DUSP11 is a member of the dual specificity phosphatase subgroup of type-I based cysteine-based protein tyrosine phosphatases (PTPs). Most members of the DUSP subgroup are active on protein substrates, but DUSP11 is atypical in that it is more active on RNA substrates [11]. Human DUSP11 (hDUSP11) (Chr2P13.1, O75319) has three isoforms, the major isoform being 330 amino acids (AA) in length. All DUSPs share a highly conserved phosphate-binding loop (P-loop) containing the consensus phosphatase sequence (HCXXXXXR; AA: 151–158 in human) [19] (Fig 1A). In addition, DUSP11 contains an R residue (R192 in hDUSP11), which has been previously suggested to be involved in the β-phosphatase activity, unique to RNA 5’-triphosphatases and lacking in mRNA capping enzymes [20]. Backed by recent insight into the immunomodulatory abilities of DUSP11, we hypothesized viruses might co-opt this activity. To this end, we set out to determine whether vertebrate virus genomes contain genes with similar features to DUSP11.

Using BLASTp analysis [21], we identified several APVs and avipox-related viruses predicted to encode proteins featuring high sequence identity (~44–60%) to human (hDUSP11) (S1 Table). Of those viruses, we initially sought to evaluate the putative viral homolog from one of the better-studied APVs, canarypox virus (CNPV), which encodes a 403 AA predicted homolog to DUSP11 (CNPV085/NP_955108.1) [44% (96/217) amino acid identity, (140/217) 64% positives (which indicates amino acids with similar chemical properties)] (Fig 1A). Canarypox viruses have been studied largely for use as a recombinant vaccine vector [22,23]. However, due to initial difficulties cloning the putative CNPV viral homolog, we pursued the closely related virus shearwaterpox virus 2 (SWPV2) as the model putative viral DUSP11 homolog. The SWPV2 genome contains a 303 AA predicted ORF (SWPV080/ARE67303.1) featuring high sequence identity to hDUSP11 [50% (94/188) amino acid identity, 67% (121/188) positives] (Fig 1A), hereafter vDUSP11. Notably, this protein (and other vDUSP11s) includes the R residue (AA: 161) corresponding to R192 of hDUSP11 (Fig 1B), suggesting a distinct role from existing mRNA capping enzymes. Using AlphaFold2 [24], we generated a structural prediction of the SWPV2 putative vDUSP11 homolog and aligned it to a partial crystal structure of hDUSP11 (PDB: 4JMJ; AA: 28–208) [25] with unsolved N-terminal and C-terminal extensions determined using AlphaFold2 (Figs 1C and S1). This structural analysis highlights that hDUSP11 and SWPV2 vDUSP11 are nearly superimposable, excluding the N- and C-terminal extensions, and that key catalytic residues (shown in green) are structurally conserved.

Using the SWPV2 vDUSP11 AA sequence as a query returns numerous host DUSP11 sequences, but not other phosphatases from related viruses, such as vaccinia VH1 (S2 Table). This reciprocal BLASTp analysis indicates that the progenitor of SWPV2 vDUSP11 was likely acquired by horizontal gene transfer (HGT) from a host copy of DUSP11. Many of the avipoxviruses that encode vDUSP11 infect hosts within the order of Passeriformes, making them a potential source of DUSP11. Further searches detected a total of 13* DUSP11 ORFs encoded by other APV/APV-related viruses, 2 of which are notably smaller (S2 and S3 Figs and S3 Table). (*Note: Our original studies identified 11 vDUSP11 homologs which are analyzed in the figures presented here. Two additional vDUSP11s were identified during database searches during the revisions and added to S3 Table). These findings show that multiple APV/APV-related viruses encode proteins that resemble DUSP11.

To further examine the acquisition of the putative vDUSP11 in an evolutionary context, we generated an inferred phylogenetic tree using AA sequences from proteins classified as atypical DUSPs from various hosts, the putative avipox viral DUSP11 sequences, poxviral VH1-like DUSPs from vaccinia virus (VACV) and corresponding avipoxvirus homologs, and the avipoxviral large capping enzyme subunit (S4 Table). The putative avipox vDUSP11s cluster with related host DUSP11s, and not with the other poxviral genes (Fig 2). These results further support the avipox vDUSP11 arose from host DUSP11 and not other viral or host DUSPs.

Based on AA sequence and predicted structural similarity, we hypothesized that these vDUSP11s possess similar catalytic activities as hDUSP11. hDUSP11 has been demonstrated to be active on 5’-triphosphate RNA, sequentially liberating both the γ- and β- phosphates, leaving a 5’-monophosphate [26]. This modification renders RNA susceptible to decay via monophosphate-dependent 5’-3’ exoribonuclease XRN1. We hypothesized that vDUSP11s are catalytically active RNA 5’-triphosphatases capable of performing similar functions as hDUSP11.

To assess this hypothesis, we applied a previously established assay to measure the ability of vDUSP11 to sensitize in vitro transcribed RNAs to decay via XRN1 [27,28]. 5’-RNA phosphatase treatment followed by XRN treatment of a hepatitis C virus (HCV) 5’-UTR RNA results in the generation of a stable quantifiable smaller fragment [13,27,28]. We generated in vitro transcribed and translated flag-tagged full-length proteins for hDUSP11, negative control hDUSP11, catalytically inactive mutant (D11-CM) (C152S), the SWPV2 vDUSP11, and the predicted SWPV2 vDUSP11 catalytically inactive mutant (SWPV-CM) (C135S). We also included an additional negative control of untagged luciferase. Protein expression was confirmed via immunoblot analysis (Fig 3A). We treated in vitro transcribed HCV 5’ UTR RNA with the in vitro transcribed/translated constructs. Following phosphatase treatment, we purified the RNA and then treated ± recombinant XRN1. We evaluated the abundance and length of RNA by using a urea-PAGE gel stained with ethidium bromide (EtBr) to visualize RNA. All conditions using treatment with XRN1 resulted in a small degree of generation of a shorter RNA fragment, suggesting low levels of susceptible monophosphate RNA present in the substrate population (Fig 3B). However, treatment with catalytically active hDUSP11 or the SWPV2 vDUSP11, but not either of the catalytic mutants (CM), resulted in a substantial reduction in the full-length HCV 5’ UTR fragment. Quantification indicated a ~ 30–50% increase in the ratio of degradation fragment to full-length RNA seen following XRN1 treatment (Fig 3B and 3C). To confirm that vDUSP11 directly provides the relevant enzymatic activity assayed here, we performed this same reaction using purified chimeric GST fusion protein containing the amino terminus of SWPV vDUSP11 encompassing the catalytic core (amino acids 1–197). This protein showed activity while the CM (C135S) negative control GST fusion protein did not (S4 Fig). These data demonstrate that vDUSP11, in the absence of any other stoichiometric host proteins, is an RNA triphosphatase. This result is what is expected for the known catalytic activity of DUSP11 to act on 5’-triphosphorylated RNA, sensitizing them to decay via XRN1. Thus, SWPV2 vDUSP11 is a catalytically active 5’-RNA triphosphatase and this activity depends on cysteine in the catalytic pocket, as previously demonstrated for hDUSP11.

We next determined if vDUSP11 shares additional functions with hDUSP11. RIG-I is a key PRR that recognizes structured or double-stranded 5’ di-/tri-phosphorylated RNAs, which, when activated, leads to the induction of antiviral signaling, including transcripts such as ISG15 and IFN-β [4–6]. We hypothesized that APVs/APV-related viruses have acquired vDUSP11 to mimic host DUSP11’s immunomodulatory properties, which alter the sensitivity of RIG-I activation and modulate RNAP III transcripts upregulated in the context of various virus infections [8,9,14].

To assay the functions of the vDUSP11s in cell culture, we transfected immunostimulatory RNAs (Fig 4A) into previously described A549 DUSP11 knockout (KO) cells [12] stably expressing various DUSP11s and measured RIG-I activity via IFN-stimulated gene (ISG) expression. Included in this analysis is a second full-length vDUSP11 from cheloniidpox virus 1 (ChePV) [44% (88/201) amino acid identity, 66% (133/201) positives to hDUSP11]. (We selected vDUSP11 from ChePV due to this virus being isolated from a non-avian host). We generated cells stably reconstituted with a control empty-vector (EV) or the following N-terminally flag-tagged proteins: 1) wild-type hDUSP11 (D11), 2) hDUSP11 catalytic-mutant (C152S, D11-CM), 3) SWPV2 vDUSP11 (SWPV2), 4) SWPV2 catalytic-mutant vDUSP11 (C135S, SWPV2-CM), 5) ChePV vDUSP11 (ChePV), 6) ChePV catalytic-mutant of ChePV (C135S, ChePV-CM), or 7) phosphatase DUSP12 (D12) (a canonical DUSP known to be active on protein substrates) as a negative control protein. Protein expression for each was confirmed via immunoblot analysis (Fig 4B).

Upon transfection of 5’-triphosphate RNA, cells expressing hDUSP11, SWPV vDUSP11, or ChePV vDUSP11, but not the empty-vector nor any catalytic mutants/negative controls, displayed a reduction in both induced ISG15 and IFN-β ISG transcript levels (Figs 4C and S5). Compared to empty vector cell lines, those containing either hDUSP11 or the SWPV and ChePV vDUSP11s showed on average a ~ 40–50-, ~ 40–90-, and ~ 30–50-percent reduction in induced ISG15 mRNA levels, respectively, and all showed a greater than a ~ 70–90-percent average reduction in induced IFNB1 mRNA levels. These results indicate that both vDUSP11s can modulate immune activation, as previously shown for hDUSP11.

The role of DUSP11 in virus infection is context-dependent, however, previous work from our lab has shown that some viruses (e.g., + ssRNA sindbis virus (SINV) and -ssRNA vesicular stomatitis virus (VSV)) benefit from DUSP11 activity during infection [14]. Both SINV and VSV have been shown to produce 5’-triphosphorylated RNA as part of their replication cycle. Additionally, host RNAs have also been implicated in activating RIG-I upon infection with some RNA viruses [9]. We hypothesized that vDUSP11 can enhance virus infection in a manner similar to hDUSP11. To investigate this, we infected A549 DUSP11 KO cells reconstituted with the various viral and human DUSP11s with M51R mutant VSV for 48 hours at an MOI of 0.01 PFU/cell. VSV M51R lacks the ability to fully block host gene expression and, as a result, stimulates a stronger interferon response [29].

Cells expressing either hDUSP11, SWPV vDUSP11, or ChePV vDUSP11, but not catalytic mutants or negative controls (D12/EV) enhanced VSV infection (Fig 5A-C). Compared to cells expressing only the empty vector, those reconstituted with hDUSP11 had, on average, a ~ 7-fold increase in the total virus they produced, consistent with previously published findings [14]. Cells expressing the SWPV vDUSP11 had, on average, a ~ 20-fold increase in viral titer, and those expressing the ChePV vDUSP11 had a ~ 3-fold increase in viral titer. These findings lend further support to evolutionarily conserved shared functionality between host and viral DUSP11s.

Given that the presence of vDUSP11 can alter the outcome of virus infection for RNA viruses as previously demonstrated for hDUSP11, we next sought to evaluate these vDUSP11 in a context closer to natural infection. A lack of a facile model system limited our ability to test these proteins during avipoxvirus infection, so we instead utilized the well-studied, related poxvirus: VACV. While VACV does not appear to have acquired a host copy of DUSP11, studies have reported interactions between the VACV protein E3 and RIG-I. E3, encoded by the gene E3L, has been reported to have many functions on the interface of pathogen-host interactions [30,31]. While a role for E3 in altering RIG-I activation has been proposed [30,32,33], the extent of this in altering immune signaling is unclear.

We first set out to determine if there were any detectible differences in infection or immune signaling in response to infection by VACV ΔE3L dependent on the presence of RIG-I. We infected previously characterized A549 RIG-I knockout (ΔRIG-I) and non-targeted (NT) control cells [34] with either a recombinant WT (WT-R) or an E3L deletion mutant of VACV (VACV ΔE3L) and quantified viral titers and mRNA levels of ISG15 and IFN-β at 8 and 16 hours post infection (hpi). Viral titers measured at 16 hpi were unaffected by the presence of RIG-I (S6A-D Fig). Infection with WT-R VACV resulted in low levels of ISG15 and IFN-β transcripts in both cell lines, with on average a 0.5 to 3-fold increase in either transcript detected in the presence of RIG-I compared to in the ΔRIG-I cells (Fig 6A and 6B). However, infection with VACV ΔE3L resulted in a 20–52-fold increase of either ISG15 or IFN-β mRNA levels detected in the presence of RIG-I compared to in the ΔRIG-I cells. These data compellingly support that E3 can function to reduce RIG-I-dependent immune activation during VACV infection.

Next, we evaluated if the presence of either hDUSP11 or vDUSP11 impacted immune signaling in response to infection by VACV ΔE3L. Overall, we observed only minor differences in virus production between cell lines at 16 hours post infection (hpi) (S6E-G Fig). We quantified ISG15 and IFN-β mRNA levels at 16 hpi from cells either mock-infected or infected with VACV ΔE3L (Fig 6C and 6D). In most cases in uninfected states, we observed that cells expressing only the empty vector (EV) showed a slight increase in expression of either ISG15 or IFN-β mRNA under mock infection conditions, consistent with an increased propensity for immune activation in the absence of DUSP11. In response to VACV ΔE3L infection, we observed that cells expressing hDUSP11 (D11) had on average a 2-fold reduction in ISG15 mRNA levels and 3.6-fold reduction in IFN-β mRNA levels compared to the EV expressing cell line. Expression of either SWPV2 vDUSP11 (SWPV) or ChePV vDUSP11 (ChePV) resulted in an on average 1.5-fold and 1.9-fold reduction in ISG15 mRNA levels, respectively, and an average 2.7-fold reduction and 1.5 in IFN-β mRNA levels, respectively (Fig 6C and 6D). These data demonstrate that expression of either hDUSP11 or vDUSP11 results in modest reduced immune signaling in response to infection by VACV ΔE3L. These data further support a model whereby vDUSP11s reduce RIG-I-dependent immune activation during avipoxvirus infection.

Previous work from our lab demonstrated that DUSP11 modulates the 5’-triphosphate status and steady-state levels of several RNAP III transcripts, including vault RNAs (vtRNAs), Y RNAs, and RMRP [12]. This activity could be advantageous to viruses as RNAP III transcripts can activate host PRRs such as RIG-I during virus infection [8,9]. Based on this, we sought to determine if the presence of vDUSP11 can impact the steady-state levels of endogenous RNAP III transcripts.

To evaluate if vDUSP11 catalytic activity can alter the abundance of endogenous RNAPIII transcripts, we harvested total RNA from A549 DUSP11 KO reconstituted cells and conducted northern blot analysis (Fig 7A). We probed for vtRNA1–1 and vtRNA2–1, using the 5’ monophosphate tRNA cysteine (tRNA-Cys) as a loading control. In the presence of either hDUSP11 or either vDUSP11, we observed ~30–70% reduced levels of vtRNA1–1 and vtRNA2–1 when compared to the empty vector negative control (Figs 7B–7D and S7). These data suggest that vDUSP11s can alter vtRNA levels comparable to hDUSP11.

RIG-I is predominantly localized to the cytoplasm [35], but RNAP III RNAs are typically distributed in both the cytoplasm and nucleus. In resting cells, hDUSP11 has been shown to predominantly reside in the nucleus [36]. However, in infected cells, DUSP11 must also be active in the cytoplasm, as both HCV and VSV transcripts are direct substrates that are restricted to the cytosol [13,14]. Poxviruses are DNA viruses, but unlike the majority of DNA viruses, they replicate in the cytoplasm. We sought to determine how vDUSP11 localization compares to hDUSP11.

We performed confocal immunofluorescence microscopy using the various hDUSP11 and vDUSP11 reconstituted A549 DUSP11 KO cell lines. As expected, resting cells expressing hDUSP11 and hDUSP11-CM displayed predominantly nuclear localization (Fig 8). In contrast, both vDUSP11s and their catalytic mutants appear to be pan-cellular with substantial abundance in the cytoplasm (Fig 8). The localization of these vDUSP11s is consistent with overlapping the site of poxviral replication and replicative intermediates.

Viral piracy of host proteins often occurs when genes have important immunological functions [2]. Due to their role at the host-pathogen interface, pirated genes are often subject to volatile evolution, which can include frequent duplications and loss [2,37]. For poxviruses in particular, genes that inhibit host defense often cluster together within the genome and often appear in regions towards the ends of the genome [38–40]. Comparing the genomic localization of these vDUSP11s can also inform the evolutionary history of vDUSP11 including whether these genes were acquired from the host multiple times or through a single event.

To assess the acquisition of vDUSP11 into the various APV/APV-related genomes, we analyzed the genomic location and adjacent genes. Using reference genome sequences in the NCBI database [41] for each APV/APV-related virus containing a vDUSP11, we analyzed genes upstream and downstream of the vDUSP11 coding region (Fig 9 and S5 Table). We found that while vDUSP11 is not encoded in the inverted terminal repeats (ITRs), in many cases, it is surrounded by other genes predicted to be involved in modulating the host response to infection (Fig 9 and S5 Table). Due to the similarity in the order of flanking genes, these data suggest that vDUSP11s are orthologous and likely acquired by a single acquisition event.

During this analysis, we noted that the vDUSP11 loci for crowpox virus and magpiepox virus 2 are substantially shorter. The coding region for the magpiepox virus 2 vDUSP11 encodes a predicted 60 AA protein retaining key residues essential for catalytic activity. However, it is unclear if this is sufficient for functionality. In the case of crowpox virus, comparing the genomic location to other closely related viruses showed a mutated start codon resulting in a protein not predicted to retain catalytic activity. The presence of these truncated forms is consistent with this locus being a site of evolutionary pressure. Thus, these evolutionary profiles are consistent with what would be expected if vDUSP11 activity is involved in modulating host defenses.

Our findings support the model of a single acquisition event of a host DUSP11 in an ancestor virus that gave rise to the modern lineage of avipox and avipox-related viruses with vDUSP11s. To decipher a potential source of acquisition, we performed reciprocal BLASTp analysis for each viral homolog to identify the most similar host DUSP11. Based on both BLASTp E-value and percent identity, we identified the most similar host DUSP11 to be from Corvus hawaiiensis (Hawaiian crow) to the penguinpox 2 vDUSP11 [53% (94/176) amino acid identity, 71% positives (126/176)] (S6 Table). Furthermore, this host is recalled for all the full-length vDUSP11, except for mudlarkpox virus for which Varanus komodoensis, (Komodo dragon) was recalled [51% (96/189) amino acid identity, 69% positives (131/189)] (S7 Table). This further supports the likelihood that the vDUSP11s were acquired from an avian species and potentially one closely related to Corvus hawaiiensis.

To further evaluate the acquisition source of APV/APV-related vDUSP11, we performed phylogenetic analysis using eight avipox vDUSP11 sequence and eighteen selected host DUSP11 AA sequences (S8 Table). Using PhyML [42] to generate an inferred phylogenetic tree, we found that the avipox vDUSP11 clusters with avian DUSP11 (S8 Fig). These data lend additional support for an avian origin for the various avipox vDUSP11. Combined, the above data demonstrate that APV/APV-related viruses have acquired a viral homolog of host RNA phosphatase DUSP11, which is enzymatically active and features characteristics of a protein involved in the modulation of the host immune response.

Viral homologs to DUSP11 were identified using BLASTp [45] with the hDUSP11 sequence as query (S1 Table). Sequences were selected based on returned BLAST parameters (total score, e-value, and percent identity). Sequence alignment was generated using FASTA AA sequences acquired from NCBI database [41] and Clustal Omega [46] was used for multisequence alignment. In Fig 1B, sequence alignment was plotted using the R package ggmsa [47]. Reverse BLASTp [45] results for various vDUSP11 are shown in S2, S6, S7, and S9 Tables.

The pcDNA3.1 puro and pLenti DUSP11-3xFLAG and DUSP11-3xFLAG-CM expression vectors were previously described [12]. The PISK-T7-HCV5’UTR vector was previously described [13]. vDUSP11 reference AA sequences were inputted into the IDT (Integrated DNA technologies) codon optimization tool to generate codon-optimized DNA sequences for expression in human cells. DNA sequences for SWPV2 vDUSP11, ChePV vDUSP11 and DUSP12 were synthesized as gBlocks (IDT) containing XhoI/XbaI restriction enzyme sites. The gBlocks were cloned into pcDNA3.1+ (puro) and plenti-EF1α (pLenti) backbones. Around-the-horn PCR [48] was utilized to add amino-terminal epitope (3xFLAG) tags to pLenti vDUSP11 and DUSP12 constructs. vDUSP11 catalytic mutant (CM) constructs were generated using overlap PCR [49] combined with restriction enzyme digest. Tagged pcDNA3.1+ (puro) 3xFLAG-tagged vDUSP11/vDUSP11-CM constructs were subcloned into pcDNA3.1 + behind a T7 promoter using restriction enzyme digest. For pGEX 4T-1 D11 CC AA: 29–205-GST, PCR was used to amplify a portion of hDUSP11 corresponding to amino acids 29–205 (Genbank accession #: NM_003584) using the pcDNA3.1-puro-3XFLAG-DUSP11 as template and this amplicon was cloned into pGEX-4T1 between the Not-I and BamHI restriction sites. For pGEX 4T-1 SWPV2 vD11 AA: 1–197-GST, PCR was used to amplify a portion of the vDUSP11 corresponding to amino acids 1–197 (Genbank accession #: ARE67303.1) using the plenti-EF1α SWPV2 vDUSP11 as template and cloned into pGEX-4T1 between the Not-I and BamHI restriction sites. For pGEX 4T-1 D11 CC-CM (C152S) AA: 29–205-GST, PCR was used to amplify a portion of hDUSP11 corresponding to amino acids 29–205 (Genbank accession #: NM_003584) using the pcDNA3.1-puro-3XFLAG-DUSP11-CM [12] as template and this amplicon was cloned into pGEX-4T1 between the Not-I and BamHI restriction sites. For pGEX 4T-1 SWPV2 vD11-CM (C135S) AA: 1–197-GST, PCR was used to amplify a portion of the viral DUSP11 corresponding to amino acids 1–197 (Genbank accession #: ARE67303.1) using the plenti-EF1α SWPV2-CM vDUSP11 as template and cloned into pGEX-4T1 between the Not-1 and BamHI restriction sites. Clones were confirmed using whole plasmid sequencing (Plasmidsaurus) or using Sanger sequencing.

A549, BHK-21 and Vero cells were obtained from ATCC. A549 DUSP11 CRISPR–Cas9 targeted cell lines were previously described [12]. Cells were maintained in DMEM supplemented with 10% (v/v) fetal bovine serum (Corning) and 1% (v/v) penicillin-streptomycin (Corning). Generation of lentiviral particles, transduction and selection with blasticidin were performed as previously described [50]. A549 DUSP11 KO reconstituted cell lines were generated via transduction of lentiviral particles generated from the following lentiviral vectors: pLenti empty-vector (EV), pLenti DUSP11-3xFLAG, DUSP11 CM-3xFLAG, pLenti SWPV2 vDUSP11-3xFLAG, pLenti SWPV2 vDUSP11 CM-3xFLAG, pLenti ChePV vDUSP11-3xFLAG, pLenti ChePV vDUSP11 CM-3xFLAG, or pLenti DUSP12-3xFLAG as a negative control. For cells used in VACV infection experiments, A549 non-targeting (A549^NT) [34], A549^∆RIG-I [34], A549^∆DUSP11+EV, A549^{∆DUSP11+DUSP11}, A549^{∆DUSP11+SPXV DUSP11}, A549^{∆DUSP11 +ChePV DUSP11}, and BHK-21 cells were cultured in Dulbecco’s modified eagle’s medium (DMEM) (Sigma, D6429) supplemented with 10% FBS (R&D, S12450), 1% antibiotics/antimycotics (Sigma, A5955), 1% MEM nonessential amino acids (Corning, 25–025-CI), and 1% L-glutamine (Corning, 25–005-CI).

Mutant VSV (rM51R-M-GFP) was provided by Douglas Lyles (Wake Forest School of Medicine, NC) and has been described previously [29]. Stocks were grown in BHK-21 cells and infectious virus concentration was determined using Vero cells. Wild-type vaccinia virus (VACV) (strain WR; obtained from Dr. Bernard Moss, NIH) and the WR VACV strain encoding a deletion of E3L (VACV ∆E3L) [31] were propagated in BHK cells using low MOI infections. Virus stocks were titered on A549 cells using an immunofluorescence-based assay that detects VACV I3L protein staining as a readout for infection. Briefly, ~ 30,000 cells were seeded onto 35 mm glass coverslips and allowed to adhere overnight. Cells were then infected with serial dilutions of each virus stock for 8 hours, fixed with methanol, washed with PBS extensively, incubated at RT with blocking buffer (PBS with 1% BSA and 0.1% Triton X-100) for 1 hour, stained with mouse-anti-VV I3L antibody [51] for 2 hours, extensively washed with blocking buffer, incubated with Alexa Fluor 488-conjugated donkey anti-mouse secondary antibodies (Invitrogen Cat# A21202) for 1 hour, extensively washed with blocking buffer and then mounted onto glass slides using ProLong Diamond anti-fade with DAPI. Images were then captured on an Olympus Fv10i-LIV confocal microscope equipped with 405 nm and 463 nm lasers using a 60x objective.

DUSP11 AA sequences and related information were retrieved from Uniprot.org (ID: O75319). Accession numbers and related information for other atypical dual-specificity phosphatases for various host species were determined using HGNC [52] and sequences were determined using NCBI [41], Uniprot.org and BLASTp [45] for confirmation (S4 Table). AA sequences were retrieved from NCBI [41]. A multiple sequence alignment (MSA) of AA sequences were performed using Clustal Omega standard parameters implemented in Geneious prime with manual adjustments as needed (S9 Fig) [46]. Phylogenetic analysis was performed using PhyML with bootstrap replicates set to 100 [42]. Model selection was performed by Smart Model Selection (SMS) integrated into PhyML [53]. FigTree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/) was used for visualization. Additional phylogenetic analysis was performed using the same parameters for host and avipox viral DUSP11 sequences (S8 Table). MSA of AA sequences was visualized using Jalview (S9 and S10 Figs) [54].

A partial crystal structure of hDUSP11 was obtained from wwPDB.org (PDB: 4JMJ) [25]. Alphafold2 was used to generate a full-length structure for hDUSP11, using standard parameters and the rank 1 structure was used for analysis. The crystal structure of hDUSP11 was superimposed on the Alphafold2 structural prediction for verification of solved regions, and the Alphafold2 structural prediction was ultimately used in depictions. Structural predictions of the SWPV2 vDUSP11 were generated using AlphaFold2 using standard parameters and the rank 1 structure was used for analysis [24]. UCSF chimera was used for visualization mapping, and analysis [55].

To generate 5’-triphosphorylated HCV 5’ UTR RNA used for liposomal delivery and the in vitro phosphatase assays, in vitro transcription was performed with the AmpliScribe T7-Flash Transcription Kit (Bioresearch Technologies) according to manufacturer’s instructions. DNA templates for the in vitro transcription reaction were previously described [13]. RNA was then purified using MicroSpin G-25 columns [GE Healthcare].

A549 DUSP11 KO cells reconstituted with various DUSP11/vDUSP11s were plated at a density of 0.1 x 10⁶ cells per well in a 12-well plate for virus infection. After 1 hour of virus adsorption with gentle shaking at 15-minute intervals, virus inoculum was aspirated from the wells and washed gently with PBS before adding growth medium. GFP images were taken at indicated time points, one representative image is shown with replicates shown in S11 Fig. 100 μL of media was collected at select time points post infection and stored at -80°C to determine the yield of infectious virus. Virus concentration was quantified by standard plaque assay on Vero cells in 6-well plate format, serially diluted in a countable range (5–250 plaques per well). 72 hours post-carboxymethyl cellulose (2.5%) overlay, plates were fixed and stained with 0.25% Coomassie blue in 10% acetic acid, 45% methanol or 0.5% methylene blue in 50% methanol.

A549 cells were either mock-infected or infected with either wild-type VACV (strain WR) or VACV∆E3L (MOI = 10) in serum-free DMEM for 1 hour. Viral inoculum was replaced after 1 hour with complete media. Cells were collected at 8 or 16 hpi for subsequent RNA extraction. Total RNA was extracted using the RNeasy Mini Kit (Qiagen, 74104) according to the manufacturer’s protocol. Subsequent RT-qPCR analysis was conducted on blinded samples.

Northern blot analysis was performed as previously described [56]. In brief, total RNA was harvested from cells using TRIzol reagent (Thermo Fisher) or PIG-B [56,57] and fractionated on 10–12% polyacrylamide gel electrophoresis (PAGE)-urea gel and transferred to Amersham Hybond-N+ membrane (GE Healthcare). Transferred membranes were crosslinked 2x with a UV crosslinker (UVP) with the RNA side of the membrane facing the UV source at 1,200 μJ/m². Crosslinked membranes were prehybridized for ~ 1 hour at 55°C in hybridization solution (Takara). DNA oligos (S10 Table) were radiolabeled with [γ-³²P] ATP (Perkin Elmer) by T4 polynucleotide kinase (New England Biolabs) for ~ 1 hour at 37°C. Radiolabeled probes were added to membranes following prehybridization and hybridized while rotating in UVP hybridization oven (HL-2000 HybriLinker UVP) overnight at 38°C. Probed membranes were exposed to a phosphor screen and visualized using Typhoon Biomolecular Imager (GE Healthcare). Membranes were stripped by incubating membrane with boiled 0.1% SDS with agitation for 15 minutes and repeated three times. Replicate northern blots shown in S12 Fig.

pGEX 4T-1 D11 CC AA: 29–205-GST, pGEX 4T-1 SWPV2 vD11 AA: 1–197-GST, pGEX 4T-1 D11 CC-CM (C152S) AA: 29–205-GST, pGEX 4T-1 SWPV2 vD11-CM (C135S) AA: 1–197-GST were expressed in NEB Express Iq Escherichia coli strain that produces large quantities of LacI repressor at 37°C. When OD600 reached 0.4, cells were induced with 0.05 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and subsequently grown for over 12 hours at 25°C. Cells were then incubated with lysis buffer (50 mM Hepes, pH 7.5, 500 mM NaCl, 10 mM imidazole, 5 mM β-mercaptoethanol, 0.1% (v/v) Triton X-100, and Roche Protease cOmplete EDTA Free Protease Inhibitor Cocktail. St.Louis, Missouri) and lysed by sonication (Q125 sonicator) on ice. Lysates were sonicated for 6 seconds three times with 1-minute breaks between each interval. The lysates were then centrifuged at 12,000 rcf for 45 minutes. The supernatant containing the GST fusion proteins were purified in batch by incubating lysates with 300 µ L glutathione sepharose 4B beads (Cytiva, Massachusetts) in PBS buffer for 4 hours at 4°C. The lysates were washed in tenfold volume of PBS for 8 rounds of washes. The washes were conducted at 4°C. Enzyme activity was initially assayed by p-nitrophenyl phosphate hydrolysis (Sigma-Aldrich. St.Louis, Missouri). The pNPP liquid was serially diluted (half dilutions) three times from the original 500mM stock concentration. 5 µ L of pNPP liquid was incubated with 5 µ L of purified protein preparations (GST, D11, SWPV, D11-CM, SWPV-CM) for 30 minutes at 37C in a 50 µ L reaction (50mM Tris pH 8.0, 50nM NaCl, 5mM DTT). 50 µ L of total reaction volume was added to each well of a 96 well plate was read by a microplate reader at an absorbance of 405 nm. The results showed an increase in absorbance for WT relative to the negative controls reactions (reaction buffer only, GST, D11-CM, SWPV-CM). Protein concentrations were estimated by comparing the intensity of the relevant Coomassie blue-stained band migrating at the appropriate size compared to known amounts of purified BSA protein. Proteins were visualized on a 12% SDS denaturing gel stained with coomassie blue. The gel was ran at 120V for 2 hours. The size migration ladder used is Color Protein Standard Broad Range (New England Biolabs, Massachusetts).

In vitro translated DUSP11, DUSP11-CM, SWPV2 vDUSP11, SWPV2 vDUSP11-CM lysates were generated using the TnT quick-coupled transcription/translation system (Promega) and the pcDNA3.1 DUSP11-3xFLAG, pcDNA3.1 DUSP11-CM-3xFLAG, pcDNA3.1 SWPV2-vDUSP11-3xFLAG, and pcDNA3.1 SWPV2-vDUSP11-CM-3xFLAG, plasmids as templates. Five micrograms of HCV 5’ UTR RNA was incubated in a 100-μL reaction [50mM Tris, 10mM KCl, 5mM DTT, 50mM EDTA, 40 units SUPERaseIn (Thermo Fisher Scientific)] with 6 μL of in vitro translated products (DUSP11-3xFLAG, DUSP11-CM-3xFLAG, SWPV2-vDUSP11-3xFLAG, SWPV2-vDUSP11-CM-3xFLAG, or a luciferase negative control) for 10 min at 37°C. EDTA was omitted from CIP control reactions as it is inhibitory of its enzymatic activity. RNA was purified using PIG-B [57]. Purified RNA from each phosphatase or control reaction was then used to set up two reactions 20 μL reactions containing 1 μg of treated RNA in NEBuffer 3.1 (New England Biolabs) with or without the addition of 1 μL (1U) recombinant XRN1 (New England Biolabs) and was then incubated for 60 minutes at 37°C as previously described [13]. The reactions were fractionated by 7.5% urea/PAGE and were then stained using ethidium bromide (~1 μg/mL) for 3–5 minutes while rocking. The Bio-Rad Molecular Imager GelDoc XR was then used to visualize gel via exposure to UV. Gel images were processed using Fiji (ImageJ variant) [58,59]. Replicate EtBr-stained urea gels shown in S13 Fig. For reactions where purified GST fusion proteins were assayed 10 µg of GST, 0.70 µg of GST-D11CC, 10 µg of GST-D11-CM, 0.50µg of GST-SWPV, or 2.25 µg of GST-SWPV-CM was added. In all experiments with purified proteins, increased amounts of CM variants and negative controls relative to WT proteins were utilized.

A549 DUSP11 KO reconstituted cells were seeded at a density of 0.1 X 10⁶ cells per well in a 12-well format. 24 hours following plating, cells were transfected with 5–10 ng of in vitro transcribed 5’-triphosphorylated HCV 5’ UTR RNA, or with Lipofectamine 2000 (Invitrogen) alone control (3 µ L). Transfection was terminated after 18 hours by the addition of either TRIzol reagent (Invitrogen) or PIG-B, followed by RNA extraction and subsequent RT-qPCR analysis.

Total RNA was extracted from cells using TRIzol reagent or PIG-B [57]. Extracted RNA was treated with DNase I (Thermo Fisher) followed by ethanol precipitation (100% ethanol, 3M sodium acetate). Precipitated RNA was resuspended in nuclease-free water (Thermo Fisher) and was quantified using the NanoDrop ND-1000 Spectrophotometer. RNA purity was assessed using the measured OD at 260/280 nm, with RNA considered acceptable having values within ~1.7-2.1. Complementary DNA (cDNA) was synthesized using SuperScript III reverse transcriptase (Invitrogen) using equal amounts of RNA. All RT-qPCR quantification and measurements were performed on a StepOnePlus Real-Time PCR system (Applied BioSystems). A standard 3-step PCR cycling protocol (95°C for 15 sec, 60°C for 15 sec, 68°C for 30 sec, 40 cycles) using PerfeCTa SYBR Green FastMix (Quantabio) was used according to the manufacturer’s instructions. At least two technical replicates per sample/condition. Gene expression of human IFNB1 and ISG15 were normalized to the expression of GAPDH (S10 Table). GAPDH values for A549 DUSP11 KO + EV cells transfected with L2K alone were used when defining the experiment negative control within the Applied BioSystems program [14].

A549 DUSP11 KO reconstituted lines were plated at 0.5 X 10⁵ cells per well in a 24-well plate containing a 12-mm glass coverslip. The next day cells were washed three times with 1X PBS before fixation with 2% paraformaldehyde for 20 minutes. Paraformaldehyde was removed and cells were washed three times with 1X PBS. Following washes, cells were permeabilized with 0.5% Triton X-100 in 1X PBS 3% BSA for 2 minutes at room temperature. Next, the permeabilization buffer was removed and cells were incubated in blocking buffer (0.2% Triton X-100, 3% BSA, 1X PBS) while rocking for 1 hour at room temperature. Following incubation, M2 FLAG antibody (Sigma Aldrich #F1804) was diluted to 1:1000 in blocking buffer and incubated overnight at 4°C. The next day cells were subjected to three 5-minute washes in blocking buffer. Secondary antibodies (S10 Table) were diluted 1:1000 in blocking buffer and added to cells to incubate for 30 minutes at room temperature in the dark. Cells were washed three times in blocking buffer for 5 minutes each. ProLong Diamond Antifade Mountant (Invitrogen #P36961) was used when mounting coverslips on slides. Cells were imaged on ZEISS confocal using the 40x objective (1 representative image shown). Images were processed to final form using ZEISS image analysis software and Image J.

Protein was extracted from cells using either RIPA lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 0.25% sodium-deoxycholate, 1% NP-40, 0.1% SDS with protease inhibitor) supplemented with Protease and Phosphatase Inhibitor Cocktail (Abcam, ab201119) or SDS page sample buffer (1X) [60]. Cell lysate was fractionated on an 8% SDS-PAGE and transferred to Amersham Protran 0.45 μm nitrocellulose membrane (GE Healthcare). Blots were incubated in blocking buffer (Intercept PBS blocking buffer, LI-COR) then probed with primary antibodies for M2 FLAG (monoclonal, 1:2000 dilution, Sigma Aldrich, F1804) and beta-Tubulin (polyclonal, 1:2000 dilution, Cell Signaling, 2146S) in blocking buffer overnight at 4°C. Blots were washed with ~10 mL 1X PBST three times for 5 minutes prior to incubation with secondary antibodies. IRDye 800CW (1:10,000 dilution, LI-COR, 926–32213) and IRDye 680LT (1:10,000 dilution, LI-COR, 926–68022) were used as secondary antibodies, diluted in blocking buffer. Blots were washed with ~ 10mL 1X PBST three times for 5 minutes, then imaged with an Odyssey CLx infrared imaging system (LI-COR). Uncropped western blots shown in S14 Fig.

NCBI was used to identify genomic locations for viral DUSP11 sequences within the viral genomes [41]. Accession numbers for adjacent coding portions corresponding to the 5 genes upstream and downstream from each viral DUSP11 were recorded (S5 Table). Reciprocal BLAST hits (RBH) were performed for coding sequences of vDUSP11-adjacent genes across virus species to identify homologs (S5 Table) [45]. Genomic database annotations and the NCBI database were used for gene identities and homology [41]. CoreGenes 5.0 was used to further verify the presence of homologous genes between selected poxvirus genomes [61].

GraphPad Prism software was used for statistical analyses. For Fig 5, a standard unpaired student t-test was used on transformed (Y = log(X)) viral titers. For Fig 6, panels A and B were analyzed using a standard unpaired students t-test, panels C and D were analyzed using one-way ANOVA followed by Dunnett’s multiple comparison test. Results for Dunnett’s multiple comparison test are indicated on corresponding figures. ANOVA results for VACV infected samples for panels C and D were P = 0.01 and P = 0.08, respectively. For S6 Fig, panels C and D were analyzed using a standard unpaired students t-test, panels G and H were analyzed using one-way ANOVA followed by Dunnett’s multiple comparison test. Results for Dunnett’s multiple comparison test are indicated on corresponding figures. ANOVA results for panels G and H were P = 0.07 and P = 0.16, respectively. For all other statistical analysis, a standard Student t-test was used with significance set at P = 0.05. Error bars were presented as mean ± standard error of the mean. Statistical consultation was provided by the Department of Statistics and Data Sciences at the University of Texas at Austin.