Proteomics-Based Approach for Detailing the Allergenic Profile of Cannabis Chemotypes

2.2. Putative Allergens of Cannabis

Preliminary analysis revealed the presence of cannabis proteins belonging to 50 distinct allergen families (Figure 2). Importantly, we identified peptides belonging to known and validated allergens of cannabis, including profilin (Can s 2), nonspecific lipid transfer protein (nsLTP, Can s 3) and Bet v 1-homologue- and pathogenesis-related protein PR-10 (Can s 5). Interestingly, we did not identify peptides for the oxygen-evolving enhancer protein 2 (OEEP2; Can s 4)) in cannabis extract samples in all chemotypes and hemp. Since allergic sensitization to cannabis can occur through multiple routes (inhalation, contact, ingestion), we examined the broad distribution of putative allergens based on the route of exposure (Figure 3 and Supplementary Table S1). The two putative contact allergens include proteins homologous to glucose–methanol–choline (GMC) oxidoreductase N-domain-containing protein (Mala s 12; skin yeast) and malate dehydrogenase (Mala f 4).

2.3. Putative Airway Allergens of Cannabis

The primary route of consumption of cannabis continues to be the inhalation of cannabis smoke. Further, in symptomatic exposure to cannabis, one major concern is the exacerbation of underlying asthma. Consequently, allergens that are known to be sensitizers in airways are likely to be important. We identified 18 putative cannabis allergens in our LC-M/MS analysis that were previously reported as typical airway allergens (Figure 4 and Supplementary Table S1). Among these are allergens that share homology with common airway allergen sources, such as Dermatophagoides pteronyssius or D. farinae (house dust mite) (tubulin α-chain, Der p 33; ferritin, Der p 30; SERPIN domain-containing protein, Der f 27; Chitin-binding type-1 domain-containing protein, Der p 18). Some airway allergens in cannabis share homology with other mites, such as ML domain-containing protein and aldehyde dehydrogenase (Blo t 2 and Tyr p 35; storage mites), or grass allergens, such as expansin (Cyn d 1; Bermuda grass), pectate lyase (Amb a 1; short ragweed) and plastocyanin (Amb a 7; short ragweed). Other allergens include those with homology to fungal allergens, such as catalase (Pen c 30; Penicillium), protein disulfide-isomerase (Alt a 4; Alternaria), aldehyde dehydrogenase (Alt a 10; Alternaria), calreticulin (Pen ch 31; Penicillium), transaldolase (Cla c 14; Cladosporium), peptidyl-prolyl isomerase (Asp f 11; Aspergillus), cytochrome c-domain-containing protein (Cur l 3; Curvularia). Collectively, our data indicate that cannabis contains allergens that share significant homology with known allergens from other environmental sources. Further, peptides for putative cannabis allergens relevant to airway exposure were detected in all chemotypes of cannabis and in hemp. The two exceptions included the X8 domain-containing protein (hemp and Mx varieties) and ML-domain-containing protein (B5 and V1-19 varieties).

2.4. Putative Food Allergens of Cannabis

The ingestion of cannabis is common and is typically consumed in the form of tea or an ingredient in baked goods, and allergic reactions (including anaphylaxis) on the ingestion of cannabis has been reported in the literature [14,15]. We identified peptides for 13 putative food allergens in cannabis (Figure 5 and Supplementary Table S1), which include Germin-like protein (Cit s 1; sweet orange), Chlorophyll a-b-binding, chloroplast (Api g 3; celery), L-ascorbate peroxidase (Mus a 6; banana), β-fructofuranosidase (Sol a 12; tomato), cysteine protease inhibitor (Sola t 3; potato), β-amylase (Tri a 17; wheat), (R)-mandelonitrile lyase (Pru du 10; almond) and agglutinin domain-containing protein Tri a 18; wheat) peptides. Other putative food allergens shared homology with common striped catfish allergens, such as fructose-bisphosphate aldolase (Pan h 3), pyruvate kinase (Pan h 9), glucose-6-phosphate isomerase (Pan h 11) and lactate dehydrogenase (Pan h 10). Peptides for all putative food allergens in cannabis were detected in all chemotypes and hemp.

2.5. Putative Contact Allergens of Cannabis

Contact exposure to cannabis can occur in both recreational and occupational settings. In LC-MS/MS analysis we identified two putative allergens in cannabis with homology to known contact allergens identified in skin yeasts (Figure 6 and Supplementary Table S1). These include malate dehydrogenase (Mala f 4) and GMC oxidoreductase N-domain-containing protein (Mala s 12). While malate dehydrogenase peptides were identified in all chemotypes of cannabis and hemp, peptides for GMC oxidoreductase N-domain containing protein were detected only in Mx and V1-19 varieties.

2.6. Putative Allergens of Cannabis with Capability to Sensitize through Multiple Routes

Some allergens have been previously reported to cause allergic reactions through multiple routes. Most putative allergens share homology to common and potent allergens from birch (Betula verrucosa), latex (Hevea brasiliensis) and peanut (Arachis hypogea). Four allergens have been previously known to act as airway and food allergens concomitantly and include glyceraldehyde 3-phosphate dehydrogenase (Pan h 13 and Tri a 34), glutathione transferase (Der p 8, Asc l 13), α-amylase (Der p 4, Hor v 16) and Bet v 1-domain-containing protein (pathogenesis-related protein, PR-10) (Bet v 1, Ara h 8) (Figure 7 and Supplementary Table S1). Some allergens have been previously reported as sensitizers through contact and oral route and include endo-1,3(4)-beta-glucanase (Hev b 2, Mus a 5) and patatin (Sola t 1, Hev b 7). While peptides for α-amylase were detected only in the B5 and V1-19 varieties of cannabis, peptides for endo-1,3(4)-beta-glucanase were detected in hemp and B5 varieties.

Five other allergens have been reported as sensitizers through all three routes and include cyclophilin (peptidyl-prolyl cis-trans isomerase) (Ara h 18, Bet v 17 and Mala s 6), thioredoxin domain-containing protein (Asp f 28, Mala s 13 and Tri a 15), superoxide dismutase (Asp f 6, Hev b 10, Pis v 4). The validated allergens of cannabis, including nonspecific lipid transfer protein (Can s 3) (with homology to Hev b 12, Amb a 6, Pru p 3) and profilin (Can s 2) (with homology to Bet v 2, Hev b 8 and Ara h 5) were also detected and are known to sensitize via multiple routes (Figure 8 and Supplementary Table S1).

4.2. Protein Extraction

Protein extraction was performed using two different methods. For the 3 strains Mx, V1-19 and B5, proteins were extracted from buds at the National Center for Natural Product Research, Research Institute of Pharmaceutical Sciences, School of Pharmacy at the University of Mississippi, using the PierceTM Plant Total Protein Extraction Kit (Thermo Fischer Scientific, Waltham, MA, USA) as per methodologies described previously [3]. Hemp protein extract was generated from buds by finely grinding buds using a mortar and pestle. One gram of hemp bud extract was transferred into a 15 mL conical tube and protein extraction buffer was performed under denaturing conditions using proprietary methods developed by InBio Inc. (Charlottesville, VA, USA).

4.3. Protein Estimation

Total protein content in cannabis plant extracts was measured using the PierceTM bicinchoninic acid (BCA) Protein Assay kit (Thermo Fisher Scientific), as per the manufacturer’s instructions. The serial dilution of cannabis protein samples and protein standards (generated with bovine serum albumin, BSA) were plated onto a 96-well plate, mixed with BCA working reagent solution mix and incubated at 37 °C for 30 min. Following incubation, plates were cooled briefly at room temperature (RT) and colorimetric absorbance signal was measured at 560 nm using a spectrophotometer. A standard curve was generated using absorbance measures at 560 nm for BSA protein standards. Absorbance signals measured for cannabis protein extracts was plotted on the standard curve to estimate protein concentrations.

4.4. Sample Preparation and Nanoscale Tandem Mass Spectrometry (MS/MS)

Cannabis protein sample preparation for LC-MS/MS analysis was performed using methods described previously [20]. All methods related to LC-MS/MS were performed by MS Bioworks (Ann Arbor, MI). Briefly, protein samples were first precipitated in tricarboxylic acid (TCA). Pellets were resolubilized in buffer containing 8 M urea, 50 mM Tris HCl, 150 mM NaCl, pH 8.0 and 1X Roche Complete Protease Inhibitor. Protein quantitation was performed using Qubit assay (ThermoFisher Scientific) and samples were normalized to the protein yield. The sample was digested overnight with trypsin (enzyme–substrate ratio of 1:20) with gentle shaking. Next, samples were reduced for 1 h at RT in 12 mM dithiothreitol (DTT) followed by alkylation for 1 h at RT in 15 mM iodoacetamide. Each sample was acidified in formic acid and subjected to solid-phase extraction (SPE) on a Waters μHLB C18 plate.

A total of 2 μg aliquot of samples was analyzed using a nano LC-MS/MS with a Waters NanoAcquity HPLC system interfaced to a ThermoFisher Fusion Lumos. Peptides were loaded on a trapping column and eluted over a 75 μm analytical column at 350 nL/min; both columns were packed with Luna C18 resin (Phenomenex). A 4 h gradient was employed. The mass spectrometer was operated in data-dependent mode, with MS and MS/MS performed (3 s cycles) in the Orbitrap at 60,000 full width at half maximum (FWHM) resolution and 15,000 FWHM resolution, respectively, with advanced peak determination (APD) activated.

4.5. Data Processing and Analysis

Data were processed through the MaxQuant software v1.6.2.3 (www.maxquant.org accessed on 23 June 2021). MS acquired data were recalibrated, and the data were filtered at 1% protein and peptide false discovery rate (FDR). LFQ Intensity was calculated from the individual peptide peak areas (extracted ion chromatograms) that were scaled to a protein value and then normalized between samples based on the total intensity per sample [21]. Data were saved on the Andromeda platform [22] using following criteria: enzyme—trypsin; database—UniProt (Cannabis sativa); fixed modification—carbamidomethyl (C); variable modifications—oxidation (M) and acetyl (protein N-terminal); and fragment mass tolerance—20 ppm. MaxQuant settings were as follows: protein and peptide FDR—0.01 FDR; minimum peptide length—7; minimum razor + unique peptides—1; minimum unique peptides—0; minimum ratio count for LFQ—1; and secondary peptide setting—TRUE. This was to account for two peptides co-transmitted and producing a composite product ion spectrum. Data were uploaded to Perseus v1.5.5.3 for data processing and analysis.

This work was supported in part by the National Institute of Allergy and Infectious Diseases (NIAID)/National Institutes of Health (NIH), grant R21AI140411, to A.P.N. The authors would like to thank and remember Steve Groff M.D. for providing hemp flower samples for proteomic analysis and for his contribution to advancing hemp accessibility for research.