The High-Risk Human Papillomavirus Type 16 E6 Counters the GAP Function of E6TP1 toward Small Rap G Proteins

Abstract

We have recently identified E6TP1 (E6-targeted protein 1) as a novel high-risk human papillomavirus type 16 (HPV16) E6-binding protein. Importantly, mutational analysis of E6 revealed a strong correlation between the transforming activity and its abilities to bind and target E6TP1 for ubiquitin-mediated degradation. As a region within E6TP1 has high homology with GAP domains of known and putative Rap GTPase-activating proteins (GAPs), these results raised the possibility that HPV E6 may alter the Rap small-G-protein signaling pathway. Using two different approaches, we now demonstrate that human E6TP1 exhibits GAP activity for Rap1 and Rap2, confirming recent findings that a closely related rat homologue exhibits Rap-specific GAP activity. Using mutational analysis, we localize the GAP activity to residues 240 to 945 of E6TP1. Significantly, we demonstrate that coexpression of HPV16 E6, by promoting the degradation of E6TP1, enhances the GTP loading of Rap. These results support a role of Rap small-G-protein pathway in E6-mediated oncogenesis.

The high-risk human papillomaviruses (HPVs) are etiologically linked to human cervical cancer (44). Two early genes of the high-risk HPV genome, E6 and E7, are essential and sufficient for oncogenic transformation of human cells in vitro (17, 25). Expression of E6 and E7 together is necessary for efficient immortalization of human cervical keratinocytes, imposing limitations on the elucidation of biological pathways selectively targeted by these two oncogenes. However, as we demonstrated earlier, E6 alone can efficiently immortalize normal mammary epithelial cells (2). This single-gene immortalization model has provided a valuable system to dissect the transformation-related biochemical pathways specifically targeted by E6. For example, the high-risk HPV E6 proteins interact with and facilitate the degradation of p53, a transcriptional activator that plays a crucial role in cellular responses to DNA damage (23, 40). Thus, by eliminating p53 function, E6 facilitates the emergence of genomic alterations that contribute to cellular transformation (23, 40). Recent studies have shown that E6 also interacts with a number of other cellular proteins, and substantial evidence suggests that some of these interactions contribute to E6-induced cellular transformation (reference 21 and references therein).

We have recently identified a novel high-risk HPV type 16 (HPV16) E6-binding protein termed E6TP1 (E6-targeted protein 1) (14). Furthermore, we have shown that high-risk HPV E6 oncoproteins target E6TP1 for degradation via the E6AP-mediated ubiquitin-proteasome pathway (12-14). Our studies revealed a strict correlation between the abilities of E6 mutants to bind to and induce the degradation of E6TP1 and their ability to immortalize mammary epithelial cells (13), consistent with a potentially important role for the loss of E6TP1 function in E6-mediated cellular transformation. Sequence analysis showed a striking homology of the E6TP1 residues 489 to 819 to GAP domains of known and putative Rap GTPase-activating proteins (GAPs) (14). The proteins with the highest degrees of homology to E6TP1 included the mammalian Rap1GAP (4, 36), SPA1 (16, 22, 39), tuberin (the tuberous sclerosis complex 2 product, TSC2) (9, 18, 41), as well as the Drosophila RapGap1 (6) and one putative RapGAP open reading frame (T27F2.2) identified in the Caenorhabditis elegans genome. Recently, a rat protein, SPAR (SPA1-related), was identified and shown to have GAP activity using in vitro GAP assays (30). SPAR has a 95% amino acid identity with human E6TP1 over its 1,783-residue length, indicating that it is the rat homologue of E6TP1, and that human E6TP1 may also function as a RapGAP.

Rap1 proteins (Rap1A, Rap1B, Rap2A, and Rap2B) constitute a distinct subfamily of small GTPases in the RAS family (19, 20, 27, 29, 31, 32). Within this subfamily, Rap1A and Rap1B show >95% sequence identity and appear to be functionally indistinguishable (27). Rap2 shares 60% amino acid identity with Rap1, whereas Rap2A and Rap2B are 90% identical (27). Rap1A was originally identified as an antagonist of Ki-Ras-induced transformation and designated K-ras revertant protein 1 or Krev-1, and its effector domain is essentially identical to that of Ras (20). As Rap1 interacted with certain Ras targets, such as c-Raf, but did not modulate their activity, it was postulated that Rap proteins antagonize Ras function by sequestering Ras effectors (7, 20). However, there is only meager evidence that Rap proteins antagonize the function of normal cellular Ras (reviewed in reference 3). Indeed, recent studies have revealed Ras-like and Ras-independent functions of Rap proteins (42). Active Rap1 was shown to specifically interact with B-raf and mediate the late, sustained phase of mitogen-activated protein kinase activation critical for nerve growth factor-induced neuronal differentiation of PC12 cells. Rap proteins have also been shown to mediate the cyclic AMP-induced PC12 differentiation (43), CD31-induced integrin activation in lymphocytes (34), and lipopolysaccharide-induced activation of β2-integrin function in macrophages (5). Rap is activated through a calcium-dependent pathway by stimuli that regulate platelet aggregation (10), and a role for Rap proteins in regulating the oxidative burst in leukocytes has also been demonstrated (11). Overexpression of wild-type Rap1 in Swiss 3T3 cells resulted in a decrease in doubling time, increased saturation density, and morphological transformation; these Rap1-overexpressing cells formed tumors when injected into nude mice (1). Thus, Rap proteins appear to play important physiological roles and may be involved in oncogenic transformation.

Similar to other small G proteins, the basic GTPase cycle of Rap proteins between active GTP-bound and inactive GDP-bound states is regulated by guanine nucleotide exchange factors and GAPs (3). Recent studies have identified several Rap-specific guanine nucleotide exchange factors (GEFs) that are activated by tyrosine kinases (C3G), Ca²⁺ and diacylglycerol (CalDAG-GEF1), or cyclic AMP (Epac) (3). A number of Rap-specific GAPs have also been identified, including Rap1GAP (36), SPA1 (16, 22, 39), GAP1^IP4BP (8), and tuberin (9, 18, 41). The Rap1-specific Rap1GAP is selectively expressed in brain (36), whereas SPA1 is predominantly expressed in lymphohematopoietic cells (16, 22) and exhibits GAP activity for Rap1 and Rap2, but not for other small GTPases (22). Tuberin is the product of familial tuberous sclerosis gene, and deletion or mutations of this gene are known to cause benign tumors (9, 18, 37, 41). The GAP activity of GAP^IP4BP, which does not show clear homology with other RapGAPs, is less clear; while one study showed its activity toward both Rap and Ras proteins (8), another study did not find any activity of the bovine homologue (reviewed in reference 3). Notably, genetic studies have demonstrated a crucial role of RapGAP in Drosophila eye development (6).

Given the significant sequence homology of E6TP1 (human) with GAP domains of known Rap-specific GAPs and the recent demonstration of GAP activity for its rat homologue, we wanted to determine whether E6TP1 is indeed a Rap-specific GAP and whether its interaction with E6 can perturb the Rap G-protein activity. Using two different in vivo approaches, we show here that E6TP1 functions as a GAP for both Rap1 and Rap2 and localize the GAP domain of E6TP1 within residues 240 to 945. Importantly, we demonstrate that HPV16 E6 protein, by promoting the degradation of E6TP1, elevates the levels of GTP-bound Rap. Our analyses suggest that HPV E6 oncoproteins may activate Rap G-protein-mediated signaling pathways during cellular transformation.

GAP activity of human E6TP1 toward Rap small G proteins.

To assess the GAP activity of E6TP1, we utilized two independent in vivo approaches. In the first approach, we transfected mammalian expression constructs encoding glutathione S-transferase (GST) fusion proteins of Rap1A or Rap2A into 293T cells, labeled the cells with [³²P]orthophosphoric acid, and recovered the transfected Rap proteins together with bound ³²P-labeled guanine nucleotides from cell lysates using a pull-down assay with glutathione-Sepharose affinity beads. GTP and GDP bound to Rap were quantified by phosphorimager analysis after thin-layer chromatography separation (15, 26). When GST-Rap1A was expressed in 293T cells by itself, a small fraction (4.3% ± 0.6%) of Rap1 was found in the GTP-bound form (Fig. 1A, lane 1). Coexpression of SPA1, a known GAP for Rap, reduced the GTP-bound form of Rap1A by about 60% (to 1.7% ± 0.4%) (Fig. 1A, lane 3), whereas NF1, a Ras-specific GAP (used as a negative control) did not affect the proportion of GTP-bound Rap1A (3.9% ± 0.3%) (Fig. 1A, lane 4). Notably, when E6TP1 was cotransfected with GST-Rap1A, the fraction of GTP-bound Rap1A was substantially reduced (to 1.9% ± 0.4%, a 56% decrease) (Fig. 1A, lane 2).

FIG. 1.

E6TP1 exhibits GAP activity toward Rap small G proteins. 293T cells (10⁶) per 60-mm-diameter dish were cotransfected with 2 μg of plasmid DNA encoding GST-Rap1A (A) or GST-Rap2A (B) and 1 μg each of the indicated constructs using Fugene-6 reagent. For a control, vector DNA (−) was added (lanes 1). After 48 h, cells were washed in Tris-buffered saline and labeled with [³²P]orthophosphoric acid for 4 h in phosphate-free medium. Cells were then lysed, and GST-Rap1A (A) or GST-Rap2A (B) was recovered using glutathione-Sepharose beads. The bound guanine nucleotides were analyzed by polyethyleneimine thin-layer chromatography and detected using a phosphorimager. The positions of GTP and GDP are indicated. The percentage of Rap1 with bound GTP was calculated expressing GTP signals as a percentage of the total GTP plus GDP signals. GDP signals (two phosphates) were equalized with GTP signals (three phosphates) by multiplying with 1.5. For each panel, a representative experiment is shown in the top blot, and the means ± standard deviations for five independent experiments done in duplicate are shown in the bar graph.

Given the ability of E6TP1-related SPA1 to regulate both Rap1 and Rap2 (22), we also examined the ability of E6TP1 to function as a GAP toward Rap2A. Consistent with high basal GTP loading of Rap2 proteins (28), a substantially higher proportion of Rap2A was in the GTP-bound form (49.2% ± 2.0%) (Fig. 1B, lane 1). As expected, coexpression of SPA1 reduced the proportion of GTP-bound Rap2A (to 34.9% ± 1%) (Fig. 1B, lane 3), whereas NF1 had no effect (48.3% ± 4%) (Fig. 1B, lanes 1 versus 4). Importantly, coexpression of E6TP1 resulted in a marked decrease in GTP-bound Rap2A (to 14.4% ± 1%).

To further corroborate the above findings, we utilized a pull-down assay with the Rap-binding domain of RalGDS (as a GST fusion protein) (10). The GST-RalGDS specifically binds to the GTP-bound form of Rap1 (10); the bound Rap1 was detected by immunoblotting. When hemagglutinin (HA)-tagged Rap1A was expressed alone in 293T cells, anti-HA immunoblotting easily detected the GTP-bound Rap1A in a GST-RalGDS-RBD pull-down assay (Fig. 2A, top blot, lane 1). Cotransfection of SPA1 or RapGAP efficiently reduced the proportion of Rap1A pulled down with GST-RalGDS-RBD (Fig. 2A, top blot, lanes 7 and 8), whereas NF1 had no effect (lane 6). Notably, coexpression of increasing levels of E6TP1 led to a dose-dependent decrease in GTP-bound Rap1A pulled down with GST-RalGDS-RBD (Fig. 2A, compare lanes 2, 3, 4, and 5 with lane 1). Anti-HA immunoblotting of whole-cell lysates demonstrated comparable levels of total Rap1A in all samples (Fig. 2A). The expected levels of expression of E6TP1, Rap1GAP, NF1, and SPA1 proteins were confirmed by immunoblotting (Fig. 2A).

FIG. 2.

E6TP1 exhibits GAP activity toward Rap1 and Rap2 as assessed by GST pull-down assay. 293T (10⁶) cells per 100-mm-diameter dish were transfected with 5 μg of pMT2-HA-Rap1A alone (A) or 1 μg of pCMV-HA-Rap2A alone (B) or in combination with increasing amounts (0.5, 1, 2.5, and 5 μg) of pCMV-myc-E6TP1 or 5 μg of NF1, SPA1, or RapGAP, as indicated above the blots by the calcium phosphate precipitation method. The total amount of DNA was kept constant by adding vector DNA in each lane. Rap-GTP was precipitated from cell lysates with RalGDS-RBD and detected by 12CA5 immunoblotting (top blots), as described in the text. The expression of total Rap is shown in the second blots. The levels of expression of E6TP1, SPA1, NF1, and Rap1GAP proteins were detected by immunoblotting, as indicated.

The RalGDS-RBD pull-down assay was also used to confirm the activity of E6TP1 toward Rap2A. As shown in Fig. 2B, coexpression of E6TP1 decreased the levels of Rap2-GTP in a dose-dependent manner (compare lanes 2, 3, 4, and 5 with lane 1). As expected, coexpression of Rap1GAP decreased the levels of GTP-bound Rap2 (lane 7), whereas NF1 had no effect (lane 6). The expected levels of expression of E6TP1, Rap1GAP, NF1, and SPA1 proteins were confirmed by immunoblotting (data not shown). Taken together, the analyses of metabolically ³²P-labeled cells and RalGDS-RBD pull-down analyses demonstrate that E6TP1 functions to regulate the GTP loading of both Rap1 and Rap2 G proteins in mammalian cells. Thus, our results with human E6TP1 are consistent with findings on the rat homologue, which also exhibits GAP activity toward Rap1 and Rap2. Importantly, the in vivo measurement of GAP activity provided suitable assays to determine the effect of E6 on the GTP loading of Rap G proteins (see below).

Identification of the minimal region of E6TP1 with GAP activity.

Sequence comparison with known and putative Rap-specific GAPs indicated that E6TP1 residues 489 to 819 correspond to a putative GAP domain. To define the boundaries of the E6TP1 region capable of functioning as a GAP toward Rap, we constructed a series of E6TP1 deletion mutants in the pCMV vector that encode N-terminally myc-tagged proteins upon transfection into 293T cells (Fig. 3A). The GAP activity of these proteins was assessed by coexpressing the proteins with GST-Rap1A in 293T cells and analyzing the proportion of GTP-bound Rap1A using [³²P]orthophosphate labeling. The lack of activity of certain mutant proteins was not a consequence of their inability to be expressed (Fig. 3B, blot). These analyses (Fig. 3) defined the residues 240 to 945 of E6TP1 as the minimal region capable of functioning as a GAP toward Rap1A. This domain is significantly larger than that predicted by homology with other members of the family (residues 489 to 819). The additional N- and C-terminal regions that are not conserved in other Rap-specific GAP proteins are required for GAP activity. Whether the requirement for these extensions reflects their role in GAP activity or simply facilitates proper folding of the minimal GAP domain remains to be determined. Further refinement of deletion analyses and in vitro approaches should help to more precisely map the GAP domain. A notable result from our deletion analyses was that the PDZ domain, located C terminal to the GAP domain and conserved in SPA1, is not required for GAP activity. As the PDZ domain serves as protein-protein interaction motifs, usually targeting signaling proteins to submembranous complexes by binding to the C termini of transmembrane proteins, PDZ domain-mediated interactions could help target the GAP activity of E6TP1 to a specific subcellular location. This will be an area of substantial interest for future investigations.

FIG. 3.

GAP domain of E6TP1. (A) Schematic diagrams of the deletion mutants of E6TP1. Mutants exhibited (+) or did not exhibit (−) GAP activity toward Rap1. (B) Rap1-GAP activity of various E6TP1 mutants was assessed by transfecting 293T cells with 2 μg of plasmid DNA encoding GST-Rap1A and 1 μg each of the indicated constructs. Analysis of bound nucleotides to Rap1 was performed as described in the legend to Fig. 1. Values are means ± standard deviations from three independent experiments. Expression of various E6TP1 mutants was analyzed by immunoblotting with anti-myc 9E10 antibody (blot).

HPV16 E6 counters the GAP activity of E6TP1 toward Rap small G protein.

As we identified E6TP1 as an HPV E6-binding protein and found that E6 targeted E6TP1 for degradation, we examined whether HPV16 E6 has an influence on GTP loading of Rap. For this purpose, we coexpressed HA-Rap1A and myc-E6TP1 with or without increasing amounts of E6 or an E6TP1 non-binding E6 mutant pCR3.1-Δ9-13 in 293T cells. GTP-bound Rap levels were assessed using the GST-RalGDS-RBD pull-down assay. The expected decrease in Rap1-GTP levels was observed when Rap1 was coexpressed with E6TP1 (Fig. 4A, compare lanes 1 and 2). Significantly, coexpression of increasing amounts of HPV16 E6 led to a dose-dependent increase in the levels of GTP-bound Rap (Fig. 4A, lanes 3 to 5 versus lane 2), whereas an E6TP1 non-binding E6 mutant had no effect on the levels of GTP-bound Rap (Fig. 4A, lanes 6 to 8 versus lane 2). Western blot analysis of cell lysates indicated the expression of similar levels of total Rap1A in all samples (middle blots), while anti-myc blotting revealed the expected levels of E6TP1. Similar results were obtained in C33A, an HPV-negative cancer cell line derived from cervical keratinocytes (Fig. 4B). These results demonstrate that HPV E6, by inducing E6TP1 degradation, can elevate the levels of Rap in its GTP-bound form.

FIG. 4.

HPV16 E6 counteracts E6TP1 activity toward Rap. (A) 293T cells (10⁶) per 100-mm-diameter dish were transfected with pMT2-HA-Rap1A (5 μg) alone or in combination with pCMV-myc-E6TP1 (E6TP1) (1 μg) with or without increasing amounts (1, 2.5, and 5 μg) of pCR3.1-HPV16 E6 (16E6) or its mutant, pCR3.1-Δ9-13 (Δ9-13) by the calcium phosphate precipitation method. The total amount of DNA was kept constant by adding vector DNA in each lane. Intracellular Rap1-GTP was detected by binding with GST RalGDS-RBD, as described in the legend to Fig. 2. (B) Similar GST-RalGDS-RBD pull-down assay was performed as described above for panel A in the C33A cell line using 6 μg of pMT2-HA-Rap1A, 3 μg of pCMV-myc-E6TP1, and increasing amounts of pCR3.1-E6 (3, 6, and 9 μg).

Given that HPVs are causally linked to human cancer (44) and the transforming ability of HPVs is mediated by E6 and E7 oncogenes, it is now well accepted that E6 and E7 oncoproteins target critical cellular biochemical pathways in order to subvert normal cell growth and differentiation to promote oncogenic transformation. Thus, identification of novel cellular targets of HPV E6 oncoprotein and elucidation of biochemical activities of these targets are likely to reveal important cell growth regulatory pathways that are rendered aberrant in cancer cells. Demonstration of the ability of E6 to counter the GAP function of E6TP1 and consequent elevation of the GTP-bound form of Rap suggests that aberrant Rap signaling could constitute a part of the oncogenic transformation process mediated by E6. Consistent with such a possibility, genetic analyses of Drosophila RapGAP and biochemical studies of mammalian Rap-specific GAPs have begun to reveal their crucial physiological roles (6, 9, 16, 18, 22, 33, 35, 39, 41). Notably, tuberin, which has been shown to possess Rap1 GAP activity, is a tumor suppressor protein whose inactivation is associated with familial benign tumors (9, 18, 37, 41). Given the recently identified role of Rap protein in activating the mitogen-activated protein kinase cascade independently of Ras (43), these findings support a potential role of Rap-specific GAPs as tumor suppressors. While considerable future efforts will be required to determine if E6TP1 is a tumor suppressor, this possibility is strongly suggested by our previous and present results, as well as by the localization of the E6TP1 gene to chromosome 14q23.2-14q24.3, a site for deletions in malignant meningiomas (24, 38).

In keeping with a potential role of E6-induced perturbation of E6TP1 in HPV E6-mediated oncogenesis, E6TP1 binding and degradation strongly correlate with immortalizing abilities of HPV16 E6 mutants. Future analyses of endogenous E6TP1 expression and its potential perturbation upon HPV infection in relevant HPV target tissues should help establish if E6TP1 is indeed involved in HPV oncogenesis. Given the similar strategies employed by different oncogenic viruses, it will also be of interest to assess if E6TP1 binds to and is inactivated by other viral oncogenes.