Complete genome analysis of a frog virus 3 (FV3) isolate and sequence comparison with isolates of differing levels of virulence

In order to determine if FV3 strains SSME, wt-FV3, and aza-C^r induced different degrees of infection, L. pipiens tadpoles were exposed to each strain. Tadpoles were exposed to FV3-infected water for 12 h before being transferred to FV3-free water for the remainder of the experiment. The tadpoles were monitored for 41 days, at which time all tadpoles had either died or reached metamorphosis. The results of the survival analysis revealed significant differences in strain effect on tadpole mortality rates (X² = 21.3, p < 0.01). Although deaths triggered by wt-FV3, aza-C^r, and SSME stocks peaked between 15 and 18 days post infection, overall mortality was higher following infection with wt-FV3 and aza-C^r virus (>96%) than with SSME (84%). Wt-FV3 induced a mortality of 97%, and aza-C^r a mortality rate of 96% (Figure 1). In contrast, SSME infected tadpoles reached a mortality plateau after 28 days, and showed a mortality rate of 84%. The control tadpoles had the lowest mortality, with a rate of 66% after 41 days (Figure 1). These data demonstrate that the closely related FV3 strains, wt-FV3 and aza-C^r[21, 22], are associated with higher mortality than SSME (Figure 1). Of concern was the high mortality seen even in control tadpoles (Figure 1). We did confirm that the control tadpoles, both those that survived and those that died, were ranavirus negative (data not shown). This mortality could be the result of overcrowding of tadpoles during infection. In addition, this mortality is occurring during the first few days of the experiment at a time when the tadpoles are very small and very weak, very sensitive to stressful conditions. At the beginning of the experiment, the tadpoles are transferred from tank to tank that may account for the high mortality rate early on in the experiment or may have resulted in high mortality of a sublethal viral dose.

In order to better understand the possible genetic basis of phenotypic variation among strains, we sequenced the SSME strain using 454 GS-FLX technology. Our sequenced genome showed high similarity to the published FV3 sequence (wt-FV3) [12], with a sequence identity of 98.79% and average genome sequencing coverage of 51x. Results revealed that SSME differed from the wt-FV3 genome length of 105,903 bps and ORF number of 98; instead, SSME had a length of 105,070 bps, and a total of 95 predicted ORFs (Table 1). Despite high overall sequence identity between SSME and FV3, marked differences were noted in several regions of the genome (Table 2).

In order to investigate the possible genetic variation within coding regions and regions of high variability between closely related viral isolates, we compared our sequenced SSME genome with four published ranavirus genomes. The genomes compared included: SSME, wt-FV3 [17], rana grylio iridovirus (RGV) [23], soft-shelled turtle iridovirus (SSTIV) [24], and tiger frog virus (TFV) [25]. We began by focusing our analysis on predicted coding regions, as variation in these areas may have consequences in pathogenicity or the kind of disease a virus causes.

Upon analysis of selected ranaviral genomes, a 757 bp deletion was discovered in SSME, deleting the entire 65L coding region and the majority of the 66L coding region (Figure 2). This deletion was not present in wt-FV3 or in related ranaviruses RGV, SSTIV, and TFV. However, sequence alignments indicated reorganization of 66L in RGV, SSTIV, and TFV, with a 139 bp insertion found in their corresponding 66L regions (Figure 2). Therefore, there are at least 3 different genomic presentations of this region in ranaviral genomes, making it highly variable among isolates.

Another coding region demonstrating variation was 50L. An in frame deletion of 13 amino acids was found in 50L when compared to wt-FV3, present only in the SSME genome (Figure 3). In addition, 50L had a single nucleotide deletion resulting in a frameshift mutation in SSME, RGV, and SSTIV when compared to wt-FV3. This frameshift mutation resulted in the loss of 50L’s stop codon, which led to the combination of 49L and 50L into a single ORF. These two ORFs merged in frame, creating 49/50L. In contrast, TFV did not have this single nucleotide deletion, nor did it display the merger of its 49L and 50L equivalent ORFs (Figure 3). This is consistent with TFV being the most divergent virus of the group.

Further ranaviral gene variation was discovered in 43R in the form of a single nucleotide deletion, resulting in a frameshift mutation. Similar deletions were found in SSME, RGV, SSTIV, and TFV, leading to pronounced differences in amino acid sequences when compared to wt-FV3 (Figure 4). It is worth noting that although SSME, RGV, and SSTIV share the same frameshift mutation the genes are not identical as there are several single amino acid changes that lead to modest variation (Figure 4). In contrast, despite having a similar nucleotide deletion, TFV’s proposed amino acid sequence was markedly different from other analyzed ranaviruses. Another single nucleotide deletion was found in 46L (Tables 1 and 2), which led to the loss of the original stop codon and the extension of the ORF by 319 bps (Tables 1 and 2). This was found in all genomes other than TFV and wt-FV3. These data demonstrate that even between very similar viral genomes, there is considerable gene variation, and that when viruses differ in pathogenicity, it could be related to variation in coding regions.

Genomic sequencing, and subsequent analysis, of SSME in comparison to wt-FV3, provided a general overview of the genetic variation between closely related FV3 isolates, particularly in coding regions. We then decided to explore possible genetic variation in highly variable repeat regions, and thus define these genetic regions with greater inter-strain variability. To achieve this, we chose three repeat regions (suspected to be polymorphic in copy number) to investigate. These regions included: Region 1 (22,499-22,574 bps); Region 2 (52,443-52,747 bps); and Region 3 (54,948-54,986 bps), based on the wt-FV3 sequence. We performed a repeat analysis on our sequenced SSME genome that we had previously evaluated for coding region variability, and reference genomes RGV, SSTIV, and TFV. We also decided to analyze the repeat regions of wt-FV3 and aza-C^r in order to check for repeat number stability between the two strains over the multiple viral passages they have undergone since the creation of aza-C^r in 1987 [22]. In addition, we expanded our analysis to include 6 environmental samples isolated from the same waterway in Manitoulin, Kagawong, ON, Canada. These isolates were designated as: E3, E4, E5, F4, F6, and G4; a lab FV3 sample (ATCC® VR-567™) was also used [26].

The information acquired through analyzing these repeat regions indicated that variation in repeat copy number is present between ranaviral isolates, including those from one geographic location (Figure 5). In the first region, repeat copy number ranged from 2–5 copies, with 2 copies being the most common and 5 copies being unique to isolate F6. Wt-FV3 and aza-C^r shared a distinctive repeat sequence, containing a G rather than an A at the 9^th nucleotide position (Figure 5). Region 2 proved to be more variable than region one, with repeat copy number ranging from 4–18 copies, while Region 3 was found to consist of two repeat sequences. These Region 3 repeats varied in copy number between isolates, with a maximum copy number of 4. However, in some instances, the first sequence of nucleotides characteristic of this region did not repeat (Figure 5).

Although copy number variation was present amongst isolates, when samples were compared across all three repeat regions, some were found to be identical; such isolates included F4, E5, and E3 (Figure 5), which were isolated from the same geographic region. In addition, isolates wt-FV3 and aza-C^r were also found to be identical across repeat regions (Figure 5), which is consistent with their common origin [22]. This data taken as a whole suggests that viral isolates, even those from one geographic location, can display both variation and identical patterns across repetitive regions of the genome. Thus, these repeats could act as a fingerprint to discriminate between viral isolates.

FV3 isolate SSME, wt-FV3, and aza-C^r, were analyzed in this study and were provided by Professor Gregory Chinchar from the University of Mississipi Medical Center (Jackson, MS, USA). SSME was isolated from a wild, spotted salamander population in Maine, USA, while aza-C^r was derived from the laboratory wt-FV3 strain through selection with azacytidine [22]. Both wt-FV3 and aza-C^r have been previously sequenced and found to be identical, with wt-FV3′s nucleotide sequence deposited into GenBank [12]. Amphibian renal cells (A6 cells) were supplied by Niels Bols of the University of Waterloo and maintained in Leibovitz’s L-15 media (Invitrogen, Burlington, ON) supplemented with 10% fetal bovine serum (FBS; Invitrogen, Burlington, ON), penicillin (100 U/ml), and streptomycin (100 μg/ml).

In order to monitor survival of FV3 infected tadpoles, we obtained Lithiobates pipiens tadpoles, approximately Gosner stage 25 [37], from the Environment Canada Atlantic Laboratory for Environmental Testing in Moncton, NB., courtesy of Paula Jackman. The animals were received two weeks prior the beginning of the experiment and were kept in 20 L tanks filled with 10 L of aged clean dechlorinated water. Each treatment group was done as 2 replicates of 25 tadpoles per treatment. Groups tadpoles were placed in dechlorinated water, with the host density (number of tadpoles per volume of water) adjusted to 1 tadpole per 250 mL to avoid any effect of density on tadpole development [38]. To infect, 25 tadpoles were placed in 50 mL of infected water containing 10,000 pfu/mL of a FV3 strain (SSME, wt-FV3, and aza-C^r). According to past experiments, such concentration is known to induce sublethal effects in these laboratory conditions [38]. Control individuals were placed within 50 ml of FV3-free water. The tadpoles were left within the infected solution overnight (12 hours) tadpoles were then transferred together with the contaminated water in 2 L plastic containers filled with 1 L of dechlorinated water (aged for three days) for the rest of the experiment. Containers were held in a climatic chamber (Thermo Incubator Model 3740) where the temperature was set to remain at 22°C with a 12 h:12 h dark:light cycle. Tadpoles were fed on a weekly basis after the water was changed with standard tadpole food (Carolina Biological Supply Company, Burlington, NC) at 30 mg/tadpole for week 1, 60 mg/tadpole for week 2, and 120 mg/tadpole for week 3 until the end of the experiment [38]. Starting on week 3 the water in each tank was replaced once a week with clean dechlorinated aged (24 h) water. As a result, exposed tadpoles were held in virus-containing water for 3 weeks, a period which is long enough for tadpoles to be in close proximity with residual infection [38]. Tanks were monitored on a daily basis. Dead tadpoles were removed to prevent any scavenging, and stored at −25°C in individual plastic vials with ethanol for subsequent analyses.. The experiment terminated when all the individuals died or reached metamorphosis. The procedures used in this experiment follow protocol #2010-04-02 approved by the Laurentian University Animal Care Committee.

In order to check for ranavirus infection, all animals (including euthanized ones) were dissected to remove the liver that was then crushed into a 1.5 ml Eppendorf tube. The resulting tissue mixture was used for DNA extraction. DNA was extracted using QIAmp DNeasy Kit following the standard protocol (Qiagen). After extraction, a double blind PCR was performed using a primer set known to successfully amplify a portion of the major capsid protein within the FV3 genome: MCP-ranavirus-F (5′-GACTTGGCCACTTATGAC-3′) and MCP-ranavirus-R (5′- GTCTCTGGAGAAGAAGAA), following the PCR conditions listed in Mao et al. [33], using 1.5 μl of template DNA and cycled 40 times. Individuals showing two positive amplifications for both PCRs were considered infected. We analyzed host survival using a survival analysis and failure time analysis following the Kaplan & Meier product limit method associated with Chi square and Gehan’s Wilcoxon tests (multiple and two sample comparisons respectively) [39]. Individuals surviving to the end of the experiment were censored to account for our lack of information about their true time of death [40].

The SSME viral isolate was propagated on a confluent monolayer of A6 cells, and grown in a 75 cm² flask, with cells infected at a multiplicity of infection (MOI) of 1 PFU/cell. The cells were harvested 5 days post infection (once cytopathic effect appeared), and viral DNA was extracted using the Purelink Viral RNA/DNA Mini Kit according to manufacturer’s protocol (Invitrogen, Burlington, ON).

Standard kits and protocols developed by the manufacturer were used to sequence the SSME sample on a 454 GS-FLX platform (Roche Diagnostics Corporation). Briefly, a Rapid Library Preparation Kit (Roche, Mississauga, ON) was used to mechanically shear 500 ng of template DNA into short fragments. A universal sequencing primer that included a short DNA sequence unique to the sample (MID tag) was then annealed to both ends of each DNA fragment. A GS Junior Titanium Emulsion PCR Kit (Roche, Mississauga, ON) was used to amplify the sample library, which was sequenced using a GS Junior Titanium Sequencing Kit (Roche, Mississauga, ON). In order to assemble a full genomic sequence, the short sequences produced by 454 sequencing were aligned with the reference FV3 genome, wt-FV3, using GS Reference Mapper (Roche, Mississauga, ON). Any gaps in the assembled genome were then sequenced using custom PCR primers specific to each gap, with sequencing performed by the Robarts Sequencing Facility (London, ON). The final genomic sequence was deposited in GenBank accession number KJ175144.

FV3 environmental samples were collected from frogs caught by hand at various sites along a lakeshore in Manitoulin, Kagawong, ON, Canada (Latitude: 45.86418, Longitude: −82.27150). The frogs were caught using disposable gloves which were changed between each animal inspection. This method is preferred to the ‘net-catching method’ as it has been suggested that cross contamination can occur via the net. Each individual was toe clipped following the protocol #2009-03-04 approved by the Laurentian University Animal Care Committee for tissue sample collection. DNA was then isolated from toe clippings using the DNeasy Blood & Tissue Kit according to the manufacturer’s protocol for total DNA extraction from animal tissues (Qiagen, Mississauga, ON). Samples then underwent PCR with primers designed to amplify specific repeat regions identified in Eaton et al. [19]. Primers used included: Region 1-F: CGTGGTCAGACTGGTCCTCG; Region 1-R: CACCTCTGTCTCTGAATCGG; Region 2-F: GAGTTTACTTGGTGGCCATG; Region 2-R: TCCTGTCAAGAGATCCCCTC; Region 3-F: CTTGCTGCTGCCGTTCAGGC; and Region 3-R: AGAGTGAAAAAGGTAAAGGC.

In a PCR reaction tube the following reactants were combined: 10X PCR buffer (Invitrogen, Burlington, ON), 50 mM MgCl₂ solution (Invitrogen, Burlington, ON), 5X TAQ DNA polymerase (Invitrogen, Burlington, ON), 10 mM deoxyribonucleotide triphosphates (dNTPs), 0.1 mM primer, 2.5 ng DNA, and water to a final volume of 50 μl. The reactions were then placed in a thermocycler under the following conditions: 94°C for 3 minutes, 94°C for 30 seconds, 56°C for 1 minute, and 72°C for 1 minute for 30 cycles. Sequences of PCR products were determined by Robarts Research Institute DNA Sequencing Facility in London, ON, and were analyzed using BioEdit v7.0.5.