Hev b 9, an enolase and a new cross-reactive allergen from Hevea latex and molds

Purification, characterization, cloning and expression


K. Hoffmann-Sommergruber, Department of Pathophysiology (formerly General and Experimental Pathology), AKH-EBO 3Q, Waehringer Guertel 18–20, A-1090 Vienna, Austria, Fax: + 43 1 40400 5130, Tel.: + 43 1 40400 5116, E-mail: Karin.Hoffmann@akh-wien.ac.at


Natural rubber latex allergy is an IgE-mediated disease that is caused by proteins that elute from commercial latex products. A complementary DNA (cDNA) coding for Hev b 9, an enolase (2-phospho- d-glycerate hydrolyase) and allergen from latex of the rubber tree Hevea brasiliensis, was amplified by PCR. The PCR primers were designed according to conserved regions of enolases from plants. The obtained cDNA amplification product consisted of 1651 bp and encoded a protein of 445 amino-acid residues with a calculated molecular mass of 47.6 kDa. Sequence comparisons revealed high similarities of the Hevea latex enolase to mold enolases that have been identified as important allergens. In addition, the crucial amino-acid residues that participate in the formation of the catalytic site and the Mg2+ binding site of enolases were also conserved. Hevea latex enolase was produced as a recombinant protein in Escherichia coli with an N-terminal hexahistidyl tag, and purified by affinity chromatography. The yield amounted to 110 mg of purified Hev b 9 per litre of bacterial culture. The recombinant allergen bound IgE from latex, as well as mold-allergic patients, in immunoblot and ELISA experiments. The natural enolase was isolated from Hevea latex by (NH4)2SO4 precipitation and ion exchange chromatography. The natural and the recombinant (r)Hev b 9 showed equivalent enzymatic activity. Patients’ IgE-antibodies preincubated with rHev b 9 lost their ability to bind to natural (n) Hev b 9, indicating the identity of the B-cell epitopes on both molecules. Cross-reactivity with two enolases from Cladosporium herbarum and Alternaria alternata was determined by inhibition of IgE-binding to these enolases by rHev b 9. Therefore, enolases may represent another class of highly conserved enzymes with allergenic potentials.

nHev b 9

natural Hev b 9


natural rubber latex


polyvinylidene difluoride

rAlt a 5

recombinant Alt a 5

rCla h 6

recombinant Cla h 6

rHev b 9

recombinant Hev b 9

IgE-mediated type 1 allergy to natural rubber latex (NRL) is caused by proteins present in diverse latex products. Latex is the milky sap produced by the laticiferous cells of the tropical rubber tree Hevea brasiliensis of the family Euphorbiaceae. Sensitization occurs via the skin, mucosal or wound contact, or inhalation of airborne allergens released from powdered latex gloves [1]. NRL-allergy has been acknowledged to impose serious health problems, especially among occupationally exposed individuals, e.g. health care workers and individuals who have to undergo repeated surgeries such as patients with spina bifida [2,3].

In NRL a variety of IgE-binding proteins with molecular masses ranging from 5 to 110 kDa were detected by two-dimensional immunoblotting [4–6]. Since 1993, a total of 10 NRL allergens have received an international nomenclature designation. With the exception of Hev b 4, full coding sequences are known for all other designated latex allergens [7–9] (S. Wagner & E. Ganglberger, unpublished data) . Some of these latex allergens have been identified in latex only, while others cross-react with homologous proteins present in fruits [10]. These cross-reactivities have been termed the ‘latex–fruit syndrome’. Immunological cross-reactivity has been reported between allergens from latex and banana, avocado, chestnut, peanut, papaya, passion fruit, fig, melon, pineapple, kiwi, and tomato [11]. In addition, cross-reactivities between allergens from NRL and Ficus benjamina, sweet pepper, grass, and weed pollen have also been described [12–14].

Prohevein from Hevea latex (Hev b 6.01) and its homologous proteins in vegetables and fruits have been implicated in cross-reactivity of latex with sweet pepper, chestnut, banana, peach, and kiwi [13]. Prohevein belongs to the group of chitin-binding proteins with a hevein domain. This group also includes plant class I chitinases [15]. Hevein (Hev b 6.02), the N-terminal 4.7-kDa domain of prohevein, shares epitopes with class I chitinases from avocado, chestnut, and banana, and represents one cause for their cross-reactivity [16–18]. Another protein that may play a role in the cross-reactivity of latex with pollen and fruits is Hev b 8, the latex profilin. Profilin was the first allergen to be named a ‘panallergen’ because of its high cross-reactivity [19]. Latex profilin was detected in a nonammoniated latex extract [20] and a certain amount of cross-reactivity to pollen proteins was shown by inhibition studies [14]. Hev b 2, a β-1,3-glucanase, was separated from nonammoniated latex and described as a latex-allergen. Hev b 2 was recognized by several atopic patients, indicating a possible cross-reactivity [21].

Based on data obtained by immunoprobing two-dimensional Hevea latex protein blots with sera from latex allergic patients and by protein microsequencing, several proteins from NRL have been described as putative allergens [4]. Two IgE-reactive protein spots shared over 90% sequence identity with the N-terminus of Ricinus communis enolase (SWISS-PROT P42896). Enolase (2-phospho- d-glycerate hydrolyase) is a highly expressed key enzyme of glycolysis and gluconeogenesis that is needed for the conversion of glycerate 2-phosphate to phosphoenolpyruvate and vice versa. Enolase occurs as a homodimer in all eukaryotic and many prokaryotic cells. Owing to its essential role as a cytoplasmic household enzyme, enolase is a highly conserved enzyme. It has been identified and characterized from diverse sources ranging from bacteria to higher vertebrates. Enolases are described as important allergens in several molds. An enolase was first identified as an allergen in Saccharomyces cerevisiae, but IgE cross-reactivity exists between enolases of Cladosporium herbarum (Cla h 6), Alternaria alternata (Alt a 5), Candida albicans, Aspergillus fumigatus, and Fusarium solani[22–25].

Here, we report the nucleotide and deduced amino-acid sequence of the new latex allergen Hev b 9, which is an enolase. Further, we describe the cloning and expression of the recombinant molecule in E. coli and its biological activity. Characterization as an 48-kDa latex allergen was shown by IgE-immunoblot and IgE-inhibition experiments with the natural Hevea enolase isolated from latex C-serum. The results from IgE-immunoblots and ELISA-inhibition experiments performed with Hevea latex enolase and the mold allergens Alt a 5 and Cla h 6 indicate a new area of cross-reactivity of latex allergens.

Patients, material and methods

Isolation of total RNA from latex

Fresh latex was collected from regularly tapped Malaysian rubber trees (H. brasiliensis, clone RRIM 600). Latex exuding from the tapped trees was collected while continuously mixing it with an equal volume of RNA extraction buffer (0.1 m Tris/HCl, 0.3 m LiCl, 1 m m EDTA, 10% SDS, pH 9.5) at ambient temperature. This mixture was centrifuged at 100 000 g for 30 min at 15 °C. The rubber plug was removed, the aqueous phase extracted with chloroform/phenol and total RNA was precipitated with LiCl.

RT-PCR amplification and sequencing of Hev b 9

First-strand cDNA synthesis was performed by reverse transcribing 0.5 µg of total latex RNA using a modified oligo(dT) primer, T25NN (5′-GGAGAAGGA(T)25(A/G/C)N-3′). For the first-round PCR 2 µL of the first-strand cDNA synthesis and for the second-round PCR 2 µL of the previous amplification reaction were used to amplify part of the coding sequence for Hev b 9. Based on sequence similarities among Ricinus communis enolase and enolases from other plants, degenerate primers were designed. The primers used to amplify a partial sequence of the Hev b 9 cDNA were Eno6 (5′-TT(C/T)AA(A/G)GA(A/G)GC(A/C/T)ATGAAGATGGG-3′) and T25NN for first-round PCR and Eno5 (5′-ATGCAGGA (A/G)TT(C/T)ATGAT(C/T)CT(C/T)CC-3′) and T25NN for the following PCR. The second-round PCR products were ligated into the pGEM®-T easy vector (pGEM®-T Easy Vector System, Promega, Wisconsin, USA) and propagated in XL-1 Blue E. coli cells according to the vendor’s instructions. The PCR products were sequenced using the Thermo Sequenase fluorescent labelled primer cycle sequencing kit (Amersham Life Science, Little Chalfont, UK) with the fluorescent labelled primers fT7 forward and fM13 reverse flanking the vector’s multiple cloning site.

To complete the missing 5′ end of the cDNA, the Marathon™ cDNA Amplification Kit (Clontech, Palo Alto, CA, USA) was used according to the vendor’s instructions. Complementary DNA synthesis was performed using the cDNA primer from the kit, and following first and second-round PCR with the kit’s cDNA adaptor primers and the gene-specific internal primers Eno9 (5′-CAGCTTTAGCAATGGCAGTCTT-3′) and Eno7 (5′-CAACATTTGTTGCATCCTGACC-3′). The second-round PCR products were cloned and sequenced as described above.

The complete cDNA sequence and its deduced amino-acid sequence were compared to the European Molecular Biology Laboratory (EMBL) and SWISS-PROT databases.

Construction of the expression plasmid

The complete coding sequence of Hev b 9 was amplified by RT-PCR from total latex RNA. The primers used in the PCR were Eno10 (5′- AGGTACCGCGATTACCATTGTCTCC-3′, KpnI site underlined) and Eno11 (5′- ACAGCTGCTAATAGGGTTCGACAGG-3′, PvuII site underlined) according to the 5′ and 3′ end of the latex Hev b 9 cDNA determined in the previous experiments. The resulting 1339 bp long amplification product was first digested with SalI, overhanging ends filled in with Klenow enzyme (Boehringer Mannheim, Germany), subsequently digested with KpnI and ligated into the pQE-30 expression vector (Qiagen, Hilden, Germany). The pQE-30 expression vector provides a sequence coding for a hexahistidyl (6 ×His) affinity tag at the 5′ end of the Hev b 9 coding sequence. The expression construct pQE-30/Hev b 9 was transformed into competent M15[pREP4]E. coli cells.

Expression of rHev b 9 in E. coli

For the expression of rHev b 9, 1 L 2 × tryptone/yeast culture medium containing both ampicillin and kanamycin was inoculated with 10 mL of an overnight culture and grown with vigorous shaking at 37 °C to an D600 of 0.8. Expression of rHev b 9 was induced by adding isopropyl thio-β- d-galactoside (IPTG) to a final concentration of 2 m m. After growing for an additional 16 h at 37 °C, the cells were harvested and cell pellets were frozen at −20 °C.

Purification of rHev b 9 under nondenaturing conditions

Cell pellets were thawed for 15 min on ice and resuspended in 30 mL lysis buffer (50 m m NaH2PO4, 300 m m NaCl, pH 8.0). Lysozyme (Merck, Darmstadt, Germany) was added to a concentration of 1 mg·mL−1 and the mixture was incubated on ice for 1 h. The highly viscous lysate was treated with DNaseI at a concentration of 200 µg·L−1 on ice for 1 h. To pellet cellular debris the lysate was centrifuged for 20 min at 4 °C at 20 000 g. The supernatant was loaded onto a packed column of 3 mL of 50% w/v Ni/nitriloacetic acid agarose (Qiagen, Hilden, Germany). The column was washed with 20 mL of lysis buffer and 10 mL of wash buffer (same composition as lysis buffer plus 20 m m imidazole). Purified rHev b 9 was eluted with lysis buffer containing 100 m m imidazole. The course of the purification was monitored by SDS/PAGE on 12% gels under reducing conditions. The separated proteins were visualized by staining with Coomassie brilliant blue R-250 (Bio-Rad Laboratories, Richmond, CA, USA).

Enzymatic activity of Hev b 9

Enolase enzyme assay was performed as described previously [26]. Enolase was added to 1 mL of assay mixture (20 m m Na2HPO4, pH 7.4, 400 m m KCl, 0.01 m m EDTA, 2 m m 2-phospho- d-glycerate). The enzymatic activity was determined at room temperature by transferring the mixture into a quartz cuvette and measuring the absorption at A240 for 3 min. For the determination of the specific activity an enolase from S. cerevisiae (Sigma-Aldrich, Steinheim, Germany) with a specific activity of 44 U·mg−1 protein was used in the enolase enzyme assay. One unit is defined as the conversion of 1 µmol of 2-phosphoglycerate to phosphoenolpyruvate per minute at pH 7.4 at 25 °C.

Cloning of enolases from Cladosporium herbarum and Alternaria alternata

C. herbarum enolase (Cla h 6) was isolated by immunological screening of a cDNA expression library as described previously [27]. Using the coding sequences of C. herbarum enolase as a probe, the A. alternata cDNA expression library was screened for a positive clone [27].

Expression and purification of rCla h 6 and rAlt a 5

Allergen coding sequences were cloned into pMW172 [28] via NdeI/EcoRI in the case of Cla h 6 and NdeI/HindIII in the case of Alt a 5. Expression of the respective constructs was performed as previously described [27]. E. coli lysates of the recombinant proteins were analysed by SDS/PAGE, Coomassie staining, and IgE immunoblots [27]. Subsequently, the recombinant enolase of A. alternata was purified to near homogeneity, whereas the recombinant enolase of C. herbarum was enriched by an inclusion body preparation [25].

Preparation of latex C-serum

Fresh Hevea latex was collected in chilled containers from rubber trees (H. brasiliensis RRIM 600 clone) and then centrifuged at 44 000 g at 4 °C for 1 h. The aqueous phase, the C-serum, was collected and centrifugation repeated. The clear aqueous C-serum was freeze dried and stored at −20 °C.

Isolation of natural Hev b 9 from latex C-serum

Two hundred milligrams lyophilized latex C-serum was dissolved in 20 m m Tris/HCl, pH 7.5. (NH4)2SO4 was added to the solution to 25% saturation and stirred for 30 min at 4 °C. The precipitated material was centrifuged and collected. (NH4)2SO4 was added to the supernatant to 50% saturation, in a final step to 75% saturation, and treated as described above. The three precipitates were dissolved in 20 m m Tris/HCl, pH 7.5, and tested for enolase activity. The redissolved pellet of the 75% saturation was desalted in 20 m m Tris/HCl, pH 7.5, by passing through a PD-10 column (Pharmacia, Uppsala, Sweden). The fraction collected in the void volume was applied to a MonoQ HR5/5 column (Pharmacia, Uppsala, Sweden), and elution performed using a linear gradient of 0–0.5 m NaCl in 20 m m Tris/HCl, pH 7.5. Each fraction was analysed for enolase activity and protein contents by SDS/PAGE.

Fraction number 9 was separated on a 8% SDS/PAGE gel and blotted onto a poly(vinylidene difluoride) membrane. The membrane was stained with Coomassie brilliant blue R-250 (Bio-Rad Laboratories, Richmond, California). The 48-kDa band was excised and N-terminal sequencing was performed using the 491 Protein Sequencer of Procise Protein Sequencing System (PE Applied Biosystems, Foster City, CA, USA).

Study population

A total of 122 individual serum samples from adult patients with positive case histories, positive radioallergosorbent tests (RAST classes ≥ 2), and characteristic type 1 allergic reactions was used in this study. Of these sera, 103 were obtained from latex allergic patients, 7 from patients with characterized allergy to latex and A. alternata or C. herbarum, and 12 from A. alternata or C. herbarum allergic patients.

The serum pool comprised selected sera from five latex allergic and five latex and mold allergic patients. As a control group, six serum samples from nonallergic patients were used.

Immunoblot analysis

Volumes of 2 µg nHev b 9, rHev b 9, rAlt a 5, and rCla h 6 per lane were separated on 12% SDS/PAGE gels, blotted onto nitro-cellulose membranes and treated with blocking buffer (40 m m Na2HPO4, 7 m m NaH2PO4, 0.5% BSA, 0.05% w/v sodium azide, 0.5% w/v Tween-20, pH 7.5). Membrane strips were then incubated with patients’ sera diluted in blocking buffer 1 : 5 overnight at 4 °C. After a washing step the blots were incubated with 125I-labeled anti-(human IgE) Ig (IBL, Hamburg, Germany) overnight. Binding of patients’ IgE to the allergens was visualized on Biomay™ MS film (Kodak, New York, USA).

For inhibition of IgE binding to nHev b 9, a pool of Hev b 9 positive sera was incubated overnight at 4 °C with 50 µg and 100 µg purified rHev b 9. The preincubated sera were then used to probe the nitrocellulose strips containing nHev b 9. Incubation of the serum-pool with 100 µg purified rHev b 10 was used as a negative control. Hev b 10 is a manganese superoxide dismutase and was produced as a recombinant protein in E. coli (S. Wagner, unpublished data).

IgE inhibition ELISA

For cross-inhibition experiments, the Hev b 9, Alt a 5 and Cla h 6 specific serum pool was diluted 1 : 6 in 150 m m NaCl/50 m m Tris, 0.05% Tween 20, 0.5% BSA, and preincubated (overnight at 4 °C) with 1, 10, 50, and 100 µg·mL−1 rHev b 9 and also rHev b 10. Microtiter plates (Nunc-Immuno Plate, Nalge Nunc International, Denmark) were coated with 4 µg·mL−1 of recombinant allergens overnight at 4 °C. After blocking with NaCl/Tris, 0.05% Tween 20, 1% BSA, preincubated sera were applied onto the coated plates and incubated overnight at 4 °C. After washing five times with 150 m m NaCl/50 m m Tris and 0.05% Tween 20, the plates were incubated with alkaline phosphatase conjugated mouse anti-(human IgE) Ig (Pharmingen, San Diego, CA, USA) for 1 h at 37 °C, and for another hour at 4 °C. Color development was performed using 0.1% azinoethylbenzthiozoline sulfonic acid substrate (Sigma-Aldrich, Steinheim, Germany) and the optical density was measured at 405 nm (550 nm as reference wavelength) after 30 min using an ELISA-reader (Dynatech MR 7000, Alexandria, VA, USA).


Cloning and sequence analysis of Hev b 9

Degenerate primers designed according to conserved regions of enolases from plants were used to amplify a partial sequence (coding for amino-acids 182–445) and the complete noncoding 3′ end of latex enolase cDNA by RT-PCR from total latex RNA. Using a protocol to amplify cDNA ends by PCR, the complete cDNA clone was obtained. The Hevea latex enolase cDNA comprises 1651 bp, containing an open reading frame from bases 30–1367. T-rich sequences dominate the last 100 bp, and the complete polyadenylation signal consists of an aggregate of multiple elements. A TTGTAA motif representing an upstream efficiency element was found 48 bp upstream and an A-rich region as a positioning element 15–18 bp upstream of the poly(A) sequence. Both motifs have been found frequently in plant sequences [29]. The complete cDNA sequence coding for latex enolase is available from the EMBL database under the accession number AJ132580.

The Hev b 9 cDNA sequence codes for a protein of 445 amino-acid residues corresponding to a predicted molecular mass of 47.6 kDa and a calculated pI of 5.6 ( Fig. 1). The deduced amino-acid sequence of Hev b 9 was compared to sequences in the SWISS-PROT database. Hev b 9 displays an overall identity of 72% to the human enolase, and approximately 60% to various fungal enolases, e.g. 60% to S. cerevisiae enolase, 62% to C. herbarum enolase, Cla h 6, and 60% to A. alternata enolase, Alt a 5, respectively ( Fig. 1). Sequence analysis of enolases from S. cerevisiae and H. brasiliensis suggested a common active site architecture [30–32]. The crucial amino-acid residues for enolase activity have been defined in S. cerevisiae as Glu 211, Lys 345, and His 373. The corresponding amino-acid residues in Hevea enolase are Glu 215, Lys 352, and His 380 ( Fig. 1). These three amino-acid residues are part of conserved regions and Lys 380 is contained within an enolase ‘fingerprint’ motif, which is highly conserved in all known enolases [31]. The three-dimensional structure determination of S. cerevisiae enolase reveals the presence of two Mg2+ in the active site. In S. cerevisiae enolase Ser 39, Asp 246, Glu 295, and Asp 320 are complexed to the magnesium ions, corresponding to Ser 42, Asp 250, Glu 300, and Asp 327 in Hevea enolase ( Fig. 1).

Figure 1.

Sequence analyses of Hev b 9. Hev b 9 shows 60% identity to A. alternata enolase (Alt a 5), 62% to C. herbarum enolase (Cla h 6), 60% to S. cerevisiae enolase, and 72% identity to the human beta enolase. Identical amino-acid residues are indicated by dashes. Important active site residues are marked in grey, the positions of the metal binding ligands shown by arrows, and the enolase ‘fingerprint’ motif is boxed.

Expression and purification of rHev b 9

The coding region of Hev b 9 (excluding the N-terminal methionine) was ligated into the expression vector pQE-30, and rHev b 9 was produced as a recombinant His-tagged protein. After induction of transformed M15 E. coli with 2 m m IPTG for 16–20 h, cells were harvested, lysed by addition of 1 mg·mL−1 lysozyme (Merck, Darmstadt, Germany) and rHev b 9 was detected by SDS/PAGE. The recombinant protein was purified by Ni/nitriloacetic acid affinity chromatography and purification was analysed by SDS/PAGE ( Fig. 2). The yield of purified, soluble rHev b 9 was 110 mg per litre culture.

Figure 2.

Purification of rHev b 9 and isolation of nHev b 9. Protein fractions at different stages of purification were loaded onto 12% SDS/PAGE and stained with Coomassie brilliant blue R-250. Molecular mass markers in kDa as indicated are shown in lane M, total protein extract from E. coli harboring pQE 30 without an insert in lane 1, total protein extract from E. coli harboring pQE 30 with the rHev b 9 coding sequence inserted in lane 2, and elution fraction from Ni/nitriloacetic acid agarose column in lane 3. Proteins from latex C-serum precipitated by 75% (NH4)2SO4 are shown in lane 4, and nHev b 9 purified by ion-exchange chromatography in lane 5.

Enolase activity

Purified recombinant Hev b 9 and S. cerevisiae enolase as a standard with a specific activity of 44 U·mg−1 enolase were used for the enzyme assay. Comparing the increases of absorption·mg−1·min−1 of the enolases resulted in a specific activity of 22 U·mg−1 enolase for rHev b 9.

Isolation of natural Hev b 9 and its biochemical and immunological analysis

Natural Hev b 9 was isolated from latex C-serum. Proteins of the C-serum were separated by (NH4)2SO4 precipitation. Subsequently, each fraction was tested for enolase activity. Specific activity was present within the 75% (NH4)2SO4 fraction. The proteins of this fraction were further separated by ion exchange chromatography, collecting 500 µL samples. Among five samples with specific enolase activity, the fraction with the highest activity (20 U·mg−1 protein) contained a distinct component migrating at 48 kDa together with additional components on SDS/PAGE ( Fig. 2, lane 5). N-terminal sequencing of the 48-kDa protein revealed 15 amino-acid residues, matching the N-terminus without a methionine residue of the Hev b 9 sequence.

A serum pool of latex allergic patients was used for immunoblot and inhibition studies. No difference in IgE binding to nHev b 9 was observed when compared to purified rHev b 9 ( Fig. 3, lanes 1 and 2). In order to demonstrate that rHev b 9 was able to inhibit IgE binding to nHev b 9, we examined the effect of preincubating the pooled sera with soluble rHev b 9 before adding the sera to blotted nHev b 9. At concentrations of 50 µg·mL−1 rHev b 9, IgE binding to nHev b 9 was significantly reduced and almost completely abolished when 100 µg·mL−1 were applied ( Fig. 3, lane 3 and 4). No inhibition could be observed when the serum pool was preincubated with 100 µg rHev b 10 ( Fig. 3, lane 5).

Figure 3.

IgE immunoblots and inhibition experiments with recombinant and natural Hev b 9. Blotted recombinant (lane 1) and natural Hev b 9 (lane 2) were probed with a serum pool of latex-allergic patients. For IgE inhibition, blotted natural Hev b 9 was probed with the serum pool preincubated with 50 µg (lane 3) or 100 µg (lane 4) recombinant Hev b 9, and 100 µg recombinant Hev b 10 (lane 5). Molecular mass markers in kDa are indicated. For controls nHev b 9 was probed with a normal human serum pool (lane N), and a buffer control without addition of serum (lane B).

IgE binding to rHev b 9, rAlt a 5, and rCla h 6

The importance of H. brasiliensis enolase as an allergen was assessed by direct binding to the recombinant allergen by IgE of latex and mold allergic patients. Sixteen out of 110 latex-allergic patients, that is 14.5%, revealed IgE binding to rHev b 9 in immunoblots ( Fig. 4A, lanes 1–16). According to case histories, four of these patients were also allergic to molds (lanes 13–16). Out of 12 patients with characterized allergy to A. alternata and C. herbarum, three showed IgE reactivity to rHev b 9 ( Fig. 4A, lanes 17–19). No IgE-binding could be detected when using a serum pool of six nonallergic individuals or a buffer control ( Fig. 4A, lane N and B). The designation Hev b 9 for the latex enolase was submitted and granted from the IUIS allergen nomenclature committee.

Figure 4.

IgE immunoblots with rHev b 9 (A),rAlt a 5 (B) and rCla h 6 (C). Blotted recombinant proteins were probed with sera from single patients, with a normal human serum pool (lane N) or a buffer control without addition of serum (lane B). Lanes 1–12 show IgE reactions of latex allergic patients, lanes 13–16 display patients with latex and mold allergy, and lanes 17–19 IgE reactions of mold allergic patients.

All 19 patients’ serum samples with IgE reactivity to rHev b 9 were further tested for IgE cross-reactivity to rAlt a 5 and rCla h 6. In the case of rAlt a 5, all 19 sera displayed specific IgE antibodies ( Fig. 4B), in the case of rCla h 6 all sera except serum number 8 displayed IgE antibodies reacting with the recombinant allergen ( Fig. 4C). All three recombinant allergens displayed different IgE binding capacities.

ELISA inhibition studies with rHev b 9, rAlt a 5, and rCla h 6

For cross-inhibition studies, 10 sera from patients displaying IgE reactivity to rHev b 9, rAlt a 5, and rCla h 6 were used as a serum-pool in ELISA experiments (sera from patients 1, 2, 5, 7, 9, 14–18 from Fig. 4). The three recombinant enolases were coated (4 µg·mL−1) and tested with aliquots of the serum-pool after preincubation with increasing amounts of rHev b 9 and rHev b 10 as a negative control. Recombinant Hev b 9 was able to diminish about 60% of IgE binding to rHev b 9 at a concentration of 1 µg·mL−1 and up to 90% at concentrations over 10 µg·mL−1. Furthermore, rHev b 9 was able to reduce IgE binding to Alt a 5 µp to 19% and to Cla h 6 µp to 24% at concentrations over 50 µg·mL−1( Fig. 5). Preincubation of the serum-pool with equivalent amounts of purified recombinant Hev b 10 did not influence IgE binding to rHev b 9 ( Fig. 5).

Figure 5.

IgE inhibition ELISA. Recombinant Hev b 9, rAlt a 5, and rCla h 6 were coated at 4 µg·mL−1. Inhibitions were determined with a Hev b 9-specific serum pool after preincubation with different concentrations of rHev b 9 or rHev b 10.


We report the cloning and sequencing of Hev b 9, a new and cross-reactive Hevea latex allergen. The Hev b 9 cDNA encodes a cytoplasmic latex protein of 445 amino-acid residues with a predicted molecular mass of 47.6 kDa and a calculated pI of 5.6. The protein displays 62% identity with Cla h 6, the enolase of the mold C. herbarum, and 60% identity with Alt a 5, the enolase of the mold A. alternata ( Fig. 1). We have produced Hev b 9 as a recombinant His-tagged protein in E. coli and demonstrated that the molecule possesses IgE reactivity equivalent to nHev b 9 ( Fig. 3). Furthermore, cross-reactivity with enolases from molds was tested using IgE-inhibition ELISA.

Enolase is a highly expressed key enzyme of glycolysis and gluconeogenesis and a highly conserved enzyme in all eukaryotic and most prokaryotic cells. Total latex RNA was used to amplify the cDNA coding for the Hevea enolase. The deduced amino-acid sequence of the enolase from H. brasiliensis shows an identity in the range of 60% to enolases from several molds. The important amino-acid residues responsible for the active site of enolases have been described for the enolase of S. cerevisiae. Interestingly, sequence alignments of enolases showed that these amino-acid residues are located in conserved regions [31], which are also present in the Hev b 9 sequence.

Enolases are described as allergens in several molds. Alt a 5, the enolase from A. alternata, is the second most important allergen of this mold, and recognized by 22% of all A. alternata allergic patients [25]. The same has been reported from the enolase of C. herbarum, which is also the second most important allergen of this mold, with 22% of the patients reacting positively [25,33]. In addition, a 46- to 48-kDa enolase is also a main allergen and cross-reacting component of C. albicans, S. cerevisiae, and C. utilis[22,24,34]. Comparing several mold protein extracts, a 46-kDa band appeared as the only dominant band present in all the protein extracts [34]. Therefore, enolases are not only important allergens but also predominant proteins in mold extracts.

Two-dimensional electrophoresis and N-terminal sequences lacking the N-terminal methionine provided the first evidence that enolase is an allergen of Hevea latex [4]. We have produced Hev b 9 as a recombinant His-tagged protein in E. coli and were able to purify rHev b 9 from E. coli cells at a yield of 110 mg per litre culture medium.

The biological activity of rHev b 9 was tested with biochemical and immunological assays to confirm that the recombinant protein was equivalent to its natural counterpart and correctly folded. Therefore, in a first step the enzymatic activity of rHev b 9 was tested in an enolase assay. In this assay, the enzyme catalyzed the conversion of glycerate 2-phosphate to phosphoenolpyruvate and the extinction coefficient of the produced phosphoenolpyruvate was measured. Recombinant Hev b 9 showed an activity of 22 U·mg−1 enolase in this enzyme assay. Thus, the recombinant protein possessed an active site because of its correct folding. Furthermore, enolases are described to be enzymatically active only as dimers with identical subunits [32]. Our rHev b 9 was enzymatically active, indicating its ability to form homodimers.

In latex C-serum natural Hev b 9 is a minor component. Detection of IgE binding to Hev b 9 in latex C-serum is difficult because a high percentage of IgE reactivity is directed to proteins in the 43–50 kDa range [5,6]. Therefore, we isolated natural Hev b 9 from C-serum. The protein showed enzymatic activity of 20 U·mg−1 protein, which is comparable with the activity found for the recombinant protein. Both natural and recombinant Hev b 9 bound IgE from a serum-pool of latex allergic individuals with comparable intensities in immunoblots. Furthermore, rHev b 9 was able to abolish IgE binding to nHev b 9 when the serum-pool was preincubated with the recombinant protein ( Fig. 3). Consequently, the highly purified and standardized recombinant protein is a superior tool compared to the natural protein.

We tested the IgE binding capacity of the recombinant Hev b 9 by IgE immunoblots using 110 individual serum samples from latex-allergic patients. Sixteen patients (14.5%) revealed IgE-binding to the recombinant Hev b 9. Four out of these 16 patients were also allergic to molds. Three out of 12 mold allergic patients showed IgE-binding to rHev b 9. All rHev b 9 positive patients were also tested for their IgE-binding to rAlt a 5 and rCla h 6. Except one patient ( Fig. 4, lane 8), all others showed IgE reactivity to the mold enolases ( Fig. 4), but IgE binding intensities to the three recombinant enolases were different when tested in IgE immunoblots. The cross-inhibition ELISA studies confirmed these results, that rHev b 9 was only able to diminish IgE binding to rAlt a 5 and rCla h 6 in the range of 19 and 24%, respectively. From these data we conclude that the three enolases share a number of epitopes. It has been described that enolases from molds contain several common IgE binding epitopes, but inhibition occurs only partially [24]. This implies that the mold enolases also have unique epitopes. Therefore, it becomes evident that sensitization to Hev b 9 may occur from Hevea latex and does not represent an epiphenomenon of sensitization to the homologous mold allergens.

Some Hevea latex allergens seem to have a high potential for cross-reactivity, including well-conserved proteins such as manganese superoxide dismutases, chitinases, profilins, and enolases. Sequence comparisons, three-dimensional structure determinations and cross-inhibition assays could help to predict immunogenic regions of the individual proteins. These cross-reactivities often occur with proteins from several fruits and were described as the ‘latex–fruit syndrome’[11]. The latex allergens involved in this syndrome are often defence-related enzymes such as Hev b 2 and class I chitinases [21]. In contrast, enolases are house-keeping enzymes that have been described as important allergens of several molds. Another important cross-reactive allergen from molds is a manganese superoxide dismutase [35,36]. MnSOD from Hevea latex is also a described allergen and has been designated Hev b 10 (IUIS Allergen Nomenclature Subcommittee: Official list of allergens available at ftp://biobase.dk/pub/who-iuis/allergen.list). Although Hev b 9 is a minor latex allergen, its cross-reactivities and the possible cross-reactivities of Hev b 10 to mold allergens indicate the possible existence of a ‘latex–mold-syndrome’ that will have to be confirmed in the future by the use of the relevant recombinant allergens.


The complete cDNA sequence coding for the latex enolase is available from the EMBL database under the accession number AJ132580

This work was supported in part by the Austrian Science Funds grant P12838-GEN (to H. B.) and by the Austrian National Bank grant 6965 (to K. H. S.)


  1. Enzymes: enolase (EC; superoxide dismutase (EC