Tripartite motif (TRIM) proteins mediate antiviral host defences by either directly targeting viral components or modulating innate immune responses. Here we identify a mechanism of antiviral restriction in which a TRIM E3 ligase controls viral replication by regulating the structure of host cell centrosomes and thereby nuclear lamina integrity. Through RNAi screening we identified several TRIM proteins, including TRIM43, that control the reactivation of Kaposi’s sarcoma-associated herpesvirus. TRIM43 was distinguished by its ability to restrict a broad range of herpesviruses and its profound upregulation during herpesvirus infection as part of a germline-specific transcriptional program mediated by the transcription factor DUX4. TRIM43 ubiquitinates the centrosomal protein pericentrin, thereby targeting it for proteasomal degradation, which subsequently leads to alterations of the nuclear lamina that repress active viral chromatin states. Our study identifies a role of the TRIM43–pericentrin–lamin axis in intrinsic immunity, which may be targeted for therapeutic intervention against herpesviral infections.
Infection by viruses of the family Herpesviridae is the cause of significant morbidity and mortality in humans worldwide, especially in immunocompromised individuals1. Furthermore, two of the eight human herpesviruses, Kaposi’s sarcoma-associated herpesvirus (KSHV) and Epstein–Barr virus (EBV), are human carcinogens and cause a range of malignancies of lymphoproliferative and endothelial origin2,3. Despite differences in pathogenesis and carcinogenicity, herpesviruses share common features in their replication strategies, most notably in the switch from a state of persistent latent infection with minimal viral gene expression to a state of lytic infection characterized by productive viral replication. While antiviral therapies targeting the lytic phase exist, a major obstacle to herpesvirus eradication is the persistence of these viruses in latently infected cells. As such, molecular insight into the factors that modulate the latent-to-lytic transition of herpesviral replication is of fundamental importance. Although major progress has been made towards identifying the viral factors that regulate the latent-to-lytic switch, our knowledge of host intrinsic factors that modulate the transition is still rudimentary.
Tripartite motif (TRIM) proteins have been increasingly appreciated as important antiviral factors that suppress the replication of a wide range of RNA viruses and DNA viruses. TRIM proteins inhibit viral replication by either directly targeting viral components or modulating the innate immune responses that result in antiviral gene expression4,5. For example, TRIM19 (also known as promyelocytic leukaemia protein, PML) restricts multiple RNA viruses and DNA viruses, including herpesviruses, by organizing PML nuclear bodies6,7. We postulated that hitherto unidentified TRIM proteins play critical roles in restricting herpesvirus lytic replication or reactivation.
Here, we have identified several TRIM proteins that suppress KSHV reactivation. Among the TRIM proteins identified, TRIM43 was distinguished by its ability to restrict a broad range of herpesviruses. TRIM43 targets the centrosomal protein pericentrin for degradation, which leads to the loss of nuclear envelope integrity and changes in viral chromatin that repress lytic replication.
Identification of TRIM43 as a herpesvirus-specific antiviral protein
To identify TRIM proteins that modulate herpesvirus reactivation, we performed an RNAi screen targeting 62 human TRIM genes to test the effect of TRIM depletion on KSHV reactivation (Fig. 1a,b). Individual TRIM genes were silenced by lentiviral transduction of a pool of three different shRNAs per TRIM in HEK 293 cells latently infected with recombinant KSHV strain 219 (293-rKSHV.219), which expresses red fluorescent protein (RFP) under the control of a lytic viral promoter, allowing high-throughput detection of viral reactivation by monitoring RFP expression8. Depletion of 15 TRIM proteins showed a significant increase in RFP expression compared to cells transduced with control non-targeting shRNA (sh.C), suggesting that these TRIM proteins suppress KSHV reactivation and/or lytic replication (Fig. 1b). In contrast, the depletion of other TRIM proteins had minor or no effects on KSHV reactivation (Fig. 1b).
A hallmark of many antiviral factors is inducible gene expression in response to the virus(es) they restrict9. Therefore, we next examined whether the expression of the 15 candidate TRIM proteins identified in our screen was inducible by KSHV reactivation (Fig. 1c). This showed that the transcript levels of TRIM43 were profoundly (>4,000-fold) upregulated following KSHV reactivation; in comparison, the mRNA levels of the other candidate TRIM proteins were much less, or not, induced by KSHV reactivation (Fig. 1c). TRIM43 mRNA expression was also strongly induced following reactivation of EBV, a gamma-herpesvirus closely related to KSHV, or following de novo infection by human cytomegalovirus (HCMV; of the Betaherpesvirinae subfamily) or herpes simplex virus type 1 (HSV-1; of the Alphaherpesvirinae subfamily) (Fig. 1d–f and Supplementary Fig. 1a). In contrast, infection with adenovirus (Ad), a dsDNA virus unrelated to herpesviruses, or with the RNA viruses vesicular stomatitis virus (VSV) or dengue virus (DV), did not induce TRIM43 mRNA expression (Fig. 1f and Supplementary Fig. 1b).
We next asked whether TRIM43 is also upregulated in herpesvirus-associated diseases in humans. TRIM43 mRNA was significantly upregulated in bronchoalveolar lavage (BAL) human patient samples that were positive for HSV-1 as compared to HSV-1-negative BAL samples (Fig. 1g). Investigation of TRIM43 protein expression in Kaposi’s sarcoma (KS) tissues derived from AIDS patients showed that TRIM43 was moderately or highly expressed in 10 of 17 cases analysed, whereas TRIM43 expression was minimal in all healthy skin tissues (Supplementary Fig. 1c). Taken together, these results show that TRIM43 gene expression is robustly induced following herpesviral infection, which prompted us to investigate whether TRIM43 has anti-herpesviral activity.
To test this, we first asked whether TRIM43 knockdown induces productive KSHV replication in cells that are usually refractory to efficient KSHV lytic replication, such as HEK 293T cells. Knockdown of endogenous TRIM43 using an shRNA with a different targeting sequence from the shRNAs used in our screen led to efficient KSHV lytic gene expression and production of infectious virions in infected HEK 293T cells (Supplementary Fig. 2a,b). Depletion of TRIM43 also enhanced KSHV lytic replication in SLK cells harbouring recombinant KSHV.219 (iSLK.219) (Supplementary Fig. 2c). To determine if TRIM43 specifically inhibits herpesviral infection, we next examined the effect of silencing TRIM43 on the replication of HSV-1, Ad, VSV and encephalomyocarditis virus (EMCV). We also tested six additional TRIM proteins (TRIM42, TRIM45, TRIM46, TRIM47, TRIM65 and TRIM66) that had demonstrated pronounced effects on KSHV reactivation in our shRNA screen (Fig. 1b) as well as TRIM25, which is known to inhibit the replication of several viruses and thus served as a positive control. siRNA-mediated depletion of endogenous TRIM43 led to a significant increase in HSV-1 replication. In contrast, knocking down TRIM43 had no or a marginal effect on the replication of Ad, VSV or EMCV (Fig. 1h and Supplementary Fig. 2d-i), whereas depletion of the other seven TRIM proteins affected the replication of multiple viruses, and each TRIM possessed a unique restriction profile (Fig. 1h), strengthening the conclusion that TRIM43 specifically restricts herpesviruses. In further support of this hypothesis, silencing of endogenous TRIM43 also induced the spontaneous reactivation of EBV from latently infected cells and enhanced lytic replication of HCMV, as indicated by the enhanced expression of the immediate–early/early genes, BMRF1 and Zta, or immediate–early-1 (IE-1), respectively (Fig. 1i–k and Supplementary Fig. 2j). Finally, we asked the inverse question of whether ectopic expression of TRIM43 suppresses herpesviral infection, and whether this activity is dependent on its RING E3 ligase activity, which is known to play an important role in the antiviral activity of other TRIM proteins5,10,11. Ectopic expression of wild-type (WT) TRIM43 suppressed HSV-1 replication, while expression of a mutant of TRIM43 that has the RING-finger domain deleted (TRIM43ΔRING) did not have an inhibitory effect on HSV-1 replication (Supplementary Fig. 2k). Collectively, these results demonstrate that TRIM43 suppresses the reactivation or lytic replication of multiple herpesviruses, and further suggest that TRIM43 antiviral activity is dependent on its RING E3 ligase activity.
TRIM43 is induced by DUX4 during herpesvirus infection as part of a germline-specific transcriptional program
As TRIM43 is a completely uncharacterized protein, we first tested its expression pattern in different human tissues (Supplementary Fig. 3a). We found ubiquitous but very low mRNA expression of TRIM43 in all human tissues tested, with the exception of the brain, thymus and testis, which showed higher expression. We next used laser-scanning confocal microscopy to assess the localization of TRIM43 protein in primary HFF cells that were either left uninfected, or infected with HSV-1 (Fig. 2a). We observed extremely low expression of endogenous TRIM43 protein in uninfected cells; TRIM43 was detectable only at high magnification, where the protein appeared as one or two small puncta per cell. However, TRIM43 protein expression was profoundly upregulated in HSV-1-infected cells, showing punctate expression throughout the cytoplasm (Fig. 2a). Ultraviolet-inactivated HSV-1 did not induce TRIM43 expression, suggesting that TRIM43 upregulation is dependent on active viral replication (Fig. 2a).
To provide insight into the mechanism of TRIM43 transcriptional upregulation, we first asked whether TRIM43 gene expression is type I interferon (IFN)-inducible, as is the case for many other TRIM proteins4,12. Treatment with IFN-α or infection by Sendai virus (SeV), known to robustly trigger IFN induction, failed to enhance TRIM43 transcript levels, while the mRNA levels of IFN-stimulated gene 15 (ISG15) were strongly upregulated (Supplementary Fig. 3b). TRIM43 was also robustly upregulated by HSV-1 infection in Vero cells (Supplementary Fig. 3c), which are known to be defective in IFN-α/β signalling13, further supporting that herpesvirus-induced TRIM43 expression is independent of type I IFNs.
Previous reports indicated that TRIM43 is efficiently expressed during early embryogenesis but minimally expressed in adult tissues14,15,16,17. Moreover, TRIM43 gene expression was found to be robustly upregulated in cells from patients with facioscapulohumeral muscle dystrophy (FSHD); in this disease, TRIM43 was shown to be a target gene of the germline transcription factor double homeobox 4 (DUX4), whose aberrant expression causes FSHD18,19,20. Therefore, we asked whether DUX4 gene expression is induced by herpesvirus infection, and whether TRIM43 expression is mediated by DUX4 in herpesvirus-infected cells. DUX4 transcripts were profoundly upregulated over the course of HSV-1 infection, which correlated closely with the induction of TRIM43 mRNA (Fig. 2b). DUX4 protein was also readily detectable in HSV-1-infected cells, where it localized to the cell nucleus, but was not detectable in uninfected cells (Fig. 2c and Supplementary Fig. 3d). In a pattern similar to that of herpesvirus-specific induction of TRIM43 expression (Fig. 1c–g), DUX4 was highly expressed only during herpesvirus infection, whereas its expression was minimal during infection with VSV or DV (Fig. 2d). To examine if DUX4 is sufficient for the induction of TRIM43, we ectopically expressed DUX4 in HEK 293T cells, which do not detectably express DUX4 under normal (uninfected) conditions, and found that ectopic expression of DUX4 robustly induced TRIM43 transcripts (Fig. 2e). Next, we depleted endogenous DUX4 in HSV-1-infected cells using specific siRNA (Fig. 2f). TRIM43 mRNA expression was highly upregulated by HSV-1 infection in cells transfected with non-targeting control siRNA (si.C), whereas knockdown of DUX4 markedly reduced TRIM43 mRNA levels in infected cells. These results indicate that DUX4 is necessary and sufficient to drive TRIM43 gene expression.
Our data show that herpesvirus infection upregulates TRIM43 gene expression in a manner that is dependent on DUX4, which is a germline transcription factor that is usually only expressed during early development, or in FSHD patients, which carry genetic modifications of the DUX4 gene locus19,21. Comparative global transcriptome analysis of HEK 293T cells transfected with DUX4 or infected with HSV-1 showed that the transcript levels of TRIM43 and also several other known DUX4-target genes (for example, members of the PRAMEF and ZSCAN protein families15,16,17,18,20) are highly upregulated by HSV-1 infection or DUX4 expression (Fig. 2g and Supplementary Fig. 4a,b). Quantitative real-time PCR analysis confirmed that these DUX4-target genes are also markedly upregulated following KSHV reactivation (Supplementary Fig. 4c). Together, these results indicate that herpesvirus infection triggers TRIM43 gene expression as part of a DUX4-dependent germline-specific transcriptional program.
TRIM43 regulates the stability of pericentrin and thereby centrosomal integrity
To elucidate the mechanism behind TRIM43-mediated herpesvirus restriction, we began by characterizing the localization of TRIM43, which under normal (uninfected) conditions appears as either one or two puncta per cell (Fig. 2a). Confocal microscopy analysis revealed that the TRIM43 puncta co-stained with the centrosomal marker γ-tubulin, suggesting that TRIM43 localizes to the centrosome (Fig. 3a). Of note, siRNA-mediated depletion of TRIM43 abolished its centrosomal staining pattern, ruling out unspecific staining of centrosomes by the polyclonal TRIM43 antibody (Supplementary Fig. 5a). Electron microscopy analysis confirmed that TRIM43 localization was restricted to the centrosome in uninfected cells, where it localized to the pericentrosomal material (Fig. 3b and Supplementary Fig. 5b). Furthermore, we observed that ectopic expression of TRIM43, but not TRIM25 (which has a similar domain structure as TRIM43 but does not localize to centrosomes11), led to a near-complete loss of centrosomal γ-tubulin staining (Fig. 3c,d). Ectopic expression of TRIM43 also led to a loss of staining for the centrosomal proteins centrobin and Sas-6 (Supplementary Fig. 5c-e), which prompted us to hypothesize that TRIM43 controls herpesviral reactivation/lytic replication by modulating the structural integrity and/or function of the centrosome.
Indeed, mass spectrometry (MS) analysis of affinity-purified TRIM43ΔRING showed that it interacted with several proteins that are important for centrosome function or structural integrity: the key structural centrosomal proteins pericentrin (PCNT) and γ-tubulin22,23, as well as the cohesin subunits SMC1, SMC3 and RAD21, which are involved in centrosome disengagement24 (Supplementary Fig. 5f). Importantly, we used in this experiment the E3 ligase-defective TRIM43ΔRING mutant because we postulated that this would allow us to purify putative TRIM43 substrates that would otherwise have been degraded.
We next asked whether TRIM43 E3 ligase activity triggers the degradation of the centrosomal proteins identified by our MS analysis. We found that ectopic expression of FLAG-TRIM43 WT effectively induced the degradation of endogenous PCNT (Fig. 4a,b and Supplementary Fig. 6a) but did not trigger the degradation of γ-tubulin, SMC1, SMC3 or RAD21 (Fig. 4a). Treatment with the proteasome inhibitor MG132 rescued TRIM43-induced degradation of PCNT (Fig. 4a,b and Supplementary Fig. 6a). In contrast to WT TRIM43, ectopic expression of the catalytically inactive TRIM43ΔRING mutant did not induce PCNT degradation, but instead led to an increase in PCNT protein abundance (Fig. 4a and Supplementary Fig. 6a), suggesting a dominant negative effect of TRIM43ΔRING. Kinetics analysis of PCNT protein abundance in HSV-1-infected or KSHV-reactivated cells showed a gradual loss of PCNT protein, which closely correlated with increased TRIM43 protein expression (Supplementary Figs. 6b and 7a). In contrast, VSV or Ad infection did not lead to changes in PCNT protein abundance (Supplementary Fig. 7b,c). These data suggested that TRIM43 interacts with PCNT at the centrosome and regulates its protein stability during herpesvirus infection.
In support of an interaction between TRIM43 and PCNT, co-immunoprecipitation (co-IP) experiments showed that endogenous PCNT interacted with FLAG-TRIM43, but not FLAG-TRIM25 (Fig. 4c). Ground state depletion combined with total internal reflection fluorescence (GSD-TIRF) super-resolution microscopy indicated that endogenous TRIM43 and PCNT form a complex at the centrosome (Fig. 4d). We next tested whether TRIM43 mediates PCNT degradation through direct ubiquitination by examining the effect of ectopically expressed TRIM43 on PCNT ubiquitination by IB. TRIM43, but not TRIM25, expression strongly enhanced PCNT polyubiquitination (Fig. 4e). Conversely, whereas endogenous PCNT ubiquitination was increased following HSV-1 infection, silencing of TRIM43 diminished PCNT ubiquitination in infected cells (Fig. 4f). Depletion of endogenous TRIM43 led to a strong increase in the abundance (both signal intensity and numbers) of PCNT puncta, further supporting that TRIM43 regulates PCNT protein stability via degradative ubiquitination (Fig. 4g,h). These data indicate that TRIM43 regulates the structural integrity of centrosomes by interacting with PCNT and inducing its proteasomal degradation.
Herpesvirus restriction by TRIM43 is dependent on PCNT degradation
To address the question whether the ability of TRIM43 to degrade PCNT is responsible for the antiviral activity of TRIM43, we used siRNA to deplete PCNT, thus mimicking the effect of TRIM43-mediated PCNT degradation. In parallel, we also silenced TRIM43, which leads to higher stability of PCNT and is therefore expected to have the opposite effect on virus replication. Knockdown of PCNT significantly decreased (by ~50%) the mRNA and protein expression of HSV-1 infected cell protein 0 (ICP-0), indicating repressed lytic replication (Fig. 5a,b), while it had no effect on Ad or VSV replication (Supplementary Fig. 7d,e). In contrast, silencing TRIM43 markedly enhanced ICP-0 mRNA and protein abundance as well as HSV-1 titres (Fig. 5a–c). Importantly, ectopic expression of a siRNA-resistant TRIM43 (TRIM43RNAi-res) construct in cells depleted of TRIM43 reversed the effect of TRIM43 knockdown on HSV-1 replication, ruling out off-target effects of the siRNA (Fig. 5d,e). Moreover, while knockdown of TRIM43 enhanced HSV-1 replication and KSHV reactivation, double-knockdown of TRIM43 and PCNT reversed this effect, strongly suggesting that the antiviral activity of TRIM43 is mediated by its effect on PCNT (Fig. 5c,f). Of note, ectopic expression of TRIM43 did not lead to changes in the cell cycle, ruling out that the effect of TRIM43 on PCNT leads to defective mitosis and thereby affects herpesvirus infection (Supplementary Fig. 8a). These results indicate that TRIM43-mediated restriction of herpesvirus replication is dependent on its ability to degrade PCNT.
PCNT degradation by TRIM43 induces nuclear lamina alterations
We next sought to determine the molecular mechanism by which TRIM43-mediated PCNT degradation affects the reactivation/lytic replication of herpesviruses, which replicate in the host cell nucleus. It is well established that the centrosome is associated with the nuclear envelope by microtubules and LINC (linker of nucleoskeleton and cytoskeleton) proteins25,26. Centrosomal proteins such as PCNT can thus influence nuclear positioning and the architecture of the nuclear envelope27,28. As the nuclear lamina has been shown to influence not only cellular chromatin but also viral chromatin, we postulated that TRIM43, through its regulation of PCNT stability, induces nuclear lamina alterations that may affect the transcriptional activity of viral chromatin and thereby modulate viral reactivation/lytic replication. Ectopic expression of TRIM43, which induces PCNT degradation, led to a range of morphology changes of the nuclear lamina, such as lamina invaginations and ruffling, and even complete breakdown of the nuclear lamina when TRIM43 was highly expressed (Fig. 6a,b and Supplementary Fig. 8b,c). In contrast, TRIM25 expression had no effect on nuclear envelope integrity, whereas TRIM43ΔRING expression induced ‘thickening’ of the nuclear lamina (Supplementary Fig. 8b), suggesting that the nuclear lamina alterations induced by TRIM43 are dependent on its E3 ligase activity. Notably, however, the observed nuclear lamina changes were not mediated by TRIM43-induced proteasomal degradation of lamin A/C or lamin B, as the overall abundance of lamin proteins remained unchanged following TRIM43 overexpression (Supplementary Fig. 8c,d). In concordance with our hypothesis that TRIM43-mediated PCNT degradation induces nuclear lamina alterations, depleting endogenous PCNT resulted in very similar lamina changes (for example, lamina invaginations/ruffling, doughnut-shaped nuclei and extranuclear lamin) (Supplementary Fig. 9a-c), which is consistent with published results27 and further strengthens that the effect of TRIM43 on nuclear lamina integrity is dependent on PCNT. Along these lines, HSV-1 infection led to the loss of PCNT, as previously reported29, as well as nuclear lamina alterations (Fig. 6c) similar to those observed in cells that ectopically express TRIM43 (Fig. 6a). In contrast, VSV or Ad infection, which did not trigger TRIM43 upregulation or PCNT degradation (Supplementary Fig. 7b,c), did not induce nuclear lamina changes (Supplementary Fig. 10a). Next, we asked whether the nuclear lamina alterations observed in HSV-1-infected cells are mediated by TRIM43. Strikingly, HSV-1-induced alterations of lamina morphology were reversed by knocking down endogenous TRIM43 (Fig. 6c). Consistent with this, TRIM43 depletion also rescued the degradation of endogenous PCNT in HSV-1-infected cells (Fig. 6c). Together, these data indicate that herpesvirus-induced TRIM43 expression leads to degradation of PCNT and the loss of centrosomal integrity, ultimately inducing alterations of the nuclear lamina.
TRIM43-induced nuclear lamina alterations suppress transcriptionally active viral chromatin states and thereby inhibit herpesvirus replication
Lamina-associated domains typically serve as anchoring platforms for cellular heterochromatin. In the case of viral chromatin, however, lamin A/C is required for targeting viral genomic DNA to the periphery of the cell nucleus30. As such, in Lmna−/− cells, viral replication compartments are reduced in size, which results in diminished virion production due to increased viral heterochromatin and transcriptional silencing of viral genes30. We next asked whether TRIM43-mediated nuclear lamina changes influence viral chromatin and thereby viral lytic replication/reactivation. To test this hypothesis, we adapted the recently developed CRISPRainbow technology31 to visualize the chromatin of HSV-1 genomes in cells that were transfected with either non-targeting control siRNA or TRIM43-specific siRNA (Fig. 6d,e). Briefly, we designed a guide RNA (gRNA) with sequence specificity for HSV-1 repeat DNA that is fused to MS2-RNA-binding sites. Co-transfection of this gRNA along with a construct encoding a catalytically defective mutant of CRISPR-associated protein 9 (dCas9) and a construct encoding MS2-RNA-binding domains fused to blue fluorescent protein (BFP) allows for recruitment of dCas9 and BFP to HSV-1 DNA and visualization of viral genomic DNA. Importantly, BFP foci co-localized extensively with ICP-8, an HSV-1-encoded single-stranded DNA binding protein that binds to replicating viral genomes, which confirmed that our system allows for specific visualization of viral genomes and not cellular chromatin (Supplementary Fig. 10b). Compared to HSV-1 genomes in cells transfected with non-targeting siRNA, the co-localization of HSV-1 genomes with the euchromatic marker histone H3 trimethyl lysine 4 (H3K4me3) was markedly enhanced in cells depleted of endogenous TRIM43 (Fig. 6d,e). Cells depleted of TRIM43 also showed a significantly enhanced association of the euchromatic marker H3K4me3 with promoters of the viral genes VP16 and thymidine kinase (TK) as compared to cells transfected with control siRNA (Fig. 6f). Collectively, these data demonstrate that TRIM43, through degradation of PCNT, induces nuclear lamina alterations that suppress transcriptionally active viral chromatin states and thereby viral lytic replication/reactivation.
It has become evident that members of the TRIM protein family play a critical role in the antiviral host response. Here we reveal a role for TRIM proteins in modulating KSHV reactivation. We found that silencing specific TRIM proteins in cells latently infected with KSHV induced spontaneous virus reactivation, suggesting that these TRIM proteins play a role in suppressing KSHV lytic replication.
Among the TRIM proteins identified by our screen, the previously uncharacterized TRIM43 protein was distinguished by its profound upregulation following infection with both KSHV and other herpesviruses and its ability to restrict a broad range of herpesviruses belonging to the α-, β- and γ-herpesvirus subfamilies. Furthermore, TRIM43 was also upregulated in herpesvirus-associated diseases in humans, including γ-herpesvirus-associated tumours. As such, TRIM43 and its transcription factor DUX4 could potentially serve as biomarkers of active herpesviral replication and pathogenesis. Moreover, it is tempting to speculate that differences in steady-state levels of endogenous TRIM43 due to cell type or cell differentiation stage could determine whether herpesviruses undergo active lytic replication or establish latency in a particular cell.
In contrast to many TRIM proteins whose expression is type I IFN-inducible, we found that the upregulation of TRIM43 during herpesvirus infection was not dependent on IFN but was instead mediated by the germline-specific transcription factor DUX4, which is itself transcriptionally upregulated during herpesviral infection via an unknown mechanism. In the absence of infection, we saw minimal to no expression of DUX4 in fully differentiated adult human cells. We further observed that DUX4 expression and herpesvirus infection transcriptionally induced a similar set of genes, some of which encode additional TRIM proteins. Intriguingly, based on genome-wide transcriptional analysis, the DUX4-dependent TRIM proteins that are upregulated during herpesviral infection are minimally or not at all expressed in normal (uninfected) differentiated cells, and their activities have yet to be characterized. Further investigation is needed to determine how herpesvirus infection upregulates DUX4 expression and how other DUX4-induced genes affect herpesviral replication. Moreover, future studies will need to determine the physiological relevance of TRIM43 in other DUX4-regulated cellular processes and diseases, such as regulation of zygote gene activation16 and FSHD19.
Our work has unveiled a mechanism of herpesviral inhibition in which a TRIM E3 ligase controls viral lytic gene expression by regulating the structure of host cell centrosomes and thereby the integrity of the nuclear lamina. As such, TRIM43 provides an attractive target for addressing latent herpesvirus reservoirs. Chemical inhibition or RNA-based depletion of TRIM43 may allow for reactivation of human herpesviruses from latent reservoirs, which could be exploited for therapeutic intervention against herpesvirus-associated diseases.
Cell culture and viruses
Primary HFF, HeLa, A549, Vero, HEK 293T and HEK 293 cells (all purchased from ATCC) as well as Huh7 (provided by M. Farzan, Scripps Institute Florida) were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Thermo Fisher Scientific) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS; Thermo Fisher Scientific), 2 mM GlutaMAX (Thermo Fisher Scientific) and 1% (vol/vol) penicillin–streptomycin (Thermo Fisher Scientific) under standard culture conditions. iSLK and iSLK.219 cells (harbouring recombinant (r) KSHV.219) were provided by D. Ganem (University of California) and maintained in DMEM supplemented with 10% (vol/vol) FBS, 2 µg ml−1 puromycin (Sigma) and 250 µg ml−1 G418 (Sigma). AKBM cells32 (provided by M. Ressing, Leiden University Medical Center) and BJAB cells (provided by J. Jung, University of Southern California) were cultured in Roswell Park Memorial Institute medium (RPMI, Thermo Fisher Scientific) supplemented with 10% (vol/vol) FBS, 2 mM GlutaMAX and 1% (vol/vol) penicillin–streptomycin. AGS-EBV cells33 (provided by N. Raab-Traub, University of North Carolina, Chapel Hill) were cultured in Ham’s F12 medium (Thermo Fisher Scientific) supplemented with 10% (vol/vol) FBS, 2 mM GlutaMAX, 1% (vol/vol) penicillin–streptomycin and 500 µg ml−1 G418 (Sigma).
HEK 293.rKSHV.219 cells were generated by infecting HEK 293 cells with rKSHV.219 (ref. 8), followed by selection with 1 µg ml−1 puromycin for at least 7 days. Lytic replication of KSHV was induced in iSLK.219 cells by plating cells in the absence of antibiotics and by adding 900 µM sodium butyrate (Sigma) and 1 µg ml−1 doxycycline (Sigma). Supernatants from induced iSLK cells containing infectious virus were harvested 3 days post-induction, and viral stocks were stored in aliquots at −80 °C until use.
Cell lines from ATCC were authenticated by ATCC and were not validated further in our laboratory. Cell lines that were obtained and validated by other groups were not further authenticated. All cell lines used in this study have been regularly tested for potential mycoplasma contamination by PCR or using the MycoAlert Kit (Lonza).
HSV-1 (strain KOS) was provided by D. Knipe (Harvard University). HSV-1-GFP (strain 17) was provided by B. Kaufer (Freie Universität Berlin) and adenovirus 12 (Ad12) was a gift from W. Doerfler (University of Erlangen-Nuremberg). Ad-GFP (type 5) was obtained from Vector Biolabs (#1060). For infection experiments with HCMV, the laboratory strain AD169 was used. VSV-GFP was a gift from S. Whelan (Harvard University) and SeV (Cantell strain) was purchased from Charles River Laboratories. DV serotype 2 (strain 16681), propagated in Aedes albopictus C6/36 cells (provided by M. Farzan, Scripps Institute Florida), has been described previously34. EMCV (strain EMC) was purchased from ATCC.
TRIM shRNA screen for KSHV reactivation
A custom-made TRIM shRNA library cloned into the pGIPZ lentiviral vector was purchased from Open Biosystems (173795 TRIM ReArray). To individually silence the 62 TRIM proteins, a pool of three shRNAs per TRIM protein with different targeting sequences were used. Lentiviral particles were generated by transfecting the respective pGIPZ plasmids encoding TRIM-specific shRNAs together with the lentiviral packaging plasmids psPAX2 and the VSV-G envelope expressing plasmid pMD2.G (both provided by D. Trono; Addgene plasmids nos. 12260 and 12259, respectively) into HEK 293T cells using Lipofectamine and Plus reagent (Thermo Fisher Scientific). After 48 h, supernatants were harvested, filtered using a 0.2 µm filter to remove cellular debris, and then stored at −80 °C until use. HEK 293.rKSHV.219 cells harbouring the rKSHV.219 genome were seeded into 24-well plates (∼50,000 cells per well). KSHV strain rKSHV.219 is a recombinant virus strain that contains a GFP under control of a CMV-immediate early promoter and a RFP under control of the viral PAN-promoter that allows monitoring of lytic replication. Cells transduced with pooled supernatants containing three different lentivirally packaged shRNAs for each TRIM protein were monitored daily for RFP expression using a Nikon TiE fluorescent microscope, and then harvested at 96 h post-transduction by detaching them using 0.1% EDTA (in PBS, pH 7.4). Cells were pelleted by centrifugation at 300g for 5 min, and then fixed with 2% (wt/vol) paraformaldehyde (PFA, in PBS) for 10 min, followed by washing cells twice with PBS. RFP-positive cells were analysed by FACS analysis using a LSRII flow cytometer equipped with a 546 nm laser (BD Biosciences), and data were analysed using FlowJo (Tree Star) software. The screen was performed in three biological replicates for each sample.
Reagents, plasmids and transfections
TRIM43A cDNA was purchased from Thermo Fisher (MGC Human TRIM43 sequence-verified cDNA, cat. no. MHS6278-202755530) and cloned into the pEF-BOS vector (containing an N-terminal FLAG-tag) using NotI and SalI restriction sites, or into the pEBG vector (containing N-terminal glutathione S-transferase, GST) using BamHI and ClaI. pEF-BOS-FLAG-TRIM43 was further used as a PCR template to generate FLAG-TRIM43 ΔRING (amino acids 53–446). Plasmid pCS2-mkg-DUX4, which encodes untagged human DUX4, was a gift from S. Tapscott (Addgene plasmid no. 21156; ref. 35). The plasmids encoding TRIM25-V5 and FLAG-TRIM25 have been described previously11,36. pLKO.1-sh.NT (sense sequence: CCTAAGGTTA AGTCGCCCTC GCTCGAGCGA GGGCGACTTA ACCTTAGG) and pLKO.1-sh.TRIM43 (cat. no. RHS3979–9601504) were purchased from Open Biosystems. FLAG-PCNT was provided by K. Rhee (Seoul National University) and has been described previously37,38. To generate the TRIM43RNAi-res plasmid, the following silent mutations (indicated in lower case) were introduced into the TRIM43 sequence targeted by the four individual siRNAs included in the Dharmacon siRNA SMARTpool: aCAacGctcgTGGGTtAAg (target sequence 1; bp 636–654), GaACgCAccGtCAgACgAA (target sequence 2; bp 281–299), TtTAtcGttTaAAtaGaTT (target sequence 3; bp 836–854) and CaGTgtTaCAcAAaGAgGA (target sequence 4; bp 563–581). The TRIM43RNAi-res-coding sequence was synthesized as an IDT gBlock and then cloned into pEF-Bos between BamHI and NotI with an N-terminal FLAG-tag using the Gibson Assembly method. All constructs were sequenced to verify 100% agreement with the original sequence.
Transfections were performed using the calcium phosphate method, GenJet (SignaGen Laboratories) or Lipofectamine and Plus reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. A goat F(ab′)2 fragment directed against human IgG (MP Biomedicals LLC) was used at a concentration of 0.1 mg ml−1 for 24 h to induce EBV reactivation in AKBM cells. Recombinant IFN-α2 was obtained from PBL Biomedical Laboratories, and MG132 was purchased from Sigma (M8699). Protease inhibitor cocktail was obtained from Sigma (P2714) and used at 1:500.
Virus replication or reactivation following siRNA-mediated TRIM knockdown
For determining HSV-1, Ad5-GFP, VSV-GFP and EMCV replication following transient knockdown of selected TRIM proteins (data shown in Fig. 1h and Supplementary Fig. 2d-i), Huh7 (used for HSV-1 or VSV-GFP) or HEK 293T (used for Ad5-GFP and EMCV) cells were seeded into 12-well plates (~5 × 105 cells per well) and, the next day, transfected with non-targeting control siRNA (si.C) or TRIM-specific siRNA (all purchased from Dharmacon) as described in detail in the following. At 36 h post-transfection, cells were infected with HSV-1 (MOI 0.1), Ad5-GFP (MOI 0.1), EMCV (MOI 1) or VSV-GFP (MOI 0.01). At 36 h (for HSV-1), 72 h (for Ad5-GFP), 24 h (for VSV-GFP) or 8 h (for EMCV) post-infection, cells were harvested by detaching them using 0.1% EDTA (in PBS, pH 7.4), and then pelleted by centrifugation at 300g for 5 min. Cells were fixed with 2% (wt/vol) PFA (in PBS) for 15 min, followed by washing cells twice with PBS. Ad5-GFP and VSV-GFP replication was assessed by determining GFP-positive cells using FACS analysis (LSRII flow cytometer, BD Biosciences).
To determine HSV-1 replication in TRIM-depleted cells, cells were incubated with blocking/permeabilization buffer (5% (vol/vol) FBS and 0.1% (wt/vol) saponin in PBS), followed by staining with primary (anti-ICP-8; 1:1,000) and secondary antibody (anti-rabbit-Alexa 488 (1:400)) on ice in blocking/permeabilization buffer for 30 min each. The percentage of ICP-8-positive cells was determined on an LSRII flow cytometer (BD Biosciences), and data were analysed using FlowJo (Tree Star) and FCS Express 3 softwares (De Novo Software).
EMCV replication was determined by TCID50 (median tissue culture infective dose) assay. Briefly, supernatants from TRIM-depleted cells that had been infected with EMCV were serially diluted and then added onto Vero cells. After 24 h, wells with dead cells were counted to determine the infectious titre according to the Reed–Muench method39.
HCMV (strain AD169) replication was determined via IE-1 fluorescence as described in detail previously40. Briefly, HFF cells transiently transfected with control non-targeting siRNA or TRIM43-specific siRNA were infected with HCMV at the indicated MOI. After 24 h, cells were fixed and stained with a monoclonal antibody (p63–27) directed against IE-1 (ref. 41). The number of IE-1-positive cells was determined and used to calculate viral titres, expressed as IE protein-forming units (IEU).
EBV reactivation in latently infected AGS-EBV cells transfected with non-targeting control siRNA or TRIM43-specific siRNA was determined by qRT–PCR analysis of viral BMRF1 gene expression using forward primer 5′-CAACACCGCACTGGAGAG-3′, reverse primer 5′-GCCTGCTTCACTTTCTTGG-3′ and probe 5′-aggaaaaggacatcgtcggaggc-3′ on a 7500 Fast Real-Time PCR System (Thermo Fisher Scientific) or by IB analysis of EBV Zta protein using an anti-Zta antibody (1:500, BZ1, Santa Cruz).
HSV-1 plaque assays
Viral stocks were serially diluted in PBS containing 0.5 mM MgCl2, 0.9 mM CaCl2, 0.1% dextrose and 5% (vol/vol) heat-inactivated FBS. Vero cells, seeded into 6- or 12-well plates, were incubated with virus dilutions for 1 h at 37 °C. The inoculum was then replaced with MEM containing 5% (vol/vol) heat-inactivated FBS and 1.5% (wt/vol) carboxymethylcellulose and incubated at 37 °C for 4 days. Cells were fixed using methanol for 5 min and then stained for 30 min with crystal violet (Sigma) diluted in 20% (vol/vol) ethanol. Cells were rinsed with water, dried, and plaques were counted.
SiRNA-mediated knockdown of TRIM proteins, PCNT and DUX4
Indicated cells were seeded into 24- or 12-well plates (~2.5 or 5 × 105 cells per well). The next day, cells were transfected with 60 nM or 120 nM of gene-specific siRNAs using Lipofectamine RNAiMAX (Thermo Fisher Scientific) according to the manufacturer’s instructions. SiRNAs targeting TRIM43 (siGENOME SMARTpool M-007127-01-0050), PCNT (siGENOME SMARTpool M-012172-01-0005), DUX4 (SMARTpool Lincode R-188080-00-0005), TRIM4 (siGENOME SMARTpool M-007101-00-0005), TRIM23 (siGENOME SMARTpool M-006523-00-0005), TRIM25 (siGENOME SMARTpool M-006585-00-0005), TRIM42 (siGENOME SMARTpool M-007173-00-0005), TRIM45 (siGENOME SMARTpool M-007073-01-0005), TRIM46 (siGENOME SMARTpool M-007071-01-0005), TRIM47 (siGENOME SMARTpool M-007106-02-0005), TRIM65 (siGENOME SMARTpool M-018490-01-0005) and TRIM66 (siGENOME SMARTpool M-026772-01-0005) as well as the non-targeting control siRNA (siGenome Control Pool D-001206-14-50) were purchased from Dharmacon. At 48 h post-transfection, the knockdown efficiency of endogenous TRIM proteins, DUX4 or PCNT was determined by qRT–PCR or IB analysis, as described in the following sections.
Analysis of TRIM43 mRNA expression in human tissues
The Human Total RNA Master Panel II (Clontech cat. no. 636643) was used to determine the absolute number of TRIM43 transcripts in different human tissues. qPCR was performed in one step using 20 ng of the individual cDNA library samples and the SuperScript III Platinum Kit (Thermo Fisher Scientific) on a 7500 Fast Real-Time PCR Machine (Thermo Fisher Scientific) according to the manufacturer’s protocols. A TaqMan probe for the TRIM43 gene was acquired as premixed Taqman MGB assay (Thermo Fisher Scientific, cat. no. 4351370) and added to the reaction. A standard of known concentration was serially diluted, which served to determine the absolute number of TRIM43 transcripts in the individual tissues.
HFF cells were grown in 6-well plates on cover glasses (Marienfeld Superior). After 24 h, cells were fixed with 4% (wt/vol) PFA (in PBS) for 5 min and then permeabilized with 0.1% (vol/vol) Triton X-100 for 3 min. Cells were blocked with 5% (vol/vol) bovine serum (in PBS) for 1 h and then stained with anti-TRIM43 antibody (Abcam 80460; diluted 1:50 in 5% (vol/vol) bovine serum (in PBS)) overnight at 4 °C. Cells were extensively washed with PBS and then stained with 5 nm gold particles coupled to protein A for 1 h at room temperature. Cells were washed and then fixed with 1.25% glutaraldehyde and 0.015% picric acid in 50 mM sodium cacodylate buffer (pH 7.4) at room temperature for 30 min. After washing with sodium cacodylate buffer (50 mM, pH 7.4), cell layers were embedded in Epon for 24–48 h at 40 °C. Perfused cells were cut ultrathin (Reichert-Ultracut-S) and then transferred into a formvar/carbon-coated copper grid for imaging. All images were acquired using the Tecnai G² Spirit BioTWIN (FEI) at the Electron Microscopy Facility of Harvard in Boston.
Pulldown assay, co-IP and IB analysis
HEK 293T cells were lysed in RIPA buffer (150 mM NaCl, 0.5–1% (vol/vol) NP-40, 1% (wt/vol) deoxycholic acid (DOC), 0.01% (vol/vol) SDS, 20 mM Tris pH 8.0) or NP-40 buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 1% (vol/vol) NP-40, protease inhibitor cocktail (Sigma) and 10 μg ml−1 proteasome inhibitor MG132), and cell debris were pelleted by centrifugation at 13,000 r.p.m. for 20 min at 4 °C. FLAG pulldowns, co-IP and western blot analyses were performed as previously described11,34. Elution of immunoprecipitated proteins was performed by heating samples in Laemmli SDS sample buffer at 95 °C for 5 min. For immunoblotting, the following primary antibodies were used: anti-FLAG (M2; 1:2,000; Sigma), anti-HA (1:2,000; clone HA-7; Sigma), anti-ICP-8 (1:5,000, provided by D. Knipe, Harvard), anti-Zta (1:500, BZ1, Santa Cruz), anti-V5 (1:5,000, R960–25, Novex), anti-GST (1:5,000, Sigma), anti-PCNT (1:1,000, clone 28144, Abcam), anti-lamin A/C antibody (clone 636, Santa Cruz Biotech), anti-γ-tubulin (1:1,000, clone GTU-88, Abcam), anti-DUX4 (1:1,000, clone 9A12, Millipore), anti-SMC1a (1:500, rabbit polyclonal, Bethyl antibodies) anti-SMC3 (1:500, ab9263, Abcam), anti-RAD21 (1:500, 05–908, Millipore), anti-α-Tubulin (1:1,000, rabbit polyclonal, Genscript), anti-ubiquitin (1:500, clone P4D1, Santa Cruz Biotech) and anti-β-actin (1:10,000; Abcam or AC-15, Sigma).
To determine endogenous PCNT ubiquitination following HSV-1 infection of cells depleted of TRIM43, HEK 293T cells were transduced with lentivirally packaged non-targeting control shRNA or TRIM43-specific shRNA and then selected using puromycin (1 µg ml−1) at 72 h post-transduction. Cells were transiently transfected with FLAG-tagged ubiquitin, and, 24 h later, either infected with HSV-1 (MOI 1) for 16 h or left uninfected. Cells were further treated with MG132 (50 µM) for 4 h to limit the degradation of PCNT triggered by HSV-1 infection. Cells were lysed in NP-40 buffer (300 mM NaCl, 10 mM Tris HCl pH 8.0, 1% (vol/vol) NP-40 and protease inhibitor cocktail (Sigma)), and cell debris were pelleted by centrifugation at 10,000 r.p.m. for 10 min at 4 °C. Lysates were incubated with anti-PCNT-coupled G magnetic beads to immunoprecipitate PCNT.
Large-scale protein purification and MS analysis
To identify cellular interaction partners of TRIM43, twelve 10 cm dishes of HEK 239T cells (~1 × 107 cells per dish) were each transfected with 15 µg of pEF-BOS empty vector, pEF-BOS-FLAG-TRIM43 ΔRING or pEF-BOS-FLAG-TRIM25 (control). After 36 h, cells were lysed with Triton X-100 buffer (150 mM NaCl, 5% EDTA, 1% (vol/vol) Triton X-100 pH 7.5, protease inhibitor cocktail (Sigma)), followed by centrifugation at 13,000 r.p.m. for 20 min at 4 °C. Lysates were pre-cleared using 100 μl Sepharose 4B (Sigma-Aldrich) to remove cell debris and then mixed with an ~50% slurry of anti-FLAG M2 Affinity Gel (Sigma) and incubated for 4 h at 4 °C. After extensive washing of the beads with lysis buffer, bound proteins were eluted by heating samples in 2× Laemmli SDS sample buffer for 5 min at 95 °C and subsequently separated on a NuPAGE 4–12% Bis–Tris gradient gel (Thermo Fisher Scientific). To stain interacting proteins, a Silverquest staining kit (Thermo Fisher Scientific) was used according to the manufacturer’s instruction. Bands that were specifically present in the FLAG-TRIM43 ΔRING sample, but not in the empty vector and FLAG-TRIM25 samples, were excised and analysed by ion-trap MS at the Harvard Taplin Biological Mass Spectrometry Facility in Boston.
HeLa or HFF cells (~2.5 × 105 cells per sample) were seeded onto cover glasses (Marienfeld Superior). The next day, cells were transfected with plasmid DNA (usually 1 µg DNA per sample) using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. At 24–36 h post-transfection, cells were fixed with 2% (wt/vol) PFA (in PBS) for 10 min, permeabilized with 0.1% (vol/vol) Triton X-100, and then blocked with 5% (vol/vol) bovine serum (in PBS) for 1 h at room temperature. Cell preparation and confocal microscopy analysis were performed as previously described11. For immunostaining, cells were incubated with the following antibodies, usually for 1 h at room temperature: anti-TRIM43 (1:400; Abcam 80460), anti-DUX4 (1:4,000; clone 9A12, Millipore), anti-ICP-8 (1:1,000; provided by D. Knipe, Harvard), anti-PCNT (1:400; clone 28144, Abcam), anti-γ-tubulin (1:400; clone GTU-88, Abcam), anti-FLAG (1:1,000; M2, Sigma), anti-lamin A/C antibody (clone 636, Santa Cruz Biotech), anti-lamin B1 (clone B10, Santa Cruz Biotech), anti-centrobin (1:500; Sigma) and anti-SAS-6 (1:400; clone G1, Santa Cruz). This was followed by incubation for 1 h at room temperature with the secondary antibodies donkey anti-rabbit Alexa Fluor 594, donkey anti-mouse Alexa Fluor 488, donkey anti-mouse Alexa Fluor 555, donkey anti-rabbit Alexa Fluor 555, donkey anti-rabbit Alexa Fluor 488, donkey anti-mouse Alexa Fluor 647 or donkey anti-rabbit Alexa Fluor 647 (all purchased from Thermo Fisher Scientific). Nuclei were stained with DAPI (Sigma). The samples were imaged using an Olympus IX8I confocal microscope or a Leica TCS SP5 confocal microscope, with 405, 488, 543 or 633 nm laser lines, scanning each channel separately under image capture conditions that eliminated channel overlap. The Fiji distribution of ImageJ was used for image analysis and quantification. For 3D reconstruction of images, stacks of confocal images from single cells imaged using a Leica TCS SP5 microscope were deconvoluted and subjected to 3D surface rendering with Huygens Professional Software (Scientific Volume Imaging).
Total RNA was extracted from cells using the HP Total RNA Kit (OMEGA bio-tek) according to the manufacturer’s instructions. Reverse transcription and qRT–PCR were performed in one step using 0.1–1 µg of RNA as template using the SuperScript III Platinum One-Step qRT–PCR kit with ROX (Thermo Fisher Scientific) on a 7500 Fast Real-Time PCR machine (Thermo Fisher Scientific) according to the manufacturer’s protocols. Primers/probes for GAPDH, HPRT, 18S rRNA, DUX4, TRIM43, TRIM4, TRIM8, TRIM15, TRIM23, TRIM33, TRIM34, TRIM41, TRIM42, TRIM43, TRIM45, TRIM46, TRIM47, TRIM54, TRIM65, TRIM66 and TRIM69 were acquired as premixed master mixes (IDT or Thermo Fisher Scientific/ABI) and added to the reaction. For detection of HSV-1 ICP-0, forward primer 5′-CGGACACGGAACTGTTCGA-3′, reverse primer 5′-CGCCCCCGCAACTGC-3′ and probe 5′-/56-FAM/CCCCATCCA/ZEN/CGCCCTG/3IABKFQ/−3′ (all purchased from IDT) were used. Expression levels for each target gene were calculated by normalizing values to cellular GAPDH, HPRT or 18S rRNA. Fold induction of each target gene relative to mock-infected or mock-treated cells was calculated using the comparative CT method (ΔΔCT method). Quantification of KSHV viral genome copies with primers specific for KSHV orf26 (orf26-fwd primer: 5′-GCTAGCAGTGCTACCCCCATT-3; orf26-rev primer 5′-GGTCAAATCCGTTGGATTCG-3 and orf26 probe: FAM/AGCCGAAAGGATTCCACCATTGTGC/TAMRA, all purchased from IDT) was performed as described previously42.
TRIM43 expression analysis in BAL samples
Anonymized BAL samples for analysis of TRIM43 transcript expression were selected from the sample repository of the diagnostics department of the Institute for Clinical and Molecular Virology at University Hospital Erlangen. Samples that were no longer needed for diagnostic purposes and assigned for disposal were exclusively used. Patient samples that tested positive for multiple viruses were excluded from the analysis, as well as samples with fewer than 120,000 albumin transcript copies in the initial nucleic acid extraction to allow for efficient RNA extraction. In total, 34 BAL samples from patients with pulmonary infections were analysed, from which 18 BAL samples were positive for HSV-1 DNA by qPCR, and 16 BAL samples were negative for HSV-1 DNA by qPCR. Total nucleic acids from 400 µl of BAL sample were extracted on an automated Qiagen EZ1 analyser using the Qiagen EZ1 virus mini kit v2.0 according to the manufacturer’s instructions, and phage M2-RNA was added as an extraction control. Extracted nucleic acids were eluted in 60 µl ultrapure water. A 6 µl volume of 10× PCR buffer (New England Biolabs) was added to the 60 µl of eluted nucleic acids, and DNA was removed by incubation with 1 U of RNAse-free DNAse (Epicentre) for 15 min at 37 °C. DNAse was inactivated at 70 °C for 15 min. qRT–PCR was performed using 8 µl of each sample in a 20 µl reaction with a 2× universal probe One-Step RT–qPCR Kit (New England Biolabs) and primer/probe mixes for HPRT (IDT, cat. no. Hs.PT.58v.45621572) and TRIM43 (Thermo Fisher Scientific, cat. no. 4351370) spanning two exons to exclude amplification of genomic DNA on an ABI 7500 real-time PCR system. Relative induction of TRIM43 transcripts was calculated by normalizing values to HPRT values using 1/2^(CtTRIM43-CtHPRT). In samples negative for TRIM43 mRNA, Ct values were set to 45 (45 PCR cycles). Statistical significance was calculated using a Mann–Whitney U test (*P < 0.05).
HEK 293T cells, seeded into 12-well plates (~500,000 cells per well), were mock-treated, transfected with 2 µg DNA of pCS2-mkg-DUX4 or infected with HSV-1 (MOI 10) for 18 h. Cells were harvested in ice-cold PBS, and total cellular RNA was extracted using standard phenol/chloroform extraction. Purified RNA was converted to a DNA library and sequenced at the Harvard Biopolymers Facility. Up to 1,000 ng total RNA was used in the Wafergen (formerly IntegenX) PrepX RNA–Seq Library Kit (cat. no. 400039) and placed into the Wafergen Apollo, and run using a standard protocol. PCR was performed on samples using indexed primers according to Wafergen instructions and then cleaned using the Wafergen Apollo. The resulting samples were run on an Agilent 2200 Tape Station on a D1000 High Sensitivity Tape with ladder provided to assess the integrity and amount of DNA. DNA libraries were run in a qPCR assay with SYBR green KAPA SYBR FAST Universal 2× qPCR Master Mix (Kapa Biosystems, cat. no. KK4602) and primers to the P5 and P7 regions of the adapters. Diluted PhiX was used for the standard curve to determine concentration. Libraries passing quality control were subjected to Illumina HiSeq 2500 sequencing. A local Galaxy server was used to process the RNA–seq data43. Raw sequence reads were quality trimmed using TRIM Galore! (Cutadapt44 and FastQC45 wrapper) and then aligned to the human genome using Tophat2 (ref. 46) using standard settings. Htseq-count was used to calculate the induction of human transcripts, and values were normalized using total count normalization47. Of note, transcripts of the three TRIM43 gene copies found in humans (TRIM43A, TRIM43B and TRIM43C) were analysed. Data were visualized in R (ref. 48) using the ggplot2 package. The RNA–seq data have been deposited in NCBI’s GEO database (no. GSE101435).
Immunohistochemical analysis of TRIM43 protein in KS tissues
KS tissues were obtained from skin biopsies of 17 HIV-1-infected individuals with AIDS in the context of medically required diagnostic or therapeutic procedures. Uninvolved tissues of the skin of five of the patients as well as cerebellum tissue of an HIV-1-negative person were obtained as controls. The use of these post-diagnostic tissues was approved by the ethics committee of University Hospital Erlangen.
Staining of PFA-fixed, paraffin-embedded tissue sections for TRIM43 and KSHV LANA-1 were performed as previously described49. A rabbit polyclonal antibody directed against human TRIM43 (ABCAM, ab80460, 1:50) and a rat monoclonal antibody against LANA-1 (Tebu-bio, cat. no. 13-210-100, 1:250) were used as primary antibodies. Primary antibody binding was detected using the respective Vectastain Elite ABC Kits (Vector Laboratories). For both antibodies, citrate buffer pH 6.0 (DakoCytomation) was used for antigen retrieval. Sections were developed using NovaRed Substrate (Vector Laboratories), counterstained with Hematoxylin Gill III (Merck), and analysed using a DM6000 B microscope (Leica).
CRISPRainbow labelling of HSV-1 genomes
A guide RNA (gRNA) directed against the repeats of HSV-1 was designed (5′-GTTACCTGGGACTGTGCGGT-3′) and cloned into pLH-sgRNA1–2XMS2 (a gift from T. Pederson, Addgene no. 75389; ref. 31). Plasmids pHAGE-TO-MCP-3XBFPnls (Addgene no. 75383; ref. 31), pHAGE-TO-dCas9 (Addgene ref. 75381; ref. 31) and pLH-sgHSV1–2XMS2 were co-transfected into HeLa cells. At 24 h post-transfection, cells were re-seeded onto cover slips (~50,000 cells per sample) and then transfected with either control non-targeting siRNA or TRIM43-specific siRNA (both Dharmacon). After 24 h, cells were infected with HSV-1 (MOI 1) for 1.5 h. Cells were incubated with primary antibodies specific for H3K4me3 (rabbit polyclonal, Abcam, 1:500) and viral ICP-0 (mouse monoclonal, clone 11060, Santa Cruz, 1:400) for 1 h at room temperature, followed by incubation for 1 h at room temperature with the secondary antibodies donkey anti-rabbit Alexa Fluor 488 and donkey anti-mouse Alexa Fluor 647 (both purchased from Thermo Fisher Scientific). Stacks of confocal images from single cells imaged using a Leica TCS SP5 microscope were deconvoluted and subjected to 3D surface rendering with Huygens Professional Software (Scientific Volume Imaging) to visualize co-localization of HSV-1 genomes with H3K4me3.
HFF cells, seeded into 10 cm dishes (~1 × 106 cells per dish), were transfected with the indicated siRNAs at a concentration of 120 nM using Lipofectamine 2000 (Thermo Scientific) according to the manufacturer’s instructions. After 36 h, cells were infected with HSV-1 (KOS strain; MOI 1) for 5 h. Cells were fixed in 1% formaldehyde for 10 min and fixation was quenched by the addition of glycine (final concentration of 0.125 M). Cell pellets were washed twice with ice-cold PBS containing protease inhibitor cocktail (Cell Signaling Technologies), resuspended in 300 µl of sonication buffer (10 mM Tris pH 8.0, 0.25% (wt/vol) SDS, and 2 mM EDTA), and then sonicated using a Bioruptor (Diagenode). Samples were diluted 1:1.5 in equilibration buffer (10 mM Tris, 233 mM NaCl, 1.66% (wt/vol) Triton X-100, 0.166% (wt/vol) sodium deoxycholate, 1 mM EDTA, and protease inhibitor cocktail), centrifuged at 13,000 r.p.m. for 10 min, and supernatants were then transferred into new tubes. Chromatin was immunoprecipitated by incubation with 5 µl of anti-trimethyl-histone3-lysine 4 antibody (rabbit polyclonal, ab8580, Abcam) overnight at 4 °C, followed by incubation with 25 µl Protein A-Dynabeads (Thermo Scientific) for 2.5 h. Beads were washed twice each with RIPA low-salt buffer (containing 140 mM NaCl), RIPA high-salt buffer (containing 500 mM NaCl) and RIPA LiCl buffer (containing 250 mM LiCl). Beads were then washed once in 10 mM Tris (pH 8.0), and chromatin eluted in 50 µl elution buffer (10 mM Tris pH 8.0, 0.4% (wt/vol) SDS, 5 mM EDTA, 300 mM NaCl and 2 µl proteinase K) by incubation for 1 h at 65 °C and then 6 h at 55 °C. Eluted DNA was purified using Agencourt AMPureXP Beads (Beckmann Coulter) in a 1.8× volume according to the manufacturer’s instructions. DNA was subjected to SYBR green qRT–PCR analysis using the following forward (fwd) and reverse (rev) primers: VP16-fwd: 5′-CGGATTGGGAAACAAAGGCACGCAACGCC-3′; VP16-rev: 5′-TTGAGGTCTTCGTCTGTG-3′; TK-fwd: 5′-ATGCTGCCCATAAGGTATCG-3′; TK-rev: 5′-GTAATGACAAGCGCCCAGAT-3′; ICP-8-fwd: 5′-CCACGCCCACCGGCTGATGAC-3′; ICP-8-rev: 5′-TGCTTACGGTCAGGTGCTCCG-3′; ICP27-fwd: 5′-CCACGGGTATAAGGACATCCA-3′; ICP27-rev: 5′-GGATATGGCCTCTGGTGGTG-3′; Actin-fwd: 5′-CGGGAAATCGTGCGTGACATTAAG-3′; Actin-rev: 5′-GAACCGCTCATTGCCAATGGTGAT-3′. Quantitative PCR was performed using Luna SYBR 2× Master Mix (New England Biolabs) on an ABI 7500 machine (Thermo Fisher Scientific). Values were calculated using the ΔCt method and normalized to input DNA.
P values for the TRIM shRNA screen were calculated using an unpaired two-tailed Student’s t-test. Three times each of 10,000 cells per sample from biological replicates were quantified. P < 0.05 was considered statistically significant. For analysis of TRIM43 expression in patient samples, statistical significance was calculated using a Mann–Whitney U test. P < 0.05 was considered statistically significant. For the data shown in Fig. 6b, statistical significance was calculated using a two-tailed χ2 test. P < 0.05 was considered statistically significant. For all other experiments, an unpaired two-tailed Student’s t-test was used to compare differences between two unpaired experimental groups in all cases. P < 0.05 was considered statistically significant.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
The data that support the findings of this study are available from the corresponding author upon request. RNA–seq data from this study are deposited in NCBI GEO under accession code GSE101435. Supplementary figures and tables are available in the Supplementary Information. Complete western blot images of all figures in the manuscript are provided in Supplementary Fig. 11.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors thank M. Ericsson from the Harvard Electron Microscopy Facility, Boston, for assistance with sample preparation and electron microscopy, and R. Tomaino (Taplin Mass Spectrometry Facility, Harvard) for mass spectrometry analysis. The authors also thank G. Förtsch (Division of Molecular and Experimental Surgery, University Hospital Erlangen) for excellent technical assistance and R. Coras (Department of Neuropathology, University Hospital Erlangen) for cerebellum tissue used as a staining control. This study was supported by the US National Institutes of Health grants R21 AI118509, R01 AI087846 and R01 AI127774 (to M.U.G.), and grants from the German Research Foundation CRC796, TP B1 and EN423/5-1 (to A.E.), CRC796 and STA357/7-1 (to T.S.), FOR 2438/subproject 2 (to M.St.), FU 949/1-1 and FU 949/2-1 (to F.F.) and SP 1600/1-1 (to K.M.J.S.). F.F. was further supported by a Marie Skłodowska-Curie Individual Fellowship from the European Union’s Framework Programme for Research and Innovation Horizon 2020 (2014–2020) under grant agreement no. 703896, and the Interdisciplinary Center for Clinical Research Erlangen (IZKF, J57). M.A.Z. received support from NIH training grant T32 GM007183. A.E. also received funding from IZKF, A66.