genase, yielding butyryl-CoA. In this cycle,
NADH or NADPH can serve as electron donors
for the reductive steps (depending on the
choice of heterologous enzymes). By iterating
this reaction sequence, the acyl chain can be
elongated by two carbon atoms per cycle.
Depending on the desired chain length and
functional groups of the products, the reverse
b-oxidation can be terminated through different enzymes. For instance, TE can release free
FA from the CoA-bound form, and aldehyde/
alcohol dehydrogenases can reduce the acylCoA esters to n-alcohols (see Sect. V).
Elongation beyond a length of four carbon
atoms requires multiple turns of reverse b-oxidation, which is hampered by a competition
between the thiolase and the termination
enzymes. Thus, although the synthesis of
minor amounts of longer-chain FA via reverse
b-oxidation could be demonstrated at the
proof-of-concept level in E. coli, the product
yields gradually decreased with each iteration
of the cycle (Dellomonaco et al. 2011; Clomburg
et al. 2012). Follow-up studies therefore focused
on improving the selectivity of reverse b-oxidation for longer-chain products. To this end, the
core pathway was expanded by thiolases that
accept acyl-CoA intermediates with various
chain lengths, such as the b-ketothiolase BktB,
which proved suitable to produce C6–C10 compounds in multiple studies (Dekishima et al.
2011; Clomburg et al. 2015; Kim et al. 2015).
The chain length specificity of termination
enzymes (i.e., TE for free FA) was shown to be
the second key determinant for selective production of longer acyl chains (Clomburg et al.
2015; Kim and Gonzalez 2018). To prevent a
premature termination of the cycle, the deletion
of endogenous TE genes was critical in E. coli
strains. These observations are likely transferrable to S. cerevisiae. To provide sufficient carbon and reducing equivalents, heterologous
reverse b-oxidation pathways were expressed
in the yeast cytosol. As a proof of concept, nbutanol production, for the synthesis of which
only one functional turn of the cycle is needed,
has been targeted. Using a variety of enzymes
from different organisms and in combination
with different precursor supply pathways (see
above), titers of up to 1 g L
À1 could be reached
(Lian and Zhao 2015; Schadeweg and Boles
2016a, b). Less effort has been dedicated to the
production of medium-chain products in S.
cerevisiae so far, but the feasibility could be
demonstrated by combining reverse b-oxidation with a medium-chain-specific TE,
CpFatB1, thereby enabling the production of
C6–C10 FA, albeit at a low yield (Lian and
Zhao 2015). It is certain that by transferring
the expanded enzyme toolbox developed in E.
coli (Kim and Gonzalez 2018) and fine-tuning
the expression of endogenous TE, many of
which were recently characterized (e.g., Kruis
et al. 2018), the production of short- and
medium-chain FA via reverse b-oxidation will
be further optimized in S. cerevisiae in the
future.
Directly compared, canonical FA biosynthesis and reverse b-oxidation both bear certain
advantages and drawbacks. The most obvious
advantage of FA biosynthesis is a rather strict
chain length control that can be easily manipulated by protein engineering of FAS, as outlined
above, whereas specific production of longer
carbon chains via reverse b-oxidation is still
very challenging. While the FAS functional
modules operate as a perfectly synchronized
machine within one macromolecular complex,
it is difficult to fine-tune the activity of individual reverse b-oxidation enzymes. At the current
stage of technology, reverse b-oxidation can
compete with FA biosynthesis in the product
range of up to ten carbon atoms. On the other
hand, FA biosynthesis is energetically more
expensive than reverse b-oxidation, since each
elongation cycle by FAS requires one MalCoA
that is synthesized at the expense of one ATP
molecule. In contrast, reverse b-oxidation has
the advantage of using only AcCoA for elongation steps, which also makes it insensitive to the
AcCoA/MalCoA ratio, a parameter that has a
significant influence on the chain length control
by FAS and overall cellular fitness. Whereas the
reverse b-oxidation can be designed to use the
easily available NADH, FAS is strictly dependent on the less abundant NADPH. Thus, the
redox cofactor supply of reverse b-oxidation
requires far less interventions into the central
350
L. Baumann et al.
NADH or NADPH can serve as electron donors
for the reductive steps (depending on the
choice of heterologous enzymes). By iterating
this reaction sequence, the acyl chain can be
elongated by two carbon atoms per cycle.
Depending on the desired chain length and
functional groups of the products, the reverse
b-oxidation can be terminated through different enzymes. For instance, TE can release free
FA from the CoA-bound form, and aldehyde/
alcohol dehydrogenases can reduce the acylCoA esters to n-alcohols (see Sect. V).
Elongation beyond a length of four carbon
atoms requires multiple turns of reverse b-oxidation, which is hampered by a competition
between the thiolase and the termination
enzymes. Thus, although the synthesis of
minor amounts of longer-chain FA via reverse
b-oxidation could be demonstrated at the
proof-of-concept level in E. coli, the product
yields gradually decreased with each iteration
of the cycle (Dellomonaco et al. 2011; Clomburg
et al. 2012). Follow-up studies therefore focused
on improving the selectivity of reverse b-oxidation for longer-chain products. To this end, the
core pathway was expanded by thiolases that
accept acyl-CoA intermediates with various
chain lengths, such as the b-ketothiolase BktB,
which proved suitable to produce C6–C10 compounds in multiple studies (Dekishima et al.
2011; Clomburg et al. 2015; Kim et al. 2015).
The chain length specificity of termination
enzymes (i.e., TE for free FA) was shown to be
the second key determinant for selective production of longer acyl chains (Clomburg et al.
2015; Kim and Gonzalez 2018). To prevent a
premature termination of the cycle, the deletion
of endogenous TE genes was critical in E. coli
strains. These observations are likely transferrable to S. cerevisiae. To provide sufficient carbon and reducing equivalents, heterologous
reverse b-oxidation pathways were expressed
in the yeast cytosol. As a proof of concept, nbutanol production, for the synthesis of which
only one functional turn of the cycle is needed,
has been targeted. Using a variety of enzymes
from different organisms and in combination
with different precursor supply pathways (see
above), titers of up to 1 g L
À1 could be reached
(Lian and Zhao 2015; Schadeweg and Boles
2016a, b). Less effort has been dedicated to the
production of medium-chain products in S.
cerevisiae so far, but the feasibility could be
demonstrated by combining reverse b-oxidation with a medium-chain-specific TE,
CpFatB1, thereby enabling the production of
C6–C10 FA, albeit at a low yield (Lian and
Zhao 2015). It is certain that by transferring
the expanded enzyme toolbox developed in E.
coli (Kim and Gonzalez 2018) and fine-tuning
the expression of endogenous TE, many of
which were recently characterized (e.g., Kruis
et al. 2018), the production of short- and
medium-chain FA via reverse b-oxidation will
be further optimized in S. cerevisiae in the
future.
Directly compared, canonical FA biosynthesis and reverse b-oxidation both bear certain
advantages and drawbacks. The most obvious
advantage of FA biosynthesis is a rather strict
chain length control that can be easily manipulated by protein engineering of FAS, as outlined
above, whereas specific production of longer
carbon chains via reverse b-oxidation is still
very challenging. While the FAS functional
modules operate as a perfectly synchronized
machine within one macromolecular complex,
it is difficult to fine-tune the activity of individual reverse b-oxidation enzymes. At the current
stage of technology, reverse b-oxidation can
compete with FA biosynthesis in the product
range of up to ten carbon atoms. On the other
hand, FA biosynthesis is energetically more
expensive than reverse b-oxidation, since each
elongation cycle by FAS requires one MalCoA
that is synthesized at the expense of one ATP
molecule. In contrast, reverse b-oxidation has
the advantage of using only AcCoA for elongation steps, which also makes it insensitive to the
AcCoA/MalCoA ratio, a parameter that has a
significant influence on the chain length control
by FAS and overall cellular fitness. Whereas the
reverse b-oxidation can be designed to use the
easily available NADH, FAS is strictly dependent on the less abundant NADPH. Thus, the
redox cofactor supply of reverse b-oxidation
requires far less interventions into the central
350
L. Baumann et al.