significant costs for aeration and carbon sources can be saved if the process is exploited for sewage treatment. Accordingly, a number of novel approaches for N-removal from wastewater combine partial nitrification (ammonia oxidation by AOB) with anammox (Third et al., 2001; van Dongen et al., 2001; Pynaert et al., 2003; van der Star et al., 2007).

All nitrifying microbes and anammox bacteria are very slow-growing organisms, which are sensitive to changes of their growth conditions. As a consequence, nitrification has often caused trouble in wastewater treatment due to unpredictable performance breakdowns. The causes of such events remain obscure in

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

many cases. Technical applications of anammox are young, and field reports are still rare, but the extremely slow growth of anammox bacteria may also become a problem when their populations decline for unknown reasons and the process fails in a full-scale WWTP. Therefore, fast methods to detect and quantify the involved organisms in activated sludge or biofilm are important tools not only for fundamental research, but could become relevant in the future also for the routine operation of nitrifying and anammox-based reactors: If a decline of any of the key populations is detected soon enough, operational countermeasures may be able to prevent a worst-case scenario. FISH certainly is the best available technique for this purpose, and a large set of rRNA-targeted probes is already available for the detection of AOB, NOB, and anammox bacteria. This chapter provides an overview of the most common nitrifiers and anammox organisms in WWTPs and of the FISH probes frequently used to detect and to quantify these microbes.

## 2.2 AMMONIA OXIDIZERS

In most nitrifying WWTPs, ammonia is oxidized by AOB of the genus Nitrosomonas (including Nitrosococcus mobilis). Most commonly found are AOB related to N. europaea, N. eutropha, N. mobilis, and N. oligotropha (Figure 2.1). Experiments using lab-scale reactors and studies of pure cultures suggested that N. oligotropha and closely related AOB are better adapted to low ammonia concentrations than N. europaea (Bollmann et al., 2002, and references therein; Limpiyakorn et al., 2007). However, detailed studies have been carried out only with a limited number of AOB strains. Furthermore, Lydmark et al. (2007) observed differential responses of N. oligotropha-related populations to ammonia shifts in a pilotscale plant. Local differences in substrate concentrations (microniches) within flocs or biofilms may also affect the distribution and activity of AOB in WWTPs (Gieseke et al., 2005). Nitrosococcus mobilis-related AOB seem to occur especially in reactors treating sludge liquor or other types of wastewater with elevated ammonia and salt concentrations such as animal rendering waste (Juretschko et al., 1998).

In activated sludge flocs and biofilms, AOB related to Nitrosomonas usually form almost spherical, compact cell aggregates (Figure 9.1, Chapter 9). Single cells within these clusters are well visible at 630*£* or 1000*£* magnification. The diameter of most AOB cell clusters is 10–50 mm. Less compact aggregates occur occasionally, where the cells are more irregularly arranged with more space between them. Single AOB cells are seldom found in activated sludge by microscopy, but they may easily be overlooked when dense floc or biofilm structures are observed.

AOB of the genus Nitrosospira have occasionally been detected in WWTPs, but these AOB are generally more common in terrestrial habitats and seem to play only minor roles for wastewater treatment. Notable exceptions are rhizoremediation plants, where Nitrosospira-related organisms seem to be more frequent (Haleem et al., 2000).

Both genera Nitrosomonas and Nitrosospira belong to the Betaproteobacteria. In marine habitats, Nitrosomonas marina and the gammaproteobacterial lineage of AOB with the two species Nitrosococcus oceani and N. halophilus are common. Due to their high salt requirements, these AOB are absent from most WWTPs with the possible exception of brackish water and saltwater systems: N. marina was found in a nitrifying trickling filter biofilm associated with a marine aquaculture system (Foesel et al., 2008). Occurrence of N. oceani and N. halophilus in WWTPs has not been reported yet.

The question of whether AOA might be important for nitrification in WWTPs still is a matter of active research and dispute. Indeed, AOA have been detected in activated sludges by means of specific polymerase chain reaction (PCR) (Park et al., 2006). However, PCR does not allow to distinguish intact, active cells from naked DNA or metabolically inactive cells that are occasionally swept into the activated

Figure 2.1. 16S rRNA-based phylogenetic tree showing main lineages of known bacterial AOB. Brackets indicate the coverage of the probes listed in Table 2.1. Dashed brackets indicate incomplete sequence information at the respective probe target site, so that precise coverage information cannot be inferred. The exact branching order of the AOB lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

sludge systems but have no relevant functions there. So far, high abundance of AOA in nitrifying WWTPs has not been confirmed by more robust methods such as quantitative FISH with rRNA-targeted probes.

## 2.2.1 Probes for the detection of AOB

The oligonucleotide probes, which have most frequently been used to detect and quantify AOB in WWTPs, are listed in Table 2.1. The tree in Figure 2.1 illustrates the specificity and coverage of the various probes. Please note that this compilation of probes is not complete (i.e. additional AOB-targeted probes have been published). However, the probes listed here have been tested in a large number of studies by different research groups. Their coverage and specificity have been re-evaluated, based on increasingly large rRNA sequence databases, since the probes were originally published.


Table 2.1. rRNA-targeted oligonucleotide probes used to detect AOB in nitrifying activated sludge and biofilm samples.

1) FA ¼ Formamide.

2) To ensure probe specificity, the unlabeled competitor must be used in equimolar amounts together with the fluorescently labeled probe.

3) NEU can also be used with 35% formamide.

If the aim is to detect and quantify most of the AOB in an activated sludge sample at once, it is convenient to use a probe mixture consisting of the probes Nso1225 þ NEU þ NmV þ Cluster6a192 in equimolar amounts (and also the respective competitor oligonucleotides). All probes except the competitors should then be labeled with the same fluorochrome. All the probes in this ''AOB-mix'' can be used with 35% formamide in the hybridization buffer. In order to differentiate the AOB in a sample, the same ''AOB-mix'' can be used, but then with probes labeled with different fluorochromes. In addition, one could apply the other probes (Table 2.1) using the respective formamide concentrations.

For unknown reasons, probe Nso1225 occasionally yields only dim fluorescence signals even though the target AOB are present in the sample and their cellular ribosome content would be high enough for a bright fluorescence signal after FISH. Additional test hybridizations with probe Nso190 should be carried out to ensure that no AOB are overlooked in a sample. The coverage of Nso190 resembles that of Nso1225, but Nso190 is more reliable in terms of probe brightness. However, due to its requirement for 55% formamide, Nso190 cannot be mixed with the other probes and has to be applied separately.

Please note that probe Nsv443 seldom yields unambiguously positive signals when applied to activated sludge. In some cases a few cells are stained by this probe in each field of view, but their density seems too low for technical relevance. Whether such cells are Nitrosospira spp. or unspecifically detected nontarget organisms remains to be shown. As mentioned already, higher densities of Nitrosospira spp. have

not frequently been observed in full-scale wastewater treatment plants. If cells are detected after FISH with Nsv443 in activated sludge, it is recommended that additional control experiments should be performed to check for autofluorescent cells, or unspecific binding of DNA or of the fluorochromes to cell surfaces. This can be accomplished by (i) using a nonsense probe, which cannot target any bacterium; (ii) using the same probe labeled with a different fluorochrome; or (iii) adding large amounts of DNA (e.g. salmon sperm DNA) to the sample prior to FISH in order to saturate any unspecific DNA binding sites on cell surfaces. Furthermore, a hierarchy of phylogenetically nested probes should be used to confirm the identity of probe-positive cells. In case of Nitrosospira this could be a combination of Nsv443 with probes Nso1225 or Nso190, respectively, which should be labeled with a different fluorochrome than Nsv443.

## 2.3 NITRITE OXIDIZERS

In the majority of nitrifying WWTPs, the dominant NOB are members of the genus Nitrospira (e.g. Juretschko et al., 1998; Schramm et al., 1998; Daims et al., 2001a; Gieseke et al., 2003). This genus is part of the distinct bacterial phylum Nitrospirae, and thus is not closely related to the other known NOB, which are all Proteobacteria (Figure 2.2). All Nitrospira are slow-growing and recalcitrant bacteria, which are very difficult to culture in the laboratory. Only three isolates have been obtained from other sources than activated sludge, whereas no Nitrospira has been isolated from any WWTP although one high enrichment was achieved (Spieck et al., 2006). Therefore, the discovery that Nitrospira (not Nitrobacter) are the key NOB in full-scale wastewater treatment plants (Juretschko et al., 1998) was based on the use of FISH and other cultivation-independent methods. Nitrospira is a diverse genus that consists of several phylogenetic sublineages (Figure 2.2). In WWTPs, the sublineages I, II and IV are found, but in many systems sublineage I Nitrospira are predominant. Evidence exists that sublineage I can outcompete sublineage II Nitrospira at (transiently) elevated nitrite concentrations (Maixner et al., 2006). Sublineage IV, which occurs in nature in marine habitats, is limited to WWTPs receiving sewage with high salt concentrations.

All known Nitrospira in WWTPs form spherical or irregularly shaped cell aggregates, which consist of several hundred or thousand cells (Figure 9.2, Chapter 9). The diameter of these aggregates is 10–100 mm, but even larger clusters are found occasionally. In particular the large Nitrospira aggregates often contain narrow ''channels'' and larger cavities (Figure 9.2, Chapter 9).

In contrast to Nitrospira, the genus Nitrobacter seems to play a minor role in wastewater treatment. Nitrobacter cells occur in many reactors and can be enriched or isolated by incubation of activated sludge in artificial nitrite media, but FISH has shown that Nitrobacter do usually not reach significant cell densities in WWTPs (Wagner et al., 1996). Notable exceptions are reactors that temporarily contain elevated nitrite concentrations, for example sequencing batch reactors treating highly concentrated sludge liquor. In such systems, Nitrobacter can also reach a relatively high abundance (Daims et al., 2001a), probably because these NOB are adapted to higher NO <sup>2</sup> concentrations whereas Nitrospira are adapted to lower NO <sup>2</sup> concentrations (Schramm et al., 1999). Nitrobacter in WWTPs form cell aggregates like the other nitrifiers, or they occur as single cells embedded in the biofilm matrix.

In flocs and biofilms, NOB often occur in the direct spatial neighborhood to AOB, which reflects the mutualistic symbiosis of these two functional groups (Maixner et al., 2006) (Figure 9.3, Chapter 9).

Other NOB comprise the marine genera Nitrococcus and Nitrospina and the only recently discovered betaproteobacterial Nitrotoga arctica (Alawi et al., 2007). To date, none of these NOB has been shown, by FISH or any other quantitative cultivation-independent assay, to be functionally important for nitrification in wastewater treatment.

Figure 2.2. 16S rRNA-based phylogenetic tree showing main lineages of known bacterial NOB. Brackets indicate the coverage of the probes listed in Table 2.2. Dashed brackets indicate incomplete coverage of the target groups. Sublineages I–IV comprise the genus Nitrospira. (-) Probe NTG840 may also bind to some nontarget Betaproteobacteria, which have no 16S rRNA sequence mismatch at the probe binding site.

## 2.3.1 Probes for the detection of NOB

The oligonucleotide probes, which have most frequently been used to detect and quantify NOB in WWTPs, are listed in Table 2.2. The tree in Figure 2.2 illustrates the specificity and coverage of the various probes. Please note that this compilation of probes is not complete (i.e. additional NOB-targeted probes have been published). However, most of the probes listed here have been tested in a large number of studies by different research groups. Their coverage and specificity have been re-evaluated, based on increasingly large rRNA sequence databases, since the probes were originally published.

In most cases, probes Ntspa662 targeting the genus Nitrospira and NIT3 targeting Nitrobacter are sufficient to detect NOB in WWTPs. Probe Ntspa662 can be combined with Ntspa712, which also targets all Nitrospira, or with Ntspa1431, Ntspa1151 and Nspmar62, which target different sublineages of this genus. The latter combinations are especially interesting if the probes are labeled with different fluorescent dyes, so that one can determine whether different Nitrospira co-exist in the same WWTP.

Probes Ntspa1026 and Nsr1156 usually yield excellent signal intensities with targeted Nitrospira, but neither probe completely covers the respective Nitrospira sublineages that are relevant for wastewater treatment. Therefore, the use of these probes is recommended only in combination with either Ntspa662 or Ntspa712 to increase the likelihood that all Nitrospira in a sample are detected.

## 2.4 ANAMMOX BACTERIA

All known anaerobic ammonium oxidizers belong to the phylum Planctomycetes (Strous et al., 1999). They are peculiar prokaryotes in many respects, but their most distinct feature is intracellular


Table 2.2. rRNA-targeted oligonucleotide probes used to detect NOB in nitrifying activated sludge and biofilm samples.

1) FA ¼ Formamide.

2) To ensure probe specificity, the unlabeled competitor must be used in equimolar amounts together with the fluorescently labeled probe.

3) Ntspa712 can also be used with 35% formamide, especially if combined with Ntspa662. 4) Ntspa1026 does not cover all members of these Nitrospira sublineages. 5) NTG840 has not yet been used with activated sludge. Nevertheless it is listed here, because Nitrotoga-like bacteria have been enriched from a WWTP (Alawi et al., 2007) and NTG840 is the only published probe targeting this organism. Please note that NTG840 is not fully specific as some nontarget bacteria have no 16S rRNA sequence mismatches at the probe binding site.

compartmentalization, which is very uncommon in bacteria (Lindsay et al., 2001). Since the discovery of anammox bacteria about a decade ago, several different representatives have been enriched from WWTPs. These are the lineages Brocadia with the candidate species ''B. anammoxidans'' (Strous et al., 1999) and ''B. fulgida'' (Kartal et al., 2008), Kuenenia with ''K. stuttgartiensis'' (Schmid et al., 2000), Scalindua with ''S. brodae'' and ''S. wagneri'' (Schmid et al., 2003), and Anammoxoglobus with ''A. propionicus'' (Kartal et al., 2007). The best studied anammox organism so far is K. stuttgartiensis whose genome was sequenced, by environmental genomics, and was used to reconstruct the intricate anammox biochemical pathway (Strous et al., 2006). Interestingly, A. propionicus co-oxidizes propionate and ammonium (Kartal et al., 2007), whereas B. fulgida co-oxidizes acetate and ammonium more efficiently than the other anammox bacteria (Kartal et al., 2008). These physiological differences suggest adaptations of the various anammox organisms to specific ecological niches, which could affect the anammox community composition in WWTPs.


Table 2.3. rRNA-targeted oligonucleotide probes used to detect anammox organisms in activated sludge and biofilm samples.

1) FA ¼ Formamide.

Figure 2.3. 16S rRNA-based phylogenetic tree showing main lineages of known anammox organisms. Brackets indicate the coverage of the probes listed in Table 2.3.

Usually, anammox bacteria occur as small to large cell aggregates (Figure 9.4, Chapter 9). Anammox bacteria have a coccoid morphology and a 'donut-like' shape after FISH with an apparent small hole in the centre of the cells. This 'hole' is caused by the absence of ribosomes (and thus, of target molecules for FISH probes) in this region. Hence, in combination with suitable FISH probes, anammox organisms are easy to identify by microscopy. However, one should keep in mind that other planctomycetes, which are morphologically similar, also occur in WWTPs (Neef et al., 1998). These planctomycetes probably are heterotrophs not involved in the anammox process, and they are also labeled by general planctomycetesspecific FISH probes such as Pla46. Thus, anammox-specific FISH probes (Schmid et al., 2005) should always be used to confirm that donut-shaped cells resembling planctomycetes really are anammox bacteria.

## 2.4.1 Probes for the detection of anammox organisms

The oligonucleotide probes, which are best for detecting and quantifying anammox bacteria in WWTPs, are listed in Table 2.3. The tree in Figure 2.3 illustrates the specificity and coverage of the various probes. Note that some additional probes targeting anammox bacteria have been published (a compilation can be found in Schmid et al., 2005). However, these other probes do not offer a better coverage and specificity than the probes listed in Table 2.3 for anammox organisms occurring in full-scale WWTPs.

3

## Identification of denitrifying microorganisms in activated sludge by FISH

## Jeppe Lund Nielsen and Aviaja A. Hansen

## 3.1 INTRODUCTION

Denitrification is defined as the anoxic process in which nitrate or nitrite is reduced to gaseous nitrogen oxides (nitrogen oxide, nitrous oxide or free nitrogen). The process has been the focus of numerous studies due to its major importance in the global N-cycle, for nitrogen removal in the wastewater industry, and lately also for potential production of the greenhouse gas N2O. Although subject for many research studies, relatively little is known about the microorganisms carrying out the process in full-scale WWTPs and in particular their phylogenetic distribution (Juretschko et al., 2002; Wagner and Loy, 2002). However, a few recent studies have revealed some important denitrifiers in full-scale WWTPs (Thomsen et al., 2007; Hagmann et al., 2008; Morgan-Sagastume et al., 2008).

The denitrifying organisms are facultative anaerobic heterotrophs and, less frequently, autotrophs. The interaction of the denitrification process and the nitrification process is the main cause for removal of nitrogen from wastewater. While nitrification primarily occurs under aerobic conditions, denitrification requires anoxic conditions with the presence of nitrate or nitrite and easily degradable organic substrates. Under oxic conditions most denitrifiers carry out full aerobic respiration. These facultative denitrifiers seem to constitute a major fraction of the whole of bacteria in the activated sludge community (Nielsen and Nielsen, 2002a; Morgan-Sagastume et al., 2008). Investigations on the denitrification rates in the presence of various electron donors reveal the presence of several specialized populations of denitrifiers in activated sludge (Thomsen et al., 2007; Hagmann et al., 2008; Morgan-Sagastume et al., 2008). Some

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

microorganisms are only partially involved in the denitrification process for example by reducing nitrate and/or nitrite to nitric oxide or nitrous oxide but these are not well described.

## 3.2 IDENTITY OF DENITRIFIERS IN WASTEWATER TREATMENT SYSTEMS

The organisms capable of denitrification are not limited to specific microbial taxa but present in many phylogenetic subgroups. Denitrifiers are found among the Proteobacteria, Firmicutes, and the Bacteroidetes covering more than 50 genera (Zumft, 1992). Previously, culture-dependent studies of denitrifiers typically described these to be members of Pseudomonas, Bacillus or Alcaligenes, but these are usually not the dominant genera in the activated sludge process. Several publications based on both culture-dependent and culture-independent approaches indicate that other denitrifiers must be dominating the activated sludge system (e.g. Heylen et al., 2006a,b; Hagman et al., 2008).

Culture-independent studies carried out directly in a WWTP treating industrial wastewater (e.g. Juretschko et al., 2002; Wagner and Loy, 2002) have shown the betaproteobacterial genus Azoarcus (Rhodocyclaceae) as the dominating denitrifier, while similar studies in treatment plants treating municipal wastewater have revealed that several betaproteobacterial genera belonging to the families Comamonadaceae and Rhodocyclaceae are present. Within the Comamonadaceae, Curvibacter-related organisms (reclassified from Aquaspirillum, Ding and Yokota, 2004) and within the Rhodocyclaceae members of the genera Zoogloea, Azoarcus, and Thauera are the dominating denitrifiers (Rossello`-Mora et al., 1995; Juretschko et al., 2002; Thomsen et al., 2004; 2007). In WWTPs with Enhanced Biological Phosphorus Removal (EBPR) with combined nitrogen removal, polyphosphate accumulating members of Accumulibacter (Betaproteobacteria) are present and some of these, but presumably not all, are capable of denitrification (e.g. Kong et al., 2004). The identity of these subgroups of denitrifying Accumulibacter is not well known and it has hitherto not been possible to phylogenetically distinguish these by application of gene probes or by other molecular methods (Seviour et al., 2003). These groups of Betaproteobacteria constitute the vast majority of denitrifying microorganisms in most WWTPs (Thomsen et al., 2007), although several other, less abundant unclassified denitrifiers also are present (Morgan-Sagastume et al., 2008). Culture-dependent studies have shown that denitrifying organisms belonging to the alphaproteobacterial genera Paracoccus (Rhodobacteraceae) and Rhizobium (Rhizobiaceae) are present in WWTPs (Heylen et al., 2006a,b). However, due to the lack of appropriate probes the abundance of these denitrifying Alphaproteobacteria has not been further evaluated.

Culture-dependent studies of wastewater treatment systems fed with methanol under denitrifying conditions frequently identify Alphaproteobacteria belonging to the genus Hyphomicrobium (Hyphomicrobiaceae) (e.g. Timmermans and Van Haute 1983), which have been further verified by dot blot analysis on the activated sludge (Layton et al., 2000).

Studies carried out in denitrifying lab-scale reactors inoculated with activated sludge and run on acetate or other single substrates reveal high diversities of potential denitrifying organisms belonging to Proteobacteria, Bacteroidetes, Chloroflexi, Firmicutes, and Actinobacteria (Yosie et al., 2001; Ginige et al., 2004; Osaka et al., 2006). This may not, however, reflect the indigenous consortia abundant in full-scale WWTPs.

## 3.3 ABUNDANT DENITRIFIERS IN FULL-SCALE PLANTS

#### Family Comamonadaceae:

Genus Curvibacter. Curvibacter has been found to be a potential dominating denitrifier in full-scale plants without external carbon addition with abundances of 11–29% of all Bacteria (Thomsen et al., 2004).

Its morphology is quite distinct and easy to recognize. It is typically large and coccoid with a diameter of 1–1.5 mm and grows as single cells or in microcolonies typically ranging from 10–15 mm in diameter, see Figure 9.5, Chapter 9.

#### Family Rhodocyclaceae:

Genus Azoarcus. The abundance of Azoarcus in some full-scale plants varies from 3–16% of the total biomass (Thomsen et al., 2007), but in plants with addition of an external carbon source in the form of methanol more than 30% have been found (Hagman et al., 2008). Different morphotypes have been observed, but usually they appear rod-shaped (0.5–1 mm *£* 1–2 mm) and grow as single cells or in relatively small microcolonies (5–20 mm in diameter), see Figure 9.6, Chapter 9.

Genus Thauera. The number of Thauera-related bacteria ranges from 2–11% of the total biomass in some full-scale plants (Thomsen et al., 2007). The cells usually grow in microcolonies of various sizes (5–40 mm), and the rod-shaped cell size is typically 1.0 *£* 2.5 mm, see Figure 9.7, Chapter 9.

Genus Zoogloea. Zoogloea-related bacteria are typically observed in small numbers except for a few plants in which up to 10% of the total biomass has been reported (Rossello`-Mora et al., 1995). Zoogloearelated bacteria have often been described as forming colonies in typical branched gelatinous matrices, the so-called Zoogloea fingers. Most cells are rod-shaped (1–1.5 mm *£* 1 mm).

Genus Accumulibacter. The presence of the potential PAO Accumulibacter usually indicates the presence of an EBPR configuration with abundances ranging from 3–15% of the total biomass. The cells are typically rod-shaped (1–1.5 mm) and almost always growing in microcolonies (Larsen et al., 2006). See also Chapter 4 about PAOs.

Other denitrifiers, such as Acidovorax, are usually observed but only in small numbers, typically less than 1–2% of the total biomass (Heylen et al., 2008). Glycogen-accumulating organisms (e.g. Competibacter) contain members that are capable of denitrification (Zeng et al., 2003), see Chapter 4.

#### Other potential denitrifiers

Denitrifiers from several genera have been isolated from activated sludge. Figure 3.1 shows a selection of the phylogenetically identified isolates. The tree has also been supplemented with sequences from three denitrifying clone libraries, derived i) from stable isotope probing of denitrifiers grown in lab-scale reactors on acetate and methanol (Ginige et al., 2004, 2005; and Osaka et al., 2006) and ii) denitrifying pure cultures from lab-scale reactors seeded with wastewater (Etchebehere et al., 2001; Yosie et al., 2001; Martin et al., 2006; Wang et al., 2007). Recently, it was found that hitherto unidentified bacteria with a potential denitrifying capability (besides from the Betaprotobacteria) may be found within the groups of Alphaproteobacteria, Gammaproteobacteria, and Actinobacteria. The study revealed that their relative abundance in the activated sludge investigated was almost as important as the Betaproteobacteria (Morgan-Sagastume et al., 2008).

## 3.4 PROBES FOR DETECTION OF DENITRIFIERS

In Table 3.1 the oligonucleotide probes that can be used for identification and quantification of denitrifiers or potential denitrifiers are listed. In the phylogenetic tree in Figure 3.1 the specificities and coverage of the probes can be seen. At present it is not possible to target all denitrifiers, but for detection of the majority of the Betaproteobacteria it is recommended to use the following probes:

Of the Comamonadaceae family, most Acidovorax can be targeted by probe ACI208 (Amann et al., 1996), but also Variovorax are targeted. Some Curvibacter are targeted by Curvi997 with a high specificity (Thomsen et al., 2004).

Figure 3.1. 16S rRNA-based phylogenetic tree showing denitrifying bacteria from activated sludge. Brackets indicate the coverage of the probes listed in Table 3.1. Dashed brackets indicate incomplete coverage of the indicated groups. <sup>a</sup> Sequences from bioreactor studies, <sup>b</sup> Sequences from full-scale and bioreactor studies.


 Renamed from Aqs997

 Renamed from AZO644

Of the family Rhodocyclaceae, most Azoarcus can be targeted by probe AZA645 (Hess et al., 1997). Probe Thau646 (formerly probe MZ1) targets Thauera (Lajoie et al., 2000) but with a low specificity and with a high proportion of outgroup hits including Dechloromonas of the Rhodocyclaceae family, Comamonas of the Comamonadaceae family and Aquabacterium of the Incertae Sedis 5 (Figure 3.1). Most Azoarcus and Thauera can be detected by probe AT1458 (Rabus et al., 1999). The probe is very specific, apart from targeting most of the Castellaniella of the Alcaligenaceae family. ZRA (identical to ZRA23a) targets a high proportion of Zoogloea-related organisms (Rossello´-Mora et al., 1995), but not all known Zoogloea strains from WWTPs (Figure 3.1).

Accumulibacter-related PAOs can be targeted by the PAOmix (Crocetti et al., 2000), but it targets also a high proportion of Propionivibrio of the Rhodocyclaceae family.

Specific probes for the denitrifying Alphaproteobacteria and Gammaproteobacteria have yet to be developed and therefore the probes available have a low coverage of the denitrifiers in WWTPs (Figure 3.1). The following probes can be used to target Alphaproteobacteria:

PAR651 targets most Paracoccus of the family Rhodobacteraceae with a high specificity (Neef et al., 1996). The denitrifying Hyphomicrobium is specifically targeted by probe HyphoCII-654 (Layton et al., 2000) (see Figure 9.9, Chapter 9). G\_Rb is a broad probe targeting most Rhodobacteraceae and some Phyllobacteriaceae, see Figure 9.10, Chapter 9. (Giuliano et al., 1999; Eilers et al., 2000) and has therefore also many non-denitrifying organism targets.

The following probe can be used to target some potential denitrifiers within the Gammaproteobacteria: Pae997 targets Pseudomonas (Amann et al., 1996) with a high specificity, see Figure 9.11, Chapter 9.

## 4

## Identification of polyphosphate-accumulating and glycogen-accumulating organisms by FISH

Per Halkjær Nielsen, Hien Thi Thu Nguyen, Simon Jon McIlroy, Artur Tomasz Mielczarek, and Robert Seviour

## 4.1 INTRODUCTION

Enhanced biological phosphorus removal (EBPR) processes have been widely used to remove soluble phosphate from domestic wastewater. Consistent and reliable operation of these systems requires successful selection and enrichment of microorganisms that biologically accumulate Pi into granules of intracellular polyphosphate.

The microbial communities in EBPR processes comprise populations considered important for EBPR [i.e. polyphosphate-accumulating organisms (PAOs)], those competing with PAOs [i.e. glycogenaccumulating organisms (GAOs)], and yet-to-be identified populations whose functions and interactions with the PAOs and GAOs remain to be further understood. PAOs in full-scale systems are mainly represented by the betaproteobacterial Candidatus Accumulibacter phosphatis (hereafter called Accumulibacter) or related species that consist of members with and without an ability for denitrification. Proliferation of glycogen-accumulating organism populations can cause failure in EBPR plants as they compete with the PAOs for organic substrates during the anaerobic period without contributing to Pi removal (Oehmen et al., 2007). Originally they were referred to as the ''G bacteria'' (Cech and Hartman, 1990) and more recently as the tetrad-forming microorganisms (TFOs) morphotype, since some appear as

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

distinctive coccal-shaped cells growing in tetrads (Tsai and Liu, 2002). The term glycogen-accumulating organisms (GAOs) was first suggested by Mino et al. (1995) defining a phenotype that stores glycogen aerobically and consumes it anaerobically for energy production for the uptake of short-chain fatty acids for PHA production. However, not all GAOs appear in tetrads. It has been reported that in the absence of PAOs, phylogenetically different GAOs can compete with each other in lab-scale reactors. Less is known about their competition in full-scale plants.

A comprehensive set of rRNA-targeted oligonucleotide probes has been applied to identify in situ such populations in full-scale EBPR systems around the world including countries like Japan, USA, Denmark, and Australia. The role of Accumulibacter-related PAOs seems less important in full-scale EBPR processes compared to laboratory-scale EBPR systems fed with acetate as a sole carbon source. Thus, it is necessary to further identify other important PAOs besides Accumulibacter in EBPR processes and to study their dynamics/interaction with GAOs and their ecophysiological traits in full-scale processes under controlled operational conditions.

## 4.2 IDENTITY OF PAOs

Accumulibacter. The most abundant and important Gram negative identified PAOs in most full-scale EBPR wastewater treatment plants belong to the genus Accumulibacter in the family Rhodocyclaceae of subclass 2 of the Betaproteobacteria. None have been grown in pure culture. Accumulibacter exhibit in situ the phenotype expected from the biochemical models for the PAO. Thus, under anaerobic conditions short-chain fatty acids are assimilated and used for synthesis of PHA with Pi release. Subsequent aerobic PHA respiration provides energy for Pi assimilation and polyphosphate synthesis. Details of their physiology are given in two reviews (Seviour et al., 2003; Oehmen et al., 2007). Accumulibacter can become highly enriched in lab-scale reactors and are also generally common in most full-scale EBPR plants, with typical relative abundances reported between 3–15% of the total biomass expressed by the EUBmix. When the functional genes of polyphosphate kinase (ppk) are used as phylogenetic markers at least five clades of Accumulibacter exist in wastewater treatment plants and natural ecosystems (Peterson et al., 2008). However, 16S rRNA-targeted FISH probes do not differentiate between these clades because of the more conserved nature of this gene. Some, but not all Accumulibacter can denitrify, although currently the precise identity of these strains is not clear, and so it is not possible yet to recognize these with gene probes or other molecular methods.

Accumulibacter cells have a typical coccobacillus shape and are almost always observed growing in microcolonies (Figure 9.15 and 9.16, Chapter 9), although in some EBPR communities they exist predominantly as single cells.

Actinobacterial PAOs. The number of Gram positive Actinobacteria is often high in EBPR plants. Their functions there are largely unknown, although some of these may have a PAO phenotype. However, these do not take up short-chain fatty acids under anaerobic conditions, and their intracellular storage polymer is not PHA, but remains unidentified. Thus they do not behave according to the current biochemical models proposed for the PAO.

Two potential actinobacterial PAO morphotypes, both closely related to members of the genus Tetrasphaera in the family Intrasporangiaceae are abundant in many full-scale plants (Kong et al., 2005) as a) cocci in tetrads and b) short rods. Both morphotypes in situ grow as relatively small cells (0.3 to 0.5 mm thick and 0.8 to 1.0 mm long) and usually form medium-sized microcolonies (Figure 9.12, Chapter 9). In some cases the cocci do not exist as the expected tetrads of Tetrasphaera, but form branches

('tree-shaped'), see Figure 9.13, Chapter 9. Typical rod-shaped Tetrasphaera are shown in Figure 9.14, Chapter 9.

Other potential PAOs. Several other bacteria grown in pure culture have been proposed to be PAOs, but none possess the expected PAO phenotype and where FISH probes have been used, were not shown to be abundant in full-scale treatment plants. These include Acinetobacter (Fuchs and Chen, 1975; Wagner et al., 1994b), Microlunatus phosphovorus (Nakamura et al., 1995; Santos et al., 1999), Lampropedia spp. (Stante et al., 1997), and Tetrasphaera spp. (Maszenan et al., 2000). FISH probe-defined uncultured species of Dechloromonas (targeted by the probe Bet135) were proposed as putative PAOs on the basis of their phenotype. They are relatively abundant in some full-scale EBPR plants, but how important they are as PAOs is not known (Kong et al., 2007).


Table 4.1. 16S rRNA oligonuclotide probes for the identification of potential PAOs.

Figure 4.1. 16S rRNA-based phylogenetic tree showing main lineages of known or putative PAOs. Brackets indicate the coverage of the probes listed in Table 4.1.

## 4.2.1 Probes for detection of PAOs

Table 4.1 lists the oligonucleotides probes available for identifying and quantifying the PAOs or putative PAOs, and the phylogenetic tree in Figure. 4.1 illustrates the specificity and coverage of these. To quantify most Accumulibacter in a sample from a full-scale plant, the recommendation is to use the PAOmix probe set since it targets most currently known Accumulibacter-related 16S rRNA sequences. The probes of Zilles et al. (2002) embrace more sequences than the probes described by Crocetti et al. (2002) but in our experience there is no difference with full-scale plant samples compared to the results obtained when the PAOmix probe set is used.

Putative PAOs closely related to members of the genus Tetrasphaera can be targeted by probe Actino221 (primarily cocci in tetrads) and probe Actino658 (primarily short rods). They are also targeted by two broad-range actinobacterial probes HGC69a and actino1011. However, occasionally filamentous bacteria probably closely related to Candidatus Nostocoida limicola (a member of the genus Tetrasphaera) also fluoresce with these probes (see Chapter 5). As for most Gram positive bacteria, it may be important to apply enzymes (lysozyme 0.5 g/liter in 100 mM Tris [pH 7.5] and 5 mM EDTA) or mutanolysin (5,000 U/ml in phosphate buffer) to permeabilize cells to allow their passage to their target site, the ribosomes.

## 4.3 IDENTITY OF GAOs

Gammaproteobacterial GAOs. These bacteria exhibit the phenotype expected of a GAO. Namely the uptake of short-chain fatty acids and/or glucose with formation of PHA under anaerobic conditions and

subsequent synthesis of glycogen but not polyphosphate under aerobic conditions. None have been grown in pure culture. Details of their physiology can be found in reviews by Seviour et al. (2003) and Oehmen et al. (2007). The most numerically abundant identified GAOs in full-scale wastewater treatment plants belong to the genus Competibacter within the Gammaproteobacteria, which are also referred to as the GB-group. The most common of these are Candidatus Competibacter phosphatis, often abbreviated to Competibacter. Competibacter can become highly enriched in lab-scale reactors but reports suggest they may also be common in full-scale EBPR plants with typical relative abundances reported up to 10% of the total cell biovolume.

More than seven subgroups of Competibacter are currently described, based primarily on 16S rRNA gene sequences from lab-scale reactor communities. Most of these are present or often abundant in fullscale EBPR plant communities too, but only limited information exists about possible differences in their distribution and ecology. Some Competibacter may denitrify but whether it is possible to distinguish these with gene probes or other molecular methods is not known.

Competibacter grow typically as coccobacilli or rods. Cells are relatively large, oval, and easily recognized under the microscope. They can form clusters of tetrads, but also appear as microcolonies (Figure 9.15, Chapter 9).

Alphaproteobacterial GAOs. Other GAOs present in some full-scale EBPR plants are members of the Alphaproteobacteria. Two distinct subgroups cluster 1 and 2, related to Defluviicoccus vanus, are distinguished based on their 16S rRNA sequences (Wong et al., 2004; Meyer et al., 2006). They are most commonly seen in lab-scale EBPR reactors and are generally rare in full-scale EBPR plants. Only members of cluster 2 have been seen as commonly in some EBPR plants as Accumulibacter and Competibacter (Burow et al., 2007). It is likely that further phylogenetic diversity exists within this group.

The Defluviicoccus-related GAOs grow as cocci and rods, and typically exhibit a TFO morphology, where they form small colonies, although filamentous forms also exist (Figure 9.16, Chapter 9).

Other potential GAOs. Several cultured bacteria have also been proposed as GAOs but none are abundant in full-scale treatment plants. These include Amaricoccus kaplicensis (Maszenan et al., 1997), also named as Tetracoccus cechii (Blackall et al., 1997), and the original 'G' bacteria of Cech & Hartman (1993), Quatrionicoccus sp., Micropruina glycogenica, Kineosphaera limosa (see review by Oehmen et al., 2007). Other probe-defined uncultured populations within Betaproteobacteria and Gammaproteobacteria were proposed as putative GAOs (targeted by probe Bet65 and Gam445) in some full-scale EBPR plants (Kong et al., 2007) but their functional importance is unclear.

## 4.3.1 Probes for detection of GAOs

The oligonucleotides probes designed for identifying and quantifying these GAOs are shown in Table 4.2, and the tree in Figure 4.2 illustrates the specificity and coverage of these probes.

For most Competibacter in an activated sludge sample, either the general GB probe or a combination of the two lower hierarchical level probes GB\_G1 probe (identical to GAOQ989) and GB\_G2 can be used. Generally, the use of the GAOmix (equal amounts of GAOQ989 and GB\_G2) is recommended, as this targets most known Competibacter-related sequences. Seven different subgroups can be targeted by GB\_1-GB\_7 (Kong et al., 2002). However, as stated above, there is little information that supports the application of these subgroup probes, as possible differences in their distribution and ecology are not known.


Table 4.2. 16S rRNA oligonuclotide probes for the identification of potential GAOs.

Figure 4.2. 16S rRNA-based phylogenetic tree showing main lineages of known GAOs. Brackets indicate the coverage of the probes listed in Table 4.2.

The alphaproteobacterial GAOs related to members of the genus Defluviicoccus can be targeted by two sets of probes. In our experience from studies of full-scale plant communities, we recommend using the DF1MIX (TFO\_DF218 plus TFO\_DF618) for cluster 1 and DF2MIX (DF988 plus DF1020 and helper probes H966 and H1038) for cluster 2 members.

## 5

## Identification of filamentous bacteria by FISH

Caroline Kragelund, Elisabeth Mu¨ller, Margit Schade, Hien Thi Thu Nguyen, Hilde Lemmer, Robert Seviour, and Per Halkjær Nielsen

## 5.1 INTRODUCTION

Filamentous bacteria are found in all types of WWTPs, where they are often responsible for bulking (e.g. inadequate separation of biosolids and liquid effluent phases) or foaming (biosolids transported to the surface of either process tank or clarifier, typically dominated by one or two filamentous morphotypes). Filamentous bacteria should be considered as normal members of the activated sludge communities, primarily involved in degradation of organic material. Under some conditions they proliferate to such an extent that they markedly affect treatment plant performance. Their relative abundance may then exceed the usual 1–3% of the total biomass. The reasons for their excessive proliferation are many and will not be covered in this chapter, since several recent reviews and books discuss this in detail (Eikelboom, 2000; Jenkins et al., 2004; Tandoi et al., 2006; Nielsen et al., 2009; Seviour et al., 2009). In order to control the growth of these problematic bacteria, their reliable identification is necessary as the factors known to promote their growth can vary considerably. These include presence of sulphides in the influent, lack of nutrients or low oxygen concentrations in the process tank. Methods used for their attempted control are described elsewhere [for an overview on causes of sludge separation problems and control measures in several countries see (Tandoi et al., 2006)]. Filamentous bacteria also inhabit biofilm systems where they are located often at their surfaces (Galvan and de Castro, 2007). Here they rarely cause problems. Bulking arising from their excessive presence may occur, but only occasionally when clarifiers are employed to

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

settle surplus biofilm sludge. Excessive growth of filamentous bacteria may also hamper settling of granules in these new reactor systems (Liu and Liu, 2006).

In 1975 Eikelboom published descriptions of a range of filamentous bacteria observed in a large number of municipal activated sludge WWTP by conventional light microscopic characterization methods (Eikelboom, 1975). He divided these into morphotypes based on morphological characteristics and staining properties, and their abundance was expressed as a numerical filament index ranging between 0 (no filamentous bacteria present) to 5 (extensive growth of filamentous bacteria), see Eikelboom (2000). Manuals containing descriptions of 26 different filament morphotypes in samples taken largely from municipal WWTPs were published subsequently and are still used routinely to identify filamentous bacteria in activated sludge (Eikelboom, 2000; Jenkins et al., 2004). Furthermore, an additional 40 morphotypes have since been detected and described in industrial treatment systems (Eikelboom and Geurkink, 2000; 2002; van der Waarde et al., 2002). Most of these industrial morphotypes have only rarely been found in municipal plants and only at low levels.

The presence of these morphotypes in full-scale WWTPs has been investigated and reported in several plant surveys around the world. Generally, bulking or foaming is observed in 30–50% of all plants at any given time, clearly indicating the widespread nature of these operational problems. One important outcome from these surveys is that some morphotypes appear to be associated only with some types of treatment systems defined in terms of their process design, sludge loading (sludge age) and type of influent wastewater. This probably explains why their occurrence and frequency often differ markedly between countries where different WWTP operating concepts are preferred, as demonstrated by the dominating morphotypes encountered in treatment systems in Denmark, Germany, and Australia (Table 5.1). These bacteria are divided into those seen in i) conventional plants with organic removal with or without nitrification at a sludge load of around 0.15–0.8 kg BOD/kg MLSS d (often called high F/M filaments) and ii) more advanced plants with denitrification and biological phosphorus removal at a sludge load of around 0.01–0.15 kg BOD/kg MLSS d (low F/M). Presence of filamentous bacteria and operating information from reactors with biofilms or granules is limited and not included here.

From molecular methods of analyses we now know that basing identification into morphotypes on microscopic characters is frequently inadequate for the reliable identification of many filamentous bacteria. One morphotype can embrace several phylogenetically very different organisms, as exemplified by the Nostocoida limicola morphotype. Studies have revealed that the N. limicola morphotypes affiliate with members of the class Alphaproteobacteria (Snaidr et al., 2002), the phyla Chloroflexi (Schade et al., 2002), Firmicutes (Liu et al., 2000), Planctomycetales (Liu et al., 2001b), and Actinobacteria (Blackall et al., 2000; Liu and Seviour, 2001). This example illustrates that identification of individual filamentous organisms in activated sludge and biofilm systems often requires further molecular analyses with techniques like FISH to confirm their true identity after the manuals based on morphology have been used.

To illustrate the broad phylogenetic diversity existing among these filamentous bacteria, a comprehensive phylogenetic tree of all commonly encountered species is given in Figures 5.1a–5.1c. Most filamentous bacteria can be identified at least to phylum level by FISH (e.g. Chloroflexi). It is then often possible to distinguish between individual genera or species. However, more specific gene probes are still needed to completely resolve a number of morphotypes (e.g. those in Mycolata and Chloroflexi). In Table 5.2, we have provided an overview of the most commonly observed morphotypes and their phylogenetic affiliations when known. Furthermore, speciation of members of these phylogenetic groups where available and the corresponding morphotypes are listed in Table 5.3.

The following sections discuss each of the filamentous morphotypes described by Eikelboom (2000; 2006) (Table 5.1). They detail their abundances, provide a short description of their morphological appearances, and discuss their current phylogenetic affiliations, which are illustrated as phylogenetic trees



 Industrial WWTPs show similar filament ranking in high and low loaded plants, decisive is the carbon sources quality

Figure 5.1a. Maximum likelihood phylogenetic tree of the 16S rRNA gene sequences of the important filamentous Proteobacteria and related sequences. All sequences were at least 1000 bp long. The scale bar corresponds to 0.1 substitutions per nucleotide position.

which include FISH probe coverage of each. Lastly, the currently available FISH probes for their in situ identification are listed and recommendations made as to which FISH probes should be used for their in situ identifications. However, it is important of course to check each probe specificity using available databases prior to use, since many new sequences are generated every day (see Chapter 7).

As morphotype abundance depends on plant operating parameters which can vary in different countries, we have chosen to present these filaments in alphabetic order. As many potential users of this handbook probably are familiar with the morphotype descriptions, we consider this to be the most appropriate method for its practical application.

## 5.2 FISH DETECTION OF FILAMENTOUS BACTERIA

#### General approach

on 04 February 2019

In order to identify the dominant filamentous organism/s in a biofilm or activated sludge sample we propose the following procedure:

1. Observe by light microscopy (phase contrast) the abundance of filamentous bacteria and, if relevant, record the filament index (use manuals of Eikelboom, 2000; Jenkins et al., 2004).

Figure 5.1b. Maximum likelihood phylogenetic tree of the 16S rRNA gene sequences of the important filamentous Chloroflexi, Planctomycetales and Bacteroidetes and related sequences. All sequences were at least 1100 bp long. The scale bar corresponds to 0.1 substitutions per nucleotide position.


Figure 5.1c. Maximum likelihood phylogenetic tree of the 16S rRNA gene sequences of the important filamentous Gram positive bacteria and related sequences. All sequences were at least 1300 bp long. The scale bar corresponds to 0.1 substitutions per nucleotide position.

## 5.3 FILAMENTOUS BACTERIA DETECTION AND IDENTIFICATION

This section provides advice on how best to select the FISH probes to detect and identify filamentous bacteria found in activated sludge, foam, and biofilms. It includes actual data on their abundance in municipal and industrial WWTPs and biofilm reactors, a brief morphological description, their phylogenetic affiliation, and recommendations for FISH probes best suited for this purpose.

For in situ monitoring purposes a semiquantitative filament count might be assessed by determining a class index according to Eikelboom (2000). This is done for both conventional microscopy and FISH in order to achieve comparable data. Class indices from 0 indicating no filamentous growth to 5 indicating extensive growth of filamentous organisms with class index steps differing by a factor of 10 turned out to be helpful for practitioners' monitoring needs. In case of FISH overall biomass concentrations should be checked by DAPI or a universal bacteria probe such as EUBmix to be similar to those examined in wet mount.


Table 5.2. Overview of the most common morphotypes and their possible phylogenetic identity.


Table 5.3. Overview of the phylogenetic affiliation, species name if possible and corresponding morphotype.


#### Table 5.3. Continued

## 5.3.1 Beggiatoa morphotype

#### Abundance, morphology and affiliation

Abundance of Beggiatoa morphotype reported in plants in Australia, the Netherlands and the US (summarized in Jenkins et al., 2004) is moderate. No bulking or foaming problems have been implied from its presence. They are often present in low numbers in biofilms (Nielsen PH and Nielsen JL, unpublished results). Beggiatoa spp. serve as a strong indicator for a presence of reduced sulfur compounds, as it may be mixotrophic, using these as energy sources.

The Beggiatoa morphotype is characterized by being mobile by gliding and exhibiting straight to bent filaments with a cell diameter of 1.5–2.5 mm. In vivo sulfur granules are often visible. Beggiatoa stain Gram negative and Neisser negative.

The genus Beggiatoa is affiliated in the family Thiotrichaceae (class Gammaproteobacteria).

#### Recommending the best probes

Three Beggiatoa strains have been isolated from activated sludge (Williams and Unz, 1985) but no phylogenetic analyses were performed on them. Only FISH probes for marine species exist (Mussmann et al., 2003; Macalady et al., 2006), and so the broad class probe GAM42a for Gammaproteobacteria might indicate the presence of Beggiatoa in activated sludge, see Table 5.4. However, Beggiatoa is identifiable routinely by phase contrast microscopy only. The probes for marine species have never resulted in a FISH positive signal for this morphotype in activated sludge samples (Nielsen JL and Kragelund C, unpublished results).

## 5.3.2 Haliscomenobacter hydrossis morphotype

#### Abundance, morphology and affiliation

The abundance of H. hydrossis is reported in many surveys based on morphology (summarized in Jenkins et al., 2004), usually without ever being present in large amounts. Surveys using FISH probes have confirmed that H. hydrossis is almost always present in minor amounts in both municipal and industrial plants, independent of plant configuration (Kragelund et al., 2008).

The H. hydrossis morphotype exists as short straight (needle-like appearance) or longer slightly bent filaments, usually protruding from the flocs. Filament length is variable, and trichome diameter is approx. 0.4 mm. Cells are surrounded by a sheath. They stain Gram negative and Neisser negative.

These distinctively thin needle-shaped bacteria usually affiliate with members of H. hydrossis in the genus Haliscomenobacter (family Saprospiraceae) within the phylum Bacteroidetes. However, other


Table 5.4. General probes covering the different phyla and classes or broader groups. Can be used together with more specific probes or, if no specific probes exist, to reveal the overall affiliation of an otherwise unidentified filamentous organism.


# Formamide 0–35%

 A. use pretreatment with achromopeptidase and lysozyme, no extended hybridization time needed

 B. use lysozyme, acid, lipase, proteinase K and extended hybridization time 15–18h

 C. Lysozyme and extended hybridization time 15–18h

filaments with the same morphology are often detected, either belonging to other members of the Bacteroidetes (Kragelund et al., 2008) or to those in the phylum Chloroflexi (Kragelund et al., 2007a). Other filamentous bacteria can demonstrate a similar appearance but they have been identified solely on morphology and their phylogenetic affiliations are currently unresolved.

#### Recommending the best probes

Table 5.5 shows those probes for identification and quantification of the H. hydrossis morphotype, and the tree presented in Figure 5.2 illustrates their specificity and coverage.

Table 5.5. Haliscomenobacter hydrossis morphotype: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability.


Figure 5.2. 16S rRNA-based phylogenetic tree showing main lineages of known H. hydrossis bacteria. Brackets indicate the coverage of the probes listed in Table 5.5. The exact branching order of the H. hydrossis lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

The most commonly used probe is HHY, targeting the species H. hydrossis which was designed against a pure culture isolate (Wagner et al., 1994a). A broader probe HHY-654 is now available which also targets an additional H. hydrossis isolate (Kragelund et al., 2008). Substantially more H. hydrossis respond to this probe in full-scale plants. Other probes have been published to detect Bacteroidetes in different habitats but these are very broad and do not target individual species (e.g. CFB286, CFB563 and CFB719, Weller et al., 2000). Even though these were designed originally to target freshwater or marine Bacteroidetes (CFB286) they also target filaments in activated sludge. Other probes detect members of the phylum Bacteroidetes (CF319 aþb, Manz et al., 1996), targeting both single cells and filamentous bacteria. The SAP-309 probe (Schauer and Hahn, 2005) targets members of the family Saprospiraceae which contain both single cells and filamentous bacteria.

Our recommendation for detection of the H. hydrossis morphotype in the Bacteroidetes in activated sludge is probe HHY-654. To detect most H. hydrossis and H. hydrossis–like bacteria, the broader targeting probes CFB719 and SAP-309 can be used in combination. Morphological appearance of probedefined H. hydrossis morphotype can be seen in FISH image Figure 9.17, Chapter 9.

## 5.3.3 Leucothrix mucor morphotype

#### Abundance, morphology and affiliation

The abundance of Leucothrix mucor has been assessed in surveys based only on morphology (Jenkins et al., 2004). It has also been detected by FISH in industrial samples (van der Waarde et al., 2002). It is observed only occasionally in WWTPs where it plays no decisive role in bulking or foaming incidences (van der Waarde et al., 2002; Jenkins et al., 2004).

The Leucothrix mucor morphotype is characterized by bent to twisted filaments, a highly variable cell diameter and a variable cell morphology of discoid, round, to rod-shaped cells. It never has intracellular S-granules and stains Gram negative and Neisser negative.

Leucothrix mucor is member of the genus Leucothrix, closely related to genus Thiothrix in the family Thiotrichaceae (class Gammaproteobacteria).

#### Recommending the best probes

Only one probe exists for L. mucor (LMU) (Wagner et al., 1994a) designed based on a pure culture isolate. It can also be targeted by the class specific probe for Gammaproteobacteria (GAM42a, see Table 5.4). The LMU probe sequence is included in Table 5.6, which contains FISH probes available for less commonly observed filament morphotypes. The tree in Figure 5.3 illustrates the specificity and coverage of these probes.

## 5.3.4 Microthrix parvicella morphotype

#### Abundance, morphology and affiliation

The Microthrix parvicella morphotype is an important bacterium responsible for bulking and foaming incidences throughout the world (Eikelboom, 2000; Jenkins et al., 2004). Two species have been identified in activated sludge. One, Candidatus M. parvicella, predominates in WWTPs with biological N and P removal treating primarily domestic wastewater (Blackall et al., 1996). The other, Candidatus M. calida, resembles this morphotype but its trichome is slightly thinner and it can grow at higher temperatures in pure culture. It has only been found so far in WWTPs treating mainly industrial waste, both with and without biological N and P removal (Levantesi et al., 2006).

Microthrix parvicella morphotype filaments are bent to curled with a cell diameter of 0.5 – 0.6 mm (M. calida: cell diameter between 0.3 and 0.6 mm). They stain Gram positive and Neisser positive with

Table 5.6. Morphotype Leucothrix mucor, Sphaerotilus natans, Leptothrix discophora, Streptococcus and 1863 morphotype: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability. Exemplified for different morphotypes which are not very commonly observed in activated sludge.


This probe is denoted Str in probeBase. C. Lysozyme and extended hybridization time 15–18h.

Figure 5.3. 16S rRNA-based phylogenetic tree showing main lineages of different morphotypes which are not very commonly observed in activated sludge. Brackets indicate the coverage of the probes listed in Table 5.6. The exact branching order of the bacterial lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

polyP granules, which may be absent, for example, after prolonged starvation periods or after treatment with polyaluminum products (Eikelboom, 2006). M. parvicella morphotype filaments can also contain lipid/PHB granules (Rossetti et al., 2005).

Candidatus M. parvicella and Candidatus M. calida both affiliate with the Actinobacteria (unclassified Actinobacteria) (Blackall et al., 1996).

#### Recommending the best probes

Table 5.7 shows the probes currently available for the morphotype M. parvicella, and the tree in Figure 5.4 illustrates their specificity and coverage.

To detect M. parvicella, cell permeabilization by pretreatment with enzymes is required to obtain a satisfactory FISH signal. Many protocols have been published for permeabilizing M. parvicella and other Gram positive bacteria (Beimfohr et al., 1993; Erhart et al., 1997; Carr et al., 2005; Kragelund et al., 2007b). That which seems to promote the strongest FISH signal is combining achromopeptidase and lysozyme (Kragelund et al., 2007b), which was developed originally for CARD-FISH (Sekar et al., 2003). An extended hybridization time of up to 15–18 hours might be needed to eliminate irregular fluorescence signal strength along the filament (Mu¨ller et al., 2007).

The phylum probe for Actinobacteria (HGC69a) designed to target their 23S rRNA (Roller et al., 1994) is not suitable for Candidatus M. parvicella, since at least two mismatches exist with it (Bradford et al., 1998). Instead the probe HGC1156 (Erhart, 1997) targeting the Actinobacteria can be used for Candidatus M. parvicella (Mu¨ller et al., 2007) and Candidatus M. calida. Several probes have been published that target the M. parvicella morphotype. Four probes MPA645, MPA223, MPA60, and MPA650 were developed for in situ detection of Candidatus M. parvicella (Erhart et al., 1997). The latter probe requires two competitor probes. Another two probes Mpa-all-1410 and Mpa-T1-1260 were developed which target Candidatus M. calida (Levantesi et al., 2006). However, Mpa-all-1410 has a weak central mismatch to Candidatus M. parvicella, but despite this mismatch, the probe clearly hybridized to Candidatus M. parvicella isolates and M. parvicella present in sludge samples.

We recommend the MPAmix (mixture of MPA645, MPA223 and MPA60) be used for in situ detection of the Microthrix morphotype including both Candidatus M. parvicella and Candidatus M. calida, together with the HGC1156 probe targeting Actinobacteria. To distinguish between similar


Table 5.7. Microthrix morphotype: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype and comments on recommendability.

A. use pretreatment with achromopeptidase and lysozyme, no extended hybridization time needed

B. use lysozyme, acid, lipase, proteinase K and extended hybridization time 15–18h

Figure 5.4. 16S rRNA-based phylogenetic tree showing main lineages of Microthrix morphotypes in the Actinobacteria. Brackets indicate the coverage of the probes listed in Table 5.7. The exact branching order of the Actinobacteria lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

looking Candidatus M. parvicella and Candidatus M. calida in a sludge sample, MPAmix should be applied first in order to detect both. Thereon Mpa-T1-1260 should be used to visualize Candidatus M. calida filaments only. FISH images of M. parvicella and M. calida morphotypes can be seen Figures 9.18 and 9.19 in Chapter 9.

## 5.3.5 Nostocoida limicola morphotypes

#### Abundance, morphology and affiliation

The abundance of the different N. limicola morphotypes (I–III) has been assessed by morphologically based surveys and these are associated with bulking and foaming (Eikelboom, 2000; Jenkins et al., 2004). Several surveys using FISH probes have been published; Probe NLIMII175 detected this morphotype frequently in WWTPs of Australia, New Zealand, France (Liu and Seviour 2001) and in a higher frequency than did the NLII65 probe in German WWTPs (Mu¨ller et al., 2007). The N. limicola affiliating with the Alphaproteobacteria seem to play a major role in the bulking events in plants with a high load of industrial waste (Levantesi et al., 2004). In municipal WWTPs in Germany only a few have been recorded (Mu¨ller et al., 2007), while in Australian EBPR plants they are more abundant (Seviour, unpublished). The Chloroflexi affiliated N. limicola morphotype has been observed in several German scums in municipal plants (Mu¨ller et al., 2007).

The N. limicola morphotype is distinguished by its long, coiled and tangled filaments. Three different morphotypes (I, II, and III) have been described based on minor differences primarily in their cell diameter (Eikelboom, 2000). Individual variable opaque coccoid cells are barely resolvable by light microscopy in thin filaments of N. limicola type I. Cells are coccoid to discoid in the thicker type II

filaments. Flattened discoid cells are found in N. limicola type III. Diameter varies from ca. 0.8 to 2 mm. Nostocoida limicola filaments are Gram variable and often but not always Neisser positive, and show a light to dark grey violet cell wall.

It is difficult (or impossible) to relate each of the morphotypes to phylogenetically distinct probedefined species. They have been identified as members of several phylogenetically distantly related species. Some affiliate with the Alphaproteobacteria, others with members of the Chloroflexi, Firmicutes, Planctomycetes, and Actinobacteria.

#### Recommending the best probes for Nostocoida limicola II

Table 5.8 shows probes that can be used for N. limicola morphotypes, and the tree in Figure 5.5 illustrates their specificity and coverage.

Nostocoida limicola morphotype II is commonly seen in municipal WWTPs. Many belong to Candidatus N. limicola affiliating with the genus Tetrasphaera in the family Intrasporangiaceae in the phylum Actinobacteria. This filamentous species is usually not observed in large quantities in bulking incidences but often in low numbers especially in municipal WWTPs. Importantly, in German plants the NLIMII175 probe has been observed to also target other morphotypes including bent sickle-shaped rods, and long thin filaments developing two different morphotypes with long rectangular cells and spherical to square-shaped cells (Schade M., unpublished). A typical N. limicola morphotype can be seen in Fig 9.20 in Chapter 9.

The N. limicola morphotype II can also be a member of the phylum Chloroflexi (class Thermomicrobia, Hugenholtz and Stackebrandt, 2004) and is targeted by probe AHW183 (designed based on 4 cultured isolates, Schade et al., 2002). A typical N. limicola morphotype can be seen in Figure 9.21 in Chapter 9.

In industrial WWTPs the most commonly observed N. limicola morphotypes II belong to the Alphaproteobacteria, and these account for many incidents of bulking (Levantesi et al., 2004). The most commonly observed filaments Candidatus Monilibacter batavus, Candidatus Alysiomicrobium bavaricum, Candidatus Sphaeronema italicum and Candidatus Alysiosphaera europaea can be detected with the general probe (ALF968) for the Alphaproteobacteria, and more specifically with probes MC2-649 and DF198, PPx3-1428, Sita-649 and Noli-644, respectively. These probes are not very specific, for example, MC2-649 (Nittami, unpublished results), PPx3-1428 and Sita-649 (Kragelund et al., 2006), but they can still be used for preliminary screening. For some of these filaments more specific probes are available such as PPx1002 for Candidatus Alysiomicrobium bavaricum, and Nost993 in combination with a helper probe for Candidatus Sphaeronema italicum (Kragelund et al., 2006). Probe DF988 designed for the GAO Defluviicoccus (see Chapter 4) also targets Candidatus Monilibacter batavus (Seviour R., unpublished results). Other alphaproteobacterial species with this morphology have been identified in activated sludge: Meganema perideroedes can be detected with probes Meg983 and Meg1028 (Thomsen et al., 2006a) and Candidatus Combothrix italica is detectable with probe Combo1013 (Levantesi et al., 2004). However, these filaments are not commonly observed in either industrial or municipal WWTPs.

For identification of these alphaproteobacterial Nostocoida limicola morphotype II we recommend the following probes listed in Table 5.8. The ALF968 probe detects all known filaments belonging to the Alphaproteobacteria (see Table 5.4). For further species differentiation the following probes are recommended: DEF198, PPx3-1428, Sita-649 and Noli-644 which target the most commonly observed species predominantly in industrial sludges (Table 5.8). However, bear in mind that more specific probes are available for some of the alphaproteobacterial species and please consult the listed references for details. For the actinobacterial species, probe NLIMII175 is recommended, but be aware that other

Table 5.8. Nostocoida limicola morphotypes: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments.


Figure 5.5. 16S rRNA-based phylogenetic tree showing main lineages of Nostocoida limicola morphotypes. Brackets indicate the coverage of the probes listed in Table 5.8. The exact branching order of the Nostocoida limicola morphotypes lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

morphotypes have been observed with this probe. Also the phylum probe HGC1156 targets the morphotype (Table 5.4). Lastly, the Chloroflexi-related N. limicola II is targeted by probe AHW183 which is also targeted by CFX mix (Table 5.4). Morphological appearance of Nostocoida limicola II morphotypes can be seen in FISH images in Figures 9.20–9.30, Chapter 9.

Other probes for morphotypes N. limicola I and III

For Nostocoida limicola morphotypes I and III affiliating to members of the Firmicutes and Planctomycetales several probes have been developed. However, there are some discrepancies between morphotype definitions and probe targets so these do not necessarily target N. limicola morphotypes I and III as described by Eikelboom. Probe NLIMI91 was designed based on sequences from three N. limicolalike isolates, affiliated with Trichococcus spp. within the phylum Firmicutes (Liu et al., 2000). It can also be targeted by the phylum probes LGC354A-C (Table 5.4). This probe was used to detect N. limicola I in Australian, Italian and French WWTPs (Liu and Seviour, 2001). However, no detection of the typical N. limicola I morphotype was observed with this probe in either municipal or industrial European plants (Eikelboom, 2006; Mu¨ller et al., 2007). Quite the contrary, NLIMI91 detected short filaments and coccoid cells in Australia and Europe (Liu and Seviour 2001, Mu¨ller et al., 2007) and long thin filaments with rectangular cells (Schade M., unpublished). Probe NLIMIII301 was designed based on five N. limicola III-like isolates belonging to the Planctomycetales. They can also be targeted by the phylum probe Pla46 (Table 5.4). The experience with this probe is rather conflicting: In contrast to the opaque cells described by Eikelboom for morphotype Nostocoida III, they possessed transparent cocci in slightly bent filaments (Liu et al., 2001b). The probe targeted single cells within transparent Isosphaera filaments in some cases (Mu¨ller et al., 2007). In German plants probe NLIMIII301-targeted cells were never present in high abundance as compared to the typical N. limicola morphotype III (Mu¨ller et al., 2007), but were highly abundant in some Australian plants treating dairy wastes appearing as long chains of coccoid cells (Seviour R., unpublished), see Figure 9.31, Chapter 9.

## 5.3.6 Nocardioform actinomycetes/Mycolata morphotype

#### Abundance, morphology and affiliation

Abundances of Nocardioforms/Mycolata have been assessed in surveys based on morphology (Eikelboom, 2000; Jenkins et al., 2004) and with available FISH probes (Eales et al., 2005; Mu¨ller, 2006; Kragelund et al., 2007b; Mu¨ller et al., 2007). Mycolata are found in municipal and industrial WWTPs with or without nutrient removal (Kragelund et al., 2007b; Seviour et al., 2008).

Two quite distinct filamentous morphotypes have been identified. Gordonia amarae-like organisms (GALOs) have filaments with perpendicular branching angles, while acute-angle branched filaments distinguish the pine tree-like organisms (PTLOs). This difference is used for their microscopic 'identification'. In addition, non-filamentous growth forms may occur, and here are referred to as NOClike organisms. Development of these often depends on environmental conditions or organism growth phase. Therefore it is difficult (or often impossible) to relate any of these to phylogenetically distinct probe-defined populations. Abundances of several species of Mycolata have been assessed using available FISH probes in foam and sludge samples. Studies have examined the abundances of Gordonia, Skermania, and Rhodococcus populations and NOC-like organisms in plant samples from all over the world (e.g. Australian, Czech, Danish, German, and Swedish plants) (Eales et al., 2005; Kragelund et al., 2007b; Mu¨ller, 2006; Mu¨ller et al., 2007).

Their cell diameters vary from 0.6–2.0 mm, depending on the morphotype present. Usually they are Gram positive, but Gram variable morphotypes have also been observed especially in industrial plants or those operating at a very high sludge age. Neisser/DAPI positive polyP granules are occasionally seen.

The Mycolata morphotypes formerly known as 'nocardioforms' all contain mycolic acid in their cell walls. They are classified as members of the suborder Corynebacterineae within the Actinobacteria (Stackebrandt et al., 1997). This group encompasses members of several genera including Corynebacterium, Dietzia, Gordonia, Millisia, Mycobacterium, Nocardia, Rhodococcus, Skermania, Tsukamurella, and Williamsia, based primarily on chemical, molecular, and morphological markers (Chun et al., 1997; Butler et al., 2005; Soddell et al., 2006). Nocardia amarae has been reclassified as Gordonia amarae (Goodfellow et al., 1994; Klatte et al., 1994; Ruimy et al., 1994). Nocardia pinensis has also been reclassified into a new genus Skermania as Skermania piniformis (Chun et al., 1997). Phylogenetic and phenotypic studies of Mycolata in WWTPs reveal a high diversity among members of this group and Dietzia, Gordonia, Mycobacterium, Rhodococcus, Tsukamurella, and S. piniformis are often involved in foaming events (Goodfellow et al., 1998; Schuppler et al., 1998; Soddell et al., 1998; Seong et al., 1999; Stainsby et al., 2002; Wagner and Cloete, 2002; Nam et al., 2003; Eales et al., 2005; Mu¨ller, 2006; Kragelund et al., 2007b).

#### Recommending the best probes

In general FISH protocols for detecting Mycolata incorporate a pretreatment permeabilization step to facilitate probe transport into the cells, as described above for the Microthrix morphotype. A lysozyme pretreatment at least is recommended. For some Rhodococcus spp. and Skermania piniformis an acid, lipase, proteinase K treatment might be necessary (Carr et al., 2005, Mu¨ller, 2006). The protocol that seems to promote the strongest FISH signal for detecting Gordonia and Skermania species is combining achromopeptidase and lysozyme (Kragelund et al., 2007b). In addition extending hybridization times to 15–18 h is recommended, as outlined earlier for Microthrix.

Table 5.9 shows the FISH probes useful for Mycolata morphotypes at group, genus, and species level, and the tree in Figure 5.6 illustrates their specificity and coverage. Firstly probes for detecting the entire Mycolata group are described, and then those targeting individual genera and species are discussed.

Mycolata group-targeted probes. No probe has been designed that targets the entire Mycolata group. Probe HGC1156 targets most known Mycolata but also many other members of the phylum Actinobacteria. However, branched filaments (GALOs and PTLOs) can to some extent be detected with the probe if no intermediate morphotypes are observed as reported in several papers (Kragelund et al., 2007b). But NOClike organisms cannot be distinguished from other non-mycolic acid-containing organisms. Therefore, additional group- and genus-targeted probes are necessary for further identification. This step is essential since individual Mycolata members can differ markedly in their physiological features, indicating that individual control strategies are required for each.

Two broad probes (Myc657 and MNP1) target most known members of the family Corynebacteriaceae where most of the genera of mycolic acid containing branched filaments are grouped (Myc657; Davenport et al., 2000; MNP1; Schuppler et al., 1998), see Table 5.4. Probe Myc657 detects more Mycolata than MNP1, but interestingly the latter probe detects NOC-like organisms better than the branched filamentous bacteria. Probe CMN119 (Erhart et al., 1997) detects a subgroup of the Corynebacterineae excluding Dietzia spp. and Tsukamurella spp. and this probe has identified GALOs which could not be detected with the other broad probes (Mu¨ller, 2006). These broad probes are grouped in Table 5.4 along with other group or phylum probes.

Gordonia (GALO)-specific probes. The probe Gor596 is suitable for detecting all known members of the genus Gordonia (de los Reyes et al., 1997). Other probes have been designed to target Gordonia spp. in sludge and foam (Gam192, G. amarae (de los Reyes et al., 1997), GLP2, different Gordonia sequences


Table 5.9. Nocardioform actinomycetes/Mycolata morphotype: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability.

A. use pretreatment with achromopeptidase and lysozyme, no extended hybridization time needed

C. Lysozyme and extended hybridization time 15–18h

(Schuppler et al., 1998), and G.am205, G. amarae (de los Reyes et al., 1998). Hierarchical probe overlap is observed with Myc657 and MNP1.

Skermania piniformis (PTLO)-specific probes. Two probes for detecting Skermania piniformis have been published: Spin1449 (Eales et al., 2006), and NPI452 (Bradford, 1997). PTLOs are commonly observed using probe Spin1449 in foaming European WWTPs. Hierarchical probe overlap is observed with Myc657.

Rhodococcus (GALO)-specific probes. Members of the genus Rhodococcus are a diverse group of organisms. Rainey et al. (1995) showed that Rhodococcus species fall into six distinct phylogenetic groups and currently contain some members whose taxonomic status is uncertain. Therefore it is not possible to design a probe targeting all members of this genus. Nevertheless, two probes Rco1 and Rco2 are described which detect most Rhodococcus spp. (Davenport et al., 1998). However, these two probes showed no positive FISH results with pure cultures of Rhodococcus fascians, R. rhodnii, R. globerulus, and R. rhodochrous (Mu¨ller E., unpublished). They were optimized under different FISH conditions than used here (e.g. different hybridization formamide concentrations and temperature) and are thus not recommended.

Three probes are available for different Rhodococcus spp. The probe R.ery619 targets R. erythropolis (Mu¨ller, 2006) (Rhodococcus group IV according to Rainey et al., 1995). Hierarchical probe overlap occurs with Myc657 and MNP1 (Mu¨ller, 2006). The other two probes, i.e. RHOa429 and RHOb183, were designed to detect two Rhodococcus clusters not affiliated to any of Rainey's Rhodococcus groups or any known Rhodococcus species. Probe RHOa429 targets Rhodococcus spp. cluster A, and RHOb183 targets Rhodococcus sp. cluster B. Sequence analyses of Rhodococcus cluster A and B members show one and two mismatches to the MNP1 and Myc657 probes, respectively. They respond only to the group-specific probe CMN119 (Erhart, 1997; Mu¨ller, 2006).

Dietzia (NOC-like organisms)-specific probes. Two probes are available against Dietzia spp. showing a NOC-like appearance in scum. Probe DLP (Schuppler et al., 1998) identifies members of the genus

Figure 5.6. 16S rRNA-based phylogenetic tree showing main lineages of Nocardioform actinomycetes/ Mycolata morphotype. Brackets indicate the coverage of the probes listed in Table 5.9. The exact branching order of the Nocardioform actinomycetes/Mycolata lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

Dietzia (excluding D. maris), which form short irregular rods (Schuppler et al., 1998). Another probe DIE993 targets D. maris, D. psychralcaliphila, D. daqingensis, D. natronolimnaea, uncultured foam bacteria (Wagner and Cloete, 2002), and scum clones (Mu¨ller, 2006), all of which are members of the genus Dietzia. The morphology of the NOC-like organisms was bent single rods either in clusters or forming short non-branched filaments. Hierarchical probe overlap is observed with Myc657 and also MNP1 (Mu¨ller, 2006).

Recommending the best probes. We recommend probe Myc657 as the general probe targeting most Mycolata. For detecting Rhodococcus spp. in sludge or foam it might be advantageous to apply CMN119 as a general probe also because some Rhodococcus species are not targeted by the Myc657 probe. At genus level the following probes Gor596 (Gordonia spp.), Spin1449 (Skermania piniformis), RHOa429, RHOb183, R.ery619 (three different Rhodococcus spp.) and DIE993 (Dietzia spp.) differentiate between some common species within the broad Mycolata group. However, these more specific probes do not cover the entire Mycolata group and so it is still not possible to identify all GALOs and PTLOs by FISH.

For reliable identification it is best to use activated sludge samples and not foam samples, as organisms in the foam often have low FISH signals from their moribundity. See FISH images, Figures 9.32–9.38 in Chapter 9.

## 5.3.7 Sphaerotilus natans and Leptothrix discophora morphotype

#### Abundance, morphology and affiliation

This morphotype was once held responsible for many bulking incidences in plants operating with high F/M ratios before other potential bulking filaments have been described (e.g. Eikelboom 1975, Jenkins et al., 1993). It is still seen, primarily in municipal WWTPs without nutrient removal, but is now considered as being rarely responsible for bulking (van der Waarde et al., 2002). However, it can be found in large quantities in some industrial WWTPs worldwide, such as the pulp and paper industries in Australia (Seviour R, unpublished) or food industries such as breweries, dairies or starch processing plants in Germany (Lemmer et al., 2000). Leptothrix discophora has been identified in activated sludge plants but is rarely abundant (Jenkins et al., 1993; van der Waarde et al., 2002). When a FISH survey was conducted using species-specific probes on samples from industrial WWTPs (van der Waarde et al., 2002), S. natans was detected in approx. 25% and L. discophora in less than 1% of these plants, but no information about their relative abundances was given. More recently S. natans was present in only 13% of 126 industrial WWTPs, and the filament index never exceeded 1.5 (Kragelund C and Nielsen PH, unpublished results).

The S. natans morphotype comprises straight or bent sheathed filaments which are distinctively falsely branched and may produce swarming gonidia. Cell diameter is approx. 1.0–1.3 mm. Cells are rectangular (filaments center) or rod-shaped (filaments ends). It stains Gram negative and Neisser negative. S. natans and L. discophora share an almost identical morphology. Thus probes are needed for their differentiation (Eikelboom, 2006).

Sphaerotilus natans belongs to the genus Sphaerotilus within the Betaproteobacteria. Leptothrix discophora is also a member of the Betaproteobacteria and is located within the genus Leptothrix.

#### Recommending the best probes

Probes SNA and LDI (Wagner et al., 1994a) for detecting the S. natans and L. discophora morphotypes, respectively, are described in Table 5.6, which contains sequences for less commonly seen filament morphotypes, while the tree in Figure 5.3 illustrates their specificity and coverage. These FISH probes were both developed with pure cultures of S. natans and L. discophora. Probe SNA for S. natans requires the CTE competitor probe, since Wagner et al. (1994a) report that Sphaerotilus spp., S. natans, type 1701, L. discophora, and L. cholodnii all hybridized with it. Moreover, type 1701 and Leptothrix spp. show a close phylogenetic affiliation with Sphaerotilus spp. In German municipal WWTPs the probe SNA even with CTE still visualized single rod-shaped and coccoid cells in addition to Sphaerotilus filaments (Schade M., unpublished). Both S. natans and L. discophora can be hybridized by the class-specific probe for Betaproteobacteria (Table 5.4). See FISH images, Fig 9.39, Chapter 9.

## 5.3.8 Streptococcus morphotype

#### Abundance, morphology and affiliation

Filamentous streptococci can be quite abundant in municipal WWTPs with EBPR, constituting 1–4% of the biovolume (Kong et al., 2008). The filaments of Streptococcus are bowed and twisted, round cells with a

diameter of approx. 0.7 mm. Septa and constrictions are clearly visible. They stain Gram positive and Neisser negative. Filamentous streptococci affiliate in the genus Streptococcusin the family Streptococcaceae (class Bacilli, phylum Firmicutes). This morphotype may also cover other Firmicutes, such as Trichococcus spp. (Liu et al., 2000), which can also be assessed as morphotype Nostocida limicola I, see section above.

#### Probes for detection of streptococci

Table 5.6 shows that probe Strept (Trebesius et al., 2000) can be used to identify filamentous streptococci (see also Chapter 6). The probe targets 44% of all streptococci, and can be used together with the broader probe LGC354C (Meier et al. 1999). The morphologies of the Strept-probed cells embrace both small cocci and filamentous organisms. The phylogenetic tree in Figure 5.3 shows the less commonly seen filament morphotypes and the probes for their detection along with their specificity and coverage.

## 5.3.9 Thiothrix and Type 021N morphotypes

#### Abundance, morphology and affiliation

These morphotypes occur in municipal and industrial WWTPs with and without nutrient removal (Eikelboom, 2000; Jenkins et al., 2004) and in biofilm reactors (Nielsen PH. and Nielsen JL., unpublished results). FISH surveys have applied several probes targeting these morphoptypes and they were seen frequently in industrial samples where they were found in 38% of samples (detected using probes TNI and 21N) (van der Waarde et al., 2002). More specific probes targeting subpopulations of morphotype 021N were used in a survey, and two of the three 021N morphotypes were frequently seen in Japanese and European WWTPs (Kanagawa et al., 2000).

The Thiothrix morphotypes contain six microscopically recognizable different forms, whose trichomes can be separated on their diameter, which may vary from 0.6–2.5 mm with rectangular cells in the filament center and rod-shaped cells at their ends. Often sulfur granules accumulate in situ (Unz and Head, 2005). They generally display a positive S<sup>0</sup> result without or after incubation with a reduced sulfur source. They are Gram and Neisser negative.

The 021N filament morphotype has a cell diameter of 1.0–2.0 mm with discoid or almost square cells in the centre and rod-shaped cells at the filament ends. Occasionally small sulfur granules are visible in situ or after incubating with a reduced sulfur source. Type 021N stains Gram and Neisser negative.

Several morphologically indistinguishable filaments have been identified: at least one belongs to the Alphaproteobacteria (Meganema perideroedes) and possibly others (Howarth et al., 1999; Thomsen et al., 2006a) and some affiliate to the Chloroflexi being not separable microscopically from T. eikelboomii (Type 021N) (Kragelund et al., 2007a). As the S<sup>0</sup> test is not always useful in identifying each, FISH probes are needed to differentiate between them.

Currently, seven species of Thiothrix in the Gammaproteobacteria have been isolated and described (Cole et al., 2005; 2008; Unz and Head, 2005) and furthermore T. nivea (Teske et al., 1996) and T. ramosa (Polz et al., 1996) have been extensively described. The Eikelboom 021N morphotype has now been shown to represent three separate species within the genus Thiothrix (T. eikelboomii, T. disciformis and T. flexilis) (Howarth et al., 1999; Kanagawa et al., 2000; Unz and Head, 2005).

#### Recommending the best probes

Table 5.10 shows the oligonucleotide probes for identifying and quantifying the Thiothrix/Type 021N group, and the tree in Figure 5.7 illustrates their specificity and coverage.


Table 5.10. Thiothrix/021N morphotype rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability.

The probe G123T with its competitor G123T-C (Kanagawa et al., 2000) targets the entire Thiothrix/ 021N group. Probes 21N and TNI (Wagner et al., 1994) target subpopulations of type 021N and Thiothrix, respectively. The latter also hybridizes with some rod-shaped and coccoid cells, presumably the gonidia from Thiothrix. Probe 21N occasionally targets much thinner filaments, which are readily distinguished from type 021N (Schade M., unpublished). Probes based on pure culture isolates allow the three type 021N species to be differentiated: G1B targets group 1 (T. disciformis), G2M, group 2 (T. eikelboomii) and G3M, group 3 (T. flexilis).

We recommend using probe G123T and its competitor probe to detect the entire Thiothrix/021N group. As these probes may target filaments other than Thiothrix/021N, simultaneous application of probe G123T

with probes TNI and 21N is recommended (Schade M., unpublished). For further differentiation between these type 021N filaments (T. eikelboomii, T. flexilis and T. disciformis) and Thiothrix species more specific probes can be used, refer to Figure 5.7. Also the class-specific probe for Gammaproteobacteria targets these morphotypes (Table 5.4). See also FISH images in Figures 9.40–9.44, Chapter 9.

## 5.3.10 0041/0675 morphotype

#### Abundance, morphology and affiliation

This morphotype has been identified in WWTPs all over the world. 0041/0675 filaments are ranked as the second to sixth most frequently seen bacteria (summarized in Jenkins et al., 2004; Tandoi et al., 2006). FISH surveys have also been published with a selection of probes targeting the 0041/0675 morphotype. For example, approximately 15% of all morphotype 0041/0675 filaments were targeted by the probe for TM7 subdivision 1 in Danish WWTPs (Thomsen et al., 2002). A German survey detected a larger fraction (approx. 40%) of TM7 affiliated morphotypes (Mu¨ller et al., 2007).

The 0041/0675 morphotype has straight or bent filaments with a cell diameter of 0.7 to 2 mm and almost square cells. Filaments often show attached growth. They stain mainly Gram positive and Neisser negative. A faint Neisser staining response might suggest phosphorus deficiency in the plant.

The phylogenetic affiliation of members of morphotype 0041/0675 is diverse. Many show affiliation with members of the Chloroflexi (Bjornsson et al., 2002; Kragelund et al., 2007a). Others affiliate with Curvibacter (recently reclassified from Aquaspirillum) within the Betaproteobacteria (Thomsen et al., 2006b) and some are embraced by the Candidate division TM7 (Hugenholtz et al., 2001). However, blast searches of the TM7 probes show that approx. 50% of the targeted sequences are in fact affiliated to the phylum of Chloroflexi. Therefore, a thorough cross check of probe results with the filament morphological details after phase contrast microscopy is highly recommended.

#### Recommending the best probes

Table 5.11 shows the oligonucleotide probes available for identifying and quantifying morphotype 0041/0675, while the tree in Figure 5.8 illustrates their specificity and coverage.

Morphotype 0041/0675 filaments are most often identified with the phylum-targeted CFXmix probes for the Chloroflexi (Bjornsson et al., 2002; Kragelund et al., 2007a). These also detect morphotypes 0092 and 1851 and some H. hydrossis-like bacteria (see other sections). For more precise identification, genusand species-targeted probes within the Chloroflexi phylum are required, but these are currently not available. Some 0041/0675 morphotypes hybridize with probe Curvi997 when applied with competitor probes Comp1curvi997 and Comp2curvi997 (Thomsen et al., 2006b) and thus belong to the genus Curvibacter. However, this probe also targets some abundant microcolony-forming bacteria present in many WWTPs (Thomsen et al., 2004 and Chapter 3) and in addition hybridizes with type 1701. Some 0041/0675 filaments are detected with probes TM7905 targeting almost the entire Candidate division TM7, and TM7305, which target members of subdivision 1. No taxonomically defined species within this phylum have been assigned so far to type 0041/0675 filaments. In addition, single cells and other filaments have been seen hybridizing with these TM7905 and TM7305 probes (Hugenholtz et al., 2001).

We recommend applying probes for each phylum in studies with type 0041/0675 i.e. probes CFX mix for Chloroflexi members, Curvi997 and competitor probes for some Betaproteobacteria and probes TM7905 and TM7305 for those in candidate division TM7. Remember to cross-check the FISH data with those obtained with phase contrast microscopy. See also FISH images in Figures 9.45–9.48, Chapter 9.

Table 5.11. Morphotype 0041/0675: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability for bacteria with attached growth.


Figure 5.8. 16S rRNA-based phylogenetic tree showing main lineages of known bacteria with attached growth. Brackets indicate the coverage of the probes listed in Tables 5.11 and 5.13–5.14. The exact branching order of the bacteria with attached growth lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

## 5.3.11 0092 morphotype

by IWA Publishing user on 04 February 2019

#### Abundance, morphology and affiliation

Type 0092 appears prominently in microscopy-based surveys of activated sludge plants around the world (Jenkins et al., 2004). An Australian FISH survey of EBPR plants where this morphotype was abundant, showed that it never responded to the EUBmix probes (Speirs et al., 2009). Such survey data suggest that it seems to be associated especially with plants operating with low F/M feeds and long sludge ages, and thus is seen frequently in activated sludge systems removing P. However, it does not stain positively for polyP granules with DAPI.

Its morphotype is distinctive, appearing as short blunt ended trichomes extruding from flocs or occasionally freely suspended in the bulk liquid. Some reports suggest it is also enriched in foams, although whether it is a causative organism is unlikely. It can be 'identified' readily microscopically by its Neisser positive staining reaction, where filaments appear distinctively 'lilac' in color. It stains Gram negative.

Reports that this morphotype has been cultured should be treated cautiously. The phylogeny of most isolates has never been resolved, and where 16S rRNA sequencing has been performed, FISH probes designed against them have never hybridized with this filament morphotype in situ.

#### Recommending the best probes

Table 5.12 shows the oligonucleotides probes recommended now for identifying and quantifying type 0092, and the tree in Figure 5.9 illustrates their specificity and coverage.

Table 5.12. Morphotype 0092: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability.


Figure 5.9. 16S rRNA-based phylogenetic tree showing main lineages of morphotype 0092 belonging to the Chloroflexi. Brackets indicate the coverage of the probes listed in Table 5.12. The exact branching order of the Chloroflexi lineages cannot yet be determined due to a limited phylogenetic resolution of the 16S rRNA marker gene. Therefore, unclear tree topology is depicted as multifurcation.

Morphotype 0092 hybridized with the GNSB941 and CFX1223 probes targeting the Chloroflexi, but not the CFX109 or CFX784 probes for Chloroflexi subgroups 3 and 1a, or CHL1851 designed against the Chloroflexi type 1851 filament. Two more specific 16S rRNA-targeted probes were designed against clone sequences recovered from one plant (Speirs et al., 2009). The CFX197 probe hybridized with the slightly thicker of two morphological variants of this morphotype in several Australian P-removal plant samples, while CFX223 probe targeted the thinner variant. Is is not clear whether these two variants represent different species or strains of this filament. With all FISH analyses, the fluorescent signal from type 0092 was localized, giving the filaments a granular appearance, suggesting that their ribosomes are in aggregates in cells.

We recommend applying the probes CFX197 and CFX223 probes for targeting morphotype 0092. However, the probes need to be applied to more biomass samples from plants around the world before this issue can be addressed confidently. However, on present evidence, type 0092 morphotypes should be screened initially with the EUBmix, GNSB941 and CFX1223 probes, and then if negative for the first and positive for the latter two probes, the CFX197 and CFX223 probes should be used. Whether they hybridize globally with morphologically different type 0092 filaments is of interest. See also FISH images in Figures 9.49–9.51, Chapter 9.

## 5.3.12 1701 morphotype

#### Abundance, morphology and affiliation

Type 1701 is only seen occasionally, particularly in US plants, and is considered to cause severe bulking problems only rarely (Jenkins et al., 1993; Thomsen et al., 2006b). Filaments occur often in bundles.

The 1701 morphotype forms short filaments which are often attached to flocs. The trichome has a diameter of 0.7–1.1 mm, and consists of rod-shaped cells. Attached growth is often present. It is usually Gram and Neisser negative.

The identity of this morphotype is unclear. S. natans was suggested as a close relative (Howarth et al., 1998) based on 16S rRNA gene sequence analyses, and it is placed within the genus Sphaerotilus. Wagner et al. (1994a) also found type 1701 was closely affiliated with Sphaerotilus spp. and was readily detected by FISH with the probe SNA in full-scale plants. Other strains of this morphotype belong possibly to Aquaspirillum delicatum within the Betaproteobacteria (Thomsen et al., 2006b), which is now classified as Curvibacter delicatus (Ding and Yokota, 2004), a notation used throughout this handbook.

As stressed with other filamentous bacteria with attached growth (types 0041/0675 and 1851) a thorough cross-check of FISH images with those from phase contrast microscopy is recommended.

#### Recommending the best probes

Table 5.13 shows the oligonucleotides probes for identifying and quantifying type 1701, and the tree in Figure 5.8 illustrates their specificity and coverage.

The probe Curvi997 (formerly Aqs997) with competitor probes Comp1curvi997/Comp2curvi997 (formerly Comp1aqs997/Comp2aqs997) were developed initially for microcolony-forming bacteria. Curvi997 was shown subsequently to target bacteria with the type 1701 morphology (Thomsen et al., 2004). Two variants were observed with this probe: one resembles type 1701 and the other is more similar to type 0041/0675. The 1701 morphotypes are frequently observed in low loaded plants with N removal and biological P removal but only occasionally cause bulking there (Thomsen et al., 2006b).

We recommend applying Curvi997 and respective competitor probes for the Curvibacter-related type 1701. However, this probe targets other filament morphotypes resembling type 0041. As type 1701 is easily confused with several other morphotypes, its staining behavior (1701 Gram-, 0041/0675 Gramþ) and morphological features under phase contrast should first be checked carefully (e.g. rod-shaped cells in short filaments (1701) or square cells in long filaments (0041). In addition probe SNA should also be applied for type 1701 strains affiliated with Sphaerotilus spp. (Wagner et al., 1994a). See also FISH images in Figures 9.52–9.53, Chapter 9.

Table 5.13. Morphotype 1701: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability for bacteria with attached growth.


## 5.3.13 1851 morphotype

#### Abundance, morphology and affiliation

The 1851 morphotype is commonly seen in WWTPs all over the world, although not associated with bulking incidences in the US (summarized in Jenkins et al., 2004). It is found in municipal and industrial WWTPs with and without nutrient removal (Eikelboom, 2000; Beer et al., 2002; Jenkins et al., 2004; Kragelund et al., 2007a), and is responsible occasionally for severe bulking and foaming incidents in German municipal low F/M WWTPs (Lemmer et al., 2000). FISH surveys have also been conducted using probes targeting morphotype 1851, where long straight to slightly bent filaments forming twisted bundles were detected predominantly in municipal WWTPs in Australia, New Zealand, the Netherlands, and France (Beer et al., 2002), Japan (Kohno et al., 2002), and in German WWTPs with scumming events (Mu¨ller et al., 2007). Morphotype 1851 has also been observed in plants with high industrial loads and can be observed both with and without attached growth (Kragelund et al., 2007a).

This morphotype shows straight or slightly bent long filaments that often form characteristic bundles. Its cell diameter is between 0.6 and 0.8 mm. Attached growth is nearly always present in filaments seen in domestic plants, but is often missing in those from industrial WWTPs. It stains Gram positive, characteristically as light blue, and is usually Neisser negative.

Several morphotype 1851 strains have been identified as Kouleothrix aurantiaca in the phylum Chloroflexi (class Roseiflexus) after 16S rRNA sequence analysis of micromanipulated filaments from activated sludge samples (Beer et al., 2002; Kohno et al., 2002; Kragelund et al., 2007a).

#### Recommending the best probes

Table 5.14 shows the oligonucleotide probes for identifying morphotype 1851, and the tree in Figure 5.8 illustrates their specificity and coverage.

Two probes (CHL1851 and EU25-1238) were designed to target this morphotype in activated sludge (Beer et al., 2002; Kragelund et al., 2007a). A survey of samples from industrial WWTPs showed that approx. 50% of probe EU25-1238 positive filaments did not hybridize with the probe CHL1851 (Kragelund et al., 2007a), even though the EU25 isolate possessed the probe target site at E. coli position 592.

Morphotype 1851 strains from predominantly municipal WWTPs are readily detected with the probe CHL1851. However, we recommend using probe EU25-1238 for detecting this morphotype in WWTPs with both high and low loadings of industrial wastewater. Many of the EU25-1238 probe-positive filamentous bacteria in industrial WWTPs have little or no attached growth. See FISH images in Figures 9.54–9.56 in Chapter 9.

## 5.3.14 1863 morphotype

#### Abundance, morphology and affiliation

Morphotype 1863 is observed in WWTPs only occasionally and seems to play no important role in bulking (Jenkins et al., 2004; van der Waarde et al., 2002). Type 1863 and unicellular Acinetobacter cells can be detected in some foams or scums since these produce often hydrophobic capsular material and are known to release biosurfactants under nutrient limitation (Lemmer et al., 2000; Mu¨ller et al., 2007). In a survey of municipal WWTPs 60% contained Gram negative coccoid cells and 30% had type 1863 filaments which were detected by the probe ACA23a (Wagner et al., 1994b) specific for the morphotype 1863 affiliating with the genus Acinetobacter in the Gammaproteobacteria (Mu¨ller et al., 2007). However, it also targets straight filaments with square-shaped cells (Schade M., unpublished results). When this probe was applied to samples from industrial WWTPs (van der Waarde et al., 2002), the morphotype 1863 was detected in approx. 16% of plants. Their relative abundances were not reported (van der Waarde et al., 2002).

Table 5.14. Morphotype 1851: rRNA-targeted oligonucleotide probes with corresponding formamide concentration, target, morphotype, and comments on recommendability for bacteria with attached growth.


However, another survey where the abundance of morphotype 1863 in 126 industrial WWTPs was reported, only 4% of the plants contained this morphotype and the filament index never exceeded 1 (Kragelund C and Nielsen PH, unpublished results).

Type 1863 is characterized by bent, curled or tangled short filaments consisting of regular cells of 0.5 to 0.7 mm diameter. Cells are spherical to ovoid, and stain Gram and Neisser negative.

The morphotype 1863 is polyphyletic. Some isolates affiliate with members of the genus Acinetobacter in the Gammaproteobacteria (also targeted by class-specific probe GAM42a, Table 5.4) (Wagner et al., 1994b), while others affiliate with Chryseobacterium spp. within the Bacteroidetes (Seviour et al., 1997). It can also be targeted by probe CFB563, Table 5.4.

#### Recommending the best probes

Table 5.6 shows the oligonucleotide probes for identifying and quantifying morphotype 1863, and the tree in Figure 5.3 illustrates their specificity and coverage.

The probe ACA23a was designed to detect Acinetobacter spp. in the Gammaproteobacteria in activated sludge (Wagner et al., 1994b). Applied to municipal activated sludge samples ACA23a visualizes rods and/ or cocci in clusters and chains (Wagner et al., 1994b), see FISH image Figure 9.57 in Chapter 9. No probes have been designed for the type 1863 affiliating with Chryseobacterium spp. within the Bacteroidetes.

Morphotype 1863 never causes severe bulking problems. As it can be involved in scum formation we recommend that the ACA23a probe for Acinetobacter-related 1863 strains be used in such cases.

## 5.3.15 0803, 0914 morphotypes and other still unidentified filamentous species

#### Abundance, morphology and affiliation

A number of morphotypes are still currently unidentified (e.g. morphotype 0211, 0411, 0581, 0803, 0914, 0961, 1702 and 1802). However, many of these (morphotypes 0211, 0411, 0581, 0961, 1702 and 1802) are only rarely observed and therefore not considered important enough for inclusion in this manual. Some morphotypes observed more often in activated sludge like morphotypes 0803 and 0914 will be discussed.

Morphotypes 0803, and 0914 have been 'identified' by microscopy in activated sludge communities around the world, and are often seen in large numbers. Thus, morphotype 0803 was ranked as the fourth most abundant filament in Czech Republic, and morphotype 0914 ranked fifth (surveys summarized in Jenkins et al., 2004 and Tandoi et al., 2006).

Both type 0803 and 0914 form short (550 mm) straight, or longer (100–300 mm) bent filaments. Filaments are attached sometimes by a holdfast to inorganic particles. The cell diameter is between 0.8 and 1.1 mm. Cells are mostly square (0914) or rectangular (0803). With sulfur stores, type 0803 becomes difficult to distinguish from type 0914 according to the descriptions of Eikelboom (2000). However, different cell shape and cell diameter are found in situ and can be used for their differentiation. These filaments can be Gram variable, but are more usually Gram negative, and occasionally feature Neisser positive granules.

The phylogenetic affiliation is still not resolved for any of these morphotypes. Partial 16S rRNA gene sequence analysis from a micromanipulated morphotype 0803 isolate showed an affiliation with the Betaproteobacteria (Bradford et al., 1996). However, no in situ hybridization with the general betaproteobacterial probe BET42a is observed (Kragelund C, unpublished results). So the phylogenetic affiliation of type 0803 is still uncertain. Morphotype 0914 is still unidentified and to our knowledge no successful attempts have been made to sequence the 16S rRNA gene (either from an isolate or from micromanipulated filaments). However, very recent results indicate that they may belong to the Chloroflexi, hybridizing with the probes GNSB941 and CLX 1223 (Seviour R, unpublished) and new probes are presently being developed.

## 6

## Identification of other microorganisms in activated sludge and biofilms by FISH

## Per Halkjær Nielsen and Jeppe Lund Nielsen

## 6.1 INTRODUCTION

Besides the functional groups described in previous chapters, many other groups of bacteria are present in wastewater treatment systems, although often in low numbers. In this chapter we have included a few groups that are present or important in many plants, they may have an important function, and relatively specific gene probes are available. However, we do not provide detailed recommendations, only up-to date references.

## 6.2 EPIPHYTIC BACTERIA INVOLVED IN PROTEIN HYDROLYSIS

In full-scale treatment plants several types of filamentous bacteria are often observed with a large number of epiphytic, attached bacteria. These are known as type 0041, type 1851, and type 1701 based on morphological identification (see Chapter 5 about filamentous bacteria). Most of these filamentous bacteria belong to the candidate phylum TM7, the phylum Chloroflexi, and the class Betaproteobacteria (related to Curvibacter) and can be detected by gene probes. Recently, it has been shown that most of their epiphytic bacteria belong to the Saprospiraceae in the phylum Bacteroidetes (Kong et al., 2007; Xia et al., 2008). A new genus Candidatus Epiflobacter was proposed for this epiphytic group. Three clusters were identified but no differences in physiology could be detected. These bacteria are specialized in protein hydrolysis and use amino acids as energy and carbon sources. They are not involved in

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

denitrification and cannot store polyphosphate or PHA. These epiphytic rods are typically perpendicularly attached onto the filaments and are 1.3–1.9 mm in length and 0.3–0.4 mm in width. They can be very abundant in BNR activated sludge treatment plants and account up to 10% of all bacterial biovolume.

Most of the epiphytic bacteria hybridize with probe SAP-309 targeting Saprospiraceae in the phylum Bacteroidetes, see Chapter 5. Probe Bac111 (Kong et al., 2007) is more specific for species found in wastewater treatment plants and targets also most of these epiphytic bacteria. Three probes exist for detection of three subgroups but as no differences in abundance or ecophysiology is known today, it is recommended to apply probe Bac111 to detect almost all epiphytic Saprospiraceae (Kong et al., 2007; Xia et al., 2008).

## 6.3 SULFATE-REDUCING BACTERIA

Sulfate-reducing bacteria are present in most activated sludge systems (Lens et al., 1995), and are physiologically active during anaerobic periods, for example, in the anaerobic reactors of EBPR plants or in sludge storage tanks (Rasmussen et al., 1994; Kjeldsen et al., 2004). Prolonged aeration may not lower their activity, and oxygen-tolerant sulfate-reducing Desulfovibrio strains isolated from activated sludge plants have been described (Kjeldsen et al., 2005; Mogensen et al., 2006). About 0.5–8% of the total bacteria have been identified as sulfate-reducers in full-scale WWTPs (Manz et al., 1998; Vester and Ingvorsen, 1998; Nielsen and Nielsen, 2002a). The dominant SRB belonged to the families Desulfovibrionaceae and Desulfobacteraceae (Manz et al., 1998).

Sulfate-reducing bacteria are a polyphyletic group with most known members in the Deltaproteobacteria. Their detection by FISH will not be described in detail in this book and the readers are recommended to use the studies by Manz et al. (1998) and Lu¨cker et al. (2007).

## 6.4 FERMENTING BACTERIA

Fermentation in activated sludge has been shown to be performed partially by bacteria belonging to the genus Streptococcus in the class Bacilli and the family Streptococcaceae (Kong et al., 2008). The physiology of the fermenting streptococci reveals a broad group of cells which are capable of partially oxidizing glucose to produce lactate, acetate, ethanol, and formate (Liu et al., 2002; Kong et al., 2008). These are facultative anaerobes, utilizing glucose under both aerobic and anaerobic conditions (but not mannose or galactose) (Nielsen PH, unpublished). They probably ferment substrates in the anaerobic reactors, thus being important in providing readily metabolizable substrates for the denitrifying and phosphate-removing communities in these processes. Streptococci are quite abundant in many municipal WWTPs with EBPR, constituting 1–4% of the biovolume (Kong et al., 2008). Some of these are filamentous, see Chapter 5.

The probe-defined consortia hybridizing with probe Str/Strept (Trebesius et al., 2000) can be used to identify fermenting Streptococci (see also Chapter 5). The probe targets 44% of all streptococci, and can be used together with the broader probe LGC354C (Meier et al., 1999). The morphologies of the Str/Strept-probed cells cover both small cocci and filamentous organisms. A broader probe, Strc498, targets most Streptococcus spp. (Franks et al., 1998), but has to our knowledge not been tested in wastewater systems.

## 6.5 ESCHERICHIA COLI AS INDICATOR ORGANISM FOR ENTERO-PATHOGENS

Fecal material from human and animals contain a large variety of pathogenic microorganisms. A systematic search of all potential strains of entero-pathogens is not feasible, thus enumeration of

gammaproteobacterial Escherichia coli has been used as indicator bacteria to evaluate the fecal content and the sanitary risk. Several attempts of using FISH to enumerate E. coli in municipal WWTPs have been performed. The need for accurate numbers of potentially active E. coli in natural waters has resulted in approaches by which E. coli has to be revived prior to FISH (Garcia-Armisen and Servais, 2004). Direct viable count (DVC) procedures involve exposing the cells to a revivification medium containing antibiotics preventing cellular division; elongated cells are then enumerated as viable cells.

Using such DVC-FISH approaches typically around 1% of the total cell number appear as E. coli (Garcia-Armisen and Servais, 2004), whereas very few studies have been able to enumerate the specific pathogens for which E. coli function as indicator (e.g. Campylobacter jejuni, Clostridium perfringens, Bacillus cereus, Enterococcus faecalis, E. coli O157:H7, Helicobacter pylori, Klebsiella pneumoniae, Legionella pneumophila, Listeria monocytogenes, Pseudomonas aeruginosa, Salmonella spp. and Staphylococcus aureus). However, from quantitative PCR these have been found to estimate several orders below that of E. coli (e.g. Shannon et al., 2007).

Escherichia in the activated sludge is believed to be coming from the incoming wastewater. To what extent these survive or even grow here is not known. The appearance under these conditions is rods, typically 1.0–2.3 mm in length and 0.4–0.5 mm in width.

Several probes have been used for the identification and quantification of E. coli. The probe colinsitu has a central mismatch towards the target site in the genera Escherichia and Shigella both within the family Enterobacteriaceae, but colinsitu has fewer outgroups hits compared to the perfect match probe Ec637 (Regnault et al., 2000). Since both probes yield good FISH signals the colinsitu probe may be recommended for detection of E. coli.

## 7

## Protocol for Fluorescence in situ Hybridization (FISH) with rRNA-targeted oligonucleotides

Jeppe Lund Nielsen

## 7.1 INTRODUCTION

Fluorescence in situ hybridization (FISH) using rRNA-targeted oligonucleotide probes has become one of the most widely used approaches when studying microorganisms directly in complex systems without prior cultivation and isolation. The method makes it possible within a relatively short time to retrieve information on phylogenetic identities of the cells directly in a sample, and since it also maintains the morphology of the cells it also supplies information on the spatial distribution as well as the number of identified organisms.

The first applications of in situ hybridization were applied using radioactively labeled oligonucleotides, which had limited microscopic resolution and needed autoradiography for visualization. The use of FISH with fluorescently labeled probes was first described in 1989 (DeLong et al., 1989) and has since become the method of choice for examining the microbial consortia in both natural and engineered systems. In wastewater microbiology, the FISH approach has been shown to be especially well-suited with high cell signals and detection sensitivities.

The principle of the FISH technique is based on hybridizing fluorescently labeled probes to ribosomal rRNA in permeabilized whole microbial cells. The probes consist of short pieces of DNA (usually 15–25 nucleotides in length) and they are designed to specifically hybridize to their complementary target

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

sequence on the rRNA structures (16S and 23S subunits are typically used for Bacteria, while for Eucarya the 18S subunit is the most used; Amann, 1995) in the target cell. From the composition of the probe, it is possible to design it to specifically target a narrow phylogenetic group (down to the species level) or any other higher phylogenetic hierarchical group (top to bottom approach, Amann et al., 1995). No probes will hybridize to those cells without target sequences. Cells containing the target sequence will on the other hand retain the hybridized probe and due to the large number of ribosomes within active cells thus become fluorescently labeled.

Several recent reviews have elegantly described the potentials of the FISH technique and new developments for improved sensitivity (e.g. Wagner et al., 2003; Zwirglmaier, 2005; Bottari et al., 2006; Amann and Fuchs, 2008). Often occurring pitfalls within the application of the FISH technique and solutions hereto have been reviewed by Moter and Go¨bel (2000), while important factors that have an influence on the sensitivity and quality of the obtained data are described by Bouvier and del Giorgio (2003).

The main obstacles associated with the FISH method are poor cell permeability, insufficient ribosome content, ribosome inaccessibility, and sample autofluorescence. Thus knowledge of the nature and applicability of the sample as well as a uniform protocol with application of the proper controls are of fundamental importance for obtaining solid and comparable information.

This chapter provides a detailed protocol for applying FISH on activated sludge samples and possible solutions to typical problems encountered. A typical flowchart for FISH in activated sludge or biofilm is shown in Figure 7.1 and includes sampling, possible pretreatment, fixation, possible enzyme treatment, hybridization, epifluorescence microscopy, and possible image analysis. Details are described below.

Figure 7.1. Schematic illustration of the FISH approach.

## 7.2 FISH PROTOCOL

## 7.2.1 Materials and solutions

#### 8% Paraformaldehyde/PBS (8% PFA)

Stir 4 g PFA in 30 mL dH2O at 60*–*C and dissolve with a drop of 2 N NaOH (in the fume hood!); add 16.6 mL 3· phosphate-buffered saline (PBS) (see below). Adjust with dH2O to 50 mL. The solution is filtrated through a 0.22 mm polycarbonate filter to remove autofluorescing particles. Should be used fresh, or stored in aliquots at 20*–*C (do not freeze thawed solutions).

#### 1:1 PBS/Ethanol (PBS/EtOH)

Mix equal volumes of 1·PBS and 96% ethanol, sterilize by filtration.

#### Gelatine solution

Dissolve 0.5 g gelatine and 0.01 g CrK(SO4)2 in 100 mL dH2O at 70*–*C. Store at 4*–*C.

### Poly-L-lysine solution

Dissolve 0.01 g poly-L-lysine in 100 mL dH2O.

#### Lysozyme

Dissolve lysozyme (Merck Chemicals, Denmark) to a final concentration of 10 mg/mL (,360 000 U/mL) in 0.05 M EDTA, 0.1 M Tris-HCl, pH 8.0 (prepare fresh when needed; keep on ice until use).

#### Mutanolysin

Dissolve mutanolysin (Sigma-Aldrich, St. Louis, MO, USA) to 5000 U/mL in 0.1 M potassium phosphate buffer, pH 6.2 (prepare fresh when needed; keep on ice until use).

#### Achromopeptidase

Prepare achromopeptidase (Sigma-Aldrich, St. Louis, MO, USA) solution to a final concentration of 60 U/mL in 0.01 M NaCl, 0.01 M Tris-HCl (pH 8.0) (prepare fresh when needed, keep on ice until use), alternatively, prepare aliquots in tubes and freeze until needed.

#### Proteinase K

Dissolve proteinase K (Tritirachium album, Sigma-Aldrich, St. Louis, MO, USA) to 20 000 U/mL in TE buffer (0.01 M EDTA, 0.1 M Tris-HCl, pH 8.0). (Prepare fresh when needed; keep on ice until use).

#### Lipase

on 04 February 2019

Dissolve lipase (Sigma-Aldrich, St. Louis, MO, USA) to 75 000 U/mL in 1·PBS (prepare as 3·PBS and dilute with dH2O) (Prepare fresh when needed; keep on ice until use).

#### 3· Phosphate-buffered Saline (3·PBS)

Mix 0.1 M NaH2PO4 with 0.1 M Na2HPO4 till pH 7.4. Use 300 mL hereof and mix with 22.8 g NaCl and adjust to 1000 mL with dH2O. Autoclave, and store at room temperature.

#### Tris-EDTA buffer (TE buffer)

0.01 M EDTA (ethylenediaminetetraaceticacid), 0.1 M Tris-HCl, pH 8.0. Sterilize by filtration and store at 4*–*C.

#### 1 M Tris-HCl, pH 8.0

Dissolve 121.1 g Tris in 800 ml dH2O, add 42 mL concentrated HCl, let cool, adjust pH and fill to 1L with dH2O. Autoclave, and store at room temperature.

## 5 M NaCl

Dissolve 292.2 g NaCl in 800 mL dH2O, fill to 1 L with dH2O. Sterilize by filtration, and store at room temperature.

## Sterile distilled H2O (dH2O)

Autoclave sterile filtered distilled H2O.

## 10% Sodium dodecylsulfate (10% SDS)

Heat 50 g SDS (electrophoresis-quality) in 400 mL dH2O to 70*–*C, adjust pH with concentrated HCl to 7.2, fill to 500 mL; no sterilization required. Store at room temperature.

## 0.5 M EDTA

Dissolve 18.6 g EDTA in 80 mL dH2O by adjusting the pH to 8.0 (ca. 2 g NaOH pellets required), fill to 100 mL with dH2O. Sterilize by filtration and store at 4*–*C.

## 7.2.2 Equipment and supplies needed for FISH


## 7.3 PROTOCOL

## 7.3.1 Sample collection and fixation

Fresh samples are collected and fixed as soon as possible (activated sludge and biofilms can usually be kept in the refrigerator 2–4*–*C for 2–3 days without major changes).

Sample preparation prior to hybridization with the fluorescently labeled probes involves fixation and cell wall permeabilization in order to inactivate the microbial cells and any enzymatic activity, to avoid growth/decay after harvest and to permeabilize the cells for probe penetration. The most commonly used fixative agent is polymerized formaldehyde (paraformaldehyde) which preserves the cell morphology, although some Gram positive cells are not adequately permeabilized with this fixative. Thus the sample is usually fixed independently for both Gram negative and Gram positive cells.

### Gram positive cells:

Mix equal volumes of sludge and 96% ethanol in a test tube. Keep the solution in the freezer (20*–*C).

### Gram negative cells:

Mix equal volumes of sludge with cold 8% PFA.

Incubate on ice for ½–3 hours, depending on sample texture and target organisms (some cells require short incubations in order to allow permeabilization of the probe, while other cells remain FISH-detectable even after several hours of fixation. Typically sludge or biofilm samples with flocculated material must be fixed for 2–3 hours to allow an even fixation of interior cells in the flocs/biofilms).

Separate PFA and biomass by centrifugation (3500 g for 8 min).

Dissolve the pellet in 5 mL cold 1:1 PBS/EtOH. Separate by centrifugation (3500 g for 8 min). Repeat this step once more.

Resuspend the pellet after the last step in 1–5 mL cold 1:1 PBS/EtOH and mix.

The fixed sample can be kept in the freezer (20*–*C) for several months.

## 7.3.2 Sample preparation

The sample can typically be applied either directly to the slide to maintain spatial resolution within the flocs, or else after gentle homogenization to get a higher resolution and thus more accurate enumerations (see also Chapter 8). Gentle homogenization can be obtained by gently rubbing two glass slides with 20 mL sample against each other, or more efficiently by a tissue grinder.

Cryosectioning can be carried out to maintain high resolution and still keep the spatial resolution. Cryosectioning can be obtained by embedding the sample in paraffin or in cold polymerizing resin (Tissue-Tek O.C.T.). Paraffin embedding usually provides accurate and homogenous slices without breakage of the biomass sample, but requires heating the sample and de-waxing with xylene, while cold polymerizing resin application can be carried out at low temperature and does not require further chemical treatment.

The biofilm/sludge samples (e.g. in the lid of an Eppendorf tube) are covered by embedding material (Tissue-Tek O.C.T.) which is allowed to migrate into the sample overnight and polymerize at 4*–*C, before transferring it to liquid nitrogen. Sectioning into 5–20 mm thin slices is carried out on a cryotome at 20*–*C (e.g. Dublier et al., 1995; Moter and Gobel, 2000; Gieseke et al., 2005). The slices are immediately placed on a slide (room temperature), where they melt. The slices are allowed to dry on the bench for 3 hours.

## 7.3.3 Immobilization of the samples on glass slides

To provide the best possible adhesion without loss of sample during the further treatments and to obtain an even and homogeneous distribution of the sample, it is recommended to use pretreated glass slides. The best results are usually obtained on a relatively hydrophilic surface, which can be obtained by acid-washed slides and slides coated with gelatin or poly-L-lysine. An acid wash before coating is recommended.

#### Acid wash procedure

Place the slides in a beaker containing preheated (60*–*C) 1 M HCl solution. Let the slides stand for at least 8 hours. After cooling the slides are carefully rinsed in dH2O. Rinse the slides in 95% ethanol and let dry.

#### Procedure for gelatine coating of microscopic slides

Rinse the slides (either normal glass slides or Teflon coated slides) in 70% ethanol, let air dry, dip in 0.5% gelatine solution for 5 min at 70*–*C.

Remove the slides and let them air dry in vertical position in a dust-free environment.

#### Procedure for Poly-L-Lysine coating of microscopic slides

The acid-washed slides are incubated in 0.01% poly-L-lysine solution for 5 min at room temperature. Remove the slides and let them air dry in vertical position in a dust-free environment.

#### Immobilization of the sample on slides

Place 5–15 mL of the sample on a slide, cover glass or in each well in a teflon-coated slide. Use the side of the pipette tip to spread out the sample. Let the samples dry in the fume hood until they are completely dry (15–30 min). Drying at elevated temperature (40–60*–*C) improves the binding of dilute samples to gelatine-coated slides.

## 7.3.4 Dehydration

Dehydration of the immobilized (untreated, homogenized or cryosectioned) sample removes water from the sample in order to increase the resolution during microscopy. Serial treatment with increasing ethanol concentrations efficiently removes the water in the sample rendering the final sample with a decreased thickness down to almost 25%. Dehydrate the slides with the immobilized sludge samples by dipping it into a container with 50% ethanol for 3 min, followed by 3 min in 80% ethanol and by 3 min in 96% ethanol. Let the slides air-dry before continuing with the permeabilization step.

## 7.3.5 Permeabilization by enzymatic or chemical treatment

Some cell types require additional treatment by enzymes or chemicals for sufficient permeabilization over the cell wall. Several enzymes have been applied and must be individually tested for new cell types, although a few guidelines exist. Lysozyme treatment makes many Gram positive cells permeable, some require treatment with acid, lipase or mutanolysin (Davenport et al., 2000). Some organisms (e.g. certain Archaea) are not readily amendable to the lysozyme permeabilization strategy and require protease treatment, e.g. protease K, mutanolysin or achromopeptidase, while others (e.g. Mycolata) require a mixture of the various enzymatic treatments and/or mild acid hydrolysis pre-treatment (e.g. Carr et al., 2005, Kragelund et al., 2007b). Digestion with enzymes, too long incubations or treatment with too harsh chemicals is critical and will cause lysis of some more sensitive cells. It might thus be incompatible to access multiple organisms in complex environments. Therefore always the gentlest treatment possible should be used.

### Procedure for permeabilization with Lysozyme

Start with immobilized samples dehydrated in EtOH series.

Apply 10–15 mL cold lysozyme (36 000–360 000 U/mL) per slide (or well on slide) and place the slide in a horizontal position, in a 50 mL polyethylene tube, containing tissue paper with 2 mL of dH2O.

Incubate the slides for 10–60 min at 37*–*C.

Wash the slide three times in dH2O, then once in absolute ethanol, and let the slide air-dry. After permeabilization, the slide can be stored at –20*–*C for several months.

### Permeabilization with Mutanolysin

Apply 10–15 mL cold mutanolysin solution (5000 U/mL) in per slide (or well on a slide) and place the slide in a horizontal position, in a 50 mL polyethylene tube, containing tissue paper with 2 mL of dH2O.

Incubate the slides for 10–30 min at room temperature.

Wash the slide three times in dH2O, then once in absolute ethanol, and let the slide air-dry. After permeabilization, the slide can be stored at –20*–*C for several months.

### Procedure for permeabilization with Proteinase K

Start with immobilized samples dehydrated in EtOH series.

Apply 10–15 mL cold proteinase K (2000–20 000 U/mL) per slide (or well on slide) and transfer the slide to a 50 mL polyethylene tube with a moisturized piece of tissue paper. Incubate the slide for 20–60 min at 37*–*C.

Wash the slide three times in dH2O, then once in absolute ethanol, and let the slide air-dry. After permeabilization, the slide can be stored at –20*–*C for several months.

#### Permeabilization with Achromopeptidase

Apply 10–15 mL cold achromopeptidase solution (60 U/mL) per slide (or well on slide) and place the slide in a horizontal position, in a 50 mL polyethylene tube, containing a piece of tissue paper with 2 mL of dH2O.

Incubate the slides for 20–60 min at 37*–*C.

Wash the slide three times in dH2O, then once in absolute ethanol, and let the slide air-dry. After permeabilization, the slide can be stored at –20*–*C for several months.

#### Permeabilization with Lipase

Apply 10 ml (75 U/mL) per slide (or well on a slide) and place the slide in a humid chamber containing adsorbent paper wetted with PBS to stop evaporation of the enzyme solution.

Incubate the slides for 60 min at 37*–*C.

The lipase solution is removed after incubation by washing the slide in dH2O. Then wash the slide once in absolute ethanol, and let the slide air-dry.

After permeabilization, the slide can be stored at –20*–*C for several months.

#### Permeabilization with mild acid hydrolysis

The microscopic slides spotted with dehydrated cells are subjected to hydrochloric acid (1M HCl) at 37*–*C for 30 min.

Wash the slide with dH2O, then once in absolute ethanol, and let the slide air-dry.

## 7.3.6 Preparation and quality check of probes

Commercially acquired probes typically arrive lyophilized and must be resuspended in TE buffer. Stock solutions are diluted to contain 5 mg/mL. Working solutions are prepared at a concentration of 50 ng/mL (by diluting stock solution 1:100 in dH2O) and stored in small portions (50 to 100 mL) at 20*–*C.

To test if the probe concentration corresponds to the concentration claimed by the manufacturer the probe is suspended in 50 mL of dH2O. The absorbance of the 1:100 diluted stock solutions (in dH2O) are then measured at 260 nm (1 A260 < 20 mg/mL of single-stranded DNA oligonucleotide).

To check the quality of the labeling of the oligonucleotide, the ratio of absorption of fluorochrome and DNA (presuming optimal labeling), i.e. A260/A550, should amount to < 1 for a Cy3-labeled 18-mer probe.

## 7.3.7 Hybridization

Hybridization of the oligonucleotide probe to the ribosome in the target cell must be carried out under high stringency, with defined conditions regarding preheating temperature, ionic strength, and proper formamide concentrations. By providing the proper conditions, sufficient stringency can be obtained to discriminate between single mismatches in the hybridized probe-ribosome complex.

Prepare 2 mL of hybridization buffer for each percentage of formamide (see Table 7.1). The formamide concentration must reflect the empirically optimized conditions for the applied gene probes. For details on many published probes see chapters elsewhere in this book or visit probeBase (http://www.probeBase.net; Loy et al., 2007).


Table 7.1. Composition of the hybridization buffer at different formamide concentrations.

Hybridization procedures are typically designed to occur at 46*–*C, but can be carried out at other temperatures by changing the matching formamide concentration according to the relation: 1% FA ¼ 0.65*–*C.

Transfer 8 mL of hybridization buffer onto the slide within a small area or into each well on a tefloncoated slide. Prepare one slide at a time to avoid evaporation of hybridization buffer and thus changed stringency.

Add 1 mL of each gene probe (work solution 50 ng/mL) and mix carefully (avoid contact with the sample) with the hybridization buffer (sterile pipette tips must be used for all work with gene probes). If more gene probes are added to the same well the order is in the exact same manner – 1 mL of each of the probes is added to the well. Equimolar concentrations of each competitor probe are added if needed.

Place the slide horizontally into a 50 mL polyethylene tube with a piece of tissue wetted with 1–2 mL of hybridization buffer. Place the tube in the hybridization oven (46*–*C) for at least 1½ hours (increased hybridization to less accessible regions of the ribosome has been shown to occur upon hybridization for up to 72 hours, see also Chapter 5).

Probes with different T<sup>d</sup> (requiring different formamide concentrations) cannot be applied together, but must be applied in a double hybridization with two subsequent hybridizations starting with the highest formamide concentration.

During hybridization prepare the washing buffer in a 50 mL polyethylene tube (formamide is replaced by NaCl, according to Table 7.2). Preheat the washing buffer in a 48*–*C water bath.

Gently rinse the slide by pouring a few millilitres of the preheated washing buffer over the sample (must be carried out in the fume hood to avoid toxic formamide fumes).

Transfer the slide to the 50 mL polyethylene tube with the remaining preheated washing buffer, and incubate for 15 min at 48*–*C (water bath).

Remove the slides from the washing buffer and dip in cold dH2O. Let the slide air-dry.


Table 7.2. Composition of the washing buffer corresponding to the formamide concentrations applied during the hybridization.

## 7.3.8 Counterstaining with DAPI

After hybridization the slide can be counterstained, for example, with the DNA stain 4<sup>0</sup> ,60 -diamidino-2 phenylindol (DAPI) to determine the fraction of FISH positive out of the total DAPI count (optional).

Add DAPI solution (1 mg/mL in distilled water) to cover the sample, and stain for 15 min at 4*–*C in the dark.

Rinse with plenty of dH2O, and let the slide air-dry.

Please note that DAPI is staining DNA only, for other counterstaining methods see Chapter 8.

## 7.4 MICROSCOPY

Evaluation of the FISH signal is performed after embedding in a small drop of mounting agent containing an antibleaching agent (e.g. Citifluor or Vectashield). Volumetric mixtures of Citifluor and Vectashield are preferred by several users (optional).

The embedding medium can also be directly amended with DAPI (final concentration 1 mg/mL). This usually yields strong signals but renders a strong background fluorescence. Stained preparations can also be stored at –20*–*C without substantial loss of signal intensity.

Upon examination of multiple stained slides always examine the longer wavelength first (less energetic) to avoid bleaching.

When examining new samples always include a nonsense probe (Non-EUB338) using the same strong fluorescing fluorochrome as the one used with the specific probes (e.g. Cy3 or Alexa-dyes). A detailed review of the factors influencing the sensitivity of FISH, effect of fluorochrome type, and stringency conditions are described by Bouvier and del Giorgio (2003).

## 7.5 RECOMMENDATIONS AND TROUBLESHOOTING

Here follows a list of typical problems encountered when working with the FISH technique such as low signals and problems related to the microscopic examination.

#### Test of new oligonucleotide probes

Several software programs exist for design of new probes (e.g. ARB software package (http://www.arbhome.de), and Primrose (Ashelford et al., 2002)). Several publications describe the choice of parameters

and pitfalls during the process of designing new probes (e.g. Amann and Ludwig, 2000) and only a few recommendations for the testing will be given here.


### Microscopy



#### Low signals or no signals



Table 7.3. Most commonly used fluorescent dyes to label oligonucleotides for FISH analysis. For optimal filters to visualize the above given dyes see the webpages of the main suppliers (e.g. http://www.zeiss.de; http://www.chroma.com).

<sup>a</sup> emission in the infrared area, it thus requires a digital camera detecting infrared light and image analysis to be observable

#### General planning


## 8

## Quantitative FISH for the cultivationindependent quantification of microbes in wastewater treatment plants

Holger Daims

## 8.1 INTRODUCTION

In many fields of microbial ecology and environmental engineering, including research related to wastewater treatment, not only the cultivation-independent detection but also the in situ quantification of uncultured microorganisms are important tasks and the main reasons to use molecular biological methods. For example, monitoring the abundances and population dynamics of selected microbes in activated sludge can give important clues on correlations between problems such as sludge bulking and the microbial community composition. As outlined in the other chapters of this book, fluorescence in situ hybridization (FISH) with rRNA-targeted probes is a very powerful tool and the method of choice for the cultivation-independent detection and identification of microorganisms in activated sludge and biofilm from wastewater treatment plants (WWTPs).

A key advantage of FISH over other methods, for example those based on PCR, is the direct visual feedback one gets during microscopic observation of the analyzed samples. This does not only give more confidence regarding the accuracy of an experiment, but it is also the basis for extensions that convert FISH into a precise method to quantify microorganisms in environmental samples. Based on the rule ''count what you see'', quantitative FISH often is less prone to biases than quantitative PCR approaches, which are quite sensitive to methodical errors (e.g. low PCR efficiency or stochastic

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

tube-tube variations) and laboratory contaminations with tiny amounts of template DNA. Furthermore, PCR-based approaches do not actually quantify microbial cells, but measure copy numbers of marker genes. The number of marker gene copies per genome, and even the number of genomes per cell, can vary among the targeted microbes (Ludwig and Schleifer, 2000). Thus, measuring these copy numbers by PCR allows only indirect conclusions on cell densities and on important parameters such as substrate turnover rates per cell. In contrast, quantification methods based on FISH directly measure cell numbers or equivalent quantities such as biovolume. A disadvantage of FISH is its lower sensitivity compared to quantitative PCR: while PCR can (theoretically) detect a single microbial cell, the detection limit of FISH is as high as 10<sup>3</sup> –10<sup>4</sup> cells per ml of suspended activated sludge. On the other hand, less frequent populations are unlikely to be functionally relevant for the wastewater treatment process. Cultivation-based methods, for example most-probable number (MPN) counting, also are potentially more sensitive than FISH but do not correctly quantify aggregated cells and fail to detect all organisms that do not grow in the offered nutrient media. Thus, FISH is the best currently available technique for quantifying microbes in lab-scale or pilot-scale reactors and in samples from full-scale WWTPs.

This chapter briefly summarizes approaches to quantify microbes by FISH with special focus on combining FISH with digital image analysis. A protocol is provided that uses a slightly modified FISH protocol and free open-source software for measuring the abundances of probe-target populations in activated sludge and biofilm samples.

## 8.2 QUANTITATIVE FISH: A BRIEF OVERVIEW

The simplest approach to quantitative FISH is direct visual counting of probe-stained microbial cells under the fluorescence microscope. For this purpose, FISH is performed according to the standard protocol as described elsewhere in this book. Subsequently, the probe-target cells are counted in several microscope fields of view (FOV). Unfortunately, the obtained numbers can hardly be related to any useful reference value, such as sludge volume, if complex samples like activated sludge or biofilm are analyzed this way. This is due to (i) difficulties of applying an exact known amount of flocculated biomass onto a glass slide prior to FISH, and (ii) variations in biomass content between different FOV. Regarding point (i), please note that by FISH only a few microlitres of sludge are analyzed in each experiment. Even though standard methods allow the determination of biomass concentration for larger volumes, variations due to pipetting errors and differences in the sizes of individual flocs become a major source of imprecision when very small volumes are handled. Therefore, most past studies determined relative rather than absolute abundances of microbes in activated sludge. In this case, the sample is hybridized to a specific probe and also to a 'universal' second probe or probe mixture, which targets all Bacteria, like the EUB-mix (Daims et al., 1999). Alternatively, the 'universal' probe(s) can be replaced by a nucleic-acid staining dye such as DAPI. Following FISH, the cells stained by the specific probe and the cells stained by the 'universal' probe or dye are counted. This is done for several FOVs and the results are summed up to determine the fraction of the specifically stained cells relative to all microbes at the end. This technique has been used, for example, to quantify non-aggregated cells in activated sludge (Wagner et al., 1994b). While simplicity is its greatest advantage, this method is very time-consuming and tedious, because several hundreds or even thousands of cells must manually be counted in several FOVs in order to obtain statistically robust results. It is not suitable to quantify cells in aggregates, because a precise counting of clustered cells, by eye and in many FOVs, is hardly feasible. The disruption of tight cell aggregates in activated sludge to facilitate counting is not possible either (Manz et al., 1994; Wallner et al., 1995).

designed a fully automated setup that linked image acquisition by microscopy to an image analysis pipeline for counting planktonic bacterial cells in ocean water samples. This sophisticated approach was much more rapid than manual counting and enabled the analysis of 28 samples within only 4 hours. Unfortunately, microbes that cluster in biofilms cannot precisely be counted by this and related methods. Prior to cell counting by image analysis, the images must be segmented by the computer – that means, the machine must detect the objects to be counted in the images. In case of planktonic cells, which do not form tight aggregates, every detected object corresponds to one single cell (or very few attached cells). In contrast, the individual cells embedded in tight cell clusters cannot be distinguished by image segmentation in most cases, because in the images there is no background that would separate these cells from each other. Thus, a fully automated quantification of cells in sludge flocs or biofilms usually is not possible.

An alternative solution is to measure the biovolume fraction of the targeted microbial population relative to the total microbial community. This approach was developed and first applied on activated sludge by Bouchez et al. (2000) and Schmid et al. (2000). The sludge sample is hybridized to a specific probe and a 'universal' probe (set) as described above for manual counting. Subsequently, digital images of either probe signal are recorded for 20–30 FOVs. This is best accomplished by using a confocal laser scanning microscope (CLSM) to avoid blurred images due to light from outside the focal plane. However, a normal epifluorescence microscope with a CCD camera can be used if low amounts of biomass are applied onto the slides to obtain thin layers of biomass, which are completely within the depth of focus of the microscope. An experienced user can record the images needed to quantify one population in about one hour. Importantly, the FOVs must be randomly selected in the x, y and z dimensions (by panning and zooming with the microscope). When all images are taken, the areas of the probe-target organisms are measured, by image analysis, in the images showing the signal of the specific probe and are summed up to obtain a total measured area of the probe-target population. Likewise, the areas of all organisms detected by the 'universal' probe are measured in the respective images. The area fraction of the specifically targeted population is then calculated as percentage of the total area of all organisms. Although these measurements are made with two-dimensional images (not with confocal z-stacks), the determined area fraction is an estimate of the biovolume fraction of the quantified population (Daims and Wagner, 2007). Notably, this approach does not require the detection of individual cells during image segmentation. Areas in images can be measured for single cells as well as for filaments or cell aggregates. This greatly simplifies image segmentation and allows even densely packed biofilm bacteria to be quantified.

The primary result of this 'biovolume-based' quantification method is a relative quantity, the biovolume of the probe-target population as percentage of the total microbial biovolume in the sample. This is sufficient for many applications such as monitoring population dynamics. However, absolute cell numbers are needed to estimate substrate turnover rates per cell, and also to compare different sludge samples with each other (because the total biomass content of these samples may be different). Absolute cell numbers can be obtained by an extended 'biovolume-based' approach called 'Spike-FISH', which uses an internal standard to convert biovolume fractions to absolute cell numbers (Daims et al., 2001b).

A more detailed description of the quantification methods mentioned here, and of additional techniques, is provided by Daims and Wagner (2007). The next section explains how the 'biovolume-based' quantification method is used in practice to determine relative amounts of probe-defined bacterial populations in activated sludge.

## 8.3 A PROTOCOL FOR QUANTITATIVE FISH AND IMAGE ANALYSIS TO MEASURE BIOVOLUME FRACTIONS

#### Sample preparation

For quantitative FISH follow the general sample fixation procedures described in this book in Chapter 7. Gram negative and Gram positive cells should be quantified in separate steps using formaldehyde-fixed or ethanol-fixed sample aliquots, respectively. Formaldehyde fixation usually applied to preserve Gram negative cells renders most Gram positive bacteria impermeable to FISH probes. In contrast, ethanol fixation used for Gram positive cells does not preserve Gram negative cells for extended time periods and when these cells start to lyse, the quantification results will most likely become biased. Moreover, if additional enzymatic pre-treatment is needed to permeabilize Gram positive cells, the Gram negative microbes are easily destroyed (lysed) by the enzymes. In cases where only ethanol-fixed samples are available, these samples should be analyzed within 1–3 days after sampling and fixation to reduce biases due to lysed Gram negative cells.

#### Sample immobilization

For standard FISH only small amounts (5–15 ml) of an activated sludge sample are immobilized on glass slides. This is sufficient for purely qualitative analyses, but can cause biases in quantifications due to single cells falling down onto the slide surface (Figure 8.1A). To avoid this problem, thick layers of sludge sample must be applied onto the slides (Figure 8.1B). First, put 15 ml of fixed sludge onto the slide and dry the sludge at 46*–*C or at room temperature. Subsequently, add another 15 ml aliquot and dry again. These steps should be repeated two to four times to obtain a thick layer of biomass (Figure 8.1B). With this large amount of sludge on the glass slide, additional measures are needed to avoid a loss of biomass during the hybridization and washing steps. First, poly-L-lysine-coated slides should be used to improve sludge adhesion to the slide surface. Second, the dried biomass should be covered with a thin agarose layer, which will act like a glue and keep the biomass attached to the slide (Figure 8.1C). For this purpose, dip the slide (with the dried biomass) horizontally into hand-warm dissolved 0.5–1% agarose, leave it for a few seconds in the agarose, and then cool the slide immediately in the horizontal position (biomass side up) on ice. When the agarose has solidified, the usual dehydration in 50, 80 and 96% ethanol is performed and the slide is dried again at 46*–*C. The combination of dehydration and drying will shrink the agarose so that a thin, glue-like agarose film is obtained on top of the sludge sample (Figure 8.1C). Finally, wipe off excess agarose from the bottom side of the slide. Although agarose may show autofluorescence, the very thin agarose layer does usually not interfere with fluorescence microscopy.

#### In situ hybridization

Hybridization, washing, drying, and embedding the agarose-covered slide in antifadent are carried out according to the standard FISH protocol. For quantitative FISH, one must apply (i) a specific probe that targets the population(s) to be quantified, and (ii) a 'universal' probe set that detects most Bacteria, which usually is the EUB-mix (Daims et al., 1999). Probes (i) and (ii) must be labeled with different fluorochromes. The 'universal' probe can be replaced by a nucleic acid stain, but be sure to use only dyes that stain both RNA and DNA. For example, DAPI is not useful for this kind of quantification, because DAPI stains DNA only. Unlike ribosomes, DNA is not evenly distributed inside most bacterial cells. Therefore, DAPI-stained cells look smaller than cells labeled by FISH and their projected area in the recorded images is smaller. Clearly, this biases a quantification technique based on measuring cell areas in the images. In contrast, SYBR Green stains RNA as well as DNA and can be used for this quantitative FISH approach.

Figure 8.1. A. Low amount of activated sludge biomass on a glass slide. During dehydration for FISH, single cells (light grey) may fall down onto the glass slide, whereas cell aggregates (dark grey) stay in place. Depending on the focal plane of the microscope (stippled line), mainly aggregates or single cells are captured in the images. B. More biomass on the slide prevents most small cells from falling down onto the glass surface. C. Covering the biomass with a thin agarose layer as 'glue' prevents biomass detachment from the slide during the dehydration, hybridization, and washing steps.

#### Image acquisition

Images for quantitative FISH should be recorded by using a CLSM. Take images of 20–30 FOVs, which must be chosen by randomly moving the slide in the x/y dimensions and by randomly adjusting the focal plane of the microscope in the z dimension. 'Randomly' means: do not choose the FOVs based on the presence of the population to be quantified! This would bias the analysis. Make sure that each FOV contains some microbial biomass (not necessarily the quantified population), but do not use any other selection criteria.

For each FOV, take one image of the population-specific probe and one image of the 'universal' probe (or nucleic acid stain). Use a low magnification (400*£* or lower) to capture as much biomass as possible in each image. This will improve statistics. Cells of the quantified population will appear in either image of an 'image pair' (specific probe and EUB-mix; Figure 8.2) in different colors. The detector of the CLSM must be adjusted properly to ensure that these cells have the same size (area) in both images. If this condition is not met, the analysis will most likely be biased. Please keep in mind that this quantification method is based on measuring the areas of cells in the images and on calculating area fractions (see above).

It makes sense to store the image pairs with numbered filenames to facilitate later analysis. For example, the two images Gamma\_001.tif and EUB\_001.tif would belong to FOV #1 (with Gammaproteobacteria as the organisms to be quantified).

#### Image analysis

on 04 February 2019

Measuring the areas of the probe-stained cells essentially means counting and summing up the pixels that belong to these cells in each image. First, the images must be segmented (see also below) to distinguish cells (¼'objects') from background. Second, the pixels that belong to objects must be counted.

Commercial image analysis programs, which are capable of automated image segmentation and calculations such as area measurements, are mostly very expensive. But for quantitative FISH there is no need to invest in such software, because some free software packages are available that can fulfil these tasks as well. Examples are the programs CMEIAS (Liu et al., 2001a), ImageJ (http://rsbweb.nih.gov/ij/),

*n i*=1 ∑ *AP i n i*=1 ∑ *AB i*

Figure 8.2. Principle of approach for quantifying the biovolume fraction of a probe-defined target population. See text for details.

and daime (Daims et al., 2006). While ImageJ is a general-purpose image analysis software, CMEIAS and daime have been developed having applications in microbiology in mind.

This chapter focuses on using daime for quantitative FISH. This program (''digital image analysis in microbial ecology''¼ daime) has been written especially for analyzing images of microbial cells as obtained by FISH and other fluorescence labeling techniques. It allows images to be segmented by using different intensity thresholding and edge detection algorithms. If needed, background can be reduced prior to segmentation. Segmented images can be opened in an 'object editor', which allows the user to manually refine the automated segmentation results and to remove fluorescent artefacts (such as plant material) from the images before the actual analysis starts (Figure 8.3). daime offers various image analysis routines: morphometry of single cells or cell aggregates, biovolume fraction measurement for quantitative FISH, fluorescence intensity measurements, automated analysis of formamide concentration series for FISH probe evaluation, and spatial statistics to determine if two microbial populations co-aggregate in biofilm samples (Daims et al., 2006; Maixner et al., 2006). In addition, daime contains a volume-rendering module that can display confocal image stacks in 3D at interactive rendering speed on current PC hardware.

More information about daime and its features can be found on the daime website (www.microbialecology.net/daime) and in the user manual. The program and the manual can be downloaded free of charge. daime runs on Windows 2000 and Windows XP (Vista not yet tested) and Linux (the source code is available for compiling daime on Linux systems). Appropriate graphics hardware and sufficient RAM memory (at least 1 gigabyte) are recommended for optimal performance.

To analyze the images of the different FOVs, which have been acquired in the previous steps, import them into daime as 2D image series. For this purpose, the image files should have been saved to disk with numbered filenames (as suggested above) and in the TIFF image format (8 bits per pixel, greyscale, or

Figure 8.3. Screenshot of the 'object editor' of the daime software. This dialog window offers several tools for identifying artefacts in images of probe-labeled microbial cells, for selecting objects based on morphological and other features for subsequent analysis, and for the manual refinement of automated image segmentation.

24 or 32 bits per pixel, RGB color). Then choose the option 'Stereology: Biovolume fraction' in the 'Analysis' menu. Indicate which image series shows the population to be quantified, and which one shows all cells detected by the 'universal' probe. The images will then be segmented automatically, and the object editor will open with both image series loaded (Figure 8.3). Push 'OK' if the segmentation results look satisfying. The areas will now be measured in all images, and the results window will open (Figure 8.4). As described above, the area fraction determined from all images is an estimate of the biovolume fraction of the quantified population. In addition to this value, daime provides some other parameters and statistics that allow the reliability of the quantification to be assessed (Figure 8.4). This depends mainly on the quality of the input images and on the success of image segmentation.

## 8.4 CONCLUDING REMARKS

With today's microscopes and computers, the FISH-based quantification of microbial populations in activated sludge is a straightforward task. However, like any other laboratory technique this approach has pitfalls and suffers from potential biases. Some of these problems have been addressed in the sections above. Here are a few general suggestions on how to optimize quantitative FISH analyses:

. Make sure that the basic method, FISH itself, works properly with your samples. Always test the applied oligonucleotide probes with samples known to contain the target organisms

Figure 8.4. Screenshot of the results window shown by the daime software after biovolume fraction measurement. Cumulative average analysis and other parameters inform on the reliability of the quantification results.

(positive control). Also use standard negative controls for FISH (e.g. nonsense probes) to test whether autofluorescence or unspecific attachment of probes to cells and to other surfaces might hamper quantification by FISH and image analysis. Typical caveats of FISH and solutions have been addressed elsewhere (Wagner et al., 2003; Daims et al., 2005).


## 9 Color image section

## Nitrifiers

Figure 9.1. Different populations of ammoniaoxidizing bacteria (AOB) simultaneously detected in a nitrifying biofilm from a sequencing batch biofilm reactor with probes NmV (Cy3; red), Nso1225 (Cy5; blue), and NEU (FLUOS; green).

Figure 9.2. Cell aggregate of nitrite-oxidizing Nitrospira detected with probe Ntspa662 (Cy3; red) in a nitrifying biofilm from a sequencing batch biofilm reactor. Bar ¼ 10 mm.

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

Figure 9.3. Three populations of nitrifying bacteria, detected simultaneously in a nitrifying sequencing batch biofilm reactor with ammonia-oxidizing bacteria (AOB) of the Nitrosomonas oligotropha lineage detected by probe Cluster6a192 (Cy5; blue), nitriteoxidizing bacteria (NOB) of Nitrospira sublineage I probe Ntspa1431 (Cy3; red), NOB of Nitrospira sublineage II probe Ntspa1151 (FLUOS; green). Bar ¼ 10 mm.

Figure 9.4. Anaerobic ammonium-oxidizing (anammox) bacteria, Candidatus Brocadia anammoxidans, detected in an anammox enrichment culture established from a wastewater treatment plant with probes Amx368 (Cy3; red) and EUB338 probe mix (FLUOS; green). Anammox cells appear yellowish due to simultaneous binding of Amx368 and EUB338, whereas other bacteria appear green due to binding of EUB338-I/II/III only. Bar ¼ 5 mm.

## Denitrifiers

Figure 9.5. Typical morphology of single cells and microcolonies of denitrifying Curvibacter spp. detected with probe Curvi997 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.6. Typical cell aggregates of denitrifying Azoarcus detected with probe Azo644 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.7. Typical cell aggregates of denitrifying Thauera spp. detected with probe Thau646 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.8. Typical morphology of cells belonging to the denitrifying Acidovorax spp. detected with probe Aci208 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.9. Typical morphology of single cells of denitrifying Hyphomicrobium spp. detected with probe HyphoCII-654 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.10. Typical cell structures of denitrifying cells belonging to the Rhodobacteraceae and Phyllobacteriaceae detected with probe GRb 208(Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.11. Typical cell aggregates of denitrifying Pseudomonas spp. detected with probe Pae997 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

## PAOs/GAOs

Figure 9.12. Typical cocci in tetrad aggregates of actinobacterial PAOs detected with probe Actino-221 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.13. Atypical cell aggregates of actinobacterial PAOs detected with probe Actino-221 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.14. Typical cell aggregates of actinobacterial PAOs detected with probe APAO-658 (Cy3; red) in activated sludge and EUBmix (FLUOS; green). Bar ¼ 10 mm.

Figure 9.15. Cell aggregates of Competibacter GAOs and Accumulibacter PAOs from a lab-scale reactor detected with probe GB þ GB\_G1 þ GB\_G2 (Cy3; red) in activated sludge and PAOmix (Cy5; blue) and EUBmix (FLUOS; green). Yellow overlay of GB þ GB\_G1 þ GB\_G2 and EUBmix. Light blue overlay of PAOmix and EUBmix. Bar ¼ 10 mm.

Figure 9.16. Cell aggregates of Defluviicoccus GAOs and Accumulibacter PAOs from a lab-scale reactor detected with probe DF2mix (Cy3; red) in activated sludge and PAOmix (Cy5; blue) and EUBmix (FLUOS; green). Yellow overlay of DF2mix and EUBmix. Light blue overlay of PAOmix and EUBmix. Bar ¼ 10 mm.

## Filamentous bacteria

Figure 9.17. Typical H. hydrossis morphotype (Bacteroidetes) detected with probe HHY (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.18. Typical M. parvicella morphotype (Actinobacteria) detected with probe MPAmix (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.19. Typical M. calida morphotype (Actinobacteria) detected with probe Mpa-T1-1260 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.20. Typical N. limicola II morphotype (Actinobacteria) detected with probe NLIMII175 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.21. Typical N. limicola II morphotype (Chloroflexi) detected with probe AHW183 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.22. Typical N. limicola II morphotype Candidatus Monilibacter batavus (Alphaproteobacteria) detected with probe DF1004 (Cy3; red), DF198 (Cy5; blue) and EUBmix (FLUOS; green) in a labscale reactor. White overlay of DF1004, DF198 and EUBmix. Bar ¼ 10 mm.

Figure 9.23. Same Candidatus Monilibacter batavus aside in phase contrast.

Figure 9.24. Typical N. limicola II morphotype Candidatus Alysiomicrobium bavaricum (Alphaproteobacteria) detected with PPx3-1428 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.25. Typical N. limicola II morphotype Candidatus Alysiomicrobium bavaricum (Alphaproteobacteria) detected with PPx3-1428 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.26. Same Candidatus Alysiomicrobium bavaricum aside in phase contrast.

Figure 9.27. Typical N. limicola II morphotype Candidatus Sphaeronema italicum (Alphaproteobacteria) detected with Sita-649 þ CompSita-649 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.28. Same Candidatus Sphaeronema italicum aside in phase contrast.

Figure 9.29. Typical N. limicola II morphotype Candidatus Alysiosphaera europaea (Alphaproteobacteria) detected with Noli-644 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.30. Same Candidatus Alysiosphaera europaea aside in phase contrast.

Figure 9.31. Typical N. limicola III morphotype Isosphaera spp. (Planctomycetales ) detected with NLIMIII301 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.32. Typical Mycolata morphotype (Actinobacteria) detected with Myc657 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.33. Typical Mycolata GALO morphotype Gordonia spp. (Actinobacteria) detected with Gor596 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.34. Typical Mycolata GALO morphotype Gordonia amarae (Actinobacteria) detected with G.am205 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.35. Typical Mycolata PTLO morphotype Skermania piniformis (Actinobacteria) detected with Spin1449 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.36. Typical Mycolata GALO morphotype Rhodococcus erythropolis (Actinobacteria) detected with R.ery619 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.37. Typical Mycolata GALO morphotype Rhodococcus cluster B (Actinobacteria) detected with RHOb183 (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.38. Typical Mycolata NOC morphotype Dietzia spp. (Actinobacteria) detected with DIE993 (Cy3; red) and EUBmix (FLUOS; green); Yellow overlay of DIE993 and EUBmix in activated sludge. Bar ¼ 10 mm.

Figure 9.39. Typical S. natans morphotype (Betaproteobacteria) detected with SNA (Cy3; red) and EUBmix (FLUOS; green) in activated sludge.

Figure 9.40. Typical Thiothrix morphotype (Gammaproteobacteria) and Meganema perideroedes (Alphaproteobacteria) detected with Meg983 þ 1028 (Cy3; red) and TNI (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.41. Typical Thiothrix morphotype (Gammaproteobacteria) detected with TNI (Cy3; red) and G123T (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.42. Typical 021N morphotype (Gammaproteobacteria) detected with G1B (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.43. Typical 021N morphotype (Gammaproteobacteria) detected with 21N (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.44. Phase contrast of the same typical 021N morphotype aside.

Figure 9.45. Typical 0041/0675 morphotype (TM7) detected with TM7305 (Cy3; red) and EUBmix (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.46. Phase contrast of the same typical 0041/0675 morphotype aside.

Figure 9.47. Typical 0041/0675 morphotype (Chloroflexi) detected with CFXmix (Cy3; red) and EUBmix (FLUOS; green), yellow overlay of CFXmix and EUBmix in activated sludge. Bar ¼ 10 mm.

Figure 9.48. Typical 0041/0675 morphotype (Chloroflexi) detected with CFXmix (Cy3; red) and EUBmix (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.49. Thinner 0092 morphotype (Chloroflexi) detected with CFX223 (Cy3; red) and EUBmix (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.50. Typical 0092 morphotype (Chloroflexi) detected with CFX197 þ competitor (Cy3; red) and EUBmix (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.51. Typical 0092 morphotypes (Chloroflexi) stained with Neisser. The arrows indicate the two different morphotypes that can be detected with probes CFX223 and CFX197. Bar ¼ 10 mm.

Figure 9.52. Typical 1701 morphotype Curvibacter spp. (Betaproteobacteria) detected with Curvi997 and competitors (Cy3; red) and EUBmix (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.53. Typical 1701 morphotype Curvibacter spp. (Betaproteobacteria) detected with Curvi997 and competitors (Cy3; red) and EUBmix (FLUOS; green) in activated sludge. Bar ¼ 10 mm.

Figure 9.54. Typical 1851 morphotype Kouleothrix aurantiaca (Chloroflexi) detected with Chl1851 (Cy3; red) and EUBmix (FLUOS; green) in a reactor. Bar ¼ 10 mm.

Figure 9.55. Typical Chloroflexi from an industrial WWTP without epiphytic growth (Chloroflexi) detected with CFXmix (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.56. Typical Chloroflexi from an industrial WWTP without epiphytic growth (Chloroflexi) detected with CFXmix (Cy3; red) in activated sludge. Bar ¼ 10 mm.

Figure 9.57. Typical 1863 morphotype Acinetobacter spp. (Gammaproteobacteria) detected with ACA23a (Cy3; red) in activated sludge. Bar ¼ 10 mm.

## 10 Reference list


*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.











## Index

Note: Page numbers with f and t refer to figures and tables, respectively.


on 04 February 2019

acid wash procedure 77 ammonia-oxidizing archaea (AOA) 9, 10 PCR-based methods for detection 10–11 ammonium oxidizers (AOB) 9, 10–11 16S rRNA-based phylogenetic tree 11f probes for detection 11–13, 12t anaerobic ammonium oxidizers see anammox anammox bacteria 9, 10, 14–17 16S rRNA-based phylogenetic tree 16f probes for detection 16t, 17

Beggiatoa morphotype 41 abundance, morphology and affiliation 41 probes for detection 41, 42–43t biovolume-based approach 86, 88–92

cell counting by image analysis 86, 87 CMEIAS 89, 90 confocal laser scanning microscope (CLSM) 87 counterstaining, with DAPI 81 culture-dependent approach 20 culture-independent approach 20

daime 90 denitrifiers 19–20, 94–96f 16S rRNA-based phylogenetic tree 22f abundance in full-scale plants 20–21 identity of 20 probes for detection 21–24, 23t

*#* 2009 IWA Publishing. FISH Handbook for Biological Wastewater Treatment: Identification and quantification of microorganisms in activated sludge and biofilms by FISH. Edited by Per Halkjær Nielsen, Holger Daims and Hilde Lemmer. ISBN: 9781843392316. Published by IWA Publishing, London, UK.

40 ,60 -diamidino-2-phenylindol (DAPI) 81 direct viable count (DVC) procedures 71 DNA microarrays 2 Enhanced Biological Phosphorus Removal (EBPR) 20, 25 epiphytic bacteria, involved in protein hydrolysis 69–70 Escherichia coli, as indicator for entero-pathogens 70–71 fermenting bacteria 70 fields of view (FOV) 86, 87 filamentous bacteria 3, 33–36, 35t, 98–109f 16S rRNA-based phylogenetic tree 36f, 37f, 38f, 61f detection and identification 36–67 of 0041/0675 morphotype 59, 60t, 105–106f of 0092 morphotype 61–63, 62t, 106–107f of 021N filament morphotype 57–59, 58t, 104–105f of 0803 morphotype 67 of 0914 morphotype 67 of 1701 morphotype 63, 64t, 107f of 1851 morphotype 65, 66t, 108f of 1863 morphotype 46t, 65, 67, 109f of Beggiatoa morphotype 41, 42–43t of Haliscomenobacter hydrossis morphotype 41, 44, 44t, 44f, 98f of Leptothrix discophora morphotype 42–43t, 56 of Leucothrix mucor morphotype 45, 46t of Microthrix parvicella morphotype 45–48, 47t, 98f of Nocardioform actinomycetes/Mycolata morphotypes 52–56, 54t, 102–103f of Nostocoida limicola morphotypes 48–52, 50t, 98–102f of Sphaerotilus natans morphotype 46t, 56, 104f of Streptococcus morphotype 46t, 56–57 of Thiothrix morphotype 57–59, 58t, 102–103f, 104f probes for detection see individual entries G bacteria 25 gelatine coating of microscopic slides procedure 77 gelatine solution 75 glycogen-accumulating organisms (GAOs) 25, 26, 97f 16S rRNA-based phylogenetic tree 31f identity of 28–30 probes for detection 29–31, 30t Haliscomenobacter hydrossis morphotype 41, 44, 44f, 45, 98f 16S rRNA-based phylogenetic tree 44f abundance, morphology and affiliation 41, 45 probes for detection 44t, 45 hybridization procedures 79–81 Leptothrix discophora morphotype 56 abundance, morphology and affiliation 56

Leucothrix mucor morphotype 45 16S rRNA-based phylogenetic tree 46f abundance, morphology and affiliation 45 probes for detection 45, 46t lipase 75 permeabilization 79 lysozyme 75 permeabilization 78 microbiology 2–4 microorganisms growth, important factors 4–6 microscopy 81, 82–83 Microthrix parvicella morphotype 45–48, 98f 16S rRNA-based phylogenetic tree 48f abundance, morphology and affiliation 45–47 probes for detection 47–48, 47t mild acid hydrolysis permeabilization 79 morphotype descriptions 39t Most-Probable Number (MPN) 1 mutanolysin 75 permeabilization 78 Mycolata morphotypes 52–56, 102–103f 16S rRNA-based phylogenetic tree 55f abundance, morphology and affiliation 52–53 probes for detection 53–56, 54t NaCl 76 nitrifiers 9–10, 93–94f ammonium oxidizers (AOB) see ammonium oxidizers (AOB) anammox bacteria see anammox bacteria nitrite oxidizers (NOB) see nitrite-oxidizing bacteria (NOB) nitrite oxidizers see nitrite-oxidizing bacteria nitrite-oxidizing bacteria (NOB) 9, 13–14 16S rRNA-based phylogenetic tree 14f probes for detection 14, 15t NOC-like organisms 52 Nocardioform actinomycetes morphotypes 52–56, 102–103f 16S rRNA-based phylogenetic tree 55f abundance, morphology and affiliation 52–53 probes for detection 53–56, 54t Nostocoida limicola morphotypes 48–52, 98–102f 16S rRNA-based phylogenetic tree 51f abundance, morphology and affiliation 48–49 probes for detection 49–52, 50t PCR-based methods 1 for AOA detection 10–11 permeabilization by enzymatic or chemical treatment 78–79 achromopeptidase permeabilization 79 lipase permeabilization 79

lysozyme permeabilization 78

on 04 February 2019

probes for detection 56

#### Index 123

mild acid hydrolysis permeabilization 79 mutanolysin permeabilization 78 proteinase K permeabilization 78–79 phylochips-based assays 2 pine tree-like organisms (PTLO) 52 plate count 1 poly-L-lysine coating of microscopic slides procedure 77 poly-L-lysine solution 75 polyphosphate-accumulating organisms (PAOs) 25, 96f 16S rRNA-based phylogenetic tree 29f identity of 26–27 probes for detection 28, 27t probes preparation and quality check 79 test for 81–82 see also under specific microorganisms proteinase K 75 permeabilization 78–79 protocol 73–74, 74f counterstaining with DAPI 81 dehydration 78 equipment and supplies 76 general planning 84 hybridization 79–81 immobilization of the samples on glass slides 77–78 materials and solutions 75–76 microscopy 81 permeabilization 78–79 preparation and quality check 79 for quantitative FISH see under quantitative FISH recommendations and troubleshooting 81–84

sample collection and fixation 76–77 sample preparation 77 quantitative FISH 85 optimization 91–92 overview 86–87 protocol 88–91 image acquisition 89 image analysis 89–91 in situ hybridization 88 sample immobilization 88 sample preparation 88 quantitative PCR 1, 86 species composition 3, 4 Sphaerotilus natans morphotype 56, 104f 16S rRNA-based phylogenetic tree 46f abundance, morphology and affiliation 56 probes for detection 46t, 56 Spike-FISH 87 Streptococcus morphotype 56–57 16S rRNA-based phylogenetic tree 46f abundance, morphology and affiliation 56–57 probes for detection 46t, 57 sulfate-reducing bacteria 70 tetrad-forming microorganisms (TFOs) 25 Thiothrix morphotype 57–59, 102–103f, 104f 16S rRNA-based phylogenetic tree 58f abundance, morphology and affiliation 57

probes for detection 57–59, 58t

The *FISH Handbook for Biological Wastewater Treatment* provides all the required information for the user to be able to identify and quantify important microorganisms in activated sludge and biofilms by using fluorescence in situ hybridization (FISH) and epifluorescence microscopy.

It has for some years been clear that most microorganisms in biological wastewater systems cannot be reliably identified and quantified by conventional microscopy or by traditional culture-dependent methods such as plate counts. Therefore, molecular biological methods are vital and must be introduced instead of, or in addition to, conventional methods. At present, FISH is the most widely used and best tested of these methods. This handbook presents all relevant information from the literature and, based on the extensive experience of the authors, advice and recommendations are given for reliable FISH identification and quantification.

The overall purpose of the book is to help scientists, consultants, students, and plant operators to get an overview of important microorganisms in biological wastewater treatment and to explain how FISH can be used for detecting and quantifying these microbes. A proper and reliable identification of dominant microorganisms is of great importance for research and new developments in the wastewater treatment industry, and it is important for optimization and troubleshooting of operational problems in present wastewater treatment plants.

The book encompasses an overview of dominant microorganisms present in the wastewater treatment systems, which oligonucleotide probes (gene probes) to select for detection of these microbes by FISH, how to perform FISH (detailed protocols), how to quantify the microbes, and how to solve common problems of FISH. The book addresses several functional groups: nitrifiers, denitrifiers, polyphosphate-accumulating organisms, glycogen-accumulating organisms, bacteria involved in hydrolysis and fermentation, filamentous bacteria from bulking sludge, and scum-forming bacteria. A comprehensive collection of FISH-images showing dominant representatives of these groups helps readers to use FISH in the context of wastewater treatment.

Editor(s): Per Halkjær Nielsen,

**FISH Handbook** 

in activated sludge and biofilms by FISH

**Wastewater Treatment**

Identification and quantification of microorganisms

Editor(s): Per Halkjær Nielsen, Holger Daims and Hilde Lemmer

**for Biological** 

Holger Daims and Hilde Lemmer

Identification and quantification of microorganisms in activated sludge and biofilms by FISH

**FISH Handbook for Biological Wastewater Treatment**

## www.iwapublishing.com

ISBN 13: 9781843392316 ISBN: 1843392316