Efficient bioconversion of crude glycerol from biodiesel to optically pure d-lactate by metabolically engineered Escherichia coli

Article abstract of DOI:10.1039/c3gc41769g

Biodiesel has attracted considerable attention as one of the best choices among alternative and renewable fuels. Large quantities of crude glycerol are produced as a main co-product with increasing biodiesel production. Currently, the problem of waste glycerol utilization needs to be crucially addressed, not only for environmental protection but also for the economy of the biodiesel industry. In this paper, the use of crude glycerol for the production of d-lactate by engineered Escherichia coli was investigated. Engineered E. coli B0013-070 with a homolactic pathway for d-lactate synthesis by elimination of byproduct pathways (ethanol, succinate, formate and acetate) could convert 20 g L^-1 of crude glycerol to 13.6 g L^-1 of d-lactate with a yield of 0.67 g g^-1 glycerol. Overexpression of d-lactate dehydrogenase by a low-copy vector in E. coli B0013-070 resulted in the increased production and yield of d-lactate, in which 14.5 g L^-1 of d-lactate was produced with a yield of 0.72 g g^-1 glycerol from crude glycerol. The effect of temperature on d-lactate fermentation by the engineered strain E. coli B0013-070-pTHldhA was also investigated, and 34 °C and 40 °C were found to be the optimal temperatures for cell growth and lactate production, respectively. The engineered strain B0013-070-pTHldhA produced 100.3 g L^-1 of d-lactate with 99.97% optical purity from 531.5 g of crude glycerol with an overall productivity of 2.78 g L^-1 h^-1 and a yield of 75.4 g per 100 g glycerol (0.77 mol mol^-1) using two phase fermentation combined with a temperature shifting strategy in a 7 L bioreactor. In summary, this paper shows that crude glycerol could be directly converted to d-lactate without any prior purification.

Full text of DOI:10.1039/c3gc41769g

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Efficient bioconversion of crude glycerol from
biodiesel to optically pure D-lactate by
metabolically engineered Escherichia coli†
Cite this: Green Chem., 2014, 16,
342
Xian-zhong Chen,*‡a,c Kang-ming Tian,‡b Dan-dan Niu,b Wei Shen,a,c
Govender Algasan,d Suren Singhd and Zheng-xiang Wang*b
Biodiesel has attracted considerable attention as one of the best choices among alternative and renew-
able fuels. Large quantities of crude glycerol are produced as a main co-product with increasing biodiesel
production. Currently, the problem of waste glycerol utilization needs to be crucially addressed, not only
for environmental protection but also for the economy of the biodiesel industry. In this paper, the use of
crude glycerol for the production of D-lactate by engineered Escherichia coli was investigated. Engineered
E. coli B0013-070 with a homolactic pathway for D-lactate synthesis by elimination of byproduct pathways
(ethanol, succinate, formate and acetate) could convert 20 g L−1 of crude glycerol to 13.6 g L−1 of
D-lactate with a yield of 0.67 g g−1 glycerol. Overexpression of D-lactate dehydrogenase by a low-copy
vector in E. coli B0013-070 resulted in the increased production and yield of D-lactate, in which 14.5 g L−1
of D-lactate was produced with a yield of 0.72 g g−1 glycerol from crude glycerol. The effect of tempera-
ture on D-lactate fermentation by the engineered strain E. coli B0013-070-pTHldhA was also investigated,
and 34 °C and 40 °C were found to be the optimal temperatures for cell growth and lactate production,
respectively. The engineered strain B0013-070-pTHldhA produced 100.3 g L−1 of D-lactate with 99.97%
optical purity from 531.5 g of crude glycerol with an overall productivity of 2.78 g L−1 h−1 and a yield of
75.4 g per 100 g glycerol (0.77 mol mol−1) using two phase fermentation combined with a temperature
shifting strategy in a 7 L bioreactor. In summary, this paper shows that crude glycerol could be directly
converted to D-lactate without any prior purification.
Received 28th August 2013,
Accepted 21st October 2013
DOI: 10.1039/c3gc41769g
www.rsc.org/greenchem
2016, which implied that approximately 4 billion gallons of
crude glycerol would be produced.2 The problem of waste gly-
Introduction
Biodiesel (fatty acid methyl esters), one of the best choices cerol utilization needs to be crucially addressed if biodiesel
among alternative fuels, has attracted considerable attention usage is to be enhanced.3 There are some chemical applications
because it has the potential to completely displace its petroleum which utilize waste glycerol, but a purification process is needed
counterpart.1 It is produced from triacylglycerols by transesteri- to use glycerol as a chemical source.4 Therefore, development of
fication with short chain alcohols that employs methanol and sustainable biorefineries based on crude glycerol is imperative
alkali catalysis. This process would generate large quantities of to favor the economy of the biodiesel industry by reducing the
glycerol into the market in the near future. It was projected that costs associated with the disposal of residues and increasing
the world biodiesel market would reach 37 billion gallons by the production of value-added chemicals.5
Microbial conversion of glycerol to high value chemicals
such as 1,3-propanediol,6 poly(3-hydroxypropionate),7 succi-
aKey Laboratory of Industrial Biotechnology, Ministry of Education,
nate8 and lactate9 is attractive in that the process is relatively
Jiangnan University, Wuxi 214122, China. E-mail: xzchen@jiangnan.edu.cn
simple and does not generate toxic byproducts. In most cases,
bKey Laboratory of Industrial Fermentation Microbiology of Ministry of Education,
refined glycerol was used to produce a value-added compound.
Tianjin University of Science & Technology, Tianjin 300457, China.
However, the interest in low-grade glycerol as a substrate is
high but there are only a few investigations examining the bio-
conversion of this substance into high value chemicals.10,11
Lactate and its derivatives have versatile applications in the
food, pharmaceutical, textile, chemical and polymer industries.12
For example, D-lactate is a critical chiral intermediate and a
E-mail: zx.wang@tust.edu.cn
cSchool of Biotechnology, Jiangnan University, Wuxi 214122, China
dDepartment of Biotechnology & Food Technology, Faculty of Applied Sciences,
Durban University of Technology, Durban, 4001, South Africa
†Electronic supplementary information (ESI) available. See DOI:
10.1039/c3gc41769g
‡These authors contributed equally to this work.
342 | Green Chem., 2014, 16, 342–350
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
monomer for the production of polylactate (PLA). The physical
properties and the rate of biodegradation of PLA are due to the
isomeric forms.13 Manufacturing polylactic acid needs large
amounts of the monomer, D-lactic acid and/or L-lactic acid.
Therefore, a greater demand for chirally and chemically pure
D-lactate is impending with the increase in the global PLA
market. However, the relatively high cost of raw materials and
the requirement of optically pure monomers of lactate pose
major limitations on the commercialization of polylactate. The
two basic approaches to overcome this are as follows: (i) to
produce D-lactate with higher optical purity using novel strains
and (ii) to use cheaper raw materials for monomer production.
The first approach is being addressed by overproducing
D-lactate using engineered strains and recently a yield of up to
125 g L−1 of D-lactate with a chiral purity of more than 99.9%
from glucose by engineered E. coli has been reported.13,14
Another strategy used is utilization of low-priced feedstock
such as crude glycerol generated from the biodiesel production
process. It was noted that wild-type E. coli exhibited a hetero-
fermentative behavior using glycerol as the substrate under
anaerobic and microaerobic conditions, which resulted in the
synthesis of significant amounts of ethanol, acetate, succinate,
and formate but only negligible production of D-lactate.15
Fig. 1 shows that in addition to ATP, NADH is also formed in
the glycerol to D-lactate pathway. Actually, this metabolic
pathway is an unbalanced-redox reaction under the conditions
of limited availability of external electron acceptors (e.g.,
anaerobic or microaerobic conditions). Therefore, it is impor-
tant to engineer E. coli strains producing D-lactate from
Fig. 1 Metabolic pathways involved in the aerobic and/or microaerobic
utilization of glycerol in E. coli and genetic modification strategies for
glycerol to eliminate byproduct formation and balance cell producing D-lactate employed in the present study. Relevant reactions
are represented by the names of the gene(s) coding for the enzymes.
aceA: isocitrate lyase; aceB: malate synthase; ackA, acetate kinase gene;
adhE, alcohol dehydrogenase gene; dhaKLM, dihydroxyacetone kinase;
dld, respiratory D-lactate dehydrogenase gene; frdABCD, fumarate
growth and product accumulation. Recently, refined glycerol
has been successfully used for the production of D-lactate.9
It has also been reported that metabolically engineered E. coli
is able to produce 50 g L−1 of L-lactate from 56 g L−1 of crude
reductase gene; fumABC, three isoenzymes of fumarases; gldA, glycerol
glycerol at a yield of 93% of the theoretical maximum.16
Recently, we isolated a wild-type E. coli B0013 strain based
on its ability to grow quickly, and engineered it for overprodu-
cing D-lactate from glucose via elimination of byproduct-
forming pathways as indicated in Fig. 1. The resulting strain
dehydrogenase; glpD, aerobic glycerol-3-phosphate dehydrogenase;
glpK, glycerol kinase; ldhA, fermentative D-lactate dehydrogenase gene;
mdh, malate dehydrogenase; pdh, pyruvate dehydrogenase complex;
pflB, pyruvate formate lyase gene; poxB, pyruvate oxidase gene; pps,
PEP synthase gene; pta, phosphotransacetylase gene; pykA/pykF, pyru-
vate kinase. Abbreviations: DHAP, dihydroxyacetone phosphate; PEP,
produced 125 g L−1 D-lactate with a high chiral purity from phosphoenolpyruvate. Red lines denote overexpression of ldhA, and red
cross bars denote the disruption of indicated genes.
glucose and reduced the yields of byproducts such as acetate,
succinate, formate, and ethanol significantly.13 The current
investigation is aimed at assessing the potential to bioconvert
crude glycerol into optically pure D-lactate using a rationally 070 (Δack-pta, Δpps, ΔpflB, Δdld, ΔpoxB, ΔadhE, ΔfrdA) was
engineered E. coli strain. Biochemical and kinetic consider- obtained from a previous study.13 E. coli B0013-070-pTHldhA
ations related to the growth of engineered strains and the (Δack-pta, Δpps, ΔpflB, Δdld, ΔpoxB, ΔadhE, ΔfrdA, pTH-ldhA)
efficiency in D-lactate production from glycerol were considered and B0013-070-pUCldhA (Δack-pta, Δpps, ΔpflB, Δdld, ΔpoxB,
and discussed. To date, there have been few reports on the ΔadhE, ΔfrdA, pUC-ldhA) were constructed in this study. Both
direct production of D-lactate using crude glycerol.
strains were stored at the CICIM-CU (Culture and Information
Center of Industrial Microorganism of China Universities at
Jiangnan University, http://CICIM-CU.jiangnan.edu.cn).
All pre-cultures were grown in LB medium. Modified
M9 medium containing glycerol was used as the fermentation
medium for lactate production. Modified M9 medium14 con-
Materials and methods
Strain, media and culture conditions
E. coli B0013 is a wild type strain isolated from a natural tained (per liter): 15.11 g Na2HPO4·12H2O, 3 g KH2PO4, 1 g
sample from which all derivatives were created. E. coli B0013- NH4Cl, 0.5 g NaCl to which a 0.1% MgSO4 solution (1 M) and
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 342–350 | 343

View Article Online
Paper
Green Chemistry
a 0.1% trace metal stock solution were added. The trace metal lactate production phase was initiated. During the cell growth
stock solution contained (per liter): 2.4 FeCl3·6H2O, phase, all flasks were incubated at 37 °C and 200 rpm except
g
0.3 g CoCl2·6H2O, 0.15 g CuCl2·2H2O, 0.3 g ZnCl2, 0.3 g where indicated. When the cell concentration reached 1.37 g L−1
Na2MoO4·2H2O, 0.075 g H3BO3, 0.495 g MnCl2·4H2O. Both LB DCW, 1.5 mL of 500 g L−1 glucose and 2 g CaCO3 were
medium and modified M9 medium were sterilized at 121 °C added to the broth, and the flasks were kept stationary at
for 20 min. The trace metal stock solution was filter-sterilized 37 °C to start the oxygen-limited phase. Fermentations ceased
through a 0.22 μm pore-size filter.
when glycerol was depleted. When the fed-batch process was
conducted, an appropriate volume of autoclaved crude glycerol
was added at the indicated time point.
Genetic manipulations
Standard molecular biology protocols were used for DNA
manipulation unless otherwise specifically mentioned.24 The
Bioreactor experiments
ldhA gene encoding the D-lactate dehydrogenase (LDH) was A modified two-phase fed-batch process was used that was
cloned from the genomic DNA of E. coli B0013 using PCR similar to that reported by Zhu et al.,27 except that the lactate
amplification and the primers 5′-CCCAAGCTTGGGCAGCCC- production phase generated by nitrogen sparging was not
GAGCGTCATCAG-3′ and 5′-CCCAAGCTTCTTGAATTCAAGCTT- applied. For each bioreactor experiment, cells were pre-
GCTGCCGGAAATCATCATTTTTT-3′. The resulting PCR frag- cultured as described above and transferred to 150 mL fresh
ment, which includes the promoter and coding frame of the modified M9 medium containing 5 g L−1 glycerol (in a 500 mL
ldhA gene, was inserted into the SmaI site of the high-copy flask) at a concentration of 0.1 g L−1 DCW. After incubation
vector pUC19 (TaKara, China) to generate the recombinant for 9 h, this seed culture was used to inoculate a 7 L bioreactor
plasmid pUC-ldhA. The ldhA fragment was obtained by digest- (New Brunswick Scientific Co., USA) containing 4 L modified
ing pUC-ldhA using the restriction enzyme HindIII and was M9 medium with 20 g L−1 glycerol at a concentration of
inserted into the corresponding site of the low-copy vector 0.06 g L−1 DCW. Growth was initiated by sparging air into the
pTH18kr25 to create pTH-ldhA. The recombinant plasmids bioreactor at 3 to 7 L min−1 and the dissolved oxygen concen-
were transformed into competent E. coli B0013-070 cells by tration was maintained above 30% saturation by agitation at
electroporation to generate the engineered strains E. coli 200 to 1000 rpm. The pH was maintained at 7 by automatically
B0013-070-pTHldhA and E. coli B0013-070-pUCldhA, respect- feeding concentrated NH4OH and 10% (vol/vol) H2SO4. The
ively. The resulting strains were used to investigate the effect microaerobic phase started when the cell concentration
of an increased copy number of ldhA on microbial growth and reached about 26 OD600 (∼10 g L−1 DCW). During the lactate
fermentation properties.
formation phase, the dissolved oxygen tension (DOT) was
controlled under 1% saturation by controlling the air flow at
3 L min−1 and agitation at 400 rpm and the pH was main-
Measurement of D-lactate dehydrogenase activity
Cells from microaerobic cultures were treated as previously tained at 7 using 25% (wt/vol) Ca(OH)2. Glycerol was added at
reported.14 Cells were harvested by centrifugation (2 min, an initial concentration of 30 g L−1 during the aerobic phase
10 000g), washed twice with 9 g L−1 NaCl, and stored as cell and diluted crude glycerol at a concentration of 600 g L−1 was
pellets at −20 °C. Cell pellets were resuspended in 0.1 M pot- automatically fed into the bioreactor to maintain the concen-
assium phosphate buffer to obtain ∼3.4 g dry cell weight tration above 10 g L−1 during the lactate formation phase. The
(DCW) L−1 and were permeabilized with chloroform. Crude two-phase fermentation stopped when glycerol was exhausted
cell extracts were prepared using the bacterial soluble total (unless otherwise stated), at which time the final product
protein preparation kit (GENMED Scientifics Inc., USA). concentrations were measured. During the process, the
The extracts were assayed for LDH activity using a kit to colori- temperatures of cell growth and D-lactate production were set
metrically determine the total bacterial LDH activity (GENMED at 34 °C and 40 °C (unless otherwise stated), respectively.
Scientifics Inc., USA). One unit of the overall LDH activity
Source of crude glycerol
was defined as the amount of enzyme required to trans-
form 1 nmol of pyruvate per minute. The protein concen- Crude glycerol was derived from alkali-catalyzed transesterifi-
tration in the crude extracts was determined by the Bradford cation of oil-bearing biomass which was used as the
method,26 and bovine serum albumin was used as the substrate for lactate production. The composition of crude
standard.
glycerol was assayed according to ref. 10 and 17. After
dilution with water and filtration, glycerol and methanol
concentrations were measured by HPLC according to the
Shake-flask experiments
Cells were pre-cultured in 100 mL LB medium in a 250 mL methods described below. The free fatty acid content was
flask. After 10 h of growth with shaking at 37 °C, cells were estimated by adjusting the pH of crude glycerol to 1.0 using
harvested by centrifugation and resuspended in modified HCl and centrifuged at 4000g for 20 min after which fatty
M9 medium. This suspension was used to inoculate 50 mL acids were carefully removed from the top layer and
fresh modified M9 medium containing glycerol at a concen- weighed.10 Ash content was determined by heating the
tration of 0.02 g L−1 DCW. A two-phase fermentation with an sample at 550 °C overnight in a furnace (Fenton, UK) before
aerobic cell growth phase, followed by an oxygen limited the ash weight was determined.17
344 | Green Chem., 2014, 16, 342–350
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Analytical methods
used to evaluate its feasibility to synthesize D-lactate from
crude glycerol.
Optical density during cell growth was measured at 600 nm.
A standard curve relating DCW to OD was constructed
(1 OD600 = 0.38 g L−1 DCW). Samples were pretreated with
H2SO4 (at 5% of the sample volume), to release organic acids
precipitated with CaCO3 or Ca(OH)2 during fermentation. After
centrifugation, the supernatant was stored at −20 °C for sub-
sequent HPLC analysis. To quantify the concentrations of gly-
cerol, lactate, acetate, formate, succinate, and ethanol,
supernatants were analyzed by ion-exclusion HPLC using a Shi-
madzu Prominence SIL 20 system (Shimadzu Scientific Instru-
ments, Inc., Columbia, MD) equipped with UV (210 nm) and
refractive index detectors, using an Agilent Hi-Plex H, 7.7 ×
300 mm, 8 µm (p/n PL1170-6830) with 0.01 M H2SO4 as an
eluent (0.6 mL min−1; 50 °C). A RI detector and a UV detector
(210 nm) were used for the detection of glycerol and organic
acids, respectively. Lactate enantiomeric excess (isomeric purity)
was measured by HPLC using a chiral column (CLC-L; Advanced
Separation Technologies Inc., Whippany, NJ, USA), at room
temperature, equilibrated with 1 mL min−1 of 5 mM CuSO4 as
the mobile phase and detected at 254 nm with a UV detector.
It has previously been reported that various kinds of impu-
rities in crude glycerol samples have a negative impact on bac-
terial growth and the bioconversion.11 In order to assess the
effect of crude glycerol on the growth performance of E. coli
B0013 and B0013-070 strains, a series of minimal media con-
taining various concentrations of crude glycerol was used to
culture E. coli in shake flasks, with pure glycerol being the
control carbon source for cell growth. There was no discernible
difference in the growth of wild type and engineered strains
when cultured at concentrations below 40 g L−1 of either crude
(40 g L−1 of crude glycerol indicates the final glycerol concen-
tration in the medium containing crude glycerol which has
63.5% glycerol, the same as below) or pure glycerol (Fig. S1†).
However, when the crude glycerol concentration increased to
60 g L−1 and 80 g L−1, cell growth was inhibited significantly
and biomass content decreased to 3.2 g L−1 and 1.3 g L−1
,
respectively (Fig. S1a†). Similarly, a high concentration of pure
glycerol exhibited little suppressive impact on both the growth
and the biomass content of both strains (Fig. S1b†). Moreover,
the glycerol consumption rate declined significantly when cul-
tured in high concentrations of glycerol (data not shown).
Fig. S1† elucidated that both wild type and engineered strains
can efficiently utilize glycerol for cell growth in the crude
samples; however, the biomass productivity of strains was
Results and discussion
Analysis of the composition of crude glycerol
Previous studies have indicated that the chemical composition inferior to that achieved with pure glycerol under higher sub-
of crude glycerol samples varied extensively based on the com- strate concentration. Presumably, the high quantities of salt in
position of the oil feedstock, the type of catalyst used to crude glycerol had a negative impact on the bacterial growth.
produce biodiesel, the transesterification efficiency, the recov- Generally, noticeable salt concentrations can inhibit bacteria
ery efficiency of the biodiesel, and the biodiesel production growth, whereas high salt concentrations can impose cell con-
processes.2,10 Thus, in this study the composition of crude traction and eventually osmotic plasmolysis, causing the
glycerol derived from oil-bearing biomass was characterized. microorganisms to shrink and die.18,19
After being sterilized by autoclaving, the treated crude glycerol
sample contained 63.5% (w/w) glycerol, 24.2% water, 1.9%
fatty acids and 2.8% ash (Table S1†). The results show a low
Evaluation of the engineered strain B0013-070 for the
capability to produce D-lactate from crude glycerol
glycerol concentration and a high water content, which Higher cell density cultures are directly related to higher volu-
may favor cell growth and downstream treatment. Methanol metric productivity, which is a major objective of any E. coli-
was undetectable after autoclaving, as has been reported based process.20 On the other hand, efficient D-lactate syn-
previously.17
thesis needs microaerobic or anaerobic conditions during the
cultivation of E. coli. Thus, we firstly addressed this issue by
manipulation of the initial glycerol concentration and fermen-
tation conditions in a shake flask. Considering that higher
Effect of crude glycerol on the growth of engineered
E. coli strains
Previously we constructed an overproducing D-lactate engi- concentrations of glycerol suppressed cell growth and sub-
neered strain of E. coli B0013-070 by deletion of byproduct- strate consumption, batch fermentations and fed-batch pro-
forming pathways including genes coding for acetate kinase cesses with different strains were carried out to investigate the
(ackA), phosphotransacetylase (pta), phosphoenolpyruvate efficiency of lactate production from glycerol. Fig. 2 and
synthase (pps), pyruvate formate lyase (pflB), FAD-binding Table 1 present a comparison of the fermentation perform-
D-lactate dehydrogenase (dld), pyruvate oxidase (poxB), alcohol ances of wild type and engineered strains in defined media
dehydrogenase (adhE) and fumarate reductase ( frdA), as indi- containing crude or pure glycerol, respectively. In a batch fer-
cated in Fig. 1. When glucose was used as the substrate, the mentation containing 20 g L−1 of crude glycerol, the wild type
resulting strain produced 125 g L−1 D-lactate with an increased strain B0013 consumed glycerol producing D-lactate, acetic
oxygen limiting lactate productivity of 0.61 g g−1 h−1 (2.1-fold acid, succinic acid and ethanol as the main primary products.
greater than E. coli B0013) and significantly reduced yields of Comparatively, the engineered strain B0013-070 converted crude
byproducts such as acetate, succinate, formate, and ethanol.13 glycerol into significantly more D-lactate and all byproducts
Therefore, it stands to reason that this engineered strain be were observed at low concentrations (Fig. 2 and Table 1). Using
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 342–350 | 345

View Article Online
Paper
Green Chemistry
Effect of overexpressing D-lactate dehydrogenase on the
strain’s fermentation performance
D-Lactate dehydrogenase (LDH, encoded by ldhA gene) cata-
lyzes pyruvate to D-lactate with an irreversible reaction. It is
likely that overexpressing ldhA improved lactate synthesis in
E. coli.9 However, high activity of D-lactate dehydrogenase leads
to lactate accumulation and inhibits cell growth during the
earlier phase of cell growth.21,22 Therefore, we constructed
various recombinant plasmids containing ldhA gene of
pTH18kr-ldhA (a low-copy vector) and pUC-ldhA (a high-copy
vector) (Fig. S2†), which were transformed into E. coli B0013-
070 to generate strains E. coli B0013-070-pTHldhA and E. coli
B0013-070-pUCldhA, respectively. The resulting strains were
evaluated for the capability to produce D-lactate from glycerol.
Fig. 3 showed the impact of amplifying the LDH pathway on
cell growth and product synthesis. Results showed that over-
expressing the ldhA gene using a low-copy vector led to an
increase of the concentration and yield of D-lactate, in which
14.5 g L−1 of lactate with a yield of 0.72 g g−1 crude glycerol
was obtained (Fig. 3b). However, the growth properties
and final cell concentration were similar to that of E. coli
B0013-070 (Fig. 3a). Increasing the expression of the ldhA gene
in E. coli B0013-070-pUCldhA significantly decreased the cell
growth rate with a lower final cell mass (Fig. 3a). Although the
final concentration of D-lactate increased (14.7 g L−1 of lactate
was produced after a 50 h fermentation period, data not
shown), the overall productivity was impeded significantly due
to prolonging of the fermentation period (Fig. 3b) and slower
glycerol consumption was observed (data not shown). Further-
more, we found that E. coli B0013-070-pUCldhA produced a
higher amount of lactate within the earlier phase of growth,
which would not only compete for more substrate for cell
growth but also depress the growth performance because of
the lower pH of the medium. We found that both engineered
strains harboring recombinant plasmids produced small
amounts of acetate and succinic acid as byproducts in the
culture medium (data not shown), similar to E. coli B0013-070
(Table 1). Moreover, LDH activity was assayed and results
showed that a 6.5 and 15.1-fold increase in the activity of LDH
was obtained from E. coli B0013-070-pTHldhA and E. coli
B0013-070-pUCldhA, respectively (Fig. S3†). The observed
increases in D-lactate yield and concentration were arguably
Fig. 2 D-Lactate production, cell growth, and glycerol utilization in
36 hour shake flask cultures of wild-type E. coli B0013 and the engi-
neered strain E. coli B0013-070 in crude glycerol-medium and/or pure
glycerol-medium. Error bars represent standard deviations for triplicate
measurements.
20.1 g L−1 of crude glycerol, B0013-070 produced 13.6 g L−1
D-lactate (0.38 g L−1
h
−1) at a yield of 0.67 g D-lactate per g
glycerol (Fig. 2 and Table 1). However, a higher concentration of
D-lactate (14.4 g L−1) with a yield of 0.76 g D-lactate per g glycerol
was obtained when pure glycerol was used as the substrate
(Fig. 2). Furthermore, a slight increase in cell mass and glycerol
consumption was also observed in batch fermentations using
pure glycerol as the substrate compared to crude glycerol
(Fig. 2), which indicated that some unknown compounds in
crude glycerol had a trivial effect on the fermentation process.
Given these results, we further examined the production of
D-lactate by the B0013-070 strain in the presence of a higher
concentration of glycerol. The strain B0013-070 produced about
27 g L−1 and 41.2 g L−1 of D-lactate from 39 and 59 g L−1 of
crude glycerol in about 60 h and 80 h, respectively (Table 1).
However, it was noted that the fed-batch process had an impact
on cell activity and viability and decreased the productivity of
D-lactate because of a long fermentation period in a shake flask.
Results indicated that the engineered strain E. coli B0013-070
could convert glycerol into D-lactate with a low concentration of
byproducts; however, the conversion efficiency was lower than
that when glucose was used as the substrate.13 It may need
further engineering to improve the strain’s performance.
Table 1 Glycerol consumption, product synthesis, and yield of cell mass and product during the two phases of aerobic–microaerobic fermentation
by wild-type and engineered strains in the defined medium containing crude glycerola
Glycerol consumed
(g L−1
DCW
Lactate
(g L−1
Succinic acid
(g L−1
Acetic acid
(g L−1
Ethanol
(g L−1
Lactate yield
Strain
)
(g L−1
)
)
)
)
)
(g g−1 glycerol)
Wild-type B0013
B0013-070
19.6 1.0
20.1 1.2
39.6 0.7
58.6 2.7
3.8 0.3
3.5 0.1
3.2 0.1
2.9 0.2
5.3 0.4
13.6 1.3
27.4 0.9
41.2 0.9
2.7 0.1
0.6 0.1
0.4 0.0
0.3 0.0
3.1 0.3
1.1 0.2
0.6 0.2
1.1 0.1
1.2 0.1
0.5 0.2
0.6 0.1
0.5 0.0
0.27 0.06
0.68 0.14
0.69 0.1
0.70 0.14
B0013-070b
B0013-070c
a Data represent the averages of three samples (mean standard deviations) taken from 36 hour shake flask cultures grown on minimal medium
supplemented with 20 g L−1 of crude glycerol, unless otherwise specified. b Cultures to which 40 g L−1 of crude glycerol was used and samples
were taken at 60 hours (99% glycerol was consumed). c Culture to which 60 g L−1 of crude glycerol (40 g L−1 initially present and 20 g L−1 added
at 48 hours) was used and samples were taken at 80 hours (∼59 g L−1 of glycerol was consumed).
346 | Green Chem., 2014, 16, 342–350
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
in a decrease in production efficiency.14 Bunch et al.23 found that
overexpressing fermentative D-lactate dehydrogenase (LDH,
encoded by the ldhA gene) to overproduce D-lactate in E. coli
resulted in very poor growth because converting the larger pyru-
vate pool to lactate by overexpression of LDH resulted in a short-
age of 3-carbon metabolic intermediates. Previously, we have
developed a two-phase fermentation process to attain a high cell
density and intensified production of D-lactate when glucose was
used as the substrate.13 However, a certain amount of lactate was
still produced during the aerobic growth phase and suppressed
the high cell density. A similar pattern of cell growth and D-lactate
accumulation in the aerobic phase was observed when using
crude glycerol as the substrate (data not shown). Therefore the
effect of temperature in relation to cell growth and D-lactate pro-
duction of the strain B0013-070-pTHldhA was investigated.
As indicated in Fig. 4, the engineered E. coli strain had
different growth rates and accumulated different amounts of
lactate during different fermentation temperatures in the cell
growth phase. At temperatures of 34 °C and 37 °C a similar
growth pattern with a higher growth rate and final cell mass
was observed (Fig. 4a). However, at higher temperatures (42 °C
Fig. 3 Evaluation of overexpression of D-lactate dehydrogenase on cell
growth and D-lactate synthesis in engineered strains. (a) Growth curves
of strains B0013-070, B0013-070-pUCldhA and B0013-070-pTHldhA.
(b) Production of D-lactate over time by strains B0013-070, B0013-070-
pUCldhA and B0013-070-pTHldhA. Three experiments were carried out
and the average values are presented.
caused by an increase in LDH activity. Results however indicated
that a moderate increase in LDH activity benefits D-lactate pro-
duction and overexpressing the ldhA gene with a high-copy
vector would lead to a negative impact on the fermentation
performance of the engineered strain.
In addition, previous studies have indicated that overexpres-
sion of the enzymes involved in the conversion of glycerol to
glycolytic intermediates of DHAP (GlpK-GlpD and GldA-DHAK
pathways) can improve the activity of two primary glycerol
transport routes and cause a modest increase in D-lactate con-
centration and yield.9 It can be observed that it is possible to
further improve the lactate fermentation performance by
engineering the glycerol assimilation pathway in E. coli.
Effect of temperature on growth and D-lactate accumulation by
engineered strains
Fig. 4 Cell growth pattern (a) and D-lactate accumulation (b) of the
strain B0013-070-pTHldhA during the aerobic growth phase on shake
flask incubation with different temperatures. Values are reported as
averages of triplicate assays (a) and error bars represent standard devi-
It has been reported that the formation of the desired product
during the cell growth phase competes for substrates with
biomass formation or inhibits cell growth directly, which results ations for triplicate measurements (b).
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 342–350 | 347

View Article Online
Paper
Green Chemistry
and/or 45 °C) there was an obvious inhibition of cell growth combined with a temperature-shifting strategy can improve
(Fig. 4a). On the other hand, the production of D-lactate exhi- D-lactate production significantly under bioreactor conditions.
bited a different trend as compared to the growth pattern
(Fig. 4b). A yield of 0.42 g g−1 glycerol was reached at a temp-
Enhanced production of D-lactate from crude glycerol in a
erature of 40 °C, which was higher when compared to other
temperatures. Increases in temperature decreased the yield,
and 0.16 g g−1 glycerol and 0.11 g g−1 glycerol of yields were
obtained at 42 °C and 45 °C, respectively. The effect of temp-
erature on D-lactate production in the micro-aerobic phase was
evaluated and the results are shown in Fig. 5. After the cultures
reached a high cell density in the aerobic phase at 34 °C
following 12 h of incubation, the temperature for D-lactate
production was optimized. Fig. 5(b) showed that temperatures
of 40 °C and 42 °C led to a higher D-lactate concentration, and
22.5 g L−1 and 22.1 g L−1 D-lactate from 40 g L−1 glycerol were
obtained under shake flask batch fermentations, respectively;
however, the cell mass decreased slightly at temperatures
above 42 °C compared to temperatures below 40 °C (Fig. 5a).
Higher temperature impaired both the cell mass and the
accumulation of the desired product (Fig. 5a and b). Therefore,
there is reason to speculate that the two-phase process
bioreactor
The above results indicated that the engineered strain B0013-
070-pTHldhA performed well when crude glycerol was used
as a carbon source, better than the wild type and B0013-070
strains. Experimentally, it is difficult to accurately control the
fermentation parameters such as dissolved oxygen and pH in
a shake flask. Therefore, a scale-up fermentation using the
B0013-070-pTHldhA strain with a 7 L bioreactor was employed
to better assess the potential for D-lactate production of this
engineered strain. The two-phase fermentation combined with
a temperature shifting process was carried out under con-
trolled conditions. Cell mass reached a maximum at ∼10 h but
lactate accumulation was very low which indicated that oxygen
supply in the aerobic phase was sufficient in the bioreactor
compared to the shake-flask fermentation (Fig. 6). Further-
more, the optimal growth temperature also significantly
reduced the lag period of the engineered strain (Fig. 6). When
the lactate formation phase was shifted by controlling the
agitation and aeration rates, at which point the temperature
was raised from 34 °C to 40 °C, lactate production increased at
a high specific rate and the final concentration reached
100.3 g L−1 with an optical purity of 99.97% (Fig. 6). After a
36 h fermentation period, 531.5 g of the total crude glycerol
was consumed with an overall lactate productivity of 2.78 g L−1 h−1
and an overall yield of 75.4 g per 100 g glycerol (Table 2),
which indicated a significant increase in lactate productivity
and concentration in the bioreactor experiment. When a
Fig. 6 Representative kinetics of D-lactate production by the engi-
neered strain B0013-070-pTHldhA in a bioreactor with the aerobic and
microaerobic fermentation processes combined with the temperature
shifting strategy. The arrow indicates the time point at which the culture
was switched from aerobic cultivation at 34 °C to the microaerobic pro-
duction phase at 40 °C. An initial concentration of 20 g L−1 of crude gly-
Fig. 5 Cell mass accumulation (a) and D-lactate production (b) of the
B0013-070-pTHldhA strain during the microaerobic fermentation phase
under different temperatures. Error bars represent standard deviations for
triplicate measurements.
cerol was used and
a high concentration of the substrate was
automatically fed into the bioreactor to maintain the concentration
above 10 g L−1 during the microaerobic phase. Coefficients of variation
(i.e., standard deviations/average × 100) were below 5% in all cases.
348 | Green Chem., 2014, 16, 342–350
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Table 2 Comparison of fermentation data of B0013-070-pTHldhA (mean range of duplicate experiments) in bioreactor experiments from pure
glycerol and crude glycerol after a period of 36 h
Yield of the overall fermentation process
Lactate volumetric (g per 100 g glycerol)
Lactate
Fermentation
processa
Glycerol
concentration productivityc
consumedb (g) (g L−1
)
(g L−1 h−1
)
Cell mass Lactate
Acetate
Succinate Pyruvate
34 °C/40 °C and crude glycerol 531.5 4.1
100.3 4.1
93.3 5.7
111.7 9.6
106.9 6.0
2.78 0.3
2.62 0.1
3.10 0.2
2.97 0.2
8.4 0.5
8.9 0.4
9.0 0.3
9.3 0.6
75.4 4.4 0.5 0.1 0.3 0.1
69.4 4.1 0.6 0.2 0.5 0.2
80.4 5.5 0.4 0.0 0.1 0.0
75.3 4.8 0.5 0.1 0.4 0.1
0.2 0.0
0.0 0.0
0.1 0.0
0.2 0.0
37 °C and crude glycerol
34 °C/40 °C and pure glycerol
37 °C and pure glycerol
538.2 15.7
555.6 16.5
568.1 15.2
a 37 °C indicated that the overall fermentation temperature was constantly set at 37 °C, 34 °C/40 °C indicated that the aerobic incubation
temperature was set at 34 °C and the microaerobic phase was set at 40 °C, and crude glycerol or pure glycerol was used as a carbon source.
b Indicated total glycerol consumed during the overall fermentation process. c Average volumetric productivity for lactate during the overall
fermentation process.
constant temperature of 37 °C was maintained during the crude glycerol as a sole carbon source. Impurities in the
overall process, a lower yield of 69.4 g per 100 g glycerol and a sample had a minor impact on the fermentation performance.
lower productivity of 2.62 g−1 L−1 h−1 were observed for D-lactate A combination of a two-phase fermentation process and a
production compared to the temperature-shifting process temperature-shifting strategy using crude glycerol improved
(Table 2), with the final concentration of D-lactate reaching only D-lactate production significantly, achieving 100.3
g
L−1
93.3 g L−1. Pure glycerol was also used to compare the potential D-lactate with an overall lactate productivity of 2.78 g L−1 h−1 and
and efficiency of lactate production from biodiesel-based crude a yield of 0.75 g g−1 glycerol (0.77 mol mol−1) from the crude
glycerol. Using similar conditions to the temperature-shifting sample. In summary, our study shows that the crude glycerol
strategy, the engineered strain showed an increase in the con- could be directly converted to D-lactate with high efficiency.
centration, yield and productivity of D-lactate, which reached
111.7 g L−1, 3.10 g L−1 h−1 and 80.4 g per 100 g glycerol, respect-
ively. It can be deduced that crude glycerol is suitable to be
used as a cheap substrate for D-lactate production using a meta-
Acknowledgements
bolically engineered strain of E. coli; however, unknown impuri-
ties in the sample did have minor effects on the efficiency of
bioconversion to D-lactate.
This work was financially supported by the National Natural
Science Foundation of China (no. 21006039), the Jiangsu pro-
vince science and technology support program (BE2012618),
the 111 Project (no. 111-2-06) and the Jiangsu province “Colla-
borative Innovation Center for Advanced Industrial Fermenta-
tion” industry development program.
It should be noted that the primary costs associated with
lactate production from crude glycerol depends on not only
D-lactate yield and productivity but also the downstream
D-lactate recovery and purification processes. Conventional
purification processes such as crystallizing, filtering, dis-
solving, and subsequently acidifying were performed for the
isolation of lactic acid from the fermentation broth. However,
huge quantities of calcium sulfate cake were generated as
waste.28 Recently, some novel purification processes such as
adsorption, electrodialysis, reactive extraction, and reactive dis-
tillation were also developed.29 Especially, the electrodialysis
process is advantageous in its capability for separation, purifi-
cation and concentration of salts, and in the conversion of these
salts into acids and bases without generating waste effluents
containing high concentrations of salts.30 This process might be
suitable for broth using crude glycerol as the substrate, which
contains high quantities of salt and some unknown impurities.
In the future, developing a robust biocatalyst which can directly
convert crude glycerol into lactate without prior autoclave could
also decrease the price of lactic acid.
References
1 P. J. Westfall and T. S. Gardner, Curr. Opin. Biotechnol.,
2011, 22, 344–350.
2 F. Yang, M. A. Hanna and R. Sun, Biotechnol. Biofuels, 2012,
5, 13.
3 J. A. Posada, L. E. Rincon and C. A. Cardona, Bioresour.
Technol., 2012, 111, 282–293.
4 M. L. Barbelli, G. F. Santori and N. N. Nichio, Bioresour.
Technol., 2012, 111, 500–503.
5 A. A. Koutinas, R. H. Wang and C. Webb, Biofuels Bioprod.
Bior., 2007, 1, 24–38.
6 X. Tang, Y. Tan, H. Zhu, K. Zhao and W. Shen, Appl.
Environ. Microbiol., 2009, 75, 1628–1634.
7 B. Andreessen, A. B. Lange, H. Robenek and A. Steinbuchel,
Appl. Environ. Microbiol., 2010, 76, 622–626.
8 M. D. Blankschien, J. M. Clomburg and R. Gonzalez,
Metab. Eng., 2010, 12, 409–419.
Conclusions
It can be concluded that a metabolically engineered strain of
E. coli could be efficiently used to produce D-lactate using
9 S. Mazumdar, J. M. Clomburg and R. Gonzalez, Appl.
Environ. Microbiol., 2010, 76, 4327–4336.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 342–350 | 349

View Article Online
Paper
Green Chemistry
10 Y. Liu, C. M. Koh and L. Ji, Bioresour. Technol., 2011, 102, 21 L. Zhou, K. M. Tian, D. D. Niu, W. Shen, G. Y. Shi, S. Singh
3927–3933. and Z. X. Wang, Biotechnol. Lett., 2012, 34, 1123–1130.
11 A. Chatzifragkou, D. Dietz, M. Komaitis, A. P. Zeng and 22 L. Zhou, W. Shen, D. D. Niu, K. M. Tian, B. A. Prior,
S. Papanikolaou, Biotechnol. Bioeng., 2010, 107, 76–84.
12 C. Gao, C. Ma and P. Xu, Biotechnol. Adv., 2011, 29,
930–939.
G. Y. Shi, S. Singh and Z. X. Wang, J. Ind. Microbiol. Bio-
technol., 2012, 39, 1209–1217.
23 P. K. Bunch, F. Mat-Jan, N. Lee and D. P. Clark, Micro-
biology, 1997, 143(Pt 1), 187–195.
13 L. Zhou, Z. R. Zuo, X. Z. Chen, D. D. Niu, K. M. Tian,
B. A. Prior, W. Shen, G. Y. Shi, S. Singh and Z. X. Wang, 24 J. Sambrook, E. F. Fritsch and T. Maniatis, Molecular
Curr. Microbiol., 2011, 62, 981–989.
Cloning.
A Laboratory Manual, Cold Spring Harbor
14 L. Zhou, D. D. Niu, K. M. Tian, X. Z. Chen, B. A. Prior,
Laboratory Press, Cold Spring Harbor, NY, 1989.
W. Shen, G. Y. Shi, S. Singh and Z. X. Wang, Metab. Eng., 25 T. Hashimoto-Gotoh, M. Yamaguchi, K. Yasojima,
2012, 14, 560–568.
15 A. Murarka, Y. Dharmadi, S. S. Yazdani and R. Gonzalez,
Appl. Environ. Microbiol., 2008, 74, 1124–1135.
A. Tsujimura, Y. Wakabayashi and Y. Watanabe, Gene,
2000, 241, 185–191.
26 M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
16 S. Mazumdar, M. D. Blankschien, J. M. Clomburg and 27 Y. Zhu, M. Eiteman, K. DeWitt and E. Altman, Appl.
R. Gonzalez, Microb. Cell. Fact., 2013, 12, 7. Environ. Microbiol., 2007, 73, 456–464.
17 D. J. Pyle, R. A. Garcia and Z. Y. Wen, J. Agric. Food Chem., 28 H. G. Joglekar, I. Rahman, S. Babu, B. D. Kulkarni and
2008, 56, 3933–3939.
18 A. L. Koch, J. Bacteriol., 1984, 159, 919–924.
19 R. E. Marquis, J. Bacteriol., 1968, 95, 775–781.
A. Joshi, Sep. Purif. Technol., 2006, 52, 1–17.
29 J. Vijayakumar, R. Aravindan and T. Viruthagiri, Chem.
Biochem. Eng. Q., 2008, 22, 245–264.
20 J. Shiloach and R. Fass, Biotechnol. Adv., 2005, 23, 30 M. A. Abdel-Rahman, Y. Tashiro and K. Sonomoto,
345–357.
Biotechnol. Adv., 2013, 31, 877–902.
350 | Green Chem., 2014, 16, 342–350
This journal is © The Royal Society of Chemistry 2014