signature=e17a25972af7fb79bfc9a87833cb8e97,A quantitative study of methanol/sorbitol co-feeding proc...
Strain and medium composition
A P. pastoris strain GS115 transformed with pSAOH5 vector bearing a pAOX1-LacZ construct was used [27, 28]. Media used were named as follows: BG (batch culture with glycerol); CS (1st continuous culture with sorbitol), TCC (2nd transient continuous culture with methanol/sorbitol co-feeding), FBM (fed-batch with methanol), and FBMS (fed-batch with methanol/sorbitol mixture). Medium compositions was as follows (per liter) : 25 mL H3PO4 85%, 1.05 g CaSO4.2H2O, 18.28 g K2SO4, 14.96 g MgSO4.7H2O, 4.15 g KOH, and 12 mL PTM1 trace elements solution (P. pastoris fermentation manual of the Invitrogen) and 3.2 C-mol/L of the carbon source. These were 0.54 mol/L glycerol, 3.2 C-mol/L sorbitol, 3.2 C-mol/L methanol/sorbitol mixture (linear change of the methanol fraction), C-mol/L methanol, and 3.2 C-mol/L methanol/sorbitol mixture (0.60/0.40 C-mol/C-mol), in BG, CS, TCC, FBM, and FBMS medium, respectively.
Bioreactor operation
All cultures were performed in 2 l bioreactor (Biostat® B plus, Sartorius AG). The temperature was maintained at 30°C during glycerol growth phase and shifted to 25°C during methanol and mixed feed induction phase. The pH was regulated at 5.8±0.2 by the addition of 25% ammonia solution. Dissolved oxygen (DO) was maintained at 30% of saturation by a PID controller . The culture history was recorded by the supervisory control and data acquisition system (MFCS/win 3.0).
Culture started in batch mode with 1.5 l of the BG medium at an initial OD600 of 0.2. After glycerol depletion (observed typically after 32 ~ 36 h with an sudden increase of pO2), a continuous culture phase started by feeding the bioreactor with the CS medium at a dilution rate of 0.023 h-1 (corresponding to a feeding rate of 46 ml/h). After a steady state was reached (i.e. after 5 residence times), the transient continuous cultures were carried out by increasing (1st transient continuous culture) or decreasing (2nd transient continuous culture) the methanol fraction in the TCC medium. The dilution rate was also maintained at 0.023 h-1 during the transient continuous cultures.
Fed-batch cultures started with a batch phase in 1.5 L BG medium at an initial OD600 of 0.2. After glycerol depletion, the cultures were fed with the FBM medium or FBMS medium. The feeding rate was fixed at 40 ml/h for two fed-batch cultures and the dilution rate was calculated based on the increase of the culture volume over time.
Quantitative analysis of biomass, metabolites, and β-galactosidase
Cell growth was monitored either by optical density at 600 nm (OD600) or dry cell weight (DCW). An OD600 value of 1 was found to correspond to 0.236 g DCW/L. Glycerol, methanol, and sorbitol were analyzed by isocratic RID-HPLC (Hewlett Packard model 1100, Waldbronn, Germany) using an Aminex HPX-87H ion-exclusion column (300×7.8 mm Bio-Rad, Hercules, USA) with 5 mmol/L H2SO4 as mobile phase at a flow rate of 0.5 ml/min at 30°C. Ammonia assay is based on the phenol-hypochlorite method [29]. β-galactosidase activities were determined as described previously [30]. All analyses were performed in duplicate. Statistical significance was accepted at p > 0.05.
Exhaust gas analysis and heat production calculation
The concentrations of oxygen and carbon dioxide in the exhaust gas were monitored in real time with a gas analyzer (EGAS-1, Advance optima, ABB). OUR (CER) were assumed to equal OTR (CTR) and were calculated by below equations:
OUR
≅
OTR
=
1000
×
O
2
in
100
−
N
2
in
×
O
2
out
100
×
N
2
out
×
Q
g
V
×
60
V
m
(1)
CER
≅
CTR
=
1000
×
N
2
in
×
C
O
2
out
100
×
N
2
out
−
C
O
2
in
100
×
Q
g
V
×
60
V
m
(2)
Heat production was calculated by the following equation [8]:
–
Δ
H
reaction
≅
–
Δ
H
o
reaction
≅
∑
j
Y
s
,
j
κ
j
substrates
-
∑
i
Y
p
,
i
κ
i
products
×
115
KJ
/
C
-
mol
(3)
Extracellular flux fextra determination
Totally extracellular fluxes of 5 species from the measurements during transient cultures could be determined as follows.
The specific biomass growth rate (μ) was determined by:
V
dX
dt
=
μXV
–
FX
≅
0
⇒
μ
≅
F
V
=
D
(4)
The specific production rate of biomass fextra,X was obtained in the unit of mmol/(g DCW· h):
f
extra
,
X
=
1000
×
μ
M
X
(5)
A standard chemical composition CH1.761O0.636N0.143[23] was used, and the ash content of biomass was assumed to be 5%. As a result, MX of 31 g DCW/mol was obtained.
The specific consumption of oxygen was obtained after OUR measurement:
f
extra
,
O
2
=
–
OUR
X
(6)
The specific production of carbon dioxide was calculated by:
f
extra
,
C
O
2
=
CER
X
(7)
The specific rates of substrate consumption were determined from the following equation:
V
d
C
i
dt
=
f
extra
,
C
i
XV
+
F
C
i
,
in
–
C
i
=
0
⇒
f
extra
,
C
i
=
–
D
C
i
,
in
–
C
i
X
i
=
methanol
,
sorbtiol
(8)
Data reconciliation and consistency check of macroscopic balances
The macroscopic balances of carbon element (C) and redox degree provided two degrees of redundancy Additional file 2. The reconciled measurement vector f *extra was the minimum variance estimate of measurement noise (error) which was assumed to distribute independently and normally with a mean value of zero:
f
extra
*
=
Arg
min
f
*
extra
f
extra
*
–
f
extra
T
W
f
extra
*
–
f
extra
(9)
s.t.
R
redundancy
f
extra
*
=
0
(10)
where Rredundancy ∈ R2 × 5 was the redundancy matrix of which each element of the first or second row was the C number or redox degree of one species molecule; fextra ∈ R5 × 1 was the measured extracellular flux vector. W was the diagonal weighting matrix of which each element wi,i equals 1/σi2 (i.e., the reciprocal of squared standard deviation of fextra,i ). In this work, σi was assumed to be a constant percent of fextra,i. The error percentage was 10% for biomass, biomass, methanol, and sorbitol measurements, and 15% for OUR and CER measurements.
The solution of Eq. (9) was obtained by the following equations [31]:
f
extra
*
=
I
–
W
–
1
R
redundancy
T
P
ε
–
1
R
redundancy
f
extra
(11)
P
ε
=
R
redundancy
W
−
1
R
redundancy
T
(12)
Additionally, a statistical consistency index hε, following a χ–square distribution was used to check the consistency of macroscopic balance constraints [31].
h
ε
=
ε
T
P
ε
–
1
ε
(13)
where ε = Rredundancy fextra was the vector of balance residual. All calculations were performed in the platform of Matlab 2011b (http://www.mathworks.com).
Intracellular metabolic flux fintra analysis
Additional file 1: Table S4 in Supplementary Material 1 shows the details of proposed metabolic model. Firstly, NADH and NADPH were considered to be equivalent or transform into each other liberally, so NADH was only used to represent both of them. Secondly, summation of a group of sequential elementary metabolic reactions without branched pathway led to an equivalency of overall reaction (such as fGAP-CO2). Finally, production of biomass was simply represented by an overall reaction, into which relevant complicated anabolic reactions from the precursor GAP were lumped macroscopically based on carbon and reduction balance.
Intracellular flux f
^
intra was estimated by solving a constrained linear system below
R
extra
·
f
^
intra
=
f
extra
*
R
intra
·
f
^
intra
=
0
s
.
t
.
f
^
intra
≥
0
(14)
The first two equations in Eq. (14) were mass balances for extracellular, and intracellular species (due to pseudo steady-state assumption and neglecting dilution effect of cell growth), respectively. And the third corresponds to the constraint of thermodynamic irreversibility. Rextra ∈ R5 × 10 and Rintra ∈ R6 × 10 were corresponding stoichimetric matrices (Additional file 2). As a whole, the proposed metabolic model has ten reactions needing estimation while five extracellular rates were measurable and six intermediates were assumed to be steady-state. As a result, the proposed model is over-determined.
Additionally, sensitivity and singularity test [31] were performed to ensure that the stoichiometric system was well-posed (i.e., the stoichiometric matrix was well-conditioned).