Mouse Schlafen Family Proteins Control MHV68 Replication in Fibroblasts

1. Introduction

The Schlafen (SLFN) protein family are a group of vertebrate-specific regulatory proteins involved in modulating several cellular responses, including viral infection, cell cycle arrest, and DNA damage [1]. Several of the SLFN genes are type I interferon (IFN)-stimulated with direct immunological and antiviral roles, which has led to the classification of many SLFNs as broad antiviral restriction factors effective against numerous RNA and DNA virus families [1,2]. All SLFN proteins encode an N-terminal core SLFN domain defined by a conserved motif called the SLFN box [3,4,5]. Additionally, the SLFN domain contains three endonuclease active site residues found in the majority of SLFN proteins. The SLFN protein family is further classified into groups I, II, or III based on overall domain structure and size [6]. SLFNs are further classified into groups I, II, or III based on overall domain structure and size [6]. While group I SLFN proteins are comprised of a SLFN core domain only, groups II and III contain a SWAVDL linker domain following their N-terminal SLFN core domain. Group III proteins have a third helicase domain at their C-terminus with implications in ssDNA binding and DNA damage [7,8,9,10]. At present, the functions of these additional domains in SLFN proteins are still being defined.

SLFN proteins are known to bind and, in some cases, cleave transfer RNAs (tRNAs) through their N-terminal domain and endonuclease active site residues; moreover, tRNA binding and cleavage has been implicated in SLFN antiviral activity [6,11,12,13,14,15]. Human SLFN11 (hSLFN11) and hSLFN13 can restrict HIV-1 infection by selectively inhibiting the production of viral proteins based on codon usage, suggesting that SLFN tRNA binding activity plays a role in its antiviral function [11,13]. Further, in vitro studies clearly demonstrate mouse Slfn8 (mSlfn8) and hSLFN11-14 can cleave mature tRNAs at D and T-loop regions, generating various tRNA fragments (tRFs) [13,15,16]. This endonuclease activity affords SLFN proteins the potential to manipulate viral translation by inactivating mature tRNAs, as well as to produce effector RNAs that can independently modulate gene expression. In support of this hypothesis, N-terminal domain mutants of hSLFN11, 12, 13, and 14 exhibit reduced antiviral activity against several RNA viruses, an activity that is attributed to their tRNA/rRNA endoribonuclease activity [2,13,17,18,19]. Finally, mature and premature tRFs induced by viral infection have been shown to silence gene expression, but whether these are generated by SLFN proteins is unknown [20,21,22]. Overall, it appears that SLFNs effectively restrict cellular tRNA functionality, which in turn inhibits viral translation.

To date, hSLFN5, hSLFN11, hSLFN12, and hSLFN14 have been shown to have antiviral activity against herpesviruses [23,24,25,26]. hSLFN5 suppresses HSV-1 transcription by directly binding to the viral genome and limiting RNA polymerase accessibility; however, hSLFN5 is not known to have tRNA cleavage activity [23,27]. hSLFN11 and hSLFN12 have both been shown to restrict beta- and gammaherpesvirus infection, respectively; however, whether restriction requires tRNA cleavage is not known [24,25]. Similarly, hSLFN14 seems to have an antiviral effect against varicella zoster virus and although hSLFN14 has been shown to cleave tRNA/rRNA, it is not clear whether its antiviral activity can be attributed to tRNA cleavage, ribosomal rRNA degradation, or another function [26,28,29]. Overall, the role of individual SLFNs in response to herpesvirus infection is not well characterized. Additionally, which mouse Slfn proteins participate in herpesvirus restriction is unknown, though mSlfn2 mutant mice (also known as elektra-mutant mice) are more susceptible to murine cytomegalovirus infection [30]. The only clear orthologs of hSLFNs are mSlfn5 (59% sequence identity with hSLFN5), mSlfn14 (70% sequence identity with hSLFN14), and mSlfn9 (63% sequence identity with hSLFN13)[2]. The fact that both mouse and human SLFN gene families have undergone extensive diversification, resulting in proteins that cleave tRNAs but also induce apoptosis, cell cycle arrest, and DNA damage repair, indicates that SLFNs may have diverse functional roles in restricting viral replication [1].

To further clarify whether mouse Slfns are induced by and respond to herpesvirus infection, we became interested in exploring mouse Slfn proteins as potential restriction factors for murine gammaherpesvirus 68 (MHV68), a model murine virus related to human Kaposi sarcoma associated herpesvirus and Epstein Barr virus. We report that mSlfn8 is upregulated in response to MHV68 infection in both mouse fibroblasts, B cells, and primary peritoneal macrophages. We find that blocking MHV68 viral DNA replication enhances mSlfn8 expression, demonstrating viral antagonism of mSlfn8 in a manner similar for other interferon-stimulated genes. Finally, we find evidence of control of MHV68 by mSlfn8, as well as family member mSlfn2, suggesting broad antiviral activity of gammaherpesvirus infection by mouse Slfn proteins.

2. Materials and Methods

2.1. Cell Lines and Culturing Conditions

NIH 3T3 (NIH/Swiss-derived) murine fibroblasts, NIH 3T12 murine fibroblasts, and HEK293T human embryonic kidney cells (ATCC) were maintained in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Cultures were screened regularly for mycoplasma (Invivogen) and cultured in the absence of antibiotics. NIH 3T12s were used for virus propagation. Peritoneal cells from C57BL/6 mice were isolated following euthanization by lavage with 10 mL of ice cold RPMI1640 + 10% FBS + 1% pen/strep and cultured in the presence of 50 ng/mL murine M-CSF. Forty-eight hours after plating, non-adherent cells were removed by washing with PBS to obtain enriched peritoneal macrophage cultures.

2.2. siRNA Transfection

For siRNA transfection, pooled siRNA (Table 1) purchased from Dharmacon was nucleofected into 5x10⁵ NIH 3T3 cells at 200 nM with the Neon™ Transfection System (Thermo Fisher) using buffer E2, 100 uL tips and the settings: 1400 volts/20 ms/2 pulses. Cells were infected 24 h post-transfection or post-nucleofection.

2.3. Virus Infections

MHV68-GFP was amplified in NIH 3T12s and titered by TCID50 on NIH 3T3s as described previously [31]. For infection, MHV68 virus was added to ½ culture volume of media, placed on NIH 3T3 cells as indicated for 60-90 min at a multiplicity of infection (MOI) of 5 or 0.05 (as calculated by TCID50) to allow viral entry. Cells were washed with PBS to remove unabsorbed virus, then replenished with fresh DMEM +10% FBS medium. The same protocol was used for infection of peritoneal macrophages, except an MOI=10 was used. For phosphonoacetic acid (PAA) treatments, PAA (Sigma) was added at a final concentration of 200 μg/ml into the media after viral absorption. Supernatants were titered by TCID50 on NIH 3T3.

2.4. RNA-Seq Datasets

Datasets from MHV68-infected NIH 3T3s and A20-HE-RITs were downloaded from the SRA database: SRR19792326, SRR19792325, SRR19792324, SRR19792321, SRR25823338, SRR25823339 [32]. MHV68-infected peritoneal macrophages were suspended in Trizol and RNA was purified and sent to Genewiz. Whole transcriptome sequencing was performed via rRNA-depletion and analysis performed by Genewiz’s basic bulk RNA-sequencing package. Sequence reads were trimmed using Trimmomatic v.0.36 and mapped to a custom combined mouse + MHV68 genome file using the STAR aligner v.2.5.2b. Differential expression analysis was performed with DESeq2. The percent-of-family calculation was done by taking ERCC normalized read counts or transcripts per million (TPM) for each of the eight mSlfn genes analyzed here (mSlfn1,2,3,4,5,8,9,14), then calculating the percent for each gene over the total for the family. Raw data is deposited under BioProject accession PRJNA1520822 (SRX35015971, SRX35015972, SRX35015973, SRX35015974, SRX35015975, SRX35015976).

2.5. Schlafen Gene Knockout

Lentiviruses expressing sgRNA were generated by cloning gRNA oligos (sequences in Table 1) into the pMCB320 lentiviral vector carrying mCherry (Addgene #89359). Lentiviruses were packaged with 3rd generation packaging plasmids in HEK293Ts as follows. HEK293T cells were seeded 24 h before transfection in. 6-well dish at density of 1.5x10⁶ cells/well. HEK293T cells were co-transfected with pMD2.G (0.31 μg), pMDL.G (0.63 μg), pRSV (0.31 μg), and pMCB320 (1.25 μg) using PolyJet (SignaGen). Lentivirus-containing supernatant was collected 48h post transfection, 0.45 μM syringe filtered, and frozen at -80 °C. Cas9-expressing NIH3T3s (Dharmacon; HD Cas9-010) were plated in a 6-well at 1x10⁶ cell/ well and transduced with sgRNA lentivirus in the presence of polybrene (4 μg/mL). Plates were spinfected for 2 h at 800 x g. The medium was changed at 48 h and antibiotic selection for pMCB320 (puromycin (5 µg/mL) and Cas9 (blastocidin (5 µg/mL)) was added. Single cell clones positive for mCherry expression were FACS sorted using a Cytek Aurora 2 Flow into a 96-well dish with the UI Flow Cytometry Facility.

2.6. Knockout Validation

Genomic DNA was isolated from Slfn knockout and non-targeting control cells using the DNeasy Blood and Tissue Kit (Zymo Research). The sgRNA target site was amplified by PCR and fragments were resolved on a 2% agarose gel (primer sequences in Table 1). Bands were excised and purified using the Zymoclean Gel DNA Recovery Kit (Zymo Research), followed by the Zymo Clean and Concentrate Kit (Zymo Research) to further purify the DNA. Purified fragments were submitted to Plasmidsaurus for Premium PCR service. The chromatograms were analyzed using the ICE analysis program for the presence of indels (insertion/deletion) [33]. The KO score was >95% for all clones except mSlfn2-KO1, which was manually verified to carry a 170 bp deletion at the cut site, which generated a premature stop codon and truncates mSlfn2 after amino acid K40 (UniProt: Q9Z0I6).

2.7. Cell Growth and Viability Assays

To measure cellular proliferation, control and mSLFN-KO cells were seeded in triplicate at 3.5x10³ cells per well in a 96-well plate. Cells were then imaged at 24, 48, and 72 h post-seeding. Images of the plate were taken with the Agilent Biotek Cytation 5 with the high contrast brightfield accessory kit. Measurements of cell counts were analyzed with Agilent BioTek Gen5 image analysis software. Cell counts were converted into fold change over time and plotted using GraphPad Prism. LDH release was measured using an LDH-Glo kit (Promega) and a Tecan plate reader.

2.8. Hydroxyurea Treatment

Control and mSLFN-KO cell lines were seeded in 2x 96-well plates at 3.5E3 cells per well in 100uL of DMEM +10% FBS. The cells were allowed to adhere for 24 h and then imaged with an Agilent Biotek Cytation 5 with the high contrast brightfield accessory kit. The media was removed from the plates and replaced with DMEM +10% FBS and 0-250 mM Hydroxyurea (Sigma H8627). After 24, 48, and 72 h of HU treatment, plates were imaged for cell count using Agilent BioTek Gen5 image analysis software. Cell counts were converted to fold change over time and plotted using GraphPad Prism software.

2.9. RT-qPCR

Total RNA was isolated from cells using TRIzol (Invitrogen) and was treated with TURBO DNase (Ambion) to remove contaminating DNA. TURBO-treated RNA was then reverse transcribed with AMV RT (Promega) primed with random 9-mers. Gene expression was measured by amplifying cDNA using iTaq SYBR Green Supermix (Biorad) on a QuantStudio 3 (Applied Biosystems), and all primers are listed in Table 1. RT-qPCR was analyzed using the delta delta Ct (ΔΔCt) method. Average fold changes were calculated relative to 18S internal controls.

2.10. Statistical Analysis

All experiments were performed in at least biological duplicate or triplicate, meaning that experiments were performed on different days with different cell populations in terms of stock vial or passage number. All RT-qPCR statistical analysis was performed using raw ΔCt values. The statistical test used is indicated in the figure legends.3. Results

This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn.

3. Results

3.1. MHV68 Infection Induces mSlfn Expression

Human SLFN proteins have been shown to play a key role in the response to human herpesvirus infection, including HSV-1 and HCMV [23,24]. Mice encode at least nine Slfn family members (mSlfn1,2,3,4,5,8,9, 10, and 14), but which of these might be antiviral apart from mSlfn2 is not known [30]. To explore which mSlfn transcripts are expressed in response to MHV68 infection, we analyzed RNA-seq data from three different cell culture models of MHV68 infection: de novo MHV68 infection (MOI=5, 18 h) in NIH 3T3 fibroblasts (Figure 1A), de novo MHV68 infection (MOI=10, 24 h) in primary murine peritoneal macrophages (Figure 1B), and MHV68 lytic reactivation (24h doxycycline treatment) in A20-HE-RIT B cells [32,34] (Figure 1C). mSlfn10 was not included in our analysis, as it is classified as a non-coding pseudogene in C57BL/6, and whether it has coding capacity is not well defined for all cell lines and mouse strains. We examined two parameters: 1) percent of family expression (i.e., the relative expression of each mSlfn as a percentage of the eight measured mSlfn family proteins during lytic replication) and 2) log2(fold-change) of each lytic condition vs control (i.e., which mSlfns are induced by MHV68 de novo infection or reactivation; X-axis). Following infection in fibroblasts, mSlfn3, mSlfn4, mSlfn5, mSlfn8, and mSlfn14 were induced 2- 200-fold by MHV68 (Figure 1A; Y-axis). However, we also found these genes to be poorly expressed relative to other mSlfns in infected NIH 3T3s (Figure 1A; X-axis). Following infection of primary peritoneal macrophages, mSlfns were more evenly expressed across the family, with mSlfn1, mSlfn4, mSlfn5, mSlfn8 showing the strongest induction with MHV68 infection (Figure 1B). Reactivation of A20-HE-RITs led to induction of a slightly different profile, including enhanced expression of mSlfn1, mSlfn4, mSlfn5, mSlfn8, mSlfn9, and mSlfn14 (Figure 1C). mSlfn8 expression made up the majority of mSlfn family expression in A20-HE-RITs (Figure 1C). Overall, this analysis suggests that a subset of mSlfn genes are induced by murine gammaherpesvirus replication, with three family members (mSlfn4, mSlfn5, and mSlfn8) induced across three cell culture models of infection, though expression level can vary across the cell types and/or infection models.

To appreciate the difference in mSlfn expression following MHV68 infection in NIH 3T3 fibroblasts, we plotted the ERCC normalized read counts for each mSlfn gene [32]. Reads were detectable (defined as >10 reads) for mSlfn2, mSlfn8, and mSlfn9 transcripts, while read counts for other members of the mSlfn gene family were extremely low (<2 reads) (Figure 2A). mSlfn9 appeared to be constitutively expressed in NIH 3T3s, on par with another host mRNA Rrbp1, and exhibited decreased expression (down 2-fold) in response to MHV68. This trend is consistent with other host mRNAs (see Rrbp1 in Figure 1A; down 3.3-fold with infection) during infection, likely a result of gammaherpesvirus-induced host shutoff [32,35]. In contrast, mSlfn2 expression was stable before and after infection, while mSlfn8 was induced 2.6-fold in response to MHV68 infection. mSlfn8 has been shown to have tRNA cleavage activity and shares approximately 60% amino acid identity with hSLFN11 and 64% with hSLFN13 [2,13]. Due to their high sequence similarity, and because hSLFN11 was shown to restrict HCMV, we investigated a potential role for mSlfn8 during MHV68 infection.

3.2. mSlfn8 Expression Is Induced by PAA and in Low MOI Infection Conditions

We next explored RNA-seq data from NIH 3T3s infected with MHV68 in the presence of phosphonoacetic acid (PAA), which inhibits the viral DNA polymerase, blocking viral DNA replication and subsequent late gene expression [36]. Surprisingly, mSlfn expression appeared to be enhanced by PAA treatment [32]. While mSlfn8 was shown previously to be stimulated by the IFN pathway [25,37], to our knowledge, PAA has not been formally reported to enhance the IFN response. We then decided to determine the kinetics of mSlfn8 expression upon MHV68 infection across a synchronized cycle of infection (MOI=5, 24 h), both in the absence and presence of 200 μg/ml PAA (Figure 2B). Although mSlfn8 was detected in the previous RNA-seq data (Figure 1A and 2A), mSlfn8 was difficult to detect by RT-qPCR (Ct>35) in both mock and MHV68-infected cells (Figure 2A). In contrast, mSlfn8 was strongly induced in the presence of MHV68 and PAA, unlike with MHV68 alone (Figure 2B), supporting RNA-seq data [32]. Next, we reasoned that PAA might more broadly enhance ISG expression by permitting accumulation of immunostimulatory viral DNA replication intermediates while simultaneously reducing expression of replication-dependent viral immune antagonists [38,39]. To examine this, we plotted the expression of host ISGs following MHV68 infection (MOI=5 for 18 h) in the presence or absence of PAA [32] (Figure 2C). As expected, a subset of the selected ISGs show enhanced expression upon MHV68 infection in fibroblasts. PAA addition further enhances the expression level of these ISGs above that seen with infection alone, suggesting that PAA addition during MHV68 infection does indeed enhance the IFN response. Based on this information, we predicted that mSlfn8 induction might be enhanced in a low MOI infection setting (MOI=0.05), where MHV68 does not infect every cell in the culture, allowing the culture to propagate a stronger interferon response [40]. We measured the expression of mSlfn8 in NIH3T3 cells infected with MHV68 at an MOI of 0.05. Expression of mSlfn8 was evident by 2 days post infection and peaked 3 days post infection (Figure 2D). Induction of mSlfn8 following a low MOI infection was much stronger (~1,000 fold) than that observed with a high MOI. These results demonstrate that mSlfn8 expression increases in response to MHV68 infection in the presence of PAA and are more robust during a low MOI infection compared to a high MOI infection.

3.3. mSlfn8/9 Knockdown with siRNAs Minimally Impacts MHV68 Titer

Next, we sought to determine whether mSlfn8 can restrict an MHV68 infection. To knockdown expression of mSlfn8, NIH 3T3 cells were treated with ON-TARGETplus SMARTpool siRNA (mSlfn8-targeting siRNA) or nontargeting control siRNA (Non-targeting siRNA) as a negative control. NIH 3T3 cells were infected 24 h post siRNA treatment with MHV68 at an MOI of 0.05. Though mSlfn8 antibodies have been made recently, this production was small scale and not commercially available [41]. We thus validated mSlfn8 knockdown by RT-qPCR, specifically at 3 dpi when mSlfn8 Ct values are reliably measured (see Figure 2C) and found that mSlfn8 expression decreased by approximately 20-fold in mSlfn8-targeting siRNA-treated cells compared to non-targeting siRNA treatment (Figure 3A). Upon checking the siRNA sequences in our SMARTpool, we discovered the siRNA pool could potentially target mSlfn9, which shares 93% mRNA identity with mSlfn8. Accordingly, we found our siRNA pool led to reduced expression of mSlfn9 in addition to reduced expression of mSlfn8 (data not shown). Depletion of mSlfn8 and, presumably, mSlfn9 resulted in 2-9-fold higher viral titers by 3 dpi, but this did not reach statistical significance (Figure 3B and C). Consistent with this result, we observed a slight increase in gB mRNA, a MHV68 late transcript, again without reaching statistical significance (Figure 3D). Together, this data suggests that the combined knockdown of mSlfn8 and mSlfn9 may restrict MHV68 replication, but that this could not be resolved by the selected siRNA knockdown approach.

3.4. mSlfnKO Cells Impact Growth in the Presence and Absence of a DNA Damaging Agent

Our previous results underscore the importance of rigorously validating siRNA-mediated knockdowns. Off-target effects, particularly when using pooled siRNAs, can confound interpretation of phenotypes. Thus, we used CRISPR/Cas9 to make clean knock outs of mSlfn2, mSlfn8, and mSlfn9 individually in NIH 3T3s. We used the Broad Institute CRISPick library to design two gene-specific guide RNAs per mSlfn gene (Table 1). Upon transduction and puromycin selection, we selected and amplified single cell clones and verified loss of coding capacity by amplicon sequencing and ICE analysis (details in Materials and Methods). We performed a growth curve using an imaging plate reader, including a single clone per guide alongside cells expressing safe-targeting guides (ST) to assess the growth of the cells following mSlfn knockout (Figure 4A). Over a 3-day growth period, we measured an average doubling time (T_d) of 25.2 hours for safe-targeting control cells. We observed enhanced growth for mSlfn-KO clones, with T_d ranging from 12-19 h. Enhanced growth was consistently robust with both mSlfn2-KO cell lines, consistent with other reports [37].

mSlfn8 and mSlfn9 proteins have been shown to have similarity to hSLFN11, known to play a role in the cellular response to DNA damage [10]. mSlfn8 and mSlfn9 are recruited to sites of DNA damage and can partially complement a hSLFN11-/- Hap1 cell line with regards to growth rate and sensitivity to DNA damaging agents like hydroxyurea (HU). Thus, to further validate our knockout cell lines, we measured the growth of mSlfn-KO cells in the presence of hydroxyurea. First, we performed a titration in the presence of 0-250 mM HU using our safe-targeting control cell line (Figure 4B). Growth of safe-targeting cells was inhibited starting at 100 mM and fully inhibited by 200 mM. We then compared the growth of safe-targeting with mSlfn-KO cell lines in the presence of 100 or 150 mM HU (Figure 4C-D). Minimal differences in growth for mSlfn2-KO cells were observed at 100 mM, however both mSlfn8-KO clones and one mSlfn9-KO clone exhibited decreased growth (Figure 4C). While safe-targeting and mSlfn2-KO clones grew, albeit more slowly, in the presence of 150 uM, mSlfn8-KO and mSlfn9-KO clones did not grow over the 3-day time course (Figure 4D). In contrast, a combined knockout of mSlfn8,9, and 10 genes in murine Ba/F3 cells was reported to enhance viability in the presence of hydroxyurea [10]. The results of these experiments highlight differential effects of mSlfn2, 8 and 9 for growth and sensitivity to hydroxyurea, but suggest that additional work should be done to clarify the role of mSlfns during the DNA damage response in fibroblasts.

3.5. mSlfn-KO Cells Support Higher Titer Replication of MHV68

Using these validated mSlfn-KO clones, we next investigated whether mSlfn2, mSlfn8, or mSlfn9 restrict MHV68 replication. Because mSlfn8 expression was most strongly induced during low-MOI infection (Figure 2D), we infected safe-targeting cells and the corresponding mSlfn-KO cell lines with MHV68 at an MOI of 0.05 and quantified infectious virus production at 3 dpi (Figure 5A). Because of the observed difference in growth (Figure 4), we seeded safe-targeting and mSlfn-KO cells at 100% confluency prior to infection to limit differential cellular growth during the experiment. Disruption of either mSlfn2 or mSlfn8 consistently resulted in substantially higher viral titers compared with safe-targeting cells, with increases ranging from approximately 6- to 10-fold. A similar increase was observed in one mSlfn9 knockout clone (mSlfn9-KO1), whereas a second clone using a separate guide RNA (mSlfn9-KO2) did not reproducibly enhance virus production (Figure 5A). Importantly, both independent mSlfn2-KO clones and both mSlfn8-KO clones, each made with a unique guide RNA, exhibited elevated MHV68 replication, providing strong evidence that loss of either gene enhances viral growth. Next, we analyzed cell viability of our cultures at the endpoint of infection by measuring LDH release into the media (Figure 5B). With our safe-targeting cell line, we observed a 2-fold increase in LDH release upon infection, consistent with cell death following MHV68 lytic replication. However, LDH release for the infected mSlfn-KO cell lines remained similar to uninfected safe-targeting cells. This assay supports other studies demonstrating that SLFNs play important roles in cell death in response to DNA damage or interferon stimulation (10, 14, 37). These findings identify mSlfn2 and mSlfn8 as host restriction factors that limit productive MHV68 infection, perhaps by promoting cell death.

4. Discussion

Our findings demonstrate that mSlfn8 expression is induced and antagonized during MHV68 infection, and that knockout of mSlfn8, along with mSlfn2, enhances viral titer in mouse NIH 3T3 fibroblasts. This observation emphasizes the role of mouse Schlafen family members in antiviral defense within the broader landscape of interferon-stimulated genes (ISGs) that restrict gammaherpesvirus replication. MHV68, like other herpesviruses, encodes multiple strategies to counteract interferon signaling and ISG activity [30]. Consistent with this model, mSlfn8 induction was low except when we infected under low-MOI conditions, where multiple rounds of replication are expected to amplify innate immune signaling in bystander cells. We also observed high levels of mSlfn8 expression when cells were infected in the presence of PAA, which we presume leads to enhanced interferon stimulation due to incomplete viral DNA intermediates or by blocking expression of viral immune antagonists. The induction of mSlfn8 suggests that it is part of the host antiviral response activated during infection. However, the observation that MHV68 simultaneously suppresses mSlfn8 expression indicates that the virus may have evolved mechanisms to evade or neutralize this restriction factor. Such antagonism is characteristic of host-pathogen interactions in which antiviral ISGs exert sufficient selective pressure to drive the evolution of viral countermeasures. The combined induction and suppression of mSlfn8 therefore supports the hypothesis that this protein represents a biologically relevant barrier to efficient MHV68 replication.

The increased viral titers observed in independent mSlfn8 knockout clones provide direct functional evidence that mSlfn8 restricts productive MHV68 infection. Notably, a similar phenotype was observed in mSlfn2-KO cells. Because the effects were reproducible across independent knockout clones, these findings strongly support a bona fide antiviral role for both proteins. In contrast, the inconsistent phenotype observed between the two mSlfn9 knockout clones prevents definitive conclusions regarding a role for mSlfn9 in restricting MHV68 infection. We predict that the mSlfn9-KO2 cell line may have accumulated suppressor mutations or acquired off-target genetic perturbations but have not yet tested for these possibilities. Additional studies using complementation approaches will be required to determine whether mSlfn9 contributes to antiviral defense. Finally, it is possible that there is functional redundancy among mSlfns and that knocking out multiple Slfns is required to enhance the antiviral phenotype observed.

Although the molecular mechanisms underlying mSlfn2- and mSlfn8-mediated restriction remain unclear, several possibilities can be considered. Schlafen family proteins have been implicated in diverse biological processes, including regulation of cellular proliferation, differentiation, translation, RNA metabolism, and innate immune signaling [1]. Mouse Schlafen proteins are less well characterized than their human SLFN counterparts, and many of their biological functions remain poorly understood. mSlfn2 is a group I SLFN protein carrying a single Schlafen core domain and shown to regulate immune cell quiescence and protect tRNAs from stress-induced cleavage [30,42]. In contrast, mSlfn8 is a group III SLFN protein consisting of a Schlafen core domain, SWADL linker, and an RNA/DNA helicase domain, and has similarities to human SLFN11/13 that serve antiviral roles and are important in the DNA damage response [10]. Not only do mSlfn2 and mSlfn8 differ by number of protein domains, but they also differ by the presence of functional endonuclease activity within their shared Schlafen core domain. While mSlfn8 retains an endonuclease active site and can cleave tRNAs in vitro, mSlfn2 lacks this activity and instead binds and protects tRNAs from cleavage [13,42]. Indeed, we found distinct roles of mSlfn2 and mSlfn8 regarding cellular growth in the presence and absence of DNA-damaging reagents. Knockout of mSlfn2, 8, and 9 all had positive effects on cell growth, with mSlfn2-KO cells showing the most robust increase in growth. Due to these findings, we performed MHV68 infection on fully confluent monolayers to avoid growth effects on viral titer. In contrast to general growth characteristics, only mSfln8 and mSlfn9, but not mSlfn2, were important for cell growth in the presence of hydroxyurea. Our result showing lack of growth for mSlfn8-KO and mSlfn9-KO cells in hydroxyurea was unexpected given results in Ba/F3 cells, where mSlfn8/9/10-/- enhanced viability under similar conditions [10]. This difference could be attributed to several profound differences between the experiments, including single vs triple mSlfn knockout, cell type (pro-B cell vs fibroblasts), hydroxyurea concentrations, and assays used (growth vs viability). This would need to be experimentally determined in future studies to define a role for mSlfn8 and mSlfn9 during DNA damage. Taken together, we posit that the differences in mSlfn2 and mSlfn8 domain structure and function suggest that these proteins likely restrict MHV68 through distinct mechanisms, which we are currently investigating.

Several limitations of the present study should be acknowledged. First, mSlfn knockouts were made and tested in NIH 3T3 fibroblasts, and the contribution of mSlfn8 to antiviral defense may differ in cell types that are naturally infected by MHV68 in vivo, including B cells, macrophages, and dendritic cells [43]. Second, although knockout studies demonstrate that mSlfn2 and mSlfn8 limit virus replication, complementation experiments restoring gene expression would further strengthen the causal relationship between Schlafen loss and enhanced viral growth. Third, given that mSlfn8 and mSlfn9 display functional redundancy, at least with regards to DNA damage response, MHV68 infection should also be characterized in mSlfn8/9 double knockout cells [10]. Finally, the viral factor or factors responsible for mSlfn8 (and other ISG) antagonism remain unknown and represent an important topic for future investigation.

In summary, this study identifies mSlfn8 as a previously unrecognized restriction factor for MHV68 and provides additional evidence that mSlfn2 also contributes to antiviral defense against gammaherpesvirus infection. The induction of mSlfn8 during infection, coupled with its apparent antagonism by MHV68 and the enhanced viral replication observed in knockout cells, supports a model in which mSlfn8 functions as part of the interferon-induced antiviral response. These findings expand the understanding of Schlafen protein function during herpesvirus infection and establish a foundation for future studies aimed at defining the molecular mechanisms by which Schlafen family members restrict gammaherpesvirus replication.