citrate-oxaloacetate shuttle, measured as an
increase in FA titers by about 20%.
An incremental improvement was achieved by introducing an additional ACL gene from Aspergillus nidulans
(a murine ACL was already present in the engineered
strain). Simultaneously the isocitrate dehydrogenase
Idp2 and the citrate-a-ketoglutarate antiporter Yhm2,
which also accepts oxaloacetate instead of a-ketoglutarate, were overexpressed. Idp2 was supposed to oxidize
citrate to a-ketoglutarate and thereby provide additional NADPH for FA synthesis. This, however, can
only occur at the expense of the cytosolic citrate and
therefore AcCoA pool. Yhm2 overexpression was
intended to increase the antiport of citrate with oxaloacetate. Unfortunately, since three genes were introduced at once as compared to the reference strain,
their individual contribution to the improvement cannot be disentangled.
A further substantial improvement of FA
synthesis in this strain background was
achieved by diverting the glucose flux from
glycolysis to the oxidative PPP to increase the
NADPH pool. To this end, the expression of the
phosphoglucose isomerase (PGI1) was downregulated, whereas the endogenous genes
encoding PPP enzymes glucose-6-phosphate
dehydrogenase (ZWF1), 6-phosphogluconate
dehydrogenase (GND1), transketolase (TKL1),
and transaldolase (TAL1) were concomitantly
overexpressed. A downregulation of the mitochondrial isocitrate dehydrogenase IDH2 to
prevent the degradation of citrate in the TCA
led to a further improvement but only in a
strain with the upregulated PPP. In its precursor strain, lowering Idh2 activity had no beneficial effect, and this notion is consistent with
previous results, obtained in a strain that
contained the citrate-oxaloacetate shuttle, but
was not engineered for an increased supply of
NADPH (Tang et al. 2013). Collectively, these
studies show that sufficient reducing power is
essential for efficient FA production.
E. Strategies to Manipulate NADPH Level
Independently of Precursor Supply Routes
Besides abovementioned possibilities of overexpressing enzymes that yield NADPH (e.g.,
Ald6, Zwf1) or creating transhydrogenase
cycles (e.g., oxaloacetate-malate-pyruvate), the
level of this cofactor can be manipulated
through interventions into glycolysis or amino
acid metabolism. One possibility to increase the
pool of NADPH is bypassing the main source of
cellular NADH—the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) reaction—by
a NADP-dependent enzyme. The expression of
a heterologous non-phosphorylating NADPdependent GAPDH in strains engineered for
polyhydroxybutyrate (Kocharin et al. 2013) or
FAEE (Shi et al. 2014a) production led to considerable product yield improvements, but not
in a strain designed to produce fatty alcohols
(d’Espaux et al. 2017). It has to be noted that—
besides different product pathways—the
genetic interventions into the carbon metabolism of these strains largely differed, suggesting
that the applicability of this strategy is contextdependent. One important consideration is that
the activity of a non-phosphorylating GAPDH
will reduce the ATP yield of glycolysis, which,
as outlined above, ultimately will lead to a
reduction of the maximally attainable yield.
The implementation of a phosphorylating
NADP-dependent GAPDH does not suffer
from this downfall, but its utility for production
of oleochemicals in S. cerevisiae could not yet
be demonstrated (d’Espaux et al. 2017). A complementary approach is reducing the NADPH
consumption through competing pathways.
Based on the observation that a large proportion of NADPH is oxidized by NADP-dependent glutamate dehydrogenase Gdh1, d’Espaux
et al. (2017) demonstrated that a deletion of the
GDH1 gene is beneficial for fatty alcohol production.
F. Engineering Malonyl-CoA Supply
MalCoA, the extender unit for FA synthesis via
FAS, is synthesized by carboxylation of AcCoA
in an ATP-dependent manner by the endogenous enzyme AcCoA carboxylase (Acc1). In
some studies, the native form of Acc1 was overexpressed to boost the production of FA derivatives, which indeed led to higher product
titers (Shin et al. 2012; Runguphan and Keasling 2014). As an alternative approach, Wang
346
L. Baumann et al.
increase in FA titers by about 20%.
An incremental improvement was achieved by introducing an additional ACL gene from Aspergillus nidulans
(a murine ACL was already present in the engineered
strain). Simultaneously the isocitrate dehydrogenase
Idp2 and the citrate-a-ketoglutarate antiporter Yhm2,
which also accepts oxaloacetate instead of a-ketoglutarate, were overexpressed. Idp2 was supposed to oxidize
citrate to a-ketoglutarate and thereby provide additional NADPH for FA synthesis. This, however, can
only occur at the expense of the cytosolic citrate and
therefore AcCoA pool. Yhm2 overexpression was
intended to increase the antiport of citrate with oxaloacetate. Unfortunately, since three genes were introduced at once as compared to the reference strain,
their individual contribution to the improvement cannot be disentangled.
A further substantial improvement of FA
synthesis in this strain background was
achieved by diverting the glucose flux from
glycolysis to the oxidative PPP to increase the
NADPH pool. To this end, the expression of the
phosphoglucose isomerase (PGI1) was downregulated, whereas the endogenous genes
encoding PPP enzymes glucose-6-phosphate
dehydrogenase (ZWF1), 6-phosphogluconate
dehydrogenase (GND1), transketolase (TKL1),
and transaldolase (TAL1) were concomitantly
overexpressed. A downregulation of the mitochondrial isocitrate dehydrogenase IDH2 to
prevent the degradation of citrate in the TCA
led to a further improvement but only in a
strain with the upregulated PPP. In its precursor strain, lowering Idh2 activity had no beneficial effect, and this notion is consistent with
previous results, obtained in a strain that
contained the citrate-oxaloacetate shuttle, but
was not engineered for an increased supply of
NADPH (Tang et al. 2013). Collectively, these
studies show that sufficient reducing power is
essential for efficient FA production.
E. Strategies to Manipulate NADPH Level
Independently of Precursor Supply Routes
Besides abovementioned possibilities of overexpressing enzymes that yield NADPH (e.g.,
Ald6, Zwf1) or creating transhydrogenase
cycles (e.g., oxaloacetate-malate-pyruvate), the
level of this cofactor can be manipulated
through interventions into glycolysis or amino
acid metabolism. One possibility to increase the
pool of NADPH is bypassing the main source of
cellular NADH—the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) reaction—by
a NADP-dependent enzyme. The expression of
a heterologous non-phosphorylating NADPdependent GAPDH in strains engineered for
polyhydroxybutyrate (Kocharin et al. 2013) or
FAEE (Shi et al. 2014a) production led to considerable product yield improvements, but not
in a strain designed to produce fatty alcohols
(d’Espaux et al. 2017). It has to be noted that—
besides different product pathways—the
genetic interventions into the carbon metabolism of these strains largely differed, suggesting
that the applicability of this strategy is contextdependent. One important consideration is that
the activity of a non-phosphorylating GAPDH
will reduce the ATP yield of glycolysis, which,
as outlined above, ultimately will lead to a
reduction of the maximally attainable yield.
The implementation of a phosphorylating
NADP-dependent GAPDH does not suffer
from this downfall, but its utility for production
of oleochemicals in S. cerevisiae could not yet
be demonstrated (d’Espaux et al. 2017). A complementary approach is reducing the NADPH
consumption through competing pathways.
Based on the observation that a large proportion of NADPH is oxidized by NADP-dependent glutamate dehydrogenase Gdh1, d’Espaux
et al. (2017) demonstrated that a deletion of the
GDH1 gene is beneficial for fatty alcohol production.
F. Engineering Malonyl-CoA Supply
MalCoA, the extender unit for FA synthesis via
FAS, is synthesized by carboxylation of AcCoA
in an ATP-dependent manner by the endogenous enzyme AcCoA carboxylase (Acc1). In
some studies, the native form of Acc1 was overexpressed to boost the production of FA derivatives, which indeed led to higher product
titers (Shin et al. 2012; Runguphan and Keasling 2014). As an alternative approach, Wang
346
L. Baumann et al.