UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Manufacture of vitamin B12 from sulfite spent liquor. Ferguson, David Kimball 1972

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1973_A7 F47.pdf [ 9.27MB ]
Metadata
JSON: 831-1.0059174.json
JSON-LD: 831-1.0059174-ld.json
RDF/XML (Pretty): 831-1.0059174-rdf.xml
RDF/JSON: 831-1.0059174-rdf.json
Turtle: 831-1.0059174-turtle.txt
N-Triples: 831-1.0059174-rdf-ntriples.txt
Original Record: 831-1.0059174-source.json
Full Text
831-1.0059174-fulltext.txt
Citation
831-1.0059174.ris

Full Text

MANUFACTURE OF VITAMIN B FROM 1 2 SULFITE SPENT LIQUOR by D A V I D K I M B A L L F E R G U S O N B.E., Nova S c o t i a T e c h n i c a l C o l l e g e , 1 9 6 U A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A P P L I E D S C I E N C E i n t h e d e p a r t m e n t o f C H E M I C A L E N G I N E E R I N G We a c c e p t t h i s t l i e s i s as c o n f o r m i n g t o t h e r e q u i r e d s t a n d a r d T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A September, 1 9 7 2 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make i t freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of C#&<+,CJc i*yfffi?^^ The University of British Columbia Vancouver 8, Canada Date '*/9nt-4. " * 3 O N E T O N O F P A P E R U S E S S E V E N T E E N T R E E S . P L E A S E R E C Y C L E T H I S T H E S I S . ## ## ## ## ## ######## ABSTRACT Up to 2 mg/l of vitamin B 1 2 were produced batch-wise using Propionibacterium freudenreichii on an ammonia based spent s u l f i t e liquor medium in 250 ml Erlenmeyer flasks and a 7 - l i t r e benchtop fermentor. The up to 6 gm/1 of bacteria u t i l i z e d the hexoses from the SSL, thus reducing the BOD by 50-80% and producing, as by-products, 3 gm/1 of acetic acid and 7 gm/1 of propionic acid. Pre-treatment of the liquor required stripping of S0 2 to below 200 ppm and adjustment of the pH to between 6.5 and 7.5; p r e c i p i t a t i o n of 1igno-su1fonate was not necessary. It was necessary to add excessive amounts (up to 75 gm/1 dry) of yeast extract or other supplementary nutrient to achieve these vitamin B i 2 y i e l d s . Solutions to this nutrient problem are suggested. The Monod fermentation model whose parameters were estimated using non-linear least squares techniques with dry bacterial c e l l concentration as the dependent variable, approximate the batch data well. Other less complicated i i models were not as s a t i s f a c t o r y . Estimating the model parameters using linear techniques was most unsatisfactory indeed. Optimum hold-up times for a single stage fermentor for Propionibacteria production, predicted from the Monod model whose parameters were estimated from batch data, were up to 125 hours. These hold-up time extrapolations are subject to large errors. Recommendations for further work on the vitamin B 1 2 process, the extension of the batch modeling work to continuous fermentation and further work on other more suitable microbiological products made from SSL are enter-tained. i i i TABLE OF CONTENTS Page ABSTRACT i i LIST OF TABLES ix LIST OF FIGURES xi i i ACKNOWLEDGEMENTS xvi Chapter 1 INTRODUCTION 1 1.1 Spent S u l f i t e Liquor 1 1.1.1 S u l f i t e pulping process and pollution 1 1.1.2 Proposed uses for s u l f i t e spent liquor 7 1.2 Vitamin B 1 2 and Propionibacteria Freudenrei chi i 11 1.3 Biochemistry of Propi oni bacteri a 14 1.4 Review of the Work of Nishakawa 18 1.5 Objectives of this Research 20 2 METHODS AND APPARATUS 22 2.1 Spent S u l f i t e Liquor Preparation 22 i v Chapter Page 2.2 Nutrient Additions to the Spent Liquor 22 2.3 Inoculum Preparation 23 2.4 Equipment and Apparatus 24 2.5 Measurement of Bacterial Growth 25 2.6 Measurement of Vitamin B 1 2 Concentration 29 2.7 Calculation of Biochemical Oxygen Demand and Chemical Oxygen Demand 31 2.8 Measurement of Reducing Sugar Concentration 36 2.9 Measurement of Acetic and Propionic Acids 36 3 RESULTS 38 3.1 I n i t i a l Yield Studies 38 3.2 Inhibitory Effect of Sulfur Dioxide on Cell Yield . 42 3.3 U t i l i z a t i o n of Spent S u l f i t e Liquor Sugars 47 3.4 7-Litre Fermentor Studies 50 3.5 Search for. a Replacement for Yeast Extract 63 3.6 Inhibitory Effect of Propionate and Hydrogen Ions . 6 7 4 YIELD COEFFICIENTS AND COST ANALYSIS OF NUTRIENTS AND PRODUCTS 70 4.1 Y x ^ s - Y i e l d of Dry Cells Based Upon Sugar Used 73 v Chapter Page 4.2 Y x / n - Y i e 1 d of Dry Cells Based Upon Yeast Extract 81 4.3 Y a^ x Determination of Acids Yield Coef f i ci ent 87 4.4 Y v / x - Y i e l d of Vitamin B 1 2 89 4.5 Nutrient Requirements and Costs for Growth of P. f reudenrei chi i 94 4.6 Comparison of the Yields of this Work with the Yields of Others 100 5 KINETIC MODEL FOR PROPIONIBACTERIA GROWTH. . . . 104 5.1 Development of Models 104 5.1.1 Zero order model 104 5.1.2 Autocatalytic or Malthus model . . . 107 5.1.3 F i r s t order substrate l i m i t i n g model 108 5.1.4 Second order autocatalytic-substrate l i m i t i n g model 110 5.1.5 Edwards and Wilke model 112 5.1.6 Monod model 113 5.2 Estimating the Parameters of the Models . . 117 5.2.1 Estimating the parameters of unintegrated rate models 118 5.2.2 Integrated linear models with time as the dependent variable 120 5.2.3 Estimation of the parameters of the non-linear models 121 v i Chapter Page 5.3 Testing the Models 124 5.3.1 Rate models 128 5.3.2 Integrated models with time as dependent variable 137 5.3.3 Non-linear models 139 5.3.4 Tentative conclusions of kinetic model parameter estimation 142 5.4 Use of these Estimated Model Parameters to Design a Continuous Fermentor 144 5.5 Results of Kinetic Model Parameter Estimation of the Other Run 148 5.6 Comparison of the Estimated Model Parameters Between Runs 159 5.6.1 Comparison between runs 159 5.6.2 Optimum hold-up time for the other runs 161 5.7 Comparison of the Results of this Work with Kinetic Work of Others 161 6 CONCLUSIONS 166 7 RECOMMENDATIONS FOR FURTHER WORK 169 7.1 SSL-Propionibacteria - Vitamin B a 2 Process 169 7.2 Kinetic Modeling and Continuous Fermentation 170 7.3 U t i l i z a t i o n of SSL to Produce a Commercial Microbiological Product 170 REFERENCES 172 NOMENCLATURE 181 v i i APPENDICES Page I CORRELATION OF DRY CELL WEIGHT AND TURBIDITY 185 II VITAMIN B 1 2 TEST. 195 III REQUIREMENTS FOR A GOOD MICROBIOLOGICAL PRODUCT TO BE MADE FROM SSL 208 IV MICROBIOLOGICAL PRODUCTS WHICH COULD BE PRODUCED FROM SPENT SULFITE LIQUOR . . . . 211 vi i i LIST OF TABLES Table Page 1. Compositions of Typical S u l f i t e Spent Liquor 5 2. Sugars in SSL From Various Sources 6 3. Proposed Uses for S u l f i t e Spent Liquor 8 4. B.O.D. Removal from Spruce Waste Liquor 32 5. Reduction of UOD for Removal of 1 gm/1 of SSL Component 35 6. I n i t i a l Test Tube Tests to Determine Yield Based on Yeast Extract 40 7. Preliminary Yeast Extract Yield Tests in 250 ml Erlenmeyer Flasks 41 8. Yiel d on Yeast Extract Using Columbia Cellulose SSL and Powder Yeast Extract 43 9. Yield on Yeast Extract Using Rayonier SSL and Powder Yeast Extract 44 10. Yield on Yeast Extract Using Rayonier SSL and Paste Yeast Extract 45 11. Test for Yield Inhibition of S u l f i t e in Fermentation Broth 46 ix Table Page 12. Growth of P. Freudenreichii on Various Wood Sugars 48 13. 7-Litre Fermentation on C l o s t r i d i a l Medium and Glucose 52 14. 7 - l i t r e Fermentation - SSL Run No. 1 54 15. 7 - l i t r e Fermentation - SSL Run No. 2 56 16. 7 - l i t r e Fermentation - SSL Run No. 3 57 17. 7 - l i t r e Fermentation - SSL Run No. 4 58 18. Growth T r i a l s Using Various Nutrients 66 19. Ef f e c t of pH Inhibition on P. Freudenreichii Growth 6 9 20. Cell Weight vs. Sugar - Data from A l l Runs . 7 4 21. Yield Coefficient Based on Sugar Usage 77 22. Dry Cell Weight vs. Yeast Extract Charged for Yeast Extract Limiting Runs 82 23. S t a t i s t i c a l Analysis of Y ^ n for Powder and Paste Yeast Extract 86 24. Yield of V o l a t i l e Acids Based on Dry Cell Weight. . 88 25. Acid Yield Coefficients 90 x Table Page 26. Vitamin B 1 2 Analyses 92 27. Vitamin B 1 2 Yield Coefficient 95 28. Raw Material Requirements and Product Yield for Vitamin B 1 2 Production 96 29. Value of Nutrients and Products in Fermentation Broth 97 30. Improvement in Net Broth Value for Some Process Changes 99 31. Comparison of Yields of this Work with that of Others 101 32. Summary of the Models to be Discussed in this Section 125 33. Data of SSL Run No. 2 for Time vs. C A Model Parameter Estimates 127 34. Calculated Data from Linear Interpolation for Rate Models 129 35. Calculated Data from Quadratic Interpola-tion for Rate Models 130 36. Estimated Parameters of the Models 131 37. Predicted Optimum Dilution Rates from Model Parameters 149 38. Estimated Values of the Parameters for a l l the 7 - l i t r e Fermentor Runs 151 xi Table Page 39. Comparison of Measured and Estimated C° Values 158 40. Predicted Optimum Hold-up Times for A l l the Runs 162 41. Determination of Dry Cell Weight vs. Turbidity Relationship 189 42. Vitamin B.i2 Test - Dilution Factors for Unknowns 198 43. Vitamin B i 2 Test - Absorbance Readings for Standard Curve 200 44. Vitamin B i 2 Test - Unknown Sample Readings 201 45. Results of Vitamin B i 2 Test Using Quadratic Model 207 xi i LIST OF FIGURES Figure Page 1. Molecular formulas of wood and SSL chemicals 3 2. Effect of vitamin B i 2 on chick growth 13 3. Vitamin B 1 2 - Cobalamin 17 4. Diagram of 7 - l i t r e NBS fermentor 26 5. Effect of S0 2 on y i e l d of Propionibacteria . . . 49 6. 7 - l i t r e fermentor run with C l o s t r i d i a l Medi um - gl ucose 53 7. 7 - l i t r e fermentor SSL Run No. 1 59 8. 7 - l i t r e fermentor SSL Run No. 2 60 9. 7 - l i t r e fermentor SSL Run No. 3 61 10. 7 - l i t r e fermentor SSL Run No. 4 62 11. Reducing sugar remaining vs. dry c e l l weight for a l l runs 78 12. Sugar used vs. change in dry c e l l weight for a l l runs combined 80 x i i i Fi gure Page 13. Dry c e l l weight vs. i n i t i a l yeast extract for a l l yeast extract l i m i t i n g runs 85 14. Acids produced vs. dry c e l l weight for al 1 runs combi ned 91 15. Vitamin B i 2 produced vs. time for a l l runs . . . 93 16. Sketches of the shapes of the various models tested 106 17. Rate models - C x vs. time showing results of parameter estimation 132 18. Rate models - rate vs. time plot showing poor quality of rate data 133 19. Li neweaver-Bu.rk plot for rate models 135 20. Expanded Lineweaver-Burk plot of Figure 19 . . . 136 21. C vs. time plots for integrated models with time as dependent variable 138 22. C vs. time plots for non-linear models with C as dependent variable 140 A 23. Residual plot for non-linear models 141 24. C vs. time plots for non-linear models -x SSL Run No. 1 152 25. C vs. time plots for non-linear models -x SSL Run No. 3 153 26 C vs. time plots for non-linear models -x SSL Run No. 4 154 xi v Fi gure Page 27. Residual plots for SSL Run No. 1 -non-linear models 155 28. Residual plots for SSL Run No. 3 -non-linear models 156 29. Residual plots for SSL Run No. 4 -non-linear models 157 30. TURB x D vs. TURB plot for dry c e l l weight test correlation 191 31. LOG(cf) vs. TURB for dry c e l l weight correlation 192 32. Dry c e l l weight vs. TURB x EXP(0.6078 x TURB) to check out or i g i n a l data points 194 33. Vitamin B.12 test standard curve -linear model 203 34. Vitamin B 1 2 test standard curve -quadratic model 205 xv ACKNOWLEDGEMENTS I wish to thank Dr. Richard Branion and Dr. Ken Pinder, under whose direction this work was undertaken, for their patience and guidence throughout this work. I also wish to thank Dr. George Strasdine of the Fisheries Research Board for his help in microbiological techni ques. Furthermore, I wish to thank Ms. Shari Haller who typed the thesis from a rather well worked over draft. F i n a l l y I wish to thank the Department of Energy, Mines and Resources — Water Resources Division, the National Research Council and the University of B r i t i s h Columbia for their f i n a n c i a l assistance in this work. xvi Chapter 1 INTRODUCTION 1.1 Spent S u l f i t e Liquor 1.1.1 S u l f i t e pulping process and po l l u t i o n . Since the discovery of the s u l f i t e pulping process in 1866 by B.C. Tilghman [1], much has been written about the water pollution caused by the wastes from s u l f i t e pulp m i l l s . Often these writings have been highly charged with emotion. Environmentalists, on the one hand, have charged that pulp m i l l wastes are poisonous to aquatic l i f e in con-centrations as low as 8 ppm [2] and that pulp m i l l s have been responsible for untold damage to the environment. On the other hand, apologists for the pulp and paper industry [3] have claimed that s u l f i t e pulp m i l l wastes cause only a s l i g h t lowering of dissolved oxygen and may, in f a c t , even be bene f i c i a l in f e r t i l i z i n g the growth of microorganisms at the lower end of the food chain and thereby eventually increasing f i s h production. 1 2 Let us then take a closer look at the s u l f i t e pulping process to see what s u l f i t e spent liquor i s and how i t pollutes. Native wood from coniferous trees consists of 42-52% a-cellulose, 10-15% hemi-eel 1ulose , 9-12% pentosans, 25-30% l i g n i n , 1-3% ethanol-benzene extract and 0.4% ash [1]. a-cellulose is a long, l i n e a r , fibrous, high molecular weight polymer of anhydroglucose units (Figure 1) linked through the 1-4 carbons. Hemi-eel 1uloses are shorter chain polyoses of, predominately xylose and mannose (Figure 1). Unlike a - c e l l u l o s e , hemi-eel 1uloses are soluble in dilute caustic soda solutions. Lignin, the 'cement' which holds the wood fibers together, is a three-dimensional polymer containing guaiacyl propane (Figure 1) and other similar monomers. The object of the pulping process i s to separate the c e l l u -lose fibers from the wood so that they can be re-arranged into a more suitable form, for example paper. In the s u l f i t e pulping process the l i g n i n matrix is dissolved in an acid b i s u l f i t e * s o l u t i o n , thus freeing the cellulose f i b e r . Because of the severity of the process most of the hemi-celluloses and some of the a-cellulose are also hydrolyzed and s o l u b i l i z e d . After the insoluble pulp, which consists • To make up t h i s b i s u l f i t e s o l u t i o n s u l f u r d i o x i d e i s d i s s o l v e d i n a b a s i c s o l u t i o n . The b a s e s u s u a l l y used a r e c a l c i u m o x i d e ( C a O ) , ammonium (NhU+) o r m a g n e s i u m o x i d e ( M g O ) . OH OH OH Guaiacylpropane Guaiacylpropane Vanillin skeleton alpha-sulfonate Figure 1 . Molecular Formulas of Wood and SSL Chemicals. 4 of a - c e l l u l o s e with traces of residual l i g n i n and hemi-cellu-lose, is removed the spent liquor is thrown out causing 'pollution. 1 Spent s u l f i t e l iquor, therefore, consists predomi-nantly of the products of the acid hydrolysis of l i g n i n and hemi-eel 1ulose with some traces of a - c e l l u l o s e residues. The l i g n i n residues are mostly sulfonates of guaiacylpropane (Figure 1). The glucose in spent s u l f i t e liquor comes mainly from the hydrolysis of cellulose while the other wood sugars probably comes from the hydrolysis of the hemi-cellulose. Table 1 shows the gross compositions of typical s u l f i t e spent liquors as determined by various workers. Table 2 breaks down the total sugars in spent s u l f i t e liquor into the individual sugar components of both deciduous and coniferous trees. Despite claims to the contrary [2], no extremely toxic compounds have yet been isolated from spent s u l f i t e liquor. Apparently the deleterious effect of the spent liquor i s caused by the 25,000 to 50,000 mg/l biochemical oxygen demand (B.O.D.) of the liquor. As w i l l be discussed in a later section (Section 2.7), this high B.O.D. i s probably primarily caused by the sugars in the s u l f i t e spent liquor -glucose, mannose, galactose, xylose and arabinose. However, Table 1 COMPOSITIONS OF TYPICAL SULFITE SPENT LIQUORS C O M P O N E N T L I B B Y [ 1 ] N I S H I K A W A [ 2 3 ] T H I S WORK M U E L L E R [ 2 5 ] Solids 100 gm/1 128 gm/1 122 gm/1 110 - 140 gm/1 Li gno-sulfoni c acids 65 79.3 - 72 - 98 Fermentable sugars 15 23.9 30.8 -Non-fermentabl e sugars 5 13.5 16.6 -Total reducing sugars 20 37.4 47.4 22 - 42 CaO 7 - - -Other 8 11.3 - 0 - 7 S0 2 - - 7.0 7 - 1 4 A l l c o n c e n t r a t i o n s g i v e n i n gm/1 l i t r e . 6 Table 2 SUGARS IN SSL FROM VARIOUS SOURCES PERCENT OF TOTAL MONOSACCHARIDES SUGAR (A) (B) (C) (D) (E) HEXOSES 76 79.8 7 1.4 26.3 11.8 Glucose 15 28.9 11.5 3.6 1.6 Galactose 10 4.2 11.5 1.9 0.0 Mannose 48 42.7 48.5 20.8 10.1 Fructose 2 4.0 PENTOSES 21 17.0 24. 3 68 .5 84 . 3 Xylose 15 20.3 68.5 79.4 Arabi nose 6 4.0 4.9 OTHER 4 3 . 2 4.4 5 .2 3.6 References: (A) 85% Western Hemlock, 15% White F i r , from Libby [1], p. 254. (B) (C) (D) Birch (E) Aspen Spruce, from Libby [1], p. 254, Spruce from Rydholm [26], p. 518, 7 even today, there is much difference of opinion as to how pulp m i l l wastes cause f i s h k i l l s . 1 . 1.2 Proposed uses for s u l f i t e spent liquor. Many diverse uses for spent s u l f i t e liquor have been proposed. These range from use as a preventative/ curative for foot and mouth disease [4] to house building material [6]. Table 3 shows a more complete l i s t of proposed and actual uses for spent s u l f i t e liquor. Some of these proposals are being carried out on an ind u s t r i a l scale. Some of these i n d u s t r i a l uses are burning as a f u e l , v a n i l l i n extraction, ethanol fermentation, and Torula yeast production. Since the worst pollution potential of spent s u l f i t e liquor is i t s high B.O.D., which is caused mainly by the wood sugars in i t , treatment of the liquor should concentrate on the removal of these sugars. Fortunately, i t is this portion of the liquor which is most readily attacked by microorganisms. Despite the almost zero cost of raw material, attempts to produce ethanol and other organic solvents from Ethanol p l a n t s at Ontario Paper Co. Ltd., T h o r o l d , Ont.; Canadian I n t e r n a t i o n a l Paper, Gatineau, P.O.; Georgia P a c i f i c , Be I I ingham , Wash., America. T o r u l a yeast p l a n t at Lake States Yeast Corp., Rhinelander, W i s e , U.S.A. Table 3 PROPOSED USES FOR SULFITE SPENT LIQUOR A. NON-FERMENTATIVE USES PRODUCT REFERENCE Burning for f u e l * Butler [27], [1]. wenzl [28], Libby Prevention of foot and mouth disease Ferenczi [4] out of Johnsen [5] V a n i l l i n (by extraction*) Libby [1] Emulsi f i e r s Libby [1] Road binders Libby [1] Reinforcement for synthetic rubbers Libby [1] Extraction of syringaldehyde for use in the production of 3 ,4,5-trimethoxyphenethy1 amine Amos [29] Building houses News item in December 21, the SUN 1970 [6] * Uses marked are i n d u s t r i a l l y important. CONTINUED CO Table 3 ( C o n t i n u e d ) B. FERMENTATIVE USES PRODUCT MICROORGANISM REFERENCE Ethanol* Sacharomyces cerevisae Kure [30], Watson [7], Mueller [25], Dahlgren [31] and Wiley [3,32]. Mushrooms Agaricus campestris, Tricholoma nudum Reusser, Spenser and Sallans [33,34]. * Torula yeast Candida utilus Wiley [32], Mueller [25]. Acetic and Propionic acids Propionibacterium arabinosum Watson [7]. P. freudenreichii Nishikawa [23,24]. Acetone and Butanol Clostridium acetobutylicum Mueller [25], Wiley [3,32]. Lactic Acid Lactabaci11 us pentosus Mueller [25], Wiley [3,32]. Fumaric acid Rhizopus sp. Mueller [25], Wiley [3,32]. Methane ? Wiley [3]. 10 spent s u l f i t e liquor have not been overwhelming successes. Why not? With respect to ethanol production from s u l f i t e spent l i q u o r , Watson [ 7 ] says: No less than 20 sources of raw material are used for producing alcohol. . . .A com-parison of manufacturing costs between the d i f f e r e n t processes reveals that synthetic ethanol (from ethylene) i s the cheapest, with ethanol from molasses and grain next in order. . . .One can e a s i l y see that the future of the s u l f i t e alcohol industry i s dependent upon the continuance of p r e v a i l i n g markets and prices. If either volume of sales or the price were to revert to pre-war (World War II) figures, alcohol manufacture from spent s u l f i t e l i q u o r would not be a competitive or economically f e a s i b l e process. . . . The primary reason ( f o r the poor compe-t i t i v e s i t u a t i o n ) i s the low concentration of alcohol obtained, and the consequently high cost of s t r i p p i n g and r e c t i f i c a t i o n . The competitive s i t u a t i o n for other simple organic solvents, which can be readily produced from feed stocks from the petroleum industry, is worse because these other organics do not enjoy the same a r t i f i c i a l government price supports and controls as ethanol does. The fermentation of spent s u l f i t e liquor and other waste carbohydrates to produce food protein for animals and/or people should increase with population. The Torula yeast process can be improved by more s p e c i f i c fermentations to produce large concentrations of s p e c i f i c essential amino 11 acids (e.g. methionine l y s i n e , leucine, i - l e u c i n e , etc.) and vitamins such as vitamin B i 2. From this beginning, the con-cept of production of complex organic compounds, which can-not be produced competitively from petroleum by-products, emerges. This work i s concerned with the production of vitamin B i 2 concentrate via Propionibacterium fermentation of s u l f i t e spent liquor. 1.2 Vitamin B i 2 and Propionibacteria Freudenreichii Vitamin B i 2 is the most complex of the B vitamins. It is found in small quantities in many foods, notably beef l i v e r (200-600 ppb), beef, pork, veal, lamb (10-25 ppb), cheese (2-9 ppb) and milk (3-5 ppb). Vitamin B i 2 has been found [8] essential for the growth of rats, mice, pigs, and + chickens. Vitamin B i 2 has been found useful in the treat-ment of an ever increasing number of c l i n i c a l conditions. Merck [8] gives an excellent review of these disorders and the effectiveness of vitamin B i 2 treatment of them. Fermentation in i t s proper sense means anaerobic b i o l o g i c a l growth. However, fermentation is a l s o commonly used f o r both anaerobic and a e r o b i c growth. In t h i s t h e s i s the term fermentation w i l l be used i n d e s c r i m i n a t e I y to d e s c r i b e both anaerobic fermentation and aerobic ' r e s p i r a t i o n . ' ^Note t h a t a l l these animals are non-ruminants. 1 2 Among these disorders are macrocytic anaemia, multiple s c l e r o s i s , l i v e r disorders, skin disorders, endocrine d i s -turbances, retarded growth and pregnancy! However, by far the greatest potential demand for vitamin B 1 2 is in i t s use as a feed supplement for livestock. In p a r t i c u l a r i t is used in poultry and swine feeds. Figure 2 , which is adapted from Williams, S t i f f e y and Jukes [ 9 ] shows the eff e c t of varying amounts of vitamin B i 2 supplement on the growth of young chicks. Among the many species of microorganisms capable of synthesizing vitamin B i 2 , the Propionibacteria species produce the highest yields of vitamin B i 2 . Let us take a closer look at one of these Propi oni bacteri a (P. freudenrei chi i) to see i f i t w i l l grow on a spent s u l f i t e liquor medium. According to Prev6t [ 1 0 ] P. freudenreichii are found in milk products such as cheese and in part i c u l a r Emmenthal . The bacteria take the form of rods 3 to 4 u long by 0 . 5 to 0 . 6 y across; they grow singly or in pairs; or even in short chains However, in certain media (including spent s u l f i t e liquor) they can take the form of cocci, singly or in chains that resemble streptococci. These bacteria are anaerobic-micro-aerophilic and grow at an optimum temperature of 3 0 ° C . Again according to Prevot [ 1 0 ] v glucose, levulose, mannose, galactose, and glycerol support the growth of Figure 2. Effect of Vitamin B i ? on Chick Growth. j i i i : i i 10 2 0 3 0 AMOUNT OF B l 2 ADDED TO CHICKS DIET i p q B|2p@r K g o f f e e d ) 14 P. f reudenreichi i with production of acid (acetic and propionic acids) and gas (C0 2). Also, nitrates are not reduced to n i t r i t e s and di-saccharides are not fermented. However, pyruvate and lactate are fermented with production of pro-pionic and acetic acids and carbon dioxide. Pantothenic acid is required by P. freudenreichii and some strains also require b i o t i n . Thiamine has also been found to stimulate growth of some st r a i n s . For building proteins P. freudenreichii require complex sources of amino acids. Acid hydrolysate of casein, corn steep liquor and yeast extract have been found [10] to be excellent for this purpose. Among the amino acids , glycine [10,11,12], and to a lesser extent asparagine [10] have been found to stimulate the growth of P. freudenreichi i . Field and Lichstein [11] have shown that the growth of Propioni-bacteria species is stimulated i f the glucose (or other car-bohydrate source), phosphate and the amino acid-vitamin complex are s t e r i l i z e d together before inoculation, rather than separately, as is often done in practice. 1.3 Biochemistry of Propionibacteria Propionibacteria are believed [10,14] to follow the well known [14] Embden-Meyerhof-Parnas pathway to convert the 15 spent s u l f i t e hexoses, glucose, galactose and mannose to pyruvic acid, thus: 6 [0] + 3 C 6 H 1 2 0 6 > 6 CH3C = 0C00H + 6H 20 ( I . I ) In order to preserve redox n e u t r a l i t y , most anae-robic microorganisms w i l l reduce pyruvic acid to l a c t i c acid [14]; most aerobes w i l l u t i l i z e atmospheric oxygen to supply the oxygen to reaction ( I . I ) . The aerobes w i l l also burn up the pyruvate produced in reaction ( I . I ) via the Krebs cycle [14], thereby extracting much more energy from the starting hexose. Propionibacteria, however, further oxidize this pyruvic acid to acetic acid and carbon dioxide, thus: 2 CH3C = 0C00H + 2 [0] > 2 CH3COOH + 2C0 2 (1.2) In order to supply oxygen to reactions ( I . I ) and ( 1 . 2 ) , Propionibacteria reduce pyruvic acid to propionic acid and oxidizing power, thus: 4 CH3C = 0C00H + 4 H20 > 4 CH3CH2C00H + 8 [0] ( I.3) 16 It is believed [14,18] that the series of reactions repre-sented by reaction ( 1.3) includes vitamin B i 2 as a co-enzyme. Reactions ( I . I ) , ( 1.2) and ( 1.3) can be added to give the overall energy metabolism reaction for Propioni-bacteri a: 3 C 6 H 1 2 0 6 2 CH3COOH + 4 CH 3CH 2C00H + 2 C0 2 + 2 H 20 ( 1 . 4 ) thus providing the 'theoretical' basis for the 2:1 molar ra t i o of propionic:acetic found by Nishikawa [23], Neronova [47], Vorob'eva [48] and Martin [56]. The mechanisms by which Propionibacteria produce vitamin B i 2 are not as well known as the energy metabolism. Some of the known steps in the synthesis of vitamin B i 2 are outlined by Greenberg [14], M i l l e r [19] and Rhem [18], Figure 3 shows the vitamin B 1 2 molecular structure. Note that this molecule is composed of three parts: the cobamide ring structure with the co-ordinated cobalt in the centre, ribose phosphate, and the nucleotide 5,6,-dimethylbenzimidole (DMBZ) linked through a 2-propanol linkage. It is int e r e s t -ing to note that in the absence of oxygen, Propionibacteria C H 2 O H Figure 3. Vitamin B 1 2 - Cobalamin. 18 w i l l not produce complete vitamin B 1 2; however, i f DMBZ i s added to the medium the true vitamin w i l l be produced. Un-fortunately, however, high levels of aeration i n h i b i t the growth of the bacteria. With s l i g h t aeration, the bacteria w i l l produce the vitamin in the absence of added DMBZ. This is the basis of the two stage fermentation process used in this work and used by Riley [22], Speedie and Hull [54], Bullerman and Berry [50,51,52], and Martin et a l . [56]. 1.4 Review of the Work of Nishikawa This work on the production of vitamin B i 2 on spent s u l f i t e liquor was begun by Nishikawa [23,24]. Propioni-bacterium arabinosum was grown on spent s u l f i t e liquor to produce acetic and propionic acids by Martin et al. [56] at Columbia Cellulose, but they did not look for vitamin B i 2 . Nishikawa used calcium base liquor which he pre-treated by addition of calcium carbonate to precipitate the s u l f i t e s , followed by addition of calcium hydroxide which precipitated much of the 1igno-sulfonates as calcium 1igno-sulfonate, followed in turn by neutralization with s u l f u r i c acid which precipitated more 1igno-sulfonates. During this pre-treat-ment process, many of the reducing sugars were also lost because they are closely bound to the 1igno-sulfonates. From 19 this pre-treated spent s u l f i t e l i q u o r , which contained approximately 10 gm/1 of reducing sugars and to which was added 10 gm/1 of yeast extract, and traces of phosphate, Nishikawa got yie l d s of from 1 to 3 gm/1 of bacterial c e l l s , from 0.5 to 1.4 mg/l of vitamin B 1 2 and from 2 to 5 gm/1 of v o l a t i l e acids with an accompanying reduction in C.O.D. of approximately 5,000-15,000 mg/l. Nishikawa was not success-* f u l in growing P. freudenreichii on untreated ammonium base or calcium base spent liquors. Calcium base liquor, while sat i s f a c t o r y for growing the bacteria, causes a l o t of processing problems. Because of the low s o l u b i l i t y of calcium sa l t s of 1igno-sulfonic acids, any r i s e in temperature, change in pH, etc. w i l l cause a precipitate to form. Because the carbohydrates in the liquor are closely associated with the ligno-sulfonates, much of the sugar can be l o s t in this way. Fortunately, however, most s u l f i t e pulp mills are switching to soluble f base processes for other reasons. Is ammonium base liquor Untreated l i q u o r s in t h i s case means th a t the S 0 2 was removed by p r e c i p i t a t i o n with CaC03 but t h a t the l i g n i n s were not p r e c i p i t a t e d by treatment with Ca(0H) 2. t One such reason i s to promote recovery of the pulp-ing chemicals by c o n c e n t r a t i o n of the l i q u o r and burning the r e s i d u e s . Another reason i s to avoid s c a l i n g problems in p r o c e s s i n g equipment caused by the p r e c i p i t a t i o n of calcium s a l t s mentioned above. 20 at least as good for growing P. freudenreichii? Is i t necessary to get r i d of the 1igno-sulfonates before good growth w i l l occur? 1.5 Objectives of this Research The objectives of this research are: 1. To determine whether ammonium base l i q u o r i s as s a t i s f a c t o r y f o r growing P. f r e u d e n r e i c h i i as calcium base l i q u o r i s . 2. To determine the minimum pre-treatment required before P. f r e u d e n r e i c h i i w i l l grow well on the l i q u o r . 3. To determine the amounts of n u t r i e n t s (carbohydrate, yeast e x t r a c t , phosphate, etc.) required f o r the growth of P. f r e u d e n r e i c h i i on spent s u l f i t e l i q u o r medium. 4. To determine which of the spent s u l f i t e l i q u o r sugars are a c t i v e in support of the growth of P. f r e u d e n r e i c h i i . 5. To determine the maximum y i e l d s of vitamin B.i 2 and v o l a t i l e o rganic acids one can expect from growth on t h i s medium in order to determine whether the production of vitamin 21 B i 2 f r o m s p e n t s u l f i t e l i q u o r i s an e c o n o m i c p r o p o s i t i o n w o r t h y o f f u r t h e r i n v e s t i g a t i o n . 6. To d e t e r m i n e f r o m b a t c h f e r m e n t a t i o n s s u f f i c i e n t k i n e t i c i n f o r m a t i o n on t h e g r o w t h o f P. f r e u d e n r e c h i i and v i t a m i n s y n t h e s i s so t h a t opt imum h o l d - u p t i m e s f o r c o n t i n u -ous f e r m e n t a t i o n p r o c e s s e s can be d e t e r m i n e d . 7. To d e t e r m i n e w h e t h e r P r o p i o n i b a c t e r i u m f e r m e n t a t i o n o f s u l f i t e s p e n t l i q u o r w i l l r e d u c e t h e p o l l u t i o n p o t e n t i a l o f t h e s p e n t l i q u o r . Chapter 2 METHODS AND APPARATUS 2.1 Spent S u l f i t e Liquor Preparation It was determined early in this work that only a minimum of preparation was necessary in order to permit growth of P. freudenreichii on ammonium base spent s u l f i t e liquor. The raw spent s u l f i t e liquor was boiled un t i l the sulfur dioxide concentration, as measured by t i t r a t i o n with N/10 iodine-potassium iodide solution, was below 200 ppm. The water lo s t through evaporation was replaced with d i s t i l l e d water to avoid concentration of the liquor. It was not found necessary to precipitate the 1igno-sulfonates , as Nishikawa [23] believed to be necessary. 2.2 Nutrient Additions to the Spent Liquor The prepared spent s u l f i t e liquor was diluted with * d i s t i l l e d water and the desired amount of yeast extract was Since the s o l u b i l i t y of yeast e x t r a c t i s 70-100 gm/1, i t was necessary to d i l u t e the spent s u l f i t e l i q u o r 22 23 added. To each l i t r e of this fermentation media was added 2 gm of ammonium sulfate ((NhV) 2 S 0 i l ) , 0.8 gm of calcium car-bonate (CaC0 3), 4 gm of potassium dihydrogen phosphate (KH2PO1J and 6.25 ppm of cobalt ion as cobalt chloride. The pH was adjusted to 7.0 with 10 N ammonium hydroxide before s t e r i l i -zation. S t e r i l i z a t i o n was accomplished by autoclaving the fermentation mixture in dir e c t contact with steam at 15 psig for 15 minutes. The pH usually dropped during s t e r i l i z a t i o n . The pH of the fermentations done in the 7 l i t r e fermentor were adjusted with 10 N ammonium hydroxide before inoculation; the runs in the Erlenmeyer flasks were not (see Section 2.4). 2.3 Inoculum Preparation § The inoculum was grown in 0X0ID CM149 Reinforced C l o s t r i d i a l Medium. The formula for this medium i s yeast extract (0X0ID L20) , 3 gm/1; beef extract, 10 gm/1; peptone (0X0ID L37), 10 gm/1; soluble starch, 1 gm/1; dextrose, 5 gm/1; w i t h d i s t i l l e d w a t e r w h e n t e s t i n g c a r b o h y d r a t e / y e a s t e x t r a c t r a t i o s o f 1:5 o r l e s s i n o r d e r t o k e e p a l l t h e n u t r i e n t i n s o l u t i o n . t P o s s i b l e r e a s o n s f o r t h i s a r e l o s s o f a m m o n i a o r h y d r o l y s i s o f I i g n o - s u I f o n a t e s . § O x o i d , L t d . ; L o n d o n , S . E . I, U . K . 24 cysteine hydrochloride, 0.5 gm/1; sodium chloride, 5 gm/1; sodium n i t r a t e , 3 gm/1 and agar-agar, 0.5 gm/1. The inoculum used was taken in the exponential growth phase which occurs from 1-3 days after starting the growth of inoculum. The amount of inoculum used varied from 1:10 to 1:100 r a t i o of inoculum to fresh medium. No true lags were noted in any of the runs using inoculum prepared this way. Stock culture was kept in the same medium with 2 gm/1 of extra agar added to make i t s o l i d , and this stock culture was kept under r e f r i g e r a t i o n at 5°C for up to six>months with no noticable deterioration. The i n i t i a l culture was started from culture obtained from the American Type Culture Collection as ATCC-6207 [35]. 2.4 Equipment and Apparatus Most of the preliminary and screening experiments were carried out in standard, 150-250 ml, Erlenmeyer flasks stopped with non-absorbent cotton. These flasks were incu-bated at 30°C in a converted re f r i g e r a t o r incubator. The temperature was controlled to within 1C° except once when the refrigerant leaked out. It was noted that i f the flasks were aerated by shaking the flasks during the f i r s t few days of a fermentation, l i t t l e or no growth occurred. It was 25 rationalized from this that i t was essential to prevent oxygen from penetrating the medium at the beginning of the fermentation but that after the bacteria were growing vigor-ously they could maintain the low oxygen tension that they l i k e , even when the fermentation medium was aerated s l i g h t l y . For the 7 - l i t r e fermentations a New Brunswick * S c i e n t i f i c model MF-07 bench top fermentor was used. A schematic drawing of this fermentor showing the geometry of the ag i t a t i o n , b a f f l e s , a i r and sample i n l e t s / o u t l e t s is shown in Figure 4. With this apparatus i t is possible to ferment up to 5 l i t r e s of broth, control the temperature to within 1°C of a set point within a range of 10°C to 50°C, control the agitation rate from 100 to 900 rpm, and to control an a i r flow rate from 1 to 16 litres/minute. In addition, the fermentor can be s t e r i l i z e d with the fermenta-tion medium in i t ; samples can be removed and aseptic addi-tions made at any time during the fermentation. 2.5 Measurement of Bacterial Growth There are many ways of measuring bacterial growth. Among them are: New B r u n s w i c k S c i e n t i f i c C o . , I n c . ; New B r u n s w i c k , N . J . , U . S . o f A . 26 Figure 4. Schematic Drawing of 7-Lttre Fermentor. 27 1. T u r b i d i t y . Bacterial c e l l s interfere with and absorb a certain portion of any l i g h t which is passed through a suspension of them. At low optical densities,* one finds that optical density is proportional to the dry c e l l weight per unit volume. At higher optical densities the t u r b i d i t y is reduced due to back scattering of the l i g h t [36]. 2. Dry C e l l Weight. The amount of dry c e l l u l a r matter is measured d i r e c t l y after drying a unit volume or weight of the fermentation broth under controlled conditions. This test method and the relationship between t u r b i d i t y and dry c e l l weight is discussed further in Appendix I. 3. Viable or Plate Count. As explained by Stanier et a l . [36], this technique measures the number of viable units in a culture, be they single c e l l s or whole colonies clustered together. That i s , i t measures the number of separable e n t i t i e s capable of reproduction. 4. C o u l t e r Counter. This remarkable piece of machinery is an electronic device which can automatically O p t i c a l d e n s i t y i s a l s o c a l l e d t u r b i d i t y o r a b s o r -b a n c e . S e e N o m e n c l a t u r e u n d e r A B S . 28 measure both c e l l numbers and r e l a t i v e size of the individual c e l l s . The method is based upon the r e l a t i v e e l e c t r i c a l conductivity of bacterial c e l l s and the media in which they are suspended as the bacterial c e l l s pass through an aperture. In this work we are primarily interested in the amount of dry c e l l matter produced per unit volume of f e r -mentation broth. However, measuring t u r b i d i t y has certain advantages over the measurement of dry c e l l weight. These advantages are: (a) Turbidity is a faster and easier test. (b) It is more accurate i f s u f f i c i e n t replications are made. (c) Smaller sample sizes are required. Therefore, while bacterial growth w i l l be reported as dry c e l l weight per unit volume, the actual tests were made turbidometrically. Dry c e l l weights per unit volume were calculated from the formu1 a DCW = 0 . 2 8 9 5 x T U R B x E X P ( Cv. 6078 . * T U R B ) x D where DCW i s d r y c e l l w e i g h t p e r u n i t v o l u m e ( g m / 1 ) , T U R B i s t h e m e a s u r e d t u r b i d i t y , and D i s t h e d i l u t i o n f a c t o r . I f t h e c e l l s w e r e d i l u t e d 10:1 w i t h d i s t i l l e d w a t e r b e f o r e t u r b i d i t y was m e a s u r e d , D = 1 0 . 29 The derivation of this equation and the s p e c i f i c test methods for dry c e l l weight and t u r b i d i t y are given in detail in Appendix I. 2.6 Measurement of Vitamin B X 2 Concentration There are two basic methods for measurement of vitamin B i 2 concentration. These are: 1. Chem-ioal Method. This method i s outlined by Nishikawa [23]. In this method the vitamin is extracted and concentrated in organic solvnets and the absorbance measured with a spectrophotometer at a c h a r a c t e r i s t i c wavelength for vitamin B i 2. 2. M l c v o b i o l o g - t c a l Method. In this method a micro-organism such as Lactobacillus leichmannii, which has an absolute requirement for vitamin B i 2 , is grown on an unknown sample and i t s growth is compared to the growth of the same microorganism on nutrient medium containing known amounts of the vitamin [37,38,39]. In this work the chemical method was found to be unsatisfactory for the following reasons: 3 0 1 . E x c e s s i v e foaming of the two phase e x t r a c t i o n mixture at the organic/aqueous i n t e r f a c e prevented the separa-t i o n of the phases and the recovery of the vitamin was not q u a n t i t a t i v e . 2. The fermentation c o n c e n t r a t i o n s of vitamin were very low and as a r e s u l t more c o n c e n t r a t i o n than is usual E233 was necessary in the e x t r a c t i o n procedure r e s u l t i n g in too much unrecovered v i t a m i n . 3 . Many substances in the fermentation broth appear to i n t e r f e r e with the spectrophotometer a b s o r p t i o n curve. It was found that the e r r o r s in the c o r r e c t i o n s f o r these sub-stances (see Nishikawa, p. 16 C2 3H f o r these c o r r e c t i o n s ) were greater than the amount of vitamin B i 2 measured. That i s , the measured vitamin c o n c e n t r a t i o n s were not s i g n i f i c a n t l y d i f f e r e n t from zero. A. Many samples were measured at the same time. Large numbers of samples are more e a s i l y handled by the micro-b i o l o g i c a l method. A l l vitamin B . i 2 test results reported in this work are done by the microbiological method. The d e t a i l s of the test method and the method used to calculate the confidence li m i t s on the microbiological test are given in Appendix II. 31 2.7 Calculation of Biochemical Oxygen Demand and Chemical  Oxygen Demand Neither b i o l o g i c a l oxygen demand (BOD) nor chemical oxygen demand (COD) were measured in this work. The accuracy of, interpretation of, and u t i l i t y of BOD and COD test results on s u l f i t e spent liquor is not f u l l y understodd. Mueller [25], for example, states that The read-fly fermentable sugars are the •primary cause for the high B0D5 of spent s u l f i t e l i q u o r , which ranges from 25,000 to 40,000 ppm. The l i g n i n frac-tion of SSL i s very r e s i s t a n t to micro-b i o l o g i c a l attach, and therefore con-t r i b u t e s l i t t l e to the B0D5. Also, Casey [40] states that the sugars of s u l f i t e spent liquor which contribute to only 20% of the total dissolved s o l i d s , contribute 75%-95% of the BOD. However, contrary to t h i s , Wiley et a l . [32] state that . . . only the hexose sugars in the spent l i q u o r are converted into alcohol by fer-mentation (by S. c e r e v i s i a e ? ) . These sugars account for 50% or less of the 5-day BOD, which means that t h i s i s a far from s a t i s factory stream improvement measure. Table 4 which shows the amount of BOD removed from spent s u l f i t e liquor by various treatments, is taken from Wiley [32] Table 4 B.O.D. REMOVAL FROM SPRUCE WASTE LIQUORS (Data are only approximate) EXTENT OF TREATMENT COMPOSITION CHANGE RESIDUAL B0D5 (mg/l) BOD REMOVED (% of i n i t i a l ) None Whole liquor 35,000 0 Stripping or pr e c i p i t a t i o n S0 2 removed 31 ,500 10 Pre-treatment alcoholi c fermentati on & S0 2 & hexoses removed 15,750 55 Pre-treatment Torula yeast treatment & S0 2, hexoses, toses & some acid removed pen-aceti c 8,750 75 T a b l e t a k e n f r o m W i l e y [ 3 2 1 . 33 who then goes on to suggest that, "the remaining 25% of the BOD in the yeast plant effluent results from carbohydrate material derived d i r e c t l y or i n d i r e c t l y from the hydrolytic breakdown of sugars," and he continues, "l a t e s t evidence points to the r e l a t i v e s t a b i l i t y of these compounds (ligno-sulfonates). Appreciable degrees of 1igno-sulfonate destruc-tion apparently do not take place b i o l o g i c a l l y in less than 20 days, and this slow rate probably does not show up in streams by present methods of measurement." I might add my opinion that such a slow b i o l o g i c a l degradation would not l i k e l y cause a dissolved oxygen dip in a moving stream. Chemical oxygen demand (COD) is a test widely used to supplement BOD testing because i t is a much shorter test (3 hours vs. 15 days) and not as tedious. This test involves oxidizing organic material with an excess of potassium dichromate and back t i t r a t i n g the mixture to determine the i n i t i a l organic reducing material present. According to Sawyer and McCarty [41], this test oxidizes carbohydrates and inorganic reducing matter r e a d i l y , i t reduces low molecular weight hydrocarbons i f s i l v e r ions are present; however, aromatic hydrocarbons and pyridine are not oxidized under any circumstances. Considering 1igno-su1fonates as aromatic hydrocarbons, i t is unlikely that 1igno-sulfonates are oxidized in this test. 34 In order to assess the effect of P. freudenreichii upon the pollution potential of the spent li q u o r , I w i l l use the ultimate oxygen demand (UOD) concept. With UOD we calculate the amount of oxygen required to completely oxidize a l l the chemical species to their highest normal oxidation states; thus: C 6 H i 2 0 6 + 6 0 2 6C0 2 + 6 H20 1067 mg 02/gm sugar S0 2 + \ 0 2 S0 3 250 mg 02/gm S0 2 Table 5 shows the UOD reductions which can be expected from the removal of a l l of the normal amount of each species present in spent s u l f i t e liquor before and after fermentation. Experiments have shown [41] that the BOD for glucose is from 75% to 90% of the UOD. The B0D5/U0D rat i o for acetic and propionic acids is probably 50-100%, and that for the 1igno-sulfonate is probably approximately zero. Therefore l i k e Wiley in Table 4, we can assume an approximate 55% reduction in B0D5 by a combination of S0 2 removal, Propioni-bacterium fermentation and subsequent recovery of the v o l a t i l e acids and the c e l l u l a r material. We also maintain that Table 5 REDUCTION OF UOD FOR REMOVAL OF 1 GM/l OF SSL COMPONENT PRODUCT AUOD/lgm/1 REMOVAL* (mg/1) USUAL AMOUNT OF COMPONENT IN LIQUOR (gm/1) NET REDUCTION IN UOD (mg/1) S0 2 250 7.0 1 ,750 Hexoses 1067 31.2 33,400 Total sugars 1067 48.0 51 ,300 Acetic acid 1067 6.95 (max.) 5.67 (norma 1) 7,400 7 ,200 Propionic acid 1520 17.1 (max.) 14.0 (normal) 26,000 21 ,300 Lignin (as vana1 1i n) (as syringa1dehyde) 2160 3010 75.0 162,000 225,000 A s s u m p t i o n s : 1. It was assumed t h a t a l l c o m p o n e n t s a r e o x i d i z e d to t h e i r h i g h e s t o x i d a t i o n s t a t e . 2. The maximum a c i d y i e l d s were c a l c u l a t e d f r o m t h e t h e o r e t i c a l e q u a t i o n 3 C 6 H i 2 0 = h CH 3CH 2 C00H + 2 CH3COOH + 2C02 + 2 H 2 0 3. The norma l a c i d s y i e l d s were c a l c u l a t e d f r o m y i e l d c o e f f i c i e n t s f o u n d i n C h a p t e r 4. If the co n c e n t r a t i o n of the component in the f i r s t column i s reduced by I gm /1 , then the UOD of the l i q u o r w i l l be reduced by the number in the second column in mg/l reduct i o n in UOD. The la s t column, Net Reduction in UOD, shows the reduction in UOD one might expect i f a I I of any p a r t i c u l a r component of the l i q u o r were e l i m i n a t e d . co 36 because of the u n r e l i a b i l i t y of the BOD and COD t e s t s , that testing of these would only obscure the issue. 2.8 Measurement of Reducing Sugar Concentration The method used by Nishikawa [23] was used with the following modifications: 1. F o r e a c h new b a t c h o f r e a g e n t s , a new s t a n d a r d c u r v e o f gm/1 o f r e d u c i n g s u g a r v s . t i t r a t i o n w a s m a d e u s i n g 0 , 1 0 , 2 0 , 3 0 , 40 a n d 50 g m / 1 o f g l u c o s e . 2 . A s m a n y a s e i g h t s a m p l e s w e r e r u n s i m u l t a n e -o u s l y , r e d u c i n g t h e n e e d f o r a s m a n y b l a n k r u n s a s n o r m a l . 2.9 Measurement of Acetic and Propionic Acids A few v o l a t i l e acids measurements were made using the method of Nishikawa [23]. These results are called 'by actual measurement' in the text. However, in the 7 - l i t r e fermentations, when the pH was adjusted to keep the pH above 5.5, the amount of 10 N ammonium hydroxide necessary to adjust the pH to 7.0 and the amount of l i q u i d in the fermen-tator were recorded. Therefore, the amount of acids pro-duced since the previous pH adjustment could be calculated. Such results are labelled in the text 'by pH adjustment.' 37 In this c alculation i t i s assumed that the contribution of carbon dioxide to the hydrogen ion strength of the fermenta-tion broth is small. This is not a bad assumption consider-ing the d i s s o c i a t i o n constants of the acids involved; the error introduced w i l l be less than 5%. Chapter 3 RESULTS Most of the results of this work require a l o t of discussion and comment. In order to f a c i l i t a t e the presen-tation of these results and the analysis, the results w i l l be presented in this section with l i t t l e or no comment. They w i l l then be reorganized and discussed in later sections. This applied p a r t i c u l a r l y to the product yields and the kinetic studies. 3.1 I n i t i a l Yield Studies In this work two sources of spent s u l f i t e liquor were used. These were: 1. Ammonia based liquor from the Columbia Cellulose p i l o t plant at New Westminster, and 2. Ammonia based liquor from Rayonier's plant at Port Angeles, Wash. Furthermore, two types of yeast extract were used as supplementary nutriets for Propionibacteriurn freudenreichii * growth: 1. Powder yeast extract, and 2. Paste yeast OXOID L 2 I powder yeast e x t r a c t , Oxoid, L td., London S.E.I, U.K. 38 39 extract. The powder yeast extract is es s e n t i a l l y 100% dry material. The paste yeast extract was tested to be 73.8% dry matter; the s o l i d content of which is claimed by the manufacturer to be ide n t i c a l to the powder yeast extract. These d i f f e r e n t sources of raw material were tested against each other for their a b i l i t y to support growth of P. freudenrei chi i . An i n i t i a l test run in 25 ml test tubes, which were 3/5 f u l l of fermentation broth, was made as a preliminary study of the yeast extract and spent s u l f i t e liquor sugars required for P. freudenreichii growth. The fermentation charge data and the results of this experiment are l i s t e d in Table 6. As this test was inconclusive, four further tests were made in 250 ml Erlenmeyer f l a s k s . These tests were: 1. A t e s t of Columbia C e l l u l o s e SSL and powder yeast e x t r a c t at four d i f f e r e n t l e v e l s of SSL and yeast e x t r a c t , the r e s u l t s of which are shown in Table 7-2. Another t e s t with Columbia C e l l u l o s e SSL and powder yeast e x t r a c t with seven l e v e l s of SSL and yeast e x t r a c t . In t h i s t e s t the f l a s k s were sampled at s i x d i f f e r e n t OXOID L 2 0 paste yeast e x t r a c t , Oxoid, L t d . , London S.E.I, U.K. Table 6 I N I T I A L TEST TUBE TESTS TO DETERMINE Y I E L D BASED ON Y E A S T EXTRACT INITIAL REDUCING SUGAR CONCENTRATION ( g m / 1 ) INITIAL YEAST EXTRACT CONCENTRATION ( g m / 1 ) DRY CELL WEIGHT (gm/1) AFTER 18£ h r . 92 h r . 308 h r . 27.6 27.3 27.0 0.0 0.9 2.2 0.243 0.317 0.340 0.384 0.431 0.424 0.384 0.417 0.458 26.4 4.4 0.321 0.414 0.511 25.4 8.3 0.370 0.542 0.607 23.5 15.4 0.521 0.610 0.672 20.4 26.6 0.625 1 .022 0.815 I n i t i a l c e l l c o n c e n t r a t i o n e s t i m a t e d t o be 0.2 g m / 1 . I n o c u l u m c o n c e n t r a t i o n was 1 : 1 5 , i n o c u I u r n : m e d i u r n . O X O I D p o w d e r y e a s t e x t r a c t was u s e d . C o l u m b i a C e l l u l o s e s u l f i t e s p e n t l i q u o r was u s e d . Table 7 P R E L I M I N A R Y Y E A S T E X T R A C T Y I E L D T E S T S I N 250 M L E R L E N M E Y E R F L A S K S INITIAL REDUCING SUGAR CONCENTRATION (gm/1 ) INITIAL YEAST EXTRACT CONCENTRATION ( gm /1 ) DRY CELL WEIGHT (gm/1) AFTER 156 hr. 297 hr. 40 0 0.140 0.127 28 9.1 0.610 0.790 20 22.6 2.417 2.537 0 45.4 0.721 1.142 I n i t i a l d r y c e l l w e i g h t : 0 . 0 9 6 6 g m / 1 . 150 ml o f f e r m e n t a t i o n b r o t h we re used i n 250 ml E r l e n m e y e r f l a s k s . OXOID p o w d e r y e a s t e x t r a c t was u s e d . I n o c u l u m c o n c e n t r a t i o n : 1 : 3 0 , i n o c u I um : f r e s h m e d i u m . C o l u m b i a C e l l u l o s e s u l f i t e s p e n t l i q u o r was u s e d . 42 t i m e s d u r i n g t h e p r o g r e s s o f t h e f e r m e n t a t i o n i n o r d e r t o g e t some p r e l i m i n a r y k i n e t i c i n f o r m a t i o n on t h e f e r m e n t a t i o n . T h e s e r e s u l t s a r e l i s t e d i n T a b l e 8. 3. A run s i m i l a r t o 2. above was made w i t h R a y o n i e r SSL and powder y e a s t e x t r a c t a t e i g h t d i f f e r e n t l e v e l s o f SSL and y e a s t e x t r a c t . T h e s e r e s u l t s a r e l i s t e d i n T a b l e 9. 4. The f i n a l run i n t h i s s e r i e s was made w i t h R a y o n i e r SSL and p a s t e y e a s t e x t r a c t . T h e s e r e s u l t s a r e l i s t e d i n T a b l e 10. 3.2 Inhibitory Effect of Sulfur Dioxide on Cell Yield Sulfur dioxide is reported to have an in h i b i t o r y e f f e c t on microorganisms. It is used as a preservative in many foods. The l i t e r a t u r e , unfortunately, i s rather vague about the s p e c i f i c amounts of S0 2 which are in h i b i t o r y to the growth of microorganisms. This experiment was intended to provide some firm information on exactly how much of the sulfur dioxide i t w i l l be necessary to remove from the spent s u l f i t e liquor before fermentation can take place. As can be seen from Table 11, various amounts of sulfur dioxide, added as sodium s u l f i t e , were added to spent T a b l e 8 Y I E L D ON Y E A S T E X T R A C T U S I N G C O L U M B I A C E L L U L O S E S S L A N D POWDER Y E A S T E X T R A C T I N I T I A L R E D U C I N G SUGAR ( g m / 1 ) 2 7 . 0 2 3 . 8 2 7 . 0 1 6 . 3 1 4 . 0 9 . 5 0 . 0 I N I T I A L Y E A S T E X T R A C T ( g m / 1 ) 1 2 . 5 1 8 . 7 5 2 5 . 0 3 1 . 2 5 3 7 . 5 4 3 . 5 7 5 0 . 0 DRY C E L L W E I G H T ( g m / i ) A F T E R 0 h r . 0 . 0 2 7 0 . 0 2 5 0 . 0 2 6 0 . 0 2 7 0 . 0 2 8 0 . 0 3 0 0 . 0 2 9 4 4 h r . 0 . 3 8 6 0 . 5 8 5 0 . 9 7 0 1 . 3 6 4 1 . 4 3 6 1 . 2 7 9 0 . 5 2 1 8 9 h r . 1 . 0 8 7 1 . 7 5 0 2 . 7 1 0 2 . 4 7 8 2 . 6 6 8 2 . 4 7 5 0 . 8 1 3 1 3 4 h r . 1 . 2 4 0 1 . 9 8 2 2 . 4 6 7 2 . 4 5 0 3 . 2 0 9 2 . 4 5 6 0 . 8 3 5 1 8 0 h r . 1 . 3 4 6 1 . 9 9 2 2 . 2 6 8 2 . 4 4 9 3 . 6 0 0 2 . 6 1 6 0 . 9 7 7 2 5 3 h r . 1 . 5 4 6 2 . 2 0 8 2 . 4 1 7 2 . 6 5 1 3 . 6 6 9 2 . 9 2 7 1 . 4 6 3 R E D U C I N G S U G A R ( g m / 1 ) A F T E R 0 h r . 3 7 . 0 2 3 . 8 2 7 . 0 1 6 . 3 1 4 . 0 9 . 5 4 4 h r . 3 8 . 3 2 6 . 0 1 0 . 5 — — 8 9 h r . 3 6 . 5 2 6 . 0 9 . 3 --1 3 4 h r . 3 6 . 5 _ _ 2 0 . 0 6 . 5 _ _ _ _ 1 8 0 h r . 3 5 . 0 2 0 . 5 — 7 . 0 • 2 5 3 h r . 3 4 . 0 _ — 2 0 . 5 _ — 6 . 0 C o l u m b i a C e l l u l o s e s u l f i t e s p e n t l i q u o r was u s e d . O X O I D p o w d e r y e a s t e x t r a c t was u s e d . I n o c u l u m c o n c e n t r a t i o n : I : 30 : : i n o c u I urn : f r e s h m e d i u m . co Table 9 YIELD ON YEAST EXTRACT USING RAYONIER SSL AND POWDER YEAST EXTRACT INITIAL REDUCING SUGAR CONCENTRATION (gm/1) INITIAL YEAST EXTRACT CONCENTRATION (gm/1) DRY CELL WEIGHT (gm/1) AFTER 137 i h r . 47 .8 0 0 0. I l l 37 .2 9 6 0. 473 23 8 21 5 2. 070 11 9 32 4 2. 774 11 9 32 4 2. 759 9 0 35 1 2. 641 5 9 37 8 2. 016 0 0 43 1 0. 678 R a y o n i e r S S L was u s e d . O X O I D p o w d e r y e a s t e x t r a c t was u s e d . I n o c u l u m c o n c e n t r a t i o n was a p p r o x i m a t e l y I : 3 0 : : i n o c u I u m : f r e s h m e d i u m . Table 1 0 YIELD ON YEAST EXTRACT USING RAYONIER SSL AND PASTE YEAST EXTRACT INITIAL REDUCING SUGAR CONCENTRATION ( g m/i) INITIAL YEAST EXTRACT CONCENTRATION ( g m/i) DRY CELL WEIGHT ( g m / 1 ) AFTER a s r e c e i v e d d r y b a s i s 145 h r . f e r m e n t a t i o n 4 7 . 8 2 1 . 8 0 . 0 2 0 . 0 0 . 0 1 4 . 7 0 . 2 6 0 1 . 0 0 2 2 3 . 8 3 0 . 0 2 2 . 1 1 . 5 6 7 1 6 . 0 1 6 . 0 1 2 . 0 4 0 . 0 4 0 . 0 4 5 . 0 2 9 . 5 2 9 . 5 3 3 . 2 2 . 2 6 1 2 . 2 5 4 2 . 4 2 6 8 . 0 0 . 0 5 0 . 0 6 0 . 0 3 6 . 8 4 4 . 3 . . 2 . 5 1 7 0 . 9 3 1 I n o c u l u m c o n c e n t r a t i o n 1 :30 i n o c u I u r n : f r e s h m e d i u m . 46 Table 11 TEST FOR YIELD INHIBITION OF SULFITE IN FERMENTATION BROTH SULFUR DIOXIDE CONCENTRATION DRY CELL WEIGHT (gm/1) AFTER (ppm) 145 h r . 0 2.171 100 2.131 200 2.085 400 1 .994 800 1 .749 1600 1 .401 3200 0.871 I n i t i a l r e d u c i n g s u g a r c o n c e n t r a t i o n - R a y o n i e r S S L 2 8 . 8 g m / I . I n i t i a l p a s t e y e a s t e x t r a c t c o n c e n t r a t i o n - 20 g m / 1 a s r e c e i v e d - 1 4 . 7 gm/1 d r y b a s i s . E s t i m a t e d r e s i d u a l S 0 2 l e v e l i n S S L - l e s s t h a n 2 0 0 p p m . E x p e r i m e n t was d o n e u s i n g 150 m l . b r o t h i n 2 5 0 m l . E r l e n m e y e r f l a s k s . 47 s u l f i t e liquor which had been previously stripped of S0 2 to less than 100 ppm and yeast extract medium. The growth of P. freudenreichii on these media was measured after 145 hours. The results of this experiment, shown in Table 11 and Figure 5, indicate that the i n h i b i t i o n effect of s u l f i t e i s linear with S0 2 concentration. Furthermore, i t is apparent that i f the sulfur dioxide concentration i s reduced below 200 ppm, there w i l l be no s i g n i f i c a n t decrease in the ultimate y i e l d of P. freu d e n r e i c h i i . Two things should be noted about this t e s t : 1) The s u l f i t e was added to the Erlenmeyer flasks before s t e r i l i z a -tion and some loss of S0 2 may have occurred during s t e r i l i z a -t i o n . 2) No rate i n h i b i t i o n was looked for. That i s , even though 200 ppm of S0 2 does not decrease the y i e l d of bacteria, there could well be some decrease in the rate at which the bateria were produced. 3.3 U t i l i z a t i o n of Spent S u l f i t e Liquor Sugars As shown in Table 2, Section 1.1, the main wood sugars present in spent s u l f i t e liquor are the hexoses-glucose, mannose and galactose and the pentoses-xylose and arabinose. Which of these sugars support the growth of P. freudenreichii? The experiment shown in Table 12 seeks the answer to this Table 12 GROWTH OF P, FREUDENREICHII ON VARIOUS WOOD SUGARS SUGAR USED RESIDUAL REDUCING SUGAR (gm/1) AFTER DRY CELL WEIGHT (gm/1) AFTER 0 h r . 97 h r . 193 h r . 97 h r . 193 h r . SSL 20.0 19.0 15.6 0.961 1 .623 SSL 19.3 18.7 17.7 0.993 1 .650 Glucose 20.5 15.3 14.2 2.406 2.620 Glucose 21.1 15.4 14.6 2.473 1 .970 Mannose 20.6 17.2 15.9 1.941 2.231 Mannose 20.8 17.6 15.9 1 .977 2.077 Galactose 19.6 15.7 13.8 1.703 1 .977 Galactose 19.6 14.3 13.2 1 .688 2.104 Arabi nose 19.9 19.0 20.7 0.525 0.410 Arabi nose 20.1 19.4 22.0 0.500 0.416 Xy1ose 21.9 22.2 23.1 0.545 0.535 Xylose 21.7 22.3 23.3 0.563 0.483 No e x t r a s u g a r ( e s t i m a t e d from r e s u l t s o f S e c t i o n 3.1) 0. 480 N O T E S : I n i t i a l p a s t e y e a s t e x t r a c t c o n c e n t r a t i o n 2 8 . 8 g m / 1 a s r e c e i v e d , 2 1 . 2 gm/1 d r y e x t r a c t b a s i s . F e r m e n t a t i o n v e s s e l 150 ml E r l e n m e y e r f l a s k s . F e r m e n t a t i o n b r o t h u s e d 100 m l . I n o c u l u m 5 ml p r e p a r e d a c c o r d i n g t o S e c t i o n 2 . 1 . Figure 5. Effect of S0 2 on Yield of Propionibacteria 0 IOOQ 2OOJ0 3 0 0 0 INITIAL S 0 2 ADDED ( m g / l ) 50 question. These f i v e sugars, at concentrations of approxi-mately 20 gm/1 , along with spent s u l f i t e liquor with a similar sugar concentration were tested for their support of the growth of the microorganism. Table 12, which shows the results of this experiment, indicates that the hexoses are u t i l i z e d by the bacteria but that the pentoses are not. More-over, the fra c t i o n of spent s u l f i t e liquor sugar which is active in supporting growth calculated from this experiment is 0.68. This figure agrees very well with other methods of calculation as discussed in Section 4.1. 3 .4 7-Litre Fermentor Studies Fermentation studies in Erlenmeyer flasks are un-sati s f a c t o r y for many purposes. Among the reasons for this are: 1. I t i s d i f f i c u l t t o t a k e e i t h e r l a r g e or many-s m a l l samples from f l a s k s f o r use i n k i n e t i c s t u d i e s . 2. I t i s d i f f i c u l t t o a d j u s t t h e pH o f f e r m e n t a -t i o n b r o t h i n f l a s k s w i t h o u t c o n t a m i n a t i n g t h e f e r m e n t a t i o n . T h u s , f o r y e a s t e x t r a c t l i m i t i n g c u l t u r e s , i t i s i m p o s s i b l e t o go t o y e a s t e x t r a c t c o n c e n t r a t i o n s g r e a t e r t h a n 40 t o 50 gm/1. At y e a s t e x t r a c t c o n c e n t r a t i o n s i n e x c e s s o f 50 gm/1, 51 there w i l l he so much a c i d produced that the growth of b a c t e r i a w i l l cease because of a c i d i n h i b i t i o n r a t h e r than from n u t r i e n t l i m i t a t i o n s . 3. It i s d i f f i c u l t to c o n t r o l a g i t a t i o n / a e r a t i o n i n Erlenmeyer f l a s k s . Since i t i s b e l i e v e d that a two-stage pro-cess, the f i r s t stage of which i s anaerobic, the second stage of which i s m i c r o - a e r o b i c , i s e s s e n t i a l f o r the p r o d u c t i o n of l a r g e amounts of v i t a m i n B12, the maximum y i e l d s of v i t a m i n Bi2 cannot be determined from experiments i n these f l a s k s . The 7 - l i t r e fermentor described in Section 2.4 was, therefore, used for these kinetic studies. Five 7 - l i t r e fermentations were run. One of these used 38 gm/1 of OXOID reinforced c l o s t r i d i a l medium with 30 gm/1 of glucose added. The purpose of this run was to check out the fermentor. The results of this run are reported in Table 13 and shown graphically in Figure 6. The four runs on spent s u l f i t e liquor are described in Tables 14, 15, 16 and 17. These results are also shown graphically in Figure 7, 8, 9 and 10. In the f i r s t two SSL runs the pH control was poor; the pH was allowed to drop below 6.0. In Run 4, with the addition of s t e r i l i z a b l e pH electrodes to the fermentor, the pH control improved greatly. Also on Run 4 the control on the vitamin B i 2 test was much closer because of better 5 2 Table 1 3 7-LITRE FERMENTATION ON CLOSTRIDIAL MEDIUM AND GLUCOSE TIME (hr) DRY CELL WEIGHT (gm/i) REDUCING SUGAR (gm/i) pH 0 0 . 0 5 0 3 1 . 5 6 0 . 0 8 0 3 3 . 2 -11 0 . 1 2 8 -- -2 2 0 . 4 2 1 _ _ _ 2 9 i 0 . 7 1 7 3 2 . 0 5 . 2 3 4 0 . 8 0 4 3 0 . 0 -4 6 1 . 2 6 7 2 7 . 5 5 3 1 . 2 4 5 — -5 9 1 . 3 7 3 2 8 . 4 -7 0 1 . 5 9 8 2 7 . 5 _ 8 0 1 . 6 0 7 2 8 . 4 -9 8 1 . 8 0 6 2 7 . 3 4 . 8 1 2 2 2 . 5 8 2 2 5 . 0 1 4 6 2 . 6 4 0 — -1 7 0 2 . 5 1 3 2 4 . 3 -1 9 4 2 . 4 0 0 2 3 . 5 -A i r was t u r n e d o n a t f l o w o f i v o l u m e o f a i r p e r v o l u m e o f f e r m e n t a t i o n b r o t h a t 102 h o u r s f e r m e n t a t i o n t i m e . A g i t a t o r s p e e d - 2 5 0 r p m . I n i t i a l c l o s t r i d a l m e d i u m c o n c e n t r a t i o n 38 g m / 1 . I n i t i a l a d d e d g l u c o s e 30 g m / 1 . I n o c u l u m c o n c e n t r a t i o n 1 : 4 0 p r e p a r e d i n t h e u s u a l m a n n e r . 53 Q. T I M E ( hours ) Figure 6. 7-Litre Fermentor Run with C l o s t r i d i a l Medium-Glucose. 54 Table 14 7-LITRE FERMENTATION - SSL RUN NO. 1 TIME (hr) DRY CELL WEIGHT (gm/l) ACIDS (gm/l) REDUCING SUGARS (gm/l) PH VITAMIN B 1 2 (mg/l) 0 0.048 _ 20.6 _ 1 0.063 - 18.0 -3 0.062 - - 6.7 8 0.078 _ 18.22 6.7 13 0.104 - - 6.65 22 0.224 - 19.52 6.55 27 0.414 18.95 6.4 32i 0.584 2.53 19.52 6.3 36 0.781 - 19.38 6.2 46 1 .644 3.02* 17.80 6.0 58 2.444 3.95 14.90 6.15 70 3.152 5.86 12.86 6.2 79 3.912 _ 10.48 6.1 96 4.811 8.25* 9.32 5.82 1.52 ± 0.20 101 4.813 5.86 7.56 6.5 120 4.849 8.71 6.44 6.5 127 4.552 9.28 6.82 7.0 143 4.518 - 7.20 7.0 2.14 ± 0.97 166 4. 535 _ 6.48 7.0 2.89 ± 1 .93! 176 4.955 - 6.44 6.95 190i 5.007 - 6.65 6.9 215 5.624 _ 6.87 6.9 223i 5.980 - - 6.9 238 6.147 9.04 6.82 6.9 3.0.1 ± 3.74! A c i d c o n c e n t r a t i o n f o u n d by t i t r a t i o n o f w h o l e b r o t h w h e n a d j u s t i n g up t o pH 7 . 0 . CONTINUED 55 Table 14 (Continued) NOTES: A i r w a s t u r n e d o n a t 124 h o u r s a t 1 /3 v o l . a i r / v o l . f e r m e n -t a t i o n b r o t h w i t h a g i t a t o r a t 3 . 3 3 h z ( 2 0 0 r p m ) . I n i t i a l y e a s t e x t r a c t : 72 g m / 1 a s r e c e i v e d , 53 g m / 1 d r y . A c i d s w e r e c a l c u l a t e d a s s u m i n g t h a t t h e y w e r e p r o d u c e d i n t h e t h e o r e t i c a l r a t i o o f 2:1 m o l e f r a c t i o n . ! M e a n s o f t h e s e r e s u l t s a r e n o t s i g n i f i c a n t l y d i f f e r e n t f r o m z e r o . 56 Table 15 7-LITRE FERMENTATION - SSL RUN NO. 2 TIME (hr) DRY CELL WEIGHT (gm/l) ACIDS (gm/l) REDUCING SUGARS (gm/l) pH VITAMIN B.i 2 (mg/l) 1 0.071 31.15 7.0 10 0.096 6.8 22i 0.242 6.7 32 0.584 6.6 51± 1 .411 31.25 6.3 0.32 ± 0.03 77 5.85 96 3.654 27.65 5.7 0.35 ± 0.07 106 3.971 5.6 118i 4.121 6.62 5.6-7.0 120 4.750 21 .63 6.75 143* 5.710 8.39 6.6-7.0 166i 6.342 9.98 16.97 6.6 2.28 ± 0.4 190i 6.743 11 .04 6.7 217 6.757 11 .89 6.8 244 6.850 12.02 6.9 263 6.811 7.0 285i 6.824 7.0 2.19 ± 0.33 A c i d s c o n c e n t r a t i o n d e t e r m i n e d by t i t r a t i o n o f w h o l e b r o t h . 2:1 m o l a r r a t i o o f p r o p i o n i c : a c e t i c a c i d s . A i r w a s t u r n e d o n a t 119 h o u r s a t a f l o w o f 1 : 3 v o l . a i r t o v o l . o f f e r m e n t a t i o n b r o t h w i t h a g i t a t o r a t 3 . 3 3 h z . I n i t i a l y e a s t e x t r a c t c o n c e n t r a t i o n : 80 g m / l a s r e c e i v e d , 59 g m / 1 d r y b a s i s . I n o c u l u m c o n c e n t r a t i o n : 1 : 5 0 . 57 Table 16 7-LITRE FERMENTATION - SSL RUN NO. 3 TIME (hr) DRY CELL WEIGHT (gm/l) ACIDS (gm/l) REDUCING SUGARS (gm/l) pH VITAMIN B.i 2 (mg/l) 0 0.062 0.0 23.7 7.0 5 0.109 23.8 6.86 18 0.216 23.6 6.77 25 0.266 6.65 42 0.678 1.63 22.2 6.4-7.0 66 1.680 2.97 6.5-7.0 0.13 ± 0.025 90 2.578 17.5 143 3.370 4.58 15.8 6.49-7.0 0.295 ±0.055 162 3.618 7.25 186 3.425 7.16 210 3.455 7.03 0.935 ± 0.635 260 3.947 13.2 6.90 330 4.487 12.8 1 .155 ± 0.16 A c i d s d e t e rm i n e d by t i t r a t i o n o f who 1 e b r o t h . A i r w a s t u r n e d o n a t 144 h o u r s a t r a t e o f 0 . 2 5 v o l / v o l w i t h a g i t a t o r a t 3 . 3 3 h z . I n i t i a l y e a s t e x t r a c t c o n c e n t r a t i o n : 102 g m / l a s r e c e i v e d , 7 5 g m / l d r y b a s i s . I n o c u l u m c o n c e n t r a t i o n : 1 : 5 0 . 58 Table 17 7-LITRE FERMENTATION - SSL RUN NO. 4 TIME (hr) DRY CELL WEIGHT (gm/l) ACIDS (gm/l) pH VITAMIN B 1 2 ( m g / l ) 0 0.088 0.0 7.0 24 0.383 0.76 6.7-7.0 48 1 .392 2.54 6.45-7.0 72 2.758 4.79 6.37-7.0 96 3.375 0.533 ± 0.055 120 4.557 6.44 6.55-7.15 144 5.071 8.40 6.6-7.0 168 5.946 192 6.018 216 5.479 1 .945 ± 0.233 240 5.493 2.183 ± 0.244 A c i d s d e t e r m i n e d by t i t r a t i o n o f w h o l e b r o t h w h e n a d j u s t i n g p H . I n i t i a l r e d u c i n g s u g a r c o n c e n t r a t i o n : 3 3 . 6 g m / l . I n i t i a l y e a s t e x t r a c t c o n c e n t r a t i o n : 100 g m / l a s r e c e i v e d , 7 3 . 8 g m / l d r y b a s i s . I n o c u l u m c o n c e n t r a t i o n : 1 : 5 0 . A i r w a s t u r n e d o n a t 96 h o u r s a t a f l o w o f 1 /4 v o l . a i r p e r v o l . o f m e d i u m w i t h a g i t a t o r a t 2 0 0 r p m . 5 9 TIME ( hours ) Figure 7. 7-Litre Fermentor SSL Run No. 1. TIME (hours) Figure 8. 7-Litre Fermentor SSL Run No. 2. C E L L S < g m / l ) ; B, 2 (mg / l ) ; SUGAR ( g m / d l ) ; pH C E L L S (gm/l) , B , 2 ( m g / l ) , p H 5 10 ACIDS (gm/ l ) 29 63 preliminary estimates of the vitamin concentrations and more sample r e p l i c a t e s , which resulted in much narrower confidence lim i t s on the test r e s u l t s . Run 3 cannot be treated as a normal run. Just prior to inoculation i t was noticed that the rubber sealing ring on the fermentor head was missing. This allowed a i r to enter the fermentor jar and was a possible source of contamination. Rather than throwing the medium out, i t was r e - s t e r i l i z e d . As fate would have i t , the yields of both bacterial c e l l s and vitamin B i 2 for this run were at y p i c a l . It could have been that the excessive heating of the medium caused these low yields or perhaps something else was wrong. No further work was done to check this out. The further studies on the kinetics and yields obtained in these runs w i l l be discussed in Chapter 4 Jand 5. 3.5 Search for a Replacement for Yeast Extract The amount of yeast extract required for good growth of P. freudenreichii in spent s u l f i t e l iquor, as we shall see in Sections 4.2 and 4.5, is excessive. The following experiments were devised in an attempt to see i f any of the common microbiological growth promoting substances would be better for the growth of P. freudenreichii than yeast extract. The growth promoting substances which were investigated were: 64 1. P a s t e y e a s t e x t r a c t - O X O I D 2. P o w d e r y e a s t e x t r a c t - N B C * 3. R e i n f o r c e d C l o s t r i d i a l M e d i u m - O X O I D 4. B e e f e x t r a c t - D I F C O - B A C T O 5. P e p t o n e - D I F C O - B A C T O 6. B 1 2 I n o c u l u m B r o t h USP - D I F C O - B A C T O These were added at concentrations of 20 and 50 gm/l as received to spent s u l f i t e liquor prepared in the usual way (see Section 2.1). The experiments were done in 250 ml. Erlenmeyer f l a s k s ; the amount of fermentation broth in the flasks was 150 ml. It should be noted that during this experiment, the temperature control on the incubator was ex-tremely e r r a t i c because the refrigerant had leaked out, and as a result the temperature varied from 30°C to 35°C. There-fore the results should not be compared too c r i t i c a l l y to the other r e s u l t s . However, these results can be taken as consistent within themselves. The ultimate growth of bacteria on each nutrient is tabulated in Table 18. NBC - - N u t r i t i o n a l B i o c h e m i c a l s C o r p o r a t i o n , C l e v e l a n d , O h i o , U . S . A . N o t e t h a t t h i s i s n o t t h e s a m e y e a s t e x t r a c t a s was u s e d e a r l i e r i n S e c t i o n s 3 .1 a n d 3.4. 65 The result of the experiment was disappointing. It was believed that some of the meat preparations, beef extract or peptone, would produce much higher c e l l growths than the yeast extracts. The reason for this b e l i e f i s that the glycine concentration in these nutrients is approximately 2% which is considerably higher than the glycine concentra-tion of yeast extract which contains approximately 0.2%. Furthermore, i t was believed that glycine is essential for the growth of P. freudenreichii for the following reasons: 1. L i m [ 1 2 ] f i n d s that 0 . 0 5 $ t o 0 . 4 $ o f g l y c i n e a d d e d t o n o r m a l m e d i a f o r p r o d u c t i o n o f P . f r e u d e n r e i c h i i i m p r o v e s v i t a m i n B i 2 y i e l d s by a s m u c h a s 100% f r o m 13 m g / l t o 23 m g / I . 2 . P r e " v o t C l O ] s t a t e s , " D i v e r s a m i n o a c i d e s i s o l e s s o n t a c t i f s q u a n d i l s s o n t a u t o c l a v e s a v e c l e m e l a n g e f I u c o s e -p h o s p h a t e e t p a r m i e u x , c ' e s t l a g l y c i n e q u i a I ' e f f e t l e p l u s m a r q u e . " 3 . G l y c i n e i s b e l i e v e d [ 1 4 , 1 6 , 1 9 3 t o be r e q u i r e d i n t h e f i r s t s t e p i n v i t a m i n B i 2 s y n t h e s i s . However, the experiment stands as a b r i l l i a n t f a i l u r e because there was, as can be seen in Table 18, very l i t t l e difference in the growth promoting a b i l i t y of these supplementry nutrients. 66 Table 18 GROWTH TRIALS USING VARIOUS NUTRIENTS NUTRIENT CONCENTRATION (gm/l - as received) DRY CELL WEIGHT (gm/l) Paste yeast extract OXOID 20.0 50.0 (14.76 (36.9 dry) dry) 1 .472 2.201 Powder yeast extract NBC 20.0 50.0 (20.0 (50.0 dry) dry) 1.728 3.093 C l o s t r i d i a l medium OXOID 20.0 50.0 1 .155 2.121 Beef extract (paste) DIFCO-BACTO 20.0 50.0 1 .134 2.097 Peptone (powder) DIFCO-BACTO 20.0 50.0 1 .078 1 .530 B i 2 inoculum medium DIFCO-BACTO 20.0 50.0 1 .167 2.1.91 I n i t i a l r e d u c i n g s u g a r c o n c e n t r a t i o n : 27 g m / l - R a y o n i e r S S L . T h i s e x p e r i m e n t w a s d o n e w i t h 150 m l . m e d i u m i n 2 5 0 m l . E r l e n m e y e r f l a s k s . I n o c u l u m c o n c e n t r a t i o n : 1 : 5 0 . D r y c e l l w e i g h t was m e a s u r e d a f t e r a p p r o x i m a t e l y 150 h o u r s . 67 3.6 Inhibitory Effect of Propionate and Hydrogen Ions What eff e c t does the accumulation of the products of Propionibacteria metabolism have on the growth of the bacteria? The metabolic products of Propionibacteria growth are hydrogen ion, propionate and acetate. The accumulation of these products can affect the growth of the bacteria in two ways. The f i n a l y i e l d of bacterial mass can be depressed, and the rate at which the bacteria are produced can be slowed down without decreasing the ultimate y i e l d of bacteria. Neronova et a l . [43] have found that between the l i m i t s of pH of 5.5 to 7.5, there is no s i g n i f i c a n t variation in the ultimate y i e l d of bacterial c e l l mass for Propioni-bacteria species. Moreover, these workers indicate that the range for optimum vitamin B i 2 synthesis is narrower, i . e . pH between 6.25 and 7.5. These workers did not indicate whether or not there was any rate i n h i b i t i o n caused by low (or high) pH within the range of 5.5 to 7.5. In order to check out the optimum pH range of 5.5 to 7.5 the following experiment was run. A mixture of various amounts of propionic acid and acetic acid in the molar ratio of 2:1::propionate : acetate was added to the standard f e r -mentation recipe of Section 2.1. The amounts of acid added are given in Table 19. The pH was not adjusted. After inoculation and 140 hours of incubation at 30°C, the 68 fermenting media were tested for t u r b i d i t y (dry c e l l weight) and pH. These results are reported in Table 19. Apparently, at an i n i t i a l acid concentration of between 24 and 48 mole/ l i t r e and an i n i t i a l pH of 4.9, growth of Propionibacteria w i l l not occur. Therefore, care was taken with the work of this thesis to maintain the pH of the fermenting media above 5.5, or preferably above 6.5. In the pH range of 5.5 to 7.5, no rate i n h i b i t i o n was noticed but the level of sophistica-tion of the experiments was not s u f f i c i e n t l y high to detect any but the most obvious rate i n h i b i t i o n s . In a later paper, Neronova and co-workers [44] claim that propionate ion i t s e l f at a pH of 6.8 to 7.0 w i l l i n h i b i t the rate of Propionibacteria shermanii - a very close r e l a t i v e of P. f r e u d e n r e i c h i i . However, upon close scrutiny, the work of this paper appears to have been done at unreal-i s t i c a l l y low (ca. 0.1 gm/l) bacterial c e l l concentrations. Furthermore, i t appears as i f there is a misplaced decimal place in some of the more important r e s u l t s , thus voiding an otherwise excellent paper. Table 19 EFFECT OF pH INHIBITION ON P. FREUDENREICHII GROWTH INITIAL ACID CONC. (mmole.1 i t r e ) DRY CELL WEIGHT (gm/l) AFTER pH AFTER ACID PRODUCED* (mmole/1) TOTAL ACID (mmole/1) 140 h r . 140 h r . 0 2.08 4.9 47.75 46. 24 0.74 4.95 14.06 38. 48 0.1 7 + 4.7 - 48. 96 0.16+ 4.45 - 96. C a 1 c u 1 a t e d f rom y i a 1d c o e f f i c i e n t Y , d e t e r m i n e d i n S e c t i o n a / x 4 . 3 . C = 1 . 6 5 2 a 6 9 . 3 x C o n s i d e r e d a s n o g r o w t h . cn co Chapter 4 YIELD COEFFICIENTS AND COST ANALYSIS OF NUTRIENTS AND PRODUCTS We have produced v o l a t i l e acids, bacterial biomass, and vitamin B i 2 from spent s u l f i t e liquor and yeast extract. What are the relationships between the substrate usages and product yields? Let us assume that each y i e l d c o e f f i c -ient is constant for a l l the conditions t r i e d in this work. Now we can define the following y i e l d c o e f f i c i e n t s : A C x Y X ^ S = - ^ Q— = c e l l s produced/sugar used ( 4 . 1 ) AC Y X ^ N = - = c e l l s produced/yeast extract used (4.2) A C a Y , = TF~ ~ r a t i o of total acids produced/ (4.3) ' x c e l l s produced AC Y . = -r-rX = ra t i o of vitamin B i 2 produced/ (4.4) ' x c e l l s produced 70 71 from which the following y i e l d relationships can be derived: Ya/s - " - Ya/x * Yx/s ( 4 - 5 ) and Ya/n = " ^ = Y a / x x Yx/n ( 4 - 6 ) AC Y = - — = Y x Y ( 4 7 ) Yv/s AC Yv/x Yx/s K q ' n A C v Y . = - -TTA = Y , x Y . (4.8) v/n AC n v/x x/n AC Ys/n " AC;= Yx/n / Yx/s ( 4:' 9 ) 72 In summary these y i e l d c o e f f i c i e n t s and ratios were found to be: Y . = 0 . 3 8 3 gm d r y c e I Is p r o d u c e d / g m s u g a r a c t u a l l y u s e d o r 0 . 2 4 8 gm d r y e e l Is p r o d u c e d / g m S S L s u g a r s . Y - n = 0 . 0 7 9 gm d r y c e l l s p r o d u c e d / g m d r y y e a s t e x t r a c t c h a r g e d . Y . = 1 . 6 5 2 gm a c i d s p r o d u c e d / g m c e l l s p r o d u c e d . or assuming a 2/1, propionate/acetate molar r a t i o Yacetate/x = ° - 4 7 6 9 m a c e t a t e / g m e e l Is v propionate/x = 1 . 1 7 6 gm p r o p i o n a t e / g m c e l l s . Y y. = 0 . 3 6 mg v i t a m i n B i 2 / g m e e l Is p r o d u c e d a n d t h e d e r i v e d y i e l d s . Y . = 0 . 6 3 2 gm a c i d s / g m s u g a r a c t u a l l y u s e d . T h i s c o m p a r e s f a v o u r a b l y w i t h t h e 0 . 7 5 g m / g m w h i c h i s e x p e c t e d f r o m e q u a t i o n ( 1 . 4 ) o f S e c t i o n 1 . 3 . T h e l o w e r v a l u e o f Y / o b t a i n e d c a n be r a t i o n a l i z e d by a / s a s s u m i n g t h e d i f f e r e n c e w e n t t o b u i l d -i n g c e l l u l a r m a t e r i a l ( b a c t e r i a ) . 73 Y g y n = 0 . 1 3 gm a c i d s / g m d r y y e a s t e x t r a c t Y - s = 0 . 1 3 8 mg v i t a m i n B i 2 / g m s u g a r a c t u a I Iy u s e d . Y . = 0 . 0 2 8 5 mg v i t a m i n B i 2 / g m d r y y e a s t e x t r a c t . Yg . = 0 . 2 0 6 gm s u g a r u s e d / g m d r y y e a s t e x t r a c t r e q u i r e m e n t s . The methods and data used to calculate these quan-t i t i e s are detailed in the following sections. The economic implications of the yi e l d s found are also discussed. and 4.1 Y x ^ 3 - Y i e l d of Dry Cells Based Upon Sugar Used The y i e l d c o e f f i c i e n t (Y . ) can be determined from a l l the sugar/dry c e l l data from a l l the runs which for convenience have been collected and presented as Table 20. The slopes of the sugar vs. dry c e l l weight plots for a l l these runs have been calculated and these slopes which are equal to Y x^ s are presented in Table 21 and are shown graphically in Figure 11. The slopes of the runs are close enough to each other that they can be combined into a series 74 Table 20 CELL WEIGHT VS. SUGAR-DATA FROM ALL RUNS RUN IDENTIFICATION TIME (hr) DRY CELL WEIGHT (gm/l) SUGAR (gm/dl) CHANGE IN SUGAR CHANGE IN CELLS S e c t i o n 3 . 1 T a b l e 8 I n i t i a l s u g a r = 37 g m / l 0 44 89 134 180 253 0.0268 0.3861 1 .087 1.240 1 .346 1.546 3.70 3.83 3.65 3.65 3.50 3.40 0.103 0.203 0.023 0.023 -0.127 -0.227 -0.921 -0.553 0.148 0.301 0.407 0.607 S e c t i o n 3 . 1 T a b l e 8 I n i t i a l s u g a r = 27 g m / l 0 44 89 134 180 253 0.025 0.970 2.710 2.467 2.268 2.417 2.70 2.60 2.00 2.00 2.05 2.05 0.467 0.367 -0.233 -0.233 -0.183 -0.183 -1 .783 -0.840 0.900 0.657 0.458 0.607 S e c t i o n 3 . 1 T a b 1e 8 I n i t i a l s u g a r = 14 g m / l 0 44 89 134 180 253 0.0285 1 .436 2.668 3.209 3.600 3.669 1 .40 1 .05 0.93 0.65 0.70 0.60 0.517 0.167 0.047 -0.233 -0.183 -0.283 -2.407 -0.999 0.233 0.774 1 .165 1 .234 S e c t i o n 3 . 3 T a b l e 12 S S L 0 97 193 0 97 193 0.02 0.961 1 .623 0.02 0.993 1.65 2.00 1 .90 1 .56 1 .93 1 .87 1.77 0.162 0.062 -0.278 0.092 0.023 -0.068 -0.858 0.083 0.745 -0.858 0.115 0.772 CONTINUED 75 Table 20 (Continued) RUN IDENTIFICATION TIME (hr) DRY CELL WEIGHT (gm/i) SUGAR (gm/dl) CHANGE IN SUGAR CHANGE : IN CELLS S e c t i o n 3 . 3 T a b l e 12 G 1 u c o s e 0 97 193 0 97 193 0.02 2.406 2.62 0.02 2.473 1 .970 2.05 1 .53 1 .43 2.11 1 .54 1 .46 0.365 -0.155 -0.265 0.425 -0.145 -0.225 -1.56 5 0.821 1 .035 -1 .56 5 0.888 - .385 S e c t i o n 3 . 3 T a b l e 12 Ma n n o s e 0 97 193 0 97 193 0.02 1 .941 2.231 0.02 1 .977 2.077 2.06 1 .76 1 .59 2.08 1 .76 1 .59 0.26 -0.08 -0.21 0.28 -0.04 -0.21 -1.358 0. 563 0.853 -1.358 0.599 0.700 S e c t i o n 3.. 3 T a b l e 12 Ga 1 a c t o s e 0 97 193 0 97 193 0.02 1 .703 1 .977 0.02 1 .688 2.104 1 .96 1 . 57 1 .38 1 .96 1 .43 1 .32 0.357 -0.033 -0.223 0.357 -0.173 -0.283 -1.232 0.451 0.725 -1.232 0.436 0.852 S e c t i o n 3 . 4 T a b l e 14 0 1 8 0.0479 0.0633 0.0779 2.06 1 .80 1 .822 0.951 0.691 0.713 -3.108 -3.093 -3.078 7 - l i t r e f e r m e n t o r S S L R u n N o . 1 27 32i 36 0.4139 0.5838 0.7805 1 .895 1 .952 1 .938 0.786 0.843 0.829 -2.742 -2.573 -2.376 46 558 70 1 .644 2.44 3.152 1 .78 1 .49 1 .286 0.671 0.381 0.177 -1 .512 -0.7124 -0.0042 79 96 101 3.912 4.811 4.8133 1 .048 0.932 0.759 -0.061 -0.177 -0.350 0.755 1 .655 1 .657 CONTINUED 76 Table 20 (Continued) TIME (hr) DRY CELL SUGAR (gm/dl) CHANGE CHANGE RUN IDENTIFICATION WEIGHT (gm/l) IN SUGAR IN CELLS 120 4.849 0.6435 -0.466 1 .693 127 4.552 0.682 -0.427 1 .395 143 4.515 0.720 -0.389 1 .359 166 4.535 0.648 -0.461 1 .378 170 4.955 0.643 -0.466 1 .798 190± 5.007 0.665 -0.444 1 .850 215 5.824 0.687 -0.422 2.668 238 6.147 0.682 -0.427 2.991 S e c t i o n 3 . 4 T a b l e 15 7 - l i t r e f e r m e n t o r 1 0.0710 3.81 5 1 .102 -3.171 51i 96 1 .411 3.637 3.125 2.765 0.412 0.0516 -1.8312 0.3947 S S L R u n N o . 2 1 20 4.750 2.164 -0.5493 1 .508 166i 6.342 1 .697 -1.0159 3.010 0 0.0621 2.42 0.463 -1.869 5 0.1086 2.44 0.483 -1.822 S e c t i o n 3 . 4 18 0.2159 2.40 0.443 -1 .715 T a b l e 16 7 - l i t r e f e r m e n t o r S S L R u n N o . 3 42 90 0.6781 2.578 2.29 1 .83 0.333 -0.127 -1.253 0.672 143 3.37 1 .61 -0.347 1 .439 260 3.947 1 .35 -0.607 2.016 330 4.487 1 .32 -0.637 2.556 77 Table 21 YIELD COEFFICIENT BASED ON SUGAR USAGE RUN AVERAGE OF DRY CELL WEIGHT RESULTS (gm/l) AVERAGE OF SUGAR RESULTS (gm/dl) YIELD Yx/s COEFFICIENT* (gm/l/gm/dl) T a b 1 e 8 3.156 1 .109 0.253 T a b l e 8 2.713 3.242 0.322 T a b l e 8 1 .958 1.931 0.260 T a b l e 1 2 3.627 0.939 0.216 T a b 1 e 1 2 2.233 1 .810 0.298 T a b l e 1 2 0.883 2.435 0.208 T a b 1 e 1 2 1 .838 0.878 0.177 T a b l e 1 4 1 .685 1 .585 0.245 T a b 1 e 1 5 1 .800 1 .378 0.202 T a b l e 1 6 1 .603 1 .252 0.293 F r o m s l o p e o f l i n e ( s ) o n F i g u r e I I . Average o v e r a l l slope = 0.26057 from Figure 12. Theref o r e y i e l d c o e f f i c i e n t Y . = 2.606 gm sugar/gm c e l l s S / X produced or a l t e r n a t e l y Y . = O . 3 8 3 gm cell/gm C"6 sugar used. 78 Figure 11. Reducing Sugar Remaining vs. Dry Cell Weight for A l l Runs. 79 of p a r a l l e l curves. This combined curve i s shown in Figure 12. In order to be able to condense these 10 lines of similar slopes, but of widely varing intercepts onto a single l i n e , the points for this graph were calculated by subtracting the average dry c e l l concentrations and the average sugar concentration from the individual dry c e l l and sugar concentrations, thus forcing a l l the points to pass through the o r i g i n . The overall slope of this l i n e is v x ^ s = 0.383 gm c e l l s produced/gm sugar used. However, of the sugars in spent s u l f i t e l iquor, only the hexoses are used by P. f r e u d e n r e i c h i i . What then is the r a t i o of fermentable sugars to the total reducing sugars in spent liquor? We have two estimates of this r a t i o . 1. In the test to see which of the wood sugars which w i l l support the growth of P. freudenreichii reported in Section 4.3, the c e l l u l a r y i e l d when SSL was used is 68% of the y i e l d when the hexoses were used, suitable correc-tions having been made for the bacterial growth on yeast extract with no sugar. 2. From the 7 - l i t r e fermentation SSL Run No. 1, reported in Table 14 in Section 3.4, in which carbohydrate was the l i m i t i n g nutrient, the f r a c t i o n of the o r i g i n a l sugar charged that was used up by the bacteria was found to be 0.65. Let us therefore assume that the f r a c t i o n of fermentable reducing sugars in this p a r t i c u l a r spent s u l f i t e liquor is 0.65. CHANGE IN DRY C E L L WEIGHT ( g m / l ) Figure 12. Sugar Used vs. Change in Dry Cell Weight for A l l Runs Combined. 81 These two y i e l d c o e f f i c i e n t s can therefore be calculated: Y x ^ s = 0 . 3 8 3 gm e e l Is p r o d u c e d / g m s u g a r s u s e d and Y , = 0 . 2 4 8 gm c e l l s p r o d u c e d / g m t o t a l S S L r e d u c i n g s u g a r s c h a r g e d . 4.2 Y x ^ n - Y i e l d of Dry Cells Based Upon Yeast Extract Unfortunately the y i e l d c o e f f i c i e n t for yeast extract is not as well defined as the sugar y i e l d . The primary reason for this i s that yeast extract is a complex substrate and i t is not known what the actual l i m i t i n g nutrient is that the yeast extract is supplying. Therefore, the time course of the l i m i t i n g nutrient during the fermentation could not be determined. It was, therefore, necessary to resort to i n -direc t methods to determine Y x^ n > Table 22 shows a re-compilation of the runs made where yeast extract was made the li m i t i n g nutrient by insuring that the other nutrients such as sugar were in excess. To 82 Table 22 DRY CELL WEIGHT VS. YEAST EXTRACT CHARGED FOR YEAST EXTRACT LIMITING RUNS RUN IDENTIFICATION INITIAL YEAST (gm/l) EXTRACT ULTIMATE DRY CELL WEIGHT (gm/l) a s r e c e i v e d d r y b a s i s 12 .5 12 5 1. 382 S e c t i o n 3. 1 18 .25 18 25 2. 045 T a b l e 8 25 .0 25 0 2. 475 37 5 37 5 3. 306 0 0 0 0 0. 111 S e c t i o n T a b l e 9 3. 1 9 21 57 53 9 21 57 53 0. 2. 473 07 32 27 32 27 2. 774 32 27 32 27 2. 759 0 0 0 0 0. 2602 20 0 14 7 1 . 002 S e c t i o n T a b l e IC 3. ) 1 30 40 40 0 0 0 21 29 29 1 5 5 1 . 2. 2. 567 261 254 45 0 33 1 2. 426 50 0 36 8 2. 517 S e c t i o n 3. 4,Tab . 1 4 72 0 53 0 5. 142 T a b . 1 5 80 0 59 0 6. 72 T a b . 1 6 100 2 74 0 3. 73 T a b . 1 7 100 0 73 8 5. 29 CONTINUED 83 Table 22 (Continued) INITIAL YEAST EXTRACT ULTIMATE DRY RUN IDENTIFICATION CELL WEIGHT (gm/l) (gm/l) as received d ry hasi s Sectio n 3.5, Paste 20.0 14.7 1 .472 Table 18 50.0 36.8 2.201 S e c t i o n 3.5, Powder 20.0 20.0 1 .728 Table 18 50.0 50.0 3.093 84 avoid confusion between powder and paste yeast extract, yeast extract concentrations and yields are a l l expressed as gm of dry yeast e x t r a c t / l i t r e . Figure 13 shows the data points of Table 22 plotted. The paste and powder yeast extract runs were analysed both separately and pooled in order to check for any s t a t i s t i c a l differences between them. The detail s of this s t a t i s t i c a l analysis are tabulated in Table 23. As can be seen from this table, the increase in the sum of the squares of the errors when the paste and powder yeast extract data are combined is not s t a t i s t i c a l l y s i g n i f i c a n t . That is S S ^ ( c o m b i n e d ) - ( S S ^ C p a s t e a l o n e ) + S S ^ ( p o w d e r a l o n e ) ) /df S S ^ C p a s t e a l o n e ) + S S ^ C p o w d e r a l o n e ) j / d f 0 522/2 7*402/19 = 0.671 w h i c h i s n o t s i g n i f i c a n t a t t h e 90% c o n f i d e n c e l e v e l Therefore the y i e l d c o e f f i c i e n t s , from the slope of the curve in Figure 13, are: Cn" INITIAL YEAST EXTRACT (gm/l dry basis) S Figure 13. Dry Cell Weight vs. I n i t i a l Yeast Extract for A l l Yeast Extract Limiting Runs. Table 23 STATISTICAL ANALYSIS OF Y x / n FOR POWDER AND PASTE YEAST EXTRACT SOURCE D of F SS MS F SIG. POWDER REGRESSION 1 9.196 9.196 56.54 A L O N E ERROR 9 1 .464 0.1626 **** TOTAL 10 10.660 P A S T E REGRESSION 1 35.228 35.228 A L O N E ERROR 10 5.938 0.954 59.33 **** TOTAL 11 41.166 SUM OF POWDER AND P A S T E ERROR 19 7.402 0.3900 POWDER & P A S T E TOG REGRESSION 1 47.024 T O G E T H E R ERROR 21 7.924 0.377 125.0 ** ** D 1 F F E R E N C E 2 0.522 0.261 0.671 N.S. B e c a u s e o f t h e i n c r e a s e i n sum o f s q u a r e s c o m p u t e d by l u m p i n g t h e p a s t e a n d p o w d e r y e a s t e x t r a c t s t o g e t h e r o v e r t h e sum o f s q u a r e s c o m p u t e d by c a l c u l a t i n g s e p a r a t e s u m s o f s q u a r e s f o r e a c h e x t r a c t i s n o t s t a t i s t i c a l l y s i g n i f i c a n t a t t h e 9 0 $ l e v e l , we a r e p e r f e c t l y j u s t i f i e d i n p o o l i n g t h e m . N o t e t h a t f o r t h i s a n a l y s i s t h e w e t p a s t e y e a s t e x t r a c t c o n c e n t r a t i o n s m u s t be c a l c u l a t e d a s d r y b a s i s . 87 0.0684 gm eelIs/gm dry powder yeast e x t r a c t charged, 0.0,83 gm eel Is/gm dry yeast e x t r a c t in paste changed, 0.079 gm eelIs/gm dry yeast e x t r a c t charged or f o r paste yeast e x t r a c t as r e c e i v e d ; 0.0581 gm eelIs/gm wet paste yeast e x t r a c t charged. 4.3 Y, / w Determination of Acids Yield Coefficients a / x The data used to determine the ratio of combined propionic and acetic acids produced to the dry c e l l u l a r matter produced are presented in Table 24. Two methods for measuring the amount of acids produced were used as explained in Section 2.9. The acids concentrations were converted from mmoles/1 to gm/l by assuming that the molar r a t i o of propionic acid produced to the acetic acid produced was the same as the theoretical [14,18,19] 2:1, as discussed in Section 1.3. An average acids 'molecular weight 1 of 69.33 was used for the conversion from mmoles/litre to gm/l. The slopes of the acid vs. c e l l s curves for each run were determined by the method of least squares with the Powder a I one: ^ x / n Paste alone : ^ x / n Comb i ned : ^ x/n v lx/n Table 24 YIELD OF VOLATILE ACIDS BASED ON DRY CELL WEIGHT RUN IDENTIFICATION TIME (hr) DRY CELL WEIGHT (gm/l) VOLATILE ACIDS (mmole/1) ACIDS TEST 118i 4.121 95.5 143i 5.71 121 .0 166* 6.342 143.9 by adjusting 190i 6.743 159.2 pH 217 6.757 171.2 244 6.856 173.4 24 0.3831 11.0 48 1 .392 36.7 72 2.758 69.05 - by adj. pH 120 4.557 92.86 144 5.071 121 .14 J 42 0.678 23.5 1 66 1.680 42.9 by adj. pH 143 3.37 66.1 J 0 0.0 0.0 1 32i 0.5838 36.51 - • by adj. pH 58 2.444 56.99 J 70 3.152 84.55 ' by actual 101 4.813 109.16 measurement 120 4.849 125.58 , 127 4.552 133.92 by adj. pH 238 6.147 130.41 ' 96 4.811 119.0 46 1 .644 43.5 J S e c t i o n 3 . 4 T a b l e 15 S S L R u n N o . 2 S e c t i o n 3 . 4 T a b l e 17 S S L R u n N o . 4 S e c t i o n 3 . 4 T a b l e 16 S S L R u n N o . 3 S e c t i o n 3 . 4 T a b l e 14 S S L R u n N o . I 89 l i n e forced to pass through the o r i g i n . These results are presented in Table 25. The slopes of these lines are close enough that the overall slope Y . = 1.652 gm acids/gm c e l l s a / x can be used. This slope with the raw data points i s pre-sented as Figure 14. Using the 2:1 propionate:acetate molar r a t i o , the individual y i e l d c o e f f i c i e n t s for propionic acid and acetic acid can be determined as: Y a c e t i c / x = ° - 4 7 6 9™ a c e t i c a c i d / g m e e l I s , and Y „ M „ „ , - ^ / „ = 1 . 1 7 6 gm p r o p i o n i c a c i d / g m c e l l s . propiOniC/X . a r r .a 4.4 Y v / X - Y i e l d of Vitamin B 1 2 The awful spread of the vitamin B i 2 data makes the evaluation of the vitamin y i e l d c o e f f i c i e n t Y y ^ x more d i f f i -c u l t . Table 26 summarizes the time-dry c e l l weight-vitamin B 1 2 data from a l l runs. These curves are shown in Figure 15. Nevertheless, the vitamin B i 2 values which were assumed to be steady state values (see Table 26) were averaged, weighting each value by the reciprocal of i t s variance as suggested by Table 25 ACID YIELD COEFFICIENT RUN IDENTIFICATION Ya/x SSL Run No. 2 1 .650 SSL Run No. 4 1 .584 SSL Run No. 3 1 .472 SSL Run No. 1 1.7025 ALL RUNS POOLED 1 .652 T h e y i e l d c o e f f i c i e n t Y g ^ x i s i n t h e u n i t s o f gm a c i d s p e r gm o f b a c t e r i a ; i t i s a s s u m e d t h a t t h e a c i d s a r e p r o d u c e d i n t h e m o l a r r a t i o o f 2:1 p r o p i o n i c : a c e t i c . 91 2.0 4 0 6.0 DRY CELL WEIGHT (gm/l) Figure 14. Acids Produced vs. Dry Cell Weight for A l l Runs Combined. 92 Table 26 VITAMIN B i 2 ANALYSES RUN IDENTIFICATION TIME (hr) DRY CELL WEIGHT (gm/i) VITAMIN B 1 2 (mg/l) 95% CONFIDENCE LIMITS S e c t i o n 3 . 4 T a b l e 14 S S L R u n N o . 1 0 96 143 166 238 0.0479 4.811 4.518 4.535 6.147 0.0411 1 .5,19 2.141 2.886* 3.014 0.0320 0.197 0.973 1 .928 3.74 S e c t i o n 3 . 4 T a b l e 15 S S L R u n N o . 2 1 51* 96 166 285i 0.0728 1 .411 3.637 6.342 6.824 0.317 0.354 2.28* 2.19* 0.033 0.074 0.40 0.33 S e c t i o n 3 . 4 T a b l e 16 S S L R u n N o . 3 66 143 210 330 1 .680 3.370 3.455 4.487 0.13 0.295 0.935 1 .155 0.025 0.055 0.635 0.16 S e c t i o n 3 . 4 T a b l e 17 S S L R u n N o . 4 96 216 240 3.375 5.479 5.493 0.533. 1 .945 2.183* 0.055 0.233 0.244 T h e d a t a t h u s m a r k e d a r e a s s u m e d t o be u l t i m a t e v a l u e s o f t h e v i t a m i n B j 2 f o r t h a t p a r t i c u l a r r u n . T h e s e d a t a w e r e u s e d t o c a l c u l a t e t h e a v e r a g e v i t a m i n B i 2 y i e l d , w h i c h i s 2 . 1 2 m g / l ± 0 . 3 6 . The average vitamin B i 2 y i e l d c o e f f i c i e n t Yv/X w a s c a l c u l a t e d from data above by weighing each point by the inverse of the 35% confidence l i m i t . The y i e l d c o e f f i c i e n t thus determined is Y y ^ x = 0.36 mg of vitamin per gm of dry c e l l u l a r m a t e r i a l . E Symbol O • A V Section S S L run 3-4 3-4 3 4 3 4 I 2 3 4 CJ CQ (M < S S L I 100 2 0 0 TI M E ( hours ) 3 0 0 Figure 15. Vitamin B 1 2 Produced vs. Time for A l l Runs to C O 94 Himmelblau [72]. Thus C~v, the weighted average of the vitamin B 1 2 concentrations designated in Table 26 as ultimate concen-tr a t i o n of vitamin, was calculated to be 2.12 mg/l ± 0.36. The wierd SSL Run No. 3 was omitted, from this calculation since this run is believed to be an anomaly for reasons dis-cussed in Section 3.4. These C~v values were compared to the ultimate c e l l concentration values for each run and Y , V / X was calculated. These y i e l d c o e f f i c i e n t s are tabulated in Table 27. The average y i e l d c o e f f i c i e n t , omitting the anomalous resu l t of SSL Run No. 3, was found to be Y , = 0.36 V / A mg vitamin Bi2/gm dry c e l l . 4.5 Nutrient Requirements and Costs for Growth of  P. freudenreichii These y i e l d c o e f f i c i e n t s can be used to calculate the raw material requirements for vitamin B i 2 production. Table 28 shows the results of this c a l c u l a t i o n . Also these y i e l d c o e f f i c i e n t s can be used to estimate the improvement in the value of the fermentation broth produced by the f e r -mentation. Table 29 shows the results of such a calculation using prices for yeast extract, acetic acid, propionic acid and vitamin B i 2 from O i l , Paint and Drug Reporter, Sept. 29, 1969. Even without including the cost of capital for building 95 Table 27 VITAMIN B YIELD COEFFICIENT 12 RUN IDENTIFICATION DRY CELL WEIGHT (gm/l) VITAMIN Bi 2. (mg/l) YIELD COEFFICIENT (mg/gm) SSL Run No. 1 5.34 2.12 0.40 SSL Run No. 2 6.98 2.12 0.305 SSL Run No. 3 4.09 1.16 0.285 SSL Run No. 4 5.80 2.12 0.367 A v e r a g e y i e l d c o e f f i c i e n t = 0 . 3 6 mg v i t a m i n p r o d u c e d p e r gm c e l l u l a r m a t e r i a l p r o d u c e d . T h i s v a l u e was c a l c u l a t e d by d i v i d i n g t h e w e i g h t e d a v e r a g e v i t a m i n B12 c o n c e n t r a t i o n , C ° = 2 . 1 2 by t h e a v e r a g e d r y c e l l w e i g h t c o n c e n t r a t i o n s . T h u s Y , = 2 . 1 2 / 5 . 9 4 = 0 . 3 6 m g / g m . 96 Table 28 RAW M A T E R I A L REQUIREMENTS AND PRODUCT Y I E L D S FOR V I T A M I N B , 2 PRODUCTION B A S I S : I l b . o f V i t a m i n B 1 2 NUTRIENT OR PRODUCT CALCULATION NUTRIENT PRODUCT REQUIREMENT OR PRODUCTION Vitamin B i 2 - 1 Lb. C e l l u l a r product ^ v / x 2780 Lb. Acids (total) Ya/x x c e l l s 4580 Lb. Acetic Propionic 1320 3260 Lb. Lb. Sugar (SSL) c e l l s / Y x / s 11700 = 31300 Lb. USG of total sugar of SSL Yeast extract c e l l s / Y x / n 39700 Lb. Table 29 VALUE OF NUTRIENTS AND PRODUCTS IN FERMENTATION BROTH NUTRIENT OR PRODUCT CONCENTRATION OF NUTRIENT OR PRODUCT (gm/l) VALUE U/lb.) CONTRIBUTION TO BROTH VALUE U / l i t r e ) Yeast extract ( d r y ) 75 27 -4.45 Cells ( d r y ) 5.97 --Spent s u l f i t e liquor Total sugars Sugars used 500 ml 23.8 15.4 — Acetic acid 2.82 9 0 .056 Propionic acid 6.95 15 0.229 Vitamin B i 2 2.13 x 10"3 363,000 1 .69 NET IMPROVEMENT -2.48 N O T E : N e g a t i v e q u a n t i t i e s i n ' B R O T H V A L U E ' c o l u m n r e p r e s e n t e i t h e r a c o s t o r i n t h e ' N E T I M P R O V E M E N T ' row - a l o s s . 98 the fermentation plant and recovery processes, i t can be c l e a r l y seen that production of vitamin B 1 2 from spent s u l f i t e liquor is uneconomic. The basic reason for this unfavourable economic position is the excessive requirement for the r e l a t i v e l y expensive yeast extract. A second cause is the low y i e l d of vitamin B i 2 . What can be done to improve the economic outlook of the production of vitamin B i 2 from spent s u l f i t e liquor? Table 30 shows the net improvement in fermentation broth value with certain process improvements. These improvements, which are as yet untested, include replacement of yeast extract by less expensive raw materials such as f i s h packing plant scraps, or soy bean extract, or improvement of the vitamin y i e l d on yeast extract charged. However, l e t us take a look at this problem from a more distant perspective. What we are e s s e n t i a l l y doing is replacing r e l a t i v e l y expensive carbohydrate such as beet or cane sugar at 4<£/lb with a waste product, spent s u l f i t e l i quor, at a cost of Oct/lb. The cost of this substitution is larger fermentation and separation equipment because of lower sugar concentrations. A l l the other process variables, except perhaps pollution load which seems to be unimportant at present in our economic system, should be the same. The economic advantage of the substitution of sugar with spent 99 Table 30 IMPROVEMENT IN NET BROTH VALUE FOR SOME PROCESS CHANGES PROCESS CHANGE NET IMPROVEMENT IN BROTH VALUE ( * / l i t r e ) Base case (Table 29). -2.48 Replacement of SSL by cane sugar at 4tt/lb. -2.62 Replacement® of yeast extract with cheaper yeast product* at lOtt/lb. + 0.36 Replacement® of yeast extract with f i s h meal at 8.45<fr/l b+. + 0.58 Replacement® of yeast extract with soy bean meal at 4.2<t/lb#. + 1 .28 Improvement of y i e l d Y . by a factor of 2. x / n -0.25 Improvement of y i e l d Y , by a factor of 2. v / x -0.79 T Sum of a l l the good things above. +2.79 A s s u m i n g t h a t ^ x / n f ° r y e a s t e x t r a c t r e p l a c e m e n t s i s t h e s a m e a s t h a t f o u n d f o r t h e y e a s t e x t r a c t ( s ) u s e d h e r e . T h e r e i s n o r e a s o n t o a s s u m e ( o r n o t t o a s s u m e ) t h i s . D r i e d y e a s t s o l i d : 9 - 3 / 4 < £ / 1 b, C & E N , J u l y 2 9 , 1963 # O i l P a i n t a n d D r u g R e p o r t e r , S e p t . 2 9 , 1969 E d w a r d s L~493, U s i n q d o u b l e d Y . , f i s h m e a l a n d d o u b l e Y . , • x / n v . / x -100 s u l f i t e liquor is approximately 0.14<£/litre of fermentation broth based upon the figures of Table 30. This saving is an order of magnitude less than the 4.45<£/litre cost of the yeast extract (or a similar nutrient). This leads to the general conclusion which w i l l be elaborated in Appendix I I I , that a fermentation process u t i l i z i n g spent s u l f i t e liquor sugars should be one in which the primary raw material cost is the sugar or carbohydrate portion, not another more expensive, complex nutrient. A rough outline of three micro-b i o l o g i c a l products, c e l l u l a s e s , g i b b e r e l l i c acid and proteases, which i t might be possible to produce from s u l f i t e l i quor, is presented as Appendix IV. 4.6 Comparison of the Yields of this Work with the  Yields of Others Table 31 shows a comparison of the yi e l d s of pro-ducts determined in this work compared with the published work of others who have worked with Propionibacteria grown on other media than spent s u l f i t e liquor. Three things are immediately apparent from this table: I. T h e y i e l d s o f d i f f e r e n t w o r k e r s i n t h i s f i e l d a r e v e r y v a r i a b l e . T a b l e 31 COMPARISON OF Y I E L D S OF T H I S WORK WITH THAT OF OTHERS YIELD COEFFICIENTS REF. CARBON SOURCE AMINO ACID/ VITAMIN CELLS (gm/l) VITAMIN Biz (mg/l) ACIDS (gm/l) SUGAR USED (gm/l) Y x / s Y x / n Y a / x Y a / s v T v / x Y v / s Y "v/n 1 SSL P a s t e y e a s t e x t r a c t 100 gm/l 7.0 2.5 11.5 20.5 0.375 0.07 1 .65 0.62 0.36 0.122 0.025 2 SSL Powder y e a s t e x t r a c t 10 gm/l 1 .25 0.625 2.2 5.0 0.25 0.125 1 .76 0.44 0.50 0.125 0.062 3 G l u c o s e 4% 8.8% Corn s t e e p l i q u o r 2 1 g / 1 - A u t o l i zed p e n n i c i l i u m 13.4 Mycelium 31.5 gm/l 13.4 3.7 14.5 - 40.0 88.0 0.153 0.425 -- 1 .08 0.093 0.165 0.176 0.460 4 Cheese whey-( l a c t o s e ) 6% Y e a s t e x t r a c t 15 gm/l -8.43 20.0 - 60 52.8 -- - - - 0.14 0.37 0.56 1 .33 5 2% g l u c o s e Y e a s t e x t r a c t 32 gm/l Corn e x t r a c t 30 gm/l 3.8 5.32 9.5 4.0 - 20.0 20.0 0.19 0.266 0.119 0.1 78 - - 2.5 0.75 0.475 0.20 0.297 0.133 6 10% g l u c o s e Corn s t e e p l i q . 39.5 gm/l - 23.0 - 100.0 - - - - - 0.23 0.58 7 10% g l u c o s e Corn s t e e p l i q . 66 gm/l & DMBZ - 25.0 - 100.0 - - - - - 0.25 0.38 CONTINUED T a b l e 31 ( C o n t i n u e d ) YIELD COEFFICIENTS REF. CARBON SOURCE AMINO ACID/ VITAMIN CELLS (gm/l) VITAMIN B i 2 (mg/l) ACIDS (gm /D SUGAR USED (gm/l) Y x / s v "x/n Y a / x Y a / s Y v / x Y v / s Y v / n 8 12% b e e t - m o l -a s s e s i n v e r t a s e Y e a s t e x t r a c t 100 gm/l - 4.5 - 120.0 - - - - - 0.0375 0.045 9 SSL ( 2 % s u g a r ) Y e a s t e x t r a c t e x t r a c t - - 16.0 21.4 - - - 0.75 - - -10 Beet mol-a s s e s & i n v e r t a s e Brewers y e a s t 23.3 gm/l 100.0 gm/l 12.9 7.4 4.47 4.5 - 100.0 120.0 0.129 0.0615 0.551 0.074 -- 0.346 0.607 0.045 0.0375 0.191 0.045 R E F E R E N C E S : I . T h i s w o r k . 2. N i s h a kawa [ 2 3 ] . 3. R i l e y et al. [ 2 2 ] . 4. B u l l e r m a n and B e r r y [ 5 0 , 5 1 , 5 2 ] . 5. S h a p o s h n i k o v and N e r o n o v a [ 5 3 ] . 6. S p e e d i e and H u l l [ 5 4 ] . 7. Norn i ne [ 5 5 ] . 8. V o r o b ' e v a [ 4 8 ] . 9. M a r t i n [ 5 6 ] . D i d n o t l o o k f o r o r ~rry 1~ o p r o d UCB v i t a m i n B i 2 < 10. S u d a r s k y a n d F i s h e r [ 5 7 ] . USP 2 8 1 6 8 5 6 ( 1 9 5 7 ) . o ro 103 2 . T h e a m o u n t o f s u p p l e m e n t r y n u t r i e n t ( y e a s t e x t r a c t , c o r n s t e e p l i q u o r , o r a u t o l y z a t e o f P e n i c i I l i u m m y c e l i u m ) i s v e r y h i g h i n a l l t h e s e w o r k s , b u t i t i s p a r t i c u -l a r l y h i g h i n t h i s w o r k . 3 . T h e y i e l d s o f e e l l u l a r m a t e r i a l a n d v i t a m i n B i 2 b a s e d u p o n y e a s t e x t r a c t f o r t h i s w o r k a r e l o w e r t h a n a l I t h e o t h e r w o r k s r e p o r t e d i n T a b l e . 3 1 . Table 31, however, is not a complete table or even a representative table of a l l the work done in this f i e l d . Only those works which reported r e l a t i v e l y high yi e l d s of vitamin B i 2 are l i s t e d in Table 31. Others, notably Neronova [43,46,47] and Vorob'eva [48] have reported lower y i e l d s . The results of other workers are also reported in the reviews of Mervyn and Smith [58] and Noyes [13]. Chapter 5 KINETIC MODEL FOR PROPIONIBACTERIA GROWTH 5 .1 Development of Models Good kinetic models for batch fermentation, properly f i t t e d to the data are invaluable for use i n : (a) the design of continuous fermentation processes from batch data, (b) the control of these processes, and (c) optimization of process designs for batch fermentations. There have been many models proposed for growth of l i v i n g systems, from the simple Malthusian exponential growth curve to the esoteric models of Frederickson, Tsuchiya and Ramkrishna [59,60,61]. In order to show the r e l a t i o n -ship between these models and to c l a r i f y the nomenclature and symbols used, a short derivation of the models to be examined in this section w i l l be given. 5.1.1 Zero order model. The simplest model considered w i l l be the zero order model 104 105 d C x — - = y (5.1) dt which upon integration assuming constant y and that at t = 0, C v = C° becomes A X C x = C° + yt (5.2) The shape of this very simple curve can be seen in Figure 16a. The model can be seen to f a i l as t goes to i n f i n i t y . Perhaps this f a i l u r e could be overcome by imposing the condi t i ons C = C° when t < t< X X C = C° + y ( t - t°) when t° < t < t F (5.3) X X C v = C when t > t r X X , c 0 0 - c° F _ 4 . 0 a. X X where t = t° + T . T Figure 16. Sketches of the Shapes of the Various Models Tested. 107 These conditions, however, require two more arbitrary param-eters, t° and t F . It may prove more f r u i t f u l to use these two parameters in continuous models. 5.1.2 Autocatalytic or Malthus model. The autocatalytic or Malthus model; d C x — = yC ( 5 . 4 ) dt x can be integrated, assuming y to be constant and C = C° X X at t = 0, to give C - C° x EXP(y x t) (5.5) X A or rearranged into pseudo-linear form t = t° + 1 x Ln(C x/C°) (5.6) 108 This equation has an exponential shape as shown in Figure 16b. It can be seen that at low t, the shape of this curve approximates very well the usual shape of bacterial growth curves. As t approaches i n f i n i t y however, the model predicts that C goes to i n f i n i t y which of course is impossible. A ' 5.1.3 F i r s t order substrate l i m i t i n g model. Assuming that the bacterial growth rate is f i r s t order in a ' l i m i t i n g ' substrate or nutrient, then d C x — = yC n ( 5 . 7 ) dt where C n is the concentration of the l i m i t i n g nutrient. Now l e t us assume that the bacterial growth rate is related to the disappearance of the l i m i t i n g substrate, thus: d C x d C n — " " Yx/n — ( 5 ' 8 ) dt x / n dt Now equation ( 5 . 7 ) can be integrated assuming that v x / n re-mains constant throughout the useful range of t, C , and C A l l thus: 109 x/n C° x - Y x/n C° n ( 5 . 9 ) or as w i l l sometimes be used for convenience x/n C - C° x x - Z ( 5 . 1 0 ) where Y / x C° x/n n and equation ( 5 . 7 ) can be written 5c dt x/n ( 5 . 1 1 ) which can be integrated making the usual assumptions to y i e l d ( 5 . 1 2 ) 110 This equation ( 5 . 1 2 ) can be line a r i z e d with respect to the parameters t° and y / Y x / n > thus t = f - u / Y x / n x L n C ^~C^~ x x ( 5 . 1 3 ) The shape of this curve is shown in Figure 16c where i t can be seen that the high end of the curve f i t s most observed bacterial growth rate data quite well indeed, but the t = 0 end does not f i t well at a l l . 5.1.4 Second order autocatalytic-substrate l i m i t i n g model The Malthus model which approximates the data at low time can be combined with the f i r s t order substrate l i m i t i n g model which approximates the data at high times, thus : dC. dt u C x C M x n ( 5 . 1 4 ) and applying equation ( 5 . 1 0 ) I l l dC. dt x C. x/n Cx - Cx - 1 ( 5 . 1 5 ) Which again can be integrated making the usual assumptions, thus: Cx * Cx + 1 C° + Z x EXP f- y / V x / n x (C° + Z) x t (5.16) or for convenience in estimating the model parameters, Cx " Cx + * z CJ • Z x EXP A x t (5.17) Again this equation can be li n e a r i z e d with respect to the parameters t° and A, thus: 112 ( 5 . 1 8 ) Note that in equation ( 5 . 1 8 ) the stoichiometric parameter, Z, and the i n i t i a l condition parameter, C° are not li n e a r -X ized by this rearrangement and must be estimated from other information before the parameters of equation ( 5 . 1 8 ) can be estimated by the methods of linear regression. The shape of this curve is shown in Figure 16d. The shape nicely approximates most bacterial growth cruves. One drawback i s , however, that the C - t curve is symmet-A r i c a l about the point where C = C° + i Z . Thus the model A A cannot approximate growth curves which sta r t up quickly and then slow down slowly or those curves which s t a r t off slowly and slow down abruptly. 5.1.5 Edwards and Wilke model. Edwards and Wilke [62] have proposed the use of the " l o g i s t i c " model Cx X 1 + EXP f ( t ) ( 5 . 1 8 . 5 ) 113 where f ( t ) = a 0 + a i t + a 2 t 2 + a 3 t 3 + a^t" + a 5 t 5 + ••• When f ( t ) = a 0 + a i t , this model becomes iden t i c a l to the second order autocatalytic model. However,when higher orders of the polynomial are used the meaning and the eff e c t of the parameters in the model become obscure and use of the model, for extrapolation in p a r t i c u l a r , becomes extremely dangerous. 5.1.6 Monod model. The model proposed by Monod [63,64] taken by analogy from the kinetic expressions for enzyme kinetics of Michaelis and Menten as described by Laidler [65] makes use of a satu-ration constant K', thus: dC u C C M x n x ( 5 . 1 9 ) dt K' + C n In this model as K' approaches 0, the rate equation tends to the Malthus model and as K1 becomes i n f i n i t e l y large, the 1 1 4 equation tends to the second order autocatalytic model where y is replaced by y/K'. Applying equation ( 5 . 1 0 ) to equation ( 5 . 1 9 ) and integrating making the usual assumptions, and l e t t i n g K = K' x Y x. , then: Cx 'c; • 1 C° + Z x EXP-A y(CJ + * z) +• r cx c° + z X K K cx ( 5 . 2 0 ) V. J \ which usually converges for K > (C° + Z) or K 7 c° + z x ( 5 . 2 1 ) C° x EXP{yt} x Z which usually converges for K < (C° + Z). Note that these A equations cannot be solved e x p l i c i t l y for C , and that i t e r a -tive techniques are necessary to solve for C . 1 1 5 For convenience l e t A = y c ; + z / K then equations (5.20) and (5.21) become c° c° + z X X C" + z x  frjol K C° + Z x EXP{- At.}4 A K C° x EXP-A A K CJ + Z - 773-c° CJ - z C u + Z x (5.22) This model can be linearized with respect to the parameters U (or A) and K. (but not C° or Z) , thus: A 116 t° + - Ln y Cx " C x " + A" L n ' I J t Cx Cx Cx - Z (5..23) The e f f e c t of the parameter K is to skew the C - t curve. A The u t i l i t y of this parameter w i l l be demonstrated later when some batch fermentation data are used to determine the parameters of this model. This concludes the section on model development. It is believed that the scatter in the data of this work lim i t s the model used to describe those data, to a maximum of two kinetic parameters (K and y), one stoichiometric parameter ( v x/ n) and perhaps one independent i n i t i a l condition parameter (C°) which can be fixed at some measured i n i t i a l A value or relaxed and allowed to find i t s own value. For more accurate data, more complex models involving more parameters can be developed. One approach would be to build a model analogous to the i n h i b i t i o n - a c t i v a t i o n models used for enzyme kinetics by Laidler [65], Bray and White [66] or Higgins [67]; or to the well known heterogeneous ca t a l y s i s models which go under various names such as Hougen-Watson and Langmuir-Hinshel1 wood, etc. Other approaches to building of kinetic models for bacterial growth based upon d i f f e r e n t growth rates for 117 di f f e r e n t parts of the bacterial c e l l have been suggested by Frederickson et al. [59], Ramkrishna et al. [60] and by Perret [68]. The physical basis of these models is the var-ia t i o n of DNA, RNA and c e l l u l a r protein in the c e l l at dif f e r e n t growth rates. This variation is very well i l l u -strated in two excellent papers by Neidhardt [69] and Kjeldgaard [70,71]. These models are much too complex to be f i t t e d to the data collected in this work. 5.2 Estimating the Parameters of the Models The s u i t a b i l i t y of the models developed above depends not only upon the models themselves, but also upon the methods used to estimate the model parameters from the experimental data. The c r i t e r i a used to determine the 'goodness of f i t ' are also important in selecting an acceptable model. In this case we can assume that of the variables c e l l concentration and time, the only s i g n i f i c a n t error is in the measurement of c e l l concentration C . Furthermore, i f we assume that A the error in C is a random variable with zero mean and A that i t is normally distributed about the mean, then the estimates of the model parameters which are chosen in such a way so as to minimize the sum of the squares of the re s i d -uals of C , that i s , £(C Y - C .)2» w i l l be the unbiased 118 * estimates of the parameters with the smallest variances of a l l the possible estimates of the parameters [72]. There-fore, our primary c r i t e r i o n of a well f i t t e d model w i l l be the sum of squares of the residuals of C , no matter how the A model parameters were actually estimated. 5.2.1 Estimating the parameters of unintegrated rate models. The simplest way to estimate the model parameters is to l i n e a r i z e the rate equations with respect to the param-eters, and then to minimize the sum of squares of the assumed dependent variable. For example, the Monod rate equation dC y c x c x n ( 5 . 1 9 ) dt K' + C n can be li n e a r i z e d to 1 (5.24) R C n S e e H i m m e l b l a u C 7 2 ] , p a g e 1 1 0 . T h i s t e x t s h o w s v e r y v i v i d l y t h e p i t f a l l s o n e c a n f a l l i n t o w h e n m o d e l b u i l d i n g . 119 or for cases where measurement of C n is not possible or is d i f f i c u l t , equation ( 5 . 1 0 ) can be applied to give RX ~ I " \ x C - C° - Y . C ( 5 . 2 5 ) ^ M x x x/n n in which case (Y x/ nC°) must be known or estimated from other i nformati on. The f i r s t , and most important problem one encounters in estimating the parameters of rate equations with batch data is in determining the rate, R , from concentration-A time data. In this work R was determined numerically using A two methods: (1) linear interpolation by joining each set of two consecutive points on the C - t curve and estimatii A R at Cj = ( C ^ + 1 + c])/2 as the slope of the l i n e . That x x x x x i s C 1 + l-C 1 ^ i = 7 T T ^ f C 5 - 2 6 ) or (2) quadratic interpolation which was done s i m i l a r l y to the linear interpolation except that a second degree polynomial 120 was passed through each set of three consecutive points, and the slope was determined at the mean C" and t. A S i m i l a r l y , the second order rate model of equation ( 5 . 1 5 ) was linearized to c x dt f R x ) x/n C° + Y , C° x x/n n x C. ( 5 . 2 7 ) x/n Given that C° and C° are known or can be estimated from other data, the model parameters of this model, u and Y x^ n, can be determined by the usual method of linear least squares minimization. The rates for use in f i t t i n g this model were determined from the batch C - t data by the same methods as x J for the Monod model case above. The consequences of using RY/C and C /R as the A A A A dependent variables and the consequences of estimating R A from batch data w i l l be discussed l a t e r . 5.2.2 Integrated linear models with time as the  dependent variable. A l l the models discussed in this section can be linea r i z e d at least with respect to the kinetic parameters 121 K1 , and y ( a l t e r n a t i v e l y K and A) by taking the log of the terms containing C as shown in equations ( 5 . 6 , 5 . 1 3 , 5 . 1 8 A a n d 5 . 2 3 ) . These models along with the linear model of equation ( 5 . 2 ) are summarized in Table 32. The linearized parameters of these models can be estimated by the methods of linear regression using time as the dependent variable. For the f i r s t order substrate l i m i t i n g model of equation ( 5 . 1 3 ) , the second order model of equation ( 5 . 1 8 ) and the Monod model of equation ( 5 . 2 3 ) i t was necessary to use an a p r i o r i estimate of the stoichiometric parameter Y x/ n-The value of Y x/ n chosen for this purpose was the value of the parameter which was later found for the 'best' model, estimated using non-linear methods. See Table 36 for this value. For some of these models the term Ln(-Z/(C - C° - Z)) A A appears. When the argument of this term is negative the data point was not used to avoid the problem of logarithms of negative numbers. This aribtrary discarding of data of course, decreases the number of degrees of freedom associated with the residual error of the model. 5.2.3 Estimation of the parameters of the nonlinear models. E a r l i e r in this section we stated that only the estimates of the parameters given by minimizing the sum of 1 22 the squares of the residuals of C would give the 'best' X estimates of the model parameters. However, except for the t r i v i a l zero order case, forcing C to be the dependent variable causes the model to become non-linear in one or more of the kinetic parameters. The parameters of two models, the second order model of equation (5 . 1 7 ) and the Monod model of equation (5.22), were estimated using non-linear least squares tech-niques. The parameters of the other models were not e s t i -mated using non-linear techniques because they are either t r i v i a l (the zero order model) or because they were shown to be inadequate by l o g i c . The method used to determine the parameters of these non-linear models was the method of l i n e a r i z a t i o n (Taylor's series) as given by Draper and Smith [73] or Himmelblau [72]. The transformation on the param-eters : Pi = Ln A P 2 = Ln K P 3 - L n Yx/n P* = Ln C° By ' b e s t ' e s t i m a t e s o f p a r a m e t e r s we m e a n t h o s e e s t i m a t e s o f t h e p a r a m e t e r s w h i c h h a v e t h e s m a l l e s t sum o f s q u a r e s o f t h e r e s i d u a l e r r o r s . 123 was used as suggested by Hunter and Mezaki [74] and by Box and Hunter [83]. This transformation r e s t r i c t s the values Mezaki also claim that this transformation makes the i t e r a -tive technique converge faster because i t makes the contours of the sum of squares surface less elongated. The University of B r i t i s h Columbia Computer Centre l i b r a r y sub-routine DPRLQF was used with the transformations noted above to do the calculations necessary to estimate the parameters of the models. To calculate the C function for the Monod A model, C was f i r s t estimated using the best f i t parameters A from the second order model. This estimate was then entered into the right-hand side of equation ( 5 . 2 2 ) and C was A improved i t e r a t i v e l y u n t i l the convergence was acceptable. The vector of p a r t i a l derivatives, (3C /3P.), necessary for use in the subroutine was calculated by perturbing the vector P., by a small amount, <5. , and calculating of C°, K, A and Y x/n to positive numbers. Hunter and ( 5 . 2 9 ) 124 Double precision arithmetic was necessary for this calculation to avoid round-off error. This and the fact that more i t e r a -tions were required than for estimating the parameters of the non-linear second order model, resulted in parameter estimation for the Monod model taking up to 10 times as much * computer time as did the f i t t i n g of the second order model. For both the Monod and the second order models, the i n i t i a l condition parameter C° , can be relaxed and A allowed to find i t s own value, or fixed at some known or measured value. Both of these were t r i e d and the best method was determined by comparison of the predicted curve with the raw data. A l l of these models and their parameters are summarized in Table 32. 5.3 Testing the Models To test these models and the estimated parameters, the data for the 7 - l i t r e fermentor Run No. 2 of Section 3.4 were used. For convenience these data and any transformed variables calculated from them are tabulated in Tables 33, W e r e I t o s t a r t t h i s a l I o v e r a g a i n , I w o u l d s e r i o u s l y c o n s i d e r u s i n g t h e m e t h o d o f M a r q u a r d t E - 75 ,7 611 f o r f i t t i n g t h i s m o d e l . T h i s m e t h o d i s a l s o o u t l i n e d b y H i m m e l b l a u [ 7 2 ] , B a l l [ 7 7 ] a n d D r a p e r a n d S m i t h [ 7 3 ] . T a b l e 32 SUMMARY OF THE MODELS TO BE DISCUSSED IN THIS SECTION PARAMETERS DETERMINED IN THE MODEL TYPE OF FIT TYPE OF MODEL EQUATION OF MODEL K 1 ' x / n C° x t° EQUATION NUMBER IN TEXT PLOTTING NUMBER R a t e R a t e Monod S e c o n d o r d e r _ A _ _ L R x *m C - C° - Y . C° m x x x / n n R y r r-7^ - = v - 2 - C°+Y , C° -T7-!2-C C x Y x / n l x x / n n> Y x / n x yes y e s no no no yes no y e s no no 5-25 5-27 8A l i n e a r 8B quad 1A l i n e a r IB quad I n t e g r a t e d w i t h t i m e as t h e d e p e n d e n t v a r i a b l e Z e r o o r d e r M a l t h u s F i r s t o r d e r S e c o n d o r d e r C° + u t x Km t = t° + -L- l n t = t° - m l n t = t° - ^ l n x / n rc c - c° - z - x x -C x C x - C x " Z Monod t = t«> + _ J - 1 n x + ^ l n  ym L x A FX , " x -Z C*"*C -C u-Z X X X y e s yes y e s no yes no no no y e s y e s no no no no no no no no no y e s no no y e s y e s yes y e s 5-2 5-6 5-13 5-18 5-23 CONTINUED ro cn T a b l e 32 (Continued) PARAMETERS DETERMINED IN THE MODEL TYPE OF FIT TYPE OF MODEL EQUATION OF MODEL vm A K K' Y x / n C x t ° EQUATION NUMBER IN TEXT PLOTTING NUMBER N o n - l i n e a r Second o r d e r C° (C° + Z) [no yes no no y e s * no 5 -17 9A C° f i x e d x C° + Z >l" A t  r „ C x ( C x + Z ) 1 no yes no no yes y e s no 9B C° r e l a x e d x C x + Z K { C 0 ' r ° + 7 o "At _X * K . Monod -K f no [no yes yes yes y e s no no y e s y e s * yes no no 5 -22 5 -22 6A C° f i x e d 6B C° r e l a x e d AK C°+Z c°+z = C° I  x fCx -Z ] A P C x- C x- Z J NOTES: I. ' y e s ' i n d i c a t e s t h a t t h e p a r a m e t e r i s d e t e r m i n e d by t h e m o d e l . —' ro 2. 'no' i n d i c a t e s t h a t t h e p a r a m e t e r i s n o t i n v o l v e d i n t h e m o d e l . °^ 3. '*' i n d i c a t e s t h a t t h e p a r a m e t e r m u s t be known b e f o r e t h e m o d e l c a n be f i t t e d . Table 33 DATA OF SSL RUN NO, 2 FOR TIME VS. C v MODEL PARAMETER ESTIMATION A TIME (hr) (gm/l) LN(C X) LN( - Z * / ( C X - C° - Z)) LN(-Z x C x / ( C ; ( C x - C° • - z ) ) ) 1 0.071 -2.645 0.0 -0.0000001788 10 0.0961 -2.342 0.003589 0.3063 2 2 i 0.2417 -1.420 0.02474 1 .250 32 0.5842 -0.5375 0.07629 2.184 51 i 1.411 0.3444 0.2130 3.202 77 2.311 0.8378 0.3866 3.869 96 3.637 1 .291 0.7142 4.650 106 3.971 1 .379 0.8171 .841 118i 4.078 1 .406 0.8522 4.903 120 4.750 1 .558 1 .108 5.311 143 5.710 1 .742 1 .645 6.033 166i 6.342 1 .847 2.279 6.771 190 6.743 1 .909 3.101 7.655 217 6.757 1 .011 3.146 7.702 263 6.810 1.018 3.342 7.906 285 6.824 1 .920 3.397 7.963 Z = 6 . 98 g m / l a s d e t e r m i n e d f r o m ' b e s t ' n o n - l i n e a r m o d e l . C ° = 0 . 0 7 1 g m / l a s m e a s u r e d 128 34 and 35. The estimated values of the parameters, sums of squares of the residuals of C and the degrees of freedom A associated with the sums of squares are tabulated in Table 36. 5.3.1 Rate models. The rate models, as can be seen from the sum of squares column in Table 36, give very poor estimates of the parameters. The Monod model i s p a r t i c u l a r l y bad. Figure 17, a plot of C vs. time shows graphically just how bad A these parameter estimates are. Even the best of these rate models missed most of the experimental points. Let us examine these rate models more closely before dismissing them to see why they f a i l to estimate reasonable values of the model parameters. Figure 18 shows the interpolated rates plotted against time. Because of the measurement errors in c e l l concentration and because of the uneven d i s t r i b u t i o n of the sampling times, samples taken quite close together such as the points at 118i hours and 120 hours, can give apparent rates much higher or much lower than the actual rates. This effect is exaggerated when the reciprocal of the rates is taken, p a r t i c u l a r l y for low rates. In order to see the effe c t of these points on the estimation of the parameters Table 34 C A L C U L A T E D D A T A F R O M L I N E A R I N T E R P O L A T I O N F O R R A T E M O D E L S TIME (hr) CELL CONC. ( g m/l) RATE ( g m/l/hr) Cx/R - i / ( c x - c; - z) C Squared A 5.5 0.0835 0.00558 14.97 0.1434 0.00698 16.25 0.1689 0.02330 7.249 0.1452 0.02853 27.25 0.4129 0.07211 5.727 0.1505 0.1705 41 .75 0.9976 0.08481 11 .76 0.1651 0.9953 64.25 1.861 0.0706 26.36 0.1925 3.464 86.5 2.974 0.1396 21 .31 0.2450 8.845 101 3.804 0.06688 56.88 0.3075 14.47 112.25 4.025 0.01702 236.5 0.3299 16.2 119.25 4.414 0.8963 4.925 0.3785 19.48 131.5 5.230 0.08344 62.68 0.5476 27.35 154.75 6.026 0.05382 112.0 0.9707 36.31 178.2 6.543 0.03414 190.6 1 .948 42.80 203.5 6.750 0.00103 6553.0 3.269 45. 56 240.0 6.784 0.00233 2911 .0 3.673 46.02 274.2 6.817 0.00117 5827.0 4.186 46.47 N O T E : C ° was t a k e n a s 0 . 0 7 1 g m / l a n d Z was t a k e n a s 6 . 9 8 6 g m / l t a k e n f r o m ' b e s t ' n o n - l i n e a r m o d e l . Table 35 CALCULATED DATA FROM QUADRATIC INTERPOLATION FOR RATE MODELS TIME (hr) CELL CONC. (gm/l) RATE (gm/l/hr) Cx/R - i / ( c x - C J - Z) C Squared A 11.17 0.1363 0.00746 18.27 0.1445 0.01857 21 .5 0.3073 0.0233 13.19 0.1482 0.09445 35.33 0.7457 0.03959 18.83 0.1585 0.556 53.5 1.435 0.03869 37.10 0.1779 2.061 74,83 2.453 0.0517 47.45 0.2173 6.018 93.0 3.307 0.0535 61 .82 0.2667 10.93 106.8 3.895 0.0205 189.9 0.3164 15.17 115.8 4.266 0.1707 24.99 0.3585 18.20 127. 2 4.845 0.1855 26.12 0.4524 23.48 143.2 5.601 0.03429 163.3 0.6871 31 .37 166.5 6.265 0.022 284.9 1 .264 39.25 191.2 6.614 0.0086 769.1 2.263 43.75 223.3 6.770 0.00087 7782.0 3.499 45.84 255.2 6.797 0.00091 7469.0 3.862 46.20 N O T E : C ° = 0 . 0 7 1 g m / l a n d Z = 6 . 9 8 6 g m / l w e r e u s e d a s d e t e r m i n e d f r o m A ' b e s t ' n o n - l i n e a r m o d e l . Table 36 ESTIMATED PARAMETERS OF THE MODELS ESTIMATED PARAMETERS OF THE MODELS MODEL EQN. NO. PLOT NO. y A K K' v 'x/n Z t° ( VC x > df Monod r a t e l i n e a r i n t e r p o l a t i on Monod r a t e q u a d r a t i c i n t e r p o l a t i on Second order r a t e l i n e a r i n t e r p . Second order r a t e q u a d r a t i c i n t e r p . 5-25 5-25 5-27 5-27 8A 8B IB 1A 0.00221 0.00139 0.01412 0.00699 0.005166 0.003613 0.1176 0.05649 3.05 2.71 2.66 2.37 0.873 0.875* 0.896 0.849 6.986* 6.986* 7.168 6.789 0.071* 0.071* 0.071* 0.071* -321 .7 322.7 89.85 11 .99 13 12 1 3 12 Zero order time as dep. var. Malthus model time as dep. var. F i r s t order model time as dep. var. Second order model time as dep. var. Monod model time as dep. var. 5-2 5-6 5-13 5-18 5-23 3 2 4 5 7 0.0258 0.02178 0.0135 0.00397 0.0255 0.0285 0.02178 0.01545 0.0321 0.01891 0.509 8.303 0.873* 0.873* 0.873* 6.986* 6.986* 6.986* 388 0.151 0.071* 0.071* 0.071* -26.64 -3.88 11 4 6471.0 36.64 5.703 2.143 14 14 14 14 13 Second order n o n - l i n e a r C° f i x e d C° r e l a x e d Monod model n o n - l i n e a r C° f i x e d C° r e l a x e d 5-17 5-17 5-22 5-22 9A 9B 6A 6B 0.00571 0.00416 0.0343 0.10564 0.04613 0.03434 0.0217 0.0286 11.14 25.65 9.725 21 .65 0.827 0.830 0.873 0.844 6.614 6.642 6.985 6.753 0.071* 0.2321 0.071 0.1911 -1 .723 0.6174 0.6961 0.5622 14 1 3 13 12 NOTES: P a r a m e t e r v a l u e s marked '*' were not d e t e r m i n e d from t h e m o d e l s . T h e s e v a l u e s w h i c h must be known b e f o r e ttve model can be f i t t e d were assumed t o be e q u a l t o t h o s e v a l u e s f o u n d f r o m t h e ' b e s t ' f i t . A Is d e f i n e d d i f f e r e n t l y a c c o r d i n g t o t h e m o d e l . See t e x t f o r t h e s e d e f i n i t i o n s . U n i t s : t° - hours Z - gm/l C° - gm/l Y . - d l / 1 . Note t h a t Z and C must be i n i d e n t i c a l u n i t s , s i n c e we have x x/n x C° - gm/dl a r b i t r a r i l y c h o s e n qm/d I f o r C , we must use d l / l f o r Y , . n ' M n' x/n K' - gm/dl K and A as v a r i o u s l y d e f i n e d i n t e x t . Type of fit rate,quad, rate,linear rate.linear rate,quad. 100 200 TIME (hours) 300 co ro Figure 17. Rate Models - C vs. Time Showing Results of Parameter Estimation 133 lO d e UJ I- — < d No. Model Type 6A Monod non-lin 9B 2d ord non-lin O linear interpolation © quadratic interp. o e> to q < d E UJ o < CD O 6 0 100 200 TIME (hours) Figure 18. Rate Models - Rate vs. Time Plot Showing Poor Quality of Rate Data. 134 of the Monod model, l e t us examine Figure 19, a Lineweaver-Burk [17] plo t , in which C /R is plotted against A A - 1/(C Y - C° - Z) or 1/C„ . This plot shows that the low A A n values of R (which are the high values of C /R) have a much A greater e f f e c t upon the slope of this plot than the high values of R. The same plot with a change in scale shown as Figure 20, demonstrates that the points which were crowded into the (0,0) corner of Figure 19 generally follow the lines 6A and 9A, which are the lines which represent the models which w i l l l a t er be found to be the best models when the parameters are estimated using non-linear techniques. Therefore, by taking the reciprocal of the poorly defined rate, R, we very heavily weight the small values of R which are the least r e l i a b l e values of R on a per cent error basis. Furthermore, even i f R were determined perfectly (by using the parameters of the Monod model estimated by non-linear techniques as an interpolation formula for example), the errors in R or C /R would not necessarily be normally X A distributed random variables as we have assumed that the errors in C x are. These d i f f i c u l t i e s in estimating the parameters of rate equations by l i n e a r i z a t i o n of the rate equation are well reviewed by K i t t r e l l [78,79]. They w i l l not be dealt with further here, except to warn that the d i f f i c u l t i e s may be more pronounced in fermentation kinetic studies because of the great variation in measurements found in such work. 135 xlO 3 No. Model Type 6A Monod non-lin. 8A Monod rate,lin. Figure 19. Li neweaver-Burk. Plot for Rate Models. Figure 20. Lineweaver-Burk Plot Expanded from Figure 19 137 5.3.2 Integrated models with time as dependent variable. The models with time as the dependent variable are more promising as can be seen in Figure 21. The zero order, Malthus, and f i r s t order substrate l i m i t i n g models were not expected to f i t well, and they certainly do not, but were put in to show their relationship to the second order model and to the Monod model. The sums of squares of the resid -uals for the second order model and the Monod model as shown in Table 36 are beginning to approach a reasonable range. However as can be seen in Figure 21, even these models f i t t e d in this way miss many of the experimental points. However, this technique cer t a i n l y selects the Monod model over the second order model. The reason why the estimates of the parameters of those models with time as the dependent variable are not s a t i s f a c t o r y is that when time is selected as the dependent variable we automatically adopt the assumption that a l l of the s i g n i f i c a n t measurement error involved in the experiments are associated with time measurement and that very l i t t l e s i g n i f i c a n t error i s associated with the independent variable, bacterial c e l l concentration. As a consequence of assuming that this measurement error is associated with time measure-ment, we weight those experimental points for which dt/dC is high. Thus, while we should expect good estimates for 8£ L 139 the model parameters where dC /dt is close to zero, i t can A be seen from Figure 21, that these models are most unreliable in the most important region of the curve where the highest fermentation rates occur. This problem is overcome when the 'true' dependent variable, C , is used as the actual A dependent variable, since the most accurate data points are those where dC x/dt , the fermentation rates are high. 5.3.3 Non-linear models. The estimation of the parameters of the models with C , bacterial c e l l mass, as the dependent variable give the least sums of squares of a l l the techniques used. Figure 22 shows the second order and Monod models on a C - t plot X for these non-linear models along with the data used to estimate the model parameters. Since i t is d i f f i c u l t to discriminate between these models on this type of plo t , the residuals (C - C ) were plotted against time in Figure 23. X X -These residual plots show much which could not be detected from the C - t plot. F i r s t l y , the point at 118i hours is X obviously an o u t l i e r . Residual plot 9A (Figure 23) of the second order model with C° fixed shows that most of the 1.723 X sum of squares of residuals comes from the points in the range from 30 - 100 hours indicating a serious lack of f i t - CD h-E U J o >-t r o CM r-0 6 A . 6 B 9 B 9 A Model Monod Monod 2 d ord 2 d ord fixed relaxed fixed relaxed S S L R u n N o . 2 100 200 TIME(hours) 300 5 * Figure 22. C vs. Time Plots for Non-Linear Models with C as Dependent Variable x x R E S I D U A L S 1 + 1 + 1 + i + 1 9 1 . 1 1 i 1 > 0 < • • > • > » 1 • 1 i • • • o e T l cz -s 1 > 4 t • n> ro _ OJ - o • • • • o i » » < 9 9 • 9 • » • o Cl _,. «a» s m » • • o » • i > « c-t- © - h E» o ro • • • - O o " o 1 > ( > » 1 r -3 0) OJ -s o C • o. CD CO « • 4 • - o o 0) 0> CD CO w > See figure 22 for legend 142 in this region. When C° is relaxed and allowed to find i t s own l e v e l , the sum of squares is reduced by almost 2/3 rd., but there is s t i l l some lack of f i t between 0 and 100 hours. Also, the fact that C° is much greater than the value of X C° measured may well indicate a systematic error at low A bacterial c e l l concentrations which causes the measured C° A values to be too low. The Monod model, plots 6A and 6B in Figure 23, eliminate most of the lack of f i t problem which occur with the second order model. The small bias remaining is prob-ably caused by the compensation allowed for the o u t l i e r at 118i hours. There s t i l l remains, however, the C° - C° bias X X which w i l l be dealt with later when the estimates of the parameters for the models of the other 7 - l i t r e fermentations are discussed. 5.3.4 Tentative conclusions of kinetic model parameter  estimati on. At this stage of the model building process certain tentative conclusions can be entertained. These are: 1. The n o n - l i n e a r l e a s t s q u a r e s t e c h n i q u e i s t h e most r e l i a b l e o f t h e t e c h n i q u e s t r i e d f o r p u r p o s e s o f model d i s c r i m i n a t i o n . 1 4 3 2. Residual p l o t s can give much a d d i t i o n a l v a l u -able information about lack of f i t of the b e t t e r models which does not show up on normal C - t p l o t s or from the sums of squares. 3. Of the models s t u d i e d , the Monod model gives the best estimates of the model parameters when compared to the raw data. The second order model with C° relaxed a l s o x estimates the model parameters w e l l . However, i t is probable that the parameter C° is compensating f o r the lack of the K. parameter in t h i s model. k. There appears to be a systematic e r r o r in C x at low c e l l c o n c e n t r a t i o n s which makes C° higher than the value of C° as measured when C° is t r e a t e d as an a d j u s t a b l e x x J parameter. 5. Rate models, when actual c o n c e n t r a t i o n s are measured with c o n s i d e r a b l e e r r o r , give very misleading r e s u l t s . 6. L i n e a r i z e d models with time as the dependent v a r i a b l e are s u i t a b l e f o r rough screening of models, but to be s a f e , n o n - l i n e a r l e a s t squares techniques are recommended for any f i n a l a n a l y s i s . T h i s c o n c l u s i o n w i l l l a t e r be s h o w n t o be w r o n g . S e e S e c t i o n 5.5. 144 5.4 Use of These Estimated Model Parameters to Design a  Continuous Fermentor Now that we have s a t i s f i e d our original goal of developing a good model, what can we do with i t ? One of our j u s t i f i c a t i o n s for doing this modeling work was that i t was necessary before a continuous fermentation system could be designed. Let us now calculate the optimum hold-up times for a single continuous s t i r r e d tank reactor used for the production of P. freudenreichii dry c e l l s from spent s u l f i t e liquor and yeast extract. The parameter estimates of the Monod and second order model parameters w i l l be the only ones used in these calculations. * The material balance equations around our fermentor can be written: dC. -gy 5- = R. - D x ( C . - C l ) (5.30) where i represents the i t h component - c e l l s , nutrients, etc. - or at steady state when dt M o s t o f t h i s c o n t i n u o u s f e r m e n t o r t h e o r y h a s b e e n w o r k e d o u t by H e r b e r t , E l s w o r t h a n d T e l l i n g L " 8 0 , 8 l j . 145 for the Monod model dC x y C x C n ~dT = K' + C - D x C x = 0 n ( 5 . 3 1 ) and dC dt x/n y C C p x n K' + C n C - C" n n (5.32) from which r _ D x K'  L n y - D (5.33) and C = Y . x x x/n D x K1 n y - D (5.34) 146 and the c e l l production rate can be written P = D x C = D x Y . x x/n C ° n D x K ' y - D (5.35) where P has the units of, for example, gm/litre/hour of bacteria, and the optimum d i l u t i o n rate, D , can be found r ' m when Then ^ = 0 dD u m ^ 1 - 1 + K . ' / C ° n (5.36) For the case of the second order model, the same relationships at steady state become and dC n u - ~ n C C - D d t Yx/n x " C n from which and C n = D/u C = Y , C - D/u x x/n I n ' M • and the optimum d i l u t i o n rate i s given 148 D m = i y C ° ( 5 . 4 1 The d i l u t i o n rates and hold-up times for the pre-dicted by the estimated parameters of the models of interest are tabulated in Table 37. As can be seen from Table 37, even these good models, well f i t t e d , predict a wide variety in hold-up times from the same data. However, a continuous fermentor designed to have a hold-up time of at least 125 hours should be adequate to at least find the optimum hold-up time. More-over, this calculation shows the f o l l y of designing continuous fermentation systems with only batch kinetic data; continuous ki n e t i c fermentation data i s necessary before continuous fermentation processes can be designed with any degree of confi dence. 5.5 Results of Kinetic Model Parameter Estimation of  the Other Runs How do the results of estimation of the parameters of the second order and Monod models to the data of the other 7 - l i t r e fermentor runs compare with the estimates of the model parameters of the run estimated in d e t a i l above? Table 37 PREDICTED OPTIMUM DILUTION RATES FROM MODEL PARAMETERS MODEL PREDICTED VALUES AT OPTIMUM PRODUCTION RATE AT STEADY STATE Monod C° fixed X Monod C° relaxed X C x ( g m / 1) C n ( g m / d l ) D ( h r ) H O L D - U P T I M E ( h r ) 3.73 3.92 4.27 4.67 0.0086 0.0141 116 71 Second order C° fixed X Second order C° relaxed X 4.135 4.15 5.0 5.0 0.0285 0.0207 35 48 A s s u m e s f e e d c o n c e n t r a t i o n o f c ' = 1 0 . 0 g m / d l . 150 Table 38 shows the estimated parameters from a l l the 7 - l i t r e fermentations for the Monod and second order models. These curves predicted from the estimated model parameters have also been plotted in growth curve plots of C vs. t in Figures A 24, 25 and 26. In order to exploit the greater powers of discrimination of the residual p l o t , the predicted curves are shown in residual plots for these runs in Figures 27, 28 and 29. As before, although there is not much difference in the sums of squares of the Monod model and the second order model, both estimated by non-linear least squares tech-niques, the superiority of the Monod model cannot be denied. In order to check out the tentative conclusion of Section 5.3.4, that there appears to be a systematic error at low values of c e l l concentration, l e t us again examine C° and C° for these other runs. These measured A A i n i t i a l c e l l concentrations C° and estimated i n i t i a l c e l l A concentrations C° are compared in Table 39. For the Monod A model parameters estimates, our previous hypothesis of Section 5.3.4 does not hold up very well. However, for the second order f i t s the phenomenon previously observed that C° > C° A X i s again observed. Therefore, i t must be concluded that c e l l concentration measurements at low concentrations are no more prone to error than high c e l l concentrations, and that the bias in C° in the second order model comes about Table 38 ESTIMATED VALUES OF THE PARAMETERS FOR A L L THE 7-LITRE FERMENTOR RUNS RUN IDENTIFICATION C x (gm/l) Z (gm/l) A K Y x / n ( d l / l ) (h y r " 1 ) K' gm/dl (C - c ) 2 x x' df MODEL TYPE 0 . 0 4 7 9 5 3 3 9 0 0 3 3 0 9 8 . 1 2 0 . 7 4 1 0 . 0 4 9 9 1 0 . 9 6 3 . 0 1 6 21 S S L R u n N o . 1 0 . 0 7 1 6 9 8 5 0 0 2 1 7 11 . 1 4 0 . 8 7 3 0 . 0 3 4 3 1 2 . 7 6 0 . 6 9 6 1 13 M o n o d m o d e l S e c . 3 . 4 0 . 0 6 2 1 4 0 9 3 5 0 01 531 4 . 9 4 9 0 . 4 0 1 0 . 0 1 8 2 1 2 . 3 4 0 . 5 1 9 10 N o n - l i n e a r f i t , n O t i r~. A 0 . 0 8 7 6 5 8 0 2 0 0 2 3 8 8 . 9 3 8 0 . 5 8 0 2 0 . 0 3 6 1 1 5 . 4 0 0 . 7 5 8 3 8 u T i x e u X 0 . 0 2 0 0 2 7 4 7 0 0 1 6 2 2 2 . 9 4 2 6 0 . 2 7 5 0 . 0 1 7 2 1 0 . 7 0 0 . 3 1 9 12 S S L R u n N o . 2 0 . 0 0 0 2 7 5 5 2 3 0 0 2 3 9 6 . 0 3 1 0 . 7671 0 . 0 3 4 2 1 0 . 3 1 2 . 8 0 20 M o n o d mode 1 S e c . 3 . 4 0 . 1 1 9 1 6 7 5 3 0 0 2 8 6 25 . 6 5 0 . 8 4 4 1 0 . 1 0 6 7 3 0 . 3 9 0 . 5 6 2 2 1 2 N o n - l i n e a r f i t , 0 . 0 0 5 1 4 2 1 2 0 0 1 3 4 4 . 5 1 3 0 . 4 1 3 0 . 0 1 4 3 1 0 . 9 3 0 . 4 7 3 9 C° r e l a x e d 0 . 1 8 2 4 5 6 4 8 0 0 2 8 5 13 . 8 8 0 . 5 6 5 0 . 0 6 7 8 2 4 . 5 7 0 . 7 2 8 9 7 X 0 . 0 4 9 7 2 2 9 0 0 0 7 2 5 _ 0 . 2 2 9 0 . 0 0 7 1 1 . 0 8 1 4 14 S S L Run N o . 3 0 . 0 4 7 9 5 0 9 4 0 0 7 6 1 - 0 . 7 0 7 0 . 0 1 0 5 - 3 . 6751 21 S e c o n d o r d e r m o d e l S e c . 3 . 4 0 . 0 7 1 6 6 1 4 0 0 4 6 1 - 0 . 8 2 7 0 . 0 0 5 7 1 - 1 . 7 2 3 14 Non1 i n e a r f i t . 0 . 0 6 2 1 3 6 6 8 0 0 5 6 1 - 0 . 3 6 0 0 . 0 0 5 4 1 - 0 . 9 3 4 6 10 C° f i x e d 0 . 0 8 7 6 5 507 0 0 5 0 9 - 0 . 5 5 1 0 . 0 0 5 0 1 - 1 . 4 4 8 2 9 x 0 . 2 2 3 4 2 3 0 6 0 0 4 1 0 3 _ 0 . 2 3 1 0 . 0 0 3 7 5 _ 0 . 4 7 6 5 13 S S L Run N o . 4 0 . 0 9 9 3 5 0 9 6 0 0 6 3 8 9 - 0 . 7 0 8 0 . 0 0 8 7 1 - 3 . 4 2 5 5 21 S e c o n d o r d e r m o d e l S e c . 3 . 4 0 . 2 3 2 1 6 6 4 3 0 0 3 4 3 4 - 0 . 8 3 0 0 . 0 0 4 1 5 - 0 . 6 1 7 4 13 N o n - l i n e a r f i t , 0 . 1 2 9 3 3 6 4 2 0 0451 7 - 0 . 3 5 7 0 . 0 0 4 2 8 - 0 . 8 5 0 7 10 C° r e l a x e d 0 . 2 7 3 9 5 2 0 2 0 0 4 0 7 6 - 0 . 5 2 0 0 . 0 0 3 8 7 - 0 . 7 5 7 4 8 x CD I-e O U l U J o >-cr Q 2 2 A & B O O 23 A a B 22A 22B 23A 23B Model Monod Monod 2d ord 2d ord c; fixed relaxed fixed relaxed SSL Run No. I 0 100 TIME ( hours) 2 0 0 ure 24. C vs. Time Plots for Non-Linear Models - SSL Run No. 1. x CD TIME (hours) Figure 25. C x vs. Time Plots for Non-Linear Models - SSL Run No. 3. 154 Tl ME ( hours ) Figure 26. C vs. Time Plots for Non-Linear A Models - SSL Run No. 4. RESIDUALS T T - o co -s n> r o 72 (D — i . Q -£ T3 O m o co _ co O 5 cn o 3 I 3 n> -s o Q -o o o o 4 i « 0 0 9 ro CD • e ro OJ e ro ro DD 0 I 0 0 0 0 ro ro See figure 24 for legend 99L 0 •I - — 2 4 A + 1 0 -I co < . z> +1 - 0 CO LU or -I + 1 0 -I -9 «-O P 9fl) 9 * • 9 ~tt—T~9 *~ -* 100 200 — J — i TIME ( hours) © 300 2 4 B -i o C M 25 A 0> C O 25 B cn Figure 28. Residual Plots for SSL Run No. 3 - Non-Linrar Models 0 -1 0 co - | _j < 3 Q + | £ o * - i +1 0 - r -* 9-o _a 2_ e e • e © • 9 9 - v V 9 0 100 200 TIME ( h o u r s ) 20A 2IA 2IB 300 -o c C P 20B -o CO CM 0) C P <E> CO Figure 29. Residual Plots for SSL Run No. 4 - Non-Linear Models 1 5 8 T a b l e 3 9 COMPARISON OF MEASURED AND ESTIMATED C° VALUES RUN , I D E N T I F I C A T I O N 1 " MONOD MODEL S E C O N D O R D E R MODEL C° X c ° * X C° X C° X C l o s t r i d i a l M e d i u m R u n 0 . 0 4 9 7 0 . 0 2 0 0 0 . 0 4 9 7 0 . 2 2 3 4 S S L R u n N o . 1 0 . 0 4 7 9 0 . 0 0 0 2 7 0 . 0 4 7 9 0 . 0 9 9 3 S S L R u n N o . 2 0 . 0 7 1 0 . 1 1 9 0 . 0 7 1 0 . 2 3 2 1 S S L R u n N o . 3 0 . 0 6 2 1 0 . 0 0 5 1 0 . 0 6 2 1 0 . 1 2 9 3 S S L R u n N o . 4 0 . 0 8 7 6 0 . 1 8 2 4 0 . 0 8 7 6 0 . 2 7 3 9 t A l l r u n s w e r e d o n e i n 7 - l i t r e f e r m e n t o r . S e e S e c t i o n 3.4. C ° i s t h e m e a s u r e d i n i t i a l b a c t e r i a l c e l l c o n c e n t r a t i o n , x C° i s t h e e s t i m a t e d i n i t i a l c o n d i t i o n p a r a m e t e r , x K B o t h a r e i n u n i t s o f g m / l o f d r y b a c t e r i a l c e l l u l a r m a t e r i a l . 1 59 because of deficiencies in the second order model. That i s , the parameter C° is performing some of the function of the A parameter K. The second order model w i l l , therefore, be re-jected on this basis and the Monod model with C° fixed at X the value measured w i l l be adopted as the best model to represent the experimental data. 5.6 Comparison of the Estimated Model Parameters Between Runs 5.6.1 Comparison between runs. How well do the f i t t e d parameters agree between runs? There are considerable differences between the e s t i -mated parameters for the d i f f e r e n t runs (see Table 38). The run which was reported in Section 3.4 in which Reinforced C l o s t r i d i a l Medium and glucose were used as fermentation medium as expected, is d i f f e r e n t from the others, as can be seen in the difference in the parameters. Only the param-eters of the Monod model f i t t e d non-1inearily with C° fixed J x w i l l be compared in this section. The f i r s t observation is that the spent s u l f i t e liquor Run No. 3 of Section 3.4 appears to be an atypical run as was previously noted. The kinetic parameter, A, i s lower for this run than for any of the others; the parameter K is also low as i s the stoichiometric parameter Y , . These 160 abnormalities were probably caused by the o v e r - s t e r i l i z a t i o n of the medium used for this run as was noted in Section 3.4. The agreement of the y i e l d parameter v x / n between runs was disappointing. The y i e l d c o e f f i c i e n t could not be predicted at these high c e l l concentrations as i t was from the low c e l l concentration data used in Section 3.1. Possible reasons for this discrepancy are: 1. D e s t r u c t i o n of some of the l i m i t i n g n u t r i e n t by over Cor under ) s t e r i l i z a t i o n of the medium. 2. These 7 - l i t r e fermentations had higher i n i t i a l c o n c e n t r a t i o n s of n u t r i e n t s -which r e s u l t e d i n higher c e l l c o n c e n t r a t i o n s , more metabolic p r o d u c t s , some of -which could w e l l be i n h i b i t o r s of c e l l growth. I t i s q_uite probable indeed, t h a t the b u i l d - u p of a c i d s i n the pH unadjusted f e r -mentations, prevented the t r u e y i e l d c o e f f i c i e n t from express-i n g i t s e l f . The problem of product i n h i b i t i o n i s d i s c u s s e d f u r t h e r i n S e c t i o n 3.6. U n d e r h e a t i n g d u r i n g s t e r i l i z a t i o n c o u l d a l s o c a u s e l o w e r y i e l d s b e c a u s e o f t h e e f f e c t d e s c r i b e d i n S e c t i o n 3.4. 161 5.6.2 Optimum hold-up times for the other runs. The optimum hold-up times for these other 7 - l i t r e runs are tabulated in Table 40. The same methods of calcu-latio n as were used in Section 5.4 were used for these runs. The conclusions which were made in Section 5.3.4 about the v a r i a b i l i t y in hold-up times depending upon the model used to predict them and the inadvisabi1ity of using batch kinetic data to design large scale continuous fermentation systems, are reinforced. 5.7 Comparison of the Results of this Work with Kinetic  Work of Others How do the kinetic parameters found in this work compare with the results of other workers in the f i e l d ? Unfortunately, there is very l i t t l e published work on the results of f i t t i n g extensive fermentation data. Some results have been published by Herbert, Elsworth and T e l l i n g [80] and by Ramanathan and Gaudy [82] and by Monod [63,64] for microorganisms other than Propionibacteria. Monod's pub-lished data for glucose l i m i t i n g media give values for the saturation parameter K', which is postulated to be controlled by the l i m i t i n g nutrient, of the order of 0.01 gm/l. For amino acid limited media his values of K' are even lower. 162 Table 40 PREDICTED OPTIMUM HOLD-UP TIMES FOR ALL THE RUNS MONOD M O D E L C ° F I X E D X RUN IDENTIFICATION Cx ( g m / 1 ) C n ( g m / d 1 ) D ( h r - 1 ) HOLD-UP TIME (hours) S S L R u n N o . 1 S e c t i o n 3 . 4 3.11 3.73 1.71 2.54 4.20 4.27 4.26 4.38 0.0138 0.0086 0.0047 0.0080 72 116 214 125 M o s t 1i k e 1 y h o l d - u p t i m e • 1 25 MONOD M O D E L C ° F I X E D X RUN IDENTIFICATION Cx ( g m / 1 ) ( g m / d 1 ) D ( h r " 1 ) HOLD-UP TIME (hours) S S L R u n N o . 2 S e c t i o n 3 . 4 1.15 3.19 3.92 1 .73 2.58 4.18 4.16 4.67 4.19 4.57 0.0048 0.0098 0.0141 0.00397 0.01064 207 102 71 252 94 M o s t 1i k e 1 y h o l d - u p t i m e 1 0 0 CONTINUED 163 Table 40 (Continued) S E C O N D O R D E R M O D E L C ° F I X E D X RUN IDENTIFICATION ( g m / l ) Cn ( g m / d 1 ) D ( h r - 1 ) HOLD-UP TIME (hours) S S L R u n N o . 3 S e c t i o n 3 . 4 1.145 3.535 4.135 1.80 2.755 5.0 5.0 5.0 5.0 5.0 0.0355 0.0525 0.0285 0.0271 0.0251 28 19 35 37 40 M o s t 1i k e 1 y h o l d - u p t i m e 40 S E C O N D O R D E R M O D E L C ° R E L A X E D x RUN IDENTIFICATION Cx ( g m / l ) Cn ( g m / d 1 ) D ( h r - 1 ) HOLD-UP TIME (hours) S S L R u n N o . 4 S e c t i o n 3 . 4 1 .155 3.84 4.15 1.78 2.6 5.0 5.0 5.0 5.0 5.0 0.0187 0.0435 0.0207 0.0214 0.0193 53 23 48 47 52 M o s t 1i k e 1 y h o l d - u p t i m e 50 NOTES: I. F e e d s t r e a m n u t r i e n t c o n c e n t r a t i o n o f C ° = 1 0 . 0 g m / d I was u s e d . 2 . T h e o p t i m u m h o l d - u p t i m e was d e f i n e d i n S e c . 5 . 4 . 3 . T h e m o s t l i k e l y h o l d - u p t i m e i s a c o m p r o m i s e b e t w e e n t h e m e a n h o l d - u p t i m e a n d a c o n s e r v a t i v e ' s a f e ' h o l d - u p t i m e s . 164 Herbert, Elsworth and T e l l i n g [80] from data on the fermen-tation of Aerobacter cloacae in a 2 0 - l i t r e continuous s t i r r e d tank fermentor, report K' values of 0.0123 gm/l; that i s , of the same order of magnitude as Monod's re s u l t s . However when their data is re-computed using non-linear least squares techniques with bacterial c e l l concentrations as the depen-dent variable, K.' is found to be 0.343 gm/l which is at least closer to the K' values found in this work. Ramanathan and Gaudy [82] have reported K' values of 0.106 gm/l from transient studies of continuous culture of mixed bacterial cultures in sewage treatment. Again, when their raw data are re-calculated using sounder s t a t i s t i c a l analyses, K' is found to be much larger, of the order of 2.33 gm/l. Why do workers in this f i e l d consciously find too low K.' values? O r i g i n a l l y , Monod adopted his model from Michaelis-Menten enzyme k i n e t i c s , where sing l e , r e l a t i v e l y simple, reaction systems were studied and K' was found to be quite small which was t h e o r e t i c a l l y sound for simple enzyme systems [65]. However, for bacterial growth pro-cesses, we no longer have these simple well defined systems. There are very many consecutive and competing enzyme reac-tions which occur both inside and outside the bacterial c e l l . The function of the parameter K1 i s no longer s t r a i g h t -forward as i t is in enzyme k i n e t i c s , but is an empirical 165 average of the effects of a great many parameters. Thus K' may serve only as a measure of the skew in the bacterial growth curve. Of course, in this work we do not know the absolute values of K or K'. Our estimated value of approximately 100 gm/l for K must be multiplied by the concentration of the actual unknown l i m i t i n g nutrient in the yeast extract. A true comparison with reported K. values of others cannot therefore be made u n t i l the actual l i m i t i n g nutrient(s) has been i d e n t i f i e d and measured. Chapter 6 CONCLUSIONS Ammonium based spent s u l f i t e liquor i s as s a t i s -factory as calcium based spent s u l f i t e liquor for growing Propionibacterium freudenreichii and production of vitamin B 1 2 . Simple steam stripping of the liquor to remove the sulfur dioxide to a concentration below 200 mg/l would be s u f f i c i e n t pre-treatment to permit Propionibacteria growth; pr e c i p i t a t i o n of the 1igno-sulfonates is not necessary. The six carbon sugars — mannose, glucose, and galactose — are u t i l i z e d by the bacteria. The pentoses — xylose and arabinose — are not u t i l i z e d . The pH range from 6.0-7.5 was found to be sa t i s f a c t o r y for Propionibacteria growth and vitamin B 1 2 synthesis. Yields of the fermentation products - vitamin B 1 2, dry bacterial c e l l u l a r material and propionic and acetic acids based upon the nutrients sugar and yeast extract were found to be constant over the ranges studied in this work. Numerical values of these y i e l d c o e f f i c i e n t s are summarized in Section 4.0. 166 1 6 7 The amount of the supplementary nutrient, yeast extract, required to produce recoverable amounts of the products (bacterial c e l l s , organic acids and vitamin B i 2 ) is excessive. Propionibacteria require much of this expensive supplementary nutrient compared with their r e l a t i v e l y inex-pensive requirements for carbohydrate. The savings to be gained by replacing the carbohydrate in the Propionibacteria fermentation medium with spent s u l f i t e liquor sugars are very small compared to the cost of the supplementary nutrient in the fermentation medium. Attempts to replace yeast extract with smaller amounts of other nutrients such as beef extract, peptone, etc., were not successful. In estimating the parameters of and discriminating between various mathematical models to describe the batch kinetic data of this work, i t was found that the method used to determine the parameters of these models was extremely important. The only acceptable model parameter estimation method was found to be dir e c t non-linear least squares e s t i -mation with bacterial c e l l concentration, the variable with the largest error associated with i t , as the dependent v a r i -able. The other parameter estimation methods investigated, l i n e a r i z a t i o n with time as the dependent variable and working on the rate equations, were not s a t i s f a c t o r y . 168 The model of Monod, with three adjustable parameters, which were estimated using non-linear techniques, was found to represent the batch kinetic data of this work s a t i s f a c -t o r i l y . A second order autocatalytic-substrate l i m i t i n g model with two adjustable parameters was a close second. Using these two models to predict the optimum hold-up time for a single continuous s t i r r e d tank fermentor gave results ranging from 40 hours for the second order model to 125 hours for the Monod model. This c l e a r l y shows the hazards of extrapolating batch fermentation data to continuous systems. Chapter 7 R E C O R D A T I O N S FOR FURTHER WORK 7 .1 SSL-Propionibacteria-Vi tamin B i 2 Process A cheaper replacement for yeast extract as supple-mentary nutrient should be investigated. Among those materials which should be investigated are f i s h plant wastes, soy bean meal (or an extract), or even pre-growing yeast on the same spent s u l f i t e liquor for this purpose. The high requirement for yeast extract could be reduced i f that component (or complex of components) which is "the l i m i t i n g nutrient" of the yeast extract could be i d e n t i f i e d and supplied cheaply. Some of these suspected components are the amino acids glycine and asparagine and the growth factors pantothenic acid, b i o t i n and thiamine. Since i t has been noted by Riley et a l . [22] and Berry and Bullerman [51] that a low oxygen level in the f i r s t 80-95% of the Propionibacteria growth i s essential for high vitamin B i 2 y i e l d s , i t is recommended that more careful control of the oxygen concentration in this stage 169 170 of the fermentation be investigated. This investigation could be begun by measuring the actual oxygen concentrations in the fermentor ( i f possible) at diff e r e n t rates of nitrogen purging and comparing this to ultimate yields of vitamin. 7.2 Kinetic Modeling and Continuous Fermentation In order to test and/or extend the kinetic models developed for batch fermentation, i t is recommended that continuous fermentation of Propionibacteria be attempted. It should also be possible to severely test these models at high bacterial c e l l concentrations by recycling enriched centrifuged bacterial c e l l s back to the feed to a continuous fermentor. 7.3 U t i l i z a t i o n of SSL to Produce a Commercial  Microbiological Product Because of the large amount of non-carbohydrate nutrient required to support the growth of Propionibacteria, i t is recommended that some other bio-synthetic products which can be produced by "carbohydrate intensive" micro-organisms be investigated for their s u i t a b i l i t y for growth on spent s u l f i t e liquor. Among these microbiological products 171 c e l l u l o s e s , proteases and g i b b e r e l l i c acid are suggested. A b r i e f discussion of the nature of these products, the microorganisms that produce them and the problems l i k e l y to be encountered in their production is presented as Appendix IV. REFERENCES [1] Libby, C.E. (ed.). 1962. Pulp and Paper Science and Technology, Vol. I, Pulp, McGraw-Hill, Toronto. [2] Wilber, CG. 1969. The B i o l o g i c a l Aspects of Water P o l l u t i o n , Thomas, S p r i n g f i e l d , 111. [3] Wiley, A.J. 1955. The Microbiology of S u l f i t e Liquor, Monograph No. 15, pp. 226-254. [4] Frenczi , S. and Z. Angew. 1912. Chem., Vol. 25, p. 2088. Taken from Johnsen [5]. [5] Johnsen, B. and R.W. Hovey. 1919. U t i l i z a t i o n of Waste S u l p h i t e L i q u o r , Kings Printer, Ottawa. [6] The Sun, Dec. 21, 1970, Vancouver yippie newspaper. [7] Watson, C.A. 1959. "Alcohol Porduction from Spent S u l f i t e Liquor," Forest Products Journal, Vol. IX, No. 3, pp. 25A-28A. [8] Merck and Co., Inc. 1958. Vitamin B i 2 , Merck and Co., Ltd., Rahway, N.J. [9] Williams, W.L., A.V. S t i f f e y , and T.H. Jukes. 1956. "Microbiological and Chick Assay of Vitamin B i 2 in Feed Supplements and Other Natural Products," J. Agr. Food Chem., Vol. 4, pp. 364-367. [10] Prevot, A.R., A. Turpin, and P. Kaiser. 1967. Les B a c t e r i e s Anaerobies, Dunod, Paris. 172 173 [11] F i e l d , M.F., and H. C. Lichst e i n . 1957. "Factors Affecting the Growth of Propionibacteria," J. B a c t e r i o l o g y 3 Vol. 73, pp. 96-99. [12] Lim, P.G. Nov. 19, 1968. U.S. Patent 3, 411, 991. Taken from Noyes [13]. [13] Noyes, R. 1969. Vitamin B 1 2 Manufacture, Noyes Development Corp., Park Ridge, N.J. [14] Greenberg, D.M. (ed.). 1967. M e t a b o l i c Pathways, 3rd e d i t i o n , Academic Press, N.Y. [15] Altman, P.L. and D.S. Dittmer. 1968. "Metabolism," B i o l o g i c a l Handbooks, Federation of Am. Soc. for Experimental Biology, Bethesda, Md. [16] Lehninger, A.L. 1956. B i o e n e r g e t i c s , Benjamin, N.Y. [17] Aiba, S., A.E. Humphrey, and N.F. M i l l i s . 1956. Biochemical Engineering, Academic, N.Y. [18] Rhem, H.-J. 1967. I n d u s t r i e l l e Mikrobiologie, Springer-Verlag, B e r l i n . [19] M i l l e r , M.W. 1961. The P f i z e r Handbook of Microbial M e t a b o l i t e s , McGraw-Hill, N.Y. [20] Perlman, D. 1959. "Microbial Synthesis of Cobamides," Adv. in Appl. Microbiology, Vol. 1. [21] Perlman, D., J.M. Barnett, and P.W. Jackson. 1961. Cobamides Synthesized by Propionibacterium Species, European Symposium on Vitamin B i 2 and I n t r i n s i c Factor, Hamburg. [22] Riley, P.B., D.R. Jackson, and P.A. Savage. 1961. Production of Vitamin B\z and an Analogue by Continuous Fermentation, SCI Monograph No. 12, Continuous Culture of Micro-organisms, SCI, London. 174 [23] Nishikawa, M. 1968. Fermentation of S u l f i t e Liquor, M.A.Sc. Thesis, University of B r i t i s h Columbia. [24] Nishikawa, M. , R.M.R. Branion, K.L. Pinder, and G.A. Strasdine. 1970. "Fermentation of Spent S u l f i t e Liquors to Produce Acetic Acid, Propionic Acid and Vitamin B i 2 , " Pulp and Paper Mag. of Can., Vol . 71 , No. 3, p. 159. [25] Mueller, J.C. 1970. "Fermentative U t i l i z a t i o n of Spent S u l f i t e Liquor; A Review and Proposal," Pulp and Paper Mag. of Can., Vol. 71, No. 22, pp. 72-76. [26] Rydholm, S.A. 1965. P u l p i n g Processes, Interscience, N.Y. [27] Butler, W.J. 1949. Pulp and Paper Mag. of Can., Vol. 50, No. 11, p. 108. [28] Wenzl , H.F.J. 1965. S u l f i t e Pulping Technology, Lockwood, N.Y. [29] Amos, D. 1945. "The Preparation of 3,4,5,-tri-methylphenethy1 amine from Eucalypt Lignin," New Zealand J. of Pharmacy, Vol. 64, p. 529. [30] Kure, A.P. 1957. "Ethyl Alcohol from Spent Waste Liquor," Can. J. of Chem. Eng., Vol. 35, pp. 86-90. [31] Dalhgren, E.H. 1964. "A Pulp M i l l s Approach to the U t i l i z a t i o n of Spent S u l f i t e Liquor," J. Wat. P o l l . Fed., Vol. 36, pp. 1543-1545. [32] Wiley, A.J., J.M. Halderby, and L.P. Huges. 1951. "Food Yeast from S u l f i t e Liquor," I & EC, Vol. 43, p. 1702. [33] Reusser, F. , J.F.T. Spencer, and H.R. Sallans. 1958. "Protein and Fat Content of Some Mushrooms Grown in Submerged Culture," App. M i c r o b i o l o g y , Vol. 6, pp. 1-4. 175 [34] Reusser, F., J.F.T. Spencer, and H.R. Sallans. 1958. "Tricholoma nudum as a Source of Microbiological Protein," Vol. 6, pp. 5-8. [35] American Type Culture C o l l e c t i o n . 1968. Catalogue of S t r a i n s , 8th edition , •Rockvilie , Md. [36] Stanier, R.Y, M. Doudoroff, and E.A. Adelberg. 1963. The M i c r o b i a l World, 2nd ed i t i o n , Prentice-Hall H a l l , Englewood C l i f f s , N.Y., pp. 322-324. [37] Guttman, H.N. 1963. "Vitamin B 1 2 and Cogeners," in Kavanagh [39], pp. 527-550. [38] Skeggs, H.R. 1963. "Lacto b a c i l l u s leichmannii Assay for Vitamin B 1 2," in Kavanagh [39], pp. 551-565. [39] Kavanagh, F. (ed.). 1963. A n a l y t i c a l M i c r o b i o l o g y , Academic Press, N.Y. [40] Casey, J.P. 1960. "Pulp and Paper Chemistry and Technology," 2nd e d i t i o n , Vol. 1, P u l p i n g and B l e a c h i n g , Interscience, N.Y. [41] Sawyer, C.N. and P.L. McCarty. 1967. Chemistry f o r S a n i t a r y Engineers, 2nd ed i t i o n , McGraw-Hill Toronto. [42] Every thesis you pick up has a number 42 reference in i t . [43] Neronova, N.M., N.D. IErysalimskii , and A.I. Anchurova. 1962 "Obrazovanie Vitamina B i 2 Propionionovokislimi bakteriyami v zavicimosti ot u s l o v i i K u l ' t i v i r o -vaniya," M i k r o b i o l o g i y a , Vol. 31 (in Russian), pp. 203-208. [44] Neronova, N.M., S.I. Ibragimova, and N.D. IErusalimskii. 1967. "Effect of the Propionate Concentration on the S p e c i f i c Growth Rate of P. Shermanii." English translation from M i k r o b i o l o g i y a , Vol. 36, No. 3, pp. 404-409. 176 [45] Oil Paint anal Drug Reporter, Sept. 29, 1 969. [46] Neronova, N . M . and N.D. IErusalimskii. 1959. Mikro-b i o l o g i y a , Vol. 28, p. 647. [47] Neronova, N.M. and S.I. Ibragimova. 1969. "Production of V o l i t i l e Acids and Vitamin B i 2 by P. Shermanii Growing in Continuous Cultures with an Excess of Lactate," trans, from Russian M i k r o b i o l o g i y a , Vol. 38, No 1 3, p. 420. [48] Vorob'eva, L.I. and N.A. Baranova. 1966. "Stimulating Effect of Mycobacterium luteum on Vitamin B i 2 Production by Propionibacteria ," trans, from M i k r o b i o l o g i y a , Vol. 35, No. 2, pp. 249-252. [49] Edwards, V.H. 1969. "The Recovery and P u r i f i c a t i o n of Biochemicals ," Adv. in Appl. Microbiol., Vol. 11 , pp. 159-210. [50] Bullerman, L.B. and E.C. Berry. 1966. "Use of Cheese Whey for Vitamin B i 2 Production: I. Whey Solids and Yeast Extract Levels," Appl. M i c r o b i o l . , Vol. 14, No. 3, pp. 353-355. [51] Berry, E.C. and L.B. Bullerman. 1966 Whey for Vitamin B i 2 Production Precursor and Aeration Levels," Vol. 14, No. 3, pp. 356-357. "Use of Cheese II. Cobalt, Appl. Microbiol., [52] Bullerman, L.B. and E.C. Berry. 1966. "Use of Cheese Whey for Vitamin B i 2 Production: III Growth Studies and Dry-weight A c t i v i t y , " Appl. M i c r o b i o l . Vol. 14, No. 3, pp. 358-360. [53] Shaposhnikov, V.N. and L.I. Vorob'eva. 1963. "Develop-ment of Propionic Acid Bacteria and Synthesis of Vitamin B X 2 in Synthetic and Natural Media," Transl. from M i k r o b i o l o g i y a , Vol. 32, No. 2, pp. 204-208. [54] Speedie, J.D. and G.W. Hull. 1960. U.S. Patent 2,951, 017. Also B r i t . Pat. 829,232, 1960. From Chem. Abstr., Vol. 54, 15821n and 25561g. 177 [55] Nomine, G. and L. Penasse. 1961. French Patent 1,264,016. From Chem. Abstr., Vol. 56, 9236b. [56] Martin, M.E., M. Wayman, and G. Graf. 1961. "Fermen-tation of Sulphite Waste Liquor to Produce Organic Acids," Can. J. of M i c r o b i o l . , Vol. 7, pp. 341-346. [57] Sudarsky, J.M. and R.A. Fisher. 1957. U.S. Patent 2,816,857. From Mervyn and Smith [58]. [58] Mervyn,L. and E.L. Smith. 1957. "The Biochemistry of Vitamin B i 2 Fermentation," Prog, i n Ihd. M i c r o b i o l . , Vol. 5, p. 151. [59] Fredrickson, A.G., D. Ramkrishna, and H.M. Tsuchiya. 1967. " S t a t i s t i c s and Dynamics of Procaryotic Cell Populations," Mathematical B i o s c i e n c e s , Vol. 1 , pp. 327-374. [60] Ramkrishna, D., A.G. Fredrickson, and H.M. Tsuchiya. 1967. "Dynamics of Microbial Propagation: Models Considering Inhibitions and Variable Cell Composition," B i o t e c h , and Bioeng., Vol. 9, pp. 129-170. [61] Tsuchiya, H.M., A.G. Fredrickson, and R. A r i s . 1966. "Dynamics of Microbial Cell Populations," Adv. in Chem. Eng., Vol. 6, p. 125. [62] Edwards, V.H. and CR. Wilke. 1968. "Mathematical Representation of Batch Culture Data," B i o t e c h . , and Bioeng., Vol. 10, pp. 205-232. [63] Monod, J. 1949. "The Growth of Bacterial Cultures," Ann. Rev. of Microbiology, Vol. 3, p. 371. [64] Monod, J. 1950. "La Technique de Culture Continue: Thdorie et Applications," Annales de I ' i n s t i t u t Pasteur, Vol. 79, p. 390. See also: Monod, J . , Ann. L'inst. Past., Vol. 68, 1942, p. 444; Vol. 71, 1945, p. 37; Vol. 69, 1943, p. 179. 178 [65] L a i d l e r , K.J. 1958. The Chemical K i n e t i c s , of Enzyme A c t i o n , Oxford. [66] Bray, H.G. and K. White. 1966. K i n e t i c s and Thermo-dynamics in Biochemistry3 Academic, N.Y. [67] Higgins, J. Sept. 1966-June, 1967. "The Theory of O s c i l l a t i n g Reactions," from ACS, "Applied Kinetics and Chemical Reaction Engineering," American Chemical Society, Washington, D.C; a c o l l e c t i o n of reprints from I & EC. [68] Perret, C.J. 1960. "A New Kinetic Model of a Growing Bacterial Population," j. Gen. M i c r o b i o l . , Vol. 22, pp. 589-617. [69] Neidhardt, F.C. and B. Magasanik. 1960. "Studies on the Role of Ribonucleic Acid in the Growth of Bacteria," Biochem. Biophys. Acta., Vol. 42, pp. 99-116. [70] Kjeldgaard, N.O. 1961. "The Kinetics of Ribonucleic Acid and Protein Formation in S. typhumurum During the Transition Between Different States of Balanced Growth," Biochem. Biophys. Acta., Vol. 49, pp. 64-76. [71] Kjeldgaard, N.O., 0. Maaloe, and M. Schaechter. 1958. "The Transition between Different Physiological States During Balanced Growth of Salmonella typhrnuriurn," J. Gen. M i c r o b i o l . , Vol. 19, pp. 607-616. [72] Himmelblau, D.M. 1970. "Process Analysis by S t a t i s -t i c a l Methods, Wiley, N.Y. [73] Draper, N.R. and H. Smith. 1966. A p p l i e d Regression A n a l y s i s , Wiley, N.Y. [74] Hunter, W.G. and R. Mezaki. 1964. "A Model Building Technique for Chemical Engineering Kinetics," AIChE. J., Vol. 10, No. 3, pp. 315-322. 179 [75] Marquardt, D.W. June, 1963. "An Algorithm for Least-squares Estimation of Non-linear Parameters," J. Soc. Indust. Appl. Math., Vol. 11, No.. 2, pp. 431-441. [76] Marquardt, D.W. 1959. "Solution of Non-linear Chemical Engineering Models," Chem. Eng. Prog., Vol. 55, No. 6, pp. 67-70. [77] B a l l , W.E. and L.C.D. Groenweghe. 1966. "Determina-tion of Best F i t Rate Constants in Chemical Kinetics," I S. EC Fundamentals, Vol. 5, No. 2, pp. 181-184. See also I & EC Fund., Vol. 6, No. 3, p. 475, 1967, for a discussion of this paper. [78] K i t t r e l , J.R., W.G. Hunter, and CC. Watson. 1965. "Non-linear Least Squares Analysis of Cata l y t i c Rate Models," AlchE.. J., Vol. 11, No. 6, pp. 1051-1057. [79] K i t t r e l , J.R., R. Mezaki , and C C Watson. 1965. "Estimation of Parameters for Non-linear Least Squares Analysis," I & EC, Vol. 57, No. 12, pp. 19-27. [80] Herbert, D., R. Elsworth, and R.C T e l l i n g . 1955. "The Continuous Culture of Bacteria: A Theoretical and Experimental Study," J. Gen. M i c r o b i o l . , Vol. 14, pp. 601-622. [81] Herbert, D. 1966. "A Theoretical Analysis of Contin-uous Culture Systems," SCI Monograph No. 12, Continuous Culture of Microorganisms, SCI, London. [82] Ramanathan, M. and A.F. Gaudy, J r . 1969. "Effect of High Substrate Concentration and Cell Feedback on Kinetic Behaviour of Heterogeneous Popula-tions in Completely Mixed Systems," B i o t e c h , and Bioeng., Vol. 11, pp. 207-237. [83] Box, G.E.P. and W.G. Hunter. 1962. "A Useful Method for Model-buiIding," Technometrics, Vol. 4, No. 3, pp. 301-318. 180 [84] Borrow, A., E.G. J e f f e r y s , R.H.J. Kessell , E.C. Lloyd, P.B. Lloyd, and I.S. Nixon. 1961. "The Metabo-lism of Gibberella f u j i k u r o i in S t i r r e d Culture," Can. J. Microbiol., Vol. 7, p. 227. [85] Borrow, A., S. Brown, E.G. J e f f e r y s , R.H.J. Kessell, E.C. Lloyd, P.B. Lloyd, A. Rothwell, B. Rothwell, and J.C. Swait. 1964. "The Kinetics of Metab-olism of Gibberella f u j i k u r o i in S t i r r e d Culture," Can. J. Microbiol., Vol. 10, p. 407. [8.6] A. Borrow., S. Brown, E.G. J e f f e r y s , R.H.J. Kessell , E.C. Lloyd, P.B. Lloyd, A. Rothwell, B. Rothwell, and J.C. Swait. 1964. "The Ef f e c t of Varied Temperature on the Kinetics of Metabolism of Gibberella f u j i k u r o i in S t i r r e d Cutlure," Can. J. Microbiol., Vol. 10, p. 445. [87] J e f f e r y s , E.G. 1970. in Adv. in Appl, pp. 283-316. "The Gibberellin Fermentation," M i c r o b i o l . , Vol. 13, a review, [88] Loginova, L.G. and Zh. Tashpulatov. 1965. "The Thermo-p h i l i c Fungus Aspergillus fumigatus Producing Active Cellulase," trans, from Russ. M i k r o b i o l o g i y a , Vol. 34, No. 2, pp. 258-264. [89] Mandels, Mary and E.T. Reese. 1957. Cellulase in Trichoderma virde Carbon Sources and Metals," J. Vol. 73, pp. 269-278. "Induction of as Influenced by Bacteriology , [90] American Chemical Society. 1969. "Cellulases and Their Applications," ACS Advances i n Chemistry, No. 95, ACS, Washington, D.C. [91] American Chemical Society. 1961 ACS Advances in Chemistry, D.C. "Gibberellins," No. 28, ACS. Washington, NOMENCLATURE A B B R E V I AT IONS ABS - Absorbance on Spectrophotometer. Also called t u r b i d i t y (TURB) and optical density (OD). ATCC - American Type Culture C o l l e c t i o n , Rockville, Md. , USA. BOD - Biochemical Oxygen Demand. A measure of the dis -solved oxygen depressing potential of a waste liquor. See Sec. 2.7. COD - Chemical Oxygen Demand. An estimate of BOD. See Sec. 2.7. DCW - Dry bacterial c e l l weight. A measure of bacterial growth. See Sec. 2.5 and Appendix I. DMBZ - 5,6-dimethylbenzimidazale. A precourser to vitamin B.i 2 . See Sec. 1.3. EMP - Emden-Meyerhof-Parnas pathway which breaks down hexoses to pyruvate. See Sec. 1.3. SSL - S u l f i t e spent liquor or spent s u l f i t e liquor. Also called waste s u l f i t e liquor. TURB - Turbidity. See ABS above. UOD - Ultimate Oxygen Demand. A high estimate of BOD. UOD i s always greater than BOD. See Sec. 2.7. 181 182 S Y M B O L S - ROMAN A - Kinetic parameter. Usually a function of u. See Sec. 5.1.4. Concentration of a product or a nutrient. Thus C is the concentration of bacterial c e l l s (x) A usually in units of gm/l cf - Correction factor used to correct t u r b i d i t y readings to dry c e l l weight. See Sec. 2.5 and Appendix I. c l . - 95% confidence l i m i t s . D - Dilution rate. Equivalent to r e c i p r i c a l of the hold-up time for a continuous s t i r r e d tank f e r -mentor. See Sec. 5.4. D is also used as the symbol for d i l u t i o n factor for t u r b i d i t y test in Sec. 2.5 and Appendix I. df - S t a t i s t i c a l number of degrees of freedom. F - F-test s t a t i s t i c . K or K' - Saturation parameter in Monod model. See Sec. 5.1.6. Pi - A general parameter in a mathematical model. See Sec. 5.2.3. q - Number of replications used for a single vitamin B 1 2 test sample. See Appendix II. R - Fermentation rate. Thus R = dC /dt = the A A fermentation rate based upon the appearance of bacterial c e l l s , usually in units of gm/l/hr. See Sec. 5.2.1 . 183 SS - S t a t i s t i c a l sum of squares. Thus SS £ i s the sum of squares of the residual error after some mathematical model has been applied to the experimental data. t or T - Time, usually in hours. - Yield c o e f f i c i e n t . Thus v x ^ n is the y i e l d of bacterial c e l l s (x) based upon yeast extract nutrient (n) usage. Y x / n = ~ A C x ^ A C n ' S e e Chapter 4. Generalized stoichiometric parameter. Usually a function of Y x^ n. See Sec. 5.1.3. SYMBOLS - GREEK A Difference operator. Thus AC = C . + 1 - C . 6 - Smaller difference operator. Thus 3C /9P. = r r X J ^x,i+l " u x , i 6C/SP. = P. . + 1 - P. . where 6P. is very small As in Appendix I. SUBSCRIPTS a - Acids products - Propionic and acetic acids together. E - Residual error. Thus see SS above. i or j - A general product, nutrient, or parameter. 184 m - A maximum such as Dm, the d i l u t i o n rate at which m maximum productivity takes place as in Section 5.4, n 1. A general nutrient or the nitrogen containing nutrient. 2. Yeast extract as nutrient. s - A general substrate and in par t i c u l a r sugar (or carbohydrate) as a substrate. v - Vitamin B 1 2 as a product. x - Bacterial c e l l s as a product. x/s - Used in conjunction with Y. Thus Y x y s i s the y i e l d of dry bacterial c e l l s based upon sugar usage as in Chapter 4. SUPERSCR I.PTS ° - I n i t i a l value. Thus C° is c e l l concentration at time = 0. - A value of a parameter estimated from a mathematical model. Thus C° is the i n i t i a l c e l l concentration A estimated from some model whereas C° i s the A measured value of the same quantity. - Average value. Thus C~ is an average c e l l concentration. «> or F or " - A steady state value, or a f i n a l value or the estimated value of a variable as time approaches i n f i n i t y . C^ = C^ = c"x. APPENDIX I CORRELATION OF DRY CELL WEIGHT AND TURBIDITY DRY CELL WEIGHT TEST The dry bacterial c e l l weigh of a given volume of fermentation broth was measured using the following procedure. 1. D r y s m a l l a l u m i n i u m w e i g h i n g d i s h e s ( a p p r o x . .13 g m . ) i n an o v e n a t I 0 0 ° C f o r t w o h o u r s ( o r m o r e ) . A f t e r c o o l i n g i n a d e s s i c a t o r , w e i g h t h e d i s h e s t o t h e n e a r e s t 0 . 0 0 0 0 1 g m . 2. P r e p a r e t h e s a m p l e ( s ) f o r t h e d r y c e l l w e i g h t t e s t u s i n g t h e t e c h n i q u e d e s c r i b e d b e l o w u n d e r " T u r b i d i t y T e s t . 1 1 3 . A d d e x a c t l y 10 m l . o f p r e p a r e d s a m p l e t o t h e a l u m i n i u m d i s h e s . 4. E v a p o r a t e m o s t o f t h e w a t e r f r o m t h e s a m p l e s o n a low t e m p e r a t u r e h o t p l a t e t a k i n g c a r e n o t t o s p l a s h 185 186 a n y o f t h e s a m p l e o u t o f t h e d i s h a n d n o t t o s c o r c h o r b u r n t h e s a m p l e by a l l o w i n g i t t o b e c o m e t o o d r y . 5 . H e a t t h e a l m o s t d r y s a m p l e i n I 0 0 ° C o v e n f o r a t l e a s t 2 i h o u r s ( s e e n o t e 3 . ) . 6 . W e i g h t o n e a r e s t 0 . 0 0 0 0 1 g m . a n d c a l c u l a t e d r y c e l l w e i g h t p e r u n i t v o l u m e . NOTES: 1. 1. Empty and sample-filled dishes must be cooled in an e f f i c i e n t dessicator before weighing. 2. It i s essential to place some drying agent (anhydrous CaSOiJ in the balance during weighing. The dried bacteria are extremely hydroscopic and w i l l pick up enough moisture so as to render the test meaningless. 3. Drying in the oven at 100° does not seem to affect the weight. Samples l e f t in for 24 hours did not show a s i g n i f i c a n t change in weight. TURBIDITY TEST 1. Centrifuge sample in IEC International Centrifuge-Universal Model UV centrifuge for 20 minutes at 4000 rpm. Save l i q u i d from this centrifugation for later sugar analysis. 187 2. Resuspend the c e l l s in d i s t i l l e d water using approximately the same amount of water that was removed by centrifugation. Recentrifuge the sample for 20 minutes at 4000 rpm. 3. Repeat wash step 2. (If the colour of the f i n a l suspension shows signs that a l l the spent s u l f i t e liquor and yeast extract were not removed, this step was repeated as many times as was necessary.) 4. Resuspend the c e l l s in d i s t i l l e d water. If the weight of the washed c e l l suspension d i f f e r s from the ori g i n a l sample weight, note both the original sample weight and the washed sample weight so that a correction be made to the t u r b i d i t y r e s u l t . This correction factor is ' f . 5. Dilute an appropriately sized aliquot of the washed sample, so that the t u r b i d i t y reading on the Bauch and Lomb Spectrometric 20 spectrophotometer (at a wavelength of 520 mu) f a l l s between 0.1 and 0.5 on the absorbance scale. Let the d i l u t i o n factor be D and the t u r b i d i t y reading be TURB. 188 * 6. F i l l at least half way, at least three cuvettes with the diluted sample. Measure the t u r b i d i t y (absorbance) at 520 my. The average of the three cuvettes is used unless one of the tests is obviously wrong in which case i t is di scarded. 7. Correct the t u r b i d i t y readings for non-linear-i t i e s in the dry c e l l vs. t u r b i d i t y relationship as explained below under "Dry Cell Weight from Turbidity," can calculate dry c e l l weight per unit volume. DRY C E L L WEIGHT FROM T U R B I D I T Y It was noted, p a r t i c u l a r l y at t u r b i d i t i e s above 0.4 on the absorbance scale, that at di f f e r e n t d i l u t i o n s , t u r b i d i t y readings on the same sample were not the same. In order to investigate this phenominon, one suspended c e l l sample was diluted 1, 1.5, 2, 3, 4 and 6 times. Let this d i l u t i o n factor be 'D'. The t u r b i d i t y readings are shown in Table 41. A c t u a l l y i n s t e a d o f t h e s t a n d a r d c u v e t t e s , K I M A K I c m . by 5 c m . m i c r o b i o l o g i c a l t e s t t u b e s w e r e u s e d f o r t h i s t e s t . I t was f o u n d t h a t t h e y w e r e b e t t e r t h a n t h e m o r e e x p e n s i v e s t a n d a r d c u v e t t e s . 189 Table 41 DETERMINATION OF DRY CELL WEIGHT VS. TURBIDITY RELATIONSHIP D TURB D x TURB 1.612 T TURB x D LN(f) 1.0 0.903 0.903 1.7852 0.57953 1.0 0.901 0.901 1.7891 0.5817 1.0 0.900 0.900 1.7911 0.5828 1.5 0.7025 1 .054 1.5298 0.4252 1.5 0.702 1 .053 1.5309 0.4258 1.5 0.700 1 .050 1.5352 0.4287 2.0 0.585 1.17 1.37778 0.3205 2.0 0.580 1.16 1.3897 0.3291 2.0 0.580 1.16 1.3897 0.3291 3.0 0.415 1 .245 1.2948 0.2583 3.0 0.418 1 .254 1.2855 0.2511 4.0 0.340 1 .36 1.1853 0.1699 4.0 0.340 1 .36 1.1853 0.1699 6.0 0.222 1 .332 1.2102 0.1908 6.0 0.255 1 .530 1.0536 0.0522 6.0 0.238 1 .428 1.1288 0.1212 1 9 0 When (TURB x D) is plotted against TURB as shown in Figure 3 0 , a straight li n e results with the intercept (TURB x D) = 1 . 6 1 2 at TURB = 0 . The physical significance of this is that at i n f i n i t e d i l u t i o n the tu r b i d i t y corrected for the d i l u t i o n factor w i l l be 1 . 6 1 2 . Assuming that 'TURB1, the t u r b i d i t y at i n f i n i t e d i l u t i o n is the correct t u r b i d i t y , then a correction factor cf can be defined by r f = TURB°° 1 . 6 1 2  C T TURB x D TURB x D In order to put this relationship into a more useful form, Log(cf) was plotted against TURB. The straight li n e shown in Figure 31 was obtained with the constraint Log(cf) = 1 . 0 at TURB = 0 . The t u r b i d i t y correction equation then becomes TURB°° = TURB x E X P ( A x TURB) = TURB x E X P ( 0 . 6 0 7 8 x TURB) How well does this relationship work? When 2 3 dry c e l l weights were plotted against corrected t u r b i d i t i e s as shown 191 I I I I I I 3 0.5 r.o TURBIDITY AT 520 m/A Figure 30. TURB x D vs. TURB Plot for Dry Cell Weight Test Correlation. TURBIDITY AT 520 m/x Figure 31. Log(cf) vs. TURB for Dry Cell Weight Correlation. 1 9 3 in Figure 3 2 , a nice straight l i n e ensued. The 'best' equa-tion of this l i n e , with the constraint at DCW = 0 ; TURB = 0 by DCW = 2 . 8 9 5 x TURB°° = 2 . 8 9 5 x TURB x E X P ( 0 . 6 0 7 8 x TURB) The approximate 9 5 % confidence l i m i t s on dry c e l l weight DCW measured/calculated this way are ± 0 . 5 gm/l. TURB x EXP( 0.6078 x TURB ) Figure 32. Dry Cell Weight vs. TURB x EXP(0.6078 x TURB) to Check Out Original Data Points. APPENDIX II VITAMIN B19 TEST The test used in this work for measurement of vitamin B i 2 is e s s e n t i a l l y the Lactobacillus leichamanii test outlined by Skeggs [38] and Guttman [37]. The spectro-photometry method developed by Nishikawa [23] was not used in this work because of reasons stated in the text in Section 2.6. In this appendix, the method used to assay for vitamin B a 2 w i l l be outlined, mainly by example, with special emphasis on the method used to calculate the confidence limits on the test r e s u l t s . The vitamin B i 2 test results of the run whose data are presented in Section 3.4 as SSL Run No. 4 w i l l be used to i l l u s t r a t e the technique. STANDARD CURVE PREPARATION The following procedure was done in t r i p l i c a t e . 195 196 1. 0, 0.02, 0.04, 0.06, 0.08, 0.1, 0.12, 0.16 * and 0.2 mug of standard vitamin B 1 2 were added to 20 ml culture tubes. 2. 5 ml of DIFCO-BACTO B 1 2 Assay Medium USP (Difco Laboratories, Detroit, Mich., America) prepared accord-ing to the instructions on the l a b e l . S u f f i c i e n t d i s t i l l e d water was added to each test tube to give a total l i q u i d level in each test tube of 10 ml. This procedure results in a vitamin B i 2 concentration in the culture tubes of 0.0, 0.002, 0.004, 0.006, 0.008, 0.01, 0.012, 0.016 and 0.02 mug/1. UNKNOWN SAMPLE PREPARATION In order to (a) insure that the Propionihacterium freudenreichii c e l l walls were broken down to release the vitamin, and to (b) convert the vitamin from the hydroxyl form to the more stable sulfato form [37,38], a 1 ml aliquot of each sample was added to 25 ml of pH = 4.5 citrate-phos-phate buffer prepared as follows: Na 2HP0 4 - 12.9 gm; c i t r i c S q u i b b R u b a m i n , l o t s 8F74I75 a n d 7F72722 w e r e d i l u t e d f o r t h i s p u r p o s e . 197 acid monohydrate - 11.47 gm; d i s t i l l e d water to 1 l i t r e to which 1 gm of sodium b i s u l f i t e was added just before use. The buffered samples were then autoclaved for 15 minutes at a steam pressure of 15 psig. 1. T h e u n k n o w n s a m p l e s w e r e d i l u t e d a s s h o w n i n T a b l e 42 t o a c o n c e n t r a t i o n b e l i e v e d t o be w i t h i n t h e t e s t r a n g e . 2 . 2 ml a n d 5 ml a l i q u o t s o f e a c h s a m p l e w e r e a d d e d t o a 20 ml c u l t u r e t u b e . F i v e r e p l i c a t i o n s w e r e d o n e f o r e a c h u n k n o w n s a m p l e . 3 . F i v e ml o f D I F C O - B A C T O B i 2 A s s a y M e d i u m a n d s u f f i c i e n t d i s t i l l e d w a t e r t o g i v e a t o t a l o f 10 ml o f l i q u i n e a c h c u l t u r e t u b e w e r e a d d e d t o e a c h c u l t u r e t u b e , a s w i t h t h e s t a n d a r d s . 4 . F i v e b l a n k s w e r e p r e p a r e d s i m i l a r l y t o t h e p r e p a r a t i o n . o f t h e s t a n d a r d s a n d t h e u n k n o w n s w i t h d i s t i l l e w a t e r s u b s t i t u t i n g f o r t h e u n k n o w n s a m p l e o r t h e s t a n d a r d v i t a m i n B i 2 . Table 42 VITAMIN B TEST - DILUTION FACTORS FOR UNKNOWNS SAMPLE GUESS OF B12 CONC. ( m g / l ) DILUTION FACTOR CORRECTION FOR CELL WASH* 96 hr. 0.5 13,000 : 1 0.997 216 hr. 2.0 65,000 : 1 0.999 240 hr. 2.0 65,000 : 1 0.982 Standard 100.0 2,500,000 : 1 -When e e l Is w e r e w a s h e d , t h e m a k e - u p w a t e r a d d e d was n o t e x a c t l y t h e s a m e a m o u n t a s was c e n t r i f u g e d o u t . T h i s c o r r e c t i o n f a c t o r c o r r e c t s f o r t h i s . S e e A p p e n d i x I f o r m e a s u r e m e n t o f d r y c e l l w e i g h t a n d t u r b i d i t y . CO oo 199 TEST ITSELF 1. A l l of the unknowns, blanks and the standards were capped with polyolefin culture caps which allow a i r to enter the culture medium but exclude stray microorganisms. The tubes were then s t e r i l i z e d by autoclaving with 15 psig steam for f i v e minutes. 2. One drop of inoculum, prepared according to the label of the DIFCO-BACTO Inoculum Broth USP used to grow the inoculum, was added a s e p t i c a l l y to a l l the culture tubes except the blanks. The microorganism used for these tests was Lactobacillus leichmanii ATCC 7830 [35], which was maintained between tests in DIFCO-BACTO B i 2 Culture Agar USP. 3. The unknowns, standards and blank culture tubes were incubated at 32°C for 40 hours. 4. After 40 hours incubation the absorbance of each culture tube (unknowns and standards) was determined at 600 my on the Bausch and Lomb Spectronic 20 using a 1 cm c e l l . The prepared blanks were used to zero the absorbance scale of the instrument. These absorbance measurements are tabulted in Table 43 for the standard curve and in Table 44 for the unknowns. 200 Table 43 ABSORBANCE READINGS FOR STANDARD CURVE A B S O R B A N C E AT 6 0 0 m y . B 1 2 IN SAMPLE ( m y g ) No. 1 No. 2 No. 3 AVERAGE 0.0 0.02 0.04 0.115 0.21 5 0.285 0.130 0.220 0.295 0.115 0.220 0.290 0.120 0.218 0.290 0.06 0.08 0.10 0.323 0.450 0.575 0.340 0.470 0.545 0.380 0.470 0.530 0.346 0.463 0. 550 0.12 0.16 0.20 0.540 0.570 0.660 0.520 0.600 0.680 0.560 0.580 0.630 0.540 0.583 0.657 Table 44 UNKNOWN SAMPLE READINGS ABSORBANCE AT 600 my. SAMPLE SIZE (ml) No. 1 No. 2 No. 3 No. 4 No. 5 AVERAGE q 96 hr 2 5 0.480 0.725 0.510 0.670 0.460 0.695 0.450 0.700 0.510 0.670 0.482 0.692 5 5 216 hr 2 5 0.365 0.595 0.350 0.585 0.368 0.590 0.362 0.600 0.365 0.362 0.592 5 4 240 hr 2 5 0.348 0.625 0.338 0.650 0.355 0.620 0.355 - 0.349 0.632 4 3 202 ANALYSIS OF THE DATA * Theoretically the standard curve plot of vitamin B i 2 concentration in the standards culture tubes vs. absor-bance should be l i n e a r . This straight l i n e , exhibited in Figure 33 is B 1 2 = -0.06524 + 0.3388 x ABSORBANCE ( H - l ) the slope of which i s s i g n i f i c a n t . The 95 per cent c o n f i -dence li m i t s of this model are given by c l . = ± 0.0412 where q i s the number of replications of the unknown sample to be tested in the assay. However, upon close examination of this 'straight' li n e in Figure 33, i t is obvious that there is a considerable lack of f i t in this model. The cause of this lack of f i t ~* B e e r - L a m b e r t l a w , e t c . 1 + q 27 1_ , (ABS - 0.4188) 0.81 (H-2) I I I I 0.2 0.4 0.6 0.8 A B S O R B A N C E AT 6 0 0 m/i. Figure 33. Vitamin B 1 2 Test Standard Curve - Linear Model. 204 is probably due to re-scattering and/or back-scattering of l i g h t in the concentrated milieu, similar to the problem(s) encountered in the bacteria t u r b i d i t y test in Appendix I. Let us, therefore, entertain a quadratic model, such as that shown in Figure 34. The equation of this line i s : B = -0.00267 + 0.4348 x (ABSORBANCE)2 (11-3) with confidence l i m i t s given by c. 1 . = ± 0.03153 This quadratic model has a smaller sum of squares of the residual error (3.659) than that of the linear model (SS = 6.251). It is also quite apparent from examination of the plotted points on the models in Figures 33 and 34 and from the confidence limits equations (11 -2) and (11-4) that the confidence limits of the quadratic model are narrower than those of the linear model. 27 (ABS2 - 0.1407) 2 0.5147 C H - 4 ) 205 CO ( I / B T / U J ) 31dSMVS NI 2,a NllrWllA Figure 34. Vitamin B 1 2 Standard Curve - Quadratic Model. 2 0 6 The vitamin B i 2 concentrations of the o r i g i n a l samples were calculated using the quadratic model and multiplied by the appropriate d i l u t i o n factors. These results are displayed in Table 45 for your reviewing pieasure. 207 Table 45 RESULTS OF VITAMIN B TEST USING QUADRATIC MODEL SAMPLE SIZE (ml) B i 2 (mg/l) 95% CONFIDENCE LIMIT 96 hr 2 0.637 0.103 5 0.533 0.055 216 hr 2 1 .762 0.497 5 1 .945 0.233 240 hr 2 1 .605 0.490 5 2.183 0.244 APPENDIX III REQUIREMENTS FOR A GOOD MICROBIOLOGICAL PRODUCT TO BE MADE FROM SSL 1. The microorganism should be able to u t i l i z e inorganic nitrogen rather than more complex nitrogen sources. If any growth factors are required they should be well de-fined and required in small amounts. The carbohydrate, or sugar requirement should account for a s i g n i f i c a n t portion of the raw material cost of the fermentation. 2. The product should be complex so that i t w i l l not be prof i t a b l e to synthesize i t from petro-chemical feed-stocks. 3. The microorganism should be able to metabolize a l l of the spent s u l f i t e sugars, or at least a l l the hexoses. 4. The product should have a wide market basis. That is to say, a product such as a livestock feed supplement 208 209 or a plant growth stimulant would have a widespread poten-t i a l , whereas a wart removing chemical would not be as useful. Conversely, i f a low volume specialty product were produced i t would be necessary to build a v e r s a t i l e multi-product plant producing a variety of products. 5. The by-products of the fermentation should not themselves be pol1utants,unless i t i s economically feas-ib l e to those products. Anaerobic fermentations nearly always produce large amounts of organic acids-acetic, l a c t i c , propionic etc. Aerobes, on the other hand, usually produce carbon dioxide (less than the amount of CO2 used by the trees used to produce the SSL from which i t was pro-duced). However, the aerobes produce more microorganism, which i f not used up by feeding i t to animals or a secondary fermentation process, w i l l become p o l l u t i o n . There i s also the expense of supplying oxygen to these aerobes. 6. The microorganisms should also, of course, be able to grow on spent s u l f i t e medium with a minimum of pre-treatment. Sulfur dioxide removal and pH adjustment are easy, whilst l i g n i n removal would be not only messy but also expensive. 210 7. It is desirable that the product of the f e r -mentation be a i n t r a - c e l l u l a r product so that the product would be e a s i l y separated from the l i g n i n and other 'impurities' in the spent s u l f i t e liquor. However, i f the whole broth could be e a s i l y transported and used (for example spraying on crops) an e x t r a - c e l l u l a r product would be quite acceptable. It is unlikely that any one microbiological product would s a t i s f y a l l these requirements. They should, never-theless, be considered in process s e l e c t i o n , p a r t i c u l a r l y the f i r s t three requirements. APPENDIX IV MICROBIOLOGICAL PRODUCTS WHICH COULD BE PRODUCED FROM SPENT SULFITE LIQUOR There is a large number of commercially useful microbiological products. Some of these products which are produced in large quantities are: vitamins and growth factors, a n t i b i o t i c s , plant growth regulators, steroids, a l k a l o i d s , pigments, enzymes, organic acids, amino acids, etc. From this large l i s t of products have been chosen two enzymes (cellulases and proteases) and one plant growth promotor (gibberel1in) as p o s s i b i l i t i e s to be looked at with respect to their s u i t a b i l i t y for production from spent s u l f i t e 1i quor. G I B B E R E L L I N ( S ) Gibberellins are a class of plant growth promoters which stimulate the inernode growth of many plants. They 211 212 also shorten germination times. Large scale experimentation into commercial scale uses of gibberellins has probably been limited by the u n a v a i l a b i l i t y of large amounts of g i b b e r e l l i n . Some of the effects of gibberellins on plant growth are discussed in an ACS [91] monograph. Gi b b e r e l l i c acid is produced in r e l a t i v e l y large concentrations by the fungus GIBBERELLA f u j i k u o r i (ATCC 12616). Borrow [84,85,86] et a l . have found that this fungus w i l l grow well on a medium of glucose, inorganic nitrogen, phosphate and trace minerals. They also found that the bulk of the g i b b e r e l l i n was produced during the stationery phase of growth when the stationary phase was brought on by the depletion of nitrogen. It is not known whether G. f u j i k u o r i w i l l metabolize galactose, mannose, xylose or arabanose but i t is known [87] that i t w i l l metabolize glu-cose, sucrose and g l y c e r o l . It i s , however, highly l i k e l y that G. f u j i k u r o i w i l l u t i l i z e the six carbon sugars of spent s u l f i t e liquor but, unfortunately, i t is unlikely that i t w i l l u t i l i z e the five carbon sugars. G. f u j i k u r o i requires oxygen (air) for growth so that the main metabolic product of carbon dioxide w i l l cause no disposal problems. Experiments by Borrow et a l . [85] have indicated that at high growth rates i t is probably oxygen that i s the l i m i t i n g nutrient. However, since 90% of 213 the production of g i b b e r e l l i n takes place after the growth of the fungus has ceased, supplying oxygen w i l l probably not be the major expense. G. f u j i k u r o i grows best in the pH range of 3.0 to 5.5. At the lower end of this range, at say 3.5, most other microorganisms w i l l not multiply, therefore contamination w i l l not be much of a problem. The main drawback in production of g i b b e r e l l i n is the long hold-up times required to build up high concentra-tions of g i b b e r e l l i n . The primary emphasis of any research program would be the optimization of the time required to attain maximum concentrations of g i b b e r e l l i c acid. Some variables which might be important are pH, oxygen tension, temperature, concentration of nutrients, concentrations of trace metals, inducers, etc. Another drawback in g i b b e r e l l i n production might be the separation of the product(s) from the fermentation medium which would contain a l o t of l i g n i n . Of course there is no a p r i o r i reason that the whole f i l t e r e d medium could not be used on crops thus solving simultaneously a s t i c k l y separation problem and a nasty spent fermentation liquor problem. (Also as good technocrats we could probably find b e n e f i c i a l uses for the l i g n i n in i t -- as a s o i l binder for example.) 214 C E L L U L A S E S Cellulases are enzymes which break down native cellulo s e to soluble cellulose (Ci c e l l u l a s e ) , then break down soluble c e l l u l o s e to cellobiose (C cellulase) and A f i n a l l y break cellobiose down into glucose ( c e l l o b i a s e ) . Cellulases could prove to be important in the future in pre-treatment of feeds before feeding to animals (including man) to break down cellulose to the more eas i l y digested glucose or in prehydrolysing cellulose wastes l i k e straw and bagasse prior to further processing. A cel l u l a s e ( p a r t i c u l a r l y C x) pre-soak of pulp wood before pulping may also prove to be b e n e f i c i a l . This seems to be ever so s l i g h t l y far fetched! To date research into the uses of cellulases has been limited by the a v a i l a b i l i t y of c e l l u l a s e preparations of known strength and s p e c i f i c i t y . There has been a recent ACS [90] publication which discusses at length the many p o s s i b i l i t i e s for c e l l u l a s e u t i l i z a t i o n . Two of the many microorganisms which produce cellulases in high concentrations are Myrothecum verrucaria (ATCC 9095) and Trichoderma viride (ATCC 1 3631 ). It is also possible to extract cellulases from cultures of Aspergillus fumigatus as was done by Loginova and Tashpulatov [88]. T. viride has been cultured on a medium consisting of glucose (5g/l), KH 2P0ij (2g/l) , and trace elements by Mandels and 215 Reese [89]. In their work cellul a s e was also produced from mediums with<glucose , cellobiose, lactose and cellul o s e as carbon sources. Arabinose, xylose, mannose, and galactose also supported growth of T. viride but, unfortunately, did not support the production of c e l l u l a s e . It may be that cellu l a s e i s induced by some type of c e l l u l o s e - l i k e poly-glucose either added to the medium or synthesized by the fungus from glucose. Unfortunately the l i t e r a t u r e i s sparse in information on the oxygen requirements and optimum pH for best growth of cellu l a s e producing microorganisms. PROTEASES Proteases are enzymes which break down proteins (and enzymes). Spe c i f i c proteases have found a wide variety of uses. For example, thermo-stable alkaline proteases were once widespread in washing powders before Nader; other proteases are used to treat burns, tenderize meats, coagulate milk to make cheese and yogurt. Further potential uses of proteases include enzyme treatment of sewage to break down organic wastes. Two of the many microorganisms which produce high concentrations of proteolytic enzymes are Streptomyces thermophi1 us, ATCC-19282 , a high temperature horseshit 216 bacteria and Aspergillus oryzae (ATCC 9362). There i s l i t t l e or no work on the minimum nutrient requirements of these protease producing microorganisms. They w i l l grow well on glucose and probably on galactose and mannose. The induc-tion of protease may require large amounts of complex protein. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0059174/manifest

Comment

Related Items