UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effect of process conditions on productivity and glycosylation of cystatin C in P. pastoris and DNA… Pritchett, Jason 2003

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

Item Metadata

Download

Media
831-ubc_2003-0287.pdf [ 15.53MB ]
Metadata
JSON: 831-1.0058680.json
JSON-LD: 831-1.0058680-ld.json
RDF/XML (Pretty): 831-1.0058680-rdf.xml
RDF/JSON: 831-1.0058680-rdf.json
Turtle: 831-1.0058680-turtle.txt
N-Triples: 831-1.0058680-rdf-ntriples.txt
Original Record: 831-1.0058680-source.json
Full Text
831-1.0058680-fulltext.txt
Citation
831-1.0058680.ris

Full Text

The Effect of Process Conditions on Productivity and Glycosylation of Cystatin C in P. pastoris and DNA Microarray Analysis of Amino Acid Limitations in S. cerevisiae Culture By Jason Pritchett Bachelor of Applied Science, University of Waterloo, 1998 A thesis submitted in partial fulfillment of the requirements for the degree of Master of Applied Science In The Faculty of Graduate Studies Department of Chemical and Biological Engineering We accept this thesis as conforming to the required standard The University of British Columbia April 2003 © Jason Pritchett, 2003 In presenting t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I further agree that permission f o r extensive copying of t h i s thesis f o r s c h o l a r l y purposes may be granted by the head of my department or by h i s or her representatives. I t i s understood that copying or p u b l i c a t i o n of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of CWow^ca\ &to\o^c&\ rj\HCJ\A.(>er\* The U n i v e r s i t y of B r i t i s h Columbia Vancouver, Canada Date A p r s \ 300.3 A B S T R A C T Human cystatin C is a cysteine-proteinase inhibitor with several potential therapeutic applications. A recombinant variant of cystatin C with two potential sites for N-linked glycosylation (Nakamura, Ogawa et al. 2000) was selected for expression in the yeast Pichia pastoris. Glycosylation has been shown to significantly improve the heat stability and activity of recombinant cystatin C (Nakamura, Ogawa et al. 1998). When using standard fermentation protocols however, most cystatin C is not glycosylated. Thus, the effects of induction pH, temperature, and nitrogen sources on cystatin C productivity and glycosylation were examined in 250 mL shake flasks and 2-litre bioreactors. The pH and temperature were studied over the ranges of 5.2 - 6.8, and 21 - 35°C respectively. Nitrogen sources examined include ammonium hydroxide, peptone and amino acid supplements. Nitrogen source was the most significant parameter. The maximum cystatin C productivity and glycosylation was obtained under conditions of 20 g-L'1 peptone, 20 g-L"1 amino acid mix, and 0 g-L'1 ammonium hydroxide resulting in 13.6 nmol-gDCW'h"1 cystatin C with 30% glycosylated species, a five-fold increase from standard fermentation conditions. In a separate study using S. cerevisiae, DNA microarray analysis was used to monitor gene expression changes resulting from amino acid limitations in fermentation media. Experiments were performed to examine leucine and glutamine limitations imposed on batch cultures. Recent literature results from Gasch (Gasch, Spellman et al. 2000) and Natarajan (Natarajan, Meyer et al. 2001) were also re-analyzed. The results show that both leucine and glutamine biosynthesis genes were differentially expressed as much as 13-fold under complete amino acid starvation and genes involved in leucine and histidine biosynthesis were significantly up-regulated under mild leucine limitation and histidine starvation with 3AT (3-aminotriazol). Additional amino acid starvation experiments revealed that leucine starvation resulted in a 2-fold up-regulation of the leucine biosynthesis genes LEU1, LEU2 and LEU4. However, changes in gene expression for glutamate biosynthesis genes under glutamic acid starvation resulted in no distinguishable trends. ii T A B L E OF CONTENTS ABSTRACT II TABLE OF CONTENTS III LIST OF TABLES VII LIST OF FIGURES IX ACKNOWLEDGEMENTS XI INTRODUCTION 1 1.1 R E C O M B I N A N T PROTEIN PRODUCTION 1 1.2 C Y S T A T I N C 2 1.3 PROCESS D E V E L O P M E N T A N D OPTIMIZATION 2 1.4 THESIS L A Y O U T 3 2 LITERATURE REVIEW 4 2.1 INTRODUCTION 4 2.2 G L Y C O S Y L A T I O N 4 2.2.1 Types of Oligosaccharides 5 2.2.2 Glycoprotein Heterogeneity ; 9 2.2.3 Analytical Methods 10 2.2.4 Host Selection 11 2.2.4.1 Bacterial 11 2.2.4.2 Yeasts 11 2.2.4.3 Insects 12 2.2.4.4 Mammalian 12 2.2.5 Process Factors Influencing Glycosylation 13 2.2.5.1 A m m o n i u m / p H 13 2.2.5.2 Temperature Growth State 16 2.2.5.3 Carbon Source 18 2.2.5.4 Culture Time 20 2.2.5.5 Other Factors 21 2.3 PlCHIA PASTORIS AS A HOST 22 iii 2 . 4 P R O C E S S C O N F I G U R A T I O N A N D O P T I M I Z A T I O N 2 3 2 . 5 P R O C E S S O P T I M I Z A T I O N U T I L I Z I N G M I C R O A R R A Y S 2 5 2.5.1 DNA microarrays 25 2.5.2 Process Optimization and Microarrays 27 2.5.3 Gasch, Spellman et al. Results 27 2.5.4 Natarajan, Meyer et al. Results 29 2.5.5 Biosynthetic Pathways 30 3 MATERIALS AND METHODS 33 3 . 1 I N T R O D U C T I O N 3 3 3 . 2 C E L L S T R A I N S 3 3 3.2.1 Pichia pastoris 33 3.2.2 Saccharomyces cerevisiae 34 3 . 3 C U L T U R E M E D I A 3 4 3.3.1 Culture Plates and Inoculums 34 3.3.2 Screening Experiments 35 3.3.3 2-litre Bioreactor Optimization 39 3.3.4 Statistical Analysis 41 3.3.5 Microarray Experiments 42 3 . 4 E X P E R I M E N T A L P R O T O C O L S 4 3 3.4.1 Screening Experiments 43 3.4.2 2-litre Bioreactor Optimization 44 3.4.3 Microarray Experiments 45 3 . 5 A N A L Y T I C A L M E T H O D S • 4 6 3.5.1 Cell Density 46 3.5.2 Cystatin C Assay 47 3.5.3 SDS gel electrophoresis 49 3.5.4 ENDOF1 Treatment 50 3.5.5 Gel Imaging Analysis ••: 51 3.5.6 RNA isolation 51 3.5.7 Microarray probe preparation and hybridization 52 3.5.8 Microarray data acquisition and analysis 52 iv 3 . 6 S T A T I S T I C A L D E S I G N A N D A N A L Y S I S 5 4 3.6.1 Screening Experimental Design 54 3.6.2 2-Litre Bioreactor Experimental Design 55 4 RESULT AND DISCUSSION 56 4 . 1 I N T R O D U C T I O N 5 6 4 . 2 R E C O M B I N A N T C Y S T A T I N C S T A B I L I T Y 5 6 4.2.1 Introduction 56 4.2.2 Results and Discussion 58 4 . 3 F E E D S T R A T E G Y 5 9 4.3.1 Introduction 59 4.3.2 Results 60 4.3.3 Discussion 63 4 . 4 E F F E C T O F P R O C E S S C O N D I T I O N S O N C Y S T A T I N C P R O D U C T I V I T Y A N D G L Y C O S Y L A T I O N 6 5 4.4.1 Screening Experiments 65 4.4.1.1 Background 6 5 4 . 4 . 1 . 2 Design Details 6 5 4 . 4 . 1 . 3 Growth Results 6 6 4 . 4 . 1 . 4 Statistical Analysis 6 8 4 . 4 . 1 . 5 Growth Discussion 7 2 4 . 4 . 1 . 6 Productivity Results 7 4 4 . 4 . 1 . 7 Statistical Analysis 7 5 4 . 4 . 1 . 8 Productivity Discussion 7 9 4 . 4 . 1 . 9 Glycosylation Results 8 1 4 . 4 . 1 . 1 0 Glycosylation Discussion 8 3 4.4.2 2-Litre Bioreactor Optimization 84 4 . 4 . 2 . 1 Background 8 4 4 . 4 . 2 . 2 Design Details 8 4 4 . 4 . 2 . 3 Growth Results 8 5 4 . 4 . 2 . 4 Productivity Results 8 6 4 . 4 . 2 . 5 Statistical Analysis 8 7 v A A.2.6 Productivity Discussion : 92 4.4.2.7 Glycosylation Results 96 4.4.2.8 Statistical Analysis 100 4.4.2.9 Verification of Glycosylated Cystatin C 105 4.4.2.10 Glycosylation Discussion 106 4.5 M I C R O A R R A Y P R O C E S S O P T I M I Z A T I O N R E S U L T S 111 4.5.1 Introduction Ill 4.5.2 Gasch, Spellman et al. Results Re-analyzed 112 4.5.3 Natarajan, Meyer et al. Results Re-analyzed. 114 4.5.4 Amino Acid Limitation Experiments 115 4.5.5 Microarray Discussion 119 4.5.5.1 Gasch, Spellman et al. Discussion 119 4.5.5.2 Natarajan, Meyer et al. Discussion 120 4.5.5.3 Leucine Limitation Experiments 121 4.5.5.4 Glutamate Limitation Experiments 123 4.5.5.5 Microarray Discussion Summary 123 4.5.5.6 Considerations for Future Microarray Experiments 124 5 CONCLUSIONS 125 6 RECOMMENDATIONS 128 7 ABBREVIATIONS 130 8 REFERENCES 133 9 APPENDIX 145 vi LIST OF T A B L E S T A B L E 1: A M M O N I U M L I T E R A T U R E 14 T A B L E 2: T E M P E R A T U R E L I T E R A T U R E 17 T A B L E 3 : G L U C O S E L I T E R A T U R E . . 19 T A B L E 4: C U L T U R E T I M E L I T E R A T U R E 20 T A B L E 5 : F E E D S T R A T E G I E S L I T E R A T U R E 24 T A B L E 6: T A B L E O B T A I N E D F R O M G E N O M I C E X P R E S S I O N P R O G R A M S I N T H E R E S P O N S E O F Y E A S T C E L L S T O E N V I R O N M E N T A L C H A N G E S ( G A S C H , S P E L L M A N E T A L . 2000) 28 T A B L E 7 : YPD M E D I A C O M P O S I T I O N 34 T A B L E 8: C O M P O S I T I O N O F B A C T O P E P T O N E - T Y P I C A L A N A L Y S I S ( D I F C O ) 35 T A B L E 9 : BMGH A N D BMMFf M E D I A C O M P O S I T I O N 36 T A B L E 10 : YNB M E D I A C O M P O S I T I O N 37 T A B L E 11: S C R E E N I N G E X P E R I M E N T S C E N T R A L C O M P O S I T E D E S I G N 38 T A B L E 12: 2 - L I T R E B I O R E A C T O R E X P E R I M E N T S F A C T O R I A L D E S I G N 39 T A B L E 13 : B A S A L S A L T S M E D I A C O M P O S I T I O N 40 T A B L E 14: PTM1 T R A C E S A L T S M E D I A C O M P O S I T I O N 41 T A B L E 15 : M I N I M A L M E D I A C O M P O S I T I O N 42 T A B L E 16 A M I N O A C I D S D R O P O U T M I X C O M P O S I T I O N 43 T A B L E 17 : F E E D S T R A T E G Y E X P E R I M E N T A L D E S I G N 45 T A B L E 18: C Y S T A T I N C A S S A Y R E A G E N T S 48 T A B L E 19: SDS PAGE E L E C T R O P H O R E S I S R E A G E N T S 49 T A B L E 20: ANOVA F O R S C R E E N I N G D E S I G N G R O W T H R E S U L T S 69 T A B L E 21: M O D E L A N A L Y S I S F O R S C R E E N I N G D E S I G N G R O W T H R E S U L T S 69 T A B L E 22 : ANOVA F O R S C R E E N I N G D E S I G N P R O D U C T I V I T Y R E S U L T S 76 T A B L E 23 : M O D E L A N A L Y S I S F O R S C R E E N I N G D E S I G N P R O D U C T I V I T Y R E S U L T S 77 T A B L E 24: ANOVA F O R 2 - L I T R E B I O R E A C T O R P R O D U C T I V I T Y R E S U L T S 88 T A B L E 25 : M O D E L A N A L Y S I S F O R 2 - L I T R E B I O R E A C T O R P R O D U C T I V I T Y R E S U L T S 89 T A B L E 26: G L Y C O S Y L A T I O N R E S U L T S F O R 2 - L I T R E B I O R E A C T O R E X P E R I M E N T S 99 T A B L E 27: ANOVA F O R 2 - L I T R E B I O R E A C T O R G L Y C O S Y L A T I O N R E S U L T S 101 T A B L E 28 : M O D E L A N A L Y S I S F O R 2 - L I T R E B I O R E A C T O R G L Y C O S Y L A T I O N R E S U L T S 101 vii T A B L E 2 9 : H I S T I D I N E B I O S Y N T H E S I S G E N E S ( N A T A R A J A N , M E Y E R E T A L . 2 0 0 1 ) 1 1 5 T A B L E 3 0 : L E U C I N E B I O S Y N T H E S I S G E N E S ( N A T A R A J A N , M E Y E R E T A L . 2 0 0 1 ) 1 1 5 T A B L E A 1 1 4 6 T A B L E A 2 1 4 6 T A B L E A 3 1 4 7 T A B L E A 4 1 4 7 T A B L E A 5 1 4 8 T A B L E A 6 1 5 0 T A B L E A 7 1 5 2 T A B L E A 8 1 5 4 T A B L E A 9 1 5 5 T A B L E A 1 0 1 5 6 T A B L E A U 1 5 7 T A B L E A 1 2 1 5 9 T A B L E A 1 3 1 5 9 viii LIST OF FIGURES F I G U R E 1 : M A M M A L I A N N - L I N K E D O L I G O S A C C H A R I D E P R O C E S S I N G 7 F I G U R E 2 : O L I G O S A C C H A R I D E S T R U C T U R E S 9 F I G U R E 3 : M I C R O A R R A Y C O N S T R U C T I O N A N D P R O B E P R E P A R A T I O N 2 7 F I G U R E 4 : L E U C I N E B I O S Y N T H E S I S S C H E M A T I C 3 1 F I G U R E 5 : G L U T A M A T E B I O S Y N T H E S I S S C H E M A T I C 3 2 F I G U R E 6 : W C W vs. D C W C A L I B R A T I O N C U R V E 4 7 F I G U R E 7 : C Y S T A T I N C S T A B I L I T Y A S A F U N C T I O N O F S T O R A G E T I M E ( F I L E S E T A L . 2 0 0 0 ) 5 7 F I G U R E 8 : C Y S T A T I N C S T A B I L I T Y A S A F U N C T I O N O F S T O R A G E T I M E 5 8 F I G U R E 9 : C E L L D E N S I T Y R E S U L T S F O R F E E D S T R A T E G Y E X P E R I M E N T S 6 1 F I G U R E 1 0 : C Y S T A T I N C Y I E L D R E S U L T S F O R F E E D S T R A T E G Y E X P E R I M E N T S 6 1 F I G U R E 1 1 : M E T H A N O L C O N S U M P T I O N R E S U L T S F O R F E E D S T R A T E G Y E X P E R I M E N T S 6 3 F I G U R E 1 2 : C E L L D E N S I T Y R E S U L T S F O R S C R E E N I N G D E S I G N 6 7 F I G U R E 1 3 : B o x P L O T F O R S C R E E N I N G D E S I G N G R O W T H R E S U L T S 6 8 F I G U R E 1 4 : R E S P O N S E S U R F A C E F O R S C R E E N I N G D E S I G N G R O W T H R E S U L T S 7 0 F I G U R E 1 5 : R E S P O N S E S U R F A C E F O R S C R E E N I N G D E S I G N G R O W T H R E S U L T S 7 1 F I G U R E 1 6 : P L O T O F R E S I D U A L S F O R S C R E E N I N G D E S I G N G R O W T H M O D E L 7 2 F I G U R E 1 7 : C Y S T A T I N C P R O D U C T I V I T Y R E S U L T S F O R S C R E E N I N G D E S I G N 7 5 F I G U R E 1 8 : B o x P L O T F O R S C R E E N I N G D E S I G N P R O D U C T I V I T Y R E S U L T S 7 6 F I G U R E 1 9 : R E S P O N S E S U R F A C E F O R S C R E E N I N G D E S I G N P R O D U C T I V I T Y R E S U L T S 7 8 F I G U R E 2 0 : P L O T O F R E S I D U A L S F O R S C R E E N I N G D E S I G N P R O D U C T I V I T Y M O D E L 7 9 F I G U R E 2 1 : S D S P A G E G E L R E S U L T S F O R S C R E E N I N G D E S I G N 8 2 F I G U R E 2 2 : C E L L D E N S I T Y R E S U L T S F O R B I O R E A C T O R E X P E R I M E N T S 8 5 F I G U R E 2 3 : C Y S T A T I N C Y I E L D R E S U L T S F O R B I O R E A C T O R E X P E R I M E N T S 8 6 F I G U R E 2 4 : C Y S T A T I N C P R O D U C T I V I T Y R E S U L T S F O R B I O R E A C T O R E X P E R I M E N T S 8 7 F I G U R E 2 5 : B o x P L O T F O R B I O R E A C T O R P R O D U C T I V I T Y R E S U L T S 8 8 F I G U R E 2 6 : R E S P O N S E S U R F A C E F O R B I O R E A C T O R P R O D U C T I V I T Y R E S U L T S 9 0 F I G U R E 2 7 : R E S P O N S E S U R F A C E F O R B I O R E A C T O R P R O D U C T I V I T Y R E S U L T S 9 1 F I G U R E 2 8 : P L O T O F R E S I D U A L S F O R B I O R E A C T O R P R O D U C T I V I T Y M O D E L 9 2 F I G U R E 2 9 : S D S P A G E G E L R E S U L T S F O R B I O R E A C T O R E X P E R I M E N T S 9 7 ix F I G U R E 30: S D S P A G E G E L R E S U L T S F O R B I O R E A C T O R E X P E R I M E N T S 98 F I G U R E 3 1 : P E A K A N A L Y S I S F O R B I O R E A C T O R G E L S 99 F I G U R E 32 : B o x P L O T F O R B I O R E A C T O R G L Y C O S Y L A T I O N R E S U L T S 100 F I G U R E 33 : R E S P O N S E S U R F A C E F O R B I O R E A C T O R G L Y C O S Y L A T I O N R E S U L T S 102 F I G U R E 34: R E S P O N S E S U R F A C E F O R B I O R E A C T O R G L Y C O S Y L A T I O N R E S U L T S . . . . 103 F I G U R E 35 : R E S P O N S E S U R F A C E F O R B I O R E A C T O R G L Y C O S Y L A T I O N R E S U L T S 104 F I G U R E 36 : R E S I D U A L S P L O T F O R B I O R E A C T O R G L Y C O S Y L A T I O N M O D E L 105 F I G U R E 37 : S D S P A G E G E L S F O R E N D O F l A N A L Y S I S 106 F I G U R E 38 : G E N E E X P R E S S I O N F O R L E U C I N E S Y N T H E S I S G E N E S ( G A S C H , S P E L L M A N E T A L . 2000) 112 F I G U R E 39 : G E N E E X P R E S S I O N F O R G L U T A M A T E S Y N T H E S I S G E N E S ( G A S C H , S P E L L M A N E T A L . 2 0 0 0 ) 113 F I G U R E 4 0 : G E N E E X P R E S S I O N F O R G L U T A M A T E S Y N T H E S I S G E N E S ( G A S C H , S P E L L M A N E T A L . 2000) 113 F I G U R E 41: G R O W T H C U R V E S F O R L E U C I N E L I M I T A T I O N E X P E R I M E N T S 116 F I G U R E 42 : G R O W T H C U R V E S F O R G L U T A M I C A C I D L I M I T A T I O N E X P E R I M E N T S 117 F I G U R E 43 : G E N E E X P R E S S I O N F O R L E U C I N E L I M I T A T I O N E X P E R I M E N T S 118 F I G U R E 44 : G E N E E X P R E S S I O N F O R G L U T A M I C A C I D L I M I T A T I O N E X P E R I M E N T S 118 x A C K N O W L E D G E M E N T S I would first like to thank Dr. Susan Baldwin for informing me of this exciting opportunity. Her knowledge and support has helped make this project a success. As my supervisor she helped me build a strong background in the biological sciences and has always provided me with valuable feedback. Next I would like to thank Dr. Jamie Piret for his support and guidance with much of the microarray work. His lab was always open and I am especially thankful for being included in his weekly lab meetings where I obtained exposure to other aspects of biotechnology and biochemical engineering making my experience at UBC invaluable. Thanks also go to Dr. Masahiro Ogawa in the Food Science Department. His expertise in protein expression in Pichia pastoris, helped bring me up to speed quickly. Clive Glover was also a very valuable resource and helped me learn the techniques involved in microarray preparation and analysis. I would also like to thank Gary Leznicki for his expert opinions in P. pastoris fermentation and for the use of his methanol sensor in the initial experiments. Dr. Nakamura, Dr. Nakai, and Dr. Ogawa also deserve thanks for their efforts in originally engineering the mutant human cystatin C used in my research efforts. I would also like to Thank Dr. Ross MacGillivray for his supreme expertise in aspects of biochemistry and for offering his lab for the biochemical analysis work. Mike Page also deserves recognition for his help with analytical techniques and troubleshooting. Finally I would like to thank the most important people in my life who have supported me long before this project began. These behind the scenes people are my family and friends. Special mention goes out to my wicked awesome wife Leah. xi Introduction 1.1 Recombinant Protein Production The field of biotechnology has seen the advent of virus vaccines (polio, measles, mumps), to monoclonal antibodies (cancer, HIV), and finally to complex structured glycoproteins (tPA, y-interferon, EPO) (Kretzmer 2002). Many of these existing therapeutic proteins, such as insulin, interferons (a,P,y) and tPA are manufactured using recombinant DNA technology. Interferon-beta (IFN-beta), which is used for the treatment of Multiple Sclerosis, is produced by this technology and has a market value of more than two billion US dollars globally. Development of screening and production systems to produce these products is necessary to obtain high titers and reduce production costs. Products such as tPA are secreted by mammalian cells at low concentrations (0.01 mg purified tPA from one uterus)(Cartwright 1992). However, with recombinant technology CHO cells can be produce tPA at 50 mg-109CHO cells'day"1 (Kretzmer 2002). For each new recombinant therapeutic, development of a specific production process is necessary for maximal expression. For example, recombinant anticoagulant (hirudin, rHVaLys47) originally expressed at levels less than 10 mg-L"1 in S. cerevisiae was optimized to obtain levels as high as 600 mg-L"1 (Mendoza-Vega, Sabatie et al. 1994). However, this result was only attained after experimenting with many variables including expression vector, host strain, feed strategy, culture medium, culture conditions and scale-up. The Variables that must be considered for recombinant protein expression include the host organism (bacteria, yeast, insect, plant or mammalian), the production method (batch, fed-batch, continuous or perfusion), and the process settings (growth medium, temperature, DO, pH and feed regime). Glycoproteins are one type of complex product produced by recombinant DNA technology. Human cystatin C is an example of a protein that has been mutated specifically to introduce sites for potential glycosylation to improve physical characteristics such as temperature stability (Nakamura, Ogawa et al. 2000). 1 1.2 Cystatin C Cystatin C is a cysteine protease inhibitor. In the body, cystatin C regulates papain-like proteases, and prevents uncontrolled protein degradation and tissue damage (Abrahamson, Barrett et al. 1986). Potential therapeutic applications for cystatin C, were reviewed in a previous thesis (Files 2000). Among its potential applications, cystatin C is involved in defence against bacteria and viruses (Bjorck, Akesson et al. 1989; Bjorck, Grubb et al. 1990) and inhibition of tumour metastasis (Cox, Sexton et al. 1999). There is also potential to use cystatin C in the treatment of sciatica, a neurological disorder (Grubb 1988). With several potential therapeutic applications, there is interest in developing a process for over expression of cystatin C. Recombinant cystatin C has been expressed in bacteria (Escherichia coif) at levels of 0.3 - 1 g-L"1 (Dalboge, Jensen et al. 1989; Berti, Ekiel et al. 1997). In E. coli, recombinant cystatin C is expressed intracellularly and thus a recovery loss of approximately 50% could be expected (Abrahamson, Dalboge et al. 1988). In the yeast Pichia pastoris, a mutated cystatin C was expressed extracellularly at approximately 0.72 g-L"1 (Files 2000). The mutated cystatin C that was expressed in the research by Files, and which was also used in this research, was a mutant form with two potential N glycosylation sites (Nakamura, Ogawa et al. 2000). These mutations have been shown to improve the stability and activity of cystatin C (Nakamura, Ogawa et al. 1998). In order to study the effects of process conditions on glycosylated proteins expressed by P. pastoris, cystatin C was used as a model protein. Under standard fermentation conditions (Mut+, MeOH-feed, pH = 6, T = 30°C) very little protein was expressed in either of the glycosylated forms (single or double) (Files 2000). One of the objectives of this project was to identify process conditions that increase the amount of glycosylated protein produced using the P. pastoris system. 1.3 Process Development and Optimization Optimization of bioprocesses is often done using statistical factorial design and analysis. However, even then it is only possible to examine a few variables at a time, making process development costly and time consuming. For example, a full factorial design 2 examining 3 variables, at 3 different levels, with replicates, requires 54 fermentation runs. Microarray technology has emerged in the past ten years and has the potential to accelerate the development process. Microarrays are used to monitor gene expression changes and can capture a snap shot of the physiological state of the culture. For example, Gasch and colleagues (Gasch, Spellman et al. 2000) used microarrays to monitor the effects of several environmental stresses on S. cerevisiae cells, which showed stereotypical expression changes for a unique set of genes. It is our hypothesis that changes in gene expression patterns can be used to identify unknown media deficiencies. If so, DNA microarrays may be used to accelerate media and fermentation optimization dramatically. This hypothesis was tested in a pilot experiment measuring gene expression changes in S. cerevisiae grown under amino acid limitations. S. cerevisiae was used because arrays are available for this organism that include the entire genome and are relatively inexpensive. 1.4 Thesis Layout This thesis is divided into 5 chapters. The second chapter identifies the applicable technology and terminology with respect to the topic and provides a thorough review of the background information and existing literature. Chapter 3 describes the fermentation procedures, analytical techniques, and cell lines used for experimentation. The results and discussion of the experiments performed are presented in Chapter 4. Chapter 5 includes a summary of the key findings of this research. Recommendations are presented in Chapter 6. A list of nomenclature and cited references follow Chapter 5. Finally, the Appendices provide details on media formulations and tables of raw data. 3 2 Literature Review 2 . 1 Introduction This chapter provides background on glycosylation and recombinant protein production. Information on glycosylation pathways, analytical techniques for glycosylation, and process factors affecting glycosylation are presented. An understanding of host system selection and the P. pastoris expression system are also introduced. A review of these topics is important to achieve an understanding of the experiments that are described in Chapter 4. 2 . 2 Glycosylation Glycosylation is one of the key post-translational modifications performed by eukaryotic cells. It involves the addition of carbohydrate residues to specific sites on the peptide backbone of proteins as they are processed in the endoplasmic reticulum (ER) and Golgi apparatus of the cell (Jenkins, Parekh et al. 1996). This processing results in the production of glycoproteins. The carbohydrate structures on glycoproteins play an integral part in the final configuration and functionality of the protein. Glycoproteins account for a very high percentage of the potential and existing therapeutic proteins (Goochee and Monica 1990). Addition of glycan structures to proteins can influence many important protein attributes including protein folding, secretion, stability, biological activity, in vivo clearance rates, antigenicity, and immunogenicity (Jenkins 1995). The importance of glycosylation related to therapeutic recombinant proteins is evident, and the FDA and CPMP are now insisting on more complex characterization and carbohydrate analysis than in the past (Jenkins and Curling 1994). Clearly it is necessary to develop an understanding of glycosylation by characterizing carbohydrate structures and identifying the fermentation or cell culture conditions affecting glycosylation, so that glycosylation can be better understood and controlled. To date most of the research in process optimization with respect to glycosylation has been directed at mammalian, insect and murine cell culture. Optimization work with these host systems has shown that there are many factors influencing glycosylation including: ammonia and glucose concentration; culture time; temperature and the growth phase. Yeast research has only 4 focused on characterization of glycosylated variants in terms of site location and micro and macroheterogeneity. Optimization efforts to examine the effect of process conditions on glycosylation in yeast have not been examined thoroughly. Yeasts have played a significant role in the development and production of recombinant therapeutic proteins including human insulin, tPA, interferons, monoclonal antibodies, HIV antigens etc. The ability to perform glycosylation makes them attractive in terms of the similarities to higher order eukaryotes (Cereghino and Cregg 2000). However, there are differences in carbohydrate composition from other eukaryotic hosts which may limit the potential for some therapeutic applications (Cereghino and Cregg 1999). Future advances may continue to increase their efficacy as a host system for therapeutic applications. 2.2.1 Types of Oligosaccharides Most eukaryotic hosts are capable of both O-linked and N-Linked glycosylation. O-linked glycosylation involves the addition of carbohydrates to the hydroxyl groups of threonine or serine on the protein backbone (Cereghino and Cregg 2000). The presence of serine or threonine does not guarantee glycosylation but rather indicates the potential for glycosylation (Jenkins, Parekh et al. 1996). O-linked oligosaccharides can be composed of simple sugars such as fucose, glucose, mannose, GlcNAc, and GalNAc, or more complex sugars such as Gaipi-3GalNAc (Jenkins 1994; Bjoern, 1991; Harris, 1991). N-linked glycosylation is carried out when carbohydrates are added to the amide nitrogen of asparagine residues when found in the consensus sequence Asn-Xaa-Thr/Ser (Bretthauer and Castellino 1999). Again, the presence of this sequence only represents a potential site for glycosylation (Jenkins, Parekh et al. 1996). An example of typical oligosaccharide processing reactions for N-linked glycosylation is presented in Figure 1. These reactions require the activity of a variety of enzymes to perform the addition and trimming of sugars from the basic protein structure. A source of carbohydrates is also necessary to supply the sugars essential for glycosylation. The sugars and enzymes required for glycosylation are identified in Figure 1. Another compound required for glycosylation is the lipid dolichol, which is used to transfer the carbohydrate structure to 5 the protein backbone. These compounds including enzymes, sugars and the lipid dolichol represent the key components necessary for glycosylation in eukaryotes. 6 ENDOPLASMIC RETICULUM Figure 1 : Mammalian N-linked oligosaccharide processing. A potential pathway of mammalian N-linked oligosaccharide processing. The reactions are catalyzed by the following enzymes: (1) oligosaccharyltransferase, (2) a-glucosidase I, (3) a-glucosidase II, (4) ER a(l,2) mannosidase, (5) Golgi a-mannosidase I, (6) N-acetylglucosaminyltransferase I, (7) Golgi ot-mannosidase II, (8) N-acetylglucosaminyltransferase II, (9) a(l,6) fucosyltransferase, (10) (3(1,4) galactosyltransferase, (11) a(2,3) sialyltransferase. The symbols are: •, N-acetylglucosamine (GlcNAc); O, mannose (Man); •, glucose (GIu); O, fucose (Fuc); •, galactose (Gal); •, sialic acid (NeuAc). Dol-P-P is dolichyldiphosphate. The co-substrates for reactions 6,8,9,10 and 11 are the energized forms of the monosaccharides where: UDP is uridine diphosphate, GDP is guanosine diphosphate and CMP is cytidine monophosphate (Godchee, Gramer et al. 1991). 7 The structure of N-linked glycosylation for all eukaryotes is based on the core structure Glc3ManaGlcNAc2, which is built from sequential addition of GlcNAc, mannose (Man), and glucose (Glc) residues onto the lipid dolichol using phosphorylated intermediates (Rosenwald, Stoll et al. 1990; Lehrman 1991; Orlean 1992). This synthesis begins in the cytoplasm and is then processed by the ER and Golgi where the core structure can be trimmed by glucosidases and mannosidases (Jenkins and Curling 1994). Further processing by exoglycosidases and glycotransferases may allow subsequent addition of outer structures to the trimmed core. These added structures may include GlcNAc, galactose, sialic acid or fucose (Cumming 1992; de Vries and van den Eijnden 1992; Goochee 1992; Kobata 1992). Complex oligosaccharide structures, especially those including sialic acid and terminal galactose residues, have been shown to play an important role in therapeutic application by affecting in vivo half life, antigenicity, and immunogenicity (Jenkins and Curling 1994). Typical glycan structures include oligomannoside, complex (bi-antennary), hybrid structures, and O-linked glycans (Figure 2) (Jenkins, Parekh et al. 1996). 8 ASN Oligomannose Complex (biantennary) ASN it Hybrid O -gly c o sylati o n Figure 2 : Oligosaccharide structures. Common oligosaccharide structures found on glycoproteins: O Man, • NeuAc, • Gal, -k GalNAc, • GlcNAc, • Fucose. 2.2.2 Glycoprotein Heterogeneity The glycosylation of recombinant proteins usually results in glycan structures with a range of molecular weights (Jenkins and Curling 1994). This variation is defined as heterogeneity. It can be further classified as macroheterogeneity, the variability of site occupancy of glycans, or microheterogeneity, the variability in glycan structures at specific sites. Heterogeneity is a concern especially for therapeutic applications. Therapeutic products must maintain a consistent level of glycosylation heterogeneity from batch to batch for clinical approval and this is highly scrutinized by the FDA (Jenkins and Curling 1994). The glycan heterogeneity of these proteins can be determined by a variety of analytical techniques. 9 2.2.3 A nalytical Met ft ods Glycan analysis can range from simple detection to complex characterization of detailed structure. The level of sophistication used for analysis is dependant on the protein of interest, its potential applications, and the glycosylation impact on product performance. Electrophoresis, such as glycan analysis by SDS PAGE, is used to determine glycan variability based on molecular weight of the oligosaccharide species (Lin, Immormino et al. 2001; Saito, Usui et al. 2002). In this analysis N-linked glycans can be released from the protein backbone using an enzyme such as Endo-(3-N-acetylglucosaminidase H (endo H), or peptide-N-glycosidase F (PNGase F) to identify the different glycosylated variants. This analysis allows for determination of macroheterogeneity of glycan structures. Imaging software can be combined with this analysis to compare relative peak intensities. This type of imaging analysis and quantification has been used in applications for protein expression in host systems including S. cerevisiae and E. coli (Zhang, Zhou et al. 1991; Pilon, Yost et al. 1996; Kerry-Williams, Gilbert et al. 1998; Cooley and Mishra 2000). Although this technique is simple to perform, it does not provide detailed information about microheterogeneity or site location of the attached glycans which may be necessary for proper characterization. Liquid chromatography can also be used for glycan analysis (Montesino, Garcia et al. 1998; Kalidas, Joshi et al. 2001). The location of glycosylation sites on the protein and the structure of each glycan group can be found by first cleaving the protein at specific sites and then processing the fragments through reverse-phase HPLC (Jenkins and Curling 1994). From this type of analysis it is possible to deduce the type of glycan structures but not the site location of the attached glycans. NMR is used to find the specific glycan structure using a database of NMR spectra (van Kuik, Hard et al. 1992). NMR is capable of assigning unambiguous structure to completely unknown oligosaccharides (Jenkins and Curling 1994). This analysis provides a complete characterization of the attached glycan; however, the cost of analysis and processing time are more significant that those techniques previously mentioned. 10 Mass spectrometry (MS) has been used to determine the glycan structure of many glycoproteins. It can also be used to quantify site occupancy based on the conversion of Asn to Asp (IDa difference) after PNGase F treatment (Jenkins and Curling 1994). This analysis also provides a complete characterization of the attached glycan, however again the cost of analysis and processing time are limiting factors for this type of analysis. The selection of the appropriate glycan analysis tool is very specific to the type of information that is desired such as site occupancy, site location and specific glycan structure. Other factors, which may influence the type of analysis required, include the protein's physical and chemical characteristics and the host expression system. 2.2.4 Host Selection 2.2.4.1 Bacterial Most bacterial systems do not have the ability to glycosylate. A few bacterial strains have shown O-glycosylation (Stimson, Virji et al. 1995). Nevertheless this lack of glycosylation ability limits the potential for bacteria to produce functional glycoproteins. 2.2.4.2 Yeasts Yeast can carry out N-linked and O-linked glycosylation. Most studied yeasts undergo glycan trimming, resulting in a core structure of MangGlcNAc2. This core structure undergoes only mannose addition of the ct-l,2-linked, a-l,3-linked, or a-l,6-linked branching (Bretthauer and Castellino 1999). S. cerevisiae, the most studied yeast, is known to add glycan structures with a very high number of mannose residues, often greater than 50 residues, known as hyperglycosylation. Hyperglycosylation is associated with antigenicity and rapid clearance from the blood by the liver, and is a problem for therapeutic protein production (Cereghino and Cregg 2000). P. pastoris in most cases does not hyperglycosylate, and does not produce glycans with a-l,3-linkages both of which make many yeast produced recombinant proteins unsuitable for human pharmaceuticals (Romanos, Scorer et al. 1992). 11 The similar glycosylation group, chain lengths observed in proteins expressed by P. pastoris and mammalian host systems make this yeast a potential organism for screening new recombinant glycoprotein candidates. These unique glycosylation features make P. pastoris a more attractive host than S. cerevisiae for the production of human therapeutics. Although P. pastoris lacks the ability to add more complex glycans such as fucose, sialic acids, and glucose, there is still potential through genetic modification to introduce the appropriate glycosyltransferases to obtain more desirable glycoforms. This type of genetic modification has already been shown successful in mammalian cell lines, and an a-2,6-sialyltransferase gene was successfully transfected and expressed in P. pastoris showing the potential in yeast hosts as well (Chotigeat, Chayanunnukul et al. 2000). 2.2.4.3 Insects Insect systems have the ability to glycosylate. Most research shows that N-glycosylation is limited to simple oligomannoside structures. Only a few systems have shown more complex glycan structures. It has been shown that some insect cell lines like Trichoplusio ni, TN-368 and BTI-Tn-5Bl-4 lines, have added terminal galactose and sialic acids (Davis and Wood 1995). These specific systems may have greater value for therapeutic applications. 2.2.4.4 Mammalian Host systems that more closely resemble human cells are desirable. When attempting to produce a recombinant human protein, one might expect human cells to be the ideal host. This is not necessarily the case. The transformation event required to form a stable cell line may alter the glycosylation profiles (Yamashita, Koide et al. 1989). The most common mammalian systems used include mouse (hybridomas) and hamster (CHO, BHK) cells. Like most mammalian cells, mouse cells differ from human cells in that they express the enzyme a-l,3-galactosyltransferase that adds Gakxl,3-Galpl,4-GlcNAc to secreted glycoproteins (Jenkins, Parekh et al. 1996). This feature invokes a human immune response to the proteins. Mouse NSO and rat YO myeloma are two specific cell 12 lines that do not have this feature and thus induce only mild human immune response (Jenkins, Parekh et al. 1996). Mouse cell lines also add NeuGc (N-glycolylneuraminic acid), which results in rapid removal of the molecule from circulation and an immune response (Jenkins, Parekh et al. 1996). Most hamster cell lines have been inactivated for a-l,3-galactosyltransferase and make only low levels of NeuGc, making them more attractive than mouse cell lines (Smith, Larsen et al. 1990), yet they also have drawbacks. They lack a functional a-2,6-sialyltransferase enzyme and hence synthesize only a-2,3 linked terminal sialic acids. However, CHO cell and mouse cell lines have been genetically modified to express the desired glycosyltransferases (Jenkins, Parekh et al. 1996). Mammalian systems are usually not the most effective host in terms of protein productivity. However these systems are essential for the production of many proteins due to their ability to perform the complex post-translational protein processing that cannot yet be performed by lower eukaryotes and prokaryotes. 2.2.5 Process Factors Influencing Glycosylation 2.2.5.1 Ammonium / pH The presence of extracellular ammonia is common in both cell culture and fermentation processes. In typical mammalian cell culture processes, glutamate is utilized as a nitrogen source and can range from 1-7 mM (Freshney 1987). This can result in ammonia concentrations up to and greater than 5 mM (Butler 1989; Hayter, Curling et al. 1991). In many large-scale fermentation applications, ammonium hydroxide or ammonium sulfate is often used as the sole nitrogen source and can exceed concentrations of 80 mM. There are a significant number of findings suggesting the presence of ammonium in culture media has an impact on protein and oligosaccharide processing. Table 1 provides a list of relevant research, identifying the significance of ammonium concentration with respect to production of recombinant glycoproteins. Listed is the host organism and recombinant protein produced, the level of ammonium concentration examined, and the affects on both microheterogeneify and macroheterogeneity. 13 Table 1: Ammonium literature H O S T (PRODUCT) R A N G E M I C R O H E T E R O G E N E I T Y M A C R O H E T E R O G E N E I T Y R E F . C H O (h-EPO) t N H 4 C I (0-40 mM) 1 sialylation •I site occupancy 31% I O-linked glycosylation (Yang and Butler 2000) C H O (h-EPO) f NH4CI (0-40 mM) Tetraantenary decreased by 60% (Yang and Butler 2000) Tn5Bl-4 (SEAP) t ammonia (0-62mM) Little or no effect Mi ld affects after 30mM Little or no effect Mi ld affects after 30mM (Donaldson, Wood et al. 1999) C H O (G-CSF) t ammonia (0-10, 50mM) I oc-2,6-linked sialic acids 4- disialo/monosialo (Andersen Dana 1995) C H O (mPL-1) T ammonia (0-9mM) (pH varied) 4- site occupancy (65%) ammonia affect is pH dependent (Borys 1994) C H O (mPL-1) t p H (6.1-8.7) (ammonia present) 4- site occupancy dependant on ammonia concentration (Borys 1994) C H O (mPL-1) t p H (6.1-8.7) 4 site occupancy (below pH 6.9) 4- site occupancy (above pH 8.2) (Borys, Linzer etal. 1993) Hybridoma (IgM) t ammonia (0-1 OmM) (50mM) •I sialylation (Thorens and Vassalli 1986) BHK-21 (IL-Mu6) t ammonia (0-15mM) i terminal galactosylation and sialylation (40%) (Gawlitzek, Ryll et al. 2000) B H K (IL-2 variant) t ammonia ( t UDPGNAc) f antennary structures (Grammatikos , Valley et al. 1998) Extracellular ammonia in the concentration range 2 mM to 50 mM has been shown to decrease the amount glycosylation and degree of sialylation (Goochee, Gramer et al. 1991; Yang and Butler 2000). For example, the production of h-EPO in CHO subjected to an increase in ammonium concentration from 4-40 mM resulted in a 60% reduction in tetra-antennary sialylated glycans, with a corresponding increase in tri and bi-antennary 14 glycans (Yang and Butler 2000). In the production of hG-CSF using CHO cells, the ratio of O-linked disialylated to monosialylated glycans decreased from 2.6 to 1 when extracellular ammonia increased from 0-1 OmM (Andersen Dana 1995). The effect of ammonia concentration also impacts the overall heterogeneity of glycoprotein production. Increasing ammonium concentration from 0-40 mM resulted in considerable changes in glycosylation with respect to macro and microheterogeneity. The degree of sialylation and the amount of O-linked glycosylation was inhibited and the total site occupancy decreased by approximately 31% (Yang and Butler 2000). Borys and colleagues determined that even low ammonium concentrations ranging from 0-9 mM had a significant impact on glycosylation, reducing site occupancy up to 65% (Borys 1994). This affect was determined to be both a function of extracellular ammonium concentration and culture pH. Even concentrations as low as 2 mM have been shown to result in decreased levels of glycosylation. It has been widely hypothesized that ammonia impacts glycosylation due to a weak base affect. Ammonium ion transport and the intracellular environment explain this hypothesis. Extracellular ammonia and ammonium ions are transported into the cell from the medium. Inside the cell, amines tend to accumulate in pH sensitive, acidic, intracellular compartments to concentrations in excess of their extracellular concentration (Goochee and Monica 1990). The trans-Golgi is one of the compartments affected, causing a localized increase in pH due to the uptake of protons by neutral amines (Goochee and Monica 1990). This accumulation and subsequent alteration in localized pH can result in the inhibition of enzyme activity and disruption of receptor ligand interactions necessary for vesicle trafficking (Dean, Jessup et al. 1984; von Zastrow, Castle et al. 1989). This shift in pH can alter the enzymatic activity of transferases involved in oligosaccharide and protein processing. P-1,4 galactosyltransferase was identified as having a pH optimum at 6.5. It has also been reported that a pH range of 7 to 7.2 for the Golgi is expected under an extracellular ammonium concentration of 10-15 mM (Gawlitzek, 2000). This evidence suggests that enzymes involved in protein processing and glycosylation may be affected by ammonium concentrations in a pH dependant manner. It has also been reported that culture conditions with greater than 15 10 mM N H 4 could cause a shift in pH that would reduce the enzymatic activity of sialyltransferase by 2-fold (Andersen Dana 1995). There are some exceptions to these observed trends. Increased ammonium levels up to 62 mM had little effect on the level of glycosylation in Tn5Bl-4 insect cell line (Donaldson, Wood et al. 1999). Culture pH conditions in the range of 6.9 to 8.2 have been shown to have no effect on glycosylation alone. However, pH indirectly affects glycosylation via ammonia transport into the cell (Borys, Linzer et al. 1993). As pH increases, the equilibrium of amines shift from ionized to the neutral form. The neutral form of the amine is transported freely across the cell membrane unlike the ionized form, which requires an ion transport system. Borys and colleagues showed that as pH was increased over the range (6.7 - 8.4) with an ammonium concentration of 9 mM site occupancy decreased by 80%. In another study a pH below 6.9 or greater than pH 8.2 resulted in a decrease the amount of glycosylation (Borys, Linzer et al. 1993). There is significant evidence to suggest that ammonium concentration is a key factor in optimizing production of recombinant glycoproteins. Still more investigation is needed to confirm suspected mechanisms, and to extend these theories to other cell lines and protein applications. 2.2.5.2 Temperature Growth State The culture temperature tends to have significant effect on both site occupancy and degree of microheterogeneity. Table 2 provides a list of relevant research identifying the significance of temperature with respect to production of recombinant glycoproteins. Listed is the host organism and recombinant protein produced, the level of temperatures examined and the affects on both microheterogeneity and macroheterogeneity. 16 Table 2 : Temperature literature HOST (PRODUCT) RANGE MICRO HETEROGENEITY MACRO HETEROGENEITY REF. CHO (t-PA) •l- Temp (37C - 33C) T site occupancy (Andersen, Bridges et al. 2000) Tn5Bl-4 (SEAP) 1 Temp (20C - 30C) t oc-l,3-linked mannose t site occupancy (Donaldson, Wood et al. 1999) CHO (t-PA) •I Temp t site occupancy (Bahr-Davidson 1995) CHO (prolactin) t cycloheximide t site occupancy (Shelikoff Marc . Sinskey 1994) Recent studies show that a temperature change from 37 to 33°C for t-PA in CHO cells resulted in a 16% increase in site occupancy (Andersen, Bridges et al. 2000). SEAP production in the Tn5Bl-4 line resulted in a decrease of longer chain oligosaccharides and an increase of a-1,3 linked mannose as temperature was decreased from 30 - 20°C (Donaldson, Wood et al. 1999). The mechanism involved in the effects of temperature on glycosylation is not well understood. Observations have identified that shifting to lower temperatures increases the fraction of cells in the G i / G 0 growth phase (Moore, Mercer et al. 1997). This suggests that temperature may be indirectly affecting glycosylation through a change in growth phase. Cell cycle inhibitors, quinidine and thymidine, have been used to increase the number of CHO cells in the G 0 / G i phase (Andersen, Bridges et al. 2000). A 20% increase in the number of CHO cells in the G 0 / G i phase corresponded roughly to a 10% increase in site occupancy of t-PA (Andersen, Bridges et al. 2000). The differences in glycosylation due to growth phase may be due to the influence of other oligosaccharide biosynthetic pathways (Andersen, Bridges et al. 2000). In yeasts it has been shown that the up-regulation of dolichol-linked oligosaccharide pathway genes occurs upon shifting into the G i from the G 0 phase (Kukuruzinska and Lennon 1994). The effect of temperature over the range of 20°C to 30°C was examined for the Production of human 17 secreted alkaline phosphatase (SEAP) in Tn5Bl-4 insect cells (Donaldson, Wood et al. 1999). The culture temperature did not result in any different oligosaccharides produced, but there was an observable change in the degree of site occupancy. Further investigation identified that there was a shift toward species which did not contain a(l,3)-linked mannose. It was suggested that higher temperatures may have had an influence on a-mannosidase II processing. Results showed an almost 2-fold increase in species containing a(l,3)-linked mannose structures. This has a significant impact on glycosylation because the structure without a(l,3)-linked mannose are dead end structures, unable to undergo further processing by glycosyltransferases (Donaldson, Wood et al. 1999). Other theories have suggested that lower temperatures lead to a reduced translation elongation rate, which may affect the level of glycosylation by allowing more time for addition of glycans resulting in increased glycan size and site occupancy (Shelikoff Marc . Sinskey 1994). The evidence indicates that temperature is an important factor to consider in optimization of recombinant glycoprotein processes. 2.2.5.3 Carbon Source Glycans are comprised of a variety of sugar units including glucose, fucose, mannose, galactose etc. Limitations associated with carbohydrate availability may affect glycosylation. It seems that that glucose starvation in many mammalian host systems may lead to the attachment of abbreviated oligosaccharide precursors and hence a decrease in glycan micro and macroheterogeneity (Goochee and Monica 1990). Table 3 provides a list of relevant research identifying the significance of temperature with respect to production of recombinant glycoproteins. Listed is the host organism and recombinant protein produced, the changes in glucose examined and the affects on both microheterogeneity and macroheterogeneity. 18 Table 3 : Glucose literature HOST RANGE MICRO HETEROGENEITY MACRO HETEROGENEITY REF CHO (y-interferon) •I glucose (starvation) 1 site occupancy (Hayter, Curling et al. 1991; Xie.Nyberget al. 1997;Nyberg, Balcarcel etal. 1999) BHK-21 (IL-1 mutant) X glucose (starvation) 4 site occupancy (Gawlitzek, Valley et al. 1995) CHO i glucose (starvation) I oligosaccharide chain length (Rearick, Chapman et al. 1981) Mouse BALB/c 3T3 ,NIL 8 and CHO •I glucose (starvation) 1 oligosaccharide chain length (Gershman and Robbins 1981) CHO (sinibus virus) i glucose X shorter chain oligosaccharides •I site occupancy (Davidson and Hunt 1985) Production of lipid-linked oligosaccharides in CHO cells in glucose free media resulted in a change from the normal synthesis of Glc3Man9GlcNAc2 to a smaller Man5GlcNAc2 species. These findings suggest a different glycosylation pathway exists under glucose starvation (Rearick, Chapman et al. 1981). CHO and NIL 8 cells under glucose starvation have produced species with predominant MansGlcNAc2 and Man2GlcNAc2 glycans (Gershman and Robbins 1981). BALB/c 3T3 cells also showed a decrease in higher mannose lipid-linked oligosaccharides but ManoGlcNAc2 was still the major species (Gershman and Robbins 1981). The mechanism for this decrease in oligosaccharide size is explained by the presence of specific mannose donors. There is evidence that GDP mannose is responsible for adding the first five mannose residues while the remaining residues are transferred by dolichol-P mannose (Rearick, Fujimoto et al. 1981). The decrease in concentration of the latter has been associated with glucose starvation and thus produces shorter oligosaccharides (Rearick, Chapman et al. 1981). Glucose starvation at levels of 0 to 80 mg/L can also cause a decrease in site occupancy of oligosaccharides at light chain asparaginyl sites that are normally glycosylated (Stark and Heath 1979). This effect was also seen in CHO cells under glucose starvation (Davidson and Hunt 1985). 19 Overall glucose limitations generally reduce glycosylation, which can lead to undesirable variable heterogeneity. 2.2.5.4 Culture Time One of the most complicated production problems with recombinant proteins and cell culture is the overall maintenance of culture conditions. Although there are many culture conditions that are monitored and can be controlled to some extent, there are many factors that cannot. The glycosylation characteristics of recombinant proteins have been monitored over time for many batch cultures and the macroheterogeneity has been shown to change significantly (Curling, Hayter et al. 1990). Table 4 provides a list of relevant research identifying the impact of culture time on glycosylation. Listed are the host organisms, the recombinant proteins produced, the culture periods examined and the affects on both microheterogeneity and macroheterogeneity. Table 4 : Culture time literature HOST RANGE MICRO HETEROGENEITY MACRO HETEROGENEITY REF CHO (t-PA) 4 growth rate (1-7 days) t site occupancy (Andersen, Bridges et al. 2000) CHO (t-PA) T G]/G0 phase (quinidine) (thymidine) t site occupancy (Andersen, Bridges et al. 2000) Tn5Bl-4 (SEAP) t (34- 120 hrs) t mannosidase resistant glycans (Donaldson, Wood et al. 1999) HepG2 (transferrin) t days (4,6,1 ldays) same protein secretion rate t site occupancy (Hahn and Goochee 1992) HepG2 (transferrin) Replating of high glycosylating cells 4 site occupancy (Hahn and Goochee 1992) CHO (Hu-IFN-y) t culture time (3 hr-195 hr) 4 site occupancy (Curling, Hayter et al. 1990) 20 In CHO cells producing t-PA , the culture was examined for seven days and an increase in site occupancy of 20% was observed (Andersen, Bridges et al. 2000). The same effect was also seen in the Hep G2 producing transferrin over a culture period of eleven days. A shift from 81% to 69% non-glycosylated protein was offset by an increase in bi and tri-antennary glycoforms (Hahn and Goochee 1992). However, this trend does not always hold true. In CHO cells the amount of non-glycosylated INF-y increased from 0 to 30% by the end of the culture period (Curling, Hayter et al. 1990). Some of the of the suggested causes for glycosylation changes over culture time include changes in cell growth phase (Andersen, Bridges et al. 2000) and culture conditions. It has also been shown that HepG2 cells producing transferrin show a 3.2 fold decrease in GlcNAc-T V activity over the culture period (Hahn and Goochee 1992). This decrease is correlated with a 3.5-fold increase in biantennary structures and suggests that regulation of glycosylases due to culture changes over time may be partly responsible for the observed results. 2.2.5.5 Other Factors The addition of growth factors, such as Interleukin-6, added to the medium of myeloma cells reduced the activity of N-acetylglucosaminetransferase III activity but increased the activity of GlcNAc-T IV and GlcNAc-T V (Nakao 1990). These changes can lead to altered glycan structures. Hormone additions have been shown to affect glycan structures in rat hepatocytes and cultured thyroid cells using dexamethasone which acts at the transcriptional level, and gonadotropin-releasing hormone (Ramey, Highsmith et al. 1987; Pos, van Dijk et al. 1988; Wang, O'Hanlon et al. 1989). These changes correspond with altered glycosyltransferase mRNA and their respective enzymes. Changes are specific to cell line and the hormones supplements used. The impact on glycosylation is thus difficult to predict. Culture conditions such as media type may also affect glycosylation. Increased levels of terminal sialic acid and galactose residues in monoclonal IgG produced by mouse hybridomas was observed in serum-free medium vs. serum (Patel, Parekh et al. 1992). However, the addition of galactosyl groups to antibodies produced in CHO cells was 21 increased in serum-containing versus serum-free media (Lifely, Hale et al. 1995). Dissolved oxygen levels from mild to severe hypoxia have shown minimal affect on glycosylation. It has been observed that CHO cells producing t-PA showed no change under mild hypoxia. Yet, production of FSH under mild hypoxia resulted in altered level of sialylation (Jenkins, Parekh et al. 1996). 2.3 Pichia pastoris as a Hos t Through genetic modification, the host organism P. pastoris can be transformed to produce recombinant proteins both for intracellular or extracellular expression. The transformation of P. pastoris results in an altered phenotype. The expression cassette containing the gene of interest is inserted into the genome producing one of three potential phenotypes, Mut+, Muts, and Muf. The selection of methanol utilization phenotype is complicated. The two most commonly used P. pastoris phenotypes are Mut+ and Muf3. The Mut+ phenotype contains the AOX1 gene that regulates synthesis of the AOX1 protein responsible for the first step of methanol oxidization (Cereghino and Cregg 2000). This enzyme allows P. pastoris the ability to grow utilizing methanol as a sole carbon source. The Muf3 phenotype has the recombinant gene integrated into the AOX1 locus and thus cannot synthesize AOX1. Nevertheless, P. pastoris does have another alcohol oxidase gene, AOX2, which still permits growth on methanol but at a much slower rate (Brierley, Bussineau et al. 1990). Although the induction time for Mut+ strains is less than half that of Muf3 strains there is often more difficultly in scale-up and large scale fermentation due to a significantly higher oxygen demand (Tschopp, Brust et al. 1987). Residual concentrations of methanol in the medium can be toxic to Mut+ strains because of increased production of formaldehyde (Hong, Meinander et al. 2002). Muf3 strains are much less sensitive to additional methanol. It has also been reported that methanol phenotype does not have a significant effect on productivity (Clare, Rayment et al. 1991). These results have also been shown for production of recombinant cystatin C (Files 2000). The Muf3 phenotype was thus selected for future optimization experiments. 22 2.4 Process Configuration and Optimization Determining the appropriate fermentation strategy is an important step in the optimization and development of recombinant protein production processes. The selection of operation mode including batch, fed batch, or continuous processes will greatly affect the productivity and stability of host as well as the ease of scale-up and optimization. The highest levels of productivity for recombinant protein production in P. pastoris have been with fed batch and continuous cultures. Lysozyme c2 production in continuous culture using a Mut+ strain resulted in a productivity of 13 mg-L"'-hr"1 (+/- 2 mg-^Lhr"1) while maintaining a cell concentration of 100 gDCW-L"1 (Digan, Lair et al. 1989). Although continuous cultures often result in higher productivities, the substrate utilization is significantly less and chance of contamination is greater due to the length of operation. Fed-batch cultures also obtained similar productivity results for lysozyme expression with cell concentrations of 120 gDCW-L"1 (Digan, Lair et al. 1989). Fed-batch protocols can be improved to obtain higher productivity by reducing the time between fermentation cycles. This has been achieved by harvesting only 90% of the culture and then refilling the reactor with fresh media to begin a batch phase growth again (Sreekrishna, Brankamp et al. 1997). This type of application is made possible due to the genetic stability of transformants in which the expression cassette is integrated into the genome. A fed-batch fermentation strategy was chosen for the optimization experiments due to ease of operation, high productivity, and short run times of approximately 6 days. The feed strategy must also be considered to optimize the process. When using a fed batch protocol one of the key parameters with respect to feed strategy is the concentration and control of methanol during induction. The following table highlights some of the recent results for different feed strategies using the P. pastoris expression system. 23 Table 5 : Feed strategies literature STRAIN FEED STRATEGY YIELD REFERENCE Mut+ As required (-0.5% methanol) ISmL-hr-'-day-1 500 mg-L"1 (galactose oxidase) (Whittaker and Whittaker 2000) Mut+ 3mL-L"' - 9 mL-L'1 1.2 g-L"1 (rAcAP-5) (Inan, Chiruvolu et al. 1999) Muts 0.2% - 0.8% (methanol) 250 mg-L"1 (bovine lysozyme) (Brierley, Bussineau et al. 1990) Muts Mixed Glycerol:MeOH, 4:1 (0.9 g-L"1 methanol) 180 mg-L"1 (bovine lysozyme) (Brierley, Bussineau et al. 1990) Muts Mixed Glycerol :MeOH, 2:1 (0.9 g-L"1 methanol) 290 mg-L"1 (bovine lysozyme) (Brierley, Bussineau et al. 1990) Muts Mixed feed limiting glycerol 2-5 g-L"' methanol 375 mg-L"1 (bovine lysozyme) (Brierley, Bussineau et al. 1990) Mut+ Limiting methanol (D.O. spike method) 450 mg-L"1 (bovine lysozyme) (Brierley, Bussineau et al. 1990) Muts Less than 1-2% methanol 2 mg-L"1 (SRAFP) (Loewen, Liu et al. 1997) Muts Mixed feed GlycerokMeOH, 5:2 Less than 0.5% methanol 30 mg-L"1 (SRAFP) (Loewen, Liu et al. 1997) Mut+ 0-1.6% methanol 37 mg-L"1 (SRAFP) (Loewen, Liu et al. 1997) Mut+ 0.034% methanol 774 mg-L"1 BoNT/A(Hc) (optimal) (Zhang, Bevins et al. 2000) Mut+ 0.12% methanol 36.9 mg-L-'-hr"1 BoNTVA(Hc) (optimal) (Zhang, Bevins et al. 2000) Mut+ 0.3% methanol 120 mg-L"1 (transferrin) (Guarna, Lesnicki et al. 1997) Muts 0.2% methanol 2.2 mg-L^-hr"1 (srAFP) (d'Anjou and Daugulis 2001) Mut+ 0.5% 130 mg-L"1 (laccase) (Hong, Meinander et al. 2002) 24 Table 5 Continued: Feed strategies literature STRAIN FEED STRATEGY YIELD REFERENCE Mut+ 1% 400 mg-L-1 (laccase) (Hong, Meinander et al. 2002) Mut+ Limiting methanol (D.O. spike method) 120 U-mL"1 (lipase) (Minning, Serrano et al. 2001) Mut+ Less than 0.25% methanol 325 U-mL"1 (lipase) (Minning, Serrano et al. 2001) The literature shows that the methanol concentrations used for P. pastoris fermentations range from less than 0.05% to greater than 2%. There are also cases where mixed feeding with glycerol and methanol during the induction phase was used. Supplying additional glycerol during the induction phase increases the cell concentration which directly affects productivity. However the advantage of increased cell mass is offset by the repressing affect of glycerol on protein production. Production of bovine lysozyme using different feed strategies shows that the concentration of glycerol used has an impact on productivity (Brierley, Bussineau et al. 1990)(see Table.5). Brierley's experiments also showed that although glycerol could be used to increase cell mass, the overall results showed that methanol alone resulted in the highest bovine lysozyme production. For each new system that is investigated these two variables must be optimized to obtain maximum protein production. 2.5 Process Optimization Utilizing Microarrays 2.5.1 DNA microarrays DNA microarrays are used to monitor changes in gene expression for a select set of genes. The number of genes examined can include the entire genome in organisms such as S. cerevisiae and E. coli. Microarrays for the entire human genome are currently in the developmental stages (Brown and Botstein 1999). Changes in gene expression have been used for applications including tumour classification (Pinkel 2000), cellular response to environmental changes (Gasch, Spellman et al. 2000), and determining metabolic and 25 genetic control of cells (DeRisi, Iyer et al. 1997). Microarrays have also been used to examine the changes in gene expression caused by over expression of recombinant proteins (Oh and Liao 2000). Microarrays are glass slides with strands of DNA covalently bonded to the glass surface in a matrix of spots. Each spot on the array represents an individual gene or expressed sequence tag (EST). Two commonly used arrays are cDNA microarrays which contain spots comprised of DNA strands which are 400 to 1000 base pairs in length and oligonucleotide microarrays which contain DNA base pairs up to 100 base pairs in length. Preparation of the microarrays for analysis is similar for both types of arrays. The procedure involves extracting mRNA from a control and experimental population of cells. These samples undergo reverse transcription to produce fragments of cDNA, which are labeled with fluorescent dyes. The control and experimental samples are mixed and hybridized onto the array. After the hybridization procedure the arrays can be read using a scanner and the resulting images are compared to analyze for differences in gene expression between the control and experimental population. A detailed outline of the procedure is presented in the Materials and Methods. A schematic representation of the array process is presented in Figure 3. 26 Array C o n s t r u c t i o n P r o b e P r e p a r a t i o n cDNA library, each cell contains a unique sequence O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O o o o o o o o o o o O O O Q . O O O O O O m icroscope slide not expressed significantly in either sample Figure 3 : Microarray construction and probe preparation Schematic representation of microarray construction and probe preparation (Glover 2002). 2.5.2 Process Optimization and Microarrays 2.5.3 Gasch, Spellman et al. Results Recently published literature has presented the genomic expression patterns for S. cerevisiae in response to a diverse set of environmental transitions including temperature shocks, hydrogen peroxide, menadione, diamide, dithiothreitol, hypo and hyper osmotic shock, amino acid starvation, nitrogen source depletion, and progression into stationary phase.(Gasch, Spellman et al. 2000). DNA microarrays were used to monitor transcript 27 levels over time for the entire S. cerevisiae genome. The strains used in the Gasch, Spellman et al. study are presented in Table 6. Table 6 : Table obtained from Genomic Expression Programs in the Response of Yeast Cells to Environmental Changes (Gasch, Spellman et al. 2000) NAME GENOTYPE SOURCE DBY7286 MATa ura3-52 GAL2 Fereaetal., 1999 DBY8768 ura3-52/ura3-52 GAL2/GAL2 Fereaet al., 1999 DBY9434 MATa ura3-52 yapl::KANMX4 GAL2 Gasch, Spellman et al., 2000 DBY9435 MATa msn2::KANMX4 msn4::URA3 GAL2 DeRisi etal., 1997 DBY9439 DBY7286 harboring pRS416 Tae Bum Shin DBY9440 DBY7286 harboring pTSl Tae Bum Shin DBY9441 DBY7286 harboring pTS2 Tae Bum Shin The microarray probe preparation used for these environmental transition experiments is very similar to the procedure presented in section 3.5.6 - 3.5.8 of the Materials and Methods. Detailed methods and raw data for this study are presented on the World Wide Web (http://www-genome.stanford.edu/yeast_stress). The Gasch, Spellman et al. data available on the web is reported in log base two and represents the ratio of the cy-3 to cy-5 spot intensities for each gene. For each gene there is also a replicate spot. An average of the replicates is used for analysis. Log base two ratios greater than 1 represent changes in gene expression greater than 2-fold from the control cells. Log base two ratios less than -1 represent changes in gene expression of less than half the original. The methods of data analysis are presented in the Materials and Methods Section. Changes in gene expression are typically considered significant if the expression changes are greater than two fold or less than half of the original intensity observed for the control cells. The results of the Gasch, Spellman et al. study identified that yeast cells respond to environmental changes by altering the expression of thousands of genes, creating a 28 genomic expression program that was customized to each environment (Gasch, Spellman et al. 2000). These genomic programs include features that are specific to each stress imposed and respond to supply the gene products that are required under the specific conditions (Gasch, Spellman et al. 2000). In addition to these specific stress responses, a subset of approximately 900 genes was found to be up-regulated or down-regulated in almost all of the stress conditions. This general response was referred to as the environmental stress response (ESR)(Gasch, Spellman et al. 2000). One of the Gasch, Spellman et al. experiments examined the effects of amino acid starvation in the culture media. Amino acid mixtures have been reported to be a preferred nitrogen source for S. cerevisiae and yeast cultures (Magasanik and Kaiser 2002). In typical S. cerevisiae cultures, amino acids are added to supplement the media to improve growth and protein production. However these components are usually added at the start of the culture and if not monitored may be consumed resulting in one or more amino acid limitations. The experiments performed by Gasch, Spellman et al. were conducted by supplying the cells with an amino acid supplement combined with a defined media (complete minimal media SCD) for growth. The cells were grown to early log phase and then exposed to the same defined media without the amino acid supplements. Samples were harvested after 0.5, 1, 2, 4, and 6 hours and compared with gene expression data from the pre-stressed genes. The Gasch, Spellman et al. study did not discuss the results of nitrogen source starvation or amino acid starvation on gene expression. However they did report that ESR genes were differentially expressed under both of these stress conditions. The results from Gasch, Spellman et al. were re-analyzed and our interpretation is presented in the Results section 4.5. 2.5.4 Natarajan, Meyer et al. Results Natarajan (Natarajan, Meyer et al. 2001) presented results with respect to amino acid gene expression profiling and amino acid starvation. In this study 3-aminotriazole (3AT) was added to S. cerevisiae culture, which imposes a histidine limitation on the cells. The compound 3AT is a competitive inhibitor of His3p, a transcriptional activator for histidine synthesis genes. The gene expression profiles were examined for cultures that 29 were exposed to 10 mM 3AT and 100 mM 3AT and histidine and leucine limitations. The leucine and histidine limitations were imposed by exchanging the original media supplemented with amino acids, with a media containing half the original concentration of those amino acids. The gene expression data for these results is available on the World Wide Web (http://www.rii.com/publications/200 l/mcb200 lMarton.htm). The results of this study showed that the transcriptional activator Gcn4p is general regulator for amino acid biosynthetic genes. Under the amino acid limitations imposed in this study Gcn4p was determined to target genes in every amino acid biosynthetic pathway except cysteine (Natarajan, Meyer et al. 2001). Gcn4p was also determined to target genes for amino acid precursors, vitamin biosynthetic enzymes, peroxisomal components, mitochondrial carrier proteins, and autophagy proteins. The Natarajan (Natarajan, Meyer et al. 2001) findings identified that Gcn4p evokes a broad transcriptional response to the stress conditions imposed. However the study did not identify any unique responses to the specific stress conditions. The results from Natarajan (Natarajan, Meyer et al. 2001) were re-analyzed and our interpretation is presented in the Results section 4.5. 2.5.5 Biosynthetic Pathways The approach to analyze the existing literature and the amino acid limitation experiments carried out in this thesis was to identify genes that are involved in amino acid synthesis and transport pathways. The amino acids selected for examination were leucine and glutamine. Leucine was selected for analysis because it can only be synthesized via one pathway making the expression analysis less complicated. The leucine biosynthesis pathway is presented in Figure 4. 30 Pyruvate » 2-Hydroxyethyl-ThPP Leucine Biosynthesis ILV2 ILVG I IILVS Isoleucine biosynthesis (same genes as Valine biosynthesis) (S)-2-Acetolactate | ILV5 j | (R)-3-Hytiroxy-3-methyl-2 oxobutanoate T Puruvate Metabolism | ILVS | ^ (R)-2,3-DihyrJroxy-3-methylbutanoate 2 Isopropytmaleate 3-lsopropylmalate ? " | L E U 1 | T 2-Oxo-4-methyl-3- <=> carboxypentanoate 4-Methyl-2-oxopentanoa1e -L-lsoleucine L-Leucine COCSO NAM2 i L-Leu-tRNA(Leu) I Protein Protein Figure 4 : Leucine biosynthesis schematic Leucine biosynthesis pathway based on data from Kegg website (www.kegg.org). The genes are outlined by boxes and intermediate compounds are also listed. From Figure 4 the production of leucine requires the genes LEU1, LEU2, and LEU4. These genes are regulated by extracellular leucine concentration at the transcriptional level (Reece 2000). The activation of LEU genes is dependant on a Zn(II)2Cys6 binuclear cluster containing a protein called Leu3p and is controlled by the concentration of a-isopropylmalate (a-IPM) and intermediate of the leucine biosynthesis pathway (Reece 2000). Glutamine was selected for analysis because it is a key amino acid involved in nitrogen regulation in yeasts (Magasanik and Kaiser 2002). Glutamine is readily converted to glutamate and thus genes involved in glutamate biosynthesis were examined. All of the pathways for the utilization of non-preferred nitrogen sources feed into a common set of 31 reactions for the production of glutamate and glutamine (Magasanik and Kaiser 2002). A schematic representation of the glutamate biosynthesis pathway is presented in Figure 5. Glutamate Biosynthesis Glucosamine-6P - O N^Acefyt D Glucosamine-6P O N-AcetylD-Glucosamine-6P Malate |GCNA| |GCH2| |GCN3] Oxaloacetale GLN3| |GCDI| IHAP2| Citrate Succinate Fumarate O o: 2-Oxogluterate [ GLT1 | |GDH3| |GDH2| |GDHI| I ' D R I H C |AAT1 I |AAT2| YLR089C L-&-glutamyl cysteine OXID O |MSE1| L-glutamyl-tRNA (Glu) succinate semialdenyde C~ L-Glutamyl-tRNA(Gln) PUT1'| Histidine metabolism Nitrogen metabolism - •» Cyanoamino acid metabolism D-Gln&D-Glu metabolism " Glutathione metabolism ^ Urea cycle * Butanoate metabolism ' C5-branched dibasic acid metabolism *- Phophyfine metabolism *Q L:1-Pyrroline 5-carboxylate Arginine & Proline 1 ^ HI y i l l l N c o c r t u ^O* - O metabolism 4-aminobutanoate Figure 5 : Glutamate biosynthesis schematic Glutamate biosynthesis pathway based on data from Kegg website (www.kegg.org). The genes are outlined by boxes and intermediate compounds are also listed. There is a much more complex network of pathways able to synthesize glutamate (Figure 5). The glutamate biosynthesis pathway is closely connected to the TCA cycle, the urea cycle, histidine, cyano amino acid, glutathione, butanoate and phophyrine metabolism. The key genes for glutamate synthesis include GLN1, GDH2 which code for glutamine synthetase, and glutamate dehydrogenase respectively and catalyze reactions to produce glutamine (Magasanik and Kaiser 2002). The genes GLT1, GDH1 and GDH3 are responsible for the synthesis of glutamate from ammonia and a-ketoglutarate. GLN3 activates the formation of enzymes also involved in glutamate biosynthesis. These genes represent many of the key regulators in glutamate biosynthesis. 32 Materials and Methods 2.6 Introduction This chapter is organized in the following manner. The P. pastoris and S. cerevisiae strains used are given. Next the culture media composition and preparation for all experiments are described. The experimental protocols for the screening, 2-litre bioreactor, and DNA microarray experiments are given in detail. Finally, the analytical techniques used for determining cystatin C assay, cell density, and glycosylation for the P. pastoris experiments, and RNA isolation, DNA microarray preparation and analysis for the S. cerevisiae experiments are described. 2.7 Cell Strains 2.7.1 Pichia pastoris The host microorganism was a wild-type P. pastoris strain known as X-33 (Mut+, His+) (Invitrogen 1998). The cells were previously transformed using electroporation and multicopy selection was performed using Zeocin (Files 2000). The pPICZa vector, which contained the gene for the cystatin C variant, was inserted into the P. pastoris genome at the AOX1 gene by homologous recombination (Files 2000). Insertion at the AOX1 gene resulted in a Muf3 phenotype. The cells selected were auxotrophic for histidine (His") which is likely due to the disruption of the histidinol dehydrogenase gene (H1S4) by gene insertion (Invitrogen 1998). This procedure is presented in detail in David Files Masters thesis (Files 2000). The cystatin C gene was mutated to introduce a consensus sequence for N-linked glycosylation (N-X-S/T) at two sites on the gene. The amino acid sequence of human cystatin C with the described mutations is presented below (Nakamura, Ogawa et al. 2000). S S P G K P P R L V G G P M D A S V E E E G V R R A L D F A V G E Y N K A S N D M Y H S R A L N S T Q V V R A R K Q I V A G V N Y F L D V E L G R T T C T K T Q P N L D N C P F H D Q P H L K R K S T A F C S F Q I Y A V P W Q G T M T L S K S T C Q D A 33 2.7.2 Saccharomyces cerevisiae The microorganism used for the microarray studies was S. cerevisiae YPH 239. This strain was created from the parent strains YPH 169 and YPH 4. YPH 169 is wild type for all loci and YPH 4 has the following modifications: lys2-801 (amber), ura3-52, ade2-lOl(ochre), his3-A200. The mating type of strain YPH 239 is mat a. The strains were supplied by the Hieter Lab 2.8 Culture Media Unless otherwise indicated, all the chemicals were from Sigma, St Louis. 2.8.1 Culture Plates and Inoculums For culture plates and inoculums the media formulations used included YPD (Yeast Peptone Dextrose) with and without Zeocin™. The formulation for the YPD media is presented in Table 7. Table 7 : Y P D media composition YPD Component Concentration (g*L_1) Yeast Extract 10 Bacto Peptone (Difco) • 20 Dextrose 20 Agar 20 Zeocin™ 0.1 - 0.5 The components in Table 7 were dissolved in distilled water and autoclaved at 121°C for 20 minutes. Agar was only added for making solid media plates. After autoclaving the mixture was cooled to 60°C before adding Zeocin™ (Invitrogen 1998). 34 Table 8: Composition of Bacto Peptone - Typical Analysis (Difco) BACTO PEPTONE (DIFCO) Component (%) Component (%) Protein 96.3 Glutamic Acid 5.73 Total Nitrogen 15.4 Histidine N / A Ash 3.8 Leucine 2.12 Moisture 2.7 Methionine 0.75 Amino Nitrogen 3.5 Phenylalanine 1.27 Calcium 1.8xlfJ-3 Serine 1.98 Potassium 0.25 Tryptophan N / A Chloride 0.90 Valine 1.69 Phosphate 0.40 Arginine 4.46 Magnesium 6.0x10"5 Cystine N / A Sodium 1.8 Glycine 15.03 Sulfate .32 Isoleucine 1.14 Alanine 5.16 Lysine 2.15 Aspartic Acid 3.34 Ornithine N / A Proline 6.88 Tyrosine 1.11 Threonine 1.17 The Components presented in table 8 represent a typical analysis of the Bacto peptone media that is used in the Y P D formulation. 2.8.2 Screening Experiments For all small scale-experiments the media formulations used include B M G H , and B M M H recommended by Invitrogen (Invitrogen 1998). The composition of the media is presented in Table 9 and 10. 35 Table 9 : B M G H and B M M H media composition B M G H / B M M H Component Concentration (g^L1) K2HP04 2.3 KH2P04 11.8 NaOH (pH adjustment) as required Y N B 13.4 Biotin 0.0004 Glycerol (for BMGH) (Fisher) 20 Methanol (for BMMH) (Fisher) 5 Histidine 0.004 36 Table 10 : Y N B media composition Y N B Component Concentration (for a total of 13.4g) Ammonium Sulfate 10 Biotin 0.002 Calcium Pantothenate 0.4 Folic Acid 0.002 Inositol 2 Niacin 0.4 P-Aminobenzoic Acid 0.2 Pyridoxine Hydrochloride 0.4 Riboflavin 0.2 Thiamine Hydrochloride 0.4 Boric Acid 0.5 Copper Sulfate 0.04 Potassium Iodide 0.1 Ferric Chloride 0.2 Manganese Sulfate 0.4 Sodium Molybdate 0.2 Zinc Sulfate 0.4 Potassium Phosphate Monobasic 1 Magnesium Sulfate 0.5 Sodium Chloride 0.1 Calcium Chloride 0.1 Three factors were varied: temperature, pH and ammonium concentration, as shown in Table 11. The pH was adjusted using NaOH. Peptone was supplemented to the media to offset the ammonium sulfate so as to maintain a constant nitrogen concentration in the medium. One gram of NH4HSO4 contains the same moles of nitrogen as 0.75 grams of peptone. However peptone is comprised of an enzymatic digest of protein and therefore supplies an additional carbon source to the media. The carbon source was not balanced for these experiments. For the experiments 3 variables were examined at 2 different levels. This resulted in a central composite design (CCD) with 8 factorial runs, 6 centre 37 points and 6 axial points with a a value of 1.68 to maintain rotatability. The different levels were denoted by -1,+1 and O with axial points denoted by the letters " A " and "a". Table 11 : Screening experiments central composite design Screening Design Run# Design Pattern Temp °C PH N H 4 H S 0 4 1 0 28 6 4.6 2 aOO 21.3 6 4.6 3 0 28 6 4.6 4 0 28 6 4.6 5 — 24 5.5 0.0 6 ++- 32 6.5 0.0 7 AOO 34.7 6 4.6 8 00A 28 6 12.3 9 -++ 24 6.5 9.1 10 -+ - 24 6.5 0.0 11 0 28 6 4.6 12 +- 32 5.5 0.0 13 +-+ 32 5.5 9.1 14 +++ 32 6.5 9.1 15 0 28 6 4.6 16 0A0 28 6.8 4.6 17 00a 28 6 0.0 18 0 28 6 4.6 19 OaO 28 5.2 4.6 20 -+ 24 5.5 9.1 Phosphate buffer (pH 5.2-6.8), histidine (400 mg-L'1) and biotin (200 mg-L-1) stock solutions were prepared in advance by mixing these components with distilled water and filter sterilizing with a 0.2 urn filter. Solutions were stored at 4°C. The YNB stock solution was prepared by dissolving the compounds of Table 9 in distilled water and filter sterilizing with a 0.2 p. filter. The BMGH/BMMH media was prepared by aseptically combining the components of Table 9. 38 2.8.3 2-litre Bioreactor Optimization Fermentation optimization was performed using two 2-litre glass fermentors with temperature, pH and D.O. control. An LH controller system and an Applikon biobundle (ADI 1010, ADI 1030) system were used. For all 2-litre bioreactor experiments, the media used was basal salts with PTM1 trace salts. Glycerol was used as the carbon source for the batch and fed batch growth phase and methanol was used for the fed batch induction protein expression phase. The compositions of the media are presented in Table 9 and 10. The three factors included NH 4OH, peptone, and amino acid mix, with the levels shown in Table 12. NH 4OH was used in the bioreactor experiments instead of NH4HSO4 to aid in pH control. Table 12 : 2-litre bioreactor experiments factorial design 2-Litre Bioreactor Optimization Run # Design Pattern N H 4 0 H Peptone (g^ 1 ) Amino Acids (g-L1) 1 +- 30 0 0 2 -+ + 0 20 20 3 -+- 0 20 0 4 +-+ 30 0 20 5 ++- 30 20 0 6 +++ 30 20 20 7 -+ 0 0 20 8 — 0 0 0 The amount of peptone and amino acids used was determined by a nitrogen balance based on the standard conditions using only ammonium hydroxide (A detailed example calculation is presented in the Appendix). The basal salts medium was prepared by dissolving the first 6 components of Table 13 in distilled water. 39 Table 13 : Basal salts media composition Basal Salts Media Component Concentration (g#L_1) H3P04(85%) 26.7 mL'L"1 CaS0 4 *H 2 0 0.93 K 2 S0 4 18.2 MgS0 4 *H 2 0 14.9 KOH 4.13 Glycerol (Fisher) 40 Methanol (Fisher) 2 PTM1 4.35 mL'L"1 Histidine 1 NH 4 OH 30 Bacto Peptone (Difco) 20 The media was autoclaved at 121°C for 20 minutes then allowed to cool under positive pressure using filtered air. Histidine was added prior to inoculation. Ammonium hydroxide (NH4OH) was used to adjust the pH of the media to obtain the desired pH conditions. PTM1 trace salts were added to the base media for the batch phase growth at 4.35 mL-L"1 of fermentation broth. For the fed batch phase, a 50% (v/v) glycerol feed with 12 mL-L"1 of PTM1 trace salts was used. For induction, a 100% methanol feed with 12 mL-L"1 or PTM1 trace salts was used. A PTM1 trace salts stock solution was prepared in advance by dissolving the compounds in Table 14 and filter sterilizing with a 0.2 um filter. 40 Table 14 : PTM1 trace salts media composition PTM1 Component Concentration C U S 0 4 * 5H 2 0 6 Nal 0.8 MnS04 * H 2 0 3 Na2Mo04 * 2H20 0.2 H3B03 0.2 CaS04 * 2H20 0.5 ZnC12 20 FeS04 * 7H20 65 Biotin 0.2 H2S04 5 mL'L" 1 2.8.4 Statistical A nalysis For both the screening central composite design and the 2-liter bioreactor full factorial design Jmp-In statistical software was used for analysis. For all variables an effect was considered significant if its F-value was greater than the critical F-value obtained from statistical tables. The critical F-value is based on the degrees of freedom of the treatment, the degrees of freedom of the error and the desired confidence level. For all statistical analysis a confidence level of a = 0.05 was selected as the minimum requirement for statistical significance. A least squares regression is performed using Jmp-in. The t-ratio and a two-tailed probability test for each of the model coefficients are calculated. Model coefficients representing interactions between variables are designated by "x" (example: VAR1 x VAR2). Comparison of p-values was used to analyze the significance of the different effects. The p-value represents the probability that the observed effect is due experimental error. A p-value of less that 0.05 represents a statistically result and is equivalent to an F statistic with a = 0.05 where F-value > F-critical. 41 2.8.5 Microarray Experiments For all microarray experiments a typical minimal media for S. cerevisiae was used. The minimal media was prepared by dissolving the compounds of Table 15 (except the drop out mix ingredients) in distilled water. The mixture was then autoclaved at 121°C for 20 minutes. Amino acid stock solutions were prepared in advance, filter sterilized (0.2 u. filter) and added to the media prior to inoculation. The composition of amino acids in the final media is presented in Table 16 (Ausubel, Brent et al. 1999). Table 15 : Minimal media composition Minimal Media (5. cerevisiae) Component Concentration (g^L1) Yeast Nitrogen Base 1.7 NH 4 S0 4 5 Dextrose 20 Drop Out Mix 1.3 42 Table 16 Amino acids drop out mix composition Nutrient Drop Out Mix Concentration in Formula (g) Media (mg^L1) adenine sulfate (hemisulfate) 2.5 40 L-arginine (HCI) 1.2 20 Laspartic acid 6 100 L-glutamic acid (monosodium salt) 6 100 L-histidine 1.2 20 L-leucine 3.6 60 L-lysine (mono HCI) 1.8 30 L-methionine 1.2 20 L-phenylalanine 3 50 L-serine 22.5 375 L-threonine 12 200 L-tryptophan 2.4 40 L-tyrosine 1.8 30 L-valine 9 150 Uracil 1.2 20 2.9 Experimental Protocols 2.9.1 Screening Experiments Screening experiments were performed in 250 mL baffled shake flasks. The inoculum was prepared by selecting one colony from YPD plates. The colony was grown in 50mL of BMGH media at 28°C in a shaking incubator (Innova 4000, New Brunswick Scientific) at 300 rpm. The culture was grown to an OD600 of approximately 4 (~7 g-L'1). This culture was used as the inoculum for the shake flasks experiments. The starter culture was used to inoculate 50 mL of BMGH, with modifications, for each of the experimental runs. The modifications to the BMGH media are discussed in section 3.3.2. The culture was grown for 20 hours in a shaking incubator (Innova 4000, New Brunswick Scientific) at 28°C and 300 rpm. After 20 hours of growth the cultures were 43 sampled to determine cell density. The cultures were then centrifuged (Eppendorf centrifuge 581 OR) at 3000 rpm. The media was exchanged and replaced with BMMH with modifications as outlined in Table 11. The cultures were then grown for 4 days at 28°C and 300 rpm in the shaking incubator. After the initial media exchange to start induction, samples were collected every 24 hours to determine cell density, cystatin C activity, and for gel analysis. Methanol was also added every 24 hours at a concentration of 0.5% to maintain methanol levels for protein induction. 2.9.2 2-litre Bioreactor Optimization The 2-litre bioreactor experiments were performed in a 2-litre baffled LH Inceltech bioreactor (series 210) and 2-litre baffled Applikon bioreactor (ADI 1010 biocontroller, ADI 1025 bioconsole). The inoculum was prepared by selecting one colony from the YPD plates. The colony was grown in 50 mL of YPD media at 28°C in a shaking incubator (Innova 4000, New Brunswick Scientific) at 300 rpm for 20 hours. This culture was used as the inoculum for the 2-litre bioreactor. One liter of basal salts medium was used for the 2-litre bioreactor experiments. Prior to inoculation, 0.1 g-L"1 of antifoam 289 (Sigma, St. Louis, MO) and 1 g-L"1 of histidine were added to the medium. The pH was adjusted to 5.0 by the addition of 30% (w/w) ammonium hydroxide. Runs where ammonium hydroxide was not used as a nitrogen source the pH was adjusted with NaOH. The pH was measured with a Mettler Toledo pH electrode (405-DPAS-SC) or an Applikon Applisense electrode (Z001023510). The temperature was maintained at 28°C and the impeller speed was set to lOOOrpm. Medical grade oxygen was fed to the bioreactor at approximately 0.3 L-min"1 to maintain the dissolved oxygen concentration above 30% of air saturation. The oxygen flow rate was controlled by the Applikon 1010 controller or LH controller. The oxygen concentration was measured with an Ingold electrode (322 756702/74091) or an Applikon Applisense electrode. The bioreactor was operated in batch mode until all the glycerol was consumed (approximately 18-20 hours). Glycerol depletion was identified by an oxygen spike. The glycerol fed-batch phase was approximately 4 hours with a glycerol feed of 50% (v/v) glycerol and 12 mL-L"1 PTM1 trace salts fed to the bioreactor at a rate of 15 mL-L'-h"1 44 (9.5 g glycerol-L/'-h-1). The glycerol fed batch was stopped after the cell density reached 200 gWCWcellsL'1. Prior to induction and methanol feeding the pH was adjusted to 6. A methanol feed of 100% methanol with 12 mL-L/1 PTM1 trace salts was then maintained for as long as 96 hours at a constant rate of 2.3 g-I/'-h-'. For methanol feed optimization experiments a methanol sensor and controller (Methanol Sensor Model 2.1, Raven Biotech) was used to maintain a constant methanol concentration in the media. The concentrations and run details are presented in Table 17. Mixed feeding was examined from 63-95 h with bioreactor F2. For mixed feeding, Glycerol was fed to the reactor at 3.4 g-L_1-h_1. Table 17 : Feed strategy experimental design Methanol Conditions (%) Bioreactor Time 0-26 h 26-70 h 63-95 h FI 0% >1% >1% F2 0% 0.05% 0.05% F3 0% 0.10% 0.20% 2.9.3 Microarray Experiments Microarray experiments were performed in 250 mL baffled shake flasks. The inoculum was prepared by selecting one colony from the YPD plates. The colony was grown in 50mL of minimal media (Table 15, 16) at 28°C in a shaking incubator (Innova 4000, New Brunswick Scientific) at 300 rpm. The culture was grown to an O D 6 0 0 of approximately 4. This culture was used as the inoculum for the experiments. The starter culture was used to inoculate 50 mL of minimal media (Table 15, 16), for each of the experimental runs. There was a replicate for each run. The culture was grown for approximately 9 hours in a shaking incubator (Innova 4000, New Brunswick Scientific) at 28°C and 300rpm. After 9 hours of growth the cultures were centrifuged (Eppendorf centrifuge 581 OR) at 3000 rpm for 5 minutes. The cells were resuspended in fresh minimal media (Table 15) without the amino acid mix to wash the cells. Finally 45 media was exchanged and replaced with fresh minimal media (Table 15, 16). For leucine and glutamate experiments, leucine and glutamic acid were excluded respectively. The cultures were then grown for another 6 hrs at 28°C and 300 rpm in the shaking incubator (Innova 4000, New Brunswick Scientific). Samples were collected prior to the initial media exchange, and at 0.5, 1, 3 and 6 hrs after the media exchange. The samples were centrifuged (Eppendorf centrifuge 5810R) at 3000 rpm for 5 minutes then immediately frozen at -80°C for future analysis. Additional samples collected prior to the media exchange were used as a control. 2.10 Analytical Methods 2.10.1 Cell Density For determination of cell density (DCW, g-L"1), 5 mL samples were dried in an oven at 100°C for 24 hrs before weighing on a Mettler Toledo AB104-S balance. Wet cell weight (WCW, g-L"1) was used to measure cell density in the bioreactor. For WCW measurements a 10 mL sample was centrifuged at 3,000 rpm (Eppendorf centrifuge 581 OR) for 3 minutes and then the supernatant was aspirated using a pipette. The combined weight of the cells and micro centrifuge tube were measured with a Mettler Toledo scale (AB104-S). The tare weight of the tube was subtracted to obtain the WCW (g-L'1). Wet cell weights were measured in duplicate. The dry cell weights (DCW, g-L"1) were calculated based on a calibration curve that is shown in Figure 6. 46 70 0 J , , , , , 1 0 50 100 150 200 250 300 Wet cell weight (g'L"1) Figure 6 : WCW vs. DCW calibration curve Calibration curve for dry cell weight (DCW) as a function of wet cell weight (WCW). The calibration curve was based on 9 10-mL samples that were dried for 24 hours at 100°C. The calibration curve was prepared using nine 10-mL samples that were dried for 24 hours at 100°C (Raw data in Appendix). 2.10.2 Cystatin C Assay Active cystatin C expressed by the P. pastoris was measured using a papain inhibition assay with Not-benzoyl-DL-arginine p-nitroanilide hydrochloride (BAPNA) as the substrate (Barret 1981). Nitroanaline liberated by the residual papain activity was measured by spectrophotometry at 410 nm (Pharmacia Biotech Ultrospec 1000). Papain and BAPNA were from Sigma (St. Louis, MO). The activity of cystatin C was determined by comparing the samples to blanks that were prepared by heating at 100°C for 10 minutes to denature the cystatin C (Nakamura, Ogawa et al. 1998). Samples were measured in duplicate. The reagents used for the cystatin C assay are presented in Table 18. 47 Table 18 : Cystatin C assay reagents Cystatin C Assay Reagents Component Concentration Substrate Stock B A P N A (in DMSO) 43.4 g ' L 1 Assay Buffer sodium phosphate buffer (pH 6.8) 1 M disodium E D T A 2mM Cysteine Base 0.61 g-L'1 Papain Solution Papain Stock (in distilled water) 20ml*L1 Papain Stock Solution Papain (activity 28 units/mg) 28 g-L-1 HgC12 0.5 mM E D T A 1 mM One unit of papain will hydrolyze 1.0 umole of Na-benzoyl-L-arginine ethyl ester per minute at pH 6.2 and 25°C. Cystatin C samples were analyzed using 1.5 mL Eppendorf tubes. Into each tube was added 0.5 mL of assay buffer, 0.1 mL of papain stock, and 0.1 mL of the cystatin C sample. The reaction was initiated by adding 50 uL of stock substrate solution. The tubes were incubated for 15 minutes at 37°C. The reaction was stopped by adding 0.32 mL of acetic acid. The nitroanaline that was liberated by enzymatic activity was measured by spectrophotometry at 410 nm (Pharmacia Biotech Ultrospec 1000) indicating the papain activity. The percentage of inhibition of papain was calculated by comparing the absorbance of the sample solution to the absorbance of the blank. Sample calculations for cystatin C concentration and productivity are presented below. . %inhibition 0.056mgpapain \mol 1,000,000 /^, g \kDA cystatm(pM) = -00 \00pLsample 2\kDapapain L lOOOwg 1.66*10"2lg 6.022;cl023 cystatin(pM) cystatin _ productivity(pM • gDCW • h ) = dry_ cell _ weight(gDCW) • culture _ time(h) 48 2.10.3 SDS gel electrophoresis Cystatin C expressed by the P. pastoris was analyzed for the presence of glycosylation using SDS gel electrophoresis. The chemicals required are presented in Table 19. Table 19 : SDS PAGE electrophoresis reagents SDS gel electrophoresis Reagents Component Concentration Separating Gel Acrylamide (45%) 150 g-L 1 Tris HCI (1.5M) (pH6.8) 0.375 M SDS (10%) 0.10% Ammonium persulfate (10%) 0.10% Temed 0.67 g«L1 Concentrating Gel Acrylamide (45%) 50 g-L"1 Tris HCI (1.5M) (pH6.8) 0.21M SDS (10%) 0.10% Ammonium persulfate (10%) 0.03% Temed 1.65 g-L 1 Running Buffer Tris HCI 6 g-L'1 Glycine 28 g-L"1 SDS 1 g-L 1 Sample Buffer Tris HCI (pH6.8) 20 mM SDS 20 g-L"1 b-Mercaptoethanol 20 g-L 1 Glycerol (Fisher) 400 g-L 1 Staining Comas sie Blue Dye 0.25 g«L' Methanol (Fisher) 50% (v/v) Acetic Acid 10% (v/v) Destaining Methanol (Fisher) 20% (v/v) Acedc Acid 10% (v/v) 49 The gel apparatus used was from Amersham Biosciences (Hoeffer SE 250 Mighty Small II). The samples were concentrated by adding Trichloroacetic Acid (TCA) to a concentration of 20% (v/v) to the supernatant samples. The samples were kept at 4°C for 10 minutes and then centrifuged at 14000 rpm for 5 minutes. The supernatant was removed by pipette and the pellet was resuspended in 300 pL of ice-cold acetone to wash the pellet. The samples were centrifuged (Eppendorf Brinkmann 5415) at 14000rpm for 5 minutes. The acetone wash step was repeated once. The final pellet was used to prepare the gel samples. Approximately 20-40 uL of sample buffer was added to the pellet. The pellet was resuspended and heated for 3 minutes at 100°C. A standard low molecular weight marker (BIORAD 161-0304) was also prepared using the same heating procedure. The gel lanes were loaded with 20 uL of sample and the system was electrophoresed at 50mA and max voltage. The running buffer used was described in Table 19. After running the gels, they were stained for 24 hours, then destained for 24 hours using the solutions outlined in Table 19. The gels were analyzed for different cystatin C species using imaging software. The single (SG) and double (DG) glycosylated bands have been confirmed as cystatin C by western blot analysis using rabbit antiserum that was raised against cystatin C (DAKO Corp., CA) (Files 2000). 2.10.4 ENDO FI Treatment ENDO FI treatment was used to cleave carbohydrates from the glycoprotein. 25 pL of denaturing solution (2% SDS and IM beta-mercaptoethanol) was added to 450 pL of substrate solution. The substrate solution was comprised of 1.1 mgmL - 1 ribonuclease B in IX reaction buffer (50 mM NaHP0 4, pH 5.5). This mixture was heated to 100°C for ten minutes then allowed to cool before adding 25 pL of Triton X-100 solution. Then 1- 5 uL of diluted enzyme (16 U-mg"1, 17 U-mL -1) was added to five 50 pL aliquots of denatured substrate. The reactions was incubated at 37°C for 5 minutes then the reaction was stopped by heating at 100°C for 5 minutes. 50 2.10.5 Gel Imaging A nalysis Scion imaging software was used for gel analysis (www.scioncorp.com). The gels were scanned using an HP flatbed scanner and the relative peak intensities were determined using the software. A typical scan resulted in an output that was displayed by 1-3 peaks. Peak analysis by evaluation of the area under the curve was used to compare relative peak intensities. Due to the close proximity of the single and unglycosylated peaks it was necessary to visually estimate the division of these peaks to perform the peak analysis. The imaging analysis was only used to detect relative differences in the quantity protein species. 2.10.6 RNA isolation Prior to RNA isolation the cells were stored in a 1.5 mL eppendorf tube in a -80°C freezer. The cells were then removed from the -80°C freezer and resuspend in 0.7 mL of acid phenol-chloroform-isoamylalcohol (25:24:1, pH 4.7; Sigma P-1944) and 0.7 mL of TES (10 mM Tris pH 7.5, 10 mM EDTA, 0.5% SDS; autoclaved). The acid phenol-chloroform-isoamylalcohol was prewarmed for 10 minutes at 65°C. The cell sample size required for this analysis was 0.02 g of dry cells. The cells were then mixed by vortexing the sample to resuspend the pellet. The sample was then incubated in a 65°C water bath for 1 hour while continuing to vortex the sample for 20 seconds every 20 minutes. After 1 hour of incubation the sample was vortexed one final time and then centrifuged at 14,000 rpm for 10 minutes at 4°C. The aqueous layer was removed by pipette and stored in a new 1.5 mL eppendorf tube. Then 750 uL of acid phenol-chloroform-isoamylalcohol (25:24:1, pH 4.7; Sigma P-1944) was added to the tube. The sample was vortexed for 20 seconds and then centrifuged at 14,000 rpm for 10 minutes at 4°C. The aqueous layer was then removed by pipette and stored in a new 1.5 mL eppendorf tube. Then 750 pL of chloroform : isoamyl alcohol (24:l)(Sigma C-0549) was added to the tube. The sample was vortexed for 20 seconds and then centrifuged at 14,000 rpm for 10 minutes at room temperature. 51 The aqueous phase was then transferred to tube with 50 uL of 3M Sodium Acetate (NaOAc) pH 5.2 and 1 mL of 100 % ethanol (pre-cooled to -20°C) was then added. The sample was then stored at -20°C for at least 30 minutes. After storage, the sample was centrifuged for 5 minutes at 14,000 rpm at room temperature. The remaining ethanol was aspirated and the sample was left to dry for 1 minute. The pellet was then washed with 500 uL of 80% ethanol (pre-cooled to -20°C). The contents of the tube were then centrifuged for 1 minute at 14,000 rpm, at room temperature. The sample was then aspirated and allowed to dry for 1 minute. The pellet was dissolved by adding DEPC treated water to the samples (approximately 50 uL). The RNA purity and quantity were measured using OD260/OD280 spectrophotometry. The final sample was then frozen at -80°C. 2.10.7 Microarray probe preparation and hybridization The arrays were purchased from the Ontario Cancer Institute Microarrays center. Each array contains 6240 complete Yeast ORF's (open reading frame) plus control spots totaling 6400 spots. Probe preparation and hybridization procedures are available on the Ontario Cancer Institute Microarray website (www.microarrays.ca/support/proto.html). The procedure was modified so that the concentration of SSC used was 0.1 x SSC instead for 1 x SSC. The protocol is also presented in the Appendix. 2.10.8 Microarray data acquisition and analysis The arrays were scanned using a Virtek ChipReader™ (www.virtek.ca). The arrays were scanned using 80% laser power and 80% detector gain. The acquisition and analysis of array images was performed using the Array Pro 4.0 software package (www.mediacy.com). Background noise was subtracted using the imaging software. The data obtained from the software was the mean cy3 and cy5 spot intensities. These values were normalized using R software version 1.5.1 (http://cran.r-project.org). The normalization technique used was a lowess (locally weighted polynomial regression) fit. With this technique the data is modeled by a polynomial weighted least squares 52 regression, giving more importance to the local data points. The following 4 step procedure is used to obtain the lowess fit (Cleveland 1979). For each " i " let h; be the distance from x; to the rth nearest neighbor of X j . Therefore |x; - X j | , for j=l,...n. For k=l,...n, let *i(*,) = W ( * * - * / ) ) (l) A value of "f' is selected between 0 and 1 and represents the fraction of Wk(xj) = 1. For the microarray analysis an f value of 0.2 was selected. A 1. For each " i " compute the estimates of /?J(XJ), j=0,.. .d, of the parameters in a polynomial regression of degree d on yk on X k , fit with weighted least squares with Wk(xk) A for (xk,yk). Thus /?J(XJ) are the values that minimize S w t . ( * i ) 0 ' * - A - M - - - A * * ) : k=l (2) and y. is defined by j=0 k=\ (3) where rk (XJ) does not depend on yi, j=l,.. .n. 2. Let B be the bisquare function that is represented by B(x) = (1-x2)2, for |x| < 1 and B(x) = 0 for |x| > 1. (4) A Let e; = yi - yi be the residuals from the current fitted values and let "s" be the median of the |e|. The robust weights are defined by 5k = B(ek/6s) (5) 53 3. Compute the new yi, for each " i " by fitting a dth degree polynomial using a weighted least squares with weights 5kWk(x;) at (xk,yk). 4. Repeat steps 2 and 3 a total of "t" times y ; = log(7cy3*cy5) Xi = log(cy3/cy5) Once the data was normalized, Microsoft Excel was used for data analysis. The log ratio of cy3 to cy5 spot intensities for each time point, were compared with the reference sample. The reference sample used for these runs was a sample collected prior to media perturbations and was designated the Ohr time point. The image fdes and excel data for the experiments performed are available on compact disc accompanying this thesis. 2.11 Statistical Design and Analysis 2.11.1 Screening Experimental Design A central composite design was selected for small-scale experimentation. The central composite design, or CCD, consists of a 2 k fractional factorial design of resolution V with 2k axial runs and a chosen number of centre runs. Unlike a 2 k design, the CCD can be used for fitting second order models. The resolution V means that no main effect or two-factor interaction effects are aliased. The selection of the axial run parameter and the number of centre points are important to obtain a model with good predictability throughout the range of experimentation. Axial runs for a CCD are specified by the distance, or a value, of the axial runs from the design centre point. An a value satisfying the equation a = (n f)1 /4 results in a rotatable design. Box and Hunter suggest that a rotatable design is necessary for a second order response surface design (Box and Hunter 1957). This design leaves the variance of the predicted response unchanged when the design is rotated about the center. The number of centre runs recommended for this type of design is between 3 and 5 (Montgomery 1997). 54 2.11.2 2-Litre Bioreactor Experimental Design A 2k factorial design was selected for large-scale experimentation. For this 2k fractional factorial design 3 variables were examined resulting in 23 design. A full model was chosen requiring 8 experiments and 1 replicate run. The full model allows for the identification of all main effects and interactions. This design was chosen to minimize the number of 2-litre experiments required, compared to a central composite which would have required 20 experiment. The factorial design however, assumes that the response is linear over the range examined, and does not allow for the prediction of second order relationships. 55 3 Result and Discussion 3.1 Introduction This chapter discusses in detail the experimental results. The stability of recombinant cystatin C was examined to identify the storage limitations. The methanol feed strategy for P. pastoris fed-batch fermentation was studied to determine the best strategy for future design experiments. Process optimization of recombinant cystatin C production focusing on productivity and glycosylation is also presented. Finally, microarrays, as a method to accelerate process development and optimization were examined. 3.2 Recombinant Cystatin C Stability 3.2.1 Introduction Human cystatin C is stable at temperature of 80°C for 10 minutes at pH 6.5 and is also stable at a pH as low as 2 at 25°C, for 10 minutes (Barrett 1984). Despite this stability data human cystatin C is susceptible to oxidation (Berti, Ekiel et al. 1997), dimerization (Abrahamson and Grubb 1994; Ekiel and Abrahamson 1996) and proteolysis (Lenarcic, Krasovec et al. 1991). These factors reduce human cystatin C stability for both short and long term storage. Fermentation runs for cystatin C production typically last 6 days, thus storage of supernatant samples is required to minimize analytical time and also reduce variability in assay sample preparation. Temperature stability studies performed on recombinant cystatin C produced in Pichia pastoris show storage temperature has a significant affect on cystatin C activity (Files 2000). The results are presented below. Samples for this stability study were collected directly from the culture medium at various time points and stored at 4°C after cell removal by centrifugation. Each curve represents a specific sample from the bioreactor monitored over a period of time up to 130 hours. The most significant decrease in activity was observed over the first 24 hours of storage at 4°C. 56 50 Storage Time (hours) Figure 7 : Cystatin C stability as a function of storage time (Files et al. 2000) Measured activity for cystatin C as a function of storage time. Samples were taken from the reactor, centrifuged, and the supernatant was stored in Eppendorf centrifuge tubes at 4°C. Each set of points represents a sample taken from the bioreactor at a particular time. Cystatin C activities were measured in duplicate. The error bars are the standard deviations of the measurements (Files 2000). It is likely that factors such as proteases, aggregation and medium composition still play a role at this temperature. Storage of medium supernatant samples at 4°C is not an adequate method to store cystatin C even for periods as short as 24 hours. It has been recommended that cystatin C can be stored in sodium phosphate buffer (50 mM, pH 6.7) with sodium chloride (0.1 M) and will remain stable for up to 3 months at 4°C (Ekiel and Abrahamson 1996). Although this sample storage technique may be applicable, it is desirable not to add any additional components to the sample to avoid interference with assay techniques. To address this existing stability problem, storage of cystatin C samples at -20°C was investigated to determine the effect on cystatin C activity. 57 3.2.2 Results and Discussion Recombinant cystatin C produced in the 2-Litre fermentor was used to examine the effects of storage at -20°C over a period of 3 weeks. The objective was to determine whether freezing and thawing the samples would affect the measured Cystatin C activity. Cystatin C samples were taken at 48 hours and 72 hours from 3 different 2-L fermentor runs. Cystatin C activity was determined using the papain inhibition assay, as described in the material and methods section, and samples were analyzed at 0 hours, 1 week, and 3 weeks of storage at -20°C. The cystatin C yield results for this experiment are presented in Figure 8. -a 15 03 U 5 3 4 5 6 Sample # j n o h r s Q 1 w e e k B 3 w e e k s Figure 8 : Cystatin C stability as a function of storage time Yield of active cystatin C for stability analysis comparing storage times at -20°C. The error bars represent the standard error for each measurement. Duplicate samples were run for each time point. Standard error was determined from the replicates and is represented by error bars. The highest standard error for the cystatin C assay was +/- 2.8 u.M. Comparison of samples over a 3-week period resulted in a 58 maximum standard error of +/- 1.9 U.M. The variability observed over the 3-week period is within the range of error of the assay. These stability results show that storage of recombinant cystatin C at -20°C for 3 weeks and the freezing and thawing of samples had little effect on Cystatin C activity. Bioreactor supernatant samples were stored at -20°C for a maximum of three weeks for all future experiments. 3.3 Feed Strategy 3.3.1 Introduction Several fermentation feed strategies are used for Muf3 strains. Invitrogen recommends a standard protocol for protein expression with a Muts strain (Invitrogen 1999). This protocol includes batch and fed-batch phases with glycerol as the carbon source. These modes of operation are used to obtain high cell density cultures. In the third phase of operation the carbon source is shifted to methanol to induce protein expression. Methanol feeding strategy is one of the most important factors for maximizing recombinant protein production (Zhang, Bevins et al. 2000). For Muf strains a methanol concentration between 0.2 - 0.8% is recommended (Higgins and Cregg 1998). However, concentrations as low as 0.05% and greater than 1% have also been used. The Invitrogen protocol also provides typical methanol feed rates for expression in the induction phase; however, the specific strain and protein being expressed may affect the optimal conditions. In addition to methanol, other carbon sources can be supplied during induction to increase cell concentration. A pure methanol feed or a mixed feed of methanol and glycerol are the two most frequently used feed strategies. A methanol sensor was obtained from the UBC Biotechnology Laboratory pilot plant to perform methanol feed optimization experiments. These experiments were used to determine the best methanol feed rates for future 2-litre bioreactor runs in which no methanol sensor was available. Three preliminary experiments were performed to examine the effects of methanol concentration on recombinant cystatin C production. 59 3.3.2 Results These experiments were designed to evaluate different methanol concentrations for the induction phase. Two different feeding scenarios were also examined. A pure methanol feed and a glycerol mixed feed. The runs were performed using induction methanol concentrations outlined in Table 17 (Materials and Methods). The mixed feeding strategy was only used for F2 from 63-95 hours. The cultures were grown in batch mode for 26 hours with basal salts medium and PTM1 trace nutrients at pH 5 and a temperature of 28°C. After 26 hours of growth on glycerol all three cultures reached approximately 200 g/L (WCW), 50 g/L (DCW). The cell yield on glycerol was 0.46 g DCW/g glycerol. Induction was started at 26 hours by increasing the pH to 6 and initiating methanol feed. The methanol concentration in the reactor was monitored and maintained at the desired set point using the methanol sensor and feed controller. Samples were taken at 26, 50, 63, 74 and 93 hours. Samples from the reactors were analyzed using the cystatin C assay procedure. The cell growth data and cystatin C yield results are presented in Figures 9 and 10. 60 0 20 40 60 Time (h) 80 100 Figure 9 : Cell density results for feed strategy experiments Cell density (gWCW-L1) as a function of culture time (hours). F l , F2, and F3 represent the bioreactor runs outlined in Table 17. 25 Figure 10 : Cystatin C yield results for feed strategy experiments Yield of active cystatin C as a function of culture time. F l , F2, and F3 represent the bioreactor runs outlined in Table 17. The error bars indicate the standard error for each measurement. 61 Growth on methanol over the induction period resulted in a negligible change in cell density except for mixed feeding conditions (Figure 9). Mixed feeding resulted in a final cell density of 284 g-L-1 compared to approximately 200 g-L"1 for pure methanol feed. The recommended feed strategy from Invitrogen (1 mL^h^L" 1 followed by 10 % increments every 30 minutes to a maximum of 3 mL*h_1) was used for FI and resulted in an accumulation of methanol in the fermentor exceeding the limits of the sensor (>1%). The cell density in FI dropped from approximately 185 g-L"1 to 165 g-L"1 over the induction phase. After 63 hours of fermentation, results indicated that the cystatin C concentration was comparable in all cultures within the range of 7.82 pM - 9.27 pM (Figure 10). This indicates that the concentration of methanol for the values examined (0.05%, 0.1% and 1% methanol) seemed to have little effect on the productivity of cystatin C in P. pastoris. At 70 hours the culture conditions were changed to examine concentration effects at 0.2% methanol in F3 and 0.05% methanol combined with glycerol mixed feeding in F2. Increasing the methanol concentration to 0.2% resulted in a greater than two-fold increase in cystatin C yield equivalent to 20 pM cystatin C in the supernatant. The specific cystatin C yield was determined to be 100 nmol-gWCW"1. For mixed feeding conditions a glycerol feed was initiated at 11.5 g-L-'-hr"1 for approximately 12 hours. The same result was also obtained for glycerol mixed feeding resulting in a final cystatin C concentration of 19.5 pM. The specific cystatin C yield was determined to be 68 nmol-gWCW"1. Based on the results obtained in the pure methanol feeding experiments, a methanol concentration of 0.2% was selected as the most favorable conditions for large-scale production of recombinant cystatin C. To determine the best methanol feed rates for future experiments, a plot of the methanol consumption vs. time was used. The results are presented in Figure 11. Using the slope information from the curve, we can estimate the feed rate required to maintain a methanol concentration of 0.2% for a given cell density (-200 gWCW-L"1). 62 1274 1258 -\ , , — — — - r - , , 1 0 2 4 6 8 10 12 Time (hrs) Figure 11 : Methanol consumption results for feed strategy experiments Methanol consumption as a function of time for experimental run F3. Details for F3 run are outlined in Table 17. The slope of the line in this figure was determined to be -1.49 indicating that the methanol feed rate necessary to obtain a methanol concentration of 0.2% was 1.49 g-L'-hr"1. This methanol feed rate was used for all future 2-litre bioreactor experiments. 3.3.3 Discussion The mixed feeding strategy has been used with success in other Pichia fermentations including the production offish antifreeze protein, which resulted in a 15-fold increase in yield (Loewen 1997). However, despite the ability of glycerol to increase culture productivity in some cases, evidence shows that glycerol above concentrations of 2% repress the AOX1 gene promoter (Gellissen, Hollenberg et al. 1995) and decrease cell specific productivity. Mixed feeding protocols can also result in the production of inhibitory levels of ethanol (Brierley, Bussineau et al. 1990). Although mixed feeding 63 results showed similar cystatin C yields and higher cell densities compared to a pure methanol feed strategy at 0.2%, a more favourable alternative would be to grow the Muf3 culture to a higher DCW prior to induction then use methanol as the sole carbon source. One approach to improve mixed feeding strategies involves investigating alternative carbon sources with a less inhibitory affect on the AOX1 expression. Sorbitol is a potential alternative to glycerol. It has been shown to increase cell mass during induction as well as the specific rate of recombinant protein production resulting in higher volumetric cell yields (Thorpe, d'Anjou et al. 1999). Sorbitol can also accumulate to concentrations as high as 5 g/L without affecting protein yield. There is potential to consider mixed feeding strategies and alternative carbon sources in future studies, however, for this work only methanol optimization was examined. The optimum methanol concentration for pure methanol feed conditions was determined to be 0.2% for the conditions examined. Lower cystatin C yields were obtained at a concentration of 1% methanol. This could be caused by production of inhibitory byproducts including formaldehyde or hydrogen peroxide inside the cell (Couderc and Barratti 1980; Murray, Duff et al. 1989). It has also been shown previously that the transcription levels, initiated by the AOX1 promoter, can be 3 to 5 times higher under growth-limiting methanol rates compared to conditions bf excess methanol (Minning, Serrano et al. 2001). At lower methanol concentrations of 0.05 to 0.1%, insufficient carbon source likely resulted in lower cystatin C yields. It is necessary to strike a balance with methanol feed concentration to maximize growth and expression. Only three experiments were performed to optimize the methanol feed strategy. Although the optimum conditions obtained in this experiment were 0.2% methanol, others have reported conditions of 0.3% resulting in five-fold increases in productivity compared to standard fermentation protocols. Methanol concentrations between 0.2% and 1% should also be examined to further optimize the process. For later experiments performed in this 1 1 thesis a methanol feed rate of 1.49 g»L" »h" was used based on the methanol 1 1 consumption rate determined from Figure 11. The feed rate of 1.49 g»L" »h" is only a guideline and does not guarantee that the methanol concentration will be maintained at 0.2% in the culture. For future experiments it is necessary to consider that the methanol 64 concentration may limiting or in excess which can have a negative affects on growth, productivity and glycosylation. 3.4 Effect of Process Conditions on Cystatin C Productivity and Glycosylation 3.4.1 Screening Experiments 3.4.1.1 Background Small-scale screening experiments are typically used to facilitate the optimization and development process. Combined with the power of statistical experimental design, researchers can minimize the experimentation required while maximizing the amount of information gathered. To screen a variety of potentially significant factors, shake flask experiments were performed for the expression of cystatin C in P. pastoris. Several factors can be investigated simultaneously by operating multiple shake flasks. The weaknesses of shake flask fermentations include the inability to control dissolved oxygen, pH and methanol concentration. Previous work with cystatin C examined the effects of pH, feed strategy, and phenotype on cystatin C productivity. However, the extent of glycosylation was not examined. The literature review revealed that temperature, pH, and nitrogen source could be factors that affect glycosylation significantly. The effect of these variables on productivity was also studied. 3.4.1.2 Design Details For the screening experiments, 250 mL shake flasks were used with media conditions, inoculum preparation, and sampling as outlined in Materials and Methods. The standard fermentation conditions for protein expression in P. pastoris suggest a temperature of 30°C (Invitrogen 1998). However 28°C was chosen as the centre point temperature from literature evidence suggesting the superior range for productivity and growth lies within the range of 26 to 30°C (Inan, Chiruvolu et al. 1999). The range of pH was selected based on results from David Files and Inan, suggesting that maximal productivity was obtained within the pH range of 5 to 7. The amount of ammonium sulfate was also 65 varied. However, nitrogen is essential for cell growth and protein expression therefore peptone was used to supplement the culture medium when ammonium sulfate was removed. The amount of peptone supplemented was based on a nitrogen balance. A completely randomized design was used for the experiments. The flasks were placed randomly in the environmental shaker and replicates were taken for each sample time. 3.4.1.3 Growth Results The growth results for each run were based on a one-time response for each of the treatments. Optical density measurements were performed for standard fermentation conditions resulting in maximum cell densities after 24 hours of growth. Based on standard run results a batch culture time of 20 hours was selected for the design runs to ensure the culture did not reach the stationary phase. The final cell densities after 20 hours of growth under the various run conditions are presented in Figure 12. 66 16 14 12 10 8 6 4 2 H i ft 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 RUN# Figure 12 : Cell density results for screening design Final cell density (DCW) as a function of run # (details outlined in Table 11). DCW was determined from experimentally determined calibration curve. Samples were collected after 20 hours of batch growth on modified BMGH. The error bars represent the standard error for each measurement. Final cell densities ranged from approximately 5.5 to 12.5 gDCWL" 1. The DCW values were determined by spectrophotometry and use of the calibration curve presented in the Materials and Methods. The maximum standard error calculated for the DCW measurements was determined to be +/- 1.67 gDCW-L - 1. The yield of cells on glycerol for the batch phase of growth ranged from approximately 0.275 gDCWg"1 glycerol to 0.625 gDCWg"1 glycerol. 67 3.4.1.4 Statistical Analysis The growth results for the small-scale experiments are summarized in the following three-dimensional box plot Figure 13. The eight vertices represent the eight different treatment combinations. The response values that are shown represent the average DCW obtained for the specific treatments. -0.0734 6.37395 3.7125i 18.0935 OJ 0.1793 2.41905 9 . 1 7.3121 17.48551 N H . H S O ^ g . L " 1 ) 21.3 Temp °C 3 4 7 Figure 13 : Box plot for screening design growth results A box plot to indicate the Growth response (g-L"1) for all combinations of three factors. Values are derived from the growth model outlined in equation (6). In brackets are the ranges for the variables examined. The value outlined with a box indicates the growth values (g-L"1). Jmp-In software was used for statistical analysis of the data. An analysis of variance was performed on the data and is presented in Table 20. 68 Table 20 : ANOVA for screening design growth results Source Nparm DF Sum of Squares F Ratio P-value Temp l l 11.27 5.63 0.04 pH l l • 0.08 0.04 0.85 N H 4 1 1 34.65 17.30 0.00 Temp x pH i i 1.23 0.62 0.45 Temp x N H 4 1 1 0.42 0.21 0.66 pH x N H 4 1 1 1.22 0.61 0.45 Temp x Temp l l 11.63 5.81 0.04 p H x p H l i 10.12 5.05 0.05 N H 4 x N H 4 1 1 0.46 0.23 0.64 Based on analysis using Jmp-In statistical software the temperature, NH4HSO4, (temperature)2, and (pH)2 terms were identified as significant effects. The effects are bolded in Table 21. Table 21 : Model analysis for screening design growth results Term Estimate Std Error t Ratio P-Value Intercept 9.45 0.56 16.90 <.0001 Temp 0.91 0.38 2.37 0.04 P H -0.08 0.39 -0.20 0.85 N H 4 -1.80 0.43 -4.16 0.00 Temp x pH 0.39 0.50 0.78 0.45 Temp x N H 4 0.23 0.50 0.46 0.66 pH x N H 4 0.39 0.50 0.78 0.45 Temp x Temp -0.90 0.37 -2.41 0.04 p H x p H -0.90 0.40 -2.25 0.05 N H 4 x N H 4 0.23 0.47 0.48 0.64 The resulting equation obtained from the response surface analysis is comprised of the 5 significant terms identified in Table 21. This equation estimates the final cell mass (gDCWL"1) after 20 hours of growth. The response equation is presented below. 69 Cell Density = 9.45 + 0.91 (T) -1.80 (NH4) - 0.90 (T)2 - 0.90 (pH)2 (6) Cell Density Units = (gDCWL 1) The model represents a second order relationship where the value of (T), (NH4), and (pH) are coded and range from -1 to 1 representing the range examined. Response surface plots based on the model are presented in Figures 14 and 15. These plots show the impact of the studied variables on growth. Figure 14 : Response surface for screening design growth results Response surfaces for growth as a function of NH 4HS0 4 concentration (g-L1), and temperature. The response surface is derived from the statistical model and is represented by Equation 6. The pH was held constant at 6 for this response surface. 70 NH 4HSO, (gL) Figure 15 : Response surface for screening design growth results Response surfaces for growth as a function of N H 4 H S O 4 concentration (g-L"1), and pH. The response surface is derived from the statistical model and is represented by Equation 6. The temperature was held constant at 28°C for this response surface. The model adequacy was confirmed by examining a plot of the residuals vs. the predicted values (Figure 16). 71 S 1.5 •§ • l-H CD 0.5 0 y -o.5 ^ .1 £ -1.5 o O -2 -2.5 • • • 8 • 10 12 • • • • -IN Growth ( g D C W - L ) Predicted Figure 16 : Plot of residuals for screening design growth model A plot of the residuals as a function of the predicted values derived from the model presented in Equation 6. The random scatter in the Figure 16 indicates that the variance is relatively constant and the model adequately represents the data. 3.4.1.5 Growth Discussion The growth results presented in Figure 12 were expected for batch fermentation of P. pastoris with glycerol as the carbon source. The yield ranged from 0.275 gDCW-g"1 glycerol - 0.625 gDCW-g'glycerol. This is a broad range but the experiment was only carried out for 20 hours at which time the medium was exchanged and induction started. This did not allow each culture to reach maximum cell density. Typically cell yields for P. pastoris grown in glycerol are approximately 0.5 gDCW-g"1 glycerol (Thorpe, d'Anjou et al. 1999). Yields higher than 0.5 gDCW-g"'glycerol were unexpected, however, peptone was used to supplement nitrogen in some run conditions. Peptone not only supplies additional nitrogen but also 72 provides an additional carbon source, since it is comprised of 50% carbon. The statistical analysis of the growth results identified 4 significant effects. The effect of nitrogen source relating to the level of ammonium sulfate has a significant affect on cell growth, which is supported by a p-value < 0.01. As evident from the response surface plot presented in Figure 14 and 15, ammonium sulfate concentration has a linear inverse relationship with growth. Although typical P. pastoris fermentation conditions utilize ammonium salts as the nitrogen source, there is evidence that more complex nitrogen sources such as peptone or individual amino acid supplements result in enhanced growth. By supplementing the media with peptone, an enzymatic digest of bacterial proteins, the cells are supplied with amino acids that can be directly used for protein synthesis. Peptone can also be broken down and utilized like an inorganic nitrogen source. Peptone has also been shown to improve cell growth in other expression systems including CHO and mouse hybridoma cell culture (Franeck 2003; Nyberg 1998; Jan 1994). Although peptone resulted in the highest final cell densities, a more thorough investigation is necessary, examining specific growth rates over the batch phase. These values would give a better indication of the affect of nitroge\i source on growth and would reduce the bias introduced by differences in carbon source quantities. The effect of temperature was found to have a significant impact on growth. A p-value of 0.04 was obtained for both the linear and second order terms relating to temperature indicating statistical significance. The surface response, Figure 14, shows that temperature has the most significant affect over the range of 21°C to 28°C increasing the final cell mass gDCWL" 1 as temperature increases. From 28°C to 35°C the effect of temperature is minor. This trend is represented by the quadratic influence of the model, and at temperatures of 35°C a decrease in growth is observed. The optimal range for temperature in terms of growth was identified as 26 - 30°C. Although maximal growth conditions were obtained over the range of 28°C to 35°C, a more important indicator of the system performance is productivity. For this reason, growth results were only used as a secondary indicator of performance. 73 Similarly, pH was also determined to have a significant effect on growth. A second order relationship with pH was supported by a p-value of 0.05. Examining the response surface plot (Figure 15) the effect of pH is described by a quadratic curve with an optimum of 6. Typical fermentation protocols recommend a pH of 6 for the P. pastoris expression system (Invitrogen 1999). Previous work by David Files (2000) also indicated that maximal productivity for recombinant cystatin C was obtained at pH of 6. 3.4.1.6 Productivity Results The productivity results for each run were based on a one-time response for each of the treatments. Cystatin C activity and yield measurements were performed using the papain inhibition assay outlined in the Materials and Methods. Samples were collected at 48 hours post-induction. Cell density measurements were also collected for 48-hour time points to determine cell specific productivity. Cystatin C productivity results for the screening design are presented in Figure 17. 74 u Q co 2 -»-> CO u 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 -\ 0 u m 1 2 3 4 5 6 7 J U QJ db U U 9 10 11 12 13 14 15 16 17 18 19 20 RUN# Figure 17: Cystatin C productivity results for screening design Yield of active cystatin C as a function of the run # (details outlined in Table 11). Cystatin C measurements were performed on supernatant samples collected at 48 hours. Error bars represent the standard error for each measurement. The Cystatin C cell specific productivity ranged from non-detectable levels (< 0.1 nmolgDCW'hr"1) to approximately 2.9 nmolgDCW'-hr - 1. The standard error for productivity results had a maximum value of +/-0.79 nmolgDCW'hr 1 . 3.4.1.7 Statistical Analys is The productivity results for the screening experiments are summarized in the following three-dimensional box plot (Figure 18). The eight vertices represent the eight different treatment combinations. The response values that are shown represent the average productivity (nmol-gDCW'-hr"1) obtained for the specific treatments. 75 -0.5006 0.78828 CO CO 2.96091 0.39541 -0.0364 1.37224 9.1 CN IT) 2.7713 0.32554 2 1 3 Temp °C 34.7 NH^HSO^ (g • L ) Figure 18 : Box plot for screening design productivity results The Box plot indicates the cystatin C productivity response (nmol-gDCW'-hr1) for all combinations of three factors. Values are derived from the Productivity model outlined in equation (7). In brackets are the ranges for the variables examined. The value outlined with a box indicates the productivity values (nmol-gDCW'-hr1). An analysis of variance was performed on the data and is presented in Table 22. Table 22 : ANOVA for screening design productivity results Source Nparm DF Sum of Squares F Ratio P-value Temp l l 1.38 7.69 0.02 P H l l 0.02 0.09 0.77 N H 4 1 1 5.11 28.36 0.00 Temp x pH l l 0.00 0.01 0.94 Temp x N H 4 1 1 1.47 8.18 0.02 pH x N H 4 1 1 0.05 0.26 0.62 Temp x Temp l i 0.44 2.42 0.15 p H x p H l l 0.26 1.45 0.26 N H 4 x N H 4 1 1 1.14 6.35 0.03 76 The factors determined to be significant were temperature, NH4HSO4, temperature x NH4HSO4. These effects are bolded in Table 23. Table 23 : Model analysis for screening design productivity results Term Estimate Std Error t Ratio P-Value Intercept 0.68 0.17 4.07 0.00 Temp -0.32 0.11 -2.77 0.02 pH -0.04 0.12 -0.30 0.77 N H 4 -0.69 0.13 -5.33 0.00 Temp x pH -0.01 0.15 -0.07 0.94 Temp x N H 4 0.43 0.15 2.86 0.02 pH x N H 4 -0.08 0.15 -0.51 0.62 Temp x Temp -0.17 0.11 -1.55 0.15 p H x p H 0.14 0.12 1.20 0.26 N H 4 x N H 4 0.36 0.14 2.52 0.03 The resulting equation obtained from the response surface analysis is comprised of the 5 significant terms identified in Table 23. This equation estimates the maximum productivity (nmolgDCW^hr1). The response equation is presented below. Productivity = 0.68 - 0.32 (T) - 0.69 (NHL,) + 0.43 (T)(NH4) + 0.36 (NH4)2 (7) Productivity Units = (nmoI-gDCW-'hr1) The model represents a second order relationship where the value of (T) and (NH4) are coded from -1 to 1 representing the range examined. A response surface plot based on the model is presented in Figure 19. This plot shows the impact of the studied variables on productivity. 77 Figure 19 : Response surface for screening design productivity results Response surfaces for growth as a function of N H 4 H S O 4 concentration (g-L1) and temperature. The response surface is derived from the statistical model and is represented by Equation 7. The pH was held constant at 6 for this response surface. The model adequacy was confirmed by examining a plot of the residuals vs. the predicted values as given below in Figure 20. 78 0.6 £ 0.4 ^ 0.2 U Q if 1 o I ^ 5 <» • 1—( > P3 - ° - 2 -0.4 o 2 -0.6 -0.8 • • 0.5 1.5 2.5 Productivity (nmol-gDCW ~ l-hr~ l) Predicted Figure 20 : Plot of residuals for screening design productivity model ' Residuals as a function of the predicted values derived from the model presented in Equation 7. The random scatter in the Figure 20 indicates that the variance is relatively constant and the model adequately represents the data. These response surfaces are based on the obtained model and indicate the relative effect of the significant variables impacting productivity. 3.4.1.8 Productivity Discussion The productivity results for the screening experiments ranged from 0 to 2.9 n m o l g D C W ' V 1 (Figure 17). Productivity was measured in terms of cell specific values to eliminate the variability in growth obtained for individual runs. The yield of recombinant cystatin C in these shake flask experiments was lower than previous cystatin C expression experiments where concentrations of approximately 3.5 nmol-gDCW"1 hr"1 were obtained (Files 2000). This earlier work, however, utilized a complex media 79 including yeast extract and peptone (YPD) resulting in higher cell densities and increased protein production. Although previous research has shown that supplementing the media with complex components such as yeast extract can improve both growth and productivity (Files 2000), defined medium recipes are preferred for large scale and industrial production applications. Defined medium reduces variability from batch to batch and increases the ease of downstream processing efforts to purify the product. For the screening experiments we utilized BMMG/BMMH medium. This defined medium containing ammonium sulfate was utilized as,the standard medium to examine nitrogen source impacts. As evident from the surface plot presented in Figure 19, the nitrogen source was a significant factor. As the ammonium sulfate concentration decreases and the nitrogen is supplemented with peptone, an increase in protein productivity is observed. The effect of nitrogen the source was confirmed with p-value of < 0.01, 0.02 and 0.03 for the main effects NH4HSO4, and the interaction terms, NF14HSO4 x temperature, and NH4HSO4 x NH4HSO4 respectively. The effect of nitrogen is less pronounced as the temperature increases from 21°C to 35°C, which is described by the temperature x NH4HSO4 interaction term. The conditions for maximum productivity were identified as 24°C and a high peptone (6.8-9.2 g-L"1) and no ammonium (Og-L"1). ThepH was determined to be insignificant and thus the standard pH conditions of 6 were used for future experiments. These findings are in agreement with reports in other host systems such as S. cerevisiae indicating that increasing ammonium concentration leads to a decrease in recombinant protein production (Chen, Yang et al. 1995; Chen, Chen et al. 1999). Other studies have corroborated this finding showing similar results in the expression of several heterologous a-amylases in S. cerevisiae (Rothstein, Lazarus et al. 1984; Rothstein, Lahners et al. 1987). Ammonia is known to inhibit glutamate dehydrogenase and, if present, may act as a negative feedback inhibitor of cellular productivity (Gawlitzek, Condradt et al. 1995). Ammonia concentration may also influence the intracellular pH of specific organelles including the ER, and the Golgi and lysosomes in murine cells (Ohkuma, Chudzik et al. 1981). These localized changes in pH can affect protein expression and potentially the secretion efficiency and cellular productivity. In many 80 cases researchers use a non-selective media containing yeast extract and peptone or casamino acids for secretion and characterization experiments. Alternative nitrogen sources, such as asparagine, have the potential to improve fermentation conditions by increasing cell mass and recombinant protein yield in S. cerevisiae (Chen, Chen et al. 1999). Peptone has been shown to increase protein productivity in other eukaryote expression systems including mouse hybridoma and CHO cells (Franeck et al. 2003; Nyberg et al. 1998; Jan et al. 1994). In one of these studies the addition of peptone to mouse hybridoma cells producing mAb resulted in an increase in protein concentration of 150% (Jan et al. 1994). The addition of di and tri-peptides to mouse hybridoma culture also resulted in similar findings increasing mAb production by 126-158%) (Franeck et al. 2003). Peptone, which was used to supplement nitrogen in these experiments, is an enzymatic digest of bacterial proteins. This digest provides a rich source of amino acids for the nitrogen source. These amino acids are precursors to protein synthesis and by ensuring sufficient concentrations we may increase the potential productivity of the cells. Peptone also improves productivity indirectly by acting as a substrate for extracellular proteases secreted by the host. The effect of these proteases, is minimized by peptone, a the competitive substrate of the recombinant product (Barr, Hopkins et al. 1993; Choi and Jimenez-Flores 2001). 3.4.1.9 Glycosylation Results To examine the effect of the variables on glycosylation, SDS PAGE (SDS polyacrylamide gel electrophoresis) gels were run to identify cystatin C species of different molecular weights. Gels were performed for each experimental run. No observable differences were noted for a most of the run conditions (data not shown). Under standard fermentation runs conditions where the nitrogen source was ammonium sulfate no observable glycosylated cystatin C was observed. Runs with high peptone however, resulted in the production of some glycosylated species. The results for these two run conditions are presented below in Figure 21. 81 Run # s t d 97.0 kDa 66.0 kDa 45.0 kDa 30.0 kDa Std . •4-DG SG UG Figure 21 : SDS PAGE gel results for screening design SDS-polyacrylamide gel electrophoresis of cystatin C. Samples were collected at 48 hours after start of fermentation. The molecular weight markers are shown in the lane marked "Std." and the corresponding molecular weights are identified to the left of the figure. Lanes 1 and 2 represent runs in which a high peptone and low ammonium concentration was used. Lane 3 represented the standard run conditions with only ammonium as the nitrogen source. Unglycosylated, single glycosylated and double glycosylated species are designated UG, SG and DG respectively. Unglycosylated recombinant cystatin C is present just above the 14.4kDa level and is indicated by (UG) (Figure 21). The single (SG) and double glycosylated (DG) bands represent species of higher molecular weight and are observed at ranges of 20 kDa and 25-30 kDa respectively (Figure 21). The single and double glycosylated species tend to appear as broader bands than the unglycosylated species due to the variability of the glycoforms attached to the protein. Comparison of the gel results shows that standard run conditions resulted in no observable production of glycosylated species (lane 3, Figure 21). However, the run conditions in which high peptone concentrations were used resulted in the production of both single and double glycosylated variants (lane 1 and 2, Figure 21). To ensure that the protein observed in the gel analysis is cystatin C, it was necessary to utilize both western blotting analysis and Endo H or FI cleavage combined with SDS PAGE analysis. These results are presented in section 3.4.2.9. 82 3.4.1.10 Glycosylation Discussion The glycosylation results show that the unglycosylated cystatin C is observed around a molecular weight slightly higher than 14.4 kDa (Figure 20). This is higher than the calculated molecular weight of 13.4 kDa. However, electrophoresis systems commonly show a band between 15 and 16 kDa for cystatin C (Abrahamson 1994). Although gels were performed for all experimental runs, no significant trends were observed for most run conditions. The gel data for the screening experimentation was difficult to interpret due to distorted bands. This may have been caused by high salt concentrations in the sample (Gomes 2002). The protein samples were obtained from supernatant. However, for the screening experiments very little protein was produced requiring the samples to be significantly concentrated up to 5-fold. Concentrating the samples led to increased salts and other media components, which in turn can affect the buffer for running the gels, resulting in broad bands. The gels for the screening experimentation were only examined for obvious visible conclusions. The most notable difference in glycosylation was observed in comparison of the standard run conditions vs. the high peptone concentration (low ammonium sulfate) conditions (lane 3 vs. lane 1 and 2, Figure 20). These results indicate that the nitrogen source plays a critical role in protein production and glycosylation. Early studies with P. pastoris have shown that addition of peptone to the medium resulted in slightly increased production of glycosylated P-casein production (Choi and Jimenez-Flores 2001). The peptone added was in addition to the existing nitrogen source, which was ammonium sulfate. Studies in mammalian cell culture suggest that a key factor affecting glycosylation is the ammonium concentration in the medium. Ammonia concentrations increasing from 5 to 50 mM resulted in decreased levels of glycosylated protein in mammalian cell culture (Goochee and Monica 1990; Yang and Butler 2000). It is unclear at this point whether the increase in glycosylation observed is due to the presence of peptone as a complex nitrogen source or whether the presence of ammonium is the overriding factor causing a decrease in glycosylation. Future work will examine the individual affects of different nitrogen sources on cystatin C glycoprotein production 83 Modifications to temperature and pH did not result in any observable differences in glycosylation over the range examined. Temperature and pH may still play a role although the impact of these variables is less significant than the nitrogen source. The optimal process conditions for glycosylation identified from this design were high concentrations of peptone, approximately 6.8 - 9.2 g-L -1, and no ammonium sulfate, 0 g-L-1. Temperature and pH were not identified as significant variables over the range examined and thus the conditions for pH and temperature determined from the productivity analysis, were used for future experiments. 3.4.2 2-Litre Bioreactor Optimization 3.4.2.1 Background An experimental design was used for large-scale experimentation to speed the optimization process by minimizing the experimentation required while maximizing the amount of information gathered. The results from the screening experiments provided the basis for the 2-Litre bioreactor optimization. From the factors investigated previously the most significant variable was determined to be the nitrogen source. The effects of ammonium and peptone were investigated, however due to the experimental design selected, the individual affects of the nitrogen sources could not be deduced. The focus of the large-scale 2-litre bioreactor experiments was to examine nitrogen sources including 1) ammonium hydroxide, 2) peptone, and 3) amino acid supplements. 3.4.2.2 Design Details A 2-litre Applikon bioreactor and 2-litre LH bioreactor were used with media conditions, inoculum preparation, and sampling as outlined in Materials and Methods. The standard fermentation conditions for protein expression in P. pastoris suggest a temperature of 30°C (Invitrogen 1999). However 28°C was chosen based on results from the screening experiments. A pH of 6 was also selected based on results from screening experiments. Ammonium sulfate or ammonium hydroxide is typically used for 2-litre bioreactor fermentations. The concentration of peptone and amino acid supplements were determined from a nitrogen balance using the typical ammonium hydroxide conditions as 84 the standard. A completely randomized design was used for the experiments and replicate samples were taken for each time point examined. 3.4.2.3 Growth Results The fermentor was operated in batch mode for approximately 17 hours after which time the initial glycerol in the medium was exhausted. The system was then switched to a fed batch protocol with the same carbon source. The glycerol fed batch phase lasted approximately 4 hours and was ended when the cell density reached 200 gDCW-L"1 (50 gDCW-L"1). A target cell density of 200 gWCW-L"! was used for all runs to eliminate any effect of cell concentration on productivity and glycosylation. Induction was started by switching the carbon source to methanol. Samples were taken at various time points to determine cell density and active cystatin C concentrations. The growth profiles obtained for the 2-litre bioreactor runs are presented in Figure 22. 250 -j Time (hours) —•— Ammonium hydroxide -A- Ammonium hydroxide + A A - A - Peptone —x— Ammonium hydroxide + Peptone —•— A A + Peptone - • - A A —o— Ammonium hydroxide + Peptone + A A Figure 22 : Cell density results for bioreactor experiments Dry cell weight as a function of culture time, (details for runs outlined in Table 11). 85 After induction very little growth was observed for all runs. Final cell densities were approximately 200 gWCWL" 1 for each run, and represent typical cell concentrations obtained in P. pastoris fed batch cultures (Lesnicki 2001). The WCW values can be converted to DCW values using the calibration curve presented in the Materials and Methods. 3.4.2.4 Productivity Results The Cystatin C activity and yield measurements were determined using the papain inhibition assay outlined in the Materials and Methods. Cell density measurements were also performed to determine cell specific productivity. The cell specific productivity results were calculated for each sample point for the induction phase of growth. Cystatin C yield results for the 2-litre bioreactor are presented in Figure 23. 0.9 -i Time (hours) "•"Ammonium hydroxide + AA -O-Ammonium hydroxide -iV" Peptone -O-Ammonium hydroxide + Peptone - • " A A + Peptone Ammonium hydroxide + AA + Peptone Figure 23 : Cystatin C yield results for bioreactor experiments Cystatin C yield (pmol-gDCW"1) as a function of culture time (hours) for 2-litre bioreactor experiments. The nitrogen sources used in each run are identified in the legend. Al l other process conditions were kept constant. Details for runs outlined in Table 11. 86 The maximum cystatin C yield obtained ranges from 0.15 p m o l g D C W 1 -0.8 nmolgDCW' 1 over the design range. The standard error for the cystatin C measurements had a maximum of +/- 0.11 umolgDCW 1 . The maximum productivity from each of the experimental runs was also determined using the cell density and cystatin C yield data. The productivity data is presented in Figure 24. 16 1 2 3 4 5 6 7 Run# Figure 24: Cystatin C productivity results for bioreactor experiments Cystatin C productivity for 2-litre bioreactor experiments. Productivity (nmoI-gDCW 1) is plotted as a function of the Run # (details for the runs are outlined in Table 11). The maximum cystatin C productivities obtained ranges from approximately 4.3 nmol-gDCW'hr' 1 - 13.5 nmol-gDCW'-hr' 1 over the design. The standard error for the productivity measurements has a maximum of +/- 2.3 n m o l g D C W ^ h r 1 . 3.4.2.5 Statistical Analysis The productivity results for the large-scale experiments are summarized in the following three-dimensional box plot (Figure 25). The eight vertices represent the eight different 87 treatment combinations. The response values that are shown represent the average productivity (nmol-gDCW1-hr"1) obtained for the specific treatments. 13.6113 o 11.6587 d '—v ST * 4.09867 0.16133 11.9687 12.9413 5.87133 20 4.85933 A m i n o A c i d s o ( g - L " 1 ) 0 30 N H 4 0 H ( g - I / 1 ) Figure 25 : Box plot for bioreactor productivity results The Box plot indicates the cystatin C productivity response (nmol-gDCW*1-hr"1) for all combinations of three factors. Values are derived from the productivity model outlined in equation (8). An analysis of variance was performed on the data and is presented in Table 24. Table 24 : A N O V A for 2-litre bioreactor productivity results Source Nparm DF Sum of Squares F Ratio P-value N H 4 1 1 5.0 30.3 0.03 Peptone i l 165.8 1003.8 0.00 Amino Acids l l 4.7 28.5 0.03 N H 4 x Peptone 1 1 6.2 37.8 0.03 N H 4 x Amino Acids 1 1 4.6 27.7 0.03 Peptone x Amino Acids 1 1 2.1 12.8 0.07 88 The variables or interactions that were identified as significant include, peptone, ammonium hydroxide, amino acids, ammonium hydroxide x peptone and ammonium hydroxide x amino acids. These effects are bolded in Table 24. The t-ratio and a two-tailed probability test for each of the model coefficients were calculated. The results from this analysis are presented in Table 25. Table 25 : Model analysis for 2-litre bioreactor productivity results Term Estimate Std Error t Ratio P-Value Intercept 8.15 0.14 58.7 0.00 N H 4 0.76 0.14 5.5 0.03 Peptone 4.40 0.14 31.7 0.00 Amino Acids 0.74 0.14 5.3 0.03 N H 4 x Peptone -0.85 0.14 -6.2 0.03 N H 4 x Amino Acids -0.73 0.14 -5.3 0.03 Peptone x Amino Acids -0.50 0.14 -3.6 0.07 The resulting equation obtained from the response surface analysis is comprised of the 6 significant terms identified in Table 25. This equation estimates the maximum productivity ( n m o l g D C W " 1 h r " ' ) . The response equation is presented below. Productivity = 8.15 - 0.76(NH4OH) + 4.40(P) + 0.74(AA) - 0.85(NH4OH)(P) -0.73(NH4OH)(AA) (8) Productivity Units = (nmolgDCW'hr1) 89 The model represents a second order relationship where the value for Peptone, ammonium hydroxide, and amino acids are coded from -1 to 1 representing the range examined. Response plots based on the model is presented in Figures 26 and 27. These plots show the impact of the significant variables on productivity. Figure 26 : Response surface for bioreactor productivity results Response plot for productivity (nmol-gDCW'-hr 1 ) as a function of Peptone (g-L"1) and ammonium hydroxide (g-L"1) concentration. The response plot is derived from the statistical model and is represented by Equation 8. 90 Peptone (g-L') Amino Aci ds (g L') Figure 27 : Response surface for bioreactor productivity results Response plot for productivity (nmol-gDCW'-hr 1 ) as a function of Peptone (g-L"1) and amino acid (g-L - 1) concentration. The response plot is derived from the statistical model and is represented by Equation 8. The model adequacy was confirmed by examining a plot of the residuals vs. the predicted values as given below in Figure 28. 91 0.4 0.3 O Q 11 o 0.2 0.1 0 -0.1 -0.2 5 10 • • -0.3 Productivity (nmol-gDCW "^ hr"1) Predicted Figure 28 : Plot of residuals for bioreactor productivity model Residuals as a function of the predicted values derived from the model presented in Equation 8. The random scatter in the Figure 28 indicates that the variance is relatively constant and the model adequately represents the data. These response surfaces are based on the obtained model and indicate the relative effect of the significant variables impacting productivity. 3.4.2.6 Productivity Discussion The recombinant cystatin C yields for the 2-litre bioreactor experiments ranged from 4.3 nmol-gDCW"1 - 13.5 nmol-gDCW"1. These results are comparable to previous cystatin C productivities obtained in 2-litre bioreactor studies (Files 2000). The productivity of recombinant cystatin C in the 2-litre bioreactor experiments was much higher than the yields obtained in the shake flask design. Maximum productivity was 2.9 nmol-gDCW"1 hr"1 in shake flask and 13.5 nmol-gDCW"1 hr"1 in the 2-litre bioreactor. This improvement in cystatin C productivity is due mainly to the superior mixing, oxygen and feed control in the 2-litre bioreactor. 92 Previous research has shown that supplementing the media with complex components such as yeast extract can improve both growth and productivity (Files 2000). Screening experiments also identified that the nitrogen source was an important factor in growth and productivity. However, these experiments failed to elucidate the individual effects of ammonium and peptone. For the 2-litre bioreactor experiments the individual impact of ammonium and peptone were examined. Amino acid supplements were also studied as an alternative nitrogen source providing the amino acids present in peptone while maintaining a defined media with no peptides. Statistical analysis of the data showed that the key variable affecting productivity was peptone. The response surface models presented in Figures 26 and 27 show a significant increase, greater than 2-fold, as peptone is varied from 0-20 g-L"1. Amino acids and ammonium hydroxide also affected the maximum productivity, however, the impact was less significant than peptone. Ammonium hydroxide concentration influences the productivity as a main effect but also interacts with the other nitrogen sources. This trend is apparent upon examination of the model equation 8 showing that ammonium has a negative influence on productivity at high levels of peptone (20 g-L"1) and amino acids (20 g-L"1). Previous findings have shown that increasing ammonium concentration lead to a decrease in HbsAg expression in S. cerevisiae (Chen, Chen et al. 1999). Other reports indicated that ammonium inhibits the secretion of a-amylase in P. pastoris (Rothstein, Lazarus et al. 1984). As mentioned previously ammonia is also known to inhibit glutamate dehydrogenase and, if present, may act as a negative feedback inhibitor onto cellular productivity (Gawlitzek, Condradt et al. 1995). Ammonium concentration may also influence the intracellular pH of specific organelles including the endoplasmic reticulum, the Golgi and lysosomes in murine cells (Orlean 1992). These localized changes in pH can affect protein expression and potentially the secretion efficiency and cellular productivity. Although most of the previous studies pertain to S cerevisiae these findings add support to the results observed in the P. pastoris system. 93 Peptone, which was found to be the most significant factor, is a common ingredient in rich media cultures known to provide improved growth and productivity compared to defined media recipes. Peptone is an enzymatic digest of bacterial proteins, which provides a rich source of amino acids to the cells in the form of both short chain oligopeptides and free amino acids (see peptone composition Table 8 Materials and Methods). Amino acids are precursors to protein synthesis and by ensuring sufficient concentrations the productivity of the cells may be increased. Peptone has been shown to improve productivity and cell growth in mammalian cell cultures including mouse hybridoma and CHO cells (Franeck et al. 2003; Nyberg et al. 1998; Jan et al. 1994). Cultures supplemented with peptone resulted in improvements in protein production as much as 150% compared (Jan et al. 1994). Peptides have also been added to mammalian cell culture to improve productivity. The production of mAbs in mouse hybridoma cells has been increased by 126% and 158% by adding tri-peptides Lys-Lys-Lys and Gly-Lys-Gly respectively (Franeck et al. 2003). These results show that it is not only the presence of peptides but the peptide composition that affect protein productivity. It has been suggested that peptide molecules act by hitting specific targets in the cell, imposing alterations that affect cell metabolism and the cell proliferation mechanism (Franeck et al. 2003). Peptone may also improve productivity indirectly by acting as a substrate for extracellular proteases secreted by the host. Proteases can degrade the recombinant product, but in the presence of peptone, a competitive substrate, this effect can be significantly reduced (Barr, Hopkins et al. 1993; Choi and Jimenez-Flores 2001). There are a significant number of additional studies performed with other eukaryotic expression systems, mainly S. cerevisiae, which corroborate these findings, showing the positive influence of peptone and peptides on growth and productivity (Mendoza-Vega, Sabatie et al. 1994; Udeh and Achremowicz 1997; Alexeeva, Ivanova et al. 2002). Although peptone has been shown to provide improved fermentation conditions for growth and protein production, defined medium recipes are still preferred for large-scale industrial applications. Defined media ensures minimal variability from batch to batch and generally increases the ease of downstream processing efforts to purify the product. For these reasons the addition of amino acids as a nitrogen source was examined. Using a recipe of individual amino acids developed for protein expression in S. cerevisiae 94 (Ausubel, Brent et al. 1999) we maintained a defined media with known concentrations. The individual amino acid supplements, unlike peptone do not contain peptides, which cause downstream processing complications. Examining the response surface in Figure 27 it is clear that addition of amino acid supplements to the culture media improves protein productivity at low peptone concentrations. When compared to the standard fermentation conditions, where only ammonium hydroxide is used, amino acids also improve productivity (Figure 27). However at high peptone concentrations amino acids have a slight negative impact on productivity (Figure 27). The negative impact on productivity at high peptone concentrations may be due to differences in osmolarity. For fermentor runs where multiple nitrogen sources are used the higher level of media salts can affect cell growth and productivity (Mendoza-Vega, Sabatie et al. 1994). This may account for part of the negative impact observed for the second order interactions at high peptone concentrations. In previous studies alternative nitrogen sources, such as asparagine, have also shown the potential to improve fermentation conditions by increasing cell mass and recombinant protein yield (Chen, Chen et al. 1999). Studies with S. cerevisiae have reported increased growth and productivity with specific amino acid supplements such as histidine (Shiba, Fukui et al. 1998). It has also been reported that amino acid supplements can be used as a defined substitute for the more complex nitrogen sources in S. cerevisiae cultures (Mendoza-Vega, Sabatie et al. 1994). In our findings the addition of amino acids to culture has been shown to improve productivity, however, the impact of amino acid supplements on productivity was not as significant as the affect of peptone. Other studies have shown that amino acid supplements are not as effective as peptone at improving productivity. The production of mAbs in mouse hybridoma cells was only increased by 50% with amino acid addition compared to an increase of 125-150% with peptone addition (Jan et al. 1994). The composition of the amino nitrogen source also plays a role. Franeck (Franeck et al. 2003) reported a broad range of productivity improvements ranging from 0-158% depending on the amino composition of the peptide supplements. Overall findings indicate that the addition of peptides to the culture media can improve protein productivity. However the impact of peptides on productivity is dependant on peptide composition. 95 Analysis of the 2-litre bioreactor results indicated that all of the nitrogen sources examined had an impact on productivity. The optimal process conditions identified based on the model presented resulted in a maximum productivity of 13.5 nmol-gDCW"1 hr"1. This result was obtained when peptone and amino acids were present at 20 g-L"1 and no ammonium hydroxide was used. It would be advantageous to investigate the variables outside the range examined to determine whether productivity could be improved further. In addition experiments examining center points for this design could also be used to elucidate any quadratic affects of the variables. The amino acid mixture used was not optimized for the P. pastoris system and more experimentation is also necessary to determine the optimal amino acids composition for protein productivity. It may also be advantageous to examine nitrogen sources such as defined peptide media as an alternative to amino acids supplements and peptone. This type of media may provide the advantages gained by the presence of peptides while still maintaining a defined medium composition. Although it has been shown that nitrogen source had a significant impact of productivity there are still other medium components, such as the carbon source and trace elements including: KH2PO4, MgSCu, NaCl and Q1SO4 that should be examined to further improve the expression of recombinant cystatin C in P. pastoris. CUSO4 in particular has been shown to increase protein production in S. cerevisiae (Vad, Moe et al. 1998; Wolff, Hansen et al. 2001). 3.4.2.7 Glycosylation Results To determine the effect of the different treatments on glycosylation, samples were collected at various time points post-induction. SDS gel electrophoresis was used to compare the relative quantity of different cystatin C species produced in culture. Due to the variability in cystatin C concentrations found for different runs, it was necessary to concentrate the samples to obtain enough protein for the SDS PAGE analysis. The samples were concentrated by using the Trichloroacetic acid (TCA) precipitation method outlined in the Materials and Methods. Samples 1 and 6 were run in duplicate with the second sample at half the concentration. The results for the SDS gel electrophoresis are 96 presented in Figures 29 and 30. 97.0 kDa 66.0 kDa 45.0 kDa 30.0 kDa 2-Litre Bioreactor •o Experiments a 1 1 1 2 3 D G Figure 29 : SDS P A G E gel results for bioreactor experiments SDS-polyacrylamide gel electrophoresis of cystatin C. Samples were collected at 48 hours after start of fermentation. The molecular weight markers are shown in the lane marked "Standard." and the corresponding molecular weights are identified to the left of the figure. The numbers at the top of the figure represent the run numbers (run details outlined in Table 12). Unglycosylated, single glycosylated and double glycosylated species are designated U G , S G and D G respectively. 97 T J C J Run # I 97.0 kDa 66.0 kDa 45.0 kDa 30.0 kDa Large Scale Fermentation s o U G Figure 30 : SDS PAGE gel results for bioreactor experiments SDS-polyacrylamide gel electrophoresis of cystatin C. Samples were collected at 48 hours after start of fermentation. The molecular weight markers are shown in the lane marked "Standard." and the corresponding molecular weights are identified to the left of the figure. The numbers at the top of the figure represent the run numbers (run details outlined in Table 12). Unglycosylated, single glycosylated and double glycosylated species are designated UG, SG and DG respectively. It is evident upon inspection of the gel results presented in Figure 29 and 30 that the cystatin C produced under different nitrogen sources results in variable macroheterogeneity. However to utilize the statistical tools available and analyze the data it was necessary to semi-quantitate the results using imaging software. Scion imaging software (www.scioncorp.com) was used for this analysis (see Material and Methods). The gels were scanned and the relative peak intensities were determined. A typical scan resulted in an output that was displayed by 1-3 peaks. Peak analysis by evaluation of the area under the curve was used to compare relative peak intensities. A typical peak analysis is presented in Figure 31. 98 Figure 31 : Peak analysis for bioreactor gels Typical peak analysis of SDS gel electrophoresis for recombinant cystatin C in culture. UG, SG, and DG species are identified. The area under the curve represents the relative peak intensities. Due to the close proximity of the single and unglycosylated peaks it was necessary to estimate the division of these peaks to perform the peak analysis. The results of the peak analysis are presented in Table 26 and represent the relative quantities of single and double glycosylated species compared to the unglycosylated form. Table 26 : Glycosylation results for 2-litre bioreactor experiments glycosylated (%) SG (%) DG (%) 6 N H 4 100 0 7 N H 4 100 0 2 30 Peptone x A A 65 35 3 26 Peptone 62 38 3 28 Peptone 62 38 4 10 NH 4 x A A 50 50 5 12 NH 4 x Peptone 54 46 6 16 Peptone x A A x NH4 52 48 7 23 A A 50 50 Some immediate trends are obvious upon examination of Table 26. Peptone in the media resulted in the highest total glycosylated species of approximately 30%. However to elucidate other less obvious effects, the data from Table 26 was analyzed using statistical software. 3.4.2.8 Statistical Analysis The glycosylation results for the large-scale experiments are summarized in the following three-dimensional box plot (Figure 32). The eight vertices represent the eight different treatment combinations. The response values that are shown represent the total amount of glycosylated species (%) obtained for the specific treatments. 29.7333 R 127.2667 0) C ^ J J I—I OH * CD O O 12.2667 -0.2667 16.2667 22.7333 9.73333 20 6.13333 A m i n o A c i d s ( g - L " 1 ) 0 30 N H 4 0 H ( g . L / 1 ) Figure 32 : Box plot for bioreactor glycosylation results Box plot to indicate the cystatin C glycosylation response (%) for all combinations of three factors. Values are derived from the Productivity model outlined in equation (8). In brackets are the ranges for the variables examined. The value outlined with a box indicates the glycosylation values (%). An analysis of variance was performed on the data presented in Table 27. 100 Table 27 : ANOVA for 2-litre bioreactor glycosylation results Source Nparm DF Sum of Squares F Ratio P-value N H 4 1 1 26.8 21.1 0.04 Peptone 1 i 621.7 490.8 0.00 Amino Acids l l 19.7 15.6 0.06 N H 4 x Peptone 1 1 64.0 50.6 0.02 N H 4 x Amino Acids 1 1 42.8 33.8 0.03 Peptone x Amino Acids 1 1 54.3 42.9 0.02 The t-ratio and a two-tailed probability test for each of the model coefficients were calculated. The variables or interactions that were identified as significant include, peptone, ammonium hydroxide, ammonium x peptone, ammonium x amino acids, and peptone x amino acids. These effects are bolded in Table 27. The response surface analysis is presented in Table 28. Table 28 : Model analysis for 2-litre bioreactor glycosylation results Term Estimate Std Error t Ratio P-Value Intercept 15.5 0.38 40.3 0.00 N H 4 -1.8 0.38 -4.6 0.04 Peptone 8.5 0.38 22.2 0.00 Amino Acids 1.5 0.38 4.0 0.06 N H 4 x Peptone -2.7 0.38 -7.1 0.02 N H 4 x Amino Acids -2.2 0.38 -5.8 0.03 Peptone x Amino Acids -2.5 0.38 -6.6 0.02 The resulting equation obtained from the response surface analysis is comprised of the 6 significant terms identified in Table 28 This equation estimates the level glycosylation for the given parameters over the range examined. The response equation is presented here. 101 Glycosylation = 15.5-1.8(NH4OH) + 8.5(P) - 2.7(NH4OH)(P) -2.2(NH4OH)(AA) - 2.5(P)(AA) (9) Glycosylation units = (%) The model represents a second order relationship where the value for peptone and amino acids are coded from -1 to 1 representing the range examined. Response plots based on the model are presented in Figure 33, 34, and 35. These plots show the impact of the variables on glycosylation. Figure 33 : Response surface for bioreactor glycosylation results Response plot for glycosylation (%) as a function of Peptone (g-L"1) and ammonium hydroxide concentration. The response plot is derived from the statistical model and is represented by Equation 9. 102 Peptone (g-L') 0 0 Amino Acids (g-L*) Figure 34 : Response surface for bioreactor glycosylation results Response plot for glycosylation (%) as a function of Peptone (g-L 1) and amino acid (g-L"') concentration. The response plot is derived from the statistical model and is represented by Equation 9. 103 Figure 35 : Response surface for bioreactor glycosylation results Response plot for glycosylation (%) as a function of amino acid (g-L"1) and ammonium hydroxide (g-L 1) concentration. The response plot is derived from the statistical model and is represented by Equation 9. The model adequacy was confirmed by examining a plot of the residuals vs. the predicted values as given below in Figure 36. 104 m o o o 0.5 -0.5 -1.5 • • • • 0 1 1 1 5 10 15 i i i 20 25 30 35 • Glycosylation (%) Predicted Figure 36 : Residuals plot for bioreactor glycosylation model Residuals as a function of the predicted values derived from the model presented in Equation 9. The random scatter in Figure 36 indicates that the variance is relatively constant and the model adequately represents the data. These response surfaces are based on the obtained model and indicate the relative effect of the significant variables impacting glycosylation. 3.4.2.9 Verification of Glycosylated Cystatin C Previous experiments were performed in which the 3 cystatin C bands identified in Figures 29 and 30 were verified by western blot analysis using rabbit antiserum that was raised against cystatin C (DAKO Corp., CA) (Files 2000). Endo FI cleavage was also performed on bioreactor samples of recombinant cystatin C to cleave the carbohydrate groups leaving only the unglycosylated species. The results for this analysis are presented in Figure 37. 105 Figure 37 : SDS PAGE gels for ENDO FI analysis SDS-polyacrylamide gel electrophoresis of cystatin C. The molecular weight markers are shown in the lane marked "Std.". Runs labelled untreated represent runs in which high peptone and low ammonium concentrations were used. Runs labelled treated represent runs in which high peptone and low ammonium concentration were used and then the sample was treated with ENDO FI to remove carbohydrate structures. Unglycosylated, single glycosylated and double glycosylated species are designated UG, SG and DG respectively. In Figure 37 an untreated sample from the bioreactor (lane 1) is compared with a sample treated with Endo FI (lane 2). The results indicate that all of the DG cystatin C was cleaved indicated by the absence of a band in the 25 - 30 kDa range (lane 2). The SG cystatin C however was only partially cleaved resulting in a significantly weaker band at 20 kDa (lane 2). 3.4.2.10 Glycosylation Discussion Previous results from screening experiments indicated that nitrogen source plays a significant role in glycosylation. Standard run conditions were compared to runs with high peptone and low ammonium hydroxide. The screening study showed that almost no glycosylated cystatin C was produced under standard run conditions. However, in the presence of peptone and low ammonium concentrations both single and double 106 glycosylated cystatin C species were produced. To further elucidate the effects of nitrogen source on glycosylation, 2-litre bioreactor design experiments were performed to examine the affects of ammonium hydroxide, peptone, and amino acids at different levels in the culture media. The results for the 2-litre bioreactor optimization represent the first report to identify and quantify the individual affects of peptone, amino acids and ammonium concentration for this host organism. The results from the imaging analysis, Table 26, show that the range of glycosylation varied from 6 - 30 % indicating that the nitrogen source has a significant impact on glycosylation. It is also evident from Table 26 that the ratio of single to double glycosylated is around roughly 50% indicating that an equal amount of single and double glycosylated species are present in the culture. For the standard run conditions however, very little glycosylated cystatin C was observed and no detectable levels of double glycosylated species were present. Repeats of the gels were performed to determine the variability in the gel and imaging analysis. Findings presented in Table 26 show that variability in the gel analysis was +/- 1-2% resulting in a standard error of 5 - 21%. Other more sophisticated techniques are available to evaluate glycosylation (see Literature Review section). However, for this study the objective was to identify differences in site occupancy and determine relative amounts of glycosylated species produced. Gel analysis provides an efficient method and as indicated previously shows consistent reproducibility. Statistical analysis of the data indicated that the key variable affecting glycosylation was peptone having a p-value less than 0.01 showing statistical significance. The response model presented in Figures 33 & 34 shows that as peptone concentration is increased from 0-20 g-L"1 glycosylation increases greater than 5-fold. The equation that represents this system has a coefficient of 8.5 for the peptone term with the second most significant term having a much lesser impact with a coefficient of 2.7. Peptone supplies a rich mixture of amino acids to the cells, which can be directly utilized for protein synthesis. Using simple nitrogen sources for fermentation requires cells to synthesize the necessary amino acids, which are precursors to protein production. This adds an additional metabolic burden to the cell culture and may influence glycosylation. Earlier studies with P. pastoris have shown that addition of peptone to the media resulted in a slight increase in glycosylated p-casein production (Choi and Jimenez-Flores 2001). However, in this 107 previous study the results were not quantified nor did they elucidate the interaction of other nitrogen sources in the media such as ammonium hydroxide. It has also been reported that the presence of peptides in peptone act on specific targets in the cell which may affect the cell metabolism and cell proliferation (Franeck et al. 2003). These cellular changes have been shown to have a positive impact on protein productivity. This positive impact observed for productivity echoes the affects that peptone has been shown to have on glycosylation in the experiments performed. The ammonium concentration also has an affect on glycosylation. However its affect is opposite of peptone and leads to a decrease in the amount of glycosylation. The affect of ammonium hydroxide is greatest under conditions of high peptone. This is due to the interaction effects of the variables. There are a significant number of findings in the reviewed literature for other expression systems that lend support the observed affect of ammonium and peptone in the P. pastoris system. It is has been reported for many mammalian cell culture systems that increased levels of ammonium often result in decreased glycosylation. In the P. pastoris expression system ammonium hydroxide alone is used as the standard nitrogen source and exists at concentrations in excess of 80 mM in culture. Ammonia concentrations in cell culture ranging from 2-50 mM have been shown to decrease the level of glycosylation h-EPO in CHO cells (Goochee, Gramer et al. 1991; Yang and Butler 2000). Borys and colleagues also showed that ammonium levels from 0-9mM reduced glycosylation as much as 65% for the production of mPL-1 in CHO cells (Borys 1994). Ammonium in mammalian cell culture has been reported to accumulate in pH sensitive, acidic, intracellular compartments, such as the trans-Golgi, to concentrations greater than their extracellular concentration (Goochee and Monica 1990). This accumulation results in a localized increase in pH which may alter the activity of enzymes associated with oligosaccharide and protein processing in these compartments (Dean, Jessup et al. 1984; von Zastrow, Castle et al. 1989). It is evident from the literature review that ammonium can negatively impact glycosylation in mammalian cell culture. Although the P. pastoris expression is significantly different in many respects to mammalian systems, many of the initial steps in glycosylation and protein trafficking are conserved. The observations with respect to 108 ammonium concentration in the expression of recombinant cystatin C in P. pastoris seem to be in agreement with ammonium effects observed in some mammalian cell culture systems. We have not been able to identify any previous reports on the effects of ammonium concentration on glycosylation in other yeast systems. Although the addition of peptone to the media resulted in the optimum glycosylation, it is still a less desirable media component, which results in downstream processing complications and media variability from batch to batch. Amino acids supplements were examined to determine if similar productivity and glycosylation could be obtained by utilizing a defined media supplying a similar composition to peptone. Observing the trends in Figures 34 and 35 we see that amino acids provide a superior nitrogen source to ammonium hydroxide alone. However the improvements in glycosylation are not as significant as the effects of peptone. The presence of amino acids at high concentrations of peptone resulted in a negative influence on glycosylation (Figure 34). However, some of the negative impacts observed for combined nitrogen source effects may be due to changes in osmolarity caused by an increased level of solutes in the media (Mendoza-Vega, Sabatie et al. 1994). The amino acid mixture used was based on a formulation that was recommended for S. cerevisiae (Ausubel, Brent et al. 1999). Studies have reported productivity improvements ranging from 0-158% depending on the amino composition of peptide supplements added to murine cell culture for mAb production (Franeck et al. 2003). The affects of amino acid composition might also apply to free amino acid supplements added to the culture medium. The medium used in the 2-litre bioreactor runs only utilized a mixture of 14 amino acids at concentrations that were designed for the host S cerevisiae and no optimization of amino acid concentrations was attempted. Although there is still potential to improve glycosylation further by optimizing the amino acid concentrations, findings have shown that amino acid supplements are not as effective as peptone at improving productivity. The production of mAbs in mouse hybridoma cells was only increased by 50% with amino acid addition compared to an increase of 125-150% with peptone addition (Jan et al. 1994). Optimization of the nitrogen source resulted in a maximum glycosylation of approximately 30 %. The addition of peptone (20 g-L"1) and amino acids (20 g-L"1) and 109 the absence of ammonium hydroxide from the standard fermentation protocol provided these conditions. The optimization of nitrogen source has been shown to play a critical role in production of glycosylated recombinant cystatin C in P. pastoris. However, there are still other process conditions, including carbon source and trace elements that must be examined to further improve this system. The effect of carbon source is a key variable in P. pastoris cultures. Previous findings have shown that glucose starvation can cause a decrease in site occupancy of oligosaccharides at light chain asparaginyl sites that are normally glycosylated (Stark and Heath 1979). This effect was also seen in CHO cells under glucose starvation (Davidson and Hunt 1985). Overall glucose limitations generally reduce glycosylation, which can lead to undesirable variable heterogeneity. The carbon source and concentration directly affect the availability of sugars that are transferred to the protein by the GDP mannose and dolichol-P mannose complexes (Rearick, 1981; Rearick, 1981). This has been shown to result in shorter oligosaccharides and decreased site occupancy. These findings are particularly important to the P. pastoris expression system. P. pastoris requires a methanol feed protocol providing limiting methanol concentrations. The use of limiting concentrations may lead to the reduced glycosylation observed in mammalian expression systems. There are some alternative feeding strategies including mixed feeding with glycerol or sorbitol, which may have an influence on glycosylation. There is potential further improve the production of glycosylated cystatin C through carbon source and feed strategy optimization. To confirm the existence of glycosylated cystatin C, Endo FI cleavage of the carbohydrate groups was performed. Examination of the gels in Figure 36 shows that the Endo FI treatment did not completely remove the band present at 20kDa (lane treated). This suggests that either the Endo FI treatment was incomplete which might be due to reaction kinetics or difficulty in cleavage at one of the glycosylation sites. There is also the possibility that another protein may exist at the same molecular weight that was present in the media. This scenario is unlikely but cannot be ruled out. More sophisticated analysis utilizing mass spectroscopy or chromatography is necessary to characterize the different glycosylated cystatin C species. 110 3 . 5 Microarray Process Optimization Results 3.5. / Introduction Process development and optimization of cell culture and fermentation processes often involves the use of statistically designed experiments. However, when using this approach, the results are empirical and the number of experiments required to evaluate a few variables can be significant. The complexity of cell physiology and the number of potential variables makes this approach time consuming and iterative. Typical variables that are monitored and optimized may include: dissolved oxygen, pH, mixing speed, temperature, osmolarity, and media composition. The complexity of media alone can make optimization a formidable project. D N A micro arrays have the potential to monitor the expression profiles of thousands of genes. This technique is very powerful because it captures the actual physiological state of the culture and has the potential to identify or examine multiple limitations in a single experiment. The objective of this research is to show that unique expression profiles can be used to identify specific media limitations in cell culture and fermentation processes. Examining the current literature on microarray technology there is evidence that different nutrient limitations result in unique and distinguishable changes in gene expression when compared to a control where nutrients are not limited. Gasch (Gasch, Spellman et al. 2000) showed that complete amino acid starvation and nitrogen depletion resulted in significant changes in global gene expression. The Natarajan (Natarajan, Meyer et al. 2001) study examined effects of a histidine starvation imposed by adding 3AT a competitive inhibitor of His3p. Starvation by competitive inhibition however, may not represent the same gene expression changes that would be observed under histidine depletion in the media. Most of the current work has focused on global changes in gene expression and the ESR (environmental stress response) (Gasch, Spellman et al. 2000). The objective of our research is to determine whether unique expression profiles might indicate a specific limitation. To examine this hypothesis the existing literature was re-analyzed to identify discernable patterns in gene expression with respect to amino acid limitations. The data from these studies were re-analyzed to determine common 111 distinguishable gene expression patterns. In addition shake flask experiments were also performed to monitor gene expression profiles in S. cerevisiae grown under specific amino acid limitations. 3.5.2 Gasch, Spellman et al. Results Re-analyzed The Gasch (Gasch, Spellman et al. 2000) data was mined for changes in gene expression for genes identified in the both leucine and glutamate biosynthesis pathways. The genes with a significant shift greater than 2 fold are presented in Figures 38, 39, and 40. 3 T 2.5 -Time (hours) 1 -OLEUl -A-LEU 2 -«-LEU4 -A-BAT 2 -*-BAT 1 \ Figure 38 : Gene expression for leucine synthesis genes (Gasch, Spellman et al. 2000) Change in gene expression (fold) as a function of culture time (hours) after media exchange. 112 Time (hours) - O - G D H l - O G D H 3 - A - G L T 1 -A -GLN1 H B - G L N 3 - • - A D E 4 - fr-CPA 1 - • - C P A 2 - * - G S H 1 Figure 39 : Gene expression for glutamate synthesis genes (Gasch, Spellman et al. 2000) Change in gene expression (fold) as a function of culture time (hours) after media exchange. -3.5 -i > 1 , — • , , , 0 1 2 3 4 5 6 Time (hours) -H-GUA 1 - « - G L N 4 - © - G C N 3 - B - G C D 1 Figure 40 : Gene expression for glutamate synthesis genes (Gasch, Spellman et al. 2000) Change in gene expression (fold) as a function of culture time (hours) after media exchange. 113 The genes involved in Leucine biosynthesis show a notable up-regulation in gene expression (Figure 38). The LEU1 and LEU2 genes show a quick increase in expression observed after 0.5 hours. The BAT1 and BAT2 genes are also initially up-regulated. The LEU2 gene results show the most significant change in expression with a 5-fold increase (log base 2 value = 2.3) observed 1 hour after changing the media. After 6 hours, the genes involved in Leucine synthesis are still up-regulated almost 2-fold (log base 2 value > 1) or greater with the exception of BAT2. The LEU genes examined here were also up-regulated under conditions of complete nitrogen starvation (Gasch, Spellman et al. 2000). These genes are not included in the ESR genes and showed no differential expression for all other stress conditions with exception of progression into stationary phase. The results presented in Figure 39 show the expression profiles for genes that are involved in glutamate biosynthesis and are up-regulated at least 2-fold. The most significant up-regulation of 13-fold was observed for the GDH3 gene. On the other hand the change in gene expression for the GUA1 gene is down regulated to 0.125 of the control (log base 2 value = -3) (Figure 40). Nitrogen depletion also affected glutamate biosynthesis genes resulting in significant changes in gene expression. Changes in gene expression profiles for glutamate biosynthesis genes were not unique to amino acid starvation. Complete nitrogen starvation as well as several other stresses including hydrogen peroxide, temperature shocks, diothiothreitol and progression into stationary phase also resulted in expression changes to glutamate biosynthesis genes. 3.5.3 Natarajan, Meyer et al. Results Re-analyzed The Natarajan, Meyer et al. data was examined for changes in gene expression. For our analysis of the data, genes associated with the specific imposed limitations were examined. The expression data for genes involved in leucine and histidine synthesis are presented in Tables 29 and 30. 114 Table 29 : Histidine biosynthesis genes (Natarajan, Meyer et al. 2001) Gene Mild HIS, L E U Limitation 10 mM 3AT 100 Mm 3AT HISl 1.0 2.1 2.4 HIS2 0.5 1.8 2.2 HIS3, HIS10 1.2 3.4 3.2 HIS4 1.6 3.0 3.7 HIS5 1.5 3.8 4.3 HIS6 0.3 0.0 -0.2 HIS7 1.2 2.7 3.3 Table 30 : Leucine biosynthesis genes (Natarajan, Meyer et al. 2001) Gene Mild HIS, L E U Limitation 10 mM 3AT 100 Mm 3AT L E U l 0.6 -0.2 -1.7 LEU2 -0.3 0.4 -1.3 LEU3 0.1 0.8 1.4 LEU4 1.3 0.5 1.4 Table 29 reveals that the genes involved in histidine synthesis are up regulated for all 3 conditions. The most significant change in gene expression was observed under treatment with lOOmM 3AT with expression changes up to 20-fold (log base 2 value = 4.3). Table 30 shows that the genes involved in Leucine synthesis are not as significantly affected for these experiments. LEU3 and LEU4 gene are mildly up regulated under lOOmM 3AT treatment but are much less affected compared to the histidine genes. 3.5.4 Amino Acid Limitation Experiments The results presented by Gasch (Gasch, Spellman et al. 200) and Natarajan (Natarajan, Meyer et al. 2001) have shown that under stress conditions there are genes which exhibit 115 a general response. However neither study explored whether specific genomic responses could be identified under the imposed stresses. To determine if a single amino acid limitation in the media could be distinguished from conditions where multiple amino acid limitations were imposed, small-scale experiments were performed. Shake flask experiments were performed using S. cerevisiae (strain YPH 239). The amino acid limitations examined were leucine and glutamate. The same gene set identified from the literature and presented in Table 29 and 30 was examined for the shake flask experiments. The growth profiles for S. cerevisiae in the leucine and glutamate limitation experiments are presented in the Figures 41 and 42. 4 6 8 10 Time (hrs p ost in o culati on) 12 14 16 "Leucine limited "Control Figure 41: Growth curves for leucine limitation experiments Cell density ( g D C W - L 1 ) as a function of culture time (hours). Arrow indicates time when medium was exchanged to impose Leucine limitation. 116 45 40 • 35 30 % 25 & 20 15 Q 10 5 0 Media Change 4 6 8 10 Time (hrspost inoculation) 12 14 16 "Glutamic acid fimited •Control Figure 42 : Growth curves for glutamic acid limitation experiments Cell density (gDCW-L1) as a function of culture time (hours). Arrow indicates time when medium was exchanged to impose Glutamate limitation. Although the medium was exchanged in the early log phase, the culture approached the stationary growth phase 6 hours after the media exchange. The samples collected for microarray analysis were taken at 0.5, 1, and 6 hours after the media was exchanged. Again the glutamate and leucine synthesis pathways were used to identify genes of interest for our analysis. The results presented in Figures 43 and 44 show genes involved in glutamate synthesis and leucine synthesis. 117 1.50 -2.00 Time (h) -OLEUl -B-LEU2 -&-LEU4 -A-BAT1 Figure 43 : Gene expression for leucine limitation experiments Change in gene expression (fold) as a function of culture time (hours) after media exchange. .2.5 J Time (h) | - » - G L N I - Q - G L N 3 - * ~ G L N 4 - 6 - G C N 4 - • - H A P 4 - O - G A D l ~| Figure 44 : Gene expression for glutamic acid limitation experiments Change in gene expression (fold) as a function of culture time (hours) after media exchange. 118 The changes in gene expression observed under leucine starvation indicate that the leucine genes are up regulated as much as two 2-fold (log base 2 value =1). Although the initially all genes were down-regulated, the final response for the leucine experiment exhibited an overall increase in expression for the genes presented. For the glutamate starvation experiment the maximum increase in expression observed was 2.6-fold (log base 2 value = 1.4). Some genes showed an initial down-regulation to less than 0.25 (log base 2 value < -2.1) compared to the control. 3.5.5 Microarray Discussion 3.5.5.1 Gasch, Spellman et al. Discussion Examining the results from the Gasch (Gasch, Spellman et al. 2000) study we expected to find up-regulation of genes involved in both glutamate synthesis and leucine synthesis. Figures 38 and 39 show that genes involved in these pathways were significantly up regulated up to 5-fold for leucine genes and 13-fold for glutamate genes. Glutamate plays an essential role in nitrogen metabolism (ter Schure, van Riel et al. 2000). Genes such as GDH1, GDH3, GLN1, and GLT1, which are directly involved in glutamate synthesis, showed significant increases in expression that were sustained for over 6 hours. Genes including GUA1 and GLN4, which are involved in the utilization of glutamate, showed an expected decrease in gene expression. When comparing these genes examined for amino acid starvation with other stress conditions in the same study, no obvious trends were observed. Leucine synthesis genes under amino acid starvation also exhibited a general increase in expression except for BAT2. For all other stress conditions examined no significant changes in gene expression were observed for the LEU genes except for progression into stationary phase. These findings show predictable changes in gene expression with respect to the amino acid synthesis genes examined. These studies by Gasch, Spellman et al. looked at the effects of limiting amino acids in fermentation media. The control for these experiments contained a complex mixture of amino acids, which was then exposed to a media environment void of amino acids. It is unclear whether the changes in gene expression observed for the genes associated with 119 glutamate and leucine presented were due to the absence of these compounds or to a general amino acid starvation response by limiting all the amino acids. To determine whether one specific limitation could be detected by monitoring gene expression data, it was necessary to study individual limitations. 3.5.5.2 Natarajan, Meyer et al. Discussion The Natarajan, Meyer et al. study examined the affects of histidine starvation via treatment with 3AT. The data showed a significant up-regulation of all histidine synthesis genes for both lOmM 3AT and lOOmM 3AT experiments (Table 29). This result shows that inhibition of histidine synthesis by 3AT results in an increase in expression of the entire histidine synthesis pathway. This response provides evidence suggesting that a single limitation may be detected by analysis of gene expression data. Although these results are promising, the limitation imposed artificially by using 3AT does not necessarily mimic histidine depletion in the media. Natarajan, Meyer et al. also conducted an experiment in which leucine and histidine limitations were imposed by reducing the amino acids concentration in the media by half. In this scenario the gene expression changes showed a consistent up-regulation for all histidine and leucine genes (Table 29, 30). These results suggest that up-regulation of the histidine pathway gives an indication of leucine and histidine limitation in the media. However, the gene expression changes observed in this experiment were not as significant resulting in a maximum change of 3.0-fold. This lesser affect may be explained by the mild limitation imposed compared to the more significant expression changes observed for the complete starvation experiment. Although the Natarajan, Meyer et al. study examined effects of a histidine starvation, this starvation was imposed by adding 3AT, a competitive inhibitor of His3p. This method of imposing a specific amino acids limitation does not necessarily mimic histidine depletion in the media. To determine whether one specific limitation could be detected by monitoring gene expression data, it was necessary to study the effect of limiting specific media components such as leucine instead of using an inhibitor like 3AT to impose the limitation. 120 3.5.5.3 Leucine Limitation Experiments The gene expression profiles for the leucine amino limitation experiments reveal a decrease up to 3.5-fold in the expression of the genes associated with leucine metabolism. Although this initial down-regulation is observed, the 1 hour and 6 hour time points indicate an increase in gene expression ranging from approximately 1.5 to greater than 2-fold. These results are comparable to the mild leucine limitation experiments performed by Natarajan, Meyer et al. It has been shown that the production of leucine is regulated by the extracellular leucine concentration at the transcriptional level (Andreadis, Hsu et al. 1984; Hsu and Schimmel 1984). There is only one pathway to synthesize leucine in S. cerevisiae, which involves the genes LEU1, LEU2, LEU4, BAT1 and BAT2 metabolism (Figure 4, Chapter 2.5.5 Biosynthetic Pathways). The activation of the LEU genes is dependant on the Leu3p transcriptional activator protein which binds upstream of the LEU genes and contains both a DNA binding and transcriptional activation domain (Friden and Schimmel 1987). The expression of Leu3p is affected by Gcn4p. The Gcn4p complex is expressed under conditions of general amino acid starvation (Zhou, Brisco et al. 1987). Increasing Gcn4p results in Leu3p synthesis and leucine biosynthesis but may also increase the expression of genes involved in other amino acid synthesis pathways. Thus up-regulation of the LEU genes may occur under conditions other than leucine starvation. A second mechanism which has been reported indicates that transcriptional control of Leu3p is also mediated by ct-isopropylmalate (a-IPM), an intermediate of the leucine biosynthesis pathway, which acts by feedback inhibition (Figure 4, Chapter 2.5.5 Biosynthetic Pathways) (Reece 2000). This response pathway is specific to leucine and does not impact other amino acid synthesis genes. These two different regulation pathways indicate that an increase in expression of leucine biosynthesis genes can occur because of a specific response to a leucine limitation that is initiated by a decrease in a-IPM, or a general response to leucine or any combination of amino acid limitations, which induces gene expression by the Gcn4p protein. The media limitation imposed on the culture resulted in an overall up-regulation in gene expression for LEU genes. 121 Although the 1 hour and 6 hour time points support the observed increase in gene expression, the initial decrease in expression observed for the 0.5-hour time point was unexpected. This observation may be due to environmental effects unrelated to the amino acid limitation imposed. The washing and medium exchange procedure exposes the cells to a different culture environment for a short period of time of less than 10-15 minutes. Yeast transcription turnover rates can range from 3 to 90 minutes and could therefore impact the observed expression profiles (Wang, Liu et al. 2002). To further validate the results of this study, the effect of the washing procedure on gene expression should be examined. Although the consistent change in expression of all LEU genes indicates an up-regulation of leucine synthesis, the change is approximately 2-fold or less. Typically an expression change must be at least 2-fold to be considered significant (Gasch, Spellman et al. 2000). Consideration should also be given to gene expression changes observed for biosynthesis of other amino acids. The Gcn4p protein has been shown to influence transcription of up to 35 genes encoding amino acid biosynthetic enzymes (Natarajan, Meyer et al. 2001). Many of the genes involved in glutamate synthesis were differentially expressed under the leucine limitation up to a maximum of 4-fold. The genes involved in glutamate metabolism are directly linked to other amino acid synthesis pathways (Figure 5, Chapter 2.5.5 Biosynthetic Pathways), which might explain the observed changes. The genes involved in leucine synthesis were unaffected by glutamate starvation and therefore leucine limitations may be detectable and distinguishable from other amino acid limitations. This is further supported by the Natarajan, Meyer et al. results showing consistent distinguishable up-regulation of LEU genes under leucine starvation. Due to the mild changes in gene expression observed it may be more prudent to use sets of genes when trying to identify a specific media limitation. To identify the leucine limitation imposed in our experiments, the gene set LEU1, LEU2 and LEU4 may be used as an indicator. 122 3.5.5.4 Glutamate Limitation Experiments Results for the glutamate experiments reveal a more ambiguous conclusion. Only the GLN1, GLN3, and GLN4 genes were observed to have a significant change in gene expression under glutamate starvation up to 2.5-fold. The GLN1 gene showed an initial decrease in expression greater than 4-fold which was unexpected. There are other genes involved in glutamate synthesis discussed in section 3.5.2 that were not affected in this experiment. Examining the glutamate biosynthesis pathway, it is evident that glutamate can be synthesized via multiple pathways metabolism (see Literature Review Figure 5). The relatively mild changes in gene expression observed for the genes involved in glutamate synthesis may be attributed this complex network. Under a glutamate limitation the cells can still obtain glutamate via the TCA cycle, arginine, proline, histidine, and nitrogen metabolism (see Literature Review Figure 5). These potential pathways suggest that a single amino acid limitation such as glutamate may not significantly impact the genes involved in glutamate synthesis. While searching for unique genes that indicate a glutamate limitation may not be possible, examining the response of a group of genes to identify limitations may provide further insight. 3.5.5.5 Microarray Discussion Summary The recent literature and experiments performed on single amino acid starvation have shown that DNA microarrays can be used to monitor physiological responses to different nutrient limitations. The ability to map these changes at the transcriptional level may aid in the development of DNA microarray based techniques as a diagnostic tool for cell culture and fermentation. The Gasch and Natarajan studies showed groups of genes with significant up-regulation responding predictably to known environmental stressed. The amino acid starvation studies with leucine also showed similar results indicating the potential to use microarrays for this type of analysis. One very recent study also reported predictable changes in gene expression for S. cerevisiae cells exposed to specific carbon, nitrogen, phosphorus and sulfur medium limitations (Boer et al. 2002). Experiments with glutamate starvation however, have revealed that the complexity of cellular networks 123 may complicate the interpretation of gene expression data as it is applied to detection of media limitations, starvation and other environmental stresses. 3.5.5.6 Considerations for Future Microarray Experiments There are several considerations that will improve future experiments. The media should be exchanged earlier at 4-6 hours post induction. In these experiments the 6 hour time point was taken when the culture was in the stationary phase. Cells in the stationary phase will likely display changes in gene expression that are not indicative of the genetic response resulting from the specific amino acid limitation. Using a lower inoculum concentration could also solve this problem. In future experiments the media should be exchanged at a time point that will ensure all samples points lay within the exponential growth phase. Secondly, the media exchange procedure takes approximately 10-15 minutes and during this time the cells can be exposed to conditions that may cause changes in gene expression profiles. To eliminate the effect of media exchange a control culture should be run in parallel. The control culture would undergo the same washing and media exchange procedure but would be supplied with fresh media without limiting any components. Finally, repeat experiments should also be performed to determine the statistical significance of the results. 124 Conclusions The primary objective of this project was to investigate process factors affecting productivity and glycosylation of recombinant mutant cystatin C expressed in the methylotrophic yeast Pichia pastoris. The factors examined in these experiments were: 1) temperature, 2) pH for expression, 3) methanol concentration, and 4) nitrogen source. Initial stability studies were performed and results showed that storage of recombinant cystatin C at -20°C for 3 weeks has no significant effect on activity. These storage conditions were used for the bioreactor optimization experiments. Feed strategy experiments with a methanol analyzer identified that maintaining a 0.2% methanol concentration during induction yielded a greater cystatin C specific productivity than 1% methanol or mixed glycerokmethanol feed (glycerol feed of 11.5 g-L'-h"1, methanol concentration of 0.05%). These experiments were also used to determine methanol feed rates for subsequent experiments that were done without a methanol analyzer. The feed strategy experiments only examined a limited range of methanol concentrations since the methanol analyzer was only available for a few experiments. Screening experiments were performed to determine the effects of nitrogen source, pH and temperature, since these were identified in the literature review as important parameters affecting glycosylation. Maximum cystatin C productivity was 2.9 nmol-gDCW1-h"1 in shake flask and 13.5 nmol-gDCW"1-hr"1 in the 2-litre bioreactor. The conditions for maximum productivity were identified as 24°C and a high peptone (9.2 g-L"1) and no ammonium (Og-L"1). ThepH was determined to be best at 6. Substitution of peptone for ammonium sulfate as a nitrogen source resulted in more visible bands in the SG and DG cystatin C regions of the SDS PAGE gels. Cursory examination of these gels suggested that the intensities of the glycosylated bands were stronger relative to the unglycosylated bands for experiments with peptone compared to experiments without peptone. No measurable trends were found with respect to 125 temperature and pH. However, low titers were obtained in the shake flask experiments necessitating concentration of samples more than 5 fold before running gels. This resulted in interference from salts and other contaminants. Since nitrogen source was found to be a significant factor affecting glycosylation bioreactor experiments were performed to further verify this and to include another potential nitrogen source; a defined amino acid mix. Substituting peptone for ammonium hydroxide as a nitrogen source resulted in an increase in cystatin C productivity greater than 2-fold. Ammonium hydroxide concentration had a negative influence on productivity at high levels of peptone (20 g-L"1) and amino acids (20 g-L"1). The addition of amino acids to standard fermentation conditions where only ammonium hydroxide is used, resulted in improved productivity with the statistical model predicting an increase from 0.16 to 4.1 nmol-gDCW"1 h"1. However, at high peptone concentrations amino acids had a slight negative impact reducing productivity from 12.9 - 12.0 nmol-gDCW1 h"1. The optimum conditions were obtained with high peptone (20g-L"!), high amino acids (20g-L"') and no ammonium hydroxide (Og-L"1) resulting in 13.6 nmol-gDCW"1 h"1. A 5-fold increase in the amount of glycosylated cystatin C produced (from 6% to 30%) resulted from the addition of peptone. Ammonium concentration had a negative effect on the extent of glycosylation reducing it from 30% (Og-L"1 NH 4 HS0 4 ) to 12% (30g-L"' NH4HSO4) under conditions of high peptone (20g-L1). The affect of ammonium hydroxide was greatest under conditions of high peptone. Adding the defined amino acid mix had a positive effect on the extent of glycosylation, however, it was not as great as that of peptone. The results suggest that the presence of peptides, versus free amino acids in the medium are responsible for the increase in glycosylated protein. Over the conditions investigated, a maximum glycosylation of approximately 30 % was measured with the addition of peptone (20g-L_1) and amino acids (20g-L_1) in the absence of ammonium hydroxide, with process conditions: 0.2% methanol during induction, reactor temperature of 28°C and pH of 6. The secondary objective of this project was to examine an alternative approach to process development and optimization, using DNA microarray technology. To achieve this goal 126 recent data from the literature was analyzed and shake flask experiments were designed to examine media limitations in S. cerevisiae fermentation processes. Two limitations were examined including glutamate and leucine. Results from the literature and the leucine limitation experiments indicate that gene expression profiles may be used to identify media limitations by examining groups of genes that show a consistent trend and that are only expressed under leucine starvation. The recent literature including studies by Gasch, Spellman et al. revealed that leucine genes were up-regulated as much as 5 fold under conditions of amino acid starvation. Our own experiments showed that a single nutrient limitation such as leucine resulted change in expression up to 2-fold. To identify a leucine limitation imposed, the gene set LEU1, LEU2 and LEU4 can be used as an indicator. Results for the glutamic acid starvation experiments were more ambiguous. There are many potential pathways in glutamate synthesis and a single amino acid limitation such as glutamic acid did not significantly impact the genes involved in glutamate biosynthesis. Results from the Gasch, Spellman et al. study indicated that genes involved in glutamate synthesis were up-regulated as much as 13-fold (GDH3) under complete amino acid starvation. However, the shake flask experiments, where a single glutamic acid limitation was imposed resulted in more subtle changes in gene expression. A maximum change of 2.5 fold was observed and many of the genes were down-regulated resulting in no clear trends. Only a few genes including GLN1, GLN3, and GLN4 were observed to have significant changes in gene expression. 127 Recommendations The methanol concentration used in these experiments was 0.2%, however, concentrations between 0.2% and 1% should also be examined to further optimize the process. Other researchers have found optimum methanol concentrations to be 0.3% (Guarna, Lesnicki et al. 1997) and 0.25% (Minning, Serrano et al. 2001). Future experimentation should also either utilize a methanol sensor or perform off-line methanol analysis to ensure methanol concentrations are constant. Limiting or excess methanol concentrations may impact the growth, productivity and glycosylation results. It would be advantageous to perform additional experiments to obtain center points for the existing experimental design. This could be used to elucidate any quadratic affects of the variables. The amino acid mixture used was developed for S. cerevisiae not for the P. pastoris system. Additional experiments are necessary to determine the amino acids composition for maximal protein productivity and glycosylation. It may be advantageous to select one amino acid such as glutamine instead of a mixture to simplify the media requirements. Although it has been shown that nitrogen source had a significant impact of productivity there are still other medium components, including the carbon source and trace elements that should be examined to further improve the expression of recombinant cystatin C in P. pastoris. The literature review identified carbon source availability as a parameter that has an impact on productivity and glycosylation. Under limiting methanol concentrations the lack of available sugars may impact the synthesis of key components involved in glycosylation including the lipid-dolichol and mannose sugar groups. There are several considerations that will improve future microarray experiments. The media should be exchanged earlier at 4-6 hours post induction. In these experiments the 6 hour time point was taken when the culture was in the stationary phase. Cells in the stationary phase will likely display changes in gene expression that are not indicative of the genetic response resulting from the specific amino acid limitation. Using a lower inoculum concentration could also solve this problem. In future experiments the media 128 \ should be exchanged at a time point that will ensure all samples points lay within the exponential growth phase. Secondly, the media exchange procedure takes approximately 10-15 minutes and during this time the cells can be exposed to conditions that may cause changes in gene expression profdes. To eliminate the affect of media exchange a control culture should be run in parallel. The control culture would undergo the same washing and media exchange procedure but would be supplied with fresh media without limiting any components. Repeat experiments should also be performed to determine the statistical significance of the results. Finally, experimentation with other amino acid supplements should be investigated to determine gene expression data can be used to detect and distinguish other amino acids limitations in addition to those already studied. 129 Abbreviations A0X1 a lcoho l o x i d a s e g e n e ade a d e n i n e A N O V A ana l ys i s of va r i ance AOX1 a lcoho l o x i d a s e protein AOX2 a lcoho l ox i dase gene AOX2 a lcoho l o x i d a s e protein A s n aspa rag ine A s p aspar t i c ac id B A L B / c 3T3 mur ine ce l l strain B A P N A Na -benzoy l -DL -a rg i n i ne p-nitroani l ide hydroch lor ide B H K baby hamste r k idney B M G H buf fered min ima l g lycero l m e d i a with hist id ine B M M H buf fered m in ima l methano l m e d i a with hist id ine C C D centra l compos i t e des ign c D N A comp l imen ta ry D N A C H O C h i n e s e hamste r ova ry C M P cyt id ine monophospha te cy3 water so lub le cyan in d y e s cy5 water so lub le cyan in d y e s D a Da l ton D C W dry ce l l weight (g/L) D E P C diethyl pyrocarbonate D N A deoxy r i bonuc le i c ac id D O d i s s o l v e d oxygen d o l - P - P do l i chy ld iphosphate E D T A e thy lene d iam ine tetra acet ic ac id E n d o H E n d o - b - N - a c e t y l g l u c o s a m i n i d a s e H E P O erythropoiet in E R e n d o p l a s m i c re te icu lum E S R env i ronmenta l s t ress response 130 EST expressed sequence tag FDA Food and Drug Administration F S H follicle stimulating hormone Fuc fucose G1 phase of cell cycle Gal galactose GalNAc N-acetylgalactosamine G - C S F granulocyte colony-stimulating factor G D P guanosine diphosphate GlcNAc N-acetylglucosamine Glu glucose Gly glycerol Go phase of cell cycle Go/G1 phase of cell cycle H E P - G 2 human hepatoblastoma His- auxotrophic for histidine HIS4 histidinol dehydrogenase gene HIV human immunodeficiency virus H P L C high performance liquid chromatography IFN interferon igG immunoglobulin G INF interferon kDa kilodalton lys lysine Man mannose MeOH methanol mRNA messenger RNA MS mass spectrometry Mut- methanol utilization negative Mut+ methanol utilization positive Muts methanol utilization slow NeuAc sialic acid NeuGc N-glycolylneuraminic acid 131 NIL 8 hamster cells NMR nuclear magnetic resonance N S O murine myeloma cell line O R F open reading frame P A G E polyacrylamide gel electrophoresis PNGaseF peptide-N-glycosidase F RNA ribonucleic acid S D S sodium dodecylsulfate S E A P secreted alkaline phosphatase Ser serine S R A F P sea raven type II antifreeze protein S S C tricarboxylic acid T C A trichloroacetic acid T E S Tris, EDTA, sorbitol Thr threonine tPA tissue-type plasminogen activator ura uracil W C W wet cell weight Xaa any amino acid YNB yeast nitrogen base Y P D yeast extract, peptone, dextrose Y P H Yeast strains DG double glycosylated UG unglycosylated S G single glycosylated 3AT 3-aminotriazole IPM isopropylmalate 132 4 References Abrahamson, M . (1994). "Cystatins." Methods Enzymol 244: 685-700. Abrahamson, M. , A. J. Barrett, G. Salvesen and A. Grubb (1986). "Isolation of six cysteine proteinase inhibitors from human urine. Their physicochemical and enzyme kinetic properties and concentrations in biological fluids." J Biol Chem 261(24): 11282-9. Abrahamson, M. , H. Dalboge, I. Olafsson, S. Carlsen and A. Grubb (1988). "Efficient production of native, biologically active human cystatin C by Escherichia coli." FEBS Lett 236(1): 14-8. Abrahamson, M . and A. Grubb (1994). "Increased body temperature accelerates aggregation of the Leu-68—>Gln mutant cystatin C, the amyloid-forming protein in hereditary cystatin C amyloid angiopathy." Proc Natl Acad Sci U S A 91(4): 1416-20. Alexeeva, Y. V., E. P. Ivanova, I. Y. Bakunina, T. N . Zvaygintseva and V. V. Mikhailov (2002). "Optimization of glycosidases production by Pseudoalteromonas issachenkonii K M M 3549(T)." Lett Appl Microbiol 35(4): 343-6. Andersen Dana, C. G. C. F. (1995). "The effect of ammonia on the O-linked glycosylation of granulocyte." Biotechnology & Bioengineering 47(1): 96-105. Andersen, D. C , T. Bridges, M . Gawlitzek and C. Hoy (2000). "Multiple cell culture factors can affect the glycosylation of Asn-184 in CHO-produced tissue-type plasminogen activator." Biotechnol Bioeng 70(1): 25-31. Andreadis, A., Y. P. Hsu, M . Hermodson, G. Kohlhaw and P. Schimmel (1984). "Yeast LEU2. Repression of mRNA levels by leucine and primary. structure of the gene product." J Biol Chem 259(13): 8059-62. Ausubel, F. M. , R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith and K. Struhl (1999). Short Protocols in Molecular Biology, 4th Edition. Bahr-Davidson, J. B. (1995). "Factors affecting glycosylation site occupancy of Asn-184 of tissue-type plasminogen activator produced in Chinese hamster ovary cells." PhD dissertation, Stanford University. Barr, K. A., S. A. Hopkins and K. Sreekrisshnna (1993). "Protocol for efficient secretion of HSA developed from Pishiapastoris.'" Pharm Eng 12: 48-51. Barret, A. J. (1981). "Cystatin, the egg white inhibitor of cysteine proteinases." Methods Enzymol 80: 771-778. Barrett, A. J. (1984). "The place of human y-trace (cystatin C) amongst the cysteine proteinase inhibitors." Biochem Biophys Res Commun 120: 931-936. 133 Berti, P. J., I. Ekiel, P. Lindahl, M . Abrahamson and A. C. Storer (1997). "Affinity purification and elimination of methionine oxidation in recombinant human cystatin C." Protein ExprPurif 11(1): 111-8. Bjoern, S., D. C. Foster, L. Thim, F. C. Wiberg, M . Christensen, Y. Komiyama, A. H. Pedersen and W. Kisiel (1991). "Human plasma and recombinant factor VII. Characterization of O-glycosylations at serine residues 52 and 60 and effects of site-directed mutagenesis of serine 52 to alanine." J Biol Chem 266(17): 11051-7. Bjorck, L., P. Akesson, M . Bohus, J. Trojnar, M . Abrahamson, I. Olafsson and A. Grubb (1989). "Bacterial growth blocked by a synthetic peptide based on the structure of a human proteinase inhibitor." Nature 337(6205): 385-6. Bjorck, L., A. Grubb and L. Kjellen (1990). "Cystatin C, a human proteinase inhibitor, blocks replication of herpes simplex virus." J Virol 64(2): 941-3. Boer, V. M. , de Winde, J. H., Pronk, J. T., Piper, M . D. W. (2003). "The Genome-wide Transcriptional Responses of Saccharomyces cerevisiae Grown in Aerobic Chemostat Cultures Limited for Carbon, Nitrogen, Phosphorus, or Sulfur." J Biol Chem 278 (5): 3265-74. Borys, M . C , D. I. Linzer and E. T. Papoutsakis (1993). "Culture pH affects expression rates and glycosylation of recombinant mouse placental lactogen proteins by Chinese hamster ovary (CHO) cells." Biotechnology (N Y) 11(6): 720-4. Borys, M . C , Linzer, D. I. H., Papoutsakis, E. T. (1994). "Ammonia affects the glycosylation patterns of recombinant mouse placental lactogen-I by Chinese hamster ovary cells in a pH-dependent manner." Biotechnology & Bioengineering 43: 505-514. Box, G. E. P. and J. S. Hunter (1957). Multifactor Experimental Designs for Exploring Response Surfaces. Bretthauer, R. K. and F. J. Castellino (1999). "Glycosylation of Pichia pastoris-derived proteins [see comments]." Biotechnol Appl Biochem 30(Pt 3): 193-200. Brierley, R. A., C. Bussineau, R. Kosson, A. Melton and R. S. Siegel (1990). "Fermentation development of recombinant Pichia pastoris expressing the heterologous gene: bovine lysozyme." Ann N Y Acad Sci 589: 350-62. Brown, P. O. and D. Botstein (1999). "Exploring the new world of the genome with DNA microarrays." Nat Genet 21(1 Suppl): 33-7. Butler, M. , Jenkins (1989). "Nutritional aspects of the growth of animal cells in culture." J Biotechnol 12: 97-110. Cartwright, T. (1992). Animal Cell Biotechnology. London, Academic Press. 134 Cereghino, G. P. and J. M . Cregg (1999). "Applications of yeast in biotechnology: protein production and genetic analysis." Curr Opin Biotechnol 10(5): 422-7. Cereghino, J. L. and J. M . Cregg (2000). "Heterologous protein expression in the methylotrophic yeast Pichia pastoris." FEMS Microbiol Rev 24(1): 45-66. Chen, D. C , S. Y. Chen, M. F. Gee, J. T. Pan and K. T. T. (1999). "A variant of Saccharomyces cerevisiae pep4 strain with improved oligotrophic proliferation, cell survival and heterologous secretion of a-amylase." Appl Microbiol Biotechnol 51: 185-192. Chen, D. C , B. C. Yang, W. P. Chen and K. T. T. (1995). "Abnormal growth induced by expression of HBsAg in the secretion pathway of S. cerevisiae pep4 mutants." Curr Genet 27: 201-206. Choi, B. K. and R. Jimenez-Flores (2001). "Expression and purification of glycosylated bovine beta-casein (L70S/P71S) in Pichia pastoris." J Agric Food Chem 49(4): 1761-6. Chotigeat, W., W. Chayanunnukul and A. Phongdara (2000). "Expression of a mammalian alpha 2,6-sialyltransferase gene in Pichia pastoris." J Biotechnol 81(1): 55-61. Clare, J. J., F. B. Rayment, S. P. Ballantine, K. Sreekrishna and M . A. Romanos (1991). "High-level expression of tetanus toxin fragment C in Pichia pastoris strains containing multiple tandem integrations of the gene." Biotechnology (N Y) 9(5): 455-60. Cleveland, W. S. (1979). "Robust Locally Weighted Regression and Smoothing Scatterplots" Journal of the American Statistical Association 74(368): 829-36. Cooley, M . and N . C. Mishra (2000). "Genetic analysis of the in vivo role of DNA polymerases in Saccharomyces cerevisiae." Curr Genet 38(5): 256-63. Couderc, R. and J. Barratti (1980). "Oxidation of methanol by the yeast Pichia pastoris: purification and properties of alcohol oxidase." Agric Biol Chem 44: 2279-2289. Cox, J. L., P. S. Sexton, T. J. Green and N. A. Darmani (1999). "Inhibition of B16 melanoma metastasis by overexpression of the cysteine proteinase inhibitor cystatin C." Melanoma Res 9(4): 369-74. Cumming, D. A. (1992). "Physiological relevance of protein glycosylation." Dev Biol Stand 76(3): 83-94. Curling, E. M. , P. M . Hayter, A. J. Baines, A. T. Bull, K. Gull, P. G. Strange and N. Jenkins (1990). "Recombinant human interferon-gamma. Differences in glycosylation and proteolytic processing lead to heterogeneity in batch culture." Biochem J 272(2): 333-7. 135 Dalboge, H., E. B. Jensen, H. Tottrup, A. Grubb, M . Abrahamson, I. Olafsson and S. Carlsen (1989). "High-level expression of active human cystatin C in Escherichia coli." Gene 79(2): 325-32. d'Anjou, M . C. and A. J. Daugulis (2001). "A rational approach to improving productivity in recombinant Pichia pastoris fermentation." Biotechnol Bioeng 72(1): 1-11. Davidson, S. K. and L. A. Hunt (1985). "Sindbis virus glycoproteins are abnormally glycosylated in Chinese hamster ovary cells deprived of glucose." J Gen Virol 66(Pt 7)(12): 1457-68. Davis, T. R. and H. A. Wood (1995). "Intrinsic glycosylation potentials of insect cell cultures and insect larvae." In Vitro Cell Dev Biol Anim 31(9): 659-63. de Vries, T. and D. H. van den Eijnden (1992). "Occurrence and specificities of alpha 3-fucosyltransferases." Histochem J 24(11): 761-70. Dean, R. T., W. Jessup and C. R. Roberts (1984). "Effects of exogenous amines on mammalian cells, with particular reference to membrane flow." Biochem J 217(1): 27-40. DeRisi, J. L., V. R. Iyer and P. O. Brown (1997). "Exploring the metabolic and genetic control of gene expression on a genomic scale." Science 278(5338): 680-6. Digan, M . E., S. V. Lair, R. A. Brierley, R. S. Siegel, M . E. Williams, S. B. Ellis, P. A. Kellaris, S. A. Provow, W. S. Craig, G. Velicelebi, M . M . Harpold and G. P. Thill (1989). "Continuous production of a novel lysozyme via secretion from the yeast, Pichia pastoris." Biochem Biophys Res Commun 232: 656-662. Donaldson, M. , H. A. Wood, P. C. Kulakosky and M. L. Shuler (1999). "Glycosylation of a recombinant protein in the Tn5Bl-4 insect cell line: influence of ammonia, time of harvest, temperature, and dissolved oxygen." Biotechnol Bioeng 63(3): 255-62. Donaldson, M. , H. A. Wood, P. C. Kulakosky and M . L. Shuler (1999). "Use of mannosamine for inducing the addition of outer arm N-acetylglucosamine onto N-linked oligosaccharides of recombinant proteins in insect cells." Biotechnol Prog 15(2): 168-73. Ekiel, I. and M . Abrahamson (1996). "Folding-related dimerization of human cystatin C." J Biol Chem 271(3): 1314-21. Files, D. A. (2000). "Expression of Human Cystatin C by the Methylotrphic Yeast Pichia pastoris.'" Masters thesis, University of British Columbia.: 1-157. Franeck, F., Eckschlager, T.,Katinger, H. (2003). "Enhancement of Monoclonal Antibody Production by Lysine-Containing Peptides." Biotechnol Prog 19(1): 169-74. 136 Franeck, F., Hohenwarter, O., Katinger, H. (2000). "Plant Protein Hydrolysates: Preparation of Defined Peptide Fractions Promoting Growth and Production in Animal Cells Cultures." Biotechnol Prog 16(5): 688-92. Freshney, R. I. (1987). Culture of animal cells: a manual of basic technique. New York, Liss. Friden, P. and P. Schimmel (1987). "LEU3 of Saccharomyces cerevisiae encodes a factor for control of RNA levels of a group of leucine-specific genes." Mol Cell Biol 7(8): 2708-17. Gasch, A. P., P. T. Spellman, C. M . Kao, O. Carmel-Harel, M . B. Eisen, G. Storz, D. Botstein and P. O. Brown (2000). "Genomic expression programs in the response of yeast cells to environmental changes." Mol Biol Cell 11(12): 4241-57. Gawlitzek, M. , H. S. Condradt and R. Wagner (1995). "Effect of different cell culture conditions on the polypeptide integrity of a recombinant model glycoprotein." Biotechnol Bioeng 46: 536-544. Gawlitzek, M. , T. Ryll, J. Lofgren and M . B. Sliwkowski (2000). "Ammonium alters N-glycan structures of recombinant TNFR-IgG: degradative versus biosynthetic mechanisms." Biotechnol Bioeng 68(6): 637-46. Gawlitzek, M. , U. Valley, M . Nimtz, R. Wagner and H. S. Conradt (1995). "Characterization of changes in the glycosylation pattern of recombinant proteins from BHK-21 cells due to different culture conditions." J Biotechnol 42(2): 117-31. Gellissen, G., C. P. Hollenberg and Z. A. Janowicz (1995). "Gene expression in methylotrophic yeasts." Bioprocess Technol 22: 195-239. Gershman, H. and P. W. Robbins (1981). "Transitory effects of glucose starvation on the synthesis of dolichol-linked oligosaccharides in mammalian cells." J Biol Chem 256(15): 7774-80. Glover, C. (2002). Microarray construction and probe preparation. Gomes, A. V. (2002). PROTEIN GEL ELECTROPHORESIS TIPS AND TROUBLESHOOTING GUIDE, http://members.tripod.com/~Aldrin/index-3.html. Goochee, C. F. (1992). "Bioprocess factors affecting glycoprotein oligosaccharide structure." Dev Biol Stand 76(6): 95-104. Goochee, C. F., M . J. Gramer, D. C. Andersen, J. B. Bahr and J. R. Rasmussen (1991). "The oligosaccharides of glycoproteins: bioprocess factors affecting oligosaccharide structure and their effect on glycoprotein properties." Biotechnology (N Y) 9(12): 1347-55. 137 Goochee, C. F. and T. Monica (1990). "Environmental effects on protein glycosylation." Biotechnology (N Y) 8(5): 421-7. Grammatikos, S. I., U. Valley, M . Nimtz, H. S. Conradt and R. Wagner (1998). "Intracellular UDP-N-acetylhexosamine pool affects N-glycan complexity: a mechanism of ammonium action on protein glycosylation." Biotechnol Prog 14(3): 410-9. Grubb, A. (1988). "Safer proteinase treatment of sciatica. A biochemical preview of chymopapain inhibitors." Acta Orthop Scand 59(1): 63-5. Guarna, M. M. , G. Lesnicki, B. M . Tam, J. Robinson, C. Z. Radziminski, H. D., A. Boraston, E. Jervis, R. T. A. MacGillivray, R. F. B. Turner and D. G. Kilburn (1997). "On-line Monitoring and Control of Methanol Concentration in Shake-Flask Cultures of Pichia pastoris." Biotechnology & Bioengineering 56(3): 279-286. Hahn, T. J. and C. F. Goochee (1992). "Growth-associated glycosylation of transferrin secreted by HepG2 cells." J Biol Chem 267(33): 23982-7. Harris, R. J., C. K. Leonard, A. W. Guzzetta and M . W. Spellman (1991). "Tissue plasminogen activator has an O-linked fucose attached to threonine-61 in the epidermal growth factor domain." Biochemistry 30(9): 2311-4. Hayter, P. M. , E. M . Curling, A. J. Baines, N. Jenkins, I. Salmon, P. G. Strange and A. T. Bull (1991). "Chinese hamster ovary cell growth and interferon production kinetics in stirred batch culture." Appl Microbiol Biotechnol 34(5): 559-64. Higgins, D. R. and J. M . Cregg (1998). Pichia Protocols. Hong, F., N. Q. Meinander and L. J. Jonsson (2002). "Fermentation strategies for improved heterologous expression of laccase in Pichia pastoris." Biotechnol Bioeng 79(4): 438-49. Hsu, Y. P. and P. Schimmel (1984). "Yeast LEU1. Repression of mRNA levels by leucine and relationship of 5'-noncoding region to that of LEU2." J Biol Chem 259(6): 3714-9. Inan, M. , V. Chiruvolu, K. M . Eskridge, G. P. Vlasuk, K. Dickerson, S. W. Brown and M. M. Meagher (1999). "Optimization of temperature-glycerol-pH conditions for a fed-batch fermentation process for recombinant hookworm (Ancylostoma caninum) anticoagulant peptide (AcAP-5) production by Pichia pastoris." Enzyme Microb Technol 24: 438-445. Invitrogen (1998). EasySelect Pichia Expression Kit. Invitrogen (1999). Pichia Fermentation Process Guidelines. Jan, D. C , Jones, S. J., Emery, A. N. , al-Rubeai, M . (1994)."Peptone, a low cost growth-promoting nutrient for intensive animal cell culture." Cytotechnology 16(1): 17-26. 138 Jenkins, N . (1995). "Monitoring and control of recombinant glycoprotein heterogeneity in animal cell cultures." Biochem Soc Trans 23(1): 171-5. Jenkins, N . and E. M . Curling (1994). "Glycosylation of recombinant proteins: problems and prospects." Enzyme Microb Technol 16(5): 354-64. Jenkins, N. , R. B. Parekh and D. C. James (1996). "Getting the glycosylation right: implications for the biotechnology industry." Nat Biotechnol 14(8): 975-81. Kalidas, C , L. Joshi and C. Batt (2001). "Characterization of glycosylated variants of beta-lactoglobulin expressed in Pichia pastoris." Protein Eng 14(3): 201-7. Kerry-Williams, S. M. , S. C. Gilbert, L. R. Evans and D. J. Ballance (1998). "Disruption of the Saccharomyces cerevisiae YAP3 gene reduces the proteolytic degradation of secreted recombinant human albumin." Yeast 14(2): 161-9. Kobata, A. (1992). "Structures and functions of the sugar chains of glycoproteins." Eur J Biochem 209(2): 483-501. Kretzmer, G. (2002). "Industrial processes with animal cells." Appl Microbiol Biotechnol 59(2-3): 135-42. Kukuruzinska, M . A. and K. Lennon (1994). "Growth-related coordinate regulation of the early N-glycosylation genes in yeast." Glycobiology 4(4): 437-43. Lehrman, M . A. (1991). "Biosynthesis of N-acetylglucosamine-P-P-dolichol, the committed step of asparagine-linked oligosaccharide assembly." Glycobiology 1(6): 553-62. Lenarcic, B., M . Krasovec, A. Ritonja, I. Olafsson and V. Turk (1991). "Inactivation of human cystatin C and kininogen by human cathepsin D." FEBS Lett 280(2): 211-5. Lesnicki, G. (2001). University of Britsh Columbia Pilot Plant Manager. Lifely, M . R., C. Hale, S. Boyce, M . J. Keen and J. Phillips (1995). "Glycosylation and biological activity of CAMPATH-1H expressed in different cell lines and grown under different culture conditions." Glycobiology 5(8): 813-22. Lin, F., R. M . Immormino, M . Shoham and M . E. Medof (2001). "Bulk production and functional analyses of mouse CD55's native and deglycosylated active domains." Arch Biochem Biophys 393(1): 67-72. Loewen, M . C , X. Liu, P. L. Davies and A. J. Daugulis (1997). "Biosynthetic production of type II fish antifreeze protein: fermentation by Pichia pastoris." Appl Microbiol Biotechnol 48(4): 480-6. Magasanik, B. and C. A. Kaiser (2002). "Nitrogen regulation in Saccharomyces cerevisiae." Gene 290(1-2): 1-18. 139 Mendoza-Vega, O., J. Sabatie and S. W. Brown (1994). "Industrial production of heterologous proteins by fed-batch cultures of the yeast Saccharomyces cerevisiae." FEMS Microbiol Rev 15(4): 369-410. Minning, S., A. Serrano, P. Ferrer, C. Sola, R. D. Schmid and F. Valero (2001). "Optimization of the high-level production of Rhizopus oryzae lipase in Pichia pastoris." J Biotechnol 86(1): 59-70. Montesino, R., R. Garcia, O. Quintero and J. A. Cremata (1998). "Variation in N-linked oligosaccharide structures on heterologous proteins secreted by the methylotrophic yeast Pichia pastoris." Protein Expr Purif 14(2): 197-207. Montgomery, D. C. (1997). Design and Analysis of Experiments (4th edition), John Wiley and Sons, Inc. Moore, A., J. Mercer, G. Dutina, C. J. Donahue, K. D. Bauer, J. P. Mather, T. Etcheverry and T. Ryll (1997). "Effects of temperature shift on cell cycle, apoptosis and nucleotide pools in CHO cell batch cultures." Cytotechnology 23: 47-54. Murray, D. W., S. J. B. Duff and P. H. Lanthier (1989). "Induction and stability of alcohol oxidase in the methylotrophic yeast Pichia pastoris." Appl Microbiol Biotechnol 32:95-100. Nakamura, S., M . Ogawa and S. Nakai (1998). "Effects of polymannosylation of recombinant cystatin C in yeast on its stability and activity." J Agric Food Chem 46: 2882-2887. Nakamura, S., M . Ogawa and S. Nakai (2000). Production and use of modified cystatins. International patent (Patent Cooperation Treaty). Nakao, H. N. A. K. T. N. T. I. M . (1990). "Modulation of N Acetylglucosaminyltransferase Iii Iv and V Activities and." Biochemical & Biophysical Research Communications 172(3): 1260-1266. Natarajan, K., M . R. Meyer, B. M. Jackson, D. Slade, C. Roberts, A. G. Hinnebusch and M. J. Marton (2001). "Transcriptional profiling shows that Gcn4p is a master regulator of gene expression during amino acid starvation in yeast." Mol Cell Biol 21(13): 4347-68. Nyberg, G. B., R. R. Balcarcel, B. D. Follstad, G. Stephanopoulos and D. I. C. Wang (1999). "Metabolism of Peptide Amino Acids by Chinese Hamster Ovary Cells Grown in a Complex Medium." Biotechnol Bioeng 62: 324-335. Nyberg, G. B., R. R. Balcarcel, B. D. Follstad, G. Stephanopoulos and D. I. C. Wang (1999). "Metabolic effects on recombinant interferon-y glycosylation in continuous culture of Chinese hamster ovary cells." Biotechnol Bioeng 62: 336-347. Oh, M . K. and J. C. Liao (2000). "DNA microarray detection of metabolic responses to protein overproduction in Escherichia coli." Metab Eng 2(3): 201-9. 140 Ohkuma, S., J. Chudzik and B. Poole (1981). "The effects of basic substances and acidic ionophores on the digestion of exogenous and endogenous proteins in mouse peritoneal macrophages." J Cell Biol 90: 656-664. Orlean, P. (1992). "Enzymes that recognize dolichols participate in three glycosylation pathways and are required for protein secretion." Biochem Cell Biol 70(6): 438-47. Patel, T. P., R. B. Parekh, B. J. Moellering and C. P. Prior (1992). "Different culture methods lead to differences in glycosylation of a murine IgG monoclonal antibody." Biochem J 285(Pt 3)(3): 839-45. Pilon, A. L., P. Yost, T. E. Chase, G. L. Lohnas and W. E. Bentley (1996). "High-level expression and efficient recovery of ubiquitin fusion proteins from Escherichia coli." Biotechnol Prog 12(3): 331-7. Pinkel, D. (2000). "Cancer cells, chemotherapy and gene clusters." Nat Genet 24(3): 208-9. Pos, O., W. van Dijk, N. Ladiges, C. Linthorst, M . Sala, D. van Tiel and W. Boers (1988). "Glycosylation of four acute-phase glycoproteins secreted by rat liver cells in vivo and in vitro. Effects of inflammation and dexamethasone." Eur J Cell Biol 46(1): 121-8. Ramey, J. W., R. F. Highsmith, W. W. Wilfinger and D. M . Baldwin (1987). "The effects of gonadotropin-releasing hormone and estradiol on luteinizing hormone biosynthesis in cultured rat anterior pituitary cells.." Endocrinology 120(4): 1503-13. Rearick, J. I., A. Chapman and S. Kornfeld (1981). "Glucose starvation alters lipid-linked oligosaccharide biosynthesis in Chinese hamster ovary cells." J Biol Chem 256(12): 6255-61. Rearick, J. I., K. Fujimoto and S. Kornfeld (1981). "Identification of the mannosyl donors involved in the synthesis of lipid-linked oligosaccharides." J Biol Chem 256(8): 3762-9. Reece, R. J. (2000). "Molecular basis of nutrient-controlled gene expression in Saccharomyces cerevisiae." Cell Mol Life Sci 57(8-9): 1161-71. Romanos, M . A., C. A. Scorer and J. J. Clare (1992). "Foreign gene expression in yeast: a review." Yeast 8(6): 423-88. Rosenwald, A. G., J. Stoll and S. S. Krag (1990). "Regulation of glycosylation. Three enzymes compete for a common pool of dolichyl phosphate in vivo." J Biol Chem 265(24): 14544-53. Rothstein, S. J., K. N . Lahners, W. E. Smith, C. M . Lazarus, D. C. Baulcombe and A. A. Gatenby (1987). "Synthesis and secretion of wheat a-amylase in Saccharomyces cerevisiae." Gene 55: 353-356. 141 Rothstein, S. J., C. M . Lazarus, W. E. Smith, D. C. Baulcombe and A. A. Gatenby (1984). "Sectretion of wheat a-amylase expressed in yeast." Nature 308: 662-665. Saito, A., M . Usui, Y. Song, H. Azakami and A. Kato (2002). "Secretion of glycosylated alpha-lactalbumin in yeast Pichia pastoris." J Biochem (Tokyo) 132(1): 77-82. Shelikoff Marc . Sinskey, A. J. (1994). "The effect of protein synthesis inhibitors on the glycosylation site." Cytotechnology 15(1-3): 195-208. Shiba, Y., F. Fukui, K. Ichikawa, N. Serizawa and H. Yoshikawa (1998). "Process development for high-level secretory production of carboxypeptidase Y by Saccharomyces cerevisiae." Appl Microbiol Biotechnol 50(1): 34-41. Smith, D. F., R. D. Larsen, S. Mattox, J. B. Lowe and R. D. Cummings (1990). "Transfer and expression of a murine UDP-Gal:beta-D-Gal-alpha 1,3-galactosyltransferase gene in transfected Chinese hamster ovary cells. Competition reactions between the alpha 1,3-galactosyltransferase and the endogenous alpha 2,3-sialyltransferase." J Biol Chem 265(11): 6225-34. Sreekrishna, K., R. G. Brankamp, K. E. Kropp, D. T. Blankenship, J. T. Tsay, P. L. Smith, J. D. Wierschke, A. Subramaniam and L. A. Birkenberger (1997). "Strategies for optimal synthesis and secretion of heterologous proteins in the methylotrophic yeast Pichia pastoris." Gene 190(1): 55-62. Stark, N . J. and E. C. Heath (1979). "Glucose-dependent glycosylation of secretory glycoprotein in mouse myeloma cells." Arch Biochem Biophys 192(2): 599-609. Stimson, E., M . Virji, K. Makepeace, A. Dell, H. R. Morris, G. Payne, J. R. Saunders, M . P. Jennings, S. Barker, M . Panico and et al. (1995). "Meningococcal pilin: a glycoprotein substituted with digalactosyl 2,4-diacetamido-2,4,6-trideoxyhexose." Mol Microbiol 17(6): 1201-14. ter Schure, E. G., N . A. van Riel and C. T. Verrips (2000). "The role of ammonia metabolism in nitrogen catabolite repression in Saccharomyces cerevisiae." FEMS Microbiol Rev 24(1): 67-83. Thorens, B. and P. Vassalli (1986). "Chloroquine and ammonium chloride prevent terminal glycosylation of immunoglobulins in plasma cells without affecting secretion." Nature 321(6070): 618-20. Thorpe, E. D., M . C. d'Anjou and A. J. Daugulis (1999). "Sorbitol as a non-repressing carbon source for fed-batch fermentation of recombinant Pichia pastoris.'''' Biotechnol Lett 21: 669-672. Tschopp, J. F., P. F. Brust, J. M . Cregg, C. A. Stilhnan and T. R. Gingeras (1987). "Expression of the lacZ gene from two methanol-regulated promoters in Pichia pastoris." Nucleic Acids Res 15(9): 3859-76. 142 Udeh, K. O. and B. Achremowicz (1997). "High-glutathione containing yeast Saccharomyces cerevisiae: optimization of production." Acta Microbiol Pol 46(1): 105-14. Vad, R., E. Moe, K. Saga, A. M . Kvinnsland and T. B. Oyen (1998). "High-level production of human parathyroid hormone (hPTH) by induced expression in Saccharomyces cerevisiae." Protein Expr Purif 13(3): 396-402. van Kuik, J. A., K. Hard and J. F. Vliegenthart (1992). "A 1H NMR database computer program for the analysis of the primary structure of complex carbohydrates." Carbohydr Res 235(3): 53-68. von Zastrow, M. , A. M . Castle and J. D. Castle (1989). "Ammonium chloride alters secretory protein sorting within the maturing exocrine storage compartment." J Biol Chem 264(11): 6566-71. Wang, X. C , T. P. O'Hanlon and J. T. Lau (1989). "Regulation of beta-galactoside alpha 2,6-sialyltransferase gene expression by dexamethasone." J Biol Chem 264(3): 1854-9. Wang, Y., C. L. Liu, J. D. Storey, R. J. Tibshirani, D. Herschlag and P. O. Brown (2002). "Precision and functional specificity in mRNA decay." Proc Natl Acad Sci U S A 99(9): 5860-5. Whittaker, M . M . and J. W. Whittaker (2000). "Expression of recombinant galactose oxidase by Pichia pastoris." Protein Expr Purif 20(1): 105-11. Wolff, A. M. , O. C. Hansen, U. Poulsen, S. Madrid and P. Stougaard (2001). "Optimization of the production of Chondrus crispus hexose oxidase in Pichia pastoris." Protein Expr Purif 22(2): 189-99. Xie, L., G. B. Nyberg, X. Gu, H. Li , F. Mollborn and D. I. C. Wang (1997). "Gamraa-interferon production and quality in stoichiometric fed-batch cultures of Chinese hamster ovary (CHO) cells under serum-free conditions." Biotechnol Bioeng 56: 577-582. Yamashita, K., N . Koide, T. Endo, Y. Iwaki and A. Kobata (1989). "Altered glycosylation of serum transferrin of patients with hepatocellular carcinoma." J Biol Chem 264(5): 2415-23. Yang, M . and M . Butler (2000). "Effect of ammonia on the glycosylation of human recombinant erythropoietin in culture." Biotechnol Prog 16(5): 751-9. Yang, M. and M . Butler (2000). "Effects of ammonia on CHO cell growth, erythropoietin production, and glycosylation." Biotechnol Bioeng 68(4): 370-80. Zhang, W., M . A. Bevins, B. A. Plantz, L. A. Smith and M . M . Meagher (2000). "Modeling Pichia pastoris growth on methanol and optimizing the production of a recombinant protein, the heavy-chain fragment C of botulinum neurotoxin, serotype A." Biotechnol Bioeng 70(1): 1-8. 143 Zhang, Y., W. Zhou, N . Liu and K. Yang (1991). "Expression of calf prochymosin gene in Escherichia coli." Chin J Biotechnol 7(3): 169^75. Zhou, K., P. R. Brisco, A. E. Hinkkanen and G. B. Kohlhaw (1987). "Structure of yeast regulatory gene LEU3 and evidence that LEU3 itself is under general amino acid control." Nucleic Acids Res 15(13): 5261-73. 144 5 Appendix Nitrogen Balance (2-Litre Bioreactor) N H 4 O H Nitrogen in NH 4 OH = N + (N + O + 5H)"1 = 14 * (14 + 16 + 5)"1 = 0.40 N for fermentation = 30 g-L"1 (27 % NH4OH) * 0.40 = 3.2 g-L"1 (Nitrogen) Peptone Nitrogen in Peptone = 0.16 (nitrogen concentration from Sigma spec sheet) Nitrogen in fermentation = 20 g-L"1 * 0.16 = 3.2 g-L"1 (Nitrogen) Amino Acids Nitrogen in Amino Acid Mix = 0.16 (nitrogen concentration from Sigma spec sheet) Nitrogen in fermentation = 20 g-L"1 * 0.16 = 3.2 g-L"1 (Nitrogen) Nitrogen Balance (Shake flask) NH4HSO4 Nitrogen in NH4HSO4 = N * ( N + 5H + S + 40)"1 = 14 * (14 + 5 + 32 + 4*(16))"1 = 0. Nitrogen for fermentation = 4.6 g-L"1 * 0.12 = 0.552 g-L"1 (Nitrogen) Peptone Nitrogen in Peptone = 0.16 (nitrogen concentration from Sigma spec sheet) Nitrogen in fermentation = 3.45 g-L"1 * 0.16 = 0.552 g-L"1 (Nitrogen) 145 Table A 1 DATA for section 3.5.1 (Analytical Methods - Cell Density) wet tare wet total dry tare dry total wet weight dry weight Sample (g) (g) (g) (g) (g«J) (g«acf) 1 13.1927 14.9044 1.2393 1.686 171.17 44.67 2 13.1225 15.036 1.2445 1.703 191.35 45.85 3 13.0937 14.4237 1.244 1.568 133 32.4 4 13.1021 14.4546 1.2315 1.549 135.25 31.75 5 13.098 14.1011 1.2476 1.473 100.31 22.54 6 13.1389 14.0777 1.2401 1.466 93.88 22.59 7 13.1015 13.7145 1.237 1.372 61.3 13.5 8 13.1824 13.63 1.2522 1.343 44.76 9.08 9 13.113 15.562 1.2489 1.849 244.9 60.01 volume 10 mL Table A 2 DATA for section 4.2 (Recombinant Cystatin C Stability) Fermentor 1 Fresh Sample Frozen Time abs.(410nm) abs.(410nm)rep abs.(410nm)blank abs.(410nm) abs.(410nm)rep abs.(410nm)blank 50 0.815 0.86 1.007 0.674 0.706 0.937 63 0.623 0.584 0.873 0.571 0.525 0.825 74 0.589 0.59 0.827 0.595 0.563 0.881 Glycerol Frozen 2 Time abs.(410nm) abs.(410nm)rep abs.(410nm)blank abs.(410nm) abs.(410nm) abs.(410nm)blank 50 0.869 0.789 0.814 63 0.808 0.774 0.78 0.462 0.479 0.664 74 0.939 0.868 0.788 0.478 0.47 0.723 % Inhibition Time Fresh Frozen Glycerol Frozen 2 50 16.8 26.4 -1.8 63 30.9 33.6 -1.4 29.1 74 28.7 34.3 -14.7 34.4 146 Table A 3 Fermentor 2 Fresh Sample Frozen Time abs.(410nm) abs.(410nm)rep abs.(410nm)blank abs.(410nm) abs.(410nm)rep abs.(410nm)blank 50 0.628 0.664 0.957 0.921 0.869 1.016 63 0.75 0.731 1.047 0.86 0.803 0.952 95 0.217 0.187 0.804 0.222 0.232 0.948 Glycerol Frozen 2 Time abs.(410nm) abs.(410nm)rep abs.(410nm)blank abs.(410nm) abs.(410nm) abs.(410nm)blank 50 0.929 0.82 0.81 63 0.945 0.864 0.865 0.721 0.796 1.04 95 0.817 0.753 0.667 0.203 0.239 0.827 % Inhibition Time Fresh Frozen Glycerol Frozen 2 50 32.5 11.9 -8.0 63 29.3 25.3 -9.1 27.1 95 74.9 76.1 -17.7 73.3 Table A 4 Fermentor 3 Fresh Sample Frozen Time abs.(410nm) abs.(410nm)rep abs.(410nm)blank abs.(410nm) abs.(410nm)rep abs.(410nm)blank 50 0.725 0.764 0.969 0.671 0.632 1.046 74 0.249 0.232 0.851 0.277 0.322 0.996 95 0.203 0.25 0.84 0.225 0.253 0.922 Glycerol Frozen 2 Time abs.(410nm) abs.(410nm)rep abs.(410nm)blank abs.(410nm) abs.(410nm) abs.(410nm)blank 50 0.826 0.73 . 0.84 74 0.823 0.717 0.724 0.321 0.309 0.934 95 0.685 0.729 0.641 0.23 0.221 0.745 % Inhibition Time Fresh Frozen Glycerol Frozen 2 50 23.2 37.7 7.4 74 71.7 69.9 -6.4 66.3 95 73.0 74.1 -10.3 69.7 147 Table A 5 DATA for section 4.3 (Feed Strategy) Time pH Rpm Temp d02 MeOH MeOH Glycerol WCW h:m pH rpm C % millivolts grams grams 0:00 4.79 400 29.9 94.9 2729 831.41 1094.19 0 1:00 5.36 400 30.1 58.3 7238 830.56 1094.19 2:00 5.36 400 30.1 43.7 7209 829.72 1094.49 3:00 5.36 475 30 39 7099 830.28 1094.49 4:00 5.35 558 29.9 39.3 6969 831.13 1094.78 5:00 5.35 617 29.8 39.4 6792 831.41 1094.78 6:00 5.33 706 30 39.3 6665 831.41 1094.78 7:00 5.31 768 30 39.3 6439 831.41 1094.78 8:00 5.27 847 30 39 6171 831.41 1094.78 9:00 5.22 995 30.3 38.3 5675 831.41 1094.78 ' 10:00 5.14 1195 30.3 38.5 4121 831.41 1094.78 38.1818 11:00 5.01 1250 30.4 40.1 2059 831.41 1094.49 12:00 5 1250 30.2 42.8 1911 831.41 1094.78 13:00 5 1250 30.1 41.2 1755 830.84 1094.78 14:00 4.99 1250 30.2 39.7 1768 831.41 1093.9 15:00 4.99 1250 30.1 ' 39.8 1735 829.44 1094.19 16:00 5.01 1250 30 39.6 1555 830.28 1090.65 17:00 4.99 1250 29.6 40.1 1583 829.72 1068.5 18:00 5 1250 30.2 39.7 1613 829.44 1049.31 19:00 5 1250 30 38.1 1637 829.44 1027.76 20:00 5 1250 30 38.9 1700 829.44 1005.91 76.3636 21:00 5 1250 29.9 39.6 1749 829.44 989.37 22:00 5 1250 30 39 2500 1327.69 979.03 23:00 5 1250 29.9 39.3 2075 1327.69 974.9 24:00 5 1250 30 39.7 2259 1327.69 974.9 25:00 5 1250 29.9 39.6 2346 1327.97 974.9 26:00 4.99 1250 29.9 39.7 2381 1328.25 975.19 127 27:00 6.28 1250 29.9 39.3 6077 1317.84 974.9 28:00 6.05 1250 30 39.4 8541 1313.9 974.9 29:00 6 1250 29.9 40.3 8913 1311.09 975.19 30:00 6.01 1250 29.9 40.1 9100 1308 975.19 31:00 6.05 1250 29.9 40.4 9272 1302.09 975.19 32:00 6.06 1250 30 40 9409 1295.62 974.9 33:00 6.07 1250 29.9 39.7 9491 1288.3 975.19 34:00 6.06 1250 29.9 39 9531 1283.52 974.9 35:00 6.05 1250 29.9 40.8 9536 1280.43 974.9 36:00 6.04 1250 29.9 40.7 9596 1271.42 974.9 37:00 6.03 1250 29.9 39 9648 1260.73 975.78 38:00 6 1250 30 40.4 9681 1249.48 974.9 39:00 5.98 1250 30 38.7 9709 1239.07 974.9 40:00 5.99 1250 30 38.4 9728 1228.94 976.97 185.75 41:00 5.99 1250 30 40 9742 1217.41 974.9 42:00 5.99 1250 29.9 40.7 9762 1207.56 975.19 43:00 6 1250 30 39.2 9777 1195.74 975.78 44:00 6 1250 30 38.6 9786 1183.93 975.19 Cystatin Cystatin % Inhibition Activity (nM) 148 Table A 5 (continued) Time pH Rpm Temp d02 MeOH MeOH Glycerol WCW h:m pH rpm C % millivolts grams grams 45:00 6 1250 30 40.3 9796 1174.08 975.19 46:00 6 1250 30 40.8 9798 1166.2 975.49 47:00 6 1250 30 39.7 9792 1166.2 976.08 48:00 6 1250 29.9 40.5 9788 1166.2 975.19 49:00 6 1250 29.9 40.1 9786 1166.2 975.78 50:00 6.01 1250 29.9 40.5 9781 1165.64 975.19 167 51:00 6.01 1250 30 40 9777 1165.64 974.9 52:00 6.02 1250 30 40.1 9772 1164.23 974.9 53:00 6.02 1250 30 40.4 9767 1165.36 974.9 54:00 6.03 1250 29.9 39.9 9760 1165.64 975.49 55:00 6.04 1250 30 39.6 9752 1165.36 975.78 56:00 6.04 1250 30 39.7 9747 1165.08 975.19 57:00 6.05 1250 30 40.2 9742 1165.64 975.19 58:00 6.06 1250 30 40.1 9732 1166.2 975.78 59:00 6.07 1250 30 39.7 9727 1165.08 975.49 60:00 6.07 1250 30 40 9718 1164.51 975.49 61:00 6.08 1250 29.9 39.8 9708 1165.36 974.9 62:00 6.09 1250 30 40.1 9703 1165.92 974.9 63:00 6.1 1250 30 40 9697 1164.51 974.9 171.5 64:00 6.11 1250 30 39.6 9683 1165.92 976.97 65:00 6.11 1250 29.5 39.4 9661 1166.2 976.67 66:00 6.12 1250 30 40.7 9664 1166.2 976.67 67:00 6.12 1250 30.3 39.8 9669 1165.92 975.78 68:00 6.13 1250 30.1 39.8 9663 1166.2 976.37 69:00 6.14 1250 29.9 40 9644 1166.2 975.19 70:00 6.15 1250 29.1 39.8 9624 1166.2 976.08 71:00 6.15 1250 29.7 39.6 9629 1166.2 975.19 72:00 6.16 1250 30 40.5 9625 1165.08 975.49 73:00 6.16 1250 30 39.7 9615 1165.36 975.19 74:00 6.17 1250 29.9 39.8 9605 1165.64 976.37 165.2 75:00 6.17 1250 30 40.1 9599 1165.36 976.08 76:00 6.18 1250 30 40.1 9587 1165.92 975.78 77:00 6.18 1250 30 39.6 9580 1165.08 975.19 78:00 6.18 1250 30 40.3 9571 1164.8 975.19 79:00 6.19 1250 29.9 40.3 9556 1164.23 975.19 80:00 6.19 1250 30 40.5 9541 1164.23 974.9 81:00 6.19 1250 30 40 9536 1164.23 974.9 82:00 6.2 1250 29.9 40.1 9523 1164.23 974.9 83:00 6.2 1250 30 40.1 9507 1164.51 975.19 166.312 84:00 6.2 1250 30 40.3 9497 1164.23 974.9 85:00 6.21 1250 30 40.3 9485 1164.23 974.9 86:00 6.21 1250 29.9 40.3 9472 1164.23 975.19 87:00 6.21 1250 30 40.4 9457 1164.23 974.9 88:00 6.21 1250 29.9 40.4 9433 1164.23 975.49 89:00 6.22 1250 29.8 39.8 9414 1164.23 975.19 90:00 6.22 1250 30 40.4 9409 1164.23 974.9 91:00 6.22 1250 29.9 40 9389 1164.23 974.9 92:00 6.22 1250 29.9 40.1 9364 1164.23 974.9 166.875 93:00 6.22 1250 29.9 40 9340 1164.23 975.49 Cystatin Cystatin 16.83 4.49361 30.87 8.24229 28.71 7.66557 28.04 7.48668 149 Table A 6 D A T A for section 4.3 (Feed Strategy) Time pH Rpm Temp d02 MeOH MeOH Glycerol h:m pH rpm C % millivolts grams grams 0:00 4.82 400 29.6 98.5 1634 951 1131.91 1:00 5 400 30.1 66.9 5238 951 1131.91 2:00 4.99 400 30 56.7 4579 951 1131.91 3:00 5 400 29.9 44.1 4520 951 1131.91 4:00 5.01 443 29.9 39.3 4446 951 1131.91 5:00 4.99 498 29.9 39.6 4334 952 1132.21 6:00 5 547 29.9 39.7 4259 951 1132.5 7:00 4.99 599 30 39.7 4139 951 1133.98 8:00 4.99 646 30.1 39.7 4012 951 1132.8 9:00 5.01 697 30.2 39.9 3841 951 1131.91 10:00 4.99 757 30 39.7 3546 951 1131.91 11:00 4.99 833 30.1 39.7 3048 951 1131.91 12:00 4.99 905 30.1 38.2 2115 951 1132.21 13:00 4.99 1046 30 39.1 1817 951 1131.91 14:00 4.99 1152 30.1 39.4 1589 951 1132.21 15:00 4.99 1207 30.3 40 1440 951 1131.91 16:00 4.99 1207 30 40.1 1221 951 1127.76 17:00 4.99 1143 29.6 40.5 1079 951 1107.03 18:00 5 1080 30 40.1 1041 951 1085.12 19:00 5 1022 30.1 40.5 993 951 1063.5 20:00 4.99 966 30 39.7 954 951 1041.29 21:00 5 912 29.8 40.3 910 951 1024.11 22:00 4.99 893 29.6 40.8 1180 1310 1011.67 23:00 4.998 1001.6 29.9 38.64 934.2 1310 1008.65 24:00 4.994 1164.6 30.22 39.22 947 1310 1009.6 25:00 4.99 1250 30 38 973 1310 1009 26:00 4.992 1250 30 37.94 1003.2 1310 1007.52 27:00 6.072 1250 30 42.52 4012 1299 1007.52 28:00 5.912 1250 29.98 37.34 4188.6 1299 1008.18 29:00 5.992 1250 29.92 45.92 4241.4 1299 1007.52 30:00 5.998 1250 29.9 42.5 4282.2 1299 1007.76 31:00 5.99 1250 29.92 38.26 4236.6 1299 1007.52 32:00 6.116 1249 29.98 36.72 4202 1299 1007.52 33:00 6.122 1105.8 29.92 40.6 4094.4 1299 1007.52 34:00 6.1 855.2 29.98 41.48 3893 1299 1007.52 35:00 6.11 720.8 29.62 40.04 5044.4 1296 1007.52 36:00 6.138 695.4 29.78 39.78 4787.8 1296 1007.52 37:00 6.14 656.8 29.9 40.36 4084.6 1296 1007.52 38:00 6.14 687.6 29.9 40.14 4353 1295 1007.52 39:00 6.14 711 29.92 40.12 4662 1293 1007.29 40:00 6.13 711.8 29.92 39.74 4841.2 1292 1007.52 41:00 6.122 704.4 29.9 39.78 4831.2 1291 1007.52 42:00 6.12 684.2 29.9 38.46 4746.6 1290 1007.52 43:00 6.11 630 29.9 41 3998.2 1290 1007.52 44:00 6.11 639.8 29.92 40.6 4286.4 1288.8 1007.52 45:00 6.11 641.4 29.9 40.02 4494.2 1287 1007.52 WCW Cystatin Cystatin g<MJ % Inhibition Activity QiM) 0 0 0 35.9091 71.8182 124 203.875 150 Table A 6 (continued) Time pH Rpm Temp d02 MeOH MeOH Glycerol h:m pH rpm C % millivolts grams grams 46:00 6.1 658.4 29.8 39.88 4724.2 1286 1007.52 47:00 6.11 623 29.9 38.52 4603.4 1287 1007.52 48:00 6.11 579.6 30 39.9 4007.6 1287 1007.58 49:00 6.11 586 29.9 40.48 4412.2 1286 1007.52 50:00 6.11 582.2 30 40.12 4676.4 1285 1007.52 51:00 6.11 536.8 29.9 40.44 3917.2 1285 1007.52 52:00 6.11 556.8 30 39.98 4451.4 1284 1007.52 53:00 6.11 541 29.9 39.74 4626.6 1283 1007.52 54:00 6.11 512.2 29.9 40.1 4104.4 1283 1007.7 55:00 6.11 520.8 29.9 40.14 4479.6 1282 1007.52 56.00 6.11 487.8 29.9 39.82 3910.4 1282 1007.52 57:00 6.11 501 29.9 39.72 4397.6 1281 1007.52 58:00 6.11 460.2 29.9 41.14 3894.2 1281 1007.52 59:00 6.11 477 29.98 39.98 4349.4 1280 1007.52 60:00 6.11 452.2 29.92 39.86 3900.6 1280 1007.52 61:00 6.11 466.6 29.9 39.88 4462.8 1279 1007.52 62:00 6.11 439.2 29.98 40.04 4131.2 1279 1007.52 63:00 6.11 445.4 29.98 38.98 4387.2 1279 1007.52 64:00 6.11 433 29.9 38.46 4063.4 1278.2 1007.52 65:00 6.11 439.6 29.88 40.36 4496.8 1277.8 1007.82 66:00 6.11 421.4 30 40.72 4321.2 1277 1007.52 67:00 6.11 401.6 30 40.36 4095.2 1277 1007.76 68:00 6.118 400 30 40.76 4337.8 1277 1007.52 69:00 6.12 400 29.9 42.56 4162 1276 1007.52 70:00 6.11 588 29.86 32.8 3892.6 1276 1007.52 71:00 6.024 1229.6 30.5 38.04 4368.8 1276 994.726 72:00 5.97 1250 30.1 40.26 3882.2 1275 982.106 73:00 5.978 1250 29.94 36.34 4455 1274 970.2 74:00 6 1223.2 29.96 42.14 4631.8 1273 960.312 75:00 5.994 1250 29.96 36.98 4151 • 1272.4 947.4 76:00 6 1176.8 29.96 51.18 3958 1272 935.08 77:00 6 1241.6 30 36.76 4093.8 1271 922.996 78:00 5.996 1250 30 34.8 4216.8 1270 910.436 79:00 6 1247.6 29.92 39.94 4054.2 1269 898.116 80:00 6 1250 29.94 36.1 3882 1268 886.038 81:00 5.99 1140.2 30 36.56 3900 1267 876.91 82:00 6 909 29.96 40.16 4721.2 1265.6 873.594 83:00 5.99 858.8 30 40.48 4385.8 1265 873.06 84:00 5.99 840.2 29.96 36.76 4423.4 1263 873.124 85:00 5.99 861.2 29.94 41.32 4308.2 1261 873.18 86:00 5.99 880.2 29.94 39.54 4674.6 1260 872.944 87:00 6 809 29.96 37.6 4434.8 1258 873.006 88:00 5.996 804.2 29.94 42.22 4156.6 1257 873.948 89:00 5.994 836 29.94 40.74 4530 1254 872.944 90:00 5.996 786.2 29.96 35.46 4429.6 1253 873.24 91:00 5.994 774.4 29.94 41.42 4281 1252 872.77 92:00 6 798.6 29.9 39.22 4746 1250 873.36 93:00 5.994 715.6 29.9 42.36 3969 1249 873.36 WCW Cystatin Cystatin % Inhibition Activity (nM) 205 32.5 8.6775 211.4 29.27 7.81509 226.4 29.18 7.79106 284.267 74.8 19.99296 151 Table A 7 DATA for section 4.3 (Feed Strategy) Time pH Rpm Temp d02 MeOH MeOH Glycerol h:m pH rpm C % millivolts grams grams 0:00 4 . 8 2 4 0 0 2 9 . 8 9 7 . 4 598 809 1 0 6 4 . 8 7 1:00 5 4 0 0 2 9 . 9 7 3 . 7 4 6 0 3 809 1 0 6 5 . 4 7 2 :00 5 4 0 0 3 0 6 6 . 7 4 6 3 2 809 1 0 6 5 . 1 7 3 :00 5 4 0 0 3 0 5 8 . 6 4 5 6 7 809 1 0 6 3 . 9 7 4 : 0 0 4 .99 4 0 0 3 0 4 7 . 4 4 4 9 0 809 1 0 6 5 . 1 7 5:00 5.01 4 1 6 2 9 . 9 39 .7 4 3 6 5 809 1 0 6 5 . 7 7 6:00 4 . 9 9 4 5 7 3 0 3 9 . 6 4 3 0 4 8 0 9 1065 .77 7:00 4 . 9 9 4 9 4 3 0 3 9 . 9 4 1 8 8 808 1065 .77 8:00 5 5 3 2 3 0 3 9 . 6 4 0 6 5 808 1065 .77 9:00 4 . 9 9 5 7 8 30.1 3 9 . 6 3 8 6 7 8 0 9 1065 .47 10:00 5 6 3 2 30 3 9 . 7 3611 8 0 9 1065 .47 11:00 5 7 0 0 30.1 3 9 . 6 3 2 2 8 809 1064 .27 12:00 5 7 7 6 3 0 39 .5 2 4 2 6 809 1064 .57 13:00 5 851 3 0 39 .6 6 4 7 809 1065 .77 14:00 ' 5 9 7 7 3 0 39 .3 ' 583 809 1064 .57 15:00 4 . 9 9 1108 3 0 . 4 3 8 . 2 • 5 8 4 8 0 9 1064 .57 16:00 5 1150 30.1 4 0 544 8 0 9 1061 .58 17:00 4 . 9 9 1107 3 0 4 0 . 3 5 3 4 808 1040 .34 18:00 5 1047 2 9 . 9 4 0 . 3 528 809 1 0 1 9 . 6 9 19:00 5 9 8 6 2 9 . 9 4 0 . 4 531 809 9 9 8 . 7 5 2 0 : 0 0 5 8 6 7 3 0 40.1 5 5 4 809 9 7 7 . 8 2 1 : 0 0 4 . 9 9 8 2 6 3 0 40.1 568 809 9 6 1 . 0 5 22 :00 4 . 9 9 811 2 9 . 8 4 0 9 2 8 1341 9 5 2 . 0 8 23 :00 5 9 4 0 . 8 2 9 . 9 6 4 0 . 4 8 691 1341 9 4 8 . 4 9 24 :00 4 . 9 9 4 9 6 1 . 2 2 9 . 9 4 3 8 . 6 4 711 .4 1341 9 4 8 . 4 9 25 :00 5 1163 .6 2 9 . 9 3 9 . 9 6 800 1341 9 4 8 . 4 9 26 :00 4 . 9 9 6 1228 2 9 . 9 4 3 9 . 2 830 .2 1341 9 4 8 . 3 7 27 :00 6 1247 .6 2 9 . 9 3 9 . 7 2 5 4 2 6 . 6 1332 9 4 9 . 9 2 4 28 :00 5 .99 1249 .6 2 9 . 9 4 1 . 7 8 5 4 2 4 . 2 1332 9 4 8 . 4 9 29 :00 5 .99 1247 2 9 . 9 4 0 . 0 8 5391 1332 9 4 8 . 4 9 30 :00 5 .996 1250 3 0 3 9 . 4 4 5 6 4 9 . 6 1331 9 4 8 . 4 9 31 :00 6 1250 3 0 3 9 . 5 2 5 6 0 9 1331 948 .31 32 :00 6 .004 1169 .6 2 9 . 9 4 0 . 2 4 5 6 0 0 . 4 1331 9 4 8 . 3 7 33 :00 6 .16 1 1 2 2 . 6 2 9 . 9 4 4 0 . 4 4 5 5 2 4 1331 9 4 8 . 2 5 3 4 : 0 0 6 .14 9 8 5 . 2 2 9 . 9 6 4 2 . 6 2 5 3 7 1 . 4 1331 948 .31 35 :00 6 .14 7 7 1 . 6 2 9 . 9 4 0 . 1 8 5 7 9 6 . 4 1329 9 4 8 . 3 7 36 :00 6 .15 7 6 0 . 8 2 9 . 9 6 3 9 . 6 8 5 8 2 0 . 2 1329 9 4 8 . 0 7 37 :00 6.15 8 3 4 . 4 2 9 . 9 6 3 9 . 1 8 5 5 0 1 . 8 1327 9 4 8 . 0 7 38 :00 6 .14 9 9 7 . 4 30 3 8 . 2 4 5 6 4 9 . 2 1325 .4 9 4 8 . 1 3 39 :00 6.11 1153 .4 2 9 . 9 4 3 9 . 8 4 5 6 2 6 . 4 1324 948 .31 4 0 : 0 0 6.08 1142 .4 2 9 . 9 3 9 . 5 8 5 7 6 6 . 4 1322 948 .01 4 1 : 0 0 6 .06 1095 .8 2 9 . 9 4 4 0 . 7 2 5 5 0 5 . 2 1319 .4 9 4 8 . 3 7 4 2 : 0 0 6 .046 1048 .8 2 9 . 9 6 3 9 . 6 4 5 5 1 9 . 2 1317 9 4 8 . 0 7 4 3 : 0 0 6 . 0 3 6 9 7 3 . 8 2 9 . 9 6 4 1 . 1 2 5 6 6 2 . 8 1315 .4 9 4 7 . 8 3 4 4 : 0 0 6 . 0 2 6 9 4 0 . 8 2 9 . 9 4 0 . 2 2 5 6 0 3 . 8 1314 948 .31 4 5 : 0 0 6 .02 9 2 1 . 8 2 9 . 9 4 4 0 . 3 6 5 5 1 5 . 8 1312 9 4 8 . 1 3 WCW Cystatin Cystatin % Inhibition Activity (uM) 35 7 0 111 185 .25 152 Table A 7 (continued) Time pH Rpm Temp d02 MeOH MeOH Glycerol h:m pH rpm C % millivolts grams grams 46:00 6.01 926 30 40.8 5439 1310 948.49 47:00 6.01 890 29.9 40.4 5365 1308 948.49 48:00 6 889 30 39.5 5792 1305 948.19 49:00 5.99 877 29.9 39.2 5762 1304 948.49 50:00 5.99 859 30 39.8 5355 1303 948.49 51:00 5.98 849 29.9 40.2 5531 1301 948.49 52:00 5.98 835 30 40.1 5656 1299 948.49 53:00 6 816 30 39.4 5778 1297 947.89 54:00 6 805 29.9 40.1 5612 1296 946.99 55:00 6 798 29.9 39 5477 1294 946.99 56:00 6 795 29.8 40.3 5628 1293 947.59 57:00 6 780 29.8 40.4 5398 1292 947.89 58:00 5.99 779 29.8 40.1 5739 1290 946.99 59:00 6 768 29.9 40.2 5609 1289 947.89 60:00 6 754 29.9 40.2 5453 1287 947.59 61:00 5.99 755 29.9 40 5676 1286 946.69 62:00 5.99 746 30 40.1 5572 1284 947.29 63:00 6 737 30 40.3 5440 1283 947.89 64:00 6 732 29.9 39.9 5357 1282 948.49 65:00 6 746 29.9 39.5 5711 1280 948.49 66:00 5.99 740 29.9 40.3 5660 1278 948.49 67:00 6 733 30 40.1 5644 1277 948.49 68:00 5.99 730 30 39.9 5602 1276 948.49 69:00 5.99 725 29.9 40.1 5469 1275 948.49 70:00 5.99 734 29.9 40.3 5431 1273 948.49 71:00 5.99 801 30 40.1 6530 1272 948.49 72:00 6 794 30 39.9 6491 1271 948.49 73:00 6 777 29.9 40.2 6429 1269 948.49 74:00 5.99 781 29.9 40.2 6446 1268 950.58 75:00 5.99 787 30 40.1 6537 1266 950.58 76:00 6 794 29.9 39.9 6752 1264 950.58 77:00 6 798 30 39.8 6666 1263 949.68 78:00 6 789 29.9 40.3 6562 1262 950.28 79:00 5.99 793 30 40.3 6463 1260 950.58 80:00 6 799 29.9 39.7 6491 1259 949.98 81:00 6 774 29.9 40.1 6116 1234 949.98 82:00 5.99 718 29.9 40.5 4972 1233 949.98 83:00 6 597 29.9 40.7 2697 1233 949.68 84:00 6.01 400 29.7 40.6 1252 1233 949.98 85:00 6.02 400 29.9 59.8 1252 1233 948.78 86:00 6.02 400 29.9 67.1 1249 1232 949.98 87:00 6.03 400 29.9 71.1 1257 1233 949.98 88:00 6.03 400 29.9 74.2 1238 1233 949.98 89:00 6.03 400 29.9 76.9 1212 1232 949.38 90:00 6.04 400 30 78.8 1228 1232 949.08 91:00 6.04 400 29.9 80.7 1200 1232 950.28 92:00 6.05 400 29.9 82.5 1157 1232 950.58 93:00 6.05 400 30 83.9 1135 1232 949.38 WCW Cystatin Cystatin % Inhibition Activity fliM) 194 23.16 6.18372 198.938 34.72 9.27024 193.6 73.03 19.49901 196.667 197.867 72.9 19.4643 153 Table A 8 DATA for section 4.4.1 (Screening Experiments) Time O hr 24 hr 48 hr 0 hr Total Cell Density 24 hr Total Cell Density 48 hr Total Cell Density Run# Tare Weight (g) (gsgtL 1) Tare Weight (g) (gsrfL 1) tare Weight (g) (g^BL 1 ) 1 0.94 1.0303 0.0903 0.9523 1.0485 0.0962 0.9345 1.0316 0.0971 2 0.9655 1.0189 0.0534 0.9572 1.0136 0.0564 0.9532 1.0065 0.0533 3 0.9644 1.0698 0.1054 0.9473 1.0493 0.1020 0.9334 1.032 0.0986 4 0.946 1.0493 0.1033 0.9456 1.0439 0.0983 0.9551 1.0534 0.0983 5 0.9647 1.0674 0.102700 0.9456 1.0458 0.1002 0.9315 1.0307 0.0992 6 0.9616 1.0762 0.1146 0.9489 1.0597 0.1108 0.9493 1.0561 0.1068 7 0.954 1.035 0.081 0.961 1.0473 0.0863 0.9325 1.0138 0.0813 8 0.9455 1.008 0.0625 0.9488 1.0210 0.0722 0.9398 1.0118 0.072 9 0.9534 1.0186 0.0652 0.9546 1.0142 0.0596 0.9444 1.0013 0.0569 10 0.9583 1.0313 0.073 0.9454 1.0108 0.0654 0.9585 1.0289 0.0704 11 0.9356 1.0379 0.1023 0.9415 1.0571 0.1156 0.9452 1.0538 0.1086 12 0.9451 1.0358 0.0907 0.9549 1.0440 0.0891 0.9482 1.0338 0.0856 13 0.9608 1.0373 0.0765 0.9458 1.0171 0.0713 0.9353 1.0068 0.0715 14 0.9621 1.0402 0.0781 0.9647 1.0489 0.0842 0.9432 1.0243 0.0811 15 0.9641 1.0381 0.074 0.9625 1.0335 0.0710 0.9412 1.0264 0.0852 16 0.9449 1.0047 0.0598 0.9483 1.0106 0.0623 0.9603 1.0293 0.069 17 0.9605 1.0863 • 0.1258 0.9563 ' 1.0669 0.1106 0.9587 1.0839 0.1252 18 0.9613 1.0474 0.0861 0.9487 1.0315 0.0828 0.9387 1.0218 0.0831 19 0.9469 1.0253 0.0784 0.9552 1.0371 0.0819 0.9527 1.0349 0.0822 20 0.946 0.9874 0.0414 0.9427 1.0008 0.0581 0.9332 0.9957 0.0625 Time 72 hr 96 hr 72 hr Total Cell Density 96 hr Total Cell Density Run# tare Weight (g) (gstfL"1) tare Weight (g) 1 0.9575 1.0516 0.0941 0.9535 1.0391 0.0856 2 0.9179 0.9742 0.0563 0.9452 0.9877 0.0425 3 0.9206 1.0108 0.0902 0.9451 1.0176 0.0725 4 0.9465 1.0466 0.1001 0.9627 1.0587 0.096 5 0.9455 1.0407 0.0952 0.9485 1.027 0.0785 6 0.9216 1.0199 0.0983 0.9531 1.0231 0.07 7 0.9297 1.0085 0.0788 0.9593 1.0294 0.0701 8 0.9247 1.0189 0.0942 0.9642 1.0438 0.0796 9 0.9187 0.9806 0.0619 0.9359 0.978 0.0421 10 0.9607 1.0059 0.0452 0.9535 0.9986 0.0451 11 0.9475 1.0379 0.0904 0.9577 1.0302 0.0725 12 0.9204 1.0153 0.0949 0.9634 1.0496 0.0862 13 0.9208 0.9876 0.0668 0.9633 1.0191 0.0558 14 0.9243 1.0026 0.0783 0.962 1.0405 0.0785 15 0.9276 1.0024 0.0748 0.9657 1.0207 0.055 16 0.9218 0.9841 0.0623 0.9537 1.0198 0.0661 17 0.9237 1.0237 0.1 0.9592 1.0488 0.0896 18 0.9544 1.0269 0.0725 0.9584 1.0283 0.0699 19 0.9276 1.0186 0.091 0.9527 1.034 0.0813 20 0.9201 0.9926 0.0725 0.9363 0.9995 0.0632 154 Table A 9 D A T A for section 4.4.1 (Screening Experiments) cystatin conversion = 13.39 mgsgfholes" methanol feed time = 48 hrs abs. (410) dilution prep final D C W cystatin cystatin cell specific run # Run # sample repeat blank samples dil gDcw^ri U M mg^gj yield umo les^DCW" ' 1 2 0.5 0.5 0.542 1 2 45.2 1.03 13.85 0.023 2 3 0.54 0.55 0.555 1 2 26.7 0.24 3.22 0.009 3 7 0.48 0.45 0.539 1 2 52.7 1.83 24.54 0.035 4 8 0.47 0.47 0.557 1 2 51.6 2.09 27.92 0.040 5 1 0.7 0.656 0.9 1 1 51.3 6.59 88.19 0.128 6 • 4 0.47 0.46 0.591 1 2 57.3 2.85 38.11 0.050 ' 7 5 0.55 0.55 0.543 1 2 40.5 0.00 0.00 0.000 8 6 0.52 0.53 0.561 1 2 31.3 0.86 11.47 0.027 9 12 0.53 0.51 0.541 1 2 32.6 0.52 6.94 0.016 10 20 1.01 0.98 1.232 1 1 36.5 5.14 68.78 0.141 11 10 0.58 0.567 0.628 1 2 51.2 1.16 15.51 0.023 12 16 0.88 0.89 0.95 1 1 45.3 1.83 24.46 0.040 13 15 0.53 0.55 0.56 1 2 38.3 0.48 6.38 0.012 14 11 0.6 0.63 0.63 1 2 39.1 0.32 4.26 0.008 15 13 0.54 0.54 0.627 1 2 37.0 1.85 24.80 0.050 16 17 0.48 0.5 0.53 1 2 29.9 1.01 13.49 0.034 17 9 0.43 0.46 0.58 1 2 62.9 3.11 41.61 0.049 18 19 0.56 0.58 0.68 1 2 43.0 2.16 28.92 0.050 19 18 0.59 0.57 0.696 1 2 39.2 2.23 29.79 0.057 20 14 0.48 0.51 0.51 1 2 20.7 0.39 5.26 0.019 Error Error jmp cell specific productivity cystatin error growth cell specific yield cell specific productivity run # Run# n m o l e s ^ C W " 1 ^ ' u M g D C W ^ J umolessfiDCW " , nmolesqgDCW " ' ^B" 1 1 2 1.711E4O0 0.00 9.03 0.000 0.000 2 3 6.723E-01 0.24 5.34 0.009 0.672 3 7 2.596E4O0 0.74 10.54 0.014 1.053 4 8 . 3.016E-HX) 0.00 10.33 0.000 0.000 5 1 9.579E400 1.31 10.27 0.025 1.899 6 4 3.711E+O0 0.23 11.46 0.004 0.295 7 5 0.000E-K10 0.00 8.10 0.000 0.000 8 6 2.047E-KJ0 0.24 6.25 0.008 0.569 9 12 1.187E+O0 0.49 6.52 0.015 1.131 10 20 I.051E+01 0.65 7.30 0.018 1.330 11 10 1.691E+00 0.28 10.23 0.005 0.403 12 16 3.009E+00 0.28 9.07 0.006 0.463 13 15 9.309E-01 0.48 7.65 0.012 0.931 14 11 6.079E-01 0.64 7.81 0.016 1.216 15 13 3.739E-K10 0.00 7.40 0.000 0.000 16 17 2.519E-+O0 0.50 5.98 0.017 1.259 17 9 3.691E-+O0 0.69 12.58 0.011 0.820 18 19 3.748E-K)0 0.39 8.61 0.009 0.681 19 18 4.240E4O0 0.38 7.84 0.010 0.731 20 14 1.417E-K30 0.79 4.14 0.038 2.835 155 Table A 10 DATA for section 4.4.2 (2-litre Bioreactor Experiments) Conditions NH4 +AA NH4 FI F2 Time growth growth (hours) g-jfj g^d 0 0 0 14 119 132 18 190 182 19 207 201 25.75 207 204 30.5 189 188 43 187 198 52.5 192 200 67.5 184 198 75 193 216 80.5 180 196 93.5 186 205 122 190 200 Conditions Peptone NH4 + Peptone FI F2 Time growth growth (hours) g<jrJ g<£TJ 0 0 0 13 26 45 16.5 64 93 17.75 75 107 19 94 130 22.5 . 130 172 24 165 200 37.5 185 212 47.5 190 182 51 192 186 62.5 190 182 69.5 185 186 86 180 182 93 183 188 Conditions AA + Peptone AA Conditions NH4 + Peptone + AA FI F2 FI Time growth growth Time growth (hours) g<ffJ g ^ (hours) g<fi} 0 0 0 0 0 12.5 59 62 14 98 14 93 98 19 203 18.5 155 149 26 186 20 202 205 37.5 195 25 190 190 46 190 36.5 205 193 63 196 40.5 200 200 86 185 58.5 207 210 68 203 202 83 208 206 156 Table A l l DATA for section 4.4.2 (2-litre Bioreactor Experiments) F2 R U N #1 abs. (410) Dilution Cystatin Error Productivity Cystatin Time (hrs) Sample Repeat Blank Samples Blank U M (+/-) HMsfDCW"1 0 0 0 0 1 1 0.00 0.00 0.00 0.00 17 0 0 0 1 1 0.00 0.00 0.00 0.00 43 0.969 1.005 1.202 1 1 9.55 0.03 0.22 0.19 52 1.09 1.042 1.381 1 1 12.18 -0.04 0.23 0.24 75 0.985 1.023 1.363 1 1 14.07 0.03 0.19 0.28 93 1.826 1.555 2.351 1 1 15.00 -0.12 0.16 0.30 F l R U N #2 abs. (410) Dilution Cystatin Error Productivity Cystatin Time (hrs) Sample Repeat Blank Samples Blank uM H M ^ D C W " 1 0 0 0 0 1 1 0 0.00 0.00 0.00 17 0 0 0 1 1 0 0.00 0.00 0.00 36 0.846 0.801 0.618 1 0.5 17.82 -0.04 0.50 0.36 45 0.611 0.621 0.711 1 0.5 30.27 0.01 0.67 0.61 58 0.493 0.506 0.763 1 0.5 35.92 0.01 0.62 0.72 68 0.473 0.521 0.769 1 0.5 36.14 0.03 0.53 0.72 89 0.76 0.753 1.065 1 0.5 34.43 0.00 0.39 0.69 F l R U N #3 abs. (410) Dilution Cystatin Error Productivity Cystatin Time (hrs) Sample Repeat Blank Samples Blank U M (+/-) HMsgDCW"' 0 0 0 0 1 1 0 0.00 0.00 0.00 17 0 0 0 1 1 0 0.00 0.00 0.00 37 0.6 0.56 0.884 0.5 0.5 18.36 -0.05 0.50 0.37 47 0.724 0.782 1.163 0.5 0.5 18.83 0.05 0.40 0.38 62 0.704 0.77 1.174 1 0.5 36.64 0.03 0.59 0.73 86 0.884 0.802 0.947 1 0.5 29.63 -0.05 0.34 0.59 93 0.854 0.888 0.9 1 0.5 27.56 0.02 0.30 0.55 F l R U N #4 abs. (410) Dilution Cystatin Error Productivity Cystatin Time (hrs) Sample Repeat Blank Samples Blank U M (+/-) H M ^ D C W " 1 0 0 0 0 1 1 0.00 0.00 0.00 0.00 17 0 0 0 1 1 0.00 0.00 0.00 0.00 43 0.99 0.966 1.201 1 1 9.92 -0.02 0.23 0.20 52 0.658 0.702 0.942 1 1 14.85 0.05 0.29 0.30 75 0.942 0.826 1.34 1 1 18.17 -0.09 0.24 0.36 93 1.389 1.316 2.071 1 1 18.53 -0.04 0.20 0.37 157 Table A l l (continued) F2 R U N #5 abs. (410) Dilution Cystatin Error Productivity Cystatin Time (hrs) Sample Repeat Blank Samples Blank U M (+/-) H M s f D C W 1 0 0 0 0 1 1 0 0.00 0.00 0.00 17 0 0 0 1 1 0 0.00 0.00 0.00 37 0.97 1.06 0.834 1 0.5 20.91 0.06 0.57 0.42 47 0.743 0.94 0.962 1 0.5 30.04 0.11 0.64 0.60 62 0.627 0.584 0.856 1 0.5 34.51 -0.03 0.56 0.69 86 0.616 0.478 0.88 1 0.5 36.80 -0.08 0.43 0.74 93 0.58 0.66 0.886 1 0.5 34.72 0.05 0.37 0.69 FI R U N #6 abs. (410) Dilution Cystatin Error Productivity Cystatin Time (hrs) Sample Repeat Blank Samples Blank U M (+/-) u M ^ i U M ^ C W " 1 0 0 0 0 1 1 0 0.00 0.00 0.00 17 0 0 0 1 1 0 0.00 0.00 0.00 37 1.057 1.169 1.462 1 1 12.75 0.08 0.34 0.25 49 0.409 0.434 0.95 1 1 29.71 0.03 0.61 0.59 71 0.188 0.187 0.791 1 1 40.74 0.00 0.57 0.81 94 0.326 0.295 0.893 1 1 34.83 -0.04 0.37 0.70 F2 R U N #7 abs. (410) Dilution Cystatin Error Productivity Cystatin Time (hrs) Sample Repeat Blank Samples Blank u M (+/-) u M ^ H M # C W " ' 0 0 0 0 1 1 0 0.00 0.00 0.00 17 0 0 0 1 1 0 0.00 0.00 0.00 36 1 1.053 0.602 1 0.5 7.87 0.05 0.22 0.16 45 0.855 0.91 0.961 1 1 4.36 0.06 0.10 0.09 58 0.7 0.685 0.738 1 1 3.29 -0.02 0.06 0.07 68 0.654 0.634 0.752 1 1 7.67 -0.03 0.11 0.15 89 0.712 0.759 0.78 1 1 3.05 0.06 0.03 0.06 158 Table A 12 MAX Final Max Final Time Run # Productivity u r v L ^ Productivity Error Error (hours) 1 0.23 0.16 0.00 -0.12 52 2 0.67 0.39 0.00 0.00 45 3 0.59 0.30 0.00 0.02 62 4 0.29 0.20 0.00 -0.04 52 5 0.64 0.37 0.00 0.05 47 6 0.61 0.37 0.00 -0.04 49 7 0.22 0.03 0.00 0.03 36 MAX Final Max Run # Productivity Productivity Error nmoles^) CW"1 ^ P 1 nmoles<^ ) CW*1 1 1 4.68 3.23 0.00 2 13.45 7.74 0.00 3 11.82 5.93 0.00 4 5.71 3.98 0.00 5 12.78 7.47 0.00 6 12.13 7.41 0.00 7 4.37 0.68 0.00 Table A 13 Run# UG SG DG Total Glycosylated (%) Conditions SG (%) DG (%) l 669 39 0 708 6 NH4 100 0 1 604 43 0 647 7 NH4 100 0 2 2021 566 302 2889 30 Peptone x A A 65 35 3 1641 369 222 2232 26 Peptone 62 38 3 965 232 140 1337 28 Peptone 62 38 4 917 53 52 1022 10 NH4 x AA 50 50 5 1248 94 81 1423 12 NH4 x Peptone 54 46 6 949 94 87 1130 16 Peptone x A A x NH4 52 48 7 1521 153 152 1354 23 AA 50 50 159 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058680/manifest

Comment

Related Items