INTRODUCTION

Cockroaches produce potent allergens, which give rise to IgE antibody (Ab)¹ responses in genetically predisposed individuals living in cockroach-infested housing (1-8). IgE Ab to cockroach are strongly associated with asthma, and sensitization to cockroach allergens is a major risk factor for hospital emergency room visits for asthma (5, 9, 10). Although cockroach may carry viral and bacterial pathogens, asthma is the only disease for which an unequivocal causal relationship with cockroach exposure has been established (9-11). Over the last few years, several allergens from the principal domiciliary cockroach species, Blattella germanica (German cockroach) and Periplaneta americana (American cockroach), have been cloned and insights into their biological function have been obtained (12-15). B. germanica allergen Bla g 2 shows sequence homology to aspartic proteases, whereas Bla g 4 is a ligand-binding protein or calycin (12, 13). In Taiwan, P. americana is an important cause of asthma, and a 72-kDa allergen, Per a 3, has recently been cloned, which shows homology to insect hemolymph proteins (arylphorins) (14).

We have identified an important B. germanica allergen with sequence homology to the glutathione S-transferase (GST) superfamily. These enzymes are involved in the detoxification of endogenous and xenobiotic toxic compounds and are widely distributed in most forms of life (16-18). In molecular biology, plasmid vectors have been constructed to express foreign proteins in Escherichia coli, as fusion proteins with the COOH terminus of GST from Schistosoma japonicum. These high level expression systems (pGEX vectors) have been widely used for purification of recombinant proteins by glutathione affinity chromatography (19).

Here, we report the sequence, purification, and recombinant expression of B. germanica GST, and demonstrate that this protein causes IgE Ab responses in ~70% of cockroach-allergic asthmatic patients. Homologous GST allergens were also identified in the house dust mite, Dermatophagoides pteronyssinus, and in the helminth parasite Schistosoma mansoni (20, 21), thus defining the importance of the GST superfamily in causing IgE antibody responses.

EXPERIMENTAL PROCEDURES

A B. germanica cDNA library was prepared in the UniZAP-XR phagemid expression vector as described previously (12, 13). Six cDNA clones were identified by screening the library with IgE Ab in a serum pool from eight cockroach-allergic patients (13). Further screening with individual sera from 20 cockroach-allergic patients and 4 non-allergic controls by plaque immunoassay revealed that the protein encoded by cDNA clone bg16 bound IgE Ab from 70% of the patients, but not from control sera. This protein was designated B. germanica allergen, 5 or Bla g 5, in keeping with the current allergen nomenclature (22). Double stranded sequencing of clone bg16 was carried out by dideoxynucleotide chain termination with a Sequenase kit (U. S. Biochemical Corp., Cleveland, OH) (23). The sequence was compared with the National Biomedical Research Foundation, Swiss-Prot, and GenBank data bases using FASTA, and the results revealed that the protein encoded by cDNA clone bg16 shared sequence homology with the GST. Sequence alignments were performed using the GCG program.

B. germanica GST was purified from cockroach whole body extract by chromatography over glutathione-Sepharose. A B. germanica whole body extract containing 289 mg of protein/ml was prepared from 12 g of cockroach, as described previously (24). The extract (1.3 ml) was passed over a 0.5-ml glutathione-Sepharose column (Pharmacia, Piscataway, NJ), and B. germanica GST was eluted with 10 mM reduced glutathione. After dialysis, further purification was carried out using a Superdex 75 HR10/20 size-exclusion column (Pharmacia). To assess purity, B. germanica GST was analyzed by silver-stained SDS-PAGE, using a PhastSystem (Pharmacia). The amino acid sequence of 35 NH₂-terminal residues (spanning residues 1-38) was determined by Edman degradation.

Recombinant B. germanica GST was expressed as a non-fusion protein using the pET21a expression vector (Novagen, Madison, WI). A 973-base pair DNA fragment containing the coding sequence for B. germanica GST was generated by polymerase chain reaction. Ten nucleotides were added at the 5-end, encoding the initiation Met and the first three NH₂-terminal amino acid residues (Ala-Pro-Ser) lacking in the original cDNA. Primers for polymerase chain reaction were designed as follows: - GGAATTCCATATGGCACCGTCTTATAAACTGACATAC-3 (sense), containing an NdeI restriction site, and 5-CTGTTGATTTTAGGAAGTCAT-3 (antisense), downstream from a HindIII site in the 3-untranslated region (nucleotides 922-927). The 973-base pair polymerase chain reaction-amplified DNA was ligated into NdeI-HindIII digested pET-21(a) vector. Expression of the 23-kDa B. germanica GST was induced in E. coli strain BL21(DE3) with 1 mM isopropyl-1thio--D-galactopyranoside at 30 °C, and recombinant protein was purified from cell lysates using glutathione-Sepharose (yield typically 3-4 mg/liter of culture). Recombinant B. germanica GST was analyzed by silver-stained SDS-PAGE and size exclusion chromatography. The amino acid sequence of the five NH₂-terminal residues was confirmed by Edman degradation.

IgE Ab to natural and recombinant B. germanica GST were compared using an antigen-binding RIA (25). Briefly, 10 and 20 µg of natural and recombinant B. germanica GST were labeled with 0.5 mCi of ¹²⁵I, using the chloramine T technique (specific activity, 35 and 23 µCi/µg), respectively. Serum samples diluted 1:2 and 1:10 were incubated with either ¹²⁵I-labeled allergen (~100,000 cpm added) for 4 h at room temperature. Immune complexes were precipitated overnight at 4 °C with 50 µl of sheep anti-human IgE (The Binding Site, San Diego, CA), using IgE myeloma serum (patient P. S.) diluted 1:200 as a carrier. Precipitates were counted in a -counter following washing with BBS. Quantitation of both assays was carried out using a control curve, constructed with serum from patient I. H., assigned to contain 1,000 units/ml IgE Ab.

A direct binding ELISA was also developed for measuring IgE Ab to recombinant B. germanica GST. Antigen was coated to a microtiter plate (0.5 µg/well) in carbonate-bicarbonate buffer, pH 9.6, overnight at 4 °C. Plates were washed with PBS-Tween and serum samples diluted 1:2 and 1:10 in horse serum were added for 1-2 h, after blocking with 1% bovine serum albumin, phosphate-buffered saline-Tween for 1 h at room temperature. Antibody binding was detected using peroxidase-conjugated goat anti-human IgE and the ABTS substrate system (26). The assay was quantitated using a control curve constructed with serum from patient I. H., assigned with 1,000 units/ml IgE ab.

Quantitative intradermal skin testing was carried out using serial 10-fold dilutions of B. germanica extract (1:20, w/v, Allergy Laboratories of Ohio, Columbus, OH) or purified recombinant B. germanica GST from 10⁵-10⁰ µg/ml, as described previously (12, 13, 24). Skin testing of human subjects using recombinant allergens was approved by the University of Virginia Human Investigation Committee.

Sera from 40 cockroach-allergic asthmatic patients had been obtained from patients who participated in previous studies on emergency room asthma (5, 9, 10, 12, 13). All patients had IgE Ab to cockroach by RAST (82-6, 400 units/ml, 1 unit~ of 0.1 ng IgE). Four non-allergic control subjects were also studied.

The enzymatic activity of B. germanica GST was assessed using the 1-chloro-2,4-dinitrobenzene (CDNB) substrate and compared with the activity of rat liver glutathione S-transferase (Sigma). The absorbance was monitored at 340 nm and activity was calculated using the extinction coefficient of 9.6 mM¹ (27).

RESULTS

The nucleotide and deduced amino acid sequence of cDNA clone bg16 is shown in Fig. 1. The 1,140-base pair sequence contains a 600-base pair open reading frame, encoding a 200-residue polypeptide. A stop codon (TAA) was found at position 603. NH₂-terminal sequencing of the natural protein revealed that the sequence encoding the initiation methionine and the first three amino-terminal residues (Ala-Pro-Ser) were lacking from the original cDNA clone (Fig. 2). The estimated molecular mass of the 203-amino acid residue protein was 23,176 Da and no potential N-linked glycosylation sites were found.

Sequence similarity searches revealed that the protein encoded by cDNA clone bg16 was a GST. These enzymes catalyze the nucleophilic addition of the thiol of reduced glutathione to organic compounds, and their active site contains two binding sites, one for glutathione (G-site) and one substrate binding (H-site). Five different gene families encode GST, including four classes of cytosolic enzymes referred to as alpha, mu, pi, and theta, and a class of microsomal enzymes. The highest degree of homology was found with the GST-2 subfamily of insect GST from Anopheles gambiae (51.5%), Musca domestica (housefly) (47.8%), Drosophila melanogaster (46.8%), and Manduca sexta (tobacco hornworm) (42.9%) (28-30). Significant homology was also found with other GST including human, rat, and mite GST (Fig. 3). In particular, 18 of the 26 invariable residues identified in mammalian GST by x-ray crystallography were present in B. germanica GST, including Tyr-8, Gln-63, and Asp-97, which are involved in glutathione binding at the G-site of the molecule (17, 18, 31-35).

Natural B. germanica GST was purified from whole body cockroach extract by glutathione affinity chromatography and size exclusion. The protein gave a single homogeneous band on silver-stained SDS-PAGE, with a molecular mass of 23 kDa (Fig. 4A). NH₂-terminal sequence analysis demonstrated that the natural and deduced amino acid sequences were identical except for one amino acid substitution, Phe to Cys, at position 9 (Fig. 2). This substitution occurs in a hydrophobic region of the molecule next to one of the glutathione binding residues (Tyr-8). The mosquito and tobacco hornworm sequences also have Phe at position 9. The amino acid change from Cys (TGT) to Phe is a conservative substitution. The results imply that there are other isoform(s) of B. germanica GST with the Phe substitution at position 9. The yields of natural allergen from cockroach extracts were very low (0.002% of the protein in the cockroach extract). Recombinant B. germanica GST was produced as a non-fusion protein in E. coli, using the pET21a vector system, and purified from bacterial lysates by glutathione affinity chromatography. The pure protein migrated as a single 23-kDa band on SDS-PAGE (Fig. 4B). Following size exclusion chromatography, recombinant B. germanica GST eluted as a single peak with an estimated molecular mass of 46 kDa, suggesting that the recombinant protein may form a homodimer. The enzymatic activity of recombinant B. germanica GST was evaluated using the CDNB substrate. The CDNB conjugation activity of recombinant cockroach-GST was 0.05 µM/min/mg (mean of two experiments), which was very low, compared with the rat liver GST control (13.65 µM/min/mg, 273 times higher).

Results of antigen binding RIA, using ¹²⁵I-labeled natural or recombinant B. germanica GST, demonstrated that both proteins had excellent IgE Ab reactivity in vitro. IgE Ab binding to natural and recombinant B. germanica GST was found in 27/40 (67.5%) and 29/40 (72.5%) of patients' sera, respectively. There was a good correlation between IgE Ab binding to natural and recombinant GST (r = 0.67, p < 0.05, Fig. 5A), and between the RIA and ELISA for measuring IgE Ab (r = 0.79, p < 0.001, Fig. 5B). Most patients who had IgE Ab to B. germanica GST also had IgG Ab to this allergen, as assessed by RIA. Only 3/40 patients had IgE Ab with no IgG Ab, and conversely only 1/40 had IgG Ab with no detectable IgE antibody. There was no significant correlation between levels of specific IgE and IgG antibodies (r = 0.41, p = 0.09, data not shown).

Quantitative intradermal skin testing was used to evaluate the reactivity of recombinant B. germanica GST in vivo. Five of the seven cockroach-allergic asthmatic patients who underwent skin testing showed a positive reaction (>8 × 8-mm wheel diameter) 15 min following the injection of 0.03 ml of B. germanica GST, at concentrations of 10⁴-10¹µg/ml (Table I). All patients with IgE Ab to B. germanica GST also had IgE Ab to whole cockroach extract, and there was a good correlation between the levels of serum IgE Ab to B. germanica GST and the concentration of allergen giving a positive skin test reaction. Three non-allergic control subjects gave negative skin tests using B. germanica GST at a concentration of 1 µg/ml.

Table I. Skin tests and serum IgE antibodies to recombinant B. germanica allergen Bla g 5  Patient Skin test to B. germanicaa Bla g 5  Cockroach RAST Skin test Serum IgE units/ml µg/ml cpm boundb J. S. 8  × 10 mm 1,395 104 38,934 S. E. 7  × 15 mm 1,290 103 49,554 T. B. 14  × 15 mm 450 103 24,384 B. S. 8  × 9 mm 270 102 17,097 R. M. 105 1,065 101 3,704 S. T. 105 1,080 >1 243 J. H. 7  × 7 mm <10 >1 NDc Controls (n = 3) Neg <10 >1 382 ± 162d a Wheal diameter on prick test or dilution of extract inducing an 8 × 8-mm wheal on intradermal test. b Measured by antigen-binding RIA. c ND, not done. d Mean ± 2 S.D.

DISCUSSION

We have reported the identification, purification, molecular cloning, and sequencing of a glutathione S-transferase from the cockroach Blattella germanica. Although it was previously known that cockroach extracts contained GST activity (36, 37), this is the first time that cockroach-GST has been purified and the structure of the molecule has been defined. Three pieces of evidence characterize the B. germanica protein as an important allergen and GST. First, the high prevalence of IgE Ab binding and potent biologic activity of the allergen in vivo (positive immediate skin tests to as little as 3 pg injected allergen). Second, the sequence homology to other GST, particularly those of insect origin (up to 51% sequence identity); and finally, the ability to bind glutathione, allowing purification of both natural and recombinant cockroach-GST.

Sequence analysis has shown that insect GST can be classified into two distinct groups, GST-1 and GST-2 (28-30, 38-42). There is 44-97% sequence homology within the GST-1 and GST-2 subgroups, but only 15-24% homology between the two groups. The GST-1 enzymes have higher identity to cytosolic mammalian theta and alpha GST classes, whereas GST-2 are more similar to the pi and mu classes, and are antigenically distinct from GST-1. Based on the degree of sequence homology, B. germanica GST has been assigned to the GST-2 subfamily. An intriguing aspect of our results was that, contrary to other insect GST, the B. germanica GST showed low activity against the CDNB substrate. This could imply that B. germanica GST has higher affinity against substrates other than CDNB, or may have unusual substrate specificity. B. germanica may also produce other GST isoenzymes with higher CDNB activity than the Bla g 5 allergen.

Insect GST are thought to play an important physiologic role in the detoxification of foods and other substances ingested by cockroach. Up-regulation of GST production in insects is associated with resistance to insecticides (particularly organophosphorates) (38, 43-45). Our findings that B. germanica GST is a potent allergen suggest that attempts to control cockroach populations in the homes of allergic patients could, paradoxically, lead to increased GST expression and environmental allergen exposure. Infestation of housing with cockroach can be difficult to control and often requires prolonged treatments with insecticides. The effect of these treatments on cockroach allergen levels are not known. The implication of our results is that insecticide treatments could increase exposure to the GST allergen which in turn could lead to increased risk of allergic symptoms. This hypothesis could be investigated by monitoring B. germanica GST levels in the environment.

It has been reported that a house dust mite (D. pteronyssinus) allergen, Der p 8, which reacts with IgE Ab in sera from ~40% mite allergic patients, is a GST (20). In addition, S. mansoni GST, a promising candidate for a vaccine against schistosomiasis, induces IgE and IgA Ab, which are thought to be beneficial in reducing parasitic infection (21, 46-48). IgE Ab responses to schistosomes may protect against human infection and vaccination with S. mansoni GST has been associated with reduced egg production and resistance to infection (in animal models). The degree of sequence homology of D. pteronyssinus GST and S. mansoni GST with B. germanica GST is low (27 and 26.2%, respectively). However, the three-dimensional structures of the alpha, mu, and pi mammalian GST classes (sharing only ~30% sequence identity) have been recently determined by x-ray crystallography and these proteins have similar folding topology (17, 18, 31, 32).

Diagnosis of cockroach allergy is carried out by immediate skin testing or serologic assays for IgE Ab, using extracts of ground cockroach bodies that are not standardized. Production of highly purified recombinant allergens will enable these reagents to be used for diagnostic purposes. A mixture of recombinant cockroach allergens will be required, based on the fact that multiple allergens cause IgE Ab responses. There is a high prevalence of IgE Ab to Bla g 2 (60%), Bla g 4 (40-60%), and Bla g 5 (GST, 70%) among cockroach-allergic patients (12, 13, 24). Each of these allergens is an important cause of sensitization to cockroach and preliminary comparisons show that the in vivo purified allergens give positive skin test reactions at 10³-10⁵ µg/ml (49). Serologic analyses of ~40 sera suggest that measuring IgE Ab to Bla g 1, Bla g 2, Bla g 4, and Bla g 5 will demonstrate sensitization in 95% of B. germanica allergic patients.² Bla g 2 and Bla g 4 are not expressed in P. americana and recent results showed that P. americana extracts did not inhibit IgE Ab binding to Bla g 5 in RIA (12, 13). Thus if P. americana produces a GST allergen, it is a different isoform or GST class to B. germanica GST. We have recently cloned the cross-reactive Bla g 1 homologue from P. americana (Per a 1) and this allergen, in addition to Per a 3, could be used for diagnosis of sensitization to Periplaneta species (14).

Over the last few years, the primary structure of most major allergens has been determined and there do no appear to be particular structural features in these molecules that selectively induce IgE Ab responses (50). There is evidence that the biochemical activities of some allergens (e.g. Der p 1, bee venom phospholipase A₂) could contribute to allergenicity (50-54). It has also recently been reported that the enzyme leukotriene C4 synthase is a GST (54, 55). However, this enzyme appears to be a unique GST, with no significant sequence homology to other members of the GST family. The possibility that cockroach-GST may potentiate IgE Ab responses in the airways through its enzymatic activity could be investigated using site-directed mutagenesis to modify cockroach-GST at conserved residues involved in binding glutathione.

In summary, we have reported the identification and cloning of a major B. germanica allergen Bla g 5, which is a member of the GST superfamily. Recombinant cockroach GST has been produced in E. coli, with comparable immunologic activity to the natural protein. Results of these studies will lead to better approaches to control environmental exposure to this ubiquitous insect, including the development of GST inhibitors. In addition, cockroach-GST will provide a model for studying both the cellular and IgE Ab responses to GST and for establishing the role of these proteins in the development of asthma and other allergic diseases.