EVM005: An Ectromelia-Encoded Protein with Dual Roles in NF-κB Inhibition and Virulence

Poxviruses contain large dsDNA genomes encoding numerous open reading frames that manipulate cellular signalling pathways and interfere with the host immune response. The NF-κB signalling cascade is an important mediator of innate immunity and inflammation, and is tightly regulated by ubiquitination at several key points. A critical step in NF-κB activation is the ubiquitination and degradation of the inhibitor of kappaB (IκBα), by the cellular SCFβ-TRCP ubiquitin ligase complex. We show here that upon stimulation with TNFα or IL-1β, Orthopoxvirus-infected cells displayed an accumulation of phosphorylated IκBα, indicating that NF-κB activation was inhibited during poxvirus infection. Ectromelia virus is the causative agent of lethal mousepox, a natural disease that is fatal in mice. Previously, we identified a family of four ectromelia virus genes (EVM002, EVM005, EVM154 and EVM165) that contain N-terminal ankyrin repeats and C-terminal F-box domains that interact with the cellular SCF ubiquitin ligase complex. Since degradation of IκBα is catalyzed by the SCFβ-TRCP ubiquitin ligase, we investigated the role of the ectromelia virus ankyrin/F-box protein, EVM005, in the regulation of NF-κB. Expression of Flag-EVM005 inhibited both TNFα- and IL-1β-stimulated IκBα degradation and p65 nuclear translocation. Inhibition of the NF-κB pathway by EVM005 was dependent on the F-box domain, and interaction with the SCF complex. Additionally, ectromelia virus devoid of EVM005 was shown to inhibit NF-κB activation, despite lacking the EVM005 open reading frame. Finally, ectromelia virus devoid of EVM005 was attenuated in both A/NCR and C57BL/6 mouse models, indicating that EVM005 is required for virulence and immune regulation in vivo.


Introduction
The NF-kB family of transcription factors activate potent proinflammatory and anti-viral immune responses that are activated by a variety of signalling pathways [1,2]. The family consists of five members, p50, p52, p65 (RelA), RelB, and c-Rel, which function as homo-or heterodimers to activate specific genes. The bestcharacterized NF-kB dimer is the p50/p65 heterodimer, which is held inactive in the cytoplasm by the inhibitor of kB (IkBa) [1,2]. Signalling cascades initiated by both tumour necrosis factor a (TNFa) and interleukin 1b (IL-1b) trigger the activation of a set of kinases known as the IkB kinase (IKK) complex, which is composed of IKKa, IKKb and IKKc/NF-kB essential modifier (NEMO) [2]. Upon activation of the IKK complex, IKKb phosphorylates IkBa on serines 32 and 36, targeting IkBa for polyubiquitination and degradation by the 26S proteasome [1,2]. The SCF (Skp1/Cul1/F-box) ubiquitin ligase recruits phospho-IkBa through the F-box domain-containing adaptor protein, b-TRCP, resulting in the degradation of IkBa, and translocation of the p50/p65 heterodimer into the nucleus [1,2].
Regulation of NF-kB signalling is common amongst most viruses, with each virus employing a combination of specifically tailored strategies [3][4][5]. For example, human immunodeficiency virus (HIV), human T-lymphotropic virus type 1 (HTLV-1), hepatitis B virus (HBV), and Epstein-Barr virus (EBV) activate the NF-kB signalling pathway [4]. Virus activation of the NF-kB pathway could serve several roles. For instance, viruses that lack anti-apoptotic mechanisms may activate NF-kB to prolong the life of the infected cell in order to complete the viral replication cycle. In the case of EBV, constitutive activation of NF-kB leads to the up-regulation of NF-kB-regulated pro-survival proteins during latency [6]. Alternatively, HIV-1 contains NF-kB binding sites in the long terminal repeat (LTR) region of the genome that mediate HIV-1 gene expression [7]. In contrast, other viruses encode proteins that specifically inhibit NF-kB signalling [3][4][5]. For example, the V and C proteins encoded by the Paramyxoviridae associate with the STAT family of transcription factors, thus inhibiting the interferon response and NF-kB activation [8]. Moreover, African Swine Fever Virus encodes a homolog to IkBa that sequesters p65 in the cytoplasm following IkBa degradation [9]. Overall, the varieties of viral proteins that manipulate NF-kB indicate the importance of the long and varied relationship with NF-kB.
Recently, we identified a family of four ankyrin/F-box proteins encoded by ECTV, EVM002, EVM005, EVM154 (recently renamed EVM159), and EVM165 (recently renamed EVM170) that interact with the cellular SCF ubiquitin ligase complex [25]. The ECTV-encoded proteins contain N-terminal ankyrin repeats in conjunction with a C-terminal F-box domain; similar viral ankyrin/F-box proteins have been found in a wide range of poxviruses [26]. To date, no cellular F-box proteins have been found in conjunction with ankyrin repeats, suggesting that poxviruses, including ECTV, have evolved a novel set of genes to regulate the cellular SCF ubiquitin ligase. Multiple orthologs for EVM002, EVM154, and EVM165 exist within the poxvirus family; however, EVM005 has only one ortholog, CPXV-BR011, in CPXV virus strain Brighton Red, suggesting that EVM005 and CPXV-BR011 may play an important role that is specific to ECTV and CPXV. Since degradation of IkBa is catalyzed by the SCF b-TRCP ubiquitin ligase, we investigated the role of EVM005 in regulation of IkBa degradation. Here, we show that cells infected with ECTV and stimulated with TNFa or IL-1b accumulate phosphorylated IkBa, indicating that IkBa is stabilized and not degraded. Ectopic expression of Flag-EVM005 inhibited both TNFa-and IL-1b-stimulated IkBa degradation and subsequent nuclear translocation of NF-kB; however, deletion of the EVM005 F-box domain resulted in activation of NF-kB. ECTV devoid of EVM005, ECTV-D005, inhibited NF-kB activation. Finally, we demonstrated that EVM005 is a critical virulence factor, since ECTV-D005 was attenuated in both A/ NCR and C57BL/6 mice compared to wild type ECTV. These data demonstrate a novel role for poxvirus-encoded ankyrin/Fbox proteins in regulation of the SCF ubiquitin ligase and NF-kB signalling.

Ectromelia virus infection blocks IkBa degradation
The NF-kB signalling cascade activates a family of transcription factors responsible for initiating the pro-inflammatory response and antiviral innate immunity [1,2]. Recent evidence indicates that many poxviruses encode proteins that tightly regulate the activation of NF-kB through the expression of secreted and intracellular factors [5,12]. Unlike strains of VACV, ECTV lacks M2, K7, B14, A49 and A52, all of which are important inhibitors of NF-kB activation [16,18,21,23,24]. Given the absence of these inhibitors, we sought to determine if ECTV infection inhibited NF-kB activation. Since the degradation of IkBa is crucial for activation of the NF-kB pathway, we examined the kinetics of IkBa degradation during infection. HeLa cells were mockinfected, or infected with ECTV or VACV for 12 hours and treated with TNFa for up to 120 minutes. Mock-infected cells treated with TNFa showed a typical pattern of IkBa degradation kinetics (Fig. 1A). As early as 10 minutes post-TNFa treatment, mock-infected cells showed phosphorylated IkBa, as indicated by a doublet, which was subsequently degraded ( Fig. 1A and B) [27,28]. ECTV-and VACV-infected cells treated with TNFa also showed obvious phosphorylation of IkBa ( Fig. 1A and B); however, the levels of both IkBa and phospho-IkBa were sustained compared to mock-infected cells ( Fig. 1A and B). Interestingly, we did observe that lower levels of phospho-IkBa accumulated in cells infected with VACV compared to ECTV, and accumulation was delayed in comparison to ECTV-infected cells (Fig. 1A). Western blotting for I5L, a late poxviral protein, and cellular b-tubulin were used as loading controls ( Fig. 1C and D) [25]. Similar observations were also seen following treatment with IL-1b ( Fig. 1E-H), indicating that members of Orthopoxvirus genera, including ECTV, sustained levels of phospho-IkBa and inhibited the degradation of IkBa.
Since IkBa appeared to be phosphorylated but not rapidly degraded during infection with ECTV or VACV, we sought to determine if NF-kB p65 was retained within the cytoplasm. HeLa cells were mock-infected or infected with ECTV or VACV, and p65 nuclear accumulation was assayed using immunofluorescence ( Fig. 2A and D). Mock-infected cells demonstrated cytoplasmic retention of p65 in the absence of TNFa or IL-1b stimulation, as expected ( Fig. 2A panels a-c and D panels m-o). In contrast,

Author Summary
Poxviruses are large dsDNA viruses that are renowned for regulating cellular pathways and manipulating the host immune response, including the NF-kB pathway. NF-kB inhibition by poxviruses is a growing area of interest and this family of viruses has developed multiple mechanisms to manipulate the pathway. Here, we focus on regulation of the NF-kB pathway by ectromelia virus, the causative agent of mousepox. We demonstrate that ectromelia virus is a potent inhibitor of the NF-kB pathway. Previously, we identified a family of four ectromelia virus genes that contain N-terminal ankyrin repeats and a C-terminal F-box domain that interacts with the cellular SCF ubiquitin ligase. Significantly, expression of the ankyrin/F-box protein, EVM005, inhibited NF-kB, and the F-box domain was critical for NF-kB inhibition and interaction with the SCF complex. Ectromelia virus devoid of EVM005 still inhibited NF-kB, indicating that multiple gene products contribute to NF-kB inhibition. Importantly, mice infected with ectromelia virus lacking EVM005 had a robust immune response, leading to viral clearance during infection. The data present two mechanisms, one in which EVM005 inhibits NF-kB activation through manipulation of the host SCF ubiquitin ligase complex, and an additional, NF-kBindependent mechanism that drives virulence.
mock-infected cells stimulated with TNFa or IL-1b showed nuclear accumulation of p65 ( Fig. 2A panels d-f and D panels p-r). Upon infection with ECTV or VACV, p65 was retained in the cytoplasm following treatment with TNFa or IL-1b, indicating that ECTV could inhibit NF-kB despite the lack of orthologs of M2, K7, B14, A49, and A52 ( Fig. 2A panels g-l and D panels s-x). These data were confirmed by Western blotting cytoplasmic and nuclear extracts from infected HeLa cells with an antibody specific for p65 ( Fig. 2B and E). As expected, p65 was absent from the nuclear extract of mock-infected cells. In contrast, mock-infected cells treated with TNFa or IL-1b showed nuclear p65 ( Fig. 2B and E). Little p65 accumulation in the nuclear extract was observed in cells infected with ECTV or VACV and treated with TNFa or IL-1b ( Fig. 2B and E). These results were also confirmed in mouse embryonic fibroblasts (MEF) (Fig. 2C and F). Together, these data indicate that NF-kB signalling is inhibited upon infection with members of the Orthopoxvirus genus. Importantly, despite the lack M2, K7, B14, A49 and A52 in ECTV, ECTV infection clearly inhibited p65 translocation to the nucleus.

ECTV-encoded ankyrin/F-box protein, EVM005, inhibits p65 nuclear accumulation
Recently, we identified a family of four ankyrin/F-box proteins in ECTV (EVM002, EVM005, EVM154 and EVM165), which interact with the cellular SCF ubiquitin ligase (Fig. S1) [25]. The poxvirus family of ankyrin/F-box proteins differs substantially from the cellular F-box proteins. In contrast to the cellular F-box proteins, the poxviral F-box domains are found at the C-terminus in combination with N-terminal ankyrin repeats [25,26,[29][30][31][32][33][34]. With the exception of EVM005, which has only one ortholog in cowpox virus Brighton Red, CPXVBR011, multiple orthologs exist for EVM002, EVM154 and EVM165. Since the SCF b-TRCP ubiquitin ligase plays an essential role degrading phospho-IkBa, we sought to determine if EVM005 could inhibit IkBa degradation and NF-kB activation during ECTV infection [27,28].
We first tested the ability of EVM005 to inhibit the nuclear accumulation of NF-kB p65. HeLa cells were mock-transfected or transfected with full length Flag-EVM005. At 12 hours posttransfection, cells were stimulated with TNFa or IL-1b for 20 minutes and nuclear accumulation of p65 was detected using immunofluorescence (Fig. 3). As expected, unstimulated HeLa cells demonstrated cytoplasmic staining of p65 ( Fig. 3A panels a-c and B panels m-o), and strong nuclear accumulation of p65 was seen following TNFa and IL-1b stimulation ( Fig. 3A panels d-f and B panels p-r). In contrast, cells expressing Flag-EVM005 that were stimulated with TNFa or IL-1b strongly inhibited p65 nuclear accumulation ( Fig. 3A panels g-i and B panels s-u). Given the importance of the F-box domain in associating with the SCF ubiquitin ligase, we wanted to determine whether this domain also contributed to the inhibition of NF-kB activation. To do this, we utilized an EVM005 mutant, Flag-EVM005(1-593), which lacks the C-terminal F-box-like domain and fails to interact with the SCF ubiquitin ligase [25]. Interestingly, cells expressing Flag-EVM005(1-593) displayed strong nuclear staining of p65 following TNFa or IL-1b stimulation ( Fig. 3A panels j-l and B panels v-x). Nuclear translocation of p65 was quantified by counting cells from three independent experiments (Fig. 3C). These data indicate that expression of Flag-EVM005 inhibited both TNFa-and IL-1binduced nuclear accumulation of p65, and that inhibition of p65 nuclear accumulation required a functional F-box domain.

EVM005 inhibits TNFa-and IL-1b-induced IkBa degradation
Since transient expression of EVM005 inhibited p65 translocation (Fig. 3), we sought to determine if EVM005 stabilized IkBa. HeLa cells were transfected with Flag-EVM005 and IkBa was ; however, upon treatment with TNFa, IkBa was degraded, suggesting that the F-box domain was necessary for EVM005 to inhibit IkBa degradation ( Fig. 4A panel u-x). To further confirm these data, HeLa cells were mock-transfected or transfected with Flag-EVM005 in the absence or presence of TNFa and immunoblotted for anti-IkBa, anti-Flag to detect EVM005, and anti-b-tubulin as a loading control. Compared to unstimulated cells, which showed a significant level of IkBa, cells treated with TNFa showed decreased levels of IkBa (Fig. 4B). Expression of EVM005 led to the stabilization of IkBa (Fig. 4B). Finally, we tested the ability of HeLa cells were mock-infected or infected with ECTV or VACV at a MOI of 5 for 12 hours followed by stimulation with 10 ng/ml TNFa (A) or 10 ng/ml IL-1b (D) for 20 minutes. Cells were stained with anti-p65 and DAPI to visualize nuclei and viral factories. (B and E) HeLa cells were mock-infected or infected with ECTV or VACV for 12 hours followed by stimulation with 10 ng/ml TNFa (B) or 10 ng/ml IL-1b (E) for 20 minutes. Nuclear and cytoplasmic extracts were collected and western blotted with anti-p65, antib-tubulin, and anti-PARP. (C and F) MEF cells were mock-infected or infected with ECTV or VACV for 12 hours followed by stimulation with 10 ng/ml TNFa (C) or 10 ng/ml IL-1b (F) for 20 minutes. Nuclear and cytoplasmic extracts were collected and western blotted with anti-p65, anti-b-tubulin, and anti-PARP. doi:10.1371/journal.ppat.1004326.g002 EVM005 Inhibits NF-kB and Contributes to Virulence PLOS Pathogens | www.plospathogens.org EVM005 to inhibit IkBa degradation by flow cytometry. HeLa cells were mock-transfected, or transfected with Flag-EVM005 or Flag-EVM005(1-593) [25]. Twenty-four hours post-transfection, cells were stimulated with TNFa or IL-1b, and fixed and stained with anti-Flag and anti-IkBa, to detect EVM005 and IkBa, respectively. Flag-positive cells were gated for analysis ( Fig. 4C panels b and e). Untransfected cells demonstrated levels of IkBa that were significantly decreased following TNFa or IL-1b stimulation, as indicated by a leftward shift on the histogram (shown in green) ( Fig. 4C panels a and d). Pre-treatment of HeLa cells with the proteasome inhibitor MG132, and subsequent TNFa or IL-1b stimulation (shown in blue) inhibited the degradation of IkBa, as expected ( Fig. 4C panel a and d) [27]. HeLa cells expressing Flag-EVM005 and stimulated with TNFa or IL-1b inhibited IkBa degradation ( Fig. 4C panels b and e); however, expression of Flag-EVM005(1-593) was unable to stabilize IkBa, resulting in degradation of IkBa (Fig. 4C panels c and f). This experiment was repeated in triplicate and these data were quantified by measuring the percentage of cells that underwent IkBa degradation (Fig. 4D). These data indicated that Flag-EVM005 strongly inhibits TNFa-and IL-1b-induced IkBa degradation, while the F-box deletion mutant failed to inhibit IkBa degradation (Fig. 4). Together, these data show that EVM005 expression blocked IkBa degradation in an F-boxdependent manner.

ECTV-D005 is still a potent inhibitor of NF-kB activation
To further examine the role of EVM005 during infection, we generated an EVM005 deletion virus, ECTV-D005. In the past, deletion of open reading frames from poxvirus genomes has been performed by inserting drug selection or fluorescent markers into the gene of interest. Instead, we used the novel Selectable and Excisable Marker system that utilizes the Cre recombinase to delete the selection markers resulting a clean deletion of the targeted open reading frame [35,36]. A cassette containing yellow fluorescent protein fused to guanine phosphoribosyl transferase (yfp-gpt) was inserted into the EVM005 locus. To generate a marker-free EVM005 deletion virus, ECTV-D005, we removed the yfp-gpt marker by infecting U20S cells that stably expressed a cytoplasmic mutant of the Cre recombinase ( Fig. S2) [36]. Additionally, two revertant viruses were generated by replacing the yfp-gpt cassette with either wild type EVM005 or EVM005(1-593), a mutant lacking the F-box domain. PCR amplification of the EVM005 locus from viral genomes was used to screen for the purity of our viral products (Fig. S2). Using a multi-step growth curve, no growth defects were detected upon infection with To determine if ECTV devoid of EVM005 could still inhibit nuclear accumulation of p65 following stimulation with TNFa, HeLa cells were mock-infected, or infected with ECTV or ECTV-D005. Immunofluorescence revealed that infection with ECTV-D005 inhibited NF-kB p65 nuclear accumulation ( Fig. 5A panels j-l). This was further supported by nuclear and cytoplasmic extracts in both HeLa (Fig. 5B) and MEF cells (Fig. 5C). We also examined the effect of ECTV and ECTV-D005 infection on the production of NF-kB-regulated transcripts. HeLa cells were mock-infected, or infected with ECTV or ECTV-D005 at a MOI of 5. At 12 hours post-infection, cells were stimulated with TNFa, and RNA samples were collected at 0, 2, 4, and 6 hours post-TNFa treatment. Samples were screened for the relative levels of RNA transcripts corresponding to TNFa, IL-1b, and IL-6; genes known to be upregulated by NF-kB [37]. All samples are presented as relative units compared to GAPDH as well as the unstimulated or 0 hour time point within each sample. Mock-infected cells displayed an increase in TNFa, IL-1b, and IL-6 transcript levels at 2 hours post-TNFa stimulation, as expected (Fig. 6). Transcript levels decreased at 4 and 6 hours post-stimulation, due to the up- regulation of NF-kB inhibitors such as IkBa (Fig. 1A) [38]. In contrast, infection with ECTV and ECTV-D005 prevented transcriptional upregulation of TNFa, IL-1b, and IL-6 ( Fig. 6). We additionally screened our 0 hour time points to compare basal levels of NF-kB transcripts between samples (Fig. S5). This analysis, demonstrated that basal levels of TNFa, IL-1b and IL-6 were higher in infected cells compared to mock infected cells, however, we were still unable to detect any significant changes between cells infected with ECTV versus cells infected with ECTV-D005 (Fig. S5). These data correlated with our previous data indicating that infection with either ECTV or ECTV-D005 inhibited the nuclear accumulation of NF-kB p65.
Finally, we looked upstream at IkBa levels. HeLa cells were infected with ECTV, ECTV-D005, or ECTV-005-rev. At 12 hours post-infection, cells were stimulated with TNFa, fixed and stained with anti-IkBa or anti-I3L, an early poxvirus protein that is indicative of infection, and analyzed by flow cytometry. Unstimulated cells (shown in black) demonstrated high levels of IkBa that decreased following TNFa stimulation (shown in green) ( Fig. 7A panel a). As expected, TNFa-stimulated mock-infected  EVM005 Inhibits NF-kB and Contributes to Virulence PLOS Pathogens | www.plospathogens.org cells that were pre-treated with the proteasome inhibitor MG132 maintained IkBa levels (shown in blue) ( Fig. 7A panel a). ECTVinfected cells stimulated with TNFa also indicated no change in the level of IkBa, lending further support that IkBa is not degraded in cells infected with ECTV ( Fig. 7A panel b). Additionally, TNFa-stimulated cells infected with ECTV-D005 or ECTV-005-rev also inhibited IkBa degradation (Fig. 7A panels  c and d). These data were quantified by measuring the percentage of cells with IkBa expression from three independent experiments to obtain standard errors (Fig. 7B). Overall, despite lacking EVM005, ECTV-D005 inhibits IkBa degradation.
Since EVM005 is one of four ankyrin/F-box proteins in ECTV, it is possible that deletion of more than one open reading frame may be necessary to render ECTV susceptible to TNFa-induced NF-kB activation and degradation of IkBa. Therefore, we used the Selectable and Excisable Marker system to excise four open reading frames, EVM002, EVM003, EVM004, and EVM005, from the left end of the ECTV genome ( Fig. S3 and S4) [36]. Notably, EVM002 and EVM003 are duplicated genes that are encoded on both ends of the ECTV genome. ECTV-D002-005 is depleted of both copies of EVM002, but only the left end copy of EVM003 (Fig. S4). EVM002 is an ECTV-encoded ankyrin/F-box protein that interacts with the SCF ubiquitin ligase and inhibits NF-kB activation by interacting with NF-kB1/p105, a member of the IkB family of proteins [25,31,39]. Deletion of EVM002 from ECTV led to slightly increased NF-kB levels in vivo, contributing to decreased virulence, potentially through low level paracrine stimulation of interferon and NF-kB in neighbouring cells [40]. Significantly, deletion of the EVM002 ortholog, CPXV006, from CPXV, rendered CPXV susceptible to NF-kB activation [39]. EVM003 encodes a vTNFR, but a copy of this gene is present at both ends of the genome. Thus, even though EVM003 was deleted from the left end of the genome, EVM003 is still expressed from the right end (Fig. S4). EVM004 encodes a BTB-only protein with unknown function [41,42]. We tested the ability of this virus, lacking two ankyrin/F-box proteins that inhibit NF-kB activation, to inhibit IkBa degradation. HeLa cells were mock-infected, or infected with ECTV, single deletion strains ECTV-D002, ECTV-D005, or the large deletion strain ECTV-D002-005, and analyzed for their ability to protect against TNFa-induced IkBa degradation using flow cytometry (Fig. 7C). Following stimulation with TNFa, ECTV, ECTV-D002, and ECTV-D005, inhibited IkBa degradation (Fig. 7C panels j-l). Additionally, IkBa was still protected from degradation in cells infected with ECTV-D002-005 ( Fig. 7C panel m). As before, staining with anti-I3L indicated virus infection (Fig. 7C panels n-r). These data were quantified by measuring the percentage of cells with IkBa expression from three independent experiments to obtain standard errors (Fig. 7D). These data indicate that deletion of more than two ankyrin/F-box proteins, and potentially other ECTV encoded NF-kB inhibitors, may be necessary to render ECTV susceptible to TNFa-induced NF-kB activation.
We also assessed the contribution of EVM005 to virulence in the A/NCR mouse strain [44,45]. Five female A/NCR mice were mock-infected, or infected with ECTV, ECTV-D005, ECTV-005rev, or ECTV-005(1-593)-rev via the footpad [44]. We infected sets of five mice with escalating 10-fold doses between 10 1 and 10 4 pfu per mouse and monitored daily changes in body weight, day of death and mortality (Fig. S7). Data from one challenge dose is displayed (Fig. 8C and D). Similar to the data observed in C57BL/6 mice ( Fig. 8A and B), ECTV-D005 was attenuated in A/NCR mice compared to wild type ECTV, ECTV-005-rev, and ECTV-005(1-593)-rev ( Fig. 8C and D). The data demonstrated that mice infected with ECTV, ECTV-005-rev, and ECTV-005(1-593)-rev succumbed to infection by day 7 post-infection. Alternatively, two of five mice infected with ECTV-D005 survived through day 21 ( Fig. 8C and D). Together, the results suggest that EVM005 is a critical virulence factor for infection of two mouse strains by two different routes of infection. Notably, mice infected with ECTV-005(1-593)-rev displayed similar mortality and weight loss profiles to mice infected with wild type ECTV and ECTV-005-rev, suggesting that although the F-box domain was necessary for inhibition of the NF-kB pathway in vitro, the ankyrin domains alone are sufficient for virulence.
Mice infected with ECTV-D005 demonstrate decreased viral spread and increased immune cell activation Though EVM005 was a potent inhibitor of NF-kB activation in tissue culture, deletion of EVM005 did not abrogate the ability of ECTV to prevent activation of NF-kB. Since tissue culture assays lack many components of the immune response, we wanted to explore the contribution of EVM005 to immune inhibition and virulence in vivo. To monitor virus spread and activation of the immune response, C57BL/6 mice were infected via intranasal inoculation with ECTV or ECTV-D005, and sacrificed at days 3, 4, and 7 post-infection (Fig. 9). Tissue from the spleen, liver, lungs, and kidneys was harvested, and the amount of virus present was determined by plaque assay (Fig. 9A and B). At 4 days postinfection, ECTV and ECTV-D005 showed no significant difference in growth rate in all organs tested (Fig. 9A); however, at 7  days post-infection, ECTV had grown to significantly higher levels than ECTV-D005 in lung, kidney and liver tissues (Fig. 9B). Notably, the decrease in viral spread correlates well with the decreased mortality previously described (Fig. 8A).
To measure the immune response, whole blood and splenocytes were harvested on days 3 and 7 post-infection and immune cell populations were quantified using flow cytometry (Fig. 9C-F). In mice infected with ECTV-D005, there was a significant increase in circulating and splenic NK cells at day 7 compared to ECTVinfected mice (Fig. 9C and E). Additionally, we observed a significant increase in circulating virus-specific CD8+ T-cells at day 7 post-infection in mice infected with ECTV-D005 compared to those infected with ECTV (Fig. 9F). Notably, we did not observe an increase in virus-specific CD8+ T-cells in the spleen on day 7 (Fig. 9D). The data suggest that the virus-specific CD8+ Tcells are being activated and expanded in non-splenic tissues before entering circulation. Finally, we assayed for transcriptional upregulation of NF-kB-regulated genes in liver and spleen on day 7 post-infection ( Fig. 9G and H). Transcriptional upregulation of TNFa, IL-1b, and IL-6 was determined by harvesting RNA from tissue samples and subjecting it to qRT-PCR. Mice infected with ECTV-D005 did not demonstrate an increase in NF-kB-regulated transcripts compared to ECTV-infected mice. These observations support a role for EVM005 in regulating virulence that is independent of its ability to inhibit NF-kB activation. Together these data indicate that mice infected with ECTV-D005 displayed boosted immune cell repertoires, increased viral clearance, and decreased mortality compared to mice infected with wild type ECTV.

Discussion
The NF-kB family of transcription factors regulate a variety of genes involved in inflammation and innate immunity [1]. Not surprisingly, viruses have evolved multiple mechanisms to regulate NF-kB [3,46], and a growing number of poxviral NF-kB inhibitors can be added to this list [5]. Previously, we identified four ankyrin/ F-box proteins in ECTV that interact with the SCF ubiquitin ligase via C-terminal F-box domains; potentially recruiting a unique set of proteins to the SCF ubiquitin ligase [25]. The NF-kB signalling pathway is dependent on the SCF ubiquitin ligase for ubiquitination and degradation of the inhibitory protein, IkBa [47]. Here we demonstrate that IkBa is phosphorylated but not degraded during ECTV infection, suggesting that signalling is inhibited at the point of IkBa ubiquitination, an event mediated by the SCF ubiquitin ligase (Fig. 1). Additionally, we demonstrate that the ECTV-encoded ankyrin/F-box protein, EVM005, inhibits p65 nuclear accumulation and IkBa degradation in a process that requires its C-terminal F-box domain ( Fig. 3 and 4). From this, we conclude that EVM005 is an inhibitor of NF-kB activation through manipulation of the SCF ubiquitin ligase. ECTV lacking the EVM005 open reading frame, ECTV-D005, was created and tested for its ability to inhibit NF-kB activation (Fig. 5-7). Even though EVM005 was deleted from the genome, ECTV-D005 still inhibited p65 nuclear accumulation (Fig. 5), the production of NF-kB-regulated transcripts (Fig. 6), and degradation of IkBa in tissue culture (Fig. 7). Significantly, ECTV lacking EVM005 was attenuated in both A/NCR and C57BL/6 mouse strains, indicating an additional NF-kB-independent mechanism for EVM0005 (Fig. 8). Interestingly, ECTV expressing a mutant of EVM005 lacking the F-box domain was still virulent, demonstrating that the ankyrin domains alone were sufficient for virulence (Fig. 8). Mice infected with ECTV devoid of EVM005 were able to mount a stronger immune response, consisting of higher numbers of NK cells and virus-specific CD8+ T-cells (Fig. 9). A strong immune response is most likely responsible for virus clearance and decreased mortality of mice, and the observed decrease in virus spread to the liver, lungs, spleen, and kidneys ( Fig. 9).
EVM005 is one of many open reading frames encoded by ECTV that inhibits NF-kB activation. Given the plethora of poxvirus-encoded inhibitors of NF-kB, the deletion of multiple open reading frames is likely required to render ECTV susceptible to NF-kB activation. In an attempt to create a strain of ECTV that was unable to inhibit TNFa-induced NF-kB activation, we deleted four open reading frames from the left end of the ECTV genome, including EVM002, EVM003, EVM004, and EVM005 [36]. Large deletion strains of VACV, such as VACV811, and modified vaccinia virus Ankara (MVA), have been tremendous tools for the characterization of novel poxvirus-host interactions [48,49]. Although VACV811 is missing 55 open reading frames, this virus is capable of inhibiting NF-kB activation [50]. MVA is an attenuated strain of VACV that has been passaged over 500 times in chicken embryonic fibroblasts and has acquired numerous gene deletions, truncations, and point mutations [49]. MVA is the only large deletion virus that has been rendered susceptible to TNFa induced NF-kB activation [51]. Our large deletion strain of ECTV, ECTV-D002-005, was able to inhibit TNFa-induced IkBa degradation ( Fig. 7C and D). EVM002 is an ankyrin/F-box protein that we have previously shown to interact with the SCF ubiquitin ligase and inhibit p65 nuclear accumulation [25,31]. These data suggest that deletion of more than two ankyrin/F-box proteins, and potentially other ECTV-encoded inhibitors of NF-kB activation, would be required to render ECTV susceptible to NF-kB activation. Creation of an ECTV strain unable to inhibit NF-kB activation would allow us to investigate how ECTV infection triggers NF-kB activation, since little is known about how poxviruses activate this pathway.
Our data demonstrate an accumulation of phospho-IkBa in ECTV-infected cells that is linked to regulation of the cellular SCF ubiquitin ligase by poxviral ankyrin/F-box proteins. The cellular F-box protein, b-TRCP, recognizes phospho-IkBa in uninfected cells and mediates ubiquitination and subsequent degradation via the 26S proteasome [52]. Though ECTV encodes four ankyrin/Fbox proteins [25,34], we tested the ability of EVM005 to inhibit NF-kB signalling, since it is unique to ECTV and CPXV [25]. Our data demonstrate that EVM005 inhibited IkBa degradation, perhaps through competition with b-TRCP for available Skp1 binding sites at the SCF ubiquitin ligase. This competition would disrupt the association between Skp1 and b-TRCP, an interaction that is required for IkBa ubiquitination and degradation. This idea is consistent with our data demonstrating the requirement of the F-box domain by EVM005 in order to inhibit degradation of IkBa ( Fig. 4 and 5). In a similar fashion, HIV-encoded Vpu disrupts the association between b-TRCP and Skp1, thus inhibiting the ubiquitination and degradation of IkBa [53]. Notably, the VACV protein A49 inhibits NF-kB signalling by binding to b-TRCP in a similar fashion to Vpu [24]. This represents a fascinating example of convergent evolution, since both EVM005 and A49 serve similar functions to inhibit NF-kB signalling, but by targeting different proteins within the SCF ubiquitin ligase. Notably, an EVM005 ortholog is not encoded by VACV, and A49 is not encoded by ECTV, demonstrating the importance of regulating NF-kB through the SCF ubiquitin ligase. Of note, our data do not rule out the possibility that EVM005 recruits substrates that are involved in NF-kB activation for ubiquitination; however, we were unable to detect degradation of IkBa, NF-kB1 p50/105, or p65 in ECTV-infected cells (N. van Buuren and M. Barry unpublished data).
Regulation of NF-kB activation by poxviral ankyrin/F-box proteins has been investigated for ECTV-encoded EVM002, CPXV-encoded proteins, CP77 and CPXV006, and the variola protein, G1R [29,31,39,40]. Similar to EVM005, these proteins interact with the cellular SCF ubiquitin ligase [29,31]. In contrast to EVM005, the mechanism by which G1R inhibits NF-kB activation does not depend on the F-box domain [31,39]. Instead, G1R and its orthologs, CPXV006 and EVM002, bind to the Nterminus of p105, an inhibitory protein similar to IkBa, to prevent TNFa-induced degradation [31]. Degradation of p105 is generally mediated by the SCF b-TRCP ubiquitin ligase following TNFa stimulation, similar to IkBa [54]. Deletion of CPXV006, a G1R ortholog encoded by CPXV, rendered CPXV susceptible to NF-kB activation [39]. Additionally, ECTV lacking EVM002 demonstrated decreased virulence and slightly increased levels of NF-kB activation in vivo [40]. We demonstrated that ECTV lacking EVM002 still inhibited IkBa degradation in tissue culture, demonstrating that the ECTV ankyrin/F-box proteins act collectively to inhibit IkBa degradation ( Fig. 7C and D). In contrast, ECTV lacking EVM005 was still a potent inhibitor of NF-kB activation in culture and in vivo (Fig. 5-7, and 9). CP77 contains a shortened F-box domain that is necessary to inhibit NF-kB activation [29]. Additionally, CP77 binds to free p65 through its ankyrin repeat domains. The model suggests that CP77 replaces the regulatory protein IkBa, following its degradation, holding the NF-kB transcription factor, p65, inactive in the cytoplasm [29]. In contrast, we were unable to detect an interaction between EVM005 with p65, p50/105 or IkBa (N. van Buuren and M. Barry, unpublished data). Similar to our data with EVM005, CP77 serves a dual role for CPXV as a host range factor in addition to its role in the inhibition of NF-kB activation [55,56]. It is clear that the ankyrin domains play a major role for most of the ankyrin/F-box proteins described to date, and this is consistent with the virulent phenotype of ECTV-005(1-593)-rev (Fig. 8). Discovery of binding partners for the ankyrin domains of EVM005 will likely provide insight to the mechanism of virulence controlled by EVM005. Together, the data suggest that the poxvirus encoded ankyrin/F-box proteins possess unique mechanisms to regulate NF-kB activation.
Although poxviral ankyrin/F-box proteins associate with Skp1 in the SCF ubiquitin ligase through their F-box domains, identification of substrates recruited for ubiquitination has eluded the field. Additionally, of the sixty-nine cellular F-box proteins encoded in the human genome, substrates have been identified for only nine [57]. The poxviral F-box proteins are suspected to function as substrate adaptor molecules for the SCF ubiquitin ligase, using their unique ankyrin domains to recruit still unidentified cellular or viral target proteins. Although binding partners, other than Skp1, have been identified for several of the poxvirus ankyrin/F-box proteins, none of these identified proteins have been characterized as bona fide substrates for ubiquitination [30,31,55]. In support of the substrate hypothesis, the ankyrin-only mutant virus, ECTV-005(1-593)-rev, was still virulent, supporting a critical role for the ankyrin domains, potentially in substrate recruitment. The data presented in this paper suggest a mechanism in which EVM005 inhibits degradation of cellular substrates, such as IkBa. This suggests an alternative mechanism for the poxviral ankyrin/F-box proteins as inhibitors of the SCF ubiquitin ligase.
Finally, we determined that EVM005 was a required virulence factor for ECTV during infection of C57BL/6 and A/NCR mice. However, ECTV-D005 was capable of inhibiting IkBa degradation, p65 nuclear accumulation and the synthesis of NF-kB regulated transcripts (Fig. 5, 6 and 7). These data suggest that an EVM005 function independent of NF-kB inhibition is responsible for mediating virulence during ECTV infection. To this end, we demonstrated that ECTV-D005 spread in vivo was suppressed compared to ECTV and that this observation correlated with increased immune cell activation (Fig. 9). It is possible that EVM005 regulates the immune response in vivo. At this time, any regulation of the immune response appears to be independent of NF-kB activation as we were unable to detect increased transcription of TNFa, IL-1b or IL-6 in spleens or livers of infected mice on day 7 (Fig. 9). We hypothesize that EVM005 is recruiting substrates to the SCF ubiquitin ligase through its ankyrin domains. Infection of mice with ECTV-005(1-593)-rev demonstrated that expression of the ankyrin domains alone was sufficient for virulence in both the A/NCR and C57BL/6 mice. Potentially the ankyrin-only mutant is still able to bind and sequester these hypothetical SCF substrates and that sequestration alone was sufficient for virulence. If the poxvirus-encoded ankyrin/F-box proteins truly function as substrate adaptors for the cellular SCF ubiquitin ligase, the identification of substrates through proteomics approaches could lead to insight into how EVM005 aids in virulence. Additionally, as mice infected with ECTV-D005 demonstrated increased immune responses, we feel that it is therefore likely that these hypothetical target substrates function in immune cell regulation.
In conclusion, our data show that ECTV-encoded EVM005 is a unique inhibitor of NF-kB activation and also suggests the existence of an NF-kB-independent mechanism for EVM005 to contribute to virulence and inhibition of immune activation. In contrast to previously characterized poxviral inhibitors of NF-kB, EVM005 requires its C-terminal F-box domain to manipulate the cellular SCF ubiquitin ligase and inhibit IkBa degradation. Further characterization of the NF-kB-independent mechanism of virulence mediated by EVM005 as well as the identification of ubiquitinated substrate proteins remains a goal of our laboratory. 50 mg/ml of streptomycin, 200 mM glutamine, and 10 mM nonessential amino acids (Invitrogen Corporation). BGMK cells were cultured in DMEM supplemented with 10% newborn calf serum, 50 U/ml of penicillin, 50 mg/ml of streptomycin and 200 mM glutamine. Vaccinia virus strain Copenhagen (VACV), and ectromelia virus strain Moscow (ECTV) were propagated in BGMK cells and harvested as previously described [58].
To create the large deletion strain of ECTV, lacking EVM002, EVM003, EVM004 and EVM005 from the left end of the genome, we used sequential insertion and Cre-mediated excision of the yfp-gpt cassette (Fig. S3). Following Cre-mediated excision, one residual loxP site remains in place of the yfp-gpt cassette. BGMK cells were infected with ECTV-D002 at a MOI of 0.01 and transfected with 10 mg of linearized pDGloxP-EVM005KO using Lipofectamine 2000 (Invitrogen Corporation). YFP-GPT positive virus was selected as described above to create ECTV-D002/D005-YFP-GPT. To delete EVM002, EVM003, EVM004 and EVM005, U20S-Cre cells were infected with ECTV-D002/ D005-YFP-GPT at a MOI of 0.01 and white foci were selected using immunofluorescence. The resulting virus, ECTV-D002-005, lacks all sequences between the residual loxP site in the EVM002 locus and the new loxP site at the right side of the EVM005 locus introduced during recombination with pDGloxP-EVM005KO. PCR analysis of the EVM002, EVM003, EVM004 and EVM005 loci confirmed the identity and purity of this large deletion strain of ECTV (Fig. S4).

Antibodies
Mouse and rabbit anti-Flag M2 were purchased from Sigma-Aldrich, anti-poly(ADP-ribose) polymerase (PARP) was purchased by (BD Biosciences) and anti-b-tubulin was purchased from ECM Biosciences. Antibodies specific for Skp1 and I5L were previously described [25,59] Antibodies recognizing the early poxvirus protein I3L were generously donated by Dr. David Evans (University of Alberta). Antibodies recognizing IkBa and phospho-IkBa were purchased from Cell Signalling Technologies. Anti-NF-kB p65 was purchased from Santa Cruz Biotechnology. Antibodies that detected cell surface markers CD45, NK1.1, CD3, and CD8 were purchased from BD Biosciences. An APC labeled tetramer specific to the immunodominant epitope for VACV B8R/CD8 T cell expression was obtained from the NIH tetramer facility.

Cytoplasmic and nuclear extracts
HeLa cells or MEF cells were mock-infected or infected with ECTV, VACV, or ECTV-D005 at a MOI of 5 for 12 hours followed by stimulation with either 10 ng/ml TNFa (Roche) or 10 ng/ml IL-1b (PeproTech Inc) for 20 minutes. Cells were harvested and lysed in cytoplasmic extract buffer containing 10 mM HEPES, 10 mM KCl, 0.1 mM EDTA (pH 8.0), 0.1 mM EGTA (pH 8.0), 1 mM dithiolthreitol (DTT) and 0.05% NP40. Samples were centrifuged at 10006 g for five minutes to remove nuclei. Supernatants were collected and resuspended in SDS sample buffer. The nuclear pellets were washed and resuspended in nuclear extract buffer containing 20 mM HEPES, 25% glycerol, 0.4M NaCl, 1 mM EDTA (pH 8.0), 1 mM EGTA (pH 8.0), and 1 mM DTT and lysis was performed on ice for 30 minutes. Samples were centrifuged at 10006g for five minutes. Supernatants were collected as nuclear extracts and mixed with SDS sample buffer.

Immunofluorescence microscopy
HeLa cells were mock-transfected or transfected with pcDNA3-Flag-EVM005 or pcDNA3-Flag-EVM005(1-593). Alternatively, HeLa cells were mock-infected or infected with ECTV, VACV, or ECTV-D005 at a MOI of 5. At 12 hours post-infection or transfection, cells were stimulated with 10 ng/ml TNFa (Roche) or 10 ng/ml IL-1b (PeproTech Inc) for 20 minutes and fixed with 2% paraformaldehyde (Sigma-Aldrich) for 10 minutes at room temperature. Cells were permeablized with 1% NP40 and blocked with 30% goat serum (Invitrogen Corporation). Cells were stained with anti-NF-kB p65 ( Whole blood or splenocytes were harvested on days 3 or 7 post infection from C57BL/6 mice infected with ECTV or ECTV-D005. Whole blood was lysed with water at a 40:1 water to blood volume ratio for ten seconds then brought to 1X with 10X PBS. The remaining white blood cells were resuspended in PBS with 2% FBS prior to staining. Spleen tissues were disrupted with the Bullet Blender (STL Scientific) for ,2 minutes at room temperature using the lowest setting in PBS. The cell suspension was pelleted and the red blood cells were lysed with BD Pharm Lyse. The remaining white blood cells were resuspended in PBS with 2% FBS prior to staining. Cells were stained for flow cytometry using Fc block and the described antibody cocktails in PBS with 2% FBS for 20-30 minutes on ice. Cells were washed twice with PBS containing 2% FBS then fixed on ice with PBS containing 2% FBS and 1% methanol free formaldehyde. Stained cells were analyzed on a BD LSRII or BD Canto. NK cells were identified as being CD45 positive, CD3 negative and NK1.1 positive. These are defined in the literature as NK lytic cells and can only be identified in C57BL/6 mice [60]. An APC labeled tetramer specific to the immunodominant epitope for VACV B8R/CD8 T cell expression was obtained from the NIH tetramer facility [61]. Virus-specific CD8+ T-cells were identified as CD45 positive, CD8 positive, and tetramer positive.

Real time PCR
HeLa cells were mock-infected or infected with ECTV or ECTV-D005. At 12 hours post infection cells were stimulated with 10 ng/ml TNFa and RNA was harvested using Trizol according to the manufacturer's protocol (Invitrogen Corp.). RNA samples were converted to cDNA using Superscript II reverse transcriptase (Invitrogen Corp.). Transcript levels were analyzed by real time PCR using the following primers, TNFa forward, 59-GGC-GTGGAGCTGAGAGATAAC- 39  Additionally, we measured transcriptional activation of TNFa, IL-1b and IL-6 in infected liver and spleen through isolation of RNA using Trizol (Invitrogen Corporation) as per manufacturer's protocol. RNA samples were subjected to qRT-PCR as described above to quantify transcriptional upregulation with reference to GAPDH.

Infection of C57BL/6 and A/NCR mice with ECTV
To determine the role of the ECTV encoded EVM005 in virulence we infected female C57BL/6 mice. Four to six week old female C57BL/6 mice were obtained from the National Cancer Institute (Frederick, MD). Groups of five mice were infected via the intranasal route with 10-fold escalating doses of wild type ECTV, ECTV-D005, ECTV-005-rev or ECTV-005(1-593)-rev. Mice were anesthetized with 0.1 ml/10 g body weight with ketamine HCL (9 mg/ml) and xylazine (1 mg/ml) by intraperitoneal injection. Anesthetized mice were laid on their dorsal side with their bodies angled so that the anterior end was raised 45u from the surface; a plastic mouse holder was used to ensure conformity. Strains of ECTV were diluted in PBS without Ca 2+ and Mg 2+ to the required concentration and slowly loaded into each naris (5 ml/naris). Mice were subsequently left in situ for 2 to 3 minutes before being returned to their cages. Mice were monitored for body weight daily for up to 21 days. Mice that demonstrated a drop in body weight to 70% of their original mass, or signs of severe morbidity were euthanized. To determine organ titers and immune activation, mice were sacrificed at 2, 3, 4, and 7 days post-infection, and tissue from the spleen, liver, lungs, and kidney were harvested in addition to blood collected by a needle stick in the heart. Tissue was homogenized using a tissue homogenizer (Next Advance), followed by dilution in PBS (10% w/v). Viral titers were determined on BSC-1 cells using a plaque assay. To prevent avoidable suffering, mice demonstrating a drop in body weight to 70% of their original mass, or signs of severe morbidity, were euthanized.
Alternatively, we infected the susceptible A/NCR strain of mice to determine the role of EVM005 during an ECTV infection. Five to ten week old female A/NCR mice were obtained from the National Cancer Institute (Frederick, MD). Groups of five mice with similar body mass were arranged into separate cages. Mice were anesthetized with CO 2 /O 2 prior to infection. ECTV, ECTV-D005, ECTV-005-rev, and ECTV-005(1-593)-rev were diluted in PBS without Ca 2+ and Mg 2+ to the required concentration and 10 ml was used to infect mice via footpad injection. Body weight, day of death and mortality were monitored daily. Mice that demonstrated a drop in body weight to 70% of their original mass, or signs of severe morbidity were euthanized.

Ethics statement
To prevent avoidable suffering, mice demonstrating a drop in body weight to 70% of their original mass, or signs of severe morbidity, were euthanized.
Mice were anesthetized with 0.1 ml/10 g body weight with ketamine HCL (9 mg/ml) and xylazine (1 mg/ml) by intraperitoneal injection. Alternatively, mice were anesthetized with CO 2 / O 2 . Mice were euthanized by first anesthetizing them with CO 2 / O 2 , followed by cervical dislocation.
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Animal experiments were performed at Saint Louis University and approved by the Institutional Animal Care and Use Committee (#IACUC 2082). Additionally, these experiments were performed in accordance with mouse ethics outlined by the Canadian Council on Animal Care and the University of Alberta. Figure S1 ECTV encodes four ankyrin/F-box proteins. (A) ECTV-encoded proteins, EVM002, EVM005, EVM154 and EVM165 contain a series of N-terminal ankyrin repeats in conjunction with a C-terminal F-box domain. EVM005(1-593) is a mutant that lacks the C-terminal F-box domain and has lost the ability to interact with the cellular SCF ubiquitin ligase [25]. (B) Cullin-1 serves as a scaffold protein for the cellular SCF ubiquitin ligase complex. Roc1 binds the C-terminus of cullin-1 and contains E3 ligase activity that recruits activated E2 enzymes that catalyze the formation of K48-linked polyubiquitin chains on substrate proteins. Substrate adaptor proteins utilize an F-box domain to interact with Skp1, the linker protein of the SCF ubiquitin ligase complex. Adaptor proteins recruit substrates through additional protein-protein interaction motifs. (TIF) Figure S2 Construction and characterization of ECTV-D005. (A) The Selectable and Excisable Marker system was used to delete the EVM005 gene from ECTV strain Moscow. We inserted a yfp-gpt fusion cassette flanked by loxP sites, and used the Cre recombinase to remove the marker following selection in order to construct a marker-free deletion virus. (B) PCR of viral genomes was used to demonstrate insertion and deletion of our yfp-gpt cassette, and purity of our virus stocks. (C) Multi-step growth curve analysis of ECTV compared to ECTV-D005, ECTV-005-rev and ECTV-005(1-593)-rev. BGMK cells were infected at a MOI of 0.05 and growth was monitored over 72 hours. The number of plaque forming units at each time point was measured by plaque assay. The average number of pfu at each time point was averaged from three independent experiments. (TIF) Figure S3 Schematic for construction of ECTV-D002-005. The Selectable and Excisable Marker System was used to delete four genes, EVM002, EVM003, EVM004, and EVM005 from the left end of the ECTV strain Moscow genome. (A) BGMK cells were infected with wild type ECTV and transfected with linearized pDGloxP-EVM002KO and YFP-GPT positive viruses were purified. (B) Excision of the yfp-gpt cassette was performed by infecting U20S-Cre cells with ECTV-D002-YFP-GPT to create ECTV-D002. (C) The yfp-gpt cassette was then inserted into the EVM005 locus by infecting BGMK cells with ECTV-D002 and transfecting linearized pDGloxP-EVM005KO and YFP-GPT positive viruses were purified. (D) Finally, we infected U20S-Cre cells with ECTV-D002/005-YFP-GPT to remove the yfp-gpt cassette. Cre recombination removed all DNA between the loxP site in the EVM002 locus and the loxP site introduced into the EVM005 locus, consisting of EVM002, EVM003, EVM004 and EVM005, to create ECTV-D002-005. (TIF) Figure S4 PCR analysis of viral genomes to verify construction of ECTV-D002-005. BGMK cells were infected with ECTV, ECTV-D002, ECTV-D002/005-YFP-GPT, and ECTV-D002-005 for 48 hours. Viral genomes were subjected to PCR analysis of the EVM002, EVM003, EVM004, and EVM005 loci as well as verifying Cre deletion from EVM002 to EVM005. The presence of PCR products near 500 bp in length represent excision of the yfp-gpt cassette by Cre recombination. Alternatively, PCR products of ,1700 bp represent an intact yfp-gpt cassette, and PCR products larger than 2000 bp represent wild type sequences for EVM002 or EVM005. The presence of EVM004 is noted by the presence of a 822 bp PCR product. The presence of EVM003 is denoted by a 963 bp PCR product. (TIF) Figure S5 Quantitative PCR analysis of cytokine induction following infection with ECTV or ECTV-D005. HeLa cells were infected with ECTV or ECTV-D005 at a MOI of 5 for 12 hours. RNA from infected cells was harvested with Trizol and converted to cDNA by reverse transcription. The relative levels of TNFa, IL-1b and IL-6 were measured by real time PCR and normalized to GAPDH as well as uninfected cell transcript levels. (TIF) Figure S6 Dose response curve to ECTV infection in C57BL/6 mice. Groups of five female C57BL/6 mice were mockinfected or infected with ECTV, ECTV-D005, ECTV-005-rev, or ECTV-005(1-593)-rev via intranasal inoculation with 10-fold escalating doses between 10 2 and 10 4 pfu per mouse. Mice were monitored daily for body weight (A-C), day of death and mortality (D-F). (TIF) Figure S7 Dose response curve to ECTV infection in A/ NCR mice. Groups of five female A/NCR mice were mockinfected or infected with ECTV, ECTV-D005, ECTV-005-rev or ECTV-005(1-593)-rev with 10-fold escalating doses between 10 1 and 10 4 pfu per mouse via footpad injection. Mice were monitored daily for body weight (A-D), day of death and mortality (E-H). (TIF) Figure S8 EVM005 is expressed early during infection. HeLa cells were infected with ECTV at a MOI of 5 in the presence or absence of the DNA replication inhibitor AraC. RNA was extracted from cells at the indicated time points post infection with Trizol and subjected to RT-PCR analysis. Primers specific for GAPDH as well as known early (EVM004) and late (EVM058) poxvirus genes served as controls. (TIF)