Sunlight-Mediated Inactivation of MS2 Coliphage via Exogenous

The second and third mechanisms, endogenous and exogenous photoinactivation, ..... In WSP water samples diluted with 50% D2O, kobs increased 1.3-fold ...
10 downloads 0 Views 141KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 192-197

Sunlight-Mediated Inactivation of MS2 Coliphage via Exogenous Singlet Oxygen Produced by Sensitizers in Natural Waters TAMAR KOHN AND KARA L. NELSON* Department of Civil and Environmental Engineering, Univeristy of California, Berkeley, Berkeley, California 94720

Pathogens in sunlit surface waters can be damaged directly by UVB light. Indirect inactivation by reactive oxygen species (ROS) generated by sunlight interacting with external sensitizer molecules may also be important, but this mechanism has not been conclusively demonstrated. To better understand the role of ROS, we investigated the inactivation of MS2 coliphage, a commonly used surrogate for human enteric viruses, in water samples irradiated with a solar simulator and containing different types of sensitizers: waste stabilization pond (WSP) constituents, Fluka humic acid (FHA), and Suwannee River humic acid (SRHA). Inactivation of MS2 by the indirect mechanism was significant for all three sensitizers, and the efficiency of the sensitizers at inactivating MS2 was FHA > SRHA > WSP. Both dissolved and particulate fractions in the WSP water contributed to inactivation. In the WSP water, the indirect process was quantitatively more important than direct damage by UVB light, due to the rapid attenuation of UVB compared to the longer wavelengths that may initiate the indirect mechanism. Singlet oxygen (1O2) was the most important ROS involved in the inactivation of MS2. The addition of histidine, a 1O2 quencher, decreased inactivation, whereas inactivation rate constants increased in solutions of D2O. Selective quenchers for other ROS showed little or no protective effect. Inactivation in WSP water was a function of the steady-state 1O2 concentration and could be described by a second-order rate expression.

Introduction Sunlight is an important factor impacting the survival of pathogens in sewage-influenced fresh and seawater (1-5). This naturally occurring water disinfection process is also exploited for various engineering applications. In waste stabilization ponds (WSPs), which can be considered eutrophic, shallow lakes, sunlight-mediated pathogen inactivation contributes significantly to improving microbial effluent water quality (6, 7). Solar disinfection of drinking water (SODIS) also relies on this process (8, 9). Three distinct mechanisms can lead to sunlight-mediated pathogen inactivation (10). The first mechanism involves the direct damage of cell components by sunlight in the ultraviolet (UV) region (11). The second and third mechanisms, endogenous and exogenous photoinactivation, are indirect processes, which potentially can be initiated by both UV and visible light. In the presence of oxygen, the excitation of * Corresponding author phone: (510) 643-5023; fax: (510) 6427483; e-mail: [email protected]. 192

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007

sensitizers (light-absorbing compounds that transfer energy to other molecules) leads to the formation of reactive oxygen species (ROS), which can damage cell constituents. In endogenous photoinactivation, sensitizers are located within the cell (e.g., flavins; ref 11), whereas in exogenous photoinactivation sensitizers are located outside the cell (e.g., humic substances; refs 6, 12, 13). ROS include singlet oxygen (1O2), superoxide (O2-), hydroxyl radical (OH·), peroxyl radical (RO2-• ), and hydrogen peroxide (H2O2) (12). Common formation pathways for ROS involve absorption of a photon by a sensitizer molecule and subsequent reaction either by energy transfer or by electrontransfer reactions with oxygen (12, 13). Direct solar UV damage to both viral and bacterial pathogens has been described in some detail (11). In contrast, the contributions of endogenous and exogenous photoinactivation have not been quantified, in part because they are difficult to separate. Davies-Colley et al. (7) reported the occurrence of all three mechanisms in WSP, although the rates and dominant mechanisms varied among organisms (7, 14). Curtis et al. (6) found evidence for both endogenous and exogenous mechanisms in WSP, and implicated H2O2 and 1O2 as the most important ROS contributing to photoinactivation of fecal coliform bacteria. In general, indirect photoinactivation has been found to depend on the dissolved oxygen concentration (7, 15), pH (6, 7), salinity (3, 16), and wavelength of light (6, 7). The goal of this study was to develop a quantitative understanding of the exogenous photoinactivation of MS2 in water and to identify the most important ROS. We chose MS2 for two main reasons. First, it is a common model for human enteric viruses due to its similar size and morphology (17). Second, the protein capsid is unlikely to contain sensitizers that efficiently form ROS when exposed to sunlight (18). Therefore, unlike bacteria, MS2 is not subject to endogenous photoinactivation, and presents a simplified model for investigating exogenous mechanisms. A better understanding of exogenous photoinactivation will ultimately contribute to the development of mechanistic models of indicator organism and pathogen inactivation in sunlit waters that account for the effects of water quality, the individual characteristics of microorganisms, and hydrodynamics.

Experimental Section Experiments were conducted in batch reactors containing water samples that were spiked with MS2 and exposed to simulated sunlight in the laboratory. The inactivation of MS2 was monitored over time. The following parameters were investigated: sensitizer type, ROS, pH, and salinity. Reagents and Organisms. A description of the reagents and organisms used in this study can be found in the Supporting Information. Experimental Setup. Irradiation experiments were performed using a solar simulator (Oriel) with a collimated beam (8 × 8 in), equipped with a 1000 W Xe lamp. The irradiance spectrum (W/m2/nm) incident to the samples was measured with spectroradiometers (RPS 200 and 380, International Light). An atmospheric attenuation filter (serial no. 81088; Oriel) was used for all experiments. In addition, a second filter was used to modify the UV portion of the spectrum. When the second filter was an atmospheric attenuation filter (serial no. 81017; Oriel) the UVB portion (290-320 nm) of the solar spectrum in Berkeley (38° N) was mimicked very closely, but the visible portion of the spectrum was underrepresented (Figure 1). The total irradiance (280-700 nm) using these two filters was 368 W/m2, while the total irradiance 10.1021/es061716i CCC: $37.00

 2007 American Chemical Society Published on Web 12/01/2006

FIGURE 1. Solar spectrum obtained in Berkeley, CA (38° N, 6/21/05, noon) (‚‚‚‚), solar simulator output with two atmospheric attenuation filters (s) and with one atmospheric attenuation and one UVB/C blocking filter (- ‚ -), and absorption spectrum of a typical WPS sample (right axis) (- -). by natural sunlight on a summer day, measured on June 21, 2005 at noon in Berkeley, CA was 510 W/m2. Alternatively, a blocking filter (serial no. 81050; Oriel) with a cutoff at approximately 320 nm could be used to eliminate the UVB portion (responsible for direct inactivation), resulting in an overall irradiance of 352 W/m2 (Figure 1). This filter also had the unfortunate effect of reducing the UVA portion (320400 nm). 150-mL glass reactors containing 100-mL samples were placed under the beam in a water bath cooled to 20 °C by a recirculating water chiller (Thermo Electron). The exterior of the reactors was painted black to prevent reflection of incident light. Dark controls for all experiments were conducted under the same conditions as illuminated samples, but in reactors covered by aluminum foil. Unless indicated otherwise, no dark inactivation was observed. The sample solutions were continuously stirred by a magnetic stir bar. Alternatively, experiments were carried out in reactors that were continuously sparged with either air or N2. This setup allowed us to investigate the influence of dissolved oxygen on inactivation, as well as to maintain a constant, airsaturated oxygen concentration in experiments involving histidine and in the pond filtrate experiments. Experiments involving DI water were performed in autoclaved solutions buffered at pH 7.5 with 5 mM phosphate and 10 mM NaCl. Sensitizers (Fluka humic acid (FHA), Suwannee River humic acid (SRHA), or Rose Bengal (RB)) were added into the reactors from filtered (0.45 µm pore size filter; Millipore) stock solutions in DI. The pH was adjusted by adding NaOH and HCl. Grab samples from WSPs were obtained periodically between May 2005 and March 2006. Samples were stored at 4 °C and used within 2 weeks. A typical WSP sample contained around 15 mg/L total organic carbon, up to 0.3 mg/L total iron, and was green in color, due to the abundance of algae. 5 mM phosphate buffer and 10 mM NaCl were added before each experiment, and the pH was adjusted to 7.5. In spite of the buffer, the pH had to be lowered periodically during the experiment, presumably to counteract photosynthesis by algae. For experiments with filtered WSP water, samples were sequentially passed through a series of filters (pore size 0.9-8 µm [prefilter], 0.8 µm, 0.45 µm, and 0.2 µm; Millipore). MS2 stock solutions were first diluted 1:100 in dilution water (DW). 1 mL of the diluted stock solution was added to the reactors, yielding an initial MS2 concentration of ∼2 × 107 plaque forming units (PFU)/mL. Sample aliquots (100

µL) were withdrawn at regular time intervals, immediately diluted into 900 µL of autoclaved DW in a 2-mL microcentrifuge tube, and stored in the refrigerator until further processing. In experiments with added H2O2, a stock solution (100 mM) of H2O2 was prepared in DI water and used immediately. The DW used for the processing of these samples contained 200 U/mL of catalase to quench the H2O2. Samples were processed within 5 h of sampling. This time delay did not adversely affect the MS2 count. Measurement of ROS Concentrations. The steady-state concentrations of 1O2 ([1O2]ss) and OH‚([OH‚]ss) were determined by measuring the decay of a selective probe compound with known quenching rate constant (kq) (19). The probe compound for 1O2 was FFA (kq) 1.2 × 108 M-1s-1; ref 19) and that for OH‚ was phenol (kq ) 1.4 × 1010 M-1s-1; ref 20). Probe compounds were added to the sample at a final concentration of 50 µM. Sample aliquots (1 mL) were extracted at regular intervals, and transferred to amber, 2-mL glass vials with a crimp seal. The decay of phenol and FFA was followed by HPLC/UV (Gynkotek), equipped with a guard column and a C18 column (Alltech). Acetonitrile and 2 mM formic acid were used as eluents. The detection wavelengths were 220 nm (FFA) and 254 nm (phenol). The minimum detectable steady-state concentrations determined from comparison with probe compound disappearance in sensitizer-free solution were 10-15 M for [1O2]ss and 10-16 M for [OH‚]ss. The concentration of H2O2 was determined colorimetrically (21), using a UV/vis spectrometer (Perkin-Elmer). The detection limit for this method was 0.8 µM. The quantification of superoxide production rates was performed via the dismutation to H2O2 as described in ref 22. The detection limit of this method was 1.6 µM. Quencher Experiments. To differentiate between the effects of the different ROS, experiments with quenchers were performed to selectively suppress the steady-state concentration or prevent the accumulation of the ROS under consideration. The quenchers used were formate (50 mM) for OH‚, and L-histidine (20 mM) for 1O2, SOD (2 U/mL) for superoxide and catalase (200 U/mL) for H2O2. In addition, the iron chelator desferrioxamine (2-15 µM) was used to inhibit the formation of OH‚ by (photo-) Fenton processes. Data Analysis. To compare MS2 inactivation under different conditions, first-order rate constants were obtained from the slope of plots of ln (PFU/mL) vs time, and are reported as kobs (h-1) (95% confidence intervals. R2 were generally higher than 0.95. To compare the ability of different sensitizers to cause exogenous inactivation by ROS, and to account for differences in light attenuation by different solutions, we also analyzed our data as a function of the average number of photons absorbed throughout the sample, -1 2 nabs,average. The resulting rate constant, κabs obs (photon m ; see Supporting Information), accounts for the fact that each photon in the wavelength range considered (320-700 nm) can potentially produce ROS. Because we did not measure the wavelength dependence of ROS production, we weighted all wavelengths equally. A correct accounting may change our results, as the wavelength dependence of ROS quantum yields may vary for different sensitizers (23, 24). Our approach also does not incorporate the wavelength dependence of other processes that may cause inactivation (e.g., direct UV damage, OH‚ formation by photolysis of nitrate). Finally, we did not account for the effects of scattering by particles, which would increase the path length of light in our reactors.

Results and Discussion Direct UV Damage, Endogenous, and Exogenous Photoinactivation. To establish the relative importance of the VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

193

FIGURE 2. Inactivation of MS2 as a function of time in 5 mM phosphate buffer illuminated with ([) and without (b) UVB light, and in phosphate-buffered WPS water, illuminated with (2) and without UVB light (1). different photoinactivation mechanisms for the inactivation of MS2, experiments with and without UVB, as well as with and without exogenous sensitizers were conducted. No significant inactivation over 12 h was observed in the absence of both sensitizers and UVB (Figure 2), confirming that endogenous inactivation initiated by wavelengths >320 nm was not a relevant process over the time scale considered. In the presence of UVB, a 2.0-log inactivation (two ordersof-magnitude decrease in culturable virus) over 8 h (kobs ) 0.56 ( 0.10 h-1) was observed if no external sensitizers were present. We believe this inactivation was a result of direct UVB damage to the nucleic acid. Although we cannot completely rule out that some of this inactivation could be from endogenous inactivation initiated by components of the protein capsid (e.g., tryptophan) sensitized by UVB, the generation of ROS by proteins from light in the solar spectrum is expected to be minimal in the absence of a bound chromophore (18). The presence of WSP constituents in the sample shown in Figure 2 led to a 2.0-log inactivation of MS2 over 12 h (kobs ) 0.39 ( 0.06 h-1) in the absence of UVB. Because endogenous inactivation did not occur under these conditions, the observed inactivation can be attributed completely to exogenous mechanisms mediated by sensitizers in the WSP water. It should be noted that the light in the UVB (and UVA) ranges that was removed with the blocking filter may also contribute to exogenous inactivation. Thus, the contribution of exogenous inactivation in experiments using the blocking filter is underestimated compared to full sunlight. A 2.6-log inactivation over 8 h (kobs ) 0.72 ( 0.09 h-1) was observed in the presence of both WSP water constituents and UVB. Because inactivation occurred faster in WSP water than in sensitizer-free phosphate buffer, it is apparent that the reduction in direct UV damage caused by the WSP water (which reduced penetration of UV into the solution) was overcome by the increase in inactivation due to exogenous mechanisms mediated by the sensitizers. To illustrate this effect, consider the light attenuation (see Supporting Information, eq S1), under the conditions shown in Figure 1. Over 99% of the UVB light at 290 nm, which directly damages nucleic acid (11), is absorbed in the top 2.5 cm. Direct UVB damage at this wavelength thus can only occur in the top of our 5 cm deep experimental reactors. Visible light at 556 nm, however, which has been found to contribute to photoinactivation of RNA phages (3), penetrates deeper (99% adsorbed at 8 cm), allowing indirect exogenous inactivation of MS2 to occur throughout our reactors. 194

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007

From these experiments we conclude that exogenous inactivation can be an important contributor to the overall inactivation of MS2 in WSP water. The other exogenous sensitizers investigated, FHA and SRHA, exhibited similar trends, but the relative importance of direct UVB damage and exogenous inactivation was concentration-dependent. The subsequent experiments reported in this paper were performed using the UVB blocking filter, which allowed us to investigate exogenous inactivation of MS2 without interference or confounding effects of direct UV damage. Influence of Sensitizer Composition. To gain insight into the sensitizers present in WSP water, experiments were conducted after serial filtration of WSP water through membranes of decreasing pore size. The inactivation rate decreased after each treatment, with the three lowest filtrates (0.22 µm, 0.45 µm, and 0.8 µm) exhibiting very similar inactivation (Figure 3a). Consistent with findings by others (7), both particulate and dissolved fractions appeared to be effective sensitizers. The particulate fraction in WSP water is mostly algal cells, which were predominately removed by the 0.8 µm filter. Both algal cells and algal exudates have been shown to enhance the indirect photolysis of organic chemicals (25), and may also play a role in the inactivation of MS2. The serial filtration altered not only the size, but also the concentration of sensitizers, which could have several competing effects. It may decrease the production of ROS, but simultaneously decrease quenching of some ROS as well as increase light penetration. To account for differences in light penetration, we normalized the curves by plotting them against nabs,average (Figure 3b). The absorption-normalized rate -25 constants, κabs ( 0.2 × 10-25 obs ranged from 2.4 × 10 photon-1m2 (kobs ) 0.63 ( 0.07 h-1) in the unfiltered sample to 1.9 × 10-25 ( 0.2 × 10-25 photon-1m2 (kobs ) 0.44 ( 0.05 h-1) in the lower filtrates. Thus, solutions containing sensitizers >0.8 µm were slightly more effective at a 95% confidence level. Experiments with 5 mg/L FHA exhibited an absorption-25 ( normalized inactivation rate constant of κabs obs ) 10 × 10 1 × 10-25 photon-1m2 (kobs ) 0.25 ( 0.02 h-1), and those with -25 ( 0.6 × 25 mg/L SRHA a rate constant of κabs obs ) 3.5 × 10 10-25 photon-1m2 (kobs ) 0.12 ( 0.02 h-1). Comparing κobs, FHA was thus a more efficient sensitizer for MS2 inactivation than SRHA, which was slightly more efficient than WSP water. Influence of Dissolved Oxygen, pH, and Salinity. No difference in inactivation rate constants was observed between WSP samples sparged with air and WSP samples sparged with N2 (data not shown). Although the dissolved oxygen concentration was 1 µm pore size (1), and unfiltered WSP water (b). Samples were illuminated using the UVB/C blocking filter. to produce ROS, or from differences in the susceptibility of the virus to photoinactivation. Neither the UV/vis absorption spectra of our solutions, nor the production of 1O2, were significantly affected by variations in ionic strength. Interestingly, it has been shown (26) that MS2 stability toward heat treatment drastically decreases between 1 and 100 mM NaCl. We speculate that the observed effect on kobs reflects the impact of ionic strength on virus susceptibility toward ROS. Assessment of the Role of ROS. Hydroxyl Radical. Hydroxyl radicals are unselective and react with many organic substrates at diffusion-limited rates (27). They are, therefore, likely to exhibit reactivity toward the components of MS2. OH‚ has been implicated to cause MS2 inactivation during photocatalytic disinfection processes (28). In our experimental system, however, the high organic matter content of the samples was an efficient quencher for OH‚ (13), yielding a [OH‚]ss below our detection limit in all samples investigated. Neither the addition of the OH‚ quencher formate to our samples (FHA and WSP), nor the presence of the iron chelator desferrioxamine, which inhibits the formation of OH‚ via (photo-) Fenton reactions, protected MS2 form inactivation. We, therefore, conclude that OH‚ did not contribute significantly to the sunlight-mediated inactivation of MS2. Superoxide. Superoxide exhibits lower reactivity than OH‚ toward most organic substrates (e.g., k < 100 M-1s-1 for most amino acids; ref 29). Its steady-state concentration in surface waters can reach 10-8 M (12), although it may be several orders of magnitude lower in the presence of trace amounts of copper or iron (e.g., 30). In our WSP samples, the production rates were too low to be measured, and addition of SOD did not protect MS2 from inactivation. Thus, we conclude that superoxide reactivity with the protein capsids or nucleic acid of MS2 was not sufficient to significantly contribute toward MS2 inactivation. Singlet Oxygen. The potential of 1O2 to disinfect viruses has been recognized and explored for medical purposes, for example in photodynamic inactivation (e.g., ref 31). Its relative importance in inactivating viruses via exogenous mechanisms in water exposed to sunlight, however, has not been studied. The measured [1O2]ss was typically on the order of 10-13 M in WSP samples and 5 × 10-14 M in FHA and SRHA samples. kobs of WSP water was 0.42 ( 0.01 h-1, and decreased to 0.24 ( 0.04 h-1 in the presence of histidine. In solutions containing 5 mg/L FHA, histidine completely inhibited MS2 inactivation. This indicates that 1O2 played an important role in the inactivation of MS2.

To confirm that this decrease in kobs was caused by the quenching of 1O2, experiments were conducted in D2O. The quenching rate constant of 1O2 with D2O is approximately 13 times lower than that of H2O (32). Because water is the most important quencher for 1O2, experiments conducted in D2O should increase [1O2]ss and the corresponding kobs accordingly. In experiments with 5 mg/L FHA in 99.8% D2O, kobs increased 3.0-fold. In WSP water samples diluted with 50% D2O, kobs increased 1.3-fold compared to samples diluted with 50% H2O. These increases in kobs are lower than their theoretical values (13-fold increase in 100% D2O; 2-fold increase in 50% D2O), but the results nevertheless confirm that 1O2 plays a significant role in MS2 inactivation. The lower-than-expected increase in kobs may be a result of association between MS2 and the sensitizer. In a recent study (33), the 1O2 concentration inside and in close proximity to the sensitizer (humic acid) was found to be several orders of magnitude higher than in the bulk solution. In addition, changing the solvent to D2O only increased the 1O2 concentration in the aqueous phase, but not within the sensitizer. Thus, if MS2 is closely associated with the sensitizer, its inactivation may be partially governed by the internal sensitizer 1O2 concentration, and we would expect the effect of D2O on kobs to be diminished. Further investigations on the effect of sensitizer-MS2 interactions are currently under way in our laboratory. Hydrogen Peroxide. The presence of catalase in our samples led to a slight decrease in MS2 inactivation. Catalase, however, also lowered [1O2]ss. Discriminating between the effect of decomposing H2O2 and quenching 1O2 was, therefore, difficult. H2O2 is the least reactive of the investigated photooxidants (12) and external addition of H2O2 in the concentration range measured in our WSP water samples (1-5 µM) did not lead to an observable inactivation of MS2 (data not shown). It is, therefore, likely that the reduced inactivation was mainly a result of the lowered [1O2]ss. As discussed below, however, synergistic action of H2O2 and 1O2 on MS2 is possible. Synergistic Effects of 1O2 and H2O2. To better differentiate between the effects of H2O2 and 1O2, experiments were performed in solutions with externally added H2O2 and Rose Bengal (RB) as a sensitizer for 1O2. The presence of RB (0.5 µM) in solution led to a [1O2]ss of 4 × 10-12 M and rapid MS2 inactivation (Figure 4), (kobs ) 5.6 ( 1.2 h-1). Illuminated solutions without RB, but with externally added H2O2 (200 µM), exhibited a kobs of 10.8 ( 1.7 h-1. If both RB and H2O2 VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

195

FIGURE 4. Synergistic effect of H2O2 and 1O2, as well as H2O2 and light on the inactivation of MS2. Experiments were conducted in 5 mM phosphate buffer, in the presence of 200µM H2O2 only (1) 0.5 mg/L Rose Bengal as a sensitizer for 1O2 only (b), or both (9). Samples were illuminated using the UVB/C blocking filter. The fourth data set ([) shows the inactivation by 0.5 mg/L Rose Bengal and 200µM H2O2 in the dark. were present, the resulting inactivation was greater than the sum of the individual effects, with a kobs of 31.6 ( 17.4 h-1, indicating that H2O2 and 1O2 act synergistically to inactivate MS2. Light and H2O2 also exhibited a strong synergistic effect (Figure 4). Inactivation by 200 µM H2O2 in the dark occurred with a 4.4-fold lower kobs of only 2.4 ( 0.7 min-1 compared to inactivation in the light. The formation of OH‚ either by photolysis of H2O2, or via photo-Fenton processes involving trace metals was not the cause of this synergistic inactivation. Photolysis of H2O2 by light in the solar spectrum is slow due to the poor overlap of sunlight with the spectrum of H2O2 (12), and the addition of an OH‚ quencher (formate) did not inhibit inactivation. A similar synergistic effect of UVA light and H2O2 on the inactivation of various DNA and RNA phages has previously been reported (34, 35). Hartman et al. (36) found that the combination of UVA and H2O2 led to the formation of protein-nucleic acid cross-links in DNA phage T7, which inhibited the complete injection of the nucleic acid into the host. Analogous processes may take place for RNA phages such as MS2 (35), and may be the cause for the observed synergistic effect of H2O2 and sunlight in our system. In actual WSP water or surface waters, the concentrations of H2O2 and 1O2 are typically at least 2 orders of magnitude lower than the concentrations used in the experiments with RB as a sensitizer. It is likely, however, that the observed synergistic effects of H2O2 and light, as well as H2O2 and 1O2, also occur over longer time scales at environmentally relevant concentrations. For example, the inactivation rate of MS2 in 5 mg/L SRHA was found to increase once H2O2 levels surpassed 4 µM, which may be a result of the synergistic effect of H2O2 with either light or 1O2 (data not shown). [1O2]ss as a Predictor for MS2 Inactivation. A compilation of kobs and [1O2]ss measured in various filtered and unfiltered WSP water samples obtained over the course of 10 months is shown in Figure 5. kobs and [1O2]ss are highly correlated (R2 ) 0.9) for the WSP investigated. Although we cannot exclude inactivation by other oxidants that may scale with [1O2]ss (e.g., excited triplet states of dissolved organic matter), [1O2]ss can be used an indicator to estimate the inactivation of MS2 in this WSP. From the data in Figure 5, we can derive a second-order inactivation rate constant k for the following rate expression: 196

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007

FIGURE 5. Dependence of MS2 inactivation rate constant in WSP water on [1O2]ss. Samples were illuminated using the UVB/C blocking filter (filled symbols) or using two atmospheric attenuation filters (empty symbols). Error bars indicate 95% confidence intervals on the inactivation rate constants. The solid line indicates the linear regression line (kobs ) -0.17 ((0.21) + 4.8 × 1012 ((1.5 × 1012) × [1O2]ss; R2 ) 0.90; S.E. ) 0.08), dotted lines indicate the 95% confidence intervals of the regression.

d[MS2] ) k[1O2]ss[MS2] dt According to the slope of the linear regression in Figure 5, the second-order rate constant k is 4.8 × 1012 ( 1.5 × 1012 M-1h-1 (or 1.3 × 109 ( 0.4 × 109 M-1s-1). This rate constant is surprisingly high, approximately 1 order of magnitude higher than those reported for the reaction of 1O2 with protein components (18), and several orders of magnitude higher than those for the subunits of RNA (37). One factor that may explain the high rate constant is that MS2 is much larger than these individual macromolecules, which would result in a higher chance of encounter with 1O2 in solution. Alternatively, the discrepancy in the magnitude of reaction rate constants may again reflect the association of MS2 with the sensitizers, which would lead to an overestimation of the second-order rate constant based on the solution-phase [1O2]ss. We can assume, however, that in the WPS under investigation [1O2]ss in the bulk solution is proportional to the concentration produced in proximity to the sensitizer and, therefore, use [1O2]ss as an indicator to estimate the inactivation of MS2. It should be noted that the derived rate expression is specific to the conditions of this study. In an actual WSP, variations in sunlight intensity during the day and mixing in the water column must be accounted for. The rate expression also may not apply to other sensitizers, because the sensitizer-MS2 association, as well as the concentration profile of 1O2 in proximity to the sensitizer, may vary. Environmental Relevance. In engineered disinfection processes that rely solely on direct UV damage, association of MS2 with particles and light screening by organic matter inhibit MS2 inactivation (38). In contrast, when WSP water was exposed to simulated sunlight, the organic matter sensitized the production of 1O2 and the overall inactivation rate of MS2 increased (Figures 2 and 3). Further research is needed to establish the relative importance of direct versus indirect inactivation in natural surface waters (e.g., eutrophic vs oligotrophic lakes). In addition, more research is needed to understand the importance of exogenous inactivation for actual enteric pathogens, including bacteria, in which endogenous mechanisms may also be important.

Acknowledgments This study was supported by a PECASE/CAREER award from the U.S. National Science Foundation. Support for TK was

provided by a postdoctoral fellowship (PBSK2s106572) of the Swiss National Science Foundation. We thank Khalid Kadir for obtaining WSP water samples and designing the experimental setup. The lab assistance of April Wong and Bonita Lee is much appreciated.

Supporting Information Available The reagents used, MS2 propagation and enumeration, and a description of the approach to calculate light-normalized inactivation rates κabs obs. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Davies-Colley, R. J.; Bell, R. G.; Donnison, A. M. Sunlight inactivation of Enterococci and fecal coliforms in sewage effluent diluted in seawater. Appl. Environ. Microbiol. 1994, 60, 20492058. (2) Sinton, L. W.; Finlay, R. K.; Lynch, P. A. Sunlight inactivation of fecal bacteriophages and bacteria in sewage-polluted seawater. Appl. Environ. Microbiol. 1999, 65, 3605-3613. (3) Sinton, L. W.; Hall, C. H.; Lynch, P. A.; Davies-Colley, R. J. Sunlight inactivation of fecal indicator bacteria and bacteriophages from waste stabilization pond effluent in fresh and saline waters. Appl. Environ. Microbiol. 2002, 68, 1122-1131. (4) Noble, R. T.; Lee, I. M.; Schiff, K. C. Inactivation of indicator micro-organisms from various sources of faecal contamination in seawater and freshwater. J. Appl. Microbiol. 2004, 96, 464472. (5) Boehm, A. B.; Grant, S. B.; Kim, J. H.; Mobray, S. L.; McGee, C. D.; Clark, C. D.; Foley, D. M.; Wellman, D. E. Decadal and shorter period variability of surf zone water quality at Huntington Beach, California. Environ. Sci. Technol. 2002, 36, 3885-3892. (6) Curtis, T. P.; Mara, D. D.; Silva, S. A. Influence of pH, oxygen and humic substances on ability of sunlight to damage fecal coliforms in waste stabilization pond water. Appl. Environ. Microb. 1992, 58, 1335-1343. (7) Davies-Colley, R. J.; Donnison, A. M.; Speed, D. J.; Ross, C. M.; Nagels, J. W. Inactivation of faecal indicator micro-organisms in waste stabilization ponds: interactions of environmental factors with sunlight. Water Res. 1999, 33, 1220-1230. (8) Wegelin, M.; Canonica, S.; Mechsner, K.; Fleischmann, T.; Pesaro, F.; Metzler, A. Solar water disinfection: scope of the process and analysis of radiation experiments. J. Water Supply Res. Technol. 1994, 43, 154-169. (9) Reed, R. H. In Advances in Applied Microbiology: Elsevier: New York, 2004; Vol. 54, pp 333-365. (10) Davies-Colley, R. J.; Donnison, A. M.; Speed, D. J. Towards a mechanistic understanding of pond disinfection. Water Sci. Technol. 2000, 42, 149-158. (11) Jagger, J. Solar-UV Actions on Living Cells; Praeger Publishers: New York, 1985. (12) Cooper, W. J.; Zika, R. G.; Petasne, R. G.; Fischer, A. M. In Aquatic Humic Substances; Suffet, I. H., McCarthy, P., Eds.; American Chemical Society: Washington, DC, 1989. (13) Hoigne´, J.; Faust, B. G.; Haag, W. R.; Sully, F. E.; Zepp, R. G. In Aquatic Humic Substances; Suffet, I. H., McCarthy, P., Eds.; American Chemical Society: Washington, DC, 1989. (14) Davies-Colley, R. J.; Donnison, A. M.; Speed, D. J. Sunlight wavelengths inactivating feacal indicator microorganisms in waste stabilization ponds. Water Sci. Technol. 1997, 35, 219225. (15) Reed, R. H. Solar inactivation of faecal bacteria in water: the critical role of oxygen. Lett. Appl. Microbiol. 1997, 24, 276-280. (16) Davies, C. M.; Evison, L. M. Sunlight and the survival of enteric bacteria in natural waters. J. Appl. Bacteriol. 1991, 70, 265274. (17) Havelaar, A. H.; Olphen, M.; Drost, Y. C. F-specific RNA bacteriophages are adequate model organisms for enteric viruses in fresh water. Appl. Environ. Microb. 1993, 59, 29562962.

(18) Davies, M. J. Singlet oxygen-mediated damage to proteins and its consequences. Biochem. Biophys. Res. Commun. 2003, 305, 761-770. (19) Haag, W. R.; Hoigne´, J. Singlet oxgyen in surface waters. 3. Photochemical formation and steady-state concentrations in various types of waters. Environ. Sci. Technol. 1986, 20, 341348. (20) Kochany, J.; Bolton, J. R. Mechanism of photodegradation of aqueous organic pollutants. 1. EPR spin-trapping technique for the determination of OH radical rate constants in the photooxidation of chlorophenols following the photolysis of H2O2. J. Phys. Chem. 1991, 95, 5116-5120. (21) Voelker, B. M.; Sulzberger, B. Effects of fulvic acid on Fe(II) oxidation by hydrogen peroxide. Environ. Sci. Technol. 1996, 30, 1106-1114. (22) Petasne, R. G.; Zika, R. G. Fate of superoxide in coastal seawater. Nature 325, 325, 516-518. (23) Haag, W. R.; Hoigne´, J.; Gassman, E.; Braun, A. M. Singlet oxygen in surface waters - Part II: quantum yields of its production by some natural humic materials as a function of wavelength. Chemosphere 1984, 13, 641-650. (24) Paul, A.; Hackbarth, S.; Vogt, R. D.; Ro¨der, B.; Burnison, B. K.; Steinberg, C. E. W. Photogeneration of singlet oxygen by humic substances: comparison of humic substances of aquatic and terrestrial origin. Photochem. Photobiol. Sci. 2004, 3, 272280. (25) Zepp, R. G.; Schlotzhauer, P. F. Influence of algae on photolysis rates of chemicals in water. Environ. Sci. Technol. 1983, 17, 462-468. (26) Verbraeken, E.; Fiers, W. Studies on the bacteriophage MS2. XX. Expansion of the virion in low salt. Virology 1972, 50, 690-700. (27) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (•OH/•O-) in aqueous solution. J. Phys. Chem. Ref. Data 1988, 17, 513886. (28) Sjogren, J. C.; Sierka, R. A. Inactivation of phage MS2 by ironaided titanium dioxide photocatalysis. Appl. Environ. Microbiol. 1994, 60, 344-347. (29) Bielski, B. H. J.; Cabelli, D. E.; Arudi, R. L.; Ross, A. B. Reactivity of HO2/O2- radicals in aqueous solution. J. Phys. Chem. Ref. Data 1985, 14, 1041-1100. (30) Voelker, B. M.; Sedlak, D. L. Iron reduction by photoproduced superoxide in seawater. Mar. Chem. 1995, 50, 93-102. (31) Floyd, R. A.; Schneider, J. E.; Dittmer, D. P. Methylene blue photoinactivation of RNA viruses. Antiviral Res. 2004, 61, 141151. (32) Rodgers, M. A. J.; Snowden, P. T. Lifetime of O2(1∆g) in liquid water as determined by time-resolved infrared luminescence measurements. J. Am. Chem. Soc. 1982, 104, 5541-5543. (33) Latch, D. E.; McNeill, K. Microheterogeneity of singlet oxygen distributions in irradiated humic adic solutions. Science 2006, 311, 1743-1747. (34) Ananthaswamy, H. N.; Eisenstark, A. Near-UV induced breaks in phage DNA - sensitization by hydrogen-peroxide (a tryptopha photoproduct). Photochem. Photobiol. 1976, 24, 439-442. (35) Eisenstark, A.; Buzard, R. L.; Hartman, P. S. Inactivation of phage by near-ultraviolet radiation and hydrogen peroxide. Photochem. Photobiol. 1986, 44, 603-606. (36) Hartman, P. S.; Eisenstark, A.; Pauw, P. G. Inactivation of phage T7 by near-ultraviolet radiation plus hydrogen peroxide: DNAprotein crosslinks prevent DNA injection. Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 3228-3232. (37) Wilkinson, F.; Helman, W. P.; Ross, A. B. Rate constants for the decay and reactions of the lowest electronically excited state of molecular oxygen in solution. An expanded and revised compilation. J. Phys. Chem. Ref. Data 1995, 24, 177-224. (38) Templeton, M. R.; Andrews, R. C.; Hofmann, R. Inactivation of particle-associated viral surrogates by ultraviolet light. Water Res. 2005, 39, 3487-3500.

Received for review July 19, 2006. Revised manuscript received October 9, 2006. Accepted October 13, 2006. ES061716I VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

197