| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Institut für Genetik und Allgemeine Biologie, Universität Salzburg, Salzburg A-5020, Austria
b EMBL-Heidelberg D-69012, Germany
c Novartis Forschungsinstitut, A-1235 Vienna, Austria
d Hewlett-Packard Analytical Group, Waldbronn D-76337, Germany
e Institut für Allgemeine und Experimentelle Pathologie, Universität Wien, Vienna A-1090, Austria
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: major birch pollen allergen • Bet v 1 mutants • IgE binding epitopes • allergen-specific T cell clones • hypoallergens
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
receptors. Cross-linking of mast cell-bound IgE antibodies by antigen (allergen) represents the signal for the release of preformed and newly synthesized inflammatory mediators and chemotactic substances, leading to the typical immediate-type allergic reactions (type I hypersensitivity). In general, antibodies recognize conformational epitopes (discontinuous) of antigens, which consist of residues that are not contiguous in the primary structure but are brought together at the protein surface by folding of the polypeptide chain (for reviews, see refs 1 and 2). The production of allergen-specific IgE requires the ‘help’ of allergen-specific T lymphocytes, which are stimulated by linear peptide fragments of antigen. These peptides are created by antigen-presenting cells through antigen processing and are displayed at the cell surface with molecules of the MHC complex (3, 4). T cell clones (TCC) established from atopic individuals with specificity for allergens could characteristically be attributed to the T helper 2 (Th2) subset (high level of IL-4 and IL-5 production), whereas TCC specific for bacterial antigens produced high levels of IFN-
upon stimulation, thus belonging to the Th1 subset (5). It has been shown that successful specific immunotherapy (SIT) is associated with a modulation of the immune response to allergens at the level of Th cells (6, 7). Recombinant allergens represent promising tools for diagnosis and therapy of type I allergy. The value of these molecules for diagnosis has been evaluated in detail, and a panel of recombinant allergens is available for routine diagnosis of certain inhalant allergies (8). The use of recombinant allergens for SIT, on the other hand, has always been hampered by the high allergenicity (IgE binding capacity) of these molecules, which of course is a desired property for a sensitive diagnosis, but leads to allergic side effects during treatment (9). In a recent publication, we proposed the use of hypoallergenic isoforms of allergens for SIT (10). These isoforms differ only in a few amino acids, but they are differentially recognized by the human immune system. Some of these isoforms display reduced allergenicity, because they obviously lack IgE-epitopes, but show good T cell antigenicity, a prerequisite of tolerance induction. Injection of these recombinant allergens could therefore modulate the allergic immune response at the level of the T helper cell, but with a substantially decreased risk of anaphylactic side effects during treatment.
In this study, we demonstrate that this goal can also be achieved by producing allergen mutants with the desired properties regarding allergenicity and antigenicity. By using a method developed to predict functional residues in proteins (11), we obtained a list of most common residues likely to influence IgE binding to Bet v 1, the major birch pollen allergen. Based on the results of this analysis, we introduced point mutations at critical positions in the sequence of Bet v 1, resulting in a molecule that was recognized by allergen-specific T lymphocytes, but display a significantly reduced IgE binding capacity.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Internal mutagenic primers: A30 5'ACCTTTGGAAcGAGATTATC3'; A57 5'TCGGGAAGtTGATCTTCTT3'; A112 5'GATGGAGGATgCATCTTGAA3'; A113 5'GATGGAGGATCCgTCTTGAA3'; A112+113 5'GATGGAGGATgCgTCTTGAA3'; A125 5'ACCAAAGGTAaCCATGAGGT3'; D30 5'ACCTTTGGAAaGAGATTATC3'; D57 5'TCAGGAAAGcTGATCTTCTT3'; D112 5'GATGGAGGATcCGTCTTGAA3'; D113 5'GATGGAGGATGCaTCTTGAA3'. Bases exchanged are indicated in lowercase.
Flanking primers for both Bet v 1a and Bet v 1d: Bet-Ncol primer 5'GGGCCATGGGTGTTTTCAATTACGA3', Ncol site, is underlined; Bet-EcoRI primer 5'GGGAATTCTTAGTTGTAGGCATCGGAGTGTGCCAA3', EcoRI site, is underlined. To generate a Bet v 1a construct carrying six-point mutations, the following approach was used: initially two mutants were created, each carrying three-point mutations. The first construct contained three mutations at the 5' half of Bet v 1a cDNA (amino acid positions 10, 30, and 57) and the second construct contained three mutations at the 3' half (amino acid positions 112, 113, and 125). After digestion with EcoRI and BglII, resulting in the release of the 3' half of each mutated pMW175/Bet v 1 construct, the mutated fragments of each construct were ligated and used to transform the Escherichia coli strain BL21.
The amino acid exchange at position 10 in the Bet v 1a (A10 primer, Thr 
Pro) was engineered by one-step PCR-mediated mutagenesis, by using the following flanking primers: A10 5'GGGCCATGGGTGTTTTCAATTACGAAACTGAGACCcCCTCTGTT3', base exchange is indicated in lowercase, NcoI site is underlined; Bet-EcoRI 5'GGGAATTCTTAGTTGTAGGCATCGGAGTGTGCCAA3', EcoRI site is underlined.
PCR was performed with 1 ng template (pMW175/Bet v 1a or pMW175/Bet v 1d constructs) and 1 mM of each primer (NcoI mutagenic primer and EcoRI primer for one-step reactions, and an internal mismatch primer plus a flanking primer for the first reaction of the three-primer mutagenesis method), by using 30 cycles of 1 min denaturation at 95°C, 1 min annealing at 42°C, and 1 min extension at 72°C. The second PCR step was performed with an aliquot of the first reaction and 1 mM of the flanking primer, under the same conditions described above. The PCR products were digested with NcoI and EcoRI and subcloned in the pMW175 expression vector. The resulting plasmids were used to transform competent E. coli BL21 cells. All PCR amplified products were sequenced according to the dideoxi chain termination method (17).
Expression and preparation of E. coli cell lysates of allergen mutants
For expression of the pMW175/Bet v 1a and pMW175/Bet v 1d mutant constructs, competent E. coli strain BL21(DE3) was transformed and selected on plates containing 100 mg/l ampicillin (14). A single transformant colony was picked and grown to an OD600 of 1.0. Isopropyl-ß-D-thiogalactopyranoside was then added to a final concentration of 1.0 mM, and incubation continued for 6 h at 37°C. After expression, cells were harvested by centrifugation and pellets resuspended in 50 mM Tris-HCl, pH 7.5, containing 220 mM NaCl (buffer A). The cells were then disrupted by freezing in liquid nitrogen, followed by thawing at 37°C. This step was repeated twice. Wild-type recombinant Bet v 1a (rBet v 1a) and mutants A2, A4, A5, and A6 were recovered in the supernatant after centrifugation at 30,000 x g/25 min/4°C, which was then used for immunoblot analysis. Wild-type rBet v 1d and mutants A1, A3, A1–6, D1, D2, D3, and D4 were recovered in the pellet by a low speed centrifugation. Insoluble inclusion bodies were solubilized in buffer A containing 6 M urea and dialyzed at 4°C against buffer A. After dialysis, the extracts were used for immunoblot experiments. We previously (18) showed by quantitative ELISA inhibition experiments that denaturation of Bet v 1 in 6 M urea and renaturation by dialysis does not affect IgE binding.
SDS-PAGE and immunoblots
E. coli lysates of allergen mutants were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) according to the method of Laemmli (19), by using 15% acrylamide gels. Proteins were visualized by staining with Coomassie brilliant blue R-250. As the level of expression varied for each isoallergen variant construct, Coomassie-stained gels were analyzed on a GS-670 imaging densitometer (Bio-Rad, Hercules, Calif.) and peak areas were measured. In this way it was possible to calculate the amount of lysate to achieve equivalent amounts for each mutant in immunoblot experiments.
For immunoblot analysis, proteins were separated by 15% SDS-PAGE and electroblotted (20) onto nitrocellulose membranes. IgE immunoblots were performed as described previously (21). Bound IgE was by detected by using 125I-rabbit anti-human IgE (Pharmacia, Uppsala, Sweden). Immunoblots using monoclonal anti-Bet v 1 antibodies, BIP 1 and BIP 4 (22), were performed as previously described (15). E. coli lysates harboring the plasmid without an insert were used as a control. In all experiments, reagents and cell lysates were from identical batches and were used in the same concentrations. Autoradiography was performed at -70°C for 12–48 h with intensifying screens.
Protein purification
Natural Bet v 1 (nBet v 1) was purified from aqueous extract of birch (Betula verrucosa) pollen (Allergon, Engelholm, Sweden) by reversed phase high-performance liquid chromatography (HPLC) (C8 Hypersil WP 300, 10 µm, 30x250 mm) by using a linear gradient of 2-propanol at room temperature (solvent A: aqueous 0.1% trifluoroacetic acid; solvent B: aqueous 90% 2-propanol/0.1% trifluoroacetic acid; gradient 0–70% B within 60 min; flow rate 5.0 ml/min). UV absorbance was monitored at 280 nm. Fractions (2 ml) were collected, dried in vacuo, and resuspended in water. Aliquots were analyzed by SDS-PAGE and immunoblots using monoclonal antibodies and allergic patients' sera.
Wild-type and mutant proteins were purified by using the method described for rBet v 1a (18), with few modifications. Briefly, bacterial lysates were prepared in 25 mM imidazole-HCl buffer containing 6 M urea (buffer C). The extracts were loaded onto a PBE-94 exchanger (Pharmacia) column equilibrated with buffer C, at room temperature. Bound proteins were eluted with aqueous 12.5% (v/v) polybuffer 74-HCl (Pharmacia), pH 4.0, containing 6 M urea. Fractions containing Bet v 1 proteins, as determined by dot blot immunoassays with monoclonal antibodies, were pooled and subjected to reversed phase HPLC, as described previously (18). Protein concentrations of all purified allergen preparations were determined by the micro-Kjeldahl method, by using glycine as standard (23).
NH2-terminal sequence analysis
E. coli lysates of Bet v 1 mutants were separated by 15% SDS-PAGE and blotted onto polyvinylidene difluoride membrane (Millipore, Bedford, Mass.). Bands corresponding to Bet v 1 mutants were excised and proteins were eluted by incubation with 40% (v/v) acetonitrile, 30% (v/v) trifluoracetic acid for 1 h at room temperature. Samples were vacuum dried, resuspended in water, and sequenced with the HP G1005A protein sequencing system (Hewlett-Packard, Palo Alto, Calif.). Protein samples were loaded directly onto the hydrophobic part of the biphasic sequencer column and subjected to Edman degradation by using the manufacturer's Routine 3.0 chemistry.
MALDI-TOF mass spectrometry
For determining intact molecular mass, we applied 0.5 µl (approx. 0.5 µg) of purified allergens, 0.5 µl protein calibration standard (horse heart myoglobin, average molecular weight 16,950.9) and 0.5 µl of a saturated solution of gentisic acid in 0.1% trifluoroacetic acid (matrix) to the target slide with intermittent drying in an airstream. Samples were analyzed with the Kompact MALDIT III mass spectrometer from Kratos Analytical in the linear flight mode. For calibrating the instrument, we used molecular peaks of myoglobin and gentisic acid were used.
Circular dichroism
Circular dichroism (CD) spectra of aqueous allergen solutions were recorded as described previously (24). Briefly, we used cuvettes with a path length of 0.1 cm for collecting the sample CD spectra on a Jasco J-710 spectropolarimeter (Japan Spectroscopic Co., Tokyo), with a response time of 0.25 s and a data point resolution of 0.2 nm. Each spectrum represents an average of five scans.
IgE inhibition assay
The relative ability of the Bet v 1/1–6 variant to interact with IgE antibodies was determined by competitive inhibition in a dot blot assay as follows: the serum from each allergic patient was preincubated overnight at 4°C with various concentrations (0.01–10 µg/ml of serum diluted 1:10) of each purified rBet v 1a and Bet v 1/1–6 mutant protein. Afterward, the solutions were added to nitrocellulose strips containing immobilized purified rBet v 1a (0.5 µg/dot). Bound IgE was detected by using 125I-labeled rabbit anti-human IgE (Pharmacia). Autoradiograms were analyzed on an imaging densitometer (model GS-670; Bio-Rad Laboratories, Richmond, Calif.).
Allergen-specific T cell clones
Allergen-specific Th clones were established from the peripheral blood of seven different individuals suffering from birch pollen allergy according to protocols described (25). Specificity and epitopes recognized by TCC were assessed as previously described (25). Briefly, 2 x 104 T cell blasts were incubated in triplicate cultures in 96-well plates in the presence of 105 autologous irradiated peripheral blood mononuclear cells (PBMC) with purified nBet v 1. Microtiter plates were then incubated for 48 h at 37°C, 5% CO2, humidified atmosphere. After pulsing for 16 h with [3H]TdR, cells were harvested and the incorporated radioactivity was measured by scintillation counting. When the stimulation index (ratio between cpm obtained in cultures with TCC plus autologous irradiated PBMC plus antigen and cpm obtained in cultures containing TCC and PBMC alone) was greater than 10, responses were considered positive. nBet v 1-positive TCC were then tested for reactivity for rBet v la, A1–6 mutant, and overlapping Bet v 1a-peptides, respectively. In all experiments, we used identical concentrations of allergen (1 µg/well 5 µg/ml).
Skin testing with purified rBet v 1a and Bet v 1/1–6 variant
A skin prick test (SPT) was performed according to guidelines described (26). For SPT, 25 µl of each recombinant allergen solution (final concentration of 0.01 µg–10 µg/ml sterile saline) and commercially available test solutions (birch pollen extract, grass pollen extract, histamine, 0.9% sodium chloride) was applied. Skin testing of birch pollen-allergic individuals with recombinant Bet v 1 molecules was approved by the Ethics Committee of the University of Vienna, Austria.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Therefore, for a global analysis of patterns of amino acid substitutions in all tree pollen isoallergens and their IgE binding activities, we used an algorithm recently developed by Casari et al. (11) to predict functional residues in proteins. The basis of this ‘sequence space’ method is a translation of sequences in vectors. Each sequence is represented as a vector point in a multidimensional space (sequence space), with residue positions and residue type as the basic dimensions. At each position in the sequence, a summary is given of which amino acid type occurs at what frequency in the protein family—in our case tree pollen group I allergens. By using this method it is possible not only to define protein subfamilies, but also to trace at the same time the principal components back to the individual residues and positions characteristic of the different groups of sequences. The differences among isoforms can be detected by this algorithm by defining a ‘functional direction’ in the sequence space. In this case, functional axes that separate isoforms according to their allergenicity were defined manually by assigning +1 for high IgE binding and -1 for low/no IgE binding activity.
For the multiple sequence alignment, we used a total of 14 closely related sequences of tree pollen isoallergens: Aln g 1 (28), four isoforms of Cor a 1 (27), and nine isoforms of Bet v 1 (12, 13). From the combined information of the IgE reactivity patterns of 9 tree pollen-allergic patients to the 14 isoallergens, we obtained the following list of the 8 most common residues/amino acid positions (given in a decreasing order of ‘predictive power’) likely to influence IgE binding: positions 113, 57, 125, 112, 30, 18, 133, and 82. We chose to substitute by site-directed mutagenesis the amino acid residues occurring in positions 113, 57, 125, 112, and 30 of a high IgE binding isoform (Bet v 1a) by those present in the same positions of low IgE binding isoforms. Similarly, the amino acid residues found in these positions of a low IgE binding isoform (Bet v 1d) were substituted by amino acid residues present in the corresponding positions of a high IgE binding isoform (Bet v 1a). All isoallergen mutants presented here and their corresponding amino acid substitutions are schematically shown in 
Fig. 1.
|
Expression, purification, and characterization of site-specific allergen mutants
Seven Bet v 1a mutants (A1–A6, and A1–6; 
Fig. 1) and four Bet v 1d mutants (D1–D4; Fig. 1), all obtained by site-directed mutagenesis, were expressed in E. coli as nonfusion proteins. High levels of expression ranging from 10 to 100 mg protein per liter of bacterial culture were obtained. Immunoblot analysis showed that all Bet v 1 mutants were recognized by a rabbit serum raised against rBet v 1a, and also by BIP1, a monoclonal antibody raised against natural Bet v 1 (22), indicating that the BIP1 epitope is not disturbed by any of the point mutations introduced. However, the substitution of a serine residue at position 112 by a cysteine in Bet v 1a (A3 mutant) abolished binding of BIP4 (22), another monoclonal antibody raised against natural Bet v 1. BIP4 bound weakly to wild-type rBet v 1d, and no significant differences were observed in reactivity to the Bet v 1d mutants (data not shown).
The A1–6 mutant was selected for experiments concerning activation of T lymphocytes from allergic individuals. For this purpose, A1–6 mutant, wild-type rBet v 1a, and natural (pollen) Bet v 1 were purified to homogeneity, as determined by SDS-PAGE (data not shown). The integrity of the primary structure of the purified recombinant proteins was evaluated by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry.
NH2-terminal sequence analysis of all Bet v 1 mutants showed that initiating methionine was removed; in the case of Bet v 1 A6 and A1–6 mutants, the substitution of a threonine by a proline residue at position 10 was also verified. To confirm all engineered amino acid exchanges at the protein level, the mutant allergens were further analyzed by MALDI-TOF mass spectrometry. The molecular mass of intact proteins was measured as follows (theoretical molecular mass based on the cDNA sequences is given in parentheses): wild-type rBet v 1a 17,437.3 (17,439.62), A1 17,386.6 (17,391.58), A2 17,464.9 (17,466.63), A3 17,453.5 (17,455.65), A4 17,424.9 (17,425.60), A5 17,438.0 (17,438.62), A6 17,436.7 (17,435.63), A1–6 17,418.3 (17,415.61), wild-type rBet v 1d 17,419.2 (17,417.68), D1 17,464.1 (17,465.72), D2 17,388.8 (17,390.67), D3 17,403.2 (17,401.65), D4 17,430.4 (17,431.70). Thus, the primary structure of all isoallergens mutants obtained by site-directed mutagenesis was confirmed by mass spectrometry analysis.
In Fig. 2,
the normalized CD spectra of wild-type rBet v 1a, wild-type rBet v 1d, and the A1–6 mutant proteins are displayed. The overall secondary structure content of the three proteins was very similar. Only in the far UV part (200–210 nm) did the spectra show a slight shift of the baseline crossing toward lower wavelengths, referring to a small change in the ratio of the secondary structure elements.
|
Influence of single amino acid substitutions on IgE binding
The IgE binding activity of six different Bet v 1a mutants containing single amino acid substitutions compared to the wild-type Bet v 1a sequence was evaluated by immunoblot analysis using sera from 14 allergic patients. Figure 3
shows the IgE binding patterns of five representative birch pollen-allergic patients with each isoallergen mutant. In general, each of the single amino acid substitution in Bet v 1a efficiently decreased IgE binding for some, but not for all, patients tested. For example, the A2 mutant was not recognized by patients 1 and 3, but for patient 5, the IgE binding activity of A2 was comparable to wild-type rBet v 1a. Two contrasting patterns can be seen for patients 3 and 4, where in one case every single amino acid exchange abolished IgE binding; in the other, none of the substitutions significantly reduced IgE binding. Similar results were obtained with the Bet v 1d mutants ( Fig. 3). Each of the single amino acid substitutions could effectively restore IgE binding activity for some, but not all, patients tested. For example, D1 and D2 mutants strongly bound IgE from patient 1, but no binding was observed for patients 3 and 5.
|
These results indicate that each of the single amino acid substitutions introduced by site-directed mutagenesis has dramatic effects on the IgE binding properties of tree pollen isoallergens. However, the pattern of recognition of each modified isoallergen among allergic individuals is not uniform.
Influence of multiple amino acid substitutions on IgE binding
Sera from 28 birch pollen-allergic patients were tested in immunoblots for IgE reactivity to the A1–6 mutant. The A1–6 mutant combines the six amino acid exchanges that were individually tested for Bet v 1a (see Fig. 1). Figure 4
shows the IgE binding activity of 13 representative allergic individuals to wild-type rBet v 1a (panel A) and A1–6 mutant (panel B). All patients tested displayed extremely low IgE binding activity to the A1–6 mutant. The residual IgE binding activity of the A1–6 mutant in immunoblots was calculated for each patient by taking the wild-type rBet v 1a as a reference (100% IgE binding activity) ( Fig. 4C).
|
The ability of the A1–6 mutant to inhibit binding of IgE to rBet v 1a was determined by comparing the amount of A1–6 mutant or rBet v 1a required to give 90–100% inhibition of IgE binding to immobilized rBet v 1a. Figure 5
shows the inhibition curves for six sera. Similar inhibition patterns were observed for all sera. To reach 90–100% inhibition of IgE binding the A1–6 mutant required 50- to 100-fold more protein than rBet v 1a.
|
The ability of the A1–6 variant and rBet v 1a to elicit cutaneous reactions was evaluated for 11 birch pollen-allergic patients by SPT ( Table 1).
With increasing concentrations (0.01–10 µg/ml) of the A1–6 protein, seven patients showed negative skin reactions with all concentrations applied. Four patients (patients 1, 5, 6, and 10) showed weak reactivity only at the highest concentrations used. When compared to rBet v 1a, the A1–6 mutant showed a 100- to 1000-fold reduced IgE reactivity in vivo. A monosensitized grass pollen-allergic patient (control 1) and a nonallergic individual (control 2) showed negative skin reactions with both rBet v 1a and the A1–6 mutant.
|
These results indicate that the amino acid positions 10, 30, 57, 112, 113, and 125 all seem to be important for the in vitro and in vivo IgE binding activity of the Bet v 1 allergen.
Activation of allergen-specific T cell clones
Nine Bet v 1-specific TCC established from the peripheral blood of birch pollen-allergic patients were tested for reactivity with nBet v 1, wild-type rBet v 1a, and the A1–6 mutant. All TCC reacted with natural Bet v 1, and 8/9 TCC reacted with the A1–6 mutant and rBet v 1a. One TCC failed to recognize Bet v 1a and the A1–6 mutant. All TCC reacting with rBet v 1a were mapped for T cell epitopes by using peptides synthesized according to the sequence. Results of proliferation assays with allergen-specific TCC are summarized in Table 2.
Identified T cell epitopes are also shown in Fig. 1. From these experiments, we concluded that it is possible to modulate the IgE binding properties of allergens without significantly changing their recognition by allergen-specific T lymphocytes.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bet v 1, our model allergen, is the single major allergenic protein of birch pollen (12, 30). More than 95% of birch pollen-allergic individuals display IgE against Bet v 1, and 60% react exclusively to this allergen (21). Detailed studies concerning T cell epitopes of Bet v 1 showed that this allergen harbors a manifold of T cell epitopes scattered over the complete amino acid sequence (25, 31). Furthermore, allergic and nonallergic individuals display a polyclonal T cell response to the epitopes of Bet v 1 (32, 33).
Until now, precise information about IgE binding structures on the Bet v 1 molecule were not available. However, there have been several indications that IgE epitopes of Bet v 1 are conformation-dependent: 1) when screening a birch pollen cDNA library with patients' sera, only complete clones of Bet v 1 were reactive with IgE antibodies (12), 2) NH2- and COOH-terminally truncated forms of Bet v 1 expressed in E. coli exhibited no IgE binding activity (34), 3) Bet v 1 fragments obtained by chemical or enzymatic treatment were all unable to bind IgE (35), 4) 75 overlapping dodecapeptides spanning the entire Bet v 1a sequence were not able to bind patients' IgE (25).
Using a method developed to predict functional residues in proteins (11), we analyzed the deduced amino acid sequences of 14 tree pollen isoallergens and their IgE binding activities tested for nine birch pollen-allergic patients. Amino acid positions predicted to influence IgE binding were experimentally tested by using site-directed mutagenesis and expression of nonfusion recombinant proteins in E. coli. The substituting amino acid residues were chosen according to the following two criteria: 1) the residues should be present in the corresponding positions of naturally occurring isoforms, and 2) the residues should have a high correlation with either high IgE binding or with low/no IgE binding activity. For example, four different amino acids (phenylalanine, isoleucine, leucine, and valine) are found at position 30 of birch and hazel pollen isoallergens. Yet only valine showed a strict correlation with low IgE binding activity. On the other hand, high IgE binding activity correlated best with a phenylalanine residue at position 30. By using this approach, which mimics ‘in vivo site-directed mutagenesis’, we expected that disturbances of the protein fold and structural characteristics caused by residue substitutions would be kept to a minimum. This, in turn, is essential in order to properly evaluate the contribution of each amino acid residue in IgE recognition. In fact, a comparison of the pairs of exchanged residues with matrices of preferred amino acid exchanges for ‘safe’ residue substitutions (least likely to disturb the protein structure) constructed by Bordo and Argos (36) showed that none of them fall into the category of substitutions to be avoided.
In this way, a collection of 11 mutants of Bet v 1 isoallergens was generated and tested for their IgE binding properties. We found that recognition of Bet v 1 by IgE antibodies from allergic patients depended on at least six crucial amino acid residues/positions. Considering the A1–A5 constructs, none of the single substitutions seemed to more frequently affect IgE recognition by different patients. This agrees with the results obtained by the computer algorithm, as the A1–A5 positions had almost the same value of ‘predictive power’.
The analysis of IgE binding properties of hazel pollen isoforms compared with Bet v 1a suggested that the substitution of threonine at position 10 in Bet v 1a for a proline in Cor a 1/16 isoform correlated with low IgE binding activity (27). However, the computer algorithm indicated this position for only one patient from nine analyzed. Our experimental data revealed that the substitution at position 10 of Bet v 1a (A6 mutant) drastically reduced IgE binding for 87% of the patients tested. In contast, the substitution at position 10 of Cor a 1/16 dramatically increased IgE binding activity for all patients tested (unpublished data). These results are consistent with the notion that both NH2 and COOH termini of different proteins often show higher-than-average antigenic activity, possibly because the terminal residues are predominantly surface oriented and have a high relative flexibility (37, 38).
Recently, the 3-dimensional structure of one Bet v 1 isoform (ALK2227) was determined both in the crystalline state (X-ray diffraction) and in solution (nuclear magnetic resonance, or NMR) (39). This isoform differed from Bet v 1a only at amino acid position 62, where a leucine substituted a phenylalanine residue present in the Bet v 1a sequence. The structural features determined by 2- and 3-dimensional NMR measurements of Bet v 1a (40) did not differ from those obtained for the ALK2227 isoform. A combined analysis of the determined protein surface and the known sequences of tree pollen isoallergens showed three patches larger than 600 Å2 of coherently connected conserved surface residues on the determined Bet v 1 structure. These surface areas are assumed to be potential candidates for harboring cross-reactive antibody binding epitopes of tree pollen allergens (39). Residues at positions 30, 112, and 113 picked by the computer algorithm and proven to influence the recognition by IgE antibodies are located in two of these patches. Every single-point mutant of Bet v 1a displayed full IgE binding activity for more than one allergic patient tested, suggesting that the single amino acid substitutions reported here caused minimal perturbations on the overall folding and structural characteristics of the proteins. In addition, our CD measurements did not show significant differnces in the overall secondary structure content of the multiple-point A1–6 mutan when compared to wild-type rBet v 1a and rBet v Id isoforms.
Taken together, our results clearly indicate that each of the six amino acid positions is involved in the formation of IgE binding epitopes of Bet v 1. The reactivity patterns of individual patients to the single-point mutants imply that there are multiple conformation-dependent epitopes recognized differently by patients' IgE. The fact that the multiple-point A1–6 mutant in vitro exhibited extremely low IgE binding activity for all tree pollen-allergic patients tested further corroborates these assumptions. In vivo (skin prick) tests showed that the potency of the A1–6 mutant to induce typical urticarial skin reactions (wheal and flare) in allergic individuals was dramatically reduced compared to rBet v 1a and commercially available test solutions ( Table 1). Thus, data obtained from our in vitro experiments correlate well with the in vivo IgE binding activity of the A1–6 mutant and demonstrate its functional properties, leading to allergic effector mechanisms.
T lymphocytes are crucial in the induction and maintenance of IgE-mediated allergies (41–43). Moreover, it has been shown that specific immunotherapy of immediate-type hypersensitivities is operative at the level of this cell type (6). Repeated administration of high doses of allergen leads to a tolerance of allergen-specific T lymphocytes and alters the pattern of cytokine production in such a way that allergic effector mechanisms are affected (44–46). This leads to an amelioration of allergic symptoms in the patient. The A1–6 allergen mutant described in this paper was designed to possess low IgE binding capacity, but conserved T cell epitopes and thus T cell activating capacity, a precondition for immunomodulation. Actually, eight of nine TCC reacting with natural Bet v 1 also proliferated with A1–6 mutant ( Table 2), and all TCC reactive with Bet v 1a recognized the A1–6 mutant. Thus, the amino acid exchanges in the multiple-point mutant obviously did not influence T cell recognition. Even when three amino acid residues were exchanged in one T cell epitope (epitope 110–128; see Fig. 1), no influence on the recognition of this peptide by the respective TCC (clone XPC33) was observed. The clone XPC33 showed better proliferation to A1–6 mutant than to rBet v 1a. According to a previous study, this seems to be a consistent finding among clones originating from different patients that are specific for the epitope 110–128 (10). One TCC failed to recognize Bet v 1a, very likely because it is specific for a distinct Bet v 1 isoform in the natural mixture (10, 47). Likewise, this TCC did not react with the mutant A1–6.
In conclusion, we propose a new concept for the development of vaccines for SIT based on the use of site-directed mutagenesis. The molecules developed possess significantly reduced allergenicity in terms of IgE binding, and therefore will not lead to anaphylactic reactions upon injection. On the other hand, such mutants are recognized by T lymphocytes and can thus modulate the allergic immune response shown to be essential for successful therapy (6, 7). This approach is probably feasible with every allergen with known amino acid sequence, irrespective of the source (pollen, food, mites, mammalian) from which it may be derived. In fact, a similar approach was suggested for indoor allergens (48). Site-directed mutagenesis was used to target the three disulfide bonds known to stabilize Der p 2, the major house dust mite allergen (49). In another study, predictive algorithms were used to target surface residues for mutagenesis, allowing the identification of other important B cell determinants on Der p 2 (50).
As demonstrated here, the identification of structural determinants critical for IgE binding might also be facilitated by characterization of naturally occurring isoforms of allergenic proteins. It has already been shown—in most cases by cDNA cloning—that several allergenic proteins consist of a mixture of closely related isoforms. Isoforms have been described for Amb a 1, the predominant allergen from short ragweed (Ambrosia artemisiifolia) pollen (51); for Poa p 9, a group of basic isoallergens from Kentucky bluegrass (Poa pratensis) pollen (52); for Car b 1, the major allergen from hornbeam (Carpinus betulus) pollen (29); for Cor a 1, the major allergen from hazel (Corylus avellana) pollen (27), for profilin, a plant pan-allergen (53); for Phl p 1 and Phl p 5, major allergens from timothy grass (Phleum pratense) pollen (54, 55); for Pha a 5, a major allergen of canary grass pollen (56); and for Par j 1, the major allergen of Parietaria judaica pollen (57). In the case of Bet v 1, the patterns of IgE reactivity toward each cloned isoform (10) were similar for all allergic patients tested irrespective of the geographic location of the sensitizing birch trees (our unpublished observations). Thus, use of the hypoallergenic isoforms or variants produced by site-directed mutagenesis in specific immunotherapy offers the possibility of a significantly safer approach for treatment of immediate-type allergies.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Abbreviations: IgE, immunoglobulin E; TCC, T cell clone; Th, T helper cell; SIT, specific immunotherapy; PCR, polymerase chain reaction; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; nBet v 1, natural Bet v 1; rBet v 1, recombinant Bet v 1; HPLC, high-performance liquid chromatography; NMR, nuclear magnetic resonance; PBMC, peripheral blood mononuclear cells; MALDI-TOF, matrix-assisted laser desorption/ionization-time of flight; CD, circular dichroism; SPT, skin prick test.
Received for publication July 14, 1997. Accepted for publication October 10, 1997.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
secretion in specific allergen-stimulated T cell cultures. J. Immunol. 154, 4187–4194[Abstract]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or A1–6 mutant (
).