LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MTE DUE MTE DUE DATE DUE 1m clam.m14 lTl ULTRAFILTRATION PROCESSING OF DRY BEAN (Phaseolus vulgaris L.) PROTEIN FRACTIONS FORMULATED INTO A BEVERAGE SUITABLE FOR SMALL CHILDREN By Albert David Bolles A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Food Science and Human Nutrition 1997 ULTRAFILTR PROTEIN FRAC New beans uhrafritration separ isolated protein fro These separations Membranes of vari and Spiral wound derelopment of a I Beans wer adjusted to pH S refrigerated temp Protein enriched E p01S’Sllifone m em wound Utrafiltr Weight Cutoff of during MP each membrane had a l ABSTRACT ULTRAFHJRATION PROCESSING OF DRY BEAN (Phaseolus vulgaris L) PROTEIN FRACTIONS FORMULATED INTO A BEVERAGE SUITABLE FOR SMALL CHILDREN By Albert David Bolles Navy beans (Phaseolus vulgan's L.) were processed using microfiltration and ultrafiltration separation technologies to produce protein enhanced fractions. Fractions of isolated protein from the retentate and permeate streams were isolated during processing. These separations were based on molecular weight and specific membrane type. Membranes of various configurations were evaluated, specifically plate and frame, tubular and spiral wound. These protein fractions were utilized as a fortification nutrient in development of a protein beverage suitable for consumption by small children. Beans were milled (13.9% moisture) to a flour, mixed with tap water at 4°C and adjusted to pH 9 (3N NaOH). The slurry was mixed for one hour and held at a refiigerated temperature for 24 hours and the supernatant was decanted to recover a protein enriched extract, which was then adjusted to pH 7 (3N HCl). Two different MW polysulfone membranes were evaluated during microfiltration (MF), tubular and spiral wound. Ultrafiltration (UF) exclusively utilized spiral wound membranes with a molecular weight cut-off of 50,000 kd (kilodaltons). The tubular and spiral wound membranes used during MF each had a molecular weight cut-ofl‘ of 200,000 kd but the spiral wound membrane had a pore size of 0.3 microns to improve selectivity. A comparison of protein recovery rates between microfiltration and ultrafiltration was conducted to establish a mass balance for the processes. Greater protein (40%) partitioned intr tubular membl yielded 34.89 0 The LT memb protein. Anal membranes bas the protein Protein technology are were added to 2 ml Tetra Slimi tested versus 1h. be aPPTOPriate MW (amino ; MOSCOW indi aseptic processin These fin and LT teChriok aPpropriate for hr assess their Effie. l popUlEiilOrls partitioned into the retentate utilizing microfiltration, 200,000 kd molecular weight cut-ofl‘ tubular membrane than using spiral wound membrane, 50,000 kd. The retentate fiaction yielded 34.8% (db) protein with associated reductions of permeate protein, 26.6% (db). The UF membrane yielded 26.6% protein versus the MF membrane which yielded 9.01% protein. Analysis of the NDND-PAGE gels indicated separation differences between membranes based on molecular weight and pore size without appreciable denaturation of the protein. Protein fi'actions suitable for food applications using membrane separation technology are technologically feasible. Isolated protein fractions for these separations were added to a formulated peach shake and aseptically processed and packaged in a 200 ml Tetra Slim® package. Composition of products formulated with legume protein were tested versus the dairy-based protein control. Protein quality was measured and found to be appropriate at the 15% of total solids (WHO guidelines) and of expected protein quality (amino acid score). The nitrogen soluble index, NDND-PAGE, and Confocal Microscopy indicated limited protein denaturation occurred in separation of protein and aseptic processing of the final product. These findings indicated that legume protein can be selectively isolated using MF and UF technology, formulated, aseptically processed and packaged into a beverage appropriate for human consumption. Protein fractions should be further characterized to assess their efi‘ectiveness as an ingredient resource for applications in other food formulations and products developed should be tested for consumer acceptability in target populations. iv To the Memory of My Father Albert W. Belles Since my youth he taught me to attain the education he never had the chance to get. I dedicate this dissertation to him. I loved him as much as any son could love a father. Thank you very much Dad. I would 1 Bennink, Dr. Wil and continued SL encouragement professional dex nontraditional 3 educational exce utmost respect A Special Patience, guidan mentor and ffien Apprecia Um‘mitl'- Spet Jackson and Jo} S“604355. ACKNOWLEDGMENTS I would like to extend my appreciation to my committee members Dr. Maurice Bennink, Dr. William Haines, Dr. George Hosfield and Dr. Mary Zabik for their guidance and continued support through my graduate program. I also appreciate my committee’s encouragement and support in my continued education and sincere interest in my professional development. This group of individuals were able to deal with a nontraditional approach to a doctorate degree and demonstrated the dedication to educational excellence, to this, I owe them my sincere thanks and they have earned my utmost respect. A special appreciation is given to my major professor, Dr. Mark Uebersax, whose patience, guidance and insight were valuable during my education. He will always be my mentor and friend. I owe him great deal for my success and thank him very much. Appreciation is also extended to Food Science graduate students at Michigan State University. Specifically, Dr. Lillian Oceana, Dr. Ahmad Shirazi, Dr. Yongsoo Chung, Josi Jackson and John Rodgers whose participation in my project was vital to my program success. I would like to also acknowledge Gerber Products Company for their funding and assistance in my doctoral program. Special thanks to Tropicana Products Inc. for allowing me the time to finish my degree and to Helen Hombeck and Christina Keys for their administrative assistance. I also would like to thank the Tetra Laval organization for providing me with the equipment and facility to produce test samples. Dr. Sandy McCurdy at the POS group in Saskatoon, Saskawaton I would like to thank for conducting {6" thank Dr. R3)" my attempts to Lastly, encouragement never lost faith conducting feasibility studies in the early stages of our program. I would also like to thank Dr. Ray St. John for his untiring attempt to improve the grammar in the text despite my attempts to violate the rules of the English language. His friendship I cherish. Lastly, I would like to thank my wife and best fiiend, Dawn. Her support, encouragement and love enabled me to see this endeavor through its completion. She never lost faith in my abilities and because of that I love her. LIST OF 1 LIST OF P INI'RODL'I LITERATI Physico- Struc S. C . Comp Nc Pr: Utrafi, Ber Fitnmn-a Techno Hist Clas Separati Micr Utra TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION LITERATURE REVIEW Physico-Chemical Composition of Dry Beans Structural Characteristics Seed Coat Cotyledon Compositional Characteristics Non-Protein Constituents Protein Constituents Ultrafiltration Application in Legumes Benefits and Tradeoffs of Ultrafiltration Technology Ultrafiltration and Membrane Separation Technology Technology Overview History Classification of Separation Concept and Theory Separation Concepts Microfiltration (MF) Ultrafiltration (UF) Reverse Osmosis (RO) Nanofiltration (NF) Processing Theory Membrane Module and System Configurations Aseptic Processing and Packaging Technology Overview Product Development Strategies Nutritionally Enhanced Products for Children Weaning Foods xii \lflUt-b-b-h-h 10 15 18 18 18 19 19 19 20 22 23 23 26 32 37 37 Qua Nutritir STUDY Ol' Protl l’lficr Introdur Material Extra M Quant El SI M; Quality Driven Product Development Traditional Process: Phase Review Nutritional Labeling and Education Act (NLEA ) STUDY ONE Protein Separation for Navy Beans Utilizing Ultrafiltration and Microfiltration Technology Introduction Materials and Methods Extraction and Separation Microfiltration Quantitative and Qualitative Analyses Electrophoreses SDS Gel Electrophoreses Micro Nutrients Water Soluble Nitrogen Light Microscopy Results and Discussion Membrane Processing Gel Electrophoresis Amino Acids Thermal Process Simulation Conclusions STUDY TWO Protein Separation and Formulation of a Fortified Beverage Suitable for Small Children Introduction Materials and Methods Process and Product Development Ultrafiltration Product Formulation, Processing and Packaging Qualitative and Quantitative Analyses Proxirnate Composition and Product Characteristics 39 41 45 47 47 48 55 56 58 60 6O 62 62 68 79 84 89 91 91 92 94 96 N Raul C and . SUMMAI RECODI): APPEN DI ( l APPENI i Pr Pr APPENDI APPE ND 1; Appendi Nebr Appendix L) as Appendix Feb I' Appendix 1 1992 Appendix 24, 19 REFERENCES ix Nutrients VItarnins Protein Characterization SDS Gel Electrophoresis NLEA Compositional Label Results and Discussion Conclusions SUMMARY AND CONCLUSION RECOMMENDATIONS FOR FURTHER RESEARCH APPENDICES APPENDD( A - Preliminary Extract and UF Processing Trial Project Summary Project Report APPENDDI B - Compositional Analysis of Preliminary UF Processed Fractions APPENDIX C - Abstracts and Results of Professional Oral Presentations Made During the Course of Study Appendix C- 1 A- Bean Improvement Cooperative (BIC), University of Nebraska, NE Nov. 4-6, 1991 Appendix C- 1 B- Evaluation of Common Dry Beans (Phasest vulgaris L.) as a Protein Source in Weaning Foods Appendix C-2 - Fourth ASEAN Food Conference, Jakarta, Indonesia, Feb. 19, 1992 Appendix C-3 - Institute of Food Technologists, New Orleans, LA, June 1992 Appendix C-4 - Bean Improvement Cooperative (BIC), Boise, ID, Nov. 2-4, 1993 REFERENCES 98 99 l 04 107 108 124 125 128 130 131 136 148 158 159 165 172 175 180 199 Prc Cor L'F Prot MF 1 Prote Micrr L’tiliz. Micro Utilizii Whole with 1. Amino Diafiltre SUmrnar years 01( Health C Calculari MF. 2.1 2.2 2.3 2.4 2-5 2.6 2.7 2.8 2-9 3.1 LIST OF TABLES Synopsis of Developments in Aseptic Technology for Food Preservation. Contrast of QFD and Phase Review Processes. UF Equipment and Tubular Specifications Used for Navy Bean Protein Fraction in Trial 1. MP Equipment and Tubular Specifications Used for Navy Bean Protein Fractionation. Microfiltration Flux Rates of the Bean Protein Extract Processed Utilizing a Spiral Wound Membrane. Microfiltration Flux Rates of the Bean Protein Extract Processed Utilizing a Tubular Membrane. Whole Bean and Extract Protein Values for Bean Flour Extracted with 1.7% NaCl at pH 9.0. Amino Acid Profiles for UF Processing and Subsequent MF Diafiltration Utilizing Tubular and Spiral Wound Membranes. Summary of Estimates of Amino Acid Requirements for Children (12 years old), as proposed in 1985 by Food and Agriculture, World Health Organization, United Nations, Protein Content and a Score Calculation for Whole Bean, the Retentate and Permeate for UF & MF. Amino Acid Score Calculations in Ascending Orderfor the Amino Acids Required by 2 year-old Children. Scores for Whole Bean: 200,000 MW UF and Diafiltration Through a 50,000 MW, MF Retentate. Percent Water Soluble Protein for Retentate Which Was Processed Utilizing Ultrafiltration (200,000 MW) Tubular Membrane and then Diafiltered with Microfiltration (50,000 MW) Spiral Wound Membrane. UF Equipment and Tubular Specifications Used for Navy Hm- Protein Fraction in Trial 1. la?” 32 45 48 53 62 64 68 80 82 83 88 92 33 38 CC Pe .Ari Se; Pet Su1 by . He Per Arr Ret Pro Pea Pro hie; Sha Frat Nut for 3.3 3.4 3.5 3.6 3.7 3.8 Composition of UF Separated Bean Protein Fractions: 200,000 MW Permeate and Retentate. Amino Acid Composition (mg/100gzmg/g of Protein) of UP Separated Bean Protein Fractions: 200,000 MW Retentate and Permeate. Summary of the Estimates of Amino Acid Requirements for Children (2 years old), as Proposed in 1985 by Food and Agriculture, World Health Organization, United Nations and Amino Acid Scores for Permeate and Retentate. Amino Acid Score Calculation, in Ascending Order, of Ten Essential Amino Acids for 2-year-old Children. Scores for both Permeate and Retentate Using UF, 200,000 MW Membrane. Proximate Analyses and Wet Chemistry Analytical Data of Standard Peach Shake and Peach Shake Fortified with UF Separated Bean Protein Fraction. Mean Values for Mineral and Vitamin Content of Standard Peach Shake and Peach Shake Fortified with UF Separated Bean Protein Fraction. Nutrient Profile of Peach Shake with Lowfat Yogurt. % U.S.RDA for Ages 12 to 48 Months - Proposed Nutritional Label. l5 O? (D 111 113 114 115 116 117 118 Beers 1.1 1.2 1.3 1.7 5c 1-8 Sc. 110 SChi 1.11 1.12 1.13 Hous Anni; Dry n: (200‘0 (50,00 eSIgn 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11 1.12 1.13 2.1 xii LIST OF FIGURES Typical Membrane Filtration System Illustrating Principle of Size Separation. Comparison of Membrane Pore Size and Molecular Weight Cut- OIT of Various Separation Technologies. Schematic of Plate and Frame Arrangement. Schematic of a Spiral-Wound (SW) Arrangement. Schematic of a Tubular (T) Module Arrangement. Schematic of a U- Shaped Fiber Membrane Module. Schematic of a Straight-Through (ST) Capillary Membrane Module. Schematic Diagram of a Typical Tetra Vertical-Form-Fill-Seal Filler Illustrated by Components of the System. Typical Process and Package Configuration Utilizing Teta Pak® Technologies. Schematic Diagram of the Aseptic Roll Stock Formed into a Package Illustrating Functionality by Barrier. Product Development: Phase Review Process. Formal Process of QFD. House of Quality Blueprint of Externally Defined Design Attributes. Dry Navy Bean Protein Fractionation Using both a Microfiltration (200,000 MW Membrane) Tubular System and Ultrafiltration (50,000 MW Membrane) Spiral Wound Membrane. Sample Code Designators are Indicated. 19 21 27 28 29 3O 31 34 35 36 42 43 49 has: Do 2 Al- 50. are M mHC DnfiPGC ECCEd 7. 2 E 2am on 2 SFE PPS 2.10 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 Dry Navy Bean Protein Fractionation Using Microfiltration (.03 micron) 200,000 MW Spiral Wound and Ultrafiltration Through a 50,000 MW Spiral Wound Membrane. Sample Code Designations are Indicated. Microfiltration Batch System Utilizing Three 200,000 MW Tubular Membrane in a Filtration Series. Batch System Utilizing a Microfiltration 200,000 MW (0.3 micron) Membrane and Ultrafiltration 50,000 MW Spiral Wound Membrane System. Method to Simulate Thermal Process Conditions. Protein Source are Sealed in Metallized Retort Pouches, Placed in an Oil Bath and Heated to 99°C (210°F) for 5 to 30 Seconds. Thermal Data are Collected On-Line. Principle of Confocal Scanning - Path Followed by Light Coming from the Focal Plane (White Band) and from Outside the Focal Plane (Dotted Line) in the Zeiss LSM. The Focal Point of the Out- of-Focus Light F ails in Front of the Pin Hole; Thus Most of the Out-of-Focus Information Does Not Reach the Detector. Dry Navy Bean Protein Fractionated Using both a Microfiltration (200,000MW Membrane) Tubular System and Microfiltration (50,000MW Membrane) Spiral Wound Membrane System. Mass Balances for Total Solids and Protein are Expressed as Percents (% db) of Original. Dry Navy Bean Protein Fractionation Using Microfiltration 200,000MW (0.3 micron) and 50,000MW Spiral Wound Membranes. Schematic for Microfiltration and Ultrafiltrations Processes with Fractionation Systems Designated. Fraction Codes and Electrophoretic Lane Codes are Designated. Non-denaturing and Non-disassociating PAGE Analysis of Bean Protein Fractionated Using a Tubular Microfiltration Membrane System. lit” 50 51 52 54 61 66 67 69 73 to ;_. [J to ‘._. b) 33 35 SIX Tut hiei SIDE Tub 3161 1 PROCESS Particle Function —> Nanofiltration (— Microfiltration —> Molecular Weight Cut-ofl' over (Kilo Daltons) 300 20,000 500,00 1,000,000 5R 0 (kd) FIGURE 1.2: Comparison of Membrane Pore Size and Molecular Weight Cut-Off of Various Separation Technologies The term molecular weight cut-ofl refers to the molecular weight of the species retained by the membrane. The early UF membranes were made of cellulose acetate, which poses many operational limitations. New membranes are increasingly resistant to organic solvents, to broad and fluctuating ranges of pH and temperature, and to oxidizing cleaning agents such as chlorine. Ultrafiltration membranes use polymers such as polycarbonate, polyvinyl chloride, polyarrrides, polysulfone, polvinylidene fluoride, copolymers of acrylonitrile and vinyl chloride to increase versatility and durability under a broad spectrum of processing conditions. 22 Reverse Osmosis (RO): Reverse osmosis is probably the most important class of the membrane filtration array. At least half of the breakthroughs in membrane technology in the past 30 years have occurred with R0. An RO membrane should be considered as a semi-permeable barrier rather than as a fine filter. The osmotic pressure of the solution to be treated has a very significant impact on the design, operation, and econonrics of RO systems. During the 1960’s engineers/researchers at Dow Chemical Company in Midland, MI, began to spin fine cellulose acetate fibers suitable for membrane applications (Mahon, 1963; Mahon et al., 1969; Mahon, 1966 a and b). Their approach was that in spite of a low water flux, the fibers might be usefirl for water desalination because of the fibers’ low cost and high packing density per cubic foot of space. Thus, by reducing the size of the fibers, they could make the wall extremely thin without increasing the potential for the collapse of the fiber. Today, with a decrease in the availability of water and a decline in the quality of water, increased numbers of R0 systems are being installed to produce drinking water. Continuous improvement of RO systems yielded the spiral wound design (Westemorland, 1968; Bray, 1968). Prior to this, most RO systems used relatively expensive modules of the plate-and—frame or tube-in-shell (tubular) types. The spiral- wound configuration provided significant savings in the cost and space required for R0 systems, thus making RO systems more commercially feasible. Dow Chenrical developed hollow fibers of cellulose triacetate and a means to modularize these fibers (Mahon, 1966; McLain, 1969; McLain et al., 1969). This approach was technically successful but was not at that time commercially competitive. 23 Richter and Hoehn (1971) at DuPont in Wilmington, DE, developed a competitive design with the spiral-wound Loeb-Sourirajan cellulose acetate membrane. DuPont chose to spin hollow fibers using an aromatic polyamide polymer. DuPont gave the name permasep(R) to this RO system, which has found wide use in water desalination. A new class of membranes and fabrication techniques emerged to challenge existing materials and methods during the 1970’s. These were the so-called “thin film composite” membranes (Rozelle et al., 1970). The process for making the membrane was to deposit a very thin layer of membrane material onto the surface of a finely porous substrate (Cadotte et al., 1970; Cadotte et al., 1971). Polysulfone membranes, which have received wide market acceptance, were developed at DuPont by Riley and his co-workers (Riley et al., 1972). N anofiltration(N F): Nanofiltration, which was developed in the 1980’s, describes a membrane that possesses performance characteristics between R0 and UF. The membrane has a high rejection of divalent ions and low rejection of monovalent ions. The membrane, often described as a “loose” RO, will reject most organics with a molecular weight above approximately 300. Salt (NaCl) rejection is in the range of 0 to 20%. Nanofiltration is characterized as having a high rejection for most chemical species while operating at pressures common to ultrafiltration. PROCESSING THEORY The fiindamental theory of membrane separation requires a feed stream to enter the membrane system and a driving force of pressure, concentration or vacuum to be applied across the membrane (Figure 1.1). Under the influence of this positive force, the membrane selectively allows components to pass it. This selectivity is the basis for the separation into two streams termed retentate (membrane-retained stream) and “permeate” (membrane-permeable stream). Thus, the fluid that passes through the membrane is termed the permeate and the fluid that is retained by the membrane is designated the retentate or concentrate. The flux, which is the rate that the permeate passes through the membrane, is expressed in units of volume of permeate per unit of time per unit of membrane area. Examples of commonly used expressions of flux rate include gal/ft sq/day (gfd) or L/m sq or (lmh). Flux rate may be mathematically expressed as firnction of membrane, process and product variables as follows (Metzner and Reed, 1955; Goldsmith and Mason, 1963; Harper and Kamel, 1966; Skelland, 1967): Flux (J) = A*(Pf-Pp)/fluid viscosity (u) o A = Membrane permeability coefficient a P = Transmembrane pressure The rejection measures the membrane’s ability to separate or retain solution components. It is the fraction of the components retained by the membrane and is usually expressed as a percentage. Rejection = 100* ((Fi-Pi)/Fi) o I = specific component or group of components 0 F= concentration of I in the feed stream 0 P = concentration OH in the permeate stream Recovery is the fraction of the feed that is recovered as permeate. ex; ci‘ ‘h‘l 25 Permeability is the rate at which components permeate the membrane. It is expressed in units of quantity: example, quantity of components times thickness per unit of time per unit of membrane area per unit of driving force. Fluid dynamics is measured by the Reynold ’3 Number (Metzner and Reed, 1955), which measures the forces acting on the fluid. Re = pVD/u where o p=density o V=Velocity o D=efl‘ective diameter 0 u=viscosity Relatively Flow Characteristics: 0 Re < 2300, laminar flow 0 23004000, turbulent flow In food systems, turbulent flow is preferred to provide rapid heat transfer since when molecular action is random greater heat transfer can occur at the geometric center. Juices and beverages typically exhibit turbulent flow whereas viscose products such as mayonnaise exhibit larrrinar flow. Turbulent flow will usually increase the flux (Baker et al., 1990; Cheryan, M. 1986). Osmotic pressure, which is related to temperature and the molar concentration of the solution, can be calculated using Van’t Hofl’s equation (Skelland, 1967). Osmotic Pressure = CRT . Where: C=molar concentation of the solute . R=universal gas constant c T=absolute temperature (degrees Kelvin) Fouling, which occurs during cross-flow membrane processing, is manifested as a decline in flux with time of operation. Foulants can be proteins, carbohydrates, oils, microorganisms and inorganic material (McBain et al., 1931). Membranes that are fouled need to be cleaned to assure eficient separation (flux) capacity (Mohr et al., 1988). Another type of fouling is caused by changes in the membrane structure. This is termed “compaction”. Compaction is a permanent physical change in a polymeric membrane that decreases its permeability and flux. Applied pressure contributes to compaction, pressure combined with high temperature can increase it. Structural and conformational changes in membranes are irreversible, and cleaning will not restore membrane permeability (Cheryan, M. 1986). These changes decrease membrane performance and limit the commercial utility (operational life) of the membrane system. Generally, increasing pressure and temperature will increase the flux whereas increasing feed concentration will decrease flux. NEMBRANE MODULE AND SYSTEM CONFIGURATIONS Membrane modules and systems are available in many configurations. These are spiral wound, tubular, plate and frame, and hollow fiber with each type designed to meet specific product requirements (Merlo et al., 1993). The plate and frame (PF) system represents the earliest of all designs. As shown in Figure 1.3, it consists of flat sheets of membranes “sandwiched” together. Below are the 27 key advantages and disadvantages from the NFPA Membrane Filtration 1‘1ka (August 1993) for plate and frame, spiral wound, tubular, and hollow fiber membranes. Permeate Feed Retentate Spacers FIGURE 1.3: Schematic of Plate and Frame Arrangement. The spacer plates serve as flow channels and are held in place by a central bolt. These are typically stainless steel in design. Advantage of PF Membranes: - System can accommodate low levels of suspended solids and viscous fluids due to its thin feed flow channels which provide high turbulent flows. Disadvantages of PF Membranes: 0 High space requirements due to low density of membrane packing. - Capital costs are high for new system. 0 Ongoing operating costs are high due to increased labor requirements for stacking. - Membranes are susceptible to fouling. The spiral wound (SW) module is an extension of the plate and flame (Merlo et al. , 1993). The membranes are sandwiched around a porous, woven permeate carrier or spacer. Three sides of this envelope are sealed, but the fourth is open and attached to a perforated tube (Figure 1.4). Module housing PM "W -—-> ——> mortar flow Collection Pipe Permeate flow ”ed “0" ’ Residual flow Feed flow Membrane Permeate flow after passing through memebrane FIGURE 1.4: Schematic of a Spiral-Wound (SW) Arrangement. The feed flows axially into the channels formed by the wrapping of the feed channel spacer. The permeate penetrates the membrane and travels up the carrier spirally to the permeate collection tube. This multileaf design is used to minimize pressure drop. These vessels are usually constructed of fiberglass. Advantages of SW Membranes: - Low space required due to high density of membrane packing. 0 Lower capital costs. Disadvantages of SW Membranes: o Fouling is more frequent when high solids material is utilized. 0 Membrane cleaning requires greater labor and operating costs. 0 Due to fouling, temperature and pressures rise, which tend to deform plastic membranes. Ar 29 Tubular (T) modules are very common (Merlo et al., 1993). These modules are able to handle products and fluids of high viscosity. The tube diameters are large, ranging from one-eighth of an inch up to one inch. The feed fluid flows through one end of the tube while the permeate passes through the membrane (Figure 1.5). Permeate Hollow, Feed thin walled —* O i 1 r i r 9" “mm“ porous tubes \ l l l l l Single Tube Retentate FIGURE 1.5: Schematic of a Tubular (T) Module Arrangement. The permeate is collected in an outer shell that encloses the tube while the retentate exits at the opposite end. Larger tubes require higher rates of pumping to reduce the risk of membrane fouling. Most of the tubular modules are either stainless steel or ceranric. Advantages: of T Membranes: 0 Less fouling with high suspended solids material. . Easy to clean. . Tubes require low maintenance (if one fails, the tube can be easily by-passed or replaced without much down time). Disadvantages of T Membranes: . High space requirements due to low density membrane packing. . Higher cost of capital for initial purchase. 0 Due to tubular design, utility costs are higher, increasing operating expenses. Hollow fiber (HF) membranes consist of hollow, hair-like fibers with an outside diameter of 50 um to about 1 mm (Merlo et al., 1993). These fibers are bundled together into either a U-shaped configuration or a straight-through configuration (“tubular like”). A U-shaped configuration is a loop of fibers inside a pressure vessel (Figure 1.6). Retentate Stream Feed stream Permeate stream FIGURE 1.6: Schematic of a U-Shaped Fiber Membrane Module. As the feed enters the shell, pressure is applied, and the permeate passes through to the center of the hollow fibers. The permeate exits the open fiber end whereas the retentate exits the module opposite the feed. 31 Straight-through (ST) modules (Merlo et al., 1993) have fibers placed inside a pressure vessel and are open at both ends, similar to a tubular system (Figure 1.7). Hollow fiber membranes Module shell / CU: Feed—> —" Product c ./ 3 l 0“P “8 Perm eate End plug FIGURE 1.7: Schematic of a Straight Through (ST) Capillary Membrane Module. Advantages of ST Membranes: 0 Low space is required due to high density of packing material. 0 Cleaning is easily accomplished by back flushing. - Due to hollow fiber design, utility costs are low, thus decreasing operating expenses. Disadvantages of ST Membranes: o Membranes are susceptible to damage. a Fouling is frequent, particularly when material with high suspended solids is utilized. 0 Ongoing capital costs are high; if module is damaged it must be replaced in its entirety. 32 Aseptic Processing and Packaging TECHNOLOGY OVERVIEW A brief history of aseptic processing and packaging during the twentieth century appears in Table 1.1 (Holdsworth SD. 1992). TABLE 1.1: Synopsis of Developments in Aseptic Technology for Food , .. Preservation. 7 I Period I Heat/Cooling System I Packaflgll’roducts 1920’s Steam injection Systems (G. Grinrod) C.O. Ball pioneering work on heat & Plate heat exchangers (R Seligrnan) penetration process with American Can 1930’s Cmnpany 1940’s Sterideal tubular heat exchangers (Stork) G Grinrod Avoset process Continental Can Company Developmmts 1950’s Use of scraped surface heat exchangers Wm. Mck. Martin develops Dole Canning (Dole) System Smith-Ball Flash 18 process Large container filling - 55 gallon drums (1958) 1960’s Use of aseptic paperboard (Tetra and R. Rausing development of Tetra Pak Scholle) (1961) and Tetra Brik aseptic (1968) Bag-in-box (1968), Scholle (1990) Single portion puddings (1969) 1970’s Use of aseptic plastic (Bosch, Hamba, Extension of technology to aseptic filling Gasti and Remy) of cups and pots, pouches and sachets, bulk tanks 1980’s Particulate sterilizers developed Ohmic 1981 FDA allows H202 sterilization of heater Jupiter, Steriglen and polyethylene single flow specific (APV Crepaco) 1985 extension to other materials Developmart of particulate fillers 1986 Chunky soups in cartons (Nestle) Product sterilization had its roots in conventional canning. Independent package sterilization was commercialized in the 1950’s with the Dole System for low acid puddings. Significant commercial development of innovative processing and packaging systems has developed since the 1980’s. Tetra Pakm (Lund, Sweden): Tetra Pak” was founded in the early 1950’s by Ruben Rausing. He envisioned an innovative method to efficiently distribute liquid a: 33 foodstuffs, in particular, milk. His idea was to form a tube from a flat web of packaging material, to fill this tube with liquid and then to seal through the liquid to form a package. This concept led to the standardized and global Tetra System. The first package was the tetrahedron (Tetra Classic), and then this led to the rectangular shape (Tetra Brik) and then to other packages during the 1960’s. Rausing’s concept was that a package should efficiently fulfill the “purpose of packaging,” which is to transport food from the site of production to wherever it is consumed. A package should in this way “save more than it costs.” From this basic vision also evolved the concept of a packaging system — not only packaging material and packaging machines but also distribution equipment including everything from straw applicators to palletizers and shrink wrappers. In 1991, Tetra Pak“ acquired Alfa Lade and as a result of this acquisition also integrated into Tetra Pak" the business of food processing. This processing primarily includes Ultra High Temperature (UHT) equipment for pasteurization and sterilization. These developments allowed Tetra Pak” to develop and supply a completely integrated (“turn-key”) process and packaging system for delivering food products globally that meet consumer needs from a taste, convenience, cost and safety perspective. Vertical-Form—Fill-Seal fillers are common in the beverage packaging industry. A typical aseptic system from Tetra Pak” appears in Figure 1.8. LS Strip Applicator Initial Folder /7 / «I U :[ Fluid Filling System / Packfizizlz’lniterial fig 1 414050 Final Folder FIGURE 1.8: Schematic Diagram of a Typical Tetra Vertical Form Fill-Seal Filler Illustrated by Components of the System. Printed roll-stock paper is fed from a material magazine and formed into a wedge. This paper is sprayed with a hydrogen peroxide (3 0% solution) as chemical sterilant and dried with heat to remove residual sterilant. Federal standards require monitoring and removal of sterilant. The product is then filled into the package and heated jaws hermetically seal both ends. An applicator strip for either a pour spout or straw is added prior to the product filling. This system allows for an entire closed-loop aseptic processing of package and product. These systems normally have excellent line eficiencies running greater than 90%. Line speeds average 100 cartons per minute. Following forming and filling, the drink box is conveyed through a series of secondary or downstream packaging operations (Figure 1.9). Sli “E 35 Typical Packaging Line Modified TBA/19 Filler Tetra Multi Shrink wrapper / / Palletizer / Tetra Cardboard Packer Figure 1.9: Typical Process and Package Configuration Utilizing Tetra which is required for normal distribution channels. Pak® Technologies“ The package is conveyed to a shrink tunnel where packages are wrapped into a 3-pack. The 3-pack is then placed into a secondary package, in this case a wrap-around tray pack dimensions depending upon customer requirements; however, 12- and 24-count trays are standard. Once completed, the finished product is palletized and stacked ready for warehousing and distribution. The individual package is usually of multi-layer material (Figure 1.10). This tray pack can be of various [lei are r a [DJ ‘2; INSIDE Internal Polyethylene .003” Adhesive Layer .0004” Aluminum Foil .00027” Polyethylene Lining .001” Paperboard .010 - .011” Print Layer ----- External Polyethylene .0007” Figure 1.10: Schematic Diagram of the Aseptic Roll Stock Formed into a Package Illustrating Functionality by Barrier. This material serves several functions such as protecting the product fi'om external stresses (e. g., oxygen), and providing necessary food to package contact areas. This material is necessary to preserve the quality of the finished product. The key multi-layer components are as follows: 1. Internal Polyethylene The internal polyethylene layer primarily provides the food product contact surface of the container. It also gives support to the aluminum foil layer. Adhesive Layer The adhesive layer bonds the aluminum foil and the internal polyethylene layer. Aluminum Foil The aluminum foil layer protects the product from external gas and light contamination. Polyethylene Lining The polyethylene lining insures adhesion between the paperboard and the aluminum foil, and it adds stability to the foil. . Paperboard The paperboard layer acts as a frame and provides strength and stability to the package. 37 6. Print Layer The print layer provides the surface for container decoration and presentation. 7. External Polyethylene The external polyethylene layer provides protection against external humidity and liquid contamination. Product Development Stratggies NUTRITIONALLY ENHANCED PRODUCTS FOR CHILDREN Weaning Foods: Wean (Old English wenian) : “to accustom a child to loss of mother ’5 milk " (Webster ’5 Third International Dictionagg, 1989). Use of the term varies. Some apply the term to complete discontinuance of breast or bottle suckling; others to the addition of supplemental foods when breast milk becomes inadequate in either protein or energy for adequate growth. Weaning foods are of primary importance as supplements to breast milk for infants and as a primary food source for small children. Valverde and Rawson (1976) found that village children in Costa Rica suffer greater morbidity effects than urban children. Developing low-cost, locally-grown supplements with high nutritional quality is necessary in many less developed countries (LDC) (Pellett and Mamarbachi, 1976; Hafex et al., 1986). Research conducted by Karyadi et al. (1990) and Olaofe (1988) indicates locally- grown vegetables can be utilized as protein-caloric foods of great quantity and high quality in Indonesia and Nigeria. The need for weaning foods in less developed countries is well documented (Adewusi et al., 1991; Devadas et al., 1984). A rural Indian mother typically produces h hr to b (Ade ”PM brfi) Kain fo’mul how mil-1mg 38 400-600 ml of breast milk daily (Devadas et al., 1984). This amount is sufficient to meet a child’s nutritional requirement only up to six months. Children 0-5 years in Indonesia have demonstrated faltering growth due to the lack of weaning food supplements to breast milk (Kusin et al., 1984). Guiro et al. (1987) developed a pearl millet weaning food in Senegal, which proved to be effective for the treatment of protein-caloric malnutrition (kwashiorkor). It is apparent that inexpensive protein-based foods must be formulated for the LDC mother to make and feed to her child to supplement breast milk. Numerous formulated products suitable for weaning foods have been proposed. Eka (1978) and Uwaegbute and Nnanyelugo (1987) found that fortified corn paps and millet porridges were nutritionally comparable to the commercial product Cerele manufactured by The Nestle Corporation in Geneva, Switzerland. These foods were fortified with maize and cowpeas locally grown in Nigeria. However, the Cerelac" formula was significamly higher in sodium, potassium and calcium (F atoki and Bamiro, 1990) because of the added milk (versus the local vegetable proteins). It is not uncommon in countries like Nigeria for high carbohydrate foods such as cassava starch (1.0% protein) to be consumed in greater quantities than sorghum, which has an 8.0% protein level (Adewusi et al., 1991). Multi-mixtures of formulas with rice, millet, pasta and bulgar and supplementary ingredients of egg, fish, chickpeas and mung bean may serve as the basis for formulating inexpensive but nutritious, infant-weaning food (Abbey and Nkanga, 1988; Ketiku and Olusanya, 1986; Mathew and Pellett, 1986; Ulloa et al., 1988). These formulations do not necessarily need to be complex blends or recipes. Marero et al. (1988) showed germination rice/mung bean formulas (7 0/30 mixture) were well accepted by infants and provided one-third of the RDA requirement of protein and energy while 39 remaining microbiologically safe for six months. When traditional Egyptian bread has been fortified with 15-20% lentil flour, it has shown significant enhancement in Protein Efiiciency Ratio (PER) values. Reddy et al. (1990) also formulated locally-grown grains into four weaning mixtures, all of which met the criteria for inexpensive and nutritious foods compared to commercially available options in India. All recommendations for foods must assure inactivation of anti-nutritional factors. For example, uncooked beans may cause a severe health risk because of complete destruction of phytohemagglutin (Korte, 1972: Cofl‘ey et al., 1985). Technology can also play a major role in the development of low-cost, nutritious food and culturally appealing products. Drum-dried mixtures of rice, cowpeas and milk, sealed in cans, were shown to have no protein loss after six months of ambient storage (Roman et al., 1987). In Costa Rica, Fernandez et al. (1980) and Sellers (1988) indicated that the installation of commercial fruit and vegetable processing facilities improved agronomic inputs and provided rural employment while decreasing the burden of food subsidies on the national economy. This industrial development has profound impact on the economics and type of products suitable for weaning foods. Developing countries can also utilize low-cost extrusion to develop highly nutritious supplementary foods from locally grown food crops (Jansen et al., 1980). QUALITY DRIVEN PRODUCT DEVELOPMENT Quality Science has become a globally focused field of study. Many forces and individuals have established philosophical and technological basis for quality. These include (1) researchers J. Demming, J. Juran and P. Crosby, (2) globally competitive markets and (3) the overwhelming impact of consumer demands for enhanced value. Quality Function Deployment (QFD) is a process that originated in Japan for managing product development which emphasizes incorporation of appropriate design criteria to assure consumer acceptance. The Mitsubishi shipyards in Kobe are credited with its development in 1972 (Kogure et al., 1983). In 1978 Toyota adopted QFD. Fuji- Xerox initiated QFD to decrease product development time. Recently, many U. S. corporations (e. g., Ford Motor Company, 1983) have incorporated QFD into their practices which enhance the Total Quality Management Systems (TQM). US. firms are experiencing the fact that this process integrates product development activities well into company-wide Quality Control programs for superior overall product management. Sullivan (1986) reported that many US. companies have developed quality philosophies in recent years. When integrated into the workplace culture, these philosophies must be deployed both vertically and horizontally as part of the company policy in order to change the thinking and ultimately the behavior of employees. The main objective of any manufacturing company is to bring superior, new products with sustainable competitive advantages to market faster than its competition, at lower cost and with improved quality. This “Excellence in Execution” is the main mechanism embraced by Japanese corporations. There are six key terms associated with QFD as reported by Sullivan (1986): 1. Quality Function Deployment (QFD) is an overall concept that provides a means of translating customer requirements into the appropriate technical requirements for each stage of product development and production. 2. The Voice of the Customer (VOC) allows customers to express each product requirement through their own. 41 3. Counterpoint Characteristics (CC) are the expression of the voice of the customer into technical language that specifies customer-required quality; counterpoint characteristics are critical final product-control characteristics. 4. Product Quality Deployment (PQD) activities translate the voice of the customer into counterpart characteristics. 5. Deployment of the Quality Function activities assures that customer-required quality is achieved with the assignment of specific quality responsibilities to specific departments. 6. Quality Tables are the series of matrices used to translate the voice of the customer into final product-control characteristics and establish the final “House of Quality” format. Traditional Process: Phase Review: Figure 1.11 represents a so-called phase-review product-development process where activities occur in series with many points of management approval (Urban et al., 1980) 42 OPPORTUNITY IDENTIFICATION GoHNoGo CONCEPT DEVELOPMENT Go H NoGo PRODUCT Reposition DESIGN Go H No Go PROCESS DESIGN Go LA NoGo COMMERCIAL PRODUCTION V FIGURE 1.11: Product Development: Phase Review Process Development proceeds sequentially, with different fimctional groups participating in different development phases, then passing their results on to the next fimctional group (Rosenthal and March, 1991). This non-interactive type of approach leads to a typical “over the wall” project where project engineers rarely talk to marketing, the company does not consider the customer and manufacturing is the last to find out. Moreover, this type of phase-review pathway has little customer feedback and no early warning system to alert management of problems. This pathway thus causes long commercial development cycles that increase project costs. In essence, cross-functionality is lost and quality is driven by internal focus rather than by customer or external focus. The primary benefit of QFD is that it is a totally externally defined customer focus done internally among cross- 43 functional departments. Grifin and Hauser (1991) show the formal process for QFD represented in Figure 1.12. QFD PROCESS Design Attributes ustomer ustomer Design Needs Matrix Engineering Features Measures D . 9'3" Operating Houaze. of Attributes Matrix Q“ rty Measures Process Steps Features Control Matrix Measures Operational Conditions Process Steps Measures FIGURE 1.12: Formal Process of Quality Function Deployment (QFD). The primary aim of this process is to incorporate the “Voice of the Customer” into the product-development and process-development requirements that profitably deliver the identified customer needs and wants (Sullivan, 1987). QFD manages across individual functional departments for new product development, providing a mechanism for a coherent process (Hauser et al., 1988; Sullivan, 1986a). Daetz (1990) indicated that it brings together all the elements of definition, design, and product delivery to meet or exceed customer needs. 44 The most familiar term with QFD is “House of Quality” (Hauser et al., 1988). As shown in Figure 1.13, “House of Quality” relates data generated from market research to be translated. I Design _ Attributes Customer Needs 5 ‘ (200-300 in hierarchy) . . Clarity Crispness of Lines 2 Relationships Distinguish Detail between Read Graphics Text 1 Customer Needs More than One Person and NEXT NEC IBM No Eye Easy to Read Text . . Strain Flicker not Noticable 4 D9318“ Attributes Customer Comfortable Eye Level Perceptions / Costs and Feasibility Importances I “Engineering” I Measures FIGURE 1.13: House of Quality Blueprint of Externally Defined Design Attributes. This represents the “blueprint” of what inputs are necessary to achieve customer- accepted design. The second QFD matrix relates potential product failures to deliver performance characteristics. Matrices three and four bring process characteristics and production requirements into relationship with Engineering and Marketing. A contrasting comparison of QFD and Phase-Review processes appears in Table 1.2. 45 TABLE 1.2: Contrast of QFD and Phase Review Processes Qharactemt‘ ics of QFD Phase Review Process Elements 0 Simultaneous development across - Sequential, iterative development functions 0 All functions participating from start 0 Function involvement by phase 0 Team empowered to make decisions 0 Management approval after each phase a Tasks shared across functions 0 Tasks assigned by function - Consensus decisions about trade-offs - Functionally led trade-ofi‘ decisions 0 Working meetings to develop results jointly - Presentation meetings to present results Perhaps the best way to summarize QFD is to look at its results at Toyota. Toyota claims that QFD virtually eliminated all warrantee problems associated with rust. It also allowed Toyota as a company to reduce product development costs and decrease time to market by as much as 40% (Eureka, 1987; Sullivan, 1986 a and b). It is proposed that principles of QFD be applied to the food industry. Nutritional Labeling and Education Act (NLEA) The NLEA, which was signed into law on November 8, 1990, represents the first comprehensive revision of the food labeling requirements of the Federal Food, Drug, and Cosmetic Act (FD&C Act) since its enactment in 1938 (Hutt et al., 1993). The new legal statute is lengthy, detailed, complex and in some cases controversial, since it attempts (I) to respond to a heightened public awareness of the efi‘ect of certain foods on health, (2) to take into account many conflicting scientific claims, and (3) to prevent manufacturers from being overburdened by new regulatory requirements while ensuring public health and well- being. The various provisions of the NLEA can be divided into six major categories: nutrition labeling, nutrient descriptors, disease prevention claims, general labeling provisions, national uniformity and state enforcement in federal courts. During 1991 and 1992, the Food and Drug Administration (FDA) proposed implementation of regulations specifying compliance directives in these areas, with focus on the statutory deadlines of November 8, 1992, when most final regulations were to be promulgated. These regulations became effective for labeled food six months later, on May 8, 1993. This new law maintains the current division of authority between the United States Department of Agriculture (USDA) and the FDA for regulating the labeling of meat, poultry, and eggs. ('1 ("D t D A STUDY ONE Protein Separation for Navy Beans Utilizing Ultrafiltration and Microfiltration Technology Introduction In this study, dry edible beans (Phaseolus vulgaris L.) were used as the protein source for evaluation of various separation techniques utilizing selective membranes based on pore size. Milled beans were passed through a series of molecular weight polysulfone membranes to selectively concentrate proteins. Two types of separation technologies were evaluated: ultrafiltration (UF) and microfiltration (MF). Two types of membranes were also selected: tubular (T) and spiral wound (SW). Protein fractionation was also determined by molecular weight cut-off of the various membranes. In this study, 200,000 (kd) and 50,000 (kd) polysulfone membranes were evaluated. The protein quality was also determined utilizing several quantitative and qualitative analyses. Protein fractions were subjected to simulated heating conditions (71- 99°C/5 seconds) in an effort to measure protein coagulation and denaturation. Qualitatively, confocal microscopy was utilized to determine denaturation based on the degree of coagulation. Quantitatively, denaturation was measured by using a nitrogen solubility index (N SI). Size partitioning was also evaluated with NDND-PAGE in an efl‘ort to determine native protein molecular weight achieved through partitioning with UP and MF membranes. Amino acid scores were calculated for each fraction to determine protein quality based on the limiting level of amino acid. 47 IX Ill usir hon HC MET IA. .‘lk [qr 48 Matefl’ and Mgggg EXTRACTION AND SEPARATION Microfiltration: Dry navy beans were hammer milled (13.9% moisture) to produce a flour meal using a standard Fitzrnill (1/8” screen). Bean flour was soaked (4:1 tap water) for 24 hours at 4°C in a 1.7% solution of NaCl in which the pH was adjusted to 9.0 using 3N sodium hydroxide. The extracted supernatant was decanted and adjusted to pH 7 with 3N HCl and filtered to remove suspended solids and produce a clear solution. This filtered material was processed through a series of ultrafiltration tubular type polysulfone membranes in a pilot scale system (Table 2.1). TABLE 2.1: UF Equipment and Tubular Specifications Used for Navy Bean Protein Fraction in Trial 1 Microfiltration System Equipment: APV Crepaco, Towanda, NY APV Crepaco Equipment Specification for BRO/BUF Membrane Filtration Pilot Unit, Serial No. 21202 Specifications: UF Configuration 9.3 it 2 (0.9 m2) Membrane Area Two PCI type B1 modules having 9.3 ft 2 membrane area/B1 module Bl module type heat exchanger with stainless steel shroud (2 feet long) Feed Pump Capacity: 7.5 gpm at 100 psi (UF) Hold-Up Volumes: BASIC UNIT (excluding modules) 1.32 gallons (estimated) TUBE SIDE - .8 gallons (estimated) PERMEATE SIDE - 2.4 gallons (estimated) ll] ”Cl? 49 Separation through a 200,000 (kd) molecular weight tubular membrane resulted in the recovery of two fraction streams: (1) permeate I and (2) retentate I. These fractions were then ultrafiltered through a 50,000 (kd) molecular weight spiral wound membrane which yielded two streams designated (1) permeate II and (2) retentate II. Retentate H was used for formulation of the final product. A detailed process flow and mass balance was prepared and is presented in Figures 2.1 and 2.2. Whole Bean Seafarer I Extraction ‘ Filzmill, 1/8” Resrdue Soak((4:l): 24 hours@?t °c 1.7% NaCl Solution, pH 9.0 f Supemant +0.8 parts 3M HCl:pH7 Solids - 4.32% Protein - 28.97% I Microfiltration 200,000 MW Membrane Tubular I r 7 Retentate Permeate (TE!) (TPl) Ultrafiltration Separator 50,000 MW Membrane Spiral Wound J i I J I l Retentate Permeate - Retentate Permeate (TR1-R2) (TRl-P2) (TPl-R2) (TPl-P2) FIGURE 2.1: Dry Navy Bean Protein Fractionation Using both a Microfiltration (200,000 MW Membrane) Tubular System and Ultrafiltration (50,000 MW Membrane) Spiral Wound Membrane. Sample Code Designations are Indicated. F—T'l HG 50 Whole Bean Smfarer I . Extraction Resume ~ (Emu, 1/8”) Soak (4:1): 24 hours @ 4 °C 1.7% NaCl Solution, pH 9.C _I Supermnt +0.8 parts 3M HClsz7 Solids - 4.44% Protein - 29.18% I Microfiltration (0. 3 micron) 200,000 MW Membrane Spiral Wound l I , I , l - Retentate Parnmte Retentate Permmte (SR1-R2) (SR1-P2) (SP1-R2) (SP1-P2) FIGURE 2.2: Dry Navy Bean Protein Fractionation Using Microfiltration (0.3 micron) 200,000 MW Spiral Wound and Ultrafiltration Through a 50,000 MW Spiral Wound Membrane. Sample Code Designations are Indicated. Process utilizing Batch (Feed) FIGURE ; Link alld I: and Sample a 51 Process schematics for MF and UF are presented in Figures 2.3 and 2.4. Microfiltration utilizing a 200,000 MW cut-ofi’ tubular membrane is presented in Figure 2.3. Batch , Recycle (W) r>O m =5: =9 : =:=:=.-=.:=.-=.<-.- Tubular Membrane Pump——> 0:. °:- -:- -:-.-."..:'-.-:-'- _> {..{..,..i..3..5.. -—> ‘ Permeate i Retentate Collection , Collection FIGURE 2.3: Microfiltration Batch System Utilizing Three 200,000 MW Tubular Membranes in a Filtration Series. This system had a series of three membranes and was batch fed from an external tank and pumped through the tubular system. Diafiltration was accomplished in a continuous system with permeate and retentate fractions captured for flux rate assessment and sample analyses. A second 53' Batch (Feed) Tank Pumpi— HGURE 2.4: WOLlnd membrane “Stem Operated in a ”filling a 50,000 [cm 52 A second system, which was also used for microfiltration, is depicted in Figure 2.4. Batch Recycle eed _ (F ) Holding Tank 200,000 MW . microfiltration Il’ump l——-| Dairy Pump 8 Wound (0,3 Retentate Filter ‘ ‘ ‘ \- z Permeate Outlet, Inlet , Pressure COIICCtlon Tefn::::re FIGURE 2.4: Batch System Utilizing a Microfiltration 200,000 MW (0.3 micron) Membrane and Ultrafiltration 50,000 MW Spiral Wound Membrane System. This system was also batch fed but utilized a 200,000 MW (0.3 micron) spiral wound membrane. The details of this membrane system are shown in Table 2.2. This system operated in a similar manner to the tubular system, except various membranes could be tested. This system was also employed to test the ultrafiltration specifications utilizing a 50,000 MW spiral wound membrane. TABLE 2.2: Microfiltration Sy; Equipment APV Crepac Work No. 2 Specifications: UF/MF Con Spir: Tubr Plate Membrane 1 Spir: B] n the : ViSt: \ 53 TABLE 2.2: MF Equipment and Tubular Specifications Used for Navy Bean Protein Fractionation. Microfiltration System Equipment: APV Crepaco for Membrane F iltrations Pilot System Work No. 21202 Specifications: UF/MF Configuration Spiral Wound (mode used for separation) Tubular Plate and Frame Membrane Area Spiral wound area approximately 20 fi2 each B1 module type heat exchanger with stainless steel shroud (2 feet long) for the 50,000 MW cut-off. The elements were manufactured by Osmonics, Vista, CA (PW2540C1077) A method was developed to simulate heating during thermal processing. A schematic of this is shown in Figure 2.5. x. Oil Medium PLC m Loop _ Data AC(luisiti \ ”GUREZS: Th fiacthnal Ar‘f Retort Pouch Sealed with Test Medium - i | N“ Pouch Holder L, R» 0‘1 01 Heat .__. 1 1 Source Medium Medium I __ PLC Loop -/\_/ Data Acquisition FIGURE 2.5: Method to Simulate Thermal Process Conditions. Protein Sources are Sealed in Metallized Retort Pouches, Placed in an Oil Bath and Heated to 99°C (210°F) for 5 to 30 Seconds. Thermal Data are Collected On-line. The fi'actionated protein material was heat sealed in a metallized retort pouch. These pouches were placed in a heated oil bath from 71°C to 99°C for 5 to 30 seconds. The pouch was then removed and air cooled to simulate heat transfer properties of pasteurization. The material was prepared in triplicate (n=3) and utilized for qualitative (microscopy) and quantitative protein analyses (Kjeldahl nitrogen and electrophoretic patterns). QUANTITATIVE Electrophoresis: Nondenatur (PAGE) was perfc acrylamide gradient ml Tris, 80 mM bl pH 8.5, with 3% ac mM Tris, 80 mM b! volume of bufl‘er (. mixed with l volum Electrophor. Node! SE 600', H voltage Power supp PA). A Constant « running gel and the trackjng dye reache. 0yemlght in 0.4% C multicular Wei ghts 55 QUANTITATIVE AND QUALITATIVE ANALYSES Electrophoresis: Nondenaturing and nondissociating (NDND)-polyacrylamide gel electrophoresis (PAGE) was performed according to Andrews (1986). The process used 10% linear acrylamide gradient gels of 0.75 or 1.5mm thickness (acrylamide to his ratio 37: 1) with 90 mM Tris, 80 mM boric acid, 2.5 mM Na-EDTA (sodium-ethylenediamine tetraacetic acid) pH 8.5, with 3% acrylamide stacking gels. The running buffer for NDND-PAGE was 90 mM Tris, 80 mM boric acid and 2.5 mM Na-EDTA, pH 8.4. Samples were mixed with 1 volume of buffer (2 volumes of 0.45 M Tris, 0.4 boric acid, and 12.5 mM Na-EDTA mixed with 1 volume of glycerol) . Electrophoresis was carried out with a Hoefi'er Vertical Electrophoresis unit (Model SE 600; Hoeffer Scientific Instruments, San Francisco, CA) using a constant voltage power supply (Fisher Biotech Electrophoresis System, Model PE 458, Pittsburgh, PA). A constant current of 20 mA was applied until the proteins migrated into the running gel and then the current was increased to 30 mA until the bromophenol blue tracking dye reached the bottom of the running gel. The gels were removed and stained overnight in 0.4% Coomassie Blue in 9/45/45 (v/v/v) acetic acid/methanol/water. Subunit molecular weights of studied proteins were estimated using a mixture of standard molecular weight protein markers (MW -ND-500 Kit, molecular weight range of 14,000- 545,000) purchased from the Sigma Chemical Corp. St. Louis, MO. The NDND protein mixture consisted of the following proteins: or-lactalbumin (14.2 kilodalton), carbonic anhydrase (29.0 kilodaltron), chicken egg albumin (45.0 kilodaltrons), bovine serum albumin (66.0 kilodaltrons for monomer and 132.0 kilodaltrons for the dimer), urease (272.0 kilodaltrons ' protein solutions We Bulletin No. MKR- (Kll) of the protein RM: A plot of relative m0 relative mobility of I estimated from the qualitatively compare were Identified by the 56 (272.0 kilodaltrons for the trimer and 545 .0 kilodaltrons for the hexamer). The standard protein solutions were prepared according to the method described in Sigma Technical Bulletin No. MKR-137 (Sigma Chemical Corp., St. Louis, MO). The relative mobility (RM) of the protein standards was calculated using the formula RM = Distance of Protein Miggation (cm) Marker Dye Distance (cm) A plot of relative mobility vs. molecular weight was constructed as a standard curve. The relative mobility of each protein subunit was calculated, and the molecular weight was estimated from the standard curve. The protein bands present on the gels were qualitatively compared using a Microtec Scamnaker (Model 600Z). The protein bands were identified by their subunit molecular weights. SDS Gel Electrophoresis: Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SD S-PAGE) of the bean protein fi'actions including albumin, GI and GH was performed on 12. 5% acrylamide running gels with 4.5% stacking gels using the system of Laemmli (1970). A protein solution of each bean protein was prepared to a final concentration of 10mg nitrogen sample per mL of buffer, heated and mildly vortexed until completely solubilized. A 25rd of each protein solution sample was applied into the sample well of stacking gel. Electrophoresis was carried out with a Hoefi‘er Vertical Electrophoresis unit (Model SE 600; Hoefi‘er Scientific Instruments, San Francisco, CA) using a constant voltage power supply (Fisher Biotech Electrophoresis System, Model FB 45 8, Pittsburgh, PA). A constant current of 12 mA was applied for 14 hours, until the bromophenol blue tracking dye reach for8 hours, in 0.49- Subunit mC standard molecular Weight Markers a1 Chemical Corp, 8‘. proteins: a-lactalbr (24 kd), carbonic a egg albumin (45 kd f0110“:ng proteins: kd). phosphorylase mndard protein sc Technical Bulletin b mOblljty (RAJ) Oflhr 57 tracking dye reached the bottom of the running gel. The gels were removed and stained for 8 hours, in 0.4% Coomassie Blue in 9/45/45 (v/v/v) acetic acid/methanol/water. Subunit molecular weights of studied proteins were estimated using a mixture of standard molecular weight protein markers (SDS-7 Dalton mark VII - L/Low Molecular Weight Markers and SDS 6H/High Molecular Weight Markers) purchased from Sigma Chemical Corp, St. Louis, MO. The SDS-7 protein mixture consisted of the following proteins: a—lactalbumin (14.2 kilodalton), soybean trypsin inhibitor (20.1 kd), trypsinogen (24 kd), carbonic anhydrase (29 kd), glyceraldehyde-3 -phosphate dehydrogenase (36 kd), egg albumin (45 kd), bovine albumin (66 kd). The SDS-6H protein mixture contained the following proteins: carbonic anhydrase (29 kd), egg albumin (45 kd), bovine albumin (66 kd), phosphorylase B (97.4 kd), B-galactosidase (116 kd) and myosin (205 kd). The standard protein solutions were prepared according to the method described in Sigma Technical Bulletin No. MWS-877C (Sigma Chemical Corp, St. Louis, MO). The relative mobility (RM) of the protein standards was calculated using the formula RM = Distance of Protein Miggation (cm) Marker Dye Distance (cm) A plot of relative mobility vs. molecular weight was constructed as a standard curve. The relative mobility of each protein subunit was calculated, and the molecular weight was estimated fiom the standard curve. The protein bands present on the gels were qualitatively compared using a Microtec Scanmaker (Model 6002). The protein bands were identified by their subunit molecular weights. This method was used with and without 5% 2-mercaptoethanol. Micro-Nutrients: Amino Acid: The fraction was determi glutamic acid, proli phenylalanine, histic samples was mixed Plfi’em oxidation of flask by repeated he sealed, and the flag hydrolyzed to amino Opened, mixed with i 58 Micro-Nutrients: Amino Acid: The concentration of eighteen amino acids of each major bean protein fraction was determined. The amino acids measured were aspartic acid, threonine, serine, glutarnic acid, proline, glycine, alanine, valine, methionine, isoleucine, leucine, tyrosine, phenylalanine, histidine, lysine, cysteine, tryptophan, and arginine. A portion of the samples was mixed with ahydrochloric acid solution in a modified Kjeldahl flask. To prevent oxidation of the amino acids, as much oxygen as possible was removed from the flask by repeated heating and freezing, under vacuum. The neck of the flask was heat sealed, and the flask heated for 20 hours at 110°C. The protein in the sample was hydrolyzed to amino acids by the hot hydrochloric acid solution. The samples are cooled, opened, mixed with internal standard, and adjusted to pH 2.2. The amino acids were separated on an ion exchange column, in an amino acid analyzer, by a pH gradient elution with controlled column temperatures. The separated amino acids were subsequently reacted with ninhydrin, forming color-complex solutions that were measured spectrophotometrically. The concentration of each amino acid was quantitiated against a standard solution of amino acids of known concentration, and internal standard, which was injected into the amino acid analyzer. Cysteine was oxidized to cysteic acid and methionine oxidized to methionine sulfone by treatment with a performic acid solution for 16 hours at 0° C. The excess performic acid was destroyed by reduction with hydrogen bromide, followed by evaporation to dryness on a rotary evaporator, removing any bromine gas that was generated. The sample was hydrolyzed with a hydrochloric acid solution for 18 hours at 110°C. The hydroly reconstituted in a sod? Cysteic acid 2 and reacted With r. mehotometricar asainst acomb'med ,1 ninth was taken thrr Methods 0fthe AOA TryptOphan V was rinsed With a 5‘ addition of the amj removed from the fl; filed, thawed and hi'droll’md to amim moled, Opened, “cu m - . - assure liquid chm] 59 110°C. The hydrolysate was spiked with an internal standard, taken to dryness, and reconstituted in a sodium citrate solution. Cysteic acid and methionine sulfone were separated on an ion exchange column and reacted with ninhydrin, forming color-complex solutions that were measured spectrophotometrically. The concentration of cysteine and methionine was quantitiated against a combined standard of cysteic acid and methionine sulfone, with internal standard, which was taken through the method and injected onto the amino acid analyzer. Oficial Methods ofthe AOAC. 1th Edition, (1995) Method 985.28, Locator # 45.4.05. Tryptophan was measured utilizing an alkaline hydrolysis. A portion of the sample was mixed with a sodium hydroxide solution in a modified Kjeldhal flask. To prevent oxidation of the amino acids as they were hydrolyzed, as much oxygen as possible was removed from the flask by freezing and evacuating under vacuum. The sample flask was sealed, thawed and heated for 20 hours at 110°C. The protein in the sample was hydrolyzed to amino acids by the hot sodium hydroxide solution. The samples were cooled, opened, neutralized and bufi‘ered to pH 4.2. The samples were injected on a high pressure liquid chromatograph (HPLC), measured by ultraviolet spectrophotometry, and quantitiated from standards of known concentration that were injected on the HPLC. Oflicial Methods of the AOAC, 16'‘1 Edition, (1995) Method 988.15, Locator #45404 (Modified). Amino Acid Score: The amino acid score involved a comparison of the test protein with an ideal reference pattern of amino acids. The amino acid score of a food protein is an evaluation of its nutritional quality and is calculated as follows: Score of Test Protein lire amino acid with limiting 311.1110 acid.” Water Soluble Nitro Water soluble Method 46-23. The tubes. TUbes were deemed throngh a ”Slag AACC MEIhOt “We B = In] back-t 0 60 Score of Test Protein Content of each indispensable amino acid in food protein (ngg of protein) x 100 Content of same amino acid in reference protein (mg/ g of protein) The amino acid with lowest score in relation to the reference pattern is termed the “first limiting amino acid.” Water Soluble Nitrogen: Water soluble nitrogen extracts were prepared by a modification of the AACC Method 46-23. The bean protein fractions were weighed (5 g samples) into centrifuged tubes. Tubes were centrifiiged for 10 minutes at 2500 rpm. The supernatant was decanted through a pipette (2.5 ml of clear liquid) and analyzed for Kjeldahl nitrogen using AACC Method 46-11. % Water soluble N = (B-S)xNx0.14x100 wt. of sample where B = ml back-titration of blank, S = ml back-titration of sample, N = normality. % Nitrogen solubility Index (N SI) = %water-soluble N x 100 % total N Light Microscopy: The samples for microscopic examination were prepared by suspending a composite sample of a heated protein fraction on a glass slide. A laser seaming confocal microscope (Zeiss 210, West Germany) was used with a red laser (Helium-neon emitting at 633 nm) and a blue laser (Argon-ion emitting at 488 nm) as shown in Figure 2.6. FIGURE 2.6: 61 Laser Beam Splitter / Collecting Lens :3: Monitor ,,,,,,,,,,,, .5, . -. ’1'," ‘ _—-———-> Confocal Unit with Pinhole Objective Lens Focal Plane Out-of-Focus ' ‘ Area Path of In-Focus Light FIGURE 2.6: Principle of Confocal Scanning - Path Followed by Light Coming from the Focal Plane (White Band) and from Outside the Focal Plane (Dotted Line) in the Zeiss LSM. The Focal Point of the Out-of-Focus Light Falls in Front of the Pin Hole; Thus Most of the Out-of-Focus Information Does Not Reach the Detector. Photomicrographs of representative specimens were obtained after viewing numerous fields and used to draw conclusions regarding the aggregation of protein bodies due to the denaturing heat treatment. MEMBRANE PROCl Microfiltration process ultrafiltration of permez TABLE 2.3: l l Microfiltratio M _\ Time I: 62 MW MEMBRANE PROCESSING Microfiltration processing of bean protein extract and the subsequent microfiltration and ultrafiltration of permeate and retentate fractions are presented in Table 2.3. TABLE 2.3: Microfiltration Flux Rates of Bean Protein Extract Processed Utilizing a Spiral Wound Membrane. Microfiltration Processing of Bean Protein Extract Through a 200,000 MW (0.3 micron) Spiral Wound Membrane. Permeate Retentate Temperature Time Pressure Flux Rate Flux Rate (Min) (psi)In/Out (ml/min) (ml/min) F C 2 85/80 200 1920 58 14.5 10 85/80 150 2000 60 15.6 50 - - - 105 2100 62 16.7 130 60/70 95 2360 68 20.0 175 50/60 95 1840 70 21 .l 200 70/90 90 1600 70 21 . l 250 60/80 88 1540 74 23.4 300 40/60 75 1400 76 24.5 Microfiltration Processing of Permeate Fractionate Using a 200,000 MW Spiral Wound Membrane and Partitioned Through a 50,000 MW Spiral Wound Membrane. Permeate Retentate Temperature Time Flux Rate Flux Rate (min) (ml/min) (ml/1mg F C 5 410 2640 60 15.6 15 400 2800 78 25.6 30 310 2440 88 31.1 Microfiltration Processing of Retentate Fractionated from 200,000 MW Spiral Wound Membrane Partitioned Through a 50,000 MW Spiral Wound Membrane. Permeate Retentate Temperate Time Flux Rate Flux Rate (min) (ml/min) (ml/min) F C 2 17 5 1800 60 15.57 10 140 1660 60 15.57 The W Of this (0.3 micron) 59“?“ WO minutes to 300 minutt fraction decreased frOI 1400 ml/min. There v minutes at 145°C t fractionated penneau clearly indicated that protein material extr made it difficult t denamration of prot. Additional 1\ Whale and subs Table 2.4. 63 The MP of this bean protein extract was accomplished utilizing a 200,000 MW (0.3 micron) spiral wound membrane. The data indicated that as time increased fi'om 2 minutes to 300 minutes a significant drop in flux occurred. The flux for the permeate fiaetion decreased from 200 ml/min to 75 ml/min and in the retentate from 1920 ml/min to 1400 ml/min. There was also a significant increase in temperature from an initial time of 2 minutes at 145°C to 300 minutes at 245°C. Subsequently, UF processing of the fractionated permeate and retentate showed similar results in loss of flux. These data clearly indicated that spiral wound membranes are extremely prone to fouling when using protein material extracted from dry beans. This decrease in membrane performance also made it dificult to achieve efl‘ective separation without mechanical and/or heat denaturation of protein caused by prolonged processing times. Additional MF processing of the bean protein extract utilizing a 200,000 tubular membrane and subsequent UF processing of permeate and retentate fractions appears in Table 2.4. TABLE 2.4: l Microfiltrati , MW Tubula Time fl; 5 4 is 30 TABLE 2.4: Microfiltration Flux Rates of Bean Protein Extract Utilizing Tubular Membrane. Microfiltration Processing of Bean Protein Extract Through a 200,000 MW Tubular Membrane. Permeate Retentate Time Pressure Flux Rate Flux Rate (min) (psi) (ml/min) (ml/ min) 5 100 560 3740 15 170 480 10,020 30 140 420 8160 45 135 370 10,320 60 135 370 9000 Permeate Fractionated from 200,000 MW Tubular (UF) Membrane Separated Through 50,000 MW Spiral Wound @117) Membrane. Permeate Retentate Temperature Time Pressure Flux Rate Flux Rate (psi) (ml/min) (ml/min) F C In/Out 5 80/80 210 1460 70 21.13 10 80/80 180 1540 70 21 . 13 Retentate Fractionated from 200,000 MW Tubular (MF) Membrane Separated Through 50,000 MW Spiral Wound (UF) Membrane. Permeate Retentate Temperature Time Pressure (psi) Flux Rate Flux Rate In/Out (ml/min) (ml/min) F C 5 100/95 280 1500 70 21.13 10 50/50 100 1500 70 21.13 The system was run for a total of 60 minutes; flux rates also decreased with time for the permeate (S60 — 370 ml/min) and retentate (10,020 - 9,000 ml/min). These flux rates were acceptable; separation was apparent afier 60 minutes since the amount of permeate recovered from the diafiltration (recycle) stream was not in appreciable quantities. These permeate and retentate fractions were also subjected to UF processing through a 50,000 MW spiral wound membrane (Table 2.4). These systems were run for only 10 minutes since 5 of diafiltration. These data in membranes, tubular is 1#38 less membrane ‘ Wound membranes. ' “‘0 Systems. The membrane material \ recoguiled that m fomfion (Chemistr Performance Chara. 65 only 10 minutes since sufiicient recovery of the permeate was achieved in the early stages of diafiltration. These data indicated that in a comparison between spiral wound and tubular membranes, tubular is a preferred system for dry bean protein fiactionation because there was less membrane fouling and faster separation with the tubular than with the spiral wound membranes. This result can be attributed to the innate design features between the two systems. The spiral wound has much more surface-area-to-volume ratios of the membrane material whereas the tubular is more of a straight-in-and-out design. It must be recognized that membrane technology is empirical and that methods of membrane formation (chemistry) and physical layout (structure and geometry) greatly influence the performance characteristics. The simpler in-line and shell-and-tube membrane designs used in this study are less likely to foul than the more complex, higher surface area spiral wound membranes. The mass balances for both processing schematics are presented in Figures 2.7 and 2. 8. These material balances include both solids (°/o) and protein (°/o) fiom extraction through final mierofiltration and ultrafiltration treatments. The solids and protein values for the whole bean, initial extract and second extract are presented in Table 2. 5. Whole Bean Solids - 86.1% Protein 49.08% l Fitzmill, 1/8” i Extraction Soak (4:1): 24 hours @ 4 °C 1.7% NaCl Solution, pH 9 I l 7 Residue Supemant 11 +0.8 parts 3M HCltpH‘I Solids - 4.44% herein - 29.18% db I Microfiltration 200,000 MW Membrme Tubular Retentate Permeate (Emmi W Solids . 6.06% Solids - 3.18% Protein - 40.00%db Protein - 9.01%dh Ultrafiltration Separator 50,000 MW Membrane Spiral Wound J l Retentate Permeate Retentate Permeate Com—amt C__raaw_nom t M 9m Solids -6.04% Solids - 3.36% Solids - 3.10% Solids - 2.97% Protein- 42.08% db Protein - 10.76% db Protein - 11.21% db Protein - 10.10% db FIGURE 2.7 : Dry Navy Bean Protein Fractionated Using Both a Microfiltration (200,000 MW Membrane) Tubular System and a Microfiltration (50,000 MW Membrane) Spiral Wound Membrane System. Mass Balances for Total Solids and Protein are Expressed as Percents (% db) of Original. 67 Whole Bean Solids - 86.1% Protein -28.97% I Fitzrnill, 1/8” I Extraction Soak (4:1): 24 hours @ 4 °C 1.7% NaCl Solution, pH 9 I L l - m 11 Resrdue Supe ant +0.8 parts 3M HClzpl-N Solids - 4.32% Protein - 28.97% db I Microfiltration 200,000 MW Membrane Spiral Wound (0.3 micron) Retentate Permeate Comment £92m! Solids — 5.72% Solids - 1.14% Protein - 34.79% db Protein - 26.60% db Ultrafiltration 50,000 MW Membrane Spiral Wound l J Retentate Permeate Retentate Permeate Component Commt Commt CM Solids -5.04% Solids - 2.78% Solids - 3.23% Solids - 1.91% Protein- 37.02% db Protein - 9.61% db Protein - 30.81% db Protein - 12.53% db FIGURE 2.8: Dry Navy Bean Protein Fractionation Using Microfiltration 200,000 MW (0.3 micron) and 50,000 MW Spiral Wound Membranes. TABLE 2.5: Commnent Whole Bean Initial Extract Second Extra: Residue Initial Extract Second Extra Residue The Whole 1 {met} rePorted in mracts for trials 1 Processing utilizing the ”mute (40.0< Plotein recovered i: 68 TABLE 2.5: Whole Bean and Extract Protein Values for Bean Flour Extracted with 1.7% NaCl at pH 9.0. Component Membrane Solids (%) Protein (%) Whole Bean 86.10 19.08 Initial Extract MF 5.89 38.96 Second Extract MF 4.44 29.18 Residue MF 22.99 1 1.25 Initial Extract UF 5.86 36.29 Second Extract UF 4.32 28.97 Residue UP 21 .89 8.08 The whole bean protein level was at 19.1%, which is consistent with previous research reported in the literature and demonstrated in Study One. The initial and second extracts for trials 1 and 2 are also consistent with previously reported work (Study One). Processing utilizing microfiltration tubular membranes yielded greater protein recovery in the retentate (40.00%) than did the spiral wound membranes (34.79%). There was less protein recovered in the permeate (9.01%) with tubular separation than the spiral wound membrane (26.60%). This difference of recovery may be due to longer processing times for the spiral wound MF than the tubular MF systems or the better partitioning of protein into the retentate with the use of the tubular membrane. Subsequent UF for both fractions through 50,000 MW spiral wound membranes essentially yielded identical results. These data only reinforce the recommendation of tubular membranes as the preferred processing system technology for separation. Gel Electrophonesis: The overall schematic for ultrafiltration (UF) and mierofiltration (MF) processes appears in Figure 2.9. The various fractionation streams were systematically coded to designate all the cells (lanes) utilized during gel electrophoresis. Microfil 200.000 MW Tubul l Retentate 0'81) 69 Microfiltration Microfiltration (0.3mieron) 200,000 MW membrane 200,000 MW membrane Tubular Spiral wormd l I I F l Retentate Permeate Retentate Permeate (TRl) (TPl) (SR1) (SP1) A1 ”I A. us 1 l l l Ultrafrltration 50,000MW membrane Spiral wound l I F l |_J—'l l_'__| Permeate Retemate Permeate Retentate Permeate Retentate Permeate Retentate ('l'Rl-PZ) (’l'Rl-RZ) (TPl-PZ) (TPl-RZ) (SR1-P2) (SR1-R2) (SP1-P2) (SP1-R2) Br Cl' Er Fr Bs Cs E: F: FIGURE 2.9 Schematic for Microfiltration and Ultrafiltrations Processes with Fractionation Systems Designated. Fraction Codes and Eletrophoretic Lane Codes are Designated. Diagrams of the electrophonesis gel fractions obtained from UP and MF appears in Figures 2. 10-2. 15. The non-denaturing and non-dissociating (NDND) PAGE analyses of UF and MF appears in Figures 2.10 and 2.13. The NDND-PAGE used in this investigation allows protein separation based on the net negative charge on the protein molecule without protein denaturation. Consequently, this approach depicts separation of proteins in their undenatured, unreduced, and non-dissociated “native” forms. The data for A1 versus D1 indicated hexarner configuration with MF tubular membrane processing of the retentate and permeate streams. This is indicative of protein separation in the native state. Larger molecular weight proteins were concentrated in the retentate than the permeate fractions. The data suggest molecular weight above the urease, hemer star permeate. Concentr and (inner proteins v protein of Phaseolu fraction and can ran and Watt, 1970; D1 111., 19818). If this may very well be r the UP Spiral wor 2.10). These resu native protein cor were associated v the gel. This co Sample. The F ”IpfiSingly high permeate gels of 70 urease, hexamer standard of 545,000 (kd) was found in the retentate but not the permeate. Concentrations below this level were not different and suggests that trimer and dimer proteins were not denatured. One reason may be the presence of the primary protein of Phaseolus vulgn's L, phaseolin. Phaseolin or vicilin is found in the globulin fraction and can range between three to five subunits fi'om 23,000 to 56,000 kd, (Pusztai and Watt, 1970; Derbyshire et al., 1976 a and b; Bolini and Chrispeels, 1978; Brown et al., 1981a). Ifthis is the case, the protein above the urease, trimer standard (272,000 kd), may very well be native phaseolin. It was also determined that no protein was found in the UF spiral wound 50,000 MW separation for permeate gels of B1 and E (Figure 2.10). These results were also reported in Figure 2.13, suggesting no differences in the native protein configurations between UF and MF processing. However, these results were associated with proteins that were fillly eluded or precipitated proteins not entering the gel. This condition could be attributed to non-homogenous mixing of the protein sample. The permeate fiaetion of the 200,000 MW tubular membrane showed surprisingly high concentrations of protein above 200,000 MW. This was found in the permeate gels of Dr and F 1‘- One or more of the several phenomena may explain this observation. Native proteins may in fact re-associate after passing through the membrane pore. The hydrophobic forces, which keep these globulins folded, may temporarily be disrupted during processing and thus allow the protein structure to unfold and then re- associate alter processing. Also, some of these proteins may be somewhat linear in nature, enabling them to pass through undenatured, or they may simply by-pass the membrane altogether and end up in the permeate at the end of processing. Nevertheless, these data 511883“ 1 PAGE electronhon SDS-PAGE 2.12 and 2.15. Thi presence of the d (airfactant), SDS ‘ polypeptide chain chains. ME furtl polypeptide chain, SDS-PAGE Figure 2.11 and disulfide bonds. '1 and Cs versus F s ( PAGE (Tables 2 hnkages This m; found in beans, pa The data 71 these data suggest that proteins are present in the permeate when analyzed with NDND- PAGE electrophoresis. SDS-PAGE with added Mercaptoethanol (+ME) analyses are shown in Figures 2.12 and 2.15. This method is utilized to estimate the size of the polypeptide chain in the presence of the detergent sodium dodecyl sulfate (SD S). As a surface active agent (surfactant), SDS binds to proteins, disrupts their structure and shape, disassociates into polypeptide chains and imposes comparable shapes and net charge densities on the chains. ME further enhances polypeptide size by attacking the sulfide bonds on the polypeptide chain, unfolding the protein moiety. SDS-PAGE without added mercapteothanol (-ME) analyses are presented in Figure 2.11 and 3.14. These data suggest there is a relatively low concentration of disulfide bonds. These major differences in retentate and permeate fractions As versus D3 and Cs versus F3 due to spiral wound membrane comparison of non-reduced and reduced PAGE (Tables 2.14 and 2.15) demonstrated low concentration of disulfide cross- linkages. This may be due to the relatively low levels of sulfur containing amino acids found in beans, particularly cysteine. The data in Figures 2.12 and 2.15 suggest some difl'erences in protein characteristics between tubular and spiral wound membranes. Comparison of gels AT versus D1 indicated not much difference in protein quality among all bands. However, when bands As versus Ds were compared, there was a clear difference in protein quantity. The permeate fiom the first MF showed less protein than the permeate fiom the tubular UF. These data suggest that the pore size of the spiral wound membrane (0. 3 micron) is smaller than the 200,000 kd MW cutofi‘ on the tubular membrane. This could indicate that the sp membrane; conseql processing is desire indicate that greate (50 kd) than the tu generating highly g fractionation utilizir Gel bands of BT , second UF membrg lower molecular v“ 72 indicate that the spiral wound membrane is more selective during MF than the tubular membrane; consequently, the spiral wound is the better membrane to select if MF processing is desired (although as mentioned earlier, fouling may be a limitation). Results indicate that greater protein fractionation and difl'erentiation occurred using spiral wound (50 kd) than the tubular membrane (200 kd). Thus, these conditions could be used for generating highly specific proteins. The permeate was also analyzed for the second fiactionation utilizing UF processing with a spiral wound a 50,000 MW cut-ofl‘ membrane. Gel bands of B1 , ET , B3 and Es showed a lesser amount of protein in all cells. The second UF membrane separated out more protein but less than the 50,000 MW due to its lower molecular weight cut-off. l 73 545K 545K 272 272 136 . 136 66 ii" 66 45 45 29 29 M AT BT CT DT E, 1:T M FIGURE 2.10: N on-denaturing and Non-disassociating PAGE Analysis of Bean Protein Fractionated Using a Tubular Microfiltration Membrane System. Sample Code Designation M: Standard AT: TR1;Retentate of MP with tubular membrane (200,000 MW) B1: TRl-P2; Permeate of AT passed through UF membrane (50,000 MW) CT: TRl-RZ; Retentate of AT retained by UF membrane (50,000 MW) Dr: TPl; Permeate of UP with tubular membrane (200,000 MW) ET: TPl-P2; Permeate of D1 passed through UF membrane (50,000 MW) FT: TPl-R2; Retentate of Dr retained by UF membrane (50,000 MW) 74 M AT 13T CT DT Er FT M FIGURE 2.11 SDS-PAGE (-ME) Analysis of Bean Protein Fractionated Using a Tubular Microfiltration Membrane System in the Absence of Mercaptoethanol. Sample Code Designation M: Standard AT: TR1;Retentate of ME with tubular membrane (200,000 MW) B1: TRl-P2; Permeate of AT passed through UF membrane (50,000 MW) CT: TRl-RZ; Retentate of AT retained by UF membrane (50,000 MW) D1: TPl; Permeate of UP with tubular membrane (200,000 MW) ET: TPl-PZ; Permeate of D1 passed through UF membrane (50,000 MW) FT: TPl-R2; Retentate of DT retained by UF membrane (50,000 MW) 75 205K 116 ass AT BT CT DT 15T FT M FIGURE 2.12 SDS-PAGE (+ME) Analysis of Bean Protein Fractionated Using a Tubular Microfiltration Membrane System in the Presence of 2% Mercaptoethanol. Sample Code Designation M: Standard AT: TRl; Retentate of MP with tubular membrane (200,000 MW) B1: TRl-P2; Permeate of AT passed through UF membrane (50,000 MW) C1: TRl -R2; Retentate of A1- retained by UF membrane (50,000 MW) D1: TPl; Permeate of UF with tubular membrane (200,000 MW) ET: TPl-PZ; Permeate ofDT passed through UF membrane (50,000 MW) FT: TPl-R2; Retentate of DT retained by UF membrane (50,000 MW) 76 M As BS C3 D3 Es Fs M FIGURE 2.13 Non-denaturing and Non-dissociating PAGE Analysis of Bean Protein Fractionated Using a Spiral Wound Microfiltration Membrane System. Sample Code Designation M: Standard As: SR1; Retentate of MP with spiral wound membrane (pore size 0.3 micron, 200,000 MW) Bs: SRl-P2; Permeate of As passed through UF membrane (50,000 MW) Cs: SRl-R2; Retentate of As retained by UF membrane (50,000 MW) D3: SP1; Permeate of UF with spiral wound membrane (pore size 0.3 micron, 200,000 MW) Es: SP1-P2; Permeate of D3 passed through UF membrane (50,000 MW) Fs: SPl-R2; Retentate of D3 retained by UF membrane (50,000 MW) l 77 FIGURE 2.14 SDS-PAGE (-ME) Analysis of Bean Protein Fractionated Using a Spiral Wound Microfiltration Membrane System in the Absence of Mercaptoethanol. Sample Code Designation M: Standard As: SR1; Retentate of MF with spiral wound membrane (pore size 0.3 micron, 200,000 MW) BS: SR1-P2; Permeate of As passed through UF membrane (50,000 MW) Cs: SRl-R2; Retentate of As retained by UF membrane (50,000 MW) Ds: SP1; Permeate of UF with spiral wound membrane (pore size 0.3 micron, 200,000 MW) Es: SP1-P2; Permeate of Ds passed through UF membrane (50,000 MW) F3: SP1-R2; Retentate of D3 retained by UF membrane (50,000 MW) 20% H61 ,1 78 R... 205K ._. 205K 15 1161 _I 116 ‘———' 24 97 l ‘h—_: i; —l 97 84 ‘—' 84 66 66 55 l I 55 45 45 36 36 29 ‘:I 29 24 _. 24 1=S M FIGURE 2.15 SDS-PAGE (+ME) Analysis of Bean Protein Fractionated Using Spiral Wound Microfiltration Membrane System in the Presence of 2% Mercaptoethanol. Sample Code Designation M: Standard As: SR1; Retentate of MF with spiral wound membrane (pore size 0.3 micron, 200,000 MW) B3: SRl-P2; Permeate of As passed through UF membrane (50,000 MW) Cs: SR1-R2; Retentate of As retained by UF membrane (50,000 MW) D3: SP1; Permeate of UP with spiral wound membrane (pore size 0.3 micron, 200,000 MW) Es: SPl-PZ; Permeate of Ds passed through UF membrane (50,000 MW) F3: SPl-R2; Retentate of D3 retained by UF membrane (50,000 MW) Amino Acids: The amin (Phaseolus vulgar units: g/ 1008, mg MT tubular memb MW membrane 3 (TRI) stream we yielded retentate ( acids compared t amino acids versu: and TRl-Rz 1.68 indicated that am] The permeate frac and Were below 0. Whole bean indicat m8'100g) had the ; Wonine (1202 which is liable to ac (”‘548/100g) 79 Amino Acids: The amino acid composition (18 essential and non-essential) of whole beans (Phaseolus vulgaris L) appears in Table 2.6. These data are expressed in three different units: g/ 100g, mg/ 100g and mg/g of protein. The whole bean is compared to fractions of MP tubular membrane and subsequent UF streams. Tubular separation through a 200,000 MW membrane yielded a retentate (TRl) and permeate (TPl) stream. The retentate (TRI) stream was firrther fractionated using the UF spiral wound membrane, which yielded retentate (TR1-R2) and permeate (TR1-P2) fractions. The total amount of amino acids compared to total protein content was as follows: whole bean 18.93 g/ 100g of amino acids versus 20.8% protein; TR] 1.73 g/ 100g of amino acids versus 1.91% protein; and TR] -R2 1.68 g/ 100g of amino acids versus 1.99% protein. These comparisons indicated that amino acid levels were consistent with total protein levels in each fraction. The permeate fi'actions TPl and TRl-PZ had low protein levels (0.40 and 0.34 g/ 100g) and were below 0.05 g/ 100g of amino acid for each amino acid analyzed. Analysis of the whole bean indicated that the acid class of glutamic (146.15 g/100g) and aspartic (112.50 mg/ 100g) had the highest concentration. As expected, the sulfur containing amino acids, methionine (12.02 g/100g) and cysteine (11.06 g/100g), were the lowest. Tryptophan, which is liable to acid hydrolysis, was also quantified and found to be present in low levels (11.54 g/lOOg). ho when :0 35: 2: ES: :1 -7 .. ~ SF: 3.42%.; an: «5 2E» QC <2 <2 8.8 8.8 8 88 <2 <2 8.8 8.8 8 8:: 8.8 2:: :8 658:0 <2 <2 8.8 8.8: 28 8:: <2 <2 8.8 8.8: 88 8:: 2.8: 88 88 2585 <2 <2 8.8 8.8 8 8:. <2 <2 88 8.8 8 88 8.8 8:: 8.8 65:85. <2 <2 8.8 8.8 o: :8 <2 <2 8.8 8.8 8: :8 8.8 8: 8.: 8:93 <2 <2 88 8.8 8 8:: <2 <2 8.8 8.8 8: 28 8.8 8:: 2.: 88:82:: <2 <2 8.8 8.8 8: :8 <2 <2 8.8 8.8 8: :8 8.8 82 8.: 8:86.: <2 <2 8.8 8.8: 8 8:: <2 <2 8.8 8.2 8 8.8 8.: 8m 8:: 8:82:62 <2 <2 8.8 8.8 8 8:. <2 <2 8.8 8.8 8 8:: 8.8 8:: 8:: 65:82 <2 <2 8.8 8.8 8 8:: <2 <2 8.8 8.8 8 8:: 8.8 o; :8 8:2,: <2 <2 8.8 8.8 8: 2.8 <2 <2 8.8 8.8 c: :8 8.8 8: 8:: 8:8 <2 <2 8.8 8.8: o: :2: <2 <2 8.8 :8: 8m 88 8.8: 88 88 6:532 see 8: a 8: so: 8: a 8: 522 8: a 8: 52% woeme m8:m so: 8: a 8: mo 3mm: mo 3?: mo MES no Ema mo mun: FE 8:25: 48.35 SE Bo< oz§< a): 88.88 83:33: .8: 8.8 2888:: a): 88.8 ":2 8.88 g 83:86.: N: .5 o8.8~ m .5 o8.o8 33:33: 89:95:52 6553 .3QO ES .3155. ”:35: stab—win m2 «5:333 can ”383?:— mi 3.. 83?:— Eu< 95:2: 6." 93:. The TR] an data but at reduced and .22g/100g) and were low for TR] ( The calculat in Table 2.7. Tabl. Whole bean and rett 81 The TR] and TRl-R2 fractions indicated consistent results with the whole bean data but at reduced concentrations. Glutarnic and aspartic acids were highest for TR] (.30 and .22g/100g) and TRl-R2 (.28 and .21g/100g), respectively. Methionine and cysteine were low for TR] (.02 g/100g) and TRl-R2 (.02 and .05 g/ 100g). The calculation of amino acid scores was conducted, and the results are presented in Table 2.7. Table 2.8 displays the amino acid scores for children (2 years old) of the whole bean and retentate fractions obtained in Process Trial H. TABLE 2.7 Sum year. Heal Calc MF. Amino Acid Refcre Protc Panel mtg/ks Methionine 27 & Qste'me Phenihlfllline 6g & TlTOSine Lime 6.. ”mm“ 3'. Van” 31 Marine 3 . Tn ' 8&1 l 2 x l Rsteamed {mm Rem TABLE 2.7 82 Summary of Estimates of Amino Acid Requirements for Children (12 years old), as Proposed in 1985 by Food and Agriculture, World Health Organization, United Nations, Protein Content and a Score Calculation for Whole Bean, the Retentate and Permeate for UF & MF. Whole Bean Retentate 200,000 MF R 200,000 MF R Permeate Amino Acrd Reference 200,000 UF 50,000 MF 50,000 MF 200,000 UF Protein (TRl) Permeate Retentate (TPl) Pattern ‘ ('I'Rl-PZ) (TRl-R2) mg/kg/day mg/g of Score mg/g of Score mg/g of Score mg/g of Score mg/g of Score protein protein protein protein protein Methionine 27 23.12 0.86 20.94 0.78 NA NA 35.18 1.30 NA NA & Cysteine Leueine 73 77.88 1.07 73.30 1.00 NA NA 65.33 0.89 NA NA Phenylalanine 69 83.65 1.21 89.01 1.29 NA NA 75.38 1.09 NA NA & Tyrosine Lysine 64 64.90 1.01 62.83 0.98 NA NA 55.28 0.89 NA NA Threonine 37 44.23 1.20 41.88 1. 13 NA NA 40.00 1.08 NA NA Valine 38 47.60 1.25 47.12 1.24 NA NA 40.20 1.06 NA NA Isoleucine 31 39.90 1.29 36.65 1.18 NA NA 35.18 1.13 NA NA Tryptophan 12.5 11.54 0.92 10.47 0.84 NA NA 25.13 2.01 NA NA ‘ Reprinted from Recommended Daily Allowance: 10th Edition, Copyright 1989 by The National Academy of Sciences. Courtesy of the National Academy Press, Washington, DC. Amino Acid WT B Methionine & ( Cysteine Tomorhan ( Lysine - lercine Thiamine Pllfllylillanine & Tyrosine Valine [501mm \ Quantitatlv e TABLE 2.8 Amino Acid Score Calculations in Ascending Order for the Amino Acids Required by 2 year-old Children. Scores for Whole Bean, the 200,000 MW UF and Diaf'rltration Through a 50,000 MW, MF Retentate. 200,000 MF Amino Acid Whole Amino Acid 200,000 MF Amino Acid Retentate Bean Retentate 50,000 UF (TRI) Retentate (IRl-Rz) Methionine & 0.86 Methionine & 0.78 Lysine 0.89 Cysteine Cysteine Tryptophan 0.92 Tryptophan 0.84 Leucine 0.89 Lysine 1.01 Lysine 0.98 Valine l .06 Leucine 1.07 Leucine l .00 Threonine l .08 Threonine 1.20 Threonine 1.13 Phenylalanine & 1.10 Tyrosine Pharylalanine & 1.22 lsoleueine 1.18 lsoleueine 1.13 Tyrosine Valine 1.25 Valine 1.24 Methionine & 1.30 Cysteine lsoleueine 1.29 Phenylalanine & 1.29 Tryptophan 2.01 Tyrosine Quantitative results are presented in ascending order in each column. As expected, the methionine and cysteine (.86 and .78) were the most limiting for both the whole bean and retentate fractions. These results are consistent with other research conducted where they were also assessed to be the limiting amino acid in the retentate with a score of .76. However, recycle using UF with a 50,000MW spiral wound membrane, TR] -R2, yielded different results. In this case, lysine and leucine were the most limiting amino acids. These amino acids associated with low molecular weight polypeptides which may have passed through the membrane and been concentrated in the permeate; however, no major levels above .05g/100g were found in the permeate to support this theory. It would be valuable to assess the position of these limiting amino acids within specific peptides and determine the p to enhance the n of highly specific THERMAL PR A method “'35 Performed or I0 99°C for a pen included as an abu shows the results f agglomeration of l bOdies were observ Partially, as the te 84 determine the potential membrane selectivity for the enriched peptides that could be used to enhance the nutritional quality of bean products through the addition of small quantities of highly specific peptide fiactions. THERMAL PROCESS SIMULATION A method to simulate actual thermal process requirements during pasteurization was performed on the bean protein retentate fractions. The heating range was fi'om 71 °C to 99°C for a period of 5 seconds. A negative control of 99°C for 30 seconds was also included as an abusive variable. Qualitatively this can be observed in Figure 2.16, which shows the results from the Confocal Scanning Microscopy. These data suggest that some agglomeration of the proteins occurred as the temperature increased. Larger protein bodies were observed during heating when compared to the non-heated raw material. In particular, as the temperature reached 99°C, the agglomeration was more evident. The negative control (99°C/3O sec) showed larger and more agglomerated protein bodies than the highest heat simulated treatment (99°C/5 sec). However, since these data provide only a qualitative measure, they can only be used as directional information. More quantitative measures are required to draw decisional conclusions. Therefore, gel electrophoresis and nitrogen solubility index tests were performed to provide quantitative results useful to assess the denaturation phenomenon. . p -. 1: ‘La ow? mt! .‘t.. . .. uh mummified... Hi FIGURE 2.16 Photographs of Fractionated Bean Protein Heat Treated at 71.1°C - 983°C for 5 to 30 Seconds 3) 71.1°C / 5 sec. h) 822°C IS sec. c) 93.3°C / 5 sec. d) 983°C / 5 sec. e) raw f) 98.8°C / 30 sec. In F igur differential heat protein occurret Two the fractio aseptic process 1 30 seconds. The denaturation occ Protein denatura fiactionate Speci: overall protein qL 86 In Figure 2.17, the NDND-PAGE analysis of bean protein fi'actions subjected to difl'erential heating is presented. The data indicated that no appreciable denaturation of protein occurred within normal pasteurization conditions. This is important since in Study Two the fractionated protein was heated at 118°C for 4.6 seconds qualified during the aseptic process and package trials. In this case, proteins were not disrupted until 99°C for 30 seconds. These data support the results obtained in Study Two that little to no protein denaturation occurred with the processing of the bean protein fortified shake. The lack of protein denaturation is important since separation technology not only was utilized to fractionate specific proteins but also was utilized as a low-heat method for improving overall protein quality. FIGU SaInplt 87 545K FIGURE 2.17 Non-denaturing and Non-dissociating PAGE Analysis of Bean Protein Fraction (TRl-RZ) Subjected to Differential Heating. Sample Code Designation M: Standard A: TRl-R2, raw B: TRl-RZ, 160°F/5 sec C: TRl-RZ, 180°F/5 sec D: TRl-RZ, 200°F/5 sec E: TRl-RZ, 210°F/5 sec F: TRl-RZ, 210°F/3O sec 88 Table 2.9 shows the percent of solubility for the retentate processed through UF (200,000 MW) tubular membranes and then diafiltered with MF (50,000 MW) spiral wound membranes. These data indicated little loss in solubility among all heat treatments. TABLE 2.9: Percent Water Soluble Protein for Retentate Which Was Processed Utilizing Ultrafiltration (200,000 MW) Tubular Membrane and then Diafiltered with Microfiltration (50,000 MW) Spiral Wound Membrane. Treatment‘ % Solubility so TR], Raw 102.7 4.1 TR], 71°C /5 seconds 103.3 1.] TR], 82°C /5 seconds 102.4 3.3 TR], 93°C /5 seconds 101.7 6.0 TR], 99°C /5 seconds 102.3 3.5 TR], 99°C /30 seconds 100.9 3.8 TR2, Raw 97.6 2.4 TR2, 71°C /5 seconds 100.] 0.7 TR2, 82°C /5 seconds 96.0 3.3 TR2, 93°C /5 seconds 96.3 0.7 TR2, 99°C /5 seconds 92.4 2.1 TR2, 99°C /30 seconds 98.4 1.8 1n=2 for all treatments, mean values and standard deviations (SD) These data provide quantitative support for the findings obtained with confocal scanning microscopy and NDND-PAGE that showed limited protein denaturation occurred during thermal processing. 89 Conclusions These results indicated membrane based separation technologies could be utilized to fractionate and isolate legume protein. These findings indicated that in a comparison between spiral wound and tubular membranes, tubular is the preferred system for dry bean protein fi’actionation. Tubular designs foul less, have faster rates of separation, are easier to operate and are generally more efiicient to clean and maintain. These benefits are applicable not only to research programs but also to commercial systems where operating and preventive maintenance costs along with higher yields would lead to a more economic usage of assets for industrial applications. Separation of native proteins from whole beans of Phaseolus vulgaris L. was also demonstrated to be feasible with microfiltration and ultrafiltration systems. Isolation of large protein agregated greater than 545 (kd) was observed in the retentate fi'actions. The ability to isolate proteins in their native state may be of importance not only in quantification of specific proteins, including fimctional property retention, but also in the overall nutritional quality of the proteins. Results of the experiment indicated that bean protein fiactions underwent limited denaturation when subjected to simulated thermal processing conditions. Protein quality was determined by utilizing both qualitative and quantitative methods. Qualitatively, Concofocal Scanning Microscopy indicated limited protein agglomeration when subjected to heating at 99° C for 5 seconds. Quantitatively, no significant changes in solubility were observed with NSI measurements among all heat treatments tested. Limited protein denaturation was also detected using NDND-PAGE analyses. Amino acid scores showed 90 that the limiting amino acids in the retentate were methionine and cysteine, which are the sulfur containing amino acids. These findings may be of importance to the development of specific fortified or the so-called “nutraceutical products.” These products deliver to consumers specific health benefits associated with a targeted nutrient, using foods and beverages as the vehicles of delivery. The ability to isolate proteins or specific peptides in their native state and formulate products that deliver necessary taste, convenience and nutrition to consumers will be paramount in the firture of new-product research. Further research needs to be conducted to optimize legume protein separation using membrane technology. This work needs to be coupled with in-depth investigation of targeted protein masses, peptides or amino acids that may have potential consumer acceptance and health benefits. Together, food technology and nutrition will drive the development of new consumer products into the twenty-first century providing industry with higher profit margin brands. This will give consumers healthier and better-tasting products and will improve the quality of the overall diet. STUDY TWO Protein Separation and Formulation of a Fortified Beverage Suitable for Small Children Introduction Ultrafiltration processing is a technology that can be utilized as a continuous method to separate or fractionate materials based on molecular weight. This technology has been employed extensively by the dairy industry over the last 30 years to fractionate and concentrate components of whole milk and whey. Application to vegetable proteins has been primarily focused on soya (gylemic species). The use of dry beans (Phasest vulgaris L.) has not been widely practiced. In this study, dry edible beans were used as the protein source for development of a nutritionally fortified beverage. Milled beans were passed through a series of 200,000 (kd) molecular weight polysulfone membranes to selectively concentrate proteins. The protein fiom the retentate was added to the beverage and aseptically processed in a 200m] Tetra Slim® package. The product design criterion were driven by a number of external factors. These criterion were followed during formulation and processing to determine product fit versus design criteria. This technique is called “Quality Functional Deployment” (QFD) and represents the best approach for defining and externally measuring quality during the design and development phases of new product introduction. This process was followed in this study of the development of a fortified beverage, suitable for small children utilizing protein fractions separated through ultrafiltration. 91 92 Materials And Methods PROCESS AND PRODUCT DEVELOPMENT Ultrafiltration: Dry navy beans (Seafarer cv.) were hammer milled (13.9% moisture) to produce a flour (particle size of 700nm) using a standard Fitzrnill (1/8” screen). The bean flour was soaked (4:1 tap water) for 24 hours at 4°C in a 1.7% solution of NaCl in which the pH was adjusted to 9.0 using 3N sodium hydroxide. The extracted supernatant was decanted and adjusted to pH 7 with 3N HCL and filtered to remove suspended solids and produce a clear solution. This filtered material was processed through a series of ultrafiltration tubular-type polysulfone membranes in a pilot scale system (Table 3.1). TABLE 3.]: UF Equipment and Tubular Specifications Used for Navy Bean Protein Fraction in Trial 1 Ultrafiltration System Equipment: APV Crepaco, Towanda, NY APV Crepaco Equipment Specification for BRO/BUF Membrane Filtration Pilot Unit, Serial No.2]202 Specifications UF Configuration 9.3 fl 2 (0.9 m2) Membrane Area Two PCI type B] modules having 9.3 ft 2 membrane area/B1 module B] module type heat exchanger with stainless steel shroud (2 feet long) Feed Pump Capacity: 7.5 gpm at 100 psi (UF) Hold-Up Volumes: Basic Unit (excluding modules) 1.32 gallons (estimated) Tube Side - .8 gallons (estimated) Permeate Side - 2.4 gallons (estimate) 93 Separation through a 200,000 kd molecular weight membrane resulted in recovery of two fraction streams: (1) permeate I and (2) retentate I. The initial permeate fiaction (permeate I) was passed through the same membrane a second time, yielding two streams designated (1) permeate H and (2) retentate II. Retentate H was used for formulation of the final product. A detailed process flow and mass balance was prepared and is presented in Figures 3.1 and 3.2. Whole Bean Protein - 19.08% F Fitzrnill, 1/8” I Extraction Soak (4: l): 24 hours @ 4 °C 1.7% NaCl Solmion, pH 9 I I l _ Supemate 11 Residue +0.8 parts 3m HCl, pH 7 I UF Separator 200,000 MW Membrane l P j Retentate J Permeate I Diafiltration 200,000 MW Membrane l I l Retentate II Permeate 11 FIGURE 3.]: Dry Navy Bean Protein Extraction and UF Fractionation with 200,000 MW Tubular Membrane. 94 Batch Recyde wC) :- -:’ 3 -:- :--:v :- -:- as: ': a: Tubular Membrane Pump 4+0 :3: 3;:1: 25:13; 2.2343138— +Cj 3;: iii 3;: 3:: ‘EE‘E 1:: 1' ”f l Permeate Retentate Collection Collection FIGURE 3.2: Ultrafiltration Batch System Utilizing Three 200,000 MW Tubular Membranes in a Series of Filtration. Product Formulation, Processing and Packaging: Formulation utilizing protein separated concentrate was achieved at pilot processing facilities maintained by the White Knight Corporation, Grand Rapids, MI. The protocol used in the peach shake formulations trial is outlined in Figure 3 .3. FIGUl 95 Blend Tank 700 It; (170 911) “Peach Shake Formulation Base pins 45.5 kg BeanP’l-otein Fractioncontalning036kgProtetnblended for 15 minutes 1 Balance Tank l Pre-Heat IM°F (823°C) J Stea- Injection 245°]? (118.4°C) I Hold Tube 4.6 Sec Vacuu- FIash Cool 186°F (816°C) 1 Homogentze 500 to 2500 psi L Regeneration Cooling 90-100°F (GB-73°C) Final Cooling ~ 70°F(2]°C) 1 I Aseptic Tank ’ l AseptieFlller FIGURE 3.3: Standard Peach Shake and Peach Shake Fortified with UF Separated Bean Protein Fraction Produced in Aseptic Tetra Slim® Packaging at White Knight Corporation in Grand Rapids, MI: Blending, Processing and Packaging Flow Chart. Bean protein retentate was directly substituted for dairy base fraction on a 1:] basis. Approximately 45.5 kg of bean protein retentate containing 0.96 kg of protein was added to a 700 kg peach shake base. The formulation was comprised of low fat yogurt, sugar, peach concentrate, and other natural flavors (W GNP) in a proprietary blend. This mixture of ingredients was blended for 15 nrinutes and transferred to a holding tank prior secor male; homt asept Slim: lhree wnfi place and: “till; merh 96 to processing and packaging. The protein fortified peach shake was pumped from the holding tank and pre-heated to 180°F (823°C) and pumped to a tubular (shell-in—tube, Alfa Laval, 150 gal/pm) pasteurizer. The material was heated to 245°F (118.4°C)/4.6 seconds and immediately flash cooled under vacuum to 186°F (856°C). The cooled material was homogenized at 2500 psi using an Alfa-Laval Homogenizer. The homogenized material was firrther cooled to 70°F (21°C) and transferred to a 500 gallon aseptic holding tank. The peach shake was subsequently aseptically transferred to a Tetra Slim® filling machine which aseptically filled 200ml of the product in a vertical-form-fill- seal paperboard package. Individual packages were passed through a shrink tunnel where three cartons were wrapped together and placed in a corrugated tray pack of 3x9 configuration yielding twenty-seven 200ml cartons per case. Cases were palletized and placed in an ambient (80°F [267°C] /80% RH) warehouse for storage prior to analytical and sensory analyses. QUALITATIVE AND QUALITATIVE ANALYSES Proximate Composition and Product Characteristics: Total Solids: For moisture content, the AVAL (1984) vacuum oven method was utilized. Eat: The soxhlet extraction procedure was used in the fat determination (AACC method 30—25). Fat% = (wg’ght of dish + fat) - (weight of dish) x 100 weight of sample Protein: Dry matter content was predetermined by drying appropriately sized samples (5 to 100g) in a forced air oven at 105°C to constant weight. Nitrogen content of 97 the dried sample was determined by an automated Kj eldahl method using the Kjel-Foss Automatic Model 16210 (AOAC method 7.02 1, 1984) and crude protein content was obtained by multiplying by the factor 6.25. fl: The dried samples obtained from the moisture determinations were incinerated at 525°C for 24 hours in a Barber-Coleman mufile firmace. The uniform white ash was cooled to room temperature in a desiccator prior to weighing (AACC method 08- 0]). Ash% = (weight of dishfi-Lash) - (weight of dis_h) x 100 (weight of dish + sample) - (weight of dish) Carbohydrate: Calculated by difference which equals % Total Solids = (°/o Fat + % Protein + % Ash). Total Diet—ary Fiber: Duplicate samples of dried foods, fat-extracted if containing >10% fat, were gelatinized with Termamyl (heat-stable a-amylase) and then enzymatically digested with protease and amyloglucosidase to remove protein and starch. (When analyzing mixed diets, always extract fat prior to determining total dietary fiber.) Four volumes of ethyl alcohol were added to precipitate soluble dietary fiber. The residue was filtered, washed with 78% ethyl alcohol, 95% ethyl alcohol, and acetone. After drying, the residue was weighed. One sample was analyzed for protein, and other was incinerated at 525°C and ash was analyzed. The formula was total dietary fiber = weight residue - weight (protein + ash), (Oficial method of analysis AOAC [1996] 985.29). m: Soluble solids (%) were determined with a ABBE Refractometer, Bausch & Lomb, United States. l Pittsburr within 1' Nutrien determir was (lire [30°13], this san Farming the Ge boiling lial wh Sleps v Progre: 98 pit A Fisher Accumet pH Meter, model 830 (Fisher Scientific Company, Pittsburg, Pennsylvania) was used to measure the pH of each sample at room temperature within 12 hours of the sample preparation. Nutrients: Amino Acid: The amino acid composition of each major bean protein fiaction was determined. The isolated bean protein sample (1 .0-1.2mg protein) in a 6 x 50mm tube was directly solubilized in 500p] performic acid (formic acid [88%]: hydrogen peroxide [30%], 9: 1, v/v) and held at 22°C for 1 hour. Following 2 hours of performic oxidation, this sample was then dried using Speed Vac Concentrators (Savant Instruments, Inc. , Farmingdale, NY). Acid hydrolysis and amino acid analysis of the oxidized samples were performed at the Gerber Products Company in Fremont, Michigan. Two hundred u] of constant boiling HC] (Pierce Chemical Co., Rockford, IL) was added to the bottom of the vacuum vial which contained 12 samples per hydrolysis. Three alternate vacuum-nitrogen flushing steps were required to ensure an oxygen-free atmosphere. Vapor phasehydrolysis was progressed at 112-116°C for 24 hours. Following hydrolysis, the samples were re-dried with ethanolzwaterztriethylamine (2:2: 1, v/v), and derivatized. The derivatization reagent consisted of a 7: 1 :1 solution of ethanol, triethylamine, water, and phenylisothiocyanate (PITC). After 10 minutes of derivatization, the samples were vacuum dried and re- dissolved in 500p] of 5mM sodium phosphate, pH 7.8. An injection volume of 40p] was used. A Waters 600 HPLC Systems (Waters Associates, Milford, MA) was used for the analysis of the derivatized amino acids. Separation was accomplished using a PicoTag reverse column (Water Associates) and a gradient mobile phase system (A, 15mM sodium 99 acetate bufl‘er pH 5.9 and 0.0% triethylamine, B, acetonitrile: H20 (60:40). Ultraviolet absorbence at 254mm was used for detection of the PITC-labeled amino acids. Vitamins: Marotene (Vitamin A) - Spectrophotometric Method : Place 2-5 g weighed sample in high-speed blender; add 40 mL acetone, 60 mL hexane, and 0,] g MgCOa , and blend 5 min. Filter with suction or let residue settle and decant into separator. Wash residue with two 25 mL portions acetone, then with 25 mL hexane, and combine extracts. Wash acetone from extract with five 100mL portions H20, transfer upper layer to 100 mL volumetric flask containing 9 mL acetone, and dilute to volume with hexane. If desired, alcohol may be used instead of acetone for extraction. Use 80 mL alcohol and 60 mL hexane in blender, other volumes same as for acetone. Pack activated magnesia-diatomaceous earth mixture (1 + 1) in chromatographic tube 22 mm od x 175 mm sealed to 10 mm od tube at bottom. To prepare column, place small glass wool or cotton plug inside tube, add loose adsorbent to 15 cm depth, attach tube to suction flask, and apply firll vacuum of H20 pump. Use flat instrument (such as inverted cork mounted on rod or tarnping rod) to gently press adsorbent and flatten surface (packed column should be ca 10 cm deep). Place 1 cm layer anhydrous Na2SO4 above adsorbent. Mth vacuum continuously applied to flask, pour extract into column. Use 50 mL acetone-hexane (1 + 9), or slightly more, if necessary, to develop chromatogram and wash visible carotene through adsorbent. Keep top of column covered with layer of solvent during entire operation (conventionally done by clamping inverted volumetric flask full of solvent above column with neck 1 to 2 cm above surface of adsorbent). 100 Transfer eluate, which has been reduced in volume by loss of vapor through H20 pump, to 100 mL volumetric flask, diluted to volume with acetonehexane (1 + 9), and determine carotene content photometrically. Determine absorbance of solution as soon as possible with spectrophotometer at 436 nm or with instrument having suitable filter system, such as Klett photo-meter with No. 44 filter, or Evelyn photoelectric colorimeter with 44 nm filter. Calibrate these instruments first with solutions of high purity B-carotene as shown by characteristic absorption curve (J. Biol. Chem. 144, 210942)). Prepare calibration charge and convert A of solution to be determined to carotene concentration fiom chart. When determinations are made with properly calibrated spectrophotometer at 436 “m, C+(Ax454)/(]96xLxW) where C = concentration carotene (mg/lb) in original sample, L = cell length in cm, and W = g sample/mL final dilution. Report results as mg B-carotene/lb. Multiply by 2.2 to give ppm or by 1667 to give International Units/lb. Oficial Methods of Analysis AOAC (1993) 941.15 Ascorbic Acid (Vitar_nin C) - Fluorescence Method: Ascorbic acid and dehydroascorbic acid were extracted from the sample with a solution of metaphosphoric acid/methanol/water. Norit was added to remove colored interferences from the sample extracts and to oxidize ascorbic acid to dehydroascorbic acid in the samples and standards prior to filtration of the extracts. Using an Alpkem RFA—300 System, aliquots of the sample solutions and standards were buffered and reacted with o-phenylenediamine, forming quinoxaline-l—one, a fluorescent condensation product of dehydroascorbic acid and o-phenylenediamine. Background or interfering fluorescence was measured by reacting aliquots of the sample and standard solutions with boric acid, which complexed 101 dehydroascorbic acid, preventing the condensation with o-phenylenediamine. Sample and standard fluorescence values were corrected for background interference, and the samples were quantitiated from the set of standards of known concentration. Official Methods of Analysis AOAC (1995), 16" Ed, Method 984.25, Locator #45116. Thiamin: (Vitamin B 1_) - Ma_n_t_la1 Fluorescence Method. Free thiamin was extracted by autoclaving with a diluted hydrochloric acid solution; bound thiamine was released by incubating overnight in an enzyme solution. Excess protein was precipitated by a pH adjustment and filtration. An aliquot of sample solution was phased with isobutyl alcohol to remove fluorescent interferences. The thiamin extract was oxidized with alkaline potassium ferricyanide to form thiochrome, which was measured fluorometrically and quantitiated fi'om a set of standards of known concentration that are taken through the procedure fiom the oxidation step. Oflicial Methods of Analysis of the AOAC, (1995) 16th Ed, Method 942.23, Locator #45.1.os (Modified). Riboflavin (Vitamin B2) - Fluoremetric Method: An accurately weighed portion (1.5 g maximum) of the sample containing ca 10 pg riboflavin was placed into a 100m] amber volumetric flask and solubilized in 50 ml., 0.1M HCl, washing down sides of flask and dispersing sample. Flasks were covered with foil and autoclaved 30 min at 121°C and air cooled to room temperature (solutions may be stored at this point). Add pre- determined aliquot of 1.25M NaOAc with swirling. Adjust unit so that when aliquot of 1.25M NaCH3C00 solution is added to 50.0 m] 0.1M HC], pH is 4.3 i 0.] (ea 6.0 mL). Add 35ml pH 4.3 HPO; bufl‘er. Check pH with meter and adjust any sample hydrolyzate differing from standards by 10.]pH unit, incubate overnight at 37°C before diluting to volume. Enzyme hydrolysis will not affect riboflavin determination, but enzyme solution 102 must also be added to standards because enzyme contains small amount of riboflavin. Dilute solutions to volume with pH 4.3 HP03 buffer solution; add drop of wetting agent and filter through glass fiber paper. Pump high standards solution (0.15 ug/mL), through system and set recorder pen at 100% with standard calibration adjustment on fluorometer. Aspirate and pump set of standards and sample filtrates through system. Use one 0.10 ug/mL standard with every series of 20 samples to correct for any drift. Ifsarnple is more concentrated than highest standard, dilute with wash solution to bring peak height into range of standards. After all samples have been run, replace NaCH3C00 with solution with Na2S204. Let run to stable baseline, adjust to original baseline, and re-sample filtrates to obtain corresponding blanks. Standards provide linear standard curve passing through origin. Subtract peak height of blank from peak height of sample and determine concentration, C, from standard curve. Mg Riboflavin/100 g = C x10/W where W = g sample. Ofiicial Methods of Analysis AOAC (1981) 981.15. m: This is an autoturb turbidirnetric microbiological method based on the observation that Lactobacillus plantarium ATCC 8014 requires niacin for growth. A basal medium, nutritionally complete in all respects except for niacin, is used as the diluent for the final dilutions of the standards and samples on the Autoturb. Following incubation, the growth response of the bacterial cultures is measured as percent transmittance on the Autoturb. A dose-response line is constructed and the sample concentrations are calculated. Oflicial Methods of Analysis of the AOAC, (1995) 16til Ed, Method 960.46, Locator #45201 and Method 944.13, Locator #45204 (Modified) 103 Vitamin Ba: This is an Autoturb turbidirnetric microbiological method based on the observation that Saccharomyces uvarum ATCC 9080 requires Vitamin B, for growth. A basal medium, nutritionally complete in all respects except for Vitamin B5, is used as the diluent for the final dilutions of the standards and samples on the Autoturb. Following incubation, the grth response of the bacterial cultures is measured as percent light transmittance on the Autoturb. A dose-response line is constructed and the sample concentrations are calculated. Official Methods of Analysis of the AOAC, (1995) 16‘“ Ed, Method 961.15, Locator #45208, (Modified). M41593: Retentate and permeate samples for the three studied membranes were analyzed for mineral content using an inductively coupled argon plasma (ICP) emission spectrometer (Jamell-Ash Model 955 Atom comp) equipped for simultaneous analysis of 19 elements (Al, As, Ca, Cd, Co, Cr, Cu, Fe, Hg, K, Mg, Mn, Mo, Na, P, Pb, Se, T1 and Zn) (The Department of Pharmacology and Toxicology, Michigan State University). All glassware and Tuff containers were acid washed for mineral analysis. All samples were brought to volume using water purification by a Millipore water purification system Class 1 (Millipore Comp, Bedford, MA). Duplicate samples of freeze-dried and spray-dried retentate and permeate flour were combined with 2ml of concentrated Baker Instant-analyzed nutric acid in 15m] screw-capped teflon vials (TufilTainen, Pierce Chemical Co., Rockford, IL). The samples were incubated at 70 to 7 5°C overnight, cooled, and qualitatively transferred to 10m] Class A volumetric flasks containing 1.0m] of 100ppm trium (internal standard). They were then diluted to a final volume of 10m] with Millipore purified water. Along with bean samples, other procedural flasks containing no sample material and a sample of 104 standard material (SRM) were prepared essentially by the same procedure and served as standards. All samples were then rapidly mixed by refluxing the sample in an acid solution I-INO32HZSO4, 5:1 (v/v) according to the method of Siemen and Brinkley (1981) and were analyzed for their mineral composition and mineral contents reported in ppm on a dry weight basis. PROTEIN CHARACTERIZATION SDS Gel Electrophoresis : Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SD S-PAGE) of the bean protein fractions including albumin, GI and GII was performed on 10% acrylamid running gels with 3% stacking gels using the system of Laemmli (1970). A protein solution of each bean protein was prepared to a final concentrate of 10 mg nitrogen sample per ml of buffer, heated, and mildly vortexed until completely solubilized. A 25 ulof each protein solution sample was applied into the sample well of stacking gel. Electrophoresis was carried out with a Hoeffer Vertical Electrophoresis unit (Model SE 600; Hoefi'er Scientific Instruments, San Francisco, CA) using a constant voltage power supply (Fisher Biotech Electrophoresis System, Model FB 458, Pittsburgh, PA). A constant current of 30mA was applied until the proteins migrated into the mnning gel, and then the current was increased to 60 mA until the bromophenol blue tracking dye reached the bottom of the running gel. The gels were removed and stained for 6 hours in 0.4% Coomassie Blue in 9/45/45 (v/v/v/) acetic acid/methanol/water. The gels were destained using 7/25/67 .5 (v/v/v) acetic acid/methanol/water until clear. The subunit molecular weights of the studied proteins were estimated using a mixture of standard 105 molecular weight protein markers (SDS-7 Dalton mark VII - L/Low Molecular Weight Markers and SDS 6H/High Molecular Weight Markets) purchased from Sigma Chemical Corp, St. Louis, MO. The SDS-7 protein mixture used in preliminary trials consisted of the following proteins: lysozyrne (14.4 kd), soybean trypsin inhibitor (21.5 kd), carbonic anhydrase (31kd), ovalbumin (45kd), bovine albumin (66.2kd), myosin (200.0kd). The SDS-6H protein mixture used to differentiate problems in process Trial 1 contained the following proteins: carbonic anhydrase (29kd), egg albumin (45kd), bovine albumin (66kd), phosphoryl B (97.4kd), li-galactosidase (116kd), and myosin (205kd). The standard protein solutions were prepared according to the method described in Sigma Technical Bulletin No. MW S-877C (Sigma Chemical Corp., St. Louis, MO). The relative mobility (RM) of the protein standards was calculated using this formula: RM = Distamze of Protein Migration (mm) Market Dye Distance (mm) 106 Gel Calibration (15% Acrylamide Gel) Standards MW Migration (mm) Myosin 200000 4.8 B-Galactosidasc 1 16250 1 1.2 Serum Albumin 66200 20.2 Ovalbumin 45000 36.0 Carbonic Anhydrase 31000 56.0 Trypain Inhibitor 21500 73.0 Lysozyme 14400 79.0 Dye Front 132.5 250 Y = 552940 * {05614 -29915 20° 1 I: R2 = 0.997 a; 150 é’ é‘ a a I ‘ g 9 100 - 2 v s‘ . O ' n 2 - ..... 50 ‘ " ' - . . , .. _ _ _ . . _ _ _ ‘ -n 0 I I I I I I Migration (mm) FIGURE 3.4: Standard Curve for SDS Electrophoresis A plot of relative mobility versus molecular weight (Figure 3 .4) was constructed as a standard curve. The relative mobility of each protein subunit was calculated, and the molecular weight was estimated fi'om the standard curve. The protein bands present on the gels were qualitatively compared using a Ehimadzu Dual Wavelength thin-layer Chromato Scanner (Model CS-93 0, Kyoto, Japan). The protein bands were identified by their subunit molecular weights. 107 NLEA COMPOSITIONAL LABEL Detailed quantitative analyses were conducted and a standardized label format prepared using guidelines of The Nutrition Labeling and Education Act (NLEA). Components of the product which were analyzed or calculated include calories, total fat, sodium, total carbohydrate, protein, vitamins A, C, thiamin, riboflavin, B6, niacin and minerals, calcium, iron and zinc. The label panel was prepared as print- ready copy. 108 Results Ed Discussion Separation of protein was accomplished in this experiment by utilizing ultrafiltration technology. This technology uses membranes which are selective and separated based on molecular weight size (kd) to yield two streams: (1) Retentate, which is retained by the membrane and (2) Permeate, which passes through the membrane. The process sequence (Trial 1) including the mass balance for fiaction weights and protein content is outlined in Figure 3.5. 109 Dry Navy Beans (100 LB) Fltzmiil, 1/8” C om ponent Mass (Ihs) (kg) Percent Total 91 .5 41 .590 100.00 Solids 78.69 35.76 86.0 Protein 17.31 7.868 22.0 I Soak (4/1): 24 hours @ 4 degrees C 1.7% NaCl Solution, pH: 9 L J 1 Residue Supernatant 172.5 lls, 365 Proc. 295II1, 17% Protein Adjust pH to 7: Filter Component Massflls) (kg). Percent Component Massah) (kg) Percent Total 295.0 134.090 100.00 Total 172.5 78.409 100.00 Solids 82.6 37.545 28.0 Solids 9.76 4.436 5.56 Protein 14.04 6.382 17.0 Protein 3.51 1.595 36.0 1 UP Separator 200.000 MW Membrane l 1 fi Retentate I (126.5 lb) Permeate I (27Ih) 42% Protein 14% Protein Component Massflh) (kg) Percent Component Mnssflls) (kg) Percent Total 126.5 57.50 100.0 Total 27.0 12.27 100.0 Solids 8.72 3.96 6.9 Solids 0.405 0.18 1.50 Protein 3.66 1.66 42.0 Protein 0.0567 0.03 14.0 UF Separator: 200,000 MW Membrane Occasional spray to snrpress foam 1 I I Retentate II (10911:) Pemnte "(5915) 47% Protein 12y. Pro“... Component Mass (111) (kg) Percent Component Mass m.) (1“) Percent Total 109 49.55 100.0 Total 59 26.82 100.0 Solids 4-86 2-21 44‘ Solids 1.03 0.47 1.74 Protein 2-28 1-04 47-0 Protein 0.123 0.06 12.0 FIGURE 3.5: Dry Navy Bean Protein Extraction and UF Fractionation (200,000 kd) Used for Process Trial 1. A two-step separation was utilized in an effort to obtain a highly concentrated protein fraction. The molecular weight separation was based on a 200,000 (kd) cutofl‘. This was an attempt to isolate the phaseolin protein fraction fiom navy beans. This exclusion size is also important to remove smaller, undesirable fractions such as the lectins (50,000 (kd)) and phytate. 110 Hammer-milled flour was hydrated in a alkaline/ salt solution for 24 hours to separate starch and protein components. Salt-soluble proteins were extracted in the supernatant and partitioned from the starch/fiber residue. The decanted meal supernatant was filtered and passed through the 200,000 (kd) cut-ofl‘ membrane for protein concentration. The data presented in Figure 3.5 indicate partitioning and separation occurred, yielding 126.5 lbs (42% protein) of retentate versus 27 lbs (14% protein) of permeate recovered during the initial fractionation. The protein concentrate from the retentate was then processed through the membrane system a second time. The mass balance indicated that 109 lbs (47% protein) of retentate versus 59 lbs (12% protein) permeate was recovered with this additional UF pass. This second phase of processing achieved the intended objective of firrther concentrating the protein fraction. The final retentate protein concentration was 47%. During UF processing, foaming was a problem due to the high level of undenatured (native) protein content and high flux rates experienced during filtration. This foam was washed with intermediate water sprays during processing to assure adequate recovery. However, ultrafiltration proved to be an efl‘ective method of separating fractions yielding a high level (47%) of protein fi'action from native bean flour which was suitable for subsequent product formulation. The composition of bean protein fiactions is presented in Table 3.2. Results indicate that protein was 13 .5 times more concentrated in the retentate than in the permeate. 111 TABLE 3.2: Composition of UF Separated Bean Protein Fractions; 200,000MW Permeate and Retentate. NUTRIENT/MEASURE (”L UNITS j PERMEATE I RETENTATE Total Solids g/ 100g 1.55 4.52 Protein g/ 100g 0.16 2.16 Brix Degree 2.39 4.8 pH 6.95 6.18 Calcium mg/lOOg 9.92 15.12 Phosphorus mg/ 100g 5.53 10.77 Iron mg/ 100g 0.08 0.29 Sodium mg/ 100g 348 364 Potassium mg/ 100g 90.52 94.82 Copper mg/l 00g 0.86 0.17 Zinc mg/ 100g 0.34 0.64 Magnesium mg/ 100g 6.07 8.64 (1) Mean Values Although minerals are soluble and of relatively low molecular weight, it was not surprising that the minerals were also found to be in greater abundance in the retentate: predominantly, calcium, phosphorus, zinc and iron. This may be due to the minerals’ ability to form covalent bonds with proteins, inhibiting them from being separated during ultrafiltration. The monovalent salts, sodium and potassium contents were nearly equivalent in concentration between theretentate and the permeate (an indication of equilibrium in the solution). Divalent minerals, such as copper, were adequately partitioned into the permeate suggesting they are perhaps less bound to proteins than calcium, phosphorus, zinc and iron. Further quantification of the protein fractions for the permeate and retentate was achieved using SDS gel electrophoresis (Figure 3.6). 112 L — 200K — 173 - '64 106 _ q 62 — 55 — = h—i 45 — — F 3] — — _ 18 — — FIGURE 3.6: SDS-PAGE of UF Fractionation (200,000 MW) for Navy Bean Proteins: Permeate (A); Retenate (B); Standards (M). The SDS-PAGE analysis utilizes detergent to separate subgroups within the native species. Thus, the appearance of numerous low molecular weight components within high relatively high Rm regions is due to the separation of the native protein subgroups. The permeate fraction (A) bad a higher concentration of protein above 50,000 (kd) molecular weight than below the 50,000 (kd) molecular weight cut-ofl‘. The retentate fraction (B) had more bands, a denser and broader spectrum of low to high molecular weight. This result was due to the exclusion membrane of 200,000 (kd) molecular weight cut-ofl‘ utilized to separate the native protein. Proteins below this size were partitioned into the permeate fi'action. Larger proteins were not able to pass through the membrane and were consequently concentrated in the retentate. ./._ 1 The amino acid profile for the permeate and retentate streams appear in Table 3 .3. 113 TABLE 3.3: Amino Acid Composition (mg/100g:mg/g of Protein) of UF Separated Bean Protein Fractions; 200,000 MW Retentate and Permeate. Permeate Retentate Classification Amino Acid mfllOOg mgg of Protein mglflg mg; of Protein Aliphatic Alanine 3.42 21.38 92.90 43.01 Hydrocarbon Glycine 2.97 18.56 85.20 39.44 lsoleueine 0.94 5.88 102.40 47.41 Leucine 2.93 18.31 175.40 81.20 Valine 1.38 8.63 115.00 53.24 subtotal 11.64 72.75 570.90 264.31 Alcohol Serine 1 .46 9. 13 129.60 60.00 Threonine 0.00 0.00 101.50 46.99 subtotal 1.46 9.13 231.10 106.99 Acid Aspartic Acid 3.13 19.56 159.30 73.75 Glutarnic Acid 23.89 149.31 286.00 132.41 subtotal 27.02 168.88 445.30 206.16 Basic Arginine 2.64 16.50 109.40 50.65 Histidine 3 .62 22.63 61.60 28.52 Lysine 2.86 17.88 149.70 69.31 Tryptophan 2.40 15.00 27.50 12.73 subtotal 11.52 72.00 348.20 161.20 Sulfur Cysteic Acid 2.10 13.13 25.30 11.71 Methionine 0.86 5.38 19.30 8.94 subtotal 2.96 18.50 44.60 20.65 Aromatic Phenylalanine 1.76 1 1.00 125.40 58.06 Tyrosine 7 .94 49.63 86.10 39. 86 subtotal 9.70 60.63 211.50 97.92 Heterocyclic Proline 2.16 13.50 79.60 36.85 subtotal 2.16 13.50 79.60 36.85 TOTAL 66.46 415.38 1931.20 894.07 N=means of duplicate analyses conducted on a single sample Eighteen amino acids were evaluated in this study. The data were reported both as mg/ 100g and mg/g of protein. The latter format is required to calculate an amino acid score for determination of protein quality. Since relatively high molecular weight protein 114 was retained by the 200,000 (kd) molecular weight cut-ofi‘ membrane, almost twice (2X) mg/g of protein was found in the retentate (894.07 mg/g of protein) versus the permeate (415.38 mg/g of protein). This difference was indicative of the concentration of protein during ultrafiltration. The calculations of the amino acid score for the permeate and retentate are outlined in Table 3 .4. TABLE 3.4: Summary of the Estimates of Amino Acid Requirements for Children (2 years old), as Proposed in 1985 by Food and Agriculture, World Health Organization, United Nations and Amino Acid Score for Permeate and Retentate. Amino Acid Referencel Permeate Score Retentate Score mg/kyday mg/ g of mg/g of Protein Protein Isoleucine 31 5.88 .19 47.41 1.53 Leucine 73 18.31 .25 81.20 1.11 Lysine 64 17.88 .28 69.31 1.08 Methionine & 27 18.51 .69 20.65 0.76 Cysteine Tyrosine & 69 60.63 .88 97.92 1.32 Phenylalanine Threonine 37 O 0 46.99 1.27 Tryptophan 12.5 15 1.2 12.73 1.02 Valine 38 8.63 .23 53.24 1.40 1Reprinted with permission from Recommended Daily Allowances: 107" Edition, Copyright 1989 by the National Academy of Sciences. Courtesy of the National Academy Press, Washington, DC. Table 3 . 5 displays the amino acid scores of the ten essential amino acids for the permeate and retentate fractions obtained in Process Trial 1. 115 TABLE 3.5: Amino Acid Score Calculation, in Ascending Order, of Ten Essential Amino Acids for 2-yr-01d Children. Scores for both Permeate and Retentate Using UF 200,000 MW Membrane. Permeate Retentate Threonine O Methionine & 0.76 Cysteine lsoleueine . 19 Tryptophan 1.01 Valine .23 Lysine 1 .08 Leucine .25 Leucine 1.1 1 Lysine .28 Threonine 1 .27 Methionine, Cysteine .68 Tyrosine & 1.32 Phenylalanine Tyrosine & .88 Valine 1.40 Phenylalanine Tryptophan 1 .2 lsoleueine 1 .53 Quantitative results are presented in ascending order with the first limiting amino acid at the top of each column. As expected, the retentate showed the sulfirr containing amino acids, methionine and cysteine (0.7 6), as the most limiting. These data are consistent with other legume analyses for navy beans. However, the permeate fraction showed a more unusual pattern with threonine at non-detectable levels. The sulfiir containing amino acids, methionine and cysteine, were not limiting in this fraction. In fact, cysteine was found at high levels in the permeate (0.49). This relatively high concentration of cysteine may be associated with the low molecular weight (less than 50 kd) of lectins. The nutritional and analytical comparison between control and bean protein fortified indicated subtle compositional differences (Table 3 .6). 116 TABLE 3.6: Proximate Analyses and Wet Chemistry Analytical Data of Standard Peach Shake and Peach Shake Fortified with UF Separated Bean Protein Fraction PEACH SHAKE FORMULATION COMPONENT (gl100 g) CONTROL BEAN PROTEIN FRACTION Nutritional Protein 3.85 3.39 Carbohydrate 18.88 17.70 F at 2.06 1.67 Saturated Fat 1.19 1.04 Ash 0.70 0.81 Total Dietary Fiber 0.90 0.76 Total Sugars 15.30 14.30 Analytical Total Solids (g/ 100g) 25.13 23.86 Brix 20.47 20.47 Total Activity (%) 0.67 0.67 pH 4.33 4.33 With the exception of carbohydrate content, all nutrient levels were essentially the same. The lower carbohydrate level in the bean protein fortified product was attributed to carbohydrate dilution inherent in difference calculations associated with the added protein. The control, consisting of a peach shake fortified with low-fat yogurt, was processed under the same conditions as the bean protein fortified shake except no bean concentrate was added. The compositional differences were assessed to be non- distinguishable based on the standard variances inherent in these standard analytical procedures. Normally, the minimum batch size for production is 250 gallons; however, the experimental objective was designed to maximize the proportion of protein concentrate isolate within the batch, which necessitated modification. To achieve this, a SOD-gallon batch of peach shake was made and approximately 3 52 gallons were 117 supplemented with the bean protein concentrate. This fortified batch was used throughout processing and packaging of the final product. From the 352 gallons, approximately 250 cases of final product was produced, which was then used for analytical analyses and sensory consumer evaluations. The specific mineral and vitamin content of the finished products are presented in Table 3.7. TABLE 3.7: Mean Values for Mineral and Vitamin Content of Standard Peach Shake and Peach Shake Fortified with UF Separated Bean Protein Fraction PEACH SHAKE FORMULATION NUTRIENT (mg/100gm) CONTROL BEAN PROTEIN FRACTION Minerals Calcium 93.60 93.30 Iron 0.24 0.24 Sodium 49.00 64.00 Zinc 1.59 1.63 Vitamins B—Carotene (IU) 93.00 80.60 Vitamin C 5.24 5.31 Thiamin 0.07 0.07 Riboflavin 0.17 0.15 Niacin 0.37 0.36 Vitamin B6 0.08 0.07 Standard Test Methodologies with defined variances It is noted that the control had a lower sodium content (49.0 mg/100 gm) versus bean protein fraction (64.0 mg/ 1 00gm). The increase in sodium is attributed to the salt added to the extraction medium, which yielded increased protein solubility and was not completely dialyzed to the permeate fiactions. The control also exhibited a higher B- 118 carotene level (93.0 IU) versus bean protein fraction (80.6 IU). This phenomenon was probably due to the dilution efi‘ect caused by bean protein fortification. The nutritional label declaration data are outlined in Table 3.8. TABLE 3.8: Nutrient Profile of Peach Shake with Lowfat Yogurt. % USRDA for Ages 12 to 48 Months - Proposed Nutritional Label Nutrient Units Actual Label ‘1’ (per 6.8 ounce [192.89] Control W serving) Calories kcal 210.00 190.00 Protein % U.S.RDA 7.00 6.00 Carbohydrate g 36.00 34.00 Fat % U.S.RDA 6.00 5.00 Calcium % U.S.RDA 20.00 20.00 Zinc % U.S.RDA 20.00 20.00 Iron % U.S.RDA 2.00 2.00 Sodium mg 95.00 120.00 Total Vitamin A % U.S.RDA less than 2% Vitamin C % U.S.RDA 10.00 10.00 Thiamin % U.S.RDA 10.00 <10.00 Riboflavin % U.S.RDA 20.00 15.00 Niacin % U.S.RDA 4.00 4.00 Vitamin B6 % U.S.RDA 8.00 8.00 (1) = NLEA Guidelines The finished product is positioned as a calorically dense food (190 kcal/serving). Serving size is based on 193ml of beverage and uses the USRDA guidelines established for children 12 to 48 months. Actual protein levels were 6% USRDA under NLEA guidelines. Calcium at 20% USRDA is slightly lower than milk (33% USRDA); but is still an excellent source according to NLEA guidelines. Sodium content is only 120 mg, which meets criteria for “low sodium” product. Product is an excellent source of calcium and zinc which are very important for growth. The nutrient of insufficient or low quantity in 119 this food system was ascorbic acid (Vitamin C) at only 10% USRDA Vitamin C content should be at 100% of the USRDA for small children since this vitamin plays a significant role of increasing iron adsorption in children, as well as providing antioxidants and other physiological properties. Fortification of the beverage will enhance Vitamin C content and is normal in the food industry. Thiamin, Riboflavin and B5 are in line with fortification target guidelines. Much higher levels of these B vitamins will cause ofilflavors and give a medicinal taste that results in decreased product acceptability. It was important not only to develop a nutritionally dense product but also to provide it in a manner that did not sacrifice taste and thus increase its probability of consumption A typical nutritional panel displaying the label declaration for the control and protein-fortified Peach Shake is outlined in the official NLEA format (Figures 3.7 and 3.8). 120 Nutrition Facts Serving Size 6fl oz (180ml) Servin 3 er Container 1 Amount per Serving Calories 210 Total Fat Sodium Total Carbohydrate Fiber Sugars Protein °/o Daily Value Children ages 14 Protein 7% Vitamin A less than 2% Vitamin C 10% Vitamin 36 8% Calcium 20% Iron 2% Thiamin <1 0% [Riboflavin 20% INiacin 4% Izmc 20% FIGURE 3.7: Label Declaration and Format Display Using NLEA Guidelines for Children Ages 12 to 48 Months for Formulated Peach Shake (Control) . FIGURE 3.8: 121 Nutrition Facts Size 6fl oz. (180m|)) Container 1 Servi 190 otal Fat um otal 1 27 n % Daily Value Children 1-4 in A less than in C 1 n BS um ron in 'boflavin iacin Label Declaration and Format Display using NLEA Guidelines for Children Ages 12 to 48 Months for Formulated Peach Shake Fortified with Bean Protein. “Nutrition Facts” labels for food specifically for children less than 4 years old do not present % Daily Value footnotes as used on the general food supply labels as outlined by The Code of Federal Regulations (CFR). Label information for children less than 2 years does not recommend labeling of calories from fat or saturated fat. This recommendation is based on the need for small children to receive sufficient caloric intake. 122 Calories fiom fat pose little health risk at this age but rather are needed to sustain the rapid grth in children’s early years. The externally defined design attributes which drive the development of a high quality beverage for small children are outlined in Figure 3 .9. Customer Needs (200—300 in hierarchy) Low Cost Nutritional 0 Protein Content 0 Meat Supple- ent(100 salaries) - Vita- in/Mineral Fortified 0 Iron Supply Locally Grown Ingredients Food Safley (pH) Taste Convenience Sanitation Stability Ready-to-Serve (no toe-I was-r seq-ha) Microbiologically Stable FIGURE 3.9: — Swuqauau A i Design i Relationships between Customer Needs and Design Attributes Costs and Feasibility Local Manufacturing am.) Local Raw Materials 0 Paper 0 Dairy 0 Legumes Local Labor Distribution Efflciences Job Creation Plant Efficiencies Low Inventories “Engineering” Measures Technology Focused and Funded 0 Ultraflitration - Aeseptic Processing Low Environmental Risks Technology Support MN and Tetra his. Regulatory Compliance (WHO, FAD. on.) Fortified Shake Utilized. Locafly Grown Ingredients. Products were nutritionaly fortified. The product was convenient, good tnsthg and safe to consume. Customer Perceptions House of Quality Blueprint of External Defined Attributes for a Formulated Beverage Suitable for Small Children. The primary objective of development was to define critical customer needs, by priority, necessary to drive the product design. Achieving customer feedback is essential for determining whether the product is fitlfilling the unmet consumer needs. In this case, children in under-developed countries were targeted; low cost and high nutritional quality 123 were more important criteria than taste and convenience. In developed countries like the United States, taste and convenience lead the hierarchy of priorities while low cost replaces the “perceived value for their money”. 124 Conclusigns These findings indicated that legume protein can be selectively isolated using ultrafiltration technology. These proteins can be used in formulation of a fortified nutritional beverage that is aseptically processed and packaged suitable for human consumption. Protein fractions should be firrther characterized to assess their efl‘ectiveness as an ingredient resource for applications in other food formulation. Ultrafiltration should also be explored fiirther, to use this technology to isolate key ingredients by their molecular weight based on desired functionality. Quality Functionality Deployment is an acceptable process to apply for development of products that satisfy consumer needs. The objective of this study was to design and develop such a product and quantitatively and quantitatively measure it versus a “known” control as the benchmark. This was achieved with the development of protein-fortified shake with utilized protein fiom navy beans. The product was acidified and aseptically packaged using “known” and globally accepted technology with Tetra Laval®. This packaging material is economical and provides a convenient mechanism to distribute and market finished packaged consumer goods products. The ability to store at ambient conditions is also important in areas where refiigeration is limited. SUMMARY AND CONCLUSIONS Dry edible navy beans (Phaseolus vulgaris L.) were utilized in this study because they are an economical protein source locally grown in Lesser Developed Countries (LCD). Small children in LCD typically survive on diets high on carbohydrates with low levels of protein and essential vitamins and minerals. The ability to afi‘ord and maintain fortified foods necessary to support life during early years can be an economic hardship in LCD. The purpose of this study was to develop a protein fortified shake utilizing local ingredients, which could provide small children (2-5 years old) with a beverage that had high nutritional quality and good taste and was convenient and was afl‘ordable. Therefore, the application of membrane separation technology to isolate legume protein that could be fortified into a beverage was the vehicle used to fillfill this unmet consumer need. Quality Functional Deployment (QFD) was the process utilized to externally define these unmet needs. This technique represents the best approach for defining and externally measuring quality parameters necessary to drive product design and subsequent development. In this case, it was found that taste, nutritional quality, safety, convenience and long shelf life were critical in designing a product to satist consumer wants and needs. Separation technology utilizing microfiltration (MP) and utrafiltration (UF) processing was evaluated to fiactionate specific proteins from dry edible beans. A series of various membranes were tested. These consisted of plate and frame, tubular and spiral wound with 50,000(kd) and 200,000(kd) molecular weight cut-011‘. It was determined that the tubular membrane was the preferred method of separating protein. The plate and frame and spiral wound both exhibited significant loss of flux during processing because of 125 126 membrane fouling. This membrane fouling caused a significant increase in pressure, temperature and overall processing inefficiencies. In conjunction with the tubular membrane, it was also determined that 200,000 (kd) cut-ofl‘ had the highest protein yield. The overall protein recovery in the retentate reached 40.0%. The amino acid scores of these proteins also indicated that the sulfur containing amino acids, methionine and cysteine, were the most limiting of the ten-essential amino acids for children (2 years old) which was consistent with levels found in the whole bean. The overall protein quality from these fractions was also measured qualitatively and quantitatively in the raw and heated state in an effort to simulate thermal heating during beverage processing. Confocal Microscopy was used as a measurement of protein aggregation during simulated thermal process conditions. These results indicated a slight increase in agglomeration when subjected to heat conditions ranging from 71° to 99°C for 5 seconds. Quantitatively, gel electrophoresis and nitrogen soluble indexes (N SI) were utilized to evaluate level of protein denaturation. During gel electrophoresis, the non- denaturing and non-dissociation (NDND)-PAGE was utilized. The result indicated that proteins above 545 (kd) were present in the retentate, an indication of native protein selection. These findings were most significant with the tubular, 200,000 MW membrane. These proteins were also subjected to the same heat treatments used for the Cofocal Microscopy and when analyzed for NSI, no significant difi‘erences in protein solubility were found among treatments, a result firrther confirming limited protein denaturtion. Further evaluations using NDND-PAGE on heated samples indicated little protein denaturation occurred during simulated thermal processing. 127 The isolated protein fractions from MF and UF were added to a peach shake beverage to fortify to 6% protein content suitable for small LCD children. This product was formulated, processed and aseptically packaged in a 200 ml Tetra Slim® paperboard package. The equipment utilized was manufactured by Tetra Laval” in Lund, Sweden, a global company, which provides industry with sanitary food processing and packaging equipment. This system provides a complete model of processing and distribution efficiencies, which enables the company to be one of world’s true low cost producers. Bean fortified and non-bean (control) samples were processed and packaged aseptically at 118°C for 4.6 seconds using a TBA-8 Tetra Laval® system. The results of theses studies are significant and show conclusively that MF and UP fractionated protein from dry edible beans (Phaseolus vulgaris L.) is acceptable for use in a formulated beverage product suitable for small children. RECOMMENDATIONS FOR FURTHER RESEARCH Further research needs to be conducted to optimize legume protein separation using membrane technology. This work needs to be coupled with in-depth investigation of targeted protein masses, peptides or amino acids that may have potential consumer health benefits. Together, food technology and nutrition can drive the development of new consumer products into the twenty-first century providing industry with higher profit margin brands. This will give consumers healthier, better tasting products, improve the quality of life, and prevent disease while improving mortality rates worldwide. Specifically, key areas of further investigation are as follows: I Spiral wound membranes showed a high degree of fouling, which made operation a difficulty. However, the data indicated that a greater protein fractionation may have occurred with the spiral wound membrane which may be thus a more selective membrane. u The focus of the research was on utilizing the retentate. The data indicated that specific protein masses were isolated in the permeate. This conclusion needs to be further quantified. u Opportunity exists to firrther investigate high cysteine concentrations in the permeate. This may be associated with lectins; thus membrane separations may be an effective technology to remove antinutritional factors without using heat. This would enable undenatured protein fractions to be available for food formulation and yet preserve their inherent functionality. I Further protein quantification such as amino acid sequencing may be helpfirl to determine if native phaseolin or vicilin is fractionated. 128 129 u The experiments were run with recycling of the permeate. It may be helpfiil to improve separation and membrane performance by employing complete diafiltration techniques through diluting with deionized water. I Optimization of the extraction process to centrifirge or decant the precipitate material will improve membrane performance. Improved protein fractionation sensitivity would be achieved if lower carbohydrates are present in the initial fiactions. The conditions for ultrafiltration and microfiltration would also be improved. I Ultimately, the product needs to be presented to children in LDC to determine acceptability. APPENDICES APPENDIX A: Preliminary Extract and UF Processing Trial This initial study was conducted in cooperation with Dr. S. McCurdy at the POS Pilot Plant in Saskatoon, Canada, to assess the potential for membrane fraction of Phaseolus vulgaris L. proteins. This location was selected due to the availability of pilot scale milling, extraction and UP membrane fractionation capability, and personnel expertise in Northern Plains crop utilization (pulses and oil seeds). This pre-trial test was undertaken to evaluate the type of membrane system suitable for fractionation (plate and frame, spiral wound, tubular or hollow fiber design) and to screen a series of difi‘erential membranes with variable molecular weight exclusion specifications (“cut-oft”). Results of this study indicated that excessive membrane failure was a primary limitation with the Phaseolus vulgaris L. protein separation when utilizing plate and frame membrane. Mass balance, total mass, protein, solids and flux rate data were used to assess fiirther system design. The tubular membrane system with a double pass diafiltration sequence was proposed for subsequent process trials. The report of these initial trials is presented in Appendix A. 130 131 PR JE SUMMARY TITLE: Extraction and Ultrafiitration of Navy Bean Protein BACKGROUND The Gerber Products Company is interested in using bean protein in product formulations. However, in the isolation of bean protein, trypsin inhibitor and other antinutritional compounds are co-extracted with the protein. The relatively low molecular weight of the undesirable compounds suggests that they can be removed fi'om the solubilized bean protein via ultrafiltration prior to spray-drying. The overall objectives were to produce small samples of extracted, ultrafiltered navy bean protein for evaluation by the client and to obtain mass balance information for the process. WORK PERFORMED Navy beans were provided by Dr. Mark Uebersax, Michigan State University, E. Lansing, Michigan, and milled to a flour in a laboratory pin mill. The project work was divided into two phases. In the first phase, mass balance data for the extraction and ultrafiltration of navy bean protein were collected in duplicate. The extraction conditions utilized a 1:7 extraction ratio, at 4°C, pH for 1 hour. The extract was diafiltered for 1 hour, then ultrafiltered using a 50,000 molecular weight cutoff membrane to reduce volume to about one-third. 132 In the second phase, navy bean protein extract was produced under the same conditions as Phase I, then subdivided and diafiltered and ultrafiltered using three different molecular weight cutoff membranes. The retentates were spray-dried to provide samples for evaluation. RESULTS Sixty percent of navy bean flour protein and 25% of the flour dry matter were solubilized under the aqueous extraction conditions employed. In Phase 1 dia- and ultra- filtration of the protein extract against a 50,000 molecular weight membrane resulted in 7% of the original dry matter and 6% of the original protein passing through the membrane into the permeate. The retentate had 6.4% solids (as-is-basis), a protein content of 7 0% (dry basis) and recovered 3 8% of the starting protein. The two replicate experiments produced very similar results. In Phase 2, dia- and ultrafiltration gave better protein yields, 46% to 49% of starting protein, and a higher protein content in the product (74% to 75% protein, dry basis). This was probably due to longer processing times, since a more concentrated retentate was produced. As the molecular weight cutoff of the membrane was increased, there was a slight increase in the amount of dry matter and protein partitioning into the permeate, from 8. 5% of the original dry matter and 7.2% of original protein (6,000MW) to 9.3% of dry matter and 9.7% of original protein (200,00MW). 133 Samples of the starchy precipitate byproduct of extraction and of the permeates were freeze-dried for evaluation by the client. Phase 1 retentates were also freeze-dried; retentates from Phase 2 were spray-dried to more closely simulate a commercial process. CONCLUSIONS Mass balance data for wet matter, dry matter and protein were obtained for the aqueous extraction (pH 9) of protein fiom navy beans and dia- and ultra-filtration processing to remove anti-nutritional components. Samples of navy bean protein ultrafiltered with three different molecular weight cutoff membranes were produced. 134 TABLE A: Comparison of Fraction Wet Weight'I and Composition Obtained for Replication 1 2 in Phase 1. Fraction, Component Replication 1 Replication 2 J Nag Bean Slim Wet Weight, G 2415 g 2414g % & g of Dry Matter 9.44%; 228.0g 10.18%; 245.73 % & g of Protein, dry 23.33%; 53.23 21.57%; 53.0g basis Sgrch & Hull Precgg' itate Wet Weight, g 567.2g 596.8g % & g Dry Matter 32.89%; 186.5g 31.30%; 186.8g % & g Protein, dry basis 11.90%; 22.2g 11.90%; 22.2g Protein Extract (supernatant) Wet Weight, g 1641.5g 1672.6g % & g Dry Matter 3.68%; 60.4g 3.59%; 60.0g % &gProtein, dry basis 54.37%; 32.8g 54.47%; 32g Permeate Wet Weight, g 1414.0g 1373.0g %&gDryMatter 1.14%; 16.1g 1.11%; 15.2g % & Protein, dry basis 21.57%; 3.5g 18.74%; 2.93 m Wet Weight, g 410.0g 475.0g % & g Dry Matter 6.39%; 26.2g 5.74%; 27.3g % & LProtein, dry basis 70.13%; 18.4g 69.53%; 19.0g ‘ Actual wet weights taken during processing and not corrected for samples taken or handling losses. 135 TABLE B: Actual Dry Matter and Protein Recoveries from Diafiltration/Ultrafiltration Processing in Phase 2. Fractions and Actug D51 Matter Data Actual Protein Data Membrane Percentage Mass Yield Percentage Mass Yield in Wet Recovered in Dry Recovered Fraction Matter Protein Extract 3.66 153.7 100 54.7 84.1 100 Permeate 6,000MW 1.24 52.1 33.9 18.9 9.84 11.7 50,000MW 1.27 55.1 35.8 18.7 10.3 12.3 200,000MW 1.42 56.4 36.7 23.5 13.3 15.8 Retentate 6,000MW 9.25 75.9 49.4 75.6 57.4 68.2 50,000MW 11.9 74.5 48.5 74.6 55.6 66.1 200,000MW 6.91 84.5 55.0 74.3 62.8 74.7 136 PROJECT REmRT TITLE: Extraction and Ultrafiltration of Navy Bean Protein BA R UND The Gerber Products Company is interested in using bean protein in product formulations. However, in the isolation of bean protein, trypsin inhibitor and other antinutritional compounds are co-extracted with the protein. The relatively low molecular weight of the undesirable compounds suggests that they can be removed fiom the solubilized bean protein via ultrafiltration prior to spray-drying. OBJE S The overall goal of this project was to produce small samples of extracted, ultrafiltered navy bean protein for evaluation by the client and to obtain mass balance information for the process. This goal was accomplished by: 1. An aqueous extraction and ultrafiltration experiment using 300g of navy bean flour, done in duplicate, to collect mass balance information (Phase 1), and 2. Preparation of 13,000g of navy bean protein extract, followed by diafiltration/ultrafiltration processing of 4200g aliquots, each through a difl‘erent molecular weight cutofi‘ membrane, and spray drying of the retentates (Phase 2). WORK PERFORMED Materials: Approximately 45kg of Seafarer navy beans were received from Dr. Mark Uebersax, Michigan State University, E. Lansing, Michigan (primary source: Hilbert Schulze, Hillman, MI) on September 30, 1991. Due to a scheduling conflict, the 137 beans were not pin-milled in the pilot plant as originally proposed. A total of 6.8kg were milled, first through a erey Mill (Model No. 2, Arthur H. Thomas Co., Philadelphia, PA) to reduce size in preparation for milling in a laboratory pin mill (Ultra-Centrifirgal Nfill, F. Kurt Retsche GmbH & Co, Haan, Germany). Particle size analysis was conducted on the flour. Ro-tap sieve analysis was used for particles above 700nm; a light scattering type particle size analyzer (microtrac H, Leeds & Northrup, North Wales, PA) was used for particles under 700nm. Proximate analysis was also conducted on the flour. PHASE 1 - Mass Balance Extraction: Three hundred g of bean flour were mixed with tap water at 4°C and the pH adjusted to 9 with 3N NaOH using an autotitrimeter (Metrohm Titroprocessor 686 and Dosirnat 665, Metrohm Ltd, Herisau, Switzerland). After the slurry mixed (300 rpm, overhead stirrer) for 1 h (held at 4°C and pH 9), it was centrifiiged (4700 xg, 15 min) to recover the protein extract. The extract was adjusted to pH 7 with 3N HCl. Diafiltration/Ultrafiltration: The extract (supernatant) was transferred to a 4L beaker in a 4°C water bath and diafiltered for about 1 h at a feed rate of 2.3 L/min using 336 cm2 of 50,000 molecular weight cutofi‘ polysulfone membrane (DDS GR51PP) in a lab scale plate and frame ultrafiltration unit (Mini-Lab 10, DDS RO-Division, Nakskov, Denmark). During diafiltration (DF), the feed volume was kept constant by adding water. In this case, a total of 400ml of water was added in 50ml increments as the permeate was drawn ofl’. After 1 h, additions of water to the feed tank were ceased, and ultrafiltration (UF) was carried out until the retentate volume reached about 4501111. During these processes, the feed was cooled by circulating cooled water around the breaker. The 138 temperature of the supernatant rose from 4°C to 31°C during the 2.5 h total processing period. Sampling: Wet weights were recorded, and triplicate samples were taken of the slurry, supernatant, precipitate, permeate (total from DF and UP), and retentate for dry matter and crude protein assays. The materials remaining after sampling were freeze— dried. Phase 2 - Ultrafiltration with Three Membranes Extraction: On Monday, October 28, 1991, 2.3kg of navy bean flour was extracted with 16.1 L of prechilled water (4°C) as described for Phase 1. This process produced 13.3 kg of navy bean protein extract (pH 70 to be divided into three aliquots and used to compare the effects of using three different molecular weight cutofl’ membranes for ultrafiltration of the extract. The extract was placed in a 20L bucket and stored at 5°C until needed. Diafiitratiou/Ultrafiitration: The filtration processing of the extract had been planned for October 29—31 following extract production, using a different membrane each day. However, a plate in the ultrafiltration unit broke at the end of the first filtration and replacement required two days. In addition, UF processing with the 6,00MW and 200,00MW membranes required about 10 h and thus was conducted over two days, so processing actually took place on October 29 (50,000MW), November 1-2 (6,000MW) and November 4-5 (200,000MW). An aliquot of well mixed extract was transferred to a 5 jacketed feed tank; chilled water from a refiigerated water bath was circulated through the tank jacket to cool the material during dia- and ultra-filtration processing. Diafiltration was carried out as 139 described previously, but starting with a 4.2kg aliquot of extract so that a total of 1000ml of water was added and the time was extended. The feed rate was 126 to 175 Mt, the pressure drop 0.1 to 0.5 bar, and processing time 2 to 3 h, depending on the membrane used. Ultrafiltration was conducted to a final retentate volume of 600 to 1200ml, depending on the membrane. For ultrafiltration, the feed rate was 171 to 207 Mr, pressure drop 0.5 to 0.7 bar, and processing time 6.5 t 10.1 h. In each of the three different membrane trials, a new polysulfone membrane (DDS GR81PP, 6,000MW; DDS GR51PP, 50,000MW; DDS GR30PP, 200,000MW) was used at the start of processing. Spray-Drying: The ultrafiltered retentates were spray dried the day after production in a laboratory scale, glass spray dryer (Buchi 190 Mini-Spray Dryer, Flawil, Switzerland) using an inlet temperature of 150°C and an outlet temperature of 75°C. Sampling: Wet weights were recorded, and triplicate samples were taken of the precipitate, extract, and each permeate and retentate for dry matter and protein assays. Portions of the precipitate and permeate were freeze-dried. An ic ' Dry matter contents were determined by drying appropriately even-sized samples (5 to 100g) in a forced air oven at 105°C to constant weight. Nitrogen content of the dried samples was determined by an automated Kjeldahl method using the Kjel-Foss Automatic Model 16210 (AOAC Method 7.021, 1984) and crude protein content obtained by multiplying by the factor 6.25. 140 RESULTS Bean Flour: The beans were milled at a fairly high moisture content (13.9%) to avoid producing a very fine flour that would be dificult to disperse. The particle size analysis is shown in Table 1. TABLE 1: Particle Size Analysis of Navy Bean Flour Seive Size % of Sample (U.S.A. Standard Seive No.) On 30-mesh ' (600m) 3 5.5 On 35-mesh ' (500nm) 7.9 On 45-mesh " (355nm) 19.9 On 6-mesh " (250nm) 10.6 On 120-mesh " (125nm) 9.8 On 325-mesh ° (45pm) 6.0 Below 325-mesh ° 11.3 ' Ro—Tap Analysis ° Particle Size Analyzer Proximate analysis of the bean flour is shown in Table 2. TABLE 2: Proximate Analysis of Navy Bean Flour Component Percentage__ Moisture, as is basis 12.98 Fat, dry basis 1.56 Ash, dry basis 4.46 Crude Fiber, dry basis 4.54 Protein, dry basis 22.52 141 Phase 1 - Mag Balance A mass balance diagram for wet mass, dry mass and protein for the aqueous extraction and ultrafiltration processing of navy bean protein is shown in Figure 1. The masses were extrapolated to compensate for sampling and are an average of results obtained for the two replicates. The two replicates were in good agreement; the raw data from the two replicates are show in Table A in the Appendix. 100 parts Whole Bean Flour 87.02 parts dry matter 100% 19.6 parts protein + 700 parts water @ 4°C 4» 5.4 parts 3M NaOH to achieve pH 9 and hold for 1 hr. 805.4 arts Bean Slur 79.01 parts dry matter 17.74 parts protein ‘Centn'fuge @, 4700 x g for 15 min. 209.3 parts Starch and 596.1 parts Bean Protein Extract Hull Precipitate 21.67 parts dry matter 67.17 arts matter . p dry 11.79partsprotem 7'99 parts protein + 1.6 parts 3M HCl to achieve pH 7 + Dialilter adding 133 parts of water A/erter against 50,000 MW membrane 554.7 parts Permeate ”50 arts Retentate 6.24 parts dry matter 10.7 parts dry matter 1.26 parts protein 7-5 parts pmwin ' Sampling of shiny was likely poor; dry matter and protein balance were improved for sampling of precipitate and extract. FIGURE 1: Mass Balance Diagram for Phase 1 Aqueous Extraction and Ultrafiltration Processing of Navy Bean Protein 142 The aqueous extraction at pH 9 solubilized 60% of the bean protein and 25% of the dry matter (Figure 1). Diafiltration and ultrafiltration processing of the extract with a 50,000MW membrane removed 7% of the dry matter and 6% of the crude protein (Figure 1). Retentate hold-up in the ultrafiltration unit is 10011115, accounting for the fairly high losses of dry matter (5.4%) and protein (15.7%) occurring at this step. prrocessing volumes were larger, it is likely that 50% of the original protein (rather than 38% as shown in Figure 1) would be recovered in the retentate. The ultrafiltered navy bean protein concentrate had an average protein content of 69.8%, dry basis (Table A, AppendiX). The flux rate during the diafiltration period remained constant, as would be expected, since no concentration of protein was taking place. The flux rate during for the first replication was 17.4 L/m2 x h; during the second replication flux rate was slightly less, between 15.2 and 16.1 L/m2 x h. The first replication utilized new membranes and these membranes were cleaned and used for the second replication. During ultrafiltration, flux rate did not drop off during processing as expected. This may be due to the short processing time (1 .3-1 .5 h) and small volume (14001111) of the material processed. Flux during ultrafiltration was similar to that observed during diafiltration. 143 Phase 2: Uitrafiltration with Three Membranes Extract: The Phase 2 mass balance for the aqueous (pH 9) extraction of bean protein is shown in Figure 2 and was calculated using the same criteria of compensating for sampling loss as in Figure 1. The mass balance for extract production was similar to that achieved in Phase 1. Ultrafiltration: The processing parameters for diafiltration/ultrafiltration of the bean protein with three different membranes are shown in Table 3. Diafiltration flux rates varied among the membranes, reflecting primarily the water flux rates published by the manufacturer. Ultrafiltration flux rates were more variable, probably reflecting some plugging of the membrane during processing. During the last 1.5 to 2 h of processing, flux rate dropped off as is typical in ultrafiltration systems as the retentate becomes more concentrated. 144 TABLE 3: Observations of Weights, Times and Rates During Diafiitration Membrane 6,000MW 50,000MW 200,000M Mass of Extract Processed 42003 42003 42003 Mass of Water Added 10003' 10003 10003' During Diafiltration Feed Rate Diafiltration 2.5 L/min 2.1 L/min 2.9 L/min Ultrafiltration 2.9 L/min 3.4 L/min 2.9 L/min Pressure Drop Diafiltration 3.0h 2.0h 2.3h Ultrafiltration 9.5h 6.5h 10.1h Flux Rate During Constant Period Diafiltration 9.217,th 16.4 Umth 11.9 L/mth Ultrafiltration 8.0, 14.9 L/mzx h“ 20.8 Um2 x b 11.10 11.0, 8.0L/m2x h" Period of decreasing 2h 1.9h 1.5h flux rate during ultrafiltration Mass of Permeate 41993 43363 39753 Recovered Mass of Retentate 8203 6283 12253 Recovered a) For these membranes, diafiltration/ultrafiltration could not be completed in 1 day, so the extra shutdown step added morewater (immeasured)tothe process, sincethetmit cannot be started dry. b) Flux rate recorded for each of the two processing periods. 145 100 parts Whole Bean Flour 87.02 parts dry matter 19.6 parts protein + 698 parts water @ 4°C + 5.3 parts 3M NaOH to achieve pH 9 V and hold for 1 hr. 803.3 parts Bean Slurgy Nmnfge @ 4700 x g for 15 min. 202.8 parts Starch and 600.5 arts Bean Protein Extract Hull Precipitate 22.0 parts dry matter - 65.8 pans dry matter 12.0 parts protein 61.3% ofProtein 7.5 parts protein 383% “m + 1.8 parts 3M HCL to achieve pH 7 * Diafiltered (adding 143 parts water) & ultrafiltered against the following membranes (corrected data) Permeate Retentate 6 000 MW 6 000 MW 602.1 parts parts 7-5 part3 dry matter- ”7'8 12.7 parts dry matter ‘. 1-4 Parts 9'0“" 9.6 parts protein 49.2% ofProtein 50 000 MW 50 000 MW 621.8 parts , 103.8 parts 7.9 Ms dry matter 12.3 parts dry matter 1-5 9th protein 9.2 parts protein @ 1> '- 200 000 MW 200 000 MW 570.0 parts 175.4 parts 8-1 partsdry matter- 12.1 parts drymatter 1.9 parts protein 9.0 parts protein FIGURE 2: Mass Balance Diagram for Phase 2 Aqueous Extraction and Ultrafiltration Processing of Navy Bean Protein 146 Mass Balance: A comparison of the actual dry matter and protein recoveries using the three different membranes to ultrafilter the extract is shown in Table B in the Appendix. The actual dry matter recoveries varied from 83 to 92%, due to the variable amount of retentate held in the UP unit. The technician became more skilled at recovering this material with the third process (200,0MW membrane). Thus, the wet weight, dry matter and protein masses shown in Figure 2 have been corrected to equalize recoveries among the three membranes so as to make the results comparable. As would be expected, more of the dry matter and protein partitioned into the permeate with the larger molecular weight cutoff membrane (Figure 2). Protein recovery in the retentate was from 26 (200,00MW) to 49% (6,000MW) of the protein in the navy bean flour starting material. The protein content (dry basis) of the retentate was 74% to 75% (Table B, Appendix). These yields and protein contents were higher than those observed in Phase 1. A list of the samples freeze-dried or spray-dried for shipment is given in Table 4. TABLE 4: List of Samples to be Shipped to Client 147 Sample Identification DryingMethod " Amount Milled Navy Bean Flour not applicable 4483 From Phase 1 Starch and Hull Residue, Rep 1 FD 149.03 Starch and Hull Residue, Rep 2 FD 149.03 Permeate, Rep 1 FD 9.83 Permeate, Kg: 2 FD 6.3g Retentate, Rep 1 FD 17.33 Retattate, Rep 2 FD 1933 From Phase 2 Starch and Hull Residue FD 286.33 Permeate, 6,000MW FD 43.13 Permeate 50,000MW FD 47.73 Permeate 200,000MW FD 48.93 Retentate, 6,000MW SD 29.53 Retentate 50,00MW SD 33.23 Retentate 200,000MW SD 43-2L ' FD = Freeze-dried; SD - Spray Dried 148 APPENDIX B: Compositional Analysis of Preliminary UF Processed Fractions The composition of freeze dried fi'actions obtained from the POS process trial (Appendix A) were characterized. UF processed fiactions (permeate and retentate) obtained from 60,000 MW, 50,000 MW and 200,000 MW) membrane cutofl‘ were analyzed using SDS-PAGE for proteins, inductively coupled plasma emission spectroscopy for mineral content, and HPLC for sugar content. Data for each UF fraction and relative compositional component shifts were noted. SDS-PAGE gels demonstrated protein shifts occurring between permeate and retentate fi'actions at each designated molecular weight cutoff. The data indicate more protein was retained or concentrated in the retentate than partitioned into the permeate. As the molecular weight cutofi' increased from 6,000 to 200,000MW, more protein was subsequently retained in the retentate. These results are consistent; however, the size of protein is difiicult to discern since SDS- PAGE was utilized and large proteins were split into smaller sub—units by the detergent action of SDS. These results would appear to suggest that larger molecular weight proteins are retained as molecular weight cut-off is increased. Further analysis with ACID-PAGE is necessary to verify these conclusions. There was little variability in the data between gel replications. Mineral partitioning was demonstrated. Generally, monovalent water soluble minerals (notably potassium, phosphorus and sodium) had a higher concentration in the permeate than in the retentate. These results are consistent with membrane dialysis expectations. However, the fact that divalent salts (particularly manganese, copper and iron) were higher in concentration within the retentate fraction suggests metal complexes with proteins and retention. This result was not observed for calcium and zinc, which 149 selectively partition with the permeate. These data appear to be inconsistent with generalized association of the divalent binding hypothesis. Partitioning of sugars was influenced by membrane size and dependent on sugar complexity. SDS-PAGE FOR UF FRACTIONS CONTAINING WHOLE BEANS: STARCH & HULL, 6,000 MW, 50,000 MW, AND 200,000 MW NIEMBRANES Gel Calibration (15% Acrylamide Gel) Standards MW Mon (mm) Myosin 200000 4.8 B-Galactosidase 1 16250 1 1.2 Serum Albumin 66200 20.2 Ovalburnin 45000 36.0 Carbonic Anhydrase 31000 56.0 Trypsin Inhibitor 21500 73 .0 Lysozyme 14400 79.0 Dye Front 132.5 250 200 - -‘ .. Y = 552940 * x'0-56'4 -29915 f» ‘. 2 g 3 1w .. \‘ R = 0.997 3 '6 - E 5 100 4 O 2 .. . . 50 ~ ‘ " ----- _. ..... . o I l I I I l Migration (mm) 150 SDS-PAGE FOR UF FRACTIONS CONTAINING WHOLE BEAN: STARCH & HULL, 6,000 MW, 50,000 MW, AND 200,000 MW MEMBRANES R R R R In m m m i l 2 i 2 2 6 _l 5 o II—II-l 6 I 3‘ '2 — 1g ' 14 a 25 g 21 9 F — 23 'fi 35 —. 32 — 7’ - if 45 35 :7;9;.u; a 3‘ , - 6 _ 42 M l - 76 M A B i C D E F M g 0 E8 E i 3 ii E3 3% 3% ag g E g $33 636’ 3’5: 9,33 «gulf § STARCH & 151 HULL, 6,000 MW, 50,000 MW, AND 200,000 MW MENIBRANES SDS-PAGE FOR UF FRACTIONS CONTAINING WHOLE BEAN K mmmmmmm m a u w m n 00:20qu _ a8 85:23” F 32 89% __ 67 H 67 6 1 N a2 8355M 1— G 32 25.8 .n E :4 N 92 Snug—om 32 80.3 g 92 338qu D 32 08.8 Haw—qfi‘ E “42 "I41 ‘ R C :33 seesaw Rm ll . - - 1 , Rm 2 6 7 m n - «5.. n. cocouoCom E 2: x .0599: 9528 n m 152 MINERAL DISTRIBUTION AMONG UF FRACTIONS OBTAINED USING 6,000 MW, 50,000 MW, AND 200,000 MW MEMBRANES ALUMINUM magnum "ml 150] “3130 120 114 100 PPM PPM .3 25 1.5 0 0 4M . . . , o . . e Whole Bean 6.000 50,000 200,000 ml. 8.01” 50.1110 2111.000 "mm Size Momma Sin EJWhoIeBo-n IPermute Omit-re lVlholeBem uPermeate 13m MOLYBIUM COLBOLT CONTENT ‘0 - 31.6 29 9 2'5 1 2-10 2.00 . ' 2 r 7° PPM 54 m .87 1 1 0.51 0 o r . , 0 o . T , Whole Bean a 000 50.000 200.000 2: 390° 50.1110 200.000 Membrane Size Mernbrate Size ElWholeBe-t IPermeate [Retentate ambient IPenneate [Instant-ta 153 MINERAL DISTRIBUTION AMONG UF FRACTIONS OBTAINED USING 6,000 MW, 50,000 MW, AND 200,000 MW NIEMBRANES MMAKSE CONTENT BARNM CONTENT s73 0.1.3 2 1 set Whole Boat 6.1!!) Whole Bean 6,000 50.000 200.000 Membrane Size Montana Size I Mole lean I Poms-ah D m I Mote learn I PM BORON CONTENT n l 2720 PP" 004 one 048 [.01 ‘31 .47 o [:1 r ' 1* E 1 fi “a In- I.” 50,”. asses Murtbrane Size IWhole been I Perms-h E1 Rebutt- Z'jff, 1.14 1.1s 1.17 'J {1'1 .1 'T ‘. 012 a. m ms 1 -- ._ 705 us . U V I V I f 7 V ' 50M 200.000 Um MINERAL DISTRIBUTION AMONG UF FRACTIONS OBTAINED USING 6,000 MW, 50,000 MW, AND 200,000 MW MEMBRANES COPPER CONTENT zss 3“ 31.1 as 25.1 t. m 721 0 1E . . —1 the Bean 6,000 50,000 200,011) Mentrane Size n m Bean u Perntsate 1:1 Rm MAGNESIUM CONTENT - 3.440 1.030 1.4103“ I.” the Bean 6,000 50.000 200.0111 Merritrane Size IWholeBe-t IPermeate am 5101 1W1 IMN CONTENT 811 415 zmc CONTENT 188.. “.7 “.8 32.1 WholeBem us ’ I] A 6,110 50 (ID 200,011) Mentrane Size IWholeBe-t IPennade um 155 MINERAL DISTRIBUTION AMONG UF FRACTIONS OBTAINED USING 6,000 MW, 50,000 MW, AND 200,000 MW MEMBRANES POTASSIUM CONTENT 00,000 - PPM ”ii E..- 0 Menbrane She000 I Whole lean I Permeate D W PIIOSPIIORIIS CONTENT 15m 1 “1°. .mjfl PPM .000 5.5!) 0 . . . Whole Beat 6 000 50 (1!) 200,000 Membrane Size IWhoie Bean I Permeate D m CALCIIM CONTENT "I‘_lm 072.1 WhobBeII 8I_l.m Mentrane Size(mo IWhoieIean IPannsaI 0m SOIIIIM CONTENT 1m l ”A. . Iili ”“3“" MWraneSfl‘m 2°°-°°° IWhoieIe-I IPennsah Elm 156 PHASE I. MINERAL and TRACE MINERAL CONTENT (ppm) P K Na Ca 1:13 Whole Bean 5580 :1: 44.6 15000 :1: 198 7.28 1:06 1670 :l: .23 1490 i .39 Starch & Hull Residue Rep 1 4840 :1: 28.1 8190 :t 46.5 1690 i 55.9 2640 i 14.9 1710 :t 12.6 Rep 2 4430 :1: 25.0 8140 :1: 40.9 1810 :1: 54.8 2370 i 21.7 1450 i 8.9 Retentate Rep 1 11500 i 40.7 69800 i 737 17300 i 51.3 1250 :l: 6.49 3190 :l: 8.06 Rep 2 12200 i 48.6 71300 :1: 826 17600 d: 22.3 1310 i 3.99 3280 :t 9.18 Permeate Rep 1 11200 :1: 6.06 71800 :1: 752 17600 :t 56.9 1120 :l: 2.4 3240 :l: 7.89 Rep 2 10400 i 7.03 70900 :1: 707 17300 i 8.75 1280 :t 5.8 2960 :1: 4.46 Fe Zn B Ba Cu Whole Bean 78.5 :t .46 32.1 i .3 8.46 i .09 1.88 d: .03 7.21 1.05 Starch & Hull Residue Rep 1 60.1 i: .32 14.2 :1: .13 8.03 :t .17 2.96 i: .03 4.22 :t .09 Rep 2 56 :t .20 13.9 :1: 1.4 7.32 i .39 2.63 :l: .01 4.25 :l: .03 Retentate Rep 1 126 j: .71 142 i .43 30.9 :t .15 .43 :t .01 5.65 d: .04 Rep 2 64.2 r: .42 156 :1: .41 29.4 i .30 .45 :l: .02 14.9 :t .09 Permeate Rep 1 103 1.09 165 j: .63 32.1 i .21 .40 i .01 5.28 i .05 Rep 2 88.4 i: .26 150 :1: .59 33.2 i .14 1.63 1.03 8.25 :t .06 Co Mo AI Cr Mn Whole Bean 0 0 7.56 i .32 0 2.05 1: .07 Starch & Hull Residue Rep 1 - - 15.9 :t .58 O 12.8 :1: .04 Rep2 - - 11.81.61 - 9361.16 Retentate Rep 1 3.92 :1: .16 58.3 i. 20 39.3 d: .23 28 i .17 20.5 i .05 Rep 2 2.41 i .09 36 i .60 94.9 :1: .22 16.9 :1: .14 12.2 :t .09 Permeate Rep 1 2.89 1:05 44.9 i. 08 23.1 :t .07 23.6 :1: .14 17 j: .07 Rep 2 2.78 1.08 39.4 i .49 83.1 :t .71 24 :1: .28 14.8 i .07 157 SOLUBLE SUGAR ANALYSIS ULTRAFILTRATION ULTRAFILTRATION RETENTATE FRACTION PERMEATE FRACTION SUGAR WHOLE 6,000 50,000 200,000 6,000 50,000 200,000 BEAN (mw) (mw) (mw) (mw) (mw) (mw) GLUCOSE 1.3 1.90 0.79 0.53 5.64 2.09 2.19 SUCROSE 0.54 6.71 0.19 0.18 0.64 4.64 5.99 RAFFINOSE 0.67 1.21 0.65 0.88 0.94 1.04 1.22 STACHYOSE 0.89 1.37 0.32 0.28 0.40 1.21 1.35 158 APPENDIX C: Abstracts and Results of Professional Oral Presentations Made During the Course of Study Selected formal scientific presentations were made during the undertaking of this dissertation research. These papers were presented at national and international conferences and provided opportunity to critically evaluate and assess the data in a technical manner. The results presented were based on the pre-test trial (POS) samples obtained on subsequent analytical data used to characterize fractions. Further, results of the peach shake product formulation (Study Two) were presented. These oral presentations are provided in slide format within Appendix C 1-4 to illustrate the scope of research discussions undertaken. C-l: Bean Improvement Cooperative (BIC), University of Nebraska, Lincoln, NE, Nov. 4-6, 1991 Paper: Composition of Ultrafiltration Produced Navy Bean Protein Fractions Poster: Evaluation of Common Dry Beans (Phaseolus Vulgaris L.) as a Protein Source in Weaning Foods 023 Fourth, ASEAN Food Conference, Jakarta, Indonesia, Feb 19, 1992 Paper: Assessment of Navy Bean Protein Fractions Produced Utilizing Ultrafiltration Technology C-3: Institute of Food Technologist, New Orleans, LA, June 1992 Paper: Ultrafiltration Process Characterization Used For Dry Bean Protein Fractionation C-4: Bean Improvement Cooperative (BIC), Boise, Idaho, Nov. 2-4, 1993 Paper: Development of a Fortified Beverage Utilizing Dry Bean Protein (Phaseolus Vulgaris L.) Isolated Using Ultrafiltration Technology 159 APPENDIX C-lA: Bean Improvement Cooperative (BIC), University of Nebraska, Lincoln, NE, Nov. 4-6, 1991 COMPOSITION OF ULTRAFILTRATION PRODUCED NAVY BEAN PROTEIN FRACTIONS Bolles, A.D., Uebersax, M.A., and Hosfieid, G.L. Gerber Products Company, Fremont, Michigan; Department of Food Science and Human Nutrition, Michigan State University, East Lansing, Michigan; Department of Crops and Soil Sciences, Michigan State University, East Lansing, Michigan. ABSTRACT - PAPER Dry beans (Phaseolus vulgaris L.) have potential for utilization as a protein source in weaning foods. Navy beans were processed using ultrafiltration (UF) to produce protein enhanced fractions (retentate). Composition of fractions was evaluated to assess functional and nutritional value as an ingredient in formulated foods. 160 115337.001 GERBER/Bouts tom/91 SLIDE r COMPOSITION OF ULTRAFILTRATION PRODUCED NAVY BEAN PROTEIN FRACTIONS 0 AD. Bolles, Gerber Products Company 0 MA. Uebersax, Michigan State Univesity o G.L. Hosfield, USDA/Michigan State University 135337.021 GERBER/nouns 10118191 SLIDE 2 WEANING BEVERAGE/CEREAL e Protein added as an alternative protein source to dairybasedcaseinand soybeanprotein sources 0 Specific for “village grown” protein type in Costa Rica 0 Protein added to 10 to 15% protein level a Product packaged in 250ml Tetra-Brik aseptic package, nitrogen flushed headspace 35331026 7 GERBE LLES 1072931 . SLIDE 3 PRODUCT DEVELOPMENT GUIDELINES RDA LEVELS Catgggry Ages Protein ( 3) Children 1 to 3 l6 4 to 6 24 7 to 10 28 ,‘135337002 GERBER/Bouts 10718791 sun“ 0 Dry beans (Phaseolis vamp-is L.) have potential for utilization as a protein source in weaning foods 0 Navy beans were processed using ultrafiltration (UF) to produce protein enhanced fractions (retentate) 0 Composition of fractions were evaluated to assess functional and nutritional value as an ingredient in formulated foods OBJECTIVES 0 To obtain mass balance information (dry matter, wet matter and nitrogen) for extraction and ultrafiltration processing of been protein 35331003 GERBER/301.138 10118191,, SLIDES 135337.004 WEB/3011.115 1M8l91 SLIDE 6 OBJECTIVES 0 To prepare approximately 503 each of the three spray-dried bean protein concentrates, prepared using one of three molecular weight cut-off membranes during ultraliltration processing '35337005 GERBER/noun 1M8191 51.10117 OBJECTIVES o Utilize isolated bean protein in product formulations to enhance nutritional value of commercially prepared infant weaning foods 35337006 GERBER/BOLLES 1071801 SLIDE 8 . PHASE I - MATERIAL PREPARATION - Seafarer navy beans (Phaseolus vulgan‘s L) were obtained from Michigan foundation seed 0 Pin mill whole bean into bean flour conducive for ultrafiltration . BS337.W7 VGERBERIBOLLES 10118191 SLIDE 9 PHASE I - STARTING MATERIAL SOKG navy beans (Seafarer) i 1 Pin mill (or hammer mill) 1 49kg whole bean flour Proximate analysis 352337.008. GERBER/3011.38 10118191 SLIDE i0 PHASE II - MASS BAIANCE Preparation of protein extract Ultrafiltration with 50,000 mw membrane to washout low molecular material - Wet fraction weights and samples for the determination of dry matter and protein content will be taken throughout the process. The bean starch fraction, the permeate and ulbafiltered bean protein will be freeze dried 161 PHASE 11 MASS BALANCE I 300g whole bean flour 1 l 2400g bean slurry I Stir for 1 hr at 40c I Centrifuge 4700 x g for 15 min I Sample for D.M. & N assays \ , , , .~ , 7 , , , , V V , W , t , , w v r. n , . y ' 'I' ‘ ‘ ' ' ' " 11 “‘“ M‘ ' ‘L ' ' ' w PHASE II- MASS BALANCE Mgwholebanflan' +21wgmterat4°C 1 mm ~Mgbeanslurry - Stirfrrlhat4°C - Centrifuge‘flngfOI'ISnin - SumfleforanzNasuys ~7fllgbeanstuch ~1650 gbmpmtdnextnct admin-dale - SunlefanM&Nmys 0 Meta-D.M81Na-Iys 0 Email)de 0 Frocudry (mmmuhgsaMm-vm ~187gn'eaedfledbelnstardi , 3&1: 13 ' PHASE II- MASS BALANCE Mgwholebeenflom' +21wgmat4°C - Adjustmpfl9 ~mgbeanslurry 0 Stirforlhat4°C - CentfihgeflflngorlSmin - SarnplcforDMJLNmys ~750gbeunstarch ~1650 gbanpmtehextnct &hullrosidue - SunflofornM.&Nmys - Sample forI).M. & N assays - malmtion (1 hr) and ultrafiltration 0 Mull-y (to4ll)ni)uing50,mmwrmmbnne ~187gfi'ewedriedbeanstarch ~1250gpermeate - SampleforD.M.&Nmys - hoaedry ~30ng 162 115337.912 . *“ GERBER/Roms 19113191. suns)“ PHASEII-MASSBALANCE ~7$gheustuth 46$ gheIi mm 82 lnl will: 0 Saw it DM &N may! - Surpbh‘nMarNaIuys - mum-law - home my (u) Mud) “mill! m m ~187 g lime died hem starch ~13) gm ~4II gm 0 Sunk it DM 6: Nays 0 Sunk l - hone dy 0 line ri'y ~60 g Iva: d'bd hen putt: wgaufldl W , 35331013 GERBER/BOLLES 10118191 SLIDE 15 , '3533‘1.014 GERBER/BOLLES 10118191 SLIDE 16 PHASE II - ULTRAFILTRATION MEMBRANE PHASE III - ULTRAFILTRATION MEMBRANE SEPARATION TEST 0 Bean protein extracts subjected to three selective size lSOOg whole bean flour mw membranes to isolate/concentrate protein I +105 kg water at 4°C . Isolated protein freeze dried and utilized as a high 1 Adjust to pH9 quality protein source in formulated infant weaning 12kg bean slurry food 0 Stir l h at 4°C 0 Centrifuge 4700 x g for 15 mi 35331015 GERBER/ROLES 10118191 SLIDE 17 PHASE II] - ULTAFILTRATION MEMBRANE TEST 1500 2 whole bean flour + 10.5 k 2 water at 4°C ' Adjust to pH 9 12 kg bean slurry 0 Stir for 1 I1 at 4°C ° Centrifuge 4700 X g for 15 min ~3750 g bean starch & hull residue 0 Freeze dry a portion 82 discard remainder ~200 2 freeze dried bean starch 163 3533’ 7.016f-§f ' icmmo“ ' ' ‘ PHASE 111 - ULTRAFILTRATION MEMBRANE TEST 1500 g whole bean Ilaar + 10.5 kgwaterat4°C - Adjust to pH 9 12 kgbeanslurry - Stirfarlhat4°C - Centrifuge 47”ngor 15min ~3750 g beaa starch ~8250 g beau protein extract &hull residue - Divideinto3portiaas (~2750geach) 0 Freeze dry a parthn & - Diafllter (1 hr) and altraflltratisa each discard remainder 2750 g portion with a dill'ereat sire UF ~200 g freeze dried bean starch membrane 322:... "115337.617 ' ' " 01111112111301.1311 111/18191 e _‘1_...,2 5111111119 PHASE II] - ULTRAFILTRATION MEMBRANE TEST 1500 a whole bean flour + 10.5 k 2 water at 4°C 0 Adjust to pH 9 12 I12 bean slurrv - Stir for 1 h at 4°C - Centrifuge 4700 X g for 15 min ~3750 g bean starch ~8250 a: bean protein extract 82 hull residue - Divide into 3 portions (~2750 2 each) - Freeze dry a portion & - Diafilter (1 hr) and ultrafiltration each discard remainder 2750 1: portion with a different size UF ~200 2 freeze dried bean star membrane ~4500 2 permeate - Freeze dry ~46 g solids 164 r ‘ ‘1: ‘7. .2‘.V“4.(;,5 ‘. 72" .3511 5,1 M... . - PHASE III- ULTRAFILTRATION TEST l500gwholebeant|ow +10.5kgwaterat4°C - Adjusttopfl9 12 Irgbeanslurry - Stirforlhat4°C - Centrifhge4700ngor15ndn 4750gbeanstareh ~8250gheanprotdlextraet & hull residm - Divide Into3 portions (~2750 g each) - Freezedryaportiona 0 I)iafllter(1hr)anduitraiiltrationeach discardrerminder 2750gportionwithadifferentsiuUF ~200gfreeaedriedbeanstarch 4500 g permeate 490 g of each retentate ' M d"! ' SP“! dry ~46 g solids ~50 g of each spray dried bean protein concern-ate fl: '1: '- : ‘1 , [0‘ V a, 5.31-'12:; 1': . - ‘ . . .. .~ ' - 33“} SEMIPERMEABLE MEMBRANE PARTITIONING ULTRAFILTRATION MEMBRANE SEPARATION o Permeate - solvent and relatively low molecular weight II F Penna“? solutes -. riei (”Imbk m salt) 0 Retentate - impermeable higher molecular weight " " ‘ I “ components Retentate (protein concentrate) ULTRAFILTRATTON SYSTEM I I U I ' PRoxiMATE ANALYSES e DDS MINILAB 10- Seafarer o Membrane: 50,000 MW cut-off (Pin-Milled) (monitored flux rate) Moisture 11.3% . Permeate: Freeze dried for analyses Protein 23.2% . Retentate: Spray dried for formulations Fat 1.1% Ash 3.9% 165 APPENDIX C-lB: EVALUATION OF COMMON DRY BEANS (Phaseolus vulgaris L.) AS A PROTEIN SOURCE IN WEANING FOODS BOLLES, A.D. UEBERSAX, M.A., AND HOSFIELD, G.L. Gerber Products Company, Fremont, Michigan; and Department of Food Science and Human Nutrition, Michigan State University, East Lansing, Michigan ABSTRACT - POSTER Potential use of dry beans in follow-up formula as weaning foods targeted for older infants and small children was calculated with established industry standards. The benefits of these bean-based products are that older babies have nutritional requirements intermediate between those met by either starter formulas or adult foods. I66 PROJECT OVERVIEW The objective of this project is to develop high quality bean protein sources suitable for the development of nutritional beverages and/or cereals. The final formulation should contain sufiicient protein to meet or exceed the Recommended Dietary Allowances (RDA) as stated by the National Academy of Science. Products to be developed will utilize existing foods culturally accepted in the Costa Rican diet - mainly wheat/rice combinations. The fact that beans (Phaseolus vulgaris L.) are culturally accepted and locally grown makes them an ideal protein source. Whole bean meal does not deliver adequate protein levels and has associated undesirable antinutritional factors and oligosaccharide fractions. Subjecting bean meal to ultrafiltration technology (UF) removes the nutritional negatives and concentrates the protein, which will be substituted in the wheat/rice foods at levels that are organoleptically accepted and economically feasible. 167 RECOMMENDED DAILY ALLOWANCES (RDA) FOR CHILDREN AT 10% ISOLATED BEAN SUBSTITUTION [ Cstagery I Age I WT HT 1 RDA ‘ Protein Percent ] (years) (kg) (cm) (8) (8’1008) (%) Children 1 to 3 13 9O 16 15.2 100 4 to 6 20 112 24 15.2 60 7 to 10 28 132 28 15.2 50 Bean Protein Substitution in Rice/Wheat Cereal 1 5.2 15 - Total Protein 9 (%) 10 _ 5 i‘ 0 .. l Rice! Wheat 10% s“"“"' Bean Substitution I Whole Bean Meal [:1 Isolated Bean Protein 23% (Protein) 90% (Protein) 168 CEREAL 355?? EQRMUEAHON H tAnalysuper 1006 g: \ Control 1 e. . . , Venue Substitution _. _ . , Protein (3) 7.5 (5) (10) (15) Fat (g) 0.7 0.7 0.7 0.7 Carbohydrates (g) 77.0 73.0 71.0 67.0 Ash (g) 3.1 3.1 3.1 3.1 Total Solids (g) 88.0 88.0 88.0 88.0 Calon'es per 1003 344.0 344.0 344.0 344.0 Hhole Bean Meal Isolated Bean Protein————————-> Blend < IEAI BASED FOODS FOR CHILDREN Product Application Concepts - Strategies ) ( Dry Rice/Wheat Flour [Product Optione] I . Flavor Ad uncts (Banana. hocolate. Vanilla) Dry Cereal D Canister ( 2 flash) u:1 (Haterzfllx) Home preparation Dry Beverage 10:1 (Hoter:Mlxl Home preparation I [Ready-To-Use Beverage] Ultra High Temperature (152 Solids) ASEDtic Brik Pak 169 CALORIC DISTRIBUTION (%) OF MACRONUTRIENTS IN CEREAL BASED FORMULATIONS IN ISOLATED BEAN BROTEDJ [Nutrient I I__ . , .%BEAN,SUB_SHTIJIION , _ . I Rice/Wheat (0) (5) (10) (15) Control Protein 8.7 13.0 17.6 22.0 Fat 1.8 1.8 1.8 1.8 Carbohydrates 89.5 85.2 80.6 76.2 AMINO ACIDS PROFILE (mg/100g) FOR CEREAL BASED INGREDIENTS L,thlgm7l«1‘hr I also lieu-l II: It. mall. ah: I. val I Wheat 16 281 446 747 223 126 534 417 Rice 66 234 282 516 234 108 300 420 Bean 70 335 437 671 577 78 429 568 CONCLUSIONS 1. It is desirable to subject whole bean meal to UP Technology in order to provide a high quality ingredient necessary for optimal food formulations. 2. Substitution of 10% isolated bean protein in wheat/rice food provides A) 100% RDA for children 1 to 3 B) 60% RDA for children 4 to 6 C) 50% RDA for children 7 to 10 3. Levels of substitution above 10% isolated bean protein yields A) Products which have excessive protein levels B) Products which are not economically feasible and are organoleptically unacceptable 4. The ability to formulate foods with isolated bean protein in alternate gain based foods allows for amino acid complementation, which improves the nutritional value of the food. 170 REFERENCES Ang, H.G., Kwik, W.L., And Lee, CK. 1986. Ultrafiltration Studies Of Foods: Part 1 - The Removal Of Undesirable Components In Soymilk And The Effects On The Quality Of The Spray-Dried Powder. Food Chemistry 20: 183-199. Harpe, Judson, M. And Tribelhom, Ronald E. 1985. Comparison Of Relative Energy Costs Of Village-Prepared And Centrally Processed Weaning Foods. Food And Nutrition Bulletin 7 (4): 54-60. Marero, L.M., Payumo, E.M., Aguinaldo, A.R., And Homma, S. 1988. Nutritional Characteristics Of Weaning Foods Prepared From Geminated Cereals And Legumes. Journal Of Food Science 53 (5): 1399-1402. Marero, L.M., Payumo, E.M., Librando, E.C., Lainex, W.N., Gopex, M.D., And Homma, S. 1988. Technology Of Weaning Food Formulations Prepared From Germinated Cereals And Legumes. Journal Of Food Science 53 (5): 1931-1395, 1455. Owen, George M., Md. 1987. Interaction Of The Infant Formula Industry Mth The Academic Community. American Journal Of Clinical Nutrition 46: 221-225. Reddy, N. Snehalatha, Wagharme, S.Y., And Pande Vrjaya. 1990. Formulation And Evaluation Of Home-Made Weaning Mixes Based On Local Foods. Food And Nutrition Bulletin 12 (2): 138-140. Redgwell, Robert J. And Turner, Noel A. 1986. Pepino (Solanum Muricatum): Chemical Composition Of Ripe Fruit. Journal Of The Science Of Food And Agriculture 37: 1217-1222. Sathe, Shridhar, K., Deshpande, S.S., And Sulunkhe, D.K Dry Beans Of Phaseolus. A Review. Part 1. Chemical Composition: Protein, Crc Critical Reviews In Food Science And Nutrition 20(1): 1-45. Scrimshaw, Nevin S. 1980. A Look At The Incaparina Experience In Guatemala. The Background And History Of Incaparina. Food And Nutrition Bulletin 2(2): 1-2. Taha, Fakhriya S., Mohamed, Samira S., And El-Nockrashy, Ahmed S. 1986. The Use Of Soya Bean, Sunflower And Lupin Seeds In The Preparation Of Protein Bases F o Nutritious Beverages. Journal Of The Science Of Food And Agriculture 37 : 1209-1216. Ulloa, J.A., Valencia, M.E., And Garcia, Z.H. 1988. Protein Concentrate From Chickpea: Nutritive Value Of A Protein Concentrate From Chickpea (Cicer Arientinum) Obtained By Ultrafiltration And Its Potential Use In An Infant Formula. Journal Of Food Science 53(5): 1396-1398. 171 Valencia, M.E., Troncoso, R., And Higuera, I. 1988. Linear Programming Formulation And Biological Evaluation Of Chickpea-Based Infant Foods. Cereal Chemistry 65(2): 101-104. 172 APPENDIX 02: Fourth, ASIAN Food Conference, Jakarta, Indonesia, Feb 19, 1992. Albert D. Bolles, Mark A. Uebersax, and George L. Hosfield Assessment of Navy Bean Protein Fractions Produced Utilizing Ultrafiltration Technology Gerber Products Company, Fremont, Michigan; Department of Food Science and Human Nutrition; and Department of Crops and Soil Sciences, Michigan State University, East Lansing, Michigan, USA. ABSTRACT - FOURTH ASIAN WORLD FOOD CONFERENCE Navy beans (Phaseohrs vulgaris L.) were processed using ultrafiltration (UF) to produce protein enhanced fractions. The fractions were assessed for physical/chemical firnctionality and nutritional potential as an ingredient in formulated foods. The results indicated that these fractions have potential application as high quality/neutral components suitable for weaning food product development. 173 A54340270‘ERB ERIBO’ 'LLEs 611/92 SLIDEZ I COMMITTEE MEMBERS Edible Dry Beans I Dr. M.R. Bennink I Dr. W. Haines The High Fiber Food I Dr. G.L. Hostield I Dr. M.A. Uebersax of the 90’s I Dr. ME. Zabik W026 GERBERIBOLLES 611192 ' SLIDE 3‘ ‘4 Asa-4.020 GERBER/B01118 611/92 ‘ SLIDE 4] I Protein Analyses I Pilot Scale Production I POS- Benchtop DDS Ultrafiltration Systan I Formulation (Membrane Screening) I Processing I MSU - Mineral Analyses I Packaging I Gerber - Protein and Sugar Analyses I Bean Cowper: CRSP I APV - Production Ultrafiltration System I Michigan State University I Purdue University I University of Costa Rica I Gerber Products Company I Fremont, USA I San Jose, Costa Rica 165434.018 GBRBER/BOLLES 6III92 ‘ SLIDBS ii A5434. 016 GERBER/301138 611192 SLIDE 6 w] . .YANDPBOCESSEDMANS ....FROSTDAMAGEDBEANS I Intrinsic (Beany Profile) I Late Season Growth I Induced (Field and Storage) I Early Frost I Extraneous (Chemical Taint) I Physiological Damage ”434.015 GERBER/ROLES 611192 SLIDE 7 "A5643“ ' GERBER/3011.38 612192 ‘ SLIDE 8 ' ...COLOR AND FLAVOR DEFECTS . IECHNICAL WMG _ .. I Chilling Injury I Rapid Cooling (fully cooked) I Cellular/Membrane Disruption I Precooked (frozen/Dehydrated) I High Metabolic Rate I Microwavable I Leaching/Soaking I Separation Technology (specialty I Browning ingredients) I Dead Sea I Frost Damaged Beans I Cull and Substandard Beans I Split Beans ‘ A5434 err-"cannaR/Bouss" “ 611192 fl SLIDE Io . . BEANPROTEIN CHARACTERISTICS I Protein Isolation and Classification I SDS-PAGE Peptide Pattern I Amino Acid Analyses A5434. ’ ” 010 GERBER/Bo" . LLBs 611192 ' "‘ SLID Err" A5434. f ”009'GERBER/BOLLES 611192 SLIDE 12 i .- , .IN-VI’I‘RO PROTEIN DIGESIIBIIJTY . .. B PROTEIN FRACTIONATION . I Albmnin (Water Soluble) I Globulin (Salt Soluble) (G l and G 11) I Prolamin (Alcohol Soluble) I Glutelin (Alkali Soluble) AOAC Method 43.265 In-Vitro Protein Digestion - for C-PER 174 A5434.“ GERBER/3011.38 611192 SLIDE 13 A5434. 006 GBRBER130LLES 611192 SLIDE 14 l IN-VIIRO PROTEIN DIGES'IIBILITY ~ . . . FLATULENCE FACTORS I 10 mg N Sample in 10 mL DD Water I Oligosaccharides (Stachyose, Raflinose) I Adjust pH to 8.0 @ 37C I Proteins I Add 1 ml. Enzyme Solution A I Indigestible Starch (Trypsin, Chymotrypsin and Peptidase) I Fiber Hold @ 37C for 10 min I Add 1 MI Enzyme Solution B (Pronase B), Hold @ 55C for 19 min A5434. 003 eBRBBwBOLLBs' 611192 Sum: 13 115434.004 GERBBmBoLLBs 611192 sum 16] I Micro I Macro Zn,Cu,Se,Mn,A1,B,Ba P,K,Na,Ca,Mg,Fe l A5434 " “ 00312131111 ' moms 611192 sun“ B17 A5434 002’ GERB‘ BRIBOLLBS 611192 SLID B 18] I Protein Analyses I Protein Analyses I Pilot Scale Production I Formulation I Processmg' I Packaging 175 APPENDIX C-3: Institute of Food Technologist, New Orleans, LA, June 1992 ULTRAFILTRATION PROCESS CHARACTERIZATION USED FOR DRY BEAN PROTEIN FRACTIONATION A.D. Bolles, MA. Uebersax, S.M. McCurdy & G.L. Hosfeld ABSTRACT: Navy beans (Phaseolus vulgaris L.) were processed using ultrafiltration to produce protein enhanced fractions. The fractions of isolated protein, trypsin inhibitor and other antinutritional compounds are co-extracted with the protein. The molecular weight of undesirable compounds suggests that they can be removed from the solubilized bean protein utilizing ultrafiltration. The subsequent isolated proteins can be finther utilized and have potential application as a high quality/neutral component suitable for food product development. Beans were pin milled (13.9% moisture) to a flour (particle size of 700m), mixed with tap water at 4°C and adjusted to pH 9(3N NaOH). The slurry was mixed for a hour and centrifuged (4700 xg, 15 min) to recover protein extract. The extract was adjusted to pH 7 (3N Hcl) . Three different Mw polysulfone membranes were evaluated: 6,000, 50,000 and 200,000 MW. Diafiltration flux rates varied among the membranes reflecting the expected water flux rates. Ultrafiltration flux rates were more variable, reflecting plugging of the membrane during processing (6,000 MW, 14.9 W x h; 50,000 Mw, 20.8 L/M x h; 200,000 MW, 8.9 W x h). The Flux rate dropped during the last 1.5 to 2 hours, as the retentate concentration increased. 176 A comparison of the dry matter and protein recoveries was conducted to establish a mass balance for the process. Greater dry matter and protein partitioned into the permeate with the larger molecular weight cutofl‘ membrane: 7.2% protein (6,000 MW) to 9.7% protein (200,000 MW). Protein recovery in the retentate ranged from 49% (6,000 MW) to 46% (200,000 MW) of the protein in the original navy bean flour. The total protein content (db) of the retentate was 74% to 75%. The results demonstrated that selection of bean protein based on molecular weight can be achieved using ultraliltration. Protein fractions should be further characterized to assess their efl'ectiveness as an ingredient resource for applications in the food formulation. 177 55434.00} cmmsouas 611192 SLIDE 1 > '. » PROCESS WW OF- DRY BEAN - WIN FRACTIONS PRODUCED BY 0 ULTRAFETRA-TION A.D. Bolles, Gerber Products Company M.A. Uebersax, Michigan State University S. McCurdy, POS Corporation G. L. Hosfield, USDA/Michigan State University 99999 3302: EDIBLE DRY BEANS THE HIGH FIBER FOODS OF THE 90‘s 'fiWflfiW? rmmmmmmammmmmummmn = ~gpum - Genetics Agronomic Production Post-Harvest Handling Processing Technology End-User Preparation Afinw7‘wcfimmmmmmaw2amm7 Latin America 46. % w. Eurooo 5 Sum- Latin American 46.5% E. Europe - North America 7.5% W. Europe 6.3% M. East 4% Africa 24.3% Asia 4.5% 9,3 ‘5‘ k3 ERM- ' 1' 31. ES. 5111928] 13' ’ 325:}; ' ‘ Dry and Process Beans Intrinsic (Beany Profile) Induced (Field and Storage) Extraneous (Chemical Taint) ”mmmwmuwawaumn . ECONOMIC ENHANCEMENT Frost Damaged Beans Cull and Substandard Beans Split Beans FROST DAMAGED BEANS Late Season Growth Early Frost Physiological Damage Color and Flavor Defects Chilling Injury Cellular/Membrane Disruption High Metabolic Rate Leaching/Soaking Browning Dead Seed Rapid Cooking (frilly cooked) Precooked (frozen/dehydrated) Microwavable Separation Technology (specialty ingedients) I78 Physical Preparation Postharvest Storage Changes Biological Antinuuients Digestibility Flatulaice Factors t». ume-«W ‘11-, “3°. '11" i8: -2 ”1‘?‘ Soaking Germination/Sprouting &. Fermentation Dehulling Roasting Extractive Pre-treatments Cooking (Autoclave; Extrusion) Drying (Freeze; Sin)", Dunn-Dying) Membrane Processing Membrane Ultrafiltration Tubular Hollow Fiber Membrane No Heat (No Sugar Carmaelization) No Oxidation Improved Flavor Retention (Delicate Flavor) Reduced Discoloration Long Term Storage Albumin (Water Soluble) Globulin (Salt Soluble) (G 1 and G l 1) Prolamin (Alcohol Soluble) Glutelin (Alkali Soluble) Protein Analyses Pilot Scale Production I Formulation I . Processrng I Packaging Protein Analyses Pilot Scale Production I Formulation I Processing I Packaging Bean Cowpea - CRSP I Michigan State University Purdue University University of Costa Rica I San Jose, Costa Rica 179 Asmara r ’ GERBERIBOLLES 612192 SLIDE 19" ”434.020 GERBER/BOLLES 61119 SLIDE 20 n ' 0110er CHARACTERISTICS , 2100mm Protein Isolation and Classification Phytic Acid SDS-PAGE Peptide Pattern Protein Amino Acid Analyses Starch CERBERIBOLLES 611192 SLIDE 21' 115434.022 i ' ' GERBER/BOLLES 611192 SLIDE 22 . 7. 021" 1 («VHROPROTEINDIGESTIBILITY , .. 1 AOAC METHOD 43.265 In-Vitro Protein Digestion — for C-PER . IN~VITRO PROTEIN DIGESTIBILITY 10 MG 11 Sample in 10M] DD Water Adjust pH to 8.0 @ 37C Add 1 mL Enzyme Solution A (Trypsin Chymotrypsin and Peptidase) Add 1 mL Enzyme Solution B (Pronase B), Hold @ 55C for 19 min ' 415434.023 “ ’ ‘ ERBER/BOLLES 612192 sum: 2:1 ‘A5434.024 GERBER/nouns 612192 sum: 24 . . ANTINUTRIENTS ‘ ._ FLATULENCE FACTORS Protease Inhibitors Oligosacchardies (Stachyose, Raflinose) Lectins Proteins Amylase Inhibitors Indigestible Starch Procyanidin Interactions Fiber Biological Factors | ”434.025 . ' GERBER/ROMS 611192 SLIDE ‘25 It ”434.026 ‘ GERBER/HOLMES 611192 SLIDE 26 ' SOLUBLE COMPONENTS OF BEANS .. MINERAL ANALYSES Sugars (Glucose, fructose, Sucrose) Inductively Coupled Plasma Oligosaccharides (raflinose, Stachyose, Varbascose) Emission Spectrometry Minerals (ICP) Phytate AS43402? ‘ GLRBBRIBOLLBSSW 31.111327 145434.018 61113111111101.1103 611192 sum: 30 . ., . ,MINERALANALYSIS _ 3. . ._MINERALCONTENT Inductively Coupled Plasma (Micro) Emission Spectrometry (ICP) Zn, CU, Se, Mn, Al, B, Ba AS4343” ' i ' GERBER/30111138612192 SLIDE” POS - Benchtop DDS Ultrafiltration System (Membrane Screening) MSU - Mineral Analyses Gerber - Protein and Sugar Analyses APV - Production Ultrafiltration system DATA AND RESULTS PRESENTED WITHIN APPENDIX A & 3 180 APPENDIX C-4: Bean Improvement Cooperative (BIC), Boise Idaho, Nov 2-4, 1993 lBolles, A.D., 2Uebersax, M.A., and 3Hosfreld, G.L. 1Tropicana Products, Inc., Bradenton, Florida, 2Department of Food Science and Human Nutrition, Michigan State University, East Lansing, Michigan; 3Department Crops and Soil Science, Michigan State University, East Lansing, Michigan DEVELOPMENT OF A FORTIFIED BEVERAGE UTILIZING DRY BEAN PROTEIN (Phaseolus vulgaris L.) ISOLATED USING ULTRAFILTRATION TECHNOLOGY, A.D. Dry beans (Phaseolus vulgaris L.) were fractionated to isolate specific proteins. This was accomplished utilizing ultrafiltration technology separating proteins by molecular weight (200,000 MW or greater). The isolated protein was incorporated in a flavored dairy based beverage at total fortification level of 12 percent. The formula was subjected to an aseptic thermal process and packaged in 200111] Tetra Pak®, which allowed for a dry shelf life of twelve months. 181 DEVELOPMENT OF A FORTIFIED BEVERAGE UTILIZING DRY BEAN PROTEIN IPHASEOLUS VULGARISI ISOLATED USING ULTRAFILTRATION TECHNOLOGY A.D. Belles VP/Director Research & Development Tropicana Products, Inc. MA. Uebersax MR. Bennink G.L. Hosfield 182 OBJECTIVES 1) To obtain mass balance information (dry matter, wet matter and nitrogen) for extraction and ultrafiltration processing of bean protein. 2) To assess compositional changes using three selective ultrafiltration membranes (6,000, 50,000 and 200,000 molecular weight). 3) To develop a shelf-stable fortified beverage suitable as a weaning food. ULTRAFILTRATION PROCESSING ....... ADVANTAGES o No heat (no sugar/Caramelization) . No oxidation 0 Improved flavor retention (delicate flavor) . Reduced discoloration . Long term storage ULTRAFILTRATION Semipermeable Membrane Partitioning o Permeate: Solvent and relatively low molecular weight solutes . Retentate: Impermeable higher molecular weight components 183 ULTRAFILTRATION MEMBRANE SEPARATION l u_p_ I ------- > PERMEATE (Soluble Sugars, Salt) I I I I I I v RETENTATE (Protein Concentrate) ULTRAFILTRATION SYSTEM . Plate and Frame: DDS Mimilab 10 . Polysulfone Membrane: 6,000, 50,000 and 200,000 MW (monitored flux rate) . Diafiltration (D): Constant volume 184 RESEARCH MATERIAL 50 KG NAVY BEANS (SEAFARERI < ————......_-_— 49 KG WHOLE BEAN FLOUR PROXIMATE ANALYSIS PROXIMATE ANALYSIS SEAFARER (PIN-MILLEDI ° MOISTURE I 12.9% 0 PROTEIN 22.5% 0 FAT 1.5% 0 ASH 4.4% 0 FIBER | 4.5% MATERIAL PREPARATION . Seafarer navy beans (Phaseolus Vulgaris L.) were obtained fiom Michigan Foundation Seed. 0 Pin mill whole bean into bean flour conductive for ultrafiltration. 185 ANALYTICAL METHODS PROTEIN: 0 SDS - Page 0 Amino Acids - Lectin (lmino-Assay) SOLUBLE SUGARS: 0 HPLC MINERALS: 9 Plasma Emission PROXIMATE ANALYSIS: 0 Moistures 0 Fat 0 Protein 0 Carbohydrates 0 Ash Molecular Weight (kD* 10) 186 Gel Calibration (15% Acrylamide Gel) Standards MW Migration (mm) Myosin 200000 4.8 B-Galactosidasc 1 16250 1 1.2 Serum Albumin 66200 20.2 Ovalbumin 45000 36.0 Carbonic Anhydrase 3 1000 56.0 Trypain Inhibitor 21 500 73 .0 Lysozyme 14400 79.0 Dye Front 132.5 250 200 ‘ 150 ‘ 100? 50‘ Y = 552940 * 1'0““ -29915 I ‘, R2 = 0.997 Migration (mm) 187 SDS - PAGE GEL A STANDARD WHOLE BEAN BAND # % IOD STANDARD MW % IOD STANDARD MW RM 8.39 200+ 13.18 2 200+ 0.56 6.15 6 183 7.51 5.76 7 173 0.78 28.18 18 74 0.76 25.40 32 48 9.52 6.65 35 46 0.35 14.67 42 41 13.03 0.99 15.32 ©W\IO\MAWN— 0.19 19.26 23.22 188 SDS - PAGE GEL A STANDARD 6,000 MW PERMEATE 6,000 MW RETENTATE cont’d BAND # % IOD RM MW % IOD RM MW % IOD RM MW 8.39 200+ 22.76 1 200+ 2.20 2 200+ 0.56 198 77.24 10 145 1.27 5 193 7.51 183 2.43 8 164 0.78 173 1.18 10 145 0.76 164 5.21 13 120 9.52 125 25.68 18 74 0.35 106 6.54 20 65 13.03 62 2.19 21 62 0.99 55 1.98 28 52 15.32 45 25.75 33 47 0.19 39 10.26 41 42 19.26 31 2.83 53 34 23.22 18 0.48 59 44 10.25 64 26 1.75 72 21 .— \OOONGUI-h-WN 189 SDS - PAGE GEL A STANDARD 6,000 MW PERMEATE 6,000 MW RETENTATE cont’d BAND # % IOD RM MW % IOD RM MW % IOD RM MW 1 8.39 2 200+ 35.29 1 200+ 6.30 2 200+ 2 0.56 4 198 50.70 10 145 2.85 6 182 3 7.51 6 183 5.61 8 164 4 0.78 7 173 1.88 10 145 5 0.76 8 164 7.62 13 120 6 9.52 12 125 28.30 19 68 7 0.35 14 106 7.46 22 60 8 13.03 21 62 2.78 24 57 9 0.99 25 55 0.82 25 55 10 15.32 35 45 2.18 29 51 11 0.19 45 39 2.44 33 47 12 19.26 56 31 5.22 35 46 13 23.22 76 18 2.65 37 44 14 5.60 41 42 15 4.62 43 40 16 1.14 48 37 17 3.43 53 34 18 0.28 55 32 19 0.93 55 30 20 4.72 65 26 21 0.53 67 24 22 0.45 70 22 23 2.21 73 20 190 SDS - PAGE GEL 3 STANDARD 200,000 UFP 200,000 UFR BAND % IOD RM MW % IOD RM MW % IOD RM MW # l 8.39 2 200+ 19.25 2 200+ 4.74 1 200+ 2 0.56 4 198 7.90 5 193 6.57 4 198 3 7.51 6 183 3.42 7 173 2.44 7 173 4 0.78 7 173 23.19 11 135 1.62 9 154 5 0.76 8 164 13.28 12 127 2.75 12 127 6 9.52 12 125 19.59 18 74 2.82 16 92 7 0.35 14 106 0.00 20 65 36.36 20 65 8 13.03 21 62 6.52 24 57 5.84 24 57 9 0.99 25 55 2.07 27 53 0.58 27 53 10 15.32 35 45 4.53 32 48 3.79 30 50 11 0.19 45 39 2.51 35 46 12 19.26 56 31 5.39 37 44 13 23.22 76 18 5.42 42 41 14 3.61 45 39 15 1.50 50 36 16 3.20 54 33 17 0.57 58 30 18 1.00 62 28 19 6.01 66 25 20 0.06 71 21 21 1.50 75 19 22 1.20 84 I3 191 AMINO ACID WHOLE 6,000 MW 6,000 MW 50,000 MW 50,000 MW 200,000 MW 200,000 MW BEAN PERMEATE RETENTATE PERMEATE RETENTATE PERMEATE RE'I‘ENTA'I'E AA (mg of protein) ALIPHATC HYDROCARBON Glycine 122.19 46.57 199.61 25.50 130.47 59.03 I 10.15 Alanine 120.39 77.32 165.74 35.22 25.29 77.67 102.59 Valine 107.67 44.98 196.00 19.89 76.05 I60.45 120.76 Leucine 149.75 79.37 219.89 37.04 75.29 86.51 125.32 lsoleueine 82.10 19.68 140.61 8.06 58.27 40.23 77.14 ALCOHOL Serine 164.08 17.46 222.30 14.19 77.69 42.34 122.08 Threonine 136.30 41.34 258.71 23.16 126.55 71.66 145.37 ACID Aspartic Acid 298.83 83.19 185.45 50.71 24.94 86.66 138.12 GIutamic Acid 486.77 540.46 509.29 316.73 133.46 439.37 293.65 BASIC Arginin 88.65 160.75 174.30 81.53 46.51 118.59 100.81 Histidine 71.12 35.35 121.99 19.39 17.22 38.88 71.4] Lysine 48.20 9.65 65.10 4.36 8.65 13.02 23.95 SULFUR Methionine 17.67 21.51 43.18 10.45 50.76 17.04 8.58 Cystine AROMATIC PhenylaIine 84.80 45.66 199.52 20.09 51.36 48.45 107.94 Trosine 34.86 469.29 77.18 212.43 52.58 15.14 81.08 HETEROCYCLIC Proline 141.03 86.77 321.96 48.60 259.37 96.03 202.12 Hydroxpmpline 5.92 9.43 7.62 4.21 5.15 8.30 3.84 200,000 MW Pilot WHOLE BEAN 200,000 MW 200,000 MW Plant Scaleup PERMEATE RETENTATE Amino Acid AA img/gof protein) AA (mg of protein) AA Qng/g of protein) ALIPHATC HYDROCARBON Glycine 122.19 1897.28 1354.89 Alanine 120. 39 3096.60 1947.27 Valine 107.67 2455.51 2475.74 Leucine 149.75 3472.66 3164.63 lsoleueine 82. I0 1738.61 2135.54 ALCOHOL Serine 164.08 1185.78 1617.29 Threonine 136.30 1591.95 1911.88 ACID Aspartic Acid 298.83 652.15 559.15 Glutarnic Acid 486.77 2948.55 2222.26 BASIC _A_rginin 88.65 2377.21 1863.55 Histidine 71 . 12 802.12 1366.67 Lysine 48.20 1960.57 820.24 SULFUR Methionine 17.67 837.50 734.67 Cystine 316.71 35.55 AROMATIC || Phenylaline 84.80 1629.41 3056.70 Trosine 34.86 664.53 1891.58 HETEROCYCLIC Praline 141.03 1716.44 2301.02 Hydroxpropline 5.92 192 LECTIN ANALYSIS FOR 200,000MW PILOT PLANT Lectin Positive Whole Bean Positive Retentate Positive Permeate Positive LECTIN ANALYSIS FOR 6,000, 50,000 AND 200,000MW MEMBRANES Ultrafiltration Permeate Fraction J Lectin Positive Whole Bean Positive 6,000 MW Retentate Positive 50,000 MW Retentate Negative 200,000 MW Retentate Negative SOLUBLE SUGAR ANALYSIS 1 Ultrafiltration Retentate Fraction Sugar 6,000(mw) 50,000(mw) 200,000(mw) 6,000(mw) 50,000(mw) Glucose 1.90 0.79 0.53 5.64 2.09 Sucrose 6.71 0.19 0.18 0.64 4.64 Raflinose 1.21 0.65 0.88 0.94 1.04 Stachxose 1.37 0.32 0.28 0.40 1.21 200,000(mw) 2.19 5.99 1.22 1.35 193 NUTRIENT PROFILE OF PEACH SHAKE WITH LOWFAT YOGURT %USRDA FOR AGES 12-48 MO - HIGH PROTEIN EFFICIENCY NUTRIENT BEAN CONTROL Calories 112.8 108.1 Protein g 3.40 3.14 CHO g 19.63 19.20 Fat g 2.30 2.08 Ash g 0.89 0.89 Total Solids 26.22 25.13 Calcium mg 105.00 100.00 Iron mg 0.31 0.19 Sodium mg 51.87 54.44 B-Carotene IU 0.00 105.6 Vit C mg 4.70 5.24 Thiamin mg 0.052 0.036 Riboflavin mg 0.185 0.180 Niacin mg 0.33 0.69 Zinc mg 2.00 1.69 Name: Aroma 194 Bae#: Aroma Aroma Fullness = no aroma 7 = intense, complex aroma Fullness Fullness Smoothness/Blendedness 1 = disjointed, unblended, inappropriate aromas 7 = harmonious blend of appropriate flavors Smoothness/Blendedness SmoothnesslBlendedness Flavor Flavor Flavor Fullness 1 = thin, watery 11v 7 = intense, complex flv Fullness Fullness Smoothness/Blendedness 1 = disjointed, unblended, inappropriate aromas 7 = harmonious blend of appropriate flavors Smoothness/Blendedness Smoothness/Blendedness Overall Freshness 1 = no fresh flavor (cooked, old, stale, immature fruit, spoiled, etc. - describe below) 7 = very fresh flavor (with no “unfrcsh” characteristics Overall Freshness Tart - Sweet 1 = Tart, no sweet 4 = Equal intensities of tart & sweet 7 = Sweet, no tart Overall Freshness Tart - Sweet Tart - Sweet Bitterness 1 = none 7 = strong Bitterness . Bitterness Viscosity 1 = water 7 = Chocolate syrup Viscosity Viscosity Other Descriptors: Other Descriptors: Other Descriptors: General formofscale: l=none 3=slight 5=moderate 7=strong I95 SENSORY DIFFERENCE TESTING (Panel research conducted at Gerber Products Company, F reemont, MI) Test Obiective: To determine if a significant exists in Peach Yogurt Shake when the protein isolate is added. Products Tested: Peach Yogurt Shake: Control: Current Standard Formula Variable: Current standard formula plus protein isolate Test Methodology: The products were served at room temperature using a dual trio test design. Red lights were used to mask a slight color difference between the products. Results: Out of a total of 30 judgments, 20 were correct. This is significant at the 95% confidence level. Conclusion: Panelists found a significant difl‘erence between the products tested. Acceptance testing can be done to determine if the products, although difi‘erent, are equally liked. I96 / Daily Key Nutrient Requirements* Per Pound Adult vs. Child (1 -3 years) Adult men Adult women Children (1—3years) 1.38 Vitamins Minerals A C B6 Iron Calcium Zinc "01°?“ (IU) (“18) (m8) (m8) (m8) (m8) 026/0.37 29 0.34 0.01 0.10 5.7 0.09 0.32/0.47 36 0.43 0.01 0.13 7.2 0.11 " Based on current U.S.RDA labeling values 1 Infant Energy Needs Calories 900 850 800 750 Baby Weight, lbs - Infant Need I Formula Only I First Foods + Formula ’orlsula = Wm“ 1 liar I97 NUTRIENT PROFILE OF PEACH SHAKE WITH LOWFAT YOGURT %USRDA FOR AGES 12 TO 48 MO - HIGH PROTEIN EFFICIENCY ammo“ ~ ~' ‘ ' BEAN" 'r — ‘ comm. , . DISTRIBUTION - f _. . Protein % 12.1 11.6 Fat% 18.3 17.3 CHO % 69.6 71.1 NUTRIENT PROFILE OF PEACH SHAKE WITH LOWFAT YOGURT %USRDA FOR AGES 12 TO 48 MO - HIGH PROTEIN EFFICIENCY Nutrient ” L 1 L N ' Actual Labell " ' . Milk] Calories 220.0 110.0 Protein g 6.0 9.0 CHO g 38.0 12.0 Fat g 5.0 2.0 Calcium mg 25.0 30.0 Iron mg 4.0 0.0 Sodium mg 100.0 13.0 Vitamin C mg 10.0 4.0 Thiamin mg 4.0 8.0 Riboflavin mg 30.0 30.0 Niacin mg 6.0 0.0 Vitamin B6 mg 8.0 0.0 Zing mg 30.0 0.0 CONCLUSIONS . Bean protein can be effectively concentrated to 75% protein with ultrafiltration. . SDS gel electrophoresis was used to characterize protein from various membrane/molecular weight cut-offs. 198 Partitioning of soluble sugars and minerals indicated increased concentration in the permeate and decreased concentration for the retentate. Ultrafiltration technology was directionally suitable for separation of active lectin. Functional bean protein was incorporated into a fortified beverage targeted for small children and achieved action standards for nutrition and quality. LIST OF REFERENCES Abbey, B.W. and Nkanga, U.B. 1988. Production of high quality weaning products from maize-cowpea-crayfish mixtures. Nutrition Reports International 37 (5): 951- 957. Abudu, LA. and Akinyele, 1.0. 1990 and Youssef 1987. Adams, M.W. 1972. On the quest for quality in the field bean. In “Nutritional Improvement of Food Legumes by Breeding.” Protein Advisory group of the United Nations System, United‘Nations, NY. Adewusi, S.R.S., Orisadare, B.O., and Oke, CL. 1991. Studies on weaning diets in Nigeria. 1. carbohydrates sources. Cereal Chemistry 68(2): 165-169. Akpapunam, M.A and Markakis, P. 1979. Oligosaccarides of 13 American cultivars of cowpeas (Vigna sinensis). J. Food Sci. 44: 1317. Anderson, J.W. 1984. Hypocholesterolemic efl‘ects of oat-bran or bean intake for hypercholesterolemic men. Am. J. clin. Nutr 40: 1146-1155. Anderson, J.W. 1985. Cholesterol-lowering efi‘ects of canned beans for hypercholesteolemic men. Clin. Res. 33: 871A (abstr). Anderson, J.W. 1986. Influence of high carbohydrate diets on glucose tolerance of normal and diabetic effect of pre-cooked instant preparations of beans and potato in normal and diabetic subjects. Am.J. Clin. Nutr. 43: 30-36. Anderson, J.W. and Chen, W.J.L. 1983. Legumes and their soluble fibers: Effect on cholesterol-rich lipoproteins, in unconventional sources of dietary fiber. Am. Chem. Soc., p. 49-57. Anderson, J.W., Story, L., Sieling, B. and Chen, W.J.L. 1984. Hypocholesterolemic efl‘ects of high-fiber diets rich in water-soluble plant fibers: long term studies with oat-bran and bean-supplemented diets for hypercholesterolemic men. J. Can. Diet Assoc, 45: 140—149. Aug, H.G., Kwik, W.L., Lee, GK, and Theng, CY. 1986, Ultrafiltration Studies of Foods: Part l-The removal of undesirable components in soymilk and the efl‘ects on the quality of the spray-dried powder. Food Chemistry 20: 183. Antunes, PL, and Sgarbieri, V.C. 1990. Effect of heat treatment on the toxicity and nutritive value of dry bean ( Phaseolus vugaris var. Rosinha G2) proteins. J. Agric Food Chem. 28: 938-938. 199 200 Antunes, P.L., Sgarbieri, V.C., and Garruti, RS. 1979. Nutrification of dry beans (Phaseolus vulgaris L.) by methionine infusion. Journal of Food Science. 44 (5): 1302-1305. Aquilera, J .M., Lusas, B.W., Uebersax, MA. 1982. Development of food ingredients fi'om navy beans (Phaseolus vulgaris L.) by roasting and air classification. J. Food Science 47: 1151-1154. Augustin, J., Beclg C.B., Kalbfleish, G., and Kagel, LC. 1981. Variation in the vitamin and mineral contents of raw and cooked commercial Phaseolus vulgaris L. classes. J. Food Sci. 46: 1701. Baker, R.W., Cussler, E.L., Eykamp, W., Koros, W.J., Riley, R.L., and Strathmann, H. 1990. “Membrane Separation Systems - A Research and Development Needs Assessment - Final Report, Volume 11,” March, U. S. Department of Energy, Office of Energy Research, Oflice of Program Analysis, DOE/ER/30133-Hl, Vol. 2 of 2. Barker, RD.J., Derbyshire, E., Yarwood, A., and Boulter, D. 1976. Purification and characterization of the major storage proteins of Phaseolus vulgaris L. seeds and their intracellular and cotyledonary distribution. Phytochem. 15: 751. Becker, R., Olson, A.C., Frederixk, DR, [(011, S., Gumbmann, M.R, and Wagner, JR, 1974. Condition for the autolysis of alpha-galactosides and phytic acid in California small white beans. J. Food, Sci. 37: 766 Bennink, MR. and Srisuma, N. 1989. Digestibility of dry legume starch and protein. Proceedings of the World Congress on Vegetable Protein Utilization in Human Foods and Animal F eedstuffs. Edited By Thomas H. Applewhite. American Oil Chemist’s Society Champaign, IL. (J. Amer. Oil Chem.) 266-272. Berot, S., Gueguen, J ., and Berthaud, C. 1987. Ultrafiltration of fababean (Vcia faba L.) protein extracts: Process parameters and functional properties of the isolates. Lebensm-Wiss. U - Technol, 20(3): 143-150. Bollini, R. and Chrispeels, M.J. 1978. Characterization and subcellular localization of vicilin and phytohemagglutinin, the two major reserve protein of Phaseolus vulgaris L. L. Planta 142: 291. Boloorforooshan, M. and Markakis, P. 1979. Protein supplementation of navy beans with sesame. J. Food Sci. 44(2): 390-391. Boulter, D. 1981. Protein composition of grains of the Leguminosae. In “Plan Proteins for Human Food,” p. 43. Kluwer Academic Publishers, Boston, MA. 201 Braharn, J.E. and Bressani, R. 1985. Effect of bean broth on the nutritive value and digestibility of beans. Journal of the Science of food Agriculture 36(10): 1028- 1034. Bray, UT. 1968. Reverse Osmosis Purification Apparatus, US. Patent 3,417,870, Dec. 24. Bressani, R., Elias, LG, and Valiente, AT. 1963. Effects of cooking and of amino acid supplementationon the nutritive value of black beans (Phaseolus vugaris, L.) Brit. J. Nutri. 17: 69. Bressani, R. 1983 a. Research needs to rip-grade the nutritional quality of common beans (Phaseolus vulgaris L. ). Qualitas Plantarum Plant Foods for Human Nutrition. 32 (2): 101-110. Bressani, R 1983b. World needs for improved nutrition and the role of vegetables and legumes. Shanhua,Taiwan, Asian Vegetable Research and Development Center. 10‘“ Amriversary monograph Series. AVRDC Publication 83-185. Bressani, R 19830. Efi‘ects of chemical changes during storage and processing on the nutritional quality of common beans. Food Nutr. Bull. 5: 23. Bressani, R. and Elias, LG. 1977. The problem of legume protein digestibility. In Nutritional Standards And Methods of Evaluation for Food Legume Breeders, Eds. J.H. Hulse, K.O. Rachie and L.W. Bressani, R. and Elias LG. 1980. The nutritional role of a polyphenols in beans. In “Polyphenols in Cereals and Legumes,” J .H. Hule (Ed), p. 61. IDRC, Canada. Bressani, R and Elias, LG 1982. Reduction of digestibility of legume proteins by tannins. J. Plant Foods 4: 43-55. Bressani, R, Elias, LG. and Braham, J .E. 1982. Reduction of digestibility of legume proteins by tannins. J. Plant Foods 4: 43-55. Bressani, R. Elias, LG, and Navarette, DA. 1961. Nutritive value of Central American beans. The essential amino acid content of samples of black beans, red beans, rice beans and cowpeas of Guatemala. J. Food Sci. 26: 525. Bressani, R, Navarrette, DA, and Elias LG. 1984. The nutritional value of diets based on starchy foods and common beans. Qualitas Plantarum Plant Foods for Human Nutrition. 34(2): 109-115. 202 Brillouet, Jean-Marc and Carre, B. 1983. Composition of cell wall fiom cotyledons of Pisum sativum, Vrva faba and Glycemic max. Phytochemistry 22: 841. Brown, J.W.S., Ma, Y. Bliss, FA, and Bliss, T.C. 1981a. Genetic variations in the subunits of globulin-1 storage protein of French bean. Theor. Appl. Geent. 59: 83. Bukovac, M.J., Rasmussen, HP, and Shull, VB. 1981. The cuticle: Surface structure and function. Scanning Electron Microsc, III: 213. Burkitt, DP. 1977. Food fiber: Benefits from a surgeon’s perspective. Cereal Foods Worlds 22: 6. Burkitt, D.P., Walker, A.RP. and Paiter, NS. 1972. Efl‘ect of dietary fiber on stools and transit times, And its role in the causation of disease. Lancet 2: 1408-1412. Burr, H.K. 1973. Efl‘ect of storage on cooking qualities, processing and nutritive value of beans. In: Nutritional Aspects of Common Beans and Other Legume Seeds as Animal and Human Foods. W.G. Jafi‘e, Ed. Arch Latioaner. Nutr. Cadotte, J .E. and Francis, PS. 1971. “Ultrathin Semipermeable Membrane,” US. Patent 3,580,841, May 25. Cadotte, J.E., Roselle, L.T., Petersen, F.J., and Francis, RS. 1070. “Water Transport Across Ultrathin Membranes of Mixed Cellulose Ester and Ether Derivatives, “John Wiley and Sons Publishers, New York, NY. Calloway, DR 197 3. Gas-forming property of food legumes. In “Nutritional Improvement of Food Legumes by Breeding” (Milner, M.(Ed) John Wiley and Sons, New York, NY, p. 263-75). Calloway, DH. and Burroughs, SE. 1969. Effect of dried beans and silicone on intestinal hydrogen and methane production in man. Gut 10: 180. Carabez, TA, Paredes, L0. and Reyes, MC. 1989. Microstructure of cotyledon cells from hard-to-cook Common beans. Starch 41 (9): 335-339 Carpenter, K.J. 1981. The nutritional contribution of dry beans (Phaseolus vulgan's L.) in perspective. Food Technol. 35: 77. Ceming-Beroard, J. Sapsonik, A., and Guibot, A 1975. Carbohydrate composition of horsebeans (Vicia faba L.) of difl'erent origins. Cereal Chem. 52: 125. Chang, KC. and Satterlee, LB. 1981. Isolation and characterization of the major protein fiom great northern beans (Phaseolus vulgaris L.). J. Food Sci. 46: 1368. 203 Chang, KC. and Satterlee, LB. 1982. Chemistry of dry bean proteins. J. Food Proc. Preserv. 6: 203. Cheryan, M. 1986. Ultrjafiltration Handbook, Technomic Publishing Co., Inc, Lancaster, PA. Cheryan, M., and Schlesser, J .E. 1978. Performance of a hollow fiber system for ultrafiltration of aqueous extracts of soybeans. Lebensm. - Wiss. u. - Techno]. 11(2): 65-69. Coffey, D.G., Uebersax, M.A., Hosfield, G.L., and Bnmner, J .R. 1985. Evaluation of the emagglutinating activity of low-temperature cooked kidney beans. J. Food Sci. 50: 58. Cofl‘ey, D.G., Uebersax, M.A., Hosfield, G.L., and Bennink, M.R. 1993. Thermal extrusion and alkali processing of dry beans (Phaseolus vulgaris L.). Journal of Food Processing and preservation. 16(6): 421-431. Coffey, D.G., Uebersax, M.A., Hosfield, G.L., and Bennink, M.R 1992. Stability of red kidney bean lectin. Journal of Food Biochemistry. 16(1): 43-57. Daetz, D. Planning for customer satisfaction with quality firnction deployment. Proceedings: Eighth International Conference of the ISQA, Jerusalem, Israel, November 1990. Davis, RR and Cockrell, CW. 1976. Effect of added calcium chloride on the quality of canned dried lima beans. Arkansas Farm Res. 25(4): 14. Derbyshire, E. and Boulter, D. 1976a. Isolation of legumin-like protein from Phaseolus aureus and Phaseolus vulgaris L. Phytochem. 15: 411. Derbyshire, E. and Boulter, D. 1976b. Legumin and vicilin, storage proteins of legume seeds. Phytochm. 15: 3. Deschamps, I. 1958. Peas and beans. In “Processed Plant Protein Foodstufl‘s,” A.M. Altschul (Ed), Academix Press, New York, NY. Deshpahde, SS. and Nielsen, SS. 1987. Nitrogenous constituents of selected grain legumes. J. Food Sci. 52: 1321. Devadas, Rajarnmal P., Chandrasekhar, Usha, and Bhooma, N. 1984. Nutritional outcomes of a rural diet supplemented with low cost locally available foods-III. Development and introduction of weaning foods for infants. The Indian Journal of Nutrition and Dietetics 21: 82-88. 204 Dieckert, J .A. and Dieckert, MC. 1985. The chemistry and biology of seed storage proteins. “In New Protein Foods,” Vol. 5., p. 1, Academic Press, Orlando, Fl. Diosday, L.L., Tzeng, Y.M., and Rubin, L]. 1984. Preparation of rapeseed protein concentrates and isolates using ultrafiltration. Journal of Food Science 49: 768- 776. Drumm, T.D., Gray, J .I., Hosfield, G.L. 1990. Variability in the major lipid components of four market classes of dry edible beans. Journal of the Science of Food Agriculture. 50(4): 485-497. Drumm, T.D., Gray, J .1., Hosfield, G.L. and Uebersax, MA. 1990. Lipid, saccharide, protein, phenolic acid and saponin contents of four market classes of edible dry beans as influenced by soaking and canning. Journal of the Science of Food and Agriculture. 51 (4): 425-435. Dull, G.G. and Leeper, ER 1975. Ultrastructure of polysaccharides in relation to texture. In “Postharvest Biology and Handling of Fruits and Vegetables,” N.F. Haard and D.K. Salunkhe (Ed), p. 55-61, AVl Publishing Co., Westport, CT. Eka, O.U. 1978. Chemical evaluation of nutritive value of Soya paps and porridges, the Nigerian weaning foods. Food Chemical 3: 199-206. Ekpenjong, TE, and Borchers, KL. 1980. Effect of cooking on the chemical composition of winged beans (Psophacarpus tetragonolobus). J. Food Sci. 45: 1559. Eureka, WE. Introduction to quality function deployment. Section III. In: Quality Function Deployment: A Collection of Presentations and AFD Case Studies. American Suppliers Institute Publication, 1987. Fatoki, ES. and Bamiro, F0. 1990. Levels of Na, K, Ca and Mg in infant formula and in corn-flour infant feeds. Food Chemistry 37: 269-273. Fernandez, R., Cooke, R.D., Quiros, R, Madrigal, L., Samuels, A., Aguilar, F ., and Ortiz, A. 1980. Fruit and vegetable processing and appropriate technology in Costa Rica: a case study. Tropical Science 22(2): 149-158. Finnigan, Timothy J.A. and Lewis, Michael J. 1989. Ultrafiltration of aqueous extracts of United Kingdom commercial rapeseed meal made at pH 2.0 and 11.0. Lebensmittel-Wissenschafi und - Tecnologie 22(5): 23 7-239. Fleming, S.E. 1981a and 1982. A study of relationships between flatus potential and carbohydrate distribution in legume seeds. J. Food Sci. 46: 794. 205 Fleming, S.E. 1981b. F latulence activity of the smooth-seeded field pea as indicated by hydrogen production in the rat. J. Food Sci. 47: 12. Fordham, J .R., Wells, CE, and Chen, LH. 1975. Sprouting of seeds and nutrient composition of seeds and spouts. J. Food Sci. 40: 552. Geervani, P. and Theophilus, F. 1982. lnfluence of legumes starches on protein utilization and availability of lisyne methionine to albino rats: a literature review. Nut. Abstr. Rev. Ser. A. 52: 1237. Goetz, A. and Tsuneishi, N. 1951. “Application of Molecular Filter Membranes to the Bacteriological Analysis of Water,” J. Amer. Water Wcfics Assoc, (43): 943. Goldsmith, H.L., and Mason, S.G. 1963. The flow of suspensions in tubes. J. Colloid Sci. 18, 237. Gomez-Brenes, R, Elias, L.G., Molina, M.R, de la Fuente, G., and Bressani, R. 1975. Changes in the chemical and nutritive value of common beans and other legumes during house cooking. Arch. latinoam. Nutr. Proceedings of Annual Meeting, 1973: 93-108. Goodwin, T.W. and Mercer, El. 1983. The Plant Cell Wall. 1n “Introduction to Plant Biochemistry,” 2nd edition, Pergamon Press, New York, NY. Griffin, Al and Hauser, J .R. 1991. The voice of the customer. University of Chicago Working Paper. Guiro, A.T., Sall, M.G., Kane, 0., Ndiaye A.M., Diarra, D., and Sy, M.T.A. 1987. Protein-calorie malnutrition in Senegalese children. Effects of rehabilitation with a pearl millet weaning food. Nutrition Reports International 36(5): 1071-1079. Hafex, Youssef S. and Elqadri, Saleh S 1986. Proposed weaning formula for developing countries. Nutrition Reports International 34(5): 915-919. Hall, T.C., McLeester, RC, and Bliss, F .A. 1977. Equal expression of the material and paternal alleles for the polypeptide subunits of the major storage protein of the bean Phaseolus vulgaris L. Plant Physiol. 59: 1122. Hall, T.C., Sun, S.M., Ma, Y., McLeester, R.C., Pyne, J.W., Bliss, F .A., and Buchbinder, B.U. 1979. The major storage protein of French bean seeds: Characterization in vivo and translation in vitro. In ”The Plant Seed: Development, Preservation, and Germination,” I. Rubenstein, R.L. Phillips, C.E. Green, and HG. Gengenbach (Ed). 206 Harper, J .C., and Karnel, I. 1966. Viscosity Relationships for Liquid Products, Am. Soc. Agr. Eng. Paper 66-837. Am. Soc Agr. Engrs., St. Jospeh, Mich. Harris, HE. 1969. Bird resistance in sorghum. Annual Corn Res. Conf. Proc. 24th. Chicago, IL. p. 113. Haslam, E. 1979. Vegetable tannins, Rec Adv. Phytochem. 12: 475. Hauser, J .R. and Clausing. D. The house of quality. Harvard Business Review 63-73 (May-June 1988). Holdsworth, SD. 1992. Elsevier Applied Food Science Series. Aseptic Processing and Packaging of Food Products. Hoover, R. and Sosulski, F. 1985. Studies on the firnctional characteristics and digestibility of starches fiom Phaseolus vulgaris biotypes. Starth 37: 181. Hunt, SM, and Grofi‘, J .L. Adyanced Nutrition and Human Metabolism West Publishing Company, 1990 Hutt, PB. 1993. Guide to US. Food Labelig, Thompson Publishing Group, NY, NY. Volume 1, Tab 600. Hymowitz, T., Collins, F.I., Panezner, J ., and Walker, W.M. 1972. Relationship between oil, protein, and sugar in soybean seed. Agron. J. 64: 613. Ishino, K. and Ortega, D.M.L. 197 5. Fractionation and characterization of major reserve proteins from seeds of Phaseolus vulgaris. J. Agric. Food Chem. 23: 529. Jansen, G. Richard and Harper, Judson M. 1980. Application of low-cost extrusion cooking to weaning foods in feeding programs. Part 2. Food and Nutrition Bulletin 6(2): 15-23. Karyadi, Darwin, Mahmud, Mien K., and Hermana. 1900. Locally made rehabilitation foods. The Malnourished Child, Robert M. Suskind and Leslie Lewinter—Suskind (Ed), p. 371. Vevey / Raven Press, Ltd., New York. Kay, EE. 1979. Haricot bean. In “Food Legumes,” DE. Kay (Ed), Tropical Products Institute, London, UK. Ketiku, AD. and Olusanya, J .O. 1986. Nutrient composition of multimixes for use as weaning foods in Nigeria. Food Chemistry 21: 47-56. 207 Koehler, H.H. and Burke, D.W. 1981. Nutrient composition, sensory characteristics, and texture measurements of seven cultivars of dry beans. J. Am. Soc. Hort. Sci. 106: 313. Kogure, M. and Akoa, Y. Quality firnction deployment and CWQC. Quality Progress: 16(10): 25-29 (October 1983). Kohnhorst 1990; Dukiandjiev 1987; Ken 1973; Kon and Wagner 1979 Borowska and Kozlowska 1988 and K011 and Booth 1974. Kerr, S. 1979. Effect of soaking temperature on cooking and nutritional quality of beans. J. Food Sci. 44: 1329. Korte, R. 1972. Heat resistance of phytohemagglutinins in weaning food mixtures containing beans (Phaseolus vulgaris L). Ecology of Food and Nutrition 1: 303- 307. Korytnyke, W. and Metzler, EA. 1963. Composition of lipids of lima beans and certain other beans. J. Sci. Food Agric. 14: 84]. Kusin, Jane A., Kardjati, Sri, de With, C., and Renqvist, U.h. 1984. Measured food intake and nutritional status of pre-school children in rural East Java, Indonesia. Nutrition reports International 30(3): 651-660. Lah, CL, and Cheryan, M. 19803. Protein solubility characteristics of an ultrafiltrated full-fat soybean product. J. Agric. Food Chem. 28: 911. Lah, CL, and Cheryan, M. 1980b. Emulsifying properties of a filll-fat soy protein product produced by ultrafiltration. Lebensm.-Wiss. u. - Technol. 13: 259. Lawhon, J .T., Hensley, D.W., Mulsow, D., and Mattil, K.F. 197 8. Optimization of protein isolate production from soy flour using industrial membrane systems. J. Food Sci. 43: 361. Lawhon, J .T., Mulslow, D., Cater, C.M., and Matti], K.F. 1977. Production of protein isolates and concentrates from oilseed flour extracts using industrial ultrafiltration and reverse osmosis systems. J. Food Sci. 42: 389. Liener, LE. and Thompson, RM. 1980. In vitro and in vivo studies on the digestibility of major storage protein of the navy bean (Phaseolus vulgaris). Qua]. Plant. 30: 13. Luh, B.S. Wang, C., and Daoud, H.W. 1975. Several factors affecting color, texture and drained weight of canned dry lima beans. J. Food Sci. 40: 577. 208 McBain, J .W. and Kistler, SS. 1931. “Ultrafiltration as a test for Colloidal Constituents in Aqueous and Nonaqueous Systems,” J. Phys. Chem, (35): 130. McLain, EA. 1969. Wound hollow fiber permeability apparatus and process of making the same, US. Patent 3,422,008, Jan 14. McLain, EA. and Mahon, HI. 1969. Permslective hollow fibers and method of making, US. Patent 3,423,491, Jan 21. Ma, Y. and Bliss, F .A. 1978. tannin content and inheritance in common bean. Crop Sci. 18: 201. Mahon, HI 1963. “Hollow Fibers as Membranes for Reverse Osmosis,” National Research Council Publication 942: 345-354. Mahon, HI. 1966. US. Patent 3,228,876, Jan 11. Mahon, H.I. 1966a. Permeability separatory apparatus, permeability separatory membrane element, method of making the same and process utilizing the same, US. Patent 3,228,8765, Jan. Mahon, H.I. 1966b. Permeability separatory apparatus and process utilizing hollow fibers, US. Patent 3,228,877, Jan. 11 Mahon, H.1., Mclain, E.A., Skiens, W.E., Green, B.J., and Davis, TE. 1969. “Hollow Fiber Membranes for Selective Permeation,” AIChE Chem. Eng; Proc. Smp. Ser. No. 91: 48-51. Marero, L.M., Payumo, E.M., Aguinaldo, ARM. and Home S. 1988. Nutritional characteristics of weaning foods prepared from germinated cereals and legumes. Journal of Food Science 53(5): 1399-1402. Mathew, Susan P. and Pellett, Peter L. 1986. Protein quality of homemade weaning food mixtures: 1. Biological evaluation and quality. Ecology of Food and Nutrition 19: 31-40. Mazliak, P. 1983. Plant membrane lipids: Changes and alteration during aging and senescence. In “Postharvest Physiology and Crop Preservation,” M. Lieberman (Ed), p. 123. Plenum Press, New York, NY. McLeester, R.C., Hall, T.C., Sun, S.M., and Bliss, FA 1973. Comparison of globulin proteins from Phaseolus vulgaris L. with those from Vicia faba. Phytochem. 12: 85 209 Meiners, C.R, Derise, N.L., Lau, H.C., Ritchey, SJ, and Murphy, B.W. 1976. Proximate composition and yield of raw and cooked mature dry legumes. J. Agric. Food Chem. 24: 1122. Merlo, C.A., Pedersen, L.D., and Rose, W.W. 1985. ‘Hyperfiltration/Reverse Osmosis, A Handbook on Membrane Filtration for the Food Industry,” July, U.S.Department of Energy, DOE/CE/40691- 1. Merlo, A.C., Rose, W.W., and Ewing, N.L. Membrane Filtration Handbook, National Food Processor’s Association. 1993 Metzner, AB, and Reed, J .C. 1955. Flow of non-Netonian fluids correlation of the laminar, tansition and turbulent-flow regions. Am. Inst. Chem. Engrs. J. 1, no. 4, 434-440. Mohr, C.M., Leeper, S.A., Engelgau, DE and Charboneau, BL. 1988. “Membrane Applications and Research in Food Processing: An Assessment,” August, US. Department of Energy, DOE/ID-10210 (DE88016250). Naivikul, O. and D’Appolonia, BL. 1978. Comparison of legumes and wheat flour carbohydrates. 1. Sugar analysis. Cereal Chem. 55: 913. Nene, S.P., Vakil, UK, and Screenivasn, A. 1975. Effect of gamma radiation on physico-chemical characteristics of red gram (Cajanus cazan) starch. J. Food Sci. 40: 943. Nichols, DJ, and Cheryan, M. 1981. Production of soy isolates by ultrafiltration: factors afl‘ecting yield and composition. J. Food Sci. 46: 367. Nielsen, SS. 1991. Digestibility of legume proteins. Food Technology. 45(9): 112- 114,118. Noah, N.D., Bender, A.E., Reaidi, GB. and Gilbert, R.J. 1980a. Food poisoning fiom raw kidney beans. Brit. Med, J. 281: 236 Noah, N.D., Bender, A.E., Reaide, G. B. and Gilbert, RJ. 1980b. Epidemology of food poisoning from raw red kidney beans. Br. Med. J. July 19: 236-237 Nordstrom, CL. and Sistrunk, WA. 1977. Efi‘ect of type bean, soak time, canning media and storage time on quality attributes and nutritional value of canned dry beans. J.Food Sci. 43(3): 795-798. 210 Nordstrom, CL. and Sistrunk, WA. 1979. Effect of type bean, moisture level, blanch treatment and storage time on quality attributes and nutrient content of canned dry beans. J. Food Sci. 44: 392. O’Dell, B.L., Burpo, CE, and Savage, J .E. 1972. Evaluation of zinc availability in foodstuffs of plant and animal origin. J. Nutr. 102: 653. Okubo, K., Waldrop, A.B., Iacobucci, GA, and Myers, D.V. 1975. Preparation of low-phytate soybean protein isolate and concentrate by ultrafiltration. Cereal Chem. 52: 263. Olaofe, O. 1988. Mineral contents of Nigerian grains and baby foods. Journal of the Science of Food and Agriculture 45: 191-194. Olsen, HS. 1978. Continuous pilot plant production of bean protein by extracting, centrifugation, ultrafiltration and spray drying. Lebensm.- Wiss. u. - Technol. 11: 57. Omosaiye, O. and Cheryan, M. 1978. Removal of oligosaccharides from soybean water extracts by ultrafiltration. J. Food Sci. 43: 354. Omosaiye, O. and Cheryan, M. 1979a. Ultrafiltration of soybean water extracts: processing characteristics and yields. J. food Sci. 44: 1027. Omosaiye, O. and Cheryan, M. 1979b. Low-phytate, full-fat soy protein product by ultrafiltration of aqueous extracts of whole soybeans. Cereal Chem. 56: 58. Osborne, TB. and Campbell, GE. 1989. Proteins of the pea. J. Am. Chem. Soc. 20: 348. Ott, AC. and Ball, CD. 1943. Some components of the seedcoats of the common bean, Phaseolus vulgaris L., and their relation to water retention. Arch. Biochem. 3: 189 Patel, K.M., Bedford, CC, and Youngs, CW. 1980. Amino acid and mineral profile of air-classified navy bean flour fractions. Cereal Chem. 57: 125 Pellett, Peter L. and Mamarbachi, Dalal. 1979. Recommended proportions of foods in home-made feeding mixtures. Ecology of food and Nutrition 7: 219-228. Porter, MC. 1975. “Selecting the Right Membrane,” Chem. Ergr. Prog, (71): 55. Powrie, W.D., Adams, M.W., and Pflug, U. 1960. Chemical anatomical and histochemical studies on the navy bean seed. Agron. J. 52: 163. 211 Pritchard, P.J., Dryburgh, EA, and Wilson, B.J. 1973. Carbohydrates of spring and winter field beans (Vicia faba L.). J. Sci. Food Agric. 24: 663. Pusztai, A. and Watt, W.B. 1970. Glycoprotein II. The isolation and characterization of a major antigenic and non-haemagglutinating glycoprotein from Phaseolus vulgaris L. Biochem. Biophys. Acta 207: 413. Quenzer, N.M., Huffman, V.L., and Burns, EB. 1978. Some factors affecting pinto bean quality. J. Food Sci. 43: 1059. Reddy, N. Snehalatha, Waghmare, S.Y., and Pande, Vijaya. 1990. Formulation and evaluation of home-made weaning mixes based on local foods. Food and Nutrition Bulletin 12(2): 138-140. Reddy, NR and Salunkhe, D.K. 1980. Changes in oligosaccharides during germination and cooking of black gram and fermentation of black gram/rice blend. Cereal Chem. 57: 356. Reddy, N.R., Piereson, M.D., Sathe, SK, and Salunkhe, D.K. 1984. Chemical, nutritional and physiological aspects of dry bean carbohydrates. A review. Food Chem. 13: 25-68 Reeve, RM. 1947. Relation of histological characteristics to texture in seed coats of peas. Food Res. 12: 10. Reeve, RM. 1970. Relation of histological structure to texture of fresh and processed fruits and vegetables. J. Texture Studies 1: 247 Richter, J .W. and Hoehn, H.H. 1971. Permselective, aromatic, nitrogen, nitrogen- containing polymetric membranes, US. Patent 3,567,632, March 2. Riley, R.L., Hightower, GR, Lonsdale, H.K., Los, J .F., Lyons, C.R., Mysels, K.J. and Want, DE. 1972. “Development of Reverse Osmosis Modules for Seawater Desalination,” Final Report for Office of Saline Water, NTIS Report No. PB214973. Roberts, AH. and Yudkin, J. 1960. Dietary phytate as a possible cause of magnesium deficiency. Nature 185: 823. Rockland, LB. and Radke, TM. 1981. Legume protein quality. Food Tech. 34: 79-82. Rockland, L.B., Zaragosa, E.M., and Jracca-Tetteh, R. 1979. Quick-cooking of winged beans (Psophocarpus tetragonolobus). J. Food Sci. 44: 1004. 212 Rockland, LB. and Jones, F .T. 1974. Scanning electron microscope studies: effects of cooking on the cellular structure of cotyledons in rehydrated large lima beans. J. Food Sci. 39: 342-346 Roman, Ana V., Bender, Arnld E., and Morton, Ian D. 1987. Storage stability of an infant food based on rice, cowpeas and skim milk. Human Nutrition: Food Sciences and Nutrition 41F: 23-28. Rosenthal, S. and March, A. Speed to market: Disciplines for product design and development. Boston University School of Management Manufacturing Roundtable Research Report, 1991. Rozelle, L.T., Cadotte, J .E., and MxClure, DJ. 1970. “Ultrathin Cellulose Acetate Membranes for Water Desalination,” John Wiley and Sons Publishers, New York, NY. Sahasrabudhe, M.R, Quinn, J .R., Paton, D., Youngs, CG, and Skura, B]. 1981. Chemical Composition of White Bean (Phaseolus vulgaris L.) and firnctional characteristics of its air-classified protein and starch fractions. J. Food Sci. 46: 1079. Salunkhe, D.K., Kadam, 8.8., and Chavan, J .K. 1985. Chemical composition. Ch. 3 in “Post harvest Biotechnology of Food Legumes,” p. 29. CRC Press, Inc, Boca Raton, FL. Sathe, S.K., Deshpande, S.S., Salunkhe, D.K. 1984a. Dry beans of Phaseolus: A review. Part 2. Chemical composition. Carbohydrates, fibers, minerals, vitamins, and lipids. CRC Crit. Rev. Food Sci. Nutr. 21(1) : 41-93. Sathe, S.K., Desphpande, S.S., Salunkhe, D.K. 1984b. Dry beans of Phaseolus: A review. Part 1. Chemical composition. Proteins, CRC Crit. Rev. Food Sci. Nutr. 20(1): 1-46 Sathe, SK. and Krishnarnurthy, K. 1953. Phytic acid and absorption of iron. Inc]. J. med. Res. 41: 453. Sathe, SK. and Salunkhe, D.K. 1981a. Studies on trypsin and chymotrypsin inhibitory activity, hemagglutinating activity and sugars in the Great Northern Bean. J. Food Sci. 46: 626-629. Sathe, SK. and Salunkhe, D.K. 1981b. Preparation and utilization of protein concentrates and isolates for nutritional and functional improvement of foods. J. Food Quality 4: 233-245. 213 Sathe, SK. and Salunkhe, D.K. l981c. Functional properties of the Great Northern (Phaseolus vulgaris L.) : proteins-«emulsion, foaming, viscosity and gelatin properties. J. Food Sci. 46: 71. Sathe, SK. and Salunkhe, D.K. l981d. Isolation, partial characterization and modification of the great northern bean (Phaseolus vulgaris L.) starch. J. Food Sci. 46: 952. Sathe, S.K., Lilley, G.G., Mason, AC, and Weaver, GM. 1987. High resolution sodium dodecyl sulfate polyacrylamide gel electrophoresis of soybean (Glycine max L.) proteins. Cereal Chem. 64(6): 380-3 84. Sefa-Dedeh, S. and Stanley, D.W. 1979b. The relationship of microstructure of cowpeas to water absorption and dehulling properties. Cereal Chem. 56: 379. Sefa-Dedeh, S. and Stanley, D.W. 1979a. Microstructure of cowpea variety (Adua ayera). Cereal Chem. 56: 367. Sellers, Stephen G. 1984. Diet patterns and nutritional intake in a Costa Rican community. Ecology of Food and Nutrition 14: 205-218. Sgarbieri, V.C. and Garruti, RS 1986. A review of some factors affecting the availability and the nutritional and technological quality of common dry beans, a dietary staple in Brazil. Can. Inst. Food Sci. Technol. J. 19: 202. Sgarbieri, V.C. and Whitaker, J .R. 1982. A review of some factors affecting the availability and the nutritional and technological quality of common dry beans, a dietary staple in Brazil. Can. Inst. Food Sci. Technol. J. 19: 202. Skelland, A.H.P. 1967. Non-Newtonian Flow and Heat Transfer. John Mley & Sons, New York. Snauwaert, F. and Markakis, P. 1976. Efl‘ect of germination of gamma irradiation on the oligosaccharides of navy bean (Phaseolus vulgaris L). Legersrnitt Wissensch Technol. 9:2, p. 93. Sosulski, F.W, and Youngs, CG. 1979. Yield and filnctional properties of air-classified protein and starch fractions from eight legume flours. J. Amer. Oi] Chem. 80. 56: 292-295. Srisuma, N., Songyos Ruengsakulrach, Uebersax, M.A. 1994. Starch characteristics and canning quality of four selected navy beans (Phaseolus vulgaris L.) cultivars. Starch/Staerke. 46(5): 187-193 214 Stanley, D.W. and Aguilera, J .M. 1985. A review of textural defects in cooked reconstituted legumes - the influence of structure and composition. Journal of Food Biochemistry 9: 277-323. Sullivan, L.P. QFD benefits. American Suppliers Institute Publication 1C, 1986a. Sullivan, L.P. Quality function deployment. Quality Progress 19(6): 39-50 (June 1986b). Sullivan, LP. The power of Taguchi methods. Quality Progress 12(6): 76-79 (June 1987) Swanson, B.G. and Raysid, A. 1984. Protein quality of soybean, red bean and corn tempeh. 44th Annual Meeting of the I.F.T., Abstract #310. Takayama, K.K., Muneta, P., and Wiese, AC. 1965. Lipid composition of dry beans and its correlation with cooking time. J. Agr. Food Chem. 13: 269. Tecklenburg, E., Zabik, M.E., Uebersax, M.A., Dietz, J .C., and Lusas, B.W. 1984. Mineral and phytic acid partitioning among air-classified bean flour fraction. J. Food Sci. 49: 569. Tobin, G. and Carpenter, K.J. 197 8. The nutritional value of the dry bean (Phaseolus vulgaris). A literature review. Nutr. Abstr. and Rev. 48: 920. Uebersax, M.A., and Bedford, CL. 1980. Navy beans processing: Effect of storage and soaking methods on quality of canned beans. Mich. State Univ. Agr. Exp. Sta, E. Lansing, MI. Res.Rpt. 410. Uebersax, M.A., Lee, J .P., and Hosfield, G.L. 1981. The effect of selected soak water additives on the quality of dry cranberry beans. Michigan Dry Bean Digest. 6(1): 16-17. Uebersax, M.A., and Ruengasakulrach, S. 1989. Structural and compositional changes during processing of dry beans (Phaseolus vulgaris). In: Quality Focus of Fruits and Vegetables. ACS Symposium. Ulloa, J .A., Valencia, ME, and Garcia, Z.H. 1988. Protein concentrate fiom chickpea: Nutritive value of a protein concentrate from chickpea (Cicer arietinum.) obtained by ultrafiltration and its potential use in an infant formula. Journal of Food Science 53(5): 1396-1398. Urban, Glenn L. and Hauswer, John R. Design and Marketing of New Products. Englewood Cliffs, NJ: Prentice-Hall, Inc, 1980. 215 Uwaegbute, AC. and Nnanyelugo, DO. 1987. Nutritive value and biological evaluation of processed cowpea diets compared with local weaning foods in Nigeria. Nutrition Reports International 36(1): 119-129. Valverde, Victor and Rawson, Ian G. 1976. Dietetic and anthropometic differences between children from the center and surrounding villages of a niral region of Costa Rica. Ecology of Food and Nutrition 5: 197-203. Van Buren, J .P. 1979. The chemistry of texture in fruits and vegetables. J. Texture Studies 10: 1. VanBuren, J .P. 1980. Calcium binding to snap bean water insoluble solids calcium and sodium concentrates. J. Food Sci. 45: 752. Vose, J .G. 1980. Production and functionality of starches and protein isolates from legume seeds (Field peas and Horsebeans). Cereal Chem. 57: 406. Vose, J .R, Basterrechea, M.J., Gorin, P.A.J., Finlayson AJ. and Youngs, CG. 1976. Air-classification of field peas and horse-bean flours: chemical studies of starch and protein fractions. Cereal Chem. 53: 928. Watt, RT. and Merrill, W. 1963. “USDA Handbook Number 8,” Agric. Res. Service, Washington, DC. Westemoreland, J. 1968. US. Patent 3,367,504, Feb. Zabik, M.E., Uebersax, M.A., Lee, J.P., Aguilera, J.M., and Lusas, B.W. 1983. Characteristics and utilization of dry roasted air-classified navy bean protein fraction. Journal of the American Oil Chemist’s Society. 60 (7): 1303-1308. Zsigrnondy, R. 1922. US. Patent 1,421,341, July 22. 1 l‘llillillti‘llIl 1