OPTIMIZATION OF RECOMBINANT HUMAN TRANSFERRIN EXPRESSION
IN INSECT CELLS BACULOVIRUS SYSTEM
CLARENCE M. ONGKUDON
A thesis submitted in fulfillment of the
requirements for the award of the degree of
Master of Engineering (Bioprocess)
Faculty of Chemical and Natural Resources Engineering
Universiti Teknologi Malaysia
NOVEMBER 2006
iii
Specially dedicated to:
my father Marcellus Ongkudon, mother Juanah Ungit,
Sisters Sibylla; Clarice; Stella; Mellisa, Brother McMarshall,
Uncle Bacon, Auntie Jane,
and my beloved Jessica @ Jess
iv
ACKNOWLEDGEMENT
First and foremost, praise to the Almighty God who from His blessings and
will has enabled me to accomplish this thesis. Next, I wish to express my heartfelt
appreciation to my supervisors, Dr. Azila Bt. Abdul Aziz and Dr. Badarulhisam B.
Abdul Rahman. They have proficiently guided and thought me how to conduct a
systematic and professional research as well as problem solving. I am also indebted
to Dr. Firdausi B. Razali, the Head of Bioprocess Engineering Department for his
motivation and knowledge that he has shared during the Bioreactor Design and
Analysis course. Thanks also to Prof. Dr. Michael Betenbaugh of John’s Hopkins
University USA for providing the recombinant baculovirus.
Many thanks also to the Malaysian Ministry of Science and Technology for
funding this research. To the Sultanah Zanariah librarians, thank you for helping me
to get accessed to all relevant literatures. I would also like to acknowledge the
support people particularly Pn. Siti Zalita, En. Mat, En. Yakub, and En. Malek for
their patience and help whenever I was in the laboratory. My fellow postgraduate
researchers should also be recognized for their ideas and help at various occasions
although it is not possible to list all of them in this limited space.
I can not thank enough to all my family members and friends for their
supports, words of courage and prayers. They have indeed helped me to face the
difficulties that I have encountered along the journey. I wish and pray for their good
health and fortune in the days to come.
v
ABSTRACT
Insect cells-baculovirus expression system is a promising new artificial
system for the production of many therapeutic glycoproteins. This system owns
many of the protein processing and folding mechanisms of mammalian cells and is
capable of expressing a large amount of recombinant proteins. This work aimed at
expressing, optimizing, and characterizing recombinant human Transferrin (rhTf), a
model glycoprotein, at a laboratory scale. In this research, 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 19.89µg/ml to
65.12µg/ml.
vi
ABSTRAK
Sistem ekspresi sel serangga-bakulovirus adalah satu sistem alternatif dalam
penghasilan pelbagai jenis glikoprotein teraputik. Sistem ini memiliki banyak
mekanisma pemprosesan dan penglipatan protein sel mamalia serta mampu untuk
menghasilkan protein rekombinan dalam kuantiti yang besar. Penyelidikan ini
bertujuan untuk mengekspresi, mengoptimum dan mencirikan model glikoprotein
iaitu Transferin manusia rekombinan (rhTf) pada skala makmal. Di dalam
penyelidikan ini, 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 parameterparameter yang telah dioptimumkan, penghasilan rhTf telah meningkat sebanyak 3kali ganda iaitu daripada 19.89ug/ml kepada 65.12ug/ml.
vii
TABLE OF CONTENTS
CHAPTER
TITLE
PAGE
BLANK PAGE
-
THESIS DECLARATION
-
SUPERVISOR’S DECLARATION
-
TITLE PAGE
i
DECLARATION OF ORIGINALITY
ii
DEDICATION
iii
ACKNOWLEDGEMENT
iv
ABSRACT
v
ABSTRAK
vi
TABLE OF CONTENTS
vii
LIST OF TABLES
xiii
LIST OF FIGURES
xv
LIST OF
xviii
SYMBOLS/ABBREVIATIONS/TERMINOLOGY
1
LIST OF APPENDICES
xxii
INTRODUCTION
1
1.1
Preface
1
1.2
Research Problem Background
3
1.3
Research Objective
4
1.4
Research Scopes
4
viii
1.5
2
Research Contributions
5
LITERATURE REVIEW
6
2.1
Insect Cells - Baculovirus Expression System
6
2.1.1
Baculovirus Characteristics
6
2.1.2
Insect Cell Lines
8
2.1.3
The Pros and Cons of the Insect Cells –
9
Baculovirus Expression System
2.1.4
2.2
2.3
2.4
Generation of Recombinant Baculovirus
10
Model Glycoprotein
11
2.2.1
Native Human Transferrin (nhTf)
11
2.2.2
Recombinant Human Transferrin (rhTf)
14
2.2.3
Biosynthesis of N-Glycans
15
Insect Cell Culture Medium
16
2.3.1
Protein Hydrolysates (Peptones)
16
2.3.2
Carbohydrates
17
2.3.3
Amino Acids
17
2.3.4
Lipids
18
2.3.5
Albumin
19
2.3.6
Serum Free Medium (SFM)
19
Optimization of Protein Expression in BEVS
20
2.4.1
20
Physical Factors that Ensure Success of
Expression
2.4.2
Optimization of Recombinant Baculovirus
21
Stock
2.4.3
2.5
Medium Optimization
23
Design of Experiments
2.5.1
Factorial
Experiments
25
in
Completely
25
Randomized Designs
2.5.2
Interactions
26
2.5.3
Coded Variables
26
ix
2.6
2.5.4
Factor Levels Combinations
27
2.5.5
Fractional Factorial Experiments
28
2.5.6
Screening Experiments
29
Analysis of Experiments
30
2.6.1
Correlation
30
2.6.2
Regression Analysis
31
2.6.3
Nonlinear
and
Higher-Order
Regression
32
Analysis
2.7
2.8
Optimization of Experiments
33
2.7.1
34
Improvements of RSM
Specialized Protein Analysis: Theories and Principles
34
2.8.1
34
Sodium Dodecyl Polyacrylamide Gel
Electrophoresis (SDS-PAGE)
2.8.2
Enzyme
Linked
Immunosorbent
Assay
38
(ELISA)
3
2.8.2.1 Basic Immunology
38
2.8.2.2 Principles of ELISA
40
RESEARCH METHODOLOGY
43
3.1
43
3.2
Materials
3.1.1 Cell line and Recombinant Baculovirus
43
3.1.2 Equipments
43
3.1.3 Chemicals
44
Insect Cells Techniques
45
3.2.1
45
The Preparation of TC100 Medium From
Powdered Formulation
3.2.2
Maintenance and Regeneration of Sf9 Cells
45
Monolayer Culture
3.2.3
Cells Freezing
46
3.2.4
Cells Thawing
46
3.2.5
Cells Counting
47
x
3.2.6
Adaptation of Sf9 Cell Culture in Serum Free
48
Medium
3.3
3.2.7
Adaptation of Sf9 Cells in Suspension Culture
48
3.2.8
Sf9 Cells Maintenance in Suspension Culture
49
Baculovirus Techniques
50
3.3.1
Viral Amplification
50
3.3.2
Viral Titration by End Point Dilution Method
51
3.3.3
Generation of Pure Recombinant Virus Stocks
52
by End Point Dilution Method
3.4
Protein Analysis Techniques
53
3.4.1
53
Sodium Dodecyl Sulphate – Polyacrylamide
Gel
Electrophoresis,
SDS-PAGE
under
Denaturing Condition
3.4.2
Bicinchoninic Acid (BCA) Assay
3.4.3
Enzyme
Linked
Immunosorbent
54
Assay,
55
ELISA-Conversion of Calibrated Data to
56
ELISA
3.4.4
Actual Product Concentration
3.5
Recombinant Human Transferrin (rhTf) Expression
57
and Optimization
3.5.1
Optimization
of
rhTf
Expression
in
57
Monolayer Culture
3.6
3.5.2
Medium Screening
57
3.5.3
Medium Optimization in Suspension Culture
58
Response Surface Methodology, RSM (Method of
59
Steepest Ascent)
4
RESULTS AND DISCUSSIONS
61
4.1 The Study of Sf9 Insect Cells Culture Growth Profiles
61
4.1.1
Sf9 Cell Growth in T-flask (Monolayer) and
Shake flask (Suspension)
61
xi
4.1.2
Development of Sf9 Suspension Culture
63
System in 24-well Plate
4.2
4.1.3
Growth Profiles of Infected Cells
65
4.1.4
Growth Analysis
67
Study on the Expression Profiles of rhTf in Infected
69
Sf9 Insect Cells Culture
4.2.1
rhTf Expression at Different MOIs
4.2.2
rhTf
Expression
at
69
Different
Seeding
71
rhTf Expression at Different Times of
73
Densities
4.2.3
Infection
4.3
Optimization of the Recombinant Human Transferrin
76
Expression
4.3.1
Recombinant Baculovirus Screening
76
4.3.2
Medium Screening
81
4.3.3
Medium
Optimization
using
Response
90
Surface Methodology
4.3.3.1 Regression Model
B
4.3.3.2 Nutrients Interactions
4.4
90
94
Characterization of the Optimized Recombinant
98
Human Transferrin Expression
5
CONCLUSIONS
106
5.1 Summaries
106
5.1.1
rhTf
Expression
in
Sf9
Insect
Cells
106
Utilization of 24-well Plates for Insect Cells
107
Monolayer Culture
5.1.2
Suspension Culture
5.2
5.1.3
Medium and Baculovirus Screening
107
5.1.4
Response Surface Methodology
108
Recommendations
108
xii
5.2.1
Large Scale Study in Bioreactor
109
5.2.2
Expression and Purification of Biologically
110
Active Glycoprotein
REFERENCES
111
APPENDICES
130
xiii
LIST OF TABLES
TABLE NO.
TITLE
PAGE
2.1
Seeding densities for typical vessel sizes
21
2.2
Example of a 4-Factor, 2-level Full Factorial Experiment
28
2.3
Example of 12-Run, 11-Factor, 2-Level, Screening
29
Design
4.1
Comparison of Sf9 growth in T-flask, Shaker, and 24-
65
well plate
4.2
rhTf yield coefficients at various seeding density, MOIs,
75
and time of infection
4.3
Concentration of rhTf in each well of a 96-well plate
77
4.4
Poisson distribution data sheet
79
4.5
Factors affecting the end point dilution method
81
4.6
Real values for the screening of 13 selected nutrients
82
using Plackett-Burman design
4.7
13-factor (nutrients), 33-run, 2-level Plackett-Burman
83
xiv
screening design
4.8
Estimated effects on rhTf yield based on the results of
87
Plackett-Burman screening experiments
4.9
Central composite design for the optimization of
91
glutamine, glucose and lipid mixtures 1000x
4.10
Analysis of Variance (ANOVA) of the CCD
93
4.11
Summary of the characteristics of optimized rhTf
98
expression
xv
LIST OF FIGURES
FIGURE NO.
TITLE
PAGE
2.1
Spodoptera frugiperda Sf9 Cells
8
2.2
Insect Cells Baculovirus Expression System
10
2.3
3D structure of the first domain of Human Transferrin
12
2.4
Samples being electrophorase in SDS-PAGE
35
2.5
Chemical structure of a polyacrylamide
36
2.6
Movement of molecules in the porous gel
36
2.7
Basic principles of Direct Sandwich ELISA
41
4.1
Growth curves of Sf9 monolayer culture in 25cm2 T-
62
flask at different seeding densities, SD
4.2
Growth curves of Sf9 suspension culture in 250ml
62
shake flask at different seeding densities, SD
4.3
Growth curves of Sf9 suspension culture in 24-well
plate at different seeding densities, SD
64
xvi
4.4
Growth curves of Sf9 suspension culture in 24-well
64
plate at different seeding densities, SD
4.5
Growth curves of Sf9 infected with AcMNPV
66
4.6
Comparison between uninfected (U), wild-type (WI),
66
and recombinant (R) virus-infected Sf9 cells
4.7
Growth rate constants of Sf9 in various cultivators and
68
at different seeding densities
4.8
Doubling time of Sf9 in various cultivators and at
68
different seeding densities
4.9
SDS PAGE analysis of rhTF expression
70
4.10
rhTf expression profiles at different MOIs
70
4.11
rhTf expression profiles at different seeding densities
72
4.12
Surface plot of figure 4.11
72
4.13
rhTf expression profiles at different times of infection
74
4.14
Surface plot of figure 4.13
74
4.15
3D plot of Table 4.3
77
4.16
Infected cells appearance in medium A (lipid mixtures
84
added) and medium B (no lipid mixtures added)
4.17
rhTf concentration at different medium compositions
based on Plackett-Burman screening experiments
84
xvii
4.18
SDS-PAGE analysis of medium screening
85
4.19
Effect of nutrients on rhTf yield
87
4.20
Amino Acids in Human Transferrin (679 residues)
89
4.21
Observed and predicted experimental data for the
92
optimization of glutamine, glucose and lipid mixtures
4.22
Glutamine (Gln) vs Glucose (Gluc) vs rhTf
95
4.23
Glutamine(Gln) vs Lipid Mixtures 1000x (Lip) vs rhTf
95
4.24
Glucose (Gluc) vs Lipid Mixtures 1000x (Lip) vs rhTf
96
4.25
Sf9 growth in controlled and optimized expression.
99
4.26
Total protein and rhTf contents in controlled and
99
optimized expression
4.27
Total protein and rhTf production rates in controlled
101
and optimized expression
4.28
Glucose and lactate concentrations in controlled and
101
optimized expression
4.29
Lactate production and glucose uptake rate in
103
controlled and optimized expression
4.30
SDS PAGE gel for non optimized medium
104
4.31
SDS PAGE gel for optimized medium
105
xviii
LIST OF SYMBOLS/ABBREVIATIONS/TERMINOLOGY
a
-
Constant
ABTS
-
2, 2’-azino-bid (3-ethylbenzthiazoline-6sulfonic acid)
AcMNPV
-
Autographa Californica Multiple Nuclear
Polyhidrosis Virus
apo-hTf
-
Low iron binding human trasferrin
Arg
-
Arginine
Asn-X-Thr/Ser
-
Asparagine-X-Threonine/Serine
ATCC
-
American Tissue Culture Collection
b
-
Constant
BEVS
-
Baculovirus Expression Vector System
b-Gal
-
Beta Galactosidase
BL1
-
Biosafety Level 1
BSA
-
Bovine Serum Albumin
BTI-Tn-5B1-4
-
Insect cell line
CDG
-
Carbohydrate-Deficient Glycoprotein
cm2
-
Surface area in centimeter square
CO2
-
Carbon dioxide
Cys
-
Cystine
DMSO
-
Dimethylsulfoxide
DNA
-
Diribonucleic Acid
doub
-
Doubling
dpi
-
Days post infection
E
-
Global error
E. coli
-
Escherichia Coli
xix
ELISA
-
Enzyme Linkage Immunosorbent Assay
ER
-
Endoplasmic Reticulum
exp
-
Exponential
FBS
-
Fetal Bovine Serum
FP
-
Few Polyhedra
Fruc
-
Fructose
g l−1
-
Gram per liter
g/cell
-
Gram per cell
g/ml
-
Gram per milliliter
GC
-
Gas Chromatography
Glc
-
Glucose
Glcp3-Manp9-GlcNAcp2
-
3(Glucose)-9(Manose)-2(NAcetylglucosamine)
Gln
-
Glutamine
Gluc
-
Glucose
H2O
-
Water
H2SO4
-
Sulphuric acid
HRP
-
Horseradish Peroxidase
Htf
-
Human Transferrin
Interc.
-
Intercept
k
-
Number of factors in experimental design
kbp
-
Kilo base pair
kDa
-
Kilo Dalton
KOH
-
Kalium Hidroxide
Lys
-
Lysine
M
-
Molar
Malt
-
Maltose
Man
-
Mannose
Man(alpha-1,6)
-
Mannose (alpha-1,6)
Manp8-GlcNAcp2
-
8(Mannose)-2(N-Acetylglucosamine)
max
-
Maximum
Met
-
Methionine
mg/ml
-
Milligram per milliliter
mM
-
Millimolar
xx
MOI
-
Multiplicity of Infection
MS
-
Mass Spectrometry
n
-
Number of possible combinations in
experimental design
NaCl
-
Sodium Chloride
NaOH
-
Sodium Hidroxide
ng
-
Nanogram
nm
-
Nanometer
NPV
-
Nuclear Polyhidrosis Virus
OD
-
Optical Density
OPD
-
O-phenylene diamine
OV
-
Occluded virus
p
-
Proportion of cultures receiving particular
number of infectious units.
p
-
Probability in analysis of variance
PCR
-
Polymerase Chain Reactor
PD
-
Proportionate Distance
PFU
-
Plug Performing Unit
pfu/cell
-
Plug performing unit per cell
pH
-
Potential hydrogen
r
-
Number of infectious units
R
-
Recombinant
RER
-
Rough Endoplasmic Reticulum
rhTf
-
Recombinant human trasferrin
SD
-
Seeding Density
SDS
-
Sodium Dodecyl Sulphate
SDS-PAGE
-
Sodium Dodecyl Sulphate Polyacrylamide Gel
Electophoresis
SEAP
-
Human Secreted Alkaline Phosphatase
Ser
-
Serine
Sf21
-
Insect cell line
Sf9
-
Insect cell line
SFM
-
Serum Free Medium
Std. Err.
-
Standard Error
xxi
t
-
Student’s test
TBS
-
Tris Buffered Saline
TCA
-
Tricarboxylic Acid
TCI
-
Time Course of Infection
TCID50
-
50 % Tissue Culture Infectious Dose
TEMED
-
Tetramethylethylenediamine
Thr
-
Threonine
TMB
-
Tetramethyl benzidine
TOI
-
Time of Infection
Tris
-
Tromethamine
Tris-HCL
-
Tromethamine and Hydrochloric Acid
Tyr
-
Tyrosine
U
-
Uninfected
unet
-
Net Growth Constant
USA
-
United State of America
UTM
-
Universiti Teknologi Malaysia
V
-
Voltage
Val
-
Valine
w/v
-
Weight per Volume
w/w
-
Volume per Volume
WI
-
Wild-Type
X
-
Cell concentrations at time t
X0
-
Cell concentrations at time 0
xi, yi
-
Data pairs
µ
-
Mean concentration of the infectious units
µl
-
Microliter
%
-
Percentage
0
-
Temperature level in degree Celcius
C
xxii
LIST OF APPENDICES
APPENDIX NO.
A1
TITLE
PAGE
Kinetics Analysis of Sf9 Insect Cells Growth at
130
Various Conditions
A2
Growth kinetic coefficients of Sf9 cells
133
monolayer culture
B1
TCID50 Calculation
134
B2
Calculation of End Point Dilution based on
135
Reed and Muench method
C1
RSM spreadsheet
136
D1
Flowchart of the major steps involved in this
138
research
CHAPTER 1
INTRODUCTION
1.1
Preface
The Internet created a sky rocketed investment that made history in this
modern era. It is a matter of time in some circles that biotechnology is the next
computer revolution that will change the world, spawn new industries and create
multi-millionaires.
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
2
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
heterologous proteins. In many cases the products are chemically, antigenically,
immunologically and functionally similar, if not identical to their authentic
counterparts (Vlak, 1997).
The baculovirus expression vector system (BEVS) is frequently a method of
choice for the expression of recombinant mammalian proteins (O’Reilly et al., 1994).
Apart from the simplicity and cost-effectiveness of this method, the insect host cells
possess many of the protein-processing and -folding mechanisms of mammalian cells
(O’Reilly et al., 1992) therefore functional and antigenic differences are rarely seen.
The technology called the BEVS for the safe, abundant and rapid production of
recombinant proteins in insect cells and insects was pioneered in the laboratory of
Dr. Max D. Summers of Texas A&M University USA in 1982.
The BEVS has become a core technology for the cloning and expression of
genes for study of protein structure, processing and function. It is also important for
the production of biochemical reagents and study of regulation of gene expression. 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.
3
Studies of proteins for the development of drug therapies, vaccines, and
insights into biological function depend upon the ability to produce large amounts of
structurally complex proteins. It is important that these proteins are biologically
active, processed correctly, assume a native shape, and locate to the proper place in
the cell. The inability to generate large quantities of structurally complex eukaryotic
proteins with these characteristics has been a major limitation for many years. Thus,
this thesis hoped to give necessary foundations on how to develop a process that will
produce greater amount of recombinant proteins for therapeutic purpose.
1.2
Research Problem Background
The development of new recombinant therapeutic proteins requires extensive
studies on the expressional host and product. In this research, human transferrin, a
model protein was chosen as the expressional product and insect cell as the
expressional host. The selection was based on many reports from other researchers
which indicate that insect cell baculovirus system is a promising new artificial
system for the production of large amount of recombinant proteins. Insect cells
Spodoptera Frugiperda (Sf9) and recombinant Autographa Californica Multiple
Nuclear Polyhydrosis Virus (rAcMNPV) were utilized in this research. Human
transferrin was chosen because it is a simple form of glycoprotein which is easier to
study than the complex and hybrid forms.
4
To complement metabolic engineering works involving the humanization of
recombinant glycoprotein, it is important that the recombinant protein can be
generated in large quantities. The understanding of the insect cells and baculovirus
behaviour is as critical as the expressional behaviour of the recombinant protein at
various settings. Various yields of recombinant human transferrin (rhTf) using the
baculovirus system have been reported. Among those were Tomiya et al., (2003)
who reported rhTf yield of 7µg/ml and Ali et al., (1996) with 20 µg/ml of rhTf.
1.3
Research Objective
The ultimate objective of this research was to optimize the expression level of
recombinant human transferrin in insect cells baculovirus expression system in terms
of its concentration (µg/ml) and protein percentage.
1.4
Research Scopes
This research focused on the optimization of expression of recombinant
human transferrin gene which had already been cloned into the baculovirus DNA.
The scopes of this research were as follows:
a) Expression and optimization of rhTf in Sf9 insect cells monolayer culture
using conventional method. Variables studied were seeding density (SD),
multiplicity of infection (MOI), time of infection (TOI), and harvest time
(HT).
b) Screening of the Sf900-II SFM insect cell culture medium and recombinant
baculovirus stock that resulted in improved production of rhTf.
5
c) Expression and optimization of rhTf in Sf9 insect cells suspension culture
using experimental design. Variables studied were dominant medium
components that were screened earlier.
1.5
Research Contributions
Some major contributions of this research are listed below.
1) Establishment of methods for optimizing recombinant protein expression in
insect cells culture.
2) Trained (hands-on-experienced) personnel in Baculovirus Insect Cells
Expression System.
3) Two research papers (proceedings) were published (Ongkudon et al., 2004;
Ongkudon et al., 2005) and two other papers are in preparation.
CHAPTER 2
LITERATURE REVIEW
2.1
Insect Cells - Baculovirus Expression System
2.1.1
Baculovirus Characteristics
Baculovirus has a large, double stranded, covalently closed, circular DNA
genome of between 88 and 200 kbp (Arif, 1986). The length of the nucleocapsid may
be 200-400 nm and the width remains constant at about 36 nm (Fraser, 1986). The
baculoviruses utilized in this research come from the group nuclear polyhydrosis
virus (NPV). The specific baculovirus is Autographa californica (AcMNPV) which is
obtained from the alfalfa looper in the form of proteinacious nuclear occlusion
bodies (Vail et al., 1971). The virus infects several orders of insects primarily the
Lepidopteran species. It can replicate in over 30 Lepidopteran species. AcMNPV is
biphasic and differs from other DNA of animal viruses.
Wild type baculoviruses exibit both lytic and occluded life cycles that
develop independently throughout the three phases of virus replication. The early
phase occurs 0.5 to 6 hours after infection. The late phase, where extracellular
7
viruses are released, occurs 18 to 36 hours after infection. In the very late phase the
polyhedrin and p10 genes are expressed. Between 24 and 96 hours after infection, the
cells start to produce occluded virus (OV). A polyhedrin protein with a molecular
weight of 29 kDa is the major structural component of the viral occlusions (Summers
and Smith, 1978). In infected Spodoptera frugiperda cell cultures, polyhedrin
accumulates to very high levels, routinely 1mg/ml per 1.0-2.0 x 106 infected cells
accounting for 50-75% of total “stainable” protein of the cell detected on SDS-PAGE
(Summers and Smith, 1988).
After prolonged passage through multiple infection cycles, the ability of the
virus to infect insect cells diminishes (Tramper et al., 1990). Therefore, mutant virus
needs wild type virus to multiply. It is however; quite rare for recombinant
baculoviruses to undergo mutations which affect recombinant protein expression
(Possee et al., 1990).The accumulation of genotype changes may be reduced by
exercising good practice in virus propagation.
Although these viruses may enter other cells types (perhaps by phagocytosis),
they are not infectious in them. For example, nucleocapsid proteins are not removed
in most human cells. In human hepatic cell lines that do remove these proteins, the
virus fails to replicate and express proteins due to the absence of insect transcription
factors. Thus, working with baculoviruses is considered safe for humans and
contamination of mammalian cell lines in shared biosafety hoods is not a problem.
Recombinant DNA guidelines recommend a BL1 biosafety level for most
baculovirus expression experiments (Frank, 1998).
8
2.1.2 Insect Cell Lines
There are few common cells lines that have been used in BEVS such as Sf9,
BTI-Tn-5B1-4, Sf21, and High Five. In this research, Sf9 which came from
Spodoptera frugiperda (fall army worm) pupal ovarian tissue (Vaughn et al., 1977)
was utilized. Sf9 can be frozen in serum containing medium and can be thawed
directly to serum free medium. Sf9 cultured in a complete growth medium will have
a doubling time of 18-24 hour and can be passed after 4-5 days. Adaptation is
complete after 5 passages. Figure 2.1 shows the general appearance of Sf9 cells.
Figure 2.1: Spodoptera frugiperda Sf9 Cells
Although there is significant scientific data on the characteristics of Sf9 cell
line, it remains to be confirmed whether it is the best line for virus or recombinant
protein production. Ongoing research suggests that different insect cell line may
support varying levels of expression and differential glycosylation with the same
recombinant protein, (Hink et al., 1991). The Sf9 cells used in this research were
obtained from American Tissue Culture Collection (ATCC).
9
2.1.3
The Pros and Cons of the Insect Cells – Baculovirus Expression System
The insect cells baculovirus system has been known to be safe to be handled
under normal laboratory settings. It can be cultured in serum free medium which
eliminates the need of the costly fetal bovine serum. The late expression of
polyhedrin is controlled by strong promoter elements which were not essential for
viral replication. Therefore, it can easily be replaced by new desired genes to produce
a commercial product in a significant amount. The rod shaped capsid of the
baculovirus can expand and accommodate large DNA inserts and is also easily
modified. Plaque purified recombinant virus can be obtained in 4-6 weeks (rapid
scale up). The baculovirus is host specific. It can propagate only in Lepidoptera cell
lines and is not known to infect animal or plant cells (Volkman and Knudson, 1986;
Grioner, 1986). It is believed that it requires strong promoters in the insect
hemolymph to multiply. Besides that, the viruses cannot survive in the environment
without their polyhedrin coats. Insect viruses have particularly strong promoters
meaning that the protein for which they encode is produced in high levels, up to 250
times that of mammalian cultures (Luckow, 1991).
As with other eukaryotic expression systems, baculovirus expression of
heterologous
genes
permits
folding,
post-translational
modification
and
oligomerization in manners that are often identical to those that occur in mammalian
cells (Miller, 1988; Miyajima et al., 1987). The insect cytoplasmic environment
allows proper folding and S-S bond formation, unlike the reducing environment of
the E. coli cytoplasm. Proteins may be secreted from cells or targeted to different
subcellular locations. Despite these potential advantages, particular patterns of posttranslational processing and expression must be empirically determined for each
construct. Differences in proteins expressed by mammalian and baculovirus infected
insect cells have been described and overcome in some cases. For example,
inefficient secretion from insect cells may be circumvented by the addition of insect
secretion signals (ex. honeybee melittin sequence). Improperly folded proteins and
proteins that occur as intracellular aggregates may be due to expression late in the
infection cycle. In such cases, harvesting cells at earlier times after infection may
10
help. Low levels of expression can often be increased with optimization of time of
expression and multiplicity of infection. Potential N-linked glycosylation sites are
often either fully glycosylated or not glycosylated at all, as opposed to expression of
various glycoforms that may occur in mammalian cells (Ailor et al., 2000; Goochee
et al., 1991). Metabolic engineering works on the glycosylation of non-glycosylated
recombinant protein expressed by insect cells are on going (Ailor et al., 2000;
Tomiya et al., 2003; Tomiya et al., 2004).
Transfer
vector
Viral DNA
Baculovirus
EcoR1
Promoter
Foreign gene
EcoR1
Cloning site
Polyhedrin gene
Transfer
vector
Transfer
vector
Homologous
recombination
Polyhedrin promoter
Transfection
Insect cell
Viral DNA
Infection
Recombinant
viral DNA
Insect cell
Transcription
Insect cell culture
Protein
Figure 2.2: Insect Cells - Baculovirus Expression System.
2.1.4 Generation of Recombinant Baculovirus
Baculoviruses infect invertebrates, including insects. During infection, two
forms of baculoviruses are formed in which one of them is surrounded by a
polyhedrin protein matrix. Upon the last stages of infection cycle, polyhedrin protein
11
is synthesized in a massive amount. It was found that replacement of polyhedrin
gene with a heterologous protein gene, followed by infection of cultured insect cells
with genetically modified baculovirus, would result in the vast production of
heterologous protein. The recombinant heterologous protein is expected to undergo
similar posttranslational modifications as in the genetically modified insects cells.
The first step to constructing the recombinant protein is to produce
recombinant baculovirus AcMNPV transfer vector (Figure 2.2). Transfer vector is an
E. coli-based plasmid which carries a segment of the AcMNPV virus DNA. Next,
insect cells in culture that have been transfected with AcMNPV DNA are transfected
with a transfer vector carrying a cloned gene. Within some of the doubly transfected
cells, a double crossover event occurs and the cloned gene becomes integrated into
AcMPNV DNA. DNA hybridization or a polymerase chain reactor (PCR) assay can
be used to detect recombinant baculovirus. Heterologous protein can be harvested
after 4 to 5 days. In this research, a recombinant AcMNPV carrying human
transferrin gene was obtained from Johns Hopkins University, USA.
2.2
Model Glycoprotein
2.2.1
Native Human Transferrin (nhTf)
Human serum transferrin (HST) belongs to the transferrin family of metalbinding 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
12
intragenic duplication that led to two homologous domains, each of which binds 1
ion of ferric iron (Figure 2.3) with both sites of glycosylation in the carboxylterminal 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
are identical when aligned by inserting gaps at appropriate positions (MacGillivray et
al., 1982).
Figure 2.3: 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
13
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 follows.
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
KASYLDCIRAIAANEADAVTLDAGLVYDAYLAPNNLKPVVAEFY
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
SAGWNIPIGLLYCDLPEPRKPLEKAVANFFSGSCAPCADGTDFP
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
STIFENLANKADRDQYELLCLDNTRKPVDEYKDCHLAQV PSHT
VVARSMGGKEDLIWEL LNQAQEHF G KDKSKEFQLFSSPHGK
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
PEAPTDECKPVKWCALSHHERLKCDEW SVNSVGKIEC VSAET
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
KSDNCEDTPEAGYFAV AVVKKSASDLTWDNLKGKKSCHTAVG
RTAGWNIPMGL LYNKINHCRFDEFFSEGCAPGSKKDSSLCKLC
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
VPQNTGGKNPDPWAKNLNEKDYELLCLDGTRKPVEEYANCHLA
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
LFRSETKDLLFRDDTVCLAKLHDRNTYEKYLGEEYVKAVGNLR
K C S T S S L L E A C T F R R P -679 (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)
syndrome and the use of genetically-determined polymorphisms for forensic
purposes.
14
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 Nglycan pattern observed in the host (Montreuil et al., 1997).
2.2.2
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 Nacetylglucosamine. Neither sialic acid nor galactose was detected on any of the N-
15
glycans. Carbohydrate analysis revealed a small fraction of Gal in oligosaccharides
obtained from N-glycans of human lactoferin was expressed in Spodoptera
frugiperda (Sf9) (Wolff et al., 1996).
2.2.3
Biosynthesis of N-Glycans
Biosynthesis of N-Glycans is independent of the protein that will be
glycosylated in a later step. The biosynthesis of the oligosaccharide precursor
molecule starts at the cytosolic site of the rough ER (RER). In general, the degree of
glycosylation is dependent on the number of available Asn-X-Thr/Ser sites in the
protein and their conformational accessibility, available amount of completely
glucosylated precursors, and oligosaccharyltransferase activity.
After the precursor molecule Glcp3-Manp9-GlcNAcp2 has been assembled ,
its transfer to Asn by oligosaccharyltransferase (inhibitors are tunicamycin and
amphomycin) is followed by further processing by two membrane-bound
glucosidases which remove the terminal Glc residues. First, the terminal Glc is
removed by the action of alpha-1,2-glucosidase I. Then, the remaining two residues
are cleaved by alpha-1,3-glucosidase II.
After the Glc residues have been removed, one Man residue is cleaved from
the Man(alpha-1,6) branch by ER alpha-1,2-mannosidase. The glycoprotein with the
linked Manp8-GlcNAcp2 (high mannose type) side chain is then transported from
the RER to the cis cisternae of the Golgi stacks. This transport is thought to be
receptor mediated with the receptor recognizing the deglucosylated oligosaccharide
chain and can be inhibited with Brefeldin A. In the Golgi stacks further processing
occurs depending on the final destination of the glycoprotein, which is solely
determined by its amino acid sequence (Frosch, 1997).
16
2.3
Insect Cell Culture Medium
2.3.1
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.
17
2.3.2
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).
2.3.3
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 (β-Gal and SEAP) with increasing infection densities (Taticek and Shuler,
18
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).
2.3.4
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 µg/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 α-tocopherol acetate
and cholesterol.
19
2.3.5
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.
2.3.6
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 protein
20
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)
2.4
Optimization of Protein Expression in BEVS
2.4.1
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%
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.
21
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)
Type of Vessel
Cell Density
Minimum Virus
Incubate in Final
Volume
Volume
96 well plate
2.0 x 104/well
10 µl
100 µl
24 well plate
6.0 x 105/well
200 µl
500 µl
60 mm2 dish
2.5 x 106/dish
1 ml
3 ml
25 cm2 flask
3.0 x 106/flask
1 ml
5 ml
75 cm2 flask
9.0 x 106/flask
2 ml
10 ml
150 cm2 flask
1.8 x 107/flask
4 ml
20 ml
(based on MOI)
(based on size)
spinners (all)
6
1.5-2.0 x 10 /ml
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
experiment to determine the optimum time for harvesting recombinant proteins; and
not to rely on data published by others.
2.4.2
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
22
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 = µre-µ/r!
…2.1
where µ is the mean concentration of the infectious units in the diluted solution.
Therefore the proportion of culture receiving no infectious units is
p = e-µ
(r = 0)
… 2.2
and the proportion of culture receiving one or more infectious units is
p = 1-e-µ
(r >= 1)
…2.3
The proportion of culture receiving only one infectious unit is
p = µe-µ
(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) = µe-µ/(1-e-µ)
…2.5
If we want to be 95% confident that the infected cultures contains only a single
infectious unit, then
µe-µ/(1-e-µ) = 0.95
…2.6
Solving this will reveal that µ=0.101. Therefore the proportion of uninfected cultures
e-µ=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
23
is to dilute the virus such that, if multiple cultures are exposed to the diluted
inoculum, any cultures that become infected will have received only a single
infectious unit (Reed and Muench, 1938).
2.4.3
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).
24
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
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
25
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).
2.5
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
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.
2.5.1
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.
26
2.5.2
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.
2.5.3
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
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, x2, and x3 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
27
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, x2, and x3. 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).
2.5.4 Factor Levels Combinations
A straightforward way to list all unique combinations of a 2 level factorial
design is as follows (Montgomery, 1996; Montgomery, 2001);
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.
28
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
Run no
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2.5.5
A
-1
-1
-1
-1
-1
-1
-1
-1
1
1
1
1
1
1
1
1
Factors
B
-1
-1
-1
-1
1
1
1
1
-1
-1
-1
-1
1
1
1
1
C
-1
-1
1
1
-1
-1
1
1
-1
-1
1
1
-1
-1
1
1
D
-1
1
-1
1
-1
1
-1
1
-1
1
-1
1
-1
1
-1
1
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).
29
2.5.6
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)
Run No.
1
2
3
4
5
6
7
8
9
10
11
12
Factor No.
1 2 3
+ + + +
+ +
+ + +
+ + + +
+ + +
+ +
-
4
+
+
+
+
+
+
-
5
+
+
+
+
+
+
-
6
+
+
+
+
+
+
-
7
+
+
+
+
+
+
-
8
+
+
+
+
+
+
-
9
+
+
+
+
+
+
-
10
+
+
+
+
+
+
-
11
+
+
+
+
+
+
-
30
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
employed to study the interaction between factors and improve the quality of a
process without having to do so much of trials and errors.
2.6
Analysis of Experiments
2.6.1
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
31
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).
2.6.2 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, yi), error, ei, is defined as the difference between the predicted value and the
actual measured value.
ei = error at data pair i = Yi - yi = axi + b - yi.
...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
32
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
2.6.3 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
33
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).
2.7
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
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.
34
2.7.1
Improvements of RSM
Further improvement is only possible by the following techniques:
¾ 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.
¾ The data is replicated. This helps to cancel the effect of random noise
(experimental error).
¾ A higher-order regression scheme is used. Note that here only a linear (firstorder) 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:
¾ 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).
¾ Overall, RSM is a very powerful technique for optimizing a response.
2.8
Specialized Protein Analysis: Theories and Principles
2.8.1
Sodium Dodecyl Polyacrylamide Gel Electrophoresis (SDS-PAGE)
Gel electrophoresis is a separation method based on the movement of charged
molecules within a fluid medium under the influence of an electric field
(Westermeier, 1993). Separation of large molecules depends upon two forces which
35
are the charge and mass. When a biological sample, such as proteins or nucleic acid,
is mixed in a buffer solution and applied to a gel, these two forces act together. The
electrical current from one electrode repels the molecules while the other electrode
simultaneously attracts the molecules (Figure 2.4). The gel matrix acts as a
"molecular sieve," separating the molecules by size. During electrophoresis,
macromolecules are forced to move through the pores when the electrical current is
applied. After staining, with suitable solution, the separated macromolecules in each
lane can be seen as a series of bands spread from one end of the gel to the other
(Rosenberg, 1996).
The rate of migration of the macromolecules depends on the strength of the
field; net charge, size and shape of the molecules and also on the ionic strength,
viscosity and temperature of the medium in which the molecules are moving.
electrical current from one
electrode repels the
molecules while the other
electrode attracts the
molecules.
-
+
Figure 2.4: Samples being electrophorase in SDS-PAGE
36
Figure 2.5: Chemical structure of a polyacrylamide
Mixture of
Macromolecules
Electrophoresis
Porous Gel
Figure 2.6: Movement of molecules in the porous gel
The gel matrix is usually made of paper, cellulose acetate, starch gel, agarose
or polyacrylamide gel. Polyacrylamide (Figure 2.5), which is easy to handle and to
make at higher concentrations, is used to separate most proteins and small
oligonucleotides that require a small gel pore size for retardation. A porous gel can
37
act as a sieve by retarding, or in some cases completely obstructing, the movement of
large macromolecules while allowing smaller molecules to migrate freely (Figure
2.6).
Utilization of SDS-PAGE:
1. Density estimation
2. Protein separation
3. Quality check (combined technique with western blotting)
4. Molecular weight determination
5. Monitoring protein purification
Proteins are amphoteric compounds which means their net charge is
determined by the pH of the medium in which they are suspended. In a solution with
a pH above its isoelectric point, a protein has a net negative charge and migrates
towards the anode in an electrical field and vice versa. The net charge carried by a
protein is independent of its size. At a given pH and under non-denaturing
conditions, the electrophoretic separation of proteins is determined by both size and
charge of the molecules (Boyer, 1993).
Sodium dodecyl sulphate (SDS) is an anionic detergent which denatures
proteins by "wrapping around" the polypeptide backbone. Since the 3D structure of
most proteins depends on interactions between hydrophobic amino acids in their
core, the detergent destroys 3D structures, transforming them into linear molecules
now coated with negative SDS groups. After boiling in SDS, proteins therefore
become elongated with negative charges arrayed down them in proportion to their
length, so they will move towards a positive electrode (Westermeier, 1993). It is
usually necessary to reduce disulphide bridges in proteins before they adopt the
random-coil configuration necessary for separation by size: this is done with 2mercaptoethanol or dithiothreitol. In denaturing SDS-PAGE separations therefore,
migration is determined not by intrinsic electrical charge of the polypeptide, but
merely by size and molecular weight (Figure 2.6).
38
Nucleic acids however, remain negative at any pH used for electrophoresis
and in addition carry a fixed negative charge per unit length of molecule, provided by
the PO4 group of each nucleotide of the nucleic acid. Electrophoretic separation of
nucleic acids therefore is strictly according to size.
A linear relationship exists between the logarithm of the molecular weight of
an SDS-denatured polypeptide, or native nucleic acid, and its Rf. The Rf is
calculated as the ratio of the distance migrated by the molecule to that migrated by a
marker dye-front. A simple way of determining relative molecular weight by
electrophoresis (Mr) is to plot a standard curve of distance migrated vs. log10MW for
known samples, and read off the logMr of the sample after measuring distance
migrated on the same gel (Bollag et al., 1996).
SDS-PAGEs are sensitive because it can separate 2% difference in mass and
requires only very little proteins (2 pmol). However it also faces some disadvantages
due to its less effectiveness in separating proteins of similar molecular weights. Since
SDS denatures proteins, there is no enzyme activity and other samples need to be
prepared for activity test (McGregor, 1991).
2.8.2
Enzyme Linked Immunosorbent Assay (ELISA)
2.8.2.1 Basic Immunology
Knowledge of immunology especially the properties of certain components of
the immune system are essential in developing ELISAs. Immunoglobulins are
globulins which function as antibody. Generally, there are four intermolecular forces
that are involved in antibody-antigen reaction: hydrogen bonds, ionic bonds, Vander
39
Waals bonds and hydrophobic bonds (Steward, 1984). Antibody detection methods
vary from precipitation which is less sensitive to ELISA which is most sensitive.
Antibody-antigen (Ab-Ag) reaction can be simplified by the following equation
Ab + Ag
ka
AbAg
…2.19
kb
where (AbAg) = bound antibody concentration, (Ab)= free antibody concentration,
(Ag) = free hapten concentration. Affinity is a thermodynamic measurement of the
strength of the Ab-Ag interaction and expressed as the equilibrium constant K
(L/mole) or ∆Go ( kilo calorie/ mole).
ka
( AbAg )
=K=
kd
( Ab )( Ag )
∆G o = − RT ln K
…2.20
…2.21
where K = equilibrium constant. From the law of mass action the following form of
Langmuir adsorption isotherm may be derived
( AbAg )
nK ( Ag )
=r=
( Ab)
1 + K ( Ag )
r
= nK − rK
( Ag )
…2.22
…2.23
where r = moles haptens bound per mole of antigen present, n = antibody valence.
Thus a plot of r/(Ag) versus r (Scatchard plot) for a range of antigen concentration
allows n, the antibody valence and K to be obtained.
40
Important factors affecting antibody affinity are immunogenic stimulus;
genetic factors; lymphocyte functions; diet and hormone; reticuloendothelial
functions; and effects of free antibody and antigen-antibody complexes (Crowther,
1995). Antibody forming cells have immunoglobulins receptors on its surface with
specificity for antigen. High affinity receptors produce high affinity antibody and
antigen selects cells with high receptor affinity thus stimulates them to produce
antibody.
From immunopathological perspective, excessive production of low affinity
antibody has been considered as an expression of immunodeficiency. Antigenantibody diseases arise from genetically controlled low affinity antibody response
which fails to eliminate antigen (Vander et al., 1994).
2.8.2.2 Principles of ELISA
ELISA by definition exploits the principles of immunology in which an
antigen is attached to an antibody and an enzyme can subsequently be attached to the
antibody (Figure 2.7). The antibody can be a monoclonal or polyclonal antibody. The
fact that proteins (enzymes and antibody) can be attached to plastics surface has been
exploited in the application of ELISA (Steward, 1984). Since one of the components
is attached to a solid phase by passive adsorption, subsequent reagents can be added,
and after a period of incubation, unreacted chemicals can simply be washed away.
Subsequent addition of relevant enzyme substrates causes a colour intensity change
which can be measured and quantified using a spectrophotometer (Ongkudon et al.,
2004). Standard curves which contain known concentrations of antigen or antibody
can be plotted and used to calibrate concentrations of test samples. ELISA also
provides a highly sensitive and precise method for estimation of biological
parameters.
41
Antibody
Blocking Reagant
Antigen
Non-developed secondary
Developed secondary
Step 1:
Coat with
antibody
Step 2:
Step 3:
Block nonAdd Sample
Specific
Binding sites
Step 4:
Add
conjugated
Secondary
Step 5:
Add
Substrate
Figure 2.7: Basic principles of Direct Sandwich ELISA
The use of passive adsorption allows a great deal of flexibility in designing an
assay. ELISA in general can be categorized under four groups: direct, indirect,
sandwich and competition. The first three will be discussed here.
In direct-labeled-antibody ELISA, an antigen attached to a solid phase is
reacted directly with enzyme-labeled antibody. This has the disadvantage that
antibody rose against different antigens all have to be labeled. In direct-labeledantigen, the antibodies are adsorbed to the solid phase, and then enzyme-labeledantigens are added and reacted with antibodies of suitable specificity. However,
antigens are rarely labeled and this assay has been used in the estimation of hormone
concentrations and is analogous to many radioimmunoassay methods (Steward,
1984).
In indirect ELISA, antibodies from a particular species react with antigen
attached to a solid phase. Any bound antibodies are detected by the addition of an
antispecies antibody labeled with enzyme. This type of assay is widely used in
42
problem diagnosis. This is extensively used for the detection and titration of specific
antibodies from serum samples. The specificity of the assay is directed by the antigen
on the solid phase which may be highly purified and characterized or vise versa.
Such assay offer an advantages over the direct ELISA since only a single antispecies
enzyme conjugate is needed to titrate antibody from many animals of a single species
(Crowther, 1995).
In direct-sandwich ELISA, the antibodies are attached to the solid phase and
allowed to capture antigen. This is then detected using an enzyme-labeled antibody
specific for the antigen (Figure 2.7). The capture antibody and the detecting antibody
can be from the same or different sources. However, the antigen must have at least
two different antigenic sites, one for capture antibody and the other one for detecting
antibody (MacGillivray et al., 1983). In indirect-sandwich ELISA, the detecting
antibody is from a different species than the capture antibody. The antispecies
enzyme-labeled antibody binds to the detecting antibody specifically and not to the
capture antibody. The advantage is that many second antibodies may be titrated with
a single conjugate (Crowther, 1995). Before making any decision on which ELISA
techniques to be used, some preliminary investigations need to be carried out such
as:
1. Purpose of the assay,
2. Information on the product sample,
3. Reagents availability,
4. Information on the reagents,
5. Whether a kit is required,
6. Whether the assay will be used in other laboratories,
7. Whether the test is for research purpose or applied use,
8. Feasibility-proof that a test can work,
9. Validation-showing that a test is stable and is evaluated over time under
different conditions, and
10. Standardization- quality control, establishment that the test is precise and can
be used by different workers in different laboratories.
CHAPTER 3
RESEARCH METHODOLOGY
3.1
Materials
3.1.1
Cell line and Recombinant Baculovirus
Insect cell line Spodoptera frugiperda (Sf-9) was purchased from ATCC cat.
No. 1711 (Rockville, MD). The recombinant baculovirus carrying the gene of
Human Transferrin was provided by Prof. Dr. Michael J. Betenbaugh of Johns
Hopkins University, USA. The wild type baculovirus AcMNPV was a gift from Prof.
Dr. Mohd. Sanusi Jangi, Universiti Kebangsaan Malaysia, Malaysia.
3.1.2
Equipments
The electrophoresis unit used was Mini-Protean II from Bio-Rad (California,
USA). Shimadzu UV-160 spectrophotometer (Minnesota, USA) was used to measure
44
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 was
used to analyze glucose and lactate contents.
3.1.3
Chemicals
Sf-900 II Serum Free Medium (SFM) and Fetal Bovine Serum (FBS) were
from GIBCO BRL (Gaithersburg, MD). Goat anti-Human Transferrin-affinity
purified, Goat anti-Human Transferrin-HRP conjugate, Human Reference Serum,
TMB (3,3’,5,5’-tetramethylbenzidene), and Peroxidase Solution B were purchased
from Bethyl Laboratories Inc. (Texas, USA). Acrylamide, bis-acrylamide, bovine
serum albumin (BSA), ammonium persulfate, 2-mercaptoethanol, dimethyl
sulphoxide (DMSO), N,N,N’,N’-tetramethylethylenediamine (TEMED), tris, and
glycine were purchased from Sigma (Missouri, USA). Trypan Blue, ethanol, acetic
acid, sodium hydroxide, hydrochloric acid, sodium dodecyl sulphate (SDS),
bromophenol blue, sodium bicarbonate, glycerol and methanol were obtained from
Fluka (Missouri, USA).
45
3.2
Insect Cells Techniques
3.2.1
The Preparation of TC100 Medium From Powdered Formulation
Initially, all glasswares were sterilized. The medium composition was TC100, 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
composed medium was filter-sterilized through a 0.22 micron filter. TC100 solution
stored at 40C has a shelf live of at least 1 year while TC100 + FBS solution stored at
40C has shelf live of at least 4 weeks.
3.2.2
Maintenance and Regeneration of Sf9 Cells Monolayer Culture
Medium and floating cells from 80% to 90% confluent monolayer were
aspirated and discarded. To each 25-cm2 flask, 5 to 10 ml of complete growth
medium equilibrated to room temperature was added. For 75-cm2 flasks, 15 ml was
added per flask. Cells were resuspended by pipetting the medium across the
monolayer with a Pasteur pipette or by shaking the T-flask vigorously. The cell
monolayer was observed using an inverted microscope to ensure adequate cell
detachment from the surface of the flask. The viable cell count of harvested cells was
determined using a haemocytometer and trypan blue dye exclusion. 5 x 105 viable
cells/ml were inoculated into 25- or 75-cm2 flasks. The cell cultures were incubated
at 27°C ± 0.5°C with loose caps to allow for gas exchange. The flasks were
subcultured when the monolayer reaches 80% to 100% confluency, approximately 2
to 4 days postplanting. The length of time needed to reach confluency before
subculturing often depended on the cell innoculums concentration.
46
To ensure adequate oxygenation, minimal medium depth and loose caps were
maintained. When the cell line was growing slowly, spent medium was replaced with
fresh medium at day 3 or 4 post-planting. The cells were subcultured when the
culture reached 80% to 100% confluency.
3.2.3
Cells Freezing
Only cells with exponential growth and more than 90% viability should be
used for preservation in liquid nitrogen. Cells were centrifuged at 1,000 g, 5 min.
The supernatant was discarded and the cells were resuspended in fresh medium at a
concentration of 1-2 X 107 cells/ml. Equal volumes of resuspended cells and freezing
medium (85% media, 15% DMSO) were mixed and placed on ice. 1 ml volume of
the cell suspension was pipetted into cryovials. The cryovials were placed into a
Nalgene Cryo 1 C Freezing Container (an isopropanol bath) and placed directly at 70°C overnight. The cryovials were then transferred into a liquid nitrogen storage
tank.
3.2.4
Cells Thawing
Cryovials were taken out from liquid nitrogen and thawed rapidly (in about 1
minute) at 37°C. The outer surface of the vials was wiped with 70% ethanol and the
cells were transferred to a T25 flask containing 4 ml complete medium. After 24
hours, the cells have adhered to the plastic. The medium was removed and the
adherent cells were fed with fresh medium to dilute the DMSO. If the cells did not
adhere, they were probably dead. In this case, another vial of cells was taken out
47
from the liquid nitrogen and thawed. After 24 hours, the cells were fed with fresh
medium. For cell maintenance, protocols for subculturing cells were followed.
3.2.5
Cells Counting
Adherent and semi adherent cells were brought into suspension. Under sterile
conditions 100-200 µl of cell suspension were removed. Equal volume of Trypan
Blue (dilution factor =2) was added and mixed by gentle pipetting. Haemocytometer
was cleaned. The coverslip was moistened with water or exhaled breath. The coverslip was then slid over the chamber back and forth using slight pressure until
Newton’s refraction rings appear (Newton’s refraction rings are seen as rainbow-like
rings under the cover-slip). Both sides of the chamber were filled with cell
suspension (approx. 5-10 µl) and viewed under a light microscope using x20
magnification.
The number of viable (seen as bright cells) and non-viable cells (stained blue)
were counted. Ideally more than 100 cells were counted in order to increase the
accuracy of the cell count (eq. 3.1). The number of squares counted to obtain the
count of more than 100 was recorded. The concentration of viable and non-viable
cells and the percentage of viable cells were calculated using the equations below.
Viable cells concentration (cells/ml)
= (total non stained cells within 8 major squares)
… 3.1
x (dilution factor) x 10000 / 8
Percentage Viability
= 100 x no. of viable cells / total no. of cells
… 3.2
48
3.2.6
Adaptation of Sf9 Cell Culture in Serum Free Medium
The procedure took about 7 days and involved gradually decreasing the
percentage of serum containing medium in the cell culture. Changing the medium
over to serum-free without adaptation could lead to greatly decreased doubling times
and cell mortality. First, adherent log phase cells that were 50% confluent were
obtained. 25% of the serum-free medium (Sf-900 II SFM) was added to 75% of the
serum-containing medium (TC100 + 10% FBS). Sf9 cells were grown to confluency.
Sf9 cells were split to 1:1 and subcultured in 50% serum-containing medium and
50% serum-free medium. Cells were let to grow to confluency. Again, the cells were
split to 1:1 and subcultured in 25% serum-containing medium and 75% serum-free
medium. Cells were again grown to confluency. Finally, the cells were subcultured to
desired density/dilution using 100% serum-free medium.
During the adaptation process it is normal for the growth rate of the cells to
slow down. Keeping the cells at no lower than 50% confluency helped to keep them
in log phase and minimized time loss due to slow growth rates.
3.2.7
Adaptation of Sf9 Cells in Suspension Culture
Because insect cells are not generally anchorage dependent, they adapt easily
to suspension culture conditions. It is however, important to proceed slowly when
adapting stationary cultures to suspension culture. A drop in viability and increased
clumping through the first three to five passages were normal. Protocol 3.2.7 has
adapted the Sf9 cell lines to suspension culture and reduced cell clumping over a
short period of time. Two to three confluent 75-cm2 monolayer flasks were sufficient
to initiate a 50-ml suspension culture.
49
Cells were dislodged from the bottom of the T-flasks. Viable cells count was
carried out. The cells suspension was diluted to approximately 5 x 105 viable cells/ml
in complete serum-supplemented or serum-free growth medium equilibrated to room
temperature. Cells were incubated at 27.0°C ± 0.5°C with a stirring rate of 100 rpm
for shake flasks or a stirring rate of 75 rpm for spinner cultures. The cells were
subcultured when the viable cell count reached 1 x 106 to 2 x 106 cells/ml (3 to 7
days post-planting). The stirring speed was increased by 5 to 10 rpm with each
subsequent passage. When cell viabilities dropped below 75%, stirring speed was
decreased by 5 rpm for one passage until culture viability recovered and was >80%.
For shake flask cultures, adaptation was completed when the constant stirring speed
reached 120 to 130 rpm. For spinner cultures, adaptation was completed when the
constant stirring speed reached 90 to 100 rpm.
3.2.8
Sf9 Cells Maintenance in Suspension Culture
The orbital shaker or stirring platform were maintained in a 27°C ± 0.5°C,
nonhumidified, non-CO2 equilibrated, ambient-air regulated incubator or warm
room. For cultures already adapted to and maintained in suspension culture, orbital
shaker was set at 135 to 150 rpm and spinner platforms at 90 to 100 rpm.
100µl sample from a 3- to 4-day-old suspension culture (in mid- exponential
growth) were removed and the viable cell concentration was determined. Cells were
diluted to 5 x 105 viable cells/ml in complete serum-free or serum-supplemented
growth medium equilibrated to room temperature. For shake flasks, stock cultures
were maintained as a 50- to100-ml culture in 250-ml Erlenmeyer flasks. For spinner
flasks, stock cultures were maintained as 150 to 175 ml cultures in 250-ml spinner
flasks. To aerate the cultures, the caps were loosen about ¼ to ½ of a turn. The
cultures were incubated until they reached 2 x 106 to 3 x 106 viable cells/ml. To
maintain consistent and optimal cell growth, suspension cultures were subcultured
50
twice weekly. Once every 3 weeks, the cell suspension was gently centrifuged at 100
x g for 5 min. The cell pellet was resuspended in fresh medium to reduce the
accumulation of cell debris and metabolic byproducts.
3.3
Baculovirus Techniques
3.3.1
Viral Amplification
A 75 cm2 flask was seeded with 5 x 106 Sf9 cells and infected with seed virus
stock at low MOI (less than 1 pfu/cell if seed virus stock had been titrated or use 0.25
– 0.5 ml if it had not). The medium from the cells monolayer was removed and
replaced with the required volume of virus inoculum. When the volume was less than
0.5 ml, medium was added to increase the volume to this amount. The innoculum
was spread out all over the cell monolayer. The inoculum was rocked gently over the
cells every 15-20 mins for 1 hour. After 1 hour at ambient temperature, the inoculum
was removed and replaced with 10 ml of fresh medium. Then the cells were
incubated at 27oC for 4-6 days until they were well infected. It was important that a
low MOI was used at all stages of virus amplification otherwise it was possible that
deletion mutants might occur. Viruses were harvested by centrifuging the infected
medium at 250 x g (1800rpm) 5 minutes and stored at 4oC. The innoculum was
titrated first before being used in an experiment. It was worthwhile freezing small
amounts of intermediate stock at -70oC (for long time storage), and stored the
remainder at 4oC (for at least 1 year).
51
3.3.2
Viral Titration by End Point Dilution Method
This procedure was originally described by Reed and Muench (1938). It was
cheap, easy and reliable. 10-fold dilutions of the virus to be titrated were prepared in
a final volume of 100 µl. The viruses were diluted with complete medium. Generally,
10-1 through 10-8 dilutions were prepared. The easiest and most practical way to do
this was by serial dilution. Dilution could be carried out in the first row of 96-well
micro plate. The last row was kept as non infected cells. 90 µl Sf9 cells suspension
(1.0 x 105 – 2.5 x 105 cells/ml) were added into each well from row 2 to 11. 10 µl of
each virus dilution were added into each well of each row. The cells-virus suspension
was mixed thoroughly but gently. The plate and a damp paper towel were placed into
a plastic bag and sealed tightly. This kept the plate and wells from drying out. The
plate was incubated for 5-7 days at 27oC, checking daily for virus infection.
Occlusion negative virus would take longer time to detect.
Calculation of virus titer was based on Reed and Muench, (1938) method.
The 50% end point was determined by interpolation from the cumulative frequencies
of positive and negative responses to occur in a dilution in which there would be
50% positive responses and 50% negative responses. According to the Poisson
distribution, the proportion (p) of cultures remaining uninfected at any given dose is
e-µ, where µ is the mean concentration of infectious particles at that dose. The
TCID50 is the dose at which 50% of the cultures become infected, that is, p=0.5.
Thus 0.5=e-µ, which implies that µ, the mean concentration of the infectious units at
that dose, is 0.69. Hence, TCID50 x 0.69 = pfu.
52
3.3.3
Generation of Pure Recombinant Virus Stocks by End Point Dilution
Method
The generation of pure virus stock involves the preparation of a stock starting
from a single infectious unit. According to Poisson distribution, 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 are infected.
The cells were diluted to a concentration of 1x106 cells/ml with complete tissue
culture medium. A tenfold serial dilutions of the virus were prepared. Dilutions of
10-6 and 10-7 were appropriate for most stocks. 10 µl of each dilution was mixed with
100 µl of cell suspension and seeded into each well of a 96 well plate. For each
dilution at least 40 replicates were tested. Therefore 2 tenfold dilutions were tested in
one plate including 4 wells for uninfected controls. Plates were incubated at 270C in
humidified environment. Length of incubation was 4-5 days for occ+ and 7 days for
occ- viruses. Each well was examined daily for virus replication and progress of
infection. All wells with sign of infection were scored as positive.
Each infected culture from the highest dilution was harvested provided that
10% or less of the whole culture in that dilution was infected. Each sample was
tested for product gene expression using ELISA. Samples that gave high level of hTf
yield were then selected to undergo the purification process twice more or until the
hTf level reached a constant yield provided that other parameters remained
unchanged for every purification round. Finally the high purity recombinant virus
was amplified to generate a large scale stock.
53
3.4
Protein Analysis Techniques
3.4.1
Sodium Dodecyl Sulphate – Polyacrylamide Gel Electrophoresis, SDSPAGE under Denaturing Condition
Chemicals required were 100 ml of 2 M Tris-HCL (pH 8.8), 100 ml of 1 M
Tris-HCL (pH 6.8), 100 ml of 10% (w/v) SDS, 100 ml of 50% (v/v) glycerol, 1 L of
Coomassie Gel Stain, 1 L of Coomassie Gel Destain, 10 ml of 1% (w/v)
bromophenol blue, 100 ml of Solution A (30 %w/v acrylamide + 0.8 %w/v bisacrylamide), 100 ml of Solution B (75 ml 2 M Tris-HCL + 4 ml 10% SDS + 21 ml
H2O), 100 ml of Solution C (50 ml 1 M Tris-HCL + 4 ml 10% SDS + 46 ml H2O), 1
L of Electrophoresis Buffer (3 g Tris + 14.4 g glycine + 1 g SDS + H2O to make 1
L volume), 5 ml of 10% ammonium persulfate (0.5 g ammonium persulfate + 5 ml
H2O) and 10 ml of 5x Sample Buffer (0.6 ml 1 M Tris-HCL + 5 ml 50% glycerol +
2 ml 10% SDS + 0.5 ml 2-mercaptoethenol + 1 ml 1% bromophenol blue + 0.9 ml
H2O).
Gel plates were assembled according to Mini-Protean II manuals. Separating
gel with X% polyacrylamide was prepared which consisted of X/3 ml of Solution A,
2.5 ml of Solution B, (7.5 –X/3) ml of H2O, 50 µl of 10% ammonium persulfate and
5 µl of TEMED.
The separating gel was poured into the gel plate until it reached 1.5 cm below
the front plate. The gel was left for 30 minutes at room temperature to polymerize.
Next, stacking gel with 5% polyacrylamide was prepared. Stacking gel was consisted
of 0.67 ml of Solution A, 1 ml of Solution C, 2.3 ml of H2O, 30 µl of 10%
ammonium persulfate and 5 µl of TEMED.
54
The stacking gel was poured into the gel plate until it reached the top of the
front plate. The comb was inserted into the stacking gel and left for 30 minutes to
polymerize. The comb was then removed and the gel plate was attached to the
electrode assembly. The gel-electrode-assembly was inserted into the electrophoresis
chamber which was filled with electrophoresis buffer.
Samples for electrophoresis were prepared by mixing 200 µl of protein
sample with 50 µl of 5x sample buffer. The samples were heated in boiling water for
2-10 minutes. The samples were spun to remove debris. Using a Hamilton syringe,
20 µl of sample was introduced into the gel well. The gel was eletrophorase at 200V
for 45 minutes. When the electrophoresis had completed, the gel was removed from
the plates and stained with 20 ml of coomassie gel staining solution for 30 minutes.
The gel was destained using 50 ml of coomassie gel destaining solution for one hour
and repeated for 24 hours.
3.4.2
Bicinchoninic Acid (BCA) Assay
This method was used to quantify the concentration of total protein in the
medium. The assay was a developed variation of the Lowry assay. It was simpler to
perform, faster and had fewer interfering substances.
Chemicals required were Reagent A (pH 11.25, 1% BCA, 2 % Na2C03.H2O,
0.16% Na2C4H4O6.2H2O, 0.4% NaOH, 0.95% NaHCO3), Reagent B (4%
CuSO4.5H2O) Standard Working Reagent (SWR) was prepared based on volume
ratio (50 volumes Reagent A : 1 volume Reagent B).
55
Standard Bovine Serum Albumin (BSA) solutions were prepared in the range
of 0.2 – 1.0 mg/ml. Samples were prepared 5 - 10 times more dilute to fall within the
standard concentration range. 1 volume of sample and 20 volumes of SWR were
mixed in the test tubes (i.e. 0.025 ml and 0.5 ml). This step was done with the test
tube immersed in cold water to delay the reaction. The test tubes were incubated at
37oC for 30 minutes or 60oC for 15 minutes. The test tubes were cooled down in cold
water. Absorbance readings were taken at 562 nm. Samples concentrations were
determined based on the absorbance of standard BSA concentrations.
3.4.3
Enzyme Linked Immunosorbent Assay, ELISA
Chemicals required were Coating Buffer (0.05 M Sodium Carbonate, pH
9.6), Wash Solution (50 mM Tris, 0.14 M NaCl, 0.05% Tween 20, pH 8.0),
Blocking/Postcoat Solution (50 mM Tris, 0.14 M NaCl, 1% BSA, pH 8.0),
Sample/Conjugate Diluent (50 mM Tris, 0.14 M NaCl, 1% BSA, 0.05% Tween 20,
pH 8.0), Enzyme Substrate (TMB, OPD or ABTS can be used), Stopping Solution (2
M H2SO4 or other appropriate solution), Coating Antibody (Goat anti-Human
Transferrin-affinity purified), Calibrator (Human Reference Serum) and HRP
Detection Antibody (Goat anti-Human Transferrin-HRP conjugate)
Each well was coated with 100µl capture antibody and incubated for 60
minutes. After incubation, the excess capture antibody was aspirated and each well
was washed. Empty site was blocked using 200µl blocking solution and incubated
for 30 minutes. After incubation, excess blocking solution was removed and each
well was washed. 100µl samples and standards were added into each well and
incubated for 60 minutes. After incubation, excess samples and standards were
removed and each well was washed. 100µl HRP detection antibody was added and
incubated for 60 minutes. After incubation, excess HRP conjugate was removed and
56
each well was washed. 100µl Enzyme substrate was added and incubated for 5-30
minutes. Reaction was stopped using 200µl 2 M H2SO4. Absorbance reading at 450
nm was taken using microtiter plate reader.
3.4.4
ELISA-Conversion of Calibrated Data to Actual Product Concentration
The actual product concentration was assumed as Y ug/ml. When 5 µl of
sample was diluted in 395 µl diluent, therefore, the dilution factor was
(395 + 5) / 5 = 80 times more dilute
…3.3
The new product concentration after dilution is
Y µg/ml x 0.005 ml = 0.0125Y µg/ml
…3.4
0.400 ml
This (0.0125Y) was the value that was calibrated from the standard curve. To obtain
the original product concentration, the calibrated value has to be multiplied by the
dilution factor,
Original product concentration = 0.0125Y µg/ml x 80 = Y µg/ml …3.5
The standard curves for ELISA were based on OD versus concentration and
not OD versus total protein. So whatever dilution made was the factor of dilution and
must be multiplied by the calibrated value to obtain the actual product concentration.
57
3.5
Recombinant Human Transferrin (rhTf) Expression and Optimization
3.5.1
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 µl 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.
3.5.2
Medium Screening
Based on the literature reviews, 13 nutrients were selected to be screened in
the Sf9 insect cells monolayer culture.
Chemicals required were D-fructose, D-glucose, Maltose, L-arginine, Lcysteine, L-glutamine, L-lysine, L-methionine, L-serine, L-threonine, L-tyrosine, Lvaline, and Lipid mixtures 1000x (already in solution form). All chemicals were
purchased from Sigma, USA.
58
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.
1.0 ml each of the 33 designed medium compositions was prepared in 2 x 24well 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 for 3 times and
the results were analyzed using the software Statistica (Statsoft, v. 5.0).
3.5.3
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 4.3.2). 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
59
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
20 µl 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 the software Statistica
(Statsoft, v. 5.0).
3.6
Response Surface Methodology, RSM (Method of Steepest Ascent)
A Taguchi design array (3 parameters and 3 levels) was generated from the
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
60
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.
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 were
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.
CHAPTER 4
RESULTS AND DISCUSSIONS
4.1
The Study of Sf9 Insect Cells Culture Growth Profiles
4.1.1
Sf9 cell growth in T-flask (Monolayer) and Shake flask (Suspension)
Growth curves of Sf9 cells monolayer at different seeding densities are
shown in Figure 4.1. At initial density of 0.4 x 106 cells/ml, the cells could multiply
efficiently thus achieving maximum density. At lower seeding densities, the cells
viability could also be maintained for a longer period. When the seeding density was
increased, the exponential phase became shorter thus unabling the cells to reach
optimum density. Once the Sf9 cells monolayer reached confluency, any futher cells
division would resulted in multilayers of cells as observed in this research. Some
cells began to float, clump and form clusters post confluency.
Mass transfer in insect cells depends on energy dependent and energy
independent diffusion of nutrients and oxygen across the cell membrane (Bailey and
Ollis, 1986). Two factors that contribute to the diffusivity were area of diffusion and
concentration gradient of molecules in the culture medium (Shuler and Kargi, 2002).
62
12
100
80
8
60
6
40
4
Viability (%)
Viable cells/ml (x10e6)
10
20
2
0
0
0
48
96 144 192 240 288 336 384 432
Time (hour)
SD=0.8x10E6 cells/ml
SD=1.6x10E6 cells/ml
SD=0.4x10E6 cells/ml
Figure 4.1: 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
100
1.2E+07
80
70
8.0E+06
60
50
6.0E+06
40
4.0E+06
Viability (%)
Viable cell density (cells/ml)
90
1.0E+07
30
20
2.0E+06
10
0.0E+00
0
0
48
96 144 192 240 288 336 384 432
Time (hour)
SD=0.4x10E6 cells/ml
SD=1.6x10E6 cells/ml
SD=0.8x10E6 cells/ml
Figure 4.2: 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
63
After confluence was reached, growth area was fully utilized and the cells made
close contact to one another, thus reducing the area of diffusion of a cell. The
concentration gradient of molecules in this region might vary for each cell due to the
absence of fluid capillary as in insect larvae and reduced homogeneity. In this
situation, the over confluent cells competed against each other for nutrients and
oxygen as the area of diffusion and concentration gradient became limited. Some
cells stopped dividing while others tend to overgrow (Freshney, 2000).
In suspension culture (Figure 4.2), the cells could achieve even higher
density (~9.0 x 106 cells/ml) than the monolayer culture (~7 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.
4.1.2
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 4.3, it could be seen that the
growth patterns were similar to Figure 4.2. Next, Sf9 cells were cultured in 1.0 ml of
SFM. In this experiment however, the cells could not propagate properly (Figure
4.4). 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
1.2E+07
100
1.0E+07
80
8.0E+06
60
6.0E+06
40
4.0E+06
Viability ((%)
Viable cells density (cells/ml)
64
20
2.0E+06
0.0E+00
0
0
48
96 144 192 240 288 336 384 432
Time (hour)
SD=0.4x10E6 cells/ml
SD=1.6x10E6 cells/ml
SD=0.8x10E6 cells/ml
Figure 4.3: 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
100
1.0E+07
80
8.0E+06
60
6.0E+06
40
4.0E+06
Viability (%)
Viable cells density (cells/ml)
1.2E+07
20
2.0E+06
0.0E+00
0
0
48
96 144 192 240 288 336 384 432
Time (hour)
SD=0.4x10E6 cells/ml
SD=0.8x10E6 cells/ml
SD=1.6x10E6 cells/ml
Figure 4.4: 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
65
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 4.1).
Table 4.1: Comparison of Sf9 growth in T-flask, Shake flask, and 24-well plate
T-flask
Shake flask
Medium volume
5 ml
50 ml
(a) 1.0 ml
(b) 0.5ml
Initial viability drop
No
Yes
Yes
Yes
Exponential growth
Slow
Fast
Slow
Fast
Life span
Long
Short
Long
Short
Category
Monolayer
Suspension
Combination Suspension
4.1.3
24 well plate
Growth Profiles of Infected Cells
During the early phase of virus infection, adsorptive endocytosis took place,
followed by DNA replication (Volkman and Goldsmith, 1985). Late phase occurred
within 6 to 24 hours post infection where budded viruses were produced
logarithmically (Knudson and Harrap, 1976). At the end of this phase, infected cells
66
100
90
1.2
80
1
70
60
0.8
50
0.6
40
viablity (%)
Viable cells/ml (x10e6)
1.4
30
0.4
20
0.2
10
0
0
0
24
48
4M OI wt
25M OI r
72
96
Time (hour)
120
144
5M OI r
Control
168
10M OI r
Figure 4.5: Growth curves of Sf9 infected with AcMNPV. wt – wild type
baculovirus; r – recombinant baculovirus. Straight lines represent cell density and
dotted lines represent cell viability
WI
R
U
Figure 4.6: Comparison between uninfected (U), wild-type (WI), and recombinant
(R) virus-infected Sf9 cells
stopped growing and budded viruses ceased its production. In this study, the infected
Sf9 cells stopped growing from 24 - 48 hours post infection (Figure 4.5). The cells
viability remained high for about 2 days and dropped to 50% at day 3. The very late
67
phase occurred when cells ceased production of budded virus and began the
assembly, production and expression of recombinant gene product. The physical
appearances of infected cells are shown in Figure 4.6. Sf9 cells infected with wild
type baculovirus shows small granules within the cells whereby cells infected with
recombinant baculovirus shows rough surfaces around the cells.
4.1.4
Growth Analysis
Growth rate constants of Sf9 cells are shown in Figure 4.7. For T-flask, the
optimum growth rate is 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 stay within close vicinities. For
shake flask, the optimum growth rate is at the seeding density of 0.4 x 106 cells/ml
and this figure stays 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)) is slightly
lower than the T-flask and shaker. In general, the growth rates are stable suggesting
that the growth of Sf9 is not really affected by its seeding density.
For 24-well plate at 1.0 ml SFM (24well (a)) however, the Sf9 growth rate is
greatly affected by its seeding density. At 0.4 x 106 cells/ml, the growth rate constant
is 0.012 hr-1 while at 1.6 x 106 cells/ml, the value is 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 4.1.1.
The doubling times of Sf9 in various cultivators and at different seeding
densities are shown in Figure 4.8. A healthy Sf9 cell doubles 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
68
Growth Constant , µ (1/hr)
0.02
0.015
Tflask
Shaker
0.01
24well(a)
24well(b)
0.005
0
0.4
0.8
1.6
Seeding Density (x10e6 cell/ml)
Figure 4.7: Growth rate constants of Sf9 in various cultivators and at different
seeding densities
300.00
Doubling Time (hour)
250.00
200.00
Tflask
Shaker
150.00
24well(a)
24well(b)
100.00
50.00
0.00
0.4
0.8
1.6
Seeding Density (x10e6 cell/ml)
Figure 4.8: Doubling time of Sf9 in various cultivators and at different seeding
densities
69
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 to Sf9 cells grown in a shake flask.
4.2
Study on the Expression Profiles of rhTf in Infected Sf9 Insect Cells
Culture
4.2.1
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 4.9). In lane 4 of Figure 4.9 (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 4.9). 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.
70
1
2
3
4
5
6
7
8
9
10
225
150
100
75
Apo-hTf
Recombinant-hTf
50
35
25
Figure 4.9: 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)
rhTf concentration ( g/ml)
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
0
1
2
5 MOI
3
4
5
6
Days post infection
10 MOI
7
8
25 MOI
Figure 4.10: 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
71
produced extensively (O’Reilly et al., 1994). From Figure 4.10, it could be seen that
during the first day, rhTf yield was still low (~1.0 µg/ml). rhTf was produced
extensively after 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 (Figure
4.5). 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.
4.2.2
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 4.11, 4.12) with
approximately 10 µg/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
72
12.0
[hTf] ( g/ml)
10.0
8.0
6.0
4.0
2.0
0.0
0
2
4
8
Time post infection (day)
SD=0.8x10e6 cells/ml
SD=3.2x10e6 cells/ml
16
SD=1.6x10e6 cells/ml
SD=5.6x10e6 cells/ml
Figure 4.11: rhTf expression profiles at different seeding densities, SD.
Experiments were done in T-flask
Figure 4.12: Surface plot of figure 4.11
73
4.2.3
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 (refer to Figure 4.1). 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 4.14).
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 4.2). 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 4.13) and the yield was minimal too (Table 4.2).
74
14.0
[hTf] ( g/ml)
12.0
10.0
8.0
6.0
4.0
2.0
0.0
0
TOI=day 0
TOI=day 4
2
4
8
Time post infection (day)
TOI=day 1
TOI=day 6
12
TOI=day 2
Figure 4.13: rhTf expression profiles at different times of infection, TOI
Figure 4.14: Surface plot of figure 4.13
75
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
rhTf concentration however, was found to continue increasing after the deceleration
phase.
Table 4.2: rhTf yield coefficients at various seeding densities, MOIs, and times of
infection. Yield was based on day 4 post infection and 103 cells .
MOI
1.0
2.5
5.0
7.5
10.0
12.5
Yield at day
4 post
infection
(ng/103cell)
1.76
2.82
1.70
1.58
1.76
1.70
Seeding
Density
(106viable
cells/ml)
0.8
1.6
3.2
5.6
Yield at day
4 post
infection
(ng/103cell))
4.71
3.18
0.83
0.50
Time of
Infection
(day post
culture)
0
1
2
4
6
Yield at day
4 post
infection
(ng/103cell)
4.66
3.82
4.87
0.65
0.38
One possible reason is that, after day 4, there were still some viable cells
existed in the culture (refer to Figure 4.5). The cells were still reproducible and
might be able to express further 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
76
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.
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 later into suspension culture.
4.3
Optimization of the Recombinant Human Transferrin Expression
4.3.1
Recombinant Baculovirus Screening
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 4.3 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 4.15). 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 4.15).
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 4.4). 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.
77
Table 4.3: Concentration (µg/ml) of rhTf in each well of a 96-well plate
Row
A
B
C
D
E
F
G
H
Column
(High MOI) Dilution factor =10-6
2
3
4
5
6
0.495 0.055 0.0
0.0
0.0
0.0
0.044 0.011 0.0
0.0
0.0
0.0
0.0
0.0
0.050
0.0
0.0
0.0
0.0
0.017
0.0
0.0
0.017 0.0
0.0
0.0
0.011 1.100 0.011 0.000
0.0
0.066 0.363 0.011 0.000
0.031 0.0
0.072 0.011 0.715
(Low MOI) Dilution factor =10-7
7
8
9
10
11
0.022 0.0 0.0
0.0
0.0
0.0
0.0 0.0
0.0
0.0
0.0
0.0 0.0
0.0
0.0
0.0
0.0 1.045 0.028 0.011
0.0
0.0 0.066 0.011 0.0
0.0
0.0 0.022 0.0
0.0
0.0
0.0 0.011 0.011 0.0
0.011 0.0 0.011 0.011 0.0
rhTf-AcMNPV Cloning (1st plate results)
1.20
Row H
Row G
Row F
Row E
Row D
Row C
Row B
Row A
1
2
3
4
5
6
7
8
9
10
11
12
rhTf (ug/ml)
1.00
0.80
0.60
0.40
0.20
0.00
Column no
Figure 4.15: 3D plot of Table 4.3
rhTf yield in the purified baculovirus stock was very low ~1µg/ml (F4 and
D9 in Table 4.3) compared to the original stock ~10µg/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 4.3). When the high purity baculovirus stock was amplified
and used to express rhTf at the same experimental conditions, the yield was
~15µg/ml. This proved that the screened or purified recombinant baculovirus had
resulted in improved production of rhTf.
78
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
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.
The cause of this effect was well 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):
exp(−α (Vl (t ) / No(r ))) x(α (Vl (t ) / No(t ))) k
P(t , k ) =
k!
…4.1
where α 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 ‘µ’.
79
exp(− µ ) x( µ ) k
P(t , k ) =
k!
…4.2
Table 4.4: Viral Screening by End Point Dilution Method (Poisson distribution data
sheet)
Proportion
Of
Virus
Culture
Purity
µ
(MOI) (/100%) Infected
0
0
0
0.001 1
0.001
0.006 0.997
0.006
0.011 0.995
0.011
0.016 0.992
0.016
0.021 0.99
0.021
0.026 0.987
0.026
0.076 0.962
0.073
0.081 0.96
0.078
0.086 0.958
0.082
0.091 0.955
0.087
0.096 0.953
0.092
0.101 0.95
0.096
0.106 0.948
0.101
0.111 0.946
0.105
0.116 0.943
0.11
0.181 0.912
0.166
0.186 0.91
0.17
0.191 0.908
0.174
0.196 0.905
0.178
0.201 0.903
0.182
0.207 0.9
0.187
0.212 0.898
0.191
0.217 0.895
0.195
0.222 0.893
0.199
0.297 0.859
0.257
0.302 0.857
0.261
0.307 0.854
0.264
0.312 0.852
0.268
0.317 0.85
0.272
0.322 0.848
0.275
Average
rhTf yield
(mg/106cell)
1.23
3.08
4.08
6.01
Proportion
Of
Virus
Culture
Purity
µ
(MOI) (/100%) Infected
0.327
0.845
0.279
0.332
0.843
0.283
0.337
0.841
0.286
0.342
0.839
0.29
0.432
0.8
0.351
0.532
0.757
0.413
0.632
0.717
0.468
0.732
0.678
0.519
0.832
0.641
0.565
0.932
0.605
0.606
1.257
0.5
0.715
1.357
0.47
0.743
1.457
0.442
0.767
1.557
0.416
0.789
1.657
0.39
0.809
1.757
0.366
0.827
2.057
0.302
0.872
2.557
0.215
0.922
3.057
0.151
0.953
3.617
0.1
0.973
4.117
0.068
0.984
4.617
0.046
0.99
5.117
0.031
0.994
5.617
0.02
0.996
6.117
0.014
0.998
9.117
0.001
1
9.617
0.001
1
10.12
0
1
10.62
0
1
11.12
0
1
11.62
0
1
Average
rhTf yield
(mg/106cell)
12.00
20.00
22.00
80
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).
Purity =
µ exp(− µ )
1 − exp(− µ )
…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 µ would be
calculated and used to find the purity of the recombinant baculovirus stock. In this
case, Table 4.4 was generated to monitor whether the definition of virus purity based
on Poisson distribution was valid. Note that the dynamic MOI (µ) in Table 4.4 is
only valid at time of infection. From Table 4.4, 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 4.3 and also by Hu and Bentley (2000). As the
virus innoculum was diluted in Table 4.3, 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 4.4). 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
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.
81
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.
Besides that, it was easy to handle as it 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 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 4.5.
Table 4.5: Factors affecting the end point dilution method
Factors
1. Virus dilution
Effect
Low dilution >High concentration
High dilution >Low concentration
2. Cells concentration High
concentration
infectivity
Low
concentration
infectivity
>
low
>
high
3. Number of cultures Many cultures with few infected
Many cultures with many infected
exposed to virus
4.3.2
Virus purity
¾ low purity
¾ high purity
¾ low purity
¾ high purity
¾ better purity
¾ low purity
Medium Screening
The lower and higher values for each nutrient for medium screening are
shown in Table 4.6. The Plackett-Burman Screening Design is shown in Table 4.7.
82
Table 4.6: Real values for the screening of 13 selected nutrients using PlackettBurman design
Nutrients
SUGARS
D-Fructose (D-Fruc)
D-Glucose
(D-Gluc)
Maltose
(Malt)
AMINO ACIDS
L-Arginine
(L-Arg)
L-Cysteine
(L-Cys)
L-Glutamine (L-Gln)
L-Lysine.HCl (L-Lys)
L-Methionine (L-Met)
L-Serine
(L-Ser)
L-Threonine (L-Thr)
L-Tyrosine
(L-Tyr)
L-Valine
(L-Val)
OTHER COMPONENTS
Lipid Mixtures (Lip)
Added Concentration
(-1)
(+1)
Unit
Lower
Higher
mg/l
0.00
1200
mg/l
0.00
3800
mg/l
0.00
3000
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
mg/l
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1200
180
2600
2425
1170
1325
625
1200
600
% v/v
0.00
0.8
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 4.7). 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
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.
83
Table 4.7: 13-factor (nutrients), 33-run, 2-level Plackett-Burman screening design
Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
DFruc
1
-1
-1
-1
-1
1
-1
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
DGluc
1
-1
-1
-1
1
-1
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
Malt
1
-1
-1
1
-1
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
LArg
1
-1
1
-1
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
LCys
1
1
-1
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
-1
LGln
1
-1
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
LLys
1
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
-1
LMet
1
-1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
1
-1
LSer
1
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
1
-1
-1
LThr
1
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
1
-1
1
-1
LTyr
1
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
1
-1
1
1
-1
LVal
1
-1
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
1
-1
1
1
1
-1
Lip
1
1
-1
-1
-1
1
1
1
1
1
-1
-1
1
1
-1
1
-1
-1
1
-1
-1
-1
-1
1
-1
1
-1
1
1
1
-1
1
-1
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 4.16). Somehow, the lipid mixtures might have
affected the ability of the cells to express recombinant protein. Figure 4.17 shows
the results of three medium screenings which were carried out based on the PlackettBurman screening designs (Table 4.6, Table 4.7). 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
84
ml were 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.
A
B
Figure 4.16: Infected cells appearance in medium A (lipid mixtures added) and
medium B (no lipid mixtures added)
30.0
25.0
Screen1
Screen2
Screen3
rhTf (ug/ml)
20.0
15.0
10.0
5.0
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Experiments no.
Figure 4.17: 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
85
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
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 4.7) 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.
Ex1 Ex4
Ex7
Ex12 M (kDa)
S Ex17
Ex24 Ex32 Ex33
225
150
Molecular Weight (kb)
100
nhTf
75
rhTf
50
35
25
Figure
4.18:
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 µl. Samples from Screen 3
86
SDS-PAGE analysis was conducted to assess the quality of rhTf and the
results are presented in Figure 4.18. 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 µg/ml and the highest was Ex7~ 22 µg/ml. Production of heterologous
proteins ranging from 25-225 kDa was observed. The thickness of each band is in
agreement with the data in Figure 4.17. Figure 4.18 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 Man3GlcNAc2 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 4.8
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
87
(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.
Table 4.8: Estimated effects on rhTf yield based on the results of Plackett-Burman
screening experiments
Factor
Mean/Interc.
LipMix
Gln
Gluc
Cys
Val
Met
Thr
Ser
Fruc
Tyr
Malt
Effect
12.18
5.59
1.90
0.92
0.79
0.72
0.35
0.21
0.11
-0.28
-0.44
-0.52
Std.Err.
0.36
0.71
0.71
0.71
0.71
0.72
0.71
0.71
0.71
0.71
0.71
0.71
T(82)
34.24
7.88
2.66
1.29
1.11
1.01
0.49
0.29
0.16
-0.40
-0.61
-0.73
p
0.00
0.00
0.01
0.20
0.27
0.32
0.62
0.77
0.88
0.69
0.54
0.47
-95%
11.48
4.18
0.48
-0.50
-0.63
-0.70
-1.06
-1.21
-1.30
-1.70
-1.85
-1.94
95%
12.89
7.00
3.32
2.33
2.21
2.15
1.77
1.62
1.52
1.13
0.98
0.89
Effects (ug rhTf/unit nutrient)
6
5
4
3
2
1
0
-1
Lip Gln Gluc Cys Val Met Thr Ser Fruc Tyr Malt
Factors (Nutrients)
Figure 4.19: 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
88
It is clearly presented in Figure 4.19 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 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 4.8). 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.
70
60
50
40
59 58 57
50
45 45 42
41 38
30
34 32
30 28
26 26
20
10
19 17
15
9 8
0
Leu
Lys
Ala
Gly
Val
Asp
Glu
Ser
Cys
Asn
Pro
Thr
Phe
Tyr
Arg
His
Gln
Ile
Met
Trp
Number of Amino Acids
89
Amino Acids
Figure 4.20: 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
4.20). 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.
90
4.3.3
Medium Optimization using Response Surface Methodology
4.3.3.1 Regression Model
The results of the optimization experiments are shown in Figure 4.21 and
Table 4.9. In the control experiment (test no. 17), where there was no additional
nutrients feeding, rhTf concentration was 19.89 µg/ml. The maximum rhTf yield
was in test no. 5 with 62.28 µg/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 4.10. 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 − 5.63x1x2 − 3.95x1x3 +
2.79x2x3 + 4.3112 + 3.21x22 − 9.99x32
…4.4
where Y is rhTf response in µg/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 4.9.
91
Table 4.9: Central composite design for the optimization of glutamine, glucose and lipid mixtures 1000x
Test No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Control
Coded
Values
Gln
-1
-1
-1
-1
1
1
1
1
-1.7
1.7
0
0
0
0
0
0
-2
Gluc
-1
-1
1
1
-1
-1
1
1
0
0
-1.7
1.7
0
0
0
0
-2
Lip
-1
1
-1
1
-1
1
-1
1
0
0
0
0
-1.7
1.7
0
0
-2
mg/L
Gln
2500
2500
2500
2500
7500
7500
7500
7500
796
9204
5000
5000
5000
5000
5000
5000
0
Real Values
mg/L
Gluc
2500
2500
7500
7500
2500
2500
7500
7500
5000
5000
796
9204
5000
5000
5000
5000
0
%v/v
Lip
0.4
1.1
0.4
1.1
0.4
1.1
0.4
1.1
0.8
0.8
0.8
0.8
0.1
1.4
0.8
0.8
0
Actual
37.89
25.77
32.65
25.76
62.28
28.46
28.62
15.47
49.41
47.93
56.38
34.60
12.66
2.05
36.84
30.22
19.89
rhTf (µg/ml)
Predicted Residual
37.17
0.72
27.55
-1.78
30.20
2.45
31.73
-5.97
58.06
4.22
32.66
-4.20
28.59
0.03
14.33
1.14
45.97
3.44
48.95
-1.02
55.03
1.35
33.53
1.07
16.29
-3.63
-4.00
6.05
35.02
1.82
35.02
-4.79
20.79
-0.90
92
rhTf yield (ug/ml)
80
70
Observed
60
Predicted
50
40
30
20
10
0
1 2
3 4 5 6
7 8 9 10 11 12 13 14 15 16 17
Test No.
Figure 4.21: Observed and predicted experimental data for the optimization of
glutamine, glucose and lipid mixtures. Medium composition was based on Table 4.9
Analysis of variance (ANOVA) was done using Statistica (Statsoft v. 5.0)
and the results are given in Table 4.10. The fisher F-test value signifies how great
the mean square of the regressed model compares to mean square of the residuals
(errors). The F value for this case is 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 SSregression to SStotal. In this case, the value of the determination coefficient
(R2 = 0.96) indicates that only 4% of the total variations are not explained by
equation 4.4. The value of the adjusted determination coefficient (Adj. R2 = 0.90) is
also very high, which indicates a high significance of the model. The correlation
coefficient (R = 0.98) shows a significant correlation between the independent
variables and the rhTf response.
93
Table 4.10: Analysis of Variance (ANOVA) of the CCD
Regression Statistics
Multiple R
R Square
Adjusted R
Square
Standard Error
Observations
0.98
0.96
0.90
5.00
17
ANOVA
9
7
16
Sum
of
Squares
3754.50
174.75
3929.24
Coefficient
35.02
0.87
4.31
-6.32
3.21
-5.97
-9.99
-5.63
-3.95
2.79
Standard
Error
2.87
1.33
1.39
1.33
1.39
1.33
1.39
1.63
1.63
1.63
Degree
Of
freedom
Regression
Residual
Total
Variables
Intercept
Gln
Gln x Gln
Gluc
Gluc x Gluc
Lip
Lip x Lip
Gln xGluc
Gln x Lip
Gluc x Lip
Mean
Square
417.17
24.96
F-value Significance F
16.71
0.00
t-value
12.22
0.66
3.09
-4.75
2.30
-4.48
-7.17
-3.45
-2.42
1.71
P-value
0.00
0.53
0.02
0.00
0.05
0.00
0.00
0.01
0.05
0.13
The significance of each coefficient was determined by student's t-test and P
values, which are listed in Table 4.10. 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 twoway 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 twoway 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 (Px32 = 0.00001 > Px1
2
= 0.0200 > Px22 =
94
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 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.
4.3.3.2 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
constructed. The third nutrient was held at its center point. The results of the surface
plots are shown in Figure 4.22, 4.23, and 4.24.
Glutamine and glucose interactions are shown in Figure 4.22 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).
95
Figure 4.22: Glutamine (Gln) vs Glucose (Gluc) vs rhTf
Figure 4.23: Glutamine (Gln) vs Lipid Mixtures 1000x (Lip) vs rhTf
96
Glutamine and lipid mixtures interactions are shown in Figure 4.23. 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 4.23, 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.
Figure 4.24: Glucose (Gluc) vs Lipid Mixtures 1000x (Lip) vs rhTf
97
Glucose and lipid mixtures interactions are shown in Figure 4.24. 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 4.24, the quadratic effect of lipid mixtures is very
significant as it is in Figure 4.23. 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 4.22 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.
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 x3=-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 µg/ml. An experiment performed using the
optimized parameters resulted in rhTf yield of 65.12µg/ml. This result therefore,
verified the trend of the predicted value and the effectiveness as well as the
usefulness of the model towards achieving the optimization.
98
4.4
Characterization of the Optimized Recombinant Human Transferrin
Expression
Table 4.11: Summary of the characteristics of optimized rhTf expression
SUMMARY
Parameter
Seeding Density
MOI
Time of Infection
Length of Cultivation Period
Length of Lag Phase
Length of Exponential Phase
Doubling Time
Specific Growth Rate, µ
Max Glucose Uptake Rate
Max Lactate Production Rate
Max rhTf Production Rate
Unit
cell/ml
pfu/cell
hr
hr
hr
hr
hr
1/hr
µg/106cell/hr
6
µg/10 cell/hr
6
µg/10 cell/hr
6
Control
Optimized
1600000
1600000
15
15
48
48
240
240
0
36
84
60
30.52
62.24
0.5451
0.2673
7699.62
6619.75
402.57
1835.04
0.0070
0.0741
6.1145
Max Protein Production Rate
µg/10 cell/hr
0.4718
Max Cell Density
cell/ml
7.71E+06 2.97E+06
Max Protein Density
µg/ml
5300.67
8665.55
Max rhTf Density
µg/ml
18.57
69.37
In Figure 4.25, the Sf9 growth rate of optimized expression is lower than the
controlled expression. The specific growth rates of controlled and optimized
expression are 0.5451 hr-1 and 0.2673 hr-1 respectively (Table 4.11). This is 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
is also observed for the optimized expression and this is assumed to be due to
medium adaptation. From Figure 4.25, 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.
99
100
14
80
10
8
60
6
40
Viability (%)
Viable Cell Density
(x10e6 cells/ml)
12
4
20
2
0
0
0
2
4
6
8
Days in culture
10
12
Cell Density (Control)
Cell Density (Optimized)
Viability (Control)
Viability (Optimized)
Figure 4.25: Sf9 growth in controlled and optimized expression. Dotted line
indicates where infection was initiated
1.2
1.0
8
0.8
6
0.6
4
0.4
2
rhTf percentage (%)
Protein Concentration in
Medium (mg/ml)
10
0.2
0
0.0
0
2
4
6
8
Days in culture
10
12
T otal Protein (Control)
T otal Protein (Optimized)
rhT f (Optimized)
rhT f (Control)
Figure 4.26: Total protein and rhTf contents in controlled and optimized expression
100
Figure 4.26 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 µg/ml
and 8665.55 µg/ml respectively. Protein concentration drops were observed after
four days of infection. This occurrence is 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 becomes 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 4.26, 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.
10
0.1
8
0.08
6
0.06
4
0.04
2
0.02
0
0
rhTf Production Rate
( g/10e6cells/hr)
Total Protein Production Rate
( g/10e6cells/hr)
101
-2
-0.02
-4
-0.04
0
2
4
6
8
Days in culture
10
12
T otal Protein (Optimized)
T otal Protein (Control)
rhT f (Optimized)
rhT f (Control)
Figure 4.27: Total protein and rhTf production rates in controlled and optimized
12
1.2
10
1
8
0.8
6
0.6
4
0.4
2
0.2
0
0
0
2
4
6
8
Days in culture
10
Lactate Concentration (g/L)
Glucose Concentration (g/L)
expression
12
Glucose (Optimized)
Glucose (Control)
Lactate (Optimized)
Lactate (Control)
Figure 4.28: Glucose and lactate concentrations in controlled and optimized
expression
102
Figure 4.27 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 4.26, 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
µg/106cell/hr) which was 13 times higher than the controlled expression (0.47
µg/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
µg/106cell/hr which was 11 times higher than the production rate of non-optimized
rhTf (0.007 µg/106cell/hr). These observations clearly showed that rhTf production
was significantly affected by the optimized medium.
Glucose 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 to medium
storage and exposure to open environment was negligible. The results are displayed
in Figure 4.28. As for glucose, the concentration was continuously depleted during
the course of cultivation. Glucose was also not a limiting factor since more than
2g/L 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
103
limiting oxygen, no lactate was produced. During this period (lactate drop), it was
also observed that Sf9 growth was in the exponential phase (Figure 4.25). 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 4.25).
Lactate Production/Glucose Uptake
Rate ( g/10e6cells/hr)
8000
6000
4000
2000
0
0
2
4
6
8
Days in culture
10
12
Lactate (Control)
Lactate (Optimized)
Glucose (Control)
Glucose (Optimized)
Figure 4.29: 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 4.29. 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 4.25).
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
104
requirements of Sf9 which needs more energy for rhTf assembly. This was
supported by the increase in rhTf concentration in optimized medium (Figure 4.26).
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.
Day0
Day2
Day4 MW(kDa) Day6
Day8
Day10
225
150
100
75
50
rhTf~73kDa
A~60kDa
B~44kDa
35
C~27kDa
25
Figure 4.30: 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 4.30 and 4.31. The thickness
of the rhTf bands agreed with the results shown in Figure 4.26. 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 4.30 and 4.31 basically reduced towards the end of
infection phase. This was because, at the very late phase of infection cycle, Sf9 cells
105
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.
Day0
Day2
Day4 MW (kDa) Day6
Day8
Day10
225
150
100
75
rhTf~73kDa
A~60kDa
50
B~44kDa
35
C~27kDa
25
Figure 4.31: 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 4.29).
CHAPTER 5
CONCLUSIONS
5.1
Summaries
5.1.1
rhTf Expression in Sf9 Insect Cells Monolayer Culture
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 11.2µg/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
107
culture. This work helped to monitor any changes that occurred when working with
suspension culture and decided on how to optimize rhTf expression.
5.1.2
Utilization of 24-well Plates for Insect Cells Suspension Culture
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 mimick 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, the miniaturized experiment could not mimick 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.
5.1.3
Medium and Baculovirus Screening
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
108
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.
5.1.4
Response Surface Methodology
The use of central composite design and response surface methodology has
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 µg/ml to 65.12 µg/ml.
5.2
Recommendations
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 the 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
109
glycoprotein-related diseases. Further study on the improvement of the recombinant
glycoprotein expression can be made as follow:
5.2.1
Large Scale Study in Bioreactor
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 as follow:
¾ oxygen uptake,
¾ nutrients depletion,
¾ power consumption,
¾ mixing time,
¾ shear rate, and
¾ heat transfer coefficient.
110
5.2.2
Expression and Purification of Biologically Active Glycoprotein
In order to ensure the expression of biologically active rhTf, an assay to
measure rhTf activity is needed. One of the techniques will be iron uptake analysis.
The ability to produce a large quantity of biologically active rhTf will be a definite
plus and this will help in the development of processes for the production of
therapeutic glycoproteins in insect cells baculovirus expression system.
Protein purification is usually a multi-step process exploiting a wide range of
biochemical and biophysical characteristics of the target protein, such as its source,
relative concentration, solubility, charge, and hydrophobicity. Ideally, purification
strives to obtain maximum recovery of the desired protein, with minimal loss of
activity, combined with maximum removal of other contaminating proteins. When
designing a purification protocol, one should aim for the following:
¾ high recovery,
¾ highly purified end product (>99.99%),
¾ reproducibility, and
¾ economical use of reagents.
REFERENCES
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.
Aisen, P. and Listowsky, I. (1980). Iron transport and storage proteins. Annu. Rev.
Biochem. 49:357-393.
Akhnoukh, R., Kretzmer, G. and Schugerl, K. (1996). On-line monitoring and
control of the cultivation of Spodoptera frugiperga Sf9 insect cells and betagalactosidase 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.
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.
112
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).
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.
Barnes, D. and Sato, G. (1980). Methods for growth of cultured cells in serum-free
medium. Anal. Biochem. 102(2):255-270.
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. 15:129138.
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.
113
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.
Bhatia, R., Jesionowski, G., Ferrance, J. and Ataai, M. M. (1997). Insect cell
physiology. Cytotechnology. 24:1-9.
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.
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. C., Godfray, H. C. J. and O'Reilly, D. R. (2001). Persistence of an OcclusionNegative Recombinant Nucleopolyhedrovirus in Trichoplusia ni Indicates
High Multiplicity of Cellular Infection. Applied and Environmental
Microbiology. 67:5204-5209.
Butler, M., Huzel, N., Barnabe, N., Gray, T. and Bajno, L. (1999). Linoleic acid
improves the robustness of cells in agitated cultures. Cytotechnology. 30:2736.
Butters, T. D. and Hughes, R. C. (1981). Isolation and characterization of mosquito
cell membrane glycoproteins. Biochem. Biophys. 640:655-671.
114
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.
Cornell, J. A. (1990). How to Apply Response Surface Methodology. U.S.A.:
American Society for Quality Control.
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.
David, P. C. and Lonnie, D. R. (1997). Molecular Biology. U.S.A.: Cache River
Press.
115
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).
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.
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. 42:697707.
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.
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.
116
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.
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 H nuclear magnetic resonance
spectroscopy. Anal. Biochem. 206:53-63.
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.
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.
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).
117
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).
Haaland, P. D. (1989). Experimental Design In Biotechnology. New York, U.S.A.:
Marcel Dekker, Inc.
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.
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.
Hsieh, P. and Robbins, P. W. (1984). Regulation of asparagine-linked
oligosaccharide processing. Oligosaccharide processing in Aedes albopitus
mosquito cells. J. Biol. Chem. 259:2975-2382.
118
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., 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.
Huebers, H. A., Finch, C. A. (1987). The physiology of transferrin and transferrin
receptors. Physiol. Rev. 67(2):520-82.
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).
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.
Jarvis, D. L. and Carrington, J. C. (1993). Method and vector for the purification of
foreign proteins. (U.S Patent 5,179,007).
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.
119
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.
King, L. A. and Possee, R. D. (1992). The baculovirus expression system: a
laboratory guide. London: Chapman and Hall.
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., 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.
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.
Kumar, S. and Miller, L. K. (1987). Effects of serial passage of Autographa
Californica Polyhedrosis Virus in cell culture. Virus Res. 7:335-350.
120
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.
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).
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, P. and Bailey, J. E. (1991). Factors influencing recombinant proteins yields in
an insect cell-baculovirus expression system: multiplicity of infection and
intracellular protein degradation. Biotechnol. Bioeng. 37:238-246.
Lineback-Zins, J. and Brew, K. (1980). Preparation and characterization of an NH2terminal fragment of human serum transferrin containing a single ironbinding 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.
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: McGrawHill. 1-24.
Luckow, V. A. and Summers, M. D. (1988). Trends in the development of
baculovirus expression vectors. Biotechnol. 6:47-55.
Lynn, D. E. (2002). Effects of temperature on the susceptibility of insect cells to
infection by baculoviruses. Methods in Cell Science. 23: 221-225.
121
MacGillivray, R. T. and Brew, K. (1975). Transferrin: internal homology in the
amino acid sequence. Science. 190(4221):1306-1307.
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., 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.
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., Shauger A., Howarth, B. and Inlow, D. (1988). Supply of lipids in
animal cell culture media. Abstr. Pap. Am. Chem. Soc. 196th (MBTD143).
Mant, C. T. and Hodges, R. S. (1991). HPLC of peptides & proteins. U.S.A.: CRC
Press.
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. and Kubelka, V. (1995). Protein glycosylation
in insects . Glycoproteins. 543-563.
122
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.
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.
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 interleukin3. Gene. 58:273-281.
Montgomery, D.C. (1996). Design and Analysis of Experiments. 4th ed. U.S.A.: John
Wiley & Sons, Inc.
123
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.
Montreuil, J., Spik, G. and Mazurier, J. (1997). Transferrin superfamily.
Glycoproteins II. 203-242.
Murphy, C. I. and Young, E. (1996). Baculovirus vectors for expression of secretory
and membrane-bound proteins. (U.S Patent 5,516,657).
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):651660.
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
expressed by
a baculovirus recombinant induces neutralizing antibodies.
Virology 173:631-637
Ohman, L. and Haggstrom, L. (2001). Culture medium for insect cells lacking
glutamine and contains ammonium salt. (U.S Patent 6,210,966).
124
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.
O'Reilly, D., Miller, L. and Luckow, V. (1992). In: Baculovirus expression vectors.
New York: W. H. Freeman and Co. pp109-215.
O’Reilly, D. R., Miller, L. K. and Luckow, V. A. (1994). Baculovirus Expression
Vectors: A Laboratory Manual. U.S.A.: Oxford University Press.
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 F68 direct interaction with insect cells: impact on shear stress protection,
recombinant protein, and baculovirus production. Enzyme Microb. Technol.
26: 324-331.
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.
125
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.
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).
Possee, R. D., Weyer, U. and King, L. A. (1990). Control of Virus Disease.
Cambridge University Press. 53-76.
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.
Reed, L. and Muench, H. (1938). A simple method for estimating fifty percent
endpoints. Am. J. Hyg. 27:493-497.
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.
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.
Rosenberg, I. M. (1996). Protein Analysis and Purification. U.S.A.: Birkhauser.
126
Rudolf, F. B. and Mcintire, L. V. (1999). Science, Engineering, and Ethical
Challenges for the 21st Century. Washington, D.C.: Joseph Henry Press.
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
Shen, C. F., Hawari, J. and Kamen, A. (2004). Micro-quantitation of lipids in serumfree 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.
Siemensma, A., Huttinga, H. and Hakkaart, M. (2002). Eliminating Animal Derived
Media. Gen. Eng. News. Vol. 22.
Steward, M. W. (1984). Antibodies: Their structure and function. London: Chapman
and Hall.
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).
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., Bradfield, J. Y. and Keeley, L. L. (1991). Lepidopteran AKH
signal sequence. (U.S Patent 5,023,328).
127
Summers, M. D., Oker-Blom and Christian, E. G. (1992). Baculovirus dual
promoter expression vector. (U.S Patent 5,169,784).
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.
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).
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.
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 Nglycans 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 Nglycan processing in mammalian cell lines to native and engineered
lepidopteran insect cell lines. Glycoconjugate Journal. 21:343-360.
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.
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.
128
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.
Vander, Sherman and Luciano. (1994). Human Physiology: Basic Cell Functions.
U.S.A.: McGraw Hill.
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.
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.
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.
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.
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).
129
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.
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.
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.
Appendix A1: Kinetics Analysis of Sf9 Insect Cells Growth at Various
Conditions
Sample calculation:
Effect of seeding density of 0.8 x 106 cells/ml
a) Exponential Growth Phase: Data pairs which were assumed to be in the
exponential phase are as follow
X
Y
(hours) (x106cells/ml)
0
0.8
48
1.17
96
3.65
144
6.41
Plot of lnY vs X yields the following graph which shows a linear fashion curve
2
1.5
1
0.5
0
-0.5
0
50
100
150
200
Values for xi, yi, xi2 and xiyi were then calculated and summed up.
n
xi=Xi
yi=lnYi
xi 2
xiyi
1
0
-0.2231
0
0
2
48
0.1598
2304
7.6727
3
96
1.2947
9216
124.2938
4
144
1.8578
20736 267.5317
288
3.0892
32256 399.4983
n
∑
i =1
131
Appendix A1: (continued)
Thus the net growth rate, unet can easily be calculated as follow
n
⎛ n ⎞⎛ n ⎞
n∑ xi yi − ⎜ ∑ xi ⎟⎜ ∑ yi ⎟
⎝ i =1 ⎠⎝ i =1 ⎠ = 4(399.4983) − (288)(3.0892 )
unet = a = i =1
2
n
4(32256) − 2882
⎛ n ⎞
2
n∑ xi − ⎜ ∑ xi ⎟
i =1
⎝ i =1 ⎠
= 0.0154/hour
where doubling time is
τd =
ln 2
µ net
=
0.693
µ net
= 45.01 hour
b) Two-fold Death Phase: Data pairs which were assumed to be in the two-fold death
phase are as follow
X
Y
(hours) (x106cells/ml)
240
5.09
288
2.8
336
1.52
432
0.15
Plot of lnY vs X yields the following graph which shows a linear fashion curve
2
1
0
-1
-2
-3
0
100
200
300
400
500
132
Appendix A1: (continued)
Values for xi, yi, xi2 and xiyi were then calculated and summed up.
n
xi=Xi
yi=lnYi
xi2
xiyi
1
240
1.6273
57600
390.5467
2
288
1.0296
82944
296.5304
3
336
0.4187
112896 140.6867
4
432
-1.8971
186624 -819.556
1296
1.1785
440064 8.2079
n
∑
i =1
Thus the net death rate, unet can easily be calculated as follow
n
⎛ n ⎞⎛ n ⎞
n∑ xi yi − ⎜ ∑ xi ⎟⎜ ∑ yi ⎟
⎝ i =1 ⎠⎝ i =1 ⎠ = 4(8.2079 ) − (1296 )(1.1785 )
unet = a = i =1
2
n
4(440064) − 1296 2
⎛ n ⎞
2
n∑ xi − ⎜ ∑ xi ⎟
i =1
⎝ i =1 ⎠
= -0.0185/hour
where the two-fold death time is
τd =
ln 2
µ net
=
0.693
µ net
= -37.47 hour (–ve means death phase)
133
Appendix A2: Growth kinetic coefficients of Sf9 cells monolayer culture.
Parameters
exp growth
doub time
max density
-exp death
halv time
Studied
u(1/h)
t/2(h)
cells/ml
u(-1/h)
t/2(h)
a) Effects of seeding density in TC100 + FBS
0.05 x 106 viable
cells/ml
0.0164
42.27
1.95E+06
0.0089
77.88
0.15 x 106
0.0162
42.79
2.36E+06
0.0123
56.35
6
0.0152
45.60
2.16E+06
0.0092
75.34
0.30 x 10
b) Effects of seeding density in SF-900 II SFM
0.2 x 106 viable
0.0115
60.27
5.90E+06
0.0046
150.68
0.8 x 106
0.0154
45.01
6.90E+06
0.0185
37.47
6
1.6 x 10
0.0121
57.28
5.45E+06
0.0156
44.43
3.2 x 106
0.0013
533.19
4.35E+06
0.0163
42.52
cells/ml
6
0
0
5.60E+06
0.0067
103.45
TC100 + 5%FBS
0.0107
64.78
1.11E+06
0.0073
94.95
TC100 + 10%FBS
0.0152
45.60
2.16E+06
0.0092
75.34
SF-900 II SFM
0.0154
45.01
6.90E+06
0.0185
37.47
5.6 x 10
c) Effects of medium
d) Effects of different MOIs
5
-
-
-
0.0273
25.39
10
-
-
-
0.0242
28.64
25
-
-
-
0.0277
25.02
50
-
-
-
0.0162
42.79
100
-
-
-
0.0181
38.30
134
Number 0f
Uninfected
Wells
Total
Number
Infected
Total
Number
Uninfected
%
Total
Infected
Above
50%?
%
Above
50%?
%
Below
50%?
Log
1.0E-01
1.0E-02
1.0E-03
1.0E-04
1.0E-05
1.0E-06
1.0E-07
1.0E-08
10
10
10
10
10
5
1
0
0
0
0
0
0
5
9
10
56
46
36
26
16
6
1
0
0
0
0
0
0
5
14
24
100.00
100.00
100.00
100.00
100.00
54.55
6.67
0.00
TRUE
TRUE
TRUE
TRUE
TRUE
TRUE
FALSE
FALSE
0.00
0.00
0.00
0.00
0.00
54.55
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
6.67
0.00
0
0
0
0
0
-6
0
0
Num. Wells
mls/well
10
0.01
54.55
6.67
-6
Prop. Dist.
Log TCID50
TCID50
1/TCID50
TCID50/ml
0.095
-6.095
8.04E-07
1.24E+06
1.24E+08
pfu/ml
8.59E+07
Volume of virus for infection
10
MOI =
=
10
pfu/cell
1.60E+06
cells =
=
1.60E+07
pfu
ml
virus
=
0.186
Dilution
Above 50%
Appendix B1: TCID50 Calculation (spreadsheet)
Number of
Infected wells
Dilutions
135
Appendix B2: Calculation of end point dilution based on Reed and Muench method
Data sample is given below;
Dilution
10-5
10-6
10-7
10-8
Positive rate
10/10
6/10
1/10
0/10
Positive number
10
6
1
0
Negative number
0
4
9
10
Positive total
17
7
1
0
Negative total
0
4
13
23
Positive rate total 17/17
7/17
1/14
0/23
% positive
63.6
7.1
0
100
Proportionate distance, PD = ((%next above 50%) – 50%) / ((%next above 50%) –
(% next below 50%))
= (63.6-50.0) / (63.6-7.1)
Log lower dilution (dilution above 50%) = -6.0
Minus the PD = -0.24
Sum (50% end point) = -6.24
= Log TCID50
TCID50=10-6.24=1/(1.74 x106)
If the virus innoculum was 0.005 ml per well cells, the final expression would be:
(1.74 x106)/0.005
= 3.48 x 108 TCID50 per ml
= 3.48 x 108 TCID50 per ml x 0.69
= 2.4 x 108 pfu/ml
136
Appendix C1: RSM Spreadsheet (Sample Calculation)
1) Increments specification
Choose increment:
dGln
dGluc
dLip
10
11
0.05
2) Calculation of response based on a Taguchi Design array
Operating
point
mimimum
maximum
Range
Mid value
original parameters, with units
Gln
Gluc
Lip
rhTf
response
y
coded
variables,
dimensionless
x1
x2
x3
2211.16
2201.16
2221.16
2221.16
2211.16
2221.16
2211.16
2211.16
2201.16
2201.16
1292.755
1292.75
1303.75
1281.75
1292.75
1292.75
1281.75
1303.75
1303.75
1281.75
0.64
0.64
0.64
0.69
0.69
0.59
0.64
0.59
0.69
0.59
47.33
46.25
46.21
46.07
46.03
46.17
46.29
46.12
45.99
46.20
0.0
-1.0
1.0
1.0
0.0
1.0
0.0
0.0
-1.0
-1.0
0.0
0.0
1.0
-1.0
0.0
0.0
-1.0
1.0
1.0
-1.0
0.0
0.0
0.0
1.0
1.0
-1.0
0.0
-1.0
1.0
-1.0
2201.16
2221.16
20
2211.16
1281.755
1303.755
22
1292.755
0.59
0.69
0.1
0.64
-1
1
2
0
-1
1
2
0
-1
1
2
0
3) Regression Analysis
SUMMARY OUTPUT
Regression Statistics
Multiple R
0.16
R Square
0.03
Adjusted
R
Square
-0.46
Standard Error
0.47
Observations
10.00
ANOVA
Regression
Residual
Total
Df
3
6
9
SS
0.04
1.30
1.34
MS
0.01
0.22
F
0.05
Significance F
0.98
MID
POINT
137
Appendix C1: (continued)
Intercept
X1
X2
X3
Coeff
46.27
0.0020
-0.0399
-0.0658
Std Error
0.15
0.19
0.19
0.19
t Stat
314.37
0.01
-0.21
-0.35
P-value
0.00
0.99
0.84
0.74
4) Vector of steepest ascent
X1
Vector of steepest ascent in coded variables =
magnitude of this vector =
X3
( 0.0020
-0.0399
-0.0658 )
( 0.0260
-0.5188
-0.8545 )
0.0770
Unit vector of steepest ascent in coded variables =
Verify: magnitude of this vector =
X2
1.0000
5) RSM movement towards optimum
Starting point:
Gln
Gluc
Lip
y
2211.16
1292.75
0.64
47.33
2211.18
1292.35
0.64
47.33
2211.20
1291.95
0.64
47.33
2211.21
1291.55
0.64
47.33
2211.23
1291.15
0.63
47.33
2211.25
1290.75
0.63
47.33
2211.27
1290.35
0.63
47.32
<<<<MAXIMUM VALUE
138
Appendix D1: Flowchart of the major steps involved in this research
Lab and
materials set
up
Growth of Sf9
cells
monolayer
Monolayer
growth profile
analysis
Viral
amplification
Optimization
of monolayer
rProtein
expression
profile analysis
Recombinant
protein
expression
Viral titration
Monolayer
optimization
analysis
Virus
screening
Virus
screening
analysis
Medium
screening
Optimization
of suspension
Suspension
growth profile
analysis
Growth of Sf9
cells
suspension
Medium
screening
analysis
Verification of
optimized
expression
Characterization
of optimized
expression
Report and
future works
SD HT MOI TOI
SD MOI TOI Gln Gluc
Lip
Suspension
optimization
analysis
•
SD, seeding density; TOI, time of infection; HT, harvest time; MOI, multiplicity
of infection; Gln, glutamine; Gluc, glucose.