vot 74192 the development and optimization of …eprints.utm.my/id/eprint/4281/1/74192.pdf ·...
TRANSCRIPT
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
10-5 Dilution 10-6 Dilution
10-7 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
SF-900 II SFM TC-100
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
Final infectivity percentage
(%) 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
density and dotted lines represent cell viability
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
density and dotted lines represent cell viability
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
0 24 48 72 96 120Time post-infection (h)
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
0 24 48 72 96 120Time post-infection (h)
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
0 24 48 72 96 120Time post-infection (h)
Infe
ctiv
ity
(%)
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
0 24 48 72 96 120Time post-infection (h)
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
Final infectivity percentage
(%)
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
0 1 2 3 4 5 6 7Elution Time (min)
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
0 1 2 3 4 5 6 7Elution Time (min)
A26
0
-0.050.000.050.100.150.200.250.300.350.400.450.500.550.60
0 1 2 3 4 5 6 7Elution 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
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)
AbsSteps ElutionrHtf
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)
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
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)
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 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)
AbsSteps ElutionrHtf
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)
AbsSteps ElutionrHtf
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.
Glycobiology. 7: 433-443.
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.
Glycobiology. 9: 645-654.
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