Characterization of Frog Virus 3 Knockout Mutants Lacking Putative Virulence Genes

Abstract

To identify ranavirus virulence genes, we engineered Frog Virus 3 (FV3) knockout (KO) mutants defective for a putative viral caspase activation and recruitment domain-containing (CARD) protein (Δ64R-FV3) and a β-hydroxysteroid dehydrogenase homolog (Δ52L-FV3). Compared to wild type (WT) FV3, infection of Xenopus tadpoles with Δ64R- or Δ52L-FV3 resulted in significantly lower levels of mortality and viral replication. We further characterized these and two earlier KO mutants lacking the immediate-early18 kDa protein (FV3-Δ18K) or the truncated viral homolog of eIF-2α (FV3-ΔvIF-2α). All KO mutants replicated as well as WT-FV3 in non-amphibian cell lines, whereas in Xenopus A6 kidney cells replication of ΔvCARD-, ΔvβHSD- and ΔvIF-2α-FV3 was markedly reduced. Furthermore, Δ64R- and ΔvIF-2α–FV3 were more sensitive to interferon than WT and Δ18-FV3. Notably, Δ64R-, Δ18K- and ΔvIF-2α- but not the Δ52L-FV3 triggered more apoptosis than WT FV3. These data suggest that vCARD (64R) and vβ-HSD (52L) genes contribute to viral pathogenesis.

INTRODUCTION

Ranaviruses such as Frog Virus 3 (FV3) are emerging pathogens that cause severe morbidity and mortality among fish, amphibians and reptiles worldwide (Chinchar et al., 2009; Duffus et al., 2015). The increase in prevalence and the expansion of host range suggests that ranaviruses are successful in overcoming host immune defenses. Although the general outlines of the FV3 replication cycle are known and 19 iridovirus genomes have been sequenced (Chinchar et al., 2009; Jancovich et al., 2015; Jancovich et al., 2010; Tan et al., 2004; Williams et al., 2005), the precise functions of most viral genes are still unknown. Previously a number of temperature-sensitive mutants were isolated and have proven useful in identifying genes essential for virus replication (Chinchar and Granoff, 1986; Goorha and Dixit, 1984; Goorha et al., 1981). In addition, transient knock down of viral gene function using antisense morpholino oligonucleotides (Sample et al., 2007) or siRNA (Whitley et al., 2011) have also elucidated the function of several viral genes. However, the random nature of temperature sensitive mutants and the inability to readily perform knock down in vivo, limits the usefulness of these approaches, especially if one wishes to target virulence genes. As an improvement on these approaches, we and others have recently used homologous recombination to directly knock out specific ranavirus genes and assess their roles in virus replication and virulence (Chen et al., 2011; Jancovich and Jacobs, 2011a).

Among putative ranavirus immune evasion genes, the viral homolog of the cellular translation factor eIF-2α (vIF-2α) has received attention as an antagonist of protein kinase R (PKR) (Beattie et al., 1995; Rothenburg et al., 2011). Several ranaviruses including Epizootic Haematopoietic Necrosis Virus (EHNV; (Essbauer et al., 2001)); Ambystoma tigrinum Virus (ATV, (Jancovich and Jacobs, 2011)); and Rana catesteiana Virus Z (RCV-Z, (Rothenburg et al., 2011)) encode full-length vIF-2α genes. Furthermore, in both ATV and RCV-Z, vIF-2α was postulated to act as a pseudo-substrate and block PKR-mediated translational inhibition and cell death (Rothenburg et al., 2011; Jancovich and Jacobs, 2011a). In addition, vIF-2α may also play a role in the degradation of PKR following ATV infection (Jancovich and Jacobs, 2011a). In contrast to the above ranaviruses, the FV3 vIF-2α gene is truncated and lacks the N-terminal PKR binding domain and the central helicase domains (Chen et al., 2011). Thus, the precise functional role of the FV3 vIF-2 remains in question.

In addition to vIF-2α, a Caspase Activation and Recruitment Domain (CARD)-containing gene (vCARD) has also garnered attention as a putative immune-evasion protein. Typically CARD motifs modulate interactions among CARD-containing cellular proteins (Kawai and Akira, 2009, 2010). Cellular signaling molecules containing CARD domains include pro-apoptotic proteins, pro-inflammatory molecules and proteins participating in the cellular interferon responses, e.g., RIG-I and MAVS (Besch et al., 2009; Meylan et al., 2005). Ranavirus vCARD is postulated to interact with one or more of these signaling molecules and to either block apoptosis or impair interferon (IFN) induction.

Lastly, a ranavirus homolog of β-hydroxysteroid dehydrogenase (vβHSD) has been identified as another possible immune evasion protein. βHSD homologs are present within poxviruses and have been shown to play a role in dampening host immune responses by elevating the levels of steroid hormones (Moore and Smith, 1992; Sroller et al., 1998). Whether the putative ranavirus homolog of βHSD functions in the same way remains to be determined.

We have recently developed a robust, dual-selection system (consisting of the puromycin-resistance gene fused to the gene for enhanced green fluorescent protein, PuroR/GFP) and successfully generated two FV3 KO mutants in which the genes encoding the 18K immediate-early protein and vIF-2α were replaced with PuroR/GFP (Chen et al., 2011). Here we use this technique to target two putative ranavirus virulence genes, vCARD (ORF 64R) and β-HSD (ORF 52L). In the first part of this report, we describe the generation and characterization of these two KO mutants, whereas in the latter sections we examine these two mutants along with two previously isolated ones lacking vIF-2α and the 18 kDa immediate early protein. The latter studies examined the ability of the KO mutants to replicate in Xenopus A6 cells, induce and respond to IFN, and trigger apoptosis. Marked differences between these four KO mutants and WT virus following infection of A6 cells suggest that the four targeted genes play critical roles in virus replication in Xenopus cells and modulate interferon and apoptotic responses.

RESULTS

Generation and isolation of FV3 knockout mutants

To generate FV3 knockout (KO) mutants, a recombination cassette was constructed in which the puromycin resistance gene fused to the EGFP gene by a five amino acid linker was placed under the control of the strong FV3 18K promoter (18Kprom-Puro-EGFP). The cassette was transfected into FV3-infected BHK cells and introduced into the FV3 genome at regions encoding vβHSD (ORF 52L;57,481 – 58,548) and vCARD (ORF 64R; 75,529 – 75,816) using homologous recombination (Fig 1A). The resulting KO mutants (Δ52L- and Δ64R-FV3; Table 1) were initially selected for growth in the presence of puromycin and subsequently plaques purified until all plaques were GFP-positive (Fig. 1B).

Figure 1. Generation and selection of FV3 KO mutants.

(A) Schematic of site-specific integration of the 18Kprom-Puro-EGFP cassette into the FV3 genome. Constructs consisting of the 18Kprom-Puro-EGFP cassette and regions (approximately 500 bp) flanking the targeted insertion sites (gray) were introduced into pBluescript SK(+). The recombination vectors (p18Kprom-Puro-EGFPs) were transfected into WT FV3-infected BHK-21 cells and recombinant FV3 generated by homologous recombination. Sequential selection was performed based on virus replication in the presence of puromycin (50 μg/ml) and expression of EGFP were performed. (B) Fluorescence (top) and phase contrast (bottom) microscopy of BHK-21 cells infected with Δ52L-FV3 and Δ64R-FV3 after six consecutive rounds of selection. All the plaques produced by these recombinant viruses were EGFP-positive indicating that were not contaminated with WT virus.

Table 1.

FV3 Knock Out Mutants

Mutant Designation	Targeted ORF	Nucleotide position	Gene name	Comment	Reference
Δ18K-FV3	82R	89,450–89,923	18K	18 kDa immediate-early protein (ICP-18); non-essential for replication in BHK, FHM, and A6 cells; deletion attenuates infection in tadpoles	Chen et al., 2011; Sample et al., 2007
ΔvlF-2α-FV3	26R	32,967–33,197	vIF-2α	Truncated elF-2α homolog; postulated to maintain translation in virus-infected cells; blocks activity of PKR	Chen et al., 2011; Rothenburg et al., 2011; Jancovich & Jacobs, 2011
Δ64R-FV3	64R	75,529–75,816	vCARD	Caspase activation and recruitment domain (CARD)-containing protein; modulates protein-protein interactions	This work
Δ52L-FV3	52L	57,481–58,548	vβHSD	β-hydroxysteroid dehydrogenase; involved glucocorticoid synthesis; dampens inflammatory responses	This work

Correct insertion of the marker cassette and replacement of the targeted gene were verified by PCR using primers flanking the gene of interest. Compared to WT FV3, both KO mutants produced PCR products consistent with the size of the Puro-GFP cassette and flanking regions (3.5 kb; Fig. 2 lanes 3 and 6), whereas PCR analysis of WT virus generated products of 2.6 kbp for 52L gene and 2.0 kbp for 64R gene (Fig. 2A, lanes 2 and 5, respectively). Furthermore, we did not detect amplicons indicative of WT gene products following analysis of either KO mutant confirming that KO mutants were not contaminated with WT virus. To further confirm that insertion of the Puro/EGFP cassette replaced the gene of interest, we performed PCR assays on each putative KO clone using primers that amplified within the two targeted genes (52L or 64R) as well as within GFP and vDNA Pol II. No 52L-specific product was amplified from Δ52L-FV3, and no 64R product was detected in Δ64R-FV3 (Fig 2B). As expected, both KO mutants contained the inserted GFP gene as well as the viral DNA polymerase gene. Lastly, the presence of correct inserts was subsequently verified by cloning and sequencing each of the amplified fragments. Together, the above results confirmed the absence of WT contamination in the two newly generated KO mutants and the loss of the targeted genes.

Figure 2. Confirmation of of gene deletion by PCR.

(A). PCR was performed for each putative KO mutant using primers specific for sequences flanking the targeted gene (Table 2). PCR amplification of a band of ~ 3 kb indicated the presence of the 18Kprom-Puro-EGFP insert, whereas smaller products (2.6 kbp for 52L and 2.0 kbp for 64R) are indicative of WT products. Lanes 1 and 4: DNA size ladder; lanes 2 and 5: WT-FV3; lane 3: Δ52L-FV3; lane 6: Δ64R-FV3. (B) Confirmation of gene-specific KO by PCR using primers specific for internal regions of 52L (expected size product 484 bp), 64R (expected size product 235 bp), GFP, and vDNA Pol II. The FV3 DNA polymerase gene (vDNA Pol II) was used as a positive control. PCR products were visualized using ethidium bromide staining.

Assessment of replication in BHK and FHM cells

To determine whether deletion of the 52L or 64R genes or insertion of the GFP cassette affected growth in susceptible BHK-21 and FHM cells, we monitored viral replication and transmission by examining both single- and multiple-step growth curves (Fig. 3A and B, respectively). WT and KO mutants exhibited nearly identical replication kinetics following single-cycle infections of BHK-21 cells and multi-step infections of FHM cells. These results indicate that the 52L (vβHSD) or 64R (vCARD) genes were not essential for viral replication and spread within susceptible cultured cells and that presence of the Puro/EGFP cassette did not adversely affect viral replication in vitro.

Figure 3. Comparison of WT and KO mutant replication in cell culture.

(A) One-step growth curve assay in BHK-21 cells. (B) Multiple-step growth curve assay in FHM cells. Equivalent numbers of cells were infected with either WT-, Δ52- or Δ64- FV3 at an MOI of 5 (BHK-21) or 0.01 (FHM). Viruses were harvested at the indicated time points and quantified by plaque assay. Differences in replication between WT and KO mutants were not statistically significant as indicated using the Kruskal-Wallis test (Graphpad Prism software 5.0).

Assessment of the infectivity of FV3 KO mutants in X. laevis tadpoles

X. laevis adults naturally resist FV3 infection and clear virus within 2–3 weeks, whereas tadpoles are considerably more susceptible to infection, and most succumb to infection within a month (Gantress et al., 2003). In order to examine the importance of the targeted genes to virulence, we infected X. laevis tadpoles by immersion in water containing 1×10⁶ plaque-forming units (PFU)/ml of WT-, Δ64R- or Δ52L-FV3 and monitored tadpole survival over the course of 30 days (Fig. 4A). Compared to WT FV3, the mortality of tadpoles infected with either Δ64R- or Δ52L-FV3 was both delayed and substantially reduced. Indeed, the majority (80%) of WT FV3-infected tadpoles succumbed to the virus by 30 days post-infection (dpi), whereas only 20 to 30% of tadpoles infected with the KO mutants died (Fig. 4A). Furthermore, at 6 dpi, kidneys from tadpoles infected intraperitoneally with 1×10⁴ PFU of either Δ64R- or Δ52L-FV3 exhibited a greater than 90% reduction in viral genome levels compared to those infected with WT-FV3, indicating that viral replication of the two FV3 KO mutants was compromised in vivo (Fig. 4B). These results provide evidence that the 64R and 52L genes encode virulence and/or immune evasion proteins, and that their loss results in marked attenuation in vivo.

Figure 4. Survival and in vivo viral replication of WT-, Δ52L-, or Δ64R-FV3-infected tadpoles.

(A) Stage 50 tadpoles (12/treatment group; N=12) were infected by immersion for 1 hr. in a total volume of 5 ml containing 5×10⁶ PFU of WT-, Δ52L-, or Δ64R-FV3 or mock infected with APBS. Tadpole survival was monitored over the course of 30 days post-FV3 infection. (B) Viral replication in vivo was monitored following i.p. infection (1×10⁴ PFU) of tadpoles at stage 54–55 (3–4 weeks post-fertilization). Viral loads in kidneys were determined at 6 dpi by absolute qPCR against the FV3 vDNA Pol. Data are means representing FV3 DNA copies ± SEM for 5 individual per group. Statistical significance from the vector control: (*) P<0.05, (**) P<0.005 and (***) P<0.0001 as determined by One-Way Analysis of Variance (ANOVA).

Replication of FV3 KO mutants in X. laevis kidney-derived A6 cells

As shown above, KO mutants targeting FV3 genes 64R and 52L replicated as well as WT virus in both BHK-21 and FHM cells, which are cell lines derived from non-host species (Fig. 3). However, both KO mutants replicated poorly in vivo following infection of X. laevis tadpoles, which resulted in a marked drop in mortality compared to infection by WT virus (Fig. 4). In view of these differences, we examined viral replication in cultured A6 kidney cells derived from a natural host species, X. laevis. A6 cells were infected at an MOI of 0.6 with WT virus as well as with Δ64R-FV3, Δ52L-FV3 and two previously characterized KO mutants (Chen et al., 2011) bearing complete deletions of the genes encoding the 18 kDa immediate early protein (ORF 82R; Δ18K-FV3) and a truncated viral homolog of eukaryotic translational initiation factor 2α (ORF 26R; ΔvIF-2α-FV3). Unlike replication in BHK-21 and FHM cells in which all KO mutants grew as well as WT-FV3, viral replication of Δ64R, Δ52L and ΔvIF-2α mutants was reduced nearly 10-fold in A6 cells (Fig. 5). In contrast, replication of Δ18K-FV3 was not significantly different from the WT-FV3 in A6 cells. We conclude that 64R (vCARD), 52L (vβ;HSD) and 26R (IF-2α) genes are dispensable for viral growth in non-host cell lines, but are required for optimal replication in cell lines derived from a natural host.

Figure 5. Replication of FV3 WT and KO mutants in A6 cells.

A6 cells were infected at an MOI of 0.6 with WT-, Δ18K-, ΔvIF-2α Δ64R- or Δ52L-FV3 for 6, 24 and 48 hrs, and FV3 DNA copy number was assessed by absolute qRT PCR. Statistical significance from the vector control: (**) P<0.005 as determined by One-Way Analysis of Variance (ANOVA).

Susceptibility of FV3 KO mutants to Xenopus laevis IFN

To determine the possible involvement of interferon (IFN) responses in the replication impairment of FV3 KO mutants in A6 cells, we first monitored the expression of type I and type III IFNs following FV3 infection. As previously reported (Grayfer et al., 2014), WT-FV3 elicited modest increase of type I IFN gene expression that was statistically significant only at 48 hpi (Fig. 6A). In contrast, infection with Δ18K-FV3 resulted in a marked and accelerated induction of type I IFN expression at 24 hpi that remained elevated at 48 hr p.i.. This was not the case with Δ52L-FV3 and ΔvIF-2α-FV3 that induced similar levels of type I IFN gene expression as WT FV3. At 24 hr p.i. WT-FV3 transiently increased type III IFN mRNA levels, whereas none of the KO mutants elicited significant changes in the expression of this cytokine (Fig. 6B). It is noteworthy that at early stages of infection (6 hpi) type I and III IFN gene expression was lower than the uninfected control for all viruses.

Figure 6. Induction of type I and type III IFN.

A6 cells were infected at an MOI of 0.6 with WT-, Δ18K-, ΔvIF-2α-, Δ52L- or Δ64R-FV3 for 6, 24 and 48 hrs and (A) type I IFN and (B) type III IFN gene expression was monitored by qPCR. All gene expression was examined relative to the GAPDH endogenous control. The dashed line indicates the level of detection by qPCR. The results are means ± SEM of gene expression from two independently performed experiments. Statistically significant differences (P<0.05) from uninfected control are indicated by (*) and from WT-FV3 by (#) as determined by One-Way Analysis of Variance (ANOVA).

To more directly address the possible involvement of 64R (vCARD), vIF-2α and possibly 18K genes in the viral evasion of IFN-mediated immunity, we next examined the ability of the KO mutants to replicate in A6 cultures following exposure to a recombinant X. laevis type I IFN (rXlIFN). We postulated that if the targeted genes were involved in antiviral immune evasion, their deletion would severely compromise their replication in the IFN-treated A6 cells. To this end, A6 cells were pre-treated for 6 hrs with either control supernatants from mock-transfected insect cells processed in parallel to rXlIFN production or with rXlIFN (100 ng/mL) and then infected for 24 hrs with WT-FV3, Δ18K-FV3, ΔvIF-2α-FV3 or Δ64R-FV3. At 24 hr p.i. viral replication was monitored by plaque assay (Fig. 7). We observed that replication of ΔvIF-2α- and Δ64R-FV3 was significantly reduced by rXlIFN pre-treatment compared to Δ18K-FV3 and WT-FV3. Although rXlIFN treatment of A6 cells reduced the growth of WT- and Δ18K-FV3 by 58% and 75%, respectively, replication of ΔvIF-2α-and Δ64R-FV3 was more severely affected with reductions of 97% and 88% respectively. These findings support the notion that unlike the 18K gene, vIF-2α and vCARD (64R) gene products are critical for evading an anti-viral interferon response.

Figure 7. Assessment of WT-, Δ18K-, ΔvIF-2α- or Δ64R-FV3 replication in rXlIFN- pretreated A6 cells.

A6 cells were pretreated with 100 ng/mL of rXlIFN or equal volumes of control supernatant for 6 hrs and infected with either WT-, Δ18K-, Δ64R-, or ΔvIF-2α-FV3 at a MOI = 0.6 PFU/cell. At 24 hpi. cells were subsequently harvested, processed, and viral yields determined by plaque assays. All experiments described above employed three A6 cultures per treatment group (N=3) and all of the results are presented as means ± SEM. The different above-head letters represent statistically different groups (P<0.05) as determined by One-Way Analysis of Variance (ANOVA). All treatment groups are compared against each other and those that are not statistically significant share the same letters, whereas those that are statistically different have different letter designations. For example, the group pre-treated with rXlIFN and infected with Δ64R-FV3 is significantly different from groups infected with WT FV3 (rXlIFN pretreated or not pretreated or not) as well as from groups infected with ΔvIF-2α-FV3 (rXlIFN pretreated or not).

Assessment of apoptosis induction by WT FV3 and KO mutants

An effective interferon response often culminates in the induction of apoptosis in virus-infected cells (Liang et al., 2015). Accordingly, we hypothesized that if the targeted FV3 genes were critical for immune evasion by impairing programmed cell death, their absence should result in increased cellular apoptosis. To this end, we utilized a flow cytometry-based apoptosis assay to assess programmed cell death of A6 cells infected 8 and 24 hrs earlier with either WT or the four KO mutants (Fig. 8). Early and late stages of apoptosis were assessed by co-staining with Annexin V and propidium iodine (PI). Cycloheximide treatment was used as a positive controls and uninfected cells served as a negative control. Although little evidence of Annexin V staining was seen in uninfected cells at either time point, WT-FV3 induced detectable early-stage apoptosis at both 8 and 24 hpi (10 and 20%, respectively). In contrast, 18K, 64R, and vIF-2α KO mutants triggered about twice as much apoptosis (Annexin V⁺/PI^neg) at both 8 and 24 hr p.i. than WT FV3 or the 52L KO mutant (Fig. 8, panels A and B). Moreover, the fraction of cells double stained with Annexin V and PI (Annexin V⁺/PI⁺; Fig. 8A) indicative of late stage of apoptosis or necrotic death was negligible, i.e., less than 3%. Overall, apoptosis induced by Δ64R-, ΔIF-2α- and Δ18K-FV3 approached but did not reach the level observed following cycloheximide treatment (70%). In contrast, apoptosis induced by infection with Δ52L-FV3 was not as marked as for other KO mutants and was comparable to that of WT-FV3. It is noteworthy that despite using a high MOI of 20, apoptosis of infected A6 cells remained below 70%. This can be explained by the resistance of A6 cells to apoptotic stimuli (e.g., cycloheximide induced only 70% apoptosis at 24 hr), but may also suggest that more than one viral gene interferes with apoptotic induction. The elevated apoptosis seen following infection with FV3 mutants defective in 64R (vCARD) as well as vIF-2α and 18K supports their role as possible virulence genes, although it appears that the 18K gene is likely associated with a pathway different from 64R and vIF-2α genes.

Figure 8. FACS analysis of early and late-stage apoptosis in WT-, Δ18K-, ΔvIF-2α-, Δ52L- or Δ64R-FV3-infected A6 cells.

A6 cells were infected for 8 and 24 hrs with WT-, Δ18K-, ΔvIF-2α-, Δ52L-, or Δ64R-FV3 (20 MOI), harvested, stained with Annexin V and propidium iodine and examined by FACS for early and late stage apoptosis, respectively. Early apoptosis is regarded as annexin V positive, PI negative and late apoptosis as annexin V positive, PI positive. (A) Representative plots of early (annexin V; X-axis) and late (PI; Y-axis) stage apoptosis of cells infected for 24 hrs. (B) Early (Annexin V⁺/PI^neg) stage apoptosis detected 8 and 24 hours post infection. Results in B are means ± SEM from 3 independent experiments. Statistical significance (P<0.05) from wild type (WT) virus infected control is indicated by (*) as determined by One-Way Analysis of Variance (ANOVA).

DISCUSSION

Because of the increasing role of ranaviruses in amphibian declines and disease outbreaks among commercially important amphibian and fish species, it is imperative to gain insight into the determinants of virulence encoded by these emerging pathogens. This study represents our current efforts at improving and utilizing a recombination-based knockout methodology to generate FV3 mutants defective for putative virulence genes. As shown here and elsewhere (Chen et al., 2011), all four FV3 KO mutants were severely compromised in their ability to infect and trigger fatal disease in Xenopus tadpoles, while our in vitro data indicate that some of the targeted genes have important and possibly overlapping roles in viral replication and immune evasion.

Although treatment of A6 cells with Xenopus IFN inhibited the replication of WT and the tested KO mutants, replication of Δ18K-FV3 was significantly less affected than that of either the Δ64R- or the ΔvIF-2-FV3. This suggests that both vCARD and vIF-2α gene products interfere with some aspect of the interferon response and contribute to subverting the antiviral IFN response within infected cells. Compared to WT FV3, deletion of vCARD did not lead to an increase in type I IFN synthesis in A6 cells (Fig. 6), but resulted in higher levels of apoptosis (Fig. 8). These results suggest that vCARD may play a critical role in blocking apoptosis in ranavirus-infected cells and are consistent with a recent report showing that over-expression of a grouper iridovirus vCARD homolog inhibited apoptosis induced by mitochondrial and death receptor signaling (Chen et al., 2015).

Since the FV3 vIF-2α gene is truncated and lacks the N-terminal PKR-binding and central helicase domains (Chen et al., 2011) that are thought to mediate the protective effects seen in ATV (Jancovich and Jacobs, 2011) and RCV-Z, (Rothenburg et al., 2011), it is surprising that its knockout has such a marked effect on viral replication and pathogenesis in vivo. Intriguingly, ΔvIF-2α-FV3 maintains high levels of viral protein synthesis and effectively replicates within BHK-21 and FHM cells. Moreover, ΔvIF-2α-FV3 exhibited severely compromised replication in tadpoles, resulting in significantly lower host mortality (Chen et al., 2011), and, as described here, displayed increased sensitivity to IFN treatment. Furthermore, infection with FV3-ΔvIF-2α resulted in markedly increased levels of apoptosis within infected cells, suggesting that the FV3 vIF-2α gene product also counteracts programmed cell death induced by viral infection, possibly through the IFN response. Perhaps the truncated FV3 vIF-2α has adopted unique interactions with PKR, and/or may function by interfering with a distinct and yet undefined cellular antiviral mechanism. It will be interesting and important to learn the precise roles of the FV3 vIF-2α gene, which may represent an indispensable viral immune evasion strategy. Generation of an FV3 knock-in mutant with a full length vIF-2 alpha may provide useful insights.

Although Δ18K-FV3 was more resistant to rXlIFN inhibition than the three other KO mutants, 18K deletion resulted in substantially increased apoptosis. This is consistent with the idea that 18K may regulate timely FV3 gene expression and release. Although the results of infection with the Δ18K KO mutant, shown here, and 18K knockdown examined earlier (Sample et al., 2007), are similar, the role of this gene remains unknown. Possibly, defective viral assembly may have led to a greater intracellular PKR activation/signaling and thus triggered the observed increases in cellular apoptosis as well as the increase of IFN-I response. It will be interesting to delineate the precise roles of 18K in controlling FV3 immune evasion and virulence.

Differences in the replication of the various FV3 KO mutants in host and non-host cell lines are striking and reminiscent of effects seen with Singapore grouper iridovirus (SGIV). In those studies, differences were observed in the ability of SGIV to induce apoptosis in different cell lines (Huang et al., 2013; Yan et al., 2013). Likewise cell line-related differences were noted in the in vitro replicative ability of vaccinia mutants defective in the expression of E3L or K3L. Although both genes encode proteins whose ultimate effect is to maintain protein synthesis in virus infected cells, they perform different functions. E3L binds dsRNA and prevents activation of PKR; K3L interacts with PKR and prevents phosphorylation of eIF-2α (Langland and Jacobs, 2002). These observations suggest that assessment of viral gene function can be influenced by the cell lines chosen for in vitro analysis.

Ranaviruses represent a growing and imminent threat to poikilothermic populations. It is our belief that by performing studies akin to this one, we can begin to devise preventative and therapeutic measures for counteracting the devastation caused by these pathogens to aquaculturally-maintained species and ecological communities. Perhaps KO mutants lacking one or more critical virulence genes may serve as vaccines capable of protecting ecologically and commercially important ectothermic vertebrates from infection and or disease. If so, this work will have important practical, as well as theoretical, value.

MATERIALS AND METHODS

Cells and viruses

Fathead minnow cells (FHM, American Type Culture Collection, ATCC No.CCL-42) and baby hamster kidney cells (BHK-21, ATCC No. CCL-10) were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (FBS; Invitrogen), penicillin (100 U/mL) and streptomycin (100 μg/mL) with 5% CO₂ at 30°C and 37°C, respectively. FV3 (Granoff et al., 1965); ATCCVR-569) was propagated on confluent monolayers of FHM cells grown in 75 cm² flasks. To propagate virus, FHM cells were infected at a multiplicity of infection (MOI) of 0.01 PFU/cell and harvested 5 days later when cytopathic effect was extensive. Virus stocks, partially purified after cell lysis by centrifugation through a 30% sucrose cushion, were resuspended in Dulbecco’s PBS (DPBS; GIBCO, Carlsbad, CA), and stored at −80°C. Virus titers were determined by plaque assay. Generation of KO mutants was carried out in BHK-21 cells in DMEM supplemented with 2.5% FBS at 30°C.

rXlIFN

Following transfection of Sf9 cells with an expression vector encoding X. laevis Type I IFN, recombinant IFN (rXL-IFN) was isolated from the tissue culture medium as described previously (Grayfer et al., 2014). As a control, medium from cells transfected with an empty vector were processed in a similar fashion.

Plaque assays

Virus was serially diluted ten-fold in DMEM supplemented with 2.5% FBS. 500 μl of each dilution was plated in duplicate onto confluent monolayers of FHM or BHK-21 cells in 6-well plates, and incubated at 30°C for 1 h with redistribution of virus inocula every 20 min. Inocula were removed by aspiration, and 2 ml of overlay media (DMEM supplemented with 2.5% FBS and 1% methyl-cellulose (Sigma) was added. Cells were incubated 7 days at 30°C in 5% CO₂. Overlay media was aspirated and the cells stained for 10 min with 1% crystal violet in 20% ethanol.

Construction of recombination vectors

To generate FV3 KO by homologous recombination, regions encompassing about 500 bp flanking the targeted genes (FV3 genomic location for 52L: 57,481–58,548 and 64R: 75,529–75,816) were amplified by PCR using FV3 genomic DNA as template and subsequently cloned into right (restriction sites XhoI and ClaI) and left (restriction sites SacI and SpeI) sides of a cassette containing the puromycin-resistance gene fused with the coding sequence of EGFP under the control of the promoter for the FV3 immediate-early (IE) gene 18K (18Kprom-Puro-EGFP cassette; (Chen et al., 2011)). Primer to amplify flanking genomic regions of 52L were for the right side: forward Xho (5′-TCAGATCTCGAGccgacagacagggagcaggccgc-3′) and reverse Cla (5′-AAGCTTATCGATaagaaaaacctacttgagggtatagattc-3′); for the left side: forward Spe1 (5′-TCTAGAACTAGTcttgtatctggtttgtgatatggaatgttgc-3′) and reverse Sac (5′-GATCTCGAGCTCgcatctgcgtaaagacgtcgccc-3′). For 64R primers were for the right side: forward Sal1 (5′-CTCGAGGTCGACcctggccgttaagatgcccccg-3′) and reverse Cla (5′-AAGCTTATCGATggcttattgtgtaaagctgggtgc-3′); for the left side: forward Spe1 (5′-TCTAGAACTAGTatagagattagggacttgtagatagataaaatacc-3′) and reverse Sac (5′-GATCTCGAGCTCccggtcccagcggggcgcc-3′).

Generation of KO mutants

Confluent BHK-21 cells in six-well plates were infected for 2 h at 30°C with FV3 at an MOI of 5. Recombination vectors were transfected into FV3-infected cells using Lipofectamine 2000, according to the manufacturer’s instructions (Invitrogen). Two days post infection (p.i.), virus was collected and used to re-infect BHK-21 cells. To select for puromycin-resistant viruses, cell cultures were treated beginning at 4 h p.i. with 50 μg/mL puromycin. Three days later viruses were collected and used to re-infect BHK-21 cells without puromycin selection. The resultant recombinant viruses were plaque purified ~ 4–5x until all plaques were positive for GFP. Additional information concerning the two KO mutants generated herein, along with two previously isolated mutants, is found in Table 1.

PCR confirmation of gene KO

Confluent BHK-21 cells were infected at an MOI of 0.01 and incubated at 30°C until cytopathic effect was apparent. At that time, cells were pelleted, lysed in buffer containing 50 mM Tris-HCl (pH 8), 100mM NaCl, 10 mM EDTA, 1% SDS, 100 μg/ml proteinase K (Roche), and incubated overnight at 56°C. DNA was extracted with phenol-chloroform and ethanol precipitated. PCR reactions were carried out using Iproof polymerase (Bio-Rad) with 250 ng DNA, 200 nM of primers flanking (Fig. 2A) or within (Fig. 2B) the genes of interest and the following cycling conditions: 95 °C for 10 min, then 35 cycles of 95 °C for 30 s, 56 °C for 45 s, 72 °C for 4 min followed by an extension reaction at 72 °C for 10 min. PCR products were separated by electrophoresis on 0.8% agarose gels in TAE buffer (40mM Tris–acetate, 1mM EDTA) and the products visualized by ethidium bromide staining. Primer pairs used in these studies are described in Table 2. In addition to size verification, PCR products from the recombinant viruses were purified by gel extraction (QIAquick, Qiagen) and sequenced to confirm loss of the targeted gene.

Table 2.

List of primer sequences

PRIMER	SEQUENCE (5′-3′)
EF-1α	F: CCTGAATCACCCAGGCCAGATTGGTG R: GAGGGTAGTGTGAGAAGCTCTCCACG
Flanking 64R	F: TGGAGTCTTACAGCGACGAGGAG R: GGAGATGTGGGCAGGACAGGAG
64R	F: AGACCCTCCAGAACCTGATTGAC R: TCTCCCTTCTTGGTGACGCTGT
Flanking 52L	F: TCGTGCCTGGCAGTCTTGCTGT R: AACTTTACGCCCGTCACCCTCA
52L	F: CACTGTCCACGATCACGTCTACC R: AACTTTACGCCCGTCACCCTCA
DNA Pol II	F: ACGAGCCCGACGAAGACTACA R: TGGTGGTCCTCAGCATCC T
GAPDH	F: GACATCAAGGCCGCCATTAAGACT R: AGATGGAGGAGTGAGTGTCACCAT
GFP	F: ACGGCCACAAGTTCAGCGTG R: GTCCATGCCGAGAGTGATCC
IFN	F: GCTGCTCCTGCTCAGTCTCA R: GAAAGCCTTCAGGATCTGTGTGT
IFNγ	F: TCCCTCCCAACAGCTCATG R: CCGACACACTGAGCGGAAA

F: Forward; R: Reverse

Single-step and multiple-step growth curves

Six well plates containing 75% confluent monolayers of BHK-21 or FHM cells were infected with either WT- or recombinant viruses at an MOI of 5 for analysis of single-step growth kinetics and at an MOI of 0.01 for multiple-step analysis. After 1 hr, the virus was removed and the monolayers washed three times with PBS. Samples were collected at various times p.i., and virus yields determined by plaque assay on BHK-21 or FHM cells as described above.

Viral infection of tadpoles

To assess survival, 12 stage 50 (i.e., 2-weeks post-fertilization) immunocompetent tadpoles (Nieuwkoop and Faber, 1994) were infected by immersion in a total volume of 5 ml containing 5×10⁶ PFU for 1 hr. Subsequently, tadpoles were transferred into 1-liter containers, and cumulative mortality was monitored over a 30-day period. To monitor viral replication in vivo, slightly older tadpoles at stage 54–55 (3–4 weeks post-fertilization) were anesthetized with 0.1% tricaine menthanesulfonate (TMS) and infected by intraperitoneal (i.p.) injection with 1 × 10⁴ plaque-forming units PFU of WT or KO mutant viruses in 5 to 10 μl PBS (Gantress et al., 2003).

A6 cell maintenance, rXlIFN treatment and FV3 infection

A6 cell cultures were maintained in the medium described above and passaged weekly. For all A6 cellular gene expression and FV3 replication studies, 5×10⁵ A6 cells were seeded into individual wells of 48 well plates and infected at an MOI of 0.6 with WT-, Δ18K-, Δ64R- or ΔvIF-2α-FV3. For gene expression, cells were harvested and processed at 6, 24 and 48 hr p.i. To monitor the impact of type I IFN on virus replication, A6 cells were grown as above, pretreated with 100 ng/mL of rXlIFN, infected with WT and recombinant FV3. At 24 hr p.i., cells were harvested, lysed by 3 freeze and thaw cycles and the lysate assayed by plaque assay.

Flow cytometry-based apoptosis assay

A6 cells (5 ×10⁴) were either mock-treated with amphibian PBS (APBS; negative control) treated with 50 mg/mL of cycloheximide (positive control; Sigma, USA), or infected at a MOI 20 with WT or KO mutants (Δ18K, Δ64R, ΔvIF-2α and Δ52L) for 8 and 24 hrs. Cells were then washed in ice-cold APBS containing 1% BSA, 0.01% NaN₃ and 2.5mM CaCl₂ and incubated for 15 minutes with APC-conjugated Annexin V and propidium iodide (PI), respectively (BD Pharmigen, USA) according to the manufacturer’s protocol. For each sample, 10,000 events were collected using a BD Accuri C6 instrument (BD Pharmigen, USA) and the data were analyzed with the FlowJo software (Tree Star Inc.).

Quantitative PCR

Relative gene expression analysis was performed by qPCR using the delta^delta CT method with the ABI 7300 Real-Time PCR System and PerfeCta SYBR Green FastMix ROX (Quanta). Gene expression was examined relative to the GAPDH endogenous control and standardized against respective uninfected control gene expression. To account for the differences in WT and recombinant FV3 infection loads, the expression data was then normalized against the respective FV3 loads and expressed as a ratio of gene expression observed at 6 hpi for WT FV3 infections. FV3 viral loads were assessed by absolute qPCR on isolated DNA using a standard curve. Briefly, an FV3 vDNA Pol II PCR fragment was cloned into the pGEM-T vector (Promega), amplified in bacteria, quantified and serially diluted to yield 10¹⁰–10¹ vDNA Pol II fragment-containing plasmid copies. These dilutions were used to generate a standard curve in subsequent absolute qPCR assays of FV3 DNA quantities. All experiments were performed using the ABI 7300 real-time PCR system and PerfeCTa® SYBR Green FastMix, ROX (Quanta). ABI sequence detection system software (SDS) was employed for all expression analysis. Primers were validated prior to use and are shown in Table 2.

Statistics

Results from PCR data were evaluated by One-Way Analysis of Variance (ANOVA) for Independent or Correlated Samples using an online database available through Vassar Stat, a website for statistical computation (http://faculty.vassar.edu/lowry//anova1u.html). The survival curves were tested by the Kaplan-Meier method, and the growth curves with the Kruskal-Wallis test, using Graphpad Prism software 5.0.

Research Highlights.

• Generation of new FV3 knockout mutants identify two virulence genes vCARD (64R) and vβ-HSD (52L)

• Δ64R- and Δ52L-FV3 induce less mortality and replicate less in Xenopus tadpoles than WT-FV3

• Δ64R- and ΔvIF-2α–FV3 are more sensitive to interferon than WT and Δ18-FV3

• Δ64R-, Δ18K- and ΔvIF-2α- but not the Δ52L-FV3 trigger more apoptosis than WT FV3

• vCARD (64R) and vβ-HSD (52L) genes contribute to viral pathogenesis

Abbreviations

ANOVA

One-Way Analysis of Variance

BHK-21

baby hamster kidney-21 cells

βHSD

ORF 52L, β-hydroxysteroid dehydrogenase

DMEM

Dulbecco’s modified Eagle’s medium

FBS

fetal bovine serum

FV3

Frog Virus 3

IE

immediate-early

i.p.

intraperitoneal injection

MOI

multiplicity of infection

PFU

plaque forming units

p.i.

post-infection

qPCR

quantitative real-time PCR

RV

Ranavirus

vCARD

ORF 64R, Caspase Activation and Recruitment Domain-containing protein

vIF-2α

ORF 26R, viral homolog of eukaryotic translation initiation factor-2 alpha

18K

ORF 82R, FV3 18 kDa immediate early protein