Legumes as food ingredient: characterization, processing, and applications

Legumes as Food Ingredient • Alfonso Clemente and Jose C. Jimenez-Lopez Legumes as Food Ingredient Characterization, Processing, and Applications Printed Edition of the Special Issue Published in Foods www.mdpi.com/journal/foods Alfonso Clemente and Jose C. Jimenez-Lopez Edited by Legumes as Food Ingredient Legumes as Food Ingredient: Characterization, Processing, and Applications Editors Alfonso Clemente Jose C. Jimenez-Lopez MDPI •Basel •Beijing •Wuhan •Barcelona •Belgrade •Manchester •Tokyo •Cluj •Tianjin Jose C. Jimenez-Lopez Spanish National Research Council (CSIC) Spain Editors Alfonso Clemente Spanish National Research Council (CSIC) Spain Editorial Office MDPI St. Alban-Anlage 66 4052 Basel, Switzerland This is a reprint of articles from the Special Issue published online in the open access journal Foods (ISSN 2304-8158) (available at: https://www.mdpi.com/journal/foods/special issues/Legumes Food Ingredient Characterization Processing Applications). For citation purposes, cite each article independently as indicated on the article page online and as indicated below: LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. Journal Name Year,Volume Number, Page Range. ISBN 978-3-0365-0614-2 (Hbk) ISBN 978-3-0365-0615-9 (PDF) Cover image courtesy of Jose R. Fernandez. © 2021 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications. The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND. Contents About the Editors ..............................................vii Preface to ”Legumes as Food Ingredient: Characterization, Processing, and Applications” .. ix Alfonso Clemente and Jose C. Jimenez-Lopez Introduction to the Special Issue: Legumes as Food Ingredient: Characterization, Processing, and Applications Reprinted from: Foods 2020,9, 1525, doi:10.3390/foods9111525 .................... 1 Mar´ıa del Carmen Mar´ın-Manzano, Oswaldo Hernandez-Hernandez, Marina Diez-Municio, Cristina Delgado-Andrade, Francisco Javier Moreno and Alfonso Clemente Prebiotic Properties of Non-Fructosylated α-Galactooligosaccharides from PEA (Pisum sativum L.) Using Infant Fecal Slurries Reprinted from: Foods 2020,9, 921, doi:10.3390/foods9070921 ..................... 5 Elena Lima-Cabello, Juan D. Alch´e and Jose C. Jimenez-Lopez Narrow-Leafed Lupin Main Allergen β-Conglutin (Lup an 1) Detection and Quantification Assessment in Natural and Processed Foods Reprinted from: Foods 2019,8, 513, doi:10.3390/foods8100513 ..................... 23 Ayato Matsuo, Kaho Matsushita, Ayano Fukuzumi, Naoki Tokumasu, Erika Yano, Nobuhiro Zaima and Tatsuya Moriyama Comparison of Various Soybean Allergen Levels in Genetically and Non-Genetically Modified Soybeans Reprinted from: Foods 2020,9, 522, doi:10.3390/foods9040522 ..................... 43 Hamid Khazaei, Maya Subedi, Mike Nickerson, Cristina Mart´ınez-Villaluenga, Juana Frias and Albert Vandenberg Seed Protein of Lentils: Current Status, Progress, and Food Applications Reprinted from: Foods 2019,8, 391, doi:10.3390/foods8090391 .................... 61 Kinnari J. Shelat, Oladipupo Q. Adiamo, Sandra M. Olarte Mantilla, Heather E. Smyth, Ujang Tinggi, Sarah Hickey, Broder R¨uhmann, Volker Sieber and Yasmina Sultanbawa Overall Nutritional and Sensory Profile of Different Species of Australian Wattle Seeds (Acacia spp.): Potential Food Sources in the Arid and Semi-Arid Regions Reprinted from: Foods 2019,8, 482, doi:10.3390/foods8100482 .................... 85 Katharina Schlegel, Anika Leidigkeit, Peter Eisner and Ute Schweiggert-Weisz Technofunctional and Sensory Properties of Fermented Lupin Protein Isolates Reprinted from: Foods 2019,8, 678, doi:10.3390/foods8120678 .................... 97 Cynthia El Youssef, Pascal Bonnarme, S´ebastien Fraud, Anne-Claire P´eron, Sandra Helinck and Sophie Landaud Sensory Improvement of a Pea Protein-Based Product Using Microbial Co-Cultures of Lactic Acid Bacteria and Yeasts Reprinted from: Foods 2020,9, 349, doi:10.3390/foods9030349 .....................113 Prit Khrisanapant, Biniam Kebede, Sze Ying Leong and Indrawati Oey A Comprehensive Characterisation of Volatile and Fatty Acid Profiles of Legume Seeds Reprinted from: Foods 2019,8, 651, doi:10.3390/foods8120651 .....................131 v Ye-Na Kim, Syahrizal Muttakin, Young-Min Jung, Tae-Yeong Heo and Dong-Un Lee Tailoring Physical and Sensory Properties of Tofu by the Addition of Jet-Milled, Superfine, Defatted Soybean Flour Reprinted from: Foods 2019,8, 617, doi:10.3390/foods8120617 .....................151 Nesli Sozer, Leena Melama, Selim Silbir, Carlo G. Rizzello, Laura Flander and Kaisa Poutanen Lactic Acid Fermentation as a Pre-Treatment Process for Faba Bean Flour and Its Effect on Textural, Structural and Nutritional Properties of Protein-Enriched Gluten-Free Faba Bean Breads Reprinted from: Foods 2019,8, 431, doi:10.3390/foods8100431 .....................161 vi About the Editors Alfonso Clemente He is a staff scientist at the Spanish National Research Council, working at the Estaci´ on Experimental del Zaid´ ın (Granada, Spain). He has been working in legume seeds for the last 20 years, being involved in several national and international related projects. Alfonso Clemente joined different labs (Institute of Food Research, 1999–2000; John Innes Centre, 2000–2002; Sainsbury Laboratory, 2003–2004) in the UK to broaden his laboratory skills and scientific knowledge. Currently, he is the President of the Spanish Legume Association (Asociaci´ on Espa˜ nola de Leguminosas, www.leguminosas.es) having strong interaction with a relevant network of scientists and agricultural associations and agri-food companies in the field. He is author of more than 120 scientific manuscripts and an editorial board member of the World Journal of Gastroenterology and Frontiers in Bioscience, among others. Jose C. Jimenez-Lopez Bs. in Biochemistry and Molecular Biology (1998) and Bs. in Biological Sciences (2001), Ms. in Agricultural Sciences (2004), University of Granada, Spain and PhD degree in Plant Cell Biology (2008) at the Spanish National Research Council (CSIC). He developed a Full-time Postdoctoral research associate at Purdue University, USA (2008-2011). Marie Curie Research Fellow (FP7-PEOPLE2011-IOF) (2012–2015) at the University of Western Australia and CSIC working in human health benefits of legume seed proteins, their allergy molecular aspects and cross allergenicity. He is a Senior Research Fellow (Ramon y Cajal research program), currently working in the functionality, health benefits, and allergy implications of proteins from reproductive tissues (pollen and seeds) in crop species of agro-industrial interest (mainly legumes). He is an Author of more than 60 peer-review journal articles, 25 book chapters. His work has been presented in more than 130 international congresses. He is an Active member of different Scientific Societies: Spanish and International Legume Society; Spanish and EU Microscopy societies. He is Editor of multiple books, Reviewer for more than 35 Peer-Review Journals of editorial as Elsevier, Springer, Wiley, Frontiers, etc, and international expert panels member for funding grants evaluation. vii Foods 2020,9, 1525 8. Schlegel, K.; Leidigkeit, A.; Eisner, P.; Schweiggert-Weisz, U. Technological Properties of Fermented Lupin Protein Isolates. Foods 2019,8, 678. [CrossRef][PubMed] 9. EI Youssef, C.; Bonnarme, P.; Fraud, S.; Peron, A.C.; Helinck, S.; Landaud, S. Sensory Improvement of a Pea Protein-Based Product Using Microbial Co-Cultures of Lactic Acid Bacteria and Yeasts. Foods 2020 ,9, 349. [CrossRef][PubMed] 10. Kim, Y.N.; Muttakin, S.; Jung, Y.M.; Heo, T.Y.; Lee, D.U. Tailoring Physical and Sensory Properties of Tofu by the Addition of Jet-Milled, Superfine, defatted Soybean Flour. Foods 2019,8, 617. [CrossRef][PubMed] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). 3 foods Article Prebiotic Properties of Non-Fructosylated α-Galactooligosaccharides from PEA (Pisum sativum L.) Using Infant Fecal Slurries María del Carmen Marín-Manzano 1, Oswaldo Hernandez-Hernandez 2, Marina Diez-Municio 2, Cristina Delgado-Andrade 1, Francisco Javier Moreno 2and Alfonso Clemente 1,* 1Estación Experimental del Zaidín (CSIC), Consejo Superior de Investigaciones Científicas (CSIC), 18008 Granada, Spain; [email protected] (M.d.C.M.-M.); [email protected] (C.D.-A.) 2 Institute of Food Science Research (CIAL, CSIC-UAM), 28049 Madrid, Spain; [email protected] (O.H.-H.); [email protected] (M.D.-M.); javier.mor[email protected] (F.J.M.) *Correspondence: [email protected]; Tel.: +34-9-5857-2757 Received: 26 April 2020; Accepted: 8 June 2020; Published: 13 July 2020 Abstract: The interest for naturally-occurring oligosaccharides from plant origin having prebiotic properties is growing, with special focus being paid to supplemented products for infants. Currently, non-fructosylated α -galactooligosaccharides ( α -GOS) from peas have peaked interest as a result of their prebiotic activity in adults and their mitigated side-effects on gas production from colonic bacterial fermentation. In this study, commercially available non-fructosylated α -GOS from peas and β -galactooligosaccharides ( β -GOS) derived from lactose were fermented using fecal slurries from children aged 11 to 24 months old during 6 and 24 h. The modulatory effect of both GOS on different bacterial groups and bifidobacteria species was assessed; non-fructosylated α -GOS consumption was monitored throughout the fermentation process and the amounts of lactic acid and short-chain fatty acids (SCFA) generated were analyzed. Non-fructosylated α -GOS, composed mainly of manninotriose and verbascotetraose and small amounts of melibiose, were fully metabolized and presented remarkable bifidogenic activity, similar to that obtained with β -GOS. Furthermore, non-fructosylated α -GOS selectively caused an increase on the population of Bifidobacterium longum subsp. longum and Bifidobacterium catenulatum/pseudo-catenulatum. In conclusion, non-fructosylated α -GOS could be used as potential ingredient in infant formula supplemented with prebiotic oligosaccharides. Keywords: galactooligosaccharides; GOS; gut microbiota; pea; prebiotic; raffinose oligosaccharides; short-chain fatty acids (SCFA) 1. Introduction The colonisation of the gastrointestinal tract (GIT) by microorganisms is an essential process in our life cycle and starts during the gestation period [ 1 , 2 ]. Evidence shows that the GIT colonisation rapidly increases after birth by aerobic microorganisms that decrease the concentration of oxygen. This lowering of redox potentially allows the colonisation of anaerobic bacteria [ 3 ]. The continued succession of microorganisms during the first years of life has an important effect on the long-term maturity of the GIT microbiota, which is affected by diet, mode of delivery, antibiotic treatments, genetics, intestinal mucin glycosylation and other factors [4]. Non-digestible dietary ingredients are capable of modulating composition and metabolic function of gut microbiota. In this context, the prebiotic concept was recently defined by the International Scientific Association for Probiotics and Prebiotics (ISAPP) as “a substrate that is selectively utilized by host microorganisms conferring a health benefit” [ 5 ]. It is now recognized that the prebiotic effect Foods 2020,9, 921; doi:10.3390/foods9070921 www.mdpi.com/journal/foods 5 Foods 2020,9, 921 extends beyond bifidobacteria and lactobacilli and includes other genera such as Akkermansia and Faecalibacterium. These bacterial groups, among others, are involved in the production of lactic acid and short-chain fatty acids (SCFA), which perform crucial physiological functions as metabolic regulators and immunomodulators, decreasing the growth of potential intestinal pathogens. The SCFA are responsible for the decrease of pH in the intestinal lumen, an increase of calcium absorption and shortening of gastrointestinal transit time. Besides, SCFA play an important role in cross-feeding processes and are major sources of energy for colonic epithelial cells and other cell types located in peripheral tissues [4,6]. The human-milk oligosaccharides (HMOs) are one of the most important dietary ingredients modulating the infant microbiome. Only certain bifidobacterial species, such as Bifidobacterium longum subsp. infantis, are able to utilize HMOs as metabolic substrates [ 7 ] with specific preferences depending on the Bifidobacterium strain and the HMOs structure [ 8 , 9 ]. Due to different clinical and social conditions, human milk is not a realistic choice for some infants and, in such cases, infant formulas are the most suitable alternative. However, access to HMOs for infant formula is extremely limited [ 10 ]. As a result, different oligosaccharides have been utilised to mimic the function of HMOs, such as fructooligosaccharides (FOS) and β -galactooligosaccharides ( β -GOS). Other natural sources have also been utilised to obtain carbohydrates for infant formulas with well reported prebiotic activities such as inulin [ 11 ], pectin-oligomers [ 6 ], raffinose [ 12 ] and resistant starch [ 13 ]. For infant formulas, research has mainly focused on inulin, fructooligosaccharides (FOS) and β -galactooligosaccharides (GOS); indeed, there is still a wide range of oligosaccharides that could mimic, at least to some extent, the functional properties of HMOs. Plant-based GOS, with α -galactosidic linkages instead of β -linkages, are water soluble carbohydrates and are particularly abundant in legume seeds as soybean, chickpea, lentil, faba bean and pea. They are commonly named raffinose family oligosaccharides (RFOS) and show a terminal sucrose unit linked by the glucose monomer to galactoses via α -(1 → 6) linkages. Like β -GOS, α -GOS are not hydrolysed in the upper gastrointestinal tract due to the absence of the enzyme α -galactosidase expressed by somatic cells in the mammalian GIT [ 14 ]. As a result, they reach the large intestine where are fermented by gut microbiota and can exert prebiotic properties [ 15 ]. However, it is well-known that RFOS cause discomfort, bloating and flatulence to consumers due to the presence of the ending fructose monomer [16,17]. Non-fructosylated α -GOS include melibiose (CAS 585-99-9), manninotriose (CAS 13382-86-0) and verbascotetraose (CAS 1111-08-6), which are essentially raffinose, stachyose and verbascose without the ending fructose units. Non-fructosylated α -GOS from peas is produced by the activity of β -fructosidases which are able to split sucrose from the α -GOS chain into glucose and fructose, being commercially available as AlphaGOS ® . This particular α -GOS mixture has claimed to reduce the post-prandial glycaemic responses, being recently approved by the European Food Safety Agency (EFSA) [ 18 ], and exert prebiotic activity in adult faecal inoculum [ 19 ]. Available scientific data suggest that the administration of prebiotic oligosaccharides to healthy infants does not raise safety concerns with regards to adverse effects. In this sense, non-fructosylated α -GOS has been proven safe in infant formula in a concentration up to 8 mg/mL after preclinical evaluation in neonatal piglets [20]. Despite the various scientific evidence available regarding non-fructosylated α -GOS derived from peas, no reports have been published so far regarding the effect of these plant oligosaccharides on infant microbiota either in vitro or in vivo . Taking into account the published data reporting the bifidogenic properties of non-fructosylated α -GOS in adults, and its safety in infant formula, we hypothesized that a similar effect could be also found in infants, despite the differences between GIT microbiota in adults and in infants. Such increase in the number of bifidobacteria might be considered a major shift in the gut microbiota towards a potentially healthier composition. Important functions have been attributed to bifidobacteria including a protective role against pathogens, promote gut epithelium integrity and modulate the host immune system. Therefore, the main aim of the study was to evaluate the effect of 6 Foods 2020,9, 921 non-fructosylated α -GOS in the microbiota present in infant faecal samples with particular attention paid to their bifidogenic properties. 2. Material and Methods 2.1. Chemicals MTBSTFA (N-terbutil-dimetil-silil-N-metil-trifluoroacetamid) was purchased by Fluka Chemie AG (Buchs, Switzerland). Short chain fatty acids (SCFA) and organic acids were obtained from Sigma-Aldrich. All other chemicals were of analytical grade. 2.2. Non-Fructosylated α-GOS and β-Galactooligosacharides (GOS) Characterization Non-fructosylated α -galactooligosaccharides ( α -GOS, commercial brand AlphaGOS ® ) derived from pea seeds were kindly provided by Olygose (Venette, France). The chemical composition of α-GOS is detailed in Table 1and shows as major components mannitriose and verbascotetraose with degree of polymerization of three (DP3) and four (DP4), respectively. Table 1. Chemical composition of AlphaGOS®. Parameter % (Dry Matter) Dry matter a96.3 Crude proteins a<0.2 Crude ash a<0.2 DP1 +Maltose +Sucrose b<0.2 DP2: Melibiose b3.9 DP3: Manninotriose b49.2 DP4: Verbascotetraose b43.0 aProvided by the supplier; bquantified by GC-FID (see Section 2.10 for details); DP: degree of polymerization. An industrially available galacto-oligosaccharide mixture derived from lactose ( β -GOS) was used in this study for comparative purposes as prebiotic reference. This mixture was initially composed by galactose (1%), glucose (21%), DP2 (37%), DP3 (22%), DP4 (11%), >DP4 (8%). Gel filtration chromatography was carried out for the removal of monoand disaccharides due to the presence of high levels of digestible carbohydrates including lactose; the DP of collected fractions was determined by electrospray ionization mass spectrometry (ESI-MS). The purified β -GOS mixture consisted in a complex mixture of oligosaccharides from DP3 to DP6 and fully free of monoand disaccharides. Thetrisaccharidefractionof β -GOS(35.2%oftotalcarbohydrates)containedmainly4  -galactosyl-lactose (15.8%), 6  -galactosyl-lactose (5.3%), as well as other minor galactobioses linked to the reducing glucose unit by β-(1→2) and β-(1→6) glycosidic linkages (Supplementary Materials Table S1) [21]. 2.3. Infant Faecal Samples Parents/guardians of children of the Los Angeles Nursery (Granada, Spain) were invited to attend a meeting where the nature of the study was explained and those who agreed to participate signed the informed consent. Parents/guardians were given a small survey to learn about the overall health and eating habits of donors. The participants fed on solid food and none of them had taken probiotics, fermented foods, antibiotics or received medical treatment that could affect the intestinal microbiota at least a month before the stool samples were collected. The study protocol was conducted in accordance with the ethical recommendations of the Declaration of Helsinki. Stoolsampleswerecollectedfromeightchildrenintheagerangebetween11–24months. Theywere taken by our laboratory staffafter deposition in nappies, avoiding the edges and the area in contact with the nappy and also avoiding urine that was absorbed in the cellulose. Immediately after deposition, fecal samples were housed in closed jars containing Anaerocult A tablets (Anaerocult ® , Darmstad, 7 Foods 2020,9, 921 Germany) in order to maintain anaerobic conditions during the transport to the laboratory. The fecal samples (~5 g), within 2 h after collection, were homogenized in a stomacher by dilution 1/10 (w/v) with anaerobic 0.1 M sodium phosphate buffer, pH 6.8, by using sterile plastic filter bags, and filtrated through Miracloth (Calbiochem, Darmstad, Germany) to remove solid residues. 2.4. In Vitro Faecal Fermentation Faecal homogenates were transferred into sterile Hungate tubes containing a specific basal medium for cultivation of infant faeces (BMIF) by dilution 1/10 (v/v)[ 22 ] and were stabilized overnight under anaerobic conditions at 37 ◦ C. After that, 0.3% (w/v) of specific substrate (non-fructosylated α -GOS and β -GOS) was added to stabilized faecal mixtures. Fermentation experiment for the fecal sample of each subject was carried out once for each GOS-type. An additional tube was kept without inoculum and without GOS as a negative control and the positive control contained inoculum but not GOS. In addition, the medium was supplemented with 1/1000 (w:v) of resazurine to control the strict anaerobiosis conditions. Fermentations were carried out in anaerobic chambers at 37 ◦ C and samples were taken at time 0, 6 and 24 h. One mL of culture was centrifuged (12,000 × gfor 15 min). The pellet was frozen at − 80 ◦ C and freeze-dried for the study of bacterial populations; the supernatant was used for direct determination of pH at the different collection times (Crison Instruments S.A., Barcelona, Spain) and in order to quantify lactic acid and SCFA. 2.5. DNA Extraction DNA was extracted from pellets of fecal batch cultures using FavorPrep ™ Stool DNA Isolation Mini Kit (Favorgene Biotech Corp, Shuttleworthstraße, Vienna, Austria). A NanoDrop ND-100 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) was used for DNA quantification; purified DNA samples were stored at −80 ◦C. 2.6. Microbial Faecal Populations Determined by Quantitative PCR (qPCR) Analysis Quantitative PCR was used to evaluate the effect of non fructosylated α -GOS and β -GOS on microbial composition in fecal homogenates after 6 and 24 h of in vitro fermentation in comparison with the control group (absence of oligosaccharides). Different microbial groups including total bacteria, Bacteroides, lactobacilli, bifidobacteria, Eubacterium rectale/Clostridium coccoides,Clostridium leptum, enterobacteria and Faecalibacterium prausnitzii were distinguished and quantified using qPCR. The 16S rRNA gene-targeted group-specific primers used in this study are listed in Supplementary Materials Table S2. qPCR conditions used in this study were as previously reported [ 23 ]. For quantitative analysis of the different bifidobacteria species (B. adolescentis,B. bifidum,B. catenulatum/pseudo-catenulatum, B. breve,B. longum subsp. infantis and B. longum subsp. longum), the PCR conditions were one cycle of 94 ◦ C for 5 min, then 40 cycles of 94 ◦ C for 20 s, 55 ◦ C for 20 s, and 72 ◦ C for 1 min. The species-specific primer pair for bifidobacteria is listed in Supplementary Materials Table S3. In the case of Eubacterium rectale/Clostridium coccoides and Clostridium leptum groups, PCR conditions were an initial denaturation step at 94 ◦ C for 5 min followed by 40 cycles at 94 ◦ C for 20 s, 50 ◦ C for 20 s and 72 ◦ C for 1 min for primer annealing and product elongation [ 23 ]. The fluorescent product was detected in the last step of each cycle. Following amplification, melting temperature analysis of PCR products was performed to determine the specificity of the PCR. The melting curves were obtained by slow heating at 0.5 ◦ C increments from 55 to 95 ◦ C, with continuous fluorescence collection. A plasmid standard containing the target region was generated for each specific primer set using DNA extracted from fecal homogenates. The amplified products were cloned using the TOPO TA cloning kit for Sequencing (Invitrogen, Barcelona, Spain) and transformed into E. coli One Shot Top 10 cells (Invitrogen). Sequences were submitted to the ribosomal RNA database to confirm the specificity of the primers. For quantification of target DNA copy number, standard curves were generated using serial 10-fold dilutions of the extracted products by using at least six non-zero standard concentrations per assay. The bacterial concentration in each sample was measured as log 10 copy number by the 8 Foods 2020,9, 921 interpolation of the C t values obtained by the fecal homogenates samples and the standard calibration curves. Each plate included triplicate reactions per DNA sample and the appropriate set of standards. 2.7. Short-Chain Fatty Acids (SCFA) and Lactic Acid Analysis SCFA and lactic acid were analyzed as described previously by Tabasco et al. [ 24 ], with some modifications. For quantitative analysis, 55 μ L of a 100 mM 2-ethylbutyric acid in distilled water was added as internal standard to 550 μ L of bacterial supernatant after in vitro fermentation. SCFA and organic acids were extracted by the addition of 275 μ L concentrated HCl and 1 mL diethyl ether followed by vortexing 1 min. Then, samples were centrifuged at 3000 × gfor 10 min and 50 μ L of the ether layer was transferred to a GC-microvial to minimize ether evaporation, adding 10 μ L of the derivatization agent MTBSTFA. Hermetically stoppered microvials were heated at 80 ◦ C for 20 min, afterwards the reaction mixture was kept for 24 or 48 h at room temperature to ensure total derivatization. An external calibration curve was built using a standard solution mixture made from pure compounds (formate 10 mM, acetate 60 mM, propionate 20 mM, butyrate 20 mM, iso-butyrate 5 mM, valerate 5 mM, iso-valerate 5 mM, lactate 17 mM, succinate 17 mM) and derivatized as described for samples. A gas chromatograph Thermo Scientific (Trace GC Ultra, Rodena, Italy) equipment was used, with a mass detector ion trap (Thermo Scientific ITQ 900, Rodena, Italy) and with an automatic sampling system (Triplus). The chromatograph was equipped with a Thermo Scientific column (TR-5MS 30 m × 0.25 mm × 0.25 μ m) and a 2.5 m × 1 μ m 0.32 ID pre-column (Teknokroma, Guard Column TR30005, San Cugat del Vall é s, Spain). The samples were introduced via split (ratio 50:1) injection with the port heated to 275 ◦ C. Helium was used as the carrier gas at a flow rate of 1.0 mL/min. The oven temperature was initially held at 63 ◦ C for 3 min, increased to 190 ◦ C with a 20 ◦ C per min rate, held at 190 ◦ C for 6 min, then raised to 230 ◦ C with a 40 ◦ C per min rate, where it was held for 1 min. The mass spectrometer interface temperature was set to 250 ◦ C. For monitoring and confirmation analysis, the electronic impact (EI, 70 eV) mode was used. For the MS/MS experiments, the adequate ion precursor and a collision energy was selected for each compound, in order to perform the quantification. Identification and quantification of SCFA and organic acids was based on the use of the relative response factors calculated for the target compounds in the standard solutions at different concentrations against the IS. Measurements were conducted at least twice. 2.8. PCR-Denaturing Gradient Gel Electrophoresis (DGGE) Analysis of Bifidobacteria The 16S rRNA genes were amplified by PCR from extracted DNA of pellets from fecal batch cultures of infants. The Bifidobacterium genus-specific primers (Bif164-F and Bif662-GC-R) and PCR amplification conditions were as previously reported [ 23 ]. PCR fragments were separated by DGGE by using a denaturing gradient of 40 to 65%. The gels were visualized by silver-staining, dried at 37 ◦ C and scanned. The total number and dendogram of similarity cluster analysis of DGGE bands were determined by Quantity One analysis software (BioRad). DGGE profiles were determined by cluster analysis using the Dice similarity coefficient and the unweighted-pair group method by means of arithmetic average clustering algorithm (UPGMA). The richness (S) of the bifidobacterial community was established from the number of electrophoretic bands in individual samples. Shannon index (H), was calculated as reported by Magurran [ 25 ] as: H = − (pi · lnpi), where pi is the abundance of every species. The evenness (E) of the bacterial community was further estimated as E =H/lnS [26]. 2.9. Excision and Sequencing of DGGE Bands The DGGE bands were excised and eluted by using sterilised distilled water at 4 ◦ C. The primers Bif164-F and Bif662-R were used to amplify the bands of interest as previously described [ 23 ]. PCRproductswerepurified and thencloned usingtheTOPOTACloning kitfor Sequencing(Invitrogen). Plasmids DNA were isolated from selected transformants with the GenElute Plasmid Miniprep kit (Sigma-Aldrich, St. Louis, USA), being inserts sequenced. Searches for sequence similarity were carried out using the BLAST algorithm [ 27 ] of the GenBank database Release 232.0 (www.ncbi.nlm.nih.gov) 9 Foods 2020,9, 921 for identification of the nearest relatives of the partial 16S rRNA sequences. A sequence similarity ≥98% of the 16S rRNA gene was used as the criterion for identification of bifidobacterial species. 2.10. Carbohydrate Quantification by Gas Chromatography Coupled to Flame-Ionization Detector (GC-FID) The carbohydrate concentration before and after in vitro fecal fermentation was analysed by GC-FID. The carbohydrate fraction was derivatized to their corresponding trimethylsilyl oximes (TMSO) following the method of Brobst and Lott [ 28 ]. The samples were dried and the oximes were formed by adding 350 μ L of hydroxylamine chloride in pyridine (2.5% w/v)at70 ◦ C for 30 min. The resulting oximes were silylated with hexamethyldisilazane (350 μ L) and trifluoroacetic acid (35 μ L) at 50 ◦ C for 30 min. The mixtures were centrifuged at 7000 × gfor 4 min. The supernatants were stored at a temperature of 4 ◦ C prior to analysis. The chromatography separation was carried out in an Agilent Technologies gas chromatograph (Mod 7890A, Santa Clara, USA) with a fused silica capillary column DB-5HT (5%-phenyl-methylpolysiloxane; 30 m × 0.25 mm × 0.10 μ m) (Agilent). The oven temperature was set to 150 ◦ C and then increased to 380 ◦ C at a rate of 3 ◦ C/min. The injector and detector temperatures were set to 280 ◦ C and 385 ◦ C, respectively. One mL/min of nitrogen was used as carrier gas and the injections were performed in split mode (1:20). Data acquisition and integration were performed using the Agilent ChemStation software. 2.11. Statistical Analysis The effect of non-fructosylated α -GOS and β -GOS on microbiota composition and SCFA amounts of fecal contents was analyzed using a linear mixed model: repeat measure (SPSS Statistics version 22.0, Madrid, Spain). The Bonferroni method with a pvalue ≤ 0.05 was used for adjustments for major effects and the time 0 h was used like a covariate. 3. Results and Discussion 3.1. Microbiota Composition after In Vitro Fermentation of Non-Fructosylated α -GOS and β -GOS with Infant Fecal Samples Using infant fecal slurries, all bacteria and up to seven different bacterial groups were analysed before and after 6 and 24 h of the in vitro fermentation of non-fructosylated α -GOS and β -GOS. Table 2 shows the population of these bacterial groups, which is the average of data resulting from eight infant donors. During treatment, the control samples (in the absence of oligosaccharides) did not show significant differences, with the exception of bifidobacteria and enterobacteria that slightly increased after 6 and 24 h. Bifidobacterium spp. increased significantly in non-fructosylated α -GOS and β -GOS groups but no differences were observed between 6 and 24 h treatment. Indeed, the bifidogenic effect in non-fructosylated α -GOS and β -GOS groups was significantly greater than that found in control samples. To the best of our knowledge, no data has been previously reported regarding the effect of non-fructosylated α -GOS in the modulation of infants’ microbiota, using either in vivo or in vitro systems. On the contrary, the bifidogenic properties of β -GOS in infants’ microbiota is well documented. Thus, the bifidogenic activity of β -GOS using infant fecal slurries in a three-stage in vitro culture method that mimics the proximal, transversal and distal colon has been reported [ 29 ]. A strong bifidogenic activity in an in vivo intervention study in infants by using a formula-fed containing β -GOS was also reported [ 30 ]. A significant decrease in enterobacteria was observed in non-fructosylated α -GOS and β -GOS samples, while a slight increase was found in the control group. C. coccoides/E. rectale group, Lactobacillus spp., C. leptum group, F. praustnizii, Bacteroides and total bacteria did not show significant differences in both non fructosylated α-GOS and β-GOS treatments (Table 2). 10 Foods 2020,9, 921 Table 2. Microbiota population from in vitro-fermented infant faecal samples with non-fructosylated α-GOS and commercial β-GOS derived from lactose. Control α-GOS β-GOS p-Values Pooled SEM Log10 Copy Number/ Gr Dry Faeces Bs 6 h 24 h Bs 6 h 24 h Bs 6 h 24 h T Prebiotics T * Prebiotics Bs All bacteria 9.78 9.83 9.75 9.77 9.79 9.79 9.89 9.77 9.83 0.871 0.316 0.600 <0.001 0.027 Bifidobacteria spp. 7.49 7.59 a7.68 a8.01 8.59 b8.62 b7.61 8.18 b8.35 b0.285 <0.001 0.823 <0.001 0.043 Clostridium coccoides/ Eubacterium rectale group 8.54 8.60 8.54 8.63 8.61 8.66 8.61 8.53 8.55 0.994 0.144 0.606 <0.001 0.020 Lactobacilli spp. 5.01 4.78 4.40 5.03 4.86 5.02 5.02 4.69 4.80 0.788 0.150 0.208 <0.001 0.068 Clostridium leptum subgroup 7.57 7.67 7.73 7.72 7.68 7.69 8.00 7.93 7.88 0.900 0.049 0.719 <0.001 0.026 Enterobacteria spp. 7.81 7.85 b7.88 b8.24 8.12 a, b 8.12 a, b 7.93 7.69 a7.70 a0817 <0.001 0.967 <0.001 0.028 F. praustnizii 8.13 8.14 8.12 8.20 8.09 8.10 8.13 8.04 7.99 0.716 0.100 0.893 <0.001 0.023 Bacteroides 9.63 9.67 9.61 9.67 9.59 9.60 9.63 9.51 9.54 0.855 0.065 0.707 <0,001 0.021 Outcomes were analyzed using a linear mixed model: repeat measure. A Bonferroni method was used for adjustments for major effects (time of treatment and type of GOS) a,b Mean value with different letters means that differences among treatments was significantly different at p-value <0.05. Unadjusted mean values n=8. Bs: Baseline like covariate. * Interaction among major effects. 11 Foods 2020,9, 921 Analysis of different bacterial groups after in vitro fermentation of non-fructosylated α -GOS and β -GOS suggest that these oligosaccharides are preferentially metabolized by bifidobacteria. Although this study is mainly focused in the bifidogenic properties of both GOS types, prebiotic targets extend beyond stimulation of bifidobacterial and lactobacilli, and recognizes that health benefits can derive from effects on other beneficial taxa. Such is the case of F. praustnizzii and the Clostridium coccoides/Eubacterium rectale bacterial group that includes species that are known butyrate-producers, thereby contributing to important processes linked to colonic health, including the protection against inflammatory bowel diseases. As reported in Table 2, in vitro fermentation of non-fructosylated α -GOS and β -GOS did not affect the growth of these butyrate-producers whilst enterobacteria numbers significantly decreased. In this sense, further studies are planned to determine the preventive role of non-fructosylated α-GOS in the colonization of pathogenic enterobacteria. The modulation of gut microbiota by RFOS has been previously evaluated using in vitro fermentation batch cultures, inoculated with a standardised human adult faecal sample; a higher increase of Bifidobacterium spp. using RFOS when compared with FOS and β -GOS was observed, being the bifidogenic effect dose-dependent [ 19 ]. However, to the best of our knowledge, there is not specific information regarding the potential prebiotic properties of non-fructosylated α -GOS. Recently, the prebiotic potential of melibiose-derived gluco-oligosaccharides linked by α -(1 → 3) and α -(1 → 6) glycosidic bonds as unique carbon source has been evaluated by testing the growth of bifidobacteria (B. breve,B. longum subsp. longum,B. animalis) and lactobacilli (L. reutieri,L. plantarum and L. rhamnosus)[ 31 ]. In agreement with our data, lactobacilli were unable to metabolize melibiose-derived oligosaccharides whereas bifidobacteria were able to utilize them. Although no data regarding the modulatory effect of non-fructosylated α –GOS in infant fecal microbiota has been previously reported, it has been suggested they could be safely used in infant formulas. In a study carried out with neonatal piglets to mimic infancy conditions, an intervention during three weeks having a daily consumption of 3.6–3.9 g non-fructosylated GOS/kg body weight demonstrated to be well tolerated with no adverse effects in terms of clinical signs, body weight, feed consumption, haematology, organ weight and histopathology [20]. 3.2. Effect of Non-Fructosylated α-GOS on Bifidobacterial Composition As previously pointed out, non-fructosylated α -GOS and β -GOS had a strong bifidogenic effect in infant inoculum after 6 h of fermentation and was kept at 24 h of treatment. To obtain a more comprehensive assessment of the impact of both GOS on the population structure of bifidobacteria in children’s faecal samples, 16S rRNA gene profiles were generated by means of genus-specific primers PCR-DGGE. Control samples reflected an inter-individual variation in faecal bifidobacterial community being grouped in two cladograms with electrophoretic bands differing from 3 up to 11 (Supplementary Materials Figure S1). Dendrogram analysis indicated a grouping of the samples by children but they were not clustered after non-fructosylated α -GOS and β -GOS treatment (Figure 1). Theelectrophoreticbandswereexcisedforsequencingandbifidobacterialspeciesidentified. Sequencing allowed us to identify B. catenulatum/pseudocatelunatum,B. longum subsp. longum and B. longum subsp. infantis as the most frequently present bifidobacterial species in infants. Whereas B. longum subsp. infantis was clearly identified, B. longum subsp. longum did not achieved such level of recognition at subspecies level and might be an experimental limitation that needs to be considered. The presence of B. adolescentis and B. bifidum was identified in only one of the infants. These species have been reported in many studies as the dominant bifidobacteria in infant faecal samples. Significant differences in bifidobacteriacommunitiesdisplay differentialmetabolic featuresandhavestrongimplicationsin infant physiology state and health [ 32 – 34 ]. Delivery mode (vaginal birth or assisted delivery) and feeding modes (breastfeeding, milk formula and mix-fed) have influences on bifidobacterial composition at species level [ 32 ]. Significant changes in the gut microbiota of infants occur when cessation of breastfeedingandintroductionof solidfoodscauseanimportantshift ingutmicrobiotacompositionand beginsto resembleastable adult-likemicrobiome[ 35 ]. Under thesecircumstances, the supplementation 12 Foods 2020,9, 921 2. Rodriguez, M.A.; Murphy, K.; Stanton, C.; Ross, R.P.; Kober, O.I.; Juge, N.; Avershina, E.; Rudi, K.; Narbad, A.; Jenmalm, M.C.; et al. The composition of the gut microbiota throught life, with emphasis on early life. Microb. Ecol. Health Dis. 2015,26, 26050. [PubMed] 3. Houghteling, P.D.; Walker, W.A. Mucosal why is initial bacterial colonization of the intestine important to the infant’s and child’s health? J. Pediatr. Gastroenterol. Nutr. 2008,6, 2166–2171. 4. Verkhnyatskaya, S.; Ferrari, M.; De Vos, P.; Walvoort, M.T.C. Shaping the infant microbiome with non-digestible carbohydrates. Front. Microbiol. Res. Top. 2019,10, 1–8. [CrossRef] 5. Gibson, G.R.; Hutkins, R.; Sanders, M.E.; Prescott, S.L.; Salminen, S.J.; Scott, K.; Stanton, C.; Cani, P.D.; Verbeke, K.; Reid, G. The International Scientific Association for Probiotics and Prebiotics (ISAPP) consensus statement on the definition and scope of prebiotics. Nat. Rev. Gastroenterol. Hepatol. 2017 ,14, 491–502. [CrossRef] 6. Akkerman, R.; Faas, M.M.; de Vos, P. Non-digestible carbohydrates in infant formula as substitution for human milk oligosaccharide functions: Effects on microbiota and gut maturation. Crit. Rev. Food Sci. Nutr. 2019,59, 1486–1497. [CrossRef][PubMed] 7. Marcobal, A.; Barboza, M.; Froehlich, J.W.; Block, D.E.; German, J.B.; Lebrilla, C.B.; Mills, D.A. Consumption of human milk oligosaccharides by gut-related microbes. J. Agric. Food Chem. 2010 ,58, 5334–5340. [CrossRef] [PubMed] 8. Ruiz-Moyano, S.; Totten, S.M.; Garrido, D.A.; Smilowitz, J.T.; Bruce German, J.; Lebrilla, C.B.; Mills, D.A. Variation in consumption of human milk oligosaccharides by infant gut-associated strains of Bifidobacterium breve.Appl. Environ. Microbiol. 2013,79, 6040–6049. [CrossRef][PubMed] 9. Wang, Y.; Jiang, K.; Ma, H.; Zeng, W.; Wang, P.G.; Yao, N.; Han, W.; Cheng, J.; Wang, W. Enzymatic production of HMO mimics by the sialylation of galacto-oligosaccharides. Food Chem. 2015 ,181, 51–56. [CrossRef] [PubMed] 10. Barile, D.; Rastall, R.A. Human milk and related oligosaccharides as prebiotics. Curr. Opin. Biotechnol. 2013 , 24, 214–219. [CrossRef][PubMed] 11. Oswari, H.; Widodo, A.D.; Handayani, F.; Juffrie, M.; Sundjaya, T.; Bindels, J.; Hegar, B. Dosage-related prebiotic effects of inulin in formula-fed infants. Pediatr. Gastroenterol. Hepatol. Nutr. 2019 ,22, 63–71. [CrossRef][PubMed] 12. Ehara, T.; Izumi, H.; Tsuda, M.; Nakazato, Y.; Iwamoto, H.; Namba, K.; Takeda, Y. Combinational effects of prebiotic oligosaccharides on bifidobacterial growth and host gene expression in a simplified mixed culture model and neonatal mice. Br. J. Nutr. 2016,116, 270–278. [CrossRef][PubMed] 13. Gopalsamy, G.; Mortimer, E.; Greenfield, P.; Bird, A.R.; Young, G.P.; Christophersen, C.T. Resistant starch is actively fermented by infant faecal microbiota and increases microbial diversity. Nutrients 2019 ,11, 1–16. [CrossRef][PubMed] 14. Hernandez-Hernandez, O.; Olano, A.; Rastall, R.; Moreno, F.J. In Vitro digestibility of dietary carbohydrates: Toward a standardized methodology beyond amylolytic and microbial enzymes. Front. Nutr. 2019 ,6, 61. [CrossRef] 15. Van den Ende, W. Multifunctional fructans and raffinose family oligosaccharides. Front. Plant. Sci. 2013 ,4, 1–11. 16. Myhara, R.M.; Nilsson, K.; Skura, B.J.; Bowmer, E.J.; Cruickshank, P.K. Gas production from melibiose, raffinose and white bean extracts by bacteria of human fecal origin. Can. Inst. Food Sci. Technol. 1988 ,21, 245–250. [CrossRef] 17. Martins, G.N.; Ureta, M.M.; Tymczyszyn, E.E.; Castilho, P.C.; Gomez-Zavaglia, A. Technological aspects of the production of fructo and galacto-oligosaccharides. Enzymatic synthesis and hydrolysis. Front. Nutr. 2019,6, 78. [CrossRef][PubMed] 18. EFSA NDA Panel (EFSA Panel on Dietetics Products, Nutrition and Allergies). Scientific Opinion on the substantiation of a health claim related to AlphaGOS ® and a reduction of post-prandial glycaemic responses pursuant to Article 13(5) of Regulation (EC) N◦1924/2006. EFSA J. 2014,12.[CrossRef] 19. Fehlbaum, S.; Prudence, K.; Kieboom, J.; Heerikhuisen, M.; van den Broek, T.; Schuren, F.H.J.; Steinert, R.E.; Raederstorff, D. In Vitro fermentation of selected prebiotics and their effects on the composition and activity of the adult gut microbiota. Int. J. Mol. Sci. 2018,19, 3097. [CrossRef] 19 Foods 2020,9, 921 20. Kruger, C.; Zhou, Y.; Thorsrud, B.A.; Morel-Despeisse, F.; Chappuis, E. Safety evaluation of α -galacto-oligosaccharides for use in infant formulas investigated in neonatal piglets. Toxicol. Res. Appl. 2017,1, 1–10. [CrossRef] 21. Hernandez-Hernandez, O.; Mar í n-Manzano, M.C.; Rubio, L.A.; Moreno, F.J.; Sanz, M.L.; Clemente, A. Monomer and linkage type of galacto-oligosaccharides affect their resistance to ileal digestion and prebiotic properties in rats. J. Nutr. 2012,142, 1232–1239. [CrossRef] 22. Arboleya, S.; Binetti, A.; Salazar, N.; Fern á ndez, N.; Solis, G.; Hern á ndez-Barranco, A.; Margolles, A.; de los Reyes-Gavil á n, C.G.; Gueimonde, M. Establisment and development of intestinal microbiota in preterm neonates. FEMS Microbiol. Ecol. 2012,79, 763–772. [CrossRef] 23. Mar í n-Manzano, M.C.; Abecia, L.; Hern á ndez-Hern á ndez, O.; Sanz, M.L.; Montilla, A.; Olano, A.; Rubio, L.A.; Moreno, F.J.; Clemente, A. Galacto-oligosaccharides derived from lactulose exert a selective stimulation on the growth of Bifidobacterium animalis in the large intestine of growing rats. J. Agric. Food Chem. 2013 ,61, 7560–7567. [CrossRef] 24. Tabasco, R.; Fern á ndez de Palencia, P.; Fontecha, J.; Pel á ez, C.; Requena, T. Competition mechanisms of lactic acid bacteria and bifidobacteria: Fermentative metabolism and colonization. Food Sci. Technol. 2014 ,55, 680–684. [CrossRef] 25. Magurran, A.E. Ecological Diversity and its Measurements; Princenton University Press: Princenton, NJ, USA, 1988; p. 179. 26. Pielou, E.C. Ecological Diversity; John Wiley & Sons, Inc.: New York, NY, USA, 1975; p. 165. 27. Altschul, S.F.; Gish, W.; Myers, E.W.; Lipman, D.J. Basic local alignment search tool. J. Biol. Mol. 1990 ,215, 403–410. [CrossRef] 28. Brobst, K.; Lott, C.E. Determination of some components in corn syrup by gas-liquid chromatography of trimethylsilyl derivatives. Cereal Chem. 1966,43, 35–43. 29. Grimaldi, R.; Cela, D.; Swann, J.R.; Vulevic, J.; Gibson, G.R.; Tzortzis, G.; Costabile, A. In Vitro fermentation of β -GOS: Impact of faecal bacterial populations and metabolic activity in autistic and non-autistic children. FEMS Microbiol. Ecol. 2017,93, 1–10. [CrossRef][PubMed] 30. Borewicz, K.; Suarez-Diez, M.; Hechler, C.; Beijers, R.; de Weerth, C.; Arts, I.; Penders, J.; Thijs, C.; Nauta, A.; Lindner, C.; et al. The effect of prebiotic fortified infant formulas on microbiota composition and dynamics in early life. Sci. Rep. 2019,9, 1–13. [CrossRef] 31. Ispirli, I.; Kaya, Y.; Dertli, E. Bifidogenic effect and In Vitro immunomodulatory roles of melibiose-derived oligosaccharides produced by the acceptor reaction of glucansucrase E81. Proc. Biochem. 2020 ,91, 126–131. [CrossRef] 32. Matsuki, T.; Watanabe, K.; Tanaka, R. Genusand species-specific PCR primers for the detection and identification of bifidobacterial. Curr. Issues Intest. Microbiol. 2003,4, 61–69. 33. Turroni, F.; Peano, C.; Pass, D.A.; Foroni, E.; Severgnini, M.; Claesson, M.J.; Kerr, C.; Hourihane, J.; Murray, D.; Fuligni, F. Diversity of Bifidobacteria Within the Infant Gut Microbiota. PLoS ONE 2012 ,7, e36957. [CrossRef] [PubMed] 34. Yang, B.; Chen, Y.; Stanton, C.; Ross, R.P.; Lee, Y.K.; Zhao, J.; Zhang, H.; Chen, W. Bifidobacterium and Lactobacillus composition at species level and gut microbiota diversity in infants before 6 weeks. Int. J. Mol. Sci. 2019,20, 3306. [CrossRef] 35. Berström, A.; Skoy, T.H.; Bahl, M.I.; Roager, H.M.; Christensen, L.B.; Ejlerskov, K.T.; Molgaard, C.; Michaelsen, K.F.; Licht, T.R. Establishment of intestinal microbiota during early life: A longitudinal, explorative study of a large cohort of Danish infants. App. Environm. Microbiol. 2014 ,80, 2889–2900. [CrossRef][PubMed] 36. Braegger, C.; Chmielewska, A.; Decsi, T.; Kolacek, S.; Mihatsh, W.; Moreno, L.; Piescik, M.; Puntis, J.; Shamir, R.; Szajewska, H.; et al. Supplementation of infant formula with probiotics and/or prebiotics: A systematic review and comment by the ESPGHAN committee on nutrition. J. Pediatr. Gastroenterol. Nutr. 2011,52, 238–250. [CrossRef] 37. Bindels, L.B.; Delzenne, N.M.; Cani, P.D.; Walter, J. Towards a more comprehensive concept for prebiotics. Nat. Rev. Gastroenterol. Hepatol. 2015,12, 303–310. [CrossRef] 38. Hutkins, R.W.; Krumbeck, J.A.; Bindels, L.B.; Cani, P.D.; Fahey, G.; Goh, Y.J.; Hamaker, B.; Martens, E.C.; Mills, D.A.; Rastall, R.A.; et al. Prebiotics: Why definitions matter. Curr. Opin. Biotechnol. 2016 ,37, 1–7. [CrossRef][PubMed] 20 Foods 2020,9, 921 39. Davis, L.M.; Martinez, I.; Walter, J.; Goin, C.; Hutkins, R. Barcoded pyrosequencing reveals that consumption of galactooligosaccharides results in a highly specific bifidogenic response in humans. PLoS ONE 2011 ,6, e25200. [CrossRef][PubMed] 40. Pokusaeva, K.; Fitzgerald, G.F.; van Sinderen, D. Carbohydrate metabolism in Bifidobacteria. Genes Nutr. 2011,6, 285–306. [CrossRef][PubMed] 41. El Kaoutari, A.; Armougom, F.; Gordon, J.I.; Raoult, D.; Henrisat, B. The abundance and variety of carbohydrate-active enzymes in the human gut microbiota. Nat. Rev. Microbiol. 2013 ,11, 497–504. [CrossRef] 42. Canfora, E.E.; Jocken, J.W.; Blaak, E.E. Short-chain fatty acids in control of body weight and insulin sensitivity. Nat. Rev. Endocrinol. 2015,11, 577–591. [CrossRef] 43. Flint, H.J.; Scott, K.P.; Louis, P.; Duncan, S.H. The role of the gut microbiota in nutrition and health. Nat. Rev. Gastroenterol. Hepatol. 2012,9, 577–589. [PubMed] 44. Ventura, M.; O’Connell-Motherway, M.; Leahy, S.; Moreno-Munoz, J.A.; Fitzgerald, G.F.; van Sinderen, D. From bacterial genome to functionality; case bifidobacteria. Int. J. Food Microbiol. 2007 ,120, 2–12. [CrossRef] [PubMed] 45. Alexander, C.; Swanson, K.; Fahey, G.C.; Garleb, K.A. Physiological importance of short-chain fatty acids from undigestible carbohydrate fermentation. Adv. Nutr. 2019,10, 576–589. [CrossRef][PubMed] 46. Ejby, M.; Fredslund, F.; Andersen, J.M.; Žagar, A.V.; Henriksen, J.R.; Andersen, T.L.; Svensson, B.; Slotboom, D.J.; Hachem, M.A. An ATP binding cassette transporter mediates the uptake of α -(1,6)-linked dietary oligosaccharides in Bifidobacterium and correlates with competitive growth on these substrates. J. Biol. Chem. 2016,291, 20220–20231. [CrossRef][PubMed] 47. Thongaram, T.; Hoeflinger, J.L.; Chow, J.; Miller, M.J. Prebiotic galactooligosaccharides metabolism by probiotic lactobacilli and bifidobacterial. J. Agric. Food Chem. 2017,65, 4184–4192. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). 21 foods Article Narrow-Leafed Lupin Main Allergen β-Conglutin (Lup an 1) Detection and Quantification Assessment in Natural and Processed Foods Elena Lima-Cabello 1, Juan D. Alché1and Jose C. Jimenez-Lopez 1,2,* 1Department of Biochemistry, Cell & Molecular Biology of Plants, Estacion Experimental del Zaidin, Spanish National Research Council (CSIC), Profesor Albareda 1, E-18008 Granada, Spain; [email protected] (E.L.-C.); [email protected] (J.D.A.) 2The UWA Institute of Agriculture and School of Agriculture and Environment, The University of Western Australia, Crawley, WA 6019, Australia *Correspondence: [email protected]; Tel.: +34-958-181-600 Received: 25 July 2019; Accepted: 4 October 2019; Published: 18 October 2019 Abstract: The increasing prevalence of lupin allergy as a consequence to the functional characteristics of a growing number of sweet lupin-derived foods consumption makes the imperious necessity to develop analytical tools for the detection of allergen proteins in foodstuffs. The current study developed a new highly specific, sensitive and accurate ELISA method to detect, identify and quantify the lupin main allergen β -conglutin (Lup an 1) protein in natural and processed food. The implementation of accurate standards made with recombinant conglutin β 1, and an anti-Lup an 1 antibody made from a synthetic peptide commonly shared among β -conglutin isoforms from sweet lupin species was able to detect up to 8.1250 ± 0.1701 ng (0.0406 ± 0.0009 ppm) of Lup an 1. This identified even lupin traces present in food samples which might elicit allergic reactions in sensitized consumers, such as β -conglutin proteins detection and quantification in processed (roasted, fermented, boiled, cooked, pickled, toasted, pasteurized) food, while avoiding cross-reactivity (false positive) with other legumes as peanut, chickpea, lentils, faba bean, and cereals. This study demonstrated that this new ELISA method constitutes a highly sensitive and reliable molecular tool able to detect, identify and quantify Lup an 1. This contributes to a more efficient management of allergens by the food industry, the regulatory agencies and clinicians, thus helping to keep the health safety of the consumers. Keywords: vicilin; 7S-globulins; food allergens; Lup an 1; sweet lupin species; food labelling; processed food 1. Introduction Lupine is a legume that belongs to the genus Lupinus and is included in the Leguminosae family, which is of great interest to the food industry, similarly for chickpeas, beans, peanuts, soya bean, lentils, and peas. Lupinus comprises between 200–600 different species [ 1 ]. However, only the four known as the sweet lupin group have gained interest since they are used in human food production [ 2 ] for their low levels of alkaloids [ 3 ] contained in their seeds. These four species include Lupinus albus (white lupine), Lupinus angustifolius (blue lupin or narrow-leafed lupin, NLL), Lupinus luteus (yellow lupine) [ 4 ], and Lupinus mutabilis (pearl or Andean lupin) that are mostly cultivated and consumed in central and South America [5]. Current interest for lupin seeds as a new functional food is growing [ 6 ], and the seeds from the sweet lupin species are becoming a crucial and alternative source of proteins for human consumption with nutritional and nutraceutical properties [ 7 – 10 ]. Sweet lupin species are a promising source of innovative ingredients for functional food, particularly those from the vicilin or β -conglutin family, Foods 2019,8, 513; doi:10.3390/foods8100513 www.mdpi.com/journal/foods 23 Foods 2019,8, 513 which are the most abundant proteins in NLL seeds [ 11 ]. Among the most frequently found, flour is one of the basic products and a common form to use lupin seeds as an ingredient in a wide range of food i.e., bread, cake, pasta, pizza, sausage, spices, cream cheese, tofu, jam [ 12 ]. In addition, lupin might often replace soya bean, egg white, and milk in vegan food [ 5 , 13 ], and is used as a functional ingredient in gluten-free food [12,14]. In the last five years and due to this increased nutraceutical knowledge, the number and range of commercially available lupin based products increased. In this regard, the rising lupin-derived products consumption, and the growing number of allergy reactions in sensitized persons have also increased. The routes of sensitization not only arise as primary sensitization to ingested lupin seed proteins, but also occur due to the cross-reactivity in atopic subjects sensitized to seed proteins from other legumes, particularly to soybean and peanut [ 15 – 17 ]. A third route appears to be the occupational allergy in people that works daily with lupin flour and lupin derived products [18]. Although the prevalence in the general population of lupin allergy is still unknown [ 19 ], it has been estimated to be in the range of 0.3% to 8%, and particularly in children being 5% [ 5 , 13 , 20 ]. It is important to identify the lupin seed proteins involved in allergy reactions [ 21 , 22 ], with storage proteins being the main lupin allergens [ 23 ], particularly from the vicilin family ( β -conglutins), which were named as Lup an 1 in NLL (WHO/IUIS Allergen Nomenclature Subcommittee). Due to the significant increase of reported cases of lupin allergy, and in order to keep safety among the population, the seeds from this legume together with soya bean and peanut were included in the European Union regulations (Regulation (EU) No. 1169/2011) as foods prone to induce allergy reactions. There is also a mandatory declaration in the labelling of the pre-packaged food ingredients, and overall, it provides information to consumers of these potential sources of food allergens. Thus, this demands very accurate and sensitive, highly specific quantitative methods to identify the main allergen protein content in food in order to identify the cross-contamination that justifies the precautionary labelling because of the presence of lupin as an ingredient. Thus, the aim of the present study was to develop a new highly specific ELISA method for the detection, identification and quantification of the lupin main allergen β -conglutin ( Lup an 1 ), and the assessment of this method in natural and processed food. Since lupin flour is being mainly used as an ingredient in bakery products, vegetarian and vegan based products, this analytical method was validated in natural (non-processed) food and processed (fermented, cooked, boiled, roasted) food such as flour, bread, and biscuits; as well as others lupin based alternative foods as spreading, sauce, cheese, pickles, drink, butter, meat. This study also includes data from these commercial foodstuffs differing in their labelling (“may contain lupin traces”, “lupine contained”, “lupine content not included on the food package”). 2. Material and Methods 2.1. Bioinformatics Analysis of β-Conglutin Protein Sequences The sequences of seven available isoforms of NLL β -conglutin protein were retrieved from the NCBI database (accession number F5B8V9, F5B8W0–F5B8W5 corresponding to β 1to β 7). A multiple sequence alignment and subsequent analysis was performed using ClustalW software (https://www. ebi.ac.uk/Tools/msa/clustalw2), based on Blosum62 matrix (BLOck SUbstitution Matrix) [ 24 ], and viewed using the Jalview viewer 2.2 (http://www.jalview.org). The Bioedit v 7.0.5.3 (http://www.mbio. ncsu.edu/BioEdit/bioedit.html) software was used to calculate the sequence identity matrices. 2.2. Analysis of the Antigenicity of β-Conglutin Proteins The prediction of antigenicity for these β -conglutin protein isoforms was performed. Antigenic epitopes were determined on the basis of the following parameters: hydrophobicity [ 25 ]; amino acid surfaceaccessibility [ 26 , 27 ]; antigenicitymethods suchas HoppandWoods hydropathy [ 28 ], Welling [ 29 ], Parker [ 30 ], BepiPred-2.0 (sequential β -cell epitope predictor, http://www.cbs.dtu.dk/services/BepiPred), and Kolaskar and Tongaonkar [ 31 ]. The results were confirmed using the antigenicity prediction of GenScript service (https://www.genscript.com/antigen-design.html). 24 Foods 2019,8, 513 2.3. β-Conglutin Proteins Structure Modelling The conglutin β 1 protein sequence (F5B8V9) was retrieved and used for searching the best structural templates in the Protein Data Bank (PDB) (http://www.rcsb.org). The suitable templates for this sequence were selected using BLAST server (http://ncbi.nlm.nih.gov). To improve the best final templates identification and selection, the software BioInfoBank Metaserver (http://meta.bioinfo.pl) specializing in fold recognition homology was used, as well as the Swiss-model server (swissmodel.expasy.org). The best four identified templates (1uij, 2phl, 3s7e, and 2eaa) were retrieved from the PDB database, and implemented for homology modelling. The conglutin β 1 protein model was built by the implementation of SWISS-MODEL via the ExPASy web server (swissmodel.expasy.org) using these top PDB closely related structural templates. The structural errors in the initial structural 3D models were identified by using ProSA (prosa.services.came.sbg.ac.at/prosa.php), obtaining the first overall quality estimation of each model using QMEAN4 (swissmodel.expasy.org/qmean/cgi/index.cgi). The final structure of β -conglutin proteins was subjected to energy minimization using GROMOS96, which was implemented in DeepView/Swiss-PDBViewerv3.7(spdbv.vital-it.ch)improvingthevan derWaals contactsand correcting the stereochemistry. The quality of the final models was assessed by assessing the protein stereology with PROCHECK (www.ebi.ac.uk/thornton-srv/software/PROCHECK) and ProSA programs, as well as the protein energy using ANOLEA (protein.bio.puc.cl/cardex/servers/anolea). The suitability of the model on the basis of the number of protein residues in the favoured regions was assessed by Ramachandran plot statistics. To define the potential functional and interacting areas/clusters in the protein, the electrostatic Poisson-Boltzmann (PB) potentials were calculated by using APBS (DeLano Scientific LLC) molecular modelling software implemented in PyMOL 0.99 (www.pymol.org). The electrostatic PB potential values are given in units of kT per unit charge (k Boltzmann’s constant; T temperature). 2.4. Construction of the Expression Plasmid The protein expression in bacteria was achieved using the pET28a(+) vector (Novagen) with some modifications, such as an N-terminal 6xHis Tag. Vector pUC57 was used for cloning a synthetic gene encoding β -conglutin protein (GenBank HQ670409, β 1), connected by restriction enzyme linker sequences, NcoI and XhoI (GenScript). The genetic construct included the expression vector pET28a(+)- conglutin β 1-6xHis-Tag, which was obtained through the digestion of the pUC57conglutin β 1 construct with NcoI and XhoI restriction enzymes, followed by ligation of the β 1 fragment into the pET28a(+) vector. 2.5. Overexpression of Conglutin β1 The final genetic construct containing the expression vector with the conglutin β 1 gene was transformed into Rosetta ™ 2(DE3) pLysS Singles ™ Competent Cells (Novagen) for β 1 expression. The protein expression was accomplished by using an auto-induction method [ 32 ]. Briefly, a colony of Escherichia coli containing the construct was isolated and grown for 24 h in ZY-medium plus kanamycin (50 μ g/mL) at 37 ◦ C in constant shaking (190 rpm). The culture was diluted 1:175 in Studier medium to grow for 7 h until the cell density reached 0.6 OD at 600 nm. The overexpression induction was achieved by adjusting the temperature to 20 ◦ C during 17 h. The bacterial cell pellet was collected by centrifugation at 5000 × gat 4 ◦ C. The bacterial pellet was washed three times with PBS, pH 7.5, and after removing the supernatant, the pellet cell was flash frozen using liquid nitrogen. The resulting pellet was stored at −80 ◦C until further use. 2.6. Purification of the Recombinant Conglutin β1 Protein Overall, β 1 purification was performed following the company’s recommendations (Qiagen) for Histidine tagged proteins. Briefly, the main steps comprised the cells breaking, followed by affinity chromatography using nickel-NTA spin columns, and linking a 6xHis-Tag at the C-terminal end 25 Foods 2019,8, 513 of the β 1 protein. The elution of this protein from the column was performed with an increasing imidazole concentration gradient (25–350 mM), collecting 2 mL fractions. All the fractions containing the protein were analyzed using SDS-PAGE, and those containing a single band with the expected molecular weight were pooled and dialyzed against PBS during 2 days with buffer changes every 12 h. The resulting samples were then aliquoted and flash frozen in liquid nitrogen to be kept at − 80 ◦ C until further uses. The purity of the protein samples was >95%. The typical yields were ~25–55 mg/mL. 2.7. Antibody Production Against β-Conglutin The performance and specificity of an antibody highly depends on the nature of the binding to its protein target. A polyclonal epitope-specific antibody was developed using a combination of a lineal epitope identification and characterization (antigenicity assessment) in the target protein sequence and an affinity capture approach involving synthesized peptides. The sequences currently available of seven NLL β -conglutin genes were retrieved from UNIPROT database(https://www.uniprot.org). Thealignment ofthesesequences wasanalyzedto finda commonly shared antigenic peptide among the seven NLL β -conglutin proteins variants following an analysis of antigenicity described in previous sections, finding the following sequence: Nt– VDEGEGNYELVGIR –Ct. A synthetic peptide with this sequence was produced (Agrisera, Vännäs, Sweden). The chosen sequence (up to 14 amino acids +terminal cysteine) was synthesized in a 15 mg immunograde purity scale without any further modifications. In order to elevate the antigenicity of the synthetic peptide, 2 mg of it was coupled to a carrier protein, KLH, which is the keyhole limpet hemocyanin copper-containing protein, with a molecular size >450 kDa. TheKLHcoupledsyntheticpeptidewasusedforanimal(rabbits)immunization. Thetypicalamount of antigen for a standard immunization protocol is ~500 μ g/animal, and a preferable concentration >1mg/mL dissolved in PBS. Five animals were subjected to a program of three immunization rounds (105 days long) with the same antigen. After 80 days, the test samples were delivered to our laboratory for evaluation, where sera, after this period of immunization, was tested to check for the capability to detect its target protein and also to show the potential background signal from these different sera in Western blot assays using recombinant and lupin seed protein extracts. The animals’ analyzed polyclonal sera having target-specific antibodies directed towards linear epitope and giving Western blot bands of correct molecular weight against β -conglutin proteins and low or no background was used for further immunization and obtaining the sera. After the third immunization, all serum samples were checked for the presence of antigen specific antibodies using ELISA. A further proof of high specificity and no background produced by the antibody was obtained by a screening of pre immune serum. Testing the pre-immune samples made sure that no background signals were detected in the molecular weight region of the protein under investigation. Furthermore, the antibody capture and clean up (affinity-purified) from antiserum was performed from the final rabbit immune serum by a liquid chromatography system against the same synthetic peptides as the affinity column tag. This step contributed to obtaining a higher specificity of the antibody, but also reduced the amount of the available antibody in the final serum. 2.8. Food and Biological Samples Used for Lupin Allergen Detection and Quantification The different commercially available foods containing variable quantity of lupin seed were purchased from different supermarkets. More detailed information of each product is described in Table 1. 26 Foods 2019,8, 513 Table 1. Summary of samples analyzed in this study. The information includes brand, the lupin content, and the food processing state. Product Number Product Product Information Related to Lupin Content Web 1 Toasted bread-Crostini Lupin protein https://www.schaer.com/en-int/p/crostini 2Gluten Free Maxi Sorrisi Chocolate Biscuit It may content traces of lupin https://www.schaer.com/en-int/p/maxi-sorrisi 3 Commercial lupin flour Sweet lupin flour https://www.arche-naturkueche.de/de/produkte/europaeische-kueche/backen-binden/s%C3%BCsslupinenmehl 4 Peanut butter Roasted peanuts (97%), palm oil, and sea salt. https://shop.wholeearthfoods.com/collections/award-winning-nut-butters/products/whole-earth-dark-roasted-peanut-butter-340-g 5Carob spread hazelnut Chocolate duo Lupin flour 5% https://greenest.ee/en/product/carobella-carobella-chocolate-duo-bio-350-g 6 Seeds (Lupinus albus L.) Seeds https://www.semillascantueso.com/ 7 Seeds (Lupinus luteus L.) Seeds https://www.semillascantueso.com/ 8 Seeds (Lupinus angustifolius L.) Seeds https://www.semillascantueso.com/ 9 Lupinen BOLOGNESE SAUCE Sweet lupin seeds cooked (8%) https://www.veggie-shop24.com/food/ready-meals/sauces-and-dips/alberts-lupine-bolognese-sauce-organic-300g 10 Fresh spread cheese Lupin protein (6.6%) https://www.veggie-shop24.com/food/cheese-alternatives/cream-cheese/made-with-luve-frisch-cremiges-streichglueck-fresh-creamyspread-bliss-herbs-150g 11 Wheat toasted bread Non lupin content http://www.bimbo.es/productos/tostados#asies 12 Lupinen—Tempeh LUPEH Boiled sweet lupin seeds 99% https://www.veggie-shop24.com/food/vegan-basics/tofu-and-tempeh/alberts-lupeh-lupine-tempeh-organic-170g 13 Lupinen BURGER—MEDITERRANEAN Boiled sweet lupin seeds 15% https://www.veggie-shop24.com/food/meat-alternatives/burger-and-grill/alberts-lupine-burger-gluten-free-organic-200g 14 Pickled lupine Lupin (L. albus)https://es.openfoodfacts.org/producto/8480000330987/altramuces-encurtidos-hacendado 15 Lupinen Drink Lupin protein (2.3%) https://www.alles-vegetarisch.de/lebensmittel/milchalternativen-und-desserts/milchersatz/made-with-luve-lupinen-drink-natur-1l 16 TOFU smoked Made with soya bean. Non lupin content. https://www.veggie-shop24.com/food/vegan-basics/tofu-and-tempeh/alberts-tofu-smoked-organic-1kg 17 Boiled lentils Non lupin content 18 Boiled chickpea Non lupin content https://www.deliberry.com/mercadonamadrid/alimentacion-general/legumbres-y-verduras-envasadas 19 Boiled faba bean Non lupin content https://www.deliberry.com/mercadonamadrid/alimentacion-general/legumbres-y-verduras-envasadas 27 Foods 2019,8, 513 2.9. Proteins Extraction The total protein (including allergen proteins) extractions were performed from the different food and biological samples. The samples were homogeneously grinded with a polytron homogenizer (Kinematica Polytron ™ PT 2500E, VWR), always keeping the samples in a water-ice cool bath. For these proteins, extractions were used as the extraction buffer containing 100 mM Tris, pH 7.4, 250 mM NaCl, 1 mM EGTA, 1 mM EDTA, 1% Triton X-100, 0.5% Sodium deoxycholate, Protease inhibitor cocktail and 1 mM PMSF. The homogenates were centrifuged at 10,000 × gto remove all the gross material and to keep the supernatant. The high NaCl concentrations contained in the extraction buffer assures specific vicilin proteins extraction [33]asβ-conglutin (Lup an 1). For the protein extraction from tissues, the samples were dissected and kept at a low temperature on ice-water to prevent proteolysis. Then, 50 mL of complete extraction buffer was added to1gof the sample in a tube to be homogenized with the polytron homogenizer. Afterward, the blades of the homogenizer were rinsed with the extraction buffer, and the homogenized sample was maintained at a constant agitation for 2 h at 4 ◦ C for the proteins’ final extraction to the media. A centrifugation for 20 min at 15,000 × gand 4 ◦ C was performed, and the supernatant was placed on ice-water to make the aliquots (containing the soluble protein extract) in a fresh, chilled tube and to store the samples at −80 ◦C. The lupin drink was analyzed as it was in the bottle. 2.10. Gel Electrophoresis and Western Blots The analyses of the protein extracts were achieved by mixing each sample individually with 6 × protein sample buffer and heated to 95 ◦ C for 5 min. The proteins were separated on commercial 4–20% gradient TGX gels (Bio-Rad, Hercules CA, USA). The molecular weight markers used for stained gels were Precision Plus Protein ™ Dual Colour Standards (BioRad). The separated protein bands were visualized in a Gel Doc ™ EZ Imager (BioRad). The proteins were electrophoretically transferred from gel to PVDF membranes. Prior to the transference, the membranes were blocked for 2 h at room temperature (RT) with 5% non-fat dry milk in PBST (phosphate-buffered saline, 0.05% Tween-20) followed by the incubation with the first antibody, goat IgG anti-beta conglutin (dilution 1:2000), overnight at 4 ◦ C in continuous agitation. After washing 5 times with PBST, the membrane was incubated with secondary antibody goat anti-IgG rabbit conjugated with horseradish peroxidase (dilution 1:10,000) in 2% non-fat dry milk in PBST for 2 h at RT. The membrane was washed 5 times with PBST and the chemiluminescence signal was developed by membrane incubation with ECL Plus chemiluminescence substrate following the manufacturer’s instructions (BioRad). The light signal was detected by exposure of the membrane to C-Digit Blot Scanner (LI-COR). 2.11. ELISA Test for the Detection and Quantification of β-Conglutin Allergen Proteins in Lupin-Containing Food The protein standards used for Lup an 1 identification and quantification was performed using purified allergen β -conglutin proteins. Coating the wells was performed by using purified conglutin β 1 (50 μ g). The plate wells containing purified proteins were used as blanks (controls without β -conglutin protein) and were simultaneously incubated (triplicate samples) overnight at 4 ◦ C. The wells on the plates were washed 5 times with 200 μ L of PBS for each well. In order to avoid the unspecific protein-binding sites, the samples were blocked in the coated wells by adding 200 μ L blocking buffer (5% non-fat dry milk/PBS) per well, and incubated for2hatRT.Thewells of the plates were then washed five times with 200 μ L PBS. The first anti-IgG β -conglutin antibody (dilution 1:1000) was incubated in each well for 2 h at RT. The solution containing the antibody was removed and each well was washed five times with 200 μ L PBS. The solutions on each well containing washing buffer were eliminated by flicking the plate over. The remaining drops on each well were eliminated by patting the plate with a paper towel. The wells on the plates were then incubated with a goat anti-IgG rabbit HRP conjugated antibody. The incubations were developed for 1 h and 30 min at RT, and the washes were performed five times with 200 μ L PBS on each well. The development of the signal was made using a compatible substrate, 100 μ L of TMB which was added to each well and incubated for 5 min at RT. 28 Foods 2019,8, 513 (1) The protein extraction protocol in the current study is highly specific for the extraction of the vicilin family of proteins ( β -conglutins), based on the presence of NaCl (0.25M) in the extraction buffer [ 33 ]. Previous studies made protein extractions using general protein extraction buffers [ 4 , 45 , 47 ], or using alternative methods from commercial kits [ 48 ] currently no longer available (Abnova, http://www.abnova.com/products/products_detail.asp?catalog_id=KA3310), displaying very limited or no information about: (i) The antibody design; (ii) antibody production and use in the ELISA detection method; (iii) the limited information about protocol for total proteins extraction, which may not be specific for β -conglutin extraction; (iv) no information about lupin species used [ 46 , 48 ] to obtain this protein extract. The last two are the main factors with high impact in the protein extract characteristics, such as a low amount or not of Lup an 1 content. These disadvantages may result in the increase of the number of false positives as a result of the detection of non-allergen proteins from a low specific antibody, or using non-appropriate protein extracts. (2) A second advantage of the current method compared with previous ones and commercial kits is the design and the production of the antibody (anti-IgG β -conglutin proteins). In the current study, the experimental animal was immunized with a synthetic peptide commonly shared by the seven NLL β -conglutin protein isoforms. This synthetic peptide constitutes a highly antigenic epitope in these proteins probed in the current study, while also exhibits a high specificity to detect the lupin main allergen Lup an 1 in the most frequently used lupin species. On the contrary, previous methods [ 4 , 45 – 47 ] have used the whole crude protein extract from lupin flour to immunize the experimental animal and obtain the antibody. The method to produce this antibody makes this antibody non-specific, and detecting a wide range of proteins, including many non-allergenic proteins, may lead to false positive detections. The detection of false positives might be also enhanced due to the implementation of a buffer inappropriate for vicilin protein extraction from lupin-derived foodstuffs. (3) The current method exhibited another advantage which was the type of standards (Figure S3) that was made for a specific quantification of Lup an 1. Previous methodological developments of standards for the quantification were based on lupin flour total protein extract [ 4 , 45 – 47 ]. This may induce variable immunization for a complex mix of proteins (allergenic and non-allergenic proteins) from the crude extract leading to an excess or lack of reactivity in the standard samples because a variable representation or content of Lup an 1 (over or under representation, depending on the extraction method). Therefore, lupin allergy reactions could be developed with high severity from primary sensitization to lupin proteins or due to the cross-reactivity with proteins from other legumes. The abundance of these proteins in many (natural or processed) foodstuffs has led to the European Union (EU Regulation No. 1169/2011) [ 49 ] to include lupin in a list of allergens with mandatory identification in all lupin containing foodstuffs. Indeed, an EU Labelling Directive involves the mandatory declaration of manufacturers regarding the presence of 14 allergenic products on pre-packaged foods [ 50 ]. Since lupin is included among these, analytical tools should be developed for the detection of lupin allergen traces in complex and processed foodstuffs. Thus, a high accurate method to detect and identify the lupin main allergen in food would be very valuable, with a particular importance due to the increasing prevalence of lupin allergy [ 16 ] among the population. The current study used highly pure (>95%) recombinant purified Lup an 1 (Figure 2), with the combination of an antibody that was developed using a specific synthetic peptide (main antigenic epitopes of the Lup an 1), while making the identification and quantification of the Lup an 1 highly specific and accurate. In order to evaluatethe applicationof the developedmethod toactual foodstuffs, several commercial samples labelled as “it may contain traces”, “with or without lupine” were tested for the presence of Lup an 1. The summarized ELISA results, together with the corresponding label information of samples, are described in Tables 1and 2. 35 Foods 2019,8, 513 The Lup an 1 lower content value in the food samples analyzed was 0.0406 ± 0.0009 ppm for toasted bread (Table 2). This result is in the comparable range of an ELISA method developed to detect soy protein content in foodstuff [ 51 ]. In this regard, the lowest eliciting dose for allergic reactions to lupin, responsible for inducing mild symptoms in peanut-sensitized patients, was 0.5 mg of lupin flour [ 17 ]. More recently, the VITAL program of the Allergen Bureau of Australia and New Zealand (ABA) established 4 mg of protein as the reference allergenic dose for lupin [ 52 ]. Taking into consideration that conglutins are the most abundant protein in lupin, being 40% of the protein seeds content [ 7 ], our ELISA method is by far capable of detecting these allergenic reference doses of lupin proteins in foodstuffs as demonstrated in this study. Overall, NLL exhibited the highest Lup an 1 value (574.7918 ± 19.8876 ng) compared to all the samples investigated. The analysis of the sample number 2 (biscuits), and these samples declaring the information “may contain traces of lupine”, showed no Lup an 1 content (Table 2), which might be due to non-contamination with lupin flour or other lupin derived component. Despite this fact, the manufacturer has implemented the common practice of the precautionary labelling. The same result was obtained from samples as number 11 (Tables 1and 2) labelled as “non lupin content”. The estimated content of Lup an 1 in commercial lupin flour displayed an intermediate value among the three lupin seed species analyzed (Tables 1and 2), which is in agreement with the labelling composition as sweet lupin attributed to three main domesticated lupin species (L. albus,L. luteus and L. angustifolius). The samples number 5, 9, 10, and 12 to 16 containing “lupine flour” or “lupine protein” among their ingredients, Lup an 1 was detected, identified and quantified in a range of allergen in accordance with their labelling (Tables 1and 2) and also, to the variable quantity of seed flour from these three different sweet lupin species. 3.4. Specificity of the Antiβ -Conglutin Antibody Tested in Processed Foodstuffs and Potential Cross-Reactive Lupin allergen proteins have been identified to be stable towards thermal treatment [ 5 , 47 ]in studies concerning the impact of the type of food processing on the allergenicity. Indeed, comparable results have been obtained in the current study for soy and chickpea-containing products, as well as for food processed under fermentation, soaking, extrusion, cooking, boiling and microwave heating conditions. These cooking processes do not affect the allergenic potential of protein extracts, while the autoclaving process drastically reduces the binding capacity of antibodies tested with serum from atopic patients [ 53 – 55 ]. This may be due to the conformational epitopes disappearing in these proteins under these processed conditions. The main factors with a significant influence in the allergen analysis with ELISA methods are food matrices, the level of food processing and the ELISA test kit selection [ 56 ]. The current study analyzed the different foodstuffs (Tables 1and 2), which were differentially processed. The ELISA method was able to detect, specifically identify and quantify Lup an 1 allergen: toasted [bread ( 8.1250 ±0.1701 ng )], fermented [Lupinen—Tempeh (152.9862 ± 7.3353 ng)], Pickle [Pickled lupin (293.5833 ± 11.0853 ng)], Pasteurized [Lupinen drink (59.8660 ± 0.8389 ng)], and cooked [BOLOGNESE SAUCE ( 3.8890 ±0.2599 ng) ] foodstuff. Furthermore, a previous study indicated that boiling affected the allergenicity of legumes, particularly soy (Alvarez-Alvarez et al., 2005). Despite this fact, this study was able to detect and quantify Lup an 1 in boiled lupin based foodstuffs like Lupinen-tempeh (152.9862 ±7.3353 ng) and LupinenBurger-Mediterranean (129.1667 ±3.6839 ng). Allergy reactions to lupin seed proteins are often triggered in atopic patients sensitized to proteins from other legumes, such as pea, lentil, soya, chickpea and peanut [ 5 , 21 , 57 ]. These allergy reactions arise against seed storage proteins from the Leguminosae family included in foodstuffs, since these proteins share similar epitopic regions in these different legume seed proteins. In this regard, atopic patients sensitized to one legume allergen protein might develop cross-reactions to proteins from another legume [ 58 , 59 ]. The most frequently described and important cross-reactivity from a clinical point of view is between lupin and peanuts [ 5 ]. Additional cross-reactivity between lupin and other legumes (lentil, pea) has been also reported, but in low percentages in children [ 5 , 57 ]. Furthermore, at 36 Foods 2019,8, 513 a molecular level, lupin IgE cross-reactivity has been reported for peanut, soya, lentil, chickpea and bean [ 5 ]: Lupin sensitized patients were reported to develop cross-reactivity between 59–72% [ 13 , 20 , 60 ] and 52–55% [ 20 , 60 ] for soya bean and pea, respectively. Allergy reactions to peanuts are currently the most frequently identified legume allergy, followed by soya bean [57]. In this regard, a high specific and reliable method to detect the presence of other legume proteins is of crucial importance in order to avoid the development of cross-reactivity allergy reactions. The method developed in the current study was able to detect the lupin main allergen Lup an 1 with high specificity compared to other previous methods. However, the method may show false negative results as a consequence of the low antibody specificity and/or the absence of detection of allergen proteins that were not extracted in enough quantity by these alternative previously developed methods. On the other hand, false positive results as consequence of the use of a non-high specific antibody developed against a whole protein extract may not have clinical relevance. This may further lead to cross-reactivity between the antibody and the target protein from a related species with similar detection results (lack of specificity), and in the end, this scenario has notable consequences for the quality of life of people. In the current study, lupin cross-reactivity with other legumes could be clinically relevant. For that reason, the authors developed an accurate, reliable, and highly specific method to detect, identify and quantify the main lupin allergen Lup an 1 that may be responsible of cross-reactivity with other legume proteins [ 11 , 21 , 22 , 36 ]. The high specificity of our antiβ -conglutin antibody is capable of avoiding false positives as a result of the cross-reactivity (Table 2). An analysis of different food products containing multiple legume proteins, such as peanut butter, chickpea, lentil, faba bean, even cereals resulted in the negative detection of homologous proteins to Lup an 1. In comparison, other ELISA methods previously developed showed low specificity of their antibodies used in the ELISA methods, since cross-reactivity showed with other legumes: Either with pea, chickpea, peanut, lentil, and soy [ 45 ], with brown bean and fenugreek (Holden et al., 2007); black bean and soy [ 46 ]; or even with non-legume proteins such as sunflower seed, cashew, almond, and pumpkin seed [ 61 ]. Furthermore, Koeberl et al. [48] used three available commercial kits to analyze cross-reactivity between lupin and other legume proteins finding that all peanut samples tested showed cross-reactivity on ELISA test kit B and C [ 48 ]. In addition, cross-reactivity was also described for the entire lentil samples analyzed, thus promoting false positive results in all analyzed legume samples. These results highlight the importance of transparency in the information provided by developed kits, at least in the characteristics of ELISA antibodies and protocols for their production, and further for the proteins extraction protocols implemented in these kits. This could lead to more uncertainty about the link between the food allergen protein identification and quantification, with the development of clinical therapies. 4. Conclusions A newly developed ELISA assay was assessed for its capability to detect, identify and quantify the lupin main allergen β -conglutin protein (Lup an 1) in natural products and processed foodstuffs (toasted, boiled, fermented, cooked, pickled, pasteurized lupin-based products). Cross-reactivity was tested using peanut and other legumes, obtaining negative results. The standards were made using recombinant purified β -conglutin proteins, showing a highly specificmethodfor particularallergenproteinscomparedtopreviousdevelopments. Thisfacthighlights the importance of particular molecular tools to develop a more reliable and highly specific analytical method for food allergen protein detection. This recombinant protein could constitute an available improvement for the quantification of food allergens, which should be approved as (certified) reference material for food allergens identification helping to implement the labelling mandatory EU regulations and as relevant proteins for clinical allergy research. Finally, this study demonstratedthat thisELISA method isbased onaccurate andreliable molecular tools, which can contribute to a more effective management of allergens by the food industry, the regulatory agencies and clinicians, thus helping to protect the health of sensitized/allergic consumers. 37 Foods 2019,8, 513 Supplementary Materials: The following are available online at http://www.mdpi.com/2304-8158/8/10/513/s1, Figure S1: Multiple alignment of the deduced amino acid sequences of β 1to β 7., Figure S2: Antigenicity assessment of conglutin β1., Figure S3: Standard curve made of the conglutin β1 protein. Author Contributions: Conceptualization: J.C.J.-L. and E.L.-C.; methodology: J.C.J.-L. and E.L.-C.; software: J.C.J.-L.; validation: J.C.J.-L. and E.L.-C.; data Analysis: J.C.J.-L. and E.L.-C. and J.D.A.; investigation: J.C.J.-L. and E.L.-C.; resources: J.C.J.-L. and J.D.A.; writing—original draft preparation: J.C.J.-L.; writing—review & editing: J.C.J.-L. and J.D.A.; funding acquisition: J.C.J.-L. and J.D.A. Funding: This research was funded by the European Research Program MARIE CURIE (FP7-PEOPLE-2011-IOF) grant number PIOF-GA-2011-30155; the Spanish Ministry of Economy, Industry and Competitiveness grants numbers RYC-2014-16536 (Ramon y Cajal Research Program) and BFU2016-77243-P; the CSIC-Intramural grant number 201540E065; And The APC was funded by the Spanish Ministry of Economy, Industry and Competitiveness grants numbers RYC-2014-16536 (Ramon y Cajal Research Program). Acknowledgments: J.C.J.-L. thanks the European Research Program MARIE CURIE (FP7-PEOPLE-2011-IOF) for through the grant ref. number PIOF-GA-2011-30155; to the Spanish Ministry of Economy, Industry and Competitiveness for the grant ref. number RYC-2014-16536 (Ramon y Cajal Research Program), and the grant ref. BFU2016-77243-P; and to CSIC-Intramural grant ref. 201540E065. Conflicts of Interest: The authors declare no conflicts of interest. References 1. Kurlovich, B.S.; Stankevich, A.K.; Stepanova, S.I. The review of the genus Lupinus L. In Lupins (Geography, Classification, Genetic Resources and Breeding); Kurlovich, B.S., Ed.; OY International North Express: St. Petersburg, Russia; Pellosniemi, Finland, 2002; Chapter 2; pp. 11–38. 2. Uauy, R.; Gattas, V.; Yáñez, E. Sweet lupins in human nutrition. World Rev. Nutr. Diet. 1995,77, 75–88. 3. Frick, K.M.; Kamphuis, L.G.; Siddique, K.H.M.; Singh, K.B.; Foley, R.C. Quinolizidine alkaloid biosynthesis in lupins and prospects for grain quality improvement. Front. Plant Sci. 2017,8, 87. [CrossRef] 4. Ecker, C.; Cichna-Markl, M. Development and validation of a sandwich ELISA for the determination of potentially allergenic lupine in food. Food. Chem. 2012,130, 759–766. [CrossRef] 5. Jappe, U.; Vieths, S. Lupine, a source of new as well as hidden food allergens. Mol. Nutr. Food Res. 2010 ,54, 113–126. [CrossRef] 6. Delgado-Andrade, C.; Olias, R.; Jimenez-Lopez, J.C.; Clemente, A. Nutritional and beneficial effects of grain legumes on human health. ARBOR Ciencia Pensamiento y Cultura 2016,192, 779. 7. Lima-Cabello, E.; Alche, V.; Foley, R.C.; Andrikopoulos, S.; Morahan, G.; Singh, K.B.; Alche, J.D.; Jimenez-Lopez, J.C. Narrow-leafed lupin (Lupinus angustifolius L.) β -conglutin proteins modulate the insulin signaling pathway as potential type 2 diabetes treatment and inflammatory-related disease amelioration. Mol. Nutr. Food Res. 2017,61.[CrossRef] 8. Lima-Cabello, E.; Morales-Santana, S.; Foley, R.C.; Melser, S.; Alche, V.; Siddique, K.H.M.; Singh, K.B.; Alche, J.D.; Jimenez-Lopez, J.C. Ex vivo and in vitro assessment of anti-inflammatory activity of seed β-conglutin proteins from Lupinus angustifolius.J. Funct. Foods 2018,40, 510–519. [CrossRef] 9. Lima-Cabello, E.; Morales-Santana, S.; Leon, J.; Alche, V.; Clemente, A.; Alche, J.D.; Jimenez-Lopez, J.C. Narrow-leafed lupin (Lupinus angustifolius L.) seed beta-conglutins reverse back the induced insulin resistance in pancreatic cells. Food Funct. 2018,9, 5176–5188. [CrossRef] 10. Lima-Cabello, E.; Robles-Bolivar, P.; Alche, J.D.; Jimenez-Lopez, J.C. Narrow-leafed lupin Beta-conglutin proteins epitopes identification and molecular features analysis involved in cross-allergenicity to peanut and other legumes. Genom. Comput. Biol. 2016,2, e29. [CrossRef] 11. Foley, R.C.; Jimenez-Lopez, J.C.; Kamphuis, L.G.; Hane, J.K.; Melser, S.; Singh, K.B. Analysis of conglutin seed storage proteins across lupin species using transcriptomic, protein and comparative genomic approaches. BMC Plant Biol. 2015,15, 106. [CrossRef] 12. Cabello-Hurtado, F.; Keller, J.; Ley, J.; Sanchez-Lucas, R.; Jorr í n-Novo, J.V.; Aïnouche, A. Proteomics for exploiting diversity of lupin seed storage proteins and their use as nutraceuticals for health and welfare. J. Proteom. 2016,143, 57–68. [CrossRef] 13. De Jong, N.W.; Van Maaren, M.S.; Vlieg-Boersta, B.J.; Dubois, A.E.J.; De Groot, H.; van Wijk, R.G. Sensitization to lupine flour: Is it clinically relevant? Clin. Exp. Allery 2010,40, 1571–1577. [CrossRef] 14. Ronchi, A.; Duranti, M.; Scarafoni, A. A real-time PCR method for the detection and quantification of lupin flour in wheat flour-based matrices. Food Chem. 2009,115, 1088–1093. 38 Foods 2019,8, 513 15. Sirtori, E.; Resta, D.; Brambilla, F.; Zacherl, C.; Arnoldi, A. The effects ofvarious processing conditions on a protein isolate from Lupinus angustifolius.Food Chem. 2010,120, 496–504. [CrossRef] 16. Sanz, M.L.; De Las Marinas, M.D.; Fern á ndez, J.; Gamboa, P.M. Lupin allergy: A hidden killer in the home. Clin. Exp. Allery 2010,40, 1461–1466. [CrossRef] 17. Peeters, K.A.; Koppelman, S.J.; Penninks, A.H.; Lebens, A.; Bruijnzeel-Koomen, C.A.; Hefle, S.L.; Taylor, S.L.; van Hoffen, E.; Knulst, A.C. Clinical relevance of sensitization to lupine in peanut-sensitized adults. Allergy 2009,64, 549–555. [CrossRef] 18. Campbell, C.P.; Yates, D.H. Lupin allergy: A hidden killer at home, a menace at work; occupational disease due to lupin allergy. Clin. Exp. Allery 2010,40, 1467–1472. [CrossRef] 19. Kaw, C.H.; Hefle, S.L.; Taylor, S.L. Sandwich Enzyme-Linked Immunosorbent Assay (ELISA) for the Detection of Lupine Residues in Foods. Food Sci. 2008,73, T135–T140. [CrossRef] 20. Hieta, N.; Hasan, T.; Makinen-Kiljunen, S.; Lammintausta, K. Lupin allergy and lupin sensitization among patients with suspected food allergy. Ann. Allergy Asthma Immunol. 2009,103, 233–237. [CrossRef] 21. Jimenez-Lopez, J.C.; Foley, R.C.; Brear, E.; Clarke, V.C.; Lima-Cabello, E.; Florido, J.F.; Singh, K.B.; Alch é , J.D.; Smith, P.M.C. Characterization of narrow-leaf lupin (Lupinus angustifolius L.) recombinant major allergen IgE-binding proteins and the natural β -conglutin counterparts in sweet lupin seed species. Food Chem. 2018 , 44, 60–70. [CrossRef] 22. Lima-Cabello, E.; Berral-Hens, J.M.; Florido, J.F.; Alche, J.D.; Jimenez-Lopez, J.C. The function of conglutin proteins from the “sweet lupin” group in food allergy. In Legumes for Global Food Security, 1st ed.; Nova Science Publishers, Inc.: New York, NY, USA, 2017; Chapter 10; pp. 249–275. 23. Goggin, D.E.; Mir, G.; Smith, W.B.; Stuckey, M.; Smith, P.M.C. Proteomic analysis of lupin seed proteins to identify conglutin βas an allergen, Lup an 1. J. Agric. Food Chem. 2008,56, 6370–6377. [CrossRef] 24. Jimenez-Lopez, J.C.; Gachomo, E.W.; Sharma, S.; Kotchoni, S.O. Genetic sequencing and next-generation sequence data analysis: A comprehensive compilation of bioinformatics tools and databases. Am. J. Mol. Biol. 2013,3, 115–130. [CrossRef] 25. Kyte, J.; Doolittle, R.F. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 1982,157, 105–132. [CrossRef] 26. Emini, E.A.; Hughes, J.V.; Perlow, D.S.; Boger, J. Induction of hepatitis A virus-neutralizing antibody by a virus-specific synthetic peptide. J. Virol. 1985,55, 836–839. 27. Karplus, P.A.; Schulz, G.E. Prediction of chain flexibility in proteins. Naturwissenschaften 1985 ,72, 212–213. [CrossRef] 28. Hopp, T.P.; Woods, K.R. Prediction of protein antigenicdeterminants from amino acid sequences. Proc. Natl. Acad. Sci. USA 1981,78, 3824–3828. [CrossRef] 29. Welling, G.W.; Weijer, W.J.; van der Zee, R.; Welling-Wester, S. Prediction of sequential antigenic regions in proteins. FEBS Lett. 1985,188, 215–218. [CrossRef] 30. Parker, K.C.; Bednarek, M.A.; Coligan, J.E. Scheme for rankingpotential HLA-A2 binding peptides based on independent binding ofindividual peptide side-chains. J. Immunol. 1994,152, 163–175. 31. Kolaskar, A.S.; Tongaonkar, P.C. A semi-empirical method forprediction of antigenic determinants on protein antigens. FEBS Lett. 1990,276, 172–174. [CrossRef] 32. Fox, B.G.; Blommel, P.G. Autoinduction of Protein Expression. Curr. Protoc. Protein Sci. 2009,56, 5–23. 33. Nadal, P.; Canela, N.; Katakis, I.; O’Sullivan, C.K. Extraction, Isolation, and Characterization of Globulin Proteins from Lupinus albus.J. Agric. Food Chem. 2011,59, 2752–2758. [CrossRef] 34. Wang, Y.; Wu, W.; Negre, N.N.; White, K.P.; Li, C.; Shah, P.K. Determinants of antigenicity and specificity in immune response for protein sequences. BMC Bioinform. 2011,12, 251. [CrossRef] 35. Palladino, C.; Breiteneder, H. Peanut allergens. Mol. Immunol. 2018,100, 58–70. [CrossRef] 36. Jimenez-Lopez, J.C.; Lima-Cabello, E.; Melser, S.; Foley, R.C.; Singh, K.B.; Alch é , J.D. Lupin allergy: Uncovering structural features and epitopes of β -conglutin proteins in Lupinus angustifolius L. with a focus on cross-allergenic reactivity to peanut and other legumes. Bioinform. Biomed. Eng. 2015,9043, 96–107. 37. Jimenez-Lopez, J.C.; Robles-Bolivar, P.; Lopez-Valverde, F.J.; Lima-Cabello, E.; Kotchoni, S.O.; Alch é , J.D. Ole e 13 is the unique food allergen in olive: Structure-functional, substrates docking, and molecular allergenicity comparative analysis. J. Mol. Graph. Model. 2016,66, 26–40. [CrossRef] 39 Foods 2019,8, 513 38. Jimenez-Lopez, J.C.; Kotchoni, S.O.; Hernandez-Soriano, M.C.; Gachomo, E.W.; Alch é , J.D. Structural functionality, catalytic mechanism modeling and molecular allergenicity of phenylcoumaran benzylic ether reductase, an olive pollen (Ole e 12) allergen. J. Comput. Aided Mol. Des. 2013,27, 873–895. [CrossRef] 39. Jimenez-Lopez, J.C.; Kotchoni, S.O.; Rodr í guez-Garc í a, M.I.; Alch é , J.D. Structure and functional features of olive pollen pectin methylesterase using homology modeling and molecular docking methods. J. Mol. Model. 2012,18, 4965–4984. [CrossRef] 40. Jimenez-Lopez, J.C.; Rodr í guez-Garc í a, M.I.; Alch é , J.D. Analysis of the effects of polymorphism on pollen profilin structural functionality and the generation of conformational, Tand B-cell epitopes. PLoS ONE 2013,17, e76066. [CrossRef] 41. Jimenez-Lopez, J.C.; Gachomo, E.W.; Ariyo, O.A.; Baba-Moussa, L.; Kotchoni, S.O. Specific conformational epitope features of pathogenesis-related proteins mediating cross-reactivity between pollen and food allergens. Mol. Biol. Rep. 2012,39, 123–130. [CrossRef] 42. Jimenez-Lopez, J.C.; Melser, S.; DeBoer, K.; Thatcher, L.F.; Kamphuis, L.G.; Foley, R.C.; Singh, K.B. Narrow-Leafed Lupin (Lupinus angustifolius) β 1and β 6-Conglutin Proteins Exhibit Antifungal Activity, Protecting Plants against Necrotrophic Pathogen Induced Damage from Sclerotinia sclerotiorum and Phytophthora nicotianae.Front. Plant Sci. 2016,9, 1856. [CrossRef] 43. Magni, C.; Scarafoni, A.; Herndl, A.; Sessa, F.; Prinsi, B.; Espen, L.; Duranti, M. Combined 2D electrophoretic approaches for the study of white lupin mature seed storage proteome. Phytochemistry 2007 ,68, 997–1007. [CrossRef][PubMed] 44. Duranti, M.; Sessa, F.; Carpen, A. Identification, purification and properties of the precursor of conglutin β , the 7S storage globulin of Lupinus albus L. seeds. J. Exp. Bot. 1992,43, 1373–1378. [CrossRef] 45. Holden, L.; Faeste, C.K.; Egaas, E. Quantitative sandwich ELISA for the determination of lupine (Lupinus spp.) in foods. J. Agric. Food Chem. 2005,53, 5866–5871. [CrossRef] 46. Taylor, S.L.; Nordlee, J.A.; Niemann, L.M.; Lambrecht, D.M. Allergen immunoassays–considerations for use of naturally incurred standards. Anal. Bioanal. Chem. 2009,395, 83–92. 47. Ecker, C.; Ertla, A.; Pulverera, W.; PalSzekely, A.N.; Petraschc, A.; Linsberger-Martin, G.; Cichna-Markl, M. Validation and comparison of a sandwich ELISA, two competitive ELISAs and a real-time PCR method for the detection of lupine in food. Food Chem. 2013,141, 407–418. [CrossRef] 48. Koeberl, M.; Sharp, M.F.; Tian, R.; Buddhadasa, S.; Clarke, D.; Roberts, J. Lupine allergen detecting capability and cross-reactivity of related legumes by ELISA. Food Chem. 2018,256, 105–112. [CrossRef] 49. European Parliament Regulation (EU). No 1169/2011 of the European Parliament and of the Council of 25 October 2011 on the Provision of Food Information to Consumers, Amending Regulations (EC) No 1924/2006 and (EC) No 1925/2006 of the European Parliament and of the Council, and Repealing Commission Directive 87/250/EEC, Council Directive 90/496/EEC, Commission Directive 1999/10/EC, Directive 2000/13/EC of the European Parliament and of the Council, Commission Directives 2002/67/EC and 2008/5/EC and Commission Regulation (EC) No 608/2004 Text with EEA Relevance L304/18 C.F.R.; EU: Brussel, Belgium, 2011. 50. EFSA. Directive 2007/68/EC Amending Annex IIIa of Directive 2000/13/EC of the European Parliament and of the Council (2007). Off. J. Eur. Union 2007,310, 11–14. 51. Pedersen, M.H.; Holzhauser, T.; Bisson, C.; Conti, A.; Jensen, L.B.; Skov, P.S.; Bindslev-Jensen, C.; Brinch, D.S.; Poulsen, L.K. Soybean allergen detection methods—A comparison study. Mol. Nutr. Food Res. 2008 ,52, 1486–1496. [CrossRef] 52. Taylor, S.L.; Baumert, J.L.; Houben, G.F.; Crevel, R.W.R.; Brooke-Taylor, S.; Allen, K.J.; The Allergen Bureau of Australia & New Zealand; Houben, G. Establishment of reference doses for residues of allergenic foods: Report of the VITAL expert panel. Food Chem. Toxicol. 2014,63, 9–17. 53. Alvarez-Alvarez, J.; Guillamon, E.; Crespo, J.F.; Cuadrado, C.; Burbano, C.; Rodriguez, J.; Muzquiz, M. Effects of extrusion, boiling, autoclaving, and microwave heating on lupine allergenicity. J. Agric. Food Chem. 2005 , 53, 1294–1298. [CrossRef] 54. Gupta, R.K.; Gupta, K.; Sharma, A.; Das, M.; Ansari, I.A.; Dwivedi, P.D. Health Risks and Benefits of Chickpea (Cicerarietinum) Consumption. J. Agric. Food Chem. 2017,65, 6–22. [CrossRef] 55. Verma, A.K.; Kumar, S.; Das, M.; Dwivedi, P.D. A Comprehensive Review of Legume Allergy. Clin. Rev. Allergy Immunol. 2013,45, 30–46. [CrossRef][PubMed] 40 Foods 2019,8, 513 56. Török, K.; Hajas, L.; Horv á th, V.; Schall, E.; Bugyi, Z.; Kem é ny, S.; Tömösközi, S. Identification of the factors affecting the analytical results offood allergen ELISA methods. Eur. Food Res. Technol. 2015 ,241, 127–136. [CrossRef] 57. Peeters, K.A.; Nordlee, J.A.; Penninks, A.H.; Chen, L.; Goodman, R.E.; Bruijnzeel-Koomen, C.A.; Knulst, A.C. Lupine allergy: Not simply cross-reactivity with peanut or soy. J. Allery Clin. Immunol. 2007 ,120, 647–653. [CrossRef][PubMed] 58. Berkner,H.; Neudecker,P.; Mittag, D.; Ballmer-Weber, B.K.; Schweimer, K.; Vieths,S.; Rosch, P.Cross-reactivity of pollen and food allergens: Soybean Gly m 4 is a member of the Bet v 1 superfamily and closely resembles yellow lupine proteins. Biosci. Rep. 2009,29, 183–192. [CrossRef] 59. Shaw, J.; Roberts, G.; Grimshaw, K.; White, S.; Hourihane, J. Lupin allergy in peanut-allergic children and teenagers. Allergy 2008,63, 370–373. [CrossRef] 60. Hoffmann, B.; Münch, S.; Schwägele, F.; Neusüß, C.; Jira, W. A sensitive HPLC-MS/MS screening method for the simultaneous detection of lupine, pea, and soy proteins in meat products. Food Control 2017 ,71, 200–209. [CrossRef] 61. Holden, L.; Moen, L.H.; Sletten, G.B.; Dooper, M.M. Novel polyclonal-monoclonal-based ELISA utilized to examine lupine (Lupinus species) content in food products. J. Agric. Food Chem. 2007 ,55, 2536–2542. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). 41 foods Article Comparison of Various Soybean Allergen Levels in Genetically and Non-Genetically Modified Soybeans Ayato Matsuo 1, Kaho Matsushita 1, Ayano Fukuzumi 1, Naoki Tokumasu 1, Erika Yano 1, Nobuhiro Zaima 1,2 and Tatsuya Moriyama 1,2,* 1Department of Applied Biological Chemistry, Graduate School of Agriculture, Kindai University, Nara 631-8505, Japan; [email protected] (A.M.); [email protected] (K.M.); [email protected] (A.F.); [email protected] (N.T.); [email protected] (E.Y.); [email protected] (N.Z.) 2Agricultural Technology and Innovation Research Institute, Kindai University, Kindai University, Nara 631-8505, Japan *Correspondence: [email protected]; Tel.: +81-742-43-8070 Received: 27 February 2020; Accepted: 16 April 2020; Published: 21 April 2020 Abstract: Several analyses of allergen levels have been reported as part of the safety assessment of genetically modified (GM) soybean; however, few comprehensive analyses have included new allergens. Thus, in this study the levels of eight major soybean allergens, including Gly m 7 (a newly reported soybean allergen), were semi-quantitatively detected in six GM soybeans and six non-GM soybeans using antigen-immobilized ELISA and immunoblotting. We also analyzed the IgE-reactivity to thesesoybeans through immunoblotting, usingsera from three soybean-allergic patients. The results showed that there were no significant differences in the levels of the major soybean allergens in the GM and non-GM soybeans. Moreover, there were no significant differences in the serum IgE-reactive protein profiles of the patients, as analyzed using immunoblotting. These results indicate that, in general, CP4-EPSPS-transfected GM soybeans are not more allergenic than non-GM soybeans. Keywords: soybean; allergens; allergenicity; genetically modified; Gly m 7 1. Introduction Food resource problems associated with climate change, environmental destruction, and population growth are of increasing concern. As a means to overcome these concerns, scientists have developed “genetic modification technology”, which alters the properties of agricultural products. Using this technology, genetically modified (GM) crops have been developed that are not only more resistant to herbicides but also contain beneficial traits such as drought tolerance, delayed ripening, bacterial disease resistance, high oleic acid levels, and pest resistance, to prepare for an increase in global demand [ 1 – 4 ]. The total area of cultivation of GM crops has increased worldwide, to 189,800,000 ha in 2017, with four major GM crops: soybeans (50%), corn (31%), cotton (13%), and rapeseed (5%) [ 5 ]. Soybeans are not only used as a raw material for soybean oil but are also widely used as foods such as tofu, fermented soybeans (natto), miso, soy sauce, and soy milk, and as additives in various processed foods in the form of soy protein isolate (SPI). However, ingestion of soybeans can cause allergic reactions, and various soybean allergens have been identified to date [ 6 ]. Soybean allergies can be divided into class 1 food allergies and class 2 food allergies based on differences in sensitization routes [ 7 , 8 ]; 7S globulin (Gly m 5) [ 9 , 10 ], 11S globulin (Gly m 6) [ 11 ], Gly m 7 [ 12 ], Gly m Bd 30K [ 13 , 14 ], Kunitz-type trypsin inhibitor [ 15 ], oleosin [ 16 ], etc. have been identified as class 1 food allergens causing class 1 allergy. It has been reported that these allergens mainly cause systemic symptoms such as urticaria, diarrhea, vomiting, atopy, and anaphylaxis. Gly m 3 (profilin) [ 17 , 18 ] and Gly m 4 (starvation-associated message 22: SAM22) [ 18 – 21 ] from soybeans Foods 2020,9, 522; doi:10.3390/foods9040522 www.mdpi.com/journal/foods 43 Foods 2020,9, 522 have been reported as class 2 allergens causing class 2 allergy (i.e., pollen–food allergy syndrome [PFAS]). Both Gly m 3 and Glym4arehomologues of Bet v 2 and Bet v 1, which are birch pollen allergens that mainly cause oral allergy syndromes (OASs) [ 18 ], although severe cases of anaphylaxis with facial swelling, airway narrowing, and breathing difficulties have also been reported [8]. Analyses of variation in the relative levels of known endogenous allergens is required to verify whether genetic transformation or transgenes adversely affect human health and whether the level of endogenous allergens is altered by genetic modification. Therefore, studies using sera from soybean allergy patients with IgE antibodies have been conducted using GM soybeans and non-GM soybeans. Lua et al. conducted an IgE-immunoblot and IgE-ELISA using GM soybeans and a closely related variety of non-GM soybeans; they found that GM soybeans had similar allergenicity to non-GM soybeans and identified no changes in the immunoblot results attributable to genetic modification [ 2 ]. Kim et al. reported that IgE-inhibition ELISA using patient serum showed equivalent inhibition in both non-GM soybean and GM soybean extracts, with IgE-immunoblots detecting the most 33 kDa bands in 50% (7/14) of the sera tested and in lanes applied with GM soybean extract and non-GM soybean extract, which were identical to P34 proteins (Gly m Bd 30K). They also argued that the allergenic risk of GM and non-GM soybeans is the same as the allergenic risk of wild-type soybeans because no specific IgE antibodies were detected against the recombinant protein that was genetically integrated into the soybeans, EPSPS (5-enolpyruvylshikimate-3-phosphate synthase) [ 22 ], which confers resistance to the herbicide glyphosate [ 23 , 24 ]. Tsai et al. also reported no significant differences in Gly m 4 levels in non-GM soybean cultivars and GM soybeans (transfected with EPSPS genes and CaMV 35S promoters) [ 25 ]. In addition, there are several reports describing the allergen levels of GM crops [ 26 – 31 ]. Thus, while some studies have explored the allergenicity of GM soybeans, there have been no studies focusing on a wide variety of soybean allergen components, and none that explore differential allergenicity of Gly m 7, a recently discovered soybean allergen. Therefore, in this study, to update the allergenicity assessment of GM soybeans, we analyzed the variability of various soybean class 1 food allergens and soybean class 2 (pollinosis-related) food allergen levels in GM soybeans and non-GM soybeans in vitro and compared the patterns of IgE-binding proteins using sera from soybean allergenic patients. 2. Materials and Methods 2.1. Materials Horseradish peroxidase (HRP)-labeled anti-rabbit and anti-mouse IgGs were obtained from Thermo Fisher Scientific (Waltham, MA, USA), and HRP-labeled anti-guinea pig IgG was obtained from Jackson ImmunoResearch (West Grove, PA, USA). HRP-labeled anti-human IgE was obtained from Kirkegaard and Perry Laboratories, Inc. (Gaithersberg, MD, USA). ECL TM Western blotting reagent and Hyperfilm TM -MP X-ray films were obtained from GE Healthcare (Piscataway, NJ, USA). PVDF membrane (ImmobilonTM-P) was obtained from Millipore (Billerica, MA, USA). 2.2. Soybean Sample Extraction GM soybeans and non-GM soybeans (controls) were obtained from an anonymous seed company. Each sample (approximately 2.5 g) was mixed with distilled water (25 mL), soaked at room temperature (25 ◦ C) for 4 h, and crushed for 30 s in a mixer. Thereafter, the mixture was squeezed with quadruple gauze to obtain a protein extract. The extract was diluted 20and 800-fold with distilled water for detection of the allergen levels. 2.3. Immunochromatography To confirm genetic modification of the GM soybean samples, the transgene CP4-EPSPS (EPSPS derived from Agrobacterium CP4 strain) was detected using the Reveal for CP4 Strip Test Kit (Neogen) according to the instruction manual. 44 Foods 2020,9, 522 Figure 7. Comparison of Gly m 5 levels in GM-and non-GM soybeans by ELISA ( a , b ) and immunoblotting ( c–h ) using rabbit-derived polyclonal antibodies. Soybean protein extracts were evaluated by ELISA ( a , b ) and immunoblotting ( c – h ) for detection of Gly m 5 levels. The ELISA data are presented in absorbance values (Abs). The α - and α ’-subunits of Gly m 5 were detected ( c – e ) separately from the β -subunit of Gly m 5 ( f – h ). The individual data from six GM-and non-GM soybeans ( a – f ) are presented as the mean ± SD of three independent replicates. The collated data ( b , d , g )are presented as the mean ± SD of all individual data points from the control (C1–C6, six non-GM soybeans) or experimental (G1–G6, six GM soybeans) groups relative to the value of control number 1 (C1). Representative immunoblots are also provided (e,h). 51 Foods 2020,9, 522 Figure 8. Comparison of Gly m 6 levels in GM-and non-GM soybeans by ELISA ( a , b ) and immunoblotting ( c – h ) using mouse-derived polyclonal antibodies. Soybean protein extracts were evaluated by ELISA ( a , b ) and immunoblotting ( c – h ) for detection of Gly m 6 levels. The ELISA data are presented in absorbance values (Abs). The acidic subunit of Gly m 6 was detected ( c – e ) separately from the basic subunit of Gly m 6 ( f – h ). The individual data from six GM-and non-GM soybeans ( c , f ) are presented as the mean ± SD of three independent replicates. The collated data ( b , d , g )are presented as the mean ± SD of all individual data points from the control (C1–C6, six non-GM soybeans) or experimental (G1–G6, six GM soybeans) groups relative to the value of control number 1 (C1). Representative immunoblots are also provided (e,h). 52 Foods 2020,9, 522 Figure 9. Comparison of Gly m 7 levels in GM-and non-GM soybeans by ELISA ( a , b ) and immunoblotting ( c – h ). Soybean protein extracts were evaluated by ELISA ( a , b ) and immunoblotting ( c – h ) for detection of Gly m 7 levels. The ELISA data are presented in absorbance values (Abs). Immunoblotting for Gly m 7 levels was performed using a rabbit-derived peptide-antibody ( c – e ) and streptavidin–HRP for the biotin moiety of Gly m 7 ( f – h ). The individual data from six GM-and non-GM soybeans ( c , f ) are presented as the mean ± SD of three independent replicates. The collated data ( b , d , g ) are presented as the mean ± SD of all individual data points from the control (C1–C6, six non-GM soybeans) or experimental (G1–G6, six GM soybeans) groups relative to the value of control number 1 (C1). Representative detections are also provided ( e , h ). ELISA and immunoblotting were performed using animal-derived antibodies or streptavidin–HRP as described in Section 2.5. 3.4. IgE-ELISA and IgE-Immunoblotting using Patient Serum The allergenicity of non-GM soybeans and GM soybeans was then compared by IgE-ELISA and immunoblotting using the sera of three commercial soybean-allergic patients (Figures 10–13). Soybean strains exhibited varying levels of allergenicity to serum IgE as determined by IgE-ELISA, but there was no significant difference between the allergenicity of the non-GM soybean and the GM soybean groups in the sera of all three patients (Figure 10). Furthermore, IgE-binding patterns were evaluated by IgE-immunoblotting and analyzed both visually and by densitometric analysis. IgE-immunoblotting revealed qualitative differences in IgE-binding patterns for different soybean strains in the sera of different patients. For example, in patient serum 1, the peak densitometric intensities of IgE-bound proteins were found at approximately 72 kDa, 37 kDa, and 18 kDa (Figure 11). In patient serum 2, the peak densitometric intensities of IgE-bound proteins were found at approximately 50 kDa and 18 kDa (Figure 12). In patient serum 3, the peak densitometric intensities of IgE-bound proteins 53 Foods 2020,9, 522 were found at approximately 75–50 kDa and 30 kDa (Figure 13). The peak densitometric intensities of IgE-bound proteins should supposedly correlated with the molecular weight of major soybean allergens, such as Gly m 5 and Gly m 6 (full list in Table 1). There were no qualitative differences in the IgE-immunoblotting results of the non-GM soybean and the GM soybean groups for all three patients’ sera; specifically, IgE-binding bands were not increased or decreased in GM soybeans (Figures 11–13). These results indicate that the patient-serum IgE does not specifically bind to the transgene product (CP4-EPSPS). Figure 10. IgE-ELISA of GM-and non-GM soybeans using patient sera. IgE-ELISA performed using sera from three soybean-allergenic patients is shown as follows: patient serum 1 ( a , b ), patient serum 2 ( c , d ), patient serum 3 ( e , f ). The collated data ( b , d , f ) are presented as the mean ± SD of all individual data points from the control (C1–C6, non-GM soybeans) or experimental (G1–G6, GM soybeans) groups for each serum sample. 54 Foods 2020,9, 522 Figure 11. IgE-immunoblotting of GM-and non-GM soybeans using patient serum 1. A representative IgE-immunoblot of GM soybeans (G1–G6) and non-GM soybeans (C1–C6) is shown ( a ). Immunoblots were analyzed on each soybean sample once by densitometry and separated into non-GM soybean ( b ) and GM soybean groups ( c ). The average IgE-immunoblot profiles of non-GM soybeans and GM soybeans were calculated (d). Figure 12. IgE-immunoblotting of GM-and non-GM soybeans using patient serum 2. A representative IgE-immunoblot of GM soybeans (G1–G6) and non-GM soybeans (C1–C6) is shown ( a ). Immunoblots were analyzed on each soybean sample once by densitometry and separated into non-GM soybean ( b ) and GM soybean groups ( c ). The average IgE-immunoblot profiles of non-GM soybeans and GM soybeans were calculated (d). 55 Foods 2020,9, 522 Figure 13. IgE-immunoblotting of GM-and non-GM soybeans using patient serum 3. A representative IgE-immunoblot of GM soybeans (G1–G6) and non-GM soybeans (C1–C6) is shown ( a ). Immunoblots were analyzed on each soybean sample once by densitometry and separated into non-GM soybean ( b ) and GM soybean groups ( c ). The average IgE-immunoblot profiles of non-GM soybeans and GM soybeans were calculated (d). 4. Discussion In immunochromatography, two lines specific to genetic recombination were detected from six GM soybeans. Twelve soybean strains (six non-GM and six GM strains) were evaluated to determine whether genetic modification affected expression of previously identified allergens or IgE allergenicity. Immunoblotting using antibodies to detect the recombinant gene product CP4-EPSPS revealed expression of EPSPS in the six GM soybean strains but not in the non-GM strains. These results confirmed that all six GM soybean species used in this study had been genetically modified and demonstrated that all six non-GM soybean species had not been genetically modified to express EPSPS (Figure 1a–c). SDS–PAGE and CBB staining showed no visible differences in the protein expression profiles of GM soybean and non-GM soybean groups as a whole; and it was speculated that no new protein bands detectable at the CBB staining levels were found to be generated, increased, decreased, or eliminated by the introduction of CP4-EPSPS. One of the major storage proteins, Gly m 5 (7S globulin: β -conglycinin), consists of three subunits ( α subunit, approximately 68 kDa; α ’-subunit, approximately 72 kDa; β -subunit, approximately 50 kDa); the α subunit was first identified as an allergen, and subsequent studies using IgE antibodies from the sera of soybean-allergic patients revealed that the α ’and β subunits were also allergens [ 35 ]. Structural homology between these three subunits is relatively high. Gly m 5, which is found in tofu, a processed soybean food, is stable against pepsin-digestion, and has been reported to be responsible for food-dependent exercise-induced anaphylaxis (FDEIA) [ 36 ]. Gly m 6 (11S globulin) is also known to be a major soybean allergen [ 11 ]. Both Glym5andGlym6areseed storage proteins that account for about 70% of all seed proteins [ 37 ]. In this study, we found that the allergen levels of these two major seed storage proteins do not differ significantly between GM soybeans and non-GM soybeans. 56 Foods 2020,9, 522 The newly discovered soybean allergen Gly m 7 is a unique seed-specific biotinylated protein (SBP) that belongs to the late embryogenesis (LEA) protein family. It was discovered by Riascos et al. in a study evaluating the allergenicity of boiled lentils. The authors generated full-length cDNA clones encoding SBPs identified in lentils from developing soybean seeds and successfully expressed the protein as His-tagged recombinant proteins (rSBP) in Escherichia coli. They succeeded in purification of naturally-derived soybean SBP (nSBP-soy, later named Gly m 7) and confirmed IgE-positive and basophil-stimulating effects between soybean and peanut-allergic sera, suggesting that Gly m 7 may cause IgE-mediated allergic reactions [ 12 ]. In the present study, we detected and compared the levels of this novel allergen by two methods (peptide-antibody and biotin-detection) and found no significant differences between its expression levels in the GM soybean and non-GM soybean groups in either case. Gly m Bd 30K is the predominant allergen found in soybeans. It is a 32 kDa protein also known as the vacuolar protein p34 in soybeans. It has been identified as an oil-body associated component of soybean seeds [ 13 ]. A Kunitz-type trypsin inhibitor was identified as a soybean allergen in 1980 [ 15 ] and was reported to be an occupational inhalant allergen [ 38 ]. There were no significant differences in protein levels between the GM soybean and non-GM soybean groups for any of these classical allergens. The soybean allergen Gly m 4, which belongs to the pathogenesis-related protein10 (PR-10) family, is a homolog of the pollen-antigen Betv1ofbirch.Glym4hasbeen widely reported to cross-react with food PR-10 proteins. Berkner et al. reported immunoblot inhibition assays using recombinant (r)Gly m 4 indicating that rBet v 1 was most inhibited by IgE-binding to rGly m 4 (100%), followed by rGly m 4, apple (rMal d 1), and cherry (rPru av 1) [ 21 ]. Gly m 3 (profilin) also cross-reacts with Bet v 2, another birch pollen-antigen, and is an actin-binding protein present in all organisms (including plants and animals), with more than 70% homology between Gly m 3 and Bet v 2. Rihs et al. reported that there were common IgE-binding epitopes in rGly m 3 and rBet v 2 as determined by EAST (enzyme allergosorbent test) inhibition assays using sera from non-soybean-allergic patients with cypress pollinosis; furthermore, preincubation of sera with rGly m 3 completely inhibited IgE binding to rBet v 2[17]. Quantification of Glym3insome soy products by indirect ELISA was also reported [39]. The immunoblotting and ELISA assays in this study indicated that the levels of these two pollinosis-related soybean allergens (Gly m 3 and Gly m 4) were not increased or decreased by genetic modification (Figures 2and 3). Interestingly, the levels of these two allergens have been reported to be significantly increased by worm wounding [ 28 ]. In particular, since Gly m 4 is a pathogenesis-related protein, it is known that its expression is induced by stresses such as disease. Therefore, it is suggested that the level of these allergens is greatly affected by the cultivation environment. Our results suggest that the level of PR protein is unlikely to be increased by the genetic modification process. Next, IgE-binding was evaluated using the serum of three soybean-allergic patients in order to evaluate the allergenic capacity of GM and non-GM soybeans from a clinical perspective. The IgE-ELISA results showed no significant differences between the GM soybean and the non-GM soybean groups when testing the sera of all three patients (Figure 10). IgE-immunoblotting revealed qualitative differences in IgE-binding patterns for different soybean strains in the sera of different patients. These IgE-binding proteins were supposed to be soybean-major allergens such as Gly m 5, Gly m 6 (Figure 11a–c, Figure 12a–c, Figure 13a–c). There were no qualitative differences in the IgE-immunoblotting results of the non-GM soybean and the GM soybean groups for all three patients’ sera, indicating that the allergen-candidate molecules did not differ between GM soybeans and non-GM soybeans (Figure 11d, Figure 12d, Figure 13d). These results also indicate that patient-serum IgE does not specifically bind to the transgene product (CP4-EPSPS). Taken together, it was concluded that the CP4-EPSPS transfected GM soybeans used in this study had similar allergen abundance levels and allergen reactivity to the non-GM soybeans. These results are similar to other GM soybean allergenicity studies conducted thus far. In this study, we found that GM technology did not increase or decrease the level of endogenous soybean allergen proteins, nor did it induce the appearance of new soybean allergens. However, further investigation of the allergenicity of GM soybeans will be necessary for more rigorous evaluation. Research on more soybean varieties, 57 Foods 2020,9, 522 changes in allergenicity due to changes in the cultivation environment, and sensitization potencies, as well as the allergen levels of GMand non-GM soybeans should be considered. Author Contributions: T.M. designed the study. A.M., K.M., and A.F. performed the experiments. A.M. and T.M. wrote the paper. N.T. and E.Y. supported the experiments. T.M. and N.Z. reviewed and edited the manuscript. All authors read and approved the manuscript. Funding: This work was supported by the Japan Society for the Promotion of Science KAKENHI (Grant-in-Aid for Scientific Research(C)) (grant numbers JP16K07756 and 19K05919 to T.M.). This study was also supported in part by a grant from NPO, “Science of Food Safety and Security (SFSS)” and by a grant from Agricultural Technology and Innovation Research Institute (ATIRI), Kindai University. Acknowledgments: We would like to thank Editage (www.editage.jp) for English language editing. Conflicts of Interest: The authors declare no conflict of interest. References 1. Chrispeels, M.J. Global production and consumption of genetically engineered crops. J. Huazhong Agric. Univ. 2014,4, 120–132. 2. Lua, M.; Jina, Y.; Weber, B.B.; Goodman, R.E. A comparative study of human IgE binding to proteins of a genetically modified (GM) soybean and six non-GM soybeans grown in multiple locations. Food Chem. Toxicol. 2018,112, 216–223. [CrossRef][PubMed] 3. Ichim, M.C. The Romanian experience and perspective on the commercial cultivation of genetically modified crops in Europe. Transgenic Res. 2019,28, 1–7. [CrossRef][PubMed] 4. Shukla, M.; Al-Busaidi, K.T.; Trivedi, M.; Tiwari, R.K. Status of research, regulations and challenges for genetically modified crops in India. GM Crops Food 2018,9, 173–188. [CrossRef][PubMed] 5. ISAAA. Global Status of Commercialized Biotech/GM Crops: Biotech Crop Adoption Surges as Economic Benefits Accumulate in 22 Years; ISAAA: Ithaca, NY, USA, 2017; pp. 100–104, ISBN 978-1-892456-67-2. 6. Ogawa, T.; Samoto, M.; Takahashi, K. Soybean allergens and hypoallergenic soybean products. J. Nutr. Sci. Vitaminol. 2000,46, 271–279. [CrossRef] 7. Yagami, T. Allergies to Cross-Reactive Plant Proteins. Latex-fruit syndrome is comparable with pollen-food allergy syndrome. Int. Arch. Allergy Immunol. 2002,128, 271–279. [CrossRef][PubMed] 8. Amlot, P.L.; Kemeny, D.M.; Zachary, C.; Parkes, P.; Lessof, M.H. Oral allergy syndrome (OAS): Symptoms of IgE-mediated hypersensitivity to foods. Clin. Allergy 1987,17, 33–42. [CrossRef] 9. Ogawa, T.; Bando, N.; Tsuji, H.; Nishikawa, K.; Kitamura, K. α -Subunit of β -Conglycinin, an allergenic protein recognized by IgE Antibodies of soybean-sensitive patients with atopic dermatitis. Biosci. Biotechnol. Biochem. 1995,59, 831–833. [CrossRef] 10. Maruyama, N.; Sato, S.; Cabanos, C.; Tanaka, A.; Ito, K.; Ebisawa, M. Gly m 5/Gly m 8 fusion component as a potential novel candidate molecule for diagnosing soya bean allergy in Japanese children. Clin. Exp. Allergy 2018,48, 1726–1734. [CrossRef] 11. Holzhauser, T.; Wackermann, O.; Ballmer-Weber, B.K.; Bindslev-Jensen, C.; Scibilia, J.; Perono-Garoffo, L.; Utsumi, S.; Poulsen, L.K.; Vieths, S. Soybean (Glycine max) allergy in Europe: Gly m 5 ( β -conglycinin) and Gly m 6 (glycinin) are potential diagnostic markers for severe allergic reactions to soy. J. Allergy Clin. Immunol. 2009,123, 452–458. [CrossRef] 12. Riascos, J.J.; Weissinger, S.M.; Weissinger, A.K.; Kulis, M.; Burks, A.W.; Pons, L. The Seed Biotinylated Protein of Soybean (Glycine max): A BoilingResistant New Allergen (Gly m 7) with the Capacity to Induce IgE Mediated Allergic Responses. J. Agric. Food Chem. 2016,64, 3890–3900. [CrossRef][PubMed] 13. Ogawa, T.; Tsuji, H.; Bando, N.; Kitamura, K.; Zhu, Y.L.; Hirano, H.; Nishikawa, K. Identification of the soybean allergenic protein, Gly m Bd 30K, with the soybean seed 34-kDa oil-body-associated protein. Biosci. Biotechnol. Biochem. 1993,57, 1030–1033. [CrossRef][PubMed] 14. Ogawa, T.; Bando, N.; Tsuji, H.; Okajima, H.; Nishikawa, K.; Sasaoka, K. Investigation of the IgE-binding protein in soybeans by immunoblotting with the sera of the soybean-sensitive patient with atopic dermatitis. J. Nutr. Sci. Vitaminol. 1991,37, 555–565. [CrossRef][PubMed] 15. Moroz, L.A.; Yang, W.H. Kunitz Soybean Trypsin Inhibitor—A Specific Allergen in Food Anaphylaxis. N. Engl. J. Med. 1980,302, 1126–1128. [CrossRef] 58 Foods 2020,9, 522 16. Cao, Y.; Zhao, L.; Ying, Y.; Kong, X.; Hua, Y.; Chen, Y. The characterization of soybean oil body integral oleosin isoforms and the effects of alkaline pH on them. Food Chem. 2015,177, 288–294. [CrossRef] 17. Rihs, H.P.; Chen, Z.; Ruëff, F.; Petersen, A.; Rozynek, P.; Heimann, H.; Baur, X. IgE binding of the recombinant allergen soybean profilin (rGly m 3) is mediated by conformational epitopes. J. Allergy Clin. Immunol. 1999 , 104, 1293–1301. [CrossRef] 18. Yagami, A.; Inaba, Y.; Kuno, Y.; Suzuki, K.; Tanaka, A.; Sjolander, S.; Saito, H.; Matsunaga, K. Two cases of pollen-food allergy syndrome to soy milk diagnosed by skin prick test, specific serum immunoglobulin E and microarray analysis. J. Dermatol. 2009,36, 50–55. [CrossRef] 19. Kleine-Tabbe, J.; Vogel, L.; Crowell, D.N.; Haustein, U.F.; Vieths, S. Severe oral allergy syndrome and anaphylactic reactions caused by a Bet v 1related PR-10 protein in soybean, SAM22. J. Allergy Clin. Immunol. 2002,110, 797–804. [CrossRef] 20. Mittag, D.; Vieths, S.; Vogel, L.; Becker, W.M.; Rihs, H.P.; Helbling, A.; Wüthrich, B.; Ballmer-Weber, B.K. Soybean allergy in patients allergic to birch pollen: Clinical investigation and molecular characterization of allergens. J. Allergy Clin. Immunol. 2004,113, 148–154. [CrossRef][PubMed] 21. Berkner,H.; Neudecker,P.; Mittag, D.; Ballmer-Weber, B.K.; Schweimer, K.; Vieths,S.; Rösch, P.Cross-reactivity of pollen and food allergens: Soybean Gly m 4 is a member of the Bet v 1 superfamily and closely resembles yellow lupine proteins. Biosci. Rep. 2009,29, 183–192. [CrossRef][PubMed] 22. Kim, S.H.; Kim, H.M.; Ye, Y.M.; Kim, S.H.; Nahm, D.H.; Park, H.S.; Ryu, S.R.; Lee, B.O. Evaluating the Allergic Risk of Genetically Modified Soybean. Yonsei Med. J. 2006,47, 505–512. [CrossRef] 23. Wu, H.; Wang, X.; Zhou, X.; Zhang, Y.; Huang, M.; He, J.; Shen, W. Targeting the middle region of CP4-EPSPS protein for its traceability in highly processed soy-related products. J. Food Sci. Technol. 2017 ,54, 3142–3151. [CrossRef][PubMed] 24. Funke, T.; Healy-Fried, M.L.; Han, H.; Alberg, D.G.; Bartlett, P.A.; Schönbrunn, E. Differential Inhibition of Class I and Class II 5-Enolpyruvylshikimate-3-phosphate Synthases by Tetrahedral Reaction Intermediate Analogues. Biochemistry 2007,46, 13344–13351. [CrossRef][PubMed] 25. Tsai, J.J.; Chang, C.Y.; Liao, E.C. Comparison of Allergenicity at Gly m 4 and Gly m Bd 30K of Soybean after Genetic Modification. J. Agric. Food Chem. 2017,65, 1255–1262. [CrossRef] 26. Selb, R.; Wal, J.M.; Moreno, F.J.; Lovik, M.; Mills, C.; Hoffmann-Sommergruber, K.; Fernandez, A. Assessment of endogenous allergenicity of genetically modified plants exemplified by soybean—Where do we stand? Food Chem. Toxicol. 2017,101, 139–148. [CrossRef] 27. Goodman, R.E.; Panda, R.; Ariyarathna, H. Evaluation of Endogenous Allergens for the Safety Evaluation of Genetically Engineered Food Crops: Review of Potential Risks, Test Methods, Examples and Relevance. J. Agric. Food Chem. 2013,61, 8317–8332. [CrossRef] 28. Graf, L.; Hayder, H.; Mueller, U. Endogenous allergens in the regulatory assessment of genetically engineered crops. Food Chem. Toxicol. 2014,73, 17–20. [CrossRef] 29. Fernandez, A.; Mills, E.N.; Lovik, M.; Spoek, A.; Germini, A.; Mikalsen, A.; Wal, J.M. Endogenous allergens and compositional analysis in the allergenicity assessment of genetically modified plants. Food Chem. Toxicol. 2013,62, 1–6. [CrossRef] 30. Panda, R.; Ariyarathna, H.; Amnuaycheewa, P.; Tetteh, A.; Pramod, S.N.; Taylor, S.L.; Ballmer-Weber, B.K.; Goodman, R.E. Challenges in testing genetically modified crops for potential increases in endogenous allergen expression for safety. Allergy 2013,68, 142–151. [CrossRef] 31. EFSA Panel on Genetically Modified Organisms (GMO). Guidance on allergenicity assessment of genetically modified plants. EFSA J. 2017,15, e04862. [CrossRef] 32. Kyhse-Andersen, J. Electroblotting of multiple gels: A simple apparatus without buffer tank for rapid transfer of proteins from polycrylamide to nitrocellulose. J. Biochem. Biophys. Methods 1984,10, 203–209. [CrossRef] 33. Hei, W.; Li, Z.; Ma, X.; He, P. Determination of beta-conglycinin in soybean and soybean products using a sandwich enzyme-linked immunosorbent assay. Anal. Chim. Acta 2012,734, 62–68. [CrossRef][PubMed] 34. Hanafusa, K.; Murakami, H.; Ueda, T.; Yano, E.; Zaima, N.; Moriyama, T. Worm wounding increases levels of pollen-related food allergens in soybean (Glycine max). Biosci. Biotechnol. Biochem. 2018 ,82, 1207–1215. [CrossRef][PubMed] 35. Krishnan, H.B.; Kim, W.S.; Jang, S.; Kerley, M.S. All three subunits of soybean beta-conglycinin are potential food allergens. J. Agric. Food Chem. 2009,57, 938–943. [CrossRef][PubMed] 59 Foods 2020,9, 522 36. Adachi, A.; Horikawa, T.; Shimizu, H.; Sarayama, Y.; Ogawa, T.; Sjolander, S.; Tanaka, A.; Moriyama, T. Soybean beta-conglycinin as the main allergen in a patient with food-dependent exercise-induced anaphylaxis by tofu: Food processing alters pepsin resistance. J. Clin. Exp. Allergy 2009,39, 167–173. [CrossRef] 37. Natarajan, S.; Khan, F.; Song, Q.; Lakshman, S.; Cregan, P.; Scott, R.; Shipe, E.; Garrett, W. Characterization of Soybean Storage and Allergen Proteins Affected by Environmental and Genetic Factors. J. Agric. Food Chem. 2016,64, 1433–1445. [CrossRef] 38. Quirce, S.; Fern á ndez-Nieto, M.; Polo, F.; Sastre, J. Soybean trypsin inhibitor is an occupational inhalant allergen. J. Allergy Clin. Immunol. 2002,109, 178. [CrossRef] 39. Amnuaycheewa, P.; Gonzalez de Mejia, E. Purification, characterisation, and quantification of the soy allergen profilin (Gly m 3) in soy products. Food Chem. 2010,119, 1671–1680. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). 60 Foods 2019,8, 391 stress, for example, in wheat [ 87 , 88 ], chickpea [ 89 ], and soybean [ 90 , 91 ]. In contrast, some other studies reported reduction of protein content in response to other environmental stresses [ 92 ]. These differences may be attributable to the intensity and duration of stresses imposed on plants. Additionally, at the beginning of seed filling stage under unfavorable conditions, proteins related to protection against stress are probably synthesized (increasing protein content), whereas a reduction in protein content is due to their hydrolysis and degradation. In lentil, heat stress is reported to reduce protein content (26%–41% [ 93 ]). Heat stress also inhibited the accumulation of globulins, albumins, glutelins, and prolamins. Excluding proline, glycine, alanine, isoleucine, leucine, and lysine, which increased under heat stress, the rest of the amino acids significantly decreased [ 93 ]. The decrease in lentil storage proteins and most of the AA composition profile resulting from high temperature stresses may be explained by the inactivity of biosynthetic enzymes [ 94 ], and by changes in nitrogen content [ 95 ]. The increase in some AA (e.g., proline and glycine) under stress conditions may be the effects of osmoregulation mechanisms [ 96 ]. Further investigation is required to determine the impacts of environmental stresses on protein content, protein fractions and AA composition, especially in legume crops such as lentil. 8. Yield and Protein Relationships and Stability For lentil breeders, a major challenge is the simultaneous increase of both yield and protein content while maintaining progress in the development of resistance to biotic and abiotic stresses. Hamdi et al. [ 14 ], using two large sets of ICARDA lentil germplasm (829 +987 accessions), found negative correlation between seed protein content and seed yield. They also reported high heritability (0.84) for protein content. Erskine et al. [ 16 ] showed the same trend in lentil for a smaller germplasm set. Later, Stoddard et al. [ 17 ] showed lack of correlation between protein and yield. More recently, Lizarazo et al. [ 97 ] reported negative relationships between protein concentration and seed yield in 14 lentil cultivars grown in a boreal growing environment. This may suggest that independent selection of both characters during breeding is challenging, i.e., the rate of gain in one trait being reduced by that in the other. Barulina [ 12 ] indicated that the protein content varied little across locations among lentil accessions. Similar results are reported for lentil from different authors [ 17 , 35 , 97 ] and for other legume crops (e.g., [ 67 , 98 ]). These observations suggest there is low G × E interaction for protein content and AA composition. This supports the hypothesis that the nitrogen fixing ability of legumes makes their protein concentration relatively stable across environments [ 17 ]. It is known that the protein content of the seed is highly affected by soil nitrogen level, so in legumes the Rhizobium bacteria may greatly enhance the percentage seed protein. For example, Ivanov [ 99 ] reported large differences between seed protein content of non-inoculated and inoculated chickpea plants, 12.6% and 31.2%, respectively. 9. Agronomic Protein Yield of Lentil Protein yield is calculated as protein fraction × grain yield. In lentil, a value of 0.33 t ha −1 protein was reported by Erskine et al. [ 16 ]. Khatun et al. [ 100 ] reported 0.2 t ha −1 for lentil protein yield grown in Bangladesh. Lizarazo et al. [ 97 ] reported protein yield of 0.4 t ha −1 for lentil grown in northern Europe, which was less than that for faba and (1.6 t ha −1 ) and narrow-leafed lupin (1.1 t ha −1 ). In pea, protein yield has been reported at 0.7 t ha−1[101] and 0.9 t ha−1[102]. The potential protein yield of lentil in specific environments has most likely not been fully explored due to limitations imposed by the narrow genetic base of most lentil breeding programs, which are not able to fully exploit the genetic potential due to adaptation bottlenecks. With the recent increasing emphasis on genomics within breeding programs, it may be possible to more fully explore the genetic potential for improvements in protein quantity and quality. The AGILE (Application of Genomic Innovation in the Lentil Economy) project (https://knowpulse.usask.ca/study/2675314), which evaluated phenotypic influences of temperature and photoperiod of 324 diverse lentil genotypes in replicated trials in the three main global agro-ecological regions of lentil production, may provide 67 Foods 2019,8, 391 deeper understanding of aspects of protein quality and protein yield potential of lentil. Genomic information of the 324 sequenced lentil genotypes grown across the three major agro-ecological zones for lentil production may further provide the understanding of the underlying genetics of protein quality and protein yield in lentil. 10. Seed Crude Protein Determination in Lentil The two most prevalent protein determination methods, Kjeldahl and Dumas combustion, are commonly used for grain and seed protein analysis in crops including lentil [ 12 , 17 , 19 ]. The methods rely on the release of nitrogen from the amine groups found in the peptide bonds of the polypeptide chains of protein. The traditional Kjeldahl method is based on oxidation to release nitrogen, while the Dumas combustion method breaks down the bonds in the peptide chains, permitting the release of nitrogen through complete combustion of the sample [ 103 ]. The released nitrogen content is multiplied by a factor to measure protein content [ 104 ]. The factor varies between crop species depending on nitrogen content of protein between 13% and 19%. For pulses, the average nitrogen (N) content of protein was found to be about 16%, which led to use the calculation N × 6.25 to convert nitrogen content into protein content [105]. In a series of experiments, most researchers employed a Kjeldahl method to determine crude protein in lentil. For example, Barulina [ 12 ] and Hawtin et al. [ 13 ] determined seed protein content in lentil accessions using a macro–Kjeldahl (Table 2). Recently, different versions of a modified Kjeldahl method were adopted to measure protein content in lentil seeds [ 15 , 17 ]. Although the Kjeldahl method was more precise and frequently used to analyze protein, this method is under threat by the challenge of safer, clearer, and faster instruments employed in the Dumas Combustion method [ 103 ] which resulted in faster, safer, and more reliable data for protein content in seeds compared to the Kjeldahl method. The Dumas combustion method was adopted to determine lentil seed protein content by Tahir et al. [ 8 ] when they compared lentil protein with other pulses in Canada using AACC (American Association of Cereal Chemists) method 46–30 to determine percent crude protein (CP; N × 6.25) through the use of a LECO CNS-2000 Nitrogen Analyzer (LECO Corporation, St. Joseph, MI, USA, Model No. 602-00-500). A rapid test method using near–infrared (NIR) spectroscopy as a complement to current protein determination using either the Kjeldahl or Dumas combustion method was also successfully applied to estimate protein content in lentil seeds using a different model of analyzer [ 14 , 16 , 19 ]. Protein measured through NIR was validated by calibrated value against measurements obtained through either Kjeldahl or Dumas methods using representative samples. The NIR method was found to be a rapid, low cost, and green complementary technique as it does not use chemicals and reagents [ 106 ]. NIR is a useful high throughput method for estimating protein content for lentil breeding programs if calibrated curves are used to validate the method. The value of protein concentration using three different methods in lentil research are illustrated in Table 2. 11. Protein Isolation Methods and Extraction Lentils are traditionally consumed as whole seed, dehulled split seeds, or as footballs (cotyledons remain attached) in salads and soups or stews commonly known as ‘dal’ [ 4 , 107 ]. Diverse and novel applications are needed to identify ways to increase the use of lentils in the food industry. Nutritional components in lentil seeds such as dietary fiber, starch and protein concentrates or isolates can be extracted and separated [ 38 ]. These can be used as ingredients in the preparation of diverse value-added food products. Isolation or separation of seed proteins from pulses is possible using wet or dry processes [ 108 – 110 ]. Dry processes, such as pin milling and then air classification, are designed to differentiate fractions of starch and protein based on size and density. Air classification separates milled lentils into a light to fine fraction (the protein concentrate) and a heavy or coarse fraction (the starch concentrate) [ 111 , 112 ]. Protein concentrates produced by air classification through dry processes generally contain 38%–68% protein [ 113 , 114 ]. In the past, air classification processes were well adapted to extraction of isolates of 68 Foods 2019,8, 391 lentils and peas because of the large diameter and fairly uniform distribution of starch granules [ 112 ]. The dry method is a relatively easy and simple process, however, efficacy of separation is not high enough to yield high protein concentration. Currently, the wet method for extraction is more widely adopted for legume protein extraction [ 108 ]. The extraction of pulse proteins through wet methods may be relatively easy and reliable, as they are highly soluble under alkaline and acidic conditions. In wet methods, protein is extracted by solubilization in an alkaline solution by dispersing pulse flour in water at pH 8–10, followed by stirring of the dispersion. Then, the insoluble material is removed by centrifugation and proteins are recovered by adjusting the, supernatant pH to a value around 4.5, where proteins are precipitated isoelectrically. The precipitation is usually carried out at the isoelectric point of the protein at which its solubility is the lowest, which for lentil protein is around pH 4.5. The final concentrate or isolated protein is then dried using spray-drum or freeze-drying methods [ 115 ]. Lentil isolates prepared with an alkaline process yield overall 80% of protein [ 108 , 116 , 117 ]. Many researchers have used various wet fractionation methods to isolate lentil protein [ 108 , 109 , 116 – 118 ] under single or multiple isoelectric pH conditions using diluted sodium hydroxide [ 36 ] (Table 5). For example, Boye et al. [ 38 ] reported they extracted lentil protein isolates from red and yellow cotyledon lentils using isoelectric precipitation at pH 9 and 25 ◦ C using a 1:10 solid to solvent ratio, resulting in protein concentrates between 78.2% and 88.6%. Similarly, Joshi et al. [ 119 ] extracted lentil protein isolate by alkaline extraction at pH 8 using 1:10 solid to solvent ratio at room temperature when they studied physicochemical characteristics of the isolated protein that obtained from three drying methods (freeze, spray and vacuum drying). In another study, Kaur et al. [120] revealed that yield of lentil protein isolate ranged from 81.7%–83.5% for Indian cultivars when they performed the protein isolation using isoelectric precipitation pH 4.5. In contrast, Alsohaimy et al. [ 117 ] found the highest protein recovery (93% and 100%) from lentil isolate at isoelectric pH of 12 with ammonium sulphate and alcohol precipitation solvent, with a 5:100 solid to solvent ratio. They compared seven different pH values ranging from 6 to 12 with three different protein recovery methods—isoelectric precipitation, ammonium sulphate preparation, and alcohol precipitation. Similarly, Lee et al. [ 115 ] reported pH 9 at 30 ◦ C as the optimum extraction condition for green lentil that yielded 56.6% protein, and pH 8.5 at 35 ◦ C for red lentil that yielded 59.3% protein when they compared five pH levels (distilled water, pH 8, 8.5, 9, and 9.5) and four temperatures (22, 30, 35, and 40 ◦ C). Johnston et al. [ 121 ] used a modified isoelectric precipitation procedure by adjusting initial pH to 9 initially and then collecting lentil protein isolate at pH 4.6 with 1:10 solid to solvent ratio. Cultivar, particle size of the flour, type of solubilizing agent, temperature, and pH of extraction medium influenced the protein yield and quality [110]. Table 5. Conditions for wet fractionation methods for extraction of lentil protein in six recent studies. Conditions for Lentil Protein Extraction % Protein Yield % Protein in Final Extracts Reference pH: 6, 7, 8, 9, 10, 11, and 12 Temperature: room Solid to solvent ratio: 5:100 80.0 21.5 Alsohaimy et al. [117] pH: 8, 8.5, 9, 9.5 Temperature: 22 ◦ C, 30 ◦ C, 35 ◦ C, and 40 ◦ C Solid to solvent ratio: 1:10 56.6–59.3 - Lee et al. [115] pH: 9 Temperature: 25 ◦C Solid to solvent ratio: 1:10 50.3–69.1 - Boye et al. [38] pH: 8 Temperature: room Solid to solvent ratio: 1:10 - - Joshi et al. [119] pH: 4.6 Temperature: room Solid to solvent ratio: 1:10 82.0 14.5 Johnston et al. [121] pH: 8, 9 and 10 Temperature: room Solid to solvent ratio: 1:10 70.3–85.7 12.3–16.5 Jarpa-Parra et al. [122] 69 Foods 2019,8, 391 In addition to air classification, a new emerging dry fractionation method, triboelectrostatic separation has emerged recently as a novel solvent-free approach to separate protein isolates in the food industry [ 123 – 125 ]. This method relies on differences in dielectric properties of flour particles instead of their size and density. The basic principle of this technique is that proteins can be electrostatically charged more than carbohydrates, because of the ionizable N-terminus and C-terminus groups in their amino acid residues [ 126 ]. Thus, an electric field can separate protein and carbohydrate rich fractions depending upon their types and magnitudes of charge. Like air classification, the main advantages of this method are that it does not use chemical reagents that render the concentrates unsafe for consumption, or that induce changes in functional character. This method is more energy efficient and effective than air classification because it effectively separates particles that are similar in size and density, but different in charge. The major limitation of this approach is that different components may exhibit similar charges under certain conditions, which reduces concentrate purity, and gravitational force may cause airborne particles [125]. Triboelectrostatic separation of legume flours pneumatically conveys the milled particles through tubes or fluidized/vibrating beds, thus imparting a positive or negative charge to the surface of the constituent protein and carbohydrate particles depending on their tribo-charging behavior and the contact medium. Upon contact charging, the oppositely charged particles are separated in a strong electric field [ 126 ]. The amount of charge gained on a particulate from triboelectric charging depends on factors such as surface conditions, area of contact, speed of rubbing, the materials involved, and humidity [ 123 ]. Electrostatic separation methods have been widely used in the mining and pharmaceutical industries, and the effectiveness of this approach is under investigation in the legume industry. For example, the development of optimization of triboelectric bio-separation using a single-stage separation of navy bean flour was successful [ 126 ]. Later, singleand multi-stage tribo-electrostatic bioseparation processes for dry fractionation of protein concentrate found that the two-stage approach resulted in a protein-rich fraction yield of 38% accounting for 60% of the total protein which was a significantly higher than that of the optimized single-stage triboelectrostatic separation [ 127 ]. Similarly, Jafari et al. [ 124 ] and Tabtabaei et al. [ 128 ] examined the physiochemical and functional properties of navy bean protein concentrated using triboelectrostatic separation. They found that electrostatically separated increased from 25.4% to 43.0% total protein yield of original navy bean flour, the protein fractions protein fractions exhibited superior solubility, superior emulsion stability, foam expansion and foam volume stability compared to the wet-fractionated navy bean protein isolate. This method has been extensively explored for fractionation of protein isolated from navy bean and other legumes, but there is limited information in the literature regarding the fractionation of protein from lentil flour using triboelectrostatic processes. However, the variation in turbocharging behavior of proteins and carbohydrates of flour extends its scope to protein isolation of lentil flours. 12. Bioactive Peptides Food proteins not only provide dietary amino acids but also supply health benefits because of the presence of bioactive peptides, short fragments of 2–20 amino acid residues that are encrypted and inactive within the sequence of the precursor protein. Bioactive peptides play an important role in human health, being released during digestion or food processing (enzymatic hydrolysis, cooking, germination, fermentation, and ripening of foods), then absorbed in the intestine and transported to target tissues where they exert specific physiological effects [129]. Proteins from pulses are considered a good source of bioactive peptides. As an example, lentil convicilin was investigated as a source of bioactive peptides using the predictive tools in the BIOPEP database (http://www.uwm.edu.pl/biochemia, see Supplementary Table S1). This storage protein contains a total of 126 peptides encrypted in the amino acid sequence. The following array of biological effects was reported: (i) inhibition of angiotensin I converting enzyme (ACE), dipeptidyl aminopeptidase III and IV, calmodulin-dependent nucleotide phosphodiesterase (CaNPDE), and renin; (ii) stimulation of the release of vasoactive substances and the uptake of glucose; (iii) antioxidant 70 Foods 2019,8, 391 activity, and (iv) regulation of secretion of gastric mucosa. This fact encourages more research studies exploring the biotechnological production of bioactive peptides from lentil proteins for functional food or nutraceutical applications. A few studies are reported with the aim of ensuring efficient peptide release from lentil proteins. Critical processing parameters, such as pH, temperature and time require optimization, and the enzymes or microorganisms used for peptide release require evaluation for efficacy, reproducibility and stability. Enzymatic hydrolysis of lentil proteins has been performed using a wide number of proteolytic enzymes including Savinase, Alcalase, Protamex, Neutrase, Flavourzyme, bromelain, and papain [ 130 , 131 ]. Savinase ® 16L was reportedly the most effective enzyme to produce bioactive peptides from cultivated lentil concentrates [130]. Bioactive peptidesderived from lentilproteinsreportedlyexhibit antihypertensive, antioxidant, and antifungal activities. Table 6summarizes the in vitro effect of protein hydrolysates and bioactive peptides produced during gastrointestinal digestion, enzymatic hydrolysis, germination and fermentation of different lentil-based raw materials. Most studies performed to date have shown that enzymatic hydrolysis of lentil proteins by food grade commercial proteases (savinase, papain, alcalase, flavourzyme, and bromelain), digestive enzymes (pepsin, trypsin, α -chymotrypsin, pancreatin) or germination of lentil seeds (30–40 ◦ C for 5 days) produce peptides with the ability to inhibit angiotensin I converting enzyme (ACE, EC. 3.4.15.1) (see Table 6). ACE is a carboxypeptidase involved in the cleavage of angiotensin I into angiotensin II, a vasoactive peptide that binds with receptors on the vascular wall to cause vasoconstriction, therefore, inhibition of ACE may reduce systolic and diastolic blood pressure [132]. Cultivated lentil proteins treated with Savinase ® produce multifunctional peptides with dual antioxidant and ACE inhibitory activities [ 133 ]. Three peptides were identified to have the highest potencies for inhibiting ACE and delay oxidation of proteins in the presence of oxygen radicals in vitro (LLSGTQNQPSFLSGF, NSLTLPILRYL, TLEPNSVFLPVLLH). The gastrointestinal digestion of these peptides greatly improved their dual biological activity, indicating that smaller peptide fragments with higher biological potency are produced at the gastrointestinal level. The antioxidant/antihypertensive activity of lentil peptides was linked to the primary structure of the C-terminal heptapeptide [ 133 ]. In particular, the ACE inhibition relies on the formation of hydrogen bonds between C-terminal residues of peptides and residues of the ACE catalytic site. The ability of these peptides to inhibit ACE is consistent with earlier studies showing that hydrophobic or aromatic residues or proline residue at the C-terminus positively contribute to the improvement of ACE inhibitory potency [134]. Lentil proteins are also sources of antifungal peptides with potential application as ingredients in the bakery industry. Recently, Rizzello et al. [ 135 ] produced a hydrolysate from a legume flour blend consisting of lentil, pea and faba bean by the combination of fermentation with Lactobacillus plantarum 1A7 and enzymatic hydrolysis with Veron. Among the antifungal compounds of the hydrolysate, four were identified as antifungal peptides derived from lentil lectin. These were purified to confirm their capacity to inhibit the development of Penicillium roqueforti conidia at a minimum inhibitory concentration of 7–9 mg/mL. Similarly, Wang and Ng [ 136 ] isolated a natural antifungal peptide from a red lentil protein extract by chromatographic fractionation. It was able to inhibit 50% of the mycelial growth of Mycosphaerella arachidicola at a concentration of 36 μM. Although bioactive peptides have been identified and isolated from lentil for potential use in functional food and nutraceutical applications, none are currently available in the market for human use. The principal obstacle to the regulatory approval of health claims is the lack of in vivo studies supporting the health and safety claims of bioactive peptides [ 137 ]. To overcome these challenges, future research should focus more on generating data on the safety, efficacy, mechanisms of action, interactions of bioactive peptides with other drugs, absorption, distribution, metabolism, and excretion of bioactive peptides in clinical trials. 71 Foods 2019,8, 391 Table 6. In vitro biological activity of cultivated lentil protein hydrolysates and peptides. Biological Activity Raw Material Processing Conditions Peptide Sequence Effect Observed Reference Antioxidant and antihypertensive Protein concentrate Enzymatic hydrolysis with Savinase 16 L (0.1 U/mg protein, pH 8, 40 ◦C,2h) LLSGTQNQPSFLSGF 1ACE 2inhibition: IC50 3=120 μM ORAC 4: 0.013 μmol Trolox eq./μmol Garcia-Mora et al. [133] NSLTLPILRYL ACE inhibition: IC50 =77.14 μM ORAC: 1.432 μmol Trolox eq./μmol TLEPNSVFLPVLLH ACE inhibition: IC50 =117.81 μM ORAC: 0.139 μmol Trolox eq./μmol Antihypertensive Sprouts Germination (30–40 ◦C for 5 days, 98% humidity) Unknown ACE inhibition: IC 50 =0.044 and 0.034 mg/mL Mamilla and Mishra [138] Antifungal Flour Fermentation with Lactobacillus plantarum (7 log cfu/g) and enzymatic hydrolysis with Veron PS (E/S 5 of 1/400, 30 ◦C, 24 h) HIGIDVNSIK Inhibition of germination of Penicillium roqueforti DPPMAF1 conidia: MIC 6=7–9 mg/mL Rizzello et al. [135] NLIFQGDGYTTK FSPDQQNLIFQGDGYTTK HIGIDVNSIK Antihypertensive Protein isolate Enzymatic hydrolysis with pepsin (E/Sof1/100, pH 2, 37 ◦C for 18 h) Unknown ACE inhibition: IC50 =606 μg/mL Boschin et al. [139] Antihypertensive Protein isolate Pepsin (250 U/mg, pH 2, 37 ◦C, 2 h) and pancreatin (0.7%, pH 7, 37 ◦C,1h) KLRT ACE inhibition: IC50 =0.13–0.02 mg/mL for different peptide fractions Jakubczyk and Baraniak [140] TLHGMV VNRLM Antihypertensive Red protein concentrates Pepsin (E/Sof1/250, for 2 h, pH 2, 37 ◦C) +Trypsin and α-chymotrypsin (E/Sof1/250 for each enzyme, 2.5 h, pH 6.5, 37 ◦C) Unknown ACE inhibition: IC50 =0.090 mg/mL Barbana and Boye [131] Papain (E/Sof1/25, pH 6.5, 4 h, 40 ◦C) Unknown IC50 =0.086 mg/mL Alcalase (1/8 for E/S ratio, pH 7, 1 h, 50 ◦C) + Flavourzyme (E/Sof1/10, pH 8, 1.5 h at 50 ◦C) Unknown IC50 =0.154 mg/mL Bromelain (E/Sof1/4, pH 8, 8 h, 40 ◦C) Unknown IC50 =0.190 mg/mL Green protein concentrates Pepsin (pepsin (E/Sof1/250, for 2 h, pH 2, 37 ◦C) + Trypsin and α-chymotrypsin (E/Sof1/250 for each enzyme, 2.5 h, pH 6.5, 37 ◦C) Unknown ACE inhibition: IC50 =0.053 mg/mL Papain (E/Sof1/25, pH 6.5, 4 h, 40 ◦C) Unknown IC50 =0.080 mg/mL Alcalase (1/8 for E/S ratio, pH 7, 1 h, 50 ◦C) + Flavourzyme (E/Sof1/10, pH 8, 1.5 h at 50 ◦C) Unknown IC50 =0.152 mg/mL Bromelain (E/Sof1/4, pH 8, 8 h, 40 ◦C) Unknown IC50 =0.174 mg/mL 72 Foods 2019,8, 391 Table 6. Cont. Biological Activity Raw Material Processing Conditions Peptide Sequence Effect Observed Reference Antihypertensive Red protein isolates Pepsin (E/Sof4/10, pH 2, 37 ◦C, 2 h) and pancreatin (E/Sof0.5/10, pH 7, 37 ◦C,2h) Unknown ACE inhibition: IC50 =0.008–0.33 mg/mL in gastric phase Akıllıo˘glu and Karakaya [141] IC50 =0.26–0.89 mg/mL in intestinal phase Antifungal Red lentil extract Chromatographic fractionation TETNSFSITKFSPDGNKLIFQGDGYTTKGK Inhibition of mycelial growth in Mycosphaerella arachidicola:IC 50 =36 μMWang and Ng [136] 1 Amino acids are coded according to their one letter abbreviation: A =alanine; C =cystine, D =aspartic acid, E =glutamic acid, F =phenylalanine, G =glycine, H =histidine, I = isoleucine, K =lysine, L =leucine, M =methionine, N =asparagine, P =proline, Q =glutamine, R =arginine, S =serine, T =threonine, V =valine, W =tryptophan, Y =tyrosine. 2 ACE, angiotensin I converting enzyme. 3 IC 50 , inhibitory concentration that reduces 50% of the original enzymatic activity. 4 ORAC, oxygen radical absorbance capacity. 5 E/S, enzyme to substrate ratio. 6MIC, minimal inhibitory concentration. 73 Foods 2019,8, 391 13. Food Applications of Lentil Proteins Lentil protein ingredients represent a viable alternative to proteins from animal-derived sources and soybean, especially since the food industry aims to diversify their formulations because of cultural, religious or ethical dietary restrictions, growing populations, availability, and cost reduction. Although commonly sold as flours, lentil protein ingredients can also be further fractionated into higher protein enriched flours (<60% protein), concentrates (60%–85% protein) or isolates (>85% protein). Enriched flours and dry concentrates are usually produced by air classification methods [ 142 ], whereas concentrates/isolates are produced using wet extraction processes, such as by alkaline extraction followed by isoelectric precipitation [ 109 ]. These fractions are known to have good nutritional and functional value. In terms of functionality, the method and conditions used to produce the ingredients can have a big impact on their functionality within food applications. 13.1. Functionality The majority of published studies have focused on the solubility and emulsifying properties of lentil proteins. Solubility of lentil proteins relates to the balance between protein-solvent and protein-protein interactions. Ladjal-Ettoumi et al. [ 143 ] reported a typical u-shaped pH-dependent solubility profile that was comparable to pea and chickpea proteins, where proteins assume a high surface charge away from their isoelectric point (net charge of 0 mV, pH 4.5) to promote more protein-solvent interactions (e.g., +30 mV at pH 2, and − 40 mV at pH 8). Minimum solubility was found near the isoelectric point (~15%), and higher solubility at pH 2 and 8 (~65%). Can Karaca et al. [ 144 ] found lentil protein isolates at pH 7 to have high solubility (91%) when produced either from alkaline extraction-isoelectric precipitation or by salt extraction. Boye et al. [ 145 ] reported lentil protein concentrates produced by alkaline extraction-ultrafiltration showed greater solubility than those produced by alkaline extraction-isoelectric precipitation, with both showing the u-shaped pH-dependent profile. Lentil protein isolates have demonstrated excellent emulsifying properties. During emulsion formation, proteins migrate to the oil-water interface and then rearrange to orient hydrophobic groups towards the oil phase and hydrophilic groups towards the water phase, lowering interfacial tension. Aggregation of absorbed proteins then creates a viscoelastic interfacial film to stabilize oil droplets from coalescence and gravitational separation. Ladjal-Ettoumi et al. [ 143 ] and Chang et al. [ 146 ] reported lentil protein-stabilized emulsions were most stable at pH levels away from their isoelectric point. Can Karaca et al. [ 144 ] indicated protein isolates prepared by alkaline extraction-isoelectric precipitation to have greater emulsion forming properties and stability than if salt extraction methods were used to prepare the isolates. Primozic et al. [ 147 ] examined the stabilizing effects of high-pressure homogenized lentil protein isolates relative to unmodified isolates. The authors found that homogenization acted to reduce the particle size, hydrophobicity, and interfacial storage moduli of the lentil proteins relative to unmodified proteins, but had no effect on their interfacial tension. Overall, modified lentil proteins showed better physical stability for prepared emulsions than unmodified proteins. Gumus et al. [ 148 ] showed that lentil protein-stabilized emulsions were also effective at inhibiting oxidative reactions within fish oil-in-water emulsions. Water holding and fat/oil absorption capacities relates to the amount of water or oil that a gram of protein material can hold, and to the surface properties (hydrophilic vs. hydrophobic groups), protein/aggregate conformation and solvent conditions used. Aryee and Boye [ 149 ] reported both water holding and fat absorption capacities were improved with wet extraction (i.e., isolate), followed by cooked flour and then raw flour. Boye et al. [ 145 ] indicated that red lentil protein concentrates produced by alkaline extraction-ultrafiltration exhibited greater water holding and fat absorption capacities than concentrates produced by alkaline extraction-isoelectric precipitation. Water holding was similar to that of yellow pea, and greater than that of chickpea. Fat absorption capacity of red lentil protein concentrates (produced from alkaline extraction-ultrafiltration) was much greater than that of other pulses. In the case of foaming, proteins migrate to the air-water interface, re-align and 74 Foods 2019,8, 391 aggregate similar as in emulsions, to form a viscoelastic lamella that entraps gas bubbles [ 39 ]. Toews and Wang [ 150 ] reported that lentil protein concentrate produced the most stable foams and had the highest foaming capacity in comparison to pea, navy bean and chickpea. Lately, investigation of the foaming properties of the pulse cooking water (known as aquafaba) is gaining some interest as an egg replacer [ 151 ]. This highlights an opportunity for food technologists to apply lentil proteins in food applications. 13.2. Challenges Like other pulses, lentil protein ingredients also have unwanted flavour compounds that limit their widespread use. Numerous flavour reduction strategies are available to reduce these compounds in pulses. For instance, Chang et al. [ 152 ] used various organic solvent (acetone, ethanol, and isopropanol) treatments to reduce flavour compounds found in lentil protein isolates, however, this resulted in negative effects on protein functionality. Shariati-Levari et al. [ 153 ] used infrared heating to reduce flavour compounds in lentils, and Ma et al. [ 154 ] reported pre-cooking (roasting and cooking) significantly reduced flavour compounds in green lentils. 13.3. Applications Lentil protein concentrates have been used to replace eggs in production of protein-enriched doughnuts [ 155 ], angel food cake, and muffins [ 156 ]. Lentil flour was used to make gluten-free crackers [ 157 ], lentil flour with transglutaminase has been used as a binding agent to make protein-enriched restructured beef steaks or beef patties [ 158 , 159 ], and lentil protein isolates have been used as an emulsifier to produce salad dressings [ 160 ]. Lentil protein isolates have also been applied as stabilizers for nano emulsion systems [ 147 , 161 ], as encapsulation agents for delivery of omega-3 rich oils [ 162 , 163 ], and in combination with zein, as anti-microbial films [ 164 ], and used to produce nanofibers [ 165 ]. Keeping in mind the many technological functions of lentil proteins, niches are emerging for their inclusion in functional foods, in nutraceuticals, or even in cosmetics. 14. Conclusions Lentil is the most rapidly expanding pulse crop for direct human consumption, and has potential for greater impact as a desirable protein source for food applications. Improvements in lentil protein quality, amino acid composition, and processing fractions will enhance the nutritional quality of this rapidly expanding pulse crop. Genetic strategies focused on increasing the concentration of limiting amino acids are required in lentil. The potential of lentil wild species in breeding programs by introgression of favourable genes for protein improvement may have potential as a long-term breeding strategy. Supplementary Materials: The following are available online at http://www.mdpi.com/2304-8158/8/9/391/s1, Table S1: List of peptides with bioactive potential encrypted in lentil (Lens culinaris Medik.) convicilin (see http://www.uwm.edu.pl/biochemia). Author Contributions: Conceptualization, H.K. and A.V.; writing—original draft preparation, H.K.; writing—review and editing, H.K., M.S., M.N., C.M-V., J.F., and A.V. Funding: This study received support from the Natural Sciences and Engineering Research Council of Canada (NSERC) Industrial Research Chair Program, from the Saskatchewan Pulse Growers (SPG), and from Institute of Food Science, Technology and Nutrition - Spanish National Research Council (ICTAN-CSIC, Madrid, Spain, funded by project AGL2013-43247-R from AEI/FEDER UE) and from the Strategic Research Chair Program of the Saskatchewan Ministry of Agriculture. Conflicts of Interest: The authors declare no conflict of interest. References 1. Krishnan, H.B.; Coe, E.B., Jr. Seed storage proteins. In Encyclopedia of Genetics; Science Direct: Amsterdam, The Netherlands, 2001; pp. 1782–1787. 75 Foods 2019,8, 391 2. Nosworthy, M.G.; Neufeld, J.; Frohlich, P.; Young, G.; Malcolmson, L.; House, J.D. Determination of the protein quality of cooked Canadian pulses. Food Sci. Nutr. 2017,5, 896–903. [CrossRef][PubMed] 3. Vandenberg, A.; Slinkard, A.E. Genetics of seed coat color and pattern in lentil. J. Hered. 1990 ,81, 484–488. [CrossRef] 4. Vandenberg, A. Lentil: Improvement in developing countries. In The Lentil: Botany, Production and Uses; Erskine, W., Muehlbauer, F.J., Sarker, A., Sharma, B., Eds.; CAB International: Oxford, UK, 2009; pp. 391–424. 5. Muehlbauer, F.J. Lentil: improvement in developing countries. In The Lentil: Botany, Production and Uses; Erskine, W., Muehlbauer, F.J., Sarker, A., Sharma, B., Eds.; CAB International: Oxford, UK, 2009; pp. 137–154. 6. Khazaei, H.; Caron, C.T.; Podder, R.; Kundu, S.S.; Diapari, M.; Vandenberg, A.; Bett, K.E. Marker-trait association analysis of iron and zinc concentrations in lentils (Lens culinaris Medik.) seeds. Plant Genome. 2017,10, 1–8. [CrossRef][PubMed] 7. Grela, E.R.; Kiczorowska, B.; Samoli´nska, W.; Matras, J.; Kiczorowski, P.; Rybi´nski, W.; Hanczakowska, E. Chemical composition of leguminous: Part I—Content of basic nutrients, amino acids, phytochemical compounds and antioxidant activity. Eur. Food Res. Technol. 2017,243, 1385–1395. [CrossRef] 8. Tahir, M.; Lindeboom, N.; Båga, M.; Vandenberg, A.; Chibbar, R.N. Composition and correlation between major seed constituents in selected lentil (Lens culinaris Medik). Can. J. Plant Sci. 2011 ,91, 825–835. [CrossRef] 9. Food and Agriculture Organization of the United Nations. FAOSTAT. Available online: http://faostat.fao.org (accessed on 2 September 2019). 10. Semba, R.D. The rise and fall of protein malnutrition in global health. Ann. Nutr. Metab. 2016 ,69, 79–88. [CrossRef] 11. Arsenault, J.E.; Brown, K.H. Effects of protein or amino-acid supplementation on the physical growth of young children in low-income countries. Nutr. Rev. 2017,75, 699–717. [CrossRef] 12. Barulina, H. Lentils of the U.S.S.R. and of other countries: A botanic-agronomical monograph. In Bulletin of Applied Botany, of Genetics and Plant-Breeding; Leningrad Vsesoiuznaia akademiia sel’sko khoziaistvennyk im: Leningrad (St. Petersburg), Russia, 1930; Suppl. S40, pp. 1–319. 13. Hawtin, G.C.; Rachie, K.O.; Green, J.M. Breeding strategy for the nutritional improvement of pulses. In Nutritional Standards and Methods of Evaluation for Food Legume Breeders; Hulse, J.H., Rachie, K.O., Billingsley, L.W., Eds.; IDRC: Ottawa, ON, Canada, 1977; pp. 43–51. 14. Hamdi, A.; Erskine, W.; Gates, P. Relationships among economic characters in lentil. Euphytica 1991 ,57, 109–116. [CrossRef] 15. Kumar, J.; Singh, J.; Kanaujia, R.; Gupta, S. Protein content in wild and cultivated taxa of lentil (Lens culinaris ssp. culinaris Medikus). Indian J. Genet. Plant Breed. 2016,76, 631–634. 16. Erskine, W.; Williams, P.C.; Nakkoul, H. Genetic and environmental variation in the seed size, protein, yield, and cooking quality of lentils. Field Crops Res. 1985,12, 153–161. [CrossRef] 17. Stoddard, F.L.; Marshall, D.R.; Ali, S.M. Variability in grain protein concentration of peas and lentils grown in Australia. Aust. J. Agric. Res. 1993,44, 1415–1419. [CrossRef] 18. Hedley, C.L. Carbohydrate in Grain Legume Seeds; CAB International: Wallingford, UK, 2001; pp. 1–11. 19. Wang, N.; Daun, J.K. Effects of variety and crude protein content on nutrients and anti-nutrients in lentils (Lens culinaris). Food Chem. 2006,95, 493–502. [CrossRef] 20. Karaköy, T.; Erdem, H.; Baloch, F.S.; Toklu, F.; Eker, S.; Kilian, B.; Özkan, H. Diversity of macro-and micronutrients in the seeds of lentil landraces. Sci. World J. 2012,2012, 1–9. [CrossRef][PubMed] 21. Zaccardelli, M.; Lupo, F.; Piergiovanni, A.R.; Laghetti, G.; Sonnante, G.; Daminati, M.G.; Sparvoli, F.; Lioi, L. Characterization of Italian lentil (Lens culinaris Medik.) germplasm by agronomic traits, biochemical and molecular markers. Genet. Resour. Crop Evol. 2012,59, 727–738. [CrossRef] 22. Alghamdi, S.S.; Khan, A.M.; Ammar, M.H.; El-Harty, E.H.; Migdadi, H.M.; Abd El-Khalik, S.M.; Al-Shameri, A.M.; Javed, M.M.; Al-Faifi, S.A. Phenological, nutritional and molecular diversity assessment among 35 introduced lentil (Lens culinaris Medik.) genotypes grown in Saudi Arabia. Int. J. Mol. Sci. 2014 , 15, 277–295. [CrossRef] 23. Heuz é , V.; Tran, G.; Sauvant, D.; Bastianelli, D.; Lebas, F. Lentil (Lens culinaris). Feedipedia, a programme by INRA, CIRAD, AFZ and FAO. 2015. Available online: https://www.feedipedia.org/node/284 (accessed on 2 September 2019). 24. Shewry, P.R. Seed storage proteins: structures and biosynthesis. Plant Cell Online 1995,7, 945–956. 76 Foods 2019,8, 391 159. Baugreet, S.; Kerry, J.P.; Brodkorb, A.; Gomex, C.; Auty, M.; Allen, P.; Hamill, R.M. Optimisation of plant protein and transglutaminase content in novel beef restructured steaks for older adults by central composite design. Meat Sci. 2018,142, 65–77. [CrossRef] 160. Ma, Z.; Boye, J.I.; Simpson, B.K. Preparation of salad dressing emulsions using lentil, chickpea and pea protein isolates: A response surface methodology study. J. Food Qual. 2016,39, 274–291. [CrossRef] 161. Primozic, M.; Duchek, A.; Nickerson, M.T.; Ghosh, S. Effect of lentil protein isolate concentration on the formation, stability and rheological behavior of oil-in-water nanoemulsions. Food Chem. 2017 ,234, 65–74. [CrossRef][PubMed] 162. Chang, C.; Nickerson, M.T. Stability and in vitro release behaviour of encapsulated omega fatty acid-rich oils in lentil protein isolate-based microcapsules. Int. J. Food Sci. Nutr. 2018,69, 12–23. [CrossRef][PubMed] 163. Chang, C.; Varankovich, N.; Nickerson, M.T. Microencapsulation of canola oil by lentil protein isolate-based wall materials. Food Chem. 2016,212, 264–273. [CrossRef][PubMed] 164. Boyaci, D.; Yemenicioglu, A. Expanding horizons of active packaging: Design of consumer-controlled release systems helps risk management of susceptible individuals. Food Hydrocoll. 2018,79, 291–300. [CrossRef] 165. Tam, N.; Oguz, S.; Aydogdu, A.; Sumnu, G.; Sahin, S. Influence of solution properties and pH on the fabrication of electro spun lentil flour/HPMC blend nanofibers. Food Res. Int. 2017 ,102, 616–624. [CrossRef] [PubMed] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). 83 foods Article Overall Nutritional and Sensory Profile of Different Species of Australian Wattle Seeds (Acacia spp.): Potential Food Sources in the Arid and Semi-Arid Regions Kinnari J. Shelat 1,2, Oladipupo Q. Adiamo 1, Sandra M. Olarte Mantilla 1, Heather E. Smyth 1, Ujang Tinggi 3, Sarah Hickey 4, Broder Rühmann 5, Volker Sieber 5and Yasmina Sultanbawa 1,* 1Queensland Alliance for Agriculture and Food Innovations, Health and Food Sciences Precinct, Cooper Plains, Brisbane 4108, QLD, Australia; [email protected] (K.J.S.); [email protected] (O.Q.A.); [email protected] (S.M.O.M.); [email protected] (H.E.S.) 2Australian National Fabrication Facility–Queensland Node, Australian Institute of Bioengineering and Nanotechnology, The University of Queensland, St. Lucia, Brisbane 4067, QLD, Australia 3Health Support Queensland, Queensland Health, Inorganic Chemistry, Forensic and Scientific Services, Coopers Plains, Brisbane 4108, QLD, Australia; [email protected].au 4Karen Sheldon Catering, PO Box 2351, Parap 0812, NT, Australia; [email protected] 5 Department of Chemistry of Biogenic Resources, Technical University of Munich, 94315 Straubing, Germany; broder[email protected] (B.R.); [email protected] (V.S.) *Correspondence: y[email protected]; Tel.: +61-7-3443-2474 Received: 24 August 2019; Accepted: 8 October 2019; Published: 11 October 2019 Abstract: Wattle seed (Acacia spp.)is a well-known staple food within indigenous communities in Australia. A detailed investigation of the overall nutritional and sensory profile of four abundant and underutilized Acacia species—A. coriacea,A. cowleana,A. retinodes and A. sophorae—were performed. Additionally, molecular weight of protein extracts from the wattle seeds (WS) was determined. The seeds are rich in protein (23–27%) and dietary fibre (33–41%). Relatively high fat content was found in A. cowleana (19.3%), A. sophorae (14.8%) and A. retinodes (16.4%) with oleic acid being the predominant fatty acid. The seeds contained high amounts of essential amino acids (histidine, lysine, valine, isoleucine and leucine). A. coriacea is rich in iron (43 mg/kg), potassium (10 g/kg) and magnesium (1.7 g/kg). Pentose (xylose/arabinose), glucose, galactose and galacturonic acids were the major sugars found in the four species. Raw seeds from A. sophorae,A. retinodes and A. coriacea have the highest protein molecular weight, between 50–90 kDa, 80 kDa and 50–55 kDa, respectively. There was variation in the sensory profile of the WS species. This study showed that the four WS species have good nutritional value and could be included in human diet or used in food formulations. Keywords: wattle seed species; nutritional profile; sensory profile; gel electrophoresis 1. Introduction As the world population increases and natural resources diminish, there has been a serious concern on available sustainable nutritious foods [ 1 ]. In addition, a majority of people from developing countries suffer from protein malnutrition, famine and different kinds of diseases due to inadequate food supply and poor quality food [ 2 ]. In order to meet these continued population growth and nutritional requirements, studies are required to examine and discover new sources of food. In the past few years, researchers have focused on the use of underutilized plant products as human food and animal feed [3–5]. The genus Acacia, commonly known as wattle, belongs to the family Fabaceae and it is a large group of woody species comprising of shrubs. Acacia subgenus Phyllodineae are naturally the most Foods 2019,8, 482; doi:10.3390/foods8100482 www.mdpi.com/journal/foods 85 Foods 2019,8, 482 common Acacia species found in Australia and are among the most promising native leguminous plants [ 6 , 7 ]. These Acacias have been reported to exhibit significant potential to lower poverty in semi-arid regions of Africa [ 8 , 9 ]. Moreover, the seeds from various Acacia species, which were used traditionally as source of food by Australian Indigenous population, have been economically revived as food additives, such as emulsifying and flavouring agents [ 10 – 12 ]. Acacia victoriae Bentham is the most common species of Acacia with high water-soluble carbohydrates and protein contents and, thus, have been reported to have significant functional properties in food systems [ 13 , 14 ]. Additionally, Acacia plants have been frequently used to treat diseases, such as fever, leucorrhoea, throat infection, diarrhoea and haemoptysis [15]. However, several Acacia species which are also widely cultivated by indigenous people in different regions of Australia have not been fully utilized in food formulations or incorporated in human diets. These includes A. coriacea and A. cowleana which occurs throughout Northern Australia as well as A. retinodes and A. sophorae that are found in Southern and Southeastern Australia. These Acacia plants, particularly A. retinodes, are mainly used for gum production and ornamental purposes [ 16 ]. Nevertheless, information on the nutritive value of these Acacia species is sparse which may limit their use in foods. Therefore, this study investigated the overall nutritional value of seeds of these abundant and native Australian Acacia species. Furthermore, the seeds were roasted and the molecular weight profiles of protein extracts before and after treatment were examined. In addition, a preliminary sensory profiling of the four wattle seed species was carried out. This study will provide information on whether or not it is advisable to incorporate these seeds into the human diet and indicate the sensory characteristics of these species. 2. Materials and Methods 2.1. Materials Mature seeds of four different species of Australian Acacia species (A. coriacea,A. cowleana, A. retinodes and A. sophorae) used in this study are shown in Figure 1.A. coriacea and A. cowleana were sourced from NATIF Australian Native Superfoods, Fruits Herbs Spices and Mixes, Victoria, Australia, and A. retinodes and A. sophorae were supplied by Valley Seeds Pty Ltd., Victoria, Australia. The seeds from each species were separately ground using a coffee grinder (power: 200 W, time: 30 s) and stored in the refrigerator until further analysis. Additionally, parts of the whole seeds were roasted at 180 ◦ C for 5 min and used to determine the molecular weight profile of the seeds protein extracts for comparison with that obtained from raw seeds. All samples were analysed at least in duplicate and the average for each parameter was reported. 86 Foods 2019,8, 482 Figure 1. Four species of Australian wattle seeds—appearance and size comparison. 2.2. Proximate Analysis The four different species of Acacia seeds were sent to Symbio Alliance Lab Pty Ltd., Eight Mile Plains, Queensland, Australia. A complete proximate analysis was performed at this accredited National Association of Testing Authorities (NATA) laboratory using AOAC [ 17 ] standard methods. The following analysis were measured: moisture (AOAC 925.10) by air oven with a measurement of uncertainty (MU) of ± 15%, ash (AOAC 923.03), crude protein (AOAC 990.03) by Dumas combustion with a MU of ± 10%, crude fat (AOAC 991.36) with a MU of ± 15% and dietary fibre (985.29) with a MU of ±15%, carbohydrate and energy by calculation using information from the Food Standards Code. 2.3. Sugar Analysis A combination of fast liquid chromatography coupled to UV and electrospray ionization trap detection (LC-UV-ESI-MS/MS) was used for the quantification of various sugars [ 18 ]. The hydrolysis was performed in duplicates by adding 6 mL of2MTFAto12mgofgrounded Acacia seeds into 15 mL glass tubes. The tubes were incubated in a heating block (VLM GmbH, EC-Model, Heideblümchenweg, Bielefeld, Germany) for 90 min at 121 ◦ C. After cooling to room temperature, the hydrolysates were neutralized to pH ~8 by adding an aqueous solution of 3.2% NH 4 OH, since light alkaline conditions are required for the subsequent derivatization of monosaccharides. A 25 μ L of neutralized hydrolysate supernatant were derivatized via the high throughput 1 phenyl-3-methyl-5-pyrazolone (HT-PMP) method [ 18 ]. The calibration standards were diluted with neutralized TFA-matrix to compensate the influence on the derivatization process. Each sample was derivatized in triplicates and the carbohydrate fingerprint was analysed. 2.4. Fatty Acid Profiles About 1 g of finely chopped seed samples were taken for initial extraction with chloroform and methanol (2:1) followed by agitation at room temperature for 1 h and centrifuged for 5 min at 3500 × g. Fatty acid profiling was performed at the School of Agriculture and Food sciences, University of Queensland laboratory. The GC-MS (Shimadzu QP2010, Shimadzu Coporation, Tokyo, Japan) was 87 Foods 2019,8, 482 used at oven temperature of 100 ◦ C, injector temperature 250 ◦ C, total program time was 39 min, and helium used as the carrier gas. The inlet pressure used for gas chromatography was 0.4 kPa, at linear gas velocity of 42.7 cm/s, column (Restek stabilwax capillary column; 30 m × 0.25 mm ID × 0.5 μ m film thickness) flow 1.10 mL/min with a split ratio of 1:1 and injection volume of 0.2 μ L. For mass spectrometry, the ion source temperature used was 200 ◦C, the interface temperature was 250 ◦C and the mass range was 35–500 atomic mass units. Identification of the compounds was done by comparing their retention times and mass spectra with corresponding data from a standard food industry FAME Mix (Restek Corporation, Bellefonte, PA, USA). 2.5. Amino Acid Analysis The samples (100 mg per replicate) were first hydrolysed with 6 M HCl at 110 ◦ C for 24 h. As asparagine is hydrolysed to aspartic acid and glutamine to glutamic acid, the reported amount of these acids is the sum of those respective components. After hydrolysis, all amino acids were analysed at the Department of Molecular Science, Australian Proteome Analysis Facility, Macquarie University, NSW, Australia, using the Waters AccQTag Ultra chemistry on a Waters Acquity UPLC. Samples were analysed in duplicate and results are expressed as an average. The coefficient of variation (CV) of the UPLC analysis of amino acids was less than 5%. 2.6. Mineral Analysis A detailed description of method used for mineral analysis is outlined as described by Carter et al. [ 19 ]. The ground seed samples were accurately weighed (0.3 g) into digestion Teflon vessels and concentrated nitric acid (4 mL) was added. The samples were digested using a microwave digestion system (MarsXpress, CEM, Matthews, NC, USA) programmed to three steps: step 1 (400 W power, 85 ◦ C, 14 min), step 2 (800 W power, 110 ◦ C, 20 min), and step 3 (1600 W power, 160 ◦ C, 10 min) and the analysis was performed using inductively coupled plasma mass spectrometry (ICP-MS 7500a, Agilent, Tokyo, Japan) and optical emission spectrometry (ICP-OES, Varian Australia, VIC, Australia). 2.7. Sodium Dodecyl Sulphate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) The SDS-PAGE analysis of lyophilized protein extracts from raw and roasted wattle seeds were conducted at an accredited Protein Expression Facility (PEF), University of Queensland, St Lucia, Queensland, Australia. The extracts were resuspended in PBS to a final concentration of 2 mg/mL. Samples were loaded onto a 4–12% Bis-Tris SDS-PAGE gel and run under denatured and reduced conditions, except where noted. Analysis was performed using a Bio-Rad Chemi-Doc TM XRS +imaging system. 2.8. Rapid Sensory Profiling A sensory tasting session was carried out to develop sensory descriptors of four WS species: A. coriacea,A. retinodes,A. sophorae and A. cowleana. Twelve trained assessors (nine females; three males) with an average age of 50 years old participated in a two-hour session where they provided descriptors for aroma, flavour and aftertaste. For sensory evaluation the seeds of the four species were roasted in the oven for 7 min at 180 ◦ C and ground when cooled. The wattle seeds were presented in two forms for sensory evaluation: as ground seeds on their own and the ground seeds mixed with semolina paste to make them more palatable for the assessors. One gram of ground seeds was presented for aroma assessment in a 30 mL plastic cup covered with a lid and labelled with a three digit blinding code. The ground seeds mixed with semolina where used to conduct a second assessment of aroma and to assess flavour and after taste. Each assessor was presented with6gof1:20 ground wattle seeds and hydrated semolina mix in a 30 mL plastic cup covered with a lid and labelled with the same three digit blinding code used for the non-mixed sample. Green apple and water were used as palate cleansers. Upon completion of the session tasting component the panel leader facilitated session where sensory 88 Foods 2019,8, 482 descriptors of each category for four wattle seed species were summarised and a consensus reached. This resulted in a preliminary sensory profile for each variety. 2.9. Statistical Analysis The data were calculated using Microsoft Excel 2013 and the results are expressed as mean of the triplicate experiments unless otherwise stated. Statistical analysis of the data was done by one-way analysis of variance procedure (ANOVA) using SPSS software (Version 23.0 IBM Corporation, Armonk, NY, USA), and means comparison was done using Duncan’s multiple range test at p<0.05. 3. Results and Discussion 3.1. Proximate Composition As shown in Table 1, all of the species had low moisture content (less than 9%) which are comparable with the 6.9% recorded for A. victoriae Bentham [ 20 ] and the range (6.3–8.0%) for A. tumida and A. colei [ 8 ]. The four species of wattle seeds (WS) showed high protein content, ranging from 22.5% in A. coriacea to 27.5% in A. retinodes. The values for crude protein obtained in this study were higher than that of A. victoriae Bentham [ 20 ] but lower than those found in different subspecies of A. saligna (28.6–32.6%) [ 3 ]. However, the protein contents were within the range (23.4–34.1%) reported for A. tumida and A. colei [8]. The results indicate that WS can serve as a source of protein in the diets of the Aboriginal population in Australia. Moreover, it can be included in food formulations as a source of protein. The crude fat content varied among the four species with relatively high values observed in A. cowleana followed by A. retinodes while A. coriacea had the least value (9.8%). These results showed that A. cowleana seed would be a good source of energy due to its relatively high crude fat content. All the four species studied had similar ash contents (3.4–3.9%) which were comparable to that found in A. victoriae Bentham [ 20 ] and A. tortilis [ 21 ]. All seeds also showed high amounts of dietary fibre ranging from 33.7% in A. cowleana to 41.4% in A. coriacea. These values were higher than that recorded for A. tortilis [ 21 ] and A. victoriae Bentham [ 20 ]. Therefore, the seeds from these Acacia species can be considered as a good source of dietary fibre. Dietary fibre has a vital function in human nutrition by maintaining the health of the gastrointestinal tract, however, in excess may bind to iron and zinc, thereby lowering their bioavailability [ 22 ]. Regarding the carbohydrate content, the seeds have a range of 12.8–15.6% carbohydrate. These levels of carbohydrate were lower than that reported for A. victoriae Bentham [20] and A. tumida [2], due to the greater amounts of crude fat, fibre and protein in the seeds. Traditionally, A. coriacea is known to lower postprandial glucose level [ 23 , 24 ], higher dietary fibre and protein content could be the reason for these potential health benefits. This suggest that using wattle seed flour can be a healthier option. Therefore, the chemical composition of these wattle seed species was found to be nutritious and adding these seeds into the human diets will enhance the nutrition status. Table 1. Proximate composition of four different species of wattle seeds (%, dry weight). Composition A. cowleana * A. coriacea A. sophorae A. retinodes A. victoriae [19] Moisture (%) 5.3 8.6 7.5 5.6 6.9 Crude protein (%) 23.0 22.5 22.7 27.5 17.5 Crude fat (%) 19.3 9.8 14.8 16.4 3.2 Ash (%) 3.4 3.9 3.5 3.7 3.5 Dietary fibre (%) 33.7 41.4 36.0 34.0 29.4 Non-fibre carbohydrate (%) 15.2 13.7 15.6 12.8 67.5 Energy (kJ) 1634.0 1310.0 1485.0 1563.0 1384.0 Values are the means of triplicate analyses. * Estimated measurement of uncertainty (MU) was calculated at 95% confidence interval. A. cowleana:Acacia cowleana. 89 Foods 2019,8, 482 3.2. Fatty Acid Composition The fatty acid composition of the four species of wattle seed are presented in Table 2. The unsaturated fatty acids (UFA) constitute the bulk of the fatty acids, as in the case of certain edible legumes, such as peanut [ 25 ]. Palmitic acid (17.69–23.83%) was found to be the predominant saturated fatty acid (SFA) in all the seeds, followed by stearic acid (4.15–10.12%) with A. coriacea and A. retinodes having significantly (p<0.05) higher values in the two fatty acids, respectively. Monounsaturated fatty acids (MUFA) contributed the larger percentage of the total UFA with values more than twice that of polyunsaturated fatty acids (PUFA) in A. coriacea,A. sophorae and A. retinodes, unlike in other legumes and wattle seed species [ 3 , 26 ]. The C18:1 was found to be the most abundant of all the MUFAs with similar values (50.82–57.57%) obtained in all the species except A. cowleana (36.24%) (p<0.05). However, greatest (p<0.05) amount of PUFA particularly linoleic acid was found in A. cowleana (34.77%) as compared to other species, with A. sophorae having significantly (p<0.05) lowest PUFA value (7.17%). These four species have lower linoleic acids as compared to that of other legumes such as broad beans (48.3%), chickpeas (56.7%) and small lentils (52.3%) [ 26 ], as well as other WS species such as A. saligna [ 27 ] and A. tortilis [ 21 ]. Oils rich in linoleic acid can have an important role in lowering the levels of blood cholesterol [ 28 ], thus the oil obtained from A. cowleana is highly nutritious when consumed regularly as a part of diet. Moreover, only trace amount of n-3 fatty acids was identified in all the seeds. Overall, the high degree of unsaturation found in all the four WS species agrees with those of common vegetable oils indicating that WS composite flour can be a healthy option for human consumption. Table 2. Fatty acid profile of four different species of wattle seeds expressed as percentage ( ± SD) of the total fatty acid profile as determined by FAME GC-MS analysis. Fatty acid (%) A. cowleana A. coriacea A. sophorae A. retinodes Lauric acid (C12:0) 0b0b0.56 ±0.03 a0b Myristic acid (14:0) 0b0b1.33 ±0.07 a0b Palmitic acid (C16:0) 21.3 ±0.35 b23.8 ±0.35 a21.6 ±0.77 b17.7 ±0.11 c Stearic acid (C18:0) 4.15 ±0.05 c4.36 ±0.17 c7.51 ±0.31 b10.1 ±0.21 a Arachidonic acid (C20:0) 0.99 ±0.05 a1.07 ±0.10 a0.86 ±0.03 b0.95 ±0.06 ab Behenic acid (C22:0) 2.49 ±0.07 a1.80 ±0.13 b0.34 ±0.03 c1.02 ±0.04 b Lignoceric acid (C24:0) 0b0b0.16 ±0.01 a0b Total SFA 28.9 ±0.52 c31.0 ±0.75 ab 32.4 ±1.25 a29.8 ±0.42 bc Palmitic acid (C16:1) 3.40 ±0.03 b0.45 ±0.01 c3.84 ±0.27 a0.46 ±0.03 c Oleic acid (C18:1) 32.8 ±0.43 c50.8 ±0.26 b57.6 ±0.49 a50.1 ±1.34 b Ecosanoic acid (C20:1) 0 c0c0.37 ±0.01 b1.57 ±0.02 a Erucic acid (C22:1) 0b0b0b0.52 ±0.01 a Total MUFA 36.2 ±0.46 c51.3 ±0.27 b61.8 ±0.77 a52.7 ±1.4 b Linoleic acid (C18:2) 34.3 ±0.08 a17.4 ±0.33 b6.76 ±0.75 c16.0 ±1.64 b Linolenic acid (C18:3) 0.49 ±0.02 b0.27 ±0.04 c0.41 ±0.04 b1.61 ±0.08 a Total PUFA 34.8 ±0.10 a17.7 ±0.73 b7.17 ±0.79 c17.6 ±1.72 b SFA: Saturated fatty acids; MUFA: Monounsaturated fatty acids; PUFA: polyunsaturated fatty acids. Means not sharing the same superscript’s letters (a, b, c, d) in a row are significantly different at p<0.05 as assessed by Duncan’s multiple range tests. 3.3. Amino Acid Composition The essential amino acid analysis of four different species of WS showed all four of them contained about 13–15% of glutamic acid, followed by about 9–11.5% of aspartic acid, 6.8–7.4% of lysine, 8.0–8.6% of leucine and about 7% of arginine, serine and alanine (Table 3). Moreover, methionine is the limiting amino acids found in all four seeds and this agrees with that reported in other Acacia seed species [ 21 , 22 ]. This is the first time that a complete detailed analysis of all essential amino acids of these four different species of wattle seeds have been investigated. Due to high amounts of protein in all these species, 90 Foods 2019,8, 482 it is important to look at different types of amino acids available in these seeds. It is observed that the glutamic acid content is slightly lower than that of about 18% in mung bean flour [ 29 ]. The amount of lysine in all four seeds is comparable to that of soybean [ 30 ]. The amount of arginine is very similar to that of winged bean and soy beans [ 31 ]. Presence of high amount of arginine potentially can help to increase physiological pool of L-arginine and may have positive impact on cardiovascular health [ 32 ]. Comparing the amino acid profile of the seed proteins with FAO reference pattern [ 33 ] can be used to justify the potential food value of the proteins. The amino acid profile of the four WS species showed that histidine, lysine, valine, isoleucine and leucine had higher levels than those listed in FAO/WHO reference pattern. Table 3. Amino acid (g/100 g dry weight) profile of four different species of wattle seeds. Amino acid A. cowleana A. coriacea A. sophorae A. retinodes FAO/WHO References [32] Essential amino acids Histidine 2.8 2.6 4.3 2.4 1.9 Threonine 4.1 4.0 4.2 4.3 3.4 Lysine 7.0 6.8 7.4 7.0 5.8 Tyrosine 2.0 2.1 2.1 2.2 6.3 Methionine 0.4 0.4 0.3 0.3 Valine 5.9 5.8 6.0 6.2 3.5 Isoleucine 4.4 4.2 4.4 4.3 2.8 Leucine 8.4 8.0 8.6 8.5 6.6 Phenylalanine 3.6 3.3 3.5 3.5 Non-essential amino acids Serine 7.0 7.7 7.1 7.1 Arginine 6.0 7.1 6.0 6.3 Glycine 9.7 10.6 9.5 10 Aspartic acid 10.5 11.5 9.4 10.7 Glutamic acid 15.3 13.3 14.2 14.2 Alanine 7.3 6.7 7.1 7.0 Proline 5.7 5.8 6.0 6.0 Results are expressed as the mean of duplicate experiments. 3.4. Mineral Composition Table 4outlines a detailed mineral analysis performed on four different species of Australian WS. As highlighted in the table, all four species are good sources of important essential minerals such as iron, potassium, magnesium, calcium and zinc. Potassium is the most abundant element in all the WS species with A. coriacea having significantly (p<0.05) highest value (11,000 mg/kg dry weight (DW)). The results obtained agrees with that reported for different WS species [ 8 , 21 , 22 ]. A significantly (p<0.05) higher iron content (195.0 mg/kg DW) was found in A. sophorae as compared to other species of WS, showing it is a good source of iron considering the Australian recommended dietary allowance (RDA) of iron is 7 and 12–16 mg/day for men and women during pregnancy, respectively [ 34 , 35 ]. Moreover, heavy metals, such as Hg, are less than 0.005 mg in all four WS species, indicating the use of all these four species of WS is safe for human consumption. As stated in the Food Standards Australia New Zealand (FSANZ), Standard 1.4.1 for contaminants and Natural Toxicants, the maximum level of Pb in legumes is set at 0.2 mg/kg and cadmium in rice is 0.1 mg/kg. The amount of Pb in all four WS species is within this range and in case of A. sophorae and A. cowleana, they have significantly (p<0.05) higher values (0.1 and 0.175 mg/kg DW, respectively) as compared to other species, but the values still complied with the food standards code [ 36 ]. Based on the above results, the WS are a good source for minerals and thus can be incorporated into foods, such as commercial baked products, that are deficient in minerals to enhance their nutritional properties. Further studies should be carried out to investigate the bioavailability of the essentials minerals, particularly potassium in the WS species. 91 Foods 2019,8, 482 Table 4. Mineral analysis of four different species of wattle seeds. Minerals (mg/kg DW) A. cowleana A. coriacea A. sophorae A. retinodes Major Ca 2300 ±0.0 c4300 ±141.4 a2600 ±424.3 c3150 ±212.1 b K8700 ±0.0 b11000 ±0.0 a7300 ±141.4 c9050 ±495.0 b Mg 1700 ±0.0 b2350 ±70.7 a1700 ±0.0 b2400 ±141.4 a Na <20 c<20 c1100 ±0.0 a940 ±70.7 b P1700 ±0.0 b2350 ±70.7 a2300 ±0.0 a2300 ±141.4 a Trace Co 0.1 ±0.0 c0.04 ±0.0 b0.365 ±0.0 a0.095 ±0.0 c Cr 2.1 ±0.2 a0.8 ±0.1 c1.95 ±0.2 a1.15 ±0.2 b Cu 4.8 ±0.1 c5.0 ±0.0 c8.65 ±0.2 a7.2 ±0.1 b Fe 75.0 ±0.0 b50.5 ±0.7 b195.0 ±35.4 a49.5 ±0.7 b Mn 14.0 ±0.0 c14.0 ±0.0 c46.5 ±2.1 b130.0 ±0.0 a Mo 0.9 ±0.2 c0.84 ±0.1 c1.75 ±0.1 b2.0 ±0.0 a Se 0.9 ±0.0 a0.68 ±0.0 b0.165 ±0.0 c0.34 ±0.0 d Zn 24.5 ±0.7 b23.0 ±0.0 c21.0 ±0.0 d34.0 ±0.0 a Other minerals Al 42.0 ±1.4 b16.5 ±2.1 c77.5 ±3.5 a3.75 ±0.4 d As ND <0.005 b0.87 ±0.2 a<0.005 b Ba 3.2 c49.0 ±1.4 a1.0 ±0.0 d7.15 ±0.1 b Cd <0.005 c<0.005 c0.054 ±0.0 a0.011 ±0.0 b Hg <0.005 a<0.005 a<0.005 a<0.005 a Ni 2.7 ±0.1 c1.1 ±0.0 d3.8 ±0.0 a3.05 ±0.1 b Pb 0.1 ±0.0 ab 0.074 ±0.0 b0.175 ±0.1 a0.0085 ±0.0 b Sb ND 0.01 ±0.0 a<0.01 a<0.01 a Sn <0.05 b0.85 ±0.1 a<0.05 b<0.05 b Sr 19.0 ±0.0 b24.0 ±1.4 a5.9 ±0.1 d13.0 ±0.0 c V0.1 ±0.0 b0.035 ±0.0 b1.65 ±0.4a0.01 ±0.0b ND: not detected. Means not sharing the same superscript’s letters (a, b, c, d) in a row are significantly different at p<0.05 as assessed by Duncan’s multiple range tests. 3.5. Sugar Profile The results in Table 5show that the monosaccharides are the major kind of sugar present in all the four WS species. Pentose sugars (xylose/arabinose) were the predominant monosaccharides found in the seeds and these were followed by galactose and glucose, while a small amount of mannose and fucose were present. Among the species, A. sophorae has a significantly (p<0.05) higher amount of pentose sugars (84.2%), while significantly (p<0.05) higher glucose and galactose contents were found in both A. sophorae and A. retinodes. Moreover, the seeds are rich in galacturonic acid but contain lesser amounts of rhamnose. All these sugars were present in substantial amount in all the species, particularly in A. retinodes,A. sophorae and A. coriacea as compared to A. cowleana. To the best of our knowledge, this study was the first to report the sugar composition present in wattle seeds. Overall, the results revealed that these wattle seed species have high amount of reducing sugar and this may be due to action of endogenous enzymes involved in hydrolysing the stored carbohydrate during maturation and storage. 92 Foods 2019,8, 678 2.3.2. Bacteria Strains and Culture Conditions Microorganisms were purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen(Braunschweig, Germany),LehrstuhlfürTechnischeMikrobiologie,TechnischeUniversität München (Freising, Germany) and Prof. Werner Back (Table 1). The microorganisms were stored as a cryo-culture in our strain collection and were recovered on MRS (De Man, Rogosa, & Sharpe) and TSYE (Trypticase soy yeast extract) agar. The selection of the microorganisms based on the results of previous experiments in which 26 microorganisms were tested for growth in LPI (data not shown). The microorganisms investigated in this study were the eight most promising microorganisms and were further processed. Table 1. Bacteria Strains. Bacteria Strain Lactobacillus reuteri DSM 20016 Lactobacillus brevis TMW 1.1326 Lactobacillus amylolyticus TL 5 Lactobacillus parabuchneri DSM 5987 Lactobacillus sakei subsp. carnosus DSM 15831 Staphylococcus xylosus DSM 20266 Lactobacillus helveticus DSM 20075 Lactobacillus delbrueckii DSM 20081 Liquid cultures were incubated at aerobic (Lactobacillus reuteri,Lactobacillus brevis,Lactobacillus amylolyticus,Lactobacillus sakei subsp. carnosus,Staphylococcus xylosus,Lactobacillus parabuchneri) and anaerobic (Lactobacillus helveticus,Lactobacillus delbrueckii) conditions in sealed tubes (Sarstedt AG & Co, Nümbrecht, Germany) without shaking. Liquid pre-cultures (15 mL) were prepared from single colonies on agar plates and incubated for 36–48 h at 30 ◦ C(L. parabuchneri,L. brevis), 37 ◦ C(L. helveticus, L. delbrueckii,L. sakei subsp. carnosus,L. reuteri,S. xylosus) and 42 ◦C(L. amylolyticus), respectively. 2.3.3. Determining Growth Conditions of Microorganisms Growth curves were recorded with a microplate reader (Infinite M1000 Pro, Tecan Group Ltd., Männedorf, Switzerland) measuring with an OD of 600 nm each 15 min with previous shaking at various temperatures. Aliquots of 200 μ L liquid medium in 96 well micro test plates were inoculated from pre-cultures. For anaerobic conditions, cultures were covered with 100 μL sterile paraffin oil. 2.3.4. Determination of pH and Viable Cell Counts The pH course during lupin fermentation was recorded for 24 h with one measurement point each 30 min with a wtw pH 3310 pH electrode (Xylem Analytics Germany GmbH, Weilheim, Germany). The viable cell counts were determined on nutrients agar from 1 mL of diluted (0.9% NaOH) culture aliquots and expressed as log10 of colony forming units per milliliter per sample (CFU/mL). 2.4. Fermentation of Lupin Protein Isolates Fermentation of LPI was carried out in a 5 L glass reaction vessel in an incubator under aerobic conditions and in a 5 L glass reaction vessel with a fermenter (Satorius) under anaerobic conditions, respectively. A 5% LPI (w/w) solution with 0.5% glucose (w/w) was pasteurized separately at 80 ◦ C for 20 min and mixed under sterile conditions. LPI solution was inoculated with the activated culture in the late exponential growth phase (10 7 CFU/mL). Anaerobic conditions were achieved by flushing the reactor with N 2 . The inoculated lupin protein isolate was incubated for 24 h without stirring and sampled at 0, 4, 18, and 24 h. Fermentation was stopped by heat treatment at 90 ◦ C for 20 min. All samples were neutralized (pH 7) with 1 M NaOH and spray-dried with a Niro Atomizer 2238 (GEA, Düsseldorf, Germany). The whole experiment was repeated in duplicate. 99 Foods 2019,8, 678 2.5. Analysis of d-Glucose For determination of d-glucose, 100 μ L of the sample was mixed with 450 μ L zinc sulphate (10%) and 450 μ L NaOH (0.5 M) and incubated for 20 min at room temperature. After incubation, the sample was centrifuged at 12,045 × gfor 10 min and the supernatant was filtrated using a 0.45 μ m nylon filter. d-Glucose was analyzed using the enzymatic d-glucose test from R-Biopharm (Darmstadt, Germany) following the manufacturer’s instructions. 2.6. Chemical Composition The protein content was estimated based on the nitrogen content according to the Dumas combustion method (AOAC 968.06) with a factor of N × 5.8 [ 15 ] using a Nitrogen Analyzer FP 528 (Leco Corporation, St. Joseph, MI, USA). The dry matter was identified according to AOAC methods 925.10 in a TGA 601 thermogravimetric system (Leco Corporation) at 105 ◦C. 2.7. Sensory Analysis 2.7.1. Panelists The panel consisted of trained volunteers recruited from Fraunhofer IVV (Freising, Germany), exhibiting no known illness at the time of examination and with normal olfactory function. The panel consisted of 10 panelists (eight female and two male). All panelists were trained in weekly training sessions with selected, super-threshold aroma solutions to correctly identify and name fragrances. 2.7.2. Sensory Evaluation For sample evaluation, 2% (w/w) solutions of the LPI and fermented LPI, respectively, and tap water were prepared by stirring. The samples were evaluated in two sessions on one day. Each panelist received five samples of 20 mL aliquots (room temperature) in the first session and four samples in the second session in covered glass vessels, with tap water and flavorless crackers used to neutralize between each sample. To obtain the retronasal aroma and taste attributes, panelists were required to open the lid of the beaker and record the retronasal aroma and taste attributes. After the discussion regarding the most prominent aroma and taste attributes, panelists were asked to select and rate attributes on a scale from 0 (no perception) to 10 (strong perception) that were perceived by at least half of the panel. In addition, the panelists were asked to rate the overall intensity of the aroma of the sample also on a scale from 0 (no perception) to 10 (strong perception), and a hedonic scaling was performed on a scale from 0 (strong dislike) to 5 (neutral) to 10 (strong like). We used supra-threshold aroma solutions for orthonasal perception as a reference for each selected aroma attribute. 2.8. Techno-Functional Properties 2.8.1. Protein Solubility The solubility (%) of LPI and fermented LPI was measured in duplicate at pH 4 and 9 according to Morr et al. [ 16 ]. For each measurement, 1.5 g of protein was suspended in 50 mL 0.1 M NaCl and the pH was adjusted with 0.1 M NaOH or 0.1 M HCl. After stirring for 1 h at room temperature, the non-dissolved fractions of the samples were separated by centrifugation (20,000 × g, 15 min, room temperature) and the supernatants were passed through Whatman No. 1 filter paper to remove any remaining particulates. The protein content of the supernatant was determined by Lowry et al. [ 17 ] using the DC Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA). The absorbance was read at 750 nm and the protein concentration was calculated using the generated BSA standard curve. The resulting amount of dissolved proteins was related to the total amount of protein and the protein solubility (%) was determined. 100 Foods 2019,8, 678 2.8.2. Foam Properties For the determination of the foaming activity, 50 mL of a 5% (w/w) protein solution (pH 7) was whipped at room temperature for 8 min in a Hobart 50-N device (Hobart GmbH, Offenburg, Germany) according to the method described by Phillips et al. [ 18 ]. The increase of foam volume was used to determine the foam activity. The percentage leftover of foam volume after 1 h was defined as the foaming stability (%). 2.8.3. Emulsifying Capacity Emulsifying capacity (EC) was identified according to Wang and Johnson [ 19 ]. Samples were dispersed in deionized water (1%, w/w), adjusted to pH 7 and stirred with an Ultraturrax at 18 ◦ C. Rapeseed oil was continuously added using a Titrino 702 SM titration system (Metrohm GmbH & Co. KG, Hertisau, Switzerland) at a rate of 10 mL/min until phase inversion was detected by means of an LF 521 meter fitted with a KLE1/T electrode (Wissenschaftlich-technische Werkstätten GmbH, Weilheim, Germany). For EC calculation, volume of oil needed to achieve the phase inversion was used (mL oil per g sample). Measurement was repeated in duplicate. 2.9. Analysis of the Molecular Weight Profiles of LPI and Fermented LPI The molecular weight distribution of the untreated and fermented LPI was determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described by Laemmli [ 20 ] with modification under reducing conditions. Untreated LPI, fermented LPI and control samples were re-suspended in 1 mL loading buffer (0.125 mol/l Tris-HCl, 4% SDS (w/v), 20% glycerol (v/v), 0.2 mol/l DDT, 0.02% bromophenol blue, pH 6.8), dissolved for 30 min at 30 ◦ C in an ultrasonic bath and boiled for 5 min at 95 ◦ C in an Eppendorf ThermoMixer C (Eppendorf AG, Hamburg, Germany). Following centrifugation at 12,045 × gfor 10 min (Mini Spin, Eppendorf AG), an aliquot of the supernatant was transferred to a fresh tube and supplemented in a ratio of 1:10 with loading buffer (see above). An aliquot of 10 μ L of each sample was transferred into the wells of pre-cast Criterion TGX stain-free 12% polyacrylamide gels (Bio-Rad Laboratories, Hercules, CA, USA). The samples were separated for 36 min at 200 V (60 mA, 100 W) (Amersham Biosciences Europe GmbH, Freiburg, Germany) at room temperature in a vertical electrophoresis cell (Bio-Rad Laboratories) with a 10–250 kDa Precision Plus Protein Unstained Standard (Bio-Rad Laboratories) alongside as size markers. Protein subunits were visualized using a Gel Doc ™ EZ Imager system (Bio-Rad Laboratories). The molecular weight distribution was determined using Image Lab software (Bio-Rad Laboratories). 2.10. Statistical Analysis Results are expressed as means ± standard deviations and for sensory evaluation (aroma profile) as median ± standard deviations. Data were analyzed using pairwise t-test to determine the significance of differences between a sample and the unfermented LPI, with a threshold of p<0.05. For the microbial growth (CFU, pH, glucose), data were analyzed using one-way analysis of variances (ANOVA) and means were generated and adjusted with Tukey’s honestly significant difference post hoc test to determine the significance of differences between all samples, with a threshold of p<0.05. Statistical analysis was performed with SigmaPlot 12.5 for Windows (Systat Software GmbH, Erkrath, Germany). 3. Results and Discussion 3.1. Chemical Composition Dry matter and protein content of LPI and fermented LPI are given in Table 2. 101 Foods 2019,8, 678 Table 2. Dry matter and protein content of unfermented lupin protein isolate (LPI) and fermented (24 h) LPI. Dry Matter Protein Content LPI (unfermented) 95.4 ±0.0% 89.6 ±0.0% L. reuteri 94.6 ±0.0% * 82.4 ±1.3% L. brevis 94.9 ±0.9% 80.1 ±0.5% * L. amylolyticus 94.7 ±0.2% 79.7 ±0.7% * L. parabuchneri 94.7 ±0.0% * 81.7 ±0.2% * L. sakei subsp. carnosus 94.2 ±0.4% 80.6 ±1.5% S. xylosus 95.3 ±1.1% 80.1 ±1.3% L. helveticus 94.7 ±0.2% 78.5 ±0.4% * L. delbrueckii 92.8 ±1.4% 78.5 ±1.7% The data are expressed as mean ± standard deviation (n=4). Means marked with an asterisk (*) within a column indicate significant differences between the individual sample and the unfermented LPI (p<0.05) following pairwise t-test. Dry matter of LPI and fermented LPI ranged from 92.8% for L. delbrueckii to 95.4% for unfermented LPI. Unfermented LPI contained the highest protein content with 89.6%. 3.2. Comparison of Microbial Growth on Lupin Protein Isolate Solutions The growing parameters consisting of CFU, pH and glucose for all eight microorganisms investigated are shown in Table 3. In addition, growth curves of four of these microorganisms were selected in Figure 1in order to highlight the temporally different transitions into the exponential phase. The results showed that all microorganisms were able to grow in LPI solution. The minimum increase in CFU/mL ( Δ E CFU ) was recorded for L. reuteri with 1.36 × 10 7 CFU/mL and the maximum for S. xylosus with 6.01 × 10 8 CFU/mL. The results of the pH curve showed that L. amylolyticus and L. helveticus appear to have the best metabolism and were adapted most rapidly to the lupine solution. The pH curves showed a direct and constant pH acidification over 24 h for L. helveticus (exemplarily shown in Figure 1a) and L. amylolyticus (similar to Figure 1a). After 24 h of fermentation, a change into the stationary phase could not be observed for both microorganisms. S. xylosus (exemplarily shown in Figure 1b) and L. delbrueckii (similar to Figure 1b) showed a lag phase and a transition into the log phase after approximately 8 h. S. xylosus reached the stationary phase after 14 h. L. delbrueckii did not show a clear transition into the stationary phase after 24 h. L. sakei subsp. carnosus (exemplarily shown in Figure 1c) and L. reuteri (similar to Figure 1c) changed from the lag phase to the log phase after 10 h and reached the stationary phase after 18 h. The largest lag phases were recorded for L. parabuchneri (exemplarily shown in Figure 1d) and L. brevis (similar to Figure 1d). Both microorganisms reached the log phase after 14 h and changed to the stationary phase after 20 h. The results of glucose concentrations showed that the added carbon source of 5 g/kg glucose was metabolized by all microorganisms. After 24 h fermentation, residues of glucose were present in all fermented samples. The degradation of glucose ( Δ E Glucose ) ranged from 3.4 g/kg with remaining 1.7 g/kg glucose after 24 h for L. amylolyticus to 4.8 g/kg with remaining 0.2 g/kg glucose after 24 h for L. reuteri, respectively. Fritsch et al. [ 21 ] and Lampart-Szczapa, Konieczny, Nogala-Kałucka, Walczak, Kossowska and Malinowska [ 9 ] showed similar results and confirmed the suitability of lupine flour and lupine protein, respectively, for lactic fermentation. 102 Foods 2019,8, 678 Table 3. Colony forming units (CFU) (a) and pH, and glucose amount (b) after 0 h, 4 h, 18 h and 24 h of fermentation. (a) CFU (CFU/mL) 0h 4h 18h 24h ΔECFU 1 L. reuteri 2.68 ×106±2.40 ×105a 6.52 ×106±2.02 ×106a 5.77 ×107±5.16 ×106a 1.63 ×107±2.83 ×106a 1.36 ×107±3.07 ×106a L. brevis 1.83 ×107±2.33 ×106a 4.20 ×107±8.91 ×106a 1.24 ×108±3.75 ×107a 1.31 ×108±1.16 ×108a 1.13 ×108±1.19 ×108a L. amylolyticus 1.38 ×107±7.38 ×106a 4.39 ×106±1.11 ×106a 5.77 ×107±2.33 ×106a 5.88 ×107±4.60 ×106a 4.50 ×107±1.20 ×107a L. parabuchneri 1.21 ×107±3.39 ×106a 2.30 ×107±1.00 ×107a 3.43 ×107±7.07 ×105a 5.91 ×107±4.62 ×107a 4.70 ×107±4.29 ×107a L. sakei subsp. carnosus 1.95 ×107±9.19 ×106a 1.74 ×107±5.09 ×106a 1.33 ×107±3.32 ×106a 3.99 ×107±1.56 ×106a 2.04 ×107±7.64 ×106a S. xylosus 1.24 ×107±1.54 ×107a 3.13 ×107±1.27 ×107a 1.86 ×108±1.91 ×108a 6.13 ×108±7.59 ×108a 6.01 ×108±7.75 ×108a L. helveticus 3.08 ×107±1.25 ×107a 3.67 ×107±2.03 ×107a 5.09 ×107±1.12 ×107a 7.44 ×107±1.94 ×107a 4.37 ×107±3.19 ×107a L. delbrueckii 3.10 ×106±9.97 ×105a 3.01 ×106±1.77 ×105a 4.34 ×107±1.51 ×107a 1.20 ×108±1.84 ×107a 1.17 ×108±1.74 ×107a (b) pH Glucose (g/kg) 0h 4h 18h 24h 0h 4h 18h 24h ΔEGlucose 2 L. reuteri 6.5 ±0.0 b6.4 ±0.0 c,d,e 4.9 ±0.0 b,c,d 4.8 ±0.0 b,c 5.0 ±0.0 a4.1 ±0.2 a,b,c 0.9 ±0.1 a0.2 ±0.2 a4.8 ±0.0 d L. brevis 6.6 ±0.0 b,c 6.5 ±0.1 d,e 5.3 ±0.4 c,d,e 5.0 ±0.3 b,c 5.0 ±0.0 a4.7 ±0.3 d3.7 ±1.1 e0.9 ±0.3 d4.1 ±0.2 b L. amylolyticus 6.6 ±0.0 c6.2 ±0.0 c5.6 ±0.3 e5.2 ±0.1 c5.0 ±0.0 a4.1 ±0.1 a,b,c 2.5 ±0.0 c,d 1.7 ±0.3 e3.4 ±0.1 a L. parabuchneri 6.6 ±0.0 c6.5 ±0.0 d,e 5.5 ±0.0 d,e 4.9 ±0.0 b,c 5.0 ±0.0 a4.3 ±0.2 b,c,d 3.3 ±0.3 d,e 0.7 ±0.1 c,d 4.3 ±0.1 b,c L. sakei subsp. carnosus 6.6 ±0.0 c6.3 ±0.0 c,d 4.7 ±0.0 a,b,c 4.7 ±0.0 b5.0 ±0.0 a4.4 ±0.4 c,d 1.4 ±0.0 a,b 0.4 ±0.1 a,b 4.6 ±0.1 c,d S. xylosus 6.6 ±0.0 c6.5 ±0.0 e4.8 ±0.1 b,c 4.8 ±0.0 b,c 5.0 ±0.0 a4.3 ±0.2 b,c,d 1.4 ±0.3 a,b 0.5 ±0.4 a,b,c 4.5 ±0.1 c,d L. helveticus 6.0 ±0.0 a5.4 ±0.1 a4.1 ±0.2 a3.9 ±0.2 a5.0 ±0.0 a3.8 ±0.1 a1.8 ±0.2 b0.6 ±0.1 b,c,d 4.4 ±0.0 b,c L. delbrueckii 6.0 ±0.1 a6.0 ±0.0 b4.2 ±0.1 a,b 4.4 ±0.0 a5.0 ±0.0 a4.0 ±0.4 a,b 2.1 ±0.2 b,c 0.8 ±0.2 c,d 4.2 ±0.1 b,c The data are expressed as mean ± standard deviation from duplicates. Values followed by different letter in a column indicate significant differences between groups (p<0.05) following one-way ANOVA (Tukey). 1ΔECFU =Growth rate over 24 h fermentation; 2ΔEGlucose =Difference of the start and end glucose content of the fermentation. 103 Foods 2019,8, 678 (a) (b) (c) (d) Figure 1. CFU/mL, glucose amount, and pH after 0 h, 4 h, 18 h, and 24 h for L. helveticus ( a ), S. xylosus ( b ), L. sakei subsp. carnosus ( c ), and L. parabuchneri ( d ). The data are expressed as mean ± standard deviation from duplicates. 3.3. Sensory Anaylsis In the retronasal sensory evaluation, the following six aroma qualities and corresponding references were selected by the panelists for the description of LPI and fermented LPI: cheesy (butanoic acid); popcorn-like,roasty (2-acetylpyrazine); earthy,moldy,beetroot-like (geosmin); pea-like, green bell pepper-like (2-isopropyl-3-methoxypyrazine); cooked potato-like (3-(methylthio-)propanal); and oatmeal-like,fatty (oatmeal). Comparative aroma profile analyses (Figure 2) of LPI, L. brevis,L. amylolyticus and S. xylosus were emphasized to highlight the increase of aroma perception in cheesy and roasty,popcorn-like as well as the reduction of the mean aroma perceptions in comparison to LPI, across the panel. The primary aroma attributes in the LPI were pea-like,green bell pepper-like, and oatmeal-like,fatty with median values of 4.5 and 3.0, respectively. Otherwise, the aroma impression was evaluated with low intensities for earthy, moldy,beetroot-like;cooked potato-like, and popcorn-like,roasty with median values of 2.5, 2.5, and 1.0, respectively, while the attribute cheesy was imperceptible (median value of 0). The dominant aroma attributes in the LPI samples obtained after fermentation with L. reuteri were described as cheesy (median of 4.0) and oatmeal-like,fatty (median value of 3.0). The aroma perception was otherwise evaluated with low intensities for earthy,moldy,beetroot-like;pea-like,green bell pepper-like, and cooked potato-like with values of 2.0 and popcorn-like,roasty with a value of 1.0. The aroma perception of the samples 104 Foods 2019,8, 678 obtained after L. brevis fermentation was evaluated with a maximum intensity of 3.0 for oatmeal-like, fatty. Followed by low intensities of earthy,moldy,beetroot-like;cooked potato-like, and popcorn-like,roasty with values of 2.0 and cheesy and pea-like,green bell pepper-like with values of 1.0. Fermentation of LPI with L. amylolyticus was described with a dominant aroma impression of popcorn-like,roasty with a value of 5.0, followed by oatmeal-like,fatty with a value of 3.5. Low intensities were judged for pea-like,green bell pepper-like with a value of 2.0, cooked potato-like with a value of 1.5, and earthy,moldy, beetroot-like with 1.0. Attribute cheesy was not perceptible (median value of 0). The aroma profile of the L. parabuchneri fermented samples showed the aroma impressions of pea-like,green bell pepper-like; cheesy; and oatmeal-like,fatty with intensities of 4.0, 3.0, and 3.0 respectively. Less dominant were the attributes popcorn-like,roasty and cooked potato-like, both with median values of 2.0, and earthy,moldy, beetroot-like with an intensity median value of 1.5. L. sakei subsp. carnosus fermentation was described with popcorn-like,roasty as main aroma impression with an intensity of 4.0, followed by oatmeal-like, fatty with an intensity of 3.0. The attributes earthy,moldy,beetroot-like;pea-like,green bell pepper-like, and cooked potato-like were described equally less intensely with values of 2.0. The attribute cheesy was imperceptible in the samples obtained after fermentation with L. sakei subsp. carnosus. Samples fermented with S. xylosus exhibited a cheesy intensity of 5.0, followed by oatmeal-like,fatty with a value of 3.0. The other attributes were less intense with pea-like,green bell pepper-like (2.5), earthy,moldy, beetroot-like (2.0), popcorn-like,roasty (2.0), and cooked potato-like (1.0). Similar to L. amylolyticus and L. sakei subsp. Carnosus fermentation, the attribute popcorn-like,roasty was described as a dominant aroma impression for the L. helveticus fermented samples with an intensity of 4.0. The aroma profile of L. helveticus was otherwise described with aroma impressions pea-like,green bell pepper-like with a value of 3.0. The attributes oatmeal-like,fatty;cooked potato-like;earthy,moldy,beetroot-like, and cheesy were rated less intensively with values of 2.5, 2.0, 1.0, and 0.5 respectively. The main aroma impression of L. delbrueckii was assessed as oatmeal-like,fatty with an intensity of 4.0. The aroma impressions popcorn-like,roasty;earthy,moldy,beetroot-like;pea-like,green bell pepper-like; and cooked potato-like were rated with an intensity of 2.0. The attribute cheesy could not be observed by the panelists. Figure 2. Retronasal aroma profile analyses of LPI, L. brevis,L. amylolyticus, and S. xylosus fermented samples on a scale from (no perception) to 10 (strong perception). The data are displayed as median values of the sensory evaluations (n=10). 105 Foods 2019,8, 678 The aroma profile showed, with the exception of the L. parabuchneri fermented samples, that the aroma perception pea-like,green bell pepper-like decreased in intensity due to the fermentation. The maximum decreases were determined for the samples obtained after L. brevis fermentation with an intensity of 1.0 compared to unfermented LPI with a value of 4.5. Furthermore, fermentation increased the intensity of the aroma impressions popcorn-like,roasty and cheesy. In the samples fermented with L. amylolyticus,L. sakei subsp. carnosus and L. helveticus the aroma impressions popcorn-like,roasty increased from 1.0 for unfermented LPI to 5.0, 4.0 and 4.0 respectively. S. xylosus and L. reuteri showed an increase in intensity of the attribute cheesy from 0 for unfermented LPI to 5.5 and 4.0 respectively. Several authors confirm for lupin, soy and pea protein, respectively, the significant modification of aroma profile by fermentation [ 6 – 8 , 10 ]. The authors described the reduction of n-hexanal content, which contributes most to the green and beany off-flavor of pea, by fermentation with lactic acid extract in pea protein extract and soy, respectively [ 6 – 8 ]. Further studies showed that the fermentation of soy protein isolate with lactic acid bacteria significantly reduced the aroma impression of beany [ 10 ]. A statement about the reduction of n-hexanal content due fermentation cannot be obtained in this study. A reduced perception of the pea-like,green bell pepper-like aroma may also be caused by masking effects. The total aroma intensities of all samples differed only slightly with median values of 5.5 for the samples fermented with L. amylolyticus, 5.0 for LPI, L. parabuchneri,S. xylosus,L. reuteri, and L. sakei subsp. carnosus fermented samples, 4.5 for L. helveticus and L. delbrueckii, and 4.0 for L. brevis fermented samples. The taste impressions bitter and salty were analyzed by the panel in the sensory evaluation and displayed in Figure 3. The bitter intensity of unfermented LPI was described with a mean value of 2.3. All fermented samples did not differ significantly (p<0.05) in bitterness compared to LPI, with the exception of L. sakei subsp. carnosus (1.0). However, the trend of the mean values of the fermented samples showed slightly lower intensities of bitterness compared to LPI. The intensity of salty was described for LPI with a mean value of 1.9. In comparison, no significant differences (p<0.05) in the intensity of the fermented samples were found. However, the samples obtained after fermentation with L. amylolyticus,S. xylosus, and L helveticus tended to be more salty. (a) (b) Figure 3. Intensity of bitter ( a ) and salty ( b ) taste perception of LPI and fermented LPI rate on a scale from 0 (no perception) to 10 (strong perception). Means marked with an asterisk (*) indicate significant differences between the individual sample and the unfermented LPI (p<0.05) following pairwise t-test. The hedonic evaluation was performed for a first indication of the prevalence of the samples and was not performed according to ISO standards. The evaluation (Figure 4) of the panel resulted 106 Foods 2019,8, 678 in a rating of 4.2 for unfermented LPI. The sample, which was most popular with the panelists (6.4), was fermented with L. sakei subsp. carnosus, followed by the samples fermented with L. helveticus (5.5) and L. amylolyticus (5.4). The most unpopular samples were the ones fermented with S. xylosus and L. reuteri with values of 2.8 and 3.0, respectively. The results showed that both saltiness and bitterness do not have a considerable effect on the acceptance of the samples. The acceptance seems to be influenced by the differences in the aroma profile. It was found that the samples with the aroma attributes popcorn like were rated as popular by all subjects. The samples with the maximum evaluation had the attribute popcorn-like,roasty in their aroma profile as dominant aroma impression and also the highest intensity in this attribute compared to LPI and all other fermented samples. Meanwhile, the aroma impression cheesy dominated in S. xylosus and L. reuteri fermented samples—the samples with the minimal hedonic rating. In addition, both of these samples had the highest intensity of this attribute compared to LPI and the other samples. Figure 4. Rate of hedonic evaluation of LPI and unfermented LPI; scale from 0 (strong dislike) to 5 (neutral) to 10 (strong like). Means marked with an asterisk (*) indicate significant differences between the individual sample and the unfermented LPI (p<0.05) following pairwise t-test. 3.4. Techno-Functional Properties 3.4.1. Protein Solubility The protein solubility of LPI and fermented LPI was determined at pH 4 and pH 7 and given in Table 4. The protein solubility of all samples was higher under neutral condition (pH 7) than under acidic condition (pH 4). Usually, protein solubility is minimal at the isoelectric point (pH 4.5) [ 4 , 22 – 24 ]. Unfermented LPI showed a significantly (p<0.05) higher protein solubility of 63.6% at pH 7 than all fermented samples. The minimum solubility was measured for the samples obtained after L. helveticus and L. delbrueckii fermentation with 23.57% and 27.48%, respectively, and the maximum solubility of the fermented samples with L. reuteri with 42.35%. Similar results have been found in other studies. Lampart-Szczapa, Konieczny, Nogala-Kałucka, Walczak, Kossowska, and Malinowska [ 9 ] determined also lower solubility of lupin proteins after fermentation than non-fermented samples and other authors observed this for fermented soy [10,11,25,26]. Lactobacilli produce organic acids during fermentation, which might have induced an irreversible coagulation of proteins and thus a reduced solubility. Further, heat treatment for the fermentation stop (90 ◦ C, 20 min) might have promoted aggregation and cross-linking of partially hydrolyzed lupin proteins [ 10 ]. In contrast, the protein solubility at pH 4 after fermentation was not significantly different (p<0.05) to unfermented LPI 107 Foods 2019,8, 678 (7.31%), with the exception of the samples fermented with L. delbrueckii (5.55%) and L. helveticus (5.92%). The highest protein solubility at pH 4 was measured for S. xylosus fermented samples with 8.11%. Table 4. Protein solubility (%) at pH 4 and pH 7 of unfermented und fermented (24 h) LPI. Samples Protein Solubility pH 4 pH 7 LPI (unfermented) 7.31 ±0.26 63.59 ±3.04 L. reuteri 7.40 ±0.69 42.35 ±3.76 * L. brevis 7.41 ±0.97 38.45 ±2.87 * L. amylolyticus 7.13 ±0.31 35.47 ±3.16 * L. parabuchneri 8.01 ±0.90 27.37 ±4.00 * L. sakei subsp. carnosus 7.17 ±1.01 37.40 ±4.53 * S. xylosus 8.11 ±0.77 * 28.04 ±3.01 * L. helveticus 5.92 ±0.92 * 23.57 ±2.99 * L. delbrueckii 5.55 ±0.41 * 27.48 ±2.51 * The data are expressed as mean ± standard deviation (n=4). Means marked with an asterisk (*) within a column indicate significant differences between sample and unfermented LPI (p<0.05) following pairwise t-test. 3.4.2. Foam Properties The foam properties (foam activity and stability) of LPI and fermented LPI are shown in Table 5. Allfermentedsamplesshowedasignificanthigher(p<0.05)foamactivitycomparedtounfermentedLPI, with the exception of the samples fermented with L. brevis and L. delbrueckii.L. brevis and L. delbrueckii fermented samples showed an increase in foam activity compared to unfermented LPI, although the differences were not significant (p<0.05). Similar results were obtained by Klupsaite et al. [ 27 ] for lupin proteins and Meinlschmidt, Ueberham, Lehmann, Schweiggert-Weisz, and Eisner [ 10 ] for soy protein isolates. Table 5. Foam activity (%) and foam stability (%) of unfermented and fermented (24 h) LPI. Samples Foam Activity (%) Foam Stability (%) LPI (unfermented) 1613 ±11 89 ±3 L. reuteri 1646 ±20 * 87 ±3 L. brevis 1683 ±57 86 ±6 L. amylolyticus 1688 ±52 * 94 ±2 L. parabuchneri 1703 ±25 * 16 ±5* L. sakei subsp. carnosus 1670 ±32 * 83 ±3 S. xylosus 1678 ±23 * 91 ±1 L. helveticus 1698 ±17 * 20 ±0* L. delbrueckii 1652 ±36 80 ±0* The data are expressed as mean ± standard deviation (n=4). Means marked with an asterisk (*) within a column indicate significant differences between sample and unfermented LPI (p<0.05) following pairwise t-test. Unfermented LPI showed a foam stability of 89%. The foam stability of all fermented samples, with exception of L. parabuchneri and L. helveticus, was above 80%. L. parabuchneri and L. helveticus showed significantly (p<0.05) lower stability of 16% and 20%, respectively, compare to unfermented LPI. 3.4.3. Emulsifying Capacity Unfermented LPI showed an emulsifying capacity of 552.9 mg/mL (Table 6). The treatments with the various microorganisms resulted in emulsifying capacities in the range of 347.7 mg/mL for L. parabuchneri up to 595.6 mg/mL for S. xylosus. Samples fermented with L. parabuchneri (347.7 mg/mL), L. delbrueckii (370.3 mg/mL), L. sakei subsp. carnosus (407 mg/mL) and L. brevis (447 mg/mL), showed a significantly (p<0.05) lower emulsifying capacity than the unfermented LPI sample. The residual samples did not show a significantly (p<0.05) lower emulsifying capacity compared to unfermented 108 Foods 2020,9, 349 enzymatic activities such as alcohol and aldehyde dehydrogenase, enzymes that have been reported to decrease green beany flavor in soybeans through the degradation of aldehydes and alcohols [ 25 ]. Yeasts are known to improve the aroma quality of fermented beverages via the formation of ester compounds with fruity and floral notes such as ethyl acetate and 2-methylbutyl acetate [ 26 ]. The aim of this study is, therefore, to investigate the impact of fermentation with yeasts in co-culture with LAB on the volatile profile associated with off-flavor in peas. The specific objectives of this study are: (1) to evaluate the impact of the addition of yeasts on the acidifying activity of LAB; and (2) to compare the flavor profile of the samples fermented with a pure culture of LAB and the ones fermented with LAB in co-culture with yeasts. 2. Materials and Methods 2.1. Raw Materials, Ingredients, and Strains Pea protein isolates of Pisum sativum L. (Purispea Tm 870, batch1708TL1) were supplied by Cargill (Chicago, IL, USA). The sample had the following characteristics: pH 7.0, moisture 4.2%, and protein content 82.1%. Sucrose from sugar cane was provided by Tereos (France). The raw materials were stored at room temperature. VEGE 047 LYO was obtained from DuPont Danisco (Dang é -Saint-Romain, France) and consisted of freeze-dried defined strains of lactic acid bacteria: Lactobacillus acidophilus NCFM ® , Streptococcus thermophilus,Lactobacillus delbrueckii subsp. bulgaricus, and Bifidobacterium lactis HN019 ™ . Using API 50 CHL medium (Biom é rieux SA; Marcy l’Etoile, France), this mixture of lactic acid bacteria and Bifidobacterium was found to be glucose (+), fructose (+), and sucrose (+). Torulaspora delbrueckii TD 291 (freeze-dried BIODIVA TM ) was provided by LALLEMAND S.A.S, France. Kluyveromyces lactis Clib 196 and Kluyveromyces marxianus 3810 were obtained from the INRA collection (UMR GMPA, Grignon, France). 2.2. Fermentation of Pea Protein Isolates 2.2.1. Inoculum Preparation Kluyveromyces lactis Clib 196 and Kluyveromyces marxianus 3810 were incubated from frozen glycerol stocks ( − 80 ◦ C) in potato dextrose broth (PDB, Becton, Dickinson & Company, Sparks, MD, USA) for 21hat30 ◦ C with an agitation of 200 rpm using an incubator (INFORS HT). When the stationary phase of growth was reached, cells were harvested by centrifugation at 4000 rpm (1699 × g) for 15 min at 10 ◦ C (Eppendorf, 5804R). The pellet was then washed and suspended in sterile physiological water (9 g of NaCl in 1 L of osmotic water) to obtain 10 7 colony forming unit/mL (CFU/mL) and used as the inoculum. For the rehydration of VEGE 047 LYO, we followed the supplier’s instructions. The freeze-dried culture was rehydrated in a 500-mL of a 4% thermally treated pea protein solution using Purispea Tm 870. The mixture was left to settle for 15 min and then agitated gently before the preparation of the cryotubes. The latter was then stored at −80 ◦C. As for Torulaspora delbrueckii, 0.25 g/L was inoculated in a pea protein solution and then transferred to cryotubes at −80 ◦C until the fermentation process. Upon fermentation, the samples were inoculated with 10 7 CFU/mL of VEGE and Torulaspora delbrueckii. 2.2.2. Preparation of Fermented Pea Protein Isolate Osmotic water was used to prepare a 4% pea protein solution with 3% sucrose. The mixture was stirred using a magnetic stirrer at room temperature for 10 min until a homogeneous solution was obtained. The solution was then thermally treated at 110 ◦ C for 15 min. This process was required to eliminate the endogenous microflora of pea proteins before fermentation. Prior to inoculation, the solution was cooled down to 30 ◦C. The initial pH of the solution was 7.1 ±0.1. 115 Foods 2020,9, 349 Four different fermented samples were inoculated with the following strains: VEGE 047 (VEGE), VEGE 047 +Kluyveromyces lactis Clib 196 (VEGE +KL), VEGE 047 +Kluyveromyces marxianus 3810 (VEGE +KM), and VEGE 047 +Torulaspora delbrueckii (VEGE +TOR). The experiments were carried out in triplicate. Fermentations were stopped at pH 4.55 (an optimal pH to ensure sanitary and textural qualities) by rapid cooling in an ice bath until the temperature reached 4 ◦ C. Microbial enumerations were performed at this point. All fermented samples were stored in 150-mL glass vials at 4 ◦ C for 7 days. At day 7, the products were sent to an expert panel for sensory evaluation and the remaining samples were frozen at − 80 ◦ C until analysis. 2.3. Fermentation Monitoring 2.3.1. Acidification Activity Measurement The Cinac system 4 (AMS, Fr é pillon France) [ 27 ] was used to measure the acidification activity of the microbial strains at 30 ◦ C. The pH of the inoculated pea protein samples was continuously measured and automatically recorded at 3-min intervals. The time to reach pH 4.55 (tpH 4.55 in min) was used as a descriptor for the acidification activity. It was calculated using Cinac 4 (version 4, release 0.4.4). Measurements were made in triplicate. 2.3.2. Microbial analyses LAB and yeast populations were determined at the beginning and at the end of the fermentation process when a pH of 4.55 was reached. The samples were diluted 1:10 with sterile physiological water (9 g of NaCl/L) and then homogenized using an Ultra Turrax ® device (Labortechnik, Germany) at 8000 rpm for 1 min. The yeast population was determined by surface plating in triplicate using yeast–glucose–chloramphenicol agar (YGCA, BIOKAR, Beauvais, France) after three days of incubation at 30 ◦ C. Lactic acid bacteria were counted by spread plating technique in triplicate on non-acidified Man Rogosa and Sharpe agar (MRS, BIOKAR, Beauvais, France) after 3 days at 42 ◦ C under anaerobic conditions (Bugbox Anaerobic System, Ruskinn, Bridgend, United Kingdom). 2.3.3. Biochemical Analysis Analyses Using HPLC–MS to Determine Sugar Content To obtain an accurate determination of the amount of sucrose, fructose, and glucose in a complex matrix, a sugar analysis was done using liquid chromatography coupled with mass spectrometry. Sugars were extracted as previously described [ 28 ]. After thawing, the samples were diluted in 50/50 (v/v)LC/MS water/acetonitrile and were quantified using high-performance liquid chromatography coupled with mass spectrometry (Waters, Beaver Dam, WI, USA). Metabolites were separated on an XBridge BEH Amide column (length: 150 mm; internal diameter: 4.6 mm; particle size: 3.5 μ m; WATERS). The column temperature was set at 75 ◦ C. The flow was 0.4 mL/min and the solvent were acetonitrile with 0.1% formic acid (D) and ultra-pure water (B) + 0.1% formic acid. The elution gradient was as follows: 0 min at 80% D +20% B, then 50% D +50% B for 23 min, level at 80% D and 20% D for 2 min. The injection volume was 5 μ L, and the injector temperature was 7 ◦C. Each analysis took 25 min. Mass spectrometric detection was performed with an ISQ ™ EC-LC Quadrupole with a heated electrospray source (HESI–II) operated in the negative ionization mode (Thermofisher Scientific). Metabolites were identified and quantified (ng/g wet weight) using Chromeleon 7.2.10 software (Thermofisher scientific, Waltham, MA, USA). 116 Foods 2020,9, 349 Analyses Using HPLC to Determine Ethanol and Lactic Acid Concentrations The concentrations of ethanol and lactic acid were determined by high performance liquid chromatography (HPLC). Similar to the preparation of sugar extracts, ethanol and lactic acid were extracted as previously described [28]. The analysis was performed using a Waters Associates chromatographic system (Alliance) equipped with a pump, an automatic injector (Waters e2695) and two columns, a pre-column of 30 ×4.6 mm (Bio-Rad Labs; Richmond, CA, USA), and an HPX-87H columns (300 × 7.8 mm; Bio-Rad Labs; Richmond, CA, USA) connected in series. The columns were operated at 35 ◦ C. The samples were eluted with 0.01 N sulfuric acid at a flow rate of 0.6 mL/min. The eluting compounds were detected by a UV detector (Model 2489). This detector was connected in series to an RI detector (Model 2414); Empower TM 3 chromatography data software (Waters Corporation) was used to integrate peak areas using calibration by an external standard solution. 2.4. Sensory Evaluation The sensory evaluation of the four products was performed using descriptive analysis. A panel of 15 trained panelists was recruited for their familiarity with plant-based products. Sensory analysis was carried out in an air-conditioned room (20 ◦ C), in individual booths, under daylight. Samples of the fermented products (80 g) were presented in plastic cups labeled with randomly selected three-digit numbers. The sample evaluation order was balanced over the panel following a Williams Latin square design to account for potential order and carry-over effects. Panelists were asked to rinse their mouths with water and crackers between samples. A one-hour session was dedicated to the generation of attributes, followed by training in the use of these attributes to obtain a quantitative description of the products. The 15 panelists generated a vocabulary of sensory attributes that covered the odor, texture, aroma, and taste of the samples. During the second session, the panelists had to rate the intensity of the 13 attributes (global intensity, sour, bitter, astringent, tangy, sparkling, green flavor/vegetal, leguminous plant, citrus fruit, nut, beer/yeast, sourdough, cultured apple cider) generated for each product on an interval scale ranging from 0 to 15 (from nonexistent to marked). Samples were presented in a monadic sequence. The panel performances were validated in a third session in terms of repeatability, using different means of analysis of variance (ANOVA) 2.5. Aroma Compound Analysis To identify the aroma compounds present in the non-fermented and fermented pea protein solutions, GC/MS analysis was performed. All analyses were performed in triplicate. Volatile compounds were extracted using the purge and trap method by means of a Gerstel Dynamic Headspace System (DHS) coupled with a Gerstel Multipurpose Sampler (MPS) Autosampler (Mulheim an der Ruhr, Denmark). Five grams of the fermented or non-fermented samples were weighed in a vial. The DHS system heated the samples to 40 ◦ C for 3 min with an agitation speed of 500 rpm. The samples were purged with a helium flow at 30 mL/min for 10 min and analytes (volatile molecules) were collected on sorbent material at 30 ◦ C. The sorbent material used for volatile molecule collection was Tenax TA (2, 6-diphenylene oxide polymer) (Gerstel). The sorbent material was dried to remove residual water vapor at 30 ◦C with a helium flow of 50 mL/min for 6 min. GCMS was performed using a 7890 Agilent GC system coupled to an Agilent 5977B quadruple mass spectrometer (Agilent, Santa Clara, CA, USA). A non-polar Agilent column DB-5MS (60 m ×0.32 mm ×1μm) was used. The injection was performed in splitless mode using helium at a flow rate of 1.6 mL/min. The oven temperature of the column was programmed as follows: temperature increase from 40 to 155 ◦ Cat4 ◦ C/min, followed by 155 to 250 ◦ Cat20 ◦ C/min. The oven temperature was then maintained at 250 ◦ C for 5 min. The gas chromatogram was recorded and analyzed for 117 Foods 2020,9, 349 volatile retention time. Volatile compounds identified by comparison with a mass spectra library (NIST database) were chosen based on their percentage of identity. For the quantification of the volatile compounds, two different standard solutions were prepared: Solution A containing the molecules responsible for the off-notes, and solution B comprising the ester molecules. The choice of the volatile compounds responsible for the off-flavor perception in pea was based on their occurrence in the literature. For Solution A, a mixture of 21 selected compounds was prepared. Pure commercial volatile standards belonging to different families of compounds were purchased from Sigma Aldrich (Milwaukee, USA). These standards were as follows: 2-methylpropanal, trans-2-methyl-2-butenal, hexanal, (E)-2hexenal, heptanal, (E)-2-octenal, nonanal, butanal, (E)-2-heptenal, decanal, 1-penten-3-ol, 1-octen-3-ol, 1-hexanol, 1-octanol, 6-methyl-5-hepten-2-one, 2-octanone, 2-nonanone, 2-n-heptylfuran, 2-ethylfuran, and 2-pentylfuran. Solution B consisted of ester compounds, major aromatic molecules resulting from yeast fermentation: isoamyl acetate, 2-methylbutyl acetate, 2-phenylethyl acetate, isobutyl acetate, ethyl octanoate, ethyl hexanoate, hexyl acetate, ethyl isobutyrate, ethyl propionate, propyl acetate, and ethyl acetate. To obtain the calibration curves, four concentrations were prepared for Solution A or B, made from two mother solutions in water and injected three times. To consider the interactions between aroma compounds and proteins, dilutions were carried out in a 4% sodium caseinate solution (Sigma Aldrich, St. Louis, MO, USA). The choice of the dairy protein is justified by the intrinsic contents of off-flavors in pea proteins, the neutral taste, and the negligible volatile profile of dairy proteins. 2.6. Statistical Analysis One-way analysis of variance (ANOVA) was performed using Xlstat sensory software (version 2019.4.1) (Addinsoft, New York, NY, USA). All tests were performed at p=0.05. The heat map was generated using Euclidean distance and the complete linkage algorithm (90) implemented in the gplots package (version 3.0.1.1) (https://CRAN.Rproject.org/package=gplots) of R software (version 3.6.1) (http://www.r-project.org/). 3. Results and Discussion 3.1. The Impact of the Addition of Yeasts on Pea Fermentation by LAB Table 1summarizes the main parameters that characterize the fermentations performed by the different cultures: the bacterial and yeast biomass at the initial and/or final time (t0 and tf, respectively, where tf is the time necessary to reach pH 4.55), the residual sugar level at tf, and finally the lactic acid and ethanol concentration at tf. To evaluate the impact of the addition of yeasts on the acidification rate of VEGE, the parameter tf was analyzed. As shown in Table 1, the addition of yeasts to VEGE did not have a significant impact on the fermentation time. This is a positive result in the aim of producing commercial products. However, we can observe a difference in the tf between two yeast species: there was an increase in the time necessary to reach pH 4.55 for VEGE +KM compared to VEGE +TOR. Nevertheless, this cannot be directly linked to the final bacteria biomass since higher concentrations were found in VEGE +TOR. 118 Foods 2020,9, 349 Table 1. Fermentation characteristics of the different cultures. Bacteria Biomass (×108CFU/mL) Yeast Biomass (×107CFU/mL) Kinetic Parameters (g/L) at tf at t0 at tf * tf (h) Total Residual Sugar at tf Lactic Acid at tf Ethanol at tf VEGE047 2.9 ±0.2 b-- 13.1 ±0.5 ab 25.4 ±0.3 a3.5 ±0.03 bVEGE047 +K. marxianus 4.2 ±0.6 ab 1.4 ±0.1 b4.4 ±0.5 a14.9 ±0.9 a13.2 ±0.1 d3.0 ±0.05 d4.8 ±0.007 a VEGE047 +K . lactis 6.0 ±1.6 a3.4 ±0.6 a4.1 ±0.7 a13.1 ±0.4 ab 20.8 ±0.6 b3.6 ±0.03 a1.7 ±0.004 c VEGE047 +T. delbrueckii 5.4 ±0.3 a0.4 ±0.1 b0.7 ±0.1 a12.3 ±0.01 b16 ±0.3 c3.3 ±0.02 c4.2 ±0.02 b Each mean is based on three independent replicates. The values with letters of the same color were compared with each other. Values with the same letters are not significantly different (p>0.05). * tf is the time needed to reach pH 4.55. 119 Foods 2020,9, 349 On the other hand, differences in the final concentration of lactic acid were observed in the presence of yeasts compared to VEGE alone: a lower concentration of lactic acid was observed for VEGE +KM and VEGE +TOR. These differences could be explained by an acidifying metabolite produced by yeasts. In fact, these two co-cultures displayed a higher production of ethanol compared to VEGE alone (in which no production was identified) or in VEGE +KL. Ethanol production paralleled carbon dioxide generation, a gas that, by dissolution, acidifies the medium. Thus, the production of CO 2 could explain the lower concentration of lactic acid generated to reach pH 4.55 in the two former conditions. As for VEGE +KL, the low final ethanol concentration could explain that the lactic acid concentration is nearly the same in this co-culture compared to VEGE alone. Since Kluyveromyces lactis is an aerobic-respiring yeast, it could be limited by the oxygen availability in our static fermentation, leading to limited growth and ethanol production compared to Kluyveromyces marxianus [29,30]. Considering the sucrose metabolism, the total residual sugars were lower in the presence of yeasts, compared to VEGE. In fact, VEGE alone consumed 6 g/L of sucrose, and no residual monosaccharides (fructose or glucose) were detected (Supplementary Table S1). The yield of lactic acid/sucrose calculated for VEGE was approximately 0.8 g/g, which is close to a homofermentative yield. In fact, VEGE contains three homofermentative species (Lactobacillus acidophilus,Streptococcus thermophilus, and Lactobacillus delbrueckii subsp. bulgaricus). In co-cultures, nearly 100% of the initial sucrose (30 g/L) was hydrolyzed into glucose and fructose, and these monosaccharides were partially metabolized (Supplementary Table S1). The consumed sugars were 9 g/L for (VEGE+KL), 14 g/L for (VEGE +TOR), and 17 g/L for (VEGE +KM). These results are in agreement with the final concentration of ethanol obtained in each condition. As a conclusion, the addition of yeasts had a weak impact on the behavior of VEGE, although slight differences were identified due to probable negative and/or positive interactions between VEGE and the yeast used. 3.2. Modification of the Sensory Perception in the Presence of Yeasts In order to evaluate the impact of yeasts, a sensory analysis was performed only on the fermented products. It was previously shown that VEGE cultures improved the sensory perception of a pea protein yogurt-like product, but not enough for consumer acceptability (General Mills, personal communication). The panelists generated typical descriptors for a plant matrix: green flavor/vegetal, leguminous plant, bitter, astringent, nut, and other descriptors such as sparkling, tangy, sour, citrus fruit, beer/yeast, and cultured apple cider. Figure 1shows the characteristics of each fermented product as the average intensity of the individual panelists’ scores, and the detailed data are shown in Table 2. Considering the main defects detected in the pea matrix, the intensities of leguminous plant and green flavor/vegetal were significantly reduced in the presence of yeasts. There is a lack of information in the literature about the impact of yeasts on the improvement of the sensory characteristics of pea proteins. However, one recent study revealed a significant decrease in the beany odor using analytical methods after the fermentation of soybean residue, okara, by K. marxianus [31]. The global intensity of aroma in the presence of yeasts was significantly higher compared to VEGE culture. This increase could be linked to the presence of trigeminal sensations such as sparkling, tangy, and sour attributes, which were significantly higher in the samples with yeasts. The sparkling attribute could be directly linked to the presence of CO 2 [ 32 ]. Considering the sour attribute, the yeasts can produce pyruvic and acetic acid, which could explain this perception [ 33 ]. In addition, it was previously shown that high levels of carbonation significantly enhanced the sourness and astringency in flavored milk beverages [34]. Finally, the fermented products obtained with yeasts were characterized by a significantly higher “beer/yeast” attribute than the products obtained with VEGE. In fermented beverages such as beer, esters are the most important set of yeast-derived aroma-active compounds, and they are responsible for their fruity character [35]. 120 Foods 2020,9, 349 The modifications of the sensory perception in the presence of yeasts could be attributed to a reduction in the concentration of pea off-notes or the generation of new notes that could modify the perception of sensory defects. Thus, to obtain a better understanding of the sensory modifications, the volatile profiles were analyzed. Table 2. Average intensity of sensory attributes determined for different fermented products using a scale ranging from 0 to 15. VEGE VEGE +K. marxianus VEGE +K. lactis VEGE +T. delbrueckii Global intensity 4.660 b 7.033 a 6.390 a 6.900 a Sour 2.257 b 5.000 a 4.250 a 5.400 a Bitter 1.837 a 3.200 a 2.840 a 2.867 a Astringent 3.730 a 4.733 a 4.243 a 4.933 a Tangy 0.417 c 5.267 a 3.500 b 5.200 a Sparkling 0.050 d 7.467 b 4.817 c 9.367 a Green flavor/Vegetal 2.637 a 0.733 b 1.533 ab 0.867 b Leguminous plant 3.723 a 1.147 b 1.500 b 0.800 b Citrus fruit 0.400 a 0.300 a 0.850 a 0.200 a Nut 0.267 a 0.167 a 0.333 a 0.200 a Beer/yeast 0.183 c 5.000 a 3.150 b 2.800 b Sourdough 0.243 b 1.260 ab 1.380 ab 1.800 a Cultured apple cider 0.200 a 1.340 a 1.067 a 1.067 a Mean values in the same row that are not followed by the same letter are significantly different (p<0.05). Figure 1. Aroma profile analyses of the four fermented pea protein isolate products. Data are displayed as mean numerical values of the sensory evaluations. 3.3. Characterization of Volatile Compounds Identified Using GC–MS Analysis To obtain a better insight into the impact of fermentation on the volatile profile, GC–MS analyses were performed on the uninoculated and fermented samples (VEGE, VEGE +KM, VEGE +KL and VEGE +TOR). 121 Foods 2020,9, 349 3.3.1. Volatile profiles of Uninoculated and Fermented Samples A total of 87 volatile molecules were detected. These compounds were grouped into five families, including aldehydes, alcohols, ketones, furans, and esters. A heat map (Figure 2) was drawn up using the proportions of each molecule (proportions calculated using the surface areas of the peaks) among the samples (Supplementary Table S2). Figure 2. A hierarchically clustered heat map showing the patterns of the different samples for the identified volatile compounds. First of all, it was observed that the uninoculated sample contained most of the aldehyde, ketone, and furan compounds, which are major families of molecules responsible for pea off-flavor. Upon fermentation with VEGE or VEGE +yeasts, the volatile profiles were strikingly modified compared to 122 Foods 2020,9, 349 the uninoculated sample. As shown in Figure 2, most of the aldehyde, furan, and ketone molecules were degraded. Moreover, the esters were only present in the fermented products with VEGE and yeasts. The concentrations of alcohol were also generally higher in these products compared to VEGE. The distribution of the proportions of the molecules revealed the presence of two main groups for the fermented cluster: VEGE and VEGE +TOR vs. VEGE +KL and VEGE +KM. These two clusters were not in agreement with the sensory results, which revealed differences in the sensory perception of products fermented with VEGE or VEGE and yeasts. In fact, there is not a direct link between the volatile compounds identified by GC–MS and the sensory descriptors. Consequently, to acquire a better understanding of the sensory perceptions, we focused on the off-flavor molecules. 3.3.2. Degradation of Off-Flavor Molecules in the Fermented Samples Twenty molecules among the 87 aroma compounds detected were reported in the literature as beingresponsiblefor the off-flavor perceptionin peas [ 9 , 36 , 37 ]. Theconcentrations of these 20 molecules are presented in Table 3. The perception thresholds were also reported for information purposes because they were determined by orthonasal olfaction in water [38]. First, it should be emphasized that most of the molecules responsible for the pea off-flavors were present in the uninoculated sample in which some of them were detected at high concentrations, including hexanal, butanal, 2-pentylfuran, and 2-ethylfuran. Hexanal has already been reported to be the major molecule responsible for the “green” and “herbal” perception in pea protein isolates [ 39 ]. Moreover, hexanal, as well as heptanal and nonanal, were identified as the main volatile compounds of soymilk flavor [40]. In the four fermented samples, there was a significant elimination of off-flavor molecules, mainly aldehydes, ketones, and furans. Previous studies showed that Lactobacilli and Streptococci could eliminate compounds related to the beany flavor of soymilk during fermentation, such as hexanal and 2-pentylfuran [ 40 , 41 ]. Moreover, lactic acid fermentation with L. plantarum or with P. pentosaceus had the potential to decrease the concentration of hexanal in lupin protein extracts [22]. Other compounds such as (E)-2-heptenal, 6-methyl-5-hepten-2-one, and trans-2-methyl-2-butenal were present in the products fermented with VEGE, but were not detectable in those fermented with VEGE and yeasts. These differences might explain the fact that the products fermented with VEGE were perceived as being greener and more leguminous by the panelists compared to the products fermented with VEGE and yeasts. This hypothesis should be supported by the determination of the odor perception thresholds by retronasal olfaction in pea. In the fermented products with VEGE or with VEGE and yeasts, the incomplete reduction or the increase in off-flavor concentrations were notable. For example, 2-pentyl-furan and 2-ethyl-furan were reduced but remained at high levels in all the products. As for 2-methylpropanal, 1-hexanol, and 1-octanol, the concentrations increased when yeasts were added. These compounds could be the result of yeast metabolism under anaerobic conditions. Previous studies showed that the level of 2-methylpropanal increased in fermented pea gels using a microbial consortium, including yeasts [ 42 ]. This compound, which contributes to malty and chocolate-like notes, is the result of valine degradation by yeasts through the Ehrlich degradation pathway [ 43 ]. As for the two other alcohols, they could have been produced by the reduction of hexanal and octanal through the action of alcohol dehydrogenase activities. It was previously suggested that fermentation of soybean residue by K. lactis drastically decreased the amount of hexanal to trace levels, with a corresponding increase in hexanoic acid and/or hexanol [ 25 – 44 ]. The products with yeasts were perceived as being less “leguminous plant and green flavor/vegetal” than the products with VEGE; it is possible that the concentrations of 2-methylpropanal, 1-hexanol, and 1-octanol could be under their perception threshold. As previously mentioned, it will be necessary to determine the odor threshold of the molecules responsible for the sensory defects in peas. 123 Foods 2020,9, 349 Table 3. Concentrations of the off-flavor molecules in the non-fermented and fermented samples (μg/L). Volatile Compounds Descriptors Uninoculated VEGE VEGE + K. marxianus VEGE + K. lactis VEGE + T. delbrueckii Detection Threshold a 2-pentylfuran Musty/earthy, mushroom, floral, buttery, rancid, green 304.3 46.4 56.7 148 55 6 Hexanal Green, grass 181.3 <DL <DL <DL <DL 4.5 2-ethylfuran Beany, earthy, malty, sweet 77.7 28.1 29.7 56.7 32.8 Butanal Pungent, green, malty, chocolate, cocoa 54.9 <DL <DL <DL <DL 1-penten-3-ol Green, vegetable, fruity 19.6 3.7 11 28.1 5.6 (E)-2-octenal Green, cucumber, musty/earthy, waxy, fatty, grass, banana, sweet 10.3 <DL <DL <DL <DL 3 Nonanal Aldehydic, fatty, green, geranium, floral, soapy, citrus, waxy 8.1 <DL <DL <DL <DL 1 Heptanal Green, fresh, fatty 5.7 <DL <DL <DL <DL 3 2-nonanone Green, earthy, grassy, fruity, sweet, 5.0 <DL <DL <DL 1.2 2-methylpropanal Aldehylic, grass, green, floral 4.0 0.7 40.9 70 39.6 6 1-octen-3-ol Mushroom, earthy, burnt, green, vegetable, stale 3.9 2.2 2.9 6.7 2.9 1 (E)-2-heptenal Pungent green, fatty 2.7 0.8 <DL <DL <DL 13 2-octanone Green, floral, soapy, fruity, fatty 2.5 0.7 <DL 0.9 0.8 Octanal Aldehylic, green, soapy, citrus-like, sweet, waxy, fruity 2.5 <DL <DL <DL <DL 0.7 (E)-2-Hexenal Tea-like, green grass, almond, cherry, juicy, rancid 2.4 <DL <DL <DL <DL 17 Decanal Fresh, marine, aldehydic, iodized, soapy, grapefruit, bitter, sweet 1.04 <DL <DL <DL <DL 0.1 6-methyl-5-hepten-2-one Nutty, moldy, green, vegetable, citrus 0.5 0.5 <DL <DL <DL 2 Trans-2-methyl-2-butenal Strong green-type odor and a fruity flavor 0.3 0.1 <DL <DL <DL 1-hexanol Green, musty/earthy, peanut hull, chemical-like, fruity, grassy <DL 126.5 175.4 500.1 116.7 500 1-octanol Moss, mushroom, green, vegetable, fatty, waxy, citrus, floral <DL <DL 4.4 <DL 2.9 Concentrations are classified as per decreasing concentrations in the initial matrix. aDetermined in water by orthonasal olfaction [38]. <DL: value inferior to the detection limit. 124 foods Article A Comprehensive Characterisation of Volatile and Fatty Acid Profiles of Legume Seeds Prit Khrisanapant 1,2, Biniam Kebede 1, Sze Ying Leong 1,2 and Indrawati Oey 1,2,* 1Department of Food Science, University of Otago, PO Box 56, Dunedin 9054, New Zealand; [email protected] (P.K.); [email protected] (B.K.); [email protected] (S.Y.L.) 2Riddet Institute, Private Bag 11 222, Palmerston North 4442, New Zealand *Correspondence: [email protected]; Tel.: +64-3-479-8735 Received: 30 October 2019; Accepted: 4 December 2019; Published: 6 December 2019 Abstract: Legumes are rich in unsaturated fatty acids, which make them susceptible to (non) enzymatic oxidations leading to undesirable odour formation. This study aimed to characterise the volatile and fatty acid profiles of eleven types of legumes using headspace solid-phase microextraction gas chromatography–mass spectrometry (HS-SPME-GC-MS) and GC coupled with a flame ionisation detector (GC-FID), respectively. Volatile aldehydes, alcohols, ketones, esters, terpenes and hydrocarbons were the chemical groups identified across all the legumes. The lipids comprised palmitic, stearic, oleic, linoleic and α -linolenic acids, with unsaturated fatty acids comprising at least 66.1% to 85.3% of the total lipids for the legumes studied. Multivariate data analysis was used to compare volatile and fatty acid profiles between legumes, which allow discriminant compounds pertinent to specific legumes to be identified. Results showed that soybean, chickpea and lentil had distinct volatileand fattyacid profiles, withdiscriminating volatilesincluding lactone, ester andketone, respectively. While all three Phaseolus cultivars shared similar volatile profiles, 3-methyl-1-butanol was found to be the only volatile differentiating them against the other eight legumes. Overall, this is the first time a multivariate data analysis has been used to characterise the volatile and fatty acid profiles across different legume seeds, while also identifying discriminating compounds specific for certain legume species. Such information can contribute to the creation of legume-based ingredients with specific volatile characteristics while reducing undesirable odours, or potentially inform relevant breeding programs. Keywords: legumes; volatiles; fatty acids; characterisation; fingerprinting; multivariate data analysis 1. Introduction The seeds of legume plants (usually referred to as ‘legumes’) are a nutritious source of proteins, carbohydrates, lipids, vitamins and minerals [ 1 ]. Unprocessed legumes have a distinct odour due to their inherent plant metabolism [ 2 ]. Regrettably, legumes are not widely utilised to their full potential due to various factors, such as low protein digestibility, their hard-to-cook-effect and undesirable odours. The undesirable odours are largely influenced by lipoxygenase-catalysed unsaturated fatty acid oxidation that occurs during several postharvest processes [ 3 , 4 ]. Trained sensory panels have used negative descriptors such as beany, musty, haylike, grassy and green to describe the odours of legumes such as soybean and peas. It has been recognised that these odours are majorly derived from volatile compounds such as hexanal and 1-octen-3-ol [ 3 – 5 ]. Previous studies [ 3 , 6 – 9 ] have reported volatile compounds and their formation mainly in soybean, and there is still a limited understanding on the volatile profile of other commercially relevant legumes such as cowpea, lentil, common bean and pea [ 10 – 13 ]. However, these studies used targeted analysis without providing a holistic picture of all low molecular weight compounds present in the legume’s volatile fraction. Advancements in volatile Foods 2019,8, 651; doi:10.3390/foods8120651 www.mdpi.com/journal/foods 131 Foods 2019,8, 651 analysis and instrument sensitivity are able to capture a wider range of volatile compounds; hence, they enable researchers to take an untargeted fingerprinting approach. By definition, fingerprinting is an untargeted analytical approach aiming to detect as many compounds as possible in a particular food fraction [ 14 ]. This untargeted fingerprinting approach has been used as a tool to study the effect of cooking on legumes [ 10 , 15 ]. To date, there is still a gap in the detailed, holistic characterisation of legume volatiles using this approach, and hence, it is essential to examine whether this analytical approach could differentiate different types of legumes. As previously mentioned, undesirable legume volatiles are primarily formed from fatty acid oxidation. In spite of this connection, no previous investigation has examined fatty acid composition of legumes in conjunction with their volatile profile. There is an opportunity here to elucidate the connection of fatty acid composition and volatile profile in the whole bean matrix. Therefore, the objective of the present study was to comprehensively characterise the volatile and fatty acid profiles in commercially relevant legumes. The headspace solid phase-gas chromatography–mass spectrometry (HS-SPME-GC-MS) approach was implemented to detect volatile compounds, whereas the fatty acid was profiled using gas chromatography flame ionisation detection (GC-FID). The novelty of this study lies on the fact that (i) the fingerprinting approach is being applied for the first time on a wide range of commercial relevant legumes, and (ii) volatile and fatty acid components are analysed in an integrated fashion using multivariate data analysis. 2. Materials and Methods 2.1. Raw Material Handling and Storage Eleven types of dry legume seeds, namely, soybean (Glycine max), pea (Pisum sativum), chickpea (Cicer arretium var. kabuli), orange lentil (Lens culinaris), mung bean (Vigna radiata), fava bean (Vicia faba), cowpea (Vigna unguiculata), adzuki bean (Vigna angularis), kidney bean (Phaseolus vulgaris), navy bean (Phaseolus vulgaris) and black bean (Phaseolus vulgaris), were purchased in a single batch from the local market in Dunedin (New Zealand). Seeds with physical damages and discolouration were discarded. The remaining seeds were vacuum packed in opaque aluminium bags and stored at 4 ◦ C until analysis. 2.2. Sample Preparation Legume seeds (30 g) were ground using a laboratory blender (Waring, Auckland, New Zealand) for 60 s, with a pause every 30 s, at room temperature (20 ± 2 ◦ C). The resulting flour was sieved to pass through an 850 μm mesh. Flour retaining between 450 and 850 μm mesh size was used for lipid analysis for consistent extraction of lipids. For volatile profiling, the grounding time was increased to 180 s to maximise flour surface area. 2.3. Moisture Determination The legume flour sample (0.2 g) was weighed and transferred into a glass petri dish (Steriplan, Kimax, Auckland, New Zealand). The dish was subsequently covered with perforated aluminium foil. The drying was carried out at 130 ◦ C in a convection oven for 16 h (Qualtex, Andrew Thom, Sydney, Australia). The samples were removed and cooled in a desiccator lined at the bottom with silica beads. The percentage of moisture content was estimated based on the weight loss after drying and cooling. The moisture content determination was conducted in five independent replicates. 2.4. Headspace Volatile Analysis with HS-SPME-GC-MS Fingerprinting Headspace solid phase micro-extraction gas chromatography mass spectrometry (HS-SPME-GC-MS) was conducted according to the work of Liu and others [ 16 ] with modifications, consisting of sample preparation, incubation, volatile extraction, injection and GC-MS analysis. Prior 132 Foods 2019,8, 651 to analysis, method parameters were optimised in order to capture a wide range of volatile compounds. The optimisation included sample weight, sample dilution, sample to salt ratio and type of GC column. Upon sample analysis, legume flour was weighed (2.5 g) into a 20 mL glass vial, and 5 mL of saturated sodium chloride solution (360 g/L) was added to increase the solution’s ionic strength and drive the legume volatiles into the headspace. The vial was then tightly sealed with PTFE-coated silicon septa screw cap (Supelco, Sigma-Aldrich, St. Louis, MO, USA). The sealed vials were then vortexed for 30 s. Using the Gerstel MPS Maestro autosampler (Gerstel, Linthicum Heights, MD, USA), each sample was incubated at 40 ◦ C for 5 min, with agitation at 250 rpm. Thereafter, headspace volatile compounds were extracted using headspace-solid phase microextraction (HS-SPME). A preconditioned (according to the manufacturer’s instructions) SPME fibber with a 30/50 μ m divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) sorptive coating (Stableflex, Supelco, Bellefonte, PA, USA) was used to extract a wide range of volatile compounds from the headspace of the vial for 30 min at 40 ◦C. For the GC-MS analysis (Agilent 6890N, Agilent Technologies, Santa Clara, CA, USA), extracted volatiles were desorbed in the injection port at 230 ◦ C for 2 min, then injected in splitless mode onto a ZB-Wax capillary column (30 m × 0.25 mm × 0.25 μ m; Agilent Technologies, Santa Clara, CA, USA) for separation with helium as the carrier gas at 1.5 mL/min. To facilitate separation and elution of the injected headspace volatile compounds, the GC oven was maintained at 50 ◦ C for 5 min before the temperature was ramped up to 210 ◦ Cat5 ◦ C/min, after which it was again ramped to 240 ◦ C at the rate of 10 ◦ C/min, for a total GC-MS run time of 37 min. For the MS, the quadrupole was set at 70 eV, and the ion sources were 150 ◦ C and 230 ◦ C, respectively, with a mass-to-charge ratio scanning range of 30–300 m/z. Thereafter, the SPME fibre was regenerated according to the manufacturer’s instruction. The same SPME fibre was used across all samples. The volatile profiling for each legume seed was conducted in five independent replicates. With regard to preprocessing of GC-MS chromatograms, volatile fingerprinting chromatograms often contain co-eluting peaks, which can confound data analysis. Therefore, an automated mass deconvolution and identification system (AMDIS; version 2.72, build 140.24, Agilent Technologies, Santa Clara, CA, USA) was used to deconvolute potential co-eluting peaks. The spectra obtained were further processed by mass profiler professional (MPP; version 14.9.1, build 1316, Agilent Technologies, Santa Clara, CA, USA), a peak filtering and alignment software. This creates aligned peaks lacking nonreproducible and background peaks. Afterwards, a table of retention time and volatile amount expressed as peak area was obtained. Tentative identification of volatile compounds was conducted manually. In the present work, three criteria were employed to increase the power of compound identification: (i) match and reverse match with the NIST library of no less than 90%; (ii) comparison of experimental retention index with RI according to literature; and (iii) matching retention time and spectra with authentic standards from chemical groups of detected volatiles (alcohol, aldehyde, terpene and acid) (See Table S1). 2.5. Determination of Fatty Acids in Legume Seeds Using FAME-GC-FID For fatty acid analysis, legume lipid was extracted and converted to fatty acid methyl esters (FAME) and detected using gas chromatography flame ionisation detection (GC-FID) according to AOAC method 963.22 [17] with modifications. 2.5.1. Total Lipid Extraction Based on Soxhlet Method Legume flour sample (2 g) was weighed and placed inside a cellulose extraction thimble (26 × 60 mm, Whatman, Buckinghamshire, UK). The filled extraction thimble was fitted onto a Soxtec distillation apparatus (Tecator, Hilleroed, Denmark). Meanwhile, aluminium cups were filled with five to ten antibumping granules and weighed before adding with 25 mL of organic solvent mix consisting of a 2:1 (v/v) ratio of chloroform (EMPARTA, Merck, Darmstadt, Germany) and methanol (Ajax Univar, 133 Foods 2019,8, 651 North Shore, New Zealand). The cups were then fitted underneath the thimble and above the heating plate of the Soxtec distillation apparatus. Continuous reflux distillation of the samples was then carried out for 1 h with the heating plate set at 160 ◦ C. After that, solvent was evaporated from the sample, and then the cups were released from the apparatus, and the residual solvent was allowed to evaporate inside a convection oven (Sanyo MOV-212F, New South Wales, Australia) set at 50 ◦C for 15 min. The lipid yield of each sample was estimated by weight difference between the weight of the empty cups (filled with antibumping granules) and the weight of the cup filled with lipids upon completion of the solvent extraction. After weighing, 15 mL of hexane (Ajax Finechem, Auckland, New Zealand) was used to resuspend the lipids in each aluminium cup. The hexane containing lipid solution was then stored in a refrigerator at 4 ◦ C in a tightly sealed glass tube before proceeding to the lipid purification step. The lipid extraction of each legume seed was conducted in four independent replicates. 2.5.2. Lipid Purification Fatty acid methyl esters (FAMEs) were obtained by purifying lipids by saponification to remove nonsaponifiable materials, followed by esterification into FAMEs. A volume of lipid containing hexane solution equivalent to 5 mg lipid was pipetted into a sealable glass tube. Thereafter, 5 mL of a solution containing potassium hydroxide (0.5 M, AnalaR, Leuven, Belgium) dissolved in methanol was added, and the tube was sealed immediately. The fatty acid saponification was carried out for 20 min at 80 ◦ C on a heating block. The tubes were removed from the heating block and allowed to cool in ambient air for 10 min. To the cooled solution, 3 mL of diethyl ether (LabServ, Auckland, New Zealand) and 7 mL of milliQ water were added, and the test tube was inverted to mix. The mixture was allowed to stand for 2 min to allow separation between water and organic solvent layers. The top layer of diethyl ether was then discarded to remove any nonsaponifiable material. The fatty acids were liberated by neutralisation to ~pH 7 with concentrated hydrochloric acid (37%; EMSURE, Merck, Darmstadt, Germany). Then, another 4 mL of diethyl ether was added and inverted to mix. The top diethyl ether layer formed was collected in a clean glass test tube for derivatisation. 2.5.3. Lipid Derivatisation to Fatty Acid Methyl Esters (FAMEs) One millilitre of boron trifluoride (14%) in methanol (Sigma-Aldrich, St. Louis, MO, USA) was promptly added as derivatisation agent. Fatty acid esterification was carried out for 20 min at 80 ◦ C on a heating block and then cooled. After cooling, 7 mL of saturated sodium chloride solution (360 g/L) was added and vortexed for 15 s. The top diethyl ether formed was then collected for FAME GC-FID analysis. 2.5.4. Fatty Acid Profiling Using GC-FID Forfattyacidanalysis, gaschromatographyflameionisationdetection(GC-FID)wasconducted[ 17 ] with modifications. A GC-FID system (6890A G1530A; Agilent Technologies, Santa Clara, CA, USA) was used. It was equipped with an autosampler (7683 series injector, Agilent Technologies, Santa Clara, CA, USA) and fitted with a BPX70 capillary column (70% Cyanopropyl Polysilphenylene-siloxane, SGE, Victoria, Australia). Samples (1 μ L) were injected in split mode (20:1 ratio) at 240 ◦ C for separation with hydrogen gas at 2.2 mL/min. To ensure good separation of fatty acid methyl esters, the GC oven temperature was increased from its initial temperature of 120 ◦ C to 225 ◦ C at the rate of 3 ◦ C/min, then ramped to 245 ◦ C at 10 ◦ C/min. Once the GC oven temperature reached 245 ◦ C, the column was held at this temperature for another 2 min. For the FID, the detector temperature was set at 250 ◦C. 2.5.5. Identification and Data Preprocessing of FAME Chromatograms obtained from GC-FID were analysed with GC ChemStation (Build 4.01, Agilent Technologies, Santa Clara, CA, USA) and individual peaks manually identified by matching retention 134 Foods 2019,8, 651 time with commercial standards (FAMQ-005, AccuStandards, New Haven, CT, USA). Following manual peak alignment and removal of interfering background compounds, the proportion of signal abundance of each fatty acid was calculated in % abundance of total signal abundance. Thereafter, a table of fatty acid profile for each legume was obtained. 2.6. Multivariate Data Analysis and Identification of Compounds Relevant to Specific Legume Type Multivariate data analysis, marker selection and marker identification were performed on the combined data sets comprising legume fatty acid and headspace volatile data sets. Using both volatile and fatty acid data, multivariate data analysis was conducted using principle component analysis (PCA), followed by partial least square discriminant analysis (PLS-DA), utilising Solo software (Version 8.2.1, Eigenvector Research, Manson, WA, USA). Firstly, PCA was used as an unsupervised, exploratory technique to determine grouping/separation in the data, as well as to detect outlier. Secondly, PLS-DA was used as a supervised technique to detect similarities and differences between different legume seeds, as well as correlation between volatile compounds and legumes. Thereafter, a bi-plot was generated as a visual representation of the information obtained (OriginPro, OriginLab, Northampton, MA, USA). Volatile compounds that showed a clear discriminant correlation with each legume were selected through determination of variable identification (VID) coefficients [ 18 ]. VID values are the corresponding correlation coefficients between X-variables (volatile compounds and fatty acids) and predicted Y-variables (Legume type). An absolute threshold value of |0.800|was selected. Therefore, volatiles with an absolute VID coefficient higher than 0.800 were plotted as bar graphs, and statistical significance between the means was determined using analysis of variance, conducted through SPSS Statistics (IBM, Version 26), followed by Tukey’s post-hoc test (p<0.05). Those compounds were considered important discriminant compounds associated with each legume seed. Discriminant volatile compounds were identified by comparing the deconvoluted mass spectra with an established mass spectra library using National Institute of Standards and Technology (NIST) Mass Spectral Search Program (Version 2.2, build June 10, 2014). The identities were also rechecked with a minimal 90% match and reverse match on NIST, as well as comparison of retention index with literature. 3. Results 3.1. Moisture Content of Legume Seeds The moisture content of legume seeds was within the range of 7.4% (soybean) to 14.9% (kidney bean). Pea, navy bean, orange lentil and chickpea contained 9.9%, 9.6%, 8.7% and 8.4% moisture, respectively. In comparison, adzuki bean, black bean, fava bean, mung bean and cowpea were slightly moister, at 12.4%, 11.3%, 11.0%, 10.3% and 10.3%, respectively. 3.2. Fatty Acid Analysis of Legume Seeds GC-FID analysis of the fatty acid methyl esters was able to detect five clearly separated peaks in eleven legume samples. The legume lipid fractions consisted of palmitic (C16:0), stearic (C18:0), oleic (C18:1), linoleic (C18:2) and α -linolenic (C18:3) acids (Table 1). A commonality of the lipid profiles is that the level of saturated fatty acids such as palmitic and stearic acids is low, with the ratio of saturated to unsaturated fatty acids at 1:2 in cowpea and mung bean, and up to 1:5.4 and 1:6 in soybean and chickpea. On the other hand, all legumes had a high level of essential polyunsaturated fatty acids, namely, linoleic and α -linolenic, ranging from 47.3% in orange lentil to 71.0% in black bean. Out of the eleven types of legumes, four of them, i.e., chickpea, orange lentil, pea and fava bean, had a high (>20%) proportion of oleic acid, a monounsaturated fatty acid. This affirms that legumes are a good source of unsaturated fatty acids [ 19 ], with their fatty acid profile favourable from a cardioprotective perspective [20]. 135 Foods 2019,8, 651 Table 1. Relative fatty acid abundance of 11 types of legume seeds, as analysed by fatty acid methyl ester gas chromatography coupled with a flame ionisation detector (FAME-GC-FID). Legumes Total Lipid Extracted (g/100g Sample) C16:0 (g/100g Lipid) C18:0 (g/100g Lipid) C18:1 (g/100g Lipid) C18:2n-6 (g/100g Lipid) C18:3n-3 (g/100g Lipid) SFAs (g/100g Lipid) MUFAs (g/100g Lipid) PUFAs (g/100g Lipid) n-6/n-3 Ratio Soybean 19.20 d±1.98 11.76 a±0.13 3.68 c,d ±0.01 19.20 c±0.28 55.15 g±0.15 8.88 b±0.02 15.44 a,b ±0.13 19.20 c±0.28 64.03 c,d ±0.14 6.21 c±0.02 Chickpea 7.73 c±0.73 10.94 a±0.20 1.80 a±0.0.16 37.87 e±0.16 45.78 f±0.42 2.33 a±0.05 12.74 a±0.35 37.87 e±0.16 48.11 a±0.46 19.67 e±0.36 Lentil 3.90 a,b ±0.21 21.40 e±0.41 2.77 a,b,c ±0.10 28.06 d±0.35 38.21 c,d ±0.55 9.07 b±0.09 24.29 e±0.40 28.06 d±0.35 47.27 a±0.64 4.21 b,c ±0.03 Cowpea 3.46 a,b ±0.10 27.68 f±0.26 4.76 e±0.53 7.35 a,b ±1.64 35.97 b,c ±0.85 23.34 e±0.48 33.34 f±0.97 7.35 a,b ±1.64 59.31 b,c ±1.05 1.54 a±0.04 Pea 3.41 a,b ±0.12 13.48 a,b ±0.35 4.50d e±0.11 34.40 e±2.30 38.66 c,d ±2.38 8.78 b±0.67 17.98 b,c ±0.44 34.40 e±2.30 47.44 a±3.04 4.41 b,c ±0.07 Mung bean 3.20 a,b ±0.15 27.04 f±1.56 5.73 f±0.12 6.54 a±37.78 43.71 e,f ±2.23 15.82 c±1.17 33.75 f±2.22 6.54 a±3.78 59.53 b,c,d ±3.18 2.76 a,b ±0.14 Fava bean 2.75 a,b ±0.19 15.25b c±0.39 3.64 c,d ±0.93 24.57 c,d ±0.34 52.68 g±0.88 3.61 a±0.57 19.14 c,d ±1.02 24.57 c,d ±0.34 56.29 b±1.17 14.59 d±2.06 Navy bean 3.87 b±0.17 18.00 d±0.22 3.12 bc ±0.13 19.49 c±0.81 28.30 a±0.45 31.09 f±0.33 21.12 d±0.15 19.49 c±0.81 59.39 b,c,d ±0.73 0.91 a±0.01 Kidney bean 3.59 a,b ±0.14 17.97 d±0.49 2.40 a,b ±0.27 12.71 b±2.00 29.02 a±1.16 35.71 g±1.39 20.49 c,d ±0.65 12.71 b±2.00 64.74 d±2.55 0.81 a±0.00 Black bean 3.20 a,b ±0.16 17.17c d±0.48 2.39 a,b ±0.29 8.29 a,b ±0.16 33.65 b±0.34 37.30 g±1.22 19.56 c,d ±0.71 8.29 a,b ±0.16 70.95 e±1.50 0.90 a±0.02 Adzuki bean 1.96 a±0.10 27.96 f±1.02 3.36 c±0.40 3.92 a±0.52 41.88 d,e ±1.12 20.31 d±0.73 31.35 f±1.45 3.92 a±0.52 62.19 c,d ±1.96 2.06 a±0.01 F-value 170.887 229.2 33.7 93.3 114.7 1047.5 148.5 93.3 46.0 218.0 Significant 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Values expressed as mean ± standard deviation (n=4). C16:0 =palmitic acid; C18:0 =stearic acid; C18:1 =oleic acid; C18:2 =linoleic acid; C18:3 = α -linolenic acid; SFAs =saturated fatty acids; MUFAs =monounsaturated fatty acids; PUFAs =polyunsaturated fatty acid; n-6/n-3 =ratio of omega-6 to omega-3 fatty acids. Means with different superscripts in the same column indicate significant difference (p<0.05). 136 Foods 2019,8, 651 3.3. Volatile Analysis of Legume Seeds The HS-SPME-GC-MS fingerprinting method was able to detect an increased number of volatile compounds, totalling 97 different volatiles across all 11 legumes. Visually, the total ion chromatograms appear to be different based on the number and intensity of the peaks present amongst the samples. Some representative total ion chromatograms of the samples are shown in Figure 1. The chemical classes of detected volatile compounds consisted of alcohols, aldehydes, ketones, esters, lactones, terpenes, hydrocarbons, furans, pyrroles and sulphur-containing compounds. Note that the percentage of specific compounds mentioned in this section refers to their relative abundance, not absolute concentration. It is also important to note that soybean and chickpea had volatiles with the highest total peak area compared to other legumes. In soybean, a total of 63 volatile compounds were detected with the headspace fingerprinting method, consisting mainlyof aldehydes, ketones, alcohols, esters, furans andsulphur and hydrocarbons. Hexanal (40.9%), 1-octen-3-ol (21.1%) and 1-hexanol (5.9%) were volatiles with the highest abundance. A total of 76 compounds were detected in chickpea with aldehydes and alcohols as two dominant chemical classes, consisting of hexanal (56.3%), nonanal (8.7%) and 1-hexanol (4.4%). Acid, ketones, furans, esters and terpenes were also present, with hexanoic acid (1.7%) being the most abundant. In cowpea, 80 headspace volatiles were detected in this study. Hexanal (22.6%), 4,1 methylethyl benzaldehyde (22.0%) and 1-hexanol (10.4%) were prominent. Other aldehydes, terpene and benzene compounds were also detected in cowpea. A total of 65 compounds were detected in pea. Alcohols and aldehydes are the majority of volatile detected in pea, such as hexanal (42.5%), 1-penten-3-ol (9.0%), 1-hexanol (7.7%) and nonanal (6.0%). In orange lentil, 82 volatile compounds were detected. Hexanal (32.7%), 1-hexanol (16.4%), 2-hexenal (5.6%) and o-cymene (4.6%) make up the top four most abundant volatile compounds, accounting for 59.3% of total volatile detected. Uniquely, lentil had the largest number (13) of terpenes of all the samples. Headspace volatiles detected in mung bean (63) comprised aldehydes, alcohols, terpenes, ketones and sulphur compounds. The top three volatiles consisted of hexanal (38.8%), 1-hexanol (13.9%) and 1, 3-dimethyl-benzene (6.3%). Interestingly, mung bean had the highest abundance of xylene (5.7%) detected out of all samples. Fava bean had the fewest (55) volatiles detected in the headspace. Aldehydes and alcohols are the main volatile detected, with terpenes, terpene derivatives, furans and esters comprising a minor fraction. Hexanal (40.4%), 3-methylbutanol (19.1%) and 3-methylbutanoic acid (10.3%) were the top three volatile compounds detected in the fava bean headspace fraction. Major volatiles detected in adzuki bean (67) headspace fraction consisted of aldehyde, alcohol and, interestingly, furans, including hexanal (21.1%), 3-furaldehyde (19.5%), 3-furanmethanol (7.5%) and 1-hexanol (6.0%). Black, navy and kidney beans had 68, 76 and 72 volatile compounds detected in the headspace, respectively, with the most abundant fraction being aldehyde and alcohol, followed by terpenes, acids and ketones. Hexanal, 1-hexanol, 1-penten-3-ol and 3-methylbutanol are similarly the major volatile compounds detected in the three Phaseolus samples. 3.4. Comparison of the Volatile and Fatty Compositions Among the Eleven Legumes and Identifying Discriminant Compounds Multivariate data analysis (MVDA), which is an advanced chemometrics technique, was used to compare the volatile and fatty acid profiles among the 11 legume samples and identify discriminating compounds. In order to investigate the interdependence and relation among the measured attributes, the volatile and fatty acid data were merged into a single data matrix and analysed with MVDA. A principle component analysis (PCA) was first used as an unsupervised exploratory technique to detect groupings, separations or outliers within the volatile and sample data. From the PCA modelling (results not shown), it was able to be determined that there is indeed some distinct grouping and separation within the samples and that there were no outliers. 137 Foods 2019,8, 651 Figure 1. Representative total ion chromatograms of soybean ( A ), chickpea ( B ), orange lentil ( C ) and black bean ( D ) obtained with the headspace solid-phase microextraction gas chromatography–mass spectrometry (HS-SPME-GC-MS) fingerprinting method. 138 Foods 2019,8, 651 Thereafter, a partial least squares discriminant analysis (PLS-DA) model was constructed using the volatile and fatty acid profiles as X-variables and the 11 types of legumes as categorical Y-variables. A bi-plot constructed using the first two latent variables (LVs) is shown in Figure 2. On the bi-plot, samples that are close to each other are considered similar, whereas samples that are further apart are considered different [ 18 ]. Figure 2clearly shows that soybean and chickpea are projected further away from other samples in their own quadrant, indicating a large difference compared to the other legumes. The third quadrant is shared by cowpea and lentil, again indicating differences from other legumes. The fourth quadrant is occupied by pea, mung bean, fava bean, adzuki bean and all three Phaseolus beans, indicating similarity between the samples, especially between kidney, navy and black beans. This similarity may be attributed to them belonging to the same species. Figure 2. A bi-plot based on partial least square discriminant analysis (PLS-DA) comparing the volatile and fatty acid profiles among the 11 types of legumes. The variance explained is (X =22%, Y =10%) and (X =17%, Y =10%) for the first and second latent variable, respectively. =Adzuki bean (ADZ) A. =Chickpea (CHI) =Black bean (BLA) =Navy bean (NAV) =Kidney bean (KID) =Fava bean (FAV) =Mung bean (MUN) =Pea (PEA) =Soybean (SOY) =Cowpea (COW) =Orange lentil (LEN). In addition to the legume samples, unfilled circles on the bi-plot represent volatile and fatty acid compounds (X-variables). The location of each circle represents its relation to other measured attributes (X-variables) or samples (Y-variables). Hence, a PLS-DA bi-plot provides a graphical representation of the relation between measured attributes and legume types. To gain further understanding into the specific volatiles or fatty acids which are clearly different between legume samples, variable selection was performed using a VID technique. The selected discriminant compounds are listed in Table 2.To illustrate the differences amongst the seeds, some representative discriminant volatiles and fatty acids are also visually presented in Figure 3, with significant difference (p<0.05) determined using analysis of variance and Tukey’s post-hoc test. Key points are discussed in Section 4. 139 Foods 2019,8, 651 Table 2. List of discriminant volatile compounds/fatty acids for individual legume samples. VID Identity RI Chemical Group Soybean (18) 0.989 α-methyl-γ-butyrolactone 1621 Ester & Lactone 0.977 1-octen-3-one 1321 Ketone 0.971 β-methyl-γ-butyro-lactone 1644 Ester & Lactone 0.967 Heptanal 1189 Aldehyde 0.964 Linoleic acid * Fatty Acid 0.959 2(Z)-heptenal 1349 Aldehyde 0.958 Stearic acid * Fatty Acid 0.941 1-octen-3-ol 1461 Alcohol 0.934 1-pentanol 1251 Alcohol 0.934 2(E)-octenal 1458 Aldehyde 0.920 Palmitic acid * Fatty Acid 0.902 3,5-octadien-2-ol 1433 Alcohol 0.889 2(Z)-penten-1-ol 1330 Alcohol 0.878 2,4-nonadienal 1710 Aldehyde 0.874 5-ethylcyclopent-1-enecarboxaldehyde 1451 Aldehyde 0.867 3-octanone 1268 Ketone 0.823 Pentanal 949 Aldehyde 0.804 Oleic acid * Fatty Acid Lentil (15) 0.984 2-butanone 873 Ketone 0.982 Pyrrole 1540 Pyrrole 0.976 Menthol 1647 Alcohol 0.974 2-methoxyethylbenzene 1519 Hydrocarbon 0.970 Anethole 1816 Hydrocarbon 0.966 Caryophyllene 1630 Terpene 0.958 o-cymene 1291 Terpene 0.953 α-copaene 1529 Terpene 0.944 Linalool 1554 Terpene 0.941 Terpinen-4-ol 1620 Terpene 0.898 α-terpinyl acetate 1705 Terpene 140