A rigorous assessment and comparison of enumeration methods for environmental viruses
Bacteriophages
Four lytic E. coli-specific phages were used in the present study: MS2 (DSM 13767), T4 (DSM 4505), T7 (DSM 4623), and ϕX174 (DSM 4497). The genomic and structural properties of the phages as well as their bacterial hosts are listed in Table 2. For preparation of the virus isolate stocks, the respective bacterial host was grown in sterile LB medium (LB broth Miller, Sigma-Aldrich, St. Louis, Missouri) until an optical density of 0.3 measured at 600 nm was reached, then inoculated with phages at a virus-to-bacteria-ratio of 0.1, followed by overnight incubation. Remaining bacterial cells were killed by the addition of 1/10 volume of chloroform for 1 h. After separation from the bacterial cell debris, virus stocks were filtered with 0.22 µm syringe filters (Millex-GP, Merck-Millipore, Billerica, Massachusetts) and filtration was repeated prior preparation of samples for measurements.
Environmental samples
Environmental samples were collected from four different aquatic habitats: the income water tank of a wastewater treatment plant (Gut Großlappen, Munich, Germany), an on-site groundwater collection well (48°13′25.8″ N 11°35′45.4″ E, Munich, Germany), a lake (Feldmochinger See; 48°12′56.0″ N 11°30′49.4″ E, Munich, Germany), and a river (Isar; 48°32′59.3″ N, 12°10′42.4″ E, Landshut, Germany). To remove particles the size of bacteria and larger, all water samples were filtered with 0.22 µm syringe filters (Millex-GP). Measurements with flow cytometer and nanoparticle tracking analysis were performed simultaneously and on the sampling day. Quantification with epifluorescence microscopy as well as DNA extraction was conducted on the next day. Samples were stored in 4 °C.
Additionally, a mixed water sample (lake and wastewater) with an approximate concentration of 108 virus-like particles per mL (VLP mL–1) was prepared. This sample was spiked with 1× 108, 5× 108 and 1× 109 T4 particles mL−1. Before the addition, phage T4 stock has been quantified with qPCR.
Viral quantification
All measurements were performed in biological and technical duplicates.
Plaque assay (PA)
The PA was performed using a soft agar overlay technique as described elsewhere24. Briefly, 0.5 mL of appropriate dilutions of phages were mixed with an equal volume of fresh cultures of the corresponding hosts, grown overnight (incubated in LB medium at 37 °C until an optical density of 0.3 measured at 600 nm was reached). The phage-bacteria-suspension was mixed with 3 mL warm soft agar (0.75% w/v agar and 2.5% w/v LB) and gently poured on a petri dish already containing an LB agar layer (1.5% w/v agar and 2.5% w/v LB) in biological and technical replicates. Upon solidification, the petri dishes were incubated bottom up for overnight at 37 °C. After 15–20 h, depending on the bacterial growth efficiency, the plaques formed were manually counted and the phage titers as plaque-forming units per mL (PFU mL–1) were calculated.
Flow cytometry (FCM)
All samples were prepared as described previously with some adaptations14. We decided on these modifications based on the publications of Tomaru and Nagasaki (2007) and Brum and colleagues (2013). More precisely, samples were not fixed with glutaraldehyde after sampling as this may decrease the fluorescence intensity as well as the viral counts. Tomaru and Nagasaki concluded, that a fixation does not necessarily improve the staining ability of the virus particles20. Besides, our samples were measured immediately on the day of sampling, thus a preservation of the viral particles was not necessary. Another step recommended by Brussaard (2004) we did not follow is the flash freezing of the viral sample in liquid nitrogen. It has been shown that nitrogen fixation hampers the preparation procedure for TEM resulting inter alia in morphology changes25. To what extent particles would be enumerated correctly after fixation and nitrogen treatment with nanoparticle tracking analysis where particle integrity would certainly play a role during the enumeration process, is also debatable. As consequence, we decided, to omit this step in order to maintain a consistent sample handling and accomplish comparable conditions for all methods.
In brief, samples were diluted appropriately with sterile, filtered PBS buffer (0.02 µm Anotop 25 syringe filter, Whatman, Maidstone, UK; Sigma Aldrich) to fulfill the instrument’s optimal concentration requirements of approximately 106 VLP mL–1 (Table 1). Fluorescent TRUCOUNT beads (BD, Becton, Dickinson and Company, Franklin Lakes, New Jersey) were added to each sample as an internal reference. The samples were stained with 1 × SYBR gold nucleic acid stain (Thermo Fisher, Waltham, Massachusetts) and incubated either for 10 min at 80 °C (FCM80) or for 1 h at 30 °C (FCM30) prior to measurement. Tomaru & Nagasaki recommended an incubation at room temperature, as higher temperatures reduced the viral counts. We chose therefore two staining temperatures, one at 80 °C, following the suggestion of Brussaard14 and one at 30 °C, following the reference of Tomaru & Nagasaki20.
All samples were measured with a FC500 flow cytometer equipped with an air-cooled 488 nm Argon ion laser (Beckman Coulter, Brea, California) in biological and technical replicates. Analysis and evaluation of the samples was performed using StemCXP Cytometer software (v2.2).
Nanoparticle tracking analysis (NTA)
Viral isolate samples were diluted appropriately with sterile phage buffer (10 mM Tris [pH 7.5], 10 mM MgSO4, and 0.4% w/v NaCl) to obtain the optimal concentration range of 107–109 VLP mL–1 (Table 1). Afterwards, samples were either untreated or stained with 1 × SYBR gold for 10 min at 80 °C or 1 h at 30 °C (NTA80 or NTA30, respectively). Each sample was injected manually into the machine’s specimen chamber with a sterile 1 mL syringe (Braun, Melsungen, Germany), and measured three times for 20 sec at room temperature in three independent preparations. Samples were measured using a NanoSight NS300 (Malvern Pananalytical Ltd., Malvern, United Kingdom) equipped with a B488 nm laser module and a sCMOS camera, following the manufacturer’s protocol. Analysis was performed with the NTA 3.1 Analytical software (release version build 3.1.45).
Epifluorescence microscopy (EPI)
Staining of the samples was carried out as described by Patel et al.26. Briefly, all samples were diluted appropriately with 0.02 µm filtered 1 × TE buffer (pH 7.5, AppliChem, Darmstadt, Germany) to a concentration of 107 particles mL–1. For environmental samples with lower concentrations, a volume of 10 mL was used.
Then, 1 mL of each diluted sample (10 mL of environmental samples) was passed through a 0.02 µm Anodisc filter (Whatman) in duplicates. After complete desiccation, the filter was stained using a drop of 2 × SYBR gold dye (Thermo Fisher) with the virus side up, and incubated at room temperature for 15 min in the dark. Stained filters were mounted on a glass slide with 20 µL antifade solution (Thermo Fisher). Slides were analyzed using an Axiolab fluorescence microscope (Carl Zeiss, Oberkochen, Germany) equipped with a 488 nm laser. A camera was used to take ten pictures per sample, which were analyzed using ImageJ (version 1.50i). Numbers of particles on the whole filter were calculated by multiplying the counts with the quotient of the area of the filter by area of the pictures.
Quantitative real-time PCR (qRT-PCR)
Prior to the DNA extraction 1 mL of sample has been treated with DNase as described previously with a modified incubation procedure for one hour at 37 °C27. The DNA extraction has been conducted from the complete volume after DNase treatment using the Wizard® PCR Preps DNA Purification Resin and Minicolumns (Promega, Madison, Wisconsin) as previously described28. RNA was extracted with a QIAmp MinElute Virus Spin Kit (total volume of 1 mL sample) (Qiagen, Hilden, Germany) and cDNA was synthesized using a DyNAmo cDNA Synthesis Kit (Thermo Fisher) according to the manufacturers protocols. For all samples, DNA or RNA was isolated in duplicates.
T4 was quantified using primers amplifying a 163 bp region of the gp18 tail protein (T4F 5′-AAGCGAAAGAAGTCGGTGAA-3′ and T4R 5′-CGCTGTCATAGCAGCTTCAG-3′)29. For T7, primers amplifying a 555 bp segment of gene 1 (T7_4453F 5′-CTGTGTCAATGTTCAACCCG-3′ and T7_5008R 5 ‘-GTGCCCAGCTTGACTTTCTC-3′)30. ϕX174 was quantified using primers specific for the capsid protein F (ϕX174F 5′-ACAAAGTTTGGATTGCTACTGACC-3′ and ϕX174R 5′-CGGCAGCAATAAACTCAACAGG-3′) resulting in a 122 bp fragment31. For MS2, primers amplifying a 314-bp fragment (MS2_2717F 5′-CTGGGCAATAGTCAAA-3′ and MS2_3031R 5′-CGTGGATCTGACATAC-3′) were used32. Quantitative PCR was performed in a total volume of 20 µL consisting of 10 µL Brilliant III Ultra-Fast QPCR Master Mix (Agilent, Santa Clara, California), 5 µL DNA template or PCR-grade water as a negative control, as well as the following optimized primer concentrations (supporting information): 0.5 µM primers T4F and T4R, 0.8 µM primers T7_4453F and T7_5008R, 0.6 µM primers ϕX174F and ϕX174R, or 0.3 µM primers MS2_2717F and MS2_3031R, respectively. The amplifications were run on a Mx3000P qPCR system (FAM/SYBR® Green I filter [492 nm–516 nm], OS v7.10, Stratagene, San Diego, California) with the following cycling conditions: T4: 95 °C for 10 min, (95 °C for 15 sec, 60 °C for 1 min, 72 °C for 1 min) for a total of 45 cycles, T7: 95 °C for 12 min, (95 °C for 30 sec, 58 °C for 30 sec, 72 °C for 1 min) for a total of 30 cycles, ϕX174: 94 °C for 3 min, (94 °C for 15 sec, 60 °C for 1 min) for a total of 40 cycles, and MS2: 95 °C for 10 min, (95 °C for 15 sec, 50 °C for 30 sec, 72 °C for 30 sec) for a total of 45 cycles. Each replicate was measured four times. Analysis of the melting curves confirmed the specificity of the chosen primer as no variations compared to the standard melting curves could be observed. Standard curves were prepared using the appropriate dilutions of gblocks gene fragments (IDT, Coralville, Iowa) of the respective viral DNA in PCR-grade water (supporting information, Tables S1 and S2). Data analysis was performed using the manufacturer’s MxPro Mx3000P software (v4.10).
TEM preparation
Although TEM may be used for quantification, only the virus morphology and integrity upon applying the staining conditions were monitored. Therefore, the phages MS2 and T7 were either incubated for 10 min at 80 °C or further processed without any temperature treatment. Ten µL of the sample were then applied to the carbon side of a carbon-coated copper grid. Excessive water was blotted dry with a filter paper and washed two times with double-distilled water. After each washing step grids were again blotted dry onto a filter paper before negative staining with 2% uranyl acetate for 20 sec. The staining liquid was blotted onto a filter paper and the grids were air-dried as described previously33. Transmission electron microscopy was carried out using a Zeiss EM 912 with an integrated OMEGA filter in zero-loss mode. The acceleration voltage was set to 80 kV and images were recorded using a Tröndle 2 k × 2 k slow-scan CCD camera (Tröndle Restlichtverstärker Systeme, Moorenweis, Germany).
Sample stability test
In order to substantiate our decision of omitting a fixative step for FCM measurements and to confirm a certain stability of the virus concentration over a short time range (few days), phage T4 and wastewater samples were measured with FCM at time 0, after 24 h and after 48 h. The samples were either kept in 4 °C or were fixed with 0.5% glutaraldehyde for 30 min in 4 °C followed by freezing in liquid nitrogen with adjacent storage at -80 °C, as suggested by Brussaard (2004). At each time point, samples were prepared for FCM as described above with two different staining procedures (30 °C and 80 °C). Additionally, a fixed T4 phage sample was prepared for NTA measurements in the same way in order to test the usability of glutaraldehyde fixation. For phage T4, measurements of the 4 °C, unfixed samples were mostly slightly higher compared to the fixed samples (Fig. S6a,b). Comparing the initial quantification with the results after 48 h, the decrease in counted particles was minor. For the wastewater samples, viral numbers of the unfixed samples were marginally lower, however, a general decline in particle numbers over time could be observed (Fig. S6c,d). This decline was in all cases less than one order of magnitude. As both, fixed and unfixed samples declined only to a small extent and no trend of a stronger decrease of viral particles in the unfixed samples could be observed, omitting the fixation with glutaraldehyde and liquid nitrogen is not supposed to have a wide influence on the enumeration within 48 h.
Statistical analysis
Statistical analysis was carried out in R (v3.4.3) and RStudio (v1.1.383). Data were log transformed and analysis of variance (ANOVA) was conducted. Normal distribution of data was confirmed by density plots and quantile–quantile plots; homogeneity of variances was confirmed with Levene’s test. Afterwards, multiple pairwise comparisons were calculated with a post-hoc Tukey honest significant differences test. In addition, similarities in viral isolate quantification methods were assessed using principal coordinate analysis. More
