RESEARCH ARTICLE
Establishment and early succession of bacterial communities in
monochloramine-treated drinking water biofilms
Randy P. Revetta1, Vicente Gomez-Alvarez1, Tammie L. Gerke2, Claudine Curioso1,
Jorge W. Santo Domingo1 & Nicholas J. Ashbolt1
1
U.S. Environmental Protection Agency, Office of Research and Development, Cincinnati, OH, USA; and 2ORISE, U.S. Environmental Protection
Agency, Cincinnati, OH, USA
Correspondence: Nicholas J. Ashbolt,
U.S. Environmental Protection Agency,
Office of Research and Development,
Cincinnati, OH 45268, USA.
Tel.: 513 569 7318;
fax: 513 569 7464;
e-mail: [email protected]
Received 27 December 2012; revised 13 June
2013; accepted 14 June 2013.
Final version published online 17 July 2013.
DOI: 10.1111/1574-6941.12170
MICROBIOLOGY ECOLOGY
Editor: Cindy Nakatsu
Keywords
monochloramine; microbial communities;
colonization; succession; Mycobacterium;
drinking water.
Abstract
Monochloramine is an increasingly used drinking water disinfectant and has
been shown to increase nitrifying bacteria and mycobacteria in drinking waters.
The potential successions and development of these bacteria were examined by
16S rRNA gene clone libraries generated from various biofilms within a water
distribution system simulator. Biofilms were obtained from in-line and off-line
devices using borosilicate glass beads, along with polycarbonate coupons from
annular reactors incubated for up to 8 months in monochloramine-treated
drinking water. No significant difference in community structures was observed
between biofilm devices and coupon material; however, all biofilm communities that developed on different devices underwent similar successions over
time. Early stages of biofilm formation were dominated by Serratia (29%),
Cloacibacterium (23%), Diaphorobacter (16%), and Pseudomonas (7%), while
Mycobacterium-like phylotypes were the most predominant populations
(> 27%) in subsequent months. The development of members of the nontuberculous mycobacteria (NTM) after 3 months may impact individuals with
predisposing conditions, while nitrifiers (related to Nitrospira moscoviensis and
Nitrosospira multiformis) could impact water quality. Overall, 90% of the diversity in all the clone library samples was associated with the phyla Proteobacteria, Actinobacteria, and Bacteroidetes. These results provide an ecological insight
into biofilm bacterial successions in monochloramine-treated drinking water.
Introduction
Ecological succession has been described as a process of
colonization of newly available habitats (i.e. primary succession) followed by a progressive replacement over time
of the established community by new species (Connell &
Slatyer, 1977). By considering the effect of early animal
species on the establishment of later succession species,
Connell & Slatyer (1977) developed models indicating
that succession may either be accelerated (facilitation), be
unaffected (tolerance), or be slowed (inhibition). Describing the rate of succession in an intertidal coastal community, Farrell (1991) noted that not all early colonizers had
the same effect on later succession species and that the
model of succession was progressively different. In
contrast to other communities, there have been limited
ecological tools to characterize patterns in microbial
community assemblage during colonization and succession. Recent studies have quantitatively examined bacterial community structure in a phylogenetic framework
using various indices to characterize species distribution
including phylogenetic diversity (PD; Faith, 2002), net
relatedness index, and nearest taxon index (NRI and NTI,
respectively; Webb et al., 2002). For example, HornerDevine & Bohannan (2006) applied NRI and NTI indices
to bacterial diversity data (based on 16S rRNA gene
sequences) to determine whether bacterial communities
were phylogenetically structured. Most of the bacteria
they examined tended to co-occur with other bacteria
that were more closely related than expected by chance.
Furthermore, NRI can provide insight into drivers of
microbial community assembly (Bryant et al., 2008; Jones
& Hallin, 2010). For example, positive NRI values indicate phylogenetic clustered communities, where abiotic
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
FEMS Microbiol Ecol 86 (2013) 404–414
405
Bacterial diversity in drinking water biofilms
selective filters (e.g. environmental conditions) cause
assemblages to consist of closely related taxa (Webb et al.,
2002; Horner-Devine & Bohannan, 2006). Negative NRI
values indicate phylogenetic overdispersed communities
of distantly related species (Webb et al., 2002). Overdispersion of the community may be caused by two possible biotic interactions: competition (e.g. competitive
exclusion) and facilitation (Webb et al., 2002; HornerDevine & Bohannan, 2006). Facilitation may play an
important role in community structure by either promoting native diversity or creating microhabitats that permit
distantly related species (e.g. invasive species) to persist
within a local assemblage (Altieri et al., 2010).
Drinking water distribution systems (DWDS) are engineered environments that are subject to frequent, variable
disturbances caused by many factors including long residence times due to dead ends and low flow rates, each
associated with loss of disinfectant residual and generating
higher levels of biofilm growth (Keevil, 2009). In addition,
flow velocity, pressure, overall system configuration, pipe
sizes, and maintenance practices, such as regular flushing,
also affect biofilm development. Nonetheless, drinking
water (DW) biofilms are recognized as complex, dynamic
microbial assemblages with extensive phenotypic diversity
supporting adaptation to different hydraulic and chemical
surface conditions (Berry et al., 2006; Deines et al., 2010;
Roeder et al., 2010; Douterelo et al., 2013). Martiny et al.
(2003) observed that the succession development of a
model DW biofilm during a 3-year period was an orderly
process resulting in a stable, well-defined community. DW
biofilms are reservoirs of sequestered enteric and environmental pathogens, both of which present a potential problem when released back into the main DW flow (Szewzyk
et al., 2000). Indeed, environmental pathogens such as
Legionella pneumophila and Mycobacterium avium have
been shown to be associated with DW biofilms (Emtiazi
et al., 2004; Falkinham et al., 2008; Liu et al., 2012). Further, environmental DWDS pathogens are known to be
associated with the presence of free-living protozoa and,
in particular, various amoebae (Thomas & Loret, 2008;
Thomas & Ashbolt, 2011; Wingender & Flemming, 2011).
In contrast, very little is known about the potential impact
bacterial diversity plays on biofilm pathogen survival. To
shed some light on this issue, it is important to understand the taxonomic identity of DW bacterial biofilm
communities using robust methods.
The purposes of the present study included testing
whether biofilm samples obtained from three types of
biofilm devices (in-line and off-line placed devices containing borosilicate glass beads, along with polycarbonate
coupons from annular reactors) display a particular community structure and how these patterns vary temporally.
Furthermore, we wanted to test whether the phylogenetic
structures (e.g. clustering or overdispersion) reveal ecological interactions at the early stages of biofilm succession, by comparison of 16S rRNA gene clone libraries
obtained from biofilm devices incubated for up to
8 months in monochloramine-treated DW.
Materials and methods
Sample collection
Biofilm samples (n = 19) corresponding to four time
periods were collected over a period of 8 months from
various biofilm devices associated with a semi-closed pipe
loop distribution system simulator (Fig. 1a), which was
fed with chlorinated municipal DW amended with
ammonia to yield a 2 mg L 1 monochloramine residual.
Water in the pipe loop was recirculated, with c. 60% of
Inside device
Annular
reactor
(a)
Water flow
(b)
Feed Tanks
Fig. 1. (a) Schematic of sample locations
within the chloraminated pipe loop system. (b)
Perforated 316 stainless steel sheeting
cylindrical baskets filled with glass beads were
inserted in the main flow and in a single-pass
side stream via a PVC manifold as in-line (⊕)
and off-line (⊗) biofilm devices, respectively.
The pipe loop water was recirculated, and
approx. 60% was replaced daily.
FEMS Microbiol Ecol 86 (2013) 404–414
pipe
Water
Loop 6
• Material: PVC
• Length: 27m
• Diameter: 150 mm
Flow meter
Ammonia &
Sodium
hypochlorite
Sample tap
Inside
Chlorinated water
from treatment
plant
Heat exchanger
Recirculation pump
To drain
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
R.P. Revetta et al.
406
the total volume replaced daily. System properties and
water quality characteristics are presented in Table 1. Biofilm devices (i.e. glass bead baskets and polycarbonate
coupons) were used to study attached communities. Glass
bead baskets consisting of perforated 316 stainless steel
sheeting, rolled, and welded to create cylindrical baskets
were filled with c. 17.5 g of 2.5-mm-diameter glass beads
(Bio Specs Products Inc., Bartlesville, OK) (Fig. 1b).
These units were suspended in the main flow (i.e. in-line
devices), while additional baskets were incubated in a single-pass side stream (100 mL min 1) via a PVC manifold
connected directly to the pipe loop (i.e. off-line devices).
Additional off-line biofilm devices consisted of polycarbonate coupons incubated within annular reactors
(BioSurface Technologies, Bozeman, MT) receiving a
single-pass (100 mL min 1) side stream from the pipe
loop. Two annular reactors were used in this study, each
with an estimated hydraulic retention time of 10 min.
Samples from biofilm devices were removed for analysis
after 1, 2, 3, and ≥ 5 months of incubation. At each sample time point, one in-line glass bead basket (surface area
1597 cm2), one off-line glass bead basket (1597 cm2),
and ten polycarbonate coupons (175 cm2) were
collected for processing and analyses.
Water quality
Temperature, pH, total chlorine residual, turbidity, and
ORP measurements were taken from online sensors.
Nitrogen compounds were analyzed from a composite of
two samples taken from two different locations in the
loop. Nitrate/nitrite and ammonia analyses were performed with a discrete colorimetric SmartChem 200
autoanalyzer (Westco Scientific, Danbury CT) using the
EPA methods 353.2 (USEPA, 1993a) and 350.1 (USEPA,
1993b). Monochloramine analyses were performed using
an indophenol method using the HACH Method 10200
(HACH, Loveland, CO) according to the manufacturer’s
instructions. Although the nitrogen compound samples
were collected and analyzed 2 years after the biofilms were
removed, the same temperature, flow rates, chloramine,
and chlorine residual levels were maintained within the
simulated distribution system since 2008. Indeed, recent
water quality data for the simulated distribution system
operated using the same standard procedures and collected
over eight consecutive months indicated that there was
little change in the amount of nitrogen compounds during
that period (data not shown). As the system’s operating
conditions have not changed for the last 5 years, we
assume that the nitrogen compound results are indicative
of conditions in the pipe loop at the time of biofilm development and sample collection.
DNA extraction, PCR, cloning, and sequencing
Biofilm samples were independently harvested, and DNA
extracts were used to generate clones utilized in sequencing reactions. Briefly, biomass was removed from glass
beads or polycarbonate coupons by vortexing for 10 min
in 200 mL of sterile water. The homogenate was removed
Table 1. Water quality values for monochloramine-treated water in the pipe loop system
Characteristics
System
PVC pipe
Flow rate [m s 1]
Length [m]
Parameters*
Temperature [°C]
pH
Total chlorine residual [mg L 1]
Turbidity [NTU]
ORP† [mV]
Pump flow rate [g min 1]
Dissolved oxygen [mg L 1]
Nitrogen compounds‡
NH2Cl [Cl2 mg L 1]
NH3-N [mg L 1]
NO2-N [mg L 1]
NO3-N [mg L 1]
Loop
(01/09 to 12/10)
Sampling time
1 month
2 months
3 months
≥ 5 months
27.7 (0.5)
8.66 (0.04)
2.04 (0.03)
0.12 (0.01)
390.2 (5.0)
87.1 (0.1)
9.36 (0.15)
25.9 (0.3)
8.67 (0.02)
1.92 (0.02)
0.68 (0.12)
387.3 (3.7)
85.5 (0.4)
9.52 (0.15)
24.5 (0.3)
8.54 (0.10)
1.80 (0.04)
0.48 (0.09)
378.1 (2.4)
84.6 (0.4)
9.70 (0.12)
24.2 (0.2)
9.11 (0.01)
2.05 (0.03)
2.79 (0.05)
281.5 (1.0)
86.8 (0.9)
10.23 (0.04)
150 mm schedule gray 80
0.30
27
23.7 (0.5)
8.39 (0.28)
1.88 (0.11)
1.45 (0.29)
319.4 (13.0)
87.1 (0.5)
10.09 (0.11)
1.80
0.78
0.02
0.89
(0.03)
(0.01)
(0.00)
(0.01)
*Average (SE) results from daily measurements.
†
ORP, oxidation reduction potential.
‡
Average result for three independent samples (4x) taken from the pipe loop two years after the biofilms were removed.
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
FEMS Microbiol Ecol 86 (2013) 404–414
407
Bacterial diversity in drinking water biofilms
and centrifuged at 10 000 g for 5 min (i.e. pellet), and
the supernatant was then filtered onto 0.2-lm polycarbonate membranes (GE Osmonics, Minnetonka, MN).
DNA was isolated from both the membrane and pellet
for each sample using UltraClean Soil DNA kit following
the manufacturer’s instructions (MoBio Laboratories, Solana Beach, CA). The 16S rRNA gene was amplified from
total community DNA using bacterial primers Eub-8f and
787r (Lane, 1991). Primer coverage at the bacterial
domain level using RDP’s Probe Match tool (Cole et al.,
2009) and SILVA’s Test Prime tool (Klindworth et al.,
2013) showed that the primer set had a 58% coverage
when no mismatches were allowed, while the coverage
increased to 85% when 1 mismatch was allowed, which
is consistent with previous studies (Winsley et al., 2012;
Klindworth et al., 2013). In addition, four hypervariable
regions of the 16S rRNA gene are covered with this primer set. The DNA extracts were PCR-amplified in triplicate reactions. The reactions were then pooled for each
sample, cleaned with a QIAquick PCR purification kit
(QIAGEN, Valencia, CA), and cloned into the pCR4.1
TOPO TA vector following the manufacturer’s instructions (InvitrogenTM, Carlsbad, CA). Transformed cells
were grown on LB agar plates containing ampicillin
(100 mg mL 1). Random colonies were screened for PCR
products of expected size using M13 primers, and selected
clones were sequenced in both directions using ABI Big
Dyeâ Terminator sequencing chemistry on an ABI 3730xl
DNA Analyzer (Applied Biosystems, Foster City, CA).
Sequence assembly was performed using SEQUENCHER version 4.10 software (Gene Codes Corp., Ann Arbor, MI).
Prior to contig assembly, sequences with overall low quality scores were removed using SEQUENCHER. Base calling
for ambiguous sequences within the assembled contigs
was performed by inspecting chromatograms for such
bases. Sequences were further processed following the
protocol of Schloss et al. (2011). Specifically, reads classified as Chloroplasts and Mitochondria, identified as chimeric or containing homopolymers (greater than eight
nucleotides) were removed from further analysis using
the MOTHUR program (Schloss et al., 2009). A total of
4636 16S rRNA gene clones were analyzed in this study
with 274 sequences identified as chimeric using UCHIME
(Edgar et al., 2011). Sequences were deposited in GenBank with the following accession numbers: JX084275 to
JX087349.
Sequence analysis
Prior to analysis, clone libraries were normalized by randomization to the smallest dataset (i.e. 156 clones).
Curated sequences were aligned and grouped with 97%
sequence identity as the cutoff point for each operational
FEMS Microbiol Ecol 86 (2013) 404–414
taxonomic unit (OTU) using the software MOTHUR version
1.22.2 (http://www.mothur.org) (Schloss et al., 2009).
Identification and nearest neighbor of sequences were performed using the Classifier tool (Ribosomal Database Project II release 10.26) (Cole et al., 2009) and BLASTN
(Altschul et al., 1997), respectively. Phylogenetic trees were
constructed from the alignments based on the maximumlikelihood method and calculated using Tamura–Nei
model (Tamura et al., 2004). The software MEGA version
5.03 (Tamura et al., 2011) was used to build trees using
100 replicates to develop bootstrap confidence values.
Species richness and phylogenetic diversity
Normalized libraries were used to calculate the rarefaction
curves and phylogenetic diversity (PD). Rarefaction
curves are a result of repeated randomizations of the
observed species accumulation curve, allowing the
observed richness to be compared among samples
(Hughes et al., 2001). The Good’s coverage estimator was
used to estimate the percent of the total species represented in a sample. The quantitative measure of PD was
obtained by calculating the sum of the branch length in a
neighbor-joining phylogenetic tree that connects all species with the root (Faith, 2002). Tree construction and
calculations were performed using the software MOTHUR
(Schloss et al., 2009).
Community structure
Nonmetric multidimensional scaling (nMDS) analysis was
used to describe the relationships among biofilm communities based on the relative distribution of OTU groups.
A nMDS ordination plot based on the Bray–Curtis similarity coefficient of the transformed data (log[x+1]) was
generated using the software PAST version 2.03 (Hammer
et al., 2001). Phylogenetic community structure was compared using the net relatedness index (NRI), which is a
standardized measure based on the mean pairwise phylogenetic distance among OTUs in each sample (Webb
et al., 2002). NRI can provide insight into drivers of
microbial community assembly (Bryant et al., 2008; Jones
& Hallin, 2010). The NRI was calculated using the PHYLOCOM version 4.2 software (http://www.phylodiversity.net/
phylocom) with abundance data, null model #2
(i.e. OTUs in each sample become random draws from
the phylogeny pool without replacement), and 999
permutations (Webb et al., 2008).
Statistical analysis of community assemblages
Analysis of similarities (ANOSIM) was used to determine
whether there were significant differences (P < 0.05)
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
R.P. Revetta et al.
408
between biofilm devices (off-line, in-line, and annular
reactor), surface material used (glass bead and polycarbonate coupons), and sampling times (biofim age) on
community assemblages (see Table 2 for factor groups).
Interaction effects were tested using a two-way crossed
ANOSIM. R values (R test statistic) near 0 indicate a true
null hypothesis of no difference between groups, whereas
those > 0 (up to 1) indicate dissimilarities between
groups (Clarke, 1993). Similarity percentage (SIMPER)
analysis based on Bray–Curtis coefficients was used to
determine which species were most responsible for the
differences observed between biofilms of different sampling times (Clarke, 1993). Analyses were performed with
the software PRIMER 6 (PRIMER-E Ltd, Plymouth, UK).
Randomization procedures based on the Fisher’s exact
test with corrected P-values (Storey’s FDR multiple test
correction approach) using the software package STAMP
version 2.0 (http://kiwi.cs.dal.ca/Software/STAMP) (Parks
& Beiko, 2010) were used to identify statistically distinct
OTUs from similar sampling times. The relationship
between NRI values and sampling times was examined
using the nonparametric Kruskal–Wallis one-way analysis
of variance by ranks with multiple comparison pairwise
analysis (post hoc) to test the difference between each
sampling time (Siegel & Castellan, 1988). Tests to determine variation in NRI values were performed with the
software STATISTICA version 6.1 (StatSoft, Inc, Tulsa, OK).
Table 2. Results of ANOSIM test based on Bray–Curtis similarity matrix
derived from the distribution of biofilm microbial communities
Source of variation*
One-way ANOSIM
Global tests
Coupons (material)
Biofilm device
Two-way crossed ANOSIM
Global tests
Biofilm device
Sampling time
Pairwise tests‡
1 month vs. 2 months
1 month vs. 3 months
1 month vs. ≥ 5 months
2 months vs. 3 months
2 months vs. ≥ 5 months
3 months vs. ≥ 5 months
Global R†
0.078
0.071
P
0.247
0.207
Permutations
999
999
0.308
0.786
0.093
< 0.001
999
999
1
0.948
1
0.438
0.683
0.508
0.029
0.029
0.003
0.029
0.003
0.003
35
35
330
35
330
330
*Global tests on the effect of biofilm devices: in-line (n = 7), off-line
(n = 8), annular reactor (n = 4); support material: polycarbonate
(n = 4), glass beads (n = 15); and sampling time: 1 month (n = 4),
2 months (n = 4), 3 months (n = 4), ≥ 5 months (n = 7).
†
R values > 0 (up to 1) indicate dissimilarities between groups,
whereas those values near 0 indicate a true null hypothesis of no
difference between groups (Clarke, 1993).
‡
Pairwise tests on the effect of sampling time. Significance was set at
a = 0.05.
Results and discussion
Composition of clone libraries
Community analysis identified 233 operational taxonomic
units (OTUs) at 97% cutoff value detected in this study.
There was not a single OTU that was present in all of the
samples collected in this study. Nevertheless, collectively
several OTUs were represented in all sampling periods.
For example, our analysis identified OTU011 (i.e. Mycobacterium) in 16 (of 19) libraries, with just one representative found during the first month of biofilm
development. This could be the result of the relative
amount of sequences (low coverage) generated by the
Sanger sequencing technology. However, we should not
discard the fact that ecological and physiological factors,
such as competition and adaptation, could influence the
presence and distribution of microorganisms during the
first month of colonization. The sequencing technology
used in this study may only provide a minimal description of the complexity of the ecosystem (Lynch et al.,
2012).
Taxonomic analysis revealed that more than 90% of
the diversity in all the clone library samples was associated with the phyla Proteobacteria, Actinobacteria, and
Bacteroidetes (Supporting Information, Fig. S1). Bacteria
closely related to Firmicutes, Nitrospira, and the TM7
phylum were detected at ≤ 1% each (Fig. S1). Our taxonomic findings are broadly consistent with one of the few
studies that has exhumed actual distribution system pipes
and investigated the presence of opportunistic bacteria
pathogens from biofilms (Storey & Kaucner, 2009).
Capturing devices and sampling time effect
The results of one-way ANOSIM tests based on Bray–Curtis
similarity matrices derived from the distribution of
biofilm microbial communities showed no significant
difference (a ≥ 0.05) between support material used
(Global R = 0.078, P = 0.247) and capturing devices
(Global R = 0.071, P = 0.207) (Table 2). However, differences in community structure were evident when samples
were grouped by sampling time (two-way ANOSIM, Global
R = 0.786, P = < 0.001) (Table 2). Hence, relationships
between biofilm communities were more evident in connection with the period of biofilm growth than with the
capturing devices or support materials used (i.e. coupon
surfaces and associated hydrodynamics) (Henne et al.,
2012). This observation has important practical implications, as ranges of different devices/coupons are used to
examine DWDS biofilms, and it has been unclear as to
the impact/relevance any device may have on interpreting
biofilm composition within the distribution system
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
FEMS Microbiol Ecol 86 (2013) 404–414
409
Bacterial diversity in drinking water biofilms
(Storey & Ashbolt, 2002). The data were analyzed according to different sampling times to highlight succession of
taxonomic groups.
Species diversity
Rarefaction analysis of observed phylotypes was used to
assess the level of richness of the clone libraries generated
in this study. For biofilms sampled at 2, 3, and
≥ 5 months of age, the number of clones did not reach
saturation (Fig. 2a). By comparison, the rarefaction curve
for biofilm samples at 1 month appears to approach a
plateau, indicating that the sequencing effort has been
reasonably adequate (Good’s coverage estimator: 0.98).
Phylogenetic diversity (PD) curves show a similar trend
of increasing diversity with sampling time (Fig. 2b). Both
patterns support the observation of a less diverse population after 1 month. In general, these results are consistent
with other biofilm studies that suggest a wide variety of
low-abundance OTUs (Hong et al., 2010; Henne et al.,
2012). Although additional sampling is needed to more
accurately describe the total community in our samples,
normalization of our libraries (e.g. random subsampling
of clones) combined with sampling time replicates
(n ≥ 4) provided sufficient statistical power to allow
further assumptions about the structure of these biofilms.
Community structure
The resulting nMDS ordination plot, confirmed by ANOSIM
(Global R = 0.786, P < 0.001), highlighted marked bacterial community differences between the sampling times
(Fig. 3), indicating a high variability of the bacterial core
communities of the biofilms. The robustness of the
nMDS ordination was evaluated using the Shepard diagram (stress = 0.13), which is used for evaluating the
goodness of fit of an nMDS plot (e.g. < 0.1 indicates a
good fitting solution) (Wickelmaier, 2003). This clustering was further confirmed by pairwise ANOSIMs of bacterial
communities associated with clone libraries within a similar sampling time (Table 2). Four (of 88) taxonomic
groups explained 54% of the dissimilarity (SIMPER
analysis, P < 0.05) within all samples (Fig. 3). Overall, 1month-old biofilms were different from the rest of the
biofilms. These differences were attributed to the abundance of a few Proteobacteria populations, for example,
Serratia, Cloacibacterium, and Diaphorobacter, while members of the order Chromatiales and an increasing distribution of Mycobacterium-related species were dominant in
older biofilms (Fig. 3 and Fig. S2).
Colonization and succession
Phylogenetic structure of biofilm communities varied
with sampling times (one-way Kruskal–Wallis, H = 11.8,
d.f. = 3, P = 0.0081), suggesting various ecological strategies during colonization and succession, that is, biofilm
community development (Webb et al., 2002; Fierer et al.,
2010; Wang et al., 2012; Whitfeld et al., 2012). Net relatedness index (NRI) revealed a pattern of rapid increase in
clustering in the community from the first month to the
second and third months, followed by a gradual decrease
thereafter (curve line; r2 = 0.60, P < 0.005) (Fig. 4). During the first month of colonization, a significant phylogenetic overdispersion [NRI = 0.85 0.38, (mean standard error), post hoc comparison, P < 0.05] of the
community was observed (Fig. S3), suggesting competition for establishment in the newly available substrate
(Fierer et al., 2010). For example, 1-month-old biofilm was
dominated by OTUs closely identified with the bacterial
genera Serratia (29%), Cloacibacterium (23%), Diaphorobacter (16%), and Pseudomonas (7%) (Supporting Information, Table S1). Species of Serratia and Pseudomonas
are common and rapid-colonizing residents of DWDS
(Wingender & Flemming, 2004; Bachmann & Edyvean,
2005) and may contain possible opportunistic pathogens,
although their presence does not necessarily indicate a
risk to public health (Figueras & Borrego, 2010). Cloacibacterium spp. have been identified as dominant primarycolonizing biofilm bacteria in water lines of paper
machines (Tiirola et al., 2009), while isolates of the genus
Diaphorobacter have been shown to perform simultaneous
nitrification and denitrification under aerobic conditions
(Khardenavis et al., 2007).
(a)
(b)
FEMS Microbiol Ecol 86 (2013) 404–414
Phylogenetic diversity
Fig. 2. Rarefaction analysis of (a) observed
phylotypes (CI) and (b) phylogenetic diversity
for each sampling time. Each curve represents
the cumulative value of 660 clones from each
sampling time.
Richness (S)
100
1 month
2 month
3 month
≥ 5 month
80
60
40
20
0
1
80 160 240 320 400 480 560 640
Clones (n)
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
1 month
2 month
3 month
≥ 5 month
1
80 160 240 320 400 480 560 640
Clones (n)
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
R.P. Revetta et al.
410
0.02
0.15
0.01
≥ 5 month
0.05
Cloacibacterium
–0.4
–0.3
Mycobacterium
–0.2
1 month
Serratia
–0.1
0.1
–0.05
0.2
3 month
–0.01
2 month
–0.15
–0.02
Clustered
2
1
0
–1
–2
m
on
th
≥
3
m
on
th
2
m
on
th
1
m
on
th
Overdispersed
–3
5
Net Relatedness Index
3
Fig. 4. Variation in net relatedness index (NRI) for biofilm
communities. The NRI reveals a pattern of a rapid increase in
clustering in the community from the first month to the second and
third months, followed by a gradual decrease thereafter (curve
line; r2 = 0.60, P < 0.005). Positive and negative index values
indicate phylogenetically clustered and overdispersed communities,
respectively. Open symbols represent community phylogenetic
structures unlikely to arise by chance (P < 0.05).
During the second and third months of biofilm development, assemblages were comprised of closely related
taxa with similar NRI values (1.59 0.23 and
1.60 0.09, respectively, post hoc comparison, P = 0.99),
suggesting phylogenetic clustering of the community
(Fig. 4) (Webb et al., 2002). At 2 months of age, the
most abundant OTUs were closely related to the order
Chromatiales (40%) and species of Mycobacterium (27%),
0.3
Fig. 3. Nonmetric multidimensional scaling
(nMDS) ordination plot of biofilm communities
(n = 19) from various sampling times (months).
The analysis was based on Bray–Curtis
similarity coefficients calculated from the
relative distribution of OTUs. Contribution of
four taxonomic groups that explained 54%
(SIMPER analysis) of the dissimilarity within all
samples (two-way ANOSIM, R = 0.786,
P < 0.001) is represented by the size and
direction of vectors. All as genera, except
Chromatiales. Legend: 1 month (○), 2 months
(■), 3 months (□), ≥ 5 months (▲) biofilm
age.
specifically M. mucogenicum (Table S1 and Fig. S3). At
3 months, biofilms were dominated by M. arupense,
M. mucogenicum (23% and 15%, respectively), and
P. aeruginosa (10%) (Table S1 and Fig. S3), species with
members known to be opportunistic pathogens (Mena &
Gerba, 2009; Moritz et al., 2010). Phylogenetic clustering
between local assemblages has been interpreted as evidence of positive interaction or habitat filtering, where
groups with suitable colonization abilities survive (Webb
et al., 2002; Whitfeld et al., 2012).
Decrease in NRI values after 5 months (0.24 0.40,
post hoc comparison, P < 0.05) could be the result of the
combination of positive (environmental filtering and
facilitation) and negative (competition) ecological interactions (Webb et al., 2002; Fierer et al., 2010). For example,
the availability of new niches creates an opportunity
for distantly related taxa of functional specific groups
(e.g. nitrifying bacteria) to co-exist with local assemblages
(Fig. S3). Furthermore, the most dominant Mycobacterium
group (OTU 011) represented 50% of the total biofilm
community, while other mycobacteria represented < 4%
of the total (Table S1 and Fig. S3). The sequence of
Mycobacterium OTU 011 was closely related to M. gordonae, slow-growing nontuberculous mycobacteria (NTM)
often isolated from potable water supplies (Torvinen
et al., 2004; Liu et al., 2012). It should be noted that M.
gordonae has been implicated in symptomatic cases from
individuals with predisposing conditions (Lalande et al.,
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
FEMS Microbiol Ecol 86 (2013) 404–414
411
Bacterial diversity in drinking water biofilms
2001); nonetheless, it is often considered a contaminant
of clinical specimens (Arnow et al., 2000). In general, the
results from this study show that slow-growing Mycobacterium is not the main primary biofilm colonizer; however, their predominance after 2 months suggests that
they can become integrated into the biofilm as important,
persistent secondary colonizers of chloraminated DW systems, consistent with previous chloraminated DW studies
(Pryor et al., 2004; Berry et al., 2010).
It is important to note that biofilms are dynamic structures, with material sloughing off and thin ‘streamers’
breaking off, both of which may redeposit on biofilms
downstream as well as stimulate more rapid biofilm
regrowth. Biofilm detachment may result from changes
in hydraulic regime, with erosion, the dominant release
mechanism under steady-state conditions, whereas sloughing predominant following the sudden increase in shear
stress (Choi & Morgenroth, 2003) and under turbulent
flow conditions (e.g. high Reynolds numbers) (Storey &
Ashbolt, 2002). Due to differential biofilm release phenomena, Lewandowski et al. (2004) only studied young
biofilms to achieve reproducibility in annular reactors,
whereas Telgmann et al. (2004) suggest that biofilm
sloughing and growth should be considered normal processes in the development of biofilms and studied
together. Our results indicate that, at the scale of biofilm
community structure, reproducibility was possible over
months, but could at least in part be due to redeposition
of previous upstream biofilms in our system with considerable recirculation. Other stressors, such as a change in
DW disinfectant concentration or type or pressure
wave (e.g. ‘water hammer’), may also trigger biofilm
detachment (Daly et al., 1998).
Public health
It has become evident that biofilms in DWDS can
become transient or long-term habitats for opportunistic
environmental pathogens (Wingender & Flemming,
2011); therefore, it is important to understand the ecology of these environments to better assess potential public health risks and ultimately to develop control options
for such pathogens. Clinically important environmental
bacteria such as species/strains of Mycobacterium, Legionella, and Pseudomonas may be adapted to oligotrophic
conditions in DWDS and thus may persist and even multiply in DWDS biofilms (Emtiazi et al., 2004). Specifically, potentially pathogenic members of the NTM and P.
aeruginosa were identified in this study. In addition, compared with bacterial cells within the planktonic fraction,
bacteria released from biofilms into the water phase may
develop phenotypic properties related to increased disinfectant resistance and infectivity (Steed & Falkinham,
FEMS Microbiol Ecol 86 (2013) 404–414
2006). Therefore, the biofilm mode of survival for opportunistic pathogens in DWDS, such as environmental
NTM, should be considered in any risk assessment of the
public health impact from these engineered environments.
In addition, a more comprehensive understanding of the
microbial ecology of DW systems relative to disinfection
regime will enable more effective management of water
quality in these engineered systems and subsequently help
to safeguard human health.
Conclusion
Differences in community structure were evident when
samples were grouped by month, suggesting that all biofilm communities underwent similar successions over
time. Early stages of biofilm formation were dominated
by phylotypes classified as members of the Serratia, Cloacibacterium, Diaphorobacter, and Pseudomonas, while
Mycobacterium-like phylotypes were the most predominant populations in subsequent months. While NTM-like
sequences did not appear to be the main primary biofilm
colonizers, their subsequent dominance is of potential
concern, given that this group has been implicated in
outbreaks targeting individuals with predisposing conditions. Previous studies of biofilms in chloraminated DW
distribution systems detected a high presence of species
related to Mycobacterium (Falkinham et al., 2001; Beumer
et al., 2010). In addition, mycobacteria are predominant
in model biofilm systems receiving chloraminated DW
(Williams et al., 2005) and within premise plumbing
(Feazel et al., 2009). Moreover, recent metagenomic
analyses of DW receiving either free chlorine or
monochloramine treatment confirmed the ubiquity of
Mycobacterium-like sequences in chloramine-treated
samples (Gomez-Alvarez et al., 2012).
The quantitative analysis of the community assemblages reveals a pattern of a rapid increase in clustering in
the community from the first month to the second and
third months, which has been interpreted as evidence of
competition for the establishment in the newly available
substrate. This was followed by a gradual decrease thereafter, suggesting the availability of new niches creates an
opportunity for distantly related taxa of functional specific groups (e.g. nitrifying bacteria) to co-exist with local
assemblages. While flow conditions (hydraulics) are
thought to be critical for different biofilm thicknesses and
structures (Douterelo et al., 2013), most previous work
has focused on nutrient-rich environments (Battin et al.,
2003; Xavier et al., 2004). Yet, in this study, neither
in-line devices with different flow conditions compared
with the off-line devices nor the substrata appeared to be
significant, whereas biofilm age was. These results provide
an ecological insight into biofilm succession of bacterial
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
R.P. Revetta et al.
412
populations in monochloramine-treated DW systems,
which should ultimately facilitate effective management
of water quality in these engineered systems and subsequently help to safeguard human health.
Acknowledgements
The authors thank Sarah Staggs and Ian Struewing for
their assistance with sample analysis and data collection.
The opinions expressed in this manuscript are those of
the authors and do not necessarily reflect the official positions and policies of the US EPA. Any mention of product or trade names does not constitute recommendation
for use by the US EPA.
References
Altieri AH, van Wesenbeeck BK, Bertness MD & Silliman BR
(2010) Facilitation cascade drives positive relationship
between native biodiversity and invasion success. Ecology 91:
1269–1275.
Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z,
Miller W & Lipman DJ (1997) Gapped BLAST and
PSI-BLAST: a new generation of protein database search
programs. Nucleic Acids Res 25: 3389–3402.
Arnow PM, Bakir M, Thompson K & Bova JL (2000) Endemic
contamination of clinical specimens by Mycobacterium
gordonae. Clin Infect Dis 31: 472–476.
Bachmann RT & Edyvean RGJ (2005) Biofouling: an historic
and contemporary review of its causes, consequences and
control in drinking water distribution systems. Biofilms 2:
197–227.
Battin TJ, Kaplan LA, Newbold JD, Cheng X & Hansen C
(2003) Effects of current velocity on the nascent architecture
of stream microbial biofilms. Appl Environ Microbiol 69:
5443–5452.
Berry D, Xi C & Raskin L (2006) Microbial ecology of
drinking water distribution systems. Curr Opin Biotechnol
17: 297–302.
Berry D, Horn M, Xi C & Raskin L (2010) Mycobacterium
avium infections of Acanthamoeba strains: host strain
variability, grazing-acquired infections, and altered dynamics
of inactivation with monochloramine. Appl Environ
Microbiol 76: 6685–6688.
Beumer A, King D, Donohue M, Mistry J, Covert T &
Pfaller S (2010) Detection of Mycobacterium avium subsp.
paratuberculosis in drinking water and biofilms
by quantitative PCR. Appl Environ Microbiol 76:
7367–7370.
Bryant JA, Lamanna C, Morlon H, Kerkhoff AJ, Enquist BJ &
Green JL (2008) Microbes on mountainsides: contrasting
elevational patterns of bacterial and plant diversity. P Natl
Acad Sci USA 105: 11505–11511.
Choi YC & Morgenroth E (2003) Monitoring biofilm
detachment under dynamic changes in shear stress using
laser-based particle size analysis and mass fractionation.
Water Sci Technol 47: 69–76.
Clarke KR (1993) Non-parametric multivariate analyses of
changes in community structure. Aust J Ecol 18: 117–143.
Cole JR, Wang Q, Cardenas E et al. (2009) The Ribosomal
Database Project: improved alignments and new tools for
rRNA analysis. Nucleic Acids Res 37: 141–145.
Connell JH & Slatyer RO (1977) Mechanisms of succession in
natural communities and their role in community stability
and organization. Am Nat 111: 1119–1144.
Daly B, Betts WB, Brown AP & O’Neill JG (1998) Bacterial
loss from biofilms exposed to free chlorine. Microbios 96:
7–21.
Deines P, Sekar R, Husband PS, Boxall JB, Osborn AM &
Biggs CA (2010) A new coupon design for simultaneous
analysis of in situ microbial biofilm formation and
community structure in drinking water distribution systems.
Appl Microbiol Biotechnol 87: 749–756.
Douterelo I, Sharpe RL & Boxall JB (2013) Influence of
hydraulic regimes on bacterial community structure and
composition in an experimental drinking water distribution
system. Water Res 47: 503–516.
Edgar RC, Haas BJ, Clemente JC, Quince C & Knight R
(2011) UCHIME improves sensitivity and speed of chimera
detection. Bioinformatics 27: 2194–2200.
Emtiazi F, Schwartz T, Marten SM, Krolla-Sidenstein P & Obst
U (2004) Investigation of natural biofilms formed during
the production of drinking water from surface water
embankment filtration. Water Res 38: 1197–1206.
Faith DP (2002) Quantifying biodiversity: a phylogenetic
perspective. Conserv Biol 16: 248–252.
Falkinham JO III, Norton CD & LeChevallier MW (2001)
Factors influencing numbers of Mycobacterium avium,
Mycobacterium intracellulare, and other mycobacteria in
drinking water distribution systems. Appl Environ Microbiol
67: 1225–1231.
Falkinham JO III, Iseman MD, deHass P & vanSoolinger D
(2008) Mycobacterium avium in a shower linked to
pulmonary disease. J Water Health 6: 209–213.
Farrell TM (1991) Models and mechanisms of succession: an
example from a rocky intertidal community. Ecol Monogr
61: 95–113.
Feazel LM, Baumgartner LK, Peterson KL, Frank DN, Harris
JK & Pace NR (2009) Opportunistic pathogens enriched in
showerhead biofilms. P Natl Acad Sci USA 106:
16393–16399.
Fierer N, Nemergut D, Knight R & Craine JM (2010) Changes
through time: integrating microorganisms into the study of
succession. Res Microbiol 161: 635–642.
Figueras MJ & Borrego JJ (2010) New perspectives in
monitoring drinking water microbial quality. Int J Environ
Res Public Health 7: 4179–4202.
Gomez-Alvarez V, Revetta RP & Santo Domingo JW (2012)
Metagenomic analyses of drinking water receiving different
disinfection treatments. Appl Environ Microbiol 78:
6095–6102.
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
FEMS Microbiol Ecol 86 (2013) 404–414
413
Bacterial diversity in drinking water biofilms
Hammer Ø, Harper DAT & Ryan PD (2001) PAST:
paleontological statistics software package for evolution and
data analysis. Palaeontol Electronica 4: 1–9.
Henne K, Kahlisch L, Brettar I & H€
ofle MG (2012) Analysis of
structure and composition of bacterial core communities in
mature drinking water biofilms and bulk water of a citywide
network in Germany. Appl Environ Microbiol 78: 3530–3538.
Hong PY, Hwang C, Ling F, Andersen GL, LeChevallier MW
& Liu WT (2010) Pyrosequencing analysis of bacterial
biofilm communities in water meters of a drinking water
distribution system. Appl Environ Microbiol 76: 5631–5635.
Horner-Devine MC & Bohannan BJM (2006) Phylogenetic
clustering and overdispersion in bacterial communities.
Ecology 87: S100–S108.
Hughes JB, Hellmann JJ, Ricketts TH & Bohannan BJM (2001)
Counting the uncountable: statistical approaches to
estimating microbial diversity. Appl Environ Microbiol 67:
4399–4406.
Jones CM & Hallin S (2010) Ecological and evolutionary
factors underlying global and local assembly of denitrifier
communities. ISME J 4: 633–641.
Keevil CW (2009) Biofilms in water systems - origins and
treatment. Culture 30: 13–16.
Khardenavis AA, Kapley A & Purohit HJ (2007) Simultaneous
nitrification and denitrification by diverse Diaphorobacter sp.
Appl Microbiol Biotechnol 77: 403–409.
Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn
M & Gl€
ockner FO (2013) Evaluation of general 16S
ribosomal RNA gene PCR primers for classical and
next-generation sequencing-based diversity studies. Nucleic
Acids Res 41: e1.
Lalande V, Barbut F, Varnerot A, Febvre M, Nesa D, Wadel S,
Vincent V & Petit JC (2001) Pseudo-outbreak of
Mycobacterium gordonae associated with water from
refrigerated fountains. J Hosp Infect 48: 76–79.
Lane DJ (1991) 16S/23S rRNA sequencing. Nucleic Acid
Techniques in Bacterial Systematics (Stackebrandt E &
Goodfellow M, eds), pp. 115–175. Wiley, New York, NY.
Lewandowski Z, Beyenal H & Stookey D (2004)
Reproducibility of biofilm processes and the meaning of
steady state in biofilm reactors. Water Sci Technol
49: 359–364.
Liu R, Yu Z, Zhang H, Yang M, Shi B & Liu X (2012)
Diversity of bacteria and mycobacteria in biofilms of two
urban drinking water distribution systems. Can J Microbiol
58: 261–270.
Lynch MD, Bartram AK & Neufeld JD (2012) Targeted
recovery of novel phylogenetic diversity from
next-generation sequence data. ISME J 6: 2067–2077.
Martiny AC, Jorgensen TM, Albrechtsen H-J, Arvin E & Molin
S (2003) Long-term succession of structure and diversity of
a biofilm formed in a model drinking water distribution
system. Appl Environ Microbiol 69: 6899–6907.
Mena KD & Gerba CP (2009) Risk assessment of Pseudomonas
aeruginosa in water. Rev Environ Contam Toxicol 201:
71–115.
FEMS Microbiol Ecol 86 (2013) 404–414
Moritz MM, Flemming HC & Wingender J (2010) Integration
of Pseudomonas aeruginosa and Legionella pneumophila in
drinking water biofilms grown on domestic plumbing
materials. Int J Hyg Environ Health 213: 190–197.
Parks DH & Beiko RG (2010) Identifying biologically relevant
differences between metagenomic communities.
Bioinformatics 26: 715–721.
Pryor M, Springthorpe S, Riffard S, Brooks T, Huo Y, Davis G
& Sattar S (2004) Investigation of opportunistic pathogens
in municipal drinking water under different supply and
treatment regimes. Water Sci Technol 50: 83–90.
Roeder RS, Lenz J, Tarne P, Gebel J, Exner M & Szewzyk U
(2010) Long-term effects of disinfectants on the community
composition of drinking water biofilms. Int J Hyg Environ
Health 213: 183–189.
Schloss PD, Westcott SL, Ryabin T et al. (2009) Introducing
mothur: open-source, platform-independent,
community-supported software for describing and
comparing microbial communities. Appl Environ Microbiol
75: 7537–7541.
Schloss PD, Gevers D & Westcott SL (2011) Reducing the
effects of PCR amplification and sequencing artifacts on 16S
rRNA-based studies. PLoS ONE 6: e27310.
Siegel S & Castellan NJ (1988) Nonparametric Statistics for the
Behavioural Sciences. McGraw-Hill, New York, NY.
Steed KA & Falkinham JO III (2006) Effect of growth in
biofilms on chlorine susceptibility of Mycobacterium avium
and Mycobacterium intracellulare. Appl Environ Microbiol 72:
4007–4011.
Storey MV & Ashbolt NJ (2002) A comparison of methods
and models for the analysis of distribution pipe biofilms.
Water Sci Technol Wat Supp 2: 73–80.
Storey MV & Kaucner CE (2009) Understanding the growth of
opportunistic pathogens within distribution systems,
Research Report No. 79; Water Quality Research Australia
Limited, Cooperative Research Centre for Water Quality
and Treatment; Adelaide, Australia.
Szewzyk U, Szewzyk R, Manz W & Schleifer K-H (2000)
Microbiological safety of drinking water. Annu Rev
Microbiol 54: 81–127.
Tamura K, Nei M & Kumar S (2004) Prospects for inferring
very large phylogenies by using the neighbor-joining
method. P Natl Acad Sci USA 101: 11030–11035.
Tamura K, Peterson D, Peterson N, Stecher G, Nei M & Kumar
S (2011) MEGA5: Molecular Evolutionary Genetics Analysis
using maximum likelihood, evolutionary distance, and
maximum parsimony methods. Mol Biol Evol 28: 2731–2739.
Telgmann U, Horn H & Morgenroth E (2004) Influence of
growth history on sloughing and erosion from biofilms.
Water Res 38: 3671–3684.
Thomas JM & Ashbolt NL (2011) Do free-living amoeba in
treated drinking water systems present an emerging health
risk? Environ Sci Technol 45: 860–869.
Thomas V & Loret J-F (2008) Biodiversity of amoebae and
amoebae-resisting bacteria in a drinking water treatment
plant. Environ Microbiol 10: 2728–2745.
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
R.P. Revetta et al.
414
Tiirola M, Lahtinen T, Vuento M & Oker-Blom C (2009)
Early succession of bacterial biofilms in paper machines.
J Ind Microbiol Biotechnol 36: 929–937.
Torvinen E, Suomalainen S, Lehtola MJ, Miettinen IT, Zacheus
O, Paulin L, Katila ML & Martikainen PJ (2004)
Mycobacteria in water and loose deposits of drinking water
distribution systems in Finland. Appl Environ Microbiol 70:
1973–1981.
USEPA (United States Environmental Protection Agency)
(1993a) EPA method 353.2. Determination of Nitrate-Nitrite
by Automated Colorimetry (O'Dell JW, ed.), pp. 353.2-1–
353.2-15. Office of Research and Development, Cincinnati,
OH.
USEPA (United States Environmental Protection Agency)
(1993b) EPA method 350.1. Methods for the Determination
of Inorganic Substances in Environmental Samples (O'Dell
JW, ed.), pp. 350.1-1–350.1-15. EPA/600/R-093/100. Office
of Research and Development, Cincinnati, OH.
Wang J, Soininen J, He J & Shen J (2012) Phylogenetic
clustering increases with elevation for microbes. Environ
Microbiol Rep 4: 217–226.
Webb CO, Ackerly DD, McPeek MA & Donoghue MJ (2002)
Phylogenies and community ecology. Annu Rev Ecol Syst 33:
475–505.
Webb CO, Ackerly DD & Kembel SW (2008) Phylocom:
software for the analysis of phylogenetic community
structure and trait evolution. Bioinformatics 24:
2098–2100.
Whitfeld TJS, Kress WJ, Erickson DL & Weiblen GD (2012)
Change in community phylogenetic structure during
tropical forest succession: evidence from New Guinea.
Ecography 35: 821–830.
Wickelmaier F (2003) An introduction to MDS. Reports from
the Sound Quality Research Unit (SQRU) No. 7, Aalborg
University, Denmark, http://homepages.uni-tuebingen.de/
florian.wickelmaier/pubs/Wickelmaier2003SQRU.pdf.
Williams MM, Santo Domingo JW & Meckes MC (2005)
Population diversity in model potable water biofilms
receiving chlorine or chloramine residual. Biofouling 21:
279–288.
Wingender J & Flemming HC (2004) Contamination potential
of drinking water distribution network biofilms. Water Sci
Technol 49: 277–286.
Wingender J & Flemming HC (2011) Biofilms in drinking
water and their role as reservoir for pathogens. Int J Hyg
Environ Health 214: 417–423.
Winsley T, van Dorst JM, Brown MV & Ferrari BC (2012)
Capturing greater 16S rRNA gene sequence diversity within
the domain Bacteria. Appl Environ Microbiol 78: 5938–5941.
Xavier JB, Picioreanu C & van Loosdrecht MC (2004)
Assessment of three-dimensional biofilm models through
direct comparison with confocal microscopy imaging. Water
Sci Technol 49: 177–185.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Table S1. Taxonomic affiliation of the top 10 most
abundant Bacteria domain representatives at each
sampling timea.
Fig. S1. Distribution of the Bacteria domain as
determined by taxonomic identification of partial 16S
rRNA gene sequencing (at class level) of biofilm samples
from each sampling time.
Fig. S2. Relative distribution (%) of four taxonomic
groups that explained 54% of the dissimilarity within
all samples.
Fig. S3. Phylogenetic relationships among the ten most
abundant phylotypes from each sampling time.
Published 2013. This article is a U.S. Government work and is in the public domain in the U.S.A.
FEMS Microbiol Ecol 86 (2013) 404–414