i
OPTIMIZATION OF RECOMBINANT AMYLASE EXPRESSION USING
RESPONSE SURFACE METHODOLOGY (RSM)
KAVITHA A/P MUNIANDY
A dissertation submitted in partial fulfillment of the
requirements for the award of the degree of
Master of Science (Biotechnology)
Faculty of Biosciences and Bioengineering
Universiti Teknologi Malaysia
DECEMBER 2010
iii
Specially dedicated to my beloved parents and brothers
iv
ACKNOWLEDGEMENT
First of all, I would like to express my highest gratitude to my supervisor Dr.
Goh Kian Mau, who willing to be my supervisor and assisted me in various aspects
throughout this project. His endless guidance, advices and dedication really gave me
valuable knowledge in completing this project.
Also, my sincere appreciation to my parents Mr & Mrs Muniandy and other
family members for their inadequate love and moral supports towards me throughout my
entire project. Their continuous positive encouragements mean a lot to me in completing
this study.
Apart from that, I would like to thank all the master students and the lab assistant
in Special Equipment Laboratory for their willingness to assist me and to share with me
their knowledge on this project. Special thanks to Mr. Ummirul Mukminin, Miss Goh
Poh Hong and Miss Chai Yen Yen who had guided me continuously until completing
this project.
Last but not least, my special appreciations to everyone who involve in this study
either directly or indirectly.
v
ABSTRACT
The Anoxybaccilus DT3-1 is a newly found bacterium that is able to express
amylase. The gene that encodes the amylase was recently cloned and expressed in E.
coli system. However, the expression level was far too low to be used. The main
objective of this study is to enhance the recombinant amylase expression level using
pET-22b vector. Another objective of this study is to determine the end product release
by the reaction of this amylase. The media optimization was carried out with five
different media i.e. LB, TB, SB, CDM 1 and CDM 2. Medium LB was found to be the
best medium to support the cell growth and amylase production (72 U/ml). Relevant
factors such as the inducer (IPTG) concentration, yeast extract concentration and
induction time (OD600nm) were optimized through two Response Surface Methodology
(RSM) methods, which were the Two-level factorial and Central Composite Design
(CCD). After the final optimization using CCD, 83 U/ml of amylase activity was
obtained with the optimal condition of 0.007 mM IPTG, 0.3% of yeast extract and
induction should be done when the cells optical density was at 1.52. Upon achieving the
optimal conditions, the end products were determined using High Performance Liquid
Chromatography (HPLC). The amylase was able to degrade various starches like rice,
corn, wheat and soluble starch and produced a wide variety of oligosaccharides such as
the glucose, maltose and isomers of maltose.
vi
ABSTRAK
Anoxybacillus DT3-1 adalah bakteria yang baru ditemui yang mampu
mengekspres amilase. Gen yang mengkod amilase ini telah diklonkan dan diekspres
dalam sistem E. coli. Namun, tahap ekspresi terlalu rendah untuk digunakan. Tujuan
utama projek ini adalah untuk meningkatkan tahap ekspresi amilase rekombinan
menggunakan vektor PET-22b. Selain itu, tujuan lain dalam projek ini adalah untuk
menentukan penghasilan produk akhir dari reaksi amilase ini. Optimasi media dilakukan
dengan menggunakan lima media yang berbeza iaitu LB, TB, SB, CDM 1 dan CDM 2.
Didapati, media LB merupakan media yang terbaik menengah untuk pertumbuhan sel
dan pengeluaran amilase (72 U / ml). Faktor lain yang relevan seperti kepekatan induser
(IPTG), kepekatan ekstrak ragi dan masa induksi (OD600nm) telah dioptimumkan
melalui dua kaedah ‘Response Surface Methodology (RSM)’ iaitu ‘Two-level Factorial’
dan ‘Central Composite Design (CCD)’. Setelah pengoptimuman terakhir menggunakan
CCD, 83 U/ml aktiviti amylase dapat diperoleh dengan keadaan optimum IPTG 0.007
mM, 0.3% ekstrak ragi dan induksi harus dilakukan ketika sel ketumpatan optik berada
pada 1.52. Setelah mencapai keadaan yang optimum, penghasilan produk akhir
ditentukan dengan menggunakan Kromatografi Cair Kinerja Tinggi (HPLC). Amilase
tersebut mampu mendegradasi pelbagai jenis kanji dari beras, jagung, keladi dan
gandum di mana ia dapat menghasilkan pelbagai oligosakarida seperti maltose, glukosa
dan isomer maltose yang lain.
vii
TABLE OF CONTENTS
CHAPTER
1
TITLE
PAGE
TITLE
i
DECLARATION
ii
DEDICATION
iii
ACKNOWLEDGEMENTS
iv
ABSTRACT
v
ABSTRAK
vi
TABLE OF CONTENTS
vii
LIST OF TABLES
x
LIST OF FIGURES
xi
LIST OF ABBREVIATIONS
xiii
LIST OF APPENDICES
xiv
INTRODUCTION
1
1.1
Introduction
1
1.2
Problem Statement
3
1.3
Objectives
3
viii
1.4
2
Scopes of the research
4
LITERATURE REVIEW
5
2.1
Starch
5
2.2
Thermostable Bacteria
7
2.2.1
Thermostable enzyme
8
2.2.2
Applications of Thermophilic
9
Enzymes
2.3
2.4
3
!-amylase ( amylase family and characteristic)
11
2.3.1
12
Reaction Mechanisms of Amylase
Design of Experiment
15
2.4.1
Response Surface Methodology (RSM)
16
2.4.2
Central Composite Design (CCD)
17
MATERIALS AND METHODS
19
3.1
Preparation of Bacterial Stock
19
3.2
Bacteria Revival and Culture
21
3.3
General Media Optimization
21
3.3.1
Composition of Each Medium
22
3.3.2
Overnight culture preparation
23
3.3.3
Optimization and Expression of E.coli BL21 24
Carrying Amylase in pET 22-b Vector
3.4
3.5
3.6
Further Optimization of Media
25
3.4.1
Intracellular Enzyme Extraction
25
3.4.2
Enzyme Assay
26
Optimization of Other Factors Using DoE
27
3.5.1
30
Analysis
End Product Analysis
31
3.6.1
31
High Performance Liquid Chromatography
(HPLC)
ix
4
RESULTS AND DISCUSSIONS
4.1
33
Optimization and Expression of E.coli BL21 Carrying 33
Amylase in pET 22-b Vector
4.2
4.3
4.4
Further Comparison of CDM 2 and LB media
37
4.2.1
Cell Growth Profiles
37
4.2.2
Enzyme Activity
38
Optimization of Relevant Factors Using DoE
44
4.3.1
Adequacy of The Model
49
4.3.2
Optimal Design from Two-level Factorial
52
Expression Optimization Using Central Composite
54
Design (CCD)
4.4.1
Selection and Validation For Significant
54
Effect
4.4.2
Analysis of Variance (ANOVA)
60
4.4.3
Model Validation
63
4.4.3.1
Normal Probability Plot
63
4.4.3.2
Residual Versus Predicted Plot
64
4.4.3.3
Outlier T Plot
65
4.4.3.4
Box-Cox Plot
66
4.4.4
4.5
5
Optimal Design Based on CCD
67
End Product Analysis
69
4.5.1
69
High Performance Liquid Chromatography
CONCLUSION
72
5.1
Conclusion
72
5.2
Future Work
74
REFERENCES
75
APPENDICES
80
x
LIST OF TABLES
TABLE NO.
TITLE
PAGE
2.1
Thermophiles and their common habitat
8
2.2
Applications of the thermophilic enzymes
10
3.1
Experimental factors for Two-level Factorial
28
3.2
Experimental factors and levels for Two-level Factorial
29
3.3
Experiment factors for CCD
29
3.4
Experimental factors and levels for second CCD
30
4.1
Process parameters and their levels for Two-level
45
Factorial
4.2
Experiment factors and responses
46
4.3
Model and coded factors
47
4.4
Comparison between actual values and predicted values
49
4.5
ANOVA analysis for extracellular activity
51
4.6
ANOVA analysis for intracellular activity
51
4.7
Process parameters and their levels for CCD
56
4.8
Experiment factors and responses for amylase activity
58
4.9
ANOVA for amylase activity
60
4.10
Model and coded factor of CCD
62
4.11
Comparison between enzyme at unoptimized
68
condition, Two-level factorial and CCD
xi
LIST OF FIGURES
FIGURE NO.
TITLE
PAGE
2.1
Structures of starch
7
2.2
Different enzymes involve in starch degradation
13
2.3
The double displacement mechanisms
14
3.1
Overview of total work flow
20
4.1
Profiles of microbial growth in five media
35
4.2
Fuwa assay on three final intervals samples in all five
35
media
4.3
Profiles of microbial growth in LB broth and CDM 2
37
for 68 hours at 25˚C
4.4
The extracellular and intracellular amylase activity
39
in LB medium for 68 Hours
4.5
The extracellular and intracellular amylase activity in
41
CDM 2 medium for 68 hours
4.6
The comparison of extracellular amylase activity
43
between LB and CDM 2 media for 68 hours
4.7
The comparison of intracellular amylase activity
43
between LB and CDM 2 media for 68 hours
4.8
Predicted vs. actual data for extracellular activity
47
xii
4.9
Predicted vs. actual data for intracellular activity
48
4.10
Ramp of extracellular amylase
52
4.11
Ramp of intracellular amylase
53
4.12
Half-normal plot of two-level factorial (extracellular
55
activity)
4.13
Response surface of first CCD design
57
4.14
Contour plots and response surfaces for the
59
effect of IPTG concentration
4.15
Predicted versus actual in CCD
62
4.16
Normal plot of residual for amylase production
64
in second CCD
4.17
Residual versus predicted plot for amylase
65
expression
4.18
Plot of Outlier T of amylase production
66
4.19
Box-Cox plot for generated model of amylase
67
expression
4.20
Ramps of various factors in CCD
68
4.21
Chromatogram of separation of standards
69
(oligosaccharides)
4.22
Chromatogram of separation of sugar components
for various kind of starch after degraded
with recombinant amylase
70
xiii
LIST OF ABBREVIATIONS
ANOVA
Analysis of Variance
CaCl2
Calcium chloride
CCD
Central Composite Design
CDM 1
Chemically Defined Medium 1
CDM 2
Chemically Defined Mediun 2
DNA
Deoxyribonucleic acid
DoE
Design of Experiment
E.coli
Escherichia coli
HPLC
High Performance Liquid Chromatography
LB Broth
Luria-bertani broth
mL
Mililiter
NaOH
Sodium hydroxide
OFAT
One Factor At Time
PCR
Polymerase chain reaction
rpm
rotary per minute
RSM
Response Surface Methodology
TAE
Tris-acecate-EDTA
Tris
Tris (hydroxymethyl) aminomethane
xiv
LIST OF APPENDICES
APPENDIX
A1
TITLE
Process parameters and responses
PAGE
80
(amylase activity) for first CCD
A2
ANOVA analysis of first CCD
A3
Selected Model Validation Analysis for First CCD 82
A4
Optimal Design Based on First CCD
B1
Sugar Separation of Various Kind of Starch After 85
Degraded with Recombinant Amylase
81
84
1
CHAPTER 1
INTRODUCTION
1.1
Introduction
Enzymes are biological catalyst that reduces the activation energy of a reaction
by providing an alternative pathway for the reaction to occur. These enzymes are the key
component in many industries that revolve around biotechnology industries. This is
mainly due to the ability of enzymes to convert the substrates to desired product with
minimal conditions at relatively lesser time and money (Gupta et al., 2003).
Starch degrading enzymes such as amylases have been in high amount for their
industrial benefits. Amylases (1, 4-!-D-Glucan glucanohydrolase) are enzymes that
hydrolyze starch molecules into smaller compounds such as oligosaccharides and
dextrins. Amylases are also able to hydrolyse starch to the very basic sugar component
which is glucose. Amylases are one of the most important enzymes in many industries
2
such as food, textiles and paper industries. Recent discoveries also reveal that amylases
have potential useful in pharmaceutical industry as well if amylases are prepared with
suitable properties (Hmidet et al., 2008).
There are many sources of amylase which varied from animal to plant and can be
find vastly in microorganisms. Current mode in industries requires the usage of
microorganisms as biotechnological sources of industrially relevant enzymes. This is
because, microbial enzymes are significantly more economical and environmental
friendly compared to chemicals. The major advantages of using microorganisms for the
production of amylases is the cost effective bulk production capacity, less time and
space required for production and microbes are relatively easy to manipulate to obtain
enzymes of desired characteristics (Pandey et al., 2000, Gupta et al., 2003 and Asgher et
al., 2007).
Due to the increasing demand for amylase enzymes in various industries, there
is enormous interest in developing enzymes with novel properties such as raw starch
degrading amylases suitable for industrial applications and their cost effective
production techniques. This increases the discovery and researches on the exploration of
extracellular enzymatic activity in several microorganisms (Gupta et al., 2003).
The classical procedure in revelation of novel species that able to produce useful
enzyme is the isolation of microbial species. By using this method, it is able to produce
novel enzymes from uniquely extreme environments such as extreme temperature or
extreme pH environment. This would also able to offer a competitive advantage over the
existing products which are more common. Eventually, characterization of these novel
extreme environmental enzymes under fermentation conditions to optimize the enzyme
production properties plays crucial role in evaluation of their industrial and economic
significance (Prakasham et al., 2007).
3
One of these novel discoveries are the founding of newly emerge species of
Anoxybaccilus from one of the hot-spring in Malaysia known as the Dusun Tua hotspring. The Anoxybaccilus which are currently named as the Anoxybaccilus DT3-1 are
found by the research team of Universiti Teknologi Malaysia (UTM), where this
microorganism is able to produce thermostable amylase enzyme.
1.2
Problem Statement
Optimizing the best recombinant amylase production is important as this is a
novel enzyme from a newly found thermostable microorganism. Once the highest
enzyme expression through optimal condition able to obtain, the industrial value for this
enzyme will increased. This ultimately provides an alternative to currently available
enzymes with less expenditures and high productivity.
1.3
Objectives
1. To determine the best media for amylase expression.
2. To optimize relevant factors that involve in amylase expression such as
absorbance value and induction time through Two-level factorial and Central
Composite Design (CCD)
3. End product determination using HPLC
4
1.4
Scopes of the research
The scopes of research are as follow:
a) General optimization of the best media
b) Precise optimization of the best media
c) Enzyme assay, protein assay and localization of cell
d) Other factors optimization using 2-level factorial
e) Further optimization using Central Composite Design (CCD)
f) High Performance Liquid Chromatography (HPLC)
5
CHAPTER 2
LITERATURE REVIEW
2.1
Starch
Starch-producing crops are economically important in most countries around the
world as the large consumption of daily food include starch component. Even though,
there are many plants that able to produce starch, only a number of these plants are
economically useful. The main crops that consist of high economical value of starch
storage include rice, maize, potato and tapioca.
Starch is the compound that is synthesized by green plants during
photosynthesis. The process of synthesizing starch commonly occurs in plastids of
leaves and also in amyloplasts that can be found in seeds and roots. The synthesis in
leaves normally take place to fulfill the short term usage while in amyloplast, starch are
synthesis to store for longer period of time. In order for longer period of storage, starch
in amyloplasts accumulates as water-insoluble granules (Maarel et al., 1994).
6
In green plants, starch is comprised of approximately 25% of amylose and the
remaining is amylopectin. Amylose is a linear structure of large amount of glucose
molecules that is join by !-1,4-glucosidic linkages. Whereas, amylopectin is compactly
branched glucose molecules that attached with linear glucose polymers in branch form
through !-1,6-glucosidic bond (Nakamura, 2002).
Starch had been used diversely in many industries. However, the main
application of starch is in food industry, where it had been used as food thickening such
as sauces and puddings. Other uses of starch in food industry include using starch as fat
replacer, as glazing agent, stabilizer and also as emulsifier. This is mainly done through
chemically or by using enzyme to harvest the starch to produce other derivatives of
starch such as fructose or cyclodextrin (Gupta et al., 2003)
Besides that, starch is also used as glue in wallpapers, stamps and envelopes. In
paper industry, starch is used to make paper stronger. Other then that, starch can also
been used as filing agent in pharmaceutical products such as in tablets and to increase
the moisture absorption in baby diapers (Maarel et al., 1994).
7
Figure 2.1
Structure of starch showing the linear linkages of amylose and the bridge
linkage form to known as amylopectin (http://www.food-info.net/uk/carbs/starch.htm)
2.2
Thermostable bacteria
Thermostable bacteria are bacteria that can withstand high temperature.
Generally, this type of bacteria can be found naturally in geothermal heated places such
as hot springs and deep sea hydrothermal vents. Thermophilic bacteria can only be
prokaryotic (Hobel, 2004). There are two groups of thermophiles which are the obligate
and facultative thermophiles. Obligate thermophiles or also known as the extreme
thermophiles are bacteria and archaea that are able to grow at temperature as high as
110˚C (Huber et al., 2000).
8
There are many sources where thermophilic bacteria can be found. According to
Hobel (2004), natural geothermal areas are vastly found around the world, but these
locations are primarily associated with tectonically active zones where the movements of
the Earth’s surface occur. This makes the geothermal heat sources restricted to few
concentrated regions. Terrestrial hot springs are one of the common habitats of
thermophilic bacteria. This terrestrial hot springs are classified to the nature of the heat
source and pH. Table 2.1 shows few examples of thermophiles and their common
habitat.
Table 2.1: Thermophiles and their common habitat
Microorganisms
Sources
References
Bacillus sp.
Sediments of hot spring Badal et al. (1989), Mamo
and hot springs
and Gessese (1999)
Bacillus stearothermophiles
Compost
Kenji et al. (1989)
Bacillus circulans
Garbage dump
Ashita et al. (2000)
Thermus sp.
Hot springs
Shaw et al. (1995)
2.2.1
Thermostable enzyme
Generally, in order to have optimum survival mode, thermophilic bacteria
consists of thermophilic enzymes where these enzymes are able to function well at high
temperature. Thermophilic enzymes have the ability to maintain their three dimensional
structure by tighten their structure folding at higher temperature (Niehaus et al., 1999).
9
In order to make their structure and function optimal at higher temperature, these
bacteria have several thermostable enzymes with different mechanisms of enzyme
thermostabilization. These mechanisms include increase in number of hydrogen bonds,
hydrophobic residues and higher mode of stabilization (Saboto et al., 1999). Apart from
that, these bacteria also increase unique interactions such as electrostatic, disulphide
bridges and hydrophobic interaction to maintain their functions in high temperature
(Kumar and Nussinov, 2002).
Besides that, the formation of cell membrane which consists of saturated fatty
acids also succor in maintaining the optimal shape and function of thermophilic bacteria.
This is because fatty acids contribute high hydrophobic environment for the cell and
retain the rigidity of the cell to live in high temperature (Herbert and Sharp, 1992).
Thermophilic bacteria also are able to withstand high temperature in the cell by having
low grade of protoplasmic organization in their cell structure (Gaughran).
2.2.2
Applications of thermophilic Enzyme
Thermophilic enzymes have wide applications in various industries. The
application of thermophilic enzymes varies from food industries to waste management.
However, the first breakthrough of commercially useable thermophilic enzymes was in
1946 where the corn syrup manufacturing was patented by Dale and Langois (Hamid et
al., 2003).
10
A prime commercial application of thermophilic enzymes is the use of the
enzyme as DNA polymerase in Polymerase Chain Reaction (PCR). The most common
enzyme used in this process is the Thermus aquaticus which is known as Taq
polymerase (Saiki et al., 1998). This enzyme was isolated from a hot spring bacterium
and been expressed as recombinant.
Apart from that, the other applications of thermophilic enzymes include
maximizing the reactions in food and paper industry, toxic waste removal and detergents
(Haki and Rakshit, 2003). Other applications of these thermophilic enzymes are
summarized in Table 2.2.
Table 2.2 : Applications of thermophilic enzymes (Haki and Rakshit, 2003)
Enzymes
Bioconversion
Applications
Amylase
Starch
Baking, brewing, starch hydrolysis, digestion
Sugar
in
milk,
saccharifying
enzymes
and
oligosaccharides
Pullulanase
Starch
Production of glucose syrups
dextrose syrups
Lipase
Fat removal, alcholysis Detergent, pharmaceuticals, waste water
and aminolysis
treatments,
oleo-chemical,
leather
and
cosmetics industry
Xylanases
Craft pulp
xylan + Pulp and paper industry including paper
lignin
Proteases
bleaching
Protein
amino acids Baking, brewing, food processing and leather
industry
DNA
DNA
restriction
polymerase
amplification
Cellulase
Cellulose
and Genetic engineering
glucose
Cellulose hydrolysis
11
2.3
!-amylase ( amylase family and characteristic)
The !-amylase family consists of a large group of starch hydrolases and related
enzymes comprising about 20 different enzyme specificities, and is currently known as
glycosyl hydrolase family 13. Enzymes in this family are multi-domain proteins that
commonly form the barrel shape of (!/")8 (Stefan et al., 1997 and Mc Gregor, 2001).
Amylases (1,4-!-D-Glucan glucanohydrolase) are one of the starch degrading
enzymes that hydrolyze starch molecules into smaller compounds such as
oligosaccharides and dextrins. Amylases are also able to hydrolyze starch to the very
basic monomer component which is glucose. The !-amylase family comprises a group
of enzymes with a variety of different specificities that all act on one type of substrate
being glucose residues linked through an !-1-1, !-1-4, !-1-6, glycosidic bonds. Amylase
generally hydrolyzes glycosidic bond in starch molecules converting it to a simple sugar
unit. The process of glycosidic hydrolysis that might occur is represented by the
hydrolysis of 1,6 glycosidic bond, hydrolysis of 1,4 glycosidic bond, transglycosylation
to form to different glycosidic bond (Kuriki, T. and Imanaka, 1999).
Amylases are first discovered in 1811 and being researched until today. One of
the major research by Ohlsson (1930), the initial proposal classified amylase into !amylase and "-amylase according to the anomeric type of sugar produces by the
enzymatic reaction. Apart from that classification, there are two main categories of
amylases which are endoamylases and exoamylases. The function of both these
categories are different where endoamylases catalyse random hydrolysis in the interior
side of starch resulting in linear and branched oligosaccharides in various length
whereas exoamylases hydrolyses at the non-reducing end of starch thus produced short
end products (Gupta et al., 2003).
12
Even though amylases can be express by many types of bacteria, it is very
challenging to obtain a suitable strain that is capable of producing commercially
acceptable yields. Choosing the suitable strain is one of the most important factors in
amylase production. Each application of amylases requires unique properties with
respect to specificity, stability, temperature and pH dependence. Screening of
microorganisms with higher amylase activities could therefore, facilitate the discovery
of novel amylases suitable to new industrial applications (Pandey et al., 2000 and
Asgher et al., 2007).
2.3.1
Reaction mechanism of amylase
According to Maarel et al. (2002), there are four groups of starch-converting
enzymes: (i) endoamylases; (ii) exoamylases;(iii) debranching enzymes; and (iv)
transferases. However, amylase can be both endoamylase and exoamylase.
Endoamylases are enzymes that are able to cleave !-1,4-glycosidic bonds that
appear in inside an amylose and amylopectin bond. Enzyme that is commonly known as
the endoamylase is !-amylase, which can be found both in bacteria and Archae (Pandey,
2000). !-amylase catalyses the degrading of starch that resulted in final products of
oligosaccharides in multiple lengths.
In the other hand, exoamylases such as "-amylases are enzymes that cleave at
external residues of glucose of amylose and amylopectin. This make the end product of
exoamylases reaction is oly glucose molecules. Similar to endoamylases, the "-amylase
cleaves the polysaccharides at !-1,4-glycosidic bond (Pandey et al., 2000).
13
Figure 2.2
Different enzymes that involve in degradation of starch (Maarel et al.,
2002)
All enzymes in the amylase super-family have the similar catalytic mechanism,
which is derived through the same catalytic residues. The common mechanism known
for amylase catalytic activity is double displacement reaction (Maarel et al., 2002). This
mechanism involves two catalytic residues in the active site such as a glutamic acid as
catalyst and an aspartate as nucleophile.
Generally, the double displacement reaction is made up of five continuous steps.
First, when a substrate binds to the active site, transfering of proton to the glycosidic
bond oxygen will take place. This will be followed by the formation of a covalent
intermediate when an oxocarbonium ion-like transition state is formed (Maarel et al.,
2002).
14
Then, the protonated glucose molecules leave the active site while a water
molecule substitutes the place in the active site. This will attack the covalent bond
between the glucose molecule and the nucleophile such as aspartate. This will be
followed by again the formation of the oxocarbonium ion-like transition state.
Finally, a base catalyst such as glutamate will accept a hydrogen from incoming
water molecule. The oxygen molecules from the water molecule will then replaces the
oxocarbonium bond and forms a new hydroxyl group at C1 position of the glucose
(Maarel et al., 2002).
Figure 2.3
The double displacement mechanism and the formation of a covalent
intermediate by which retaining glycosylhydrolases act (Maarel et al., 1994)
15
2.4
Design of Experiment (DoE)
In order to optimize the design of a new discovery, it is vital to identify which
factors that have greatest influences and value where optimal production is. The
common approach of designing the optimization is by doing trial-and-error using
conventional one factor at a time (OFAT) strategy. Nevertheless, this classic approach is
very time consuming and not feasible in large scale of activity. To overcome this,
experiments with multiple factors is a better approach that can be used. This multiple
factors experiments are commonly known as the factorial design (Mohammadi et al.,
2004).
All DoE softwares rely on statistical method which is used to probe the outcome
of the experiments and also to determine differences in variation each factors contribute.
An example of such statistical analysis is the Analysis of Variance (ANOVA). In the
classical ‘one-factor-at-a-time’ method, one independent variable will be studied while
the values of other variables (factors) are kept constant. This is an extremely time
consuming and also unreliable method. Apart from that, this method also does not
guarantee the determination of optimal conditions since experiments are done on each
factor which makes it not possible to identify the correct interactions that occur between
the factors (Li et al., 2006). However, in the DoE approach, the use of ANOVA analysis
will be able to overcome such weaknesses.
In order to avoid these complicities, the statistical method which is known as
Response Surface Methodology (RSM) is a better choice. RSM uses quantitative data
from conducted experiments to ascertain and concurrently decipher multivariate
equations. RSM also can be described as a collection of statistical techniques for
designing experiments, model building and analyzing optimum conditions of factors for
desirable responses (Li et al., 2006).
16
2.4.1
Response Surface Methodology (RSM)
Response surface methodology (RSM) was first described by Box and Wilson
(1951). This is an experimental strategy for seeking the optimum condition for a
multivariable system. RSM is a concise way of describing and predicting response of a
system of variables (Murphy, 1977). Moreover, it defines the effect of the independent
variables, alone or in combinations, on the process and generates a mathematical model
that accurately describes the overall process.
Gohel et al. (2007) described, RSM offers advantages through which one can
understand and correlate the effect of the nutrient at varying concentrations and give
significant reduction in total number of experiments resulting in saving time, glassware
and chemicals. RSM had been successfully employed for the optimizing medium
ingredients and operating in many bioprocesses (Lee, 2002).
In most RSM related limitations, the form of relationship between the response
and the independent variables is unknown. Therefore, the preliminary step in RSM is to
find a suitable approximation for the true functional relationship between the response
and independent variable. Following this, the RSM can be performed using the fitted
surface. A regression design is normally employed to model a response as a
mathematical function of a few continuous factors and best model parameter estimates
are delivered (Montmogery, 1997).
RSM is regularly used to build models for making predictions. Thus, the
prediction variance is considerable important in evaluating or comparing between
designs and within a design. Two-dimensional contour plots as well as the threedimensional response surface plots of prediction variance provide a good profile of the
17
prediction variance in the total experimental region. Following the software package
calculation, the optimization of process parameters can be determined.
A successful experimental factorial design and RSM was already applied in
various fields and it is well suited with the study of the main and interaction of the factor
in bioconversion yield. At a basic biological level, recent studies have indicated the use
of RSM for analyzing effects of different factors on proteolytic activity and optimization
of zylanase production. This study is an attempt to evaluate the effects of several factors
on the production of an industrial amylase (Nawani, 2004).
2.4.2
Central Composite Design (CCD)
Central Composite Design (CCD) is one of the approaches that use to build a
second-order response surface model. This is due to the ability of CCD that can be run
sequentially where the first subset estimates linear and two-factor interaction effect
while the second subset estimates curvature effects. This makes that the second subset is
not needed if the data from the first subset denote the absence of significant curvature
effects (Montgomery, 1997).
The CCD model was first described by Box and Wilson in year 1951. Since then,
CCD had been the highest frequency usage model and highly recommended under the
RSM design. Therefore, the CCD is selected in this study to allow determination of
levels of various parameters to be carried out with the interrelation of levels of variance
parameters evolved concurrently (Montgomery, 1997).
18
The design of CCD is very efficient due to the dense information provided by the
design on effect of experiment variables and total experimental error in a minimum
numbers of required laboratory work. CCD is also a robust and flexible design where the
availability of few varieties of CCD enables their function under various experimental
regions and operability.
In general, a CCD design starts with a factorial design with centre points that are
augmented with a group of star points and centre points where these points allows the
estimation of curvature. Precisely, a CCD model requires five coded levels of each
factors which are the plus or minus one (factorial points), plus or minus alpha (axial
points) and the all zero level known as centre points.
19
CHAPTER 3
MATERIALS AND METHODS
3.1
Preparation of bacterial stock
The desired enzyme which is !-amylase from Anoxybacillus sp. DT3-1 was
obtained from another post graduate student. The full length gene which encode the
protein has been cloned in expression system E.coli BL21 using pET 22b (Novagen) as
the vector (unpublished data).
Stock culture was prepared by inoculating a single colony from the given E.coli
strain and incubated overnight at 37 ˚C. The following day, 700 #L of the overnight
culture was pipetted into a microcentrifuge tube and then added with 300 #L of 80%
glycerol. The mixture was vortexed vigorously and stored in -80 °C for long term
storage.
20
!"#"$%&'(")*%'+,-(*.%-+#'/0'(")*%1''
2$"3*4")'(")*%'+,-(*.%-+#'/5'(")*%1''
678$%3"&&9&%$'%#)'*#8$%3"&&9&%$'3"&&4'+:8%*#*#;'
6#.<("'%3-=*8<')"8"$(*#%-+#'/>9?%1'
@?+A&"="&'B%38+$*%&'
C8%-4-3%&'%#%&<4*4'
D"#8$%&'D+(,+4*8"'E"4*;#'/DDE1'
C8%-4-3%&'%#%&<4*4'
6#)',$+)938')"8"$(*#%-+#F'G2HD'''
Figure 3.1
Overview of the total work flow
21
3.2
Bacteria revival and culture
Amylase in E.coli BL21 was revived from glycerol stock at -80°C. Streaking was
done on an agar plate and spread plate technique was used to plate the bacteria. It was
then incubated for 24 hours at 37°C and stored at 4°C as stock. Agar plates were then
incubated for 24 hours at 37°C and stored at 4°C as stock.
3.3
General Media optimization
As shown in Figure 3.1, the main stage in this experiment starts with optimizing
the best media for amylase expression. In order to accomplish this, the expression of !amylase in variety of media was done. The selected media were Luria-Bertani Broth
(LB), Terrific Broth (TB), Super Broth (SB), Chemically Defined Media 1 (CDM1)
which is composed of glucose and Chemically Defined Media 2 (CDM2). The
preparation and composition of each medium are shown as the followings.
Each medium were added with ampicilin as the antibiotic. Ampicillin
(25mg/mL) was prepared by dissolving 1.25g of powdered ampicillin in 50 mL of
distilled water. The solution was then filter sterilized using a 0.22 #M syringe filter.
22
3.3.1
Composition of each medium
a) Luria-Bertani Broth (LB)
To prepare 500ml of this media, 5.0g of tryptone, 2.5g of yeast extract and 5.0g
of sodium chloride were added into 500ml of distilled water and stirred until the powder
fully dissolved. The solution was autoclaved and kept at 4°C.
b) Terrific Broth (TB)
To prepare 500ml of this media, 6.0g of tryptone, 12.0g of yeast extract and 2
mL of glycerol were added into 400ml of distilled water and stirred until the powder
fully dissolved. The solution was then adjusted to 450mL and autoclaved. Upon
autoclaved, the broth was adjusted to 500mL and pH 6.0 by adding filter sterilized
0.17M KH2PO4 and 0.72M K2HPO4. The solution was kept at 4°C for further use.
c) Super Broth (SB)
To prepare 500ml of this media, 16.0g of tryptone, 10.0g of yeast extract, 2.5g of
sodium chloride and 2.5 mL of 1.0M NaOH were added into 500ml of distilled water
and stirred until the powder fully dissolved. The solution was autoclaved and kept at
4°C.
23
d) Chemically defined media 1
To prepare 500ml of this media, 10g of peptone, 2.5g of soluble starch, 1.5g of
K2HPO4 and 0.5g of MgSO4.7H2O were added into 500ml of distilled water and stirred
until the powder fully dissolved. The solution was autoclaved and kept at 4°C (Nikerel
et al., 2006)
e) Chemically defined media 2
To prepare 500ml of this media, 4.75g of KH2PO4, 1.5g of (NH4)2HPO4 and were
added into 400ml of distilled water and stirred until the powder fully dissolved. The
solution was then autoclaved. Upon autoclaved, 7.5g of filter sterilized glucose solution
and 12.5g of filter sterilized MgSO4.7H2O solution was added. The final volume was
then adjusted to 500 mL by adding distilled water. The solution was kept at 4°C for
further use (Tabandeh et al., 2008).
3.3.2
Overnight culture preparation
A single colony from the stored plate was inoculated and cultured overnight at
room temperature with shaking in a flask that contained 100 mL of LB broth and
100mg/ml ampicillin. The next day, 10% of the overnight culture which was 10 mL, was
poured into a centrifuge tube and centrifuged at room temperature for 15 minutes at
4000 rpm.
24
After centrifugation, supernatant was discarded and immediately after that, the
cell at the bottom of the tube was resuspended in fresh LB broth containing ampicillin.
The cell was totally resuspended in the fresh broth by pipetting the solution.
This solution was then used as inoculum for the cell growth. The inoculum was
mixed into a flask that contained 100 mL of fresh LB broth with ampicillin. The mixture
was then left to grow in a shaker with 200 rpm at room temperature until the absorbance
reading of OD600nm of the culture reached the value of 1.
3.3.3
Optimization and Expression of E.coli BL21 Carrying Amylase in pET 22Vector
Five flasks were prepared with each containing 100 mL of selected five
different media of LB broth, SB broth, TB broth, Chemically Defined Media 1 and 2.
Upon reaching the absorbance value, OD 600 of 1, 10mL of the culture was transferred
into each of the flasks. Then, 1 mL of IPTG (100mM) was added into each flask. The
flasks were then incubated at 25˚C with 200 rpm of shaking.
The flasks were continued to incubate with shaking for 30 continuous hours.
The reading of microbial optical density (measured at 600nm) and 4 mL of culture
samples was collected every 2-6 hours for 30 hours.
25
The samples of cultures were harvested at 4000 rpm for 15 minutes. The
supernatant was kept at -20˚C while the cells were kept at 4˚C for further use and
analysis. Besides that, the cell growth profile was plotted and the best growth medium
was determined.
3.4
Further Optimization of Media
A more precise media optimization was done to select the best media for amylase
expression between LB broth and CDM 2 which contain glucose. This precise
optimization was done by profiling the cell activity and its expression for 68 hours.
The overnight culture preparation and IPTG induction was done similarly to the
previous general optimization method. At each time interval, optical density was read at
absorbance 600nm and cell was centrifuged. The supernatant and cell pellet was stored
for further experiments on enzyme assay and protein assay.
3.4.1
Intracellular enzyme extraction
Cell wall of the stored pellet was broken in order to extract the intracellular
content of the harvested cell at each time intervals. Bacterial cells were centrifuged for
10 minutes at 4000 rpm to obtain the bacterial pellet.
26
After centrifugation, supernatant was frozen for further use. 4 mL of B-PER
Bacterial Protein Extraction Reagent (Thermo Scientific) were added for every 1 gram
of cell pellet. Together with that, 2 #L of lysozyme and 2 #L of DNase were added for
every 1 mL of the B-PER Reagent. The suspension was pipetted up and down until it
was fully homogeneous.
The solution was then left at room temperature for approximately 15 minutes.
Then, the solution was centrifuged for 10 minutes at 6500 rpm to separate the soluble
proteins from the insoluble proteins. The supernatant was then frozen for further analysis
and the pellet which is known as the inclusion body was discarded.
3.4.2
Enzyme assay
The analysis of amylase activity using FUWA assay was done according to
Goyal et al. (1995) with slight modifications. The iodine reagent for this analysis was
prepared by adding 0.2% of Iodine with 2% Potassium Iodide and the mixture was
dissolved in distilled water. The reagent mixture was then kept in a Scott bottle that had
pre-wrapped using aluminium foil to avoid exposure to light.
250 #L of enzyme was added with 250#L of 0.2% of soluble starch which was
dissolved in 0.1M phosphate buffer (pH 6.5). The combination of this solution was
incubated at 60˚C for 30 minutes. Immediately after incubation, 250 #L of 1M HCL was
added to the solution to stop the reaction between the enzyme and substrate. This is
followed by the addition of 250 #L of 0.2% Iodine in 2% Pottasium Iodide and also
4mL of distilled water.
27
The changes of colour in the solution were then measured by reading the
absorbance through spectrophotometry at 690nm wavelength. Enzyme activity was then
calculated according to following equation where blank was unreacted starch where 1
unit of activity was defined as the amount of amylase needed to reduce the colour of
starch-iodine compound for 1 %.
Activity (units/mL) =
OD790 (blank) – OD790 (sample)
____________________________
OD790 (blank)
X 100
3.5 Optimization of other factors using Design of Experiment
The Design of Experiment (DoE) method is an effective technique to optimize few
factors at a same time. The factors that taken into consideration in this project were the
optical density absorbance value when IPTG was induced, the percentage of yeast
extract used in the LB broth and also the amount of IPTG induced in each flask.
Through this method, the optimal conditions able to be obtained by conducting lesser
experiments compared to the classical one-factor-at time method.
28
RSM uses an experimental design such as the 2-level factorial and Central
composite design (CCD) to fit a model by least square technique. Adequacy of the
proposed model is then revealed using the diagnostic checking tests provided by analysis
of variance (ANOVA). The factors, the range and the levels of the variables investigated
in this project were summarized in Table 3.1 and Table 3.2. These parameters and their
range were selected based on various journals and publications.
The Design Expert Version 6.0.4 software was used to develop the experimental
plan for both TWO-level factorial and CCD. The software was also used to analyze the
data collected by performing ANOVA studies. The 2-level factorial was designed with
three factors whereas the CCD was designed with two significant factors chose from the
2-level factorial.
Table 3.1: Experimental factors for Two-level factorial
Factors
Low level (-1)
High level (+1)
A) Absorbance (OD)
0.3
1.5
B) Yeast extract (%)
0.05
0.3
C) IPTG concentration (mM)
0.01
0.1
29
Table 3.2 : Experiment factors and levels for Two-level Factorial
Run
A(%)
B(mM)
C(OD)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0.05
0.30
0.05
0.30
0.05
0.30
0.05
0.30
0.17
0.17
0.17
0.17
0.17
0.17
0.01
0.01
0.10
0.10
0.01
0.01
0.10
0.10
0.06
0.06
0.06
0.06
0.06
0.06
0.30
0.30
0.30
0.30
1.50
1.50
1.50
1.50
0.90
0.90
0.90
0.90
0.90
0.90
From Two-level Factorial, the significant factors and their best ranges were
chosen to proceed with Central Composite Design (CCD). Optimization using CCD was
done twice with different ranges in each factor. In the first CCD, as shown in Table 3.3,
the ranges of each factor were set according to the result from Two-level factorial and
these ranges were set at –! and +! in the program. While for the second run, also as
shown in Table 3.3, the ranges were further narrowed down and set at zero level.
Table 3.3 : Experimental factors for CCD
Factors
1
2
-!
-1
0
+1
+!
OD
0.3
0.52
1.05
1.58
1.8
IPTG (mM)
0.005
0.01
0.03
0.04
0.05
OD
1.3
1.37
1.55
1.73
1.8
IPTG (mM)
0.007
0.008
0.011
0.014
0.015
30
Table 3.4 : Experiment factors and levels for second CCD
Run
IPTG(mM)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0.008
0.008
0.014
0.014
0.011
0.011
0.007
0.015
0.011
0.011
0.011
0.011
0.011
0.011
3.5.1
Absorbance(OD)
1.37
1.73
1.37
1.73
1.30
1.80
1.55
1.55
1.55
1.55
1.55
1.55
1.55
1.55
Activity(U/mL)
52.025
36.759
57.246
43.567
59.222
38.933
41.149
50.160
82.345
82.082
82.411
82.806
82.543
82.148
Analysis
Two-level factorial was done with consideration of factors from Tables 3.1 and
3.2. The preparation of overnight culture and method of induction was similar to media
optimization method. Cultures were incubated at 25˚C for 50 hours with 200 rpm
shaking after induction for harvesting. After 50 hours, cells were harvested and analysis
such as enzyme activity and protein analysis were done for both extracellular and
intracellular cells as described earlier.
31
Total cell supernatant with highest enzyme activity were frozen in larger amount.
This supernatant was used to obtain the concentrated enzyme. To concentrate the
enzyme, first the frozen supernatant was thawed. Then, the thawed solution was
centrifuged for 30 minutes at 8000 rpm. After centrifuge, the supernatant was transferred
into the ‘U-tube Concentrators’ which was purchased commercially by Merck.
The Concentrators were centrifuged for 45 minutes with 12 000 rpm to
concentrate the enzymes. The concentrated enzyme was kept at 4˚C for short term
storage as the concentrated enzyme was further used in High Performance Liquid
Chromatography (HPLC).
3.6 End Product Determination
End product determination was done in order to identify the final sugar compound
that was released by the hydrolysis of soluble starch of this amylase. This determination
was carried out through High Performance Liquid Chromatography (HPLC).
3.6.1
High Performance Liquid Chromatography (HPLC)
HPLC was carried out using Waters HPLC machine with filtered deionized water
as the mobile phase. A single time plot reaction was carried out between this amylase
and variety of starch types namely the amylopectin, sagu, wheat, rice, tapioca, potato
and soluble starch.
32
The selected starch types were reacted with amylase for 18 hours at 60 °C water
bath. After incubation, the reaction was stopped by boiling the reaction mixtures for 10
minutes. The reaction mixtures were then filtered prior injecting to the HPLC machine.
The time frame for each injection was 15 minutes and the standards used were varieties
of sugar compounds that were purchased commercially.
The sugar separation of the standards and samples were generated by the HPLC
machine in the form of chromatograms. These chromatograms were compared to
identify the sugar released.
33
CHAPTER 4
RESULTS AND DISCUSSION
4.1
Optimization and Expression of E.coli BL21 Carrying Amylase in pET 22-b
Vector
The main objective of this study is to enhance the expression of a recombinant
amylase in E.coli system. The amylase enzyme had been cloned from a newly
discovered Anoxybacillus named as Anoxybacillus DT3-1. This Anoxybacillus DT3-1
was a discovery by fellow UTM researcher and was derived from a hot spring known as
Dusun Tua hot spring in Malaysia (Chai et al., 2010). The gene sequence has not been
deposited in NCBI database currently. The expression level of the recombinant amylase
using pET 22-b/ E.coli BL21 was low. Therefore, the intention of this wok is to improve
its expression. This was done by first selecting the best known-medium, followed by
optimization of the inducer concentration (IPTG) and induction time using Design
Expert software.
34
General media optimization was done by growing the E.coli BL21 in all the five
chosen media (Luria-Bertani broth (LB), Super broth (SB), Terrific broth (TB),
Chemically Defined Media 1 (CDM1) and Chemically Defined Media 2 (CDM2)) for 36
hours at 25˚C. According to Lo et al. (2007), a general temperature for enzyme
expression for E.coli was 25˚C, so this temperature was maintained throughout all the
optimization process. The culture in all media was induced similarly with IPTG of 1mM
unless specified.
Samples were collected and harvested for time interval between two to six hours
to determine the cell growth. Cell growth was checked through spectrophotometer at
OD600nm absorbance while collected samples were centrifuged. Then the supernatant was
used for enzyme assay (Fuwa assay) to quantify the enzyme activity. Figure 4.1 shows
the growth profile of E.coli in all the five media.
From Figure 4.1 clearly can be seen that the cell growth in CDM 2 medium is
relatively higher than that of the growth in other media. It had been assumed that, the
high cell growth could be due to the presence of glucose as one of the components in
CDM 2. Glucose is the primary and simplest carbon source for E.coli propagation. In
contrast, the cell growth in CDM 1 medium was relatively low compare to growth in
other media whereas the remaining three media had almost similar cell growth.
35
MK'
Absorbance (600nm)
MJ'
M5'
MI'
@O'
L'
HO'
K'
DEP'5'
J'
DEP'M'
CO'
5'
I'
I'
0'
MI'
M0'
5I'
50'
NI'
N0'
Time interval (hour)
Figure 4.1
Profiles of microbial growth in five different media at 25˚C
QI'
!"#$%&'()*+%&,--./0(
KI'
0I'
JI'
5J'R+9$4'
NI'
NI'R+9$4'
5I'
NK'R+9$4'
MI'
I'
AMI'
@O'
Figure 4.2
HO'
DEP'5'
DEP'M'
CO'
1'23,(45(.36%7(
Fuwa assay on three final intervals samples in all five media
36
Fuwa assay was used to quantify the enzyme activity and was carried out at 24,
30 and 36 hours which happen to be the log phase of the growth. From Figure 4.2, two
most prominent media that contribute in amylase expression was the CDM 2 (which
contains glucose) and LB medium. Apart from these two media, other media did not
have any promising activity signs.
Amylase activity was not found in TB, CDM 1 and SB media. In CDM 1 media,
there are mass content of trace elements which contributed to high metal ion content.
High content of ions such as Mg2+ as in MgCl (one of the trace element), have
possibility to act as inhibitor thus destabilize the enzyme production. Other than that,
high concentration of nutrients in SB and TB may contribute to lack of enzyme
production in both the media where this speeds up cell propagation and growth but
reduces enzyme production (Goh, 2009).
However, these very low and negative activities might be also due to the culture
duration. Most probably, 36 hours were not enough for the culture to reach the stationary
phase. Since the cultures were merely still in log phase at 36 hours, the cells in the
culture were only rapidly growing rather than expressing the enzyme. Yet, the LB broth
and CDM 2 showed good activities. High activity in both of these media could be due
to the fact that both media only have adequate nutrients with no unnecessary minerals,
thus only focusing in cell growth rather than enzyme productivity. One of the reasons of
lesser growth in CDM 2 could be because of the high glucose content in the medium.
This is because, upon reaching the stationary phase in the culture high concentration of
glucose cause the metabolic inhibition in the cells (Tabandeh et al., 2008). Therefore,
more precised optimizations were carried out with longer culture duration between these
two media to choose the best media.
37
4.2
Further Comparison of CDM 2 and LB Media
4.2.1
Cell Growth Profiles
During the earlier pre-screening (section 4.1), the incubation of cells was
restricted to 36 hours. However, based on Figure 4.1, the culturing period seemed to be
insufficient. The culture in CDM 2 medium has not reached stationary phase. Due to
this, the experiment was repeated with extended time up to 68 hours. As only culture in
LB and CDM 2 media gave positive result in the earlier pre-screening, the other media
were excluded in this repetition.
!8,4987+"3():;;+.0(
Q'
K'
0'
J'
N'
HO'
5'
DEP'5'
M'
I'
I'
0'
MI' M0' 5I' 50' NI' N0' JI' J0' 0I' 00' KI' K0' QI'
1%.3()<4*90(
Figure 4.3
25˚C
Profiles of microbial growth in LB broth and CDM 2 for 68 hours at
38
The differences in cell growth between both media can be clearly seen in Figure
4.3. The graph showed that, the cell growth which were determined by the optical
density at 600nm (OD600) in LB medium is much higher than in CDM 2. The biomass
production in both the media increased continuously until they reached the stationary
phase at 35th hour of incubation. However, the cell turbidity increment in LB was
significantly higher where it increased from 0.078 at zero hour to 7.361 after 68 hours.
In contrast, the growth in CDM 2 was only recorded as 0.108 at the early state and
reached up to 2.708 after 68 hours. This showed that, the culture were more favorable to
grow and produce higher biomass in LB broth.
4.2.2
Enzyme activity
Hetero protein expression using E.coli could be localized either intra or
extracellularly. Cell culturing conditions could affect the localization of amylase.
Extracellular amylase was directly quantified using the cell-free supernatant. The cell
pellet which contained the intracellular enzyme was lysed before this fraction could be
quantified using the standard Fuwa assay.
39
SI'
Activity (units/mL)
LI'
QI'
KI'
0I'
JI'
"78$%3"&&9&%$'%3-=*8<'
NI'
*#8$%3"&&9&%$'%3-=*8<'
5I'
MI'
I'
I'
0' MI' M0' 5I' 50' NI' N0' JI' J0' 0I' 00' KI' K0' QI'
Time (hours)
Figure 4.4
The extracellular and intracellular amylase activity in LB medium for 68
hours
The results of the enzyme assay for LB and CDM 2 were displayed in Figure 4.4,
4.5, 4.6 and 4.7. Figure 4.4 and 4.5 shows the extracellular and intracellular enzyme
activity in LB and CDM 2 respectively while Figure 4.6 and 4.7 showed the comparison
of activities between the two media. Both LB and CDM 2 had higher activity of
intracellular amylase compared to extracellular amylase but in LB media, the activity of
extracellular amylase increased drastically from 20th hour until end of the culture
duration.
Basically, in LB media the intracellular enzyme production fluctuates throughout
the incubation period until it reaches the stationary phase at 50th hour with enzyme
activity of 78.36 U/mL from 42.81 U/mL at zero hour. Once the cell reaches the
stationary phase, the intracellular amylase production was decreased gradually until it
reached 49.70 U/mL of activity after 68 hours of incubation. This could occur due to the
excretion of the enzyme out of the cells after the culture reached stationary phase.
40
This by chance also made the increase in extracellular enzyme to be consistent
even after the stationary phase which showed the gradual increase of extracellular
enzyme production from 24.86 U/mL at zero hour up to 78 U/mL after 68 hours of
incubation. Cultures in both LB and CDM 2 media showed presence of activity at zero
hour for both intracellular and extracellular enzymes. This occurs due to the fact the
activity was calculated with multiplication of 100 which shows high activity even at
initial phase. Other than that, the induction for both media was done once a certain OD
was reached and sample were transferred to fresh medium upon inducing. This shows
that, the culture that transferred to the fresh medium able to produce immediate activity,
thus exhibits activity at zero hour.
On contrary, the enzyme production in CDM 2 medium was rather different
compared to LB medium. Even though, intracellular enzyme production was higher
compared to extracellular enzyme production but both the enzyme production reduced
gradually until the end of incubation period. The intracellular amylase activity at zero
hour was 47.44 U/mL and the activity increased until about 15 hours where the activity
was at peak with 75.06 U/mL of activities and followed by fluctuated decrease until it
reached 18.03 U/mL after 68 hours of incubations.
On the other hand, the extracellular amylase activity was 26.23 U/mL at zero
hours and remains almost the same until it reached 35 hours of incubation and reduced
drastically to 8 U/mL activities. The activity of extracellular amylase continues to
decrease until after 45 hours of incubations, the activity became almost zero and
remained until after 68 hours of incubation. One of the reasons for higher intracellular
activity was the environment or composition of CDM 2 medium that prevented or
slowed down the movement of expressed amylase out of the cell. This is also could be
due to the osmotic pressure of the medium was higher than inside the cells as this
medium contained high concentration of glucose.
41
LI'
Activity (units/mL)
QI'
KI'
0I'
JI'
"78$%3"&&9&%$'%3-=*8<'
NI'
*#8$%3"&&9&%$'%3-=*8<'
5I'
MI'
I'
I'
0'
MI' M0' 5I' 50' NI' N0' JI' J0' 0I' 00' KI' K0' QI'
Time (hour)
Figure 4.5
The extracellular and intracellular amylase activity in CDM 2 medium
for 68 hours
Figure 4.6 and 4.7 shows the comparison between the activity in both media for
extracellular amylase and intracellular amylase respectively. According to Figure 4.6,
the productions of extracellular enzyme in both media were almost the same at the
beginning of incubation which were around 25 U/mL. However, when the incubation
reached 20 hours, the enzyme production in LB increased consistently until the end of
68 hours with highest enzyme activity of 78 U/mL. Whereas, extracellular enzyme
production in CDM 2 media were stagnant with about 25 U/mL of activity at the
beginning of experiment until the culture reached 30 hours. After that, gradual decrease
in enzyme activity from 25 U/mL to almost zero occur until the culture reached 45 hours
and remained almost zero activity until end of 68 hours. In Figure 4.6 clearly shown that
LB media was able to produce higher amount of extracellular amylase compare to CDM
2 media.
42
However, the intracellular amylase production in both media had a different
pattern which can be clearly seen in Figure 4.7. Although the enzyme activity was
almost equal at zero hour which was approximately 45 U/mL, most enzymes were
expressed in CDM 2 medium where the intracellular amylase activity reached the peak
at 15th hour and relatively higher compared to LB medium. After that, the production
was decreased gradually until end of incubation.
In contrast, intracellular amylase production in LB medium fluctuated until 25th
hour. At 25th hour, the enzyme production was almost equal again between both the
media which were approximately 48 U/mL. After that, the production of intracellular
amylase in LB medium boost up gradually until reached the peak at 50th which was the
stationary phase of the culture and proceeded with slight decreased until the end of 68
hours where the activity was only 49 U/mL. Although both media have decrease in
intracellular enzyme production, it was obvious that LB media able to express relatively
higher amylase compare to CDM 2 media.
43
LI'
Activity (units/mL)
QI'
KI'
0I'
JI'
HO'
NI'
DEP'5'
5I'
MI'
I'
I'
0'
MI' M0' 5I' 50' NI' N0' JI' J0' 0I' 00' KI' K0' QI'
Time (hours)
Figure 4.6
The comparison of extracellular amylase activity between LB and CDM
2 media for 68 hours
SI'
Activity (units/mL)
LI'
QI'
KI'
0I'
JI'
DEP'5'
NI'
HO'
5I'
MI'
I'
I'
0'
MI' M0' 5I' 50' NI' N0' JI' J0' 0I' 00' KI' K0' QI'
Time (hours
Figure 4.7
The comparison of intracellular amylase activity between LB and CDM
2 media for 68 hours
44
LB broth was more favorable to be used in expressing amylase both extracellular
and intracellular. This is because, LB medium is able to produce higher cell biomass as
well as enzyme expression. Besides, LB medium is a rich medium and contained largely
required nutrients that made it suitable for variety of microorganisms to grow in it
optimally. Yeast extract as one of the ingredient in LB broth also contribute vastly in
this because yeast extract contained a variety of types of organic molecules that needed
for bacterial growth (Sezonov, 2007). Therefore, LB broth was chosen as the best
medium and used for subsequent optimization using Response Surface Methodology
(RSM).
4.3
Optimization of Relevant Factors Using Design of Experiment
Response Surface Methodology (RSM) have experimental design such as Twolevel Factorial and Central Composite Design (CCD) in order to fit a model by the least
square technique. Accuracy of the proposed models was studied using the diagnostic
checking tests provided by Analysis of Variance (ANOVA). These plots can be used to
verify the surface and optimal conditions of factors.
The main purpose of this part of research was to obtain the best condition and
feasibility of the Design of Experiment (DoE) method on the optimization of the
selected parameters in both Two-level factorial and CCD. In order to find the most
significant parameters and their condition range, the factors were chosen based on
literature readings of various journals and papers. Upon completing Two-level factorial,
the two most significant parameters were chosen and the conditions were further
precised using CCD.
45
In order to determine the optimal design of experiment, three factors, namely the
OD at induction time, the percentage of yeast extract in broth and the amount of IPTG
used to induce were chosen. The parameters and their levels were shown in Table 4.1
and Table 4.2. These were accomplished by designing a series of compression tests by
utilizing the Two-level factorial design method followed by evaluating the output data
using ANOVA method. Two-level factorial was first carried out because it was easier to
eliminate the less significant factor and easier to be done. Other than that, this method
was also chosen because it is the most efficient method to design the compression test
which involves two or more factors each at two levels. There were 23 = 8 different
treatment combinations (A, B, C, AB, AC, BC and ABC) in this design.
Table 4.1 : Process parameters and their levels for Two-level Factorial
Factors
Low level (-1)
High level (+1)
A) Induction Absorbance (OD)
0.3
1.5
B) Yeast extract (%)
0.05
0.3
C) IPTG concentration (mM)
0.01
0.1
Fourteen runs of experiments were carried out and the observed responses were
tabulated in Table 4.3. Six runs at centre point of the design were also done. The results
were analyzed using Design Expert 6.0.4 software to obtain the regression analysis and
also to estimate the coefficients of the regression equation. Two replicates were used for
each experiment. In each complete replication of the experiments, all possible
combinations of the levels were studied in order to determine the main effects and
interactions between the parameters.
46
Table 4.2 : Experiment factors (A= yeast extract, B= IPTG and C= absorbance density)
and responses for extracellular and intracellular amylase activity
Run
A(%)
B(mM)
C(OD)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0.05
0.30
0.05
0.30
0.05
0.30
0.05
0.30
0.17
0.17
0.17
0.17
0.17
0.17
0.01
0.01
0.10
0.10
0.01
0.01
0.10
0.10
0.06
0.06
0.06
0.06
0.06
0.06
0.30
0.30
0.30
0.30
1.50
1.50
1.50
1.50
0.90
0.90
0.90
0.90
0.90
0.90
Activity(extra)
Activity(intra)
58.1731
50.9014
28.2452
35.4567
71.8266
79.0093
65.5108
58.2663
57.7694
57.7694
57.7694
57.7694
57.7068
57.7694
72.5962
75.1202
70.0721
67.5481
67.5542
65.3251
66.2539
68.5449
68.1704
67.9825
68.1078
68.1704
67.8571
68.2331
In Table 4.2, A, B and C are the coded values of the percentage of yeast extract
supplied as medium, the amount of IPTG used to induce (mM) and the OD at induction
time respectively. Using the two-level factorial, the three variables were studied and
correlated with the response which is the activity of the enzyme both extracellular and
intracellular. For an example, in one condition (denoted as ‘Run 1’ in Table 4.2), 0.05,
0.01 and 0.3 were used for yeast extract percentage, IPTG (mM) concentration and OD
at induction respectively. The activity for intracellular amylase activity and extracellular
enzyme activity were 58.17 units/mL and 72.60 units/mL respectively.
The regression models for the enzyme activity both extracellular and intracellular
suggested by the Design Expert 6.0.4 software in terms of coded factors were tabulated
in Table 4.3. The coefficients with one factor indicate the effect of that particular
variable whereas the coefficients with more than one factor indicate the interaction
between the variables. The positive sign in front of the equations represent synergistic
effects whereas the negative signs represent antagonistic effects.
47
Model
Extracellular activity
Intracellular activity
Table 4.3 : Model and coded factor
Coded Factors
= +55.92 - 9.05* B + 12.73* C + 2.29 * B * C
- 3.61* A * B * C
= + 69.13 - 1.02 * B - 2.21* C + 1.50* B * C
+1.20 * A * B * C
The value of R squared (correlation coefficient) for the extracellular activity and
intracellular activity were 1.0000 and 0.9976 respectively. The data collected in this
study was good as a regression model with R square value for than 0.9 is considered as a
good model (Li Y et al., 2006). Other than that, the closer the R square value to 1, the
higher the correlation between the experimental and predicted values.
Figure 4.8
Predicted vs. actual data for extracellular activity
48
Figure 4.9
Predicted vs. actual data for intracellular activity
The evidence of the fact that the plot of the predicted versus actual extracellular
and intracellular activity in Figure 4.8 and 4.9 are close to y = x indicated that the
prediction of data is comparable with experimental data.
The reasonable value for the
model indicated that there were good agreements between the predicted and
experimental values from the model. From the regression model, it implied that 100%
(extracellular) and 99.76% (intracellular) of the total variation in the observed response
were attributed to the experimental value. This alternatively shown that the calculated R
squared value for both activities are acceptable which therefore visualize the
performance of the model. Table 4.4 shows the comparison between experimental values
and predicted values for extracellular and intracellular activities. The variations or the
residuals between experimental and predicted values are very small or insignificant.
49
Table 4.4 : Comparison between actual values and predicted values of extracellular and
intracellular activities
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Extracellular activity (U/mL)
Intracellular activity (U/mL)
________________________________________________________________
Actual
Predicted
Residual
Actual
Predicted
Residual
58.17
58.15
0.022
72.60
72.63
-0.036
50.90
50.92
-0.022
75.12
75.08
0.036
28.25
28.24
8.033E-003
70.19
70.04
0.16
35.46
35.46 -8.033E-003
67.43
67.58
-0.16
71.83
71.80
0.022
67.55
67.67
-0.11
79.01
79.03
-0.022
65.33
65.21
0.11
65.51
65.50
8.489E-003
66.25
66.17
0.081
58.27
58.27 -8.489E-003
68.54
68.63
-0.081
57.77
57.76
0.010
68.17
68.09
0.084
57.77
57.76
0.010
67.98
68.09
-0.10
57.77
57.76
0.010
68.11
68.09
0.021
57.77
57.76
0.010
68.17
68.09
0.084
57.71
57.76
-0.052
67.86
68.09
0.23
57.77
57.76
0.010
68.23
68.09
0.15
4.3.1
Adequacy of the model
Run
To obtain a good model, several tests such as tests for lack-of-fit, tests for
significance on individual model coefficients and other tests were carried out. The
ANOVA (Analysis of Variance) in Table 4.5 and 4.6 summarize the probability value
(p-value) and F-value for each term and interaction. F-value is the ratio of mean square
of regression with regression of mean square of residual. F-value also denotes the degree
of significance of each controlled factor on tested models. On the other hand, the p-value
related to F-value shows that the probability of differences between calculated and
predicted values were only due to random experimental error (Zulfikri et al., 2007).
50
The ANOVA of the regression model shows that the model is significant for both
extracellular and intracellular activities as response. This was proved with the calculated
F-value at 95% of confidence level and low probability value (P < 0.0001), which
indicated that both models were correct and adequate.
The lack-of-fit parameter was insignificant for both of the models. The values of
“Prob > F” for both of the models were less than 0.05, thus indicated that the models
were significant. This is desirable as it showed that the terms in the model have a
significant effect on the response. The P-values are used as a tool to check the
significance of each coefficient where this is vital in understanding the pattern of the
mutual interactions between the best variables. Smaller P-values result in larger
significance of the corresponding coefficient.
The significant terms for extracellular enzyme activity responses were
represented in Table 4.5. It can be seen clearly that the linear coefficients (B and C), the
two level interactions (BC) and three level interactions (ABC) were the most significant.
These significances were concluded by the very small value of P-value (P < 0.05).
Lower F-values of lack-of-fit in both extracellular and intracellular activities which were
1.15 and 1.49 indicated that both the models had statistically insignificant and
acceptable models.
51
Table 4.5 : Analysis of Variance (ANOVA) for extracellular activity
Source
Model
B
C
BC
ABC
Sum of squares
2098.53
655.79
1296.34
41.92
104.48
DF
4
1
1
1
1
Mean2
544.63
655.79
1296.34
41.92
104.48
F value
Prob>F
7.598E+005 <0.0001
9.497E+005 < 0.0001
1.877E+006 < 0.0001
60716.5 < 0.0001
1.513E+005 < 0.0001
Significant
Curvature
11.55
1
11.55
16725.09 <0.0001 Significant
Residual
5.524E-003
8 6.905E-004
Lack of Fit
2.252E-003
3 7.508E-004
1.15 0.4153 Not significant
Pure Error
3.272E-003
5 6.543E-004
Cor Total
2110.08
13
_________________________________________________________________________
Std. Dev.
Mean
C.V.
PRESS
0.026
56.71
0.046
0.021
R-Squared
Adj R-Squared
Pred R-Squared
Adeq Precision
1.0000
1.0000
1.0000
2952.749
Table 4.6 : Analysis of Variance (ANOVA) for intracellular activity
Source
Model
B
C
BC
ABC
Sum of squares
77.41
8.36
38.98
18.05
12.03
DF
4
1
1
1
1
Mean2
19.35
8.36
38.98
18.05
12.03
F value
820.64
354.40
1652.90
765.30
509.98
Prob>F
<0.0001
< 0.0001
< 0.0001
< 0.0001
< 0.0001
Significant
Curvature
3.71
1
3.71
157.24
< 0.0001 Significant
Residual
0.19
8
0.024
Lack of Fit
0.089
3
0.030
1.49
0.3233 Not significant
Pure Error
0.099
5
0.020
Cor Total
81.30
13
_______________________________________________________________________
Std. Dev.
0.15
R-Squared
0.9976
Mean
68.68
Adj R-Squared
0.9964
C.V.
0.33
Pred R-Squared
0.9904
PRESS
0.78
Adeq Precision
98.186
52
4.3.2
Optimal Design Based on Two-Level Factorial
The Design Expert software is able to predict or suggest the optimum design
based on the provided experimental data. Figure 4.10 and 4.11 show the ramp model of
extracellular and intracellular enzyme respectively. The box of activity ranges and the
highest activity was also shown in this ramp. In both figures, the factors and their
ranges, the best conditons and the desirabilities were shown. The ‘circle’ mark on each
factors represent the optimal conditon with high desirabilites (Bhunia et al.,2008).
Figure 4.10
Ramp of extracellular amylase
According to the model in Figure 4.10, the highest activity that can be obtained
from extracellular amylase was 79.0181 units/mL if the optimal conditons of 0.30% of
yeast extract, 0.01mM concentration of IPTG inducer and induced at OD600 1.50 was
done. On the other hand, for the intracellular amylase expression the optimal conditon of
0.30% of yeast extract, 0.01mM concentration of IPTG inducer and induced at OD600
0.30 could produce the highest activity of 75.0833 units/mL.
53
Figure 4.11
Ramp of intracellular amylase
As mentioned earlier, recombinant amylase could be expressed extracellularly or
intracellularly. From this set of experiment, the optimum IPTG concentration and yeast
extract were 0.01 and 0.30 respectively, regardless the cell localization of the expressed
amylase. Interestingly, if the inducer IPTG is added at early stage of the young culture
(OD 0.3, see Figure 4.11), most amylase would stayed inside the cells. However,
inducing at higher OD of 1.5 (older cells) promoted the secretary of the enzyme to the
culture broth.
After going through all the analysis and thorougly studied the models of
extracellular and intracellular enzyme production at optimal condition proposed by Twolevel factorial, further analysis was decided to be carried out. The analysis, Central
Composite Design (CCD) was only ran for extracellular amylase since the overall
activity of extracellular amylase was relatively much higher compared to intracelluar
amylase.
54
Beside, expresing higher extracellular amylase is more relatively cost effective
and less time consuming in large scale. This is because, the process of breaking of cell
wall to obtain the intracellular enzyme is very tedious if done physically or if carried out
chemically. Based on all this factors, only extracellular enzyme was chosen to further
optimize using CCD.
4.4
Expression optimization using Central Composite Design (CCD)
Upon completing all the statistical analysis from Two-level factorial, the next
stage of experiment was Central Composite Design (CCD). The CCD approach is one of
the most popular RSM designs and it is available in Design Expert 6.0.4 software.
Generally for a CCD design, there are a total of five levels which were - !, -1, 0, +1 and
+!. As in Two-level factorial, only two levels (-1 and +1) were used.
4.4.1
Selection and validation for Significant Effect
In the earlier Two-level factorial design, three factors which were yeast extract
percentage, induction OD and inducer concentration were used. However, the CCD
design, yeast extract percentage was excluded because it played a less significant role in
the amylase expression. The range of the other two factors were narrowed down in the
CCD study. This had been achieved through observing the half-normal plot of the twolevel factorial design.
55
From Figure 4.12, the selected factors can be clearly seen. The selection was
done by clicking on the variables consecutively to align the line as near as possible to
zero. The factors that align near or along to the line was excluded and other factors
together with their cross-interactions are significant. According to Abdul-Wahab et al.
(2007), the factors that were aligned along the line and not significant were used to
estimate the experimental error. While the significant factors were chosen in the model
and proceeded as variables in CCD.
Based on Figure 4.12, the selected factors were C, B, ABC and BC. Since the
most significant factors were B ( concentration of IPTG) and C (absorbance at induction
time), thus yeast extract percentage in LB medium was omitted in CCD. The yeast
extract composition in LB medium for CCD was maintained at 0.3% as the highest
enzyme activity was when the LB medium contained 0.3% of yeast extract.
Figure 4.12
Half-normal plot of two-level factorial (extracellular activity)
56
The selected factors and their ranges were listed in Table 4.7 (a) and (b).
Optimization using CCD method was done twice, one after the other. The range for the
factors were different in these two trials. In the first CCD design (Tabble 4.8 (a)), the
ranges of each variables were set according to the range set in Two-level factorial and
these values were set at –! and +! in the program. While for the second run, the optimal
points suggested from the Two-level factorial were set at zero level instead of setting at
the outer range (-! and +!) in the case for the first design.
Therefore, only the CCD runs and statistical analysis of the second run was
deeply discussed. This is because, the second run was the optimal run and provided the
best activity with a perfect model design.
Table 4.7: Process parameters and their levels for first and second CCD
(a)
Factors
-!
OD
0.3
0.52
1.05
1.58
1.8
0.005
0.01
0.03
0.04
0.05
IPTG
Low level (-1)
Zero level (0)
High level (+1)
+!
(b)
Factors
-!
OD
1.3
1.37
1.55
1.73
1.8
0.007
0.008
0.011
0.014
0.015
IPTG
Low level (+1)
Zero level (0)
High level (-1)
+!
57
Figure 4.13 showed the response surfaces design of first CCD run. From the first
run, the results indicated that the ANOVA analysis was good. Other important
verification parameters such as lack-of-fit, R-squared and model validations were also
satisfying (see apendix). Nevertheless, curvature (peak of the plot) was not observed in
the 3-D plot as shown in Figure 4.13. A flat 3-D plot was obtained and the optimum
point was not able to determine. This may be due to the selected range for the factors set
at the beginning of the design. The ranges are most probably not suitable in this case of
study.
Figure 4.13
Response surface of first CCD design
Therefore, some modification in setting the range was done in the second CCD
design. The result for this design will be discussed in detail here. The experimental runs,
their respective parameters level and amylase activity for each run were indicated in
Table 4.8.
58
Table 4.8 : Experiment factors and responses for amylase activity
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IPTG(mM)
0.008
0.008
0.014
0.014
0.011
0.011
0.007
0.015
0.011
0.011
0.011
0.011
0.011
0.011
Absorbance(OD)
1.37
1.73
1.37
1.73
1.30
1.80
1.55
1.55
1.55
1.55
1.55
1.55
1.55
1.55
Activity(U/mL)
52.025
36.759
57.246
43.567
59.222
38.933
41.149
50.160
82.345
82.082
82.411
82.806
82.543
82.148
Figure 4.14 indicates both the contour plots (2D) and the response surfaces (3D)
graphs. These plots show the interactions between both the variables in order to obtain
maximum activity. The top of the concaved shape in the 3D plot indicated the highest
activity produced (82 U/mL). On the other hand, the elliptical shape of the contour plots
signified good interaction occuring between the two independent variables
corresponding to the response surfaces (Reddy et al., 2008).
59
Figure 4.14
Contour plots and response surfaces for the effect of IPTG concentration
and the time of induction in amylase expression
60
4.4.2
Analysis of Variance (ANOVA)
Table 4.9 reviews the ANOVA output for the linear regression model of amylase
expression in the second CCD design. The model that was generated from the second
CCD was observed to be highly significant with very low p-values (Probability values).
A p-value is a measure of the frequency of evidence one has against the null hypothesis
(Joseph, 2008). A p-value of less than 0.05 rejects the null hypothesis thus accept the
model. Apart from that, the model F-value of 17642.75 showed that there was only
0.01% chance that such a large “Model F-value” could occur due to noise. From these
analysis, it can be summarized that the proposed model can be used to explain the
experimental data with full confident.
Table 4.9 : Analysis of Variance (ANOVA) for amylase activity
Source
Model
A
B
A2
B2
AB
Residual
Lack of Fit
Pure Error
Cor Total
Sum of squares
4705.89
415.30
76.71
2046.55
2489.00
0.63
0.43
0.074
0.35
4706.31
DF
5
1
1
1
1
1
Mean2
941.18
415.30
76.71
2046.55
2489.00
0.63
F value
17642.75
7784.94
1437.88
38363.45
46657.21
11.79
0.053
0.025
0.071
0.35
8
3
5
13
Prob>F
<0.0001
< 0.0001
< 0.0001
< 0.0001
< 0.0001
0.0089
0.7926
Significant
Not significant
_________________________________________________________________________
Std. Dev.
Mean
C.V.
PRESS
0.23
62.39
0.37
1.03
R-Squared
Adj R-Squared
Pred R-Squared
Adeq Precision
0.9999
0.9999
0.9998
302.270
61
Table 4.9 also shows the statistical analysis of the R-squared, predicted and
adjusted R-squared values. The R-squared value (also known as the coefficient of
determination) for this model was calculated as 0.9999 at 95% confidence level.
Apart from that, adequate precision is a quantify measurement of the range of the
predicted responses compare to its associated error. Any ratio more than 4 is desirable
(Mason et al., 2003). The adequate precision ratio for this experiment was 302.270
showed a sufficient signal as controllable factors. Thus, indicating that the generated
model can be used.
The adequacy of the model was further verified through the lack-of-fit F-tests
which described as the variation of the data surrounded the fitted model. The lack-of-fit
parameter of this model was insigificant. This proved that the generated model was fit
enough to be used. Besides, the low value of coefficient of variation (CV) which was
0.37 also showed excellent precision and reliability of the experiment as suggested by
Ahmad et al. (2005).
Apart from that, the coefficients of the full regression model equation and their
statistical significance were also determined and analyzed. The regression models for the
enzyme activity both extracellular and intracellular in terms of coded factors were
tabulated in Table 4.10.
62
Table 4.10: Model and coded factor of second CCD
Model
Coded Factors
Activity of second CCD
= +82.39 - 7.21* A + 3.10* B - 16.65 * A 2
- 18.36 * B2 +0.40 * A * B *
Figure 4.15
Predicted versus actual plot
63
4.4.3
Model validation
Model validation was done through various types of diagnosis that was also
provided by the Design-Expert 6.0.4 software. This step was essential in determining the
availibity of the generated model as well as validate the significance of the model. Few
diagnosis such as the normal probability plot, residual plot and Box-Cox plot were
shown in this report.
4.4.3.1 Normal Probability Plot
Normal probabilty plot was analysed to check the competency of the generated
model. Defiance of the model competency was easily done by the examination of
residuals. For a model to be competent, the residual alignment supposed to be without a
specific orientations or obvious patterns. According to Figure 4.16, there were no stern
notification of abnormality as well as no obvious evidence of possible outliers. This
showed that, the generated model was normally distributed which resembles almost a
straight line.
64
Figure 4.16
Normal plot of residual for amylase production in second CCD
4.4.3.2 Residual versus predicted plot
Figure 4.17 shows the residual versus predicted graph of amylase expression.
This plot analysed the assumption of constant variance. A plot of a good generated
model should scattered randomly but with all the points were within the constant range
of residuals throughout the graph. According to Figure 4.17, the random distribution of
the residuals of the generated model showed absence in any specific trend and the
impartiality of variance was not stern. This showed that the generated model was
competent and there were consistent between the actual and predicted values of
responses.
65
Figure 4.17
Residual versus predicted plot for amylase expression
4.4.3.3 Outlier T plot
Outlier T which is also known as the Externally Studentized Residual is the plot
that shows the number of actual values that deviates from the predicted value. Outlier T
affect the statistical inference as it inflates the estimated experimental error variance and
ultimately interfered in the estimated mean value. Generally, in order to obtain a good
model, all the points in a Outlier T plot should scattered within the border lines.
According to Figure 4.18, the border lines for the plot were +3.50 and -3.50. Since all
the points were inside the border lines, this can be concluded that the model was a
competent model.
66
Figure 4.18
Plot of Outlier T of amylase production
4.4.3.4 Box-Cox Plot
A Box-Cox plot contributed the principle of selecting the correct power law
transformation. In this plot, a recommended transformation was listed based on the best
lamda value. This value can be found at the minimum point of the generated curve by
the natural log of the sum of squares of the residuals. However, no specific
transformation would be recommended if the 95% confidence interval around this lamba
include 1 (Stat-Ease Inc., 2000).
Figure 4.19 shows the Box-Cox plot of this generated model which could test the
application of any transformation. According to the analysis, there was no suggestion of
transformation on model displayed so it maintain at lamba equals to 1.
67
Figure 4.19
4.4.4
Box-Cox plot for generated model of amylase expression
Optimal Design Based on CCD
Optimal design for highest amylase expression was provided in Figure 4.20 with
high desirability value of 1.00. In order to obtain the maximum amylase expression level
which was 83.1471 unit, the values of the factors should induce the E.coli cells with
0.01mM of IPTG when absorbance reading of the cultures at 600nm has reached 1.52.
68
Figure 4.20
Ramps of various factors in CCD
All the analysis had brought together to achieve the objective of the study which
was to find the optimal condition for highest amylase yield. This can be clearly seen
from the comparisons in Table 4.11. Once the optimal conditions were obtained, the
experiment was continued to determine the end product components that able to produce
by this amylase.
Table 4.11: Comparison between enzyme at unoptimized condition, Two-level factorial
and CCD
Unoptimized
Two-level
CCD
condition
factorial
Enzyme activity (unit/mL)
72
79
83
IPTG concentration (mM)
1.00
0.1
0.007
Yeast extract (%)
0.5
0.3
0.3
Induction absorbance (OD600nm)
1.00
1.5
1.52
69
4.5
End Product Analysis
4.5.1
High Performance Liquid Chromatography (HPLC)
HPLC was done with the aid of Waters HPLC machine and the mobile phase
used was filtered deionized water.
Name
1
2
3
4
5
6
maltohexaose
maltopentaose
maltotetraose
maltotriose
maltose
glucose
Figure 4.21
Retention
Area
% Area Height
Int
Time
Type
11.867 321205
3.64
9685 BV
12.522 789118
8.94 20475 VB
14.708 1377020 15.60 28862 BV
15.774 3079962 34.89 62604 VV
16.705 2604003 29.50 63382 VV
18.125 655910
7.43 13582 VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Chromatogram of separation of standards (oligosaccharides)
70
Standard (Fig.4.20)
soluble
tapioca
amylopectin
nnnnnnnn
sago
potato
corn
rice
wheat
Figure 4.22
Chromatogram of separation of sugar components for various kind of
starch after degraded with recombinant amylase
71
Figure 4.21 and 4.22 show the chromatogram of the separations that were
generated by the waters HPLC machine. The standard used was the oligosaccharides
standard kit (Sigma Aldrich). The amylase was able to hydrolyse various types of starch
to produce end products such as glucose, maltose and other isomers of maltose. Figure
4.22 display an overlay of multiple chromatograms which show the final products of all
the reacted starch. Majority of the sugar that were produced by all of the stach were
maltohexaose, maltopentaose and maltotetraose. However, the resolution of each peak
was not distinct due to their very close retention time. Thus, the results only serve as
preliminary qualitative analysis.
72
CHAPTER 5
CONCLUSION
5.1
Conclusion
The optimization of this newly derived amylase was very helpful in increasing
the industrial usage of thermostable amylase. Since the cloning of amylase gene of
Anoxybacillus DT 3-1 was only recently done (unpublish data, manuscript in
preparation), this made the optimization of enzyme expression in E.coli DH5! a
challenging task. Therefore, this study revolved on optimization as the main
optimization through Response Surface Methodology (RSM) and followed by a subobjective on determining the end products that are able to produce by this amylase
through hydrolyzing starch by implementing the High Performance Liquid
Chromatography (HPLC).
The first part of study emphasized on choosing the best media for amylase
expression and it was carried out with five different media and other analysis such as
enzyme activities determination and growth profiles were done to verify the finding.
73
From the media optimization study, it had been concluded that the best media for this
amylase expression was LB media. This is because microbial growth was highest in LB
broth and had the highest enzyme activity.
For the second objective of this study. In this stage of experiment, a software
known as Design-Expert 6.0.4 was used to assist in obtaining the best result. Two types
of optimization had been carried out namely the Two-level Factorial Design and the
Central Composite Design (CCD). The chosen factors were the composition of yeast
extract in LB medium, the concentration of IPTG and the time or the absorbance density
for induction. Upon this, the yeast extract composition showed no significant changes in
amylase expression through Two-level Factorial Design, thus the best percentage which
was 0.3% was maintained and was not further tested in CCD. After analysis of CCD was
carried out, the optimal condition of amylase expression had been determined. In
conclusion, the optimal condition for the highest enzyme activity was inducing the
culture with 0.007mM of IPTG when the absorbance density (OD600nm) was at 1.52.
Ultimately, through this study, the enhancement of expression improved from 70 U/mL
to 83 U/mL.
This optimization was followed by the determining of the end products of
enzymatic reaction. The end products were the components that were produced by
hydrolyzing starch with the recombinant amylase. The method used to determine this
was the HPLC. From the analysis and comparison that had been carried out, it had been
concluded that the amylase degraded starch to form a wide range of oligosaccharides.
In summary, the optimization of this newly derived amylase was successfully
carried out through this approach. This ultimately concludes that the objectives of this
study were completely achieved.
74
5.2
Future Work
Since all the proposed objectives of this study had been achieved, the future
work on optimization of this amylase can be focused on other approaches. Optimization
of physical factors such as the incubation temperature and the agitation speed of the
incubator shaker also can be done. This is because physical parameters also play vital
role in producing high activity of enzyme.
75
REFERENCES
Ahmad, A. L., Ismail, S. and Bhatia, S. (2005). Optimization of coagulation-flocculation
process for palm oil mill effluent using response surface methodology.
Environmental Science and Technology. 39: 2828-2834.
Ashita, D., Gupta, J. and Khanna, S. (2000). Enhanced production, purification and
characterization of novel cellulose-poor thermostable, alkali tolerant xylanase
from Bacillus circulans AB 16. Process Biochemistry. 35:849-856
Badal, C., Gwo-Jenn, S., Kailash, C., Lloyd, W. and Gregory, Z. (1989). New
thermostable !-amylase-like pullulanase from thermophilic Bacillus sp. 3183.
Enzyme and Microbial Technology. 11: 760-764
Bhunia, P. and Ghangrekar, M. M. (2008). Statistical modeling and optimization of
biomass granulation and COD removal in UASB reactors treating low strength
wastewaters. Bioresource Technology. 99: 4229-4238.
Chai, Y. Y., Kahar, U. M., Md. Salleh, M., Md. Illias, R and Goh, K. M. (2010).
Anoxybacillus sp. SK3-4 and Anoxybacillus sp. DT3-1 are novel pullulandegrading thermophilic bacteria isolated from Malaysia hot springs.
Che Radzi, N. (2007). Separation of hydrochloric acid and glucose using
electrodialysis. Masters Thesis. Universiti Sains Malaysia.
76
Goh, P.H. (2009). Enhancement of recombinant mutant cyclodextrin glucanotransferase
H43T/Y87F protein expression level. Bachelor of Science with Honours Thesis.
Universiti Teknologi Malaysia.
Goyal, N. Sidhu, G. S., Chakrabarti, T. and Gupta, J. K. (1995). Short communication:
Thermostability of !-amylase produced by Bacillus sp. E2-a thermophilic
mutant. World Journal of Microbiology and Biotechnology. 11: 593-594.
Gupta, R., Gigras, P., Mohapatra, H., Goswami, V. K. and Chauhan, B. (2003).
Microbial !-amylases: a biotechnological perspective. Process Biochemistry. 118
Haki, G. D. and Rakshit, S. K. (2003). Developments in industrially important
thermostable enzymes: A review. Bioresource Technology. 89:17-34
Hamid, N. S.A., Zen, H. B., Tein, O. B., Halifah, Y. M., Saari, N. and Abu Bakar, F.
(2003).
Screening
and
identification
of
extracellular
lipase-producing
thermophilic bacteria from a Malaysian hot spring. World Journal of
Microbiology and Biotechnology. 19:961-968.
Hobel, C. F. V. (2004). Access to Biodiversity and New Genes from Thermophiles by
Special Enrichment Methods. Doctor Philosophy, University of Iceland,
Reykjavik.
Huber, R., Huber, H. and Stette, K. O. (2000). Towards the ecology of
hyperthermophiles: biotopes, new isolation strategies and novel metabolic
properties. FEMS Microbiology Reviews. 24:615-623
Kahar, U. M. (2009). Purification and characterization of recombinant cyclodextrin
77
glucanotransferase (CGTase) using chromatography approach. Bachelor of
Science with Honours Thesis. Universiti Teknologi Malaysia.
Kenji, S., Akira, Y., Hajime, K., Mamoru, W. and Mitsuaki, M. (1994). Purification and
characterization of three thermostable "-N-acetlyhexoseaminidase of Bacillus
stearothermophilus CH-4 isolated from chitin-containing compost. Applied and
Environmental Microbiology. 60: 2911-2915
Kumar, S. and R Nussinov. (2002). Relationship between ion pair geometries and
electrostatic strengths in proteins. Biophysiology Journal. 83(3):1595-1612
Lee, H. J.(2002). Designing of an electrolysis desalination plant. Desalination. 142: 267286
Li, Y., Cui, F., Liu, Z., Xu, Y. and Zhao, H. (2006). Improvement of xylanase
production by Penicillium oxalicum ZH-30 using response surface methodology.
Enzyme Microbiology Technology. 40: 1381-1388
Ling, C. Y. (2008). Optimization of gamma-cyclodextrin production using mutant
CGTase H43T/Y87F. Bachelor of Science with Honours Thesis. Universiti
Teknologi Malaysia.
Lo, P. K., Hassan, O., Ahmad, A., Mahadi N. M. and Md. Illias, R. (2007). Excretory
over-expression of Bacillus sp. G1 cyclodextrin glucanotransferase (CGTase) in
Escherichia coli: Optimization of the cultivation conditions by response surface
methodology. Enzyme and Microbial Technology. 40:1256-1263.
Mamo , G. and Gessese, A. (1999). A highly thermostable amylase from a newly
isolated thermophilic Bacillus sp. WN11. Journal of Applied Microbiology.
86:557-560
78
Mohammadi, T., Razmi, A. and Sadrzadeh, M. (2004). Effect of operating parameters
on Pb2+ separation from wastewater using electrodialysis. Desalination. 167:
379-385
Montmogery, D. C. (1997). Design and analysis of experiments. Fourth edition. New
York : John Wiley.
Muniandy, K. (2009). Cloning and molecular characterization of DNA fragments
obtained from 5’-RACE of starch branching enzyme from Metroxylon sagu.
Bachelor of Science with Honours Thesis. Universiti Malaysia Sarawak.
Niehaus, F., Bertoldo. C., Kahler. M. and Antranikian. G. (1999). Extremophiles as a
source of novel enzymes for industrial application. Applied Microbiology
Biotechnology. 51:711-729.
Nikerel, E., Oner, E. T., Kirdar, B. and Yildirim, R. (2006). Optimization of medium
composition for biomass production of recombinant Escherichia coli using
response surface methodology. Biochemical Engineering Journal. 32: 1-6.
Ong, R. M. (2005). Molecular cloning of cyclodextrin glucanotransferase gene from
Bacillus sp G1. Master. Universiti Teknologi Malaysia.
Pandey, A., Nigam, P., Soccol, C. R., Soccol, V. Y., Singh, D. and Mohan, R. (2000).
Advances in microbial amylases. Biotechnology Applied Biochemistry. 31: 135152
Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis,
K. B. and Erlich, H. A. (1998). Primer-directed enzymatic amplification of DNA
with thermostable DNA polymerase. Science. 239:487-491
79
Sharp, R. J., Riley, P. W. and White, D. (1992). Heterotrophic thermophilic bacilli.
Thermophilic Bacteria. 19-50. Edited by Kristjansson, J. K. Boca Raton: CRC
Press
Shaw, J., Pang, L., Chen, S. and Chen, I. (1995). Purification and properties of an
extracellular !-amylase from Thermus sp.. Botanical Bulletin Academia Science.
36: 195-200
Stat-Ease, Inc. (2000). Design Expert User’s Guide, The Stat-Ease Inc., USA
Tabandeh, F. Khodabandeh, M., Yakhchali, B., Habib-Ghomi, H. and Shariati, P.
(2008). Response surface methodology for optimizing the induction conditions
of recombinant interferon beta during high cell density culture. Chemical
Engineering Science. 63: 2477-2483
Van der Maarel, M.J.E.C, van der Veen, B, Uitdehaag, J.C.M, Leemhuis, H. and
Dijkhuizen, L. (2002). Properties and applications of starch-converting enzymes
of the !-amylase family. Journal of Biotechnology. 94:137-155
Zulfikri, A. R., Rosli, M. I., Mahadi, N. M. and Hassan, O. (2007). Experimental design
to optimization of beta cyclodextrin production from ungelatinized sago starch.
Eur. Food. Res. Technol.
80
APPENDIX A
RESULTS AND DISCUSSION (CCD)
Appendix A1: Process parameters and responses (amylase activity) for first CCD
81
Appendix A2: ANOVA analysis of first CCD
82
Appendix A3: Selected Model Validation Analysis for First CCD
1) Normal Residual Plot
83
2) Residual versus Predicted Plot
3) Cook’s Distance Plot
84
4) Predicted versus Actual Plot
Appendix A4: Optimal Design Based on First CCD
85
86
Appendix B1: Sugar Separation of Various Kind of Starch After Degraded With
Recombinant Amylase
Name
2
3
4
5
6
7
Retention Time
10.512
11.324
11.892
12.416
13.403
15.247
Figure B-1
Area
3940120
3554165
1947051
2338499
4882987
230380
% Area
23.02
20.76
11.38
13.66
28.53
1.35
Height
108134
139914
68318
67524
117893
6041
Int Type
VV
VV
VV
VV
VV
VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Peak Codes
Chromatogram of sugar separation in soluble starch
87
Name
4
5
6
7
8
9
10
11
12
Retention Time
10.480
11.362
11.986
12.498
13.445
15.356
16.533
16.733
17.018
Figure B-2
Area
4216105
3150447
1042635
1846223
4928295
217711
5995
15126
81540
% Area
27.14
20.28
6.71
11.89
31.73
1.40
0.04
0.10
0.52
Height
115858
128329
37995
52834
115315
5669
779
1624
3112
Int Type
VV
VV
VV
VV
VB
BB
BV
VV
VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Peak Codes
I06
I06
Chromatogram of sugar separation in tapioca starch
88
Name
Retention Time
10.227
10.544
11.401
11.893
12.486
13.455
15.383
16.567
17.020
Figure B-3
Area
2526525
2718874
3811844
2412671
1747392
5772288
280814
3669
61324
% Area
12.71
13.68
19.18
12.14
8.79
29.04
1.41
0.02
0.31
Height
64627
106505
148060
77437
54326
115303
6777
526
2103
Int Type
VV
VV
VV
VV
VV
VV
VB
BV
VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Peak Codes
I06
Chromatogram of sugar separation in amylopectin
89
Name
2
1
3
4
5
6
7
8
Retention Time
11.459
10.547
11.924
12.564
13.502
15.384
16.533
16.965
Figure B-4
Area
4060596
4723847
843880
1914271
6155857
350694
6092
81758
% Area
22.39
26.05
4.65
10.55
33.94
1.93
0.03
0.45
Height
152049
94045
36280
52788
116943
7815
826
2752
Int Type
VV
BV
VV
VV
VV
VB
BV
VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Peak Codes
I06
Chromatogram of sugar separation in sago starch
90
Name
Retention Time
11.430
10.612
11.898
12.551
13.512
15.404
16.905
18.400
18.600
18.830
Figure B-5
Area
4421963
3881762
1185897
1940749
5695627
479357
153177
482
1311
3131
% Area
24.89
21.85
6.68
10.93
32.06
2.70
0.86
0.00
0.01
0.02
Height
158464
151460
46723
53743
103843
11011
4088
94
180
226
Int Type
VV
VV
VV
VV
VV
VV
VB
BB
BV
VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Peak Codes
I08
I08
Chromatogram of sugar separation in potato starch
91
Name
2
1
3
4
5
6
7
8
9
Retention Time
11.388
10.267
11.759
12.577
13.477
15.377
16.715
18.950
19.324
Figure B-6
Area
3987373
399604
201452
1140259
5707013
255426
135450
1762
16861
% Area
33.66
3.37
1.70
9.63
48.18
2.16
1.14
0.01
0.14
Height
129912
12043
13332
31343
93098
5811
3640
236
674
Int Type
VV
VV
VV
VV
VV
VV
VB
BV
VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Peak Codes
I06
I08
Chromatogram of sugar separation in corn starch
92
Name
1
2
3
4
5
6
7
Retention Time
9.233
10.513
11.280
12.552
13.492
15.376
16.558
Area
2452
2347032
2055209
1765842
3310129
193463
109375
% Area
0.03
23.99
21.01
18.05
33.83
1.98
1.12
Height
408
63406
85267
34746
78632
4206
3098
Int Type
BV
VV
VV
VV
VV
VV
VB
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Peak Codes
I06
Figure B-7
Chromatogram of sugar separation in rice starch
93
Name
1
2
3
4
5
6
7
8
Retention Time
9.167
10.597
11.299
12.621
13.505
14.238
15.439
16.657
Figure B-8
Area
2901
1289571
2189932
1333826
3282693
2351716
176444
158019
% Area
0.03
11.96
20.31
12.37
30.44
21.81
1.64
1.47
Height
447
31628
85715
31665
80525
70342
4050
4261
Int Type
BV
VV
VV
VV
VV
VV
VV
VB
Amount
Units
Peak Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Chromatogram of sugar separation in wheat starch
Peak Codes
I06