Mitg
lied
der H
elm
holtz
-Gem
eins
chaf
t
Launching an automated microtiter cultivation platform for enhanced bioprocess optimization
18th April 2013 | Forschungszentrum Jülich GmbH IBG-1: Biotechnology Institute of Bio- and Geosciences (IBG) Bioprocesses & Bioanalytics Group Leo-Brandt-Str. / 52428 Jülich / Germany www.fz-juelich.de/ibg/ibg-1
P. Rohe, O. Schweissgut, R. Freudl, W. Wiechert, M. Oldiges
Folie 2
Outline
Introduction: What is so complicated about heterologous protein expression and why do we need higher troughput ? Results:
• JuBOS: Juelich Bioprocess Optimization System – A smart platform for small scale cultivation • Bioprocess optimization of C. glutamicum as expression system
− Strain selection − Induction profiling − Medium composition − Fed-Batch feedrate
Summary:
Slide 3
Replication
Transcription
Translation
Folding Secretion SEC-Path TAT-Path
Proteolysis Agglomeration
SP = signal peptide
Active protein
SP SP
Induction +
Affected by bacterial metabolism
Introduction: Optimizing protein production
Requires testing of different biological and bioprocess engineering variables
Slide 4
Hosts
Enzymes
Chaperones
Signal peptides
Temperature
Feedrate
Induction strength
Induction biomass
Media component
E. coli B. subtilis P. putida C. glutamicum R. capsulatus C. utilis
1 2 3 4 5 6 7 8 9 10 11 12 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Process phase Growth phase Production phase
680.400 com
binations 1.2*10
17
combinations
8*1022 com
binations
Enzyme Toolboxes (qualitative factors)
Bioengineering parameters (quantitative factors)
Introduction: Optimizing protein production
Slide 5
Outline
Introduction: What is so complicated about heterologous protein expression and why do we need higher troughput ? Results:
• JuBOS: Juelich Bioprocess Optimization System – A smart platform for small scale cultivation • Bioprocess optimization of C. glutamicum as expression system
− Strain selection − Induction profiling − Medium composition − Fed-Batch feedrate
Summary:
Slide 6
media preparation (Liquid handling)
cultivation (biolector)
storage (cooling device)
cell separation (mtp centrifuge)
photometric assay
(mtp reader)
Juelich Bioprocess Optimization System (JuBOS)
Integration of Cultivation and Robotic setup
• Installed in laminar flow cabinet • Automated media preparation • Pipetting (sampling/dosing/harvest)
• Programmable pipetting events triggered by (CDW, pO2, Time, pH) • Glucose fed-batch medium
Slide 7
Outline
Introduction: What is so complicated about heterologous protein expression and why do we need higher troughput ? Results:
• JuBOS: Juelich Bioprocess Optimization System – A smart platform for small scale cultivation • Bioprocess optimization of C. glutamicum as expression system
− Strain selection − Induction profiling − Medium composition − Fed-Batch feedrate
Summary:
Folie 8
Why is Corynebacterium glutamicum interesting ?
gram positive soil bacterium tempopt ≈ 30 °C / pHopt ≈ 7.2 discovered in Japan (1957) tolerates high glucose conc.
… is used for industrial amino acid production since many decades:
5 µm
L-Glutamate, L-Lysine, L-Iso-Leucine L-Leucine L-Valine …
… was also discovered for efficient protein secretion … very low extracellular protease background activity
Slide 9
Catalytic triad • Cutinase from Fusarium solani pisi • Lipase model-enzyme • Degrading cutin (insoluble matrix covering plant surface) • 22 kDa, no cofactors required • Application: Textile, Dairy, Detergents • Costs: 5000 €/g purified protein (www.vtt.fi)
• Library of 180 SEC- signal peptides (SP)
• Each SP is fused to cutinase
• Screening for best SP in MTP approach
SP Ranking: Brockmeyer 2006, J Mol Biol. 362(3):393-402
Target for Bioprocess development: Optimize secretory production of cutinase using C. glutamicum
Secretory cutinase production in B. subtilis
Transfer library of signal peptides for SEC transport to C. glutamicum Optimize secretory production of cutinase in C. glutamicum
Slide 10
Strain screening
media optimization
induction optimization
feedrate screening
feed optimi- zation
scale-up
clone- library
Premliminary tests
Increasing throughput via MTP-based cultivations similar to bioreactor conditions Optimization by modular „Step-by-step“ approach
Typical sequential steps during bioprocess optimization Application to C. glutamicum as expression system
Juelich Bioprocess Optimization System (JuBOS)
Slide 11
Time [h]0 2 4 6 8 10 12 14 16
CD
W [g
. L-1]
lip. a
ct. [
U. m
L-1]
02468
10121416
CDWlip. act.
Time [h]0 2 4 6 8 10 12 14 16
CD
W [g
. L-1]
lip. a
ct. [
U. m
L-1]
02468
10121416
CDWlip. act.
NprE YwmC YpjP Empty
spec
ific
activ
ity [U
. mg-1
]
0.00.20.40.60.81.01.21.4
NprE YwmC YpjP Empty
spec
ific
activ
ity [U
. mg-1
]
0.00.20.40.60.81.01.21.4
C.glutamicum ATCC 13032 pEKEX2::SP-Cutinase T=30°C, 1200 rpm, 3 mm, media: CG XII, 0.5 mM IPTG
CDW [mg.mL-1]0 2 4 6 8 10 12 14
lipol
ytic
act
ivity
[U. m
L-1]
02468
101214
1mL MTP-scale BioLector (NprE-Cutinase):
1.05 +/- 0.06 U.mg-1
µ = 0.4 h-1
µ = 0.4 h-1
CDW [mg.mL-1]0 2 4 6 8 10 12 14
lipol
ytic
act
ivity
[U. m
L-1]
02468
101214
1.07 +/- 0.03 U.mg-1 spec
.lip.
act.
[U. m
g-1 ]
spec
.lip.
act.
[U. m
g-1 ]
How comparable are results from 1 L lab-scale bioreactor and 1 mL MTP-scale BioLector cultivation ?
Perfect match of process characteristics with scale-up factor 1000 Similar µ, YX/S, YP/X and SP performance observed at MTP-scale
1L lab-scale Bioreactor (NprE-Cutinase): Rohe et al. Microbial Cell Factories 2012, 11:144.
(highly accessed)
Slide 12
time [h]
0 2 4 6 8 10 12 14 16 18
CDW
[g. L
-1]
02468
101214 sampling
time [h]
0 2 4 6 8 10 12 14 16 18
CDW
[g. L-1
]
02468
101214
time [h]
0 2 4 6 8 10 12 14 16 18
CDW
[g. L
-1]
02468
101214
4 h
start timer
sampling
4 h4 h
lip. a
ctiv
ity [U
. mL-
1 ]
02468
1012141618
epr amyE YwmC
lip. a
ctiv
ity [U
. mL-
1 ]
02468
1012141618
epr amyE YwmC
I. Synchronous sampling at 15 h (Over night culture):
C.glutamicum ATCC 13032 pXMJ 19: SP-Cutinase T=30°C, 1200 rpm, 3 mm, 1 mL media: CG XII, 0.5 mM IPTG
± 30 %
± 10 %
How to analyse fast and slow strains on a fair basis ?
II: Biomass triggered standardized sampling events
Strain specific standardization of individual growth/production time neccessary Biomass triggered sampling events improve reproducibility down to 10%
N > 11 technical replicates 3 different SP´s (epr, amyE, YwmC)
Slide 13
B. subtilis
C. glutam
icum
lip.act [U.mL-1]C.glutamicum
0 2 4 6 8 10 12
lip.a
ct. [
U. m
L-1 ]
B
.sub
tilis
0
2
4
6
lip.act [U.mL-1]C.glutamicum
0 2 4 6 8 10 12
lip.a
ct. [
U. m
L-1 ]
B
.sub
tilis
0
2
4
6
lip.act [U.mL-1]C.glutamicum
0 2 4 6 8 10 12
lip.a
ct. [
U. m
L-1 ]
B
.sub
tilis
0
2
4
6
lip.act [U.mL-1]C.glutamicum
0 2 4 6 8 10 12
lip.a
ct. [
U. m
L-1 ]
B
.sub
tilis
0
2
4
6
lip.act [U.mL-1]C.glutamicum
0 2 4 6 8 10 12
lip.a
ct. [
U. m
L-1 ]
B
.sub
tilis
0
1
2
3
4
5
6
7
- + ~equal
+ -
1
Y-Data: Brockmeyer 2006, J Mol Biol. 362(3):393-402 X-Data (this work): C.glutamicum ATCC 13032 pXMJ 19: SP-Cutinase 1 mL; T=30°C; 1200 rpm, 3 mm, CG XII, 0.5 mM IPTG
Good and bad signal peptides (SP): Comparing protein secretion performance in B. subtilis vs. C. glutamicum
Non-predictable secretion performance when transferring SP´s to C. glutamicum De-novo search for best SP seems to be obligatory !
Slide 14
CG XIIlip activity [U.mL-1]
0 2 4 6 8 10
BHI
lip a
ctiv
ity [U
. mL-1
]
0
1
2
3
4
5
Complex vs. Minimal medium
30°Clip activity [U.mL-1]
0 2 4 6 8 10
23°C
lip a
ctiv
ity [U
. mL-1
]
02468
101214
30°C vs. 23°C
Are good signal peptides always good ones : Uniqueness of the best SP (or unique only for chosen condition)
Complex vs minimal medium: only a few SP performance changes, but tendency for higher performance in minimal medium
Temperature change results in complete mix up of SP performance and no SP shows good performance under both conditions
SP´s are best suited only at chosen cultivation conditions !!
(com
plex
)
C.glutamicum ATCC 13032 , pXMJ 19: SP-Cutinase 1 mL, 1200 rpm, 3 mm, media: CG XII, 0.5 mM IPTG
Slide 15
Strain screening
media optimization
induction optimization
feedrate screening
feed optimi- zation
scale-up
clone- library
Premliminary tests
Results: media optimization
Juelich Bioprocess Optimization System (JuBOS)
Folie 16
IPTG induction: Optimal time point vs. concentration
Aktivität [U/mL]
Optimum clearly identified at 1 h and 200 µM IPTG
Optimization of time point of IPTG addition and induction strength for NprE-Cutinase (pEKEX2) in C. glutamicum with CGXII medium
IPTG was automatically added by the robotic setup and samples were harvested after cultivation
Mesh of 8 x 6 = 48 data points (= 1 MTP)
Slide 17
Strain screening
media optimization
induction optimization
feedrate screening
feed optimi- zation
scale-up
clone- library
Premliminary tests
What is the best minimal medium composition for C. glutamicum as expression system ?
Juelich Bioprocess Optimization System (JuBOS)
Slide 18
NH3 vs Harnstoff
Time [h]0 10 20 30
CD
W [g
. L-1]
0
5
10
15
ManualRobot
Glucose
Biotin
PCA Kanamycin
IPTG
FeSO4
MnSO4
ZnSO4
CuSO4
NiCl2
K2HPO4 / KH2PO4
(NH4)2SO4
CaCl2
MgSO4
MOPS
Urea
Betain
Volumes in .csv-Worklist
6 technical replicates
6 technical replicates
C.glutamicum DSM 13032 T = 30°C, 1 mL 1200 rpm, 3 mm, media: CG XII
Time for 1 mtp: 2.5 h
Time for 1 mtp: 15 min
What is the best minimal medium composition for C. glutamicum as expression system ?
Error: 10%
Error: 7%
Initial testing of all single components showed four interesting candidates Automated media optimization with Design of Experiments (DoE)
Med
ium
com
pone
nts
of C
GXI
I med
ium
Slide 19
ZnSO4
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
Ammonium
CuS
O4
ZnSO4
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
Ammonium
CuS
O4
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
Ammonium
CuS
O4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
Ammonium
CuS
O4
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
Ammonium
CuS
O4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
0 20 40 60 80 100
Bet
ain
0
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 10 20 30 40
Ammonium sulfate [g.L-1]
0 0.
15
0.30
0.
45
0.60
copp
er s
ulfa
te [m
g.L-
1 ] Optimization of four parameter with evolutive response surface algorithm:
Evolutive Algorithm
Algorithm: Schweissgut & Wiechert, Proceedings of 7th EUROSIM Congress, Prague 2010
Medium optimization assisted by computational methods
BioLector experiment
Only 240 experiments from 625 were neccessary by using the evolutive response surface algorithm for DoE
Slide 20
ZnSO4
0 20 40 60 80 100
Beta
in
0
20
40
60
80
100
0 20 40 60 80 100
Beta
in
0
20
40
60
80
100
0 20 40 60 80 100
Beta
in
0
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 100
Beta
in
0
20
40
60
80
100
0 20 40 60 80 100
Beta
in
0
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
ZnSO4
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 20 40 60 80 1000
20
40
60
80
100
0 10 20 30 40
Ammonium sulfate [g.L-1]
0 0.
15
0.30
0.
45
0.60
copp
er s
ulfa
te [m
g.L-
1 ]
0.8 1.0 1.2 1.4 1.6
Classic media: CG XII
New media: CG ExXII
GFP
FL
(Glc
-lim
it, b
iom
ass
spec
.)
response surface function:
Medium optimization assisted by computational methods: Old CGXII and the new CGExXII
New CGExXII medium: factor 1.6 increased GFP fluorescence factor 1.5 increased cutinase activity
Slide 21
Strain screening
media optimization
induction optimization
feedrate screening
feed optimi- zation
scale-up
clone- library
Classical and MTP-based glucose feedrate optimization
Premliminary tests
Juelich Bioprocess Optimization System (JuBOS)
Slide 22
Start Feed
0
10
20
30
40
50
60
70
0 20 40 60
EA [k
U/L
]
CDW [g.L-1]
0
1
2
3
4
0 5 10 15
EAsp
ec [k
U/g
] Feed [gGlc
.L-1.h-1]
Start Feed
benchmark (batch)
DO
[%]
NprE-Cutinase
Classical feedrate optimization in 1 L bioreactor
Glucose feed rate is a critical design parameter for fed-batch Performance increase (3x) with optimal fed-batch setup
Different signal peptides showed different optimal fed-batch feedrates !
Slide 23
Strain screening
media optimization
induction optimization
feedrate screening
feed optimi- zation
Summary: Bioprocess optimization of cutinase expression in C. glutamicum using JuBOS
Premliminary tests
scale-up
14 days 4 days 16 days 4 days
240 cultivations
96 cultivations
384 cultivations
96 cultivations
Bioprocess optimization with 816 cultivations Possible with 1 person in ~6 weeks in MTP based cultivations
Juelich Bioprocess Optimization System (JuBOS)
Slide 24
Summary: Bioprocess optimization of cutinase expression in C. glutamicum using JuBOS
Strain screening
media optimization
induction optimization
feedrate screening
Premliminary tests
• Severe cross-relationship between different optimization modules oberved
• Initial strain screening conditions must be similar to production conditions
• Arrangement of the modules have direct effect on the final optimal set of biological and bioprocess parameters (local vs. global optimum) • Finding optimal standard conditions is a very unlikely scenario for protein secretion in C. glutamicum (also in other microbial systems ?) Strong demand for target protein specific de-novo optimization approach ! • JuBOS setup provide powerful technical environment for enhanced bioprocesss optimization under conditions very similar to lab-sale bioreactor conditions Outlook: • Implementation of parallelized fed-batch option and pH control in micro plate Strong demand for design of experiments
Folie 25
Acknowledgements
Forschungszentrum Jülich (IBG-1) : Prof. Bott Dr. Britta Kleine
Dr. Frank Kensy Dipl.-Ing. Carsten Müller
Project partners:
Heinrich-Heine University Düsseldorf: Prof. K.E. Jäger Prof. Pietruzska Dr. Ulrich Krauss Dr. Britta Kleine Kathrin Scholz
Project funding : Collaboration:
Mark your calendar: 15-19 July 2013, Berlin
#Quantitative Biology: Current concepts and tools for strain and process developments
www.QBio-SummerSchool.de
# Organized by former and current members of the Young Researchers Network (Zukunftsforum Biotechnologie) of the German Society for
Chemical Engineering and Biotechnology.