Profilins are cytoskeletal proteins that sequester G-actin and bind to membrane-associated phosphatidylinositol-4, 5-bisphosphate,1-7 thereby affecting both cell morphol- ogy and signal transduction. Recently, the existence of human profilin multimers has been reported, in which tetramers may be the high-affinity actin-binding form.8 Profilins have also been identified and purified from tree, grass and weed pollens and have been produced recombinantly.9-14 European studies have revealed that profilin isoforms isolated from various plant sources may act as pan-allergens9, 10, 15, 16 and that approximately 20% of all pollen-allergic patients display IgE reactivity to recombinant birch profilin. The recombinant birch profilin, as well as natural profilins from birch, timothy grass and mugwort, are able to elicit IgE-mediated histamine release from basophils of pollen-allergic patients. Thus, profilin may be a pan-allergen as well as being responsible for the sensitization and maintenance of a significant number of type I allergic patients.

Structural analysis has revealed remarkable homology between profilins from different plant species.9-11 Plant profilins appear to share common IgE epitopes and IgE from all sensitive patients cross-reacts with different profilins. In addition, rabbit polyclonal antibodies raised against recombinant birch profilin16 cross-react with virtually all plant profilins studied thus far. The available data indicate that profilin from one plant source can cross-sensitize an individual to several plant species and may explain why some patients with type I hypersensitivities display reactions to a wide range of distantly related pollens and foods.

Available data regarding profilin as a pan-allergen have thus far been drawn from European patients, whereas sensitivity of the North American population to plant profilin is unknown. Furthermore, the ability of plant profilin to form multimers has not been directly studied. However, the existence of multimers should be considered when investigating the allergenic properties of profilin, because larger antigens would likely present additional epitopes to elicit a greater IgE-mediated histamine release. Therefore, recombinant plant profilin multi- merization was studied and immunoassays were developed to assess IgE reactivity of individuals to plant profilin. The correlation between type I hypersensitivities and reactivity to plant profilin within the local population was subsequently examined.

The cDNA encoding an isoform of profilin derived from the pollen of Zea mays (ZmPRO1; courtesy of Dr Christopher Staiger, Purdue University, West Lafayette, IN, USA)17 was provided in a transfection vector (pET-23a; Novagen, Madison, WI, USA) with an isopropyl beta- D-thiogalactopyranoside (IPTG)-inducible promoter; polyclonal rabbit IgG that recognizes the protein product encoded by ZmPRO1 cDNA was also provided. Cyanogen bromide (CNBr)-activated sepharose 4B was purchased from Pharmacia (Piscataway, NJ, USA) and poly(L-proline) (PLP; 10000-30000 MW) was purchased from Sigma Chemical Co. (St Louis, MO, USA). Horseradish peroxidase (HRP)-conjugated monoclonal antibodies (goat antirabbit IgG, goat antihuman IgE) and silver staining kits were purchased from Pierce Chemical Co. (Rockford, IL, USA).

Human serum samples from routine blood draws were kindly provided by the University of Illinois College of Medicine at Rockford, Office of Family Practice (Rockford, IL, USA) with appropriate patient consent; patient-declared allergies were annotated on each serum container. Serum isolated from whole blood by standard centrifugational methods was either stored at 4°C (and used within 1 week) or aliquoted and stored at -20°C. The samples were categorized into one of three groups: (i) no declared allergies; (ii) miscellaneous reactions (i.e. non-plant allergens, such as dust, adhesive tape, synthetic materials etc.); or (iii) classical type I allergies to plant pollens.

Pre-thawed competent BL21(DE3) Escherichia coli cells (Novagen Inc., Madison, WI, USA) were transformed with ZmPRO1/pET-23a by a modified protocol from the manufacturer and essentially as described for plant11, 13, 18 and human profilins.19 The DNA content and quality from lysates of various transformed E. coli clones were analyzed by standard spectrophotometric measurement (i.e. 260 nm, concentration; 260 nm/280 nm, relative nucleotide purity vs protein) and agarose (0.7%) gel electrophoresis. The E. coli clones expressing the highest concentrations of ZmPRO1 cDNA were selected for profilin production.

Transformed E. coli initially grown in 10 mL L-broth (in g/L: 10 tryptone, 5 bacto yeast extract, 10 NaCl + 0.15 ampicillin) at 37°C for 10 h were brought to a final 1 L volume of L-broth and incubated for an additional 2 h (37°C, gentle mixing at 100 r.p.m.) prior to the addition of either IPTG (0.4 mmol/L final concentration) or vehicle for an additional 6 h incubation. The cultures were centrifuged (1000 g for 30 min at 22°C) to yield pellets that were resuspended in 5 volumes of ice-cold lysis buffer (0.01% Triton X-100, 2 µmol/L leupeptin, 1 µmol/L aprotinin, 0.2 µmol/L pepstatin, 5 mmol/L This-HCl, pH 7.2) and sonicated (continuous output control setting 2 × 10 s, Sonifier cell disruptor, Branson Sonic Power Co., Danbury, CT, USA). The lysates were centrifuged (12000 g for 30 min at 4°C) and supernatants poured ontoa poly( L-proline)-sepharose 4B (i.e. PLP bead) affinity column, as described previously.8, 20 Briefly, a step-wise elution gradient with urea was used to collect and purify profilin for overnight dialysis (at 4°C) against 2 mmol/L N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), pH 7.2/0.1 mmol/L CaCl₂ and concentration by centrifugation (centriplus-3, Amicon Inc., Beverly, MA, USA) to a final concentration of approximately 1 mg/mL, which was stored at -20°C. In some cases, ZmPRO1 profilin was isolated by a co-incubation of E. coli lysate:PLP bead slurry (1:4 vol; 4-16 h at 4°C, gentle shaking), followed by centrifugation to pellet and wash the profilin-PLP bead complexes (3 times with 100 mmol/L NaCl, 100 mmol/L glycine, 0.01 mmol/L DTT, 10 mmol/L Tris base, pH 7.8). The final pellet was suspended and boiled in sample buffer, with or without β-mercaptoethanol (BME).

Proteins isolated by either PLP bead slurry (e.g. Figure 1; initial purification and validation that profilin was made in E. coli) or column chromatography were analyzed by standard silver-stained sodium dodecyl sulfatepolyacrylamide gel electropheresis (SDS-PAGE; 15% acrylamide) techniques. Profilin was further characterized by western immunoblotting, as previously described.8 The immunoblot was developed by incubation with rabbit anti- ZmPRO1 primary antibody (1:1000) and goat antirabbit secondary antibody conjugated with horseradish peroxidase (1:500). Proteins were visualized with either a fluorescent substrate or enhanced metal substrate (Super Signal or metal diaminobenzidine tetrahydrochloride (DAB), Pierce Chemical Co., Rockford, IL, USA).

Purified ZmPRO1 product (50 ng/well) or a control vehicle (tris-buffered saline (TBS), pH 7.4) was added to designated wells of a 96-well immunoassay plate (Immulon-2, Dynatech Laboratories Inc., Chantilly, VA, USA) that was stored overnight at 4°C. The general sequence to develop the appropriate wells for the ELISA was as follows: (i) block non-specific sites (4% non-fat powdered milk, 0.1% bovine serum albumin (BSA), 0.02% NaN₃ in TBS, 20% SuperBlock from Pierce Chemical Co. for 2 h at 4°C); (ii) 1 × TBS wash and incubate with either serum samples or a control vehicle (TBS or heat-inactivated fetal calf serum; overnight at 4°C); (iii) discard samples and add either TBS to sample wells or primary rabbit IgG antiplant profilin (1:1000 dilution in TBS, 0.01% Tween-20, 0.01% BSA for 1.5 h) into control wells to ensure ZmPRO1 profilin coating; and (iv) 1 × TBS wash of all wells, add appropriate secondary antibodies (either goat antihuman IgE-HRP in wells that contained serum or goat antirabbit IgG-HRP in control wells; 1:500 dilutions for 2 h at 4°C). The plate was extensively washed then developed using 2,2'-azino-bis (3-ethylbenzthiazoline-6-sulfonic acid) as a colorimetric method substrate (although other established HRP substrates were also successful in preliminary work) and optical densities were measured with a microtiter platereader (λ_{570 nm}). Serial dilution assay of profilin standards indicated that the assay was linear between 0.1 and 100 ng additions of profilin per well. Serum samples were considered reactive with profilin (i.e. positive) if there was at least one standard deviation difference between the optical densities obtained from the wells containing profilin versus those without profilin. Serum identified as positive gave a linear increase in signal intensity at concentrations between 10 and 100%.

In some instances, a dot filtration immunoblot assay (Bio Rad, Hercules, CA, USA) using supported nitrocellulose (0.2 µm pore size) was a necessary alternative to the ELISA to determine the allergenic potential of ZmPRO1 monomers versus multimers. Profilin was either placed under reducing (4.5% BME at 95°C for 3 min) or non-reducing conditions (to favor monomeric or multimeric conditions, respectively) and subsequently allowed to adhere to the dot immunoblot (50 ng/well for 2 h) prior to vacuum filtration to remove the medium from the membrane. The BME was then removed by thorough washing with TBS and the dot immunoblot was developed with antibodies similar to the ELISA method, but with an enhanced metal DAB as the HRP substrate. Quantitative values for the intensity of immunorecognition (i.e. darkness) were obtained by computer scanning the dot- filtration immunoblot and using Adobe Photoshop (Adobe Systems, San Jose, CA, USA) software program (under histogram, black channel). The average brightness value was obtained from the fixed number of pixels (486) that covered each dot and the corresponding darkness value was calculated by 100 × the inverse of the brightness (i.e. increased relative value represents increased darkness or immunoreactivity). Comparisons between the means of different treatments were made by Student's t-test.21

Previous characterization of native human profilin has revealed the formation of functionally relevant tetramers.8 Immunoblot and SDS-PAGE analyses have shown that approximately 14.8 kDa human profilin monomers form dimers and tetramers. Furthermore, functional significance has been inferred from preferential actin binding to the tetrameric form. In light of the human profilin findings, the ability of plant profilin to form multimers was examined. Purified recombinant protein encoded by the ZmPRO1 cDNA was visualized by silver-stained SDS-PAGE separation (Figs. 1, 2a). A predominant band at 14.8 kDa was identified in transformed E. coli, particularly after IPTG induction. Other extraneous bands (> 30 kDa) presumably include multimeric profilin and/or cellular proteins that remain with the more convenient PLP affinity bead slurry. It is apparent that ZmPRO1 profilin separated by column chromatography yields a cleaner preparation that was also studied under both reduced and non-reduced conditions (Fig. 2a). A predominant band appeared as expected for monomeric ZmPRO1 plant profilin (approximately 14.8 kDa), but, consistent with human profilin, higher molecular weight proteins remained that were resistant to reducing agents. The band at approximately 60 kDa (Fig. 2a, + BME) suggests the formation of a tetramer resistant to reducing agents, in addition to the distinct aggregation of proteins near the top of the gel (> 97 kDa). The larger proteins become more pronounced under non-reducing conditions (Fig. 2a, - BME) and are associated with a corresponding loss of monomeric profilin. The presence of stained proteins that remained in the stacking gel (not shown) further support the idea of natural protein (i.e. profilin) aggregation/multimerization. Any faint protein represented at 14.8 kDa in the absence of reducing agent becomes more evident on development of corresponding immunoblots with sensitive substrates (Fig. 2b). However, western immunoblotting gave inconsistent positive identification for the higher molecular weight profilin multimers; presumably this is due to the different efficiencies in transfer of various protein sizes, alterations in net charge that may occur on proteinprotein interactions (e.g. ionic bonds as discussed by Babich et al.8), difficulty for > 90 kDa profilin multimers in migration from the stacking gel to the separating gel or a lack of antibody recognition due to epitope masking when profilin aggregates/multimerization occurs. Indeed, differences in the observed western immunoblot band intensities (monomer > tetramer >> higher orders) support these explanations. Collectively, the results suggest that, similar to native human profilin, immunologically distinct Zea mays profilin was produced and purified and preferentially forms multimers.

An ELISA was developed to further immunologically identify the purified recombinant protein and to provide a means to study whether human serum from allergic individuals recognizes ZmPRO1 profilin. Six representative control wells are pictured (Fig. 3), of which only those coated with the purified protein elicited a significant colorimetric response. In addition, rabbit antihuman profilin IgG did not recognize ZmPRO1 profilin (not shown), which is similar to the inability of rabbit antiplant profilin IgG to recognize human profilin.18 Thus, a method was established with a clear signal-to-noise ratio that was selective for ZmPRO1 profilin and further verified the production of immunologically distinct plant profilin.

The developed ELISA was subsequently used to measure type I allergy patient IgE recognition of ZmPRO1 profilin. Sera from more than 30 individuals were used in preliminary work to establish conditions for the ELISA and other methods. Due to depletion of many of the sample volumes during the course of assay development and the lack of signal from the negative patient pool, sera selected for assay in the final study included a greater number of miscellaneous and positive samples (eight negative, 14 miscellaneous, nine positive, shown in Table 1). There was one positive reactivity to profilin among patients with no known allergies (one of eight; but no response from > 20 negative samples in the preliminary work to establish the assay), minimal reactivity in those with miscellaneous allergies (three of 14) and significant reactivity among those declaring type I allergies to pollen (three of six). The raw data from the three samples that gave a positive reaction to profilin (Table 2) show a strong IgE reactivity to profilin with minimal background (i.e. a relatively high ratio of the optical densities when serum was added to profilin-coated wells vs non-profilin coated wells). The results agree with previous work that type I allergy patients are immunoreactive to plant profilin and indicate that the current approach may be useful to screen for patients with type I allergies.

The allergenic potential of profilin monomers and multimers was tested by dot-filtration immunoblot analysis (Fig. 4). Although the ELISA is more sensitive and quantitative, the presence of reducing agents used to favor a monomeric profilin state would hinder the adsorption of profilin to plastic wells. Thus, a dot blot filtration apparatus was used to remove the reducing agents. In all instances, a greater response was measured from IgE recognition of + BME/profilin (i.e. profilin as a predominantly multimeric form) compared with BME/profilin. In contrast to the relatively weak signal from negative control patients (i.e. serum(-)), IgE from the positive serum category displayed a significantly greater recognition of profilin (regardless of ± BME), thereby implicating higher profilin orders as allergens.

The present work describes the production and purification of recombinant ZmPRO1 profilin, which yields multimeric forms. The protein was subsequently identified to be immunologically distinct through western immunoblot analysis and an ELISA that was developed to assess allergy patient IgE reactivity. An apparent tetrameric 60 kDa profilin was found in addition to higher multimeric orders (> 97 kDa) that become amplified under non-reducing conditions. Although relatively high percentage acrylamide gels were used to visualize and study monomeric or lower multimeric orders of ZmPRO1 profilin, it appears that fastidious high profilin multimers persist, despite exposure to heat and reducing agents. The findings are remarkably similar to the results obtained on characterization of native human profilin,8 which has demonstrated tetramers to be the predominant functional multimeric form under identical non-reducing conditions. The recent identification of various plant profilin homopolymers and heteropolymers in vitro22 is also consistent with the present proposal that plant profilin multimers exist and are the relevant allergenic form. The combination of near-capacity protein loading and a relatively more sensitive SDS-PAGE staining procedure appears requisite to identify the additional protein bands, compared with typical reports with Coomassie blue protein staining (e.g. Babich et al.8). In addition, plant and human profilins may be similar in their ability to resist chemical reduction. Computer-based molecular modeling of human profilin suggests a profilinprofilin interaction may occur that protects some of the disulfide bonds from harsh reducing agents (Babich et al., unpubl. data, 1998; see also Mitterman et al.22). Modeling studies with plant profilin are needed to further the molecular comparison between plant and human profilin multimers and to identify relevant epitopes.

The existence of plant profilin multimers is supported by studies on birch pollen allergens. Induction of mouse and monkey IgE antibodies against recombinant Betv2 (a profilin form) has been demonstrated in vivo,12 albeit to a lesser extent than another pollen, Betv1. Immunoblot and SDS-PAGE analyses have revealed that Betv2 self-associates through disulfide bonds, which have been suggested to decrease the type I immune response to the protein. In contrast, the present work is consistent withthe idea that larger molecules may elicit an immune response, such as by increased epitope presentation. Perhaps the ZmPRO1 structure differs from Betv2, in that epitopes are not blocked by multimerization and/orthe quaternary structure presents novel epitopes to the immune system. Further comparisons of the allergenic potential of various plant profilins are complicated, because it is also unknown what conformation is assumed when acquired or introduced in vivo (i.e. whether they form multimers in vivo or in solution8). In support of allergenic profilin multimers, recent enzyme allergosorbent tests and immunoblot inhibition studies have revealed that IgE from birch pollen allergic patients is very reactive with 34 and 35 kDa allergens,23 of which one was thought to be a Betv1 dimer. Future analyses of known24 and perhaps alternative epitopes, combined with computer-based molecular modeling, may shed light on optimum conformations for the allergenic potential of profilin.

Several points in support of plant profilin multimerizaton may also be raised from the recent finding of a 60 kDa cross-reactive allergen from mugwort.25 Similar to human profilin tetramers:8 (i) the size is alike; (ii) it is resistant to chemical breakdown; (iii) a protein 'ladder' has been detected (e.g. approximately 15, 30, 60 kDa) that is also recognized by pollen allergic patient IgE; and (iv) the N-terminus is blocked (also supported by Babich et al., unpubl. data, 1998). Furthermore, (v) trypsin treatment abolishes IgE reactivity against profilin (monomer) and the 2830 kDa allergen (dimer?), but not against the60 kDa protein, which, if the protein were to be a profilin tetramer, is consistent with the notion of a larger protein being more allergenic; (vi) consistent with the idea that the larger 60 kDa protein represents profilin multimers, monoclonal antibodies raised against the 60 kDa allergen also recognize smaller, but not larger, proteins from three different plant extracts; (vii) pre- incubation of patient IgE with the 60 kDa allergen (IgE-inhibition experiments) reduces subsequent IgE binding to > 30 kDa proteins from plant extracts, but has little effect on IgE binding to the smaller MW species (e.g. 14-15 kDa or profilin): the latter point would also be consistent with the present notion that the 60 kDa protein is comprised of smaller proteins (e.g. 2 × 30 kDa dimers) and is more allergenic; and (viii) complementary DNA were not obtainable, which would be expected if the protein were truly a multimer(a point recognized by the investigators). Collectively, these points are consistent with the present conclusions; however, they may only represent a strong coincidence. Future studies should elucidate this issue.

The results support and extend the literature, which indicates that approximately 20% of pollen allergic individuals will show IgE reactivity to plant profilin.9, 10 Of the nine patients who declared type I allergies to plant pollens, three showed significant reactivity to ZmPRO1 profilin with the developed ELISA. Three additional serum samples from the miscellaneous category (e.g. dust) also yielded a positive response. This might be expected, considering that many type I allergy patients often include dust as an allergen; furthermore, the identification of profilin as an IgE-binding component in latex26 raises the issue that allergens considered outside the conventional spectrum of type I candidates may indeed be recognized by type I allergy patient IgE.

The relatively greater recognition of ZmPRO1 profilin multimers by IgE reveals novel aspects of plant profilin as a proposed pan-allergen. Greater recognition of profilin multimers is not due to simple additive effects, because the same amount of total profilin was added to each well used in the dot-immunoblot experiments. Thus, it appears that profilin multimers act in synergy to either sterically facilitate access to binding sites or to present unique epitopes. In the absence of a larger sample size and defined conditions to assay specifically for profilin multimerIgE interactions, the data thus far also suggest that: (i) more type I allergy patients than previously estimated have IgE that recognize profilin; and (ii) profilin multimers are causative agents for type I allergies. However, conditions to produce only monomers versus multimers first need to be defined to pursue these hypotheses. In addition, the current studies (including preliminary work) and those of others (e.g. Vrtala et al.,11, 12 Pauli et al.,14 Astwood et al.27) indicate that other methods of detection are feasible and should be considered in future basic and clinical applications. Finally, it is recognized that, in order to truly establish a correlation between type I allergies and reactivity to profilin (particularly multimers or the plant profilin source), studies with a large number of patients, carefully screened for allergies by objective methods, are required.

The current work provides a framework for future research in the areas of allergy diagnosis and treatment (e.g. Valenta and Kraft16). Further verification of profilin as a pan-allergen would raise the issue of whether profilin is the causative agent in a proportion of type I allergies, as opposed to being only a correlative agent. Conceivably, therapeutic strategies may be developed to block profilin epitopes and to prevent or reverse in vivo self-association, thereby targeting the cause of the disease.

ACKNOWLEDGEMENTS

The ZmPRO1 cDNA and rabbit IgG anti-ZmPRO1 were kindly provided by Dr Christopher Staiger (Purdue University, West Lafayette, Indiana). The excellent technical assistance of Lisa RP Foti and constructive advice from Dr Ramaswamy Kalyanasundaram (University of Illinois College of Medicine at Rockford) were appreciated. This work was funded by a grant from the American Lung Association (National and Illinois Affiliate) and awarded second place at the 1998 Midwest Student Biomedical Research Forum (Omaha, NE, USA).

REFERENCES

1
Carlsson L, Nystrom LE, Lindberg U. Crystallization of a non-muscle actin. J. Mol. Biol. 1976; 105: 353-366.
 

2
Theriot JA, Mitchison TJ. The three faces of profilin. Cell 1993; 75: 835-838.
 

3
Sohn RH, Goldschmidt-Clermont PJ. Profilin: At the crossroads of signal transduction and the actin cytoskeleton. Bioessays 1994; 16: 465-472.
 

4
Goldschmidt-Clermont PJ, Janmey PA. Profilin, a weak CAP for actin and RAS. Cell 1991; 66: 419-421.
 

5
Baatout S. Profilin. An update. Eur. J. Clin. Chem. Clin. Biochem. 1996; 34: 575-577.
 

6
Lassing I, Lindberg U. Specific interaction between phosphatidylinositol 4,5bisphosphate and profilactin. Nature 1985; 314: 472-474.
 

7
Valenta R, Ferreira F, Grote M et al. Identification of profilin as an actin-binding protein in higher plants. J. Biol. Chem. 1993; 268: 22777-22781.
 

8
Babich M, Foti LRP, Sykalik LL, Clark CR. Profilin forms tetramers that bind to G-actin. Biochem. Biophys. Res. Commun. 1996; 218: 125-131.
 

9
Valenta R, Duchene M, Ebner C et al. Profilins constitute a novel family of functional plant pan-allergens. J. Exp. Med. 1992; 175: 377-385.
 

10
Valenta R, Duchene M, Vrtala S et al. Profilin, a novel plant pan-allergen. Int. Arch. Allergy Immunol. 1992; 99: 271-273.

11
Vrtala S, Wiedemann P, Mittermann I et al. High-level expression in Escherichia coli and purification of recombinant plant profilins: Comparison of IgE-binding capacity and allergenic activity. Biochem. Biophys. Res. Commun. 1996; 226: 42-50.
 

12
Vrtala S, Mayer P, Ferreira F et al. Induction of IgE antibodies in mice and rhesus monkeys with recombinant birch pollen allergens -Different allergenicity of Bet v 1 and Bet v 2. J. Allergy Clin. Immunol. 1996; 98: 913-921.
 

13
Susani M, Jertschin P, Dolecek C et al. High level expression of birch pollen profilin (Bet v 2) in Escherichia coli: Purification and characterization of the recombinant allergen. Biochem. Biophys. Res. Commun. 1995; 215: 250-263.
 

14
Pauli G, Oster JP, Deviller P et al. Skin testing with recombinant allergens Rbet v 1 and birch profilin, Rbet v 2 -Diagnostic value for birch pollen and associated allergies. J. Allergy Clin. Immunol. 1996; 97: 1100-1109.
 

15
Valenta R, Duchene M, Pettenburger K et al. Identification of profilin as a novel allergen: IgE autoreactivity in sensitized individuals. Science 1991; 253: 557-560.
 

16
Valenta R, Kraft D. Type I allergic reactions to plant-derived food -A consequence of primary sensitization to pollen allergens. Curr. Opin. Immunol. 1995; 7: 751-756.
 

17
Staiger CJ, Goodbody KC, Hussey PJ, Valenta R, Drobak BK, Lloyd CW. The profilin multigene family of maize: Differential expression of three isoforms. Plant J. 1993; 4: 631-641.
 

18
Karakesisoglou I, Schleicher M, Gibbon BC, Staiger CJ. Plant profilins rescue the aberrant phenotype of profilin-deficient dictyostelium. Cell Mot. Cytoskel. 1996; 34: 36-47.
 

19
Gieselmann R, Kwiatkowski DJ, Janmey PA, Witke W. Distinct biochemical characteristics of the two human profilin isoforms. Eur. J. Biochem. 1995; 229: 621-628.
 

20
Janmey PA. Polyproline affinity method for purification of platelet profilin and modification with pyrone-malamute. Meth. Enzymol. 1991; 196: 92-99.
 

21
Sokal RR, Rohlf FJ. Biometry, 2nd edn. San Francisco: WH Freeman, 1981.

22
Mittermann I, Fetrow JS, Schaak DL et al. Oligomerization of profilins from birch, man and yeast. Profilin, a ligand for itself? Sexual Plant Reprod. 1998; 11: 183-191.

23
Wellhausen A, Schoning B, Peterson A, Vieths S. IgE binding to a new cross reactive structure A 35 kDa protein in birch pollen exotic fruit and other plant foods. Zeitschrift Ernahrungswissenschaft 1996; 35: 248-355 (in German).

24
Weidemann P, Giehl K, Almo SC et al. Molecular and structural analysis of a continuous birch profilin epitope defined by a monoclonal antibody. J. Biol. Chem. 1996; 271: 29915-29921.

25
Heiss S, Fischer S, Muller WD et al. Identification of a 60 kD cross-reactive allergen in pollen and plant-derived food. J. Allergy Clin. Immunol. 1996; 98: 938-947.
 

26
Vallier P, Balland S, Harf R, Valenta R, Deviller P. Identification of profilin as an IgE-binding component in latex from Hevea brasiliensis -Clinical implications. Clin. Exp. Allergy 1995; 25: 332-339.
 

27
Astwood JD, Leach JN, Fuchs RL. Stability of food allergens to digestion in vitro. Nat. Biotech. 1996; 14: 1269-1273.