UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Bitterness in enzymatically-produced hydrolysates of commercial shrimp (Pandalopsis dispar) processing… Cheung, Imelda Wing Yan 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-0357.pdf [ 9.15MB ]
Metadata
JSON: 831-1.0100797.json
JSON-LD: 831-1.0100797-ld.json
RDF/XML (Pretty): 831-1.0100797-rdf.xml
RDF/JSON: 831-1.0100797-rdf.json
Turtle: 831-1.0100797-turtle.txt
N-Triples: 831-1.0100797-rdf-ntriples.txt
Original Record: 831-1.0100797-source.json
Full Text
831-1.0100797-fulltext.txt
Citation
831-1.0100797.ris

Full Text

Bitterness in Enzymatically-Produced Hydrolysates of Commercial Shrimp (Pandalopsis dispar) Processing Waste by  Imelda Wing Yan Cheung  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF  M A S T E R OF SCIENCE in The Faculty of Graduate Studies (Food Science) THE UNIVERSITY OF BRITISH C O L U M B I A October 2007  © Imelda Wing Yan Cheung, 2007  A B S T R A C T  The shrimp processing industry produces a huge amount of waste which can potentially be converted into value-added products by enzymatic hydrolysis. Nevertheless, bitterness in hydro lysates is a common problem that needs to be addressed. In this study, 16 protein hydrolysate samples were produced from commercial shrimp (Pandalopsis dispar) processing wastes using Taguchi's L (4 ) experimental design. Four factors, namely—5  16  water-to-substrate ratio, percent enzyme, time of hydrolysis and type of protease, were investigated. The properties of the shrimp waste hydrolysates were assessed by three responses: product yield, degree of hydrolysis and bitterness. It was found that the type of protease had the most significant impact on hydrolysate properties. Hydrolysates produced from the proteases had significantly higher soluble product yields compared to the controls incubated without added protease. Moreover, the yield from treatments with Alcalase or Protamex reached over 30 %, which was six times higher than the control samples. In terms of degree of hydrolysis (DH), Alcalase, Flavourzyme and Protamex gave higher DH than bromelain, which had a DH similar to the controls. Despite the high soluble yields and DH for Alcalase and Protamex hydrolysates, their bitterness was intense and the 10 % (w/v) solutions of these samples were evaluated to be greater than 2000 ppm caffeine. Bromelain and Flavourzyme samples had significantly lower bitterness close to 1500 ppm caffeine but these samples were still significantly more bitter than the controls, which contained nearly no bitterness. A sample (blended in 2:1 water-to-substrate ratio and hydrOlyzed with 4 % Alcalase for 4 hours) that gave a soluble yield of 37.64 %, a DH of 2.30 meq/g and bitterness of 2300 ppm was selected for further fractionation by size and hydrophobicity to study the characteristics of the bitter substances. Results showed that bitter substances were small having molecular weight under 3 kDa and contained a large amount of hydrophobic amino acid residues such as Tyr, Phe, Leu, He and Lys. Therefore, it was concluded that Flavourzyme had the best potential to be used to produce protein hydrolysates from shrimp processing discards and small hydrophobic peptides were the major contributors to the bitterness.  n  T A B L E  OF  C O N T E N T S  Abstract  ii  Table of Contents  iii  List of Tables  vii  List of Figures  x  Acknowledgements  xiii  Co-Authorship Statement  xiv  CHAPTER I  Introduction  1.1  Study Context  1  1.2  Protein Hydrolysates from Shrimp Processing Discards  5  1.2.1  Shrimp Waste  5  1.2.2  Protein Hydrol ys ates  6  1.2.3  Type of Hydrolysis  1.2.4  Factor Consideration for Enzyme-Assisted Hydrolysis Reactions  1.2.5  Degree of Hydrolysis  1.3  1.4  1.5  Bitter Taste Sensation  10 12 20 21  1.3.1  Bitter Compounds and Their Structures  21  1.3.2  Bitter Taste Receptor  21  1.3.3  Bitter Taste Transduction Mechansim  24  1.3.4  Similarities between Sweet and Bitter  24  Bitter Peptides in Protein Hydrolysates  30  1.4.1  Bitter Peptides  30  1.4.2  Methods of Debittering or Minimizing Bitterness  38  Peptide Fractionation  41  1.5.1  Fractionation by Size  41  1.5.2  Fractionation by Hydrophobicity  43  1.5.3  Fractionation using Two-Dimensional Methods  43  1.6  Hypothesis and Objectives  47  1.7  References  49  iii  CHAPTER H  Use of Taguchi's Experimental Design to Investigate the Hydrolysis Conditions for Production of Protein Hydrolysates from Shrimp  (Pandalopsis  dispar)  Processing Discards 2.1  Introduction  60  2.2  Materials and Methods  62  2.2.1  Materials  62  2.2.2  Sample Preparation  63  2.2.3  Proximate Analysis of Shrimp Processing Waste  63  2.2.4  Shrimp Waste Hydrolysate Production  63  2.2.5  Degree of Hydrolysis (DH) Determination  2.2.6  Informal Sensory Evaluation of Shrimp Waste Hydrolysates  68  2.2.7  Formal Sensory Evaluation of Shrimp Waste Hydrolysates  68  2.2.8  Proximate Analysis of Shrimp Waste Hydrolysates  71  2.2.9  Amino Acid Analysis of Shrimp Waste Hydrolysates  71  2.2.10  Statistical Analysis  2.3  71  Results and Discussion  2.3.1  2.4  67  73  Proximate Analysis of Shrimp Processing Waste, Solid Waste Blend and Liquid Waste Blend  73  2.3.2  Preliminary Studies with Liquid Waste Blend  .76  2.3.3  Hydrolysate Production from Whole Shrimp Waste using Taguchi's Fractional Factorial Design  80  2.3.4  Sensory Evaluation  90  2.3.5  Hydrolysate Production Reproducibility  2.3.6  Proximate Analysis of Soluble Shrimp Waste Hydrolysates  104  2.3.7  Amino Acid Analysis of Soluble Shrimp Waste Hydrolysates  106  2.3.8  Conclusions  References  102  110 112  iv  CHAPTER III  Bitter Fractions from Enzymatically-Produced Hydrolysates of Commercial Shrimp Processing Waste  3.1  Introduction  118  3.2  Materials and Methods  120  3.2.1  Materials  120  3.2.2  Gel Filtration Chromatography of Soluble Hydrolysates Produced from Liquid Waste Blend  3.3  121  3.2.3  Fractionation of Soluble Hydrolysates from Whole Shrimp Waste  121  3.2.4  Sensory Evaluation of Fractionated Shrimp Waste Hydrolysates  125  3.2.5  Amino Acid Analysis of Fractionated Shrimp Waste Hydrolysates  125  Results and Discussion  •.  126  3.3.1  Preliminary Studies with Liquid Waste Blend Hydrolysates  126  3.3.2  Fractionation of Bitter Shrimp Waste Hydrolysates by Ultrafiltration... 133  3.3.3  Further Fractionation of Ultrafiltered Fractions from SWA4 Soluble Shrimp Waste Hydrolysates by HIC  3.3.4  136  Further Fractionation of HIC Fractions from Ultrafiltered Fractions of SWA4 Soluble Shrimp Waste Hydrolysates by GF  3.3.5  138  Fractionation of Soluble Shrimp Waste Hydrolysates of S WA4 by GF and Amino Acid Compositions of Bitter Fractions Isolated from  3.3.6 3.4  GF  144  Conclusions  148  References  ...  150  v  CHAPTER IV  Conclusions  4.1  Study Findings and Implications  153  4.2  Areas for Further Research and Application  156  4.2.1  Isolation and Identification of Bitter Peptides using TwoDiiriensional High Performance Liquid Chromatography and Mass Spectrometry  4.2.2  156  Sensory Evaluation of Shrimp Waste Hydrolysates on Shrimp Aroma and Shrimp Flavour  4.2.3  156  Study of the Effectiveness of Different Strategies to Mask the Bitterness and the Use of Enzymatic Approach to Debitter Shrimp Waste Hydrolysates  4.3  References  157 .  160  APPENDICES Appendix A  162  Appendix B  163  Appendix C  164  Appendix D  165  Appendix E  166  Appendix F  167  Appendix G  168  Appendix H  169  Appendix I  170  Appendix J  171  vi  LIST O F T A B L E S Table 1.1 Essential amino acid composition of shrimp waste hydrolysates and their recommended daily intake by WHO  7  Table 1.2 Essential amino acid composition of shrimp waste hydrolysates obtained by different researchers  8  Table 1.3 Examples of dietary uses of protein hydrolysates in human nutrition as protein supplements or for specific clinical conditions by Clemente (2000)  9  Table 1.4 Summary of origin, optimal pH arid temperature, protease type and preferential cleavage sites of Alcalase, bromelain, Flavourzyme and Protamex  18  Table 1.5 Contribution (%) of basic tastes to the tastes of free amino acid proposed by Fuke(1994)  23  Table 1.6 Bitter peptides of proteins and protein hydrolysates isolated by different researchers  .-31  Table 1.7 Fractionation methods and their separation characteristics summarized by Issaq and others (2002) and Stasyk and Huber (2004)  42  Table 1.8 Some examples of published fractionation methodology for isolation of bitter peptides  45  Table 2.1 The type of protease and the amount of enzyme used in the hydrolysis of liquid waste blend in the preliminary study and the pH recorded for the supernatant of the resulting shrimp waste hydrolysates  65  Table 2.2 Summary of Taguchi's Li6 (4) experimental design with 4 factors and 4 5  levels and additional experiments for hydrolysate production from whole shrimp waste  66  Table 2.3 Proximate composition of solid waste blend, liquid waste blend and whole shrimp waste  74  Table 2.4 Comparison of proximate composition of shrimp processing discards in this current study with existing literature  75  Table 2.5 a-Amino group content and informal sensory results of soluble shrimp waste hydrolysates from liquid waste blend  79  Table 2.6 The pH of whole shrimp waste before and after hydrolysis  82  Table 2.7 Summary of percent soluble yield, degree of hydrolysis and bitterness for each shrimp waste hydrolysate produced from the Taguchi's experimental design  83  Table 2.8 ANOVA General Linear Model results showing p-values of each factor investigated from the Taguchi's Li6 (4) design on the selected responses ....84 5  Table 2.9 Sensory scores of first sensory evaluation on soluble whole shrimp waste hydrolysates evaluated using a scale based on 0, 750, 1500 and 3000 ppm caffeine standard solutions  92  vin  Table 2.10 Sensory scores of second sensory evaluation on soluble whole shrimp waste hydrolysates evaluated using samples with scores (900, 1300, 1400, 1900 and 2300 ppm caffeine from Section 2.2.7.1 as references points and a sample with score 1300 pprii caffeine was used as the reference, marked at mid-point of the 15-cm line scale  95  Table 2.11 Reproducibility of shrimp waste hydrolysate production  103  Table 2.12 Proximate composition of two soluble hydrolysate samples produced by Alcalase (SWA4) and control incubation (SWC1)  105  Table 2.13 Amino acid composition of soluble shrimp waste hydrolysates compared to the starting shrimp waste blend  107  Table 3.1 Sensory qualities of fractions isolated from each protease-hydrolyzed sample of soluble liquid waste blend collected according to absorbance at 280 nm from a column of Sephadex G-25  131  Table 3.2 Amino acid composition of fraction 1 of SWA-1, fraction 2 of SWB-1 and fraction 1 of SWF-1 isolated from soluble liquid waste blend hydrolysates according to absorbance at 280 hm from a column of Sephadex G-25  132  Table 3.3 Product yields and their relative distribution of some soluble shrimp waste hydrolysates fractionated using ultrafiltration of 1 % (w/v) sample solutions through 10, 3 and 1 kDa membranes with a flow rate of 220 mL/min  135  Table 3.4 Amino acid composition of some bitter fractions collected from Sephadex G25 chromatography of SWA4 soluble hydrolysate  146  ix  LIST OF FIGURES Figure 1.1 Structures of some bitter compounds by Frank and others (2004)  22  Figure 1.2 The structure of T2R bitter taste receptors proposed by Montmayeur and Matsunami (2000)  25  Figure 1.3 Schematic of the bitter taste transduction cascades proposed by Lindemann (2001)  26  Figure 1.4 Schematic of sweet and bitter taste transduction mechanism proposed by Margolskee (2002)  28  Figure 1.5 Some examples of sweet and bitter compounds that share similarity in their structures adapted from Walters (1996)  29  Figure 2.1 Relative soluble and insoluble percent yields of shrimp waste hydrolysates produced from liquid waste blend  77  Figure 2.2 Main effect plots of mean percent soluble yield as a function of water-tosubstrate ratio, percent enzyme, time of hydrolysis and protease  .87  Figure 2.3 Boxplot of percent soluble yield as a function of protease  88  Figure 2.4 Main effect plots of mean degree of hydrolysis as a function of water-tosubstrate ratio, percent enzyme, time of hydrolysis and type of protease  89  Figure 2.5 Plot of degree of hydrolysis as a function of type of protease, percent enzyme and time of hydrolysis  91  x  Figure 2.6 A plot of z-score of first sensory evaluation of soluble shrimp waste hydrolysates  94  Figure 2.7 A plot of bitterness in ppm caffeine of second sensory evaluation of soluble whole shrimp waste hydrolysates  96  Figure 2.8 Boxplot of bitterness (ppm caffeine) of duplicate sensory scores from 8 panelists each for both sensory evaluations for soluble shrimp waste hydrolysate samples, as a function of time  98  Figure 2.9 Main effect plots of mean bitterness (ppm) as a function of water-tosubstrate ratio, percent enzyme, time of hydrolysis and type of protease Figure 2.10 Boxplot of bitterness (ppm) as a function of protease  99 101  Figure 3.1 Schematic of separation and analysis procedures adapted for fractionation of bitter peptides from soluble whole shrimp waste hydrolysate in the current study  122  Figure 3.2 Molecular weight distribution of various soluble liquid waste blend 1-hour hydrolysates at 214 nm from a column of Sephadex G-25  127  Figure 3.3 Molecular weight distribution of various soluble liquid waste blend 1-hour hydrolysates at 280 nm from a column of Sephadex G-25  128  Figure 3.4 Hydrophobic interaction chromatogram of soluble shrimp waste hydrolysate of SWA4 UF > 10 kDa on Hi-Trap™ Phenyl (low sub) column  137  Figure 3.5 Molecular weight distribution of HIC fraction 1 of SWA4 UF > 10 kDa from a column of Sephadex G-25  139  xi  Figure 3.6 Molecular weight distribution of HIC fraction 2 of SWA4 UF > 10 kDa from a column of Sephadex G-25  140  Figure 3.7 Molecular weight distribution of HIC fraction 3 of SWA4 UF > 10 kDa from a column of Sephadex G-25  141  Figure 3.8 Molecular weight distribution of HIC fraction 7 of SWA4 UF > 10 kDa from a column of Sephadex G-25  142  Figure 3.9 Molecular weight distribution of SWA4 soluble shrimp waste hydrolysate from a column of Sephadex G-25  145  ACKNOWLEDGEMENTS  I would like to express rriy sincere thanks to Dr. Eunice Li-Chati, my supervisor, for her continual guidance and thoughtful advice throughout the program.  Also, I would like to thank my committee members Dr. Tim Durance, Dr. Christine Seaman and Dr. Alex Yousif for their valuable comments and suggestions to improve my research approach. As well, thanks to Albion Fisheries Ltd. and Neova Technologies Inc. for sponsoring the commercial shrimp processing waste and the food grade proteases, respectively, and to the technicians Dr. Pedro Aloise and Val Skura for their laboratory expertise.  Thanks to my colleagues, Anusha Samaranayaka, Sooyeun Moon, Judy Chan, Wendy Lo, Andrea Liceaga, Tom Ho, Crystal Cinq-Mars, Melanie Lam, Andrea Goldson, Monica Purnama, Xiumin Chen, and Katie Hu for their support.  xiii  CO-AUTHORSHIP STATEMENT  This project was initiated by discussion between Dr. Eunice C.Y. Li-Chan (supervisor) and Imelda Wing Yan Cheung. All the experimental work and data analysis were performed by Imelda Wing Yan Cheung while the manuscript was prepared under the guidance of Dr. Eunice C.Y. Li-Chan.  xiv  CHAPTER I  1.1  Introduction  Study Context  Protein hydrolysates are one of the extensively studied functional foods. They usually contribute to several purposes: medical diets (Schmidl and others 1994; Clemente 2000), protein supplementation (Frokjaer 1994; Clemente 2000) and food use (Lahl and Braun 1994). Because of their broad application, there have been many studies to investigate the production of protein hydrolysates from different sources. Casein (Chobert and others 1988; Gallagher and others 1994; Morato and others 2000; Pedroche and others 2004), whey (Jost and Monti 1977; Smyth and FitzGerald 1998; Doucet and others 2003) and soy (Deeslie and Cheryan 1988; Aaslyng and others 1998; Lee and others 2001) have been excellent resources while other materials, such as chickpea (Clemente and others 1999), fish (Beddows and others 1976; Shahidi and others 1995; Diniz and Martin 1996; Imm and Lee 1999; Ravallec-Ple and others 2000), fish waste (Benjakul and Morrissey 1997; Liaset and others 2000; Guerard and others 2001; Guerard and others 2002; Liaset and others 2003; Deng and others 2004; Aspmo and others 2005; Nilsang and others 2005; Dumay and others 2006), shrimp (Simpson and others 1998; Mukhin and Novikov 2001; He and others 2006) and shrimp waste (Jaswal 1989; Balogun and AkegbejoSamsons 1992; Cancre and others 1999; Syndwiecki and Al-Khateeb 2000; Gildberg and Stenberg 2001;  Mizani  and others  2005; deHolanda and Netto 2006;  Ruttanapornvareesakul and others 2006; Coward-Kelly and others 2006; Cavalheiro and others 2007) have also gained popularity lately. The Canadian fisheries and oceans industries are huge with total commercial catches of a million tonnes and total value of $2 billion in 2005 (Fisheries and Oceans Canada 2007). Seafood processing, one of the principal economic components of British Columbia's fisheries, contributes to approximately one-third of this revenue. Of all the BC seafood production, wild shellfish production accounts for about 20 % and shrimp/prawns together constitute a wholesale value of over 45 million dollars in 2003 (Ministry of 1  Agriculture and Lands 2004). During shrimp processing, approximately 50 - 75 % of the total landings composed of shrimp heads, tails and shells are discarded as waste, which can cause environmental problems (Mandeville and others 1991; Goldsmith and others 2003; Islam and others 2004). This large amount of waste has been shown to be an excellent resource for proteins (Lekshmy and Prabhu 1989; Shahidi and Synowiecki 1991; Mandeville and others 1992a; Shahidi 1994; Synowiecki and Al-Khateeb 2000; Mizani and others 2005; deHolanda and Netto 2006) so hydrolysate production with these materials can reduce the disposal cost and increase the overall profit for food processors. Although protein hydrolysates hold many positive attributes, their sensory qualities can pose major concerns. Protein hydrolysates usually carry a very undesirable bitter taste that detracts customers from consuming them (Aubes-Dufau and others 1995; Lee and Warthesen 1996; Kim and others 1999; Maehashi and others 1999; Kim and others 2003; Cho and others 2004). Many attempts have been made to reduce or eliminate the bitterness of the hydrolysates and yet, success has been limited (Lalasidis and Sjoberg 1978; Minagawa and others 1989; Yeom and others 1994; Pedersen 1994; Izawa and others 1997; Stevenson and others 1998; Saha and Hayashi 2001; FitzGerald and O'Cuinn 2006). Therefore, it is crucial to understand the substances that produce the bitter taste and devise a method to specifically target the removal of the bitter compounds. Two important substances, specifically lipid oxidation compounds and peptides, have been proposed to cause the bitterness in protein hydrolysates. Occurrence of lipid oxidation is a problem during processing and storage arid it was suggested that shaking the reactor during the hydrolysis reaction would enhance the production of bitter taste, especially for starting materials rich in lipids (Liu and others 2000). However, lipid oxidation could be easily controlled by the addition of antioxidants or the use of nitrogen gas (Liu and others 2000). On the contrary, the elicitation of bitter taste from peptides is more difficult to control. The bitter taste Of proteins is usually released after extensive hydrolysis; thus peptides are believed to be responsibleforthis undesirable characteristic of the protein hydrolysates (Kirimura and others 1969; Ishibashi and others 1988a; Aubes-Dufau and others 1995; Lee and Warthesen 1996; Gomez and others 1997; Kim  2  and others 1999; Cho and others 2004; Fallico and others 2005). Different commercially available proteases have different specificities for cleavage sites; the peptide mixtures that are produced from various enzymes will not be the same. Peptide length (Asao and others 1987; Aubes-Dufau and others 1995; Cho and others 2004), peptide hydrophobicity (Ishibashi and others 1987; Asao and others 1987; Ishibashi and others 1988a; Aubes-Dufau and others 1995; Gomez and others 1997; Kim and others 1999; Fallico and others 2005) and amino acid sequence (Kim and others 1999; Kim and others 2003) have been suggested to be important factors contributing to the bitterness; therefore, the molecular weight profiles, the abundance of hydrophobic amino acids and the primary structure of the peptides present in the hydrolysate mixtures will provide information relevant to the bitter taste caused by peptides. Different hydrolysis conditions have been suggested to affect the quality of the final products. Type of proteases (Shahidi and others 1995; Simpson and others 1998; Liaset and others 2000; Nilsang and others 2005; Aspmo and others 2005; Dumay and others 2006; deHolanda and Netto 2006), enzyme-to-substrate ratio (Diniz and Martin 1996; Simpson and others 1998; Ravallec-Ple and others 2000; Guerard and others 2002; Nilsang and others 2005; Aspmo and others 2005) and incubation time (Lee and others 2001; Nilsang and others 2005) are some factors that have been extensively explored. These experimental parameters are important factors to consider for devising a methodology for the production of protein hydrolysates. Although many studies have been done to explore the use of shrimp processing wastes to produce protein hydrolysates, there has been little effort to study the different experimental parameters and their impact oh the yields and sensory qualities of the final products. Therefore, in the current study, Taguchi's experimental design, which is well known for obtaining maximum amount of information using the least number of experiments (Charteris 1992), was used to investigate four experimental factors (waterto-substrate ratio, percent enzyme, time of hydrolysis and type of proteases) in regards to their effects on soluble yield, degree of hydrolysis and bitterness of the final shrimp waste hydrolysates. The goal was to discover process conditions by which shrimp waste  3  hydrolysates could be made with high soluble yield and degree of hydrolysis but with low or no bitterness. Further, through fractionation by size and hydrophobicity, bitter fractions were isolated. These fractions could provide information on the molecular weight and hydrophobicity of the bitter peptides. The amino acid compositions of the enzymatically-produced hydrolysates of commercial shrimp processing discards and some isolated bitter fractions were also compared to examine if certain amino acids were major contributors to the bitter flavour. Understanding these characteristics of the bitter peptides could provide insights to develop methodology to produce non-bitter protein hydrolysates from shrimp processing wastes.  4  1.2 1.2.1  Protein Hydrolysates from Shrimp Processing Discards Shrimp Waste  In Canada, shrimp/prawn processing is a big sector of the seafood processing industry. In 2003, the total landings of shrimp/prawns reached 34,000 tonnes which represented a wholesale value of 47 million dollars (Ministry of Agriculture and Lands 2004). Although the high shrimp/prawn production generates a considerable amount of revenue to the country, the huge amount of processing waste, which constitutes approximately 50 - 75 % of the total landings composed of cephalothorax and exoskeleton, can lead to serious environmental problems (Mandeville and others 1991; Goldsmith and others 2003; Islam and others 2004). In order to dispose of these processing wastes, food processing industries are facing high financial burdens due to the enforcement of pollution laws that prohibit direct dumping of discards to the oceans or landfill sites (Mandeville and others 1991). Because of this, numerous studies have shown that these discards contain high contents of proteins (Lekshmy and Prabhu 1989; Mandeville and others 1992a; Synowiecki and Al-Khateeb 2000; Mizani and others 2005) which could be further utilized to produce value-added products. Shrimp processing discards contain 70 - 80 % moisture, 40 - 45 % proteins (dry basis, db), 20 - 30 % ash (db), 5 - 10 % fats (db) and other minor components (Shahidi and Synowiecki 1991; Shahidi 1994; Synowiecki and Al-Khateeb 2000). Some studies have shown higher protein contents (Mandeville and others 1992a; Gildberg and Stenberg 2001; Mukhin and Novikov 2001; Mizani and others 2005; Coward-Kelly and others 2006), lower lipids (Mandeville and others 1992a; Gildberg and Stenberg 2001; Mukhin and Novikov 2001), higher lipids (Coward-Kelly and others 2006) and lower ash (Mandeville and others 1992a; Mukhin and Novikov 2001; Coward-Kelly and others 2006). These variations in proximate composition have been suggested to be associated with seasonal changes, species differences, as well as diet (Gordon and Roberts 1977; Mandeville and others 1992a).  5  Aside from the major components, extensive research has been done on utilizing a nitrogen-containing polysaccharide component, chitin (Synowiecki and Al Khateeb 2000; Gildberg and Stenberg 2001; Diaz-Rojas and others 2006), in shrimp processing wastes. Chitin can be converted to chitosah, which has many usages in cosmetics and hair conditioning products. Enzymes such as chitinase, alkaline phosphatase and hyaluronidase have also been isolated from shrimp for other applications (Shahidi 1994). Flavour-active components, such as fatty acid esters, long chain alcohols, aldehydes, ketones, carboxylic acids and hydrocarbons, have been extracted from shrimp processing discards to produce seafood flavourings (Mandeville and others .1991; Mandeville and others 1992b). Furthermore, shrimp wastes contain many peptides and free amino acids, which are essential for humans (Table 1.1), and the production of protein hydrolysates has attracted great interest by researchers (Jaswal 1989; Ferrer and others 1996; Synowiecki and Al-Khateeb 2000; Gildberg and Stenberg 2001; Mizani and others 2005). During the hydrolytic process, proteins and large peptides are continuously cleaved into smaller peptides or free amino acids, which increase the nutritive values of the final products (Table 1.2). Therefore, many studies have been performed to hydrolyze the processing discards to produce potential protein hydrolysates (Jaswal 1989; He and others 2006) and animal feeds (Jaswal 1989; Gildberg and Stenberg 2001; Mizani and others 2005; Coward-Kelly and others 2006; Cavalheiro and others 2007). 1.2.2  Protein Hydrolysates  Protein hydrolysates, with their improved nutritional and functional qualities, can be used as protein supplements (Frokjaer 1994; Clemente 2000), medical diets (Schmidl and others 1994; Clemente 2000) as well as general food use (Lahl and Braun 1994). They can not only benefit patients with disorders such as digestion difficulty, absorption problems and poor amino acid metabolism, but can also help to reduce allergic reactions in hypersensitive individuals (Clemente 2000). Table 1.3 lists some examples of dietary uses of protein hydrolysates in human nutrition.  6  Table 1.1  Essential amino acid composition of shrimp waste and their recommended daily intake by WHO.  Reference  Coward-Kelly and others 2006  He and others 2006  Penaeus indicus Acetes chinensis Shrimp Waste Content (g/lOOg)  3-4 months old  2-5 years old  10-12 years old  Adult  Essential  Shrimp Waste Content (g/lOOg)  Histidine  1.78  1.450.  28  n.d.  n.d.  8-12  Isoleucine  4.54  2.219  70  31  30  10  Leucine  8.30  5.118  161  73  45  14  Lysine  5.63  5.266  103  64  60  12  Methionine  1.84  2.789  Cysteine  n.d.  4.340  58  27  27  13  Phenylalanine  4.93  3.290  Tyrosine  3.15  2.316  125°  69  27  14  Threonine  4.06  2.813  87  37  35  7  Tryptophan  n.d.  n.d.  17  12.5  4  3.5  Valine  5.77  2.823  93  38  33  10  40.00  32.424  714  352  261  84  From World Health Organization (1985). Value includes both methionine and cysteine. °Value includes both phenylalanine and tyrosine, n.d. represents values not determined. b  a  Shrimp Species  Total Essential Amino Acids a  Recommended Daily Intake (mg/kg/day)  b  b  c  b  c  b  c  Table 1.2  Essential amino acid composition of shrimp waste hydrolysates obtained by different researchers.  Reference  3  Jaswal 1989  Synowiecki and Al-Khateeb 2000  Gildberg and Stenberg 2001  Ruttanapornvareesakul and others 2006  Mizani and others 2005  Pandalus Metapenaeus Penaeus Penaeus endeavouri monodon semisulcatus eous  Crangon crangon  Pandalus borealis  mg/g dry weight  g/lOOg amino acids  g/100 g amino acids  g/kg protein  g/100 g dry matter  g/100 g dry matter  g/100 g dry matter  Histidine  6.6  5.01 ±0.08  3.12 ±0.00  13.72 ±0.20  2.82  2.74  2.62  Isoleucine  12.2  13.2 ±0.04  5.77 ±0.03  32.21 ±0.11  2.45  2.82  2.80  Leucine  19.1  8.86 ±0.09  49.80 ±0.51  4.90  5.64  5.67  Lysine  18.8  6.60 ± 0.02  8.31 ±0.32  53.79 ±0.10  5.07  5.84  6.02  Methionine  6.3  2.99 ±0.10  3.30 ±0.04  30.70 ± 0.50  2.42  2.10  2.12  Phenylalanine  12.1  4.93 ± 0.07  5.55 ± 0.09  59.20 ± 0.22°  3.69  4.67  4.50  Threonine  13.2  5.19 ±0.03  6.04 ± 0.02  33.02 ±0.20  3.43  3.54  3.71  Tryptophan  n.d.  1.20 ±0.03  n.d.  n.d.  n.d.  n.d.  n.d.  Valine  14.2  5.89 ±0.10  6.72 ± 0.04  35.54 ± 0.61  3.15  3.54  3.46  Shrimp Species Units  Pandalus borealis  b  :  Enzymatic hydrolysis in all cases except Jaswal (1989) in which hydrolysates were made from 6 hours hydrolysis with 5 N HCI. "Reported values included cysteine. °Reported values included tyrosine, n.d. represents value not determined. a  Table 1.3  Examples of dietary uses of protein hydrolysates in human nutrition as protein supplements or for specific clinical conditions by Clemente (2000).  Protein Supplementation  Clinical Use  Energy drinks  Phenylketonuria (PKU)  Geriatric products  Hypoallergenic infant formula  Sport nutrition  Acute and chronic liver disease  Weight-control diets  Short bowel syndrome Crohn's disease Pancreatitis Ulcerative colitis  9  As mentioned in Section 1.1, protein hydrolysates have been produced from various sources such as casein, whey and soy. In addition, researchers have also explored other materials such as fish, fish waste, shrimp and shrimp waste. The use of shrimp and fish waste products has become a very attractive alternative to the food industry because the utilization of processing waste materials to produce a food product will not only reduce the disposal costs but also increase the profits. 1.2.3  Type of Hydrolysis  Scientists have developed different processes to produce hydrolysates such as autolytic hydrolysis, acidic hydrolysis, alkaline hydrolysis and enzymatic hydrolysis (Kristinsson and Rasco 2000). 1.2.3.1  Autolytic Hydrolysis  The action of the endogenous digestive enzymes in the foodstuffs is termed autolytic hydrolysis. Autolysis is very complex because the enzymes involved each have different activity requirements. Further, in some food sources such as fish, the type of enzymes as well as their concentration varies with seasons, gender, age and species; hence, it is very difficult to control the hydrolysis (Kristinsson and Rasco 2000). These complications also result in poor functionality of the hydrolysates due to the inability to predict the molecular weight profiles of the peptide mixture; as a result, the market value of autolytically-produced hydrolysates is poor. Nevertheless, this process has been widely used to produce fish sauce and fish silage. Although fish sauce produced by autolytic hydrolysis does not improve the nutritional value of the product, the shelf life and organoleptic properties are usually enhanced (Kristinsson and Rasco 2000).  10  1.2.3.2  Acid and Alkaline Hydrolysis  Acid hydrolysis usually involves the use of hydrochloric acid or sulfuric acid to extensively hydrolyze the proteins at high temperature (Kristinsson and Rasco 2000). Since the final products usually contain very small peptides or single amino acids, the major advantage is their enhanced solubility. This process has been used in the production of vegetable hydrolysates as food flavouring (Aaslyng and others 1998). Although this process is rather efficient and inexpensive, the downside of this method are the high salt content after neutralization and the destruction of an essential amino acid, tryptophan (Kristinsson and Rasco 2000). Several studies have shown the potential of using acid hydrolysis to produce protein hydrolysates from shrimp wastes (Jaswal 1989; Ferrer and others 1996; Coward-Kelly and others 2006; Cavalheiro and others 2007). Alkaline hydrolysis, on the other hand, hydrolyzes proteins primarily with sodium hydroxide. It is less preferred than acidic and enzymatic hydrolysis in hydrolysate production because of the many side reactions observed. Some examples include racemization of L-amino acids to D-amino acids, and P-elimination reaction that decreases the content of cysteine, serine and threonine (Kristinsson and Rasco 2000). These reactions not only reduce the amount of amino acids but also possibly introduce toxic substances. 1.2.3.3  Enzymatic Hydrolysis  Of all the hydrolytic processes, enzymatic hydrolysis seems to be the most promising for producing protein hydrolysates of improved nutritive values. Enzymatic hydrolysis involves the use of exogenous enzymes to assist in hydrolyzing proteins. The preferred cleavage sites, the optimal activity and the working conditions of many proteases have been reported and thus, make it easier to predict the extent of hydrolysis and the peptide profiles of the resulting products. Hydrolysis with added enzymes produces hydrolysates with peptides of characteristic amino acid composition and size that could be beneficial to  11  individuals with hypersensitivity, digestion difficulty as well as absorption problems (Clemente 2000). Many researchers have documented the use of enzymatic hydrolysis to hydrolyze food materials. Some examples include hydrolysis of soy protein (Aaslyng and others 1998; Cho and others 2004), red hake (Imm and Lee 1999), capelin (Shahidi and others 1995), chickpea (Clemente 1999), cod (Ravallec-Ple and others 2000) and shrimp (Simpson and others 1998). Although hydrolytic processes can increase the use of underutilized food substances and improve the nutritive values of the product, the negative aspect of the protein hydrolysates is their undesirable sensory qualities. Very often protein hydrolysates will exhibit a bitter flavour that is very unappealing to consumers. 1.2.4  Factor Consideration for Enzyme-Assisted Hydrolysis Reactions  When planning a hydrolysis reaction, many factors need to be considered. Enzymatic hydrolysis involves the use of proteases to hydrolyze materials to smaller peptides or single amino acids; hence, factors that can affect the activity of the chosen protease will have an impact on the hydrolysis reaction. Enzyme performance depends on pH and temperature because each protease has its own optimal operating condition at which its activity can reach the highest. Other factors such as the nature of the substrate, water-tosubstrate ratio, amount of enzyme used, incubation time as well as the proteases can affect the overall performance of the hydrolysis reaction. Many studies have been carried out to investigate the optimal conditions for hydrolysate production through enzymatic hydrolysis. Diniz and Martin (1996) have explored the effect of pH, temperature and enzyme-to-substrate ratio on hydrolyzing dogfish muscle while Nilsang and colleagues (2005) have studied the impact of temperature, type of protease, enzyme concentration and incubation time on hydrolysis of fish soluble concentrate. Another study looked at the effect of temperature, pH, enzyme-to-substrate ratio and incubation time on casein hydrolysate production (Morato and others 2000) 12  whilst temperature, time and enzyme concentration were examined for shrimp meat hydrolysis (Simpson and others 1998). Influence of type of protease was studied by several researchers (Shahidi and others 1995; Aspmo and others 2005; Dumay and others 2006) while effect of temperature, time, pH and enzyme concentration on hydrolysis of cod muscle was explored by Ravallec-Ple and others (2000). 1.2.4.1  Temperature and pH  Temperature and pH are important factors affecting the activity of the protease; more specifically, each protease functions optimally at a certain temperature and pH range. Several studies have looked at the temperature and pH effect on hydrolysate production. Shahidi and colleagues (1995) have focused on the effect of temperature for production of capelin hydrolysate and it was found that the optimal temperature was associated with incubation time. When the substrate was incubated for one hour, the optimal temperature for the hydrolysis was found to be 60 °C while for a 2-hour hydrolysis, the optimal temperature was 55 °C. Another study performed by Ravallec-Ple and others (2000) indicated that the closer the experimental condition was to the optimal pH of the protease, the higher the degree of hydrolysis was achieved. The protease chosen for the study was Alcalase, which had an optimal pH of 8.0. It was shown that as pH was increased from 7-8, the DH increased from 8.2 % to 24.3 % while further increase of pH from 8 to 8.2 did not increase the DH. Even though sufficient evidence has suggested temperature and pH are essential parameters to consider for hydrolysis conditions, the energy expenditure and labour costs for temperature and pH adjustments should not be overlooked. Depending on the nature of the substrate, the pH may need to be adjusted by acid or base to reach the optimal pH of the protease, in which case a considerable amount of salt can be generated during the neutralization reaction. Salt may not be easily removed from the final product so acceptability of the final product may vary depending oh the applications. 13  1.2.4.2  Water-to-substrate Ratio  Even though water-to-substrate ratio is not as influential as other factors in the hydrolysis reaction, several studies have documented different values indicating its potential importance in affecting hydrolytic performance. Numerous studies have shown the use of lower water-to-substrate ratios such as 1:1 (Shahidi and others 1995; Gildberg and Stenberg 2001; Aspmo and others 2005; Dumay and others 2006), 2:1 (Synowiecki and Al-Khateeb 2000; Ruttanapornvareesakul and others 2006) and 4:1 (Nilsang and others 2005), while others have used higher ratios of 9:1 (Ravallec-Ple and others 2000), 10:1 (Simpson and others 1998) and 50:1 (Clemente and others 1999). The value of water-tosubstrate ratio used can relate to the nature of the substrate. However, none of these work looked at the effect of substrate concentration on the quality of final hydrolysate products. A study done by Benjakul and Morrissey (1997) investigated the effect of substrate/buffer ratio on hydrolysis and nitrogen recovery. It was shown that an increase in the ratio resulted in an increased nitrogen recovery up to a ratio of 1:3, after which no further increase was observed. In 2005, Spellman, O'Cuinn and FitzGerald (2005) performed an experiment to study the effect of substrate concentration on whey protein hydrolysates. They examined a range of different total solid levelsfrom50 g/L to 300 g/L and found that the DH increased with a decrease in total solid levels while the bitterness of samples having the same DH values were significantly lower when hydrolysis was performed at a lower substrate concentration. 1.2.4.3  Percent Enzyme  The amount of proteases in a hydrolytic reaction can dramatically affect the final products. Enzymes act as cleaving agents in the hydrolysis reaction to break down protein molecules into smaller peptides or single amino acids. An increase in the amount of protease added will increase the efficiency of the hydrolysis.  14  Diniz and Martin (1996) studied the effect of the enzyme-to-substrate ratio (E/S) of 2.0, 3.0 and 4.0 on the hydrolysis of dogfish muscle. They found that the DH increased with an increase of the E/S from 2.0 to 3.0; however, when the E/S was raised to 4.0, a decrease of DH was observed, which could possibly be due to the occurrence of enzyme inhibition. In a study performed by Guerard and others (2002) to explore the effect of varying enzyme concentration from 0.1 to 1.5 % (w/w protein) when hydrolysing tuna waste, an increase of DH was observed with an increase of enzyme concentration, but the increase of DH was not as rapid beyond 1 % enzyme concentration. When Nilsang and others (2005) hydrolyzed fish soluble concentrate using different enzyme concentration from 1 - 5 % (w/w protein), it was shown that enzyme concentration and temperature showed an increase of DH, with 5 % enzyme concentration achieving the highest DH. The effect of enzyme concentration on the recovery of soluble matters was examined during the hydrolysis of Atlantic cod viscera (Aspmo and others 2005). As enzyme concentration was increased from 0 to 2 % (w/w dry matter)^ an increase in soluble solid recovery was observed. Even though ah increase in enzyme-to-substrate ratio often offers higher degree of hydrolysis and higher recovery of soluble matters, a high amount of proteases will imply a higher cost of production. Therefore, it is necessary to maintain a balance between the amount of protease used and the desired level of hydrolysis to be achieved to keep a low cost of production. 1.2.4.4  Time of Hydrolysis  Time of hydrolysis is another widely studied factor in controlling hydrolysis reaction. A long hydrolysis time will lead to the breakdown of protein molecules into smaller peptides and single amino acids. Rebeca and others (1991) have studied the incubation time of 0, 1, 2 and 3 hours to hydrolyze eviscerated mullet. They found that soluble nitrogen increased from 3 % in the 15  substrate to about 80 % after hydrolysis for 3 hours. An increase of nitrogen recovery with time was observed for all the proteases that were studied. Similarly, in a study done by Nilsang and others (2005), it was found that as time of incubation was increased from 1 hour to 6 hours, the DH increased. Simpson and others (1998) examined the effect of hydrolysis time (1, 2 arid 3 hotirs) on aroma quality of shrirrip meat hydrolysates. It was shown that an increase in duration of hydrolysis increased the aroma quality and intensity. However, similar to percent enzyme, incubation time involves energy expenditure so investigation on the hydrolysis time required to produce the products is crucial. 1.2.4.5  Type of Protease  In choosing proteases for use in food processing, pH optimum, the heat stability, the presence of activators or inhibitors, the cost and availability, and the specificity of the enzyme are all crucial factors to consider (Anson and Others 1966; Aspmo and others 2005). Several food-grade enzyme preparations have evidence suggesting their efficiency in the enzymatic hydrolytic process. Acitinidin (Aspmo and others 2005), Alcalase® (Shahidi and others 1995; Diniz and Martin 1996; Benjakul and Morrissey 1997; Clemente and others 1999; Ravallec-Ple and others 2000; Synowiecki and Al-Khateeb 2000; Lee and others 2001; Guerard and others 2001; Aspmo and others 2005; Dumay and others 2006), bromelain (Beddows and others 1976; Aspmo and others 2005), Flavourzyme® (Aaslyng and others 1998; Clemente and others 1999; Imm and Lee 1999; Lee and others 2001; Nilsang and others 2005; Dumay and others 2006), Kojizyme™ (Nilsang and others 2005), Neutrase® (Shahidi and others 1995; Aspmo and others 2005), papain (Shahidi and others 1995; Aspmo and others 2005), Protamex® (Liaset and others 2003; Aspmo and others 2005; Dumay and others 2006) and Savorase M (Imm and Lee 1999) have all demonstrated their abilities in producing protein hydrolysates.  16  Alcalase has shown its superior ability to produce high soluble solid yield and degrade large proteins to small peptides; likewise, bromelain has been demonstrated to produce a high yield of alpha amino groups (Aspmo and others 2005). Hydrolysates produced with Flavourzyme have been shown to have low or no bitterness (Aaslyng and others 1998; Imm and Lee 1999) and a high DH (Aaslyng and others 1998) while Protamex has been suggested to produce non-bitter hydrolysates (Liaset and others 2003). For the purpose of this study, these four commercially-available food grade proteases were chosen based on their ability to solubilize, give high DH and/or produce non-bitter final hydrolysates. The origin, optimal pH and temperature as well as protease type of these enzymes are summarized in Table 1.4. 1.2.4.5.1  Alcalase®  Alcalase® 2.4 L FG (E.C. 3.4.21.62), also called subtilisin, is a serine protease derived from Bacillus licheniformi (Novozymes A/S 2001a). It has an activity of 2.4 Anson Units per gram (AU/g) and functions best at pH 8.0 - 8.5 and at temperature 55 - 60 °C (Novozymes A/S 2001a). The preferential cleavage sites of this enzyme have been suggested to be at hydrophobic side chains near the carboxyl side and Doucet, Otter, Gauthier and Foegeding (2003) have shown that this enzyme has specificity for Phe, Trp, Tyr, Glu, Met, Leu, Ala, Ser and Lys residues. Alcalase has been used in different types of hydrolysis reactions mainly because of its ability to produce smaller peptides in comparison to other commercially available enzymes (Smyth and FitzGerald 1998; Aspmo and others 2005). 1.2.4.5.2 Bromelain  Bromelain (E.C. 3.4.22.32) is a cysteine protease derived from pineapple stem and its activity is 2000 Gelatin Digestion Unit per gram (GDU/g) (Nutriteck 2005). It prefers to cleave at Lys, Arg, Phe and Tyr residues (Adler-Nissen 1986). This enzyme operates over a broad pH range from 3.8 to 8.0 and a temperature range of 45 - 60 °C (Nutriteck 2005).  17  Table 1.4  Summary of origin, optimal pH and temperature, protease type and preferential cleavage sites of Alcalase, bromelain, Flavourzyme and Protamex.  Origin  Alcalase  Optimal pH  Bacterial 8.0-8.5  a  Optimal Temperature (°C)  Type of Protease  55-60  Serine  a  Preferential Cleavage Sites  Phe, Trp, Tyr, Glu, Met, Leu, Ala, Ser, Lys b  Bromelain  Plant  3.8-8.0°  45 -60°  Cysteine  Lys, Arg, Phe, TyrCOOH d  Flavourzyme  Protamex  Fungal  5.5-7.5°  50-55°  Bacterial 5.5-7.5  35-60  f  f  Exopeptidase/ endoprotease complex  Endoprotease complex  Not known  N kn o t  0 w n  From Novozymes A/S (2001a). From Doucet and others (2003). °From Nutriteck (2005). From Adler-Nissen (1986). From Novozymes A/S (2001b). From Novozymes A/S (2001c).  a  b  d e f  18  1.2.4.5.3 Flavourzyme® Flavourzyme® 1000 L is an exopeptidase/eridopeptidase complex derived from the fungus Aspergillus oryzae and it has an activity of 1000 Leucine Amino Peptidase Unit per gram (LAPU/g) (Novozymes A/S 2001b). This enzyme preparation contains several exopeptidases and endopeptidases that are active at different pH. The endoproteases have both neutral and acid proteases, giving a broad operating pH (Pommer 1995). Flavourzyme's optimal pH and temperature are at 5.5 - 7.5 and 50 - 55 °C, respectively (Novozymes A/S 2001b). In contrast to other proteases, it functions better at a natural drifting pH because different peptidases in the complex will be activated at different pHs making the hydrolysis more efficient (Pommer 1995; Imm and Lee 1999). This nature of the enzyme preparation not only eliminates the trouble of adjusting the pH during the hydrolytic reaction, but also produces non-bitter hydrolysates with very diverse peptide mixtures which are mostly small in size (Aaslyng and others 1998; Novozymes A/S 2001b). 1.2.4.5.4 Protamex® Protamex® is a protease complex from a strain of Bacillus and its activity is reported to be 1.5 AU/g (Novozymes A/S 2001c). The optimal conditions of this enzyme are at pH 5.5 - 7.5 and a temperature of 35 - 60 °C (Novozymes A/S 2001c). It has been suggested to produce hon-bitter hydrolysates (Novozymes A/S 2001c; Liaset and others 2003).  19  1.2.5  Degree of Hydrolysis  The degree of hydrolysis (DH) is used to represent the extent of hydrolysis during a catalytic reaction. DH represents the percentage of peptide bonds cleaved during the hydrolysis reaction and it can be described using the following formula: DH = li/htot*100%  where h is called the hydrolysis equivalent expressed as milliequivalent per gram of protein which represents the number of peptide bonds cleaved during a hydrolytic process while h t can be determined by calculating the amount of free amino acids per gram of to  protein (Adler-Nissen 1986). Several methods have been used to determine the hydrolysis equivalents of the hydrolytic reactions.  Some  examples  are  formol  titration,  ninhydrin reaction and  trinitrobenzenesulphonic acid (TNBS) method. The TNBS is a widely used method because the reaction of free amino groups with TNBS gives chroniophores which allow the use of spectrophotometric techniques to monitor the reaction progress (Adler-Nissen 1979).  20  1.3  1.3.1  Bitter Taste Sensation  Bitter Compounds and Their Structures  Bitter substances are generally regarded as unfavourable by most humans despite some such as tea and coffee are widely accepted. Several compounds have been shown to elicit bitterness sensation and they are divided into two groups: ionic and nonionic. As shown in Figure 1.1, sucrose octaacetate and caffeine are examples of nonionic stimuli while quinine hydrochloride and denatonium benzoate are ionic bitter compounds. Besides these chemical structures, some amino acids have also been known to elicit bitterness. As shown in Table 1.5, proline, lysine, phenylalanine, histidine, arginine, valine, leucine and methionine are perceived as bitter. In addition, compounds produced during lipid oxidation have also been associated with the generation of bitterness. The production of lipid oxidation was shown to be enhanced by shaking of the reactor, which was usually applied during hydrolysis (Liu and others 2000). 1.3.2  Bitter Taste Receptor  Recently, the receptors responsible for the sensation of bitterness were identified using molecular and genomic approaches. The receptors belong to the T2R family, which contains approximately 30 genes (Margolskee 2002; Montmayeur and Matsunami 2002; Zhang and others 2003; Scott 2004). The diversity of the T2R receptors and their coexpression in one cell population on the tongue offer them the capability of responding to a wide range of structurally different bitter substances (Scott 2004). These 30 members of T2R receptors share 25 - 90 % identity within the gene family (Margolskee 2002; Montmayeur and Matsunami 2002). Their similarities are greatest in the three cytoplasmic loops and the adjacent transmembrane segments while their differences peak 21  This figure was removed for copyright reason; original source can be found in Frank ME, Bouverat BP, MacKinnon BI, Hettinger TP. 2004. The distinctiveness of ionic and nonionic bitter stimuli. Physiology and Behavior 80: 421-431.  Figure 1.1  Structures of some bitter compounds by Frank and others (2004).  22  T a b l e 1.5  Contribution (%) of basic tastes to the tastes of free amino acid proposed by Fuke (1994). Numbers in bold represent the major taste(s) of the corresponding amino acids.  This table was removed for copyright reason; original source can be found in Fuke S. 1994. Taste-active components of seafoods with special reference to umami substances. In: Shahidi F, Botta JR, editors. Seafoods: chemistry, processing technology and quality. 1 ed. Glasgow: Blackie Academic and Professional, p 115-139. st  23  in the extracellular regions, which are essential for interaction with a wide range of bitter structures (Margolskee 2002) (Figure 1.2). These receptors are predominantly expressed in circumvallate and foliate papillae while very few fungiform papillae contain bitter receptors (Margolskee 2002). 1.3.3  Bitter Taste Transduction Mechanism  The sensation of taste is very important to humans. Different tastes signal the intake of different food substances. Sweet compounds represent the intake of nutritionally rich food sources while bitter materials send alerts ori toxins and contaminants (Zhang and others 2003). When bitter substances come in contact with the human tongue, two proposed types of mechanisms can be activated (Figure 1.3). Some compounds, like quinine and denatoniurn, can permeate the cell membrane and directly trigger the ct-subunit of the G protein gustducin without passing through the receptors. Other substances, such as caffeine and other methyl-xanthines, can permeate the cell membrane and cause blockage of the phosphodiesterase (PDE) leading to an increase in guanidine 3',5'-cyclic monophosphate (cGMP) (Lindemann 2001). In contrast, some bitter compounds require the binding to the receptors to activate the taste transduction mechanism. The (3 and y subunits of the G protein gustducin, once bonded with a bitter substance, can activate phospholipase Cp2 (PLCP2) causing an increase in the concentration of inositol-1,4,5-trisphosphate (IP3) (Gilbertson and others 2000). This will trigger the release of intracellular calcium ions and lead to neurotransmission. 1.3.4  Similarities between Sweet and Bitter  A lot of research has been conducted to study the similarities between sweet and bitter tastes. There are evidences showing that bitter and sweet transduction mechanisms share  24  This figure was removed for copyright reason; original source can be found in Montmayeur JP, Matsunami H. 2002. Receptors for bitter and sweet taste. Current Opinion in Neurobiology 12: 366-371.  Figure 1.2  The structure of T2R bitter taste receptors proposed by Montmayeur and Matsunami (2000). Individual amino acids are shown as spheres. The darkcoloured spheres represent conserved residues while light-coloured spheres show amino acids that are variable.  25  This figure was removed for copyright reason; original source can be found in Lindemann B. 2001. Receptors and transduction in taste. Nature 413: 219-225.  Figure 1.3  Schematic of the bitter taste transduction cascades proposed by Lindemann (2001). R, multiple G-protein-coupled receptors; a, a-subunit of gustducin; Py,  G-protein subunits  P3 and yl3; P L C 0 2 ,  phospholipase C subtype;  IP3, inositol-1,4,5-triphosphate; PDE, taste-specific phosphodiesterase; cAMP, cyclic adenosine monophosphate; cGMP, cyclic guansine monophosphate; sGC, soluble guanylate cyclase; NO, nitric oxide; NOS, NO synthase.  26  many common components. Both sweet and bitter transductions contain ion channels in which compounds can pass through without the need of the G-protein (Gilbertson and others 2000; Zhang and others 2003). In addition, the G-protein receptor coupled to PLC (GPCR-PLC) pathway adopted by artificial sweeteners is extremely similar to that used by bitter compounds (Figure 1.4). Both mechanisms involve the binding of the substance to the G-protein with the activated py subunits triggering the reaction of PLC. Both pathways induce the production of IP3 and diacylglycerol (DAG) that will lead to an increase in calcium ion concentration to produce a neural signal (Walters 1996; Gilbertson and others 2000; Lindemann 2001; Margolskee 2002). Besides the similarities iii transduction mechanisms, researchers have demonstrated that sweet-tasting compounds and bitter substances are very similar. A small variation in the structure can alter the sweet taste of a compound to bitter, and vice versa (Walters 1996). Figure 1.5 shows some examples of bitter and sweet compounds that share similar structures. The stereoisomer of aspartame, an artificial sweetener, is perceived as bitter while the sulfur-substitute analog of saccharin also gives a bitter taste. In addition, the bitter compound neohesperidin, upon cleavage of the C-0 bond, produces an extremely sweet substance (Walters 1996).  27  This figure was removed for copyright reason; original source can be found in Margolskee RF. 2002. Molecular mechanisms of bitter and sweet taste transduction. The Journal of Biological Chemistry 277(1): 1-4.  Figure 1.4  Schematic of sweet and bitter taste transduction mechanism proposed by Margolskee (2002). AC, adenylyl cyclase; AP, action potentials; a, (3, y, subunits of gustducin; cAMP, cyclic adenosine monophosphate; cNMP, cyclic nucleotide monophosphate; DAG, diacylglycerol; IP3, inositol triphosphate; PDE, phosphodiesterase; PKA, protein kinase A; PLC, phospholipase C; PLC(32, phospholipase C subtype; NT, neurotransmitter.  This figure was removed for copyright reason; original source can be found in Walters DE. 1996. How are bitter and sweet tastes related? Trends in Food Science and Technology 7: 399-406.  Figure 1.5  Some examples of sweet and bitter compounds that share similarity in their structures adapted from Walters (1996).  29  1.4  1.4.1  Bitter Peptides in Protein Hydrolysates  Bitter Peptides  Most proteins, in intact forms, do not give a bitter sensation. The explanation is that the hydrophobic side chains are usually buried in the interior part of the protein due to hydrophobic interaction. The stereochemistry, hydrOphobicity as well as size of the peptides, are considered crucial to the sensation of bitterness (Ramos de Armas and others 2004). As a result of hydrolysis, protein molecules are degraded, exposing the hydrophobic side chains to interact with the taste buds and generate a bitter taste. Peptides have been shown to be responsible for the elicitation of bitter taste in protein hydrolysates (Aubes-Dufau and others 1995; Lee and Warthesen 1996; Kim and others 1999; Maehashi and others 1999; Kim and others 2003; Cho and others 2004). During the hydrolytic process, the protein molecules are hydrolyzed into smaller peptide chains or even single amino acids while at the same time, bitterness is gradually elicited and enhanced during the process. The sensation of bitterness has been postulated to associate with degradation of protein during hydrolysis to expose hydrophobic amino acid residues to come in contact with the bitter taste buds (Pedersen 1994). These bitter peptides have been suggested to fall between the molecular weight range of 1000 to 6000 Daltons (Aubes-Dufau and others 1995; Kristinsson and Rasco 2000; Cho and others 2004) but other studies have found that molecular weight of bitter peptides could be under 1 kDa (Kim and others 1999; Kim and others 2003). Numerous papers have also suggested that bitterness in peptides is associated with hydrophobicity (Ishibashi and others 1987; Ishibashi and others 1988a; Ishibashi and others 1988b; Nishitoba and others 1988; Aubes-Dufau and others 1995; Gomez and others 1997; Kim and others 1999). Many scientists have isolated and identified bitter peptides from protein hydrolysates produced using different sources, with the earliest work done dating back to the 1950's. Some examples are listed in Table 1.6. Most of these discovered peptides contain a  30  Table 1.6  Bitter peptides of proteins and protein hydrolysates isolated by different researchers.  Source  Peptides  Trypsin hydrolysate of casein  Gly-Pro-Phe-Pro-ValIle Phe-Phe-Val-Ala-ProPhe-Pro-Glu-Val-PheGly-Lys Phe-Ala-Leu-Pro-GlnTyr-Leu-Lys  Bacterial proteinase hydrolysate of casein Beer yeast residue  Pepsin hydrolysate of zein  Ney's A Q (kcal/mol)  Reference  1.833  Matoba and others 1970  a  1.733 1.875  Arg-Gly-Pro-Pro-PheIle-Val  1.729  Minamiura and others 1972  Trp-Phe  2.600  Trp-Pro Leu-Pro-Trp  2.400 2.567  Matsusita and Ozaki 1993  Ala-Ile-Ala Ala-Ala Leu Gly-Ala-Leu Leu-Gin-Leu Leu-Glu-Leu Leu-Val-Leu Leu-Pro-Phe-AsnGln-Leu Leu-Pro-Phe-Ser-GlnLeu  1.567 1.000 2.900 1.300 2.100 2.100 2.667  Wieser and Belitz 1975  1.733 1.783  Pepsin hydrolysate of hemoglobin  Val-Val-Tyr-Pro-TrpThr-Gln-Arg-Phe  1.756  Aubes-Dufau and others 1995  Alkalase hydrolysate of hemoglobin  Val-Val-Tyr-Pro-Trp  2.160  Aubes-Dufau and Combes 1997  31  Source Soy Protein  Bovine p-CN (Cterminal) Pepsin hydrolysate of soybean  Isolated from cheese  Pepsin hydrolysate of soy globulin  Peptides  Ney's AQ (kcal/mol)  Reference  Arg-Leu  2.000  Arg-Leu-Leu Ser-Lys-Giy-Leu  2.300 1.250  Fujimaki and others 1971  Arg-Gly-Pro-Phe-ProIle-Ile-Val  1.850  Takahashi and others 1995  Ser-Lys-Gly-Leu  1.250  Pedersen 1994  Phe-Ile-Gln-Gly-Val  1.540  Glu-Val-Leu-Asn Asn-Glu-Asn-Leu-Leu Ala-Pro-Phe-Pro-GluVal-Phe  1.375 1.220  Gly-Leu  1.450  Leu-Lys Leu-Phe Phe-Leu Tyr-Phe-Leu Gln-Tyr-Phe-Leu Phe-(Ile, Leu)-GlnGly-Val Pyr-Gly-Ser-Ala-IlePhe-Val-Leu Phe-(Arg, Asp , Gln , Gly, He, Leu, Lys , Pro, Set, Thr)-Trp(Ala, Arg, Asp, Gly, Val)-Gln-Tyr-Phe-Leu  2.400 2.600 2.600 2.267 1.825  a  2  b  2  Pedersen 1994  1\729 Fujimaki and others 1968; Fujimaki and others 1970; Yamashita and others 1969; Arai and others 1970  1.929 1.614  2  2  Trypsin hydrolysate of casein  Gly-Pro-Phe-Pro-IleIle-Val  1.300  1.957  Matoba and others 1969; Matoba and others 1970  32  Source  Subtilisin hydrolysate of casein  Peptides  Ney's AQ (kcal/mol)  Leu-Val-Pro-Arg-TyrPhe-Gly  1.714  Arg-Gly-Pro-Pro-PheIle-Val Val-Tyr-Pro-Phe-ProPro-Gly-Ile-Asn-His  a  Reference  Ichikawa and others 1959, Minamiura and others 1972  1.729 1.570  Papain hydrolysate of casein  Ala-Gln-Thr-Gln-SerLeu-Val-Tyr-Pro-PhePro-Gly-Pro-Ile-ProAsn-Ser-Leu-Pro-GlnAsn-Ile-Pro-Pro-LeuThr-Gln  1.441  Clegg and others 1974  Butterkaese  Pro-Phe-Pro-Gly-ProIle-Pro-Asn-Ser  1.411  Huber and Klostermeyer 1974  Arg-Leu  2.000  Arg-Leu-Leu Phe-Ile-Ile-Glu-GlyVal  2.300  Fujimaki and others 1968, Fujimaki and others 1970  Pepsin hydrolysate of soya protein  Cheddar cheese  Pro-Phe-Pro-Gly-IlePro Pro-Phe-Pro-Gly-ProIle-Pro-Asn-Ser Gln-Asp-Lys-Ile-HisPro-Phe-Ala-Gln-ThrGln-Ser-Leu-Val-TyrPro-Phe-Pro-Gly-ProIle-Pro  1.733 1.783  Hamilton and others 1974  1.411  1.505  Calculated from estimated hydrophobic effect, side chain burial in kcal/mol taken from Li-Chan (2004). Pyr stands for pyroglutamyl.  33  high content of hydrophobic side chains with large alkyl groups or aromatic groups. These groups are believed to be essential in causing bitter sensation. Different methods have been investigated in terms of predicting or evaluating the bitter taste of the peptides. The most famous scheme is the Q rule proposed by Ney in 1971 (Ney 1979). Ney postulated a correlation between the hydrophobicity of peptides and the bitterness by calculating the average free energy for the transfer of the amino acid side chains from ethanol to water, the Q value (Adler-Nissen 1986). He found that most bitter peptides had Q values larger than 1400 cal/mol while non-bitter ones were less than 1300 cal/mol. The Q values for the peptides isolated by different researchers listed in Table 1.6 were calculated and it is evident that most of these peptides have a Q-value greater than 1400 cal/mol, which further confirms Ney's hypothesis. However, when this rule Was extended to a mixture of peptides, it failed to give any meaningful correlation (Adler-Nissen 1986). In terms of predicting the hydrophobicity of a mixture of peptides, a lot of factors need to be considered. The hydrophobic characteristics of the intact protein cannot be used alone to estimate the bitterness of the resulting peptide mixture. This only accounts for an average hydrophobicity of the whole peptide mixture but not the distributive function of the Q value in the mixture (AdlerNissen 1986). Some peptides in the mixture can contain extremely high Q values while others are low. The overall hydrophobicity of the mixture can be very moderate but the peptide mixture can be very bitter because a small amount of bitter peptides can elicit extremely strong bitterness. The failure in predicting the bitterness in peptide mixture reduces the usefulness of the method because mostly peptide mixtures are used in the food industry instead of single peptide, due to the complexity of isolating individual peptides. A lot of researchers have attempted to study the mechanism of bitterness production in peptides. Ishibashi and colleagues have performed a series of studies to investigate the role of peptides in generating bitterness. In 1987, they synthesized peptides containing Phe and Tyr residues and evaluated the bitterness using a caffeine equivalent scale. They  34  explained the bitterness elicited with these peptides by stating that hydrophobic side groups could act as bitter taste determinant sites interacting with the bitter taste receptors to generate bitter sensation (Ishibashi and others 1987). In a later study, Ishibashi and others, through studying synthetic peptides, suggested that a minimum number of three carbons were needed in the side chain skeleton in order for the peptides to be bitter (Ishibashi and others 1988a). In the same year, they also postulated that two sites were required for the peptides to be bitter and that steric configuration was also essential. The two sites have been proposed to be spaced approximately 4.1 A apart. The primary site, also called the binding unit, consisted of hydrophobic side chains with a minimum of three carbons and was responsible for the bitterness of peptide. The secondary site, the stimulating site, contained either a bulky basic group or a hydrophobic group and acted as a determinant site for bitterness (Ishibashi and others 1988b). These two sites together with the correct steric configuration could interact with the bitter taste receptor to elicit bitterness. Although these results could explain the possible cause of bitterness, it was still difficult to evaluate the intensity Of bitter taste of peptides. Sensory analysis is commonly utilized in terms of rating the bitterness of the peptides. Several studies have documented detailed analysis of peptide bitterness in protein hydrolysates. Aubes-Dufau, Seris and Combes (1995) have carried out a study on producing peptic hemoglobin hydrolysates and evaluating the product bitterness. The soluble portion of the bovine hemoglobin solution was hydrolyzed by pepsin using different enzyme-to^ substrate ratios to produce hydrolysates. The hydrolysates were- freeze-dried and characterized using the TNBS method, gel filtration chromatography, reversed phase chromatography as well as sensory evaluation using reference to a quinine sulfate concentration scale. A hydrolysate sample with a DH of 13 % resulted in the strongest bitterness and thus was subjected to different fractionation procedures. Ultrafiltration, organic solvent extraction and reversed phase chromatography were used to isolate the bitter fractions. Ultrafiltration separated the hydrolysate sample into 4 fractions, namely < 0.5 K, 0.5 - 5 K, 5 - 10 K and > 10 K while solvent extraction isolated the bitter fraction  35  in the organic layer. Reversed phase chromatography slowed down the elution of the hydrophobic substances due to hydrophobic interaction and therefore was an effective method to separate hydrophobic fraction from others. From this study, the conclusion was that peptic hemoglobin hydrolysates exhibited a strong bitter taste and the intensity of bitterness increased with increasing DH from 6 % to 13 %. The bitter fraction was found to have a molecular weight lower than 5 K. In addition, several organic solvents such as hexane, toluene, methyl acetate, ethyl acetate, ethanol, 2-butanol, 1-propanol and 2propanol were used to extract the bitter peptides. It was found that 2-butanol was the best solvent because it extracted majority of the bitter taste without significantly reducing the product yield. For reversed phase chromatography, a linear gradient of H2O - 0.1 % trifluoroacetic acid (TFA) and CH3CN - 0.1 % TFA was used and fractions were collected at 0 - 20, 20 - 30, 30 - 40, 40 - 50, and > 50 % acetonitrile. Results showed that 30 - 40% CH3-CN fraction contained most of the bitter compounds. Even though these results showed promising strategy to remove bitterness in the hydrolysates, the complexity of the method made its application less favourable for large scale commercial processing. In the study done by Cho and others (2004), commercial soy protein hydrolysates, Supro 710 and FP 900, were used to prepare the peptide solution for bitterness evaluation. A 10 % soy protein hydrolysate solution was prepared, heated, centrifuged and the soluble fractions were subjected to further peptide isolation procedures. The soluble fraction was fractionated by ultrafiltration with molecular weight cutoffs (MWCO) at 1, 3, 5 and 10 K to produce 5 fractions, specifically <1K, 1-3 K, 3-5 K, 5-10K and > 10 K fractions. These fractions were freeze-dried and were subjected to sensory evaluation. The fractions for Supro 710 were prepared at 1.0 % protein solution while the ones for FP 900 were at 0.15 % protein. Caffeine solutions with concentration 0, 200, 500 and 1000 ppm were used as standards and the bitterness of the samples were evaluated as caffeine equivalent (CE) value. The peptide fractions from ultrafiltration, the whole protein samples and the soluble protein samples were all analyzed using High Performance Liquid Chromatography (HPLC) and amino acid analysis to obtain the corresponding molecular weight profiles and amino acid composition. Ney's Q value was also calculated to  36  correlate the hydrophobicity of the amino acids to hydrolysate bitterness. It was observed that UF fractions isolated from FP 900 had much stronger intensity of bitterness ranging from 1340 - 5200 ppm CE than those from Supro 710 (274 - 787 ppm CE) due to the fact that FP 900 achieved higher DH resulting in more peptides with lower molecular weights. The 5 - 10 K fraction of Supro 710 contained the strongest bitterness (787 ppm CE) and it was significantly more bitter than the parent whole Supro 710 which was rated 307 ppm CE. The most bitter fraction of FP 900 was also in the 5 - 10 K fraction. Its bitterness was given a score of 5200 ppm, which was significantly higher than 3720 ppm CE of the parent FP 900. The amino acid compositions of all UF fractions for both Supro 710 and FP 900 were obtained. Most UF fractions for both hydrolysates contained mostly proteins except for the < 1 K fraction. The amino acid profiles for fractions from Supro 710 did not vary. For FP 900, there was a stronger effect of molecular weight on amino acid profiles. The amount of some amino acids, namely Glu, Arg, Cys, His, Lys and Pro, decreased as molecular weight of the fraction decreased while others, like Ala, He, Leu, Phe, Ser, Thr, Tyr and Val, showed an increase with a decrease in molecular weight. It was interesting to observe that even though a large number of hydrophobic amino acids were found in the smaller molecular weight fractions yet, bitterness was low. Further, Qvalues were calculated for each UF fractions of both Supro 710 and FP 900. It was shown that all Q-values had values smaller than 1200 cal/mol, which did not show correlation to the bitterness level in the hydrolysate. Due to the labour-intensive nature of sensory evaluation, efforts have been continuously made to explore other means to evaluate bitterness in peptides. Recently, new methodologies have been proposed to study the quantitative structure-activity relationship of bitter peptides. Kim and Li-Chan (2006a) explored the use of partial least squares regression analysis to correlate the properties or the position of the primary sequence of the peptides with their bitterness. In another study, they looked at the use of Fourier transform Raman spectroscopy in predicting bitterness in peptides (Kim and Li-Chan 2006b).  37  1.4.2  Methods of Debittering of Minimizing Bitterness  Different methods which have been adapted to remove, reduce or mask the level of bitterness include treatment with activated carbon (FitzGerald and O'Cuinn 2006), azeotropic extraction (Lalasidis and Sjoberg 1978; FitzGerald and O'Cuinn 2006), chromatographic separation (Lalasidis and Sjoberg 1978; FitzGerald and O'Cuinn 2006), acetylation of bitter amino acids (Tamura and others 1990; Yeom and others 1994), masking of bitterness (Tamura and others 1990; FitzGerald and O'Cuinn 2006) and enzymatic treatment (Izawa and others 1997; Minagawa and others 1989). 1.4.2.1  Treatment with Activated Carbon, Azeotropic Extraction and Chromatographic Separation  Activated carbon acts as a hydrophobic adsorbent binding hydrophobic peptides and amino acids (Pedersen 1994). Since hydrophobic amino acids and peptides are suggested to produce bitterness, this binding mechanism will significantly suppress the bitter taste of the overall hydrolysate products. However, this treatment results in a great loss of essential amino acids such as Trp, Phe and Arg so supplementation is required to retain the nutritional quality of the hydrolysates (Saha and Hayashi 2001; Liu and others 2004). Azeotropic extraction involves the use of two or more solvents to separate the bitter tasting peptides from the rest. The most commonly used solvents are secondary butyl alcohol, aqueous ethanol, or aqueous isopropyl alcohol and the alcohol phase usually contains most of the bitter peptides (Lalasidis and Sjoberg 1978). Nevertheless, this extraction method not only uses a large amount of solvent which increases cost and hazards, but reduces the overall product yield due to the removal of the bitter fractions. Hydrophobic interaction chromatography (HIC) separates by hydrophobicity, which has been proposed to contribute to bitterness in peptides. HIC is made of resin that will interact with hydrophobic and aromatic substances, thus effectively delays the elution of  38  bitter fractions (Saha and Hayashi 2001). Nonetheless, this method is not practical in an industrial setting due to its complexity. It not only requires an installation of an extremely large column for the massive separation of peptides, but also demands time and labour. 1.4.2.2  Masking of Bitterness  Masking of bitterness, on the other hand, is a fairly simple method. It involves the addition of different substances, such as cc-cyclodextrin, monosodium glutamate, polyphosphates, gelatin, and glycine etc., to the hydrolysates to mask the bitter taste (Pedersen 1994; Saha and Hayashi 2001). Further, adding sweet compounds can effectively mask the bitterness but the end product usually carries a strong sweet taste (Walters 1996). Therefore, depending on the food application, the use of additives to mask bitterness may bring in unfavourable sensory attribute. 1.4.2.3  Enzymatic Treatment and Acetylation of Bitter Amino Acids  Enzymatic treatment involves the use of exogenous enzymes to aid in producing nonbitter hydrolysates. Some enzymes, for example, aminopeptidase, alkaline and neutral protease, carboxypeptidase, have been suggested to reduce the bitterness of protein hydrolysates (Saha and Hayashi 2001). Studies performed by Minagawa and others (1989) and Izawa and others (1997) explored the use of aminopeptidase to lower the bitterness in protein hydrolysates. In a more recent study by Deng and others (2004), the use of polyenzymatic method has been adapted to the production of protein hydrolysate using offal protein from Harengula Zunasi Bleeker. The poly-enzymatic method involves the use of two or more enzymes to hydrolyze the protein and a combination of endopeptidases and exopeptidases are usually used. Endopeptidases cleave peptide bonds ih the interior part of the chain; exopeptidases, on the other hand, prefer to hydrolyze terminal peptide bonds (Anson and others 1966). This combination methodology will ensure a considerable degree of hydrolysis and a selective cleavage of hydrophobic amino acids by specific  39  exopeptidases; therefore, the resulting hydrolysate products will contain minimal bitter tastes. Acetylation of bitter amino acids involves chemical modification of the amino acid. Since specific amino acids have been suggested to be bitter, chemically altering the structure of these groups can reduce bitterness. Yeom and others (1994) have studied the use of acetylation to modify the lysine residues to reduce bitterness of soy protein hydrolysates while Tamura and others (1990) examined the effect of using acetyl, aspartyl and glutamyl residue to block the amino groups of valine, leucine and isoleucine on reduction of bitterness. Both studies showed promising results indicating the potential of using these methods to reduce bitterness in protein hydrolysates.  40  1.5  Peptide Fractionation  In order to isolate bitter peptides from the hydrolysate mixture, a series of fractionation processes has to be carried out. As shown in Table 1.7, different strategies such as electrophoresis and chromatographic separation have been developed for fractionating proteins and peptides (Issaq and others 2002; Stasyk and Huber 2004). 1.5.1  Fractionation by Size  Proteins and peptides are usually fractionated by size followed by other types of separation. Size exclusion chromatography, ultrafiltration as well as gel electrophoresis are methods available to separate by size. Size exclusion chromatography, most frequently used for analytical purpose, can give a molecular weight profile of a sample. This technique involves the use of small polymer particles containing a network of uniform pores for which the small molecules have to pass through, thus slowing down the elution of small substances (Skoog and others 1998). Similarly, gel electrophoresis is mainly used to determine molecular weight of unknown samples or to confirm if the sample of known molecular weight is isolated. It requires a small sample size and the small substances will move faster towards the electrode producing the weight separation. Also, separation can be completed within a few hours so it can help to establish methodology for peptide isolation using other techniques if the molecular weight range of the samples is unknown. On the contrary, ultrafiltration is normally used as a preparative means. Different MWCO membranes can be used for different applications. MWCO membranes of 10 K, 5 K, 3 K and 1 K were used to fractionate the bitter peptides in the soy protein hydrolysates (Cho and others 2004) while 10 K, 5 K and 0.5 K MWCO membranes were used to separate the peptides in the peptic hemoglobin hydrolysates (Aubes and others 1995). This technique fractionates the samples in bulk quantity into different molecular weight range  41  Table 1.7  Fractionation methods and their separation characteristics summarized by Issaq and others (2002) and Stasyk and Huber (2004).  Fractionation Method  Separation Characteristics  Ultracentrifugation  Density  Size-Exclusion Chromatography  Size  Hydrophobic Interaction Chromatography  Hydrophobicity  Reversed-phase Chromatography  Hydrophobicity  Ion-exchange Chromatography  Charge  Affinity Chromatography  Binding to particular substances  Ultrafiltration  Size  Gel Electrophoresis  Size or charge  Isoelectric Focusing  Isoelectric point  42  fractions and hence, it is suitable for large-scale production. 1.5.2  Fractionation by Hydrophobicity  In terms of fractionating bitter substances, separation by hydrophobicity is commonly used because hydrophobic substances have been suggested to cause bitter taste. Several techniques, such as reversed phase high performance liquid chromatography (RP-HPLC) and hydrophobic interaction chromatography (HIC), can be used for this purpose. Reversed-phase chromatography uses a non-polar stationary phase and a relatively polar mobile phase (Skoog and others 1998). In this case, the hydrophobic substances will interact with the stationary phase thus stay in the column for a longer time. Therefore, polar materials will elute first and hydrophobic substances will come out last. Hydrophobic interaction chromatography, likewise, separates hydrophobic substances by their interactions with chromatographic sorbents of hydrophobic nature (Harris 1982). This method is more suitable for food-grade purpose because organic solvents are usually used with RP-HPLC. RP-HPLC has been used to separate bitter peptides in Ragusano cheese (Fallico and others 2005), cheese (Lee and Warthesen 1996) and cheese made from pasteurized and raw milk (Gomez and others 1997). 1.5.3  Fractionation using Two-Dimensional Methods  Two-dimensional gel electrophoresis has been available for the proteome research. It involves first separating the proteins or peptides by their isoelectric point (pi) and then further separation by size in the second dimension. The resolved bands can then be analyzed by mass spectrometry (MS) to identify the proteins and peptides (Issaq and others 2002). However, this technique has its drawbacks such as the complexity of isolating individual band from the 2-D gel for MS analysis and the inability to separate samples with extreme pi and molecular weight. Recently, two-dimensional chromatography has appeared which seems to be superior to the 2-D gel technique. The 2-dimensional chromatographic method involves fractionating in the first dimension by  43  means of charge or size and second dimension by hydrophobicity (Stasyk and Huber 2004). The second column, usually reversed phase column, is coupled to a tandem mass spectrometer to identify the peptides. The disadvantage of this method, in comparison to the 2-D gel electrophoretic technique, is its cost of operation. Many researchers have adapted some of these techniques in fractionating peptides from proteins and protein hydrolysates of different materials. Some of their strategies are summarized ih Table 1.8.  44  Table 1.8  Some examples of published fractionation methodology for isolation of bitter peptides.  Fractionation steps  Equipment  Solvent  Ultrafiltration  MWCO at 1, 3, 5 and 10 K  High Performance Liquid Chromatography  TSKG3000SWXL column and GPC Peptide column  6M guanidine hydrochloride with dithiothreitol in 0.1M phosphate buffer  Gel-Permeation High Performance Liquid Chromatography  Superdex Peptide HR 10/30 column  Distilled water  Reversed Phase High Performance liquid Chromatography  Supelcosil Cig column  0.1 % HCI in water; 0.1 % HCI and 90 % ethanol in water  Reversed Phase High Performance liquid Chromatography  Delta-Pak Cig column  0.1 %HC1 in 1 % EtOH; 0.1 % HCI in 90% EtOH  Reversed Phase High Performance liquid Chromatography  TSK ODS 80™ column  0.1 %HC1 in 1 % EtOH; 0.1 %HC1 in 70% EtOH  Substrate  Reference  Soy protein hydrolysates  Cho and others 2004  Hydrolysate of soybean US glycinin  Kim and others 2003  Fractionation steps  Gel Permeation High Performance Liquid Chromatography  20 mM potassium phosphate Superdex peptide HR 10/30 buffer containing 250 mM NaCl column  Supelcosil semi-preparative Reversed Phase Cig column Chromatography High Performance Liquid Chromatography Reversed Phase Chromatography Reversed Phase-High Performance Liquid Chromatography  Ultrafiltration  TSK ODS 80TM analytical column  4^  Oy  Reference  Soybean Proglycinin Subunit  Kim and others 1999  Cheese  Lee and Warthesen 1996  Hemoglobin hydrolysates  Aubes-Dufau and others 1995  0.1% HCI in 1% EtOH; 0.1% HCI in 70% EtOH  18  YM10, YM5, YC05 2-butanol, 1-propanol, 2-propanol, ethanol Lichroprep RP18  Substrate  0.1% HCI in water; 0.1% HCI and 90% ethanol in water  Vydac C Model 201TP54 0.1 % trifluoroacetic acid (TFA) in water; 90 % acetonitrile and analytical column 0.1 % TFA in water  Organic solvent extraction Reversed Phase Chromatography  Solvent  Equipment  H O-0.1%TFA; CH CN-0.1%TFA 2  3  1.6  Hypothesis and Objectives  The overall objective of this thesis is to produce protein hydrolysate samples from shrimp processing waste using the Taguchi's experimental design Lie (4 ), to investigate if 5  water-to-substrate ratio, percent enzyme, hydrolysis time and/or type of protease have any effect on product yield, degree of hydrolysis as well as bitterness level. In addition, the characteristics of bitter substances will be examined using fractionation by size and hydrophobicity followed by amino acid analysis to provide potential information to establish methodology to reduce or completely eliminate bitterness in shrimp waste protein hydrolysates. Therefore, the hypotheses of this thesis are 1. Peptides are responsible for the bitter taste in the shrimp waste hydrolysates. 2. There is ah optimal condition from the experimental conditions examined in this study at which shrimp waste hydrolysates can be produced with high product yield but low or no bitterness in comparison to the controls. 3. Bitter peptides in shrimp waste hydrolysates are in the molecular weight range of 1 - 6 kDa. 4. Bitter peptides in shrimp waste hydrolysates are strongly hydrophobic. To test these hypotheses, the objectives of the study are to: a. Hydrolyze the shrimp processing waste using conditions designated using a 4factor 4-level Taguchi's Li6 (4) experimental design to produce shrimp waste 5  hydrolysates.  47.  b. Monitor the product yield (both soluble and insoluble), degree of hydrolysis and bitterness level of each hydrolysate sample to examine if the four selected factors for the experimental design have any effects on these parameters. c. Fractionate and isolate the bitter fractions using several separation techniques including ultrafiltration (UF), hydrophobic interaction chromatography (HIC) and gel filtration (GF) to reveal characteristic property and molecular size of bitter peptides of the selected sample. d. Analyze selected soluble shrimp waste hydrolysate samples and bitter fractions by amino acid analysis to examine if certain amino acid residues dominate in the hydrolysates containing strong bitterness.  48  .7  References  Aaslyng MD, Martens M, Poll L, Nielsen PM, Flyge H, Larsen LM. 1998. Chemical and sensory characterization of hydrolyzed vegetable protein, a savory flavoring. Journal of Agricultural and Food Chemistry 46: 481-489. Adler-Nissen J. 1979. Determination of the degree of hydrolysis of food protein hydrolysates by trinitroberizenesulfonic acid. Journal of Agricultural and Food Chemistry 27(6): 1256-1262. Adler-Nissen J. 1986. Enzymic hydrolysis of food proteins. New York: Elsevier. 427p. Anson ML, Chichester CO, Mrak EM, Stewart GF. 1966. Enzymes in food processing. New York: Academic Press. 483p. Arai S, Noguchi M, Kurosawa S, Kato H, Fujimaki M. 1970. Applying proteolytic enzymes on soybean. 6. Deodorization effect of Aspergillopeptidase A and debittering effect of Aspergillus acid carboxypeptidase. Journal of Food Science 35: 392-395. Asao M, Iwarhura H, Akamatsu M, Fujita T. 1987. Quantitative structure-activity relationships of the bitter thresholds of amino acids, peptides, and their derivatives. Journal of Medicinal Chemistry 30: 1873-1879. Aspmo SI, Horn SJ, Eijsink VGH. 2005. Enzymatic hydrolysis of Atlantic cod (Gadus morhua L.) viscera. Process Biochemistry 40: 1957-1966. Aubes-Dufau I, Combes D. 1997. Effect of different proteases on bitterness of hemoglobin hydrolysates. Applied Biochemistry and Biotechnology 67: 127138. Aubes-Dufau I, Seris JL, Combes D. 1995. Bitter peptide from hemoglobin hydrolyzate: isolation and characterization. FEBS Letters 364: 115-119. Aubes-Dufau I, Seris JL, Combes D. 1995. Production of peptic hemoglobin hydrolysates: bitterness demonstration and characterization. Journal of Agricultural and Food Chemistry 43: 1982-1988. Balogun AM, Akegbejo-Samsons Y. 1992. Waste yield, proximate and mineral composition of shrimp resources of Nigeria's coastal waters. Bioresource Technology 40: 157-161. Beddows CG, Ismail M, Steinkraus KH. 1976. The use of bromelain in the hydrolysis of mackerel and the investigation of fermented fish aroma. Journal of Food Technology 11: 379-388. 49  Benjakul S, Morrissey MT. 1997. Protein hydrolysates from Pacific whiting solid wastes. Journal of Agricultural and Food Chemistry 45: 3423-3430. Cancre I, Ravallec R, Wormhoudt AV, Stenberg E, Gildberg A, Gal YL. 1999. Secretagogues and growth factors infishand crustacean protein hydrolysates. Marine Biotechnology 1: 489-494. Cavalheiro JMO, deSouza EO, Bora PS. 2007. Utilization of shrimp industry waste in the formulation of tilapia (Oreochromis niloticus Linnaeus) feed. Bioresource Technology 98: 602-606. Charalambous G. 1980. The analysis and control of less desirable flavors in foods and beverages. New York: Academic Press. 358p. Charteris W. 1992. Taguchi's system of experimental design and data analysis: a quality engineering technology for the food industry. Journal of the Society of Dairy Technology 45: 33-49. Cho MJ, Unklesbay N, Hsieh FH, Clarke AD. 2004. Hydrophobicity of bitter peptides from soy protein hydrolysates. Journal of Agricultural and Food Chemistry 52: 5895-5901. Chobert MJ, Sitohy MZ, Whitaker JR. 1988. Solubility and emulsifying properties of casein modified enzymatically by Staphylococcus aureus V8 protease. Journal of Agricultural and Food Chemistry 36: 220-224. Clegg KM, Lim CL, Manson W. 1974. The structure of a bitter peptide derived from casein by digestion with papain. Journal of Dairy Research 41: 283-287. Clemente A. 2000. Enzymatic protein hydrolysates in human nutrition. Trends in Food Science and Technology 11: 254-262. Clemente A, Vioque J, Sanchez-Vioque R, Pedroche J, Bautista J, Millan F. 1999. Protein quality of chickpea (Cicer arietinum L.) protein hydrolysates. Food Chemistry 67: 269-274. Coward-Kelly G, Agbogbo FK, Holtzapple MT. 2006. Lime treatment of shrimp head waste for the generation of highly digestible animal feed. Bioresource Technology 97: 1515-1520. Deeslie WD, Cheryan M. 1988. Functional properties of soy protein hydrolysates from a continuous ultrafiltration reactor. Journal of Agricultural and Food Chemistry 36: 26-31. Deng SG, Peng ZY, Chen F, Yang P, Wu T. 2004. Amino acid composition and antianaemia action of hydrolyzed offal protein from Harengula Zunasi Bleeker. Food Chemistry 87: 97-102.  50  Diaz-Rojas EI, Arguelles-Monal WM, Higuera-Ciapara I, Hernandez J, LizardiMendoza J, Goycoolea FM. 2006. Determination of chitin and protein contents during the isolation of chitin from shrimp waste. Macromolecular Bioscience 6: 340-347. Diniz FM, Martin AM. 1996. Use of response surface methodology to describe the combined effects of pH, temperature and E/S ratio on the hydrolysis of dogfish (Squalus acanthias) muscle. International Journal of Food Science and Technology 31: 419-426. Doucet D, Otter DE, Gauthier SF, Foegeding EA. 2003. Enzyme-induced gelation of extensively hydrolyzed whey proteins by alcalase: peptide identification and determination of enzyme specificity. Journal of Agricultural and Food Chemistry 51: 6300-6308. Dumay J, Donnay-Moreno C, Barnathan G, Jaouen P, Berge JP. 2006. Improvement of lipid and phospholipids recoveries from sardine (Sardina pilchardus) viscera using industrial proteases. Process Biochemistry 41: 2327-2332. Fallico V, McSweeney PLH, Home J, Pediliggieri C, Harmon JA, Carpino S, Licitra G. 2005. Evaluation of bitterness in Ragusano cheese. Journal of Dairy Science 88: 1288-1300. Ferrer J, Paez G, Marmol Z, Ramones E, Garcia H, Forster CF. 1996. Acid hydrolysis of shrimp-shell wastes and the production of single cell protein from the hydrolysate. Bioresource Technology 57: 55-60. Fisheries and Oceans Canada. 2007. Coastal and Water Statistics. Fast Facts http://www.dfo-mpo.gc.ca/cornmunic/facts-info/facts-info_e.htm FitzGerald RJ, O'Cuinn G. 2006. Enzymatic debittering of food protein hydrolysates. Biotechnology Advances 24: 234-237. Frank ME, Bouverat BP, MacKinnon Bl, Hettinger TP. 2004. The distinctiveness of ionic and nonionic bitter stimuli. Physiology and Behavior 80: 421-431. Frokjaer S. 1994. Use of hydrolysates for protein supplementation. Food Technology 48: 86-88. Fujimaki M, Kato H, Arai S, Tamaki E. 1968. Applying proteolytic enzymes on soybean. 1: Proteolytic enzyme treatment of soybean protein and its effect on the flavor. Food Technology 22: 889-893. Fujimaki M, Kato H, Arai S, Yamashita M. 1971. Application of microbial protease to soybean and other materials to improve acceptability, especially through the formation of plastein. Journal of Applied Bacteriology 34: 119-131.  51  Fujimaki M, Yamashita M, Okazawa Y, Arai S. 1970. Applying proteolytic enzymes on soybean. 3. Diffusable bitter peptides and free amino acids in peptic hydrolyzate of soybean protein. Journal of Food Science 35(3): 215-218. Fuke S. 1994. Taste-active components of seafoods with special reference to umami substances. In: Shahidi F, Botta JR, editors. Seafoods: chemistry, processing technology and quality. 1 ed. Glasgow: Blackie Academic and Professional, p 115-139. st  Gallagher J, Kanekanian AD, Evans EP. 1994. Hydrolysis of casein: a comparative study of two proteases and their peptide maps. International Journal of Food Science and Technology 29: 279-285. Gilbertson TA, Damak S, Margolskee RF. 2000. The molecular physiology of taste transduction. Current Opinion in Neurobiology 10: 519-527. Gildberg A, Stenberg E. 2001. A new process for advanced utilisation of shrimp waste. Process Biochemistry 36: 809-812. Goldsmith P, Amankwah F, Gunjal K, Smith J. 2003. Financial feasibility of producing value-added seafood from shrimp waste in Quebec. Journal of Aquatic Food Product Technology 12(4): 39-61. Gomez MJ, Garde S, Gaya P, Medina M, Nunez M.T997. Relationship between level of hydrophobic peptides and bitterness in cheese made from pasteurized and raw milk. Journal of Dairy Research 64: 289-297. Gorden DT, Roberts GL. 1977. Mineral and proximate composition of Pacific coast fish. Journal of Agricultural and Food Chemistry 25(6): 1262-1268. Guerard F, Dufosse L, Broise DDL, Binet A. 2001. Enzymatic hydrolysis of proteins from yellowfin tuna (Thunnus albacares) wastes using Alcalase. Journal of Molecular Catalysis B:Enzymatic 11: 1051-1059. Guerard F, Guimas L, Binet A. 2002. Production of tuna waste hydrolysates by a commercial neutral protease preparation. Journal of Molecular Catalysis B: Enzymatic 19-20: 489-498. Hamilton JS, Hill RD, van Leeuwen H. 1974. A bitter peptide from cheddar cheese. Agricultural and Biological Chemistry 38: 375-379. Harris DC. 1982. Quantitative chemical analysis. 5th ed. United States of America: W.H. Freeman and Company. 899p. He H, Chen X, Sun C, Zhang Y, Gao P. 2006. Preparation and functional evaluation of oligopeptide-enriched hydrolysate from shrimp (Acetes chinensis) treated with crude protease from Bacillus sp. SM98011. Bioresource Technology 97: 385-390. 52  deHolanda HD, Netto FiVI. 2006. Recovery of components from shrimp (Xiphopendeus Kroyeri) processing waste by enzymatic hydrolysis. Journal of Food Science 71(5): 298-303. Huber L, Klostermeyer H. 1974. Isolation and identification of a bitter peptide from Butterkaese cheese. Milchwissenschaft 29: 449-455. Ichikawa K, Yamamoto T, Fukumoto J. 1959. Studies on the bitter taste peptides produced by proteinases. 1. The formation of bitter taste peptides by "neutral proteinase" subtilis and isolation of the peptides. Nippon Nogeikagaku Kaishi 33: 1044-1048. Imm JY, Lee CM. 1999. Production of seafood flavor from red hake (Urophycis chuss) by enzymatic hydrolysis. Journal of Agricultural and Food Chemistry 47: 2360^23666. Ishibashi N, Ono I, Kato K, Shigenaga T, Shinoda I, Okai H, Fukui S. 1988a. Role of the hydrophobic amino acid residue in the bitterness of peptides. Agricultural and Biological Chemistry 52(1): 91-94. Ishibashi N, Kouge K, Shinoda I, Kanehisa H, Okai H. 1988b. A mechanism for bitter taste sensibility in peptides. Agricultural and Biological Chemistry 52(3): 819827. Ishibashi N, Sadamori K, Yamamoto O, Kanehisa H, Kouge K, Kikuchi E, Okai H, Fukui S. 1987. Bitterness of phenylalanine- and tyrosine-containing peptides. Agricultural and Biological Chemistry 51(12): 3309-3313. Islam MS, Khan S, Tanaka M. 2004. Waste loading in shrimp and fish processing effluents: potential source of hazards to the coastal and nearshore environments. Marine Pollution Bulletin 49: 103-110. Issaq HJ, Conrads TP, Janini GM, Veenstra TD. 2002. Methods for fractionation, separation and profiling of proteins and peptides. Electrophoresis 23: 30483061. Izawa N, Tokuyasu K, Hayashi K. 1997. Debittering of protein hydrolysates using Aeromonas caviae aminopeptidase. Journal of Agricultural and Food Chemistry 45(3): 543-545. Jaswal AS. 1989. Methodology investigations for the production of amino acid hydrolysate from shrimp waste. Canadian Institute of Food Science and Technology Journal 22(5): 460-463. Jost R, Monti JC. 1977. Partial enzymatic hydrolysis of whey protein by trypsin. Journal of Dairy Science 60: 1387-1393.  53  Kim MR, Choi SY, Kim CS, Kim CW, Utsumi S, Lee CH. 1999. Amino acid sequence analysis of bitter peptides from a soybean proglycinin subunit synthesized in Escherichia coli. Bioscience, Biotechnology, and Biochemistry 63(12): 2069-2074. Kim MR, Kawamura Y, Lee CH. 2003. Isolation and identification of bitter peptides of tryptic hydrolysate of soybean 11S glycinin by reverse-phase high performance liquid chromatography. Journal of Food Science 68(8): 24162422. Kim H, Li-Chan ECY. 2006a. Quantitative structure-activity relationship study of bitter peptides. Journal of Agricultural and Food Chemistry 54: 10102-10111. Kim H, Li-Chan ECY. 2006b. Application of Fourier transform Raman spectroscopy for prediction of bitterness of peptides. Applied Spectroscopy 60(11): 12971306. Kirimura J, Shimizu A, Kimizuka A, Ninomiya T, Katsuya N. 1969. Journal of Agricultural and Food Chemistry 17(4): 689-695. Kristinsson HG, Rasco BA. 2000. Fish protein hydrolysates: production, biochemical, and functional properties. Critical Reviews in Food Science and Nutrition 40(1): 43-81. Lahl WJ, Braun SD. 1994. Enzymatic production of protein hydrolysates for food use. Food Technology 48: 68-71. Lalasidis G, Sjoberg LB. 1978. Two new methods of debittering protein hydrolysates and a fraction of hydrolysates with exceptionally high content of essential amino acids. Journal of Agricultural and Food Chemistry 26(3): 742-749. Lee JY, Lee HD, Lee CH. 2001. Characterization of hydrolysates produced by mildacid treatment and enzymatic hydrolysis of defatted soybean flour. Food Research International 34: 217-222. Lee KD, Warthesen JJ. 1996. Mobile phases in reverse-phase HPLC for the determination of bitter peptides in cheese. Journal of Food Science 61(2): 291294. Lekshmy NA, Prabhu PV. 1989. Studies on the chemical and nutritional quality of protein powders isolated from shrimp waste. Fishery Technology 26: 56-59. Li-Chan ECY. 2004. Properties of proteins in food systems: an introduction. In: Yada RY, editor. Proteins in food processing. New York: CRC Press LLC. p 2-26. Liaset B, Julshamn K, Espe M. 2003. Chemical composition and theoretical nutritional evaluation of the produced fractions from enzymic hydrolysis of salmon frames with Protamex . Process Biochemistry 38: 1747-1759. M  54  Liaset B, Lied E, Espe M. 2000. Enzymatic hydrolysis of by-products from the fishfilleting industry; chemical characterisation and nutritional evaluation. Journal of the Science of Food and Agriculture 80: 581-589. Lindemann B. 2001. Receptors and transduction in taste. Nature 413: 219-225. Liu CC, Morioka K, Itoh Y, Obatake A. 2000. Contribution of lipid oxidation to bitterness and loss of free amino acids in the auto lytic extract from fish wastes: Effective utilization of fish wastes. Fisheries Science 66: 343-348. Liu CC, Morioka K, Itoh Y, Obatake A. 2004. Debittering autolytic extract from fish wastes and application of the extract in flavouring surimi-based product. Journal of Shanghai Jiaotong University (Agricultural Science) 22(1): 98-103. Maehashi K, Matsuzaki M, Yainamoto Y, Udaka S. 1999. Isolation of peptides from an enzymatic hydrolysate of food proteins and characterization of their taste properties. Bioscience, Biotechnology, and Biochemistry 63(3): 555-559. Mandeville S, Yaylayan V, Simpson BK. 1992a. Proximate analysis, isolation and identification of amino acids and sugars from raw and cooked commercial shrimp waste. Food Biotechnology 6: 51-64. Mandeville S, Yaylayan V, Simpson B. 1992b. GC/MS analysis of flavor-active compounds in cooked commercial shrimp waste. Journal of Agricultural and Food Chemistry 40: 1275-1279. Mandeville S, Yaylayan V, Simpson B, Pare JRJ. 1991. Gas chromatography-mass spectrometry analysis of flavour-active compounds from raw commercial shrimp waste. Spectroscopy 9: 61-72. Margolskee RF. 2002. Molecular mechanisms of bitter and sweet taste transduction. The Journal of Biological Chemistry 277(1): 1-4. Matoba T, Hayashi R, Hata T. 1970. Isolation of bitter peptides in tryptic hydrolysates of casein and their chemical structures. Agricultural and Biological Chemistry 34: 1235-1243. Matoba T, Nagayasu C, Hayashi R, Hata T. 1969. Bitter peptides in the tryptic hydrolyzate of casein. Agricultural and Biological Chemistry 33(11): 16621663. Matsusita I, Ozaki S. 1993. Bitter peptides from beer yeast extracts. Peptide Chemistry 31: 165-168. Minagawa E, Kaminogawa S, Tsukasaki F, Yamauchi K. 1989. Debittering mechanism in bitter peptides of enzymatic hydrolysates from milk casein by aminopeptidase T. Journal of Food Science 54(5): 1225-1229.  55  Minamiura N, Matsumura Y, Yamamoto T. 1972. Bitter peptides in the casein digests with bacterial proteinase: II. Bitter peptide consisting of tryptophan and leucine. Journal of Biochemistry (Tokyo) 72: 841-848. Ministry of Agriculture and Lands. 2004. Seafood processing. Fisheries Statistics http://www.agf.gov.bc.ca/fish_stats/statistics.htm Mizani M, Aminlari M, Khodabandeh M. 2005. An effective method for producing a nutritive protein extract powder from shrimp-head waste. Food Science and Technology International 11(1): 49-54. Montmayeur JP, Matsunami H. 2002. Receptors for bitter and sweet taste. Current Opinion in Neurobiology 12: 366-371. Morato AF, Carreira RL, Junqueira RG, Silvestre MPC. 2000. Optimization of casein hydrolysis for obtaining high contents of small peptides: use of subtilisin and trypsin. Journal of Food Composition and Analysis 13: 843-857. Mukhin VA and Novikov VY. 2001. Enzymatic hydrolysis of proteins from crustaceans of the Barents Sea. Applied Biochemistry and Microbiology 37(5): 538-542. Ney KH. 1979. Bitterness of peptides: amino acid composition and chain length. In: Boudreau JC, Editor. Food Taste Chemistry, ACS Symposium Series vol. 115. Washington, DC: American Chemical Society, p 149-173. Nilsang S, Lertsiri S, Suphantharika M, Assavanig A. 2005. Optimization of enzymatic hydrolysis of fish soluble concentrate by commercial proteases. Journal of Food Engineering 70: 571-578. Nishitoba T, Sato H, Sakamura S. 1988. Bitterness and structure relationship of the triterpenoids from Ganoderma lucidum (Reishi). Agricultural and Biological Chemistry 52(7): 1791-1795. Novozymes A/Z. 2001a. Alcalase® 2.4 L FG product sheet. Denmark: Novozymes. Novozymes A/Z. 2001b. Flavourzyme® 1000 L product sheet. Denmark: Novozymes. Novozymes A/Z. 2001c. Protamex® product sheet. Denmark: Novozymes. Nutriteck. 2005. Bromelain product and ingredients information. Canada: Ultra BioLogics Inc. Pedersen B. 1994. Removing bitterness from protein hydrolysates. Food Technology 48: 96-98.  56  Pedroche J, Yust MM, Lqari H, Gifon-Calle J, Vioque J, Alaiz M, Millan F. 2004. Production and characterization of casein hydrolysates with a high amino acid Fischer's ratio using immobilized proteases. International Dairy Journal 14: 527-533. Pommer K. 1995. New proteolytic enzymes for the production of savory ingredients. Cereal Foods World 40: 745-748. Ramos de Armas R, Diaz HG, Molina R, Gonzalez MP, Uriarte E. 2004. Stochasticbased descriptors studying peptides biological properties: modeling the bitter tasting threshold of dipeptides. Bioorganic and Medicinal Chemistry 12: 4815-4822. Ravallec-Ple R, Gilmartin L, Wormhoudt AV, Gal YL. 2000. Influence of the hydrolysis process on the biological activities of protein hydrolysates from cod (Gadus morhud) muscle. Journal of the Science of Food and Agriculture 80: 2176-2180. Rebeca BD, Pena-Vera MT, Diaz-Castaneda M. 1991. Production of fish protein hydrolysates with bacterial proteases: yield and nutritional value. Journal of Food Science 56(2): 309-314. Ruttanapornvareesakul Y, Ikeda M, Hara K, Osatomi K, Osako K, Kongpun O, Nozaki Y. 2006. Concentrations-dependent suppressive effect of shrimp head protein hydrolysate on dehydration-induced denaturation of lizardfish myofibrils. Bioresource Technology 97: 762-769. Saha BC, Hayashi K. 2001. Debittering of protein hydrolyzates. Biotechnology Advances 19: 355-370. Schmidl MK, Taylor SL, Nordlee JA. 1994. Use of hydrolysate-based products in special medical diets. Food Technology 48: 77-80,85. Scott K. 2004. The sweet and the bitter of mammalian taste. Current Opinion in Neurobiology 14: 423-427. Shahidi F. 1994. Seafood processing by-products. In: Shahidi F, Botta JR, editors. Seafoods: chemistry, processing technology and quality. 1 ed. Glasgow: Blackie Academic and Professional, p. 320-334. st  Shahidi F, Han XQ, Synowiecki J. 1995. Production and characteristics of protein hydrolysates from capelin (Mallotus villosus). Food Chemistry 53: 285-293. Shahidi F, Synowiecki J. 1991. Isolation and characterization of nutrients and valueadded products from snow crab (Chinoecetes opilio) and shrimp (Pandalus borealis) processing discards. Journal of Agricultural and Food Chemistry 39: 1527-1532.  57  Simpson BK, Nayeri G, Yaylayan V, Ashie INA. 1998. Enzymatic hydrolysis of shrimp meat. Food Chemistry 61: 131-138. Skoog DA, Holler FJ, Nieman TA. 1998. Principles of instrumental analysis. 5th ed. Philadelphia: Saunders College Pub.: Harcourt Brace College Publishers. 849p. Smyth M, FitzGerald RJ. 1998. Relationship between some characteristics of WPC hydrolysates and the enzyme complement in commercially available proteinase preparations. International Dairy Journal 8: 819-827. Spellman D, O'Cuinn G, FitzGerald RJ. 2005. Physicochemical and sensory characteristics of whey protein hydrolysates generated at different total solids levels. Journal of Dairy Research 72: 138-143. Stasyk T, Huber LA. 2004. Zooming in: fractionation strategies in proteomics. Proteomics 4: 3704-3716. Stevenson DE, Ofman DJ, Morgan KR, Stanley RA. 1998. Protease-catalyzed condensation of peptides as a potential means to reduce the bitter taste of hydrophobic peptides found in protein hydrolysates. Enzyme and Microbial Technology 22: 100-110. Synowiecki J, Al-Khateeb NAAQ. 2000. The recovery of protein hydrolysate during enzymatic isolation of chitin from shrimp Crangon crangon processing discards. Food Chemistry 68: 147-152. Takahashi M, Nakata T, Nakatani M, Kataoka S, Nakamura K, Okai H. 1995. Conversion of bitterness of C-terminal'Octapeptide of bovine P-casein (ArgGly-Pro-Phe-Pro-Ile-Ile-Val) into sweetness. Peptide Chemistry 32: 281-284. Tamura M, Mori N, Miyoshi T, Koyama S, Kohri H, Okai H. 1990. Practical debittering using model peptides and related compounds. Agricultural and Biological Chemistry 54(1): 41-51. Walters DE. 1996. How are bitter and sweet tastes related? Trends in Food Science and Technology 7: 399-406. Wieser H, Belitz HD. 1975. Bitter peptides isolated from corn protein zein by hydrolysis with pepsin. Zeitschrift fur Lebensmitteluntersuchung und Forschung A 159: 329-336. World Health Organization. 1985. Energy and protein requirements.World Health Organization Technical Report Series 724 http://www.fao.org/docrep/003/aa040e/AA040E00.HTM  58  Yamashita M, Arai S, Fujimaki M. 1969. Applying proteolytic enzymes on soybean. 4. A ninhydrin-negative bitter peptide in peptic hydrolyzate of soybean protein. Agricultural and Biological Chemistry Japan 33: 321-330. Yeom HW, Kim KS, Rhee JS. 1994. Soy protein hydrolysate debittering by lysineacetylation. Journal of Food Science 59(5): 1123-1126. Zhang Y, Hoon MA, Chandrashekar J, Mueller KL, Cook B, Wu D, Zuker CS, Ryba NJP. 2003. Coding of sweet, bitter, and umami tastes: different receptor cells sharing similar signaling pathways. Cell 112: 293-301.  59  C H A P T E R II  Use of Taguchi's Experimental Design to Investigate the Hydrolysis Conditions for Production of Protein Hydrolysates from Shrimp (Pandalopsis dispar) Processing Discards  2.1  1  Introduction  The production of protein hydrolysates have become very popular over the past decades because of their improved nutritive value, enhanced functional properties and potential biological activity. Protein hydrolysates are produced from hydrolysis of protein molecules with acids, bases or enzymes and the end products contain shorter peptides and/or single amino acids. These short peptides and single amino acids have been suggested to potentially contain biological activity, reduce allergenicity of the original protein molecules and increase digestibility for those suffering from absorption problems (Frokjaer 1994; Schmidl and others 1994; Clemente 2000). Because of all the advantages they hold, over the years, there has been continual research to explore the use of different sources to produce protein hydrolysates. As mentioned in Section 1.1,  casein, whey and soy are some of the most extensively studied materials  while recently, many under-utilized food sources such as fish waste and shrimp processing discards begin to draw researchers' interest because of their availability as well as low costs. The fisheries and oceans industries are huge in Canada generating revenue of up to two billion dollars annually (Fisheries and Oceans Canada 2007). Shrimps, one of the biggest contributors to this profit among all wild shellfish, are mostly processed. During shrimp processing, approximately 50 - 70 % of the total landings are usually discarded as wastes 1  A version of this chapter may be submitted for publication.  60  (Mandeville and others 1991; Goldsmith and others 2003; Islam arid others 2004). These massive amounts of discards axe normally disposed directly into the ocean or landfill dumping sites, which lead to environmental problems. Recently, a new pollution law has prohibited processors from discarding these materials in such a way (Mandeville and others 1991); therefore, the processing industry needs to pay a high cost to dispose of the wastes properly. In order to reduce disposal cost, it is necessary to investigate methods to utilize these waste materials to produce value-added products. Several attempts have been made to extract chitins (Synowiecki and Al Khateeb 2000; Gildberg and Stenberg 2001; DiazRojas and others 2006) and produce protein hydrolysates for use in functional foods (Jaswal 1989; He and others 2006) and animal feeds (Jaswal 1989; Gildberg and Stenberg 2001; Mizani and others 2005; Coward-Kelly and others 2006; Cavalheiro and others 2007) from shrimp processing wastes. However, most studies have only explored the use of one specific protease and experimental condition to produce protein hydrolysates (Synowiecki and Al-Khateeb 2000; Gildberg and Stenberg 2001; deHolanda and Netto 2006) but several factors, such as pH, temperature, water-to-substrate ratio, amount of enzymes, incubation time and/or the type of protease, have been shown to affect hydrolysis products (Shahidi and others 1995; Diniz and Martin 1996; Simpson and others 1998; Ravallec-Ple and others 2000; Lee and others 2001; Guerard and others 2002; Nilsang and others 2005; Aspmo and others 2005; Spellman and others 2005; Dumay and others 2006). Therefore, in this study, Taguehi's Li6(4 ) experimental design was used to generate 16 5  experimental conditions to explore the effect of four factors on the protein hydrolysates produced from shrimp processing waste. These factors, specifically the water-to-substrate ratio, percent enzyme, time of hydrolysis and type of protease, each comprising four levels, were studied to examine their impact on soluble product yield, degree of hydrolysis as well as bitterness of the final products. Further, amino acid compositions of chosen samples were compared to find out if certain amino acid residues dominated in hydrolysate samples that tasted bitter.  61  2.2  2.2.1  Materials and Methods  Materials  Commercial processing discards (also referred to as shrimp wastes) were composed of shrimp shells, tails and heads from cooked shrimps (Pandalopsis dispar). These shrimps were washed, cooked for 2 minutes at 100 °C, cooled and hand-peeled before the wastes were supplied in frozen form by Albion Fisheries Ltd. (British Columbia, Canada) on May 12, 2005 for the preliminary study and on Feb 17, 2006 for the main thesis study. The shrimp processing waste was stored in a -25 °C walk-in freezer. The frozen shrimp waste was thawed at 4 °C overnight and the wet samples were allocated by weight to 1000-gram packages and stored at -25 °C until use. Four commercial food grade proteases were donated by Neova Technologies Inc. (British Columbia, Canada). Alcalase® 2.4 L FG (protease from Bacillus licheniformis, 2.4 AU/g), Flavourzyme® 1000 L (protease from Aspergillus oryzae, 1000 LAPU/g) and Protamex® (1.5 AU/g, protease from Bacillus amyloliquefaciens and Bacillus licheniformis) were from Novozymes North America Inc. (North Carolina, USA) while  bromelain (protease from pineapple stem, 2000 GDU/g) was from Ultra Bio-Logics Inc. (Quebec, Canada). Trichloroacetic acid (TCA) was purchased from Sigma Aldrich (Ontario, Canada) while disodium tetraborate decahydrate and caffeine were purchased from BDH Inc. (Ontario, Canada). Trinitrobenzenesulfonic acid (TNBS) was purchased from Fluka Biochemika (Ontario, Canada). Petroleum ether, sodium phosphate monohydrate and sodium sulfite were purchased from Fisher Scientific Co. (Ontario, Canada).  62  2.2.2  Sample Preparation  For the preliminary study, thawed shrimp waste samples were blended with 50 °C ddHiO in a 2:1 water-to-substrate ratio using an Osterizer blender (Sunbeam Corporation Ltd.; Ontario, Canada). The waste blend was sieved through a 2-mm mesh to separate the solid mass (referred to as solid waste blend) and the liquid portion (referred to as liquid waste blend). For the thesis study, thawed shrimp wastes were blended with the designated water-to-substrate ratio for two minutes using the Osterizer. 2.2.3  Proximate Analysis of Shrimp Processing Waste  The solid waste blend, liquid waste blend and the wet whole shrimp waste were placed in aluminum pans and dried in a VWR® Vacuum Oven model 1430 (VWR Scientific Products; Ontario, Canada) at 70 °C and 762 mm Hg overnight. The moisture content was determined as the weight difference before and after drying and averaged among six replicates. The dried samples were ground together and three replicates were subjected to dry ashing in a Thermolyne Type F62700 Furnace (Barnstead/Thermolyne Corporation; Iowa, USA) at 570 °C overnight and fat extraction with petroleum ether using the Labconco Goldfish solvent extractor (Labconco Corporation; Missouri, USA). Four replicates were used to determine the total crude protein by 6.25 X the total nitrogen content determined using the Leco FP-428 instrument (Leco Instruments Limited; Michigan, USA). The carbohydrate content was calculated by difference. These methodologies are described in Current Protocols in Food Analytical Chemistry, Volume 1 (Wrolstad and others 2003). 2.2.4  Shrimp Waste Hydrolysate Production  For the preliminary study, liquid waste blend was used in the hydrolysis. Four commercial food grade proteases, namely Alcalase®, bromelain, Flavourzyme® and Protamex®, were used. Liquid waste blend portions, each 300 g, were distributed in glass  63  containers with lids and 800 mg enzymes were added (Table 2.1). The resulting mixtures were placed in a pre-heated 50 °C water bath and were shaken periodically during the reaction. After incubation for 1 hour, the flasks were removed from the water bath. The lids were replaced by aluminum foil and the flasks were heated at 90 °C for 30 minutes in a steam kettle. The whole hydrolysates were centrifuged at 1690Xg for 20 minutes at 4 °C using the Beckman GS-6 Centrifuge with the GH-3.7 Horizontal Rotor (Beckman Instruments Inc.; California, USA). The pH of the supernatant was measured and the supernatants and pellets were stored separately in a -25 °C walk-in freezer until freezedrying. The yields of dried soluble hydrolysates and insoluble pellets were recorded and the products were then stored in a -25 °C walk-in freezer until further analysis. In the main thesis study, Taguchi's design Li6(4 ) was generated using Minitab statistical 5  program (MINITAB 14.12.0, Minitab Inc.; Pennsylvania, USA). Using this design, 16 experiments were conducted under varying conditions of the four factors, namely waterto-substrate ratio, the percent enzyme, time of hydrolysis and type of protease, each at four levels (Table 2.2). Additional experiments were performed to include controls without addition of exogenous proteases and two additional replicates of sample SWA4 to assess process reproducibility. For each experiment, 1000 g of whole shrimp wastes were thawed overnight at 4 °C. The wet wastes were gently rinsed twice with tap water, drained, and blended with the appropriate amount of 50 °C distilled deionized water for 2 minutes using the Osterizer blender. A 20 mL aliquot of the waste blend was taken for pH measurement while to the rest was added the designated amount of protease and the whole mixture was placed in a preheated 50 °C Maqni Whirl constant temperature bath (Blue M Electric Company; Illinois, USA) equipped with a mechanical stirrer (Talboys Instrument Corp.; New Jersey, USA). The mixture was allowed to reach 50 °C before timing for incubation. After incubation for the selected duration, the hydrolysate was immediately pasteurized at 95 °C for 10 minutes in a steam kettle with constant stirring using a mechanical stirrer (Eastern Industries Inc.; Connecticut, USA) controlled by Powerstat variable transformers  64  Table 2.1  The type of protease and the amount of enzyme used in the hydrolysis of liquid waste blend in the preliminary study and the pH recorded for the 3  supernatant of the resulting shrimp waste hydrolysates.  Sample Code  Type of Protease  Amount of enzyme (mg)  pH  SWA-1  Alcalase  800  7.48  SWB-1  Bromelain  800  7.84  SWF-1  Flavourzyme  800  7.86  SWP-1  Protamex  800  7.44  SWC-1  -'  -  8.39  Shrimp wastes were blended with 2-fold ddH 0; 300 g liquid waste blend was used for each hydrolysis reaction.  a  2  65  Table 2.2  Summary of Taguchi's Li6 (4 ) experimental design with 4 factors and 4 5  levels and additional experiments for hydrolysate production from whole shrimp waste.  1  1:1  Percent Enzyme (%) 1  2  1:1  2  4  Bromelain  SWB4  3  1:1  3  8  Flavourzyme  SWF8  4  1:1  4  24  Protamex  SWP24  5  1.5:1  1  4  Flavourzyme  SWF4  6  1.5:1  2  1  Protamex  SWP1  7  1.5:1  3  24  Alcalase  SWA24  8  1.5:1  4  8  Bromelain  SWB8  9  2:1  1  8  Protamex  SWP8  10  2:1  2  24  Flavourzyme  SWF24  11  2:1  3  1  Bromelain  SWB1  12  2:1  4  4  Alcalase  SWA4  13  2.5:1  1  24  Bromelain  SWB24  14  2.5:1  2  8  Alcalase  SWA8  15  2.5:1  3  4  Protamex  SWP4  16  2.5:1  4  1  Flavourzyme  SWF1  17  2:1  4  4  Alcalase  SWA4-1  18  2:1  4  4  Alcalase  SWA4-2  19  1:1  0  24  Control  SWC24  20  1.5:1  0  1  Control  SWC1  21  2:1  0  8  Control  SWC8  22  2.5:1  0  4  Control  SWC4  Experiment #  Water-toSubstrate Ratio 3  a b  b  Time of Hydrolysis (hr)  Type of Proteases  Sample Code  1  Alcalase  SWA1  Based on 1000 g of wet shrimp waste. Based on the assumption that there were 20 % proteins in the wet shrimp waste. 66  (Superior Electric Bristol; Connecticut, USA). After pasteurization, a 20 mL aliquot of the whole hydrolysate was taken for further analysis of pH and degree of hydrolysis. The rest of the mixture was sieved through a 2-rrim mesh immediately to separate the liquid from the solids; the liquid portion was further centrifuged in closed centrifuge bottles at 3300Xg for 20 minutes at room temperature using Du Pont Sorvall Centrifuge RC 5B (Mandel Scientific Co. Ltd.; Ontario, Canada). The supernatants constituted the soluble fraction of the protein hydrolysate while the pellets and the solids from the sieving step were combined to form the insoluble fraction. These fractions were stored in a -25 °C walk-in freezer and freeze-dried to obtain yields. Percent soluble yield was calculated according to the following equation: Percent Soluble Yield (%) =  Yield of dry soluble solid (g)  * 100 %  Yield of dry soluble solid + insoluble solid (g) 2.2.5  Degree of Hydrolysis (DH) Determination  For the preliminary study, 1 g of soluble hydrolysates were dissolved in 10 mL ddH 0 2  and centrifuged at 2530Xg for 20 minutes using the Beckman GS-6 Centrifuge. The supernatants were used as the sample for the DH determination and the DH was expressed as milliequivalent of leucine per gram of soluble hydrolysate. For the main thesis study, whole hydrolysates were used to determine the degree of hydrolysis and DH was expressed as milliequivalent of leucine per gram of dry shrimp waste (meq leu/g). Five milliliters of 24 % TCA was added to 5 mL of sample and the resulting white slurry was centrifuged at 12100Xg for 10 minutes using Du Pont Sorvall Centrifuge RC 5B. The supernatants were diluted 100-fold with ddH 0. Triplicates were performed for the 2  following steps. Two milliliters of 0.05 M sodium tetraborate buffer (pH=9.2) were added to 0.2 mL aliquots of diluted supernatants, followed by 1.0 mL of 4.0 mM TNBS. The resulting mixtures were vortexed and incubated in the dark at room temperature for 30 minutes, followed by addition of 1.0 mL of 2.0 M NaHP04 containing 18 mM 2  NaS03, vortexed briefly and the absorbance readings at 420 nm were immediately 2  recorded using the Shimadzu UV-1700 PharmaSpec UV-Visible Spectrophotometer  67  (Mandel Scientific Company Inc.; Ontario, Canada). A standard curve was constructed using leucine solutions of concentrations 0.2 - 3.0 mM and the DH was expressed as meq leu/g because the htai for the shrimp waste protein hydrolysate is not known. This to  trinitrobenzenesulfonic acid method to determine the DH was first proposed by AdlerNissen (1979) and further modified by Liceago-Gesualdo and Li-Chan (1999).  2.2.6  Informal Sensory Evaluation of Shrimp Waste Hydrolysates  For the preliminary study, the hydrolysate solutions were prepared at 10 % (w/v) using ddH20  at 40 °C. Four panelists participated in this informal tasting. They were asked to  evaluate the samples according to the following criteria: shrimp aroma, shrimp taste, bitterness and consumer preference (i.e. the likelihood of purchasing the products if it were sold in market). The evaluation was done using a 4-point scale where 0 represented the least and 4 represented the most. Panelists were asked to evaluate the samples relative to each other (for example, a sample with score 2 in the bitterness category would be considered relatively more bitter than a sample with score 1). Each panelist was provided with 10 % (v/v) lemon water, water and crackers. Panelists were asked to evaluate the shrimp aroma first by sniffing close to the sample holder. They were then instructed to plug their nose and taste the sample for 5 seconds to evaluate the bitterness after which they were asked to assess the shrimp flavour. Based on the three qualities evaluated, the panelists would then determine the preference for the products. The results from this evaluation did not give any quantitative data but provided a general idea On the sample quality.  2.2.7  Formal Sensory Evaluation of Shrimp Waste Hydrolysates  Two sensory evaluations were performed a year apart. Eight panelists were recruited for each evaluation.  68  2.2.7.1  First Sensory Evaluation (June 2006)  In the first sensory evaluation, panelists attended 2 training sessions and 7 tasting sessions. In the first training session, each panelist was provided four caffeine standards at 0, 500, 1000 and 1500 ppm and 2 samples at 10 % (w/v) solutions prepared using ddHiO. Panelists were instructed to plug their nose to block off any interfering aroma and injected the solution using a 1 mL syringe to the rear end of their tongue where bitterness sensation is the most prominent. After each tasting, panelists were asked to rinse the mouth once with 10 % (v/v) lemon water and twice with water. If bitterness persisted, they were asked to take some unsalted cracker to remove the bitter taste. The conclusion from this training session was to expand the bitterness scale to 3000 ppm due to the fact that samples were rated close to or higher than 1500 ppm, which was the upper limit of the current scale. In addition, panelists were unable to detect differences between 0 and 500 ppm caffeine standards so the standard of 750 ppm was used to represent the lowest detectable limit. In the second training session, caffeine standards of 0, 750, 1500 and 3000 ppm and 2 samples at 10 % (w/v) were provided to the panelists. Panelists were able to identify bitterness in the standards as well as their differences. They could also detect the bitter taste in the samples and rated the bitterness within the scale. Panelists all agreed upon using 1500 ppm caffeine standard as the reference for the tasting sessions. In the tasting sessions, panelists were seated in individual booths, and provided with 10 % (v/v) lemon water, water, crackers as well as an evaluation form (Appendix A). Five samples prepared at 10 % (w/v) and two 1500 ppm caffeine standards were given to the panelists. They were asked to evaluate the samples in random order and to refer to the standards whenever necessary. The sensory scores were analyzed using statistical analysis as described in Section 2.2.10.  69  2.2.7.2  Second Sensory Evaluation (August 2007)  In the second sensory evaluation, panelists attended 2 training sessions and 4 tasting sessions. In the first training session, each panelist was provided 4 caffeine standards at 0, 750, 1500 and 3000 ppm and 2 samples at 10 % (w/v). Panelists were asked to follow procedures as described in Section 2.2.7.1 to evaluate the samples. Some panelists could detect bitterness in the samples but failed to sense bitter taste in some of the standards. Therefore, the conclusion from this training session was to use samples from the previous evaluation (Section 2.2.7.1) as the standards not only to ensure consistent rating of the bitterness of the samples but also to evaluate the samples such that the scores from the two sessions could be compared. In the second training session, 5 samples prepared at 10 % (w/v) with an average bitterness score of 900, 1300, 1400, 1900 and 2300 ppm based on the first sensory evaluation results were provided to the panelists. The sample with score 1300 ppm was selected to be the reference and the panelists were asked to rate the other four samples according to the reference. They were able to identify bitterness in all the samples and their differences. Panelists evaluated the samples very similarly and they agreed to use the sample with a score 1300 ppm as the reference for the tasting sessions. In the tasting sessions, panelists were provided with six samples prepared at 10 % (w/v) and two reference samples (1300 ppm) and were asked to follow procedures to evaluate the bitterness of the samples as described in Section 2.2.7.1. The sensory scores were calculated by measuring the distance in centimeters from point zero to the marked point on the 15-cm line scale, followed by conversion of the distance to ppm caffeine through multiplication by 200 (3000 ppm / 15 cm). The sensory scores were then adjusted by subtracting 200 ppm because the reference samples provided were 1300 ppm instead of 1500 ppm. The adjusted scores were analyzed using statistical analysis as described in Section 2.2.10.  70  2.2.8  Proximate Analysis of Shrimp Waste Hydrolysates  Moisture, total nitrogen, fat, ash and carbohydrate content of the freeze-dried soluble shrimp waste hydrolysates were determined according to the procedures described earlier  in Section 2.2.3.  2.2.9  Amino Acid Analysis of Shrimp Waste Hydrolysates  Some soluble hydrolysate samples were sent to the Advanced Protein Technology Centre Amino Acid Analysis Facility (Hospital for Sick Children; Ontario, Canada) for amino acid analysis. Cysteine and tryptophan were not quantified because they were easily destroyed under the acidic conditions that were used in amino acid analysis.  2.2.10  Statistical Analysis  Product yields, degree of hydrolysis as well as bitterness were the responses monitored for the hydrolysis. Analysis of Variance General Linear Model ( A N O V A - G L M ) from the Minitab statistical program was used to determine whether the four factors (water-tosubstrate ratio, percent enzyme, time of hydrolysis and type of protease) had an effect on the responses. The analysis was performed at a = 0.05.  Sensory results from Section  2.2.6 were first normalized using the z-transformation test.  The normalized scores were analyzed using two-way A N O V A from the Data Analysis Tool in Microsoft® Excel 2002 and differences between samples were analyzed using Least Significant Difference test (LSD).  Sensory data from  Section 2.2.7.1 was analyzed using A N O V A - G L M with the three  factors being sample, panelist and replication. The panelist difference was adjusted using z-transformation test; the z-scores were analyzed by one-way A N O V A and differences between samples were analyzed by Tukey's test from the Minitab statistical program.  71  Sensory data f r o m Section  2.2.7.2  was a n a l y z e d u s i n g A N O V A - G L M after w h i c h a one-  w a y A N O V A w a s p e r f o r m e d arid differences between samples w e r e a n a l y z e d b y T u k e y ' s test.  72  2.3  2.3.1  Results and Discussion  Proximate Analysis of Shrimp Processing Waste, Solid Waste Blend and Liquid Waste Blend  Shrimp processing discards consist of heads, tails and shells. Depending on the processing methods, some shrimp meats can be found in the waste as well. Shrimp peeling by machinery usually leaves some shrimp meats as part of the wastes while handpeeling reduces the amount of shrimp meats in the waste materials. Due to the inhomogeneity of the starting materials, proximate analysis was performed on whole shrimp waste, solid waste blend as well as liquid waste blend to find out the differences in moisture, crude protein, fat, ash and carbohydrate (by difference) content. The proximate compositions are tabulated in Table 2.3. The proximate content of the solid waste blend contained approximately 75 % moisture, 12 % protein (wet basis, wb), 3 % fat (wb), 6 % ash (wb) and 4 % carbohydrate (wb) while the liquid waste blend composed 97 % moisture, 1.5 % protein (wb), 1.1 % fat (wb), 0.2 % ash (wb) and 0.4 % carbohydrate (wb). It was clear that the solid waste blend contained higher amount of solids; however, when the dry mass was compared between solid waste blend and liquid waste blend, it was found that the crude protein and carbohydrate content did not vary. The liquid waste blend contained a high amount of fat (35 %, db) whereas the solid waste blend was composed of more ash (25 %, db). When the proximate composition of the whole shrimp waste was examined, it was interesting to observe that the composition did not differ from the solid waste blend except for the fat content, with the whole shrimp waste having a higher content (20 %, db) than the solid waste blend (12 %, db). The proximate composition of the whole shrimp waste ih this study was compared with those reported in the literature in Table 2.4. The moisture content of whole shrimp waste in this study was 75.1 %, which fell in the range of 71 - 83 % from the literature values reported for whole shrimp waste and the shrimp head waste. The protein content varied  73  Table 2.3  Proximate composition of solid waste blend, liquid waste blend and whole shrimp waste.  Moisture  3  (%) Solid Waste Blend  wb  Liquid Waste Blend  wb  74.7 ±0.5  db  97 ± 1  Crude Protein (%)  b  Fat (%)  c  (%)  Carbohydrate (%) [by difference]  6.3 ±0.1  4.2 ±0.2  Ash  c  11.8 ± 0.2  3.0 ±0.1  46.7 ± 0.4  11.8 ±0.4 24.85 ± 0.09  16.6 ±0.6  1.5 ±0.6  1.1 ±0.5  0.2 ±0.1  0.4 ± 0.2  db  47.3 ±0.3  35 ± 1  6.1 ±0.7  12 ± 1  Whole wb Shrimp Waste db  75.1 ±0.9 11.1 ±0.4  5.0 ±0.2  5.5 ±0.3  3.3 ±0.3  44.6 ±0.5  20.0 ±0.4  21.9 ±0.6  13.4 ±0.9  Results are mean values of six replicates ± standard deviation. Results are mean values of four replicates ± standard deviation. °Results are mean values of three replicates ± standard deviation. a  b  74  Table 2.4  Comparison of proximate composition of shrimp processing discards in this current study with existing literature. Current Study  Shrimp Species Waste Parts  Water (%) Proteins (%, db) Chitin (%, db) Lipids (%, db) Ash (%, db) a  Pandalo -psis dispar Whole  75.1 ±0.9 44.6 ±0.5  Gildberg and Stenberg 2001  Mukhin and Novikov 2001  Mizani and others 2005  Coward -Kelly and others 2006  n.d.  n.d.  Heads only  Whole  Whole  Whole  Whole  Whole  Heads only  72.10 ±0.20 44.12 ±0.79 40.40 ± 0.48 8.39 ±0.80 29.03 ±0.43  71.12 ± 1.21 40.6 ±5.43 17.8 ±0.91 9.95 ±0.22 27.5 ±0.13  75.61 ±0.20 41.9 ±0.20 17.0 ±0.25 10.23 ± 0.41 29.2 ±0.20  73.5  76.8  80.5  78.54  74.0  78.0  63.9  64.10  6.5  8.8  0.4  5.2  n.d.  18.00  26.0  10.3  23.2  17.20  82.58  94.61  4.15 1.30  db, dry basis. Based on deproteinized shells, n.d. indicates value not determined. a  b  Synowiecki and AlKhateeb 2000; Shahidi and Synowiecki 1991  Shahidi 1994  Crangon crangon  b  20.0 ±0.4 21.9 ±0.6  Synowiecki and AlKhateeb 2000  Mandeville and others 1992  Pandulas Pandulas Pandalus borealis borealis borealis  Penaeus Penaeus semisul- indicus catus Heads i  only  the most ranging from 40 - 95 % (db) because the materials composing the waste can significantly alter the protein content in the sample. The lipid content of shrimp waste used in this study was high at 20 % (db) but data from Coward-Kelly and others (2006) indicated that this was not unusual. The ash content was around 20-30 % in general but some studies have shown low ash content as well (Mandeville and others 1992; Mukhin and Novikov 2001). The variation in the proximate composition could be due to difference in species, seasons as well as diet (Gordon and Roberts 1977; Mandeville and others 1992) in addition to the nature of the material in the waste. 2.3.2  Preliminary Studies with Liquid Waste Blend  The relative percent yields of soluble and insoluble fractions of the liquid waste blend hydrolysates are shown in Figure 2.1. It was obvious that Alcalase, Flavourzyme and Protamex samples achieved high relative soluble yields of about 80 % while the bromelain sample had a relative soluble yield of approximately 45 %, which was similar to the control sample. Alcalase, Flavourzyme and Protamex have been shown to efficiently solubilize substrate during hydrolysis (Aspmo and others 2005; Dumay and others 2006). Even though bromelain has also been shown to produce high soluble yield by Aspmo and others (2005), the result in this preliminary study showed otherwise. In terms of food application, solubility is a very important factor affecting the product's usage. Therefore, a protease that produces a high yield of soluble hydrolysate is preferred. During hydrolysis, proteases continuously cleave the proteins to generate smaller peptides, which are more soluble. It was evident that for the fixed amount of enzymes used in this study, Alcalase, Flavourzyme and Protamex seemed to more efficiently assist in solubilizing the hydrolysates. However, it was not conclusive to suggest that Alcalase, Flavourzyme and Protamex had higher hydrolyzing ability because the proteases used in this study did not have the same activity unit. In addition to the relative product yields, the a-amino group content of the supernatant of the soluble liquid waste blend hydrolysates was also measured. The a-amino group  76  SWA-1  Figure 2.1  SWB-1  SWF-1 Sample  SWP-1  Relative soluble and insoluble percent yields of shrimp waste hydrolysates produced from liquid waste blend. Relative percent yield is expressed as 100% * gram of soluble or insoluble yield / gram of total solid yield (refer to Table 2.1  —i —i  for sample code details).  content was high for the Flavourzyme sample, followed by the control sample (Table 2.5). Hydrolysates produced by bromelain and Protamex had relatively lower contents of a-amino groups while Alcalase sample had the lowest. This result contradicted many literature reports in which Alcalase has been shown to produce protein hydrolysates with high a-amino group contents (Shahidi and others 1995; Aspmo and others 2005). However, since the content of a-amino groups in this preliminary study was expressed as milliequivalent of leucine per gram of soluble hydrolysate, the results only reflected on the amount of free amino groups in the soluble hydrolysate fraction, and not in the whole hydrolysate. The results showed that Flavourzyme produced a larger number of small peptides and amino acids in the soluble fractions than the other proteases, which was in agreement with the work carried out by Aaslyng and others (1998). Nevertheless, the results did not show the effect of each enzyme on the DH of the whole liquid waste blend hydrolysates. Besides product yield and a-amino group recovery, an informal sensory evaluation was performed on the shrimp waste hydrolysates. The results did not provide quantitative data for statistical analysis. However, the sensory attributes of the samples could be ranked relatively. As shown in Table 2.5, the Alcalase sample was the most unfavourable product with low shrimp aroma, low shrimp flavour and high bitterness whereas hydrolysate produced by bromelain was the most preferred due to its strong shrimp aroma and shrimp flavor. Interestingly, the strength in shrimp aroma and flavour was valued over the level of bitterness. Flavourzyme sample gave the lowest bitterness; however, the relatively weak shrimp aroma and flavour contributed to the lower preference of the product. Protamex sample was in the mid-range in each category. However, since the whole shrimp waste was blended with 2-fold ddH 0 and sieved to 2  separate the liquid waste blend from the solid waste blend, the proximate contents distributed unevenly. From the proximate results shown in Section 2.3.1, the liquid waste blend, which was the substrate for this preliminary study, contained only a small proportion of the protein (1-2 %, wb) in comparison to 11 % (wb) of the whole shrimp waste. This yield was too low to be considered profitable for the food industry.  78  Table 2.5  a-Amino group content and informal sensory results of soluble shrimp waste hydrolysates from liquid waste blend (refer to Table 2.1 for sample code details).  a-Aihino group  3  Shrimp Aroma  Shrimp Flavour  b  Bitterness  Preference  b  5  b  SWA-1  1.67  1  1  4  1  SWB-1  2.71  4  4  2/3  4  SWF-1  3.83  2  3  1  3  SWP-1  2.88  3  2  2/3  2  SWC-1  3.25  n.d.  n.d  n.d  n.d  meq L-leucine / gram soluble shrimp waste hydrolysate, measured using TNBS method. The sensory scores were ranked relatively with 1 being the lowest and 4 being the highest. n.d. represents value not determined. a  b  79  As a result, these data from the preliminary study were used as the basis to set experimental conditions for the main thesis study. 2.3.3  Hydrolysate Production from Whole Shrimp Waste using Taguchi's Fractional Factorial Design  Four factors, namely water-to-substrate ratio, percent enzyme, time of hydrolysis as well as type of protease, were explored in this study. Water-to-substrate ratio may affect the solubility or accessibility of the substrate hindering the efficiency of the proteases. Percent enzyme is another major factor because in general, a higher amount of proteases will increase the degree of hydrolysis. However, there is usually a saturation level at which the addition of more enzymes will not enhance the degree of hydrolysis. Also, the cost of proteases is another point to consider because it can significantly increase the production cost. Therefore, processors usually attempt to explore conditions where the minimum amount of enzymes can be used to produce a high degree of hydrolysis and high product yield. Time of hydrolysis, in combination with the above factors, determines the sensory, textural and functional properties of the final products. The longer the substrates are hydrolyzed, the smaller the peptides are generated until no hydrolysable substrates exist in the hydrolysis mixture. These shorter peptides are sometimes favourable because they are more easily absorbed in the human body and can potentially carry biological activity. However, aside from possible oxidation and other side reactions that may occur during prolonged hydrolysis, a long incubation time also indicates a higher consumption of energy because hydrolysis involves incubation at ah elevated temperature and the process is enhanced with constant stirring. This will in turn increase the production cost. Type of protease is important in that different proteases target different cleavage sites, which can lead to completely different peptide profiles in the final products and impact on the sensory properties including bitterness. Furthermore, since different proteases have different activities, proteases with high activities are generally favoured because a smaller amount is needed to achieve a high degree of hydrolysis.  80  In order to explore these four factors, Taguchi's experimental design was chosen because it is well known to produce maximum amount of results with the least number of experiments (Charteris 1992). As shown earlier, whole shrimp waste contained a high amount of proteins (45 %, db); thus, whole shrimp waste was used as the substrate for the hydrolysis reaction. For 1000 g of whole shrimp waste, approximately 110 g of crude proteins could be obtained. In order for these waste materials to be used in an industrial setting, the processing method should be as simple and economical as possible. As a result, the pH was not controlled during the hydrolysis. Controlling pH is a laborintensive process and the adjustment of pH using acid/base usually generates salts which need to be removed at the end of the process. A similar study on Atlantic cod viscera had also adopted hydrolysis method without the control of pH and results showed that pH change was not drastic (Aspmo and others 2005). To investigate if there were major changes in pH during hydrolysis in the current study, the pH of the shrimp waste blend and the pasteurized hydrolysates were recorded (Table 2.6). The pH did not fluctuate drastically during hydrolysis and remained within the range of 8.3 - 9.2. This range did not appear to agree with the preliminary results where the pH range was 7.4 - 8.4. It could be explained by the fact that in the preliminary study, liquid waste blend was used as a substrate and 97 % of the substrate was moisture. On the contrary, the whole shrimp waste blend contained more solid matters, which could have affected the pH. Although the pH range fell outside the optimal pH range of 7.5 - 8.5 of some of the selected proteases, the advantage of not having to adjust the pH was still preferred in a large-scale setting. Soluble yield in percentage, degree of hydrolysis as well as bitterness in ppm caffeine were parameters measured for the samples produced from each of the hydrolysis conditions (Table 2.7). These parameters were analyzed using ANOVA-GLM tofindout if water-to-substrate ratio, percent enzyme, time of hydrolysis, and/or type of protease had an influence on their values. The p-values tabulated in Table 2.8 showed that percent soluble yield and bitterness (ppm caffeine) were significantly affected by the type of protease used in the hydrolysis. Degree of hydrolysis was dependent on both the percent enzyme as well as the type of protease.  81  Table 2.6  The pH of whole shrimp waste before and after hydrolysis (refer to Table 2.2 for sample code details).  Sample Code  Initial pH  Final pH  SWA1  8.35  8.42  SWB4 SWF 8  8.50 8.55  8.61 8.41  SWP24  8.75  8.70  SWF4  8.62  8.44  SWP1  8.87  8.54  SWA24  8.53  8.47  SWB8 SWP8 SWF24  8.51  8.33 8.71  a  8.91  b  8.46  SWB1  8.57 n.d..  SWA4  8.61  8.54  SWB24  8.57  8.36  SWA8  8.72  8.34  SWP4  8.86  8.49  SWF1  8.59  8.34  SWA4-1  8.72  8.76  SWA4-2  8.79  8.85  SWC24  8.71  9.18  SWC1  8.53  8.43  SWC8  8.59  8.50  SWC4  8.63  8.47  8.18  Measured after blending with ddFkO and right before addition of proteases. Measured after pasteurization of shrimp waste hydrolysates. n.d. indicated value not determined. a  b  82  Table 2.7  Summary of percent soluble yield, degree of hydrolysis and bitterness for each shrimp waste hydrolysate produced from the Taguchi's experimental design (refer to Table 2.2 for sample code details).  Sample Code  Soluble Yield (%)  Degree of Hydrolysis *  Bitterness  SWA1 SWA4  28.47 37.64 36.04 38.82 36.31 31.36  1.97 2.30 2.09 2.64 2.45 3.02  2300 2300 2400 2500 2400 2300  17.35 14.29 17.03 18.96  0.56 0.64 0.68 0.67  1300 1100 1400 1900  21.05 13.04 18.75 19.09  2.31 1.43 3.05 2.69  1500 1300 1600 1600  30.83 35.79 34.52 39.10  1.51 1.75 1.22 2.22  2100 2200 2100 2200  4.98 5.43 5.76 5.64  0.40 0.31 0.45 0.31  900 900 800  SWA4-1 SWA4-2 SWA8 SWA24 SWB1 SWB4 SWB8 SWB24 SWF1 SWF4 SWF8 SWF24 SWP1 SWP4 SWP8 SWP24 SWC1 SWC4 SWC8 SWC24  a  1  0  n.d.  Expressed as 100% * gram of soluble yield / gram of total solid yield. meq L-leucine / gram dry shrimp waste weight, measured using TNBS method. Reported to the nearest hundred ppm caffeine equivalents, averaged among all panelists over two sensory evaluations, n.d. indicated value not determined. b c  83  Table 2.8  ANOVA General Linear Model results showing p-values of each factor investigated from the Taguchi's Li6 (4 ) design on the selected responses. 5  P-value Soluble Yield  Degree of Hydrolysis  Bitterness  Water-to-substrate ratio  0.054  0.274  0.548  Percent enzyme  0.063  0.035*  0.921  Time of hydrolysis  0.204  0.053  0.500  Protease  0.001*  0.002*  0.020*  * Values bearing an asterisk are significant factors at P<0.05.  84  Percent soluble yield depended on the type of protease because different proteases have different activity which affects the hydrolyzing efficiency. A protease with a high activity hydrolyzes proteins more efficiently. When proteins are hydrolyzed, they become smaller peptides and are more soluble. Therefore, proteases that have high activities usually give higher soluble yields. A study performed by Rebeca and others (1991) has shown that enzyme concentration and incubation time also affected the soluble recovery where an increase in enzyme concentration or duration of hydrolysis increased soluble recovery. However, in the current study, percent enzyme and time of hydrolysis were not significant factors. This could be because the effect of type of protease on soluble yields was so strong that influence from other factors became negligible. Similarly, type of protease affects level of bitterness for the reason that bitter peptides are usually small (Aubes-Dufau and others 1995; Kristinsson and Rasco 2000; Cho and others 2004). Proteases with high activities can cleave proteins to smaller peptides at a shorter time, hence generate stronger bitterness. However, the concept of causing bitterness is more complex than simply size, involving the structural property of the peptides. Researchers have suggested that hydrophobic and bulky groups are essential to elicit bitterness (Ney 1979; Charalambous 1980; Ishibashi and others 1987; Ishibashi and others 1988a; Ishibashi and others 1988b; Nishitoba and others 1988; Aubes-Dufau and others 1995; Gomez and others 1997; Kim and others 1999). Therefore, proteases with different cleavage specificities will also yield peptide mixtures of varying bitterness. Degree of hydrolysis, likewise, depends on the type of protease because a protease of high activity reaches high DH faster. Percent enzyme, on the other hand, affects the DH by allowing more proteases to cleave the substrate at the same time, thereby decreasing the time required to reach high degree of hydrolysis. Therefore, at a set duration of hydrolysis, a higher percent of enzyme used will give a higher degree of hydrolysis.  85  Figure 2.2 shows the main effect plot of percent soluble yield of shrimp waste hydrolysates. Results from ANOVA-GLM indicated that water-to-substrate, percent enzyme and time of hydrolysis did not vary with the change in levels. However, an effect could be observed with the different type of protease. Alcalase and Protamex samples appeared to have higher percent soluble yields than those hydrolyzed by bromelain and Flavourzyme. In Figure 2.3, it was shown that hydrolysates produced by Alcalase and Protamex had percent soluble yields above 30 %, which were significantly higher than the bromelain, Flavourzyme and control samples. The bromelain and Flavourzyme samples achieved approximately 15 % soluble yields, which were significantly higher than the control samples that had around 5 % yields. It was clear that by using proteases to hydrolyze the shrimp processing discards, the percent soluble yields increased significantly. The high soluble yields of protein hydrolysates produced from proteolytic proteases have also been shown by other researchers (Aspmo and others 2005; Dumay and others 2006). The main effect plot of the DH is shown in Figure 2.4. Percent enzyme and type of protease were the two factors that played significant roles in the variation observed in the DH. Similar results have also been shown by other researchers (Diniz arid Martin 1996; Benjakul and Morrissey 1997; Guerard and others 2001; Dumay and others 2006). The DH did not vary as a function of water-to-substrate from 1:1 to 2.5:1. Even though Nilsang and others (2005) have shown the time effect on DH, in this study, incubation time did not seem to cause an effect on DH. When the effect of percent enzyme was examined, it was interesting to find that as percent enzyme added was increased from 1 % to 3 %, there was a gradual increase in DH; however, when it was increased from 3 % to 4%, the DH did not continue to rise. It illustrated that the amount of protease had reached a saturation point at which further addition of enzymes would not cause any effect. This phenomenon has also been observed in the study performed by Diniz and Martin (1996). In terms of type of protease, it was apparent that Alcalase and Flavourzyme excelled in hydrolyzing the shrimp processing discards according to the amount added, which is consistent with several studies showing their effectiveness for hydrolyzing other food proteins (Shahidi and others 1995; Aaslyng and others 1998; Smyth and FitzGerald 1998).  86  Percent Enzyme  Water to Substrate Ratb 35  ....  • ^*  2 >  20  15-  '.'so.:' s3s;  #  c  i  i  i  i  i  1.0  1.5  2.0  2.5  1  Time of Hydrolysis (hr)  Protease  .i  i  3  4  •  35  re _  i 2  30  •  •  •  •  •  25  •  20  i 15  Figure 2.2  i  i  i 24  i  Alcalase  •i  Bromelain  i  Flavourzyme  i  Protamex  Main effect plots of mean percent soluble yield as a function of water-tosubstrate ratio, percent enzyme, time of hydrolysis and protease. Horizontal line in each plot represents the mean value of all 16 experiments.  87  40-  30 A  B B  ? 20-  L_  3 O  10-  Alcalase  Figure 2.3  Bromelain  Control iProtease  Flavourzyme  Protamex  Boxplot of percent soluble yield as a function of protease. Bars bearing different letters indicate significant difference at P < 0.05 analyzed using Tukey's test with individual confidence level = 99.24 %.  Percent Enzyme  Water to Substrate Ratio  *  LO  *---».^  —  l.b  »  2.0  2.b  Time of Hydrolysis (hr)  •  •  1  2  3  Protease  •  4  •  •  • •  ••"•1  Figure 2.4  4  8  -  24  Alcalase  Bromelain  Flavourzyme  Protamex,^  Main effect plots of mean degree of hydrolysis as a function of water-tosubstrate ratio, percent enzyme, time of hydrolysis and type of protease. Horizontal line in each plot represents the mean value of all 16 experiments.  89  Protamex did not perform as well but still achieved a high DH. Bromelain did not seem to hydrolyze well because the DH of bromelain samples was low. This finding contradicted the results from Aspmo and others (2005) where bromelain was shown to give very high yield of a-amino groups from Atlantic cod. When the DH was plotted as a function of type of protease, percent enzyme and time of hydrolysis, the influence of each factor on each hydrolysate sample could be observed (Figure 2.5). Alcalase, Flavourzyme and Protamex samples all had higher degree of hydrolysis than the controls. Hydrolysates produced by bromelain, on the other hand, did not vary with percent enzyme and time of hydrolysis and remained low in DH comparable to the control. Among the experimental conditions investigated in this study, Alcalase samples appeared to be more affected by incubation time showing an increased DH with time. Hydrolysates produced by Protamex, on the other hand, were shown to be more influenced by the amount of enzyme involved in the hydrolysis. As percent enzyme increased, the DH also increased. Both percent enzyme and time of hydrolysis had an impact on the DH of Flavourzyme samples. They required a relatively long hydrolysis time coupled with a relatively high amount of enzyme to reach a high DH. When the whole shrimp waste was hydrolyzed with 1 % Flavourzyme and for 4 hours, the DH was approximately 1.5 meq/g whereas when the percent enzyme was increased to 3 % and the incubation time was extended to 8 hours, the DH reached 3.0 meq/g. The effect of type of protease on bitterness level will be discussed in Section 2.3.4. 2.3.4  Sensory Evaluation  The two sets of sensory data were analyzed independently. Table 2.9 summarizes the raw data from the first sensory evaluation. After ANOVA-GLM analysis of three factors including sample, panelist and replication, it was shown that replicates were not significantly different. As a result, the replicate scores were averaged and the data was analyzed by two-way ANOVA which showed significant effect on both samples and panelists. Significant panelist effect means that panelists were not using the line scale the 90  !  3.0 + i  i  !  2.5+.£  2.0 +  "! I f S  Br  l i i  -  l l t R  s l  ! - T-  •i  f"  i  i i i i i  i i  i i  B  ! 1  o  I I  I I  1.5+™-  •1 !  I I  I i  2  0.5+ - T  o4__  0-  Bromelain  •Alcalase; i  T  l  !  s  l  !  I  3 4  i  !  „  j _ _  i  l  !  l  j  1 1  1  1 ! I _  I  I  I  I l i  .  1  1 1  i  I  !  !  I  I  J  _  J  !  I !  t  I I  I  I  I  !  S 1  I I  S i.  1  I I  F " 1 I  !  i  Flavourzyme  [ Prptamex  ! ! I  I  !  -i  12  1  1  •  4^  i  l  !  I  r  1  1  3 4  r 2 3 4  I S f ( I  -f  j  I _  i  "T"  r — r  12  •  i _ _ i  1 I I  -Control I I  T  !  i i i i  r  A...  •i--r-  -r"  (A  S i.o  .>_4  I  i  I i i i  I  I l l  S l l  l I  l f  !_  r—i  +  1 2 3 4  Percent Enzyme Figure 2.5 vo  Plot of degree of hydrolysis as a function of type of protease, percent enzyme and time of hydrolysis (• 1-hour • 4hour A 8-hour • 24-hour).  Table 2.9  Sensory scores of first sensory evaluation on soluble whole shrimp waste hydrolysates evaluated using a scale based on 0, 750, 1500 and 3000 ppm caffeine standard solutions (refer to Table 2.2 for sample code details). #7  #8  2800 1900  2700 2700  2400 2500  2800 2700  1300 2200  1900 3000  2200 2500  2500 2100  2700 2800  1300 1000  2500 2400  2300 200  2600 2500  1200 2000  2500 2600  1600 700  2900 3000  2000 2500  1100 1300  1300 1400  1200 1200  1800 1200  700 800  1600 1800  1300 1700  1200 800  1200 900  1400 1600  1600 1300  1300 1400  1200 400  1800 2100  1600 1900  SWB24 SWB24  1700 1300  2500 1400  2400 1700  2200 1700  1700 1700  1400 2000  2100 2500  1100 2300  SWC1 SWC1  1100 900  1000 500  1000 1000  200 200  900 700  600 600  2200 1300  1400 600  SWC4 SWC4  900 1000  500 700  800 900  200 400  900 700  600 600  1200 1200  1200 600  SWC8 SWC8  1200 500  600 400  1000 900  200 100  900 900  600 100  1300 1300  1000 400  SWF1 SWF1  1300 1200  700 1600  1600 1600  1000 700  1900 2000  900 700  1500 1500  2000 2600  SWF8 SWF8  1800 1100  1300 500  1800 1800  1500 1400  1400 1600  1000 800  2400 2200  1600 1300  SWF24 SWF24  1400 1200  1700 1300  1700 2100  1000 900  1600 1300  1300 2100  2300 2200  2300 2100  Panelist Sample SWA1 SWA1  #1  #2  #3 #4 #5 #6 Bitterness (ppm caffeine)  1000 1600  2200 2400  2500 2800  2100 2200  2500 2700  SWA4 SWA4  2500 2200  2500 1800  3000 2200  1900 1300  SWA8 SWA8  2100 2800  2700 2700  2800 2800  SWA24 SWA24  1600 2200  1900 2400  SWB1 SWB1  800 800  SWB8 SWB8  92  same way; as a result, z-transformation was performed on the sensory scores to eliminate this difference. The z-scores were then analyzed using one-way ANOVA and Tukey's comparison test to investigate differences between samples. As shown in Figure 2.6, control samples were not significantly different among each other and had low degrees of bitterness. Bromelain samples hydrolyzed for 1 and 8 hours did not appear to differ significantly from the 1-hour control but were slightly more bitter than the 4-hour and 8-hour control samples. Among the hydrolysates produced by bromelain, the 24-hour sample was more bitter than the 1-hour hydrolysate. The Flavourzyme samples were not significantly different from each other and the bromelain samples but were more bitter than the controls. The hydrolysates of Alcalase were high in bitterness and the bitterness levels did not seem to vary among the samples. The Alcalase samples were significantly more bitter than most of the bromelain and Flavourzyme samples as well as the controls except for the 24-hour hydrolysates produced by bromelain and Flavourzyme, which were not significantly different from some or all Alcalase samples. These sensory results on soluble whole shrimp waste hydrolyates parallel those acquired from the preliminary data for hydrolysates from liquid waste blend as well as those reported in the literature, in that Flavourzyme produced samples of low bitterness while Alcalase samples had a strong bitter taste (Aaslyng and others 1998; Imm and Lee 1999; Gildberg and others 2002). Since additional experiments were performed after the first sensory evaluation, a second sensory evaluation was carried out to obtain sensory data for the new samples. Three samples from the first sensory evaluation were included in the second evaluation to assure results were comparable for the two data sets. Raw data of the second sensory evaluation is shown in Table 2.10. ANOVA-GLM with three factors comprising sample, panelist and replication were performed and results showed that sample was the only significantly different factor. Therefore, a one-way ANOVA followed by Tukey's comparison was performed and the results are shown in Figure 2.7.  93  1.5  AB  l-l+l 1 l i 11 555  0.5  8 o 91  SWA1  SWA4  SWA8  SWA24  AC BCD  CD  SWB1  SWH8  SWB24  SKVQ1  SKvd 8  IsWI I  SVVl'8  SWF24  CD  CE  -0.5  SKVQ4  DEF  EF  -1.5 Sample  Figure 2.6  A plot of z-scores of first sensory evaluation of soluble shrimp waste hydrolysates (refer to Table 2.2 for sample code details). Bars with different letters indicate significant difference at P < 0.05 analyzed using Tukey's test with individual confidence level = 99.24 %.  Table 2.10  Sensory scores of second sensory evaluation on soluble whole shrimp waste hydrolysates evaluated using samples with scores 900, 1300, 1400, 1900 and 2300 ppm caffeine from Section 2.2.7.1 as reference points and sample of 1300 ppm caffeine was used as the reference, marked at midpoint of the 15-cm line scale (refer to Table 2.2 for sample code details). #15  #16  2800 2900  2700 3000  2900 2400  3000 3000  2800 2400  3000 2900  2900 2400  2900 2700  2700 2800  3000 2600  3000 3000  2800 2700  1100 1200  2100 1300  1200 1200  2000 900  1600 100  1100 1400  1400 1500  2600 2000  1200 1300  2100 1600  1800 1300  1800 1500  2100 1300  1800 2000  900 1200  1300 1600  1300 1700  2000 900  1400 1800  1100 1100  1500 1500  SWC4 SWC4  1600 1500  1500 1300  500 2000  1600 1800  800 100  1200 900  100 390  1700 1600  SWP1 SWP1  2500 2300  2600 2500  2400 1600  1300 2700  2100 1900  2800 2100  2400 2400  2800 2300  SWP4 SWP4  2700 2600  2900 3000  1300 2600  2400 1600  200 2900  2700 2400  2700 2500  2900 2700  SWP8 SWP8  2400 2300  2200 2900  2200 1700  2000 1900  2500 2400  2200 2500  1800 2200  2700 2700  SWP24 SWP24  1700 2200  2600 2900  2300 1600  2400 2600  2400 2600  3000 2900  2400 1300  2400 2500  Panelist Sample SWA4 SWA4  #9  #10  #11 #12 #13 #14 Bitterness (ppm caffeine)  2800 2100  1800 2900  2200 2600  1900 1700  3000 3000  SWA4-1 SWA4-1  2400 2800  3000 2800  2100 2600  2100 2300  SWA4-2 SWA4-2  2400 2600  3000 2900  2100 2300  SWB4 SWB4  1300 900  1300 1400  SWF1 SWF1  2100 1800  SWF4 SWF4  95  3000  SWA4  F i g u r e 2.7  SWA4-1  SWA4-2  SWB4  SWC4  SWF1  Sample  SWF4  SWP1  SWP4  SWP8  SWP24  A plot of bitterness in ppm caffeine of second sensory evaluation of soluble whole shrimp waste hydrolysates (refer to Table 2.2 for sample code details). Bars with different letters indicate significant difference at P < 0.05 analyzed using Tukey's test with individual confidence level = 99.86 %.  as.  The control, bromelain and Flavourzyme samples in this sensory evaluation did not differ significantly in bitterness. It should be noted that the bromelain and Flavourzyme samples included in this sensory evaluation were hydrolyzed for 1 hour and 4 hours, which tended to produce hydrolysates of lower bitterness than those from extensive hydrolysis. The hydrolysates produced by Alcalase and Protamex were significantly higher in bitterness and they did not differ significantly from each other. This result was interesting because Protamex has been suggested to produce non-bitter hydrolysates (Liaset and others 2003), yet the hydrolysates produced had bitterness similar to samples hydrolyzed with Alcalase, which has been shown to produce bitter hydrolysates (Gildberg and others 2002). The three samples of Alcalase in this sensory evaluation were replicates done using the same experimental condition. As could be observed, the bitterness did not differ significantly, which indicated that the process was reproducible in terms of the bitterness level generated by the hydrolysates. Since the two sets of data needed to be combined for the analysis of effect of the four factors on bitterness, the three samples, which were evaluated both in the first and the second sensory evaluation, were used to validate the compilation of both data sets. A two-way ANOVA with factors being sample and time was performed on the data and no significant difference was found between the two sensory evaluations. As shown in Figure 2.8, the sensory scores were plotted against time and samples. It was clear that the scores from the two sensory evaluations did not differ from each other and hence, the data was compiled. The average scores of the compiled data were tabulated in Table 2.7. From the ANOVAGLM analysis, type of protease was the only significant factor. This agreed with the results from Gildberg and others (2002) in that different proteases produced different level of bitterness. In the main effect plot (Figure 2.9), it was obvious that bitterness did not change as a function of water-to-substrate ratio, percent enzyme and time of hydrolysis according to the levels studied. However, bitterness in samples hydrolyzed with various proteases differed significantly.  97  3000 H  2500 H  •'-0-  1  '  ^—^ Time  1  Sample  Figure 2.8  2 JSWA4  '  i^er 1  2 SWC4  '  =  r  1  2  SWFl  Boxplot of bitterness (ppm caffeine) of duplicate sensory scores from 8 panelists each for both sensory evaluations for soluble shrimp waste hydrolysate samples, as a function of time (time 1 indicates sensory evaluation done in 2006; time 2 represents sensory evaluation performed in 2007).  98  Water to Substrate Ratio  Enzyme to Substrate Ratio  2250 2000 |  1750-  Q.  u c  1500 i 1.0  _  r 1.5  i 2.0  1  1  1  2.5  1  2  Time of Hydrolysis (hr)  _  1750  Protease  • ••  •  1500 24  Figure 2.9  1  4  •  V  2000-  1  3  •  5 2250  2  ,  Alcalase  •  Bromelain  Flavourzyme  Protamex  Main effect plots of mean bitterness (ppm) as a function of water-tosubstrate ratio, percent enzyme, time of hydrolysis and type of protease. Horizontal line in each plot represents the mean value of all 16 experiments.  99  In Figure 2.10, the effect of protease on bitterness was more clearly shown. Alcalase and Protamex samples had bitterness levels of above 2000 ppm caffeine, indicating that the samples were very bitter. In addition, the spread of scores for these two samples was small indicating that intense bitterness could be easily identified and evaluated by most panelists. These samples were significantly stronger in bitterness than the hydrolysates produced by bromelain and Flavouryzme, which had values of about 1500 ppm caffeine. Since these samples contained mild level of bitterness, the spread of scores was larger. Each panelist had different sensitivity to the bitterness in the samples. Some panelists were more sensitive so they gave a higher score to the samples whereas others might not detect as well the level of bitterness in these samples resulting in the bigger variation in the scores. These hydrolysate samples in turn were more bitter than the controls, which had bitterness of approximately 800 ppm caffeine. It should be noted that the threshold value for bitterness detection for the panelists was 750 ppm caffeine, meaning that the control samples having bitterness of 800 ppm were not very bitter. Since the controls had nearly no bitterness, it was not difficult for the panelists to evaluate the level of bitterness in these samples so the spread of scores was not huge. From these data, it was not surprising to find that Flavourzyme samples had low level of bitterness because Flavourzyme is known to produce non-bitter hydrolysates (Pommer 1995; Aaslyng and others 1998). Nevertheless, Protamex, a protease complex known to produce non-bitter hydrolysate (Liaset and others 2003), produced samples that had bitterness as strong as Alcalase samples. This disagreement could possibly be related to the nature of the substrate as well as the hydrolysis conditions. Since the control samples were not intensely bitter and hydrolysates produced by different protease generated different level of bitterness, it was evident that peptides were the substances in the shrimp waste hydrolysates that were mainly responsible for the bitterness observed in the hydrolysates in this study. Even though lipid oxidation compounds (Liu and others 2000) and peptides (Aubes-Dufau and others 1995; Lee and Warthesen 1996; Kim and others 1999; Maehashi and others 1999; Kim and others 2003; Cho and others 2004) have both been proposed to contribute to the bitterness in protein  100  2500  2000-^  E o.  a. (A Ul 01  1500  . -1000 ^  Alcalase  Figure 2.10  Bromelain  Control Protease  Flavourzyme  Protamex  Boxplot of bitterness (ppm) as a function of protease. Bars bearing different letters indicate significant difference at P < 0.05 analyzed using Tukey's test with individual confidence level = 99.24 %.  101  hydrolysates, in this study, the controls did not contain much bitterness which strongly suggested that lipid oxidation compounds were not the main contributor to the bitterness in these shrimp waste hydrolysates. For lipid oxidation compounds to be primarily responsible would require all samples to have relatively the same level of bitterness because the lipid content would be similar for all of the starting solid waste materials used in these hydrolysates and different proteases should not cause any effect on the amount of lipid oxidation compounds. Furthermore, it has been suggested that during hydrolysis reaction, continual shaking or stirring of the reaction mixtures will increase bitterness (Liu and others 2000). Since the control sample was not bitter even after 8 hours of incubation with stirring, lipid oxidation compounds could be ruled out as the main substances eliciting bitterness in the hydrolysates. Oh the contrary, peptides could be the main contributor to the bitterness in hydrolysates. Samples hydrolyzed with different proteases contained varying levels of bitterness, which were significantly higher than the controls. Proteases usually target at different sites for cleavage so it was reasonable for different enzymes to generate products of different level of bitterness depending on the peptide profile. 2.3.5  Hydrolysate Production Reproducibility  Depending on the application of the protein hydrolysates, reproducibility could be a factor affecting product yield and sensory quality of the samples, which in turn, influence the profitability of the process. Therefore, one experimental condition (2:1 water-tosubstrate, 4 % enzyme, 4-hour hydrolysis, Alcalase), was selected from the Taguchi's design to test for repeatability due to its relatively high product yield. The hydrolysis conditions and the measured responses for the triplicate runs are shown in Table 2.11. The pH remained in the range of 8.5 - 8.8 during the hydrolysis for the three replicates despite the fact that pH was hot controlled. The soluble yield of the whole shrimp waste hydrolysates ranged from 56 - 63 % while the insoluble yield was 37 - 44 %. The degree of hydrolysis ranged from 2.09 - 2.64. Bitterness in all three samples was rated very high, from 2300 to 2500 ppm, indicating their strong bitter taste. 102  Table 2.11  Reproducibility of shrimp waste hydrolysate production.  Experiment  SWA4  SWA4-1  SWA4-2  Average ± %CV  Mass of shrimp waste (g)  1000.0  1000.1  1000.0  -  2000  2000  2000  -  Mass of protease (g)  8.0  8.0  8.0  -  Initial pH  8.61  8.72  8.79  8.71 ± 1%  Final pH  8.54  8.76  8.85  8.72 ± 2%  50  50  50  -  Soluble Hydrolysate Relative Yield (%)  60.35  56.34  63.45  60.05 ± 6%  Insoluble Hydrolysate Relative Yield (%)  39.65  43.66  36.55  39.95 ± 9%  Degree of Hydrolysis (meq/g)  2.30  2.09  2.64  2.34 ± 12%  Bitterness (ppm)  2300  2400  2500  2400 ± 4%  Volume of water (mL)  Temperature (°C)  103  Even though the experimental conditions were the same for the triplicates, it was not expected that the results would be identical. The percent coefficient of variation (%CV) for each measured parameters is included in Table 2.11. It could be observed that %CV was under 10 % for pH, product yields and bitterness of the hydrolysate samples. Even though the variation in DH was higher giving a %CV of 12 %, it was still considered reasonable because of the non-homogeneous nature of the starting materials consisting of heads, shells and tails. 2.3.6  Proximate Analysis of Soluble Shrimp Waste Hydrolysates  Two samples, SWA4 and SWC1, were selected to be analyzed by proximate analysis. The Alcalase sample was high in bitterness (2300 ppm caffeine) while the control sample was low (900 ppm caffeine). As shown in Table 2.12, the proximate compositions of the two soluble hydrolysates were generally similar. They both contained less than 10 % moisture and over 80 % crude protein (based on 6.25 X total nitrogen content), which were similar to those reported by deHolanda and Netto (2006). Hydrolysate SWA4 contained slightly more crude proteins, less moisture, less ash and less fat than sample SWC1. These results suggested that the protein hydrolysates produced from shrimp processing discards contained high crude protein contents which could be incorporated into food product to increase nutritive values. According to the proximate composition of the soluble shrimp waste hydrolysates, it could be estimated that approximately 60 g of crude proteins were recovered from 1000 g of wet whole shrimp waste under the assumption that 30 % soluble yield was obtained. Since there was about 110 g of proteins in 1000 g of wet whole shrimp waste, it showed that more than 50 % of the crude proteins from the waste were recovered in the soluble shrimp waste hydrolysates.  104  Table 2.12 Proximate composition of two soluble hydrolysate samples produced by Alcalase (SWA4) and control incubation (SWC1).  SWA4  db  db wb  Carbohydrate  (%)  (%)  5.5 ±0.2  7.3 ±0.1  87.3 ±0.2 0.2 ±0.1  6.8 ±0.1  82.5 ± 0.2  0.2 ±0.1  4.9 ±0.3  10.4 ±0.5  82.0 ±0.7  2.7 ±0.9  5± 1  9.4 ±0.5  74.4 ± 0.6 2.5 ±0.8  4± 1  wb  SWC1  Fat  Ash  3  Sample  Crude Protein  Moisture  9.26 ±0.09  a  b  8  (%)  (%)  (%)  [by difference] 5.2 ±0.3  Results are mean values of duplicates ± standard deviation. Results are mean values of triplicates ± standard deviation.  a  b  105  2.3.7  Amino Acid Analysis of Soluble Shrimp Waste Hydrolysates  Even though 80 % of the soluble hydrolysates were crude proteins, as determined from the total nitrogen content, it was not clear how much could be attributed to peptides and free amino acids. Therefore, amino acid analysis was performed on six samples, namely SWA4, SWB4, SWF4, SWP4, SWC4 and SW, where SW represents the dried shrimp waste blend starting material without any incubation. As shown in Table 2.13, samples SWC4 and SW had very similar amino acid compositions with total amino acid contents in soluble hydrolysates being 41.91 % and 39.09 %, respectively. When this was compared to the crude protein content (82 %) obtained from proximate analysis, amino acids only accounted for half. It indicated that these samples contained many non-protein nitrogen compounds so the crude protein content was overestimated during the conversion from total nitrogen content to total crude proteins because the conversion for fish muscle was used. When the total amino acid content of sample SWC4 and SW were compared, SWC4 contained relatively more glycine and proline, which are considered sweet and bittersweet respectively. Sample SW contained high amounts of Asp, Glu, Arg, Leu and Lys, which were similar to the results obtained from He and others (2006). However, the free amino acid content of SW (2.11 %) was lower than that of SWC4 (10.32 %). It could be possible that residual endogenous protease activity existed in the shrimp waste; therefore, more free amino acids were released during the incubation leading to a higher content of free amino acids in sample SWC4. Sample SWC4 contained a high amount of glycine, arginine and proline in free forms, which have been suggested to be important in affecting shrimp flavour (Matsumoto and Yamanaka 1990). All samples produced with proteases had around 60 % of soluble hydrolysates being amino acids, which were higher than the control sample and the starting waste blend. The total amino acid contents of these hydrolysates accounted for about three-quarter of the crude protein content determined from the proximate analysis so these proteasehydrolyzed samples contained less non-protein nitrogen compounds in comparison to the 106  Table 2.13  Amino acid composition of soluble shrimp waste hydrolysates compared to the starting shrimp waste blend (refer to Table 2.2 for sample code details). Amino Acid Content, g/100 g dry sampl e SWA4 SWB4 SWF4 SWP4 SWC4 SW a  Amino acid Asp Glu Ser Gly His Arg Thr Ala Pro Tyr Val Met He  Leu Phe Lys Total (Total Free)  5.64 (0.05) 9.02 (0.28) 2.80 (0.22) 3.29 . (0.61) 1.78 (0.12) 5.66 (0.74) 2.83 (0.36) 3.79 (0.55) 2.74 (0.43) 3.10 (0.31) 3.61 (0.32) 2.04 (0.48) 3.50 (0.15) 5.14 (0.48) 3.37 (0.35) 6.18 (0.27) 64.50 (6.15)  5.82 (0.37) 9.56 11.16 (0.96) (1.00) 2.91 2.51 (0.82) (0.87) 4.32 3.56 (1.34) (1.27) 1.75 1.56 (0.42) (0.43) 6.62 5.46 (2.38) (2.58) 2.33 2.73 (0.62) (0.64) 3.84 4.40 (1.47) (1.56) 3.36 2.77 (0.72) (0.72) 2.30 2.15 (0.94) (0.95) 3.01 3.37 (1.05) (1.04) 1.76 1.60 (0.66) (0.61) 2.80 2.68 (0.85) (0.83) 5.03 4.71 (2.47) (2.43) 2.61 2.47 (1.12) (1.09) 6.30 5.90 (1.35) (1-34).  5.81 (0.05) 9.48 (0.15) 2.82 (0.15) 3.52 (0.72) 1.78 (0.08) 5.71 (0.76) 2.88 (0.28) 3.91 (0.43) 2.93 (0.47) 3.01 (0.22) 3.59 (0.25) 2.02 (0.32) 3.40 (0.17) 5.06 (0.53) 3.30 (0.37) 6.11 (0.24)  3.34 3.82 (0.15) (0.02) 5.88 5.97 (0.27) (0.04) 1.58 1.71 (0.18) (0.03) 4.77 2.08 (2.22) (0.43) 0.83 1.01 (0.10) (0.02) 4.51 3.46 (1.88) (0.43) 1.51 1.48 (0.23) (0.05) 2.27 2.74 (0.78) (0.16) 3.14 1.79 (1.53) (0.33) 1.94 1.34 (0-34) (0.05) 2.15 1.95 (0.38) (0.08) 0.94 1.20 (0.20) (0.01) 2.11 1.59 (0.28) (0.05) 2.83 3.06 (0.49) (0.09) 1.52 1.93 (0.31) (0.06) 3.34 3.20 (0.50) (0.10)  68.42 59.92 (19.63) (20.14)  65.32 (5.52)  41.91 (10.32)  7.00 (0.34)  Major taste of free amino acid  b  Sour Sour Sweet Sweet Bitter Bitter Sweet Sweet Sweet and Bitter Bitter  0  Bitter Bitter Bitter  0  Bitter Bitter Sweet and Bitter  39.09 (2.11)  Amino acid content after acid hydrolysis; free amino acid contents by analysis without hydrolysis are shown in parenthesis. "From Fuke(1994). °From Kirimura and others (1969).  a  107  control sample and the starting waste blend. Since the total amino acid contents of SWA4, SWB4, SWF4 and SWP4 hydrolysate samples were higher than SWC4 and SW, it was expected that there would be higher amounts of each amino acid for each proteasehydrolyzed samples. However, several amino acids, namely aspartic acid, glutamic acid and lysine, were relatively higher in the hydrolysates produced by different proteases. Asp and Glu have been reported to taste sour while Lys tastes both sweet and bitter. When the amino acid compositions of each hydrolysate sample were examined, it was found that amino acid contents such as Tyr, Val, Met, He, Leu, and Phe tended to be higher in the more bitter samples. These hydrophobic and bulky amino acids have been reported to be associated with bitterness elicitation (Ishibashi and others 1987; Ishibashi and others 1988a; Ishibashi and others 1988b). According to results from the sensory evaluation, Alcalase and Protamex samples were the most bitter, followed by samples of bromelain and Flavourzyme and finally the least bitter were the controls. This trend was similarly observed from the hydrophobic amino acid contents with the most bitter samples (SWA4 and SWP4) containing more of Tyr, Val, Met, He, Leu and Phe, followed by relatively lower amounts in samples SWB4 and SWF4 and the lowest values in the control sample. Therefore, it was shown that the higher contents of hydrophobic amino acid contents could be responsible for the bitterness in the soluble whole shrimp waste hydrolysates. The free amino acid content of samples hydrolyzed by different proteases was different. Bromelain and Flavourzyme samples had high amounts (20 %) of free amino acids whereas Alcalase and Protamex samples contained much less free amino acids (6 %). It was reasonable for Flavourzyme sample to contain many free amino acids because Flavourzyme contains exopeptidase activity which cleaves at terminal amino group. However, it was interesting to see that bromelain sample also had high amount of free amino acids. Both Flavourzyme and bromelain samples had high contents of Arg, Leu and Phe in free forms, which indicated that less amount of these amino acids were found in peptide form. Since arginine, leucine and phenylalanine are important components in the elicitation of bitterness in peptides (Ishibashi and others 1988b), a higher content of  108  these amino acids in free forms could further explain why these samples were not high in bitterness. On the contrary, Alcalase and Protamex samples had low contents of free amino acids. Therefore, the high amounts of hydrophobic and bulky amino acid side chains that were observed in the total amino acid contents were mainly found in peptide forms. Having many hydrophobic amino, acids in the peptides would increase the bitterness of the peptides; as a result, these samples tasted intensely bitter. Based on the assumption that a soluble yield of 30 % was achieved from hydrolyzing whole shrimp waste with proteases and in the soluble hydrolysate, 60 % of the content was peptide/amino acid materials, it could be estimated that in 1000 g of wet whole shrimp waste, there would be approximately 45 g of peptides/amino acids. This value was considered high because it represented about 40 % of the original protein materials found in the whole shrimp waste.  109  2.3.8  Conclusions  To summarize, through this study, it was shown that shrimp processing discards had great potential to be used for production of protein hydrolysates due to their high protein contents. Protein hydrolysates of shrimp processing wastes had higher soluble yields when the hydrolysis was done with the assistance of proteases with Alcalase and Protamex samples giving the highest yields with the amount of enzymes used. Hydrolysates produced by Alcalase and Flavourzyme achieved high degree of hydrolysis yielding smaller peptides or even single amino acids. Despite the high yield and high DH, Alcalase samples were not ideal to be used in food products due to their high bitterness unless an economical debittering technique was coupled. Therefore, Flavourzyme samples appeared to be the most promising in producing higher soluble yields and total amino acid contents than the controls but remained relatively similar in bitterness to the controls. The sample SWF8, hydrolyzed with 1:1 water-to-substrate ratio, 3 % Flavourzyme and incubated for 8 hours, produced soluble yield of 18.75 %, high degree of hydrolysis (3.05 meq/g) and low bitterness (1600 ppm caffeine). With this hydrolysis condition, for every 1000 g wet whole shrimp waste, 28 g of peptide/amino acid mixtures could be recovered. This value represented approximately a quarter of the original protein content in the shrimp waste. However, it is necessary to further optimize the conditions required to produce protein hydrolysates of low bitterness using shrimp processing discards. If soluble yield is the main concern of the processor and the application does not require favourable sensory attributes, then Alcalase would be recommended to hydrolyze the shrimp processing discards with a hydrolysis time of 1 hour and 1 % enzyme since time of hydrolysis and percent enzyme were not significant factors affecting soluble yield, DH and bitterness. On the other hand, if the protein hydrolysates are to be used as food flavouring, then Flavourzyme would be suggested to be utilized ih the hydrolysis. Furthermore, sensory evaluation on the shrimp aroma and flavour should also be examined to see which hydrolysis conditions could produce the highest shrimp aroma and flavour but the least bitterness. Experiments could be performed using response surface methodology to 110  explore different factors such as time of incubation, percent enzyme and/or temperature to optimize the process to produce high soluble yield, high DH, high shrimp aroma and flavour but low bitterness. Shahidi and others (1995) have used this methodology to find the optimal temperature and incubation time to produce a high DH in making capelin hydrolysates while Diniz and Martin (1996) explored the effect of pH, temperature and enzyme-substrate ratio on the DH of dogfish muscle hydrolysates using response surface methodology. Even though the bitterness level in Flavourzyme samples could be tolerated in some food application, further studies may need to be carried out to couple the production of protein hydrolysates with simple debittering techniques such as addition of salt or a-cyclodextrin to investigate if bitterness in the samples can be significantly reduced, as suggested by Tamura and others (1990). By having a simple and economical debittering technique, it could broaden the application of the shrimp waste hydrolysates in the food industry.  Ill  .4  References  Aaslyng MD, Martens M, Poll L, Nielsen PM, Flyge H, Larsen LM. 1998. Chemical and sensory characterization of hydrolyzed vegetable protein, a savory flavoring. Journal of Agricultural and Food Chemistry 46: 481-489. Adler-Nissen J. 1979. Determination of the degree of hydrolysis of food protein hydrolysates by trinitrobenzenesulfonic acid. Journal of Agricultural and Food Chemistry 27(6): 1256-1262. Aspmo SI, Horn SJ, Eijsink VGH. 2005. Enzymatic hydrolysis of Atlantic cod (Gadus morhua L.) viscera. Process Biochemistry 40: 1957-1966. Aubes-Dufau I, Seris JL, Combes D. 1995. Production of peptic hemoglobin hydrolysates: bitterness demonstration and characterization. Journal of Agricultural and Food Chemistry 43: 1982-1988. Benjakul S, Morrissey MT. 1997. Protein hydrolysates from Pacific whiting solid wastes. Journal of Agricultural and Food Chemistry 45: 3423-3430. Cavalheiro JMO, deSouza EO, Bora PS. 2007. Utilization of shrimp industry waste in the formulation of tilapia (Oreochromis niloticus Linnaeus) feed. Bioresource Technology 98: 602-606. Charalambous G. 1980. The analysis and control of less desirable flavors in foods and beverages. New York: Academic Press. 358p. Charteris W. 1992. Taguchi's system of experimental design and data analysis: a quality engineering technology for the food industry. Journal of the Society of Dairy Technology 45: 33-49. Cho MJ, Unklesbay N, Hsieh FH, Clarke AD. 2004. Hydrophobicity of bitter peptides from soy protein hydrolysates. Journal of Agricultural and Food Chemistry 52: 5895-5901. Clemente A. 2000. Enzymatic protein hydrolysates in human nutrition. Trends in Food Science and Technology 11: 254-262. Coward-Kelly G, Agbogbo FK, Holtzapple MT. 2006. Lime treatment of shrimp head waste for the generation of highly digestible animal feed. Bioresource Technology 97: 1515-1520. Diaz-Rojas EI, Arguelles-Monal WM, Higuera-Ciapara I, Hernandez J, LizardiMendoza J, Goycoolea FM. 2006. Determination of chitin and protein contents during the isolation of chitin from shrimp waste. Macromolecular Bioscience 6: 340-347. 112  Diniz FM, Martin AM. 1996. Use of response surface methodology to describe the combined effects of pH, temperature and E/S ratio on the hydrolysis of dogfish (Squalus acanthias) muscle. International Journal of Food Science and Technology 31:419-426. Dumay J, Donnay-Moreno C, Barnathan G, Jaouen P, Berge JP. 2006. Improvement of lipid and phospholipids recoveries from sardine (Sardina pilchardus) viscera using industrial proteases. Process Biochemistry 41: 2327-2332. Fisheries and Oceans Canada. 2007. Coastal and Water Statistics. Fast Facts http://www.dfo-mpo.gc.ca/communic/facts-info/facts-info_e.htm Frokjaer S. 1994. Use of hydrolysates for protein supplementation. Food Technology 48: 86-88. Fuke S. 1994. Taste-active components of seafoods with special reference to umami substances. In: Shahidi F, Botta JR, editors. Seafoods: chemistry, processing technology and quality. 1 ed. Glasgow: Blackie Academic and Professional, p115-139. st  Gildberg A, Stenberg E. 2001. A new process for advanced utilisation of shrimp waste. Process Biochemistry 36: 809-812. Gildberg A, Arnesen JA, Carlehog M. 2002. Utilisation of cod backbone by biochemical fractionation. Process Biochemistry 38(4): 475-480. Goldsmith P, Amankwah F, Gunjal K, Smith J. 2003. Financial feasibility of producing value-added seafood from shrimp waste in Quebec. Journal of Aquatic Food Product Technology 12(4): 39-61. Gomez MJ, Garde S, Gaya P, Medina M, Nunez M. 1997. Relationship between level of hydrophobic peptides and bitterness in cheese made from pasteurized and raw milk. Journal of Dairy Research 64: 289-297. Gorden DT, Roberts GL. 1977. Mineral and proximate composition of Pacific coast fish. Journal of Agricultural and Food Chemistry 25(6): 1262-1268. Guerard F, Dufosse L, Broise DDL, Binet A. 2001. Enzymatic hydrolysis of proteins from yellowfin tuna (Thunnus albacares) wastes using Alcalase. Journal of Molecular Catalysis B:Enzymatic 11: 1051-1059. Guerard F, Guimas L, Binet A. 2002. Production of tuna waste hydrolysates by a commercial neutral protease preparation. Journal of Molecular Catalysis B: Enzymatic 19-20: 489-498.  113  He H, Chen X, Sun C, Zhang Y, Gao P. 2006. Preparation and functional evaluation of oligopeptide-enriched hydrolysate from shrimp {Acetes chinensis) treated with crude protease from Bacillus sp. SM98011. Bioresource Technology 97: 385-390. deHolanda HD, Netto FM. 2006. Recovery of components from shrimp (Xiphopenaeus Kroyeri) processing waste by enzymatic hydrolysis. Journal of Food Science 71(5): 298-303. Imm JY, Lee CM. 1999. Production of seafood flavor from red hake (Urophycis chuss) by enzymatic hydrolysis. Journal of Agricultural and Food Chemistry 47: 2360-23666. Ishibashi N, Oho'I, Kato K, Shigenaga T, Shinoda I, Okai H, Fukui S. 1988a. Role of the hydrophobic amino acid residue in the bitterness of peptides. Agricultural and Biological Chemistry 52(1): 91-94. Ishibashi N, Kouge K, Shinoda I, Kanehisa H, Okai H. 1988b. A mechanism for bitter taste sensibility in peptides. Agricultural and Biological Chemistry 52(3): 819827. Ishibashi N, Sadamori K, Yamamoto O, Kanehisa H, Kouge K, Kikuchi E, Okai H, Fukui S. 1987. Bitterness of phenylalanine- and tyrosine-containing peptides. Agricultural and Biological Chemistry 51(12): 3309-3313. Islam MS, Khan S, Tanaka M. 2004. Waste loading in shrimp and fish processing effluents: potential source of hazards to the coastal and nearshore environments. Marine Pollution Bulletin 49: 103-110. Jaswal AS. 1989. Methodology investigations for the production of amino acid hydrolysate from shrimp waste. Canadian Institute of Food Science and Technology Journal 22(5): 460-463. Kim MR, Choi SY, Kim CS, Kim CW, Utsumi S, Lee CH. 1999. Amino acid sequence analysis of bitter peptides from a soybean proglycinin subunit synthesized in Escherichia coli. Bioscience, Biotechnology, and Biochemistry 63(12): 2069-2074. Kim MR, Kawamura Y, Lee CH. 2003. Isolation and identification of bitter peptides of tryptic hydrolysate of soybean US glycinin by reverse-phase high performance liquid chromatography. Journal of Food Science 68(8): 24162422. Kirimura J, Shimizu A, Kimizuka A, Ninomiya T, Katsuya N. 1969. The contribution of peptides and amino acids to the taste of foodstuffs. Journal of Agricultural and Food Chemistry 17(4): 689-695.  114  Kristinsson HG, Rasco BA. 2000. Fish protein hydrolysates: production, biochemical, and functional properties. Critical Reviews in Food Science and Nutrition 40(1): 43-81. Lee JY, Lee HD, Lee CH. 2001. Characterization of hydrolysates produced by mildacid treatment and enzymatic hydrolysis of defatted soybean flour. Food Research International 34: 217-222. Lee KD, Warthesen JJ. 1996. Mobile phases in reverse-phase HPLC for the determination of bitter peptides in cheese. Journal of Food Science 61(2): 291294. Liaset B, Julshamn K, Espe M. 2003. Chemical composition and theoretical nutritional evaluation of the produced fractions from enzymic hydrolysis of salmon frames with Protamex . Process Biochemistry 38: 1747-1759. M  Liceaga-Gesualdo AM, Li-Chan ECY. 1999. Functional properties offishprotein hydrolysate from herring (Clupea harengus). Journal of Food Science 64(6): 1000-1004. Liu CC, Morioka K, Itoh Y, Obatake A. 2000. Contribution of lipid oxidation to bitterness and loss of free amino acids in the autolytic extract from fish wastes: Effective utilization of fish wastes. Fisheries Science 66: 343-348. Maehashi K, Matsuzaki M, Yamamoto Y, Udaka S. 1999. Isolation of peptides from an enzymatic hydrolysate of food proteins and characterization of their taste properties. Bioscience, Biotechnology, and Biochemistry 63(3): 555-559. Mandeville S, Yaylayan V, Simpson BK. 1992. Proximate analysis, isolation and identification of amino acids and sugars from raw and cooked commercial shrimp waste. Food Biotechnology 6: 51-64. Mandeville S, Yaylayan V, Simpson B, Pare JRJ. 1991. Gas chromatography-mass spectrometry analysis of flavour-active compounds from raw commercial shrimp waste. Spectroscopy 9: 61-72. Matsumoto M, Yamanaka H. 1990. Changes in contents of glycolytic metabolites and free amino acids in the muscle of kuruma prawn during storage. Nippon Suisan Gakkaishi 56: 1515-1520. Mizani M, Aminlari M, Khodabandeh M. 2005. An effective method for producing a nutritive protein extract powder from shrimp-head waste. Food Science and Technology International 11(1): 49-54. Mukhin VA and Novikov VY. 2001. Enzymatic hydrolysis of proteins from crustaceans of the Barents Sea. Applied Biochemistry and Microbiology 37(5): 538-542.  115  Ney KH. 1979. Bitterness of peptides: amino acid composition and chain length. In: Boudreau JC, Editor. Food taste chemistry, ACS Symposium Series vol. 115. Washington, DC: American Chemical Society, p 149-173. Nilsang S, Lertsiri S, Suphantharika M, Assavanig A. 2005. Optimization of enzymatic hydrolysis of fish soluble concentrate by commercial proteases. Journal of Food Engineering 70: 571-578. Nishitoba T, Sato H, Sakamura S. 1988. Bitterness and structure relationship of the triterpenoids from Ganoderma lucidum (Reishi). Agricultural and Biological Chemistry 52(7): 1791-1795. Pommer K. 1995. New proteolytic enzymes for the production of savory ingredients. Cereal Foods World 40: 745-748. Ravallec-Ple R, Gilmartin L, Wormhoudt AV, Gal YL. 2000. Influence of the hydrolysis process on the biological activities of protein hydrolysates from cod (Gadus morhud) muscle. Journal of the Science of Food and Agriculture 80: 2176-2180. Rebeca BD, Pena-Vera MT, Diaz-Castaneda M. 1991. Production of fish protein hydrolysates with bacterial proteases: yield and nutritional value. Journal of Food Science 56(2): 309-314. Schmidl MK, Taylor SL, Nordlee JA. 1994. Use of hydrolysate-based products in special medical diets. Food Technology 48: 77-80,85. Shahidi F. 1994. Seafood processing by-products. In: Shahidi F, Botta JR, editors. Seafoods: chemistry, processing technology and quality. 1 ed. Glasgow: Blackie Academic and Professional, p. 320-334. st  Shahidi F, Han XQ, Synowiecki J. 1995. Production and characteristics of protein hydrolysates from capelin (Mallotus villosus). Food Chemistry 53: 285-293. Shahidi F, Synowiecki J. 1991. Isolation and characterization of nutrients and valueadded products from snow crab (Chinoecetes opilio) and shrimp {Pandalus borealis) processing discards. Journal of Agricultural and Food Chemistry 39: 1527-1532. Simpson BK, Nayeri G, Yaylayan V, Ashie LNA. 1998. Enzymatic hydrolysis of shrimp meat. Food Chemistry 61: 131-138. Smyth M, FitzGerald RJ. 1998. Relationship between some characteristics of WPC hydrolysates and the enzyme complement in commercially available proteinase preparations. International Dairy Journal 8: 819-827.  116  Spellman D, O'Cuinn G, FitzGerald RJ. 2005. Physicochemical and sensory characteristics of whey protein hydrolysates generated at different total solids levels. Journal of Dairy Research 72: 138-143. Synowiecki J, Al-Khateeb NAAQ. 2000. The recovery of protein hydrolysate during enzymatic isolation of chitin from shrimp Crangon crangon processing discards. Food Chemistry 68: 147-152. Tamura M, Mori N, Miyoshi T, Koyama S, Kohri H, Okai H. 1990. Practical debittering using model peptides and related compounds. Agricultural and Biological Chemistry 54(1): 41-51. Wrolstad RE, Acree TE, An H, Decker EA, Penner MH, Reid DS, Schwartz SJ, Shoemaker CF, Sporns P. 2003. Current Protocols In Food Analytical Chemistry. John Wiley & Sons, Inc.  117  C H A P T E R III  Bitter Fractions from EnzymaticallyProduced Hydrolysates of Commercial Shrimp Processing Waste  3.1  2  Introduction  Shrimp processing discards contain a number of flavour-active components that can be incorporated into foods to produce seafood flavour (Mandeville and others 1991; Mandeville and others 1992). Moreover, approximately 40 % of the shrimp waste dry weight is protein (Lekshmy and Prabhu 1989; Shahidi and Synowiecki 1991; Shahidi 1994; Synowiecki and Al-Khateeb 2000; deHolanda and Netto 2006), which increases its nutritive value. Despite these many advantages of using shrimp processing waste as food flavouring, the intense bitterness that is usually elicited after enzymatic hydrolysis is generally unacceptable to consumers. Numerous studied have proposed that bitter peptides are responsible for the bitterness in protein hydrolysates (Aubes-Dufau and others 1995; Lee and Warthesen 1996; Kim and others 1999; Maehashi and others 1999; Kim and others 2003; Cho and others 2004). Bitter peptides are produced after extensive hydrolysis because the hydrophobic residues, which are originally buried in the interior of the protein molecules, become exposed and interact with human taste receptors (Pedersen 1994). Several criteria have been suggested to cause bitterness sensation in humans. Bitter peptides are generally small in size with molecular weight between 1-6 kDa (Aubes-Dufau and others 1995; Kristinsson and Rasco 2000; Cho and others 2004). In addition, hydrophobicity of amino acid residues has been reported to be essential to elicit bitterness (Ishibashi and others 1987; Ishibashi and others 1988a; Ishibashi and others 1988b; Nishitoba and others 1988; Aubes-Dufau and others 1995; Gomez and others 1997; Kim and others 1999). Furthermore, there have been two proposed determinant sites for bitter taste. A primary site, which is important for binding, usually contains bulky hydrophobic group of three or more carbons while a 2  A version of this chapter may be submitted for publication.  118  secondary determinant site, which is estimated to be 4.1 A away from the primary site, often has either a bulky basic group or a hydrophobic group (Ishibashi and others 1988b). Since bitterness in hydrolysates is a prominent problem that needs to be addressed, several studies have been done to isolate and identify the bitter peptides (Aubes-Dufau and others 1995; Kim and others 1999; Kim and others 2003; Cho and others 2004) while others have attempted different debittering techniques (Lalasidis and Sjoberg 1978; Minagawa and Others 1989; Pedersen 1994; Yeom and others 1994; Izawa and others 1997; Saha and Hayashi 2001; FitzGerald and O'Cuinn 2006). Even though these studies provided evidence and potential strategies to produce hydrolysates that are non-bitter, the nature of the substrate can greatly affect the peptide profile that is generated after hydrolysis. Therefore, in this study, the goal is to isolate bitter fractions, through fractionation by size and hydrophobicity, from an enzymatically produced protein hydrolysate of shrimp processing waste to understand the characteristics of the peptides causing strong bitterness in the sample. The amino acid composition of some bitter fractions is compared to investigate if certain amino acid residues are responsible for the bitterness.  119  3.2  3.2.1  Materials and Methods  Materials  Hydrolysate samples were produced as described in Section 2.2.4. Sephadex G-25 media was purchased from Amersham Biosciences (Uppsala, Sweden). The Acrodisc syringe filter with 0.45 urn HT Tuffryn Membrane was purchased from VWR International (Ontario, Canada). The Prep/Scale Spiral Wound ultrafiltration modules including the Prep/Scale holder and three different PL regenerated cellulose Prep/Scale-TFF-1 membranes with molecular weight cut-off (MWCO) of 10, 3 and 1 kDa were purchased from Millipore Corporation (Massachusetts, USA). A HiTrap™ HIC selection kit consisting of 6 X 1 mL HIC columns (Phenyl Sepharose™ 6 Fast Flow (low sub), Phenyl Sepharose™ 6 Fast Flow (high sub), Phenyl Sepharose™ High Performance, Butyl-S Sepharose™ 6 Fast Flow, Butyl Sepharose™ 4 Fast Flow and Octyl Sepharose™ 4 Fast Flow), and 25 mL of Phenyl Sepharose™ 6 Fast Flow (low sub) media, were purchased from GE Healthcare Bio-sciences Inc. (Quebec, Canada). Ammonium sulfate was purchased from EMD Chemicals Inc. (Damstadt, Germany). Sodium phosphate monobasic USP/FCC and sodium phosphate dibasic USP were purchased from Fisher Scientific Co. (Ontario, Canada). Sodium hydroxide was purchased from BDH Inc. (Ontario, Canada). Lysozyme (14300 Da), vitamin B12 (1355.4 Da) and glycine (75.07 Da) were purchased from Sigma Aldrich (Ontario, Canada) while Type I anti-freeze protein (AFPI, 3240 Da) was donated by Dr. G. Fletcher from A/F Protein Canada Inc. (Newfoundland, Canada).  120  3.2.2  Gel Filtration Chromatography of Soluble Hydrolysates Produced from Liquid Waste Blend  Sephadex G-25 media was swollen in degassed distilled deionized water at 4°C overnight, then packed into a column with a diameter of 2.6 cm and a height of 39.0 cm to yield 207 mL bed volume. Dried soluble liquid waste blend hydrolysate samples were prepared in 5.0 mL ddFL:0, with a sample concentration of 50 mg/mL. The sample solution was filtered through an Acrodisc syringe filter with 0.45 pm HT Tuffryn Membrane before loading to the column, which was then filled manually with distilled deionized H2O, the eluent, before connecting to the pump with a flow rate of 1.5 mL/min and fraction collector with a collection rate of 4.0 min/tube. A hundred test tubes were collected for each sample and ultraviolet absorbance at 214 and 280 nm was measured using the Shimadzu UV-160 UV-Visible Recording Spectrophotometer (Mandel Scientific Company Inc.; Ontario, Canada). The absorbance readings were plotted and the molecular weight profile of each sample was obtained. Specific fractions were collected and freeze-dried for further analysis, including amino acid analysis as described in Section 2.2.9 and informal sensory evaluation by two panelists to isolate specific bitter  fractions by tasting dried solids or dried solids dissolved in 1 mL of ddFkO. 3.2.3  Fractionation of Soluble Hydrolysates from Whole Shrimp Waste.  A flow chart of the fractionation and analysis procedures used in the main thesis study to isolate bitter fractions from soluble whole shrimp waste hydrolysates is outlined in Figure 3.1. 3.2.3.1  Ultrafiltration  Ten grams of freeze-dried soluble shrimp waste hydrolysates (specifically SWA4, SWA8, SWB8, SWB24, SWC8 and SWF8) was dissolved in 1 L of ddH 0. These samples were 2  121  Dried and Soluble Whole Shrimp Waste Hydrolysate SWA4, SWA8, SWB8, SWB24, SWC8, SWF8 SWA4 Ultrafiltration (1, 3 and 10 kDa membranes)  Gel Filtration Chromatography (Sephadex G-25)  SWA4 Hydrophobic Interaction Chromatography (Phenyl low sub)  I  Informal Sensory Evaluation  Amino Acid Analysis  Gel Filtration Chromatography (Sephadex G-25)  Figure 3.1  Schematic of separation and analysis procedures adapted for fractionation of bitter peptides from soluble whole shrimp waste hydrolysate in the current study (refer to Table 2.2 for sample code details).  122  selected because they had varying levels of bitterness ranging from 800 to 2400 ppm caffeine (Table 2.7). The sample solution was ultra-filtered through the Prep/Scale Spiral Wound ultrafiltration modules including the Prep/Scale holder and three different PL regenerated cellulose Prep/Scale-TFF-1 membranes with molecular weight cut-off (MWCO) of 10, 3 and 1 kDa. Ultrafiltration was performed at a flow rate of 220 mL/min with a peristaltic pump (Cole-Parmer Instrument Company; Quebec, Canada) and was continued until the sample solution volume was lowered to about 10 % of the original volume. The retentate from each MWCO membrane was kept as a fraction while the permeate was consecutively ultra-filtered through 3 kDa and 1 kDa membranes, yielding 4 fractions: >10, 3-10, 1-3, and <1 kDa. These fractions were separately freeze-dried and the yields were recorded. 3.2.3.2  Hydrophobic Interaction Chromatography (HIC)  A soluble whole shrimp waste hydrolysate SWA4 (2:1 water-to-substrate ratio, 4 % enzyme, 4-hour hydrolysis, Alcalase) with a rated bitterness of 2300 ppm was selected as the test sample to examine the ability of each of the columns in the HiTrap™ HIC selection kit to separate components in the sample according to hydrophobicity. A sample solution of 50 mg/mL was prepared in 1.0 M ammonium sulfate salt in 50 mM sodium phosphate buffer at pH 7.0 and syringe filtered. The column was equilibrated with 10 mL of 1.0 M ammonium sulfate in 50 mM sodium phosphate buffer at pH 7.0 at a flow rate of approximately 4 mL/min after which 1 mL of the sample solution was loaded to the column. The column was then washed with 10 mL of buffer containing 1.0 M ammonium sulphate, followed by stepwise elution with buffers (5 mL each containing 0.5 M, 0.25 M, 0.10 M, or no ammonium sulphate) and finally with 5 mL of 70 % ethanol. The whole fractionation procedure was carried out at a flow rate of 1 mL/min and a collection rate of 1 mL/tube. At the end of each run, the column was washed with 10 mL of  ddH20  followed by 10 mL 0.01 N NaOH and re-equilibrated with 10 mL of buffer with 1.0 M ammonium sulfate. An aliquot of 200 uL of each collected fraction was pipetted into a 96-well UV Flat Bottom microtiter plate (Thermo Electron Corporation; Massachusetts^ USA) to measure ultra violet absorbance at 214 and 280 nm using the Varian Cary 50 Bio  123  UV-Visible Spectrophotometer with the Cary 50 MPR microplate reader and the Cary WinUV Software (Varian Australia Pty Ltd.; Victoria, Australia). The absorbance readings were plotted and separation profile of each sample was obtained. For five of the columns, most of the sample components appeared in the fraction eluting with 1.0 M ammonium sulfate salt in 50 mM sodium phosphate buffer at pH 7.0, which indicated no binding to the columns. Hi-Trap™ Phenyl (low sub) was the only column that showed potential binding with the samples; therefore, Phenyl Sepharose™ 6 Fast Flow (low sub) media was used to build a column with a diameter of 1.1 cm and a height of 23 cm with a bed volume of 22 mL. One gram of freeze-dried ultrafiltered fraction was dissolved in 20 mL of 1.0 M ammonium sulfate salt in 50 mM sodium phosphate buffer at pH 7.0 and the eluting profile was the same as mentioned above with several exceptions. The volume used in the elution was scaled up 20 times because of the larger bed volume of the preparative column. The operating flow rate was adjusted to 4.0 mL/min and the fraction collection rate to 8 mL/tube. Specific fractions were collected and freeze-dried for further analysis. The only exception was the ethanol fraction, which was alternatively rotary evaporated with the BUCHI Rotavapor R-l 14 system with the BUCHI waterbath B-480 (Brinkmann Instruments Inc.; New York, USA) set at 50 °C and coupled to the Welch self-cleaning dry vacuum system model 2026 (Gardner Denver Welch Vacuum Technology; Illinois, USA) to remove most of the ethanol before freeze-drying. 3.2.3.3  Gel Filtration Chromatography (GF)  Sephadex G-25 media was swollen as described in Section 3.2.2 and was packed with a diameter of 2.6 cm, a height of 55 cm and a bed volume of 290 mL. Fractions isolated from hydrophobic interaction chromatography were prepared in 10 - 20 mL ddH 0 to 2  reach a sample concentration rangingfrom10 - 410 mg/mL. The concentration was calculated based on total solid contentsfromeach fraction and most fractions contained a large amount of salt. The sample solution was filtered through a syringe filter and loaded into the column, which was then filled manually with distilled deionized H 0, the eluent, 2  124  to the top before connecting to the pump with a flow rate of 1.0 mL/min and fraction collector (Bio-rad Laboratories Ltd.; Ontario, Canada) with a collection rate of 3.0 min/tube. 150 test tubes were collected for each sample. Absorbance at 214 and 280 nm was measured and the absorbance readings were plotted to give the molecular weight profile of each sample. Specific fractions were collected and freeze-dried for further analysis. Molecular weight markers lysozyme (14300 Da), anti-freeze protein I (3240 Da), vitamin Bn (1355.4 Da) and glycine (75.07 Da) were used to calibrate the column. 3.2.4  Sensory Evaluation of Fractionated Shrimp Waste Hydrolysates  Due to the limited amount of sample fractions available for tasting, this sensory evaluation was done informally by two panelists to investigate if specific fractions carry bitterness. For most of the fractions, dried solids were tasted but some were re-dissolved in 0.5 mL of ddFFiO because negligible amounts of solids were recovered. This test only evaluated whether bitterness was tasted in the samples but provided no quantitative information. 3.2.5  Amino Acid Analysis of Fractionated Shrimp Waste Hydrolysates  Selected hydrolysate samples and fractions were sent to analyze the amino acid composition as described in Section 2.2.9.  125  3.3  3.3.1  Results and Discussion  Preliminary Studies with Liquid Waste Blend Hydrolysates  Four shrimp waste hydrolysate samples, SWA-1, SWB-1, SWF-1 and SWP-1, were produced using liquid waste blend (Section 2.2.4). These samples were fractionated using gel filtration chromatography. Figure 3.2 shows the gel filtration chromatograms of the four hydrolysates monitored by absorbance at 214 nm. In general, there were two main areas of interest, one appeared earlier in the elution while the other eluted towards the end. Since molecular weight markers were not used at this stage, the only information regarding molecular size was the separation range for Sephadex G-25, which was 1-5 kDa. The high molecular weight fraction for all hydrolysate samples had very high absorbance, indicating the presence of many peptides. When the absorbance of the different samples was compared, it was found that SWF-1 appeared to have more small peptides and/or free amino acids, which was reasonable because Flavourzyme is known to contain exopeptidase activity that would lead to smaller peptides or amino acids (Pommer 1995). SWA-1 and SWP-1 also had high absorbance earlier in the elution but relatively lower absorbance towards the end of elution in comparison to Flavourzyme sample. SWB-1, on the other hand, had relatively lower absorbance beyond 200 mL so most materials in the hydrolysate produced by bromelain were large. When the chromatograms were examined based on the absorbance at 280 nm, there were some variation in molecular weight distribution among samples that could potentially explain why some samples were more bitter than others (refer to Table 2.5 for relative bitterness rank for each hydrolysate sample). Absorbance at 280 nm usually indicates the existence of aromatic compounds such as tyrosine and phenylalanine, which have been suggested to be involved in bitterness elicitation (Ishibashi and others 1987). The chromatograms were divided into 6 regions as shown in Figure 3.3. Region 1 represents substances with high molecular weight. As can be observed, SWA-1 and SWP-1 had the  126  Figure 3.2  Molecular weight distribution of various soluble liquid waste blend 1-hour hydrolysates at 214 nm from a column of Sephadex G-25. Sample of concentration of 50 mg/mL was loaded in one 5 mL fraction and elution was performed at  to  1.5 mL/min with each fraction containing 6 mL.  0  50  100  150  200  250  300  350  400  450  Volume (mL)  Figure 3.3  Molecular weight distribution of various soluble liquid waste blend 1-hour hydrolysates at 280 nm from a column of Sephadex G-25. Sample of concentration of 50 mg/mL was loaded in one 5 mL fraction and elution was performed at 1.5 mL/min with each fraction containing 6 mL.  highest absorbance in region 1, followed by SWB-1 and SWF-1 with the lowest. This showed that Alcalase and Protamex samples had a lot of large peptides with aromatic side chains, which could possibly explain why these samples were high in bitterness. Aromatic amino acids are bulky and hydrophobic, which are the main property associated with bitterness (Ishibashi and others 1987; Ishibashi and others 1988a; Ishibashi and others 1988b; Nishitoba and others 1988; Aubes-Dufau and others 1995; Gomez and others 1997; Kim and others 1999). Flavourzyme, on the other hand, contained fewer amounts. Although absorbance in region 2 for different hydrolysate samples showed differences, the intensities were very low. Regions 3 and 4 showed high absorbance at 280 nm and different samples gave different elution behaviour. Bromelain and Flavourzyme samples had relatively lower absorbance than samples hydrolyzed by Alcalase and Protamex in these regions indicating samples SWB-1 and SWF-1 contained comparatively less aromatic residues in this weight separation range, which could be why these samples were not as bitter as Alcalase and Protamex samples. Two peaks could be observed for all the samples in these regions but Protamex sample showed the most distinct peaks. Peaks in regions 5 and 6 represented materials of small molecular weight. It was clear that bromelain sample did not contain many aromatic compounds that were very small. Alcalase sample had some aromatic residues but Flavourzyme and Protamex samples contained slightly more. When both Figures 3.2 and 3.3 were looked at together, some interesting trends could be observed. In region 2, which was the elution volume of around 100- 145 mL, all samples showed low absorbance at 280 nm but high absorbance at 214 nm indicating that even though a lot of peptides were present in this molecular weight range, most of them were non-aromatic. The same trend was observed for region 5, with elution volume of about 200 - 230 mL. When examining the elution volume of 270 - 400 mL, some difference between samples could be seen. Sample SWF1 had the strongest absorbance at 214 nm among all samples but its absorbance at 280 nm was not higher than other samples. It showed that although Flavourzyme hydrolyzed proteins into smaller peptides or amino acids, it did not produce many peptides that had aromatic residues. While Protamex sample had relatively lower absorbance at 214 nm than the hydrolysate produced by  129  Flavourzyme, it showed slightly stronger absorbance at 280 nm. Therefore, Protamex sample had more peptides or amino acids that contained aromatic compounds. Alcalase and bromelain samples showed lower absorbance at 214 and 280 nm. In order to examine if the difference in molecular weight profile of constituents in the hydrolysates affected the sensory qualities of the soluble liquid waste blend hydrolysates, an informal sensory tasting was done on the fractions. Tasting results were tabulated in Tables 3.1. Fraction 6 for all samples and fraction 5 for some samples were not tasted due to extremely low product recovery. Fraction one contained stale shrimp or shrimp-like flavour for all hydrolysate samples with flavour in SWF-1 being more favourable. For Alcalase sample, fractions 2 - 4 contained varying level of bitterness whereas only fraction 3 was bitter in the hydrolysate produced by bromelain. Fractions 1 - 4 of the Flavourzyme sample all had shrimp flavour while fractions 2, 3 and 5 possessed a low level of bitterness. Protamex sample, interestingly, had fractions (3 - 4) that were only salty. In addition to sensory evaluation, some fractions, specifically SWA4-1 fraction 1, SWB-1 fraction 2 and SWF-1 fraction 1, of soluble liquid waste blend hydrolysates produced by different enzymes collected according to absorbance at 280 nm from preliminary gel filtration chromatography were freeze-dried and sent for amino acid analysis. Results of the analysis are tabulated in Table 3.2. It was interesting to see that fraction 1 isolated from SWA4-1 contained considerable amounts of amino acids while fraction 1 from SWF-1 only had a quarter of the total solids being amino acids. SWA4-1 fraction 2, which was evaluated as bitter with stale shrimp taste, contained approximately 62 % amino acids in the sample supporting the hypothesis that bitterness is mainly associated with the amino acids and/or peptides. Fraction 1 isolated from SWB-1 appeared to contain no detectable flavour but this fraction comprised around 45 % amino acids whereas the fraction 1 from SWF-1 with shrimp flavour and aroma consisted of only 25 %. Shrimp taste is usually caused by flavour-active compounds, which are not peptides (Mandeville and others 1992) although some amino acids such as Ala, Arg, Gly, Glii and Pro have been associated with shrimp flavour (Matsumoto and Yamanaka (1990).  130  Table 3.1  Sensory qualities of fractions isolated from each protease-hydrolyzed sample of soluble liquid waste blend collected according to absorbance at 280 nm from a column of Sephadex G-25 (refer to Table 2.1 for sample code details).  Fraction Sample  1  2  SWA-1  Stale shrimp  Stale shrimp, bitter  SWB-1  Stale shrimp  No taste  3  4  Slight bitter Bitter, salty Bitter  SWF-1  Shrimp  Strong Slight bitter, shrimp, salty, sweet, slight shrimp bitter  SWP-1  Stale shrimp  Slight bitter  Salty  5  n/d  No taste  n/d  Shrimp  Slight bitter, undesirable unknown taste  Salty  n/d  n/d means value not determined.  131  Table 3.2  Amino acid composition of fraction 1 of SWA-1, fraction 2 of SWB-1 and fraction 1 of SWF-1 isolated from soluble liquid waste blend hydrolysates according to absorbance at 280 nm from a column of Sephadex G-25.  Amino acid  SWA-1 SWB-1 SWF-1 Fraction #2 Fraction #2 Fraction #1 Bitter, stale No taste Shrimp shrimp taste Amino Acid Content (g/100 g dry sample)  Major taste of free amino acid a  Asp  8.73  6.11  3.89  Sour  Glu  11.89  8.85  4.93  Sour  Ser  2.29  2.08  1.39  Sweet  Gly  2.93  2.13  1.63  Sweet  His  1.24  1.11  0.54  Bitter  Arg  4.32  3.95  1.75  Bitter  Thr  2.44  1.89  1.33  Sweet  Ala  2.98  2.71  1.43  Sweet  Pro  4.06  2.10  1.86  Sweet and Bitter  Tyr  0.95  1.56  0.88  Bitter"  Val  4.20  2.47  1.21  Bitter  Met  1.26  1.05  0.37  Bitter  He  3.74  1.92  0.91  Bitter"  Leu  4.26  3.10  1.17  Bitter  Phe  1.46  1.62  0.55  Bitter  Lys  4.96  2.52  0.70  Sweet and Bitter  Total  61.73  45.14  24.57  From Fuke(1994). From Kirimura and others (1969).  a  b  132  Therefore, it was reasonable that the fraction containing shrimp flavour had less amino acid materials. When the amino acid composition of the fractionated hydrolysate samples was compared, it was remarkable to find that the bitter fraction consisted of high amounts of valine, isoleucine and lysine. These amino acids contain long hydrocarbon side chains which contribute to their high hydrophobicity. Also, they have been associated with bitter taste (Fuke 1994). Since hydrophobicity and bulkiness have been suggested to be responsible for bitterness in peptides, it was not surprising that the bitter fraction contained highly hydrophobic amino acids (Ishibashi and others 1987; Ishibashi and others 1988a; Ishibashi and others 1988b; Nishitoba and others 1988; Aubes-Dufau and others 1995; Gomez and others 1997; Kim and others 1999). On the contrary, when the fraction with shrimp flavour was examined, it was notable that this fraction contained high contents of Ser, Gly, Thr and Pro, which are amino acids that are considered sweet. These sweet amino acids together may enhance the sweet shrimp taste of the fraction. These preliminary results illustrated the potential of using separation technique to isolate bitter peptides and that hydrophobic peptides were present in larger amount in the bitter hydrolysates. These data together served as the basis for experimental setup for the fractionation work. 3.3.2  Fractionation of Bitter Shrimp Waste Hydrolysates by Ultrafiltration  Six soluble whole shrimp waste hydrolysate samples, specifically SWA4 (2300 ppm), SWA8 (2400 ppm), SWB8 (1400 ppm), SWB24 (1900 ppm), SWC8 (800 ppm) and SWF8 (1600 ppm), were selected based on their bitterness scores for further fractionation. These samples were ultrafiltered to yield 4 fractions: >10 kDa, 3-10 kDa, 1 - 3 kDa and < 1 kDa.  133  The percent total solid recovery for the hydrolysate samples after ultrafiltration were similar with an average value of 65 ± 4 % (Table 3.3). The two samples hydrolyzed by Alcalase did not differ ih terms of solid yields recovered from ultrafiltration, suggesting that an increase in incubation time from 4 hours to 8 hours for soluble whole shrimp waste with Alcalase did not continue to hydrolyze further. This could imply that the enzyme had already cleaved all the peptides that it could possibly hydrolyze. Since each protease has its specific cleavage sites, it cannot hydrolyze further if none of the peptides in the mixture contain amino acid residues at the appropriate C- or N-terminus. Samples SWA4 and SWA8 had over 80 % relative yield in fractions smaller than 10 kDa while most solids were recovered in the greater than 3 kDa fractions for bromelain samples. The control sample had over 50 % solids in the > 10 kDa fraction, which was rational because the control sample was not expected to be hydrolyzed unless residual endogenous enzyme activity existed. The soluble hydrolysate produced by Flavourzyme gave a relative percent yield of 41 % in the < 1 kDa fraction, which was the highest among all the samples examined. This further supported its possession of exopeptidase activity. After ultrafiltration, it was expected that bitterness would reside in one of the fractions. However, informal sensory tasting showed otherwise, with most fractions from UF containing some bitterness. As a result, only one sample, SWA4 soluble hydrolysate (2:1 water-to-substrate, 4 % enzyme, 4-hour hydrolysis and Alcalase), was selected to be further fractionated by chromatographic methods. According to the findings in chapter II, Alcalase samples gave high product yields and DH, which led to great potential of using these samples for industry-scale production for high profits. However, it is necessary to understand the characteristics of the peptide mixtures in these samples to devise proper methodology to lower their bitterness. Therefore, through further fractionating by hydrophobicity and size, the peptides responsible for bitterness could be characterized, as reported by several studies (Aubes-Dufau and others 1995; Kim and others 1999; Kim and others 2003; Cho and others 2004).  134  Table 3.3  Product yields and their relative distribution of some soluble shrimp waste hydrolysates fractionated using ultrafiltration of 1 % (w/v) sample solutions through 10, 3 and 1 kDa membranes with a flow rate of 220 mL/min (refer to Table 2.2 for sample code details). Product yield in grams (relative percent yield)  3  > 10 kDa  3-10 kDa  1-3kDa  <lkDa  Sample  Total yield (g)  Percent total solid recovery (%) b  SWA4  1.24(19)  1.65 (25)  1.69 (25)  2.11 (32)  6.69  67  SWA8  1.20(19)  1.68 (26)  1.71 (26)  1.89 (29)  6.48  65  SWB8  2.63 (37)  2.98 (42)  0.58 (8)  0.91 (13)  7.10  71  SWB24  2.03 (32)  2.76 (43)  0.82(13)  0.78 (12)  6.39  64  SWC8  2.65 (56)  0.68(14)  0.45(10)  0.93 (20)  4.71  59  SWF8  1.51 (24)  1.22 (20)  0.95 (15)  2.55 (41)  6.23  62  Calculated based on total solid recovery. Calculated from dividing total yield (g) from UF by solid content in the 1% (w/v) sample solution of soluble shrimp waste hydrolysates (10 grams each, except SWC8, which only contained 8 grams). b  135  3.3.3  Further Fractionation Of Ultrafiltered Fractions from SWA4 Soluble Shrimp Waste Hydrolysates by Hydrophobic Interaction Chromatography (HlC)  Initial screening indicated that the Phenyl-substituted column with low ligand activity was able to separate the hydrophobic substances from the soluble shrimp waste hydrolysates. It could be observed that there were three main areas in the HIC chromatogram of UF > 10 kDa fraction of SWA4 (Figure 3.4). The first region, which eluted with the equilibration buffer containing 1.0M (NH^SO/t, showed very strong absorbance at both 214 and 280 nm. The second region displayed the elution pattern of decreasing (NH ) S04 4  2  concentration while the third region showed the strongly bound materials that  required the use of 70 % ethanol to elute. Fractions 1 and 7 absorbed strongly at 280 nm indicating the presence of aromatic compounds. Fractions 2-6 only showed absorbance at 214 nm so they contained substances containing carbonyl group but no aromatic constituents. Seven fractions were collected but solids could be recovered only from fractions 1, 2, 3 and 7. Fractions 1-3 could not be tasted to evaluate the bitterness due to their high salt content. Therefore, all four fractions were further separated by gel filtration chromatography in an attempt to purify the peptides as well as to remove salt from these fractions. HIC chromatograms of Other UF fractions of soluble shrimp waste hydrolysate of SWA4 could be found in Appendices B - D.  136  V o l u m e (mL) Figure  3.4  Hydrophobic interaction chromatogram of soluble shrimp waste hydrolysate of SWA4 UF > 10 kDa on Hi-Trap  1M  Phenyl (low sub) column. One gram of soluble hydrolysate was loaded in a volume of 20 mL of 1 .OM ( N H 4 ) 2 S 0 4 in sodium phosphate buffer at pH 7 eluting stepwise with decreasing concentration of ( N H 4 ) 2 S 0 4 from 1.0, 0.5, 0.25, 0.10 ^  to 0 M, followed by 70% ethanol. The elution was performed at 4 mL/min and fractions were collected every 8 mL.  3.3.4  Further Fractionation of HIC Fractions from Ultrafiltered Fractions of SWA4 Soluble Shrimp Waste Hydrolysates by Gel Filtration Chromatography  Fractions 1, 2, 3 and 7 from HIC of the SWA4 UF > 10 kDa sample were further fractionated by gel filtration chromatography and their molecular weight profiles are shown in Figures 3.5-3.8. HIC fraction 1 gave three major areas with an increase in absorbance with decreasing molecular weight (Figure 3.5). Based on calibration of the GF column with standard markers, GF fraction 1 had molecular weight of 3300 - 3700 Da while fraction 2 contained materials of size 2800 - 3200 Da. GF fraction 3 consisted of substances of molecular weight 1900 - 2800 Da. Most materials were eluted in the third area that had molecular weight of less than 1500 Da. Fractions 4-6 showed high absorbance at 280 nm, which indicated the presence of aromatic compounds. HIC fraction 2, as expected from having very low absorbance at 280 nm during HIC fractionation, did not contain any aromatic substances. There were two major peaks, one fell in the molecular weight range of 1700 - 3300 Da and the other had mass below 1700 Da. HIC fraction 3 displayed two main areas in the GF chromatogram, one of 3200 - 3900 Da while the other below 1500 Da. These isolated materials did not show absorbance at 280 nm. HIC fraction 7 gave nine fractions from gel filtration chromatography and it was interesting that peaks were observed after elution of one full column volume, which was around 230 mL. Distinct peaks were found at molecular weight of around 4000 Da and 1400 Da. It should be noted that all peaks in this GF chromatogram absorbed strongly at  138  50  100  150  200  250  300  350  400  450  V o l u m e (mL)  Figure 3.5  Molecular weight distribution of HIC fraction 1 of SWA4 UF > 10 kDa from a column of Sephadex G-25. Sample of concentration of 272 mg/mL was loaded in one 20 mL fraction and elution was performed at 1.0 mL/min with each  £  fraction containing 3 mL for a total of 150 fractions.  0  50  100  150  200  250  300  350  400  450  Volume (mL)  Figure 3.6  Molecular weight distribution of HIC fraction 2 of SWA4 UF > 10 kDa from a column of Sephadex G-25. Sample of concentration of 409 mg/mL was loaded in one 20 mL fraction and elution was performed at 1.0 mL/min with each  4^ O  fraction containing 3 mL for a total of 150 fractions.  0.4  0.3  0.2  0.1  \k\iW \ V  50  100  150  200  250  300  350  400  450  Volume (mL)  Figure 3.7  Molecular weight distribution of HIC fraction 3 of SWA4 UF > 10 kDa from a column of Sephadex G-25. Sample of concentration of 271 mg/mL was loaded in one 20 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  ure 3.8 Molecular weight distribution of HIC fraction 7 of SWA4 UF > 10 kDa from a column of Sephadex G-25. Sample of  concentration of 533 mg/mL was loaded in one 1.5 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  280 nm so the compounds in these GF fractions were highly aromatic. However, no observable solids were recovered due to the low yield from the hydrophobic separation. Gel filtration chromatograms of the wash-off and the ethanol fractions from HIC of other UF fractions (Appendices E - J) were similar to those for the > 10 kDa UF fraction shown in Figures 3.5 - 3.8. The wash-off fractions generally showed two main areas, a less abundant fraction of molecular weight 1700 - 3200 Da and a bigger fraction of weight below 1700 Da that contained aromatic compounds. The ethanol fraction contained two major regions, one of molecular weight around 3500 - 4000 Da while another containing molecules with masses smaller than 2300 Da. These findings from gel filtration chromatography suggested that peptides from the SWA4 soluble hydrolysate samples were less than 4 kDa. Besides separation by size, gel filtration chromatography also functions as a de-salting column when the sample constituents have molecular weight larger than the V where salt would elute. However, t  most materials in the whole shrimp waste hydrolysate samples were small, especially compounds that were potentially bitter. Fractions observed at lower molecular weight range showed absorbance at 280nm, which was indicative of the presence of aromatic compounds that are usually associated with bitterness. Hence, ammonium sulfate salt from HIC was eluted concurrently with the GF fractions of small molecular weight. Some of these GF fractions were informally tasted to find out if there was bitterness in any fractions. However, none of the tasted fractions was intensely bitter or had bitterness comparable to the non-fractionated soluble hydrolysates. There could be several possibilities. During the process of fractionation, bitterness could be lost or filtered. This had been tested by filtering the soluble hydrolysate solution through the syringe filter and tasting the filtered and unfiltered solutions. No difference was noted between the two solutions. Another possibility is that the bitterness resided in fractions containing high salt which could not be tasted. If this was the case, elimination of the hydrophobic interaction chromatographic separation which required high salt in the buffers would be recommended. This would be feasible because from this study, it did not occur that the  143  hydrophobic column separated the soluble hydrolysates well using decreasing ammonium sulfate concentration in neutral buffer. Another alternative is that after several fractionation steps, the bitter peptides, which originally aggregated by hydrophobic interaction, were separated in different fractions. The synergistic effect, if any, of these peptides could have been reduced; thus, none of the isolated fractions contained intense bitterness. 3.3.5  Fractionation of Soluble Shrimp Waste Hydrolysate of SWA4 by Gel Filtration Chromatography and Amino Acid Compositions of Bitter Fractions Isolated from GF  Since it was difficult to taste the fractionated sample due to the intense saltiness, a nonfractionated soluble hydrolysate sample of SWA4 was separated directly by gel filtration chromatography, without first performing UF and HIC, and fractions were collected for informal tasting. The GF chromatograph of SWA4 is shown in Figure 3.9 and eight fractions were collected. Fractions 4-7, which had molecular mass less than 3 kDa were the dominant areas of the graph. All eight fractions were all freeze-dried and the solids were tasted informally. Informal sensory tasting suggested that some weak bitter taste was found in fractions 2 and 5 while stronger bitterness was found in fractions 4, 6, 7 and 8. However, the bitterness in any of these fractions was not comparable to the bitterness detected in the non-fractionated soluble hydrolysates of SWA4. GF fractions 4, 6, 7 and 8 that had bitter taste were sent for total amino acid analysis. The amino acid compositions of the bitter fractions collected from gel filtration chromatography are shown in Table 3.4. GF fraction 4 had 66 % of its component being amino acids while fractions 6-8 contained close to or less than 50 %. When the amino acid composition of the four fractions were compared, GF4 not only comprised a large amount of aspartic acid and glutamic acid, but also contained a lot of Ser, Ala, Thr and  144  0.2  40  ,  ,  50  100  1  1  150  200  < 250  '  '  '  '  300  350  400  450  1  500  Volume (mL)  * Measurement was made after 20 time dilution Figure 3.9  Molecular weight distribution of SWA4 soluble shrimp waste hydrolysate from a column of Sephadex G-25. Sample of concentration of 50 mg/mL was loaded in one 10 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  LTi  Table 3.4  Amino acid composition of some bitter fractions collected from Sephadex G-25 chromatography of SWA4 soluble hydrolysate . Amino Acid Content, g/100 g dry sample 3  Amino acid Asp Glu Ser Gly His Arg Thr Ala Pro Tyr Val Met He Leu Phe Lys Total (Total Free)  Major taste of free amino acid  GF4  GF6  GF7  GF8  5.35 (0.04) 9.87 (0.35) 3.16 (0.15) 2.91 (0.25) 1.52 (0.06) 5.06 (0.29) 3.29 (0.51) 4.49 (0.58) 2.12 (0.42) 1.76 (0.06) 4.11 (0.39) 2.32 (0.47) 3.96  1.43  0.84  Sour  1.83  0.71  Sour  1.01  0.52  3.94  3.29  Sweet  3.73  2.23  Bitter  3.60  4.56  Bitter  1.09  0.40  Sweet  1.22  0.76  Sweet  0.62  0.40  11.98  14.44  Bitter  1.30  0.55  Bitter  1.79  0.52  Bitter  1.23  0.57  6.38 (0.14) 2.71 (0.15) 6.64 (0.38)  2.68 (0.07) 3.47 (0.30) 2.40 (0-53) 3.98 (1.86) 3.02 (0.33) 6.35 (1.36) 1.63 (0.54) 2.97 (0.96) 1.45 (0.74) 7.54 (0.57) 1.82 (0.41) 2.09 (0.83) 1.84 (0.44) 3.20 (0.84) 6.44 (1.01) 2.35 (0.25)  1.81  0.66  8.23  5.52  1.92  1.34  65.63 (4.90)  53.23 (11.89)  46.74  37.31  (0.1.4)  b  Sweet  Sweet and Bitter 0  Bitter' Bitter Bitter Sweet and Bitter  Amino acid content after acid hydrolysis; free amino acid contents by analysis without hydrolysis are shown in parenthesis. From Fuke (1994). °From Kirimura and others (1969).  a  b  146  Pro which are sweet amino acids. Even though it also had many bitter amino acids such as arginine, leucine, valine, isoleucine and lysine, the bitterness in the fraction was still weak. The content of bitter as well as sweet amino acids could have been in the same peptide which would not elicit as strong bitterness as peptides containing mostly hydrophobic residues (Ishibashi and others 1988b). When the compositions of fractions 6 - 8 were examined, it was interesting to find a very high amount of tyrosine and phenylalanine, which are aromatic amino acids. These amino acid side chains have been associated with bitterness elicitation in peptides (Ishibashi and others 1987). Therefore, even though fractions 4, 6, 7 and 8 all tasted bitter, from the amino acid compositions, it could be observed that the bitterness could be contributed by different amino acid residues. Since fractions 6-8 eluted late in the gel filtration chromatography, it was predicted that some of the Tyr and Phe were free amino acids because Alcalase is known to cleave at these two amino acid sites (Doucet and others 2003). Although insufficient samples were available for free amino acid analysis of fractions 7 arid 8, the free amino acid contents of fractions 4 and 6 are shown in Table 3.4. Fraction 4 contained only 5 % free amino acids while fraction 6 had about 12 %. Most of the materials in fraction 4 were peptides, which was reasonable because fraction 4 had a higher molecular weight. Even though fraction 6 contained more free amino acids, the free amino acids did not comprise primarily of phenylalanine and tyrosine. It showed that in fraction 6, there was a high possibility that many small peptides consisted of Phe and Tyr residues. It has been suggested that if these amino acid side chains were located at the C-terminus of the peptide, a strong bitterness was detected (Ishibashi and others 1987). In addition, an increase in the number of these side chains in the peptides would greatly enhance the bitter taste (Ishibashi and others 1987).  147  3.3.6  Conclusions  To conclude, peptide fractionation is a very complicated process especially when evaluation of taste qualities is involved. Several factors have to be considered when the methodology is developed. Fractionation by size and hydrophobicity is common when isolation of bitterness is required because bitter peptides have been proposed to have smaller molecular weight and high hydrophobicity (Aubes-Dufau and others 1995; Kim and others 1999; Kim and others 2003; Cho and others 2004). In this study, soluble hydrolysates of shrimp processing discards were used to isolate bitter fractions because bitterness in the hydrolysates is usually an unfavourable attribute in the product. Through understanding the characteristics of the bitter substances in these hydrolysates, it may be possible to devise a scheme to produce non-bitter hydrolysates via the use of proteases with specific cleavage preference or other strategies such as solvent extraction or the addition of masking agents. The bitter fractions in the soluble whole shrimp waste hydrolysate (SWA4) in this study were found to be of molecular weight smaller than 3 kDa, which was similar to values reported by other researchers (Kim and others 1999; Kim and others 2003; Cho and others 2004; Aspmo and others 2005). Although separation by hydrophobicity was not successful using the phenyl (low sub) hydrophobic interaction column eluting with decreasing ammonium sulfate salt concentration under neutral condition, it remained a possibility that other elution conditions such as the use of a different pH (acidic or basic) or a different salt could lead to better binding Of hydrophobic compounds to the column. The amino acid analysis data provided evidence that hydrophobic and bulky amino acid side chains such as tyrosine and phenylalanine dominated in the fractions that contained bitterness. Further studies should involve fractionating the soluble protein hydrolysates of shrimp processing discards by two-dimensional high performance liquid chromatography to  148  increase sensitivity and resolution. Fractions containing bitterness could be analyzed using liquid chromatography tandem mass spectrometry to reveal the peptide sequence.  149  .4  References  Aspmo SI, Horn SJ, Eijsink VGH. 2005. Enzymatic hydrolysis of Atlantic cod (Gadus morhua L.) viscera. Process Biochemistry 40: 1957-1966. Aubes-Dufau I, Seris JL, Combes D. 1995. Production of peptic hemoglobin hydrolysates: bitterness demonstration and characterization. Journal of Agricultural and Food Chemistry 43: 1982-1988. Cho MJ, Unklesbay N, Hsieh FH, Clarke AD. 2004. Hydrophobicity of bitter peptides from soy protein hydrolysates. Journal of Agricultural and Food Chemistry 52: 5895-5901. Doucet D, Otter DE, Gauthier SF, Foegeding EA. 2003. Enzyme-induced gelation of extensively hydrolyzed whey proteins by alcalase: peptide identification and determination of enzyme specificity. Journal of Agricultural and Food Chemistry 51: 6300-6308. FitzGerald RJ, O'Cuinn G. 2006. Enzymatic debittering of food protein hydrolysates. Biotechnology Advances 24: 234-237. Fuke S. 1994. Taste-active components of seafoods with special reference to umami substances. In: Shahidi F, Botta JR, editors. Seafoods: chemistry, processing technology and quality. 1 ed. Glasgow: Blackie Academic and Professional, p 115-139. st  Gomez MJ, Garde S, Gaya P, Medina M, Nunez M. 1997. Relationship between level of hydrophobic peptides and bitterness in cheese made from pasteurized and raw milk. Journal of Dairy Research 64: 289-297. deHolanda HD, Netto FM. 2006. Recovery of components from shrimp (Xiphopenaeus Kroyeri) processing waste by enzymatic hydrolysis. Journal of Food Science 71(5): 298-303. Ishibashi N, Ono I, Kato K, Shigenaga T, Shinoda I, Okai H, Fukui S. 1988a. Role of the hydrophobic amino acid residue in the bitterness of peptides. Agricultural and Biological Chemistry 52(1): 91-94. Ishibashi N, Kouge K, Shinoda I, Kanehisa H, Okai H. 1988b. A mechanism for bitter taste sensibility in peptides. Agricultural and Biological Chemistry 52(3): 819827. Ishibashi N, Sadamori K, Yamamoto O, Kanehisa H, Kouge K, Kikuchi E, Okai H, Fukui S. 1987. Bitterness of phenylalanine- and tyrosine-containing peptides. Agricultural and Biological Chemistry 51(12): 3309-3313.  150  Izawa N, Tokuyasu K, Hayashi K. 1997. Debittering of protein hydrolysates using Aeromonas caviae aminopeptidase. Journal of Agricultural and Food Chemistry 45(3): 543-545. Kim MR, Choi SY, Kim CS, Kim CW, Utsumi S, Lee CH. 1999. Amino acid sequence analysis of bitter peptides from a soybean proglycinin subunit synthesized in Escherichia coli. Bioscience, Biotechnology, and Biochemistry 63(12): 2069-2074. Kim MR, Kawamura Y, Lee CH. 2003. Isolation and identification of bitter peptides of tryptic hydrolysate of soybean 1 IS glycinin by reverse-phase high performance liquid chromatography. Journal of Food Science 68(8): 24162422. Kirimura J, Shimizu A, Kimizuka A, Ninomiya T, Katsuya N. 1969. The contribution of peptides and amino acids to the taste of foodstuffs. Journal of Agricultural and Food Chemistry 17(4): 689-695. Kristinsson HG, Rasco BA. 2000. Fish protein hydrolysates: production, biochemical, and functional properties. Critical Reviews in Food Science and Nutrition 40(1): 43-81. Lalasidis G, Sjoberg LB. 1978. Two new methods of debittering protein hydrolysates and a fraction of hydrolysates with exceptionally high content of essential amino acids. Journal of Agricultural and Food Chemistry 26(3): 742-749. Lee KD, Warthesen JJ. 1996. Mobile phases in reverse-phase HPLC for the determination of bitter peptides in cheese. Journal of Food Science 61(2): 291294. Lekshmy NA, Prabhu PV. 1989. Studies on the chemical and nutritional quality of protein powders isolated from shrimp waste. Fishery Technology 26: 56-59. Maehashi K, Matsuzaki M, Yamamoto Y, Udaka S. 1999. Isolation of peptides from an enzymatic hydrolysate of food proteins and characterization of their taste properties. Bioscience, Biotechnology, and Biochemistry 63(3): 555-559. Mandeville S, Yaylayan V, Simpson B. 1992. GC/MS analysis of flavor-active compounds in cooked commercial shrimp waste. Journal of Agricultural and Food Chemistry 40: 1275-1279. Mandeville S, Yaylayan V, Simpson B, Pare JRJ. 1991. Gas chromatography-mass spectrometry analysis of flavour-active compounds from raw commercial shrimp waste. Spectroscopy 9: 61-72. Matsumoto M, Yamanaka H. 1990. Changes in contents of glycolytic metabolites and free amino acids in the muscle of kururria prawn during storage. Nippon Suisan Gakkaishi 56: 1515-1520. 151  Minagawa E, Kaminogawa S, Tsukasaki F, Yamauchi K. 1989. Debittering mechanism in bitter peptides of enzymatic hydrolysates from milk casein by aminopeptidase T. Journal of Food Science 54(5): 1225-1229. Nishitoba T, Sato H, Sakamura S. 1988. Bitterness and structure relationship of the triterpenoids from Ganoderma lucidum (Reishi). Agricultural and Biological Chemistry 52(7): 1791-1795. Pedersen B. 1994. Removing bitterness from protein hydrolysates. Food Technology 48: 96-98. Pommer K. 1995. New proteolytic enzymes for the production of savory ingredients. Cereal Foods World 40: 745-748. Saha BC, Hayashi K. 2001. Debittering of protein hydrolyzates. Biotechnology Advances 19: 355-370. Shahidi F. 1994. Seafood processing by-products. In: Shahidi F, Botta JR, editors. Seafoods: chemistry, processing technology and quality. 1 ed. Glasgow: Blackie Academic and Professional, p. 320-334. st  Shahidi F, Synowiecki J. 1991. Isolation and characterization of nutrients and valueadded products from snow crab (Chinoecetes opilio) and shrimp (Pandalus borealis) processing discards. Journal of Agricultural and Food Chemistry 39: 1527-1532. Synowiecki J, Al-Khateeb NAAQ. 2000. The recovery of protein hydrolysate during enzymatic isolation of chitin from shrimp Crangon crangon processing discards. Food Chemistry 68: 147-152. Yeom HW, Kim KS, Rhee JS. 1994. Soy protein hydrolysate debittering by lysineacetylation. Journal of Food Science 59(5): 1123-1126.  152  C H A P T E R IV  4.1  Conclusions  Study Findings arid Implications  In summary, the work performed in this study has shown that shrimp processing discards are excellent sources of proteins, with the potential of being used as protein hydrolysates for food application. The results have shown that the type of protease was the strongest factor affecting soluble product yield, degree of hydrolysis as well as bitterness. The amount of enzyme addition to the substrate had only affected the degree of hydrolysis but water-to-substrate ratio and duration of hydrolysis did not show any influence on the responses assessed. Overall, Alcalase samples gave high soluble product yield (> 30 %) and degree of hydrolysis (> 2.0 meq/g), yet bitterness in samples (> 2300 ppm caffeine) was intense. Protamex samples achieved similar soluble product yield as samples hydrolyzed by Alcalase but the DH of the samples was lower, ranging from 1.2 - 2.2 meq/g and the bitterness was shown not to be significantly different from Alcalase samples. The samples hydrolyzed by Alcalase and Protamex were shown to have higher soluble product yields and degree of hydrolysis than bromelain, Flavourzyme and control samples. They also had a DH higher than hydrolysates produced by bromelain and controls but similar to Flavourzyme samples. Hydrolysates produced by Flavourzyme, on the other hand, gave soluble product yields of around 20 %, which was still four times higher than the controls (5 %). They also reached significantly higher degree of hydrolysis, ranging from 1.4-3.1 meq/g, than the control samples, which only had a DH of approximately 0.4 meq/g. These samples hydrolyzed by Flavourzyme did not contain very strong bitterness (-1500 ppm caffeine). Although bromelain samples had similar soluble product yield and bitterness to those of Flavourzyme samples, the DH were significantly lower (-0.6 meq/g) and did not differ from the controls.  153  It was concluded that shrimp processing wastes hydrolyzed with Flavourzyme at 1:1 water-to-substrate ratio, 3 % enzyme and incubated for 8 hours (SWF8) produced the most favourable products out of the 16 hydrolysate samples produced using the Taguchi's design. It possessed high soluble yield (18.75 %), high DH (3.05 meq/g) but mild bitterness (1600 ppm caffeine). This level of bitterness could be tolerable depending on the application of the product. However, a response surface methodology could be used to further optimize the experimental conditions to produce shrimp waste hydrolysates. From this study, it was shown that an 8 hour incubation time coupled with 3 % (w/w protein) enzyme was the best combination. Nevertheless, it remained possible that shorter incubation time with the same enzyme concentration or other combinations of the experimental conditions could produce similar or even better product quality. Therefore, the results from this current study could act as the basis for further optimization of the process so shrimp waste hydrolysates could be produced in large-scale in industrial setting to minimize production cost and maximize quality and profits of the final products. According to the sensory evaluation data, all soluble hydrolysates produced by different proteases contained varying levels of bitterness with Alcalase and Protamex samples carrying the strongest bitter taste. However, protein hydrolysates produced without addition of exogenous enzymes (the controls) were shown to contain no or very low level of bitterness (800 - 900 ppm caffeine), which supported the hypothesis that bitter flavour was mainly caused by peptides and not lipid oxidation compounds. For lipid oxidation compounds to be primarily responsible for the bitterness found in the hydrolysate samples, each sample should contain similar level of bitterness because lipid content of the starting materials should not vary. The amino acid composition of several soluble hydrolysate samples showed that hydrolysates produced by proteases had higher contents of total amino acids than the controls and the starting shrimp waste blend, which did not differ in terms of amino acid contents. In addition, bromelain and Flavourzyme samples, which were evaluated as moderately bitter, contained around 20 % free amino acids. Since bitter amino acids in free form did not elicit as strong a bitter taste than in peptides, it was reasonable that  154  these samples were less bitter than the samples hydrolyzed by Alcalase and Protamex, which contained only around 6 %freeamino acids. Several amino acid residues such as Tyr, Val, Met, He, Leu and Phe showed slight variation in their contents, which appeared to associate with the bitterness in the samples. These amino acid side chains have been proposed to be responsible for bitterness in hydrolysate samples either in free form or as binding sites in peptide form (Ishibashi and others 1987; Ishibashi and others 1988a; Ishibashi and others 1988b; Kim and others 1999). Peptide fractionation study on the soluble whole shrimp waste hydrolysate of SWA4 suggested that most bitter peptides had molecular masses less than 3 kDa. Even though hydrophobic interaction chromatographic separation did not show binding, amino acid analysis on some isolated bitter fractions from gel filtration chromatography indicated that several amino acid side chains, specifically Tyr and Phe, dominated in the bitter fractions. Therefore, hydrophobicity has also been shown to relate to bitterness in protein hydrolysate products of shrimp processing discards. In conclusion, the results from this study served two purposes. First, it provided a rationale to produce protein hydrolysatesfromshrimp waste using Flavourzyme so low bitterness in the final products could be acceptable dependent on the food application. Secondly, this work suggested that small peptides of masses less than 3 kDa containing hydrophobic side chains were mainly responsible for bitterness in soluble shrimp waste hydrolysates. It also appeared that samples containing high contents of hydrophobic side chains in peptide forms were more bitter than those with more free amino acids. This information could serve as the foundation to identify the bitter peptides in the hydrolysates so the methodology could be tailor-designed for producing non-bitter shrimp waste hydrolysates.  155  4.2  4.2.1  Areas for Further Research and Application  Isolation and Identification of Bitter Peptides using TwoDimensional High Performance Liquid Chromatography and Mass Spectrometry  Since fractionation results from the current study did not provide samples for proper sensory evaluation of the bitterness in the final fractions due to the high salt content, it is necessary to develop other methods to isolate the bitter fractions for sensory tasting and to identify the sequence of the bitter peptides. Two-dimensional high performance liquid chromatography has been proposed to have high sensitivity, high reproducibility and high speed with further advantages of being automated and quantitative (Issaq and others 2002; Stasyk and Huber 2004). For the purposes of isolating bitter peptides, the best approach will be to first separate by size using size-exclusion chromatography and the peaks from the first column will be automatically injected to the second column to fractionate peptides of similar size by hydrophobicity using increasing concentration of ethanol in water by reversed-phase chromatography (Kim and others 1999; Kim and others 2003). The purified fractions can be evaluated on their sensory qualities and the fractions that taste bitter can be analyzed by mass spectrometry to reveal the peptide sequence. For the best resolution and separation, it is necessary for the sample to have low density (Issaq and others 2002). Therefore, a pre-fractionation step by size can reduce the amount of substances in the sample so that the separation performance by two-dimensional high performance liquid chromatography can be enhanced. 4.2.2  Sensory Evaluation of Shrimp Waste Hydrolysates on Shrimp Aroma and Shrimp Flavour  In order for the protein hydrolysates of shrimp processing discards to be considered for use in food flavouring, it is essential to ensure that the hydrolysate samples contained  156  highly favourable shrimp aroma and flavour. Several studies have shown that shrimp waste contains many flavour-active compounds (Mandeville and others 1991; Mandeville and others 1992) that can be used in food flavouring. However, there has not been any reported literature showing the sensory quality of the shrimp waste hydrolysates. Therefore, it is crucial to investigate if favourable aroma and flavour can be detected in the hydrolysates and whether these aroma and flavour-active components are comparable to the shrimp flavouring available in the market. 4.2.3  Study of the Effectiveness of Different Strategies to Mask the Bitterness and the Use of Enzymatic Approach to Debitter Shrimp Waste Hydrolysates  Several strategies, such as treatment with activated carbon, solvent extraction, the use of masking agents and the use of enzymatic approach, have been reported to effectively remove or lower bitterness in protein hydrolysates (Pedersen 1994; Saha and Hayashi 2001). However, their usefulness also depends on the application of the final products. Treatment with activated carbon has been shown to reduce the amount of several amino acids in the hydrolysates (Cogan and others 1981) while extraction with several solvents have resulted in significant loss of yields and essential amino acid contents (Lalasidis and Sjoberg 1978). However, there has not been reported nutritive loss On the use of masking agents or enzymatic approach. 4.2.3.1  The Use of Additives to Mask Bitterness in Shrimp Waste Hydrolysates  Several additives such as a-cyclodextrins arid gelatinized starch have been reported to effectively mask the bitterness (Tamura and others 1990). Cyclodextrins have been proposed to enclose the hydrophobic peptides so bitterness was reduced. Likewise, starch can be used to wrap the hydrophobic amino acid residues into the net structure upon heating to 100 °C so the bitterness could be diminished.  157  Several acidic amino acids have been reported to reduce bitterness (Tamura and others 1990). Two amino acids, namely aspartic acid and glutamic acid, and taurine, were studied regarding their effects on debittering hydrolysates as well as peptides. It was shown that bitterness in hydrolysate could be reduced by these compounds but peptide bitterness could not. However, the hydrolysate debittered using these amino acids would lead to strong sourness in the samples. Since many masking agents have been proposed to reduce or eliminate bitterness, it is necessary to study how these masking agents affect the bitterness in the shrimp waste hydrolysates. Selected masking agents could be added pre-hydrolysis and post-hydrolysis. Experiments could be designed such that the masking agents will be added at the beginning of the hydrolysis to investigate if bitterness will be decreased in the end product. On the other hand, masking agents can be added after hydrolysates are produced to see if the effect of reducing bitterness can be observed. 4.2.3.2  The Use of Enzymatic Approach to Produce Non-Bitter Shrimp Waste Hydrolysates  Besides the use of masking agents, enzymatic approaches have also been reported to reduce bitterness in hydrolysates. Several researchers have investigated the use of aminopeptidases to hydrolyze protein materials to produce non-bitter protein hydrolysates (Minagawa and others 1989; Izawa and others 1997). Some studies have reported the use of poly-enzymatic method (Aaslyng and others 1998; Deng and others 2004) or sequential protease addition (Clemente and others 1999) to hydrolyze proteins so the final products were not bitter. In the current study, Protamex, a protease complex comprising multiple endopeptidase activities that was proposed to produce non-bitter hydrolysate, produced shrimp waste hydrolysates of intense bitterness. Even though Flavourzyme, an enzyme known to contain both endo- and exopeptidase activities, produced relatively less bitter hydrolysates, significantly higher bitterness than the control samples was still observed. 158  As a result, it appears that combined enzymatic approach should be attempted. Since the current study demonstrated that Alcalase-hydrolyzed shrimp waste produced high product yield and high DH but also high bitterness, it would be worth designing experiment to investigate the use of Alcalase to initially cleave proteins into shorter peptides after which Flavourzyme will be added so the exopeptidase activities can assist in removing hydrophobic amino acid residues from the peptides to produce non-bitter hydrolysates. An incubation time of 4 hours with 4 % enzyme applied to a 2:1 water-to-substrate mixture should be chosen for the initial incubation with Alcalase because it has been shown in this work to produce a relatively high soluble yield product with high DH. After the first incubation, Alcalase in the hydrolysate mixture should be inactivated by heat and Flavourzyme should be added to further hydrolyze the mixture for another hour. The product of the hydrolysis should be subjected to sensory evaluation to examine if there is bitterness in the hydrolysate and whether or not the bitterness in the hydrolysate sample is significantly lower than the Alcalase and Flavourzyme samples.  159  3  References  Aaslyng MD, Martens M, Poll L, Nielsen PM, Flyge H, Larsen LM. 1998. Chemical and sensory characterization of hydrolyzed vegetable protein, a savory flavoring. Journal of Agricultural and Food Chemistry 46: 481-489 Cogan V, Moshe M, Mokady S. 1981. Debittering and nutritional upgrading of enzymic casein hydrolysates. Journal of the Science Food and Agriculture 32: 459-466. Clemente A, Vioque J, Sanchez-Vioque R, Pedroche J, Bautista J, Millan F. 1999. Protein quality of chickpea (Cicer arietinum L.) protein hydrolysates. Food Chemistry 67: 269-274. Deng SG, Peng ZY, Chen F, Yang P, Wu T. 2004. Amino acid composition and antianaemia action of hydrolyzed offal protein from Harengula Zunasi Bleeker. Food Chemistry 87: 97-102. Ishibashi N, Ono I, Kato K, Shigenaga T, Shinoda I, Okai H, Fukui S. 1988a. Role of the hydrophobic amino acid residue in the bitterness of peptides. Agricultural and Biological Chemistry 52(1): 91-94. Ishibashi N, Kouge K, Shinoda I, Kanehisa H, Okai H. 1988b. A mechanism for bitter taste sensibility in peptides. Agricultural and Biological Chemistry 52(3): 819827. Ishibashi N, Sadamori K, Yamamoto O, Kanehisa H, Kouge K, Kikuchi E, Okai H, Fukui S. 1987. Bitterness of phenylalanine- and tyrosine-containing peptides. Agricultural and Biological Chemistry 51(12): 3309-3313. Izawa N, Tokuyasu K, Hayashi K. 1997. Debittering of protein hydrolysates using Aeromonas caviae aminopeptidase. Journal of Agricultural and Food Chemistry 45(3): 543-545. Issaq HJ, Conrads TP, Janini GM, Veenstra TD. 2002. Methods for fractionation, separation and profiling of proteins and peptides. Electrophoresis 23: 30483061. Kim MR, Choi SY, Kim CS, Kim CW, Utsumi S, Lee CH. 1999. Amino acid sequence analysis of bitter peptides from a soybean proglycinin subunit synthesized in Escherichia coli. Bioscience, Biotechnology, and Biochemistry 63(12): 2069-2074.  160  Kim MR, Kawamura Y, Lee CH. 2003. Isolation and identification of bitter peptides of tryptic hydrolysate of soybean US glycinin by reverse-phase high performance liquid chromatography. Journal of Food Science 68(8): 24162422. Lalasidis G, Sjoberg LB. 1978. Two new methods of debittering protein hydrolysates and a fraction of hydrolysates with exceptionally high content of essential amino acids. Journal of Agricultural and Food Chemistry 26(3): 742-749. Mandeville S, Yaylayan V, Simpson B. 1992. GC/MS analysis of flavor-active compounds in cooked commercial shrimp waste. Journal of Agricultural and Food Chemistry 40: 1275-1279. Mandeville S, Yaylayan V, Simpson B, Pare JRJ. 1991. Gas chromatography-mass spectrometry analysis of flavour-active compounds from raw commercial shrimp waste. Spectroscopy 9: 61-72. Minagawa E, Kaminogawa S, Tsukasaki F, Yamauchi K. 1989. Debittering mechanism in bitter peptides of enzymatic hydrolysates from milk casein by aminopeptidase T. Journal of Food Science: 54(5): 1225-1229. Pedersen B. 1994. Removing bitterness from protein hydrolysates. Food Technology 48: 96-98. Saha BC, Hayashi K. 2001. Debittering of protein hydrolyzates. Biotechnology Advances 19: 355-370. Stasyk T, Huber LA. 2004. Zooming in: fractionation strategies in proteomics. Proteomics4: 3704-3716. Tamura M, Mori N, Miyoshi T, Koyama S, Kohri H, Okai H. 1990. Practical debittering using model peptides and related compounds. Agricultural and Biological Chemistry 54(1): 41-51.  161  Appendix A  A sample of the sensory evaluation form used in the formal sensory evaluation of shrimp waste hydrolysates.  SENSORY EVALUATION F O R M For the given samples, please evaluate according to Bitter Taste. Simply indicate your evaluation by drawing a vertical line labelled with the sample number on the scale provided. Instruction: For each sample, with your nose plug, inject the sample solution in one aliquot onto the back of your tongue. Let it stay in your mouth for at least 5 seconds before swallowing or discarding. In between samples, please rinse once with lemon water and twice with water.  Caffeine Equivalent Scale R 0  •  750  — — •  1500  • —•  ••  —+-  3000  —  - •—  •  •  Other comments: Is there any off-flavor in the samples? Could you describe? YES  NO  162  Appendix B  Hydrophobic interaction chromatogram of soluble shrimp waste hydrolysate of SWA4 UF 3 - 10 kDa on Hi-Trap  1M  Phenyl (low sub) column. One gram of soluble hydrolysate was loaded in a volume of 20 mL of 1.0M (NHO2SO4 in sodium phosphate buffer at pH 7 eluting stepwise with decreasing concentration of (NH4)2S04 from 1.0, 0.5, 0.25, 0.10 to 0 M followed by 70 % ethanol. The elution was performed at 4 mL/min and fractions were collected every 8 mL.  V o l u m e (mL)  ppendix C  TM  Hydrophobic interaction chromatogram of soluble shrimp waste hydrolysate of SWA4 UF 1 - 3 kDa on Hi-Trap  1  Phenyl (low sub) column. One gram of soluble hydrolysate was loaded in a volume of 20 mL of 1.0M (NH4)2S04 in sodium phosphate buffer at pH 7 eluting stepwise with decreasing concentration of (NH4)2S04 from 1.0, 0.5, 0.25, 0.10 to 0 M followed by 70 % ethanol. The elution was performed at 4 mL/min and fractions were collected every 8 mL.  1.5  0.5  48  88  128  168  208  248  288  328 368  Volume  (mL)  408  448  488  528  568 608  Appendix D  Hydrophobic interaction chromatogram of soluble shrimp waste hydrolysate of SWA4 UF < 1 kDa on Hi-Trap  1M  Phenyl (low sub) column. One gram of soluble hydrolysate was loaded in a volume of 20 mL of 1 .OM (NH4)2S04 in sodium phosphate buffer at pH 7 eluting stepwise with decreasing concentration of (NH ) S04 from 1.0, 0.5, 0.25, 4 2  0.10 to 0 M followed by 70 % ethanol. The elution was performed at 4 mL/min and fractions were collected every 8 mL.  V o l u m e (mL) Oy  Appendix E  Molecular weight distribution of HIC fraction 1 of SWA4 UF 3 - 10 kDa from a column of Sephadex G-25. Sample of concentration of 250 mg/mL was loaded in one 10 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  Appendix F  Molecular weight distribution of HIC fraction 6 of SWA4 UF 3 - 10 kDa from a column of Sephadex G-25. Sample of concentration of 9 mg/mL was loaded in one 10 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  Appendix G  Molecular weight distribution of HIC fraction 1 of SWA4 UF 1 - 3 kDa from a column of Sephadex G-25. Sample concentration of 249 mg/mL was loaded in one 13 mL fraction and elution was performed at 1.0 mL/min with each  -r  Da  4  3.5  Da  fraction containing 3 mL for a total of 150 fractions.  «s a •  1  J  I  i «  £  •  Id  *  -  -  214nm  -  214nm lOx dil  • •  3  11  •  5  •  2.5 -  # ^ \ L I f \J\ 1  •  I  •  • «  111  s i  1.5  »  I  1 f  /  1  I /  |  1  I I I  t  * /  0.5  0  1  50  100  150  *  11 11  / 1  f  1  1  *  1 11  I  a  *  200  *  1  \  V.  I  1 1 1 1  % V  '  250  Volume (mL)  300  350  400  450  Appendix H  Molecular weight distribution of HIC fraction 5 of SWA4 UF 1 - 3 kDa from a column of Sephadex G-25. Sample of concentration of 10 mg/mL was loaded in one 10 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  Volume (mL)  Appendix I Molecular weight distribution of HIC fraction 1 of SWA4 UF < 1 kDa from a column of Sephadex G-25. Sample of concentration of 297 mg/mL was loaded in one 13 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  Appendix J Molecular weight distribution of HIC fraction 5 of SWA4 UF < 1 kDa from a column of Sephadex G-25. Sample concentration of 8 mg/mL was loaded in one 10 mL fraction and elution was performed at 1.0 mL/min with each fraction containing 3 mL for a total of 150 fractions.  Volume (mL)  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100797/manifest

Comment

Related Items