Attachment Re Mobilization and Inactivation of Bacterioph 2022 Journal of
Attachment Re Mobilization and Inactivation of Bacterioph 2022 Journal of
Attachment Re Mobilization and Inactivation of Bacterioph 2022 Journal of
A R T I C L E I N F O A B S T R A C T
Keywords: Viruses, including human pathogenic viruses, can persist in water. For producing drinking water from surface
Virus transport water via bank filtration, natural attenuation capacities and the fate of viruses during the passage of aquatic
MS2 sediments are of particular interest. Moreover, the increasing frequency of extreme hydrological events neces
Bank filtration
sitate re-evaluation of the sustainability and efficacy of processes removing viruses. For this purpose, we per
Inactivation
Hydrological extremes
formed bank sediment filtration experiments using a mesocosm in a technical-scale experimental facility that
Climate change simulates a field situation under more tightly controlled conditions. We used the bacteriophage MS2 as a sur
rogate for enteric viruses to study the transport of different viral loads through the bank sediment. Additionally,
we simulated a heavy rain event to investigate the re-mobilization of initially attached virus particles. We
quantified the abundance of infectious MS2 phages by plaque assay and the total number of MS2 particles by
qPCR. Also, we differentiated pore water concentrations by depths of the sediment column and investigated
attachment to the sediment matrix at the end of the individual experimental phases. Bank filtration over a
vertical distance of 80 cm through sandy sediment revealed a virus removal efficiency of 0.8 log10 for total MS2
particles and 1.7 log10 for infectious MS2 particles, with an initial phage concentration of 1.84 × 108 gene copies
mL− 1. A low load of infectious MS2 (1.9 × 106 plaque forming units mL− 1) resulted in a greater removal effi
ciency (3.0 log10). The proportion of infectious MS2 phages of the total MS2 particle mass steadily decreased over
time, i.e., in the course of individual breakthrough curves and with sediment depth. The simulated pulse of
rainwater caused a front of low ionic strength water which resulted in pronounced phage remobilization. The
high proportion of infectious MS2 among the detached phages indicated that attachment to the sediment matrix
may substantially conserve virus infectivity. Therefore, the re-mobilization of previously attached viruses owing
to hydrological extremes should be considered in water quality assessment and monitoring schemes.
1. Introduction than 150 years (Kuehn and Mueller, 2000; Nagy-Kovács et al., 2019).
For example, in Germany, the country in Europe with most sites of
Bank filtration of river and lake water has proven to be of great value induced bank filtration, 9%–16% of drinking water is produced in this
for drinking water production by combining treatment and abstraction way (Gillefalk et al., 2018). However, if the efficiency of the natural
at low costs (Ahmed and Marhaba, 2017; Massmann et al., 2004; Krauss subsurface attenuation processes is insufficient or highly variable, bank
and Griebler, 2011; Kuehn and Mueller, 2000). The infiltrating surface filtrate may be contaminated with pathogens, including viruses
water is physically, chemically, and microbiologically altered during (Schijven et al., 2002; Tufenkji et al., 2002).
sediment passage, and this generally improves water quality (Jaramillo, The contamination of raw and drinking water with pathogens is a
2011). At some sites, bank filtration plants have been operating for more widespread source of waterborne disease outbreaks (Yang et al., 2012).
* Corresponding author.
E-mail address: christian.griebler@univie.ac.at (C. Griebler).
https://doi.org/10.1016/j.jconhyd.2022.103960
Received 1 July 2021; Received in revised form 2 January 2022; Accepted 10 January 2022
Available online 13 January 2022
0169-7722/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Previous studies found a range of human enteric viruses at densities of because of its similar size (26 nm) and shape (icosahedral) to enteric
102 to 103 gc (gene copies) L− 1 in surface waters (Lodder et al., 2010; viruses (especially enteroviruses), as well as its similar behavioral
Prevost et al., 2015); these were introduced either by wastewater or via characteristics (persistence, resistance, no replication in aquatic envi
surface run-off from agricultural land where wastewater and manure ronments, decay rate in surface water) that determine its fate in the
were applied (Fong and Lipp, 2005). The reduction of pathogenic viruses subsurface (Boehm et al., 2019; Harvey and Ryan, 2004; Jofre et al.,
during bank filtration is of particular interest because compared to most 2016; Pitol et al., 2018). Two scenarios in the main experiment focused
pathogenic bacteria, many viruses are more persistent in the aquatic on mimicking the infiltration of heavy viral contamination and a sub
environment. It is not rare for viruses to remain infective for more than sequent rainstorm event. Both scenarios are likely to occur frequently in
100 days in sediment, freshwater, and sewage (Fong and Lipp, 2005). fairly densely settled areas where bank filtrate is obtained from rivers
Moreover, owing to their smaller size, viruses may travel longer dis and lakes that carry treated wastewater, e.g., by upstream discharge
tances through the porous subsurface, as underground physical attenu from sewage treatment plants or by stormwater overflow from mixed
ation processes may be less effective than for bacteria (Betancourt et al., sewers. Unlike other studies, we not only looked at the fate of either total
2014; Bition and Harvey, 1992). Viruses have been reported to be or infectious virus particles suspended in sediment pore water but also
transported through subsurface media over distances as far as 1600 m evaluated their combined dynamics over space and time and included
and to a depth of 64 m below the land surface (Bales et al., 1989). For total and infectious MS2 phages attached to the sediment matrix after
these reasons, a comprehensive understanding of the attenuation effi each treatment phase, i.e., after the contamination event and after the
ciency of bank filtration particularly for viruses is important for plan heavy rain event. Moreover, key physical and chemical parameters, i.e.,
ning sustainable and safe drinking water production. This requires electrical conductivity (EC), dissolved oxygen concentration (DO), pH,
addressing the impact of dynamic changes in the hydrological cycle on redox potential, and temperature, and basic microbial parameters, i.e.,
virus transport and retention, including the impact of extended periods adenosine triphosphate concentration (ATP) and total prokaryotic cell
of droughts followed by heavy rain events and floods (Eckert et al., counts (TCC), were monitored in sediment pore water throughout the
2008). entire experiment.
Climate change is predicted to lead to an increase in extreme hy
drological events, including more frequent and extended droughts and 2. Material and methods
heavy rainfall. Understanding the impact of changes in hydrological
conditions on the distribution of pathogens during bank filtration is 2.1. Experimental facility
therefore important for public health (Sprenger et al., 2011). Previous
studies have already concluded that bank filtrate quality may be The study was conducted using a sediment column embedded in the
compromised by viral contamination induced by extreme precipitation facility for the simulation of riverbank and slow sand filtration (Fig. S1)
owing both to shorter travel time in the subsurface and to elevated viral at the German Environment Agency’s field station in Berlin-Marienfelde.
contamination in surface water (Rose et al., 2000; Sprenger et al., 2011). This is fed by water from a storage pond in which groundwater (from
Additionally, Drayna et al. (2010) proposed that the re-mobilization which Fe and Mn were removed through microbial precipitation) was
(desorption) of formerly attached viruses in sediments can be a rele exposed to the atmosphere (including natural rainfall and sunlight) so
vant process in the contamination of groundwater from infiltrating that natural flora and fauna could develop. After a residence time of 2–3
precipitation. Variation in the ionic strength of the sediment pore water weeks this water entered the slow sand filtration basin, mimicking
may exert an important influence on virus attachment-detachment and surface water with a pH of 8 ± 0.2 (29 replicate measurements) and a
(re-) mobilization (Landry et al., 1979). It is therefore important to calculated ionic strength of 13.3 mM. Selected inorganic hydrochemical
understand how heavy rain events causing water with a low ionic parameters of the infiltrating water are summarized in Table 1. Mean
strength to percolate through bank sediments affect virus transport. concentrations of total organic carbon recorded in the months of
Although bank filtration has been studied for many years, the focus November/December of preceding years amounted to 2.0 mg L− 1 (three
was chiefly on the fate of a variety of chemical pollutants and pathogenic replicates). A water-saturated sand column, similar to a lysimeter with a
bacteria. The transport of pathogenic viruses through aquatic sediment diameter of 113 cm (i.e., corresponding to a surface area of 1 m2) and a
has been less well studied, mainly owing to the challenges of virus height of 130 cm (Fig. 1) was embedded in the slow sand infiltration
detection and quantification (Dash et al., 2010; Doussan et al., 1997; basin (dimensions 10 × 10 m) and fed with surrounding pond water. The
Wyn-Jones and Sellwood, 2001). Moreover, studies have so far almost supernatant, i.e., a water column of 10 cm (corresponding to 100 L of
exclusively addressed the infectious viruses (i.e., detection by plaque water) could be connected and disconnected to the surrounding pond
assay or cell lines) suspended and transported in sediment pore water, water. The column was packed with 100 cm of natural coarse-grained
while the transport of those which are no longer infectious has scarcely medium sand (grain size distribution in Table S1) with a porosity of
been addressed. However, including them provides information on virus 41% (d10 = 0.14 mm, d60 = 0.33 mm, uniformity coefficient Cu = 2.36),
inactivation, virus removal by irreversible attachment to sediment or on top of a supporting layer of 30 cm of gravel. As the ratio of the column
decay, and virus retardation by attachment and re-mobilization (Love diameter (dcol) to the effective particle diameter (d10) of the media was
land et al., 1996). Limiting research to the study of infectious virus
particles may further underestimate the number of viruses involved in
aggregation. In contrast, exclusively addressing total viruses, as is Table 1
currently common with molecular tools, misses viral infectivity and Inorganic hydrochemical parameters of the infiltrating pond water in mg L− 1
inactivation (Hassard et al., 2016). Combining and comparing the fate of and mM (mean values of 14–27 analyses conducted over a period of 7 years).
both the total virus load and the fraction which remains infectious is Parameter Concentration ± standard Concentration ± standard
therefore the most comprehensive approach for elucidating the mech deviation (mg L− 1) deviation (mM)
anisms of virus transport during sediment passage. Na+ 50.5 ± 4.6 2.2 ± 0.2
The primary objective of the current study was to elucidate the ef K+ 3.7 ± 0.9 0.09 ± 0.02
ficacy and contribution of individual natural attenuation processes as Ca2+ 123 ± 8 3.1 ± 0.2
Mg2+ 17.3 ± 1.4 0.71 ± 0.06
viruses travel through aquatic sediment, including irreversible attach Fetot 0.018 ± 0.019 0.0003 ± 0.0003
ment, retardation (attachment and re-mobilization), inactivation, and Mntot 2.2 ± 1.6 0.04 ± 0.03
decay. This included, in particular, quantification of the re-mobilization Cl− 89 ± 6 2.5 ± 0.2
of formerly attached viruses caused by rainwater infiltration. To achieve NO3− > 0.1 > 0.002
SO42− 222 ± 12 2.3 ± 0.2
the aims of this study, bacteriophage MS2 was used as a surrogate
2
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
far greater than 50, a potential wall effect can be neglected (Knappett collected before and after the respective treatments at a high temporal
et al., 2008). Sampling ports for sampling along the flowpath were and spatial resolution (i.e., the supernatant, 40 cm, 80 cm, and the
installed at depths of 20 cm, 40 cm, and 80 cm. The system had been in column outlet at 130 cm depth; Fig. 1). Since this pre-experiment was
operation for over 10 years, sufficient for the establishment of a natural conducted two years before the main experiment, there was sufficient
bacterial community (Degenkolb et al., 2018). time for the sediment system to be flushed by MS2-free water.
Pre-experiments included (i) batch tests on the inactivation/decay of In the first of the two experimental phases, a 100 L pulse of pond
MS2 phages in sterile surface water and water-saturated sediment; and water containing a high load of 1.84 × 108 gc mL− 1 MS2 particles, as
(ii) a virus transport experiment in the sediment column with a mod determined by quantitative real-time polymerase chain reaction (qPCR),
erate virus load. and 1.5 × 108 PFU mL− 1 of infectious phage, as determined by the
For the batch tests, MS2 was added to filter-sterilized (0.22 μm) plaque assay, was supplied to the column with a pore water velocity of
Danube River water (pH: 8.19; EC: 401 μS cm− 1; DO: 12.76 mg L− 1, n = 2.1 m d− 1. Nineteen days after the high virus load pulse, experimental
3) or autoclaved water-saturated sand (natural bank sediment; grain size phase 2 was initiated to simulate a heavy rainfall event (corresponding
between 63 and 630 μm; porosity of 36%). Duplicate water samples to 37 mm/h) via infiltration of a pulse of 100 L of deionized water.
contained an initial total MS2 concentration of 6.47 × 108 particles Deionized water was used because its low ionic strength is similar to that
mL− 1. Duplicate sediment samples, each water-saturated, consisted of of natural rainwater. After each of the pulses, the column was recon
17 g dry weight sediment with approximately 3.6 mL of pore water and nected to the surrounding pond water just before the pulse solution was
received an initial total MS2 concentration of 1.52 × 108 particles mL− 1. completely infiltrated. Sampling depths for pore water included the
These samples were incubated at 7 ◦ C in the dark for 19–23 days. surface (i.e., the supernatant), 20 cm (withdrawal rate: 1.3 L/h), 40 cm
The impact of a moderate virus load on MS2 transport was assessed 2 (withdrawal rate: 1.4 L/h), 80 cm (withdrawal rate: 1.4 L/h), and the
years before the main experiment, also in November and with the same column outlet (at 130 cm depth, withdrawal rate of 33.4 L/h; Fig. 1).
column system. A pulse of pond water (200L) with a moderate virus From the intermediate ports and at the column outlet water was with
load, i.e., a final concentration of 1.9 × 106 plaque forming units (PFU) drawn continuously. Samples for infectious MS2, prokaryotic cell counts
mL− 1 of infectious MS2 (total amount of infectious MS2 of 3.8 × 1011 and microbial activity were collected every hour in the first 7 h and 12 h
PFU), was supplied with flow rate adjusted to a filter velocity of 1.1 m after the dose of MS2 pulse and the pulse of deionized water, respec
d− 1 (corresponding to a pore water velocity of 2.2 m d− 1). After the tively. Subsequently, the sampling intervals increased sequentially with
pulse, the column was re-connected to the surrounding pond water just time (see Tables S2 and S3 for details). At the mesocosm outlet in
before the pulse solution was completely infiltrated. Water samples were experimental phase 1, water samples for biological monitoring were first
3
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
collected 23 h after the supply of MS2 particles. Samples for the analyis 2.5. ATP and TCC analyses
of water chemistry were taken once a day during the first 10 days after
MS2 dosing and for the entire experimental phase 2 after simulated To study the relationship between virus reduction and the amount
rainfall. Additional water chemistry samples were taken at 15 and 16 and activity of autochthonous microorganisms, microbial activity and
days after MS2 donation. prokaryotic cell counts were monitored in both water and sediment
Sediment samples were collected after the high virus load pulse samples throughout the experiment. The microbial activity in water and
(experimental phase 1) and the simulated rain event (experimental sediment samples was estimated using ATP concentration measure
phase 2) using a double-wall sediment corer (high-density polyethylene, ments using the BacTiter-Glo Microbial Cell Viability Assay kit (Prom
10 cm outer diameter, 4 cm inner diameter, and 100 cm height), which ega) following the protocol described in Hammes et al. (2010), but with
was pushed into the sediment to a depth of 70 cm at the end of each minor modifications. Water samples and the BacTiter-Glo reagent, pre
experimental phase (day 15 and day 23 after the MS2 pulse). The inner pared in accordance with the manufacturer’s instructions, were pre-
core was withdrawn carrying the sediment, whereas the outer tube warmed separately at 38 ◦ C for at least 2 min before 1 mL of sample
remained in place for the remainder of the experiment to maintain un was mixed with 50 μL of reagent. The mixture was incubated at 38 ◦ C for
disturbed hydrological conditions. The cored sediment samples were 1 min and luminescence measured on a GloMax 20/20 Luminometer
divided into portions of 5 cm depth and transferred into falcon tubes (50 (Promega). Concentrations were determined by comparison with
mL) for the subsequent quantification of total and infectious MS2 par external ATP standards dissolved in ATP-free water (Fisher Scientific)
ticles. All experiments were conducted at the end of the calendar year using ATP-free water as blank. To correct for the contribution of
(November/December). extracellular ATP in the samples the measurements were performed on
an untreated sample fraction representing the total ATP concentration as
2.4. Cultivation and quantification of MS2 well as a fraction of the same sample that was filtered (0.1 μm syringe
filter, Merck Millipore) to remove microbial cells. The concentration of
Bacteriophage MS2 (DSMZ 13767) and MS2 host strain Escherichia intracellular ATP was then calculated by subtracting the extracellular
coli (DSMZ 5695) were obtained from the DSMZ (German Collection of concentration from the total. All measurements were conducted in
Microorganisms and Cell Cultures GmbH). Isolation and purification of triplicate.
MS2 phages after propagation with a co-incubation of E. coli 5695 were The total ATP in sediment samples was determined as follows: 50 μL
performed following the protocol described in Feichtmayer (2019) to of 0.1 μm filtered Milli-Q water was added to 200 μL of homogenized
produce a high-density phage suspension in buffer solution. sediment. Next, 100 μL preheated (38 ◦ C) BacTiter-Glo reagent was
For the pre-experiment with the moderate virus load, infectious added and the sample was mixed using a Thermomixer (1400 rpm) at
phages were quantified with the plaque assay in accordance with DIN 38 ◦ C for 2.5 min. Subsequently, 900 μL of 0.1 μm filtered Milli-Q water
EN ISO 10705–1:2001. For the pre-experiment batches and the main was added and the sample was briefly mixed using a vortex mixer and
experiments, this was adapted to follow the double-agar overlay method centrifuged. Finally, triplicate measurements of the supernatant were
of Kropinski et al. (2009). In brief, a solid agar base (10 mL of 1.5% agar, performed using a similar protocol to that for the water samples (see
w/v) was overlaid with a mixture of soft agar (4 mL of 0.75% agar, w/v) above). The extracellular ATP in sediment samples was quantified using
and 1 mL of a well-mixed solution consisting of the (diluted) sample and a similar setup, but without the addition of 100 μL preheated (38 ◦ C)
host bacteria in the logarithmic growth phase (1:1, v/v). Serial dilutions BacTiter-Glo reagent to lyse the cells.
of the mixture were incubated at 37 ◦ C overnight and PFUs were counted TCC was quantified using flow cytometry (FC 500 CYTOMICS,
in duplicate. The titer of infectious phages in the sediment samples was Beckman Coulter) as described in Fillinger et al. (2019) for pore water
estimated after the elution of 0.5 mL of sediment with 1 mL of deionized samples and as detailed in Bayer et al. (2016) for sediment samples.
water and vigorous shaking on a vortex mixer (3 times, each for 1 min,
with 30 s between each mixing step). The limit of detection (LOD) for the 2.6. Physicochemical conditions
plaque assays was 0.5 PFU mL− 1.
Triplicate quantification of the total number of phages was per Pore water temperature and pH were measured on-site using a
formed using qPCR analysis of phage RNA transcribed into cDNA. Before temperature data logger (Voltcraft) and microprocessor pH meter,
phage RNA extraction, all water samples were concentrated to a volume respectively. EC, redox potential, and DO were measured on-site using
of 1 mL using Amicon® Ultra-15 Centrifugal Filter Units (Merck). Phage field sensors (WTW).
RNA was then extracted using the AllPrep PowerViral DNA/RNA Kit
(Qiagen) and cDNA synthesis was performed using the Maxima First 2.7. Data analysis
Strand cDNA Synthesis Kit (ThermoFisher). The isolation of phage RNA
and the reverse transcription was conducted following the manufac The recovery of total and infectious MS2 during experimental phase
turer’s instruction. The qPCR targeted a 315 bp fragment using the MS2- 1 and phase 2 was calculated to a depth of 80 cm. The deeper zone of the
specific primer sequences MS2-2717Fw (5′ -CTG-GGC-AAT-AGT-CAA-A- mesocosm including the gravel layer was not included, thus avoiding
3′ ) and MS2-3031Rv (5′ -CGT-GGA-TCT-GAC-ATA-C-3′ ),and using the inconsistency due to its large sediment heterogeneity and the risk of
Brilliant III UltraFast SYBR Green QPCR Master Mix (Agilent). Then, 2 having missed the MS2 peak concentration at the mesocosm’s outlet. For
μL of cDNA were quantified in triplicate; each 25 μL reaction contained samples with a total MS2 concentrations below the detection limit, the
12.5 μL of qPCR master mix, 0.375 μL of reference dye (1:500 v/v infectious MS2 concentration was used instead for calculation. The
dilution in RNase-free water) and 0.5 μL of each primer (10 μM, Euro number of total and infectious sediment-attached MS2 was calculated as
fins). The qPCR analyses were performed in a Stratagene MX3000P the product of the total or infectious MS2 concentration detected at
qPCR cycler (Agilent, R2 > 0.99) and a Light Cycler 480 SW 1.5 (Roche, various depths of the sediment cores and the volume of the corre
R2 > 0.98) following a protocol modified from Dreier et al. (2005). sponding column layer. The phage particle concentrations collected at
Initial denaturation (95 ◦ C, 10 min) was followed by 40 cycles of depths between 65 cm and 70 cm were assumed to be representative of
denaturation (95 ◦ C, 15 s), annealing (50 ◦ C, 30 s) and elongation (72 ◦ C, the depth interval of 65–80 cm.
30 s). Afterwards, the melting curve was recorded between 55 ◦ C and To test the relationships between different biological and physico-
95 ◦ C to verify the specificity of the amplification products. Owing to the chemical variables, a Spearman rank correlation analysis was per
different sample volumes available for concentration, there was no formed with the Hmisc package in R (Harrll et al., 2018) and applied to
single LOD; instead, the LOD range was determined as 5.8 × 102–5.6 × all available variables collected for pore water and sediments encom
103 gc mL− 1. passing data from both experimental phases (Tables S4 and S5).
4
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Student’s t-test (“t.test” in R) and the Wilcoxon test (“wilcox.test” in R) 3.1.2. Transport of a moderate MS2 pulse through the sediment column
were used to analyze the significance of difference between samples In the pre-experiment that lasted for 50 h, we observed a significant
using t with a significance level of 0.05. decrease in the peak concentrations of the infectious virus breakthrough
curves with depth, from 1.9 × 106 PFU mL− 1 at the surface to 2.1 × 103
3. Results and discussion PFU mL− 1 at a depth of 80 cm (Fig. 3). After 80 cm of sediment passage,
the attenuation or removal of infectious MS2 amounted to approxi
3.1. Pre-experiments mately 3.0 log10. No significant difference in the concentration of in
fectious MS2 was detected at the depth of 80 cm and the outlet at the
3.1.1. Inactivation and decay of MS2 in batch incubations depth of 130 cm (t-test, t = 0.97, df = 9.7, P = 0.35).
Following the MS2 inoculum of 6.5 × 108 gc mL− 1 in sterilized
surface water for 23 days in batch experiments (Fig. 2) showed the 3.2. Main experiments
concentration of total MS2 particles, as measured by qPCR, to remain
constant, i.e., without any significant reduction (Wilcoxon test, n = 12, P 3.2.1. Dynamics of physico-chemical water quality
= 0.5; Fig. 2). This indicates that no MS2 decay occurred during incu During the entire course of the main experiment, the pond water
bation. However, the fraction of infectious MS2 showed some decline at infiltrating into the bank sediment differed significantly in temperature
19 days after inoculation, i.e. from 5.6 × 108 to 9.8 × 107 PFU mL− 1, (Wilcoxon test, n = 34, P < 0.05), DO concentration (t-test, t = − 3.6, df =
accounting for 15% of the total MS2 and corresponding to a 0.8-log 31, P < 0.01) and pH (t-test, t = − 4.0, df = 24.6, P < 0.001) from the
reduction of the initial infectious MS2 concentration. This decrease in water collected at the outlet after 130 cm of sediment passage.
infectious MS2 during incubation is assumed to be due to inactivation Compared to the surface water temperature (meansurface = 5.9 ◦ C ±
and implies an inactivation rate of MS2 in the sterilized surface water at 2.4 ◦ C) water temperature at the column outflow was higher and more
7 ◦ C was 0.045 log10 day− 1 (Fig. 2), similar to the inactivation rate of constant (meanoutlet = 7.8 ◦ C ± 2.0 ◦ C; Fig. 4). This difference was
MS2 (0.043 log10 d− 1) determined at 4 ◦ C in raw groundwater by partially due to the transition from autumn to winter and the delayed
Ogorzaly et al. (2010). response of the sediment temperature to the decrease in air temperature
Similar to the experiments only in sterilized water, no reduction in during the course of the experiment. The mean DO concentration
total MS2 was observed after inoculation into water-saturated sediments decreased from 13.9 ± 1.8 mg L− 1 (meansurface) to 11.5 ± 2.2 mg L− 1
(Wilcoxon test, n = 12, P = 1). The initial concentration of attached total (meanoutlet) during the passage of pore water through the column, which
MS2 in the sterile water-saturated sediment mixture was 1.5 × 108 gc was likely due to higher oxygen solubility at lower temperature and
mL− 1. Infectious MS2 attached to the sediment with an initial concen microbial processes, e.g., aerobic respiration (Diem et al., 2013). A peak
tration of 8.9 × 107 PFU mL− 1 declined at an inactivation rate of 0.035 value of 17.2 mg L− 1 DO in the infiltrating surface water on day 9 cor
log10 d− 1 (Fig. 2). The assumed inactivation rate of attached MS2 phages responded to a sudden drop in temperature (Fig. 4). The pH decreased
was lower than that of the suspended MS2, indicating that although from a mean value at the surface of 8.1 ± 0.1 to a mean value at the
attachment is an important process for attenuating MS2 during the outlet of 7.8 ± 0.3 (Fig. 4). Heterotrophic biological processes result in
subsurface passage, phage activity is preserved. Prolonged survival of DO consumption and oxidation of organic matter with the production of
the virus by attachment to sediment was also observed by Melnick et al. CO2 and/or organic acids which generally reduces pH (Kuehn and
(1980) and Schijven and Hassanizadeh (2000). Mueller, 2000). During sampling, the mean redox potential at the sur
face (350 ± 71 mV) was not significantly different from that at the
column outlet (347 ± 67 mV; Wilcoxon test, n = 32, P > 0.5). As redox
Fig. 2. The variation of total and infectious bacteriophage MS2 in sterilized surface water or sterilized water-saturated sediment (the abundance of total MS2 at time
0 in water or in sediment was recorded as N0 and the abundance of total or infectious MS2 detected at the corresponding time point in water or in sediment was
defined as Nx).
5
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Fig. 3. Breakthrough of infectious MS2 at different depths of the column, as determined from the transport pre-experiment.
conditions remained oxic throughout the entire experiment effects on moderate virus load applied in the pre-experiment, which was run under
the survival of the virus by a redox shift are likely to have been negli the same conditions as the main experiment, a 3.0 log unit removal of
gible (Feichtmayer et al., 2017). EC remained constant during the infectious MS2 in pore water was found, exceeding the removal in the
experiment (meansurface = 878 ± 223 μS cm− 1, meanoutlet = 879 ± 192 main experiment.
μS cm− 1) and was independent of sediment depth, with one exception, The lower removal efficiency in the main experiment can be
namely the supply of deionized water at the beginning of phase 2 which explained by a conceptual model for colloid transport derived from
caused a temporary shift throughout the sediment column. The lowest column studies and micromodel observations, as suggested by Bradford
EC caused by deionized water was 15.1 μS cm− 1 at the surface; a higher et al. (2009). At high initial input concentrations, a lower relative mass
value of 135.8 μS cm− 1 was detected at the outlet. The simulated heavy transfer to the solid phase occurs because weakly associated colloids are
rainfall, as intended, led to a reduction in the ionic strength of the knocked off the solids’ surface owing to more frequent collisions.
sediment pore water. Overall, the differences in DO and pH indicated the Moreover, Bradford et al. (2009) found that only a small fraction of the
activity of biological processes during bank filtration (Kuehn and porous media play an active role in colloid retention, the reason being
Mueller, 2000; Diem et al., 2013). that colloid retention is highest near grain–grain contact where low-
velocity areas form. Furthermore, another possible explanation is the
3.2.2. Experimental phase 1 – virus contamination enhanced aggregation caused by the higher virus concentration, leading
to less interaction of the individual viruses with the grain surfaces.
3.2.2.1. Fate of MS2 during transport through the bank sediment. The Indeed, all virus particles within one aggregate must be inactivated
peak concentrations of total MS2 in the breakthrough curves decreased before this aggregate group will not appear as active in the plaque assay
with increasing depth and distance to the surface (Fig. 5), i.e., from 1.84 (Schijven and Hassanizadeh, 2000). The high virus load we applied re
× 108 gc mL− 1 in the surface water to 3.2 × 107 gc mL− 1 at 80 cm. Thus, flects a worst-case scenario with a high share of sewage in river flow, as
the peak of total virus concentration was reduced by 0.8 log units raw sewage is reported to contain between 109 and 1012 infectious co
through 80 cm of sediment passage. A total amount of 1.75 × 1013 MS2 liphages per L (Lodder and de Roda Husman, 2005). Applying a high
particles was estimated to be transported through the 80 cm sediment virus load also enabled the reliable detection of total and infectious MS2
column (Table 2), i.e., at the end of experimental phase 1, only 5% of the particles over time and space.
total MS2 particles remained in the column. At 219 h after the inocu 19 days after the major MS2 pulse 0.1% of MS2 particles that were
lation with MS2, 103 total MS2 particles mL− 1 were detected in the transported through the 80 cm sediment column were still infective.
surface water, corresponding to a log decrease rate of 0.0185 h− 1. We Compared to the infectious MS2 fraction of the sterilized water in batch
attribute these remaining MS2 particles to kinetic phage attachment to experiments, a much higher proportion of infectious MS2 particles lost
and detachment from the top sediment, taking into account the mixing their infectivity during the sediment passage in the main experiment.
and non-uniform infiltration of the pond water (Frohnert et al., 2014; We assumed that the higher fraction of inactivated MS2 phages in the
Schijven and Hassanizadeh, 2000). column pore water relative to the sterile batch tests was mainly attrib
As with the concentration of total MS2 phage particles, the peak utable to the interplay of pore water flow and attachment, as well as to
concentration of infectious MS2 in pore water decreased with increasing autochthonous microorganisms and their activity. For example, Hurst
distance from the surface, i.e., from 1.5 × 108 PFU mL− 1 in the surface et al. (1980) and Schijven et al. (2002) found that the presence of aer
water to 2.9 × 106 PFU mL− 1 measured at depth of 80 cm (Fig. 5) cor obic microorganisms, especially sediment-attached microorganisms,
responding to a removal of approximately 1.7 log units. With a reduction could affect virus activity and may promote virus decay. Yates et al.
in peak concentration of 0.8 log units for total MS2 particles and 1.7 log (1990) provided an early review of the mechanisms through which
units for infectious MS2 particles, the overall reduction of virus in our microorganisms may cause the inactivation of viruses. For example,
main experiment was in the range of that reported in other studies using certain bacteria release proteolytic enzymes that destroy the protein
MS2 as a model agent, i.e., a 0.1–1.9 log unit reduction over a filtration coating of virus particles (Cliver and Herrmann, 1972). Deng et al.
distance of 0.4–1.5 m as reviewed by Bauer et al. (2011). With the (2014) indicated that the number of infectious MS2 phages may be
6
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Fig. 4. Physico-chemical conditions in the surface water and column outlet over the course of the main virus transport experiment. T: temperature; DO: dissolved
oxygen; EC: electrical conductivity.
reduced by grazing protozoa. More recently, Feichtmayer et al. (2017) layer experienced significant detachment owing to the ongoing infil
provided a comprehensive review of the antagonistic microbial pro tration of MS2-free surface water whereas sediments below 15 cm
cesses inactivating and/or eliminating viruses in aquatic ecosystems. experienced a MS2 saturation with a balance between the attachment
Importantly, the average percentage of infectious MS2 in pore water and detachment of MS2 particles.
decreased with increasing distance from the point of infiltration (Fig. 5). The distribution of infectious MS2 particles followed a slightly
Moreover, the proportion of infectious MS2, 0.4%–79%, also changed different trend, with a stable concentration in the top 45 cm and a
over time in the breakthrough water. The highest proportion of infec moderate decrease in concentration and proportion between 50 cm and
tious MS2 particles was found near or at the peak of the individual 70 cm. The percentage of infectious MS2 phages was thus high at the top
breakthrough curves, independently of sediment depth (i.e., 79% at 0 (15% at 5 cm), decreased with depth, and became lowest at the bottom
cm, 59% at 20 cm, 47% at 40 cm, and 30% at 80 cm). of the column (0.8% at 70 cm). Both the attached and suspended in
fectious MS2 decreased with depth, suggesting MS2 experienced more
3.2.2.2. Fate of MS2 attached to the sediment in phase 1 core samples. The inactivation than decay during the sediment passage. This result also
concentration of sediment-attached total MS2 particles was distributed suggests that the retention of infectious phage particles is higher than
rather evenly with depth in the sediment column at the end of phase 1 (i. that for the total MS2 particles because inactivation reduces particle size
e., 15 days after the inoculation with MS2) (Fig. 6). Starting with a total and alters the physico-chemical characteristics of the phage surface,
concentration of attached MS2 phages of 1.5 × 105 gc mL− 1 in the up facilitating transport (Ghanem et al., 2018).
permost sediment layer, the concentration increased to 4.8 × 105 gc
mL− 1 to a depth of 15 cm and then remained constant, independently of 3.2.2.3. Microbial biomass and activity in phase 1. The abundance of
further sediment depth. This pattern indicated that the top sediment prokaryotic cells and microbial activity in the pore water at individual
7
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Fig. 5. Breakthrough of total and infectious MS2 particles at various sediment depths in pore water during experimental phases 1 and 2.
sediment depths remained stable during phase 1 (Fig. 7). This indicates studies (as reviewed by Hassard et al., 2016), while the stable microbial
that the inoculation of surface water with MS2 did not significantly alter activity throughout the column below 10 cm the presence of a well-
other microbial patterns in sediment pore water. As expected, the con established depth-independent active bacterial community (Degenkolb
centration of prokaryotic cells in the pore water decreased with sedi et al., 2018). The activity of indigenous microbes is assumed to have a
ment depth, from 8.12 × 105 cells mL− 1 in the surface water to 5.39 × negative impact on the survival of allochthonous viruses and bacteria,
104 cells mL− 1 at 100 cm. A similar pattern was observed for pore water particularly those that do not have a host in the environment, such as
microbial activity (Fig. 7). Prokaryotic cell counts and microbial activity human pathogenic viruses (Sobsey et al., 1980; Feichtmayer et al.,
of the attached microbial communities from core sampling were also 2017).
highest in the top-most sediments (Fig. 8). Decreasing prokaryotic cell
numbers with increasing sediment depth were consistent with previous
8
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Table 2 similar to that of the total MS2 phages (Fig. 5). The average percentage
Estimate of total and infectious MS2 during the main experiment down to a of infectious MS2 in pore water samples at different depths was in the
column depth of 80 cm (Phase 1: dose of MS2; Phase 2: pulse of deionized range of 13%–40% and hence significantly higher than in the first
water). experimental phase (where the range was 7.6%–21%). Notably, in
Experimental Samples Total Distribution of Proportion of experimental phase 2, 3.3 × 1010 MS2 particles in total were transported
phase MS2 inoculum infectious to through the 80 cm sediment column after the application of deionized
(gc) total MS2
water, and a significant portion (on average 69%) were still infectious
Inoculum
1.84 ×
100% 79%
(Table 2). This suggested that attachment somehow conserves infec
1013 tivity of phage particles better than transport in suspension.
Sediment- 3.3 ×
Phase 1
attached 1011
1.8% 4% Similar findings were obtained in the batch experiments (Fig. 2) and
Flow through 1.75 × in earlier studies (Yates et al., 1987; Melnick et al., 1980; Schijven and
95% 6.7%
80 cm 1013 Hassanizadeh, 2000). Gerba (1984) proposed that the mechanisms
Retained–MS2
3.3 ×
1.8% 4% protecting viruses attached to sediment might include shielding from
1011
proteolytic enzymes or other substances that inactivate and degrade
Sediment- 4.2 ×
Phase 2
attached 1011
2.4% 2.9% viruses. The attachment of virus particles to the soil and sediment sur
Flow through 3.3 × face is mediated by a combination of electrostatic and hydrophobic in
0.2% 69%
80 cm 1010 teractions (Gerba, 1984). A decrease in pore water salt concentration is
known to increase the surface potential of the attached viruses owing to
the expansion of the electrostatic double layer surrounding the virus
3.2.3. Experimental phase 2 – Phage re-mobilization through a simulated particles and soil particle surfaces, causing virus detachment (Quanrud
heavy rainfall event et al., 2003). A rapid shift to low ionic strength conditions has a strong
As described in Section 2.2, the simulated heavy rainfall event was desorption effect on phages and other particles from sediment surfaces
initiated 19 days after the supply of a high load of MS2. The infiltration (Yates et al., 1987; Gerba, 1999). Deionized water, as shown in this
of 100 L of deionized water was equivalent to a recharge of precipitation experiment, can cause a quantitative re-mobilization of previously
of 37 mm/h, which represents an extreme rainfall for the Berlin area. immobilized virus and thus a second pulse of contamination (Guber
The pulse of deionized water caused a significant decrease in EC, i.e., a et al., 2006; Landry et al., 1979). Although the reactive front of the low
reactive front of low ionic strength water migrating down through the ionic strength water pulse initiated the desorption of part of the attached
sediment column (Fig. 4). After 1 day, the EC values were fully recov phages, the phage peak in the column outflow was broader. We assumed
ered, not only in the surface water but also at the outlet of the column. that some virus particles underwent re-attachment further down the
sediment column, as observed by Wellings et al. (1975).
3.2.3.1. Re-mobilization of attached MS2 phages. Although the number
of total MS2 particles had dropped below or close to the detection limit 3.2.3.2. Phages remaining attached to sediment after simulated heavy
in sediment pore water immediately before the simulated rainfall event, rain. A significant number of total MS2 particles (around 1011 gc)
the pulse of deionized water caused a new moving front of MS2 phages remained attached to the sediment after both phases of the experiment
re-mobilized from the sediment migrating through the sediment column. (Table 2). After the artificial rain event, the abundance of sediment-
In contrast to experimental phase 1, the peak concentrations of total attached total MS2 particles was moderately lower at the top of the
MS2 particles in pore water increased with depth, reaching concentra sediment column than in experimental phase 1, down to a depth of 20
tions as high as 105 gc mL− 1 (Fig. 5). cm, but below that it remained constant and similar to the numbers
The fraction of infectious MS2 phages followed a distribution pattern found at the end of experimental phase 1 (Figs. 6 and 8). Indeed, the
Fig. 6. Distribution of total and infectious sediment-attached MS2 particles with depth (a: sediment samples taken from core samples after 15 days, at the end of
experimental phase 1; b: sediment samples taken from core samples 23 days after the application of deionized water, at the end of experimental phase 2; the grey
solid line models the proportion of infectious MS2 collected from the sediment core against the sediment depth).
9
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Fig. 7. Spatio-temporal distribution of bacteriophage MS2 (total and infectious), microbial activity, and prokaryotic cell concentration in sediment pore water of the
column for both experimental phases (virus pulse and simulated rain event) and in various depths of pore water.
number of sediment-bound infectious MS2 phages after the simulated 3.2.3.3. Effects of rain on the distribution of bacteria and microbial
rainfall event was not significantly different from that in experimental activity. The artificial heavy rainfall not only mobilized sediment-
phase 1 (t-test, t = 0.80, df = 26, P = 0.43). Infectious MS2 particles attached MS2 particles but also caused prokaryotic cells to detach
attached to the sediment matrix were distributed more evenly with (Fig. 7). This result supported the findings of earlier studies, which re
depth than total phage particles, ranging from 7.6 × 103 to 2.5 × 104 ported that the contamination of groundwater wells was promoted by
PFU mL− 1. Modelling the proportion of infectious MS2 collected from the detachment of bacteria from surrounding sediments in periods of
the sediment core against the sediment depth showed that the decrease heavy rainfall (Bition and Harvey, 1992; Zyman and Sorber, 1988).
in the proportion of infectious MS2 phages with depth was less steep Interestingly, in our study, no apparent changes were observed in mi
after the artificial rainfall than in experimental phase 1 (Fig. 6), indi crobial activity down to a depth of 40 cm, indicating that cells with no or
cating that infectious MS2 particles were preferentially re-mobilized. low activity were predominantly dislodged from the sediment following
rainfall. However, at deeper zones of the sediment column, the microbial
10
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
Fig. 8. Vertical distribution of sediment-attached bacteriophage MS2 (total: a and infectious: b), prokaryotic cell concentration (c), and microbial activity (d) in the
enclosure at the end of experimental phase 1 (virus pulse) and phase 2 (simulated rain event).
activity appeared to follow the dynamics of the bacterial numbers retarded by attachment, and for the peak concentrations of total MS2
(Fig. 7). The peak concentration of prokaryotic cells in the sediment pore and infectious MS2 phage particles the reductions achieved amounted
water following the simulated rainfall was more than 10 times higher only to of 0.8 and 1.7 log units, respectively, for a sediment passage of
than the pore water concentrations detected in experimental phase 1. 80 cm. In contrast, for the lower initial MS2 concentration in the supply
However, at 91 h after the simulated rainfall, lower concentrations of water, the sediment removal efficiency was much higher, amounting to
prokaryotic cells were detected in pore water compared with phase 1, 3.0 log10. If we only consider infectious virus particles to be a health
but the overall pattern was similar: fewer prokaryotic cells were present hazard, our results are more promising. We showed that with travel
further down the column. distance through the sediment and water residence time, the proportion
As in experimental phase 1, the sediment-attached prokaryotic cell of non-infectious virus particles steadily increased. After a sediment
concentration was highest at the top of the column (5 cm sediment passage of 80 cm, 1.2 log10 of the total MS2 particles were inactivated
depth) and 10 times lower at a depth of 70 cm. After the rainfall, and could no longer be detected by the plaque assay. Indeed, compared
although the numbers of sediment-attached microorganisms were lower with MS2, other viruses, particularly many human pathogenic viruses,
mainly in the top sediments, the overall abundance of bacteria that tend to be attached more efficiently, especially within the first centi
remained attached to the sediment was lower in experimental phase 2 meters of sediment passage (Schijven and Hassanizadeh, 2000). Thus,
than in experimental phase 1. Although a breakthrough of prokaryotic the breakthrough of viruses of health concern is likely to be lower than
cells migrating through the sediment column was observed following for the bacteriophage MS2.
the simulated rain event, the total number of sediment-attached cells did The attachment of a virus particle does not imply that it will be
not change significantly (Wilcoxon test, n = 28, P = 0.38), as shown by inactivated and decay. Our data showed that re-mobilized virus particles
the cell count, which was two orders of magnitude lower in pore water were characterized by a higher proportion of infectious particles (0.3%–
than attached to the sediment matrix. Similar patterns were found for 89.4%) than that found in the suspended virus fraction during the initial
microbial activity (Fig. 8). The simulated rainfall significantly increased contamination, i.e., experimental phase 1 (Fig. 5). The inactivation of
the activity of the sediment-attached microbial consortia in the top viruses during attachment is reported to be virus-specific (Frohnert
sediment. Apart from this exception, most of the sediment-bound mi et al., 2014). Yates et al. (1987) reviewed further factors that could
crobial activity was lower in phase 2, with a significant difference be affect the transport and inactivation of viruses during soil and sediment
tween phase 1 and phase 2 revealed (Wilcoxon test, n = 28, P = 0.003). passage, including temperature, pore water pH, sediment organic matter
content, autochthonous microbial activity, and ionic strength. While
3.3. A comprehensive perspective on the main experiment earlier studies commonly agree that viruses remain infective and intact
for longer at low temperatures (Yates et al., 1987; Feng et al., 2003) and
Although it may be assumed that the virus concentrations applied in inactivation rates increase with an increase in temperature (John and
the experiment rarely occur in natural environments, in face of the Rose, 2005; Krauss and Griebler, 2011), during our experiment, tem
rather small sample volumes available and the LOD of the methods perature showed a positive correlation to both infectious MS2 and total
applied for the quantification of total and infectious MS2, the seeding of MS2 phage concentrations in pore water (Table S4). However, this
high phage numbers allowed the spatio-temporal resolution of several correlation is caused by a decrease of virus particles over time and a
processes involved in the retardation and attenuation of virus in bank simultaneous seasonal decrease of ambient temperature in the sediment
sediment infiltration, including attachment, detachment, and inactiva mesocosm. Moreover, water temperatures detected during our experi
tion. Our study revealed that high peak concentration of viruses in ment remained <10 ◦ C, a temperature range characterized by a very low
recharge water can lead to exceedance of the retention capacity of bank inactivation rate of MS2 (Schijven and Hassanizadeh, 2000). EC was
sediments, presumably caused by the displacing interactions between negatively correlated with infectious MS2 phage concentrations. This
attached MS2 and suspended MS2 in the water phase. In our main supports the view that compared with total MS2, the desorption of in
experiment, only a minor fraction of the supplied virus particles was fectious MS2 is more sensitive to ionic strength. For the sediment
11
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
samples tested, prokaryotic cell abundance and microbial activity were Appendix A. Supplementary data
significantly negatively correlated with the concentration of total MS2
particles rather than the concentration of infectious MS2 particles Supplementary data to this article can be found online at https://doi.
(Table S5). This indicates an antagonistic relationship between the org/10.1016/j.jconhyd.2022.103960.
autochthonous microbial communities and the allochthonous phages.
Microbial mediated antagonistic processes that lead to a reduction of References
allochthonous viruses and microbes have been reported (Feichtmayer
et al., 2017). A further possibility is competition between total viruses Ahmed, A.K.A., Marhaba, T.F., 2017. Review on river bank filtration as an in situ water
treatment process. Clean Techn. Environ. Policy 19, 349–359. https://doi.org/
and autochthonous bacteria for attachment sites in the aquifer medium. 10.1007/s10098-016-1266-0.
The simulated recharge of rainwater clearly showed that a significant Bales, R.C., Gerba, C.P., Grondin, G.H., Jensen, S.L., 1989. Bacteriophage transport in
proportion of the initially attached viruses was re-mobilized, resulting in Sandy soil and fractured tuff. Appl. Environ. Microbiol. 55, 2061–2067. https://doi.
org/10.1128/aem.55.8.2061-2067.1989.
a second pulse of virus-contaminated pore water moving through the Bauer, R., Dizer, H., Graeber, I., Rosenwinkel, K.H., López-Pila, J.M., 2011. Removal of
bank sediment and thus demonstrating that extreme hydrological events bacterial fecal indicators, coliphages and enteric adenoviruses from waters with high
indeed have the potential to cause a temporary impact on water quality. fecal pollution by slow sand filtration. Water Res. 45, 439–452. https://doi.org/
10.1016/j.watres.2010.08.047.
As the low EC of the simulated rainwater exceeded 130 cm of sediment Bayer, A., Drexel, R., Weber, N., Griebler, C., 2016. Quantification of aquatic sediment
passage (Fig. 4), the re-mobilization of attached viruses can be expected prokaryotes-a multiple-steps optimization testing sands from pristine and
to have an impact over a distance of several meters, depending on the contaminated aquifers. Limnologica 56, 6–13. https://doi.org/10.1016/j.
limno.2015.11.003.
time and intensity of the rain event, water flow velocity, and pore-water
Betancourt, W.Q., Kitajima, M., Wing, A.D., Regnery, J., Drewes, J.E., Pepper, I.L.,
ionic strength (Quanrud et al., 2003). Gerba, C.P., 2014. Assessment of virus removal by managed aquifer recharge at three
full-scale operations. J. Environ. Sci. Health - Part A 49, 1685–1692. https://doi.org/
4. Conclusions 10.1080/10934529.2014.951233.
Bition, G., Harvey, R.W., 1992. Transport of pathogens through soils and aquifers.
Environ. Microbiol. 19, 103–123.
Attachment, inactivation, and decay are important processes in the Boehm, A.B., Silverman, A.I., Schriewer, A., Goodwin, K., 2019. Systematic review and
attenuation of viruses, as exemplified with the bacteriophage MS2, Meta-analysis of decay rates of waterborne mammalian viruses and Coliphages in
surface waters. Water Res. 164, 114898. https://doi.org/10.1016/j.
during bank sediment passage. Even when challenged with an extreme watres.2019.114898.
virus concentration of 108 particles mL− 1, at the upper end of loads to be Bradford, S.A., Kim, H.N., Haznedaroglu, B.Z., Torkzaban, S., Walker, S.L., 2009.
expected in rivers, bank filtration can be efficient in reducing a viral Coupled factors influencing concentration-dependent colloid transport and retention
in saturated porous media. Environ. Sci. Technol. 43, 6996–7002. https://doi.org/
contamination by 1–2 log units after only a short transport distance of 10.1021/es900840d.
less than one meter. At lower concentrations of viruses infiltrating bank Cliver, D.O., Herrmann, J.E., 1972. Proteolytic and microbial inactivation of
sediments and a lower water flow velocity (i.e., high water residence enteroviruses. Water Res. 6, 797–805. https://doi.org/10.1016/0043-1354(72)
90032-2.
time), and/or higher temperatures, the natural attenuation efficiency is Dash, R.R., Prakash, E.B., Kumar, P., Mehrotra, I., Sandhu, C., Grischek, T., 2010. River
likely to be much higher, in the range of 3–4 log units. Equally impor bank filtration in Haridwar, India: removal of turbidity, organics and bacteria.
tant, only a small fraction of virus particles remained infectious after the Hydrogeol. J. 18, 973–983. https://doi.org/10.1007/s10040-010-0574-4.
Degenkolb, L., Metreveli, G., Philippe, A., Brandt, A., Leopold, K., Zehlike, L., Vogel, H.J.,
sediment passage, indicating that virus inactivation was a quantitatively
Schaumann, G.E., Baumann, T., Kaupenjohann, M., Lang, F., 2018. Retention and
important process in bank filtration. Extreme hydrological events, such remobilization mechanisms of environmentally aged silver nanoparticles in an
as heavy rainfall, will cause the re-mobilization of previously attached artificial riverbank filtration system. Sci. Total Environ. 645, 192–204. https://doi.
virus particles and autochthonous microorganisms from the sediment org/10.1016/j.scitotenv.2018.07.079.
Deng, L.I., Krauss, S., Feichtmayer, J., Hofmann, R., Arndt, H., Griebler, C., 2014.
matrix. As such, rain and flood events may pose a hazard for drinking Grazing of heterotrophic flagellates on viruses is driven by feeding behaviour.
water production, and they require particular attention in risk assess Environ. Microbiol. Rep. 6, 325–330. https://doi.org/10.1111/1758-2229.12119.
ment and management. Diem, S., Cirpka, O.A., Schirmer, M., 2013. Modeling the dynamics of oxygen
consumption upon riverbank filtration by a stochastic-convective approach.
J. Hydrol. 505, 352–363. https://doi.org/10.1016/j.jhydrol.2013.10.015.
Author statement Doussan, C., Poitevin, G., Ledoux, E., Detay, M., 1997. River bank filtration: Modelling of
the changes in water chemistry with emphasis on nitrogen species. J. Contam.
Hydrol. 25, 129–156. https://doi.org/10.1016/S0169-7722(96)00024-1.
He Wang: Sample and data analysis, visualization, writing of the MS. Drayna, P., McLellan, S.L., Simpson, P., Li, S.H., Gorelick, M.H., 2010. Association
Judith Kaletta: Field work, sample analysis. Sigrid Kaschuba: Field work, between rainfall and pediatric emergency department visits for acute
sample analysis. Sondra Klitzke: Field work, sample analysis, writing of gastrointestinal illness. Environ. Health Perspect. 118, 1439–1443. https://doi.org/
10.1289/ehp.0901671.
the MS. Ingrid Chorus: Writing of the MS. Christian Griebler: Concep Dreier, J., Stormer, M., Kleesiek, K., 2005. Use of bacteriophage MS2 as an internal
tualization, supervision, writing of the MS. control in viral reverse transcription-PCR assays. J. Clin. Microbiol. 43, 4551–4557.
https://doi.org/10.1128/JCM.43.9.4551-4557.2005.
Eckert, P., Lamberts, R., Wagner, C., 2008. The impact of climate change on drinking
Declaration of Competing Interest
water supply by riverbank filtration. Water Sci. Technol. Water Supply 8 (3),
319–324. https://doi.org/10.2166/ws.2008.077.
The authors declare that they have no known competing financial Feichtmayer, J., 2019. Bacteria-Phage Interactions: Insights into Quorum Sensing-
interests or personal relationships that could have appeared to influence Induced Anti-Phage Defense, Phage Therapy and the Pulmonary Human Virome
Composition. Doctoral dissertation. Technische Universität München, p. 211. https:
the work reported in this paper. //mediatum.ub.tum.de/?id=1464582.
Feichtmayer, J., Deng, L., Griebler, C., 2017. Antagonistic microbial interactions:
Acknowledgments contributions and potential applications for controlling pathogens in the aquatic
systems. Front. Microbiol. 8, 1–14. https://doi.org/10.3389/fmicb.2017.02192.
Feng, Y.Y., Ong, S.L., Hu, J.Y., Tan, X.L., Ng, W.J., 2003. Effects of pH and temperature
We thank Silke Pabst and Lothar Beetz for technical support and on the survival of coliphages MS2 and Qβ. J. Ind. Microbiol. Biotechnol. 30,
assistance during the transport experiment. Lothar Beetz substantially 549–552. https://doi.org/10.1007/s10295-003-0080-y.
Fillinger, L., Hug, K., Trimbach, A.M., Wang, H., Kellermann, C., Meyer, A.,
supported the technical setup of the experiment and Silke Pabst sup Bendinger, B., Griebler, C., 2019. The D-A-(C) index: a practical approach towards
ported on-site analytical measurements. Funding by the German the microbiological-ecological monitoring of groundwater ecosystems. Water Res.
Research Foundation (DFG; grant number GR 2107/2-1 and Se 508/2), 163, 114902. https://doi.org/10.1016/j.watres.2019.114902.
Fong, T.T., Lipp, E.K., 2005. Enteric viruses of humans and animals in aquatic
the German Ministry of Education (grant 02WT0944), and the Deutsche environments : health risks, detection, and potential water quality assessment tools.
Bundesstiftung Umwelt (grant 33252/01-33/2) is acknowledged. Microbiol. Mol. Biol. Rev. 69, 357–371. https://doi.org/10.1128/MMBR.69.2.357.
Frohnert, A., Apelt, S., Klitzke, S., Chorus, I., Szewzyk, R., Selinka, H.C., 2014. Transport
and removal of viruses in saturated sand columns under oxic and anoxic conditions -
12
H. Wang et al. Journal of Contaminant Hydrology 246 (2022) 103960
potential implications for groundwater protection. Int. J. Hyg. Environ. Health 217, Massmann, G., Knappe, A., Richter, D., Pekdeger, A., 2004. Investigating the influence of
861–870. https://doi.org/10.1016/j.ijheh.2014.06.004. treated sewage on groundwater and surface water using wastewater indicators in
Gerba, C.P., 1984. Applied and theoretical aspects of virus adsorption to surfaces. Adv. Berlin, Germany. Acta Hydrochim. Hydrobiol. 32, 336–350. https://doi.org/
Appl. Microbiol. 30, 133–168. https://doi.org/10.1016/S0065-2164(08)70054-6. 10.1002/aheh.200400543.
Gerba, C.P., 1999. Virus survival and transport in groundwater. J. Ind. Microbiol. Melnick, J.L., Gerba, C.P., Berg, G., 1980. The ecology of enteroviruses in natural waters.
Biotechnol. 22, 535–539. https://doi.org/10.1038/sj.jim.2900663. Crit. Rev. Environ. Sci. Technol. 10, 65–93. https://doi.org/10.1080/
Ghanem, N., Trost, M., Sánchez Fontanet, L., Harms, H., Chatzinotas, A., Wick, L.Y., 10643388009381677.
2018. Changes of the specific infectivity of tracer phages during transport in porous Nagy-Kovács, Z., Davidesz, J., Czihat-Mártonné, K., Till, G., Fleit, E., Grischek, T., 2019.
media. Environ. Sci. Technol. 52, 3486–3492. https://doi.org/10.1021/acs. Water quality changes during riverbank filtration in Budapest, Hungary. Water 11,
est.7b06271. 1–14. https://doi.org/10.3390/w11020302.
Gillefalk, M., Massmann, G., Nützmann, G., Hilt, S., 2018. Potential impacts of induced Ogorzaly, L., Bertrand, I., Paris, M., Maul, A., Gantzer, C., 2010. Occurrence, survival,
bank filtration on surface water quality: a conceptual framework for future research. and persistence of human adenoviruses and F-specific RNA phages in raw
Water 10, 1240. https://doi.org/10.3390/w10091240. groundwater. Appl. Environ. Microbiol. 76, 8019–8025. https://doi.org/10.1128/
Guber, A.K., Shelton, D.R., Pachepsky, Y.A., Sadeghi, A.M., Sikora, L.J., 2006. Rainfall- AEM.00917-10.
induced release of fecal coliforms and other manure constituents: comparison and Pitol, A.K., Bischel, H.N., Boehm, A.B., Kohn, T., Julian, T.R., 2018. Transfer of enteric
modeling. Appl. Environ. Microbiol. 72, 7531–7539. https://doi.org/10.1128/ viruses adenovirus and coxsackievirus and bacteriophage MS2 from liquid to human
AEM.01121-06. skin. Appl. Environ. Microbiol. 84, 1–13. https://doi.org/10.1128/AEM.01809-18.
Hammes, F., Goldschmidt, F., Vital, M., Wang, Y., Egli, T., 2010. Measurement and Prevost, B., Lucas, F.S., Goncalves, A., Richard, F., Moulin, L., Wurtzer, S., 2015. Large
interpretation of microbial adenosine tri-phosphate (ATP) in aquatic environments. scale survey of enteric viruses in river and waste water underlines the health status
Water Res. 44, 3915–3923. https://doi.org/10.1016/j.watres.2010.04.015. of the local population. Environ. Int. 79, 42–50. https://doi.org/10.1016/j.
Harrell Jr., F.E., Dupont, C., et al., 2018. Hmisc: Harrell Miscellaneous. https://CRAN. envint.2015.03.004.
R-project.org/package=Hmisc. Quanrud, D.M., Carroll, S.M., Gerba, C.P., Arnold, R.G., 2003. Virus removal during
Harvey, R.W., Ryan, J.N., 2004. Use of PRD1 bacteriophage in groundwater viral simulated soil-aquifer treatment. Water Res. 37, 753–762. https://doi.org/10.1016/
transport, inactivation, and attachment studies. FEMS Microbiol. Ecol. 49, 3–16. S0043-1354(02)00393-7.
https://doi.org/10.1016/j.femsec.2003.09.015. Rose, J.B., Daeschner, S., Easterling, D.R., Curriero, F.C., Lele, S., Patz, J.A., 2000.
Hassard, F., Gwyther, C.L., Farkas, K., Andrews, A., Jones, V., Cox, B., Brett, H., Jones, D. Climate and waterborne disease outbreaks. Am. Water Works Assoc. 92, 77–87.
L., McDonald, J.E., Malham, S.K., 2016. Abundance and distribution of enteric https://doi.org/10.1002/j.1551-8833.2000.tb09006.x.
bacteria and viruses in coastal and estuarine sediments-a review. Front. Microbiol. 7, Schijven, J.F., Hassanizadeh, S.M., 2000. Removal of viruses by soil passage: overview of
1692. https://doi.org/10.3389/fmicb.2016.01692. modeling, processes, and parameters. Crit. Rev. Environ. Sci. Technol. 30, 49–127.
Hurst, C.J., Gerba, C.P., Cech, I., 1980. Effects of environmental variables and soil https://doi.org/10.1080/10643380091184174.
characteristics on virus survival in soil. Appl. Environ. Microbiol. 40, 1067–1079. Schijven, J., Berger, P., Miettinen, I., 2002. Removal of pathogens, surrogates, indicators,
https://doi.org/10.1128/aem.40.6.1067-1079.1980. and toxins using riverbank filtration. In Riverbank Filtrat. 73–116. https://doi.org/
Jaramillo, M., 2011. Riverbank filtration: an efficient and economical drinking-water 10.1007/0-306-48154-5_7.
treatment technology. DYNA 79, 148–157. Sobsey, M.D., Dean, C.H., Knuckles, M.E., Wagner, R.A., 1980. Interactions and survival
Jofre, J., Lucena, F., Blanch, A.R., Muniesa, M., 2016. Coliphages as model organisms in of enteric viruses in soil materials. Appl. Environ. Microbiol. 40, 92–101. https://
the characterization and management of water resources. Water 8, 1–21. https://doi. doi.org/10.1128/aem.40.1.92-101.1980.
org/10.3390/w8050199. Sprenger, C., Lorenzen, G., Hülshoff, I., Grützmacher, G., Ronghang, M., Pekdeger, A.,
John, D.E., Rose, J.B., 2005. Review of factors affecting microbial survival in 2011. Vulnerability of bank filtration systems to climate change. Sci. Total Environ.
groundwater. Environ. Sci. Technol. 39, 7345–7356. 409, 655–663. https://doi.org/10.1016/j.scitotenv.2010.11.002.
Knappett, P.S., Emelko, M.B., Zhuang, J., McKay, L.D., 2008. Transport and retention of a Tufenkji, N., Ryan, J.N., Elimelech, M., 2002. Peer reviewed: the promise of bank
bacteriophage and microspheres in saturated, angular porous media: effects of ionic filtration. Environ. Sci. Technol. 36, 422A–428A. https://doi.org/10.1021/
strength and grain size. Water Res. 42, 4368–4378. https://doi.org/10.1016/j. es022441j.
watres.2008.07.041. Wellings, F.M., Lewis, A.L., Mountain, C.W., Pierce, L.V., 1975. Demonstration of virus in
Krauss, S., Griebler, C., 2011. Pathogenic Microorganisms and Viruses in Groundwater. groundwater after effluent discharge onto soil. Appl. Microbiol. 29, 751–757.
Kropinski, A.M., Mazzocco, A., Waddell, T.E., Lingohr, E., Johnson, R.P., 2009. https://doi.org/10.1128/aem.29.6.751-757.1975.
Enumeration of bacteriophages by double agar overlay plaque assay. Methods Mol. Wyn-Jones, A.P., Sellwood, J., 2001. Enteric viruses in the aquatic environment. J. Appl.
Biol. 501, 69–76. https://doi.org/10.1007/978-1-60327-164-6. Microbiol. 91, 945–962. https://doi.org/10.1046/j.1365-2672.2001.01470.x.
Kuehn, W., Mueller, U., 2000. Riverbank filtration: an overview. J. Am. Water Works Yang, K., LeJeune, J., Alsdorf, D., Lu, B., Shum, C.K., Liang, S., 2012. Global distribution
Assoc. 92, 60–69. https://doi.org/10.1002/j.1551-8833.2000.tb09071.x. of outbreaks of water-associated infectious diseases. PLoS Negl. Trop. Dis. 6, e1483
Landry, E.F., Vaughn, J.M., Thomas, M.Z., Beckwith, C.A., 1979. Adsorption of https://doi.org/10.1371/journal.pntd.0001483.
enteroviruses to soil cores and their subsequent elution by artificial rainwater. Appl. Yates, M.V., Yates, S.R., Wagner, J., Gerba, C.P., 1987. Modeling virus survival and
Environ. Microbiol. 38, 680–687. https://doi.org/10.1128/aem.38.4.680-687.1979. transport in the subsurface. J. Contam. Hydrol. 1, 329–345. https://doi.org/
Lodder, W.J., de Roda Husman, A.M., 2005. Presence of noroviruses and other enteric 10.1016/0169-7722(87)90012-X.
viruses in sewage and surface waters in the Netherlands. Appl. Environ. Microbiol. Yates, M.V., Stetzenbach, L.D., Gerba, C.P., Sinclair, N.A., 1990. The effect of indigenous
71, 1453–1461. https://doi.org/10.1128/AEM.71.3.1453-1461.2005. bacteria on virus survival in ground water. J. Environ. Sci. Health Part A 25, 81–100.
Lodder, W.J., Van Den Berg, H.H.J.L., Rutjes, S.A., de Roda Husman, A.M., 2010. https://doi.org/10.1080/10934529009375541.
Presence of enteric viruses in source waters for drinking water production in the Zyman, J., Sorber, C.A., 1988. Influence of simulated rainfall on the transport and
Netherlands. Appl. Environ. Microbiol. 76, 5965–5971. https://doi.org/10.1128/ survival of selected indicator organisms in sludge-amended soils. J. Water Pollut.
AEM.00245-10. Control Fed. 60, 2105–2110.
Loveland, J.P., Ryan, J.N., Amy, G.L., Harvey, R.W., 1996. The reversibility of virus
attachment to mineral surfaces. Colloids Surf. A Physicochem. Eng. Asp. 107,
205–221. https://doi.org/10.1016/0927-7757(95)03373-4.
13