RIP1/RIP3 Binding to HSV-1 ICP6 Initiates Necroptosis to Restrict Virus Propagation in Mice

Transcription

1 Article RIP1/ Binding to ICP6 Initiates Necroptosis to Restrict Virus Propagation in Mice Graphical Abstract Authors Zhe Huang, Su-Qin Wu,..., Chunfu Zheng, Jiahuai Han Correspondence (Y.L.), (J.H.) In Brief Necroptosis plays a role in restricting propagation of certain viruses. Huang et al. find that murine but not human RIP1/ directly senses ICP6 to initiate necroptosis. These findings suggest that necroptosis is a host defense mechanism against and that has evolved to escape necroptosis-mediated restriction in humans. Highlights d -mediated necroptosis plays a vital role in restricting propagation in mice d d d In mice, ICP6 binds RIP1/3 and promotes their interaction, triggering necroptosis ICP6 initiates necroptosis in mouse cells but inhibits necroptosis in human cells Virus-host RHIM interactions can be either pro-necroptotic or anti-necroptotic Huang et al., 2015, Cell Host & Microbe 17, February 11, 2015 ª2015 Elsevier Inc.

2 Cell Host & Microbe Article RIP1/ Binding to ICP6 Initiates Necroptosis to Restrict Virus Propagation in Mice Zhe Huang, 1,4 Su-Qin Wu, 1,4 Yaoji Liang, 1,4, * Xiaojuan Zhou, 1 Wanze Chen, 1 Lisheng Li, 1 Jianfeng Wu, 1 Qiuyu Zhuang, 1 Chang an Chen, 1 Jingxian Li, 1 Chuan-Qi Zhong, 1 Weixiang Xia, 1 Rongbin Zhou, 2 Chunfu Zheng, 3 and Jiahuai Han 1, * 1 State Key Laboratory of Cellular Stress Biology, Innovation Center for Cell Signaling Network, School of Life Sciences, Xiamen University, Xiamen, Fujian , China 2 School of Life Science, Innovation Center for Cell Signaling Network, University of Sciences and Technology of China, Hefei, Anhui, China 3 Soochow University, Institutes of Biology and Medical Sciences, Suzhou, Jiangsu, China 4 Co-first author *Correspondence: yaojiliang@xmu.edu.cn (Y.L.), jhan@xmu.edu.cn (J.H.) SUMMARY Necroptosis is a form of programmed necrosis that is mediated by signaling complexes containing the receptor-interacting protein 3 () and RIP1 kinases. We show that and its interaction with the herpes simplex virus type 1 () protein ICP6 triggers necroptosis in infected mouse cells and limits viral propagation in mice. ICP6 interacts with RIP1/ through its RHIM domain and forms dimers/oliogmers by its C-terminal R1 domain. These binding events result in RIP1- hetero- and - homo-interactions and subsequent necroptosis of -infected mouse cells. However, ICP6 RHIM cannot trigger necroptosis and even inhibits TNFinduced necroptosis in human cells. As the RHIM domain in murine cytomegalovirus protein vira can inhibit necroptosis in both human and mouse cells, these data suggest that both viral and host RHIM sequences determine whether the virushost RHIM interaction is pro- or anti-necroptotic and that some viruses may evolve to escape this restriction. INTRODUCTION Like the production of anti-viral cytokines such as interferon, death of virus-infected cells is also a host defense mechanism to control viral spread (Mocarski et al., 2012). Apoptosis depends on activation of caspases, and many viruses encode caspase inhibitors to suppress infection-induced apoptosis. Necrosis can also be used by host to clear invading pathogens (Han et al., 2011; Mocarski et al., 2012). Necroptosis is a type of programmed necrosis mediated by signaling complexes called necrosomes (Christofferson and Yuan, 2010; Han et al., 2011; Vandenabeele et al., 2010). The core components of necrosomes are receptor-interacting protein kinase 1 (RIP1) (Holler et al., 2000) and (Cho et al., 2009; He et al., 2009; Zhang et al., 2009), both of which are protein kinases containing an RIP homotypic interaction motif (RHIM). Our knowledge of necroptosis was mostly obtained by studying tumor necrosis factor-a (TNF-a)-induced necroptosis. TNF induces formation of necrosomes containing RIP1,, caspase-8, and adaptors TRADD and FADD in cells with high levels of expression and/or caspase-8 inhibition (Han et al., 2011; Vandenabeele et al., 2010). Phosphorylation of in necrosomes leads to the recruitment of mixed-lineage kinase domain-like protein (MLKL) (Sun et al., 2012; Zhao et al., 2012). The interaction with leads to MLKL phosphorylation and translocation to the plasma membrane, which is followed by an ion influx and cell membrane disruption (Cai et al., 2014; Chen et al., 2014; Dondelinger et al., 2014; Wang et al., 2014a). -mediated necrosis plays an important role in immune defense against some viruses, including vaccinia virus and murine cytomegalovirus (MCMV) (Cho et al., 2009; Upton et al., 2010). is a double-stranded DNA virus that has evolved numerous strategies to infect a wide range of hosts, including human and mice (Karasneh and Shukla, 2011). In comparison with in human cells, the replication of most strains is less efficient in mouse cells (Lopez, 1975). ICP6, also known as ribonucleotide reductase (RNR) subunit 1 (R1), is encoded by the UL39 gene of. RNR acts to synthesize the four deoxyribonucleotides required for DNA synthesis. With an RHIM-like domain (Lembo and Brune, 2009), ICP6 has been reported to be important for protecting -infected cells against death receptor (DR)-mediated apoptosis by interacting with caspase-8 (Dufour et al., 2011). We have examined the propagation of several viruses in knockout (KO) murine cells and found that replication was markedly elevated after deletion. We unveiled that this elevation was due to the lack of necroptosis in -infected KO cells. Further mechanistic studies revealed that protein ICP6 directly interacts with RIP1 and through its RHIM domain, and these interactions lead to RIP1- and - interaction and subsequent necroptosis. Our findings suggest that the previously observed restriction of in mouse cells may be partly due to -mediated necroptosis and provide insight into -dependent necrosis in sensing and defending certain viruses. Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. 229

3 RESULTS Is a Cellular Factor that Restricts Propagation In consideration of the role of in MCMV-induced necroptosis (Upton et al., 2010), we examined the role of cellular in the propagation of a few other viruses. Wild-type (WT) and KO cell lines from murine fibroblast cell line L929 (Chen et al., 2014) were infected with the KOS strain (hereinafter referred to as ), VSV, or NDV. (Figure 1A), but not VSV or NDV (Figures S1A and S1B), propagates much more in KO L929 cells than in WT L929 cells. The titer of in KO L929 cells was markedly higher than that in WT L929 cells after the cells were infected with (1 moi) for 24 hr or longer time. propagation was restricted when the expression of was reconstituted (Figure 1B). Consistently, the accumulation of glycoprotein D (gd), a envelope glycoprotein, in KO cells was much higher than that in WT L929 cells (Figure 1C). This phenotype was reversed when expression was reconstituted in KO cells (Figure 1D). We also used 1 moi GFP- (F strain), a GFP-tagged, to infect the cells and observed that the GFP intensity of KO cells was much higher than that of WT L929 cells (Figure 1E). Reconstitution of expression suppressed GFP- propagation in KO cells (Figure 1F). The restriction of propagation by endogenous or ectopically expressed was also observed in other cell lines, such as mouse embryonic fibroblast (MEF) and NIH 3T3 cells (Figures S1C and S1D). Collectively, these data demonstrate that restricts propagation in some cell lines. Infection Leads to -Dependent Cell Death To determine how restricts, we investigated the role of from the following three aspects: entry, genomic DNA replication, and spreading. To analyze cell penetration by, WT and KO L929 cells were infected with or without at 200 moi for 2 hr in the presence of cycloheximide (CHX) to block the synthesis of proteins including proteins. genomic DNA levels in the infected cells were analyzed by measuring the levels of gd, ICP22, and ICP47 DNA with qpcr, and no difference was found between WT and KO cells (Figure 2A). We also examined the amounts of tegument protein VP16, an virion protein internalized into infected cells, and found that the amounts of VP16 were the same in WT and KO L929 cells (Figure 2B). To analyze genomic DNA replication, genomic DNA level within the first life cycle of infection (0 10 hr) was measured. -infected KO L929 cells did not exhibit higher genomic DNA replication efficiency in comparison with WT cells (Figure 2C). Thus, deletion does not affect entry into the cells or genomic DNA replication. We then used GFP- to monitor the infection of in WT and KO L929 cells (Figure 2D). At 12 hr post-infection, the numbers of infected cells (GFP + cells) are the same in the two cell lines. But at later time points, the number of GFP + KO L929 cells continued to increase while that of GFP + WT L929 cells declined. We monitored the viability of GFP + WT and KO L929 cells by time-lapse fluorescence microscopy in the presence of propidium iodide (PI), which can stain dead cells (Figure 2E). Fifty GFP + cells in each of the two cell lines had been monitored for 42 hr. All of the GFP + WT L929 cells turned PI positive while only 3 of the 50 GFP + KO cells became so. The animation of time-lapse pictures showed that the -infected (GFP + ) WT L929 cells became round and finally ruptured (PI positive) after GFP intensity reached a certain level, whereas GFP + KO L929 cells were still viable up to 42 hr after infection (Movies S1 and S2), suggesting that the accumulation of viral proteins might be required for -induced cell death. Although we observed death of -infected WT L929 cells (Figure 2E), we did not see cell loss. A possible explanation is that 1 moi virus only infected a small amount of cells, and the cell loss was soon counterbalanced by the proliferation of newborn cells. To confirm that infection can induce -dependent necrosis in L929 cells, we treated WT and KO L929 cells with high-dose (20 moi), which would result in much higher infection efficiency than 1 moi. The infection of WT but not KO L929 cells led to cell loss (Figure 2F). The cell death was caspase independent, since pancaspase inhibitor zvad did not inhibit but rather enhanced -induced cell death (Figure 2G), indicating that the cell death was not caused by apoptosis. Taking into consideration the morphology of dying cells, the requirement of, and the independence of caspase, -induced cell death is necroptosis. Similar data were obtained in other murine cell lines, such as MEF and NIH 3T3 cells (Figures S2A and S2B). Since human is a natural host of, we examined whether also plays a role in suppressing propagation in human cells. HT-29 is a human cell line that expresses and undergoes necroptosis upon TNF plus zvad plus Smac mimetic stimulation (He et al., 2009)(Figure S2C). We infected HT-29 cells with but did not detect -dependent necroptosis (Figure S2D). Thus, -induced necroptosis might be species dependent. We also examined the viability of L929 cells after VSV or NDV infection. VSV induced L929 cell death whereas NDV infection did not affect the viability of L929 cells. The levels of VSVinduced cell death were comparable in WT and KO cells (Figure S2E). Gene deletion of also did not affect the viability of NDV-infected L929 cells (Figure S2F). Together with the results of viral replication shown in Figures S1A and S1B, these data indicate that is involved in murine cell defense against but not VSV or NDV. -Induced Necroptosis Is Independent of TNFR1, Partially Dependent on RIP1, and Fully Dependent on and MLKL Since infection induces inflammatory cytokines including TNF, which can induce -dependent necroptosis via TNF receptor 1, we ought to clarify whether TNF plays any roles in HSV- 1-induced necroptosis. We generated a TNFR1 KO L929 cell line by CRISPR/Cas9 gene editing technique (Figure S3A), and the loss of TNFR1 s function was confirmed by treating the cells with TNF-a, zvad, or TNF-a plus zvad (Figure S3B). WT and TNFR1 KO L929 cells were infected with 20 moi, and their viability was measured from 0 to 24 hr (Figure 3A). The cell viability of TNFR1 KO L929 cells only increased slightly, excluding the possibility that -induced necroptosis is an autocrine effect of TNF. To examine whether other cytokines 230 Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc.

4 A 1500 WT L KO h 18h 24h Infection Time 36h B Viral titer( 10 3 pfu/ml) 0h 18h 200 KO+Vector KO h Infection Time 36h C WT L929 KO D 1 moi(h) KO+Vector KO+ gd SE 1 moi(h) gd LE gd GAPDH GAPDH E GFP- 42h (1 moi) WT L929 KO F KO+Vector GFP- 42h (1 moi) KO+ GFP GFP Light Light Figure 1. Is a Cellular Factor that Restricts Propagation (A) deficiency leads to elevation of propagation. WT and KO L929 cells were infected with (1 moi) for 2 hr, washed twice with PBS, and replaced with fresh medium. Cells together with the supernatants were harvested at 18, 24, and 36 hr after infection. After three repeated freeze-thaw cycles, the samples were centrifuged at 6,000 rpm for 30 min, and then the supernatants were collected to determine the viral titer by a standard viral plaque assay. pfu, plaque-forming units. (B) Reconstitution of expression in KO L929 cells restores the restriction in replication. -reconstituted KO L929 cells and the control cells were infected with (1 moi) over the indicated time periods. The viral titers were measured as described in (A). (C) The accumulation of gd in KO L929 cells is higher than that in WT cells. WT and KO L929 cells were infected with (1 moi) over the indicated time periods, and the whole-cell lysates were subjected to SDS-PAGE and analyzed by immunoblotting with the indicated antibodies. SE, short exposure; LE, long exposure. (D) The same as in (C), except -reconstituted KO cells and control cells were used. (E) WT L929 cells are more resistant to GFP- infection compared with KO L929 cells. WT and KO L929 cells were infected with GFP- (1 moi) for 42 hr and then imaged using a fluorescent microscope. The top panel is light microscopy and the lower panel is fluorescent microscopy (GFP). Scale bar, 5 mm. (F) The same as in (E), except -reconstituted KO cells and control cells were used. Data are represented as mean ± SD of triplicate samples (A and B), and similar results were obtained in 3 5 independent experiments (A F). See also Figure S1. were involved, we collected conditional media from the cell culture of L929 at different time points after infection, removed by ultracentrifugation, and then used these conditional media to treat newly prepared L929 cells. The conditional media did not induce necroptosis (Figure S3C). Since RIP1 and MLKL are also important in TNF-induced necroptosis, we determined whether RIP1 and MLKL were also involved in -induced necroptosis. RIP1 and MLKL KO L929 cells Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. 231

5 A B WT KO Mock Mock VP16 * β-actin C D E GFP- (1 moi) 10h 14h 18h 22h 26h 30h KO WT F G 120 L929 WT L929 KO Survival(%) Mock z-vad +z-vad Figure 2. Infection Leads to -Dependent Cell Death (A) Comparable entry into WT and KO cells as indicated by genomic DNA levels. WT and KO L929 cells were infected with or without at 200 moi for 2 hr in the presence of CHX (10 mg/ml). After being washed twice with PBS, the cells were incubated with 0.25% trypsin and 0.025% EDTA for 5 min to remove viruses absorbed to the cell surface. Then the cells were harvested and genomic DNA was isolated for qpcr analysis of gd, ICP22, and ICP47 genomic DNA levels. (B) The experiment was performed similarly as in (A), except that the cell lysates were analyzed by immunoblotting with anti-vp16 antibody. *, non-specific band. (C) does not replicate more in KO cells during the first life cycle of infection. WT and KO L929 cells were infected with (1 moi) for 2 hr, washed with PBS twice, and then changed with fresh medium. At the indicated hours post-infection, cells were harvested and the genomic DNA was extracted for qpcr analysis of GD, ICP22, and ICP47 genomic DNA levels. (D) WT and KO L929 cells were infected with GFP- (1 moi) over the indicated time periods, and the percentage of GFP-positive cells was measured by flow cytometry analysis. (E) Infection of causes cell death. Shown are representative time-lapse images (merged) of WT and KO L929 cells upon GFP- (1 moi) infection. Cells were incubated with PI, infected with GFP-, and then monitored by a fluorescent microscope. Green and red fluorescence denote GFP- ( infected) and PI-positive (dead) cells, respectively. The arrows in the same color between the adjacent images point at the same cell. Scale bar, 2 mm. 232 Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. (legend continued on next page)

6 A B C D h 6h 12h 18h 24h 40 0 GAPDH High-Dose Time KO L929 WT L929 TNFR1 KO MLKL-3 Flag G Mock E F IP: WCL Flag- Flag-2A 0h 5h 10h 0h 5h 10h P- IP:Flag Flag WCL RIP1 MLKL Flag RIP1 MLKL GAPDH KO WT-L929 0h 2h 4h 6h 8h 10h Flag DAPI Phase Merged Figure 3. -Induced Necroptosis Is Independent of TNFR1, Partially Dependent on RIP1, and Fully Dependent on and MLKL (A) WT and TNFR1 KO L929 cells were infected with high-dose (20 moi) over the indicated time course, and cell viability was determined by PI exclusion. (B) WT, KO, RIP1 KO, and MLKL KO L929 cells were infected with high-dose (20 moi) for 20 hr, and the cell viability was measured. (C) Viability of WT, KO, RIP1 KO, and MLKL KO L929 cells upon (20 moi) plus zvad (20 mm) treatment over the indicated time periods. (D) -induced cell death is dependent on the RHIM domain, kinase activity, and T231/ S232 phosphorylation. KO L929 cells that were reconstituted with an empty vector, WT, RHIM-mutated (RHIM mut ), kinase-dead (D143N), or T231/S232 phosphorylation sitemutated (2A) were infected with high-dose (20 moi) for 15 hr. The cell viability was determined by PI exclusion (lower panel). Upper panel shows the protein amounts of mutants expressed in KO L929 cells. (E) infection leads to phosphorylation on T231/S232 sites. KO L929 cells stably expressing Flag-tagged WT or T231A/S232A (2A)- mutated were infected with (20 moi) for 0, 5, and 10 hr. The cell lysates were immunoprecipitated with anti-flag M2 beads, then the phosphorylation state of (P-) and protein levels were determined by immunoblotting with anti-phospho- (T231/S232) and anti- antibodies, respectively. (F) infection induces necrosome formation. WT L929 cells were infected with high-dose (20 moi) for 0, 2, 4, 6, 8, and 10 hr, cells were lysed to perform immunoprecipitation with anti- antibody-coupled protein A/G-sepharose beads. The levels of RIP1 and MLKL in necrosome were determined by immunoblotting with corresponding antibodies. IP, immunoprecipitates; WCL, whole-cell lysates. (G) MLKL is translocated to the plasma membrane upon infection. MLKL KO L929 cells reconstituted with C-terminal 3 3 Flag-tagged MLKL were treated with or without (20 moi) for 12 hr, and then immunostained for Flag and counterstained with DAPI. Among 52 monitored infected cells, 36 of them exhibited MLKL plasma membrane localization as presented. Scale bar, 4 mm Data are represented as mean ± SD of triplicate samples (A D) and were representative of three independent experiments. See also Figure S3. were generated previously (Chen et al., 2014) (Figure S3D). Like KO cells, MLKL deficiency totally blocked -induced necroptosis (Figure 3B). However, RIP1 deletion only partially blocked -induced necroptosis (Figure 3B). To exclude the possibility that the cell viability was affected by apoptosis, we included zvad throughout infection and obtained the same results (Figure 3C). The viral propagation was also elevated in the RIP1 KO or MLKL KO cells when the cells were infected with 1 moi (Figure S3E). The partial dependence on RIP1 in -induced necroptosis suggests that part of -induced necroptosis is through RIP1 to initiate - dependent cell death. RHIM domain-dependent RIP1- hetero-interaction, - homo-interaction, and s autophosphorylation on its T231/S232 are the sequential steps in TNF-induced necroptosis (Chen et al., 2013; Cho et al., 2009; He et al., 2009; Wu et al., 2014; Zhang et al., 2009). We expressed RHIM-mutated, kinase-dead (D143N), T231/S232-mutated (T231A/S232A or 2A), (F) High-dose infection leads to -dependent cell death. WT and KO L929 cells were infected with high-dose (20 moi) for the indicated time periods, and the cell viability was determined by PI exclusion. (G) zvad does not inhibit but rather enhances -induced cell death. WT and KO L929 cells were treated with medium (mock), zvad (20 mm), (20 moi), or + zvad for 14 hr, and the cell viability was analyzed by PI exclusion. Data are represented as mean ± SD of triplicate samples (A, C, D, F, and G), and similar results were obtained from three independent experiments. See also Figure S2 and Movies S1 and S2. Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. 233

7 A C Intensity B D m/z Da G KO + 3 Flag- E F FLAG-ICP6 GAPDH FLAG-ICP6 GAPDH IP: Flag RIP1 MLKL Flag- P- Survival (%) WCL Myc RIP1 MLKL Flag- H I L929 J Survival(%) 0h 6h 12h 18h h 10h 20h 30h WT L929+ BAC (F strain) WT L929+ ICP6 RHIM mut KO+ BAC (F strain) KO+ ICP6 RHIM mut Time (h) Figure 4. ICP6 Protein of Is a Trigger of Necroptosis (A) Time-lapse effect of ActD on -induced cell death. WT L929 cells were infected with, and ActD was added at different time points after infection. At 15 hr post-infection, cells were harvested and cell viability was determined. (B) Time-lapse effect of CHX on -induced cell death. The experiment was performed as in (A), except that CHX was used. (C) Identification of ICP6 as a binding protein with the highest affinity for among viral proteins. KO L929 cells stably expressing Flag-tagged were infected with (20 moi) for 8 hr; cells were lysed and subjected to immunoprecipitation with anti-flag M2 beads. The eluted complexes were then analyzed by mass spectrometry to identify proteins in the complex. The most abundant viral peptide in the complex is shown. (D) Sequence alignment of the RHIM domains of murine RIP1,, and ICP6. The four amino acid motifs crucial for RHIM function are boxed. (E) ICP6 overexpression induces L929 cell death in an RHIM-dependent manner, and zvad cannot inhibit but rather enhances ICP6-induced cell death. L929 cells were infected with lentivirus encoding nothing (Vector), ICP6, or RHIM-mutated ICP6 (ICP6 RHIM mut ). zvad was included in one of the ICP6 overexpression samples. The cell viability was determined over the indicated time course (lower panel). The protein amounts of Flag-tagged WT and RHIM-mutated ICP6 were shown in the upper panel. (F) Viability of WT, KO, RIP1 KO, and MLKL KO L929 cells upon ICP6 overexpression (lower panel). Upper panel showed ICP6 protein amounts. (legend continued on next page) 234 Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc.

8 or WT in KO L929 cells and found that none of these mutants can act as WT to restore the sensitivity of KO L929 cells to -induced necroptosis (Figure 3D), indicating that RHIM domain, kinase activity, and T231/S232 phosphorylation of are necessary for -induced necroptosis. We also showed that induces T231/S232 phosphorylation (Figure 3E) and the RHIM mutant, kinase dead, or 2A mutant cannot suppress propagation in KO L929 cells (Figure S3F). infection induces necrosome formation, as RIP1 and MLKL were detected in the immunocomplex from L929 cells infected with high-dose (Figure 3F). infection also leads to MLKL translocation to plasma membrane (Figure 3G). Summing up the above data, the mechanism of -induced necroptosis is the same as that of TNF-induced necroptosis downstream of, whereas the upstream signaling pathways of these two appear to be different. Protein ICP6 Is a Trigger of Necroptosis is a double-stranded DNA virus that might be sensed by DAI (Takaoka et al., 2007), cyclic GMP-AMP synthase (cgas) (Sun et al., 2013), and possibly RIG-I (Choi et al., 2009). However, -induced necroptosis is not sensed by DAI, cgas, and RIG-I, because the sensitivities of DAI KO, cgas KO, and RIG-I KO L929 cells to -induced necroptosis are the same as that of WT L929 cells (Figure S4). Although DAI may not sense dsdna in other cells (Ishii et al., 2008), its deletion in L929 cells abolished dsdna-induced IFN-b expression (Figure S4F) (Takaoka et al., 2007). Since viral proteins were accumulated in infected WT cells before the rupture of plasma membrane (Movie S1), the accumulation of viral protein(s) might be required for -induced necroptosis. To test this hypothesis, we treated WT L929 with high-dose in the presence of actinomycin D (ActD) to block mrna or CHX to block protein synthesis. Impressively, ActD or CHX treatment completely blocked - induced necroptosis (Figures 4A and 4B), indicating that newly synthesized protein(s) is/are required for to induce necroptosis. As the deletion of cgas impaired -induced type I interferon expression, but not cell death (Figures S4D and S4G), we thought that viral protein(s), but not HSV- 1-induced cellular protein(s), is/are likely to be required for -induced necroptosis. viral proteins can be divided into three sets that are produced in three sequential rounds of transcription: a (immediate early), b (early), and g (late) proteins (Whitley and Roizman, 2001). Adding ActD at 3 hr or later time points after infection cannot block -induced necroptosis (Figure 4A). Similarly, addition of CHX at 4 hr or later time cannot block - induced necroptosis (Figure 4B). Considering that it costs certain time for the drugs to reach their targets, the viral gene product(s) that can induce necroptosis should be a or b protein(s). Because -induced necroptosis could result from directly targeting, we performed mass spectrometry analysis of immunocomplex isolated from -infected L929 cells and identified ICP6 as the highest-affinity -binding protein among all viral proteins (Figure 4C). Noticeably, ICP6 is a leaky b protein (Goldstein and Weller, 1988b) encoded by UL39 gene and contains an RHIM-like domain (Figure 4D) (Lembo and Brune, 2009). A significant amount of ICP6 has been synthesized 4 hr after infection, and the amount of ICP6 correlated well with -induced necroptosis (Figures S5A and S5B). To verify the involvement of ICP6 and its RHIM domain in - induced necroptosis, Flag-tagged WT ICP6 and RHIM-mutated ICP6 were delivered into L929 cells by a lentiviral vector. The overexpression of ICP6 resulted in L929 cell death, whereas the overexpression of its RHIM mutant did not (Figure 4E). The ICP6- induced cell death was not inhibited but enhanced by zvad, indicating that the cell death is necroptosis (Figure 4E). ICP6-induced cell death has nothing to do with TNF autocrine (Figure S5C). We further tested the effect of ICP6 overexpression in RIP1 KO, KO, and MLKL KO L929 cells and found that as with infection, shown in Figures 3B and 3C, ICP6-induced necroptosis was partially blocked by RIP1 deletion and completely blocked by or MLKL deletion (Figure 4F). The necrosome formation and phosphorylation at T231/S232 were also observed in ICP6- induced cell death (Figure 4G). Thus, ICP6 overexpression mimics infection in the induction of necroptosis. To confirm that ICP6 is indeed required for to induce necroptosis, we employed a mutant termed ICP6D (Goldstein and Weller, 1988a) whose ICP6 gene was deleted. We infected WT and KO L929 cells with or ICP6D and measured their cell viabilities and gd protein levels. The deletion of ICP6 significantly reduced -dependent necroptosis (Figure 4H). Meanwhile, propagation significantly increased (Figure S5D). To determine the function of ICP6 RHIM domain in -induced necroptosis in vivo, we generated ICP6 RHIM mutant via p BAC (F strain) (Li et al., 2011). Unlike ICP6D, the propagation ability of this mutant virus (ICP6 RHIM mut ) is comparable to that of WT control ( BAC [F strain]). The ICP6 RHIM mut cannot induce necroptosis in L929 cells (Figure 4I). The mutation of ICP6 RHIM (G) ICP6 overexpression induces necrosome formation and phosphorylation at T231/S232 sites in an RHIM-dependent manner. KO L929 cells stably expressing 3 3 Flag- were infected with lentiviral virus encoding nothing (Vector), Myc-tagged ICP6, or Myc-tagged RHIM-mutated ICP6 in the presence of zvad for 30 hr. Whole-cell lysates were subjected to immunoprecipitation with anti-flag M2 beads. The total cell lysates and immunoprecipitates were immunoblotted with the indicated antibodies. (H) ICP6 deletion in impairs s ability to induce L929 cell death. WT and KO L929 cells were infected with (20 moi) and ICP6D (50 moi) over the indicated time periods, and cell viability was determined. A higher dose of ICP6D was used because under this experimental condition the expression of gd in ICP6D-infected KO cells is at a level comparable with that in -infected KO cells (Figure S5D). (I) ICP6 RHIM mutation (RHIM mut ) in impairs s ability to induce L929 cell death. WT and KO L929 cells were infected with BAC (F strain) (20 moi) and ICP6 RHIM mut (20 moi) over the indicated time periods, and cell viability was determined. (J) Viral titer in the samples described in (I) is shown. Samples were prepared as described in Figure 1A, and viral titer was determined by a standard viral plaque assay. Data are represented as mean ± SD of triplicate wells (A, B, E, F, and H J), and are representative of three independent experiments (A, B, and E J). See also Figures S4, S5, S6, and S7. Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. 235

9 domain largely eliminated -mediated restriction on propagation in L929 cells (Figures 4J and S5E). A similar result was obtained when low-dose (1 moi) was used in infection (Figure S5F). Collectively, our data demonstrated that the induction of necroptosis by is largely dependent on the RHIM domain of ICP6. ICP6 Directly Interacts with RIP1 and Since ICP6 RHIM domain is necessary for its overexpressioninduced cell death, we tested whether ICP6 interacts directly with RIP1 and. Flag-ICP6 was expressed in WT, RIP1 KO, or KO L929 cells, and its association with RIP1 and in those cells was determined by co-immunoprecipitation (co-ip). ICP6 interacted with both RIP1 and in WT cells, in RIP1 KO cells, and RIP1 in KO cells (Figure 5A). By expressing RHIM-mutated ICP6 (ICP6 RHIM mut ), N-terminal domain (ND), ND with RHIM mutation (ND RHIM mut ), and C-terminal R1 homology domain (RD) (Figure 5B), as well as Myc-RIP1 or Myc- in different combinations in HEK293T cells, we show that the RHIM domain but not the RD in ICP6 is indispensable for ICP6 to interact with RIP1 and (Figures 5C and 5D), and the RHIM domain in RIP1 and is also required for their interaction with ICP6 (Figures 5E and 5F). ICP6 Homo-Interaction Initiates the Necroptotic Process Although the ND only of ICP6 can interact with RIP1 and (Figures 5C and 5D), we found that none of the ND, ND RHIM mut, and RD overexpression can initiate necroptosis in L929 cells (Figure 6A). As it is known that TNF-induced necrosis is initiated by the formation of RIP1- hetero-interaction and then - homo-interaction (Wu et al., 2014), we reasoned that ICP6 might form dimers/oligomers and then recruit RIP1 and or and to initiate a similar necroptotic process. To test this possibility, we co-expressed Flag-tagged ICP6 or truncated ICP6 with Myc-tagged ICP6 in HEK293T cells and analyzed the protein interactions between ICP6 and ICP6 or its mutants. We found that Flag-ICP6 interacted with Myc-ICP6 in a C-terminal RD- but not RHIM-dependent manner (Figure 6B). Consistently, RD alone is sufficient to execute RD homo-interaction (Figure 6C). Since neither the ND nor RD alone is sufficient to trigger necroptosis (Figure 6A), it is possible that the RD in ICP6 is responsible for the interaction between ICP6 proteins and the ND can initiate necroptotic signal from ICP6 complex. To test this hypothesis, we determined whether artificial dimers of ICP6 ND would lead to necroptosis. We used 4-hydroxytamoxifen (4-OHT) to induce the hormone-binding domain G521R (HBD*)-based dimer formation to address the question (Chen et al., 2014). We fused HBD* to the C terminus of ICP6 ND (termed ND-HBD*) or ICP6 ND RHIM mutant (termed ND RHIM mut -HBD*) and expressed them in TNFR1 KO L929 cells, in which the influence of TNF autocrine on cell death can be well avoided. After 4-OHT or 4-OHT plus zvad being added, TNFR1 KO cells expressing ND-HBD* underwent cell death, whereas the control cells did not, demonstrating that HBD*-mediated interaction of ICP6 ND is sufficient to trigger cell death (Figure 6D). As anticipated, the RHIM domain of ND is required for its function. Taken together, these data suggest that the formation of dimers or oligomers of ICP6s initiates RIP1- hetero- and - homo-interaction. -Induced Necroptosis Is Initiated through Both RIP1-Dependent and RIP1-Independent Mechanisms We then explored the partial dependence of -induced necrosis on RIP1. As ICP6 can directly interact with both RIP1 and, an ICP6 dimer/oligomer might recruit RIP1 and to form RIP1- interaction. In the meantime, an ICP6 dimer/ oligomer might directly recruit more than one to initiate - interaction. Indeed, sequential immunoprecipitations detected both ICP6-RIP1- and ICP6-- complexes (Figure 7A). The co-existence of ICP6-RIP1- and ICP6-- complexes still cannot explain well the partial dependence on RIP1 in -induced necroptosis, because the formation of ICP6-- should be able to largely overcome the effect of RIP1 knockout unless there is a preference of the formation of ICP6-RIP1- complex. To evaluate this possibility, we compared ICP6 s affinity to RIP1 and. Flag-ICP6, HA- RIP1, and HA- were individually expressed in HEK293T cells, and the binding of ICP6 to RIP1 and were compared by an in vitro pull-down assay. When comparable amounts of HA-RIP1 and HA- are used, ICP6 pulls down much more RIP1 than (Figure 7B), indicating that affinity of ICP6 to RIP1 is much higher than to. High affinity between ICP6 and RIP1 makes the formation of ICP6-RIP1- quicker than that of ICP6--, which explains the partial dependence on RIP1 in -induced necroptosis. Restricts Propagation in Mice To determine the requirement of in host defense against infection in vivo, we infected WT and / mice with WT via the intravenous (i.v.) route and found / mice were more sensitive to -induced death (Figure 7C). In contrast, there is no significant difference between WT and / mice in virus-induced death when ICP6D was used (Figure 7D), indicating that the difference in mortality between -infected WT and / mice is an ICP6-dependent pathogenesis process. Consisting with the data of mortality, / mice exhibited more severe body weight loss in comparison with WT mice (Figure 7E); the levels of DNA in the brain of / mice were higher than those of WT mice (Figure 7F). In addition, we also found that / mice exhibited higher viral loads in trigeminal ganglia (TG) than WT mice when the mice were inoculated via cornea (Figure 7G). To examine the role of ICP6 RHIM in pathogenesis in mice, ICP6 RHIM mut and its WT counterpart were used to infect WT mice via i.v. injection. The viral loads in brain and TG were determined (Figures 7H and 7I). ICP6 RHIM mut propagated more efficiently than WT in WT mice. Consistently, ICP6 RHIM mut infection also leads to higher lethality in WT mice (Figure 7J). These data demonstrate that senses RHIM of ICP6 to restrict propagation in mice. DISCUSSION infection leads to -dependent necroptosis in murine cells, which restricts propagation in mice. Our model of ICP6-initiated necroptosis is as follows: ICP6 forms dimers/oligomers via its C-terminal RD; the ICP6 RHIMs in the dimer/oligomer 236 Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc.

10 A B C D E F Figure 5. ICP6 Directly Interacts with RIP1 and (A) ICP6 directly interacts with endogenous RIP1 and. WT, RIP1 KO, and KO L929 cells were infected with lentiviral virus encoding nothing (Vector) or Flag-tagged ICP6 for 36 hr and immunoprecipitated with anti-flag M2 beads. The immunoprecipitates and whole-cell lysates were analyzed by immunoblotting to determine the protein levels of and RIP1. (B) Schematic representation of WT and mutated ICP6. ICP6-RHIM mut, CP6 containing RHIM mutation; ND, N-terminal domain of ICP6; ND-RHIM mut,nd containing RHIM domain mutation; RD, C-terminal R1 homology domain of ICP6. (C) Mapping the RIP1-binding domain in ICP6. HEK293T cells were co-expressed with Myc-tagged mrip1 and Flag-tagged ICP6 or its different mutants for 36 hr. Co-IP assay was performed with anti-flag M2 beads. (D) Mapping the -binding domain in ICP6. The experiments were performed similarly as in (C), except that Myc-tagged m was used. (E) RIP1 interacts with ICP6 in an RHIM-dependent manner. Flag-tagged WT or RHIM-mutated mrip1 was co-expressed with Myc-tagged ICP6 in HEK293T cells for 36 hr. Co-IP assay was performed with anti-flag M2 beads. (F) interacts with ICP6 in an RHIM-dependent manner. The same as in (E), except that Flag-tagged m was used instead of Flag-tagged mrip1. recruit RIP1 and or two to result in RIP1- heterointeraction or - homo-interaction; RIP1- heterointeraction favors further recruitment of additional and thus also leads to - homo-interaction (Wu et al., 2014); and the pathway downstream of is the same as that in TNF-induced necroptosis, which involves recruiting MLKL by Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. 237

11 A B Flag-ICP6 Myc-ICP6 Flag-ICP6 Flag-RD Flag-ND IP:Flag Vector ICP6 ICP6-RHIM mut ND ND-RHIM mut RD Myc-ICP6 WCL Vector ICP6 ICP6-RHIM mut ND ND-RHIM mut RD Myc-ICP6 Flag-ICP6 Flag-ICP6 Flag-RD Flag-ND Actin C IP: Myc WCL Vector ICP6 ICP6-RHIM mut ND RD ND-RHIM mut Flag Myc Flag Myc Flag-RD Vector Myc-RD Figure 6. ICP6 Homo-Interaction Initiates Necroptotic Process (A) The abilities of different ICP6 mutants in inducing cell death. L929 cells were infected with lentiviruses encoding nothing (Vector), ICP6, ICP6 RHIM mut, ND, ND RHIM mut, or RD for 60 hr, and the cell viability was determined (left panel). The protein amounts of ICP6 and its mutants were determined by immunoblotting (right panel). (B) ICP6 forms dimers or oligomers via its C-terminal RD. Flag-tagged ICP6 or its mutant was coexpressed with Myc-tagged ICP6 in HEK293T cells for 36 hr. Co-IP assay was performed with anti-flag M2 beads. (C) C-terminal RD alone is sufficient to form dimers/oligomers. Flag-tagged RD of ICP6 was coexpressed with nothing or Myc-tagged RD in HEK293T cells for 36 hr. Co-IP assay was performed with anti-myc beads. (D) Artificially forced dimerization of ICP6 ND leads to necroptotic cell death in an RHIM-dependent manner. TNFR1 / L929 cells stably expressing Flag-tagged ND-HBD* or ND RHIM mut -HBD* were treated with EtOH (ethanol, control), EtOH + zvad, 4OHT (4-hydroxytamoxifen), and 4OHT + zvad for 72 hr, and the cell viability was determined (left panel). The protein amounts of Flag-tagged ND- HBD* and ND RHIM mut -HBD* (right panel) were determined by immunoblotting. Data are representative of the mean ± SD of triplicate wells (A and D), and similar results were obtained from three independent experiments. D Vector TNFR1 KO L929 ETOH 4-OHT ND-HBD* ETOH+z-VAD 4-OHT+z-VAD ND RHIM mut -HBD* Flag Actin and the translocation of MLKL to plasma membrane and subsequent membrane disruption (Figure S7A). ICP6 in has been shown to be important for propagation especially in non-dividing cells (Goldstein and Weller, 1988a, 1988b). ICP6D replicated more slowly than WT in culture cells, and the growth rate of ICP6D can be restored in cells expressing ICP6. The RHIM domain in ICP6 is dispensable for propagation because the propagation of ICP6 RHIM mut is the same as WT, and the replication rate of ICP6D in cells expressing RHIM-mutated ICP6 is comparable to that in cells expressing WT ICP6 (Figure S5G). The catalytic activity of ICP6 is important for propagation because the activity-deficient mutant of ICP6 (ICP6-C793A) cannot restore the growth rate of ICP6D (Figure S5G). Overexpression of ICP6-C793A still can induce cell death in L929 cells, indicating that ICP6 s enzymatic activity is not required for ICP6-induced necrosis (Figure S5H). ICP6 can block DR-induced apoptosis by interacting with caspase-8 (Dufour et al., 2011). We were able to see the inhibition of caspase-8 by ICP6 in L929 cells (data not shown). Since inhibition of caspase-8 promotes necroptosis, the inhibiting effect on casapse-8 by ICP6 should further enhance necroptosis. Although ICP6-induced necroptosis does not require DAI (Figure S4G), ICP6 still might have functional linkage with DAI, since ICP6 can inhibit DAI overexpression-induced NF-kB activation in an RHIM-dependent manner (data not shown) and DAI can suppress replication (Pham et al., 2013). ICP10, the R1 subunit of HSV-2 RNR, also contains an RHIM-like domain and an R1 domain (Lembo and Brune, 2009). Similar to, HSV-2 can also induce -dependent cell death (Figure S2G). Different from ICP6, the viral inhibitor of RIP activation (vira), which is an RHIM-containing protein encoded by the M45 gene of MCMV, has been reported to be catalytically inactive (Lembo and Brune, 2009). It is important to note that ICP6 overexpression triggers necroptosis in L929 cells whereas vira does not 238 Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc.

12 A Flag-ICP6 HA-mRIP1 HA-m Myc-m WCL 1 st IP Anti-Flag Flag Myc HA Flag Myc B Vector Flag-ICP6 HA-mRIP1 HA-m Flag-ICP6 IP: Flag HA-mRIP1 HA-m * C D * 2 nd IP Anti-HA HA Flag Myc HA Flag-ICP6 WCL HA-mRIP1 HA-m Survival(%) ns E F G ** ** * +/+ -/- H BAC F strain (n=9) Brain ** ICP6 RHIM mut (n=12) I Relative Genomic DNA level -BAC F strain (n=9) * ICP6 RHIM mut (n=12) J * Figure 7. -Induced Necroptosis Is Initiated through Both RIP1-Dependent and Independent Mechanisms, and Restricts Propagation in Mice (A) ICP6 can initiate both RIP1- hetero-interaction and - homo-interaction. HEK293T cells were transfected with expression vectors as indicated, and a sequential immunoprecipitation with anti-flag and anti-ha was performed as described in the Supplemental Experimental Procedures. (B) ICP6 has higher affinity for RIP1 compared with. The in vitro pull-down assay was performed as described in Experimental Procedures. (C) WT and / mice were infected with ( pfu) via i.v. and tested for survival rate. (D) WT and / mice were infected with ICP6D ( pfu) via i.v. and tested for survival rate. ns, no significant difference. (E) WT and / mice were infected with ( pfu) via i.v., and body weight losses (4 days after infection) were recorded. (F) WT and / mice were infected with ( pfu) via i.v., and genomic DNA levels in the brains of infected mice (harvested 3 days after infection) were measured by qpcr analysis. (G) WT and / mice were anesthetized and infected with ( pfu/eye) on the right eyes following scarification (20 times) of the cornea with a needle. Three days after infection, trigeminal ganglia (TG) were isolated from mice and tested for genomic DNA levels by qpcr analysis. (H) WT mice were infected with pfu ICP6 RHIM mut or its WT counterpart via i.v., and genomic DNA levels in the brains of infected mice (harvested 2 days after infection) were measured by qpcr analysis. (I) WT mice were infected with pfu ICP6 RHIM mut or its WT counterpart via i.v., and genomic DNA levels in the trigeminal ganglia (TG) of infected mice (harvested 2 days after infection) were measured as in (F). (J) WT mice were infected with pfu ICP6 RHIM mut or its WT counterpart via i.v. and tested for survival rate. Data are represented as mean ± SEM (E I). *p < 0.05, **p < 0.01 (Student s t test). The survival curve was generated by Kaplan-Meier methods. Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. 239

13 (Figure S6A). While vira can disrupt the RHIM-dependent RIP1- kinase complex (Upton et al., 2010) and thus inhibits TNFinduced necroptosis in L929 cells, ICP6 cannot (Figure S6B). Similar to ICP6, vira forms dimers/oligomers independent of the RHIM domain (Figure S6C) and interacts with via the ND (1 278) in an RHIM-dependent manner (Figure S6D). Different from ICP6, vira exhibits an inhibitory effect on the formation of murine RIP1-, -, or -MLKL complexes (Figure S6E). Consistently, when vira was co-expressed with murine in HEK293T cells, the autophosphorylation of was markedly reduced (the shift bands of and p- in lanes 2 and 8 of the whole-cell lysates shown in Figure S6D). ICP6 s enhancing effect on phosphorylation (band shift) is difficult to detect in this system, since ectopically expressed murine in 293T cells already has basal phosphorylation. We found that the difference between ICP6 and vira in inducing necroptosis in L929 cells lies in their distinct RHIM domains, since the change of the RHIM domains of ICP6 and vira (ICP6 vira-r and vira ICP6-R ) with each other (Figure S6F) blocked ICP6 s function to induce necrosis while enabling vira to do so (Figure S6G). These data demonstrate that there are at least two different RHIM-RHIM interactions that play opposite roles in -mediated necroptosis. The anti-necroptosis RHIM could be evolved from the pro-necroptosis RHIM, since this change allows the virus to evade the restriction by necroptosis. While our manuscript was under revision, Wang et al. reported that can trigger /MLKL-dependent necrosis (Wang et al., 2014b), which is similar to our observation in mouse cells. However, the authors of that report overlooked the fact that infection did not induce -dependent necroptosis in human cells. In addition to HT-29 cells, we also obtained data from another human cell line, HeLa. HeLa cells do not express and undergo apoptosis upon TNF treatment. Since ectopic expression of in HeLa cells converts TNF-induced apoptosis to necroptosis (He et al., 2009), we infected -expressing and control HeLa cells with and found that the death of -expressing cells was slightly more than that of control cells. However, the enhancing effect of on cell death was independent of ICP6, because the same result was obtained when ICP6D was used (Figure S2H). Since the autocrine effect of TNF upon infection contributes a little in - induced cell death (Figure 3A), we speculated that the expression of in HeLa cells enhanced TNF autocrine-induced cell death. The data obtained using HeLa cells support the conclusion that cannot induce necroptosis in human cells. ICP6 overexpression cannot induce cell death in HT-29 cells, but we still detected RHIM-dependent interaction between ICP6 and human RIP1 or human (data not shown), suggesting that the interaction with ICP6 does not promote human RIP1 and/or human to mediate necroptotic signals. Moreover, ICP6 can depend on the RHIM domain to inhibit TNF-induced necroptosis in human HT-29 cells (Figure S7B), which is contrary to the result in mouse L929 cells (Figure S7C). Consistently, pre-treatment suppresses TNF-induced necrosis in HT- 29 cells (Figure S7D). In a companion to our study, Guo et al. reported the same function of ICP6 in human cells (Guo et al., 2015). The minor structural differences between human and mouse RHIM domains should be responsible for the pro- or anti-necroptosis effect of ICP6. It is known that the propagation of in mouse cells is less efficient than in human cells. The induction of necroptosis by ICP6 in mouse but not in human cells might partly account for this phenomenon. Since is a natural human pathogen, we speculate that the RHIM of has already evolved to evade human RIP1 and human -mediated necroptosis, but has not done so to evade mouse cells. Sensing ICP6 by RIP1 and might also serve as a species barrier to limit s natural infection in mouse organism. EXPERIMENTAL PROCEDURES Cell and Cell Lines L929, HEK293T, HeLa, and Vero cells were obtained from ATCC. RIP1 KO, KO, MLKL KO, and 3 3 Flag-tagged MLKL-reconstituted MLKL KO L929 cells were established as described (Chen et al., 2014). KO HT-29 cells, TNFR1 KO, DAI KO, cgas KO, and RIG-I KO L929 cells were generated by using CRISPR/Cas9 or TALEN gene editing technique as described in the Supplemental Information. The KO cells were determined by the sequencing of targeted loci. +/+ and / MEFs were isolated from mouse embryos at 13.5 dpc. NIH 3T3 A and N cells were described in our previous work (Zhang et al., 2009). Preparations of Viruses (KOS strain) and HSV-2 (G strain) were obtained from ATCC, and GFP- (F strain) was generated as previously described (Li et al., 2011); all were propagated in Vero cells. ICP6D (KOS strain) and its WT counterpart were provided by Dr. Sandra Weller (University of Connecticut Health Center). ICP6D was propagated in Vero cells stably expressing ICP6. ICP6 RHIM mut was generated via p BAC (F strain) as previously described (Li et al., 2011). VSV- and NDV-GFP were provided by Dr. Zhengfan Jiang (Peking University). Reagents, Antibodies and Constructs The detailed information of reagents, antibodies, and constructs is described in the Supplemental Experimental Procedures. Cell Viability Assay The cell viabilities of L929, NIH 3T3, HT-29, and HeLa cells were determined by flow cytometry with two parameters: plasma membrane integrity and cell size. The plasma membrane integrity was tested by the ability of cells to exclude PI. Cells were trypsinized and collected with the dead cells in the media by centrifugation, washed once with PBS, and resuspended in PBS containing 5 mg/ml PI. The levels of PI incorporation were quantified on a FACScan flow cytometer. The cell viability of MEFs was determined by MTT assay as described in the Supplemental Experimental Procedures. and HSV-2 Titration For the quantification of and HSV-2, viral plaque assay was performed on Vero cells. The detailed method is described in the Supplemental Experimental Procedures. Confocal Microscopy Confocal microscopy analysis was performed as previously described (Chen et al., 2014). Sequential Immunoprecipitation HEK293T cells were transfected as indicated in Figure 7A for 36 hr. Then cells were harvested and the whole-cell lysates were subjected to the first round of immunoprecipitation with anti-flag M2 beads. After this, the immunoprecipitates were eluted with the 3 3 Flag peptide and diluted with 1 ml of the lysis buffer. This solution was then subjected to the second round of immunoprecipitation with anti-ha beads, followed by immunoblotting with corresponding antibodies. In Vitro Pull-Down Assay HEK293T cells were transfected with an expression vector encoding nothing (empty vector), Flag-ICP6, HA-mRIP1, or HA-m for 36 hr. The cells 240 Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc.

14 were lysed and the whole-cell lysates from the cells expressing HA-mRIP1 or HA-m were diluted with the lysates from empty vector-transfected cells to equalize the concentration of HA-mRIP1 and HA-m in the two cell lysates. Equal volumes of the equalized HA-mRIP1- and HA-m-containing cell lysates were mixed with the cell lysates from the cells transfected with vector or Flag-ICP6-expressing plasmid, as indicated in Figure 7B. The mixed cell lysates were then subjected to immunoprecipitation with anti-flag M2 beads, followed by immunoblotting analysis. Viral Infection in Mice Mice were housed in a specific pathogen-free environment. All experiments were conducted in compliance with the regulations of Xiamen University. Mice were infected i.v. with. The detailed method is described in the Supplemental Experimental Procedures. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures, seven figures, and two movies and can be found with this article online at AUTHOR CONTRIBUTIONS Z.H., Y.L., and J.H. conceived and designed the experiments. Z.H., S.-Q.W., Y.L., X.Z., W.C., L.L., J.W., Q.Z., C.C., J.L., C.-Q.Z., W.X., R.Z., and C.Z. performed the experiments. Z.H., S.-Q.W., Y.L., X.Z., and J.H. analyzed the data. Y.L. and J.H. wrote the paper. ACKNOWLEDGMENTS We thank Dr. Sandra Weller for ICP6 null mutant and Drs. Bernard Roizman and Edward Mocarski for helpful discussion. This work was supported by the 973 Program 2015CB553800, the National Major Project 2013ZX , the NSF of China ( , , , ), the 111 Project B12001, Funding from Xiamen City (No. 3502Z ), and the NSF of China for Fostering Talents in Basic Research (Grant No. J ). Received: June 23, 2014 Revised: November 6, 2014 Accepted: December 23, 2014 Published: February 11, 2015 REFERENCES Cai, Z., Jitkaew, S., Zhao, J., Chiang, H.C., Choksi, S., Liu, J., Ward, Y., Wu, L.G., and Liu, Z.G. (2014). Plasma membrane translocation of trimerized MLKL protein is required for TNF-induced necroptosis. Nat. Cell Biol. 16, Chen, W., Zhou, Z., Li, L., Zhong, C.Q., Zheng, X., Wu, X., Zhang, Y., Ma, H., Huang, D., Li, W., et al. (2013). Diverse sequence determinants control human and mouse receptor interacting protein 3 () and mixed lineage kinase domain-like (MLKL) interaction in necroptotic signaling. J. Biol. Chem. 288, Chen, X., Li, W., Ren, J., Huang, D., He, W.T., Song, Y., Yang, C., Li, W., Zheng, X., Chen, P., and Han, J. (2014). Translocation of mixed lineage kinase domainlike protein to plasma membrane leads to necrotic cell death. Cell Res. 24, Cho, Y.S., Challa, S., Moquin, D., Genga, R., Ray, T.D., Guildford, M., and Chan, F.K. (2009). Phosphorylation-driven assembly of the RIP1- complex regulates programmed necrosis and virus-induced inflammation. Cell 137, Choi,M.K.,Wang,Z.,Ban,T.,Yanai,H.,Lu,Y.,Koshiba,R.,Nakaima, Y., Hangai, S., Savitsky, D., Nakasato, M., et al. (2009). A selective contribution of the RIG-I-like receptor pathway to type I interferon responses activated by cytosolic DNA. Proc. Natl. Acad. Sci. USA 106, Christofferson, D.E., and Yuan, J. (2010). Necroptosis as an alternative form of programmed cell death. Curr. Opin. Cell Biol. 22, Dondelinger, Y., Declercq, W., Montessuit, S., Roelandt, R., Goncalves, A., Bruggeman, I., Hulpiau, P., Weber, K., Sehon, C.A., Marquis, R.W., et al. (2014). MLKL compromises plasma membrane integrity by binding to phosphatidylinositol phosphates. Cell Rep. 7, Dufour, F., Sasseville, A.M., Chabaud, S., Massie, B., Siegel, R.M., and Langelier, Y. (2011). The ribonucleotide reductase R1 subunits of herpes simplex virus types 1 and 2 protect cells against TNFa- and FasL-induced apoptosis by interacting with caspase-8. Apoptosis 16, Goldstein, D.J., and Weller, S.K. (1988a). Factor(s) present in herpes simplex virus type 1-infected cells can compensate for the loss of the large subunit of the viral ribonucleotide reductase: characterization of an ICP6 deletion mutant. Virology 166, Goldstein, D.J., and Weller, S.K. (1988b). Herpes simplex virus type 1-induced ribonucleotide reductase activity is dispensable for virus growth and DNA synthesis: isolation and characterization of an ICP6 lacz insertion mutant. J. Virol. 62, Guo, H., Omoto, S., Harris, P.A., Finger, J.N., Bertin, J., Gough, P.J., Kaiser, W.J., and Mocarski, E.S. (2015). Herpes Simplex Virus Suppresses Necroptosis in Human Cells. Cell Host Microbe 17, this issue, Han, J., Zhong, C.Q., and Zhang, D.W. (2011). Programmed necrosis: backup to and competitor with apoptosis in the immune system. Nat. Immunol. 12, He, S., Wang, L., Miao, L., Wang, T., Du, F., Zhao, L., and Wang, X. (2009). Receptor interacting protein kinase-3 determines cellular necrotic response to TNF-alpha. Cell 137, Holler, N., Zaru, R., Micheau, O., Thome, M., Attinger, A., Valitutti, S., Bodmer, J.L., Schneider, P., Seed, B., and Tschopp, J. (2000). Fas triggers an alternative, caspase-8-independent cell death pathway using the kinase RIP as effector molecule. Nat. Immunol. 1, Ishii, K.J., Kawagoe, T., Koyama, S., Matsui, K., Kumar, H., Kawai, T., Uematsu, S., Takeuchi, O., Takeshita, F., Coban, C., and Akira, S. (2008). TANK-binding kinase-1 delineates innate and adaptive immune responses to DNA vaccines. Nature 451, Karasneh, G.A., and Shukla, D. (2011). Herpes simplex virus infects most cell types in vitro: clues to its success. Virol. J. 8, 481. Lembo, D., and Brune, W. (2009). Tinkering with a viral ribonucleotide reductase. Trends Biochem. Sci. 34, Li, Y., Wang, S., Zhu, H., and Zheng, C. (2011). Cloning of the herpes simplex virus type 1 genome as a novel luciferase-tagged infectious bacterial artificial chromosome. Arch. Virol. 156, Lopez, C. (1975). Genetics of natural resistance to herpesvirus infections in mice. Nature 258, Mocarski, E.S., Upton, J.W., and Kaiser, W.J. (2012). Viral infection and the evolution of caspase 8-regulated apoptotic and necrotic death pathways. Nat. Rev. Immunol. 12, Pham, T.H., Kwon, K.M., Kim, Y.E., Kim, K.K., and Ahn, J.H. (2013). DNA sensing-independent inhibition of herpes simplex virus 1 replication by DAI/ ZBP1. J. Virol. 87, Sun, L., Wang, H., Wang, Z., He, S., Chen, S., Liao, D., Wang, L., Yan, J., Liu, W., Lei, X., and Wang, X. (2012). Mixed lineage kinase domain-like protein mediates necrosis signaling downstream of kinase. Cell 148, Sun, L., Wu, J., Du, F., Chen, X., and Chen, Z.J. (2013). Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science 339, Takaoka, A., Wang, Z., Choi, M.K., Yanai, H., Negishi, H., Ban, T., Lu, Y., Miyagishi, M., Kodama, T., Honda, K., et al. (2007). DAI (DLM-1/ZBP1) is a cytosolic DNA sensor and an activator of innate immune response. Nature 448, Upton, J.W., Kaiser, W.J., and Mocarski, E.S. (2010). Virus inhibition of - dependent necrosis. Cell Host Microbe 7, Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc. 241

15 Vandenabeele, P., Galluzzi, L., Vanden Berghe, T., and Kroemer, G. (2010). Molecular mechanisms of necroptosis: an ordered cellular explosion. Nat. Rev. Mol. Cell Biol. 11, Wang, H., Sun, L., Su, L., Rizo, J., Liu, L., Wang, L.F., Wang, F.S., and Wang, X. (2014a). Mixed lineage kinase domain-like protein MLKL causes necrotic membrane disruption upon phosphorylation by. Mol. Cell 54, Wang, X., Li, Y., Liu, S., Yu, X., Li, L., Shi, C., He, W., Li, J., Xu, L., Hu, Z., et al. (2014b). Direct activation of /MLKL-dependent necrosis by herpes simplex virus 1 () protein ICP6 triggers host antiviral defense. Proc. Natl. Acad. Sci. USA 111, Whitley, R.J., and Roizman, B. (2001). Herpes simplex virus infections. Lancet 357, Wu, X.N., Yang, Z.H., Wang, X.K., Zhang, Y., Wan, H., Song, Y., Chen, X., Shao, J., and Han, J. (2014). Distinct roles of RIP1- hetero- and - homo-interaction in mediating necroptosis. Cell Death Differ. 21, Zhang,D.W.,Shao,J.,Lin,J.,Zhang,N.,Lu,B.J.,Lin,S.C.,Dong,M.Q., and Han, J. (2009)., an energy metabolism regulator that switches TNF-induced cell death from apoptosis to necrosis. Science 325, Zhao, J., Jitkaew, S., Cai, Z., Choksi, S., Li, Q., Luo, J., and Liu, Z.G. (2012). Mixed lineage kinase domain-like is a key receptor interacting protein 3 downstream component of TNF-induced necrosis. Proc. Natl. Acad. Sci. USA 109, Cell Host & Microbe 17, , February 11, 2015 ª2015 Elsevier Inc.

16 Cell Host & Microbe Supplemental Information RIP1/ Binding to ICP6 Initiates Necroptosis to Restrict Virus Propagation in Mice Zhe Huang, Su-Qin Wu, Yaoji Liang, Xiaojuan Zhou, Wanze Chen, Lisheng Li, Jianfeng Wu, Qiuyu Zhuang, Chang an Chen, Jingxian Li, Chuan-Qi Zhong, Weixiang Xia, Rongbin Zhou, Chunfu Zheng, and Jiahuai Han

17 Supplemental Information Extended Experimental Procedures Reagents and antibodies Mouse TNFα was purchased from ebioscience. zvad was obtained from Calbiochem. 4- hydroxytamoxifen (4-OHT) and propidium iodide (PI) were purchased from Sigma. All dyes and DAPI were provided by Invitrogen. Mouse anti- gd (sc-21719), mouse anti- VP16 (sc-7545), mouse anti-ha (F-7), rabbit anti-myc and mouse anti-gapdh antibodies were purchased from Santa Cruz Biotechnology, Inc. Mouse anti-flag M2 and mouse anti-ha beads, mouse anti-flag (M2) and mouse anti-β-actin (C-15) antibodies were purchased from Sigma. Mouse anti-rip1 antibody was obtained from BD Biosciences. Anti-MLKL and anti-icp6 polyclonal antibodies were raised in rabbits using E. coli-expressed GST-MLKL ( amino acids) and GST-ICP6 ( amino acids) respectively. Anti- and anti-phospho- (T231/S232) antibodies were previously described (Chen et al., 2013; Zhang et al., 2009). Rabbit anti-human (h) antibodies were purchased from Abcam. The dsrna analog poly (I:C) was purchased from InvivoGen. The double-stranded DNA of VACV (dsvacv) and ISD were prepared as previously described (Stetson and Medzhitov, 2006; Unterholzner et al., 2010). Constructs The ICP6 cdna was PCR-amplified from a reverse-transcribed cdna library derived from L929 cells infected with (20 moi) for 10 hours. The MCMV vira cdna was from MCMV infected L929 cell cdna library. Full-length or mutated cdnas of ICP6 and vira were cloned into BamHI and XhoI sites of the modified lentiviral vector pbob using the Exo IIIassisted ligase-free cloning method. ND-HBD* of ICP6 and its RHIM mutant were created by 1

18 PCR recombination and cloned into the pbob vector. RIP1, and MLKL cdnas were also cloned into the pbob vector. All plasmids were verified by DNA sequencing. The details of the sequences are available upon request. Lentivirus preparation and infection For lentivirus production, HEK293T cells were transfected with lentiviral vectors and viruspacking plasmids by calcium phosphate precipitation. The virus-containing medium was harvested hours later and added to the L929 cells as indicated with 10 µg/ml of polybrene. Infectious medium was changed with fresh medium 12 hours later. MTT Assay For MEF cells, cell survival rates were determined by MTT assay. At the time point for cell viability analysis, sterile filtered (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) (MTT, Sigma) stock solution (5 mg/ml) in phosphate buffered saline (PBS, ph 7.4) was added into the culture medium to 0.5 mg/ml. 4 hours later, unreacted dye was removed by aspiration and the insoluble formazan crystals were dissolved in dimethylsulfoxide (Sigma) and measured in spectrophotometer at a wavelength of 490 nm. The spectrophotometer was calibrated to zero absorbance [A] using culture medium without cells. The relative cell viability (%) to control wells containing cell culture medium with mock treatment was calculated by [A] test/ [A] control 100. /2 viral plaque assay For the quantification of virus, Vero cells were seeded in the 12-well plates at a density of in the normal medium 12 to 24 hours before /2 infection. After the cells reached ~90-100% confluent, they were infected with serial dilutions of /2 containing supernatants by incubation for 2 hours at 37 C. The plates were rocked every 30 min to ensure the even distribution of virus. 2 hours post infection, the cells were washed twice with PBS and then 2 ml of plaquing 2

19 medium (DMEM with 0.4% agarose and ~1% FBS) was added. The plaquing medium was prepared by mixing a sterile solution of 4% agarose in H2O and 1% FBS containing DMEM. The 4% agarose solution was incubated in a 65 C water bath to remain liquid and the 1% FBScontaining DMEM was also warmed to 37 C before being mixed. After the plaquing medium was added, the plates were left in the level hood at RT for 15 minutes or more to let the agar overlay turn solid and then were incubated at 37 C. After 3-5 days incubation, until the plaques became visible, they were stained with MTT (5 mg/ml in PBS) for 2-4 hours, and the viral plaques were counted. Immunoprecipitation and Western blotting Immunoprecipitations were performed using anti-flag M2 beads, anti-ha beads, anti-myc beads or anti- antibody with Protein A/G agarose beads as described (Zhang et al., 2009). Western blotting of the cell lysates and immunoprecipitates was performed using anti- gd, anti- VP16, anti-, anti-flag, anti-ha, anti-myc and other antibodies as indicated. genomic DNA isolation To extract genomic DNA, we used standard phenol-chloroform extraction methods. Briefly, the tissue or cell culture samples infected with were resuspended in 500 μl of buffer (25 mm EDTA; 75 mm NaCl) and incubated for 10 min with 5 μl of RNase (20 mg/ml). Afterwards, 500 μl of lysis buffer (10 mm EDTA; 10 mm Tris-HCl, ph 8.0; 1% SDS) containing Proteinase K (200 μg/ml) was added into the samples. The samples were then shaken overnight at 56 C, extracted twice with a phenol-chloroform-isoamyl alcohol mixture (25:24:1) and twice with 1 ml of chloroform-isoamyl alcohol mixture (24:1). Aqueous phase (1 ml) was precipitated with 2.5 ml of ice-cold ethanol (100%). Then the DNA pellet was washed briefly with ethanol (70%), air-dried for a few minutes, resuspended in μl of 1 TE and then stored at 4 C. 3

20 RNA extraction and Q-PCR analysis RNA extraction and Q-PCR analysis were performed as described previously (Li et al., 2014). Primers used were as follows: for genomic DNA analysis: gd: 5 - acgactggacggagattaca-3 and 5 -ggagggcgtacttacaggag-3 ; ICP22: 5 -gtgcaagcttccttgtttgg- 3 and 5 -ggtggcatcggagatttcat-3 ; ICP47: 5 -ggtgtggcacatcgaaga-3 and 5 - aacgggttaccggattacg-3. For IFN-β mrna analysis: mifn-β: 5 - ACGCCTGGATGGTGGTCCGA-3 and 5 - TGCCTGCAACCACCACTCATTCT-3 ; mgapdh: 5'-TGTGTCCGTCGTGGATCTGA-3' and 5'-CCTGCTTCACCACCTTCTTGA-3'. CRISPR/Cas9 and TALEN gene editing techniques The CRISPR/Cas9 and TALEN gene editing techniques were described previously (Cong et al., 2013; Li et al., 2014; Mali et al., 2013; Zhang et al., 2011). The Cas9-target sites are as follows: TNFR1 (L929 cells): 5 -GCTTCAACGGCACCGTGACA-3 and human (HT29 cells): GTCGTCGGCAAAGGCGGGTT. The TALEN-target sites are as follows: DAI (L929 cells): 5 - TCTTCTCTGGGTTCCT-3 and 5 -TGCAGGATCTTTTGCT-3 ; cgas (L929 cells): 5 - TCCAGCAAGGGCCACT-3 and 5 -TCCATGGCCGAGGGCT-3 ; RIG-I (L929 cells): 5 - TGACAGCGGAGCAGCGG-3 and 5 -TCTTGATATAGTCTCTG-3. Mass spectra analysis Mass spectra analysis was performed as previously described (Zhang et al., 2009). Confocal microscopy For fixed cell imaging, cells were fixed with freshly prepared 4% paraformaldehyde (PFA) in PBS. The cells were then permeabilized in 0.2% Triton X-100/PBS, blocked with 3% BSA in PBS, stained with anti-flag (rabbit, 1:200, Sigma) and labeled with goat anti-mouse or rabbit AlexaFluor 488 (1:1 000, Invitrogen). Cells were counterstained with DAPI to visualize the nuclei. All 4

21 images were captured and processed using identical settings in the Zeiss LSM 780 laser scanning confocal microscope with a 60 oil objective. Duplicate cultures were examined, and similar results were obtained in at least three independent experiments. Experimental Animals and In Vivo Virus Infection C57BL6/J knockout mice were generated by TALEN gene editing technique, which have the same phenotype as the KO mice described previously (Lin et al., 2013; Newton et al., 2004). Mice (8 10 weeks old) were infected with ( pfu /mouse) by intravenous injection. The weight of the mice was monitored accordingly. genomic DNA levels were determined by Q-PCR analysis of gd DNA levels in the brain or trigeminal ganglia of infected mice. In all the experiments, wild-type and -/- mice were littermates with matched age and sex. 5

22 Supplemental Figures Fig. S1 deficiency results in highly elevated propagation, Related to Figure 1 A. WT and KO L929 cells were infected with VSV (1 moi) over the indicated time periods, and then the cells were harvested to analyze the level of VSV-G protein by immunoblotting. B. WT and KO L929 cells were infected with NDV-GFP (1 moi) over the indicated time periods, and then cells were harvested to analyze the GFP protein level by immunoblotting. C. The accumulation of gd in +/+ and -/- MEF cells. +/+ and -/- MEF cells were infected with different doses of for 36 hours, and the whole cell lysates were analyzed by immunoblotting with the indicated antibodies. D. The overexpression of in NIH3T3 A cells suppresses replication. NIH3T3 A cells infected with lentiviruses encoding nothing (Vector) or were treated with or without HSV- 1 (1 moi) for 36 hours, and the whole cell lysates were subjected to SDS-PAGE and immunoblotting with the indicated antibodies. Fig. S2 infection induces -dependent cell death in mouse but not human cells, Related to Figure 2 A. Viability of the +/+ and -/- MEF cells treated with or without high-dose (20 moi) HSV- 1 for 12 hours. The cell viabilities were measured by MTT assay as described in the extended experimental procedures. Upper panel indicates protein level by immunoblotting. β-actin was used as a loading control. B. overexpression sensitizes NIH3T3 A cells to high-dose infection-induced cell death. NIH3T3 A cells were infected with lentiviruses encoding nothing (Vector) or. A cells, N cells, A cells expressing and control cells were treated with (20 moi) for 6

23 14 hours. The cell viabilities were measured by PI exclusion. protein level in these cells was determined by immunoblotting (upper panel). C. knockout blocks TNF plus zvad plus Smac mimetic-induced necroptosis in HT29 cells. WT and KO HT29 cells were treated with TNF (30 ng/ml) plus zvad (20 μm/ml) plus Smac mimetic (50 μm/ml) for 20 hours and the cell viabilities were measured by PI exclusion. The upper panel shows protein levels in WT and KO HT29. T+S+Z denotes TNF plus zvad plus Smac mimetic. D. deficiency does not block high-dose -induced cell death in HT29 cells. WT and KO HT29 cells were treated with zvad, (20 moi) or plus zvad for 48 hours, and the cell viabilities were determined. E. WT and KO L929 cells were infected with 20 moi VSV for 48 hours, and then the cell viabilities were determined by PI exclusion. F. WT and KO L929 cells were infected with 20 moi NDV-GFP for 48 hours, and then the cell viabilities were determined by PI exclusion. NDV-GFP infection did not lead to L929 cell death even at high-dose infection (20 moi). G. HSV-2 induces dependent cell death. WT, TNFR1 KO and KO L929 cells were treated with HSV-2 (20moi) or HSV-2+zVAD for 20 hours, and the cell viabilities were determined by PI exclusion. H. protein ICP6 does not trigger -dependent cell death in HeLa cells expressing. HeLa cells stably expressing human and control cells were treated with (20 moi), ICP6Δ (20 moi), +zvad or ICP6Δ +zvad for 15 hours. The cell 7

24 viabilities were determined by PI exclusion. expression level was determined by inmmunoblotting (right panel). Data are represented as mean ±SD of triplicate samples (A-H). Fig. S3 Effects of RIP1, MLKL and different mutants on restricting propagation, Related to Figure 3 A. Genomic DNA sequences of TNFR1 locus in TNFR1 KO L929 cells. - indicates none nucleotide. (Note: as the karyotype of L929 cell line is triploid, there are three CRISPR alleles presented in the sequencing results. This also goes for Fig. S4.) B. Viability of WT and TNFR1 KO L929 cells upon TNF (10 ng/ml), zvad (20 µm) or TNF+zVAD treatment. Data are represented as mean ±SD of triplicate samples. C. The effect of conditional media on necroptosis. L929 cells were infected with (20 moi) for 6, 12 and 18 hours, and the conditional media were collected. Ultracentrifugation (3,9000g, 2hrs) was used to remove viruses. These conditional media were then used to treat newly prepared L929 cells for 18 hours. (20 moi) treatment was included as a positive control. The cell viabilities were measured by PI exclusion. Data are represented as mean ±SD of triplicate samples. D. The protein amounts of, RIP1 and MLKL in WT, -/-, RIP1 -/- and MLKL -/- L929 cells used in Fig. 3B. E. WT, RIP1 -/- and MLKL -/- L929 cells as in D were infected with (1 moi) over the indicated time periods. gd and VP16 protein levels were determined by immunoblotting. β-actin was used as a loading control. 8

25 F. The abilities of different mutants in controlling propagation. The cells as in Fig. 3D were infected with 1 moi (low-dose) over the indicated time periods, and the HSV- 1 gd level was measured by immunoblotting. β-actin was used as a loading control. Asterisk denotes non-specific band. Fig. S4 DAI, cgas and RIG-I are not required for -induced cell death in L929 cells, Related to Figure 4 A. Genomic DNA sequences of DAI locus in DAI KO L929 cells. - indicates none nucleotide. B. Genomic DNA sequences of cgas locus in cgas KO L929 cells. - indicates none nucleotide. C. Genomic DNA sequences of RIG-I locus in RIG-I KO L929 cells. - indicates none nucleotide. D. cgas knockout blocks ISD-induced type I interferon production in L929 cells. WT and cgas KO L929 cells were transfected with ISD (5 μg/ml) using Lipofectamin 2000 for 6 hours, and IFNβ mrna level was analyzed by Q-PCR. Data are represented as mean ±SD of triplicate samples. E. RIG-I knockout blocks poly (I:C)- induced type I interferon production in L929 cells. WT and RIG-I KO L929 cells were transfected with poly (I:C) (10 μg/ml) using Lipofectamin 2000 for 0, 4, 6 and 9 hours, and the IFNβ mrna level was analyzed by Q-PCR. F. DAI knockout blocks double-stranded DNA (dsvacv)- induced type I interferon production in L929 cells. WT and DAI KO L929 cells were transfected with dsvacv (5 μg/ml) using Lipofectamin 2000 for 8 hours, and the IFNβ mrna level was analyzed by Q-PCR. Data are represented as mean ±SD of triplicate samples. G. -induced cell death is independent of DAI, cgas and RIG-I. WT, DAI KO, cgas KO and RIG-I KO L929 cells were infected with (20 moi) for 22 hours, and the cell 9

26 viabilities were determined. KO L929 cells were used as control. Data are represented as mean ±SD of triplicate samples. Fig. S5 ICP6 protein of is a trigger of necroptosis, Related to Figure 4 A. Viability of L929 cells upon or plus CHX treatment. L929 cells were treated with (20 moi) or +CHX over the indicated time periods, and the cell viability was determined. B. The protein levels of ICP6 in the samples described in A. C. ICP6 overexpression-induced cell death is TNFR1-independent and can be enhanced by zvad treament. TNFR1 KO L929 cells were infected with lentiviruses encoding nothing (Vector) or ICP6 in the presence or absence of zvad (20 µm) for 60 hours, and the cell viabilities were determined. D. The protein levels of gd in the samples described in Figure 4H. E. The protein levels of gd and ICP6 in the samples described in Figure 4I. F. ICP6 RHIM mutation impairs -mediated restriction on propagation. WT and KO L929 cells were infected with BAC (F strain) (1 moi) and ICP6 RHIM mut (1 moi) over a period of 30 hours, and the indicated protein levels were determined. G. The effects of ICP6 RHIM mutant and enzymatic activity mutant on ICP6Δ propagation in KO L929 cells. KO L929 cells, stably expressing nothing (Vector), WT ICP6, RHIM mutated ICP6 (ICP6 RHIM mut ) or enzymatic inactive mutated ICP6 (ICP6-C793A), were infected with ICP6Δ (5 moi) for 24 hours. The viral titers were determined by viral plaque assay. The protein levels of ICP6 mutants were shown in the right panel. H. The effects of ICP6 RHIM mutant and enzymatic inactive mutant on ICP6 overexpressioninduced necrosis. L929 cells were infected with lentiviruses encoding nothing (Vector), WT 10

27 ICP6, RHIM mutated ICP6 (ICP6 RHIM mut ) or enzymatic inactive mutated ICP6 (ICP6-C793A) for 60 hours. And then the cell viability was determined by PI exclusion. The protein levels of ICP6 mutants were shown in the right panel. Data are represented as mean ±SD of triplicates samples (A, C, G & H). Fig. S6 Comparison between ICP6 and vira, Related to Figure 4 A. ICP6 overexpression induces L929 cells death whereas vira overexpression does not. L929 cells were infected with lentiviruses encoding nothing (Vector), Flag-tagged ICP6 or Flagtagged vira for 50 hours and the cell viabilities were determined (left panel). The protein levels of ICP6 and vira were shown in the right panel. B. vira, but not ICP6, inhibits TNF plus zvad-induced necroptosis in WT L929 cells. WT L929 cells were infected with lentiviruses expressing ICP6 or vira for 16 hours and then treated by TNF plus zvad for another 3 hours. The cell viabilities were determined and shown. Data are represented as mean±sd of triplicate wells. (Note: ICP6 overexpression-induced necrosis in L929 cells cannot be detected in the above described experiments, because it takes more than 24 hours for necrosis to occur after the L929 cells were infected with ICP6-encoding lentiviruses.) C. vira forms dimers/oligomers in an RHIM-independent manner. HEK293T cells were transfected with different plasmids as indicated for 36 hours, and coimmunoprecipitation was performed with anti-flag M2 beads. The proteins in the whole cell lysates and the immunoprecipitates were analyzed by immunoblotting with indicated antibodies. D. vira interacts with via its N-terminal domain (ND) in an RHIM-dependent manner. HEK293T cells were transfected with different plasmids as indicated for 36 hours, and 11

28 coimmunoprecipitation was performed with anti-flag M2 beads. The proteins in the whole cell lysates and the immunoprecipitates were analyzed by immunoblotting with indicated antibodies. E. vira inhibits the formation of murine RIP1-, - or -MLKL complexes. HEK293T cells were transfected with different plasmids as indicated for 36 hours, and coimmunoprecipitation was performed with anti-flag M2 beads. The proteins in the whole cell lysates and the immunoprecipitates were analyzed by immunoblotting with indicated antibodies. F. Schematic representation of chimeric ICP6 and vira with swapped RHIM domain. G. Swapping RHIM domain between ICP6 and vira converts their functions in inducing necroptosis in L929 cells. WT L929 cells were infected with lentiviruses encoding Flag-tagged ICP6, ICP6 vira-r, vira or vira ICP6-R for 48 hours and the cell viabilities were determined. Fig. S7 The working model of -induced necroptosis and the effects of ICP6 and vira on TNF-induced necroptosis, Related to Figure 4 A. The proposed mechanism of -induced necroptosis in mouse cells. B. HT29 cells were infected with lentiviruses encoding the indicated proteins respectively. 16 hours later, cells were treated with or without TNF plus zvad plus Smac mimetic for an additional 20 hours and then the cell viability was determined by PI exclusion. T+S+Z: TNF+Smac+zVAD. C. L929 cells were infected with lentiviruses encoding the indicated proteins respectively. 16 hours later, cells were treated with or without TNF plus zvad for an additional 3 hours and then the cell viability was determined by PI exclusion. 12

29 (Note: ICP6 overexpression-induced necrosis in L929 cells cannot be detected in the above described experiments, because it takes more than 24 hours for necrosis to occur after the L929 cells were infected with ICP6-encoding lentiviruses.) D. pre-treatment suppresses TNF-induced necrosis in HT29 cells. HT29 cells were treated with T+S+Z, +T+S+Z or with for 8 hours and then with +T+S+Z for the indicated time periods. Cell viability and ICP6 protein amounts were determined. T+S+Z: TNF+Smac+zVAD. 5 moi was used. Data are represented as mean ±SD of triplicate samples (B-D). 13

30 Supplemental Movies Movie S1 -infected WT L929 cells undergo necroptotic cell death, Related to Figure 2. WT L929 cells were incubated with PI and infected with GFP- (1 moi), and then monitored by time-lapse fluorescence microscopy for 42 hours. Green and red fluorescent denote GFP and PI positive cells respectively. After GFP- infection, the GFP intensity continued to increase before plasma membrane rupture, indicating that the viral proteins accumulation is involved in cell death. The morphology of cell rupture suggests that the cell death is necroptosis. Movie S2 -infected KO L929 cells are resistant to necroptotic cell death, Related to Figure 2. The experiment was performed as in Movie S1 except that KO L929 cells were used. Different from WT cells, the morphology of GFP + KO cells did not change during the accumulation of viral GFP protein, indicating that the accumulation of viral protein did not induce death of KO cells. 14

31 Supplemental References Chen, W., Zhou, Z., Li, L., Zhong, C.Q., Zheng, X., Wu, X., Zhang, Y., Ma, H., Huang, D., Li, W., et al. (2013). Diverse sequence determinants control human and mouse receptor interacting protein 3 () and mixed lineage kinase domain-like (MLKL) interaction in necroptotic signaling. J. Biol. Chem. 288, Cong, L., Ran, F.A., Cox, D., Lin, S., Barretto, R., Habib, N., Hsu, P.D., Wu, X., Jiang, W., Marraffini, L.A., and Zhang, F. (2013). Multiplex genome engineering using CRISPR/Cas systems. Science 339, Li, L., Chen, W., Liang, Y., Ma, H., Li, W., Zhou, Z., Li, J., Ding, Y., Ren, J., Lin, J., et al. (2014). The Gbetagamma-Src signaling pathway regulates TNF-induced necroptosis via control of necrosome translocation. Cell Res. 24, Lin, J., Li, H., Yang, M., Ren, J., Huang, Z., Han, F., Huang, J., Ma, J., Zhang, D., Zhang, Z., et al. (2013). A role of -mediated macrophage necrosis in atherosclerosis development. Cell Rep 3, Mali, P., Yang, L., Esvelt, K.M., Aach, J., Guell, M., DiCarlo, J.E., Norville, J.E., and Church, G.M. (2013). RNA-guided human genome engineering via Cas9. Science 339, Newton, K., Sun, X., and Dixit, V.M. (2004). Kinase is dispensable for normal NF-kappa Bs, signaling by the B-cell and T-cell receptors, tumor necrosis factor receptor 1, and Toll-like receptors 2 and 4. Mol. Cell. Biol. 24, Stetson, D.B., and Medzhitov, R. (2006). Recognition of cytosolic DNA activates an IRF3- dependent innate immune response. Immunity 24, Unterholzner, L., Keating, S.E., Baran, M., Horan, K.A., Jensen, S.B., Sharma, S., Sirois, C.M., Jin, T., Latz, E., Xiao, T.S., et al. (2010). IFI16 is an innate immune sensor for intracellular DNA. Nat Immunol 11, Zhang, D.W., Shao, J., Lin, J., Zhang, N., Lu, B.J., Lin, S.C., Dong, M.Q., and Han, J. (2009)., an energy metabolism regulator that switches TNF-induced cell death from apoptosis to necrosis. Science 325, Zhang, F., Cong, L., Lodato, S., Kosuri, S., Church, G.M., and Arlotta, P. (2011). Efficient construction of sequence-specific TAL effectors for modulating mammalian transcription. Nat. Biotechnol. 29,

32 Figure S1 A B WT L929 KO WT L929 KO VSV 1 moi (h) NDV-GFP 1 moi (h) VSV-G GFP GAPDH Actin C D 0.5 moi 1 moi 2 moi Mock (1 moi) MEF gd NIH3T3 A Cell gd β-actin GAPDH

33 Figure S2 A WT-1 KO-1 WT-2 KO-2 B A cell C WT HT29 KO β-actin GAPDH h β-actin Survival (%) Mock (20 moi) D E F Survival (%) Mock z-vad +z-vad G 120 L929 WT L929 TNFR1 KO KO H 120 HeLa Vector h HeLa Survival (%) h Vector h 0 Mock HSV-2 HSV-2+zVAD 0 Mock ICP6 +zvad ICP6 +zvad GAPDH

34 Figure S3 A TNFR1-wt TNFR1 KO-allele-1 TNFR1 KO-allele-2 TNFR1 KO-allele-3 B C L WT L929 Survival (%) Survival (%) KO 0h 6h 12h 18h 24h 0 6h Conditional Medium 12h Conditional Medium 18h Conditional Medium (20 moi) D F RIP1 1 moi(h) gd * KO+Vector KO+ KO+RHIM mut KO+D143N KO+2A MLKL β-actin β-actin E WT L929 RIP1 KO WT L929 MLKL KO 1moi(h) moi(h) gd gd VP16 VP16 β-actin β-actin

35 Figure S4 A DAI-wt DAI KO-allele-1 DAI KO-allele-2 DAI KO-allele-3 B cgas-wt cgas KO-allele-1 cgas KO-allele-2 cgas KO-allele-3 C RIG-I-wt RIG-I KO-allele-1 RIG-I KO-allele-2 RIG-I KO-allele-3 D E ISD Transfection Poly(I:C) Transfection 30 Mock ISD Poly(I:C) 0h Poly(I:C) 4h Poly(I:C) 6h Poly(I:C) 9h 0 WT L929 cgas KO 0 WT L929 RIG-I KO F G Relative IFN mrna level WT L929 DAI-/- KO Survial(%) Mock (20 moi)

36 Figure S5 A B 0h 2h 4h 8h 12h 24h Survival (%) ICP6 +CHX +CHX +CHX +CHX +CHX +CHX 0 2h 4h 8h 12h 24h GAPDH C D - 6h 12h 18h Survival(%) gd -- ICP6 ICP6 ICP6 ICP6 ICP6 ICP6 WT KO WT KO WT KO WT KO Mock z-vad GAPDH E F WT L929 KO WT L929 KO WT L929 KO WT L929 KO BAC F strain 20 moi (h) ICP6 RHIM mut 20 moi (h) BAC F strain 1 moi (h) ICP6 RHIM mut 1 moi (h) ICP6 ICP6 gd GAPDH gd GAPDH G Viral titer(pfu/ml) Vecrot Flag-ICP6 Flag-ICP6-RHIM mut Flag-ICP6-C793A Flag GAPDH Vector Flag-ICP6 Flag-ICP6-RHIM mut Flag-ICP6-C793A H Survival(%) Vector WT L929 Flag-ICP6 Flag-ICP6-RHIM mut Flag-ICP6-C793A Flag GAPDH Vector Flag-ICP6 Flag-ICP6-RHIM mut Flag-ICP6-C793A

37 Figure S6 A B C Flag GAPDH Vector Flag-vIRA Myc-vIRA Flag vira RHIM mut Survival (%) Vector ICP6 vira IP:Flag WCL Flag-vIRA Myc-vIRA Flag-vIRA Myc-vIRA D Vector Flag-ICP6 Flag-vIRA E Myc-m Flag Vector Flag-ICP6 Flag-ICP6- RHIM mut Flag-vIRA Flag-vIRA- RHIM mut Flag-vIRA -ND Myc-vIRA Myc-ICP Flag-m HA-mRIP HA-m HA-mMLKL IP:Flag m IP: Flag HA Flag WCL Flag m WCL Myc HA p-m F ICP6 ICP6- RD RHIM G 120 L929 ICP6 vira R vira- RD RHIM vira vira ICP6 R vira- RD RHIM ICP6- RD RHIM Vector Flag-ICP6 Flag-ICP6 vira-r Flag-vIRA Flag-vIRA ICP6-R

38 Figure S7 A? ICP6 ICP6 RHIM RD mmlkl RIP1 mrip1 RHIM m RHIM ICP6 ICP6 RIP1 P P P P Necroptosis MLKL MLKL Anti-virus B C D 4h 8h 12h ICP6 GAPDH