A Novel Family of Viral Death Effector Domain-containing Molecules That Inhibit Both CD-95- and Tumor Necrosis Factor Receptor-1-induced Apoptosis*

Molluscum contagiosum virus proteins MC159 and MC160 and the equine herpesvirus 2 protein E8 share substantial homology to the death effector domain present in the adaptor molecule Fas-associated death domain protein (FADD) and the initiating death protease FADD-like interleukin-1 (cid:98) -converting enzyme (FLICE) (caspase-8). FADD and FLICE participate in generating the death signal from both tumor necrosis factor recep-tor-1 (TNFR-1) and the CD-95 receptor. The flow of death signals from TNFR-1 occurs through the adaptor molecule tumor necrosis factor receptor-associated death domain protein (TRADD) to FADD to FLICE, whereas for CD-95 the receptor directly communicates with FADD and then FLICE. MC159 and E8 inhibited both TNFR-1-and CD-95-induced apoptosis as well as killing mediated by overexpression of the downstream adaptors TRADD and FADD. Neither viral molecule, however, inhibited FLICE-induced killing, consistent with an inhibitory ac-tion upstream of the active death protease. These data suggest the existence of a novel strategy employed by viruses to attenuate host immune killing mechanisms. Given that bovine herpesvirus 4 protein E1.1 and Kaposi’s sarcoma associated-herpesvirus protein K13 also possess significant homology to the viral inhibitory molecules

Cell suicide is a defense mechanism employed by host cells to inhibit viral replication and persistence. As a consequence, viruses have evolved numerous strategies to attenuate apopto-sis (1). For example, the Epstein-Barr virus (EBV) encodes BHRF1, a homolog of the mammalian anti-apoptosis molecule bcl-2, and the cowpox virus encodes a serpin-like protein, CrmA, that blocks apoptosis by inhibiting proteases belonging to the caspase family.
Molluscum contagiosum virus (MCV) is the only poxvirus family member still associated with human disease (2). It usually causes asymptomatic cutaneous neoplasms that can spontaneously regress. However, with the advent of immunocompromised populations, particularly those afflicted with acquired immunodeficiency syndrome, MCV infection has become a clinical challenge (3). Unfortunately, due to the inability to grow the virus in tissue culture cells and the lack of a suitable animal model, little is known about host-virus relationships (4). Equine herpesvirus 2 (EHV2) is a member of the ␥-herpesvirus subfamily that also includes herpesvirus saimiri, EBV, Kaposi's sarcoma-associated herpesvirus (KSHV), and bovine herpesvirus 4 (5,6). Although EHV2 is ubiquitously distributed and has been implicated as a pathogen in immunosuppressed states, its mode of evading the host immune response is uncertain. However, the recent availability of the MCV and EHV2 genome sequences has begun to identify genes that suggest potential pathogenic mechanisms (7,8).
MCV, surprisingly, does not encode many of the immunoregulatory molecules present in other poxviruses, especially those that antagonize the host cytokine-mediated inflammatory response. These include CrmA and a soluble TNFR-like 1 molecule (7). In contrast, EHV2 encodes an interleukin-10-like factor that may attenuate the host immune response (8). Regardless, MCV and EHV2 do not encode previously identified inhibitors of apoptosis (1). Instead, MCV and EHV2 encode novel members of an emerging family of molecules characterized by the presence of a death effector domain (DED) originally identified in signaling molecules engaged by the death receptors TNFR-1 and CD-95 (7,8).
Both TNFR-1 and CD-95 contain a stretch of approximately 60 -80 amino acids within their cytoplasmic domains termed the death domain. Upon activation the receptor death domains bind to corresponding death domains within the adaptor molecules TRADD (for TNFR-1) and FADD (for CD-95) (9 -12). Utilizing the same mechanism, TRADD is able in turn to recruit FADD to the TNFR-1 signaling complex (13). FADD appears to play a central role as a conduit for death signals from both receptors as dominant negative versions that retain the death domain but lack the amino-terminal segment effectively attenuate both TNFR-1-and CD-95-induced killing (14). Since it is likely that the amino-terminal domain of FADD functions to engage downstream components of the death pathway, it has been termed the DED (14). The importance of this domain was dramatically underscored by the discovery of its presence within the prodomain of the death protease FLICE (15)(16)(17). It appears that the DED of FADD binds to the corresponding DED motif within the FLICE prodomain and thereby recruits this death protease to the receptor signaling complex. There-* This work was supported by National Institutes of Health Grants AG13671 and ES08111. 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.
‡ Equal contribution was made by the first two authors. fore, a homophilic binding mechanism involving DEDs is responsible for assembly of the receptor death signaling complex. Disruption of such a complex by DED-containing viral gene products could potentially abrogate propagation of the death signal.
In Vitro Binding Assay-Full-length FADD and truncated N-FADD were expressed as GST-fusion proteins as described previously (18).
Transfection, Coimmunoprecipitation, and Western Analysis-Transient transfections of 293T cells were performed as described previously (19). Cells were harvested 40 h following transfection, immunoprecipitated with the indicated antibodies, and analyzed by immunoblotting.
Cell Death Assay-For CD-95, TRADD, and FLICE killing, experiments were performed in 293-EBNA cells and in 293 cells for TNFR-1 and FADD killing. cDNAs encoding putative apoptosis inducers (0.5-0.8 g) and potential inhibitors (2.5 g) were cotransfected in each experiment together with the reporter plasmid pCMV ␤-galactosidase. Cells were fixed and stained 24 -30 h following transfection. The percentage of apoptotic cells was determined by calculating the fraction of round membrane-blebbed blue cells as a function of total blue cells. All assays were evaluated in duplicate and the mean and standard deviation calculated.

DED-containing Viral Inhibitors of Apoptosis 9622
of these viral proteins contains a highly conserved module RXDL/I(L) (X is any amino acid) that is also conserved in the DEDs of FADD and FLICE. It appears that many other herpesviruses also encode DED-like molecules. Examples include the herpesvirus saimiri protein VG71 and human herpesvirus 6 protein U15 (22,23).
E8 and MC159 Inhibit TRADD and FADD Killing but Not FLICE Killing-Additional studies were undertaken to delineate the point at which MC159 and E8 were exerting their inhibitory effect on the TNFR-1-and CD-95-induced death pathways. As shown in Fig. 3, both MC159 and E8 significantly blocked both TRADD and FADD killing, suggesting that these inhibitors must function downstream of these adaptor molecules. In contrast, MC159 and E8 did not inhibit FLICE-induced death, suggesting that they must act upstream of active FLICE. The overexpression of FLICE zymogen results in autoactivation to the active protease that is potently inhibited by the viral serpin CrmA (Fig. 3C).
To demonstrate the association of the viral inhibitory molecules with FADD or FLICE in vivo, 293 cells were transiently transfected with expression constructs encoding epitope-tagged versions of the respective molecules (Fig. 4). Consistent with the in vitro binding results, MC159 precipitated with FADD (Fig. 4B), but not with FLICE (data not shown). Conversely, E8 strongly associated with FLICE (Fig. 4C), but not with FADD (data not shown). This binding specificity of MC159 and E8 suggested that distinct mechanisms were employed by these two inhibitors. MC159 binds to FADD and presumably blocks its interaction with FLICE. The reverse is probably true for E8 in that it binds FLICE and inhibits its interaction with FADD. However, when FLICE is overexpressed (upon transfection), the binding of E8 is unable to overcome the propensity of this caspase to autoactivate (Fig. 3C). Therefore, once FLICE is active, E8 has no inhibitory influence. Regardless, either mechanism would disrupt the assembly of the receptor⅐FADD⅐ FLICE signaling complex and abrogate activation of downstream caspases. Further studies will be needed to substantiate these proposed mechanisms.