Copyright is owned by the Author of the thesis.  Permission is given for 
a copy to be downloaded by an individual for the purpose of research and 
private study only.  The thesis may not be reproduced elsewhere without 
the permission of the Author. 
 



DEVELOPMENT OF OPTIMAL 
FERMENTATION EXPRESSION 
SYSTEMS FOR RECOMB INANT 

PROTEINS 

A thesis submitted in partial fulfilment 
of the requirements for the degree of 

Masters 
ill 

Technology 

at Massey L1ru\'ersity, 
Palmerston orth, 

ew Zealand 

Daniel Manderson 

2004 



CERTIFICATE OF REGULATOR COMPLIANCE 

This is to certify that the research carried out in the Masterate Thesis entitled 

Development of Optimal Fem1entation Expression Systems for Recombinant Proteins 

in the Institute of Technology and Engineering at Massey University, ew Zealand: 

(a) is the original work of the candidate, except as indicated by appropriate 

attribution in the text and/or in the acknowledgement; 

(b) that the text, excluding appendices/ annexes, does not exceed 40,000 words; 

(c) all the ethical requirements applicable to this study have been complied with 

as required by Massey University, other organisations and/or committees 

which had a particular association with this study, and relevant legislation 

Please insert Ethical Authorisation code(s) here: (if applicable) _______ _ 

Candidate's Name: 

Daniel Manderson 

Signature: 

Date: 

Supervisor's Name: 

Yusuf Chisti 

Signature: 

Date: 



MASSEY UNIVERSITY 

ABSTRACT 

DEVELOPMENT OF OPTIMAL 
FERME TATION EXPRESSION 
SYSTE 1S FOR REC01IBI A T 

PROTEI S 

by Daniel Manderson 

This research set out to maximise the titre of four recombinant protein 

products (i.e. Eg95 vaccine antigen against Echinococctts gra1111/os1fS", a aspartyl 

protease inhibitor homologue, AJpin; a secreted cytok:ine granulocyte colony 

stimulating factor (G-CSF); a secreted gonadotropin ovine follicle stimulating 

factor (oFSH)) and de,,elop parameters for the expression of those proteins in 

a small scale stirred tank bioreactor. 

Production of Eg95 as inclusion bodies in E. coli was influenced by the 

medium, feeding strategy, induction timing and dissolved oxygen 

concentration. Expression was greatest using the medium Terrific Broth. 

Higher Eg95 titres were favoured using exponential feeding, a low dissolved 

o>..7gen concentration and with cells induced in mid-exponential growth. A 

maximum titre of 1.73 g/L of Eg95 was produced in a fed-batch fermentation 

controlled at 37°C, pH 7.0 and 30% dissolved o>..7gen. Induction with 0.1 

mM of IPTG added four hours after inoculation, was optimal. The 

maximum titre attained, was a 360% improvement on fermentations prior to 

this research. 



Aspin was used to investigate the culture conditions for maximizing the 

production of soluble protein in E.coli. Soluble Aspin production was 

favoured at low expression rates. A volumetric titre of 0.220 g/ L of soluble 

AJpin was attained in batch fermentation by inducing with 2 g/ L of L­

arabinose, with the temperature reduced from 37°C to 23°C and by 

maintaining a low dissolved oxygen (DO) concentration. This yield was 

relatively high compared to previous reports [1-3]. 

G -CSF production in the yeast Pichia pasloris was influenced by the 

medium, pH and methanol-to-cell ratio. A maximum titre of 0.028g/ L of G­

CSF was produced in shaker flasks of enhanced yeast extract H y-Soy dextrose 

medium (YEHD), maintained at 200 rpm, 30°C, pH 6.0 and with 1 % (v / v) 

methanol fed per day. Cells were resuspended to an optical density of 8 prior 

to induction. No improvement in G-CSF was achieved in the fermenter, 

likely due to an inhibition by toxic materials. The optimised shaker flask yield 

was consistent \vith preYious reports [4-6] . 

Production of oFSH in insect cells was influenced by the cell density at 

inoculation and rate of agitation. 0.001g/ L of oFSH was produced in shaker 

flasks inoculated at a density of 1 x 106 cells / mL, cultured at 27°C and 

agitated at 140rpm. This represented an improvement over previous yields 

[7]. 



TABLE OF CONTENTS 

Table of Contents ............ ...... .. ............................................................. .. ..... ... ............ .... ... .. i 
List of Figures .... .. ..... ....... ... ......... .. .. ......... .. .... ............................ .... ....... ............................... i 
List of Tables ... ......... ........... ................................................................................................. i 
List Of Abbreviations and Symbols ... .. ..... ....... ........ .... ... ............ .... .. ..... .... ......... .... ......... i 
Ackno,vledgments ................................ ..... ... ...... .. ... ... ........ ..... ... ...... .... ....... .... ....... .. .. ... ..... ii 
1 Literature Review ......... ........... .......... .... ...... ..... ..... .... ..... .......... ...... ................. .... .... ..... 1 

1.1 Host Cell Line .... ......... ............. ...... ......... ......... ....... ......... ..... ... ......... .. ... ... ..... 1 
1.2 Plasmid Expression ... .......................... ................ ... .. ... ... .. ... ...... ...... ... ........... 3 

1.2.1 Expression \ Tector .. ..... ...... ....................... ........ ............. .. ......... .............. 3 
1.2.2 Fusion Partners .... ............ ......... .............................. ... ............ .... .. ....... .... 8 
1.2.3 Induction .... ... .. ........ .... ....... ....................... ......... ... ... .. ..... .... ... ... .. ... .. ... ... 10 
1.2.4 Plasmid Stability ............. ............. .... ..... ...... ..... ..... .................... ............. 12 

1.3 Process Conditions ................. ...... ... ... .... ......... ............... ............................ . 14 
1.3.1 Media Composition .. ... ...... .... ..... ....... .... .......... ... .. ... ........ ... ....... .......... . 14 
1.3.2 Feeding Regime ................... ..... ..... .. .... .. .. ... .... ......... ................ ...... ....... 17 
1.3.3 Temperature .. .. .. .. ............. ... ... .. ........... .. ...... ... .. .... ......... .... .. .. ........... .. ... 19 
1.3.4 Agitation ......... .. .. ................. ... ........... ........... ................ ...... ................. ... 21 
1.3.5 pH Conditions .... ... ................ .. ................................................. ............ 22 
1.3.6 Dissoh·ed 0""·-ygen .. .. .. ........... ...... ...... ................ ... ... ..... .... .. .... ........... ... 24 

1.4 D o\vnstream Processing .................. ... ...... ... ............................................... 26 
1. 4 .1 Clarification ............ ... .. ........... ............................................................... 28 
1.4.2 Cell Disruption .... .. ... .......... ...... ................ ...... .......... .............. .... ........... 29 
1.4.3 Initial Purification ........... ...... .. ........................ ...... ... ... .......................... 32 
1.4.4 Concentration ..... .................... ........ .... .. .. ........ ........ ........ ........ ............... 36 
1.4.5 Renaturation .. ......................................................... ........... .... .......... ...... 37 
1.4.6 Final Purification ....... ........................................................... ... ............ . 38 
1.4.7 Dehydration ... .. .. ...... .... .. .... ..... ....... ..... .. ... ...... ......... .. .... ........... .. ....... ... .. 42 

1.5 Assays .... ........ ..... ... ... ........ .... .... .. ............................ ......... ..... .. ......... .......... ..... 44 
1.5.1 Cell Density .... ... ..... .... ...... ............. ........................................................ 44 
1.5.2 Protein .r\ ssay .......................... ....... ........................................................ 45 
1.5.3 Protein Activity .... .. ........................................ ..... .. ..... ..... ..... ................. 47 

1.6 Conclusion ................ ................... ........ .... ..... .... .... ........................................ 50 
2 Insoluble Protein Production in E.coli ...... .. .. ......... ....... ........................ .... ............ 53 

2.1 Introduction ........ .. .. .. .............. ... .. ........ .. .......... ...... ..... .... ........ ..... .... ...... .. ..... 53 
2.2 Materials and Methods .. ....... .. .................................. ... .. ........ .... .... .... .. ....... 55 

2.2.1 Strain and Plasmid ................................................................................ 55 
2.2.2 Media and Chemicals ................. .................. .... .. ...... ... .. ... .. ........ .. ........ 55 
2.2.3 Inoculum Development .... ...................... ....... ... .. ..... ...... ...... ........ ...... . 56 
2.2.4 Medium Trials ... ... .... .... ... ............ .............. ...... ....... .. ... .... .. .. ... .. .. .... .. .. ... 56 
2.2.5 Inducer Concentration Trial... ...... ........... .... ................. ........ ... .. ..... .... 57 
2.2.6 Induction Time Trial.. .... .... ...... .............. ... .................................... ....... 57 



2.2.7 Fermenter Trials .... ..... ................................................ .................... .. .... 57 
2.2.8 Protein Extraction ................... ... ..... .... ....... .... .. ... ... ......... ..................... 59 
2.2.9 Analytical Methods .. .................................... ........................................ . 61 

2.3 Results and Discussion ............ ...... ... ....... .......... ... .... ............ .. ....... ............. 63 
2.3.1 Media Trials ................. ... ............. ..... .............. ....... .......... ............ .......... 63 
2.3.2 Inducer Concentration Trial... ...................................................... .... .. 67 
2.3 .3 Induction Time .............. ...... .. .... ........... ... .. ...... ....... ....... ..... .................. 69 
2.3.4 DO concentration ................... .......... ............................................. .... .. 71 
2.3.5 Feeding Strategy ... ...... ..... .. ................. ......... .............................. ........ ... . 74 

2.4 Conclusion ........................... .. ... ... .. ........... ...... .. ......... ... .... .. .. ..... .......... ......... 78 
3 Soluble Protein Production in E.coli .......... ... .............. ....... ... ... .. .. ... .. ........... ......... 79 

3.1 Introduction .................. ..... .. .... ..... ...... ..... ... .. ......... .. ..... .... .. ... .............. ......... 79 
3.2 Materials and Methods .......... ...... .... .......... .......... ... ...... ........ ...... ........... ..... 81 

3.2.1 Strain ...... ....... ... ...................... .. ........................................ .............. ....... .. 81 
3.2.2 Media and Materials .......................... .................................... ..... ... ...... . 82 
3.2.3 Environmental Screening Trial... ...... .... ........ ................ .... ................. 83 
3.2.4 Shaker Flask Trials ............................... ........... ........ .. .......... ...... .. ....... .. 83 
3.2.5 Protein Extraction ... ..... ....... ...... ... ... ...... ............ .................... .............. . 84 
3.2.6 Analysis .... .................................... ..... .......... .. ........ .. ...... ... ................... .. .. 85 
3.2.7 Fennenter Setup .................... .............................. .. .................. ... .. ........ 85 

3.3 Results and discussion ......................................................... .... ..... ... ........... 87 
3.3.1 Factor Screening Trials ........................................................................ 87 
3.3.2 Inducer Concentration .... .................................................................... 92 
3.3.3 Temperature .... ....... ................. ........... .... ..... ... ... ................................ .... 95 
3.3.4 Fermenter Trials ....... ..... .... ................................................................... 97 

3.4 Conclusion ................... ........ ................ ........... ............ ........ ..... ... ...... ......... . 107 
4 Pichia pastoris ................. ................... ... ... .. ................... ... ... ...................................... 108 

4.1 Introduction .................... ...................... ... ... .. .... .... .............. ........... .. ..... ...... 108 
4.2 Material and Methods ............. ...... .... .............. ... ............ .. .... ... ....... .... ....... 110 

4.2.1 Strain ............................................................... .. .................................... 110 
4.2.2 -1edia .. ..... ............................................... ..... ................ ... .... .................. 110 
4.2.3 Inoculum Developmcnt ....................... ............................... ........ .. .. .. 111 
4.2.4 Medium Trials ..................................................................................... 111 
4.2.5 Environmental Screening Trial... ..... .............................. .. ... .... ... ...... 112 
4.2.6 Optical Density Trial .... ................................ .... ................................ . 113 
4.2.7 pH Trial... ...... ... .......... .. ..... ......... .......................... .. .. ....... ..... .. ............. . 113 
4.2.8 Methanol Concentration Trial... ............................................. ..... ..... 114 
4.2.9 Fermentation ..... ... ........ ......... .... ...... ... ... ... ... .. .. ....... ........ .. ..... ... ...... ..... 114 
4.2.10 Cell Density ........... .. .... .. ..... .. .............................. .. ........................... 115 
4.2.11 Protein Assays .............................................................. ................ .. 115 

4.3 Results & Discussion .... ....... ..... ................................................................ 117 
4.3.1 Media Trials ....................................... ... ... .. ....... .... ... ............. .... ... ........ 117 
4.3.2 Environmental Screening Trial ......... ......................... ..... ......... ........ 120 
4.3.3 Optical Density at Induction .......................... ..... ............................. 124 
4.3.4 pH Trial ......... ........ ........ ...................................... ................................. 125 
4.3.5 Methanol Concentration ........ ....... ..... .. .... .. .. ......... .... .. .... ..... ... .... .. ... . 127 

11 



4.3.6 Feimentation .. .. .. ..... ....... ...... ...... ....... ........ ................. ......................... 129 
4.3.7 P.pastoris oFSH Fe1mentation .... ........................ .. .... .... ............ ..... . 135 

4.4 Recomendations ..... .... .... .. .... ..... .... ............................................................ 137 
4.5 Conclusion .......................... ..................................... ... ......... ...... ....... .... .... .. 138 

5 Insect cells .................................................................. ...... ...... .. ...... ... ..... .. .... ....... ... ... 140 
5.1 Materials and Methods ...... .. .. .... .... .................... ..... ... ................ .... ........... 142 

5.1.1 Strain and l\1edia ............................................... .................................. 142 
5.1.2 Adaptation of Cell line .... .. .. .... ............... ..... ............. ... .. ....... ........... ... 142 
5.1.3 Cell Density and Growth rate ......... ............ .......................... ........... 143 
5.1.4 Environmental Factor Screening Trial .... ....... ..... ........ ...... .... ...... ... 144 
5.1.5 Inoculation Cell Density Trial... ........ ... .............. ... ............ .. .... ......... 144 
5.1.6 Agitation Rate Trial... ......... ......... ........ ....... .. ..... ... ... .... ..... ....... .. ...... ... 145 
5.1.7 Bioreactor .................... ................ ..................... ...... ... .............. ............. 145 
5.1.8 Western Blot ..... ..... ........ ....... .... .......... ............ ..................................... 146 

5.2 Results and discussion .. .... ... .... .. .. ............ .... ............................................. 14 7 
5.2.1 Environmental Screening Trial... .... ............ ..................................... 147 
5.2.2 Inoculation Cell Density .. .. .. ......... ... ..... .. ..... .......................... ........... 151 
5.2.3 Agitation Rate ................... ... ...... .. ............ .... ....... ...... ...... ... ... .. ....... .... .. 153 
5.2.4 Bioreactor .................. ................... .... .... ... .. ..... ......... ............................. 156 

5.3 Recommendations ....... ...... ..... ....... .... .... ...... .... ...... .................................... 162 
5.4 Conclusion .. .... .... ................. ...... ...... ... ...... ........ .... .................... ..... .... .... ..... 163 

Concluding Remarks ......................... ............ .............................. ................................... 164 
Bibliography ........................................ ................................................. ....... ... ........... ....... 167 

111 



LIST OF FIGURES 

Number 

Figure 1: pET102 vector map showing the various genes (source 
lnvitrogen). The lacZ gene is shown here as the gene of interest. 
The ribosome binding site (RBS) proceeds the gene of interest 
which is expressed with fusion partners HP thioredoxin, EK 
recognition, VS epitope and poly-histidine domains. The plasmid 
contains the T7 promoter regulated by the operator (/acO). A T7 
translational terminator stabilises the mRNA and ampicillin 
resistance gene (bla) which aids clone selection. An origin of 

Page 

replication ( ori) determines the plasmid copy number ................................. S 
Figure 2: pPICZ vector map showing the various genes (source lnvitrogen) . 

The pPICZ vector was used for the expression of colony 
granulocyte stimulating factor (G-CSF). The gene of interest was 
cloned into the various cloning sites (Stu 1, EcoR I, Pml 1. ... ). 
Expression is under the control of the alcohol oxidase (AOX1) 
promoter and the gene expressed with the fusion partners myc­
epitope and poly-histidine tag to aid protein detection and 
purification. The plasmid contains the TEF1 and EM7 promoters 
which control the expression of the Zeocin ™ resistance gene. The 
AOX1 translational terminator stabilises the mRNA and the pUC 
origin of replication (ori) aids replication and maintenance of the 
plasmid in E. coli ....... ....... .... ..... ......... .......................................... ..... ... ........ .......... 7 

Figure 3: pMIB vector map showing the nrious genes (source Invitrogen). 
The pMIB vector was used for the expression of ovine follicle 
stimulating factor (oFSH). The gene of interest was cloned into the 
various cloning sites (Sph 1, Hind III, A .ip 7181. . .. ). Expression is 
under the control of the OpIE2 promoter and the gene expressed 
with the fusion partners VS epitope and poly-histidine tag. The 
plasmid contains the OplE1 and EM7 promoters which control the 
expression of the blasticidin and ampicillin resistance gene. The 
AOX1 translational terminator stabilises the mRNA and the pUC 
origin of replication (ori) aids replication and maintenance of the 
plasmid in E.coli .. ....... ................. ................ .... ... ....................... ..... ... ..... ... ........... 8 

Figure 4: General stages of downstream processing for protein production, 
adapted from Perry's Engineering Handbook [136] . ...... ........ .................... 27 

Figure S: Techniques of cell disruption showning the various chemical, 
biological and physical methods ...... ....... .... ................. .............. ......... ... ... ...... 30 

Figure 6: Comparison of various separation processes showing the range of 
particle and molecular size covered and the primary factor 



governing the separation process. Source Unit Operations Of 
Chemical Engineering [164] .................... ....... ... .... .... ... .... ................................ 33 

Figure 7: Equation for exponential feeding regime [93]. F(t) is the flow rate 
of feed at time t, µis the desired specific growth rate 01·1), V(t) the 
reactor volume at time t (L), S1. the substrate concentration in feed 
(g/ L) and S(t) is the substrate concentration in culture at time t 
(g/ L), X(t) cell concentration at time t (g cell dry weight/ L), Y ,1, 

cell titre on glucose (g dry cell weight/ g), tis time Q1) ... ...... .. ..................... 58 
Figure 8: Vessel specification for the 3300mL BioFlo 3000 bioreactor (New 

Brnnswick Scientific., Edison, NJ, USA). All dimensions are in 
mm ......... .. ...... .......................................................................................... .... ...... ... 59 

Figure 9: Equation for the calculation of the specific growth rate (µ). X, is 
the optical density at time t, X,1 is the optical density at time 0. t is 
the time period between 0 and t ........... ..... ..... ... ..... .... ... ......... .. ...................... 61 

Figure 10: Formula for the calculation of the dry cell weight. All 
measurements are in grams. Wet cell weight (WCW), dry cell 
weight (DC\'V') .. ............................ .............. ....... .. ... .............. ......... .. ....... ............ 61 

Figure 11: Broth pH variation in shake flask cultures of various media: 
Terrific Broth (TB), Super Broth (SB), S.O.B, Luria Broth (LB) and 
M9 Mi.t1in1al Medium (M9). 1mL of inoculum was added to 50mL 
of each of the five media in separate 250mL shaker flasks. The 
cultures were then incubated at 180rpm and 37°C. After 4 hours 
cultures were induced with 0.1 mM IPTG with the pH measured 
every two hours thereafter ......... ...... .......... ........ .................... ... ....... .. .. ........... . 65 

Figure 12: Comparison of the final cell density and Yolumetric titre of Eg95. 
11edia used included: Terrific Broth (TB), Super Broth (SB), S.O.B, 
Luria Broth (LB) and M9 Minimal Medium (M9). 1 mL of inoculum 
was added to 50mL of each of the five media in separate 250mL 
shaker flasks . The cultures were then incubated at 180rpm and 
37°C. After 4 hours cultures were induced with 0.1mM IPTG. 
Total protein, recombinant protein content and final cell density 
were measured after 8 hours ............... ........... ...... .... ............. ........................... 66 

Figure 13: E ffect of IPTG concentration on the specific and volumetric 
titres of Eg95. 1mL of overnight seed was added to twelve 250mL 
shaker flasks containing 50mL of Terrific Broth. Duplicate shaker 
cultures were induced four hours after inoculation with; 0.0001, 
0.001, 0.01, 0.1, 0.5 and 1.mM of IPTG. All cultures were incubated 
at 180rpm and 37°C. Total protein, recombinant protein content 
and final cell density were measured after 8 hours ......... .......... .. ............. ... . 68 

Figure 14: Effect of induction time on the volumetric and specific titres of 
Eg95. lmL of overnight seed was added to eight 250mL shaker 
flasks containing 50mL of Terrific Broth. Duplicates cultures were 
induced at 0, 2, 4 and 6 hours after inoculation with 0.1.mM of 
IPTG. All cultures were incubated at 180rpm and 37°C. Total 
protein, recombinant protein content and final cell density were 
measured after 8 hours . ................................ ............................... ............ ......... 69 

11 



Figure 15: Effect of dissolved oxygen concentration on biomass growth in 
fed-batch fermentations. Fermentations were inoculated with 
100mL of seed OD~1.5 in 1.4L of Terrific Broth. The dissolved 
m-.7gen (DO) was controlled at 30%, pH at 7.0 and temperature at 
37°C. All fermentation cultures used a feed solution of 315g/ L of 
glycerol and 315g/ L of yeast extract. Feeding was controlled by the 
automated program BioCommand (New Brunswick Scientific). 
Feed was supplied initially at a rate of 15mL/ h, increasing by 
0.15mL/ h for every minute above their respective set point. 
Fermentations were induced with 0.1mM IPTG after 4 hours and 
maintained until two sequential reductions in cell density were 
measured as indication the culture was going into stationary phase .. ...... 73 

Figure 16: 15% SDS-PAGE gel of Talon column elusions of varying 
concentrations of Aspin. Aspin fragments are found between 20-
30kDa and at 6.5kDa (personal correspondence, Richard Shaw, 
AgResearch, Upper Hutt, New Zealand). Lane 1 shows the low 
molecular weight marker. .. .... ... ............... ....... ... ....... .... ... ...... .................. ......... 82 

Figure 17: Effect of various concentrations of L-arabinose on the cell 
specific total protein titre. Fourteen 250mL shaker flasks were 
inoculated with 1mL of overnight seed culture in 25mL of Terrific 
Broth. Cultures were incubated for 4 hours at 37°C and 180rpm in 
a shaking incubator. The temperature was reduced to 30°C and the 
cultures induced in duplicate with 0.05, 0.1, 0.5, 1, 2.5, 5, 7.5, 1 Og/ L 
of L-arabinose respectively. Cultures were tested for total protein 
production 12 hours after inoculation ................ ..... ...... ...... ... ..... ............. ..... 93 

Figure 18: Effect of various concentrations of L-arabinose on the soluble 
and insoluble titre of AJpin. Fourteen 250mL shaker flasks were 
inoculated with 1mL of overnight seed culture in 25mL of Terrific 
Broth. Cultures were incubated for 4 hours at 37"C and 180rpm in 
a shaking incubator. The temperature was reduced to 30°C and the 
cultures induced in duplicate with 0.05, 0.1, 0.5, 1, 2.5, 5, 7.5, 1 Og/ L 
of L-arabinose respectively. Cultures were tested for soluble and 
insoluble protein production 12 hours after inoculation .............. ... .... .... .. 95 

Figure 19: E ffect of temperature on the specific soluble and insoluble titres 
of Aspin. Twelve 250mL shaker flasks were inoculated with 1mL of 
overnight seed culture in 25mL of Terrific Broth. Cultures were 
incubated for 4 hours at 37°C and 180rpm in shaking incubators. 
The culture temperatures were adjusted in duplicate to 10, 16, 19, 
23, 27, 37°C respectively and induced with 2g/L L-arabinose. 
Cultures were tested for soluble and insoluble protein production 
12 hours after inoculation .. ..................... ........ ... ... ... .......... ......... .. ... .... .. ... .. ..... 96 

Figure 20: Fermentation output of various measured parameters of the 
second Aspin fermentation . Measured variables include; acid pump 
speed [%], air ~/min], base pump speed [%], carbon source pump 
speed [%], dissolved oxygen [%], inducer pump speed [%], pH, 
temperature rq and agitation speed [rpm]. The fermenter was 

111 



inoculated with 100mL of cells at OD-2 into 1.4L Terrific broth. 
An exponential feeding regime was begun after 2 hours with a 
solution of 315g/L of glycerol and 315g/L yeast extract at a rate of 
0.3h-1 according the formula outlined previously [87]. The DO was 
controlled at 30% through agitation between 200-800rpm and 
aeration 0.5-2 L/ min, temperature 37°C and pH 6.8. The culture 
temperature was reduced to 25°C after eight hours and induced 
with 2g/ L of L-arabinose ......... ...... .......... .. ...................................................... 99 

Figure 21: E ffect of the post-induction growth rate on the soluble Aspin 
titre and inclusion body production for the fermentation cultures. 
All the fermentations were inoculated with lOOmL of overnight 
seed into 1.4L of Terrific Broth at 37°C, pH 6.8. The DO was 
controlled at 30% through agitation between 200-800rpm and 
aeration 0.5-2 L/ min except for two of the fermentations which 
used constant aeration of 0.8 L/ min and agitation at 350 and 
250rpm respectively. The cultures were all induced with 2g/ L of L­
arabinose after eight hours and harves ted after 24 hours. Three of 
the fermentations used an exponential feeding regime which was 
begun after 2 hours with a solution of 315g/ L of glycerol/ yeast 
extract at a initial rate of 0.3h 1 according the formula outlined in 
previously [87] ............................................... ...... .. .. ................................ ......... 101 

Figure 22: Fermentation output of ,-arious measured parameters for the 
fifth Aspin fennentation. The fermenter was inoculated with 
lOOmL of cells at OD-2 into 1.4L Terrific Broth. o pH or DO 
control was used. Agitation was held constant at 350rpm and 
aeration 0.8 L/ min. The culture temperature was initially held at 
37"C and then reduced to 25"C after eight hours and induced with 
2g/ L of L-arabinose ........................................................................................ 103 

Figure 23: E ffect o f resuspending cells to ,-arious optical densities prior to 
induction on the volumetric and specific titres of G-CSF. 200mL 
of enhanced YEHD medium was inoculated with lmL of frozen 
seed in a 2L shaker flask. This culture was placed in a shaking 
incubator at 30°C and 200rpm for 48 hours. The culture was then 
centrifuged at 3000g for 5 minutes. Pelleted cells were resuspended 
in sixteen 250mL shaker flasks to OD of 0.25, 0.5, 1, 2, 4, 8, 10, 12 
in YEHD medium with 0.5% (v / v) methanol replacing the 
dextrose. G-CSF production was measured three days after 
induction .................................................................................................... ....... . 125 

Figure 24: E ffect of various pH on the volumetric and specific titres of G­
CSF after 3 days. 25mL of enhanced YEHD medium was 
inoculated with 1 mL of seed in a 250mL shaker flask. This culture 
was placed in a shaking incubator at 30°C and 200rpm for 24 hours. 
The culture was then centrifuged at 3000g for 5 minutes. Pelleted 
cells were resuspended in fourteen 25mL cultures at an optical 
density of 5 in YEHD medium 'vith 0.5% (v / v) methanol replacing 
the dextrose and supplemented with 0.1M phosphate buffer. The 

lV 



cultures were adjusted in duplicate to a pH 3, 4, 5, 6, 7, 8, 9 with 2N 
sodiwn hydroxide and 2M hydrochloric acid. Cultures were grown 
for three days at 30°C, 200rpm with 0.5% (v /v) methanol added 
each day . ...... ........ ...... ................. ......... .. .... ...... ... ....... ........ ...... ... ..... ... ............. .. 127 

Figure 25: Effect of -rnrious concentrations of methanol on the Yolumetric 
and specific titre of G-CSF after 3 days. 25mL of enhanced YEHD 
mediwn was inoculated with 1rnL of seed in a 250mL shaker flask. 
Tius culture was placed in a shaking incubator at 30°C and 200rpm 
for 24 hours. The culture was then centrifuged at 3000g for 5 
minutes. Pelleted cells were resuspended in ten 25mL cultures to 
an optical density of 5 in enhanced YEHD medium with; 0.125, 
0.25, 0.5, 1.0, 2.0% (v/v) methanol replacing the dextrose. Cultures 
were grown for three days with the same amount of methanol 
added to the culture each day .............................. ......... .. ................. ..... ......... 128 

Figure 26: Comparison between chen'lical compounds in the basal salt 
medium with 150mL of dextrose feed and 200mL of methanol feed 
with the cellular composition of a yeast culture at 1 OOg/ L DCW 
[236) .... ...... ... ....... .. ........... .............. ........ ... .. .... ........ .. .... .... ... ..... ........ ........ ... ..... .. 134 

Figure 27: Typical P.pastoris fe1mentation profile of various measured 
variables. Acid/ base and carbon source (either glycerol or 
methanol) pwnp feed rate (1.5mL/%). Sparging airflow rate, 
culture dissoh-ed m .. ·ygen concentration as a percentage of air 
saturation, pH, temperature and agitation rate. 1250mL of basal 
salt medium was inoculated with 150rnL of oYenught seed at OD 
of approximately 5. Agitation was manually stepped up to 
1 OOOrpm to maintain the DO abo,-e 30%, aeration was constant at 
2 L/ min with pure oxygen blended at rugh cell-densities to 
maintain the DO above 30%. The culture temperature was 
maintained at 30°C and pH at 5.0. Fed-batch feeding of the culture 
was carried out according to protocol outline pre,-iously [21] . 
Cultures were grown till the glycerol in the initial medium was 
completely exhausted as indicated by a DO spike (~18 hours). A 
continuous feed of 50% glycerol containing 1.2% (v / v) of Pid1ia 
Trace Metal (PTiI1) solution was then started at 1.8% (v / v) for 
four hours to accumulate biomass. The glycerol feed was stopped 
for a half hour starvation period before induction. Induction was 
initiated using a continuous feed of methanol containing 1.2% 
(v/ v) of PTM 1 solution at an initial rate of 3mL/ h wluch was 
increased over 28 hours to 12rnL/ h at wruch level it was retained at 
for the remainder of the culture ....................................... ....... ................. ..... 136 

Figure 28: Equation for the calculation of the specific growth rate (µ). , is 
the cell number at time t, N 11 is the cell number at time 0. t is the 
time period between 0 and t . ..... ......... ..... .... ....... ........... .... .. .. .... .. .. ..... ........ ... 144 

Figure 29: E ffect of inoculation cell density on the growth rate of High 
Five cells, volumetric and cell specific expression of oFSH in 
Express Five. Six 250mL shaker flasks were inoculated with 0.4, 

v 



0.6 0.8, 1, 2 x 106 cells / mL in 25mL of Express Five. Cultures were 
then incubated at 27°C at 1 OOrpm for 48 hours. The cell density 
and recombinant protein production was measured after 48 hours ...... 152 

Figure 30: Effect of agitation on the growth rate of High Five cells, 
volumetric and cell specific expression of oFSH in Express Five. 
250mL shaker flasks inoculated with 1 x 106cells / mL in 25mL of 
Express Five. Cultures were incubated at 27°C and agitated at 80, 
100, 120, 140, 160 rpm respectively for 48 hours. The cell density 
and recombinant protein production was measured after 48 hours ...... 154 

Figure 31: Western blot of aggregated oFSH protein at 70kDa. Proteins 
were separated on a 13.5% SDS-PAGE gel. ......................... ...... .. ........... .. 155 

Figure 32: Profile of measured variables for the bioreactor culture of oFSH 
in High Five cells. The measured variables are; cell density [106 

cells / mL], pH, pure o:-..·-ygen in sparging gas [%] and culture DO 
[%]. 1L of Express Five medium was inoculated with 400mL of 
cells at 0.8 x 106 cells/mL. The DO was controlled at 50%, 
temperature 27°C, agitation 150rpm. Feeding with fresh media at a 
rate of 0.02h·1 was begun after 120 hours with supernatant removed 
to retain constant volume ............................................ .. ...... .... ....................... 157 

Figure 33: Profile of the specific and volumetric titres of oFSH for 
bioreactor culture of High Five cells. 1L of Express Five medium 
was inoculated with 400mL of cells at 0.8 x 106cells / mL. The DO 
was controlled at 50%, temperature 27"C, agitation 150rpm. 
Feeding with fresh media at a rate of 0.02h·1 was begun after 120 
hours with supernatant removed to retain constant volume ................... 160 

V1 



LIST OF TABLES 

Table 1: Production characteristic of E.coli expressing EfJS grown on 
various media. Media included; Terrific Broth (TB), Super Broth 
(SB), S.O.B, Luria Broth (LB) and M9 Minimal Medium (M9). 
1mL of inoculum was added to SOmL of each of the five media in 
separate 250mL shaker flasks. The cultures were then incubated at 
180rpm and 3 7°C. After 4 hours cultures were induced with 
0.1mM IPTG. Production characteristics were measured Sh after 
inoculation ... ...... ........ ....... ... ...... ... ........ ........... .. ..... .. ... .. ......... ....... .... ...... ............ 64 

Table 2: Effect of induction time on the final cell density, specific growth 
rate of cells, specific and volumetric titres of EfJS. lmL of 
overnight seed was added to eight 250mL shaker flasks containing 
SOmL of Terrific Broth. Duplicates cultures were induced at 0, 2, 4 
and 6 hours after inoculation with O. lmM of IPTG. All cultures 
were incubated at 180rpm and 37°C. Total protein, recombinant 
protein content and final cell density were measured after 8 hours ......... 71 

Table 3: Effect of dissolved oxygen concentration on biomass and protein 
production o f EfJS in fermentation cultures. Fermentations were 
inoculated with 100mL of seed OD~1.S in 1.4L of Terrific Broth. 
The dissolved o::-..7gen (DO) was controlled at 30% , pH at 7.0 and 
temperature at 37°C. All fermentation cultures used a feed solution 
of 315g/ L of glycerol and 315g/ L of yeast extract. Feeding was 
controlled by the automated program BioCommand (New 
Brunswick Scientific). Feed was supplied initially at a rate of 
1 SmL/ h, increasing by 0.1 SmL/ h for every minute above their 
respecuve set point. Fennentations were induced with O. lmlv1 
IPTG after 4 hours and maintained until two sequential reductions 
in cell density were measured as indication the culture was going 
into stationary phase .............. ....... ... ............. ............. ... ...... ........................... .. .. 7 4 

Table 4: Relative biomass, total protein and specific Ej)S production using 
various feeding regimes. Fermentations were inoculated with 
lOOmL of seed, OD~1.S in 1.4L of Terrific Broth. Dissolved 
m..7gen (DO) was controlled at 30%, pH at 6.8 and temperature at 
37°C. All fermentation cultures used a feeding solution of 315g/ L 
of glycerol and 315g/L yeast extract. The DO-stat and pH-stat 
were controlled by the automated program BioCommand (New 
Brunswick Scientific). Feed was supplied initially at a rate of 
1 SmL/ h, increasing by 0.1 SmL/ h for every minute above their 
respective set point. For the exponential feeding regime the rate of 
feeding was calculated according to the equation described 
previously [93], with a desired specific growth rate of 0.1 Sh-1

• 

Fermentations were induced with 0.lmM IPTG after 4 hours and 



maintained until two sequential reductions in cell density were 
measured as indication the culture was going into stationary phase ........ 76 

Table 5: Comparison of the previous best production characteristics prior 
to this study with the production characteristic achieved in this 
study ..................... ... ......... ................... ................ ...................... ..... .... .. ..... ... ..... ... 77 

Table 6: Plackett and Burman trial assessing the significance of various 
environmental factors on the volumetric titre of Aspin . ............................ 89 

Table 7: F-test of the significance of various environmental factors on the 
final cell density, specific titre and immunological activity of Aspin 
produced ................................ ................................................... .. .... .. ..... .... ........ .. 92 

Table 8: Summary of growth conditions and production characteristics of 
the Aspin fermentations. The shaker flask experiment was 
inoculated with 1mL of overnight seed culture in 25mL of Terrific 
Broth and incubated for 4 hours at 37°C and 180rpm in a shaking 
incubator. At induction the temperature was reduced to 28°C and 
2g/ L of L-arabinose added. The production characteristics were 
measured after 12 h. All the fermentations were inoculated with 
100mL of overnight seed into 1.4L of Terrific broth except 
fermentation 4 which used defined E. coli medium. The DO was 
controlled at 30% through agitation between 200-800rpm and 
aeration 0.5-2 L/ min, temperature 37°C and pH 6.8. The culture 
temperatures were reduced to there respective post-induction 
temperatures after eight hours and induced with 2g/ L of L­
arabinose. For fermentations 2-4 an exponential feeding regime was 
begun after 2 hours with a solution of 315g/ L of glycerol and 
315g/ L of yeast extract at a initial rate of 0.3h-' according the 
formula outlined in previously [87). Fermentation 5 and 6 used a 
constant aeration of 0.8 L / min and agitation of 350 and 250rpm 
respectively ... .. .. .. ..... ...... ..... ..... .. .. .. .. .. .. ......... ........ .... ......................................... 105 

Table 9: Final biomass, specific and volumetric titres of G -CSF produced 
after 3 days on various media. Minimal Glycerol (l\1GY), Buffered 
tv1inimal Glycerol (BMG), Buffer Glycerol-complex medium 
(BMGY), Yeast E xtract Peptone Dextrose medium (YEPD), 
enhanced Yeast Extract Hy-soy Dextrose medium (YEHD) . 
Dextrose media (YEHD). Cultures were inoculated with 1mL of 
overnight seed at OD-10 in 25mL of media. All cultures were 
grown in 250mL baffled shaker flasks at 30°C, 200rpm in a shaking 
incubator for 24 hours. Each culture was then centrifuged at 3000g 
for 5 minutes and the pellet resuspended to an optical density of 5 
in the same medium with glycerol or dextrose replaced with 0.5% 
(v / v) of methanol. Cultures were grown for a further three days 
with 0.5% (v / v) methanol added each day ..................... ............................ 119 

Table 10: Plackett and Burman trial assessing the significance of various 
environmental factors on the volumetric titre of G-CSF ........................ 123 

Table 11: Fermentation protocols used for P.pastons G-CSF ............................. ... 130 

11 



Table 12: Plackett and Burman ttial assessing the significance of various 
environmental factors on the volumetric titre of oFSH. .. ............ .......... .. 150 

111 



LIST OF ABBREVIATIONS AND SYMBOLS 

AOX alcohol oxidase 

bla ampicillin resistance gene 

bsdblasticidin resistance gene 

BCA bicinchoninic acid 

BMG buffered minin1al glycerol 
medium 

BMM buffered minin1al 
methanol medium 

BMGY buffered complex 
glycerol medium 

BMMY buffered complex 
methanol medium 

CARE continuous absorption 
recycle extraction 

CDW cell dry weight 

DO dissolved oxygen 

DTT dithiothreitol 

EDTA ethylene diamine tetra­
acetic acid 

ELISA enzyme-linked 
immunosorbent assay 

F(t) flow rate of feed at time t (h-
1) 

GFP green fluorescence protein 

G-CSF granulocyte colony 
stimulating factor 

GST. glutathione-s-transferase 

HCI hydrochloride 

HIC hydrophobic interaction 
chromatography 

His6 polyhistidine epitope tag 

HRP horse radish peroxidase 

Ig G Immunoglobulin G 

IMAC immobilised metal 
chelating resin 

INF interferon 

IPTG isopropyl ~-D-

thiogalactopyranoside 

LB Luria-Bertani broth 

LPM litres per minute 

MBP maltose binding protein 

MGY minimal glycerol medium 

MM minin1al methanol medium 

MOI multiplicity of infection 



N aOH sodium hydroxide 

NMW nominal molecular weight 

OD optical density 

PID proportional integral 

derivative 

PTM P .pastons trace metal 

solution 

RBS ribosome binding site 

REC reverse phase 

chromatography 

RO reverse osmosis 

SB superbroth 

SD Shine-Dalgarno site 

SD S-PAGE sodium dodecyl 

sulphate polyacrylamide gel 

electrophoresis 

Sf Spodoptera fmgiperda 

SF substrate concentration in feed 

(g/L) 

Y x/s cell titre of substrate (g CDW/ g) 

µ specific growth rate (h-1) 

11 

SEC size exclusion 

chromatography 

S(t) substrate concentration at 

time t (g/ L) 

TB Terrific Broth 

TBS tris buffered solution 

TMB 3, 3', 5, S'-tetramethyl­

benzidine 

TRX thioredoxin gene 

TSB tryptic soy broth 

UF ultrafiltration 

V(t) reactor volume at time t (L) 

WCW wet cell weight (g/ L) 

X(O) cell concentration at time t 

X(t) cell concentration at time t 

(gCDW/ L) 

YEPD yeast extract peptone 

dextrose medium 



ACKNOWLEDGMENTS 

The author wishes to thank D r Robert Dempster (Group Leader 

Fermentation, AgResearch, Wallaceville), Professor Yusuf Chisti (Institute 

of Technology and Engineering, Massey University) for supervising this 

work and all the staff at the Wallaceville Animal Research Centre, Upper 

Hutt, are thanked for their guidance and support. 

11 



1 Literature R eview 

The supply of many eukaryotic proteins which have potential clinical or 

industrial use is often limited by their low natural availability. The advent of 

biotechnology and gene cloning has sought to provide an alternative supply 

for eukaryotic protems. Agresearch has extensive expenence ill 

biotechnology products designed to improve the welfare of animals. 

Scientists at the \Xlallaceville facility of Agresearch have developed a number 

of recombinant protein products which have the potential as animal 

pharmaceuticals. The objective of this research was firstly to maximise the 

titre of four recombinant proteins products to improve their profitability and 

marketability to potential manufactures. The research then looks at 

developing parameters for the expression of those proteins in a small scale 

bioreactor to aid the possible transfer of the technology to higher volwne 

production. Finally tl1e research serves to provide a basis of fermentation 

knowledge to aid the development of other recombinant protein products. 

The recombinant protein products investigated included a vaccine antigen 

(EtJS) produced as inclusion bodies in E. coli, a aspartyl protease inhibitor 

homologue (AJpin) produced in E.coli, a secreted cytokine (G-CSF) produced 

in Pichia paslonj· and a secreted gonadotropin (oFSH) produced in High Five TM 

insect cells. 

1.1 HOST CELL LINE 

Three different types of micro-organism hosts are investigated in this 

research. They are EschendJZa coli, Pichia pastoris and the insect cell line High 

1 



Five™. The host cell line appropriateness is dependant on a number of 

factors including the price and form of the product, protein structure/ size 

and whether post-translational modifications such as glycosylation are 

required. 

Escherichia coli is one of the most commonly utilised organisms for producing 

recombinant products. This can be mainly attributed to the wealth of 

knowledge of the organism's biology, ability to grow to high-cell-densities on 

inexpensive media and the robustness of its growth. H owever the use of 

E.coli has many drawbacks including; the production of endotoxins, difficulty 

in secreting proteins, improper folding, and lack of post- translation 

modification [1, 8-15]. E ukaryotic proteins expressed in E.coli also often form 

inclusion bodies which add a number of steps to the downstream processing, 

making recovery more expensive and complicated [10, 16-24]. 

In contrast, the yeast Pichia pasloris can grow to high cell concentrations on 

mexpens1ve chemically defined media, can secrete proteins extracellularly 

simplifying downstream processing and has the ability to perform some 

eukaryotic post-translational modifications [4, 25-28]. H owever, proteins 

produced in P pasloris are not always folded correctly, may be over 

glycosylated and titres are variable with reports o f recombinant protein levels 

vary from milligrams to grams per litre [4, 26]. P pasloris also grows more 

slowly than E.coli witl1 typical culture time o f days instead o f hours [29]. 

Insect cells can perform complex post-translation modifications, produce 

high active protein concentrations and most proteins are secreted simplifying 

purification [30]. However, insect cells are extremely delicate, require high 

2 



seeding ratios, can be infected by mammalian viruses and are significantly 

slower to grow than bacteria and yeasts cells [23, 31, 32]. The most common 

insect cell line used is Spodoptera fmgiperda (Sf9) which is derived from the 

pupal ovarian tissue of fall army worms. High Five™ is a cell line which 

originates from the ovarian cells of the cabbage looper Tnd10plusia ni. High 

Five™ cells have a typical doubling time 18-24 hours, can be grown in 

suspended culture in serum free media and produce 5-28 fold more protein 

than Sf9 [31-33]. However insect cell media are generally more complex and 

expensive than prokaryotic media, cells do not grow to high cell densities and 

protein production is generally in tl1e milligram per litre range [32, 34-37]. 

1.2 PLASMID EXPRESSION 

1.2.1 Expression Vector 

The construction of express10n vectors reqwres several elements whose 

configuration must be carefully considered to ensure high levels of protein 

synthesis [38]. A typical E sdJerichia coli expression vector is shown in Figure 1. 

The gene of interest is preceded by a promoter which binds RNA polymerase 

initialising transcription. A large number of promoters are available for E. coli 

with the most commonly used being the lac, tac, T7, PL(A.) and araBAD. 

Desirable traits for a promoter include that it leads to the accumulation of 

high levels of protein and is tightly regulated to allow the growth of cultures 

to high densities with minimal protein expression. The lac and tac promoters 

3 



are considered weak promoters with only moderate protein titres and a high 

basal expression [13, 39]. In contrast the araBAD, T7, and pL(A) promoters 

are considered relatively strong promoters with high levels of protein 

production and low levels basal of expression [22, 40, 41]. The lac and 

araBAD systems are repressed in the presence of glucose [22]. Other 

desirable traits include the promoter needs to be easily transferable to allow 

the testing of many strains for protein titres. If the product is destined for 

large scale production then the simplicity and cost of induction must also be 

considered. The pET102 plasmid shown in Figure 1 contains the T7 

promoter controlled by the lac operator (lacO) which binds a repressor protein 

(lad) in the absence of isopropyl ~-D-thiogalactopyranoside (IPTG) 

preventing transcription [13]. IPTG is commonly used in laboratory 

experiments, however is costly at large scale [1, 42-44]. Less expensive 

options for large scale production include the PL(J...) promoter which is 

thermally induced or the araBAD promoter which is induced using L­

arabinose [16, 45, 46]. Following on from the promoter is the ribosome 

binding site (RBS) which consists of a Shine-Dalgamo (SD) site which 

interacts with the rRNA during translation initiation and a translational spacer 

before the start codon of the gene of interest [47]. The RBS plays an 

important role in the efficiency of transcription initiation. Surrounding the 

gene of interest the pET102 plasmid contains the coded sequences for the 

fusion partners HP thioredoxin, EK recognition site, VS epitope and His-6 

genes which are discussed in the next section. Downstream of the gene of 

interest a transcription terminator is located to stop transcription and also to 

protect the newly created mRNA. A transcription terminator consists o f a 

region of amino acid symmetry which protects the mRNA from 

exonucleolytic degradation by creating a loop structures in the RNA. Figure 1 

shows the plasmid also contains an ampicillin resistance gene (bla). A gene 

that confers antibiotics resistance is used to provide selective pressure on 

plasmid containing cells. Common antibiotics used in E.coli include 

ampicillin, kanamycin, tetracycline and chloramphenicol. Finally the origin of 

4 



replication (ori) gene determines the plasmid copy number. The use of high 

copy numbers has been shown to in some cases lead to higher titres of 

recombinant protein [48-50). H owever higher plasmid copy numbers have 

also been observed to lead to lower cell viability [22, 51). 

pET102/D/ JacZ 

Figure 1: pl:T1 02 vector map showing the \'a riou,; gene,; (>ource I nvitrogcn). The lac'/. gene is ,;hown 
here a,; the gene o f interc>t. The ribo,;ome binding site (RRS) proceed ,; the gene of interc,; t which i,; 
expre,;,;ed with fusion partners I IP thiorcdoxin , l:K recognition. \ '5 epitope and poly-histidine 
domains. The plasmid contains the T7 promoter regulated b y the operator (/mO). ,\ T7 translational 
terminato r stabilises the mlL'\! .\ and ampicillin re,;istance gene (bla) which aids clone selection . . \n 
origin of replica tio n (ori) determines the plasmid copy number. 

Pidna pastoris is a methylotrophic yeast capable of utilising methanol as its sole 

source of carbon. The first step in the metabolism of methanol is the 

oxidation to formaldehyde by the enzyme alcohol oxidase (AOX). The 

promoter regulating the production of AOX can be used to control 

recombinant protein expression in P .pastoris. The AOX promoter is tightly 

regulated in the presence of methanol and strongly repressed in the presence 

5 



of glucose [5, 26, 52, 53]. Two genes AOX1 and AOX2 code for alcohol 

oxidase [54, 55]. Isolation of these genes has been utilised to produce two 

different strains, Mut+ which contains the AOX1 gene and Mut5 which 

contains the AOX2 gene [56]. The AOX1 gene is responsible for a majority 

of the AOX production in P.pastoris and produces fast methanol metabolising 

strains while AOX2 isolates are much slower at metabolising methanol and 

produce recombinant protein more slowly [54]. A vector map of the pPICZ 

plasmid which was used in this research for the expression of granulocyte 

colony stimulating factor (G-CSF) is shown in Figure 2. Transcription is 

controlled by the AOX1 promoter and the gene of interest cloned into the 

multiple cloning site (Stu 1, EcoR I, Pm/ 1. .. . ). The gene of interest is 

expressed fused to the c-myc epitope and poly-histidine tag (6xHis). The 

pPICZ plasmid contains a resistance gene for the antibiotic Zeocin TM to aid 

the selection of transformed cells. The EM7 and TEF1 promoters 

constitutinly drive the Zeocin ™ resistance gene. In Pilhia pasl01is the 

recombinant plasmid can be incorporated into the genome by homologous 

recombination. Incorporating the gene of interest into the organism genome 

has been shown to improve clonal stability [55]. A CYC1 translation 

terminator is used to aid efficient mRNA processing and the pUC origin of 

replication allows maintenance and replication of the plasmid when used in 

Eslherichia coli. Recombinant protein titres using the pPICZ vector vary 

between 0.007 g/ L to 2 g/ L [57-62]. 

6 



>ases 1-94 1 

pPICZA,B,C 
3.3 kb 

• 

h gure 2: pPI CZ vector map showing the various genes (source lnvitrogen). The pPI C/'. vector was 
used fo r the expre,;sion o f colony granulocyte stimulating factor (G-CS I'). The gene o f interes t was 
cloned into the various cloning sites (Stu 1, EcoR I, Pml I. .. . ) . Expressio n is under the control o f the 
alcohol oxidase (. \ O X1 ) promoter and the gene cxprc,;sed with the fu sion partner,; myc-epi tope and 
poly- histidine tag to aid pro tein detection and purification. T he plasmid contains the TI ·: F1 and lo.\17 
p romoter> which control the expression of the /'.cocinTM re,;istance gene. The ,\ O X1 translational 
terminator stabilises the mRN ,\ and the p UC origin of repl ica tion (o ri) aids replication and 
maintenance o f the plasmid in E. coli. 

Production of recombinant pro tein in insect cells is initiated by the infection 

o f cells with a baculovirus containing the sequence of D NA for the 

recombinant pro tein. This research uses the pMIB vector from the 

baculovirus Org;yia pseudotsugata multicapsid nuclear polyhedrosis virus 

(O pMNPV) to express ovine follicle stimulating hormone (oFSH). Shown in 

Figure 3 the pMIB vector contains the OpIE2 promoter which provides 

constitutive expression of the gene of interest. The gene o f interest is 

expressed with a poly-histidine (6 x histidine) tag and a VS epitope to simplify 

purification and detection. A honeybee melitten secretion signal sequence 

directs the secretion of the protein of interest into the culture medium. The 

plasmid also contains genes that confer resistance to the antibio tics blasticidin 

(bsd) and ampicillin (bla). The antibiotic resistance genes are constitutively 

7 



expressed under the control of the EM7 and OpIE 1 promoters. The plasmid 

is transfected into insect cells using lipid-mediated transfection with stable 

cloned cells selected using blasticidin. 

-CD -> !.... • 

+u+t• ~i~~~~~i~~~~~~ taww®~m e 

:>r pMIBN5-His A 

pMIBNS-His 
A,B,C 

3.6 kb 

h gure 3: p\!l B vector map ,; howing the variou,; genes (rnurcc I nvitrogen). The p\!IB vector was 
u,;ed for the expression of ovine follicle ,;timulating factor (oFS I I). The gene of intere,;t was cloned 
into the variou,; cloning ,;ite,; (Sph 1, Hind 111 , Asp718 1. . .. ). Expre,;,;ion i,; under the control of the 
Op! E2 promoter an<l the gene expre,;sed with the fusion partners \ '5 epitope an<l poly-hi,;tidine tag. 
The plasmid contain,; the Op! I \ 1 and I ·: \ 17 promoters which control the expre,;,;ion of the bla;;ticidin 
and ampicillin re,;i;;tancc gene. The ,\ OX1 translational terminator stabili,;e;; the mRN .\ and the pUC 
origin of repEcation (ori) aid;; replication and maintenance of the plasmid in I ·:.coli 

1.2.2 Fusion Partners 

Recombinant proteins are often fused to one or more proteins "fusion 

partner" to aid purification, identification and/ or to improve solubility. 

8 



One group of fusion proteins consist of small lengths of peptides that create 

antigenic sites (epitope tags). Commonly used epitope tags include poly­

histidine (His6), green fluorescence protein (GFP), c-Myc and VS. The His6 

epitode tag is probably the most widely used fusion partner in molecular 

biology [63] . The His6 tag binds to immobilised metal ion chelating (Il\1AC) 

resin under mild solvent conditions providing a quick and relati,·ely 

inexpensive mode of purification compared to other affinity chromatograph 

techniques [64]. Its small size (si.x amino acids) reduces the possibility of 

steric hindrance and antibodies to the His6 tag can be used to identify the 

recombinant protein quickly. The epitope tag GFP produces a fluorescence 

signal which allows for the expression of recombinant protein to be measured 

using a spectrometer [6S]. VS and c-Myc are both highly antigenic tags for 

which commercially available antibodies can be used to identify, immuno­

precipitate or immuno-affinity purify the protein. 

Another group of fusion partners has been developed to improve the 

solubility of expressed recombinant proteins. PreYiously it has been found 

that fusing recombinant protein to highly soluble proteins improves the 

m·erall solubility of the fused protein [66]. The fusion partner glutathione-S­

transferase (GST) is a highly soluble protein deriYed from SdJislosoma j aponicum 

which has been \videly used to improve the solubility of eukaryotic proteins 

expressed in E.coli [9, 13, 66]. However GST large size (1 Sk:Da) can hinder 

proper folding of the recombinant thus effecting their activity [67] . For this 

reason fusion proteins often contain a sequences such as the Xa site which 

allows the fusion partner to be cleaved once the protein has been purified 

[67] . Glutathione-sepharose resin can used to purify GST fused protein and 

commercial antibodies to GST are available. It has been suggested previously 

that the insolubility of many eukaryotic proteins may be due to the redox state 

in the E.coli cytoplasm, as it tends to be more oxidative than in mammalian 

9 



cells [8]. The fusion partner thioredoxin has been used to alter the redox state 

of bacterial cytoplasm increasing recombinant solubility [8, 11, 66-68]. 

Maltose-binding protein (MBP) is another common highly soluble protein 

that has been found to improve translation efficiency and solubility of 

eukaryotic proteins expressed in E.coli [67-69]. Both thioredoxin and MBP 

have commercially available resin which can be used to purify these fusion 

proteins [38] however hybrids are now available which contain His6 sequence 

allowing less expensive IMAC purification [67]. 

1.2.3 Indudion 

Recombinant protein production in E.coli fermentations has been reported to 

be proportional to the specific growth rate at induction [39, 70, 71]. Inducing 

at the maximum growth rate has the advantage that cell metabolism is at it 

highest leading to a high cell specific protein production. However, several 

other studies have found that inducing the cells as late as possible in the 

exponential growth phase led to the higher expression as higher cell densities 

could be obtained [72-7 4]. A balance between cell density and cell specific 

titre which produces the maximum volumetric titre needs to be empirically 

determined as the optimum is likely to be a result of a combination of the 

recombinant characteristics. Care must be taken if the second approach is 

favoured as induction at high cell densities can shock cultures into stationary 

phase early [75]. Generally cultures are not induced in stationary phase as 

culture viability is reduced and a large number of proteases are produced 

which can degrade the protein of interest [76]. 

10 



For the production of soluble protein in E.colz~ cultures are induced early in 

the exponential phase to slow the culture growth rate. It has been extensively 

reported that the production of soluble proteins in E.coli is inversely related to 

the specific growth rate of the cultures [4, 22, 73, 77-79]. At high growth 

rates protein expression can overwhelm chaperones and foldases required for 

protein folding. This leads to the accumulation of protein intermediates 

which may form inclusion bodies [10, 66, 80, 81]. However, again a balance 

must be achieved as inducing the culture earlier reduces the potential to 

achieve high-cell-densities [9]. Another approach used to increase soluble 

protein production is to reduce the concentration of inducing agent [81]. 

Under the control of a strong promoter the concentration of inducing agent 

can be used to slow the rate of recombinant protein expression. Lim et al. 

[45] found using 10.7m1Vf of L-arabinose only 10% of interferon-ex (I F-o:) 

produced in E.coli was soluble but by reducing the concentration of L­

arabinose to 2.7mM the fraction of soluble IFN-o: increased to above 90% . 

In Pichia pastons cultures methanol serves both as a carbon source and an 

inducer agent. During induction the concentration of methanol must be 

maintained within a relatively narrow range to avoid growth and expression 

inhibition. Guarna et al. [52] observed that recombinant protein production 

decreases above 1.0% (v /v) methanol while other researchers have observed 

growth inhibition at concentrations above 3% (v / v) [5, 52, 82]. Both these 

limits are only slightly higher than the 0.5% (v / v) methanol concentration 

suggested in the Pichia Expression Kit (Invitrogen, San Diego, CA) to initiate 

induction. Probes that measure the residual methanol concentration are 

available however in their absence shaker flasks are generally induced with 

0.5-1.0% (v / v) metl1anol per day to avoid inhibition [5, 52, 82-86]. In high­

cell-density fermentations the rate of methanol consumption can vary 

between 0-10 % (v/v) per hour [87] . To avoid the accumulation of methanol 

11 



periodic starvation periods are used, in which methanol feeding is halted until 

methanol depletion occurs as indication by a rapid increase in the dissolved 

m.7gen concentration [5, 52]. 

Cell density has been shown to have a dramatic effect on the production of 

recombinant protein in Tnd1oplusia ni cell line with specific recombinant 

protein production decreasing with increasing cell density [31, 34, 35, 88]. 

Wickham et al. [31] reported a 6-fold reduction in specific production of ~­

galactosidase with an 8 fold increase in cellular concentration. For this reason 

infection of Tni cells is generally performed at low densities. It has been 

suggested that cellular contact may inhibit the production of Yiral D A [31] . 

Another important aspect in considering when and how to induce is the 

degradation of recombinant protein. Recombinant protein production 

tnggers stress responses such as the activation of proteases which may 

damage or destroy the protein of interest [48]. Ramirez and Bentley [75] 

found that at high rates of induction cells produce more proteases. By 

reducing the rate of induction either by reducing the amount of inducing 

agent, inducing at lower cell-densities and over a longer period culture 

metabolic stress is reduced consequentially reducing the production of 

proteases. 

1.2.4 Plasmid S tabili(y 

Instability of recombinant plasmids remains one of the most important 

problems in the commercial production of recombinant products. The 

instability is affected by both genetic and environmental factors including tl1e 

properties of the plasmids themselves, degree of expression, cultivation 

12 



temperature, nutrient levels, growth rate and fermentation mode [89]. Culture 

age increases with fermenter size increasing the chance of reversion to wild­

type strains. Unfamiliar conditions of pH, temperature or substrate 

concentration place added pressures on the organism [90]. In fermentations a 

distribution of cell copy numbers exist (Gaussian distribution). Cells with 

lower copy numbers exert less metabolic pressure and therefore have a 

greater long term stability (51]. Yeast divide by budding therefore are even 

more likely to form plasmid free cells [91 ]. To increase plasmid stability often 

antibiotics are added to a culture or a amino acid dependant strain is used to 

provide selective pressure on the recombinant population [49]. However, 

both these methods add expense and some manufacturers do not allow the 

use of antibiotics. Other methods include the careful selection of appropriate 

plasmids and their constructs to reduce metabolic stress [50]. 

13 



1.3 PROCESS CONDITIONS 

During fermentation, nutrient concentration, agitation, temperature, dissolved 

m.7gen and pH can be utilised to increase titres and protein express10n. 

Recombinant organisms generally show far more sensitivity to process 

conditions and require m ore accurate control than wild-type strains [89]. 

1.3.1 Media Composition 

All micro-organisms require a source of carbon, nitrogen, sulphur/ phosphate 

and a variety of trace minerals, vitamins and growth factors for optimal 

growth and protein synthesis. 

Glucose has the highest energy potential of all carbohydrates and is the 

preferred carbon source for most micro-organisms [92]. Other common 

primary carbon sources include glycerol, lactose, sucrose and in the case o f 

P paston·s methanol. E.coli grown on glucose has been observed to grow more 

quickly and produce higher titres of foreign protein compared to o ther 

carbon sources [46, 78, 93, 94]. However the excess supply of glucose causes 

metabolic overflow and the production of undesirable by-products such as 

acetic acid [20, 95]. Acetic acid production by E.coli has been extensively 

studied and shown to be a major inhibitor of both growth and recombinant 

protein production [72, 74, 96-102]. The choice of carbon source is also 

important in the control of plasmid expression as both the lac and araBAD 

14 



promoters are repressed in the presence of glucose and the AOX1 promoter 

by both glycerol and glucose [26, 56, 87, 103]. Prior to induction, the 

inhibitor substrate must be either removed by replacing the medium with one 

free of glucose or sufficient time allowed for the culture to consume all the 

substrate [4, 104]. Extra care must be taken with chemically defined media as 

induction/ repression ratios have been found to be several magnitudes higher 

than for complex media [22]. 

Nitrogen can be supplied either from a complex source such as a protein/ 

vegetable digest or from a synthetic source such as ammonium sulphate, 

ammonium hydroxide or ammonium chloride. In the production of nccines 

a synthetic nitrogen source is often used in preference to a complex source 

due to concerns over the transmission of disease and variation between 

batches. Generally growth is slower and recombinant protein titres lower 

us111g synthetic nitrogen sources as cells must synthesize their own amino 

acids [105, 106]. Synthetic media supplements which combine complex 

mixtures of trace metals and growth factors are available to supplement 

media; however, generally yeast extract is added to cover all the minor 

nutrients required for cellular growth [107]. The source of sulphur is usually 

drawn from the available amino acids in complex media or as part of sulphate 

salts in synthetic media [108]. Korz et al. [98] found that to increase E.coli to 

high biomass titres phosphate was required to be supplemented. Phosphate is 

required by cells for the synthesis of D A, RNA, protein and in cellular 

respiration [109]. Phosphate is often made available through the use of a 

phosphate buffer system in the medium. Medium buffering is important to 

prevent pH fluctuations which may cause culture stress and destabilise 

secreted proteins [11 O]. Low concentrations of potassium and sodium are 

provided by the addition of sodium chloride and potassium chloride [70]. 

These salts also help balance the osmolarity between the cells and the 

15 



medium. Antibiotics such as ampicillin, kanamycin, chloramphenicol and 

gentamycin are often added to the medium to provide selective pressure on 

the recombinant to retain its foreign genes and to provide limited protection 

from contamination by other organisms [38]. 

EschendJZa coli is not fastidious and will grow on almost any medium. For the 

growth of high-cell-density cultures of recombinant E.coli one of the most 

common medium used is Luria-Bertani (LB) broth [18, 73, 111]. Other 

common media include M9 minimal medium [45, 75), Terrific Broth (TB) 

[45] and Super Broth (SB) [112]. 

Pichia pasloris can also be grown on a wide range of media. Cultures are 

initially grown in media containing glycerol or dextrose then transferred to an 

equivalent medium containing methanol. Conunon media include; MGY / 

MM (minimal glycerol or minimal methanol medium) [27, 53, 84, 86], BMG/ 

BMM (buffered minimal glycerol or methanol medium) [26, 52, 82, 83, 113-

115], BMGY / BM~fY (buffered complex glycerol or methanol medium) [5, 

26, 82-84, 86, 115] and basal salt media which are commonly used in large 

scale fermentation [5, 52, 114-117]. 

Insect cells have much more complex nutritional requirements compared to 

those of bacteria or yeas t cells. This is reflected in the large nwnber of 

components in the insect culture media. Additional to the basal salt media 

which provides salts and trace elements, yeastolate, lipid-sterol emulsions and 

shear protectants are commonly added. Yeastolate is an autolytic digest of 

yeas t which serves as a source of amino acids, vitamins and nucleotides. A 

lipid-sterol emulsion is added to provide cholesterol which insect cells can not 

16 



synthesize, essential fatty acids and cx-tocopherol which is an antioxidant. 

Pluronic F-68 is a common shear protectant used to protect the cell from 

damage in suspended culture. A nwnber of media optimised for the 

cultivation of High Five cells are commercially available including; ExCell 405 

GRH Bioscience), Express Five (Life Technologies), IS BAC (Irvine 

Scientific), and CCM3 (Hyclone). 

1.3.2 Feeding Regime 

The traditional batch fermentation requires the entire nutrient allocation to be 

added at the beginning of the fermentation. This leads to large concentration 

gradients which can cause cellular stress and may inhibit growth. Other 

techniques such as continuous culture, perfusion and fed-batch culture 

overcome these concentration limits by adding nutrient thtoughout the 

fermentation as they are consumed. In continuous fermentation, whole broth 

is withdrawn from the fermenter throughout the culture and fresh medium 

added to supplement the volume. Continuous culture uses fermenter time 

more efficiently as production can be maintained for long periods making it 

appropriate for products which are required to be produced in large volumes. 

However, continuous culture is not widely used in pharmaceutical 

manufacture as production volumes are generally too small to warrant the 

added expense and also due to concerns over the long term clonal stability of 

recombinant products [69]. Perfusion is a technique where broth is 

continuously withdrawn from a fermenter thtough a filter which retains the 

cells [34, 79, 118]. The filtered bro th is collected and fresh medium added to 

supplement the culture. Using perfusion culture it is possible to obtain cell 

densities 10 to 30 times higher than the maximum cell density in batch culture 

[119]. Perfusion however is not widely used in industry due to difficulties in 

17 



maintaining large scale operations. To overcome these problems most small 

scale fermentations incorporate aspects of both continuous and batch culture 

by using fed-batch systems [73, 100, 120-123]. In fed-batch cultures, growth 

limiting nutrients are added as they are consumed to maintain growth for 

extended periods. The addition of nutrients is controlled either by an 

empll.1.cal formula (constant and exponential feeding being the most common) 

or in response to process measurements (chemostat, pH-stat, DO-stat) [73]. 

Chemostats relies on the monitoring of substrate or metabolite concentration 

to control culture feeding [7 4, 108, 124, 125]. Common compounds 

monitored include acetate, glucose, phosphate or the off gas makeup [98]. A 

Yessel mass balance or empirical relationship is then used to relate the amount 

of feed required to achieYe a desired growth profile [9]. On-line measurement 

can significantly increase the cost and complexity of production and is 

therefore not extensively utilised for small scale production. A pH-stat 

feeding system monitors the pH of the fe1111entations for indications of acetic 

acid production [69, 71, 74, 126, 127]. The feed is increased until the 

maximum growtl1 rate is reached as indicated by metabolic overflow causing a 

decrease in tl1e pH [72, 101]. The pH-stat has two main ad,·antages m·er 

chemostat in that it does not require additional monitoring equipment as a 

pH probe is a standard accessory of most fermenters and in that it ensures 

that the growth rate is always close to its maximum. One of the most 

common feeding systems is the DO-stat [39, 44, 74]. The DO-stat regime 

utilises the fact specific m .. ·-ygen uptake reaches a maximum at the maximum 

growth rate [96]. Using pulses of feed tl1e response of the organisms oxygen 

uptake is used to detel111ine whetl1er the organism is at its maxin1um growth 

rate [72]. The popularity of the DO-stat system can be mainly attributed to its 

simplicity and the quick response time (response observed in less than 30 

seconds in E.coli cultures [96]). However, DO-stat systems are liable to 

18 



control distortions caused by changes in the volumetric oxygen transfer 

coefficient. Changes can be caused by a number of things including DNA 

release, antifoam addition, protein production and agitation [94]. Online 

feeding regimes such as DO-stat and pH-stat are particularly useful during 

development of fermentation processes as they are easy to implement and 

adaptive to change howe,-er can cause high cellular stress due to fluctuations 

in culture conditions [89, 121] . Exponential and constant feeding regimes use 

empirical relationships to regulate the feed rate [25, 26, 74, 78, 93, 126]. In 

contrast to methods that use online measurement, considerable knowledge of 

the culture growth kinetics and a robust fermentation process are required to 

use these regimes. Accumulation or depletion of nutrients may occur if the 

relationship is a poor fit or if the sys tem undergoes an unanticipated change 

which affects the utilisation of substrates [72]. An advantage o f an 

exponential feeding regime is that it can be used to manipulate the growth 

rate of the cells and changes in substrate concentrations are gradual reducing 

cellular stress [126] 

1.3.3 Temperature 

Fermentation temperature affects the kinetics o f chemical reactions within the 

cells. At low temperatures organisms are sluggish and the growth rate is 

reduced as energy production slows. At very low temperatures the micro­

organisms may stop growing altogether. In contrast at high temperatures 

enzymes in the organism may be denatured and lose their facilitative ability. 

The optimal growth temperature of wild-type E.coli is 37°C [128]. The 

temperature of recombinant fermentations are often lowered to restrict the 

19 



expression rate [9, 12, 13, 48, 112]. Restriction of the expression rate is 

desirable as it has been extensively demonstrated that the production of active 

protein is increased at lower expression rates [12-15, 29, 81, 93]. 

Recombinant strains of E.coli have been found to have an optimal protein 

titre using post-induction temperatures between 20 and 30°C [4, 44, 48, 98, 

112]. Maintaining a lower temperature has also been found to increase the 

amount recombinant protein by reducing the activity proteases [9]. 

A majority of Ppasloris fermentations are controlled at 30°C [4, 5, 26, 27, 83-

87, 104, 113-117, 129, 130] . Tight control of temperature is required as above 

32°C protein expression is inhibited [5]. During methanol utilisation P pastoris 

can produce considerable amounts of heat. Coupled with the capability to 

grow to high cell densities, this means that large cooling systems are often 

required to control the temperature [117]. Reducing the fermentation 

temperature below 28°C has been reported to increase the activity of some 

proteins expressed by Ppastoris [5, 82, 117, 129]. Another advantage of 

lowering the temperature is that at the lower temperatures P pas/ons secretes 

less low molecular weight proteins reducing downstream processing [5]. 

Insect cell have been successfully grown over a range of temperatures 

between 25-30°C however the optimal temperature for cell growth and viral 

infection has been demonstrated to be 27°C [131, 132]. Reuveny et al. [133] 

found that insect cells produced similar titres of recombinant protein in the 

range 22 and 27°C, increasing the temperature led to earlier production and an 

increase in the proportion of recombinant protein secreted in the culture 

medium. Most recombinant protein productions in Tni are grown at 27°C 

[37, 79, 118, 134]. Above 30°C protein production and cell viability have been 

observed to decrease rapidly [133]. No literature was found that used lowered 

20 



temperatures to increase recombinant protein solubility or activity in insect 

cells. 

1.3. <-/ Ag it a lion 

Cells require a continuous supply of nutrients and oxygen for growth. The 

primary role of agitations in fermentation is to improve the transport of these 

materials in the medium. Agitation helps to create homogeneous conditions 

throughout the fermenter for consistency and control as well as improves the 

heat transport in the fluid. For organisms that produce toxic metabolites, 

agitation also helps to disperse toxins. 

During fermentation an operator needs to be aware of the fragility of the 

organism they are working with. E.coli and P.pasloris are relatively robust. In 

shaker flasks cultures of P.pasloris and E.coli agitation rates of between 150-

250rpm are commonly used [4, 5, 85, 115]. E.coli fermentation generally use 

speeds of between 600-1000rpm [74, 93, 97, 98]. High-cell-density cultures of 

P.pasloris require high agitation rates in excess of 1000rpm to attain the o>..7gen 

transfer rates required to maintain culture dissolved o:-..7gen levels [5, 113, 

116]. faintaining the dissolved oxygen is especially important during the 

expression of recombinant protein as P pastoris cannot grow on methanol in 

the absence of oxygen. 

Conversely, insect cells are very shear sensitive as they do not have cell walls, 

only a thin cell membrane to protect them against rupture [79, 90, 135]. A 

21 



balance must be established between the conflicting demands of nutrient 

transport and shear sensitivity. Wild-type Tni cell are adherent however 

commercially available strains such as High Five have been selectively 

passaged to obtain cells that can growth in suspension. Despite this, cells 

must still be slowly adapted to suspended culture by increasing the agitation 

rate over a number of passages [3 7]. Agitation speeds of between 100-

160rpm are used for Tni cultures in shaker flasks [30, 37, 88, 134]. In 

bioreactors marine propellers or fixed matrix supports can be used to reduce 

the shear on the cells [136]. At low agitations speeds step down gearing may 

also be required to provide accurate control of propeller movement [30]. 

Additives are often supplemented to insect cell media to protect cells agains t 

shear. Shear protecting additives including Pluronic® F68, poly (ethylene 

glycol), poly vinyl alcohol and foetal calf serum [137]. 

1.3.5 pl-l Conditions 

The pH of fermentations affects enzyme mediated reactions within the cell as 

well as the redox reactions of the cellular transport systems. Expressed 

protein folding and final conformation are also affected by the pH as different 

amino acid residues are exposed under various oxidative conditions. 

The optimal pH range for growth of E.coli is between 6.4 and 7.2 [109, 138]. 

The pH of E.coli fermentations are generally controlled through the addition 

of 2-3M sulphuric acid and 29% (w /v) ammonium hydroxide (ammonium 

solution) or SM sodium hydroxide [139] . In E.coli the solubility o f expressed 

proteins has been reported to be effected by pH [140] . Strandberg and 

Enfors [140] found that amount of recombinant protein expressed as 

22 



inclusion bodies increased with decreasing pH The specific acuv1ty of 

recombinant protein has an optimum pH during expression which maybe 

outside the range of pH for growth [109]. For this reason new recombinant 

proteins must be empirically tested over a range of pH with a balance sought 

between activity and bulk protein production. 

P.pastoris is able to grow satisfactory over a wide range of pH (3-7) [11 7]. This 

can prove to be useful in optimising expression conditions and protecting 

secreted proteins from proteolysis. During the accumulation of biomass 

P pasloris is generally maintain at pH 5.0, however post-induction the pH is 

often changed to optimise the conditions of protein expression [4, 115] . A 

number of studies have found that the titre of recombinant protein 111 

P pasloris fermentations can be increased by reducing the pH to between 3-5 

post-induction [4, 5, 83, 84, 87, 113, 115, 117, 141]. The increase in titre of 

recombinant protein at lowered pH has been attributed mainly to the 

inactintion of neutral proteases [110, 11 7]. In shaker flasks the same effect is 

achie,·ed by the use of unbuffered media such t fGY / MM which allows the 

pH to fall as cell metabolites such as ammonium accumulate in the culture [4-

6, 87, 11 7]. 

Insect cells are very sensitive to oxidative conditions. For this reason insect 

cell media is o ften buffered at 6.2 using a buffer system such as carbonate and 

bicarbonate. Under the carbonate/ bicarbonate system small changes in pH 

can be achieved through the sparging of carbon dioxide which applies 

pressure on the carbonate/ bicarbonate buffer equilibrium. In addition 0.1 M 

sodium hydroxide and hydrochloric acid can be used however the low 

intensity of agitation in cell cultures can cause localised acid and base pools 

that may damage the cells [142]. In adjusting the pH of fermentation media, 

23 



care must also be taken not to cause precipitation of medium components. 

Magnesium sulphate or chelating agents such as ethylene diamine tetra-acetic 

acid (EDTA) are generally added to synthetic media to prevent the 

precipitation of salts during fei_mentation [45, 71, 101]. 

1.3.6 Dissolved 0-'rygen 

The availability of m.-ygen can affect the nature and rate o f metabolic 

reactions within cells. Due to the low solubility of Oh.)'gen in water, o:--.7gen 

transfer in large scale fermentations is often the main limiting factor in aerobic 

microbial growth [135]. 

During culturing, the operator generally has three methods of controlling the 

dissoh-ed oxygen concentration; by the agitation speed, sparging rate and the 

o:--.1'gen concentration in the sparged gas. Furthermore the dissoh-ed m .. -ygen 

may be indirectly controlled by regulating the oxygen uptake of the micro­

organism by either reducing the culture temperature or rate of feed. Se\'eral 

studies have found that a reduction in ox-ygen transfer leads to an increase in 

specific recombinant protein through tl1e reduction of culture growth rate 

freeing up cellular resources to protein expression [39, 94, 109, 143, 144]. 

E.coli is generally grown in aerobic conditions, as during anaerobic growth it 

produces a number of metabolic products including; acetate, succinate, 

formate, lactate, and ethanol. The production of these products reduces the 

available energy for otl1er process such as growth and protein synthesis [101]. 

Maintaining a residue ox7gen concentration is also important as a number of 

proteases which may degrade the protein of interest are produced under 

oxygen starved conditions [111]. 

24 



P paslon·s is capable of extremely high cell densities (greater than to 150g 

DCW / L) [4]. Oxygen is required for the first step of methanol catabolism 

therefore is extremely important to ensure that P pasloti.s grows on methanol 

[113] . Generally P.pasl01is fermentations are controlled at 30-35% of air 

saturation [5, 26, 87, 114, 117]. At high-cell-densities agitation speeds of up to 

2000rpm, 1 vvm of air and supplementation with pure oxygen maybe required 

to maintain the dissolve m .. ; rgen concentration [113, 115-117]. 

Due to a lack of cell wall insect cells are sensitive to shear from agitation and 

the sparging of gases [79]. As a result the rate of m .. ; rgen transfer in cultures is 

often the limiting factor in expression of recombinant protein [118]. To avoid 

damage from bubble caYi.tations, gas is often m·erlayed on the surface of the 

culture [145]. This dramatically limits the 0)...'}gen transfer rate so pure o:-..;·gen 

supplementation may be required to increase the solubilization driving force. 

Other methods to reduce the shear while increasing O)...)'gen transfer include 

the use of tubing which diffuse microscopic bubbles through the culture or 

fi.xed mattix supports which cells are adhered to pre,·ent contact.[136] 

25 



1.4 DOWNSTREAM PROCESSING 

Fermentation products are invariantly found in low concentration in complex 

and ill defined solutions. The goals of downstream processing include removal 

of unwanted impurities, bulk-volume reduction with concomitant concentration 

of the desired protein and the transfer of the product to a stable and active 

environment [136]. 

Figure 4 shows the general downstream processing steps for protein production. 

All recombinant protein purifications begin with the clarification of cells from 

the fe1memation broth. For extracellular secreted protein the pelleted cells are 

simply disposed and the protein of interest purified and concentrated from the 

culture supernatant. \Xlith recombinant proteins expressed intracellularly the 

pelleted cells are then be ruptured and the supernatant filtered or centrifuged to 

remm"e the cell debris. Denatured proteins must be solubilised before 

purification and concentration. Depending on the desired form tl1e product 

may then be lyophilised or packaged in liquid form. 

26 



Intracellular products I Bioreactor I Extracellular products 

• I I .. 
Biomass Removal 

Cell H arvest Ccnt:ri fugation 
Centrifugation i\ licrofiltration 
i\!icrofil tration Ultrafiltration 
Ultra filtration 

u I 

Cell Dis ruption • Homogenisation Initial Purification 
Bead mill ing - Precipitation 
Soni cation Salt 
ChcrnicaJ Polvm cr 

Solnom • J ·:xtraction 
!'oh-mer/ poll'mer 

Cell Debris Removal Polymer/ salt 
Centrifugation .\dsorption 
\ !icrofiltration C. \R E 
Ultra filtration fapandcd bed 

I 

Denatured • 
products Concentration ,, Ultra fi ltration 

Renaturation 
J·:,·aporation 

Solubilisation ... Rc,·crsc Osmosis 
~ Precipitation 

Reoxidation 
CrYsta!lisatJon Final Purification 
Extraction Chrorrntography 
.\dsorption G cl permeation 

Ion exchange 

I ~ 
I k drophobic I ntcraction 

~ Re,·erse phase 
.\ffuuty 

Diafiltration 
I ·:kctrodialysis 

Dehydration o r I electrophoresis 
Solvent Removal 
Spray drying -.... 
l'rccze drying 
Fluid bed drving 

,, 
I Final Product I 

Figure 4: General stages o f downstream processing fo r p rote in p roduction, adapted from Perry's 
Engineering I landbook 11361-

27 



1.4.1 Clarijication 

Clarification helps to reduce process volwnes and remove major contaminants 

which may clog the more sensitive/selective operations. Shown in Figure 4, at 

several stages in the downstream processing of recombinant protein products 

clarification is required. Clarification is used to separate cells from the 

fermentation broth after cell rupture to remove cell fragments and to change 

buffers solutions between purification steps. Cell separation processes are 

almost exclusively performed by the mechanical methods centrifugation and 

filtration [146]. The choice of method depends on the physical properties of the 

broth (temperature, pH, ionic strength), mediwn components (cells, polymers, 

polyvalent cations, presence of other particles) and on the desired final state of 

the product [147]. 

Filtration relies on the retention of particles by a membrane based on size, with 

the driYing force for separation created by the pressure across a semi-permeable 

membrane. Filtration is the best established and most versatile method of 

removing insoluble material when the particles are dilute, large and rigid 

howe,·er many biological suspensions are difficult and slow to filter as they 

produce gels [136] . Generally cross-flow operations are used as the tangential 

flow of materials helps to prevent build-up of gel layers on the membrane [147]. 

Filter aids can also be used to assist filtration however these often complicate 

downstream concentration and purification steps [146]. The two main types of 

filtration membranes used for cell/ debris remove are microfiltration and 

ultrafiltration [148, 149]. Micro filtration membranes are generally used for 

particles with average nominal molecular weights (NMW) greater than 500,000 

while ultrafiltration is used for proteins below 100,000 NMW [136]. The main 

advantage of filtration over centrifugation is that it produces a retentate with 

lower water content reducing process volwnes [150]. 

28 



Centrifugation relies on the enhanced secliinentation of particles of different 

densities under an externally applied centrifugal force. Centrifugation generally 

requires more expensive equipment than filtration but is more effective in 

removing small particles which tend to clog filters [136]. Centrifuges are 

especially useful for removing cells from fermentation broths, however the 

retentate produced by centrifugation generally has a higher water content than 

that produced by filtration [150] . Many different types of centrifuges exist, the 

most corrunon of which are the tubular and disk centrifuge, scroll conveyor and 

basket centrifuge [136, 151]. For the removal of E.coli and P.pasloris cells from 

fermentation broths generally cenui.fugal forces of 2,000-5,000g for 5-20 

minutes are used [15, 71, 140]. For insect or mammalian cells which are 

sensitive to shear lower centrifugal speeds of 200-1 OOOg are used to prevent cell 

rupture [30, 79, 88, 145]. If the protein is intracellular the removal of finely 

dispersed particle after cell disruption may require forces of up to 30,000g [9, 16, 

39, 71, 120]. During centti.fugation often a low concentration of denaturant 

such as Tti.ton X-100, deo:-..7cholate or urea is added to the wash buffer to 

prevent crude impurities adhering to the surface of the protein aggregates [12, 

17, 18, 152]. 

1.4.2 Ce// Disruption 

The release of intracellular proteins is achieved through the disruption of the cell 

walls of the organism using; mechanical, non-mechanical, biological, or chemical 

lysis. 

29 



Chemical 

.\ci<l 

.\lkaline 
Soh-cnt 
Duergcnt 

Biological 

lonzyrrntic 
Phage 
.\ntibiotics 

Cell Disruption 

Physical 

Mechanical 

I ligh pressure 
homogenisation 
Bca<lmill 
Sonication 

Non-mechanical 

Osmotic shock 
Freeze thaw 
Sprar <lrying 

foigure 5: Techniques o f ceLI di>ruption :;howning the variou:; chemical , biological and physical method:; 

Physical clismption methods include high-pressure homogenisation, sorucation 

and bead milling. On the large scale homogenisation is the most common 

method of cell dismption [153]. High pressure homogenisation is a liquid shear 

method which utilises rapid changes in pressure and velocity as the cells pass 

through a small orifice to mpture cells [9, 154, 155]. Operation at high 

pressures is desirable for effecti,-e dismption with units working at pressures of 

up to 1200 bar [156]. The use of high pressures is limited by the resultant 

temperature rise in the liquid across the vah-e and by the avaliability of materials 

to prevent erosion of the valve [157]. Cooling is required at high pressures to 

prevent denaturation of the protein of interest. During cell lysis a large amount 

of D A and RNA material is released into the solution. This causes an increase 

in Yiscosity that can be a problem in downstream processing. Further passes in 

a homogeniser or the use of D Aase and RNAase help to break down the 

nucleic acids and reduce the viscosity [94, 153, 156]. 

Bead mills use agitation of a cell suspension with abrasives such as glass or steel 

beads to dismpt the cell walls [158]. Dismption of the target cells is due to a 

combination of collision between the beads, cavitations, and the generation of 

high shear forces . Bead mills are commonly used in industrial applications due 

30 



to ease of scale-up and the ability of the process to achieve high efficiencies with 

short disruption times [156]. Bead milling like high-pressure homogenisation is 

not as delicate as enzymatic or chemical lysis and can cause extensive 

fragmentation of the organism. This causes downstream processing problems 

with the fragments clogging filtration membranes and adsorption columns [154, 

159]. Shear forces during disruption also produce large amounts of heat that 

can denature the protein of interest therefore cooling of the bead chamber is 

required. 

Sonication is a liquid shear method of cell disruption [43, 160, 161]. The 

method utilises ultrasonic waves at 15-20 kHz which create micro-bubbles in the 

medium that on cavitation have the potential to rupture cell walls. Sonication 

has low operating costs, does not require sophisticated equipment or extensive 

staff training and the equipment can be easily cleaned [162]. However some 

proteins are inactivated by shear and heat produced in the process, and fine cell 

debris is produced which can hamper downstream processing [163]. On the 

laboratory scale sonication is extensively used however few industrial processes 

exist due to problems with localised heating [136]. 

On a laboratory scale a number of other methods are available including the use 

of detergents, solvents and enzymes to disrupt cells. Few of these have found 

large scale applicability as they are often too costly and may contaminate the 

final product, thus increasing the amount of downstream processing [157]. 

Chemical lysis looks to solubilise the walls of the cells. Common solubilising 

agents include detergents such as Triton X-100, sodium dodecyl sulphate (SDS) 

or chaotropes such as urea and guanidine hydrochloride [15, 46, 112]. The 

choice of solubilising agent is important as many proteins will be denatured in 

the presence of these compounds and they also may complicate downstream 

31 



processmg. Enzymatic digestion of the cell wall is also possible using lysozyme 

or phages [15, 16]. This method is extremely innovative in that it reduces the 

amount of downstream processing in removing cell walls however this method 

is currently too expensive for large scale application. Other non-mechanical 

metl1ods of disrupting cells include osmotic shock in which the cell are exposed 

to high salt concentrations which causes high osmotic pressures across tl1e cell 

membrane rupturing them. Sinlliarly cycling of freezing and thawing [39, 68] or 

pressurising the broth with nitrogen gas and then rapidly releasing it can be 

used. These methods are not widely used in industry but on the laboratory scale 

are often used in addition to mechanical metl1ods to improve disruption 

efficiency [136]. 

1.4.3 Initial Purification 

Purification ain1s to separate the product from materials witl1 sinlliar properties. 

This can be achieved using a range of techniques that utilise differences in 

molecular size, diffusivity, solubility, charge and density (see Figure 6). Methods 

used in initial purification include; precipitation, extraction and adsorption. 

32 



PRl "4Ai:IY ,-AC TOA 
.U-l""(CTl..0 SEf'AIU.r~ 

S I ZE 

01i:i:us1v1n 

IONIC Cl·IAi:IG( 

llAPQR IE!"f P 
PRE SSUi:IE 

S0LU91LllY 

SUi:lf'"AC( AC Tllll I l 

O(NSllY 

Anqsfroms 

Micrometer~ 10 . • 

USEFUL RANGES OF" VARIOUS SEPARATION PROCESSES 

11111 
1111 1 

II 

J! 

"'"" I I ' ] l 1111 
I , , : : 11!1 1 

Iii 

11 

I 1 I Ill 

11
1 

: I 

I~~~ il! l11 11: 

I 

io 
io · 3 

10 z 
10 . 1 

~~~~--1..-:......Jll 
~~~:111 

10 J 

10 . , 

GR..t.VllY 5(01,,.Elfl.Afl0'4 
1 , r 1 Ii " 1 • li fiT 1 r 1 11 11 11 , , , , , 

10. 

i a 
10 ~ 

10 

ll' 
~} 

., , 
10 j 

~ I ... CROM(1(R I "''"( l COAA'S[ I 
10,.tC ""'ACRO,..Ol( C'J lAA I PAlHICl( l'ARTICt£..J.__ PO'f10.( 

IUl'tGE 
1
- lt,UcG( -1 RANG( ~ AA.HQ( I AAIOQIC l 

l 

l'igure 6: Comparison of various separation processes showing the range of particle and molecular size covered and the primary factor governing th e separation process. Source 
Unit Operations Of Chemical I ·: nginccring j164j . 

33 



Precipitation is a simple well documented method of purifying proteins which is 

widely used in laboratories [162). Precipitation can be induced by the addition 

of solvents, salts, polymers and by adjusting the pH or temperature. The most 

common precipitate used in biological separation is ammonium sulphate as it is 

inexpensive and causes little or no denaturation [2, 162). Ammonium sulphate 

works by hydrating proteins, causing them to expose their hydrophobic regions 

which aggregate together causing the protein to precipitate. After precipitation, 

the pellet is separated by either centrifugation or filtration and washed in 

ammonium sulphate solution. Other potential precipitants include ethanol, 

acetone, polyetl1ylene glycol and polyethylene imine polyacrylic acid [136). 

Precipitation is generally used in the early stages o