Chlamydia trachomatis, an obligate intracellular bacterial species, is known to inhibit host cell apoptosis. However, the chlamydial antiapoptotic mechanism is still not clear. Because NF-κB activation is antiapoptotic, we tested the potential role of NF-κB activation in chlamydial antiapoptotic activity in the current study. First, no obvious NF-κB activation was detected in the chlamydia-infected cells when these cells were resistant to apoptosis induced via either the intrinsic or extrinsic apoptosis pathways. Second, inhibition of NF-κB activation with pharmacologic reagents failed to block the chlamydial antiapoptotic activity. Finally, NF-κB p65 gene deletion did not prevent chlamydia from inhibiting host cell apoptosis. These observations together have demonstrated that NF-κB activation is not required for the chlamydial antiapoptotic activity.
As an obligate intracellular bacterial pathogen with a distinct biphasic replication cycle, Chlamydia trachomatis consists of many serovars with tissue tropism mainly for either ocular or urogenital epithelia. The ocular infection with C. trachomatis is a major cause of the blinding trachoma in developing nations whereas the urogenital infection with C. trachomatis is a leading cause of sexually transmitted bacterial diseases in the developed world. It is thought that the inflammatory responses initiated and sustained by the chlamydia-infected cells may play a critical role in chlamydia-induced pathologies in humans (1, 2, 3, 4, 5). C. trachomatis has evolved various immune evasion strategies for achieving long-term survival in the infected hosts, including inhibition of phagolysosomal fusion (6, 7, 8, 9), sequestration of its own biosynthesis within host cell cytoplasmic vacuoles (6), and interference with host defense mechanisms (10, 11, 12, 13, 14) and signaling pathways (15, 16).
Apoptosis is an innate ability for all cells to orderly commit suicide and can be activated in response to either intracellular stress signals (intrinsic pathway; see Ref.17) or cell surface death receptor activation (extrinsic pathway; see Ref.18). In the intrinsic pathway, the intracellular death signals induce mitochondria to release apoptotic factors such as cytochrome c that can further participate in the activation of downstream caspases including caspase 9, caspase 6, and caspase 3. The activated caspase 3 can cleave DNA repairing enzymes such as poly(ADP)-ribose polymerase (PARP)3 and activate endonucleases, both of which lead to nuclear apoptosis. In the extrinsic pathway, the receptor-transmitted death signals can trigger the activation of upstream caspases such as caspase 8 and caspase 10. The activated upstream caspases can either enter the intrinsic pathway via the generation of tBid (19) or directly activate the downstream effector caspases such as caspase 3. Therefore, both the intrinsic and extrinsic apoptosis pathways can involve mitochondria and mitochondrial release of cytochrome c can be used to monitor both pathways. Besides its apparent role in tissue development and homeostasis, apoptosis is also recognized as an important effector mechanism in host defense against microbial infections (20). In innate immunity, epithelial cells can undergo apoptosis in response to intracellular invasion via the intrinsic apoptosis pathway so that the initial infection is prevented from spreading to the rest of the infected host. In adoptive immunity, Ag-specific T cells can induce the infected cells to undergo apoptosis via the extrinsic apoptosis pathway so that the intracellular pathogens are contained and cleaned up by competent phagocytes. Because of the critical roles of apoptosis in host defense, many intracellular pathogens have evolved mechanisms to counteract the host apoptosis responses (21, 22). We and others have recently shown that chlamydia has evolved the ability to prevent the infected cells from undergoing apoptosis (13, 14, 23, 24, 25). The chlamydial antiapoptotic activity was correlated with chlamydia-induced blockade of mitochondrial cytochrome c release (13) and inhibition of the mitochondrial Bax and Bak protein activation (26). However, the precise mechanisms of the chlamydial antiapoptotic activity remain unclear.
The transcription factor NF-κB regulates the expression of many genes (27), including genes coding for inflammatory cytokines and antiapoptotic factors such as the apoptosis inhibitor survivin, antiapoptotic adaptor molecules TNFR-associated factor (TRAF)1 and TRAF2, caspase inhibitors cIAPs, and the mitochondrial antiapoptotic Bcl-xL protein (28, 29). Up-regulation of these antiapoptotic factors may mainly contribute to the NF-κB antiapoptotic activity. Although activation of TNFR1 can lead to upstream caspase activation and apoptosis through the extrinsic pathway (30), TNF-α is not cytotoxic to most cells. This is because activation of TNFR1 also induces the activation of the NF-κB pathway. It is now known that activation of NF-κB serves as a major mechanism to protect cells against apoptosis induced by TNF-α and inhibition of NF-κB function with pharmacologic reagents including the protein synthesis inhibitor cycloheximide, dominant negative constructs of IκBα or NF-κB gene deletion can greatly increase cellular susceptibility to apoptosis induction by TNF-α (31, 32, 33). Physiological regulation of NF-κB function occurs at multiple levels (34). NF-κB is normally sequestrated in cell cytosol by associating with IκB. Upon IκB phosphorylation, ubiquitination, and finally degradation in response to upstream signals such as the signal generated from TNFR1 activation, NF-κB translocates into the nucleus to exert its transactivation activity. The antiapoptotic activity of NF-κB has been shown to be important in the inhibition of host cell apoptosis induced by various pathogens including Rickettsia (35, 36) and Helicobacter (37). We tested whether host NF-κB activation is also required for the C. trachomatis antiapoptotic activity in the current study.
Materials and Methods
Cell culture, chlamydial infection, NF-κB inhibition, and apoptosis induction
The C. trachomatis L2 serovar was prepared as previously described (38) and used throughout the study. HeLa 229 (a human cervical carcinoma epithelial cell line; American Type Culture Collection), mouse embryonic fibroblast (MEF) cells or MEF with NF-κB p65 gene knockout (MEFp65ko) cells (33) were infected with chlamydial organisms at a multiplicity of infection (MOI) of 1 (for immunofluorescence microscopy) or 5 (for Western blot) or as indicated in individual experiments for different times as described in different experiments. The infected cells were treated with various chemicals/cytokines for inhibiting NF-κB activity and/or inducing apoptosis. The following three NF-κB inhibitors used in the current study were all from EMD Biosciences. Kamebakaurin (KA, catalog no. 420340) inhibits NF-κB by directly targeting the DNA-binding activity of p50 and blocks the expression of antiapoptotic NF-κB target genes (39). Parthenoline (PA, catalog no. 512732) is a potent nonantioxidant inhibitor of NF-κB (40). Helenadin (HE, catalog no. 374000) inhibits NF-κB DNA binding activity by selectively alkylating the p65 subunit (41). These chemicals were first titrated for both cytotoxicity and their abilities to inhibit TNF-α-induced NF-κB nuclear translocation in pilot experiments (see Fig. 3 and data not shown). The concentrations that showed no toxicity but maintained significant NF-κB inhibitory effects were used to evaluate the role of NF-κB activation in chlamydial antiapoptotic activity. In these experiments, KA (at a final concentration of 30 μM) and PA (10 μM) were used to pretreat the cells for 2 h before apoptosis induction. For activating the intrinsic apoptosis pathway, staurosporine (Sigma-Aldrich), a kinase inhibitor, was used to treat the cultures at a final concentration of 1 μM for 4–5 h. For activating the extrinsic apoptosis pathway, TNF-α (20 ng/ml; R&D Systems) plus cycloheximide (10 μg/ml; Sigma-Aldrich) were used to treat the HeLa cell cultures for 4–5 h whereas TNF-α alone (20 ng/ml) was used to treat the MEFp65ko cells for 4 h. At the end of these treatments, cultures were harvested for analyses as described below.
Western blot assay
The Western blot was conducted as previously described (10). Cell samples were harvested either with a SDS sample buffer (2% SDS, 50 mM DTT, and 0.1 mM sodium orthovanadate in 62.5 mM Tris-HCl, pH 6.8) for direct gel loading or by scraping off the cells into cold PBS for cell fractionation. A hypotonic buffer-based fractionation method was used to separate cell cytosolic proteins from nuclear proteins (42). Briefly, cells collected in cold PBS were pelleted at 800 × g for 4 min and gently resuspended in a hypotonic lysis buffer (10 mM HEPES at pH 7.5, 10 mM KCl, 3 mM MgCl2, 0.05% Nonidet P-40, 1 mM EDTA, 1 mM DTT, 0.1 mM sodium orthovanadate and various protease inhibitors) by inversing the tubes repeatedly. After incubation on ice for 30 min, the cell suspensions were pelleted as described and the supernatants were collected as cytoplasmic (C) fraction. The pellets were rinsed with the same hypotonic lysis buffer except with Nonidet P-40 at 0.1%. The final pellets were dissolved in SDS sample buffer (as described) as the nuclear (N) fraction. The genomic DNA either in the nuclear fractions or whole cell lysates was sheered via sonication on ice three times with 5 s each. The final cell samples were loaded onto SDS polyacrylamide gels for electrophoresis separation. The separated protein bands were blotted onto nitrocellular membranes for detection with the following primary Abs: mouse anti-caspase 3 for HeLa cell samples (catalog no. C31720, 1/1000; BD Transduction Laboratories), rabbit anti-caspase 3 for MEF cell samples (catalog no. 9665, 1/2000; Cell Signaling Technology), rabbit anti-caspase 6 (catalog no. 9761, 1/1000; Cell Signaling Technology), rabbit anti-caspase 9 (catalog no. 9501, 1/1000; Cell Signaling Technology), rabbit anti-PARP (catalog no. 9542, 1/2000; Cell Signaling Technology), mouse anti-HSP70 (catalog no. SC-24, 1/6000; Santa Cruz Biotechnology), goat anti-NF-κBp65 (catalog no. 372G, 1/500; Santa Cruz Biotechnology), mouse anti-histone (catalog no. MAB052, 1/500; Chemicon International), rabbit anti-total IκBα (catalog no. 371, 1/1000; Santa Cruz Biotechnology), mouse anti-phosphorylated (ser32/36) IκBα (catalog no. 9246, clone 5A5, 1/200; Cell Signaling Technology) and a mouse anti-chlamydial major outer membrane protein (clone MC22; G. Zhong, unpublished data). The primary Ab bindings were detected with the corresponding secondary Abs conjugated with HRP (Jackson ImmunoResearch Laboratories) and visualized using the standard ECL (Santa Cruz Biotechnology).
The immunofluorescence microscopy was conducted as described (15). Cells grown on coverslips were processed for labeling with the following primary Abs: mouse anti-cytochrome c (clone 6H2B4 at 1/3000; BD Pharmingen), rabbit anti-chlamydia (D395 at 1/2000; J. Sharma and G. Zhong, unpublished data), mouse anti-NF-κB (catalog no. 8008, 1/800; Santa Cruz Biotechnology), rabbit anti-NF-κB (catalog no. 372, 1/2000; Santa Cruz Biotechnology). For costaining with multiple primary Abs, the primary Ab species were matched so that corresponding Cy2-conjugated (green) or Cy3-conjugated (red) secondary Abs (Jackson ImmunoResearch Laboratories or Molecular Probes) can be used to distinguish the different primary Abs. The DNA was always labeled blue with the Hoechst dye (Sigma-Aldrich). The images were acquired with a Hamamatsu CCD camera under an AX-70 Olympus fluorescence microscope as single color in gray and overlayed as multicolor images using the software Simple PCI. In some experiments, the coverslips were subjected to cell counting. For each coverslip, five random views with a total of ∼150 cells were counted for scoring apoptosis rates in different cell populations. Three independent experiments were conducted for acquiring the quantitation data. The data were expressed as mean + SE.
C. trachomatis infection inhibits host cell apoptosis mediated by both intrinsic and extrinsic pathways
We measured the C. trachomatis antiapoptotic activity in HeLa cells infected with serovar L2 and treated with either staurosporine (to activate the intrinsic) or TNF-α plus cycloheximide (to activate the extrinsic death pathway). As shown in Fig. 1,A, C. trachomatis infection significantly blocked the cleavage of PARP, a measurement of caspase 3 activity and activation of various downstream caspases including the caspase 3, caspase 6, and caspase 9 induced by either staurosporine (Fig. 1, lane 3 vs lane 4) or TNF-α plus cycloheximide (Fig. 1, lane 5 vs lane 6). The chlamydial inhibition of these downstream caspase activities was further correlated with chlamydial blockade of mitochondrial cytochrome c release on an immunofluorescence microscopy assay (Fig. 1,B). Because cytochrome c is normally localized in the intermembrane space of mitochondria, cytochrome c staining revealed the structure of mitochondria in normal cells (Fig. 1,Bg, red). Although infection with chlamydia did not alter the mitochondrial localization of cytochrome c (Fig. 1,Bh), treatment with staurosporine or TNF-α plus cycloheximide induced both nuclear apoptosis (nuclear condensation; Fig. 1,B, c and e) and mitochondrial cytochrome c release (Fig. 1,B, i and k, red smear staining) in the normal HeLa cell cultures. However, in the infected cultures, only the uninfected but not the infected cells were induced to undergo apoptosis responses (Fig. 1 B, d, j, f, and l). These observations have confirmed our previous findings that chlamydial infection can prevent host cell apoptosis via the inhibition of mitochondrial cytochrome c release (13).
NF-κB is not significantly activated in C. trachomatis-infected cells when these cells are resistant to apoptosis
To test the role of NF-κB activation in the C. trachomatis antiapoptotic activity, we evaluated whether C. trachomatis infection caused NF-κB nuclear translocation, an indicator of NF-κB activation. To our surprise, no significant migration of cytosolic NF-κB molecules into nuclei was detected in HeLa cells infected with the C. trachomatis serovar L2 from 12 to 60 h as revealed by an immunofluorescence assay (Fig. 2,A, a–e and h–l). As a positive control, stimulation with TNF-α at 20 ng/ml for 30 min caused most cytosolic NF-κB to enter nuclei (Fig. 2,A, f and m). Interestingly, TNF-α also induced efficient nuclear localization of NF-κB in chlamydia-infected cells (Fig. 2,A, g and n), suggesting that C. trachomatis infection does not affect the TNF-α-induced NF-κB activation signaling pathway. A cell fractionation plus Western blot detection approach was used to confirm this observation (Fig. 2,B). In this experiment, cells with or without C. trachomatis infection and TNF-α stimulation were fractionated into nuclear (N) and cytosolic (C) fractions and the NF-κB p65 molecules in these fractions and whole cell lysates (W) were detected on a Western blot. Histone was used as a nuclear marker. We found that histone was only detected in either the whole cell lysates (Fig. 2,B, lanes 1, 4, 7, and 10) or nuclear fraction (Fig. 2,B, lanes 3, 6, 9, and 12) but not the cytosolic fractions (Fig. 2,B, lanes 2, 5, 8, and 11), indicating a clear separation of cytosolic fraction from nucleic contents. An equal amount of HSP70 protein was detected in each lane, indicating equal total protein loading. Despite C. trachomatis infection, the majority of NF-κB molecules were consistently detected in the nuclear fraction of the TNF-α-stimulated cells (Fig. 2,B, lanes 6 and 12) but in the cytosolic fraction of the unstimulated cells (Fig. 2,B, lanes 2 and 8). This observation has confirmed that C. trachomatis infection does not cause NF-κB nuclear localization nor affect the TNF-α-induced NF-κB nuclear migration. As an independent measurement of NF-κB activation, we detected IκBα phosphorylation and degradation in C. trachomatis-infected cells (Fig. 2,C). Neither IκBα phosphorylation nor degradation was detected in chlamydia-infected cells over an infection course of 60 h (Fig. 2,C, lanes 1-5), which is consistent with the earlier observations that C. trachomatis infection alone does not activate the NF-κB signaling pathway. Furthermore, there was no significant difference in the TNF-α-induced phosphorylation and degradation of IκBα between C. trachomatis-infected and uninfected cells (comparing Fig. 2 C, lanes 8 and 9 with lanes 6 and 7), which is consistent with the above observations that C. trachomatis infection does not affect the TNF-α-induced NF-κB signaling pathway.
Inhibition of NF-κB fails to block the C. trachomatis antiapoptotic activity
Various chemical inhibitors, including KA, PA, and HE known to inhibit NF-κB activation, were evaluated for their ability to block TNF-α-induced NF-κB activation in HeLa cells (Fig. 3,A). Pretreatment of HeLa cells with any of these reagents at the nontoxic concentrations significantly blocked the NF-κB nuclear translocation induced by TNF-α. We then used the same treatment conditions to evaluate the effects of the inhibitors on the C. trachomatis ability to inhibit apoptosis induced by staurosporine or TNF-α plus cycloheximide (Fig. 3,B). Both mitochondrial cytochrome c release (Fig. 3,B, a–j, red) and nuclear condensation (Fig. 3,B, u–ad, blue) were used to monitor host cell apoptosis and NF-κB localization was simultaneously monitored (Fig. 3,B, k–t, green). The single color images were displayed in gray and the overlaid tricolor images (Fig. 3,B, ae–an, bottom row) are shown. Activation of the intrinsic apoptosis pathway with staurosporine induced most normal HeLa cells to undergo apoptosis (Fig. 3,B, c, w, and ag). However, in the chlamydia-infected cultures, only the uninfected cells were induced to undergo apoptosis and the infected cells completely resisted the apoptosis induction (Fig. 3,B, d, x, and ah). Pretreatment with either KA (30 μM) or PA (10 μM) did not alter the C. trachomatis ability to inhibit apoptosis in the infected cells (Fig. 3,B, e, y, ai, f, z, and aj). Similarly, activation of the extrinsic death pathway with TNF-α plus cycloheximide induced apoptosis of the uninfected HeLa cells in both normal culture (Fig. 3,B, g, aa, and ak) and the infected culture (Fig. 3,B, h, ab, and al) but failed to induce apoptosis of the chlamydia-infected cells (Fig. 3,B, h, ab, and al). Unlike the staurosporine, TNF-α plus cycloheximide induced NF-κB nuclear localization in the infected cells (see arrows, Fig. 3,B, r and al). Pretreatment with either KA or PA, although significantly blocking the TNF-α plus cycloheximide-induced NF-κB nuclear localization (arrows, Fig. 3,B, s and t), failed to prevent the chlamydial inhibition of apoptosis in the infected cells (Fig. 3,B, i, ac, am, j, ad, and an). We further quantitated the effects of the inhibitors KA and PA on the C. trachomatis antiapoptotic activity by counting five random views per coverslip from three independent experiments (Fig. 3,C). Staurosporine induced ∼80–90%, whereas TNF-α plus cycloheximide induced ∼70–80%, of the uninfected cells in either normal cultures (Fig. 3,C, samples 3 and 7, □) or infected cultures (Fig. 3,C, samples 4–6 and 8–10, □) to undergo apoptosis. However, the apoptosis rate of the infected cells was <5% regardless of the apoptosis inducers and NF-κB inhibitors used (Fig. 3 C). These quantitative data have confirmed that the C. trachomatis antiapoptotic activity is not altered by inhibition of NF-κB function.
Cells that are both infected with C. trachomatis and deficient in the NF-κB p65 gene still resist apoptosis induction by TNF-α
We used an MEF cell line deficient in the NF-κB p65 knockout gene (MEFp65ko; see Ref.33) to further evaluate the role of NF-κB activation in C. trachomatis antiapoptotic activity (Fig. 4). It has been previously shown that MEFp65ko cells can be induced to undergo apoptosis with TNF-α alone (33). Indeed, TNF-α alone induced both PARP cleavage and caspase 3 activation in the MEFp65ko cells (Fig. 4,A, lane 3). Infection with C. trachomatis significantly reduced these apoptosis events (Fig. 4,A, lane 4). The wild-type MEF cells did not show any apoptosis responses regardless of infection and apoptosis induction (Fig. 4,A, lanes 5–8), demonstrating that neither the infection nor TNF-α alone or in combination is cytotoxic to MEF cells. The C. trachomatis antiapoptotic activity detected in the MEFp65−/− cells was further evaluated using an immunofluorescence assay at a single cell level (Fig. 4,B). No significant apoptosis was detected in MEFp65ko cells alone or cells with C. trachomatis infection (Fig. 4,B, a and b). TNF-α treatment induced apoptosis of most uninfected cells in both the normal culture and the infected culture (Fig. 4,B, c and d). However, the infected cells consistently resisted the apoptosis induction as demonstrated by a lack of both mitochondrial cytochrome c release and nuclear condensation (Fig. 4,Bd). The C. trachomatis antiapoptotic activity in MEFp65ko cells was quantitated by counting five random views per coverslip in three independent experiments (Fig. 4,C). Despite the fact that TNF-α induced >70% apoptosis in the uninfected cells from both the uninfected and infected cultures (Fig. 4,C, sample c and d), <5% apoptosis was detected in the C. trachomatis-infected cells (Fig. 4 C, sample d). These observations have demonstrated that the C. trachomatis antiapoptotic activity is independent of NF-κBp65.
Since the C. trachomatis antiapoptotic activity was first described over 5 years ago (13), extensive efforts have been made to understand the mechanisms. Although the antiapoptotic activity has been correlated with the C. trachomatis-induced blockade of mitochondrial cytochrome c release (13) and inhibition of Bax and Bak, two proapoptotic members of the mitochondrial Bcl-2 family (26), the precise molecular mechanisms of C. trachomatis antiapoptotic activity are still unknown. Because the NF-κB antiapoptotic activity is broad and activation of NF-κB can lead to the inhibition of death receptor-activated caspases (28), we tested whether C. trachomatis antiapoptotic activity is dependent on host NF-κB activation. We have presented compelling evidence to demonstrate that the C. trachomatis antiapoptotic activity does not require host NF-κB activation. First, neither significant NF-κB nuclear localization nor IκBα phosphorylation/degradation was detected in C. trachomatis-infected cells, indicating that C. trachomatis infection does not cause obvious NF-κB activation. Second, inhibition of NF-κB activation with pharmacological reagents did not affect the C. trachomatis ’ ability to inhibit host cell apoptosis. Third, deletion of the NF-κBp65 gene did not block the C. trachomatis antiapoptotic activity. These results are consistent with the observations that chlamydial infection did not induce significant up-regulation of the NF-κB-regulated antiapoptotic genes such as TRAFs, Bcl-2, Bcl-xL, and survivin although various microarray-based assays have revealed expression alternations of many other host genes during chlamydial infection (43, 44, 45). Previous studies from our own group have also shown that neither Bcl-2 nor Bcl-xL protein expression is up-regulated in C. trachomatis-infected cells (Ref.13 and data not shown). In fact, the NF-κB antiapoptotic activity is not always required for pathogen-induced inhibition of host cell apoptosis. For example, Although NF-κB activation seems to play an essential role in the inhibition of host apoptosis induced by Helicobacter pylori (37) and Rickettsia rickettsii (35), the antiapoptotic activity of HSV-1 is clearly not dependent on NF-κB activation (42).
Apparently, the C. trachomatis antiapoptosis mechanism is different from that of R. rickettsii, another obligate intracellular pathogen. R. rickettsii infection induces NF-κB activation (46) whereas NF-κB activation has not been clearly demonstrated in C. trachomatis-infected cells and our current study shows that NF-κB is not activated from 12 to 60 h after C. trachomatis infection (Fig. 2). Furthermore, the antiapoptotic activity induced by R. rickettsii infection disappeared when NF-κB activity was inhibited (35, 36) whereas inhibition of NF-κB or deletion of NF-κB p65 gene did not alter the C. trachomatis antiapoptotic activity (Figs. 3 and 4 of the current study). Interestingly, infection with Chlamydia pneumoniae induced activation of NF-κB and expression of NF-κB-regulated genes in a monocytic cell line (47). Furthermore, the C. pneumoniae-induced NF-κB activation was correlated with the survival of C. pneumoniae-infected Mono Mac 6 cells (48). These studies appear to suggest a role of NF-κB in C. pneumoniae antiapoptotic activity, which, however, has not been directly demonstrated (25, 49, 50). The fact that inhibition of host protein synthesis with cycloheximide, as routinely used for growing up C. pneumoniae organisms, does not increase the apoptosis of C. pneumoniae-infected cells, indicates that NF-κB-activated gene expression is not required for the C. pneumoniae-infected cells to remain alive. Furthermore, the C. pneumoniae-induced NF-κB activation is not dependent on live organism infection (51) whereas the C. pneumoniae antiapoptotic activity clearly requires live organism replication (25). Finally, due to the difficulty in achieving a high infection rate and establishing a productive infection with C. pneumoniae organisms in the absence of cycloheximide, it is not clear at this moment whether the NF-κB activation detected in the C. pneumoniae-infected cultures came from the infected or the uninfected cells. However, the C. pneumoniae-antiapoptotic activity is known to restrict to the infected cells only (25). Therefore, more studies are required to determine the relevance of NF-κB activation in C. pneumoniae antiapoptotic activity.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by National Institutes of Health Grants R01AI47997, R01HL64883, and R21AI57450 (to G.Z.).
Abbreviations used in this paper: PARP, poly(ADP)-ribose polymerase; TRAF, TNFR-associated factor; KO, knockout; MEF, mouse embryonic fibroblast; MOI, multiplicity of infection.