Bio-P Microbiology

FEMS Microbiology Reviews 27 (2003) 99^127
www.fems-microbiology.org
The microbiology of biological phosphorus removal in

activated sludge systems
Robert J. Seviour , Takashi Mino, Motoharu Onuki
Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

Institute of Environmental Studies, Graduate School of Frontier Sciences, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113, Japan
Received 11 December 2002; received in revised form 18 February 2003; accepted 20 February 2003
First published online 20 March 2003
Abstract
Activated sludge systems are designed and operated globally to remove phosphorus microbiologically, a process called enhanced
biological phosphorus removal (EBPR). Yet little is still known about the ecology of EBPR processes, the microbes involved, their
functions there and the possible reasons why they often perform unreliably. The application of rRNA-based methods to analyze EBPR
community structure has changed dramatically our understanding of the microbial populations responsible for EBPR, but many
substantial gaps in our knowledge of the population dynamics of EBPR and its underlying mechanisms remain. This review critically
examines what we once thought we knew about the microbial ecology of EBPR, what we think we now know, and what still needs to be
elucidated before these processes can be operated and controlled more reliably than is currently possible. It looks at the history of EBPR,
the currently available biochemical models, the structure of the microbial communities found in EBPR systems, possible identities of the
bacteria responsible, and the evidence why these systems might operate suboptimally. The review stresses the need to extend what have
been predominantly laboratory-based studies to full-scale operating plants. It aims to encourage microbiologists and process engineers to
collaborate more closely and to bring an interdisciplinary approach to bear on this complex ecosystem.
3 2003 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
Keywords : Activated sludge; Enhanced biological phosphorus removal; Phosphorus accumulating organism ; ‘G-bacteria’;
Glycogen accumulating organism ; Rhodocyclus
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
2. Some history of EBPR processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3. Which chemical transformations characterize EBPR biomass? . . . . . . . . . . . . . . . . . . . . . 101
4. The biochemistry of EBPR systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.1. EBPR metabolism with acetate as sole carbon source . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2. EBPR metabolism with substrates other than acetate . . . . . . . . . . . . . . . . . . . . . . . . . 104
5. The nature and role of polyP storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6. The microbiology of EBPR ^ the culture-dependent approach . . . . . . . . . . . . . . . . . . . . . 106
7. The microbiology of EBPR ^ the culture-independent approach: chemical markers . . . . . . 107
8. The microbiology of EBPR ^ the culture-independent approach: molecular microbial ecology
of EBPR communities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
9. What have molecular techniques revealed about the nature of EBPR microbial communi-
ties? .............................................................. 108
* Corresponding author. Present address: Biotechnology Research Centre, La Trobe University, Bendigo, Vic. 3552, Australia.
E-mail address : r.seviour@latrobe.edu.au (R.J. Seviour).
Abbreviations : CFB, Cytophaga^Flavobacterium^Bacteroides division; DAPI, diamidino phenyl indole; DGGE, denaturing gradient gel electrophoresis;
DNPAO, denitrifying polyP accumulating organisms ; EBPR, enhanced biological phosphorus removal; FISH, £uorescence in situ hybridization; GAO,
glycogen accumulating organisms ; MAR, microautoradiography ; PAO, phosphorus accumulating bacteria ; PCR, polymerase chain reaction; PHA, poly-L-
hydroxyalkanoate; PHB, poly-L-hydroxybutyrate; polyP, polyphosphate; SBR, sequencing batch reactor ; SSCP, single strand conformation
polymorphism; UCT, University of Cape Town; VFA, volatile fatty acids
0168-6445 / 03 / $22.00 3 2003 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
doi:10.1016/S0168-6445(03)00021-4
FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

100 R.J. Seviour et al. / FEMS Microbiology Reviews 27 (2003) 99^127
9.1. Diversity of EBPR community populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

9.2. Possible candidates for PAO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
9.3. The case for Rhodocyclus-related organisms as PAO . . . . . . . . . . . . . . . . . . . . . . . . . 110
10. Structure:function relationships within EBPR communities . . . . . . . . . . . . . . . . . . . . . . . . 111
10.1. Using DAPI staining to indicate PAO identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.2. Using FISH/staining and FISH/MAR to indicate PAO identity . . . . . . . . . . . . . . . . 112
10.3. Using di¡erential centrifugation to indicate PAO identity . . . . . . . . . . . . . . . . . . . . . 113
11. Studies with full-scale EBPR plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
12. Denitrifying EBPR systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
13. Why do EBPR systems often behave unpredictably? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
13.1. In£uence of in£uent substrate composition on EBPR performance? . . . . . . . . . . . . . 114
13.2. What are these ‘G-Bacteria’ and GAO? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
13.3. Evidence for a role for GAO/‘G-Bacteria’ in EBPR systems from mixed culture
studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

13.4. The identity of the ‘G-Bacteria’ and GAO: culture-dependent approaches . . . . . . . . 116
13.5. The identity of the ‘G-Bacteria’ and GAO: culture-independent approaches . . . . . . . 116
13.6. Structure:function studies with non-EBPR communities and the role of ‘G-Bacteria’
and GAO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
14. How is EBPR biochemistry regulated? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
15. Mathematical models for EBPR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
16. Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
1. Introduction ceous material produce treated e¥uents with high residual

P levels, although some N removal does occur in these
Eutrophication is a global problem, in which blooms of systems [3,6,11]. In the past 30 years many plants have
cyanobacteria and eukaryotic algae occur as a conse- been designed and built around the world to deliberately
quence of the breakdown of community homeostasis reduce not just organic carbon and N levels but also P,
from continued pollution of oligotrophic aquatic bodies using microbially based methods [11^14]. This removal is
with nutrients like nitrogen (N) and phosphorus (P) at achieved by encouraging the accumulation of P in bacte-
levels which exceed growth-limiting concentrations for rial cells in the form of polyphosphate (polyP) granules in
these photosynthetic organisms [1,2]. Of these nutrients, excess of the levels normally required to satisfy the meta-
P input is considered more critical [3,4] since many of bolic demand for growth, a storage process commonly
the cyanobacteria are diazotrophic, capable of satisfying referred to as ‘enhanced biological phosphorus removal’
their N requirements from ¢xation of atmospheric N2 . The (EBPR) [15,16]. This P uptake and polyP storage in or-
problems caused by eutrophication have been discussed ganisms known as polyP accumulating organisms (PAO)
frequently in the literature, e.g. [5,6], and range from a occur under particular plant operational conditions that
decrease in the aesthetics of the a¡ected receiving body are reasonably well understood, even though how these
of water to probable oxygen depletion and stagnation fol- conditions encourage the accumulation of polyP in PAO
lowing the eventual death of these phototrophs [7]. An is not [17,18]. When the cells are removed from the process
increase in water turbidity from increased levels of phyto- during sludge wasting then so is the P.
plankton also reduces light penetration and photosynthetic Since the initial reports of the EBPR phenomenon,
oxygen generating activity [8]. It is now clear also that the many di¡erent full-scale plant con¢gurations have been
consumption of eutrophied water poses a serious medical designed and built, but these have evolved in an empirical
threat since some of the cyanobacteria are toxigenic, and manner, with little or no real understanding of the micro-
the symptoms caused by exposure to these toxins while biology involved. In the absence of this vital information,
often trivial, may be very serious, leading to death in it is quite feasible that these plants may not be designed
some cases [9,10]. for, or operate under optimal conditions for EBPR. Even
Increasing evidence points to wastewater treatment though they are used increasingly frequently to treat do-
plants as major point sources for both N and P, which mestic waste around the world, especially in land locked
in the latter case originates from fecal material, industrial communities discharging their wastes into environmentally
and commercial sources and synthetic detergents and oth- sensitive receiving bodies of water, most still operate un-
er cleaning products [5]. Conventional activated sludge reliably. Such unreliable performance persuaded the US
plants designed principally to remove organic carbona- Water Environment Research Foundation to fund projects

R.J. Seviour et al. / FEMS Microbiology Reviews 27 (2003) 99^127 101
in 2001 to investigate technology yielding consistently low cally respire and deplete the supply of organic sub-
P levels in treated e¥uents. Although P concentrations in strates, making them no longer available for the PAO,
the in£uent will vary with its source and the cultural be- 3. strictly maintaining anaerobic conditions in the anaer-
havior of individual societies, and especially whether low P obic zone,
detergents are used, most plant e¥uents need to satisfy 4. recycling the biomass through alternating anaerobic:
discharge licence agreements and contain 6 1 mg l31 P aerobic conditions, although some of the ¢rst EBPR
[19]. Thus, chemical P removal by precipitation, with seri- systems did not incorporate this design feature.
ous environmental problems of its own, is often necessar- The reasons for these requirements and how critical they
ily incorporated to produce a treated e¥uent that meets are to EBPR will be discussed later in this review, but they
the required standards [19], and these will probably be- were developed in an empirical manner with no under-
come more stringent as governments become more envi- standing of their microbiological bases, and this is still
ronmentally sensitive and respond increasingly to lobbying largely the case. Thus, like many other aspects of EBPR,
from articulate environmental groups. our knowledge of the engineering requirements has ad-

It is likely that EBPR technology will never ful¢ll its vanced much more rapidly than our microbiological
true potential until we have a more comprehensive under- understanding. Since the ¢ve-stage Bardenpho process
standing of the microbial ecology of P removal. Several was introduced, many other con¢gurations have been de-
reviews have already been written [14,17,20^22], but in the signed and built, in attempts to simplify plant con¢gura-
past few years the application of molecular RNA-based tion and decrease construction costs, and to minimize or
culture-independent methods [23^25] has had a profound eliminate the problems of nitrate recycling into the anaer-
impact on how we now view EBPR microbiology. Many obic zone [6,11,14]. These stimuli led to the design of the
of the questions raised in the earlier reviews concerning three-stage Bardenpho or Phoredox system, which can be
the identity of the putative PAO have been addressed considered the progenitor of most of the EBPR plants
[26], but there is still much to be learned even here. There- now used, like the Johannesburg and modi¢ed University
fore, it seems timely to look at the topic again to summa- of Cape Town (UCT) processes, which were both con-
rize our current state of understanding of EBPR micro- ceived to deal better with anaerobic nitrate recycling prob-
biology and to speculate on what still needs to be lems, and are described in detail elsewhere [6,11,14].
understood if we are to design and run better EBPR sys-
tems in the future.
3. Which chemical transformations characterize EBPR
biomass?
2. Some history of EBPR processes
Many studies have described the gross chemical changes
Several reviews have discussed how EBPR technology that occur during the anaerobic/aerobic cycling of the bio-
developed from its accidental discovery [27] more than 40 mass in EBPR systems, and these have been critically re-
years ago. In an often provocative review, the early events viewed [17,21]. Most have been carried out on biomass
showing the biological basis for EBPR and the importance from laboratory scale reactors fed synthetic sewage and
of the South African work by Barnard are dealt with single carbon sources, where some control can be exercised
chronologically [21]. Barnard, with others, was responsible over parameters like feed rates, hydraulic retention times
for designing an evolving series of plant con¢gurations for and sludge age. So they operate under quite di¡erent con-
EBPR, many of which are still used [6,11,14]. Most of ditions to those usually encountered in full-scale systems,
these evolved from his Bardenpho plant that was designed especially in terms of the selective pressures brought to
for more e⁄cient N removal than existing con¢gurations bear on the microbial populations [29,30], and have been
like the modi¢ed LudzakÊttinger con¢guration [28], but criticized accordingly. However, the general agreement is
adding an anaerobic reactor to this system gave reliable P that in the anaerobic zone, some bacteria very rapidly
removal in the ¢ve-stage Bardenpho, or Phoredox, system. assimilate short chain volatile fatty acids (VFA) formed
His subsequent work, and that of others (reviewed in by the action of chemoorganotrophic fermentative bacte-
[6,11,14]) suggested that the following operational condi- ria [31]. These are not used to support an increase in cell
tions were needed for successful EBPR: growth rates, but instead are stored as poly-L-hydroxyal-
kanoates (PHA), whose chemical composition depends on
1. an anaerobic reactor which receives the initial in£uent the carbon source assimilated [32,33]. Hence with acetate
containing the readily degradable carbon and energy assimilation, poly-L-hydroxybutyrate (PHB) is synthe-
sources, sized, while proprionate leads to the production of mainly
2. a need to limit the concentration of nitrate entering this poly-L-hydroxyvalerate. Often mixtures of PHA are found
anaerobic zone by incorporating an anoxic zone in the in biomass from a single plant, although the identity of the
design, since in its presence EBPR fails because the PHA synthesizing bacteria in activated sludge is still
denitrifying bacteria were considered able to anaerobi- poorly understood. During this anaerobic stage, biomass

good EBPR performance is often associated with appear-

ance of large tight clusters in the biomass.
4. The biochemistry of EBPR systems
4.1. EBPR metabolism with acetate as sole carbon source
Several empirical models have been described [16,17,38]

to attempt to explain the chemical changes, mainly from
data from studies analyzing community chemistry in re-
sponse to acetate as the sole carbon source, and these
models have been critically examined in earlier reviews

[17,21]. Their main features are given in Fig. 1. Although
there is general agreement in the interpretation of the ma-
jor changes, the models di¡er in certain important aspects
of EBPR. It should be emphasized that the chemical data
upon which all the models are based are gross re£ections
of the overall activities of communities of unknown micro-
biological nature, and so are impossible to interpret in
terms of the activities of individual populations. The mod-
els also assume that the detected uptake and release of
polyP, and PHA synthesis and reutilization all occur in
a single population of PAO cells, all sharing the same
metabolic attributes [39]. This may or may not be the
case, but seems unlikely from what we now know about
Fig. 1. A summary of the major features of the biochemical models for EBPR communities (see Section 9).
EBPR. The changes thought to take place during the aerobic and anaer- The models all explain successful EBPR as a conse-
obic stages are shown. The insert shows the characteristic microscopic
quence of the PAO achieving dominance under anaero-
appearance of the biomass taken from the aerobic zone, with clustered
cells staining black with the Neisser stain for polyP. bic/aerobic recycling conditions by having a selective ad-
vantage over the other bacterial populations present in
their abilities to synthesize intracellular storage com-
PHA levels increase in parallel with the assimilation of pounds under the ‘feast-famine’ conditions which charac-
acetate [34]. Intracellular polyP content falls and an in- terize EBPR systems. Thus, under anaerobic conditions,
crease in phosphate levels in the bulk liquid can be de- the PAO are thought to rapidly assimilate organic sub-
tected following its release from the biomass [35]. strates like acetate and use these to synthesize PHA using
Then in the subsequent aerobic stage, PHA levels in the stored polyP as an energy source, and the orthophosphate
biomass fall concomitant with a decrease in phosphate generated from polyP degradation is released into the bulk
levels in the bulk liquid and an increase in biomass P levels liquid (Fig. 1). Then in the absence of any organic com-
as polyP, which can be s 15% of cell dry weight [18]. If a pounds in the aerobic zone, organisms with stored PHA
biomass sample taken from the aerobic reactor is exam- are able to use these as carbon and energy sources to grow
ined microscopically, cells staining for polyP are abun- and to assimilate phosphate to synthesize polyP. Thus,
dant, but no intracellular PHA can be detected by stain- PAO achieve dominance under the prevailing anaerobic:
ing. On the other hand, cells in biomass from the aerobic conditions because they alone can grow aerobi-
anaerobic reactors stain for PHA but not polyP. These cally in the absence of any exogenous source of carbon
are common features of EBPR plants, although some and energy, by using the PHA accumulated anaerobically.
show patterns of chemical changes quite di¡erent to these, The general view is that the PAO are probably better able
as discussed below. Furthermore, the usual experience is to survive in ecosystems exposed to these ‘feast-famine’
that cells appearing as large coccobacilli in pairs showing regimes, with a capacity for rapid substrate assimilation
these staining responses are commonly organized into and storing energy and carbon taking precedence over
large £oc-associated clusters (Fig. 2), although they may high growth rates [40,41]. Synthesis of PHA requires a
sometimes be loosely dispersed throughout the biomass source of reducing power, and the various models pro-
[18,36,37] for reasons not understood. It is assumed that posed di¡er in what is used for this. Suggestions were
these chemical changes are implemented only by the PAO that the electrons required were derived from the anaero-
arranged in this distinctive way. The consequence of PAO bic operation of the TCA cycle [38,42], although Mino et
cell cluster organization on EBPR is not clear, although al. produced evidence [17,22] that glycogen, an intracellu-


Fig. 2. A tree showing the phylogenetic relationships among the putative PAO and ‘G-Bacteria’/GAO. This is based on that given in [26]. Only sequen-
ces of more than 1400 bp are included. Red: putative PAO, Blue: ‘G-Bacteria’/GAO.
lar storage compound synthesized aerobically by the PAO, evidence suggests that glycogen availability, and not intra-
was catabolized anaerobically to generate electrons for cellular polyP levels, may ultimately limit the ability of
PHA synthesis (Fig. 1). Evidence in favor of a key role cells to assimilate substrates anaerobically under shock
for glycogen in EBPR has increased since then and the loading conditions [44,45].
Mino model is now preferred. However, it has been pro- A similar debate has surrounded the bioenergetics of
posed that, as glycogen degradation alone would not sat- anaerobic substrate assimilation and PHA synthesis in
isfy the demands for reducing power for PHA synthesis, PAO. Early studies claimed that adenosine triphosphate
then the TCA cycle provides the remainder [43,44]. Gly- (ATP) from polyP degradation was the only energy source
cogen is thought to play a key role in EBPR by regulating used anaerobically by the PAO for substrate (i.e. acetate)
the redox balance of the PAO [17], which may be neces- assimilation, PHA synthesis and other biosynthetic re-
sary for permitting the anaerobic assimilation and metab- actions, and cell maintenance. However, glycogen cata-
olism of the diverse range of readily biodegradable sub- bolism is now also thought to provide ATP for PHA
strates encountered in full-scale systems (see below). Some production, but how much energy is made available de-

pends on the pathway used for glycogen catabolism data, Hesselman et al. [44] proposed a modi¢ed biochem-
[17,44,45]. ical model for EBPR to accommodate their ideas.
Chemical analysis of these enriched cultures can reveal However, little experimental biochemical data are yet
the bulk chemical changes taking place but contributions available to validate any of these empirical models, and
from less dominant populations may be masked, and often even NMR data are unable to explain fully the behavior
interpretation of the data is di⁄cult. In vivo nuclear mag- of the di¡erent communities, since the structure/function
netic resonance (NMR) studies have also been used to relationships of the populations involved are completely
follow metabolic £ux in mixed communities in response unknown. Thus, some of the di¡erences in chemical per-
to the supply of particular substrates under speci¢ed con- formance of biomasses from the di¡erent studies may be
ditions. This approach has allowed clari¢cation of some of explained by the presence of communities of quite di¡er-
the confusions and contradictions arising from the data ent composition developing in reactors run under very
from the chemical studies, as well as pointing out how similar conditions, a situation now documented in the lit-
poorly understood the biochemistry of EBPR communities erature [26,46]. It is strongly recommended that all future

still is [43^45]. With NMR, Pereira et al. [43] clearly studies of this nature should involve parallel molecular
showed anaerobic operation of the TCA cycle in their microbial community studies if their purpose is to clarify
biomass, and glycogen was con¢rmed as a source of re- and not to confuse.
ducing power for PHA synthesis, where glycogen appeared
to be degraded via the Entner Doudoro¡ (ED) pathway. 4.2. EBPR metabolism with substrates other than acetate
Importantly, both acetate and glycogen appeared to con-
tribute to anaerobic PHA synthesis, and the PHA was The other di⁄culty with studies of this kind is that full-
then used aerobically for glycogen production [45]. scale activated sludge systems deal with a much more di-
Some data support these earlier interpretations. Thus, verse range of substrates than acetate or other short chain
Hesselman et al. [44] provided evidence that the TCA cycle VFA, and yet relatively little is known about their in£u-
could act as a partial source of reducing power for PHA ence on the chemical changes described above, although
production, and that glycogen catabolism occurred via the many organic compounds both singly and in combination
ED pathway. However, some of their other data raise appear to support EBPR [47,48], possibly by being con-
questions about many of the previous assumptions made verted to VFA ¢rst [15]. It has been suggested [49] that the
about EBPR biochemistry, especially its bioenergetics. In chemical nature of the VFA supplied, by directly deter-
particular, they suggested, on theoretical ground as well, mining the amount and chemical composition of the PHA
that cells assimilated acetate not by active transport, as synthesized anaerobically, in turn determined aerobic P
assumed earlier, but by passive or facilitated di¡usion, assimilation. However, once again no community analyses
requiring no expenditure of energy. Furthermore, acetate were carried out under the di¡erent feed conditions used
activation, the most energetically demanding step in an- and no convincing explanation for the results was pro-
aerobic EBPR metabolism apparently occurs with no in- vided.
volvement of the enzyme acetate kinase, which had been It has been well documented in the literature that glu-
assumed to be important in all earlier models [17]. Instead cose as sole carbon source can lead to a failure of EBPR
acetyl CoA synthetase was used. [50,51] and the possible microbiological reasons for why
A role was proposed for involvement of energy sources this might happen will be discussed later. However, bio-
in anaerobic EBPR additional to those suggested in the chemical models for EBPR with glucose have been pro-
earlier models. These included pyrophosphate, released by posed and supported by laboratory scale studies [52^56],
the action of the polyP degrading enzyme detectable at which have convincingly shown that glucose may support
high levels of activity in their biomass, pyrophosphate: stable EBPR in laboratory scale reactors. For example,
phosphoenolpyruvate:AMP phosphotransferase, partici- from 13 C-NMR data it has been suggested [55] that the
pating in substrate level phosphorylation and the estab- lactic acid producing bacteria play a crucial role in the
lishment of a proton motive force. In addition, e¥ux of anaerobic zone. This population has been held responsible
MgHPO4 was coupled to proton translocation via a sec- for anaerobic glucose assimilation leading to lactic acid
ondary transporter system. As with many other similar production and eventually glycogen storage [55^57]. In
studies using laboratory scale reactors, the ratio of anaer- the model described in [55] it was hypothesized that gly-
obic P released to acetate assimilated was not a constant. cogen was transformed anaerobically into PHA by the
The literature suggests that this ratio depends on the PAO, while the lactic acid bacteria used glycogen to syn-
mixed liquor pH a¡ecting the energy demand for acetate thesize a lactate storage polymer. This model allowed P
transport [17], but Hesselman et al. [44] considered it as uptake aerobically by the PAO, as predicted by the models
re£ecting directly both the polyP and glycogen content of discussed above (Fig. 1). Unfortunately no microbiological
their biomass. It may also vary of course with di¡erent evidence was provided to suggest whether the community
EBPR communities, not all behaving precisely as these structure would support the model, apart from scant mor-
biochemical models suggest they should. From their phological descriptions of the main bacteria seen in the

biomass [55]. A di¡erent model to explain EBPR with meric substances associated with £ocs [64]. Its organiza-
glucose was proposed in [56], which was much more sim- tion inside PAO is less certain, with various claims sup-
ilar mechanistically to the Mino and Comeau^Wentzel porting its location in the cytoplasm, the periplasm or
suggestions, but also with little or no direct microbiolog- associated with intracellular membranes [65^67], and pos-
ical or biochemical evidence to support it. sibly complexed with proteins and DNA and RNA [18,61].
How EBPR biomass might be supported by amino acids These may re£ect stages in its synthesis, but all may be
as possible substrates has attracted surprisingly little inter- possible if PAO contain several di¡erent populations, as
est since their presence in wastewater is certain; but both now seems likely (see Section 10), and each stores polyP
glutamate and aspartate as sole carbon sources could sup- di¡erently.
port EBPR in a laboratory scale reactor [57,58]. It was Several attempts have been made to extract this polyP
suggested that anaerobic substrate storage again occurred, to determine the distribution of chain lengths in EBPR
but not all the glutamate used as substrate was stored as sludge and to fractionate it to understand more about P
intracellular PHA. Instead, storage was thought to involve £ux during its anaerobic:aerobic release and uptake. For

synthesis of some storage polypeptide yet to be chemically example, a high and a low molecular mass fraction was
characterized and identi¢ed. obtained by acid extraction, and it was proposed that the
It is also clear that the existing biochemical models still low molecular mass fraction was involved in anaerobic
fail to account for all the experimental data from studies polyP degradation [68], results which disagree with those
of this kind. For example, biomass has been shown to using di¡erent polyP characterization methods including
31
assimilate substrates anaerobically [59] with neither polyP P-NMR [69^71]. Other methods [72,73] have been used
hydrolysis nor conversion of glycogen to PHA, and EBPR to fractionate polyP into short chain polyP (SCP), long
has also been achieved with glucose as carbon source, but chain polyP (LCP) and granular polyP from EBPR bio-
in the absence of any anaerobic storage of PHA [60]. mass. Evidence has been presented [74] to suggest that P
These outcomes emphasize how poorly we still understand was exchanged between these pools during EBPR and that
the basic biochemistry of EBPR and how di⁄cult it is to most of this exchange in a UCT plant involved the LCP
interpret such chemical trends in any comparative manner fraction. However, this pattern of biomass P £ux was
in the absence of any knowledge about the structure and markedly in£uenced by plant con¢guration and in plants
function of the responsible microbial communities. It is where nitrate was found in the anaerobic zone, intracellu-
probable that these models will remain speculative until lar P location shifted mainly to the SCP fraction. No ex-
the biochemistry of pure cultures of authentic PAO is planation was given for these trends, but the study does
examined under conditions that closely simulate those illustrate the complexity of EBPR and reinforces the di⁄-
found in EBPR systems. culties in interpreting data from mixed communities of
unknown microbiological status.
It is clear that polyP needs the presence of counter ions
5. The nature and role of polyP storage to stabilize it and neutralize the strong negative charge it
carries. Energy dispersive X-ray spectroscopy has revealed
PolyP is a polyanionic polymer consisting of many or- the presence of Mg2þ , Ca2þ and Kþ in varying amounts
thophosphate monomers linked by phosphoanhydride depending on their availability in the bulk liquid, associ-
bonds of high energy, and the level of its polymerization ated with polyP granules in EBPR biomass. Several stud-
is known to vary between di¡erent organisms. All cells ies [75^78] have suggested that Mg2þ ions but not Ca2þ
contain polyP, which suggests that this polymer is essential are particularly important for stable EBPR, and increasing
for cell function, although it may not be visualized or the available Mg2þ substantially increased the P removal
easily detected there [22,61]. Yet surprisingly little is capacity of biomass in a bench scale pilot plant [77]. These
known about its biochemistry, especially in EBPR bio- results are interesting but not easy to test with full-scale
mass, where it can be readily detected as intracellular systems. However, evidence from some Australian EBPR
granules after staining with basic stains like methylene plants that fail routinely during the summer implicates a
blue and toluidine blue. These granules stain pink with fall in the subterranean water table and a measurable drop
methylene blue, and so are referred to as metachromatic. in the in£uent Mg2þ levels [79]. It also appears that Mg2þ
DAPI (diamidino phenyl indole), which imparts a yellow and Kþ are released and taken up simultaneously with P
£uorescence to polyP granules [62,63], has frequently been in EBPR systems, but Ca2þ is not [77], and that Ca2þ ,
used to detect polyP in EBPR studies described below. which is important in stabilizing polyP granules, may
Little recent work has been published on its organization play no apparent role in anaerobic P release. Very similar
and intracellular location in relationship to EBPR bio- results were obtained with pure cultures of Acinetobacter
mass. Earlier data convincingly showed that P in EBPR johnsonii strain 210A [80].
biomass is organized as intracellular polymeric polyP and It is not clear what function(s) polyP performs in cells
is not chemically precipitated on the surface of cells as that accumulate it, but several functions have been pro-
inorganic deposits or associated with extracellular poly- posed, supported by experimental data from pure cultures

of bacteria, especially Escherichia coli, and are summa- cation methods used, but the presence of many previously
rized in [22,61]. Known functions include acting as a undescribed Acinetobacter spp. in activated sludge systems
source of ATP following the action of the enzymes poly- is now established [93,94]. Most of the pure cultures ob-
P :AMP phosphotransferase and adenylate kinase, by re- tained from EBPR systems could synthesize polyP and
placing ATP as a phosphate donor in the phosphorylation PHA aerobically at high levels from acetate, although
of sugars and a reserve of inorganic P, and as an intra- not all isolates synthesized both, and this ability was not
cellular bu¡er. By acting as a reservoir for Mg2þ polyP associated with any particular species [18]. A. johnsonii
may also serve to supply cells with Mg2þ under Mg2þ - strain 210A shared many of these features and its physi-
limiting conditions and to play a vital role in heavy metal ology and biochemistry have been examined most closely
(e.g. cadmium) tolerance by excreting the metal via a met- [22,89,95,96]. However, even though low levels of anaero-
al phosphate transport system [81]. PolyP synthesis in bic P release were detectable in it and many other isolates,
some bacteria occurs in response to nutritional and os- none of the isolates obtained then (or since) could assim-
motic stress, and serves as a signal for production of stress ilate acetate and synthesize PHA anaerobically concomi-

proteins. It is also involved in regulating di¡erentiation tant with P release, as the biochemical models demand,
and morphogenesis in both eukaryotes and prokaryotes even under conditions set up to simulate EBPR plants
[61]. Whether these functions apply to PAO in EBPR sys- [97]. There are several explanations for such results [97].
tems is not clear, although polyP is assumed to act as an These include the possibility that the biochemical models
energy source for the anaerobic substrate assimilation and are incomplete or incorrect, that Acinetobacter behaves
PHA synthesis, and convincing evidence is available that di¡erently in pure culture than in EBPR systems, or that
polyP may bu¡er EBPR biomass under alkaline condi- Acinetobacter is not a major PAO responsible for carrying
tions, regulating the intracellular pH of PAO by its deg- out the chemical changes so characteristic of EBPR
radation [82]. sludge.
The ability to synthesize intracellular polyP granules is
widespread among bacteria [61], and not too surprisingly
6. The microbiology of EBPR ^ the culture-dependent other Gram-negative and Gram-positive bacteria cultured
approach from EBPR plants showing P removal have also been
considered as potential PAO [98,99]. These include isolates
Fuhs and Chen [15] were the ¢rst to isolate bacteria of members of the actinobacterial genus Tetrasphaera
from biomass with a high P removal capacity and they [100,101]. Some ¢lamentous bulking and foaming bacteria
identi¢ed their isolates as members of the genus Acineto- like Microthrix parvicella and Nostocoida limicola II, com-
bacter in the Q-Proteobacteria. For the next decade and monly seen in EBPR systems [102], both stain positively
more, these observations from culture-dependent tech- for polyP granules in pure culture and in situ [62,103,104].
niques largely determined the course of EBPR microbio- However, in most cases there is no evidence that these and
logical research. Many other groups around the world also the other isolates behave according to the predictions of
successfully isolated and grew bacteria from laboratory the models, because of lack of relevant information. The
and full-scale EBPR plants using di¡erent selective media, Gram-positive coccus Microlunatus phosphovorus that ac-
and in nearly all cases the majority was identi¢ed as Aci- cumulates polyP to a very high level [105] conforms par-
netobacter spp. [83,84]. So the conclusion was enthusiasti- tially to the models by assimilating P aerobically and re-
cally reached that the PAO in EBPR systems were mem- leasing it anaerobically. Yet M. phosphovorus cannot
bers of this genus. Micromanipulation [85] was used to try assimilate acetate anaerobically, although other substrates
to culture the clustered cells from EBPR plants, believing including glucose are used to synthesize polyglucose stor-
these were certain to be authentic PAO, but very few clus- age compounds. NMR studies have con¢rmed that the
tered cells grew to give colonies. Of those that did grow, metabolism of M. phosphovorus is not that expected of a
many did not identify as Acinetobacter spp. [86,87]. It has PAO [106]. Anaerobically assimilated glucose was not
been suggested [18] that the clustered cells may have been used for growth but instead was fermented to acetate,
representatives of bacterial genera that are not easily cul- which was stored unmodi¢ed with no conversion to
tured. In retrospect, because of the prevailing belief that PHA, but then respired aerobically to support growth.
acinetobacters were the PAO, sadly little e¡ort was made Molecular studies carried out so far (see Section 9) would
to identify the other colonies that grew from the isolated suggest that M. phosphovorus is not a dominating popula-
clusters in these studies. tion in EBPR systems.
Considerable e¡ort was also directed at resolving the An EBPR isolate ‘identi¢ed’ as a Lampropedia sp. only
taxonomy of these EBPR isolates of Acinetobacter. Several from its morphological features [107], which incidentally
studies showed that although most belonged to Acineto- were di¡erent to those usually associated with PAO in
bacter junii, Acinetobacter lwo⁄i and especially A. johnso- EBPR systems, came closer to ¢tting the expected metab-
nii, for reasons that are still not clear [87^90], many could olism of a PAO. It assimilated acetate anaerobically with
not be identi¢ed [91,92]. This partly re£ected the identi¢- corresponding PHA synthesis and P release, although the

acetate uptake to P release ratio, which is known to vary trends. For example, ubiquinone Q-8, diagnostic of the
considerably in di¡erent EBPR biomasses [17,21,44], was L-Proteobacteria was the most abundant quinone in both
lower than that predicted by the models. Disappointingly, EBPR and non-EBPR plant biomass samples from both
no further data on its physiology or identity are available full-scale and laboratory scale systems, and not Q-9 asso-
in the literature. ciated with the Q-Proteobacteria which include Acineto-
bacter spp. [114]. The Actinobacteria and K-Proteobacteria
possessing ubiquinone Q-10 were also thought to be
7. The microbiology of EBPR ^ the culture-independent present in large numbers, and then the Gram-positive
approach: chemical markers low mol% G+C bacteria, the Planctomycetes and the Cy-
tophaga^Flexibacter^Bacteroides (CFB) division in smaller
In our view the importance of obtaining pure cultures of proportions. Very few (3%) of the Actinobacteria appeared
EBPR bacteria cannot be overstated [108,109], because to be M. phosphovorus. In fact there were only small di¡er-
these are essential if any basic understanding of how these ences in quinone pro¢les between the communities of the

systems work and how they might then be manipulated is EBPR and non-EBPR systems analyzed, similar to other
to be obtained. However, the disadvantages and inherent experiences [116]. This trend could be interpreted as mean-
biases associated with analyzing natural microbial com- ing that only a small proportion of the total bacterial
munities using culture-dependent methods, sometimes re- community is involved in EBPR. Similar ubiquinone pro-
ferred to as the ‘plate count anomaly’ have been discussed ¢les were reported in the EBPR biomass samples
extensively in the literature [23,110], where the risk of end- [117,118], again with Q-8 and Q-10 more dominant than
ing up with what Amann and Ludwig [111] refer to as Q-9. However, in the latter study, the levels of Q-9 in-
‘laboratory weeds’ is well documented. Consequently creased and of Q-8 decreased in parallel with an increase
culture-independent techniques have been applied to in P removal, which suggested an increase in the cell num-
EBPR processes, and in almost every case the data ob- bers of the bacteria containing Q-9, i.e. the Q-Proteobac-
tained have raised further doubts about the importance teria, associated with EBPR. Earlier ubiquinone Q-9 was
of Acinetobacter spp. However, the evidence presented is shown to dominate in EBPR sludge [119], and so the evi-
not always convincing. For example, when £uorescent dence against a possible role for Acinetobacter under some
antibodies of an unknown speci¢city were prepared circumstances from these data is still equivocal. Menaqui-
against Acinetobacter [112] their application to EBPR none pro¢les may also change during EBPR, and have
systems only revealed very small percentages (3%) of been suggested as possible useful chemical indicators for
members of this genus. It could be argued that 3% of monitoring P removal [120].
the total number of cells present still represents several
million cells per gram of biomass. If all were accumulating
polyP at the levels acinetobacters are known to be capable 8. The microbiology of EBPR ^ the culture-independent
of, then they would represent a signi¢cant contribution to approach: molecular microbial ecology of EBPR
EBPR [29]. communities
The other approach has been to use chemotaxonomic
markers to indicate community composition, where the It is platitudinous to state that the impact of applying
presence of certain cell components may indicate the pres- molecular methods of analysis to activated sludge micro-
ence and relative abundances of particular bacterial pop- biology including EBPR in the past decade has been pro-
ulations [113]. These are considered [114] to lack the biases found. As well as showing how unsatisfactory culture-de-
inherent in the rRNA-based culture-independent methods pendent techniques are for providing a true representation
discussed below. They all extrapolate laboratory data to of microbial community biodiversity, this approach in
activated sludge systems, by assuming that the chemical some cases has changed completely the way we now
markers exploited are present in the same proportions and view processes like N removal [26,110]. Yet there still re-
same populations in activated sludge as in the pure cul- mains the challenge of using this information productively
tures from where the chemical data were originally ob- to help design and operate plants better [26,121]. The prin-
tained. Whether this extrapolation is justi¢able in all cases ciples and problems associated with the basic methodology
is not entirely clear. However, diaminopropane was used of the so-called 16S rRNA approach and its application to
as a marker for members of the genus Acinetobacter [115] activated sludge have been reviewed, where some of the
and shown to be present in low amounts in biomass from possible biases inherent in the steps of DNA extraction,
full-scale plants with high P removal, although detectable polymerase chain reaction (PCR), including chimera for-
at much higher levels in those with moderate EBPR ca- mation and cloning are discussed critically [25,26,110,122].
pacity. Little information is available on which DNA extraction
Where respiratory quinones have been used as markers method for activated sludge samples is least subject to bias
with both laboratory scale and full-scale EBPR systems, [26,110], but in many studies the numbers of sequences
most communities studied have given similar population found for the Actinobacteria were much lower than the

numbers of these bacteria detected by methods like £uo- 9. What have molecular techniques revealed about the
rescence in situ hybridization (FISH) described below, nature of EBPR microbial communities?
suggesting that their DNA was not as readily obtained
as that from other bacteria (e.g. [123,124]). One sensible It is clear from the literature that almost all studies
approach is to use a combination of di¡erent extraction applying these molecular methods have used laboratory
procedures and pool the DNA obtained for analysis, as scale reactors often fed with arti¢cial sewage, and have
adopted in [30] where it was also shown from clone library undertaken either FISH and/or clone library community
analysis that each extraction method sampled a phyloge- analysis. These studies are summarized in Table 1. As
netically di¡erent community. Regardless, the biases to- pointed out [30], only two published studies have used a
gether prevent the application of such methods to quanti- full cycle rRNA approach to full-scale plants [30,124], and
tative community analyses. They apply equally to neither were EBPR systems. The data invariably derive
techniques like denaturing gradient gel electrophoresis from analysis of a single sample of biomass, and very
(DGGE) [125], single strand conformation polymorphism few studies have followed changes in population composi-

(SSCP) [126] and T-restriction fragment length polymor- tion in response to changes in operational performances of
phism (T-RFLP) [127], all of which are molecular ¢nger- EBPR systems. Where this was done using DGGE pro¢l-
printing methods which have been applied to following ing, data were generated [136] suggesting that the commu-
changes in the compositions of activated sludge commun- nity composition in the sequencing batch reactor (SBR)
ities over time or in response to changing operational con- biomass studied became more highly specialized and less
ditions [24,26,110]. diverse as EBPR improved. All published studies have
Therefore, if semi-quantitative data are required, it is ignored the possible presence and importance of the Ar-
necessary to use techniques like FISH with £uorescently chaea in EBPR systems, although methanogenic bacteria
tagged 16S rRNA or 23S rRNA targeted probes [111, were detected using DGGE in conventional aerated acti-
128]. Probes can now be readily custom designed using vated sludge plants [137].
software packages like ARB [26,111] for targeting
populations at various phylogenetic levels from domain 9.1. Diversity of EBPR community populations
to species and below using 16S rRNA or 23S rRNA se-
quences derived from either pure cultures or cloned se- The data now available show that EBPR communities
quences. These techniques also have their limitations, are phylogenetically very diverse (see [26,110]), as are the
which are discussed in [24,25,111,128], and all depend non-EBPR activated sludge communities that have been
heavily on the sequence databases used for their design. analyzed [26,30,110,124]. Furthermore, all the 16S rRNA
It is now clear that many of the earlier probes designed clone library data suggest that most of the EBPR clones
from limited numbers of sequences are not speci¢c for the are very di¡erent to sequences from cultured bacteria,
organisms against which they were originally designed. although di¡erent studies have often yielded clones with
These include the ALF1b and EUB338 probes [26,111, similar but rarely identical sequences to each other
129]. [26,123,124,138]. Clone library data have often di¡ered
While revealing the population structure of activated from data from FISH analysis of the same communities,
sludge communities, by themselves these techniques tell as mentioned above. For example, several studies with
us little about the functions of these populations in situ. EBPR systems have detected a much higher proportion
This is critically important information if we hope to of the CFB division in their clone libraries than revealed
understand how these systems work. FISH in combination by FISH analyses [116,123,138,139]. This may be because
with microautoradiography (FISH/MAR) is one approach the probes presently available for the CFB group do not
to obtaining this information, where speci¢c populations encompass all naturally occurring strains [140], but the
can be identi¢ed with probing and their in situ physiology dangers of using clone library data quantitatively have
under various conditions determined by their abilities to already been mentioned. These are common trends, and
metabolize individually supplied radioactively labeled sub- may re£ect biases in DNA extraction e⁄ciency, as may the
strates of interest [130,131]. The other approach is to ana- di¡erences in the levels of phylogenetic diversity detected
lyze communities for detection of functional genes encod- in di¡erent activated sludge communities [110,126]. How-
ing enzymes important in metabolic processes of interest ever, in one study [29] with FISH, the phylogenetic diver-
in activated sludge, like nitri¢cation or denitri¢cation sity of the EBPR communities also appeared to depend on
[132^135]. As this methodology involves the same practical the in£uent composition. Using the same DNA extraction
protocols of DNA extraction and PCR discussed above, it method, a higher diversity, similar to that seen in full-scale
is also subject to the same biases, and the detection of a EBPR systems, was detected in a SBR community receiv-
particular gene is not by itself evidence that the gene is ing a complex feed containing several carbon sources, than
being expressed in situ. Furthermore, it is only appropriate in one being fed acetate alone. Thus, with acetate as sole
if each gene of interest is known together with its se- carbon source, the community was dominated by L-Pro-
quence. teobacteria in the L-2 subclass and very few Actinobacteria

Table 1
A summary of studies on the population composition of EBPR communities using culture-independent molecular techniques
Reactor con¢guration Reactor feed Methods of community analyses Major populations detected PAO identi¢ed Ref.
( s 10% of total)
Full-scale municipal plant complex FISH L-Proteobacteria n.d. [141]
Laboratory scale anaerobic:aerobic synthetic sewage with acetate as 16S rRNA clone library K-Proteobacteria (13%) n.d. [123]

SBR carbon source
L-Proteobacteria (33%)
Planctomycetes (13%)
Laboratory scale anaerobic:aerobic synthetic sewage with acetate as carbon 16S rRNA clone library K-Proteobacteria (16%) n.d. [145]
continuous £ow source
Actinobacteria (37%)
Laboratory scale anaerobic:aerobic synthetic sewage with acetate and peptone FISH L-Proteobacteria (45%) n.d. [147]
SBR as carbon sources
L-2-Proteobacteria (55%)
R.J. Seviour et al. / FEMS Microbiology Reviews 27 (2003) 99^127

Laboratory scale anaerobic:aerobic synthetic sewage with acetate as carbon FISH L-Proteobacteria (64%) n.d. [142]
SBR source
FEMSRE 768 3-4-03
Laboratory scale anaerobic:aerobic synthetic sewage with acetate as carbon 16S rRNA clone library, FISH L-Proteobacteria (89%) Rhodocyclus-related organisms, Candidatus [29]
SBR source ‘Accumulibacter phosphatis’
Laboratory scale anaerobic:aerobic complex synthetic sewage 16S rRNA clone library, FISH L-Proteobacteria (34%) Rhodocyclus-related organisms, Candidatus [29]
SBR ‘Accumulibacter phosphatis’
CFB group (16%)
Laboratory scale anaerobic:aerobic synthetic sewage with acetate and peptone quinone pro¢les, DGGE and 16S K-Proteobacteria (n.d.) n.d. [116]
SBR as carbon sources rRNA clone library
L-Proteobacteria (n.d.)
Cyaan Magenta Geel Zwart
Actinobacteria (n.d.) (with

quinone pro¢ling)
Laboratory scale anaerobic:aerobic synthetic sewage with acetate as carbon 16S rRNA clone library, FISH K-Proteobacteria (12%) Rhodocyclus-related organisms, Candidatus [139]
SBR source ‘Accumulibacter phosphatis’
L-2-Proteobacteria (80%)
Terrabacter-like CFB group (14%)
Laboratory scale anaerobic:aerobic synthetic sewage with acetate as carbon 16S rRNA clone library, SSCP CFB group (39%) Rhodocyclus-related organisms, Candidatus [138]
SBR source pro¢les ‘Accumulibacter phosphatis’
Laboratory scale continuous anaero- synthetic sewage with acetate, proprionate 16S rRNA clone library, DGGE, L-Proteobacteria (41%) Rhodocyclus-related organisms, Candidatus [62]
bic :aerobic reactor and peptone as carbon sources FISH ‘Accumulibacter phosphatis’
Pilot plant (A/O), no N removal municipal sewage with substrate FISH/MAR L-Proteobacteria Rhodocyclus-related organisms, Candidatus [37]
supplementation ‘Accumulibacter phosphatis’ but others
suspected
Q-Proteobacteria varied over time
Pilot plant (UCT) with N removal complex municipal sewage with substrate FISH/MAR L-Proteobacteria Rhodocyclus-related organisms, Candidatus [37]
supplementation ‘Accumulibacter phosphatis’ but others
suspected
Q-Proteobacteria
Laboratory scale anaerobic:aerobic synthetic sewage with mixture of VFA as FISH K-Proteobacteria (25%) Rhodocyclus-related organisms, Candidatus [156]
SBR carbon sources ‘Accumulibacter phosphatis’
Q-Proteobacteria (12%)
Full-scale (UCT) complex municipal sewage FISH K-Proteobacteria (10%) Rhodocyclus-related organisms, Candidatus [63]
‘Accumulibacter phosphatis’
Full-scale aerobic:anoxic complex municipal sewage FISH L-Proteobacteria (12%) Rhodocyclus-related organisms, Candidatus [63]
(orbal process with N removal) ‘Accumulibacter phosphatis’
109
n.d., no data.
were detected, while with a complex feed far more Actino- Although similarities were seen between the EBPR and
bacteria and members of the CFB division were present non-EBPR communities, Bond et al. [123] highlighted the
and fewer L-Proteobacteria were detected. increased presence of the Rhodocyclus group in clone li-
braries from their EBPR biomass. Without supplying any
9.2. Possible candidates for PAO direct proof, they implied that these organisms might play
a role in EBPR. Acinetobacters were again present in very
With these molecular methods Wagner et al. [141] were low numbers. Later, using FISH they showed [147] that
the ¢rst to raise further serious doubts about an important the L-Proteobacteria, especially members of the L-2 sub-
role for Acinetobacter spp. in EBPR systems, but left some class, together with the Actinobacteria dominated their
questions unanswered. Using a genus-speci¢c FISH probe EBPR community, adding further weight to the view
ACA23a, they suggested that Acinetobacter spp. were un- that both of these may be involved in EBPR, and support-
likely to be PAO, because they represented 6 10% of the ing data from other similar studies with FISH [141,148]
total bacteria present, and the L-Proteobacteria and Acti- and quinone pro¢ling [114]. However, other FISH studies

nobacteria were the dominant groups detected. These re- have suggested that the K-Proteobacteria, albeit detected
sults agree with the quinone pro¢ling study [114]. How- with the ALF1b probe now known to be of dubious spec-
ever, no mention was made of the identity of the clustered i¢city [26], may be more important than the L-Proteobac-
cells, so diagnostic of EBPR systems (see above), and 10% teria in some EBPR systems [142,143], and DGGE failed
of the total cell number may still represent enough cells to to show the presence of any L-Proteobacteria in the SBR
contribute markedly to P removal [29]. Furthermore, the EBPR community examined [116].
ACA23a probe was designed from a database containing It is also clear from clone library data that full-scale
sequences almost exclusively from clinical isolates of Aci- EBPR systems can contain large populations of phyloge-
netobacter, and so any unsequenced environmental strains netically diverse and novel ‘Chloro£exi’ or green non-sul-
may not necessarily have responded to it. In fact, sequen- fur bacteria [149], as apparently can conventional acti-
ces of two activated sludge Acinetobacter spp. from a full- vated sludge processes [30]. These populations appear to
scale treatment plant described in [124] did not match with be predominantly ¢lamentous in morphology and some of
the ACA23a probe sequence, and two additional 16S the morphotypes seen in both studies using FISH probes
rRNA targeted probes were described for this genus. designed against these clone sequences were novel, as well
However, when a range of activated sludge isolates repre- as being phylogenetically di¡erent to other ¢lamentous
senting several new Acinetobacter species [93,94] was tested ‘Chloro£exi’ reported previously in activated sludge like
against each of these probes, all responded to the ACA23a Eikelboom type 1851 [150] and N. limicola (II?) [151].
probe but only a few to the other probes reported in [124]. The clone libraries from both communities contained pre-
The general community population trends reported in dominantly the ‘Choro£exi’ 1 group but FISH analyses
[141] have been supported by much subsequent FISH revealed more ‘Chloro£exi’ 3 group bacteria in the
work on EBPR systems, especially in their conclusions EBPR communities [152,153]. Whether these were accu-
that Acinetobacter spp. are numerically minor populations mulating polyP was not discussed, but other indirect evi-
[37,142^144]. However, the use of biomass samples ob- dence suggests that some ‘Chloro£exi’ may have this abil-
tained from di¡erent experimental systems in each of the ity [142].
individual studies has meant that no universal view on
what might represent a typical EBPR community, and 9.3. The case for Rhodocyclus-related organisms as PAO
whether in fact it exists, is yet possible. While bacteria in
the L-Proteobacteria consistently emerge as major popula- Studies of the kind described above have been invalu-
tions, as they do with non-EBPR plant communities able in helping us understand EBPR microbiology, but
[26,30,110,141], the relative importance of the Actinobac- semi-quantitative community population analysis by itself
teria varies considerably, from being the most dominant in will not allow a precise identi¢cation of the nature of the
a continuous laboratory scale reactor (e.g. [145]) to only PAO. Elucidating structure^function relationships are re-
being detected in very small numbers in an SBR (e.g. quired, where the in situ ability of speci¢c populations to
[123]). Such di¡erences cannot always re£ect DNA extrac- accumulate P can be resolved, using techniques like FISH/
tion biases because both these studies used essentially the MAR [130,131] or FISH/histochemical staining [154]. In
same protocols [123,145]. Instead these data indicate that 1999 a link was ¢rst suggested between speci¢c bacterial
considerable variations may exist in EBPR community populations and their abilities to accumulate polyP in
compositions under di¡erent operational conditions [29], EBPR biomass [29]. Clone libraries were derived from
and also that several di¡erent populations may be in- biomass from an SBR fed acetate, using Rhodocyclus-spe-
volved in EBPR. As mentioned above, of the Actinobac- ci¢c primers. These were presumably selected because of
teria detected in EBPR systems, only very small numbers the predominance of these bacteria in other EBPR clone
have ever responded to the FISH probe designed against libraries [123], and because the dominating L-Proteobacte-
M. phosphovorus [37,146]. ria in FISH analysis were nearly all in the L-2 subgroup.

Table 2
Sequences and hybridization conditions for rRNA targeted probes for putative PAO and ‘G-Bacteria’/GAO in activated sludge
Trivial probe Probe sequence (5P^3P) rRNA target site Reported speci¢city Stringency : formamide Ref.
name (E. coli numbering) concentration (%)
MP2 GAGCAAGCTCTTCTGAACCG 16S rRNA (68^87) M. phosphovorus PAO?? 10 [146]
RHC175 TGCTCACAGAATATGCGG 16S rRNA (175^192) Rhodocyclus cluster 30 [29]
RHC439 CNATTTCTTCCCCGCCGA 16S rRNA (439^456) Rhodocyclus cluster 30 [29]
PAO462 CCGTCATCTACWCAGGGTATTAAC 16S rRNA (462^485) Candidatus ‘Accumulibacter 35 [139]
phosphatis’
PAO651 CCCTCTGCCAAACTCCAG 16S rRNA (651^668) Candidatus ‘Accumulibacter 35 [139]
phosphatis’
PAO846 GTTAGCTACGGCACTAAAAGC 16S rRNA (846^866) Candidatus ‘Accumulibacter 35 [139]
phosphatis’
PAO462b CCGTCATCTRCWCAGGGTATTAAC 16S rRNA (462^485) Rhodocyclus tenuis group 35 [63]

PAO846b GTTAGCTACGGYACTAAAAGG 16S rRNA (846^866) R. tenuis group 35 [63]
AMAR839 CTGCGACACCGAACGGCAAGCC 16S rRNA (839^860) Amaricoccus spp. ‘G-Bacteria’ 20 [198]
DEF438 CGCCTGAGACGATGATGAC 16S rRNA (438^456) De£uvicoccus vanus ‘G-Bacteria’ 20 [194]
MIC184 CATTCCTCAAGTCTGCC 16S rRNA (438^456) M. glycogenica GAO 20 [194]
KSB531 TTTCACTCCCGACGCAACAGAC 16S rRNA (438^460) clone sbr-gs28 20 [194]
actinobacterial ‘G-Bacteria’
TET63 GCTCCAGGGTCACCGTTC 16S rRNA (63^80) Tetrasphaera spp. ‘G-Bacteria’ 20 [194]
actino_1011 TTGCGGGGCACCCATCTCT 16S rRNA (1011^1029) T. elongata clones Ebpr19 and 20 30 [62]
GB CGATCCTCTAGCCCACT 16S rRNA (612^628) Q-Proteobacteria GAO ? 35 (70) [199]
GB1 and 2 GGCTGACTGACCCATCC 16S rRNA (587^604) Q-Proteobacteria GAO ? 20 (23) [199]
GB_2 GGCATCGCTGCCCTCGTT 16S rRNA (64^81) Q-Proteobacteria GAO ? n.d. [199]
GB_3 CCACTCAAGTCCAGCCGT 16S rRNA (642^659) Q-Proteobacteria GAO ? n.d. [199]
GB_4 GGCTCCTTGCGGCACCGT 16S rRNA (1020^1037) Q-Proteobacteria GAO ? 35 (40) [199]
GB_5 CTAGGCGCCGAAGCGCCC 16S rRNA (69^86) Q-Proteobacteria GAO ? n.d. [199]
GB_7 CATCTCTGGACATTCCCC 16S rRNA (1000^1017) Q-Proteobacteria GAO ? 35 [199]
GAOQ431 TCCCCGCCTAAAGGGCTT 16S rRNA (431^448) Candidatus ‘Competibacter 35 [200]
phosphatis’ GAO?
GAOQ989 TTCCCCGGATGTCAAGGC 16S rRNA (989^1006) Candidatus ‘Competibacter 35 [200,228]
phosphatis’ GAO?
n.d., no data.
These clones revealed sequences closely related to mem- ized with the probes and, crucially, these cells also con-
bers of the genus Rhodocyclus. FISH probing showed that tained polyP, as revealed directly by staining with methyl-
those cells hybridizing with their RHC439 and RHX991 ene blue [139]. Furthermore, when these probes were
probes designed against these sequences (see Table 2) had applied to full-scale EBPR systems, a direct correlation
the same morphology as cells staining aerobically with between the P removal capacity and the numbers of Rho-
DAPI for polyP and anaerobically with Nile blue A for docyclus-related organisms was demonstrated.
PHA, thus behaving in situ as the biochemical models
require [29]. Simultaneous DAPI or Nile blue A staining
and FISH could not be carried out, a problem encoun- 10. Structure:function relationships within EBPR
tered in other laboratories (see below) and so the link communities
made between Rhodocyclus-related bacteria and polyP
synthesis was still tentative. The putative Rhodocyclus-re- These two reports [29,139] have since led to intense
lated PAO could not be cultured, but they were named research activity, often using a range of methods, to ex-
Candidatus ‘Accumulibacter phosphatis’, according to the amine structure^function relationships in EBPR commun-
rules of the International Code for Bacterial Nomencla- ities. All have been aimed at determining the importance
ture. A tree showing the phylogenetic relationships among of these Rhodocyclus-related bacteria as PAO. It is now
the suggested PAO is given in Fig. 2. clear from such studies that these bacteria are important
Further weight was soon added to the hypothesis that populations in EBPR, and where their presence has been
these Rhodocyclus-related bacteria might be important sought they have usually been found. Thus, Rhodocyclus-
PAO [139]. Clone libraries from an EBPR SBR were ex- related organisms were detected in an EBPR SBR com-
amined and a group of clones also closely related to Rho- munity by SSCP ¢ngerprinting [138]. However, this story
docyclus was found. Again FISH using three probes has not yet reached a wholly satisfactory conclusion, and
PAO462, PAO651 and PAO846 designed against these much remains to be achieved before the ecology of the
sequences (Table 2) showed that clustered cells all hybrid- PAO is comprehensively understood.

and hence non-polyP accumulating populations using £ow

cytometry [142]. These fractionated cells can then be se-
quenced to generate clone libraries and hence be identi¢ed.
With this approach, Kawaharasaki et al. [142] showed that
although the L-Proteobacteria from FISH analyses were
the major populations in their biomass, very few £uo-
resced bright yellow, while almost all the K-Proteobacteria,
and many CFB, both present as minor components of the
community, did. No evidence for a role in EBPR for the
Actinobacteria was found. Later, with the same method-
ology it was suggested that Rhodocyclus-related bacteria,
although detected by DAPI staining as PAO, were not
necessarily the dominant PAO in their community, and

other bacteria may also be involved in EBPR [155].
10.2. Using FISH/staining and FISH/MAR to indicate

PAO identity
Combining FISH with DAPI and Sudan black staining

for detection of polyP and PHA has revealed [62] that the
Rhodocyclus-related PAO in their biomass accumulated
both polyP aerobically and PHA anaerobically, while the
actinobacterial N. limicola II [104] and Tetrasphaera elon-
gata [101], also present in large numbers, accumulated
polyP aerobically but not PHA in situ. However, in this
biomass, no cell inclusions were detected in either the
K-Proteobacteria or the CFB populations detected with
FISH. Also with FISH/staining, [156] could show their
Rhodocyclus-related organisms behaved according to the
biochemical models in an EBPR SBR system, synthesizing
PHA anaerobically and polyP aerobically. Staining meth-
ods together with FISH/MAR were employed on biomass
from an SBR operating under conditions with high and
low P removal [144]. The FISH analysis revealed that the
Fig. 3. Fluorescence micrographs of EBPR biomass showing typically
numbers of L-Proteobacteria and Rhodocyclus-related or-
clustered cells staining positive with DAPI for polyP (a), the same cells
£uorescing with the PAO462 FISH probe for Rhodocyclus-related organ- ganisms in the community corresponded directly to the
isms (b), and a composite of the two (c) showing that the cells £uoresc- level of P removal, and that most but not all the Rhodo-
ing with this probe are those containing polyP. cyclus-related organisms, which appeared as pleiomorphic
cells mainly in clusters, assimilated acetate and synthesized
10.1. Using DAPI staining to indicate PAO identity PHA anaerobically. Aerobically, most of the cells re-
sponding to the Rhodocyclus probes assimilated 33 Pi , ex-
Several studies have taken advantage of the ability of actly as the models require. FISH/MAR also demon-
DAPI to impart bright yellow £uorescence to polyP strated that some Actinobacteria could assimilate 33 Pi but
[62,84], so allowing the PAO to be detected by double these did not include the Tetrasphaera spp.
staining with FISH, a technique not without its problems Very similar results were reported in [37] using FISH/
in some laboratories [29,63,136]. This approach was used MAR, where both the Actinobacteria and Rhodocyclus-re-
[136] to show sequences of many of the dominant bands lated organisms were shown to accumulate P aerobically.
obtained by DGGE pro¢ling of EBPR SBR biomass were Uneven acetate and 33 Pi assimilation was also noted with
closely related to the Rhodocyclus-related clones obtained the Rhodocyclus-related PAO, and it was proposed that
by [139], and cells responding to the probes were similar in either some of these cells are physiologically inactive in
morphology to the DAPI staining clustered cells. A cluster situ or there is considerable physiological variation among
of PAO after simultaneous DAPI staining and FISH the populations detected by the probes used. It was also
probing with the PAO462 probe, showing the presence suggested [37] that not all clustered cells characteristic of
of polyP in these cells, is presented in Fig. 3. EBPR biomass are necessarily Rhodocyclus-related PAO,
Alternatively, DAPI stained cells can be enriched after and FISH showed that while large grape-like clustered
separation and recovery from most other non-£uorescing cells were L-Proteobacteria, the irregular clusters contained

Actinobacteria and the rounded clusters consisted of the UCT plant clustered closely with Candidatus ‘Accumu-
K-Proteobacteria. These interesting observations need to libacter phosphatis’, the clones from the other plant
be con¢rmed with other biomass samples, preferably formed an independent branch 6 97% similar to the
from full-scale plants. Furthermore, because the changes others, and thus possibly were not representatives of Can-
in population levels of Rhodocyclus-related PAO did not didatus ‘Accumulibacter phosphatis’. Two additional
correspond to changes in the P removal capability of their probes (Table 2) were designed from these sequences,
plants, it was concluded that other bacteria are likely to be but have yet to be applied to other EBPR biomass samples
involved in EBPR. to determine if these or other phylogenetically di¡erent
Rhodocyclus-related organisms may be present.
10.3. Using di¡erential centrifugation to indicate A full-scale plant was also deliberately selected in [158]
PAO identity because of reservations the authors held about the validity
of using laboratory scale systems for studies of this kind.
Another approach has been to separate and enrich the The PAO detected using DAPI staining in their biomass

PAO by buoyant density centrifugation in Percoll, based were typical but very large clustered cells with an unusual
on the assumption that PAO cells because of their intra- response to the Gram stain, and after their reactions to
cellular polyP, PHA and glycogen granules have a higher exposure to a series of stains, antibiotics and enzymes,
density than cells with no inclusions [84]. These fraction- these cells were considered to possess properties not of
ated enriched PAO cells may then be analyzed phyloge- bacteria but of eukaryotic yeast spores. They were success-
netically. With this approach, sequence analyses of bands fully cultured, but surprisingly no molecular or other
from DGGE ¢ngerprints of 16S rDNA fragments in the methods were applied to disprove unequivocally their pro-
enriched populations from laboratory and pilot scale karyotic nature or to con¢rm they were eukaryotic cells.
EBPR systems, revealed most cells belonged to the Q-Pro- Therefore their nature remains unresolved and the skeptics
teobacteria [157]. The L-Proteobacteria were poorly repre- unsilenced.
sented. It is most unlikely that a complete separation of
PAO is achievable by either centrifugation or £ow cyto-
metry used in the studies described above, bearing in mind 12. Denitrifying EBPR systems
how the biomass is aggregated into £ocs, and so some
caution is needed in interpreting quantitatively data from Earlier in this review the importance of ensuring that no
studies of this kind. nitrate was allowed to enter the anaerobic zone to stable
EBPR was mentioned, and the in£uence this requirement
has had on the design of EBPR plants emphasized. The
11. Studies with full-scale EBPR plants presence of nitrate is thought to provide denitrifying bac-
teria with an opportunity to remove the selective advan-
Apart from the work by Croceti et al. [139], most of the tage the PAO have in these systems by out-competing
discussion so far has dealt with laboratory scale systems, them for the metabolizable substrates available there
and the role of Rhodocyclus-related PAO in full-scale [159^163]. However, the inhibitory e¡ect of nitrate may
EBPR systems has received comparatively little attention be more direct because nitric oxide, an intermediate in
(Table 1). However, both £ow cytometry of DAPI stained denitri¢cation has been shown to prevent anaerobic P re-
cells and buoyant density centrifugation have been applied lease in EBPR sludge by inhibiting the adenylate kinase
to biomass samples from full-scale plants [63]. Rhodocy- (see Section 14) involved in polyP degradation [164].
clus-related PAO were plentiful and so probably are im- However, it has frequently been demonstrated that in
portant in full-scale EBPR. This may not always be the the absence of any exogenous carbon sources in the an-
case and their in£uence on EBPR may depend on the aerobic zone, EBPR may in fact occur in the presence of
plant con¢guration. Thus, in a UCT-con¢gured plant, nitrate. This is presumed to happen by organisms using
73% of PAO, representing about 20% of the total bacterial intracellular storage compounds like PHA as carbon and
population were Rhodocyclus-related organisms, and all energy sources to assimilate P and synthesize polyP, as
were storing polyP [63]. However, in aerobic :anoxic PAO do, but using nitrate not oxygen as their terminal
EBPR plants only 26% of the PAO, representing about electron acceptor [17]. Consequently two groups of PAO
6% of all cells, were Rhodocyclus-related organisms. One are now recognized, the PAO and so-called denitrifying
possible reason given for this lower ¢gure was the absence PAO or denitrifying polyP accumulating organisms
of a strictly anaerobic zone in this plant. About 50% of (DNPAO) [165]. Some evidence is also available to suggest
these Rhodocyclus-related organisms contained no DAPI that nitrite may be used by DNPAO in P uptake [166,167].
staining polyP, and thus probably played no role in The attractions of exploiting such populations in waste-
EBPR, which is similar to the ¢ndings in [37]. In their water treatment include the possibility of achieving simul-
generated clone libraries from the two plants, Zilles et taneous removal of N and P, producing less sludge and
al. [63] also showed that while most of the clones from with no need for aeration, having a process that is cheaper

to run [161]. Furthermore, a lower requirement for metab- common bands was taken to imply that some populations
olizable substrates than in conventional EBPR systems were capable of both aerobic and anoxic EBPR, and two
overcomes a major problem associated with treating mu- of these bands were recovered and sequenced. Both were
nicipal wastes using EBPR [21,161]. Some systems taking L-Proteobacteria, closely related to Rhodocyclus and Dech-
advantage of these have been conceived. For example, lorimonas, known also to appear frequently in clone libra-
Bortone et al. [168] described the DEPHANOX process ries from aerobic:anaerobic EBPR type plants [26,111].
based on DNPAO activities, and with a separate aerobic FISH probing showed the Rhodocyclus-related population
reactor for nitri¢cation, were able to achieve very high P was dominant in both communities but Dechlorimonas was
uptake rates in their anoxic reactor. These bene¢ts have not. It would have been interesting to see what emerged if
not convinced all, and it has been argued strongly on the PCR protocol had used primers speci¢c for functional
theoretical grounds [169] that because of their unpredict- genes for denitrifying enzymes like the nirS and nirK genes
able nature and reduction in EBPR capacity, denitrifying [132,133], and FISH/MAR had been applied to these com-
EBPR systems are not worth bothering with. munities to assist in identifying the denitrifying popula-

Even less is known about the community structure and tions in situ [167]. SSCP was used in [138] to follow the
identity of DNPAO than about the microbiology of anaer- changes in microbial community structure following a
obic:aerobic EBPR systems discussed above, and very few shift from an anaerobic:aerobic EBPR system to an anaer-
microbiological studies on denitrifying EBPR systems obic:anoxic one. Because the pro¢les obtained were very
have been reported. Culture-dependent techniques have similar in each, many of the populations there were
been used to study communities in an anaerobic :anoxic thought to carry out EBPR under aerobic and anoxic
SBR which stored P anoxically [170,171]. A range of conditions, agreeing with the conclusions of Ahn et al.
Gram-positive and Gram-negative bacteria were isolated [174].
with techniques selective for denitrifying bacteria and
some of these isolates could assimilate P aerobically in
pure culture [170]. Later it was reported that two of the 13. Why do EBPR systems often behave unpredictably ?
cultures obtained, Agrobacterium tumefaciens and Aqua-
spirillum dispara, used both oxygen and nitrate as electron As already mentioned, both laboratory scale and full-
acceptors for P assimilation, but polyP production in these scale EBPR systems are notorious for their unreliability.
isolates was not reported [171]. The problems associated One of the main motivating factors for elucidating the
with using culture-dependent methods in studies of this microbiology of EBPR systems is to understand why this
kind have been discussed earlier, and molecular methods deterioration in performance occurs and how it can be
directed at detecting functional nirS or nirK genes remedied. There may be reasons other than microbiolog-
[132,133] in these communities would have clearly en- ical ones for this, and the dependence of stable EBPR on a
hanced the value of the study. ready supply of Mg2þ has already been discussed as one
The physiology of other isolates from denitrifying sys- possible example. The other possibility, which has at-
tems has also been reported. For example, Paracoccus de- tracted considerable interest, is that under certain operat-
nitri¢cans synthesized polyP in the presence of both nitrate ing conditions the PAO community is harmed or disad-
and oxygen, but only with an external carbon source vantaged in some way and so EBPR eventually
[172,173]. Thus, although this organism produced PHA, deteriorates. There are many publications dealing with
it could not be used anoxically to support polyP forma- how certain operational conditions like over-aeration
tion, which is fundamentally di¡erent to the classic view of [175] and aerobic :anaerobic contact times [176,177] may
PAO described above. Furthermore, simultaneous P re- a¡ect EBPR performance. Unfortunately these are mainly
moval and denitri¢cation occurred in the absence of anae- of little more than anecdotal interest because they include
robic:aerobic cycling, again considered previously to be an no structure^function analyses of the changes that are
essential prerequisite for conventional EBPR (but see be- thought to have taken place in the microbial community,
low). It was also suggested that the ability to accumulate and their general applicability to EBPR operation is un-
polyP appeared to be widespread among denitrifying bac- likely.
teria, and Pseudomonas strain JR12 was shown to store
polyP without synthesizing PHA, implying considerable 13.1. In£uence of in£uent substrate composition on
variations may exist in the biochemistry of DNPAO. EBPR performance?
How relevant this information is to understanding denitri-
fying EBPR activated sludge systems is not clear. Most attention has been focused on the impact of in£u-
A molecular approach using PCR-DGGE analysis was ent substrate composition, especially the presence of glu-
adopted in [174] to study community structure of their cose on EBPR failure, as discussed below, although many
DNPAO community. Marked di¡erences in DGGE pro- reports of successful EBPR in the presence of glucose have
¢les of 16S rDNA fragments in biomasses from DNPAO appeared in the literature, as mentioned earlier [52,55,59,
and PAO communities were revealed. The presence of 178,179], and empirical biochemical models have been

published outlining how this may occur (e.g. [55]). Un-

fortunately, these studies have provided no microbial com-
munity compositional data. Particular attention has been
paid to the role, if any, in such EBPR failure of the so-
called ‘G-Bacteria’. This term describes a morphotype of
cocci, which often appear in activated sludge and are now
known to be very diverse phylogenetically (see Fig. 2),
containing both Gram-positive and Gram-negative mem-
bers. Cells are distinctively arranged in tetrads or in clus-
ters, as shown in Fig. 4, and have been referred to as
tetrad-forming organisms (TFO) in [180]. Much of what
is known about their taxonomy and physiology has been
reviewed [181].

13.2. What are these ‘G-Bacteria’ and GAO?
First reported in activated sludge in [182], their possible

involvement in EBPR failure came initially from the ideas
of Cech and Hartman [50,51], who noticed large numbers
of ‘G-Bacteria’ in a plant fed glucose and showing poor
EBPR. They hypothesized from the chemical transforma-
tions taking place with their mixed enriched culture that Fig. 4. Light micrographs of the microscopic appearance typical of ‘G-
these ‘G-Bacteria’ may out-compete the PAO in anaerobi- Bacteria’/GAO in EBPR systems. a: Individual tetrads of cocci. b: Tet-
c :aerobic EBPR systems under some operational condi- rads arranged in sheets.
tions. The ‘G-Bacteria’ were thought to achieve this by
more e¡ectively assimilating some substrates anaerobical-
ly, such as when the feed contained glucose as well as nence or otherwise of the GAO in EBPR communities,
acetate. The hypothesis was that the ‘G-Bacteria’ assimi- and at the appropriate value both the GAO and PAO
lated glucose anaerobically better than the PAO and used could coexist there.
it eventually for PHA production, which could then be There is no question in our view that the GAO exists in
metabolized under subsequent aerobic conditions for gly- EBPR plants as a group physiologically distinct from the
cogen formation. Thus, these ‘G-Bacteria’ were selectively PAO. However, what they are and whether the ‘G-Bacte-
favored and became dominant populations, and as they ria’, often seen microscopically to dominate the biomass in
now synthesized no polyP aerobically, then EBPR gradu- failing EBPR plants, are GAO is still largely unknown,
ally diminished. since in no case have their abilities to accumulate glycogen
Similar experiments were carried out by us [33,183] and in activated sludge been directly demonstrated [181]. In-
similar conclusions were reached. Anaerobic assimilation stead, much of the evidence for GAO having a functional
of glucose and other sugars could occur in the absence of role in EBPR failure comes from an interpretation of
any P release, consistent with polyP not being used as trends in chemical transformations using enriched mixed
energy source for substrate uptake, and hence di¡erent communities, similar to the development of biochemical
to that proposed by the EBPR biochemical models. It models for EBPR described above, and are subject to
was suggested that this re£ected the dominance of a phys- the same criticisms. With just a few exceptions, these
iological group of cells using stored intracellular glycogen chemical changes alone have been used to interpret the
synthesized aerobically as the energy source for anaerobic likely population structures of these communities, and
substrate assimilation. Some of the bacteria that appeared no accompanying detailed microbial community analyses
as large cocci in their biomass samples were considered have been performed. Consequently, contradictions and
responsible for these chemical changes because of their inconsistencies abound in the literature, and clear, gener-
dominance. These cells were described not in terms of a ally applicable principles are hard to ¢nd among the pleth-
morphotype but as a phenotype, the so-called glycogen ora of data. Most of the studies have also been carried out
accumulating organism (GAO). This is a descriptive in laboratory scale reactors using a synthetic sewage feed.
term that has been adopted in many later studies [183^
186], but in a sense it is confusing because according to 13.3. Evidence for a role for GAO/‘G-Bacteria’ in EBPR
the EBPR biochemical models, the PAO also accumulate systems from mixed culture studies
glycogen. It was suggested [183] that the chemical nature
of the substrate was less important than the phosphorus/ Filipe et al. [187,188] have suggested that the pH of the
carbon (P/C) ratio of the feed in determining the promi- anaerobic and aerobic zones is important in determining

whether PAO or GAO will dominate, and hence whether EBPR failure occurs in the absence of any glucose in the
EBPR is successful. In the anaerobic zone, acetate uptake feed [37,138,139,192]. Glucose dosing of a reactor [37] nei-
rates were considered to be adversely a¡ected in GAO but ther led to this morphotype appearing in larger numbers in
not in PAO at high pH, while in the aerobic zones ^ their biomass nor caused EBPR to fail, but instead glucose
although growth rates, PHA consumption rates and P enhanced it. Hence the question is still unanswered. How-
uptake rates in GAO were una¡ected by pH ^ they were ever, with a few exceptions, the ‘G-Bacteria’ in these
markedly lowered for PAO at a lower pH (6.5). So, while plants have not been obtained in pure culture, and so
the GAO were advantaged at low pH, the PAO out-com- we still know little about what they are or what they do
peted the GAO at pH 7^7.5. Jeon et al. [178] also followed there. Bearing in mind their known phylogenetic diversity
the in£uence of pH on EBPR in an SBR, and interpreted [181], as shown in Fig. 2, even if these ‘G-Bacteria’ are not
their chemical data as follows. At a controlled pH of 7, microscopically distinguishable from each other, their sig-
GAO dominated ; with no control, the pH rose, the PAO ni¢cance in plants may be di¡erent, and conclusions about
took over and good EBPR was achieved. They also sug- performance of plants containing them based only on mi-

gested that pH might be used to manipulate the competi- croscopic appearance of the biomass, should be treated
tion between the two groups, but this would only be val- cautiously until more is known about their identity and
uable if generally applicable to all EBPR systems. in situ physiology [181].
Unfortunately the opposite situation has been found Isolates of the original ‘G-Bacteria’ morphotype seen by
[189], where a PAO metabolism dominated that of GAO Cech and Hartman [51] were placed in a new genus Ama-
at pH 7. Possible reasons for such disagreements are ricoccus, in the K-Proteobacteria [193]. Since then other
many, but one likely explanation is that the communities K-and L-proteobacterial, as well as several actinobacterial
in each study were quite di¡erent. Whether this is the case ‘G-Bacteria’ have been grown and the pure cultures char-
is not known, but this example reveals the problems in acterized [181], yet their role in EBPR failure is also not
usefully interpreting data of this type. clear. For example, while isolates of Amaricoccus exam-
Similar studies have also suggested that plant operating ined so far fail to synthesize polyP, they do produce gly-
temperatures can determine EBPR stability by a¡ecting cogen aerobically [194], as expected of GAO, but no gly-
the balance between the PAO and GAO populations cogen synthesis could be detected under anaerobic
[190]. Thus, at 20‡C the PAO were thought to dominate, conditions. In addition, pure cultures of Amaricoccus ka-
while at 30‡C GAO were considered more in£uential, and plicensis could assimilate neither glucose nor acetate an-
when this 30‡C biomass was examined microscopically, aerobically, and PHA synthesis, while occurring aerobi-
many ‘G-Bacteria’ morphotypes were in fact visible. Ear- cally, did not take place anaerobically [195]. Thus, as
lier, community fatty acids had been used as biomarkers with Acinetobacter and the EBPR biochemical models,
for these two physiological groups of bacteria [191], Amaricoccus does not behave in pure culture as the models
although their suitability for this purpose in the absence for GAO/‘G-Bacteria’ demand [50,51].
of any knowledge of their identity was not tested. Fatty
acid pro¢les were then used to ¢ngerprint their total mixed 13.5. The identity of the ‘G-Bacteria’ and GAO:
communities, and the interpretation reached was that dif- culture-independent approaches
ferent PAO species were capable of using acetate or glu-
cose, and that GAO and PAO grown with the same sub- Partly as a consequence of unsuccessfully trying to cul-
strate also represented di¡erent species. In addition the ture these ‘G-Bacteria’, culture-independent methods have
GAO could successfully compete with PAO only at long been applied in attempts to identify them in EBPR bio-
sludge retention times. These are all interesting ideas but mass. FISH could show that the abundant large coccoba-
in the absence of relevant microbiological information and cilli seen in biomasses with poor EBPR were members of
no identi¢cation of the GAO populations, the studies pro- the L-Proteobacteria, but were not identi¢ed further [147].
vide us with little real understanding of EBPR community They were described as GAO, but with no biochemical
function. basis. From the ideas in [50,51], it was proposed that sev-
eral di¡erent bacterial populations might be involved in
13.4. The identity of the ‘G-Bacteria’ and GAO : EBPR deterioration as no K-proteobacterial ‘G-Bacteria’
culture-dependent approaches (i.e. Amaricoccus) could be detected by FISH in the bio-
mass studied [147]. This idea receives some support from
What is known about the microbial nature of the ‘G- the data in [192], where members of the Q-Proteobacteria,
Bacteria’ or GAO and their possible role in glucose-in- identi¢ed from DGGE pro¢ling of community DNA, were
duced EBPR deterioration? It is now clear that cells ¢tting detected in a deteriorating EBPR plant. Similarly, SSCP
this morphotype description (Fig. 4) can be found, often analysis of a community with poor EBPR capacity de-
in large numbers, in many anaerobic:aerobic EBPR plants tected Q-Proteobacteria with sequences almost identical
[181]. These cells sometimes occur in communities in the [138] to those found by Nielsen et al. [192]. FISH probes
absence of any evidence of EBPR failure [29,145] or where designed against the partial sequences of these fragments

[192] imparted £uorescence to large coccoid cells seen fore it would seem that populations, all very closely re-
dominating their biomass. These cells were similar in ap- lated phylogenetically and often indistinguishable by FISH
pearance to those reported in [29], which by FISH analy- may behave physiologically quite di¡erently to each other
ses were also members of the Q-Proteobacteria, although in activated sludge.
this community showed good EBPR performance, indicat- The work reported in [144,180,194] supports this. The
ing that these populations may be common in most anaer- tetrad-forming bacteria that dominated SBRs fed glucose
obic:aerobic systems. and with no EBPR [180] were members of the L-Proteo-
Similar results were reported in [62], where sequences bacteria, Q-Proteobacteria and Actinobacteria, and many of
from large cocci were also detected in clone libraries these were distinctive large cocci. As well as phylogenetic
from an EBPR SBR. These were considered to represent diversity, they also demonstrated considerable in situ
two physiologically di¡erent populations with a shared physiological variation in this morphotype. Thus, the
morphology. FISH showed that cells responding to the L-Proteobacteria stained DAPI 3ve and PHA +ve, which
probes in [192] and those designed using clone library di¡ers from the behavior of the cocci reported earlier by

sequences from this community (given in Table 2), were the same authors [62], while the Q-Proteobacteria were
also large and coccoid in shape. It is unlikely that they are both DAPI and PHB 3ve. Some Actinobacteria, also ap-
the same bacterial populations as those reported in [196] in pearing as large cocci in tetrads, were both PHA and
an EBPR system fed glucose, or in [29] (see below). High DAPI 3ve or PHA 3ve and DAPI +ve. In the latter
numbers of tetrad-forming ‘G-Bacteria’ in a biomass also case this implies an ability to synthesize polyP with no
with good EBPR capacity [145] would add to the view that formation or degradation of PHA, which is not as the
the presence of this morphotype does not necessarily in- biochemical models predict.
dicate potential EBPR deterioration, or vice versa [197]. None of these populations was identi¢ed further, but
These ‘G-Bacteria’ were not identi¢ed precisely, but li- subsequent FISH analysis [199] using probes (Table 2)
brary analysis yielded a few clones similar to Amaricoccus, designed against clones from systems with good and
which is known from FISH analysis to be common in poor EBPR capacity showed the Q-Proteobacteria were
many full-scale plants [198]. Most were Actinobacteria large coccoid/rod-shaped bacteria. These were widely dis-
closely related to Tetrasphaera spp. [145], reported to be tributed in both laboratory and full-scale EBPR as well as
capable of growing as tetrads and synthesizing polyP conventional systems. These seem to be very similar phy-
[100]. logenetically to the bacteria reported in [156,200] in both a
laboratory scale SBR and full-scale systems, which were
13.6. Structure :function studies with non-EBPR named Candidatus ‘Competibacter phosphatis’. These
communities and the role of ‘G-Bacteria’ and GAO stained for PHA but not polyP, and were assumed to be
GAO, although their ability to synthesize glycogen was
As with the PAO, knowing the identity of these ‘G- not reported, but implied [186]. They also appear to be
Bacteria’ is not su⁄cient to understand their role in anaer- the same populations originally identi¢ed using FISH (Ta-
obic:aerobic EBPR systems. For example, if the hypoth- ble 2) as L-Proteobacteria in [147], a confusion that may
esis that these may out-compete the PAO is to be tested, re£ect the di⁄culties di¡erentiating between members of
then the questions of which organic substrates they are these two divisions with the currently available FISH
able to assimilate anaerobically, and what intracellular probes for them [199].
storage compounds are synthesized in situ need to be ad- Several methods including FISH/MAR and FISH/stain-
dressed. A clearer picture of what they might be doing ing were employed to try to understand the community
there comes from applying techniques like those described structure and function in an anaerobic:aerobic SBR fed
earlier for analyzing EBPR communities that reveal struc- acetate and glucose, where the biomass contained a high
ture^function relationships of communities containing glycogen level and was dominated by ‘G-Bacteria’ [194].
them. It is clear now that many can synthesize storage From FISH analysis, this community contained mostly
polymers (Table 3). For example, the large cocci described large and small cocci belonging mainly to the K-Proteo-
in [196], which were not probed by FISH, contained both bacteria, Q-Proteobacteria, Actinobacteria and the low
polyP and glycogen granules from TEM studies, but mol% G+C Gram-positive bacteria. The latter consisted
whether any PHA was produced is not clear. FISH/stain- of both large cocci typical of ‘G-Bacteria’ in tetrads and
ing showed that the Q-Proteobacteria seen in [29] produced small clustered cocci. These populations were held respon-
PHA anaerobically but not aerobically, and no polyP was sible for the production of considerable amounts of lactic
detected in cells exposed to aerobic conditions. It was acid, a feature of other SBRs fed glucose [52,55,59], and,
suggested that these characteristics might mean that such like the PAO and ‘G-Bacteria’, were thought to be selec-
bacteria could compete anaerobically with the PAO for tively advantaged by a glucose feed together with anaero-
acetate. Thus, they behaved in a similar way to the Q-Pro- bic:aerobic biomass cycling. Very few L-Proteobacteria
teobacteria in [192] but not like the large cocci reported in could be detected in this community, and the Q-Proteobac-
[62], which could synthesize both PHA and polyP. There- teria seen there did not respond to the FISH probes in

[156,186,200]
[192]. Clone library analysis and DGGE pro¢ling of this
[193,195]
[194,201]
community were used in attempts to identify these popu-
lations, but none of the Proteobacteria or low mol% G+C
[106]
[180]
[180]
[180]
[194]
Ref.
[62]
[29]
[62]
Gram-positive bacteria could be identi¢ed, However,
among the Actinobacteria, Micropruina glycogenica [201]
Synthesis
A summary of the known anaerobic substrate assimilation patterns and intracellular storage polymer production by the ‘G-Bacteria’/GAO in anaerobic :aerobic activated sludge systems
and a clone with a sequence almost the same as that

polyP
from clone sbr-gs28 from an SBR fed glucose [55,196]

3ve
3ve
3ve
3ve
3ve
3ve
3ve
+ve
+ve
+ve
Z
were detected. FISH using probes (Table 2) designed
against these populations showed that they accounted
glycogen
Polymer
for almost all of the Actinobacteria in this community.

n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
+ve
+ve
No Amaricoccus spp. could be detected by FISH [198],

raising further questions as to their role in out-competing
Aerobic
the PAO for substrates anaerobically in EBPR deteriora-

PHA
3ve
3ve
3ve
3ve
3ve tion [50,51].

n.d.
n.d.
+ve
+ve
+ve
Considerable physiological variation was noted in the

anaerobic substrate assimilation abilities of the ‘G-Bacte-
Synthesis
ria’ belonging to some of these divisions [199]. Thus, not

polyP
3ve
3ve
3ve
3ve
3ve
3ve
n.d.
n.d.
n.d.
n.d.
n.d.
all of the K-Proteobacteria took up acetate or glucose,

although almost all the Q-Proteobacteria assimilated ace-
tate but not glucose. The low mol% G+C Gram-positive
glycogen
Anaerobic Polymer
small cocci, assimilated only glucose, consistent with them

3ve
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
+ve
being lactic acid producers, while the large cocci assimi-

lated both acetate and glucose, as did M. glycogenica and
those cocci with the sbr-gs28 clone sequence. All except
PHA
3ve
3ve
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
+ve
+ve
+ve
the low mol% G+C Gram-positive bacteria and a few

K-Proteobacteria synthesized PHB anaerobically, and
glucose, not acetate in pure culture;
glucose and acetate in pure culture
none stored polyP aerobically. This evidence pointed to

none in pure culture; n.d. in situ
Anaerobic substrate assimilation
none in pure culture or in situ
M. glycogenica as the true GAO in this community, espe-

acetate and glucose in situ
cially since its ability to utilize glucose anaerobically for

glycogen synthesis in pure culture is known [201].
A similar approach was used to examine structure^func-
tion relationships in an SBR reactor fed acetate as the sole
glucose in situ
glucose in situ
glucose in situ
acetate in situ
acetate in situ
acetate in situ
carbon source over a range of phosphorus/carbon ratios

n.d. in situ
and in situ
where good or very poor EBPR was detected [144]. FISH

analysis revealed a marked shift in community structure as
EBPR capacity decreased. The community changed from
one where the K-Proteobacteria and to a lesser extent the
large coccoid rods
large coccoid rods
large coccoid rods
large coccoid rods
large coccoid cells
large coccoid cells

Cell morphology
cocci in clusters
cocci in clusters
cocci in tetrads
cocci in tetrads
cocci in tetrads
Q-Proteobacteria, both of which were predominantly the

morphotype of ‘G-Bacteria’, gradually took over from
in sheets
the L-Proteobacteria and Rhodocyclus-related organisms

as EBPR fell. No Acinetobacter was detected in any of
these communities, and neither were any low mol% G+C
Gram-positive bacteria, M. glycogenica [201] or clone sbr-
K-Proteobacteria
L-Proteobacteria
gs28, populations that dominated a glucose fed SBR [194].

Q-Proteobacteria
Q-Proteobacteria
Q-Proteobacteria
Q-Proteobacteria
Actinobacteria
Actinobacteria
Actinobacteria
Actinobacteria
Actinobacteria
Phylogenetic
The increase in the proportions of the K-Proteobacteria

Determined by direct chemical analysis.
a⁄liation
paralleled the increases in biomass glycogen content, but

again FISH probing (Table 2) could not detect Amaricoc-
cus spp. in high numbers. FISH/MAR showed that both
Determined using staining.
K- and L-Proteobacteria assimilated acetate anaerobically

Candidatus ‘Competibacter
Determined using MAR.
in situ but not P aerobically, and no anaerobic glucose

Organism/clone name
assimilation was detected in this community. These Pro-

Amaricoccus spp.a;b
M. phosphovorusa;b
Tetrasphaera spp.a
M. glycogenicab;c
teobacteria could not be identi¢ed further, but the data

Clone sbr-gs28c
indicated that they were physiologically quite di¡erent to

HGC-TFO1a
phosphatis’
Unnameda
Unnameda
those detected in the SBR with a glucose feed described

L-TFO1a
Q-TFO1a
Table 3
above. The data also con¢rmed that these ‘G-Bacteria’

morphotypes may exist in communities showing good
b
a

EBPR capacity, and that in the community with poor Rhodocyclus-related organisms. The mRNA transcribed
EBPR, glycogen formation using acetate was probably from it could also be detected. The enzyme was apparently
from the activities of the K-Proteobacteria, in the absence associated with the particulate fraction of lysed sludge
of M. glycogenica. What is known about anaerobic sub- [208], which might explain the lack of success in detecting
strate assimilation and intracellular storage polymer syn- it in earlier studies. Furthermore, the requirement of this
thesis from these di¡erent studies with the ‘G-bacteria’/ enzyme for Mg2þ could be one reason why this cation is
GAO is summarized in Table 3. essential for successful EBPR [77].
More studies are needed where communities with well- This enzymatic information is very important, since it
de¢ned chemical properties are analyzed in this way to should now provide the tools to help understand how
clarify the considerable confusion about the nature of PPK synthesis and expression in this PAO population is
the ‘G-Bacteria’ and GAO that currently exists. It is im- a¡ected in situ by factors like plant operating conditions,
portant to identify more clearly the major functional pop- and provide the means to monitor plant performance in a
ulations responsible for the chemical transformations tak- highly relevant way. More studies like this are needed to

ing place, using the whole range of culture-independent analyze for the presence of other ppk genes in samples,
methods now available, to design more speci¢c FISH including those from full-scale EBPR plants. Some evi-
probes for their detection, and to examine their in situ dence is available on regulation of synthesis of this enzyme
physiology by MAR, before the important questions in pure cultures of Acinetobacter spp., where transcription
raised above can be answered. Then these techniques of the ppk gene is apparently induced by Pi starvation, a
must be applied to full-scale systems to see whether the condition which also induces the synthesis of an exopoly-
populations considered important from studies with labo- phosphatase, PPX, involved in polyP degradation [205].
ratory scale systems, where the selective pressures are con- Whether this is the situation in any EBPR PAO is not
siderably di¡erent, are also important there. known, but linking synthesis of these two enzymes ^ one
involved in polyP synthesis and the other in its degrada-
tion ^ was thought to explain how PAO store polyP
14. How is EBPR biochemistry regulated ? aerobically to be used later anaerobically. The suggestion
was made [205] that the Pi levels should receive more
We have made a strong case for the need to understand consideration as possible key regulators of EBPR, because
the structure and function of the microbial communities in of their ability to a¡ect the expression of several enzymes
these systems. Equally important is the need to understand that may be involved in polyP metabolism.
how the biochemistry of EBPR is regulated. For example, There is some evidence to suggest that the polyP :AMP
what is known about the enzymes involved in the synthesis phosphotransferase:adenylate kinase system is associated
and degradation of polyP, PHA and glycogen in PAO, with anaerobic polyP degradation and ATP synthesis dur-
and what a¡ects their synthesis and levels of activity ? ing EBPR. High levels of its activity were detected in an
We know very little about this, and interpretation of EBPR biomass [44], and it has been characterized in
what is known, like much of EBPR, assumes that all A. johnsonii strain 210A [202,203]. However, while levels
PAO behave identically, which seems increasingly unlikely, o this enzyme corresponded well to polyP formation in
bearing in mind the evidence that PAO populations are some studies with Acinetobacter spp. [70,203,209], in
probably phylogenetically diverse. others the correlation was poor [210]. Also, whether it is
The enzymes that are thought to be involved in polyP the only enzyme involved in polyP degradation in EBPR
metabolism and their known properties have been de- systems is not known. Studies similar to those carried out
scribed [22,61], and some have been puri¢ed and charac- with PPK [208] are clearly required to search in activated
terized from several bacteria, including A. johnsonii [202^ sludge for gene fragments encoding for this and other
205]. However, evidence for their presence in activated enzymes thought to play a role in polyP metabolism. In
sludge and thus their role in EBPR is less convincing. Of the absence of any sequence information, the method us-
those thought to be involved in aerobic polyP synthesis, ing gene cassette PCR [211] to recover complete open
polyP glucokinase has been detected in isolates of polyP reading frames from enriched EBPR communities may
synthesizing coryneform bacteria from activated sludge well provide vital information on which genes are in-
[70], but its activity in activated sludge was not reported. volved.
Most interest has focused on polyP kinase (PPK) and its Even less is known about the regulation of PHA and
role in polyP synthesis in EBPR. It can be detected in pure glycogen in EBPR communities, or the diversity of bacte-
cultures of many bacteria [61], but although very low lev- ria synthesizing these intracellular storage polymers (see
els of activity were claimed in activated sludge [206], other above). Our current understanding of their biosynthesis
attempts [207] have been less successful. Yet using PCR comes from studies with pure cultures, but knowing now
and degenerate primers fragments of genes encoding for the sequences of some of the biosynthetic genes involved
this enzyme were recovered from biomass from an EBPR [212,213] should allow PCR-based methods to be tested
SBR system [208], and some were related to the gene in and applied to EBPR communities. Other methods, like

gene cassette PCR [211] or RNA stable isotope probing accepted mathematical representation of EBPR is that
[214^216] using 13 C-labeled substrates like acetate or glu- adopted in ASM2 and ASM2d, where the basic stoichi-
cose, might help to reveal the identity of these active pop- ometry is taken from what is known about PAO metabo-
ulations. Although, if these substrates are used mainly for lism. Although glycogen clearly plays an important role in
storage and not for growth and DNA/RNA synthesis in anaerobic substrate uptake and storage by PAO
the PAO, as is believed to be the case from the biochem- [17,34,223,224] its metabolism is not modeled in ASM2
ical models, then this approach may not be appropriate. or ASM2d, since the IAWQ group deliberately sought to
Alternatively, FISH/staining already described for PAO keep both as simple as possible. They also take no account
and GAO may also provide useful information. Some in- of possible ecological interactions between di¡erent popu-
formation from studies with pure cultures of Acinetobacter lations in EBPR systems. Of special interest are the com-
sp. is available on regulation of the synthesis and de- petitive interactions between the PAO and GAO/‘G-Bac-
gradation of PHA [212,213]. Oxygen limitation is a key teria’ discussed in Section 13. Acquiring such information
trigger in PHA formation in many bacteria, arising would make a valuable contribution to EBPR modeling

from inhibition of the TCA cycle and accumulation of since deterioration of EBPR systems due to GAO prolif-
acetyl CoA, which in turn stimulates enzymes involved eration could then be predicted [225]. On the other hand
in PHA synthesis [217]. The same regulatory process detailed information on the structure^function relation-
may occur in the anaerobic zone of EBPR systems. How ships of individual populations in EBPR plants may not
glycogen synthesis and degradation are regulated and necessarily lead to more precise prediction of process be-
which enzymes are involved in EBPR is not understood havior. The complexity of such models may mean that
at all. estimating the essential parameters needed for such appli-
cations becomes a very di⁄cult task. However, modelers
would agree that increasing our understanding of the mi-
15. Mathematical models for EBPR crobiological nature of EBPR would make a valuable
contribution to improving interpretation of model predic-
Developing mathematical models to describe the behav- tions.
ior of activated sludge processes and to assist in process
design, operation and control and education of engineers
has attracted considerable interest. However, to de¢ne 16. Future work
such models for a wide range of di¡erent system con¢g-
urations operating under various conditions requires sub- In this review we have tried to summarize the present
stantial information on the physiology of individual mi- ‘state of the art’ in progress towards understanding the
crobial populations responsible for performing key microbiology of EBPR. We can now speculate on where
functions. Therefore an essential key to modeling EBPR research should be focused. The most pressing need is the
is an understanding of the physiology of PAO. same as that emphasized previously [17], and the priorities
E¡orts to integrate PAO metabolism into activated have changed little. The critical challenge is to obtain in
sludge models began in the 1980s at the University of pure culture as many PAO as possible. These will provide
Cape Town [218], but their work was taken over by the invaluable material for understanding their physiology and
IAWQ Task Group on Mathematical Modeling for De- biochemistry [108,109]. The crucial question is how can we
sign and Operation of Biological Wastewater Treatment. successfully achieve this. It is now clear that knowing the
This international group developed a more comprehensive phylogeny of organism does not often help in predicting
EBPR model, the IAWQ Activated Sludge model No. 2 its physiology [26,149], as exempli¢ed by Rhodocyclus spp.,
(ASM2) [219], in which anaerobic uptake of carbon sour- which are phototrophic bacteria and clearly quite di¡erent
ces and their storage as PHA were modeled mechanisti- to the Rhodocyclus-related PAO, which are heterotrophs
cally. A modi¢ed version, ASM2d [220], incorporated a [29]. However, FISH/MAR might provide clues on the
capability for denitri¢cation by the PAO (see Section nutritional requirements of PAO [130] by allowing insight
13). A totally di¡erent approach was adopted by the into their physiology in situ. Further useful information
group at the Delft University of Technology [39,221, may come from screening either enriched PAO popula-
222]. Their EBPR model was a structured metabolic one, tions (obtained by £ow cytometry of DAPI stained bio-
based on the stoichiometry and bioenergetics of all the mass, Percoll centrifugation) or micromanipulated PAO
known and postulated metabolic reactions underlying clusters through systems like BIOLOG, which reveal their
EBPR. Introducing stoichiometric and kinetic information substrate utilization patterns [226]. It may not be possible
for all the individual reactions markedly reduces the num- to grow these as axenic cultures but only as mixed com-
ber of parameters to be determined, even though the nec- munities, as was achieved with a four-organism consorti-
essary kinetic and stoichiometric information is not readily um, and in which no Rhodocycus-related organisms ap-
obtainable. However, this model is much more complex peared to play a role in EBPR [227].
than the IAWQ versions. Consequently, the most widely It is also very important to keep an open mind in EBPR

studies and not to repeat the mistake, made with Acineto- Acknowledgements
bacter for example, of becoming blinkered and concentrat-
ing all the e¡ort on one population type, important R. Seviour was a Visiting Professor at the University of
though Rhodocyclus-related organisms increasingly appear Tokyo while this review was written and wishes to ac-
to be. There is now considerable evidence that EBPR com- knowledge the ¢nancial support provided by the Univer-
munity structures in di¡erent systems may vary markedly. sity of Tokyo during his stay. We also thank Dr. Helen
Equally, although the biochemical models are very helpful Stratton for supplying some of the ¢gures.
in suggesting how EBPR might work, we see no absolute
necessity for an organism to behave as these models de-
mand to be considered a PAO. It is possible for example References
that polyP synthesizing organisms like Acinetobacter may
not play a primary role in EBPR, but be crucial in scav- [1] Winkler, M. (1984) Biological control of nitrogenous pollution
in wastewater. In: Topics in Enzyme and Fermentation Biotech-
enging for P in the aerobic zone and, importantly, be

nology Vol. 8 (Wiseman, A., Ed.), pp. 31^125. Ellis Horwood, Chi-
responsible for producing an e¥uent of very low P con- chester.
centration. Furthermore, the need for anaerobic:aerobic [2] Conley, D.J. (2000) Biogeochemical nutrient cycles and nutrient man-
cycling for successful EBPR may not be essential. Indeed, agement strategies. Hydrobiologia 410, 87^96.
the authors are aware of a stable fully aerobic EBPR [3] Kortstee, G.J.J., Appeldorn, K.J., Bonting, C.F.C., van Niel, E.W.J.
and van Veen, H.W. (1994) Biology of phosphate-accumulating bac-
system producing an e¥uent with very low P, in which
teria involved in enhanced biological phosphorus removal. FEMS
incidentally Acinetobacter spp. (identi¢ed in situ by a ge- Microbiol. Rev. 15, 137^153.
nus-speci¢c probe targeting the PHA synthase gene) ap- [4] Nixon, S.W. et al. (1996) The fate of nitrogen and phosphorus at the
pear to be a dominant population (R.C. Bayly and J.W. land^sea margin of the north Atlantic ocean. Biogeochemistry 35,
May, personal communication). We still need to under- 141^180.
[5] Arvin, E. (1985) Biological removal of phosphorus from wastewaters.
stand how the activities of the PAO and their population
CRC Crit. Rev. Environ. Cont. 15, 25^64.
densities might be a¡ected by operational parameters. [6] Meganck, M.T.J. and Faup, G.M. (1988) Enhanced biological phos-
Although progress has been made in exploring a possible phorus removal from waste waters. Biotreat. Syst. 3, 11^204.
role of the GAO and ‘G-Bacteria’ in EBPR performance [7] Argaman, Y. (1991) Biological nutrient removal. In: Biological Deg-
[186,200], many other possible detrimental biological in£u- radation of Wastes (Martin, A.M., Ed.), pp. 85^101. Elsevier Applied
Science, Newfoundland.
ences including PAO-speci¢c bacteriophages [228] have re-
[8] Mackenthun, K.M. (1973) Eutrophication and biological associa-
ceived little or no attention. tions. In: Environmental Phosphorus Handbook (Gri⁄th, E.J., Bee-
Hence, there is still much to be learned about EBPR ton, A., Spencer, J.M. and Mitchell, D.T., Eds.), pp. 613^622. Wiley,
and PAO and, as we are reminded [26], eventually the New York.
knowledge acquired should be useful in helping design [9] Bryan, A. (1993) Biological Nutrient Removal. Queensland Depart-
ment of Primary Industries, Brisbane, Qld.
and run these activated sludge systems better, and in re-
[10] Codd, G.A. (1995) Cyanobacterial toxins: occurrence, properties and
sponding to EBPR failure in a rational and scienti¢c man- biological signi¢cance. Water Sci. Technol. 32, 149^156.
ner. Bioaugmentation, with appropriate application tech- [11] Lindrea, K.C. and Seviour, R.J. (2002) Activated sludge ^ the pro-
nology [229,230], is one example of how such information cess. In: Encyclopaedia of Environmental Microbiology (Bitton, G.,
might be applied [26]. Although often tried in the past, this Ed.), pp. 74^81. Wiley Scienti¢c, New York.
[12] Barnard, J.L. (1975) Biological nutrient removal without the addition
strategy usually fails with activated sludge because of lack
of chemicals. Water Res. 9, 485^490.
of understanding of which organisms are responsible for [13] Barnard, J.L. (1976) A review of biological phosphorus removal in
the particular operational feature sought e.g. EBPR. An- activated sludge. Water SA 2, 136^144.
other suggestion is to exploit the microbiological knowl- [14] Toerien, D.F., Gerber, A., Lo‹tter, L.H. and Cloete, T.E. (1990) En-
edge of community structure and function to design plants hanced biological phosphorus removal in activated sludge. Adv. Mi-
crob. Ecol. 11, 173^230.
which will stimulate increased species diversity among the
[15] Fuhs, G.W. and Chen, M. (1975) Microbiological basis of phosphate
PAO, creating greater stability and therefore less chance of removal in the activated sludge process for the treatment of waste-
plant failure [26]. water. Microb. Ecol. 2, 119^138.
These are for the future, but the application of commu- [16] Wentzel, M.C., Lo«tter, L.H., Loewenthal, R.E. and Marais, G.v.R.
nity ¢ngerprinting methods like DNA microarrays to a (1986) Metabolic behaviour of Acinetobacter spp. in enhanced bio-
logical phosphorus removal ^ a biochemical model. Water SA 12,
wide range of di¡erent activated sludge systems with their
209^224.
ability to detect many genes simultaneously [231,232] may [17] Mino, T., van Loosdrecht, M.C.M. and Heijnen, J.J. (1998) Micro-
greatly increase the rate at which the information neces- biology and biochemistry of the enhanced biological phosphate re-
sary for these applications is acquired. This in turn will moval process. Water Res. 32, 3193^3207.
involve a close cooperation between microbiologists and [18] Bond, P.L. and Rees, G.N. (1999) Microbiological aspects of phos-
phorus removal in activated sludge systems. In: Microbiology of
engineers, professions not always in empathy in the past,
Activated Sludge (Seviour, R.J. and Blackall, L.L., Eds.), pp. 227^
but without such cooperation the e¡orts of both to devel- 256. Kluwer Academic Publishers, Dordrecht.
op EBPR systems that perform better and more reliably [19] Jenkins, D. and Hermanowicz, S.W. (1991) Sources of phosphorus in
will not be realized. wastewater. In: Phosphorus and Nitrogen Removal from Municipal

Wastewater: Principles and Practice (Sedlak, R.I., Ed.), pp. 91^112. Biological model for enhanced biological phosphorus removal. Water
Lewis Publishers, Boca Raton, FL. Res. 20, 1511^1521.
[20] Jenkins, D. and Tandoi, V. (1991) The applied microbiology of en- [39] Smolders, G.J.F., van der Meij, J., van Loosdrecht, M.C.M. and
hanced biological phosphate removal ^ accomplishments and needs. Heijnen, J.J. (1994) Model of the anaerobic metabolism of the bio-
Water Res. 25, 1471^1478. logical phosphorus removal process ^ stoichiometry and pH in£u-
[21] Van Loosdrecht, M.C.M., Hooijmans, C.M., Brdjanovic, D. and ence. Biotechnol. Bioeng. 43, 461^470.
Heijnen, J.J. (1997) Biological phosphorus removal processes. Appl. [40] van Loosdrecht, M.C.M., Pot, M.A. and Heijnen, J.J. (1997) Impor-
Microbiol. Biotechnol. 48, 289^296. tance of bacterial storage polymers in bioprocesses. Water Sci. Tech-
[22] Kortstee, G.J.J., Appeldoorn, K.J., Bonting, C.F.C., van Niel, E.W.J. nol. 35, 41^47.
and van Veen, H.J. (2000) Ecological aspects of biological phospho- [41] Majone, M., Dirks, K. and Beun, J.J. (1999) Aerobic storage under
rus removal in activated sludge systems. Adv. Microb. Ecol. 16, 169^ dynamic conditions in activated sludge process. The state of the art.
200. Water Res. 39, 61^73.
[23] Amann, R., Ludwig, W. and Schleifer, K.-H. (1995) Phylogenetic [42] Wentzel, M.C., Lo«tter, L.H., Ekama, G.A., Loewenthal, T.E. and
identi¢cation and in situ detection of individual microbial cells with- Marais, G.v.R. (1991) Evaluation of biochemical models for biolog-
out cultivation. Microbiol. Rev. 59, 143^169. ical excess phosphorus removal. Water Sci. Technol. 23, 567^576.

[24] Head, I.M., Saunders, J.R. and Pickup, R.W. (1998) Microbial evo- [43] Pereira, H., Lemos, P.C., Reis, M.A.M., Crespo, J.P.S.G., Carrondo,
lution, diversity and ecology: a decade of ribosomal RNA analysis of M.J.T. and Santos, H. (1996) Model for carbon metabolism in bio-
uncultivated microorganisms. Microb. Ecol. 35, 1^21. logical phosphorus removal processes based on in vivo 13 C-NMR
[25] Theron, J. and Cloete, T.E. (2000) Molecular techniques for deter- labeling experiments. Water Res. 30, 2128^2138.
mining microbial diversity and community structure in natural envi- [44] Hesselmann, R.P.X., von Rummel, R., Resnick, S.M., Hany, R. and
ronments. Crit. Rev. Microbiol. 26, 37^57. Zehnder, A.J.B. (2000) Anaerobic metabolism of bacteria performing
[26] Loy, A., Daims, H. and Wagner, M. (2002) Activated sludge ^ mo- enhanced biological phosphate removal. Water Res. 34, 3487^3494.
lecular techniques for determining community composition. In: En- [45] Maurer, M., Gujer, W., Hany, R. and Bachmann, S. (1997) Intra-
cyclopaedia of Environmental Microbiology (Bitton, G., Ed.), pp. cellular carbon £ow in phosphorus accumulating organisms from
26^43. Wiley Scienti¢c, New York. activated sludge systems. Water Res. 31, 907^917.
[27] Srinath, E.G., Sastry, C.A. and Pillai, S.C. (1959) Rapid removal of [46] Kaewpipat, K. and Grady, C.P.L. (2002) Microbial population dy-
phosphorus from sewage by activated sludge. Experientia 15, 339^ namics in laboratory-scale activated sludge reactors. Water Sci. Tech-
340. nol. 46, 19^27.
[28] Ludzack, F.J. and Ettinger, M.B. (1962) Controlling operation to [47] Gerber, A., Mostert, E.S., Winter, C.T. and de Villiers, R.H. (1986)
minimize activated sludge e¥uent nitrogen. J. Water Pollut. Cont. The e¡ect of acetate and other short-chain carbon compounds on the
Fed. 34, 920^931. kinetics of biological nutrient removal. Water SA 12, 7^12.
[29] Hesselmann, R.P.X., Werlen, C., Hahn, D., van der Meer, J.R. [48] Tasli, R., Artan, N. and Orhon, D. (1997) The in£uence of di¡erent
and Zehnder, A.J.B. (1999) Enrichment, phylogenetic analysis and substrates on enhanced biological phosphorus removal in a sequenc-
detection of a bacterium that performs enhanced biological phos- ing batch reactor. Water Sci. Technol. 35, 75^80.
phate removal in activated sludge. Syst. Appl. Microbiol. 22, 454^ [49] Randall, A.A. and Liu, Y.-H. (2002) Polyhydroxyalkanoates form
465. potentially a key aspect of aerobic phosphorus uptake in enhanced
[30] Juretschko, S., Loy, A., Lehner, A. and Wagner, M. (2002) The biological phosphorus removal. Water Res. 36, 3473^3478.
microbial community composition of a nitrifying^denitrifying acti- [50] Cech, J.S. and Hartman, P. (1990) Glucose induced breakdown of
vated sludge from an industrial sewage treatment plant analyzed by enhanced biological phosphate removal. Environ. Technol. 11, 651^
the full-cycle rRNA approach. Syst. Appl. Microbiol. 25, 84^99. 656.
[31] Wentzel, M.C., Dold, P.L., Ekama, G.A. and Marais, G.v.R. (1985) [51] Cech, J.S. and Hartman, P. (1993) Competition between polyphos-
Kinetics of biological phosphorus release. Water Sci. Technol. 17, phate and polysaccharide accumulating bacteria in enhanced biolog-
57^71. ical phosphate removal systems. Water Res. 27, 1219^1225.
[32] Lemos, P.C., Viana, C., Salguero, E.N., Ramos, A.M., Crespo, [52] Carucci, A., Majone, M., Ramadori, R. and Rossetti, S. (1994) Dy-
J.P.S.G. and Reis, M.A.M. (1998) E¡ect of carbon source on the namics of phosphorus and organic substrates in anaerobic and aero-
formation of polyhydroxyalkanoates (PHA) by a phosphate-accumu- bic phases of a sequencing batch reactor. Water Sci. Technol. 30,
lating mixed culture. Enzym. Microb. Technol. 22, 662^671. 237^246.
[33] Liu, W.-T., Mino, T., Nakamura, K. and Matsuo, T. (1996) Glyco- [53] Nakamura, K. and Dazai, M. (1986) Growth characteristics of batch-
gen accumulating population and its anaerobic substrate uptake in cultured activated sludge and its phosphate elimination capacity.
anaerobicâerobic activated sludge without biological phosphorus J. Ferment. Technol. 64, 433^439.
removal. Water Res. 30, 75^82. [54] Sudiana, I.M., Mino, T., Satoh, H., Nakamura, K. and Matsuo, T.
[34] Mino, T., Tsuzuki, Y. and Matsuo, T. (1987) E¡ect of phosphorus (1999) Metabolism of enhanced biological phosphorus removal and
accumulation on acetate metabolism in the biological phosphorus non-enhanced biological phosphorus removal sludge with acetate and
removal process. In: Biological Phosphate Removal from Waste- glucose as carbon source. Water Sci. Technol. 39, 29^35.
waters (Ramadori, R., Ed.), pp. 27^38. Pergamon Press, Oxford. [55] Jeon, C.O. and Park, J.M. (2000) Enhanced biological phosphorous
[35] Mino, T., Kawakami, T. and Matsuo, T. (1984) Location of phos- removal in a sequencing batch reactor fed with glucose as a sole
phorus in activated sludge and function of intracellular polyphos- carbon source. Water Res. 34, 2160^2170.
phates in biological phosphorus removal process. Water Sci. Technol. [56] Wang, N., Peng, J. and Hill, G. (2002) Biochemical model of glucose
17, 93^106. induced enhanced biological phosphorus removal under anaerobic
[36] Buchan, L. (1981) The location and nature of accumulated phospho- condition. Water Res. 36, 49^58.
rus in seven sludges from activated sludge plants, which exhibited [57] Satoh, H., Mino, T. and Matsuo, T. (1990) Biochemical mechanism
enhanced phosphorus removal. Water SA 7, 1^7. of anaerobic glutamate uptake in the enhanced biological phosphorus
[37] Lee, N., la Cour Janssen, J., Aspegren, H., Henze, M., Nielsen, P.H. removal process. In: Proc. IAWPRC Biennial Conf. Kyoto, pp. 531^
and Wagner, M. (2002) Population dynamics in wastewater treatment 534.
plants with enhanced biological phosphorus removal operated with [58] Satoh, H., Mino, T. and Matsuo, T. (1998) Anaerobic uptake of
and without nitrogen removal. Water Sci. Technol. 46, 163^170. glutamate and aspartate by enhanced biological phosphorus removal
[38] Comeau, H., Hall, K.J., Hancock, R.E.W. and Oldham, W.K. (1986) activated sludge. Water Sci. Technol. 37, 579^582.

[59] Carucci, A., Lindrea, K.C., Majone, M. and Ramadori, R. (1999) [78] Ro«ske, I., Scho«nborn, C. and Bauer, H.-D. (1995) In£uence of the
Di¡erent mechanisms for the anaerobic storage of organic substrates addition of di¡erent metals to an activated sludge system on the
and their e¡ect on enhanced biological phosphate removal (EBPR). enhanced biological phosphorus removal. Int. Rev. Ges. Hydrobiol.
Water Sci. Technol. 39, 21^28. 80, 1^17.
[60] Carucci, A., Ku«nhi, M., Brun, R., Carucci, G., Koch, G., Majone, [79] Lindrea, K.C., Pigdon, S.P., Boyd, B. and Lockwood, G.A. (1994)
M. and Siegrist, H. (1999) Microbial competition for the organic Biomass characterization in a NDEBPR plant during start up and
substrates and its impact on EBPR systems under conditions of subsequent periods of good and poor phosphorus removal. Water
changing carbon feed. Water Sci. Technol. 39, 75^85. Sci. Technol. 29, 91^100.
[61] Kornberg, A., Rao, N.N. and Ault-Riche, D. (1999) Inorganic poly- [80] Bonting, C.F.C., Kortstee, G.J.J., Boekestein, A. and Zehnder,
phosphate: a molecule of many functions. Annu. Rev. Biochem. 68, A.J.B. (1993) The elemental composition dynamics of large polyphos-
89^125. phate granules in Acinetobacter strain 210A. Arch. Microbiol. 159,
[62] Liu, W.-T., Nielsen, A.T., Wu, J.-H., Tsai, C.-S., Matsuo, Y. and 428^434.
Molin, S. (2001) In situ identi¢cation of polyphosphate- and polyhy- [81] Van Veen, H.W., Abee, T., Kortstee, G.J.J., Pereira, H., Konings,
droxyalkanoate-accumulating traits for microbial populations in a W.N. and Zehnder, A.J.B. (1994) Generation of a proton motive
biological phosphorus removal process. Environ. Microbiol. 3, 110^ force by the excretion of metal-phosphate in the polyphosphate-ac-

122. cumulating Acinetobacter johnsonii strain 210A. J. Biol. Chem. 269,
[63] Zilles, J.L., Peccia, J., Kim, M.-W., Hung, C.-H. and Noguera, D.R. 29509^29514.
(2002) Involvement of Rhodocyclus-related organisms in phosphorus [82] Bond, P.L., Keller, J. and Blackall, L.L. (1999) Anaerobic phosphate
removal in full-scale wastewater treatment plants. Appl. Environ. release from activated sludge with enhanced biological phosphorus
Microbiol. 68, 2763^2769. removal. A possible mechanism of intracellular pH control. Biotech-
[64] Levin, G.V. and Shapiro, J. (1965) Metabolic uptake of phosphorus nol. Bioeng. 63, 507^515.
by wastewater organisms. J. Water Pollut. Cont. Fed. 37, 800^ [83] Deinema, M.H., van Loosdrecht, M. and Scholten, A. (1985) Some
821. physiological characteristics of Acinetobacter spp. accumulating large
[65] Hill, W.E., Bene¢eld, L.D. and Jing, S.R. (1989) 31 P-NMR spectros- amounts of phosphate. Water Sci. Technol. 17, 119^125.
copy characterization of polyphosphates in activated sludge exhibit- [84] Streichan, M., Golecki, J.R. and Scho«n, G. (1990) Polyphosphate-
ing enhanced phosphorus removal. Water Res. 23, 1177^1181. accumulating bacteria from sewage plants with di¡erent processes for
[66] Streichan, M. and Scho«n, G. (1991) Periplasmic and intracytoplasmic biological phosphorus removal. FEMS Microbiol. Ecol. 73, 113^124.
polyphosphate and easily washable phosphate in pure cultures of [85] Skerman, V.B.D. (1968) A new type of micromanipulator and micro-
sewage bacteria. Water Res. 25, 9^13. forge. J. Gen. Microbiol. 54, 287^297.
[67] Mino, T., Kawakami, T. and Matsuo, T. (1985) Location of phos- [86] Duncan, A., Vasiliadis, G.E., Bayly, R.C., May, J.W. and Raper,
phorus in activated sludge and function of intracellular polyphos- W.G.C. (1988) Genospecies of Acinetobacter isolated from activated
phates in biological phosphorus removal processes. Water Sci. Tech- sludge showing enhanced removal of phosphate during pilot-scale
nol. 17, 93^106. treatment of sewage. Biotechnol. Lett. 10, 831^836.
[68] Mino, T., Kawakami, T. and Matsuo, T. (1985) Behaviour of intra- [87] Beacham, A.M., Seviour, R.J., Lindrea, K.C. and Livingstone, I.
cellular polyphosphate in the biological phosphate removal process. (1990) Genospecies diversity of Acinetobacter isolates obtained from
Water Sci. Technol. 17, 11^21. a biological nutrient removal pilot plant of modi¢ed UCT con¢gu-
[69] Appeldoorn, K.J., Boom, A.J., Kortstee, G.J.J. and Zehnder, A.J.B. ration. Water Res. 24, 23^29.
(1992) Contribution of precipitated phosphates and acid soluble poly- [88] Kim, M.H., Hao, O.J. and Wang, N.S. (1997) Acinetobacter isolates
phosphate to enhanced biological phosphate removal. Water Res. 26, from di¡erent activated sludge processes: characteristics and neural
453^460. network identi¢cation. FEMS Microbiol. Ecol. 23, 217^227.
[70] Bark, K., Sponner, A., Ka«mpfer, P., Grund, S. and Dott, W. (1992) [89] Bonting, C.F.C., Willemsen, B.M.F., Akkermans-van Vliet, W., Bou-
Di¡erences in polyphosphate accumulation and phosphate adsorp- vet, P.J.M., Kortstee, G.J.J. and Zehnder, A.J.B. (1992) Additional
tion by Acinetobacter isolates from wastewater producing polyphos- characteristics of the polyphosphate-accumulating Acinetobacter
phate: AMP phosphotransferase. Water Res. 26, 1379^1388. strain 210A and its identi¢cation as Acinetobacter johnsonii. FEMS
[71] Jing, S.R., Bene¢eld, L.D. and Hill, W.E. (1992) Observations relat- Microbiol. Ecol. 102, 57^64.
ing to enhanced phosphorus removal in biological systems. Water [90] Ka«mpfer, P., Bark, K., Busse, H.J., Auling, G. and Dott, W. (1992)
Res. 26, 213^223. Numerical and chemotaxonomy of polyphosphate accumulating Aci-
[72] Clark, J.E., Beegan, H. and Wood, H.G. (1986) Isolation of intact netobacter strains with high polyphosphate:AMP phosphotransferase
chain of polyphosphate from Propionibacterium shermanii grown on (PPAT) activity. Syst. Appl. Microbiol. 15, 409^419.
glucose or lactate. J. Bacteriol. 168, 1212^1225. [91] Soddell, J.A., Beacham, A.M. and Seviour, R.J. (1993) Phenotypic
[73] Blonda, M., Brunetti, A., Morrone, S., Ramadori, R. and May, J.W. identi¢cation of non-clinical isolates of Acinetobacter species. J. Appl.
(1994) Determination of orthophosphates in activated sludges from Bacteriol. 74, 210^214.
wastewater treatment plants showing enhanced biological phosphate [92] Knight, G.C., McDonnell, S.A., Seviour, R.J. and Soddell, J.A.
removal. Water Res. 28, 155^159. (1993) Identi¢cation of Acinetobacter isolates using the Biolog iden-
[74] Lindrea, K.C., Lockwood, G. and Majone, M. (1998) The distribu- ti¢cation scheme. Lett. Appl. Microbiol. 16, 261^264.
tion and movement of polyphosphate and associated cations in [93] Carr, E., Ward, A., Gu«rtler, V. and Seviour, R.J. (2001) Pyrolysis
sludges from NDEBPR plants in di¡erent con¢gurations at pilot mass spectrometry (PyMS) and 16S-23S rDNA spacer region ¢nger-
plant scale. Water Sci. Technol. 37, 555^562. printing suggests the presence of novel acinetobacters in activated
[75] Imai, H., Endoh, K. and Kozuka, T. (1988) Magnesium requirement sludge. Syst. Appl. Microbiol. 24, 430^442.
for biological removal of phosphate by activated sludge. J. Ferment. [94] Carr, E.L., Ka«mpfer, P., Patel, B.K.C., Gurtler, V. and Seviour, R.J.
Technol. 66, 657^666. (2003) Seven novel species of Acinetobacter isolated from activated
[76] Rickard, L.F. and McClintock, S.A. (1992) Potassium and magnesi- sludge. Int. J. Syst. Evol. Microbiol., in press.
um requirements for enhanced biological phosphorus removal from [95] Van Veen, H.W., Abee, T., Kortstee, G.J., Konings, W.N. and
wastewater. Water Sci. Technol. 26, 2203^2206. Zehnder, A.J. (1993) Characterization of two phosphate transport
[77] Scho«nborn, C., Bauer, H.-D. and Ro«ske, I. (2001) Stability of en- systems in Acinetobacter strain 210A. J. Bacteriol. 175, 200^206.
hanced biological phosphorus removal and composition of polyphos- [96] Groenestijn, J.W., Zuidema, M., van der Worp, J.J.M., Deinema,
phate granules. Water Res. 35, 3190^3196. M.H. and Zehnder, A.J.B. (1989) In£uence of environmental param-

eters on polyphosphate accumulation in Acinetobacter sp. Antonie sludge for enhanced phosphate removal. Appl. Environ. Microbiol.
van Leeuwenhoek 55, 67^82. 64, 992^998.
[97] Tandoi, V., Majone, M., May, J. and Ramadori, R. (1998) The [115] Auling, G., Pilz, F., Busse, H.-J., Karrasch, M., Streichan, M. and
behaviour of polyphosphate accumulating Acinetobacter isolates in Schon, G. (1991) Analysis of the polyphosphate-accumulating mi-
an anaerobic :aerobic chemostat. Water Res. 32, 2903^2912. cro£ora in phosphorus-eliminating, anaerobicâerobic activated
[98] Brodisch, K.E.U. and Joyner, S.J. (1983) The role of microorgan- sludge systems by using diaminopropane as a biomarker for rapid
isms other than Acinetobacter in biological phosphate removal in estimation of Acinetobacter spp. Appl. Environ. Microbiol. 57,
activated sludge process. Water Sci. Technol. 15, 117^125. 3685^3692.
[99] Lo‹tter, L.H. and Murphy, M. (1985) The identi¢cation of hetero- [116] Liu, W.-T., Linning, K.D., Nakamura, K., Mino, T., Matsuo, T.
trophic bacteria in activated sludge with particular reference to pol- and Forney, L. (2000) Microbial community changes in biological
yphosphate accumulation. Water SA 11, 179^184. phosphate-removal systems on altered sludge phosphate content.
[100] Maszenan, A.M., Seviour, R.J., Patel, B.K.C., Schumann, P., Burg- Microbiology (UK) 146, 1099^1107.
hardt, Y., Tokiwa, Y. and Stratton, H.M. (2000) Three isolates of [117] Sudiana, I.M., Mino, T., Satoh, H. and Matsuo, T. (1998) Morphol-
novel polyphosphate accumulating gram positive cocci, obtained ogy, in situ characterization with rRNA targeted probes and respi-
from activated sludge belong to a new genus Tetrasphaera gen. ratory quinone pro¢les of enhanced biological phosphorus removal

nov. and description of two new species Tetrasphaera japonica sp. sludge. Water Sci. Technol. 38, 69^76.
nov. and Tetrasphaera australiensis sp. nov. Int J. Syst. Evol. Micro- [118] Lin, C.-K., Katayama, Y., Hosomi, M., Murakami, A. and Okada,
biol. 50, 593^603. M. (2000) The relationship between isoprenoid quinone and phos-
[101] Hanada, S., Liu, W.-T., Shintani, T., Kamagata, Y. and Nakamura, phorus removal activity. Water Res. 34, 3607^3613.
K. (2002) Tetrasphaera elongata sp. nov., a polyphosphate-accumu- [119] Okada, M., Lin, C.K., Katayama, Y. and Murakami, A. (1992)
lating bacterium from activated sludge. Int. J. Syst. Evol. Microbiol. Stability of phosphorus removal and population of bio-P-bacteria
52, 883^887. under short-term disturbances in sequencing batch reactor activated
[102] Seviour, E.M., Williams, C.J., Seviour, R.J., Soddell, J.A. and Lin- sludge process. Water Sci. Technol. 26, 483^491.
drea, K.C. (1990) A survey of ¢lamentous bacterial populations [120] Fujita, M., Chen, H. and Furamai, H. (1999) An investigation on
from foaming activated sludge plants in eastern states of Australia. microbial population dynamics in enhanced biological phosphorus
Water Res. 24, 493^498. removal SBR using quinone pro¢le and PCR-DGGE techniques.
[103] Erhart, R., Bradford, D., Seviour, R.J., Amann, R. and Blackall, In: Proceedings of 7th IAWQ Asian Paci¢c Conference, Taipai.
L.L. (1997) Development and use of £uorescent in situ hybridization [121] Seviour, R.J. and Maszenan, A.S. (2002) Activated sludge ^ the
probes for the detection and identi¢cation of ‘Microthrix parvicella’ microbial community. In: Encyclopaedia of Environmental Micro-
in activated sludge. Syst. Appl. Microbiol. 20, 310^318. biology (Bitton, G., Ed.), pp. 68^73. Wiley Scienti¢c, New York.
[104] Blackall, L.L., Seviour, E.M., Bradford, D., Rossetti, S., Tandoi, V. [122] Wintzingerode, F.V., Goebel, U.B. and Stackebrandt, E. (1997) De-
and Seviour, R.J. (2000) ‘Candidatus Nostocoida limicola’, a ¢la- termination of microbial diversity in environmental samples: pitfalls
mentous bacterium from activated sludge. Int J. Syst. Evol. Micro- of PCR-based rRNA analysis. FEMS Microbiol. Rev. 21, 213^229.
biol. 50, 703^709. [123] Bond, P.L., Hugenholtz, P., Keller, J. and Blackall, L.L. (1995)
[105] Nakamura, K., Hiraishi, A., Yoshimi, Y., Kawaharasaki, M., Mat- Bacterial community structures of phosphate-removing and non-
suda, K. and Kamagata, Y. (1995) Microlunatus phosphovorus gen. phosphate-removing activated sludges from sequencing batch reac-
nov. sp. nov., a new gram-positive polyphosphate accumulating tors. Appl. Environ. Microbiol. 61, 1910^1916.
bacterium isolated from activated sludge. Int. J. Syst. Bacteriol. [124] Snaidr, J., Amann, R., Huber, I., Ludwig, W. and Schleifer, K.-H.
45, 17^22. (1997) Phylogenetic analysis and in situ identi¢cation of bacteria in
[106] Santos, M.M., Lemos, P.C., Reis, M.A.M. and Santos, H. (1999) activated sludge. Appl. Environ. Microbiol. 63, 2884^2896.
Glucose metabolism and kinetics of phosphorus removal by the [125] Muyzer, G. and Smalla, K. (1998) Application of denaturing gra-
fermentative bacterium Microlunatus phosphovorus. Appl. Environ. dient gel electrophoresis (DGGE) and temperature gradient gel elec-
Microbiol. 65, 3920^3928. trophoresis (TGGE) in microbial ecology. Antonie van Leeuwen-
[107] Stante, L., Cellamare, C.M., Malaspina, F., Bortone, G. and Tiche, hoek 73, 127^141.
A. (1997) Biological phosphorus removal by pure culture of Lam- [126] Schwieger, F. and Tebbe, C.C. (1998) A new approach to utilize
propedia spp. Water Res. 31, 1317^1324. PCR-single-strand-conformation polymorphism for 16S rRNA
[108] Palleroni, N.J. (1997) Prokaryotic diversity and the importance of gene-based microbial community analysis. Appl. Environ. Micro-
culturing. Antonie van Leeuwenhoek 72, 3^19. biol. 64, 4870^4876.
[109] Amann, R. (2000) Who is out there? Microbial aspects of biodiver- [127] Osborn, A.M., Moore, E.R.B. and Timmis, K.N. (2000) An evalua-
sity. Syst. Appl. Microbiol. 23, 1^8. tion of terminal-restriction fragment length polymorphism (T-
[110] Schramm, A. and Amann, R. (1999) Nucleic-acid based techniques RFLP) analysis for the study of microbial community structure
for analyzing the diversity, structure and dynamics of microbial and dynamics. Environ. Microbiol. 2, 39^50.
communities in activated sludge. In: Environmental Processes 1 [128] Lipski, A., Friedrich, U. and Altendorf, K. (2001) Application of
(Winter, J., Ed.), pp. 85^108. Wiley-VCH, Weinheim. rRNA-targeted oligonucleotide probes in biotechnology. Appl. Mi-
[111] Amann, R. and Ludwig, W. (2000) Ribosomal RNA-targeted nu- crobiol. Biotechnol. 56, 40^57.
cleic acid probes for studies in microbial ecology. FEMS Microbiol. [129] Daims, H., Bruhl, A., Amann, R., Schleifer, K.-H. and Wagner, M.
Rev. 24, 555^565. (1999) The domain speci¢c probe EUB338 is insu⁄cient for the
[112] Cloete, T.E. and Steyn, P.L. (1988) Combined membrane ¢lter-im- detection of all Bacteria : development and evaluation of a more
muno£uorescent technique for the in situ identi¢cation and enumer- comprehensive probe set. Syst. Appl. Microbiol. 22, 434^444.
ation of Acinetobacter in activated sludge. Water Res. 22, 961^ [130] Lee, N., Nielsen, P.H., Andreasen, K.H., Juretschko, S., Nielsen,
969. J.L., Schleifer, K.-H. and Wagner, M. (1999) Combination of £uo-
[113] Hiraishi, A., Masamune, K. and Kitamura, H. (1989) Character- rescent in situ hybridization and microautoradiography ^ a new tool
ization of the bacterial population structure in an anaerobic :aerobic for structure^function analyses in microbial ecology. Appl. Environ.
activated sludge system on the basis of respiratory quinone pro¢les. Microbiol. 65, 1289^1297.
Appl. Environ. Microbiol. 54, 197^210. [131] Gray, N.D. and Head, I.M. (2001) Linking genetic identity and
[114] Hiraishi, A., Ueda, Y. and Ishihara, J. (1998) Quinone pro¢ling of function in communities of uncultured bacteria. Environ. Microbiol.
bacterial communities in natural and synthetic sewage activated 3, 481^492.

[132] Braker, G., Fesefeldt, A. and Witzel, K.-P. (1998) Development of [148] Ka«mpfer, P., Erhart, R., Beimfohr, C., Bohringer, J., Wagner, M.
PCR primer systems for ampli¢cation of nitrite reductase genes and Amann, R. (1996) Characterization of bacterial communities
(nirK and nirS) to detect denitrifying bacteria in environmental sam- from activated sludge : culture-dependent numerical identi¢cation
ples. Appl. Environ. Microbiol. 64, 3769^3775. versus in situ identi¢cation using group- and genus-speci¢c rRNA-
[133] Hallin, S. and Lindgren, P.-E. (1999) PCR detection of genes encod- targeted oligonucleotide probes. Microb. Ecol. 32, 101^121.
ing nitrite reductase in denitrifying bacteria. Appl. Environ. Micro- [149] Bjo‹rnsson, L., Hugenholtz, P., Tyson, G.W. and Blackall, L.L.
biol. 65, 1652^1657. (2002) Filamentous Chloro£exi (green non-sulfur bacteria) are abun-
[134] Bothe, H., Jost, G., Schloter, M., Ward, B. and Witzel, K.-P. (2000) dant in wastewater treatment processes with biological nutrient re-
Molecular analysis of ammonia oxidation and denitri¢cation in nat- moval. Microbiology (UK) 148, 2309^2318.
ural environments. FEMS Microbiol. Ecol. 24, 673^690. [150] Beer, M., Seviour, E.M., Kong, Y., Cunningham, M., Blackall, L.L.
[135] Braker, G., Zhou, J., Wu, L., Devol, A.H. and Tiedje, J.M. (2000) and Seviour, R.J. (2002) Phylogeny of the ¢lamentous bacterium
Nitrite reductase genes (nirK and nirS) as functional markers to Eikelboom type 1851, and design and application of a 16S rRNA
investigate diversity of denitrifying bacteria in paci¢c northwest ma- targeted oligonucleotide probe for its £uorescence in situ identi¢ca-
rine sediment communities. Appl. Environ. Microbiol. 66, 2096^ tion in activated sludge. FEMS Microbiol. Lett. 207, 179^183.
2104. [151] Schade, M., Beimfohr, C. and Lemmer, H. (2002) Phylogenetic and

[136] Onuki, M., Satoh, H. and Mino, T. (2002) Analysis of microbial physiological characterization of a ‘Nostocoida limicola’-like organ-
community that performs enhanced biological phosphorus removal ism isolated from activated sludge. Water Sci. Technol. 46, 91^97.
in activated sludge fed with acetate. Water Sci. Technol. 46, 145^ [152] Hugenholtz, P., Goebel, B.M. and Pace, N.R. (1998) Impact of
154. culture-independent studies on the emerging phylogenetic view of
[137] Gray, N.D., Miskin, I.P., Kornilova, O., Curtis, T.P. and Head, bacterial diversity. J. Bacteriol. 180, 4765^4774.
I.M. (2002) Occurrence and activity of Archaea in aerated activated [153] Garrity, G.M. and Holt, J.G. (2001) Chloro£exi phy. nov. In: Ber-
sludge wastewater treatment plants. Environ. Microbiol. 4, 158^168. gey’s Manual of Systematic Bacteriology, 2nd Edn., Vol. 1 (Boone,
[138] Dabert, P., Sialve, B., Delgenes, J.P., Moletta, R. and Godon, J.J. D.R. and Castenholz, R.W., Eds.), pp. 427^446. Springer, New
(2001) Characterisation of the microbial 16S rDNA diversity of an York.
aerobic phosphorus-removal ecosystem and monitoring of its tran- [154] Sera¢m, L.S., Lemos, P.C., Levantesi, C., Tandoi, V., Santos, H.
sition to nitrate respiration. Appl. Microbiol. Biotechnol. 55, 500^ and Reis, M.A.M. (2002) Methods for detection and visualization of
509. intracellular polymers stored by polyphosphate-accumulating micro-
[139] Crocetti, G.R., Hugenholtz, P., Bond, P., Schuler, A., Keller, J., organisms. J. Microbiol. Methods 51, 1^18.
Jenkins, D. and Blackall, L.L. (2000) Identi¢cation of polyphos- [155] Kawaharasaki, M., Manbome, A., Kanagawa, T. and Nakamura,
phate-accumulating organisms and design of 16S rRNA-directed K. (2002) Flow cytometric sorting and RFLP analysis of phosphate
probes for their detection and quanti¢cation. Appl. Environ. Micro- accumulating bacteria in an enhanced biological phosphorus remov-
biol. 66, 1175^1182. al system. Water Sci. Technol. 46, 139^144.
[140] Weller, R., Glo«ckner, F.O. and Amann, R. (2000) 16S rRNA oli- [156] Levantesi, C., Sera¢m, L.S., Crocettii, G.R., Lemos, P.C., Rossetti,
gonucleotide probes for the in situ detection of members of the S., Blackall, L.L., Reis, M.A.M. and Tandoi, V. (2002) Analysis of
phylum Cytophaga-Flavobacterium-Bacteroides. Syst. Appl. Micro- the microbial community structure and function of a laboratory
biol. 23, 107^114. scale enhanced biological phosphorus removal reactor. Environ. Mi-
[141] Wagner, M., Erhart, R., Manz, W., Amann, R., Lemmer, H., Wedi, crobiol. 4, 559^569.
D. and Schleifer, K.-H. (1994) Development of an rRNA-targeted [157] Schuler, A., Onuki, M., Satoh, H. and Mino, T. (2002) Density
oligonucleotide probe speci¢c for the genus Acinetobacter and its separation and molecular methods to characterize enhanced biolog-
application for in situ monitoring in activated sludge. Appl. Envi- ical phosphorus removal system populations. Water Sci. Technol.
ron. Microbiol. 60, 792^800. 46, 195^198.
[142] Kawaharasaki, M., Tanaka, H., Kanagawa, T. and Nakamura, K. [158] Melasniemi, H. and Hernesmaa, A. (2000) Yeast spores seem to be
(1999) In situ identi¢cation of polyphosphate-accumulating bacteria involved in biological phosphate removal : a microscopic in situ case
in activated sludge by dual staining with rRNA-targeted oligonu- study. Microbiology (UK) 146, 701^707.
cleotide probes and 4P,6-diamidino-2phenylindole (DAPI) at a poly- [159] Kerrn-Jespersen, J.P. and Henze, M. (1993) Biological phosphorus
phosphate-probing concentration. Water Res. 33, 257^265. uptake under anoxic and aerobic conditions. Water Res. 27, 617^
[143] Mudaly, D.D., Atkinson, B.W. and Bux, F. (2001) 16S rRNA prob- 624.
ing for the determination of the family level community structure [160] Kuba, T., Smolders, G., van Loosdrecht, M.C.M. and Heijnen, J.J.
implicated in enhanced biological nutrient removal. Water Sci. (1993) Biological phosphorus removal from wastewater by anaero-
Technol. 43, 91^98. bicânoxic sequencing batch reactor. Water Sci. Technol. 27, 241^
[144] Kong, Y., Beer, M., Rees, G.N. and Seviour, R.J. (2002) Functional 252.
analysis of microbial communities in aerobic :anaerobic sequencing [161] Kuba, T., van Loosdrecht, M.C.M. and Heijnen, J.J. (1996) Phos-
batch reactors fed with di¡erent phosphorus/carbon ratios. Micro- phorus and nitrogen removal with minimum COD requirement by
biology (UK) 148, 2299^2307. integration of denitri¢cation in a two-sludge system. Water Res. 30,
[145] Christensson, M., Blackall, L.L. and Welander, T. (1998) Metabolic 1702^1710.
transformations and characterization of the sludge community in an [162] Jorgensen, K.S. and Pauli, A.S.L. (1995) Polyphosphate accumula-
enhanced biological removal system. Appl. Microbiol. Biotechnol. tion among denitrifying bacteria in activated sludge. Anaerobe 1,
49, 226^234. 161^168.
[146] Kawaharasaki, M., Kanagawa, T., Tanaka, H. and Nakamura, K. [163] Barker, P.S. and Dold, P.L. (1996) Denitri¢cation behaviour in bio-
(1998) Development and application of 16S rRNA-targeted probe logical excess phosphorus removal activated sludge systems. Water
for detection of the phosphate accumulating bacterium Microlunatus Res. 30, 769^780.
phosphovorus in an enhanced biological phosphorus removal pro- [164] Van Niel, E.W.J., Appeldoorn, K.J., Zehnder, A.J.B. and Kortstee,
cess. Water Sci. Technol. 37, 481^484. G.J.J. (1998) Inhibition of anaerobic phosphate release by nitric
[147] Bond, P.L., Erhart, R., Wagner, M., Keller, J. and Blackall, L.L. oxide in activated sludge. Appl. Environ. Microbiol. 64, 2925^
(1999) Identi¢cation of some of the major groups of bacteria in 2930.
e⁄cient and none⁄cient biological phosphorus removal activated [165] Mino, T., Liu, W.T., Kurisu, F. and Matsuo, T. (1995) Modeling
sludge systems. Appl. Environ. Microbiol. 65, 4077^4084. glycogen storage and denitri¢cation capability of microorganisms in

enhanced biological phosphate removal processes. Water Sci. Tech- anaerobicâerobic activated sludge with a minimized polyphosphate
nol. 31, 25^34. content. J. Ferment. Bioeng. 77, 535^540.
[166] Meinhold, J., Arnold, E. and Isaacs, S. (1999) E¡ect of nitrite on [185] Satoh, H., Mino, T. and Matsuo, T. (1994) Deterioration of en-
anoxic phosphate uptake in biological phosphorus removal acti- hanced biological phosphorus removal by the domination of micro-
vated sludge. Water Res. 33, 1871^1883. organisms without polyphosphate accumulation. Water Sci. Tech-
[167] Nielsen, J.L. and Nielsen, P.H. (2002) Quanti¢cation of functional nol. 30, 203^211.
groups in activated sludge by microautoradiography. Water Sci. [186] Blackall, L.L., Crocetti, G.R., Saunders, A.M. and Bond, P.L.
Technol. 46, 389^395. (2002) A review and update of the microbiology of enhanced bio-
[168] Bortone, G., Saltarelli, R., Alons, V., Sorm, R., Wanner, J. and logical phosphorus removal in wastewater treatment plants. Antonie
Tilche, A. (1996) Biological anoxic phosphorus removal ^ the DE- van Leeuwenhoek 81, 681^691.
PHANOX process. Water Sci. Technol. 34, 119^128. [187] Filipe, C.D.M., Daigger, G.T. and Grady, C.P.L. (2001) E¡ects of
[169] Hu, Z.-R., Wentzel, M.C. and Ekama, G.A. (2002) The signi¢cance pH on the rates of aerobic metabolism of phosphate-accumulating
of denitrifying polyphosphate accumulating organisms in biological and glycogen-accumulating organisms. Water Environ. Res. 73,
nutrient removal activated sludge systems. Water Sci. Technol. 46, 213^222.
129^138. [188] Filipe, C.D.M., Daigger, G.T. and Grady, C.P.L. (2001) pH as a

[170] Merzouki, M., Delgenus, J.P., Bernet, N., Moletta, R. and Benlem- key factor in the competition between glycogen-accumulating and
lih, M. (1999) Polyphosphate accumulating and denitrifying bacteria phosphate-accumulating organisms. Water Environ. Res. 73, 223^
isolated from anaerobicânoxic sequencing batch reactors. Curr. 232.
Microbiol. 38, 9^17. [189] Schuler, A.J. and Jenkins, D. (2002) E¡ects of pH on enhanced
[171] Merzouki, M., Bernet, N., Delgenes, J.-P., Moletta, R. and Benlem- biological phosphorus removal metabolisms. Water Sci. Technol.
lih, M. (1999) Kinetic behaviour of some polyphosphate-accumulat- 46, 171^178.
ing bacteria isolates in the presence of nitrate and oxygen. Curr. [190] Whang, L.-M. and Park, J.K. (2002) Competition between poly-
Microbiol. 38, 300^308. phosphate- and glycogen-accumulating organisms in biological
[172] Barak, Y. and van Rijn, J. (2000) Atypical polyphosphate accumu- phosphorus removal systems ^ e¡ect of temperature. Water Sci.
lation by the denitrifying bacterium Paracoccus denitri¢cans. Appl. Technol. 46, 191^194.
Environ. Microbiol. 66, 1209^1212. [191] Wang, J.C., Park, J.K. and Whang, L.M. (2001) Comparison of
[173] Barak, Y. and van Rijn, J. (2000) Relationship between nitrite re- fatty acid composition and kinetics of phosphorus-accumulating or-
duction and active phosphate uptake in the phosphate-accumulating ganisms and glycogen-accumulating organisms. Water Environ. Res.
denitri¢er Pseudomonas sp. strain JR 12. Appl. Environ. Microbiol. 73, 704^710.
66, 5236^5240. [192] Nielsen, A.T., Liu, W.-T., Filipe, C., Grady, L., Molin, M. and
[174] Ahn, J.A., Daidou, T., Tsuneda, S. and Hirata, A. (2002) Character- Stahl, D.A. (1999) Identi¢cation of a novel group of bacteria in
ization of denitrifying phosphate-accumulating organisms cultivated sludge from a deteriorated biological phosphorus removal reactor.
under di¡erent electron acceptor conditions using polymerase chain Appl. Environ. Microbiol. 65, 1251^1258.
reaction-denaturing gradient gel electrophoresis assay. Water Res. [193] Maszenan, A.M., Seviour, R.J., Patel, B.K.C., Rees, G.N. and
36, 403^412. McDougall, B.M. (1997) Amaricoccus gen. nov., a gram-negative
[175] Brdjanovic, D., Slamet, A., van Loosdrecht, M.C.M., Hooijmans, coccus occurring in regular packages and tetrads, isolated from ac-
C.M., Alaerts, G.J. and Heijnen, J.J. (1997) Impact of excessive tivated sludge biomass, and descriptions of Amaricoccus veronensis
aeration on biological phosphorus removal from wastewater. Water sp. nov., Amaricoccus tamworthensis sp. nov., Amaricoccus macauen-
Res. 32, 200^208. sis sp. nov. and Amaricoccus kaplicensis sp. nov. Int J. Syst. Bacte-
[176] Matsuo, Y. (1994) E¡ect of the anaerobic solids retention time on riol. 47, 727^734.
enhanced biological phosphorus removal. Water Sci. Technol. 30, [194] Kong, Y., Beer, M., Seviour, R.J., Lindrea, K.C. and Rees, G.N.
193^202. (2001) Structure and functional analysis of the microbial community
[177] Wang, J.C. and Park, J.K. (2001) E¡ect of anaerobicâerobic con- in an aerobic:anaerobic sequencing batch reactor (SBR) with no
tact time on the change of internal storage energy in two di¡erent phosphorus removal. Syst. Appl. Microbiol. 24, 597^609.
phosphorus-accumulating organisms. Water Environ. Res. 73, 436^ [195] Falvo, A., Levantesi, C., Rossetti, S., Seviour, R.J. and Tandoi, V.
443. (2001) Synthesis of intracellular storage polymers by Amaricoccus
[178] Jeon, C.O., Lee, D.S., Lee, M.W. and Park, J.M. (2001) Enhanced kaplicensis, a tetrad forming bacterium present in activated sludge.
biological phosphorus removal in an anaerobicâerobic sequencing J. Appl. Microbiol. 91, 299^305.
batch reactor. Water Environ. Res. 73, 301^306. [196] Jeon, C.O., Lee, D.S. and Park, J.M. (2000) Morphological charac-
[179] Jeon, C.O., Lee, D.S. and Park, J.M. (2001) Enhanced biological teristics of microbial sludge performing enhanced biological phos-
phosphorus removal in an anaerobicâerobic sequencing batch re- phorous removal in a sequencing batch reactor fed with glucose as a
actor: characteristics of carbon metabolism. Water Environ. Res. sole carbon source. Water Sci. Technol. 41, 79^84.
73, 295^300. [197] Fang, H.H.P., Zhang, T. and Liu, Y. (2002) Characterization of an
[180] Tsai, C.-S. and Liu, W.-T. (2002) Phylogenetic and physiological acetate-degrading sludge without intracellular accumulation of poly-
diversity of tetrad-forming organisms in deteriorated biological phosphate and glycogen. Water Res. 36, 3211^3218.
phosphorus removal systems. Water Sci. Technol. 46, 179^184. [198] Maszenan, A.M., Seviour, R.J., Patel, B.H.C. and Wanner, J. (2000)
[181] Seviour, R.J., Maszenan, A.M., Soddell, J.A., Tandoi, V., Patel, A £uorescently labeled r-RNA targeted probe for the in situ detec-
B.K.C., Kong, Y. and Schumann, P. (2000) Microbiology of the tion of G-bacteria of the genus Amaricoccus in activated sludge.
‘G-Bacteria’ in activated sludge. Environ. Microbiol. 2, 581^593. J. Appl. Microbiol. 88, 826^835.
[182] Takii, S. (1977) Bacterial characteristics of activated sludges treating [199] Kong, Y., Ong, S.L., Ng, W.J. and Liu, W.T. (2002) Diversity and
carbohydrate wastes. Water Res. 11, 85^89. distribution of a novel proteobacterial group found in anaero-
[183] Liu, W.-T., Nakamura, K., Matsuo, T. and Mino, T. (1997) Internal bic:aerobic activated sludge processes. Environ. Microbiol. 4, 753^
energy-based competition between polyphosphate- and glycogen-ac- 757.
cumulating bacteria in biological phosphorus removal reactors ^ [200] Crocetti, G.R., Ban¢eld, J.F., Keller, J., Bond, P.L. and Blackall,
e¡ect of P/C feeding ratio. Water Res. 31, 1430^1438. L.L. (2002) Glycogen-accumulating organisms in laboratory-scale
[184] Liu, W.-T., Mino, T., Nakamura, K. and Matsuo, T. (1994) Role of and full-scale wastewater treatment processes. Microbiology (UK)
glycogen in acetate uptake and polyhydroxyalkanoate synthesis in 148, 3353^3364.

[201] Shintani, T., Liu, W., Hanada, S., Kamagata, Y., Miyaoka, S., [217] Dawes, E.A. (1992) Storage polymers in prokaryotes. In: Prokary-
Suzuki, Y. and Nakamura, K. (2000) Micropruina glycogenica gen. otic Structure and Function : A New Perspective (Mohan, S., Dow-
nov., sp. nov., a new gram-positive glycogen-accumulating bacte- and, C. and Cole, J.A., Eds.), pp. 81^122. Cambridge University
rium isolated from activated sludge. Int J. Syst. Evol. Microbiol. Press, Cambridge.
50, 201^207. [218] Wentzel, M.C., Ekama, G.A., Lowenthal, R.E., Dold, P.L. and
[202] Van Groenestijn, J.W., Bentvelsen, M.M., Deinema, M.H. and Marais, G.v.R. (1989) Enhanced polyphosphate organism cultures
Zehnder, A.J. (1989) Polyphosphate-degrading enzymes in Acineto- in activated sludge ^ part III : kinetic model. Water SA 15, 89^102.
bacter spp. and activated sludge. Appl. Environ. Microbiol. 55, 219^ [219] Henze, M., Guyer, W., Mino, T., Wentzel, M.C. and Marais,
223. G.v.R. (1995) The activated sludge model no. 2. Scienti¢c and Tech-
[203] Bonting, C.F.C., Kortstee, G.J.J. and Zehnder, A.J.B. (1991) Prop- nical Report No. 3, International Association on Water, London.
erties of polyphosphate : AMP phosphotransferase of Acinetobacter [220] Henze, M., Guyer, W., Mino, T., Wentzel, M.C., Marais, G.v.R.M.
strain 210A. J. Bacteriol. 173, 6484^6488. and van Loostrecht, M.C.M. (1999) Activated sludge model no. 2d,
[204] Resnick, S. and Zehnder, A.J. (2000) In vitro ATP regeneration ASM 2D. Water Sci. Technol. 39, 165^182.
from polyphosphate and AMP by polyphosphate: AMP phospho- [221] Smolders, G.J.F., van der Meij, J., van Loostrecht, M.C.M. and
transferase and adenylate kinase from Acinetobacter johnsonii 210A. Heijnen, J.J. (1994) Stoichiometric model of the aerobic metabolism

Appl. Environ. Microbiol. 66, 2045^2051. of the biological phosphorus removal process. Biotechnol. Bioeng.
[205] Trelstad, P.L., Purdhani, P., GeiMdo«rfer, W., Hillen, W. and Keas- 44, 837^848.
ling, J.D. (1999) Polyphosphate kinase of Acinetobacter sp. strain [222] Smolders, G.J.F., van der Meij, J., van Loostrecht, M.C.M. and
ADP1 : puri¢cation and characterization of the enzyme and its role Heijnen, J.J. (1995) A structured metabolic model for the anaerobic
during changes in extracellular phosphate levels. Appl. Environ. and aerobic stoichiometry and kinetics of the biological phosphorus
Microbiol. 65, 3780^3786. removal process. Biotechnol. Bioeng. 47, 277^287.
[206] Vasiliadis, G.A., Duncan, A., Bayly, R.C. and May, J.W. (1990) [223] Satoh, H., Mino, T. and Matsuo, T. (1992) Uptake of organic sub-
Polyphosphate production by strains of Acinetobacter. FEMS Mi- strates and accumulation of polyhydroxyalkanoates linked with gly-
crobiol. Lett. 70, 37^40. colysis of intracellular carbohydrates under anaerobic conditions in
[207] T’Seyen, J., Malnou, D., Block, J.C. and Faup, G. (1985) Polyphos- the biological excess phosphate removal process. Water Sci. Tech-
phate kinase activity during phosphate uptake by bacteria. Water nol. 26, 933^942.
Sci. Technol. 17, 43^56. [224] Liu, W.T., Mino, T., Nakamura, K. and Matsuo, T. (1994) Role of
[208] McMahon, K.D., Dojka, M.A., Pace, N.R., Jenkins, D. and Keas- glycogen in acetate uptake and polyhydroxyalkanoate synthesis in
ling, J.D. (2002) Polyphosphate kinase from activated sludge per- anaerobicâerobic activated sludge with a minimized polyphosphate
forming enhanced biological phosphorus removal. Appl. Environ. content. J. Ferment. Bioeng. 77, 535^540.
Microbiol. 68, 4971^4978. [225] Satoh, H., Mino, T. and Matsuo, T. (1994) Deterioration of en-
[209] GeiMdo«rfer, W., Ratajczak, A. and Hillen, W. (1998) Transcription hanced biological phosphorus removal by the domination of micro-
of ppk from Acinetobacter sp. strain ADP1, encoding a putative organisms without polyphosphate accumulation. Water Sci. Tech-
polyphosphate kinase, is induced by phosphate starvation. Appl. nol. 30, 203^211.
Environ. Microbiol. 64, 896^901. [226] Preston-Mafham, J., Boddy, L. and Randerson, P.F. (2002) Analy-
[210] Bayly, R.C., Duncan, A., May, J.W., Schembri, M., Semertijs, A., sis of microbial community functional diversity using sole-carbon-
Vasiliadis, G. and Raper, W.G.C. (1991) Microbiological and genet- source-utilization pro¢les ^ a critique. FEMS Microbiol. Ecol. 26,
ic aspects of the synthesis of polyphosphate by species of Acineto- 1^14.
bacter. Water Sci. Technol. 23, 747^754. [227] Hollender, J., Dreyer, U., Kornberger, L., Ka«mpfer, P. and Dott,
[211] Stokes, H.W., Holmes, A.J., Nield, B.S., Holley, M.P., Nevalainen, W. (2002) Selective enrichment and characterization of a phospho-
K.M.H., Mabbutt, B.C. and Gillings, M.R. (2001) Gene cassette rus-removing bacterial consortium from activated sludge. Appl. Mi-
PCR: sequence-independent recovery of entire genes from environ- crobiol. Biotechnol. 58, 106^111.
mental DNA. Appl. Environ. Microbiol. 67, 5240^5246. [228] Khan, M.A., Satoh, H., Katayama, H., Kurisu, F. and Mino, T.
[212] Schembri, M.A., Bayly, R.C. and Davies, J.K. (1994) Cloning and (2002) Bacteriophages isolated from activated sludge processes and
analysis of the polyhydroxyalkanoic acid synthase gene from an their polyvalency. Water Res. 36, in press.
Acinetobacter. FEMS Microbiol. Lett. 118, 145^152. [229] Bouchez, T., Patureau, D., Dabert, P., Juretschko, S., Dore¤, J.,
[213] Schembri, M.A., Bayly, R.C. and Davies, J.K. (1995) Phosphate Delgene's, P., Moletta, R. and Wagner, M. (2000) Ecological study
concentration regulates transcription of the Acinetobacter polyhy- of a bioaugmentation failure. Environ. Microbiol. 2, 179^190.
droxyalkanoic acid biosynthetic genes. J. Bacteriol. 177, 4501^4507. [230] Watanabe, K., Miyashita, M. and Harayama, S. (2000) Starvation
[214] Boschker, H.T.S. and Middelburg, J.J. (2002) Stable isotopes and improves survival of bacteria introduced into activated sludge. Appl.
biomarkers in microbial ecology. FEMS Microbiol. Ecol. 40, 85^95. Environ. Microbiol. 66, 3905^3910.
[215] Mane¢eld, M., Whiteley, A.S., Gri⁄ths, R.I. and Bailey, M.J. [231] Cho, J.-C. and Tiedje, J.M. (2002) Quantitative detection of micro-
(2002) RNA stable isotope probing, a novel means of linking microbial genes by using DNA microarrays. Appl. Environ. Microbiol.
bial community function to phylogeny. Appl. Environ. Microbiol. 68, 1425^1430.
68, 5367^5373. [232] Loy, A., Lehner, L., Lee, N., Adamczyk, J., Meier, H., Ernst, J.,
[216] Radajewski, S., Webster, G., Reay, D.S., Morris, S.A., Ineson, P., Schleifer, K.-H. and Wagner, M. (2002) Oligonucleotide microarray
Nedwell, D.B., Prosser, J.I. and Murrell, J.C. (2002) Identi¢cation for 16S rRNA gene-based detection of all recognized lineages of
of active methylotroph populations in an acidic forest soil by stable- sulfate-reducing prokaryotes in the environment. Appl. Environ.
isotope probing. Microbiology (UK) 148, 2331^2342. Microbiol. 68, 5064^5081.

Bio-P Microbiology

Загружено:

Сведения о документе

Исходное описание:

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Bio-P Microbiology

Загружено:

Авторское право:

Доступные форматы

FEMS Microbiology Reviews 27 (2003) 99^127

The microbiology of biological phosphorus removal in

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

First published online 20 March 2003

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

9.1. Diversity of EBPR community populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

1. Introduction ceous material produce treated e¥uents with high residual

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

good EBPR performance is often associated with appear-

4. The biochemistry of EBPR systems

4.1. EBPR metabolism with acetate as sole carbon source

Several empirical models have been described [16,17,38]

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

R.J. Seviour et al. / FEMS Microbiology Reviews 27 (2003) 99^127

Actinobacteria (n.d.) (with

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

and hence non-polyP accumulating populations using £ow

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

10.2. Using FISH/staining and FISH/MAR to indicate

Combining FISH with DAPI and Sudan black staining

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

published outlining how this may occur (e.g. [55]). Un-

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

First reported in activated sludge in [182], their possible

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

FEMSRE 768 3-4-03 Cyaan Magenta Geel Zwart

and a clone with a sequence almost the same as that

from clone sbr-gs28 from an SBR fed glucose [55,196]

for almost all of the Actinobacteria in this community.

No Amaricoccus spp. could be detected by FISH [198],

the PAO for substrates anaerobically in EBPR deteriora-

Downloaded from https://academic.oup.com/femsre/article-abstract/27/1/99/500670 by guest on 21 June 2019

3ve tion [50,51].

Considerable physiological variation was noted in the

ria’ belonging to some of these divisions [199]. Thus, not

all of the K-Proteobacteria took up acetate or glucose,