Science of the Total Environment 518-519 (2015) 130–138
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Attenuation and colloidal mobilization of bacteriophages in natural sediments under anoxic as compared to oxic conditions Sondra Klitzke a,⁎, Jendrik Schroeder a, Hans-Christoph Selinka b, Regine Szewzyk b, Ingrid Chorus c a b c
Federal Environment Agency, Section Drinking Water Treatment and Resource Protection, Schichauweg 58, D-12307 Berlin, Germany Federal Environment Agency, Section Microbiological Risks, Corrensplatz 1, D-14197 Berlin, Germany Federal Environment Agency, Department of Drinking Water and Swimming Pool Hygiene, Schichauweg 58, D-12307 Berlin, Germany
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• MS2 and PhiX174 did not attach to colloids irrespective of pH and redox potential. • Phage removal efficiency in anoxic environments depends on organic matter dynamics. • Organic matter may block Fe/Al-oxides in anoxic sediments.
a r t i c l e
i n f o
Article history: Received 27 November 2014 Received in revised form 6 February 2015 Accepted 8 February 2015 Available online 5 March 2015 Editor: D. Barcelo Keywords: Colloid mobilization Phage Riverbank filtration Sesquioxide Virus Zeta potential
a b s t r a c t Redox conditions are known to affect the fate of viruses in porous media. Several studies report the relevance of colloid-facilitated virus transport in the subsurface, but detailed studies on the effect of anoxic conditions on virus retention in natural sediments are still missing. Therefore, we investigated the fate of viruses in natural flood plain sediments with different sesquioxide contents under anoxic conditions by considering sorption to the solid phase, sorption to mobilized colloids, and inactivation in the aqueous phase. Batch experiments were conducted under oxic and anoxic conditions at pH values between 5.1 and 7.6, using bacteriophages MS2 and PhiX174 as model viruses. In addition to free and colloid-associated bacteriophages, dissolved and colloidal concentrations of Fe, Al and organic C as well as dissolved Ca were determined. Results showed that regardless of redox conditions, bacteriophages did not adsorb to mobilized colloids, even under favourable charge conditions. Under anoxic conditions, attenuation of bacteriophages was dominated by sorption over inactivation, with MS2 showing a higher degree of sorption than PhiX174. Inactivation in water was low under anoxic conditions for both bacteriophages with about one log10 decrease in concentration during 16 h. Increased Fe/Al concentrations and a low organic carbon content of the sediment led to enhanced bacteriophage removal under anoxic conditions. However, even in the presence of sufficient Fe/A-(hydr)oxides on the solid phase, bacteriophage sorption
Abbreviations: PFU, plaque forming units; Corg, organic carbon; DOC, dissolve organic carbon. ⁎ Corresponding author. E-mail addresses: [email protected] (S. Klitzke), [email protected] (J. Schroeder), [email protected] (H.-C. Selinka), [email protected] (R. Szewzyk), [email protected] (I. Chorus).
http://dx.doi.org/10.1016/j.scitotenv.2015.02.031 0048-9697/© 2015 Elsevier B.V. All rights reserved.
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
131
was low. We presume that organic matter may limit the potential retention of sesquioxides in anoxic sediments and should thus be considered for the risk assessment of virus breakthrough in the subsurface. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Human pathogenic viruses which are transmitted via the faecal–oral route, such as enteric viruses, are frequently found in surface water (Hot et al., 2003; Lodder and De Roda Husman, 2005; Albinana-Gimenez et al., 2006; Ngazoa et al., 2008; Westrell et al., 2006). Riverbank filtration, artificial groundwater recharge and slow sand filtration are common methods for the natural (pre-)treatment of drinking water. Riverbank filtration and slow sand filtration have both been reported to remove viruses or virus surrogates (i.e. model viruses such as bacteriophages MS2 or PhiX174) during subsurface passage (Weiss et al., 2005; Bauer et al., 2011). The elimination of viruses is thereby mainly governed by hydraulic settings and physico-chemical parameters of water and sediment along with the surface characteristics of viruses itself (Jin and Flury, 2001). Redox conditions are reported to play a major role in the transport and inactivation of bacteriophages (Schijven et al., 2000; Van der Wielen et al., 2008). Both authors observed lower virus removal in anoxic than in oxic zones of a natural sandy aquifer which they attributed to lower inactivation and adsorption rates under anoxic conditions. They attribute the reduced virus adsorption to the dissolution of Fe(hydr)oxides. Similarly, Frohnert et al. (2014) observed lower virus removal under anoxic conditions in a laboratory column experiment during a 4-week-period. A number of studies indicate the relevance of colloid-associated virus transport along preferential flow paths in sediments (Seta and Karathanasis, 1997; Jin et al., 2000). Colloids are small (1 nm to 10 μm; Brady and Weil, 2002; Stumm and Sigg, 1979) and have a large surface to volume ratio (Kretzschmar et al., 1999). Therefore, colloids tend to be stable in suspensions and are hardly filtered by porous media (Seta and Karathanasis, 1997). These properties make colloids a possible carrier for sorbing viruses (McGechan and Lewis, 2002) which would otherwise be retained by the sediment. Jin et al. (2000) showed that colloid-facilitated MS2-phage transport in porous media was clearly correlated with the amount of transported clay colloids and the number of viruses adsorbed to these. Failure to consider this pathway may result in an underestimation of virus travel velocity, distances and concentrations (McCarthy and Zachara, 1989). To predict the fate and transport of viruses in contaminated waters and sediments, it is important to understand the dynamics of colloid release from sediments and the governing mechanisms of colloid–virus-interactions. Sorption of viruses to colloids and sediments is controlled by their surface properties, with the surface charge being a key parameter. The charge of individual viruses at a given pH is a function of their isoelectric point (IEP), which varies for different virus types (Jin et al., 1997; pHiep: 3.9 for MS2, and pHiep 6.6 for PhiX174; Dowd et al., 1998). Hence, virus charge is controlled by pH and ionic strength of the solution. Similarly, colloid charge is also a function of IEP (determined by colloid composition) and solution chemistry, such as pH, concentrations of dissolved organic carbon and ionic strength (Kretzschmar et al., 1999). Emelko and Tufenkje (2010) highlight the need to understand the fate of colloidal pathogens such as viruses in the subsurface (i.e. riverbank filtration systems). However, so far, only few studies looked into the fate of viruses under anoxic conditions (for instance Frohnert et al., 2014). There are many publications on virus retention as a function of pH (Schijven and Hassanizadeh, 2000; Jin and Flury, 2001; Walshe et al., 2010), but they all refer to oxic conditions. In addition, the role of colloid-facilitated virus mobilization or virus transport has never been addressed in this context. Moreover, sesquioxides (i.e. Feand Al-(hydr)oxides), which often occur in close association with each other, and their role on virus retention under anoxic conditions is still
unclear: non-redox-sensitive sorption sites such as Al-(hydr)oxides could improve virus retention under all redox conditions, but if they are released from the sediment due to break-up of bonds through the dissolution of Fe they may also enhance virus transport if viruses sorb to them. Therefore, the aim of our study was to understand (i) the dynamics of colloid release from natural flood plain sediments with different sesquioxide contents under anoxic conditions as compared to oxic conditions. (ii) the fate of viruses in these natural system such as sorption to the solid phase, sorption to mobilized colloids, and inactivation in the aqueous phase by testing the influence of pH changes on the attenuation of two model viruses, PhiX174 and MS2, in laboratory batch experiments. These two bacteriophages were chosen as they differ in their isoelectric points (Dowd et al., 1998), which is reported to lead to different sorption behaviour (Jin and Flury, 2001). 2. Material and methods 2.1. Origin and characterization of sediments and water Two sediments were obtained from the Danube floodplain. Sediment 2011 was sampled in an isolated river section with little hydrological connection to the Danube River. This section was silted up with medium-grained fine sand and overgrown with reed and riverine vegetation. Sediment 2016 was sampled in a river section with a pronounced hydrological connection to the Danube, characterized by regular erosion and sedimentation processes and a texture of silty sand. Both samples were mixed and homogenized individually prior to their storage at 4 °C under saturated conditions (i.e. the sediment was always covered with a few centimetres of river water). Both sediments differ strongly in their concentration of organic carbon, Fe and Al (Table 1): sediment 2011 has a high organic carbon (Corg) and a low Fe/Al content, sediment 2016 has a low Corg and a high Fe/Al content. For our experiments we used treated groundwater (i.e. by removal of Fe and Mn through microbial precipitation) abstracted from a quaternary aquifer on the experimental field site of the Federal Environment Agency in BerlinMarienfelde, Germany. In order to obtain water of low ionic strength, this groundwater was diluted 1:10 with deionised water, yielding a final ionic strength of about 2 mM, comprising Ca2+ and SO2− concen4 trations of about 0.40 mM and 0.25 mM, respectively. The water was autoclaved (20 min at 121 °C) prior to its use to remove any dissolved oxygen and is termed “test water” in the following. 2.2. Bacteriophages Bacteriophages MS2 (DSM 13767) and PhiX174 (DSM 4497) were used as surrogates of RNA and DNA viruses, respectively. Both viruses develop icosahedral virions with sizes of 25–30 nm. MS2 is an RNA bacteriophage of the family Leviviridae with a single-stranded RNA genome, the genome of phage PhiX174 (family Microviridae) consists of singlestranded DNA. The isoelectric points given in literature are about 3.9 for phage MS2 and about 6.6 for phage PhiX174 (Michen and Graule, 2010). To achieve stock suspensions with high bacteriophage concentrations, host bacteria were grown up to exponential growth phase and subsequently inoculated with the respective phages, as described in standardized procedures (DIN EN ISO, 10705-1, 2001; DIN EN ISO, 10705-2, 2001). 10% v/v chloroform was added to destroy the bacteria. After incubation and settling, supernatants were centrifuged (3000 ×g
132
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
Table 1 Investigated parameters of the tested sediments of the Danube River.
Location pH Electrical conductivity Organic carbon Inorganic carbon Oxalat-extractable Fe Oxalat-extractable Al Dithionite-extractable Fe Dithionite-extractable Al
[μS/cm] [mg/kg] [g/kg] [mg/kg] [mg/kg] [mg/kg] [mg/kg]
Sediment 2011
Sediment 2016
Isolated wetland (reed and coastal vegetation); strong aggradation 7.6 139.7 17,938 13.3 293 30 395 116
Wetland with connection to Danube River; frequent erosion and sedimentation 7.9 146.6 4,963 26.3 1866 75 2,952 237
for 5 min) to separate bacteria residues from bacteriophage particles. Supernatants of centrifugations were frozen in aliquots at (−80 ± 10) °C until use. Phage concentrations amounted to approximately 1012 pfu/mL (MS2) and about 1010 pfu/mL (PhiX174), respectively. Stock suspensions of both bacteriophages were individually diluted in decimal steps using peptone saline (8.5 g/L NaCl, 1 g/L peptone) up to a concentration of 109 pfu/mL. Further dilution was performed in test water to avoid influence of residual peptone on the experiments. Test water was inoculated with bacteriophages from the diluted suspensions to yield concentrations of approximately 107 pfu/mL for both bacteriophages. These suspensions in test water were used for inoculation of the batch samples described below. For anoxic experiments, peptone saline and the test water were purged with nitrogen gas for 15 min after autoclaving and transferred into the nitrogenflushed glove box where the phage suspensions were prepared. To determine the initial phage concentrations, aliquots of these suspensions were analysed according to DIN EN ISO 10705-1 (2001) and DIN EN ISO 10705-2 (2001). 2.3. Design of batch experiments We used field-moist sediment and set up triplicate batch series at 3 to 4 different pH values (see below; oxic experiments: plastic tubes, BD; anoxic experiments: glass flasks, Schott) with a soil to water ratio of 1:10, i. e. 8 g sediment mass and a total suspension volume of 80 mL. This ratio proved to be wide enough to prevent any artificial generation of colloids by abrasion (Klitzke et al., 2008). For the comparison of different pH values, appropriate amounts of dilute nitric acid (oxic experiments) or dilute hydrochloric acid (previously purged with N2; anoxic experiments), respectively were added. The following pH values were adjusted: 5.5 ± 0.4, 6.5 ± 0.1, 7.3 ± 0.1, 7.6 ± 0.1 for oxic experiments, and 5.1 ± 0.1, 6.5 ± 0.2, 7.3 ± 0.3 for anoxic experiments, with the latter ones being the natural pH values of the sediments. Batch samples i.e. 8 g of sediment and 70 mL of test water, were left to stand for a day at room temperature to allow equilibration following acid addition. In the anoxic experiments, samples were purged with nitrogen for 15 min prior to their transfer into the glovebox. 2.3.1. Oxic experiments with sediment We added 10 mL of phage suspension (c = 107 pfu/mL; see Section 2.2.) to each batch sample at the end of the equilibration period to yield a final phage concentration of approximately 106 pfu/mL and a
total sample volume of 80 mL. Samples were put on an incubating shaker (Thermoshake, Gerhardt) at 110 rpm for 16 h. The pH values were checked before and after the shaking period (Table 2). To separate material larger than 2 μm, samples were centrifuged at 100 g for 4 min (based on the assumption of an average density of 2.65 g/cm3, i.e. corresponding to the density of quartz and representing an estimated average of organic and mineral material; ISO, 11277, 1998). The centrifugation parameters were chosen in accordance with Stoke's Law (Tanner and Jackson, 1948). The removal of particles N 2 μm was confirmed using dynamic light scattering measurements. Supernatants were used to determine total concentrations (i. e. the sum of (i) dissolved and colloidal element concentrations, (ii) “free” and colloid-associated bacteriophage concentrations) of organic C, Fe, Al, Mn as well as bacteriophages. An aliquot was centrifuged (Avanti J-26 XP Centrifuge, Beckman Coulter) at 6800 g for 80 min to remove any colloidal material larger than 100 nm, assuming an average density of 1.5 g/cm3 to remove as much as possible of the suspended material. This low density was chosen as organic colloids may have a density much below the average (1.1 g/cm3 (Citeau et al., 2001), 1.6 g/cm3 (Bethwell, 2004). The supernatants of these centrifugates were defined as “dissolved” for the determined elements, and as “free” for the respective bacteriophages. The differences of the metal, C and bacteriophage concentrations between these two supernatants (i.e. total and dissolved/“free”) account for colloid-associated fractions (defined operationally). Organic C, Fe, Al, and Ca, concentrations of both phages, optical density and the zeta potentials were determined as described in the section analysis (see Section 2.4.). 2.3.2. Anoxic experiments with sediment The samples were transferred into a nitrogen-flushed glove box (O2 content below 0.5–1%) through an airlock and left to stand for 14 days (sediment 2016) or 31 days (sediment 2011) to allow the redox potential to drop in consequence of the mineralization of organic matter. Initial redox potentials of the batch samples were −120 ± 53 mV (sediment 2016) and −140 ± 36 (sediment 2011). Aliquots of phage suspensions in test water (see Section 2.2) were added to each sample at the end of the equilibration period to yield a final phage concentration of 106 pfu/mL and a total sample volume of 80 mL. pH and redox potential were checked and, if necessary, pH was re-adjusted (Table 3). Samples were closed with an airtight lid and put on an incubating shaker (Thermoshake, Gerhardt) at 110 rpm for 16 h. Redox potentials remained fairly constant during shaking, i.e. in most cases they did not
Table 2 pH values before and after shaking of the sediments under oxic conditions (n.d. not determined). Sediment 2016 Target pH pH Redox potential
Before shaking After shaking Before shaking After shaking
5.5 5.0 5.4 339 431
Sediment 2011 6.5 6.2 6.4 355 415
7.3 7.0 7.4 355 393
7.6 7.8 7.7 n. d. n. d.
5.5 5.4 5.9 356 382
6.5 6.2 6.4 356 395
7.3 6.9 7.3 355 382
7.6 7.6 7.5 n. d. n. d.
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
differ by more than 20 mV (Fig. Supp1a and Supp1b in the Supporting material). Subsequently, samples were re-introduced into the glove box, followed by checks of pH and redox potentials (Table 3). Samples were transferred into centrifuge tubes and centrifuged at conditions specified in Section 2.3.1. For the removal of the supernatants after each centrifugation step samples were put back into the glove box. Organic C, Fe, Al, and phage concentrations were determined as described in Section 2.4.
2.3.3. Determination of bacteriophage distribution in sediment samples Concentrations of phages were determined in the initial suspensions in test water (see Section 2.2.) and in the supernatants of the two different centrifugation steps described in Section 2.3.1. Bacteriophage concentrations in the supernatant of the first centrifugation step (containing only particles smaller than 2 μm) represent suspended bacteriophages following the removal of larger sediment material and comprise both “free” and colloid-associated bacteriophages (termed as “total” in the following) while bacteriophages in the supernatant of the second centrifugation step (containing only particles smaller than 100 nm) comprise only “free” bacteriophages (termed as “free” in the following). Differences in bacteriophage concentrations between the initial suspensions and the supernatants of the first centrifugation step (b2 μm) may be ascribed to bacteriophage sorption to larger sediment material and/ or to inactivation. Differences in concentrations between the supernatants of the first centrifugation step (i.e. b 2 μm) and the second centrifugation step (i. e. b 100 nm) may be ascribed to the concentration of colloid-associated bacteriophages.
2.3.4. Anoxic experiments without sediment To differentiate between the two key processes of bacteriophage elimination – sorption and inactivation – an additional set of batch samples containing only anoxic pore water was prepared to determine survival of phages. Samples were prepared in triplicate as described in Section 2.3. and subsequently transferred into a nitrogen-flushed glovebox, where they were left to stand until the redox potentials dropped below 0 mV (incubating time: sediment 2016: 28 days (− 3 to −88 mV); sediment 2011: 5 days (−72 to −136 mV)). To remove any possible sorption sites the samples were transferred into centrifuge tubes, sealed with a lid, centrifuged at 6800 g for 80 min (cut-off 100 nm, based on an assumed density of 1.5 g/cm3). Supernatants were removed in the glovebox and 35 mL of each supernatant was inoculated with 5 mL of phage suspension. The flasks were closed with an airtight lid and put on an incubating shaker outside the glovebox at 110 rpm of 16 h. Afterwards, samples were re-introduced into the glovebox for subsequent check of pH and redox potential (5.5 ± 0.1, 6.5 ± 0.1 (both sediments), 7.3 (sediment 2011) and 8.2 (sediment 2016); 272 ± 14 mV (sediment 2011) and 186 ± 21 mV (sediment 2016)). The measured redox potential in the pore water was higher than in the sediment–water system, which may be ascribed to concentration effects of redox-sensitive ions in the suspension caused by the removal of the sediment (Santschi et al., 1990). Assuming no bacteriophage attachment to colloids, the combination of the results obtained in the absence and in the presence of sediments allows for the quantification of the total elimination (elimtot) according to Eq. (1), with sorptsolid being the sorption to the solid phase and Table 3 pH values and redox potentials before and after shaking of the sediments under anoxic conditions. Sediment 2016 Target pH pH
Sediment 2011
5.1 6.5 7.3 5.1 6.5 7.3 Before shaking 5.0 6.4 7.2 5.0 6.2 7.7 After shaking 5.0 6.6 7.0 5.1 6.7 7.5 Redox potential Before shaking −2 −94 −179 −41 −157 −147 After shaking 27 −60 −150 −11 −166 −167
133
inactaq being the inactivation in the aqueous phase: Elimtot ¼ sorptsolid þ inactaq :
ð1Þ
2.4. Analysis Organic C was measured with a vario TOC cube (Elementar). Iron, Al, and Ca concentrations of centrifugates were analysed in diluted and acidified samples (3% HNO3) by ICP-OES. Optical density determined as light absorption at 525 nm of the supernatant from the first centrifugation step (b 2 μm) of oxic samples was taken as a measure for the relative amount of dispersible particles as specified by Kretzschmar et al. (1997) on a spectrophotometer (Spektrophotometer PM6, Zeiss, Germany). As colloid-free solutions also showed some light absorbance at 525 nm (presumably caused by dissolved organic matter) we corrected for this interfering absorbance by subtracting the absorbance of the supernatant from the second centrifugation step (b 100 nm). Zeta potential and size of colloids were determined in the (bacteriophagefree) suspensions obtained from sediment suspended in the same matrix as samples containing bacteriophages. Zeta potential was calculated from the electrophoretic mobility (using the Helmholtz–Smoluchowski equation), which was analysed by a Zetasizer 2000 photon correlation spectrometer (Malvern Instruments). Particle size in the supernatant of bacteriophage-containing and -free suspensions was determined by dynamic light scattering (DLS, Malvern Instruments). The bacteriophages MS2 and PhiX174 were quantified by a double layer agar method according to DIN EN ISO 10705-1 (2001) and DIN EN ISO 10705-2 (2001). Each sample was analysed in duplicates. Samples with higher concentrations of bacteriophages were diluted in decimal steps using peptone saline (8.5 g/L NaCl, 1 g/L peptone). Detection limits were 1 pfu/mL for 1 mL sample volumes and 0.2 pfu/mL for 5 mL volumes, respectively. 3. Results and discussion 3.1. Experiments performed under oxic conditions Under oxic conditions, batch experiments with sediments were conducted at different pH values to determine the distribution of bacteriophages between (i) larger sediment particles and (ii) colloids as well as the amount of free bacteriophages. 3.1.1. Colloid properties The optical density of both sediments increased with increasing pH (Fig. Supp2 in the Supporting material), suggesting a mobilization of sediment colloids with increasing pH. This observation is supported by higher colloidal concentrations of organic C (Fig. 1a), and Fe, and Al (Fig. 1b), which increase with increasing pH (except for organic C in sediment 2016; Fig. 1a). The zeta potential, a relative measure of the particle charge, clearly decreased with increasing pH (Fig. 2). The resulting repulsion between negatively charged colloids leads to a stronger mobilization of colloids. Both findings are in accordance with classical theories of colloidal behaviour described in the literature (Kretzschmar et al., 1999). Sizes of dispersed pedogenic colloids ranged between 259 and 1530 nm (sediment 2016) and between 477 and 1760 nm (sediment 2011). Size distribution of dispersed particles in the supernatant (b 2 μm) following bacteriophage addition ranged between 256 and 1820 nm (sediment 2016) and between 466 and 1850 nm (sediment 2011). This suggests that the addition of bacteriophages does not affect the size distribution of dispersed colloids. 3.1.2. Fate of bacteriophages in sediment batch samples At the end of the batch experiments, in both sediments at pH 7.3 and 7.6, total phage concentrations showed the same concentrations as in initial suspensions (Fig. 3a and b, e and f), suggesting no phage sorption
134
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
Total and free concentrations of both bacteriophages did not markedly differ from each other for both sediments at all investigated pH values (e.g. Fig. 3a and b, e and f) suggesting no binding of bacteriophages to dispersed colloids. This non-attachment is explainable for samples with equal charges of colloids (Fig. 2) and bacteriophages, i.e. NpH 3.9 for MS2 and NpH 6.6 for PhiX174, as bacteriophages are then negatively charged and thus repelled from negatively charged colloids. It remains unclear why there is no bacteriophage attachment to colloids when charges are attractive, i.e. at pH 6 to 6.6 for PhiX174 (for both sediments) or pH b 6 for MS2 (only sediment 2016). One reason could be the association of organic matter. Both sediments contain considerable amounts of organic matter, which often acts as partial coating of mineral phases (McCarthy and Zachara, 1989). These coatings may block potential sorption sites for viruses on dispersed colloids (Schijven and Hassanizadeh, 2000) and hence prevent attachment. 3.2. Experiments performed under anoxic conditions
Fig. 1. pH-dependent colloidal concentrations of organic carbon (a), as well as Fe and Al (b) in oxic batch samples of sediments 2011 and 2016.
to sediment material and/or inactivation. However, at pH 6.5, total concentrations of both phages in the supernatants were 2 to 3 log10 lower than in initial suspensions in sediment 2011 (Fig. 3e and f), indicating substantial sorption to sediment material and/or inactivation. This effect was even more pronounced for sediment 2016, in which total phage concentrations were reduced in the supernatant to a level below the limit of detection (Fig. 3a and b), corresponding to a removal through sorption to sediment material and/or inactivation of at least 5 to 6 log10. The higher removal in sediment 2016 may be ascribed to the higher Ca concentrations compared to sediment 2011 at pH 6.5 (Fig. 4), leading to enhanced phage aggregation due to a compression of the thickness of the diffuse double layer (Walshe et al., 2010), which may facilitate sorption. An enhanced phage removal at elevated Ca concentrations has also been observed by Walshe et al. (2010). At pH 5.5 (Fig. 3a and b, e and f), total phage concentrations were below the quantification limit for both sediments, suggesting complete or almost complete removal through sorption to sediment material and/or inactivation — with sorption being presumably facilitated by high Ca concentrations (Fig. 4).
Fig. 2. Zeta potentials of colloids dispersed at various pH values under oxic conditions (without the addition of bacteriophages). Bars depict one standard deviation.
Under anoxic conditions, batch experiments with sediments were conducted at different pH values to determine the distribution of bacteriophages between (i) larger sediment particles and (ii) colloids as well as the amount of free bacteriophages. Furthermore, the survival of bacteriophages was determined in batch samples using anoxic pore water. Up to now, there are only few studies which look into the fate of viruses in the subsurface under anoxic conditions (Van der Wielen et al., 2008; Schijven et al., 2000), which are based on field-scale observations. Therefore, our investigations aiming at a mechanistic understanding of the fate of viruses close an important knowledge gap. Redox potentials in anoxic batch samples of both sediments did not show any considerable changes (i.e. smaller than 20 mV in most cases) before and after shaking (Fig. Supp1a and Supp1b in the Supporting material) confirming that no oxygen penetration was possible in our experimental setup outside the glove box. 3.2.1. Colloids in sediment batch samples While Al colloids were only dispersed at very low concentrations, both sediments showed elevated concentrations of dispersed Fe colloids (Fig. Supp3 in the Supporting material). However, in most cases, bacteriophages did not attach to them (see also Sections 3.2.2. and 3.2.3.). Colloidal Fe concentrations under anoxic conditions did not differ from concentrations under oxic conditions (data not shown) at pH 5.1/5.5 (both sediments; both p-values 0.44) and pH 6.5 (sediment 2011; p-value 0.63). At pH 7.3 (both sediments; p-value 0.04) and pH 6.5 (sediment 2016; p-value 0.07) concentrations were slightly higher in anoxic samples. This is surprising as one could assume the dissolution of dispersed Fe colloids under anoxic conditions. One reason explaining our observation could be the blockage of Fe surfaces by organic matter as mineral colloids are often coated with organic matter (Kretzschmar et al., 1999). Organic matter is reported to reduce the specific surface area of minerals (Kaiser and Guggenberger, 2003) and thus to lead to ‘pore clogging’ of mineral phases (Mikutta et al., 2004). Therefore, we propose the following conceptual model as most likely mechanism which could have prevented the reduction of Fe: organic matter may have shielded the Fe oxide surfaces from microorganisms or even from small redox-reactive organic molecules (so called “redox mediators”) released by microorganisms (Patil et al., 2012). 3.2.2. Fate of bacteriophages in sediment 2016 batch samples At pH 7.3, total concentrations of both bacteriophages were about 2 to 3 log10 lower than in initial suspension (Fig. 3c and d), respectively, suggesting sorption to sediment material and/or inactivation under anoxic conditions. This is in contrast to the results under oxic conditions where no sorption to sediment material and/or inactivation was detected at this pH. We suggest the following concept for enhanced sorption under anoxic conditions: Initially, in these sediments, organic matter was present as organic coating of the sediment. Such coatings are
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
135
Fig. 3. Bacteriophage concentrations in the supernatants of oxic and anoxic batch samples of sediments 2016 and 2011: bacteriophages PhiX174 (a, c, e, g) and MS2 (b, d, f, h) under oxic (a, b, e, f) and anoxic (c, d, g, h) conditions. Note that in sediment 2016 oxic experiments were also conducted at pH 6.5, but concentrations were below the limit of detection. Dotted lines: initial phage concentrations of PhiX174 in the batch sample; solid lines: initial phage concentration of MS2 in the batch samples. Bars depict one standard deviation. n.d.: not determined.
reported to cover mineral surfaces (Maurice and Namjesnik-Dejanovic, 1999), blocking their reactive sites if organic matter concentrations are low (Kaiser and Guggenberger, 2003). The increase in DOC concentrations (Fig. 5) in comparison to oxic conditions gives rise to the assumption that during the pre-incubation period used to attain anoxic conditions, such coatings were in part subject to mineralization. Their (partial) break-up may thus render reactive sorption sites on the mineral phase accessible, leading to enhanced retention of bacteriophages. As the organic
carbon content of this sediment was low, the effect may be relevant. Even though dissolved Fe concentrations showed a tendency to be higher in anoxic than in oxic samples (Fig. Supp4 in the Supporting material) – suggesting the dissolution of some of the Fe-oxides – enough binding sites probably remained, as concentrations of Fe and Al are high in this sediment. This model remains to be tested by future research. At pH 6.5 under anoxic conditions, PhiX174 showed a similar pattern as at pH 7.3 i.e. a 2 log10 decrease in concentration, which may be
136
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
suggest that the dissolution of Fe decreases the number of available Fe sorption sites, thereby further diminishing the already low Fe content of the sediment. However, even though there were still considerable amounts of Fe left on sediment 2011 following Fe reduction, bacteriophage retention on this sediment was low (1 to 2 log10). Therefore, we presume that the high organic carbon content may block sorption sites available on sesquioxides as discussed above for sediment 2016. As expected, dissolved Al concentrations in oxic and unoxic samples did not differ from each other (except for at pH 5.1/5.5; Fig. 6). Similarly to the effects observed with sediment 2016, concentrations between total and free bacteriophages did not differ from each other in any samples, i.e. bacteriophages did not adsorb to colloids.
Fig. 4. Dissolved Ca concentrations in oxic sediment batch samples. Bars depict one standard deviation.
attributed to sorption to sediment/inactivation (Fig. 3c). This is in contrast to the experiments under oxic conditions (pH 6.5) when PhiX174 is completely eliminated due to sorption to sediment/inactivation (Fig. 3a). This pH is close to the isoelectric point of PhiX174, at which bacteriophages are more prone to aggregation. Under oxic conditions, the high Ca concentration (Fig. 4) may have facilitated additional aggregation. As DOC is known to complex Ca ions (McBride, 1994), the increased DOC release under anoxic conditions (Fig. 5) may weaken the aggregating effect of the Ca ion. The same concept may also be ascribed to both phages at pH 5.1 (oxic)/pH 5.5 (anoxic), respectively. MS2 was eliminated by 6 log10 units under both oxic and unoxic conditions at pH 6.5 (Fig. 3b and d). Concentrations of total and ‘free’ bacteriophages did not show marked differences between any of the samples, suggesting that bacteriophages did not adsorb to colloids. 3.2.3. Fate of bacteriophages in sediment 2011 batch samples At pH 7.3, total concentrations of both bacteriophages were nearly as high as initial concentrations (Fig. 3g and h), indicating that there was no substantial sorption to sediment material and/or inactivation under anoxic conditions. A similar behaviour was detected under aerobic conditions. At this pH, there was no difference in DOC concentrations between oxic and anoxic samples (data not shown), suggesting no additional DOC mineralization. Alternatively, the long incubation time (31 days) needed to render samples anoxic may have lead to a high degree of DOC mineralization. At pH 6.5 elimination by sorption to sediment material and/or inactivation of PhiX174 was higher than at pH 7.3 (about 0.5 log10 units), while MS2 did not show any significant difference. The elimination of both phages was lower than under oxic conditions (2 to 3 log10). Higher dissolved Fe-concentrations in anoxic than in oxic samples (Fig. 6)
Fig. 5. DOC concentrations in the supernatants of oxic and anoxic samples (sediment 2016 — low Corg, high Fe/Al content). Bars depict one standard deviation.
3.2.4. Fate of bacteriophages in pore water samples obtained from sediment 2016 Concentrations of PhiX174 and MS2 in the supernatants of anoxic pore water (see Section 2.3.4.) decreased by 1 to 1.5 log10 at pH 5.5 and 6.5 during 16 h of incubation (data not shown). This shows that inactivation of phages was low. At pH 8.2, concentrations of both bacteriophages were as high as in initial suspensions (data not shown), suggesting no bacteriophage inactivation. Even though the pH in these samples was higher than in samples containing sediment (i.e. pH 7.3), one may conclude that bacteriophages show equally good survival in anoxic pore water at pH 7.3, as survival is normally worse at alkaline than at neutral pH (Feng et al., 2003). As a consequence one may conclude that MS2 removal in the sediment–water-system is dominated by sorption to sediment material, inactivation in water only plays a minor role (Fig. 7a; Eq. (1)). Sorption was highest at pH 6.5. For PhiX174 inactivation is relatively more important due to the lower overall removal of only 2 log10 due to sorption on sediments/inactivation. PhiX174 removal at pH 5.1 is mainly attributed to inactivation, with sorption becoming more dominant as pH increases (Fig. 7b; Eq. (1)). The increase in sorption with increasing pH may be explained by the isoelectric point (Jin and Flury, 2001). 3.2.5. Fate of bacteriophages in pore water samples obtained from sediment 2011 Concentrations of PhiX174 in the supernatants of anoxic pore water (see Section 2.3.4) remained more or less constant over the 16 h incubation at all investigated pH values. This shows that inactivation was very low (data not shown). Concentrations of MS2 in the supernatants decreased slightly by 0.6 to 1 log10 (data not shown), demonstrating limited inactivation.
Fig. 6. Dissolved Fe and Al concentrations in the supernatants of oxic (+) and anoxic (−) samples (sediment 2011 — high Corg, low Fe/Al content). Bars depict one standard deviation.
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
137
Fig. 7. Removal processes for bacteriophages MS2 and PhiX174 under anoxic conditions in sediment 2016 (a, b) and in sediment 2011 (c, d). Bars depict standard deviations.
Regarding the overall removal, inactivation seems to play the main role in MS2 removal at pH 6.5 and 7.3, but sorption dominates at pH 5.1 (Fig. 7c; Eq. (1)), i.e. the pH closest to the isoelectric point. For PhiX174 removal, however, inactivation is of minor importance: elimination is mainly governed by sorption, which decreases with increasing pH (Fig. 7d; Eq. (1)). Based on the isoelectric point of PhiX174, one would assume an opposite sorption behaviour (Jin and Flury, 2001). This is contrary to the fate of PhiX174 in sediment 2016. For MS2, the fraction removed through sorption in this sediment is lower than in sediment 2016, suggesting that MS2 retention is predominantly governed by sediment properties (see below) rather than by bacteriophage properties. These differences in bacteriophage sorption between the two sediments cannot be explained by bacteriophage properties. Therefore, we presume that they might be explained by the different ratios of Corg and Fe/Al contents. In sediment 2016, the high Fe/Al (and low Corg) content may provide sufficient sorption sites for PhiX174 adsorption, while sediment 2011 has a much lower Fe/Al and a higher Corg content. These results may suggest that Fe/Al sorption sites are blocked by organic matter as already described by Kaiser and Guggenberger (2003), thus limiting the amount of bacteriophages sorbed and may hence also override the effect of the isoelectric point. 4. Conclusions Even though sufficient mineral colloids were dispersed under both oxic and anoxic conditions and charge conditions were partly favourable, bacteriophages did not attach to them. The blockage of reactive sorption sites by organic matter is the most likely explanation for this observation. These results suggest that mobilized colloids do not uniformly act as carriers for viruses. Their composition and surface characteristics, which may vary under the environmental conditions, are crucial parameters which determine colloid–virus interactions.
For the elimination of bacteriophages under anoxic conditions, sorption dominates over inactivation. Inactivation was low and ranged only from 0.3 to 1.7 log10 (corresponding to an inactivation rate of 0.45 to 2.55 log10/day). In most cases, the MS2 phage showed a more sensitive response to anoxic conditions than phage PhiX174. Furthermore, the results suggest the following key mechanism: (i) In a sediment with a low organic carbon content (and high Fe/Al content), the break-up of organic coatings under anoxic conditions may render reactive sorption sites on sesquioxides available which would otherwise be blocked by organic matter, thus leading to enhanced bacteriophage removal under anoxic than under oxic conditions. (ii) However, in a sediment with high organic matter content (and low contents of sesquioxides) organic matter may block sorption sites available on sesquioxides. This in turn suggests that even if some Fe/Al-(hydr)oxides remain on the solid phase of anoxic sediments, they may be insufficient for effective bacteriophage immobilization.
Thus, the ratio between organic matter and sesquioxides may affect the impact of redox conditions on phage immobilization. To the best of our knowledge, this is the first systematic study on the fate of colloid-associated viruses in sediments under anoxic conditions. Our results do not confirm the assumption found in the literature that phage retention under anoxic conditions is generally lower than under oxic conditions (Schijven et al., 2000; Van der Wielen et al., 2008; Frohnert et al., 2014). Instead, they showed quite clearly that geochemical processes induced by decreasing redox potentials (for instance incomplete degradation of sediment-borne organic matter results in increasing DOC concentrations, which may increase DOCcomplexed Ca ions compared to free Ca ions) may have a pronounced
138
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
effect on bacteriophage distribution in the sediment–water-system. Furthermore, in such a system, the effect of pH is overruled by the dynamics of organic matter. Considering the different origins of the sediments, our findings suggest that substrates from hydrologically little connected and therefore strongly silted-up sites do not seem very conducive to bacteriophage retention and thus do not provide an efficient barrier. Furthermore, the dynamics of organic matter turnover are largely influenced by the composition of microbial communities, which may develop differently under the respective conditions in the tested sediments. The overall implication for the assessment of the risk of virus breakthrough in anoxic systems is that the dynamics of organic matter need to be included when considering the retention potential of sesquioxides for viruses. Acknowledgements The authors would like to thank Silke Meier, Teresa Bley, Iris Schilling, Sami Manandhar, Christine Arndt, Sven Zander, and Doris Pötzsch for their help with batch experiments and analysis and Andreas Farnleitner (Vienna University of Technology, Austria) for providing the sediment samples. We very much appreciated the detailed and constructive comments of one anonymous reviewer. The funding provided by the German Research Foundation (DFG; Se 508/2) is acknowledged. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2015.02.031. References Albinana-Gimenez, N., Clemente-Casares, P., Bofill-Mas, S., Hundesa, A., Ribas, F., Girones, R., 2006. Distribution of human polyomaviruses, adenoviruses, and hepatitis E virus in the environment and in a drinking-water treatment plant. Environ. Sci. Technol. 40 (23), 7416–7422. Bauer, R., Dizer, H., Graeber, I., Rosenwinkel, K.H., López-Pila, J.M., 2011. Removal of bacterial fecal indicators, coliphages and enteric adenoviruses from waters with high fecal pollution by slow sand filtration. Water Res. 45 (2), 439–452. Bethwell, C., 2004. Organische Substanz und ihre partikulären und adsorbierten Anteile im Feinsediment eines bergbaubeeinflussten Fließgewässers. In: Rücker, J., Nixdorf, B. (Eds.), Gewässerreport Nr. 8. BTUC-AR 3/2004 - Eigenverlag der BTU Cottbus, pp. 109–118 (http://www.tu-cottbus.de/fakultaet4/de/gewaesserschutz/downloads/ akt-2004.html - verified 26th February 2015). Brady, N.C., Weil, R.R., 2002. The Nature and Properties of Soils. 13th ed. Prentice-Hall, Upper Saddle River, NJ. Citeau, L., Lamy, I., van Oort, F., Elsass, F., 2001. Nature des sols et nature des colloïdes circulant dans les eaux gravitaires: une etude in situ. C. R. Acad. Sci. II A 332, 657. DIN EN ISO 10705-1, 2001. Water Quality – Detection and Enumeration of Bacteriophages – Part 1: Enumeration of F-specific RNA Bacteriophages (in German). DIN EN ISO 10705-2, 2001. Water Quality – Detection and Enumeration of Bacteriophages – Part 2: Enumeration of Somatic Coliphages (in German). Dowd, S.E., Pillai, S.D., Wang, S., Corapcioglu, M.Y., 1998. Delineating the specific influence of virus isoelectric point and size on virus adsorption and transport through sandy soils. Appl. Environ. Microbiol. 64 (2), 405–410. Emelko, M., Tufenkje, N., 2010. Transport and fate of colloids and microbes in granular aqueous environments. Water Res. 44 (4), 1027. Feng, Y.Y., Ong, S.L., Hu, J.Y., Tan, X.L., Ng, W.J., 2003. Effects of pH and temperature on the survival of coliphages MS2 and Qβ. J. Ind. Microbiol. Biotechnol. 30, 549–552. 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 — po-
tential implications for groundwater protection. Int. J. Hyg. Environ. Health 217 (8), 861–870. Hot, D., Legeay, O., Jacques, J., Gantzer, C., Caudrelier, Y., Guyard, K., Lange, M., Andréoletti, L., 2003. Detection of somatic phages, infectious enteroviruses and enterovirus genomes as indicators of human enteric viral pollution in surface water. Water Res. 37 (19), 4703–4710. ISO 11277, 1998. Soil Quality – Determination of Particle Size Distribution in Mineral Soil Material – Method by Sieving and Sedimentation. Jin, Y., Flury, M., 2001. Fate and transport of viruses in porous media. Adv. Agron. 77, 39–102. Jin, Y., Yates, M.V., Thompson, S.S., Jury, W.A., 1997. Sorption of viruses during flow through saturated sand columns. Environ. Sci. Technol. 31 (2), 548–555. Jin, Y., Pratt, E., Yates, M.V., 2000. Effect of mineral colloids on virus transport through saturated sand columns. J. Environ. Qual. 29 (2), 532–539. Kaiser, K., Guggenberger, G., 2003. Mineral surfaces and soil organic matter. Eur. J. Soil Sci. 54 (2), 219–236. Klitzke, S., Lang, F., Kaupenjohann, M., 2008. Increasing pH releases colloidal lead in a highly contaminated forest soil. Eur. J. Soil Sci. 59 (2), 265–273. Kretzschmar, R., Hesterberg, D., Sticher, H., 1997. Effects of adsorbed humic acid on surface charge and flocculation of kaolinite. Soil Sci. Soc. Am. J. 61 (1), 101–108. Kretzschmar, R., Borkovec, M., Grolimund, D., Elimelech, M., 1999. Mobile subsurface colloids and their role in contaminant transport. Adv. Agron. 66, 121–193. Lodder, W.J., De Roda Husman, A.M., 2005. Presence of noroviruses and other enteric viruses in sewage and surface waters in the Netherlands. Appl. Environ. Microbiol. 71 (3), 1453–1461. Maurice, P.A., Namjesnik-Dejanovic, K., 1999. Aggregate structures of sorbed humic substances observed in aqueous solution. Environ. Sci. Technol. 33 (9), 1538–1541. McBride, M.B., 1994. Environmental Chemistry of Soils. Oxford University Press, New York. McCarthy, J.F., Zachara, J.M., 1989. Subsurface transport of contaminants. Mobile colloids in the subsurface environment may alter the transport of contaminants. Environ. Sci. Technol. 23 (5), 496–502. McGechan, M.B., Lewis, D.R., 2002. Transport of particulate and colloid-sorbed contaminants through soil, part 1: general principles. Biosyst. Eng. 83, 255–273. Michen, B., Graule, T., 2010. Isoelectric points of viruses. J. Appl. Microbiol. 109 (2), 388–397. Mikutta, C., Lang, F., Kaupenjohann, M., 2004. Soil organic matter clogs mineral pores: evidence from 1H NMR and N2 adsorption. Soil Sci. Soc. Am. J. 68 (6), 1853–1862. Ngazoa, E.S., Fliss, I., Jean, J., 2008. Quantitative study of persistence of human norovirus genome in water using TaqMan real-time RT-PCR. J. Appl. Microbiol. 104 (3), 707–715. Patil, S.A., Hägerhäll, C., Gorton, L., 2012. Electron transfer mechanisms between microorganisms and electrodes in bioelectrochemical systems. Bioanal. Rev. 4 (2–4), 159–192. Santschi, P., Höhener, P., Benoit, G., Buchholtz-ten Brink, M., 1990. Chemical processes at the sediment–water interface. Mar. Chem. 30 (C), 269–315. Schijven, J.F., Hassanizadeh, S.M., 2000. Removal of viruses by soil passage: overview of modeling, processes, and parameters. Crit. Rev. Environ. Sci. Technol. 30 (1), 49–127. Schijven, J.F., Medema, G., Vogelaar, A.J., Hassanizadeh, S.M., 2000. Removal of microorganisms by deep well injection. J. Contam. Hydrol. 44 (3–4), 301–327. Seta, A.K., Karathanasis, A.D., 1997. Stability and transportability of water-dispersible soil colloids. Soil Sci. Soc. Am. J. 61 (2), 604–611. Stumm, W., Sigg, L., 1979. Colloid-chemical Basis of Phosphate Removal by Chemical Precipitation, Coagulation and Filtration. 12 (2), 73–83 (in German). Tanner, C.B., Jackson, M.L., 1948. Nomographs of sedimentation times for soil particles under gravity or centrifugal acceleration. Soil Sci. Soc. Am. J. 12, 60–65. Van der Wielen, P.W.J.J., Senden, W.J.M.K., Medema, G., 2008. Removal of bacteriophages MS2 and ΦX174 during transport in a sandy anoxic aquifer. Environ. Sci. Technol. 42 (12), 4589–4594. Walshe, G.E., Pang, L., Flury, M., Close, M.E., Flintoft, M., 2010. Effects of pH, ionic strength, dissolved organic matter, and flow rate on the co-transport of MS2 bacteriophages with kaolinite in gravel aquifer media. Water Res. 44 (4), 1255–1269. Weiss, W.J., Bouwer, E.J., Aboytes, R., LeChevallier, M.W., O'Melia, C.R., Le, B.T., Schwab, K.J., 2005. Riverbank filtration for control of microorganisms: results from field monitoring. Water Res. 39 (10), 1990–2001. Westrell, T., Teunis, P., van den Berg, H., Lodder, W., Ketelaars, H., Stenström, T.A., de Roda Husman, A.M., 2006. Short- and long-term variations of norovirus concentrations in the Meuse river during a 2-year study period. Water Res. 40 (14), 2613–2620.
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Attenuation and colloidal mobilization of bacteriophages in natural sediments under anoxic as compared to oxic conditions Sondra Klitzke a,⁎, Jendrik Schroeder a, Hans-Christoph Selinka b, Regine Szewzyk b, Ingrid Chorus c a b c
Federal Environment Agency, Section Drinking Water Treatment and Resource Protection, Schichauweg 58, D-12307 Berlin, Germany Federal Environment Agency, Section Microbiological Risks, Corrensplatz 1, D-14197 Berlin, Germany Federal Environment Agency, Department of Drinking Water and Swimming Pool Hygiene, Schichauweg 58, D-12307 Berlin, Germany
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• MS2 and PhiX174 did not attach to colloids irrespective of pH and redox potential. • Phage removal efficiency in anoxic environments depends on organic matter dynamics. • Organic matter may block Fe/Al-oxides in anoxic sediments.
a r t i c l e
i n f o
Article history: Received 27 November 2014 Received in revised form 6 February 2015 Accepted 8 February 2015 Available online 5 March 2015 Editor: D. Barcelo Keywords: Colloid mobilization Phage Riverbank filtration Sesquioxide Virus Zeta potential
a b s t r a c t Redox conditions are known to affect the fate of viruses in porous media. Several studies report the relevance of colloid-facilitated virus transport in the subsurface, but detailed studies on the effect of anoxic conditions on virus retention in natural sediments are still missing. Therefore, we investigated the fate of viruses in natural flood plain sediments with different sesquioxide contents under anoxic conditions by considering sorption to the solid phase, sorption to mobilized colloids, and inactivation in the aqueous phase. Batch experiments were conducted under oxic and anoxic conditions at pH values between 5.1 and 7.6, using bacteriophages MS2 and PhiX174 as model viruses. In addition to free and colloid-associated bacteriophages, dissolved and colloidal concentrations of Fe, Al and organic C as well as dissolved Ca were determined. Results showed that regardless of redox conditions, bacteriophages did not adsorb to mobilized colloids, even under favourable charge conditions. Under anoxic conditions, attenuation of bacteriophages was dominated by sorption over inactivation, with MS2 showing a higher degree of sorption than PhiX174. Inactivation in water was low under anoxic conditions for both bacteriophages with about one log10 decrease in concentration during 16 h. Increased Fe/Al concentrations and a low organic carbon content of the sediment led to enhanced bacteriophage removal under anoxic conditions. However, even in the presence of sufficient Fe/A-(hydr)oxides on the solid phase, bacteriophage sorption
Abbreviations: PFU, plaque forming units; Corg, organic carbon; DOC, dissolve organic carbon. ⁎ Corresponding author. E-mail addresses: [email protected] (S. Klitzke), [email protected] (J. Schroeder), [email protected] (H.-C. Selinka), [email protected] (R. Szewzyk), [email protected] (I. Chorus).
http://dx.doi.org/10.1016/j.scitotenv.2015.02.031 0048-9697/© 2015 Elsevier B.V. All rights reserved.
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
131
was low. We presume that organic matter may limit the potential retention of sesquioxides in anoxic sediments and should thus be considered for the risk assessment of virus breakthrough in the subsurface. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Human pathogenic viruses which are transmitted via the faecal–oral route, such as enteric viruses, are frequently found in surface water (Hot et al., 2003; Lodder and De Roda Husman, 2005; Albinana-Gimenez et al., 2006; Ngazoa et al., 2008; Westrell et al., 2006). Riverbank filtration, artificial groundwater recharge and slow sand filtration are common methods for the natural (pre-)treatment of drinking water. Riverbank filtration and slow sand filtration have both been reported to remove viruses or virus surrogates (i.e. model viruses such as bacteriophages MS2 or PhiX174) during subsurface passage (Weiss et al., 2005; Bauer et al., 2011). The elimination of viruses is thereby mainly governed by hydraulic settings and physico-chemical parameters of water and sediment along with the surface characteristics of viruses itself (Jin and Flury, 2001). Redox conditions are reported to play a major role in the transport and inactivation of bacteriophages (Schijven et al., 2000; Van der Wielen et al., 2008). Both authors observed lower virus removal in anoxic than in oxic zones of a natural sandy aquifer which they attributed to lower inactivation and adsorption rates under anoxic conditions. They attribute the reduced virus adsorption to the dissolution of Fe(hydr)oxides. Similarly, Frohnert et al. (2014) observed lower virus removal under anoxic conditions in a laboratory column experiment during a 4-week-period. A number of studies indicate the relevance of colloid-associated virus transport along preferential flow paths in sediments (Seta and Karathanasis, 1997; Jin et al., 2000). Colloids are small (1 nm to 10 μm; Brady and Weil, 2002; Stumm and Sigg, 1979) and have a large surface to volume ratio (Kretzschmar et al., 1999). Therefore, colloids tend to be stable in suspensions and are hardly filtered by porous media (Seta and Karathanasis, 1997). These properties make colloids a possible carrier for sorbing viruses (McGechan and Lewis, 2002) which would otherwise be retained by the sediment. Jin et al. (2000) showed that colloid-facilitated MS2-phage transport in porous media was clearly correlated with the amount of transported clay colloids and the number of viruses adsorbed to these. Failure to consider this pathway may result in an underestimation of virus travel velocity, distances and concentrations (McCarthy and Zachara, 1989). To predict the fate and transport of viruses in contaminated waters and sediments, it is important to understand the dynamics of colloid release from sediments and the governing mechanisms of colloid–virus-interactions. Sorption of viruses to colloids and sediments is controlled by their surface properties, with the surface charge being a key parameter. The charge of individual viruses at a given pH is a function of their isoelectric point (IEP), which varies for different virus types (Jin et al., 1997; pHiep: 3.9 for MS2, and pHiep 6.6 for PhiX174; Dowd et al., 1998). Hence, virus charge is controlled by pH and ionic strength of the solution. Similarly, colloid charge is also a function of IEP (determined by colloid composition) and solution chemistry, such as pH, concentrations of dissolved organic carbon and ionic strength (Kretzschmar et al., 1999). Emelko and Tufenkje (2010) highlight the need to understand the fate of colloidal pathogens such as viruses in the subsurface (i.e. riverbank filtration systems). However, so far, only few studies looked into the fate of viruses under anoxic conditions (for instance Frohnert et al., 2014). There are many publications on virus retention as a function of pH (Schijven and Hassanizadeh, 2000; Jin and Flury, 2001; Walshe et al., 2010), but they all refer to oxic conditions. In addition, the role of colloid-facilitated virus mobilization or virus transport has never been addressed in this context. Moreover, sesquioxides (i.e. Feand Al-(hydr)oxides), which often occur in close association with each other, and their role on virus retention under anoxic conditions is still
unclear: non-redox-sensitive sorption sites such as Al-(hydr)oxides could improve virus retention under all redox conditions, but if they are released from the sediment due to break-up of bonds through the dissolution of Fe they may also enhance virus transport if viruses sorb to them. Therefore, the aim of our study was to understand (i) the dynamics of colloid release from natural flood plain sediments with different sesquioxide contents under anoxic conditions as compared to oxic conditions. (ii) the fate of viruses in these natural system such as sorption to the solid phase, sorption to mobilized colloids, and inactivation in the aqueous phase by testing the influence of pH changes on the attenuation of two model viruses, PhiX174 and MS2, in laboratory batch experiments. These two bacteriophages were chosen as they differ in their isoelectric points (Dowd et al., 1998), which is reported to lead to different sorption behaviour (Jin and Flury, 2001). 2. Material and methods 2.1. Origin and characterization of sediments and water Two sediments were obtained from the Danube floodplain. Sediment 2011 was sampled in an isolated river section with little hydrological connection to the Danube River. This section was silted up with medium-grained fine sand and overgrown with reed and riverine vegetation. Sediment 2016 was sampled in a river section with a pronounced hydrological connection to the Danube, characterized by regular erosion and sedimentation processes and a texture of silty sand. Both samples were mixed and homogenized individually prior to their storage at 4 °C under saturated conditions (i.e. the sediment was always covered with a few centimetres of river water). Both sediments differ strongly in their concentration of organic carbon, Fe and Al (Table 1): sediment 2011 has a high organic carbon (Corg) and a low Fe/Al content, sediment 2016 has a low Corg and a high Fe/Al content. For our experiments we used treated groundwater (i.e. by removal of Fe and Mn through microbial precipitation) abstracted from a quaternary aquifer on the experimental field site of the Federal Environment Agency in BerlinMarienfelde, Germany. In order to obtain water of low ionic strength, this groundwater was diluted 1:10 with deionised water, yielding a final ionic strength of about 2 mM, comprising Ca2+ and SO2− concen4 trations of about 0.40 mM and 0.25 mM, respectively. The water was autoclaved (20 min at 121 °C) prior to its use to remove any dissolved oxygen and is termed “test water” in the following. 2.2. Bacteriophages Bacteriophages MS2 (DSM 13767) and PhiX174 (DSM 4497) were used as surrogates of RNA and DNA viruses, respectively. Both viruses develop icosahedral virions with sizes of 25–30 nm. MS2 is an RNA bacteriophage of the family Leviviridae with a single-stranded RNA genome, the genome of phage PhiX174 (family Microviridae) consists of singlestranded DNA. The isoelectric points given in literature are about 3.9 for phage MS2 and about 6.6 for phage PhiX174 (Michen and Graule, 2010). To achieve stock suspensions with high bacteriophage concentrations, host bacteria were grown up to exponential growth phase and subsequently inoculated with the respective phages, as described in standardized procedures (DIN EN ISO, 10705-1, 2001; DIN EN ISO, 10705-2, 2001). 10% v/v chloroform was added to destroy the bacteria. After incubation and settling, supernatants were centrifuged (3000 ×g
132
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
Table 1 Investigated parameters of the tested sediments of the Danube River.
Location pH Electrical conductivity Organic carbon Inorganic carbon Oxalat-extractable Fe Oxalat-extractable Al Dithionite-extractable Fe Dithionite-extractable Al
[μS/cm] [mg/kg] [g/kg] [mg/kg] [mg/kg] [mg/kg] [mg/kg]
Sediment 2011
Sediment 2016
Isolated wetland (reed and coastal vegetation); strong aggradation 7.6 139.7 17,938 13.3 293 30 395 116
Wetland with connection to Danube River; frequent erosion and sedimentation 7.9 146.6 4,963 26.3 1866 75 2,952 237
for 5 min) to separate bacteria residues from bacteriophage particles. Supernatants of centrifugations were frozen in aliquots at (−80 ± 10) °C until use. Phage concentrations amounted to approximately 1012 pfu/mL (MS2) and about 1010 pfu/mL (PhiX174), respectively. Stock suspensions of both bacteriophages were individually diluted in decimal steps using peptone saline (8.5 g/L NaCl, 1 g/L peptone) up to a concentration of 109 pfu/mL. Further dilution was performed in test water to avoid influence of residual peptone on the experiments. Test water was inoculated with bacteriophages from the diluted suspensions to yield concentrations of approximately 107 pfu/mL for both bacteriophages. These suspensions in test water were used for inoculation of the batch samples described below. For anoxic experiments, peptone saline and the test water were purged with nitrogen gas for 15 min after autoclaving and transferred into the nitrogenflushed glove box where the phage suspensions were prepared. To determine the initial phage concentrations, aliquots of these suspensions were analysed according to DIN EN ISO 10705-1 (2001) and DIN EN ISO 10705-2 (2001). 2.3. Design of batch experiments We used field-moist sediment and set up triplicate batch series at 3 to 4 different pH values (see below; oxic experiments: plastic tubes, BD; anoxic experiments: glass flasks, Schott) with a soil to water ratio of 1:10, i. e. 8 g sediment mass and a total suspension volume of 80 mL. This ratio proved to be wide enough to prevent any artificial generation of colloids by abrasion (Klitzke et al., 2008). For the comparison of different pH values, appropriate amounts of dilute nitric acid (oxic experiments) or dilute hydrochloric acid (previously purged with N2; anoxic experiments), respectively were added. The following pH values were adjusted: 5.5 ± 0.4, 6.5 ± 0.1, 7.3 ± 0.1, 7.6 ± 0.1 for oxic experiments, and 5.1 ± 0.1, 6.5 ± 0.2, 7.3 ± 0.3 for anoxic experiments, with the latter ones being the natural pH values of the sediments. Batch samples i.e. 8 g of sediment and 70 mL of test water, were left to stand for a day at room temperature to allow equilibration following acid addition. In the anoxic experiments, samples were purged with nitrogen for 15 min prior to their transfer into the glovebox. 2.3.1. Oxic experiments with sediment We added 10 mL of phage suspension (c = 107 pfu/mL; see Section 2.2.) to each batch sample at the end of the equilibration period to yield a final phage concentration of approximately 106 pfu/mL and a
total sample volume of 80 mL. Samples were put on an incubating shaker (Thermoshake, Gerhardt) at 110 rpm for 16 h. The pH values were checked before and after the shaking period (Table 2). To separate material larger than 2 μm, samples were centrifuged at 100 g for 4 min (based on the assumption of an average density of 2.65 g/cm3, i.e. corresponding to the density of quartz and representing an estimated average of organic and mineral material; ISO, 11277, 1998). The centrifugation parameters were chosen in accordance with Stoke's Law (Tanner and Jackson, 1948). The removal of particles N 2 μm was confirmed using dynamic light scattering measurements. Supernatants were used to determine total concentrations (i. e. the sum of (i) dissolved and colloidal element concentrations, (ii) “free” and colloid-associated bacteriophage concentrations) of organic C, Fe, Al, Mn as well as bacteriophages. An aliquot was centrifuged (Avanti J-26 XP Centrifuge, Beckman Coulter) at 6800 g for 80 min to remove any colloidal material larger than 100 nm, assuming an average density of 1.5 g/cm3 to remove as much as possible of the suspended material. This low density was chosen as organic colloids may have a density much below the average (1.1 g/cm3 (Citeau et al., 2001), 1.6 g/cm3 (Bethwell, 2004). The supernatants of these centrifugates were defined as “dissolved” for the determined elements, and as “free” for the respective bacteriophages. The differences of the metal, C and bacteriophage concentrations between these two supernatants (i.e. total and dissolved/“free”) account for colloid-associated fractions (defined operationally). Organic C, Fe, Al, and Ca, concentrations of both phages, optical density and the zeta potentials were determined as described in the section analysis (see Section 2.4.). 2.3.2. Anoxic experiments with sediment The samples were transferred into a nitrogen-flushed glove box (O2 content below 0.5–1%) through an airlock and left to stand for 14 days (sediment 2016) or 31 days (sediment 2011) to allow the redox potential to drop in consequence of the mineralization of organic matter. Initial redox potentials of the batch samples were −120 ± 53 mV (sediment 2016) and −140 ± 36 (sediment 2011). Aliquots of phage suspensions in test water (see Section 2.2) were added to each sample at the end of the equilibration period to yield a final phage concentration of 106 pfu/mL and a total sample volume of 80 mL. pH and redox potential were checked and, if necessary, pH was re-adjusted (Table 3). Samples were closed with an airtight lid and put on an incubating shaker (Thermoshake, Gerhardt) at 110 rpm for 16 h. Redox potentials remained fairly constant during shaking, i.e. in most cases they did not
Table 2 pH values before and after shaking of the sediments under oxic conditions (n.d. not determined). Sediment 2016 Target pH pH Redox potential
Before shaking After shaking Before shaking After shaking
5.5 5.0 5.4 339 431
Sediment 2011 6.5 6.2 6.4 355 415
7.3 7.0 7.4 355 393
7.6 7.8 7.7 n. d. n. d.
5.5 5.4 5.9 356 382
6.5 6.2 6.4 356 395
7.3 6.9 7.3 355 382
7.6 7.6 7.5 n. d. n. d.
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
differ by more than 20 mV (Fig. Supp1a and Supp1b in the Supporting material). Subsequently, samples were re-introduced into the glove box, followed by checks of pH and redox potentials (Table 3). Samples were transferred into centrifuge tubes and centrifuged at conditions specified in Section 2.3.1. For the removal of the supernatants after each centrifugation step samples were put back into the glove box. Organic C, Fe, Al, and phage concentrations were determined as described in Section 2.4.
2.3.3. Determination of bacteriophage distribution in sediment samples Concentrations of phages were determined in the initial suspensions in test water (see Section 2.2.) and in the supernatants of the two different centrifugation steps described in Section 2.3.1. Bacteriophage concentrations in the supernatant of the first centrifugation step (containing only particles smaller than 2 μm) represent suspended bacteriophages following the removal of larger sediment material and comprise both “free” and colloid-associated bacteriophages (termed as “total” in the following) while bacteriophages in the supernatant of the second centrifugation step (containing only particles smaller than 100 nm) comprise only “free” bacteriophages (termed as “free” in the following). Differences in bacteriophage concentrations between the initial suspensions and the supernatants of the first centrifugation step (b2 μm) may be ascribed to bacteriophage sorption to larger sediment material and/ or to inactivation. Differences in concentrations between the supernatants of the first centrifugation step (i.e. b 2 μm) and the second centrifugation step (i. e. b 100 nm) may be ascribed to the concentration of colloid-associated bacteriophages.
2.3.4. Anoxic experiments without sediment To differentiate between the two key processes of bacteriophage elimination – sorption and inactivation – an additional set of batch samples containing only anoxic pore water was prepared to determine survival of phages. Samples were prepared in triplicate as described in Section 2.3. and subsequently transferred into a nitrogen-flushed glovebox, where they were left to stand until the redox potentials dropped below 0 mV (incubating time: sediment 2016: 28 days (− 3 to −88 mV); sediment 2011: 5 days (−72 to −136 mV)). To remove any possible sorption sites the samples were transferred into centrifuge tubes, sealed with a lid, centrifuged at 6800 g for 80 min (cut-off 100 nm, based on an assumed density of 1.5 g/cm3). Supernatants were removed in the glovebox and 35 mL of each supernatant was inoculated with 5 mL of phage suspension. The flasks were closed with an airtight lid and put on an incubating shaker outside the glovebox at 110 rpm of 16 h. Afterwards, samples were re-introduced into the glovebox for subsequent check of pH and redox potential (5.5 ± 0.1, 6.5 ± 0.1 (both sediments), 7.3 (sediment 2011) and 8.2 (sediment 2016); 272 ± 14 mV (sediment 2011) and 186 ± 21 mV (sediment 2016)). The measured redox potential in the pore water was higher than in the sediment–water system, which may be ascribed to concentration effects of redox-sensitive ions in the suspension caused by the removal of the sediment (Santschi et al., 1990). Assuming no bacteriophage attachment to colloids, the combination of the results obtained in the absence and in the presence of sediments allows for the quantification of the total elimination (elimtot) according to Eq. (1), with sorptsolid being the sorption to the solid phase and Table 3 pH values and redox potentials before and after shaking of the sediments under anoxic conditions. Sediment 2016 Target pH pH
Sediment 2011
5.1 6.5 7.3 5.1 6.5 7.3 Before shaking 5.0 6.4 7.2 5.0 6.2 7.7 After shaking 5.0 6.6 7.0 5.1 6.7 7.5 Redox potential Before shaking −2 −94 −179 −41 −157 −147 After shaking 27 −60 −150 −11 −166 −167
133
inactaq being the inactivation in the aqueous phase: Elimtot ¼ sorptsolid þ inactaq :
ð1Þ
2.4. Analysis Organic C was measured with a vario TOC cube (Elementar). Iron, Al, and Ca concentrations of centrifugates were analysed in diluted and acidified samples (3% HNO3) by ICP-OES. Optical density determined as light absorption at 525 nm of the supernatant from the first centrifugation step (b 2 μm) of oxic samples was taken as a measure for the relative amount of dispersible particles as specified by Kretzschmar et al. (1997) on a spectrophotometer (Spektrophotometer PM6, Zeiss, Germany). As colloid-free solutions also showed some light absorbance at 525 nm (presumably caused by dissolved organic matter) we corrected for this interfering absorbance by subtracting the absorbance of the supernatant from the second centrifugation step (b 100 nm). Zeta potential and size of colloids were determined in the (bacteriophagefree) suspensions obtained from sediment suspended in the same matrix as samples containing bacteriophages. Zeta potential was calculated from the electrophoretic mobility (using the Helmholtz–Smoluchowski equation), which was analysed by a Zetasizer 2000 photon correlation spectrometer (Malvern Instruments). Particle size in the supernatant of bacteriophage-containing and -free suspensions was determined by dynamic light scattering (DLS, Malvern Instruments). The bacteriophages MS2 and PhiX174 were quantified by a double layer agar method according to DIN EN ISO 10705-1 (2001) and DIN EN ISO 10705-2 (2001). Each sample was analysed in duplicates. Samples with higher concentrations of bacteriophages were diluted in decimal steps using peptone saline (8.5 g/L NaCl, 1 g/L peptone). Detection limits were 1 pfu/mL for 1 mL sample volumes and 0.2 pfu/mL for 5 mL volumes, respectively. 3. Results and discussion 3.1. Experiments performed under oxic conditions Under oxic conditions, batch experiments with sediments were conducted at different pH values to determine the distribution of bacteriophages between (i) larger sediment particles and (ii) colloids as well as the amount of free bacteriophages. 3.1.1. Colloid properties The optical density of both sediments increased with increasing pH (Fig. Supp2 in the Supporting material), suggesting a mobilization of sediment colloids with increasing pH. This observation is supported by higher colloidal concentrations of organic C (Fig. 1a), and Fe, and Al (Fig. 1b), which increase with increasing pH (except for organic C in sediment 2016; Fig. 1a). The zeta potential, a relative measure of the particle charge, clearly decreased with increasing pH (Fig. 2). The resulting repulsion between negatively charged colloids leads to a stronger mobilization of colloids. Both findings are in accordance with classical theories of colloidal behaviour described in the literature (Kretzschmar et al., 1999). Sizes of dispersed pedogenic colloids ranged between 259 and 1530 nm (sediment 2016) and between 477 and 1760 nm (sediment 2011). Size distribution of dispersed particles in the supernatant (b 2 μm) following bacteriophage addition ranged between 256 and 1820 nm (sediment 2016) and between 466 and 1850 nm (sediment 2011). This suggests that the addition of bacteriophages does not affect the size distribution of dispersed colloids. 3.1.2. Fate of bacteriophages in sediment batch samples At the end of the batch experiments, in both sediments at pH 7.3 and 7.6, total phage concentrations showed the same concentrations as in initial suspensions (Fig. 3a and b, e and f), suggesting no phage sorption
134
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
Total and free concentrations of both bacteriophages did not markedly differ from each other for both sediments at all investigated pH values (e.g. Fig. 3a and b, e and f) suggesting no binding of bacteriophages to dispersed colloids. This non-attachment is explainable for samples with equal charges of colloids (Fig. 2) and bacteriophages, i.e. NpH 3.9 for MS2 and NpH 6.6 for PhiX174, as bacteriophages are then negatively charged and thus repelled from negatively charged colloids. It remains unclear why there is no bacteriophage attachment to colloids when charges are attractive, i.e. at pH 6 to 6.6 for PhiX174 (for both sediments) or pH b 6 for MS2 (only sediment 2016). One reason could be the association of organic matter. Both sediments contain considerable amounts of organic matter, which often acts as partial coating of mineral phases (McCarthy and Zachara, 1989). These coatings may block potential sorption sites for viruses on dispersed colloids (Schijven and Hassanizadeh, 2000) and hence prevent attachment. 3.2. Experiments performed under anoxic conditions
Fig. 1. pH-dependent colloidal concentrations of organic carbon (a), as well as Fe and Al (b) in oxic batch samples of sediments 2011 and 2016.
to sediment material and/or inactivation. However, at pH 6.5, total concentrations of both phages in the supernatants were 2 to 3 log10 lower than in initial suspensions in sediment 2011 (Fig. 3e and f), indicating substantial sorption to sediment material and/or inactivation. This effect was even more pronounced for sediment 2016, in which total phage concentrations were reduced in the supernatant to a level below the limit of detection (Fig. 3a and b), corresponding to a removal through sorption to sediment material and/or inactivation of at least 5 to 6 log10. The higher removal in sediment 2016 may be ascribed to the higher Ca concentrations compared to sediment 2011 at pH 6.5 (Fig. 4), leading to enhanced phage aggregation due to a compression of the thickness of the diffuse double layer (Walshe et al., 2010), which may facilitate sorption. An enhanced phage removal at elevated Ca concentrations has also been observed by Walshe et al. (2010). At pH 5.5 (Fig. 3a and b, e and f), total phage concentrations were below the quantification limit for both sediments, suggesting complete or almost complete removal through sorption to sediment material and/or inactivation — with sorption being presumably facilitated by high Ca concentrations (Fig. 4).
Fig. 2. Zeta potentials of colloids dispersed at various pH values under oxic conditions (without the addition of bacteriophages). Bars depict one standard deviation.
Under anoxic conditions, batch experiments with sediments were conducted at different pH values to determine the distribution of bacteriophages between (i) larger sediment particles and (ii) colloids as well as the amount of free bacteriophages. Furthermore, the survival of bacteriophages was determined in batch samples using anoxic pore water. Up to now, there are only few studies which look into the fate of viruses in the subsurface under anoxic conditions (Van der Wielen et al., 2008; Schijven et al., 2000), which are based on field-scale observations. Therefore, our investigations aiming at a mechanistic understanding of the fate of viruses close an important knowledge gap. Redox potentials in anoxic batch samples of both sediments did not show any considerable changes (i.e. smaller than 20 mV in most cases) before and after shaking (Fig. Supp1a and Supp1b in the Supporting material) confirming that no oxygen penetration was possible in our experimental setup outside the glove box. 3.2.1. Colloids in sediment batch samples While Al colloids were only dispersed at very low concentrations, both sediments showed elevated concentrations of dispersed Fe colloids (Fig. Supp3 in the Supporting material). However, in most cases, bacteriophages did not attach to them (see also Sections 3.2.2. and 3.2.3.). Colloidal Fe concentrations under anoxic conditions did not differ from concentrations under oxic conditions (data not shown) at pH 5.1/5.5 (both sediments; both p-values 0.44) and pH 6.5 (sediment 2011; p-value 0.63). At pH 7.3 (both sediments; p-value 0.04) and pH 6.5 (sediment 2016; p-value 0.07) concentrations were slightly higher in anoxic samples. This is surprising as one could assume the dissolution of dispersed Fe colloids under anoxic conditions. One reason explaining our observation could be the blockage of Fe surfaces by organic matter as mineral colloids are often coated with organic matter (Kretzschmar et al., 1999). Organic matter is reported to reduce the specific surface area of minerals (Kaiser and Guggenberger, 2003) and thus to lead to ‘pore clogging’ of mineral phases (Mikutta et al., 2004). Therefore, we propose the following conceptual model as most likely mechanism which could have prevented the reduction of Fe: organic matter may have shielded the Fe oxide surfaces from microorganisms or even from small redox-reactive organic molecules (so called “redox mediators”) released by microorganisms (Patil et al., 2012). 3.2.2. Fate of bacteriophages in sediment 2016 batch samples At pH 7.3, total concentrations of both bacteriophages were about 2 to 3 log10 lower than in initial suspension (Fig. 3c and d), respectively, suggesting sorption to sediment material and/or inactivation under anoxic conditions. This is in contrast to the results under oxic conditions where no sorption to sediment material and/or inactivation was detected at this pH. We suggest the following concept for enhanced sorption under anoxic conditions: Initially, in these sediments, organic matter was present as organic coating of the sediment. Such coatings are
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
135
Fig. 3. Bacteriophage concentrations in the supernatants of oxic and anoxic batch samples of sediments 2016 and 2011: bacteriophages PhiX174 (a, c, e, g) and MS2 (b, d, f, h) under oxic (a, b, e, f) and anoxic (c, d, g, h) conditions. Note that in sediment 2016 oxic experiments were also conducted at pH 6.5, but concentrations were below the limit of detection. Dotted lines: initial phage concentrations of PhiX174 in the batch sample; solid lines: initial phage concentration of MS2 in the batch samples. Bars depict one standard deviation. n.d.: not determined.
reported to cover mineral surfaces (Maurice and Namjesnik-Dejanovic, 1999), blocking their reactive sites if organic matter concentrations are low (Kaiser and Guggenberger, 2003). The increase in DOC concentrations (Fig. 5) in comparison to oxic conditions gives rise to the assumption that during the pre-incubation period used to attain anoxic conditions, such coatings were in part subject to mineralization. Their (partial) break-up may thus render reactive sorption sites on the mineral phase accessible, leading to enhanced retention of bacteriophages. As the organic
carbon content of this sediment was low, the effect may be relevant. Even though dissolved Fe concentrations showed a tendency to be higher in anoxic than in oxic samples (Fig. Supp4 in the Supporting material) – suggesting the dissolution of some of the Fe-oxides – enough binding sites probably remained, as concentrations of Fe and Al are high in this sediment. This model remains to be tested by future research. At pH 6.5 under anoxic conditions, PhiX174 showed a similar pattern as at pH 7.3 i.e. a 2 log10 decrease in concentration, which may be
136
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
suggest that the dissolution of Fe decreases the number of available Fe sorption sites, thereby further diminishing the already low Fe content of the sediment. However, even though there were still considerable amounts of Fe left on sediment 2011 following Fe reduction, bacteriophage retention on this sediment was low (1 to 2 log10). Therefore, we presume that the high organic carbon content may block sorption sites available on sesquioxides as discussed above for sediment 2016. As expected, dissolved Al concentrations in oxic and unoxic samples did not differ from each other (except for at pH 5.1/5.5; Fig. 6). Similarly to the effects observed with sediment 2016, concentrations between total and free bacteriophages did not differ from each other in any samples, i.e. bacteriophages did not adsorb to colloids.
Fig. 4. Dissolved Ca concentrations in oxic sediment batch samples. Bars depict one standard deviation.
attributed to sorption to sediment/inactivation (Fig. 3c). This is in contrast to the experiments under oxic conditions (pH 6.5) when PhiX174 is completely eliminated due to sorption to sediment/inactivation (Fig. 3a). This pH is close to the isoelectric point of PhiX174, at which bacteriophages are more prone to aggregation. Under oxic conditions, the high Ca concentration (Fig. 4) may have facilitated additional aggregation. As DOC is known to complex Ca ions (McBride, 1994), the increased DOC release under anoxic conditions (Fig. 5) may weaken the aggregating effect of the Ca ion. The same concept may also be ascribed to both phages at pH 5.1 (oxic)/pH 5.5 (anoxic), respectively. MS2 was eliminated by 6 log10 units under both oxic and unoxic conditions at pH 6.5 (Fig. 3b and d). Concentrations of total and ‘free’ bacteriophages did not show marked differences between any of the samples, suggesting that bacteriophages did not adsorb to colloids. 3.2.3. Fate of bacteriophages in sediment 2011 batch samples At pH 7.3, total concentrations of both bacteriophages were nearly as high as initial concentrations (Fig. 3g and h), indicating that there was no substantial sorption to sediment material and/or inactivation under anoxic conditions. A similar behaviour was detected under aerobic conditions. At this pH, there was no difference in DOC concentrations between oxic and anoxic samples (data not shown), suggesting no additional DOC mineralization. Alternatively, the long incubation time (31 days) needed to render samples anoxic may have lead to a high degree of DOC mineralization. At pH 6.5 elimination by sorption to sediment material and/or inactivation of PhiX174 was higher than at pH 7.3 (about 0.5 log10 units), while MS2 did not show any significant difference. The elimination of both phages was lower than under oxic conditions (2 to 3 log10). Higher dissolved Fe-concentrations in anoxic than in oxic samples (Fig. 6)
Fig. 5. DOC concentrations in the supernatants of oxic and anoxic samples (sediment 2016 — low Corg, high Fe/Al content). Bars depict one standard deviation.
3.2.4. Fate of bacteriophages in pore water samples obtained from sediment 2016 Concentrations of PhiX174 and MS2 in the supernatants of anoxic pore water (see Section 2.3.4.) decreased by 1 to 1.5 log10 at pH 5.5 and 6.5 during 16 h of incubation (data not shown). This shows that inactivation of phages was low. At pH 8.2, concentrations of both bacteriophages were as high as in initial suspensions (data not shown), suggesting no bacteriophage inactivation. Even though the pH in these samples was higher than in samples containing sediment (i.e. pH 7.3), one may conclude that bacteriophages show equally good survival in anoxic pore water at pH 7.3, as survival is normally worse at alkaline than at neutral pH (Feng et al., 2003). As a consequence one may conclude that MS2 removal in the sediment–water-system is dominated by sorption to sediment material, inactivation in water only plays a minor role (Fig. 7a; Eq. (1)). Sorption was highest at pH 6.5. For PhiX174 inactivation is relatively more important due to the lower overall removal of only 2 log10 due to sorption on sediments/inactivation. PhiX174 removal at pH 5.1 is mainly attributed to inactivation, with sorption becoming more dominant as pH increases (Fig. 7b; Eq. (1)). The increase in sorption with increasing pH may be explained by the isoelectric point (Jin and Flury, 2001). 3.2.5. Fate of bacteriophages in pore water samples obtained from sediment 2011 Concentrations of PhiX174 in the supernatants of anoxic pore water (see Section 2.3.4) remained more or less constant over the 16 h incubation at all investigated pH values. This shows that inactivation was very low (data not shown). Concentrations of MS2 in the supernatants decreased slightly by 0.6 to 1 log10 (data not shown), demonstrating limited inactivation.
Fig. 6. Dissolved Fe and Al concentrations in the supernatants of oxic (+) and anoxic (−) samples (sediment 2011 — high Corg, low Fe/Al content). Bars depict one standard deviation.
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
137
Fig. 7. Removal processes for bacteriophages MS2 and PhiX174 under anoxic conditions in sediment 2016 (a, b) and in sediment 2011 (c, d). Bars depict standard deviations.
Regarding the overall removal, inactivation seems to play the main role in MS2 removal at pH 6.5 and 7.3, but sorption dominates at pH 5.1 (Fig. 7c; Eq. (1)), i.e. the pH closest to the isoelectric point. For PhiX174 removal, however, inactivation is of minor importance: elimination is mainly governed by sorption, which decreases with increasing pH (Fig. 7d; Eq. (1)). Based on the isoelectric point of PhiX174, one would assume an opposite sorption behaviour (Jin and Flury, 2001). This is contrary to the fate of PhiX174 in sediment 2016. For MS2, the fraction removed through sorption in this sediment is lower than in sediment 2016, suggesting that MS2 retention is predominantly governed by sediment properties (see below) rather than by bacteriophage properties. These differences in bacteriophage sorption between the two sediments cannot be explained by bacteriophage properties. Therefore, we presume that they might be explained by the different ratios of Corg and Fe/Al contents. In sediment 2016, the high Fe/Al (and low Corg) content may provide sufficient sorption sites for PhiX174 adsorption, while sediment 2011 has a much lower Fe/Al and a higher Corg content. These results may suggest that Fe/Al sorption sites are blocked by organic matter as already described by Kaiser and Guggenberger (2003), thus limiting the amount of bacteriophages sorbed and may hence also override the effect of the isoelectric point. 4. Conclusions Even though sufficient mineral colloids were dispersed under both oxic and anoxic conditions and charge conditions were partly favourable, bacteriophages did not attach to them. The blockage of reactive sorption sites by organic matter is the most likely explanation for this observation. These results suggest that mobilized colloids do not uniformly act as carriers for viruses. Their composition and surface characteristics, which may vary under the environmental conditions, are crucial parameters which determine colloid–virus interactions.
For the elimination of bacteriophages under anoxic conditions, sorption dominates over inactivation. Inactivation was low and ranged only from 0.3 to 1.7 log10 (corresponding to an inactivation rate of 0.45 to 2.55 log10/day). In most cases, the MS2 phage showed a more sensitive response to anoxic conditions than phage PhiX174. Furthermore, the results suggest the following key mechanism: (i) In a sediment with a low organic carbon content (and high Fe/Al content), the break-up of organic coatings under anoxic conditions may render reactive sorption sites on sesquioxides available which would otherwise be blocked by organic matter, thus leading to enhanced bacteriophage removal under anoxic than under oxic conditions. (ii) However, in a sediment with high organic matter content (and low contents of sesquioxides) organic matter may block sorption sites available on sesquioxides. This in turn suggests that even if some Fe/Al-(hydr)oxides remain on the solid phase of anoxic sediments, they may be insufficient for effective bacteriophage immobilization.
Thus, the ratio between organic matter and sesquioxides may affect the impact of redox conditions on phage immobilization. To the best of our knowledge, this is the first systematic study on the fate of colloid-associated viruses in sediments under anoxic conditions. Our results do not confirm the assumption found in the literature that phage retention under anoxic conditions is generally lower than under oxic conditions (Schijven et al., 2000; Van der Wielen et al., 2008; Frohnert et al., 2014). Instead, they showed quite clearly that geochemical processes induced by decreasing redox potentials (for instance incomplete degradation of sediment-borne organic matter results in increasing DOC concentrations, which may increase DOCcomplexed Ca ions compared to free Ca ions) may have a pronounced
138
S. Klitzke et al. / Science of the Total Environment 518-519 (2015) 130–138
effect on bacteriophage distribution in the sediment–water-system. Furthermore, in such a system, the effect of pH is overruled by the dynamics of organic matter. Considering the different origins of the sediments, our findings suggest that substrates from hydrologically little connected and therefore strongly silted-up sites do not seem very conducive to bacteriophage retention and thus do not provide an efficient barrier. Furthermore, the dynamics of organic matter turnover are largely influenced by the composition of microbial communities, which may develop differently under the respective conditions in the tested sediments. The overall implication for the assessment of the risk of virus breakthrough in anoxic systems is that the dynamics of organic matter need to be included when considering the retention potential of sesquioxides for viruses. Acknowledgements The authors would like to thank Silke Meier, Teresa Bley, Iris Schilling, Sami Manandhar, Christine Arndt, Sven Zander, and Doris Pötzsch for their help with batch experiments and analysis and Andreas Farnleitner (Vienna University of Technology, Austria) for providing the sediment samples. We very much appreciated the detailed and constructive comments of one anonymous reviewer. The funding provided by the German Research Foundation (DFG; Se 508/2) is acknowledged. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2015.02.031. References Albinana-Gimenez, N., Clemente-Casares, P., Bofill-Mas, S., Hundesa, A., Ribas, F., Girones, R., 2006. Distribution of human polyomaviruses, adenoviruses, and hepatitis E virus in the environment and in a drinking-water treatment plant. Environ. Sci. Technol. 40 (23), 7416–7422. Bauer, R., Dizer, H., Graeber, I., Rosenwinkel, K.H., López-Pila, J.M., 2011. Removal of bacterial fecal indicators, coliphages and enteric adenoviruses from waters with high fecal pollution by slow sand filtration. Water Res. 45 (2), 439–452. Bethwell, C., 2004. Organische Substanz und ihre partikulären und adsorbierten Anteile im Feinsediment eines bergbaubeeinflussten Fließgewässers. In: Rücker, J., Nixdorf, B. (Eds.), Gewässerreport Nr. 8. BTUC-AR 3/2004 - Eigenverlag der BTU Cottbus, pp. 109–118 (http://www.tu-cottbus.de/fakultaet4/de/gewaesserschutz/downloads/ akt-2004.html - verified 26th February 2015). Brady, N.C., Weil, R.R., 2002. The Nature and Properties of Soils. 13th ed. Prentice-Hall, Upper Saddle River, NJ. Citeau, L., Lamy, I., van Oort, F., Elsass, F., 2001. Nature des sols et nature des colloïdes circulant dans les eaux gravitaires: une etude in situ. C. R. Acad. Sci. II A 332, 657. DIN EN ISO 10705-1, 2001. Water Quality – Detection and Enumeration of Bacteriophages – Part 1: Enumeration of F-specific RNA Bacteriophages (in German). DIN EN ISO 10705-2, 2001. Water Quality – Detection and Enumeration of Bacteriophages – Part 2: Enumeration of Somatic Coliphages (in German). Dowd, S.E., Pillai, S.D., Wang, S., Corapcioglu, M.Y., 1998. Delineating the specific influence of virus isoelectric point and size on virus adsorption and transport through sandy soils. Appl. Environ. Microbiol. 64 (2), 405–410. Emelko, M., Tufenkje, N., 2010. Transport and fate of colloids and microbes in granular aqueous environments. Water Res. 44 (4), 1027. Feng, Y.Y., Ong, S.L., Hu, J.Y., Tan, X.L., Ng, W.J., 2003. Effects of pH and temperature on the survival of coliphages MS2 and Qβ. J. Ind. Microbiol. Biotechnol. 30, 549–552. 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 — po-
tential implications for groundwater protection. Int. J. Hyg. Environ. Health 217 (8), 861–870. Hot, D., Legeay, O., Jacques, J., Gantzer, C., Caudrelier, Y., Guyard, K., Lange, M., Andréoletti, L., 2003. Detection of somatic phages, infectious enteroviruses and enterovirus genomes as indicators of human enteric viral pollution in surface water. Water Res. 37 (19), 4703–4710. ISO 11277, 1998. Soil Quality – Determination of Particle Size Distribution in Mineral Soil Material – Method by Sieving and Sedimentation. Jin, Y., Flury, M., 2001. Fate and transport of viruses in porous media. Adv. Agron. 77, 39–102. Jin, Y., Yates, M.V., Thompson, S.S., Jury, W.A., 1997. Sorption of viruses during flow through saturated sand columns. Environ. Sci. Technol. 31 (2), 548–555. Jin, Y., Pratt, E., Yates, M.V., 2000. Effect of mineral colloids on virus transport through saturated sand columns. J. Environ. Qual. 29 (2), 532–539. Kaiser, K., Guggenberger, G., 2003. Mineral surfaces and soil organic matter. Eur. J. Soil Sci. 54 (2), 219–236. Klitzke, S., Lang, F., Kaupenjohann, M., 2008. Increasing pH releases colloidal lead in a highly contaminated forest soil. Eur. J. Soil Sci. 59 (2), 265–273. Kretzschmar, R., Hesterberg, D., Sticher, H., 1997. Effects of adsorbed humic acid on surface charge and flocculation of kaolinite. Soil Sci. Soc. Am. J. 61 (1), 101–108. Kretzschmar, R., Borkovec, M., Grolimund, D., Elimelech, M., 1999. Mobile subsurface colloids and their role in contaminant transport. Adv. Agron. 66, 121–193. Lodder, W.J., De Roda Husman, A.M., 2005. Presence of noroviruses and other enteric viruses in sewage and surface waters in the Netherlands. Appl. Environ. Microbiol. 71 (3), 1453–1461. Maurice, P.A., Namjesnik-Dejanovic, K., 1999. Aggregate structures of sorbed humic substances observed in aqueous solution. Environ. Sci. Technol. 33 (9), 1538–1541. McBride, M.B., 1994. Environmental Chemistry of Soils. Oxford University Press, New York. McCarthy, J.F., Zachara, J.M., 1989. Subsurface transport of contaminants. Mobile colloids in the subsurface environment may alter the transport of contaminants. Environ. Sci. Technol. 23 (5), 496–502. McGechan, M.B., Lewis, D.R., 2002. Transport of particulate and colloid-sorbed contaminants through soil, part 1: general principles. Biosyst. Eng. 83, 255–273. Michen, B., Graule, T., 2010. Isoelectric points of viruses. J. Appl. Microbiol. 109 (2), 388–397. Mikutta, C., Lang, F., Kaupenjohann, M., 2004. Soil organic matter clogs mineral pores: evidence from 1H NMR and N2 adsorption. Soil Sci. Soc. Am. J. 68 (6), 1853–1862. Ngazoa, E.S., Fliss, I., Jean, J., 2008. Quantitative study of persistence of human norovirus genome in water using TaqMan real-time RT-PCR. J. Appl. Microbiol. 104 (3), 707–715. Patil, S.A., Hägerhäll, C., Gorton, L., 2012. Electron transfer mechanisms between microorganisms and electrodes in bioelectrochemical systems. Bioanal. Rev. 4 (2–4), 159–192. Santschi, P., Höhener, P., Benoit, G., Buchholtz-ten Brink, M., 1990. Chemical processes at the sediment–water interface. Mar. Chem. 30 (C), 269–315. Schijven, J.F., Hassanizadeh, S.M., 2000. Removal of viruses by soil passage: overview of modeling, processes, and parameters. Crit. Rev. Environ. Sci. Technol. 30 (1), 49–127. Schijven, J.F., Medema, G., Vogelaar, A.J., Hassanizadeh, S.M., 2000. Removal of microorganisms by deep well injection. J. Contam. Hydrol. 44 (3–4), 301–327. Seta, A.K., Karathanasis, A.D., 1997. Stability and transportability of water-dispersible soil colloids. Soil Sci. Soc. Am. J. 61 (2), 604–611. Stumm, W., Sigg, L., 1979. Colloid-chemical Basis of Phosphate Removal by Chemical Precipitation, Coagulation and Filtration. 12 (2), 73–83 (in German). Tanner, C.B., Jackson, M.L., 1948. Nomographs of sedimentation times for soil particles under gravity or centrifugal acceleration. Soil Sci. Soc. Am. J. 12, 60–65. Van der Wielen, P.W.J.J., Senden, W.J.M.K., Medema, G., 2008. Removal of bacteriophages MS2 and ΦX174 during transport in a sandy anoxic aquifer. Environ. Sci. Technol. 42 (12), 4589–4594. Walshe, G.E., Pang, L., Flury, M., Close, M.E., Flintoft, M., 2010. Effects of pH, ionic strength, dissolved organic matter, and flow rate on the co-transport of MS2 bacteriophages with kaolinite in gravel aquifer media. Water Res. 44 (4), 1255–1269. Weiss, W.J., Bouwer, E.J., Aboytes, R., LeChevallier, M.W., O'Melia, C.R., Le, B.T., Schwab, K.J., 2005. Riverbank filtration for control of microorganisms: results from field monitoring. Water Res. 39 (10), 1990–2001. Westrell, T., Teunis, P., van den Berg, H., Lodder, W., Ketelaars, H., Stenström, T.A., de Roda Husman, A.M., 2006. Short- and long-term variations of norovirus concentrations in the Meuse river during a 2-year study period. Water Res. 40 (14), 2613–2620.
