Extending carbon chain length of 1-butanol pathway for 1-hexanol synthesis from glucose by engineered Escherichia coli

Article abstract of DOI:10.1021/ja203814d

An Escherichia coli strain was engineered to synthesize 1-hexanol from glucose by extending the coenzyme A (CoA)-dependent 1-butanol synthesis reaction sequence catalyzed by exogenous enzymes. The C4-acyl-CoA intermediates were first synthesized via acetyl-CoA acetyltransferase (AtoB), 3-hydroxybutyryl-CoA dehydrogenase (Hbd), crotonase (Crt), and trans-enoyl-CoA reductase (Ter) from various organisms. The butyryl-CoA synthesized was further extended to hexanoyl-CoA via β-ketothiolase (BktB), Hbd, Crt, and Ter. Finally, hexanoyl-CoA was reduced to yield 1-hexanol by aldehyde/alcohol dehydrogenase (AdhE2). Enzyme activities for the C6 intermediates were confirmed by assays using HPLC and GC. 1-Hexanol was secreted to the fermentation medium under anaerobic conditions. Furthermore, co-expressing formate dehydrogenase (Fdh) from Candida boidinii increased the 1-hexanol titer. This demonstration of 1-hexanol production by extending the 1-butanol pathway provides the possibility to produce other medium chain length alcohols using the same strategy.

Full text of DOI:10.1021/ja203814d

COMMUNICATION
pubs.acs.org/JACS
Extending Carbon Chain Length of 1-Butanol Pathway for 1-Hexanol
Synthesis from Glucose by Engineered Escherichia coli
Yasumasa Dekishima,†,^ Ethan I. Lan,†,‡ Claire R. Shen,† Kwang Myung Cho,†,§ and James C. Liao*,†,‡,§
†Department of Chemical and Biomolecular Engineering, ‡Biomedical Engineering Inter-departmental Program, and
§Institute for Genomics and Proteomics, University of California, Los Angeles, California 90095, United States
^Mitsubishi Chemical Group Science and Technology Research Center, Inc., Yokohama 227-8502, Japan
S Supporting Information
b
Previously, Zhang et al.6 reported 1-hexanol production in
ABSTRACT: An Escherichia coli strain was engineered to
synthesize 1-hexanol from glucose by extending the coen-
zyme A (CoA)-dependent 1-butanol synthesis reaction
sequence catalyzed by exogenous enzymes. The C4-acyl-
CoA intermediates were first synthesized via acetyl-CoA
acetyltransferase (AtoB), 3-hydroxybutyryl-CoA dehydro-
genase (Hbd), crotonase (Crt), and trans-enoyl-CoA re-
ductase (Ter) from various organisms. The butyryl-CoA
synthesized was further extended to hexanoyl-CoA via β-
ketothiolase (BktB), Hbd, Crt, and Ter. Finally, hexanoyl-
CoA was reduced to yield 1-hexanol by aldehyde/alcohol
dehydrogenase (AdhE2). Enzyme activities for the C6
intermediates were confirmed by assays using HPLC and
GC. 1-Hexanol was secreted to the fermentation medium
under anaerobic conditions. Furthermore, co-expressing
formate dehydrogenase (Fdh) from Candida boidinii in-
creased the 1-hexanol titer. This demonstration of 1-hexanol
production by extending the 1-butanol pathway provides
the possibility to produce other medium chain length
alcohols using the same strategy.
E. coli from glucose by extending the 2-ketoacid pathway devel-
oped by Atsumi et al.7 The 2-ketoacid pathway for 1-hexanol
synthesis is based on a mutated leucine biosynthesis pathway,
which employs one acetyl-CoA to extend the chain length by one
carbon unit. On the other hand, the CoA-dependent pathway
elongates the carbon chain by two carbon units at the expense of
one acetyl-CoA. Thus, the latter has a higher carbon yield.
Currently, there is no report of 1-hexanol production by the
CoA-dependent 1-butanol pathway in engineered microorgan-
isms. In this work, we extended the carbon chain of this pathway
to produce 1-hexanol as the first trial of medium chain length
alcohol production using the CoA-dependent pathway described
in Scheme 1. We confirmed enzyme activities and demonstrated
direct 1-hexanol synthesis from glucose in engineered E. coli.
Since Shen et al. demonstrated efficient synthesis of 1-butanol
in E. coli,3 the first step in constructing the 1-hexanol pathway was
to express the potential genes and detect the enzymatic activities
for extending the carbon chain length. β-Ketothiolase (BktB)
from Ralstonia eutropha and Clostridium acetobutylicum Hbd and
Crt were expressed to catalyze the first three steps after butyryl-
CoA. Each enzyme was expressed as the N-terminal fusion
protein with His6 and purified by Ni-affinity column chromatography.
Owing to the unavailability of 3-ketohexanoyl-CoA and
3-hydroxyhexanoyl-CoA, we tested the enzymatic activities of
these three reactions in the reverse direction using hexenoyl-CoA
as the substrate. Hexenoyl-CoA was chemically synthesized8 and
partially purified by preparative HPLC. The substrate was then
incubated with Crt, Hbd, and BktB with appropriate cofactors in
a stepwise fashion, and the resulting products were analyzed by
HPLC (Figure 1). Hexenoyl-CoA was detected at 10.3 min. In
the presence of Crt, hexenoyl-CoA was consumed, and a new
peak at 8.7 min appeared (Figure 1B), which was presumably
3-hydroxyhexanoyl-CoA. In the next step, the addition of Hbd
and NAD+ showed a very similar pattern on the chromatogram as
the previous reaction (Figure 1C). After the addition of BktB and
CoA, acetyl-CoA and butyryl-CoA were detected as two peaks
resolving at 4.2 and 8.9 min, respectively (Figure 1D).
n recent years, synthesis of higher chain volatile alcohols by
engineered microorganisms has received increasing interest
I
because of energy and sustainability concerns. Naturally pro-
duced by Clostridium species, 1-butanol has attracted the most
attention, and many groups have reported its production in
recombinant organisms such as Escherichia coli1,2 by expressing
the genes found in Clostridium and other organisms to catalyze
the 1-butanol synthesis reactions. In this pathway, 1-butanol
synthesis starts from acetyl-CoA and is catalyzed by thiolase/
acetyl-CoA acetyltransferase (Thl/AtoB), 3-hydroxybutyryl-
CoA dehydrogenase (Hbd), crotonase (Crt), butyryl-CoA dehy-
drogenase (Bcd), and alcohol/aldehyde dehydrogenase (AdhE2).
Shen et al.3 recently achieved high-titer production of 1-butanol
(15À30 g/L) using this series of reactions by replacing Bcd with
trans-enoyl-CoA reductase (Ter), which directly utilizes NADH
as the reducing equivalent, and by artificially building up NADH
and acetyl-CoA as driving forces. On this basis, here we extend
this 1-butanol production pathway to longer chain alcohols such
as 1-hexanol. Native organisms, such as Clostridium kluyveri, have
been reported to ferment ethanol and acetate not only to butyric
acid, hexanoic acid, and H2 but also to 1-butanol and 1-hexanol.4,5
However, the enzyme(s) involved in biosynthesis for 1-hexanol
production is not defined.
The appearance of the expected products (acetyl-CoA and
butyryl-CoA) suggested that all enzymes were active toward C6
substrates. These results were consistent with the previous
findings that BktB from R. eutropha has dissociative activity
toward 3-ketohexanoyl-CoA9 and Crt from C. acetobutylicum
Received: April 25, 2011
Published: June 27, 2011
r
2011 American Chemical Society
11399
dx.doi.org/10.1021/ja203814d J. Am. Chem. Soc. 2011, 133, 11399–11401
|

Journal of the American Chemical Society
COMMUNICATION
Scheme 1. Artificial Pathway for 1-Hexanol Biosynthesis in
Escherichia colia
Figure 1. (Left) HPLC analysis of Crt, Hbd, and BktB reaction with
hexenoyl-CoA. Hexenoyl-CoA was incubated with (A) no enzymes nor
cofactors as a control; (B) Crt; (C) Crt, Hbd, and NAD; or (D) Crt,
Hbd, BktB, NAD+, and CoA. (Right) HPLC analysis of EgTer and
TdTer reaction with hexenoyl-CoA. Hexenoyl-CoA was incubated with
(E) no enzyme as control; (F) EgTer and NADH; or (G) TdTer
and NADH.
Treponema denticola (TdTer), which was used in 1-butanol
productions,3 was reported to not possess reductive activity
against C6 substrates.12 To test these enzymes, we incubated
hexenoyl-CoA with purified EgTer or TdTer and analyzed the
products. Interestingly, HPLC data showed that both EgTer and
TdTer were able to reduce hexenoyl-CoA to hexanoyl-CoA
using NADH as the reducing cofactor, suggesting the broad
substrate specificity of these enzymes (Figure 1EÀG).
The next step is catalyzed by an aldehyde/alcohol dehydro-
genase. AdhE2 from C. acetobutylicum is known to reduce
butyryl-CoA to butyraldehyde and to 1-butanol,13 but its activity
toward longer chain length acyl-CoA substrates was not re-
ported. We tested this activity by a spectrophotometric assay.
As shown in Table 1, AdhE2 has a higher activity toward butyryl-
CoA compared to acetyl-CoA (consistent with previous data1)
and also has a significant activity for hexanoyl-CoA (15 ( 9.5
mU/mg). Interestingly, activity toward octanoyl-CoA (36 ( 6.9
mU/mg) is higher than toward hexanoyl-CoA. To confirm that
AdhE2 converts hexanoyl-CoA to 1-hexanol (or hexaldehyde)
more directly, hexanoyl-CoA was incubated with cell lysate
containing AdhE2 and NADH, and the reaction solution was
analyzed by GC. As shown in Figure S2, 1-hexanol was detected
only in the reaction mixture containing AdhE2.
Having characterized BktB, Hbd, Crt, Ter (TdTer and
EgTer), and AdhE2 activities toward C6 substrates, we then
tested the production of 1-hexanol by fermentation from glucose.
The host strain used was JCL166 or JCL299 (Table S1), both of
which have adhE, ldhA, and frdAB genes deleted to increase the
NADH availability to drive the reaction. In addition, JCL299 has
an additional pta gene deletion, which increases the acetyl-CoA
availability.3 The genes for 1-hexanol synthesis were expressed
from various plasmids: pEL11 (expressing atoB, adhE2, crt, and
hbd), pIM8 (expressing Tdter), and pEL102 (expressing Tdter,
bktB, and Egter) (Tables S1 and S2). The strains JCL166/
pEL11/pIM8, JCL166/pEL11/pEL102, and JCL299/pEL11/
pEL102 were cultivated in TB + 2% glucose media (5 mL) under
anaerobic condition for 68 h. Upon centrifugation of broths to
exclude the cells, supernatants were analyzed by GC.
a Abbreviations: AtoB, acetyl-CoA acetyltransferase; Hbd, 3-hydroxy-
butyryl-CoA dehydrogenase; Crt, crotonase; Bcd-EtfAB, butyryl-CoA
dehydrogenase-electron transfer protein; Ter, trans-enoyl-CoA reduc-
tase, AdhE2; alcohol/aldehyde dehydrogenase, BktB; β-ketothiolase.
shows activity toward hexenoyl-CoA.10 Considering that the
three-enzyme combination was able to dissociate hexenoyl-CoA
into acetyl-CoA and butyryl-CoA and these enzyme reactions are
reversible, we expect that these enzymes will be able to produce
hexenoyl-CoA using acetyl-CoA and butyryl-CoA under intra-
cellular conditions.
Since E. coli AtoB is in the same enzyme family as BktB, we also
characterized AtoB activity toward the C6 substrate. However,
no significant difference was detected between reactions with and
without AtoB (Supporting Information, Figure S1A,B), implying
that AtoB could not catalyze the condensation reaction of acetyl-
CoA and butyryl-CoA efficiently.
Shen et al.3 succeeded in the production of 1-butanol in high
concentrations by recruiting trans-enoyl-CoA reductase (Ter) in
lieu of C. acetobutylicum Bcd for the crotonyl-CoA reduction step.
Ter from Euglena gracilis (EgTer) was previously reported to be
able to convert hexenoyl-CoA to hexanoyl-CoA.11 In contrast,
11400
dx.doi.org/10.1021/ja203814d |J. Am. Chem. Soc. 2011, 133, 11399–11401

Journal of the American Chemical Society
COMMUNICATION
Table 1. Specific Activities of AdhE2 for Acetyl-CoA, Butyryl-
that of butyryl-CoA. Therefore, it would be important to increase
the intercellular hexanoyl-CoA concentration to enhance 1-hex-
anol production. Nevertheless, the strategy described in this
study may be extended further for the production of other even-
number longer chain alcohols.
CoA, Hexanoyl-CoA, and Octanoyl-CoAa
specific activities (mU/mg protein)
strain/plasmid acetyl-CoA butyryl-CoA hexanoyl-CoA octanoyl-CoA
JCL166
1.1
5.1
ND
3.7
’ ASSOCIATED CONTENT
JCL166/
30 ( 7.2
41 ( 6.9
15 ( 9.5
36 ( 6.9
pCS38_adhE2
S
Supporting Information. Experimental procedure, strains
b
a Cells were cultured under anaerobic condition. JCL166 was tested as a
and plasmids, HPLC analysis, MS spectrum, and time course of
1-hexanol production. This material is available free of charge via the
Internet at http://pubs.acs.org.
control. U = μmol/min. ND = not detectable.
’ AUTHOR INFORMATION
Corresponding Author
liaoj@seas.ucla.edu.
’ ACKNOWLEDGMENT
This work is supported by The KAITEKI Institute, Inc., Japan.
We thank Dr. Hidevaldo B. Machado and Dr. Hao Luo for helpful
discussion and Hyun-Jung Lim for her technical assistance.
Figure 2. GC analysis of 1-hexanol production in engineered E. coli:
(A) JCL299/pEL11/pEL102, (B) JCL166/pEL11/pEL102, and (C)
JCL166/pEL11/pIM8 cultures. I.S. = internal standard (2-methyl-1-
pentanol).
’ REFERENCES
(1) Atsumi, S.; Cann, A. F.; Connor, M. R.; Shen, C. R.; Smith, K. M.;
Brynildsen, M. P.; Chou, K. J.; Hanai, T.; Liao, J. C. Metab. Eng. 2008,
10, 305–311.
(2) Inui, M.; Suda, M.; Kimura, S.; Yasuda, K.; Suzuki, H.; Toda, H.;
Yamamoto, S.; Okino, S.; Suzuki, N.; Yukawa, H. Appl. Microbiol.
Biotechnol. 2008, 77, 1305–1316.
(3) Shen, C. R.; Lan, E. I.; Dekishima, Y.; Baez, A.; Cho, K. M.; Liao,
J. C. Appl. Environ. Microbiol. 2011, 77, 2905–2915.
(4) Barker, H. A.; Taha, S. M. J. Bacteriol. 1942, 43, 347–363.
(5) Thauer, R. K.; Jungermann, K.; Henninger, H.; Wenning, J.;
Decker, K. Eur. J. Biochem. 1968, 4, 173–180.
(6) Zhang, K.; Sawaya, M. R.; Eisenberg, D. S.; Liao, J. C. Proc. Natl.
Acad. Sci. U.S.A. 2008, 105, 20653–20658.
(7) Atsumi, S.; Hanai, T.; Liao, J. C. Nature 2008, 451, 86–89.
(8) Kopp, F.; Linne, U.; Oberthur, M.; Marahiel, M. A. J. Am. Chem.
Soc. 2008, 130, 2656–2666.
(9) Slater, S.; Houmiel, K. L.; Tran, M.; Mitsky, T. A.; Taylor, N. B.;
Padgette, S. R.; Gruys, K. J. J. Bacteriol. 1998, 180, 1979–1987.
(10) Waterson, R. M.; Castellino, F. J.; Hass, G. M.; Hill, R. L. J. Biol.
Chem. 1972, 247, 5266–5271.
(11) Hoffmeister, M.; Piotrowski, M.; Nowitzki, U.; Martin, W.
J. Biol. Chem. 2005, 280, 4329–4338.
(12) Tucci, S.; Martin, W. FEBS Lett. 2007, 581, 1561–1566.
(13) Fontaine, L.; Meynial-Salles, I.; Girbal, L.; Yang, X.; Croux, C.;
Soucaille, P. J. Bacteriol. 2002, 184, 821–830.
(14) Nielsen, D. R.; Leonard, E.; Yoon, S. H.; Tseng, H. C.; Yuan, C.;
Prather, K. L. Metab. Eng. 2009, 11, 262–273.
In the culture of JCL166/pEL11/pIM8, there was no detect-
able amount of 1-hexanol, but in cultures of JCL166/pEL11/
pEL102 and JCL299/pEL11/pEL102, 23 ( 11 and 27 ( 15 mg/
L of 1-hexanol were detected in 68 h, respectively (Figure 2). In
order to confirm the chemical identity of 1-hexanol, we further
analyzed these samples by GC-MS. The retention time and MS
spectra of the samples were identical to those of the 1-hexanol
standard (Figure S4).
Nielsen et al.14 and Shen et al.3 reported that overexpression of
fdh gene enhanced 1-butanol production by increasing the
intercellular NADH pool. So in the next step, we overexpressed
fdh using plasmid pCS138 in addition to the 1-hexanol synthesis
genes, and the resulting strain JCL299/pEL11/pEL102/pCS138
was examined. Furthermore, in order to replenish the carbon
supply and maintain neutral pH during fermentation, we re-
freshed the media every 24 h by extracting 10% of broth and
feeding same amount of fresh TB media containing adequate
amounts of glucose and NaOH. During the 7 h of aerobic
culturing, the cells grew up to 2.4 of OD600, but no 1-hexanol
production was detected. After switching to anaerobic condi-
tions, production of 1-hexanol started and reached at 47 mg/L, 48 h
(Figure S5). This strain also produced 5.1 and 6.5 g/L of
1-butanol at 48 and 68 h, respectively.
Further improvement of 1-hexanol production requires im-
proving the enzymatic activities toward C6 substrates. Currently,
1-butanol is still the main product, indicating that BktB (and/or
Hbd, Crt, Ter) is not active enough to produce hexanoyl-CoA.
Interestingly, AdhE2 has 70À80% activity toward acetyl-CoA
compared to butyryl-CoA; however, ethanol production is much
lower (∼0.2 g/L) than that of 1-butanol. This result implies that
the intercellular concentration of acetyl-CoA is much lower than
11401
dx.doi.org/10.1021/ja203814d |J. Am. Chem. Soc. 2011, 133, 11399–11401