Transcriptional regulation of the major histocompatibility ... - Nature

0 downloads 0 Views 181KB Size Report
class I heavy chain, TAP1 and LMP2 genes by the human papillomavirus. (HPV) type 6b, 16 and 18 E7 oncoproteins. Nikolaos T Georgopoulos1, Joanne L ...
ã

Oncogene (2000) 19, 4930 ± 4935 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc

SHORT REPORT

Transcriptional regulation of the major histocompatibility complex (MHC) class I heavy chain, TAP1 and LMP2 genes by the human papillomavirus (HPV) type 6b, 16 and 18 E7 oncoproteins Nikolaos T Georgopoulos1, Joanne L Prott1,2 and G Eric Blair*,1 1

School of Biochemistry and Molecular Biology, University of Leeds, Leeds LS2 9JT, UK

We have examined the possibility that the E7 proteins of the high-risk human papillomavirus (HPV) type 16 and 18 and the oncogenic adenovirus (Ad) type 12 E1A protein share the ability to down-regulate the expression of components of the antigen processing and presentation pathway, as a common strategy in the evasion of immune surveillance during the induction of cell transformation. Expression of the HPV 18 E7 oncoprotein, like Ad 12 E1A, resulted in repression of the major histocompatibility complex (MHC) class I heavy chain promoter, as well as repression of a bidirectional promoter that regulates expression of the genes encoding the transporter associated with antigen processing subunit 1 (TAP1) and a proteasome subunit, low molecular weight protein 2 (LMP2). HPV 16 E7 also caused a reduction in class I heavy chain promoter activity, however it did not have any signi®cant e€ect on the activity of the bidirectional promoter. Interestingly, expression of the low-risk HPV 6b E7 protein resulted in an increase in MHC class I heavy chain promoter activity, while repressing the TAP1/LMP2 promoter. Interference with the class I pathway could also explain the ability of low-risk HPVs in inducing benign lesions. Oncogene (2000) 19, 4930 ± 4935. Keywords: human papillomavirus; E7; MHC class I; TAP1; LMP2; transcriptional repression The major histocompatibility complex (MHC) class I genes (H-2K, D and L in the mouse, HLA-A, B and C in humans) encode transmembrane glycoproteins which are involved in the process of immune recognition. Class I molecules are heterodimers consisting of a polymorphic heavy chain non-covalently associated with an invariant light chain (termed b2-microglobulin) and they function by binding intracellularly processed peptides, presenting them on the cell surface to cytotoxic T lymphocytes (Williams et al., 1996). During viral infection or malignant transformation, a spectrum of peptides derived from viral or transformationassociated antigens is displayed by class I molecules. These class I-peptide complexes are recognised by the T cell receptors on a speci®c repertoire of CD8+ cytotoxic T lymphocytes, resulting in the elimination

*Correspondence: GE Blair 2 Current address: Covance Laboratories Ltd, Harrogate, North Yorkshire, UK Received 16 February 2000; revised 23 May 2000; accepted 9 August 2000

of the infected or malignant cell. Class I molecules therefore play a crucial role in immune recognition of virally infected and transformed cells, as a reduction in cell surface class I molecules renders peptide antigen presentation inecient and results in evasion of cellular immune surveillance (Heemels and Ploegh, 1995). Consequently, a reduction in the level of surface class I molecules may be an important process in the multistep pathway to oncogenesis. Transport of class I molecules to the cell surface is dependent on their binding to antigenic peptides generated in the lumen of the endoplasmic reticulum (ER) (Townsend et al., 1990). These peptides, which arise from proteolytic cleavage in the cytosol, are delivered to the ER by the action of a protein complex referred to as the transporter associated with antigen processing or TAP (reviewed by Elliott, 1997). TAP consists of two protein subunits, TAP1 and TAP2, which are encoded by separate genes located in the MHC class II cluster of the MHC. The functional signi®cance of TAP1/2 in the antigen presentation pathway has been established by the study of TAPde®cient cell lines, where MHC class I expression is absent or reduced (Baas et al., 1992). Antigenic peptides are generated by the action of the proteasome, a multi-subunit protein complex that carries the bulk of cytosolic proteolytic activity. The ®rst line of evidence implicating the proteasome in antigen processing came from the discovery of two genes termed LMP2 and LMP7 (Low Molecular weight Protein). These are very closely linked to TAP1 and TAP2 genes in the MHC class II region and they have been shown to be the two major IFN-g inducible components of the proteasome. The LMP2 and LMP7 subunits replace their constitutively expressed counterparts giving rise to the `immunoproteasome' (reviewed by van Endert, 1999). Cells in which LMP2/7 are deleted are still capable of antigen presentation, however they possess reduced levels of class I molecules and present certain antigenic peptides ineciently (Cerundolo et al., 1995). There are more than 90 recognized serotypes of human papillomaviruses (HPVs), the main aetiological agents of human cervical cancer. At least 25 of these are speci®c for lesions of the anogenital tract and they can be divided into two main groups: the high-risk serotypes, such as HPV 16, 18, 31 and 33, that are associated with highly invasive anogenital and cervical carcinomas, and the low-risk serotypes, such as 6 and 11, which are the principal causative agents of benign genital warts. The HPV 16 and HPV 18 E7 early proteins have been directly implicated in the main-

Regulation of MHC gene expression by HPV E7 NT Georgopoulos et al

tenance of the malignant phenotype (Smotkin and Wettstein, 1986; Schneider-Gaedicke and Schwartz, 1986). It has also been demonstrated that they constitute the major viral oncoproteins in cervical tumours as well as cancer-derived cell lines harbouring HPV 16 or 18 DNA, and in some cases only the E7 coding sequence is retained (Wilczynski et al., 1988). Research carried out in the past decade as well as recent ®ndings have revealed a diverse array of physical interactions of E7 with various cellular regulatory proteins, which appear to render this multifunctional protein a potent immortalizing and transforming agent (reviewed by Zwerschke and Jansen-DuÈrr, 2000). There is clear evidence suggesting that surface MHC class I molecules and components of the antigen presentation pathway are down-regulated in cervical tumours (Stern, 1996). Adenoviruses, another major group of DNA tumour viruses, have been shown to interfere with the expression of the major components of the antigen presentation pathway. The highly oncogenic adenovirus (Ad) 12 E1A represses transcription driven from both the class I heavy chain promoter (Prott et al., 1994) and the TAP1/LMP2 bidirectional promoter (Prott and Blair, 1997). The E7 protein of the high-risk HPVs exhibits a signi®cant degree of structural and functional similarities to the Ad E1A protein. In addition, both Ad E1A and E7 contain another conserved region, which can bind the product of the retinoblastoma tumour suppressor gene product, p105-RB (Dyson et al., 1989). It has been suggested that oncogenic DNA viruses (such as HPVs and adenoviruses) transform cells by similar mechanisms. It is therefore reasonable to address the question of whether speci®c reduction of surface MHC class I expression or general interference with the expression of the major components of the antigen presentation pathway constitutes a part of the `strategy' of HPVs in the induction of tumours. Loss of MHC class I molecules is a frequent event in both premalignant and malignant cervical lesions (Connor and Stern, 1990; Cromme et al., 1993a,b) and loss of the peptide transporter subunit TAP1 has also been found in malignant and metastatic lesions (Cromme et al., 1994b; Keating et al., 1995). It appears that there is no clear correlation between the presence of HPV DNA, or HPV 16 E7 transcripts and the MHC phenotype in cervical lesions (Bartholomew et al., 1995; Cromme et al., 1993b). However, HPV 18 has been associated with more advanced tumours (Arends et al., 1993), increased risk for progression to malignancy and a poorer prognosis (Burger et al., 1996) in early stage cervical cancer (cervical intraepithelial neoplasia or CIN), whereas epidemiological and molecular biological data have suggested that HPV 18 may be associated with `rapid transit' tumours that progress too rapidly to be detected by cytology screening programmes (Cox, 1995). Therefore, signi®cant di€erences may exist between HPV types 16 and 18 on the basis of malignant properties and mechanism of cell transformation, which merits further study into the ability of the high-risk HPVs to modulate MHC class I gene expression. The low-risk HPV serotypes, such as 6 and 11, are generally associated with benign tumours and low-grade CIN. We therefore investigated the e€ect of the expression of high- and low-risk HPV

4931

Figure 1 E€ect of expression of the HPV 6b, 16 and 18 E7 genes on the activity of the H-2Kb and Ad E2A promoters. 3Y1 cells were co-transfected by the calcium phosphate DNA coprecipitation method as described previously (Herrmann et al., 1987). A total of 11 mg of DNA was used, comprising 1 mg pRSVluc reference plasmid along with 5 mg of reporter gene plasmid and 5 mg of test vector (or control plasmid). The reporter constructs pH2KCAT (a) and pE2ACAT (b) have been previously described (Prott et al., 1994; Dery et al., 1987, respectively) and contain the 2 kb 5' region of the mouse H-2Kb gene and the Ad 2 E2 early promoter, respectively, cloned upstream of the chloramphenicol acetyltransferase (CAT) reporter gene. These were co-transfected with the HPV E7 expression plasmids pMo6bE7, pMo16E7 and pMo18E7 that contain the HPV types 6b, 16 and 18 E7 ORFs, respectively (Barbosa et al., 1991), as well as pAsc 4.7, which is the AccI-H fragment of Ad 12 (comprising the E1A coding region plus the 5' end of the E1B gene) cloned between the AccI and EcoRI sites of pBR322 (Byrd et al., 1982). pIBI30, a pUC18-based cloning vector (International Biotechnologies Inc., New Haven, USA), was used as a control plasmid. At 40 h post-transfection, the cells were harvested and CAT activities were assayed using standard methods (Gorman, 1985). To correct for variation in transfection eciency, the volume of cell extract used in the CAT assays was adjusted by performing luciferase reporter gene assays according to the manufacturer's protocol (Promega, UK). The CAT activity shown represents fold activity compared with pIBI30 vector alone and the activity of this vector was normalized to a value of one. The experiments were repeated at least three times with more than one batch of each plasmid and variance between individual assays with each plasmid was not greater than 10% ± representative basal CAT activities (obtained after co-transfection with pIBI30) were 48.5 and 4.5% in pH2KCAT (a) and pE2ACAT (b) cotransfections, respectively. The results of the CAT assays were quanti®ed by molecular imaging analysis using a Fuji BAS 1000 phosphorimager and luciferase reporter assays were performed in a Lumac Biocounter1 M2010 luminometer. 3Y1 cells are established Fischer rat embryo ®broblasts (Kimura et al., 1975). Cells were cultured in Dulbecco-modi®ed Eagle's medium (DMEM) supplemented with 10% (v/v) foetal calf serum (FCS), in a humidi®ed atmosphere containing 5% CO2 at 378C Oncogene

Regulation of MHC gene expression by HPV E7 NT Georgopoulos et al

4932

Figure 2 E€ect of expression of the HPV 6b, 16 and 18 E7 genes on the activity of the bidirectional TAP1/LMP2 promoter. CAT reporter gene assays were performed in 3Y1 ®broblasts 40 h after co-transfection of the pTAP1CAT or pLMP2CAT constructs (a), containing the rat TAP1/LMP2 intergenic region subcloned in either orientation upstream of the CAT reporter gene in pBLCAT3 (Prott and Blair, 1997), along with the HPV 6b, 16, 18 E7 or Ad 12 E1A expression vectors (b). A total amount of 11 mg of DNA was used, comprising 1 mg pRSVluc plasmid along with 5 mg of reporter gene vector and 5 mg of test plasmid (or pIBI30 control plasmid) and luciferase assays were performed to normalize for CAT activity as described in Figure 1 ± basal CAT activity, obtained as an average from at least three di€erent experiments, was 45.0 and 25.6% in pTAP1CAT and pLMP2CAT co-transfections (b), respectively. Dual LuciferaseTM Reporter assays (Promega, UK) were carried out after the cotransfection of pTAP1LUC or pLMP2LUC constructs (c), generated by subcloning the TAP1/LMP2 region in both orientations upstream of the Fire¯y luciferase reporter gene in the pGL3-Basic reporter vector (Promega, UK), along with the HPV E7 and Ad E1A expression vectors (as described above) (d). Oncogene

E7 genes on promoters regulating antigen processing and presentation gene expression. To assess the e€ect of the HPV 16, 18 as well as HPV 6b E7 oncoproteins on the MHC class I heavy chain promoter, transient expression assays were performed. These involved the co-transfection of the pH2KCAT construct, containing the 2 kb 5' ¯anking region of the mouse H-2Kb gene fused to the CAT gene in the pBLCAT3 reporter vector (Prott et al., 1994), along with HPV 6b, 16 or 18 E7 (Barbosa et al., 1991) or Ad 12 E1A (Byrd et al., 1982) expression vectors in established rat 3Y1 ®broblasts (Figure 1a). Expression of either HPV 16 or 18 E7 or Ad 12 E1A resulted in a 2.5 ± 3-fold down-regulation of reporter gene expression, whereas HPV 6b E7 caused a 1.5 ± 2-fold increase in CAT activity. As a positive control and to ensure that the e€ects observed in all of the experiments presented were due to the presence of functional E7 protein, we exploited the ability of E7 to transactivate the Ad E2 promoter. For this purpose, co-transfections were performed with either the E7 or the E1A expression vectors and the pE2ACAT plasmid, in which the Ad 2 E2 early promoter is cloned upstream of the CAT reporter gene (Dery et al., 1987). All E7 constructs and E1A activated the E2 promoter (as shown in Figure 1b). Therefore, like Ad 12 E1A, HPV 16 and 18 E7 can mediate a reduction in MHC class I promoter activity as eciently as Ad 12 E1A, which suggests further functional similarities between highrisk HPV E7 and Ad 12 E1A. In spite of the potential association of HPV 16 with post-transcriptional control of class I heavy chain expression previously described (reviewed by Stern, 1996), these data demonstrate that HPV 16 E7 can also cause repression of the H-2Kb promoter. In addition, like nononcogenic Ad 2 and Ad 5 E1A (Blair and Hall, 1998), expression of low-risk HPV 6b E7 leads to an increase in class I promoter activity. Reductions in TAP gene expression associated with down-regulation of MHC class I molecules in cervical lesions have been demonstrated in several studies. Complete or heterogeneous loss of class I molecules associated with loss of TAP1 expression has been demonstrated (Cromme et al., 1994a; Keating et al., 1995), whereas a signi®cant increase in class I downregulation, accompanied by loss of TAP1 expression, was observed in metastatic cells in comparison to primary tumours (Cromme et al., 1994b). Transcriptional repression of other major components of the antigen presentation pathway by Ad 12 has been reported. Expression of TAP1 and TAP2 peptide transporter polypeptides as well as the inducible proteasome subunits LMP2 and LMP7 is repressed

In this case, pCMV-RL, which comprises the Renilla luciferase reporter gene (Promega), was used as the reference plasmid in normalizing Fire¯y luciferase activity. Total DNA was kept constant at 11 mg, CAT or luciferase activities shown are expressed as fold activity compared with pIBI30 vector alone and the activity of the control vector is normalized to a value of one. Dual LuciferaseTM Reporter assays were carried out 24 h post-transfection according to the manufacturer's protocol. Transfections were performed at least three times and with more than one batch of each plasmid by the calcium phosphate DNA co-precipitation method

Regulation of MHC gene expression by HPV E7 NT Georgopoulos et al

4933

Figure 3 Down-regulation of the H-2Kb promoter by HPV 18 E7: identi®cation of a promoter element targeted by HPV 18 E7. Transfection of DNA was performed using the LipofectAMINETM reagent (Life Technologies, Inc.) following the manufacturer's instructions. CAT assays were carried out 40 h after co-transfection of the HPV 18 E7 (pMo18E7) and Ad 12 E1A (pAsc 4.7) expression vectors, as well as the control plasmid (pIBI30), with the reporter constructs p(71.07/71.44)tkCAT (a), p(71.07/ 71.18)tkCAT (b), p(71.18/71.44)tkCAT (c), p(71.26/71.15)tkCAT (d), p(71.35/71.23) (e) and p(71.44/71.32)tkCAT (f), which have been previously described (Prott et al., 1994). They contain the 71.07/71.44, 71.07/71.18, 71.18/71.44, 71.26/ 71.15, 71.35/71.23 and 71.44/71.32 sequences, respectively, of the 5' ¯anking region of the H-2Kb gene cloned into pBLCAT2, a reporter vector containing the CAT gene under the control of a herpes simplex virus thymidine kinase (tk) basal promoter. Plasmids were transfected into 3Y1 cells and CAT activities were assayed and expressed as described in Figure 1. The results illustrate experiments performed at least three times, variance between individual transfections was less than 10% and representative basal CAT activities (in the presence of pIBI30) were respectively: 20.5% (a), 7.5% (b), 53.6% (c), 10.7% (d), 36.6% (e) and 18.3% (f)

by the highly oncogenic Ad 12 (Blair and Hall, 1998). We have previously shown that the TAP1 and LMP2 genes share a common bidirectional promoter that is repressed in Ad 12-transformed cells (Prott and Blair, 1997) and that Ad 12 E1A expression results in reduction of TAP1/LMP2 promoter activity in transient expression assays performed in 3Y1 ®broblasts (P Cotterrell and GE Blair, unpublished results). We

therefore decided to examine whether HPV E7 has a similar mode of action to Ad 12 E1A in this respect. Transient expression assays were carried out in 3Y1 cells that involved the co-transfection of pTAP1CAT or pLMP2CAT constructs, which contain the rat TAP1/LMP2 intergenic region subcloned in either orientation upstream of the CAT reporter gene (Prott and Blair, 1997), along with the HPV E7 and Ad 12 Oncogene

Regulation of MHC gene expression by HPV E7 NT Georgopoulos et al

4934

Oncogene

E1A expression vectors (Figure 2a). The results from these studies showed that HPV 18 E7 expression resulted in a 2 ± 2.5- and a 2.5 ± 3-fold down-regulation of the TAP1 and LMP2 promoters, respectively, similar to that obtained with Ad 12 E1A (Figure 2b). In contrast, HPV 16 E7 did not have any signi®cant e€ect on the expression of the bidirectional promoter in either orientation. Interestingly, expression of HPV 6b E7 resulted in a twofold reduction of promoter activity. The TAP1/LMP2 intergenic region was also subcloned in either orientation into luciferase reporter vectors to obtain the pTAP1LUC and pLMP2LUC constructs (Figure 2c). Similar transient expression assays were carried out as before and the results obtained were almost identical (Figure 2d). The data obtained from both series of studies indicate that HPV 18 E7, like Ad 12 E1A, mediates the transcriptional down-regulation of the TAP1 and LMP2 genes and can therefore interfere with the expression of genes encoding major components of the antigen presentation pathway as a part of their strategy for the reduction of surface MHC class I molecules. HPV 16 E7, however, does not repress the TAP1 promoter and it does not have a signi®cant e€ect on the bidirectional promoter in the LMP2 orientation, while, as described above, it causes repression of the H-2Kb promoter. Taken together, these ®ndings suggest a potential di€erence in the e€ects of HPV 18 and 16 E7 proteins, which could account for the biological di€erences between these two serotypes described above. These results also show that HPV 6b E7 causes the repression of the TAP1/LMP2 bidirectional promoter expression. This suggests that interference with these components of the class I pathway may form part of the strategy of the low-risk HPVs (such as 6b and 11) in inducing benign as well as malignant lesions (Wilczynski et al., 1993). We have previously demonstrated that the target site for Ad 12 E1A-mediated repression of the H-2Kb promoter lies within the 71.07 to 71.44 kb region (Prott et al., 1994). To establish whether the downregulation of class I heavy chain promoter by HPV 18 E7 could be ascribed to the Ad 12 E1A targeted region of the class I promoter, co-transfection assays were carried out using the p(71.07/71.44)tkCAT, p(71.07/ 71.18)tkCAT and p(71.18/71.44)tkCAT constructs that contained the PCR-generated 71.07/71.44, 71.07/71.18 and 71.18/71.44 fragments of the upstream class I regulatory region (Figure 3), linked to a basal thymidine kinase (tk) promoter driving expression of the CAT reporter gene (Prott et al., 1994), together with the HPV 18 E7 and Ad 12 E1A expression plasmids. As shown in Figure 3, expression of E7 resulted in an approximately 3.5-fold decrease in CAT activity of p(71.07/71.44)tkCAT, comparable with the fourfold repression mediated by E1A (Figure 3a), whereas CAT activity directed by the p(71.07/ 71.18)tkCAT was slightly elevated above basal levels (Figure 3b), suggesting that this sequence was not involved in repression of the H-2Kb promoter by HPV 18 E7, in agreement with previous ®ndings for Ad 12 E1A (Prott et al., 1994). In contrast, p(71.18/ 71.44)tkCAT directed a ®vefold reduction in reporter activity (Figure 3c), indicating the existence of a major transcriptional repression element located within the 71.18/71.44 region. To further dissect and ®nely map

this site, reporter constructs were used containing a series of three overlapping, PCR-generated fragments 110 ± 120 bp in length that spanned this 260 bp region (as shown in Figure 3) cloned in pBLCAT2 (Prott et al., 1994). Although they all displayed a decrease in CAT activity (Figure 3d ± f), none of these constructs exhibited a reduction as substantial as that obtained with p(71.18/71.44)tkCAT (Figure 3c). This is suggestive of multiple interactions of E7 with more than one element over the entire 260 bp region and that these elements may function synergistically to e€ect maximal repression of H-2Kb transcription. The data presented in this study are suggestive of a direct interference by HPV 18 E7 oncoprotein with the expression of MHC class I heavy chain, TAP1 and LMP2 proteins at the level of transcription, functional properties that have been demonstrated for Ad 12 E1A protein (Ackrill and Blair, 1988; Prott et al., 1994; Prott and Blair, 1997). HPV 18 E7 mediates transcriptional down-regulation of the H-2Kb class I heavy chain (Figure 1a) which seems to be targeted, at least in part, through a minimal element (Figure 3) previously identi®ed as the target of Ad 12 E1A mediated transcriptional repression of the class I promoter (Prott et al., 1994). HPV 18 E7 also mediates the down-regulation of the bidirectional TAP1/LMP2 promoter, as demonstrated by two di€erent reporter gene assays (Figure 2b,d), analogous to Ad 12 E1A (Prott and Blair, 1997). Analysis of transactivation and transformation functions of E7 in comparison with Ad 12 E1A (Phelps et al., 1991), supported by the existence of established properties shared by these two proteins such as cooperation with c-ras to transform primary rodent cells (Matlashewski et al., 1987) and primary BRK cells (Phelps et al., 1988; Storey et al., 1988) as well as direct binding to the RB tumour suppressor (Dyson et al., 1989), has previously suggested the existence of common mechanisms of action. As well as its ability to function as a transactivator, high-risk HPV E7 has also been shown to mediate transcriptional repression of various cellular genes, such as smooth muscle a-actin and human ®bronectin (Rey et al., 2000), while a transcriptional repressor domain has been identi®ed in the carboxy-terminus of E7, suggestive of possible interaction of E7 with transcriptional repressors (reviewed by Zwerschke and Jansen-DuÈrr, 2000). Our results directly support the hypothesis that transcriptional repression of components of the class I antigen presentation pathway may represent an important step in an escape from immune surveillance for the induction of cellular immortalization and transformation by HPVs and suggest the existence of common mechanisms involved in MHC class I repression shared by the major transforming gene products (E7 and E1A, respectively) of the highly oncogenic DNA tumour viruses HPV 18 and Ad 12. The existence of multiple regulatory elements in the H-2Kb promoter that can potentially be targeted by E1A, such as two 6/7 non-canonical matches with the consensus AP-1 sequence (Prott et al., 1994), as well as the presence of an NF-kB site in the TAP1/LMP2 promoter conserved in rat, mouse and humans (Prott and Blair, 1997), suggest that, like E1A, the regulatory e€ects of E7 could be pleiotropic and could involve the repression of a series of cellular transcription

Regulation of MHC gene expression by HPV E7 NT Georgopoulos et al

factors, such as NF-kB or AP-1. It has been suggested that Ad E1A and HPV E7 can repress AP-1-driven transcriptional activation by inhibiting RB-mediated transactivation of the AP-1 complex (Nishitani et al., 1999). Such changes would result in the downregulation of multiple components of the antigen presentation pathway, followed by a reduction in the expression of surface MHC class I molecules.

Acknowledgments We would like to thank JT Schiller (National Cancer Institute, USA) for kindly providing the HPV 6b, 16 and 18 E7 expression vectors. The pE2ACAT construct was a kind gift from MB Mathews (University of Medicine and Dentistry, New Jersey at Newark, USA). We would also like to thank Lindsay J Stanbridge and Luke Stockwin for helpful discussions. This work was supported by Yorkshire Cancer Research.

4935

References Ackrill AM and Blair GE. (1988). Oncogene, 3, 483 ± 487. Arends MJ, Donaldson YK, Duvall E, Wyllie AH and Bird CC. (1993). Human Pathol., 24, 432 ± 437. Baas EJ, van Santen H-M, Kleijmeer MJ, Geuze HJ, Peters PJ and Ploegh HL. (1992). J. Exp. Med., 176, 147 ± 156. Barbosa MS, Vass WC, Lowy DR and Schiller JT. (1991). J. Virol., 65, 292 ± 298. Bartholomew J, Tinsley JM and Stern PL. (1995). Modulation of MHC Antigen Expression and Disease: MHC Expression in HPV-associated cervical cancer. Blair GE, Pringle CR, Maudesley DJ (eds). Cambridge University Press: Cambridge, pp. 233 ± 250. Blair GE and Hall KT. (1998). Semin. Virol., 8, 387 ± 397. Burger RA, Monk BJ, Kurosaki T, Antonculver H, Vesilev SA, Berman ML and Wilczynski SP. (1996). J. Natl. Cancer Inst., 88, 1361 ± 1368. Byrd PJ, Chia W, Rigby PWJ and Gallimore PH. (1982). J. Gen. Virol., 60, 279 ± 293. Cerundolo V, Kelly A, Elliott T, Trowsdale J and Townsend A. (1995). Eur. J. Immunol., 25, 554 ± 562. Connor ME and Stern PL. (1990). Int. J. Cancer, 46, 1029 ± 1034. Cox JT. (1995). Bailliere's Clin. Obstet. Gynecol., 9, 1 ± 37. Cromme FV, Meijer CJLM, Snijders PJF, Uyterlinde A, Kenemans P, Helmerhorst T, Stern PL, van den Brule AJC and Walboomers JMM. (1993a). Br. J. Cancer, 67, 1372 ± 1380. Cromme FV, Snijders PJF, van den Brule AJC, Kenemans P, Meijer CJLM and Walboomers JMM. (1993b). Oncogene, 8, 2969 ± 2975. Cromme FV, Airey J, Heemels M-T, Ploegh HL, Keating PJ, Stern PL, Meijer CJLM and Walboomers JMM. (1994a). J. Exp. Med., 179, 335 ± 340. Cromme FV, van Bommel P, Walboomers JMM, Gallee MPW, Stern PL, Kenemans P, Helmerhorst T and Meijer CJLM. (1994b). Br. J. Cancer, 69, 1176 ± 1181. Dery CV, Herrmann CH and Mathews MB. (1987). Oncogene, 2, 15 ± 23. Dyson N, Howley PM, MuÈnger K and Harlow E. (1989). Science, 243, 934 ± 937. Elliott T. (1997). Adv. Immunol., 65, 47 ± 109. Gorman C. (1985). DNA Cloning, Vol. II: A Practical Approach I. Glover DM (ed). IRL Press: Oxford, pp. 143 ± 190.

Heemels M-T and Ploegh H. (1995). Annu. Rev. Biochem., 64, 463 ± 491. Herrmann CH, Dery CV and Mathews MB. (1987). Oncogene, 2, 25 ± 35. Keating PJ, Cromme FV, Duggan-Keen M, Snijders PJF, Walboomers JMM, Hunter RD, Dyer PA and Stern PL. (1995). Br. J. Cancer, 72, 405 ± 411. Kimura G, Itagaki S. and Summers J. (1975). Int. J. Cancer, 15, 694 ± 706. Matlashewski G, Schneider J, Banks L, Jones N, Murray A and Crawford L. (1987). EMBO J., 6, 1741 ± 1746. Nishitani J, Nishinaka T, Cheng C-H, Rong W, Yokoyama KK and Chiu R. (1999). J. Biol. Chem., 274, 5454 ± 5461. Phelps WC, Yee CL, MuÈnger K and Howley PM. (1988). Cell, 53, 539 ± 547. Phelps WC, Bagchi S, Barnes JA, Raychaudhuri P, Kraus V, MuÈnger K, Howley PM and Nevins JR. (1991). J. Virol., 65, 6922 ± 6930. Prott JL, Sharma E and Blair GE. (1994). Nucleic Acids Res., 22, 4779 ± 4788. Prott JA and Blair GE. (1997). FEBS Lett., 400, 141 ± 144. Rey O, Lee S and Park NH. (2000). J. Virol., 74, 4912 ± 4918. Schneider-Gaedicke A and Schwartz E. (1986). EMBO J., 5, 2285 ± 2292. Smotkin D and Wettstein FO. (1986). Proc. Natl. Acad. Sci. USA, 83, 4680 ± 4684. Stern PL. (1996). Adv. Cancer Res., 69, 175 ± 211. Storey A, Pim D, Murray A, Ostborn K, Banks L and Crawford L. (1988). EMBO J., 6, 1815 ± 1820. Townsend A, Elliott T, Cerundolo V, Foster L, Barber B and Tse A. (1990). Cell, 62, 285 ± 295. van Endert PM. (1999). Curr. Opin. Immunol., 11, 82 ± 88. Wilczynski SP, Pearlman L and Walker J. (1988). Virology, 166, 624 ± 627. Wilczynski SP, Oft M, Cook N, Liao SY and Iftner T. (1993). Hum. Pathol., 24, 96 ± 102. Williams DB, Vassilakos A and Suh W-K. (1996). Trends Cell Biol., 6, 267 ± 273. Zwerschke W and Jansen-DuÈrr P. (2000). Adv. Cancer Res., 78, 1 ± 29.

Oncogene