Glutathione S-transferase (GST) of American Cockroach, Periplaneta americana: Classes, Isoforms, and Allergenicity

Insect glutathione S-transferases (GSTs) play important roles in insecticide/drug resistance and stress response. Medically, GSTs of house dust mites (Dermatophagoides pteronyssinus and Blomia tropicalis) and German cockroach (Blattella germanica) are human allergens. In this study, classes, isoforms and B-cell and allergenic epitopes of GST of American cockroach, Periplaneta americana, the predominant species in the tropics and subtropics were investigated for the first time. Enzymatically active native and recombinant P. americana-GSTs bound to IgE in sera of all P. americana allergic patients that were tested. By gel-based proteomics and multiple sequence alignments, the native GST comprises three isoforms of delta and sigma classes. All isoforms interacted with serum IgE of the cockroach allergic subjects. Molecularly, the protein contains six B-cell epitopes; two epitopes located at β1-α1 and β4-α3 regions bound to patients’ serum IgE, indicating that they are allergenic. P. americana are ubiquitous and their GST can sensitize humans to allergic diseases; thus, the protein should be included in the allergen array for component resolved diagnosis (CRD) of allergic patients, either by skin prick test or specific IgE determination. The GST is suitable also as a target of environmental allergen detection and quantification for intervention of cockroach sensitization and allergic morbidity.


SDS-PAGE and Western blot analysis (WB).
SDS-PAGE and WB were performed as described previously 17 . A 4% stacking and 12.5% separating polyacrylamide gels were used in the SDS-PAGE which was performed in a Mini-PROTEAN ® 3 Cell (Bio-Rad). Separated proteins in the gels were either stained or electro-blotted onto a nitrocellulose membrane (NC) for WB. For WB, empty sites on the blotted NC were blocked with 3% BSA in PBS before placing the membrane in a solution of mouse anti-6× His tag (Abcam, UK).   Anti-mouse immunoglobulin-alkaline phosphatase (AP) conjugate (Dako Cytomation, Denmark) and BCIP/ NBT substrate (KPL, MD, USA) were used to reveal the 6× His-tagged-rGST band.
Protein identification by LC-MS/MS. Native and recombinant GSTs were verified by LC-MS/MS as described previously 39 . The generated ion spectra of the peptides from tryptic-digested GSTs were interpreted by using the Turbo SEQUEST algorithm in the BioWorks TM 3.1SR1 software package (Thermo Fisher Scientific) and the nr.fasta database. The protein search parameter was performed as described previously 39 , which included mass tolerance of 1.25 amu, a fragment mass tolerance of ±0.4 amu, methionine (M) oxidation, and threonine (T) or serine (S) phosphorylation. The identified peptides were further evaluated using charge state versus cross-correlation numbers (X corr ). The criteria for a positive identification of the peptides were X corr > 1.5 for singly charged ions, X cor > 2.0 for doubly charged ions, and Xcorr > 2.5 for triply charged ions. A delta correlation (∆Cn) of >0.08 was used as a cut-off for peptide acceptance.
Determination of enzymatic activity of GST preparations. Enzymatic activities of the nGST and the rGST were determined by using glutathione S-transferase assay kit (Sigma-Aldrich., MO, USA) which the 1-chloro-2,4-dinitrobenzene (CDNB) was used as the enzymatic substrate. GST standard (0.25 mg/ml) was provided with the test kit. Active GST catalyzes the conjugation of L-glutathione to CDNB via a thiol group of the glutathione. The reaction product, GS-DNB conjugate, could be detected at absorbance 340 nm (A 340nm ). The amount of the product is directly proportional to the GST activity in the sample. In a reaction mixture, 2 µl (0.5 µg) of native, recombinant, and standard GST were mixed individually with 1 ml of substrate solution containing 980 µl Dulbecco's PBS, 10 µl of 200 mM reduced L-glutathione, and 10 µl of 100 mM CDNB in a 1-ml cuvette. Absorbance 340 nm of the reaction mixture was determined by spectrometer against blank (1 ml non-enzymatic conjugation substrate solution alone) using BioMate TM 3 series spectrophotometer (Thermo Fisher Scientific). A kinetic program was set for every 30 seconds over a period of 5 minutes after a lag time of 1 minute. Specific catalytic activities of both nGST and rGST (µmol of GST/ml/minute) were calculated.  Table 3. Orthologous proteins of database that contained peptides matched with peptides of proteins in gel plugs 1-5 of 2DE-separated P. americana-native GST.

P. americana GST peptide no. Amino acid sequence and residue numbers
Location on the P. americana-GST molecule α8-α9 Table 4. Sequences of P. americana synthetic peptides 1-6 which encompassed the 6 predicted P. americana-GST B-cell epitopes. These peptides were used for determining B-cell and allergenic (IgE-binding) epitopes.  Allergenicities of nGST and rGST. Allergenicities (IgE binding frequencies) of the nGST and the rGST were determined by IgE-ELISA. The assay was performed as described previously 17 . The nGST and rGST (5 μg/ ml carbonate-bicarbonate buffer, pH 9.6) were added to separate wells (100 μl/well) of a microtiter plate (Costar, MA, USA) and kept at 37 °C until dried. All GST-coated wells were washed and blocked with 200 μl of a blocking solution (1% BSA in PBS) before adding with serial two-fold dilutions of individual sera and the plate was incubated for 3 hours. Wells added with only the serum diluent served as blank. All wells were washed and added with 100 μl of mouse anti-human IgE-biotin conjugate (Southern Biotech, AL, USA; diluted 1:1,000 in PBS-T). Streptavidin-horseradish peroxidase (HRP) conjugate (Dako Cytomation) and ABTS substrate solution (KPL) were used for color development. Absorbance at 405 nm (OD 405nm ) of the content in each well was determined (ELISA reader, MultiscanEX, Labsystem, Helsinki, Finland) against the blank. Cut-off OD 405nm between positive and negative IgE-ELISA was arbitrarily set at ≥mean OD 405nm of non-allergic sera + 2 standard deviations (SD).

Identification of nGST isoforms and their IgE reactivity.
Isoforms of nGST were determined by a gel-based proteomics. Purified nGST was subjected to 2DE as described previously 39 . For the first dimensional electrophoresis, 7 cm-IPG strips and 0.5% pH 3-10 IPG buffer (GE Healthcare) were used. The electrophoresed-IPG strips were then subjected to 12% SDS-PAGE and proteins in the gel were stained by CBB. Gel pieces containing proteins of ~21 kDa were excised from the stained gel and subjected to in-gel tryptic digestion and LC-MS/MS, respectively. Protein orthologues were identified by comparing the peptide sequences of the P. americana-GST generated from the mass spectrometry with the Arthropoda/Insecta database sequences. The nGST isoforms were checked for their reactivity to IgE in the pool of CR allergic patients' sera by 2DE IgE-immunoblotting. The 2DE-separated nGST was electro-transblotted onto an NC, blocked with BSA, and the blot was allowed to react with the CR allergic patients' serum pool. After keeping at 4 °C overnight, the NC was washed with TBS-T before placing in a solution of appropriately diluted mouse anti-human IgE-biotin conjugate (Southern Biotech) and kept at 25 °C on a rotating platform for 3 hours. Spots of the nGST isoforms bound by the specific serum IgE were revealed by using streptavidin-AP conjugate (Dako Cytomation) and BCIP/NBT substrate (KPL).

Identification of P. americana-GST B cell and allergenic epitopes. Linear B-cell epitopes of P.
americana-GST were predicted by using BepiPred 1.0 server. The amino acid sequence of P. americana-GST was submitted to the server. Three methods available at the server, i.e., BCPred 40 , AAP 41 , and FBCPred 42 were used for the epitope prediction. Specific threshold and the epitope length were set at 85% and 20-25 residues, respectively. All predicted peptides obtained from individual methods were aligned with the P. americana-GST sequence to obtain consensus sequences (predicted linear B-cell epitopes). Peptides containing the potential linear B-cell epitopes of P. americana-GST were synthesized. Moreover, in the case that two of the predicted epitope sequences were close to each other (peptides 2 and 4 of this study), the overlapped peptide between the two sequences was also synthesized (peptide 3 of this study). Binding of the synthetic peptides to antibodies in a pool of P. americana sensitized human sera (for determining B-cell epitopes) and to IgE in individual allergic patients' sera (for determining allergenic epitopes) were determined by dot-ELISAs.

Dot-ELISAs.
For detecting binding of the synthetic peptides to antibodies in the pool of the CR allergic patients' sera, individual synthetic peptides (1 μg) were dotted onto one cm-NC squares. PBS was used as negative peptide control. The NC pieces were blocked with 1% BSA in PBS-T and then placed in the allergic patients' serum pool. After keeping at 25 °C for 3 hours, the membranes were washed and allowed to react with goat-anti-human IgG-AP conjugate and BCIP/NBT substrate, respectively, with appropriate incubation and washing with PBS-T between the steps. The enzymatic reaction was stopped by rinsing the membranes with distilled water. Color appeared at the peptide-dotted spots indicated that the peptide contained B-cell epitope.
For determining the IgE reactivity of the synthetic peptides, individual peptides were dotted separately onto NC squares (1 μg/dot) and let air-dried. They were blocked with 1% BSA in PBST and probed with individual serum samples (diluted 1:4 in PBST). Serum samples of normal subjects were included in the experiments. After incubating and washing, all NC squares were reacted sequentially with biotin-labeled-goat-anti-human IgE, streptavidin-AP conjugate, and BCIP/NBT substrate, respectively. The peptides that gave the typical colored spots indicated that they contained allergenic (IgE binding) epitopes.
Locations of the IgE-binding (allergenic) epitopes on the P. americana GST three dimensional (3D) structural model. Deduced amino acid sequence of the cloned P. americana GST was submitted to I-TASSER server 43 . The templates used for GST 3D modeling (selected by the server) were PDB ID: 3WYW, 3VK9, 3AY8 and 4PNF. The allergenic peptides were mapped on the 3D modeled structure of the P. americana-GST.

Results
CR allergic patients. Dermographic data of the 15 P. americana allergic patients and 5 normal subjects as well as clinical diagnosis and results of the skin prick tests and the specific IgE levels among the patients are shown in Table 1.

P. americana nGST and rGST.
Gene sequence coding for full-length P. americana-GST (648 bp) was cloned into pKRX-T cloning vector and subcloned into pET20b + protein expression vector. Supplementary Fig. 1 illustrates nucleotide and deduced amino acid sequences of the P. americana-GST of this study (accession number MG255130). Fig. 1  Sequence Alignment) and percent identity of amino acid sequence of P. americana-rGST of this study (accession number MG255130) with GSTs of other insects are shown in Supplementary Fig. 2. The P. americana-rGST was purified from the gst-pET20b + -transformed E. coli (Fig. 2, lane 2) and the protein was verified by the mass spectrometry ( Table 2).
The nGST prepared from the frozen P. americana whole body powder and purified by using GST trap column revealed a protein band at ∼23 kDa (Fig. 2, lane 1). LC-MS/MS verified that the protein band was the nGST ( Table 2).

Catalytic activities of nGST and rGST.
Enzymatic activities of the nGST and the rGST were 36.65 and 2.93 μmol/ml/min, respectively. The nGST had a much higher enzymatic activity than the rGST. Serum IgE reactivities of P. americana nGST and rGST. The mean + 2 SD of IgE-ELISA OD 405nm of the non-allergic control sera tested against the nGST and rGST were 0.460 and 0.649, respectively. These values were used arbitrarily as cut-off levels between positive and negative IgE-ELISA and it was found that all of the 15 allergic patients' sera gave positive IgE-ELISA results to both nGST and rGST proteins (Fig. 3). P. americana nGST isoforms. The 2 DE-pattern of purified P. americana nGST stained by CBB is shown in Fig. 4A. There are 5 protein spots at ∼21 kDa. The gel pieces containing the 5 protein spots were subjected to LC-MS/MS. The results (Table 3) revealed that tryptic peptides generated from spot nos. 1, 2, and 4 matched with peptides of P. americana delta variant 1 (accession number Gi 359326557), which the molecular mass was 24606 Da and the pI was 6.44. This protein has possibility of 14 phosphorylation sites and a possible O-glycosylation at residue 116. Peptides of spot no. 3 matched with sigma GST of Locusta migratoria (migratory locust) (accession number Gi 565341529) with molecular masses of 23459 Da and pI 6.19. This protein also has 14 possible phosphorylation sites; albeit most sites are different from the delta GST of spots 1, 2, and 4. Peptides of spot no. 5 matched with peptides of B. germanica sigma GST (accession number Gi 359326585) which has a molecular mass of 23377 Da and pI 6.84 with possible 16 phosporylation sites. Thus, the nGST of P. americana of this study has three isoforms which belong to delta (one isoform) and sigma classes (two isoforms). By the 2 DE-IgE immunoblotting, all three isoforms of the nGST bound to IgE in a serum pool of the P. americana allergic patients (Fig. 4B).

Linear B-cell epitopes and allergenic epitopes of P. americana-GST. Linear B-cell epitopes of P.
americana-GST predicted by using BepiPred 1.0 server are shown in Supplementary Fig. 3. The BCPred, AAP, and FBCPred methods predicted 2, 3, and 4 epitopic sequences, respectively. All of the predicted peptides were aligned with the P. americana-GST sequence. They were found to match with five regions (marked in red in the Supplementary Fig. 3) of the GST including β1-α1, α2-β3-β 4, α3-α4, α5-α6, and α8-α9. The consensus peptides of these locations were synthesized (Pep 1, 2, 4, 5, and 6; marked in blue in the Supplementary Fig. 3 and peptides 1, 2, 4, 5, and 6 in Table 4). Because the matched regions α2-β3-β4 and α3-α4 were close to each other, an overlapped peptide of the two regions was also synthesized (Pep 3 in the Supplementary Fig. 3 and peptide 3 in Table 4).
All synthetic peptides gave positive binding to antibodies in a pool of 10 P. americana allergic subjects (Fig. 5A), verifying the computerized results that the peptides contained B-cell epitopes. Sera of all P. americana allergic patients gave positive IgE-dot-ELISA to peptide 1 (1MTIDFYYLPGSAPCRSVLLA20) located between β1 and α1 and peptide 3 (61GFCLWESRAILSYLADQYGK80) located between β4 and α3 of the P. americana-GST molecule (Fig. 5B) and did not give positive IgE binding with peptides 2, 5, and 6 (data not shown). Figure 6 illustrates locations of the two IgE-binding (allergenic) epitopes on the 3D modeled structure of the P. americana-GST.

Discussion
American cockroach, P. americana, is a predominant species and a major source of indoor allergen causing type 1 hypersensitivity in atopic inhabitants of subtropics (such as Taiwan) and tropics (such as Brazil, Malaysia, Singapore, and Thailand), especially in crowded and unhygienic urban cities where their infestation is enhanced 1,44 . Several P. americana-derived proteins have been recognized as important (major) allergens as they sensitized >50% of cockroach allergic subjects. Glutathione S-transferases (GSTs) are indispensable enzymes which insects use to protect themselves against oxidative damage and stress and insecticide toxicity 30 . Most interest on the insect GSTs has been focused on their role in insecticide and drug resistance while other attributions of this protein family seem to be neglected. In this study, three P. americana-GST isoforms belonging to delta and sigma classes were identified. The roles of the protein in causing human allergy, i.e., allergenicity (specific serum IgE-binding frequency) as well as B-cell epitopes and allergenic epitopes were investigated. To our knowledge, this is the first report on the allergenic attribution of the American cockroach, P. americana-GST. The protein that was cloned from the GenBank database (accession number AY792949; UniProt ID Q1M0Y4) has been designated Per a 5 (Per a 5.0101) by the IUIS Allergen Nomenclature. The other isoforms reported in this study were based on the tryptic digested peptides generated from the 2DE protein spots that matched with the orthologous proteins of the database. Because the complete amino acid sequences of these isoforms were not available, they were not submitted to the IUIS Allergen Nomenclature.
Enzymatically active native and recombinant P. americana-GSTs were produced. On the equal weight of both proteins (0.5 μg), the native protein was about 12.5 times more active than the recombinant counterpart. The difference may be because the nGST contains several isoforms belonging to different GST classes whereas the recombinant one is produced from only one cDNA sequence. The multiple isoforms in the nGST might confer ScientiFic RePORTS | (2018) 8:484 | DOI:10.1038/s41598-017-18759-z additive or synergistic enzymatic activity. Besides, the rGST produced from the transformed BL21 (DE3) E. coli contains additional 6× His tag (useful for subsequent protein detection and purification) but lacks putative post-translational modifications such as glycosylation and disulfide bridge formation which might impact on the protein folding compared to the native state and hence the less enzymatic activity. After subjecting the nGST and rGST to SDS-PAGE and protein staining, the molecular masses of the native and recombinant proteins were ∼23 and ∼25 kDa, respectively. The larger size of the latter should be from the plasmid franking regions and the 6× His tag. The rGST produced in this study (accession number MG255130) contained 216 amino acids which are similar in residue number to the cloned 24614 Da B. germanica GSTD1 45 . The P. americana-rGST has 99.5 and 98.1% sequence identity to the previously reported P. americana-GSTs (UniProt ID G8XWU4 and Q1M0Y4) and 32.9-82.4% identity to the delta GSTs of other organisms and less so to the other GST classes. Thus, the recombinant P. americana-GST (accession number MG255130) should belong to the delta class.
Both native and recombinant P. americana GSTs reacted with IgE in sera of all cockroach allergic patients when tested by indirect ELISA, indicating that the GST is a novel and important (major) allergen of the P. americana. The ELISA using rGST as antigen gave higher background signal than when the nGST was used. This could be due to the contamination of residual proteins of the E. coli used as the rGST expression host in the rGST preparation which reacted to the anti-E. coli that exists naturally in the human sera.
Data on the allergenic repertoire of a protein are useful for understanding the patients' allergenic response, cross-reacting allergenic determinants among allergens, as well as for properly designing of an engineered therapeutic allergen vaccine/diagnostic material, particularly for personalized immunotherapy and component resolved diagnosis (CRD) 46,47 . Several methods have been used for gaining information on B-cell and allergenic epitopes of an allergen. These include the use of overlapping synthetic peptides or fragments of recombinant allergens 47,48 ; mimotope mapping 49 ; peptide microarray immunoassay 50,51 ; X-ray crystallography and nuclear magnetic resonance techniques 47,52-55 , computerized prediction [56][57][58] and specific monoclonal antibody binding and IgE competition assay 58-60 . In this study, a combination of in silico methods and synthetic peptide based-immunoassays was used for predicting and determining B-cell and allergenic epitopes of the P. americana-GST. For the in silico prediction of B cell epitopes, three different methods of the BepiPred 1.0 server, i.e., BCPred, AAP, and FBCPred were used. All methods gave a conformed prediction of a peptide located at α2-β3-β4; the AAP and PBCPred methods predicted another peptide at α3-α4; the BCPred and the FBCPred methods predicted another peptide at α5-α6; and only the AAP method predicted two more peptides at β1-α1 and α8-α9. The different results so-obtained suggest that several in silico methods should be used for increasing the possibility of finding the potential B-cell epitopes of a particular protein. From the in silico prediction, consensus peptides encompassed the potential B-cell epitopes were synthesized and used in the dot-ELISAs for detecting peptides bound by antibodies (B-cell epitopes) and IgE (allergenic epitopes) in sera of P. americana sensitized subjects. By using the dot-ELISAs, all six synthetic peptides that contained predicted GST B-cell epitopes reacted with antibodies in a serum pool of P. americana exposed-subjects, implying that the six peptides were part or contained GST B-cell epitopes. Among them were two IgE-binding epitopes located at the β1-α1 and β4-α4 regions of the GST molecule.
In summary, this study provides an insight into characteristics and medically important role of the P. americana-GST beyond the previously established physiologic roles of the protein in the host defense against toxic substances and stress conditions. Both native and recombinant GSTs of the P. americana were bound by IgE in sera of cockroach allergic subjects and thus the GST is a novel and major P. americana allergen. Because cockroaches are ubiquitous, attention should be paid on reducing the insect derived-allergenic proteins from human environment for intervention of allergic sensitization of the naives and clinical aggravation of the sensitized subjects.