Human papillomavirus E5 oncoprotein: function and potential target for antiviral therapeutics

Human papillomavirus

Human papillomaviruses (HPV) are small dsDNA viruses that infect both mucosal and cutaneous epithelia and cause a range of lesions from benign warts and verrucas to cancer of the anogenital and oropharyngeal tracts. A small subgroup of HPV are considered high-risk due to their association with cancer. The 8-kb HPV genome codes for six early (E1, E2, E4, E5, E6 and E7) and two late (L1 and L2) proteins. The major oncoproteins encoded by HPV are E6 and E7, which have been extensively studied, however, there is also a minor oncoprotein termed E5, whose functions remain to be fully elucidated. This review article will summarize our current understanding of the roles of E5 in virus replication, describe the recognized interactions with host signaling pathways and discuss the validity of E5 as a potential target for direct-acting antiviral therapeutics.

General characteristics of HPV E5

While E6 and E7 are thought to provide the primary transforming activities of high-risk HPV, E5 expression can augment their function and has been shown to contribute to tumor progression in vivo [1]. HPV E5 proteins are small, membrane bound and highly hydrophobic proteins, between 40 and 93 amino acids long, expressed by a subset of papillomaviruses [2]. The subcellular localization of HPV E5 is not clear, given that there is currently no E5-specific antibody reagent available. However, overexpression studies using epitope-tagged E5 have shown colocalization with markers associated with the endoplasmic reticulum, Golgi apparatus, perinuclear regions and plasma membrane [3]. E5 proteins encoded by high-risk HPV16 and HPV18 are 9.4 and 8.3 kDa, respectively, and contain three putative hydrophobic regions with α helical structure, which are proposed to function as transmembrane domains [4]. HPV16 E5 has been shown to oligomerize both in vitro and in cells, with oligomer formation driven not by the presence of disulphide linkages between cysteine residues, rather by hydrophobic interactions between individual E5 monomers [5,6]. E5 oligomerization results in the formation of a channel structure shown to permit the flow of fluorescent dyes from the lumen of artificial liposome membranes in vitro [5]. The channel forming abilities of E5 imply that it belongs to a growing family of channel-forming viral membrane proteins, termed viroporins, which have been shown to be important for a number of viruses including Influenza A virus (M2), HIV-1 (Vpu) and Hepatitis C virus (p7) [7]. These channels play a critical role in virus entry and egress and are thought to mediate these processes by modulating cellular ion homeostasis.

Despite its initial identification as a bona fide protein product in 1988 there is a remarkable lack of information concerning the nature of this often ignored protein. This is in part due to the extreme hydrophobic nature of E5, which precludes high levels of expression of recombinant E5 expression in bacteria or mammalian cells [5]. To further confound analysis, we lack effective antibody detection reagents against E5 preventing a comprehensive understanding of the E5 expression profile during the virus life cycle. Attempts to understand the temporal nature of E5 expression using mRNA expression profiles are limited given the mechanism of protein translation utilized by HPV. Despite being labeled as an early protein, E5 is located at the 3′ end of the early region of the virus genome and is the fourth ORF to be translated from polycistronic early transcripts using leaky ribosome scanning. Accordingly, little E5 protein is likely to be produced from such early transcripts [8]. Upon differentiation, E5 is the second ORF present on transcripts from the late promoter, suggesting an increased likelihood of greater expression levels and pointing toward a greater role in the differentiation-dependent stages of the HPV life cycle.

While a complete understanding of the role played by E5 in the HPV life cycle is not available, sequence analysis reveals an absence of a recognizable E5 ORF in several HPV types (β, γ and μ), indicating that functions fulfilled by E5 are fulfilled by alternative virus proteins in these specific types. In mucosal types where E5 is found it can be subdivided into four different groups based upon the E5 sequence, termed α to δ [9]. High-risk HPV types associated with malignant carcinomas (e.g., HPV16 and HPV18) all contain a conserved E5 protein of around 80 amino acids termed E5 α. Whereas the HPV types associated with benign lesions (e.g., HPV6 and HPV11) contain two putative E5 proteins, termed gamma and delta, which have little sequence conservation [9].

The role of E5 in the HPV life cycle

HPV is an epitheliotropic virus requiring the differentiation-dependent environment of keratinocytes to complete its life cycle. To begin to understand the role of E5 in the HPV life cycle researchers have taken advantage of available systems to study keratinocyte differentiation in vitro. These include monolayer differentiation protocols that rely on increasing the calcium concentrations in the media or suspending the cells in a viscous methylcellulose media to prevent cell to cell contacts and induce differentiation [10,11]. These protocols have revealed important insights into HPV biology, but cannot fully recapitulate the life cycle. Instead, organotypic raft cultures must be used. These skin equivalents require keratinocytes to be grown on a collagen matrix in the presence of fibroblast feeder cells, and raised to the air–liquid interface to induce keratinocyte differentiation [12,13]. Mutant genomes containing stop codons within the E5 coding sequence, which ablate E5 protein expression, have been used to interrogate the roles of the HPV16 and HPV31 E5 proteins [14,15]. These studies showed that E5 plays no apparent role in the early maintenance of HPV genome copy number or in proliferation of undifferentiated keratinocytes. Instead, E5 plays a clear role in the differentiation-dependent stages of the HPV life cycle. HPV31 E5 was shown to regulate differentiation-dependent genome amplification and viral gene transcription from the late promoter. Expression of E1⁁E4 was also reduced in the absence of E5 [15]. Importantly, E5 was also necessary for maintaining suprabasal cells in mitosis and active cell division following the normal differentiation-dependent exit from cell cycle. This is an essential component of the HPV life cycle to ensure continued HPV replication and virus production in the stratified epithelium [15]. These data point toward a critical role for E5 in essential stages of the HPV31 life cycle. However, an E5 knockout failed to reproduce all of these findings in the context of a complete HPV16 genome [14]. In this instance, a more subtle phenotype was observed in which E5 contributed toward host unscheduled DNA synthesis in the suprabasal layers of the epidermis but was not necessary for virus genome amplification or late gene expression [14]. Given that E7 expression is essential for maintaining cell cycle progression and HPV genome amplification in the differentiating environment, and that genomes lacking E7 display a similar but more pronounced defect in DNA synthesis, it has been postulated that E5 may co-operate with E7 to reprogram differentiated cells to allow vegetative DNA replication [16,17]. While this is a tempting explanation, our own unpublished observations using HPV18 genomes lacking E5 demonstrate similar defects in unscheduled DNA synthesis in suprabasal layers without adverse impacts on E7 protein expression, suggesting that E5 may itself be driving changes to keratinocyte proliferation pathways [Wasson CW et al., Unpublished Data]. Combined, these results indicate that E5 is important for maintaining an active cell cycle in differentiated keratinocytes, and furthermore that E5 is expressed upon keratinocyte differentiation because no defects were observed in monolayer culture in these studies. A question raised by these studies is, why are different phenotypes observed? One explanation may be due to differences in the cell types used. Both the studies of Fehrmann and Wasson (HPV31 and 18) utilized primary neonatal foreskin keratinocytes, whereas Genther's experiments (HPV16) where performed in immortalized NIKS (near-diploid immortalized keratinocytes) cells. It is noteworthy that none of these models are based on primary cervical keratinocytes, which, however, are main targets of HPV infection in vivo. Might subtle differences in the cellular environment account for the results obtained? A more provocative explanation can be found in the knockout strategy. Fehrmann et al. engineered the stop codon at the beginning of the HPV31 E5 sequence and so created a complete E5 knockout, whereas a partial HPV16 E5 knockout mutant was created in which the stop codon was placed after amino acid 28 within the E5 sequence. It is therefore possible that this truncated E5 species was able to perform a vestige of its functions within the life cycle. Alternatively, it may be that E5 proteins fulfill different roles in divergent HPV types. The 30% sequence variation between HPV16 and HPV31 may provide for distinct functional properties such as facilitating binding of unique cellular partner proteins. Further analysis using specific E5 point mutations or knockout genomes from a wider number of HPV types will provide greater insight and may explain these differences.

Major host pathways regulated by E5

Beyond the potential viroporin activity, no known intrinsic enzyme activity has been assigned to E5 proteins. As such, it is assumed that E5 exerts its effects by a number of interactions with cellular binding partners (Table 1). In fact, the expression of HPV16 E5 in cell culture affects the expression of between 25 and 179 host genes depending on the cell type and experimental conditions [18,19]. This indicates a significant but poorly investigated impact of E5 on host cells. Roles for E5 have been identified in cellular transformation, cell surface protein expression, mitogenic signaling and apoptosis (Figure 1). The myriad of cellular processes impacted suggests that E5 may interact with a significant number of host proteins.

The role of E5 in cellular transformation

In comparison to high-risk HPV types, whose major transforming proteins are E6 and E7, the major transforming protein of bovine papillomavirus (BPV) appears to be E5, which mediates its effects via an association with the PDGF receptor (PDGFR), resulting in constitutive activation [2]. Interestingly, HPV E5 proteins show little homology with BPV E5 and do not associate with the PDGFR. Despite this, studies were prompted to determine the oncogenic potential of HPV E5.

HPV E5 proteins are weakly oncogenic and are most likely not directly essential for carcinogenesis since they are not expressed in all HPV-positive tumors [25]. The first evidence of the weak transforming property of an HPV E5 protein came from HPV6 E5, which induced anchorage-independent growth in murine NIH3T3 cells and formation of small colonies in C127 cells [26]. Shortly afterwards, HPV16 E5 was shown to also induce anchorage-independent growth in a number of cell lines [27–29] and to induce mitogenic effects in primary human foreskin epithelial cells [30]. Further insights into the transforming abilities of E5 were provided when it was demonstrated to enhance the oncogenic abilities of the major transforming protein E7 in primary baby rat kidney cells [16,31]. This effect was validated in a series of elegant experiments performed in transgenic mouse models in which individual HPV oncoproteins, or combinations thereof, were expressed in mice under the control of the epithelial specific keratin-14 (K14) promoter [32]. Tumor formation was greater in mice expressing E5/E6 or E5/E7 compared with E6 or E7 alone. Interestingly, tumor formation was only observed in mice expressing E5 alone upon treatment with the hormone estrogen, suggesting that E5 may contribute to the promotion and progression of carcinogenesis rather than initiation.

E5 proteins target growth factor receptors

Encouraged by observations of the oncogenic potential of E5, the search began to uncover the molecular basis for transformation. A number of in vitro studies have since demonstrated the importance of the EGF receptor (EGFR) for E5 transformation [27,29–30,33]. These have been reinforced by genetic studies demonstrating that skin tumor formation in transgenic mice expressing E5 from an epidermis-specific keratin-14 promoter is dependent on the presence of a functional EGFR [34]. Thus the contribution of E5 toward host cell transformation appears dependent on manipulation of growth factor receptors and their downstream signaling pathways. Indeed, 70–90% of cervical cancers display increased EGFR expression, suggesting that manipulation of the EGFR pathway may be a crucial requirement for transformation [35–37].

The precise mechanism by which E5 manipulates the EGFR is unclear. The initial, widely accepted, hypothesis is based on an identified interaction between E5 and the 16 K subunit of the vacuolar H⁺-ATPase (v-ATPase) [3]. Different groups have reported that distinct regions of E5 mediate this interaction, which is thought to abrogate the normal v-ATPase-dependent endosomal acidification process, resulting in reduced EGFR degradation after growth factor stimulation [30,38–40]. Concomitantly, subversion of normal endocytic degradation pathways would induce a switch to recycling pathways, resulting in increased surface EGFR expression. Of note, the 16 K interaction is also shared with the BPV1 E5 protein, suggesting that targeting of the v-ATPase may be essential for conserved E5 functions across the papillomaviruses.

While this appealing mechanism is now dogma for E5 function within the community, work from several laboratories casts doubt on the importance of 16 K for manipulation of EGFR signaling by E5. Development of a novel 16 K antibody allowed determination of the amounts of endogenous 16 K bound by E5, which appeared too low to account for the substantial changes to endosome acidification observed [41]. Instead, E5 appears to retard fusion of early endosomes to late endosomes, interfering with receptor trafficking. The host determinants of this process are currently unclear, although it may require reorganization of the actin cytoskeleton [42]. In addition to altering membrane fusion, a plethora of additional mechanisms for manipulation of EGFR signaling have been proposed. These include E5 promotion of EGFR phosphorylation, which leads to increased expression of the prostaglandin E2 receptor EP4 [43]; direct binding of E5 to the EGFR, potentially directly modulating receptor activation [24] and disruption to the normal negative regulation processes that serve to prevent aberrant EGFR activation. E5 prevents the normal interaction between activated EGFR and the casitas B-lineage lymphoma (c-cbl) ubiquitin ligase [44]. This reduces EGFR ubiquitylation and subsequent receptor degradation, and may aid in the switch to increased receptor recycling to increase mitogenic signaling. HPV16 E5 has been shown to modulate further EGFR members, although the precise mechanisms and physiological outcomes are still poorly defined [45,46]. EGFR2 and EGFR3, but not EGFR4 (ErbB2, ErbB3 and ErbB4), display increased EGF-dependent phosphorylation in cells expressing E5. In contrast, in cervical intraepithelial neoplasia (CIN) samples expression of E5 correlated with expression of ErbB4 and EGFR [47]. Later it was revealed that the ErbB4 (JM-b/CYT-1) isoform forms a complex with HPV16 E5, leading to ligand independent activation of downstream signaling pathways [23].

While EGFR activation is necessary for keratinocyte proliferation, members of the FGFR family are required for the transition to differentiation. FGFR2IIIb (KGF receptor [KGFR]) is an FGFR2 splice variant expressed exclusively on epithelial cells [48]. In human skin KGFR is expressed at very low levels in the basal layers but more evenly distributed in the suprabasal layers [49]. KGFR levels are increased upon keratinocyte differentiation suggesting that the receptor plays a role in controlling the proliferative and differentiation program during transition from the basal to the suprabasal layers. Indeed, mice lacking epithelial KGFR display aberrant keratinocyte proliferation and can develop papillomas [50]. Importantly, this phenotype closely resembles that seen in transgenic mice expressing E5 [32] and it is now clear that KGFR is a target for manipulation by E5. Levels of both KGFR mRNA and protein are reduced in cells expressing HPV16 E5, by an unknown mechanism, resulting in reduced activation of downstream signaling pathways, including PI3-kinase–Akt/PKB, which are essential for regulating keratinocyte differentiation [51,52]. Interestingly, it was also shown that KGFR is able to induce a ligand-independent decrease in p63 expression [52]. P63 is expressed in the basal layers of the epithelium and is important in maintaining keratinocyte proliferation. Its expression is reduced as keratinocytes differentiate [53]. HPV 16 E5 targets p63 repression by blocking the host microRNA responsible for preventing p63 expression [54]. Inhibition of KGFR signaling may be an additional mechanism by which E5 is able to maintain p63 levels in order to delay differentiation and maintain proliferation.

Deregulation of mitogenic signaling & cell cycle control

In transient expression assays HPV16 E5 can activate MAPK signaling by EGF-dependent and -independent mechanisms [55]. In A31–3T3 fibroblasts expression of HPV16 E5 increased MAPK activation in response to EGF stimulation [56]; however, HPV16 E5 also promoted the translocation of PKC to cellular membranes, resulting in MAPK activation independent of EGF [45,57]. Activated MAPKs translocate to the nucleus of the cell and phosphorylate a range of transcription factors, eliciting profound alterations in the transcriptional profile of the cell. Expression of HPV11 E5a in genital wart samples appeared to correlate with increased expression of the MAPK target c-Jun [57]. In accordance with this, HPV16 E5 also activated a range of transcription factors including c-Fos and c-Myc in A31–3T3 fibroblasts and human keratinocytes [56,58]. These changes to mitogenic signaling can potentially impact cell cycle progression and increase transcription of the major HPV oncogenes E6 and E7 [25].

Cell cycle progression is mediated by decreased expression of tumor suppressor and checkpoint proteins. A number of E5 proteins have been shown to down-regulate transcription of p21^WAF1/Cip1 in NIH3T3 cells and human keratinocytes [59]. It is conceivable that this is achieved by the increased c-Jun expression, since this is known to repress p21^WAF1/Cip1 transcription [60]. This effect is enhanced by stimulation with EGF.

Major histocompatability class I trafficking

Establishment of a persistent infection is crucial for the papillomavirus life cycle and requires evasion of the host immune system. The cell-mediated arm of the adaptive immune response, which is essential for the removal of virus infected cells, functions by recognizing the presence of foreign protein epitopes found on the surface of an infected cell in complex with the host protein complex MHC class I [61]. In order to avoid cytotoxic killing of an infected cell by T cells, viruses have evolved a number of elegant methods to avoid presentation of their epitopes on MHC class I. Elimination of HPV requires T-cell activation, and as such HPV has evolved to prevent antigen recognition using a number of mechanisms and down-regulation of cell surface MHC class I can be observed in squamous cell carcinomas of the cervix [62,63]. The downregulation of MHC class I observed has been shown by a number of studies to be mediated by the E5 protein [64]. The function appears highly conserved across all papillomavirus E5 proteins tested (2a, 16, 83, BPV1, BPV4) [22,64–66]. The impact of E5 on cell surface protein expression appears to be highly specific since only certain MHC class I alleles are downregulated in cells expressing E5. In particular E5 targets HLA-A and HLA-B and removes them from the cell surface. These are the major T-cell activating MHC class I alleles, it makes sense to reduce their expression in infected cells to avoid cell killing. Alternative alleles such as HLA-C and HLA-E are left unaffected by E5. As these alleles function as inhibitory ligands to prevent aberrant NK cell activation, and it is clear that from an evasion perspective that it would benefit HPV persistence to maintain their cell surface expression.

E5 expression may retain MHC class I molecules in the Golgi apparatus to prevent them from trafficking to the cell surface. Studies reveal a direct interaction between the first transmembrane domain in E5 and MHC class I, which may function as a retention mechanism [66,67]. Others propose the formation of a ternary complex consisting of E5, MHC class I heavy chain and the chaperone calnexin [68], and indeed, E5 is unable to downregulate surface MHC class I expression in calnexin-deficient cells. Additionally, the observed interaction between E5, MHC class I chaperone B-cell receptor-associated protein 31 (BAP-31) and its binding partner A4 could play a role in this process but that has yet to be confirmed [11,20]. From a functional perspective, reduced MHC class I observed in E5 expression cells was shown to be able to prevent efficient killing of target cells by cytotoxic T cells [69]. Notably, HPV E5 does not appear to reduce expression levels of MHC class I heavy chain. In contrast, BPV4 E5 represses expression of MHC class I heavy chain in addition to binding to heavy chain and preventing its transport through the Golgi [70,71]. In this case the total MHC class I expression levels could be rescued by treatment with interferon but cell surface expression could not [72]. Additional studies have shown that E5 can also target other immunologically relevant cell surface proteins including MHC class II, which would subvert signaling to T helper cells [73], and CD1d [74]. Manipulation of these receptor proteins is likely to contribute toward the immune evasion properties of HPV.

E5 proteins modulate host apoptosis

The HPV life cycle exerts stress on the infected cell, by repressing cell cycle check points and the natural host response is to induce programmed cell death or apoptosis. HPV16 E5 is able to prevent apoptosis from a number of stimuli including immunological death ligands such as Fas ligand and TRAIL [75]. This is achieved by downregulation of the Fas receptor and impaired formation of DISC, respectively. Importantly, these findings have been confirmed in the more physiologically relevant raft culture model [76].

HPV16 E5 can also reduce apoptosis after exposure to UV-B irradiation using a pathway that requires ERK1/2 MAPK and PI3-kinase [77] and reactive oxygen species induced apoptosis was also ablated by targeting of the pro-apoptotic regulator BAX for proteasomal degradation upon hydrogen peroxide treatment [78]. While most effects appear to be mediated by targeting of pro-apoptotic proteins, co-localization between HPV16 E5 and the anti-apoptotic regulator Bcl-2 has been observed, however, neither the interaction nor functional consequences were investigated [79].

E5 may also modulate the ER stress response, either inhibiting the pathway by decreasing expression of COX-2, the transcription factor XBP-1 and the serine-threonine protein kinase IRE1a [19], or activating the pathway through increased expression of COX-2 [19,80]. In addition to these conflicting findings, others have shown no observable alterations to the ER stress pathway response in cells expressing E5 [81]. Due to the lack of a reliable E5 specific antibody (see below), various methods of overexpression and epitope tagging have been employed in these studies. It is conceivable that the exogenous E5 proteins localize in cellular compartments deviating from their natural localization which can induce varying ER stress responses. The advent of E5 specific detection reagents will allow determining how or if E5 modulates the ER stress response.

E5 as a manipulator of intracellular trafficking

Observations on the mechanism by which E5 manipulates both growth factor receptor and immunological receptor pathways indicate that it may function by modifying intracellular trafficking pathways (Figure 2). Importantly, the putative sub-cellular localization of E5 at either the ER or Golgi predestines it for roles in trafficking. In addition, immunoprecipitation studies identified an interaction between HPV16 E5 and KNβ3, a protein that plays a crucial role in exocytic trafficking [21]. Despite no current observed functional relevance for this interaction, it adds to the body of evidence to support a role for E5 in trafficking. Bravo et al. discovered that expression of HPV16 E5 alters the lipid composition of cellular membranes and propose that this effect of E5 precedes the modification of cellular trafficking [82]. Deregulated trafficking may also explain the increased cell surface expression of caveolin-1 and ganglioside GM1 that has been observed [83]. Elevated ganglioside expression promotes mitogenic signaling by the EGFR at low concentrations [84]. Thus it is possible that altered trafficking may augment ligand-independent EGFR signaling in multiple ways. Despite the potential for this role of E5 there remains a frustrating lack of concrete evidence and identification of potential host proteins that may mediate these effects. Future work must focus on identifying these proteins and understanding the molecular basis by which E5 manipulates trafficking pathways.

Is E5 a potential target for antiviral therapeutics?

New HPV infections can be prevented with the application of recently introduced prophylactic vaccines but these vaccines do not have a therapeutic effect in the millions infected with HPV. The treatment of these patients is limited to interventionist surgery and thus an effective and specific antiviral agent is extremely desirable. Targeting HPV proteins directly is a potential avenue of research, but is E5 a valid target? HPV-associated cancer is often associated with integration of the HPV genome into host chromosomes. In many cases this results in loss of the E5 ORF [86]. As such E5 is unlikely to provide a suitable therapeutic target for later stage disease. However, given that mRNA analysis indicates high E5 expression in precancerous lesions [47] coupled with the critical role of E5 in productive infection and modulation of proliferation pathways, E5 or constituents of the affected pathways may provide potential targets to prevent virus persistence and lesions from progressing into invasive cancers. The recent observation that E5 encodes ion channel activity highlights a potential therapeutic opportunity [5]. Virus-encoded ion channels (viroporins) of other viruses have been shown to be viable drug targets. In particular, the M2 channel of influenza A virus sets the precedent as a target protein, being susceptible to inhibition by the antiviral drugs amantadine and rimantadine [87]. HPV16 E5 is also susceptible to inhibition by these compounds as well as a number of other recognized viroporin inhibitor molecules [5]. Interestingly, inhibitors of viroporin activity prevented EGF-dependent EGFR hyperactivation in E5-expressing cells, suggesting that channel function may be linked to E5-mediated transformation [5]. This suggests that targeting the viroporin function of E5 may overcome the oncogenic nature of E5. Tumor vaccine strategies have also focused on E5 expressing cells. These are capable to prevent the formation of tumors but are not suitable to reduce existing tumors since a proportion of them do not express the E5 oncoprotein. Thus, vaccination with adenovirus expressing E5 sequences showed some protection against tumor development when challenged with E5 containing tumor cells in preliminary mouse studies [88,89]. An alternative strategy would be to focus on essential host pathways targeted by E5. This could provide protection against a range of HPV types and would presumably bypass concerns of virus-evolved resistance. An obvious target would be the EGFR-MAPK pathway, shown to be activated by E5 and upregulated in HPV-associated cancers, since small molecule inhibitors targeting this pathway are already in clinical development and trials.

Conclusion

Despite their small size, the E5 proteins of HPVs display a remarkable range of functions during the virus life cycle. They are necessary for maintaining a cellular environment conducive for vegetative replication by modifying epithelial proliferation and differentiation. To achieve this E5 must enhance growth promoting pathways, which presumably includes EGFR signaling, while interceding in pathways necessary for initiating keratinocyte differentiation. Deregulation of these host processes may in a minority of cases also contribute toward cellular transformation and malignancy. How does E5 modulate such a range of host processes? Since E5 lacks direct enzymatic activity, the most logical explanation is that it must engage with host proteins to perturb signal transduction. Despite a slow start, a growing number of diverse binding partners have now been identified (Table 1). However, the small size and hydrophobic, membrane imbedded, nature of E5 would argue against E5 being able to bind directly to such a large number of partners. Alternatively, E5 may interact with specific constituents of large multiprotein complexes to perturb their biological function. In this model, E5 need only bind directly to a limited number of proteins to alter signal transduction. Importantly, as many of the pathways targeted by E5 regulate host transcription, minor alterations to the signaling networks could culminate in more profound changes to global host gene expression. Thus small sequence variations between E5 proteins from different HPV types may impact their ability to bind to these ‘hub-proteins’ and direct a plethora of signaling events. Indeed, differential interactions with host proteins may help to explain the phenotypes observed in different HPV types lacking E5. Validation of this hypothesis awaits a detailed analysis of the E5-interactomes from a wider range of high- and low-risk HPV types, coupled to studies on the effects of specific E5 mutants on the HPV life cycle.

Transgenic mouse models have revealed E5 to be a more potent oncogene than originally assumed. We must therefore renew our interest in the molecular basis of transformation. While the EGFR pathway is likely to be a primary target for E5, the molecular mechanisms employed to deregulate this pathway are still unclear. Provocative evidence points to a role for E5 in deregulating host trafficking; however, the point in these pathways at which E5 interpolates is not known. Work over the coming years should focus on the identification and characterization of host targets that regulate trafficking and growth factor signaling. This may provide us with the general mechanism by which E5 is able to influence the host cell, since defects in intracellular trafficking can also explain the observed effects of E5 on apoptosis and antigen presentation. The viroporin function of E5 should now also be considered when studying the effects of E5 on virus replication and pathogenesis. Perturbation of ion flow would be expected to have a profound influence on the biology of infected keratinocytes. The precise role of the ion channel function awaits the characterization of specific channel-ablating mutations in the context of the whole HPV genome.

Future perspective

Ultimately, the resource that would make the most difference to our understanding of E5 is a specific, reliable antibody that functions in immunoblot and immunofluorescence/histochemistry. Lack of such a reagent limits the study of E5 and promotes an over-reliance on transient expression systems using epitope tagged E5 proteins, potentially producing aberrant findings. Specific E5 detection reagents could be used to highlight the temporal expression profile of E5 during a productive infection and could be used as a screening tool to determine E5 expression patterns in HPV-associated cancers.

Summary

A resurgence of interest in this ‘forgotten oncoprotein’ has reinforced its critical role in the virus life cycle and transformation. New host partners have been identified and novel pathways manipulated by this virus protein. Over the coming years we need to increase our understanding of these processes and translate the findings into potential therapeutic interventions.