Environmental Microbiology (2010) 12(1), 192206

doi:10.1111/j.1462-2920.2009.02060.x

Microbial community structure in autotrophic nitrifying

granules characterized by experimental and simulation
analyses
emi_2060

192..206

Shinya Matsumoto,1 Mayu Katoku,1 Goro Saeki,1

Akihiko Terada,2 Yoshiteru Aoi,3 Satoshi Tsuneda,1*
Cristian Picioreanu4 and Mark C. M. van Loosdrecht4
1
Department of Life Science and Medical Bioscience,
Waseda University, Wakamatsu-cho 22, Shinjuku-ku,
Tokyo 162-8480, Japan.
2
Institute of Environment and Resources, Technical
University of Denmark, DK-2800 Lyngby, Denmark.
3
Waseda Institute for Advanced Study, Waseda
University, Wakamatsu-cho 2-2, Shinjuku-ku, Tokyo
162-8480, Japan.
4
Department of Biotechnology, Delft University of
Technology, Julianalaan 67, 2628 BC Delft, the
Netherlands.
Summary
This study evaluates the community structure in
nitrifying granules (average diameter of 1600 mm)
produced in an aerobic reactor fed with ammonia
as the sole energy source by a multivalent approach
combining molecular techniques, microelectrode
measurements and mathematical modelling. Fluorescence in situ hybridization revealed that ammoniaoxidizing bacteria dominated within the first 200 mm
below the granule surface, nitrite-oxidizing bacteria a
deeper layer between 200 and 300 mm, while heterotrophic bacteria were present in the core of the
nitrifying granule. Presence of these groups also
became evident from a 16S rRNA clone library. Microprofiles of NH4+, NO2, NO3 and O2 concentrations
measured with microelectrodes showed good
agreement with the spatial organization of nitrifying
bacteria. One- and two-dimensional numerical biofilm
models were constructed to explain the observed
granule development as a result of the multiple
bacteriasubstrate interactions. The interaction
between nitrifying and heterotrophic bacteria was
evaluated by assuming three types of heterotrophic
bacterial growth on soluble microbial products from

Received 3 April, 2009; accepted 11 August, 2009. *For correspondence. E-mail stsuneda@waseda.jp; Tel. (+81) (0)3 5369 7325; Fax
(+81) (0)3 3341 2684.

2009 Society for Applied Microbiology and Blackwell Publishing Ltd

nitrifying bacteria. The models described well the

bacterial distribution obtained by fluorescence in situ
hybridization analysis, as well as the measured
oxygen, nitrite, nitrate and ammonium concentration
profiles. Results of this study are important because
they show that a combination of simulation and
experimental techniques can better explain the interaction between nitrifying bacteria and heterotrophic
bacteria in the granules than individual approaches
alone.
Introduction
The low growth rate and low fraction of autotrophic
bacteria in activated sludge processes make nitrification
the rate-limiting step in most wastewater treatment processes. Thus, a simpler and effective method of immobilizing nitrifying bacteria within wastewater treatment
processes is desired. Granulation without using any carriers has been proposed for immobilizing nitrifying bacteria in inorganic wastewater treatment processes (De Beer
et al., 1997; Tay et al., 2002; Tsuneda et al., 2003). Our
group successfully produced nitrifying granules using an
aerobic upflow fluidized bed (AUFB) reactor (Tsuneda
et al., 2003). Furthermore, the time required for producing
nitrifying granules could be greatly reduced by use of
pre-aggregating seed sludge with hematite (Tsuneda
et al., 2004). However, detailed information on granule
formation and nitrogen conversion in primarily autotrophic
systems is still lacking.
To date, there are reports indicating ecophysiological
interaction between nitrifying bacteria and heterotrophic
bacteria in autotrophic nitrifying suspended cultures (Rittmann et al., 1994) and biofilms (Kindaichi et al., 2004;
Okabe et al., 2005) grown without an external organic
carbon source. In such systems, nitrifying bacteria may
consume inorganic carbon not only to form cell mass but
also to excrete organic carbon (soluble microbial products, SMP), which will support heterotrophic bacterial
growth (Rittmann et al., 2002). Although very much
needed, a detailed analysis on microbial ecology of nitrifying granules including nitrifying and heterotrophic bacteria has never been reported. Better understanding on
the formation of nitrifying granules structure (especially

Microbial community structure in autotrophic nitrifying granules 193

the microbial community distribution) and microscale
distribution of activities of each bacterial species is
necessary for improving process performance and
further application of the nitrifying granules to wastewater
treatment.
In this study, we used experimental and simulation
analyses in combination to study the microbial community
of nitrifying granules. Since community structure in the
granules is determined by a complex interplay of various
factors including the concentration of chemical species,
presence of other bacteria and their physiology, mathematical modelling provides a logical framework for the
exploration of processes within granules. A large variety of
mathematical models have been developed and widely
applied to assess the multi-species biofilm dynamics on
support materials (see a review in Wanner et al., 2006).
Traditional models consider the dominant one dimensional (1-d) gradients of solutes and biomass concentrations (e.g. Wanner and Gujer, 1985; Okabe et al., 1996).
More complex two- or three-dimensional (2-d and 3-d)
models describe the dynamics of multiple microbial populations by various approaches: grid-based biomass (e.g.
cellular automata based by Noguera et al., 1999; Picioreanu et al., 1999; Bell et al., 2005), gridless individualbased models (IbM, by Kreft et al., 2001; Picioreanu et al.,
2004), continuum (Alpkvist and Klapper, 2007) or a hybrid
individual/continuum (Alpkvist et al., 2006). Extracellular
polymeric substances (EPS) were also included in all
approaches (in cellular automata: Laspidou and Rittmann,
2004; in IbM: Kreft and Wimpenny, 2001; Xavier et al.,
2005a,b, Xavier and Foster, 2007; in hybrids: Matsumoto
et al., 2007). Some of these biofilm models have been
adapted to represent granular biofilms (e.g. Xavier et al.,
2007 for aerobic granular sludge and Picioreanu et al.,
2005; Batstone et al., 2006 for anaerobic granular sludge)
or autotrophic/heterotrophic growth of immobilized
biomass in gel beads (Vogelsang et al., 2002). We used in
this study an IbM to describe the microbial ecology of the
granules (following directly the approach reported in Kreft
et al., 2001; Picioreanu et al., 2004; Xavier et al., 2005a;
2007).
This study is new and relevant because it combines
three different research approaches, usually applied
alone in the investigation of microbial biofilms and granules. First, community structure and spatial organization
in nitrifying granules were analysed by molecular (clone
libraries of 16S rRNA genes) and microscopic imaging
techniques (lectin-binding analysis and fluorescence in
situ hybridization, FISH). Second, concentration profiles
of essential chemical species (NH4+, NO2, NO3 and O2)
along the granule radius were measured by microelectrodes. Finally, 1-d and 2-d numerical biofilm models
focusing on eco-physiological interaction between nitrifying and heterotrophic bacteria in the nitrifying granules

were developed in order to evaluate and interpret the

experimental measurements.

Results
Phylogenetic analysis
A 16S rRNA gene clone library of bacteria was constructed
from the nitrifying granules with a diameter of 1600 mm
(Fig. 1). Totally, 93 clones were analysed, approximately
600 bp per clone was sequenced and the clones were
grouped into 39 operational taxonomic units (OTUs) on the
basis of more than 97% sequence similarity within an OTU.
Clone sequences analysed were distributed over seven
major lineages of the domain Bacteria: members of the
Cytophaga-Flavobacterium-Bacteroides division (CFB),
genus of Nitrosomonas of the b-subclass ammoniaoxidizing bacteria (AOB), member of the phylum
Verrucomicrobia, g-Proteobacteria, Chloroflexi, dProteobacteria, genus Nitrobacter of the a-subclass nitrite
oxidizing bacteria (NOB) and a-Proteobacteria except for
Nitrobacter. Nitrospira-associated sequences were not
detected in the nitrifying granule. Almost all of the clones
belonging to putatively heterotrophic bacterial groups had
a similarity of less than 96% to sequences determined in
other environmental diversity surveys.

Microbial community structure in nitrifying granules

The in situ spatial organization of heterotrophic bacteria,
nitrifying bacteria and EPS was visualized by FISH with
group-specific probes and lectin staining. The samples
used for the experiment were the nitrifying granules cultured with a diameter of 1600 mm. Both groups of nitrifying
bacteria, i.e. AOB and NOB, were detected with the
Nso190 and NIT3 probes respectively (Fig. 2A and B).
The AOB created clusters, whereas NOB were present
mostly in the form of single scattered cells near the AOB
[Fig. 2C; the 3-d image was constructed by the software
Daime (Daims et al., 2006)]. Although CFB division in the
granule was not stained with the extensively used probe
CF319a/b, the CFB719 probe successfully detected bacteria related to CFB division. CFB719-stained bacteria
were thin filamentous shaped (Fig. 2D). GNSB-941stained bacteria were long, rod-shaped (Fig. 2E). Bacteria
related to phylum Verrucomicrobia were not clearly
detected with probe EUB338III due to low signal intensity
and low abundance. Figure 2F shows the spatial distribution over a section cut through the middle of the granule of
bacterial cells and EPS stained by propidium iodide (PI)
and fluorescein isothiocyanate (FITC)-labelled lectin
respectively. The main habitat of bacteria stained with PI
was at the edge of the granule, whereas the inner part of
the granule was dominantly occupied by EPS. Moreover,

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

194 S. Matsumoto et al.

Fig. 1. Phylogenetic trees based on 16S rRNA gene sequences for Proteobacteria (A) and Verrucomicrobia, Chloroflexi and the CFB group
(B). The trees were constructed using the neighbor-joining method. The number on the nodes indicates the number of times the species
(shown on the right) grouped together in 1000 bootstrap samples. Bootstrap values below 500 are not shown. The root of the tree was
determined using the 16S rRNA gene of Aquifex pyrophilus as an outgroup. Scale bar indicates the 10% estimated difference in nucleotide
sequence position.

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

Microbial community structure in autotrophic nitrifying granules 195

Fig. 1. cont.

it appears that the bacteria and EPS formed neat concentric layers along the granule radius.
The microbial nitrifying community composition in the
granules with a diameter of 1600 mm was determined by
quantitative FISH with various sets of probes (Table 1
and Fig. 3). Members of the AOB (detected with probe
Nso190) and NOB (the genus Nitrobacter belonging to the
a-Proteobacteria) accounted for 68.0% of the total bacte-

ria detected with EUB338 probe mix. The ratio of all nitrifiers (AOB plus NOB) to all heterotrophs detected was
approximately 3:1.
The spatial distributions of populations in the nitrifying
granule were determined by results from quantitative 2-d
image analysis of confocal image series (Fig. 4). Most
bacteria detected with EUB338 probe mix at the outer
edge of the granule were nitrifying bacteria; AOB

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

196 S. Matsumoto et al.

Fig. 2. Confocal laser scanning microscope
images showing the spatial organization of
nitrifying bacteria (AOB and NOB),
heterotrophic bacteria and EPS within the
nitrifying granule cultured to be diameter of
about 1600 mm.
A. FISH with FITC-labelled probe EUB338
and Cy3-labelled Nso190 (specific for AOB).
The yellow cell clusters are the
Nso190-stained AOB dually labelled with
EUB338 (green) and Nso190 (red) probes.
B. FISH with FITC-labelled EUB338 and
Cy3-labelled NIT3 (specific for Nitrobacter).
The yellow cells are the NIT3-stained NOB
dually labelled with EUB338 (green) and NIT3
(red) probes.
C. Three-dimensional image of FISH with
FITC-labelled Nso190 (green) and
Cy3-labelled NIT3 (red).
D. FISH with FITC-labelled EUB338 and
Cy3-labelled CFB719 (specific for the CFB
group). The yellow cells are the
CFB719-stained bacteria dually labelled with
EUB338 (green) and CFB719 (red) probes.
E. FISH with FITC-labelled EUB338 and
Cy3-labelled GNSB-941 and CFX1223
(specific for GNSB). The yellow cells are the
GNSB-941 and CFX1223-stained bacteria
dually labelled with EUB338 (green),
GNSB-941 and CFX1223 (red) probes.
F. Nucleic acid stained with PI (red) and EPS
stained with FITC-labelled lectin (green). Bars
indicate (A)(E) 10 mm and (F) 200 mm
respectively. (A)(C) show the outer edge of
the granule. (D) and (E) show the inner
granule. The surface of the granule was
indicated by white broken lines in (A) and (B).

dominated within first 200 mm below the granule surface,

whereas NOB were most abundant in the layer between
200 and 300 mm below the granule surface. Bacteria
related to CFB division and Chloroflexi were detected
mainly in the zone where nitrifying bacteria exist and in
the inner part of the granules respectively. Proteobacteria were present throughout the granules. It should be
noted that absolute abundance of bacteria (shown by
open circle in Fig. 4) at the inner part of granules is
much lower than at the surface of granules. Thus, even
though Chloroflexi occupy inner part of the granules,
absolute amount of Chloroflexi is much lower than that
of AOB.

Concentration profiles in nitrifying granules

Local concentrations of NH4+-N, NO2-N, NO3-N and
dissolved oxygen (DO) in the nitrifying granules with a
diameter of 1600 mm were measured under steady-state
conditions (Fig. 5). The profiles of NH4+-N, NO2-N and
NO3-N indicated that NH4+ was oxidized to NO2 mainly
down to 250 mm from the granule surface. Little nitrite
accumulation detected shows that NO2 was immediately oxidized to NO3. The active ammonium- and
nitrite-oxidizing zones were in good agreement with the
spatial distribution of AOB and NOB detected by FISH
(Fig. 4). The DO concentration at the granule surface

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

Microbial community structure in autotrophic nitrifying granules 197

Table 1. Probe sequences and formamide concentrations in the hybridization buffer required for specific in situ hybridization.
Probe

Specificity

Sequence of probe (5 to 3)

FA (%)a

Reference

EUB338
EUB338 II
EUB338 III
ALF1b

Most bacteria
Planctomycetales
Verrucomicrobiales
a-Proteobacteria,
including Nitrobacter
b-Proteobacteria
g-Proteobacteria
Phylum Chloroflexi
CF cluster
Groups of the
Bacteroidetes phylum
Ammonia-oxidizing
g-Proteobacteria
Nitrobacter spp.
Competitor for NIT3
Nitrospira moscoviensis

GCTGCCTCCCGTAGGAGT
GCAGCCACCCGTAGGTGT
GCTGCCACCCGTAGGTGT
CGTTCGYTCTGAGCCAG

_b
_b
_b
20

Amann et al. (1990)

Daims et al. (1999)
Daims et al. (1999)
Manz et al. (1992)

GCCTTCCCACTTCGTTT
GCCTTCCCACATCGTTT
AAACCACACGCTCCGCT
TGGTCCGTRTCTCAGTAC
AGCTGCCTTCGCAATCGG

35
35
35
35
30

Manz et al. (1992)

Manz et al. (1992)
Gich et al. (2001)
Manz et al. (1996)
Weller et al. (2000)

CGATCCCCTGCTTTTCTCC

35d

Mobarry et al. (1996)

CCTGTGCTCCATGCTCCG
CCTGTGCTCCAGGCTCCG
AGCACGCTGGTATTGCTA

40

Wagner et al. (1996)

Wagner et al. (1996)
Juretschko et al. (1998)

BET42a
GAM42ac
GNSB-941
CF319a/b
CFB719
Nso190
NIT3e
Comp NIT3
Ntspa1026
a.
b.
c.
d.
e.

20

Formamide concentration in the hybridization buffer.

Used at any formamide concentration.
Unlabeled probe BET42a was used as a competitor to enhance specificity.
Although originally 55% formamide should be used, we referred to Pynaert et al., 2003.
Unlabelled probe comp NIT3 was used as a competitor to enhance specificity.

was 7 mg l-1 and oxygen penetrated about 300 mm into

the granule.
Numerical simulation analyses
One-dimensional (with gradients along the granule
radius only) and 2-d numerical models were developed
to describe the measured concentration profiles and
microbial distribution within the granules (see Experimental procedures). The 1-d models combine the advantages
of simplicity and faster computations, while the 2-d
models give a more detailed granule structure and
solutes spatial distribution. Both types of models used
identical parameters (see Tables S15) and were compared in terms of their predictions of: (i) solutes evolution
in time in the reactor bulk liquid and (ii) solutes and
microbial distributions in the granule. Figure 6 shows 2-d
simulation results of the microbial distribution in the
nitrifying granule developing in time. The comparison
of dynamics of solute concentrations in the bulk liquid

provided by 1-d simulation and three 2-d simulations is

shown in Fig. 7. Figure 8 shows the comparison of
steady-state solute concentrations of solutes in the
biofilm (day 100) from 1-d and 2-d simulations, and
experimental microelectrode data of NH4+-N, NO2-N,
NO3-N and DO. Figure 9 shows the comparison of the
steady-state biomass concentrations in the biofilm (day
100) from 1-d and 2-d simulations. The animations of
simulated granule formation in time are given in Supporting information and can also be downloaded from: http://
www.biofilms.bt.tudelft.nl/NitrifyingGranules/index.html.
In the early stages of granule formation, due to sufficient supply of ammonia and oxygen from the bulk liquid,
the granule is mainly constituted of nitrifiers (Fig. 6A). As
the granule size increases, an anoxic zone is created at
the inner part of the granule. That condition is prefarable
for heterotrophic bacteria to denitrify consuming the
excreted organic material (SMP) as an electron donor.
Therefore, in the later stage of granule formation, heterotrophic bacteria exist at the inner part of the granule,
Fig. 3. Microbial community composition for
the nitrifying granule as determined by
quantitative FISH. The a-Proteobacteria is the
bacterial group that hybridized with probe
ALF1b, excluding the genus Nitrobacter that
hybridized with probe NIT3.

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

198 S. Matsumoto et al.

Fig. 4. Spatial distributions of bacteria along
the radius of a nitrifying granule as detected
by FISH. Abundance ratio of each bacterium
was quantified in 50-mm-thick shells starting at
the granule surface (the left side of the
graph). Abundance data presented are
averages of six replicate measurements. The
total bacterial volumetric occupancy in the
granule is derived from fluorescence of
EUB338 mix-tagged cells, as shown with
symbols () along the R-axis. The
a-Proteobacteria is the bacterial group that
hybridized with probe ALF1b, excluding the
genus Nitrobacter that hybridized with probe
NIT3.

while nitrifying bacteria dominate at the outer edge. In this

later stage, the inner part of the granule becomes dominated by EPS produced by the heterotrophic bacteria
(Fig. 6C and D).
Because the 2-d model contains stochastic elements
(i.e. the initial spatial distribution of microbial types and
the direction of cell division and EPS production), results
obtained with identical sets of parameters are not identical, but only very similar. This is why three 2-d simulations
were perfomed in the same conditions. The concentrations of N-containing species in the bulk liquid in time
calculated in three 2-d simulations shown in Fig. 7 are
comparable with 1-d simulation results (which are fully
deterministic). For granules smaller than the maximum
size imposed (until day 20), the concentrations computed
with 1-d and 2-d models are identical. However, although
the calculated steady state values of ammonium, nitrite
and nitrate concentrations in the bulk liquid are similar
with 1-d and 2-d models (from approximately day 70 on),
the 2-d results show more oscillations (noise) around the
average value. This is because unlike the microbial distribution in concentric shells inherent in 1-d models, in 2-d

there are radially developed colonies or cell clusters of

one microbial type (see Fig. 6C and D). After the granule
has reached the maximum imposed size, in 2-d whole
colonies can be quickly detached from the granule
surface, while in 1-d this detachment process is completely smooth. For example, if a large patch of NOB is
eliminated from the granule in a short interval of time, the
nitrite concentration will suddenly increase while the
nitrate will decrease. Because of this fluctuations, a true
steady state is never reached in 2-d simulations, neither
for solutes nor for microbial compositions. We can speak
about steady values only by comparing values averaged
over an interval of time of about 510 days.
The steady-state solute concentrations along the
granule radius calculated with the 1-d and 2-d models
describe well the experimentally obtained microelectrode
data (Fig. 8). Moreover, the concentration profiles
obtained by averaging on concentric shells the 2-d distributions of solutes match the 1-d calculated profiles. These
results show that for a description of the main solute
concentrations and fluxes in the biofilm (ammonium,
nitrite, nitrate and oxygen), a 1-d model is sufficient.
However, for low concentration solutes (the organic substrates UAP and BAP), the 2-d profiles are more dependent on the spatial distribution of different microbial
groups, which leads to different profiles from those calculated with a 1-d model (Fig. 9). Nevertheless, this is the
first direct quantitative comparison between solute concentrations in the biofilm calculated by a multidimensional
model and those measaured by microelectrodes (for
NH4+-N, NO2-N, NO3-N and DO).

Discussion
Bacterial community structure of nitrifying granules

Fig. 5. Microprofiles of NH4+, NO2, NO3 and DO in the nitrifying

granule determined by microsensor analysis. Mean value and
standard deviations out of three replicates are shown in the
experimental results section.

The phylogenetic analysis revealed that AOB present in

nitrifying granules are affiliated with Nitrosomonas europaea and Nitrosococcus mobilis. N. mobilis is known to be
a halophilic bacteria easily overgrown by N. europaea in
low-salt environments (Juretschko et al., 1998). The

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

Microbial community structure in autotrophic nitrifying granules 199

Fig. 6. Two-dimensional simulation results of spatial distributions of microbial populations and EPS in a nitrifying granule at different moments
in time. In this visualization the biomass particles with a size of about 2 mm each have colours corresponding to their microbial composition.

synthetic wastewater supplied to nitrifying granule reactor

in this study was high in salinity (42.6 mM as Na2SO4).
Thus, N. mobilis and N. europaea would coexist in nitrifying granules. Nitrite-oxidizing bacteria detected in nitrifying granules were affiliated with Nitrobacter sp. Although
Nitrospira sp. are widespread in nature and earlier known
from, for example, freshwater, brackish waters, marine
waters, soils and wastewater treatment plants, Nitrobacter outcompete Nitrospira under halophilic condition
(Cebron and Garnier, 2005). Heterotrophic bacteria such
as Chloroflexi, CFB and Proteobacteria were detected
from nitrifying granules cultured with inorganic substrate.
These results are consistent with the previous studies
demonstrating that heterotrophic bacteria are present in
biofilms cultivated with inorganic substrate (Nogueira

et al., 2002; Kindaichi et al., 2004; Okabe et al., 2005).

Nitrifying bacteria are known to release soluble products
(SMP) from substrate metabolism and biomass decay,
which can provide a supplementary organic substrate for
heterotrophic bacteria (Rittmann et al., 1994; Barker and
Stuckey, 1999). A large amount of SMP could be accumulated within nitrifying granule because sludge retention
time of nitrifying granule is extremely long (Tsuneda et al.,
2003). Since organic carbon was not added to the substrate for nitrifying granule reactor, the SMP released by
nitrifying bacteria would be the sole organic substrates for
heterotrophic bacteria. Although there are several considerations with regard to EPS production by nitrifiers, we
hypothesized that EPS production by nitrifuing bacteria is
extremely poor following the report that polysaccharides

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

200 S. Matsumoto et al.

NO2-N and NO3-N) in the reactor liquid also indicated
that denitrifcation occurred (Fig. 3). Nucleic acid and
lectin staining revealed that inner parts of a granule are
occupied with EPS, which is produced by heterotrophic
bacteria present throughout the granule. The similar EPS
distribution is reported in the aerobic granular sludge cultured by substrate containing organic carbon (McSwain
et al., 2005).
Uptake patterns for organic substrates derived from
nitrifying bacteria

Fig. 7. Comparison of dynamics of solute concentrations in the

bulk liquid calculated by 1-d (thick lines) and 2-d (thin lines) models
with identical sets of parameters. Three different 2-d simulation
results are shown here for comparison, because the stochastic
elements included in the 2-d model lead to slightly different results
each model run.

production was not confirmed from the supernatant of

N. europaea and Nitrobacter winogradskyi (Tsuneda
et al., 2001).
Two different types of AOB affiliated with N. europaea
and N. mobilis (more than 98% similarity) were detected
by phylogenetic analysis. Thus, oligonucleotide probe
Nso190 that can detect both types of AOB was applied
to FISH analysis to determine the bacterial distribution
within nitrifying granules. Considering the FISH results
with direct comparison with microelectrode results, AOB
produce nitrite while consuming ammonia and NOB
produce nitrate while consuming nitrite at the outer layer
of granules (up to 250 mm from the surface). Sufficient
excess ammonia was measured all along the granule
depth. Therefore, oxygen supply appeared to be the limiting factor for nitrification in nitrifying granules. Since only
small amount of nitrite was detected by microelectrode
measurements, nitrite is promptly oxidized to nitrate. The
first reason for the rapid oxidation of nitrite to nitrate is
that, in the nitrifying granule, NOB exist close to AOB
(Fig. 2C), which prevents accumulation of nitrite produced
by AOB. Furthermore, it was reported that the specific
nitrite oxidation rate of Nitrobacter is at most 40 times
higher than the specific ammonia oxidation rate of
N. europaea (Prosser, 1989). Regarding heterotrophic
bacteria, CFB and g-Proteobacteria were present at outer
layer of granule, Chloroflexi were inner granule and
a-Proteobacteria were throughout the granule. At the
inner parts of the granule (deeper than 250 mm from the
surface), oxygen is depleted and nitrate produced by NOB
is present. Furthermore, as the profile of nitrate is not
perfectly linear, denitrification relying on SMP would
occur. The measured loss of total N-species (NH4+-N,

There is a limited number of studies reporting about

organic substrate uptake pattern of heterotrophic bacteria
detected in nitrifying granules. OSullivan and colleagues
(2002) reported that some members of CFB group uptake
various biomacromolecules, such as cellulose, chitin,
DNA, lipids, and proteins, which may accumulate in biofilms. Sekiguchi and colleagues (1999) reported that
yeast extract-like nutrient is essential for growth of
Chloroflexi. Cottrell and Kirchman (2000) reported that the

Fig. 8. Comparison of the steady-state solute concentrations in

the biofilm calculated in 1-d (thick lines) and 2-d (thin lines)
simulations, with the experimental microelectrode data (open circles
for NH4+, NO2, NO3 and O2). The model results are at day 100. To
obtain the comparable profiles along the radius, the 2-d
concentration distributions were averaged in concentric shells with
different radii.

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

Microbial community structure in autotrophic nitrifying granules 201

Fig. 9. Comparison of the steady-state biomass concentrations in

the biofilm (day 100) calculated by 1-d (thick lines) and 2-d (thin
lines) models with identical sets of model parameters. To obtain
the comparable profiles along the radius, the 2-d concentration
distributions were averaged in concentric shells with different radii.

a- and g-Proteobacteria in marine environments utilize

amino acids. Furthermore, Okabe and colleagues
(2005) reported that the member of the Chloroflexi
utilize the decaying nitrifying bacteria cell materials (i.e.
biomass associated products; BAP), the member of
the CFB utilize substrate utilization-associated products

(UAP), and the members of the a-Proteobacteria and

g-Proteobacteria utilize low-molecular-weight organic
matter (Org) produced by hydrolysis of EPS. In the
present model, based on those studies, we integrated
three types of heterotrophic bacteria: as HetU, HetB and
HetO that utilize UAP, BAP and Org respectively. Although
those hypotheses are not reflecting all of the bacterial
uptake patterns for organic substrates derived from nitrifying bacteria, the results of this study can be plausibly
explained on the basis of hypotheses. Detailed 2-d simulated distribution of heterotrophs in the granule is shown
in Fig. 10. HetU is present at the surface of nitrifying
granule where sufficient UAP is produced by active nitrifying bacteria (Fig. 10A). HetB is mainly present in the
inner parts of granule where BAP is produced from inactivated nitrifying bacteria because of depletion of oxygen
(Fig. 10B). HetO are present throughout the granule
because EPS produced by heterotrophic bacteria spread
broadly in granule (Fig. 10C). Although the model simulations performed in this study were not intended to provide
quantitative matches to the experimental data, the simulation results indicated that distributions of HetU, HetB
and HetO are consistent with those of CFB, Chloroflexi
and Proteobacteria, respectively, obtained by FISH analysis (1-d profiles of biomass are shown in Fig. 9). Thus,
it can be concluded that the implemented organic
substrate uptake patterns of heterotrophic bacteria were
plausible.
The model simulations used in this study reveal interactions leading to the properties of biofilm systems that
are generally not tractable by experimental approaches.
So far, there have been only a very few reports of experimental verification of the 2-d or 3-d multi-species biofilm
model predictions (Xavier et al., 2005a; Matsumoto et al.,
2007). As far as we know, this paper is the first report
that evaluated biofilm phenomena concerning microbial
ecology by combination of experimental data such as

Fig. 10. Detailed insight into the spatial localization of HetU (A), HetB (B) and HetO (C) provided by a 2-d simulation, at day 100. White
dotted line shows the granule surface.

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

202 S. Matsumoto et al.

FISH and microelectrode, and 2-d multi-species biofilm
model.
In conclusion, we have combined experimental analyses using molecular techniques (i.e. 16S rRNA-cloning
analysis and FISH) and microelectrode measurements for
NH4+-N, NO2-N, NO3-N, and O2 concentration profiles
with a 2-d biofilm model of granule formation. It is now
clear that the concerted use of these techniques helps
more understanding the development of microbial populations in nitrifying granules than the use of individual
techniques alone. The experimental analyses make it
possible to understand the spatial organization of bacteria
and in situ bacterial activity. The multidimensional (here
2-d, but directly extensible to 3-d) biofilm model can be
used to more accurately describe microbial interactions
such as syntrophy and competition for space and substrates, all of which are more difficult to be explained by
previously developed 1-d models. However, 1-d models
seem to be sufficient for the prediction of solute concentrations and overall nutrient fluxes, and preferable in
this case because of their simplicity and computational
efficiency.

Experimental procedures
Wastewater composition
Ammonia-rich synthetic inorganic wastewater based on that
discharged from a thermal power station was used in this
study. Synthetic wastewater used as influent consisted of
2375 g m-3 (NH4)2SO4 (as 500 g N m-3), 6044 g m-3 Na2SO4,
0.5 g m-3 FeSO47H2O and 0.4 g m-3 KH2PO4. Besides these
components, NaHCO3 (1511 g m-3) was added to the influent
both for pH adjustment and as an inorganic carbon source for
autotrophic nitrifying bacteria. Total dissolved solid was
9930 g m-3. Details are described elsewhere (Tsuneda et al.,
2003).

Reactor configuration and operating conditions

The granules used in this study were produced in an AUFB
reactor with a diameter of 50 mm and a height of 3.2 m
(effective volume: 6.3 l) where the synthetic wastewater had
been continuously fed for approximately 600 days (Tsuneda
et al., 2003). Granules were then inoculated to a continuous
stirred tank reactor with an effective volume of 1.0 l
(Tsuneda et al., 2006) and had been cultured for another
600 days to ensure that the granules attain steady state.
The average diameter of mature granules reached about
1600 mm. Although the size of granule is relatively larger
compared with normal biofilms or flocs, Tsuneda and colleagues (2006) produced nitrifying granule with the same
procedure and reported that the diameter reached 1200 mm.
The initial suspended solid concentration was 12700 g m-3.
The volumetric aeration rate was set to 1.0 l min-1 and the
DO was around 6 g m-3. The hydraulic retention time of the
reactor was 1 day.

DNA extraction, PCR amplification, cloning and

sequencing of the 16S rRNA gene
DNA was extracted from a granule sample (approximately
0.5 g wet weight granule) with a DNA extraction kit
(ISOPLANT II, NIPPON GENE, Tokyo, Japan), as described
in the manufacturers instructions. Then, 16S rRNA gene
fragments from isolated total DNA of the granule sample were
amplified using bacterial PCR primer sets 11f (Kane et al.,
1993) and 1492r (Weisburg et al., 1991). Purified PCR products were cloned using a Qiagen PCR cloningplus kit (Qiagen,
Valencia, CA, USA). The DNA insert was amplified and used
as template DNA in a cycle sequencing reaction with a Big
Dye Terminator v3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA, USA) by using bacterial PCR primer
sets 341f and 907r (Muyzer et al., 1997). Nucleotide
sequencing was performed with an automatic sequencer
(Prism 2100 avant genetic analyser, Applied Biosystems).
Sequences were aligned by ChromasPro version 1.4.1
(Technelysium, QLD, Australia). Partial sequences (approximately 600 bp) were compared with similar sequences of the
reference organisms by a BLAST search (Altschul et al.,
1990). In the next step, all clones having a 16S rRNA
sequence similarity of more than 97% with each other were
grouped into an OTU. Phylogenetic trees were constructed
with ClustalW using the neighbor-joining method, 1000 times
bootstrap and distance model Kimura. All partial 16S rRNA
gene sequences were deposited in the DDBJ/EMBL/
GenBank nucleotide sequence databank under the accession numbers AB505163 to AB505201.

Fixation and cryosectioning of granule samples

The granule sample was fixed with freshly prepared 4%
paraformaldehyde solution for 18 h at 4C. The sample was
embedded in Tissue-Tek OCT compound (Sakura Finetek,
Tokyo, Japan) overnight to infiltrate the OCT compound into
the granular biofilm, as described earlier (Okabe et al., 1999).
After rapid freezing at -20C, 20 mm thick slices through
the middle of the granule were prepared with a cryostat
(CM1850, Leica, Heidelberg, Germany) and placed on a
gelatin-coated slide (Matsunami, Osaka, Japan). After air
drying overnight, the slices were dehydrated by successive
passage through 50%, 80% and 98% ethanol washes (for
3 min each), air-dried and stored at room temperature (Aoi
et al., 2000).

In situ hybridization
The sequences of all oligonucleotide probes used in this
study are summarized in Table 1. The detailed information
about all these probes is found in probeBase (Loy et al.,
2003). All in situ hybridizations were performed according to
the protocol (Amann et al., 1990) in hybridization buffer at
46C for 23 h. The buffer contained 0.9 M NaCl, 20 mM Tris
hydrochloride (pH 7.2), 0.01% sodium dodecyl sulfate and
formamide with concentrations listed in Table 1. Subsequently, a stringent washing step was performed at 48C for
15 min in 50 ml of washing solution (NaCl dependent on
formamide concentration, 20 mM Tris hydrochloride at pH 7.2

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

Microbial community structure in autotrophic nitrifying granules 203

Fig. 11. Spatial scales in the 2-d model of nitrifying granules.

A. Representative biomass granule comprised in a square computational domain.
B. The square grid elements discretizing the space, each containing several biomass particles.
C. Individual biomass particles of different possible biomass types. All biomass particles within a single grid element experience the same
substrate concentrations. The biomass concentrations within a grid element are calculated from the mass of all individual biomass particles
within the element volume.

and 0.01% sodium dodecyl sulfate). The slides were then

rinsed briefly with ddH2O and allowed to dry. After slides were
mounted in FluoroGuard Antifade Reagent (Bio-Rad, CA,
USA), probe-stained cells were detected and recorded by a
confocal laser scanning microscope (IX71, Olympus, Tokyo,
Japan) equipped with an Ar ion laser (488 nm) and a HeNe
laser (543 nm).
The spatial distributions of bacteria were quantitatively
analysed by using probes specific to the several phylogenetic
groups of bacteria. The quantitative analysis was conducted
along radial transects with the constant z position through
the granules by scanning adjacent fields of about
140 mm 1 mm. Six radial transects (about 40 confocal laser
scanning microscope images) from different sections were
analysed for each probe. For determination of bacterial distributions, the average surface areas of the specific-probehybridized cells and EUB338 probe mix hybridized cells were
quantified by simultaneous in situ hybridization with various
probe sets. Threshold values were defined to exclude background fluorescence and the probe-positive cell areas were
quantified by the software daime (Daims et al., 2006).

Nucleic acid and lectin staining

Propidium iodide was used to stain the nucleic acids of bacterial cells in the granules. Concanavalin (ConA) and WGA
lectin labelled with FITC were used to stain most glycoconjugate fractions of EPS (Bockelmann et al., 2002; McSwain
et al., 2005). Since these stains may also bind with protein
and glycoconjugate groups associated with cell walls, a
counter stain with PI was used to distinguish EPS from cells.

Microelectrode measurements
The DO profiles were obtained with a Clark-type microelectrode with a tip diameter of 1015 mm (Unisense, Aarhus,
Denmark). The DO microelectrode was calibrated according
to the manufacturers instructions. Liquid ion-exchanging
membrane microsensors for NH4+, NO2 and NO3 were pre-

pared and calibrated according to previous reports (De Beer

et al., 1997; Okabe et al., 1999). An intact whole granule was
fixed in a loop of a nylon thread connected to a loader tip and
placed in the flow cell (Hille et al., 2005). All the measurements were performed in a water chamber containing 300 ml
of medium taken from the reactor at 20C as described elsewhere (Matsumoto et al., 2007).

Biofilm model description

The general structure of the granule models used in this
study is based on biofilm models previously reported. Therefore, only a few specific details newly introduced here will be
described. The 1-d model is based on the work of Wanner
and Gujer (1985), and implemented to the well-established
AQUASIM simulation software (Reichert, 1998). The 2-d ibM
follows the principles described in Picioreanu and colleagues
(2004), and Xavier and colleagues (2005a; 2007).
A schematic diagram of the 2-d model spatial scales is
shown in Fig. 11. The development of a single granule is
simulated in a square computational domain defined in Cartesian coordinates, with size of 1800 mm. The computational
domain consists of three spatial subdomains: bulk liquid,
diffusion boundary layer and the granular biofilm. The bulk
liquid subdomain is considered completely mixed and, thus,
has a uniform distribution of the component concentrations.
The boundary layer is defined at constant thickness from the
granule surface. In the granule and the boundary layer subdomains, the time-dependent concentrations of solute components are calculated from diffusion-reaction equations.
Bulk concentrations of solute and particulate components are
calculated from overall mass balances in the system, including inlet-outlet flows, reactions in the bulk and mass transfer
to/from the granules. These mass balances are explained in
detail in Picioreanu and colleagues (2004) and Xavier and
colleagues (2005a).
The granule structure is represented by a collection of
discrete non-overlapping hard cylinders of solid (also called
particulate) components (see Picioreanu et al., 2004; Bat-

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

204 S. Matsumoto et al.

stone et al., 2006). The initial biomass particles are distributed within a small circle with the centre in the middle of the
square computational domain. Each biomass particle contains active biomass of a single microbial type and inert
material biomass and is surrounded by an EPS capsule produced by the biomass within the particle (see Xavier et al.,
2005a). Each biomass particle grows and produces EPS
according to a rate defined by concentrations of solutes and
particulates. The biomass particles are divided into two
daughter particles when their size exceeds a critical size. An
EPS-only particle is excreted when the mass fraction of EPS
relative to total particle mass reaches a critical value. After
division and EPS excretion, particles are redistributed
(shoved, Kreft et al., 2001) not to overlap each other. In this
sequence of steps the whole granule size develops. A
maximum granule diameter is set to be 1600 mm, which is a
typical average value found in the AUFB reactor. All biomass
particles growing outside this size are removed from the
computational domain, a computational operation similar to
biomass detachment.
The kinetic model for the biological processes assumes
that organic carbon for the growth of the heterotrophic bacteria was derived from three reactions: autotrophic growth
(UAP), biomass lysis (BAP) (Rittmann et al., 2002) and
hydrolysis of EPS (Org) (Laspidou and Rittmann, 2002) (see
Supporting information). Although Laspidou and Rittmann
(2002; 2004) defined that BAP is produced from the hydrolysis of EPS, in this study, to distinguish organic carbon derived
from biomass lysis and EPS lysis, we called them BAP and
Org respectively. In this particle-based biomass description,
all microbial groups (i.e. AOB, NOB, UAP-uptaking heterotrophic bacteria HetU, BAP-uptaking heterotrophic bacteria HetB and Org-uptaking heterotrophic bacteria HetO) and
EPS as well are particles defined by an internal composition
(mass of one or more particulate substances), by size and by
location in space. EPS is excreted by HetU, HetB and HetO,
but not by either AOB or NOB.
All model parameters are presented in the Supporting
information (Table S14) together with the complete description of all microbial processes considered. The source code
of the 2-d model was written in C++ and the model was
executed on a personal computer with Intel(R) Xeon CPU 3.0
GHz processor.

Acknowledgements
The authors would like to thank Professor S. Okabe and Dr Y.
Nakamura of Hokkaido University (JP) for their kind advice to
construct microelectrodes and measure solute concentration
microprofiles in biofilms. The work at TU Delft was financially
supported by the Netherlands Organization for Scientific
Research (NWO, VIDI Grant 864.06.003 to C.P.).

References
Alpkvist, E., and Klapper, I. (2007) A multidimensional multispecies continuum model for heterogeneous biofilm development. Bull Math Biol 69: 765789.
Alpkvist, E., Picioreanu, C., van Loosdrecht, M.C.M., and
Heyden, A. (2006) Three-dimensional biofilm model with

individual cells and continuum EPS matrix. Biotechnol

Bioeng 94: 961979.
Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman,
D.J. (1990) Basic local alignment search tool. J Mol Biol
215: 403410.
Amann, R.I., Krumholz, L., and Stahl, D.A. (1990)
Fluorescent-oligonucleotide probing of whole cells for
determinative, phylogenetic, and environmental studies in
microbiology. J Bacteriol 172: 762770.
Aoi, Y., Miyoshi, T., Okamoto, T., Tsuneda, S., Hirata, A.,
Kitayama, A., and Nagamune, T. (2000) Microbial ecology
of nitrifying bacteria in wastewater treatment process
examined by fluorescence in situ hybridization. J Biosci
Bioeng 90: 234240.
Barker, D.J., and Stuckey, D.C. (1999) A review of soluble
microbial products (SMP) in wastewater treatment
systems. Water Res 33: 30633082.
Batstone, D.J., Picioreanu, C., and Van Loosdrecht, M.C.M.
(2006) Multidimensional modelling to investigate interspecies hydrogen transfer in anaerobic biofilms. Water Res 40:
30993108.
Bell, A., Aoi, Y., Terada, A., Tsuneda, S., and Hirata, A.
(2005) Comparison of spatial organization in top-down and
membrane-aerated biofilms: a numerical study. Water Sci
Technol 52: 173180.
Bockelmann, U., Manz, W., Neu, T.R., and Szewzyk, U.
(2002) Investigation of lotic microbial aggregates by a
combined technique of fluorescent in situ hybridization
and lectin-binding-analysis. J Microbiol Methods 49: 75
87.
Cebron, A., and Garnier, J. (2005) Nitrobacter and Nitrospira
genera as representatives of nitrite-oxidizing bacteria:
detection, quantification and growth along the lower Seine
River (France). Water Res 39: 49794992.
Cottrell, M.T., and Kirchman, D.L. (2000) Natural assemblages of marine proteobacteria and members of the
Cytophaga-Flavobacter cluster consuming low- and highmolecular-weight dissolved organic matter. Appl Environ
Microbiol 66: 16921697.
Daims, H., Bruhl, A., Amann, R., Schleifer, K.H., and Wagner,
M. (1999) The domain-specific probe EUB338 is insufficient for the detection of all Bacteria: development and
evaluation of a more comprehensive probe set. Syst Appl
Microbiol 22: 434444.
Daims, H., Lucker, S., and Wagner, M. (2006) daime, a novel
image analysis program for microbial ecology and biofilm
research. Environ Microbiol 8: 200213.
De Beer, D., Schramm, A., Santegoeds, C.M., and Kuhl, M.
(1997) A nitrite microsensor for profiling environmental biofilms. Appl Environ Microbiol 63: 973977.
Gich, F., Garcia-Gil, J., and Overmann, J. (2001) Previously
unknown and phylogenetically diverse members of the
green nonsulfur bacteria are indigenous to freshwater
lakes. Arch Microbiol 177: 110.
Hille, A., Neu, T.R., Hempel, D.C., and Horn, H. (2005)
Oxygen profiles and biomass distribution in biopellets of
Aspergillus niger. Biotechnol Bioeng 92: 614623.
Juretschko, S., Timmermann, G., Schmid, M., Schleifer, K.H.,
Pommerening-Roser, A., Koops, H.P., and Wagner, M.
(1998) Combined molecular and conventional analyses of
nitrifying bacterium diversity in activated sludge: Nitroso-

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

Microbial community structure in autotrophic nitrifying granules 205

coccus mobilis and Nitrospira-like bacteria as dominant
populations. Appl Environ Microbiol 64: 30423051.
Kane, M.D., Poulsen, L.K., and Stahl, D.A. (1993) Monitoring
the enrichment and isolation of sulfate-reducing bacteria
by using oligonucleotide hybridization probes designed
from environmentally derived 16S rRNA sequences. Appl
Environ Microbiol 59: 682686.
Kindaichi, T., Ito, T., and Okabe, S. (2004) Ecophysiological
interaction between nitrifying bacteria and heterotrophic
bacteria in autotrophic nitrifying biofilms as determined by
microautoradiography-fluorescence in situ hybridization.
Appl Environ Microbiol 70: 16411650.
Kreft, J.U., and Wimpenny, J.W.T. (2001) Effect of biofilm
structure and function as revealed by an individual-based
model of biofilm growth. Water Sci Technol 43: 135141.
Kreft, J.U., Picioreanu, C., Wimpenny, J.W., and van Loosdrecht, M.C.M. (2001) Individual-based modelling of biofilms. Microbiology 147: 28972912.
Laspidou, C.S., and Rittmann, B.E. (2002) Non-steady state
modeling of extracellular polymeric substances, soluble
microbial products, and active and inert biomass. Water
Res 36: 19831992.
Laspidou, C.S., and Rittmann, B.E. (2004) Modeling the
development of biofilm density including active bacteria,
inert biomass, and extracellular polymeric substances.
Water Res 38: 33493361.
Loy, A., Horn, M., and Wagner, M. (2003) probeBase: an
online resource for rRNA-targeted oligonucleotide probes.
Nucleic Acids Res 31: 514516.
McSwain, B.S., Irvine, R.L., Hausner, M., and Wilderer, P.A.
(2005) Composition and distribution of extracellular polymeric substances in aerobic flocs and granular sludge.
Appl Environ Microbiol 71: 10511057.
Manz, W., Amann, R., Ludwig, W., Wagner, M., and Schleifer, K.H. (1992) Phylogenetic oligonucleotide probes for
the major subclasses of proteobacteroa: problems and
solutions. Syst Appl Microbiol 15: 593600.
Manz, W., Amann, R., Ludwig, W., Vancanneyt, M., and
Schleifer, K.H. (1996) Application of a suite of 16S rRNAspecific oligonucleotide probes designed to investigate
bacteria of the phylum cytophaga-flavobacter-bacteroides
in the natural environment. Microbiology 142: 1097
1106.
Matsumoto, S., Terada, A., Aoi, Y., Tsuneda, S., Alpkvist, E.,
Picioreanu, C., and van Loosdrecht, M.C.M. (2007) Experimental and simulation analysis of community structure of
nitrifying bacteria in a membrane-aerated biofilm. Water
Sci Technol 55: 283290.
Mobarry, B.K., Wagner, M., Urbain, V., Rittmann, B.E., and
Stahl, D.A. (1996) Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. Appl
Environ Microbiol 62: 21562162.
Muyzer, G., Brinkhoff, T., Nbel, U., Santegoeds, C.,
Schfer, H., and Wawer, C. (1997) Pdenaturing gradient
gel electrophoresis (DGGE) in microbial ecology. In
Molecular Microbial Ecology Manual, Vol. 3.4.4. Akkermans, A.D.L., Van Elaas, J.D., and de Bruijn, F.J. (eds).
Dordrecht, the Netherlands: Kluwer Academic Publishers,
pp. 127.
Nogueira, R., Melo, L.F., Purkhold, U., Wuertz, S., and
Wagner, M. (2002) Nitrifying and heterotrophic population

dynamics in biofilm reactors: effects of hydraulic retention

time and the presence of organic carbon. Water Res 36:
469481.
Noguera, D.R., Pizarro, G., Stahl, D.A., and Rittmann, B.E.
(1999) Simlation of multispecies biofilm development in
three dimensions. Water Sci Technol 39: 123130.
OSullivan, L.A., Weightman, A.J., and Fry, J.C. (2002) New
degenerate Cytophaga-Flexibacter-Bacteroides-specific
16S ribosomal DNA-targeted oligonucleotide probes reveal
high bacterial diversity in River Taff epilithon. Appl Environ
Microbiol 68: 201210.
Okabe, S., Hiratia, K., Ozawa, Y., and Watanabe, Y. (1996)
Spatial microbial distributions of nitrifiers and heterotrophs
in mixed-population biofilms. Biotechnol Bioeng 50: 2435.
Okabe, S., Satoh, H., and Watanabe, Y. (1999) In situ analysis of nitrifying biofilms as determined by in situ hybridization and the use of microelectrodes. Appl Environ Microbiol
65: 31823191.
Okabe, S., Kindaichi, T., and Ito, T. (2005) Fate of 14Clabeled microbial products derived from nitrifying bacteria
in autotrophic nitrifying biofilms. Appl Environ Microbiol 71:
39873994.
Picioreanu, C., Van Loosdrecht, M.C.M., and Heijnen, J.J.
(1999) Discrete-differential modelling of biofilm structure.
Water Sci Technol 39: 115122.
Picioreanu, C., Kreft, J.U., and Van Loosdrecht, M.C.M.
(2004) Particle-based multidimensional multispecies
biofilm model. Appl Environ Microbiol 70: 30243040.
Picioreanu, C., Batstone, D.J., and Van Loosdrecht, M.C.M.
(2005) Multidimensional modeling of anaerobic granules.
Water Sci Technol 52: 501507.
Prosser, J.I. (1989) Autotrophic nitrification in bacteria. Adv
Microb Physiol 30: 125181.
Pynaert, K., Smets, B.F., Wyffels, S., Beheydt, D., Siciliano,
S.D., and Verstraete, W. (2003) Characterization of an
autotrophic nitrogen-removing biofilm from a highly loaded
lab-scale rotating biological contactor. Appl Environ Microbiol 69: 36263635.
Reichert, P. (1998) AQUASIM 2.0Computer Program for
the Identification and Simulation of Aquatic Systems.
Version 2.0. Dubendorf, Switzerland: EAWAG.
Rittmann, B.E., Regan, J.M., and Stahl, D.A. (1994) Nitrification as a source of soluble organic substrate in biological
treatment. Water Sci Technol 30: 18.
Rittmann, B.E., Stilwell, D., and Ohashi, A. (2002) The
transient-state, multiple-species biofilm model for biofiltration processes. Water Res 36: 23422356.
Sekiguchi, Y., Kamagata, Y., Nakamura, K., Ohashi, A., and
Harada, H. (1999) Fluorescence in situ hybridization using
16S rRNA-targeted oligonucleotides reveals localization of
methanogens and selected uncultured bacteria in mesophilic and thermophilic sludge granules. Appl Environ
Microbiol 65: 12801288.
Tay, J.H., Liu, Q.S., and Liu, Y. (2002) Characteristics of
aerobic granules grown on glucose and acetate in sequential aerobic sludge blanket reactors. Environ Technol 23:
931936.
Tsuneda, S., Park, S., Hayashi, H., Jung, J., and Hirata, A.
(2001) Enhancement of nitrifying biofilm formation using
selected EPS produced by heterotrophic bacteria. Water
Sci Technol 43: 197204.

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206

206 S. Matsumoto et al.

Tsuneda, S., Nagano, T., Hoshino, T., Ejiri, Y., Noda, N., and
Hirata, A. (2003) Characterization of nitrifying granules produced in an aerobic upflow fluidized bed reactor. Water
Res 37: 49654973.
Tsuneda, S., Ejiri, Y., Nagano, T., and Hirata, A. (2004)
Formation mechanism of nitrifying granules observed in an
aerobic upflow fluidized bed (AUFB) reactor. Water Sci
Technol 49: 2734.
Tsuneda, S., Ogiwara, M., Ejiri, Y., and Hirata, A. (2006)
High-rate nitrification using aerobic granular sludge. Water
Sci Technol 53: 147154.
Vogelsang, C., Schramm, A., Picioreanu, C., van Loosdrecht,
M., and Ostgaard, K. (2002) Microbial community analysis
by FISH for mathematical modelling of selective enrichment of gel-entrapped nitrifiers obtained from domestic
wastewater. Hydrobiologia 469: 165178.
Wagner, M., Rath, G., Koops, H.P., Flood, J., and Amann, R.
(1996) In situ analysis of nitrifying bacteria in sewage treatment plants. Water Sci Technol 34: 237244.
Wanner, O., and Gujer, W. (1985) Competition in biofilms.
Water Sci Technol 17: 2744.
Wanner, O., Eberl, H., Morgenroth, E., Noguera, D., Picioreanu, C., Rittmann, B., and Van Loosdrecht, M.C.M. (2006)
Mathematical Modeling of Biofilms. London, UK: IWA
publishing.
Weisburg, W.G., Barns, S.M., Pelletier, D.A., and Lane, D.J.
(1991) 16S ribosomal DNA amplification for phylogenetic
study. J Bacteriol 173: 697703.
Weller, R., Glockner, F.O., and Amann, R. (2000) 16S rRNAtargeted oligonucleotide probes for the in situ detection
of members of the phylum Cytophaga-FlavobacteriumBacteroides. Syst Appl Microbiol 23: 107114.
Xavier, J.B., and Foster, K.R. (2007) Cooperation and conflict

in microbial biofilms. Proc Natl Acad Sci USA 104: 876

881.
Xavier, J.B., Picioreanu, C., and van Loosdrecht, M.C.M.
(2005a) A framework for multidimensional modelling of
activity and structure of multispecies biofilms. Environ
Microbiol 7: 10851103.
Xavier, J.B., Picioreanu, C., and van Loosdrecht, M.C.M.
(2005b) A general description of detachment for multidimensional modelling of biofilms. Biotechnol Bioeng 91:
651669.
Xavier, J.B., de Kreuk, M.M., Picioreanu, C., van Loosdrecht
and M.C.M. (2007) Multi-scale sindividual-based model of
microbial and bioconversion dynamics in aerobic granular
sludge. Environ Sci Technol 41: 64106417.

Supporting information
Additional Supporting Information may be found in the online
version of this article:
Fig. S1. Scheme of biological transformations taken into
account in the model.
Table S1. Stoichiometric matrix for microbial reactions.
Table S2. Kinetic rate expressions for microbial reactions.
Table S3. Stoichiometric parameters for microbial reactions.
Table S4. Kinetic parameters for microbial reactions.
Table S5. Other parameters used in the simulations.
Please note: Wiley-Blackwell are not responsible for the
content or functionality of any supporting materials supplied
by the authors. Any queries (other than missing material)
should be directed to the corresponding author for the
article.

2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 12, 192206