Peptide Mimotopes Selected from a Random Peptide Library for Diagnosis of Epstein-Barr Virus Infection
http://www.100md.com
微生物临床杂志 2006年第3期
Cooperative Research Centre for Diagnostics
Department of Biochemistry, La Trobe University, VIC 3086, Australia
Queensland Medical Laboratory, P.O. Box 5410, West End, QLD 4101, Australia
ABSTRACT
Epstein-Barr virus (EBV) is a ubiquitous, worldwide infectious agent that causes infectious mononucleosis, affecting >90% of the world's population. Currently, enzyme-linked immunosorbent assay, mostly with purified preparations of EBV cell extracts to capture immunoglobulin M (IgM) antibodies in patients' serum, is used for primary diagnosis. Our objective was to determine whether a small set of peptides could contain sufficient immunogenic information to replace solid-phase antigens in EBV diagnostics. Using monoclonal antibodies, we selected four peptides that mimic different epitopes of EBV from a phage-displayed random peptide library. To assess their diagnostic value, we screened a panel of 62 individual EBV IgM sera for their reactivities with the peptides alone. For all peptides, there was a clear distinction between the EBV-positive and the EBV-negative samples, resulting in 100% specificity. The sensitivities were 88%, 85%, 71%, and 54% for peptides F1, A3, gp125, and A2, respectively. Any combination of peptides increased the sensitivity, indicating that individual peptides react with different subsets of antibodies. Furthermore, when the F1 and the gp125 peptides were coupled to bovine serum albumin and screened against 216 serum samples, there were dramatic improvements in sensitivities (95% and 92%, respectively) and little cross-reactivity with the other peptides encountered during acute viral infections, including rheumatoid factor. This study shows the potential for the use of peptide mimotopes as alternatives to the complex antigens used in current serodiagnostics for EBV infection.
INTRODUCTION
Epstein-Barr virus (EBV) is the primary causative agent of infectious mononucleosis and in severe cases is involved in the pathogenesis of certain types of lymphomas, nasopharyngeal carcinoma, and posttransplantation lymphoproliferative disease (2). The acute infection can be diagnosed serologically by measuring immunoglobulin M (IgM) antibodies primarily against viral capsid antigen (VCA). It is common, however, to augment this test with others to measure the levels of IgM, IgG, or IgA antibody to VCA, EBV nuclear antigen (EBNA), and early antigen (EA-D) to assess the staging and the severity of the disease (18).
Enzyme-linked immunosorbent assay (ELISA) is the most sensitive method for the diagnosis of EBV infection, and current commercial tests usually use native antigens purified from EBV-producing cultures to capture IgM or IgG antibodies in patients' serum (15). Some ELISAs use recombinant antigens to EBNA and EA-D (5); however, affinity-purified native antigens have repeatedly shown the best performance for commercial VCA assays (7). Furthermore, since all EBV antigens are typically complexes of two to six proteins that are often highly glycosylated, this has hindered their use in commercial diagnostics. The development of a simpler and more uniform diagnostic test is desirable; and it would be highly advantageous if, instead of the whole protein, only a peptide representing the antibody-binding site involved in the recognition could be used. The use of peptide epitopes as diagnostic antigens may allow a focus on relevant single peptide specificities and avoid the diagnostically unimportant epitopes present in crude extracts or complex antigens. Peptides have other significant advantages: peptides of high quality and stability can be cheaply and reproducibly produced and easily applied to ELISA as well as other formats, and unlike recombinant antigens, they can mimic carbohydrate epitopes (17).
The aim of this work was to generate peptides corresponding to immunodominant antigenic determinants of EBV. Combinations of these peptides may be ideally suited as replacements for the complex protein antigens used in diagnostics. One method commonly used to identify immunodominant regions is to synthesize overlapping peptides spanning whole proteins (24). However, these peptides represent mainly continuous epitopes or linear antibody-binding sites. Since a large number of functionally important serum antibodies are reactive with conformational or discontinuous epitopes, full coverage of antigenic sites will not be achieved. To circumvent this, we have chosen a phage-display approach to identify peptides from a large library that bind to antibody-binding sites, thereby mimicking the three-dimensional conformational features of both linear and conformational epitopes. These peptides are defined as "mimotopes," as they mimic the essential features of the epitope but do not necessarily have the same sequence (8). For example, a small peptide showing no apparent similarity to the native protein sequence was shown to mimic the action of erythropoietin, with the activity being almost similar to that of the parent molecule (12). This approach requires antibodies only for screening purposes and does not require any structural data on the target antigen. In fact, peptides have been identified by using the phage-display technology where the causative target antigen was unknown (6). The concept of using mimotopes as diagnostic reagents is based on the likelihood that antibodies are directed to a few immunodominant epitopes of a pathogen. Recent work from many laboratories, including our own, has clearly shown that mimotopes selected from antibodies against pathological antigens can be important probes for the detection of antibodies produced during infection (3, 10, 13, 25). Mimotopes can be identified by screening phage-displayed libraries against sera with high titers of disease-specific antibodies or screening against monoclonal antibodies (MAbs) raised against a panel of immunodominant epitopes. Furthermore, peptides identified from phage-display libraries have been used successfully as antigenic probes in diagnostic immunoassays for hepatitis B and C (16, 19). The success of this approach is dependent on the ability of peptides to mimic immunodominant epitopes within native antigens.
We aim to select peptides that mimic the immunodominant epitopes of EBV to achieve total coverage of the serum antibodies generated after typical infection with EBV. The use of a combination of several synthetic peptide mimics could provide a good diagnostic method and information regarding the key immunodominant residues of EBV.
Gp125 (BALF 4) is a major antigen of VCA that has been proposed for use as a single antigen for the diagnosis of infectious mononucleosis (23). Currently, EBV crude cell extracts are purified by using an immunoaffinity column prepared with an antibody to gp125, and this is used as a solid-phase antigen to capture serum antibodies in commercial diagnostic tests (15). Other kits use p18, an 18-kDa immunodominant region of VCA (4). Gp125 is clearly an important diagnostic epitope, and here we describe a peptide selected from our random peptide library that mimics this epitope. To extend the coverage of EBV diagnostic epitopes we produced EBV MAbs and selected peptides that mimic these epitopes. Thus, each peptide effectively represents a different epitope of EBV and therefore recognizes a different subset of antibodies. In this study, we evaluated our panel of peptide mimotopes for their utility for the diagnosis of primary EBV infection.
MATERIALS AND METHODS
Human sera. Individual human serum samples were provided by Queensland Medical Laboratory (Brisbane, Australia). The positive sera were collected from individuals with recent or the early stage of active infectious mononucleosis and were tested for the presence of IgM antibodies to EBV VCA by using a commercial diagnostic test. The negative sera were collected from patients with no previous exposure to EBV infection and were defined as seronegative by using the commercial diagnostic test.
(i) Screening of peptides alone. We screened 41 VCA IgM-positive samples, 17 VCA IgM-negative samples, and 4 additional samples that were IgM seropositive for cytomegalovirus (CMV) and negative for EBV IgM.
(ii) Screening of bovine serum albumin (BSA)-conjugated peptides. We screened 64 VCA IgM-positive samples, 110 VCA IgM-seronegative samples, and 8 serum samples from patients with past EBV infection that were either seropositive for VCA IgG and/or seropositive for EBNA and negative for VCA IgM. Putative cross-reactive sera were also screened: 19 CMV-positive samples, 3 varicella-zoster virus-positive samples, 3 parvovirus-positive samples, 3 herpes simplex virus (HSV) -positive samples, and 6 rheumatoid factor (RF)-positive samples.
Monoclonal antibodies. Anti-gp125 antibody L2 was purchased from Applied Biosystems. F1, A2, and A3 MAbs were raised in-house by immunizing mice with partially purified EBV-infected cell extract or crude EBV (Applied Biosystems). This crude EBV preparation was shown to have a high degree of reactivity with antibodies to VCA, EBNA, and EA-D. Mice were immunized subcutaneously with 20 μg EBV extract emulsified in 0.5 ml Freund's complete adjuvant. Two biweekly booster doses diluted 1:1 in Freund's incomplete adjuvant, followed by final double-dose booster doses without adjuvant 3 days before the mice were culled, were administered. Spleen cells were harvested, and hybridomas were generated by standard procedures (9).
Phage library and selection. We constructed a linear peptide library of 20 random amino acids displayed as N-terminal fusions to protein III of filamentous phage M13, using the Fuse 5 vector (21). The library of >5 x 108 random peptides (3) was screened for peptides that bound to the gp125, F1, A2, and A3 MAbs by using microtiter plates, as described previously (1), with the following modifications. Coated wells were blocked with 5% BLOTTO (milk powder diluted in phosphate-buffered saline [PBS]), and the 20-mer phage peptide library was preincubated in 1% BLOTTO for 15 min to remove any milk-binding phage before it was added to the MAb-coated wells. The phages were amplified and titrated and DNA sequencing was performed by procedures similar to those described previously (21).
ELISAs. Phage ELISAs were performed by coating 10 μg/ml of antibody in PBS to a microtiter plate (Maxisorp; Nunc) overnight at 4°C and subsequent blocking with 10% BLOTTO for 2 h. Phage dilutions (100 μl) were prepared in PBS, transferred in duplicate to the coated blocked wells, and incubated for 1 h with shaking. The wells were washed five times with PBS containing 0.05% Tween 20 (PBST), and 100 μl of anti-M13 antibody conjugated to horseradish peroxidase (HRP; Pharmacia) at a 1/5,000 dilution was added to each well. After 1 h incubation and washing as described above, the bound phages were detected with o-phenylenediamine (Sigma). For competition experiments with crude EBV and p18 antigens, the same concentration of phage (50 μl) was mixed with various concentrations of competitor (50 μl) for 30 min before it was added to the coated wells and the same procedure described above was followed.
A similar format was used for EBV antigen ELISAs, with crude EBV, p18, EBNA, EA-D, and VCA antigens used for coating (PanBio Ltd.). Anti-human IgM-HRP conjugate (PanBio Ltd.) or sheep anti-human IgM-HRP conjugate (Chemicon) was then added at a 1/5,000 dilution in fish gelatin diluent for 1 h. Binding was detected with 3,3',5,5'-tetramethylbenzidine substrate (Sigma).
ELISAs with synthetic peptides were performed by using peptide immobilizer plates (Exiqon, Vedbaek, Denmark). Briefly, the wells were coated with 10 μg/ml peptide in 0.1 M sodium carbonate buffer overnight at 4°C with gentle agitation. Control wells without any peptide coating were used for each sample. The plate was washed three times with PBS-Tween 20 (0.1%); and serum samples (100 μl) at a 1/100 dilution in PBS containing 2% fish gelatin (Sigma), 1% Tween 20, and 1% bovine serum albumin were added in triplicate for 1 h. The remaining steps were the same as those described above for the EBV antigen ELISA.
For ELISAs with peptides conjugated to BSA, the conjugates were coupled to Maxisorp plates at 5 μg/ml overnight at 4°C. The plates were washed three times with PBS, and serum samples (1/100) in PBST were added. The rest of the procedure was the same as that described above for the synthetic peptides alone.
All ELISAs were performed in duplicate or triplicate, and the assays were repeated to ensure reproducible results.
Peptide synthesis and conjugation to BSA. Peptides were synthesized to >70% purity by AusPep Pty. Ltd. For the A3 peptide, an intramolecular disulfide bond was formed between the cysteine residues. The gp125 and F1 peptides were dissolved in dimethyl formamide, the A3 peptide was dissolved peptide in 20% ethanol, and the A2 peptide was dissolved in dimethyl sulfoxide, all at 1 mg/ml; and the peptides were stored in aliquots at –20°C.
Peptides F1 and gp125 were synthesized with four additional glycine residues and a cysteine residue at the C' terminus to allow conjugation to the heterobifunctional cross-linker succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC; Pierce). Briefly, a 100 molar excess of SMCC was added to 5 mg BSA (Sigma) in 0.1 M sodium phosphate-0.1 M sodium chloride buffer, and the components were mixed for 2 h at room temperature. To remove excess linker, the mixture was desalted with a PD-10 column equilibrated with the same buffer containing 0.1 M EDTA. For conjugation, 1 mg of peptide was incubated with 1 to 2 mg of BSA-SMCC in the presence of 20% dimethyl sulfoxide for 2 h. The final BSA-conjugated peptide was desalted by using a PD-10 column into PBS.
RESULTS
MAbs covering different EBV epitopes. We raised MAbs by immunizing mice with crude EBV, and in order to select for hybridomas expressing antibodies to various EBV epitopes, screening involved the selection of clones with different binding specificities. The F1 (IgG2b subclass), A2 (IgM), and A3 (IgG1) MAbs, which exhibited different profiles of binding to EBV components (Fig. 1), were selected for use as screening reagents. The A2 MAb reacted predominantly with p18, whereas the A3 MAb recognized only an epitope present in crude EBV. In contrast, the F1 MAb bound strongly to crude EBV and VCA. The commercial anti-gp125 MAb L2 (IgG1) reacted predominantly with VCA, with a lower level of binding to crude EBV.
Screening for EBV mimotopes. Peptides reacting with the gp125, F1, A2, and A3 MAbs were isolated by screening a 20-amino-acid random linear peptide library by using multiple rounds of panning on MAb-coated microtiter plates. To bias selection toward the peptides that bind with a high affinity, the stringency of washing was increased with the successive rounds of panning. For each selection, increased numbers of bound phages were detected after the third round of panning, with a further increase in round 4. An example is shown for the A2 MAb in Fig. 2A. Similar binding patterns were observed for the F1, gp125, and A3 selections.
DNA sequencing of 20 clones from rounds 3 and 4 revealed only one peptide that was isolated for the gp125, F1, and A2 MAbs, whereas three different sequences were isolated for the A3 MAb (Fig. 2B). No consensus sequence was apparent for any of the A3 clones, although all three peptides contained two cysteine residues, which could form an intramolecular disulfide bond. Clone 2 (C2) repeatedly demonstrated the highest level of binding compared with those of C1 and C3 at the same phage concentration (Fig. 3A); therefore, A3 C2 was selected for further study. Clones isolated from F1, A2 (C2), A3, and gp125 bound only to the corresponding MAb, indicating that binding was specific for individual epitopes. Importantly, none of the phage clones selected was reactive with the relevant isotype control MAb (Fig. 3B), and binding to the parent MAb was inhibited by crude EBV antigen. An example of this is shown for the gp125 clone in Fig. 3C, where phage binding to the gp125 MAb decreased with an increasing amount of competing crude EBV. Interestingly, clone A2 competed with the p18 antigen but weakly with the EBV antigen (data not shown). This is consistent with the specificity of the parent A2 MAb, which bound strongly to p18 and weakly to crude EBV (Fig. 1). Thus, the peptides that we have selected appear to be true epitope mimics which share the antibody-binding specificity of the corresponding epitope on EBV or p18 in the case of A2.
Peptides corresponding to each of the selected clones were found to bind to the corresponding MAb (data not shown), confirming that the MAb epitopes were contained within the peptide sequence and were not dependent on the phage context.
EBV peptide mimics as diagnostic reagents. To assess the diagnostic potential of our panel of peptides, we assayed their ability to recognize EBV-specific antibodies in clinical samples. An ELISA was designed to capture EBV antibodies in the patients' serum by coating peptides in separate wells as solid-phase antigens. To ensure that blocking was adequate, each serum sample was analyzed for binding to wells without peptide. We analyzed the reactivities of 41 EBV-seropositive and 21 EBV-seronegative serum samples for all four peptides individually. The cutoff level was evaluated for each data set by taking the mean optical density (OD) at 450 nm of the negative samples and adding 3 standard deviations (SDs). By using this criterion, any sample with a value above the cutoff was defined as positive and any sample with a value below this was defined as negative. Each individual absorbance reading is plotted in Fig. 4. Although the majority of absorbance readings were <0.8 OD units, the data show a difference between the readings for the EBV-positive and the EBV-negative sera. For the negative population, there were no false-positive results, resulting in 100% specificity for all four peptides. Importantly, the peptides did not recognize serum samples that had previously tested positive for reactivity with CMV, which is a common cross-reaction for existing EBV diagnostics (18). For the same set of serum samples, we also measured the abilities of the peptides to recognize IgG antibodies. By use of the ELISA conditions described here, the results revealed comparatively low ELISA signals and no clear distinction between positive and negative samples, resulting in a very low specificity of detection for IgG (data not shown). Therefore, no further diagnostic analysis was performed by use of the peptides for the detection of EBV IgG.
The sensitivity of detection for the IgM-positive samples is shown in Table 1. As a single peptide, F1 had the highest sensitivity of 88%, correctly detecting IgM antibodies in 36 of the 41 positive individual serum samples. The A3 peptide had the next highest sensitivity (85%), followed by the gp125 peptide (71%) and, lastly, the A2 peptide (54%). Only one of the EBV IgM-positive serum samples did not give an ELISA signal with any of the peptide mimics; however, this sample gave a low signal in the commercial EBV diagnostic test.
A combination of any of the peptides resulted in an increase in the sensitivity (Table 1), since each peptide effectively recognized a different subset of antibodies, thus indicating that antibodies to different epitopes are being detected. The best combination of two peptides for the recognition of EBV IgM antibodies was gp125 and A3, which achieved a maximum sensitivity of 97.5%. Overall, the sensitivity was generally higher by combining three peptides, but the use of all four peptides in combination did not further improve the sensitivity.
Improving the diagnostic sensitivity by conjugation of peptides to BSA. In order to improve the presentation of peptides on a solid phase for diagnostic ELISA, we selected two of the peptides, gp125 and F1, for conjugation to a carrier molecule, BSA. A larger panel of serum samples was analyzed for reactivity with gp125-BSA and F1-BSA conjugates; the panel included a higher number of putatively cross-reactive IgM-positive serum samples (Fig. 5). To correct for background levels, the reactivity to BSA alone was also measured; and these values were subtracted from the BSA-peptide readings. There was a dramatic improvement in sensitivity when the peptides were conjugated to BSA (Table 2). For example, by using the F1 peptide alone, only 12 of 41 OD readings were >0.4; however, when F1 was attached to BSA, 59 of 64 of the positive OD readings were >0.4. Both BSA-conjugated peptides failed to recognize eight serum samples that were VCA IgG positive and/or EBNA positive and VCA IgM negative, typical of past EBV infection, again demonstrating the high degrees of specificity of the mimotopes for the recognition of IgM antibodies in clinical samples. Furthermore, there was little cross-reactivity with other viral IgM antibodies and rheumatoid factor, which is a common cross-reactive species in current commercial diagnostic tests.
DISCUSSION
Diagnosis of EBV-related diseases relies on detection of EBV-specific antibodies in the serum of infected individuals. Current commercial tests usually use native antigens purified from EBV-producing cultures to capture serum antibodies. As possible replacements for these complex antigens we have generated EBV antigen mimics by selecting peptides from a phage display library. In this study we demonstrated that our panel of peptides can detect IgM antibodies typically produced in human subjects after infection with EBV. One of the peptides was selected against a commercial MAb to the gp125 antigen of VCA, since this antigen is known to be diagnostically important. The other peptides (peptides F1, A2, and A3) were isolated by screening the library against novel MAbs, generated by immunizing mice with the crude EBV antigen; and although the exact antigen that the corresponding peptides mimic is not known, they all appear to be diagnostically important.
The highest sensitivity of detection of IgM in EBV-positive sera was achieved by using the F1 peptide alone (88%), and this was further improved when F1 was coupled to the carrier molecule BSA (95%). It is perceivable that F1-BSA could be used alone for diagnosis; however, further characterization of the identity of the epitope and, importantly, a much larger set of clinical samples must be screened before combinations of peptides can be eliminated from our studies.
A high sensitivity of detection was also achieved with the A3 peptide (85%). Although the identity of the A3 epitope is also unknown, these data indicate that the antigen that this peptide mimics is important diagnostically. Future studies are planned to reveal the identity of this antigen. The sensitivity for the gp125 peptide was slightly less than 71%, and again, this was significantly improved when the peptide was coupled to the carrier protein BSA (92%). It is known that this epitope on VCA is important diagnostically, and the high level of sensitivity reflects this. Peptide A2 had the lowest sensitivity of detection of 54%; this peptide competed only with p18 and not with the EBV complex, indicating specificity for this antigen. Serum antibodies reactive with this epitope could be less common, resulting in the lower detection rate. The parent antibody, A2, is an IgM isotype, and this subclass tends to have lower affinity and often higher cross-reactivity than IgG subclasses. Therefore, these characteristics could also be reflected in the corresponding A2 peptide.
Exposure to EBV infection will stimulate the immune system to generate antibodies to multiple exposed surface epitopes, and individual serum samples are likely to have antibodies to several immunodominant regions of EBV. An effective diagnostic method would aim to cover as many of these immunodominant epitopes as necessary to recognize every EBV-positive sample. Each peptide effectively represents a different epitope and will therefore recognize a different subset of antibodies; therefore, use of a combination of the peptides should lead to greater coverage of these important epitopes and an increase in the sensitivity of detection. Our results showed that this phenomenon was observed, as any peptide combination increased the sensitivity (Table 1).
The particular peptide conformation that antibodies will preferentially recognize is not predictable; therefore, prospective diagnostic peptides need to be tested in a variety of formats. The ELISA plates that we used in this study were specifically designed for immobilization of peptides; the free amino groups were adsorbed via an ethylene glycol spacer. Our preliminary experiments showed that this was preferable to adsorption directly onto microtiter plates (data not shown). However, when peptides are displayed on phages, they may adopt a more rigid conformation than the free peptide, as they are anchored to the pIII protein at the C terminus, leaving the N terminus free. We therefore chose to couple two of the peptides via the C terminus to BSA to resemble their orientations on the phages. We assume that the combination of an increased number of mimotopes attached to BSA and the fact the configuration of the mimotope was somehow more accessible to antibody binding both contributed to the increase in diagnostic sensitivity. Experiments are in progress to optimize solid-phase coupling of all of the peptides to BSA alone or in combination to achieve the greatest sensitivity. Other solid-phase formats such as latex or nitrocellulose, which are used in rapid diagnostic tests, are also being investigated.
It is also unlikely that all 20 amino acid residues of the peptides will be involved in binding, and it is possible by alanine scanning to identify the residues essential for binding (11). Once the essential amino acids have been identified, nonessential residues could be removed, resulting in shorter peptides that could also be combined. A similar technique was used for a peptide representing the p18 epitope. A "mixotope" representing a collection of mimotopes was synthesized and improved the sensitivity and specificity of detection of EBV antibodies in clinical samples (22).
When the optimal assay format has been selected, a more accurate assessment of the diagnostic value of these peptides with a larger panel of positive, negative, and putative cross-reactive sera can be completed. Here we have shown that when the F1 and gp125 peptides are attached to BSA, they do not recognize sera from patients with past infections, which are typically characterized by elevated VCA IgG and/or EBNA levels but which have a low VCA IgM titer. In a larger study it would be beneficial to study the ability to identify reactivation infections or the phase of infection (acute or convalescent) of a given patient to evaluate the true diagnostic value of these mimotopes.
In this study the peptides appeared to be EBV-specific mimotopes and did not cross-react with the sera IgM positive for CMV, parvovirus, varicella-zoster virus, HSV, or RF. However, since one of the purposes of a peptide ELISA is to increase specificity by eliminating potentially cross-reactive epitopes present in the full-length protein or antigen extract and in remaining contaminants in the purified protein, if these peptides are to be developed for a diagnostic test, further screening against all typically cross-reactive species is required. The most common cross-reactions of current EBV diagnostic kits are rheumatoid factor and other herpesviruses. Routine diagnostics have been developed to remove cross-reacting rheumatoid factor by preabsorption with anti-Ig. Therefore, it would be a major advantage if the mimotopes would prove to have a higher specificity than existing commercial tests and would not require the removal of rheumatoid factor. We have shown here that the F1 and gp125 mimotopes do not recognize 6 serum samples positive for rheumatoid factor and 34 cross-reactive serum samples, and in this study there were no false-positive results (100% specificity). However, further analyses with a much larger panel of putatively cross-reactive sera in a follow-up study are required to confirm this finding.
To select for further peptides that could recognize serum antibodies that were previously undetectable by our peptides, the library could be screened against this patient's serum. Since MAbs represent only a small portion of the immune response to the antigen, polyclonal sera may cover different epitopes simultaneously, including the epitopes missed in our experiments. Our preliminary data have shown that many peptides can be selected by this approach (data not shown) and could also lead to the identification of peptides not only specific for EBV IgM antibodies but also specific for EBV IgG antibodies. Other authors have successfully used this approach for selection of diagnostic peptides for Lyme disease (10). Furthermore, this simple process of identifying generic peptides could be applicable to any infectious disease for which antibodies are available. In addition, this technology does not necessarily require knowledge of the original ligand (6), and recently, an antigen thought to be involved in metastatic disease progression was identified by using peptide mimics against circulating antibodies in the serum of cancer patients (14).
In conclusion, the data presented here indicate that our panel of peptides can be used for the detection of IgM antibodies typically generated after infection with EBV. This study proves that the use of EBV peptide mimotopes in an ELISA system has a high diagnostic potential, even though there is no prior knowledge of the antigens. This technology is also applicable to many infectious diseases and can be important for vaccine development if the peptides effectively mimic protective antibodies (3, 13, 20, 25). It also has significant advantages over existing commercial tests for EBV, as the peptides are of relatively low cost, highly stable, and easily synthesized. Moreover, peptides can mimic carbohydrate epitopes present on many EBV antigen complexes and allow a focus on single subspecificities, thereby avoiding dilution with noninformative epitopes.
ACKNOWLEDGMENTS
We thank Joan Hoogenraad and Rosella Masciantonio (La Trobe University) for technical assistance and Edward Kachab (PanBio Ltd) for providing reagents.
This study was supported by the Cooperative Research Centre for Diagnostics (Australia).
Present address: Uniseed Pty. Ltd., Cumbrae-Stewart Building, Research Road, University of Queensland, St Lucia, QLD 4072, Australia.
REFERENCES
Adda, C., L. M. Tilley, R. F. Anders, and M. Foley. 1999. Isolation of peptides that mimic epitopes on a malarial antigen from random peptide libraries displayed on phage. Infect. Immun. 67:4679-4688.
Bharadwaj, M., A. Suhrbier, A. Elliott, and D. J. Moss. 2000. Epstein-Barr virus and associated cancers. Today's Life Sci. 13:43-45.
Casey, J. L., A. M. Coley, R. F. Anders, V. J. Murphy, K. S. Humberstone, A. W. Thomas, and M. Foley. 2004. Antibodies to malaria peptide mimics inhibit Plasmodium falciparum invasion of erythrocytes. Infect. Immun. 72:1126-1134.
Chan, K. H., R. X. Luo, H. L. Chen, M. H. Ng, W. H. Seto, and J. S. M. Peiris. 1998. Development and evaluation of an Epstein-Barr virus (EBV) immunoglobulin M enzyme-linked immunosorbent assay based on the 18-kilodalton matrix protein for diagnosis of primary EBV infection. J. Clin. Microbiol. 36:3359-3360.
Faber, I., P. Wutler, P. Wohlrabe, H. Wolf, W. Hinderer, and H. H. Sonneborn. 1993. Serological diagnosis of infectious mononucleosis using three anti-Epstein-Barr virus recombinant ELISAs. J. Virol. Methods 42:301-308.
Folgori, A., R. Tafi, A. Meola, F. Felici, G. Galfre, R. Cortese, P. Monaci, and A. Nicosa. 1994. A general strategy to identify mimotopes of pathological antigens using only random peptide libraries and human sera. EMBO J. 13:2236-2243.
Gartner, B. C., R. D. Hess, D. Bandt, A. Kruse, A. Rethwilm, K. Roemer, and N. Mueller-Lantzsch. 2003. Evaluation of four commercially available Epstein-Barr virus enzyme immunoassays with an immunofluorescence assay as the reference method. Clin. Diagn. Lab. Immunol. 10:78-82.
Geysen, H. M., S. J. Barteling, and R. H. Meloen. 1984. Use of peptide synthesis to probe viral antigens for epitopes to a resolution of a single amino acid. Proc. Natl. Acad. Sci. USA 82:3998-4002.
Harlow, E., and D. Lane (ed.). 1988. Antibodies: a laboratory manual, p. 139-244. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Kouzmitcheva, G. A., V. A. Petrenko, and G. P. Smith. 2001. Identifying diagnostic peptides for Lyme disease through epitope discovery. Clin. Diagn. Lab. Immunol. 8:150-160.
Li, F., A. Dluzewski, A. M. Coley, A. W. Thomas, L. M. Tilley, R. Anders, and M. Foley. 2002. Phage-displayed peptides bind to the malarial protein apical membrane antigen 1 and inhibit the merozoite invasion of host erythrocytes. J. Biol. Chem. 277:50303-50310.
Livnah, O., E. A. Stura, D. L. Johnson, S. A. Middleton, L. S. Mulcahy, N. C. Wrighton, W. J. Dower, L. K. Joliffe, and I. A. Wilson. 1996. Functional mimicry of a protein hormone by a peptide agonist: the EPO receptor complex at 2.8. Science 273:464-471.
Lundin, K., A. Samuelsson, M. Jansson, J. Hinkula, B. Wahren, H. Wigzell, and M. A. Persson. 1996. Peptides isolated from random peptide libraries on phage elicit a neutralizing anti-HIV-1 response: analysis of immunological mimicry. Immunology 89:579-586.
Mintz, P. J., J. Kim, K. A. Do, X. Wang, R. G. Zinner, M. Cristofanilli, M. A. Arap, W. K. Hong, P. Troncoso, C. J. Logothetis, R. Pasqualini, and W. Arap. 2003. Fingerprinting the circulating repertoire of antibodies from cancer patients. Nat. Biotechnol. 21:57-62.
Mitchell, J. L., C. M. Doyle, M. V. Land, and P. L. Devine. 1998. Comparison of commercial ELISA for detection of antibodies to the viral capsid antigen (VCA) of Epstein-Barr virus (EBV). Disease Markers 13:245-249.
Motti, C., N. Maurizio, A. Meola, G. Galfre, F. Felici, R. Cortes, A. Nicosia, and P. Monaci. 1994. Recognition by human sera and immunogenicity of HbsAg mimotopes selected from an M13 phage display library. Gene 146:191-198.
Oldenburg, K. R., D. Loganathan, I. J. Goldstein, P. G. Schultz, and M. A. Gallop. 1992. Peptide ligands for a sugar-binding protein isolated from a random peptide library. Proc. Natl. Acad. Sci. USA 89:5393-5397.
Pearson, G. R. 1988. ELISA tests and monoclonal antibodies for EBV. J. Virol. Methods 21:97-104.
Puntoriero, G. A., A. Meola, A. Lahm, S. Zucchelli, B. B. Ercole, R. Tafi, M. Pezzanera, M. U. Mondelli, R. Cortese, A. Tramontano, G. Galfre, and A. Nicosia. 1998. Towards a solution for hepatitis C virus hypervariability: mimotopes of the hypervariable region 1 can induce antibodies cross-reacting with a large number of viral variants. EMBO J. 17:3521-3533.
Scala, G., X. Chen, W. Liu, J. N. Telles, O. J. Cohen, M. Vaccarezza, T. Igarashi, and A. S. Fauci. 1999. Selection of HIV-specific immunogenic epitopes by screening random peptide libraries with HIV-1 positive sera. J. Immunol. 162:6155-6161.
Smith, G. P., and J. K. Scott. 1993. Libraries of peptides and proteins displayed on phage. Methods Enzymol. 217:228-257.
Tranchand-Bunel, D., C. Auriault, E. Diesis, and H. Gras-Masse. 1998. Detection of human antibodies using ‘convergent’ combinatorial peptide libraries or ‘mixotopes’ designed from a nonvariable antigen: application to the EBV viral capsid antigen p18. J. Peptide Res. 52:495-508.
Van Grunsven, W. M. J., W. J. M. Spaan, and J. M. Middeldorp. 1994. Localization and diagnostic application of immunodominant domains of the BFRF3-encoded Epstein-Barr virus capsid protein. J. Infect. Dis. 170:121-127.
Van Regenmortel, M. H. V. 1998. Synthetic peptides help in diagnosing viral infections. ASM News 64:332-338.
Zhong, G., G. P. Smith, J. Berry, and R. C. Brunham. 1994. Conformational mimicry of a chlamydial neutralization epitope on filamentous phage. J. Biol. Chem. 269:24183-24188.(Joanne L. Casey, Andrew M)
Department of Biochemistry, La Trobe University, VIC 3086, Australia
Queensland Medical Laboratory, P.O. Box 5410, West End, QLD 4101, Australia
ABSTRACT
Epstein-Barr virus (EBV) is a ubiquitous, worldwide infectious agent that causes infectious mononucleosis, affecting >90% of the world's population. Currently, enzyme-linked immunosorbent assay, mostly with purified preparations of EBV cell extracts to capture immunoglobulin M (IgM) antibodies in patients' serum, is used for primary diagnosis. Our objective was to determine whether a small set of peptides could contain sufficient immunogenic information to replace solid-phase antigens in EBV diagnostics. Using monoclonal antibodies, we selected four peptides that mimic different epitopes of EBV from a phage-displayed random peptide library. To assess their diagnostic value, we screened a panel of 62 individual EBV IgM sera for their reactivities with the peptides alone. For all peptides, there was a clear distinction between the EBV-positive and the EBV-negative samples, resulting in 100% specificity. The sensitivities were 88%, 85%, 71%, and 54% for peptides F1, A3, gp125, and A2, respectively. Any combination of peptides increased the sensitivity, indicating that individual peptides react with different subsets of antibodies. Furthermore, when the F1 and the gp125 peptides were coupled to bovine serum albumin and screened against 216 serum samples, there were dramatic improvements in sensitivities (95% and 92%, respectively) and little cross-reactivity with the other peptides encountered during acute viral infections, including rheumatoid factor. This study shows the potential for the use of peptide mimotopes as alternatives to the complex antigens used in current serodiagnostics for EBV infection.
INTRODUCTION
Epstein-Barr virus (EBV) is the primary causative agent of infectious mononucleosis and in severe cases is involved in the pathogenesis of certain types of lymphomas, nasopharyngeal carcinoma, and posttransplantation lymphoproliferative disease (2). The acute infection can be diagnosed serologically by measuring immunoglobulin M (IgM) antibodies primarily against viral capsid antigen (VCA). It is common, however, to augment this test with others to measure the levels of IgM, IgG, or IgA antibody to VCA, EBV nuclear antigen (EBNA), and early antigen (EA-D) to assess the staging and the severity of the disease (18).
Enzyme-linked immunosorbent assay (ELISA) is the most sensitive method for the diagnosis of EBV infection, and current commercial tests usually use native antigens purified from EBV-producing cultures to capture IgM or IgG antibodies in patients' serum (15). Some ELISAs use recombinant antigens to EBNA and EA-D (5); however, affinity-purified native antigens have repeatedly shown the best performance for commercial VCA assays (7). Furthermore, since all EBV antigens are typically complexes of two to six proteins that are often highly glycosylated, this has hindered their use in commercial diagnostics. The development of a simpler and more uniform diagnostic test is desirable; and it would be highly advantageous if, instead of the whole protein, only a peptide representing the antibody-binding site involved in the recognition could be used. The use of peptide epitopes as diagnostic antigens may allow a focus on relevant single peptide specificities and avoid the diagnostically unimportant epitopes present in crude extracts or complex antigens. Peptides have other significant advantages: peptides of high quality and stability can be cheaply and reproducibly produced and easily applied to ELISA as well as other formats, and unlike recombinant antigens, they can mimic carbohydrate epitopes (17).
The aim of this work was to generate peptides corresponding to immunodominant antigenic determinants of EBV. Combinations of these peptides may be ideally suited as replacements for the complex protein antigens used in diagnostics. One method commonly used to identify immunodominant regions is to synthesize overlapping peptides spanning whole proteins (24). However, these peptides represent mainly continuous epitopes or linear antibody-binding sites. Since a large number of functionally important serum antibodies are reactive with conformational or discontinuous epitopes, full coverage of antigenic sites will not be achieved. To circumvent this, we have chosen a phage-display approach to identify peptides from a large library that bind to antibody-binding sites, thereby mimicking the three-dimensional conformational features of both linear and conformational epitopes. These peptides are defined as "mimotopes," as they mimic the essential features of the epitope but do not necessarily have the same sequence (8). For example, a small peptide showing no apparent similarity to the native protein sequence was shown to mimic the action of erythropoietin, with the activity being almost similar to that of the parent molecule (12). This approach requires antibodies only for screening purposes and does not require any structural data on the target antigen. In fact, peptides have been identified by using the phage-display technology where the causative target antigen was unknown (6). The concept of using mimotopes as diagnostic reagents is based on the likelihood that antibodies are directed to a few immunodominant epitopes of a pathogen. Recent work from many laboratories, including our own, has clearly shown that mimotopes selected from antibodies against pathological antigens can be important probes for the detection of antibodies produced during infection (3, 10, 13, 25). Mimotopes can be identified by screening phage-displayed libraries against sera with high titers of disease-specific antibodies or screening against monoclonal antibodies (MAbs) raised against a panel of immunodominant epitopes. Furthermore, peptides identified from phage-display libraries have been used successfully as antigenic probes in diagnostic immunoassays for hepatitis B and C (16, 19). The success of this approach is dependent on the ability of peptides to mimic immunodominant epitopes within native antigens.
We aim to select peptides that mimic the immunodominant epitopes of EBV to achieve total coverage of the serum antibodies generated after typical infection with EBV. The use of a combination of several synthetic peptide mimics could provide a good diagnostic method and information regarding the key immunodominant residues of EBV.
Gp125 (BALF 4) is a major antigen of VCA that has been proposed for use as a single antigen for the diagnosis of infectious mononucleosis (23). Currently, EBV crude cell extracts are purified by using an immunoaffinity column prepared with an antibody to gp125, and this is used as a solid-phase antigen to capture serum antibodies in commercial diagnostic tests (15). Other kits use p18, an 18-kDa immunodominant region of VCA (4). Gp125 is clearly an important diagnostic epitope, and here we describe a peptide selected from our random peptide library that mimics this epitope. To extend the coverage of EBV diagnostic epitopes we produced EBV MAbs and selected peptides that mimic these epitopes. Thus, each peptide effectively represents a different epitope of EBV and therefore recognizes a different subset of antibodies. In this study, we evaluated our panel of peptide mimotopes for their utility for the diagnosis of primary EBV infection.
MATERIALS AND METHODS
Human sera. Individual human serum samples were provided by Queensland Medical Laboratory (Brisbane, Australia). The positive sera were collected from individuals with recent or the early stage of active infectious mononucleosis and were tested for the presence of IgM antibodies to EBV VCA by using a commercial diagnostic test. The negative sera were collected from patients with no previous exposure to EBV infection and were defined as seronegative by using the commercial diagnostic test.
(i) Screening of peptides alone. We screened 41 VCA IgM-positive samples, 17 VCA IgM-negative samples, and 4 additional samples that were IgM seropositive for cytomegalovirus (CMV) and negative for EBV IgM.
(ii) Screening of bovine serum albumin (BSA)-conjugated peptides. We screened 64 VCA IgM-positive samples, 110 VCA IgM-seronegative samples, and 8 serum samples from patients with past EBV infection that were either seropositive for VCA IgG and/or seropositive for EBNA and negative for VCA IgM. Putative cross-reactive sera were also screened: 19 CMV-positive samples, 3 varicella-zoster virus-positive samples, 3 parvovirus-positive samples, 3 herpes simplex virus (HSV) -positive samples, and 6 rheumatoid factor (RF)-positive samples.
Monoclonal antibodies. Anti-gp125 antibody L2 was purchased from Applied Biosystems. F1, A2, and A3 MAbs were raised in-house by immunizing mice with partially purified EBV-infected cell extract or crude EBV (Applied Biosystems). This crude EBV preparation was shown to have a high degree of reactivity with antibodies to VCA, EBNA, and EA-D. Mice were immunized subcutaneously with 20 μg EBV extract emulsified in 0.5 ml Freund's complete adjuvant. Two biweekly booster doses diluted 1:1 in Freund's incomplete adjuvant, followed by final double-dose booster doses without adjuvant 3 days before the mice were culled, were administered. Spleen cells were harvested, and hybridomas were generated by standard procedures (9).
Phage library and selection. We constructed a linear peptide library of 20 random amino acids displayed as N-terminal fusions to protein III of filamentous phage M13, using the Fuse 5 vector (21). The library of >5 x 108 random peptides (3) was screened for peptides that bound to the gp125, F1, A2, and A3 MAbs by using microtiter plates, as described previously (1), with the following modifications. Coated wells were blocked with 5% BLOTTO (milk powder diluted in phosphate-buffered saline [PBS]), and the 20-mer phage peptide library was preincubated in 1% BLOTTO for 15 min to remove any milk-binding phage before it was added to the MAb-coated wells. The phages were amplified and titrated and DNA sequencing was performed by procedures similar to those described previously (21).
ELISAs. Phage ELISAs were performed by coating 10 μg/ml of antibody in PBS to a microtiter plate (Maxisorp; Nunc) overnight at 4°C and subsequent blocking with 10% BLOTTO for 2 h. Phage dilutions (100 μl) were prepared in PBS, transferred in duplicate to the coated blocked wells, and incubated for 1 h with shaking. The wells were washed five times with PBS containing 0.05% Tween 20 (PBST), and 100 μl of anti-M13 antibody conjugated to horseradish peroxidase (HRP; Pharmacia) at a 1/5,000 dilution was added to each well. After 1 h incubation and washing as described above, the bound phages were detected with o-phenylenediamine (Sigma). For competition experiments with crude EBV and p18 antigens, the same concentration of phage (50 μl) was mixed with various concentrations of competitor (50 μl) for 30 min before it was added to the coated wells and the same procedure described above was followed.
A similar format was used for EBV antigen ELISAs, with crude EBV, p18, EBNA, EA-D, and VCA antigens used for coating (PanBio Ltd.). Anti-human IgM-HRP conjugate (PanBio Ltd.) or sheep anti-human IgM-HRP conjugate (Chemicon) was then added at a 1/5,000 dilution in fish gelatin diluent for 1 h. Binding was detected with 3,3',5,5'-tetramethylbenzidine substrate (Sigma).
ELISAs with synthetic peptides were performed by using peptide immobilizer plates (Exiqon, Vedbaek, Denmark). Briefly, the wells were coated with 10 μg/ml peptide in 0.1 M sodium carbonate buffer overnight at 4°C with gentle agitation. Control wells without any peptide coating were used for each sample. The plate was washed three times with PBS-Tween 20 (0.1%); and serum samples (100 μl) at a 1/100 dilution in PBS containing 2% fish gelatin (Sigma), 1% Tween 20, and 1% bovine serum albumin were added in triplicate for 1 h. The remaining steps were the same as those described above for the EBV antigen ELISA.
For ELISAs with peptides conjugated to BSA, the conjugates were coupled to Maxisorp plates at 5 μg/ml overnight at 4°C. The plates were washed three times with PBS, and serum samples (1/100) in PBST were added. The rest of the procedure was the same as that described above for the synthetic peptides alone.
All ELISAs were performed in duplicate or triplicate, and the assays were repeated to ensure reproducible results.
Peptide synthesis and conjugation to BSA. Peptides were synthesized to >70% purity by AusPep Pty. Ltd. For the A3 peptide, an intramolecular disulfide bond was formed between the cysteine residues. The gp125 and F1 peptides were dissolved in dimethyl formamide, the A3 peptide was dissolved peptide in 20% ethanol, and the A2 peptide was dissolved in dimethyl sulfoxide, all at 1 mg/ml; and the peptides were stored in aliquots at –20°C.
Peptides F1 and gp125 were synthesized with four additional glycine residues and a cysteine residue at the C' terminus to allow conjugation to the heterobifunctional cross-linker succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC; Pierce). Briefly, a 100 molar excess of SMCC was added to 5 mg BSA (Sigma) in 0.1 M sodium phosphate-0.1 M sodium chloride buffer, and the components were mixed for 2 h at room temperature. To remove excess linker, the mixture was desalted with a PD-10 column equilibrated with the same buffer containing 0.1 M EDTA. For conjugation, 1 mg of peptide was incubated with 1 to 2 mg of BSA-SMCC in the presence of 20% dimethyl sulfoxide for 2 h. The final BSA-conjugated peptide was desalted by using a PD-10 column into PBS.
RESULTS
MAbs covering different EBV epitopes. We raised MAbs by immunizing mice with crude EBV, and in order to select for hybridomas expressing antibodies to various EBV epitopes, screening involved the selection of clones with different binding specificities. The F1 (IgG2b subclass), A2 (IgM), and A3 (IgG1) MAbs, which exhibited different profiles of binding to EBV components (Fig. 1), were selected for use as screening reagents. The A2 MAb reacted predominantly with p18, whereas the A3 MAb recognized only an epitope present in crude EBV. In contrast, the F1 MAb bound strongly to crude EBV and VCA. The commercial anti-gp125 MAb L2 (IgG1) reacted predominantly with VCA, with a lower level of binding to crude EBV.
Screening for EBV mimotopes. Peptides reacting with the gp125, F1, A2, and A3 MAbs were isolated by screening a 20-amino-acid random linear peptide library by using multiple rounds of panning on MAb-coated microtiter plates. To bias selection toward the peptides that bind with a high affinity, the stringency of washing was increased with the successive rounds of panning. For each selection, increased numbers of bound phages were detected after the third round of panning, with a further increase in round 4. An example is shown for the A2 MAb in Fig. 2A. Similar binding patterns were observed for the F1, gp125, and A3 selections.
DNA sequencing of 20 clones from rounds 3 and 4 revealed only one peptide that was isolated for the gp125, F1, and A2 MAbs, whereas three different sequences were isolated for the A3 MAb (Fig. 2B). No consensus sequence was apparent for any of the A3 clones, although all three peptides contained two cysteine residues, which could form an intramolecular disulfide bond. Clone 2 (C2) repeatedly demonstrated the highest level of binding compared with those of C1 and C3 at the same phage concentration (Fig. 3A); therefore, A3 C2 was selected for further study. Clones isolated from F1, A2 (C2), A3, and gp125 bound only to the corresponding MAb, indicating that binding was specific for individual epitopes. Importantly, none of the phage clones selected was reactive with the relevant isotype control MAb (Fig. 3B), and binding to the parent MAb was inhibited by crude EBV antigen. An example of this is shown for the gp125 clone in Fig. 3C, where phage binding to the gp125 MAb decreased with an increasing amount of competing crude EBV. Interestingly, clone A2 competed with the p18 antigen but weakly with the EBV antigen (data not shown). This is consistent with the specificity of the parent A2 MAb, which bound strongly to p18 and weakly to crude EBV (Fig. 1). Thus, the peptides that we have selected appear to be true epitope mimics which share the antibody-binding specificity of the corresponding epitope on EBV or p18 in the case of A2.
Peptides corresponding to each of the selected clones were found to bind to the corresponding MAb (data not shown), confirming that the MAb epitopes were contained within the peptide sequence and were not dependent on the phage context.
EBV peptide mimics as diagnostic reagents. To assess the diagnostic potential of our panel of peptides, we assayed their ability to recognize EBV-specific antibodies in clinical samples. An ELISA was designed to capture EBV antibodies in the patients' serum by coating peptides in separate wells as solid-phase antigens. To ensure that blocking was adequate, each serum sample was analyzed for binding to wells without peptide. We analyzed the reactivities of 41 EBV-seropositive and 21 EBV-seronegative serum samples for all four peptides individually. The cutoff level was evaluated for each data set by taking the mean optical density (OD) at 450 nm of the negative samples and adding 3 standard deviations (SDs). By using this criterion, any sample with a value above the cutoff was defined as positive and any sample with a value below this was defined as negative. Each individual absorbance reading is plotted in Fig. 4. Although the majority of absorbance readings were <0.8 OD units, the data show a difference between the readings for the EBV-positive and the EBV-negative sera. For the negative population, there were no false-positive results, resulting in 100% specificity for all four peptides. Importantly, the peptides did not recognize serum samples that had previously tested positive for reactivity with CMV, which is a common cross-reaction for existing EBV diagnostics (18). For the same set of serum samples, we also measured the abilities of the peptides to recognize IgG antibodies. By use of the ELISA conditions described here, the results revealed comparatively low ELISA signals and no clear distinction between positive and negative samples, resulting in a very low specificity of detection for IgG (data not shown). Therefore, no further diagnostic analysis was performed by use of the peptides for the detection of EBV IgG.
The sensitivity of detection for the IgM-positive samples is shown in Table 1. As a single peptide, F1 had the highest sensitivity of 88%, correctly detecting IgM antibodies in 36 of the 41 positive individual serum samples. The A3 peptide had the next highest sensitivity (85%), followed by the gp125 peptide (71%) and, lastly, the A2 peptide (54%). Only one of the EBV IgM-positive serum samples did not give an ELISA signal with any of the peptide mimics; however, this sample gave a low signal in the commercial EBV diagnostic test.
A combination of any of the peptides resulted in an increase in the sensitivity (Table 1), since each peptide effectively recognized a different subset of antibodies, thus indicating that antibodies to different epitopes are being detected. The best combination of two peptides for the recognition of EBV IgM antibodies was gp125 and A3, which achieved a maximum sensitivity of 97.5%. Overall, the sensitivity was generally higher by combining three peptides, but the use of all four peptides in combination did not further improve the sensitivity.
Improving the diagnostic sensitivity by conjugation of peptides to BSA. In order to improve the presentation of peptides on a solid phase for diagnostic ELISA, we selected two of the peptides, gp125 and F1, for conjugation to a carrier molecule, BSA. A larger panel of serum samples was analyzed for reactivity with gp125-BSA and F1-BSA conjugates; the panel included a higher number of putatively cross-reactive IgM-positive serum samples (Fig. 5). To correct for background levels, the reactivity to BSA alone was also measured; and these values were subtracted from the BSA-peptide readings. There was a dramatic improvement in sensitivity when the peptides were conjugated to BSA (Table 2). For example, by using the F1 peptide alone, only 12 of 41 OD readings were >0.4; however, when F1 was attached to BSA, 59 of 64 of the positive OD readings were >0.4. Both BSA-conjugated peptides failed to recognize eight serum samples that were VCA IgG positive and/or EBNA positive and VCA IgM negative, typical of past EBV infection, again demonstrating the high degrees of specificity of the mimotopes for the recognition of IgM antibodies in clinical samples. Furthermore, there was little cross-reactivity with other viral IgM antibodies and rheumatoid factor, which is a common cross-reactive species in current commercial diagnostic tests.
DISCUSSION
Diagnosis of EBV-related diseases relies on detection of EBV-specific antibodies in the serum of infected individuals. Current commercial tests usually use native antigens purified from EBV-producing cultures to capture serum antibodies. As possible replacements for these complex antigens we have generated EBV antigen mimics by selecting peptides from a phage display library. In this study we demonstrated that our panel of peptides can detect IgM antibodies typically produced in human subjects after infection with EBV. One of the peptides was selected against a commercial MAb to the gp125 antigen of VCA, since this antigen is known to be diagnostically important. The other peptides (peptides F1, A2, and A3) were isolated by screening the library against novel MAbs, generated by immunizing mice with the crude EBV antigen; and although the exact antigen that the corresponding peptides mimic is not known, they all appear to be diagnostically important.
The highest sensitivity of detection of IgM in EBV-positive sera was achieved by using the F1 peptide alone (88%), and this was further improved when F1 was coupled to the carrier molecule BSA (95%). It is perceivable that F1-BSA could be used alone for diagnosis; however, further characterization of the identity of the epitope and, importantly, a much larger set of clinical samples must be screened before combinations of peptides can be eliminated from our studies.
A high sensitivity of detection was also achieved with the A3 peptide (85%). Although the identity of the A3 epitope is also unknown, these data indicate that the antigen that this peptide mimics is important diagnostically. Future studies are planned to reveal the identity of this antigen. The sensitivity for the gp125 peptide was slightly less than 71%, and again, this was significantly improved when the peptide was coupled to the carrier protein BSA (92%). It is known that this epitope on VCA is important diagnostically, and the high level of sensitivity reflects this. Peptide A2 had the lowest sensitivity of detection of 54%; this peptide competed only with p18 and not with the EBV complex, indicating specificity for this antigen. Serum antibodies reactive with this epitope could be less common, resulting in the lower detection rate. The parent antibody, A2, is an IgM isotype, and this subclass tends to have lower affinity and often higher cross-reactivity than IgG subclasses. Therefore, these characteristics could also be reflected in the corresponding A2 peptide.
Exposure to EBV infection will stimulate the immune system to generate antibodies to multiple exposed surface epitopes, and individual serum samples are likely to have antibodies to several immunodominant regions of EBV. An effective diagnostic method would aim to cover as many of these immunodominant epitopes as necessary to recognize every EBV-positive sample. Each peptide effectively represents a different epitope and will therefore recognize a different subset of antibodies; therefore, use of a combination of the peptides should lead to greater coverage of these important epitopes and an increase in the sensitivity of detection. Our results showed that this phenomenon was observed, as any peptide combination increased the sensitivity (Table 1).
The particular peptide conformation that antibodies will preferentially recognize is not predictable; therefore, prospective diagnostic peptides need to be tested in a variety of formats. The ELISA plates that we used in this study were specifically designed for immobilization of peptides; the free amino groups were adsorbed via an ethylene glycol spacer. Our preliminary experiments showed that this was preferable to adsorption directly onto microtiter plates (data not shown). However, when peptides are displayed on phages, they may adopt a more rigid conformation than the free peptide, as they are anchored to the pIII protein at the C terminus, leaving the N terminus free. We therefore chose to couple two of the peptides via the C terminus to BSA to resemble their orientations on the phages. We assume that the combination of an increased number of mimotopes attached to BSA and the fact the configuration of the mimotope was somehow more accessible to antibody binding both contributed to the increase in diagnostic sensitivity. Experiments are in progress to optimize solid-phase coupling of all of the peptides to BSA alone or in combination to achieve the greatest sensitivity. Other solid-phase formats such as latex or nitrocellulose, which are used in rapid diagnostic tests, are also being investigated.
It is also unlikely that all 20 amino acid residues of the peptides will be involved in binding, and it is possible by alanine scanning to identify the residues essential for binding (11). Once the essential amino acids have been identified, nonessential residues could be removed, resulting in shorter peptides that could also be combined. A similar technique was used for a peptide representing the p18 epitope. A "mixotope" representing a collection of mimotopes was synthesized and improved the sensitivity and specificity of detection of EBV antibodies in clinical samples (22).
When the optimal assay format has been selected, a more accurate assessment of the diagnostic value of these peptides with a larger panel of positive, negative, and putative cross-reactive sera can be completed. Here we have shown that when the F1 and gp125 peptides are attached to BSA, they do not recognize sera from patients with past infections, which are typically characterized by elevated VCA IgG and/or EBNA levels but which have a low VCA IgM titer. In a larger study it would be beneficial to study the ability to identify reactivation infections or the phase of infection (acute or convalescent) of a given patient to evaluate the true diagnostic value of these mimotopes.
In this study the peptides appeared to be EBV-specific mimotopes and did not cross-react with the sera IgM positive for CMV, parvovirus, varicella-zoster virus, HSV, or RF. However, since one of the purposes of a peptide ELISA is to increase specificity by eliminating potentially cross-reactive epitopes present in the full-length protein or antigen extract and in remaining contaminants in the purified protein, if these peptides are to be developed for a diagnostic test, further screening against all typically cross-reactive species is required. The most common cross-reactions of current EBV diagnostic kits are rheumatoid factor and other herpesviruses. Routine diagnostics have been developed to remove cross-reacting rheumatoid factor by preabsorption with anti-Ig. Therefore, it would be a major advantage if the mimotopes would prove to have a higher specificity than existing commercial tests and would not require the removal of rheumatoid factor. We have shown here that the F1 and gp125 mimotopes do not recognize 6 serum samples positive for rheumatoid factor and 34 cross-reactive serum samples, and in this study there were no false-positive results (100% specificity). However, further analyses with a much larger panel of putatively cross-reactive sera in a follow-up study are required to confirm this finding.
To select for further peptides that could recognize serum antibodies that were previously undetectable by our peptides, the library could be screened against this patient's serum. Since MAbs represent only a small portion of the immune response to the antigen, polyclonal sera may cover different epitopes simultaneously, including the epitopes missed in our experiments. Our preliminary data have shown that many peptides can be selected by this approach (data not shown) and could also lead to the identification of peptides not only specific for EBV IgM antibodies but also specific for EBV IgG antibodies. Other authors have successfully used this approach for selection of diagnostic peptides for Lyme disease (10). Furthermore, this simple process of identifying generic peptides could be applicable to any infectious disease for which antibodies are available. In addition, this technology does not necessarily require knowledge of the original ligand (6), and recently, an antigen thought to be involved in metastatic disease progression was identified by using peptide mimics against circulating antibodies in the serum of cancer patients (14).
In conclusion, the data presented here indicate that our panel of peptides can be used for the detection of IgM antibodies typically generated after infection with EBV. This study proves that the use of EBV peptide mimotopes in an ELISA system has a high diagnostic potential, even though there is no prior knowledge of the antigens. This technology is also applicable to many infectious diseases and can be important for vaccine development if the peptides effectively mimic protective antibodies (3, 13, 20, 25). It also has significant advantages over existing commercial tests for EBV, as the peptides are of relatively low cost, highly stable, and easily synthesized. Moreover, peptides can mimic carbohydrate epitopes present on many EBV antigen complexes and allow a focus on single subspecificities, thereby avoiding dilution with noninformative epitopes.
ACKNOWLEDGMENTS
We thank Joan Hoogenraad and Rosella Masciantonio (La Trobe University) for technical assistance and Edward Kachab (PanBio Ltd) for providing reagents.
This study was supported by the Cooperative Research Centre for Diagnostics (Australia).
Present address: Uniseed Pty. Ltd., Cumbrae-Stewart Building, Research Road, University of Queensland, St Lucia, QLD 4072, Australia.
REFERENCES
Adda, C., L. M. Tilley, R. F. Anders, and M. Foley. 1999. Isolation of peptides that mimic epitopes on a malarial antigen from random peptide libraries displayed on phage. Infect. Immun. 67:4679-4688.
Bharadwaj, M., A. Suhrbier, A. Elliott, and D. J. Moss. 2000. Epstein-Barr virus and associated cancers. Today's Life Sci. 13:43-45.
Casey, J. L., A. M. Coley, R. F. Anders, V. J. Murphy, K. S. Humberstone, A. W. Thomas, and M. Foley. 2004. Antibodies to malaria peptide mimics inhibit Plasmodium falciparum invasion of erythrocytes. Infect. Immun. 72:1126-1134.
Chan, K. H., R. X. Luo, H. L. Chen, M. H. Ng, W. H. Seto, and J. S. M. Peiris. 1998. Development and evaluation of an Epstein-Barr virus (EBV) immunoglobulin M enzyme-linked immunosorbent assay based on the 18-kilodalton matrix protein for diagnosis of primary EBV infection. J. Clin. Microbiol. 36:3359-3360.
Faber, I., P. Wutler, P. Wohlrabe, H. Wolf, W. Hinderer, and H. H. Sonneborn. 1993. Serological diagnosis of infectious mononucleosis using three anti-Epstein-Barr virus recombinant ELISAs. J. Virol. Methods 42:301-308.
Folgori, A., R. Tafi, A. Meola, F. Felici, G. Galfre, R. Cortese, P. Monaci, and A. Nicosa. 1994. A general strategy to identify mimotopes of pathological antigens using only random peptide libraries and human sera. EMBO J. 13:2236-2243.
Gartner, B. C., R. D. Hess, D. Bandt, A. Kruse, A. Rethwilm, K. Roemer, and N. Mueller-Lantzsch. 2003. Evaluation of four commercially available Epstein-Barr virus enzyme immunoassays with an immunofluorescence assay as the reference method. Clin. Diagn. Lab. Immunol. 10:78-82.
Geysen, H. M., S. J. Barteling, and R. H. Meloen. 1984. Use of peptide synthesis to probe viral antigens for epitopes to a resolution of a single amino acid. Proc. Natl. Acad. Sci. USA 82:3998-4002.
Harlow, E., and D. Lane (ed.). 1988. Antibodies: a laboratory manual, p. 139-244. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Kouzmitcheva, G. A., V. A. Petrenko, and G. P. Smith. 2001. Identifying diagnostic peptides for Lyme disease through epitope discovery. Clin. Diagn. Lab. Immunol. 8:150-160.
Li, F., A. Dluzewski, A. M. Coley, A. W. Thomas, L. M. Tilley, R. Anders, and M. Foley. 2002. Phage-displayed peptides bind to the malarial protein apical membrane antigen 1 and inhibit the merozoite invasion of host erythrocytes. J. Biol. Chem. 277:50303-50310.
Livnah, O., E. A. Stura, D. L. Johnson, S. A. Middleton, L. S. Mulcahy, N. C. Wrighton, W. J. Dower, L. K. Joliffe, and I. A. Wilson. 1996. Functional mimicry of a protein hormone by a peptide agonist: the EPO receptor complex at 2.8. Science 273:464-471.
Lundin, K., A. Samuelsson, M. Jansson, J. Hinkula, B. Wahren, H. Wigzell, and M. A. Persson. 1996. Peptides isolated from random peptide libraries on phage elicit a neutralizing anti-HIV-1 response: analysis of immunological mimicry. Immunology 89:579-586.
Mintz, P. J., J. Kim, K. A. Do, X. Wang, R. G. Zinner, M. Cristofanilli, M. A. Arap, W. K. Hong, P. Troncoso, C. J. Logothetis, R. Pasqualini, and W. Arap. 2003. Fingerprinting the circulating repertoire of antibodies from cancer patients. Nat. Biotechnol. 21:57-62.
Mitchell, J. L., C. M. Doyle, M. V. Land, and P. L. Devine. 1998. Comparison of commercial ELISA for detection of antibodies to the viral capsid antigen (VCA) of Epstein-Barr virus (EBV). Disease Markers 13:245-249.
Motti, C., N. Maurizio, A. Meola, G. Galfre, F. Felici, R. Cortes, A. Nicosia, and P. Monaci. 1994. Recognition by human sera and immunogenicity of HbsAg mimotopes selected from an M13 phage display library. Gene 146:191-198.
Oldenburg, K. R., D. Loganathan, I. J. Goldstein, P. G. Schultz, and M. A. Gallop. 1992. Peptide ligands for a sugar-binding protein isolated from a random peptide library. Proc. Natl. Acad. Sci. USA 89:5393-5397.
Pearson, G. R. 1988. ELISA tests and monoclonal antibodies for EBV. J. Virol. Methods 21:97-104.
Puntoriero, G. A., A. Meola, A. Lahm, S. Zucchelli, B. B. Ercole, R. Tafi, M. Pezzanera, M. U. Mondelli, R. Cortese, A. Tramontano, G. Galfre, and A. Nicosia. 1998. Towards a solution for hepatitis C virus hypervariability: mimotopes of the hypervariable region 1 can induce antibodies cross-reacting with a large number of viral variants. EMBO J. 17:3521-3533.
Scala, G., X. Chen, W. Liu, J. N. Telles, O. J. Cohen, M. Vaccarezza, T. Igarashi, and A. S. Fauci. 1999. Selection of HIV-specific immunogenic epitopes by screening random peptide libraries with HIV-1 positive sera. J. Immunol. 162:6155-6161.
Smith, G. P., and J. K. Scott. 1993. Libraries of peptides and proteins displayed on phage. Methods Enzymol. 217:228-257.
Tranchand-Bunel, D., C. Auriault, E. Diesis, and H. Gras-Masse. 1998. Detection of human antibodies using ‘convergent’ combinatorial peptide libraries or ‘mixotopes’ designed from a nonvariable antigen: application to the EBV viral capsid antigen p18. J. Peptide Res. 52:495-508.
Van Grunsven, W. M. J., W. J. M. Spaan, and J. M. Middeldorp. 1994. Localization and diagnostic application of immunodominant domains of the BFRF3-encoded Epstein-Barr virus capsid protein. J. Infect. Dis. 170:121-127.
Van Regenmortel, M. H. V. 1998. Synthetic peptides help in diagnosing viral infections. ASM News 64:332-338.
Zhong, G., G. P. Smith, J. Berry, and R. C. Brunham. 1994. Conformational mimicry of a chlamydial neutralization epitope on filamentous phage. J. Biol. Chem. 269:24183-24188.(Joanne L. Casey, Andrew M)