Decay of infectious adenovirus and coliphages in freshwater habitats is differentially affected by ambient sunlight and the presence of indigenous protozoa communities

Ambient water (~ 15 L) was collected from William H. Harsha Lake (Batavia, OH: 39.0252°N, − 84.1303° W). Immediately after collection, 50% of the sample was passed through a 0.80 μm filter to remove indigenous protozoa. Filtration of water to remove protozoa is a common method and more effective than other techniques such as chemical treatments [58‐62]. To minimize any changes in microbial populations, filtered and unfiltered water was stored in dark at 4 °C until the beginning of the experiment (< 48 h). In order to closely mimic ambient conditions by in situ incubation (at William H. Harsha Lake), a SAM was used to conduct the study. The SAM was constructed as previously described [54, 63‐66] and samples were contained using regenerated cellulose dialysis bags (75 mm flat width, 13–14 kD pore size molecular weight cut-off, Spectrum Labs, Rancho Dominguez, CA). The first day of the experiment, both filtered and unfiltered ambient water was spiked with somatic and F+ coliphages and adenovirus and stirred for 15 min to ensure proper distribution of the spikes within the sample. Measured portions of either spiked filtered or spiked unfiltered ambient water (200 mL) were used to fill each dialysis bag. One half of the dialysis bags containing each water type was attached at the top portion (approximately 2–5 cm below the water surface for the light exposure treatment), while the other half was placed at the bottom portion (approximately 25–30 cm below the water surface underneath the heavy-duty black plastic tarp for shaded treatment). For the study, four different treatments were as follows: A: exposure to ambient sunlight and indigenous microbiota including protozoa (top level, unfiltered water), B: exposure to indigenous microbiota including protozoa, (bottom level, unfiltered water), C: exposure to ambient sunlight only (top level, filtered water) and D: exposure to neither variable (bottom level, filtered water). During each sampling event, triplicate dialysis bags for each treatment were processed for the enumeration of somatic and F+ coliphages, as well as infectious adenovirus (as described below). Concentrations of all viruses were obtained immediately after the inoculum preparation (day 0) and after one and eight days of exposure. Two additional time points (days 3 and 5) were processed for both coliphage types.

Somatic and F+ coliphage were enumerated using double agar layer (DAL) procedure, as previously described [67]. If necessary, decimal dilution series were prepared using 1X phosphate buffered saline solution (PBS: 0.0425 g/L KH₂PO₄ and 0.4055 g/L of MgCl₂; pH 7.2 Sigma Aldrich, St. Louis, MO). Briefly, 1 mL of sample was added to 5 mL of “molten” top tryptic soy agar (TSA) layer (0.7% agar) containing 0.1% of appropriate antibiotic stock solution (100 μg/ mL nalidixic acid for somatic and 15 μg/ mL streptomycin/ampicillin for F+ coliphage) (Fisher Scientific, Waltham, MA), followed by addition of 200 μl of appropriate E. coli host (CN-13 ATCC#700609 [somatic] of F_amp ATCC#700891 [F+], American Type Culture Collection, Manassas, VA) in mid-log growth phase. The soft agar overlay mixture was mixed and poured on bottom agar TSA plates (1.5% agar and containing 0.1% of appropriate antibiotic stock solution). Plates were incubated at 37 °C for 16–18 h. The following day characteristic plaque forming units (PFU) for each coliphage type were enumerated and data were expressed as PFU per 1 ml. Method blank (sample substituted with 1X PBS) and media sterility negative controls were performed on each day of the experiment. For the duration of the study, no plaques were observed on any of the negative controls indicating absence of contamination.

Human lung cells (A549, ATCC® CCL-185) were propagated in Dulbecco’s Minimum Essential Medium (DMEM high glucose with HEPES, Greiner, Monroe, NC) supplemented with 10% fetal calf serum (Fisher Scientific) and 1% sodium pyruvate (Fisher Scientific) under 5% CO₂ atmosphere and at 37 °C. Test cultures of A549 cells were planted and grown to 90% confluency for 4 days in 25 cm² filter capped flasks at 37 °C using a maintenance medium (as described above) except: 1) the addition of antibiotic-antimycotic solution (1% v/v, Fisher Scientific) and 2) reduced fetal calf serum amount of 2% v/v. Prior to inoculation with samples, test cultures were washed with 10 mL of Earle’s Balanced Salt Solution per flask (EBSS, Fisher Scientific) supplemented with 1% antibiotic-antimycotic solution. Decimal dilution series of samples were created using 1X PBS and five replicate flasks per dilution were utilized. In addition, ten negative control flasks (containing 10 mL of 1X PBS instead of the sample) were ran with each sample batch. Following inoculation, flasks were placed on a rocker for 90 min to allow for viral attachment/infection to occur. Flasks were then supplemented with 10 mL of maintenance medium and incubated at 37 °C for 3 weeks [68]. During the incubation time, flasks were examined weekly for the formation of cytopathic effects (CPE). Concentrations of adenovirus were estimated using EPA’s Most Probable Number (MPN) calculator Version 2.0 (https://​cfpub.​epa.​gov/​si/​si_​public_​record_​report.​cfm?​Lab=​NERL&​dirEntryId=​309398). Resulted are reported as MPN per 1 mL.

Human adenovirus 2 (ATCC® VR-846) was obtained from ATCC and propagated in A549 cells to generate higher titers. Briefly, A549 cells were infected with adenovirus as described above for the samples. Following the development of CPE (typically in < a week), cells underwent three freeze-thaw cycles, followed by centrifugation at 2500 x g for 30 min to pellet cellular debris. The supernatant was syringe filtered (0.22 μm pore size), titered (as described above for cell culture samples) and stored in dark at 4 °C until the beginning of the experiment.

For the duration of the study, hourly light intensity (lum/ft²) and temperature (°C) measurements were recorded at both upper and lower SAM levels using HOBO® UA 002–08 data loggers (Onset Computer Corporation, Bourne, MA). The temperature at the top level (16.67 ± 1.18 °C) was slightly higher (paired t-test, p = 0.0394) compared to the bottom level 16.59 ± 0.88 °C), but the light intensity was considerably greater (paired t-test, p < 0.0001) at the top (54.34 ± 146.73 lum/ft²) compared to the bottom level (9.47 ± 19.15 lum/ft²).

All concentration data were log₁₀ transformed prior to data analyses. Log₁₀ reductions were calculated by subtracting concentrations obtained on day “n” (where “n” represents days 1, 3, 5 or 8) from concentration at the beginning of the experiment (day 0). GraphPad Prism version 7.01 (GraphPad Software, La Jolla, CA) was used to conduct a two-way analysis of variance (ANOVA with interactions) with Tukey’s multiple comparison test to evaluate the effects of two factors (indigenous microbiota including protozoa and sunlight) on decay. This software was also used to conduct the paired t-tests, one-way ANOVA and Pearson product momentum correlation to assess significant differences in light temperature measurements, across different virus measurements and to identify potential correlations trends in decay patterns, respectively.

Overall, average log₁₀ reduction on days one and eight for all treatments was greatest for adenovirus (1.48 ± 0.99), followed by F+ (0.79 ± 1.53) and somatic (0.61 ± 1.21) coliphages, although these differences were not statistically significant (p = 0.0794). After 8 days, exposure to sunlight and indigenous microbiota (Treatment A) resulted in the greatest decay for all three organisms (log₁₀ reductions of 2.92 ± 0.39, 4.48 ± 0.38, 3.41 ± 0.19 for somatic coliphages, F+ coliphages and adenovirus, respectively) (Table 1, Figs 1, 2 and 3). Exposure to sunlight only (Treatment C) resulted in log₁₀ reductions of 2.31 ± 0.20, 1.17 ± 0.01 and 1.54 ± 0.24 for somatic coliphages, F+ coliphages and adenovirus respectively, while shaded treatments (Treatments B and D) typically yielded the least decay (log₁₀ range: 0.05–1.11) (Table 1, Figs 1, 2 and 3). Decay pattern of all viruses was strongly correlated (r² range: 0.754–0.881, p value range: 0.0002 - < 0.0001), although it was the most noticeable for F+ coliphage and adenoviruses (r² = 0.881, p < 0.0001).

After 1 day of exposure, neither variable (ambient sunlight and protozoan grazing) had a significant effect on decay, and log₁₀ reductions for all treatments were negligible. Over the course of the next 48 h (day three), unfiltered treatments containing indigenous protozoa exhibited greater decay (log₁₀ reduction values: 1.17 ± 0.04 and 1.40 ± 0.12; (Treatments A and B, respectively) compared to filtered treatments (< 1 log₁₀ reduction; Treatments C and D) (Table 1, Fig. 2). While presence of protozoa was the only significant variable affecting decay at day three (Table 2) and contributing ~ 86% to variation in the data set, the interaction between variables was also significant (p = 0.0026) indicating that the effect of protozoa was dependent on sunlight exposure (Table 2). Forty-eight hours later (day five), effects of protozoa began to diminish (Fig. 1, Table 1), and sunlight irradiation became the dominant factor affecting the decay (log₁₀ reduction values: 1.34 ± 0.17 and 1.66 ± 0.19 and for Treatments A and C, respectively) and contributing ~ 94% to the observed variation in decay (Table 2). At day eight, which was the final time point, solar irradiation continued to be the dominant factor (~ 95% contribution to the observed variability in the dataset) and was the only influential variable (Table 2) causing the log₁₀ reduction values of 2.92 ± 0.39 and 2.32 ± 0.20 for Treatments A and C, respectively (Table 1, Fig. 1).

Similar to somatic coliphages, decay of F+ coliphages was minimal within the first 24 h of exposure and neither variable had significant effect on decay (Tables 1 and 2, Fig. 2). At day three, 48 h later, exposure to indigenous microbiota had a significant effect on decay contributing ~ 50% to the observed variations in the data set with minimal interactions (Table 2).

The greatest decay occurred in Treatment A (exposure to sunlight and biota; log₁₀ 1.62 ± 0.22), followed by Treatment B (exposure to biota only; log₁₀ 1.14 ± 0.17) and finally Treatments C and D (exposure to sun only and exposure to neither variable; < 1.00 log₁₀ reduction each) (Table 1, Fig. 2). Presence of protozoa continued to significantly affect the decay of F+ coliphages at day five, but its contribution to the variability was less (~ 18%). Exposure to sunlight became a dominant significant variable on day five, contributing nearly 50% to the observed variation in the dataset (Table 2). Overall, the greatest decay occurred for the treatment containing both, indigenous protozoa and sunlight (log₁₀ reduction: 3.16 ± 0.11), indicating that the effect of variables was co-dependent (Table 1). During the final time point (day eight), the effect of both variables (as well as their interaction) continued to be statistically significant and their contribution to the decay remained similar to that at day five (p = 0.0001; Table 2). Decay continued to be the most pronounced when F+ coliphage were exposed to both variables (log₁₀ reduction Treatment A: 4.48 ± 0.38), followed by exposure to sunlight only (log₁₀ reduction Treatment C: 1.43 ± 0.10) while the decay in the remaining two treatments was negligible (Table 1, Fig. 2).

Although decay data for adenovirus is limited, similar to both coliphages, neither variable had a significant effect on decay within the first 24 h of exposure (Table 2, Fig. 3). Over the course of 8 days, both sunlight and indigenous microbiota were significant factors contributing to the decay of adenovirus (Table 2, Fig. 3). Sunlight was more important variable contributing nearly 50% to the observed variations in the data set, followed by interactions between the variables (~ 24%) and indigenous biota (~ 19%) (Table 2). In sunlight treatments, adenovirus reduction in presence of protozoa (Treatment A) was approximately 2 log₁₀ greater compared to the reductions in their absence (Treatment C) (3.41 ± 0.19 vs 1.54 ± 0.24) (Table 1, Fig. 3), whereas the reduction in dark treatments was ~ 1 log_10.