Herpesviruses mimic zygotic genome activation to promote viral replication

DUX4 is a germline transcription factor and a master regulator of zygotic genome activation. During early embryogenesis, DUX4 is crucial for maternal to zygotic transition at the 2–8-cell stage in order to overcome silencing of genes and enable transcription from the zygotic genome. In adult somatic cells, DUX4 expression is silenced and its activation in adult muscle cells causes the genetic disorder Facioscapulohumeral Muscular Dystrophy (FSHD). Here we show that herpesviruses from alpha-, beta- and gamma-herpesvirus subfamilies as well as papillomaviruses actively induce DUX4 expression to promote viral transcription and replication. We demonstrate that HSV-1 immediate early proteins directly induce expression of DUX4 and its target genes including endogenous retroelements, which mimics zygotic genome activation. We further show that DUX4 directly binds to the viral genome and promotes viral transcription. DUX4 is functionally required for herpesvirus infection, since genetic depletion of DUX4 by CRISPR/Cas9 abrogates viral replication. Our results show that herpesviruses induce DUX4 expression and its downstream germline-specific genes and retroelements, thus mimicking an early embryonic-like transcriptional program that prevents epigenetic silencing of the viral genome and facilitates herpesviral gene expression.


Introduction
Herpesviruses are a major health burden worldwide, with a prevalence of 25 -100% dependent on the virus species and geography 1,2 .Herpesviruses cause a number of prevalent diseases, like oral and genital herpes, chickenpox, shingles, infectious mononucleosis and encephalitis.Life-threatening infections in healthy humans are rare, however severe disease is a signi cant risk in e.g.immunocompromised patients, transplant recipients or newborns 3,4 .Some herpesviruses such as Epstein-Barr virus (EBV) and Kaposi's sarcoma-associated herpesvirus (KSHV) are also known to be the causative agent of particular cancer types 5 .Due to their lifelong persistence in the host, the risk of reactivation and developing a lytic infection is constantly present.Treatment options are limited and therapies able to eliminate the virus from the body are unavailable.A better understanding of the mechanisms involved during herpesviral infection and persistence is therefore essential.
In a recent study, we observed that the upregulation of the cellular protein Tripartite-motif 43 (TRIM43) upon HSV-1 infection is dependent on the embryonic transcription factor DUX4 6 , which was also con rmed by Friedel et al. 7 .DUX4 is a germline transcription factor exclusively expressed during the 2cell to the 8-cell state of embryogenesis, a short phase during human embryonic development [8][9][10] .There, DUX4 is critical for zygotic genome activation (ZGA), a transcriptional activation event that leads to the induction of hundreds of target genes and allows the embryo to proceed further in development 11,12 .
Concomitantly, it induces retroelements such as LTRs and ERVs 9,10,13 which function as alternative promotors, regulate ZGA-related genes and mediate pluripotency 14,15 .After this limited period of activation, DUX4 is silenced throughout all adult tissues, except for spermatocytes 16 .An exception to this is the genetic disorder Facioscapulohumeral Muscular Dystrophy (FSHD).In FSHD, DUX4 silencing is lost and aberrant DUX4 expression in adult muscle cells leads to apoptosis and degeneration of the affected muscles 17,18 .The DUX4 gene resides within the macrosatellite repeat array D4Z4 19 , that harbors up to 100 copies of DUX4, with only the most distal copy encoding for a protein 17 .During early embryonic development, this macrosatellite array is transformed into heterochromatin after the 8-cell stage 8,10 and its repression is strictly controlled by chromatin remodeling factors and repressors which have not been fully investigated 16,19 .
Given the peculiar expression pattern of DUX4 and its upregulation in herpesviral infection, we aimed to determine whether DUX4 plays a role during herpesviral infection and to shed light on its potential functions.We show that DUX4 expression during herpesviral replication is a general mechanism that is conserved among herpesviral subfamilies and can also be observed in patient samples.For HSV-1, we could demonstrate that the viral proteins ICP0 and ICP4 induce DUX4 expression and hundreds of DUX4 target genes, including several classes of endogenous retroelements, and that DUX4 is essential for HSV-1 and HSV-2 replication.This points at so far unknown endogenous functions of DUX4 and extends our knowledge of this unique transcription factor.

Results
We con rmed the expression of DUX4 in various cell lines infected with HSV-1 by Western Blot and reanalysis of published RNA-seq.datasets (Fig. 1A-B) 6,20,21 .Comparison of three independent RNA-seq datasets in 3 different cell types showed no expression of DUX4 and DUX4 target genes (e.g.FRG1) in uninfected cells (Fig 1B).Upon infection with HSV-1, however, transcription is induced for DUX4 as well as known DUX4 target genes (Fig. 1B).Interestingly, we also con rmed the expression of HSAT-II satellite repeat transcripts.HSAT-II repeats are known DUX4 targets 10,22 and have recently been shown to play a role in HCMV replication 23 (Fig. S1a).Moreover, the reanalysis of a published Precision Run-On sequencing (PRO-seq.)dataset 24 showed no RNA-polymerase-II occupancy at the DUX4 locus in uninfected cells, indicating that the locus is completely silenced under normal conditions (Fig. 1B).In contrast, HSV-1 infection induced a robust PRO-seq signal at the DUX4 locus.DUX4 expression was also induced upon infection of primary human melanocytes with HSV-1, emphasizing the physiological relevance of DUX4 induction (Fig. S1b).Western Blot as well as qRT-PCR experiments further con rmed the expression of DUX4 and known DUX4-target genes TRIM48, TRIM49, ZSCAN4, ZSCAN5a, ZSCAN5d and RFPL4A upon HSV-1, HCMV and KSHV-infection (Fig. 1C-E, Fig. S2A-G) 25 in different experimental settings.This indicates that DUX4 is functional and acts as a transcriptional activator in herpesvirus infected cells.Reanalysis of published RNA-seq.datasets could not con rm expression of DUX4 in cells infected with other RNA-viruses or DNA-viruses like Adenoviruses or Poxviruses (Fig. S3A and data not shown).However, we could detect expression of DUX4 and DUX4 target genes in single cell RNA-seq.datasets from human Papillomavirus (HPV) positive head and neck cancer patients (Fig. 1F).By reanalyzing single cell RNA-Seq.-datafrom patients with EBV-pos.nasopharyngeal carcinoma, we could further demonstrate that DUX4 target gene expression was only detected in EBV-pos.tumor cells but not in healthy tissue from the same patients (Fig. 1G).Taken together, these data show that the induction of DUX4 expression is a common feature of human herpesviruses and papillomaviruses and of relevance in vivo.
Next, we wanted to address the signi cance of herpesviral DUX4 induction for the infection itself, since DUX4 is a germline transcription factor that is not expressed in healthy adult tissue 8,10 .We rst hypothesized that the induction of DUX4 might be part of an antiviral response of the cell to viral infection triggered by a herpesviral pathogen associated molecular pattern (PAMP), or by herpesviral induction of a cellular DNA-damage response (DDR).We reanalyzed several published RNA-sequencing datasets, investigating DUX4 mRNA expression in response to cellular changes, but found no evidence for DUX4 expression (data not shown).In addition, triggering a DNA-damage response by treating cells with Bleocin or Etoposide did not induce DUX4 expression (Fig. S3A).The only evidence for DUX4 expression came from a Chromatin-IP-sequencing (ChIP-Seq.)dataset 26 investigating the binding of the HSV-1 infected cell protein 4 (ICP4) to the cellular genome.ICP4 is a potent activator of viral transcription, part of the viral tegument, and essential for replication of HSV-1 27,28 .Our reanalysis of this dataset showed a very strong binding of the ICP4 protein to the DUX4 locus (Figure 1B).Of note, the peak intensity of ICP4 binding at the DUX4 locus was among the highest in the entire host genome.This ICP4 binding was a strong indication for an active induction of DUX4 by HSV-1 infection.Moreover, experiments with phosphonoacetic acid (PAA), an inhibitor of the viral DNA-polymerase, showed that herpesviral DNA replication is dispensable for DUX4 expression (Fig. S3B), indicating that induction of the DUX4 gene takes place at the immediate-early stage of the viral gene expression cascade.Even after infection with UV-inactivated virus, DUX4 protein was still induced, indicating that incoming components of the virion contribute to DUX4 induction (Fig. S3B).Analysis of the kinetics of DUX4 expression upon HSV-1 infection further demonstrated that DUX4 protein could be detected as early as 4h post infection with protein levels constantly increasing over the course of infection (Fig. 2A).Reanalysis of RNA-sequencing data from Rutkowski et al. 21further demonstrated that known DUX4 target genes have variable dynamics during infection with most DUX4 target genes being upregulated at 3-4 hours post infection (Fig. S4).To elucidate the involvement of HSV-1 tegument proteins in the induction of DUX4 expression we performed infection experiments with HSV-1 mutants depleted of the immediate early proteins ICP0, ICP4 and gamma-34.5.We could show that wildtype (wt) HSV-1 infection as well as infection with HSV-1-Δ γ34.5 resulted in DUX4 expression, whereas DUX4 expression is abrogated in cells infected with HSV-1 lacking either ICP0 or ICP4 (Fig. 2B and Fig. S5A).In addition, the transient co-expression of ICP0 and ICP4 is su cient to induce DUX4 expression in the absence of viral infection (Fig. 2C), whereas an E3-ligase de cient mutant of ICP0 (ICP0 FXE) was not able to induce expression of DUX4 when coexpressed together with ICP4.This con rms that ICP0 and ICP4 are necessary and su cient for inducing DUX4 expression.
DUX4 is a transcription factor active in the nucleus where it executes its physiological function by activating its target genes.We used a recombinant HSV-1 expressing ICP4-YFP and VP26-RFP in order to visualize progression of infection and localization of DUX4.ICP4 is an immediate-early gene and part of the tegument, whereas VP26, the small capsid protein of the virus, is expressed late in the infection cycle.By co-staining of cells with a DUX4 antibody, we could demonstrate that only ICP4+/VP26-cells (YFP+/RFP-cells) show nuclear expression of DUX4, indicating activation of DUX4 within early stages of infection (Figure 2E).In contrast, ICP4+ and VP26+ cells (YFP+ and RFP+ cells) show either no DUX4 expression or an aberrant cytoplasmic localization of DUX4.This indicates that DUX4 is only brie y activated by ICP0 and ICP4 during the early phase of infection.To test this, we treated cells with PAA, which arrests viral infection at the stage of viral DNA replication, i.e. after immediate-early gene expression but before late gene expression (Figure 2D and Fig. S5B).PAA treated cells showed a strong increase in expression of DUX4 and its target genes by qRT-PCR analysis of HDF cells infected with HSV-1 (Figure 2D), as well as the DUX4 protein by WB analysis in 293T cells (Figure S3B), supporting the notion of a transient activation of DUX4 during the early stages of HSV-1 infection.
It is known that the physiological function of DUX4 during embryonic genome activation at the 2-8 cell stage is to directly bind to DNA and activate genes and retroelements that are necessary for developmental progression 9,10 .In particular endogenous retroelements act as promoters for downstream genes, and binding of DUX4 to retroelements activates transcription of respective genes 14,29 .In order to analyze the role of DUX4 in herpesviral replication, we performed endogenous DUX4 ChIP-seq.as well as DUX4 CUT&Tag experiments.The DUX4-ChIP conditions optimized for DUX4 binding to the cellular genome resulted in a high background for the viral genome, most likely due to different physical properties that affect sonication and differences in chromatin accessibility.However, using CUT&Tag we could detect direct binding of DUX4 to the HSV-1 genome at about 4h post infection (Fig. 3A).We observed several DUX4 binding sites over the HSV-1 genome, with the biggest peak located in between the UL55 and UL56 genes and increased binding over time.Analysis of the binding sites from our endogenous DUX4 ChIP-Seq, CUT&Tag and comparison with the DUX4 ChIP-Seq performed by Young et al. 14 with overexpressed DUX4 protein in muscle cells showed almost identical consensus binding sites (Fig. 3B).A computational analysis found between 4-9 potential DUX4 consensus motifs in the genomes of HSV-1, HSV-2, HCMV, EBV and KSHV (Fig. 3C), suggesting that DUX4 possibly binds to viral genomes of all herpesviruses.Next, we established an electrophoretic mobility shift assay to con rm binding of DUX4 to regions of the HSV-1 genome (Fig. 3D).To this end, full-length DUX4 protein was expressed in E.coli (Fig. S6), puri ed and about 600bp-long fragments of the HSV-1 genome ampli ed by PCR.After coincubation of recombinant DUX4 with uorescently labeled PCR fragments we observed binding of DUX4 to a fragment containing one copy of the consensus DUX4 DNA-binding motif, whereas the exchange of T->C at position 9 of the binding motif completely abrogated binding (Fig. 3D).For the host genome, the ChIP-Seq and the CUT&Tag experiments revealed about 11000 DUX4 binding sites for ChIP-Seq and 3700 DUX4 binding sites for CUT&Tag within the host genome.Surprisingly, most DUX4 binding sites were not at transcriptional start sites or within exon regions of genes, but within intronic and intergenic regions (Fig. 3E).Analysis of intronic / intergenic binding sites showed a strong preference of DUX4 binding to repetitive genetic elements (Fig 3F).We found predominant binding of DUX4 to longterminal repeat (LTR-) elements and short interspersed nuclear elements (SINE) and only a small fraction of binding sites in actual genes (Fig. 3F).During early embryonic development and in particular ZGA, it is known that activation of endogenous retroelements can generate alternative promoters for expression of genes, resulting in transcription of developmental genes that are essential for further embryonic development 15,30 .Analysis of our RNA-seq and DUX4 ChIP-seq.data sets showed a DUX4-mediated activation of a speci c subset of LTR-elements that are expressed during ZGA (Fig. 4A).Both HSV-1 infection as well as DUX4 overexpression leads to an induction of the MLT-and THE-class of retroelements, indicating that the expression upon herpesviral infection is driven by DUX4 (Fig. 4A).In contrast, only HSV-1 infection induced the LTR12C class of retroelements.There expression was independent of DUX4 (Fig. 4A), Of note, the previously described C10rf159 antisense transcript induced upon infection 20 starts at an LTR12C element within the rst intron of its host gene and could thus be connected to retroelement activation.Moreover, a comparison of upregulated genes from HSV-1 infection, 8-cell stage of human development and FSHD patients showed that 843 host genes are signi cantly induced both during herpesviral infection and during the 8-cell stage of early human development (Fig. 4B) 31 .This indicates that herpesviral induction of the germline-speci c transcription factor DUX4 activates a transcriptional program of ZGA-speci c genes and retroelements.
Our data revealed a robust expression of DUX4 during lytic replication of all human herpesvirus subfamilies.We thus wanted to address the physiological consequences concerning herpesviral gene expression and replication.DUX4 is located within the 4q35 region of chromosome 4 32,33 and the gene is particularly complicated to target by CRISPR/Cas9.In healthy individuals the locus consists of 10-100 repeat units, and a copy of DUX4 resides in every unit.However, only one of the D4Z4 repeat arrays, the most distal one adjacent to the telomeres is encoding for the functional protein 17 .In addition to multiple repeats on chromosomes 4 (4q allele) there is also another heterochromatic repeat array on chromosome 10 (10q allele) 34 which is likely to interfere with CRISPR/CAS9 based knockout strategies 35 .We nally managed to generate knockout cells in the haploid cell line HAP1 (Fig. 5A).Upon infection of HAP1 DUX4 knockout (ko) cells, we could not detect a DUX4 speci c band in the Western Blot, which is present in the infected wildtype (wt) HAP1 cells.Interestingly, the expression of the viral proteins ICP0 and glycoprotein D (gD) was also strongly reduced in DUX4 ko cells compared to wt cells.In order to assess the effect of DUX4 knockout on the transcription of the entire viral genome, RNA-seq.from HSV-1 infected knockout and wildtype cells 8h post infection was performed.Comparison of host gene expression from DUX4-wt and -ko cells showed that DUX4 target genes are signi cantly downregulated in the ko cells, whereas transcription of other genes is not altered (Fig. 5B).Unsupervised clustering showed altered expression of most HSV-1 genes in ko cells compared to wt cells (Fig. 5C, Fig. S7).Whereas the expression of early and late genes is higher in wildtype cells at 8h post infection, the expression of immediate-early genes like UL54 or US1 is lower in the wildtype and higher in the knockout cells, indicating that DUX4 is required for the later stages of HSV-1 infection (Fig. 5C).Infection experiments with GFP-expressing HSV-1 and HSV-2 showed that the infection does not proceed in DUX4 ko cells compared to wt cells where the virus replicate to normal levels (Fig. 5D,E).In order to con rm the results from our DUX4 ko HAP1 cells, we also used a transient approach in 293T cells to knockout DUX4 at the population level, which resulted in a complete knockout of DUX4 for a short time period.Infection of DUX4 ko 293T cells resulted in an almost complete loss of most HSV-1 genes tested in Western Blot, like ICP0, ICP4, ICP27 and VP16 compared to wt 293T cells (Fig. 5F).This con rmed the results obtained in the HAP1 cell line and shows that DUX4 is critical for HSV-1 and HSV-2 replication.

Discussion
In humans the female oocyte is produced during female gametogenesis in the embryo and is then stored in prophase I of the meiosis for up to 50 years.Oocyte transcription is halted by epigenetic mechanisms, and stored mRNAs mostly control development 36 .After fertilization the zygote is formed, maternal to zygotic transition (MZT) takes place and with the onset of ZGA the zygotic genome starts to control transcription 36 .In order to induce transcription, the embryo has to overcome silencing of the genome, which is regulated by repressive epigenetic modi cations like DNA methylation, histone modi cations and a shortage of the cellular transcription machinery 36,37 .DUX4 has been shown to be important for ZGA by activating hundreds of genes that are necessary for further development, and several endogenous retroelements [8][9][10]13 . Altough the detailed function of most of DUX4 target genes remains elusive, it is thought that DUX4 induces several important factors that are involved in the creation of a permissive environment that allows transcription from the newly formed diploid genome.
Upon herpesviral infection the viral genome enters the nucleus and gets chromatinized, although the degree of chromatinization of the viral genome is discussed controversially, in particular for HSV-1.However, it is widely accepted in the eld that herpesvirus genomes are subjected to epigenetic silencing, and that they evolved strategies to prevent epigenetic silencing of their genome in order to allow transcription necessary for viral replication and virus transmission 38 .We show that herpesviruses from all human subfamilies as well as Papillomaviruses induce a robust expression of DUX4 upon lytic infection (Fig. 1A-E).Considering that alpha-, beta-, and gamma-herpesviruses were split into three separate lineages about 100-200 million years ago 39 , this hints at a highly conserved mechanism in the coevolution of herpesviruses with their respective hosts.For HSV-1, we demonstrate that DUX4 expression is induced by viral tegument/immediate-early proteins ICP0 and ICP4, indicating that this is an active induction by herpesviruses (Fig. 2C).ICP4 directly binds to the DUX4 locus, however, for full DUX4 induction the E3-ligase activity of ICP0 is needed (Fig. 2C), indicating that ICP0 degrades a cellular protein that mediates silencing of the DUX4 locus.ICP0 and ICP4 induce expression of DUX4 at early stages of infection for a brief period, indicated by nuclear staining of DUX4 in ICP4+/VP26-cells (Figure 2F).In ICP4+/VP26+ double positive cells we either observed no DUX4 staining or cytoplasmic DUX4 staining (Figure 2F).Blocking viral DNA replication with PAA results in higher levels of DUX4 and its target genes, suggesting DUX4 behaves similarly to the strong HSV-1 beta genes, which are upregulated early in infection and are shut down during the switch to viral DNA replication and late gene expression (Figure 2D).Interestingly, this very short induction of DUX4 resembles the mechanism of DUX4 induction during ZGA, where a short burst of DUX4 is su cient for target gene activation and reprogramming of the embryo but prevents toxicity mediated by DUX4.
Moreover, we could demonstrate that herpesviral DUX4 induction is essential for e cient herpesviral gene expression and replication.Depletion of DUX4 from cells results in a strong reduction of most herpesviral gene and protein expression, as shown for HSV-1 (Fig. 5A, C and F).In addition, we show that DUX4 can directly bind to the HSV-1 genome (Fig. 3A) and that HSV-1 as well as HSV-2 show drastically reduced viral replication in the absence of DUX4 (Fig. 5D and E).We hypothesize that herpesviruses evolved to partially mimic ZGA by actively inducing DUX4 expression.ZGA is conserved in all animals, and due to its signi cance in embryonic development there is very little room for the host to antagonize this viral mimicry of ZGA 36 .Any interference with DUX4 function, for example by mutations in the coding sequence or by preventing DUX4 expression could lead to drastic changes in the embryonic development that are incompatible with life.Thus, from a viral perspective, it is bene cial to exploit a host gene which is essential for development for its own purpose in order to limit mutations that affect viral replication.
In addition to its role in ZGA and in the development of FSHD, it was published that DUX4 also plays an important role in a variety of human cancers 40 .Reanalysis of almost 10.000 cancer transcriptomes from The Cancer Genome Atlas (TCGA) showed DUX4 re-expression in many human cancers 40 .The authors speculate that expression of DUX4 and DUX4 target genes may contribute to tumorigenesis 40 .Interestingly, the human gamma-herpesviruses, Epstein-Barr Virus (EBV) and KSHV are classi ed as human carcinogens by the WHO and cause cancer in humans 5 .We observe expression of DUX4 target genes in EBV-positive nasopharynx carcinoma cells, but not in healthy tissue from the same donors (Fig. 1G).The same holds true for patient cells of HPV-positive cancer (Fig. 1F).This indicates that the observed DUX4 induction is also of relevance in vivo.It is worth speculating but beyond the scope of this manuscript to investigate whether DUX4 expression induced by EBV, KSHV and HPVs also contributes to viral oncogenesis.In addition, Chew et al. showed that DUX4 expression results in a downregulation of MHC-Class I expression by interfering with STAT1 signaling, hinting at an immune evasion mechanism 40,41 that might also be important for viral infections.
We demonstrate that herpesviral DUX4 expression also leads to expression of several endogenous retroelements, including LTR-containing retrotransposons, LINE-1 elements and Alu-elements.Whereas cell type speci c differences in the induction of retroelements exist, with more retroelements being transcribed in tumor cell lines than in primary cells, we identi ed an Alu-element 5' of the ZSCAN4 gene that serves as a binding site for DUX4 and drives ZSCAN4 expression in all cell lines tested.In addition it is known for years that herpesviral infection leads to the induction of endogenous HERV-W retroviruses and also to the reactivation of the HIV-LTR, hinting at a role of herpesviral DUX4 in respective processes.It is hypothesized that some herpesviruses including HSV-1 have evolved a high GC-content in their genome in order to prevent insertion of endogenous retroelements that have a bias for AT-rich sequences.The active induction of DUX4 by HSV-1 proteins and the importance of DUX4 for herpesviral gene expression / replication supports this possible explanation for the high GC-content of HSV-1 genomes.It may help to preserve herpesviral genome integrity despite the DUX4-mediated induction of endogenous retroelements that could integrate into the viral genome.
Our data points at a very important if not essential role for DUX4 in herpesviral gene expression and replication.We could demonstrate that depletion of DUX4 in cells impairs viral replication.As such, it is tempting to speculate that DUX4 and DUX4 downstream genes could be targeted for therapy of herpesvirus-associated diseases.Blocking DUX4 means preventing herpesviral gene expression and subsequent viral replication from the beginning.Targeting a cellular protein has the major advantage that viral escape mutants and resistance-formation are unlikely, but usually this comes with the risk of possible side-effects.However, DUX4 is not expressed in adult somatic tissue, therefore side-effects should be negligible and DUX4 might serve as an attractive target for anti-herpesviral therapy.

Declarations
Reagents, plasmids and transfections Transfections were performed with GenJet (SignaGen Laboratories) or Lipofectamin 2000 (Thermo Fisher Scienti c) according to the manufacturer's protocol.Plasmids are listed in Table 3.
Samples were diluted with Laemmli-SDS sample buffer and heated for 5 min at 95°C.Antibodies used are listed in Table 4.
qRT-PCR RNA was extracted using the Direct-zol RNA Miniprep Plus kit from Zymo Research according to manufacturer's instructions.Reverse transcription was performed using the Super Script IV Kit (Thermo Fisher Scienti c) according to manufacturer's instructions.qRT-PCR was carried out using TaqMan™ Universal PCR Master Mix I (Applied Biosystems, Thermo Fisher Scienti c) 0,1 µg as template on a 7500 Fast Real-Time PCR machine.Or reverse transcription and qRT-PCR were conducted in one step using 0.1 µg RNA as template with the Luna Universal Probe One-Step RT-qPCR kit (New England Biolabs) following manufacturer's protocols.Primers/probes are listed in Table 1.Expression levels for each gene were obtained by normalizing values to HPRT1 or VTRNA and fold induction was calculated using the comparative CT method (ΔΔCT method).

CRISPR and sgRNAs
All sgRNAs used in this study were previously described and are listed in Table 2. sgRNAs were cloned into LentiCRISPRv2 plasmid gifted from F. Zhang (Addgene plasmid #52961) and veri ed by sequencing.
Lentiviruses were packaged with pMD2.G (Addgene plasmid #12259) and psPAX2 (Addgene plasmid #12260) (both gifted from D. Trono) into HEK 293T.HEK 293T were seeded into 12 well plates and lentiviral supernatants added at 70-80% con uency.Plates were centrifuged at 1200 rpm for 2 min after centrifugation culture medium was added and cells incubated over night at 37°C.The medium was changed to normal culture medium the next day and selection with 2 µg/ml puromycin in normal culture medium started on day 3.

DUX4 protein puri cation
The DUX4 protein was puri ed using the Intein Mediated Puri cation with an A nity Chitin-binding Tag (IMPACT) system (NEW ENGLAND BioLabs INC).The coding sequence of DUX4 (Addgene: Plasmid #21156) was subcloned into the pTYB12 vector using EcoR1 and Sap1 restriction sites, with an intein-CBD tag added to the N-terminus of DUX4.Protein expression was induced by adding 0.4 mM of IPTG to ER2566 cells at an OD600 = 0.5 overnight at 18°C.Bacterial pellets were then resuspended in Lysis Buffer (20 mM Na-HEPES, 500 mM NaCl, 1 mM EDTA, 0.1% Triton X-100 and protease inhibitors (cOmplete™ Proteasehemmer-Cocktail)) and lysed a French Press.The lysates were centrifuged at 15000 g at 4°C for 30 min, and the clari ed lysate was slowly loaded onto the chitin column for puri cation using the ӒKTA pure™ chromatography system.The beads were then washed with 50 bed volumes of Column Buffer (20 mM Na-HEPES, 500 mM NaCl, 1 mM EDTA, and 0.1% Triton X-100) before protein cleavage with washing buffer containing 50 mM DTT.Finally, the eluate was further puri ed with Size Exclusion Chromatography using a Superdex 200 Increase 10/300 GL column (Cytiva).

Next generation Chip-Seq
For CHIPmentation primary HFF cells were seeded in T175 asks infected with HSV-1 KOS ( MOI 10)   for different time points.CHIPmentation was conducted as previously described 44 .Cells were sonicated using the Bioruptor (Diagenode) for 30 cycles.For the immunoprecipitation protein G Dynabeads (Thermo Fisher Scienti c) were used.Samples were incubated with either 2.5 µg Anti-DUX4 (E5-5) (Abcam) or Normal Rabbit IgG (Cell Signaling Technology) as control.Samples were puri ed using AMpureXP beads (Beckman Coulter) according to manufacturer's description.Libraries were sequenced on HiSeq 4000 System (Illumina) Next generation RNA-Seq HEK 293T wildtype cells and HEK 293T CRISPR/Cas knockout cells were seeded in T25 asks.Cells infected for different time points with HSV-1 KOS (MOI 10).HAP1 and HAP1-DUX4 ko cells were Cells were lysed in Trizol (Life Technologies by Thermo Fisher Scienti c), and total RNA was isolated using the RNA clean and concentrator kit (Zymo Research), according to the manufacturer's instructions.
Sequencing libraries were prepared using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (NEB) with 9 cycles PCR ampli cation, and sequenced on a HiSeq 4000 device with 1x50 cycles.For quanti cation of viral gene expression alignments were done using hisat2 49 on the HSV-1 genome (strain 17, genbank accession no.NC_001806) and readcounts per gene quanti ed using quasR 51 .

CUT&Tag
Primary HFF cells were infected with HSV-1 GFP (MOI 1) in PBS supplemented with 0.1% Glucose and 1% FCS. the remaining virus was washed away with a low pH buffer (40mM Citric acid, 10mM KCl, 135mM NaCl, pH3) after one hour at 37°C.CUT&Tag was performed according to the manufacturing protocol (CUT&Tag-IT Assay Kit, Active Motif) using 2,5 µg Dux4 E5-5 (Abcam) or 2,5 µg rabbit IgG (Cell Signalling Technology).In short, cells were bound to concanavalin A-coated magnetic beads and permeabilized for subsequent incubation with primary and secondary antibodies.Speci c cutting and addition of adapters was mediated by a protein A-Tn5 transposase fusion protein.Libraries were sequenced on HiSeq 4000 System (Illumina) EMSA Viral DNA of HSV-1 GFP and HSV-1 KOS was isolated using the Quick-DNA MiniPrep from Zymo Research.Two short regions containing either the DUX4 binding motif or no DUX4 binding motif were ampli ed by PCR: about 600bp long fragments were ampli ed from HSV-1 genomes using the primers for Binding site 1 (Dux4 motif) BS1-fwd:gtgtaccactgctgtcg, BS1-rev:gtctgatcatgccccatacc, and Binding site 2 (No Dux4 motif) BS2-fwd:cgtgaaccaaagacgagggc, and BS2-rev:ccacgttgagaagctcgtcg.The EMSA was performed as described by Lee et al. 45 using 50ng of ampli ed viral DNA, which was incubated 30 min at room temperature with 1µg puri ed DUX4 protein and 2µL sperm DNA (D7656, Sigma-Aldrich).
Flow cytometry HAP1 wt and ko cells were simultaneously seeded and infected with HSV-1 GFP or HSV-2 GFP at a MOI of 0.1.At day 1, 2, 3 and 4 post infection, cells were detached by scraping, xed in 1% PFA for 20 min and resuspended in FACS buffer (PBS supplemented with 2% FCS and 0.5mM EDTA).GFP expression was measured with BD LSRFortessa and the data was analyzed with FlowJo.

Bioinformatic analysis
ChIP-seq processing was done using the PiGx-ChIP-seq pipeline (https://doi.org/10.1093/gigascience/giy123). In short, adapters and low quality bases were trimmed from reads using Trim-galore.The reads were mapped on the hg19 version of the human genome, combined with HSV-1 genome, using Bowtie2 with k = 1 parameter.bigWig tracks were created by extending reads to 200, collapsing them into pileups, and normalizing to reads per million.Peak calling was done with MACS2 (https://github.com/taoliu/MACS)using the default parameters.Motif discovery was done using MEME 47 with the default parameters, on the top 100 peaks (sorted by q value), in a region of +/-50bp around the peak center.Peak annotation was done using the hg 19 ENSEMBL GTF le, downloaded on 17.03.2017.from the ENSEMBL database 48 .Peaks were annotated based on the following hierarchy of functional categories: tss -> exon -> intron -> intergenic (eg.if a peak overlapped multiple categories, it was annotated by the class that is highest in the hierarchy).Peaks were overlapped with the hg19 Repeatmasker repeat annotation, downloaded from the UCSC database on 03.02.2015.
For CUT&Tag analysis, the data was mapped to the hg19 version of the human genome with the viral genome using Bowtie2 with k = 1 parameter.bigWig tracks were created by extending reads to 200, collapsing them into pileups, and normalizing to reads per million.Peak calling was executed by MACS2 (https://github.com/taoliu/MACS)using the default parameters.The obtained peaks were ltered for being present in two consecutive time points and not present in the 2h time point.Motif discovery was Statistical analysis P values were calculated using an unpaired Student's test.P <0.05 was considered statistically signi cant.Tables Table 1 Oligonucleotides Real time quantitative PCR probes DUX4L: 5'-TCAGCCAGAATTTCACGGAAG-3'

Figures Figure 1 A
Figures