Proteomic profiling of the weed feverfew, a neglected pollen allergen source

Feverfew (Parthenium hysterophorus), an invasive weed from the Asteraceae family, has been reported as allergen source. Despite its relevance, knowledge of allergens is restricted to a partial sequence of a hydroxyproline-rich glycoprotein. We aimed to obtain the entire sequence for recombinant production and characterize feverfew pollen using proteomics and immunological assays. Par h 1, a defensin-proline fusion allergen was obtained by cDNA cloning and recombinantly produced in E. coli. Using two complementary proteomic strategies, a total of 258 proteins were identified in feverfew pollen among those 47 proteins belonging to allergenic families. Feverfew sensitized patients’ sera from India revealed IgE reactivity with a pectate lyase, PR-1 protein and thioredoxin in immonoblot. In ELISA, recombinant Par h 1 was recognized by 60 and 40% of Austrian and Indian sera, respectively. Inhibition assays demonstrated the presence of IgE cross-reactive Par h 1, pectate lyase, lipid-transfer protein, profilin and polcalcin in feverfew pollen. This study reveals significant data on the allergenic composition of feverfew pollen and makes recombinant Par h 1 available for cross-reactivity studies. Feverfew might become a global player in weed pollen allergy and inclusion of standardized extracts in routine allergy diagnosis is suggested in exposed populations.


cDNA cloning of Par h 1
Pollen grains from feverfew were ground and total RNA purification was performed using TRIzol (Ambion/RNA, Invitogen, Carlsbad, CA, USA). The reverse transcription was achieved using oligo dT primer and the Super-Script III First-strand Synthesis System (Invitrogen). Amplification of the cDNA was obtained after nested PCR with forward degenerated primers (5'-TGGTTYGGNAAYTGYAARGA-3' and 5'-TGYAARGAYACNGARAARTGYGA-3') based on the partial Par h 1 amino acid sequence previously described 2 . The amplified products were cloned into the pGEM-T Easy Vector (Promega, Madison, WI, USA). The 5'-UTR and the signal peptide sequence were identified using the 5' RLM-RACE protocol and the gene specific primers 5'-GGACCTGGATTCTTCTTAGGATCA-3' and 5'-CTTAGGATCACAATCAAAGTAGC-3'. The Cterminal part of the sequence was obtained using a forward gene specific primer localized in the signal peptide 5'-ATGGCGAAGAGTTCAACTTCTTACTTAGT-3' and oligo dT.

Production of recombinant allergens
Par h 1 mature protein sequence was cloned into pHisParallel2 vector using 5'-GAGACATATGGGTAAAGTATGTGA-3' as forward primer, and 5'-TCTCCTCGAGCTAAGCAGCTGGTG-3' as reverse primer with NdeI and XhoI restriction sites respectively. In order to increase the stability during expression, a glycine residue was introduced at the N-terminus. Recombinant Par h 1 was expressed as non-fusion protein in E. coli Rosetta-gami B (DE3) pLysS (Novagen, Gibbstown, NJ, USA). The bacterial culture was induced with 0.4 mM isopropyl-b-D-thiogalactopyranoside and grown at 20 °C for 16 hours. Bacterial cells were harvested by centrifugation at 4000g for 20 minutes. Cells were lysed by cycles of freezing and thawing followed by sonication (2 min, pulsing 3x at 50%) in lysis buffer; 10 mM sodium phosphate pH 7.0, 300 mM sodium chloride and 2.5 µg/mL DNAse final concentrations. Lysate supernatants were precipitated stepwise with ammonium sulfate at 40% of saturation (1.6 M) at 4 °C. Par h 1 enriched supernatant obtained after the protein precipitation, was subjected to two steps of hydrophobic interaction chromatography using 10 mM sodium phosphate pH 7.0, 300 mM sodium chloride and ammonium sulfate 1.6 M as binding buffer and 10 mM sodium phosphate pH 7.0, 10 mM sodium chloride and 8% isopropanol as elution buffer. First, the protein solution was passed through 1 mL low substitution phenyl column using 20% of the elution buffer and unbound proteins were subsequently loaded onto 5 mL high substitution phenyl column. Proteins were then eluted using an increasing gradient of elution buffer, 20-100% in 60 min at a flow rate of 1 mL/min. Fractions containing Par h 1 were pooled and the final purification was performed by size exclusion using a Superdex 75 10/300 GL column (GE Healthcare, Little Chalfont, UK) in 5 mM of ammonium carbonate pH 7.8 at a flow rate of 0.3 mL/min. The purified protein was stored at -20 °C until further use.
Art v 1 from mugwort was expressed as described above for Par h 1 and bacterial lysate were obtained in the same way. The lysis buffer for Art v 1 was 10 mM ammonium hydrogen carbonate pH 7.8, 10 mM sodium chloride and 2.5µg/mL DNAse final concentration. Soluble fraction containing Art v 1 was ultra-filtrated through a membrane with 50 kDa cut-off (Vivacell 250 insert, Sartorius, Germany).
The ultrafiltration process was followed by one step of cation exchange chromatography (SP Sepharose Fast Flow, (GE Healthcare) where 50 mM sodium acetate pH 5.35 was the binding condition and for elution sodium chloride concentration was gradually increased to 1M. Fractions containing Art v 1 were pooled and the final purification was performed by size exclusion using a Superdex 75 10/300 GL column (GE Healthcare) in 5 mM of ammonium carbonate pH 7.8 at a flow rate of 0.3 mL/min. The purified protein was stored at -20 °C until further use. Recombinant purified allergens from mugwort Art v 3, Art v 4 and Art v 5 were obtained as described 3-5 . Natural Amb a 1 from ragweed was obtained as previously described 6 .

Physicochemical characterization of purified Par h 1
Amino acid analysis was performed with the protein in duplicates following the Pico- Tag  For intact mass measurements of purified Par h 1, the sample was desalted with C 18 ZipTips (Merck, Millipore, Billerica, MA, USA) and directly infused into the Q-Exactive mass spectrometer (Thermo Fisher Scientific) at a flow rate of 1 µl/min, using the nano electrospray head. Raw data obtained from intact proteins were processed with Protein Deconvolution 2.0 (Thermo Fisher Scientific).
Circular dichroism spectra to study the secondary structure were recorded in 10 mM of potassium phosphate pH 7.0 with a JASCO J-815 spectropolarimeter (Jasco, Tokyo, Japan). Spectra were also recorded when the protein was heated up to 95 °C and cooled down to 20 °C. Far UV spectra (190-260 nm) were baseline subtracted and results are presented as the mean residue molar ellipticity.

1D and 2D-gel electrophoresis
Feverfew pollen extract, bacterial lysates and purified Par h 1 were analyzed by reducing polyacrylamide gel electrophoresis using 15% gels. Proteins were visualized with Coomassie Brilliant Blue R-250 staining (Bio-Rad, Hercules, CA, USA). For 2D gel electrophoresis, 30mg of feverfew pollen grains were ground to fine powder in liquid nitrogen and the pollen extract was obtained as described 7 . The vacuum dried pellet was re-suspended in 300 µL of isoelectric focusing buffer and proteins were loaded into ReadyStrip IPG strip pH 3-10 (Bio-Rad Laboratories) by rehydration at 50 mA for 16 hours at room temperature. The buffer compositions and the isoelectric focusing conditions are the same as described by 1 . Afterwards, the strip was loaded onto 15% acrylamide gels, and proteins were separated by reducing SDS-PAGE. Proteins were first visualized with Coomassie Brilliant Blue R-250 (Bio-Rad Laboratories) and the obtained spots were excised from gels for further use.

Mass spectrometry of feverfew proteins
For proteomic analyses of feverfew pollen, two complementary methods were used. The first approach relied on generating a tryptic digest of the pollen extract which was directly analyzed by reverse-phase liquid chromatography mass spectrometry (LC-MS/MS). In this case the protein extract from 30 mg of pollen was prepared using the same extraction method described by 7 . The other method consisted in separation of the pollen proteins by two-dimensional electrophoresis (2-DE), as described above,

Database search and Gene Ontology annotations
Fragment spectra obtained from LC-MS/MS experiments were searched against the NCBInr database with PEAKS Studio 7.5. A target-decoy search strategy was employed and the false discovery rate (FDR) was calculated by the decoy-fusion method. Protein scores were set based on the desired FDR of 1%. Identifications were only considered when at least one unique peptide, for in-solution digested samples or three for in-gel digested samples were present. Additionally, the data were manually inspected, proteins identified with the same group of peptides were merged with the one with the higher score, and the amino acids leucine and isoleucine were considered as equal. Afterwards, proteins with unknown description were also excluded from the analysis in both methods.
In order to eliminate double identifications arising from highly conserved proteins across plant species, a protein blast strategy was performed as follows. Societies (IUIS) for allergen nomenclature sub-committee. In summary, an identified protein was considered an allergenic protein when the homology with a known allergen gave an identity score ≥ 50% (according to Allergome) and was officially recognized by the IUIS allergen nomenclature subcommittee.
In order to describe the entire proteome of feverfew pollen a final list of proteins was created by eliminating redundant proteins. The final list of proteins was annotated to Gene Ontology using the commercially available software Blast2GO. Pie chart graphs were created for biological process with level 2, cellular component with level 2 and molecular function with level 3.      Images were taken with ChemiDoc MP Imaging System (Bio RAD) and acquired with the software Image Lab. 4.0.1. MW, molecular weight; E, pollen extract; P, recombinant Par h 1; NHS non-atopic human sera; BC, buffer control. The images correponding to Fig. 3 were cropped (in order to get straight lines) using the image tools of the same software. Processing like changing brightness and contrast were applied to all the stripes belonging to the same images. Coomassie staining Nonexposed Nonexposed