Research Article

Immunology and Cell Biology (2000) 78, 580–585; doi:10.1046/j.1440-1711.2000.00949.x

Computer-assisted analysis of molecular mimicry between human papillomavirus 16 E7 oncoprotein and human protein sequences

Costanzo Natale1, Teresa Giannini2, Alberta Lucchese2 and Darja Kanduc2

  1. 1General Surgery Division, United Hospitals, University of Foggia, Foggia, Italy
  2. 2CARSO Cancer Research Center and Department of Biochemistry and Molecular Biology, University of Bari, Bari, Italy

Correspondence: D Kanduc, Department of Biochemistry and Molecular Biology, University of Bari, Via Orabona 4, 70126 Bari, Italy. Email:

Received 15 February 2000; Accepted 1 May 2000.



The immunology of human papillomavirus (HPV) infections has peculiar characteristics. The long latency for cervical cancer development after primary viral infection suggests mechanisms that may aid the virus in avoiding the host immunosurveillance and establishing persistent infections. In order to understand whether molecular mimicry phenomena might explain the ability of HPV to avoid a protective immune response by the host cell, sequence similarity between HPV16 E7 oncoprotein and human self-proteins was examined by computer-assisted analysis. Data were obtained showing that the HPV16 E7 protein has high and widespread similarity to several human proteins involved in a number of critical regulatory processes. In addition, multiple identical and different E7 peptide motifs are present in the same human protein. Thus, sharing of common motifs between viral oncoproteins and molecules of normal cells may be one cause underlying the scarce immunogenicity of HPV infections. The hypothesis is advanced that synthetic peptides harbouring viral motifs not and/or scarcely represented in the host's cellular proteins may represent a valuable immunotherapeutic approach for cervical cancer treatment.


E7 oncoprotein, human papillomavirus 16, molecular mimicry, sequence similarity analysis



Cervical cancer causes 15% of deaths from cancer in women worldwide. Epidemiological data indicate the sexually transmitted high oncogenic risk human papillomaviruses (HPV) 16 and 18 as aetiological agents.1, 2 In particular, the expression of the viral E6 and E7 oncoproteins appears to contribute to malignant progression. Consequently, great efforts have been directed towards designing therapeutic vaccines against HPV-induced cervical carcinoma by using the HPV16/18 E6 and E7 proteins as tumour-associated antigens.3, 4

However, the time when a vaccine will have an impact on the prevalence of papillomavirus-induced dysplasia and cancer seems to be still some way off.5 For example, a successful papillomavirus vaccine strategy should evoke CTL able to recognize antigenic fragments of internal viral proteins. Therefore, a reason for the failure of the immune system to control HPV infection may also reside in the scarce antigenicity, that is, scarce 'non-self' character, of the viral peptides exhibited by the MHC groove.

The ability of the immune system to distinguish between self and non-self molecules is an important property in maintaining tissue/organism integrity. Breakage of this self- tolerance is one of the main bases for autoimmune diseases.6, 7 However, the tolerance mechanisms used to prevent autoimmune destruction could be the main basis through which tumour-associated-antigens escape from functional antigen-specific immune recognition. To gain a better understanding of the recognition of HPV protein antigens by the immune system, we have analysed the sharing of amino acid (aa) motifs between human 'self' molecular determinants and the HPV16 E7 oncoprotein by similarity sequence studies.


Materials and Methods

The analysed HPV16 E7 oncoprotein sequence was as reported by Seedorf et al.8 (Medline accession no. K02718). As controls, we analysed the sequences of the following proteins: (i) small t antigen (SWISS-PROT accession no. P03081) from simian virus 40 (SV40); (ii) the non-structural protein NS2A (Medline U89339) from yellow fever virus (YFV); and (iii) three fragments from the haemagglutinin-neuramidase (HN) protein (EMBL accession no. X79092) from Newcastle disease virus (NDV). The three polypeptide fragments from the haemagglutinin-neuramidase protein were approximately 33 aa long each, for a total of 100 aa, and were spaced at almost regular intervals along the entire protein sequence. The fragments were: aa 176–208, fragment 1; aa 337–369, fragment 2; and aa 467–499, fragment 3. Sequence similarity analyses were conducted by using the MEDLINE, FASTA, BLAST, PIR, SWISS-PROT and PRINTS sequence analysis programs.9, 10, 11 The SYFPEITHI program ( was used as database of HLA ligands and peptide motifs.12



Molecular mimicry is defined as similar motifs/structures shared by molecules from dissimilar genes or by their protein products.13 To investigate the molecular mimicry between HPV16 E7 oncoprotein sequence and human proteoma, a systematic study of sequence similarity was done by dissecting the E7 oncoprotein sequence into 7, 6, and 5 aa motifs that were used as sequence probes.

As controls, we analysed the sequences of the following viral proteins: (i) three fragments from the NDV HN protein; (ii) YFV NS2A; and (iii) SV40 small t antigen. Sequences from the NDV HN protein were examined because of the high immunogenic potential shown by the ssRNA NDV. In fact, it has repeatedly reported that treatment with lysates of NDV-infected allogeneic human tumour is able to elicit humoral immune responses against tumour cell-associated antigens, thus breaking the tumour immune tolerance.14, 15 As a further control, NS2A sequence from the YFV was examined, because seroepidemiological surveys in African populations have shown some seropositivity for YFV antibodies, thus indicating the ability of this ssRNA virus to elicit an antibody response.16 Therefore, a low degree of similarity to human protein sequences was expected in YFV NS2A and in NDV HN protein sequences compared with HPV16 E7. In addition, the cell growth regulatory small t antigen from the dsDNA virus SV4017 was also analysed in order to have a genome/function-based control, by HPV16 being a dsDNA virus and E7 a growth regulatory protein.

By using 7-mer sequence probes, it was found that the E7 protein 7 aa motif QLNDSSE gives one human match corresponding to Na+/Pi transport protein 4 (SwissProt O00476). The E7 SSEEEDE motif is present in xeroderma pigmentosum group G (XP-G) complementing protein (SwissProt P28715). Not unexpectedly, the same motif is also present in retinoblastoma binding protein 1 (RBBP-1; SwissProt P29374), a critical cell-cycle regulatory protein. In contrast, no human polypeptide has 7-mer motifs in common with the control SV40 small t antigen, NDV HN or YFV NS2A proteins.

These data provided the incentive for a thorough analysis of the E7 motifs present in human proteoma. Because 5–6 aa are the minimum requisite to induce an antibody response, the oncoprotein sequence and the control sequences were dissected into five-mer motifs that were used as sequence probes. Figure 1 illustrates the similarity sequence data we obtained. It can be seen that all four proteins examined here present motifs in common with the human proteoma. However, the highest number of matches was found in the E7 oncoprotein sequence (Figure 1a). The SV40 small t antigen sequence showed similarity to 5-mer portions of a number of human proteins (Figure 1b), thus indicating the tendency of dsDNA viruses to 'borrow' genetic information and, consequently, sequence similarity from their hosts in a way that ssRNA viruses cannot. At the same time, it is evident that long viral sequences in SV40 small t antigen have no matches at all to human proteoma, thus offering possible epitopic determinants unknown to the host. The three HN control fragments from the immunogenic NDV had the lowest number of human matches (Figure 1c). In addition, YFV NS2A showed fewer human matches than E7 oncoprotein (Figure 1d).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact or the author

Plot of human matches versus viral 5-mer peptide probes derived from (a) E7 oncoprotein, (b) SV40 small t antigen, (c) Newcastle disease virus haemagglutinin-neuramidase polypeptide fragments and (d) yellow fever virus NS2A protein sequence.

Full figure and legend (43K)

Further computer-assisted analysis showed that a number of human proteins harboured multiple HPV16 E7 4-mer motifs of both identical and different peptide sequences. Three basic examples are reported in Table 1.

To determine the immunological potencies of shared and unshared peptide sequences, the ability to bind HLA I and II molecules by the E7 sequences that were found to be similar and dissimilar to human proteins was also examined. To this aim, we analysed two E7 fragments: EQLNDSSEEEDEIDGPAGQAE (aa 26–46), which has a high level of similarity to human proteoma (total number of 5-mer human matches, 290), and AEPDRAHYNIVTFCCKCDSTL (aa 45–65), which has a low level of similarity to human proteins (total number of 5-mer human matches, 14; see Figure 1). The two fragments were analysed for potential T-cell epitopes taking into consideration the amino acids in the anchor and auxiliary anchor positions by using SYFPEITHI program.12 In this program, the HLA-binding potential score is calculated by giving the amino acids of a certain peptide a specific value depending on whether they are anchor, auxiliary anchor or preferred residues. Amino acids that are regarded as having a negative effect on the binding ability are also evaluated by a negative value. Table 2 illustrates the data we obtained by submitting the two E7 viral polypeptide sequences to SYFPEITHI program analysis. On the whole, the table shows that peptides derived from the high-similarity E7 sequence EQLNDSSEEEDEIDGPAGQAE show a general tendency to bind to HLA-A type molecules with higher strength than peptides from the low-similarity E7 polypeptide AEPDRAHYNIVTFCCKCDSTL. In contrast, unshared sequences have higher binding potential to HLA-B-type molecules than shared motifs.



Molecular mimicry has been thoroughly studied as a phenomenon underlying autoimmune responses and diseases. When linear and/or conformational aa sequences are shared by microbial/viral agents and 'self' molecules, autoimmunity may occur if the host immune response against the infectious agents cross-reacts with the host's 'self' sequences.13 However, molecular mimicry induced autoimmunity can occur when the non-self and host determinants are similar enough to cross-react, yet different enough to break immunological tolerance. When high degrees of similarity are present between non-self and self molecules, the breaking of the powerful self-tolerance mechanisms that avoid harmful self-reactivity seems unlikely. Thus, sharing of epitopes with the host's molecules may represent an elective viral mechanism to escape immune attack. We suggest that this may be the case for HPV16 on the basis of the database searches and experimental results presented in this study. The HPV are viruses of low immunogenicity and numerous studies have demonstrated that the proliferation and malignant phenotype of human cervical carcinoma cell culture depend on continuous expression of the high-risk type HPV oncogenes E6 and E7.18 In particular, the oncoprotein E7 is believed to play a major role in cervical neoplasia.2, 19 Some studies have reported that antibodies to HPV16 E6 and E7 proteins are found more frequently in patients with invasive cervical cancer, while others have reported an inconsistent antibody response among patients with invasive HPV16 cervical cancer.20 Consequently, it sounds logical to postulate that the success of HPV infection in terms of cancer disease depends on the avoidance of the host's immune surveillance system that should destroy the foreign viral oncoproteins. From this point of view, the high number of E7 matches to human proteins (see Figure 1 and Table 1) makes HPV16 successful, as far as protection by molecular mimicry of the E7 product is regarded.

Although epitopes are often defined by 3-D structure, we have used linear probes of differing lengths to discover peptide sharing potentially involved in molecular mimicry. Given this caveat and remembering that important biological functions are carried out by short linear motifs (e.g. RGDS, KFERD, KDEL motifs are signals for integrin binding, lysosomal targeting and endoplasmic reticulum retention),21, 22 the high level of molecular mimicry between the viral oncoproteins and human proteins appears of extreme interest. It could be predicted that antibody production directed against the E7 oncoprotein may also result in hitting the XP-G complementing protein, which shares a 7-mer motif with E7. That would result in the inhibition of excision repair by normal cells, as already demonstrated in vitro.23 Analogously, nuclear phosphoprotein RBBP-1 might also become a target of anti-E7 antibodies. In such a case, derangement of cell-cycle regulation would be an obvious consequence. In addition, it has to be observed that detection of Ki-67 antigen during cervical carcinogenesis may be hampered by the multiple and different motifs present in both E7 and Ki-67 antigen (see Table 1).24, 25

Finally, the present study may have important implications for a possible vaccine strategy. It has been reported that synthetic peptides consisting of peptide 48–54 (DRAHYNI) linked to major B-cell epitopes on the E7 molecule form immunogens capable of eliciting strong antibody responses to HPV16 E7.26 It has also been demonstrated that a CTL epitope-based peptide vaccine of HPV16 E7 aa 44–62 elicits protection against outgrowth of HPV16-transformed tumour cells in mice.27, 28 We have observed here that the E7 sequence 44–62 is also the E7 motif hosting the lowest number of matches to human proteins. Moreover, the preliminary analysis for possible T-cell epitopes shows the tendency of unshared E7 peptide sequences to bind preferentially to HLA-B type molecules (see Table 2). These data suggest that presumably the molecular mimicry phenomenon may modulate the immunogenicity of unshared peptides depending on the type of bound HLA molecules. In this context, molecular mimicry may be one of the postulated relevant factors other than HLA affinity that affect immunogenicity.29, 30

In conclusion, by considering that relatively short, linear peptides can often induce useful reactive antibodies, probably as result of induced fit,31 we advance the hypothesis that artificial peptides made by short motifs absent and/or under-represented in host's self proteins may be a profitable immunotherapeutic approach to cervical cancer treatment.



  1. zur Hausen H. Papillomavirus infections – A major cause of human cancers. Biochim. Biophys. Acta 1996; 1288: F55–78. | Article | PubMed | ChemPort |
  2. Zehbe I, Ciccolini F, Dell'Orco M et al. E7 protein of human papillomaviruses and its interaction with cellular pathways. In: Bannasch P, Kanduc D, Papa S, Tager JM (eds). Cell Growth and Oncogenesis. Basel: Birkhäuser Verlag,1998; 97–107.
  3. Ji H, Chang EY, Lin KY, Kurman RJ, Pardoll DM, Wu TC. Antigen-specific immunotherapy for murine lung metastatic tumors expressing human papillomavirus type 16 E7 oncoprotein. Int. J. Cancer 1998; 78: 41–5. | Article | PubMed | ISI | ChemPort |
  4. Steller MA, Gurski KJ, Murakami M et al. Cell-mediated immunological responses in cervical and vaginal cancer patients immunized with a lipidated epitope of human papillomavirus type 16 E7. Clin. Cancer Res 1998; 4: 2103–9. | PubMed | ChemPort |
  5. Rowen D, Lacey C. Toward a human papillomavirus vaccine. Dermatol. Clin. 1998; 16: 835–8. | PubMed | ChemPort |
  6. Baum H, Davies H, Peakman M. Also... Immunol. Today 1997; 18: 252–3. | ChemPort |
  7. Tung KS. Mechanism of self-tolerance and events leading to autoimmune disease and autoantibody response. Clin. Immunol. Immunopathol. 1994; 73: 275–82. | Article | PubMed | ChemPort |
  8. Seedorf K, Kraemmer G, Dürst M, Suhai S, Roewekamp WG. Human papillomavirus type 16 DNA sequence. Virology 1985; 145: 181–5. | Article | PubMed | ISI | ChemPort |
  9. George DG, Barker WC, Mewes H-W, Pfeiffer F, Tsugita A. The PIR-international protein sequence database. Nucleic Acids Res. 1996; 24: 17–20. | Article | PubMed | ChemPort |
  10. Bairoch A, Apweiler R. The SWISS-PROT protein sequence data bank and its new supplement TREMBL. Nucleic Acids Res. 1996; 24: 21–5. | Article | PubMed | ChemPort |
  11. Attwood TK, Beck ME, Flower DR, Scordis P, Selley J. The PRINTS protein fingerprint database in its fifth year. Nucleic Acids Res. 1998; 26: 304–8. | Article | PubMed | ChemPort |
  12. Rammensee H, Bachmann J, Emmerich NP, Bachor OA, Stevanovic S. SYFPEITHI: Database for MHC ligands and peptide motifs. Immunogenetics 1999; 50: 213–19. | Article | PubMed | ISI | ChemPort |
  13. Oldstone MBA. Molecular mimicry and immune-mediated diseases. FASEB J. 1998; 12: 1253–65. | PubMed |
  14. Savage HE, Rossen RD, Hersh EM, Freedman RS, Bowen JM, Plager C. Antibody development to viral and allogeneic tumor cell-associated antigens in patients with malignant melanoma and ovarian carcinoma treated with lysates of virus-infected tumor cells. Cancer Res. 1986; 46: 2127–33. | PubMed | ChemPort |
  15. von Hoegen P, Weber E, Schirrmacher V. Modification of tumor cells by a low dose of Newcastle disease virus. Augmentation of the tumor-specific T cell response in the absence of an anti-viral response. Eur. J. Immunol. 1988; 18: 1159–66. | PubMed | ChemPort |
  16. Paix MA, Poveda JD, Malvy D, Bailly C, Merlin M, Fleury HJ. Serological study of the virus responsible for hemorrhagic fever in an urban population of Cameroon. Bull. Soc. Pathol. Exot. Filiales 1988; 81: 679–82. | PubMed | ChemPort |
  17. Sontag E, Sontag JM, Garcia A. Protein phosphatase 2A is a critical regulator of protein kinase C zeta signaling targeted by SV40 small t to promote cell growth and NF-kappaB activation. EMBO J. 1997; 16: 5662–71. | Article | PubMed | ISI | ChemPort |
  18. von Knebel Doeberitz M, Rittmüller C, Aengeneyndt F, Jansen-Dürr P, Spitkovsky D. Reversible repression of papillomavirus oncogene expression in cervical carcinoma cells: consequences for phenotype and E6–p53 and E7–RB interactions. J. Virol. 1994; 68: 2811–21. | PubMed | ChemPort |
  19. Tommasino M, Jansen-Dürr P. E7 Protein. In: Tommasino M (ed.). Papillomaviruses in Human Cancer: The Role of E6 and E7 Oncoproteins. Austin: Landes Bioscience, 1997; 103–36.
  20. Frazer IH. Immunology of papillomavirus infection. Curr. Opin. Immunol. 1996; 8: 484–91. | Article | PubMed | ISI | ChemPort |
  21. Baum H, Davies H, Peakman M. Molecular mimicry in the MHC: Hidden clues to autoimmunity? Immunol. Today 1996; 17: 64–70. | Article | PubMed | ChemPort |
  22. Roudier C, Anger I, Roudier J. Molecular mimicry reflected through database screening: Serendipity or survival strategy? Immunol. Today 1996; 17: 357–8. | PubMed | ChemPort |
  23. O'Donovan A, Wood RD. Identical defects in DNA repair in xeroderma pigmentosum group G and rodent ERCC group 5. Nature 1993; 363: 185–8. | Article |
  24. Dellas A, Schultheiss E, Leivas MR, Moch H, Torhorst J. Association of p27Kip1, cyclin E and c-myc expression with progression and prognosis in HPV-positive cervical neoplasms. Anticancer Res. 1998; 18: 3991–8. | PubMed | ISI | ChemPort |
  25. Devictor B, Bonnier P, Piana L et al. c-myc protein and Ki-67 antigen immunodetection in patients with uterine cervix neoplasia: Correlation of microcytophotometric analysis and histological data. Gynecol. Oncol. 1993; 49: 284–90. | Article | PubMed | ChemPort |
  26. Tindle RW, Fernando GJ, Sterling JC, Frazer IH. A 'public' T-helper epitope of the E7 transforming protein of human papillomavirus 16 provides cognate help for several E7 B-cell epitopes from cervical cancer-associated human papillomavirus genotypes. Proc. Natl. Acad. Sci. USA 1991; 88: 5887–91. | Article | PubMed | ChemPort |
  27. Feltkamp MC, Smits HL, Vierboom MP et al. Vaccination with cytotoxic T lymphocyte epitope-containing peptide protects against a tumor induced by human papillomavirus type 16-transformed cells. Eur. J. Immunol. 1993; 23: 2242–9. | PubMed | ISI | ChemPort |
  28. Sarkar AK, Tortolero-Luna G, Nehete PN, Arlinghaus RB, Mitchell MF, Sastry KJ. Studies on in vivo induction of cytotoxic T lymphocyte responses by synthetic peptides from E6 and E7 oncoproteins of human papillomavirus type 16. Viral Immunol. 1995; 8: 165–74. | PubMed | ChemPort |
  29. Feltkamp MC, Vierboom MP, Kast WM, Melief CJ. Efficient MHC class I-peptide binding is required but does not ensure MHC class I-restricted immunogenicity. Mol. Immunol. 1994; 31: 1391–401. | Article | PubMed | ChemPort |
  30. Ochoa-Garay J, McKinney DM, Kochounian HH, McMillan M. The ability of peptides to induce cytotoxic T cells in vitro does not strongly correlate with their affinity for the H-2Ld molecule: Implications for vaccine design and immunotherapy. Mol. Immunol. 1997; 34: 273–81. | Article | PubMed | ChemPort |
  31. Sutcliffe JG, Shinnick TM, Green N, Lerner RA. Antibodies that react with predetermined sites on proteins. Science 1983; 219: 660–6. | Article | PubMed | ISI | ChemPort |


We thank Daniela Dellino for excellent secretarial help. TG is a PhD student supported by the European Community. Part of the present study was supported by Assessorato all'Agricoltura, Regione Puglia.