VOT 74192 THE DEVELOPMENT AND OPTIMIZATION OF …eprints.utm.my/id/eprint/4281/1/74192.pdf ·...

VOT 74192

THE DEVELOPMENT AND OPTIMIZATION OF PROCESSES FOR THE

EXPRESSION OF SIALYLATED RECOMBINANT HUMAN

THERAPEUTIC GLYCOPROTEIN IN INSECT CELL-BACULOVIRUS

SYSTEM

DR. AZILA ABDUL AZIZ

PUSAT PENGURUSAN PENYELIDIKAN

UNIVERSITI TEKNOLOGI MALAYSIA

2007

ii

ACKNOWLEDGEMENTS

We would like to express our appreciation to all those who have assisted us,

whether directly or indirectly, in conducting this research. There are too many names

to list down. However, special thanks goes to the research assistants, Yap Wei Ney,

Clarence Ongkudon and Wee Chen Chen for their dedications and tireless efforts in

making this research work a success.

iii

ABSTRACT

The objectives of this research were to determine the optimal parameters

(culture conditions, transferases and sugar nucleotides content) for the expression of

complete recombinant human glycoprotein and develop an optimal processing

condition for the production of human like glycoprotein in an artificial system by the

manipulation of metabolic engineering and process engineering approach. In the

early part of the study, fundamental works were carried out to optimize Spodoptera

frugiperda (Sf-9) cells growth and mock infection. Serum concentration, different

type of media, cell subculturing condition, initial cell density and spent medium

carry over had been found to significantly influence the growth kinetics of Sf-9 cells.

The optimized parameters were then used to evaluate the expression of recombinant

hTf and �1,4-GalT in Sf-9 cells. Time course expression profiles of rhTf at various

multiplicities of infection (MOI), seeding densities (SD), times of infection (TOI),

and harvest times (HT) were studied. Screening experiments were conducted to

identify the medium components in Sf900-II SFM and the recombinant baculovirus

stock that resulted in improved production of rhTf. Finally, Response Surface

Methodology (RSM) was employed to hunt for optimum medium composition. The

results showed that the optimum HT for rhTf was between 24 to 72 hours post

infection, at SD of 1.6 x 106

viable cells/ml, TOI of day 2 post seeding, and MOI of 5

pfu/cell. Glucose and glutamine were found to have the most positive effect on rhTf

production with more than 95% significance. In addition to that, the best

recombinant baculovirus stock was identified at 98.7% purity. With the optimized

parameters, rhTf production had increased three-fold from 19.89�g/ml to

65.12�g/ml. Subsequently, native UDP-Gal levels at normal and upon baculovirus

infection produced in Sf-9 cells were monitored using Reverse Phase High

iv

Performance Liquid Chromatography. UDP-Gal concentration was discovered to

decrease gradually once infected with the recombinant baculovirus. Finally,

baculovirus coinfection study was carried out to evaluate the recombinant

glycoprotein quality. However, lectin binding analysis using Ricinus communis

agglutinin-I, revealed that co-expression between rhTf and �-1,4GalT (in vivo) did

not show encouraging result due to the reduction of UDP-Gal upon baculovirus

infection. This finding suggested that the introduction of �-1,4GalT alone was not

sufficient for successful galactosylation. However, another strategy was used to

overcome the problem. Commercial GalT and UDP-Gal were introduced artificially

to the rhTf after it was secreted from cell culture. It was found that the in vitro

strategy promoted better N-glycan quality in insect cells. Last stage of the research

was based on rhTf purification, to get a pure rhTf with improved recovery. Steps of

purification were hydrophobic chromatography, dialysis and ion exchange

chromatography. Elution strategy, flowrate and rhTf loading capacity of phenyl

sepharose 6 fast flow were optimized. Batch purification in reduced sized was used

to select suitable anion exchange matrix, pH and concentration of equilibration

buffer. 74.6% of rhTf was recovered from phenyl sepharose, 86.8% recovered after

dialysis, and 52.5% recovered from Q-sepharose and the overall recovery of pure

rhTf was 34%.

v

ABSTRAK

Objektif kajian ini ialah menentukan parameter yang optima (keadaan kultur,

transferase dan kandungan gula nukleotida) untuk menghasilkan rekombinan

glikoprotein manusia yang sempurna dan membentuk suatu keadaan pemprosesan

yang optima untuk menghasilkan glikoprotein yang mimik manusia dalam sistem

tiruan dengan memanipulasikan kejuruteraan metabolik dan kejuruteran. Dalam

kajian awal, kerja asas mengenai pengoptimuman telah dilakukan bagi pertumbuhan

sel dan jangkitan bakulovirus tanpa membawa gen tertentu dalam sel Spodoptera

frugiperda (Sf-9). Kepekatan serum, medium yang berbeza, keadaan sel subkultur,

ketumpatan sel awal dan medium telah-guna telah memberi kesan yang ketara

terhadap kinetik pertumbuhan sel Sf-9. Semua parameter yang telah dioptimumkan

telah digunakan untuk menilai ekpresi bagi rekombinasi hTf and �1,4-GalT dalam

sel Sf-9. Kajian dilakukan ke atas profil ekspresi lawan masa bagi rhTf pada pelbagai

gandaan jangkitan (MOI), kepekatan pembenihan (SD), masa jangkitan (TOI) dan

masa penuaian (HT). Eksperimen penyaringan dilakukan untuk mengenalpasti

komponen dalam medium Sf900-II SFM dan juga stok bakulovirus rekombinan yang

dapat meningkatkan lagi penghasilan rhTf. Akhirnya, Metodologi Permukaan

Tindakbalas (RSM) dijalankan untuk mencari komposisi medium yang optimum.

Hasil kajian mendapati bahawa, nilai optimum untuk HT ialah pada 24 hingga 72

jam selepas jangkitan pertama, SD sebanyak 1.6 x 106

sel produktif/ml, TOI pada

hari ke-2 selepas pembenihan pertama dan MOI sebanyak 5 pfu/ml. Glukosa dan

glutamina didapati mempunyai kesan yang paling positif terhadap penghasilan rhTf

dengan nilai signifikan melebihi 95%. Stok bakulovirus rekombinan yang terbaik

dikenalpasti pada 98.7% ketulinan. Melalui parameter-parameter yang telah

dioptimumkan, penghasilan rhTf telah meningkat sebanyak 3-kali ganda iaitu

vi

daripada 19.89ug/ml kepada 65.12ug/ml. Seterusnya, tahap UDP-Gal semulajadi

pada normal dan atas jangkitan bakulovirus yang dihasilkan dianalisis dengan

menggunakan Fasa Terbalik Kromatografi Cecair Pertunjukkan Tinggi. Didapati

bahawa kepekatan UDP-Gal menurun secara perlahan sebaik sahaja dijangkiti

dengan rekombinasi bakulovirus. Akhirnya, jangkitan serentak bakulovirus telah

dilakukan bagi menilai kualiti glikoprotein rekombinasi. Tetapi, analisis lektin

perlekatan dengan menggunakan Ricinus communis agglutinin-I, menunjukkan in

vivo galaktosilasi tidak cukup berkesan disebabkan kekurangan UDP-Gal semasa

jangkitan bakulovirus. Keputusan yang menarik ini mencadangkan bahawa

penambahan �1,4-GalT sahaja tidak cukup untuk menjayakan galaktosilasi. Oleh itu,

strategi lain telah digunakan untuk mengatasi kelemahan ini. GalT mamalia dan

UDP-Gal yang diperolehi secara komersil diperkenalkan kepada supernatan hTf yang

dikumpul. Didapati bahawa kaedah ini berjaya meningkatkan kualiti N-glikan

dengan baik. Peringkat terakhir untuk kajian ini ialah penulenan rhTf untuk

mendapat rhTf yang tulen dengan pembaikan produktiviti. Langkah penulenan

termasuk hydrophobik kromatografi, dialisis, penukaran ion kromatografi. Strategi

elusi, kadar alir dan muatan kapasiti phenyl sepharose 6 fast flow untuk rhTf

dioptimumkan. Longgok penulenan secara saiz kecil telah dilakukan untuk memilih

matriks penukar anion, pH dan kepekatan larutan penampan. 74.6% rhTf telah

diperolehi daripada phenyl sephaorse, 86.8% diperolehi daripada dialisis dan 52.5%

diperolehi daripada Q-Sepharose dan keseluruhannya, 347% rhTF tulen diperolehi.

vii

TABLE OF CONTENTS

CHAPTER TITLE PAGE

TITLE i

ACKNOWLEDGEMENTS ii

ABSTRACT iii

ABSTRAK v

TABLE OF CONTENTS vii

LIST OF TABLES xv

LIST OF FIGURES xvii

1 INTRODUCTION

1.1 Research Background 1

1.2 Objective 2

1.2 Scopes 3

2 LITERATURE REVIEW

2.1 Recombinant Protein Manufacturing Technologies 4

2.2 Glycosylation 6

2.2.1 N-Linked Glycosylation 7

2.2.2 O-Linked Glycosylation 7

viii

2.3 Glycoprotein 7

2.4 Insect Cell Baculovirus Expression System 8

2.4.1 Insect Cell Lines 9

2.4.2 Baculoviruses 9

2.4.2.1 Baculoviruses Replication 11

2.4.2.1.1 In Vivo Replication 11

2.4.2.1.2 In Vitro Replication 11

2.5 Advantages of Baculovirus Expression System 17

2.6 Model Glycoprotein 18

2.6.1 Native Human Transferrin (nhTf) 18

2.6.2 Recombinant Human Transferrin (rhTf) 21

2.7 Enzyme Immobilization 22

2.7.1 Protein Hydrolysates (Peptones) 22

2.7.2 Carbohydrates 23

2.7.3 Amino Acids 23

2.7.4 Lipids 24

2.7.5 Albumin 24

2.7.6 Serum Free Medium (SFM) 25

2.8 Optimization of Protein Expression in BEVS 25

2.8.1 Physical Factors that Ensure Success of 25

Expression

2.8.2 Optimization of Recombinant Baculovirus Stock 27

2.8.3 Medium Optimization 28

2.9 Design, Analysis and Optimization of Experiments 29

2.9.1 Design of Experiments 29

2.9.1.1 Factorial Experiments in Completely 30

Randomized Designs

ix

2.9.1.2 Interactions 30

2.9.1.3 Coded Variables 30

2.9.1.4 Factor Levels Combinations 31

2.9.1.5 Fractional Factorial Experiments 32

2.9.1.6 Screening Experiments 33

2.9.2 Analysis of Experiments 34

2.9.2.1 Correlation 34

2.9.2.2 Regression Analysis 35

2.9.2.3 Nonlinear and Higher-Order Regression 36

Analysis

2.9.3 Optimization of Experiments 36

2.9.3.1 Improvements of RSM 37

2.10 Glycosylation in Insect Cells 37

2.11 Glycosyltransferases and Glycosidases Involved in 40

N-glycan Processing in Insect Cells

2.11.1 α-Glucosidase I, II and α-Mannosidase I 40

2.11.2 N-Acetylglucosaminyltransferase I (GlcNAcT-I) 40

and α-mannosidase II

2.11.3 N-Acetylglucosaminyltransferase II (GlcNAcT-II) 41

2.11.4 �-1,4-Galactosyltransferase (�1,4-GalT) 41

2.11.5 Core α-1,3- and α-1,6-Fucosyltransferases (FucT) 41

2.11.6 �-N-Acetylglucosaminidase 42

2.11.7 Sialyltransferase (SiaT) 42

2.12 Sugar Nucleotides Involved in N-glycan Processing in 43

Insect Cells

2.12.1 Endogenous Sugar Nucleotide Levels in 43

Lepidopteran Insect Cells

2.12.2 Enzymes Involved in Sialic Acid and CMP-Sialic 43

Acid Synthesis

x

2.13 Engineering of N-glycan Processing Pathway 44

2.13.1 Improvement of N-Acetylglucosaminylation of the 46

Manα(1,3)-Branch

2.13.2 Improvement of Galactosylation 47

2.13.3 Production of Biantennary Complex-Type 47

N-glycans

2.13.4 Formation of Sialylated N-glycans 47

2.13.5 Synthesis of CMP-NeuNAc 48

2.14 Galactosylation in N-Glycan Processing in Insect Cells 48

2.14.1 Sugar acceptor 49

2.14.2 Substrate Donor 49

2.14.3 Enzyme 52

2.15 Purification of Transferrin 53

2.15.1 Hydrophobic Interaction Chromatography (HIC) 53

2.15.1.1 Factors affecting HIC 54

2.15.2 Ion Exchange Chromatography 58

2.15.2.1 Factor affecting IEX 59

3 MATERIALS AND METHODS

3.1 Materials 63

3.2 Chemicals 63

3.3 Equipments 64

3.4 Spodoptera frugiperda (Sf-9) Insect Cells 65

3.4.1 The Preparation of TC100 Medium From 65

Powdered Formulation

3.4.2 Cells Thawing 65

3.4.3 Cells Maintaining 66

3.4.4 Cells Freezing 66

3.4.5 Adapting serum contain culture to serum free 66

culture

xi

3.4.6 Adapting Monolayer Cells to Suspension culture 67

3.4.7 Maintaining suspension culture 67

3.5 Wild Type and Recombinant Baculovirus 68

3.5.1 Virus Propagation 68

3.5.2 Virus Titration (End-Point Dilution) 68

3.5.3 Generating Pure Recombinant Virus Stocks (End 69

Point Dilution)

3.6 Optimization of Recombinant Human Transferrin (rhTf) 70

Expression

3.6.1 Optimization of rhTf Expression in Monolayer 70

Culture

3.6.2 Medium Screening 70

3.6.3 Medium Optimization in Suspension Culture 71

3.7 Response Surface Methodology, RSM (Method of 72

Steepest Ascent)

3.8 Optimized Expression of rhTf 73

3.8.1 Preparation of Optimized Medium 73

3.8.2 Adapting suspension culture in medium SFM900II 73

to optimized medium

3.8.3 Expression of rhTf 74

3.9 Characterization of rhTf 74

3.9.1 Sodium Dodecyl Sulfate - Polyacrylamide Gel 74

Electrophoresis

3.9.1.1 Silver Staining 75

3.9.1.2 Coomassie Blue Staining 75

3.9.2 Western Blot 76

3.9.3 Enzyme Linked Immunosorbent Assay 76

A

xii

3.10 Characterization of nutrients consumption and 77

substances release

3.10.1 Analysis of glucose, lactic acid, glutamine 77

3.10.2 Ammonia test 78

3.11 Protein Analysis Techniques 79

3.11.1 Bicinchoninic Acid (BCA) Assay 79

3.12 Recombinant _1,4-Galactosyltransferase Detection 79

3.12.1 Thin Layer Chromatography 79

3.12.2 Lectin Binding Assay 80

3.13 Native Uridine-5_-diphosphogalactose (UDP-Gal) Level 81

3.13.1 UDP-Gal Extraction 81

3.13.2 Reverse Phase High Performance Liquid 81

Chromatography (RP-HPLC) Analysis

3.14 Coexpression of Recombinant Human Transferrin and 81

β1,4-Galactosyltransferase

3.15 Purification 82

3.15.1 Hydropbobic interaction Chromatography 82

3.15.2 Dialysis 84

3.15.3 Ion Exchange Chromatography 84

3.15.4 Batch Purification 85

4 RESULTS AND DISCUSSION

4.1 The Study of Sf9 Insect Cells Culture Growth Profiles 86

4.1.1 Fundamental Study of Sf9 Cells Growth 86

(Monolayer)

4.1.2 Sf9 cell growth in Shake flask (Suspension) and 97

Comparison with growth in T-flask (Monolayer)

4.1.3 Development of Sf9 Suspension Culture System in 98

xiii

24-well Plate

4.1.4 Growth Analysis 100

4.2 Establishment of Baculovirus Expression Vectors System 102

(BEVS)

4.2.1 Mock Infection Optimization 102

4.3 Study on the Expression Profiles of rhTf in Infected Sf9 109

Insect Cells Culture

4.3.1 rhTf Expression at Different MOIs 109

4.3.2 rhTf Expression at Different Seeding Densities 111

4.3.3 rhTf Expression at Different Times of Infection 113

4.4 Optimization of the Recombinant Human Transferrin 116

Expression

4.4.1 Recombinant Baculovirus Screening 116

4.4.2 Medium Screening 121

4.4.3 Medium Optimization using Response Surface 130

Methodology

4.4.3.1 Regression Model 130

4.4.3.2 Nutrients Interactions 133

4.5 Characterization of the Optimized Recombinant Human 138

Transferrin Expression

4.6 Study of Galactosylation 146

4.6.1 Recombinant �1,4-Galactosyltransferase 146

Expression

4.6.1.1 Time Course Expression of �1,4- 147

Galactosyltransferase

4.6.1.2 The Development of 151

�1,4-Galactosyltransferase Assay

4.6.2 Native Uridine-diphosphogalactose (UDP-Gal) 154

Monitoring at Normal and Upon Baculovirus

xiv

Infection

4.6.3 Baculovirus Coinfection Study 164

4.7 Purification 170

4.7.1 Profile of Sample Elution from hydrophobic 170

Interaction Chromatography

4.7.2 Hydrophobic Interaction Chromatography 171

Optimization

4.7.2.1 Optimization of Elution Method 171

4.7.2.2 Optimization of elution flowrate 173

4.7.2.3 Optimization of rHtf loading capacity 176

4.7.3 Batch Purification 179

4.7.4 Anion Exchange Chromatography 181

4.7.5 Characterization of rhTf purification 182

5 CONCLUSIONS AND RECOMMENDATIONS

5.1 Fundamental Study of Sf9 Cells Growth 186

5.2 Mock Infection and the Expression Profile of rhTf 187

5.3 Strategic Optimization of the Baculovirus Insect Cell 188

Expression System.

5.4 Study of Galactosylation 188

5.5 Study of Purification 190

5.6 Recommendations 190

REFERENCES 192

APPENDICES 229

xv

LIST OF TABLE

TABLE TITLE

PAGE

2.1 Seeding densities for typical vessel sizes (O’Reilly et al., 1994) 26

2.2 Example of a 4-Factor, 2-level Full Factorial Experiment 32

2.3 Example of 12-Run, 11-Factor, 2-Level, Screening Design 33

(Not Randomized)

2.4 Physical Properties of some solvent used in HIC 58

2.5 Functional groups used on ion exchangers 61

2.6 Capacity data for Sepharose Fast Flow ion exchangers 62

2.7 Characteristics of Q, SP, DEAE and CM Sepharose Fast Flow. 62

3.1 Suitable culture volume 67

3.2 Specification of YSI calibrator 78

4.1 Growth Kinetics of Sf-9 Cells at Different Parameters 95

4.2 Comparison of Sf9 growth in T-flask, Shake flask, and 24-well plate 100

4.3 Growth Kinetics of Sf-9 Cells After Mock Infection 108

4.4 rhTf yield coefficients at various seeding densities, MOIs, and times 115

of infection.

4.5 Concentration (�g/ml) of rhTf in each well of a 96-well plate 117

4.6 Viral Screening by End Point Dilution Method (Poisson distribution 120

data sheet)

4.7 Factors affecting the end point dilution method 121

4.8 Real values for the screening of 13 selected nutrients using Plackett- 122

Burman design

4.9 13-factor (nutrients), 33-run, 2-level Plackett-Burman screening 123

design

xvi

4.10 Estimated effects on rhTf yield based on the results of Plackett- 127

Burman screening experiments

4.11 Central composite design for the optimization of glutamine, glucose 131

and lipid mixtures 1000x

4.12 Analysis of Variance (ANOVA) of the CCD 133

4.13 Summary of the characteristics of optimized rhTf expression 138

4.14 Optimization of step-wise elution method for achieving higher 172

recovery of rhTf.

4.15 Optimization of elution flowrate 174

4.16 Optimization of rhTf loading capacity 176

4.17 Summary of the characteristic of purification of rhTf 182

xvii

LIST OF FIGURES

FIGURE TITLE PAGE

2.1 N-linked protein glycosylation 6

2.2 (a) High Mannose, (b) Complex and (c) Hybrid 8

2.3 A few insect species used for glycoprotein production 9

2.4 Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV) 10

2.5 A) Baculovirus particles, or polyhedra; B) Cross-section of a 11

polyhedron; C) Diagram of polyhedron cross-section

2.6 In vivo baculovirus infection and replication cycle 12

2.7 In vitro baculovirus infection and replication 14

2.8 Structural compositions of the two baculovirus phenotypes, budded 15

virus (BV), and the occlusion derived virus (ODV)

2.9 (a) A typical infected Sf-9 cells showing the presence of polyhedra is 16

indicated by the arrow (Steven Howard); (b) Electron micrograph of

AcMNPV infected Sf-9 Cell (Greg V. Williams); (c) A portion of the

nucleus containing enveloped virions in the process of being occluded

into a developing polyhedron is shown. (From the Carstens' Lab at

Queen's University, Canada)

2.10 3D structure of the first domain of Human Transferrin 19

2.11 The amino acids sequence of human transferrin gene 20

(MacGillivray et al., 1982; MacGillivray et al., 1983).

2.12 Protein N-glycosylation pathways in insect and mammalian cells 39

2.13 CMP-Neuraminic acid synthesis pathway 44

2.14 General strategy for humanization of glycoprotein produced by 46

lepidopteran cell-baculovirus expression system

2.15 Structure of a nucleotide sugar that can serve as a sugar 51

xviii

donor in a glycosyltransferase reaction. UDP, uridine diphosphate.

2.16 Transporters for sugar nucleotides, PAPS, and ATP are located 51

in the Golgi membranes of mammals, yeast, protozoa, and plants

2.17 Different hydrophobic ligands coupled to cross-linked agarose 55

matrices. (Amersham Bioscience, 1993)

2.18 The Hofmeister series on the effect of some anions and cations in 56

precipitating proteins

2.19 Relative effects of some salts on the molal surface tension of water. 56

2.20 Effect of pH on protein net charge 59

2.21 Ion exchanger types 60

3.1 Virus Titer Procedures – End Point Dilution 69

4.1 Sf-9 insect cells growth in monolayer culture at 3 different serum 88

concentrations

4.2 Sf-9 insect cells growth in monolayer culture for 2 types of media 90

4.3 Sf-9 insect cells growth in monolayer culture for 3 different initial 91

cell density

4.4 Sf-9 insect cells growth in monolayer culture at 3 different 93

subculturing conditions,

4.5 Sf-9 insect cells growth in monolayer culture at 3 different spent 94

medium carry over percentage

4.6 Growth curves of Sf9 monolayer culture in 25cm2 T-flask at 97

different seeding densities, SD

4.7 Growth curves of Sf9 suspension culture in 250ml shake flask at 98

different seeding densities

4.8 Growth curves of Sf9 suspension culture in 24-well plate at 99

different seeding densities, SD. Volume of medium was 0.5 ml

4.9 Growth curves of Sf9 suspension culture in 24-well plate at 99

different seeding densities, SD. Volume of medium was 1.0 ml

4.10 Growth rate constants of Sf9 in various cultivators and at 101

different seeding densities

4.11 Doubling time of Sf9 in various cultivators and at different seeding 102

densities

4.12 The effect of initial cell density on Sf-9 insect cells infected with 104

wild type AcMNPV viruses at MOI 10

xix

4.13 The effect of spent medium carry over on Sf-9 insect cells 105

infected with wild type AcMNPV viruses at MOI 10

4.14 The effect of MOI on Sf-9 insect cells infected in the stationary 107

phase with wild type AcMNPV Viruses

4.15 SDS PAGE analysis of rhTf expression 110

4.16 rhTf expression profiles at different MOIs 110

4.17 rhTf expression profiles at different seeding densities, SD 112

4.18 Surface plot of figure 4.11 112

4.19 rhTf expression profiles at different times of infection, TOI 114

4.20 Surface plot of figure 4.13 114

4.21 Comparison between uninfected (U), wild-type (WI), and 117

recombinant (R) virus-infected Sf9 cells

4.22 3D plot of Table 4.3 118

4.23 Infected cells appearance in medium A 124

4.24 rhTf concentration at different medium compositions based 124

on Plackett-Burman screening experiments

4.25 SDS-PAGE analysis of medium screening 125

4.26 Effect of nutrients on rhTf yield 127

4.27 Amino Acids in Human Transferrin (679 residues) 129

4.28 Observed and predicted experimental data for the optimization 131

of glutamine, glucose and lipid mixtures.

4.29 Glutamine (Gln) vs Glucose (Gluc) vs rhTf 134

4.30 Glutamine (Gln) vs Lipid Mixtures 1000x (Lip) vs rhTf 135

4.31 Glucose (Gluc) vs Lipid Mixtures 1000x (Lip) vs rhTf 136

4.32 Sf9 growth in controlled and optimized expression 139

4.33 Total protein and rhTf contents in controlled and optimized 139

expression

4.34 Total protein and rhTf production rates in controlled and 141

optimized expression

4.35 Glucose and lactate concentrations in controlled and optimized 141

expression

4.36 Glutamine concentrations in optimized expression 142

4.37 Lactate production and glucose uptake rate in controlled 144

and optimized expression

xx

4.38 SDS-PAGE gel for non optimized medium 145

4.39 SDS PAGE gel for optimized medium 146

4.40 Detection of �1,4-GalT by using chromatogram TLC 148

4.41 Time course of chromatogram of TLC. Layer 149

4.42 SDS-PAGE (9%) time course of �1,4-GalT production 150

4.43 Standard curve for the determination of �1,4-GalT activity from 152

the lectin binding assay values

4.44 Time course of �1,4-GalT enzyme accumulation in supernatants 153

detected using lectin binding assay

4.45 RP-HPLC chromatogram for UDP-Gal standard at different 156

concentrations

4.46 Standard curve for UDP-Gal 157

4.47 RP-HPLC chromatogram for native UDP-Gal sample with 158

spiking and without spiking.

4.48 RP-HPLC Chromatogram for the time course of native UDP-Gal 159

level upon infection

4.49 RP-HPLC chromatogram for time course of native UDP-Gal level 160

upon infection in 3D diagram

4.50 RP-HPLC Chromatogram for the time course of native UDP-Gal ……..161

level upon infection

4.51 RP-HPLC chromatogram for time course of native UDP-Gal 162

level upon infection in 3D diagram

4.52 Native UDP-Gal concentration in �M at normal and upon time of 162

infection.

4.53 Verification of UDP-Gal fractions from RP-HPLC analysis using 163

TLC. Layer

4.54 Gal�1�4GlcNAc linkage binding values at 450nm for the time 165

course upon coinfection between recombinant baculovirus hTf and

�1,4-GalT

4.55 Effect of the mammalian GalT on the rate of in vitro galactosylation 166

process

4.56 Gal�1�4GlcNAc linkage binding values at 450 nm for the different 168

levels of galactosylation process

4.57 Relationships among the three main elements in in vivo 169

xxi

galactosylation process

4.58 Steps and gradient elution of rhTf from column HIC. 171

4.59 HIC chromatogram for the optimization of step wise elution method. 173

4.60 HIC chromatogram for the optimization of elution flowrate 175

4.61 Line graph show the relationship between recovery percentage 177

and loading capacity

4.62 HIC chromatogram for the optimization of rhTf loading capacity 178

4.63 SDS-PAGE characterized the elution profile of rhTf 179

4.64 Binding capacity of two anion exchange matrix with Tris and 180

phosphate buffer used as equilibration buffer.

4.65 Binding capacity of Q-Sepharose with equilibration buffer of 180

different pH.

4.66 Binding capacity of Q-Sepharose with different concentration of 181

buffer Tris-HCl buffer, pH8.5 as equilibration buffer.

4.67 HIC chromatogram characterized the separation and elution profile 183

of sample

4.68 SDS-PAGE characterized the separated protein from column 183

phenyl sepharose 6 fast flow

4.69 Anion exchange chromatogram characterized the separation and 184

elution profile of sample of after HIC and after dialysis.

4.70 SDS-PAGE characterized the separated protein from column 184

Q-Sepharose.

4.71 SDS-PAGE characterized the sample pooled from each purification 185

step.

CHAPTER 1

INTRODUCTION

1.1 Research Background

Global manufacturing of biopharmaceuticals has increased significantly over

the last decade due to a number of reasons. Biopharmaceuticals offer several

advantages such as highly effective and potent action, fewer side effects and the

potential to actually cure diseases rather than merely treating the symptoms. These

advantages, combined with the increasing number of new diseases that can be treated

with biopharmaceuticals, are driving enhanced production of these drugs worldwide.

According to a report by PRNewswire, London dated November 30th

2004, the global

manufacturing capacity of biopharmaceuticals was around 2.27 million liters in

2004. This included the capacity held by both captive use and contract

manufacturers. It is expected to increase to 3.69 million liters in 2011 at a compound

annual growth rate (CAGR) of 7.2 per cent.

A variety of systems can be employed to produce biopharmaceuticals. The

most important ones are derived from bacteria and yeasts, but eukaryotic systems

become more and more important because the proteins produced are almost similar

to native proteins. In the recent past, the baculovirus insect cell system has attracted

wide attention as vectors for high level and faithful expression of a variety of

2

heterologous proteins. In many cases the products are chemically, antigenically,

immunologically and functionally similar, if not identical to their authentic

counterparts (Vlak, 1997). It has a wide application in the commercial exploration,

development and production of vaccines, therapeutics and diagnostics; drug

discovery research; as well as exploration and development of safer, more selective

and environmentally compatible biopesticides consistent with sustainable agriculture.

The potential production of therapeutic glycoproteins in baculovirus insect

cell system has stimulated the desire to monitor the glycosylation pattern of specific

insect-cell-produced glycoproteins and the glycosylation potential of insect cells in

general. The glycan moieties can significantly affect a protein’s stability, biological

activity, antigenicity, immunogenicity, solubility, cellular processing, secretion and

pharmacokinetic behaviour such as in vivo metabolic clearance rate (Takeuchi et al.,

1990, Takeuchi and Kobata, 1991, Munk et al., 1992). N-glycans found in

recombinant glycoproteins expressed by lepidopteran cells using the baculovirus

vector are predominantly high mannose type glycans and short truncated glycans

(paucimannose) with �1,3/ �1,6-linked fucose residue on its asparagines-bound N-

acetylglucosamine (GlcNAc) (Jarvis and Summers, 1989; Wathen et al., 1991;

Grabenhorst et al.,1993; Yeh et al., 1993; Manneberg et al., 1994; Ogonah et

al.,1995; Hsu et al.,1997; Opez et al., 1997). In contrast, mammalian cells usually

produce sialylated complex-type N-glycans. Generation of complete forms of

sialylated complex-type N-glycans in insect cells would increase the value of insect

cell derived products as vaccines, therapeutic and diagnostics.

1.2 Objective

The objectives of this work are as follows:

I. To determine the optimal parameters (culture conditions, transferases and

sugar nucleotides content) for the expression of complete recombinant

human glycoprotein.

3

II. To develop an optimal processing condition for the production of human

like glycoprotein in an artificial system by the manipulation of metabolic

engineering and process engineering approach.

1.3 Scopes

To optimize the expression level of recombinant human transferrin in insect cells

baculovirus expression system,

I. Expression and optimization of rhTf in Sf9 insect cells monolayer culture

using conventional method.

II. Expression and optimization of rhTf in Sf9 insect cells suspension culture

using experimental design. Variables studied were dominant medium

components that were screened earlier.

To develop a method for the expression of galactosylated recombinant hTf in insect

cells and optimize the expression of the galactosylated recombinant hTf, the

following scopes of study were investigated:

I. Monitoring of native UDP-Gal level at normal and upon baculovirus

infection

II. Evaluation of the quality of the glycoprotein obtained through baculovirus

coinfection study to coexpress _1,4-GalT and hTf (in vivo study) and the

artificial introduction of commercial GalT and UDP-Gal to secreted hTf (in

vitro study)

To develop an optimal processing condition (laboratory scale production and

purification), the following scopes of study were investigated::

I. Small scale production of rHtf using optimized medium

II. Optimization of separation process to achieve higher recovery of rhTf.

CHAPTER 2

LITERATURE REVIEW

2.1 Recombinant Protein Manufacturing Technologies

A variety of protein expression systems have been developed which are

currently focusing on the therapeutic purpose that is for human use. In

biopharmaceutical, the recombinant protein produced must achieving and

maintaining several important criteria such as efficacy, safety, immunogenic reaction

and blood circulation time.

Common bacterial expression system such as Escherichia coli (E.Coli) is the

simplest recombinant protein manufacturing process. However, the bacterial

fermentation is associated with several drawbacks. For example, the products are not

glycosylated like natural human proteins and are therefore likely to cause side effects

in the therapeutic use. In addition, bacterial proteins tend to have more sequence

translation errors and fold less consistently than glycosylated proteins, so biological

activity can be variable. Furthermore, bacteria cannot be used to manufacture very

large protein such as erythropoeitin or multiprotein assemblies such as antibodies.

Yeast expressions offer a simple production process with high yield, powerful

secretary pathways, and some limited post-translational modifications. However, the

5

glycosylation often has to adjust after purification to produce a closer match with

human glycosylation patterns.

Mammalian cell culture is a slow and expensive process. Chinese hamster

ovary (CHO), mouse-human hybridomas, myelomas, and human cell lines are some

examples. Currently, all commercial antibodies are produced via mammalian cell

culture. The main advantages of mammalian products are they can be engineered to

produce more than one protein simultaneously and the cells have near-correct human

glycosylation but do not maintain complete glycosylation under production lines.

The cell culture process takes 3 to 4 months and required a special serum-free,

chemically defined medium as well as temperatures maintained precisely in the range

of 37-42 0C. The acidity of the culture must be kept very close to neutral; this

process involves gentle bubbling of nitrogen or carbon dioxide gases through the

fermenter to make slight adjustments. Also, mammalian cells are more fragile than

bacterial or yeast cells and can shear or break open if subjected to rough mixing or

bubbling. Yields are lower than from either yeast or bacterial fermentation.

Transgenic animals are being studied as an alternative to traditional CHO cell

production processes. Transgenic animals provide a potentially less expensive

source of production for proteins compared to traditional cell culture systems.

Although the transgenic expression systems may solve the problem of protein

production yields and may lower the cost, they do not solve the problem of protein

glycosylation.

Recently, baculovirus-mediated expression in insect cells has become well-

established for the production of recombinant glycoproteins. Its frequent use arises

from the relative ease and speed with which a heterologous protein can be expressed

on the laboratory scale and the high chance of obtaining a biologically active protein.

In addition to Spodoptera frugiperda Sf-9 cells, which are probably the most widely

used in insect cell line, other mainly lepidopteran cell lines are exploited for protein

expression. Recombinant baculovirus is the usual vector for the expression of

foreign genes but stable transfection of – especially dipteran-insect cells presents an

interesting alternative. Insect cells can be grown on serum free media which is an

advantage in terms of cost as well as of biosafety. For large scale culture, conditions

6

have been developed which meet the special requirements of insect cells. With

regards to protein folding and post-translational processing, insect cells are second

only to mammalian cell lines. Evidence is presented that many processing events

known in mammalian systems do also occur in insects (Altmann et al., 1999).

However, on protein glycosylation, particularly N-glycosylation, which is insect,

differs in many respects from that in mammals.

2.2 Glycosylation

Glycosylation is the most common post-translational modifications made to

proteins by eukaryotic cells, and can significantly affect biological activity and is

particularly important for recombinant glycoproteins in human therapeutic

applications. Glycosylation is a process where oligosaccharides, or sugar chain are

covalently linked to proteins. The predominant sugars found on human

glycoproteins, include galactose, mannose, fucose, N-acetylgalactosamine (GalNAc),

N-acetylglucosamine (GlcNAc) and N-acetylneuraminic acid (the human form of

sialic acid). There are two types of glycosylation, which are N-glycosylation and O-

glycosylation (Figure 2.1).

Figure 2.1: (a) N-linked protein glycosylation. The N-linked amino acid consensus

sequence is Asn-any AA- Ser or Thr. The middle amino acid can not be proline

(Pro); (b) O-linked protein glycosylation. Most O-linked carbohydrate

covalent attachments to proteins involve a linkage between the monosaccharide N-

(a) (b)

7

Acetylgalactosamine and the amino acids serine or threonine (Adapted from

Altmann, 1996).

2.2.1 N-Linked Glycosylation

N-linked glycosylation is the oligosaccharide which link to the amino group

of asparagine (N) and have a core of Asp-GlcNAc-GlcNAc-Man-(Man)2 derived

from dolichol. The processing of N-glycans occurs co-translationally in the lumen of

the endoplasmic reticulum (ER) and continues in the Golgi apparatus. It can be bi-,

tri- and tetraantennary or if it is a poly-N-acetyl lactosamine type, it can be branched

or unbranched. There are three types of N-linked glycosylation which are complex,

hybrid, and high mannose (Figure 2.2 (a), (b) and (c)). The major distinguishing

feature of the complex class is the presence of sialic acid, whereas the hybrid class

contains no sialic acid. In contrast to the step-wise addition of sugar groups to the O-

linked class of glycoproteins, N-linked glycoprotein synthesis requires a lipid

intermediate that is dolichol phosphate. Dolichols are polyprenols (C80-C100)

containing 16 to 20 isoprene units, in which the terminal unit is saturated.

2.2.2 O-Linked Glycosylation

O-linked glycosylation most commonly links GalNAc to the hydroxyl group

of serine or threonine and occurs post-translationally in the Golgi apparatus. There

are no consensus sequence, no preformed intermediate and starts in trans Golgi. In

comparison with the N-linked glycosylation, most other O-linked glycans are highly

branched. Sulphates can be on Gal, GalNAc and GlcNAc and phosphate can be on

Man or Xyl.

2.3 Glycoprotein

Glycoproteins are the proteins that include complex carbohydrates as part of

their structure. The carbohydrate components of glycoproteins affect the

8

functionality of the molecule by determining protein folding, oligomer assembly and

secretion processes. Without the proper shape, the ability of the protein to interact

correctly with its receptor is affected, possibly affecting function. There are three

types of glycoproteins (Figure 2.2).

Figure 2.2: (a) High Mannose, (b) Complex and (c) Hybrid Structures of

carbohydrates on the 3 major classes of glycoprotein. (Adapted from Altmann, 1996)

2.4 Insect Cell Baculovirus Expression System

The baculovirus-insect cell expression system is a binary system consisting of

a recombinant baculovirus vector and its host, an insect cell. The virus delivers the

gene encoding a glycoprotein of interest to the cell, then the gene is expressed by

virus-encoded transcription factors and the protein is translated and glycosylated by

host cell machinery during viral infection.

(a) (b) (c)

9

2.4.1 Insect Cell Lines

Spodoptera Estigmene Mamestra Bombyx frugiperda acrea brassicae mori

Figure 2.3: A few insect species used for glycoprotein production (Adapted from

Tomiya et al., 2003)

Sf-9 and Sf-21 cells from the fall army worm Spodoptera frugiperda are the

most frequently used cell lines used in the heterologous expressions. However, quite

a number of other cell lines has been established, including cell lines stem from

Trichoplusia ni, i.e., TN-368 and BT1-TN-5B1-4, Bombyx mori (Bm-N), Mamestra

brassicae (e.g., MB0503) and Estigmene acrea (Ea). In general, stable cell lines are

usually obtained from embryonic cells and thus represent essentially undifferentiated

cells. An alternative strategy but there appears to be few data is to infect whole

larvae with recombinant baculovirus (Reis et al., 1992; Maeda et al., 1982; Korth et

al., 1993).

2.4.2 Baculoviruses

The most widely used vectors for the production of foreign proteins in insect

cells or larvae are recombinant baculoviruses, such as Autographa californica

multicapsid nucleopolyhedrovirus (AcMNPV) as shown in Figure 2.4, which infects

lepidopteran cells (David, 1994; Luckow, 1995). Baculoviruses are viral pathogens

that cause fatal disease in insects, mainly in members of the families Lepidoptera,

Diptera, Hymenoptera and Coleoptera. More than 600 baculovirus isolates have

been described, categorized in two subfamilies, (a) nucleopolyhedroviruses and (b)

granuloviruses (Murphy et al., 1995).

10

Baculoviruses are characterized by the presence of rod shape nucleocapsids,

that are enveloped singly or in bundles by a unit membrane (Figure 2.4 and 2.5). The

virus particles usually embedded into large protein capsules or occlusion bodies

(OB), also called polyhedra c.q. granula. These OBs, 0.1-10 �m in diameter, provide

protection of the virus particle and enhance the persistence of the virus in the

environment. The occlusion body-derived virus particles (OBV) are the infectious

entities of the OBs.

The major constituent of OBs is a single protein (polyhedron c.q granulin)

with a subunit molecular weight of approximately 30 kDa. The amino acid sequence

is highly preserved among baculoviruses (Vlak and Rohrmann, 1985).

Baculoviruses contain a double stranded, circular DNA molecule as genetic element.

This DNA varies in size between 100 and 200 kbp and is able to code for more than

70 average-sized proteins. Physical maps of various baculovirus DNAs have been

established, the most detailed one being of the prototype baculovirus AcMNPV,

whose genome has been entirely sequenced (Ayres et al., 1994). About forty

functional genes have been mapped on the AcMNPV genome, including polyhedrin.

Approximately thirty of these have also been sequenced transcriptionally and

analyzed (Blissard and Rohrmann, 1990; Kool and Vlak, 1993).

Figure 2.4: Autographa californica multicapsid nucleopolyhedrovirus

(AcMNPV). This high magnification electron micrograph shows a negatively-stained

baculovirus virion. Note the asymetric capsid structure and the presence of an

envelope with surface projections (peplomers). (From the Carstens' Lab at Queen's

University, Canada)

11

Figure 2.5: A) Baculovirus particles, or polyhedra; B) Cross-section of a

polyhedron; C) Diagram of polyhedron cross-section. (Electron micrographs (A&B)

by Jean Adams, graphic (C) by V. D'Amico)

2.4.2.1 Baculoviruses Replication

2.4.2.1.1 In Vivo Replication

Viruses are unable to reproduce without a host because they are obligate

parasites. Baculoviruses are no exception. The cells of the host's body are taken

over by the genetic message carried within each virion, and forced to produce more

virus particles until the cell, and ultimately the insect, dies. Most baculoviruses

cause the host insect to die in a way that will maximize the chance that other insects

will come in contact with the virus and become infected in turn. Figure 2.6 shows

the infection by baculovirus begins when an insect eats virus particles on a plant -

perhaps from a sprayed treatment. The infected insect dies and "melts" or falls apart

on foliage, releasing more virus. This additional infective material can infect more

insects, continuing the cycle.

12

Figure 2.6: In vivo baculovirus infection and replication cycle. (Adapted from Vlak,

1997)

2.4.2.1.2 In Vitro Replication

Baculovirus infection starts when a susceptible insect larva ingests

baculovirus occlusion bodies (Figure 2.7). The midgut lumen of lepidopteran larvae

constitutes a highly alkaline environment in which OBs dissolve and the occlusion

derived virions are released into the gut lumen. These virions pass through the

peritrophic membrane and fuse with the microvillar membrane of the midgut

epithelial cells whereafter they are transported to the nucleus, initiating the first

replication cycle. Baculoviruses have a biphasic replication cycle, in which two

genetically identical, but phenotypically distinct virus types are formed.

The newly formed budded viruses (BVs) are initially released by budding

through the plasma membrane of the infected cell. The insect tracheal system and

the hemolymph play a major role in the transport of the BVs to other organs and

tissues (Volkman, 1997; Barrett et al., 1998). Budded virions differ in several

aspects from ODVs (Figure 2.8) which are formed later in infection. (In cell culture

BVs are 1000-fold more infectious than ODVs.) Budded virions are responsible for

the systemic infection; ODVs facilitate viral spread from one individual insect to

13

others. Budded virions enter the cell by endocytosis, followed by the fusion of the

viral envelope and the endosome membrane. The fusion process is mediated by a

virus encoded essential glycoprotein, gp64, which is exclusively found in BVs

(Blissard, 1996). The ODVs are not released by budding, but acquire an envelope

inside the nucleus, followed by occlusion in polyhedra. Finally, the infected cell

ruptures and the lyses of both the nuclear and cellular membranes allow the release

of the newly formed, mature polyhedra. The polyhedra are surrounded by an

envelope composed of carbohydrates and specific proteins (Zuidema et al., 1989).

Baculovirus diseases are primarily diseases of the larval stages, and the progression

and signs of disease depend on several factors including the instars initially infected,

infection dose, nutrition, temperature, degree of compatibility of the virus with its

host, and the physical characteristics of the larva.

In typical nucleocapsid nuclear polyhedrovirus (NPV) infections, there are

very limited signs of disease during the first 3 days post infection. At about the

fourth day of the infection, larvae show reduced motor functions. They also respond

more slowly to tactile stimuli than healthy larvae. Their feeding begins to slow and

virtually ceases by day 6 or 7. At day 4 or 5 the larva will begin to appear swollen,

the cuticle will take on a pale creamy coloration. This is due to the presence of

polyhedra accumulating in epidermal and fat body cell nuclei. The hemolymph of

infected larvae at this stage is cloudy owing to the circulation of large numbers of

infected hemocytes and polyhedra released into the hemolymph as a result of lysis of

cells in various tissues during advanced stages of disease. Following this, larvae will

die within one or two days. Larvae of many lepidopteran species will crawl up to the

top of the plant on which they were feeding, and then die. After death, the larvae

become black, lose their turgor and become flaccid. The cuticle ruptures, releasing

billions of polyhedra. Figure 2.9 showed a typical infected Sf-9 cells photo

containing the presence of polyhedra.

14

Figure 2.7: In vitro baculovirus infection and replication. (A) Ingestion of

polyhedra and solubilization by digestive juices in the insect gut. (B) Fusion of the

viral envelope of the released virus with the plasma membrane of a midgut cell. (C)

Entry of the nucleocapsid into the nucleus. (D) Formation of virogenic stroma where

virus replication and assembly of progeny nucleocapsids occurs. (H) Departure of

nucleocapsids from the nucleus and formation of non-occluded virus particles (NOV)

by acquisition of an envelope from the nuclear (I) or cellular (J) membrane by

budding. (K) Systemic infection of cells from other tissues by adsorption

endocytosis. (E) Envelopment of single or multiple nucleocapsids in membrane de

novo synthesized in the nucleus. (F) Occlusion of singly and multiply enveloped

virus particles into polyhedra. (L) Formation of cytoplasmic and nuclear inclusions

(fibrillar structures) with unknown function. (G) Release of polyhedra from deceased

insect larvae. (Adapated from Vlak, 1997)

15

Figure 2.8: Structural compositions of the two baculovirus phenotypes, budded virus

(BV), and the occlusion derived virus (ODV). (Adapted from Blissard, 1996).

Proteins common to both virus types are indicated in the middle of the Figure 2.6.

Proteins specific to either BV or ODV are indicated on the left and right respectively.

The polar nature of the baculovirus capsid is indicated in the diagram with the claw-

like structure at the bottom and the ring-like structure at the top of the capsid. The

possible location of p74 is indicated by a dashed line. Lipid composition of the BV

and ODV envelopes derived from AcMNPV infected Sf-9 cells (Braunagel and

Summers, 1994) are indicated. (LPC, lysophosphatidylcholine; SPH, sphingomyelin;

PC, phospahetidylcholine; PI, phosphatidylinositol; PS, phosphatidylserine; PE,

phosphatidyl-ethonalamine)

5

Figure 2.9: (a) A typical infected Sf-9 cells showing the presence of polyhedra is indicated by the arrow (Steven Howard); (b) Electron

micrograph of AcMNPV infected Sf-9 Cell (Greg V. Williams); (c) A portion of the nucleus containing enveloped virions in the process of being

occluded into a developing polyhedron is shown. (From the Carstens' Lab at Queen's University, Canada)

(a)

(a) (b)

(c)

16

17

In Figure 2.9 (b), the polyhedra (P) containing occluded virus are visible in

the nucleus of an infected Sf-9 cell at 36 hour post infection. One occlusion body (*)

is sectioned in a plane where no occluded virus is evident. The occlusion body calyx

(C) is visible. Calyx precursors (CP) are present both in association with p10 fibrous

bodies (F) and free in the nucleoplasm. A calyx precursor is seen attaching to an

occlusion body. Fibrous bodies are visible in both the nucleoplasm (F) and the

cytoplasm (Fc). The nuclear membrane (N) is indicated. Short open-ended

membrane profiles (M) are present near the nuclear periphery as are the remnants of

the virogenic stroma (vs) assembled nucleocapsids are seen in association with the

membrane profiles in the process of being enveloped to form PDV(�)

2.5 Advantages of Baculovirus Expression System

Since 1983, when BEVS technology was introduced, the baculovirus system

has become one of the most versatile and powerful eukaryotic vector systems for

recombinant protein expression (Smith et al., 1983). More than 600 recombinant

genes have been expressed in baculoviruses to date. Since 1985, when the first

protein (IL-2) was produced in large scale from a recombinant baculovirus, use of

BEVS has increased dramatically (Smith et al., 1983). Baculoviruses offer the

following advantages over other expression vector systems.

(a) Safety: Baculoviruses are essentially nonpathogenic to mammals and plants

(Ignoffo, 1975). They have a restricted host range, which often is limited to

specific invertebrate species. Because the insect cell lines are not transformed

by pathogenic or infectious viruses, they can be cared for under minimal

containment conditions. Helper cell lines or helper viruses are not required

because the baculovirus genome contains all the genetic information.

(b) Ease of Scale Up: Baculoviruses have been reproducibly scaled up for the

production of biologically active recombinant products.

18

(c) High Levels of Recombinant Gene Expression: In many cases, the

recombinant proteins are soluble and easily recovered from infected cells late in

infection when host protein synthesis is diminished.

(d) Accuracy: Baculoviruses can be propagated in insect hosts which post-

translationally modify peptides in a manner similar to that of mammalian cells.

(e) Use of Cell Lines Ideal for Suspension Culture: AcMNPV is usually

propagated in cell lines derived from the fall armyworm Spodoptera frugiperda

or from the cabbage looper Trichoplusia ni. Cell lines are available that grow

well in suspension cultures, allowing the production of recombinant proteins in

large-scale bioreactors.

(f) Very High Expression of Recombinant Proteins: In many cases, the

recombinant proteins produced are antigenically, immunogenically and

functionally similar to their native counterparts

2.6 Model Glycoprotein

2.6.1 Native Human Transferrin (nhTf)

Human serum transferrin (HST) belongs to the transferrin family of metal-

binding proteins that transport iron and provide bacteriostatic functions in a wide

variety of physiological fluids in vertebrates (Aisen and Listowsky, 1980; Huebers

and Finch, 1987). It is a single-chain glycoprotein of 679 amino acids containing two

asparagine-linked glycan chains each capped with a terminal sialic acid residue, with

a glycosylation-dependent molecular mass in the range of 76–81 kDa (MacGillivray

et al., 1982; MacGillivray et al., 1983). Transferrin is the product of an ancient

intragenic duplication that led to two homologous domains, each of which binds 1

ion of ferric iron (Figure 2.10) with both sites of glycosylation in the carboxyl-

terminal domain at positions 413 and 611 (MacGillivray et al., 1983). The two

domains are comprised of residues 1-336 and 337-678, in which 40% of the residues

19

are identical when aligned by inserting gaps at appropriate positions (MacGillivray et

al., 1982).

Figure 2.10: 3D structure of the first domain of Human Transferrin. Adapted from

NCBI Chemical Data (ASN)

The view that transferrin consists of two homologous domains, each

associated with one metal binding site is supported by the demonstration of internal

homology in a partial sequence for human transferrin (MacGillivray and Brew, 1975)

and by the production of fragments of various transferrins by partial proteolysis that

have approximately half the molecular weight of the native protein and single sites

for Fe3+

binding (Lineback-Zins and Brew, 1980). The functional significance of the

presence of two domains with separate Fe-binding sites is uncertain. Although the

two sites have some distinguished physical properties (Aisen and Listowsky, 1980),

present evidence indicates that in human transferrin, there is no difference in the in

vivo behavior of the sites with respect to iron uptake and delivery to cells (Huebers et

al., 1981). The amino acids sequence of human transferrin gene is given as figure

2.11.

20

1- V P D K T V R W C A V S E H E A T K C Q S F R D H M K S V I P S D G P S V A C V K

K A S Y L D C I R A I A A N E A D A V T L D A G L V Y D A Y L A P N N L K P V V A E F Y

G S K E D P Q T F Y Y A V A V -100- V K K D S G F Q M N Q L R G K K S C H T G L G R

S A G W N I P I G L L Y C D L P E P R K P L E K A V A N F F S G S C A P C A D G T D F P

Q L C Q L C P G C G C S T L N Q Y F G Y S G A F K C L K D G A G -200- D V A F V K H

S T I F E N L A N K A D R D Q Y E L L C L D N T R K P V D E Y K D C H L A Q V P S H T

V V A R S M G G K E D L I W E L L N Q A Q E H F G K D K S K E F Q L F S S P H G K D L

L F K D S A H -300- G F L K V P P R M D A K M Y L G Y E Y V T A I R N L R E G T C P E

A P T D E C K P V K W C A L S H H E R L K C D E W S V N S V G K I E C V S A E T T E D

C I A K I M N G E A D A M S L D G G F V Y I A G -400- K C G L V P V L A E N Y N K S D

N C E D T P E A G Y F A V A V V K K S A S D L T W D N L K G K K S C H T A V G R T A G

W N I P M G L L Y N K I N H C R F D E F F S E G C A P G S K K D S S L C K L C M G -

500- S G L N L C E P N N K E G Y Y G Y T G A F R C L V E K G D V A F V K H Q T V P Q

N T G G K N P D P W A K N L N E K D Y E L L C L D G T R K P V E E Y A N C H L A R A P

N H A V V T R K D K E A C V H K I -600- L R Q Q Q H L F G S N V T D C S G N F C L F R

S E T K D L L F R D D T V C L A K L H D R N T Y E K Y L G E E Y V K A V G N L R K C S

T S S L L E A C T F R R P -679

Figure 2.11: The amino acids sequence of human transferrin gene (MacGillivray et

al., 1982; MacGillivray et al., 1983).

Transferrin carries iron from the intestine, reticuloendothelial system, and

liver parenchymal cells to all proliferating cells in the body. It carries iron into cells

by receptor-mediated endocytosis (Fielding and Speyer, 1974; Karin and Mintz,

1981). One ninth of hTf have iron bound at both sites, four ninth have iron bound at

one site and other four ninth have no iron bound. Iron is dissociated from transferrin

in a nonlysosomal acidic compartment of the cell. Provision of intracellular iron for

synthesis of ribonucleotide reductase, an enzyme that catalyzes the first step leading

to DNA synthesis, is required for cell division. After dissociation of iron, transferrin

and its receptor return undegraded to the extracellular environment and the cell

membrane, respectively. Human transferrin cDNA has been isolated, its

characterization and the chromosomal localization of its gene have also been done.

Transferrin isoform pattern has a number of diagnostic applications such as diagnosis

of alcohol abuse, diagnosis of inherited carbohydrate-deficient glycoprotein (CDG)

21

syndrome and the use of genetically-determined polymorphisms for forensic

purposes.

Previous works have shown that the human transferrin glycoforms are

comprised of species having terminally sialylated bi-, tri-, and, tetrantennary

oligosaccharides (Leger et al., 1989; Fu and van Halbeek, 1992). The most

pronounced glycoform includes biantennary oligosaccharides located at both

asparagine positions , although changes in physiological conditions can affect the N-

glycan pattern observed in the host (Montreuil et al., 1997).

2.6.2 Recombinant Human Transferrin (rhTf)

Recently, there has been much interest in expressing recombinant human

serum transferrin (HST) and mutants thereof for structural and functional studies (Ali

et al., 1996). There have also been many reports on the expression of recombinant

human transferrin in BEVS. Majority of their concerns are on the posttranslational

processing of protein particularly glycosylation (Ailor et al., 2000; Tomiya et al.,

2003) and production of biologically and functionally active (Ali et al., 1996)

recombinant human transferrin. Ali et al., (1996) reported amino acid sequence that

matches the native human transferrin and is identical to the correctly processed

protein as predicted from the DNA sequence of the cloned gene used for expression.

Unlike mammalian cells, however, the oligosaccharide processing pathway in insect

cells is not well characterized (Marz et al., 1995; Altmann et al., 1999).

Experimental evidence suggests that glycoprotein produced in insect cells

possess N-linked oligosaccharides are principally comprised of high mannose and

truncated low mannose (paucimannocidic) structures (Butters and Hughes, 1981;

Hsieh and Robbins, 1984; Kuroda et al., 1990; Chen and Bahl, 1991; Kulakosky et

al., 1998). Ailor et al. (2000) reported that the attached oligosaccharides of human

transferrin expressed in Trichoplusia ni included high mannose, paucimmanosidic,

and hybrid structures with over 50% of these structures linked to the Asn-linked N-

acetylglucosamine. Neither sialic acid nor galactose was detected on any of the N-

glycans. Carbohydrate analysis revealed a small fraction of Gal in oligosaccharides

22

obtained from N-glycans of human lactoferin was expressed in Spodoptera

frugiperda (Sf9) (Wolff et al., 1996).

2.7 Insect Cell Culture Medium

2.7.1 Protein Hydrolysates (Peptones)

The main disadvantages of serum and/or serum components as supplements

for cell growth are their high cost and possible contamination risk (bovine viral

diarrhea; rednose, infectious bovine rhinotracheitis; parainfluenza 3; foot and mouth

disease; prion; blue tongue disease; and mycoplasma). Development of serum free

medium was started back in the ‘70s with the use of animal derived protein

hydrolysates (peptones), which were produced with animal derived enzymes, and/or

animal or human derived purified proteins in serum free medium.

Hydrolysates or peptones are complex mixtures of oligopeptides,

polypeptides and amino acids that are produced by enzymatic or chemical digestion

of casein, albumin, plant or animal tissues or yeast cells. Hydrolysates are being

widely used to prepare insect cells culture medium and feeds. Lactalbumin

hydrolysates are used as one of the peptides and amino acids sources. The most

widely used hydrolysates in insect cells culture is without doubt yeastolates. It is not

known which component of yeastolate is responsible for its growth enhancing effect

(Wu and Lee, 1998) and no new report has been found until this thesis is written.

Protein hydrolysates for pharmaceutical applications have at least two

functions. Peptides in the hydrolysate are used directly as an amino acid source

(replacement of free amino acids), and/or indirectly as a stimulator of growth and/or

production (serum replacement). Although this is a step in the right direction, it is not

sufficient because the potential risk of introducing adventitious agents is still present.

23

2.7.2 Carbohydrates

Glucose is now considered the most important carbohydrate for insect cell

growth (Bhatia et al., 1997). High levels of glucose can result in high levels of

lactate through glycolysis. Lactate accumulation can reduce the pH throughout the

culture, and low pH can be detrimental to cell viability and productivity (Hassell et

al., 1991).

2.7.3 Amino Acids

Insect cells can utilize amino acids for both biosynthesis and energy. Amino

acids such as glutamine, glutamate, aspartate, serine, arginine, asparagine, and

methionine are used for energy production (Drews et al., 1995). Cysteine, Tyronine,

Serine, Arginine, Valine, Lysine, Tyrosine and Methionine are required for optimal

growth and therefore are included in the optimization (Ferrance et al., 1993). It has

been assumed that the majority of amino acids are not synthesized by insect cells

(Bhatia et al., 1997). Supplementation of methionine and tyrosine was found to

retard cell death in Sf9 culture (Mendonca et al., 1999). Cystine was the only amino

acid to be depleted in high density culture of Sf9 cells (Vaughn and Fan, 1997).

Glutamine is an indispensable amino acid for optimal growth of most cell and

tissue cultures. High levels of glutamine in culture media can cause ammonia to

accumulate. The ammonia results from either metabolic hydrolysis to glutamic acid

or from spontaneous deamidation as a result of medium storage. Ammonia has also

been shown to affect glycosylation of a recombinant protein (Yang and Butler,

2000). Enriched oxygen environment and an increased glutamine concentration

(9.9mM) could support increasing volumetric production of two recombinant

proteins (�-Gal and SEAP) with increasing infection densities (Taticek and Shuler,

1997). On the other hand, one time addition of a combination of yeastolate ultra

filtrate and an amino acids mixture could have the same effect on protein production

as medium replacement (Bedard et al., 1994).

24

2.7.4 Lipids

Cholesterol was proven to be essential for successful expression of proteins

using BEV system (Gilbert et al., 1996). However, it was not required for cells

growth. Different mixtures were suggested, including natural mixtures such as olive

oil (Liu et al., 1995). Lipids (fatty acids) at a concentration range of 10–100 �g/L are

essential components included in most serum-free cell culture medium formulations

(Shen et al., 2004).

Gas chromatography coupled with mass spectrometry (GC/MS) has been

extensively used for the quantitation of lipids through fatty acid analysis (Christie,

1989). The fatty acid concentration in Sf-900 II was also examined and the fatty acid

profile was similar to that found in the IPL 41 medium (Shen et al., 2004). The lipid

concentrations in serum-free insect cell culture media were much higher than that

found in mammalian cell culture media. These results were consistent with the lipid

concentrations usually reported for insect cell culture media (Inlow et al., 1989). The

lipids added into insect cell culture medium usually also include �-tocopherol acetate

and cholesterol.

2.7.5 Albumin

Albumin is the most important protein of all animal sera, with various

functions showing why this molecule is included in many serum free medium. In

principle, albumin assures transport functions for many different groups of

substances, such as lipids, hormones, some amino acids, peptides, and globulins, as

well as heavy metals (Wu and Lee, 1998). Transport functions are advantageous

when highly purified albumin is used, because albumin can be used to solubilize

fatty acids and hormones in cell culture medium and can transport these substances

to the cultured cells. In addition, toxic substances such as heavy metals, endotoxins,

or free fatty acids can be detoxified by adsorption to albumin.

25

2.7.6 Serum Free Medium (SFM)

The chief advantage of using SFM for culture of insect cells is that

purification protocols are simplified because contaminating proteins are reduced.

Following that, the analysis of product becomes much easier and accurate (Wu et al.,

1998). One disadvantage is the possible proteolytic degradation of proteins when

concentrating product (Yamaji et al., 1999). Sf-900 II SFM (GIBCOTM

) is

specifically designed for large-scale production of recombinant proteins. They

contain optimized concentrations of amino acids, carbohydrates, vitamins, and lipids

that reduce or eliminate the effect of rate-limiting nutritional restrictions or

deficiencies. The optimized formulations offer the following advantages over sera:

• Eliminate the need for costly fetal bovine serum and other animal serum

supplements

• Increase cell and product yields

• Eliminate issues related to serum sensitivity (eg. mad cow disease)

• Purification is simplified due to reduction of contaminating protei

However serum free cultures may be more sensitive to agitation than the

serum supplemented culture. A high cell growth but decreased recombinant protein

production was observed in serum free culture (Caron et al., 1990)

2.8 Optimization of Protein Expression in BEVS

2.8.1 Physical Factors that Ensure Success of Expression

Success with the baculovirus expression system is dependent on the ability to

infect cells efficiently with AcMNPV, thus obtaining maximum virus replication and

hence optimum production of the desired protein (King and Possee,

1992).Recombinant proteins have been produced as fusion or nonfusion proteins at

levels ranging from 1-500 mg/L (Luckow and Summers, 1988). The polyhedrin

protein expression depends on the use of log phase Sf9 cells which are at least 97%

26

viable, a multiplicity of infection (MOI) of at least 5-10, and high quality medium

and fetal bovine serum.

Mock-infected and wild-type virus-infected cells are essentials in each

experiment as controls to ensure infection procedures are effective, as well as having

useful controls for DNA, and protein gels. Double checking by means of light

microscopy prior to virus infection are equally important to confirm that all is well,

i.e. that cells have attached well and have formed an even monolayer that is not too

sparse, overcrowded or clumped. Insufficient amount of cells will result in

insufficient amount of sample for analysis. On the other hand, clumped cells will

result in inefficient viral infection.

Table 2.1 gives approximate seeding densities for typical vessel sizes.

Infection at these densities will usually give high virus titers (>1.0 x 108 PFU/ml);

however, for maximum levels of recombinant proteins, higher densities (>3.0 x 106

cells/ml) may be desirable (Summers and Smith, 1988).

Table 2.1: Seeding densities for typical vessel sizes (O’Reilly et al., 1994)

Some cells are infected later than others and as a result, reach maximum

expression at a later time. Therefore, it is important that sufficient virus is used to

ensure synchronous infection of all insect cells in a culture. Some proteins may not

be stable in virus-infected cells. If these proteins are harvested too late, considerable

amounts may be lost (King and Possee, 1992). It is therefore important to perform an

27

experiment to determine the optimum time for harvesting recombinant proteins; and

not to rely on data published by others.

2.8.2 Optimization of Recombinant Baculovirus Stock

Optimization of recombinant baculovirus stock generally refers to generation

of a pure virus stock. This involves the preparation of a stock starting from a single

infectious unit. Virus particles in solution are distributed according to the Poisson

distribution. According to Poisson distribution, the proportion (p) of cultures

receiving a particular number of infectious units (r) is given by the equation,

p = �re

-�/r! …2.1

where � is the mean concentration of the infectious units in the diluted solution.

Therefore the proportion of culture receiving no infectious units is

p = e-�

(r = 0) … 2.2

and the proportion of culture receiving one or more infectious units is

p = 1-e-�

(r >= 1) …2.3

The proportion of culture receiving only one infectious unit is

p = �e-�

(r = 1) …2.4

The ratio of culture receiving only one infectious unit to the total number of infected

culture is

(r=1) / (r>=1) = �e-�

/(1-e-�

) …2.5

If we want to be 95% confident that the infected cultures contains only a single

infectious unit, then

�e-�

/(1-e-�

) = 0.95 …2.6

Solving this will reveal that �=0.101. Therefore the proportion of uninfected cultures

e-�=0.90. Therefore, to be at least 95% confident that the infected cultures are

generated from a single infectious unit, the virus stock has to be diluted until only

10% or less of the total cultures infected. Values at different levels of confidence can

also be calculated and generated as a guideline. In end point dilution method, the aim

is to dilute the virus such that, if multiple cultures are exposed to the diluted

28

inoculum, any cultures that become infected will have received only a single

infectious unit (Reed and Muench, 1938).

2.8.3 Medium Optimization

Medium can be optimized by partial medium replenishment, as spent medium

may contain secreted growth promoting factors with a positive effect on protein

production (Jesionowski and Ataai, 1997).

A feeding strategy as an alternative to medium replacement is the

supplementation of essential nutrients either at time of infection or several times

during the post infection period. Reuveny et al., (1993) have shown that selected

nutrient addition can increase recombinant protein production, even after medium

replacement. Their supplement contained glucose, L-glutamine, and yeastolate.

Glucose and lactate were measured by YSI analyzer (YSI Inc.). Medium

concentrations of glucose and lactate were also determined using the Analox GM7

analyzer. The concentration of ammonia was determined spectrophotometrically

using an enzymatic reaction (Sigma). Concentrations of amino acids and

carbohydrates were determined using the Dionex-AAA method.

It was found that the spent medium collected from a culture close to the

stationary growth phase could provide full support for insect cell growth through

another batch culture after fortification with suitable nutrients (yeastolate, glucose

and glutamine) and a small fraction (15–20%) of fresh medium (Wu et al., 1998).

Glucose and glutamine feeding sustained culture viability for 36 hours post

infection (hpi). It can be seen that glucose is required for a productive infection, and

that glucose feeding by itself is sufficient to increase up to 10 times the yield of

recombinant protein (Palomares et al., 2001). It was also shown that the productivity

of cells that had been maintained in the absence of glucose for over 18 h can be

“rescued” if glucose was fed at the time of virus addition. Glucose feeding has

advantages over medium replacement. On one hand, expensive culture medium is

economized. On the other, medium replacement requires cell separation prior to

29

infection, which can be impractical and expensive at large scale. A high cell density

culture (18 × 106

cells/ml) was obtained using a glucose concentration of 10 g l−1

(Drews et al., 1995).

Glutamine feeding further increased recombinant protein yield, although its

effect was not as pronounced as glucose feeding (Palomares et al., 2004). Neerman

and Wagner, (1996) have shown that up to 15% of glutamine and 59% of glucose

consumed by uninfected insect cells are metabolized to CO2.

It was concluded that protein production in a high-cell-density culture was

limited by nutrient depletion in the culture medium, and hence the nutritional

capacity of the medium could be determined as the viable cell density multiply the

integral at which the maximum product yield was attained. Production of a

recombinant protein in a culture with medium replacement at the time of infection

can be optimized if the cells were infected at a high MOI (1 pfu/cell) and at a cell

density such that the viable cell density time integral reached the nutritional capacity

just as the protein production was completed (Yamaji et al., 1999). �

A parallel line of research could be the use of factorial experiments for the

design of new media or the screening of supplements. Factorial design is a unique

way to detect interactions between the parameters tested (Montgomery and Runger

1999) and it can greatly reduce the number of experimental runs needed. Thus its use

can result in great time and cost savings. In insect cell culture, a fractional factorial

experiment was employed for the screening of several hydrolysates, and subsequent

full factorial experiment for the optimization of the selected hydrolysate (yeastolate

and Primatone RL) concentration (Ikonomou et al., 2001).

2.8 Design, Analysis and Optimization of Experiments

2.9.1 Design of Experiments

The design and analysis of experiments involves a broad range of statistical

as well as mathematical methods. The main purpose of statistical design and analysis

30

of experiments is to gain better understanding of a process through some statistical

approaches. This will help scientists to systematically plan and conduct their

experiments. This section will review some of these methods.

2.9.1.1 Factorial Experiments in Completely Randomized Designs

A complete factorial experiment includes all possible factor level

combinations in the experimental design. One of the most straightforward designs to

implement is the completely randomized design. Randomization affords protection

from bias by tending to average the bias effects over all levels of factors in the

experiment (Haaland, 1989). When comparisons are made among levels of a factor,

the bias effects will tend to cancel out and the true factor effects will remain. Again,

randomization is not a guarantee of bias-free comparisons, but it is certainly an

inexpensive assurance.

2.9.1.2 Interactions

An interaction exists among two or more factors if the effect of one factor on

a response depends on the levels of other factors (Haaland, 1989). The presence of

interactions requires that factors be evaluated jointly rather than individually. It

should be clear that one must design experiments to measure interactions. Failing to

do so can lead to misleading, even incorrect conclusions. Factorial experiments

enable all joint factor effects to be estimated. If one does not have any evidence that

interaction effects are absent, factorial experiments should be seriously considered.

2.9.1.3 Coded Variables

In general, the units of parameters (a, b, c etc.) involved in an experiment

differ from each other. Therefore, regression analysis can not be performed on the

physical (dimensional) parameters themselves (Montgomery, 1996). Instead,

normalization method is applied to parameters a, b, and c before performing a

31

regression analysis. The normalized variables are called coded variables. In other

words, instead of using values of a, b, and c directly in the regression analysis, coded

variables, x1, x

2, and x

3 are used as the independent variables in the regression

analysis. A coded variable must be defined for each of the actual variables such as:

x1

is defined for parameter a

x2

is defined for parameter b

x3

is defined for parameter c

Each of the coded variables is forced to range from -1 to 1, so that they all

affect response y more evenly, and so the units of parameters a, b, c, etc. are

irrelevant. To convert a parameter to its coded variable x1, the following formula is

applied to each value of a in the data set:

…2.7

where amid value

is the middle value of ‘a’ in the data set,

…2.8

and arange

is the range of parameter a, i.e. from its minimum to its maximum,

…2.9

Regression analysis is then performed on y as a function of x1, x

2, and x

3. The slopes

with respect to these coded variables are used to determine the direction of steepest

ascent. When using coded variables, the vector of steepest ascent must then be

converted back to the original, physical (uncoded) parameters, using the inverse of

the above equations so that the optimization process can be performed on physical

variables (Mason et al., 2003)

2.9.1.4 Factor Levels Combinations

A straightforward way to list all unique combinations of a 2 level factorial

design is as follows (Montgomery, 1996; Montgomery, 2001);

32

1. Designate one level of each factor as -1 (low level value) and the other level as +1

(high level value)

2. Lay out table with column headings for each of the factors A, B, C… K.

3. Let n=2k, where k is the number of factors, and n is the number of possible

combinations.

4. Set the first n/2 of the levels for factor A equal to -1 and the last n/2 equal to +1.

Set the first n/4 levels of factor B equal to -1, the next n/4 equal to +1, the next n/4

equal to -1, and the last n/4 equal to +1. Set the first n/8 of the levels for factor C

equal to -1, the next n/8 equal to +1, etc. Continue in this fashion until the last

column (for factor K) has alternating -1 and +1 signs (Table 2.2).

Table 2.2: Example of a 4-Factor, 2-level Full Factorial Experiment

2.9.1.5 Fractional Factorial Experiments

Fractional factorial experiments are alternatives to complete factorial

experiments. Whenever fractional factorial experiments are conducted, some effects

are confounded with one another. The goal in the design of fractional factorial

experiments is to ensure that the effects of primary interest are either unconfounded

with other effects or if that is not possible, confounded with effects that are not likely

to have appreciable magnitudes (Haaland, 1989).

33

2.9.1.6 Screening Experiments

Screening experiments are conducted when a large number of factors are to

be investigated but limited resources mandate that only a few test runs be conducted.

Screening experiments are conducted to identify a small number of dominant factors,

often with the intent to conduct a more extensive investigation involving only the

dominant factors (Montgomery, 2001).

A special class of two-level fractional factorial experiments that is widely

used in screening experiments was proposed by Plackett and Burman. These

experiments have resolution III when conducted in completely randomized designs

and are often referred to as Plackett-Burman designs (Kalil et al., 1999). The designs

discussed by Plackett and Burman are available for experiments that have the

number of test runs equal to a multiple of four. Table 2.3 shows an example of a

screening design.

Table 2.3: Example of 12-Run, 11-Factor, 2-Level, Screening Design (Not

Randomized)

The results obtained from a full factorial design, fractional factorial design,

and screening design can further be analyzed using analysis of variance to determine

the significance of each factor effect and interaction effect. Regression analysis can

be used to find the coefficient for each factor and its interaction. Thus an equation to

relate all the factors can be made. Eventually, response surface analysis can be

34

employed to study the interaction between factors and improve the quality of a

process without having to do so much of trials and errors.

2.9.2 Analysis of Experiments

2.9.2.1 Correlation

A linear correlation coefficient is used to determine if there is a trend between

measured output, y and controlled parameter, x. If there is a trend, regression

analysis is used to find an equation for y as a function of x which provides the best fit

to the data. The linear correlation coefficient, rxy

, is defined as

…2.10

The mean value of x and the mean value of y are defined as

…2.11

By definition, rxy

must always lie between -1 and 1, i.e.

…2.12

The linear correlation coefficient is always nondimensional, regardless of the

dimensions of x and y. If rxy

= 1, it means that y increases with x in a linear fashion,

with no scatter. If rxy

= -1, it means that y decreases with x in a linear fashion, with

no scatter. The closer rxy

is to 1 or -1, the less scatter in the data. If rxy

= 0, it means

that y is uncorrelated with x, and there is no trend (Montgomery, 2001).

35

2.9.2.2 Regression Analysis

Linear regression analysis is also called linear least-squares fit analysis. The

goal of linear regression analysis is to find the "best fit" straight line through a set of

y vs. x data (Mason et al., 2003). An equation for a straight line that attempts to fit

the data pairs is normally chosen as

Y = ax + b …2.13

where ‘a’ is the slope, and b is the y-intercept when x=0. An upper case Y is used for

the fitted line to differentiate this from the actual data values, y. For each data pair

(xi, y

i), error, e

i, is defined as the difference between the predicted value and the

actual measured value.

ei = error at data pair i = Y

i - y

i = ax

i + b - y

i. ...2.14

A global measure of the error associated with all n data points can also be

defined. E is defined as the sum of the squared errors,

…2.15

Therefore, the best fitted model which can explain a set of data pairs is the one for

which E is the smallest (Montgomery and Runger, 1999). In other words, coefficients

‘a’ and ‘b’ need to be found which minimize E. To find a minimum (or maximum) of

a quantity, that quantity is differentiated, and the derivative is set to zero. Here, two

partial derivatives are required, since E is a function of two variables, ‘a’ and ‘b’.

…2.16

Finally, the following equations are derived for coefficients ‘a’ and ‘b’:

…2.17

…2.18

36

2.9.2.3 Nonlinear and Higher-Order Regression Analysis

Not all data are linear, and a straight line fit may not be appropriate. For some

data, a good curve fit can be obtained using a polynomial fit of some appropriate

order. The order of a polynomial is defined by the maximum exponent in the x data:

Excel can be manipulated to perform least-squares polynomial fits of any

order n, since Excel can perform regression analysis on more than one independent

variable simultaneously. To the right of the x column, new columns for x2, x3 ... xn

are added. All the data cells (x, x2, x3 ... xn) are chosen as the "Input X Range" in

the Regression window. Excel will treat each column as a separate variable. The

output of the regression analysis will be a y-intercept, and also a least-squares

coefficient for each of the columns. The coefficient for "X Variable 1" is a1,

corresponding to the x column. The coefficient for "X Variable 2" is a2,

corresponding to the x2 column. The coefficient for "X Variable n" is an,

corresponding to the xn column. Finally, the fitted curve is constructed from the

equation y = b + a1x + a2x2 + a3x3 + ... + anxn (Mason et al., 2003).

2.9.3 Optimization of Experiments

The conventional method of optimization involves varying one parameter at a

time and keeping the others constant. This often does not bring about the effect of

interaction of various parameters as compared to factorial design (Cochran and Cox,

1992). Response surface methodology (RSM) is a useful model for studying the

effect of several factors influencing the responses by varying them simultaneously

and carrying out a limited number of experiments.

RSM consists of a group of empirical techniques devoted to the evaluation of

relations existing between a cluster of controlled experimental factors and the

measured response, according to one or more selected criteria. The goal of RSM is to

efficiently hunt for the optimum values of a, and b such that y is maximized. RSM

works by the method of steepest ascent (Montgomery, 2001; Cornell, 1990). The

37

parameters are varied in the direction of maximum increase of the response until the

response no longer increases.

A prior knowledge and understanding of the process and the process variables

under investigation are necessary for achieving a more realistic model (Adinarayana

and Ellaiah, 2002). This can be achieved through thorough readings and

experimental observations.

2.9.3.1 Improvements of RSM

Further improvement is only possible by the following techniques:

� Data around a much smaller region are taken in the vicinity of the current

operating point. This increases the accuracy of the calculation of the direction

of steepest ascent.

� The data is replicated. This helps to cancel the effect of random noise

(experimental error).

� A higher-order regression scheme is used. Note that here only a linear (first-

order) regression analysis has been used. One can instead use a second-order

or higher-order regression analysis. Some RSM schemes have been devised

which can even take into account cross-talk between variables.

A caution about response surface methodology must be given here:

� RSM will always find a local maximum response. If there is more than one

peak in the function, one of the other peaks may have a larger value of y. In

other words, the local maximum response determined by RSM may not

necessarily be the optimum response (Cornell, 1990).

� Overall, RSM is a very powerful technique for optimizing a response.

2.10 Glycosylation in Insect Cells

Studies on the N-glycan structures produced by mosquito cells provided

earliest views of the insect protein N-glycosylation pathway (reviewed by Marchal et

38

al., 2001; Marz et al., 1995). The results showed that there were indeed some

striking differences, as the structures of the N-glycans on the glycoproteins produced

by insect cells have the significant differences from those produced by mammalian

cells (Figure 2.12). They are (i) inability to synthesize sialylated complex-type N-

glycans in contrast to mammalian cells (Marz et al., 1995; Altmann et al., 1999;

Marchal et al., 2001) and (ii) the presence of potentially allergenic structure,

Fuc�(1,3)GlcNAc-Asn.

It is clear that the inability of most lepidopteran insect cells to produce

mammalian-type N-glycan are attributable to extremely low levels of N-

acetylglucosaminyltransferase II (GlcNAcT II), �1,4-galactosyltransferase (�1,4-

GalT) activities and no detectable �2,6-sialytransferase (�2,6-ST) activities (Stollar

et al., 1976; Butters et al., 1981; Altmann et al., 1993; van Die et al., 1996; Hooker

et al., 1999). Furthermore, some insect cells have an N-acetylglucosaminidase,

which removes the terminal GlcNAc residue from GlcNAcMan3GlcNAc2-Asn and

eliminates the intermediate required for complex N-glycan production (Licari et al.,

1993; Altmann et al., 1995; Wagner et al., 1996; Marchal et al., 1999). Finally, it

has been reported that there is no detectable CMP-sialic acid, which is the donor

substrate required for sialoglycoprotein synthesis, in one insect cell line (Hooker et

al., 1999). Consequently, the major processed N-glycan typically produced by insect

cells is the paucimannose structure, as shown in Figure 2.12.

39

Figure 2.12: Protein N-glycosylation pathways in insect and mammalian cells.

Monosaccharides are indicated by their standard symbolic representations, as defined

in the key. The insect and mammalian N-glycan processing pathways share a

common intermediate, as shown. The major products derived from this intermediate

are paucimannose and complex N-glycans in insect and mammalian cells,

respectively. (Adapted from Jarvis, 2003)

Asn

Asn

Asn

Asn

Asn

α1,2-glucosidase I α1,3-glucosidase II

α-mannosidase I (RER) α-mannosidase I (Golgi)

N-Acetylglucosaminyltransferase I (GlcNAcT-I) --UDP-GlcNAc

α-mannosidase II Fucosyltransferase (FucT) --GDP-Fuc

Asn

�-N-Acetylglucosaminidase

Asn

N-Acetylglucosaminyltransferase II

N-Acetylgalactosyltransferase Sialytransferase Galactosyltransferase

Sialytransferase

INSECT MAMMALIAN

“PAUCIMANNOSE”

“COMPLEX”

Asn

Insect and mammalian N-glycan processing pathways share a common intermediate

Asn

N-Acetylglucosamine (GlcNAc) Mannose (Man) Fucose (Fuc) Galactose (Gal) N-Acetylgalactosamine (GalNAc) Sialic acid (Neu5Ac) Glucose (Glc)

Key to Symbols

40

2.11 Glycosyltransferases and Glycosidases Involved in N-glycan Processing

in Insect Cells

The processing pathway of N-glycans in lepidopteran insect and mammalian

cells is shown as Figure 2.12 (Jarvis, 2003). A number of studies have suggested

that initial processing of N-glycans in insect cells is similar or identical to that of

mammalian cells. However, insect cells appear to lack some of the processing

pathways of mammalian cells but contain additional glycosylation activities absent in

mammalian cells.

2.11.1 αααα-Glucosidase I, II and αααα-Mannosidase I

Glc3Man9GlcNAc2 is processed by α-glucosidase I, II and α-mannosidase I

to generate Man5GlcNAc2 structure. Many glycoproteins produced by lepidopteran

insect cells have high mannose type glycans. For example, N-glycans on human IgG

and hTf produced by Tn-5B1-4 cells included various high mannose type and

paucimannosidic glycans, with some incomplete complex-type glycans (Hsu et al.,

1997; Ailor et al., 2000). Expression of α-glucosidase I and II in several

lepidopteran insect cells appears adequate (David et al., 1993). In addition, α-

mannosidase I has been purified from Sf-21 cells (Ren et al., 1995) and cloned from

Sf-9 cells (Kawar et al., 1997), and its substrate specificity has been characterized

(Kawar et al., 2000). These results suggest that lepidopteran insect cells contain

ample α-glucosidase I, II and α-mannosidase I.

2.11.2 N-Acetylglucosaminyltransferase I (GlcNAcT-I) and αααα-mannosidase II

First of all, GlcNAc is added to Manα(1,3) branch of Man5GlcNAc2 by N-

Acetylglucosaminyltransferase I (GlcNAcT-I). Thereafter, two Man residues are

removed by α-mannosidase II. Substantial levels of GlcNAcT-I activities were

observed in several insect cell lines including Sf-9, Sf-21, Mb0503, and Bm-N

(Velardo et al., 1993). Like its counterpart from mammalian cells, α-mannosidase II

41

from insect cells requires GlcNAc on the Manα(1-3) branch for its activity (Altmann

et al., 1995). These studies suggest that lepidopteran insect cells have high levels of

α-mannosidase II and GlcNAcT-I in order to generate the precursor glycan required

for the formation of complex-type N-glycans.

2.11.3 N-Acetylglucosaminyltransferase II (GlcNAcT-II)

In mammalian cells, the product N-glycan of α-mannosidase II reaction

serves as an acceptor for the next reaction catalyzed by N-

Acetylglucosaminyltransferase II (GlcNAcT-II), which adds another GlcNAc to the

Manα(1,6) branch. However, lepidopteran insect cells, including Sf-9, Sf-21,

Mb0503, and Bm-N cells, have been shown to have only 1% or less of the

endogeneous GlcNAcT-II activity present in mammalian cells.

2.11.4 �-1,4-Galactosyltransferase (�1,4-GalT)

A terminal GlcNAc on either Man-branch is usually galactosylated by �1,4-

galactosyltransferase (�1,4-GalT) in mammalian cells. In contrast, galactosylated N-

glycans are rarely found in glycoproteins from lepidopteran cells. In fact, negligible

levels of �1,4-GalT activity were detected in Sf-9, Tn-5B1-4 and Mb0503 cells

(Hollister et al.,1998, van Die et al., 1996, Hollister et al., 2001). �1,4-GalT

activities in Sf-9 and Tn-5B1-4 cells were reexamined using an Eu-fluorescence

assay method (Abdul Rahman et al., 2002). Sf-9 did not contain any detectable

levels of �1,4-GalT activity (Abdul Rahman et al., 2002).

2.11.5 Core αααα-1,3- and αααα-1,6-Fucosyltransferases (FucT)

N-glycans with one or two GlcNAc on Man3-core can be further modified by

core fucosyltransferases. Both core Fuc-T’s require the presence of GlcNAc �(1,2)

on the Man α(1,3) branch for its action (Staudacher et al., 1998). N-glycans

42

containing either one or both of Fucα(1,3) and Fucα(1,6) attached to the Asn-linked

GlcNAc were identified on the membrane glycoproteins from Mb0503, Sf-21, and

Bm-N cells, in which glycoproteins from Mb0503 cells containing highest levels of

α-1,3-fucosylated N-glycans (Kubieka et al., 1994). Fuc-T C6, but not Fuc-T C3

were easily detected in Sf-9 cells.

2.11.6 �-N-Acetylglucosaminidase

A �-N-Acetylglucosaminidase specific for the terminal GlcNAc on the

Manα(1,3) branch was found in Sf-21, Bm-N and Mb0503 cells (Altman et al.,

1995), and it was suggested that this enzyme was localized in the microsome-like

membrane fraction in Sf-21 cells (Altman et al., 1995). Similar enzymatic activity

was also detected in the cell lysates and cell culture supernatant of insect cell derived

from Spodoptera frugiperda, Trichoplusia ni, Bombyx mori, or Malacosoma disstria

(Licari et al., 1993). Structural analysis of N-glycans from human IgG (Hsu et al.,

1997) and hTf (Ailor et al., 2000) expressed in Tn-5B1-4 cells suggested the

presence of such a �-N-Acetylglucosaminidase in Tn-5B1-4 cells. The further

removal of additional Man residues by α-mannosidase(s) can lead to the generation

of structures with fewer than three Man residues, as has been observed in several

studies.

2.11.7 Sialyltransferase (SiaT)

Sialytransferase (SiaT) adds N-acetylneuraminic acid to the terminal Gal

residues on N-glycans in mammalian cells. However, SiaT activity has yet to be

detected in Sf-9 (Hollister, 2001; Lopez et al., 1999; Hooker et al., 1999), Sf-21

(Hooker et al., 1999), Tn-5B1-4 (Lopez et al., 1999), Mb0503 (Lopez et al., 1999),

and Ea4 (Hooker et al., 1999) cells, even using highly sensitive assays with

radiolabeled CMP-NeuNAc or fluorescent CMP-NeuNAc derivatives as the donor

substrate.

43

2.12 Sugar Nucleotides Involved in N-glycan Processing in Insect Cells

2.12.1 Endogenous Sugar Nucleotide Levels in Lepidopteran Insect Cells

All glycosyltransferases in the synthetic pathway for complex-type N-glycans

require respective sugar nucleotides as substrate donor. Examination of the sugar

nucleotide concentrations in lepidopteran insect cells demonstrated the presence of

substantial levels of UDP-hexose, UDP-N-acetylhexosamine, GDP-Fuc, and GDP-

Man in Sf-9, Mb0503, and Tn-5B1-4 cells (Lopez et al., 1999). However, no CMP-

NeuNAc was detected in the same study (Lopez et al., 1999). Similar results were

obtained on the sugar nucleotide levels in Sf-9 and Tn-5B1-4 cells (Tomiya et al.,

2001).

2.12.2 Enzymes Involved in Sialic Acid and CMP-Sialic Acid Synthesis

Of particular significance is the absence in lepidopteran insect cells of the

CMP-NeuNAc necessary for sialylation of N-glycans. In mammalian cells, sialic

acids are synthesized from UDP-GlcNAc through multiple enzymatic reactions as

shown as Figure 2.13.

The bifunctional enzyme, UDP-N-acetylglucosamine (UDP-GlcNAc) 2

epimerase / N-acetylmannosamine (ManNAc) kinase, is believed to be a key enzyme

in the biosynthesis of NeuNAc in rat liver (Hinderlich et al., 1997). This enzyme

converts UDP-GlcNAc-6P to ManNAc-6-P, which is further converted to N-

acetylneuraminic acids (NeuNAc) by N-acetylneuraminate-9-phosphate synthase

(SAS) and N-acetylneuraminate-9 phosphate phosphatase. NeuNAc is then

converted to CMP-NeuNAc by CMP-NeuNAc synthase (CMP-SAS).

Effertz et al. (1999) reported that the UDP-GlcNAc 2-epimerase activity in

Sf-9 cells was about 30 times less (in term of specificity activity) than that in rat liver

cytosol fraction. Interestingly, Sf-9 cells had 50 times higher ManNAc kinase

activity compared with the 2-epimerase activity (Effertz et al., 1999). It was

44

reported that Sf-9 cells contained negligible levels of neuraminic acids, and no

detectable N-acetylneuraminic-9-phosphate synthase activity was present in the

lysate of Sf-9 cells (Lawrence et al., 2000). It was found that Sf-9 cells do not have

detectable CMP-sialic acid synthase activity (Lawrence, 2001).

Figure 2.13: CMP-Neuraminic acid synthesis pathway. The dotted arrow indicates

pathways which are insufficient in lepidopteran insect cells. (Tomiya et al., 2003)

2.13 Engineering of N-glycan Processing Pathway

The general strategy for humanizing glycoproteins produced by the insect

cell-baculovirus expression system is shown in Figure 2.14. The goal of engineering

N-glycan processing is to develop a new insect cell-baculovirus expression vector

system(s) that can express human-like sialylated multi-antennary complex-type N-

glycans. As described in the earlier sections, several lines of evidence suggest that

majority of lepidopteran insect cells currently used for protein expression apparently

lack several enzymes for such a goal. Moreover, lepidopteran insect cells contain the

undesirable �-N-acetylglucosaminidase and Fu-T C3. The former diminishes the key

glycans containing GlcNAc � (1,2)Manα(1,3) which stunt the normal growth of

UDP-

ManNAc-

NeuNAc-9-

NeuNA

CMP-

ManN Bifunctional UDP-GlcNAc 2-epimerase/ManNAc kinase

N-acetylneuraminic acid 9-phosphate synthase (SAS)

N-acetylneuraminic acid 9-phosphate phosphatase

CMP-neuraminic acid synthase (CMP-SAS)

N-acetylmannosamine kinase

Abbreviations:

UDP-GlcNAc - Uridine-5’-diphopho-N-acetylglucosamine ManNAc - N-acetylmannosamine ManNAc-6-P - N-acetylmannosamine-6-phosphate NeuNAc-9-P - N-acetylneuraminic acid-9-phosphate NeuNAc - 5-acetamido-3,5-dideoxy-D-glycero-D-galacto-2-nonulosonic acid

(N-acetylneuraminic acid) CMP-NeuNAc - Cytidine-5’-monopho-N-acetylneuraminic acid

45

complex-type N-glycans, and the latter generates potentially allergenic N-glycans.

Therefore, the N-glycan processing pathways need to be altered in the insect cells by

enhancing or suppressing respective processing pathways.

Many lines of evidence have indicated that the inability of the vast majority

of lepidopteran cells to synthesize mammalian type N-glycans. The inability to

obtain such N-glycans in lepidopteran cells can be attributed to the insufficient levels

of �1,4-GalT, GlcNAcT-II, SiaT, UDP-GlcNAc 2 epimerase / ManNAc kinase ,

UDP-N-acetylneuraminate-9-phosphate synthase, CMP-NeuNAc synthase activities.

�-N-Acetylglucosaminidase, which removes GlcNAc on the Man�(1,3) branch, have

been detected in several lines of lepidopteran cells. This enzyme apparently prevents

synthesis of complex-type N-glycans by removing the key intermediate glycan

containing GlcNAc�(1,2)-Man (1,3). In addition, FucT C3 generates the potentially

allergenic glycan structure, Fuc�(1,3)GlcNAc-Asn, on glycoproteins expressed in

lepidopteran cells.

46

Figure 2.14: General strategy for humanization of glycoprotein produced by

lepidopteran cell-baculovirus expression system. (Tomiya et al., 2003)

2.13.1 Improvement of N-Acetylglucosaminylation of the Manαααα(1,3)-Branch

�-N-acetylglucosaminidase was implicated as a problem in N-glycan

elongation by its absence of Estigme acrea cells, which produced N-glycans

containing terminal N-acetylglucosamine residues (Wagner et al., 1996). Sf-9 cells

are known to contain high levels of �-N-acetylglucosaminidase (Wagner et al.,

1996). Using Sf-9 cells, Wagner et al. succeeded in N-glycan elongation by

coexpression of human �-N-acetylglucosaminyltransferase I and fowl plague virus

hemagglutinin. Watanabe et al. (2001) examined the effect of a �-N-

acetylglucosaminidase inhibitor, that is 2-acetamide-1,2-dideoxynojirimycin (2-

ADN) on bovine interferon-� (bIFN- �) on production in Tn-5B1-4 cells. Watanabe

et al. (2001) speculated that the inhibitor enhanced accumulation of substrates

Lacking some enzymes: 1. �1,4-GalT 2. GlcNAcT-II 3. SiaT 4. N-acetylneuraminate-9-

phosphate synthase 5. CMP-NeuNAc synthase

Problem

Solution

Genetic Engineering:

1. Recombinant baculovirus

2. Transgenic Insect cells

Metabolic Engineering:

1. Sugar Feeding

2. Chemical Inhibitor

Evaluation

Glycoproteins Expression: 1. Glycosyltransferases activity 2. Glycosidases activity 3. Sugar nucleotides 4. Glycan analysis

47

processing a �(1,2)-linked GlcNAc, thereby leading to further elongation by �1,4-

GalT and SiaT to form sialylated N-glycans. However, the overall increase of N-

glycan containing �(1,2)-linked GlcNAc was not determined.

2.13.2 Improvement of Galactosylation

Expression of a �1,4-GalT by a baculovirus vector increased galactosylation

of glycoprotein (Jarvis et al., 1996), indicating that the mammalian enzyme

expressed by baculovirus infection could function in the infected lepidopteran cells

and that it could compete with the �-N-acetylglucosaminidase activity insect cell

(Jarvis et al., 1996). Similar results were obtained when human serum transferrin

(hTf) was expressed by Tn-5B1-4 cells infected with two baculoviruses, one

encoding a gene for hTf and the other encoding a gene for a mammalian GalT (Ailor

et al., 2000). In this study, 13% of the total N-glycans were galactosylated, and

protection of GlcNAc on Manα(1,3) branch against �-N-acetylglucosaminidase by

galactosylation was confirmed (Tomiya et al., 2003).

2.13.3 Production of Biantennary Complex-Type N-glycans

Production of biantennary complex-type N-glycans was achieved recently by

expressing a mammalian N-acetylglucosaminyltransferase II (GlcNAcT-II) in

lepidopteran cells using a transgenic insect cell, SfSWT-1 (Hollister et al., 2002), or

using baculovirus expression vector system (Tomiya et al., 2003).

2.13.4 Formation of Sialylated N-glycans

Sialylation of N-glycans was in Tn-5B1-4 cells when the cells were cultured

in the presence of a hexosaminidase inhibitor (2-ADN) (Watanabe et al., 2001). This

result is particular intriguing since Tn-5B1-4 cells lack �1,4-GalT, SiaT and CMP-

NeuAc synthase. Unfortunately, analysis was only by lectin blot and not by

quantitative chemical analysis of the exact structures. Sialylation was also detected

48

in virion glycoprotein, gp64, when Sf-9 cells were infected with a recombinant

baculovirus vector encoding mammalian �1,4-GalT and α2,6-sialyltransferase (α2,6-

SiaT), while no sialylation was detected in the absence of either �1,4-GalT or α2,6-

SiaT (Jarvis et al., 2001).

2.13.5 Synthesis of CMP-NeuNAc

The processing steps catalyzed by UDP-N-acetylglucosamine 2 epimerase /

N-acetylmannosamine kinase, N-acetylneuraminate-9-phosphate synthase, and CMP-

NeuNAc synthase represent bottlenecks in the CMP-NeuNAc synthesis pathway of

lepidopteran cells (Tomiya et al., 2003). To overcome this problem, Tomiya et al.

(2003) reported that they cloned mammalian N-acetylneuraminic-9-phosphate

synthase and CMP-NeuNAc synthase (Lawrence et al., 2001), and expressed these

enzymes in Sf-9 cells. When Sf-9 cells were infected with a recombinant baculovirus

expression vector encoding N-acetylneuraminate-9-phosphate synthase and were

cultured in a medium supplemented with N-acetylmannosamine (ManNAc), Sf-9

cells produced high levels of N-acetylneuraminic acid (NeuNAc) (Lawrence et al.,

2000).

2.14 Galactosylation in N-Glycan Processing in Insect Cells

Galactosylation is a process which links the galactose sugar to the end of the

GlcNAc (�1,2)Man (�1,3) chains. In the in vivo galactosylation, mammalian �1,4-

GalT is being introduced artificially to the cell culture which secretes the protein.

This is also known as coinfection, which is the simultaneous infection of a single

host cell by two types of different virus particles. Another new technology being

established is the in vitro galactosylation, which uses mammalian �1,4-GalT to add

the missing galactose sugar units to the carbohydrate chains of the protein via UDP-

Gal , a donor sugar known as sugar nucleotides. The process is performed after the

protein is expressed and secreted by the host cell.

49

There are three main factors that are involved in galactosylation, which are

sugar acceptor, sugar donor and enzyme. In this study, human serum transferrin was

used as the sugar acceptor. Human serum transferrin was used as the model protein

due to its simplicity of biantennary N-glycan structure. Uridine-diphosphogalactose

(UDP-Gal) was used as the substrate donor and mammalian �1,4-GalT was used as

the enzyme.

2.14.1 Sugar acceptor

Human serum transferrin (hTf) is a serum glycoprotein found in the

physiological fluids of vertebrates (Aisen, 1989; Thorstensen and Romslo, 1990) and

insect larva (Bartfeld and Law, 1990) that is responsible for carrying Fe+3 to all cells

in the body. When bound to iron, the circulating transferrin is recognized by a

specific surface receptor on cells and internalized to release iron into the cytoplasm

(Trowbridge et al., 1984). Serum transferrin also plays a role in host defense by

depriving any circulating microorganism of essential iron (Bullen et al., 1990). HTf

is a single-chain glycoprotein of 679 amino acids containing two potential N-linked

glycosylation sites in its carboxy-terminal domain at Asn413 and Asn611

(MacGillivray et al., 1983), with a glycosylation-dependent molecular mass in the

range 76-81 kDa (MacGillivray et al., 1982). Previous studies have shown that the

transferrin glycoforms present in human serum are comprised of species having

terminally sialylated bi-, tri-, and tetraantennary oligosaccharides (Leger et al., 1989;

Fu and van Halbeek, 1992). The most dominant glycoform includes bianntenary

oligosaccharides located at both asparagines positions, although, changes in

physiological conditions can affect the N-glycan pattern observed in the host

(Montreuil et al., 1997).

2.14.2 Substrate Donor

UDP-Gal is a substrate also known as sugar nucleotide (Figure 2.15), used by

galactosyltransferase for extension of sugar chain of glycoproteins. Once the sugar

nucleotides are synthesized in the cytosol, they are topologically mislocalized, since

50

most glycosylation occurs in the ER and Golgi. Their negative charge prevents them

from simply diffusing across membranes into these compartments. To overcome this

problem, cells have devised a set of nonenergy-requiring sugar nucleotide

transporters, actually antiporter, that deliver sugar nucleotides into the lumen if these

organelles with simultaneous exit of nucleotide monophosphates which must first

derived from the nucleotide diphosphates (Figure 2.16).

Nucleotide sugar transporters are membrane proteins localized in the

endoplasmic reticulum and Golgi apparatus. They play an indispensable role in

constructing the sugar chains of glycoconjugates. The transporters carry sugars into

the endoplasmic reticulum and Golgi apparatus, in which they are used by specific

transferases as precursors of sugar chains (Kawakita et al., 1998; Hirachberg et al.,

1998; Berninsone et al., 2000; Gerardy-Schahn et al., 2001; Hirschberg, 2001).

More than simply functioning as a passive entrance route of nucleotide sugars into

the organelles, the transporters may regulate the amounts of nucleotide sugars

available in the lumen of the endoplasmic reticulum or Golgi apparatus and

consequently may affect the sugar chain composition of a cell (Kumamoto et al.,

2001).

For most glycosylation reaction, the sugar nucleotide donates the sugar,

resulting in the formation of nucleoside diphosphate, which must be converted into a

monophosphate by the nucleoside diphosphatase that occurs in the Golgi lumen.

Exchange through the antiporters is electroneutral, since the sugar nucleotide with

two negative charges (one on each phosphodiester) enters and the nucleoside with a

single phosphomonoester exits.

51

Figure 2.15: Structure of a nucleotide sugar that can serve as a sugar donor in a

glycosyltransferase reaction. UDP, uridine diphosphate.

Figure 2.16: Transporters for sugar nucleotides, PAPS, and ATP are located in the

Golgi membranes of mammals, yeast, protozoa, and plants. These proteins are

actually antiporters, and the corresponding nucleoside monophosphate is carried into

the cytosol with sugar nucleotide transport. Since most glycosylation reactions

produce a nucleoside diphosphate, this requires conversion to the nucleoside

monophosphate. (Adapted from Hirschberg et al., 1998)

O

N

O N O

H O

CHO P O

O- O P

O

O- O

OH

CHO

H O

UDP Galactose

Uridin

Ribose

52

2.14.3 Enzyme

The golgi or endoplasmic reticulum glycosyltransferases constitute a

functional family of approximately 300 membrane-bound enzymes that, in general,

synthesizes complex carbohydrates of glycoconjugates of cells by transferring a

sugar moiety of a sugar nucleotide to an acceptor sugar (Roseman, 2001; Hill, 1979).

The galactosyltransferase family is the subset of the glycosyltranferases, in the

presence of the metal ion, transfers galactose from UDP-Gal to an acceptor sugar

molecule. To date, three subfamilies, �1,4-, �1,3-, and �1,3-, have been well

characterized (Amado et al., 1998) and they generate �1,4-, �1,3-, and �1,3- linkages

between galactose and the acceptor sugar, respectively. Cloning has identified the

presence in each family of several members that have sequence homology within the

family members (Amado et al., 1998). The �1,4-galactosyltransferase (Gal-T)

family, which was the first one to be cloned (Narimatsu et al., 1986; D’Agostaro et

al., 1989; Shaper et al., 1986), consists of at least seven members, �1,4Gal-T1 to

�1,4Gal-T7 (Amado et al., 1998), with a 25 to 55% sequence homology. These

enzymes are expressed in different tissues and show differences in the

oligosaccharide acceptor specificity (Lo et al., 1998; Guo et al., 2001).

In the mammary gland, only �1,4Gal-T1 is expressed (Shaper et al., 1998)

and it interacts with the calcium binding protein, �-lactalbumin, that is expressed in

the mammary gland during lactation, to form the lactose synthase complex. The

formation of this complex alters the substrate specificity of �1,4Gal-T1 such that

glucose at physiological concentrations can serve as the acceptor sugar, resulting in

the synthesis of lactose (Gal�1,4Glc).

The protein domain structure of �1,4Gal-T1 consists of a short NH2-terminal

cytoplasmic domain, a single transmembrane domain, a stem region, and a large

lumenal, catalytic domain which contains the metal (Mn2+), UDP-Gal, and the

acceptor sugar binding sites (Paulson et al., 1989; Aoki et al., 1990; Yadav et al.,

1990).

53

2.15 Purification of Transferrin

The purity of a protein is a pre-requisite for its structure and function studies

or its potential application. For structure studies or therapeutic applications, protein

of high degree is required. A wide variety of protein purification techniques like gel

filtration chromatography, ion-exchange chromatography, affinity chromatography

and hydrophobic interaction chromatography (HIC), are available. Every separation

technique is as important and the application is dependent on target proteins which

vary in biological and physico-chemical properties: molecular size, net charge,

biospecific characteristics and hydrophobicity (Kennedy, 1990; Garcia and Pires,

1993).

2.15.1 Hydrophobic Interaction Chromatography (HIC)

Hydrophobic interactions have a great importance in the biological systems.

They are the dominant force in protein folding and structure stabilization (Privalov

and Gill, 1988; Dill, 1990a; Murphy et al., 1990; Makhatafze and Privalov, 1995)

and the maintenance of the lipid bilayer structure of biological membranes (Tanford,

1973). Proteins comprise of a number of hydrophobic amino acids, with different

distribution and hydrophobicity. Hence, a specific separation can be possible with

hydrophobic supports or matrices (Ochoa, 1978; Vogel et al., 1983; Lindahl and

Vogel, 1984). Although HIC exploits nonspecific affinities, it has been successfully

used for separation purposes as it displays binding characteristics complementary to

other protein chromatographic techniques (Janson and Rydén, 1993).

Many theories have been proposed for HIC are essentially based on the

interactions between hydrophobic solutes and water, but none of them has got

universal acceptance. Tiselius (1948) was the first to use the term ‘salting-out

chromatography’. Hjertén et al. (1974) synthesized charge-free hydrophobic

adsorbents and demonstrated that the binding of proteins was enhanced by high

concentrations of neutral salts, as previously observed by Tiselius (1948), and that

elution of the bound proteins was achieved simply by washing the column with salt-

free buffer or by decreasing the polarity of the eluent (Hofstee, B.H.J. 1973; Porath,

54

J. 1973; Hjertén, S. 1974). According to Melander (1984), the most important

parameters that determine the effect of salt on the retention in HIC are the salt

molality and the molal surface increment of the salt. An increase in salt molality in

the mobile phase or a change of salt to one of greater molal surface increment will

promote an enhancement in surface tension with an increased retention of proteins in

HIC

Srinivasan and Ruckenstein (1980); Van Oss et al. (1986) have proposed that

HIC is due to van der Waals attraction forces between proteins and immobilized

ligands. The van der Waals attraction forces between protein and ligand increase as

the ordered structure of water increases in the presence of salting out salts. Van der

Waals force is much weaker compare to ionic force and specific affinity force.

Hence, biological activity of the biomolecules is maintained and the structural

damage of using HIC is minimum relative to affinity, ion-exchange or reversed-

phase chromatography (RPC) (Fausnaugh et al., 1984; Regnier, 1987).

The commercial availability of well-characterized HIC adsorbents opened

new possibilities for purifying a variety of biomolecules such as serum proteins

(Janson, J-C., 1978; Hrkal, Z., 1982), membrane-bound proteins (McNair, R.D.,

1979), nuclear proteins (Comings, D.E., 1979), receptors (Kuehn, L. 1980), cells

(Hjertén, S. 1981), and recombinant proteins (Lefort, S., 1986; Belew, M., 1991 in

research and industrial laboratories. The principle for protein adsorption to HIC

media is complementary to ion exchange chromatography and gel filtration. HIC can

separate the pure native protein from other forms HIC has also found use as an

analytical tool to detect protein conformational changes.

2.15.1.1 Factors affecting HIC

The main factors affecting HIC are: 1) Ligand type and degree of

substitution, 2) Type of base matrix, 3) Type and concentration of salt, 4) pH, 5)

Temeprature and 6) Additives (Amersham Bioscience, 1993).

55

The type of immobilized ligand determines primarily the protein adsorption

selectivity of the HIC absorbent. HIC contain alkyl or aryl chains of any size, and in

practice, most separation employ phenyl and butyl group. Fig. 3 showed the glycidyl

ether coupling HIC media, which produces charge free gels and only have

hydrophobic interactions with proteins. At constant substitution, the protein binding

capacities of HIC absorbents, hydrophobicity and the strength of interaction would

increase, but the adsorption selectivity would decrease with increased alkyl chain

length. The protein binding capacities of HIC adsorbents also increase with increased

degree of substitution of immobilized ligand. The apparent binding capacity of the

adsorbent would remains constant, but the strength of interaction would increase, at

sufficient high degree of ligand substitution or n-alkyl chain length (Jennissen, H.P,

1975; Rosengren, J., 1975; L��s, T., 1975; Maisano, F., 1985). This will cause the

bound solutes difficult to elute due to multi-point attachment and extreme elution

condition will be required.

Figure 2.17: Different hydrophobic ligands coupled to cross-linked agarose

matrices. (Amersham Bioscience, 1993)

The two most widely used types of support are strongly hydrophilic

carbohydrates, e.g. cross-linked agarose, or synthetic copolymer materials. The

selectivity of a copolymer support can change in function of the different type of

supports eventhough same type of ligand is used. To achieve the same type of results

56

on an agarose-based matrix as on a copolymer support, it may be necessary to

modify adsorption and elution conditions.

The effects of salts in HIC can be accounted for by reference to the

Hofmeister series for the precipitation of proteins or for their positive influence in

increasing the molal surface tension of water (Figure 2.18, Figure 2.19). The salts at

the beginning of the series promote hydrophobic interactions and protein

precipitation (salting-out or sntichsotropic), are considered to be water structuring;

whereas salts at the end of the series (salting-in or chaotropic ions) randomize the

structure of the liquid water and thus tend to decrease the strength of hydrophobic

interactions (Porath, 1987). Salts such as sodium, potassium or ammonium sulfates

are the most effective to promote ligand protein interactions. Magnesium sulphate

and magnesium chloride do not enhance the protein retention despite the fact that

they increase the surface tension of water. Type of salt in the eluent not only altered

the overall retention of the proteins, but also affects selectivity of the separations

(Rippel and Szepesy, 1994).

Figure 2.18: The Hofmeister series on the effect of some anions and cations in

precipitating proteins.

Figure 2.19: Relative effects of some salts on the molal surface tension of water.

The concentration of salt strongly influences the selectivity in protein

adsorption and the influence is different and dependent both on the stationary phase

and the buffer salts (Oscarsson, S., Kårsnås, P., 1998). In HIC, the use of high salt

concentration on the equilibration buffer and sample solution promotes the ligand–

protein interactions and consequently the protein retention. As the concentration of

57

such salts is increased, the amount of bound proteins also increases almost linearly

up to a specific salt concentration and continues to increase in an exponential manner

at still higher concentrations. The adsorbed proteins are eluted by stepwise or

gradient elution at decreasing salt concentration in the eluent. The viscosity, UV

transparency and stability at alkaline pH values are other important factors for

choosing the neutral salts (Narhi et al., 1989).

In general, an increase in pH weakens hydrophobic interactions (Porath, J.,

1973; Hjertén, S., 1973); a decrease in pH results in an apparent increase in

hydrophobic interactions. This is probably due to changing of charged groups at

different pH and thereby leading to an increase in the hydrophilicity or

hydrophobicity of the proteins. Proteins which do not bind to a HIC adsorbent at

neutral pH bind at acidic pH (Halperin, G., 1981). Hjertén et al. (1986) found that the

retention of proteins changed more drastically at pH values above 8.5 and/or below 5

than in the range pH 5–8.5. These findings suggest that pH is an important separation

parameter in the optimization of hydrophobic interaction chromatography.

In HIC, increasing the temperature enhances protein retention and lowering

the temperature generally promotes the protein elution (Hjerte´n et al., 1974). Van

der Waals attraction forces, which operate in hydrophobic interactions (Srinivasan,

R., 1980) increase with increase in temperature (Parsegian, V.A., 1970). However, an

opposite effect was reported by Visser & Strating (1975). This apparent discrepancy

is probably due to the differential effects exerted by temperature on the

conformational state of different proteins and their solubilities in aqueous solutions.

(Amersham Bioscience, 1993).

Additives can be used in HIC, not only to improve protein solubility or to

modify protein conformation, but also to promote the elution of the bound proteins.

The most widely used are water-miscible alcohols (e.g. ethanol and ethylene glycol)

and detergents. Additives decrease the surface tension of water thus weakening the

hydrophobic interactions to give a subsequent dissociation of the ligand-solute

complex (table 2.4). The non-polar parts of alcohols and detergents and bound

proteins compete and displace the others for the adsorption sites on the HIC media.

The separation mode involved charged group of detergent is a mixed ion-exchange

58

hydrophobic interaction process (Janson and Ryde´n, 1993). Elution using additive

could lead to denaturation of protein, so it is only applied when other milder

conditions do not promote protein recovery. However, if strongly hydrophobic

proteins bind to the stationary phase, additives can be used in cleaning up HIC

columns.

Table 2.4: Physical Properties of some solvent used in HIC (Amersham

Biosciencesa, 1993)

Solvent Viscocity

(centipoises) Dielectric Constant

Surface tension

(dynes/cm)

Water 0.89 78.3 72.00

Ethylene glycol 16.9 40.7 46.70

Dimethyl Sulphoxide 1.96 46.7 43.54

Dimethyl Formamide 0.796 36.71 36.76

n-propanol 2.00 20.33 23.71

2.15.2 Ion Exchange Chromatography

Ion exchange is probably the most frequently used chromatographic

technique for the separation and purification of proteins, polypeptides, nucleic acids,

polynucleotides, and other charged biomolecules (Bonnerjera, J., 1986). The reasons

for the success of ion exchange are its widespread applicability, its high resolving

power, its high capacity, and the simplicity and controllability of the method.

Separation in ion exchange chromatography depends upon the reversible adsorption

of charged solute molecules to immobilized ion exchange groups of opposite charge.

Separation is obtained since different substances have different degrees of interaction

with the ion exchanger due to differences in their charges, charge densities and

distribution of charge on their surfaces. These interactions can be controlled by

varying conditions such as ionic strength and pH. The differences in charge

properties of biological compounds are often considerable, and since ion exchange

chromatography is capable of separating species with very minor differences in

properties, e.g. two proteins differing by only one charged amino acid, it is a very

powerful separation technique.

59

The separation using ion exchange is based primarily on differences in the

ionic properties of surface amino acids. Thus, at a given pH, protein posses an

overall net charge. The relationship of the protein and the net charge can be

visualized as a titration curve (Figure 2.20). This curve reflects how the overall net

charge of the protein changes according to the pH of the surroundings. The

isoelectric point (pI) of each protein is the pH at which the protein has zero surface

charge. The net charge will be more positive at a pH lower than pI protein; more

negative at a higher pH. Proteins with different pI can be separated by being passed

through a chromatofocusing. Selected working pH is 1 unit away from the pI of

protein.

Figure 2.20: Effect of pH on protein net charge

2.15.2.1 Factor affecting IEX

Matrix of IEX may be based on inorganic compounds, synthetic resins or

polysaccharides. The characteristics of the matrix determine its chromatographic

properties such as efficiency, capacity and recovery as well as its chemical stability,

mechanical strength and flow properties. The nature of the matrix will also affect its

behaviour towards biological substances and the maintenance of biological activity.

The first ion exchangers designed for use with biological substances were the

cellulose ion exchangers developed by Peterson and Sober (1956), then Ion

60

exchangers based on dextran (Sephadex), followed by those based on agarose

(Sepharose) and cross-linked cellulose (Sephacel). Hydrophilic nature of cellulose

had little tendency to denature protein, but it had low capacities and had poor flow

properties due to their irregular shape.

An ion exchanger consists of covalently bound charged group to an insoluble

matrix. The charged groups are associated with mobile counter ions which can be

reversibly exchanged with other ions of the same charge without altering the matrix.

Positively charged exchangers have negatively charged counter-ions (anions)

available for exchange and are called anion exchangers; negatively charged

exchangers have positively charged counter-ions (cations) and are termed cation

exchangers.

Figure 2.21: Ion exchanger types.

The presence of charged groups is a fundamental property of an ion

exchanger. The type of group determines the type and strength of the ion exchanger;

their total number and availability determines the capacity. Table 2.5 show some

funtional groups which have been chosen for use in ion exchangers. Sulphonic and

quaternary amino groups are used to form strong ion exchangers; the other groups

form weak ion exchangers. Strong ion exchangers are completely ionized over a

wide pH range whereas with weak ion exchangers, the degree of dissociation and

thus exchange capacity varies much more markedly with pH. For cation exchanger,

carboxymethyl- and sulfo- group show significant differences when pH below 5.

Carboxymethyl- group begin to protonated below pH 5. Region of operation for

carboxymethy- is at around pH 4.5. Sulfo groups which remain fully charged right

down to pH1 is needed for low pH operation. As for anion exchanger, DEAE- groups

61

become uncharged at high pH, and not suitable for use above pH8.5. DEAE- and Q-

groups are highly charged at low pH, so they also suitable to purify low pI protein.

Hence, strong ion exchangers like Q and S sepharose which charged at very wide

range of pH.

Table 2.5: Functional groups used on ion exchangers (Amersham Bioscience)

The pH in the micro environment of an ion exchanger is not exactly the same

as eluting buffer because Donnan effect can repel or attract protons within the

adsorbent matrix. In general, pH in the matrix is up to 1 unit higher than that in the

surrounding buffer in anion exchanger and 1 unit lower in cation exchanger. The

lower the ionic strength of the buffer, the larger of the Donna effect. This phenomena

is very important considering the stability of enzymes as a function of pH. The

Donnan effect limits the operational pH range of ion exchangers, especially in the

mildly acid range.

The charge, the nature of the matrix particles in terms of bead size, flow rate

required, capacity determine the choice of adsorbent. Table 2.6 and Table 2.7 show

the capacity data and the characteristic of 4 common commercial ion exchange

matrix.

62

Table 2.6: Capacity data for Sepharose Fast Flow ion exchangers (Ammersham Bioscience)

N.D. = Not determined *For anion exchangers (DEAE and Q) the starting buffer was 0.05 M Tris, pH 8.3 and for cation exchangers (CM and S) 0.1 M acetate buffer, pH 5.0. Limit buffers were the respective start buffers containing 2.0 M NaCl. Table 2.7: Characteristics of Q, SP, DEAE and CM Sepharose Fast Flow

(Ammersham Bioscience).

* working pH range refers to the pH range over which the ion exchange groups remain charged and maintain consistently high capacity. ** pH stability, long term refers to the pH interval where the gel is stable over a long period of time without adverse effects on its subsequent chromatographic performance. pH stability, short term refers to the pH interval for regeneration and cleaning procedures

CHAPTER 3

MATERIALS AND METHODS

3.1 Materials

Spodoptera frugiperda (Sf-9) insect cells was purchased from ATCC cat. No.

1711 (Rockville, MD). The recombinant baculovirus containing the gene coding for

Human Transferrin and �1,4-galactosyltransferase were provided by Prof Dr Michael

J. Betenbaugh of Johns Hopkins University, USA. Wild type recombinant virus

AcMNPV was a gift from Prof Dr Mohd Sanusi Jangi, UKM, Malaysia.

3.2 Chemicals

Sf-900 II Serum Free Media (SFM) and Fetal Bovine Serum (FBS) were

from GIBCO BRL (Gaithersburg, MD). Goat anti-Human Transferrin-affinity

purified, Goat anti-human transferrin-HRP conjugate, Calibrator-Human Reference

Serum and TMB (3,3’,5,5’-tetramethylbenzidene) Peroxidase Substrate and

Peroxidase Solution B (water soluble) were obtained from Bethyl Laboratories Inc

(Texas). TMB Stabilized Substrate for Horseradish Peroxidase (water insoluble) was

purchased from Promega, (Madison, WI). Asialofetuin, �-galactosidase (from

bovine), peroxidase-labeled RCA 1, uridine-5’-diphosphogalactose disodium salt

(UDP-Gal), uridine 5’-Triphosphate sodium (UTP), acrylamide, bis-acrylamide,

64

bovine serum albumin (BSA), ammonium persulfate, citric acid, 2-mercaptoethanol,

silver nitrate, triton X-100, dimethyl sulphoxide (DMSO), �-lactalbumin, N,N,N',N'-

tetramethylethylenediamine (TEMED), anisaldehyde, 4-Morpholinepropanesulfonic

acid (MOPS), tris, glycine, lactose, glucose, manganese chloride and ammonium

phosphate were purchased from Sigma (Missouri, USA). Trypan blue, ethanol,

acetic acid, ethylenediamine tetraacetic acid disodium salt dehydrate (EDTA), 38%

formaldehyde, sodium chloride, sodium hydroxide, hydrochloric acid, sodium

dodecyl sulfate (SDS), bromophenol blue, sodium bicarbonate, tween 20, glycerol,

phosphoric acid, methanol, skimmed milk, potassium chloride, potassium phosphate

dibasic, potassium dihydrogen phosphate, zink sulfate 7-hydrate, barium hydroxide,

tetrabutylammonium hydrogen sulfate (TBAS) and dichloromethane were from

Fluka (Missouri, USA). Ammonium hydroxide, butanol, diethyl ether and

glutaldehyde were purchased from Merck (New Jersey, USA). NADPH, α-

ketaglutarate, triethanolamine, glutamate dehydrogenase (GLDH), Ammonia were

from Randox Laboratories (Antrim, UK). D-Glucose, L-Lactate, L-Glutamine and L-

Glutamate calibrator were from YSI laboratory (Ohio, USA)

3.3 Equipments

The High Performance Liquid Chromatography (HPLC) System used was

LC-10Dvp HPLC system and a NovaPac C18 column (3.9 x 150mm, 4 �m) with a

NovaPac C18 guard cartridge. The electrophoresis system used was Mini-Protean II

from Bio-Rad (California, USA). Western blot analysis was done using Trans-Blot

Electrophoretic Transfer Cell (Bio-Rad Laboratories, Melville, NY). Shimadzu UV-

160 spectrophotometer (Minnesota, USA) was used to measure absorbance at

562nm, 450nm and 280nm. The electrophoresis unit used was Mini-Protean II from

Bio-Rad (California, USA). (Minnesota, USA) was used to measure light absorbance

in colorimetric assays. The slow rotary shaker was purchased from Bellco

Biotechnology (New Jersey, USA). Biological safety cabinet (laminar flow hood)

was from Telstar Bio-II-A (Germany). Inverted phase contrast microscope and

compound microscope were from Zeiss Instruments (Germany). Incubator was

purchased from Memmert (Germany). Biochemical analyzer YSI 2700 Select (Ohio,

USA) was used to analyze glucose, lactate and glutamine contents

65

3.4 Spodoptera frugiperda (Sf-9) Insect Cells

3.4.1 The Preparation of TC100 Medium From Powdered Formulation

Initially, all glassware were sterilized. The medium composition was TC-100,

10% FBS, and supplements. For powdered medium, the medium was dissolved in

about 800 ml deionized water. For TC-100 medium, 0.03 g/L sodium bicarbonate

was added. pH was adjusted to 6.2 with 1 M KOH/NaOH (about 20-30 ml).

Deionized water was added to make up a total volume of medium of 1 L. The

medium was filter-sterilized through a 0.22 micron filter. TC100 solution stored at

4oC has a shelf live of at least 1 year while TC100 + FBS solution stored at 4

oC has a

shelf live of at least 4 weeks.

3.4.2 Cells Thawing

A vial of Spodoptera frugiperda (Sf-9) insect cells was taken out from the

liquid nitrogen (-196 oC). Then the vial was placed in a 37 oC water bath and gently

swirled until the cell was completely thawed. A bottle of Sf-900 II SFM medium

was removed from the cold room and placed in a 37oC water and was allowed to

acclimate to room temperature before using. Laminar flow hood was turned on and

the working surface was wiped down with 70 % ethanol. Two 25 cm2 T-flasks were

pre-wetted by coating the adherent surface with 4 ml of fresh media. The 1 ml of cell

suspension was directly transferred into a centrifuge tube (containing 4ml of media)

and 100 µl was then taken out for viability determination. The suspension was

centrifuged at 1000 rpm for 5 min to remove DMSO. The pellet was collected and

resuspended in 1 ml of fresh media and divided into the two T-flasks. The T-flasks

were transferred to a 27 oC incubator for cells attachment and propagation.

66

3.4.3 Cells Maintaining

Sf-9 insect cells were maintained in 25 cm2 tissue culture flasks in a

humidified 27oC incubator. Regular passage of cells was performed every 2 days

with fresh medium by gently dislodging the confluent monolayer, transferring of a

fraction of the suspension to sterile culture flasks, and adding of fresh medium to a

final cell density 5x105 cells/ml with the viability above 90 %. Cell viability was

determined using the trypan blue exclusion test and cell counts were performed using

an inverted microscope.

3.4.4 Cells Freezing

The cells were counted using a hemacytometer. Cells should be 90 % viable

and 80-90% confluent. It was recommended to freeze down several vials as low a

passage number as possible at a cell density 1x107 cells. Sterile cryovials were set up

in ice and labeled. The cells were centrifuged at 1000 rpm for 10 min at room

temperature. The supernatant was removed. The cells were resuspended to a given

density in the freezing medium (90 % FBS and 10 % DMSO). 1 ml of cell

suspension was transferred to sterile cryovials. The vials were placed at 4oC for

15min, -20oC for 30 min and –180oC for 60 min. The vials were stored in liquid

nitrogen.

3.4.8 Adapting serum contain culture to serum free culture

At the next routine passage, the cell was transferred into medium consisting

of 75% serum and 25% serum-free components. The cells were allowed to become

confluent. At the next passage, the cells were transferred into a medium consisting of

an equal mixture of serum and serum free components. The cells were allowed to

grow to confluent. If the cells grew slowly, previous step was repeated. At the next

passage, the cells were transferred into a medium mixture consisting of 75% serum-

free and 25% serum components. Let the cells grew until confluent. The cells were

67

then transferred into 100% serum free medium. The cells would take another two to

three passages to grow to optimum densities.

3.4.9 Adapting Monolayer Cells to Suspension culture

Insect cells were dislodged from the bottom of flasks. Confluent cells from

two units of 75cm2 T-flask would be sufficient to initiate a 50ml suspension culture.

After cell count, cell suspension was diluted to 5x105cells/ml in serum free growth

medium. Suspension culture was maintained in shaker flask or spinner flask. Stirring

rate for shaker flask and spinner cultures was started at 100rpm and 75rpm. The cells

were subcultured when the viable cell density reached 1-2x106cells/ml. Stirring rate

was increased by 5-10rpm with subsequent passage until constant stirring speed

reached 130-150rpm for shaker culture and 90-100rpm for spinner culture. If the

viability dropped below 75%, stirring speed would be decreased by 5rpm for one

passage till the culture viability recover to >80%

3.4.10 Maintaining suspension culture

Insect cell culture was incubated at 27oC, non CO2 aerated incubator for both

adherent and suspension cultures. Generally, suspension culture were subcultured

twice weekly; centrifuged at 1000rpm for 5min, and resuspended in fresh medium

once every 3 weeks. For each subculture, confluent cells (2-3x106cells/ml) were

diluted to 5x105cells/ml in serum free medium. Stirring rate maintained at 130rpm-

150rpm for shaker and 90rpm-100rpm for spinner flask. Suitable volume for

respective flask size was shown in Table. 3.1 The caps of flask were loosen about ¼

to ½ of a turn to main the aeration of cultures.

Table 3.1: Suitable culture volume

Flask Size (ml) Shaker Flask Culture Volume (ml)

Spinner Flask Culture Volume (ml)

125 250 500

1000 3000

25-50 50-125

125-200 200-400 400-800

50-100 150-200 200-300

300-1000 2000-3000

68

3.5 Wild Type and Recombinant Baculovirus

3.5.1 Virus Propagation

A 25cm2 T-flask was seeded with cells with a density of 5x105 cells/ml and

higher than 90 % viability. The cells were inoculated with virus stock by simply

adding 20 µl inoculum to the cells. The infectious culture was incubated at 27oC for

7 days, and was visually examined daily to ensure the cells were well infected. To

collect extracellular virus, the infected cells were transferred to a centrifuge tube and

spinned at 1000 x g for 15 min. The supernatant was transferred to a fresh centrifuge

tube and stored at 4oC. For long term storage, virus inocula should be kept at -80oC.

The virus stock concentration was determined by end-point dilution.

3.5.2 Virus Titration (End-Point Dilution)

Tenfold serial dilutions of the virus stock were prepared. Dilutions of 10-5,

10-6, 10-7, 10-8 should be appropriate in most cases. The cells with the viability

higher than 90% were diluted in a concentration of 1 x105 cells/ml with fresh

medium. 10 µl aliquots of each virus dilution was mixed with 100 µl aliquots of the

cell suspension, and seeded into 96 wells plate. 4 wells were seeded with 100 µl of

cells, as uninfected controls. The plate was incubated at 27oC. To avoid

dehydration, the plate was incubated in a humidified environment. The plate was

sealed in a plastic bag with damp paper towel. The plate was incubated for one

week. Each well was examined for virus replication.

69

Figure 3.1: Virus Titer Procedures – End Point Dilution

3.5.4 Generating Pure Recombinant Virus Stocks (End Point Dilution)

Sf-9 insect cells with the viability of higher than 90 % were diluted with fresh

medium to a concentration of 5 x105 cells/ml. A tenfold serial dilutions of the virus

were prepared and the dilutions of 10-6 and 10-7 were appropriate for most stocks.

10µl of each dilution was mixed with 100 µl of the cell suspension and seeded into

each well of a 96 wells plate. For each dilution at least 46 replicates were tested.

Therefore 2 tenfold dilutions were tested in one plate including 4 wells for uninfected

controls. Plate was incubated at 27oC in humidified environment. Each well was

examined daily for virus replication and progress of infection. All wells with sign of

infection were scored as positive and tested for product gene expression using

Enzyme Linked Immunosorbent Assay (ELISA). Samples that gave high levels of

recombinant protein production yield were then selected to undergo the purification

process twice further or until the recombinant protein level reached a constant yield

27oC The plate is sealed in plastic bag. Incubate for 7

Examine the virus replication Calculate TCID

Cell Concentration 1 x 105 cells/ml

10-5, 10-6, 10-7, 10-8 dilution of virus

100ul cell suspension 10ul virus

10-5 Dilution 10-6 Dilution

10-8 Dilution



1 2 3 4 5 6 7 8 9 10 11 12

Replicate

10-7 Dilution

control

70

provided that other parameters remain unchanged for every purification round.

Finally the high purity recombinant virus was amplified to generate large stock.

3.6 Optimization of Recombinant Human Transferrin (rhTf) Expression

3.6.1 Optimization of rhTf Expression in Monolayer Culture

All experimental works were conducted at Sf9 cells viability of at least 90%.

This was to reduce any variation due to non viable cells. Each 25 cm2

T-flask was

seeded with 4 x 106

Sf9 cells. When the cells had attached to the surface, the spent

medium was removed. Virus innoculums at different MOI ranging from 1-100 MOI

were tested. After 1 hour, the innoculum was removed and replaced with 5 ml fresh

SF-900 II medium. 100 �l of each flask sample was collected every 2 days for cells

counting and undergone an ELISA analysis for rhTf expression. For the expression at

different seeding densities, a range between 0.8-5.6 x106

Sf9 cells/ml was studied

using 5 MOI viruses. For the expression at different time of infection, the virus

innoculum was introduced only at certain times post culture. A range between 0-6

days time of infection were investigated.

3.6.2 Medium Screening

Based on literature reviews, 13 nutrients were selected for screening. Nutrients

selected were D-fructose, D-glucose, Maltose, L-arginine, L-cysteine, L-glutamine,

L-lysine, L-methionine, L-serine, L-threonine, L-tyrosine, L-valine, and Lipid

mixtures 1000x (already in solution form).

The first step was the preparation of 10.0 ml of concentrated nutrient

(excluding lipid mixtures) solutions using the original Sf900-II SFM as a diluent.

The concentration for each nutrient used for the preparation of different medium

compositions was 25g/L. 33 different combinations of nutrients at two levels of

added concentrations were generated using Statistica software.

71

1.0 ml each of the 33 designed medium compositions was prepared in 2 x 24-

well plates. Each medium composition was prepared by adding certain volumes of

the concentrated nutrients (25g/l each) into each well of the 24-well plate. Sf900-II

SFM was added to make up the total volume of 1.0 ml. Each of the medium

composition was observed for any physical changes. A total of 4 x 105

cells were

inoculated into each well of another 2 x 24 well plates. The cells were incubated for

2 hours to form attached monolayers after which the old medium was removed and

replaced with the designed medium. Virus innoculums of 0.36 MOI were added into

the monolayer and incubated at 27oC. Samples were harvested at day 4 and 10 post

infection by centrifuging the infected culture at 1000 rpm. The samples were

analyzed using SDS-PAGE and ELISA. The screening was repeated 3 times and the

results were analyzed using Statistica (Statsoft, v. 5.0).

3.6.3 Medium Optimization in Suspension Culture

After the medium screening was completed, three dominant factors had been

identified (lipid mixtures 1000x, glutamine and glucose) (see section 4.4.2). These

factors were further optimized in the suspension culture. The first step was the

preparation of 10.0 ml of concentrated glucose and glutamine solutions using the

original Sf900-II SFM as a diluent. The concentration for each nutrient was 25g/l. A

series of 17 central composite design (CCD) matrix experiments were conducted

which incorporated eight 2-level factorial experiments, six extreme level

experiments, two experiments at the center point and one control. Experiments were

done in duplicates to obtain the error regions for rhTf concentration. 1.0 ml each of

the 17 designed medium compositions was prepared in 2 x 24 well plates. A total of

8 x 105

Sf9 cells were inoculated into each well of another 2 x 24 well plates.

The cells were incubated for 30 minutes for them to settle to the bottom of the

wells after which the old medium was removed slowly and replaced with the

designed medium. The plates were placed on a shaker and rotated at 125-130 rpm.

After two days in culture, virus inoculums of 15 MOI were added directly into the

Sf9 cell culture. Cells density of each well was determined prior to infection. Only

72

20 �l of cell suspension was aliquoted for each cell counting. This was to maintain

the culture in suspension. Samples were harvested at day 8 post infection by

centrifuging the infected cultures at 1000 rpm. Samples were kept in appendorf tubes

at -78oC for ELISA analysis. The results were analyzed using Statistica (Statsoft, v.

5.0).

3.7 Response Surface Methodology, RSM (Method of Steepest Ascent)

A Taguchi design array (3 parameters and 3 levels) was generated from

Statistica (Statsoft, v. 5.0) and used to generate real and coded variables. The original

operating condition, although not part of the Taguchi array, was also included in the

regression analysis, since that data point was available. Increment for each variable

was chosen first. The increment size could be as large as maximum concentration of

added nutrient. It was presumed that the calculated optimum values would center

around the maximum values of the Central Composite Design experiment. Therefore,

small increment would suffice.

The response y was calculated using the regression coefficients which were

obtained from the medium optimization experiment. Regression analysis was

performed using Microsoft Excel (Tools-Data Analysis-Regression) with x1 through

x3 as the independent variables, and y as the dependent variable. Note that coded

(i.e. normalized) variables x1, x2 and x3 were used for the regression analysis

instead of real values Gln, Gluc, and Lip.

The vectors were the regression coefficients obtained after regression analysis

was performed. Magnitude of the vector was calculated. Since coded variable x3, had

the largest magnitude, the increment of its uncoded value Lip was chosen. The

increments of the other two parameters were calculated, based on the direction of

steepest ascent. Using ratios, based on the direction of steepest ascent, increment in

x1, x2 and x3 was calculated and converted to Gln, Gluc and Lip.

73

The response y was marched "uphill" from the previous middle point until y

started to decrease. RSM was repeated, using the current maximum value as a new

operating/mid point. This time, smaller increments around the operating point was

used, since the optimum value was closer. It was however, not necessary to exactly

center around the operating point, for convenience. Optimum value was obtained

when the response no longer increased.

3.8 Optimized Expression of rhTf

3.8.1 Preparation of Optimized Medium

Optimized Medium is SFM900II added with 2211.2mg/ml of Glutamate,

1291.95mg/ml of Glucose and 0.64% (v/v) of lipid mixture 1000x. (Refer to section

4.4.3.2) Powder of Glutamate and Glucose were dissolved in SFM900II and filtered

with nitrocellulose membrane, 0.22µm. Original stock of glutamate and glucose was

prepared in 25g/l respectively. A define volume of optimized medium was prepared

by mixed the calculated volume of glucose solution, glutamate solution, lipid mixture

and SFM900II.

3.8.2 Adapting suspension culture in SFM900II to optimized medium

When the suspension culture has reached more than 2x106cells/ml, split the

culture and added in equal volume of optimized medium. Then, wait another 2-3

days for the cells to become confluent again. Split the cells again and added in equal

volume of optimized medium. Repeat this for few passages. Finally, the suspension

culture was centrifuged then transferred into 100% optimized medium. Always seed

the cell at densities 1.0 x 106 cells/ml when optimized medium was used as the

growth medium. Low seeding densities would cause the denaturation of cells.

74

3.8.3 Expression of rhTf

Suspension culture adapted to optimized medium was seeded at

1x106cells/ml. When the suspension culture reached 1.6x106cells/ml, the culture was

resuspended in fresh optimized medium. The culture was infected with amplified

rhTf baculovirus at day 2, at MOI 15. The infected culture was harvested at day 8 or

day 6 post infection. The product was harvested by centrifuging at 500g (4000rpm),

5min.

3.9 Characterization of rhTf

3.9.1 Sodium Dodecyl Sulfate - Polyacrylamide Gel Electrophoresis

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)

was performed using a Mini-protean II apparatus (Bio-Rad Laboratories, Melville,

NY). Two phases of polyacrylamide gel were prepared in advance. The gel was

divided into two parts, a stacking gel for the concentration of the protein samples

before separation and a separating gel for the separation of the protein samples. The

working solutions for both phases are shown in appendix 1,2,3..

The separating gel was mixed well, poured between two plates and overlaid

with water to keep the gel surface flat and left to polymerize for 1 hour. After the gel

had polymerized, a distinct interface appeared between the separating gel and the

water. The water was rinsed off with fresh distilled water and the stacking gel was

prepared. The stacking gel was then poured on the top of the separating gel. A comb

was carefully inserted into the top of the stacking gel, so that no bubbles would be

trapped on the ends of the teeth. The gel was allowed to polymerize for 30 min.

Once polymerized, the gel was attached to the electrode assembly of Bio-Rad

Mini-Protean II Gel System and inserted into an electrophoresis tank that was filled

with 1 x Tris-glycine electrophoresis buffer, afterwards. Then the comb was

removed. Subsequently, the sample solution (combine protein sample, 20 �l and 1x

A

75

sample buffer, 20 �l and heated at 100oC for 5 min) as well as the protein marker

(consisting of 7 precisely sized recombinant proteins, ranging from 15 kDa to

220kDa), were introduced into the wells on the stacking gel using a Hamilton

syringe. The electrophoresis was carried out on a vertical slab gel using 9%

acrylamide gel at a constant voltage of 100 V for 90 min at room temperature.

Following electrophoresis, protein bands were visualized using silver stain.

3.9.1.1 Silver Staining

The polyacrylamide gel was soaked in 5:1:4 ratios of methanol, acetic acid

and water for at least 1 hour with 2 or 3 changes of the solution with gentle shaking.

The gel was then soaked for 30 min with water with at least 3 changes of water.

Solution A (0.8 g silver nitrate in 4 ml distilled water), solution B (21 ml 0.36 %

NaOH mixed together with 1.4 ml of 14.8 M ammonium hydroxide) and solution C

(solution A was added to solution B with constant stirring and later water was added

to make up a total volume of 100 ml) were freshly prepared. The gel was stained in

solution C for 15 min with gentle and constant agitation. After rinsing the gel twice

in deionized water, the gel was placed in solution D that was freshly prepared (0.5 ml

1% citric acid was added to 50 µl 38 % formaldehyde and later water was added to

make up a total volume of 100 ml) and was shaken until the bands appear.

Development was stopped with 1 % acetic acid.

3.9.1.3 Coomassie Blue Staining

Coomassie blue staining and destaining solution was prepared. The

formulation was shown in appendix 4. The gel was soaked in staining solution

enough to cover the whole gel and agitated on orbital shaker for 15minutes. Longer

staining is required if recycled staining solution was used. After that, the solution

was discarded, and the gel was rinsed with distilled water. The gel was destained

overnight using destaining solution. Frequently changing the destaining solution

would help to destain the gel faster.

76

3.9.2 Western Blot

Proteins were separated on SDS-PAGE gels (9 % polyacryamide, 90 min,

100V) and electroblotted (4oC, 1 hour and 100 V) onto 0.45 µm nitrocellulose

membrane using a Trans-Blot Electrophoretic Transfer Cell. The transfer buffer

consisted of 15.6 mM Tris and 120 mM glycine and 10 % methanol, pH 8.1 – 8.4.

Membranes were incubated in blocking buffer (Phosphate Buffer Saline (PBS),

pH7.4 containing 5 % skimmed milk) at 4oC overnight. The following day,

membranes were washed 3 x 10min with washing solution (PBS containing 0.05 %

Tween 20). Membranes were incubated with primary antibody - Goat anti-Human

Transferrin (Bethyl Laboratories Inc, Texas) in blocking solution with gentle

agitation for 2 hours. This was followed by 3 x 10 min washings with washing

solution. Membranes were incubated as before with Horseradish Peroxidase (HRP)-

conjugate secondary antibody, that is Goat anti-human transferrin-HRP conjugate

(Bethyl Laboratories Inc, Texas) in blocking solution. This was followed by 3 x

10min washings with washing solution. Bound antibody was detected using TMB

(3,3’,5,5’-tetramethylbenzidene) Stabilized Substrate for HRP (Promega, Madison,

WI).

3.9.3 Enzyme Linked Immunosorbent Assay

Direct Sandwich Enzyme Linked Immunosorbent Assay (ELISA) were

performed in 96 well microtiter plates (TPP, Switzerland) which were coated with

primary antibody and incubated at 4oC overnight. The following day, the plate was

washed with washing solution (Tris-buffered Saline (TBS), pH 8.0 containing 0.05%

Tween 20) 3 times. The plate was blocked with blocking solution (TBS, pH 8.0

containing 1% BSA) for 30 min and washed with washing solution 3 times. The

plate was subsequently incubated with serial dilutions of standards - Human

Reference Serum (Bethyl Laboratories Inc, Texas) and samples in sample/conjugate

diluent (TBS, pH 8.0 containing 1% bovine serum albumin (BSA) and 0.05% Tween

20) for 60 min and washed with washing solution 5 times. Next, the plate was

incubated with HRP-conjugated secondary antibody in sample/conjugate diluent for

60 min at 37oC and washed with washing solution 5 times. Color development by

77

the enzyme substrate reaction was performed by adding to each well 100 �l of equal

volumes of Trimethyl Benzene (TMB) Peroxidase Substrate and Peroxidase Solution

B (H2O2). After 5-30min, the reaction was stopped with 100 �l of 1 M Phosphoric

Acid (H3PO4). The absorbance at 450 nm was determined.

3.10 Characterization of and nutrients consumption and substances release.

3.10.1 Analysis of glucose, lactic acid and glutamine

Buffer concentrate from YSI was reconstituted in distilled water, 500ml per

packet. Buffer was poured into the supply bottle. Then, the bottle lid was replaced

and the electrical lead was reconnected. Another electrical lead from the sensor in

the lid was assembly onto the new YSI calibrations standard. Then, the lid was

screwed on and the calibrator was placed in the instrument compartment. The

electrical lead in the lid would be rinsed thoroughly with distilled water whenever the

calibrator was changed. YSI immobilized enzyme membrane was installed by gently

assembled onto the probe face and then returns the probe to the sample chamber.

When the instrument was in run mode, the buffer pump will operate through

two cycles and the instruments will “initialize the baseline current” and ready to

calibrate. Analysis of sample was started by placing sample of about 500µl in

appendorf tube at station 2 once a stable calibration was established. The unit was

self-calibrates every 15 minutes or every 5 samples. The samples were diluted with

distilled water if the sample was out of the detection range. The detection range was

varied with different standards.

78

Table 3.2: Specification of YSI calibrator

Standards Calibration Point Detection Range

D-Glucose 2.5g/L 0-9g/L

L-Lactate 0.5g/L 0-2.67g/L

L-Glutamate 5.00mmol/L 0-10mmol/L

L-Glutamine 5mmol/L 0-8mmol/L

3.10.2 Ammonia test

Randox’s kit was used to check ammonia content. Enzymatic UV method

was applied. Ammonia combines with �-ketoglutarate and NADPH in the presence

of glutamate dehydrogenase (GLDH) to yield glutamate and NADP+. The

corresponding decrease in absorbance at 340nm is proportional to the plasma

ammonia concentration.

�-ketoglutarate + NH3+ NADPH →GLDH glutamate + NADP+……(3.1)

Each vial of reagent 1 of the kits (0.26mM NADPH/3.88mM �-ketoglutarate)

was reconstituted with 5ml of 0.15M triethanolamine buffer, pH8.6. 0.1ml of water

as blank, standard and samples were pipetted into different cuvettes. Duplicate set

was prepared. Then, 1ml of reagent 1 was added to the cuvettes. The mixture was

mixed and allowed to stand for 5 minutes. The absorbance of the mixture was read

at 340nm. Then, 10µl of GLDH was added to each cuvette . The solution was

mixed and left to stand for 5 minutes. Finally, the absorbance at 340nm was read

once more. Concentration of ammonia is

294tan

xAA

AA

blankdards

blanksample

−−

µmol/l ……(3.2)

blankA = Absorbance (1) for Blank – Absorbance (2) for Blank ……(3.3)

dardsA tan = Absorbance (1) for Standard – Absorbance (2) for Standard …(3.4)

sampleA = Absorbance (1) for Sample – Absorbance (2) for Sample ……(3.5)

79

3.11 Protein Analysis Techniques

3.11.1 Bicinchoninic Acid (BCA) Assay

BCA protein assay kit was used to quantify total protein. BCA working

reagent was prepared by mixing reagent A with reagent B at ratio 50:1. Sufficient

volume of working reagent was prepared for duplicate set of standards and samples.

Serials dilution method was used to prepare the standards. 0.05ml of each standard

and unknown sample replicate was placed into an appropriately labeled test tube. 1.0

ml of the working reagent was added to each tube and well mixed. For working

range between 20-2,000 µg/ml, test tubes was incubated in water bath at 37°C for 30

minutes; working range between 5-250 µg/ml, was incubated in water bath at 60°C

for 30 minutes. After that, all the tubes were cool to room temperature. With the

spectrophotometer set to 562 nm, the reading was auto-zeroed with cuvettes filled

only with water. Subsequently, the absorbance of all the samples was measured

within 10 minutes

3.12 Recombinant _1,4-Galactosyltransferase Detection

3.12.1 Thin Layer Chromatography

A typical incorporation mixture contained the following in a final volume of

0.1 ml: 5 �mole of Tris-HCl, pH 7.4, 0.04 �mole of UDP-Gal, 2.0 �mole of glucose,

4 �mole of MnCl2, 0.5 �mole of UTP, 0.14 �mole of �-lactalbumin and sample �1,4-

galactosyltransferase. After 30 min of incubation at 37oC, the reaction was stopped

by adding 0.2 ml of 0.3 N Ba(OH)2 to ice-cooled mixture. This was neutralized with

1.5 volumes of a 5% solution of ZnSO4.7H2O and the precipitate was removed by

centrifugation. The supernatant was analysed using thin layer chromatography.

Silica plate (Whatmann, 200 �m) was marked with a light straight line

parallel to the short dimension of the plate, about 1 cm from one end of the plate. A

80

few small marks were made lightly perpendicular to this line to serve as a guide for

placing the substance spots. The substances were loaded on the plate and developed

in a trough chamber containing mobile phase: n-butanol-acetic acid-diethyl ether-

water (9:6:3:1) to a depth of about 5 mm. The migration time was about 120 min.

The chromatogram was freed from the mobile phase and dipped in the solution

containing 8 ml concentrated sulfuric acid, 0.5 ml anisaldehyde, 85 ml methanol and

10 ml glacial acetic acid for 2 seconds. After drying for several minutes in cold air,

the plate was heated to 120oC for 15 min.

3.12.2 Lectin Binding Assay

25 mg of asialofetuin was dissolved in 1 ml of 0.2 M sodium phosphate

buffer (pH 4.5) containing 0.1 M citric acid. 0.4 units (2.2mg) of bovine �-

galactosidase was added to the mixture. The mixture was incubated for 72 h at 37oC

to remove galactose residues. The sample was diluted at least 20 times with 0.1M

sodium phosphate-buffered saline (0.15 M NaCl, pH 7.2) containing 1 mM CaCl2

and 1 mM MnCl2 and concentrated using Amicon Model 8010 UF with MWCO of

10 000. This procedure was repeated three times to remove any remaining sugars

which had been released from protein by the enzyme treatments. The protein

produced was asialoagalactofetuin.

Each well of the 96 wells plate was coated with 100 �l of the

asialoagalactofetuin (1 µg/ml in 0.05 M Sodium Carbonate, pH 9.6 containing 2%

glutaldehyde) at room temperature for 1 hour, washed 3 times with washing solution

(PBS containing 0.05% Tween 20, PBST) and then blocked with 1% BSA in PBST

at room temperature for 1 hour. The plate was washed and the enzyme reaction was

started by adding to each well 100 µl of enzyme-donor substrate mixture (10 mM

MnCl2, 0.1% BSA, 0.32 mM UDP-Gal and sample �1,4-GalT) in 30 mM Mops

(pH7.4). The plate was incubated at 37oC for 1 hour. The reaction was stopped by

discarding the reaction mixture. The plates were washed and incubated with

peroxidase-labeled RCA 1 (0.16 mg/ml in PBST containing 1% BSA) at 37oC for 90

min. Color development by the enzyme substrate reaction was performed by adding

to each well 100 �l of equal volumes of TMB Peroxidase Substrate and Peroxidase

81

Solution B (H2O2). After 5-30 min, the reaction was stopped with 100 �l of 1M

Phosphoric Acid. The absorbance at 450 nm was determined.

3.13 Native Uridine-5_-diphosphogalactose (UDP-Gal) Level

3.13.1 UDP-Gal Extraction

Sf-9 cells or infected Sf-9 cells (about 1 x 106 cells) were collected by

centrifugation (1000 x g, 15 min at 4oC). Pellets were washed with PBS buffer,

pH7.4. Cells were lysed in ice-cold 75% ethanol (300 �l) by freeze-thawing and

homogenizing. Soluble fractions were obtained by centrifugation (16 000 rpm x

10min at 4oC). Supernatant was filtered through 10 000 MWCO membranes.

3.13.2 Reverse Phase High Performance Liquid Chromatography (RP-HPLC)

Analysis

RP-HPLC elution was carried out at 1ml/min and the column was kept at

30oC. UDP-Gal was detected by absorbance at 260 nm. The ion pair RP-HPLC was

carried out using a LC-10Dvp HPLC system and a NovaPac C18 column (3.9 x 150

mm, 4 �m) with a NovaPac C18 guard cartridge. The following two solvents were

used as eluents: 5 mM tetrabutylammonium sulfate (TBAS) – 50 mM ammonium

phosphate, pH 5.0 (E1) and 5 mM TBAS-methanol (E2). A portion of the cell

extract was injected into a NovaPac C18 column equilibrated with a mixture (98:2,

v/v) of E1 and E2.

3.14 Coexpression of Recombinant Human Transferrin and ββββ1,4-

Galactosyltransferase

Two Sf-9 insect cell cultures were infected with AcMNPV-hTf. One of the

cultures was coinfected with recombinant baculovirus carrying the gene for �1,4-

82

GalT (in vivo). For the in vitro analysis, 2.0 mU/ml commercial mammalian GalT

and 0.32 mM commercial UDP-Gal were added to the harvested AcMNPV-hTf

supernatant. There were three negative controls in this experiment which were Sf-9

cell culture infected with AcMNPV-�1,4-GalT, uninfected Sf-9 insect cell culture

and harvested uninfected Sf-9 cell culture mixed with 2.0 mU/ml commercial

mammalian GalT and 0.32 mM commercial UDP-Gal. All samples were harvested

at time 24 hours PI. As for the time course upon coinfection between recombinant

baculovirus hTf and �1,4-GalT, the medium from each of the coexpressed cell

culture supernatants were collected at time intervals of every 24 hours PI until 120

hours PI.

Each well of the ELISA plate was coated with 100 �l of the glycoprotein

(1µg/ml in 0.05M Sodium Carbonate, pH 9.6 containing 2% glutaldehyde) at room

temperature for 1 hour, washed 3 times with washing solution (PBST) and then

blocked with 1% BSA in PBST at room temperature for 1 hour. The plate was

washed and the enzyme reaction was started for the in vitro galactosylation samples

by adding to each well 100 µl of enzyme-donor substrate mixture (10 mM MnCl2,

0.1% BSA, 0.32 mM UDP-Gal and 2.0 mU/ml of �1,4-GalT) in 30mM Mops (pH

7.4). The plate was incubated at 37oC for 1 hour, and then the reaction was stopped

by discarding the reaction mixture. The plates were washed and incubated with

peroxidase-labeled RCA 1 (0.16 mg/ml in PBST containing 1% BSA) at 37oC for 90

min. Color development by the enzyme substrate reaction was performed by adding

to each well 100 �l of equal volumes of TMB Peroxidase Substrate and Peroxidase

Solution B (H2O2). After 5-30 min, the reaction was stopped with 100 �l of 1M

Phosphoric Acid. The absorbance at 450 nm was determined.

3.15 Purification

3.15.1 Hydropbobic interaction Chromatography

Slurry of Phenyl Sepharose 6 fast flow was prepared by decanting 20%

ethanol solution and replacing it with water or other low ionic strength buffer in a

83

ratio of 50–70% settled gel to 50–30% packing solution. The gel was de-gassed

using a vacuum pump filter system.

Column was flushed with distilled water to eliminate air from the column

dead spaces. A few centimeters of water was allowed to remain in the column and

then the column was closed. The slurry was poured into the column in one

continuous motion using a glass rod held against the wall of the column. The rest of

column was filled with distilled water until an upward meniscus was formed at the

top. The column was packed using a flow adaptor. The flow adaptor which was

connected to the pump was flushed and fully filled with distilled water. After

removing all bubbles, the pump was stopped and the adaptor was inserted into the

top of the column at an angle until it reaches the gel slurry. The adaptor o-ring was

kept tight to give a sliding seal on the column wall. The bottom outlet of the column

was opened and the pump was set at the desired flow rate. Ideally, Phenyl Sepharose

6 Fast Flow matrices are packed at a constant pressure of 0.15 MPa (1.5 bar) or flow

rate less than 400 cm/h. The packing flow rate was maintained for 3 bed volumes till

a constant bed height was reached. The pump was closed, the bottom outlet was

close and the adaptor was repositioned and locked on the surface of the matrix. The

column is ready for used when the bed medium is stable.

The column was equilibrated with starting buffer (1.2M Ammonium

Sulphate/ 0.4M Sodium citrate buffer, pH6) for 3 column volumes. Sample was

filtered through 0.45 µm membrane, mixed with 2x starting buffer (2.4M

Ammonium Sulphate/ 0.8M Sodium citrate buffer, pH6), and loaded into column

using pump. Then, 3 column volumes of starting buffer was used to wash away

unbound protein. Elution buffer is the mixture of starting buffer and deionized water

and percentage of the elution buffer was the same as the percentage of starting

buffer. For each step elution, three to four column volumes of elution buffer was

applied. Gradient elution was monitored using 2 pumps which drew deionized water

to starting buffer and to the column after homogenously mixing the solution. All the

equilibrating and operating flowrates were the same. On the other hand, flowrates,

steps elution and gradient elution and loading capacity were varied as a strategy of

optimization.

84

The column was washed with three column volumes of deionized water at

flowrates of 4ml/minute, and re-equilibrate with starting buffer after each run. For

the cleaning in place, precipitated proteins was removed by washing the column with

80ml of 1M NaOH solution at a flow rate of 1.2-1.4ml/min, followed immediately

with 40-60ml of deionized water and re-equilibrated with 100ml of starting buffer.

Strongly hydrophobically bound proteins was removed by washing the column with

80ml of 70% ethanol, followed by with water and re-equilibrated with starting

buffer. The column was stored in 20% ethanol in distilled water at 4oC when not in

used.

3.15.2 Dialysis

Dialysis was used for desalting, buffer exchange and removal of small

molecular weight contaminants in samples. Snake SkinTM pleated dialysis tubing

with 10,000 nolecular weight cut off was used. The already-open tubing was pulled

from the stick to the required length. The amount of tubing can be calculated using

3.7ml sample per cm of dry tubing. 2-3 inches of one end of the tubing was briefly

dipped into water and tied tightly in the wetted end of the tubing. Sample was added

into the open end of the tubing. Then, one knot was tied securely in the other open

end. Finally, the tubing was immersed in 2 liters 20mM Tris/HCl buffer, pH 8.5,

with constant stirring for 24 hours. The buffer was changed with fresh one after 12

hours.

3.15.3 Ion Exchange Chromatography

Matrix Q-Sepharose fast flow was settled in starting buffer (20mM Tris

Buffer, pH8.5) and packed as mentioned in 3.15.1. The column was equilibrated with

starting buffer for 3 column volumes. Sample after dialysis in starting buffer was

loaded into the column using a pump. Then, 2 column volumes of starting buffer was

used to wash away unbound protein. The elution method was a combination of

gradient and steps elution. It was started with a linear gradient elution where the

percentage of buffer B (0.5M NaCl/ Tris Buffer, pH8.5) was increased from 10% to

85

20% in nine column volumes and followed by step elution with 20% of buffer B and

lastly 100% of buffer B. Regeneration of Q-sepharose fast flow was performed by

washing 1M NaCl and followed by re-equilibrating in 100ml of starting buffer at

flow rates of 4–5 ml/min Column was stored in 20% ethanol in distilled water at 4oC.

3.15.4 Batch Purification

200µl anion exchange matrix was transferred into appendorf tube. Appendorf

tube was centrifuged at 500 × g for 3–5 min to sediment the matrix. The supernatant

was discarded carefully. The matrix was washed five times with 3 matrix volumes of

equilibration buffer. For each time, the slurry was centrifuged at 500×g for 3–5 min

and the equilibration buffer was discarded carefully. 500µl of sample was added to

the matrix. It was estimated that 1ml of matrix could bind approximately 30mg of

protein. Sample after HIC and after dialysis was incubated in the matrix and agitated

gently on a shaker for 2 hrs at room temperature. After that, the appendorf tube was

centrifuged at 500 × g for 3–5 min to sediment the matrix. The supernatant was

collected, and the rhTf in the supernatant was determined using ELISA. Binding

capacity was calculated by minusing the rhTf in supernatant from the total loaded

rhTf.

CHAPTER 4

RESULT AND DISCUSSION

4.1 The Study of Sf9 Insect Cells Culture Growth Profiles

4.1.1 Fundamental Study of Sf9 Cells Growth (Monolayer)

Insect cell growth can be significantly improved by paying close attention to

the conditions used in the inoculum stages. Serum concentration, different type of

media, cell subculturing conditions, initial cell density and spent medium carry over

significantly influenced the growth kinetics of Sf-9 cells. Efficient operation of

insect cell culture requires full assessment of these factors which are expected to

have significant influence on the cell metabolic activity.

During cell infection, as cell division stops, other cellular activities such as

respiration still continue as does the transcriptional and translational machinery that

are being switched to viral multiplication and expression of its genes. It is therefore

necessary that cells are held in a healthy physiological state, free of nutrient

limitation, if high recombinant protein yield are to be achieved. Consequently,

comprehensive data are needed on the effects of these important environmental and

physiological parameters that can influence growth, metabolism, cell infection, viral

multiplication and recombinant protein expression in insect cells.

87

In the early fundamental work, a few parameters which control the growth

rate of Sf-9 cells culture including initial cell density, effect of cell subculturing

conditions and spent medium were investigated. For the mock- and recombinant

baculovirus infection, the interaction of the infection parameters especially MOI and

spent medium with the above culture parameters were also examined.

As shown in Figure 4.1, higher viable cell numbers were attained in the

media (TC-100 and SF900 II SFM) containing higher FBS concentration. Higher

viable cell density in insect cell culture (Luis Maranga et al., 2002) because FBS

could replace insect cell hemolymph as the source of vitamins, growth factors and

other undefined compounds.

88

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

0 2 4 6 8 10 12 14 16 18Time (Day)

Via

ble

Cel

l Den

sity

(x 1

0 5 ce

lls/

ml)

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

110.0

0 2 4 6 8 10 12 14 16 18Time (Day)

Via

ble

Cel

l Den

sity

(x 1

0 5 c

ells

/ml)

Figure 4.1: Sf-9 insect cells growth in monolayer culture at 3 different serum

concentrations. (a) TC-100 and (b) SF-900 II SFM. Error bars indicate ±S.D of

duplicates data.

(b)

0% serum 5% serum 10% serum

SF-900 II

TC-100 0% serum 5% serum 10% serum

(a)

89

Different types of media also resulted in different cell growth. In this study,

two types of media have been used, TC-100 insect medium and SF-900 II SFM. As

presented in Figure 4.2, Sf-900 II SFM support higher cell densities compared to TC-

100 regardless of whether the medium was serum enriched or not. Based on this

finding, Sf-900 II SFM was used for the rest of the experiments. However, even

though serum affected cell growth positively (Figure 4.2 (b) and (c)), serum free

media were used for the rest of the experiments as serum contained trace amount of

sugar nucleotides and enzymes which may interfere with hTf and �1,4-GalT assay.

The effect of seeding density was investigated at three different cell

concentrations i.e. at 0.20, 1.20 and 2.33 x 105 cells/ml respectively. As shown in

Figure 4.3, the lowest cell concentration resulted in the lowest maximum viable cell

number achieved. This observation is in contrast with Kioukia et al. (1995) which

found that the maximum cell number achieved was highest for the lowest density.

However, the maximum growth rate, µ , was similar in all three cell concentrations at

about 0.004 to 0.011 h-1 (Table 4.1).

90

0.0

10.0

20.0

30.0

40.0

50.0

60.0

0 2 4 6 8 10 12 14 16 18Time (Day)

Via

ble

Cel

l Num

ber (

x 10

5 cel

ls/m

l)

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

0 2 4 6 8 10 12 14 16 18Time (Day)

Via

ble

Cel

l Num

ber (

x 10

5 cel

ls/m

l)

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

110.0

0 2 4 6 8 10 12 14 16 18Time (Day)

Via

ble

Cel

l Num

ber (

x 10

5 cel

ls/m

l)

Figure 4.2: Sf-9 insect cells growth in monolayer culture for 2 types of media. (a)

without serum; (b) with 5% serum and (c) with 10% serum. Error bars indicate ±S.D

of duplicates data.

SF-900 II SFM TC-100



(a)

(b)

(c)

91

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

0 2 4 6 8 10 12 14 16Time (Day)

Via

ble

Cel

l Den

sity

(x 1

0 5 c

ells

/ml)

Figure 4.3: Sf-9 insect cells growth in monolayer culture for 3 different initial cell

density, i.e. 0.2, 1.2 and 2.33 x 105 cells/ml. Error bars indicate ±S.D of duplicates

data.

0.2 x 105

1.2 x 105 2.33 x 105

Initial Cell Density

92

Cell subculturing condition was also investigated. Three inocula at fixed

density of 1.6 x 105 cells/ml were seeded at three different phases i.e. early

exponential, late exponential and stationary phase. Early exponential phase

subculturing resulted in the fastest cell growth rate (0.014/h) compared to the other

two phases; while those from stationary phase were obviously unsatisfactory (growth

rate and maximum viable cell number were 0.006/h and 12.55 x 105 cells/ml) as

shown in Figure 4.4. In related work (Kiokia et al., 1995), it has been shown that

there were higher proportions of G1 and S phase cells in the early exponential than in

the other growth phases. It was reported that in insect cells, the resting phase is G2

where most cells are accumulated when nutrient is depleted (Fertig et al., 1990).

This could explain the observation that cells from the early exponential phase set off

faster and achieved the highest growth rate.

The effect of spent medium on cell growth was also investigated. Spent

medium carry over has also been considered as the factor that can affect cell growth.

Cells were inoculated at fixed density, 1.5 x 105 cells/ml as shown in Figure 4.5. The

effect on growth became more significant for the spent medium percentage of 100%

and 50% which resulted in reduction in growth rate and maximum cell number

compared to the negligible percentage (0%). The result is expected because cells

prefer to survive in a rich nutrient culture for their metabolism, kinetics, respiration,

viral multiplication and protein expression.

93

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

0 2 4 6 8 10 12 14Time (Day)

Via

ble

Cel

l Num

ber (

x 10

5 cel

ls/m

l)

Figure 4.4: Sf-9 insect cells growth in monolayer culture at 3 different

subculturing conditions, i.e. early exponential, late exponential and stationary phase.

Error bars indicate ±S.D of duplicates data.

Early Exponential Late Exponential Stationary

94

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

0 2 4 6 8 10 12Time (Day)

Via

ble

Cel

l Den

sity

(x 1

0 5 c

ells

/ml)

Figure 4.5: Sf-9 insect cells growth in monolayer culture at 3 different spent

medium carry over percentage, i.e 100%, 50% and 0%. Error bars indicate ±S.D of

duplicates data.

100 50 % 0%

Spent Medium Percentage:

95

Table 4.1: Growth Kinetics of Sf-9 Cells at Different Parameters

1Experiments

Maximum viable cell

density x 105

cells/ml

2 )1

(h

µ 3 )(htd 4 GI 5 )(_

htd

Final infectivity percentage

(%)

INSECT CELL CULTURE Fig. 4.1: Effect of serum (a) TC-100 without serum TC-100 with 5% serum TC-100 with 10% serum (b) SFM without serum SFM with 5% serum SFM with 10% serum

51.88 74.88 81.03

57.65 86.63

102.43

0.011 0.017 0.019

0.010 0.010 0.013

63.4 39.7 37.4

66.5 70.2 51.4

24.7 35.7 38.6

27.5 41.3 48.8

83.0 74.5 63.8

70.3 53.7 51.4

ND ND ND

ND ND ND

Fig. 4.2: Two types of media comparison (a) TC-100 without serum SFM without serum (b) TC-100 with 5% serum SFM with 5% serum (c) TC-100 with 10% serum SFM with 10% serum

51.88 57.65

74.88 86.63

81.03

102.43

0.011 0.010

0.017 0.010

0.019 0.013

63.4 66.5

39.7 70.2

37.4 51.4

24.7 27.5

35.7 41.3

38.6 48.8

83.0 70.3

74.5 53.7

63.8 51.4

ND ND

ND ND

ND ND

Fig. 4.3: Effect of initial density of seeding inoculum - 0.2 x 105 cells/ml - 1.2 x 105 cells/ml - 2.33 x 105 cells/ml

15.00 17.43 23.00

0.004 0.011 0.011

165.3 62.8 64.8

75.0 14.5 9.9

46.2 62.2 58.1

ND ND ND

Fig. 4.4: Effect of cell subculturing conditions - Early exponential - Late exponential - Stationary

19.25 16.25 12.55

0.014 0.010 0.006

49.6 67.4

121.7

10.9 12.8 8.4

55.6 52.1 94.0

ND ND ND

96

Table 4.1: Growth Kinetics of Sf-9 Cells at Different Parameters (continue)

1All experiments were carried out in 25cm3 T-flask.

2Maximum growth rate, )1

(h

µ = 12

12 lnlntt

XX−−

where X2 = viable cell number at t2

X1 = viable cell number at t1

3Doubling time, )(htd=

µ2ln

4Growth index, GI =

densitycellinitialdensitycellimummax

5Average doubling time, )(_

htd =

)ln

(

2ln

maxtGI

,

where tmax = time at the maximum viable cell density

6Spent medium was prepared by centrifugation of medium from a 12 day culture at death phase. The

low viability percentage, 18.5% indicated that most of the cells were lysed and the breakage generated

a lot of impurities in the suspension.

ND = Not Determine

1Experiments

Maximum viable cell

density x 105 cells/ml

2 )1

(h

µ 3 )(htd 4 GI 5 )(_

htd


(%) Fig. 4.5: Effect of 6spent medium carry-over Spent medium percentage: - 100 % - 50% - 0 %

5.45 12.10 15.35

0.006 0.013 0.017

118.7 54.9 40.3

4.4 9.8

12.5

111.8 72.8 39.5

ND ND ND

97

4.1.2 Sf9 cell growth in Shake flask (Suspension)

In suspension culture (Figure 4.7), the cells could achieve even higher density

(~9.0 x 106

cells/ml) than the monolayer cultures (Figure 4.6) (~7.0 x 106

cells/ml).

In monolayer cultures, maximum density did not necessarily indicate optimum

nutrients consumption. The highest density in monolayer cultures might point to

diffusion limitation rather than nutrient depletion. In suspension culture however,

nutrient capacity could be determined at a higher confidence level. Therefore, for

medium optimization, nutrients (sugars, amino acids and lipids) utilization by insect

cells could be assessed more accurately in suspension culture.

Figure 4.6: Growth curves of Sf9 monolayer culture in 25cm2

T-flask at different

seeding densities, SD. Volume of medium was 5 ml. Straight lines represent cell

density and dotted lines represent cell viability

98

Figure 4.7: Growth curves of Sf9 suspension culture in 250ml shake flask at

different seeding densities, SD. Volume of medium was 50 ml. Straight lines

represent cell density and dotted lines represent cell viability

4.1.3 Development of Sf9 Suspension Culture System in 24-well Plate

In this work, experiments were carried out to check whether cell culture

cultivation in a suspension form could be done at a smaller volume in 24-well plates.

Initially, Sf9 cells were cultured in 0.5 ml of SFM. The agitation was maintained at

130 rpm which was a moderate rotation. At higher than 150 rpm, the risk of medium

overspill to adjacent wells was high. Based on Figure 4.8, it could be seen that the

growth patterns were similar to Figure 4.7. Next, Sf9 cells were cultured in 1.0 ml of

SFM. In this experiment however, the cells could not propagate properly (Figure

4.9). The cells tend to clump and settle down to the bottom of the well centrally and

thus resulted in mass transfer problem. The only way to overcome this bottleneck

was to increase the agitation speed but this would lead to spillage problem. Another

99

Figure 4.8: Growth curves of Sf9 suspension culture in 24-well plate at different

seeding densities, SD. Volume of medium was 0.5 ml. Straight lines represent cell


Figure 4.9: Growth curves of Sf9 suspension culture in 24-well plate at different

seeding densities, SD. Volume of medium was 1.0 ml. Straight lines represent cell


100

interesting observation was that despite the presence of a few 24-well plates filled

with water as humidifiers, the losses in volume from evaporation were still

noticeably large. Losses of volume were recorded between 10 – 20 %v/v within 7

days of cultivation. It was also observed that evaporation led to increase in cell

concentration due to reduction of total volume.

Overall, the monolayer culture maintained high viability (>80%) only for a

short period of time. However, its exponential and dead phases were slower than the

suspension culture. In addition to that, the life span was short too. Suspension culture

remained at high viability (>80%) the longest among all cultures. The exponential

growth and death phases were faster, and the life span was longer than the monolayer

culture (Table 4.2).

Table 4.2: Comparison of Sf9 growth in T-flask, Shake flask, and 24-well plate

4.1.4 Growth Analysis

Growth rate constants of Sf9 cells are shown in Figure 4.10. For T-flask, the

optimum growth rate was 0.016 hr-1

at the seeding density of 0.8 x 106

cells/ml. For

other seeding densities of T-flask, the growth rates stayed within close vicinities. For

shake flask, the optimum growth rate was at the seeding density of 0.4 x 106

cells/ml

and this figure stayed in the vicinity of 0.014 hr-1

for the other seeding densities. The

growth rate of Sf9 cells in a 24-well plate at 0.5 ml SFM (24well (b)) was slightly

lower than the T-flask and shaker. In general, the growth rates were stable suggesting

that the growth of Sf9 was not really affected by its seeding density.

101

For 24-well plate at 1.0 ml SFM (24well (a)) however, the Sf9 growth rate

was greatly affected by its seeding density. At 0.4 x 106

cells/ml, the growth rate

constant was 0.012 hr-1

while at 1.6 x 106

cells/ml, the value was 0.003 hr-1

. This

dramatic drop may be the result of mass transfer problem when using 1.0 ml SFM in

each well of the 24-well plate as explained in section 4.1.3.

The doubling times of Sf9 in various cultivators and at different seeding

densities are shown in Figure 4.11. A healthy Sf9 cell doubled in about 48 hours. In

summary, the doubling times for Sf9 cells cultured in T-flask, shake flask, and

24well(b) were close to 50 hours. In 24-well(a) however, the doubling time

Figure 4.10: Growth rate constants of Sf9 in various cultivators and at different

seeding densities

102

Figure 4.11: Doubling time of Sf9 in various cultivators and at different seeding

densities

increased as the seeding density increased. The longer the doubling time, the slower

the growth. Based on the characteristics that were discussed earlier, it was obvious

that Sf9 cells that were cultured in 0.5 ml SFM in each well of the 24-well plate

mimicked closely Sf9 cells grown in a shake flask.

4.2 Establishment of Baculovirus Expression Vectors System (BEVS)

4.2.1 Mock Infection Optimization

Three factors were investigated in the mock infection optimization including

effect of initial cell density, spent medium and MOI. In all experiments, Sf-9 cells

were infected with wild type baculovirus (AcMNPV) during the early exponential

phase.

Three different initial cell densities, i.e. 0.95, 2.05 and 5.13 x 105 cells/ml

respectively were infected with AcMNPV at MOI 10 in fresh medium. The cell

103

densities used in this experiment were relatively low to ensure oxygen and nutrient

will not be the limiting factor. As shown in Figure 4.12, the rates of infection were

similar and reached a maximum infectivity of 98.5%, 99.5% and 100% for 0.95, 2.05

and 5.13 x 105 cells/ml respectively by 120 h post-infection (PI). The results

indicated that initial cell density alone was not a critical parameter for determining

cell infectivity.

The finding on the effect of spent medium carry over on viral infectivity is

interesting. As shown in Figure 4.13, this observation suggested that the medium

must be replenished before viral infection in order to achieve highest protein

expression.

.

104

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0 24 48 72 96 120Time post-infection (h)

Infe

ctiv

ity

(%)

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0


Via

ble

Cel

l num

ber (

x 10

5 cel

ls/m

l)

Figure 4.12: The effect of initial cell density on Sf-9 insect cells infected with wild

type AcMNPV viruses at MOI 10. (a) Infectivity percentage versus time post-

infection (TPI) and (b) Viable cell number versus TPI. Error bars indicate ±S.D of

duplicates data.

0.95 x 105 2.05 x 105 5.13 x 105

Initial Cell Density

0.95 x 105 2.05 x 105 5.13 x 105

Initial Cell Density (a)

(b)

105

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0 24 48 72 96 120

Time post-infection (h)

Infe

ctiv

ity

(%)

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0


Via

ble

Cel

l Num

ber (

x 10

5 cel

ls/m

l)

Figure 4.13: The effect of spent medium carry over on Sf-9 insect cells infected with

wild type AcMNPV viruses at MOI 10. (a) Infectivity percentage versus TPI and (b)

Viable cell number versus TPI. Error bars indicate ±S.D of duplicates data.

100% 50% 0%

Spent Medium Percentage: (a)

(b) 100% 50% 0%

Spent Medium Percentage:

106

It is expected that increasing the added amount of virus in cultures can

intensify the process of cell infection. Therefore, by increasing the number of virus

per cell (MOI), a reduction in the time of cell infection can be achieved. To study

this phenomenon, different MOIs were used to infect the stationary phase cell culture

using fresh media as cell infectivity and viral yields in stationary phase have been

found to be strongly dependent on the MOI (Licari et al., 1991). The behavior is

different to that when cells were infected in the exponential phase (Maiorella et al.,

1988, Schorp et al., 1990). As shown in Figure 4.14, in the stationary phase, higher

MOI will enhance the rate of infection (rate of polyhedra development as observed

microscopically). However, final infectivity for each MOI used was similar and this

might probably be due to the availability of nutrient in the fresh media allowing the

cells to survive long enough to be infected by viruses released from primary infected

cells.

107

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0


Infe

ctiv

ity

(%)

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0


Via

ble

Cel

l Den

sity

(x 1

0 5 cel

ls/m

l)

Figure 4.14: The effect of MOI on Sf-9 insect cells infected in the stationary phase

with wild type AcMNPV Viruses. (a) Infectivity percentage versus TPI and (b)

Viable cell number versus TPI. Error bars indicate ±S.D of duplicates data.

(a)

(b)

1 15 50

MOI :

100

1 15 50

MOI :

100

108

Table 4.3: Growth Kinetics of Sf-9 Cells After Mock Infection

1All experiments were carried out in 25cm3 T-flask.

2Maximum growth rate, )1

(h

µ = 12

12 lnlntt

XX−−

where X2 = viable cell number at t2

X1 = viable cell number at t1

3Doubling time, )(htd=

µ2ln

4Growth index, GI =densitycellinitial

densitycellimummax

5Average doubling time, )(_

htd =

)ln

(

2ln

maxtGI

,

where tmax = time at the maximum viable cell density 6Spent medium was prepared by centrifugation of medium from a 12 day culture at death phase. The

low viability percentage, 18.5% indicated that most of the cells were lysed and the breakage generated

a lot of impurities in the suspension.

ND = Not Determined

MOCK INFECTION

1Experiments

Maximum viable cell

density x 105 cells/ml

2 )1

(h

µ 3 )(htd 4 GI 5 )(_

htd


(%)

Fig. 4.6: Effect of Initial density - 0.95 x 105 cells/ml - 2.05x 105 cells/ml - 5.13 x 105 cells/ml

1.98 7.23

17.63

0.006 0.014 0.020

110.9 50.7 34.6

2.1 3.5 3.4

90.9 39.6 40.4

98.5 99.5

100.0

Fig. 4.7: Effect of Spent medium carry-over

% Spent medium: - 100 % - 50% - 0 %

3.23 8.23

14.60

0.016 0.034 0.038

43.2 20.5 18.4

3.4 8.7

15.4

40.8 23.1 18.3

72.5 93.8 99.8

Fig. 4.8: Effect of MOI

- MOI = 1

- MOI = 15 - MOI = 50

- MOI = 100

34.30 31.80 21.58 12.63

0.021 0.016 0.015 0.008

32.7 44.3 46.0 88.3

4.6 4.3 2.9 1.7

32.6 34.3 46.8 62.8

97.3 97.3 99.5 99.5

109

4.3 Study on the Expression Profiles of rhTf in Infected Sf9 Insect Cells

Culture

4.3.1 rhTf Expression at Different MOIs

In uninfected Sf9 cells, heterologous proteins ranging from 25-225 kDa of

molecular weights were observed. The SDS PAGE analysis showed that Sf9 cells

infected with AcMNPV resulted in the decline or shut off of host gene expression

(Figure 4.15). In lane 4 of Figure 4.15 (Sf9 infected with wild-type baculovirus),

almost all of the major proteins in healthy Sf9 cells were not expressed. The down

regulated mechanism of the host gene expression was not fully understood. It was

believed that it required late expression for most of the host genes. Certain viral

genes (dnapol, hel(ts8), and pcna) that control viral replication might also influenced

the time of host gene expression (O’Reilly et al., 1994).

For recombinant baculovirus infection, the analysis had shown that rhTf was

expressed as the major protein. However, the molecular weight of the rhTf was

slightly lower than that of its native counterpart (apo-hTf) (Figure 4.15). This might

be due to incomplete glycosylation compared to its native counterpart and the

absence of iron bound to the transferrin molecule (Ailor et al., 2000). The absence of

particular glycosylation-related enzymes could be the reason for incomplete

glycosylation in insect cells. However glycosylation problem could be tackled by

introducing these enzymes in vivo or in vitro by means of metabolic engineering.

110

Figure 4.15: SDS PAGE analysis of rhTf expression. Marker kDa (1, 6, 10), apo hTf

(2, 9), uninfected Sf9 (3), wild type virus infected Sf9 at 4 MOI (4), recombinant

virus infected Sf9 at 10, 50,100 MOI (5, 7, 8)

Figure 4.16: rhTf expression profiles at different MOIs

During the first day of infection, the DNA of recombinant AcMNPV carrying

hTf gene would recombine with the Sf9 DNA resulting in the shutting down of Sf9

genes expression. During this phase, budded viruses were also being produced

extensively (O’Reilly et al., 1994). From Figure 4.16, it could be seen that during the

first day, rhTf yield was still low (~1.0 �g/ml). rhTf was produced extensively after

111

day one post infection and this was extended to another one to two days after which

the production rate decelerated. The deceleration phase might be due to the dramatic

drop in Sf9 cells viability after 48 hours post infection After the deceleration phase,

rhTf concentration was still increasing in a relatively small quantity.

It was also shown that infection at low MOI produced higher rhTf yield. (in

this case at 5 MOI). At lower MOIs, the Sf9 cells were allowed to propagate further,

thus increasing the concentration of rhTf.

4.3.2 rhTf Expression at Different Seeding Densities

Studies were also conducted to see the effect of different seeding densities on

rhTf production using Sf9 monolayer culture grown in T-flask. The optimum yield

was found at the seeding density of 1.6 x 106cells/ml (Figures 4.17, 4.18) with

approximately 10 �g/ml rhTf. At this density, the Sf9 cells monolayer was 100%

confluent. This observation suggested that the infection with recombinant AcMNPV

was optimum as well as the mass transfer of medium through the cells membrane. At

seeding densities higher than 1.6 x 106cells/ml, the cells monolayer was already over

confluent when infection was initiated. In this case, the virus distribution would be

inefficient and problems regarding mass transfer might become more evident.

Although the cells were at higher concentration, only a portion was able to express

rhTf. Thus, it could be concluded that the relationships between rhTf yields and

seeding densities higher than 1.6 x 106cells/ml were not dependent on MOI and

nutrient consumption. In suspension culture, the optimum seeding density for rhTf

expression was also 1.6 x 106cells/ml.s

112

Figure 4.17: rhTf expression profiles at different seeding densities, SD. Experiments

were done in T-flask

Figure 4.18: Surface plot of figure 4.17

113

4.3.3 rhTf Expression at Different Times of Infection

Effects of different times of infection on the rhTf production had also been

investigated. Again, the data showed clearly that initiation of infection at day 1 and 2

post culture gave significant rhTf yield. The initial cells densities for these tests were

0.8 x 106

viable cells/ml. At day 2 post culture, the cells had reached approximately

1.6 x 106

viable cells/ml (Figure 4.6). This was two times the initial density. When

recombinant AcMNPV was introduced into the culture during this time, rhTf yield

was maximum compared to time of infection at day 0, 1, 4, and 6 (Figure 4.20).

For time of infection of day 2 (cells density at time of infection = 1.6 x 106

cells/ml), rhTf yield was also found to be higher than at time of infection of day 0

(with same cells density = 1.6 x 106 cells/ml) (Table 4.4). This finding suggested that,

for a fixed cells density at time of infection, cells which had adapted into the culture

environment would produce higher rhTf yield. The spent medium might contain

secreted growth promoting factors with a positive effect on protein production

(Jesionowski and Ataai, 1997). When the Sf9 cells were first cultured in the T-flask,

adaptation process took place and synthesis of some growth or expression promoting

factors might still be at a low level. When infection was initiated at this time, the

rhTf yield was not really good. When the cells were infected at a later time, when

enough growth promoting factors were available, the rhTf yield was higher.

When the cells were infected at a later time, they were actually allowed to

propagate further thus achieving higher density. Higher density allowed these cells to

express more rhTf and therefore increased the rhTf yield even further. If the cells

were infected too late, the cells would become over confluent. This would reduce the

mass transfer efficiency. Furthermore, some of the nutrients might have been fully

consumed. Eventually rhTf expression could not reached higher concentration

(Figure 4.19) and the yield was minimal too (Table 4.4).

114

Figure 4.19: rhTf expression profiles at different times of infection, TOI

Figure 4.20: Surface plot of figure 4.19

It was thought that the rhTf concentration would reach a stationary phase due

to loss of viability and decline towards the end because of protein degradation. The

115

rhTf concentration however, was found to continously increase even after the

deceleration phase.

Table 4.4: rhTf yield coefficients at various seeding densities, MOIs, and times of

infection. Yield was based on day 4 post infection and 103

cells

One possible reason was that, after day 4, there were still some viable cells in

the culture. The cells were still reproducible and might be able to express rhTf.

Supplementation of methionine and tyrosine was found to retard cell death in Sf9

culture (Mendonca et al., 1999). The SF-900 II medium used in this study might

have these supplements or other death retarding nutrients that supported the cells to

remain viable for a longer period thus enabling the cells to produce more

recombinant proteins.

Another explanation is that, the recombinant rhTf might have undergone a

process where its molecular structure was degraded into smaller structures which

could still be identified by the goat-anti-hTf in the ELISA analysis. Since the ELISA

analysis was only a quantitative analysis, it accounted for whatever forms of rhTf in

the culture. Therefore, the rhTf produced very late in the culture might be the

biologically inactive or degraded ones. Biological activity of hTf is the ability to

bind iron from the extracellular fluid and release it into the intercellular fluid. The

binding of iron occurs only in the company of an anion that serves as a network

between iron and hTf (Aisen and Listowsky, 1980). The detachment of iron from hTf

depends on hTf receptor mediated endocytosis (Karin and Mintz, 1981). For this

study, biological activity analysis of rhTf was beyond the scope.

116

A significant problem encountered when infecting cells in T-flasks was the

difficulty in maintaining the homogeneity of the cells, medium and virus due to the

lack of a mixing device. However, this problem could be tackled by introducing the

cells into suspension culture.

4.4 Optimization of the Recombinant Human Transferrin Expression

4.4.1 Recombinant Baculovirus Screening

The intial step for baculovirus screening was by visual checking . As shown

in Figure 4.21, Sf9 cells infected with wild type baculovirus showed small granules

within the cells whereby cells infected with recombinant baculovirus showed rough

surfaces around the cells. At the beginning of the experiment, dilution at 10-6

(higher

MOI) and 10-7

(lower MOI) were suspected to give <10% of infected cultures. Table

4.5 displayed the concentration of rhTf in each well. For dilution at 10-6

, 17 or 42.5%

wells were found to have rhTf at different concentrations but only four (10%) wells

were significant (A3, F4, G4 and H6 in Figure 4.22). For dilution at 10-7

, 12 (30%)

wells were found to have rhTf at different concentrations but only one (2.5%) well

was significant (D9 in Figure 4.22).

If all infected wells were to be taken into account, the purity of recombinant

virus in the 10-6

diluted stock was 74.9% and for 10-7

, the stock purity was 83.3 %

(Table 4.6). If only the significant rhTf yield was to be taken into account, the purity

of recombinant virus in the 10-6

diluted stock was 94.8% and for 10-7

, the stock purity

was 98.7 %. Thus, the 98.7% purified rhTf-AcMNPV was the best stock and was

further amplified to generate large scale virus stock.

117

Figure 4.21: Comparison between uninfected (U), wild-type (WI), and recombinant

(R) virus-infected Sf9 cells

Table 4.5: Concentration (�g/ml) of rhTf in each well of a 96-well plate

rhTf yield in the purified baculovirus stock was very low ~1�g/ml (F4 and

D9 in Table 4.5) compared to the original stock ~10�g/ml. This was because the

virus screening was performed by diluting the original virus stock to 10-6

to 10-7

of

dilution factor (Table 4.5). When the high purity baculovirus stock was amplified

and used to express rhTf at the same experimental conditions, the yield was

~15�g/ml. This proved that the screened or purified recombinant baculovirus had

resulted in improved production of rhTf.

Another observation made from this finding was that although homogenous

virus and cell stock were used, the concentration of rhTf in each well was not equal.

When the virus stock was diluted at very high dilution factors (i.e. up to 10-7

), the

118

chances of productive and non productive viruses to land on different well of the

plate varied. In this case, the effect of non productive virus could be seen based on

the varying concentrations of rhTf in each well.

Figure 4.22: 3D plot of Table 4.5

The cause of this occurence was known as the effect of serial passages of

recombinant virus stock. Serial passaging of undiluted virus stocks (eg. high MOIs)

result in the accumulation of defective interfering particles. These particles are not

infectious but interfere with virus replication (Kool et al., 1991). Because the region

deleted from the genomes of these particles includes polh, significant declines in the

level of heterelogous gene expression will occur if care is not taken in the passaging

of virus stocks. Even extended passage of viruses in cell culture with low MOIs

results in a few polyhedra (FP) mutants (Kumar and Miller, 1987).

The purity of the virus stock was defined and calculated from the Poisson

distribution equation. At any given time, the fraction of cells, P(t,k) infected by k

number of virus particles was assumed by a Poisson distribution (Licari and Bailey,

1992; O'Reilly et al., 1992; Tsao et al., 1996) to be :

…4.1

119

where � was a factor describing the effectiveness of infection (virus bound and

gained entry into cell by endocytosis). V1(t)/No(t) was defined as the “dynamic

MOI” (dm(t)) which changed with time as viruses were taken up by cells and were

subsequently budded from infected cells (Hu and Bentley, 2000). As a consequence,

the infection curve shifted during the course of infection. To simplify this equation,

the dynamic MOI was represented by ‘�’.

…4.2

In this study, screening for pure recombinant virus stocks involved the

preparation of a stock from a single infectious recombinant virus. The proposed

recombinant virus purity was defined as the ratio of proportion of cells receiving

only one infectious recombinant virus to the proportion of cells receiving one and

more infectious units (equation 4.3).

…4.3

It was almost impossible to identify whether a culture has received only a

single infectious recombinant virus. However, the proportion of infected wells to the

total number of wells could be determined experimentally based on the result of

ELISA. Any well with rhTf would have received at least one recombinant virus unit

and was scored positive. From this proportion, therefore, the value for � was

calculated and used to find the purity of the recombinant baculovirus stock. In this

case, Table 4.6 was generated to monitor whether the definition of virus purity based

on Poisson distribution was valid. Note that the dynamic MOI (�) in Table 4.6 is

only valid at time of infection. From Table 4.6, it is clearly seen that as the virus

purity increases, the MOI and proportion of infected culture decreases which is in

agreement with the results in Table 4.5 and also by Hu and Bentley (2000). As the

virus innoculum was diluted in Table 4.5, the MOI decreased and fewer cells were

infected. This resulted in lower proportion of infected culture.

For synchronous infection where almost all or 99-100% of the cells are

infected, the MOIs needed for infection are about 5-10 MOI (Table 4.6). This data

was well documented by many researchers in BEVS including O'Reilly et al.,

(1992); Hu and Bentley, (2000) and Nishikawa et al., (2003). It is also possible that

the Poisson distribution could be manipulated to determine virus titer (pfu/ml) since

120

Table 4.6: Viral Screening by End Point Dilution Method (Poisson distribution data

sheet)

volumes of innoculum, cells number, number of infected cultures, and MOIs are

known. Thus, the definition of virus purity is valid based on the above

considerations.

The end point dilution method based on Poisson distribution was very useful

and could rapidly screen and determine virus purity at certain degree of confidence.

It also required less materials. This method of virus screening has strong fundamental

and principle. Virus purity can be defined in two ways. 100% purified virus consists

121

of only one single infectious unit and 100% purified recombinant virus consists of

only one single infectious unit that carries the gene of interest. The more dilute the

virus stock, the more chances a single infectious particle will land into a specific

well. If the single virus particle carries the gene of interest, the well has a pure

recombinant virus. Since the virus cannot be seen with bare eyes, the Poisson

distribution gives an estimate of the virus purity with certain confidence level. The

factors that affect the end point dilution method are shown in Table 4.7.

Table 4.7: Factors affecting the end point dilution method

4.4.2 Medium Screening

The lower and higher values for each nutrient for medium screening are

shown in Table 4.8. The Plackett-Burman Screening Design is shown in Table 4.9.

During medium preparation, precipitates formed in almost all wells except for

well no 10, 22, 25, 27, 28, 30, 32 and 33 where the solutions were clear (refer to

Table 4.9). All of the clear solutions contained no arginine and lysine while the

cloudy solutions contained either arginine or lysine. This was verified by repeating

the medium preparation without arginine and lysine. All solutions were found to be

clear and once arginine or lysine was added into the clear solutions, they formed

precipitates. The characteristics of amino acids might explain this phenomenon.

Amino acids are grouped into three types of classes mainly neutral, acidic and basic

122

Table 4.8: Real values for the screening of 13 selected nutrients using Plackett-

Burman design

amino acids. Arginine and lysine are basic amino acids with functional side chains of

guanidine and amine respectively. The pKa values of both side chains are more than

10. Therefore, arginine and lysine display very strong base characteristics. The

Sf900-II SFM cell culture medium had a pH value 6-7 which was acidic. Both side

chains of arginine and lysine might have undergone acid base reaction where the

products of reaction were salts that precipitated out of the solution. However,

arginine and lysine could still be added into the medium provided that they were in

the form of neutral, acidic or free base amino acids. Cysteine, glutamine, methionine,

serine, threonine, tyrosine, and valine are all neutral amino acids.

123

Table 4.9: 13-factor (nutrients), 33-run, 2-level Plackett-Burman screening design

During infection with recombinant baculovirus, some of the sf9 cells culture

showed a different morphology compared to the control. Cells in a medium with

added lipid mixtures exhibited a granular appearance which looked like a wild type

baculovirus infection (Figure 4.23). Somehow, the lipid mixtures might have affected

the ability of the cells to express recombinant protein. Figure 4.24 shows the results

of three medium screenings which were carried out based on the Plackett-Burman

screening designs (Table 4.8, Table 4.9). All three screenings were carried out at the

same experimental conditions. The same baculovirus stock, MOI=0.36, time of

infection= 0 hr, cell initial density= 4 x 105/ml, and volume of medium = 1.0 ml were

124

used. Different mother cultures and harvest time were used. The main reason for this

was to check for any significant changes in the patterns due to prolonged infection.

For screen 1, samples were harvested at 4 days post infection (dpi), screen 2 and 3 at

10 dpi.

Figure 4.23: Infected cells appearance in medium A (lipid mixtures added) and

medium B (no lipid mixtures added)

Figure 4.24: rhTf concentration at different medium compositions based on Plackett-

Burman screening experiments. Experiment no 1-33 represent 33 different medium

compositions in 33 different wells of Sf9-AcMNPV culture

Results show very similar patterns in all three screenings, except for wells 24

-33. These differences were not fully understood, but were most probably caused by

125

uneven distribution of cells in the wells which resulted in unsynchronized

expression. Control experiment was in well 33 in which there was no nutrient

addition but 100% Sf900-II SFM. There was a big gap between the low and high

value of rhTf concentration which indicated significant nutrient effects towards rhTf

expression.

Some wells (no 3, 4, 5, 11, 12, 15, 17, 18, 19, 20, 21, 22, 23, 27, and 29)

displayed lower rhTf yield than the control which indicated that some nutrients might

have negative effect on rhTf expression. Some wells displayed higher rhTf yield than

the control which indicated that some nutrients might have positive effect on rhTf

expression. A rapid observation on the designed medium composition (Table 4.9)

revealed that lipid mixtures existed in all of the higher peaks. Other nutrients with

positive effect on rhTf yield were not known at this stage. Statistical analysis of the

Plackett-Burman screening experiments were conducted to gain more information.

Figure 4.25: SDS-PAGE analysis of medium screening. Ex (selected

experiment/well no.), r (recombinant), n (native), M (protein marker), S (human

serum). Loading volume is 25 �l. Samples from Screen 3

SDS-PAGE analysis was conducted to assess the quality of rhTf and the

results are presented in Figure 4.25. Samples from screen 3 with higher rhTf levels

(Ex1, Ex7, Ex24 and Ex32) and lower rhTf levels (Ex4, Ex12 and Ex17) than the

control (Ex33) were selected and analyzed with SDS-PAGE. The lowest value was

Ex4 with ~9 �g/ml and the highest was Ex7~ 22 �g/ml. Production of heterologous

126

proteins ranging from 25-225 kDa was observed. The thickness of each band is in

agreement with the data in Figure 4.24. Figure 4.25 also shows that rhTf was the

major protein expressed in the infected Sf9 cells.

The molecular weight of rhTf was slightly lower than that of native human

transferrin (nhTf). This might be due to lack of complex type oligosaccharides

attached to the polypeptide as discussed by Ailor et al., (2000); Tomiya et al., (2003)

and Ali et al., (1996). Most recombinant glycoproteins produced in the baculovirus-

insect cell system have highly trimmed glycans consisting of Man3-GlcNAc2 in

place of the fully extended complex glycans found on native mammalian

glycoproteins. This difference reflects the absence of high levels of terminal

glycosyltransferase activities in insect cells or the presence of competing, membrane

bound N-acetylglucosaminidase activities. Adjacent bands closed to rhTf were also

observed and had lower molecular weight than rhTf. This might be the results of

proteolysis. Harvesting rhTf very late post infection might be the cause of protease

accumulation in the cell culture. Hu and Bentley, (2000) reported that harvesting

cells with viability near 50% could avoid further cell lysis and the release of protease

into the medium, which would worsen the degradation process.

The estimated effects of the nutrients on rhTf yield are shown in Table 4.10

for a 95% confidence level. Based on the analysis, addition of 7 nutrients (lipid

mixtures, L-glutamine, glucose, L-cysteine, L-valine, L-methionine, L-threonine, and

L-serine) were found to give an increase in rhTf yield. Meanwhile, addition of

fructose, L-tyrosine and maltose caused rhTf concentration to decrease. Based on the

p value, lipid mixtures and L-glutamine effect had the highest significance level

(p<0.05) followed by glucose, L-cysteine ans L-valine (p<0.5). Therefore lipid

mixtures and L-glutamine were chosen for further optimization.

127

Table 4.10: Estimated effects on rhTf yield based on the results of Plackett-Burman

screening experiments

Figure 4.26: Effect of nutrients on rhTf yield. Effect was calculated based on the

amount of increment/reduction of rhTf yield due to nutrient feeding in Plackett

Burman design

It is clearly presented in Figure 4.26 that lipid mixtures had the most

significant effect on rhTf production (p<<0.05) and this result confirmed the rapid

observation done earlier. According to Inlow et al., 1989, lipid concentration in

insect cell serum-free media is in the range of 1000 g/L. Lipids are the main

components of membranes and they form permeability barriers that are essential for

cell survival and function. Most serum-free cell culture medium formulations include

128

essential fatty acids to replace the growth-promoting properties of the lipid

components of serum (Barnes and Sato, 1980; Maiorella et al., 1988). Supplements

of fatty acids in serum-free cell culture media have been recognized as essential to

stimulate cell growth (Rose and Connoly, 1990) and to improve the robustness of

cells in agitated cultures (Butler et al., 1999).

It can be seen that glucose was the only carbon source utilized at the highest

significance level during the infection (Table 4.10). Fructose and maltose were not

important in this process. Although they exhibited certain degrees of effects, they

were not significant (p>>0.05). Reuveny et al., (1993) reported that only glucose,

fructose and maltose were used as carbon sources in insect cells culture. In another

report, glucose was identified as the preferred energy and carbon source (Drews et

al., 1995). Fructose and maltose were only consumed after glucose was depleted. It

can be concluded that the glucose in the Sf-900 II was not fully utilized when the

infection completed, therefore fructose and maltose were not consumed.

All amino acids that were screened gave positive effects except for tyrosine.

L-glutamine effect was the most significant (>95% significance). The results of batch

cultivations showed that glucose was the preferred energy and carbon source limiting

the cell density. However, even in the presence of glucose, significant amounts of

Asp, Gln, Asn, Ser, Arg and Met were utilized for energy production (Drews et al.,

1995). Glutamine feeding played a major role to sustain culture viability for 36 hours

post infection (hpi) (Palomares et al., 2004). The consumption of His, Lys, Thr, Gly,

Val, Leu, Phe, Tyr, Trp and Ile by the growing Sf-9 cells was almost equal to their

concentration in the biomass (Drews et al., 1995). All these amino acids can provide

energy through the tricarboxylic acid (TCA) cycle.

129

Figure 4.27: Amino Acids in Human Transferrin (679 residues)

The effects of amino acids involved in the screening experiments might be

correlated to the amount of amino acids in the human transferrin molecule (Figure

4.27). However, there are only 17 glutamines in human transferrin, which are the

fourth lowest, while its effect on rhTf yield was the most significant among the

amino acids screened. This result proposed that the glutamine consumed by Sf9 cells

were not significantly used for rhTf assembly but more for cells metabolism. The

energy produced from the tricarboxylic acid (TCA) cycle could enhance the cells

ability to express rhTf. This is in agreement with the findings by Drews et al., (1995)

and Palomares et al. (2004).

In addition to glutamine, valine and cysteine were also found to have

significant effects. They exist in significant amount in the rhTf molecule with 45

residues for valine and 38 residues for cysteine. This suggests that the consumptions

of valine and cysteine are for rhTf assembly. The effect of methionine was quite low.

With only 9 methionine residues for every rhTf molecule, it can be suggested that

methionine had little effect on both metabolism and expression of rhTf. More than 25

residues of threonine, serine, and tyrosine are present in rhTf molecule. On the other

hand, their effects were very low. Therefore, their role in promoting a successful

production of rhTf is not as pronounced as the other amino acids.

130

4.4.3 Medium Optimization using Response Surface Methodology

4.4.3.1 Regression Model

The results of the optimization experiments are shown in Figure 4.28 and

Table 4.11. In the control experiment (test no. 17), where there was no additional

nutrients feeding, rhTf concentration was 19.89 �g/ml. The maximum rhTf yield was

in test no. 5 with 62.28 �g/ml. This indicates that nutrients feeding had successfully

increased rhTf yield.

To further understand the relationship among nutrients and rhTf

concentration, a multiple regression analysis was conducted on the experimental

data. The results are given in Table 4.12. The parameters’ coefficients were used to

construct the second-order polynomial model which explained the correlation of each

nutrient and their second-order interactions with rhTf production. The equation

obtained is:

Y = 35.02 + 0.87x1

− 6.32x2

− 5.97x3

− 5.63x1x

2 − 3.95x

1x

3 +

2.79x2x

3 + 4.31

1

2 + 3.21x

2

2 − 9.99x

3

2 ... 4.4

where Y is rhTf response in �g/ml, x1

is the coded value of glutamine, x2

is the coded

value of glucose and x3

is the coded value of lipid mixtures . The quadratic model in

equation 4.4 with nine terms contains three linear terms, three quadratic terms and

three, two-factor interactions. All terms are included in the model to give the

optimum fit of the experimental data. Equation 4.4 was used to predict the output of

rhTf concentration with planned parameters and compared with observed values. The

observed and predicted experimental values are given in Table 4.11.

131

Table 4.11: Central composite design for the optimization of glutamine, glucose and

lipid mixtures 1000x

Figure 4.28: Observed and predicted experimental data for the optimization of

glutamine, glucose and lipid mixtures. Medium composition was based on Table

4.11

132

Analysis of variance (ANOVA) was done using Statistica (Statsoft v. 5.0) and

the results are given in Table 4.12. The fisher F-test value signifies how the mean

square of the regressed model compares to mean square of the residuals (errors). The

F value for this case was 16.71. The greater the F value, the more efficient the model.

The significance of F value or sometimes referred to as P value is the probability to

get large F value by chance alone. A very low probability (Pmodel > F = 0.00001)

demonstrates a very high significance for the regression model. This shows that F

value is too large to have arisen by chance alone. The fitness of the model was

checked by the determination coefficient (R2) which is the ratio of SS

regression to

SStotal

. In this case, the value of the determination coefficient (R2

= 0.96) indicated

that only 4% of the total variations were not explained by equation 4.4. The value of

the adjusted determination coefficient (Adj. R2

= 0.90) was also very high, which

indicated a high significance of the model. The correlation coefficient (R = 0.98)

show a significant correlation between the independent variables and the rhTf

response.

The significance of each coefficient was determined by student's t-test and P

values, which are listed in Table 4.12. The larger the magnitude of the t- value and

the smaller the P- value, the more significant the corresponding coefficient. As a rule

of thumb, coefficients with P<0.05 are considered significant (Kalil et al., 1999).

Almost all effects are significant except for first order effect of glutamine and two-

way effect of glutamine and lipid mixtures. The quadratic effects of glutamine and

glucose are both positive, which indicate that there are minimum values for their

concentrations. Meanwhile, the quadratic effect of lipid mixtures signifies that there

is an optimum value for its concentration. The effects of linear, quadratic and two-

way interaction can be arranged according to their ascending order of P value.

Generally, the quadratic effect of lipid mixtures (x3) is the most significant as is

evident from the respective P-values (Px3

2 = 0.00001 > P

x1

2 = 0.0200 > P

x2

2 = 0.0500)

with the first order main effects (Px3

= 0.00001 > Px2

= 0.0001 > Px1

= 0.5300) and

two-way main effects (Px1x2

= 0.0100 > Px1x3

= 0.0500 > Px2x3

= 0.13). All of these

values suggest that the concentration of glutamine, glucose and lipid mixtures have a

133

Table 4.12: Analysis of Variance (ANOVA) of the CCD

direct correlation on the expression of rhTf. The magnitudes of the coefficients are

evenly large which indicate that all of the coefficients have significant contribution

to rhTf concentration. The comparison of the predicted and observed experimental

data gives a standard deviation, Se = 3.3049, which signifies that none of the

residuals exceed twice the magnitude of Se. Thus, all of the above considerations

show excellent adequacy of the regression model.

4.4.3.2 Nutrients Interactions

To study the effect of nutrient interactions on rhTf expression, three surface

plots involving two nutrients as X-axis and Y-axis with rhTf as Z-axis were

134

constructed. The third nutrient was held at its center point. The results of the surface

plots are shown in Figure 4.29, 4.30, and 4.31.

Glutamine and glucose interactions are shown in Figure 4.29 using the

regressed equation. It can be seen that at a lower glucose concentration (coded value

= -2.5), an increase in glutamine concentration will result in increased rhTf yield. At

a higher glucose concentration however, an increase in glutamine concentration will

result in decreased rhTf yield. It is also clearly seen that at a lower glutamine

concentration, an increase in glucose concentration will increase the rhTf yield. At a

higher glutamine concentration however, an increase in glucose concentration will

decrease the rhTf yield. These interactions signify that rhTf yield is improved when

using lower concentration of glucose and higher concentration of glutamine (glucose

= -2.5, glutamine = 2.0) and vice versa (glucose = 2.0, glutamine = -2.5).

Figure 4.29: Glutamine (Gln) vs Glucose (Gluc) vs rhTf

135

Figure 4.30: Glutamine (Gln) vs Lipid Mixtures 1000x (Lip) vs rhTf

Glutamine and lipid mixtures interactions are shown in Figure 4.30. It seems

that these two nutrients have less significant interactions compared to the previous

figure. Each nutrient tends to follow its own patterns regardless of the concentration

of the other nutrients. For example in Figure 4.30, an increase in lipid mixtures

concentration will improve rhTf yield until at a certain point where rhTf yield will

start to decrease. These patterns are observed in all regions of glutamine

concentration. The quadratic effect of lipid mixtures is also more pronounced than

the quadratic effect of glutamine. This gives an optimum value of lipid mixtures at

around the middle value (coded value = 0). For glutamine, there are two rhTf peaks

observed. The first peak is at lower concentration and the second peak is at higher

concentration of glutamine. For cost effective purpose, the lower concentration of

glutamine is preferable.

136

Figure 4.31: Glucose (Gluc) vs Lipid Mixtures 1000x (Lip) vs rhTf

Glucose and lipid mixtures interactions are shown in Figure 4.31. These

nutrients also have insignificant interaction as evident by its P-value in the ANOVA.

Each nutrient has the same patterns over the concentration range of the other

nutrients. For example in Figure 4.31, the quadratic effect of lipid mixtures is very

significant as it is in Figure 4.30. These patterns again, are observed in all regions of

glucose concentration. This also gives an optimum value of lipid mixtures at around

its middle value (coded value = 0). For Glucose, one rhTf peak is observed in the

region of its lower concentration.

Optimum values for glucose and glutamine have been observed in the lower

concentration region. It is however observed in Figure 4.29 that rhTf yield will

improve when glucose concentration is greater than glutamine concentration or vice

versa. Based on these considerations, the optimum values for glucose and glutamine

are presumably in the lower concentration region where concentration of glutamine

is higher than glucose concentration. In addition to that, the optimum value for lipid

mixtures is in the center point region of its coded value.

137

Response Surface Methodology (RSM) based on the method of steepest

ascent was carried out to hunt for actual optimum point of rhTf yield using the

regression model. The optimum values of the test variables in coded values are x1=-

1.1155, x2=-1.4832, and x

3=-0.2933 with the corresponding response Y=47.33. The

real values of the test variables are glutamine=2211.20 mg/L, glucose=1291.95

mg/L, and lipid mixtures=0.64 %v/v .The predicted rhTf yield using the optimized

concentration of the nutrients is 47.33 �g/ml. An experiment performed using the

optimized parameters resulted in rhTf yield of 65.12�g/ml. This result therefore,

verifies the trend of the predicted value and the effectiveness as well as the

usefulness of the model towards achieving the optimization.

138

4.5 Characterization of the Optimized Recombinant Human Transferrin

Expression

Table 4.13: Summary of the characteristics of optimized rhTf expression

In Figure 4.32, the Sf9 growth rate of optimized expression was lower than

the controlled expression. The specific growth rates of controlled and optimized

expression were 0.5451 hr-1

and 0.2673 hr-1

respectively (Table 4.13). This was

probably due to lipid mixtures feeding which slows down the Sf9 growth

(observation made from early screening). Lipid mixtures however, stimulate rhTf

expression. Lag phase was also observed for the optimized expression and this was

assumed to be due to medium adaptation. From Figure 4.32, it can be concluded that

the maximum cell density for the optimized expression (2.97 x 106

cells/ml) was

lower than the control (7.71 x 106

cells/ml). Life span and viability drop were similar

for both cases.

139

Figure 4.32: Sf9 growth in controlled and optimized expression. Dotted line

indicates where infection was initiated

Figure 4.33: Total protein and rhTf contents in controlled and optimized expression

140

Figure 4.33 shows total protein concentration and rhTf percentage in

optimized and controlled expression. The profiles of each characteristic were similar

in their patterns. Maximum protein productions for controlled and optimized

expression were observed at day six (four days post infection) with 5300.67 �g/ml

and 8665.55 �g/ml respectively. Protein concentration drops were observed after four

days of infection. This occurrence was presumably due to degradation of protein

such as proteolysis and oxidation as a result of prolonged infection and medium

storage. Proteins might also be consumed for cells maintenance and production of

metabolic by-products. After day eight post culture, protein concentration increased

to certain levels. These increases were assumed to be the results of cell lyses where

intracellular proteins were released into the medium. Another reason could be

evaporation which became more evident as infection prolonged. As discussed earlier,

evaporation tended to concentrate the cells because of reduced volume of medium.

On the other hand, rhTf percentage (%) increased throughout the 10 days of

cultivation. rhTf percentage was the ratio of rhTf concentration to the total protein

concentration. rhTf % could be utilized to identify maximum production time. Here,

the optimum rhTf yield was defined as a good balance between highest rhTf

concentration and highest rhTf %. In Figure 4.33, rhTf% increased from day two to

day eight and then remained at a relatively small deviation (+0.01%). It also could be

seen that the maximum rhTf% in optimized expression (0.84%) was higher than the

control (0.36%). Based on these considerations, the optimum production time of rhTf

was identified at day eight (day six post infection). There was no benefit to prolong

infection since this would lead to degradation problems, accumulation of by-products

and complicated purification process.

141

Figure 4.34: Total protein and rhTf production rates in controlled and optimized

expression

Figure 4.35: Glucose and lactate concentrations in controlled and optimized

expression

142

10

11

12

13

14

15

16

17

18

19

20

21

22

0 1 2 3 4 5 6

Day

Con

cent

rati

on o

f Glu

tam

ine

(mM

)

Figure 4.36: Glutamine concentrations in optimized expression

Figure 4.34 shows total protein and rhTf production rates for both optimized

and controlled expression. Negative values were observed for total protein

production in the first two days of cultivation. These negative values signify protein

consumption as can be seen in Figure 4.33, probably for adaptation process. After

infection was initiated (dotted line), protein production began to take off. The highest

protein production rate for optimized expression was at day six (6.11 �g/106cell/hr)

which was 13 times higher than the controlled expression (0.47 �g/106cell/hr).

Production rate of rhTf was obviously higher than the controlled experiment. At day

eight of cultivation, production rate of optimized rhTf was 0.074 �g/106cell/hr which

was 11 times higher than the production rate of non-optimized rhTf (0.007

�g/106cell/hr). These observations clearly showed that rhTf production was

significantly affected by the optimized medium.

Glucose, glutamine and lactate concentration profiles were studied to

characterize optimized expression. In this study, it was assumed that glucose

depletion and lactate production were solely due to cells metabolism. Glycolysis due

143

to medium storage and exposure to open environment was negligible. The results are

displayed in Figure 4.35 and Figure 4.36. As for glucose and glutamine, the

concentration was continuously depleted during the course of cultivation. Glucose

and glutamine were also not a limiting factor since more than 2g/L and 15mM

remained at the end of the cultivation period. Glucose concentration in optimized

medium remained higher than controlled medium because of low cell density and

higher glucose concentration at day zero.

Low lactate level was observed at the end of day 10 for optimized medium.

Low lactate level has been known to maintain pH level and thus improve

productivity (Gorfien et al., 2003). The increase in lactate concentration at day two

of cultivation could possibly be due to lactate carry over from the virus inoculums.

After day two, lactate level dropped for two to four days. During this time, oxygen

transfer and cell growth were assumed to be efficient. Therefore, the drop in lactate

level after day two was caused by significant oxidation of lactate to carbon dioxide

and water (Chiou et al. 2000). Ikonomou et al. (2001) also reported that under non

limiting oxygen, no lactate was produced. During this period (lactate drop), it was

also observed that Sf9 growth was in the exponential phase (Figure 4.32). Lactate

level began to take off at day four in controlled medium and day six in optimized

medium. Lactate accumulation caused impaired cells density (Gorfien et al., 2003).

This was supported by decreased viable cell density at the same time (Figure 4.32).

Concentration of ammonia remained in the culture was lower than 2mM, which ould

not affect the growth of sf9 (Bedard, C. et. al, 1993)

144

Figure 4.37: Lactate production and glucose uptake rate in controlled and optimized

expression

To further explore glucose and lactate profiles, their production and uptake

rate in controlled and optimized expression were further assessed. The results are

shown in Figure 4.37. Generally, lactate production and glucose uptake rates

increased for two days and decreased onwards. Glucose in controlled medium was

consumed at higher rates compared to optimized medium for the first two days. Cells

in controlled medium were denser than optimized medium (Figure 4.32). Therefore

cells in controlled medium consumed more glucose for the first two days. After day

two, glucose uptake rate in optimized medium exceeded the glucose uptake rate in

controlled medium. This transition was assumed to be due to glucose requirements of

Sf9 which needs more energy for rhTf assembly. This was supported by the increase

in rhTf concentration in optimized medium (Figure 4.33).

Lactate production rates increased in optimized medium for the first two

days. As discussed earlier, this might be due to glutamine feeding or adaptation

process. The rate decreased afterwards which suggested that oxidation was more

efficient and cell death has reduced the rate of lactate production.

145

Figure 4.38: SDS-PAGE gel for non optimized medium. A, B, C – by products. MW

(molecular weight marker)

Finally, clear bands of rhTf were detected on the SDS-PAGE gel of the

optimized expression. The results are shown in Figure 4.38 and Figure 4.39. The

thickness of the rhTf bands agreed with the results shown in Figure 4.33. Protein

contents for day 0, 2, 4, and 6 were generally similar for both gels. However, protein

content in non optimized medium was higher as compared to optimized medium.

This was due to high cell density in non optimized medium that expressed the host

protein. The protein content in Figure 4.38 and 4.39 basically reduced towards the

end of infection phase. This was because, at the very late phase of infection cycle,

Sf9 cells could no longer expressed its host protein. Therefore, the existing protein

might have been consumed for rhTf expression.

In optimized medium, by-product A could be clearly seen at day 8 and 10

while in non-optimized medium, by-product A could hardly be seen or not expressed

at all. These bands could be the results of viral protein expression that secreted

during the very late phase of infection or products of proteolysis as discussed earlier

in this chapter.

146

Figure 4.39: SDS PAGE gel for optimized medium. A, B, C – by products. MW

(molecular weight marker)

Another interesting observation was the presence of by-product B and C. In

optimized medium, by-product B and C showed sudden decrease of intensity at day 8

and 10. These proteins were extracellular fluid component since they existed from

day zero of cultivation. These proteins might be consumed most probably to enhance

further expression of rhTf or when the cells were in the state of nutrient starvation.

These could probably explain why glucose requirement in optimized medium was

higher than non optimized medium after two days of infection (Figure 4.37).

4.6 Study of Galactosylation

4.6.1 Recombinant �1,4-Galactosyltransferase Expression

The supernatant of the Sf-9 cell culture infected with recombinant baculovirus

carrying the gene for �1,4-GalT was harvested at 120 hours PI. Following lactose

synthetase assay, recombinant protein from infected cells was analyzed using thin

layer chromatography (TLC). �1,4-GalT catalyzed UDP-Gal in the biosynthesis of

lactose in the presence of �-lactalbumin as the specifier protein and glucose as the

147

substrate (Brodbeck and Ebner, 1966; Brodbeck et al., 1967). �-lactalbumin is

termed the “Specifier Protein” because it modifies the substrate specificity of

galactosyltransferase from N-acetylglucosamine to glucose and allows the synthesis

of lactose in the presence of glucose (Brodbeck and Ebner, 1966; Brodbeck et al.,

1967). In the reaction, a typical mixture contained the following in a final volume of

0.1 ml: 5 �mole of Tris-HCl, pH 7.4; 0.04 �mole of MnCl2; 0.50 �mole of UTP; 0.14

�mole of �-lactalbumin and supernatant harvested at 120 hours PI. As shown in lane

3 of Figure 4.40, lactose was produced and glucose concentration decreased after the

reaction suggesting that �1,4-GalT was expressed in the supernatant.

4.6.1.1 Time Course Expression of �1,4-Galactosyltransferase

The supernatants of infected Sf-9 cell culture with recombinant baculovirus

encoded with �1,4-GalT, were harvested at 24 h intervals. Lactose synthetase assays

were performed and analyzed using TLC (Figure 4.41). Silver stained electrophoresis

gel of the supernatants revealed a 45 kDa protein (Figure 4.42) in infected cells.

Thus it can be concluded that protein expression was started as early as hour 24. The

enzyme �1,4-GalT accumulation in supernatants increased steadily until hour 144.

148

Figure 4.40: Detection of �1,4-GalT by using chromatogram TLC. Layer:

Whatmann, 200µm. Solvent : n-butanol-acetic acid-diethyl ether-water (9:6:3:1).

Detection Method: Anisaldehydyde-sulfuric acid reagent. Lane 1 represents standard

lactose; lane 2 represents standard glucose; lane 3 represents reaction mixture

containing the following in a final volume of 0.1 ml: 5 �mole of Tris-HCl, pH 7.4;

0.04 �mole of MnCl2; 0.50 �mole of UTP; 0.14 �mole of �-lactalbumin and the

sample was the source from 120 hours PI culture supernatant.

1 2 3

Start

Glucose

Lactose

149

Figure 4.41: Time course of chromatogram of TLC. Layer: Whatmann, 200µm.

Solvent : n-butanol-acetic acid-diethyl ether-water (9:6:3:1). Detection Method:

Anisaldehydyde-sulfuric acid reagent. Lane 1 represents standard lactose; lane 2

represents standard glucose; lane 3 represents standard lactose and glucose mixture;

lane 4 - 8 represent supernatants harvested at hour 24, 48, 72, 96 and 120

respectively.

1 2 3

Start

Glucose

Lactose

4 5 6 7 8

150

Figure 4.42: SDS-PAGE (9%) time course of �1,4-GalT production. A silver

stained gel revealed lane 1-7 represent cell culture supernatants harvested at every 24

hours intervals from 0 hour to 144 hours, respectively. M, molecular weight marker.

Arrow indicate the position of �1,4-GalT at molecular weight 45 kDa (Product).

kDa M 1 2 3 4 5 6 7

225 150 100 75

50

35

25

151

4.6.1.2 The Development of �1,4-Galactosyltransferase Assay

A number of methods are available for the measurement of

glycosyltransferase. In radiochemical assays, the radioactivity incorporated into

substrate acceptor from radiolabeled sugar nucleotide donors will be proportional to

the amount of enzyme present. However, this approach has some drawbacks such as

high costs and inevitable disposal problem of the radiochemical wastes. In this

respect, many nonradioactive ELISA-based methods for glycosyltransferase

activities have been developed (Stult and Macher, 1990; Taki et al., 1990; Zatta et

al., 1991; Keshvara et al., 1992; Keusch et al., 1995). All these methods are

essentially based on the same principle. First, either a glycolipid or glycoprotein is

used as an acceptor substrate on the solid surface. Second, the reaction products are

identified by either monoclonal antibodies or specific lectins labeled with enzymes or

fluorescent compounds.

In the study, a lectin binding assay similar to an ELISA-based method was

performed. Asialofetuin was digested with bovine �-galactosidase and the

asialoagalactofetuin produced was used as an acceptor substrate. The

asialoagalactofetuin was coated onto ELISA 96 wells plate, whereas peroxidase

labeled-Ricinus communis agglutinin-I (RCA-I) lectin, which recognized galactose

residues was used to recognize Gal�1,4�GlcNAc linkage on N-glycan of

glycoprotein.

Using the optimal conditions determined for the substrate donor as well as

Mn2+ concentration (Oubihi et al., 1998), the relationship between ELISA values as

the peroxidase labeled-RCA 1 binding signal and the enzyme reaction time was

assessed with different concentrations of �1,4-GalT. As illustrated in Figure 4.43,

sufficient linearity (R2 = 0.9935) was obtained for the �1,4-GalT activity between 0.5

to 2 mU/ml. As shown in Figure 4.44, enzyme expression was seen starting from

hour 24. �1,4-GalT accumulation in cell culture supernatants increased until hour

120.

152

y = 0.025x + 0.0017

R2 = 0.9935

0.000

0.010

0.020

0.030

0.040

0.050

0.060

0.0 0.5 1.0 1.5 2.0 2.5

β 1,4-GalT (mU/mL)

A45

0nm

Figure 4.43: Standard curve for the determination of �1,4-GalT activity from the

lectin binding assay values. Asialoagalactofetuin was used as the substrate acceptor,

UDP-Gal was used as the substrate donor and peroxidase-labeled RCA 1 was used as

the lectin to recognize the Gal �1,4-GlcNAc linkage. Error bars indicate ±S.D of

duplicates data.

153

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 24 48 72 96 120 144Time of Infection (hours)

β1,

4-G

alT(

mU

/ml)

Figure 4.44: Time course of �1,4-GalT enzyme accumulation in supernatants

detected using lectin binding assay. Asialoagalactofetuin was used as an acceptor,

UDP-Gal was used as the donor substrate and Ricinus communis agglutinin 1 (RCA

1) was used as the lectin. Error bars indicate ±S.D of duplicates data.

154

4.6.2 Native Uridine-diphosphogalactose (UDP-Gal) Monitoring at Normal

and Upon Baculovirus Infection

Three main elements to ensure successful protein galactosylation process, are

the presence of sufficient amount of hTf as the substrate acceptor, �1,4-GalT as the

enzyme and UDP-Gal as the substrate donor. In order to achieve galactosylation

effectively, one of the strategies is the introduction of �1,4-GalT artificially. Ailor et

al. (2000) revealed that the oligosaccharide structures of hTf produced in insect cells

infected with GalT baculovirus can alter the glycoforms of the expressed transferrin.

However, another issue is whether the use of native UDP-Gal is sufficient. Tomiya

et al. (2001) had applied High Performance Anion Exchange Chromatography

(HPAEC) method to determine the intracellular sugar nucleotide level of cultured Sf-

9 insect cells at normal level which was found to be around 1400 pmol/mg protein,

but not at the infection level. Thus, the introduction of �1,4-GalT artificially during

the expression of the hTf would not guarantee the success of the galactosylation

process.

As illustrated previously, after infection, recombinant hTf and �1,4-GalT

expression increased over time . This can be explained by the nature of baculovirus

infection cycle. Upon infection, the cells’ mechanism will be shifted to viral

multiplication and expression of its genes and thus the recombinant protein secretion

will increase upon time of infection and will be secreted into the environment.

However, the effect of sugar nucleotide content upon baculovirus infection has never

been reported. This finding is very important to ensure the success of the

galactosylation process. This study is very interesting as it can unlock the potential of

BEVS for the production of recombinant biopharmaceuticals.

In order to establish and monitor the native UDP-Gal at normal and upon

infection, a RP-HPLC analysis was carried out as described in section 3.13.2. Ion

pair RP-HPLC is known to be the one of the most effective methods for separating

sugar nucleotide and nucleotide (Ryll and Wagner, 1991). In this study a RP-HPLC

column (NovaPac C18, Waters) with tetrabutylammonium sulfate as an ion-pairing

reagent was used. A series of different concentrations of standard UDP-Gal was

monitored by RP-HPLC at a flowrate 1 ml/min and UV 260 nm. From Figure 4.45

155

(a) and (b), it was observed that the UDP-Gal peaks were eluted at around 4.8 min.

The heights of the standard peaks were proportional to the substrate donor

concentrations. From the HPLC chromatogram, a standard curve which plotted the

area against the concentration was shown in Figure 4.46. A satisfactory linearity of

0.9994 was obtained.

156

0 0.6 1.2 1.8 2.4 3 3.6 4.2 4.8 5.3 5.9

0.00

013

0.00

025

0.00

038

0.00

050

0.00

063

0.00

075

0

0.02

0.04

0.06

0.08

0.1

A26

0 nm

Elution Time (Min)Concentration (umole)

-0.005

0.005

0.015

0.025

0.035

0.045

0.055

0.065

0.075

0.085

0.095

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

0.00013 umole

0.00025 umole

0.00038 umole

0.00050 umole

0.00063 umole

0.00075 umole

Figure 4.45: RP-HPLC chromatogram for UDP-Gal standard at different

concentrations. 50 mM ammonium phosphate, pH 5.0 containing 5 mM

tetrabutylammonium sulfate (E1) and methanol containing 5 mM tetrabutylammonium

sulfate (E2) were used as the eluents. Standard UDP-Gal portion (10 �l) was injected

into NovaPac C18 column (ø3.9 x 150mm) equilibrated with the mixture of E1 and E2

(98:2, v/v). (a) 2D diagram; (b) 3D diagram.

(a)

(b)

157

y = 5E+09xR2 = 0.9994

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

4000000

0 0.0002 0.0004 0.0006 0.0008

µmole

Are

a

Figure 4.46: Standard curve for UDP-Gal. Area was plotted against the

corresponding UDP-Gal concentration in �mole.

In order to verify the reliability of the assumption that the eluted peak at 4.8

min was UDP-Gal, the native UDP-Gal sample was spiked with 10 �mole of

commercial UDP-Gal. As observed in Figure 4.47, at the elution time of 4.8 min the

peak of the native sample with the spiking was higher compared to the native sample

without the spiking thus confirming the assumption.

In order to monitor the level of native UDP-Gal at normal and upon infection,

extraction of Sf-9 cells and infected Sf-9 cells with AcMNPV-hTf were prepared as

described in section 3.13.1. From the RP-HPLC chromatograms as shown in Figure

4.48, 4.49, 4.50 and 4.51, the UDP-Gal peaks which eluted at around 4.8 min were

significantly becoming smaller starting from 0 hour PI until 120 hour PI at 24 hour

intervals. Using the standard curve generated in Figure 4.46, a graph of UDP-Gal

concentration in �M versus time of infection in hours is illustrated in Figure 4.52.

158

The UDP-Gal level was 15 �M at the beginning and dropped to almost zero upon

five days recombinant baculovirus infection.

To further confirm that the disappearing peak was UDP-Gal, the UDP-Gal

fractions from the RP-HPLC analysis were collected and verified with another assay,

TLC. The methods are as described in section 3.12.1. The time course assay

confirmed the reduction of UDP-Gal content upon baculovirus infection occurs as

observed in Figure 4.53.

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

Native Native with Spiking

UDP-Gal

Figure 4.47: RP-HPLC chromatogram for native UDP-Gal sample with spiking and

without spiking. In spiking, 10 �mole commercial UDP-Gal was introduced into the

native sample to further confirm the elution time of UDP-Gal peak.

159

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

-0.05

0.05

0.15

0.25

0.35

0.45

0.55

0 1 2 3 4 5 6 7Elution Time (min)

A26

0

-0.05

0.05

0.15

0.25

0.35

0.45

0.55


A26

0

Figure 4.48: RP-HPLC Chromatogram for the time course of native UDP-Gal level

upon infection. Infection time at (a) 0h (Normal); (b) 24h; (c) 48h; (d) 72h; (e) 96h

and (f) 120h. (Set Data 1)

(a) (b)

(c) (d)

(e) (f)

UDP-Gal UDP-Gal

UDP-Gal UDP-Gal

UDP-Gal UDP-Gal

160

0

0.6

1.2

1.69

31 2

2.5

3.1

3.7

4.3

4.7

5

5.4

0 hr

(N

ativ

e) 24 h

rs

48 h

rs 72 h

rs 96 h

rs

120

hrs0

0.1

0.2

0.3

0.4

0.5

0.6

Abs

orba

nce

260

nm

Elution Time (Min)Time of In

fection

Figure 4.49: RP-HPLC chromatogram for time course of native UDP-Gal level

upon infection in 3D diagram. Arrow indicated the UDP-Gal peak eluted at time 4.8

min. (Set Data 1)

UDP-Gal

161

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55


A26

0

-0.050.000.050.100.150.200.250.300.350.400.450.500.550.60


A26

0

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

Figure 4.50: RP-HPLC Chromatogram for the time course of native UDP-Gal level

upon infection. Infection time at (a) 0h (Normal); (b) 24h; (c) 48h; (d)72h; (e) 96h

and (f) 120h. (Set Data 2)

-0.050.000.05

0.100.150.200.250.300.35

0.400.450.50

0 1 2 3 4 5 6 7

Elution Time (min)

A26

0

(a) (b)

(c) (d)

(e) (f)

UDP-Gal UDP-Gal

UDP-Gal UDP-Gal

UDP-Gal UDP-Gal

162

0

0.6

1.2

1.69

31 2

2.5

3.1

3.7

4.3

4.7

5

5.4

0 hr

(Nat

ive) 24

hrs

48 h

rs 72 h

rs 96 h

rs 120

hrs0

0.1

0.2

0.3

0.4

0.5

0.6A

bsor

banc

e 26

0 nm

Elution Time (Min) Time of Infectio

n

0

2

4

6

8

10

12

14

16

0 24 48 72 96 120

Time of Infection (Hours)

UD

P-G

al C

once

ntra

tion

( µµ µµM

olar

)

Figure 4.51: RP-HPLC chromatogram for time course of native UDP-Gal level

upon infection in 3D diagram. Arrow indicated the UDP-Gal peak eluted at time 4.8

min. (Set Data 2)

Figure 4.52: Native UDP-Gal concentration in �M at normal and upon time of

infection. The UDP-Gal concentrations at 0, 24, 48, 72, 96 and 120 h PI were

derived from the chromatograms from Figure 4.48 and 4.23. Error bars indicate ±S.D

of duplicates data.

UDP-Gal

163

Figure 4.53: Verification of UDP-Gal fractions from RP-HPLC analysis using TLC.

Layer: Whatmann, 200µm. Solvent : n-butanol-acetic acid-diethyl ether-water

(9:6:3:1). Detection Method: Anisaldehydyde-sulfuric acid reagent. Lane 1: Lactose;

Lane 2: Glucose; Lane 3 lane to 8: Time course infection of UDP-Gal level at 0

(normal), 24, 48, 72, 96, 120h PI respectively.

1 2 3 4 5 6 7 8

Start

Glucose

Lactose

164

4.6.3 Baculovirus Coinfection Study

Baculovirus coinfection study was carried out in order to evaluate the

recombinant glycoprotein quality i.e whether the hTf oligosaccharides included a

terminal Galactose residue. As mentioned in Chapter 2, a lot of experimental

evidences suggested that glycoproteins produced in insect cells comprised of

incomplete N-glycans structures. Ailor et al. (2000) analyzed the N-glycan of human

serum transferrin produced in insect cells using metabolic radiolabeling of the

intracellular and extracellular protein fractions, followed by three-dimensional

HPLC. The attached oligosaccharides included high mannose, paucimannocidic

(Butters and Hughes,1981; Hsieh and Robbins, 1984; Kuroda et al., 1990; Chen and

Bahl, 1991; Kulakosky et al., 1998), and some hybrid structures (Hard et al., 1993;

Kubelka et al., 1994; Davidson et al., 1990; Ogonash et al., 1996) with over 50% of

these structures containing one fucose, �(1,6)- or two fucoses, �(1,6)- and �(1,3)-,

linked to the Asn-linked N-acetylglucosamine. Neither sialic acid nor galactose was

detected on any of the N-glycans.

One possible reason for the limitation in N-glycan processing of

glycoproteins in insect cells is the deficiency in the enzymes necessary for the

production of complex oligosaccharides. In order to determine if altering the

intracellular level of an enzyme in the oligosaccharide processing pathway can

promote the elongation of hTf N-glycan, recombinant �1,4-GalT was overexpressed

in Sf-9 insect cells in conjuction with hTf.

To evaluate the extent of glycosylation by coinfection strategy, an assay was

established. In this binding assay, glycoprotein was absorbed onto the ELISA plate

surface. A lectin known as Ricinus communis agglutinin-I (RCA-I) labeled with

peroxidase, was added and allowed to recognize Gal�1�4GlcNAc group on N-

glycan of glycoprotein. Unbound lectin was removed by washing and the bound

lectin determined by adding equal volumes of TMB and Peroxidase Solution B,

which can be measured by the appropriate color reaction.

In this study, the time course upon coinfection between recombinant

baculovirus hTf and �1,4-GalT was investigated. The medium from each of the

165

0.00

0.05

0.10

0.15

0.20

0.25

0.30

24 48 72 96 120Time of Infection (hours)

Bin

ding

(450

nm

)

coexpressed cell cultures was collected every 24 hour post infection until 120 hour PI

as shown in Figure 4.54. The coinfection strategy was a success. However, as

shown in Figure 4.54, in vivo galactosylation efficiency decreased gradually upon

infection due to the limitation of the substrate donor concentration, UDP-Gal, to

construct the Gal�1�4GlcNAc linkage at the end of the N-glycan hTf. The trend of

Fig. 4.54 is obviously similar to that of Fig. 4.52.

Since the coexpression between the hTf and �1,4-GalT (in vivo) did not

achieved satisfactory results in improving glycoprotein quality due to the reduction

of UDP-Gal upon bacolovirus infection, another alternative, in vitro galactosylation

was proposed to overcome this problem. Commercial mammalian GalT and UDP-

Gal were introduced artificially to the hTf after it was secreted from Sf-9 cell culture.

To determine the optimal conditions for the reaction, different concentrations of

Figure 4.54: Gal�1�4GlcNAc linkage binding values at 450nm for the time course

upon coinfection between recombinant baculovirus hTf and �1,4-GalT. Supernatants

from each of the coexpressed cell culture were collected at hour 24, 48, 72, 96 and

120 respectively. Error bars indicate ±S.D of duplicates data.

166

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.0 1.0 2.0 3.0 4.0

β 1,4-GalT (mU/ml)

Bin

ding

(450

nm

)

GalT were added to the hTf in the presence of excess amount of sugar nucleotide

(0.32 mM UDP-Gal). Lectin RCA-1 was latter added to recognize the �1,4-linkage

on the N-glycan of hTf, followed by TMB color reaction measurement. As shown in

Figure 4.55, the galactosylation reaction reached a saturation point once the

concentration of the mammalian GalT achieved 2.0 mU/ml or above. One

explanation is that the concentration of glycoprotein bound to the every solid surface

was now constant. Furthermore, excess enzyme responsible for transferring the

galactose from UDP-Gal to the N-glycan chain was washed away during the washing

procedure of the assay. For the following works, 2.0 mU/ml mammalian GalT was

selected in order to ensure sufficient amount of the enzyme in the reaction. Also, the

substrate donor will always be in the excess amount for the in vitro galactosylation

process.

Figure 4.55: Effect of the mammalian GalT on the rate of in vitro galactosylation

process. Commercial mammalian GalT and UDP-Gal were introduced to the hTf

after it was secreted from Sf-9 cell culture. Different concentrations enzyme were

added to the hTf in the presence of excess amount of sugar nucleotide, 0.32 mM.

Lectin RCA-1 was later added to recognize the Gal�1,4�GlcNAc linkage on the N-

glycan of hTf, followed by TMB color reaction measurement. Error bars indicate

±S.D of duplicates data.

167

To represent different levels of galactosylation, several conditions were

investigated which include positive controls, negative controls, in vivo and in vitro

galactosylation as shown in Figure 4.56. The positive control in this experiment was

the commercial apo hTf with two N-glycosylation sites which include galactose

residues at each branch. As for the negative controls, there were the Sf-9 cell culture

infected with AcMNPV-�1,4-GalT, uninfected Sf-9 insect cell culture as well as 2.0

mU/ml commercial mammalian GalT and 0.32 mM commercial UDP-Gal which

were added to harvested uninfected Sf-9 cell culture. For the samples, two Sf-9

insect cell cultures were infected with AcMNPV-hTf. One of the cultures was

coinfected with AcMNPV-�1,4-GalT (in vivo). For the in vitro analysis, 2.0 mU/ml

commercial mammalian GalT and 0.32 mM commercial UDP-Gal were added to the

harvested AcMNPV-hTf supernatant. For this part of the study, all cell cultures were

harvested at time 24 hour PI. The concentration of glycoprotein absorbed onto the

ELISA plate was constant, which was 1 µg/ml. A lectin, peroxidase labeled-RCA 1,

was used to interact with the Gal�1,4GlcNAc-linkage on the N-glycan of

glycoprotein, which can be measured by the TMB color reaction.

As observed in Figure 4.56, the various data for the conditions described as

above represent different level of glycosylation. Apo hTf was used as a guide for the

comparison with other conditions due to its N-glycan oligosaccharide chains with

Galactose residues. AcMNPV-hTf produced in insect cell culture did not achieved

satisfactory galactosylation, as expected, because it was missing one important

element that was the enzyme to construct the N-glycan chain. However, this

phenomenon took a positive turn once the AcMNPV-hTf was coinfected with the

AcMNPV-�1,4-GalT. The absorbance value for the in vivo coexpression was 12.5

times higher compared to the AcMNPV-hTf culture. This result showed that the

success of galactosylation did not depend only on the presence of substrate acceptor

and substrate donor only, but strongly also on the enzyme needed to elongate the

chains. As mentioned in section 4.6.2 regarding the native UDP-Gal level upon

baculovirus infection, it was found that the sugar nucleotide content decreased upon

infection. Thus, it was important to consider whether the reduction of UDP-Gal will

have any effect on the galactosylation. Hence, in vitro galactosylation was studied,

where harvested AcMNPV-hTf supernatant at time 24 hour PI was introduced with

GalT and UDP-Gal artificially. Surprisingly, the absorbance for the in vitro culture

168

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

A B C D E F GGroup

Bin

ding

(450

nm)

(0.30-0.35) was higher compared to the in vivo culture (0.22-0.29). This observation

indicated that in addition to GalT, sufficient amount of sugar nucleotide was another

key factor in guaranteeing the success of the galactosylation process. As for the

negative controls, they were showed to be insignificantly galactosylated.

For the galactosylation process to successfully occur, it needs the cooperation

of the substrate acceptor, substrate donor and enzyme. The relationships among the

three main elements of the in vivo galactosylation process, which were hTf as the

substrate acceptor, �1,4-GalT as the enzyme and UDP-Gal as the substrate donor is

illustrated in Figure 4.57. Sf-9 cell culture infected with AcMNPV-hTf and

coinfected with the AcMNPV-�1,4-GalT was used to demonstrate the relationship.

Figure 4.56: Gal�1�4GlcNAc linkage binding values at 450 nm for the different

levels of galactosylation process. A: Apo human transferrin (Standard); B:

AcMNPV-rhTf supernatant harvested at 24h PI; C: Coexpression of rhTf and �1,4-

GalT supernatant harvested at 24h PI (in vivo); D: Introduction of artificial GalT and

UDP-Gal to harvested AcMNPV-rhTf supernatant at 24h PI (in vitro); E: Uninfected

Sf-9 cell culture; F: Introduction of artificial GalT and UDP-Gal to uninfected Sf-9

cell culture; G: AcMNPV- �1,4-GalT supernatant harvested at 24 h PI. Error bars

indicate ±S.D of duplicates data.

169

0

2

4

6

8

10

12

14

16

18

0 24 48 72 96 120

Time of Infection (Hours)

UD

P-G

al C

onc.

( µM

olar

)

0

5

10

15

20

25

30

35

40

45

50

Recom

binant Protein Production (ug/ml)

UDP-galrGalTrhTf

The supernatants and lysates were collected every 24 hr. The hTf and �1,4-GalT

expression was determined using ELISA and lectin binding assay as described in

section 3.9.3 and 3.12.2. Meanwhile, the UDP-Gal monitoring was performed using

RP-HPLC analysis as described in section 3.13. As expected, UDP-Gal

concentration decreased gradually once the Sf-9 cells were coinfected with hTf and

�1,4-GalT. The pattern of the curve showed the same trend as Figure 4.52, even

though the cell culture was coinfected with hTf and �1,4-GalT. Figure 4.57 shows

that �1,4-GalT and hTf accumulation rates increased proportional to the time of

infection. However, not all hTf was galactosylated due to the limitation of UDP-Gal

(refer to Fig. 4.26). Thus, it can be concluded that even though hTf and �1,4-GalT

accumulation increased upon the time of coinfection, the gradual decrease of sugar

nucleotide (UDP-Gal) still affect the effectiveness of the galactosylation process.

Figure 4.57: Relationships among the three main elements in in vivo galactosylation

process. hTf as the sugar acceptor, �1,4-GalT as the enzyme and UDP-Gal as the

substrate donor in the process. Error bars indicate ±S.D of duplicates data.

170

4.7 Purification

Purification of all kinds of transferrin have been reported in quite a number of

paper. Among these, Stuart Ali. et. al. (1996) and Eric Ailor et. al. (2000) had

purified recombinant transferrin from sf9 and Tn cells. Purification involving Phenyl

sepharose and Q-Sepharose can give up to 95% pure transferrn. Phenyl sepharose

together with affinity gel had been used to purify testicular transferrin from rat sertoli

cells and the 100% pure transferrin give 28% overall recovery (Michael K., 1984).

Viera. A.V (1993) reported that single step purification of avian transferrin, using

HIC gave a yield of 80% purified serotransferrin. Different sources of transferrin, or

recombinant from different expression system need different extent of purifying;

some need simple purification, some need few steps to purify. The insect

cell/baculovirus system is not considered a “clean” secretion system. (Altmann, F.,

1999). Methods reported in Stuart Ali. et. al. (1996) and Eric Ailor et. al. (2000) were

used as main references. Since the method can give up to 95% rhTf, purity,

optimization of the purification would mostly focus on recovery or productivity of

the columns.

4.7.1 Profile of Sample Elution from hydrophobic Interaction

Chromatography

A combination of step and gradient elution of transferrin from Phenyl

Sepharose was done to understand the elution profile of sample from HIC column.

Sample was loaded at 15µg rhTf/ml of gel. Column was equilibrated with 100%

buffer 1.2M Ammonium Sulphate/0.4M Sodium Citrate (buffer A), pH6.0. Proteins

was eluted with 3 column volumes of 50% 1 buffer A, followed by gradient elution

using 50% to 25% buffer A in 10 column volume and lastly 3 column volume of

water. Figure 4.58 show that a large amount of unwanted protein was hydrophilic

and washed out by equilibration buffer. The third and second large peaks in figure

4.58 showed that 50% buffer A eluted protein without rhTf and 0% buffer A or water

(buffer B) eluted strong bound proteins which contained a small amount of rhTf,

respectively. Elution of a broad rhTf peak started when the buffer contained less

171

than 35% buffer A and stopped before 25% buffer A. rhTf pooled from fractions 80

to fraction 111 gave 34% recovery and did not give satisfactory selectivity.

0

0.2

0.4

0.6

0 20 40 60 80 100 120 140 160Fraction

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml

AbsSteps & Gradient ElutionrhTf

Figure 4.58: Steps and gradient elution of rhTf from column HIC. Sample was

loaded at 15µg rhTf /ml of gel. Unbound protein was washed out when the column

was equilibrated with 100% buffer A. Elution started with 3 column volumes of 50%

buffer A and followed by gradient elution using 50% to 25% buffer A in 10 column

volume. rhTf was pooled from fraction 80 to fraction 111 and the recovery was 34%.

4.7.2 Hydrophobic Interaction Chromatography Optimization

HIC can be used as the capture step. where recovery of target proteins is

more important compared to resolution. Type of gel, type of buffer and pH of buffer

which affect resolution was maintained as mentioned previously (Stuart Ali. et. al.,

1996; Eric Ailor et. al., 2000). Elution method, flowrates and loading capacity were

optimized in order to improve capacity, recovery and ease of use.

4.7.2.1 Optimization of Elution Method

Gradient elution may give high resolution but step wise elution is technically

simpler, reproducible and also able to elute interest protein in a more concentrated

172

form. (Amersham Bioscience). A few approach using steps wise elution have been

done to increase the resolution in the area where the peak of interest elutes without

affecting the recovery of rhTf. Strength of elution buffer is optimized to elute all less

strongly bound compounds, but must not exceed the level where peak of interest start

to co-elute. As mentioned in section 4.7.1, rhTf was eluted between 35% to 25% of

buffer A, which mean 1st elution step can be optimized using buffer containing 50%-

35% buffer A. The second step elution was fixed at 25% buffer A because 25%

buffer A is expected to give complete elution of rhTf with minimum unwanted

compound (Figure 4.58).

HIC using 3 different stepwise elutions involving 50% buffer A, 45% buffer

A and 35% buffer A was studied as the 1st elution buffer at a fixed flowrate of

0.5ml/min. 32±2µg rhTf per bed volume was loaded. Recovery of rhTf was highest

when 50% buffer A was used as 1st elution buffer, 25% buffer A as 2nd elution buffer

(Table 4.14). Chromatograms and SDS-PAGE in figure 4.59 characterized the

elution profile of the 3 different step wise elution methods. Although 1st step elution

with lower percentage of buffer A increased the elution of unwanted compound

(Figure 4.59) but the recovery of rhTf was low (Table 4.14). Eluted rhTf didn’t show

significant differences in resolution for the various step elutions applied (Figure

4.59). Hence, step wise elution using 50% buffer A as 1st elution buffer and 25%

buffer A as 2nd elution buffer which gave 64% recovery was chosen as the best

elution method.

Table 4.14: Optimization of step-wise elution method for achieving higher recovery

of rhTf. Different step elutions were optimized at a fixed flowrate, 0.5ml/min.

Loaded rhTf per bed volume was fixed at 32±2µg/ml. Recovery of rhTf was

percentage of pooled rhTf over total loaded rhTf.

No Step elution Flowrates

(ml/min)

Recovery of

rhTf (%)

a 50% Buffer A and 25% Buffer A 0.5 64

b 45% Buffer A and 25% Buffer A 0.5 42

c 35% Buffer A and 25% Buffer A 0.5 29

173

0

0.1

0.2

0.3

0.4

0.5

0 20 40 60 80Fraction

Abs

at U

V 2

80nm

0

20

40

60

80

100

120

Step

s(Pe

rcen

tage

of

Buf

fer

A(%

) &

C

oncn

etra

tion

of r

Htf

in µ

g/m

l

AbsSteps ElutionrhTf

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 20 40 60 80 100 120 140Fraction

Abs

at U

V 2

80nm

0

20

40

60

80

100

120

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml

Abs

Steps Elution

rHtf

0

0.1

0.2

0.3

0.4

0.5

0.6

0 20 40 60 80 100 120Fraction

Abs

at U

V 2

80nm

0

20

40

60

80

100

120

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml

AbsSteps ElutionrHtf

*Volume for the first 24 fractions collected during equilibration was double compared to other

fractions

Figure 4.59: HIC chromatogram for the optimization of elution method. Each

chromatogram, (a) to (c) showed the rhTf elution profile of the respective study

mentioned in Table 4.14. The peak characterizing the elution of rhTf at 25% Buffer

A was pooled. Lane of SDS-PAGE, (a) to (c) characterized the sample, which was

pooled from experiment (a) to (c); m is marker, s is standard hTf

4.7.2.2 Optimization of elution flowrate

Flow rate and sample load are interrelated. Flow rate and sample load are

optimized to find highest productivity where resolution is still high enough to meet

the predefined purity requirement. For this optimization of elution flowrate, an

average loading capacity of rhTf, 33±5µg/ml of gel and the optimized step elution

m a s b c *a

b c

174

buffer was fixed. The prime consideration when optimizing for highest possible

productivity is to find the highest possible sample load over the shortest possible

sample application time with acceptable loss in yield.

The elution flowrate which gave highest recovery was 1ml/min and followed

by 0.5ml/min and 2ml/min (Table 4.15). Figure 4.60 shows the elution profile of

rhTf at different elution flowrates. High elution flowrate will always give a decrease

in dynamic binding capacity, affect elution profile and recovery of step elution.

Eluting at low flowrate without compensating with high loading will result in loss of

protein due to dilution. This may be the reason for the lower recovery of rhTf for

elution flowrates at 2ml/min and 0.5ml/min. Hence, in this study, optimized elution

flowrate is 1ml/min which give 74.6% recovery.

Table 4.15: Optimization of elution flowrate. Elution flowrate at 0.5ml/min,

1ml/min and 2 ml/min were studied at fix steps elution. Loaded rhTf per bed

volume was fixed at 33±5µg/ml. Recovery of rhTf was percentage of pooled rhTf

over total loaded rhTf.

No Flowrates

Step

Elution Recovery,%

A 0.5ml/min 50/25 64

B 1ml/min 50/25 74.6

C 2ml/min 50/25 62.1

175

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 20 40 60 80Fraction

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

m)


0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 20 40 60Fraction

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml)

AbsSteps Elut ionrhTf

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 10 20 30 40 50Fraction

Abs

at U

V 2

80nm

0

20

40

60

80

100

120

Perc

enta

ge o

f B

uffe

r A

(%)

&

C

once

ntra

tion

of r

-Htf

in m

g/m

l

AbsSteps ElutionrhTf

Figure 4.60: HIC chromatogram for the optimization of elution flowrate. Each

chromatogram, (a) to (c), showed the elution profile of rhTf, representing the

respective study mentioned in Table 3. The peak characterizing the elution of rhTf at

25% Buffer A was pooled.

a

c

b

176

4.7.2.3 Optimization of rHtf loading capacity

After fixing the elution mode and the elution flowrates, rhTf loading capacity

was also be optimized to get the optimum level that gives highest recovery. Table

4.16 and figure 4.61 showed that maximum loading capacity of rhTf at optimized

flowrates and steps elution is 55µg/ml gel. Figure 4.61 shows the relationship

between loading capacity and recovery percentage. Loading of rhTf between 30-

60µg/ml gel is expected to result in the recovery of more than 70%. Figure 4.62 and

Figure 4.63 characterize the elution profile of rhTf. Elution profile of rhTf was

affected by loading capacity. Binding strength of rhTf become weaker and was

eluted earlier, using higher percentage of buffer A. Hence, the recovery of rhTf was

decrease even though all the rhTf was bound to the gel during equilibration stage

(Figure 4.62d).

Table 4.16: Optimization of rhTf loading capacity. Loading capacity of rhTf was

studied at optimized flowrates and steps elution. Recovery of rhTf was percentage of

pooled rhTf over total loaded rhTf.

No Flowrates Step Elution Total rhTf/ml of gel Recovery

A 1ml/min 50/25 15 µg/ml 67.5%

B 1ml/min 50/25 38 µg/ml 74.6%

C 1ml/min 50/25 55 µg/ml 79.7%

D 1ml/min 50/25 74 µg/ml 40%

177

30

40

50

60

70

80

90

10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0

Loaded rhTf per bed volume (µg/ml)

Rec

over

y pe

rcen

tage

(%

)

Figure 4.61: The relationship between recovery percentage and loading capacity

178

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 20 40Fract ion

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

110

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml)


0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 20 40 60Fract ion

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml)

AbsSteps Elut ionrHtf

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 20 40 60Fract ion

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

110

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml)


0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 20 40 60 80Fraction

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml)


Figure 4.62: HIC chromatogram for the optimization of rhTf loading capacity.

Chromatograms (a), (b) ,(c) & (d) characterize the elution profile of rhTf, from the

respective study mentioned in Table 4.16. The peak characterizing the elution of rhTf

at 25% buffer A was pooled.

a b

c d

179

(a) (b) (c)

(d) (e)

Figure 4.63: SDS-PAGE characterizing the elution profile of rhTf. (a) impurity

eluted at 0% buffer A; (b), (c), (d) & (e) showed the eluted fractions at 25% buffer A

of experiment A to D mentioned in Table 4.16. SDS-PAGE (e), showed more

unwanted impurity compare to the others. It was predicted that early elution of

unwanted protein, which supposed to be eluted at 0% buffer A, was due to high

loading.

4.7.3 Batch Purification

Batch purification was carried out to screen the rhTf binding capacity of

weak and strong anion exchanger in tris and phosphate buffer and the effect of pH

and concentration buffer upon rhTf binding capacity. Q-Sepharose with 20mM

Tris/HCl buffer, pH8.5 gave highest binding capacity (figure 4.64, 4.65, 4.66)

180

0.0

0.5

1.0

1.5

2.0

2.5

3.0

DEAE Sephadex A-25 Q-Sepharose

Type of Buffer

Bin

ding

Cap

acity

of

Ani

on E

xcha

nge

Mat

rix/

(mg

htf/

ml) 20mM Tris-HCl buffer, pH7.5

20mM Sodium Phosphate

Figure 4.64: Binding capacity of two anion exchange matrix with Tris and

phosphate buffer used as equilibration buffer.

2.40

2.60

2.80

3.00

3.20

3.40

3.60

6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0pH Buffer

Bin

ding

Cap

acity

(m

g H

tf/m

l of

mat

rix)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Binding capacity/(ug ofbinded htf/ml of matrix)Concentration of Htf inSupernatant/(ug/ml)

Figure 4.65: Binding capacity of Q-Sepharose with equilibration buffer of

different pH. Optimization buffer was carried out to achieve higher recovery of

rhTf.

181

2.60

2.70

2.80

2.90

3.00

3.10

3.20

0 20 40 60 80 100Concentration of Buffer/mM

Bin

ding

Cap

acity

( µg

Htf

/ml o

f m

atri

x)

-0.05

0.00

0.05

0.10

0.15

0.20

Con

cent

ratio

n of

Htf

in S

uper

nata

nt (

µg/m

l)

Binding capacity/(ug ofbinded htf/ml of matrix)

Concentration of Htf inSupernatant/(ug/ml)

Figure 4.66: Binding capacity of Q-Sepharose with different concentration of

buffer Tris-HCl buffer, pH8.5 as equilibration buffer.

4.7.4 Anion Exchange Chromatography

Anion exchange chromatography with Q sepharose was used as the polishing

step in rhTf purification in which 20mM Tris/HCl buffer, pH8.0 was used as

equilibration buffer and gradient elution of 0-100% KCl (Stuart Ali. et. al. 1996; Eric

Ailor et. al., 2000). In this work, the relationship between matrix and buffer with

rhTf binding capacity was obtained (Section 4.7.3). Parameters which show higher

rhTf binding capacity in the screening steps are similar with the parameters of Q-

Sepharose Chromatography mentioned in Stuart Ali. et. al, (1996) except pH 8.5 was

used instead of pH8.0. Since the condition resulted in rhTf of high purity (Stuart Ali.

et. al, 1996), not much further optimization work was done to improve selectivity

except for applying shallow gradient elution and slow flow rates. In this section, Q-

Sepharose was carried out with shadow gradient elution by increasing 50mM NaCl

per bed volume and slow flow rate, 0.5ml/min. 100% pure rhTf was obtained (table

4.17). Characteristic of the rhTf elution profile from Q-Sepharose is shown in Figure

4.69 and Figure 4.70.

182

4.7.5 Characterization of rhTf purification

After optimization, a compete sequence of rhTf purification was carried out

using the optimized parameters. Crude sample with 0.5% yield of rhTf, was

harvested at day 6 post infection. Sample was loaded to column Phenyl Sepharose 6

fast flow at 38µg rhTf/ml of gel and at a flowate of 1ml/min. Column was

equilibrated with 100 buffer A, protein was eluted with a step wise sequence profile

of 50% buffer A, 25% buffer A and water. 74.56% of rhTf was recovered. After

dialysis for 24 hours, sample was loaded to Q-Sepharose fast flow. 20mM Tris-HCl,

pH 8.5 was used as the equilibration buffer. Gradient elution was initiated with

increasing of 50mM NaCl to100mM NaCl in 5 column volume. 100% pure rhTf with

34% overall recovery was achieved (Table 4.17). Figures 4.67 and 4.68 characterize

the elution profile of rhTf from Phenyl Sepharose column. Figures 4.69 and 4.70

characterize the elution profile of rhTf from Q-Sepharose. Figure 4.71 qualified the

improve purity of sample from crude to final purification step.

Table 4.17: Summary of the characteristic of purification of rhTf

Parameter Sample Phenyl

Sepharose Dialysis Q Sepharose

Volume (ml) 23.00 59.00 79.50 36.20

Htf Concentration (µg/ml) 31.00 9.01 5.81 6.69

Protein Concentration

(µg/ml) 6200.00 - 24.47 6.62

Total Htf (µg) 713.00 531.63 461.50 242.08

Total Protein (µg/ml) 142600.00 - 1945.66 239.70

Purity (%): 0.5 23.72 100.99

Recovery (%) 74.56 86.81 52.46

Overall Recovery (%) 74.56 64.73 33.95

183

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 20 40 60Fraction

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

70

80

90

100

Perc

enta

ge o

f B

uffe

r A

(%)

&

Con

cent

ratio

n of

r-H

tf in

µg/

ml)


Figure 4.67: HIC chromatogram characterizing the separation and elution profile of

sample. Chromatography was carried out using optimized flowrate, step elution and

suitable loading.

m. marker

a. Fraction-40

b. Fraction-42

c. Fraction-43

d. Fraction-45

e. Fraction-47

f. Fraction-49

g. Fraction-51

h. Fraction-53

i. Fraction-55

Figure 4.68: SDS-PAGE characterizing the separated protein from phenyl sepharose

6 fast flow column. Protein in fractions 40, 42, 43, 45, 47, 49, 51, 53, 55 were shown

in 9%, silver staining, SDS-PAGE. M is molecular weight standards.

a b c m d e f g h i

184

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

0 20 40 60 80 100 120 140Fraction

Abs

at U

V 2

80nm

0

10

20

30

40

50

60

Perc

enta

ge o

f B

uffe

r B

/% &

Con

cent

ratio

n of

Htf

in µ

g/m

l

Abs

Steps & Gradient Elution

rhTf

Figure 4.69: Anion exchange chromatogram characterizing the separation and elution

profile of sample of after HIC and after dialysis. Q-Sepharose Chromatography was

carried out with 20mM Tris HCl, pH 8.5 as the equilibration buffer.

m- marker

a. Fraction-69

b. Fraction-71

c. Fraction-73

d. Fraction-75

e. Fraction-77

f. Fraction-79

g. Fraction-81

h. Fraction-83

i. Fraction-85

Figure 4.70: SDS-PAGE characterizing the separated protein from Q-Sepharose

column. Protein in fractions 69, 71, 73, 75, 77, 79, 81, 83, 85 were shown in 9%,

silver staining, SDS-PAGE. M is molecular weight standards.

a m b c d e f g h i

185

Figure 4.71: SDS-PAGE characterizing the sample pooled from each purification

step. M is molecular weight standards, (a) is supernatant sample harvest at day 6 post

infection, (b) is sample after hydrophobic interaction chromatography and after

dialysis and (c) is pure rhTf after anion exchange chromatography.

225kDa 150kDa 100kDa 75kDa 50kDa 55kDa 35kDa

m a b

CHAPTER 5

CONCLUSION AND RECOMMENDATIONS

5.3 Fundamental Study of Sf9 Cells Growth

A good inoculum is a prerequisite for successful cell growth and cell

infection with wild type and recombinant baculoviruses. Thus, in the first part of this

study, parameters which optimize the growth rate of Sf-9 cells culture were

investigated. The parameters investigated were the effects of serum, different types

of media, initial cell density, cell subculturing conditions as well as spent medium

carry-over. Serum affected viable cell numbers positively. However, since serum

contained trace amount of sugar nucleotides and enzymes which may interfere with

protein assay, serum free media was used for the rest of the experiments. In this

study, SF-900II SFM was found to support cell growth better than TC-100. In

addition, high concentration of inoculum, subculturing at early exponential phase and

fresh medium without spent medium carry-over resulted in an insect culture with

high viable cell numbers and fast growth rate.

A low-cost 24-well plate insect cell culture technique was utilized to aid in a

high throughput optimization of insect cell growth and recombinant protein

expression. The growth of Sf9 cells in 24-well plates was found to mimic the growth

in shake flasks. By performing the optimization in 0.5mL culture volumes in

standard 24-well plates, the cost and time associated with optimization process and

the amount of baculovirus required for optimization were greatly reduced. However,

187

the miniaturized experiment could not mimic exactly the output of a large-scale

production, and the results did not guarantee economical feasibility. Nevertheless,

the data obtained from the miniaturized experiments were in overall alignment with

the results produced in larger scale. Thus, the small-scale optimization evaluations

had provided a very helpful direction in terms of virus infections, cell densities, time

point of infection, harvest time and protein integrity which were all necessary for

large-scale production.

5.4 Mock Infection and the Expression Profile of rhTf

For the mock baculovirus infection, the interaction of the infection factors

especially multiplicities of infection (MOI) and spent medium carry-over with the

above parameters were also investigated. In order to achieve higher viral infectivity,

the MOI range should be within the range of 1 to 15. Furthermore, the medium must

also be replenished during the exponential phase before viral infection.

This research has greatly contributed to the knowledge of the behaviour of

rhTf expression in baculovirus insect cells expression system using Sf9 cells

monolayer and serum free medium. MOI, time of infection, seeding density, and

harvest time were found to significantly affect the production of rhTf. The maximum

rhTf obtained in the monolayer culture was approximately 11.2�g/ml. As for

induction, no specific inducers were added as it occurred through natural infection

and gene expression has been observed (Ailor et al., 2000; Tomiya et al., 2003). rhTf

yield in the infected monolayer culture was still low when compared to the average

yield of other recombinant proteins expressed in this system. A very good reason for

this was because the Sf9 cells were not able to propagate further due to monolayer

disadvantages. Further study on the expression and optimization of rhTf expression

had been carried out based on the results obtained from monolayer culture. This

work helps to monitor any changes that would occur when working with suspension

culture and decides on how to optimize rhTf expression.

188

5.3 Strategic Optimization of the Baculovirus Insect Cell Expression System.

Screenings of culture medium and recombinant baculovirus should be the

first steps towards a strategic optimization of the baculovirus insect cell expression

system. Plackett-Burmann screening design had successfully identified a few

candidates that displayed significant effects towards rhTf production. Although there

were many variables to screen, Plackett-Burman screening design allowed a feasible

number of experiments to be conducted and enough information were gathered for

analysis. For recombinant baculovirus screening, the method of end point dilution

was practically easy and the results were reliable. The purpose of conducting

baculovirus screening was to ensure virus integrity for subsequent optimization

works.

The use of central composite design and response surface methodology had

been demonstrated to be useful in optimizing an output of a biological process. The

effect of the test variables could be studied simultaneously, thus maximizing the

amount of information gathered for limited time and number of experiments. The

regression model obtained in this work was highly effective and the nutrients had

significant effects on rhTf production. This work had successfully increased the rhTf

yield by three-fold from 19.89 �g/ml to 65.12 �g/ml.

5.4 Study of Galactosylation

Three main elements to ensure successful protein galactosylation are the

presence of sufficient amount of hTf as the substrate acceptor, �1,4-GalT as the

enzyme and UDP-Gal as the substrate donor. Unfortunately, the limitation in the

elongation of the N-glycan processing of hTf in insect cells occurs due to the lack of

�1,4-GalT needed to produce galactosylated hTf. Thus, in this current study, a

proposed strategy for the production of galactosylated hTf is the introduction of GalT

artificially to the cell cultures infected with AcMNPV-hTf. This can be

accomplished through in vivo or in vitro manners.

189

Analysis of secreted AcMNPV-hTf and AcMNPV-�1,4-GalT expression had

showed that their production rates increased over the time of infection. These were

confirmed by numerous analyses including SDS-PAGE, western blot, TLC, ELISA

and lectin binding assay. A simple explanation is the nature of baculovirus infection

cycle itself. Upon infection, the cells’ mechanism will be shifted to viral

multiplication and expression of its genes. Hence, the recombinant protein secretion

will increase upon time of infection and will be secreted into the environment. To

examine another element involved in galactosylation processing, native UDP-Gal

level at normal and upon AcMNPV-hTf infection had been monitored using RP-

HPLC. It revealed that substrate donor concentration decreased upon time of

infection.

After the three elements’ expression and monitoring analyses were

successfully established, the next step was to perform different levels of

galactosylation. Apo hTf containing two N-glycosylation sites that included Gal

residues was used as a standard for comparison with others. Since AcMNPV-hTf

produced in insect cell culture was not satisfactorily galactosylated due to the

deficiency of the enzyme to construct the N-glycan chain, a strategy involving the

introduction of artificial enzyme was investigated. To examine this strategy, in vivo

galactosylation was conducted using coexpression of AcMNPV-hTf and AcMNPV-

�1,4-GalT in cultured cell. Also, in vitro study was carried out by the introduction of

commercial GalT and UDP-Gal to the harvested AcMNPV-hTf supernatant.

Although coexpression of AcMNPV-hTf and AcMNPV-�1,4-GalT resulted in

galactosylated recombinant hTf, the reduction of UDP-Gal upon infection still limit

the extent of galactosylation process. On the other hand, for the in vitro

galactosylation, the commercial UDP-Gal was able to provide sufficient amount of

sugar nucleotide in the processing pathway.

The relationships among the three main elements in in vivo galactosylation

process are found to be very interesting. As expected, UDP-Gal concentration

decreased gradually once the Sf-9 cells were coinfected with the baculovirus coding

the genes for hTf and �1,4-GalT. The hTf accumulation rate increased proportional

to the time of infection, but not all of the hTf were galactosylated due to the

limitation of UDP-Gal. This was proven by the time course analysis of UDP-Gal

190

upon coexpression of hTf and �1,4-GalT. The conclusion from this study was that

even though the model protein hTf and enzyme �1,4-GalT accumulation increased

upon the time of coinfection, the gradual decrease of sugar nucleotide still affect the

effectiveness of the galactosylation process.

5.5 Study of Purification In this work, HIC and IEX had been used to purify rhTf from sf9, HIC as

capture- intermediate steps and IEX as polishing step. Column optimizations had

been performed to improve capacity and recovery. For the HIC column, the

separation matrix was Phenyl Sepharose 6 Fast flow (high Sub), equilibration or

application buffer was 1.2M Ammonium Sulphate/ 0.4M Sodium Citrate, pH6.0 in

water and running temperature was 27oC. Step wise elution method, flow rates and

loading capacity which were closely related each other were studied and optimized.

The maximum loading capacity of rhTf at optimized flowrates, 1ml/min and

optimized steps elution (50% buffer A as 1st elution buffer; 25% buffer A as 2nd

elution buffer) was 55µg/ml gel. Loading capacity between (30-60) µg/ml of matrix

is suggested. As for IEX, the matrix was Q-Sepharose fast flow, equilibration buffer

was 20mM Tris-HCl, pH 8.5, gradient elution with 10%-20% 0.5M NaCl in 5

column volume and flowrate was 0.5ml/min. 100% pure rhTf with 34% overall

recovery was achieved

5.6 Recommendations

To produce recombinant glycoprotein in insect cells with a more

“humanised” form, there are several recommendations for further studies:

(a) N-glycans of the insect cells may be improved using in vitro

glycosylation, which utilizes the specific enzyme to transfer the sugar

to the protein after it is secreted from cell culture.

(b) More sensitive, high-throughput, and detailed analytical techniques

for the detection of enzyme activities, glycan structures, donor sugar

191

nucleotides, intracellular metabolites and extracellular structures need

to be established.

(c) Engineering of N-glycan processing pathway by means of genetic

manipulation to include the necessary processing enzymes.

(d) Use of an alternative insect cell line that may contain mammalian-like

N-glycan processing capabilities.

Most proteins released to human blood system or other human fluids are

glycoproteins. Glycoproteins are involved in the reproductive system and

metabolism of human being. This research is hoped to assist in the effective

expression of human glycoprotein for the benefit of biopharmaceutical industry and

of course human race. The characterization of the recombinant glycoproteins will

help further study on human proteins and therefore contribute to the cure of

glycoprotein-related diseases. Further study on the improvement of the recombinant

glycoprotein expression can be made as follow:

For commercialization purpose, a product should be produced in vast

quantities to meet the demand. Scale up can be simply defined as a procedure for the

design and construction of a large scale system on the basis of the results of small

scale experiments. Engineering efforts have been focused on maintaining the

volumetric oxygen transfer constant when scaling up. Other than that, it is also

important that culture medium especially serum free medium is available at a large

amount at any time. One can study the design of a personalized medium for large

scale culture to meet the glycoprotein requirements. Other variables that may affect

scale up are oxygen uptake, nutrients depletion, power consumption, mixing time,

shear rate and heat transfer coefficient.

REFERENCES

Abdul-Rahman B., Ailor E., Jarvis D. L., Betenbaugh M. J., Lee Y. C. (2002). �-1,4-

galactosyltransferase activity in native and engineered insect cells measured

with time-resolved Eu-fluorescence. Carbohydrate Research. 337: 2181-2186.

Adinarayana, K. and Ellaiah, P. (2002). Response Surface Optimization of the

Critical Medium Components for the Production of Alkaline Protease by a

newly isolated Bacillus sp. J. Pharm. Pharmaceut. Sci. 5(3):272-278.

Ailor E., Takahashi N., Tsukamoto Y., Masuda K., Rahman, B.A., Jarvis D.L., Lee

Y.C., and Betenbaugh M. J. (2000). N-glycan patterns of human transferrin

produced in Trichoplusia ni insect cells: effects of mammalian

galactosyltransferase. Glycobiology. 10: 837-847.

Ailor, E., Takahashi, N., Tsukamoto, Y., Masuda, K., Rahman, B. A., Jarvis, D. L.,

Lee, Y. C. and Betenbaugh, M. J. (2000). N-glycan patterns of human

transferrin produced in Trichoplusia ni insect cells: effects of mammalian

galactosyltransferase. Glycobiology. 10:837-847.

Aisen P. (1989). Physical biochemistry of the transferrin: update, 1984-1988. In

Lochr, T.M. (ed.), Iron Carriers and Iron Proteins. VCH Publishers, New

York, pp. 353-371.

Aisen, P. and Listowsky, I. (1980). Iron transport and storage proteins. Annu. Rev.

Biochem. 49:357-393.

193

Akhnoukh, R., Kretzmer, G. and Schugerl, K. (1996). On-line monitoring and

control of the cultivation of Spodoptera frugiperga Sf9 insect cells and beta-

galactosidase production by Autographa californica virus vector. Enzyme

Microb. Technol. 18: 220-228.

Ali, S. A., Joao, H. C., Csonga, R., Hammerschmind, F. and Steinkasserer, A.

(1996). High yield production of functionally active human serum transferrin

using a baculovirus expression system and its structural characterization.

Biochem. J. 319:191-195.

Altman F., Schwihla H., Staudacher E., Glossl J., Marz L. (1995). Insect cells

contain an unusual, membrane-bound beta-N-acetylglucosaminidase probably

involved in the processing of protein N-glycans. J. Biol. Chem. 270: 17344-

17349.

Altmann F. (1996). N-glycosylation in insects revisited. Trends Glycosci. Glycotech.

8: 101-114.

Altmann F., Kornfeld G., Dalik T., Staudacher E., and Glossl J. (1993). Processing of

asparagine-linked oligosaccharides in insect cells. N-

acetylglucosaminyltransferase and II activities in cultured lepidopteran cells.

Glycobiology. 3: 619-625.

Altmann F., Marz L. (1995). Processing of asparagines-linked oligosacharides in

insect cells: evidence for α-mannosidase II. Glycoconjugate. 12: 150-

155.Altmann F., Staudacher E., Wilson I. B., Marz L. (1999). Insect cells as

hosts for the expression of recombinant glycoproteins. Glycoconjugate J. 16:

109-123.

Altmann, F., Staudacher, E., Wilson, B. H. and März, L. (1999). Insect cells as hosts

for the expression of recombinant glycoproteins. Glycoconjugate Journal.

16:109-123.

194

Amado M., Almeida R., Schwientek T., and Clausen H. (1998). Identification and

characterization of large galactosyltransferase gene families:

galactosyltransferases for all functions. Biochem. Biophys. Acta. 1473:35-53.

Amersham Biosciences. Hydrophobic interation chromatography. Principles and

methods. 1993

Amersham Biosciences. Ion exchange chromatography. Principles and methods.

Aoki D., Appert H. E., Johnson D., Wong S. S., and Fukuda M. N. (1990). Analysis

of the substrate binding sites of human galactosyltransferase by protein

engineering. EMBO. J. 9: 3171-3178.

Arif, B. (1986). Molecular Biology of Baculoviruses. Berlin: Springer. 31-50.

Attie, A. D., Gretch, D. G., Sturley, S. L. and Beckage, N. E. (1995). Production of

recombinant proteins in insect larvae. (U.S Patent 5,472,858).

Aumiller J. J., and Jarvis D.L. (2002). Expression and functional characterization of

a nucleotide sugar transporter from Drosophila melanogaster: relevance to

protein glycosylation in insect cell expression systems. Protein Expr. Purif.

26: 436-448.

Ayres M. D., Howard S. C., Kuzio J., Lopez-ferber M. and R. D. Possee. (1994). The

complete DNA sequences of Autographa californica nuclear polyhedrosis

virus. Virology. 202: 586-605.

Bahia, D., Cheung, R., Buchs, M., Geisse, S. and Hunt, I. (2004). Optimisation of

insect cell growth in deep-well blocks: development of a high-throughput

insect cell expression screen. Protein Expr. Purif. 39:61-70.

Bailey, J. E. and Ollis, D. F. (1986). Biochemical Engineering Fundamentals. 2nd ed.

Singapore: McGraw Hill.

195

Barnes, D. and Sato, G. (1980). Methods for growth of cultured cells in serum-free

medium. Anal. Biochem. 102(2):255-270.

Barrett A. J., Mavroudis D., Tisdale J., Molldrem J., Clave E., Dunbar C., Cottler-

Fox M., Phang S., Carter C., Okunnieff P., Young N.S., Read E.J. (1998). T

cell-depleted bone marrow transplantation and delayed T cell add-back to

control acute GVHD and conserve a graft-versus-leukemia effect. Bone

Marrow Transplant. 21(6): 543-551.

Bartfeld N.S. and Law J.H. (1990). Isolation and molecular cloning of transferrin

from the tobacco hornworm, Manduca sexta. J.Biol. Chem. 265: 21684-

21691.

Bédard, C., Tom, R. and Kamen, A. A. (1993). Growth, nutrient consumption and

end product accumulation in Sf-9 and BTI-EAA insect cell cultures: insights

into growth limitation and metabolism. Biotechnol. Prog. 9:615–624.

Bedard, C., Kamen, A. A., Tom, R. and Massie, B. (1994). Maximization of

recombinant protein yield in the insect cell/baculovirus system by one time

addition of nutrients to high density batch culture. Cytotechnology. 15:129-

138.

Bédard, C., Perret, S. and Kamen, A. A. (1997). Fed-batch culture of Sf-9 cells

supports 3.3 x 107 cells per ml and improves baculovirus-expressed

recombinant protein yields. Biotechnology Letters. 19:629-632.

Beeley, J. G. (1987). Glycoprotein & proteoglycan techniques. Belgium: Elsevier.

Beljelarskaya, S. N. (2002). A Baculovirus Expression System for Insect

Cells. Molecular Biology. 36( 3): 281-292.

Belew, M., Yafang, M., Bin, L., Berglöf, J., Janson, J-C. (1991) Purification of

recombinant hepatitis B surface antigen produced by transformed Chinese

hamster ovary (CHO) cell line grown in culture. Bioseparation. 1:397–408.

196

Belyaev, A. S., Hails, R. S. and Roy, P. (1995). High-level expression of five foreign

genes by a single recombinant baculovirus. Gene. 156:229-233.

Bentley, W. E., Wang, M. Y. and Vakharia, V. N. (1994). Development of an

efficient bioprocess for poultry vaccines using high-density insect cell

culture. Annals of the New York Academy of Sciences. 745:336-359.

Berninsone P. M., and Hirachberg C.B. (2000). Nucleotide sugar transporters of the

Golgi apparatus. Curr. Opin. Struct. Biol. 10: 542-547.

Bhatia, R., Jesionowski, G., Ferrance, J. and Ataai, M. M. (1997). Insect cell

physiology. Cytotechnology. 24:1-9.

Blissard G. W., and G. F. Rohrmann. (1990). Baculovirus diversity and molecular

biology. Annual review of Entomology. 53: 127-155.

Blissard G.W. (1996). Baculovirus--insect cell interactions. Cytotechnology. 20:73-

93.

Blissard, G. W. and Monsma, S. C. (1998). Baculovirus cloning system. (U.S Patent

5,750,383).

Bollag, D. M., Rozycki, M. D. and Edelstein, S. J. (1996). Protein Methods. 2nd ed.

U.S.A.: Wiley-Liss, Inc.

Bonnerjera, J., Oh, S., Hoare, M., Dunhill, P. (1986). The right step at the right time.

Bio/Technology. 4:954-958.

Boyce, F. M. (1999). Use of a baculovirus to express and exogenous gene in a

mammalian cell. (U.S Patent 5,871,986).

Boyer, R. F. (1993). Modern Experimental Biochemistry. USA.: Benjamin Inc. Bull,

J.

197

Braunagel and Summers. (1994). Autographa californica nuclear polyhedrosis virus,

PDV, and ECV viral envelopes and nucleocapsids: structural proteins,

antigens, lipid and fatty acid profiles. Virology. 202(1) : 315-328.

Breitbach K., and Jarvis D.L. (2001). Improved glycosylation of a foreign protein by

Tn-5B1-4 cells engineered to express mammalian glycosyltransferases. J.

Biotechnol. Bioeng. 74: 230-239.

Brew Vanaman T.C., Hill R.L. (1968). The role of alpha-lactalbumin and the A

protein in lactose synthetase: a unique mechanism for the control of a

biological reaction. Proc Natl Acad Sci U S A. 59(2): 491-497.

Brodbeck U. and Ebner K. E. (1966). The subcellular distribution of the A and B

proteins of lactose synthetase in bovine and rat mammary tissue. The Journal

of Biological Chemistry. 241: 5526-5532.

Brodbeck U., Denton W. L., Tanahashi N., Ebner K. E. (1967). The isolation and

identification of the B protein of lactose synthetase as alpha-lactalbumin. J.

Biol Chem. 242(7): 1391-1397.

Bullen J. J., Ward C. G., and Rogers H. J. (1990). Iron, infection and the role of

bicarbonate. FEMS Microbiol. Lett. 71: 27-30.

Butler, M., Huzel, N., Barnabe, N., Gray, T. and Bajno, L. (1999). Linoleic acid

improves the robustness of cells in agitated cultures. Cytotechnology. 30:27-

36.

Butters T. D. and Hughes R. C. (1981). Isolation and characterization of mosquito

cell membrane glycoproteins. Biochim, Biophys Acta. 640: 655-671.

Butters T. D., Hughes R. C., and Vischer P. (1981). Steps in the biosynthesis of

mosquito cell membrane glycoproteins and the effects of tumicamycin.

Biochim. Biophys. Acta. 640: 672-686.

198

C., Godfray, H. C. J. and O'Reilly, D. R. (2001). Persistence of an Occlusion-

Negative Recombinant Nucleopolyhedrovirus in Trichoplusia ni Indicates

High Multiplicity of Cellular Infection. Applied and Environmental

Microbiology. 67:5204-5209.

Caron, A. W., Archambault, J. and Massie B. (1990). High Level Recombinant

Protein Production in Bioreactors Using the Baculovirus Insect Cell

Expression System. Biotechnol. Bioeng. 36:1133-1140.

Cerutti, M., Croizier, G., Croizier, L. and Devauchelle, G. (1996). Modified

baculovirus, its preparation process and its application as a gene expression

vector. (U.S Patent 5,583,023).

Chen, W. and Bahl, O. P. (1991). Recombinant carbohydrate and selenomethionyl

variants of human choriogonadotropin. J. Biol. Chem. 266:8192-8197.

Chiou, T. W., Hsieh, Y. C. and Ho, C. S. (2000). High density culture of insect cells

using rational medium design and feeding strategy. Bioprocess Engineering.

22:483-491.

Christie, W. W. (1989). Gas Chromatography and Lipids: A Practical Guide.

Glasgow: The Oily Press Ltd. Christie, W. W. (1997). Analysis of Fatty

Acids by High Performance Liquid Chromatography. Lipid Technology.

9:124-126.

Cochran, W. G. and Cox, G. M. (1992). Experimental Designs. 2nd edn. New York:

John Wiley and Sons. 335-375.

Comings, D.E., Miguel, A.G., Lesser, H.H.. (1979). Nuclear proteins. VI.

ractionation of chromosomal non-histone proteins using hydrophobic

chromatography. Biochim. Biophys. Acta. 563:253–260.

Cornell, J. A. (1990). How to Apply Response Surface Methodology. U.S.A.:

American Society for Quality Control.

199

Crowther, J. R. (1995). ELISA. U.S.A.: Humana Press, Inc. Dalal, N. G. and

Bentley, W. E. (1999).Mathematical characterization of insect cell (Sf-9)

death in suspended culture. Biotechnology Letters. 21: 325-329.

D’Agostaro G., Bendiak B., and Tropak M. (1989). Eur. J.Biochem. 183: 211-217.

David R. O’Reilly, Lois K.Miller, Verne A. Luckow. (1994). Baculovirus expression

vectors: a laboratory manual. New York, Oxford, Oxford University Press.

David R. O’Reilly, Rosemary S.Hails, Thomas J. Kelly. (1998). The impact of host

developmental status on baculovirus replication. Journal of Invertebrate

Pathology. 72: 269-275.

David T. R., Schuler M. L., Granados R. R., Wood H. A. (1993). Comparison of

oligosaccharide processing among various insect cell lines expressing a

secreted glycoprotein. In Vitro Cell Dev Biol Anim. 29A: 842-846.

David, P. C. and Lonnie, D. R. (1997). Molecular Biology. U.S.A.: Cache River

Press.

Davidson D. J., Fraser M. J., and Castillino F. J. (1990). Oligosaccharide processing

in the expression of human plaminogen cDNA by lepidopteran insect

(Spodoptera frugiperda) cells. Biochemistry. 29: 5584-5590.

Delong, B., Albee, A., Deeds, Z., Gifford, J., Ross, S., Kao, K. and Caple, M. (2004).

Development of an Efficient Medium Optimization Kit for Factorial Matrix

Design – A Statistical Approach to Increase Cell Growth and Productivity of

Recombinant CHO Cells. Sigma-Aldrich Biotechnology. unpublished.

Devauchelle, G., Cerutti, M. and Cahoreau, C. (1996). Modified baculovirus and

baculovirus expression vectors. (U.S Patent 5,571,709).

Dill, K.A.. (1990a.) Dominant forces in protein folding. Biochemistry. 29:7133–

7155.

200

Drews, M, Paalme, T. and Vilu, R. (1995). The growth and nutrient utilization of the

insect cell line Spodoptera frugiperda Sf9 in batch and continuous culture. J.

Biotechnol. 40:187-198.

Effertz K., Hinderlich S., Reutter W. (1999). Selective loss of either the epimerase or

kinase activity of UDP-N-acetylglucosamine 2-epimerase/N-

acetylmannosamine kinase due to site-directed mutagenesis based on

sequence alighments. J. Biol Chem. 274: 28771-28778.

Er-el, Z., Zaidenzaig, Y., Shaltiel, S. (1972) Hydrocarbon-coated Sepharoses. Use in

the purification of glycogen phosphorylase. Biochem. Biophys. Res.

Commun. 49:383-390.

Faulkner, P. (1993). Insect Cell Culture Engineering. U.S.A.: Marcell Dekker.

Ferrance, J. P., Goel, A. and Ataai, M. M. (1993). Utilization of glucose and amino

acids in insect cell cultures: quantifying the metabolic flows within the

primary pathways and medium development. Biotechnol. Bioeng. 42:697-

707.

Fertig G., Kloppinger M. and Miltenburger H. G.. (1990). Cell cycle kinetics in

insect cell cultures compared to mammalian cell cultures. Exp. Cell. Res.

189: 208-212.

Fielding, J. and Speyer, B. E. (1974). Iron transport intermediates in human

reticulocytes and the membrane binding site of iron-transferrin. Biochem

Biophys Acta. 363(3):387-396.

Frank, M. B. (1998). Baculovirus Summary. Oklahoma City. unpublished.

Frank, M. B. (1998). Insect Cell Culture & Baculovirus Infections. Oklahoma City.

unpublished.

201

Fraser, M. J. (1986). Ultrastructural observations of virion maturation in Autographa

californica nuclear polyhedrosis virus infected Spodoptera frugiperda cell

cultures. Journal of Ultrastructure and Molecular Structure Research. 95:189-

195.

Freshney, R. I. (2000). Culture of Animal Cells. U.S.A.: Wiley-Liss.

Frosch, C. (1997). Biosynthesis of N-Glycans. Institute of Toxicology University of

Mainz. unpublished.

Fu D. and van Halbeek, H. (1992). N-glycosylation site mapping of human

serotransferrin by serial lectin affinity chromatography, fast atom

bombardment-mass spectrometry and 1H nuclear magnetic resonance

spectroscopy. Anal. Biochem. 206: 53-63.

Garcia, F.A.P., Pires, E.M.V. (1993). Chromatography. In: Kennedy, J.F., Cabral,

J.M.S. eds. Recovery Processes for Biological Materials. London: Wiley.

415–451.

Gerardy-Schahn R., Oelmann S., and Bakker H. (2001). Biochimic (Paris). 83: 775-

762.

Gilbert, R. S., Nagano, Y., Yokota, Y., Hwan, S. F., Fletcher T. and Lydersen, K.

(1996). Effect of lipids on insect cell growth and expression of recombinant

proteins in serum free medium. CytoTechnology. 22:211-216.

Glick, B. R. and Pasternak, J. J. (1990). Molecular Biotechnology: Principles and

Applications of Recombinant DNA. U.S.A.: American Society For

Microbiology.

Goochee, C. F., Gramer, M. J., Andersen, D. C., Bahr, J. B. and Ragmussen, J. R.

(1991). The oligosaccharides of glycoproteins: bioprocess factors affecting

oligosaccharide structure and their effect on glycoprotein properties.

Bio/Technol. 9:1346-1355.

202

Gorfien, S. F., Paul, W., Judd, D., Tescione, L. and Jayme, D. W. (2003). Optimized

Nutrient Additives for Fed-Batch Cultures. Biopharm. International. 34-40.

Gotoh, T., Chiba, K. and Kikuchi, K. I. (2004). Oxygen consumption profiles of Sf-9

insect cells and their culture at low temperature to circumvent oxygen

starvation. J. Biochem. Eng. 17:71–78.

Grabenhorst E., Hofer B., Nimtz M., Jager V. and Conradt H. S. (1993). Biosynthesis

and secretion of human interleukin 2 glycoprotein variants from baculovirus-

infected Sf-21 cells. Characterization of polypeptides and posttranslational

modifications. Eur J Biochem. 32: 11087-11099.

Granados, R. R. and Hashimoto, Y. (1995). Gene coded for a polypeptide which

enhances virus infection of host insects. (U.S Patent 5,475,090).

Grioner, A. (1986). Specificity and safety of baculoviruses. In: Granados, R. A. and

Federici, B. A. (Eds.). The biology of baculoviruses. Boca Raton, FL: CRC

Press. 1:1-37.

Guarino, L. A. and Jarvis, D. L. (1991). Use of baculovirus early promoters for

expression of foreign genes in stably transformed insect cells. (U.S Patent

5,077,214).

Guarino, L. A. and Jarvis, D. L. (1992). Use of baculovirus early promoters for

expression of foreign genes in stably transformed insect cells. (U.S Patent

5,162,222).

Guo S., Sato T., Shirane K., and Furukawa K. (2001). Galactosylation of N-linked

oligosaccharides by human ß-1,4-galactosyltransferases I, II, III, IV, V, and

VI expressed in Sf-9 cells. Glycobiology. 11: 813-820.

Haaland, P. D. (1989). Experimental Design In Biotechnology. New York, U.S.A.:

Marcel Dekker, Inc.

203

Halperin, G., Breitenbach, M., Tauber-Finkelstein, M., Shaltiel, S. (1981).

Hydrophobic chromatography on homologous series of alkyl agaroses. A

comparison of charged and electrically neutral column materials. J.

Chromatog. 215:211–228.

Hard K., van Dorn J. M., Thomas-Oates J. E., Kamerling J. P. and van der Horst D.J.

(1993). Structure of the ASN-linked oligosaccharides of apolipoprotein III

from the insect Locusta migraforia. Chemistry. 32: 766-775.

Harrison, R. G. (1994). Protein Purification Process Engineering. U.S.A.: Marcell

Dekker.

Hasmann, F. A., Cortez, D.V., Junior, A. P. and Roberto, I. C. (2003). Optimization

of beta-xyloxidase recovery by reversed micelles using response surface

methodology. Electronic Journal of Biotechnology. 6(2):20-27.

Hassell, T., Gleave, S. and Butler, M. (1991). Growth Inhibition in Animal Cell

Culture: The Effect of Lactate and Ammonia. Appl. Biochem. Biotechnol.

30:29-41.

Hill R.L. (1979). Glycosyltransferases in analysis of oligosaccharide structure and

function. UCLA Forum Med. Sci. 21: 63-86.

Hinderlich S., Stasche R., Zeitier R., Reutter W. (1997). A bifunctional enzyme

catalyzes the first two steps in N-acetylneuraminic acid biosynthesis of rat

liver. Purification and characterization of UDP-N-acetylglucosamine 2-

epimerase/N acetylmannosamine kinase. J. Biol Chem. 272: 24313-24318.

Hink, W. F, Thomsen, D. R., Davidson, D. J., Meyer, A. L. and Castellino F. J.

(1991). Expression of three recombinant proteins using baculovirus vectors in

23 insect cell lines. Biotechnol. Prog. 7:9-14.

Hirschberg C. B. (2001). Golgi nucleotide sugar transport and leukocyte adhesion

deficiency II. J.Clin. Invest. 108: 3-6.

204

Hirschberg C. B., Robbins P. W., and Abeijon C. (1998). Transporters of nucleotide

sugars, ATP and nucleotide sulfate in the endoplasmic reticulum and Golgi

apparatus. Annu. Rev. Biochem. 67: 49-69.

Hjertén, S. (1973). Some general aspects of hydrophobic interaction

chromatography. J. Chromatog. 87:325–331.

Hjertén, S., Rosengren, J., Påhlman, S. (1974). Hydrophobic interaction

chromatography. The synthesis and the use of some alkyl and aryl derivatives

of agarose. J. Chromatog. 101:281–288.

Hjertén, S. (1981). Hydrophobic interaction chromatography of proteins, nucleic

acids, viruses and cells on non-charged amphiphilic gels. In: D. Glick, ed.

Methods of Biochemical Analysis. John Wiley & Sons, Inc. 89–108.

Hjertén, S., Yao, K., Eriksson, K.-O., Johansson, B. (1986). Gradient and isocratic

High Performance Hydrophobic Interaction Chromatography of proteins on

agarose columns. J. Chromatog. 359:99–109.

Hofstee, B.H.J. (1973). Hydrophobic affinity chromatography of proteins. Anal.

Biochem. 52:430–448.

Hofstee, B.H.J, Otillio, N.F. (1978). Non-ionic adsorption chromatography of

proteins. J. Chromatog. 159:57–69.

Hollister J. R., Shaper J. H., Jarvis D. L. (1998). Stable expression of mammalian

beta 1,4-galactosyltransferase extends the N-glycosylation pathway in insect

cells. Glycobiology. 8: 473-480.

Hollister J. R. and Jarvis D. L. (2001). Engineering lepidopteran insect cells for

sialoglycoprotein production by genetic transformation with mammalian beta

1,4-galactosyltransferase and alpha 2,6-sialytransferase genes. Glycobiology.

11: 1-9.

205

Hollister J. R., Grabenhorst E., Nimtz M., Conradt H., and Jarvis D. L. (2002).

Engineering the protein N-glycosylation pathway in insect cells for

production of biantennary, complex N-glycans. Biochemistry. 41: 15093-

15104.

Hooker A. D., Green N. H., Baines A. J., Bull A. T., Jenkins N., Strange P. G., James

D. C. (1999). Constraints on the transport and glycosylation of recombinant

IFN-gamma in Chinese hamster ovary and insect cells. Biotechnol. Bioeng.

63: 559-572.

Hrkal, Z., Rejnkova, J. (1982). Hydrophobic interaction chromatography of serum

proteins on Phenyl-Sepharose CL-4B. J. Chromatog. 242: 385–388.

Hsieh P. and Robbins P. W. (1984). Regulation of asparagines-linked

oligosaccharide processing. Oligosaccharide processing in Aedes albopitus

mosquito cells. J. Biol. Chem. 259: 2975-2382.

Hsu T. A., Takahashi N., Tsukamoto Y., Kato K., Shimada I., Masuda K., Whiteley

E. M., Fan J. Q., Lee Y. C., Betenbaugh M. J. (1997). Differential N-glycan

patterns of secreted and intracellular IgG produced in Trichoplusia ni cells. J

Biol Chem. 272(14): 9062-9070.

Hu, Y. C., and Bentley, W. E. (2000). A kinetic and statistical-thermodynamic model

for baculovirus infection and virus-like particle assembly in suspended insect

cells. Chemical Engineering Science. 55:3991-4008.

Hu, Y. C., Lu, J. T. and Chung, Y. C. (2003). High-density cultivation of insect cells

and production of recombinant baculovirus using a novel oscillating

bioreactor. Cytotechnology. 42:145-153.

Huebers, H. A., Finch, C. A. (1987). The physiology of transferrin and transferrin

receptors. Physiol. Rev. 67(2):520-82.

206

Huebers, H., Csiba, E., Josephson, B., Huebers, E. and Finch, C. (1981). Interaction

of human diferric transferrin with reticulocytes. Proc Natl Acad Sci.

78(1):621-625.

Iatrou, K. (1998). Methods of expressing proteins in insect cells and methods of

killing insects. (U.S Patent 5,759,809).

Iatrou, K., Farrell, P. J. and Hashimoto, Y. (2000). Baculovirus artificial

chromosomes and methods of use. (U.S Patent 6,090,584).

Ignoffo C. M. (1975). Entomopathogens as insecticides. Environ Lettert. 8(1): 23-40.

Ikonomou, L., Bastin, G., Schneider, Y. J. and Agathos, S. N. (2001). Design of an

efficient medium for insect cell growth and recombinant protein production.

In Vitro Cell Dev. Biol. Animal. 37:549-559.

Inlow, D., Shauger, A. and Maiorella, B. (1989). J. Tissue Cult. Methods. 12:13-16.

Janson, J.-C., Ryde´n, L. (1993). Protein separation and purification. In: Rehm, H.-J.,

Reed, G. eds. Biotechnology, vol. 3. VCH. Weinheim. 617–642.

Janson, J-C., Låås, T. (1978). Hydrophobic interaction chromatography on Phenyl-

and Octyl-Sepharose CL-4B. In: Roger, E. Ed. Chromatography of synthetic

and biological macromolecules. Chichester, England: Ellis Horwood Ltd.

Jarvis D. L. (1997). Baculovirus Expression Vectors. In Miller, L.K. (ed.). The

Baculoviruses. Plenum, New York, pp. 389-431.

Jarvis D. L. (2003). Developing baculovirus-insect cell expression systems for

humanized recombinant glycoprotein production. Virology. 310: 1-7.

Jarvis D. L. and Summers M. D. (1989). Glycosylation and secretion of human tissue

plasminogen activator in recombinant baculovirus-infected insect cells. Mol

Cell Biol. 9: 214-223.

207

Jarvis, D. L. and Carrington, J. C. (1993). Method and vector for the purification of

foreign proteins. (U.S Patent 5,179,007).

Jarvis D. L., Finn E. E. (1996). Modifying the insect cell N-glycosylation pathway

with immediate early baculovirus expression vectors. Nat. Biotechno.l. 14:

1288-1292.

Jennissen, H.P., Heilmeyer, I.M.G. (1975). General aspects of hydrophobic

chromatography. Adsorption and elution characteristics of some skeletal

muscle enzymes. Biochemistry. 14:754–760.

Jesionowski, G. A. and Ataai, M. M. (1997). An efficient medium for high protein

production in the insect cell/baculovirus expression system. Biotechnol Prog.

13:355-360.

Kalil, S. J., Maugeri, F. and Rodrigues, M. I. (1999). Response Surface Analysis and

Simulation as a Tool for Bioprocess Design and Optimization. Process

Biochem. 35(6):539-550.

Kang, C. Y. (1993). Baculovirus expression system capable of producing foreign

gene proteins at high levels. (U.S Patent 5,194,376).

Karin, M. and Mintz, B. (1981). Receptor-mediated endocytosis of transferrin in

developmentally totipotent mouse teratocarcinoma stem cells. J Biol Chem.

256(7):3245-52.

Kawakita M., Ishida N., Miura N., Sun Wada G. H., and Yoshioka S. (1998). O-

Acetylserine and O-acetylhomoserine sulfhydrylase of yeast. Further

purification and characterization as a pyridoxal enzyme. J.Biochem. 123:

777-785.

Kawar Z., Herscovics A., Jarvis D. L. (1997). Isolation and characterization of an

alpha 1,2-mannosidase cDNA from the lepidopteran insect cell line Sf-9.


208

Kawar Z., Romero P. A., Herscovics A., Jarvis D. L. (2000). N-glycan processing by

a lepidopteran insect alpha 1,2-mannosidase. Glycobiology. 10: 347-355.

Kennedy, R.M., 1990. Hydrophobic chromatography. In: Deutscher, M.P. (Ed.).

Guide to Protein Purification, Methods Enzymology, vol. 182. London:

Academic Press Inc.339–343.

Keshvara L. M., Newton E. M., Good A. H., Hindsgaul O., and Palcic M. M. (1992).

Enzyme-Linked Immunosorbent Assays for the Measurement of Blood

Group A and B Glycosyltransferase Activities. Glycoconjugate J. 9: 16-20.

Keusch J., Lydyard P. M., Isenberg D. A., and Delves P.J. (1995). �1,4-

galactosyltransferase activity in B cells detected using a simple ELISA-based

assay Glycobiology. 5: 365-370.

King, L. A. and Possee, R. D. (1992). The baculovirus expression system: a

laboratory guide. London: Chapman and Hall.

Kioukia N., Nienow A. W., Emery A.N., Al-Rubeai M.. (1995). Physiological and

environmental factors affecting the growth of insect cells and infection with

baculovirus. Journal of Biotechnology. 38: 243-251.

Knudson, D. and Harrap, K. (1976). Replication of nuclear polydydrosis virus in a

continuous cell culture of Spodoptera frugiperda: microscopy study of the

sequence of events of the virus infection. J. Virol. 17:254-268.

Kobayashi, M., Kato, S., Omasa, T., Shioya, S. and Suga, K. (1994). Enhancement

effects of BSA and linoleic acid on hybridoma cell growth and antibody

production. Cytotechnology.15:51-56.

Kool M., Vlak J. M. (1993). The structural and functional organization of the

Autographa californica nuclear polyhedrosis virus genome. Arch Virol.

130(1-2): 1-16.

209

Kool, M., Voncken, J. W., Vanlier, F. L. J., Tramper, J. and Vlak, J. M. (1991).

Detection and analysis of Autographa Californica Polyhedrosis Virus mutants

with defective interfering properties. Virology. 183:739-746.

Korth K. L., Levings C. S. (1993). Baculovirus expression of the maize

mitochondrial protein URF13 confers insecticidal activity in cell cultures and

larvae. Proc Natl Acad Sci USA. 90: 3388-3392.

Kubelka V., Altmann F., Kornfeld G. and Marz L. (1994). Structures of the N-linked

oligosaccharides of the membrane glycoproteins from three lepidoptran cell

lines (Sf-21, IZD-Mb-0503, Bm-N). Arch. Biochem. Biophysc. 308: 148-157.

Kuehn, L., Meyer, H., Reinauer, H. (1980). Hydrophobic interaction chromatography

as a tool in insulin receptor study. Proc. 2nd. Intl. Insulin Symp. 243–250.

Kulakosky, P. C., Hughes, P. R. and Wood, H. A. (1998). N-linked glycosylation of

a baculovirus-expressed recombinant glycoprotein in insect larvae and tissue

culture cells. Glycobiology. 8:741-745.

Kumamoto K., Goto Y., Sekikawa K., Takenoshita S., Ishida N., Kawashita M., and

Kannagi R. (2001). Increased Expression of UDP-Galactose Transporter

Messenger RNA in Human Colon Cancer Tissues and Its Implication in

Synthesis of Thomsen-Friedenreich Antigen and Sialyl Lewis A/X

Determinants. Cancer Res. 61: 4620-4627.

Kumar, S. and Miller, L. K. (1987). Effects of serial passage of Autographa

Californica Polyhedrosis Virus in cell culture. Virus Res. 7:335-350.

Kuroda K., Geyer R., Doergler W. and Klenk H. D. (1990). The oligosaccharides of

influenza virus hemmaglutinin expressed in insect cells by a baculovirus

vector. Virology. 174: 418-428.

210

Kuroda, K., Geyer, H., Geyer, R., Doerfler, W. and Klenk, H. D. (1990). The

oligosaccharides of influenza virus hemmaglutinin expressed in insect cells

by a baculovirus vector. Virology. 174:418-329.

L��s, T. (1975). Agar derivatives for chromatography, electrophoresis & gel-bound

enzymes. IV. Benzylated dibromopropanol cross-linked Sepharose as an

amphophilic gel for hydrophobic salting out chromatography of enzymes

with special emphasis on denaturing risks. J. Chromatog. 111:373–387.

Lawrence S. M., Huddleston K. A., Pitts L. R., Nguyen N., Lee Y. C., Vann W. F.,

Coleman T. A., Betenbaugh M. J. (2000). Cloning and expression of the

human N- acetylneuraminic acid phosphate synthase gene with 2-keto-3-

deoxy-D-glycero-D galacto-nononic acid biosynthetic ability. J Biol Chem.

275: 17869-17877.

Lawrence S. M., Huddleston K. A., Tomiya N., Nguyen N., Lee Y. C., Vann W. F.,

Coleman T. A., Betenbaugh M. J. (2001). Cloning and expression of human

sialic acid pathway genes to generate CMP-sialic acids in insect cells.

Glycoconjugate J. 18: 205-213.

Lee, S. C., Leusch, M. S., Luckow, V. A. and Olins, P. O. (1994). Method of

producing recombinant eukaryotic viruses in bacteria. (U.S Patent 5,348,886).

Lefort, S., Ferrara, P. (1986). Hydrophobic adsorbants for the isolation and

purification of biosynthetic human growth hormone from crude fermentation

mixtures. J. Chromatog. 361:209–216.

Leger, D., Campion, B., Decottignies, J. P., Montreuil, J. and Spik, G. (1989).

Physiological significance of the marked increased branching of the glycans

of human serotransferrin during pregnancy. Biochem. J. 257:231-238.

Licari R. and Bailey J. E. (1991). Factors influencing recombinant protein yields in

an insect cell-baculovirus expression system: multiplicity of infection and

intracellular protein degradation. Biotech. Bioeng. 37: 238-246.

211

Licari P. J., Jarvis D. L., Bailey J. E. (1993). Insect cell hosts for the baculovirus

expression vectors contain endogeneous exoglycosidase activity. Biotechnol

Prog. 9: 146-152.

Lindahl, L., Vogel, H.J. (1984). Metal-ion-dependent hydrophobic interaction

chromatography of a-lactalbumins. Anal. Biochem. 140:394–402.

Lineback-Zins, J. and Brew, K. (1980). Preparation and characterization of an NH2-

terminal fragment of human serum transferrin containing a single iron-

binding site. J Biol Chem. 255(2):708-713.

Liu, H. S., Wang, Y. C. and Chiou, T. W. (1995). Olive oil addition in insect cell

cultivation. Biotechnol. Tech. 9:457-460.

Lo N. W., Shaper J. H., Pevsner J., and Shaper N. (1998). The expanding beta 4-

galactosyltransferase gene family: messages from the databanks.

Glycobiology. 8: 517-526

Lokesh J., Michael L. S., and H.Alan Wood. (2001). Production of a sialylated N-

linked glycoprotein in insect cells. Biotechnol Prog. 17: 822-827.

Lopez M., Coddeville B., Langridge J., Plancke Y., Sautiere P., Chaabihi H., Chirat

F., Harduin-Lepers A., Cerutti M., Verbert A. and Delannoy P. (1997).

Microheterogeneity of the oligosaccharides carried by the recombinant

bovine lactoferrin expressed in Mamestra brassicae cells. Glycobiology.

7:635-651.

Lopez M., Tetaert D., Juliant S., Gazon M., Cerutti M., Verbert A., Delannony P.

(1999). O-glycosylation potential of lepidopteran insect cell lines. Biochim

Biophys. Acta. 1427: 49-61.

Luckow V. A. (1995). In Baculovirus Expression Systems and Biopesticides (Shuler

ML, Wood HA, Granados RR, Hammer DA, eds) pp 51-90. New York: John

Wiley Sons.

212

Luckow, V. A. (1991). Cloning and Expression of Heterologous Genes in Insect

Cells with Baculovirus Vectors. In: Ho, C., Prokop, A. and Bajpai, R. (Eds.).

Recombinant DNA Technology and Applications. New York, N.Y: McGraw-

Hill. 1-24.

Luckow, V. A. and Summers, M. D. (1988). Trends in the development of

baculovirus expression vectors. Biotechnol. 6:47-55.

Luis Maranga, Ana S. Coroadinha, and Manuel J. T. Carrondo. (2002). Insect cell

culture medium supplementation with fetal bovine serum and bovine serum

albumin: effects on baculovirus adsorption and infection kinetics. Biotechnol

Prog. 18: 855-861.

Lynn, D. E. (2002). Effects of temperature on the susceptibility of insect cells to

infection by baculoviruses. Methods in Cell Science. 23: 221-225.

MacGillivray, R. T., Mendez, E., Shewale, J. G., Sinha, S. K., Lineback-Zins, J. and

Brew, K. (1983). The primary structure of human serum transferrin. The

structures of seven cyanogen bromide fragments and the assembly of the

complete structure. J. Biol. Chem. 258(6):3543-3553.

MacGillivray, R. T., Mendez, E., Sinha, S. K., Sutton, M. R., Lineback-Zins, J. and

Brew, K. (1982). The complete amino acid sequence of human serum

transferrin. Proc. Natl. Acad. Sci. U S A. 79(8):2504-2508.

MacGillivray, R. T. and Brew, K. (1975). Transferrin: internal homology in the

amino acid sequence. Science. 190(4221):1306-1307.

Maeda S., Kawai T., Obinata M., Fujiwara H., Horiuchi T., Saeki Y., Sato Y.,

Furusawa M. (1985). Production of human alpha-interferon in silkworm

using a baculovirus. Nature. 315: 592-594.

213

Maeda, S., Furusawa, M., Marumoto, Y., Horiuchi, T., Sato, Y. and Saeki, Y. (1992).

Method of producing peptides using baculovirus vectors in cultured cells.

(U.S Patent 5,110,729).

Mahat, M. K., Illias, R. M., Rahman, R. A., Rashid, N. A. A, Mahmood, N. A. N.,

Hassan, O., Aziz, S. A. and Kamaruddin K. (2004). Production of

cyclodextrin glucanotransferase (CGTase) from alkalophilic Bacillus sp.

TS1-1: media optimization using experimental design. Enzyme Microb.

Technol. 35:467-473.

Maiorella B., Inlow D., Shauger A. and Harano D. (1988). Large scale insect cell

culture for recombinant protein production. Biotechnology. 6: 1406-1410.

Maiorella, B., Shauger A., Howarth, B. and Inlow, D. (1988). Supply of lipids in

animal cell culture media. Abstr. Pap. Am. Chem. Soc. 196th (MBTD143).

Maisano, F., Belew, M., Porath, J. (1985). Synthesis of new hydrophobic adsorbents

based on homologous series of uncharged alkyl sulphide agarose derivatives.

J. Chromatog. 321:305–317.

Makhatafze, G.I., Privalov, P.L. (1995). Energetics of protein structure. Adv. Protein

Chem. 47:307–325.

Manneberg M., Friedlein A., Kurth H., Lahm H. W. and Fountoulakis M. (1994).

Structural analysis and localization of the carbohydrate moieties of a soluble

human interferon gamma receptor produced in baculovirus-infected insect

cells. Protein Sci. 3: 30-38.

Mant, C. T. and Hodges, R. S. (1991). HPLC of peptides & proteins. U.S.A.: CRC

Press.

Marchal I., Jarvis D. L., Cacan R., Verbert A. (2001). Glycoproteins from insect

cells: sialylated or not? Biol. Chem. 382: 151-159.

214

Marchal I., Mir A. M., Kmiecik D., Verbert A., and Cacan R. (1999). Use of

inhibitors to characterize intermediates in the processing of N-glycans

syntherized by insect cells: a metabolic study with Sf-9 cell line.


Marteijn, R. C. L., Jurrius, O., Dhont, J., de Gooijer, C. D., Tramper, J. and Martens,

D. E. (2002). Optimization of Feed Medium for Fed Batch Culture of Insect

Cells Using Genetic Algorithm. Biotechnology and Bioengineering. Vol. 81.

Marz L., Altmann F., Staudacher E., Kubelka V. (1995). Protein glycosylation in

insects, in Montreuil J, Vliegenthart J F G, Schachter H (Eds.),

Glycoproteins, Vol 29a. Elsevier, Amsterdam, pp 543-563.

Mason, R. L., Gunst, R. F. and Hess J. L. (2003). Statistical Design and Analysis of

Experiments. 2nd ed. U.S.A.: John Wiley & Sons, Inc.

Matsuura, Y., Yasui, K. and Sato, T. (1993). Recombinant baculovirus. (U.S Patent

5,229,293).

Maulik, S. and Patel, S. D. (1997). Molecular Biotechnology: Therapeutic

Applications and Strategies. U.S.A.: Wiley-Liss.

McGregor, W. C. (1991). Purification and Analysis of Recombinant Proteins.

U.S.A.: Marcell Dekker.

McNair, R.D., Kenny, A.J. (1979). Proteins of the kidney microvillar membrane. The

amphipathic form of dipeptidyl peptidase IV. Biochem. J. 179:379-395.

Mel, M., Sarmidi, M. R., Aziz, R. A., Zain, Z. M. and Jamaluddin, A. A. (1999). A

Study of the Cell Growth of Haemophilus Paragallinarum: Optimization of

pH and Temperature. Malaysian National Biotech. Seminar Proceeding.

215

Melander, W., Horvath, C. (1977). Salt effects on hydrophobic interactions in

precipitation and chromatography of proteins: an interpretation of the

lyotropic series. Arch. Biochem. Biophys. 183:200–215.

Mendonça, R. Z., Palomares, L. A. and Ram´ýrez, O. T. (1999). An insight into

insect cell metabolism through selective nutrient manipulation. J. Biotechnol.

72:61–75.

Miller, L. K. (1988). Baculoviruses as gene expression vectors. Ann. Rev. Microbiol.

42:177-199.

Miller, L. K. (1993). Baculovirus expression vectors. (U.S Patent 5,244,805).

Miyajima, A., Schreurs, J., Otsu, K., Kondo, A., Arai, K. I. and Maeda, S. (1987).

Use of the silkworm, Bombadrix mori, and an insect baculovirus vector for

high-level expression and secretion of biologically active mouse interleukin-

3. Gene. 58:273-281.

Montgomery, D. C. (2001). Design and Analysis of Experiments. 5th ed. U.S.A.:

John Wiley & Sons, Inc.

Montgomery, D. C. and Runger, G.C. (1999). Applied statistics and probability for

engineers. 2nd edn. U.S.A.: John Wiley & Sons, Inc.

Montgomery, D.C. (1996). Design and Analysis of Experiments. 4th ed. U.S.A.:

John Wiley & Sons, Inc.

Montreuil. J., Spik G., and Mazurier J. (1997). Transferrin superfamily. In Montreuil,

J., Vliegenthart, J.F.G. and Schacter H. (eds.) Glycoproteins II. Amsterdam:

Elsevier. pp 203-242

Munk K., Pritzer E., Kretzschmar E., Gutte B., Garten W. and Klenk H. D. (1992).

Carbohydrate masking of an antigenic epitope of influenza virus

216

haemagglutinin independent of oligosaccharide size. Glycobiology. 2: 233-

240.

Murphy, C. I. and Young, E. (1996). Baculovirus vectors for expression of secretory

and membrane-bound proteins. (U.S Patent 5,516,657).

Murphy, K.P., Privalov, P.L., Gill, S.J. (1990). Common features of protein

unfolding and dissolution of hydrophobic compounds. Science. 247:559–561.

Narhi, L.O., Kita, Y., Arakawa, T. (1989). Hydrophobic interaction chromatography

in alkaline pH. Anal. Biochem.182: 266–270.

Narimatsu H., Sinha S., Brew K., Okayama H., and Qasba P.K. (1986). Cloning and

sequencing of cDNA of bovine N-acetylglucosamine (ß1-4)

galactosyltransferase. Proc Natl. Acad. Sci. USA. 83: 4720-4724.

Naveena, B. J., Altaf. M., Bhadrayya, K., Madhavendra, S. S. and Reddy, G. (2004).

Direct Fermentation of Starch to L(+) Lactic Acid in SSF by Lactobacillus

amylophilus GV6 Using Wheat Bran as Support and Substrate: Medium

Optimization Using RSM. Process Biochem. 40(2):681-690.

Nawani, N. N. and Kapadnis, B. P. (2004). Optimization of Chitinase Production

Using Statistics Based Experimental Design. Process Biochem. 40(2):651-

660.

Neerman, J. and Wagner, R. (1996). Comparative analysis of glucose and glutamine

in transformed mammalian cell lines, insect and primary liver cells. J. Cell.

Physiol. 166:152–169.

Nester, Roberts and Nester, (1995). Microbiology. U.S.A.: WCB Publishers.

Nishikawa, K., Fukuhara, N., Liprandi, F., Green, K., Kapikian, A. Z., Chanock, R.

and Gorziglia, M. (1989). VP4 protein of porcine rotavirus strain OSU

217

expressed by a baculovirus recombinant induces neutralizing antibodies.

Virology 173:631-637

O’Reilly, D. R., Miller, L. K. and Luckow, V. A. (1994). Baculovirus Expression

Vectors: A Laboratory Manual. U.S.A.: Oxford University Press.

Ochoa, J.L. (1978). Hydrophobic (interaction) chromatography. Biochimie. 60:1–15.

Ogonah O. W., Freedman R. B., Jenkins N. and Rooney B. C. (1995). Analysis of

human interferon-gamma glycoforms produced in baculovirus infected insect

cells by matrix assisted laser desorption spectrometry. Biochem Soc Trans.

23: 100S.

Ogonash O. W., Freedman R. B., Jenkins N., Kirit P., and Rooney B. C. (1996).

Isolation and characterization of an insect cell line to perform complex N-

linked glycosylation on recombinant proteins. Bio/techniques. 14: 197-202.

Ohman, L. and Haggstrom, L. (2001). Culture medium for insect cells lacking

glutamine and contains ammonium salt. (U.S Patent 6,210,966).

Öhman, L., Ljunggren, J. and Haggstrom, L. (1995). Introduction of a metabolic

switch in insect cells by substrate-limited fed batch cultures. Appl. Microbiol.

Biotechnol. 43:1–8.

Ongkudon, C. M., Rahman, B. A. and Aziz, A. A. (2004). Preliminary Study on the

Optimization of Recombinant Human Transferrin Expression in Insect Cells

Baculovirus System. Proceedings of the 1st National Postgraduate

Coloquium. December 8-9. Penang, Malaysia: 306-311.

Ongkudon, C. M., Rahman, B. A. and Aziz, A. A. (2005). Screening of the Sf900-II

SFM Insect Cell Culture Medium and Recombinant Baculovirus for the

Expression of Recombinant Human Transferrin. Proceedings of the 15th

MSMBB Scientific Meeting. April 4-5. Kuala Lumpur, Malaysia: pp11.

218

O'Reilly, D., Miller, L. and Luckow, V. (1992). In: Baculovirus expression vectors.

New York: W. H. Freeman and Co. pp109-215.

Oscarsson, S., Kårsnås, P. (1998). Salt-promoted adsorption of proteins onto

amphiphilic agarose-base adsorbents. II. Effects of salt and salt concentration.

J. Chromatogr. A. 803:83–93.

Oubihi M., Kitajima K., Kobayashi K., Adachi T., Aoki N., Matsuda T. (1998).

Development of an enzyme-linked immunosorbent assay-based method for

measuring galactosyltransferase activity using a synthetic glycopolymer

acceptor substrate. Anal Biochem. 257(2):169-175.

Pa°hlman, S., Rosengren, J., Hjerte´n, S. (1977). Hydrophobic interaction

chromatography on uncharged Sepharose® derivatives. Effects of neutral

salts on the adsorption of proteins. J. Chromatogr. 131:99–108.

Page, M. J. and Rodgers, B. C. (1992). Baculovirus vectors and methods of use. (U.S

Patent 5,147,788).

Palomares, L. A., Gonzalez, M. and Ram´ýrez, O. T. (1999). Evidence of Pluronic F-

68 direct interaction with insect cells: impact on shear stress protection,

recombinant protein, and baculovirus production. Enzyme Microb. Technol.

26: 324-331.

Palomares, L. A., López, S. and Ram´ýrez, O. T. (2004). Utilization of oxygen

uptake rate to assess the role of glucose and glutamine in the metabolism of

infected insect cell cultures. J. Biochem. Eng. 19:87–93.

Palomares, L. A., Pedroza, J. C. and Ram´ýrez, O. T. (2001). Cell size as a tool to

predict protein productivity of the insect cell-baculovirus expression system.

Biotechnol. Lett. 23:359–364.

Parsegian, V.A., Ninham, B.W. (1970). Temperature-dependent van der Waals

forces. Biophys. J. 10:664–674.

219

Paulson J. C., and Colley K. J. (1989). Glycosyltransferases. Structure, localization,

and control of cell type-specific glycosylation. J. Biol Chem. 264: 17615-

17618.

Peakman, T. C., Page, M. J. and Charles, I. G. (1994). Procaryotic leader sequence in

recombinant baculovirus expression system. (U.S Patent 5,322,774).

Peterson, E.A., Sober, H.A. (1956). Chromatography of Proteins. I. Cellulose ion

exchange adsorbents. J. Amer. Chem. Soc. 78:751-755.

Porath, J. (1986). Salt-promoted adsorption: recent developments. J. Chromatog.

376:331–341.

Porath, J. (1987). Metal ion-hydrophobic, thiophilic and IIelectron governed

interactions and their application to salt-promoted protein adsorption

chromatography. Biotechnol. Prog. 3, 14–21.

Porath, J., Sundberg, L., Fornstedt, N., Olson, I. (1973). Salting-out in amphiphilic

gels as a new approach to hydrophobic adsorption. Nature 245:465–466.

Possee, R. D., Weyer, U. and King, L. A. (1990). Control of Virus Disease.

Cambridge University Press. 53-76.

Privalov, P.L., Gill, S.J. (1988). Stability of protein structure and hydrophobic

interaction. Adv. Protein Chem. 39:191–234.

Radford, K. M., Cavegn, C., Bertrand, M. Bernard, A. R., Reid, S. and Greenfield, P.

F. (1997). The indirect effects of multiplicity of infection on baculovirus

expressed proteins in insect cells: secreted and non-secreted products.

Cytotechnology. 24:73-81.

Reddy, P. R. M., Ramesh, B., Mrudula, S., Reddy, G. and Seenayya, G. (2002).

Production of Thermostable beta-amylase by Clostridium thermosulfurogenes

SV2 in Solid State Fermentation: Optimization of Nutrient Levels Using

Response Surface Methodology. Process Biochem. 39(3):267-277.

220

Reed, L. and Muench, H. (1938). A simple method for estimating fifty percent

endpoints. Am. J. Hyg. 27:493-497.

Reis U., Blum B., Von Specht B. U., Domdey H., Collins J. (1992). Antibody

production in silkworm cells and silkworm larvae infected with a dual

recombinant Bombyx mori nuclear polyhedrosis virus. Biotechnology. 10:

910-912.

Ren J., Bretthuer R. K., and Castellino F. J. (1995). Purification and properties of a

Golgi-derived (alpha 1,2)-mannosidase I from baculovirus-infected

lepidopteran insect cells (IPLB-SF21AE) with preferential activity toward

mannose 6-N-acetylglucosamine 2. Biochemistry. 34: 2489-2495.

Reuveny, S, Kim, Y. J., Kemp, C. W. and Shiloach, J. (1993). Production of

recombinant proteins in high density insect cell cultures. Biotechnol. Bioeng.

42:235-239.

Rippel, G., Szepesy, L. (1994). Hydrophobic interaction chromatography of proteins

on an Alkyl-Superose column. J. Chromatogr. A. 664:27–32.

Rose, D. P. and Conolly, J. M. (1990). Effects of fatty acids and inhibitors of

cicosanoid synthesis on the growth of a human breast cancer cell line in

culture. Cancer Res. 50:7139-7144.

Roseman S. (2001). Reflections on glycobiology. J.Biol. Chem. 276: 41527-41542.

Rosenberg, I. M. (1996). Protein Analysis and Purification. U.S.A.: Birkhauser.

Rosengren, J., Påhlman, S., Glad, M., Hjertén, S. (1975). Hydrophobic interaction

chromatography on non-charged Sepharose derivatives. Binding of a model

protein, related to ionic strength, hydrophobicity of the substituent, and

degree of substitution (determined by NMR). Biochim. Biophys. Acta.

412:51–61.

221

Rudolf, F. B. and Mcintire, L. V. (1999). Science, Engineering, and Ethical

Challenges for the 21st Century. Washington, D.C.: Joseph Henry Press.

Ryll T., and Wagner R. (1991). Improved ion-pair high-performance liquid

chromatographic method for the quantification of a wide variety of

nucleotides and sugar-nucleotides in animal cells. J. Chromatography. 570:

77-88.

Sanderson, C. S., Barford. J. P., Barton, G. W., Wong, T. K. K. And Reid, S. (1999).

A structured, dynamic model for animal cell culture: application to

baculovirus/insect cell systems. J. Biochem. Eng. 3:219-229

Schorp B, Howaldt M. W. and Bailey J. E. (1990). DNA distribution and respiratory

activity of Spodoptera frugiperda population infected with wild-type and

recombinant AcMNPV. J.Biotechnol. 15: 169-185.

Shaltiel, S., Er-el, Z. (1973). Hydrophobic chromatography: Use for purification of

glycogen synthetase. USA: Proc. Nat. Acad. Sci. 70:778–781.

Shaper N., Shaper J. H., Meuth J. L., Fox J. L., Chang H., Kirsch I. R., and Hollis G.

F. (1986). Bovine Galactosyltransferase: Identification of a Clone by Direct

Immunological Screening of a cDNA Expression Library. Proc Natl. Acad.

Sci. USA. 83: 1573-1577.

Shen, C. F., Hawari, J. and Kamen, A. (2004). Micro-quantitation of lipids in serum-

free cell culture media: a critical aspect is the minimization of interference

from medium components and chemical reagents. J. Chromatography.

810:119–127.

Shuler, M. L. and Dee, K. U. (1998). Methods and culture media for inducing single

cell suspension in insect cell lines. (U.S Patent 5,728,580).

Shuler, M. L. and Kargi, F. (2002) . Bioprocess Engineering: Basic Concepts. 2nd

ed. U.S.A.: Prentice Hall.

222

Siemensma, A., Huttinga, H. and Hakkaart, M. (2002). Eliminating Animal Derived

Media. Gen. Eng. News. Vol. 22.

Smith G. E., Summers M. D., Fraser M. J. (1983). Production of human beta

interferon in insect cells infected with a baculovirus expression vector. Mol

Cell Biol. 3(12): 2156-2165.

Srinivasan, R., Ruckenstein, E.(1980). Role of physical forces in hydrophobic

interaction chromatography. Separation & Purification Methods. 9:267–370.

Staudacher E. and Marz L. (1998). Strict order of (Fuc to Asn-linked GlcNAc)

fucosyltransferases forming core-difucosylated structures. Glycoconjugate J.

15: 355-360.

Steward, M. W. (1984). Antibodies: Their structure and function. London: Chapman

and Hall.

Stollar V., Stollar B. D., Koo R., Harrap K. A., and Schlesinger R. W. (1976). Sialic

acid contents of sindbis virus from vertebrate and mosquito cells.

Equivalence of biological and immunological viral properties. Virology. 69:

104-115.

Stult C. L. M., and Macher B.A. (1990). Measurement of beta-Galactosyltransferase

activity in Cell extracts with an ELISA-based assay. Arch. Biochem.

Biophys. 280: 20-26.

Summers M. D., Smith G. E. (1987). A manual method for baculovirus vectors and

insect cell culture procedures. Texas Agricultural Experiment Station Bulletin

No. 1555. Texas A & M University, College Station, Texas.

Summers, M. D. (1992). Insect signal sequences useful to improve the efficiency of

processing and secretion of foreign genes in insect systems. (U.S Patent

5,155,037).

223

Summers, M. D. (1994). Method to improve the efficiency of processing and

secretion of foreign genes in insect systems. (U.S Patent 5,278,050).

Summers, M. D. and Smith, G. E. (1978). Baculovirus structural polypeptides.

Virology. 84: 390-402.

Summers, M. D. and Smith, G. E. (1988). A manual of methods for baculovirus

vectors and insect cell culture procedures. U.S.A.: MicroGeneSys, Inc.

Summers, M. D., Bradfield, J. Y. and Keeley, L. L. (1991). Lepidopteran AKH

signal sequence. (U.S Patent 5,023,328).

Summers, M. D., Oker-Blom and Christian, E. G. (1992). Baculovirus dual promoter

expression vector. (U.S Patent 5,169,784).

Suzuki, T., Usami, A., Oda, K., Mori, H. and Kanaya, T. (1998). Cysteine protease

gene defective baculovirus, process for its production, and process for the

production of economic protein by using the same. (U.S Patent 5,753,220).

Takeuchi M. and Kobata A. (1991). Structures and functional roles of the sugar

chains of human erythropoietins. Glycobiology. 1: 337-346.

Takeuchi M., Takasaki S., Shimada M. and Kobata A. (1990). Role of sugar chains

in the in vitro biological activity of human erythropoietin produced in

recombinant Chinese hamster ovary cells. Journal of Biological Chemistry.

265: 12127-12130.

Taki T., Nishikawa S., Ishii K., and Handa. (1990). A simple and specific assay of

glycosyltransferase and glycosidase activities by an enzyme-linked

immunosorbent assay method, and its application to assay of

galactosyltransferase activity in sera from patients with cancer. J. Biochem.

107: 493-498.

224

Tanford, C. (1973). The Hydrophobic Effect: Formation of Micelles and Biological

Membranes. New York: Wiley.

Taticek, R. A. and Shuler, M. L. (1997). Effect of elevated oxygen and glutamine

levels on foreign protein production at high cell densities using the insect

cells baculovirus expression system. Biotechnol. Bioeng. 54:142-152.

Thorstensen K. and Romslo I. (1990). The role of transferrin in the mechanism of

cellular iron uptake. Biochem. J. 271: 1-10.

Tiselius, A. (1948). Adsorption separation by salting out. Arkiv för Kemi,

Mineralogi Geologi. 26B:1–5.

Tomiya N., Ailor E., Shawn M. Lawrence, Betenbaugh M. J., Yuan C. L. (2001).

Determination of nucleotides and sugar nucleotides involded in protein

glycosylation by high-performance anion-exchange chromatography: sugar

nucleotide contents in cultured insect cells and mammalian cells. Analytical

Biochemistry. 293: 129-137.

Tomiya N., Betenbaugh M. J. and Yuan C. L. (2003). Humanization of Lepidopteran

insect cell produced glycoproteins. Acc. Chem. Res. 36: 613-620.

Tomiya, N., Howe, D., Aumiller, J. J., Pathak, M., Park, J., Palter, K. B., Jarvis,

D.L., Betenbaugh M. J. and Lee Y. C. (2003). Complex-type biantennary N-

glycans of recombinant human transferrin from Trichoplusia ni insect cells

expressing mammalian beta-1,4-galactosyltransferase and beta-1,2-N

acetylglucosaminyltransferase II. Glycbiology. 13:23-34.

Tomiya, N., Narang, S., Lee, Y. C. and Betenbaugh, M. J. (2004). Comparing N-

glycan processing in mammalian cell lines to native and engineered

lepidopteran insect cell lines. Glycoconjugate Journal. 21:343-360.

225

Tramper, J., van den End, E. J., de Gooijer, C. D., Kompier, R., van Lier, F. L. J.,

Usmany, M. and Vlak, J. M. (1990). Annals of the New York Academy of

Sciences. 589:423-430.

Trowbridge I. S., Newman R. A., Domingo D. L. and Sauvage C. (1984). Transferrin

receptors: structure and function. Biochem. Pharmacol. 33: 925-932.

Tsao, E., Mason, M., Cacciuttolo, M., Bowen, S. and Wasserman, G. (1996).

Production of parvovirus B19 vaccine in insect cells coinfected with double

baculovirus. Biotechnology and Bioengineering. 49:130-138.

Vail, P. V., Sutter, G., Jay, D. L. and Gough, D. (1971). Reciprocal infectivity of

nuclear polyhedrosis viruses of the cabbage looper and alfalfa looper. Journal

of Inverterbrate Pathology. 17(3):383-388.

van Die I., van Tetering A., Bakker H., van den Eijnden D. H., and Joziasse D. H.

(1996). Glycosylation in lepidopteran insect cell: identification of a �1,4-N-

acetylgalactosaminyltransferase involved in the synthesis of complex-type

oligosaccharide chains. Glycobiology. 6: 157-164.

Van Oss, C.J., Good, R.J., Chaudhury, M.K. (1986). Nature of the antigen-antibody

interaction. Primary and secondary bonds: optimal conditions for association

and dissociation. J. Chromatogr. 376:111–119.

Vander, Sherman and Luciano. (1994). Human Physiology: Basic Cell Functions.

U.S.A.: McGraw Hill.

Vaughn, J. L. and Fan, F. (1997). Differential requirements of two insect cell lines

for growth in serum free medium. In Vitro. Cell. Dev. Biol. 33:479-482.

Vaughn, J. L., Goodwin, R. H., Tompkins, G. J. and McCawley, P. (1977). The

establishment of two insect cell lines from the insect Spodoptera frugiperda

(Lepidoptera: Noctuidae). In Vitro Cellular and Development Biology.

13:213-217.

226

Velardo M. A., Bretthauer R. K., Boutaud A., Rinhold B., Reinhold V. N., Castellino

F. J. (1993). The presence of UDP-N-acetylglucosamine: Alpha-3-D-

mannoside beta 1,2-N-acetylglucosaminyltransferase I activity in Spodoptera

frugiperda cells (IPLB-Sf-21AE) and its enhancement as a result of

baculovirus infection. J.Bio Chem. 268: 17902-17907.

Vieira A.V., Schinieder WJ (1993). One-step chromatographic method for the

purification of avian serotransferrin. Protein Expr Purif. 4(2):110-3.

Visser, J., Strating, M. (1975). Separation of lipoamide dehydrogenase isoenzymes

by affinity chromatography. Biochim. Biophys. Acta 384:69–80.

Vlak J. M. and G. F. Rohrmann. (1985). Baculovirus polyhedrins. In: viral and

microbial pesticides (KE Sherman & K Maramorosch) Academic Press, N Y.

489-542.

Vlak, J. M. (1997). Baculoviruses and Insect Cells: A Versatile System for the

Production of Recombinant Protein including Antibodies. Advanced

International Biotechnology Course on: Advances and trends in antibody

engineering and manufacture, Leiden, Netherlands.

Vogel, H.J., Lindahl, L., Thulin, E. (1983). Calcium-dependent hydrophobic

interaction chromatography of calmodulin, troponin C and their proteolytic

fragments. FEBS Lett. 157:241–246.

Volkman L. E. (1997). Nucleopolyhedrovirus interactions with their insect hosts.

Adv Virus Res. 48: 313-348.

Volkman, L. E. and Goldsmith, P.A. (1985). Mechanism of neutralization of budded

AcMNPV by a monoclonal antibody: inhibition of entry by adsorptive

endocytosis. Virology. 143:185-195.

Volkman, L. E. and Knudson, D. L. (1986). Persistent baculoviral infections. In:

Granados, R. A. and Federici, B. A. (Eds.). The biology of baculoviruses.

Boca Raton, FL: CRC Press. 1:109-128.

227

Wagner R., Geyer H., Geyer R., Klenk H. D. (1996). N-acetyl-beta-glucosaminidase

accounts for differences in glycosylation of influenza virus hemagglutinin

expressed in insect cells from a baculovirus vector. J. Virology. 70: 4103-

4109.

Walsh, G. and Headon, D. (1994). Protein Biotechnology. U.S.A.: John Wiley &

Sons.

Wang, C. H., Lo, C. F., Kou, G. H., Huang, C. J. and Chou, C. M. (1996).

Oligonucleotides for detection of baculovirus infection. (U.S Patent

5,521,299).

Watanabe A., Takehiro Kokuho, Hitomi Takahashi, Masashi Takahashi, Takayuki

Kubota and Shigeki Inumaru. (2001). Sialylation of N-glycans on the

recombinant proeins expressed by a baculovirus-insect cell system under �-N

-acetylglucosaminidase inhibition. The Journal of Biological Chemistry. 277:

5090 -5093.

Wathen M. W., Aeed P. A. and Elhammer A. P. (1991). Characterization of

oligosaccharide structures on a chimeric respiratory syncytial virus protein

expressed in insect cell line Sf-9. Biochemistry. 30. 2863-2868.

Westermeier, R. (1993). Electrophoresis in Practice. German: VCH Press.

Wolff, M. W., Murhammer, D. W., Jarvis, D. L. and Linhardt, R. J. (1996).

Electrophoretic analysis of glycoprotein glycans produced by lepidopteran

insect cells infected with an immediate early recombinant baculovirus

encoding mammalian beta-1,4-galactosyltransferase. Glycoconjugate Journal.

16:753-756.

Wong, T. K. K., Nielsen, L. K., Greenfield, P. F. and Reid, S. (1994). Relationship

between oxygen uptake rate and time of infection of Sf9 insect cells infected

with a recombinant baculovirus. Cytotechnology. 15:157–167.

Wu, J. and Lee, K. L. D. (1998). Growth promotion by yeastolate and related

components on insect cells. Biotechnol. Tech. 12:67-70.

228

Wu, J., Ruan, Q. and Lam, H. Y. P. (1998). Evaluation of spent medium recycle and

nutrient feeding strategies for recombinant protein production in the insect

cell baculovirus process. J. Biotech. 66:109-116.

Xu, H. Q., Wu, P., Jeff, C. F., Tidwell, C. and Wang A.Y. (2002). A smooth

response surface algorithm for constructing a gene regulatory network. J.

Physiol. Genomics. 11:11-20.

Yadav S. P., and Brew K. (1990). Identification of a region of UDP-galactose: N-

acetylglucosamine beta 4-galactosyltransferase involved in UDP-galactose

binding by differential labeling. J.Biol. Chem. 265: 14163-14169.

Yamaji, H., Tagai, S.I. and Fukuda H. (1999). Optimal production of recombinant

protein by the baculovirus-insect cell system in shake-flask culture with

medium replacement. J. Biosci. Bioeng. 87: 636-641.

Yang, M. and Butler, M. (2000). Effect of Ammonia on the Glycosylation of Human

Recombinant Erythropoietin in Culture. Biotechnol. Prog. 16:751-759.

Yeh J. C., Seals J. R., Murphy C. I., Van Halbeek H. and Cummings R. D. (1993).

Site-specific N-glycosylation and oligosaccharide structures of recombinant

HIV-1 gp120 derived from a baculovirus expression system. Biochemistry.

32: 11087-11099.

Yon, R.J. (1972). Chromatography of lipophilic proteins on adsorbents containing

mixed hydrophobic and ionic groups. Biochem. J. 126:765–767.

Zatta P. F., Nyame K., Cormier M. J., Mattox S. A., Smith D. F., and Cummings R.

D. (1991). A Solid-Phase Assay for b1,4 Galactosyltransferase in Human

Serum Using Recombinant Aequorin. Anal. Biochem. 194: 185-191.

Zuidema D., Klinge-Roode E. C., van Lent J. W., Vlak J. M. (1989). Construction

and analysis of an Autographa californica nuclear polyhedrosis virus mutant

lacking the polyhedral envelope. Virology. 173(1): 98-108.

APPENDIX

APPENDIX 1 Stock Solution for SDS-PAGE

1. 2M Tris-HCl (pH8.8), 100ml

� Weight out 24.2g Tris-base and add to 50ml distilled water.

� Add HCl slowly to pH 8.8

� Add distilled water to total volume 100ml.

2. 1M Tris-HCl (pH 6.8), 100ml

� Weight out 12.1g Tris base and add to 50ml distilled water.

� Add HCl slowly to pH 6.8.

� Add distilled water to total volume 100ml.

3. 10% SDS(w/v), 100ml

� Weight out 10g SDS

� Add distilled water to a total volume 100ml.

4. 50% glycerol (v/v), 100ml

� Pour 50ml 100% glycerol

� Add 50ml distilled water.

5. 1% bromophenol blue (w/v),10ml

� Weight out 100mg bromophenol blue

Bring to 10ml with distilled water, stir until dissolved

230

APPENDIX 2 Working Solution for SDS-PAGE

1. Solution A (Acrylamide Stock Solution), 100ml

� 30% (w,v) acrylamide, 0.8% (w/v) bis-acrylamide

� Weight out 29.2g acrylamide and 0.8g bis-acrylamide and make total

volume to 100ml.

2. Solution B (4x separating gel buffer), 100ml

� 75ml 2M Tris-HCl (pH8.8)

� 4ml 10% SDS

� 21ml distilled water

3. Solution C (4x stacking gel buffer), 100ml

� 50ml 1M Tris-HCl (pH6.8)

� 4ml 10% SDS

� 46ml distilled water

4. 10% ammonium persulfate

� 0.05 g in 0.5ml distilled water

5. Electrophoresis buffer, 1L

� 3g Tris

� 14.4g glycine

� 1g SDS

� Add distilled water to make 1L.

6. 5x sample buffer, 10ml

� 0.6ml 1M Tris-HCl (pH6.8)

� 5ml 50% glycerol

� 2ml 10% SDS

� 0.5ml 2-mercaptoethanol

� 1ml 1% bromophenol blue

� 0.9ml distilled water

231

APPENDIX 3 Separating and Stacking Gel Preparation

1. Separating Gel X% Preparation (see note)

Solution A x/3 ml

Solution B 2.5 ml

H2O (7.5-x/3) ml

10% ammonium persulfate 50ul

TEMED 10ul

2. Stacking Gel Preparation

Solution A 0.67 ml

Solution C 1.0 ml

H2O 2.3 ml

10% ammonium persulfate 30ul

TEMED 10ul

Note: Optimal Resolution Ranges (adapted from Hames, B.D. pp 1-91 in Hames,

B. D. and D. rickwood, eds. 1981. Gel Electrophoresis of Proteins: a Practical

Approach. 290 pages. IRL Press, Oxford and Washington, D.C.)

Acrylamide Percentage Separating Resolution

15 % Gel 15 – 45 kDa

12.5% Gel 15 – 60 kDa

10% Gel 18 – 75 kDa

7.5% Gel 30 – 120 kDa

5% Gel 60 – 212 kDa

232

APPENDIX 4 Coomassie Blue Staining Solution

1. Coomassie blue Stainning solution, 1 liter

1.0 g Coomassie Blue R-250

450ml methanol

450ml distilled water

100ml glacial acetic acid

2. Coomassie blue Destainning solution, 1 liter

100ml methanol

100ml glacial acetic acid

800ml distilled water

VOT 74192 THE DEVELOPMENT AND OPTIMIZATION OF …eprints.utm.my/id/eprint/4281/1/74192.pdf ·...

Documents

Transcript of VOT 74192 THE DEVELOPMENT AND OPTIMIZATION OF …eprints.utm.my/id/eprint/4281/1/74192.pdf ·...