ORIGINAL PAPER
pH and base counterion affect succinate production in dual-phaseEscherichia coli fermentations
Shiying Lu Æ Mark A. Eiteman Æ Elliot Altman
Received: 6 March 2009 / Accepted: 11 May 2009 / Published online: 30 May 2009
� Society for Industrial Microbiology 2009
Abstract Succinate production was studied in Esche-
richia coli AFP111, which contains mutations in pyruvate
formate lyase (pfl), lactate dehydrogenase (ldhA) and
the phosphotransferase system glucosephosphotransferase
enzyme II (ptsG). Two-phase fermentations using a defined
medium at several controlled levels of pH were conducted
in which an aerobic cell growth phase was followed by an
anaerobic succinate production phase using 100% (v/v)
CO2. A pH of 6.4 yielded the highest specific succinate
productivity. A metabolic flux analysis at a pH of 6.4 using13C-labeled glucose showed that 61% of the PEP parti-
tioned to oxaloacetate and 39% partitioned to pyruvate,
while 93% of the succinate was formed via the reductive
arm of the TCA cycle. The flux distribution at a pH of 6.8
was also analyzed and was not significantly different
compared to that at a pH of 6.4. Ca(OH)2 was superior to
NaOH or KOH as the base for controlling the pH. By
maintaining the pH at 6.4 using 25% (w/v) Ca(OH)2, the
process achieved an average succinate productivity of
1.42 g/l h with a yield of 0.61 g/g.
Keywords Succinic acid � CO2 � Calcium
Introduction
Succinic acid (succinate) and its derivatives are widely
used as specialty chemicals in foods, pharmaceuticals and
cosmetics [10], and it can serve as a starting material for
many commercially important products [37]. Some anaer-
obic bacteria, such as Anaerobiospirillum succiniciprodu-
cens [13, 14], Actinobacillus succinogenes [11, 31] and
Mannheimia succiniciproducens [15], produce succinate as
the major fermentation product. During anaerobic fer-
mentation, these organisms fix the greenhouse gas CO2 via
carboxylation reactions and convert C3 to C4 metabolites.
Recombinant Escherichia coli can also generate a high
concentration of succinate. For example, an anaerobic
process generates 15.6 g/l succinate with a yield of 0.85 g/g
glucose through overexpression of pyruvate carboxylase in
an alcohol dehydrogenase and lactate dehydrogenase
mutant [25]. Aerobically, E. coli generates 6.7 g/l succi-
nate at a yield of 0.71 g/g glucose, although in this case
CO2 is not fixed [16, 17]). A dual-phase E. coli process
(aerobic growth followed by an anaerobic succinate
production phase) generates nearly 100 g/l succinate at a
productivity of 1.3 g/l h and a yield of 1.1 g/g glucose
[34].
A few studies have investigated the effect of pH on
succinate production. For example, Van der Werf et al.
[32] found that succinate was produced in the pH range of
6.0–7.4 by Actinobacillus sp., but pH was not controlled.
Samuelov et al. [24] reported that for A. succiniciprodu-
cens a pH of 6.2 was better than a pH of 7.2 using NaOH.
Optimal anaerobic succinate production by E. coli was
reported to occur with a 0.2 M sodium phosphate buffer
having a pH of 6.5 at the beginning of the fermentation [1],
though pH was not controlled. Similarly, a 0.2 M sodium
phosphate buffer at an initial pH of 6.5 was found to pro-
vide the optimal initial conditions for succinate production
by Bacteroides fragilis [12]. Other studies select a pH
rather than optimize it, and universally use either Na2CO3
or NaOH for pH control, resulting in sodium ion
accumulation.
S. Lu � M. A. Eiteman (&) � E. Altman
Center for Molecular BioEngineering, University of Georgia,
Athens, GA 30602, USA
e-mail: [emailprotected]
123
J Ind Microbiol Biotechnol (2009) 36:1101–1109
DOI 10.1007/s10295-009-0594-z
The highest reported succinate concentration has been
generated by E. coli AFP111 and its derivatives using a
dual-phase process [34]. We selected this strain to study the
influence of pH, and also compared its metabolism at two
different values of pH using flux analysis with 13C-labeled
glucose. The formation of succinate requires the use of
base in order to maintain the pH. When NaOH was used in
initial experiments, we observed a decrease in the rate of
succinate formation during the course of the anaerobic
phase. To determine whether the Na? specifically is det-
rimental to succinate formation, we compared three com-
mon bases [NaOH, KOH and Ca(OH)2] and elucidated
their effects as base counterions.
Materials and methods
Bacterial strain
Escherichia coli AFP111 (F? k- rpoS396 (Am) rph-1
DpflAB::Cam ldhA::Kan ptsG) was used in this study
[3, 5].
Media and fermentation conditions
All fermentations used a defined medium containing
(per liter): 40.0 g glucose, 3.0 g citric acid, 3.0 g Na2
HPO4�7H2O, 8.00 g KH2PO4, 8.00 g (NH4)2HPO4, 0.20 g
NH4Cl, 0.75 g (NH4)2SO4, 0.84 g NaHCO3, 1.00 g
MgSO4�7H2O, 10.0 mg CaCl2�2H2O, 0.5 mg ZnSO4�7H2O, 0.25 mg CuCl2�2H2O, 2.5 mg MnSO4�H2O, 1.75 mg
CoCl2 6H2O, 0.12 mg H3BO3, 1.77 mg Al2(SO4)3�xH2O,
0.5 mg Na2MoO4�2H2O, 16.1 mg Fe(III) citrate, 20 mg
thiamine HCl and 2 mg biotin.
Dual-phase fermentations operating in batch mode with
an initial volume of 1.2 l in 2.5-l fermentors (Bioflow III,
New Brunswick Scientific, Edison, NJ) were inoculated
using 50 ml grown for 10–12 h to an optical density
(OD600) of about 2–4 in the same medium in 250-ml shake
flasks. Oxygen-enriched air as necessary was sparged at
1.0 l/min with an agitation of 200–500 rpm to maintain the
dissolved oxygen (DO) above 40% as measured by an on-
line probe (Mettler-Toledo Process Analytical Instruments,
Wilmington, MA). During growth, the pH was controlled
at 7.0 with 20% (w/v) NaOH, and the temperature was
maintained at 37�C. When the OD600 reached about 20, the
aerobic growth phase was terminated by switching the inlet
gas composition to 100% (v/v) CO2. Simultaneously, the
total flow rate was reduced to 500 ml/min (dry basis, 0�C
and 1 atm), the agitation reduced to 200 rpm, and 120 ml
of 550 g/l glucose was added. During this anaerobic phase,
the pH was controlled with either 25% (w/v) NaOH, 25%
(w/v) KOH or 25% (w/v) Ca(OH)2 as indicated.
Analysis of glucose and products
Samples were diluted by 1% H2SO4 as necessary and
centrifuged (10,0009g for 10 min at 4�C) and the super-
natant analyzed for glucose, succinate, pyruvate, acetate
and ethanol by high performance liquid chromatography
(HPLC) as previously described [6]. When calcium succi-
nate was generated, the total succinate (dissolved plus
precipitate) was determined by diluting the sample with 1%
H2SO4 and sonicating prior to analysis.
For metabolic flux analysis involving [1-13C] glucose,
dual-phase fermentations were repeated as described
above except the initial volume was 0.6 l in 1.0-L fer-
mentors (Bioflow III, New Brunswick Scientific, Edison,
NJ), the inoculum volume was reduced to 25 ml, and the
initial glucose concentration was 30 g/l. The glucose
concentration was monitored using a glucose analyzer
(YSI 2700 SELECTTM, Yellow Springs Instrument, Inc,
Yellow Springs, OH). When the glucose concentration
reached about 2 g/l, the system was switched to anaerobic
conditions by sparging CO2. When the glucose concen-
tration reached about 1 g/l, 10 ml of 450 g/l [1-13C]
glucose (99%, Cambridge Isotope Laboratories, Andover,
MA) was added into the fermentor, and the pH was
reduced within a couple minutes to the desired value with
20% (v/v) H2SO4. Samples were then collected every
30 min, centrifuged at 0�C (10,0009g for 10 min) and the
supernatants stored at -4�C for later LC-MS and NMR
analyses.
LC-MS analysis of succinate
The supernatant was filtered by a 0.2-lM syringe filter
(Cole-Parmer Instrument Co., Vernon Hills, IL) prior to
LC-MS analysis. The liquid chromatography was per-
formed at 25�C (140B Solvent Delivery System, Applied
Biosystems, Foster City, CA) using a C18 column (Kro-
masil, 250 9 1 mm, 5-lm particles, 100-A pore space,
Keystone Scientific, Inc., Bellefonte, PA). The mobile
phase at 50 ll/min was a gradient of aqueous acetic acid
(0.1% w/v, pH 3.23) and methanol. Electrospray ionization
mass spectrometry was performed using a single quadru-
pole instrument with an electrospray ion source (PE Sciex
API 1 Plus, Concord, ON, Canada). The operating condi-
tions were: nebulizer gas (N2) at a flow rate of 0.6 l/min
and pressure of 30 psi, curtain gas (N2) at a flow rate of
0.8 l/min and interface temperature 50�C. Data were
acquired in negative ion mode with a capillary voltage of
3,500 V. Mass peak heights were determined using the
BioTool Box Version B software package (Applied Bio-
systems/PE Sciex, Foster City, CA). The relative concen-
trations of M ? 0, M ? 1, M ? 2 and M ? 3 (M ? 0
represents succinate without 13C label having a m/z of 117,
1102 J Ind Microbiol Biotechnol (2009) 36:1101–1109
123
etc.) were calculated by the fractions of each peak heights
after correcting for naturally occurring 13C (1.109%) and18O (0.20%) [22].
Metabolic modeling and flux analysis
An overall flux balance was developed using stoichiome-
tric analysis and the pseudo-steady-state condition for
intracellular metabolites [27]. The balance equations for
E. coli AFP111 metabolism have been developed previously
[34], and this model was further extended by including the
pentose phosphate pathway (Fig. 1). Flux partitions (/) were
defined as fractional fluxes (m) into one of two key branches.
The fraction of flux into the pentose phosphate pathway (/PPP)
is defined as
/PPP ¼v3
v2 þ v3
The fraction of flux through anaplerotic carboxylation at
the PEP node (/PPC) is defined as:
/PPC ¼v13
v13 þ v14
The model contained 27 fluxes and 21 metabolites (Fig. 1
and Appendix A). Since insignificant growth occurs during
the anaerobic process [34], glucose consumption and
product formation rates were essentially constant (e.g.,
Fig. 2); thus, the process exhibited a metabolic pseudo-
steady state. Although five fluxes are known, the balance
equations represent an underdetermined set. In order to
calculate the optimal solution for the flux model, LC-MS
results with 1-13C-glucose were used. Specifically, we
wrote analogous balance equations for isotopomers of each
metabolite [28]. We calculated the distribution of succinate
isotopomers for a given set of fluxes and calculated the
mole fraction (Xcalc) of each isotopomer i. These calculated
values were compared with the mole fractions of each
isotopomer observed from the LC-MS results (Xobs). The
optimal solution for the flux model was then determined by
minimizing the weighted sum of squared residuals over the
four mass peaks [33]:
error ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X
3
i¼0
Xcalc;i � Xobs;i
� �2
s2i
v
u
u
t
where s is the standard deviation of the observed mole
fraction determined by three LC-MS analyses of each
sample. The isotopomer analysis cannot determine whether
flux occurred through the malic enzyme and/or pyruvate
oxidase because these pathways do not alter the13C-labeled distribution compared, respectively, with PEP
carboxylase and acetate formation via acetyl CoA. There-
fore, the model did not include these two pathways.
NMR analysis of succinate
The sample supernatant was filtered by a 0.2-lM syringe
filter, and then the filtered supernatant was mixed with 15%
volume of deuterium dioxide (D2O) for NMR analysis.
Proton-coupled 13C NMR spectra (500 NMR spectroscopy;
Varian Inc., Palo Alto, CA) at 125.7 MHz were obtained with
the following spectral parameters: 45� pulses, 31.4-kHz
spectral width and 45-s relaxation delay. Field stabilization
was achieved by locking on the D2O frequency. 13C chemical
shift assignments for succinate were determined by compa-
rison with the natural abundance standard.
NMR was not used to calculate fluxes, but was
employed to complement the LC-MS results. Specifically,
since the LC-MS results did not provide information about
the position of the label, we were able to use NMR to
observe the total 13C enrichment at C-1 position (S1) and
that at C-2 position of succinate (S2). The enrichment ratio
of S2/S1 was calculated from the ratio of the peak area of
the methylene group to the carboxyl group of succinate.
This observed enrichment ratio was compared to the
enrichment ratio calculated from the optimized metabolic
flux model. We also calculated the redox ratio as the fluxes
through NAD(P)H formation steps in the pathways divided
by fluxes through NAD(P)H consumption steps.
We chose for all metabolic flux analyses the interval
between the time 1-13C-glucose was added and the time
that the succinate concentration reached about 4 g/l
(approximately 2 h).
Results
Effect of pH on succinate production
E. coli AFP111 accumulates significant amounts of succi-
nate during an anaerobic non-growth phase after growing to
a high cell density under aerobic conditions at a pH of 7.0.
During succinate generation, CO2 is incorporated into the
central metabolism through the action of the enzyme PEP
carboxylase (PPC) [8]. We first examined the effect of pH
on succinate formation. We controlled the pH at a constant
level in the range of 5.8–7.0 using 25% Ca(OH)2 as the
neutralizing agent. Figure 2 shows the generation of several
products during the anaerobic process when the pH was
maintained at 6.4. Over the course of 14–16 h, succinate
was generated to 25–30 g/l, pyruvate to 7–10 g/l, ethanol to
about 1 g/l and acetate to less than 1 g/l. For each pH
studied, the glucose consumption rate (qG), the volumetric
succinate productivity (Qs), specific succinate productivity
(qS) and mass product yields of succinate, acetate, ethanol
and pyruvate were calculated during the 14 h of the
anaerobic phase, and the values reported are the means of
J Ind Microbiol Biotechnol (2009) 36:1101–1109 1103
123
two to three experiments (Table 1). In the pH range of 5.8–
6.4, qG and qS increased with increasing pH, but the yields
of succinate and other products did not significantly change.
When controlled pH was greater than 6.6, qG and qS
decreased. Fermentations in which the pH was controlled at
6.4 resulted in the highest specific succinate productivity.
Moreover, the volumetric succinate productivity at a pH of
6.4 or 6.6 remained high throughout the course of the
anaerobic production phase (about 1.2–1.8 g/l h), resulting
in the highest mean succinate productivity (Table 1). For a
pH above 6.6, the productivity declined over the course of
the anaerobic phase (from about 2.0 to 0.5 g/l h). Since pure
(acidic) CO2 was sparged into the fermenter during the
anaerobic phase and three acid products were formed, base
was required to maintain the pH (Fig. 3). Indeed, an unac-
ceptably large quantity of base was needed above a pH of
about 6.6. The observed reduction in specific productivity,
which would account for any ‘‘dilution effect,’’ demon-
strates that the cells generated a maximal rate of succinate at
the intermediate pH. Because of these results, a pH con-
trolled at 6.4 during the anaerobic production phase was
selected for the subsequent study.
Metabolic flux analysis
Succinate is formed through two pathways: the reductive
arm of the tricarboxylic acid cycle (TCA) via the anaple-
rotic enzyme PPC and the glyoxylate shunt [35]. To
glucose-6P
glucose
fructose-6P
glyceraldehyde-3P
ribulose-5P
sedoheptulose-7P
erythrose-4P
ribose-5P
xylulose-5P
glyceraldehyde-3P
PEP
pyruvate
isocitrate
succinate
fumarate
malate
glyoxylate
oxaloacetate
CO2
4
6
52
3
acetate
citrate
10
11
9
8
7
12
16
13
CO 2
15
17 18
3P-glycerate
fructose-1,6P2
acetyl CoA
CO2
19
succinate (ext)
acetate (ext)
ethanol (ext)
20
14
21
26
2325
27
24
22
pyruvate (ext)
1
Fig. 1 Biochemical pathways
for the synthesis of succinate
from glucose in E. coli. Not all
enzymatic steps or
intermediates are shown. Key
enzymes in the pathways are as
follows: (1) gluco*kinase,
(2) phosphoglucoisomerase;
(3) 6-phosphogluconate
dehydrogenase; (4)
phosphopentose epimerase; (5)
phosphopentose epimerase; (6)
transketolase; (7) transaldolase;
(8) transketolase; (9)
phosphofructokinase; (10)
fructose biphosphate aldolase;
(11) glyceraldehyde 3-
phosphate dehydrogenase and
phosphoglycerate kinase; (12)
phosphoglycerate mutase and
enolase; (13) PEP carboxylase;
(14) pyruvate kinase; (16)
pyruvate dehydrogenase
complex; (17)
phosphoacetyltransferase; (18)
acetate kinase; (19)
acetaldehyde dehydrogenase
and alcohol dehydrogenase;
(20) citrate synthase; (21)
aconitase; (22) isocitrate lyase;
(23) malate synthase; (24)
malate dehydrogenase; (25)
fumarase; (26) fumarate
reductase
1104 J Ind Microbiol Biotechnol (2009) 36:1101–1109
123
understand whether pH affects the distribution of these two
pathways, we compared the metabolic fluxes at a pH of 6.4
to that at a pH of 6.8. We first determined the flux distri-
bution at a pH of 6.4. Table 3 compares the observed mass
distribution of succinate and the calculated mass distribu-
tion by the optimal metabolic model (i.e., least error as
defined in ‘‘Materials and methods’’). Two methods were
used to validate the metabolic model. First, we compared
the enrichment ratio S2/S1 observed in the NMR results
with that ratio predicted from the metabolic model obtained
from independent LC-MS results (Table 2). In addition, we
calculated the redox balance based on the metabolic model,
and the value was very close to 1 (Table 2). The resulting
metabolic fluxes for the process operating at a pH of 6.4
showed that 93% of the 1.25 mol succinate formed per
mole glucose was generated via the reductive arm of the
TCA cycle and 7% via the glyoxylate shunt (Fig. 4a). At
the PEP node, /PPC was 0.61 (i.e., about 61% of PEP
partitioned to the reductive branch of the TCA cycle).
Although 39% of the PEP formed pyruvate, most of the
pyruvate ultimately became by-products (external pyru-
vate, acetate and ethanol), and only 12% of the pyruvate
became succinate via the glyoxylate shunt (either directly
or from malate).
During succinate accumulation CO2 was released via
6-phosphogluconate dehydrogenase and pyruvate dehy-
drogenase (or pyruvate oxidase) but sequestered via PPC
(Fig. 1). At a pH of 6.4, the net CO2 consumption rate was
1.10 mmol/g h. The overall stoichiometric coefficient for
CO2 (i.e., the ratio of net CO2 consumption to glucose
consumption) was 0.62 (Fig. 4a).
We similarly determined the metabolic fluxes at a pH of
6.8 (Table 2). Although the higher pH significantly reduced
the glucose consumption rate and succinate production rate
(Fig. 4b), it did not alter the carbon partitioning compared
to a pH of 6.4: succinate yield was 1.24 mol/mol glucose,
/PPP of 0.15 and /PPC of 0.62 (Table 2). Also, 38% of
the intermediate PEP formed pyruvate, and about 10% of
the pyruvate became succinate via the glyoxylate shunt; the
CO2 stoichiometric coefficient was 0.59 (Fig. 4b).
Time (h)
0 2 4 6 8 10 12 14 16 18
Glu
cose
(g/
L)
10
20
30
40
50
60
70
Pro
duct
s (g
/L)
5
10
15
20
25
30
Fig. 2 Production of succinate (filled square), pyruvate (opendiamond), ethanol (open triangle) and acetate (open square) from
glucose (filled circle) during the anaerobic non-growth production
phase for E. coli AFP111 using 25% Ca(OH)2 to control pH at 6.4
Table 1 The volumetric productivity (Q), specific consumption or production rates (q) and product mass yields (Y) of E. coli AFP111 during
14 h of an anaerobic non-growth phase using Ca(OH)2 as the base to control pH
pH qG (mg/g h) QS (g/l h) qS (mg/g h) YS (g/g) YA (g/g) YP (g/g) YE (g/g)
5.8 108.7a 0.67a 76.5a 0.71a 0.01a 0.24a 0.00a
6.0 189.3b 1.02b 126.0b 0.67a 0.01a 0.21a 0.02a
6.2 192.6b 1.18bc 125.7b 0.65a 0.02a 0.24a 0.02a
6.4 284.6c 1.42c 174.3c 0.61a -0.01a 0.24a 0.03a
6.6 257.4bc 1.49c 169.3c 0.66a 0.00a 0.16a 0.03a
6.8 166.3b 0.99b 119.4ab 0.72a 0.01a 0.19a 0.03a
7.0 85.7d 0.36d 47.4d 0.55b 0.06b 0.05b 0.02a
Different letters were statistically significantly different at the 90% confidence level
G glucose, S succinate, A acetate, P pyruvate, E ethanol
pH
5.6 5.8 6.0 6.2 6.4 6.6 6.8 7.0 7.2
Bas
e C
onsu
mpt
ion
(mL)
200
400
600
800
1000
1200
1400
Fig. 3 Volume of 25% Ca(OH)2 consumption during 14 h of an
anaerobic non-growth production phase for E. coli AFP111 under
different levels of controlled pH
J Ind Microbiol Biotechnol (2009) 36:1101–1109 1105
123
Effect of base counterion
Using a pH of 6.4 and 100% CO2 in the gas phase, we next
examined the effect of the type of neutralizing agent used
on succinate production. Three different bases were com-
pared: 25% KOH, 25% NaOH or 25% Ca(OH)2. The
effects of K? and Na? appeared to be similar, with the
succinate productivity declining during the anaerobic phase
(from about 1.7 to 0.3 g/l h). The difference in volume
used (Table 3) was due to the unequal molar base con-
centrations, and the total moles of KOH or NaOH added
during the 14-h anaerobic phase were similar. When
Ca(OH)2 was used for pH control, some calcium succinate
precipitated, and the volumetric succinate productivity
remained high during 14 h of an anaerobic production
phase (about 1.2–1.8 g/l h), resulting in the highest mean
Table 2 Comparison of the mass distributions of succinate formed by E. coli AFP111 as observed by mass spectrometry and as calculated by the
optimal metabolic model
pH uPPP uPPC M ? 0 M ? 1 M ? 2 M ? 3 (S2/S1)a R/O
6.4 0.12 0.61 Observed 0.585 ± 0.006 0.399 ± 0.009 0.014 ± 0.008 0.002 ± 0.002 19.56
Calculated 0.591 0.384 0.022 0.003 18.32 0.97
6.8 0.15 0.62 Observed 0.580 ± 0.008 0.395 ± 0.009 0.018 ± 0.005 0.009 ± 0.006 18.76
Calculated 0.576 0.398 0.022 0.003 18.79 1.06
The enrichment ratio (S2/S1) was observed from NMR results. The redox ratio R/O was calculated from the optimal metabolic model
Observed values are shown as mean ± standard deviation from three analysesa The standard deviation for NMR measurements was 4–10%
glucose-6P
glucose
fructose-6P
glyceraldehyde-3P
ribulose-5P
sedoheptulose-7P
erythrose-4P
ribose-5P
xylulose-5P
glyceraldehyde-3P
PEP
pyruvate
isocitrate
succinate
fumarate
malate
glyoxylate
oxaloacetate
CO2
acetate
citrate
CO2
3P-glycerate
fructose-1,6P2
acetyl CoA
CO2
succinate (ext)
acetate (ext)
ethanol (ext)
pyruvate (ext)
1.77
1.56
0.21
0.140.07
0.070.070.07
1.70
3.40
3.47
3.47
1.34
2.13 0.52
0.31
0.82
0.31
0.350.08
0.08
0.08
0.080
2.05
2.13
2.13
2.21
glucose-6P
glucose
fructose-6P
glyceraldehyde-3P
ribulose-5P
sedoheptulose-7P
erythrose-4P
ribose-5P
xylulose-5P
glyceraldehyde-3P
PEP
pyruvate
isocitrate
succinate
fumarate
malate
glyoxylate
oxaloacetate
CO2
acetate
citrate
CO2
3P-glycerate
fructose-1,6P2
acetyl CoA
CO2
succinate (ext)
acetate (ext)
ethanol (ext)
pyruvate (ext)
0.83
0.71
0.12
0.080.04
0.040.040.04
0.79
1.58
1.62
1.62
0.62
1.00 0.23
0.26
0.39
0.26
0.070.03
0.03
0.03
0.03
0.97
1.00
1.00
1.03
A B
Fig. 4 Metabolic fluxes (mmol/g h) during an anaerobic non-growth production phase of E. coli AFP111 at a pH of 6.4 (a) and a pH of 6.8 (b)
1106 J Ind Microbiol Biotechnol (2009) 36:1101–1109
123
succinate productivity (qS) (Table 3). The small reduction
in succinate productivity over the course of the anaerobic
phase for Ca(OH)2 can be attributed to a dilution of fer-
menter contents (i.e., adding base to a non-growing popu-
lation). Twenty-five percent Ca(OH)2 and 25% NaOH have
similar OH- molar concentrations (0.67 M for Ca(OH)2
and 0.63 M for NaOH). However, during the 14-h anaer-
obic phase about twice the amount of Ca(OH)2 was con-
sumed compared to NaOH (Table 3), which is consistent
with the higher qS for Ca(OH)2. The values of qG and YS
using Ca(OH)2 were also higher than those at NaOH or
KOH as a base counterion (Table 3).
Discussion
Between the pH range of 5.8–6.8, the specific succinate
formation and glucose consumption rates achieved their
maximum values, although the fermentation product yields
were not significantly affected by the external pH
(Table 1). Most aerobic and facultatively anaerobic bac-
teria stringently regulate the cytoplasmic pH (pHi), and the
pHi of E. coli is unaffected by large variations in the
medium pH (pHex) [2]. Olsen et al. [21] found that the pHi
of E. coli (7.0–8.0) did not vary significantly over the pHex
ranging from 5.5 to 8.0. However, the pH gradient
(DpH = pHi - pHex) is approximately 1.5 at a pHex of 5.5
and decreases with increasing pHex, ultimately reaching 0
at a pHex of 8.0 [21]. A large DpH is associated with more
active transport of H? and other ions across the membrane
to maintain pH homeostasis [2]. One might anticipate that
the increased maintenance requirement resulting from
reduced pH might reduce the product yield. Although no
product yield reduction was observed, the burden of
maintaining DpH at low pH may be responsible for the
reduced rates of substrate utilization and succinate
formation.
The succinate formation rate did not increase indefi-
nitely with increasing pH; it decreased above a pH of 6.4.
The quantity of buffer required to maintain the pH
increased with increasing pH (Fig. 3). Thus, when the pH
was above 6.4, the ionic strength increased more quickly
during the succinate formation phase, an effect that may
have been detrimental to the rate of succinate formation.
Two mechanisms may exist that result in the observed
optimum pH: a DpH effect that reduces succinate forma-
tion at low pH and an ionic strength effect that reduces the
succinate formation rate at higher pH. Interestingly, the
optimal pH for succinate productivity observed in this
study (6.4) corresponds closely with the pKa of the car-
bonic acid/bicarbonate equilibrium (6.35). Our optimal pH
for succinate production by E. coli AFP111 near 6.4 is also
consistent with other succinate-producing bacteria,
including A. succiniciproducens at a pH of 6.2 [24] and
Bacteroides fragilis at a pH of 6.5 [12].
The results demonstrate that 13C-labeling can be suc-
cessfully applied to calculate metabolic fluxes during non-
growth succinate production. For this process, the reductive
branch of the TCA cycle was the most important pathway
for succinate formation. This result is consistent with other
succinate-producing E. coli strains, including a ptsG
mutant TUQ2 [36] and E. coli strains that overexpress
Lactococcus lactis pyruvate carboxylase [26]. Surprisingly,
the flux distribution/carbon distribution was not signifi-
cantly different between a pH of 6.4 and 6.8. In particular,
the pH of the medium, which in this pH range would have a
large impact on the CO2
�
HCO�3 ratio [7, 29], did not affect
the flux partition at the PEP node (Fig. 4). Similar to our
results with E. coli, the distribution of fermentation prod-
ucts by Actinobacillus sp. 130z [32] did not differ in the pH
range of 6.0–7.4. The key enzyme PPC uses bicarbonate as
the form of CO2 [20], and flux through this pathway could
be expected to depend on the external availability of
CO2
�
HCO�3 . Indeed, in the presence of limiting concen-
trations of CO2, E. coli AFP111 formed less succinate (data
not shown); however, in this study 100% CO2 was used.
Under these circ*mstances, the yield and flux results sup-
port the conclusion that CO2 was not limiting despite the
dissociation between the dissolved gas and bicarbonate in
this pH range. Consequently, under these non-limiting
conditions, increasing the total quantity of CO2 (i.e., CO2
plus HCO�3 ) as a result of increasing the pH would not lead
to an increase in the PPC flux and associated succinate
formation. Furthermore, as the previous research described
above would suggest, pHi is likely unchanged over the pH
range studied, and pH would therefore not affect the
Table 3 The volumetric productivity (Q), specific consumption or production rates (q) and product mass yields (Y) of E. coli AFP111 during
14 h of an anaerobic non-growth phase using three different bases to control pH at 6.4
Base (25% w/v) qG (mg/g h) QS (g/l h) qS (mg/g h) YS (g/g) YP (g/g) YA (g/g) YE (g/g) Vbase (ml)
Ca(OH)2 248.6 ± 31.8 1.42 ± 0.19 174.3 ± 40.2 0.61 ± 0.13 0.24 ± 0.04 -0.01 ± 0.02 0.03 ± 0.02 135 ± 18
KOH 177.9 ± 19.0 0.88 ± 0.07 95.3 ± 5.8 0.54 ± 0.08 0.37 ± 0.01 -0.01 ± 0.02 0.01 ± 0.03 113 ± 7
NaOH 176.2 ± 28.2 0.99 ± 0.19 101.6 ± 23.9 0.57 ± 0.04 0.21 ± 0.08 0.02 ± 0.01 0.00 ± 0.02 66 ± 15
Data were presented as mean ± standard deviation from two to three replications
G glucose, S succinate, A acetate, P pyruvate, E ethanol
J Ind Microbiol Biotechnol (2009) 36:1101–1109 1107
123
activities of the various enzymes in the glucose to succinate
metabolic pathways.
The results demonstrate that Ca2? is superior to Na? or
K? as a base counterion to control the pH during the fer-
mentation. Although the specific cause for this difference is
unknown, the calcium succinate has a solubility of only
11.8 g/l at 40�C [19], which is far lower than sodium
succinate or potassium succinate. Thus, the use of Ca(OH)2
would have the double benefit of removing both the cation
Ca2? and the anion succinate from the solution. The
observed removal of succinate by the precipitation of cal-
cium succinate could relieve the inhibition of PPC [4] and
isocitrate lyase [18], two key enzymes in succinate for-
mation [35]. Removal of the cation would reduce osmotic
stress. At the end of the processes studied, the concentra-
tion of Na? or K? was about 0.5 M, while the remaining
concentration of dissociated Ca2? would have been only
about 0.18 M. High ion strength leads to osmotic stress,
and under aerobic conditions E. coli exports some Na? as a
response [23]. However, under anaerobic conditions, Na?
extrusion activity may decrease [30]. Moreover, in our
study the high ionic strength only occurred under anaerobic
non-growth conditions, potentially limiting the cellular
ability to respond to that stress. A recent study on lactate
generation similarly demonstrated that monovalent cations
reduce acid formation by E. coli compared to Ca2? [38].
Gouesbet et al. [9] found that about 0.24 M KCl (osmo-
larity of 820 mOsm) had inhibitory and repressive effects
on anaerobic enzymes and the corresponding genes in
E. coli due to osmotic stress.
In summary, although the pH does affect the production
rate, the pH within the range of 5.8–6.8 does not affect
yield or the flux distribution of products in a two-phase
succinate production process using E. coli. The negative
impact of osmotic stress during succinate accumulation can
at least partly be alleviated by using calcium as the product
counterion.
Acknowledgments Financial support from the US Department of
Energy (DE-FG26-04NT42126) is gratefully acknowledged. We
acknowledge Drs. G.P. Wylie, T.M. Andacht and D.R. Phillips for
NMR and MS analyses. We also acknowledge S.A. Lee for technical
assistance and Y. Zhu for helpful discussion.
Appendix A: Metabolic reactions for succinate
production
m1 Glucose ? ATP � glucose 6-phosphate ? ADP
m2 Glucose 6-phosphate � fructose 6-phosphate
m3 Glucose 6-phosphate ? 2 NADP? ? H2O � ribulose
5-phosphate ? CO2 ? 2 NADPH ? 2 H?
Appendix continued
m4 Ribulose 5-phosphate � ribose-5-phosphate
m5 Ribulose 5-phosphate � xylulose 5-phosphate
m6 Ribose 5-phosphate ? xylulose 5-phosphate � glyceraldehyde
3-phosphate ? sedoheptulose 7-phosphate
m7 Glyceraldehyde 3-phosphate ? sedoheptulose-7-phosphate �
fructose 6-phosphate ? erythrose 4-phosphate
m8 Xylulose 5-phosphate ? erythrose 4-phosphate � fructose
6-phosphate ? glyceraldehydes 3-phosphate
m9 Fructose 6-phosphate ? ATP � fructose 1,6-
bisphosphate ? ADP
m10 Fructose 1,6-bisphosphate � dihydroxyacetone
phosphate ? glyceraldehyde 3-phosphate
m11 Glyceraldehyde 3-phosphate ? NAD? ? ADP ? Pi �
3-phosphoglycerate ? NADH ? H? ? ATP
m12 3-Phosphoglycerate � PEP ? H2O
m13 PEP ? CO2 ? H2O � oxaloacetate ? Pi
m14 Phosphoenolpyruvate ? ADP � pyruvate ? ATP
m15 Pyruvate (intracellular) � pyruvate (extracellular)
m16 Pyruvate ? NAD? ? CoA � acetyl-
CoA ? CO2 ? NADH ? H?
m17 Acetyl-CoA ? ADP ? Pi � acetate ? CoA ? ATP
m18 Acetate (intracellular) � acetate (extracellular)
m19 Acetyl-CoA ? 2 NADH ? 2 H?� ethanol ? 2 NAD?
m20 Oxaloacetate ? acetyl-CoA ? H2O � citrate ? CoA
m21 Citrate � isocitrate
m22 Isocitrate � glyoxylate ? succinate
m23 Glyoxylate ? acetyl-CoA ? H2O � malate ? CoA
m24 Oxaloacetate ? NADH ? H?� malate ? NAD?
m25 Malate � fumarate ? H2O
m26 Fumarate ? NADH ? H?� succinate ? NAD?
m27 Succinate (intracellular) � succinate (extracellular)
References
1. Agarwal L, Isar J, Meghwanshi GK, Saxena RK (2006) A cost
effective fermentative production of succinic acid from cane
molasses and corn steep liquor by Escherichia coli. J Appl
Microbiol 100:1348–1354. doi:10.1111/j.1365-2672.2006.02894.x
2. Booth IR (1985) Regulation of cytoplasmic pH in bacteria.
Microbiol Rev 49:359–378
3. Chatterjee R, Millard CS, Champion K, Clark DP, Donnelly MI
(2001) Mutation of the ptsG gene results in increased production
of succinate in fermentation of glucose by Escherichia coli. Appl
Environ Microbiol 67:148–154. doi:10.1128/AEM.67.1.148-154.
2001
4. Corwin LM, Fanning GR (1968) Studies of parameters affecting
the allosteric nature of phosphoenolpyruvate carboxylase of
Escherichia coli. J Biol Chem 243:3517–3525
5. Donnelly MI, Millard CS, Clark DP, Chen MJ, Rathke JW (1998)
A novel fermentation pathway in an Escherichia coli mutant
producing succinic acid, acetic acid, and ethanol. Appl Biochem
Biotechnol 70–72:187–198. doi:10.1007/BF02920135
6. Eiteman MA, Chastain MJ (1997) Optimization of the ion-
exchange analysis of organic acids from fermentation. Anal Chim
Acta 338:69–75. doi:10.1016/S0003-2670(96)00426-6
1108 J Ind Microbiol Biotechnol (2009) 36:1101–1109
123
7. Frahm B, Blank HC, Cornand P, Oelssner W, Guth U, Lane P,
Munack A, Johannsen K, Portner R (2002) Determination of
dissolved CO2 concentration and CO2 production rate of mam-
malian cell suspension culture based on off-gas measurement.
J Biotechnol 99:133–148. doi:10.1016/S0168-1656(02)00180-3
8. Gokarn RR, Eiteman MA, Altman E (2000) Metabolic analysis of
Escherichia coli in the presence and absence of carboxylating
enzymes phosphoenolpyruvate carboxylase and pyruvate car-
boxylase. Appl Environ Microbiol 66:1844–1850. doi:10.1128/
AEM.66.5.1844-1850.2000
9. Gouesbet G, Abaibou H, Wu LF, Mandrand-Berthelot MA,
Blanco C (1993) Osmotic repression of anaerobic metabolic
systems in Escherichia coli. J Bacteriol 175:214–221
10. Guettler MV, Jain MK, Soni BK (1998) Process for making
succinic acid, microorganisms for use in the process and methods
of obtaining the microorganisms. US patent 5,723,322, 3 March
1998
11. Guettler MV, Rumler D, Jain MK (1999) Actinobacillus succin-ogenes sp. nov., a novel succinic-acid-producing strain from the
bovine rumen. Int J Syst Bacteriol 49:207–216
12. Isar J, Agarwal L, Saran S, Saxena RK (2006) Succinic acid
production from Bacteroides fragilis: process optimization and
scale up in a bioreactor. Anaerobe 12:231–237. doi:10.1016/
j.anaerobe.2006.07.001
13. Lee PC, Lee WG, Kwon S, Lee SY, Chang HN (2000) Batch and
continuous cultivation of Anaerobiospirillum succiniciproducensfor the production of succinic acid from whey. Appl Microbiol
Biotechnol 54:23–27. doi:10.1007/s002530000331
14. Lee PC, Lee WG, Lee SY, Chang HN (2001) Succinic acid
production with reduced by-product formation in the fermenta-
tion of Anaerobiospirillum succiniciproducens using glycerol as a
carbon source. Biotechnol Bioeng 72:41–48. doi:10.1002/1097-
0290(20010105)72:1\41::AID-BIT6[3.0.CO;2-N
15. Lee SJ, Song H, Lee SY (2006) Genome-based metabolic engi-
neering of Mannheimia succiniciproducens for succinic acid
production. Appl Environ Microbiol 72:1939–1948. doi:10.1128/
AEM.72.3.1939-1948.2006
16. Lin H, Bennett GN, San KY (2005) Genetic reconstruction of the
aerobic central metabolism in Escherichia coli for the absolute
aerobic production of succinate. Biotechnol Bioeng 89:148–156.
doi:10.1002/bit.20298
17. Lin H, Bennett GN, San KY (2005) Metabolic engineering of
aerobic succinate production systems in Escherichia coli to
improve process productivity and achieve the maximum theo-
retical succinate yield. Metab Eng 7:116–127. doi:10.1016/
j.ymben.2004.10.003
18. MacKintosh C, Nimmo HG (1988) Purification and regulatory
properties of isocitrate lyase from Escherichia coli ML308.
Biochem J 250:25–31
19. Miczynski ZN (1886) Uber die Bestimmung der Loslichkeit
einiger Sauren und Salze der Oxalsaurereihe in Wasser bei ver-
schiedenen Temperaturen. Monatsh Chem 7:255–272. doi:
10.1007/BF01516575
20. O’Leary MH (1982) Phosphoenolpyruvate carboxylase: an
enzymologist’s view. Annu Rev Plant Physiol 33:297–315. doi:
10.1146/annurev.pp.33.060182.001501
21. Olsen K, Budde B, Siegumfeldt H, Rechinger K, Jakobsen M,
Ingmer H (2002) Noninvasive measurement of bacterial intracel-
lular pH on a single-cell level with green fluorescent protein and
fluorescence ratio imaging microscopy. Appl Environ Microbiol
68:4145–4147. doi:10.1128/AEM.68.8.4145-4147.2002
22. Rosman KJR, Taylor PDP (1998) Isotopic compositions of
the elements. Pure Appl Chem 70:217–230. doi:10.1351/pac
199870010217
23. Sakuma T, Yamada N, Saito H, Kakegawa T, Kobayashi H
(1998) pH dependence of the sodium ion extrusion systems in
Escherichia coli. Biochim Biophys Acta 1363:231–237. doi:
10.1016/S0005-2728(97)00102-3
24. Samuelov N, Lamed R, Lowe S, Zeikus JG (1991) Influence of
CO2–HCO3- levels and pH on growth, succinate production, and
enzyme activities of Anaerobiospirillum succiniciproducens.
Appl Environ Microbiol 57:3013–3019
25. Sanchez AM, Bennett GN, San KY (2005) Efficient succinic acid
production from glucose through overexpression of pyruvate
carboxylase in an Escherichia coli alcohol dehydrogenase and
lactate dehydrogenase mutant. Biotechnol Prog 21:358–365. doi:
10.1021/bp049676e
26. Sanchez AM, Bennett GN, San KY (2006) Batch culture char-
acterization and metabolic flux analysis of succinate-producing
Escherichia coli strains. Metab Eng 8:209–226. doi:10.1016/
j.ymben.2005.11.004
27. Savinell JM, Palsson BO (1992) Network analysis of intermedi-
ary metabolism using linear optimization. I. Development of
mathematical formalism. J Theor Biol 154:421–454. doi:10.1016/
S0022-5193(05)80161-4
28. Schmidt K, Carlsen M, Nielsen J, Villadsen J (1997) Modeling
isotopomer distributions in biochemical networks using isotopo-
mer mapping matrices. Biotechnol Bioeng 55:831–840. doi:
10.1002/(SICI)1097-0290(19970920)55:6\831::AID-BIT2[3.0.
CO;2-H
29. Stumm W, Morgan J (1996) Dissolved carbon dioxide. In:
Aquatic chemistry, John Wiley, New York, p 192
30. Trchounian A, Kobayashi H (1999) Fermenting Escherichia coliis able to grow in media of high osmolarity, but is sensitive to the
presence of sodium ion. Curr Microbiol 39:109–114. doi:
10.1007/s002849900429
31. Urbance S, Pometto AIII, Dispirito A, Denli Y (2004) Evaluation
of succinic acid continuous and repeat-batch biofilm fermentation
by Actinobacillus succinogenes using plastic composite support
bioreactors. Appl Microbiol Biotechnol 65:664–670. doi:
10.1007/s00253-004-1634-2
32. Van der Werf MJ, Guettler MV, Jain MK, Zeikus JG (1997)
Environmental and physiological factors affecting the succinate
product ratio during carbohydrate fermentation by Actinobacillussp. 130z. Arch Microbiol 167:332–342. doi:10.1007/s0020300
50452
33. Van Dien SJ, Strovas T, Lidstroml ME (2003) Quantification of
central metabolic fluxes in the facultative methylotroph methy-lobacterium extorquens AM1 using 13C-label tracing and mass
spectrometry. Biotechnol Bioeng 84:45–55. doi:10.1002/bit.
10745
34. Vemuri GN, Eiteman MA, Altman E (2002) Succinate production
in dual-phase Escherichia coli fermentations depends on the time
of transition from aerobic to anaerobic conditions. J Ind Micro-
biol Biotechnol 28:325–332. doi:10.1038/sj.jim.7000250
35. Vemuri GN, Eiteman MA, Altman E (2002) Effects of growth
mode and pyruvate carboxylase on succinic acid production by
metabolically engineered strains of Escherichia coli. Appl
Environ Microbiol 68:1715–1727. doi:10.1128/AEM.68.4.1715-
1727.2002
36. Wang Q, Chen X, Yang Y, Zhao X (2006) Genome-scale in silico
aided metabolic analysis and flux comparisons of Escherichiacoli to improve succinate production. Appl Microbiol Biotechnol
73:887–894. doi:10.1007/s00253-006-0535-y
37. Zeikus JG, Jain MK, Elankovan P (1999) Biotechnology of
succinic acid production and markets for derived industrial
products. Appl Microbiol Biotechnol 51:545–552. doi:10.1007/
s002530051431
38. Zhu Y, Eiteman MA, DeWitt K, Altman E (2007) hom*olactate
fermentation by metabolically engineered Escherichia colistrains. Appl Environ Microbiol 73:456–464. doi:10.1128/AEM.
02022-06
J Ind Microbiol Biotechnol (2009) 36:1101–1109 1109
123
