In?. J. Cancer: 46, 703-711 (1990) 0 1990 Wiley-Liss, Inc.
Publication of the International Union Against Cancer Publication de I’Union lnternationale Contre 1e Cancer
MAPPING OF LINEAR EPITOPES OF HUMAN PAPILLOMAVIRUS TYPE 16: THE E l , E2, E4, E5, E6 AND E7 OPEN READING FRAMES Joakim DILLNER Department of Virology, Karolinska InstisUte, S-10521 Stockholm, Sweden.
Certain types of human papillomavirus (HPV), especially HPV type 16, are associated with proliferative lesions of the cervix uteri that can progress to malignancy. In order to map the linear epitopes of the HPV-encoded proteins, we have synthesized the predicted amino acid sequences of the open reading frames (ORFs) in the early region of HPV 16, as a set of 94 synthetic 20-residue peptides overlapping each other with 5 amino acids. The peptides were tested for reactivity with IgA, IgG and IgM antibodies in the sera of 30 patients with HPV 16-carrying cervical neoplasia. The E l ORF had only low immunoreactivity, but several relatively minor epitopes were identified in the carboxyterminal part. The E2 ORF was found to contain several epitopes that were highly immunoreactive with a majority (up to 87%) of the cervical cancer patients’ sera. The E4 ORF had one major, regularly IgA- and IgG-reactive epitope, whereas the ES and E6 ORFs had only a few minor epitopes. The E7 ORF had several epitopes that were highly immunoreactive, but only with a minority of patients’ sera. The 10 most immunoreactive peptides were also analyzed for immunoreactivity with 60 control sera, of which 22 were derived from patients with parotid gland tumors and 38 were derived from healthy volunteers. Most of the peptides were also immunoreactive with the control sera. However, the IgA antibodies, and to some extent the IgG antibodies, were found at much lower levels among the controls.
The human papillomaviruses are a widespread family of viruses that cause various proliferative diseases in the infected epithelium (Pfister, 1984). Different types of HPV are associated with different diseases and different anatomical sites of infection (Pfister, 1984, 1987). Several types, notably HPVs 6, 11, 16, 18, 31, 33, 35 and 52, infect the genital tract (Pfister, 1987; Shah and Buscema, 1988). HPV types 6 and 11 are mainly associated with elevated, pointed, wart-like lesions, condyloma acuminata (Gissman et al., 1983; Shah and Buscema, 1988). In contrast, HPV types 16, 18, 31, 33, 35 and 52 are mainly associated with flat or “inverted” inconspicuous lesions, that are designated cervical intra-epithelial neoplasia (CIN) (Crum et al., 1985; Shah and Buscema, 1988). These lesions can progress, albeit at low frequency, to cervical carcinoma (Nasiell and Nasiell, 1986; zur Hausen, 1989). More than 90% of cervical carcinomas carry some type of HPV, with HPV 16 alone being found in about 60% of tumors (Durst et al., 1983; Boshart et al., 1984; Crum et al., 1984; Pfister, 1987). All HPV genomes have at least 8 potentially proteinencoding regions, open reading frames (ORFs) that are numbered L1, L2 and El-E8 (Pfister, 1987; Baker, 1987). The L1 and L2 ORFs encode late proteins mainly expressed in virusproducing tissue, but not in HPV-carrying tumor cells (Li et al., 1987; Komly et al., 1986). The E l ORF is the longest ORF of the papillomavirus ORFs and is also one of the best conserved ORFs (Baker, 1987). Its function and protein expression are unknown for HPV, but for bovine papillomavirus (BPV) the E l ORF has been assigned 3 distinct functions. The aminoterminal part encodes a negative regulation factor, “M”, for the BPV episomal replication (Berg et al., 1986; Lusky and Botchan, 1986), the carboxyterminal part of E l encodes a positive regulation factor, “R”, for the episomal replication (Lusky and Botchan, 1986) and the full-length E l ORF en-
codes a repressor of viral transcription and transformation (Schiller et al., 1989). The predicted amino acid sequence of the E l ORF has a strong homology to the large T antigens of the polyoma viruses, especially for the regions known to be involved in the nucleotide binding and ATPase activities of the large T antigens (Clertant and Seif, 1984). The E2 ORF codes for nuclear DNA-binding proteins (Mallon et al., 1987; Li et al., 1988; Lambert et al., 1989) that can transactivate cellular and viral genes (Spalholz et al., 1985). The E2 ORF greatly enhances the transforming ability of BPV, possibly due to its transactivating effect (Sarver et al., 1984; DiMaio et al., 1986). A truncated E2 protein encoded by the carboxyterminal part of the E2 ORF can act as a viral repressor instead of transactivator (Lambert et al., 1987). The E4 ORF does not contain an initiation codon in HPV 16, but spliced PV mRNAs have been described where an initiation codon and a short aminoterminal sequence of the protein are provided by upstream exons in the E l (Rotenberg et al., 1989) or E6 ORFs (Pettersson et al., 1987). E4 proteins have been identified for HPV 1 (Doorbar et al., 1986, 1988) and HPV 11 (Brown et al., 1988). The major protein is an extremely abundant 16- to 17-kDa doublet phosphoprotein, whose expression correlates to the productive stage of infection (Doorbar et al., 1986, 1988; Grand et al., 1989) and may be translated from the El-E4 spliced mRNA (Doorbar et al., 1988). Minor E4 proteins of molecular weights, 10, 11,21,23,32 and 34 kDa have also been described (Doorbar et al., 1988; Brown et al., 1988; Seedorf et al., 1987). The sequence of the E4 O W is poorly conserved among different HPV types, although the predicted proteins have physical characteristics in common (Doorbar et al., 1989). Little is known about the function and expression of the E5 region of HPV. The E5 ORF of BPV encodes a short (7-kDa) protein that is primarily localized to the Golgi complex (Bergman, 1989). The E5 gene has a strong transforming effect on mouse fibroblasts (Schiller et al., 1986; Bergman et al., 1988) and the E5 protein of BPV induces DNA replication when injected into the nuclei of frog oocytes (Green and Loewenstein, 1987). The principal viral mRNA transcripts in HPV-carrying tumor cells are from the E6, E7 region (Schwartz et al., 1985). The E6 ORF codes for a 16-kDa zinc-binding (Barbosa et al., 1989) protein that, together with the E7 OW, is necessary for HPV- 16-induced transformation of human keratinocytes or fibroblasts (Munger et al., 1989; Watanabe et al., 1989). E6encoded proteins of the bovine papillomavirus (BPV) may be localized both in the nucleus and in the cell membrane (Androphy et al., 1985). The E7 ORF codes for a 21-kDa zincbinding (Barbosa et al., 1989) nuclear (Sato et al., 1989) phosphoprotein (Smotkin and Wettstein, 1987) that has transforming ability (Phelps et al., 1988; Tanaka et al., 1989), possibly due to the fact that the E7 protein can form complexes with the retinoblastoma-associated tumor suppressor protein, rb (Dyson et al., 1989).
Received: June 1, 1990.
Although the genetic functions of these ORFs have been extensively studied, their immunological properties are virtually unknown. Jenison et al. (1988) reported that only 2 of 72 condyloma patients had IgG antibodies to an E2-derived bacterial fusion protein of HPV 6, and that none of these patients had antibodies to E6 and E7-derived fusion proteins. The IgA response to linear epitopes of the EZencoded nuclear antigen of HPV 16 (J. Dillner et al., 1989) and to the group-specific epitopes in the capsid proteins (L. Dillner et al., 1989) predominates over the IgG response to these proteins in patients with CIN, suggesting that IgA-specific immunoassays using linear epitopes of HPV proteins could be used to study the epidemiology of HPV in relation to CIN (L. Dillner et al., 1989; J. Dillner et al., 1989). In order to obtain an epitope map of the HPV 16 proteins, we synthesized the predicted amino acid sequences of the HPV 16 ORFs as a set of 20 amino-acid synthetic peptides with an overlap of 5 amino acids and tested the peptides for reactivity with IgA, IgG or IgM antibodies in the sera from patients with HPV-16-canying cervical neoplasia. This report concerns the ORFs of the early region, whereas the linear epitopes of the late-region ORFs were reported previously (J. Dillner et ul., 1990). MATERIAL AND METHODS
Peptide synthesis Ninety-four 20 amino-acid peptides, with an overlap of 5 amino acids, representing the deduced amino acid sequence of the E l , E2, E4, E5, E6 and E7 ORFs of HPV 16 (Seedorf et al., 1985) were synthesized using t-Boc amino acids (Bachem, Bubendorf, Switzerland) and p-methylbenzhydrylamine resin (Fluka, Buchs, Switzerland) according to the multiple solidphase peptide synthesis method (Houghten, 1985). Removal of the protecting groups from the formyl-tryptophane and methionine sulfoxide residues was achieved by cleavage with 25% hydrogen fluoride (Tam et al., 1983). The peptides were then cleaved from the resin with liquid hydrogen fluoride using a multi-vessel apparatus (Houghten et al., 1986). In order to denote the position in the protein, the putative initiation codon was assigned to be amino acid number 1. Synthetic peptide number 1 corresponds to amino acids 2-21 in the predicted protein, peptide number 2 to amino acids 17-36 and so on until the carboxyterminal peptide. However, the 2 peptides at the carboxyterminus of E6 and E7 (E6:9, E6:10, E7:5 and E7:6) were instead synthesized as 21-residue peptides and the aminoterminal peptide of E2, E2:1, was synthesized as a 19residue peptide. The overlap between the peptides was 5 amino acids. The E4 ORF does not contain any initiation codons and the first E4 peptide does therefore correspond to the
first 20 amino acids in the ORF. In the case of both the E l and the E5 ORFs, the original sequence reported by Seedorf et al. (1985) has been corrected by subsequent studies (Halbert and Galloway, 1988; Baker et al., 1987). The amino acid sequences of the peptides from these twg ORFs were therefore: E 1: ADPAGTNGEEGTGCNGWFYV, GWFYVEAVVEKKTGDAISDD, AISDDENENDSDTGEDLVDF, DLVDFIVNDNDYLTQAETET, AETETAHALFTAQEAKQHRD, KQHRDAVQVLKRKYLGSCIE, GSCIEKQSRAAKRRLFESED, FESEDSGYGNTEVETQQMLQV, LQVEGRHETETPCSQYSGG, QYSGGSGGGCSQYSSGSGGE, GSGGEGVSERHTICQTPLTN, TPLTNILNVLKTSNAKAAML, KAAMLAKFKELYGVSFSELV, FSELVRPFKSNKSTCCDWCI, CDWCIAAFGLTPSIADSIKT, DSIKTLLQQYCLYLHIQSLA, IQSLACSWGMVVLLLVRYKC, VRYKCGKNRETJEKLLSKLL, LSKLLCVSPMCMMIEPPKLR, PPKLRSTAAALYWYKTGISN, TGISNISEVYGDTPEWIQRQ, WIQRQTVLQHSFNDCTFELS , TFELSQMVQWAYDNDIVDDS, IVDDSEIAYKYAQLADTNSN, DTNSNASAFLKSNSQAKIVK, AKIVKDCATMCRHYKRAEKK, RAEKKQMSMSQWIKYRCDRV, RCDRVDDGGDWKQIVMFLRY, MFLRY QGVEFMSFLTALKRF,ALKRFLQGIPKKNCILLYGA, LLYGAANTGKSLFGMSLMKF, SLMKFLQGSVICFVNSKSHF, SKSHFWLQPLADAKIGMLDD, GMLDDATVPCWNYIDDNLRN, DNLRNALDGNLVSMDVKHRP, VKHRPLVQLKCPPLLITSNI, ITSNINAGTDSRWPYLHNRL, LHNRLVVFTFPNEFPFDENG, FDENGNPVYELNDKNWKS , KNWKSFFSRTWSRLSLHE , LSLHEDEDKENDGDSLPT, DSLPTFKCVSGQNTNTL. E5: TNLDTASTTLLACFLLCFCV, LCFCVLLCVCLLIRPLLLSV, LLLSVSTYTSLIILVLLLWI, LLLWITAASAFRCFIVYIIF, VYIIFVYIPLFLIHTHARFLIT. The amino acid sequences of the most immunoreactive peptides are listed in Table I. Sera Thirty sera from patients with HPV-16-carrying cervical carcinoma, 22 sera from patients with mixed carcinoma of the parotid gland and 21 sera from healthy volunteers were collected by Dr. D. DiLuca at the University of Ferrarra, Italy within the framework of a previous collaborative study (J. Dillner et al., 1990). Seventeen sera from healthy female laboratory workers were collected at the Virology Dept., National Bacteriological Laboratories, Stockholm, Sweden. ELISA ELISA was performed as described (J. Dillner et al., 1990). Briefly, the synthetic peptides were diluted to 20 kg/ml in 10 m~ carbonate buffer, PH 9.6 and kept overnight in half-area (50 pl/well) microtiter plates (Costar, Cambridge, MA). After
TABLE I - DETECTION OF SIGNIFICANTLY ELEVATED ANTIBODY TITERS AGAINST HPV 16 SYNTHETIC PEPTIDES AMONG SERA FROM 30 PATIENTS WITH HPV-16-CARRYING CERVICAL NEOPLASIA AS COMPARED TO A CONTROL GROUP OF 60 SERA FROM PATIENTS WITH OTHER TUMORS OR FROM
E1:33 E1:39 E2:9 E2:13 E217 E2:19 E4:4 E6:6 E7:1 E7:4
63% 43% 70% 87% 70% 30% 67% 23% 23% 30%
37% 18% 62% 68% 55% 20% 30% 13% 2% 2%