tion was reported when a bacterial PDH or the
native PDH components lacking the mitochondrial targeting sequence were overexpressed in
yeast cytosol (Lian et al. 2014).
It has to be noted, however, that the cytosolic PDH
activity was not directly demonstrated in this study.
In another study (Kozak et al. 2014a), it has been
shown that the activity of PDH in the yeast cytosol is
strictly dependent on supplementation of lipoic acid, a
cofactor that is normally synthesized inside mitochondria and must be covalently linked to the E2 subunit of
PDH. Moreover, overexpression of enzymes involved in
lipoylation of E2 in the cytosol is crucial (Kozak et al.
2014a). Since Lian et al. (2014) did neither supplement
lipoic acid nor express the components of the lipoylation machinery, the mechanism by which cPDH led to
an improvement of n-butanol production in their work
remains enigmatic.
Of particular interest for FA synthesis via
the (NADPH-dependent) FAS would be an
implementation of NADPH-yielding cPDH variants. Indeed, a mutated (NADP
+ -accepting)
PDH was developed (Bocanegra et al. 1993)
and implemented to produce the polyketide
triacetic acid lactone in yeast cytosol. Although
the cytosolic PDH activity was not directly
measured, this strategy led to a significantly
increased NADPH/NADP ratio and AcCoA
levels in whole cell extracts as well as to
improved product titers in fermentations (Cardenas and Da Silva 2016). Since this approach
has the same cofactor yield (one NADH and one
NADPH per molecule AcCoA) but a better
energetic balance compared to the PDH bypass,
it could potentially improve the production of
FA. However, the study did not address the
abovementioned requirement for lipoylation,
and this aspect will therefore require further
investigation before this approach can be
employed for FA production.
C. Phosphoketolase Pathway
Among all individually considered alternative
AcCoA pathways, the phosphoketolase/transacetylase variant was predicted to enable the
highest yield of FAS-derived FA (87% of the
theoretical maximum), when the carbon flux
is rerouted from glycolysis to the oxidative
pentose phosphate pathway (PPP) to increase
the NADPH pool (van Rossum et al. 2016b). In
this scheme, glucose-6-phosphate is converted
to xylulose 5-phosphate (X5P) through a
consecutive action of the endogenous enzymes
glucose-6-phosphate
dehydrogenase,
6phosphogluconolactonase,
6-phosphoglu
conate dehydrogenase, and D-ribulose-5-phosphate 3-epimerase, whereby the first and the
third reaction yield one NADPH molecule
each (for a review of the PPP, see Stincone
et al. 2015). X5P is then converted to acetylphosphate (AcP) and glyceraldehyde-3-phosphate (GAP) by a heterologous phosphoketolase (PK). Subsequently, AcP is directly
converted to AcCoA by a heterologous phosphotransacetylase (PTA). The PK/PTA pathway was expressed in a strain engineered to
produce FAEE and indeed led to a significant
increase of product titers compared to the reference strain containing only the native PDH
bypass (Jong et al. 2014).
Another variant to convert PKA-derived AcP to AcCoA
involves the acetate kinase (ACK), which yields acetate
and ATP by transferring the phosphate residue from
AcP to ADP. Although ATP is produced in this reaction, the production of AcCoA via this route is energetically less favorable than via PTA since two ATP have to
be invested into the subsequent activation of acetate by
ACS (see above). The implementation of this strategy
led to a significant improvement of polyhydroxybutyrate (Kocharin et al. 2013) and, to a lower extent than
with the PK/PTA variant, FAEE production (Jong et al.
2014).
In one recent study (Meadows et al. 2016),
the PK/PTA pathway was combined with the AALD to produce the isoprenoid farnesene, as
the combinatorial configuration was calculated
to lead to the highest possible product yields
(van Rossum et al. 2016b). Interestingly, the
authors found that AcP produced by PK is
partly dephosphorylated by endogenous (promiscuous) glycerol-3-phosphate phosphatases
and their activity needs to be reduced to favor
the PTA reaction. This study stands out as the
only one in which the endogenous PDH bypass
was fully substituted (by ald6 acs1 acs2 deletion) with heterologous AcCoA synthesis routes
344
L. Baumann et al.
native PDH components lacking the mitochondrial targeting sequence were overexpressed in
yeast cytosol (Lian et al. 2014).
It has to be noted, however, that the cytosolic PDH
activity was not directly demonstrated in this study.
In another study (Kozak et al. 2014a), it has been
shown that the activity of PDH in the yeast cytosol is
strictly dependent on supplementation of lipoic acid, a
cofactor that is normally synthesized inside mitochondria and must be covalently linked to the E2 subunit of
PDH. Moreover, overexpression of enzymes involved in
lipoylation of E2 in the cytosol is crucial (Kozak et al.
2014a). Since Lian et al. (2014) did neither supplement
lipoic acid nor express the components of the lipoylation machinery, the mechanism by which cPDH led to
an improvement of n-butanol production in their work
remains enigmatic.
Of particular interest for FA synthesis via
the (NADPH-dependent) FAS would be an
implementation of NADPH-yielding cPDH variants. Indeed, a mutated (NADP
+ -accepting)
PDH was developed (Bocanegra et al. 1993)
and implemented to produce the polyketide
triacetic acid lactone in yeast cytosol. Although
the cytosolic PDH activity was not directly
measured, this strategy led to a significantly
increased NADPH/NADP ratio and AcCoA
levels in whole cell extracts as well as to
improved product titers in fermentations (Cardenas and Da Silva 2016). Since this approach
has the same cofactor yield (one NADH and one
NADPH per molecule AcCoA) but a better
energetic balance compared to the PDH bypass,
it could potentially improve the production of
FA. However, the study did not address the
abovementioned requirement for lipoylation,
and this aspect will therefore require further
investigation before this approach can be
employed for FA production.
C. Phosphoketolase Pathway
Among all individually considered alternative
AcCoA pathways, the phosphoketolase/transacetylase variant was predicted to enable the
highest yield of FAS-derived FA (87% of the
theoretical maximum), when the carbon flux
is rerouted from glycolysis to the oxidative
pentose phosphate pathway (PPP) to increase
the NADPH pool (van Rossum et al. 2016b). In
this scheme, glucose-6-phosphate is converted
to xylulose 5-phosphate (X5P) through a
consecutive action of the endogenous enzymes
glucose-6-phosphate
dehydrogenase,
6phosphogluconolactonase,
6-phosphoglu
conate dehydrogenase, and D-ribulose-5-phosphate 3-epimerase, whereby the first and the
third reaction yield one NADPH molecule
each (for a review of the PPP, see Stincone
et al. 2015). X5P is then converted to acetylphosphate (AcP) and glyceraldehyde-3-phosphate (GAP) by a heterologous phosphoketolase (PK). Subsequently, AcP is directly
converted to AcCoA by a heterologous phosphotransacetylase (PTA). The PK/PTA pathway was expressed in a strain engineered to
produce FAEE and indeed led to a significant
increase of product titers compared to the reference strain containing only the native PDH
bypass (Jong et al. 2014).
Another variant to convert PKA-derived AcP to AcCoA
involves the acetate kinase (ACK), which yields acetate
and ATP by transferring the phosphate residue from
AcP to ADP. Although ATP is produced in this reaction, the production of AcCoA via this route is energetically less favorable than via PTA since two ATP have to
be invested into the subsequent activation of acetate by
ACS (see above). The implementation of this strategy
led to a significant improvement of polyhydroxybutyrate (Kocharin et al. 2013) and, to a lower extent than
with the PK/PTA variant, FAEE production (Jong et al.
2014).
In one recent study (Meadows et al. 2016),
the PK/PTA pathway was combined with the AALD to produce the isoprenoid farnesene, as
the combinatorial configuration was calculated
to lead to the highest possible product yields
(van Rossum et al. 2016b). Interestingly, the
authors found that AcP produced by PK is
partly dephosphorylated by endogenous (promiscuous) glycerol-3-phosphate phosphatases
and their activity needs to be reduced to favor
the PTA reaction. This study stands out as the
only one in which the endogenous PDH bypass
was fully substituted (by ald6 acs1 acs2 deletion) with heterologous AcCoA synthesis routes
344
L. Baumann et al.