Optimization of Hemp (Cannabis sativa L.) Protein Extraction and Characterization of Protein Structure, Function, and Nutritional Quality across Different Cultivars

Abstract

The global population is predicted to reach 9.7 billion people in 2050, presenting a significant challenge of producing enough nutritious food in a sustainable way. Specifically, the demand for protein is increasing, with marked increase in the demand for plant protein ingredients. Current plant protein sources (soy, wheat, pea) have limitations with respect to consumer perspective or from a functional or flavor perspective, necessitating exploration of novel sources.Hemp (Cannabis sativa L.), which has been cultivated for thousands of years, is an environmentally friendly crop. However, legal restrictions due to the presence of the psychotropic component delta-9 tetrahydrocannabinol (THC) have antagonized the hemp market for decades. In 2018, hulled hemp seeds, hemp seed protein powder, and hemp seed oil achieved “generally recognized as safe” GRAS status in the U.S. Hemp seeds contain high amounts of oil (~ 30%) and protein (~25%). Research on feasible production of hemp protein isolate (HPI) and on the functional properties for food applications is minimal. Determining optimal protein extraction procedures to produce HPI will be instrumental in the adoption as a desirable protein ingredient. Additionally, identifying differences in seed protein characteristics in different cultivars is needed to initiate breeding strategies to improve the prospects for food applications. Therefore, the objectives of this work were to 1) optimize protein extraction from hemp seed, following alkaline solubilization coupled with isoelectric precipitation and salt solubilization coupled with membrane filtration, to produce a protein isolate with acceptable color, purity, yield, structural and functional characteristics, and nutritional quality, and 2) evaluate HPI produced from four industrial cultivars for differences in color and protein structural, functional, and nutritional properties. Whole hemp seeds from one cultivar (CFX-2) harvested in 2016 were dehulled using an impact dehuller and further separated using sieves, an aerator, a gravity separator table, and manual separation. Dehulled seeds were pressed using a cold hydraulic press, ground, defatted with hexane, and milled to 50 mesh prior to protein extraction. Two methods of protein extraction were tested – alkaline solubilization coupled with isoelectric precipitation and salt solubilization coupled with membrane filtration. Optimal protein extraction conditions were determined by evaluating different parameters including solubilization pH, precipitation pH, salt (NaCl) solubilization concentration, and heating and assessing protein purity and yield. HPIs produced using both methods (pH-HPI and salt-HPI) were characterized in comparison to commercial soy protein (cSPI) and pea protein isolates (cPPI). Color was measured using a colorimeter. Structural analysis was performed using SDS-PAGE for protein profile, DSC for protein denaturation, a spectrophotometric method for surface hydrophobicity, zeta potential for surface charge, and Fourier transform infrared spectroscopy (ATR-FTIR) for protein secondary structure. Protein functionality including solubility, gel strength, water-holding capacity, and emulsification and foaming properties were evaluated. Structural and functional testing were performed in water and in 0.5 M NaCl due to sedimentation of salt-HPI in water. Nutritional quality was determined by calculating the protein digestibility-corrected amino acid score (PDCAAS) based on amino acid analysis and the pH drop in vitro protein digestibility assay. Whole hemp seeds from four industrial cultivars (CFX-2, Grandi, Joey, Picolo) harvested in 2019 were dehulled, separated, defatted, and milled, as described above. Four HPI samples were produced from the cultivars following the optimized pH extraction method. Color and protein structural, functional, and nutritional properties were characterized as described above. Structural and functional testing was only performed in water. Both the optimized pH-assisted (solubilization at pH 11, precipitation at pH 5) and salt-assisted (solubilization in 0.75 M NaCl at 50°C followed by ultrafiltration/diafiltration) protein extraction methods produced HPI with high protein purities (87 – 88% protein) and remarkable yields (> 80%). The use of dehulling prior to defatting and protein extraction resulted in HPIs with desirable light and bland colors. Both extraction methods produced isolates with similar protein profiles, but pH-HPI exhibited some protein polymerization and was partially denatured. These structural differences appeared to improve HPI properties. Specifically, salt-HPI needed to be dispersed in a dilute salt solution to form a gel, while pH-HPI formed a gel in water at relatively low protein concentration (10% protein). Overall, HPI was less functional than cSPI and cPPI, but had similar solubility to cSPI at acidic pH in water, and superior solubility (P < 0.05) and gel strength at neutral pH in 0.5 M NaCl. While the use of 0.5 M NaCl for solubilization improved HPI gel strength and solubility at neutral pH, it negatively impacted water holding capacity and foaming stability. The PDCAAS of HPI (0.58 for pH-HPI and 0.54 for salt-HPI) was within the range previously reported for whole hemp seeds, dehulled hemp seeds, and hemp seed meal (0.48 – 0.61). In general, alkaline solubilization coupled with isoelectric precipitation was determined to produce a more functional and nutritious HPI. Minimal structural differences among HPI extracted from the four cultivars were observed, which contributed to only slight differences in functionality and nutritional properties. All HPIs had similar solubility to cSPI at acidic pH, and three cultivars (Grandi, Joey, Picolo) produced significantly stronger gels (P < 0.05) than cSPI. There were no significant differences in in vitro digestibility among the four HPIs, but differences in amino acid profile led to significant yet minor differences in PDCAAS. This work demonstrated that protein can be successfully extracted from dehulled hemp seeds to produce an HPI with high protein purity, yield, and acceptable color. Protein extraction using pH produced an HPI with some promising functional attributes comparable or in some cases superior to cSPI and cPPI. This study was the first to optimize protein extraction parameters for hemp and to provide a comprehensive structural, functional, and nutritional comparison between pH-extracted and salt-extracted HPI. Additionally, this study was the first to examine the impact of cultivar on HPI properties. Since minimal differences among HPIs from four industrial cultivars were observed, more cultivars should be tested for differences that would warrant a hemp breeding program for improved functionality and nutritional quality. Future work is also needed to understand how HPI functions within a food matrix.