The interaction between human papillomavirus type 16 and FADD is mediated by a novel E6 binding domain

Abstract

High-risk strains of human papillomavirus, such as types 16 and 18, have been etiologically linked to cervical cancer. Most cervical cancer tissues are positive for both the E6 and E7 oncoproteins, since it is their cooperation that results in successful transformation and immortalization of infected cells. We have reported that E6 binds to tumor necrosis factor receptor 1 and to Fas-associated death domain (FADD) and, in doing so, prevents E6-expressing cells from responding to apoptotic stimuli. The binding site of E6 to FADD localizes to the first 23 amino acids of FADD and has now been further characterized by the use of deletion and site-directed mutants of FADD in pull-down and functional assays. The results from these experiments revealed that mutations of serine 16, serine 18, and leucine 20 obstruct FADD binding to E6, suggesting that these residues are part of the E6 binding domain on FADD. Because FADD does not contain the two previously identified E6 binding motifs, the LxxphiLsh motif, and the PDZ motif, a novel binding domain for E6 has been identified on FADD. Furthermore, peptides that correspond to this region can block E6/FADD binding in vitro and can resensitize E6-expressing cells to apoptotic stimuli in vivo. These results demonstrate the existence of a novel E6 binding domain.

FIG. 1.

Mutation of specific amino acids in the N terminus of FADD DED can block E6 binding. (A) E6 expression protects cells from Fas-induced apoptosis. Control cells not expressing E6 or cells expressing E6 under the control of the tetracycline/doxycycline response element, U2OSE6tet24, were grown in the presence of the indicated concentrations of doxycycline for 48 h. Cells were then treated with 50 ng of anti-Fas/ml in the presence of 5 μg of cycloheximide/ml. Cells were incubated for an additional 16 h prior to measuring cell viability via the MTT assay. Measurements were made in triplicate, and the error bars represent the standard deviation. (B) The protein sequences of the mutants that were created based on the sequence of D1 (lacking amino acids 23 to 62) are shown, along with the results of in vitro pull-down assays with E6 complexed to GST beads. The amino acids that were mutated in each construct are highlighted in red. The SELT, SSLS, and SLT2 names were chosen based on amino acids in the regions where the mutations were made.

FIG. 2.

The SLT2 set of five mutations in FADD DED that localize to the E6 binding domain occupy a patch on the surface of the FADD protein. The PDB has published the three-dimensional structure of FADD DED (accession number 1a1w). This structure was viewed utilizing the Viewer Lite program. The structure of the wild-type protein is shown (A), and the amino acids mutated in the various mutant constructs are highlighted in yellow (B, C, and D).

FIG. 3.

Two combinations of three amino acid mutations in FADD inhibit E6 binding. (A) Schematic representation of the additional FADD DED mutants used to map the region of E6 binding. The protein sequence of the mutants that were created based on the sequence of D1 (lacking amino acids 23 to 62) are shown, along with the results of in vitro pull-down assays with E6 complexed to GST beads. The amino acids that were mutated in each construct are highlighted in red. (B) Mutating three amino acids in the N terminus of FADD DED inhibits E6 binding to FADD. Glutathione bead-bound GST was used to pull-down bacterially expressed and purified wild-type FADD (D1) protein (lane 1); GST-E6 was used to pull down purified wild-type D1 and FADD variants SLT2, SLT3, and SLT4 (lanes 2 to 5), as described previously. For the upper panel, blots were probed, following SDS-PAGE separation of proteins and transfer to membranes, with antibodies to FADD. For the lower panel, the same membrane was stripped and reprobed with antibodies to GST. (C) AlphaScreen technology verifies that E6 cannot bind to the FADD mutants SLT2, SLT3, and SLT4. GST-tagged E6 at 10⁻³ μM was incubated for 1 h with 0.5, 0.1, or 0.02 μM His-tagged D1 protein at room temperature. Glutathione-coated donor beads and nickel-coated acceptor beads were then added, and plates were read on an EnVision multilabel plate reader after an overnight incubation period in the absence of light.

FIG. 4.

Two combinations of three amino acid mutations in FADD DED inhibit E6-mediated FADD degradation, while leaving other tested biological functions of FADD intact. (A) E6 cannot bind to SLT3 or SLT4 and thus does not mediate the accelerated degradation of the mutant FADD proteins. For the upper panel, U2OSE6tet24 cells cultured with 0 or 100 ng of doxycycline/ml were transfected with plasmids encoding wild-type FADD, the SLT3 mutant, or the SLT4 mutant FADD protein. At 24 h posttransfection, cell lysates were prepared for immunoblotting, and subsequent membranes were blotted with antibodies against FADD to analyze FADD content. For the lower panel, transfection experiments were repeated three times, and densitometric analysis of the resultant protein bands was performed by using a ChemiImager 4400 (AlphaInnotech Corp.). Error bars represent the standard deviations obtained from three separate experiments. (B) The three amino acid mutations in SLT4 do not inhibit the ability of FADD to undergo normal turnover. U2OS cells were transfected with pcDNA FADD wild type (lanes 1 to 3), pcDNA SLT3 (lanes 4 to 6), or pcDNA SLT4 (lanes 7 to 9), as described before. At 24 h posttransfection, cycloheximide (5 μg/ml) was added to inhibit de novo protein synthesis, and cells were then lysed at the indicated time points and used for immunoblotting. The subsequent membrane was blotted with antibodies against FADD to analyze FADD content (top panel). For the center panel, the same membrane was stripped and reblotted with antibodies against β-actin to normalize for loading. For the bottom panel, the same membrane was stripped and reblotted with antibodies against green fluorescent protein (GFP) to demonstrate transfection efficiency. (C) The mutations made in SLT3 and SLT4 do not inhibit the ability of FADD to trigger apoptosis. U2OS cells were transfected with either pcDNA (empty vector), pcDNA FADD, pcDNA SLT3, or pcDNA SLT4. At 24 h posttransfection, cell viability was measured via the MTT assay. Measurements were made in triplicate, and the error bars represent the standard deviation. (D) The amino acids mutated in SLT3 and SLT4 do not eliminate the ability of FADD DED to bind to procaspase 8 DED. Sequences encoding procaspase 8 DED, wild-type FADD DED, the SLT3 mutant, or the SLT4 mutant FADD DED were cloned into the bait or prey plasmids of a mammalian two-hybrid kit (Stratagene). The indicated combination of plasmids, along with a plasmid encoding the luciferase reporter gene, was transfected into cells that express E6 under the control of the tetracycline/doxycycline response element. Cells were treated with the indicated concentrations of doxycycline to regulate E6 expression. Luciferase expression was measured by using a luminometer to detect chemiluminescence. Measurements were made in triplicate, and the error bars represent the standard deviation.

FIG. 5.

A specific combination of amino acids in the N terminus of FADD DED mediates E6/FADD binding. (A) Various mutant FADD constructs designed to localize the amino acids which facilitate E6 binding to FADD DED. Mutation of different combinations of amino acids in the N-terminal 23 amino acids of FADD DED lead to a reduction in the ability of E6 to bind FADD. (B) The three-dimensional structure of FADD DED as viewed with the Viewer Lite program with the amino acids mutated in SLT4 depicted in blue and those mutated in SLT2 but not SLT4 shown in yellow.

FIG. 6.

Peptides can be used to block the interaction between E6 and FADD DED in vitro. (A) Peptides synthesized and tested for the ability to obstruct binding between E6 and FADD. Peptides A and B correspond to the proposed E6 binding domain on FADD. Peptides C and D correspond to the E6 binding domain on E6AP. (B) Peptides which mimic the E6 binding domain on FADD can inhibit E6/FADD interaction. Peptides A, B, C, and D at 0, 10, or 25 μM was added to a mixture of purified His-tagged D1 (0.4 μM) and GST-tagged E6 proteins (10⁻³ μM) (as described for Fig. 3C). After the specified incubation period, the addition of beads, and an overnight incubation in the absence of light, the plates were read on an EnVision multilabel plate reader. (C) Peptide A does not inhibit E6/E6-AP binding. Peptide A at 0, 10, or 25 μM was added to a mixture of purified His-tagged D1 and GST-tagged E6 proteins (as described for Fig. 3C) or to a mixture of purified His-tagged E6AP and GST-tagged E6 proteins. Peptides C and D, known inhibitors of E6/E6AP binding, at 0, 10, or 25 μM were added to mixtures of His-tagged E6AP and GST-tagged E6. (His-D1 at 0.5 μM, His-E6AP at 0.5 μM, and μM GST-E6 at 10⁻³ μM were used.) Plates were read as in panel B after an overnight incubation. (D) A peptide containing mutations in the amino acids implicated in mediating E6 binding to FADD supports our findings of a novel E6 binding domain. For the top panel, peptides were synthesized and tested for the ability to obstruct binding between E6 and FADD (Mimotopes). For the bottom panel, peptide 1 (containing a short fragment of the wild-type FADD protein sequence) and peptide 2 (containing a short fragment of the FADD sequence with the 5 amino acid changes introduced in the SLT2 construct) at 0, 50, 100, or 400 μM were added to mixtures of His-tagged D1 (0.4 μM) and GST-tagged E6 (10⁻³ μM). Plates were read as in panel B after an overnight incubation.

FIG. 7.

Overexpressing the region of FADD implicated in E6 binding in cells blocks the interaction between E6 and FADD DED and resensitizes cells to apoptosis inducing stimuli. (A) The pcDNA3 vector encoding the 23 amino acid residues on FADD implicated in E6 binding expresses the novel E6 binding domain. The amino acids in the E6 binding domain of FADD, based on the mutations introduced in SLT4, are highlighted in red. (B) Overexpression of the peptide corresponding to the E6 binding domain in cells interferes with E6-mediated FADD degradation. U2OS cells expressing either the sense (U2OSE612) or antisense (U2OSE6 AS) version of E6 were either transfected (lanes 1 and 3) or not transfected (lanes 2 and 4) with pcDNA 23 aa. At 24 h posttransfection, cell lysates were prepared, proteins separated by SDS-PAGE and transferred to a membrane. The membrane was then blotted with antibodies against FADD to analyze FADD expression (top panel). The same membrane was stripped and reblotted with antibodies against β-actin to demonstrate that equivalent amounts of lysate were used for analysis (bottom panel). (C) Expression of the implicated E6 binding domain on FADD DED resensitizes E6-expressing cells to Fas-induced apoptosis. HPV-negative C33A cells or HPV-positive Caski and SiHa cells were transfected with pcDNA 23 aa or with pcDNA vector alone. At 24 h posttransfection, cells were treated with 50 ng of anti-Fas/ml in the presence of 5 μg of cycloheximide/ml. Cells were incubated for an additional 16 h prior to measuring cell viability via the MTT assay. Measurements were made in triplicate, and the error bars represent the standard deviation. The pcDNA 23 aa plasmid expresses detectable levels of our protein of interest (inset). U2OS cells were (lane 2) or were not (lane 1) transfected with the pcDNA 23 aa plasmid. At 48 h posttransfection, cell lysates were used for Western blot analysis with antibodies directed against FADD. (D) The 23-amino-acid region on FADD implicated in E6 binding does not interfere with the ability of E6 to bind and degrade p53. U2OS cells expressing (U2OSE612) or not expressing E6 (U2OSE6 AS) were transfected with pcDNA 23 aa or with pcDNA vector alone. At 24 h posttransfection, cells were treated with 2 μg of mitomycin C/ml for 16 h prior to measuring p53 levels with ELISA. Measurements were made in triplicate, and the error bars represent the standard deviation.