ATP drives direct photosynthetic production
of 1-butanol in cyanobacteria
Ethan I. Lana,band James C. Liaoa,b,c,1
aDepartment of Chemical and Biomolecular Engineering,
Proteomics, University of California, Los Angeles, CA 90095
bBiomedical Engineering Interdepartmental Program, and
cInstitute for Genomics and
Edited by Lonnie O. Ingram, University of Florida, Gainesville, FL, and approved February 21, 2012 (received for review January 3, 2012)
While conservation of ATP is often a desirable trait for microbial
production of chemicals, we demonstrate that additional consump-
tion of ATP may be beneficial to drive product formation in a
nonnatural pathway. Although production of 1-butanol by the fer-
organisms, the first step of the pathway, condensation of two
molecules of acetyl-CoA to acetoacetyl-CoA, is thermodynamically
unfavorable. Here, we show that artificially engineered ATP con-
sumption through a pathway modification can drive this reaction
forward and enables for the first time the direct photosynthetic pro-
duction of 1-butanol from cyanobacteria Synechococcus elongatus
aldehyde/alcohol dehydrogenase (AdhE2) with separate butyralde-
hyde dehydrogenase (Bldh) and NADPH-dependent alcohol dehy-
drogenase (YqhD) increased 1-butanol production by 4-fold. These
results demonstrated the importance of ATP and cofactor driving
forces as a design principle to alter metabolic flux.
biofuel ∣ malonyl-CoA ∣ metabolic engineering ∣ synthetic biology
In particular, 1-butanol has received increasing attention be-
cause it is a potential fuel substitute and an important chemical
feedstock. 1-Butanol can be produced by two distinct routes: the
coenzyme A (CoA)-dependent pathway (1, 2) and the keto acid
pathway (3–5). The CoA-dependent pathway follows the chem-
istry of β-oxidation in the reverse direction, in which acetyl-CoA
is condensed to form acetoacetyl-CoA and then reduced to 1-bu-
tanol. The keto acid pathway utilizes 2-ketobutyrate as an inter-
mediate, which goes through keto acid chain elongation to
2-ketovalarate using the leucine biosynthesis enzymes. 2-Ketova-
larate is then decarboxylated and reduced into 1-butanol. In each
case, the pathway can be extended to produce 1-hexanol and
other higher alcohols (6–8).
The CoA-dependent reverse β-oxidation is a natural fermen-
tation pathway used by Clostridium species (9–11) and has been
transferred to various recombinant heterotrophs, resulting in
1-butanol titers ranging from 2.5 mg∕L to 1.2 g∕L with glucose
as the substrate (12–16). One of the challenges in transferring this
pathway to other organisms lies in the hydrogenation of crotonyl-
CoA to butyryl-CoA catalyzed by the butyryl-CoA dehydrogen-
ase/electron transferring flavoprotein (Bcd/EtfAB) complex.
Bcd/EtfAB complex is difficult to express in recombinant systems,
is presumably oxygen sensitive (12, 17), and possibly requires re-
duced ferredoxin as the electron donor (18). This difficulty was
overcome by expressing trans-2-enoyl-CoA reductase (Ter) (19,
20), which is readily expressed in Escherichia coli and directly re-
duces crotonyl-CoA using NADH. This modified 1-butanol path-
way (Fig. 1; outlined in blue) is catalyzed by five enzymes: thiolase
(AtoB), 3-hydroxybutyryl-CoA dehydrogenase (Hbd), crotonase
(Crt), Ter, and bifunctional aldehyde/alcohol dehydrogenase
(AdhE2). Simultaneously expressing these enzymes and engi-
neering NADH and acetyl-CoA accumulation as driving forces,
1-butanol production with a high titer of 15 g∕L and 88% of
iological production of chemicals and fuel from renewable
resources is an attractive approach to a sustainable future.
theoretical yield has been achieved using E. coli in flasks without
product removal (19). This result demonstrates the feasibility of
transferring the CoA-dependent pathway to nonnative organisms
for high-titer 1-butanol fermentation from glucose.
However, the success of the CoA-dependent pathway in
E. coli is not directly transferrable to photoautotrophs. By
expressing the same enzymes in cyanobacteria Synechococcus
elongatus PCC 7942, photosynthetic 1-butanol production from
CO2 was barely detectable (21). 1-Butanol production was
achieved by this strain only when internal carbon storage made
by CO2fixation in light conditions was fermented under anoxic
conditions (21). We hypothesized that both the acetyl-CoA and
NADH pools in this organism under photosynthetic conditions
may be insufficient to drive 1-butanol formation. Acetyl-CoA is
the precursor for fermentation pathway and the TCA cycle, both
of which are not active in light conditions. Furthermore, photo-
synthesis generates NADPH, but not NADH, and the intercon-
version between the two may not be efficient enough. Without a
significant driving force against the unfavorable thermodynamic
gradient, 1-butanol production cannot be achieved. The diffi-
culty of direct photosynthetic production of 1-butanol is in sharp
contrast to the production of isobutanol (450 mg∕L) and isobu-
tyraldehyde (1;100 mg∕L) by S. elongatus PCC 7942 (22), which
has an irreversible decarboxylation step as the first committed
reaction to drive the flux toward the products. This difference
suggests the importance of driving forces in altering the direc-
tion of metabolic flux.
We reason that instead of the acetyl-CoA pool, ATP may be
used to drive the thermodynamically unfavorable condensation
of two acetyl-coA molecules under photosynthetic conditions.
Thus, we engineered the ATP-driven malonyl-CoA synthesis and
decarboxylative carbon chain elongation used in fatty acid synth-
esis to drive the carbon flux into the formation of acetoacetyl-
CoA, which then undergoes the reverse β-oxidation to synthesize
1-butanol. We further replaced the subsequent NADH-depen-
dent enzymes with NADPH-dependent ones and successfully
achieved 1-butanol synthesis under photosynthetic conditions.
In theory, excess ATP consumption in the cell might cause a de-
crease in biomass. Thus, with notable exceptions (23–26), most
metabolic engineering design do not choose to increase ATP con-
sumption. Although many natural examples of microbes using
ATP to drive reactions, most of them are highly regulated. There-
fore, it is unpredictable whether it is feasible to use ATP con-
sumption to push flux in a nonnative pathway, for which no
Author contributions: E.I.L. and J.C.L. designed research; E.I.L. performed research;
E.I.L. and J.C.L. analyzed data; and E.I.L. and J.C.L. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Freely available online through the PNAS open access option.
1To whom correspondence may be addressed: E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/
6018–6023 ∣ PNAS ∣ April 17, 2012 ∣ vol. 109 ∣ no. 16www.pnas.org/cgi/doi/10.1073/pnas.1200074109
Incorporating an ATP Driving Force in 1-Butanol Pathway Design. The
thiolase-mediated condensation of two acetyl-CoA molecules
is reversible but strongly favors the thiolysis of acetoacetyl-CoA.
To examine the thermodynamic property of this reaction, we
overexpressed and purified E. coli thiolase (AtoB) and used an
in vitro assay to determine its equilibrium constant and ΔG∘0.
The result showed that the condensation reaction is unfavorable
(Fig. 2) with equilibrium constant (Keq) of ð1.1 ? 0.2Þ × 10−5at
pH 8.0, within the optimum pH range for cyanobacteria (27). This
experimentally determined Keqapproximately corresponds to a
ΔG∘0of 6.8 kcal∕mol, consistent with the literature reported Keq
using partially purified thiolase from pig heart protein homoge-
nate (28). Therefore, without a sufficiently large acetyl-CoA pool
or an efficient product trap, there is no driving force for the for-
mation of acetoacetyl-CoA. Although the irreversible hydrogena-
tion of crotonyl-CoA catalyzed by Ter provides a driving force, it
is insufficient to drive the reaction forward without large pools of
acetyl-CoA and reducing equivalent (19).
Instead of using the direct condensation of acetyl-CoA, we
contemplated an alternative route through the ATP-driven mal-
onyl-CoA synthesis. Malonyl-CoA is synthesized from acetyl-
formation of malonyl-CoA is effectively irreversible due to
ATP hydrolysis. In fatty acid synthesis, malonyl-CoA is then con-
verted into malonyl-acyl carrier protein (malonyl-ACP) and acts
as the carbon addition unit for fatty acid synthesis. For 1-butanol
synthesis, malonyl-CoA can react with acetyl-coA in a decarbox-
−and ATP by acetyl-CoA carboxylase (Acc). The
ylative condensation to form acetoacetyl-CoA, in a reaction ana-
logous to ketoacyl-ACP synthase III (KAS III) that catalyzes the
irreversible condensation of malonyl-ACP and acetyl-CoA to
synthesize the four carbon intermediate 3-ketobutyryl-ACP.
We note that the energy release from ATP hydrolysis (ΔG∘0of
−7.3 kcal∕mol) would compensate for the energy required for
condensation of acetyl-CoA into acetoacetyl-CoA. By combining
the reaction catalyzed by thiolase with ATP hydrolysis (Fig. S1),
the net reaction is thermodynamically favorable (ΔG∘0< 0),
which would reduce the need for high concentration of acetyl-
CoA required to push the reaction forward. More importantly,
CO2released from the second step, decarboxylative chain elon-
gation, shifts the reaction toward the formation of acetoacetyl-
CoA. Fatty acid and polyketide syntheses have naturally evolved
this mechanism to enable the thermodynamically unfavorable
formation of 3-ketoacyl-ACP. By taking advantage of this me-
chanism, it is possible to push the carbon flux to the 1-butanol
pathway without the acetyl-CoA driving force artificially con-
structed in E. coli (19).
Expression of Acetoacetyl-CoA Synthase Enables Photosynthetic Pro-
duction of 1-Butanol. Therefore, we bioprospected for a KAS III
that utilizes malonyl-CoA rather than malonyl-ACP for conden-
sation with acetyl-CoA. Because both ACP and CoA carry phos-
phopantetheine, which forms thioester bond with the malonyl
moiety, KAS III and KAS III-like enzymes may also react with
malonyl-CoA. We cloned a variety of KAS III and KAS III-like
enzymes from different organisms and examined their expression
in E. coli (Fig. S2). After His-tag purification, we assayed their
activity for condensing malonyl-CoA with acetyl-CoA (Table 1).
Among the enzymes tested, NphT7 (29) was the most active.
Other enzymes such as Bamb6244, GOX0115, and PAE-FabH2
were also active while the rest showed no detectable activity. The
condensation reaction (Fig. S3) catalyzed by NphT7 using mal-
onyl-CoA and acetyl-CoA is irreversible and accumulates acetoa-
cetyl-CoA as the product. At low starting concentrations of
malonyl-CoA, conversion yield to acetoacetyl-CoA is higher than
high starting substrate concentrations. This result is likely due to
the fact that NphT7 also catalyzes malonyl-CoA self condensa-
tion and is particularly useful for 1-butanol synthesis, because
both malonyl-CoA and acetyl-CoA pools are expected to be low
in S. elongatus PCC 7942.
Next, we constructed a plasmid harboring genes nphT7, hbd,
crt, and adhE2 under an IPTG inducible promoter PLlacO1with
Neutral Site II (NSII) recombination sequences flanking the
genes and kanamycin resistance marker (Fig. 3). By DNA homo-
logous recombination, we integrated these genes into the genome
of S. elongatus strain EL9 [expressing ter at Neutral Site I (NSI)
R. eutropha; CA, C. acetobutylicum; AC, A. caviae; TD, T. denticola; CS, C. saccharoperbutylacetonicum N1-4; CL190, Streptomyces sp. strain CL190.
Variations in the CoA-dependent 1-butanol pathway. The fermentative CoA 1-butanol pathway is in blue. Alternative routes are in red. EC, E. coli; RE,
mediated reaction. The equilibrium constant (Keq) was determined from the
equilibrium concentrations. E. coli AtoB was cloned, purified, and used in an
in vitro assay. AcCoA, acetyl-CoA; AcAcCoA, acetoacetyl-CoA; CoA, coenzyme
A. Detailed conditions and methods are listed in SI Text.
Determination of equilibrium concentrations for the thiolase (AtoB)
Lan and LiaoPNAS
April 17, 2012
under the control of another IPTG inducible promoter Ptrc] at
NSII and selected for successful transformant on kanamycin con-
taining BG-11 plates. The successful transformant strain EL20
was then analyzed for in vitro enzyme activity and 1-butanol pro-
duction. As shown in (Fig. 4A), crude extract from strain EL20
catalyzed the formation of acetoacetyl-CoA by condensation of
malonyl-CoA and acetyl-CoA and did not catalyze the thiolysis
of acetoacetyl-CoA (Fig. 4B). On the other hand, crude extract
from strain EL14 expressing atoB along with hbd, crt, ter, and
adhE2 catalyzed thiolysis (Fig. 4B) much more efficiently than
the condensation reaction (Fig. 4A). The two strains EL20 and
EL14 exhibited nearly identical growth rate (Fig. 4C). However,
Strain EL20 produced 6.5 mg∕L (Fig. 4D) of 1-butanol while
Strain EL14 produced only barely detectable amounts (detection
limit of about 1 mg∕L) of 1-butanol (Fig. 4D). This result indi-
cated that ATP-driven acetoacetyl-CoA formation is required for
photosynthetic production of 1-butanol using the CoA-depen-
Substitution of NADPH Utilizing Enzymes Aids 1-Butanol Production.
Another useful driving force in 1-butanol synthesis is the reducing
equivalent (19). Cyanobacteria produce NADPH as the direct re-
sult of photosynthesis. Intracellular NADþand NADPþlevels
exist in a ratio of about 1∶10 (30) in S. elongatus PCC 7942. Thus
NADH utilizing pathway may be unfavorable in cyanobacteria.
To synthesize 1 mol of 1-butanol from acetoacetyl-CoA requires
4 mol of NADH. Therefore, changing the cofactor preference to
NADPH may aid the production of 1-butanol.
As depicted in Fig. 1 (outlined in red), we identified enzymes
that utilize NADPHor both NADPHand NADHby bioprospect-
ing. Acetoacetyl-CoA reductase (PhaB) (31) was used to replace
Hbd. PhaB from Ralstonia eutropha is an enzyme found in the
poly-hydroxyalkanoate biosynthetic pathway for reducing 3-keto-
butyryl-CoA to 3-hydroxybutyryl-CoA using NADPH. However,
PhaB produces the (R)-stereoisomer of 3-hydroxybutyryl-CoA
instead of the (S)-stereoisomer produced by Hbd. As a result,
Crt cannot be used for the subsequent dehydration to produce
crotonyl-CoA. Therefore, a different crotonase is necessary for
dehydration of (R)-3-hydroxybutyryl-CoA. (R)-specific enoyl-
CoA hydratase (PhaJ) (32) is found in Aeromonas caviae and
dehydrates (R)-3-hydroxybutyryl-CoA into crotonyl-CoA. To-
gether, PhaB and PhaJ can replace Hbd and Crt, respectively,
Table 1. Specific activities of acetoacetyl-CoA synthases (μmol∕min∕mg)
EnzymeSpecific activityEnzymeSpecific activity
Burkholderia ambifaria BAMB6244
Gluconobacter oxydans GOX0115
Helicobacter pylori HP0202
Listeria monocytogenes LMO2202
0.0116 ± 0.0002
0.0099 ± 0.0011
Pseudomonas aeruginosa PAE-FabH2
Streptomyces avermitilis SAV-FabH4
Streptomyces coelicolor SCO5858
Streptomyces sp. strain CL190 NphT7
0.0140 ± 0.0010
6.02 ± 0.25
NSI, (B) atoB, adhE2, crt, and hbd at NSII in the genome of S. elongatus. Dif-
ferent combinations of alternative genes nphT7, bldh, yqhD, phaJ, and phaB
can replace their counterpart enzymes to recombine into NSII. (C) List of
strains with different combinations of overexpressed genes used in this study.
Schematic representation of recombination to integrate (A) ter at
extracts of wild-type S. elongatus PCC 7942, strain EL14 and EL20. (B) In vitro
assay for the thiolysis of acetoacetyl-CoA using crude extracts of wild-type
S. elongatus PCC 7942, strain EL14 and EL20. (C) Cell density and (D) 1-butanol
accumulation as a function of time of strain EL14 and EL20. 1-Butanol pro-
duction by strain EL14 was near detection limit of about 1 mg∕L.
(A) In vitro assay for the synthesis of acetoacetyl-CoA using crude
www.pnas.org/cgi/doi/10.1073/pnas.1200074109Lan and Liao
to change the cofactor preference of 3-ketobutyryl-CoA reduc-
tion to NADPH.
To replace AdhE2, NADP-dependent alcohol dehydrogenase
(YqhD) (33) from E. coli has been demonstrated to aid the
production of higher chain alcohols (22). In addition, we needed
a CoA-acylating butyraldehyde dehydrogenase (Bldh) to replace
the aldehyde dehydrogenase function of AdhE2. We thus bio-
prospected for enzymes catalyzing reduction of butyryl-CoA
to butyraldehyde. Bldh was found in high butanol producing
Clostridium species including C. beijerinckii NCIMB 8052 (34),
C. saccharobutylicum ATCC BAA-117, and C. saccharoperbuty-
lacetonicum NI-4 (35). In particular, Bldh from C. beijerinckii
has been purified and demonstrated activity in vitro with both
NADH and NADPH as a reducing cofactor.
Using the sequence of C. beijerinckii Bldh, we searched
by homology and cloned additional Bldh-like enzymes from var-
ious organisms including C. saccharoperbutylacetonicum NI-4,
C. saccharobutylicum ATCC BAA-117, Geobacillus thermogluco-
sidasius, Clostridium kluyveri, and E. coli. We assessed the perfor-
mance of these Bldh’s by 1-butanol production in recombinant
E. coli. As shown in Fig. 5, the E. coli strain expressing C. sac-
charoperbutylacetonicum NI-4 Bldh along with rest of the CoA
1-butanol pathway produced the highest titer of 1-butanol, ex-
ceeding the 1-butanol produced by E. coli strain expressing
AdhE2 by nearly threefold.
To test the effect of cofactor utilization, we constructed various
combinations of different routes by overexpressing different
genes in S. elongatus PCC 7942 (Fig. 3C). We constructed plas-
mids containing different genes and recombined them into the
genome of S. elongatus strain EL9. We then assayed the activity
of overexpressed enzymes to confirm expression (Fig. 6). Of the
strains tested, strain EL22 expressing the NADPH utilizing en-
zymes produced the highest amount of 1-butanol (29.9 mg∕L)
exceeding that of EL20 (6.4 mg∕L) by more than fourfold. This
result reinforced the importance of cofactor as driving force.
In a metabolic system involving multiple pathways, the direction
and rate of each reaction are determined by kinetics, regulated by
the enzyme expression levels and metabolite pool sizes. Typically,
a reaction with a large positive ΔG∘0is considered practically un-
feasible in the forward direction because it requires high concen-
trations of the substrate pool to drive the reaction forward. The
condensation of two molecules of acetyl-CoA to acetoacetyl-CoA
is such an example.On the otherhand, the reverse direction, thio-
lysis of acetoacetyl-CoA, is readily achievable and used as the last
step in the β-oxidation. However, some fermentative organisms,
such as Clostridium species, use direct acetyl-CoA condensation
for 1-butanol synthesis. These organisms accomplish this thermo-
dynamically unfavorable reaction presumably through a large
pool of acetyl-CoA and high reducing equivalents that drive
the subsequent reactions. This situation was recreated in E. coli
expressing the enzymes for 1-butanol synthesis (19). Unfortu-
nately, this strategy cannot be readily implemented in photosyn-
thetic organisms for multiple reasons: Acetyl-CoA is a precursor
for fermentative pathways and the oxidative TCAcycle, which are
not active under photosynthetic conditions. Thus, the acetyl-CoA
pool size is not expected to be high and is difficult to modulate.
Instead of using high acetyl-CoA pool as a driving force, here
ATP and the evolution of CO2effectively drive the reactions to-
ward 1-butanol synthesis. This strategy is used in fatty acid synth-
esis in nature, and is used to couple the fatty acid synthesis to the
energy status in the cell.
ATP consumption has been used by cells to drive various ther-
modynamically unfavorable reactions. In engineered E. coli, ATP
consumption has been used to stimulate glycolysis by futile cy-
cling (23, 24) or by deletion of membrane-coupling subunits in
(F1F0)-ATP synthase (25). Increased ATP consumption by over-
expressing enzymes that promote malonyl-CoA biosynthesis also
increased production yield of compounds downstream of malo-
nyl-CoA (26). However, additional ATP consumption in hetero-
logous pathways may cause adverse effects in the cell and may
result in reduced biomass formation. As such, most metabolic en-
gineering design models have primarily focused on maximizing
carbon yield and minimizing ATP expenditure. Here we provide
a distinct example using ATP to alter the thermodynamics of the
CoA-dependent 1-butanol pathway, which naturally does not
require additional ATP consumption. By incorporating an ATP-
dependent step into the CoA-dependent 1-butanol synthesis
pathway, we demonstrated for the first time the direct photosyn-
thetic production of 1-butanol from CO2.
Direct photosynthetic 1-butanol production from CO2 is
desirable because it reduces the number of processing steps.
S. elongatus PCC 7942 is naturally competent and therefore is
an attractive model organism for engineering. The DNA recombi-
nation method used in this study has also been broadly practiced
for engineering cyanobacteria for the production of various chemi-
cals. By metabolic engineering, cyanobacteria has also enabled the
production of Isobutyraldehyde (1;100 mg∕L) (22), isobutanol
(450 mg∕L) (22), ethanol (550 mg∕L) (36), ethylene (451 nl∕nl∕
h∕OD730) (37), isoprene (0.05 mg∕g dry cell weight) (38), sugars
(45 mg∕L) (39), lactic acid (56 mg∕L) (39), fatty alcohols
(0.2 mg∕L) (40), and fatty acids (194 mg∕L) (41) from CO2. The
pathways for the relatively high production of isobutyraldehyde,
isobutanol and ethanol naturally involve a decarboxylation step
as the first committed reaction. The loss of CO2is considered ir-
reversible and serves as a driving force to the product formation.
Our result is also consistent with this phenomenon that decarbox-
ylation aids in directing the carbon flux.
Reducing cofactor preference is another important aspect
of pathway design. Depending on the production condition and
organisms’ natural metabolism, changing cofactor preference is
necessary to achieve high flux production. For example, changing
NADPH-dependent enzymes into NADH-dependent increases
the isobutanol productivity and yield under anaerobic condition
in recombinant E. coli (42). Replacing the Bcd-EtfAB complex
that requires an unknown electron donor with an NADH-depen-
dent Ter is helpful in 1-butanol production (19, 20). In contrast,
pathways utilizing NADPH are preferred in cyanobacteria be-
cause NADPH is more abundant. By utilizing NADPH-depen-
dent enzymes, our 1-butanol production enhanced from 6.5 mg∕
L to 29.9 mg∕L (Fig. 6).
Current limitation of our 1-butanol production using cyano-
bacteria may be the synthesis of malonyl-CoA. Compared to
the high flux production of isobutanol and isobutyraldehyde in
cyanobacteria, the carbon flux through our 1-butanol pathway
is suboptimal. Malonyl-CoA biosynthesis is considered as the lim-
iting step in fatty acid synthesis (43). Therefore, increasing carbon
flux toward the synthesis of acetyl-CoA and malonyl-CoA may be
necessary to increase 1-butanol production. Intracellular acetyl-
CoA and malonyl-CoA supply may be increased by increasing
CoA biosynthesis (44), overexpression of Acc (45–48), glycolytic
JCL299 expressing CoA-dependent 1-butanol pathway with YqhD from E. coli
and Bldh from different organisms. Dashed line represents the baseline pro-
duction by using AdhE2. Detailed production procedure is listed in SI Text.
Production of 1-butanol and ethanol by recombinant E. coli strains
Lan and LiaoPNAS
April 17, 2012
enzymes such as phosphoglycerate kinase and glyceraldehyde-
3-phosphate dehydrogenase (26), and inhibition of fatty acid
biosynthesis (49). With these approaches, the 1-butanol produc-
tion in cyanobacteria may be further improved.
Materials and Methods
For details, see SI Text.
Culture Medium and Condition. All S. elongatus PCC 7942 strains were grown
on modified BG-11 (1.5 g∕L NaNO3, 0.0272 g∕L CaCl2·2H2O, 0.012 g∕L ferric
ammonium citrate, 0.001 g∕L Na2EDTA, 0.040 g∕L K2HPO4, 0.0361 g∕L
MgSO4·7H2O, 0.020 g∕L Na2CO3, 1;000 × trace mineral (1.43 gH3BO3,
0.905 g∕L MnCl2·4H2O, 0.111 g∕L ZnSO4·7H2O, 0.195 g∕L Na2MoO4·2H2O,
0.0395 g CuSO4·5H2O, 0.0245 g CoðNO3Þ2·6H2O), 0.00882 g∕L sodium citrate
dihydrate (50)) agar (1.5% w∕v) plates. All S. elongatus PCC 7942 strains were
cultured in BG-11 medium containing 50 mM NaHCO3in 250 mL screw cap
flasks. Cultures were grown under 100 μE∕s∕m2light, supplied by four Lumi-
chrome F30W-1XX 6500K 98CRI light tubes, at 30 °C. Cell growth was mon-
itored by measuring OD730with Beckman Coulter DU800 spectrophotometer.
Production of 1-Butanol. A loopful of S. elongatus PCC 7942 was used to in-
oculate fresh 50 mL BG-11. 500 mM IPTG was used to induce the growing
culture at cell density OD730 nmof 0.4 to 0.6 with 1 mM IPTG as final concen-
tration. 5 mL of growing culture was sampled for cell density and 1-butanol
production measurements every 2 d for up to day 8 after which sampling
time was switched to every 3 d. After sampling, 5 mL of fresh BG-11 with
500 mM NaHCO3, appropriate antibiotics, and IPTG were added back to
the culture. Method for 1-butanol quantification is listed in SI Text.
Enzyme Assays. Thiolase activity was measured via both condensation and
thiolysis direction. The enzymatic reaction was monitored by the increase
or decrease of absorbance at 303 nm. Acetoacetyl-CoA synthase activity
was measured by monitoring the increase of absorbance at 303 nm, which
corresponds to appearance of acetoacetyl-CoA. For details, see SI Text.
Alcohol Production by E. coli Expressing Butyraldehyde Dehydrogenase. Trans-
formed E. coli strain JCL299 (ΔadhE, ΔldhA, Δfrd, Δpta) expressing different
Bldh and the CoA-dependent pathway were cultured in Terrific broth
(TB; 12 g tryptone, 24 g yeast extract, 2.31 g KH2PO4, 12.54 g K2HPO4,
4 mL glycerol per liter of water) supplemented with 20 g∕L glucose. Fermen-
tation was conducted for 2 d, after which samples were taken for measure-
ment of 1-butanol concentration. For details, see SI Text.
DNA Manipulations. For details, see SI Text.
Plasmid Constructions. For details, see SI Text.
Strain Construction and Transformation. For details, see SI Text.
Protein Purification and SDS/PAGE. For details, see SI Text.
1-Butanol Quantification. For details, see SI Text.
ACKNOWLEDGMENTS.This research was supported by the Kaiteki Institute and
partially supported by National Science Foundation Grant NSF MCB1221392.
The authors would like to thank Dr. Hao Luo, Dr. Claire R. Shen, Yasumasa
Dekishima, Dr. Hidevaldo Machado, and Dr. Kwang Myung Cho for their
1. Papoutsakis ET (2008) Engineering solventogenic clostridia. Curr Opin Biotechnol
2. Ezeji TC, Qureshi N, Blaschek HP (2007) Bioproduction of butanol from biomass: From
genes to bioreactors. Curr Opin Biotechnol 18:220–227.
3. Atsumi S, Hanai T, Liao JC (2008) Non-fermentative pathways for synthesis of
branched-chain higher alcohols as biofuels. Nature 451:86–89.
4. Atsumi S, Liao JC (2008) Directed evolution of Methanococcus jannaschii citramalate
synthasefor biosynthesisof1-propanol and1-butanolbyEscherichiacoli.Appl Environ
5. Shen CR, Liao JC (2008) Metabolic engineering of Escherichia coli for 1-butanol and
1-propanol production via the keto-acid pathways. Metab Eng 10:312–320.
6. Dekishima Y, Lan EI, Shen CR, Cho KM, Liao JC (2011) Extending carbon chain length of
1-butanol pathway for 1-hexanol synthesis from glucose by engineered Escherichia
coli. J Am Chem Soc 133:11399–11401.
7. Dellomonaco C, Clomburg JM, Miller EN, Gonzalez R (2011) Engineered reversal of the
beta-oxidation cycle for the synthesis of fuels and chemicals. Nature 476:355–359.
8. Zhang K, Sawaya MR, Eisenberg DS, Liao JC (2008) Expanding metabolism for
biosynthesis of nonnatural alcohols. Proc Natl Acad Sci USA 105:20653–20658.
enables direct photosynthetic production of 1-butanol under oxygenic condition. Strains EL21 and EL22 expressing bldh and yqhD achieved the highest
1-Butanol production and enzyme activity of strains expressing different enzymes. Strain genotype is listed in Fig. 3C. Expression of nphT7
www.pnas.org/cgi/doi/10.1073/pnas.1200074109Lan and Liao