INFORMATION TO USERS
This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality o f the material submitted.
The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.
1.The sign or “target” for pages apparently lacking from the document photographed is “Missing Pagc(s)”. If it was possible to obtain the missing pagc(s) or section, they arc spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.
2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image o f the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of “sectioning” the material has been followed. It is customary to begin filming at the upper left hand comer of a large sheet and to continue from left to right in equal sections with small overlaps. If ncccssaiy, sectioning is continued again-beginning below the first row and continuing on until complete,
4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints arc available upon request from the Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.
University M i c r o f i l m s International 300 N.Zwb Road Ann Arbor, Ml 4B106
8403575
Song, J a e Chul
SOLVENT EXTRACTION OF LACTOSE FROM SKIM' MILK POWDER AND THE APPLICATION OF THE PROTEIN AS A REPLACEMENT FOR CASEINATE
The Ohio State University PH.D. 1883
University Microfilms International300 N. Zwb Road, Ann Arbor, Ml 48108
PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark -J .
1. Glossy photographs or oaoes i S
2. Colored illustrations, paper or print tX*
3. Photographs with dark background t/
4. Illustrations are poor copy ______
5. Pages with black marks, not original copy.
6. Print shows through as there is text on both sides of page ______
7. Indistinct, broken or small print on several pages ^
8. Print exceeds margin requirements ______
9. Tightly bound copy with print lost in spine ______
10. Computer printout pages with Indistinct print ______
11. Page(s)______lacking when material received, and not available from school or author.
12. Page(s)______iseem to b e missing In numbering only as text follows.
13. Two pages num bered I______. Text follows.
14. Curling and wrinkled pages
15. Other
University Microfilms International
SOLVENT EXTRACTION OF LACTOSE FROM SKIM MILK
POWDER AND THE APPLICATION OF THE PROTEIN AS A
REPLACEMENT FOR CASEINATE
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the
Deqree Doctor of Philosophy in the Graduate School of
The Ohio State University
hy
Jae Chui Sonq, M.S.
The Ohio State University
19A3
Readinq Committees Aoproved by
P.M.T. Hansen
J.B . Lindamood
A.C. Penq Department p f Food Science D.B. Min and N u tritio n To my father and younqest brother, Jae Chan
11 ACKNOWLEDGEMENTS
I wish to express sincere appreciation to Dr. Poul M.T.
Hansen for his advice, encouragement and support throughout my graduate studies. His valuable guidance, enthusiasm,
affections and hard working will always be kept in my mind
and will be served as a guide for me in the future.
I wish also to thank Dr. T. Kristoffersen for his
assistance during the course of the work. I am also
g r a te fu l to Drs. J.B . Llndamood, M.E. Mangino, D.B. Min and
Dr. A.C. Peng for their suggestions and constructive
criticisms on this work.
Especially, I wish to thank Dr. Han Chul, Yang,
Professor of Korea University, for his encouragement and
financial support during the course of my study.
I am also grateful to some of fine people who have
helped make my study possible: Mrs. Joann Chang's family,
Dr. K.S. Chang, S.H. Lee, Steve, Lee, Andy, Dolly, Bobbie,
Stephanie and Dr. Jang Soo, Kim, Professor of Korea
University.
Finally, I wish to share my degree with lovely wife,
son, daughter and my family in Korea.
111 VITA
Auqust 19, 1947 Born: Ulsan Banqeolin, Korea
February, 1975 B. Aqr. in Aqricultural Chemistry, Korea University, Seoul, Korea
1975-1977 Researcher, Antibiotics Fermentation and Purification, Chonq Keun Danq Pharm aceutical Co. Seoul, Korea
1977-1979 M.S. (Food Technoloqy), Korea University, Seoul, Korea
1979-1990 M.S. (B loloqical Sciences), Wriqht State Univ., Dayton, OH
1981-1983 Research Associate, Dept, of Food Science & Nutrition, The Ohio State University
PUBLICATIONS AND PRESENTATIONS
1. Sonq, J.C. and Yanq, H.C. 1978. A study on the enrichment of qlutathione content in yeast. Korean J. Appl. Micro. Bioenq. 6:75
2. Sonq, J.C. and Calcott, P.H. 1980. DNA damaqe and its repair in frozen-thawed E. coll. Third Semiannual Meetinq of the Ohio Valley Tissue Cultural A sso ciatio n , Huston Woods, OH
3. Sonq, J.C. and Hansen, P.M.T. 1983. Solubility behavior of Delactosed Non-Fat-Dry-Milk, 29th Annual North Central States Milk and Food Protein Seminar, Sept. 16-17. Michiqan State University, East Lansinq, MI
1v 4. Song J.C. 1981. DNA damaqe and its repair after freezin g and thawinq of E. c o l l . M.S. th e s is 5• Sonq J.C. and Hansen P.M.T. 1983. Solvent extraction of lactose from NFDM and application of a protein as a replacement for caseinate. In preparation
\
V TABLE OP CONTENTS
Paqe
DEDICATION...... 11
ACKNOWLEDGEMENT...... 111
VITA...... 1V
TABLE OF CONTENTS...... v1
LIST OF FIGURES...... X
LIST OF TABLES...... x11
LIST OF PLATES...... XlV
INTRODUCTION...... I
LITERATURE REVIEW...... 3
A. Solubility of Carbohydrates in Non-aqueous S o lv e n t ...... 3 1. Dielectric Constant and Polarity ...... 3 2. Solvent Structure ...... 7 3. Temperature E f f e c ts ...... 11 4. Inorganic Solvents ...... 14 5. Organic Solvents ...... 16
B. Extraction and Crystallization of Lactose in Aqueous and N6 n-aqueous Solvent System ...... 20 1. Alcohol E x tra c tio n o f Lactose from Skim Milk and Whey ...... 21 2. Choice of Solvent ...... 24 3. Parameters Affecting Selective Extraction of Lactose in Alcohol/Water Mixtures ...... 25 4. Factors Affecting Yield ...... 27 5. Conversion of Lactose in Selective Solvents 28
C. Influence of Solvent on Proteins ...... 30 1. Non-aqueous Solvents for Proteins ...... 32 a) Stronqly Protic Solvents ...... 34 b) Weakly Protic Solvents ...... 35 v1 2. Protein Denaturation in Non-aqueous Solvents 38 D. ManufactureManufacture and and Food Application of Milk P ro te in s ...... 46 1. Acid Casein.•.. 46 2 . Spray Dried Casleinates 47 3. Co-precipitates 49 4. Food A pplicatio n of Caseinates ...... 51
E. Formulation and Characteristics of Imitation Processed Chees 54
MATERIAL AND METHODS 59
A. Samples 59
B. Lactose Extraction and Extraction Systems. 59 1 . Sinqle and Double Extraction System ...... 59 2 . Soxhlet Extraction System ...... 60
C. Chemical Analysis 60 1 pH ...... 60 2 M oisture ...... 63 3 Ash ...... 63 4 F a t...... 63 5 L acto se ...... 64 6 P ro te in ...... 66 7 M ineral ...... 66 8 Sialic Acids...... 66
D. Physical and Functional Properties 68 1 S o lu b ility ...... 68 2 Colloidal Stability ...... 70 3 Water Absorption and V iscosity.... 70 4 Content of a-ai^ino Nttroqen ...... 71 5 Particle Size Distribution ...... 72 6 Bulk Density...... 72 7 Meltinq Characteristics ...... 73
E. Urea Starch Gel Electrophoresis 73 F. Microscopical E xamination. 74 1 . Powder Samples 74 2 . Imitation Proce ssed Cheese. 75
G. T extural Examin ation by Instron. 76
vH H. Formulation of Imitation Processed Cheese.'; 77 I. Statistical Analysis ...... 80 RESULTS...... 81 A. Extraction of Lactose from NFDM ...... 81 1. Solubility Characterisation of Pure Lactose 81 a) Lactose Solubility in Various Solvents ...... 81 b) Combination of Selected Solvents...... 84 c) Effects of Aqitation Time on Solubility 84 2. Extraction Conditions of Lactose from NFDM 88 a) Evaluation of Azeotropes ...... 88 b) Extraction Temperature ...... 92 c) Agitation Time ...... 92 d) Effect of Repeated Solvent Treatment ...... 95 3. Extraction of Lactose from NFDM In a Pilot System ...... 99 4. Protein Loss from NFDM durinq Solvent T reatm ent ...... 100
B. Physicochemical/Functional Analysis of DENFDM ...... 103 1. Analysis of Compositions ...... 103 2. Free Amino Groups ...... 110 3. Size Distribution and Bulk Density ...... 110 4. Colloidal Stability ...... 114 5. Water Absorption and Viscosity Profile ...... 114 6 . Examination of Particle Structure by Light Microscopy and Scanninq Electron Microscopy 118
C. Solubility Profile of DENFDM ...... 125 1. E ffec t o f pH on S o lu b ility ...... 125 2. Effect of Urea on Solubility ...... 128 3. Effect of Salt and Suqar on Solubility ...... 130 4. Effect of Aqitation on Solubility ...... 136 5. Effect of Sequestrants on Solubility ...... 136 6 . Effect of k-carraqeenan on Protein Solubility o f DENFDM...... 138 7. Effect of Sialic Acid on Solubility ...... 143 a) Extraction of Sialic Acid ...... 143 b) Effect of Methanol Concentration on Loss of Sialic Acid ...... 146 c) Effect of Feedback of Extracted Solids on Solubility...... 146
v111 D. Characteristics of IPC Formulated from DENFDM...... 148 1. Formulation and Preparation for the IPC ...... 148 2. M eitinq A rea ...... 150 3. Meitinq Temperature ...... 153 4. Textural Profile Analysis ...... 156 5. Microstructure of IPC Analoqs ...... 163
DISCUSSION...... 174
A. Extraction of Lactose from NFDM by Various Solvent Systems ...... 176
B. Solubility Behavior of DENFDM ...... 180
C. Phycal/Chemical and Functional Properties o f DENFDM...... 185
D. MorphologicaL/Textural Properties of IPC ...... 189
SUMMARY...... 191
REFERENCES...... 193
APPENDIX...... 202
A. Particle Size Distribution ...... 202
ix LIST OP FIGURES
Fiqure Title Paqe
1. Solvation of alkoxy alcohols ...... 10
2. A comparison of sucrose solubility in alco h o ls and th e i r tu -alkoxy d e r iv a tiv e s ...... " 13
3. Lactose conversion in methanolic and alkaline methanolic solution ...... 31
4. Schematic representation of the conflquration of protein in various solvents ...... 36
5. Mechanism of protein denaturation with water and alcohols ...... 41
6 . Schematic representation of the alcohol denaturation of proteins ...... 44
7. Experimental procedure for extraction of lactose from NFDM...... 61
8 . Soxhlet extraction system ...... 62
9. Standard curve of pure lactose ...... 65
10. Standard curve of sialic acid...... 69
11. Typical TPA curve and data obtained from the c u rv e ...... 78
12. Manufacture process of imitation processed cheese analoqs preparation ...... 79
13. Effect of aqitation time on pure lactose solubility in selected solvents ...... 87
14. Effect of aqitation time on solubility of l a c t o s e ...... 94
X Goodness of the fit for predictinq lactose s o l u b i l i t y ...... 96 Effect of repeated treatment on lactose extraction from NFDM ...... 98 Effect of aqitation time on extraction of lactose by usinq 6 ?.% methanol in a pilot system ...... 101
Effect of temperature on protein loss from NFDM...... 104
Effect of concentration of methanol in protein loss ...... 105 d-amino nitroqen determined in 1% dispersion by formol titration ...... 111 Water absorption and viscosity profile by viscoamyloqraph ...... 117
Effect of pH on solubility ...... 127 Effect of salt and suqar on solubility of DENFDM...... 131
Goodness of fit for predictinq solubility ...... 133
3-D plot of solubility of DENFDM ...... 134
Effect of aqitation time upon the solubility o f DENFDM...... 137
Effect of sequestrants on protein solubility o f DENFDM...... 139
Effect of concentration of sodium metaphosphate on protein solubility of DENFDM ...... 141
Effect of k-carraqeenan addition on protein solubility of DENFDM ...... 142
Effect of concentration of methanol on loss of sialic acid ...... 147
Effect of feedback of extracted solids on protein solubility of DENFDM ...... 149
x1 LIST OF TABLES
Table Title Paqe
1. Dielectric constants of various solvents at specific temperatures ...... 5 2. Solubility of several suqars in aqueous alcohol at 2 0 %...... 19
3. Solubility of hydrated lactose in water ...... 22
4. Approximate percentaqe composition of commercial casein and caseinate products ...... 52
5. Food use o f c a s e in a te s ...... 53
6 . Solubility of pure lactose in different solvent systems ...... 82
7. Effect of combination of selected solvents on pure lactose solubility ...... 85
8 . Analysis of variance: Combination of selective solvents on pure lactose solubility at 25 °C 86
9. Effect of solvents on extraction of lactose from NFDM in a sin q le treatm en t ( a t 2 5 °C)...... go
10. Effect of reaction time on reflux extraction of lactose usinq azeotropic solvents ...... 91
11. Effect of temperatures on lactose extraction in 62% methanol ...... 93
12. Analysis of variance by GLM: Aqitation time on lactose extraction from NFDM ...... 93
13. Mass balance of lactose content ...... 97
14. Extraction conditions for lactose removal from NFDM in a pilot system ...... 102
xi 1 15* Extraction efficiency of lactose at each step in a pilot system ...... 102
16. Protein loss in solvent extraction of lactose from NFDM in a pilot system ...... 106
17. Chemical analysis of products ...... 109
18. Particle size distribution and bulk density of experimental samples ...... 112
19. Colloidal stabilities and hydrability ...... 115
20. Extraction of sialic acid ...... 144
21. Mass balance of sialic acid durinq methanol e x tr a c tio n ...... 145
22. Comparison of meitinq area of various IPC a n a lo q s ...... 151
23. Analysis of variance: Meitinq area of various IPC analoqs depended upon a protein source...... 151
24. Tukey's studentized ranqe test: Meitinq area of various IPC analoqs depended upon a protein source ...... 152
25. Comparison of meitinq temperature of various IPC a n a lo q s ...... 154
26. Analysis of variance: Initial meitinq temperature of various IPC analoqs depended upon a protein source ...... 154
27. Tukey's studentized ranqe test: Initial meitinq temperature of various IPC analoqs depended upon a protein source ...... 155
28. F value of instron examination of IPC analoqs... 157
29. Instron evaluation of IPC analoqs made with different protein sources ...... 158
30. Main trends of protein source on IPC texture.... 162
31. Particle size distribution...... 202
xi11 LIST OF PLATES
Plate Title Paqe
I. Identification of protein loss durinq solvent treatment by urea starch qel electrophoresis... 107
II. Microstructure of Ca-caseinate ...... 119
I I I . M icro stru ctu re of Non Fat Dry M ilk ...... 120
IV. M icro stru ctu re of DENFDM...... 121
V. Microstructure of first extract ...... 123’
VI. M icrostructure o f second e x t r a c t ...... 124
VII. Microstructure of recovered lactosecrystal.... 126
VIII. Effec . of urea concentrations on solubility o f DEN.'DM...... 129
IX. M icrostructure o f a commercial IPC ...... 164
X. Microstructure of Ca-caseinate IPC analoq 166
XI. Microstructure of DENFDM IPC analoq ...... 167 XII. Microstructure of combination of Ca-caseinate/ DENFDM IPC analoq ( 1 : 1 ) ...... 169
XIII. Microstructure of combination of Ca-caseinate/ DENFDM IPC analoq ( 2 : 1 ) ...... 170
XIV. Microstructure of combination of Ca-caseinate/ Na-caseinate IPC analoq (1:1) ...... 172 XV. Microstructure of combination of Ca-caseinate, Na-caseinate and DENFDM IPCanaloq (1:1:1) ...... 173
x1v INTRODUCTION
Casein and caseinate are qood sources of food protein
of excellent qual ity and are used as maior inqredients for
imitation dairy products, includinq imitation processed
cheese* There ha,s been a substantial qrowth in production
and marketinq c f Imitation processed cheese to over
m illio n pound ann ually In U.S.A.
In these products non-fat-dry-milk (NFDM) is not
interchanqeable vi ith caseinates because NFDM contains a hlqh
content of lactos e Which is undesirable in substitute cheese
. manufacture. The properties of casein and caseinates are
characterized by hiqhly desirable physicochemical/functional
properties. Since NFDM has only limited application as a
replacement for caseinates, development of a method for
production of del actosed NFDM was attempted in this study.
The primary purpose of this study has been to devise a
solvent extractic n method for removal of lactose from NFDM
in a manner which would be economical and practical and to
investiqate the physical and functional properties of the
delactosed NFDM.
Finally, an attempt has been made to evaluate the properties of delactosed NFDM for practical application to the substitute processed cheese industry. LITERATURE REVIEW
A. Solubility of Carbohydrates in Non-aqueous Solvent
Systems
Systematic and experimental studies utilizlnq aqueous and non-aqueous solvents have been conducted to isolate carbohydrate and eliminate fat from various food stuffs. In the suqar industry, solvents have been employed to raise extraction efficiency of carbohydrate which is the most important outcome for development of the carbohydrate ester surfactants (Domovs and Freund, 1960).
In the selection of an appropriate solvent system, consideration must be qiven to polarity, inertness, toxicity and cost. In addition solution rate, selectivity and completion of reactions are important for desiqn of the extraction process.
1. Dielectric Constant and Polarity The dielectric constant is one of the important parameters for the solvent properties of a liquid and is related to the dipole moment. Water is an excellent solvent 4 for polar substances and has an unusually hiqh dielectric constant. Accordinq to Zinqaro (1968):
Two principal factors which determine the d ie le c tr ic co n sta n t are the d ip o le moment and hydroqen bondinq. As mentioned previously, water, which possesses a larq e d ip o le moment and Which is very stronqly hydroqen bonded, has an unusually hiqh dielectric constant. In [Table ID are listed some common solvents and the values of their dielectric constants at specified temperatures. The value of the dielectric constant, itself, can be used as basis for the classification of solvents. When a chemist refers to water as "a solvent of hiqh dielectric constant” or to phosqene as a "medium o f low d i e l e c t r ic c o n s ta n t,” these statements convey very definite meaninq. An extremely important physical process Which helps us to understand the solubility of ionic crystals and the stability of solutions of electrolytes in solvents of hiqh dielectric constant is the solvation of ions. Every ion, because it is a charqed particle, creates about itself an electric field. Because of this the dipoles of the solvent molecule will orient themselves about the ion. This is known as ion-dipole attraction, (p.5)
In Table 1 are listed some solvents and the values of their dielectric constant at specific temperature.
For the purpose of classification, Dack (1971) distinquished between solvents by their dielectric constant
(D) as follows:
For the common solvents, D ranqes from 7.o (cyclohexane) to 78 (water )j alcoholic and other solvents with acidic or basic properties qenerally possess moderate or hiqh values on this scale. On 5
Table 1 Dielectric constants of various solvents at specific temperatures (From Lange, 1956)
Solvent D ielectric constant (D) Temperature water 78,54 25 methanol 32.63 25 ethyl alcohol 24.30 25 ammonia 22.20 -33.4 acetone 20.70 25 n-propyl alcohol 20.10 25
1so-propyl alcohol 18.30 25
•1so-butyl alcohol 17.70 25 n-butyl alcohol 17.10 25 benzyl alcohol 13.10 20
1so-th1ocyanate 10.40 20 chloroform 4.81 20 ethyl ether 4.34 20 phosgene 4.34 22 cyclo-hexane 2.02 20 6
the o th e r hand, so lv en ts Which are almost,.devoid of these properties appear to fall into two distinct qroups. Aliphatic and aromatic hydrocarbons are amonq solvents in a qroup of very low dielectric constant (mainly between 7 and in ), while acetone, acetonitrile, and nitrobenzene are included in a second qroup Which have moderate values of D (between 7.1 and 36). (p.l)
Water is an excellent solvent for disaccharides and many polysaccharides, so solubility behavior of carbohydrates in water has been the subject for several reviews (Hudson, 1908: Seidell, 1919). Surprisinqly, alcohols do not compare with water as solvents for carbohydrates. Accordinq to Moye (1972):
On a simple basis of polarity, it would be expected that simple suqars would be more soluble in such polar solvents as alcohols than they actually are. It would appear, from work conducted by Moye and Smythe, that crystal-lattice enerqy resultinq from multiple hydroqen-bond associations plays a vital role in limiting solubility. In terms of such solvents as alcohols, solvation is, therefore, probably not the major factor, but rather the enerqy innut needed to disrupt the crystal lattice, (pp.86-87)
In qeneral, the solubility of carbohydrates in solvent systems are influenced by several factors such as solvent structure, solvent temperature, p H, concentration of solvent, and reaction time. 7
2. Solvent Structure
Fey et al. (1951) studied sucrose solubility In aqueous qlycerol and propylene qlycol, the structures of which are similar to sucrose, to prepare pharmaceutical
syrup. Moye and Smythe, in 1965, found that structures containinq the qroup, R-O-C-C-OH, were qood so lv en ts for
sugars. According to their studies:
The g ly co l e th e rs (I) and the fu rfu ry l alcohols (II) contain the structural grouping R-O-C-C-OH. The sy n th esis o f nine fu rth e r compounds containing this grouoing provided strong evidence to support the theory that it conferred good s o lu b ilis in g powers for sugars and many o f their derivatives. Whilst I,4-anhydroerythritol (III) also contains the ethylene gycol grouoing, the 1,2- and 1,3-0-methyleneglycerols (IV) and (V), respectively, and the furan (VI) and tetrahydrofuran (VII) derivatives do not. All proved to be excellent solvents for glucose, fru c to se , and su cro se. The hexoses, in particular, are miscible with the solvents at the boiling point of the solvent, and crystallise readily from the cooled solutions. Due to their instability, the O-methyleneglycerols could not be examined above 1 0 0 . 2 -formyloxy- and 2 -acetoxy-ethanol, on the other hand, are poorer solvents for sucrose, and th is sugqests th a t Where "R" in (I) is an electron-withdrawing qroup, the availability of electrons on the ether oxygen of the solvent is reduced, and the ability of this oxygen to form hydrogen bonds is diminished accordingly, (p.285) 8
a - o - C M , — cm,—o h
D
I R = alkyl# for ethylene qlycol monoalkyl ethers (alkyl Cellosolves). R « R'O-CI^-CHg- Ill IV V SOCMj CHj OCWj -CM,- R • alkyl VI VI ir VII VII 9 On the other hand, althouqh pyridine and N-N-dimethyl formamide{Moye, 1972), for example, were known as qood solvents for suqars, they were not qenerally available. According to Moye and Smythe (1965): The chemistry and technoloqy of carbohydrates is limited by their qeneral insolubility in orqanic solvents. In recent years, the very meaqre list of orqanic solvents for suqars, of which pyridine is the best known, has been supplemented by a number of less-common orqanic solvents. These include N,N-dimethyl formamide, dimethyl sulphoxide, sulpholane {tetrahydrothiophen 1,1-dioxide), morpholine and substituted morpholines, substituted pyrrolidones, and Y-butyrolactone. Certain of these solvents are now commercially available, but are nevertheless relatively expensive, (p.284) Piqure 1 depicts the solvation structure between a carbohydrate and a solvent possessing the R-O-C-C-OH qrouo. In 1965, Moye and Smythe found in their studies on non-aqueous suqar solvents that w-methoxy derivatives of alcohols were better solvents for sucrose than the corresponding monohydric alcohols, and they predicted the mechanism in the following statement: The effect of an extension of chain lenqth on solvent properties was also examined, and the u-methoxy derivatives of propanol and butanol were found to be better solvents than the corresponding monohydric alcohols. Solvent power is possibly due primarily to the formation of hydrogen bonds between solvent xo X \ Figure 1 Solvation of alkoxy alcohols u and suqar hydroxyl qroups, at the expense of hydroqen bonds between suqar molecules. The hlqher boilinq points of these solvents would assist this process by enabllnq the necessary enerqy to be provided to disrupt the latter intermolecular hydroqen bonds. Solvent-suqar hydroqen bonds could be formed in a variety of ways, but the relationship of the ether and the hydroxyl qroups in the solvent must be important. Several possibilities are shown in Fiq.l, where only hydroqen bondinq with one suqar molecule ha3 been considered. The solvent could equally well be shown as b rld q in q two suqar m olecules, bu t we feel that this is less likely, as each end of the solvent molecule would be functioninq more or less independently (especally for u) -methoxybutanol), and the BOlvent would not be expected to be a better solvent than its simpler, monofunctional r e l a t i v e s . 3. Temperature Effects Kononenks and Herstein (1956) examined the relationship between sucrose solubility and temperature for water as a solvent and developed the followinq linear eq u a tio n : e ■ a(b - T) where e is the weiqht of solvent needed to dissolve 1 q of sucrose (experimentally determined), and a and b are parameters calculated from their results by the method of least squares. For those solvents whose dielectric constants were known, a plot of loq a versus the dielectric co n sta n t qave a smooth curve. The siq n ific a n c e o f this result was not appreciated, but it is interestinq that some form of relationship appeared to exist. This result led the authors to suqqest that the solubility of sucrose in a non-aqueous solvent could be calculated for any temperature, provided that the solubility at one 12 temperature and the dielectric constant of' the solvent were known. ( p.89) On the basis of the linear equation described by Kononenks and Herstein, the solubility of other disaccharide in a solvent could be predicted by conductlnq the same experiment. Presumably one would draw an equation of solubility in any case of solvent by axamininq the solubility and the dielectric constant in the laboratory system and thereby postulate the solubility of carbohydrate in the specific solvent by an experimental curve between the solubility and the dielectric constant. However, carbohydrates are not directly soluble in solvents such as alcohols. After aqitation, those mixtures that appear homoqeneous are liable to have resulted from solute-solvent covalent interactions. The boilinq point of solvents is extremely important in the comparative solubilities of carbohydrates as shown in Fiqure 2 with some explanations {Moye and Smythe, 1965): The properties of solutions of suqars in the (J-ether solvents were favourable for their subsequent use in physical and chemical processes. The solubility of qlucose and fructose in the simple qlycol ethers, and in tetrahydrofurfuryl alcohol, was very hiqh at temperatures above the meltinq points of the suqars, and , as already mentioned, complete miscibility resulted near the boilinq point of the solvents. The solutions were not unduly viscous, and the suqars crystallized readily on coolinq. 13 Moye (1972) has further explained: >"tfMCTMODtVjL s " '!' B U t t M O L T«mp, Fiqure 2 A comparison of sucrose solubility in alcohols and their uJ-alkoxy derivatives. (Moye and Smythe, 1965)( p .286) Thus with ethanol* the solubility of sucrose at 80 is only 0.5% (w/w)(solvent), but this is raised to 8.6% at 140°* a significant deqree of solvation occurrinq after lattice disruption. Of greater interest* perhaps, is the unexpected discovery that a wide ranqe of solvents containinq the s tr u c tu r a l qroupinq R-O-C-C-OH are b e t t e r solvents for suqars than would be anticipated from structural considerations. 2-methoxyethanol, for example* was found to be a far better solvent for sucrose than diethyl ether or ethanol. Solvents containinq this qroupinq are shown in CFlq.lil, with the proposed solvation structure depicted in [Fig.2]. It was proposed that the solvents have favorable boilinq points (ranqinq from 120 to 200 ) that facilitate disruption of the lattice, and that the particular solvent structure enqaqes in solvation. The hiqh boiling-polnts are undoubtedly Important in relation to input of enerqy to disrupt the lattice, and this feature miqht be inferred to be of qreater importance than structure* as such solvents as 3-methoxyoropanol and 4-methoxybutanol solvate sucrose appreciably at elevated temperatures* even if sliqhtly less efficiently than 2-methoxyethanol. (p.87) 4. Inorganic Solvents Some inorganic solvents have been investigated develop excellent solvents in the suqar industry. (1972), in his publication, stated; Liquid ammonia has considerable potential as a solvent for carbohydrates, as it readily d isso lv es simple suqars and many p o ly sacch arid es. Franklin and Kraus conducted an extensive study of the solubility of a wide ranqe of compounds in liquid ammonia. Amonqst these were a number of carbohydrates. D-glucose* D-fructose, sucrose, lactose, maltose, and raffinose were found to be very soluble, and L-arabinose and D-galactose were described as being moderately soluble, (p.91) 15 Moye (1972) stated the effects of mineral salts on the solubility of carbohydrates! Mineral salts have also received attention* Domovs and Freund found the solubility of a wide ranqe of carbohydrates in methanol to be markedly enhanced by the presence of calcium chloride. Mono-, di-, and tri-saccharides in qeneral were found soluble. In the presence of 1.5 molecules of calcium chloride and 12 molecules of methanol per hexose unit, maltose, L-sorbose, melezitose, D-galactose, and sucrose were soluble, whereas D-fructose, raffinose, D-glucose, and lactose were soluble but crystallized out later, (p.95) The details for the solubility in methanol in the presence of calcium chloride has been summarized by Domovs and Freund (1960): Lactose, sucrose, maltose, qalactose, fru c to s e , and many o th e r carbohydrates and derivatives are found hiqhly soluble in cold absolute methanol containinq sufficient calcium chloride. From a hiqhly concentrated viscous solution prepared from lactose, calcium chloride, and methanol in a molar ratio such as 1:3:24, a complex of 0-lactose, calcium chloride, and methanol in the molar ratio of 1:1:4 slowly crystallizes, (p.1216) Recently removal of lactose from aqueous solution usinq qroup IIA (alkaline-earth) metal chlorides and sodium hydroxide has been investigated by Quickert and Bernhard (1982). They abstracted their studies as follows: 16 Recovery of lactose from 5% (w/v) aqueous solutions usinq strontium, maqnesium, and calcium chlorides and sodium hydroxide as precipltatinq aqents was in v e s tiq a te d . The e f f e c t o f sodium hydroxide to metal ion and metal ion to lactose molar ratios, temperature, time and order of addition of reaqents were examined. Maximum recoveries weret 26* for strontium, 55* for maqnesium and 97* for calcium. Recovery data were analyzed by multiple, nonlinear reqression and response surface methodoloqy to obtain plots of the relationship between the variables. Lactose recovery with meta hydroxides appears to involve a complexation process, and is not strictly an a d so rp tiv e p ro cess, ( p .1705) These findinqs, theoretically, were explained by Blaedel and Meloche (1963) as an inert electrolyte effect: An inert electrolyte increases the solubility of an ionizable precipitate (in this case, a metal hydroxide) due to the increase in the ionic strenqth of the solution which causes a decrease in the activity coefficient of the ions. (p.173) An alternative reaction mechanism for the recovery of lactose from solution usinq metal bases has been advanced by Quickert and Bernhard (1982). Accordinq to their study, a complex of lactose and metal hydroxides would be involved in lactose precipitation. 5. Orqanic Solvents Some studies on the solubility of carbohydrates in orqanic solvents have been conducted accordinq to Moye (1972): 17 Behr used ethanol as a recrystallizinq solvent for D-glucose, Wolfrotn and Wood recrystallized alcohols in oreparlnq D-qlucose and D-fructose from sucrose. In preparinq D-mannose from ivory-nut shavinqs, Isbell used methanol and iso-propyl- alcohol to aid crystallization, but water was present in the system. The most qenerally useful recrystallizinq solvents amonq the alcohols were, however, found by Moye and Smythe to be those co n tain in q the R-O-C-C-OH qroupinq. (p.96) Systematic studies on the solubility of carbohydrates in methanol have been carried out since 1897. Trey (1995) studied the solubility of D-qlucose and its effect on the boilinq point of methanol. Leviton (1949) used methanol to extract lactose from sklmmilk powder. His studies were associated with separation of lactose and whey oroteins by methanol extraction. Between 1973 and 1976, matty fundamental studies on lactose solubility in various solvent systems were extensively carried out (Lim et al., 1973: Nickerson et al., 1974: Maid et al., 1976). In a similar fashion, a few studies on solubility of carbohydrates in an ethanol system were carried out at various temperatures accordinq to Moye (197?): Trey found the solubility of D-qlucose in ethanol to be 0.25% w/v at 20° and 1.42% w/v at the boilinq point of the solution. Levine and coworkers quoted Hudson and Yanovsky as havinq qiven the solubility of D-qlucose in 100% ethanol at 20° as 1.6-1.7% w/v, but the present author could find no mention o f t h i s . The l a t t e r authors qave the solubility of o-L-rhamnose hydrate as 18 8.6% w/v initially, at 20 , risinq after mutarotation to 8.5% w/v. Authors are in better aqreement as to the solubility of sucrose in ethanol, with the exception of Scheibler, whose values would be expected to be less accurate, havlnq been obtained by extrapolation from aqueous ethanolic mixtures. Both Schrefeld and Pellet and Pellet qave 0.00% w/v at 14 and 15-16°, respectively, whereas Reber found the solubility to be 0.051% w/v at 75 . Because of the low solubility reported, Moye Qand Smythe conducted studies over the ranqe of 80 to 140° and found solubilities ranqinq from 0.5 to 5.6% w/v. (p p .97-98) The solubility of several suqars in aqueous alcohols at 20 °C are depicted in Table 2. Some studies were also made on the solubility of carbohydrates in a wide ranqe of related alcohols such as butyl alcohol (Moye et al., 1965), propylene alcohol (Fey et al., 1951: Merten et al., 1964) and iso-butyl alcohol (Trey, 1895) at various temperatures. One would predict that suitable solvents should be inert and should not carry health restrictions. Alcohols and ethers may fit the descriptions and contain both hydroqen acceptor sites and donor functions considered necessary for a qood solvent. Table 2 Solubility of several sugars In aqueous alcohols at 20°C (compiled from Hudson and Yanovsky, 1917) Gms. Anhydrous Sugar Sugar Solvent per 100 cc. Solution ' Initial Final Solubility Solubility 00 a Arablnose o C2 H5OH 0.74 1.94 3 Cellose 20% it 3.2 4.7 3 Fructose 80% II 13.4 27.4 3 ” 95% II 1.8 4.2 3 " Methyl alcohol 5.2 11.1 a Galactose 60% C2 H5OH 1.1 3.1 o " 80% II 0.27 0.65 3t a Glucoheptose 20% II 4 4.5 a Glucose 80% II 2 4.5 a " Methyl alcohol 0.85 1.6 a Glucose hydrate 80% C2 H5OH 1.3 3 3 Glucose 80% II 4.9 9.1 a Lactose hydrate 40% II 1.1 2.4 a Lyxose 90% II 5.4 7.9 3 Maltose hydrate 60% II 3 4.75 3 Mannose 80% II 2.4 13 3 " Methyl alcohol 0.78 4.4 3 Mellibose d1hydrate 80% C2 H5OH 0.76 1.3 a Rhamnose hydrate 100% II 8.6 9.5 a 11 11 70% II 8.2 9.6 a Xylose 80% II 2.7 6.2 Sucrose 80% II 3.7 3.7 Trehalose d1hydrate 70% II 1.8 1.8 Rafflnose pentahydrate 50% II 1.4 1.4 20 B. Extraction and Crystallization of Lactose in Aqueous and NOn-aqueous Solvent System Water is an excellent solvent for suqars including lactose. However, Seidell (1919) has referred in his publicatin on solubilities of inorqanic and orqanic compounds to a slow attainment of saturation for lactose in w ater: It was found that the saturation point was reached very slowly with this compound. Prom the results, it was concluded that aqueous solutions of milk-suqar contain two substances in equilibrium and that the mutarotation of milk-suqar results from the slow establishment, in cold solutions, of the equilibrium of the balanced reaction, Cj 2 H2»tQi2 (hydrate) ■ H 2O + C12H22O11 (^-anhydride), (p.696) Accordinq to Hudson (190B), the equilibrium reaction is more complex involvinq a two-step reaction: It seemB quite likely however from the similarity of milk-suqar to qlucose that the above reaction equation is only a portion of what occurs in milk-suqar solutions and that the full reaction should be written 12 H22O 11+ H, 0 = {**-Anhydride) Anhydride)+ jri$ 5 2\ 0 The velocity of the first of these balanced reactions must be instantaneous compared with the rate of the second one. This view reconciles the two theories of the mutarotation of the suqars which have been much discussed in the last few 21 ears, the stereoisomer and the hydrate theories, Tp . 1768) The solubility of lactose in water from unsaturated and supersaturated conditions at various temperatures is shown in Table 3. 1. Alcohol Extraction of Lactose from Skimmilk and Whey The possibility of utilizinq selective solvent extraction of lactose from skimmilk has been reported by Leviton (1949): Work previously reported on the separation of so lu b le p ro te in s and la c to se from whey powder has been extended to skim powder and skimmilk. Precipitation of the proteins of skimmilk by methanol at subzero temperatures permits the recovery of practically all the soluble proteins of the oriqinal milk. Spectroturbidimetric examination of the redisoersed protein indicates no siqnificant chanqe in particle size distribution. Spray-processed skimmilk powder treated with 62% methanol at -15 °C . Yields lactose and a "soluble" protein product. The crude lactose is of exceptionally hiqh quality. The protein product, comprisinq 42.2% of the solids of skimmilk, contains 74.1% protein. Of this protein, 81% is casein. Small losses in solubility result from both extraction and dryinq. Extraction at room temperature leads to equally qood results. The solubility of the protein product in cold water, lost in part as a result of the increase in processinq temperature, is recovered by heat treatment of the reconstituted suspension, ( p . 1351) 22 Table 3 Solubility of hydrated lactose In water Solubility (%) Temperature ( C) pro|T) prom Average Lactose In unsaturation saturation 100 g water (mM) 0 10.6 10.7 . 10.6 34.8 15 14.4 14.6 14.5 49.7 25 17.8 17.8 17.8 63.4 39 24.0 24.1 24.0 92.7 49 29.7 29.9 29.8 124.0 64 39.7 39.8 39.7 193.0 74 46.2 46.5 46.3 253.0 89 58.1 58.3 58.2 407.0 (From Hudson, 1908) Morr and Lin {1970) have used this approach for preparinq a whey protein concentrate from liquid or dried Whey: Aqueous methanol, ethanol, n-propanol, and n-butanol were investigated for precipitatinq Whey proteins from liquid and dried Whey to prepare a whey p ro te in c o n cen trate. Ethanol was more satisfactory than the other alcohols; methanol and n-propanol Whey protein concentrates were sliqhtly less soluble in water and neutral •buffer than ethanol Whey protein concentrate, and n-butanol Whey protein concentrate contained qross amounts of lactose and was practically insoluble in water. The 72% ethanol treatment precipitated 45 to 60% of the Whey proteins and produced a Whey protein concentrate with a fourfold protein enrichment on a solids basis compared to whey. Minimum coprecipitation of minerals and lactose with the protein concentrate was produced by treatlnq calcium-free acid whey with 72% (w/w) ethanol at 0° to 5 °C. Ethanol extraction of freeze-dried Whey was less satisfactory than liquid whey treatments because of substantial coextraction of proteins with the lactose and minerals. Fifty to 70% of the freeze-dried whey protein concentrate was resolubilized at pH 6.6, whereas it was totally soluble at pH 8 to 9. The protein concentrate, which exhibited normal zonal electrophoretic properties at pH 8.6, contained 7 to 3 times as much bovine serum albumin and proportionately less tJ.-lactalbumin than control Whey proteins. Pretreatments to remove calcium and other minerals from whey and use of t 72% (w/w) ethanol durinq preparation improved the solubility of the protein concentrates. Possible mechanisms for ethanol-induced Whey protein denaturation and precipitation are presented. ( p .1162) The separation efficiency for lactose and protein in the study by Morr and Lin (1970) was, however, not 24 comparable to that reported by Leviton. Accordinq to Morr and Lin (1970): Neither of the experimental procedures in our study fractionated the lactose and Whey proteins as thorouqhly as reported by Leviton and Leiqhton for spray-dried cheese whey. They quantitatively extracted lactose from WPC with little loss of protein into the extractant. The reasons for these qross differences are not known, but may be due to variations in the physical structure and porosity, crystallinity of the lactose, and mineral content of these whey systems as influenced by the method of manufacture. It is also possible that the proteins in the spray-dried Whey were heat denatured, thus renderlnq them non-extractable by ethanol. (p.H6R) 2. Choice o f Solvent Short chain alcohols have found application for extraction of lactose from skimmilk and whey. Ethanol appears to be less effective than methanol both with reqard to solubility and stability of supersaturated solutions of lactose. For example, Leviton (1949) stated: Whey powder, Which is difficult to disperse in dilute ethanol, is readily dispersible in dilute methanol. Furthermore, the stability of the supersaturated solutions of lactose in methanol is much qreater than correspondinq solutions in ethanolj consequently, additional time becomes available with the use of methanol for the separation of coaqulated proteins, ( p .1949) 25 Majd and Nickerson (1976) studied lactose solubility in methanol and ethanol at different concentrations and found th a t : Lactose solubility at hiqh concentrations of alcohol qenerally decreased as the chain lenqth of the alcohol increased. The few instances where that was not observed, e.g., 70% (v/v) alcohol, can be attributed to differences in rate of crystallization in the alcoholic solutions, ( p .1027) They stated further with reqard to lactose solubility at different concentrations: At low concentrations of alcohol the solubility of lactose was reduced sliqhtly but siqnificantly in both 2% and 5% methanol and ethanol. At these low concentrations of alcohol, chain lenqth had no effect on solubility. (p.1038) 3. Parameters Affecting Selective Extraction of Lactose in Alcohol/Water Mixtures. To characterize the extraction system, it must be recoqnized that the main effect on lactose solubility in alcohols is associated with concentration of solvent and the ratio of powder-to-solvent. Accordinq to Leviton and Leighton (1938): 26 - - the first effect on lactose of the addition of alcohol to whey powder, apart from the effect of chanqe of solvent, is a dilutinq action Whereby a solution supersaturated with respect to lactose is formed. The stability of the supersaturated solution is such that complete extraction of the lactose and separation from insoluble material by filtration is easily realized. The lactose subsequently crystallizes from the filtrate and is recovered, (p.1305) Complete removal of lactose was obtained at the powder-to-solvent ratio of 5:95 at -15 °C (Leviton, 1949). For whey system s, usinq th is r a tio , a l l la c to se was extracted by treatment with 70.7% alcohol within 3 minutes after mixinq of the inqredients. Leviton (1949) has explained the solubility of supersaturated solutions with time as follows: The resultinq supersaturated solution is stable for 7 and 15 minutes, respectively, dependinq upon Whether or not continuous aqitation is employed. Because the inqredients of whey which are insoluble in the solvent behave as a qranular precipitate upon filtration, separation, within the interval durinq which the filtrate supersaturated with respect to the lactose is stable, is quite feasible, (p.1306) There is stronq evidence that the alcohol concentration in the solvent has a pronounced effect on lactose extraction. Leviton (1949) has explained: 27 At hiqh methanol concentrations relatively small quantities of lactose are extracted. This appears to be due to incomplete absorption of solvent by the powder qrains. Between 65 and 75% methanol, the extraction process chanqes abruptly, thorouqh penetration of the qrain particles takes place, and extraction of lactose is practically complete, (p.1352) Time and tem perature fo r la c to se e x tra c tio n has a pronounced effect on solubility, stability and crystallinity of lactose. The relationship between temperature and extraction/stability has been considered by Leviton (1949): In qeneral, it is true that the stability of the supersaturated lactose solution increases with increasinq temperature, while the time required to dissolve the amorphous lactose decreases; consequently at low temperatures the extraction becomes less efficient, (p.1307) 4. Factors Affectinq Yield The quantity of solvent-to powder must be expected to exert a considerable effect of the yield of lactose in any one extraction. Leviton (1949) concluded in his studies on methanol extraction of lactose: The percentaqe of lactose extracted diminished slowly as the powder-solvent ratio was increased, but even in the most concentrated suspension (30 qrams of powder per 100 ml of solvent) e x tra c tio n was 92.6% com plete, ( p . 1352) 28 This observation may be related to the physical state of lactose. In the alcohol treatment, amorphous lactose seems likely to dissolve to yield a supersaturated solution and its degree of saturation is likely to depend upon the powder-solvent ratio. The proportion of lactose extracted was not complete at the powder-solvent ratio of 3:10. Part of the lactose may have a crystalline form, Which will not dissolve in the solvent. In addition, some of the lactose may crystallize before filtration is complete. Leviton predicted the incomplete extraction was associated with the followinq phenomena: (1) At low powder-solvent ratios all of the substances responsible for the optical rotation of milk are extracted with the exception of a small quantity which appears to be a part of the precipitated complex; (2) Even in the most hiqhly supersaturated filtrate (containinq approximately 10 qrams of excess lactose per 100 ml of solution at -15 C) the onset of crystallization is delayed at least 15 minutes beyond the time required for f i l t r a t i o n , ( p .1353) 5. Conversion of Lactose in Selective Solvents Treatment with methanol for several hours at room temperature has been reported to convert a-lactose monohydrate to a stable, anhydrous form (Ross, 197R), which is qenerally regarded identical to the stable anhydrous a-lactose (Lim et al., 1973) produced by the heating 29 procedure of Sharp (1943). Addition of more water to 95% ethanol or methanol increased the amount of crystalline a-lactose hydrated form. Since alcohol decreases solubility of lactose. it would be expected to accelerate crystallization by increasinq suoersaturation. Choi et al. (1951) noted the effect of water in alcohol on conversion of lactose. Similar observations were also made by Olano (1978): Water above 4% (w/v) in the ethanol, propanol, or butanol preparations qreatly impaired the conversion of a-lactooe hydrate to 0-lactose. It was possible to obtain 0-lactose in methanol, provided the water content was below 10% (w/v). ( p .1623) Herrinqton (1934) reported that in alcohol precipitation a-lactose was precipitated more readily than 0-lactose. Choi et al. (1951) studied the behavior of lactose crystallization in alcohol and stated as follows: The results indicate that very little of the lactose is crystallized as the a-hydrate from treatment with 95 per cent ethanol. Similar results were obtained with methanol, which, because of its low water content (0.5 per cent), would not be expected to promote the formation of crystalline a-lactose hydrate. Addition of more water to 95 par cent ethanol, however, increased the amount of crystalline a-l^ctose hydrate formed, (pp.852-853) In a similar fashion, Olano and R iO B (1978) depicted 30 the conversion of lactose in methanolic and alkaline methanolic solution, as presented in Fiqure 3, and qave the followinq explanation* The difference between the conversion of lactose in methanol and alkaline methanol can be summarized as follows* in methanol the mutarotation of lactose is slower than the crystallization of o( -lactose* therefore, the ^-lactose concentration will not reach saturation, and only anhydrous oi-lactose will crystallize (I). Since the equilibrium between the two forms is e s ta b lis h e d ra p id ly in an a lk a lin e medium in the presence of sodium hydroxide, the 4 -lactose hydrate dissolves and mutarotation takes place faster than the crystallization of the ct-lactose. The ^-lactose concentration in solution will reach saturation and will crystallize (II). (p.303) C. Influence of Solvent on Proteins Proteins exhibit a qreat deal of differences in their solubility characteristics and the classical nomenclature for proteins was, in fact, established on the basis of differences in solubility behaviors. Bezkorovainy (1970) has summarized* Proteins are classified, especially in the older literature, into simple, coniuqated, and derived proteins. The simple proteins constitute the larqest class of proteins, and contain no compounds other than amino acids. They are further subdivided into albumins, qlobulins, qlutelins, prolamines, scleroproteins, histones, 31 a-Lactose*H20 a-Lactose 0-Lactose Crystallization a-Lactose4 fast (I) B a-Lactose-I^O ►a-Lactose fast 0-Lactose 4- CrystalHzatlon 8-Lactose 4— (II) ( p.302) Figure 3 Lactose conversion 1n methanol1c and alkaline methanollc solution (From Olano and R1os, 1978) (A) methanollc solution (B) alkaline methanollc solution 32 and protamines. Albumins are proteins soluble in water, serum albumin and eqq albumin beinq typical examples. Globulins are proteins vrtiose solubility in water is increased by small amounts of salts. They can be further divided into euqlobulins or proteins completely insoluble in water, and pseudoqiobulins that are sparlnqly soluble in water. (3 -lactoqlobulin of bovine milk is a typical euqlobulin. Glutei ins are plant proteins that are insoluble in water, but are soluble in dilute acids and alkalis. Prolamines are also plant proteins, which contain larqe quantities of proline ad ammonia. They are insoluble in water and absolute alcohol, but are soluble in a 70 to 80 per cent solution of alcohol. Scleroproteins are fibrous proteins insoluble in water. Amonq the better-known scleroproteins are collaqen, elastin, and the hair keratins, soluble in water, and usually associated with nucleic acids. ( p p .13-14) 1. Non-aqueous Solvents for Proteins The use of solvents as media in treatment of proteins may be accomplished with less damaqe to the proteins concerned if proper recoqnition is qiven to possible conformational chanqes in the protein. In qeneral, proteins treated by solvents miqht have entirely different conformations (Tanford, 1969) from their native ones, and as a result, be functionally altered or inactive (Dubrow at al., 1973). The effect of solvent on protein is complex and a cause for diverse interactions. The protein solubility and stability (Herskovits et al., 1970)(Horne et al., 1981, 1982) after treatment of various solvents are depended upon 33 solvent varieties, contact time, pH and ionic strength of so lv e n t. Proteins are qenerally not directly soluble in the common nonpolar solvents, or in the usual polar solvents such as alcohols. However, they are directly soluble in stronqly protic solvents such as phenol, formic and acetic a c id . In qeneral, addition of organic compounds (McKenzie, 1970) has a dissociating effect on proteins. The dissociation of protein is associated with interchain linkages such as the noncovalent bondB (e.g. hydrogen bonds, hydrophobic bonds and salt linkaqes) and the covalent bonds (e.g. -s-s- bonds). Under favorable circumstances, organic compounds disrupt the noncovalent interchain bonds, dissociating the protein into its monomer units. These monomer units may be sinqle chains, or the units may consist of several chains with interpaptide covalent linkage. In order to examine the potential damaqing effects of solvents on protein properties it may be useful to consider the characteristics of good non-aqueous solvents for proteins. According to Singer (1965): It is clear that certain general requirements must be met by a non-aqueous solvent, if it is to be useful. Since proteins contain chemically reactive qroups, a first requirement of a satisfactory solvent is chemical inertness. That is, the solvent must not cause the rupture or 34 formation of any covalent bonds in the macromolecule other than those with hydroqen atoms. This places a severe restriction on the number of useful non-aqueous solvents, since oxidizinq or reducinq, alkylatinq or acylatinq. etc., solvents are eliminated. This requirement also places some not-so-obvious limitations on solvents Which miqht otherwise appear to be satisfactory. For example, small amounts of residual water may produce scission of peptide linkaqes in proteins dissolved in acidic or basic solvents. Such residual water may be very difficult to remove from both the solvent and the macromolecule in question, (p.3) a) Stronqly Protic Solvents These solvents are characterized by beinq both proton donors and acceptors, and are often qood solvents for proteins. The effect of these solvents on proteins has been explained by Sinqer (1965): For stronqly protic solvents, three factors may be involved in the disruption of the native structures of protein molecules and their conversion to hiqhly unfolded conformations: (a) The stronq hydroqen bonds formed between the solvent and the protein solute, particularly between the solvent as donor and the amide C«0 qroups of the protein as acceptor. This factor has already been discussed in connection with synthetic polypeptides CSection IV,D]. (b) A marked decrease, compared to water, in lyophobic interactions towards the nonpolar residues of the protein. Of the common stronqly protic liquids, only hydrofluoric acid, formic acid, and hydrazine are not completely miscible with simple hydrocarbons, While the others, such as ethylenediamine, dichloroacetic and trifluoroacetic acids, are completely miscible with them. Even in the cases of hydrofluoric acid, formic acid, and hydrazine, the solubility of hydrocarbons is much greater than in water. 35 Thus, a saturated solution of benzene in formic acid at 25 °C contains 0.088 mole fraction of benzene, compared to 0.00035 mole fraction of benzene in water [see Section IV,B,3 and Table IV]. (c) An increase, compared to water, in intramolecular electrostatic interactions within protein molecules in the case of stronqly protic solvents of hiqh dielectric constant (>*M0), such as hydrofluoric acid, formic acid, and hydrazine. Protein molecules are hiqhly ionized as a result of the abstraction of protons from the acidic solvent or the donation of protons to the basic solvent. On the other hand, in stronqly Drotic solvents of low dielectric constant (>^15), such as dichloroacetic and trifluoroaceti'c acids and ethylenediamine, these ionic charqes on protein molecules must certainly be extensively paired to solvent counterions [see Section IV,B,1], and electrostatic repulsive interactions must be of neqliqible importance in destabilizinq the native conformation, (p.40) b) Weakly Protic Solvents Such solvents are characterized by beinq only weakly active as proton donors or acceptors, and are correspondinqly less effective in protein solubilization. Yanq and Doty (1957) were the first to obtain clear evidence for conformational changes in protein molecules in weakly protic solvents. The solubility of a protein miqht well depend on conformation. Therefore it is Important to realize that the native conformations of proteins are always altered in solvents. For example, schematic diaqrams illustratinq both the native and the intermediate co n fiq u ratio n are shown in Fiqure 4. 36 •O Q Figure 4 Schematic representations of the configuration of rlbonuclease 1n various solvents: (A) 8M aqueous urea; (B) water at room temperature; (C) almost pure 2-chloro- ethanol; (D) water-chloroethanol with 80-90 mole % (50- 70X by weight) water. The dark circles represent the four disulfide bonds of rlbonuclease, located at approximately the correct positions along the polypeptide backbone. (From Weber and Tanford, 1959, p.3257) 37 B seems to be native conformation In water and D represents an unfolded conformation retaininq the net helical content of the native form but with hydrophobic regions disrupted. C exhibits an extensively helical conformation. Weber and Tanford (1959) predicted these protein conformational ohanqes miqht Involve a chanqe in the helical content. The first chanqe (B D) miqht involve a unfoldinq of the protein molecule with little chanqe in helical content: the second (D * C) miqht then involve a refoldinq of the molecule into riqht-handed Ol-helicai regions. Sinqer (1965) was in aqreement with their representation, but stressed the role of electrostatic interactions: Increased helical content in proteins dissolved in such solvents is probably connected with (a) decreased hydrogen-bonding capacity of the solvent compared to water, and (b) decreased electrostatic repulsive interactions between the fixed charqes on the protein molecule in the low dielectric constant solvent as compared to water, due primarily to counterion binding, (p.40) Singer (1965) has further stated: In the last decade, however, it has become generally recoqnized that the solvent water olays an exceedingly important role in determining and stabilizing the characteristic structure that a protein molecule exhibits in an aqueous environment. It is in connection with this problem of the role of solvent in regulating 38 macromolecular structure that non-aqueous solutions of proteins have been most frequently studied in the last few years. In an aqueous protein solution, it is well known that any of a wide variety of treatments, such as an increase of temperature of exposure to pH extremes, leads to denaturation of the protein. Denaturation is qenerally accompanied by a disorqanization and partial randomization of the secondary and tertiary structure of a protein molecule. Therefore, we have been conditioned to think of the native aqueous CONFIGURATION of a particular protein molecule as the most hiqhly ordered form which its covalent structure permits it to attain. On of the most interesting developments of studies of proteins in non-aqueous solvents, therefore, has been the findinq that in certain systems an apparently more hiqhly ordered conformation (greater helical content) of the protein molecule exists than it native aqueous form. On the other hand, there are other non-aqueous protein solutions in which the macromolecular conformation becomes hiqhly disorganized, (p.43) Z. Protein Denaturation in NOn-aqueous Solvents P ro tein d e n a tu ra tio n is commonly defined as any noncovalent chanqe in the protein structure. This chanqe may also alter the secondary, tertiary or quaternary structure of the molecule. As mentioned previously, denaturation is qenerally accompanied by a disorqanization and partial randomization of the secondary and tertiary structure of a protein molecule. Kauzmann (1959) defined protein denaturation as: 39 There is little doubt that When one is confronted with specific examples of What protein chemists would aqree to call denaturation, one almost invariably finds on close inspection that there have been changes in the way the polypeptide chains are arranqed within the protein molecule. One can be reasonably sure, therefore, that one is talking about the same phenomenon as is the protein chemist if denaturation is used to denote a process (or sequence of processes) in Which the spatial arranqement of the polypeptide chains within the molecule is chanqed from that typical of the native protein to a more disordered arranqement. (The terms configuration, conformation, and state of foldinq can be substituted for spatial arranqement in this definition), (p. 2 ) Therefore, the characteristic feature of denaturation might be specified as the rupture of intramolecular bonds leadinq to the disordering of the native protein structure. Fukushima (1969) investigated systematically the denaturinq action of various orqanic solvents on soy proteins and discussed the mechanisms of the denaturation with orqanic solvents, water and the solvent-water mixtures. According to his studies: A system atic study was made o f the denaturinq ability of about 30 kinds of orqanic solvents toward soybean proteins. In qeneral, the denaturinq ability of orqanic solvents depended on their hydrophobicities and their deqree of dilution by water. Hiqhly hydrophobic solvents possessed little denaturinq power toward proteins, even at hiqh temperature. The denaturinq power of solvents increased with addition of water, whereas that of water also increased with addition of solvents. Consequently, the water-solvent mixture had high denaturinq abilities that could not be 40 attained by the individual components alone. Lower alc o h o ls were much stro n q er d en atu ran ts than other solvents examined. The denaturinq ability of alcohols at low concentration increased with the hydrophobicities of alcohols * the reverse was found at hiqh concentrations. The denaturation mechanisms of soybean proteins with orqanic solvents and with water are discussed from the standpoint of the three-dimensional structure of soybean protein molecules, (p.156) Fiqure 5 illustrates schematically the mechanism of protein denaturation with water and alcohols. Fukushima explained quite satisfactorily the .various phenomena as follow s t The lack o f d enaturinq a b i l i t y by water-immiscible hydrophobic solvents toward soybean proteins, even in hiqh temperature, can be explained by the idea that the solvents are prevented from enterinq into the hydrophobic reqion by the surroundinq hydrophilic shell and the outside hydration layer, which cannot be disrupted by hydrophobic solvent alone. On the other hand, water can destroy these hydrophilic shells easily, owinq to its stronq hydroqen bond-forminq ability. The hydrophobic reqions of the molecules, however, cannot be disrupted directly thouqh the indirect rupture of the hydrophobic reqion miqht occur, accompanylnq the unfoldinq of polypeptide chainB caused by disruption of the hydrophilic shell. Thus, the hydrophilic shell will be the principal reqion altered as a result of denaturation with water alone. On the other hand, disruption of the hydrophobic reqion will be easily brouqht about by the addition, to water, of water-soluble solvents which possess both hydrophobic and hydrophilic radicals. The hydrophobic portions of the added solvents can penetrate into the hydrophobic reqion after disruption of the surroundinq hydrophilic shell. In this case, the denaturinq abilities of water should increase with the amounts of added 41 B Figure 5 Mechanism of protein denaturation with water and alcohols. Top (A), one globular protein molecule surrounded by water (HOH) In which a hydrophobic region (a) and a hydrophilic region (b) are present. B and C, disruption of hydrogen bond by water and of the hydrophobic bond by alcohol. Thick lines 1n B and C, polypeptide backbone of proteins; • 1n C, one molecule of water; -R 1n C, amino acid side chains. {From Fukushima, 1969, p .163) 42 s o lv e n ts , end, fu rth e r, w ith the same concentrations of added solvents their effect' should increase in the order of hydrophobicities of the added solvents* However, addinq too much of these solvents to water will weaken the action of water to breakinq the hydrophilic shell of the protein molecules, (pp.162-163) The addition of a solvent to the system will lower the dielectric constant. This will tend to increase the strenqth of all electrostatic interactions between molecules that were in contact with water only. In this case many of the protein hydroqen bonds are effectively removed from the solvent. The presence of the less Dolar solvent will also have the effect of weakeninq the hydrophobic bonds of the proteins. The denaturinq ability of alcohols on protein molecules at low concentration was suqqested by Haqerdal (1978): Low concentrations of orqanic solvents (< 30%) cause an unfoldinq of the orotein structure in dilute solutions at pH different from pi. When the concentration of orqanic solvent is increased, the amount of secondary structure will increase compared to the native protein molecule [Weber and Tanford, 1959J Tandord et al., I960*, Herskowitz et al., 1970; Herskowitz and Solli, 1975]. Nozaki and Tanford [1971] havesuqqested that this is due to the stronq preference of the backbone peptide unit for a hydrophilic environment, which causes the formation of intramolecular hydroqen bonds and thus, miqht explain the increase in secondary structure. If, however, protein solutions are exposed to increasinq concentrations of an orqanic solvent at or near pi, there is an increase in protein aqqreqation [Tanford, 1969]. This was also found here when myoqlobin was exposed to increasinq 43 concentrations of acetone. It also appeared from the DSC and water solubility analyses that only a fraction of the myoqlobin preparations obtained at both low and hiqh acetone concentrations undervent irreversible transition. This restricted irreversible transition at low and hiqh acetone concentrations should be of different oriqin because of the differences in these acetone-water media, (p.25) In a study of the structual stability and solvent denaturation of proteins. Herskovits et al. (1970) postulated: The effects of the water-miscible straiqht chain and branched alcohols and qlycols on the native conformation of sperm Whale myoqlobin, cytochrome c, and •L-chymotrypsinoqen have been investigated by spectral, difference spectral, and optical rotatory dispersion methods. Based on the midpoints of the denaturation transitions, that is, the amount of alcohol or qlycol required to produce 50% denaturation at 25°, it is concluded that the effectiveness of the alcohols as protein denaturants increases with increasinq chain lenqth or hydrocarbon content, in conformance of What is expected of the disorganization of the hydrophobic interior of these proteins revealed by their detailed three-dimensional x-ray structure. As a rule, branchinq of the hydrocarbon portion of the alcohols tends to reduce their effectiveness as protein denaturants. The qlycols are found to be less effective than the corresponding alcohols, suqqestlnq that increased polarity or hydroqen-bondinq capacity is of secondary importance When compared with the effects of increasing hydrocarbon content. (p.25BR) Fiqure 6 presents mechanism of alcohol denaturation of qlobular proteins. As predicted from Fiqure 6 , the hydrophobic group becomes largely exposed in the cause of 44 native Jr B 'unfolded jr helical alcollc Figure 6 Schematic representation of the alcohol denaturation of proteins. In the folded native form of the protein the burled tyrosyl and tryptophyl residues are represented by the shaded hexagons. These groups are exposed 1n both the unfolded and thp ordered helical form of the protein at high alcohol content, and are represented by the uniltadad haxagom, From Herskovlts et al., 1970, p.2591) 45 the unfoldinq process. If a hydrophobic qroup is exposed, it will have a low state of enerqy. This seems likely to be a continuinq process until random fluctuations in the protein structure occurs. Denaturation of protein invariably results in a marked decrease in solubility in aqueous solvents, and the loss of solubility depends on the reactive affinity of the protein molecule toward the solvent. In qeneral, as proteins are very sensitive to extraction conditions, protein solubility measurements have been used to determine the extent of denaturation takinq place durinq the process. Denaturation always leads to the rupture of intramolecular bonds, so that the denaturated protein suffers a loss of solubility in aqueous solvents. In another sense, denaturation of protein makes it susceptible to proteolysis and causes a loss of bioloqical activity. There are some applications for utilizinq the relationship between solvent and protein in the dairy industry. Horne and Parker (1981, 1982) have examined the basis of traditional alcohol stability test as a indicator of heat stability of milk. The state of stability of protein dispersion in milk was mainly depended on the deqree of hydration and zeta-potential (Puri et al., 1965). 46 D. Manufacture and Pood Application of Milk Proteins Caseins are a family of related phosphoproteins comprisinq about 80* of total proteins in milk. They possess useful functional properties for food applications (Chakraborty# 1981) and have been widely used as a maior inqredient in a number of imitation and non-dairy products (Milner et a l., 1978). Since world war II, casein production has been markedly changed in various countries (Muller, 1971) due mainly to pricinq policies which have tended to favor alternative uses of skimmilk. Today, the food application of caseins has been rapidly expanded to 70 - 80* of total production (Southward and Walker, 1980). 1. Acid Casein (iso-electric casein) This type of casein is a standard product which is manufactured from skimmilk by acid precipitation followed by washing, pressinq or centrifuqation, dryinq and qrindinq (Omeara and Munso, lSS?). The red u ctio n o f pH is commonly achieved by the action of lactic acid bacteria or by the addition of a mineral acid to the skimmilk. Adiustment of temperature is conducted to aid the expulsion of whey and to produce a curd of optimum firmness for separation from the whey. Acid casein is often differentiated on the basis of 47 the particular acid used for precipitation, e.g. lactic acid, hydrochloric acid and sulfuric acid. The dry, qround material is relatively dense and remains Insoluble in water unless further modified. 2. SDray Dried Caseinate These products are drypowders produced by neutralizing standard acid casein with alkali and spray drying the s o lu b iliz e d m ateria l (Fox, 1970s M uller, 1971). Soluble caseinates of importance contain sodium, ammonium, potassium or magnesium as the cation. Sodium caseinate is by far the most commonly used in foods. It is produced by spray drying acid casein curd which is dissolved in sodium hydroxide (Muller, 1971: Fox, 1970: Southward and Goldman, 1975). Sodium caseinate is soluble but not easily dispersed because of a tendency to form qelatineous lumps (Hokes, 1983). Caseinates lack the bulk density of acid casein, but are soluble in water to a varyinq degree, depending upon the type of alkali used for neutralization/solubilization. In comparison with sodium caseinate, calcium caseinate is less soluble. Hokes (1982) has commented on the properties of calcium caseinate as follows: 48 The chemical reactions involved in the manufacture of calcium caseinate are slower and more complex than in the sodium form. Insoluble isoelectric casein must be converted to a colloidal calcium salt. In general, calcium caseinate is produced by isoelectrically precipitatinq the casein from skimmilk, washinq thorouqhly to remove lactose, minerals and Whey proteins, dewaterinq, conversion with Ca(OH)_ or CaCl* and then spray dryinq. Variability In the final product can be obtained in a number of ways: 1) the type of acid used and the pH endpoint for precipitation, 2 ) thoroughness of the washinq, 3) amount of moisture in the precipitate after draininq, 4) method of preparinq the precipitate for conversion i.e. colloid milL or ammonia solubilization, 5) temperature of conversion, 6 ) spray dryinq parameters, (p.3) Chakraborty (1981) has discussed the functionality and food use of commercial caseinates as follows: Commercial caseinates, prepared as Na- and K-caseinates exhibit improved solubility and fu n c tio n a lity compared to C a -c a se in a te . Lower solubility of Ca-caseinates is due to larqer sized and stronqly interacting aqqregates promoted by the cross linkinq of divalent cations. In any case, it is the soluble caseinate, particularly Na-caseinate, which constitute the maior form in Which c a se in finds i t s food use. Maior food use of caseinates is summarised in CTable 5 (Milner et al., 1978)]. It is evident from the data presented that the food use of casein preparations are based on their excellent functional properties, (pp.4-5) Calcium caseinate is poorly soluble and forms milky suspensions of relatively low viscosity related to the stronq interactions of casein with calcium. Where the specifications for a calcium caseinate call for higher pH 49 values with a limit for calcium, an alkali, such as ammonium hydroxide, is used to increase pH to the value required (Muller, 1971). Ammonium caseinate is not readily available in commercial form but is technically feasible. Ammonium c a s e in a t e , p rep aredfrom a c id c a s e in by ammonium hy d ro x id e, is somewhat acidic but still hiqhly soluble in water. The ammonia is loosely bound and the product may revert to the f u l l y a c id ic form, i f th e ammonia e s c a p e s . Ammonia caseinate may be prepared by exposinq dry qround acid casein to anhydrous ammonia vapor without the need for intermediate solubilization and subsequent SDray dryinq (Glrdhar and Hansen, 1974). 3. C o -p re c ip ita te In the manufacture of casein, only some 90% of the protein of skimmilk can be recovered because the whey proteins are soluble at pH 4.6 (Muller, 1971). However, procedures for co-precipitation of casein and whey proteins have been developed to improve protein recovery and retain or improve the desired functional properties for food application. Co-precipitated casein is made in a similar manner as acid casein, but from extensively heated skimmilk (Howard et al., 1954), which promotes co-precipitation of a substantial amount of whey protein with casein to the extent of 10-15% of the final product. Casein and Whey protein co-precipitates are produced by a variety of processes that involve heatinq skimmilk to >90°C to denature Whey protein and complex them with the K-casein component of casein micelles by disulphide interchange, thereby renderinq the entire milk protein system precipitable uoon acidification to pH 4.6 or by addition of calcium ion. The level of calcium in the co-precipitate has been found to affect the functional properties (Muller et al., 1967). Muller (1971) has commented on the motivations of manufacture of co-precipitate as follows: In the 1950's i t was observed th a t the proteins precipitated by acid or calcium chloride from heated milk comprised both casein and Whey proteins, i.e., a co-precipitate. The development of these observations into commercial manufacture of co-precipitate was motivated by the followinq three considerations: i) the desire to improve recovery of protein from m ilk-recovery is when casein is manufactured. ii) the concept that the ranqe of functional properties in milk proteins for foods could be in c re a se d . iii) the hope that the sliqhtly improved nutritive value of the casein-whey protein combination could increase the role of milk proteins nutritionally. ( p . 668 ) In a paper presented by Chakraborty (19B1), co-precipitates were defined as a modification of caseins 51 due to nutritional and functional propertiest The formation of co-precipitates of casein and Whey proteins may also be considered as a modification of the caseins. Co-precipitates have hiqher nutritive value In comparison to the caseins. The solubility of the co-precipitates is someWhat le s s than the c a sein s and is dependent on the calcium level (hiqh, medium, and low calcium co-precipitates). Co-precipitates find use in breakfast foods and other products of hiqh n u t r i t i v e value Where s o lu b ility is not an important functional attribute, (p. 8) The compositional differences between different casein and caseinates are shown in Table 4. 4. Food Application of Caseinates Caseins and caseinates contribute excellent nutritional value to food products but their principal value is related to their functional contributions. They have found extensive application due to their moisture absorption capacity, fat emulcification ability, whipplnq/viscosity/thickeninq properties, solubility at neutral and alkaline pH ranqes, qel forminq/texturization ability, bland flavor, and neutral color backqround (Emmons, 4 1981: Morr, 1979: Milner et a l., 1978: Jonas, 1973). Maior food use of caseinate is summarized in Table 5. Recent developments in cheese analoq technooqy utilize the unique textural properties of casein that simulate those 52 Table 4 Approximate percentage composition of commercial casein and caseinate products Component Sodium Calcium Acid Rennet Co caseinate caseinate caseinate caseinate preslpltate Protein, N x 6.38 94 93.5 95 89 89-94 Ash (Xmax) 4.0 4.5 2.2 7.5 4.5 Sodium (X) 1.3 0.05 0.1 0.02 - Calcium (X) 0.1 1.5 0.08 3.0 - Phosphorous (X) 0.8 0.8 0.9 1.5 t* Lactose (Xmax) 0.2 0.2 0.2 - 1.5 Fat (Xmax) 1.5 1.5 1.5 1.5 1.5 Moisture (Xmax) 4.0 4.0 10 12 5.0 pH 6.6 6.8 - 7.0 6.8 (From Morr, 1982) Table 5 Food use of caseinates Type of Food Approximate Properties usage level(%) ut111zed(a) Meat emulsions 3 - 5 2.3 (welners, sausage, etc) Breads, ro lls, buns, other 2 - 4 1.2.3 yeast doughs Cakes and cookies 2 - 5 1.2.3 Pancakes, waffles, other 2 - 5 1.2.3 chemically leavened butters cereals 2-6 1 Whipped toppings, synthetic creams 5 - I0 3.5.6.7 Ice cream, frozen desserts 4 - 6 Pharmaceutical preparations 2 - 5 1.4.8 Baby foods wide range 1.4 (From Milner : al., 1978) (a) Properties used 1. Nutritive value 2. Moisture absorption 3. Emulsifying property 4. Heat stability 5. Whipping ability 6. Coating ability 7. Bodying ab ility 8. Keeping quality (b) on dry basis (c) on total sol Ids basis 54 of natural type cheeses. Casein-whey protein co-precipitates are useful in the formation of breakfast cereal, snack and pasta products since their low solubility allows them to contribute characteristic textural properties to these products. Regular casein or co-precipitated caseins are also beinq used in comminuted meat products where they contribute to the water- and fat-bindinq properties of the meat proteins. Other food product applications include dairy spreads, cultured dairy products where they contribute viscosity, texture and other desirable properties (Southward and Goldman, 1975), E. Formulation and Characteristics of Imitation Processed Cheese ' Most imitation processed cheeses have been produced for institutional and manufacturing purposes and made to resemble to the standardized or non-standardized mozzarella type or American type of natural or processed cheeses (Kautter et al., 1981). The production of reqular cheese is covered in most countries under some form of "Standard of Identity" and as such can only be made from milk with no foreiqn substance added, with the exception of certain specified functional inqredients, such as salt and rennet. Thus, the substitution of any constituent of standardized 55 varieties in the U.S.A. mandates that the product be la b e lle d as an IMITATION. An acceptable imitation cheese, functionally identical to its natural counterpart, has been produced with suitable oils and suspension of caseinate-whey for fortification by similar procedures used for natural cheese with less expensive inqredients (Petka, 1976). A number of patents have been issued for the development of imitation processed cheese recent, for example: simulated cheese products (Freck and Kondrot, 1974), protein food products resemblinq cheese (Boyer, 1975(a) 1975(b): Wynn, 1978(a)), imitation mozzarella cheese (Hansen et al., 1981), simulated cheese preparation with calcium caseinate (Wynn, 1978(b)) and imitation cheese (Wynn, 1978(c)). Hokes (1982) has considered the formulation of a processed cheddar-like imitation cheese product in more d e t a i l s : The order of addition of inqredients, mixinq temperature and shear rate become important in determininq the final product quality. Caseinate can be blended and heated with the oil-soluble constituents to which the water soluble material is added under reasonable shear rates capable of forminq the emulsion (Petka, 1976). Another method is to first form an emulsion between the aqueous and oil phase in a mixer at 180 F, deaerate to minimize air holes in the product before addinq the caseinate (Wynn, 1978(b)). The caseinate cannot be added directly to the water 56 due to the limited water available in the system. In such a product it is the protein which is responsible for any unique functional properties. Processinq and formulation must be evaluated accordinq to the protein Which represents the qreatest influence on variability, (p.7) The quality of the imitation processed cheese is depended upon the formulation (Thomas et al., 1980) and manipulation of processinq conditions (Lazarldis and Rosenau, 1980). For aid in the study of the meltinq characteristics of proteins durinq meltinq# a lipid free model system was used by Hokes (1987.). Based on these studies# Hansen et a l. (1981) have suqqested: The thermomeltinq properties of imitation cheese are related to the protein matrix formed from stronq surface interactions between calcium caseinate particles# qenerated by the influence of chanqinq the dielectric constant of the solvent mixture, (p.7) The presence of caseinates in imitation cheese mainly affects the physical properties of the product, includinq shred, melt, slice, texture and stretch characteristics unique to the imitation processed cheese (Petka, 1976). Accordinq to Hokes (1982): The calcium salt of casein differs in physical properties from the sodium salt. Calcium caseinate is poorly soluble in water where it forms a colloidal dispersion. Sodium caseinate, on the other hand# is water soluble forming 57 viscous solutions at low concentrations. A critical ratio of the monovalent and divalent metal ions is important to the functionality of natural cheese (Keller et al., 1974: Sirry and Shipe, 1959) and processed cheese (Vakaleries et al., 1967). If only calcium caseinate is used in an IPC formulation, a poorly emulsified and low melt product results due to its low solubility. Sodium caseinate is a better emulsifier but tends to produce irreqular meltinq patterns, undesirable in cheese products, (p. 10) More recently, incorporation of whey protein retentates (Covacevich and Kosikowski, 1978: Jolly and Kosikowski, 1975) from p la in and enzyme tre a te d Whey (Kumar and Kosikowski, 1977) into manufacture of imitation processed cheese has been attempted. In processed Cheddar cheese, 80% plain retentates substitution revealed an undesirable lonq qrain texture and bland flavor (Sood and Kosikowski, 1979) while additions of double, diafiltrated retentates produced cheese of poor meltinq qualities. The salt balance in the formulation is very complex in its effect on meltinq characteristics. The addition of calcium chloride may shift the balance in such a way that the pH is lowered throuqh reaction with citrate and phosphate. Phosphates and citrates influence the meltinq characteristics includinq meltability, free fat separation and appearance of melted cheese (Lazaridis and Rosenau, 1980). Monovalent sodium salts of phosphates and citrates dissolve the paracasein qel by removinq the bivalent calcium 58 in the casein complex (Sood and Kosikowski, 1979). Thomas (1970) reported that processed cheese made with polyphosphates had a low melting index. The meltinq index can be controlled larqely by inqredients (Arnott et al., 1957) and emulsifiers (Rayan et al., 1980: Lazarldis and Rosenau, 1980). Soluble protein (Thomas, 1970), chelatinq agent and acidulants (Shehata et a l., 1967) have also been found to Influence the meltinq characteristics. Imitation processed cheeses are particularly suitable for use in such dishes as pizza, tacos, sandwiches, sauces and other prepared foods. Therefore, It is important that the products posses thermomeltinq qualities which are appropriate for their use in these products. MATERIALS AND METHODS A. Samples Pasteurized qrade "A" low-temperature spray non fat dry milk {NFDM) powder from valley Lea Dairies Inc. was used throuqhout for lactose extraction. The spray dried calcium caseinate (# HP Dill, identified as havinq "qood" functional properties in imitation cheese) examined was manufactured in New Zealand while sodium caseinate was purchased from Matheson Coleman & Bell Inc. for formulation of imitation processed cheeBe analoqs. B. Lactose Extraction and Extraction System I. Slnqie and Double Extraction System Five parts of NFDM powder were extracted by addinq 95 parts of solvent solution in a qlass beaker and aqitatinq viqorously by mechanical stirrer at specified temperature and time in a sinqle system. For the double extraction system in a pilot plant, 5 parts of commercial NFDM was incorporated in 95 parts of 62% methanol solution 59 60 (g rad etT ech n ical). The m ixture was f i r s t a q ita te d by wire whip type a g ita to r (H orbart Model A-200-2 mixer) a t room temperature for 2 hours. The first extracts were separated by gravity sedimentation technique at room temperature. For the removal of excess lactose, the extracts were remixed with 62% methanol in the ratio of 20t 8D (w/w) respectively, for 4 hours at room temperature. Portion of each extract was removed and used for analysis for lactose and other components. Experimental outline of double extraction system is shown In Flqure 7. 2. Soxhiet Extraction System. This system was designed to determ ine the optimum time for efficient extraction of lactose in a continuous manner. Five parts of NFDM oowder were placed Into the c e llu lo s e e x tra c tio n thim ble (Whatman 33x80 mm) and 95 D arts of various solvent solution were put in the round flask in the soxhiet extraction apparatus depicted in Fiqure 8 . The samples were refluxed at 79°C and portions of the distillate were removed after specified time for analysis. C. Chemical Analysis 1. pH pH was measured a t room tem perature on samples with 61 NFDM + Solvent-Water Mixture 1 1st Agitation X Separation Supernatant "p|)t X Addition of Solvent-Water Mixture X 2nd Agitation 1 Separation X I Supernatant ppt Washing wl^h Solvent X Evaporator Remove Solvent i Frozen i ------X Solvent Recovery Lactose Recovery Freeze Drying i Recycling Remove Solvent Delactosed NFDM Freeze Drying L a c lo s e Figure 7 Experimental procedure for extraction of lactose from NFDM 62 Figure B Soxhiet extraction system 63 a standard pH meter (pHM 62, Radiometer-Cooenhaqen). The samples were reconstituted to •>% total solids for pH measurement (Sosulski et al., 1976). 2. Moisture NFDM and delactosed NFDM powders were weiqhed into a tared weiqhed porcelain evaporatinq dishes equipped with a l i d . The samples were placed in a vacuum oven a t 95°C and dried overniqht and then removed to a desiccator and cooled for I hour before weiqhinq. Moisture was calculated as the loss in weiqht. 3. Ash The samples ware weiqhed into porcelain evaporatinq dishes from the dryinq oven and placed Into a muffle furnace at 559°C for overniqht. The samples were treated with concentrated HCl if they were not White. After coolinq the dishes were weiqhed and the ash content calculated as percentaqe of the dry weiqht. 4. Fat The d ir e c t method o f e th e r e x tr a c t was u tilis e d (AOAC, 1965). Two qram samples were weiqhed and dried usinq the A.O.A.C. air oven method. The dried sample was transferred to an extraction thimble with porosity 64 permitting a rapid flow of diethylether. Fat was extracted in a soxhiet extractor for about 4 hours at a rate of 2-3 drops per second* After beinq removal of the ether by cautious evaporation of the contents of the soxhiet flask, the sample was dried in air oven at 100°C for I hour, cooled, weiqhed and fat content expressed as the percerit weiqht loss. 5. Lactose The phenol-sulfuric acid method by Marier and Boulet (1959) for analysis of lactose was utilised. The color development is based on the combined action of phenol and sulfuric acid with lactose. Two ml sample solution was placed into a test tube. A 0.1 ml 90% phenol solution was added and 6 ml conc. sulfuric acid was added slowly alonq the side of the tube. The samples were mixed quickly with care and incubated for color development for 10 minutes at room temperature. After coolinq, optical density was obtained a t 490 nm usinq a double beam Spectrophotom eter (Perkin-Elmer/Coleman Model 124). A standard curve for lactose determination was constructed from pure lactose solution (0-80 jjq/ml) • Figure 9 depicted the standard curve for lactose with a reqression coefficient of 0.9993, a slope of 0.0135 abs.unit/pq/ml, and a Y-intercept of 0.0060 abs.unit at 490 nm. iue Sadr cre f ue lactose pure of curve Standard 9 Figure 0.2 0.4 ABS. 0.6 0.8 1.0 1.2 20 ATS CONC(j/ l) C.(|jg/m N O C LACTOSE 40 60 80 b O T 65 66 6 . P ro tein Total n itro g en was determ ined by the K jeldahl and microKjeldahl method (A.O.A.C., 1970) utilizinq a digestion mixture of K 2SOV and HgO In concentrated sulfuric acid. Total protein was estimated by multiplying the percentage of total nitrogen by the conversion factor 6 . 38. 7. Mineral Samples were analyzed for elemental profiles by flame emission spectroscopy in the Research Extension Analytical Laboratory at OARDC, Wooster, Ohio. 8 . Sialic Acids The Warren TBA assay (Warren, 1959) of sialic acids was selected for use in this study. This colorimetric method measures free sialic acid only and consists of preliminary mild acid hydrolysis of the sample, followed by oxidation, color development and color extraction. a) Reagents i. Periodate solution - prepared by dissolving 4.28 qof sodium meta periodate in 40 ml of distilled water, and adjustinq the volume to 100 ml with 85% H 3PO1,. ii. Arsenite solution - prepared by dissolving lOq of sodium arsenite, and 7.1 q of sodium sulfate in 0 .1 N 67 and adjusting the volume to 100 ml. 111. TBA solution - prepared by dissolvinq 0.6 q of 2-thiobarbituric acid and 7.1 q of sodium sulfate in distilled water and adiustlnq the volume to 100 ml. b) Procedures P eriodate so lu tio n (0.1 ml) was added to a 0.2 ml sample in a screw-capped test tube. The contents of the tube was mixed thoroughly and placed in a water bath at 95°G for 70 m inutes. One ml o f a rs e n ite so lu tio n was added to destroy excess periodate at the end of incubation. After the addition of 3.0 ml of freshly prepared TBA solution, the sample was placed in a boilinq water bath for 15 minutes and subsequently cooled in ice water for 10 m inutes. o Cyclohexane was used to extract the chromophore at 0-5 C. The solvent la y e r was separated by c e n trifu q a tio n for 5 minutes at 3000 rpm (International Clinical Centrifuqe, Model CL). The solvent lay e r was then tra n sfe rre d into a tube and allowed to reach room tem perature. The s i a l i c acid was obtained by comparing the absorbance of the chromophors by a standard curve; prepared from crystalline * N-acetyl-neuraminic acid (NANA) (Kehaqias and Hansen, 1076). Preparation of stock solution was described by Warren (1950) and Marier et a l. (1963). A standard curve was co n stru cted from 68 N-acetyl-neuraminlc acid (NANA) ranqing from 0 to 100 ;jg/ml, o hydrolyzed in 0.1 N I^SO/* for 60 minutes at BO C. Figure 10 * showed the standard curve with a goodness fit of regression coefficient (0.999), a slope of 0.059 abs.unit/jig/ml and a Y-intercept of - 0.0027,0 abs.unit at 549 nm. If standard curves are used, precautions should be taken to insure that the standard sialic acid is cure (Aprahamian, 1973). D. • Physical and Functional Properties 1. S o lu b ility As proteins are very sensitive to extraction conditions, protein solubility measurements have been used to determine the extent of denaturation taking place during the process (Cunningham et al., 1975). Solubility of protein wasdetermined by modifying the method of Lu et a l. (1972) and Akobundu et al. (19B2). One qram of delactosed NFDM was suspended in fifty ml of deionized water with addition of sequestrants or solubilizinq aqents with or without adiustment of pH and aqitated with a maqnetic s t i r r e r for 30 minutes a t room tem oerature. The susoension was cen trifu q ed for 30 minutes at 6,000 x q and room temperature (Sorvall RC 2-B, AutomaticRefrigerated Centrifuqe with GSA rotor). The supernatant was decanted and filtered with Whatman No.42 filter paper. Five ml of 69 1.2 0.8 3 0.4 0 10 20 CONC.OF SIALIC ACS) (Hf/mO Figure 10 Standard curve of sia lic acid 70 filtrate was analyzed for nitroqen by the microKieldahl method (A.O.A.C., 1970). 2. Colloidal Stability The percentaqe of insoluble material in the delactosed NFDM dispersion and the hydration of the * in so lu b les were measured accordinq to the method o f Thompson et al. (1968). Dispersion of each sample were prepared at 1.0, 2.5, 5.0% (w/w). Eiqht qrams of each dispersion were weiqhed into 50 ml polycarbonate tubes and centrifuqed for 90 minutes a t 25,000 rpm. The supernatant was discarded and the weiqht of the hydrated pellet recorded. Excess moisture was removed by placinq the centrifuqe tubes upside down to d rain u n til a co n stan t weiqht was reached. The d iffe re n c e s were recorded as the deqree of hydration expressed as the ratio of water retained to the dry weiqht of samples. The fraction of insolubles was measured as dry weiqht after weiqhinq the dry pellet. 3. Water Absorption and Viscosity The water absorption of delactosed NFDM was measured accordinq to the method described by Hokes (1982). Twenty qrams of delactosed NFDM were suspended and stirred in 300 grams double distilled deminerallized water for 45 minutes. Thirty qrams of Mira-Clear Starch (Staley 71 Industrical Products) were incorporated into the delactosed NFDM suspension. The mixture was placed in the measurinq bowl of a Brabender Viscoamyloqraph and the beaker rinsed with 120 qrams of water. The initial temperature was equilibrated to ?5°C and then the temperature was increased o a t 1.5 C oer m inutes for 35 m inutes a t 75 rpm. The fin a l temperature of 78 °C was maintained for 5 minutes before rapid coolinq was s ta r te d . The decrease in v isc o p ity o f the starch due to the presence of the protein appears to be a function of the hydration of the protein. Maximum viscosity of the samples, in Brabender Units (B.U.), was measured usinq a starch slurry without the addition of protein. 4. Content of a-amino Nitroqen The amount of a-amino nitroqen was determined by formal titration accordinq to the method of Kuehler and Stine (1974). One-half qram of each sample was placed in 50 ml of deionized water and slurried at room temperature. Fifteen ml of a 1% suspension were titrated to pH 8.5 with 0.02 N sodium hydroxide. Two ml of neutralized formaldehyde (37%) were added to release the hydroqen ion from the end amino qroups. The hydroqen ions were then titrated to pH 8.5. Results were expressed as mq a-amino nitroqen per qram of sample from the titration data (Hokes, 1982) and converted to mq a-amino nitroqen per qram of protein. 72 mq a-amino N = (ml NaOH to pH 8.5)*(0.02N NaOH)*(14 mq N/meq N) 5. Particle Size Distribution The particle size distribution was conducted by surface illuminatinq microscopy (Nikon Profile Projector, Model 6 C) accordinq to the method of Janzen et al. (1953). For measurements, the powder particles were randomly dispersed and fields were selected at random, with all particles in each field beinq classified. Particle diameter were measured usinq Martin's diameter, the dimension, parallel to the occular scale, that divides a randomly oriented particle into two equal areas (Stockham and Fochtman, 1978). The mean volume diam eters were computed accordinq to the method of Hayashi (1969). 6 . Bulk Density A calibrated 50 ml qraduated cylinder was slowly filled with sufficient powder to fill the cylinder. The weiqht of the product to fill the cylinder without tappinq or with tappinq was used to calculate the loose or closely packed bulk density, respectively. Bulk density was determined by caiculatinq the ratio of weiqht to volume of powder. A similar method has been described by Kinq (1965). 73 7. Meltinq Characteristics Prepared cheese samples were cut and shaped into 1 cm diameter Which were weiqhed and placed in a oetri dish* The samples were melted in a microwave oven (Sanyo, EM 8600) for 20 seconds at full powder (650 watts)* The area of the melted curd was measured with a planimeter* Meltinq area was expressed as area per qram of sample. Meltinq temperature was measured by modifyinq method of Catslmpoolas et al. (1970). The cheese samples were sliced into 1 cm diameter and 3.0 cm heiqht cylindrical shape and Dlaced in test tube. The test tube with a thermometer were qradually heated in water bath until the samples were melted. The meltinq temperature at which the oriqinal shape of sample was d is to r te d was recorded as i n i t i a l m eltinq tem perature. The fin a l m eltinq tem perature was measured when the samples reached the bottom of the tube. E. Urea Starch Gel Electrophoresis Vertical urea starch qel electrophoresis was conducted to justify the components in 11.4% starch qel (Electro S tarch Company, Madison, WI) by th e method o f Cooney and Morr (197?) and Morr (1974). The qel buffer consisted of pH 8 .6 Tris-citrate buffer (0.076 M) containinq 7 M urea and about 0.1% (v/v) 2-mercaptoethanol. A discontinuous pH R .6 sodium borate buffer (0.3 M) was used in the electrode chamber. Electrophoresis was conducted in a o to 5 °C cold room for about 16 hours usinq a constant current of 25 mA. Twenty five milliqrams of each sample were dispersed in 1.0ml of the above pH fl .6 Tris-citrate buffer contalnlnq 7 M urea. 25 ml of the dispersion, were placed into the top of the shot. Two drops of 2-meroaptoethanol and one drop of a 0.07% (w/w) Amido Black, 50% (w/w) sucrose so lu tio n In the same pH 8 .6 stock qel buffers without urea were added to each dspersion iust before electrophoresis. The sliced qel (3 mm th ic k slab) was stain ed for a t le a s t 1 hour in a 0.7% Amido Black in 15% acetic acid solution. The staine'-’. qel completely destained by soakinq in a destaininq solution (qlycerolswatersacetic acid (5:5:1)) after partially destaininq with fresh acetic acid solution. 0.5 ml of a diluted rennet solution (1:100, Dairyland Food Laboratories) was added to 10% sample solutions to measure the clottinq time at 30°C described by Kehaqias and Hansen (1976). F. Microscopical Examination 1. Powder Samples The powder samples were examined both by liqht microscopy and by scanninq electron microscopy. A small amount of powder for liqht microscopy was placed on a qlass 75 slide with immersion oil (refractive index; 1.5150) and viewed throuqh a Leitz Dialux Microscope equiooed with polarizing filters. For scanninq electron microscopy, specimen holders (metal SEM stubs) were covered with double-faced adhesive tape (Buma and Henstra, 1071) and the powder particles sprinkled onto the tape. Excess particles were removed from the tape by shakinq to prevent overlapping of the particles and charqinq artifacts. The powder particles were coated with carbon and qold by soutter coating prior to examination in a Cambrldqe S4-10 Scanning Electron Microscope. 2. Imitation Processed Cheese a) Reagents i. Buffered neutral formalin solution 37-40* formalin 100 ml DM w ater 000 ml NaH2P0 ij, 1 H20 4 qm Na2HP0ij, anhydrous 6.5 qm ii. Methylene blue solution (Protein stain) Methylene blue 0.5 qm Ethanol 30 ml K0H, 0.1 N 2 ml Water 70 ml 76 iii. Oil red 0 (Pat stain) or Sudan IV Oil red O 1 qm E thanol, 70% 50 ml Acetone 50 ml b) Procedures The imitation processed cheese analoqs for liqht microscopy were cut into small cubes and fixed in 10% buffered neutral formaline solution for ?4 hours at 3°C. The samples were then sectioned to 10-20 pm on a freezinq microtone (American Optical Corp.) and stained with Sudan IV or oil red for fat stain and methylene blue for protein stain. The samples were examined under a Leitz Dialux Microscope equipped with a briqht field condensor. For scanninq Electron Microscopy, the cheese samples were fixed in 2% qlutaraldehyde (adiusted to pH 6.7 with 2,4,6-collidine) for one hour at room temperature. Samples ware dried in a freeze drier overniqht. The dry Dieces of cheese were fractured and mounted on specimen stubs usinq liquid qraphite. They were coated with carbon and qold and examined in a Cambridqe S4-10 Scanninq Electron Microscope. G. Textural Examination by Instron All samples were equilibrated at 25°C for 1 hour and 77 cut into cylindrical shapes of 7 cm diameter and I cm heiqht prior to test. Textual properties such as fracturabllity, hardness, adhesiveness, cohesiveness, sprinqness, qumness and chewness were evaluated usinq the Instron Universal Testlnq Instrument (Model 1000) with a load cell ranqe of 0-50 kq. A Dlunqer of 6.0 cm diameter was attached to the movlnq crosshead. The speed o f the crosshead was s e t a t 7.0 cm/min. In both upward and downward d ire c tio n s . The plunqer was i n i t i a l l y ad ju sted to 1.0 cm from a flat holdinq plate and set for maximum deformation of 0.75 cm with automatic control for first and second bite. A diaqram of a typical compression curve is depicted in Fiqure 11 identlfylnq each characteristic. H. Formulation of Imitation Processed Cheese Twenty-three qrams of veqetable oils and 445 qrams water were incorporated to a mixture of emulsifyinq and bufferinq aqents (Fiqure I?) in the bowl of a douqh mixer. A salt-caseinate mixture (1.5 qrams salt and 76 qrams calcium caseinate) was added to the warm water-fat blend w ith ap p ro p riate mixinq. The curd was packaqed in alumimum foil and kept in refrlqerator at 4-6°C. The cheese samples was cooled down before evaluation of characteristics. 78 DEFORMATION ft7gcm..frftzscni Figure 11 Typical TPA curve and data obtained from the curve where: yj ■ fracturabllity y2 ■ hardness A*, A«» Ag » areas under the force- deformatlon curve A2 “ cohesiveness A1 Ag ■ adhesiveness {j£ = elasticity (hardness)(cohesiveness) « gumminess (gumminess)(elasticity ) ■ chewiness 79 Water 46.5* Fat 23.0% Lactic acid 1.0% Heating up to 90°C c»if M4vt.ma Sodium phosphate 0.2% Salt Mixture «- Sod1um J1tr; te oa% 1st Mix 2-3 minutes I Scape down I 2nd Mix 3-4 minutes i Molding Keeping at 4-6°C l Evaluation Figure 12 Manufacture process of the Imitation processed cheese analogs preparation 80 I. Statistical Analysis The P-test procedure was carried out to analyze all data by SAS (1982 system). Multiple comparison of means was also performed usinq Tukey's Studentized Ranqe Test by SAS. The significance difference at 1% and 5ft level was obtained from Ott (1977). RESULTS A. Extraction of Lactose from NFDM The e x tra c tio n o f la c to se from NFDM Is la rg e ly Influenced by the extraction conditions and the nature of the solvents employed In the extraction procedure. The physical/chemlcal and functional properties of delactosed NFDM (DENFDM) may be dependent on the final composition as well as on the extraction methods utilized. This study has been conducted as an exploration of several approaches for extraction of lactose by non-aqueous solvents. The underlylnq purpose has been to develop extraction conditions Which may selectively seoarate lactose from NFDM with minimal damaqe to the functional properties of the resultlnq DENFDM. 1. Solubility Characteristics of Pure Lactose a) Lactose Solubility in Various Solvents The Bolublllty of lactose In various solvent systems* Includlnq azeotroplc mixtures at 25°C and 5°C, are Illustrated In Table 6 . Water Is the best solvent for 81 82 Table 6 Solubility of pure lactose 1n different solvent systems Solvent Cone. (%) Soluble lactose (mg/ml) 25°C 5°C n-propanol 100 90 72(c) 5.744 5.033 1so-propanol 100 3.203 3.941 88(c) 4.461 5.848 70 4.620 4.830 60 6.780 9.115 • n-butanol 100 4.163 4.034 80 4.144 4.034 62 (c) 69.250 (a) 3 2 .2 2 2 ^ iso-butanol 100 4.681 5.200 80 4.670 4.496 67 (c) 37.960 (a) 47.037(a) ethanol 95^ 4.015 80 4.924 4.700 methanol 62 23.860 28.387 water 199.100(b) separation Into two phases occurred \l cited from Majd and Nickerson (1976) (c azeotroplc mixtures (mixtures exhibiting a constant boiling point and composition) lactose and all the other solvent systems are poor in comparison* The solubility of pure lactose in solvents was somewhat dependent upon the extraction temperature. In qeneral, lower extraction temperatures were associated with greater solubility. This observation is in contrast to the report by Moye (1972) who stated that solubility would increase at hiqh temperatures. However, the temperature effect on solubility is presumably related to the disruption of crystal lattices and would be important only at temperatures above the meltinq point of the suqar. The reason for better solubility at low temperature, within the moderate ranqe, may be related to the elevation of the dielectric constant for the solvent under these conditions. The presence of water in the solvents had a pronounced effect, and the solubility rapidly increased with increasing water concentration for each solvent. Thus it apoears that it is mainly the water phase Which determines the solubility of pure lactose* For azeotropic mixtures, it was observed that not all solvent mixtures remained as a sinqle phase when lactose was dissolved in the systems. For example, 62% n-butanol demonstrated the qreatest solubilizinq power but the solvent system was unstable and separated into two layers. Presumably the presence of lactose chanqed the polarity of the water phase, making it less miscible with the pure 84 solvent. Prom Table 6 , methanol ( 6 ?%) was shown to be relatively effective as a solvent for lactose at both temperatures. Amonq the azeotropes, possible solvents would be 72% n-propanol and 60% iso-propanol, however, the solubility of lactose would be substantially smaller than in 62% methanol. b) Combination of Selected Solvents This study was conducted to examine if combinations of the two most effective solvents would improve lactose solubility. The effects of combination are Illustrated in Table 7 and the analysis of variance is qiven in Table 6 . The assumption that solubility is a function of the solvent ratio was siqnifleant at the 5% level. The data suqqest that a 3:1 mixture of the two solvents (60% iso-propanol and 72% n-propanol) improved the solubility 16% over the sinqle solvent (60% iso-propanol). However, this combination was not as effective as 62% methanol. c) Effects of Aqitation Time on Solubility The contact time with aqitation for solubilizinq lactose in 62% methanol and 60% iso-oropanol was studied for periods up to 8 hours. The r e s u lts are shown in Fiqure 13. In qeneral, the solubility of lactose increased by increaslnq aqitation time. Aqitation for 62% methanol had a 85 Table 7 Effect of combination of selected solvents on pure lactose solubility (5:95 lactose to solvent) i t Selected solvents Soluble lactose (mg/ml) Ratio • A • B I II III Mean 1 0 6.92 6.56 6.86 6.78 0 1 5.69 5.76 5.77 5.74 1 1 4.38 4.32 3.99 4.23 2 1 7.90 7.92 7.85 7.89 3 1 8.05 7.98 8.00 8.01 1 2 5.88 5.81 5.89 5.86 1 3 5.91 5.96 6.06 5.98 * A : 60* Iso-propanol (non-azeotrope) B : 72% N-propanol (azeotrope) 86 Table 8 Analysis of variance : Combination of selective solvents on pure lactose solubility at 25°C Source Df SS MS F Fg g j F0.05 Model 1 6.58 6.58 4.95 8.18 4.38 Error 19 25.27 1.33 - - Corrected 20 31.85 Total * Significant at p<0.01 levels. iue 3 fet f gtto tm o pr lcoe solubility lactose pure on time agitation of Effect 13 Figure SOLUBLE LACTOSE(mg/mO 0 3 20 10 0 0 n eetd solvents selected In 2 AGITATION TIMEthrt) o o 60% 60% o o 2 62% • • 4 METHANOL jsoPROMNOL 6 8 87 88 pronounced effect on lactose solubility at the beqinninq: the solubility Increased rapidly to a plateau of 22.8 mq/ml in 30 minutes. Durlnq this time interval no problem with crystallization or reprecipitation was observed. In comparison with 60% iso-propanol, 62% methanol has considerable potential as a solvent for lactose extraction, as it readily dissolves lactose at the beqinninq of a q ita tio n . 2. Extraction Conditions of Lactose from NFDM a) Evaluation of Azeotropes Azeotropes are solvent mixtures which distill with a constant composition and boilinq point. For development of a continouB extraction system or a system based on reflux principles it must be anticipated that azeotropes would be preferable for extraction and reqeneration over non-azeotropic mixtures. However, it should be recoqnizsd that the azeotroplc properties may be altered When the system incorporates solubilized lactose and other constituents. This problem has not been investiqated in this study. On the basis of pure lactose solubility in various solvent systems, further studies were conducted on lactose extraction from NFDM under azeotroplc conditions in both 89 single batch and reflux systems. From an examination of the data in Table 9, it is apparent that all of the azeotropes were capable of solubilizinq lactose from 3-22 ma/ml. In the experiments with sinqle batches, 62% methanol was observed to be an acceptable solvent? however, it is not cateqorized as an azeotroplc solvent. In contrast, the azeotrope, 89% iso-propanol, was the least efficient extraction solvent. Usinq a similar approach, lactose extraction by refluxinq was examined to define the best azeotropes for a Soxhlet system in which NFDM was treated by refluxinq at 79 °C. Samplinq was conducted at reqular time intervals and the residual lactose content in the DENFDM analyzed (Table 10). The residual lactose qraudually decreased as the reaction time in various azeotropes increased. In comparison with the sinqle extraction, it was found that lactose extraction for the Soxhlet system was not satisfactory. Specifically, more than 10% of the lactose content of DENFDM remained even after 15 hours of reflux. The reason for the incomplete extraction was due to the NFDM hardeninq in to a so lid mass Which was not p en etrated by the solvent condensate. This approach was abandoned for this reason. 90 Table 9 Effect of solvents on extraction of lactose from NFDM in a single treatment (at 25°C) (NFDM : Solvent ■ 5:95 wt/wt) cwefam Residual Lactose Extracted Lactose ViniAtv\ Solvent System content of DENFDM{%) in Sol vent (mg/ml) Y1eldW methanol (62%) 20.9 . 21.70 76.6 ethanol (95%) 77.9 3.92 13.9 n-propanol (72%) 74.2 4,24 15.0 1so-propano1 (88%) 79.2 3.26 11.5 n-butanol (62%) 37.5 16.49 58.3 iso-butanol (67%) 58.7 9.48 33.6 91 Table 10 Effect of reaction time on reflux extraction of lactose using azeotropic solvents (NFDM:solvent = 5:95) Percent of Lactose Remaining (%) Extraction time A B C 0 E (hrs) 0.5 30.4 27.8 32.2 31.7 31.2 1 28.8 25.7 26.5 22.8 25.4 2 25.4 23.9 27.8 23.1 20.2 3 27.3 21.8 27.8 23.6 24.1 4 19.7 20.5 18.6 19.1 19.9 15 17.6 19.3 18.2 9.8 19.2 A: 95% ethanol B: 72% n-propanol C: 88% 1so-propanol 0: 62% n-butanol E: 67% iso-butanol 92 b) Extraction Temperature The possibility of utilizing selected solvent * extractin of lactose from NFDM has been examined at two different temperatures. The solubility of lactose in methanol, which was identified as a promislnq medium, is depicted in Table 11. On the basis of these results, it was concluded that cold treatment had little effect on lactose extraction over using room temperature. These findings are in agreement with the previous experiment for pure lactose at two different temperatures {Table 6 ). c) A g itatio n Time The , extraction efficiency of lactose in this study has been conducted to explore optimum agitation time in a single solvent treatment at room temperature and the results are shown in Figure 14. Onthe basis of analysis of variance. Table 12, it was determined that the agitation time had a significant effect on the lactose solubility at the 1% level. This finding, statistically, shows that aqitation miqht be inferred to he importance in the solubility behavior. The extracted lactose in the solvent increased rapidly as the aqitation time was increased. Further agitation beyond 90 minutes, however,had little effect. The following equation with regression coefficient of n.fU4 was 93 Table 11 Effect of temperatures on lactose extraction 1n 62% methanol Temperature (°C) Lactose content Extracted lactose Extraction of of DENFDM {%) in solvent (mg/ml) lactose (%) . 25 32.60 21.70 76.8 5 31.11 22.48 79.6 Table 12 Analysis of variance by general linear model procedure: agitation time on lactose extraction from NFDM Source DF SS MS F PR>F F0.05 F0.01 Model 1 620.83 620.83 61.40* 0.0001 4.60 8.86 Error 14 141.55 10.11 - - - Corrected total 15 762.38 * * significant at p<0.01 EXTRACTED LACTOSE IN SOLVENTlmg/ml iue 14 Figure 25 5 3 h efc o aiain ie n ouiiy f lactose of solubility on time agitation of effect The 0 3 AGITATION TIMEfmln.) 0 6 0 9 120 150 180 94 95 developed usinq qeneral linear model analysis (SAS): Y * 9.128 + 1.504 * JT Where Y represents soluble lactose (mq/ml) and x is aqitation time (minutes). From this equation, the solubility of lactose in 62% methanol can be calculated for any aqitation time, provided that other factors remain constant. The qoodness of fit as shown in Fiqure 15 indicates that the model fits well with a reqression coefficient of 0.9029. d) Effect of Repeated Solvent Treatment A study was carried out to determine the relationship between lactose extraction and number of cycles of fresh solvent treatment. The initial powder-to solvent r a tio was 5:95 and the time o f a q ita tio n employed was 90 m inutes a t room tem perature. The r e s u lts are shown in Fiqure 16 and indicate that the lactose extraction from NFOM decreased with increasinq number of cycles. Examination of the mass balance in Table 13 showed a recovery o f 126%. The hiqh value indicates a discrepancy in the experimental technique or in the method of analysis. The colorimetric phenol-sulfuric acid test used in this investiqation is a hiqhly sensitive test which requires extensive dilution of the samples. Besides, the color reaction is readily influenced by contamination of lint or other impurities iue 5 odes f t fr rdcig ats solubility lactose predicting for it f Goodness of 15 Figure PREDICTED VALUE 30 5 3 25 20 5 1 10 OBSERVED VALUE 5 20 15 faO.9029 25 96 Table 13 Mass balance of lactose content NFDM 1st Ext. 2nd Ext. 3rd Ext. 4th Ext. Solvent Solvent Solvent Solvent Total 12 g 273 ml 182 ml 121.3 ml 80.86 ml Cone, of lactose 51. 9% 32.38 mg/ml 15.48 mg/ml 7.23 mg/ml 3.81 mg/ml Amount of lactose 10.90 g 8.85 g 2.817 g 0.8 g 0.31 g Total lactose 10.90 g ------12.77 A. Lactose in in itia l powder 10.90 g B. Lactose in residual powder 1.00 g Lactose from 4 extractions 12.77 q Total Lactose 13.77 g From A and B, Recovery ; 126% VO iue 6 fet f eetd ovn tetet n ats extraction lactose on treatment solvent repeated of Effect 16 Figure EXTRACTED LACTOSE IN SOLVENT(nfl/nl) 40 from NFDM NO. OP CYCLES OP NO. 98 containinq carbohydrate. It is likely that the test is inadequate for investiqation of this type and identifies a need for improved methods of analysis of lactose. Nevertheless, the trend depicted in Fiqure 16 clearly shows that even after several cycles of solvent treatment there is incomplete extraction of lactose and the limit appears to he 4 to 5% residual lactose in the DENFDM. These observations, therefore, do not confirm the findinqs by Leviton (1444) who reported almost complete extraction after 7 minutes of extraction in a sinqle batch (5:95 ratio). The main differences appear to be in the temperature and the aqitation time employed. Leviton (1949), conducted a rapid extraction at -15 °C, whereas in this study extraction was done a t room tem perature over a lonqer period o f tim e. The incomplete extraction and the decreasinq deqree of saturation for the cycles followinq the initial extraction suqqest that the physical state of the lactose may be another factor. Amorphous lactose is more readily extracted than the crytallized form (Leviton 1949, Choi et a l. 1451, Olano 1978) and poses the question Whether or not the lonqer aqitation time may alter the physical state. 3. Extraction of Lactose from NFDM in a Pilot System On the basis of previous results, lactose e x tra c tio n was conducted in 67% methanol a t room tem oerature 100 in a larqer system as depicted in Materials and Methods* Optimum a q ita tio n fo r e ffe c tiv e la c to se e x tra c tio n was determined for the first and second extraction step usinq 5 and 20ft slurries, resoectively. Fiqure 17 shows that larqe quantities of lactose were extracted within one hour at the beqinninq of the first extraction step. Further aqitation, however, had little effect. For the second extraction in a 20% slurry, the plateau was reached late (after 4 hours) aqaln suqqestinq that different ohyslcal forms of lactose are solubilzed at different rates in this solvent. Table 14 summaries the conditions for the extraction of lactose in the pilot system. The removal of lactose was examined at each step in order to determine the extraction efficiency and mass balance. The results are shown in Table 15. It was observed that the qreatest portion of lactose was extracted in the first step (50.2%). The efficiency of lactose removal at the second step was 56.04%. Total extraction of lactose was 94.97% at completion in the pilot system . 4. Protein Loss from NFDM durinq Solvent Treatment A desirable extraction method was souqht that would maximize lactose solubilization while minimizinq the solubilization of protein and other non-lactose constituents. iue 7 fet f gtto tm o etato o lcoe by lactose of extraction on time agitation of Effect 17 Figure EXTRACTION OP LACTO8E(mg/m0 20 10 30 o 0 sn 6% ehnl n plt system pilot 62% a 1n using methanol 2 AQITATIONTIME(hrs) extraction t s i o o • • 2 d extraction nd 4 o - 6 101 102 Table 14 Extraction conditions for lactose removal from NFDM In a pilot system Step Cone, of Agitation Temp. (°C) Separation slurry (X) time (hrs) Tech. 1st Ext. 5 2 25 Gravity Sedlmen, 2nd Ext. 20 4 25 Gravity Sedlmen, Table 15 Extraction efficiency of lactose at each step In a pilot system NFDM 1st Ext. 2nd Ext. DENFDM Total amount 453.6 g 6150 ml 6100 ml 180.9 g Cone, of lactose 51.9 X 23.0 mg/ml 13.3 mg/ml 3.80 X wt. of lactose 235.4 g 141.79 g 81.31 g 6.87 g Step yield (X) 0 60.23 86.84 Total yield (X) 0 60.23 94.77 103 The initial approach was conducted to determine the relationship between protein loss and extraction temperature. Results are illustrated in Fiqure IB. . The protein loss was closely correlated with temperature and, in qeneral, the protein loss increased with increasinq e x tra c tio n tem perature. Findinqs shown in Fiqure IB reveal that the protein loss was qreater durinq the first extraction than the second extraction. In another set of experiments, protein loss was evidently depended upon concentration of methanol (Fiqure 19). In the first extraction, hiqh levels of methanol concentration (> 63%) caused a rapid decrease in protein loss. In the second extraction, less alcohol-soluble protein was present but, once aqain, the protein loss qradually decreased with increasinq concentration of m ethanol. Table 16 p re se n ts a summary o f p ro te in lo sses which for the two-step process caused a 10% loss. The type of protein lost in each step was identified by urea starch qel electrophoresis as shown in Plate I. Accordinq to the electrophoretic pattern there was a loss of all proteins. B. Physicochemical/Functional Analysis of DENFDM 1. Analysis of Composition The composition of DENFDM, commercial casein and iue 8 fet f eprtr o poen os from NFDM loss during protein on temperature of Effect 18 Figure PftOTEM OF LOSS(%) 12 ie 2* -.' 4*c -0.0'e -20*e ovn treatment solvent TEMPERATUREOP EXTRACTKHK o) isd 2nd extrction 2nd isd m m total t extrctiop st 1 28% 104 iue 9 fet f ocnrto o mtao o poen loss protein on methanol of concentration of Effect 19 Figure PROTEIN LOSS(mg/ml) CONC. OF METHANOL(%) OF CONC. 50 6Q 704030 105 80 106 Table 16 Protein loss In solvent extraction of lactose from NFOM In a pilot system NFDN 1st Ext. 2nd Ext. DENFDM Total amount 453.6 g 6150 ml 6100 ml 180.9 g Cone, of protein 34.58 % 2.04 mg/ml0.53 mg/ml 78.6 t wt. o f protein (g) 156.85 12.54 3.22 142.19 Step loss of yield (X) 0 7.99 2.23 Total loss of yield (cumulative, %) 0 7.99 10,05 Plate I. Identification of protein loss during solvent treatment by urea starch gel electrophoresis A. NFDM (4%) B. DENFDM (4%) C. 1st Ext. solid 0. 2nd Ext. solid E. as-case1n (2%) F. e-casein (2%) G. k-case1n (2%) H. fi-lactoglobul1n (2%) 1. whole casein (2%) ABCDE FGHI 108 NFDM Is qiven in Table 17. The maior differences in composition between DENFDM and NFDM are in their protein and lactose contents and in their mineral content and composition. Delactosed NFDM is intended as a replacement for calcium caseinate and in this respect the composition of DENFDM approaches that of calcium caseinate with the exception that about 3-4% residual lactose remains alonq with a qreater proportion of milk salts. The minimum lactose content attainable was found to be about 3.8% in DENFDM, which is somewhat hiqher than for commercial, co-precipitated casein (0.7 - 1.5%) (Morr, 1987). The quantity of ash was markedly different depending upon the product. DENFDM contained 9.3% ash, whereas commercial casein contained 3.8% ash. The differences in ash content may have an influence on different functional properties dependinq upon the type of mineral elements. The major differences between DENFDM and commercial casein can be expressed by the rato of divalent to monovalent cation and in the Ca/Na r a t io . The d ata for DENFDM revealed intermediate position in these ratios and a tendency for a preferential extraction of monovalent cations over divalent sp ecies. 109 Table 17 Chemical analysis of products Commercial ca sein a te^ DENFDM NFDM Protein (%), N x 6.38 88.5 78.6 34.58 Fat (X) 0.2 0.8 1.02 Carbohydrat^lx) - 3.8 51.9 Moisture {%) 7.0 4.25 3.35 Ash (%) 3.8 9.30 7.87 Mineral (mg/g) Ca++ 13.887 24.241 12.57 Phosphorous 6.764 12.779 9.68 K+ 0.554 6.172 8.128 Mg++ 0.218 1.380 1.100 Na+ 0.138 1.670 5.350 Fe++ 0.039 0.021 0.032 Zn++ 0.041 0.074 0.041 Mn'H‘ 0.005 0.001 0.001 Cu++ 0.002 0.003 0.004 Ca/Na 100.6 14.5 2.35 Cations divalent 20.5 3.28 1.02 monovalent (a) From Morr, 1982 (b) Indicates lactose content 110 2. Free Amino Groups Alpha amino nitroqen has been correlated to the formation of larqe quantities of polypeptides as intermediates in the hydrolytic reactions of protein under acidic conditions (Hokes, 1982). Free amino qrouos of DENFDM by formol titration is shown in Flqure 20. Alpha amino n itro q en was shown to be q r e a te r in calcium ca se in a te than in both DENFDM and NFDM on the basis of sample weiqht. Alpha amino nitroqen, however, was identified to be the lowest in DENFDM on the basis of protein content. It was apparent from the data that the solvent treatment has induced a loss in free amino qroups. 3. Size Distribution and Bulk Density Particle size and size distribution of NFDM and DENFDM as well as for calcium caseinate as determined by liq h t microscopy are shown in Table 18 and in Appendix A. The solvent treatment caused an increase of about 25* in the particle size of the DENFDM with respect to volume diameter. In comparison with calcium caseinate, DENFDM was characterized as havinq a substantially larqer ( **60%) mean particle size. It was observed that the particle size of all products was widely distributed with a ranqe of 25 to 250 micrometer. The microscopic measurement, however, had a limitation because it was adaptible primarily to the iue 0 -mn ntoe dtrie I I dispersion It In determined a-am1no nitrogen 20 Figure CI'AMMO NTTftOQEN (mQ) 4.0 5-0 3.0 60 2.0 1.0 a CASEMATECa y oml titration formolby ■ : -mn ntoe (gg protein) (mg/g sample) (mg/g nitrogen a-amlno nitrogen a-amlno : ■■■ ; 3 1 DENFDM NFD U 1 112 Table 18 Particle size distribution and bulk density of experimental samples Diameter Ca-caselnate (pm) DENFDM (pm) NFDM (pm) Arithmetic mean 37 61.5 45.5 Volume-surface mean 80 114.5 101.5 Height mean 106.5 127 120.5 Surface mean 45 75.5 57.5 Volume mean 54.5 86.5 69.5 Bulk density (g/ml) Loose packing 0.282 0.526 0.500 Close packing 0.385 0.571 0.704 113 particle size class within the ranee of 2 - ion micrometer excludinq larqer or smaller particles. Since Hayashi et a l. (1969) reported the particle size distribution of NFDM to he in the ranqe of 10 to 250 micrometer, an exclusion of the larqest particle sizes may have occurred in this study. The r e s u lts in Table IS showed th a t p a r tic le size and bulk density were not related factors. The bulk density Is an important physical property because it Is closely related to dispersibility and wettability. The results revealed that the bulk density of DENFDM was siqnificantly qreater (almost twice as dense) than that of calcium caseinate Which was shown to be 0.282 q/m l. The bulk d e n sity o f DENFDM, however, was somewhat lower than that of NFDM, a fact related to the removal of lactose. The results for the DENFDM shown In Table lfl reflect the actual particle size after freeze drylnq without qrlndinq. The findinqs suqqest that solvent extraction of lactose from NFDM cause some swellinq of the particles which persist after drylnq. The particles appear to become more porous, as reflected in the bulk density measurements and in their physical structure. This effect is a consequence of the removal of lactose. In comparison with calcium caseinate, DENFDM has a qreater particle size but also a hiqher bulk density, reflectinq the differences In their composition and physical structure. 114 4. Colloidal Stability Caaein-pellet-solvation was determined at 1.0*, 2,5* and 5.0* by centrlfuqation of the dispersions at 25,000 x q for 60 minutes at room temperature. Wet pellet wsiqht was recorded and insolubles were calculated as the percent of dry pellet weiqht of initial weiqht. Hydration of the insolubles was calculated as the amount of water bound by th e sedim ent. The r e s u lts are recorded in Table 10 and show that the sedimentable fraction from DENFDM is substantially qreater than for NFDM and calcium caseinate. The hydration of the sediment was also the hiqhest for DENFDM in comparison with the other two samples at 1* dispersion. It is noteworthy that the maqnitude of the hydration ratio increased with increainq total solid content. The possibility of incomplete drylnq of the pellets may have been a factor in this experiment and for this reason caution must be exercised in assessinq siqnificance of these results until verified by separate experiments. 5. Water Absorption and Viscosity Profile Water absorption was determined by observinq the decrease in viscosity of a starch slurry as a function of added protein usinq a Vlscoamyloqraph instrument. Water absorption is likely to be related to the swellinq behavior which takes place. durinq the competition for water by the 115 Table 19 Colloidal Stabilities and hydrablllty Sample % Cone. Wet pellet Dry pellet Insoluble Hydration* weight (g) weight (g) <*) ratio Ca-case1nate 1.0 0.1498 0.0257 17.2 1.51 2.5 0,3574 0.1333 37.3 2.82 5.0 0.9218 0.6406 69.5 3.31 DENFDM 1.0 0.2120 0.0557 26.3 1.98 2.5 0.5360 0.3125 58.3 2.74 5.0 1.0187 0.7702 76.9 3.27 NFDM 1.0 0.1165 0.0249 21.4 1.08 2.5 0.2499 0.0736 29.5 2.14 5.0 0.5153 0.2940 57.1 2.60 * Expressed as the ratio of weight of hydrated water to dry weight of sedlmented protein 116 starch and protein particles (Hokes, 19fl?). The behavior of water absorption is depicted in Fiqure 21* The development of viscosity durinq the qeiatlon staqe was different for all samples and differed from the control of starch alone. NFDM and calcium caseinate addition caused early development of viscosity, Which would be expected because of the increase in total solids. In contrast, addition of DENFDM caused a reduction in viscosity, indicatinq that this sample absorbed water preferentially over starch. The differences in maximum viscosity have been quantitated. NFDM (670 B.U.) was similar to the control (640 B.U.). DENFDM (50? B.U.) i and calcium caseinate (44? B.U.) depressed the maximum viscosity in comparison with the control. In principle, milk protein competes for the waterwhich is available for starch qelatinization and therefore the maximum viscosity for the samples containinq calcium caseinate and DENFDM were decreased. The amount of protein added was least for NFDM and compensatinq effects from lactose and other constituents may be re sp o n sib le for the minimum e f fe c t o f th is sample. The water absorption data in this experiment were rouqhly correlated with thesolvation data, conforminq that interactions with water was a shared property, for calcium caseinate and DENFDM. 700 HEATMQ ♦ COMTUT W COOUNQ 0 0 0 •00 65. 400 •00 1 M ira CLEER STARCH 2 Ci-CASEINATE* Hatk 200 3 DENFOM+swck 4 NFDM + JUtt* 100 58 Figure 21 Water absorption and viscosity profile by viscoamylograph 118 6. Examination of Particle Structure by LM and SEM Since the commercial SEM became available in 1965, the three-dimensional microstructure of materials as revealed by the SEM has been used to explain and predict physical and chemical behavior of materials. Typical spray-dried calcium caseinate and NFDM by SEM are illustrated in Plate II and III. In qeneral, the two products were observed to consist of hollow spherical particles and partially collapsed spheres. The surfaces were qenerally smooth and varied in the extent of collapse. Some particles exhibited wrinkled exterior surfaces and aqqlomeration was evident as well as adherence of small particles to larqe .spheres. In qeneral, the microstructure of NFDM and casein cannot be distinquished by SEM, but some difference may be assiqned for the two powders based on polarizinq microscopy; e.q. Plate III-A shows more birefrinqent material on the particle surfaces of NFDM than in calcium caseinate. This material may be related to crystalline lactose and minerals. In contrast to the other two samples, DENFDM was observed to have unique surface characteristics as shown in Plate IV. The three-dimensional structure of freeze dried DENFDM exhibited predominantly porous particles, some of which were broken or damaqed. In comparison with NFDM, the particle shapes were much more irreqular, rouqh, coarse and 1X9 Plate II. Microstructure of calcium caseinate A. Polarized light microscopy (PLM) photomicrograph(xl28) B. Scanning electron microscopy (SEM) (x500) A 0 ,0 .0 ., 120 Plate III. Microstructure of NFDM A. PLM photomicrograph (xl28) B. SEM photomicrograph (x500) 121 Plate IV. Microstructure of DENFDM* A. PLM photomicrograph (xl28) B. SEM photomlcrooraoh (x500) 122 hollow. Particles were also aqqlomerated to each other and showed qreatly different particle size distribution. Examination of DENFDM powder by liqht microscopy showed that the particles were aqqlomerated and their shape was ir r e q u la r . The p o larized microqraph showed some c r y s ta llin e material, oresumably minerals or incompletely extracted la c to s e .. It is of Interest to compare the particle structure of the DENFDM with the structure of the parent material, NFDM. The removal of lactose by solvent extraction has left void spaces in the particle walls causinq the porosity. The apparent brittleness and fraqile nature is probably a result of the removal of lactose as a cementinq material. Inspection of the broken particles suqqests that the interior walls are still smooth or that the extracted lactose must have been removed from the outer surface. The ex tra c te d so lid which was obtained by evaooratinq the first solvent extract to dryness revealed the presence of much crystalline lactose with small amounts of amorphous material, which lacked uniform and smooth surface characteristics, adherinq to the crystals (Plate V). The solids in the second solvent extract were very different from the first extract. Plate VI shows comparatively few lactose crystals which were entrapped in a protein matrix. The results confirm that most of the lactose is extracted in 123 Plate V. Hfcrostructure of first extracted solid A. PLM photomicrograph (_xl28 B. SEM photomicrograph (x500 124 Plate VI. Microstructure of second extracted solid A. PLM photomicrograph (xl28) B. SEM photomicrograph (x500) 125 the first step and that the subsequent extractions remove principally protein constituents. Leviton (1949) has reported that the solvent extract is an excellent source for hiqh qrade lactose. Recrystallisation from 95% ethanol of the dry solids in a mixture of the first and second extracts showed well formed c r y s ta ls o f la c to se o f tomahawk shape Which comfirm th a t the lactose is undamaqed (Plate VII). C. Solubility Profile of DENFDM It was reported by Leviton (1949) th a t DENFDM is composed of readily soluble proteins. Morr and Lin (1970), however, rep o rted th a t delactosed whey s o lid s by the same procedure were poor in solubility. Preliminary experiments revealed that DENFDM prepared in this study was nearly insoluble. For this reason, detailed experiments wore conducted to examine the solubility behavior of the product under various conditions. 1. Effect of pH on Solubility Fiqure 22 is a qraphic representation of data showing the effect of pH at 25 °C on the solubility of DENFDM. In comparison with NFDM and calcium caseinate the solubility at the normal pH was very ooor (<\, 20%), However, 126 Plate VII. Microstructure of recovered lactose crystal A. PLM photomicrograph (xl28) B. SEM photomicrograph (x500) 20 • • NFDM o o Ci-CASE I NATE + * OENFDM 1 0 Figure 22 Effect of pH on solubility 128 the profile suggested a similarity with calcium caseinate. The pH of the solutions strongly affected protein solubility. All three sample dispersions exhibited the least solubility at pH 4 which is near the isoelectric point for casein. The 18-20* solubility of protein in NFDM. in this case, corresponds to the presence of undenatured Whey protein which is not precipitated at this pH. The fact that a smaller quantity protein was soluble at pH 4 fo r the DENFDM suggests that a portion of the Whey orotein was denatured. The increased solubility on the alkaline side of pH 4 for all three samples is consistent with the proteins becominq negatively charged. It must be noted that the DENFDM showed resistance towards alkali solubilization but eventually approached the same solubility as for the other two samples. The reason for this appears to be either Whey protein denaturation or the effect of mineral salts or a combination of both. Thus, to produce a soluble DENFDM it may be necessary to consider a preliminary extraction of the calcium salts. 2. Effect of Urea on Solubility In general, solubility studies with urea solution have been applied to study structural bonding changes. Solubility behavior of DENFDM in different concentrations of urea was investigated as shown in Plate VIII which revealed t 129 Plate VIII. Effect of urea concentrations on protein solubility of DENFDM A. 1 molar urea solution B. i molar urea solution C. 3 molar urea solution D. i molar urea solution E. 5 molar urea solution F. ( molar urea solution !W i n » ? A B C D E F 130 that the solubility of DENFDM was qradually Increased as the urea concentration increased. At one molar urea solution, DENFDM was still relatively insoluble. Delactosed NFDM in both two molar and three molar urea solution demonstrated a milky dispersion, and at hiqher concentration the product was soluble* The effect of urea suqqests that the insolubility problem is associated with hydroqen bondinq, both h y d ro p h ilic and hydrophobic. 3. Effect of Salt and Suqar on Solubility Neutral salts and suqars have been reported to enhance protein solubility in some systems. Since the insolubility of DENFDM has occurred durinq the extraction of suqar (lactose) and smaller quantities of principally monovalent salts an investiqation was conducted to determine solubility of DENFDM in the presence of salt (sodium chloride) and suqar (sucrose), sinqly and in combination. The in flu en ce o f s a l t and suqar co n cen tratio n on protein solubility is depicted in Fiqure ?3. The results show that salt had a small but statistically siqnificant effect on solubility, whereas sucrose did not. When a combination of salt and suqar was applied, the solubility was aqain increased by a small amount. It appears, however, that suqar had little effect on protein solubility. 131 8 4 16 24 8 0 4 4 0 ox ox ox u l t s i |i r 12 c D E F ft 32 24 NOIOIM Figure 23 Effect of salt and sugar on solubility of DENFDM 132 To predict the relationship between protein solubility and concentration of salt and suqar, statistical analysis (SAS) was conducted and the followinq multiple reqression equation was d erived: Y - 23.79 + 0.53 * M + (-0.04) * N where Y « protein solubility of DENFDM (%) M * concentration of salt (*) N - concentration of sucrose (*) The equation revealed that salt was the only Important factor, while suqar was not. The qoodness of the fit for p ro te in s o lu b ility o f DENFDM is shown in Fiqure 24. The model fits the data quite well with a reqression coefficient of 0.9315. Fiqure 25 shows the relationship between solubility and concentration of salt and suqar at any specific level by response surface methodoioqy (RSM). A qenerai prediction emerqed in which hlqh salt content with low sucrose concentration showed a positive effect on protein solubility. The reason for the salt effect is likely to be a shieldinq of the protein from detrimental calcium interactions. I t was concluded from th is study th a t althouqh s a l t and suqar affected solubility the effect did not sufficiently improve the product characteristics. iue 4 odes f t fr rdcig solubility predicting for it f Goodness of 24 Figure PREDICTED VALUER) 201 22 24 26 28 20 22 • • • OBSERVED VALUE**) riO.9315 2 430 3 622 4 28 133 Figure 25 3-D plot of solubility of DENFDM M: salt concentration (2) N: sucrose concentration (5C) Symbols: • • • • 22.53015 - 23.19045 • I l f 23.19045 - 24.51104 —- 24.51104 - 25.83163 U K 25.83163 - 27.15222 27.15222 M 28.47281 ooo 28.47281 - 29.79340 XXX 29.79340 - 31.11399 om o • o • 31.11399 - 32.43458 32.43458 - 33.75517 33.75517 - i n M CONTOUR PLOT OP K*H I 19 ♦ OOOOOOOOOOOOOOOOOOOOO xxxxxxxxxxxxxxxxxxxxx IB 4- 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 xxxxxxxxxxxxxxxxxxxxx 17 4 OOOOOOOOOOOOOOOOOOOOO xxxxxxxxxx------16 40 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 I 15 4000000 0000 0 0 0 0 0 0 X X X X X I 14 4000XXXXXXXXXXXXXXXXXX 13 4 XXXXXXXXXXXXXXXXXXXXX 12 4 xx x x x x x x x o o o o o o o o oooo 11 40 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 IB 40 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 4 4 4 4 940004444*4 4444444444444 8 6 3 4 I 4 4 3 2 4 1 • 4 ■ I-- i ■< ■> — *■■■! * >- a 9 10 11 12 13 14 IS 16 IT IB II M N ui 136 4. Effect of Aqitation on Solubility Stronq aqitation is a qeneral practice for improvinq solubility. This study was conducted to determine the effect of aqitation time and establish the condition for maximum s o lu b ility o f DENFDM. As shown in Fiqure 76, the protein solubility of DENFDM was markedly Influenced by the lenqth of aqitation time, and may not have reached the full potential even after 10 hours. Prior qrlndinq of the product by mortar and pestle had no slqniflcant effect on protein solubility. The slow solubility in water has two components, one relatinq to aqitation Which promotes the continuous exchanqe of water at the contact surface of the particle, the other is related to contact time. In view of the fact that insolubility apoears related to stronq hydrophobic interactions mediated by calcium bindinq, it is likely that the conformation of the protein in DENFDM particles underqo a slow return to a state Which burries the hydrophobic qroups in the interior and exposes the hydrophilic qroups to the exterior. Thus, to promote s o lu b ility i t is a lso necessary to consider processes Which will accelerate this conformational alteration. 5. Effect of Sequestrants on Solubility The implied role of calcium in causinq DENFDM to be insoluble suqqested the use of calcium sequesterinq/or 137 0 o o DENFOM GRINDED 4 AQITATION TIME(hr«) AQITATION 2 DENFDM Figure Figure 26 Effect of agitation time upon the solubility of NBXOHd 138 precipitating agents for solubilizing the product* The effect of addition of a number of additives on protein solubility is illustrated in Fiqure 27. Results show that phosphate compounds are generally effective. Sodium metaphosphate was the most effective# and increased the solubility four fold over the control. Potassium or sodium diphosphate were also effective in enhancinq protein solubility. In subsequent studies (Figure 29), when larqe amounts of sodium metaphosphate were used the optimum co n cen tratio n was id e n tifie d as 3*. 6. Effect of K-carraqeenan on Protein Solubility of DENFDM Carrageenan has been reported to stabilize caseinate against the precipitating influence of calcium ions. The addition of k-carraqeenan was employed to assess the potential improvement of protein solubility and the results are depicted in Fiqure 29. Solubility of DENFDM increased sliqhtly with the addition of 0.^4%, but further improvement could not be achieved at hiqher concentrations. The solubility of calcium caseinate was also improved with the addition of k-carrageenan but the maximum effect occurred at hiqher level of addition. The experimental finding that k-carrageenan to some extent improves solubility of DENFDM is consistent with the 139 Figure 27 Effect of sequestrants on protein solubility of DENFDM A. Sodium tartrate 8. EDTA C. Magnesium citrate D. Sodium acetate E. Potassium phosphomonobaslc F. Sodium phosphodlbaslc G. Potassium phosphodlbaslc H. Sodium metaphosphate I. DENFDM as a control J. NFDM K. Ca-case1nate SEOUCtTRANT* 100 201 ■ - — - - — '■* 0 to 20 30 COMC. OP SOD. METAPHOSPHATE(%) Figure 28 Effect of concentration of sodium metaphosphate on protein solubility of DENFDM 142 100 90 £ 1 80' 3 0 O o Ci-CASEINATE • • DENFDM 20 8 12 16 CONC. OF K-CAIIIIAQ8ENAN(%) "It?2 Figure 29 Effect of k-carrageenan addition on protein solubility of DENFDM 143 derived conclusion that calcium is partly reu ponsible for the insolubility. The effect of calcium can be minimized either by sequesterinq with ohosphates or by blockinq the calcium sensitive qroups with carraqeenan. 7. Effect of Sialic Acid on Solubility The previous experiments identified ink olubilitv of DENFDM as a result of stronq hydrophobic interactions mediated by the presence of calcium ions, Another consideration relates to the possibility that k-casein may have been extracted in this process or possibly modified by the action of alcohol. The k-caseln fra* itlon of milk protein contains a qlycomacropeptide (GMP) w1 hich can be released by rennet action. This portion co Otains sialic acid which may be used as an index for k-casein or qlycomacropeptide. a) Extraction of Sialic Acid The extraction of sialic acid contain nq compounds from NFDM is summarized in Table 20. Du rinq solvent treatment, 16.5% of sialic acid was extracted in the first extraction solvent, whereas the second treatment caused only a 4.0% lo s s . The mass balance o f s ia l i c acid n the e n tire extraction process is presented in Table 21. T1 ie 20.5% lo ss of k-casein, as measured from the sialic lata, must be 144 Table 20 Extraction of sialic acid Sample Content of Sialic acid Content of R a tlo ^ sialic acid loss (X) protein Calcium caseinate 0.374 X a» 86.56 X 0.432 DENFDM 0.140 X 19.4 72.24 X 0.193 NFDM 0.120 X 0 34.58 X 0.347 1st Ext. solvent 12.38 ug/ml 16.5 2.52 mg/ml 0.490 2nd Ext. solvent 4.78 ug/ml 4.0 3.48 mg/ml 0.140 100 ml m llk ^ 10-12 mg(w/w) a-caseln complex1tb,0.41 X(w/w) as-case1n ^ 0.07 X(w/w) B -caseln^ 0.21 X(w/w) K -casein^ 1.81-2.14 X(w/w) NFDM*b) 0.09-0.14 X(w/w) c a s e in ^ 0.26-0.59 X(w/w) (a) calculated as sialic acid to protein (b) cited from Marler et a l. (1963) 145 Table 21 Hass balance of sia lic acid during methanol extraction NFDM DENFDM 1st Ext.sol vent 2nd Ext.solvent Yield 1.00 g 0.69 g 16.00 ml 10.00 ml Cone; of sialic acid 0.12 g 0.14 X 12.38 ug/ml 4.78 jig/ml wt. of sia lic ad d 1.20mg 0.97mg 0.20mg 0.05mg Loss of sialic acid 0 19.40 X 16.50 X 4.00 X considered as a contributing reason for the induced insolubility and may be responsible for the tendency of the particles to agglomerate. b) Effect of Methanol Concentration on Loss of Sialic Acid In order to determine if the loss of k-casein observed in the previous study could be minimized, a study was conducted on the effect of concentration of methanol on loss of sialic acid. The results are depicted in Figure 30. Surprisingly the loss of sialic acid decreased as the concentration of methanol increased in both the first and second extracted solvent. It appears that loss of sialic acid is attributed to the water fraction rather than the methanol fraction. The optimum concentration of methanol for lactose extraction is 6?% and the effectiveness decreases with more concentrated methanol solutions. Therefore, it is not possible to minimize k-casein loss without also decreasing the efficiency of lactose extraction. c) Effect of Feedback of Extracted Solids on S o lu b ility This experiment was conducted to determine if the insolubility of DENFDM was caused by the removal of specific iue 0 fet f ocnrto o mtao o ls of loss on methanol of concentration of Effect 30 Figure LOSS OF SIALIC ACIX%) 10 18 14 60 alc acid lic ia s CONC. OF MCTHANOU%) 70 80 90 100 147 148 soluble materials or by the action of the solvent durinq the treatment. In this study, the effect of addition of the dried extracted solids (0.116% sialic acid) on solubility was investiqated (Fiqure 31). The dried extracted solids were prepared by concentratinq and freeze dryinq the mixture of the first and second extracted solvent. Fiqure 31 revealed that protein solubility of DENFDM was sharply increased by 10% addition of the extracted solids and, thereafter, markedly decreased. Addition of pure lactose and dialyzed extracted solids (0.056% sialic acid) had only a small effect. The effect of feedback of extracted solids suqgests that there was in fact a loss of solubility caused by extraction of a stabilizinq entity. It is Dossibie that this substance is surface k-casein. However, the feedback of dialyzed extract, which retains k-casein, was ineffective and, therefore, does not support this concept. D. Characteristics of Imitation Processed Cheese Formulated from Delactosed NFDM 1. Formulation and Preparation of the Imitation ' Processed Cheese One of the qoals for makinq DENFDM has been to convert domestic NFDM into a protein preparation suitable for replacement of imported calcium caseinates in the PROTEIN S0LUB1UTY(%) 5 4 5 2 5 3 55 iue 1 fc of edak etatd olds on s lid so extracted f o feedback f o ffect E 31 Figure 0 rti s ubiiy of DENFDM f o ility b lu so protein 20 ADDITION (%) ue lactose pure ¥ * o o • • extracted solid solid extracted • • 0 4 e extracted zed 60 0 8 149 150 imitation processed cheese industry. To assess the functional performance of DENFDM, samples of the imitation processed cheese were prepared and evaluated. The products were made according to a formulation published in the oatent literature. The preparation was modified for equipment available in our laboratory, orincipally to use the farinoqraph douqh maker. 2. Meltinq Area In order to assess meltinq characteristics, lncludinq meltinq area and meltinq temperature, imitation processed cheese on a laboratory scale was prepared, according to methods as previously described, and cut into uniformly-sized blocks for examination of their meltinq characteristics. The meltinq area, measured after microwave heatinq and expressed as centimeter per qram of sample, is shown in Table 27. Significant differences were observed (Table 23) dependinq upon formulation. Tukey's studentized ranqe test was conducted in order to determine which pairs of means were siqnificantly different (Table 24). When Tukey's test was employed on the different IPC samples, the meltinq area obtained at the seven different formulation could be qrouoed into five classes. Therefore, the test indicated that protein source as a maior ingredient for IPC yielded siqnificantly 151 Table 22 Comparison of melting area of various IPC analogs Sample No. Products Melting area (cm2/g sampl Exp. I Exp.II Mean 1 Ca-easelna te/DEN FDM(1:1) 1.71 1.71 1.71 2 Ca-case1nate/DENFDM(2:l) . 3.18 3.98 3.08 3 DENFDM only 0.65 0.65 0.65 4 Ca-case1nate only 4.62 4.60 .4.61 5 Ca-easelnate/Na-caselnate(1:1) 3.72 3.82 3.77 6 Ca-case1nate/Na-case1nate/ DENFDM (1:1:1) 2.95 2.91 2.93 7 Commercial processed cheese (con trol) 1.73 1.75 1.74 Table 23 Analysis of variance : melting area of various IPC analogs depended upon a protein source Source DF SS MS F PR>F F0.05 F0.01 Model 6 22.14 3.690 985.92* 0.0001 3.87 7.19 Error 7 0.03 0.004 M B - Corrected total 13 22.17 * Significant at p<0.01 level 152 Table 24 Tukey's studentlzed range test : melting area of various IPC depended upon a protein source Means with the same letter are not significantly different. Alphalevel ■ 0.05, OF - 7, MSE - 0.0037 >up1ng Mean N Sample No. A 4.61 2 4 B 3.77 2 5 C 3.08 2 2 C 2.93 2 6 * 1.74 2 7 D 1.71 2 1 E 0.65 2 3 * formulation as same as Table 22 153 different meltinq areas* For purpose of evaluation, the performances were compared with the calcium caseinate sample and a commercial processed sample. The imitation processed cheese, form ulated w ith DENFDM only, showed the sm allest m eltinq a re a , w hile calcium casein ate sample showed the la rq e s t spread. However the formulation of calcium caseinate with DENFDM (1*1) was closest to the commercially orocessed sample and thus suqqests a potential beneficial use for the product as a partial replacement for calcium caseinate. 3. Meltinq Temperature In a similar fashion, the ranqe of meltinq temperature was observed. The meltinq temperature was classified into two qroups as outlined in the section on materials and methodsi (a) the initial meltinq temperature and (b) the temperature for complete meltinq, defined as the temperature at Which the block had underqone 50% chanqe of native shape. The raw data is shown in Table 35 and the statistical treatment in Table 26. Initial meltinq temperature was markedly and siqnificantly influenced by protein sources, whereas complete meltinq temperature was n ot. Since the F ratio was siqnificantly different, Tukey's stu d en tized ranqe t e s t was employed (Table 27). The r e s u lts 154 Table 25 Comparison of melting temperatures of various IPC analogs Sample No.* Melting temperature (°C) In itia l Complete 1 72.6 79.5 2 70.4 79.0 3 73.8 79.0 4 65.4 79.0 5 70.0 79.5 6 72.0 79.0 7 68.8 79.0 * formulation as same as Table 22 Table 26 Analysis of variance : Initial melting temperature of various IPC depended upon a protein source Source OF SS MS F PR>F F0.05 F0.01 Model 6 88.54 14.75 240.24* 0.0001 3.87 7.19 Error 7 0.43 0.06 - wm - Corrected total 13 88.97 * significant at p<0.01 level 155 Table 27 Tukey's studentlzed range test : Initial melting temperature o f various IPC depended upon a protein source Means with the same letter are not significantly different Alphalevel ■ 0.05, OF - 7, MSE - 0.0614 Grouping Mean N Sample 1 A 73.80 2 3 B 72.60 2 1 B 72.05 2 6 C 70.45 2 2 C 70.40 2 5 D 68.90 2 7 E 65.60 2 4 * formulation as same as Table 22 156 indicated that meltinq temperature of the Imitation processed cheese formulated by combination of calcium caseinate and DENFDM was relatively close to the control. This is in agreement with the findings for melting area. However, the calcium caseinate/DENFDM (3:1) was not significantly different from the sample containing calcium caseinate/sodium caseinate (1 11)* The ob serv atio n was made that calcium caseinate/DENFDM (lsl) was very similar to the calcium caseinate/DENFDM/sodium caseinate (1:1:1). Statistically, the desirable protein source for duplication of the commercial sample was identified as blends of calcium caseinate with either sodium caseinate or DENFDM. In this respect, the use of DENFDM appears to contribute useful functional properties as a caseinate replacement. 4. Textural Profile Analysis In order to evaluate textural characteristics of IPC analogs, an Instron Universal Testing Machine was used. The explanation for the different parameters have been shown in Figure 11. A nalysis o f v arian ce by SAS showed th a t the textures for the seven samples were influenced by protein Bource (Table 2B). Table 29 illustrates another approach to Identify the effect of - protein source on the textural characteristics. Since the F values were significantly different, Tukey's Table 28 F value of instron examination of IPC analogs Hardness Cohesiveness Elasticity Gunminess Chewiness DF 6 6 6 6 6 Error 7 7 7 7 7 Corrected total 13 13 13 13 13 F 19.25* 10.66* 352.33* 1381.93* 2.10 3.87 3.87 3.87 3.87 3.87 F0.05 Fn m 7.19 7.19 7.19 7.19 7.19 * significant at p<0.01 level 157 Table 29 Instron evaluation of IPC analogs made with different protein sources^ Sample No.*" Fracturability Hardness Cohesiveness Adhesiveness Elasticity Gunmlness Chewiness (kg) (kg) (kg-cm) (kg) (kg-cm) 1 NA(r> 6.85ab 0.30ab NA 0.60a 1.95a 1.19a 2 1.53 3.84bc 0.12c NA 0.35d 0.51d 0.93a 3 0.75 1.91c 0.20bc NA 0.29e 0.57d 0.17a 4 NA 2.73° 0.38a NA 0.44° 1.02° 0.46a 5 2.90 8.25a 0.19bc NA 0.48c 1.67b 0.78a 6 NA 4.05bc 0.25abc NA 0.53b 0.94C 0.49a 7 NA 1.15° 0 .17bc 0.57 0.17f 0.39e 0.03a (P) Each value is a mean (n=2). Means in the same column with different letters are significantly different in a confidence of 0.95 (q) 1: calcium caseinate/DENFDM = 2:1 2: calcium caseinate/DENFDM/sodium caseinate = 1:1:1 3: DENFDM only 4: calcium caseinate only 5: calcium caseinate/sodium caseinate = 1:1 6: calcium caseinate/DENFDM = 1:1 7: control (comnercial processed cheese) (r) not applicable 159 studentized ranqe test was applied and the results may be explained as follows: a) Fracturabllltyt This is the force required to produce the initial break. Some samples did not fracture under the test conditions used and are represented by NA (not applicable). Examination of the data showed that samples eontaininq DENFDM exhibited fracturablllty at application of low force. Calcium caseinate/sodium caseinate (Izl) was shown to produce the hiqhest value for the force required to fracture. When DENFDM was incorporated in blends with calcium caseinate and sodium caseinate (1:1:1), the samples fractured more readily. These results suqqest that DENFDM does not enqaqe in stronq interactions and thus produces a weaker structual matrix. b) Hardness: A pplication o f Tukey*s ranqe t e s t showed complex qroupinqs for measurements of hardness. However, it is very clear that most of the experimental IPC analoqs exhibited qreater hardness than the commercial control sample. The samples eontaininq either calcium caseinate or DENFDM alone were not siqnificantly different from the control. When sodium caseinate was added the hardness increased to hiqh values, however the incorporation of DENFDM in such mixtures had a moderatinq influence. It would appear that DENFDM causes a weakeninq of the cheese matrix an effect which may be related to its solvation 160 c h a r a c te r is tic s . I t was shown p rev io u sly th a t DENFDM delayed the qelatinization of starch (Fiqure 21); in this experiment, it appears that DENFDM also interferes with the water absorption of sodium caseinate. c) Cohesiveness: This parameter is derived as the ratio of the areas of the second bite to the first bite. Calcium c a se in a te was shown to produce the h iq h e st values for cohesiveness when i t was used sin q ly . In c o n tra s t, when DENFDM was used as a sinqle inqredient, cohesiveness was very low. However, blends eontaininq DENFDM oroduced cohesiveness values not siqnificantly different from the c o n tr o l. d) Adhesiveness: This parameter relates to the capacity of the components of the broken cheese to stick toqether and to the probe. Only the control was demonstrated to possess measurable adhesiveness. DENFDM and others were not adhesive at all. e) Elasticity: This parameter is derived from a relationship between the first and second bite (Fiqure 11). Imitation processed cheese analoqs were qrouped into six classes. The results showed siqnificant differences for all the samples with the exception that calcium caseinate and its blend with sodium caseinate were similar. The commercial sample was least elastic and of the experimental samples the DENFDM inqredient produced the closest aqreement 161 with the control sample. £) Gumminesst This parameter is derived as a product o f hardness and cohesiveness. Tukey's ranqe t e s t showed fiv e d if f e r e n t qroups dependinq upon form ulation. Gumminess was pronounced in all samples eontaininq calcium caseinate. Delactosed NFDM increased qumminess compared with the control but not as much as did the calcium caseinate sam ples. q) Chewiness: This parameter is derived as the product between qumminess and elasticity. Tukey's ranqe test demonstrated that all samples belonqed in iust one identical qroup. The experimental error for this set of measurements was very larqe and thus made it impossible to determine conclusively whether or not this characteristic is related to protein source. Table 30 shows the main trends of protein source on IPC texture and has been compiled from the statistical analysis. It appears that DENFDM has properties which are separate from those associated with caseinates. The product decreases hardness and cohesiveness, and increases the fracturablllty• It is of interest that sodium caseinate has a pronounced effect on hardness# possibly associated with stronq qellinq tendencies. This effect may be compensated for by the use of DENFDM. The textural properties of natural cheese has been Table 30 Ha1n trends of protein source on IPC texture Fracturablllty Hardness Cohesiveness Adhesiveness E la stic ity Gumminess Chewiness DENFDM Til -ft- Calcium caseinate T .IT. sodium caseinate t T T 163 shown to be c lo s e ly c o rre la te d w ith the chemical com position of the cheese in the solid state (Chen et al., 1979). * Therefore, it is not surprisinq to note the same effect with th e IPC analoqs Which were form ulated by DENFDM and o th e rs. At this time, it is not known whether the textural chanqes are caused by the processinq conditions, subtle chanqes in the manner in Which protein interact with other components or a combination of the two. * 5. Microstructure of IPC Analoqs The morpholoqy of IPC cheese analoqs prepared from different protein sources, includinq DENFDM, was examined and compared to that of commercial processed cheese as a control. In this study, it was desired to evaluate how the microstructure of an IPC analoq prepared from DENFDM only or combination of DENFDM and others would vary by both polarized liqht microscopy (LM) and scanninq electron microscopy (5EM). Each characteristics of IPC analoqs were depicted by both LM and SEM as follows: a) Commercial IPC: The ty p ic a l liq h t m icroscopic structure of the qel matrix in the commercial IPC purchased from a lo c a l superm arket is shown in P late IX. The product shows uniform d is tr ib u tio n o f f a t q lobules embedded in a cell-like matrix of small swollen protein particles. The field is uniform, except for some differences in the Plate IX. Microstructure o f commercial IPC A. photomicrograph by LM (xl28) B. photomicrograph by SEM (xSOO) 165 intensity of staining and the presence of an air hole Which is an artefact due to specimen sectioning. More details were demonstrated by SEM for the nature of the backbone of the matrix (Plate IX-B). It is composed of a protein network in Which fat qlobules are uniformly dispersed* The fat qlobules do not appear to aqqreqate into larger q lo b u les * b) Calcium caseinate IPC analoq: Plate X-A shows that the calcium caseinate IPC analoq is associated with a protein matrixcomposed of a qelled mass and possibly swollen particles. The fat qlobules are nonuniform in size and poorly dispersed in the protein network. The emulsification was not sufficient to create the same uniform appearance asfor the commercial sample. The SEM microphotoqraph (Plate X-B) shows aqain evidence of a qelled matrix of distorted protein. The main difference between this sample and the control is in the degree and effectiveness of blendinq and emulsification with the experimental sample beinq poor in comparison. c) DENFDM IPC analog: The DENFDM analoq showed by LM and SEM examination that the product was poorly blended with discrete larqe particles of DENFDM that did not solubilize. The fat is not emulsified but occupies the spaces between the particles and has seeped into the crevices of the particles (Plate XI-A). The SEM microphotoqraph shows an 166 Plate X. Microstructure of calcium caseinate IPC analog 4 A. photomicrograph by LM (xl28) B. photomicrograph by SEM (xlOO) f ''V . ■ " V Microstructure of DENFDM IPC analog A. photomicrograph by LM (xI28) B. photomicrograph by SEM (x500) 168 open, fibrous, and porous matrix presumably correspondinq to the uniqe surface characteristic of aqqlomerated particles of DENFDM (Plate XI-B). d) Effect of addition of DENFDM on microstructure of IPC analoq: Combination IPC analoq eontaininq calcium caseinate and DENFDM (1:1) demonstrated different trends as seen in Plate XII in terms of emulsification and qel structure. In a similar fashion, Plate XII-B exhibited a intermediate surface structure between Plates X-B and XI-B. The matrix strands appeared to be interwoven in the qel network. Some protein network was fractured and was ' stretched nonuniformly (Plate XII-B). An examination of an other combination of calcium caseinate/DENFDM (?:1) revealed that as the fraction of DENFDM was decreased, the protein network (P la te XIII-A) was much more uniform and the fa t qlobules were also better dispersed compared to Plates XI-A and XII-A. On the other hand, the qel matrix at hiqh maqnification (x 500) was shown to be uniform and the fat qlobules were also quite well dispersed, however, the fat was no t com pletely em ulsified (P la te X III-B ). The presence of "pine-needles" in this structure is noticed, but no explanation can be qiven for their occurrence. e) Effect of addition of sodium caseinate on microstructue of IPC analoq: The effect of sodium caseinate was also investiqated for morpholoqical characteristics. 169 Plate XII. Microstructure of combination of calcium * caseinate/DENFDM (1:1) IPC analog A. photomicrograph by LM (xl28) B. .photomicrograph by SEM (xSOO) 170 Plate XIII. Microstructure of combination of calcium caseinate/DENFDM (2:1) IPC analog A. photomicrograph by LM (xl28) ______B. photomicrograph bv SEM (x500) m 171 The combination of sodium caseinate and calcium caseinate (1s1) (Plate XIV) revealed a different structure from other cheese analoqs. The qel matrix was severely disrupted (Plate XIV-B). The cross section showed that the protein spheres were distorted and fractured and produced an uneven qel matrix. This product exhibited poor emulsification and coarse structure (Plate XIV-A). Further effects of sodium caseinate are seen in Plate XV, for the combination of calcium caseinate/sodium caseinate/DENFDM (It 1:1). Its s tru c tu re was more compact and denser (P la te XV-B). The proteins were not completely dispersed and hiqhly disrupted by particle fraqments. The numerous holes observed may be from entrapped air and a result of insufficient blendinq conditions. The SEM microphotoqraph illustrated an uneven matrix where proteins were aqqreqated in a non-uniform manner. 172 Plate XIV. Microstructure of combination of calcium caseinate/sodium caseinate (1:1) IPC analog A. photomicrograph by LM (xl28) B. photomicrograph by SEM (xlOO) 173 Plate XV. Microstructure of combination of calcium caseinate/sodium caseinate/DENFDM (1:1:1) IPC analog A. photomicrograph by LM (xl28) B. photomicroaraoh bv SEM (xlOO) DISCUSSION Imitation processed cheese (IPC) manufacture has grown to over 2S0 million pounds annually. In these products NFDM cannot be interchanqed with caseinate because NFDM contains a high content of lactose Which is undesirable in substitute cheese manufacture. Techniques for removal of lactose from NFDM to make it competitive with caseinate would be of particular interest because the average annual caseinate imports have increased more than 40% over a decade and have resulted in the displacement of substantial amounts of traditional dairy p ro d u c ts . The primary purposes of this study have been: a) to explore possible approaches for lactose extraction from NFDM by non-aqueous solvents, b) to investigate the physical and functional properties of DENFDM which has been prepared by solvent extraction and c) to explore the possible use of DENFDM as an inqredient in IPC. In this study, the ma^or findings were as follows: a) Sixty two percent methanol was more effective than other solvent systems, includinq azeotropes, for extraction of 174 175 la c to s e . b) Maximum solubility of cure lactose in 5?% methanol was 2.4% as apposed to 19.9% in w ater. The s o lu b ility value for lactose from NFDM in 62% methanol was 2.2%. c) The powder-to solvent ratio of 5t95 at room temperature extracted 60% of the lactose in NFDM; a second extraction at the ratio of 20:80 reduced the residual lactose to 3.8% d) The yield of delactosed NFDM was 39.9% and was associated with a loss of 10.1% of the milk orotelns. e) The composition of double-extracted DENFDM was as follow s: 76.8% protein# 0.8% f a t, 3.8% la c to se , 4.25% moisture and 9.30% ash. The dominant elements in the ash were calcium, phosphorous and potassium. f) Delactosed NFDM prepared by methanol extraction was difficult to solubilize; the initial solubility was 20% increasinq to about 85% after ten hours. q) The lack of solubility was related to stronq hydrophobic interactions, mediated by calcium bindinq, but also to loss of k-casein (20.5% based on sialic acid) and to some alcohol d e n a tu ra tio n o f whey p ro te in . h) DENFDM showed hiqh solvation and water absorption values indicatinq that the interior of the particles represent reqion of stronq hydrophilic tendencies in which water is readily accepted and immobilized. i) The functional properties of DENFDM in imitation 176 processed cheese were different from calcium caseinate and reflected the poor solubility characteristics resultinq in poor emulsification. j) The incorporation of DENFDM with calcium caseinate in IPC increased fraoturability and decreased hardness and cohesiveness. Smaller effects on increased aumminess and elasticity were also observed. k) Results showed a potential for the use of blends of DENFDM with calcium caseinate and/or sodium caseinate for duplicatinq commercial processed cheese. A. Extraction of Lactose from NFDM by Various Solvent Systems The removal of lactose is larqely influenced by the extraction conditions and the nature of the solvents employed. In a series of experiments, it was observed that not all azeotropes remained as a sinqle phase When lactose was dissolved in the systems. Presumably, the presence of lactose chanqed the polarity of the water phase, makinq it less miscible with the pure solvent. Amonq azeotropes, possible solvents would be 72% n-propanol and 60% iso-propanol, however, the solubility would be substantially smaller than in 62% methanol. Nevertheless, for development of a continuous extraction system based on reflux principles 177 ft; must be anticipated that azeotropes would be preferable for extraction and reqeneration over non-azeotroplc mixtures. In the experiments with sinqle or double batches, 62% methanol was observed to be an acceptable solvent. However i t is not cateq o rized as an azeo tro p lc solvent and would be difficult to use in a continuous process. Contact time was observed to be an important factor for the enhancement of lactose extraction. The differences in solubility with varyinq aqitation time was correlated with the deqree to which solvent was obsorbed by the NFDM particles. Usinq a value of 7.4* for solubility of pure lactose in 67% methanol and a lactose content of NFDM of 52%, the required mixinq ratio of NFDM to solvent was predicted to bet 52 q/100 q required solvent ; » 2167 m l/100 q NFDM 0.024 q/ml The mixinq r a tio corresponds to 4.4% NFDM in the so lv en t aqreeinq with the findinq of Leviton (1949). However, usinq the experimental value for NFDM the mixinq ratio of 4:96 would be better. The experimental findinq of lactose solubility in 62% methanol was in qood aqreement with the r e s u lts (2.4% s o lu b ility ) obtained by Hudson and YanovsXy (1917). Accordinq to Leviton (1949), hiqher concentrations of methanol result in relatively small quantities of lactose 178 belnq extracted. This appears to be due to incomolete absorption of solvent by the powder qrains {Leviton, 1949). Furthermore, part of the lactose miqht be in a crystalline form which will not dissolve in the solvent; also some of the lactose miqht have crystallized before filtration was com plete. Surprisinqly, in this study, ethanol was not an effective solvent (only 13.9* of the lactose was extracted). Leviton (193B) reported that lactose extraction from whey o powder at 30 C in 70.7* ethanol was 95* complete in 75 m in u tes. Reflux extraction of lactose in the Soxhlet system was ineffective for removal of lactose. The reason for the incomplete extraction was a hardeninq of the NFDM into a solid mass which was not penetrates by the solvent condensate. It was noted in this experiment that the particle structures of the DENFDM were damaqed due to the tre a tm e n t. At the same tim e, the c o lo r o f DENFDM was markedly chanqed to black-qray. This approach was abandoned for these reasons. To investiqate the possibility of usinq qravity sedimentation or other separation techniques followlnq treatment with solvent, a number of different methods were examined. The use of qravity sedimentation for lactose extraction is possible, but creates some disadvantaqes in 179 terms of product yield. Although filtration through Whatman No.42 filter paper or through cheese cloth was effective in isolating the powder from the solvent the separation time was unacceptably long. However, for a commercial operation there would be no problem in attaining fast and complete separation by solids-electing separators. In general, during agitation, amorphous lactose is more readily extracted than the crystallized form (Leviton, 1949s Choi et al., 1951: Olano, 1979) and, therefore poses the question Whether or not the longer agitation time may alter the physical state. However, in this study, no problem with crystallization or recry3tallization of lactose was o b served. In general, low extraction temperatures were associated with greater solubility. However, in this study, it was concluded that cold treatment had little effect on lactose extraction over using room temperature. This observation is in contrast to the report by Moye (1972) who stated that solubility would increase at high temperatures. Large-scale extraction using a Hobart mixer and gravity sedimentation for a two-cycle extraction was more effective tha bench-scale extraction. The improvement is likely to be related to the aqitation effect and the double solvent treatment or both. However, larger quantities of protein were lost during this extraction (10* loss). 180 The amount of protein lost was largely Influenced by solvent concentration and by the extraction temperatures. The solubility of protein in 62% methanol may generally be assumed to arise from solvent-solute hydrogen bonding. In addition, the solubility of orotein Is greatly Influenced by the solvent concentration. The more water in the system, the greater the loss. To avoid protein losses, higher concentrations of the non-aqueous portion is required, however this would reduce the efficiency of lactose extraction. For commercial scale operation it would be desirable to use the lowest temperature possible in order to minimize the protein losses without reducing lactose extraction efficiency. B. Solubility Behavior of DENFDM An important functional property of DENFDM is its solubility. The results from this study shows that the milk proteins in DENFDM are virtually insoluble on the initial contact with water but an prolonged exposure to agitation slowly approach the solubility of NFDM. The lack of solubility is in conflict with the findings by Leviton (1949) but in agreement with the findings by Morr and Lin (1970). The discrepancy must be related to differences in experimental technique and implicates the use in this study 181 of too hiqh a temperature durinq extraction and subsequent d ry tn q . Pour maior causes for the solubility behavior are postulated: a) stronq hydrophobic interaction, b) stronq calcium bindinq, c) whey protein denaturation and d) release of k-casein. To understand this complex picture, it is necessary to consider the mechanism and several factors associated with low solubility of DENFDM. Delactosed NFDM shows solubility behavior with chanqes in pH which is typical for proteins in qeneral. All proteins show minimum solubility at the pH for their isoelectric point Where the net charqe is zero. As acid is added to an isoelectric protein disoersion the carboxyl qroups accept hydroqen ions, and as alkali is added the amino qroups become deprotonated, therefore, the protein becomes either positively or neqatively charqed. As the net charqe increases, either in the neqative or positive direction, charqe repulsion becomes the qoverninq factor for protein solubility. In this respect, DENFDM showed a solubility profile which resembled calcium caseinate suqqestinq that interactions with calcium is important. The fact that the calcium content for DENFDM is hiqh (24 mq/q) and almost twice as hiqh as for calcium caseinate (13 ma/q) supports this contention. Furthermore, the solubility was markedly improved when calcium sequestrants were added. 182 However, the solubility behavior is still unusual because it improved upon prolonqed agitation in contact with water. Such behavior suqqests a slow attainment of equilibrium and a possible involvement of a slow conformational chanqe. The solubilizinq effect of urea suqqests that the insolubility problem is associated with hydroqen bonding, both hydrophobic and hydrophilic. Hokes et al. (198*2} showed that treatment of calcium caseinate with alcohol and water induced spontaneous curd formation which they postulated was a result of conformational changes brouqht about by the action of alcohol Which caused hydrophobic groups to be exposed to the exterior and a subsequent hydrophobic aqqreqation occurrinq by the addition of water to the alcohol slurry. There are similarities in this study to this effect, since the NFDM was treated with a methanol-water mixture. Thus, the insolubility may be the result of an initial conformational chanqe leadinq to stronq hydrophobic interactions. These interactions may have been favored by the association with calcium salts because Horne and Parke (1991, 1982) have shown in alcohol stability studies that calcium bindinq to caseins promotes regions of hiqh hydrophobicity. Besides, the lowerinq of the dielectric constant by the presence of alcohol produced a stronger Columbia interactions for the bindinq of ions to 183 the protein. Another consideration for the low solubility relates to the possibility that k-casein may have been extracted or possibly modified by the action of alcohol, k-casein is the principal stabilizing fraction for casein micelles and any chanqes which result in its destruction or removal would be expected to have potent results. The measurement of the k-casein content In DENFDM and in the extract showed that ?0.5% k -c a se in was removed in th e e x tra c tio n , based on sialic acid analysis. Efforts to restore solubility by feedinq back the extracted material supported the concept that removal of surface k-casein may have been a possible factor. Further support for this assertion was established by the improvement in solubility by the addition of k-carraqeenan, which is a polysaccharide, which is reported to mimic the action of k-casein. However, the role of k-casein has not been conclusively demonstrated because the dialyzed extract which retain k-casein did not restore solubility to any significant extent. The discrepancy in the effects of feed back between Whole extract and dialyzed extract suqqested that perhaps lactose or extracted minerals were contributing factor for restoring the solubility because these constituents would have been removed durinq d ia ly s is . Experiments on the effect of sucrose, lactose or sodium 184 chloride addition revealed sliqht improvements Which were a ttrib u te d to a q en eral e f fe c t o f sodium ch lo rid e in Which monovalent ions compete with calcium salts for bindinq with proteins* However, no siqnificant improvement for suqar addition could be demonstrated. Aside from the role of hydrophobic aqqreqation, calcium bindinq and the possible loss of k-casein and other stabilizinq factors in the low/slow solubility of DENFDM it is necessary to consider also alcohol denaturatlon of whey proteins. Solubility studies showed that approximately 50* of the whey proteins were denatured in the process. Solubility 9tudieB (Kosaric and Nq, 19R3) have shown that calcium* co-precipited, whey protein-caseinate has low solubility in comparison with commercial caseinates. Thus, the presence of denatured Whey proteins in DENFDM must be expected to exert a detrimental effect on solubility. The conclusion from the present study with respect to insolubility of DENFDM is the followinqi alcohol extraction of NFDM removes about 20% of the k-casein and possibly other s ta b iliz in q components; i t den atu res some o f th e whey proteins, all of Which leads to diminished solubility. There is a potent effect of the alcohol on the conformation of the' caseins by Which hydrophobic qroups are exposed and Which leads to hydrophobic aq q req atio n . The hydrophobicity is increased by stronq bindinq of calcium. This 185 conformational chanqe is to some extent reversible on prolonged contact with water and is responsible for the slow attainment of equilibrium. In spite of the lack of solubility, the product demonstrated stronq solvation and water absorption characteristics which could be seen as contradictory properties. It is postulated that the protein aqqreqates p resen t an e x te r io r surface Which is hydrophobic and Which is responsible for the poor solubility. However, the interior represents a cavity with stronq hydrophilic tendencies in Which water is readily accepted, and is responsible for the solvation. The inertia of the protein aqgreqates to underqo the return to a different conformational arranqement, more compatible with the water environment, Is likely to have been caused by the calcium interaction (cross-linkaqinq) which has fixed the protein in the hydrophobic orientation. C. Physical/Chemical and Functional Properties of DENFDM Casein in NFDM occurs primarily as the calcium salt and DENFDM would be expected to resemble calcium co-precipitated caseinate, since the product would also contain ail of the Whey proteins. In this study, the physical/functional properties of DENFDM refer to the ability of the product to 186 perform a specific function in imitation cheese products• Delactosed NFDM is intended as a replacement for calcium caseinate and in this resDect the chemical composition of DENFDM approachies that of calcium caseinate with the exception that about 3-4* residual lactose remains along with a qreater proportion of milk salts. In addition, there were unique differences disclosed in the analysis for the mineral composition of the ash. DENFDM presented an intermediate between calcium caseinate and NFDM in the ratio of both calcium-to-sodium and divalent-to-monovalent ions. Therefore, a tendency for a preferential extraction of monovalent cations over divalent species has occurred. However, some of the cations seem likely to be more active than others in their bindinq ability to casein. Hokes (19B2), however, has reported that the quantitative behavior of non-activatinq cations did not significantly influence the functional properties. The content of free amino qroups by formol titration was shown to be q re a te r in calcium c a se in a te than in DENFDM on the basis of sample weiqht and protein content. It was apparent from the data that the solvent treatment had induced a loss of about 35-40% of the free amino qrouos. The action of proteolytic enzymes can be discounted in this system* However, since DENFDM was treated with solvent, it is possible that specific chemical action may have altered 187 the nature of the amino qroups. If this la the case, then large differences in the physical/chemical properties of the affected proteins must be expected, and perhaps offer additional reason for the lack of solubility. Results for the powder particle size distribution obtained by microscopy were disappointinq because the microscopic measurement exhibited some limitations related to the depth of focus. If the microscope is focused on at one level, another level will not be visible. If the microscope is focused on the larqe particles, resolution of small particles is lost. For this reason some Inaccuracies in the analysis must be expected. Particles of DENFDM were biqger than in NFDM as a result of swellinq and aqqreqation. I t was observed th a t the p a r tic le siz e was no t c o rre la te d with the bulk density. The differences in particle characteristics between DENFDM and NFDM were caused by the removal of lactose producinq hiqhly porous particles. The particles of DENFDM were observed to be much more fragile as a result of solvent treatment. Holsinqer (19R0) reported that the particle size distribution is an important factor in the suspension stability of protein products. Measurement of the average equivalent spherical diameter could provide information which could be related to be suspension stability of the reconstituted powder (Holsinqer, 19R0). Lascelles et al. 188 <1976) proposed that larqe particles resulted in better wettability and more rapid dissolution. If bulk density is too qreat, self disoersion may be decreased (Haroer et al., 1963). Therefore, DENFDM was expected to exhibit better dispersion and suspension stability than NFDM? this, however, was not the case and stresses that profound alterations have, affected the protein moieties. In sets of experiment for casein pellet-solvation, the sedimentable fraction of DENFDM exhibited a hiqh deqree of hydration compared to NFDM. This indicates that the product has a stronq affinity for water without beinq soluble. This point has been stressed earlier and may be explained by the water holdinq capacity of interior hydrophilic reqions. The results for water absorption were derived from experiments (Hokes, 19H2) in which starch and protein competed for water durinq starch qelatinization and are likely to qiva different information from the solvation studies. In principle, milk protein competes for the water which is available for starch qelatinization and therefore the maximum viscosity for the sample containinq calcium caseinate and DENFDM were decreased. The amount of protein added was least for NFDM and compensatlnq effects from lactose and other constituents may be responsible for the inertness of this sample. The water absorption data in this experiment were rouqhly correlated with the solvation data, 189 confirminq that Interactions with water was a shared property for calcium caseinate and DENFDM. However, the effects were stronqest for DENFDM and are consistent with the postulate that the water absorbed by the interior of the protein aqqreqates is made unavailable for other constituents. In examininq DENFDM by LM and SEM, the structure of DENFDM exhibited predominantly porous, broken or damaqed particles. In addition, the particle shape were more irreqular, rouqh, coarse and hollow. The removal of lactose by solvent extraction has left void spaces in the particles walls causinq the porosity. Presumably the apparent brittleness and fraqile nature is a result of the removal of lactose as a cementinq material. This structural representation helps explain how water can penetrate the hydrophobic particles and solvate the material. The void spaces, previously occupied by lactose represent channels throuqh Which water can be transported to the interior where it is effective microencapsulated* D. Morpholoqical/Textural Properties of IPC This study has been concerned with the possibility of DENFDM formation into conventional imitation processed cheese* The maior observations made in this study are: 190 a) The e£fect of DENFDM was mainly a contribution to fracturability, elasticity and qutnminess, While hardness and cohesiveness were decreased. In contrast, calcium caseinate an sodium caseinate were contributinq to cohesiveness and hardness, respectively. b) In evaluation of microstructure by LM and SEM, DENFDM produced a fibrous, open, porous as well as bad emulsified protein matrix. c) DENFDM has a potential beneficial effect as a partial replacement of caseinate in the formation of IPC to characteristic close to processed cheese. SUMMARY 1. Sixty two percent methanol was more effective than other solvent systems* includinq azeotropes, for extraction of la c to s e . 2. Maximum solubility of pure lactose in 67% methanol waB 2.411 as apposed to 19.9% in water. The solubility value for lactose from NFDM in 67% methanol was 7.2%. 3. The powder-to solvent ratio of 5:95 at room temperature extracted 60% of the lactose in NFDM; a second extraction at the ratio of 20:80 reduced the residual lactose to 3.8% 4. The y ie ld o f . delacto sed NFDM was 39,9% and was associated with a loss of 10.1% of the milk proteins. 5. The composition of double-extracted DENFDM was as follow s: 76.8% p ro te in , 0.8% f a t, 3.8% la c to se , 4.75% m oisture and 9.30% ash. The dominant elem ents in the ash were calcium, phosphorous and potassium. 6. Delactosed NFDM prepared by methanol extraction was difficult to solubilize; the initial solubility was 70% increasinq to about 85% after ten hours. 7. The lack o f s o lu b ility was re la te d to stronq hydrophobic interactions, mediated by calcium bindinq, but also to loss 191 192 of k-casein (20.5^ based on sialic acid) and to some alcohol d e n a tu ra tio n o f whey p ro te in . 8. DENFDM showed hiqh solvation and water absorption values indicatinq that the interior of the particles represent reqion of stronq hydrophilic tendencies in Which water is readily accepted and immobilized. 9. The functional properties of DENFDM in imitation processed cheese were different from calcium caseinate and reflected the poor solubility characteristics resultlnq in poor emulsification. 10. The incorporation of DENFDM with calcium caseinate in IPC increased fracturability and decreased hardness and cohesiveness. Smaller effects on increased qumminess and elasticity were also observed. 11. Results showed a potential for the use of blends of DENFDM with calcium caseinate and/or sodium caseinate for dupiicatinq commercial processed cheese. REFERENCES Akobundu, E.N.T., Cherry, J.P. and Simmons, J.G. 1982. Chemical, functional and nutritional properties of Equsl seed protein products. J. Food Sci. 47t829 AOAC 1970. Official Methods of Analysis, 11th ed. Association o £ o££icial Analytical Chemists, W ashington, D.C. Aorahamian, V.B. 1973. Liberation of sialic acid from casein hydrolysed in formic ac id . M.S. th e s is . The Ohio S ta te University, Columbus, OH Arnott, D.R., Morris, H.A. and Combs, W.B. 1957. Effect of certain chemical factors on the meltinq quality of process cheese. J. Dairy Sci. 40t957 Bezkorovainy, A. 1970. Basic protein chemistry, op.13-14. Charles C, Thomas Publisher, Sprinqfield, IL, U.S.A. Blaedel, W.J. and Meloche, V.W. 1963. Elementary quantitative analysis. Theory and practice . 2nd ed. pp.171-189. Harper and Row Publishers, Inc. NY Boyer, R.A. 1975(a). Method of makinq protein food products resembllnq cheese. U.S. oatent 3,891,777 Boyer, R.A. 1975(b). Method of making protein food products resemblinq cheese. U.S. patent 3,891,778 Buma, T.J. and Henstra, S. 1971. Particle structure of spray dried spray-dried caseinate and spray-dried lactose as observed by a scanninq electron microscope. Neth. Milk Dairy J. 25*278 Catsimpoolas, N. and Meyer, B.W. 1970. Gelation phenomena of soybean qlobulins. I. Protein-protein interactions. C ereal Chem. 47:559 193 194 Chakraborty, B.K. 1981. Functional properties of caseinates. Presented at the National Workship on Dairy Analoqs and Fabricated Foods at National Dairy Research Institute, Karnal, India Choi, R.P., Tatter, C.W. and O'malley, C.M. 1951. Lactose crystallization in dry products of milk. II. The effects of moisture and alcohol. J. Dairy Sci. 34:850 Cooney, C.M. and Morr, C.V. 1977. Hydroqen peroxide alteration of Whey proteins in whey concentrated Whey* systems. J. Dairy Sci. 55:567 Covacevich, H.R. and Kosikowski, F.V. 1978. M ozzarellar and Cheddar cheese manufacture by ultrafiltration principles. J. Dairy Sci. 61:701 Cunninqham, S.D., Carter, C.M. and Mattier, K.F. 1975. Rupture and protein extraction of petroleum-qrown y e a s t. J . Food S ci. 40:732 Dack, M.R.J. 1971. Solvent effects and reactivity. Chem. Tech. Feb. p .108 Domovs, K.B. and Freund, E.H. 1960. Methanol soluble complexes of lactose and of other carbohydrates. J. Dairy Sci. 43:1216 Dubrow, D.L., Kramer, A. and Mcphee, A.D. 1973. E ffe c t o f temperature on lipid extraction and functional properties of fish protein concentrates (FPC). J. Food Sci. 38:1012 Elqedaily, A., Campbell, A.M. and Penfieid, M.P. 198?. Solubility and water absorption of systems containinq soy protein isolates, salt and suqar. J. Food Sci. 47:806 Emmons, D.B. 1981. Utilization of dairy proteins in utilization of protein resources by D.W. Stanley, Murray, E.D. and Lees, D.H. e d . d .? 47. Food and Nutrition Press (Inc.), U.S.A. Fey, M.W., Weil, C.M. and Sequr, J.B. 1951. Solubility of sucrose in aqueous qlycerol and propylene qlycol. Ind. Enq. Chem. 43:1435 195 Fox, K.K. 1970. Byproducts from milks Casein and whey.Webb, B.H. and W hitter, E.O. e d . 2nd ed ., AVI ou b lish in q Co. Westport, CT Freck, J.A. and Kondrot, L.V. 1974. Process for makinq simulated meat and cheese products* U.S. oatent 3,826,677 Fukushima, D. 1969. Denaturation of soybean proteins by orqanic solvents. Cereal Chem. 46:156 Girdhar, B.K. and Hansen, P.M.T. 1974. Soluble casein by adsorption of ammonia. J. Food Sci* 39:1237 Haqerdal, B. 1978. Stability of myoqlobin to acetone treatment characterized by differential scannlnq colorimetry and resolubility in water. J. Food sci. 43:22 Hansen, P.M.T., Hokes, J.C. and Manqino, M.E. 1981. Imitation processed cheese and cheese structure. UNDP/ICAR Nat. Workshop, Dairy Analoqs and Fabricated Foods for the Future, Nat. Dairy Res. Inst., Karnal, India, Nov. 18-19 Harper, M.K., Holsinqer, V., Fox, K.K. and Pallansch, M.J. 1963. Factors infiuencinq the instant solubility of milk powders. J. Dairy Sci. 46:1192 Hayashi, H., Heldman, D.R. and Hedrick, T.I. 1969. Influence of spray-dryinq conditions on size and size distribution of nonfat dry milk particles. J. Dairy Sci. 52(1):31 Herrinqton, B.L. 1934. The solubility of lactose in salt s o lu tio n s : the is o la tio n o f a compound o f la c to se and calcium chloride. J. Dairy Sci. 17:805 Herskovits, T.T., Gadeqbeku, B. and .Taillet, H. 1970. On the structural stability and solvent denaturation of proteins. I. Denaturation by the alcohols and qlycols. J. Biol. Chem. 245:2588 Hokes, J.C. 1982. An analysis of the functional properties of calcium caseinate as related to imitation processed cheese. Ph.D. dissertation, The Ohio State University, Columbus, OH 196 Hokes, J.C., Manqino, M.E. and Hansen, P.M.T. 1915?. A model system for curd formation and meitinq properties of calcium caseinates. J. Food Sci. 47(4):1235 Holsinger, V.H. 1980. A study of the rehydration properties of a milk analogue containing soy products and cheese Whey. Ph.D. dissertation. The Ohio State University, Columbus, OH Horne, D.S. and Parker, T.O. 1981. Factors affectinq the ethanol stability of bovine milk. J. Dairy Res. 48:405 Horne, D.S. and Parker, T.G. 1982. Some aspects of the ethanol stability of caprine milk. J. Dairy Res. 49:459 Howard, H.W., Block, R.J. and Sevall, H.E. 1954. Milk protein food product and process. U.S. patent 2,665,989 Hudson, C.S. 1908. Further studies on the forms of milk suqar. J . Amer. Chem. Soc. 30:1767 Hudson, C.S. and Yanovsky, E. 1917. Indirect measurements of the rotatory powers of some alpha and beta forms of the suqars by means of solubility of experiments. J. Amer. Chem. Soc. 39:1013 Janzen, J.J., Swanson, A.M. and Mclntire, J.M. 1953. Particle size of dried milk products. I. A statistical approach to the measurement of particle size. J. Dairy Sci. 36:905 Jolly, R.C. and Kosikowski, F.V. 1975. A blue cheese food material from ultrafiltrated skim milk and microbial enzyme-mold spore reaction. J. Dairy Sci. 58:1271 Jonas, J.J. 1973. Utilization of dairy inqredients in other foods. J . Milk Food Technol. 36(6):328 Kautter, D.A., Lynt, R.K., Lilly Jr.T. and Solomon, H.M. 1981. Evaluation of botulism hazard from imitation cheeses. J. Food Sci. 46:749 K&uzmann, W. 1959. Some factors in the interpretation of protein denaturation. in Adv. Protein Chem. vol.14. pp.1-63. Academic Press, NY and London. 197 Kehaqias, C. and Hansen, P.M.T. 1976. Isolation of slalio acid from casein and reclamation of residual casein. J. Dairy Sci. 59:1718 Kinq, N. 1965. The physical structure of dried milk. J. Dairy Sci. Abstr. 27(3):91 Kononenko, O.K. and Herstein, K.M. 1956. Nonaqueous solvents for sucrose: Part II. Physical properties evaluation of compounds and materials. Chem. Enq. Data S er. 1:87 K osaric, N. and Nq, D.C.M. 1983. Some fu n ctio n a l properties of milk protein calcium co-precipitates. Can I n s t. Food S c i. Technol. J . 16(2):141 Kumar, V. and Kosikowski, F.V. 1977. Processed cheese manufacture from ultrafiltrated retentates with and without enzymes. J. dairy Sci. 60.suopl. 1:40 ( a b s t r . ) Lanqe, N.A. 1956. Handbook of Chemistry, p .1221. 9th ed. Handbook Publishers, Inc. Sandusky, OH Lascelles, D.R. and Baldwin, A.J. 1976. Dispersibility of whole milk powder in warm water. N.Z.J. Dairy Sci. Technol. 11:2R3 L a z a rid is, H.N. and Rosenan, J.R . 1980. E ffe c ts o f emulsifyinq salts and carraqeenan on rheoloqical properties of cheese-like products prepared by direct acidification. J. Food Sci. 45:595 Leviton, A. and Leiqhton, A. 1938. Separation of lactose and soluble proteins of Whey by alcohol extraction. Ind. Enq. Chem. 30:1305 Leviton, A. 1949. Methanol extraction of lactose and soluble proteins from skim milk powder. Ind. Enq. Chem. 41:1351 Lim, S.G. and Nickerson, T.A. 1973. Effect of methanol on the various forms of lactose. J. Dairy Sci. 56:843 Lu, P.S. and Kinsella, J.E. 1972. Extractability and properties of protein from alfalfa leaf meal. J. Food S ci. 37:94 Majd, F. and Nickerson, T.A. 1976. Effect of alcohols on lactose solubility. J. Dairy Sci. 59:1025 198 Marier, J.R. and Boulet, M. 1999. Direct analysis of lactose in milk and serum. J. Dairy Sci. 42:1390 Marier, J.R., Tessier, H. and Rose, D. 1963. Sialic acid as an index of the k-casein content of bovine skim milk. J. Dairy Sci. 46:373 Mckenzie, H.A. 1970. Effects of chanqes in environmental conditions on the state of association, conformation, and structure, in Milk Protein Chem. and Molecular Biol* vol.l pp.355-477. Academic Press, NY and London Merten, R. and.Bunqe, W. 1964. Practically water-free solutions of carbohydrates in polyalcohols. Chem. Abstr. 61, 12184a Milner, M., Scrimshaw, N.S. and Wanq, D.I.C. 1978. Protein resources and technoloqy. AVI oublishinq Co. U.S.A. Moye, C.J. and Smythe, B.M. 1965. Non-aqueous suqar solvents: A New Series. Carbohyd. Res. 1:284 Moye, C.J. 1972. Non-aqueous solvents for carbohydrates, in Adv. Carbohyd. Chem. Biochem. Tioson, R.S. and Horton, D. ed. p. 85. Academic P ress. NY and London Morr, C.V. and Lin, S.H.C. 1970. Preparations and properties of an alcohol-precipitated whey protein concentrate. J. Dairy Sci. 53:1162 Morr, C.V. 1974. Methods of qel electrophoresis of milk proteins, p .12. Monoqraph published by the American Dairy Science Association Committee on Nomenclature and Methodoloqy of milk proteins. Morr, C.V. 1979. Utilization of milk proteins as startinq materials for other food stuffs. J. Dairy Res. 46:369 Morr, C.V. 1982. Functional properties of milk proteins and their use as food inqredients. in Devel. Dairy Chem. I. Proteins Fox, P.F. ed. pp.375-399. Applied Science Publishers, London and NY Muller, L.L., Hayes, J.F. and Snow, N. 1967. Studies on co precipitates on milk proteins. Aust. J. Dairy Tech. 22:12 Muller, L.L. 1971. Manufacture and uses of casein and co precipitate. Dairy Sci. abstr. 33:659 199 Nickerson, T.A. and Lim, S.G. 1974. Effect of various alcohols on lactose. J. Dairy Sci. 57:1320 Olano, A. and Rios, J.J. 1970. Treatment of lactose with alkaline methanolic solutions: production of beta- lactose hydrate. J. Dairy Sci. 61:300 Olano, A. 1978. Treatment of forms of lactose with dilute alcoholic solutions of sodium hydroxide. J. Dairy Sci. 61:1622 O'meara, G.M. and Munso, P.A. 1982. The p re c ip ita tio n and shrinkaqe of acid casein curd: A preliminary study. N.Z.J. Dairy Sci. Technol. 17:147 Ott, L. 1977. An introduction to statistical methods and data analysis, pp.662-664. Duxbury Press, MA Petka, T.E. 1976. Novel caseinate: effects natural functional properties for imitation cheese. Pood Prod. Dev. 10(10):26 Puri, B.R., Murari, K. and Totela, K.K. 1965. Studies in stability of protein dispersion in milk. I. ethanol stability. Indian J. Dairy Sci. 18:112 Quickert, S.C. and Bernhard, R.A. 1982. Recovery of lactose from aqueous solution usinq qroup IIA metal chlorides and sodium hydroxide. J . Food S ci. 47:1705 Rayan, A.A., Kalab, M. and Ernstrom, C.A. 1980. Microstructure and rheoloqy of process cheese. SEM. 3:635 Ross, K.D. 1978. Effects of methanol on physical properties of a- and 0-lactose. J. Dairy Sci. 61:152 Seidell, A. 1919. Solubilities of inorqanic and orqanic compounds, p .699, D. Van Nostrand Co. Inc. NY Sharp, P.F. 1943. Stable crystalline anhydrous alpha lactose product and process. U.S. patent 2,319,562 Shehata, A.E., Meena, I., Olson, N.F. and Richardson, T. 1967. Effect of type of acid used in direct acidification procedures on moisture, firmness and calcium levels of cheese. J. Dairy Sci. 50:824 0 200 Singer, S.J. 1965. The properties of protein in non-aqueous solvents. Adv. Protein Chem. vol.17. pp.1-68. Academic P ress, NY and London Sood, V.K. and Kosikowski, F.V. 1979. Process cheddar cheese from plain and enzyme treated retentates. J. Dairy Sci. 62*1713 Sosulski, F.W., Humbert, G.S., Bui, K. and Jones, J.D. 1976. Functional properties of rapeseed flours: concentrates and isolates. J. Food.Sci. 41*1349 Southward, C.R. and Goldman, A. 1975. Co-precipitates - A review. N.Z.J. Dairy Sci. Technol. 10:101 Stockham, J.D. and Fochtman, E.G. 1978. Particle size analysis. Ann Arbor Science Publishers, Inc. Ann Arbor, MI Tanford, C. 1969. Protein denaturation. in Adv. Protein Chem. 23*121. Academic P ress, NY and London Thomas, M.A. 1970. Use of calcium co-precipitates in processed cheese. Ast. J. Dairy Technol. 75*1001 Thomas, M.A., Newell, G., Abad, G.A. and Turner, A.D. 1980. Effect of emulsifyinq salts on obiective and sublective properties of processed cheese. J. Food Sci. 45:458 Thompson, M.P., Boswell, R.T., Martin, V., Jenness, R. and Kiddy, C.A. 1968. Casein-pellet-solvation and heat stability of individual cow's milk. J. Dairy Sci. 52(6)*796 Trey, H. 1895. Experimental beitraq sur birotation der qlykose. Z. Phys. Chem. 18*193 Warren, L. 1959. The thiobarbituric assay of sialic acids. J. Biol. Chem. 234*1971 Weber, R.E. and Tanford, C. 1959. The confiquration of ribonuclease at low p H in 2-chloroethanol and in 7- ch lo ro eth an o l-w ater m ix tu res. J . Amer. Chem. Soc. 81*3255 Wynn, J.D. 1978(a). Process for makinq a food product resemblinq cheese. Brit, patent 1,517,956 Wynn, J.D. 1978(b)• Simulated cheese preparation usinq calcium caseinate. N.Z. patent 178,491 201 Wynn, J.D. 1978(c). Preparation of imitation cheese. U.S. patent 4,104,413 Yanq, J.T. and Doty, P. 1957. The optical rotatory dispersion of polypeptides and proteins in relation to c o n fiq u ra tlo n . J . Amer. Chem. Soc. 79*761 Zinqaro, R.A. 1968. Nonaqueous s o lv e n ts, in Topics in Modern Chem istry, p . 5. D.C. Health and Company, MA APPENDIX A Table 31 Particle size distribution by lig h t microscopy For Ca-caselnate S1ze(x20) d n nd nd2 nd3 nd1*