JOURNAL OF VIROLOGY, Nov. 1998, p. 9121–9130 0022-538X/98/$04.0010 Copyright © 1998, American Society for Microbiology. All Rights Reserved.
Vol. 72, No. 11
Receptor Binding Sites and Antigenic Epitopes on the Fiber Knob of Human Adenovirus Serotype 3 HERBERT LIEBERMANN,1* RENATE MENTEL,1 ULRIKE BAUER,1 PATRICIA PRING-ÅKERBLOM,2 ¨ LLING,3 SUSANNE MODROW,4 AND WERNER SEIDEL1 RUDOLF DO Institut fu ¨r Medizinische Mikrobiologie, Ernst-Moritz-Arndt-Universita ¨t Greifswald, D-17487 Greifswald,1 Institut fu ¨r Virologie und Seuchenhygiene, Medizinische Hochschule Hannover, D-30623 Hanover,2 Biosyntan, D-13125 Berlin,3 and Institut fu ¨r Medizinische Mikrobiologie und Hygiene, Universita ¨t Regensburg, D-93042 Regensburg,4 Germany Received 19 February 1998/Accepted 8 July 1998
The adenovirus fiber knob causes the first step in the interaction of adenovirus with cell membrane receptors. To obtain information on the receptor binding site(s), the interaction of labeled cell membrane proteins to synthetic peptides covering the adenovirus type 3 (Ad3) fiber knob was studied. Peptide P6 (amino acids [aa] 187 to 200), to a lesser extent P14 (aa 281 to 294), and probably P11 (aa 244 to 256) interacted specifically with cell membrane proteins, indicating that these peptides present cell receptor binding sites. Peptides P6, P11, and P14 span the D, G, and I b-strands of the R-sheet, respectively. The other reactive peptides, P2 (aa 142 to 156), P3 (aa 153 to 167), and P16 (aa 300 to 319), probably do not present real receptor binding sites. The binding to these six peptides was inhibited by Ad3 virion and was independent of divalent cations. We have also screened the antigenic epitopes on the knob with recombinant Ad3 fiber, recombinant Ad3 fiber knob, and Ad3 virion-specific antisera by enzyme-linked immunosorbent assay. The main antigenic epitopes were presented by P3, P6, P12 (aa 254 to 269), P14, and especially the C-terminal P16. Peptides P14 and P16 of the Ad3 fiber knob were able to inhibit Ad3 infection of cells. peptides, and the corresponding antisera for the determination of antigenic and immunogenic epitopes. For the localization of the hemagglutination binding domain on subgenus B:2 Ad fibers, they used mutagenesis of the fibers. Eiz and PringÅkerblom (13) localized the type-specific determinant on the fibers of Ad9 and Ad19, also with mutated recombinant fiber proteins. It is known that the soluble Ad fiber and fiber knob are able not only to inhibit the attachment of Ad to cells (29, 30, 39) but also to inhibit the infection (7, 19). However, it is unknown whether selected peptide sequence motifs of the Ad fiber knob are also able to inhibit the Ad infection of cells. For the determination of receptor binding sites, synthetic peptides should be useful. The requirement for an interaction of peptides with receptor proteins or antibodies is normally the existence of continuous (linear) determinants. Peptide scanning (pepscan) has been used up to now for the mapping of linear epitopes and recently also of discontinuous epitopes (17, 37, 42). We decided to use the pepscan for the mapping of receptor binding sites on the Ad3 fiber knob. In this study, we evaluated the reaction of labeled cell plasma membrane proteins with synthetic peptides covering the knob of the Ad3 fiber and compared it to the binding of polyclonal antibodies specific for the fiber knob, fiber, and/or virus, also by using a pepscan system (4, 14, 47). The aim was the determination of possible cell receptor binding sites and of antigenic epitopes on the fiber knob. We also studied the influence of some of these peptides on Ad infection of cells.
Human adenoviruses (Ad) have been divided into six subgenera, A to F, according to their biologic, immunologic, biochemical, and oncogenic properties (45, 51). Ad3, Ad7, and Ad21 of subgenus B:1 cause outbreaks of acute respiratory infections. Ad are icosahedral, nonenveloped, double-stranded DNA viruses. The fiber is located at the 12 vertices of the viral icosahedron. It is responsible for the specific high-affinity binding of virions to the cell receptor(s) and thus determines the tissue tropism. This is essential for the use of Ad as vectors for gene therapy (22, 24, 25). The fiber of Ad3 is already well characterized. The amino acid sequence of the fiber protein of Ad3 was published previously (46). The length of the fiber of human Ad3, terminating in a knob, is 16 nm (6, 44). The N terminus of the fiber polypeptide chains, located in the tail region, interacts with the penton base, and the C terminus is in the knob (or head) (9, 52). The fibers are trimers of identical subunits (1, 50). The first step in the interaction of the Ad with a permissive cell is the attachment of the virus, via the head domain of the fiber, to a primary cell plasma membrane receptor and via the RGD sequence motif of the penton base to a second receptor (2, 3, 11, 20, 30, 31, 33, 38, 49). The primary receptor of Ad of subgroup C is different from that of subgroup B (7, 48). Despite the extensive study of the fiber, it is not known which sequence motifs of the fiber knob of Ad3 participate in the virus attachment and which are antigenic epitopes. The receptor binding sites on the fiber head of Ad2 are also unknown. Only antigenic and partly immunogenic determinants have been evaluated, using synthetic peptides (15, 28). Recently, Mei and Wadell (34) used recombinant fiber proteins, synthetic
MATERIALS AND METHODS Fiber knob-derived peptides. In addition to the already available 15- to 20-mer peptides of the Ad3 fiber knob, P8 (amino acids [aa] 210 to 224), P13 (aa 267 to 283), P15 (aa K-291 to 306), and P16 (aa 300 to 319) (Table 1) (28), three peptides (P6 [aa 187 to 200], P11 [aa 244 to 256], and P14 [aa 281 to 294]) were selected on the basis of the structural model of the Ad5 fiber and by amino acid sequence comparisons (53, 54). They were synthesized by a solid phase synthesizer and purified by preparative high-pressure liquid chromatography as previously described (28). The other overlapping 13- to 17-mer peptides were synthe-
* Corresponding author. Mailing address: Abteilung Virologie, Institut fu ¨ r Medizinische Mikrobiologie, Ernst-Moritz-Arndt-Universita¨t Greifswald, Martin-Luther-Str. 6, D-17487 Greifswald, Germany. Phone: 49-3834-865551. Fax: 49-3834-865568. E-mail:
[email protected]. 9121
9122
LIEBERMANN ET AL.
sized simultaneously (SYRO robotic system; MultiSynTec GmbH, Bochum, Germany) on Tentagel RAM resin (Rapp Polymere GmbH, Tu ¨bingen, Germany) by means of the 9-fluorenylmethoxycarbonyl (Fmoc) strategy. Peptides were synthesized at the C terminal as amides and modified at the N terminal with biotin-6-aminohexane carbonic acid. They were analyzed by analytical reversedphase high-pressure liquid chromatography and by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (MALDI I; Shimadzu/Kratos). Cells. For cell membrane protein (CM) preparations, FL or HeLa cells were grown in monolayer cultures in Eagle’s minimal essential medium containing 7 to 10% heat-inactivated fetal calf serum and subcultured every 5 to 6 days. These cells in medium containing 2% fetal calf serum were also used for virus propagation. Labeling of cell surface proteins with biotin or DIG and their preparation. The labeling and preparation steps were performed similarly to those described by De Verdugo et al. (10) and Krah (23). The cell monolayers were washed with phosphate-buffered saline (PBS) (pH 7.2) and incubated in 10 mM EDTA in PBS for 12 min at 37°C, and the rest of the monolayer was scraped off, washed twice with PBS, and resuspended in PBS (approximately 4 3 107 cells per ml); then 70 ml of freshly prepared D-biotinoyl-ε-amidocaproic acid N-hydroxysuccinimide ester or 82 ml of freshly prepared digoxigenin-3-O-succinyl-ε-aminocaproic acid-N-hydroxysuccinimide ester (DIG-NHS) (Boehringer, Mannheim, Germany) (2 mg per 100 ml of dimethyl sulfoxide) was slowly added to 1 ml of cells, and the mixture was incubated on ice for 30 min with gentle shaking. Then 10 ml of 0.05 M Tris in PBS was added, and the cells were washed twice with PBS. After resuspension (approximately 4 3 106 cells per ml) in lysis buffer (1.5% octyl glucoside and the protease inhibitors phenylmethylsulfonyl fluoride [85 mg/ml], aprotinin [1 mg/ml], and leupeptin [2 mg/ml] in PBS) and incubation on ice for 1 h, the lysate was centrifuged at 400 3 g for 5 min at 4°C to remove nuclei and large cell debris. The supernatant was then centrifuged at 100,000 3 g for 20 min to 1 h at 4°C, and appropriate dilutions of the supernatant were used as cell membrane proteins for peptide scanning. For unlabeled cell membrane protein preparations, higher concentrations of washed cells were resuspended in lysis buffer (approximately 4 3 107 cells per ml), incubated, and centrifuged. After dilution of the sample and separation of the octyl glucoside by Sephadex G-25 (Pharmacia, Freiburg, Germany) gel filtration, a calculated dilution was used for specific blocking experiments (see below). Radioactive labeling of cell proteins. For [35S]methionine-[35S]cysteine labeling, 25-cm2 flasks with confluent FL or HeLa cell cultures were washed three times with PBS (pH 7.4) and 2 ml of methionine and cysteine-free medium and ca. 40 mCi of [35S]methionine-[35S]cysteine (specific activity, 1,175 Ci per mmol) (NEN, Du Pont) were added. After incubation overnight at 37°C, the cells were collected for membrane protein preparation as described above. Antigens for antiserum production. Virus, propagated on FL or HeLa cells, was purified by pretreatment with Freon or chloroform and two or three CsCl density gradient centrifugations (28). After expression in Escherichia coli with the pQE31 or pQE32 vector (Qiagen, Hilden, Germany) as recently reported (41), recombinant Ad3 fiber knob (rFK3) and Ad7 fiber knob, tagged with six histidine were purified by metal affinity chromatography. We used the native conditions as described by the manufacturer (Clontech, Heidelberg, Germany). The primers for the Ad7 fiber knob were GGACTTACATTCAATTC (forward) and TCAG TCGTCTTCTCTAA (reverse); the other primers are described by Pring-Åkerblom et al. (41). Natural Ad2 fiber was purified from infected HeLa cells by chromatography as described previously (36). Western blot analysis. The samples of rFK3 and recombinant Ad7 fiber knob (rFK7) were verified by a combined denaturing (boiled in 1% sodium dodecyl sulfate [SDS])-nondenaturing (unboiled samples) SDS-polyacrylamide gel electrophoresis (PAGE) followed by Western blot analysis by standard techniques (8, 26). Briefly, proteins were electrotransferred onto a nitrocellulose membrane (Bio-Rad, Munich, Germany), which was blocked in Super Block (Pierce, Rockford, Ill.) and incubated with mouse-anti-RGS six-histidine serum (Qiagen) for 1 h. The serum was diluted 1:1,000 to 1:2,000 with 10% Super Block in PBS. The reaction was detected with peroxidase-conjugated anti-mouse immunoglobulin (Amersham, Braunschweig, Germany) and the ECL chemiluminescence kit (Amersham). Antisera. Rabbit polyclonal sera directed against the Ad3 virion were produced by two immunizations of rabbits with purified virus (10 mg per injection) together with Freund’s complete and incomplete adjuvants. The anti-recombinant Ad3 fiber rabbit serum was a generous gift from J. Chroboczek, Grenoble, France. It was produced by using purified Ad3 fiber, expressed by recombinant baculovirus (1). Polyclonal antibodies against rFK3 were obtained by repeated rabbit immunization with rFK3. Anti-Ad2 fiber serum was prepared with purified Ad2 fiber. ELISA. The enzyme-linked immunosorbent assay (ELISA) was performed as previously described (28). Briefly, the wells of polystyrene microplates with high adsorption capacity (Greiner, Frickenhausen, Germany) were coated with approximately 50 ml of peptides in water (0.3 nmol per well) and incubated at 37°C overnight to dryness. After nonspecific blocking with 1% fetal calf serum or Super Block, duplicate samples of three or four serial dilutions of antisera specific for Ad3, recombinant Ad3 fiber, or rFK3 were tested with horseradishperoxidase-labeled anti-rabbit immunoglobulin G (Amersham) as the second antibody and o-phenylenediamine–H2O2 as a substrate. After the enzyme reac-
J. VIROL. tion was stopped with 0.85 M H2SO4, the optical density was measured at 492 nm. Preimmune serum dilutions were included as controls. The experiments were carried out in duplicate or triplicate. Binding of CM to peptides. When biotinylated CM were tested instead of antisera, only unbiotinylated peptides were preadsorbed and three serial dilutions (1:300 to 1:2,700) of the stock solution of labeled cell surface proteins were added to the wells, which were then incubated for 2 h at room temperature (RT) or overnight at 4°C. For dilution, blocking buffer in PBS (Super Block) containing 1 mM MgCl2, 1 mM MnCl2, and protease inhibitors as mentioned above or the protease inhibitor cocktail complete, EDTA free (Boehringer), was used. Horseradish peroxidase-labeled streptavidin (Amersham) was used with ophenylenediamine for the detection. To test the binding of DIG-labeled cell surface proteins to all the immobilized peptides, anti-DIG antibody from sheep (Fab fragments), conjugated with horseradish peroxidase (Boehringer), was used. [35S]methionine-[35S]cysteine-labeled cell proteins with 30,000 to 60,000 cpm per well were also incubated for 2 h at RT with preadsorbed peptides; then the first and second supernatants obtained after rinsing with 50 ml of PBS per well were pooled. The washing buffer used after this (two lots of 240 ml per well) was discarded. The activity of the unadsorbed proteins was measured with Rotiscint, eco plus (Fa. Roth, Karlsruhe, Germany) in a liquid scintillation counter. After incubation with 0.5 M NaOH and drying on filter paper, the peptide-bound activity was determined by counting in toluene-based scintillation counters. Analogous to the measurements with DIG-labeled CM, mixtures of labeled and unlabeled CM (more than the 50-fold concentration of labeled protein) were used in parallel, or the unlabeled CM were preincubated for 2 h at RT. The measured bound activity was in this case the value for the nonspecific binding (adsorption). The difference between the measured bound activities without and with unlabeled CM is the specific bound activity. However, only the specific bound activity was evaluated as positive, if the quotient of the values without and with blocking by unlabeled CM was $2. In a second step, the binding of labeled CM to selected peptides was measured after their preincubation with Ad3 virion for 1 h at RT. Focus reduction assay. A focus reduction assay in two modifications was conducted. The method was similar to one recently described (35). In brief, dilutions of the peptides were mixed with 50 ml of FL cells (300,000 cells/ml) in chamber slides for tissue culture (Lab Tek, Nunc). After incubation for 30 min at 37°C, 50 ml of Ad suspension was added. For the serum neutralization test, 100 ml of antiserum dilutions and 50 ml of Ad suspension were incubated for 2 h. After this, 50 ml of the respective cell suspensions were added. In both experiments, the virus doses were selected in such a way that the virus control was represented by approximately 40 FFU. Cultures were incubated for 48 h at 37°C. After drying, fixation, and staining with fluorescein isothiocyanate-labeled antibodies against Ad, the number of foci was counted with a fluorescence microscope.
RESULTS Characterization of the recombinant Ad3 fiber knob protein and the corresponding antibody preparation. Before pepscan was performed, it was necessary to characterize our recently produced recombinant Ad3 fiber knob antigen and the corresponding antiserum. The rFK3 preparation was purified by affinity chromatography and then analyzed by SDS-PAGE. As shown in Fig. 1A, a protein of 23 kDa was identified under denaturing conditions. The 23-kDa band represents the monomeric form of the fiber knob. This result agreed with the molecular mass predicted by the amino acid sequence of the fiber knob (including the RGS six-histidine epitope) (41). Under nondenaturing conditions, a component of approximately 53 kDa was obtained. Instead of the 53-kDa band, a 66- to 69kDa band was expected as a trimeric form. The Western blot displayed a reaction with the monomeric and the oligomeric rFK3 (Fig. 1B). The oligomeric protein was used as the immunogen for the preparation of polyclonal antibodies. The focus reduction assay was performed to determine the neutralizing capacity of the antisera directed against rFK3 and Ad3 virion, which were to be used in the pepscan. These antisera were compared with anti-recombinant Ad7 fiber knob serum and an antiserum against the natural Ad2 fiber. All antisera displayed neutralization titers of .2,000 (Fig. 2). Antigenic epitopes and receptor binding sites on the Ad3 fiber knob. Epitope mapping permits location of all the (linear) epitopes (43) in the FK3 protein. One method for epitope mapping based on ELISA involves the use of overlapping pep-
VOL. 72, 1998
FIG. 1. Analysis of purified recombinant fiber knob proteins. (A) SDSPAGE of the affinity-purified rFK3 and rFK7; the samples in 0.1% SDS were not boiled before electrophoresis (10% polyacrylamide gel). The gel was stained with Coomassie brilliant blue. Lanes: M, molecular size markers; 1 and 2, rFK3 samples 1 and 2; 3 and 4, rFK7 samples 1 and 2. (B) Immunoblot reaction of rFK3 with mouse-anti-RGS six-histidine serum (1:2,000) and detection by using with peroxidase-conjugated anti-mouse immunoglobulin and the ECL kit. Lanes: 1, monomer of rFK3 (23 kDa) (boiled before SDS-PAGE in 1% SDS sample buffer); 2, oligomer of rFK3 (53 kDa) (without boiling). The sizes of the markers are indicated in kilodaltons.
tides, spanning the entire amino acid sequence of FK3, and antisera directed to this protein. The domain of a viral surface protein which interacts with a cellular receptor may be defined as a receptor binding site. To identify such sites, pepscan could be applied. We extended this method, normally used for the epitope mapping, to measure the binding of labeled cell surface proteins to overlapping peptides derived from the FK3 protein. By this method, it should be possible to obtain a map of possible receptor binding sites. We preferred the use of labeled CM preparations to the use of labeled cells, because there may be more steric hindrance in the binding of whole cells than of membrane protein to small peptides, bound on a solid phase.
RECEPTOR BINDING SITES ON ADENOVIRUS FIBER
9123
Originally, 186 overlapping 9-mer peptides of the FK3 protein, bound to pins, were evaluated for their reaction with antibodies. However, the values of optical density (extinction) obtained with the antisera were not higher than twice the values of the preimmune sera as controls (data not shown). Therefore, free 13- to 20-mer overlapping peptides also spanning the entire amino acid sequence of FK3 were used for further analyses. At first, peptides P6, P8, P11, and P13 to P16 (Table 1) were evaluated for the reactivity with antisera and cell membrane proteins. Peptides P6, P11, and P14 were selected because they span the D, G, and I b-strands (see Fig. 6). These regions should participate in the binding of the cell receptor(s) (53, 54). The antibodies specific to recombinant Ad3 fiber and Ad3 virion reacted with the seven peptides in the ELISA (data not shown). The peptides were then tested with labeled cell proteins. Two kinds of labeling of the proteins were performed, biotinylation and metabolic labeling with [35S]methionine-[35S] cysteine. Peptides P6, P11, and P14 displayed a stronger reactivity with biotinylated FL cell surface proteins than the others did (Fig. 3A). The reactivity pattern was very similar when [35S] methionine-[35S]cysteine-labeled FL cell proteins were used (Fig. 3B). After these promising results, experiments were extended to include all further peptides derived from the Ad3 fiber knob (Table 1). Map of antigenic epitopes of the Ad3 fiber knob. All the overlapping peptides (Table 1) were immobilized on microplates and the reactivity with the antisera specific to recombinant Ad3 fiber, recombinant Ad3 fiber knob, and Ad3 virion were evaluated in ELISAs. As shown in Fig. 4, the antiserum directed against the recombinant Ad3 fiber protein reacted strongly with peptides P3, P6, P12, P14, and P16. These peptides contain main epitopes. The binding of anti-Ad3 virion antibodies was generally low. This effect was expected, because the 12 fibers on the virion represent only ca. 1/100 of the mass of the virion protein, and consequently the induced antibody level can be very low. However, the reactivity patterns with anti-Ad3 virion and anti-recombinant Ad3 fiber sera were similar. Surprisingly, P16, which showed the strongest reactivity with the recombinant Ad3 fiber-specific serum, did not react with antibodies directed to the Ad3 virion. Taking into account the generally low values with anti-Ad3 virion antibodies, the data
TABLE 1. Synthesized peptides of the Ad3 fiber knob and their purity
FIG. 2. Neutralization assay. Focus reduction by antibodies against rFK3 is shown in comparison with rFK7-, Ad2 fiber (F2)-, and Ad3 virion-specific sera and preimmune serum, tested with the respective homologous virus. Standard deviations SD were within 6 to 12%. The assay was performed as described in Material and Methods.
Peptide (positions)
Sequence
% Puritya
P1 (aa 131–145) P2 (aa 142–156) P3 (aa 153–167) P4 (aa 164–178) P5 (aa 175–189) P6 (aa 187–200) P7 (aa 197–213) P8 (aa 210–225) P9 (aa 221–235) P10 (aa 233–245) P11 (aa 244–256) P12 (aa 254–269) P13 (aa 267–283) P14 (aa 281–294) P15 (aa K-291–306) P16 (aa 300–319)
NTLWTGPKPEANCII NCIIEYGKQNPDSKL DSKLTLILVKNGGIV GGIVNGYVTLMGASD GASDYVNTLFKNKNV KNVSINVELYFDAT FDATGHILPDSSSLKTD LKTDLELKYKQTADFS TADFSARGFMPSTTA TTAYPFVLPNAGT G T H N E N Y I F G NbC Y QCYYKASDGALFPLEV LEVTVMLNKRLPDSRTS RTSYVMTFLWSLNA KSLNAGLAPETTQATLI TTQATLITSPFTFSYIREDD
50 80 55 90 80 .90 55 64 50 60 .90 60 63 .90 64 87
a
The purity was determined by HPLC. This residue corresponds to the sequence published by Xi et al. (53) and is Q in reference 46. b
9124
LIEBERMANN ET AL.
FIG. 3. Reactivity of selected peptides derived from the Ad3 fiber knob with labeled FL CM. (A) Binding of biotinylated FL cell surface proteins to the immobilized peptides (0.25 mg of 15-mer peptide per well). Horseradish peroxidase-labeled streptavidin was used together with o-phenylenediamine for the detection, as described in Materials and Methods. OD, optical density. (B) Bound activity of [35S]methionine-[35S]cysteine-labeled FL cell proteins. Radioactively labeled proteins were added at 50,000 cpm per well. Each datum point is the average for four wells.
indicate that there may be a minor difference in the conformations of the recombinant Ad3 fiber and the natural viral counterpart used for antisera production. The reaction pattern with rFK3-specific antibodies was different from the others, especially for P15. This serum reacted only with P10, P12, P15, and P16, suggesting that the conformation of the rFK3, expressed in E. coli is not identical to that of the natural FK3 or to that of the knob of the Ad3 fiber protein, expressed by recombinant baculovirus. The biological variability of the rabbits used for antisera production could also have influenced the reactivity pattern. Subsequently, we compared the antigenicities of the peptides determined in the ELISA by using antibodies against recombinant Ad3 fiber and the predicted secondary FK3 protein structure overlaid with the values for the antigenic index of $1.2 (5, 21). They corresponded very well (Fig. 4B), indicating
J. VIROL.
that the obtained map of antigenic epitopes is probably correct. The different purities of the peptides did not significantly change the results (data not shown). Considering the different purities of the peptides in these ELISAs, the quantities of the peptides per well were changed as follows: for example, 50 ml of a 15-mer peptide (10 mg/ml) with 90% purity and 90 ml of a 15-mer peptide with 50% purity were added to the wells and dried. Map of receptor binding sites. Analogous to the mapping of antigenic epitopes, the binding of labeled FL or HeLa CM to the immobilized peptides, derived from the Ad3 fiber knob, was evaluated. For the determination of the specific binding of viruses to cells, the attachments of the labeled virus were usually specifically blocked by unlabeled virus in great excess (7, 28, 29). Analogously, the binding of labeled CM should be blocked by the unlabeled membrane proteins. The specific bound activity was determined as described in the legend to Fig. 5A. The strong binding of DIG-labeled HeLa CM to peptides P2, P3, P6, P11, P14, and P16 was significantly reduced by using unlabeled HeLa-CM or FL-CM. To distinguish whether the binding of the CM to these peptides was Ad3 virion specific, labeled CM were preincubated with Ad3 virion. An inhibitory effect by Ad3 virion at a concentration of 0.1 mg/ml could be detected for P2, P3, and P6. However, Ad3 virion at 0.7 mg/ml completely prevented the binding of CM to peptides P2, P3, P6, P14, and P16 and partially prevented binding to P11 (Fig. 5B). The blocking effect by Ad3 virion means that the binding of the cell surface proteins to these peptides was Ad3 virion specific. Considering the different purity of the FK3 peptides, the reactivity pattern was essentially the same (data not shown). A comparison of the results with [35S]methionine-[35S]cysteine cell proteins and DIG-labeled CM is illustrated in Fig. 5C. Only for P2 and P11 were the levels different. One reason could be that during radioactive labeling, many other proteins besides the cell surface proteins were labeled. Summarizing the results with DIG-labeled, radioactively labeled, and biotinylated CM, P6, to a lesser extent P14, and probably P11 interacted specifically with CM, indicating that these peptides present cell receptor binding sites of the Ad3 fiber knob. Moreover, the reaction of the cell surface proteins with P2 and P3 was relatively strong and that with P16 was (very) moderate. These sequence motifs are in the V-sheet of the knob, which faces the virus (53, 54) (see Fig. 7). Therefore, P2, P3, and P16 probably cannot present real receptor binding sites. It was previously reported by Di Guilmi et al. (11) that in the presence of 10 mM EDTA no virus or fiber binding to membrane proteins was observed in an overlay binding assay. For this reason, a possible influence of EDTA on the binding of CM to the FK3 peptides was also investigated. In our experiments, 10 mM EDTA did not prevent the binding (data not shown). Comparing the reactivity pattern with antisera and labeled CM, the antibodies and also FL CM or HeLa CM clearly bound to peptides P3, P6, and P14, whereas P11 reacted only with CM and P16 reacted strongly only with antiserum. The peptides presenting possible receptor binding sites and the main antigenic epitopes of the FK3 region are displayed in Fig. 6. Peptides P6, P11, and P14 span the D, G, and I b-strands, respectively. Surprisingly P13 (aa 267 to 283), which contains nearly the entire H strand, reacted only weakly both with labeled CM and with antibodies to recombinant Ad3 fiber and Ad3 virion. Inhibition of the Ad3 infection of cells by fiber knob peptides. The fact that peptides of structural proteins of other viruses can block the virus attachment to cells and also inhibit the infection (16, 27, 55) led to the question whether selected peptide sequence motifs of the Ad3 fiber knob protein may
VOL. 72, 1998
RECEPTOR BINDING SITES ON ADENOVIRUS FIBER
9125
FIG. 4. Antigenic epitopes on FK3. (A) Reaction of anti-rFK3 antibody (1/600), anti-recombinant Ad3 fiber (rF3) (1/900), and anti-Ad3 virion (1/100) sera in the ELISA with overlapping FK3 peptides P1 to P16 immobilized on microplates (0.5 mg of a 15-mer peptide per well). OD, optical density. Preimmune sera were used at dilutions of 1/100, 1/600 and 1/900. The results shown are an average of three separate experiments. Only optical densities larger than twice the value of the preimmune serum are shown. (B) Plot structure of FK3. The prediction (5, 21) for the secondary protein structure overlaid with the values for the antigenic index of $1.2 is shown. The positions of the antigenic peptides are displayed. The sizes of the letters (of the peptides) correlate with the reactivity. The knob domain starts at position 134.
also be able to inhibit the infection. To determine the influence of FK3 peptides on Ad infection of cells, a focus assay was performed. First, rFK3 and rFK7 were tested (Fig. 7A). Both inhibited only the infectivity of the homologous virus. When concentrations of fiber protein of $10 mg/ml were used, toxic effects were observed. The inhibitory effect of the rFK3 was
low. The quality of the rFK7 preparation was obviously higher than that of the rFK3 preparation. In the following experiments, the inhibitory activity of the FK3 peptides P6, P11, and P14, which displayed a higher binding of cell surface proteins and also (except for P11) contained major epitopes, was studied. In addition, P16 was included in the assay, since it showed
9126
LIEBERMANN ET AL.
J. VIROL.
FIG. 5. Receptor binding sites on FK3. (A) Binding of DIG-labeled HeLa CM (HeLa-CM) to overlapping FK3 immobilized on ELISA plates. The experiments were conducted as described in Materials and Methods. The optical density (OD) of the bound labeled proteins was measured without and with unlabeled HeLa CM (.50 fold concentration as labeled). The difference in the values is the peptide-specific binding of HeLa CM. However, OD values for the specific binding (specif. bind.) are shown, which were higher than twice the values with unlabeled CM. These data are averages of two experiments with two values each. The variation coefficients were within 7 to 15%. The different purity of the peptides was considered. (B) Binding of DIG-labeled HeLa cell surface proteins (DIG HeLa-CM) to FK3 peptides without and after preincubation of DIG HeLa CM with Ad3 virions. Two virion concentrations, c1 (0.1 mg/ml) and c2 (0.7 mg/ml) were used. The results are averages of two experiments with two values each. The different purities of the peptides were considered. (C) Reaction of [35S]methionine-[35S]cysteine-labeled FL CM (35S FL-CM) with FK3 peptides in comparison with DIG-labeled HeLa CM (DIG HeLa-CM). For this, the maximum optical density and the maximum bound radioactivity obtained in these experiments were set as 100% (at P3) and the values for the other peptides were displayed as a percentage of the maximum. The displayed values of the specific binding of DIG-labeled HeLa CM to the peptides (equimolar without consideration of the purity) results from three experiments (with two and four values each, respectively). The displayed reactivities of the immobilized peptides with radioactively labeled FL CM are means of four values each.
VOL. 72, 1998
RECEPTOR BINDING SITES ON ADENOVIRUS FIBER
9127
FIG. 5—Continued.
the highest reactivity with anti-recombinant Ad3 fiber serum. As a comparison, P13 with a lower binding capacity was included (Fig. 7B). Only P14 (aa 281 to 294) and P16 (aa 300 to 319) were able to inhibit the infection at high concentrations in comparison to the whole fiber knob. When 100 mg of peptides P13 and P16 per ml was used, toxic effects were observed. As shown, much higher concentrations of peptides than of fiber protein were necessary to inhibit the infection. For 50% reduction of the focus number by P14, approximately 30 mg of peptide per ml was necessary, in contrast to approximately 3 mg of rFK3 per ml and 0.3 mg of rFK7 per ml. This is 130- to 1,300-fold higher on a molar basis. DISCUSSION Receptor binding sites. To obtain information about the regions of the Ad fiber knob which participate in the binding of the Ad to the primary receptor(s) of permissive cells, we first investigated the interaction of labeled cell surface protein and overlapping chemically synthesized peptides of the fiber knob of Ad3, a member of subgenus B:1. This approach seems to be applicable for determining receptor binding sites. We found that three peptides, P6 (aa 187 to 200), P11 (aa 244 to 256), and P14 (aa 281 to 294), displayed cell membrane protein binding capacity. These peptides span the D, G, and I bstrands, respectively (Fig. 6). This corresponds to the prediction of Xia et al. (53) that these regions could take part in binding of the cell receptor(s). In a sequence alignment of the Ad3 and Ad5 fiber knobs (data not shown), our peptide P6 of FK3 (Fig. 5 and 6) corresponded fully to the C-terminal part of the core motif (aa 445 to 462) of the Ad5 fiber knob, which binds to the major histocompatibility complex class I alpha 2 domain at the surface of human epithelial and B-lymphoblastoid cells (20). This suggests that the probability of the participation of a sequence motif of the FK3 D b-strand is high. In the structural model of the trimeric knob, the b-strand G is slightly hidden in a hollow (53, 54). This may make it difficult
FIG. 6. Primary sequence of FK3 and sequence motifs containing possible receptor binding sites (underlined) and antigenic main epitopes (overlined). The ranges of D*, G*, H*, and I* indicate the b-strands of the receptor sheet (53). (Stars were added for differentiation from the b-strands of the V-sheet). The ranges of A, B, C, and J form the other b-sheet, which is partially buried in the trimeric structure and named the virus (V)-sheet. The small b-strands E and F are parts of the D-G loop. A high intensity in the reaction is indicated by double lines, and a more moderate reaction is indicated by full, hatched, or dotted lines.
9128
LIEBERMANN ET AL.
FIG. 7. Influence of recombinant fiber knobs and peptides on Ad infection of cells. (A) Reduction of the Ad3 and Ad7 infection of FL cells by recombinant rFK3 and rFK7, respectively. rFK3 and rFK7 inhibited the infection of the homologous virus (rFK3/Ad3 and rFK7/Ad7) but not that of the heterologous virus (rFK7/Ad3 and rFK3/Ad7). The recombinant proteins were used after native purification. (B) Influence of peptides of FK3 on infection by Ad3. The focus reduction assay was performed as described in Material and Methods. Results are average of three or two separate experiments; the standard deviation is indicated for P14 and P16, which are able to inhibit the infection.
for the cell receptor to reach and could mean that not only sequence motifs of peptide P2 and P3 of the V-sheet of the Ad3 fiber knob but also sequence motifs of P11 (Fig. 5 and 6) cannot participate in the receptor binding. On the other hand, the receptor binding site of rhinoviruses forms a canyon reachable by the receptor, intercellular cell adhesion molecule 1 (ICAM-1) (18). As mentioned above, we preferred the use of labeled CM preparations to labeled cells in the modified pepscan, because there may be more steric hindrance in the binding of whole cells than of membrane protein to small peptides, bound on a solid phase. Also, it is easier to handle CM than whole living cells in the procedures described. After our experience, it was necessary to use only fresh preparations of CM. We use now sulfosuccinimidyl biotin (NHSS-biotin) (Pierce) instead of NHS-biotin, because the sulfo group prevents the passage of the reagent through the cell membrane. Our methodological variant, using DIG-labeled instead of biotinylated cell surface proteins, seems to be nearly optimal, because we can use streptavidin-coated microplates for biotinylated pep-
J. VIROL.
tides. Another advantage in comparison to radioactive labeling is that only the cell surface proteins and not all the other cell proteins were labeled when we used NHSS-DIG. Our result that 10 mM EDTA did not prevent the binding of CM to FK3 means that divalent cations are obviously not necessary for the binding of CM to the immobilized peptides. This must not contradict the earlier finding that in the presence of 10 mM EDTA no virus or fiber binding to membrane proteins was observed in an overlay binding assay (11). Di Guilmi et al. (11) used the trimeric fiber protein, but we used the monomeric peptides. Epitopes. In previous studies, antisera against FK3 peptides (aa 210 to 225, aa 267 to 283, and aa K-291 to 306) showed neutralization titers of 32 and 16, indicating that those peptides may span minor neutralizing epitopes (28). These immunogenicity results correspond to the antigenicity determined. Using recombinant fiber proteins, eight synthesized peptides, and the corresponding antisera, Mei and Wadell (34) were able to evaluate antigenic and immunogenic epitopes, together with the hemagglutination binding domain on Ad fibers of subgenus B:2. They found that an epitope with low-level immunogenicity is present in the C-terminal 20 aa of the fiber. This corresponds to our previous results describing low immunogenicity for the highly antigenic C-terminal peptide P16 (aa 300 to 319) of the Ad3 fiber knob (28), which is very similar in sequence. Mei and Wadell (34) mentioned that rabbit antisera against Ad11p, Ad34a, and Ad35p virions were also analyzed to detect epitopes and, unfortunately, displayed extremely low levels of reactivity. For this reason, we optimized the conditions for using virion-specific sera, and our anti-Ad3 serum confirmed the results with anti-rF3 serum, with the exception of P16 (aa 300 to 319). Until now, results have been published only concerning epitope mapping of the Ad2 fiber knob, not of the Ad3 fiber knob (15). A comparison of the amino acid sequences of both virus types revealed that the epitopes found by Fender et al. (15) correspond well to the epitopes of the Ad3 fiber knob (data not shown). Since we used 13- to 20-mer peptides overlapping by only 4 aa residues, the resolution of the epitope location was low. However, the early experiments involving the pin-bound 9-mer peptides of Ad3 fiber knob with an offset of 1 aa gave disappointing results. The general disadvantage of such pin- or membrane-bound peptides is that the purity of the peptide preparation cannot be determined. Ad2 and Ad5 use the same primary receptor and therefore may have a very similar receptor binding site(s) on the fiber knob. Recently, Hong et al. (20) found a corresponding knob region of Ad5 between Val 438 and Asp 462 with a core motif (aa 445 to 462) for the neutralizing monoclonal antibody 1D 6.3. Our initial results show that the corresponding peptide (aa 449 to 463) of the Ad2 fiber knob contains a main epitope and a receptor binding site (unpublished data). Focus reduction by peptides. Previous studies have reported only on the inhibition of infection by fiber proteins (7, 11, 19). Approximately 0.1 mg of the recombinant knob domain of Ad5 per ml reduced the Ad5 infection of human cells by 50% (19). In our experiments, approximately 0.3 mg of recombinant Ad7 fiber knob per ml reduced the focus number in HeLa cells by 50% (Fig. 7A). It was interesting to evaluate peptides in this field. We found that two of the FK3 peptides may inhibit the infection (Fig. 7B). However, approximately 30 mg of P14 (aa 281 to 294) per ml (14 mM) was necessary to obtain 50% inhibition of infection. A possible explanation could be that in contrast to the peptides, the fiber proteins used were trimeric. An additional reason could be the incorrect conformation of
VOL. 72, 1998
RECEPTOR BINDING SITES ON ADENOVIRUS FIBER
these peptides compared with the conformation of these sequence motifs in the whole trimeric fiber knob protein. As mentioned by Porta and Lomonossoff (40), free peptides can adopt a variety of conformations, most of which will not exist in their native environment. In comparison, about 20 mM RGD-containing peptide of VP1 of foot-and-mouth disease virus and approximately 5 mM selected peptides of the fusion protein of human parainfluenza virus were needed for 50% plaque reduction (27, 55). It is also interesting that to protect cattle against foot-and-mouth disease virus, a 1,000-fold more immunogenic tandem peptide of VP1 in comparison to the inactivated virus was necessary (12). A common reason for these findings is probably the different conformations of free peptides and the corresponding sequence motifs in the viral proteins. Summarizing, we conclude that the epitope mapping serotypespecific epitopes should be determined. The evaluated receptor binding sites also give information that will be useful for the construction of new vehicles for the gene therapy. First, the critical part of each binding site should be determined. We are currently producing antisera against peptides P6, P12 and P14, presenting probable receptor binding sites and/or strong(er) antigenic determinants to evaluate the blocking effect on the virus attachment or the neutralization. The determination of receptor binding sites and antigenic epitopes on the fiber knobs of Ad2 and Ad5 of subgenus C and of Ad8 and Ad15 of subgenus D is under way. Future studies will also investigate whether stimulated lymphocytes (36) use the same receptor binding sites and receptor molecules as HeLa or FL cells. ACKNOWLEDGMENTS We thank J. Chroboczek, Institut de Biologie Structurale, Grenoble, France, for the gift of anti-recombinant Ad3 fiber serum; M. Liebermann and M. Wolfgramm for excellent technical assistance; F. Kernchen, Biosyntan Berlin-Buch, for the synthesis of pin-bound and biotinylated peptides; U. Wegner for help in the neutralization test; and K. Lotz for help in the last pepscans. This work was supported by grants from the Ministerium fu ¨r Kultur, Schwerin (EMAU/12/94), and the Deutsche Forschungsgemeinschaft (Se700/1-3).
11. 12. 13. 14.
15. 16. 17. 18.
19. 20. 21. 22. 23. 24. 25.
26. 27.
REFERENCES 1. Albiges-Rizo, C., A. Barge, R. W. H. Ruigrok, P. A. Timmins, and J. Chroboczek. 1991. Human adenovirus serotype 3 fibre protein: comparison of native and recombinant proteins. J. Biol. Chem. 266:3961–3967. 2. Bai, M., B. Harfe, and P. Freimuth. 1993. Mutations that alter an Arg-GlyAsp (RGD) sequence in the adenovirus type 2 penton base protein abolish its cell-rounding activity and delay virus reproduction in flat cells. J. Virol. 67:5198–5205. 3. Bergelson, J. M., J. A. Cunningham, G. Droguett, E. A. Kurt-Jones, A. Krithivas, J. Hong, M. S. Horwitz, R. L. Crowell, and R. W. Finberg. 1997. Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science 275:1320–1323. 4. Bichko, V. V., S. M. Lemon, J.-G. Wang, S. Hwang, M. M. C. Lai, and J. M. Taylor. 1996. Epitopes exposed on hepatitis delta virus nucleoproteins. J. Virol. 70:5807–5811. 5. Chou, P. Y., and G. D. Fasman. 1978. Empirical predictions of protein conformation. Annu. Rev. Biochem. 47:251–276. 6. Chroboczek, J., R. W. H. Ruigrok, and S. Cusack. 1995. Adenovirus fiber, p. 39–46. In W. Doerfler and P. Boehm (ed.), The molecular repertoire of adenoviruses I. Springer-Verlag KG, Berlin, Germany. 7. Defer, C., M.-T. Belin, M.-L. Caillet-Boudin, and P. Boulanger. 1990. Human adenovirus-host cell interactions: comparative studies with members of subgroups B and C. J. Virol. 64:3661–3673. 8. Dernick, R., and J. Heukeshoven. 1989. Proteins and peptides, p. 261–347. In A. Mayr, P. A. Bachmann, B. Mayr-Bibrack and G. Wittmann (ed.), Virological methods, vol. III, Fischer, Jena, Germany. (In German.) 9. Devaux, C., M.-L. Caillet-Boudin, B. Jacrot, and P. Boulanger. 1987. Crystallization, enzymatic cleavage and the polarity of the adenovirus type 2 fiber. Virology 161:121–128. 10. De Verdugo, U. R., H.-C. Selinka, M. Huber, B. Kramer, J. Kellermann, P. H. Hofschneider, and R. Kandolf. 1995. Characterization of a 100-kilo-
28.
29.
30. 31. 32.
33. 34. 35. 36. 37.
9129
dalton binding protein for the six serotypes of coxsackie B viruses. J. Virol. 69:6751–6757. Di Guilmi, A. M., A. Barge, P. Kitts, E. Gout, and J. Chroboczek. 1995. Human adenovirus serotype 3 (Ad3) and the Ad3 fiber protein bind to a 130-kDa membrane protein on HeLa cells. Virus Res. 38:71–81. Di Marchi, R., G. Brooke, C. Gale, V. Cracknell, T. Doel, and N. Mowat. 1986. Protection of cattle against foot-and mouth disease by a synthetic peptide. Science 232:639–641. Eiz, B., and P. Pring-Akerblom. 1997. Molecular characterization of the type-specific g-determinant located on the adenovirus fiber. J. Virol. 71: 6576–6581. Estepa, A., and J. M. Coll. 1996. Pepscan mapping and fusion-related properties of the major phosphatidylserine-binding domain of the glycoprotein of viral hemorrhagic septicemia virus, a salmonid rhabdovirus. Virology 216: 60–70. ¨ berg, E. Drouet, and J. Chroboczek. Fender, P., A. H. Kidd, R. Brebant, M. O 1995. Antigenic sites on the receptor-binding domain of human adenovirus type 2 fiber. Virology 214:110–117. Fox, G., N. Parry, P. Barnett, B. McGinn, D. Rowlands, and F. Brown. 1989. The cell attachment site on foot-and-mouth disease virus includes the amino acid sequence RGD (arginine-glycine-aspartic acid). J. Gen. Virol. 70:1–12. Geysen, H. M., R. H. Meloen, and S. J. Barteling. 1984. Use of peptide synthesis to probe viral antigens for epitopes to a resolution of a single amino acid. Proc. Natl. Acad. Sci. USA 81:3998–4002. Greve, J. M., and M. G. Rossmann. 1994. Interaction of rhinovirus with its receptor, ICAM-1, p. 195–213. In E. Wimmer (ed.), Cellular receptors for animal viruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. Henry, L. J., D. Xia, M. E. Wilke, J. Deisenhofer, and R. D. Gerard. 1994. Characterization of the knob domain of the adenovirus type 5 fiber protein expressed in Escherichia coli. J. Virol. 68:5239–5246. Hong, S. S., L. Karayan, J. Tournier, D. T. Curiel, and P. A. Boulanger. 1997. Adenovirus type 5 fiber knob binds to MHC class I a2 domain at the surface of human epithelial and B lymphoblastoid cells. EMBO J. 16:2294–2306. Jameson, B. A., and H. Wolf. 1988. The antigenic index: a novel algorithm for predicting antigenic determinants. Comput. Appl. Biosci. 4:181–186. Jolly, D. 1994. Viral vector systems for gene therapy. Cancer Gene Ther. 1: 51–64. Krah, D. L. 1989. Characterization of octyl glucoside-solubilized cell membrane receptors for binding measles virus. Virology 172:386–390. Krasnykh, V. N., G. V. Mikheeva, J. T. Douglas, and D. T. Curiel. 1996. Generation of recombinant adenovirus vectors with modified fibers for altering viral tropism. J. Virol. 70:6839–6846. Krasnykh, V. N., I. Dimitriev, G. Mikheeva, C. R. Miller, N. Belousova, and D. T. Curiel. 1998. Characterization of an adenovirus vector containing a heterologous peptide epitope in the HI loop of the fiber knob. J. Virol. 72: 1844–1852. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685. Liebermann, H., R. Do¨lling, D. Schmidt, and G. Thalmann. 1991. RGDcontaining peptides of VP1 of foot-and-mouth disease virus prevent virus infection in vitro. Acta Virol. 35:90–93. Liebermann, H., R. Mentel, L. Do¨hner, S. Modrow, and W. Seidel. 1996. Inhibition of cell adhesion to the virus by synthetic peptides of fiber knob of human adenovirus serotypes 2 and 3 and virus neutralization by anti-peptide antibodies. Virus Res. 45:111–121. Liebermann, H., U. Bauer, P. Pring-Åkerblom, M. Derbas, R. Mentel, U. Wegner, and W. Seidel. 1996. Interaction of recombinant fiber protein of human adenoviruses serotypes Ad2, Ad3, Ad9 and Ad15 with receptors of permissive cells, abstr. PW33/84, p. 202. In Abstracts of the Xth International Congress of Virology. Louis, N., P. Fender, A. Barge, P. Kitts, and J. Chroboczek. 1994. Cellbinding domain of adenovirus serotype 2 fiber. J. Virol. 68:4104–4106. Mathias, P., T. Wickham, M. Moore, and G. Nemerow. 1994. Multiple adenovirus serotypes use av integrins for infection. J. Virol. 68:6811–6814. Martens, W., I. Greiser-Wilke, T. C. Harder, K. Dittmar, R. Frank, C. ¨ rvell, V. Moennig, and B. Liess. 1995. Spot synthesis of overlapping pepO tides on paper membrane supports enables the identification of linear monoclonal antibody binding determinants on morbillivirus phosphoproteins. Vet. Microbiol. 44:289–298. Mayr, G., and P. Freimuth. 1997. A single locus on human chromosome 21 directs the expression of a receptor for adenovirus type 2 in mouse A9 cells. J. Virol. 71:421–418. Mei, Y.-F., and G. Wadell. 1996. Epitopes and hemagglutination binding domain on subgenus B:2 adenovirus fibers. J. Virol. 70:3688–3697. Mentel, R., E. Matthes, M. Janta-Lipinski, and U. Wegner. 1996. Fluorescent focus reduction assay for the screening of anti-adenoviral agents. J. Virol. Methods 59:99–104. Mentel, R., G. Do¨pping, U. Wegner, W. Seidel, H. Liebermann, and L. Do ¨hner. 1997. Adenovirus-receptor interaction with human lymphocytes. J. Med. Virol. 51:252–257. Modrow, S., and H. Wolf. 1990. Use of synthetic peptides as diagnostic
9130
38. 39. 40. 41. 42. 43. 44. 45. 46.
LIEBERMANN ET AL.
reagents in virology, p. 83–101. In M. H. V. van Regenmortel and A. R. Neurath (ed.), Immunochemistry of viruses II. The bases for serodiagnosis and vaccines. Elsevier Biomedical Press, Amsterdam, The Netherlands. Persson, R., C. Wohlfart, U. Svensson, and E. Everitt. 1985. Virus receptor interaction in the adenovirus system. Characterization of the positive cooperative binding of virions on HeLa cells. J. Virol. 54:92–97. Philipson, L., K. Lonberg-Holm, and U. Pettersson. 1968. Virus-receptor interaction in an adenovirus system. J. Virol. 2:1064–1075. Porta, C., and G. P. Lomonossoff. 1998. Scope for using plant viruses to present epitopes from animal pathogens. Rev. Med. Virol. 8:25–41. Pring-Åkerblom, P., A. Heim, and F. E. J. Trijssenaar. 1998. Molecular characterization of hemagglutination domains on the fibers of subgenus D adenoviruses. J. Virol. 72:2297–2304. Reineke, U., R. Sabat, D. Volk, and J. Schneider-Mergener. 1996. Mapping discontinuous epitopes using cellulose-bound peptide libraries. Immunobiology 196:96. Rodda, S. J., N. J. Maeji, and G. Tribbick. 1996. Epitope mapping using multipin peptide synthesis, p. 137–147. In G. E. Morris (ed.), Epitope mapping protocols. Humana Press Inc., Totowa, N.J. Ruigrok, R. W. H., A. Barge, C. Albiges-Rizo, and S. Dayan. 1990. Structure of adenovirus fibre. II. Morphology of single fibres. J. Mol. Biol. 215:589– 596. Shenk, T. 1996. Adenoviridae: the viruses and their replication, p. 2111– 2148. In B. N. Fields, D. M. Knipe, and P. M. Howley (ed.), Virology, vol. 2. Lippincott-Raven Publishers, Philadelphia, Pa. Signa ¨s, C., G. Akusja ¨rvi, and U. Pettersson. 1985. Adenovirus 3 fiber polypeptide gene: implications for the structure of the fiber protein. J. Virol. 53:672–678.
J. VIROL. 47. Spencer, J. V., B. L. Trus, F. P. Booy, A. C. Steven, W. W. Newcomb, and J. C. Brown. 1997. Structure of the herpes simplex virus capsid: peptide A862H880 of the major capsid protein is displayed on the rim of the capsomer protrusions. Virology 228:229–235. 48. Stevenson, S. C., M. Rollence, B. White, L. Weaver, and A. McClelland. 1995. Human adenovirus serotype 3 and 5 bind to two different cellular receptors via the fiber head domain. J. Virol. 69:2850–2857. 49. Tomko, R. P., R. Xu, and L. Philipson. 1997. HCAR and MCAR: the human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proc. Natl. Acad. Sci. USA 94:3352–3356. 50. Van Oostrum, J., and R. M. Burnett. 1985. Molecular composition of the adenovirus type 2 virion. J. Virol. 56:439–448. 51. Wadell, G., W. Doerfler, W. C. Russel, and G. R. Kitchingman. 1994. Adenoviruses, p. 1–23. In R. G. Webster and A. Granoff (ed.), Encyclopedia of virology, vol. 2. Academic Press, Ltd., London, United Kingdom. 52. Weber, J. M., B. G. Talbot, and L. Delorme. 1989. The orientation of the adenovirus fiber and its anchor domain. Identified through molecular mimicry. Virology 168:180–182. 53. Xia, D., L. J. Henry, R. D. Gerard, and J. Deisenhofer. 1994. Crystal structure of the receptor-binding domain of adenovirus type 5 fiber protein at 1.7 A resolution. Structure 2:1259–1270. 54. Xia, D., L. J. Henry, R. D. Gerard, and J. Deisenhofer. 1995. Structure of the receptor binding domain of adenovirus type 5 fiber protein, p. 39–46. In W. Doerfler and P. Boehm (ed.), The molecular repertoire of adenoviruses I. Springer-Verlag KG, Berlin, Germany. 55. Yao, Q., and R. W. Compans. 1996. Peptides corresponding to the heptad repeat sequence of human parainfluenca virus fusion protein are potent inhibitors of virus infection. Virology 223:103–112.