Bioresource Technology 156 (2014) 232–239

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Construction of reductive pathway in Saccharomyces cerevisiae

for effective succinic acid fermentation at low pH value
Daojiang Yan a,b, Caixia Wang a,b, Jiemin Zhou a,b, Yilan Liu a,b, Maohua Yang a, Jianmin Xing a,⇑
a
National Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, P.O. Box 353, Beijing 100190, PR China
b
University of Chinese Academy of Sciences, Beijing 100049, PR China

h i g h l i g h t s

 Metabolic engineering was developed to improve succinate production in yeast.

 Reductive pathway was efficient for improving succinate yield.
 Succinate production was improved by the deletion of GPD1.
 Succinate yield was regulated by urea and biotin levels.
 Succinic acid could be effectively produced at pH 3.8 in batch bioreactor.

a r t i c l e i n f o a b s t r a c t

Article history: Succinic acid is an important precursor for the synthesis of high-value-added products. Saccharomyces
Received 31 October 2013 cerevisiae is a suitable platform for succinic acid production because of its high tolerance towards acidity.
Received in revised form 13 January 2014 In this study, a modified pathway for succinate production was established and investigated in S. cerevi-
Accepted 15 January 2014
siae. The engineered strain could produce up to 6.17 ± 0.34 g/L of succinate through the constructed path-
Available online 24 January 2014
way. The succinate titer was further improved to 8.09 ± 0.28 g/L by the deletion of GPD1 and even higher
to 9.98 ± 0.23 g/L with a yield of 0.32 mol/mol glucose through regulation of biotin and urea levels. Under
Keywords:
optimal supplemental CO2 conditions in a bioreactor, the engineered strain produced 12.97 ± 0.42 g/L
Succinic acid
Reductive pathway
succinate with a yield of 0.21 mol/mol glucose at pH 3.8. These results demonstrated that the proposed
Reducing power engineering strategy was efficient for succinic acid production at low pH value.
Low pH Ó 2014 Elsevier Ltd. All rights reserved.
Saccharomyces cerevisiae

1. Introduction Well-established manipulation tools are also available for this

Succinic acid, a 1,4-dicarboxylic acid, is widely used in the food,
pharmaceutical and chemical industries (Zeikus et al., 1999). It
serves as an important precursor for the synthesis of high-value-
added products with a potential global market of over 2 billion
USD annually (McKinlay et al., 2007). Nowadays, succinic acid is
primarily derived through petroleum-based processes, which leads
to high prices and environmental problems. On the other hand,
microbial fermentation is considered a green process that can pro-
duce succinate from renewable resources. It has seen great devel-
opments in recent years due to increases in petroleum prices and
environmental concerns (Liu et al., 2012, 2013; Liang et al., 2013).
Saccharomyces cerevisiae is a robust and important industrial
microorganism with a thoroughly researched genetic background.
⇑ Corresponding author. Address: P.O. Box 353, Zhongguancun Bei-er-tiao 1,
Haidian District, Institute of Process Engineering, Chinese Academy of Sciences,
Beijing 100190, PR China. Tel./fax: +86 10 62550913.
E-mail address: jmxing@home.ipe.ac.cn (J. Xing).
organism. Recently, the use of S. cerevisiae in production of bio-
chemicals besides the traditional ethanol has been intensively
investigated (Zelle et al., 2008; Kim et al., 2013). Current produc-
tion practices for succinic acid primarily utilize Actinobacillus suc-
cinogenes and genetically engineered Escherichia coli (Liu et al.,
2013; Song and Lee, 2006). However, in these cases, large quanti-
ties of neutralizing compounds need to be added during the fer-
mentation process to maintain a neutral environment. Compared
with prokaryotes, S. cerevisiae is highly tolerant of low pH values,
making it superior for succinic acid production. Fermentation at
low pH values markedly reduces the demand for alkaline neutral-
izers, prevents bacterial pollution and facilitates the downstream
process (Li et al., 2010).
Since succinate is not normally produced at high levels in
S. cerevisiae, metabolic engineering strategy provides an important
approach to improve succinate production. All studies to date
focused on methods utilizing the TCA cycle in the oxidative direc-
tion or glyoxylate shunt to produce succinate (Raab et al., 2010;

0960-8524/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.biortech.2014.01.053
 D. Yan et al. / Bioresource Technology 156 (2014) 232–239
Otero et al., 2013 and Agren et al., 2013). At present, the highest le-
vel achieved for succinate production is 0.11 mol/mol glucose with
a productivity of 0.022 g/L/h in S. cerevisiae (Raab et al., 2010).
However, if a reductive TCA pathway is used, the maximum theo-
retical yield of succinate is 1.714 mol of per mol of glucose, supe-
rior to values achieved by either the oxidative TCA cycle or
glyoxylate shunt (Li et al., 2013). Moreover, this process results
in net CO2 fixation instead of release which has benefits for reduc-
ing global warming. The reductive TCA pathway for succinate pro-
duction contains four steps: pyruvate carboxylation, oxaloacetate
reduction, reversible hydrated translation of malate and fumarate,
and fumarate reduction (Zeikus et al., 1999). However, there are
some obstacles in succinate production via this pathway in S. cere-
visiae. First, yeast fumarase (FUM) FUM1p exhibits characteristi-
cally irreversible catalysis of conversion of fumarate to malate,
which represents a major obstacle for production via the reductive
pathway (Pines et al., 1996). Second, the fumarate reductase (FRD)
genes FRDS1 and OSM1 are only expressed under anaerobic condi-
tions (Muratsubaki and Enomoto, 1998). In addition, production of
1 mol of succinic acid consumes 2 mol of NADH via the reductive
pathway, while production of malic acid and fumaric acid needs
less. Due to the limited NADH supply in the cytosol, the production
of succinic acid is more difficult than that of malic acid and fumaric
acid through the reductive TCA pathway in S. cerevisiae.
This study aims to improve succinate production by using met-
abolic engineering and to achieve effective fermentation of succi-
nate at low pH values. The TAM strain, a pyruvate decarboxylase
(pdc)-deficient S. cerevisiae strain that produces substantial
amounts of pyruvate, was used as the parent strain (Maris et al.,
2004). The effects of reductive pathway through overexpression
of pyruvate carboxylase PYC2p, cytosolic retargeted MDH3p,
FRDS1p and E. coli FumCp on succinate production were investi-
gated. And the resultant effectiveness of the altered pathway on
succinate production was evaluated. Furthermore, GPD1 was de-
leted to induce accumulations of carbon reserves and reduction po-
tential. The metabolic engineering strategy was shown in Fig. 1.
Besides, the levels of neutralizing agent, nitrogen source and biotin
were optimized to further improve succinic acid production. Final-
ly, fermentation by the engineered strain at low pH in a bioreactor
setting was explored.
2. Methods
2.1. Strains, media, and growth conditions
Yeast strains used in this study are derived from the evolved
pdc-deficient S. cerevisiae strain TAM (Maris et al., 2004). All strains
and plasmids used in this study were listed in Table 1. During yeast
strain construction, cultures were grown aerobically at 30 °C,
200 rpm in SD media consisting of 20 g/L glucose, demineralized
water, 6.7 g/L YNB (yeast nitrogen base without amino acids,
Difco). After the solution was sterilized for 20 min at 115 °C,
0.1 g/L amino acids (uracil or/and histidine) was added. E. coli
DH5a was used as a vector host and cultivated in Luria–Bertani
medium (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl) at 37 °C.
2.2. Construction of plasmids for overexpressing genes
For limited restriction sites, fusion PCR was used to construct
the expression cassettes of target genes (Fig. S1). All primers used
in this study were listed in Table S1. PYC2 expression cassette was
digested with Xba I, Sal I and ligated to YIPmcherry, resulting in
YIP-PYC2. The same strategy was adopted to introduce the MDH3R
(retargeted MDH3 which lacking of the last 9 base pairs) expression
cassette into YIP-PYC2. The resulting plasmid was designated as
233
YIP-PYC2MDH3R. The expression cassettes of EcFumC, and FRDS1
were generated and introduced into pRS313 by the same strategies
described above. EcFumC were cloned from E. coli DH5a genome.
FRDS1 was cloned from TAM genome. Promoter, terminator,
PYC2, MDH3R, and FRDS1 were cloned from the genome of TAM.
Restriction endonucleases (Takara, Japan) and DNA polymerase
(KOD-Plus-Neo, Toyobo, Japan) were used according to the instruc-
tions. The plasmids were electroporated into yeast cells using a
Gene Pulser Xcell Electroporation System (Bio-Rad, USA) and se-
lected on solid SD auxotrophic plates with 1.5% agar powder.
2.3. PCR-mediated seamless gene deletion method for gene knockout
To allow for the introduction of additional plasmid, HIS3 was
deleted as a second auxotrophic selection marker. HIS3 was deleted
according to the PCR-mediated seamless gene deletion and marker
recycling method (Akada et al., 2006). The single clone was
checked on SD plate with or without histidine. FUM1 and GPD1
were deleted using the same strategy. The resulting strain was
checked by PCR for losing of the gene locus.
2.4. Cell extracts and enzyme activity assays
Yeast cell cultures were collected at 24 h after inoculation. Cell
extracts were carried out according to the method used by Remize
et al. (2000). Activities of pyruvate carboxylase, malate dehydroge-
nase, isocitrate lyase and malate synthase were assayed based on
the method described previously (Zelle et al., 2008). The method
used for determining fumarase activity (conversion of L-malic acid
to fumaric acid) was identical to the method of Kanarek and Hill
(1964). The fumarase activity using fumaric acid as the substrate
was assayed according to the method described by Xu et al.
(2012). The fumarate reductase assay condition differed from ma-
late dehydrogenase assay in that 1 mmol/L fumarate was added as
the substrate. BSA was used as a standard to determine the total
protein concentrations of the crude extract (Lowry et al. 1951).
The results of enzyme activity analyses were carried out in
triplicate.
2.5. Shake flask fermentation conditions
Synthetic media contained 3 g/L KH2PO4, 6.6 g/L K2SO4, 0.25 g/L
MgSO4, 15 mg/L EDTA, 4.5 mg/L ZnSO4
7H2O, 0.3 mg/L CoCl2
6H2O,
1 mg/L MnCl2
4H2O, 0.3 mg/L CuSO4
4H2O, 4.5 mg/L CaC12
H2O,
3 mg/L FeSO4
7H2O, 0.4 mg/L NaMoO4
2H2O, 1 mg/L H3BO4,
0.1 mg/L KI. Vitamin solution and indicated urea (both were
filter-sterilized) were added to the media after sterilization. Final
vitamin concentrations per liter were: biotin, 0.05 mg; calcium
pantothenate, 1 mg; nicotinic acid, 1 mg; inositol, 25 mg; vitamin
B1, 1 mg; vitamin B6, 1 mg and para-aminobenzoic acid, 0.2 mg.
Glucose was separately sterilized and added to the media. Accurate
weight of CaCO3 was sterilized by heating the solid in glass tube at
160 °C and 6 h.
Single clone was inoculated to SD media and cultured for 48 h
in a 100 mL flask containing 20 mL medium. Then, the yeast cells
were collected by centrifugation, and resuspended with fresh syn-
thetic medium. The cells were further inoculated to a 500 mL flask
containing 100 mL medium with an initial OD660 at 0.7 ± 0.05 and
addition of indicated CaCO3. Fermentation was carried out at
30 °C and 200 rpm in a rotary shaker. Fermentation samples were
collected every 6 or 12 h. At the last few hours of fermentation,
samples were taken out every 2 h. All shake flask fermentation re-
sults were performed by three independent experiments.
234 D. Yan et al. / Bioresource Technology 156 (2014) 232–239

Glucose

NAD+ NADH
Dihydroxyacetone
Glycerol-3-phosphate x phosphate
Glyceraldehyde-3-phosphate
GPD1p NAD+

Glycerol NADH
1,3-Bisphosphoglycerate
Acetate
NAD+ NADH PDC
Ethanol Acetaldehyde x Pyruvate

PYC2 PDH
CO2 Acetyl-CoA
Oxaloacetate
NADH
MDH3R
NAD+
TCA cycle
Malate Malate

FumC x FUM1p x FUM1p

Fumarate
Fumarate
NADH
FRDS1
NAD+ Mitochondria
Succinate
Fig. 1. Metabolic engineering method of S. cerevisiae used in this study for succinate production. PDC, pyruvate decarboxylase; PDH, pyruvate dehydrogenase.

Table 1
Strains and plasmids used in this study.

Strain or plasmid Genotype Source

Strains
TAM MATa ura3-52; Dpdc1; Dpdc5; Dpdc6 Maris et al. (2004)
TAMH MATa ura3-52; Dhis3; Dpdc1; Dpdc5; Dpdc6 {YIP, pRS313} This study
TAMHf MATa ura3-52; Dhis3; Dfum1; Dpdc1; Dpdc5; Dpdc6 {YIP, pRS313} This study
TAMHfg MATa ura3-52; Dhis3; Dfum1;Dgpd1; Dpdc1; Dpdc5; Dpdc6 {YIP, pRS313} This study
PMCF MATa ura3-52; Dhis3; Dpdc1; Dpdc5; Dpdc6 {YIP-PYC2MDH3R, pRS313CF} This study
PMCFf MATa ura3-52; Dhis3; Dfum1; Dpdc1; Dpdc5; Dpdc6 {YIP-PYC2MDH3DSKL, pRS313CF} This study
PMCFfg MATa ura3-52; Dhis3; Dfum1;Dgpd1; Dpdc1; Dpdc5; Dpdc6 {YIP-PYC2MDH3R, pRS313CF} This study
Plasmids
YIPmcherry Integration expression vector, URA3 marker Lab collection
pRS313 Centromere vector, HIS3 marker Lab collection
YIP-PYC2 Integration expression vector, URA3 marker, PGPD1-PYC2 This study
YIP-MDH3R Integration expression vector, URA3 marker, PGPD1-PYC2, PTEF2-MDH3R This study
YIP-PYC2MDH3R Integration expression vector, URA3 marker, PGPD1-PYC2, PTEF2-MDH3R This study
pRS313C Centromere vector, HIS3 marker, PPGK1-EcFumC. This study
pRS313F Centromere vector, PTDH3-FRDS1 This study
pRS313CF Centromere vector, PPGK1-EcFumC, PTDH3-FRDS1 This study

2.6. Bioreactor batch fermentation conditions
Aerobic batch cultivation was conducted at 30 °C with a
medium volume of 1 L in a 3-L Bioflo 110 fermentor (New
Brunswick Scientific, Edison, NJ, USA). Synthetic medium was
identical to description above except that 0.8 g/L urea, 100 g/L
glucose, 1 mg/L biotin and 0.55 g/L CaCl2 were added. The pH
was controlled by automatic addition of 10 M KOH. The biore-
actors sparged with 2 vvm gas and stirred at 800 rpm, which
maintained the dissolved oxygen concentrations above 30% of
air saturation as measured by an online probe (Mettler-Toledo
Process Analytical Instruments, Wilmington, MA). Pure CO2
was mixed with air (Qianxijingcheng Co., Beijing, China) when
gaseous CO2 concentrations of 5%, 10% and 15% (CO2-enriched
fermentations) were needed. Antifoam (Methylsilicone oil 201,
Xilong Chemical Co. Ltd, China) was used to control foaming.
Precultures were grown on SD media in 500 mL round-bottom
flasks (200 mL media volumes) at 30 °C, 200 rpm. The bio-
masses were centrifuged, washed and inoculated in bioreactors
with initial OD660 at 1 ± 0.05. Samples were taken out at every
6 or 12 h and stored at 20 °C. Three independent experiments
were performed.
 D. Yan et al. / Bioresource Technology 156 (2014) 232–239
2.7. Analytical methods
Optical density (OD) of the fermentation broth was detected by
a spectrophotometer (723N, Shanghai Precision & Scientific Instru-
ment Co. Ltd, China) at 660 nm. Extracellular concentrations of glu-
cose, pyruvate, fumarate, succinate, a-ketoglutarate, malate and
glycerol were analyzed by high performance liquid chromatogra-
phy using a Bio-Rad Aminex HPX-87H column (7.8 mm  300 mm)
and HP1200 chromatography working station system (Agilent, Co.
Ltd USA) equipped with UV absorbance and refractive index detec-
tors. The column was eluted with 5 mM H2SO4 at a flow rate of
0.6 mL/min and at 50 °C.
3. Results and discussion
3.1. Enzyme activity assays of the constructed reductive pathway
As the functionality of the reaction chain as a whole was essen-
tial for succinate production through the reductive TCA pathway,
all four candidate enzymes were overexpressed and assayed. Pyru-
vate carboxylase (PYC) exhibits low activity (0.02 ± 0.01 U/mg) in
the parent strain (Table 2). Its activity must be amplified to en-
hance the reductive carboxylation flux. PYC2p was targeted be-
cause it exhibits a higher affinity towards pyruvate compared
with its isomerase PYC1p in S. cerevisiae (Stucka et al., 1991). The
enzyme activity of PYC increased from 0.02 ± 0.01 U/mg to
0.34 ± 0.05 U/mg protein upon overexpression of PYC2 (Table 2).
Malate dehydrogenase (MDH) exhibited obvious activity in the
parent strain, TAMH. However, the activity detected was actually
due to mitochondrial MDH1p and peroxisomal MDH3p
(McAlister-Henn et al., 1995). MDH3p with its last three amino
acid residues removed could be re-targeted to the cytosol, where
it exhibits catalytic activity (McAlister-Henn et al., 1995). Activity
of MDH improved to 22.76 ± 3.57 U/mg protein when the cytosol
retargeted MDH3p (MDH3Rp) was overexpressed (Table 2). The
activities of fumarase in TAMH were 0.19 ± 0.03 (malate to fuma-
rate) and 4.87 ± 0.52 U/mg protein (fumarate to malate) in TAMH.
When EcFumC was overexpressed, the respective enzyme activities
increased to 0.51 ± 0.07 and 5.29 ± 0.64 U/mg protein, respectively.
However, the enzyme activities of FUM in conversion of malate to
fumarate were still far below the reverse direction in PMCF. In-
deed, native FUM1p exhibits substantially higher affinity towards
fumarate than towards malate (Pines et al., 1996). A reversible bal-
ance of the catalysis reaction between malate and fumarate was
eventually achieved by overexpression of EcFumC and deletion of
FUM1 in the resulted strain PMCFf. At the same time, FRD activity
was visibly improved to 1.39 ± 0.28 U/mg protein by overexpres-
sion of FRDS1p which located in the cytosol. Enzyme activity as-
says indicated that the constructed pathway was efficient and
capable of producing succinate from pyruvate.
Table 2
Enzyme activity assays of key enzymes in the control strains and engineered strains.
Strainsa Enzyme activityb (U/mg protein)c
PYC MDH
TAMH 0.02 ± 0.01 2.47 ± 0.82
PMCF 0.34 ± 0.05 22.76 ± 3.57
TAMHf 0.02 ± 0.01 2.23 ± 0.62
PMCFf 0.32 ± 0.08 24.42 ± 2.06
a
Yeast strains were grown in shake flasks on glucose SD media at 24 h.
b
PYC, pyruvate carboxylase; MDH, malate dehydrogenase; FUM, fumarase; FRD, fumarate reductase. N.D. not detected (less than 0.001 U/mg protein).
c
1 U represents the amount of enzyme required for 1 lmol substrate conversion in 1 min.
235
3.2. Effect of the reductive pathway on succinate production
There was no obvious improvement in succinate level when the
constructed pathway was introduced into the FUM1 undeleted
strain (Table 3). However, the malate titer increased from
0.13 ± 0.05 g/L to 1.22 ± 0.14 g/L in PMCF indicating that FUM
was the major bottleneck for the reductive pathway. When FUM1
was deleted, malate accumulation was undetectable while the suc-
cinate titer in PMCFf improved to 6.17 ± 0.34 g/L. Succinate
yield likewise improved 65-fold, from 0.003 ± 0.001 to
0.197 ± 0.01 mol/mol glucose. At the same time, as shown in
Table 3, the titers of pyruvate, fumarate, and glycerol decreased
from 23.82 ± 0.41 g/L, 0.86 ± 0.03 g/L and 5.57 ± 0.79 g/L in the par-
ent strain to 18.01 ± 0.74 g/L, 0.21 ± 0.02 g/L and 4.24 ± 1.28 g/L,
respectively, in PMCFf. Biomass and glucose consumption rates
were slightly higher than those of the reference strain.
Succinate is produced through reductive pathway with phos-
phoenolpyruvate carboxylase (PPC) and phosphoenolpyruvate
carboxykinase (PCK) in prokaryotic native producers such as A. suc-
cinogenes and E. coli et al. (Zhang et al., 2009). However, S. cerevisiae
lacks an encoding gene for PPC, and its native PCK can only catalyse
oxaloacetate decarboxylation (Zelle et al., 2010). In this study, the
key step of pyruvate carboxylation was enhanced by overexpres-
sion of PYC2p. This strategy effectively directs the carbon flux from
pyruvate into the reductive TCA pathway. Then, the three other
overexpressed enzymes (MDH3Rp, EcFumCp and FRDS1p) enable
catalytic conversion of oxaloacetate to succinate in the cytosol.
Through this combined strategy, an effective reductive TCA path-
way for succinate was constructed.
3.3. Identification of succinate synthetic route
In PMCFf, succinate may be produced through three routes: the
reconstructed reductive pathway, the TCA oxidative pathway and
the glyoxylate shunt (Fig. S2). The succinate production route
was investigated in order to elevate the efficiency of the reductive
pathway. For the glyoxylate cycle route, the enzyme activities of
isocitrate lyase (ICL) and malate synthase (MAS) were assayed.
When PMCFf was grown on ethanol-amended media in the control
condition, the activities of ICL and MAS were 0.07 ± 0.03 and
0.16 ± 0.07 U/mg protein; however, their activities fell below
detectable levels when glucose was used as the sole carbon source.
In wild type S. cerevisiae, ICL and MAS are repressed by glucose
catabolites and activated in glucose-negative media (Dduntze
et al., 1969). These results confirmed that glyoxylate shunt did
not contribute to succinate production in PMCFf.
To determine whether succinate was derived from the con-
structed reductive TCA pathway, the effects of individual or com-
bined overexpression of PYC2, MDH3R, EcFumC and FRDS1 on
succinate production were investigated. As shown in Fig. 2, succi-
nate titers were not influenced by individual overexpression of
FUM FRD
(Malate to fumarate) (Fumarate to malate)
0.19 ± 0.03 4.87 ± 0.52 N.D.
0.51 ± 0.07 5.29 ± 0.64 1.38 ± 0.36
N.D. N.D. N.D.
0.34 ± 0.04 0.36 ± 0.07 1.39 ± 0.28
236 D. Yan et al. / Bioresource Technology 156 (2014) 232–239

Table 3
Shake flask cultivation characteristics of metabolites in reference strains and engineered strains.

Straina Glucose consumed (g/L)b Organics (g/L) in culture supernatantb Yield of succinate (mol/mol glucose) Biomass (OD660)
Pyruvate Malate Fumarate Glycerol Succinate
TAMH 48.42 ± 0.38 23.92 ± 0.19 0.13 ± 0.05 0.27 ± 0.04 5.72 ± 0.82 0.32 ± 0.05 0.010 ± 0.001 14.61 ± 0.63
PMCF 49.46 ± 0.22 22.78 ± 0.23 1.22 ± 0.14 0.29 ± 0.06 5.43 ± 0.64 0.33 ± 0.07 0.010 ± 0.001 14.97 ± 0.54
TAMHf 48.34 ± 0.46 23.82 ± 0.41 0.15 ± 0.09 0.86 ± 0.03 5.57 ± 0.79 0.11 ± 0.04 0.003 ± 0.001 12.88 ± 0.26
PMCFf 49.73 ± 0.09 18.01 ± 0.74 N.D. 0.21 ± 0.02 4.24 ± 1.28 6.17 ± 0.34 0.197 ± 0.010 15.19 ± 0.38
a
Fermentation experiments were carried out in a 500 ml shake flasks with 100 ml synthetic medium, 50 g/L initial glucose, 1 g/L urea and 30 g/L CaCO3.
b
Glucose consumption and concentrations of organics were assayed at 70 h after inoculation when glucoses were nearly exhausted.

7 50 20
(a)
Malate, fumarate, succinate (g/L)

18
6
40 16
5
14

Biomass (OD660)
Glucose (g/L)
4 30 12
10
3
20 8
2
6
1 10 4
2
0
_ _ _ _ _ _ _ + PYC2
+ + + + 0 0
_ _ _ _ _ + + MDH3R
+ + + + + 0 10 20 30 40 50 60 70
_ _ _ _ _ _ _ + + + + FumC
+ Times after inoculation (h)
_ _ _ _ _ + + _ + + + FRDS1
+
10 10
Fig. 2. Concentrations of three key products (malate, fumarate, and succinate)
obtained in shake flask experiments of different strains overexpressing different
(b)
combinations of pyruvate carboxylase, malate dehydrogenase, fumarase and 8 8
fumarate reductase. Concentrations of organic acids were assayed at 70 h after
inoculation (synthetic medium with 50 g/L initial glucose, 1 g/L urea and 30 g/L
CaCO3). Grey columns represent malate; blank columns represent fumarate; dark

Glycerol (g/L)
Succinate (g/L)

6 6
columns represent succinate. A plus sign indicates that the gene is overexpressed.
The minus sign means the particular gene is not overexpressed.
4 4

PYC2, MDH3R or EcFumC. However, fumarate was accumulated

when FUM1 was deleted in the reference strain as demonstrated 2 2
by Arikawa et al. (1999). The fumarate titer further increased upon
overexpression of PYC and MDH. Similar to the results reported by 0 0
Xu et al., enhanced fumarate level is also observed by overexpres- 0 10 20 30 40 50 60 70
sion of RoPYC, RoMDH and RoFUM1 in a fum1-deficient S. cerevisiae Times after inoculation (h)
strain (Xu et al., 2013). Interestingly, the succinate titer increased
3-fold (0.32 ± 0.13 g/L) by individual overexpression of FRDS1 and Fig. 3. Comparisons of glucose consumption, cell growth, succinate production and
glycerol titer during shake flask fermentation in PMCFf, PMCFfg and reference strain
up to 0.92 ± 0.22 g/L by further co-expression of PYC2 and MDH3R. TAMHfg. (a) Filled symbols represent glucose consumption curve, opened symbols
This increase in succinate levels arose because of the absence of represent growth kinetics curve; (b) filled symbols represent succinate formation
FUM, the interrupted TCA pathway, and further reduction by FRD curve, opened symbols represent glycerol formation curve. Circles represent
in the cytosol. Simultaneous expression of MDH3R, EcFumC and reference strain TAMHfg, up triangles represent PMCFfg and down triangles
represent PMCFf.
FRDS1 raised succinate levels to 2.12 ± 0.18 g/L. A maximum succi-
nate titer of 6.17 ± 0.34 g/L was reached by co-overexpression of
the four enzymes. These results together demonstrated that the
reductive pathway was an effective strategy for enhancing succi- reductive pathway was introduced, glucose consumption rate
nate production. and biomass formation in PMCFfg were rescued. Succinate titer
also improved to 8.09 ± 0.28 g/L with a yield of 0.258 ±
3.4. Effects of the deletion of GPD1 on succinate production 0.008 mol/mol glucose in PMCFfg. Meanwhile, the glycerol titer
was further reduced to 2.18 ± 0.37 g/L.
In the reductive TCA pathway, production of 1 mol of succinate Wild-type yeast achieves redox balance through the formation
consumes 2 mol of NADH. However, excess NADH is largely con- of ethanol. In a pdc-deficient strain though, the deletion of pdc re-
sumed by glycerol formation and respiration to maintain the redox sults in imbalanced redox metabolism and high concentrations of
balance in the evolved pdc-deficient strain TAM. Glycerol synthesis accumulated glycerol even under aerobic conditions (Maris et al.,
is mainly controlled by GPD1p under aerobic conditions. Therefore, 2003). Beneficial effects on cell growth and glucose consumption
GPD1 deletion as a potential mean of further improving succinate are also been observed by overexpression of the E. coli pyridine
production was investigated. As shown in Fig. 3, although glycerol nucleotide transhydrogenase UdhA, which converts NADH/NADP+
levels were reduced to 3.51 ± 0.17 g/L, glucose consumption and to NAD+/NADPH, in pdc-deficient yeast (Wang et al., 2012). The
biomass visibly decreased more in TAMHfg. However, when the NADH-consuming reductive pathway in this study had similar
 D. Yan et al. / Bioresource Technology 156 (2014) 232–239 237

influences on cell growth and glucose consumption rate. Accord-

49.71 ± 0.36
17.24 ± 0.55
19.82 ± 0.49
2.93 ± 0.37
8.62 ± 0.21
0.27 ± 0.01
ingly, sufficient supplies of NADH and carbon flux were beneficial
for succinate production.

2
3.5. Effects of CaCO3, urea and biotin levels on succinate production

49.43 ± 0.22
16.73 ± 0.39
18.54 ± 0.25
2.19 ± 0.41
9.18 ± 0.23
0.29 ± 0.01
PMCFfg produced copious amounts of organic acids, which made
it necessary to add neutralizing compounds. MgCO3 and NaHCO3

1
are commonly used for neutralizing succinic acid in prokaryotic
microorganisms. However, the high basicity of MgCO3 and NaHCO3

49.81 ± 0.14
15.66 ± 0.78
17.93 ± 0.19
2.08 ± 0.83
9.41 ± 0.26
0.30 ± 0.01
(pH > 8) made them unsuitable for use in cultivating yeast. Interest-
ingly, the low buffering capacity of CaCO3 had a surprisingly strong

0.75
delaying effect on yeast growth (Fig. 4a). The log phases of the yeast
were delayed to 30, 48 and 60 h by direct addition of 30, 40 and

49.76 ± 0.44
14.12 ± 0.51
17.07 ± 0.63
1.84 ± 0.32
9.98 ± 0.31
0.32 ± 0.01
50 g/L CaCO3, respectively, after inoculation. However, immediate
addition of 5 g/L CaCO3 followed by further addition of 30, 40 or
50 g/L CaCO3 after 24 h of cultivation had little influence on cell

0.5
growth, compared with KOH control conditions. At the same time,
succinate titers and productivities also exhibited notable improve-

49.63 ± 0.19
13.96 ± 0.42
17.42 ± 0.48
2.13 ± 0.52
9.32 ± 0.25
0.30 ± 0.01
ment by the addition of CaCO3 compared with the KOH control
(Fig. 4b). Succinate production increased to 8.17 ± 0.24 g/L with a

0.25
productivity of 0.117 ± 0.003 g/L/h by addition of 40 g/L CaCO3.
CaCO3 gives enriched concentrations of Ca2+ and bicarbonate,

Biotin (mg/L)

41.82 ± 0.76
11.58 ± 0.16
16.02 ± 0.38
3.18 ± 0.74
3.42 ± 0.27
0.13 ± 0.01
which influences the production of dicarboxylic acids. Indeed, an
enriched extracellular CO2 concentration decreases the free energy
barrier of pyruvate carboxylase. Additionally, Ca2+ has been

0
49.58 ± 0.32
18.76 ± 0.82
21.64 ± 0.54
4.39 ± 0.44
5.88 ± 0.39
0.19 ± 0.01
18
(a)
2

16
14
49.03 ± 0.29
14.16 ± 0.43
18.79 ± 0.61
2.54 ± 0.48
8.72 ± 0.09
0.28 ± 0.00
12
Biomass (OD660)

10
0.5

8
48.84 ± 0.66
13.84 ± 0.56
18.42 ± 0.83
2.42 ± 0.74
9.04 ± 0.13
0.29 ± 0.00

6
4
0.4

2
Effects of nitrogen and biotin levels on succinate production in engineered strain PMCFfg.

0
39.76 ± 0.22
12.63 ± 0.73
16.38 ± 0.51
2.12 ± 0.82
7.64 ± 0.23
0.31 ± 0.01

0 10 20 30 40 50 60 70
Times after inoculation (h)
0.3

10
(b) 0.30
18.42 ± 1.38
5.96 ± 0.62
6.76 ± 0.49
0.96 ± 0.57
3.72 ± 0.46
0.32 ± 0.04
Productivity of succinate (g/L/h)

8
0.25
0.1
Succinate (g/L)

6 0.20
3.27 ± 0.93
1.48 ± 0.59
1.07 ± 0.04
Urea (g/L)

0.15
4
0.10
–
–
–
0

2
Yield of succinate (mol/mol glucose)

0.05

0 0.00
Control 5+30 5+40 5+50 30 40 50
CaCO3 concentration (g/L)
Glucose consumed (g/L)

Fig. 4. Effect of CaCO3 on cell growth and succinate production in PMCFfg. (a) Filled
Biomass (OD660)

circles represent pH was adjusted to 5.8 with KOH every 4 h, opened up triangles,
Succinate (g/L)
Pyruvate (g/L)
Glycerol (g/L)

opened down triangles and opened squares represent 30 g/L, 40 g/L and 50 g/L
Parameters

CaCO3 was added after inoculation, respectively. Filled up triangles, filled down
triangles and filled squares represent 30 g/L, 40 g/L and 50 g/L CaCO3 was added
Table 4

after 24 precultured with 5 g/L initial CaCO3, respectively; (b) succinate titer and
productivity with different CaCO3 at 70 h cultivation. Dark columns represent
succinate titer, shadow columns represent succinate productivity.
238 D. Yan et al. / Bioresource Technology 156 (2014) 232–239

observed to strongly influence the PYC activity in Torulopsis glabrata
(Liu et al., 2007).
Nitrogen limitation accordingly limits the amount of biomass
formed and the consumption of catabolic energy substrates
(Larsson et al., 1993). To determine whether succinate production
could be regulated by the nitrogen level, effects of varying levels of
urea (the nitrogen source) within 0.1–2 g/L were investigated. As
urea concentrations fell across this range, cell growth and glucose
consumption likewise decreased (Table 4). However, although the
yields of succinate were higher when 0.1 and 0.3 g/L urea were
added (0.32 ± 0.04 and 0.31 ± 0.01 mol/mol glucose), glucose could
not be completely consumed at these concentrations by further
extending the fermentation period. Succinate titer improved to
9.04 ± 0.13 g/L upon addition of 0.4 g/L urea, the observed opti-
mum concentration. From these results, reduced urea levels
proved to be beneficial for succinate production, as similarly ob-
served from the results of heightened fumarate production in engi-
neered S. cerevisiae (Xu et al., 2013) and Rhizopus oryzae (Ding et al.,
2011).
The key step of conversion of pyruvate to oxaloacetate is cata-
lyzed by PYC which is a biotin-containing enzyme. Hence, the ef-
fects of biotin levels on succinate production were investigated
with 0.4 g/L urea used as nitrogen source. As shown in Table 4,
the succinate titer substantially decreased without any addition
of biotin. When biotin levels were increased from 0 to 0.5 mg/L,
levels of pyruvate, glycerol and biomass decreased, whereas the
succinate titer increased to 9.98 ± 0.31 g/L. Levels of biomass and
pyruvate were 10% and 5% higher at 0.75 mg/L biotin, compared
with levels upon addition of 0.5 mg/L biotin. However, succinate ti-
ter was dropped when further increased the biotin levels from
0.75 mg/L to 2 mg/L. In summary, succinate production could be
effectively regulated by the addition of appropriate levels of CaCO3,
urea and biotin.
3.6. Succinate production in bioreactors at low pH value
CaCO3-buffered cultivation in a shake flask is a process done in a
changing environment (pH 6.7–5.7). Therefore, succinate fermen-
tation at low pH values in a bioreactor was further investigated.
The pH values and exogenous CO2 concentrations were varied to
find the optimal levels for succinic acid production. As shown in
Fig. 5, maxima for pyruvate titer (48.21 ± 1.54 g/L), cell density
(47.74 ± 1.31 at OD660) and glucose consumption rate were at-
tained at pH 4.8 which was the optimal pH value for S. cerevisiae
growth. However, only 1.69 ± 0.26 g/L succinate was obtained at
this pH. Interestingly, similar succinate titers were obtained
(6.84 ± 0.29 and 6.31 ± 0.33 g/L) when fermentations were con-
ducted at pH 5.8 and pH 3.8, respectively. Although the fermenta-
tion period was extended from 85 ± 1 to 97 ± 2 h at these pH levels,
concentrations of the byproduct pyruvate significantly decreased
from 42.66 ± 1.29 to 29.19 ± 1.06 g/L with 100 g/L initial glucose
and air sparging. However, when the pH was further decreased
to 3.0, nearly half of the glucose could not be consumed
(Fig. S3a). Hence, further exploration on the effects of CO2 supple-
mentation was conducted at pH 3.8. When the CO2 concentration
was raised from 5% to 10%, the succinate titers increased from
9.77 ± 0.49 g/L to 12.97 ± 0.42 g/L. Glucose consumption period
was extended from 108 ± 0 to 120 ± 1 h by increasing exogenous
CO2 concentrations. However, the decreased levels of pyruvate
(from 25.36 ± 1.12 g/L to 21.62 ± 1.24 g/L) and biomass (from
32.52 ± 1.35 to 28.16 ± 1.42 at OD660) were observed. The succinate
titer was not improved by further increasing the CO2 concentration
to 15% (Fig. S3b).
Low extracellular pH and the presence of organic acids result in
reduced cytosolic pH (Carmelo et al., 1997). This condition leads to
generation of excess bicarbonate at equilibrium due to excess
carbon dioxide. However, surfeit bicarbonate is harmful to the

100 (a) 50 (b)

Biomass (OD660)

80 40
Glucose (g/L)

60 30

40 20

20 10

0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Times after inoculation (h) Times after inoculation (h)
50
(c) 14 (d)
40 12
Pyruvate (g/L)

Succinate (g/L)

10
30
8

20 6

4
10
2

0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Times after inoculation (h) Times after inoculation (h)

Fig. 5. Comparisons of the fermentation profile of PMCFfg under different pH values and gas sparging in bioreactors. (a) Glucose consumption; (b) growth curve; (c) pyruvate
formation curve; (d) succinate production curve. Filled symbols represent sparging with air under pH 3.8 (down triangles), pH 4.8 (circles), pH 5.8 (up triangles). Opened
symbols represent fermentation under pH 3.8 and sparging with 5% CO2 (down triangles) and 10% CO2 (circles).
 D. Yan et al. / Bioresource Technology 156 (2014) 232–239
metabolism of the organism and the activity of carboxylase (Wang
et al., 2009). Although the succinate yield obtained at pH 3.8 was
slightly below that obtained in the shake flask experiments, fer-
mentation at low pH greatly reduced the demand for active neu-
tralization and the risk of potential bacterial contamination. Most
importantly, about 50% of succinate exited in the form of free acid
at pH 3.8 (Li et al., 2010). Besides, the concentrations of byproducts
aside from pyruvate were very low (fumarate, <0.25 g/L;
a-ketoglutarate, <0.3 g/L; other organic acids, undetected) in the
bioreactor experiments. Low titers of byproducts and high levels
of succinate in the free acid form would greatly reduce the cost
of downstream processing of succinic acid.
4. Conclusion
In the present study, enhanced production of succinate was
achieved through the overexpressing of PYC2, MDH3R, E. coli FumC
and FRDS1 in a pdc and fum1-deficient strain of S. cerevisiae. Succi-
nate was mainly produced through the constructed reductive path-
way. In addition, succinate production was further improved by
the deletion of GPD1. Regulation of urea, CaCO3 and biotin levels
were beneficial for succinate production in shake flask experi-
ments. Bioreactor experiments also confirmed that the engineered
strain could effectively produce succinic acid under low pH condi-
tions with specific levels of CO2 enrichment.
Acknowledgements
Authors are thankful to Prof. J.T. Pronk in Delft University of
Technology for providing the S. cerevisiae TAM strain. The authors
gratefully acknowledge Prof. Wang Qinhong for providing the
plasmids pRS313, YIPmcherry and kindly guidance. This work
was supported by National High Technology Research and Devel-
opment Program of China (863 Project, Nos. 2012AA101807,
2011AA02A203 and 2012AA022301).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.biortech.2014.
01.053.
References
Agren, R., Otero, J.M., Nielsen, J., 2013. Genome-scale modeling enables metabolic
engineering of Saccharomyces cerevisiae for succinic acid production. J. Ind.
Microbiol. Biotechnol. 40, 735–747.
Akada, R., Kitagawa, T., Kaneko, S., Toyonaga, D., Ito, S., Kakihara, Y., Hoshida, H.,
Morimura, S., Kondo, A., Kida, K., 2006. PCR-mediated seamless gene deletion
and marker recycling in Saccharomyces cerevisiae. Yeast 23, 399–405.
Arikawa, Y., Kuroyanagi, T., Shimosaka, M., Muratsubaki, H., Enomoto, K., Kodaira,
R., Okazaki, M., 1999. Effect of gene disruptions of the TCA cycle on production
of succinic acid in Saccharomyces cerevisiae. J. Biosci. Bioeng. 87, 28–36.
Carmelo, V., Santos, H., Sá-Correia, I., 1997. Effect of extracellular acidification on
the activity of plasma membrane ATPase and on the cytosolic and vacuolar pH
of Saccharomyces cerevisiae. Biochim. Biophys. Acta 1325, 63–70.
Dduntze, W., Neumann, D., Atzpodien, W., Holzer, H., Gancedo, J., 1969. Studies on
the regulation and localization of the glyoxylate cycle enzymes in
Saccharomyces cerevisiae. Eur. J. Biochem. 10, 83–89.
Ding, Y., Li, S., Dou, C., Yu, Y., Huang, H., 2011. Production of fumaric acid by
Rhizopus oryzae: role of carbon–nitrogen ratio. Appl. Biochem. Biotechnol. 164,
1461–1467.
Kanarek, L., Hill, R.L., 1964. The preparation and characterization of fumarase from
swine heart muscle. J. Biol. Chem. 239, 4202–4206.
Kim, S.J., Seo, S.O., Jin, Y.S., Seo, J.H., 2013. Production of 2,3-butanediol by
engineered Saccharomyces cerevisiae. Bioresour. Technol. 146, 274–281.
Larsson, C., von Stockar, U., Marison, I., Gustafsson, L., 1993. Growth and metabolism
of Saccharomyces cerevisiae in chemostat cultures under carbon-, nitrogen-, or
carbon- and nitrogen-limiting conditions. J. Bacteriol. 175, 4809–4816.
239
Li, Q., Wang, D., Wu, Y., Li, W., Zhang, Y., Xing, J., Su, Z., 2010. One step recovery of
succinic acid from fermentation broths by crystallization. Sep. Purif. Technol.
72, 294–300.
Li, Y., Li, M., Zhang, X., Yang, P., Liang, Q., Qi, Q., 2013. A novel whole-phase succinate
fermentation strategy with high volumetric productivity in engineered
Escherichia coli. Bioresour. Technol. 149, 333–340.
Liang, L., Liu, R., Li, F., Wu, M., Chen, K., Ma, J., Jiang, M., Wei, P., Ouyang, P., 2013.
Repetitive succinic acid production from lignocellulose hydrolysates by
enhancement of ATP supply in metabolically engineered Escherichia coli.
Bioresour. Technol. 143, 405–412.
Liu, L., Li, Y., Zhu, Y., Du, G., Chen, J., 2007. Redistribution of carbon flux in Torulopsis
glabrata by altering vitamin and calcium level. Metab. Eng. 9, 21–29.
Liu, R., Liang, L., Cao, W., Wu, M., Chen, K., Ma, J., Jiang, M., Wei, P., Ouyang, P., 2012.
Succinate production by metabolically engineered Escherichia coli using
sugarcane bagasse hydrolysate as the carbon source. Bioresour. Technol. 135,
574–577.
Liu, R., Liang, L., Li, F., Wu, M., Chen, K., Ma, J., Jiang, M., Wei, P., Ouyang, P., 2013.
Efficient succinic acid production from lignocellulosic biomass by simultaneous
utilization of glucose and xylose in engineered Escherichia coli. Bioresour.
Technol. 149, 84–91.
Lowry, O.H., Rosebrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurement
with the Folin phenol reagent. J. Biol. Chem. 193, 265–275.
Maris, A.J.A., Geertman, J.M.A., Vermeulen, A., Groothuizen, M.K., Winkler, A.A.,
Piper, M.D.W., van Dijken, J.P., Pronk, J.T., 2004. Directed evolution of pyruvate
decarboxylase-negative Saccharomyces cerevisiae, yielding a C2-independent,
glucose-tolerant, and pyruvate-hyperproducing yeast. Appl. Environ. Microbiol.
70, 159–166.
McAlister-Henn, L., Steffan, J.S., Minard, K.I., Anderson, S.L., 1995. Expression and
function of a mislocalized form of peroxisomal malate dehydrogenase (MDH3)
in yeast. J. Biol. Chem. 270, 21220–21225.
McKinlay, J., Vieille, C., Zeikus, J.G., 2007. Prospects for a bio-based succinate
industry. Appl. Microbiol. Biotechnol. 76, 727–740.
Muratsubaki, H., Enomoto, K., 1998. One of the fumarate reductase isoenzymes
from Saccharomyces cerevisiae is encoded by the OSM1 Gene. Arch. Biochem.
Biophys. 352, 175–181.
Otero, J.M., Cimini, D., Patil, K.R., Poulsen, S.G., Olsson, L., Nielsen, J., 2013. Industrial
systems biology of Saccharomyces cerevisiae enables novel succinic acid cell
factory. PLoS One 8, e54144.
Pines, O., Even-Ram, S., Elnathan, N., Battat, E., Aharonov, O., Gibson, D., Goldberg, I.,
1996. The cytosolic pathway of L-malic acid synthesis in Saccharomyces
cerevisiae: the role of fumarase. Appl. Microbiol. Biotechnol. 46, 393–399.
Raab, A.M., Gebhardt, G., Bolotina, N., Weuster-Botz, D., Lang, C., 2010. Metabolic
engineering of Saccharomyces cerevisiae for the biotechnological production of
succinic acid. Metab. Eng. 12, 518–525.
Remize, F., Andrieu, E., Dequin, S., 2000. Engineering of the pyruvate dehydrogenase
bypass in Saccharomyces cerevisiae: role of the cytosolic Mg2+ and mitochondrial
K+ acetaldehyde dehydrogenases Ald6p and Ald4p in acetate formation during
alcoholic fermentation. Appl. Environ. Microbiol. 66, 3151–3159.
Song, H., Lee, S.Y., 2006. Production of succinic acid by bacterial fermentation.
Enzyme Microb. Technol. 39, 352–361.
Stucka, R., Dequin, S., Salmon, J.M., Gancedo, C., 1991. DNA sequences in
chromosomes II and VII code for pyruvate carboxylase isoenzymes in
Saccharomyces cerevisiae: analysis of pyruvate carboxylase-deficient strains.
Mol. Gen. Genet. 229, 307–315.
van Maris, A.J., Luttik, M.A., Winkler, A.A., van Dijken, J.P., Pronk, J.T., 2003.
Overproduction of threonine aldolase circumvents the biosynthetic role of
pyruvate decarboxylase in glucose-limited chemostat cultures of Saccharomyces
cerevisiae. Appl. Environ. Microbiol. 69, 2094–2099.
Wang, D., Li, Q., Li, W., Xing, J., Su, Z., 2009. Improvement of succinate production by
overexpression of a cyanobacterial carbonic anhydrase in Escherichia coli.
Enzyme Microb. Technol. 45, 491–497.
Wang, Z., Gao, C., Wang, Q., Liang, Q., Qi, Q., 2012. Production of pyruvate in
Saccharomyces cerevisiae through adaptive evolution and rational cofactor
metabolic engineering. Biochem. Eng. J. 67, 126–131.
Xu, G., Chen, X., Liu, L., Jiang, L., 2013. Fumaric acid production in Saccharomyces
cerevisiae by simultaneous use of oxidative and reductive routes. Bioresour.
Technol. 148, 91–96.
Xu, G., Liu, L., Chen, J., 2012. Reconstruction of cytosolic fumaric acid biosynthetic
pathways in Saccharomyces cerevisiae. Microb. Cell Fact. 11, 1–10.
Zeikus, J., Jain, M., Elankovan, P., 1999. Biotechnology of succinic acid production
and markets for derived industrial products. Appl. Microbiol. Biotechnol. 51,
545–552.
Zelle, R.M., de Hulster, E., van Winden, W.A., de Waard, P., Dijkema, C., Winkler, A.A.,
Geertman, J.M., van Dijken, J.P., Pronk, J.T., van Maris, A.J., 2008. Malic acid
production by Saccharomyces cerevisiae: engineering of pyruvate carboxylation,
oxaloacetate reduction, and malate export. Appl. Environ. Microbiol. 74, 2766–
2777.
Zelle, R.M., Trueheart, J., Harrison, J.C., Pronk, J.T., van Maris, A.J., 2010.
Phosphoenolpyruvate carboxykinase as the sole anaplerotic enzyme in
Saccharomyces cerevisiae. Appl. Environ. Microbiol. 76, 5383–5389.
Zhang, X., Jantama, K., Moore, J.C., Jarboe, L.R., Shanmugam, K.T., Ingram, L.O., 2009.
Metabolic evolution of energy-conserving pathways for succinate production in
Escherichia coli. PNAS 106, 20180–20185.