The Full-Length Isoform of Human Papillomavirus 16 E6 and Its Splice Variant E6* Bind to Different Sites on the Procaspase 8 Death Effector Domain

Human papillomavirus 16 is a causative agent of most cases of cervical cancer and has also been implicated in the development of some head and neck cancers. The early viral E6 gene codes for two alternatively spliced isoforms, E6_large and E6*. We have previously demonstrated the differential effects of E6_large and E6* binding on the expression and stability of procaspase 8, a key mediator of the apoptotic pathway. Additionally, we have reported that E6 binds to the FADD death effector domain (DED) at a novel E6 binding domain. Sequence similarities between the FADD and procaspase 8 DEDs suggested a specific region for E6_large/procaspase 8 binding, which was subsequently confirmed by mutational analysis as well as by the ability of peptides capable of blocking E6/FADD binding to also block E6_large/caspase 8 binding. However, the binding of the smaller isoform, E6*, to procaspase 8 occurs at a different region, as deletion and point mutations that disrupt E6_large/caspase 8 DED binding do not disrupt E6*/caspase 8 DED binding. In addition, peptide inhibitors that can block E6_large/procaspase 8 binding do not affect the binding of E6* to procaspase 8. These results demonstrate that the residues that mediate E6*/procaspase 8 DED binding localize to a different region on the protein and employ a separate binding motif. This provides a molecular explanation for our initial findings that the two E6 isoforms affect procaspase 8 stability in an opposing manner.The relationship between viruses and cancers is reflected in the observation that viral infections account for approximately 10 to 15% of the cancer burden worldwide (6, 60). This makes viral infections one of the preventable risk factors of cancer. Viruses are associated with several human malignancies, including hepatitis B and C virus-associated hepatocellular carcinomas (48), Epstein-Barr virus-associated nasopharyngeal carcinomas and lymphomas (36), and human T-cell leukemia virus-associated adult T-cell leukemia (8, 28). Although there is a correlation between infection and the onset of cancer, the frequency of infection supersedes the incidence of cancer inception, suggesting that the presence of the virus alone is not sufficient to trigger carcinogenesis. Progression from viral infection to tumor development therefore requires additional environmental and cellular factors in addition to the expression and activity of virus-encoded proteins (40).High-risk strains of human papillomavirus (HPV) (high-risk HPV HR-HPV]) such as HPV16 and HPV18 have been implicated in most cases of cervical cancer and also in a subset of head and neck cancers (24, 26, 39). Infection with oncogenic strains of HPV represents up to 75% of all infections. Furthermore, 1/10 of all deaths among women worldwide can be attributed to HR-HPV-related cancers (44, 45). The key players in promoting cell transformation and immortalization following HPV infection are the viral early proteins E6 and E7. These proteins are well known for their ability to interact with the tumor suppressor p53 or members of the retinoblastoma family of proteins including pRb, p107, and p130, respectively (3, 17, 41). In addition to p53, HR-HPV E6 (HR-E6) binds to a number of cellular proteins involved in various aspects of cell proliferation and virus survival (reviewed in references 34 and 53). Our laboratory has reported that E6 binds to key mediators of the apoptotic pathway including tumor necrosis factor (TNF) R1 (22), the FADD death effector domain (DED) (21), and the procaspase 8 DED (20) and, in doing so, impedes apoptosis from taking place.As noted above, HR-E6 binds to TNF R1, blocking the adaptor molecule TRADD from binding to the membrane receptor. Similarly, the binding of HR-E6 to the FADD DED, a molecule common to the TNF-, Fas L-, and TRAIL-mediated extrinsic pathways of apoptosis, leads to the accelerated degradation of FADD and thereby inhibits the binding of additional downstream molecules necessary for programmed cell death. Additionally, we have reported that HR-E6 binds to procaspase 8, another molecule common to all three receptor-mediated pathways. The importance of procaspase 8 can be demonstrated by the many proteins produced by viruses to either inactivate or inhibit this apoptotic mediator in order to evade clearance by the host immune response. Such proteins include the herpes simplex virus R1 subunit that interferes with caspase 8 activation (31); the molluscum contagiosum virus MC159 protein that binds to the DEDs of both FADD and procaspase 8, thereby inhibiting their interaction (25); the human herpesvirus 8 FLICE protein that obstructs procaspase 8 cleavage and prevents its activation (4); and the cowpox virus serpin CrmA, which, along with the human cytomegalovirus UL136 proteins, inhibits caspase 8 activation (50, 56). In a like manner, HR-HPV16 produces the early protein E6 that binds to procaspase 8. Interestingly, however, we have found that the two splice products of the E6 gene, E6_large, a protein of about 16 kDa, and E6*, a protein less than half the size of E6_large, bind to and affect procaspase 8 stability differentially. While the large isoform accelerates the degradation of procaspase 8, leading to its destabilization, the short isoform leads to the stabilization of protein expression and an increase in activity. These observations suggest that the bindings of these two E6 isoforms have different functional consequences and may well localize to different regions on procaspase 8.We have previously identified a novel E6 binding site on the FADD DED (54). Based on sequence comparisons between the DEDs of FADD and procaspase 8, we proposed that the binding motif that mediates oncoprotein binding to both proteins would be similar. To test this possibility, we performed a series of mutational and peptide competitor-based experiments and discovered that the motifs on caspase 8 and on FADD that mediate binding between E6 and its cellular partner are indeed similar. Interestingly, however, the motif by which E6* binds to procaspase 8 is located in another region of the protein. These findings provide a molecular explanation for our previously reported observations concerning the differential effects of the binding of each isoform to the procaspase 8 DED. These findings also demonstrate the ability of peptide inhibitors to successfully impair E6 binding to its cellular targets and contribute to the discovery of therapeutic agents that are effective against cervical cancer.