Xie et al. Cell Biosci (2019) 9:5 https://doi.org/10.1186/s13578-018-0268-5 Cell & Bioscience

RESEARCH Open Access Cell type‑specifc function of TRAF2 and TRAF3 in regulating type I IFN induction Xiaoping Xie1, Jin Jin2, Lele Zhu1, Zuliang Jie1, Yanchuan Li1, Baoyu Zhao3, Xuhong Cheng1, Pingwei Li3 and Shao‑Cong Sun1,4*

Abstract Background: TRAF3 is known as a central mediator of type I interferon (IFN) induction by various pattern recognition receptors, but the in vivo function of TRAF3 in host defense against viral infection is poorly defned due to the lack of a viable mouse model. Results: Here we show that mice carrying conditional deletion of TRAF3 in myeloid cells or dendritic cells do not have a signifcant defect in host defense against vesicular stomatitis virus (VSV) infection. However, whole-body inducible deletion of TRAF3 renders mice more sensitive to VSV infection. Consistently, TRAF3 was essential for type I IFN induction in mouse embryonic fbroblasts (MEFs) but not in macrophages. In dendritic cells, TRAF3 was required for type I IFN induction by TLR ligands but not by viruses. We further show that the IFN-regulating function is not unique to TRAF3, since TRAF2 is an essential mediator of type I IFN induction in several cell types, including mac‑ rophages, DCs, and MEFs. Conclusions: These fndings suggest that both TRAF2 and TRAF3 play a crucial role in type I IFN induction, but their functions are cell type- and stimulus-specifc. Keywords: TRAF2, TRAF3, Type I interferon, Antiviral immunity

Background IPS-1) serves as an adaptor of RLRs, whereas stimulator Type I interferon (IFN) plays a crucial role in innate of interferon (STING) mediates type I IFN induc- immunity against viral infections [1, 2]. Induction of tion by cGAS and other cytoplasmic DNA sensors [2, type I IFNs is typically mediated by pattern-recognition 4]. A common downstream target of the diferent PRR receptors (PRRs), including toll-like receptors (TLRs), pathways is the kinase TBK1, which upon activa- RIG-I-like receptors (RLRs), NOD-like receptors, and the tion phosphorylates and activates the transcription factor cytosolic GAMP synthase (cGAS), which detect various interferon regulatory factor 3 (IRF3) [4, 5]. Phosphoryl- molecular patterns associated with viral genomes or rep- ated IRF3 forms a dimer and translocates to the nucleus lication products [2, 3]. In response to stimulation, to participate in the induction of type I IFN . the PRRs transduce signals via specifc signaling adaptors. Activation of TBK1 and IRF3 by the PRRs also TIR-domain-containing adapter-inducing interferon-β requires members of the recep- (TRIF) mediates type I IFN induction by TLR3 and TLR4 tor (TNFR)-associated factors (TRAFs) family [3, 6]. upon stimulation by double-stranded RNA and lipopol- TRAFs are adaptor or E3 ubiquitin ligases ysaccharide (LPS), respectively. Mitochondrial antivi- that transduce signals from TNFR superfamily as well ral-signaling protein (MAVS, also known as VISA and as other immune receptors, including the PRRs [7]. While the typical functions of TRAFs include activa- tion of NF-kB and MAP kinase signaling pathways, *Correspondence: [email protected] TRAF3 has been shown to mediate activation of IRF3 1 Department of Immunology, The University of Texas MD Anderson Cancer Center, 7455 Fannin Street, Box 902, Houston, TX 77030, USA and implicated as a common signaling adaptor for type Full list of author information is available at the end of the article I IFN induction [2, 8, 9]. However, whether TRAF3 is a

© The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Xie et al. Cell Biosci (2019) 9:5 Page 2 of 10

specifc TRAF member that mediates type I IFN induc- type I IFN induction. Te results provide new insight into tion is in debate, since in some systems other TRAF the physiological function of TRAF proteins in mediating members are also involved in the type I IFN induction antiviral innate immunity. [10, 11]. Te in vivo function of TRAF3 in mediating antiviral host defense is also poorly defned, since cur- Results rent studies have been largely relied on in vitro sys- In vivo function of TRAF3 in host defense against viral tems using immortalized mouse embryonic fbroblasts infection (MEFs) or other cell lines. Germline deletion of TRAF3 causes postnatal lethality, To study the physiological function of TRAF3 in anti- which has hampered studies to examine in vivo func- viral innate immunity, we employed TRAF3 conditional tions of TRAF3 [12]. To study the in vivo role of TRAF3 knockout (KO) mice carrying TRAF3 defciencies in dif- in regulating antiviral innate immunity, we generated ferent immune cell types or with inducible deletion of myeloid cell-conditional Traf3 KO ­(Traf3f/fLyz2-Cre; TRAF3 in adult mice. We obtained genetic evidence that hereafter called Traf3MKO) mice by crossing Traf3-fox whole-body inducible deletion of TRAF3 in adult mice mice [13] with Lyz2-Cre [14]. We then challenged these attenuated their host defense against vesicular stomati- mice with VSV, a frequently used model for studying tis virus (VSV) infection. Surprisingly, however, TRAF3 host defense against RNA virus infections [15]. VSV was dispensable in both myeloid cells and dendritic cells is known to induce type I IFN induction via the cyto- for this innate immune function. Using primary cells, plasmic RNA-responsive PRR RIG-I [15]. Surprisingly, we further demonstrated that the function of TRAF3 in TRAF3 deletion in myeloid cells did not sensitize mice to mediating type I IFN induction is cell type- and stimulus- VSV infection, since the Traf3MKO and wild-type control specifc. Moreover, this innate immune function is not mice displayed comparable survival rate following intra- unique for TRAF3, since TRAF2 is equally important for venous (i.v.) infection with VSV (Fig. 1a). Since dendritic

a b 100 P=0.5377 100

) n=8 ) n=8 80 80 (% (% 60 60 40 40 P=0.3362 Survival 20 WT Survival 20 WT Traf3MKO Traf3DC-KO 0 0 024 6 8 10 12 14 024 6 8 10 12 14 Days post infection Days post infection c Spleno- d cytes 100 LN cells ) n=15 80 P=0.035 WT IKO WT IKO (% 60 TRAF3 40

Survival 20 WT Tubulin Traf3IKO 0 024 6 8 10 12 14 Days post infection Fig. 1 Survival of Traf3 conditional KO mice following VSV infection. a Age-matched (6–8 weeks old) WT and Traf3MKO mice were infected i.v. with VSV (1 107 PFU per mouse). Data are presented as 8 animals/group. b Traf3DC-KO and WT littermate control mice were infected i.v. with × VSV (2 107 PFU per mouse). Survival was recorded and presented as 8 mice per group. c Immunoblot analysis of the indicated proteins in the × extracts of splenocytes or LN cells derived from Tamoxifen-treated WT and TRAF3IKO mice. d WT and TRAF3IKO mice (n 15) were infected with VSV = (2 107 PFU per mouse) via tail vein. Data representative of two independent experiments, and statistical signifcance was determined by log- × and Gehan-Wilcoxon tests. p < 0.05 indicates signifcantly diferent Xie et al. Cell Biosci (2019) 9:5 Page 3 of 10

cells (DCs) also play a crucial role in regulating antiviral (Fig. 2f), ISD (Fig. 2g) or DMXAA (Fig. 2h). Consist- immunity, we next examined the in vivo role of TRAF3 ently, while TRAF3 defciency promoted the activation in DCs by generating DC-conditional Traf3 KO (Traf3f/ of STING/TBK1/IRF3 signaling pathways by HSV-1 fCd11c-Cre; here after called Traf3DC-KO) mice. Simi- (Fig. 2i), it attenuated these signaling events induced by lar to the Traf3MKO mice, the Traf3DC-KO mice did not cGAMP (Fig. 2j). Tese results suggest that TRAF3 may show enhanced sensitivity to VSV infection compared play diferent roles in regulating STING activation and to their age-matched wild-type control mice (Fig. 1b). upstream signaling steps in the DNA-sensing pathway. Tese results suggest that TRAF3 is dispensable in innate To examine the role of TRAF3 in innate immune cells, immune cells for host defense against VSV infections. we prepared primary bone barrow derived macrophages To further investigate the antiviral innate immune (BMDMs) from Traf3MKO and wild-type control mice. In function of TRAF3, we employed an inducible KO sys- contrast to the results obtained with MEFs, TRAF3 dele- tem allowing TRAF3 deletion in diferent cell types in tion in BMDMs had no efect on the induction of Ifna and adult mice. We crossed the Traf3-fox mice with trans- Ifnb genes by VSV or SeV (Fig. 2k, l). On the other hand, genic mice expressing a tamoxifen-inducible Cre recom- as seen in MEFs, the TRAF3 defciency in macrophages binase, ER-Cre [16]. We then injected the Traf3f/fERCre/+ promoted type I IFN induction by the DNA virus HSV and control Traf3f/f mice with tamoxifen to generate (Fig. 2m). Furthermore, TRAF3 was completely dispensa- the Traf3 inducible KO (Traf3IKO) and wild-type con- ble for the induction of Ifna and Ifnb by the TLR ligands trol mice. Immunoblot analysis using splenocytes and LPS and polyI:C (Fig. 2n, o). Consistently, loss of TRAF3 lymph node cells revealed efcient deletion of TRAF3 in in macrophages did not afect LPS-stimulated phospho- the Traf3IKO mice (Fig. 1c). Importantly, the whole-body rylation of TBK1 or its homolog IKKi (Fig. 2p). Tese deletion of TRAF3 in adult mice signifcantly attenuated results suggest cell type-specifc functions of TRAF3 in the host defense against VSV infection (Fig. 1d). Tese mediating type I IFN induction. results suggest that TRAF3 plays an important role in mediating antiviral immunity, but it may have cell type- Inducer‑specifc function of TRAF3 in DCs specifc functions. DCs also serve as an important innate immune cell type mediating host defense against viral infections. We TRAF3 is required for type I IFN induction in MEFs thus examined the role of TRAF3 in regulating type I but not in macrophages IFN induction in primary bone marrow derived DCs To better understand the cell type-specifc functions (BMDCs). Compared to the wild-type BMDCs, the of TRAF3, we prepared primary MEFs as well as innate TRAF3-defcient BMDCs had a signifcant reduction immune cells from TRAF3-defcient or wild-type control in type I IFN gene induction by both LPS and polyI:C mice. In agreement with previous studies [8, 9], TRAF3 (Fig. 3a, b). Consistently, the TRAF3-defcient DCs also defciency in primary MEFs severely attenuated the Ifna had impaired induction of TBK1 phosphorylation by LPS and Ifnb induction by the RNA viruses VSV and Sendai (Fig. 3c). Interestingly, however, TRAF3 was completely virus (SeV) (Fig. 2a, b). In addition, the TRAF3-defcient dispensable for induction of Ifna and Ifnb by the RNA MEFs were also defective in type I IFN induction by viruses VSV and SeV (Fig. 3d, e). Similarly, TRAF3 dele- transfected polyI:C known to mimic RNA viruses and tion in DCs had no signifcant efect on type I IFN induc- stimulate the RIG-I signaling pathway (Fig. 2c). Con- tion by the DNA virus HSV (Fig. 3f). To further confrm sistently, parallel immunoblot assays revealed reduced these results, we also directly stimulated the RIG-I and phosphorylation of TBK1 and its target transcription fac- STING pathways using transfected polyI:C and cGAMP, tor IRF3 in TRAF3-defcient MEFs stimulated by VSV respectively. Once again, induction of type I IFN gene (Fig. 2d). Tus, TRAF3 is required for type I IFN induc- expression by these inducers did not require TRAF3 in tion by the RIG-I pathway in primary MEFs. DCs (Fig. 3g, h). Tus, the function of TRAF3 in mediat- A recent study suggests that TRAF3 negatively regu- ing type I IFN induction is dependent on inducers as well lates DNA virus-induced type I IFN induction with as cell types. a mechanism that involves STING activation by the TRAF3-controlled kinase NIK [17]. Consistent with this TRAF2 is an essential mediator of type I IFN induction report, we found that the TRAF3-defcient MEFs had in multiple cell types greatly enhanced type I IFN induction by a DNA virus, TRAF3 has been frequently cited in the literature as herpes simplex virus 1 (HSV-1) (Fig. 2e). Surprisingly, a central mediator of antiviral innate immunity; how- however, the TRAF3 defciency did not promote, but ever, it is unclear whether this function is unique for rather signifcantly attenuated, the induction of type I TRAF3 or also for other TRAF members. To address IFN induction by non-viral STING agonists, cGAMP this question, we examined the role of TRAF2 by Xie et al. Cell Biosci (2019) 9:5 Page 4 of 10

c a 2.5 2.5 b 1.5 1.5 5 * 6 d **** **** 4 * Traf3+/+ Traf3–/– NA 2.0 2.0 *** *** * 1.0 NA 1.0 * 4 VSV (h) 048048 1.5 1.5 *** 3 * +/+ mR mRNA Traf3 mRNA mR mRNA 2 mRNA Traf3–/– p-TBK1 1.0 1.0 0.5 0.5 2 Ifna Ifna Ifna

Ifnb 1 Ifnb ns TBK1 ns 0.5 Ifnb 0.5 CE 0 0 0 0 Lipo/ 0 0 TRAF3 VSV (h) 06 9 06 9 Sev (h) 06 9 06 9 PolyI:C(h) 069 069 HSP60 e 1.5 f 2.0 2.5 g 2.0 3 1.5 *** ** **** **** p-IRF3 ** 1.5 NA 2.0 ** 1.5 **** +/+ IRF3 1.0 1.0 1.5 2 **** Traf3 mR mRNA mRNA mRNA – – NE

mRNA / mRNA 1.0 1.0 Traf3

a p-IRF7 0.5 0.5 * 1.0 1

0.5 Ifnb Ifnb 0.5 Ifn Ifna Ifnb Ifna **** *** 0.5 ** LaminB 0 0 Lipo/ 0 * 0 0 0 HSV-1 (h) 036 036 cGAMP(h) 048 048 Lipo/ISD (h) 036 036

h i j l Traf3-WT Traf3+/+ Traf3–/– 1.5 1.5 ns 5 5 Lipo/ Traf3+/+ Traf3–/– Traf3-MKO **** **** HSV1 (h) 048 048 ns ns 4 4 cGAMP(h) 036036 1.0 1.0 p-STING p-STING mRNA 3 3 mRNA ns mRNA mRNA 2 0.5 0.5 2 STING STING Ifnb Ifna Ifnb Ifna 1 1 p-TBK1 p-TBK1 0 0 0 0 Sev(h) 069 069 DMXAA(h) 036 036 TBK1 TBK1 2.0 ns m 1.5 k 1.5 1.5 p-IRF3 p-IRF3 ns 1.5 * *** 1.0

1.0 1.0 IRF3 IRF3 mRNA mRNA 1.0 * mRNA mRNA ns ns TRAF3 0.5 0.5 TRAF3 0.5 ns Ifnb

0.5 Ifna Ifnb Ifna Actin Actin 0 0 0 0 VSV (h) 069 069 HSV-1 (h) 06 9 069

Traf3- MKO n 2.5 o p WT 1.5 2.0 ns 2.0 ns ns ns LPS (min) 01015 01015 2.0 1.5 1.5 p-TBK1 1.0 1.5 Traf3-WT TBK1 mRNA mRNA 1.0

ns mRNA Traf3-MKO

mRNA 1.0 1.0 ns 0.5 p-IKKi Ifnb Ifna Ifna 0.5 Ifnb 0.5 ns 0.5 ns IKKi 0 0 0 0 TRAF3 LPS (h)0 26 026 PolyI:C(h) 0 26 026 Fig. 2 Cell type-specifc function of TRAF3 in regulating type I IFN induction. qRT-PCR analysis of Ifna and Ifnb mRNA expression in WT and Traf3–/– primary MEFs infected with VSV (a) or SeV (b) or treated with transfected with poly(I:C) (c) for the indicated times. d Immunoblot analysis of the indicated phosphorylated (P ) and total proteins in the cytoplasmic (CE) and nuclear (NE) extracts of WT and Traf3–/– primary MEFs infected − with VSV (M.O.I 10) for the indicated time period. qRT-PCR analysis of relative Ifna and Ifnb mRNA amounts in WT and Traf3–/– MEF cells infected = with HSV-1 (e) or treated with lipofectamine-transfected cGAMP (f), lipofectamine-transfected ISD (g), or stimulated with DMXAA (h) for the indicated time periods. i, j Immunoblot analysis of the indicated proteins in whole-cell lysates of WT or Traf3–/– MEFs infected with HSV-1 (i) or lipofectamine-transfeted cGAMP (j) for the indicated time periods. qRT-PCR analysis of Ifna and Ifnb mRNA relative level in BMDMs derived from WT or Traf3MKO mice infected with VSV (k), SeV (l), HSV-1 (m) or stimulated with LPS (n) or poly(I:C) (o). p Immunoblot analysis of the indicated proteins in whole-cell lysates of WT or Traf3MKO BMDMs stimulated with LPS for the indicated time periods. All qRT-PCR data are presented as fold relative to the amount of an internal control, Actb. Data are representative of three to four independent experiments, and statistical analyses are presented as mean SEM. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. ns nonsignifcant ± generating primary cells from TRAF2-conditional KO similar and diferent functions from TRAF3 in the reg- and wild-type control mice. Interestingly, like TRAF3, ulation of type I IFN induction. TRAF2 was essential for type I IFN induction by the We next examined the role of TRAF2 in regulating RNA virus SeV in primary MEFs (Fig. 4a). Tis fnding type I IFN induction by TRIF-dependent TLRs, TLR3 was intriguing, since it suggests that deletion of either and TLR4, using macrophages. BMDMs prepared for the TRAF2 or TRAF3 in MEFs blocks type I IFN induction, Traf2MKO mice had a severe defect in LPS-stimulated Ifna thus suggesting non-redundant functions of TRAF2 and Ifnb expression and also a signifcant reduction in and TRAF3. Remarkably, TRAF2 was also required polyI:C-stimulated Ifnb expression (Fig. 4c, d). Consist- for type I IFN induction by the DNA virus HSV-1 ently, the TRAF2 defciency attenuated, although did not (Fig. 4b), which was in sharp contrast to the negative completely block, LPS-stimulated TBK1 phosphorylation role of TRAF3 in mediating the DNA-sensing pathway (Fig. 4e). TRAF2 defciency did not appreciablely alter (Fig. 2e). Tese results suggest that TRAF2 has both the activation of IKKi or other signaling factors, suggest- ing the involvement of TRAF2 in regulating the TBK1 Xie et al. Cell Biosci (2019) 9:5 Page 5 of 10

Traf3 ab1.5 2.0 1.5 c WT DC-KO **** 1.5 ** ** ns * LPS (min) 01015010 15 A 1.5 N 1.0 1.0 1.0 p-TBK1 R mRNA mRNA mRNA m 1.0 TBK1 b a a 0.5 0.5

ns Ifn Ifnb 0.5 Ifn Ifn 0.5 TRAF3 ns ns 0 0 0 0 Hsp60 LPS (h) 0 2 6 0 2 6 PolyI:C (h) 0 2 6 0 2 6

ns de3.0 2.0 ns f 1.5 ns 2.5 1.5 1.5 ns ns

A ns 2.0 1.5 2.0 ns ns N 1.0 ns R

mRNA 1.0 ns 1.5 1.0

m 1.0 mRNA mRNA mRNA 0.5 1.0 mRNA Ifnb 1.0 b

Ifna 0.5 0.5 0.5 Ifnb

Ifna ns Ifn 0.5 Ifna 0 0 ns VSV (h) 0 69 0 069 SeV(h) 0 69 0 0 0 0 6 9 HSV-1(h) 0 2 6 0 2 6

g 1.5 1.5 h 1.5 1.5 ns ns ns ns ns 1.0 1.0 1.0 1.0 mRNA mRNA mRNA

mRNA Traf3 WT 0.5 0.5 * 0.5 0.5 Traf3 DC-KO Ifnb Ifna Ifnb Ifna

0 0 0 0 Lipo/PolyI:C 069 069 Lipo/cGAMP 04 04 Fig. 3 Stimulus-specifc function of TRAF3 in regulating type I IFN induction in BMDCs. a, b qRT-PCR analysis of Ifna and Ifnb expression in BMDCs derived from WT or Traf3DC-KO mice stimulated with LPS (a) or poly(I:C) (b). c Immunoblot analysis of the indicated proteins in whole-cell lysates of WT or Traf3DC-KO BMDCs stimulated with LPS for the indicated time points. qRT-PCR analysis of Ifna and Ifnb mRNA expression in WT and Traf3DC-KO BMDCs infected with VSV (d), Sev (e) or HSV-1 (f) or treated with lipofectamine-transfected poly(I:C) (g) or cGAMP (h) for the indicated time periods. Data are representative of three independent experiments, and statistical analyses are presented as mean SEM. *p < 0.05, **p < 0.01, ***p < 0.001 ± and ****p < 0.0001. ns nonsignifcant

signaling axis. Parallel studies using the RNA virus SeV demonstrated that this function of TRAF3 was cell type- and the DNA virus HSV-1 did not reveal signifcant dif- and stimulus-specifc (Table 1). Surprisingly, deletion of ferences in Ifna and Ifnb induction between the Traf2MKO TRAF3 in either myeloid cells or dendritic cells had no and wild-type control macrophages (Fig. 4f, g). efect on host defense against VSV infection, but whole- To investigate the role of TRAF2 in DCs, we gener- body inducible deletion of TRAF3 impaired this innate ated DC-conditional TRAF2 KO ­(Traf2DC-KO) mice. immune function. Consistent with these in vivo results, BMDCs prepared from the ­Traf2DC-KO mice displayed a cell culture studies revealed that TRAF3 was dispen- severe defect in type I IFN induction by LPS and polyI:C sable for type I IFN induction by both TLR ligands and (Fig. 5a, b). In contrast to the TLR pathways, induction viruses in macrophages. Furthermore, TRAF3 deletion in of type I IFNs by the RNA viruses, SeV and VSV, and DCs had no efect on virus-induced type I IFN induction, the DNA virus HSV-1 was not afected by the TRAF2 although it inhibited IFN induction by TLR3 and TLR4 defciency in DCs (Fig. 5c–e). Collectively, these results ligands. In contrast, TRAF3 was essential for IFN induc- suggest that like TRAF3, TRAF2 has an essential role in tion by RNA viruses in MEFs. Tese results suggest that mediating type I IFN induction in an inducer-specifc TRAF3 is dispensable for antiviral responses in innate manner. immune cells but may play an essential role in other cell types, such as fbroblasts. Discussion Our fnding that TRAF3 is dispensable in myeloid cells Te results presented in the present study provide for innate immunity against VSV infection is in agree- genetic evidence that TRAF3 mediates antiviral host ment with a previous study that myeloid cell-specifc defense in mice. Our data support the previous reports deletion of TRAF3 has no efect on LPS-induced IFNβ that TRAF3 plays a crucial role in regulating type I IFN production in vivo [18]. However, in contrast to our pre- induction [8, 9]. However, by employing diferent types sent fnding that TRAF3 is dispensable for LPS-stim- of primary cells as well as Traf3 conditional KO mice, we ulated type IFN expression in macrophages, this prior Xie et al. Cell Biosci (2019) 9:5 Page 6 of 10

a 2.5 2.5 b 2.0 2.5 Traf2+/+ e Traf2- **** – – **** **** **** Traf2 / MKO A 2.0 ** 2.0 *** 1.5 2.0 WT N * LPS (min) 01020010 20 R 1.5 **** 1.5 1.5 mRNA mRNA mRNA m 1.0 p-TBK1 1.0 1.0 1.0 Ifnb

Ifnb TBK1

Ifna 0.5 Ifna 0.5 0.5 0.5 0 0 0 0 p-IKKi 4 8 0 4 8 HSV-1(h) 0 4 8 0 4 8 Sev(h) 0 IKKi c 1.5 2.0 d 1.5 2.0 Traf2-WT p-Erk1/2 ns * ** *** Traf2-MKO Erk1/2 A 1.5 1.5 N 1.0 1.0 R mRNA mRNA mRNA p-JNK1/2

m 1.0 1.0 0.5 0.5 ** Ifnb Ifnb **** 0.5 Ifna ns 0.5 Ifna p-STAT1 (S727) * 0 0 0 0 p-STAT3 (S727) LPS (h) 0 2 6 0 2 6 PolyI:C(h)0 2 6 0 2 6 STAT3 f g TRAF2 2.0 ns 2.5 1.5 ns 2.0 ns ns ns A Hsp60

N 1.5 2.0 1.5 ns

R 1.0 ns 1.5

mRNA ns mRNA mRNA m 1.0 1.0 1.0 0.5 Ifnb Ifna Ifnb Ifna 0.5 0.5 0.5 0 0 0 0 Sev (h) 048 0 4 8 HSV-1 (h) 0 4 8 0 4 8 Fig. 4 Role of TRAF2 in regulating type I IFN Induction. qRT-PCR analysis of Ifna and Ifnb expression in WT or TRAF2–/– primary MEFs infected with Sev (a) or HSV-1 (b). c, d qRT-PCR analysis of Ifna and Ifnb expression in BMDMs derived from WT or Traf2MKO mice stimulated with LPS (c) or poly(I:C) (d). e Immunoblot analysis of the indicated proteins in whole-cell lysates of WT or TRAF2MKO BMDMs stimulated with LPS for the indicated time points. f, g qRT-PCR analysis of Ifna and Ifnb mRNA expression in WT and Traf2MKO BMDM cells infected with Sev (f) or HSV-1 (g). Data are representative of three independent experiments, and statistical analyses are presented as mean SEM. *p < 0.05, **p < 0.01, ***p < 0.001 and ± ****p < 0.0001. ns nonsignifcant

a 2.0 2.0 b 0.8 2.0 ** *** **** * 1.5 1.5 0.6 1.5 Traf2-WT mRNA mRNA mRNA 0.4 1.0 mRNA 1.0 1.0 Traf2-DC-KO b b

a ** Ifn Ifn Ifna

Ifn 0.5 0.5 0.2 0.5 ** ** 0 0 ** 0 0 LPS (h)20 6 0 26 PolyI:C (h)20 6 0 26 c de 1.5 ns 1.5 1.5 2.0 ns 1.5 ns 1.5 ns ns ns A

N 1.0 ns ns 1.5 R 1.0 1.0 1.0 1.0

m ns * mRNA ns ns mRNA mRNA a 1.0 b mRNA mRNA a b a b

Ifn 0.5 Ifn 0.5 0.5 0.5 0.5

Ifn Ifn 0.5 Ifn Ifn 0 0 0 0 0 0 SeV(h)0 48 0 48 VSV (h) 0448 0 8 HSV-1 (h) 0448 0 8 Fig. 5 Stimulus-specifc function of TRAF2 in regulating type I IFN induction in DCs. a, b qRT-PCR analysis of Ifna and Ifnb mRNA expression in WT and Traf2DC-KO BMDCs stimulated with LPS (a) or poly(I:C) (b) or infected with Sev (c), VSV (d) or HSV-1 (e). Data are representative of three independent experiments, and statistical analyses are presented as mean SEM. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. ns ± nonsignifcant Xie et al. Cell Biosci (2019) 9:5 Page 7 of 10

Table 1 Cell-type specifc efect of TRAF3 defciency and TRAF3 is currently unclear, but it may involve on type I IFN induction functional redundancy with other TRAF members. Pathways Inducers Cell types However, it is remarkable that TRAF2 and TRAF3 both had non-redundant roles in type I IFN induction in the MEFs BMDM BMDC same cell types, including MEFs and DCs, since dele- RIG-I pathway VSV Inhibited No efect No efect tion of either TRAF2 or TRAF3 impaired the induc- Sev Inhibited No efect No efect tion of type I IFNs by the same types of inducers. Tis Lipo/PolyI:C Inhibited NA No efect fnding suggests that TRAF2 and TRAF3 may function STING pathway HSV1 Upregulated Upregulated No efect cooperatively, as seen in the regulation of the nonca- Lipo/cGAMP Inhibited NA No efect nonical NF-kB kinase NIK [19], or at diferent steps of Lipo/ISD Inhibited NA NA the IFN induction pathways. DMXAA Inhibited NA NA A recent study suggests that TRAF3 negatively regu- TRIF pathway LPS NA No efect Inhibited lates type I IFN induction by DNA viruses, such as PolyI:C NA No efect Inhibited HSV-1, through controlling the noncanonical NF-κB NA not analyzed inducing kinase NIK [17]. Accumulation of NIK due to TRAF3 defciency promotes STING activation and induction of type I IFNs [17]. In support of this report, we found that TRAF3 deletion in MEFs and Table 2 Cell-type specifc efect of TRAF2 defciency macrophages promotes HSV-1-stimulated expres- on type I IFN induction sion of Ifna and Ifnb genes. However, the mechanism Pathways Inducers Cell types by which TRAF3 regulates STING pathway appears to MEFs BMDM BMDC be complex. First, while TRAF3 defciency promoted type I IFN induction by HSV-1, it inhibited type I IFN RIG-I pathway VSV NA NA No efect induction by the STING stimulators cGAMP, ISD, and Sev Inhibited No efect No efect DMXAA. Tese results suggest that TRAF3 may have STING pathway HSV1 Inhibited No efect No efect dual functions in DNA-stimulated type I IFN produc- TRIF pathway LPS NA Inhibited Inhibited tion. It is possible that TRAF3 may positively regulate PolyI:C NA Inhibited Inhibited STING activation of TBK1 or IRF3 and negatively reg- NA not analyzed ulates an upstream step in the DNA-sensing pathway. Second, TRAF2 defciency, which also causes NIK accu- mulation [20], did not promote type I IFN induction study suggests that TRAF3 is required for in vitro type by HSV-1, which argues for the involvement of a NIK- I IFN induction by LPS in macrophages. Te reason for independent mechanism. In fact, deletion of TRAF2 in this discrepancy is unclear, but it could be due to the use MEFs largely blocked the induction of type I IFN genes of two diferent Traf3-fox mouse strains or diferences by HSV-1. Tese results, along with the fnding that in experimental conditions. Nevertheless, we found that TRAF3 defciency blocks cGAMP-induced type I IFN TRAF3 deletion in macrophages also did not afect type induction, further suggest a positive role for TRAF2 I IFN induction by another TLR ligand, polyI:C. Future and TRAF3 in regulating STING-mediated signaling. studies will further study the role of TRAF3 and other HSV-1 has evolved multiple strategies to evade host TRAF members in TLR-stimulated type I IFN induction. antiviral responses. Some HSV-1-encoded gene prod- Although TRAF3 has been implicated as a spe- ucts have been identifed inhibitors targeting upstream cifc antiviral signaling adaptor, our data suggest that steps of the STING/TBK1/IRF3 pathway. For example, TRAF2 is equally important for mediating type I IFN the HSV-1 tegument protein UL41 inhibits type I IFN induction. In fact, TRAF2 displayed an essential role induction by selectively degrading the mRNA of cGAS, in mediating type I IFN induction in several cell types, a key DNA sensor mediating cGAMP synthesis and including MEFs, macrophages, and DCs (Table 2). In STING activation by DNA viruses [21]. Another HSV-1 macrophages and DCs, TRAF2 appeared to be more tegument protein, UL37, deamidates mouse cGAS, important for the TRIF-dependent TLR pathways, impairing its ability to catalyze cGAMP synthesis [22]. since TRAF2 deletion in these cells inhibited type I IFN Moreover, an HSV-1 virulence factor, γ134.5, inhibits induction by TLR3 and TLR4 ligands but not by RNA STING activation by interfering the translocation of or DNA viruses. Te molecular mechanism underlying STING from endoplasmic reticulum to Golgi apparatus the cell type- and inducer-specifc functions of TRAF2 [23]. Future studies will examine whether TRAF3 plays Xie et al. Cell Biosci (2019) 9:5 Page 8 of 10

a role in regulating the interplay between these HSV-1 Santa Cruz Biotechnology. p-IRF3 was purchased from components and host STING signaling pathway. TermoFisher Scientifc. LPS (derived from Escheri- chia coli strain 0127:B8) were from Sigma Aldrich. Poly Conclusions (I:C) was from Amersham. cGAMP was synthesized as Te fndings of the present study suggest that in addi- previously described [25]. Recombinant murine M-CSF tion to TRAF3, TRAF2 is a crucial mediator of type I IFN and GM-CSF were from Peprotech. Murine STING induction and that the function of TRAF2 and TRAF3 is ligand DMXAA and ISD naked were purchased from cell type- and stimulus-specifc. InvivoGen.

Materials and methods Cell preparation and stimulation Mice Bone marrows were prepared from the femurs of adult Traf2-fox and Traf3-fox mice were provided by Dr. Rob- mice and cultured in a M-CSF conditional medium for ert Brink (Garvan Institute of Medical Research) [13]. BMDM diferentiation. BMDCs were generated by cul- Traf3-fox mice were further crossed with EIIa-Cre mice, tivating bone marrow cells in RPMI medium supple- lysozyme 2-Cre (Lyz2-Cre) mice or CD11c-Cre mice to mented with GM-CSF (10 ng/ml) for 7 days, and the produce age-matched Traf3 germ line knockout mice diferentiated BMDCs were stained with Pacifc blue– (termed Traf3–/–), ­Traf3f/f ­lyz2Cre/+ (termed Traf3MKO), conjugated anti-CD11c and purifed by fow cytometric ­Traf3f/f ­CD11cCre (termed Traf3DC-KO) mice and Traf3­ f/ cell sorting as described [26]. To prepare primary MEFs, f (termed Traf3+/+) mice. Also, Traf2-fox mice were we bred heterozygous mice for obtaining the KO and WT crossed with EIIa-Cre mice, Lyz2-Cre mice or CD11c- embryos from the same pregnant female mice. Te MEFs, Cre mice to produce age-matched Traf2 germ line knock- BMDMs, and BMDCs were starved overnight in medium out mice (termed Traf2–/–), Traf2f/f ­lyz2Cre/+ (termed supplemented with 0.5% FCS before being stimulated Traf2MKO), ­Traf2f/f ­CD11cCre (termed Traf2DC-KO) mice with LPS (1 μg/ml for immunoblot experiments and and Traf2f/f (termed Traf2+/+) mice. Traf3-fox mice 100 ng/ml for cytokine induction experiments), poly(I:C) were also crossed with CAGG-Cre-ER mice to generate (20 μg/ml), DMXAA (50 μg/ml), Lipofectamine-trans- age matched tamoxifen inducible Traf3­ f/f ER­ Cre/+ mice fected poly(I:C) (20 μg/ml), cGAMP (10 μg/ml) or ISD and ­Traf3f/f mice; these mice were then injected intra- (1 μg/ml) for the indicated times. Total and subcellular peritoneally with 100 μl tamoxifen (20 mg/ml in corn oil extracts were prepared for immunoblot assays, and total solution) for a total of 5 consecutive days with 24 h inter- RNA was prepared for QPCR assays. vals, creating Traf3 inducible KO (Traf3IKO) and wild- type control mice. Seven days after the last tamoxifen Quantitative RT‑PCR injection, the mice were used for viral infection experi- Total RNA was isolated using Trizol reagent (Molecular ments. All mouse strains were in C57BL/6 background. Research Center, Inc.) and subjected to cDNA synthesis Te mice were maintained in specifc pathogen-free using RNase H-reverse transcriptase (Invitrogen) and facility of Te University of Texas MD Anderson Cancer oligo (dT) primers. qRT-PCR was performed in tripli- Center, and all animal experiments were conducted in cates, using iCycler Sequence Detection System (Bio- accordance with protocols approved by the Institutional Rad) and iQTM SYBR Green Supermix (Bio-Rad). Te Animal Care and Use Committee. expression of individual genes was calculated by a stand- ard curve method and normalized to the expression of Viruses, antibodies, and reagents Actb. Te mouse gene-specifc PCR primers are used as follows: Actb forward primer: 5′CGT​GAA​AAG​ATG​ACC​ VSV (Indiana strain) was provided by Glen Barber (Uni- CAG​ATCA, Actb reverse primer: 5′CACAGC​ CTG​ GAT​ ​ versity of Miami), and a VSV variant harboring a point GGC​TAC​GT3′; Ifnb forward primer: 5′AGCTCC​ AAG​ ​ mutation in the M gene (AV1) was provided by John AAAGGA​ CGA​ ACAT3​ ′, Ifnb reverse primer: 5′GCCCTG​ ​ Bell (Ottawa Hospital Research Institute) [24]. Vero TAG​GTG​AGG​TTG​ATCT3′; and Ifna forward primer: cells were used to propagate and determine the titers of 5′TGACCT​ CAA​ AGC​ CTG​ TGT​ GATG3​ ′, Ifna reverse HSV-1 (KOS), and BHK21 cells were used for generat- primer: 5′AAGTAT​ TTC​ CTC​ ACA​ GCC​ AGCAG3​ ′. ing VSV WT and VSV-AV1 Mutant viruses. Antibodies for STING, p-STING (Ser365), TBK1, p-TBK1, p-IKKe, Immunoblot IKKi, p52, p-IRF7, p-Erk, p-STAT3 (Ser727), p-STAT1 (S727), STAT3 and Tubulin were purchased from Cell Whole-cell lysates or subcellular extracts were prepared Signaling Technology Inc. Antibodies for Hsp60, TRAF2, and subjected to immunoblot assays as described [27, 28]. TRAF3, Erk, LaminB and IRF3 were purchased from Te samples were resolved by 8.25% SDS-PAGE. After electrophoresis, separated proteins were transferred Xie et al. Cell Biosci (2019) 9:5 Page 9 of 10

onto polyvinylidene difuoride membrane (Millipore). Competing interests The authors declare that they have no competing interests. For IB assays, the polyvinylidene difuoride membrane was blocked with 5% non-fat milk. After incubation Consent for publication with specifc primary antibody, horseradish peroxidase- Not applicable. conjugated secondary antibody was applied. Te positive Data availability immune reactive signal was detected by ECL (Amersham The datasets generated during the current study are available from the cor‑ Biosciences). responding author on reasonable request. Ethics approval and consent to participate Viral infection Not applicable. For gene induction and signaling analysis, MEFs were 5 Funding seeded into 12-well plates (5 × 10 cells per well) and This research was supported by grants from the US National Institutes of infected with VSV-AV1 mutant, HSV-1 (KOS) or Sev Health (AI64639, AI057555, and GM84459). strains in serum-free medium for 1 h. Te cells were washed once and cultured in growth medium for the Publisher’s Note indicated time periods and then collected for immuno- Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional afliations. blot or QPCR assays. For mouse infection, the indicated age-matched (6–8 weeks’ old) Traf3 conditional KO or Received: 14 November 2018 Accepted: 29 December 2018 WT control mice were housed in a biosafety level 2 facil- 7 ity and infected i.v. with VSV (2 × 10 PFU per mouse in 200 μl). Te infected mice were monitored for survival up to 14 days. References 1. Yan N, Chen ZJ. Intrinsic antiviral immunity. Nat Immunol. 2012;13(3):214–22. Statistical analysis 2. McNab F, Mayer-Barber K, Sher A, Wack A, O’Garra A. Type I interferons in Prism software was used for two-tailed unpaired t-tests infectious disease. Nat Rev Immunol. 2015;15(2):87–103. 3. Tan X, Sun L, Chen J, Chen ZJ. Detection of microbial infections and data are presented as mean ± SEM. Log-rank and through innate immune sensing of nucleic acids. Annu Rev Microbiol. Gehan-Wilcoxon tests were performed for survival 2018;72:447–78. curves. p-values < 0.05 and 0.01 are considered signifcant 4. Hu H, Sun SC. Ubiquitin signaling in immune responses. Cell Res. 2016;26(4):457–83. and very signifcant, respectively. 5. Hiscott J. Triggering the innate antiviral response through IRF-3 activa‑ tion. J Biol Chem. 2007;282:15325–9. 6. Hacker H, Tseng PH, Karin M. Expanding TRAF function: TRAF3 as a tri- Abbreviations faced immune regulator. Nat Rev Immunol. 2011;11(7):457–68. IFN: interferon; VSV: vascular stomatitis virus; SeV: Sendai virus; HSV-1: herpes 7. Shi JH, Sun SC. Tumor necrosis factor -associated factor regula‑ simplex virus 1; MEF: mouse embryonic fbroblast; PRR: pattern-recognition tion of nuclear factor kappaB and mitogen-activated protein kinase receptor; TLR: toll-like receptor; RLR: RIG-I-like receptor; cGAS: cytosolic GAMP pathways. Front Immunol. 2018;9:1849. synthase; TRIF: TIR-domain-containing adapter-inducing interferon-β; LPS: 8. Häcker H, Redecke V, Blagoev B, Kratchmarova I, Hsu LC, Wang GG, et al. lipopolysaccharide; MAVS: mitochondrial antiviral-signaling protein; STING: Specifcity in Toll-like receptor signalling through distinct efector func‑ stimulator of interferon gene; IRF3: interferon regulatory factor 3; TNFR: tumor tions of TRAF3 and TRAF6. Nature. 2006;439:204–7. necrosis factor receptor; TRAF: TNFR-associated factor; KO: knockout; DC: 9. Oganesyan G, Saha SK, Guo B, He JQ, Shahangian A, Zarnegar B, et al. dendritic cell; BMDC: bone marrow derived DC; BMDM: bone marrow derived Critical role of TRAF3 in the Toll-like receptor-dependent and -independ‑ macrophage. ent antiviral response. Nature. 2006;439:208–11. 10. Liu S, Chen J, Cai X, Wu J, Chen X, Wu YT, et al. MAVS recruits multi‑ Authors’ contributions ple ubiquitin E3 ligases to activate antiviral signaling cascades. eLife. XX designed and performed the experiments, prepared the fgures, and wrote 2013;2:e00785. part of the manuscript; JJ, LZ, ZJ, YL, and XC contributed to the performance 11. Zeng W, Xu M, Liu S, Sun L, Chen ZJ. Key role of Ubc5 and lysine-63 poly‑ of the experiments; and SCS supervised the work and wrote and edited the ubiquitination in viral activation of IRF3. Mol Cell. 2009;36(2):315–25. manuscript. All authors read and approved the fnal manuscript. 12. Xu Y, Cheng G, Baltimore D. Targeted disruption of TRAF3 leads to post‑ natal lethality and defective T-dependent immune responses. Immunity. Author details 1996;5:407–15. 1 Department of Immunology, The University of Texas MD Anderson Cancer 13. Gardam S, Sierro F, Basten A, Mackay F, Brink R. TRAF2 and TRAF3 signal 2 Center, 7455 Fannin Street, Box 902, Houston, TX 77030, USA. Life Sci‑ adapters act cooperatively to control the maturation and survival signals 3 ences Institute, Zhejiang University, Hangzhou 310058, China. Department delivered to B cells by the BAFF receptor. Immunity. 2008;28:391–401. of Biochemistry and Biophysics, Texas A&M University, College Station, TX, USA. 14. Clausen BE, Burkhardt C, Reith W, Renkawitz R, Forster I. Conditional 4 The University of Texas Graduate School of Biomedical Sciences at Houston, gene targeting in macrophages and granulocytes using LysMcre mice. Houston, TX 77030, USA. Transgenic Res. 1999;8(4):265–77. 15. Nakhaei P, Genin P, Civas A, Hiscott J. RIG-I-like receptors: sensing and Acknowledgements responding to RNA virus infection. Semin Immunol. 2009;21:215–22. We thank G Barber for VSV, J Bell for VSV AV1, and R Brink for Traf2-fox and 16. Feil S, Valtcheva N, Feil R. Inducible Cre mice. Methods Mol Biol. Traf3-fox mice. We also thank the personnel from the NIH/NCI-supported 2009;530:343–63. resources (fow cytometry, animal facility, and histology core facilities) under award number P30CA016672 at The MD Anderson Cancer Center. Xie et al. Cell Biosci (2019) 9:5 Page 10 of 10

17. Parvatiyar K, Pindado J, Dev A, Aliyari SR, Zaver SA, Gerami H, et al. A 24. Stojdl DF, Lichty BD, tenOever BR, Paterson JM, Power AT, Knowles S, et al. TRAF3-NIK module diferentially regulates DNA vs RNA pathways in VSV strains with defects in their ability to shutdown innate immunity are innate immune signaling. Nat Commun. 2018;9(1):2770. potent systemic anti-cancer agents. Cancer Cell. 2003;4:263–75. 18. Lalani AI, Moore CR, Luo C, Kreider BZ, Liu Y, Morse HC, et al. Myeloid cell 25. Guo X, Shu C, Li H, Pei Y, Woo SL, Zheng J, et al. Cyclic GMP-AMP amelio‑ TRAF3 regulates immune responses and inhibits infammation and tumor rates diet-induced metabolic dysregulation and regulates proinfamma‑ development in mice. J Immunol. 2015;194(1):334–48. tory responses distinctly from STING activation. Sci Rep. 2017;7(1):6355. 19. Sun SC. The noncanonical NF-kappaB pathway. Immunol Rev. 26. Jin J, Xie X, Xiao Y, Hu H, Zou Q, Cheng X, et al. Epigenetic regulation of 2012;246(1):125–40. the expression of Il12 and Il23 and autoimmune infammation by the 20. Grech AP, Amesbury M, Chan T, Gardam S, Basten A, Brink R. TRAF2 deubiquitinase Trabid. Nat Immunol. 2016;17(3):259–68. diferentially regulates the canonical and noncanonical pathways of NF- 27. Uhlik M, Good L, Xiao G, Harhaj EW, Zandi E, Karin M, et al. NF-kappaB- kappaB activation in mature B cells. Immunity. 2004;21:629–42. inducing kinase and IkappaB kinase participate in human T-cell 21. Su C, Zheng C. Herpes simplex virus 1 abrogates the cGAS/STING-medi‑ leukemia virus I Tax-mediated NF-kappaB activation. J Biol Chem. ated cytosolic DNA-sensing pathway via its virion host shutof protein, 1998;273:21132–6. UL41. J Virol. 2017. https​://doi.org/10.1128/JVI.02414​-16. 28. Xiao G, Harhaj EW, Sun SC. NF-kappaB-inducing kinase regulates the 22. Zhang J, Zhao J, Xu S, Li J, He S, Zeng Y, et al. Species-specifc deamida‑ processing of NF-kappaB2 p100. Mol Cell. 2001;7:401–9. tion of cGAS by Herpes Simplex Virus UL37 protein facilitates viral replica‑ tion. Cell Host Microbe. 2018;24(2):234 e5–248 e5. 23. Pan S, Liu X, Ma Y, Cao Y, He B. Herpes Simplex Virus 1 gamma134.5 pro‑ tein inhibits STING activation that restricts viral replication. J Virol. 2018. https​://doi.org/10.1128/JVI.01015​-18.

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions