Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Characterization of Natural and Affected Environments
Occurrence of Legionella spp. in Water-Main Biofilms from Two Drinking Water Distribution Systems Michael Waak, Timothy M. LaPara, Cynthia Halle, and Raymond M Hozalski Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b01170 • Publication Date (Web): 14 Jun 2018 Downloaded from http://pubs.acs.org on June 15, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 33
Environmental Science & Technology
Occurrence of Legionella spp. in Water-Main Biofilms from Two Drinking Water Distribution Systems Michael B. Waak1, Timothy M. LaPara1,2, Cynthia Hallé3, Raymond M. Hozalski1,2*
1
Department of Civil, Environmental, and Geo-Engineering, University of Minnesota, 500
Pillsbury Dr. SE, Minneapolis, Minnesota 55455, U.S.A. 2
BioTechnology Institute, University of Minnesota, 1479 Gortner Ave., Saint Paul, Minnesota
55108, U.S.A. 3
Department of Civil and Environmental Engineering, Norwegian University of Science and
Technology, S.P. Andersens veg 5, Trondheim NO-7491, Norway
*Corresponding author; phone: +1 (612) 626-9650; e-mail:
[email protected] Abstract = 169 words 5113 words (main text) + 300 * 1 small figure + 600 * 3 large figures/tables = 7213 words
ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 33
1
ABSTRACT
2
The maintenance of a chlorine or chloramine residual to suppress waterborne pathogens
3
in drinking water distribution systems is common practice in the United States but less common
4
in Europe. In this study, we investigated the occurrence of Bacteria and Legionella spp. in water-
5
main biofilms and tap water from a chloraminated distribution system in the United States and a
6
system in Norway with no residual using real-time quantitative polymerase chain reaction
7
(qPCR). Despite generally higher temperatures and assimilable organic carbon levels in the
8
chloraminated system, total Bacteria and Legionella spp. were significantly lower in water-main
9
biofilms and tap water of that system (p < 0.05). Legionella spp. were not detected in the
10
biofilms of the chloraminated system (0 of 35 samples) but were frequently detected in biofilms
11
from the no-residual system (10 of 23 samples; maximum concentration = 7.8×104 gene
12
copies cm−2). This investigation suggests water-main biofilms may serve as a source of
13
legionellae for tap water and premise plumbing systems, and residual chloramine may aid in
14
reducing their abundance.
ACS Paragon Plus Environment
2
Page 3 of 33
Environmental Science & Technology
15
INTRODUCTION
16
After entering the drinking water distribution system (DWDS), drinking water may spend hours,
17
days, or even weeks before reaching consumer taps. In addition to the above-ground
18
infrastructure (i.e., pumps, water towers, and reservoirs), the predominant components of the
19
DWDS are the hundreds to thousands of kilometers of water mains buried underground. Because
20
of the relatively high surface area-to-volume ratio and often harsh environment of the bulk water,
21
more than 95% of the drinking water microbiome—the sum of all microorganisms within the
22
DWDS—exists in thin biofilms on the walls of the water mains.1 Biofilm bacteria may
23
exacerbate corrosion of iron water mains, produce unpleasant tastes and odors, decrease residual
24
concentrations of disinfectant (commonly chlorine and/or chloramine), and shed viable bacterial
25
cells into treated drinking water.2-5 Of particular concern is the role of biofilms as reservoirs of
26
waterborne pathogens, including opportunistic microbes like Legionella spp.6-8
27
Infections caused by Legionella spp. are known as legionellosis, which can manifest as
28
two clinical syndromes. Legionnaires’ disease is a severe pneumonia that is fatal for 1 in 10
29
cases, and Pontiac fever involves milder respiratory symptoms, similar to influenza.9 While all
30
legionellae are believed to be pathogenic, 90–95% of clinical cases of Legionnaires’ disease have
31
been attributed to L. pneumophila, with serogroup 1 causing up to 70% of cases.9 Legionellosis
32
is caused by inhalation of contaminated water droplets; drinking tap water that contains
33
Legionella spp. is not known to cause illness.9 Outbreaks have been associated with
34
contaminated premise plumbing (e.g., pipes and showerheads), spas and decorative water
35
features, and cooling towers,10 although public water supplies have been long suspected as a
36
means for the dispersal of Legionella to such environments.11,12
ACS Paragon Plus Environment
3
Environmental Science & Technology
37
Page 4 of 33
To minimize the risk of pathogen exposure via tap water, a residual disinfectant is used in
38
many parts of the world, particularly in the United States.13 The residual disinfectant is intended
39
to provide continuous suppression of the DWDS microbiome and prevent pathogen growth. Two
40
common disinfectants in drinking water are free chlorine (HOCl/OCl–) and combined chlorine
41
(primarily as monochloramine, NH2Cl).14 Despite their widespread usage, chlorine and
42
chloramine have well-known drawbacks. Chlorine-based disinfectants impart an unpleasant taste
43
and odor to tap water. In addition, the reaction of chlorine or chloramine with organic matter in
44
the water produces disinfection byproducts, which may have adverse health effects.15
45
Furthermore, chlorine and chloramine may enhance leaching of lead and other metals from
46
plumbing if adequate corrosion prevention has not been implemented.16,17
47
In contrast, disinfectant residuals are often low or nonexistent in many European cities.
48
The reduced reliance on chlorination may be due to the ability to obtain water from very high-
49
quality sources, the implementation of rigorous treatment barriers to remove bacteria and
50
nutrients from the water, as well as aggressive programs to maintain the DWDS infrastructure.18
51
A recent review of epidemiological data concluded that incidences of waterborne diseases were
52
no higher in European cities compared to cities in the United States, which suggests that safe
53
drinking water can be attained without the need for residual dinsinfection.13
54
In the present study, we collected water-main biofilms and tap water samples from a
55
chloramine-containing DWDS in the United States and a DWDS in Norway that does not
56
maintain a residual disinfectant. We initially used real-time quantitative polymerase chain
57
reaction (qPCR) targeting 16S ribosomal RNA (rRNA) genes to assess bacterial biomass in
58
biofilms and tap water. After Legionella-like operational taxonomic units were identified by
ACS Paragon Plus Environment
4
Page 5 of 33
Environmental Science & Technology
59
high-throughput sequencing of PCR-amplified 16S rRNA gene fragments, additional qPCR was
60
performed to quantify legionellae-specific gene markers in both DWDSs.
61
MATERIALS & METHODS
62
Drinking Water Distribution Systems. Water-main sections and tap water were obtained at
63
various locations in two drinking water distribution systems: a chloramine-containing system in
64
the United States (chloraminated system) and a system in Norway that does not maintain a
65
disinfectant residual (no-residual system). Water mains and tap water samples were collected
66
May–November 2014 in the chloraminated system and over several trips in June 2014, October
67
2014, and May 2015 in the no-residual system. The chloraminated system withdraws raw water
68
directly from a river, and treatment includes lime softening, recarbonation, alum coagulation,
69
sedimentation, filtration, and disinfection. Primary disinfection is achieved using free chlorine
70
and then the residual is quenched with ammonia to form chloramines before distribution to
71
consumers. The total chlorine concentration in the finished water is 3.8±0.1 mg Cl2 L–1 (mean ±
72
standard deviation). In contrast, the no-residual system treats water from a lake in a protected
73
and relatively pristine watershed. The soft raw water is hardened by passage through beds of
74
granular calcium carbonate (CaCO3), and then dosed with free chlorine to achieve a minimum of
75
0.05 mg Cl2 L–1 total chlorine after a 30-min contact time. The water is further disinfected by UV
76
radiation (40 mJ cm−2 using medium-pressure lamps) and then distributed to consumers with a
77
free chlorine residual of 0.08±0.01 mg Cl2 L–1.
78
Water-Main Biofilms. Sections of water mains were removed from the distribution systems
79
during routine maintenance activities. Photographs of several representative water mains are
80
provided in Figure 1. The sample collection sites were dictated by utility valve or main
81
replacement schedules. Only water mains with diameters of 15–20 cm (or 6–8 in) were collected;
ACS Paragon Plus Environment
5
Environmental Science & Technology
Page 6 of 33
82
larger mains were avoided to keep the transport and handling of sampled sections manageable.
83
Water mains were taken out of service by closing adjacent gate-valves no more than 12
84
hours prior to removal. To limit contamination of the pipe interior, the water was allowed to
85
remain in the main such that water evacuated when the main was cut. To mitigate contamination
86
of the sample while cutting, the exterior of the water main was brushed to remove adhered soil
87
and then rinsed with a chlorine-bleach solution (400 mg L−1) using a hand-operated sprayer. Cast
88
iron pipes made of gray iron or cement-lined ductile iron were cut with a hydraulic pipe cutter to
89
produce samples of 30–60 cm longitudinal length. As the hydraulic pipe cutter simply crushed
90
unlined ductile iron water mains, a hand-held chop saw was used for those mains. In these cases,
91
longer samples (75–100 cm) were collected and scraped for biofilms at least 20 cm from a cut
92
end to minimize the risk of biofilm disturbance or contamination from the chop saw.
93
A flame-sterilized steel microspatula was used to gently scrape three to four internal
94
surfaces of the water mains to sample surface biofilms (median area 1.3 cm2; range 0.4–8.5 cm2);
95
each biofilm scraping was treated as an individual biological replicate. The microspatula was
96
dipped and vigorously swirled in lysis buffer (5% sodium dodecyl sulfate, 120 mM sodium
97
phosphate, pH 8) to release sampled biofilms, with flame sterilization of the microspatula
98
between uses. Each sample set was accompanied by three negative controls, in which the flame-
99
sterilized microspatula was dipped in lysis solution. These controls were extracted for DNA
100
simultaneously with the biofilm samples.
101
Ten sampling sites were utilized for biomass sampling in the chloraminated DWDS (C1–
102
C10) and seven in the no-residual DWDS (N1–N7). Ideally, corresponding samples of water and
103
water-main biofilms would have been obtained from various locations throughout each system.
104
Although water can be readily collected throughout most systems, sites for water-main collection
ACS Paragon Plus Environment
6
Page 7 of 33
Environmental Science & Technology
105
could not be selected by our team but were dictated by utility DWDS maintenance schedules.
106
Unfortunately, because of accessibility and other issues it was not always possible to obtain
107
water from areas near water-main sampling and vice versa. In total, nine water mains were
108
collected from nine sites in the chloraminated DWDS (sites C2–C10), and six water mains were
109
collected from three sites (N5–N7) in the no-residual system (Table S1 in the Supporting
110
Information, SI). Water mains from the same site typically came from different sides of a 3- or 4-
111
way pipe intersection, may have had different pipe ages and materials, and were designated with
112
lower-case letters (e.g., a and b).
113
Tap Water Samples. Tap water samples were collected either directly from the water main via
114
pitot valves (chloraminated system) or via nearby public or private faucets (no-residual system).
115
Taps were flame-sterilized and flushed up to 1 min for pitot valves and 5 min for faucets prior to
116
sample collection to evacuate stagnant water. Water was then sampled using autoclave-sterilized
117
glass bottles. Three to four biological replicates were collected in separate bottles at each
118
location (approximately 1 L each) and transported to the laboratory in a cooler. Samples were
119
individually vacuum-filtered through separate 0.22-µm nitrocellulose membrane filters (47-mm
120
diameter; EMD Millipore; Bellerica, MA) within an hour of collection. The median volume of
121
filtrate was 1000 mL (range 747–1265 mL). Filters were directly submerged in microcentrifuge
122
tubes containing lysis buffer to initiate DNA extraction. Each set of tap water samples was
123
accompanied by three negative control filters, which were prepared by vacuum-filtering 2.0 mL
124
of PCR-grade water (Sigma-Aldrich; St. Louis, MO). In total, water was collected from six sites
125
in the chloraminated DWDS (sites C1, C2, C4, C6, C7, and C9) and seven sites in the no-
126
residual DWDS (N1–N4, N6, and N7) plus the treatment plant of that system (N0; Table S2).
ACS Paragon Plus Environment
7
Environmental Science & Technology
Page 8 of 33
127
DNA Extraction. To recover DNA after submersion in lysis buffer, the samples were subjected
128
to three freeze-thaw cycles followed by a 90-min incubation at 70°C. DNA was extracted using
129
the FastDNA SPIN Kit (MP Biomedicals; Santa Ana, CA) according to the manufacturer’s
130
instructions. Extracted DNA samples were stored at −20°C until further analysis.
131
Real-time Quantitative PCR (qPCR). Real-time quantitative polymerase chain reaction
132
(qPCR) was used to quantify several target genes. Primers targeting the V3 region of the 16S
133
rRNA gene (341F/534R) were used to measure total bacterial biomass, as described previously.19
134
Similarly, qPCR was used to quantify three different gene targets as measures of Legionella
135
biomass: ssrA for Legionella spp. (primers PanLegF/PanLegR and probe PanLegP),20 mip for L.
136
pneumophila (primers LpF/LpR and probe LpP),20 and wzm for L. pneumophila serogroup 1
137
(primers P65/P66).21 Full primer/probe sequences, PCR reaction concentrations, and PCR
138
thermoprofiles are described in Table S3. The oligonucleotides used for creating qPCR standard
139
curves are summarized in Table S4. Due to background amplification of 16S rRNA genes in
140
reaction reagents, there was no limit of detection (LOD) for this assay, and the limits of
141
quantification (LOQ) were defined as 10 times the gene copy number observed in no-template
142
controls, or 1.3×104 copies per reaction. The LOD and LOQ for the qPCR reactions targeting
143
ssrA, mip, and wzm were 5 and 10 copies per reaction, respectively, as previously described.22
144
Amplification efficiencies, LOQs, LODs, and standard curves are summarized in Table S5. The
145
method limits of detection and quantification (LODm and LOQm, respectively) for each sample
146
were defined as the LOD or LOQ normalized to the surface area (biofilms) or filtrate volume
147
(water). PCR reaction chemistry and quality control protocols are described in the SI Materials &
148
Methods.
ACS Paragon Plus Environment
8
Page 9 of 33
Environmental Science & Technology
149
Characterization of Legionella-like 16S rRNA Genes. The V3 region of the 16S rRNA gene
150
was PCR-amplified using modified versions of the 341F/534R primers and sequenced using the
151
MiSeq platform (Illumina, Inc.; San Diego, CA) to gather microbial community data, as
152
previously described.23 Samples with detectable but non-quantifiable 16S rRNA gene copy
153
numbers via qPCR were omitted from analysis to avoid biases from reagent contamination.24
154
Raw sequence reads were processed via the ‘metagenomics-pipeline’ version 1.5.25 Briefly, after
155
quality trimming, filtering, stitching of paired-end reads, and identification of chimeric
156
sequences, operational taxonomic units (OTUs) were clustered at 99% similarity for ‘species’-
157
level analysis26,27 and assigned consensus taxonomy using the SILVA 99% reference sequences
158
and taxonomy (release 128).28,29 Singleton OTUs and alignment failures were removed prior to
159
data analysis. Using the number of amplicon sequences per sample, we defined a practical limit
160
of detection (LODseq) as 1/[sequence depth + 1]. A description of sample preparation prior to
161
sequencing and detailed information about the bioinformatics pipeline are provided in the SI.
162
To confirm taxonomic identity of Legionella-like OTUs, representative sequences of the
163
5 most frequently observed Legionella-like OTUs in each sample type from both DWDSs (19
164
OTUs total when pooled) were manually searched using the National Center for Biotechnology
165
Information (NCBI) BLAST web utility30,31 with the GenBank32 nucleotide collection. A
166
multiple sequence alignment was then performed with MAFFT33 using the 19 Legionella-like
167
OTU sequences in addition to published 16S rRNA gene sequences (trimmed to the 341F/534R
168
V3 region) of 27 Legionella spp. and 4 phylogenetically related non-legionellae species (i.e.,
169
class Gammaproteobacteria). A phylogenetic tree was constructed with FastTree34 and rooted
170
using the midpoint.
ACS Paragon Plus Environment
9
Environmental Science & Technology
Page 10 of 33
171
Water Quality Analyses. Total chlorine in water samples was determined immediately after
172
sample collection using the N,N-diethyl-1,4-phenylenediamine (DPD) method and a Pocket
173
Colorimeter II (Hach Company; Loveland, CO). The water utilities provided daily total chlorine
174
concentrations for finished water for 2014 and average daily temperatures of the raw water at the
175
treatment plants for 2014–2015. Temperatures of distributed water at 15 monitoring taps across
176
the chloraminated DWDS were also provided. Comparable data were not available for the no-
177
residual DWDS, so temperature measurements were taken at four locations during May–June
178
2017. Assimilable organic carbon (AOC) was measured using the P-17/NOX method35,36 at three
179
locations of varying distances from the treatment plant in each DWDS in 2015 (chloraminated
180
system) or 2016–2017 (no-residual system). A detailed description of the method is provided in
181
the SI.
182
The chloraminated water utility provided daily pH as well as hardness, total ammonia
183
(NH3+NH4+ plus combined chlorine), free ammonia (NH3+NH4+), nitrate (NO3–), and total
184
phosphorus (P) concentrations for the treated water during 2014. The no-residual utility
185
measures pH daily while hardness, ammonium (NH4+), and nitrate are measured much less
186
frequently (i.e., several times per year) due to highly consistent raw and treated water quality.
187
Total and free ammonia were calculated from 2014 ammonium concentrations based on pH and
188
water temperature, as previously described.37 The no-residual utility does not monitor
189
phosphorus, so total phosphorus (Hach LCK 349) was determined in May 2017 for raw and
190
treated water. Methods used by the water utilities for water quality analyses are summarized in
191
Table S6.
192
Data Analysis and Statistics. Statistical tests and data transformations were performed with R
193
software.38 Hypothesis testing was performed using generalized Wilcoxon rank sum tests39 on
ACS Paragon Plus Environment
10
Page 11 of 33
Environmental Science & Technology
194
log10-transformed gene concentrations, with non-detects (i.e.,
