Accepted Article

Title: Fatty acid biosynthesis: Chain length regulation and control

Authors: Martin Grininger, Christina S. Heil, S. Sophia Wehrheim, and

Karthik S. Paithankar

This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.

To be cited as: ChemBioChem 10.1002/cbic.201800809

Link to VoR: [Link]

A Journal of

[Link]
ChemBioChem 10.1002/cbic.201800809

Fatty acid biosynthesis: Chain length regulation and control

Christina S. Heil, S. Sophia Wehrheim, Karthik S. Paithankar, and Martin

Grininger*

Author affiliation
Institute of Organic Chemistry and Chemical Biology, Buchmann Institute for Molecular Life

Accepted Manuscript
Sciences, Goethe University Frankfurt, Max-von-Laue-Str. 15, 60438 Frankfurt am Main,
Germany.

Correspondence
*grininger@[Link]

Abstract

De novo biosynthesis of fatty acids undergoes an iterative process with strict regulation of the
length of the produced fatty acids. In this review, we focus on the factors determining the chain length
in fatty acid biosynthesis. In a nutshell, the process of chain length regulation can be understood as a
competition between the output of a chain elongating C-C bond forming reaction and a terminating
fatty acid release function. At the end of each cycle in the iterative process, the synthesizing enzymes
need to decide whether the growing chain is to be elongated through another cycle or released as
mature fatty acid. Recent research has shed light onto the decision makers of fatty acid chain length
and has also achieved the control of chain length for the production of the technologically interesting
short (C4-C8) and medium (C10-C14) chain length fatty acids.

Keywords: fatty acid metabolism, multidomain enzymes, compartmentalized synthesis, protein

engineering, biocatalysis, biofuels, cerulenin

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

1. Introduction

In order to introduce fatty acid synthesis and its molecular mechanisms of chain length
regulation, a comparison to the well understood synthesis of proteins and nucleic acids is worthwhile.
Protein and nucleic acid synthesis are guided via the mRNA and DNA templates both in the choice of
the monomeric unit and in the determination of the length of the polymer. Fatty acid biosynthesis is
simpler with respect to selection of the monomeric unit, because it does not vary the polymerizing unit.
However, the determination of fatty acid chain length is comparably more challenging, because it is
defined by the intrinsic properties of the protein, the fatty acid synthase (FAS), itself. Understanding
the mechanism of chain length regulation can be seen as the search for the intrinsic molecular ruler,

Accepted Manuscript
which constantly monitors the fatty acid during biosynthesis and decides in favor of product release
when the appropriate chain length is achieved. As a central process in biosynthesis, chain length
[1]
regulation has been in the purview of the field since decades. Chain length regulation was posited to
occur by the competitive action of a chain elongating condensation and the chain terminating release
reaction, and experimental data support this concept. Studies in the recent years continued to
discover further details on fatty acid synthesis and contributed with high-resolution structures of key
enzymes, so that the understanding and knowledge on the mechanism of chain length regulation have
[2]
also improved.

1.1. Basic principles of fatty acid biosynthesis

Most commonly, FASs release palmitic acid (C16) and stearic acid (C18), or the coenzyme A
(CoA) derivatives thereof, with the typical amphiphilic properties originating from the hydrophobic tail
and the hydrophilic head group. Such properties allow fatty acids to act as building blocks of
membranes (e.g. as part of the phospholipids), serve as covalently tethered membrane anchors of
proteins, be packed tightly into storage fat (as part of the triacylglycerides), or be processed to
[3]
signaling molecules (e.g. eicosanoids). FASs are divided into two systems based on their protein
architecture. Bacteria and plants perform fatty acid biosynthesis with separate, monofunctional
[2a, 4]
enzymes in a dissociated fatty acid biosynthesis system, which is referred to as type II FAS. In
eukaryotes and distinct bacteria of the subgroup of Corynebacterineae (Corynebacteria, Mycobacteria
and Nocardia (CMN-bacteria)) fatty acid biosynthesis occurs in large multifunctional FASs, termed
[2b, 3b, 5]
type I FASs, that integrate all the required enzymatic activities into multidomain complexes. De
[6]
novo fatty acid biosynthesis is also found in mitochondria resembling the bacterial type II FAS.
Although FASs occur in different kingdoms of life and different structural frames, the chemistry of fatty
acid biosynthesis is highly conserved. A fatty acid is always synthesized by repetitive reactions, in
which the initial priming step is performed with an acetyl moiety, which is then successively elongated
with C2 units, that originate from a malonyl moiety, until an acyl chain of the desired chain length is
received (Fig. 1). The C-C bond formation proceeds in a decarboxylative Claisen condensation
reaction catalyzed by the condensing ketosynthase (KS) to yield a β-ketoacyl intermediate bound to
the acyl carrier protein (ACP). After condensation, the β-ketoacyl intermediate is modified through

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

sequential processing by a NADPH-dependent ketoreductase (KR), a dehydratase (DH) and another

NADPH-dependent enoylreductase (ER) to eventually yield the fully saturated acyl intermediate. The
acetyl and malonyl moieties are provided as CoA-esters and loaded by specific acyltransferases
(ATs). The largest variation in fatty acid biosynthesis is found in the termination of the repeating
elongation. Type I systems of fungal and CMN-bacterial FASs terminate synthesis by the AT-mediated
transfer of the acyl moiety from ACP to CoA. Type I animal FASs (including human FAS) terminate
synthesis with a thioesterase (TE) releasing the free acid as product. Type II systems in bacteria and
[7]
mitochondria terminate synthesis by forwarding acyl-ACP to lipid biosynthetic pathways as well as
[8]
lipoic acid synthesis and electron transport chain (ETC) complex assembly , respectively. Type II
[9]
synthesis in plants is terminated by TE-mediated hydrolysis.

Accepted Manuscript
Figure 1: General cycle of fatty acid biosynthesis. Multiple catalytic reactions are performed during chain
elongation, which are conserved between species. The growing chain is first extended by a two carbon (C2) unit
and then processed in a cascade of three reactions (reduction – dehydration – reduction) that blur the traces of
the actual C-C bond forming condensation reaction resulting in a saturated aliphatic chain. The extended acyl
chain can be further elongated by running through the same set of reactions until the desired length is received.
1
R refers to the growing alkyl chain, which in the first cycle represents CH3 and in the other cycles a multiple
thereof (counted in C2H5 steps).

1.2. ACP shuttling

All substrates and intermediates accepted by the FAS and occurring during synthesis are
[10]
shuttled by the ACP that is an inherent protein of FAS type II or domain of type I systems. ACPs are
small, monomeric, helical bundle proteins, and post-translationally modified by a phosphopantetheine
moiety at the active serine (Fig. 2A). ACPs accept a broad range of acyl moieties and even unnatural
cargos, which are all covalently attached upon thioester formation to the phosphopantetheine
prosthetic group.

After getting loaded by ATs, the ACPs successively carry the cargo from one
enzyme/enzymatic domain to another during the iterative biosynthetic cycles, and thus play a central
role in fatty acid biosynthesis. It is therefore clear that the interactions of ACP with the constituent FAS
enzymes/enzymatic domains are critical for the FAS function. Each of these interactions must show

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

appropriate specificity to allow for both loading and unloading the substrates, but at the same time
[11]
enable a rapid turnover.

In type I FASs, the ACPs are tethered by flexible linkers to the multienzymes, while in type II
FASs the ACPs are discrete proteins and diffuse through the cytosol to transport the growing fatty acid
chain to the reaction partners. In order to protect their cargo from thioester hydrolysis, side-reactions
or premature product release, ACPs of some FAS systems sequester the reactive intermediates within
[12]
the hydrophobic core of the fold. This substrate sequestration primarily occurs in type II systems
[13]
and has not been observed for type I systems. Crystallographic studies on E. coli ACP showed that
[14]
the protein can accommodate the growing acyl chain in an expandable hydrophobic pocket. To

Accepted Manuscript
deliver its cargo to appropriate reaction partners at the proper time, the sequestered intermediates
must be released upon protein-protein interaction. The underlying mechanism of chain translocation
[15]
has been termed chain-flipping (formerly also termed switchblade mechanism) (Fig. 2B). A first
structural glimpse of this chain-flipping mechanism was captured for a crosslinked protein complex of
[16]
Escherichia coli ACP and the corresponding DH FabA (PDB ID: 4KEH). Another structure of E. coli
ACP bound to the reductase FabI was solved with help of molecular dynamic simulations (PDB ID:
[17]
2FHS). It is assumed that ACP may communicate its acylation status through cargo-mediated
conformational changes to prevent unproductive interactions or to trigger downstream catalytic steps
[15a]
and enhance productive interactions.
Sequestration may also play a role in chain length regulation and should therefore also be
considered in engineering approaches on the modulation of chain length. NMR studies on a spinach
ACP support this assumption, showing that the ACP was able to sequester an acyl chain up to C10,
[12f]
but not the full-length product of C18. The hydrophobic cleft may act as molecular ruler,
discriminating between chain lengths. The final products are exposed to the cytosolic surrounding and,
as such, more likely to be hydrolyzed. Several reviews are available, which discuss the ACP and its
[2a, 10a, 18]
protein-protein interactions in more detail.

Figure 2: ACP fold and chain flipping. A) Superposition of E. coli ACP (pink) with covalently bound decanoyl-chain
[14] [13a]
(PDB ID: 2FAE) and ACP of baker’s yeast type I FAS (grey) in its apo-form (PDB ID: 2ML8) . The latter is
comprised of the canonical domain and an additional helical fold involved in docking (termed structural domain).
The star indicates the active serine, which is post-translationally modified and carries the acyl load. B) Bacterial
FAS type II ACP sequesters the acyl-chain in a central cavity during transfer between active sites. Upon docking
of the ACP to the enzyme/enzymatic domain, the acyl chain sleeves into the hydrophobic pocket for turnover at
the active site. E. coli ACP is shown in surface representation (PDB ID: 2FAE). The phosphopantetheine in the
“released”-form (without bound acyl chain) has been modeled from Rhodopseudomonas palustris ACP structural
[19]
data (PDB ID: 2LL8) .

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

1.3. The structure of type I FAS

Type I FASs occur in essentially two structural frames emerging from separate evolutionary
[5, 20]
developments. Whereas the fungal and CMN-bacterial FASs form about 2.6 MDa large barrel-
shaped complexes with enclosed reaction chambers, the animal FASs are dimers of multidomain
polypeptides in an overall open structural arrangement. An exhaustive review on the three-
[2b]
dimensional structures of type I FASs was recently published elsewhere.

There is clear scientific consensus that multifunctional type I proteins evolved from the type II of
[3c, 20b, 21]
separate proteins. During the course of evolution, two different multifunctional FAS

Accepted Manuscript
architectures emerged; the homodimeric animal FASs and the oligomeric fungal/CMN-bacterial FASs,
of which the fungal/CMN-bacterial fold resembles the more elaborate, higher developed fold.
Multifunctional architectures provide an advanced regulation. The close proximity of the domains
improves enzymatic efficiency, minimizes unspecific interactions and shelters putatively reactive
[22]
intermediates in reaction chambers of clefts.
It has been suggested that duplication of primordial genes and successive gene fusion at an early
evolutionary state, prior to the divergence of prokaryotes and eukaryotes, gave rise to the
[3c, 21]
multifunctional type I FASs. Continuous integration of domains led to a single gene encoding a
complete FAS as evolutionary stable system. This gene fusion hypothesis is supported by evidence of
FAS-like gene clusters, operons or partially fused genes in present-day organisms, e.g. the E. coli fab
[23]
gene cluster encoding a type II FAS , the Rhizobium meliloti fix-23 gene cluster encoding a FAS-like
[24]
polypeptide , and the Mycobacterium tuberculosis mycocerosic acid synthase, which is encoded by
[25]
a single gene and shares significant sequence and structural similarity with animal FAS type I. All
three examples show an order of clustered or fused genes similar to the domain order found within
animal FAS. A remarkable sequence and functional similarity can also be found for animal FAS and
[2b, 26]
polyketide synthases (PKSs). A fully methylating iterative PKS was identified as common
evolutionary ancestor, underlined by the non-functional evolutionary relict of a methyltransferase (MT)
[20b, 27]
domain in the animal FAS scaffold, which only fulfills structural tasks.
Fungal and CMN-bacterial FASs show a domain arrangement different from that of animal FASs.
Phylogenetic and structural characterization identified bacterial monofunctional proteins from fatty acid
and polyketide metabolism as ancestors. Interestingly, these proteins had already some of the
scaffolding elements implemented that are now involved in stabilizing the intricate barrel-shaped
complex. Domains were integrated until a single-gene multidomain precursor occurred, that resembles
CMN-bacterial-like FASs. Further scaffolding mainly occurred in the fungal protein, likely driven by
advantages in catalytic efficiency originating from an increased rigidity of the barrel-shaped structure.
The last fungal common ancestor was encoded on one single gene. At a later evolutionary state
splitting in two genes at various sites occurred and led to a diversified fungal FAS system with gene
[2b, 20a]
topological variants.
Alternative scenarios hypothesize that multifunctional FASs first emerged in eukaryotes as products of
intronic recombination and at later stage spread into prokaryotes by horizontal gene transfer. Although
this alternative evolutionary route cannot be entirely discounted, the gene fusion hypothesis has
[2b, 3c, 20a, 21, 28]
become the dominating doctrine as of today.

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

FAS from baker’s yeast is the best characterized fungal FAS with initial low resolution
[29]
structures solved by electron microscopy (EM) several decades ago. Crystals of yeast FAS were
[30]
grown as early as 1969. However, the detailed insights into the structure of yeast and other fungal
FASs took about four decades, when in 2006 X-ray crystallographic structures of Thermomyces
[15b, 31]
lanuginosus FAS (3.1 Å) and baker’s yeast FAS (3.1–4 Å) were published. Fungal FASs are
comprised of an intertwined rigid wheel-like structure that forms the central platform for two
conformationally less rigid domes at either side. Remarkably, the fungal FAS utilizes about 50% of the
[5]
total polypeptide length into non-catalytic scaffolding elements that support the elaborate structure.
Shortly thereafter, a first complete model of yeast FAS was published, that included the
phosphopantetheine transferase (PPT) domain at the perimeter of fungal FAS, an integral domain in

Accepted Manuscript
[32]
fungal FAS type I catalyzing the phosphopantetheinylation of the ACP domain (Fig. 3A). Cryo-EM
reconstruction at 5.9 Å resolution allowed the direct localization of the ACP domain during the catalytic
[33]
cycle. The baker’s yeast FAS was recently used as a model protein for studying protein
denaturation at water air interfaces, as occurring during cryo-EM sample preparation. An optimized
[34]
sample preparation protocol delivered a 4.0 Å map from 28,000 particles. The molecular
mechanism of ACP-mediated substrate shuttling has been analyzed in further detail by molecular
[35]
dynamic and coarse-grained molecular modeling. Fungal FASs vary slightly in their polypeptide
topology. Next to the baker’s yeast-type α6β6 heterododecameric FAS, also homohexameric variants
[20a, 28, 36]
of fused α- and β-chains exist. All of these complexes share the high structural similarity of a
[37]
barrel-shaped overall fold, and assembly via a highly conserved pathway. The evolutionarily related
CMN-bacterial FASs are encoded on a single gene and show a similar barrel-shaped homohexameric
fold as fungal FASs, but with less non-catalytic scaffolding elements (only 35% of the total polypeptide
[28]
length) and without C-terminally attached PPT domains (Fig. 3B).

Figure 3: Structure and domain organization of the microbial type I FAS. A) Fungal FAS type I. X-ray
[32]
crystallographic model of baker’s yeast FAS type I (PDB ID: 3HMJ) . The α- and β-chains are colored as
indicated in the domain overview. The PPT domain, which is part of fungal FAS, is attached at the perimeter of
the α-wheel structure. Different chains in the oligomeric assembly are indicated by primes. Slices of the barrel, in
orientation as indicated, show the inner volume of the barrel-shaped fold, which harbor the reaction chambers of
fatty acid synthesis. The ACP is stalled at the KS under the crystallization conditions. B) Bacterial FAS type I. X-
[38]
ray crystallographic model of Mycobacterium smegmatis FAS type I (PDB ID: 4V8L) . Bacterial FAS type I is
evolutionarily related to fungal FAS type I, and can be seen as minimal fungal FAS type I fold missing the PPT

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[28]
domains, subdomains and insertion elements. The ACP domain could not be traced in electron density and is
omitted from the model.

[3c]
The animal FAS comprises two multifunctional polypeptide chains each of about 270 kDa.
Owing to its open fold and conformational variability, structural characterization took considerably
[39]
longer and was initially driven by mutational analysis. The cryo-EM study on human FAS was an
[40]
initial milestone in providing a low-resolution map. The breakthrough came in 2006 when the
structure of porcine FAS was determined by X-ray crystallography to 4.5 Å and subsequently to 3.2 Å
[27, 41]
resolution. In both of these structures, the loosely tethered ACP and TE domains could not be

Accepted Manuscript
traced in electron density. However, these domains were previously structurally characterized as
[13b, 42]
isolated proteins. The X-shaped animal FAS structure is segregated into a condensing and a
processing part that are connected via an intertwined short linker (Fig. 4A). The condensing part is
termed after the condensing KS domain, which forms the central dimeric unit. The KS domain is
connected to the malonyl/acetyl transferase (MAT) via the linker domain LD. The processing wing
comprises the domains KR, DH and ER for the processing of the β-carbon. The two wings do not
interact permanently via interfaces, but undergo large relative positional changes, as initially observed
[43]
in a negative stain EM and recently shown by high-speed atomic force microscopy (Fig. 4B). X-ray
structures have also been determined for the condensing wings of human FAS and recently for murine
[44]
FAS with bound substrates. Computational molecular docking studies on a human FAS model
analyzed the type I animal FAS substrate shuttling and shed light on the overall conformational
[45]
dynamics of this protein.

Figure 4: Structure and domain organization of animal type I FAS. A) X-ray crystallographic model of porcine FAS
[27]
type I (PDB ID: 2VZ8 ). Each chain is colored as indicated in the domain overview. The TE and ACP domains
are flexibly attached, and could not be traced in electron density. Nonfunctional domains that fulfill structural tasks
are indicated in italic letters in the domain overview. B) Surface representation of a full-length FAS model with
[42] [13b]
domains TE and ACP from PDB entry codes 1XKT and 2PNG , respectively. Condensing and processing
wings have been moved apart by about 30 Å to better indicate the inherent conformational variability of the
[43]
protein.

1.4. The structure of type II FAS

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Type II fatty acid biosynthesis is present in bacteria, mitochondria and plants, and is in the following
briefly outlined for E. coli. E. coli type II FAS produces fatty acids in an iterative cycle via the
consecutive action of 13 discrete proteins (Fig. 5). Although the chemical reactions of type II FASs are
similar to that of the multienzyme type I system, there are some marked differences. Most importantly,
type II FAS systems are capable of producing a diversity of products for cellular metabolism, e.g.
different chain lengths of saturated as well as unsaturated fatty acids, iso- and anteisobranched fatty
[4]
acids and even hydroxyl fatty acids.
The E. coli type II FAS cycle can be divided into an initiation and an elongation module. The initiation
module comprises the enzymes ACP synthase (AcpS), CoA carboxylase (ACC), malonyl-CoA:ACP
transacylase (FabD) and β-ketoacyl-ACP synthase III (FabH). Activation of the ACP is catalyzed by

Accepted Manuscript
AcpS, which acts as the PPT and transfers the phosphopantetheine moiety from the CoA onto the
2+
ACP. The reverse reaction is catalyzed by the Mn -dependent phosphodiesterase (AcpH), which
hydrolyzes the phosphopantetheine moiety from the ACP. Due to rapid turnover of the AcpS catalyzed
reaction, ACP is almost exclusively detected in its active form. ACC catalyzes the carboxylation of
acetyl-CoA to malonyl-CoA, which is subsequently loaded onto ACP by FabD, an AT that is highly
specific for malonyl-CoA and continuously supplies malonyl-ACP as elongation substrate for the FAS
cycle. As last enzyme of the initiation module, FabH catalyzes the initial step in acyl chain formation
[46]
solely being responsible for the first condensation reaction of acetyl-CoA and malonyl-ACP.
Next to FabH, two more KS enzymes are present in the type II FAS cycle performing the condensation
reactions in the elongation module, β-ketoacyl-ACP synthase I and β-ketoacyl-ACP synthase II (FabB
[47]
and FabF, respectively). While FabH is highly selective for acetyl-CoA, FabB and FabF are
promiscuous for acyl chains from C4 to C14, and FabF even accepts C16. In the elongation cycle
ketoacyl-ACP is first successively processed by the β-ketoacyl-ACP reductase (FabG), β-hydroxyacyl-
ACP dehydratases (FabA or FabZ) and enoyl-ACP reductase (FabI) before subsequent rounds of
elongation are initiated by the condensing enzymes FabB or FabF.
Unsaturated fatty acids, found in the product spectrum of E. coli type II FAS, originate from the unique
property of FabA, which not only possesses a DH function, but is also able to isomerize trans-2-
[4]
decenoyl-ACP to cis-3-decenoyl-ACP, which can subsequently be elongated by FabB or FabF. The
condensing enzymes from the elongation module (FabB and FabF) show differential acceptances for
unsaturated substrates; i.e. FabB is elongating the cis-3-decenoyl-ACP three times, while FabF
[2a, 4, 48]
condenses the cis-7-hexadecenoyl-ACP up to cis-11-octadecenoyl-ACP. The final product of
bacterial fatty acid biosynthesis is an ACP-tethered acyl moiety that is subsequently harnessed by
[9]
phosphate acyltransferases either for acyl-phosphate or lysophosphatidic acid (LPA) production.
Also intermediates of the FAS type II cycle are used in the synthesis of key cellular constituents, e.g.
for synthesis of lipoic acid or quorum-sensing molecules, by diverting the ACP intermediates into other
[4]
biosynthetic pathways. A more comprehensive description of type II synthesis can be found
[2a, 4]
elsewhere.

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Accepted Manuscript
Figure 5: Bacterial fatty acid cycle (E. coli). An ensemble of 13 discrete enzymes catalyzes the fatty acid
biosynthesis in E. coli. Three different KS related enzymes with distinguishable substrate specificities are involved
in the fatty acid cycle. A portion of unsaturated fatty acids occurs upon isomerization of β-hydroxyacyl
intermediates by a side reaction of the designated DH domain FabA, which can be further elongated by the
designated KSIII FabH. TEs are not directly involved in the fatty acid cycle of E. coli. The acyl-ACP products are
rather directly transported in the lipid biosynthetic pathway by PlsX/PlsY and PlsB.

Current understanding of mitochondrial fatty acid biosynthesis is as follows (Fig. 6).

[49]
Mitochondrial FASs act as a bacterial-like type II systems composed of free-standing and
monofunctional proteins. Overall fewer proteins are involved in synthesis as compared to bacteria; e.g.
a single KS is responsible for all the condensation steps in mitochondria. The following description
[6a]
refers to mitochondrial fatty acid synthesis in humans and Saccharomyces cerevisiae. While
transfer of malonyl moieties to the ACPs is catalyzed by the specific malonyl-CoA:ACP transacylase
[50]
(Mct1 (yeast), MCAT (human)) , no mitochondrial acetyl-transferase has been described yet, but
acetyl-ACP may be generated by transfer of acetylated phosphopantetheine from CoA to ACP via the
[51]
PPT. A single condensing enzyme, the mitochondrial ketoacyl-synthase (mtKAS), specifically
[51b, 52]
condenses acyl-ACPs (CEM1 (yeast), OXSM (human)) . Within three processing steps, catalyzed
[50a, 53]
by the 3-ketoacyl reductase (KAR; OAR1 (yeast), KAR1 (HSD17B8 and CBR4 (human)) , the 3-
[54]
hydroxyacyl thioester dehydratase (HTD; HTD2 (yeast), HsHTD2 (human)) and the enoyl thioester
[55]
reductase (ETR; ETR1 (yeast), MECR (human)) , a saturated acyl chain is obtained. The co-
existence of the mitochondrial FAS with the cytoplasmic FAS type I system raises the question to the
benefit of preserving mitochondrial synthesis and its detailed function. Deletion of mitochondrial FAS is
[51b, 56]
lethal, which states the importance of this pathway for the function of mitochondria. Studies on
mitochondrial FAS report a bimodal chain length spectrum of fatty acids; i.e. C8 and longer fatty acids
[51b, 57]
such as C16 and C18. While the short chain C8 fatty acids are the precursors of lipoic acid
[58] [59]
synthesis , the fate of long fatty acids produced in mitochondria is just poorly understood.
Recently, it was shown that mitochondrial ACPs carrying long acyl chains are accessory subunits of
[8, 60]
the electron transport chain (ETC) complex I.

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Accepted Manuscript
Figure 6: Mitochondrial fatty acid cycle. Mitochondrial fatty acid biosynthesis is performed by a type II FAS system
of dissociated catalytic enzymes. Different from E. coli, mitochondria utilize only a single KS enzyme (KAS). The
iterative mechanism of chain elongation is closely related to that of type I systems. The mitochondrial system
displays a bimodal product spectrum dominated by C8 and C16-C18 chains. Octanoyl-ACP (C8) is an important
precursor for lipoic acid synthesis, while long chain products are involved in electron transfer chain (ETC)
complex assembly.

In plants, fatty acids are predominantly synthesized in the plastids. Plants harbor a type II FAS system
[61]
composed of separated enzymes. Similar to E. coli, three different KSs with varying substrate
[62]
specificities are present. The products of fatty acid synthesis in plants are hydrolyzed by chain
[63]
length specific TEs, FatA and FatB, which determine the chain length of the released products. In
most plants, the major product of fatty acid synthesis is oleic acid (C18:1), but in some seed oils a high
amount of short and medium chain fatty acids (C8-C14) is found.

2. KS, AT and TE domains/enzymes – key players in chain length control

In order to measure fatty acid chain length during fatty acid synthesis, two principal mechanisms can
be envisioned. On the one hand, the condensing KS provides limited space for the growing chain, so
that the elongation is impaired when the chain exceeds this space. On the other hand, the extending
hydrophobic tail of the growing chain induces affinity for the terminating transferase or hydrolase
function that is subsequently releasing the acyl chain as CoA ester or free fatty acid. In 1969, Sumper,
Oesterhelt and Lynen have precisely postulated the interplay of those functions as being responsible
for chain length regulation in yeast FAS. They posited that at long chain lengths the KS domain acts
as a repressor of the chain extension reaction and the off-loading AT domain as an inducer of the
cleavage reaction. At a time when cloning was not yet a tool of molecular biology, they proved their
hypothesis by modulating the yeast FAS product chain length as a function of the malonyl to acetyl
[1]
relative concentration. The structural information collected during the last few decades on the
[2b, 5]
condensing and terminating enzymatic functions of type I and type II FAS provides new basic

10

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

understanding of the molecular mechanisms of chain length control and refines the initial concept by
Sumper et al.. This information is summarized in the following.

2.1. KS domain/enzyme: The condensing KS is responsible for the formation of the C-C bond,
and thus for the extension of the growing chain. The KS follows a ping-pong mechanism: In the ping
step, an acyl moiety, usually delivered by an ACP, is loaded into the KS active site, where it binds
covalently to the active site cysteine. The actual C-C bond-formation occurs in the pong step, where
malonyl, again shuttled by ACP, decarboxylates to yield the enolate nucleophile that condenses with
[31c, 62, 64]
the cysteine-bound acyl electrophile to form β-ketoacyl-ACP. It should be noted that the
sequence of events in the decarboxylative Claisen condensation and mechanistic details of KS-

Accepted Manuscript
mediated condensation are still a matter of debate. The overall consensus regarding the mechanism
for the condensation favors the sequential reaction of the initial decarboxylation to the enolate that
subsequently attacks the acylated cysteine to generate the product under C-C bond formation (Fig. 7,
mechanism 1). In support of the sequential reaction, decarboxylation and C-C bond formation can be
uncoupled, monitored by acetyl production from malonyl, when mutating an active site histidine in
[62]
animal type I FAS to omit its stabilizing role for the malonyl moiety and the enolate. A similar
[65]
uncoupling was observed in baker’s yeast type I FAS, when the active site cysteine is blocked.
[66]
Another well-studied KS enzyme is the M. tuberculosis KasA, because it is a relevant drug target.
[67] [68] [69]
KasA has been analyzed in structure and function , and recently also by computational means.
Also for the KasA-mediated condensation, decarboxylation of malonyl-ACP is considered to precede
C-C bond formation. Calculated protonation states of key amino acids at the active site imply the
release of CO2 during condensation, while experimental data on animal FAS indicated release of
[62]
bicarbonate.

Two alternative routes have been reported for the decarboxylative Claisen condensation, i.e.
[64]
decarboxylation and C-C bond formation occurring in a concerted manner (Fig. 7, mechanism 2),
[70]
and stepwise reaction of an initial enolization/bond-formation followed by decarboxylation (Fig. 7,
mechanism 3). A biomimetic amine base-catalyzed decarboxylative aldol reaction has been recently
developed, and mechanistic studies established the reversible formation of a
[71]
postaddition/predecarboxylation intermediate following mechanism 3.

11

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Accepted Manuscript

Figure 7: Proposed mechanistic pathways for the decarboxylative Claisen condensation. The active site histidines
[62]
have been labeled according to rat FAS type I. The ping step is well understood, except that the molecular
[72]
foundation of cysteine deprotonation is still controversial. For the pong step, several mechanisms are
discussed (outlined in 1A, 1B, 2 & 3). We note that other KS enzymes and domains have been studied, e.g. M.
[69]
tuberculosis KasA, in a computational approach. The protonation state of His 331 is under debate, similar as
the release of CO2 or bicarbonate (mechanism 1A vs. 1B). Mechanism 1B involves a tetrahedral transition state
[62]
that collapses under bicarbonate release, which is not shown here.

12

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Structural evidence indicates the KS-mediated regulation of chain length by limited space of the
binding channel. Numerous X-ray structures of KSs from type II FAS systems have been determined
with substrates or inhibitors bound, allowing the comparison of free vs. loaded states. Such 3D-
[73]
structures are for example available for E. coli enzymes FabH, FabB and FabF , and human
[57, 74]
mtKAS. In line with its function, E. coli FabH accepts acetyl-CoA, while being less active on
[73a, 75]
substrates longer than C3 as well as on ACP bound acyl-moieties. In the other KS enzymes, a
hydrophobic acyl chain binding pocket extends towards the dimer interface. In E. coli FabF, the distal
[73e]

Accepted Manuscript
ends of the dodecanoyl chains in the dimer are 12 Å apart (PDB ID: 2GFY) . From a comparison of
free and loaded states, it is evident that the KS binding channel just marginally responds to the
substrate loading. This is particularly true for the FabB binding channel, in which essentially just the
aromatic phenylalanine at the entrance of the binding tunnel rotates upon acyl chain binding (Phe392
in E. coli FabB; e.g. the dodecanoyl (C12) chain in the models PDB IDs: 1EK4 and 2BXY vs. apo-state
[73c, 73d]
in PDB ID: 2BUH). Also in FabF, the aromatic ring of the conserved phenylalanine re-arranges
to accommodate acyl chains (Phe400 in E. coli FabF, equivalent to Phe392 of E. coli FabB). In
contrast to FabB, FabF additionally features a gatekeeper residue Ile108 at half of the distance of the
binding channel, which adopts a different rotamer conformation when binding compounds with acyl
chains longer than C6 (e.g. the dodecanoyl chain in PDB ID: 2GFY, cerulenin in PDB ID: 1B3N vs.
[47b, 73b, 73e]
apo-state in PDB IDs: 2GFW and 1KAS) (Fig. 8A&B). MtKAS shows structural properties
similar to bacterial KS enzymes. As analyzed for mtKAS from human, no significant side-chain re-
arrangements in the binding channel occur upon acyl chain binding, except of Phe447, the equivalent
residue to E. coli FabB Phe392 and E. coli FabF Phe400, changing conformation upon binding. Similar
to FabF, mtKAS features a gatekeeping residue in the binding channel (Met154 for human mtKAS),
[57]
which needs to rotate upon binding of acyl chains longer than C7 (see Fig. 8A).

The gatekeeper residues FabF and mtKAS have long been suggested to be involved in chain
length regulation. The Ile108Phe variant of E. coli FabF, presumably increasing the rotational barrier of
[76]
the gatekeeper, revealed compromised elongation efficiency with substrates longer than C6 .
Further, a low preference of human mtKAS for C8 substrate was suggested to originate from the
energy barrier for the rotational movement of Met154. As such, Met154 was posited to be responsible
[51b]
for the bimodal product spectrum of mitochondrial fatty acid synthesis. Finally, structural
[31c]
information of an inhibitor bound vs. free KS binding channel is available for the type I yeast FAS.
As the inhibitor cerulenin contains a C8 acyl chain, cerulenin-bound states are suited mimics for
studying acyl binding. Structural data on yeast FAS with bound cerulenin revealed Met1251 as
[31c]
gatekeeper in the binding tunnel (Fig. 8C). Changing the rational barrier of Met1251 by decreasing
its conformational variability by a Gly1250Ser mutation and/or altering its spatial claim by a
[77]
Met1251Trp mutation resulted in altered chain length spectra. Overall, data suggests that
gatekeepers can restrict the space provided for the embedment of the acyl chain substrate, when
constrained in rotational variability, and, by this, interfere in chain length regulation.

13

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Accepted Manuscript
Figure 8: Three-dimensional structures of KS domains/enzymes. A) E. coli FabF (upper panel) and human mtKAS
(lower panel) in superposition of loaded and unloaded binding pockets. (Upper panel) Upon binding of long acyl
chains or the inhibitor cerulenin, the gatekeeper Ile108 rotates, in order to accommodate the aliphatic tail. The
[73e]
cerulenin-inhibited X-ray structure in blue (residues in yellow, cerulenin in orange) (PDB ID: 1B3N) is
[47b]
superimposed with an unloaded FabD (grey) (PDB ID: 1KAS) . The cavity is shown for the inhibited/loaded
structure. (Lower panel) For mtKAS, X-ray structures of the hexanoyl-loaded mtKAS (blue, residues in yellow)
[57] [57]
(PDB ID: 2IWZ) and the unloaded mtKAS (grey) (PDB ID: 2IWY) are superimposed. The cavity is shown for
the hexanoyl-bound structure. Met154 does not rotate at this acyl-chain length. Cerulenin (shown in orange),
modeled from the inhibited FabD (PDB ID: 1B3N), indicates the necessary rearrangement of the Met154 at longer
acyl-chain length. The rotational barrier is important for the bimodal product spectrum of mitochondrial fatty acid
synthesis. B) Alignment of KS sequences of selected organisms showing regions involved in the catalytic function
(indicated by stars) and acyl chain accommodation. FAS type I sequences have been selected unless otherwise
indicated. The color code corresponds to the sequence similarity according to Clustal Omega. Residue numbers
are given for E. coli KASII and baker’s yeast FAS I. C) X-ray crystallographic structure of baker’s yeast FAS.
[32]
(Upper panel) Key residues of the cerulenin-inhibited (residues in yellow, cerulenin in green) and the wildtype
[15b]
FAS (residues in grey) in superposition. For orientation, the complete structure and the α-wheel of baker’s
yeast FAS are shown as inset. The yellow spot highlights one of the six KS active sites. (Lower panel) KS binding
channel of baker’s yeast FAS. The KS domain of the inhibited FAS (“open”) is shown in surface representation in
blue with relevant residues in yellow and cerulenin in green. The wildtype FAS (“closed”) is shown in surface
representation in grey. The gatekeeper residue of the inhibited and the wildtype KS as well as cerulenin are
shown in both panels. The binding cavities of the inhibited and the non-inhibited KS domain differ. The
gatekeeper residue Met1251 blocks the inner volume of the binding pocket in the closed state, but swings out in
[78]
the open state. All structural figures were prepared with the program PyMOL.

14

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

It is still debated whether the ping or the pong step of the condensation reaction imposes specificity
and is responsible for chain length regulation. For the FAS-related PKS systems, there is indication of
[79]
chain length regulation occurring during the ping as well as the pong step. This is plausible,
because the restricted conformational variability of the KS binding channel suggests that the initial
binding of acyl-ACP, the reversible transacylation (ping step) and the overall irreversible pong step
impose similar spatial constraints. Since chain length regulation is the collaborative output of the
[1]
actions of the condensing KS and the terminating TE or AT , the molecular ruler function needs to be
imposed during reversible steps of the enzymatic reactions. Therefore, chain length regulation can
easily be embedded during the reversible steps of substrate binding and transacylation. In contrast,

Accepted Manuscript
the overall irreversible C-C bond forming pong step could impose a molecular ruler function just in an
initial reversible step before decarboxylation. Accordingly, the widely accepted pong step mechanism
of decarboxylative enolate formation followed by bond formation (see Fig. 7 mechanisms 1B) includes
reversibility just when assuming that malonyl decarboxylation is strictly constrained by the length of
the acyl moiety bound to cysteine. If not the case, this mechanism would be at danger in wasting
malonyl-CoA and prompting an ACC-mediated carboxylation/KS-mediated decarboxylation futile
cycle. Other pong step mechanisms, i.e. the sequential reaction of an initial enolization/bond-formation
followed by decarboxylation (see Fig. 7 mechanism 2) as well as the concerted decarboxylation and
C-C bond formation reaction (see Fig. 7 mechanism 3), embed reversibility more intuitively.

2.2. AT domain/enzyme: ATs catalyze the transfer of acyl-CoA to ACP by following a ping-pong
mechanism. The AT catalyzed transacylation is different to the related KS mechanism, as this process
is reversible over both steps and therefore able to load and unload CoA and ACP. In AT-mediated
loading, the active serine residue attacks the carbonyl group of the acyl-CoA under formation of an
AT-bound oxoester intermediate and the release of CoA (ping step). In the subsequent pong step, the
thiol group of the phosphopantetheine prosthetic group of ACP accepts the acyl moiety, leading to the
formation and the respective acylated ACP and the regeneration of the AT domain/enzyme. The
nucleophilicity of the active site serine is increased by a conserved histidine. The transition states are
stabilized by an oxyanion hole at the active site, created by two backbone amide groups (Fig. 9A).
FASs essentially use acetyl-CoA and malonyl-CoA as substrates. Just in rare cases ATs accept
propionyl and short branched acyl units for priming, as e.g. for monomethyl-branched fatty acids in
[80]
adipose tissues , and methylmalonyl for elongation, e.g. for the production of branched fatty acid in
[81]
certain glands. It has been shown during several decades of FAS research that the actual substrate
specificity of FAS does not primarily originate from ATs. Particularly ATs of type I FASs have been
characterized as substrate-tolerant proteins that are able to load even functionalized noncanonical
[44b, 82]
CoA-esters. In that light, the role of AT domains and enzymes in chain length regulation is more
subtle than for KSs. ATs act on chain length in loading FAS depending on their substrate specificity
and on cellular acyl-CoA concentrations, whereas particularly the relative concentrations of acetyl- to
[1, 83]
malonyl-CoA influences chain length. Fungal and CMN-bacterial FAS type I feature ATs with an
additional function. The AT malonyl/palmitoyl-transferase (MPT) is responsible for loading the malonyl

15

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

moiety as well as off-loading the mature acyl chain as CoA-ester, and directly acts on chain length
[84]
regulation in this role.

Accepted Manuscript
Figure 9: AT mechanism and structure. A) Mechanistic pathway for AT domain mediated substrate transfer (upper
sequence). The active site residues follow the numbering of murine FAS type I. Hydrolysis occurs at lower
[44b] [44b]
reaction rates (lower sequence). B) Condensing part of murine FAS type I. Domains are in coloring as
introduced in Figure 1, except that the ferredoxin-like subdomain of MAT is shown in brown for clarity. A model of
animal FAS type I is shown as inset. C) Superposition of the four KS-MAT polypeptide chains in the asymmetric
unit of murine FAS X-ray structural data (aligned via the KS domains). The superposition indicates the
conformational variability between KS and MAT domains, but also between α/β hydrolase subdomain and the
ferrdoxin-like fold of MAT. The conformational variability likely accounts for substrate tolerance. D) Binding site of
murine FAS MAT (see grey highlight in A) in complex with malonyl-CoA. E) Malonyl-loaded enzyme states of
murine FAS MAT and E. coli FabD. The hydrogen bond network is shown in dashed lines. Arg606 of murine MAT
is in side-on bidentate coordination with the carboxyl group of serine bound malonyl. The malonyl-specific
[46a]
transferase FabD (PDB ID: 2G2Z) shows end-on bidentate coordination of the guanidinium group of the active
site arginine and the free carboxyl group of serine bound malonyl. All structural figures were prepared with the
[78]
program PyMOL.

Bacterial type II FAS contains a malonyl-specific AT (FabD), but no acetyl-specific AT, since
[85]
FabH directly binds acetyl-CoA and primes fatty acid synthesis. FabD has been shown to be
[46a, 86]
specific for malonyl loading , largely imposed by the active site arginine (Arg117; E. coli FabD

16

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

numbering) that, via the guanidinium moiety, forms a bidentate salt bridge with the free carboxyl group
of the malonyl moiety. The bidentate interaction between arginine and the malonyl moiety is end-on,
[87]
as it has also been found in the malonyl-specific AT of the PKS DynE8. In contrast to bacterial type
II FAS, type I systems feature AT domains that load acetyl- and malonyl-CoA. Animal FASs type I
share a single AT that transfers both the priming acetyl- and the elongating malonyl moiety (termed
[3c, 82b]
MAT). Transfer of both substrates with similar efficiencies is possible, since a conserved active
site arginine (Arg606; murine FAS numbering) complexes the free carboxyl group of the malonyl
moiety without counter-selecting against acetyl-transfer. Mutating Arg606 to residues such as lysine or
alanine compromised binding of malonyl, while leaving the efficiency of acetyl transfer essentially
[44b, 88]
unaffected. A recent X-ray structural analysis of a malonyl:MAT complex and enzyme kinetic

Accepted Manuscript
analysis with the native substrates and multiple noncanonical CoA-esters have characterized MAT of
murine FAS as a polyspecific enzyme that is also able to transfer longer acyl chains, as hexanoyl (C6)
[44b]
and palmitoyl (C16) chains, although the latter with lower efficiencies. Polyspecificity presumably
results from the conformational flexibility of the binding pocket that is embedded between the α/β-
hydrolase subdomain and the ferredoxin-like fold of the MAT domain (Fig. 9B & C). The guanidinium
moiety of the active site arginine (Arg606; murine FAS numbering) holds the free carboxyl group of the
malonyl moiety in side-on coordination, which may also contribute to substrate polyspecificity (Fig. 9D
[44b]
& E). In a recent combined quantum mechanics/molecular mechanics approach, human MAT-
mediated substrate transfer was analyzed mechanistically, and a free energy profile of the ping step
was presented. The results provided insight into the individual roles of active site residues in substrate
[89]
transfer, as e.g. the contribution of Arg606 (murine FAS numbering) for malonyl-specificity. Fungal
and CMN-bacterial FASs harbor two ATs, which are responsible for loading substrates and releasing
[90]
the final products that are typically palmitoyl- (C16) and stearyl-CoA (C18). Priming with the acetyl
[91]
moiety is initiated by acetyl-specific ATs. Intriguingly, in fungal and CMN-bacterial FASs, the
elongating malonyl moiety is loaded by an AT that is also responsible for off-loading the final acyl-CoA
[84]
(termed MPT). For off-loading, the mature acyl chain is transferred in the “reverse” direction from
the ACP to free CoA. The MPT has been characterized as being tolerant in the transfer of acyl
[82a]
moieties of a broad variety in lengths as well as in the transfer of functionalized CoA-esters. In this
property, it resembles the MAT of animal FAS type I.

2.3. TE domain/enzyme: TE domains and enzymes are involved in chain length regulation by
terminating fatty acid synthesis by hydrolyzing acyl-thioesters. Mechanistically, they again follow a
ping-pong reaction sequence (Fig. 10A). TE-mediated regulation of chain length is provided by a
growing affinity for the extending hydrophobic tail of the acyl chain. Additionally, similar to the KS, the
limited space of a TE domain binding channel plays a crucial role in defining the chain length product
spectrum. Multiple structural studies have supported these two principal mechanisms, and molecular
docking studies on the human type I FAS TE domain further elucidated how the TE domain
[92]
discriminates acyl chain lengths. The tight substrate specificity of the TE domain is supported by
biochemical studies. Palmitic acid (C16) is released from the enzyme as major product of fatty acid
biosynthesis with small amounts of also stearic acid (C18), while activities of the TE domain on
[93]
substrates longer than C18 or shorter than C16 are drastically decreased. Deletion of the TE

17

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

domain in human type I FAS leads to a loss in chain length specificity, yielding longer chain lengths of
[94]
C20 and even C22, but also diminishes the enzyme’s ability to release fatty acids.

It is noted that in certain tissues, palmitic acid is not the exclusive product of fatty acid
biosynthesis. For example, in sebaceous glands, like harderian and meibomian glands and avian
preen glands, methyl-branched or odd carbon-number fatty acids can be generated by animal FAS
[3c, 81b, 81c, 95]
using alternative priming and elongating substrates. The main termination products of odd
[81c]
carbon-number fatty acid synthesis are pentadecanoic acid (C15) and margaric acid (C17) , which
align well with the substrate specificity imposed by the TE domain. Main termination products of
methyl-branched fatty acid synthesis are 2,4,6,8-tetramethyldecanoic acid or 2,4,6,8-

Accepted Manuscript
[81b]
methylundecanoic acid, dependent on the priming unit used. The shorter chain lengths most likely
result from the methyl groups that additionally crowd the KS binding channel and induce pre-mature
[3c]
termination of synthesis. A broad product spectrum of C4 to C18 fatty acids is found in lactating
[96]
mammary glands. The origin for the pre-mature termination of fatty acids is different in ruminant and
non-ruminant animals. In ruminants, the pre-mature release of medium chain length acyl moieties is
achieved by an extended substrate specificity of the AT domain, which can unload C6-, C8-, and C10-
[97]
CoA from the type I FAS. In non-ruminant mammals, pre-mature chain termination is mediated by a
discrete monofunctional TE type II. Here, TE type II is able to access the growing fatty acid chain
[98]
bound to ACP during catalysis, and hydrolyzes lauric acid (C12) and myristic acid (C14). The
[98a, 99]
function of TE type II has been extensively studied. Comparing substrate specificities of the TE
domain from human type I FAS and TE type II, TE type I domain shows highest activity for C16
substrates followed by C18, with comparable activities for acyl-ACP and acyl-CoA esters, whereas TE
type II shows highest activity for C14 substrates followed by C16, preferring ACP bound acyl chains
over CoA esters. Moreover, TE type II appears to be more active against acyl-ACP than TE type I
domain and incubation of intact full-length human FAS with TE type II resulted in a product spectrum
enriched in caprylic acid (C8), capric acid (C10) and lauric acid (C12). Thus, TE type II efficiently
outcompetes the endogenous TE type I domain within FAS and releases fatty acids of shorter chain
length, although intramolecular hydrolysis by the intrinsic TE type I domain should be favored based
[98b]
on proximity.

18

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Accepted Manuscript
Figure 10: TE mechanism and structure. A) Mechanistic pathway for the TE-mediated hydrolysis. The active site
[42]
residues are numbered according to human FAS type I. B) Structure of the TE domains of human type I FAS
[100]
with a covalently bound γ-linolenyl tail of methyl γ-linolenylphosphonate (MGLP) (PDB ID: 3TJM). Domains are
colored as introduced in Figure 1, except that the TE subdomain is shown in grey for clarity. A model of animal
FAS type I is shown as inset. Active site residues and the γ-linolenyl tail are shown in stick representation. The TE
domain is comprised of a α/β hydrolase fold and an inserted helical subdomain. As depicted in zoom-in panels,
ligand binding causes the structuring of an amphiphatic α-helix (residues 2342–2355), termed the “helix flap”, and
the emergence of two binding pockets. Site 1 is formed between the two subdomains and present also in the apo-
form. C) Human type II TE (colored in blue) in superposition with the apo-structure of the TE domain of human
FAS type I (yellow). Type II TE is comprised of the canonical α/β hydrolase subdomain and a small capping
[98b]
domain (PDB ID: 4XJV). The structure is in closed conformation and cannot adopt an acyl-chain without
clashes (bottom panel). Please note the missing helix flap of the apo-TE domain (yellow). All structural figures
[78]
were prepared with the program PyMOL.

The X-ray structure of TE domain from human FAS type I has been determined, and is a
[42]
representative for the evolutionarily highly conserved animal FAS type I TEs (PDB ID: 1XKT and
[101]
PDB ID: 4Z49 ) (Fig. 10B). Additionally, an X-ray structure of human FAS type I TE domain in
[100]
complex with a polyunsaturated acyl chain is available (PDB ID: 3TJM). The structure of the
monomeric 32 kDa TE type I domain comprises two subdomains. The larger subdomain (23 kDa) is
arranged in an α/β hydrolase-like fold, while the smaller subdomain (9 kDa), which is inserted between
two β-strands of the hydrolase fold, shows an all α-helical motif sitting on top of α/β hydrolase-like fold
like a flexible cap. The interface between the two subdomains forms a highly hydrophobic groove
close to the catalytic triad, which was identified as Ser2308, Asp2338 and His2481 by site-directed
[42, 102]
mutagenesis experiments (human FAS numbering). An extensive hydrogen bonding network
stabilizes the catalytic triad and reaction intermediates in the hydrolysis of the thioester bond. The
binding modes of fatty acyl chains ranging from C12 to C20 in length were analyzed in a molecular

19

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[92]
docking study and underline the substrate specificity of the TE. Palmitate, the substrate with highest
preference, fits perfectly into the groove with its hydrophilic head, the carboxyl group, interacting with
the catalytic triad, and the long aliphatic tail held in place by overall hydrophobic residues reaching into
a distal binding pocket. The hydrophobic interactions between the TE groove and the aliphatic tail of
the substrate are the dominant driving force for the binding. While this binding mode still tolerates C18
chain length, longer chains than C18 lose their interaction with the catalytic triad and can no longer be
accommodated by the TE domain. The X-ray structure of human type II TE revealed the conserved
α/β hydrolase fold and a mobile capping domain, which exhibited greater similarity to type II TE
domains of PKS systems (29% sequence identity to RifR from rifamycin synthesis pathway (PDB ID:
[103] [104]
3FLB) and 26% to RedJ from prodiginine pathway (PDB ID: 3QMV) ) than to the human FAS

Accepted Manuscript
[98b]
type I TE domain (18% sequence similarity (PDB ID: 4XJV) ). In the available crystal structure, TE
type II shows a closed conformation with the active site covered by the mobile capping domain. Taking
into comparison the human FAS type I TE domain structure in complex with the inhibitor orlistat (PDB
[105]
ID: 2PX6) , the capping domain of TE type II must undergo considerable structural changes in order
to bind acyl chains. These changes are most likely induced by specific interactions with the
[98b]
phosphopantetheine moiety and the surface of ACP (Fig. 10C).

Although TEs are not considered as integral part of bacterial type II fatty acid biosynthesis, they are
able to interfere in chain length regulation. In E. coli, two different TEs were identified, the 22 kDa
[106]
thioesterase I (also known as TesA or TAP) and the 122 kDa thioesterase II (also known as
[107]
TesB) , which can catalyze hydrolysis of both acyl-CoA and acyl-ACP substrates, though the latter
[108]
with slower rates. Mostly, the products of type II fatty acid biosynthesis in bacteria are readily
catabolized in metabolic pathways, leaving little room for chain length control of those highly specific
TEs in their native system. However, upregulation of native TEs in E. coli or heterologous expression
of bacterial TEs combined with metabolic engineering is extensively used to produce high amounts of
[109]
technologically relevant short and medium chain free fatty acids.

[110] [111]
In the past decade, about a dozen structures of TesA (PDB ID: 1IVN , PDB ID: 5TIC ) and also
[112]
one structure of TesB (PDB ID: 1C8U ) have been determined by X-ray crystallography.
Furthermore X-ray crystal structures of TesA from M. tuberculosis (PDB ID: 6FW5 & 6FVJ) and from
[113]
Pseudomonas aeruginosa (PDB ID: 4JGG) are published. The two E. coli TEs differ not only in
structure but also in their functional diversity. TesB is a monofunctional protein, which forms a
homodimer, with each monomer showing a double hotdog fold. Structural analysis and site-directed
mutagenesis revealed residues Asp204, Gln278 and Thr228 as catalytic triad. The active site is
stabilized by an extensive hydrogen bonding network, which orients a water molecule between
Asp204 and Thr228 for nucleophilic attack on the substrate. An oxyanion hole stabilizes the
tetrahedral intermediate formed during catalysis, and the negative charge of Asp204 promotes
[112]
expulsion of the negatively charged reaction product. TesA on the other hand is a multifunctional
enzyme, which acts as thioesterase, protease and lysophospholipase. The monomeric TesA protein
folds into a central β-sheet flanked by helices and is a member of the SGNH-hydrolase family, which is
underlined by several indications (topology of the catalytic triad, absence of a nucleophilic elbow,
[110]
identical invariant residues, structural homology and a conserved hydrogen bonding network).

20

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Structural analysis revealed residues Ser10, Asp154 and His157 as the catalytic triad and the active
site is stabilized by an extensive hydrogen bonding network (E. coli TesA numbering). A highly
hydrophobic groove is located on top of the central β-sheet, which can accommodate long acyl chains
[110]
(>C10). The catalytic mechanism of SGNH-hydrolases is still poorly understood, but X-ray
[114]
structures with bound octanoic acid (C8) (PDB ID: 1U8U) and the inhibitor diethyl p-nitrophenyl
[110]
phosphate (DEP) (PDB ID: 1J00) , supported by site-directed mutagenesis and molecular dynamics
[111]
simulations , elucidated the substrate binding mode of TesA. The acyl chain binds with its
hydrophilic head, the carboxyl group, interacting with the catalytic triad. The long aliphatic tail is held
by overall hydrophobic residues, facing the gatekeeper residue Arg108. The tetrahedral intermediate
formed during catalysis is stabilized by an oxyanion hole and a negatively charged aspartate residue

Accepted Manuscript
[110]
near the active site facilitating expulsion of the negatively charged reaction product. Upon substrate
[114]
binding, a conformational change within a flexible loop, the so-called switch loop, is induced. This
switch loop movement is chain length dependent and determines substrate specificity of TesA. Much
like in the animal FAS type I TE domain, hydrophobic interactions between the groove and the
aliphatic tail of the substrate are the dominant driving forces in binding and switch loop movement.
The switch loop stabilizes the enzyme-substrate complex during catalysis building a continuous
[111, 114]
hydrophobic surface. For short acyl chains, hydrophobic interactions are partly lost, disturbing
the continuous hydrophobic surface which does not induce switch loop movement, as seen in the
[114]
TesA structure with the inhibitor DEP bound.

Two classes of TE type II were found in plants, FatA and FatB, with FatA showing high
preference for oleic acid (C18:1) and minor preference for stearic acid (C18) and palmitic acid (C16),
and FatB showing preference for saturated acyl chains ranging from C8-C16 in length. FatB can be
further divided into two groups, one showing specificity for long and the other for medium chain
[63, 115]
substrates. Best characterized plant TE is FatB from Umbellularia californica (termed UcFatB),
[9, 116]
which has has been extensively studied since almost three decades , and its X-ray structure has
[117]
been recently determined (PDB ID: 5X04). UcFatB forms a homodimer, with each monomer
showing a tandem hotdog fold. Structural analysis and site-directed mutagenesis revealed the
substrate binding pocket and residues Asp281, Asn283, His285, and Glu319 forming the catalytic
network. Although structural comparison of UcFatB to TesB showed a similar hotdog backbone,
different active sites suggest diverse catalytic mechanisms. The proposed catalytic mechanism of
UcFatB requires a conformational change from a closed to open structure, uncapping the substrate
binding pocket, induced by binding to acyl-ACP. The substrate specificity of the UcFatB is explained
by limited space of the binding pocket. Residue Thr137 is located near the bottom of the binding
pocket and its bulky side-chain restricts binding of acyl chains longer than C12, making lauric acid
(C12) the main product. In contrast, the rationally designed mutant Thr137Gly shifted the product
[117]
spectrum towards longer fatty acids of C14-C16 length.

3. Regulation of fatty acid chain length in type I FAS

21

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Fatty acid biosynthesis can be divided into three stages with each key player in chain length regulation
acting primarily on one stage: chain initiation is dominated by substrate loading catalyzed by the AT,
chain elongation is predominantly depending on the condensation reaction of the KS, and chain
termination occurs upon hydrolysis by the TE (in animal type I and plant type II FAS) or by the off-
loading by the MPT in type I fungal and CMN-bacterial FAS.

3.1. Animal type I FAS: First mechanisms of chain length regulation can already be found at the
chain initiation stage of the fatty acid cycle. Translocation of acyl chains is a reversible process
catalyzed by the promiscuous MAT domain. The MAT domain not only loads both the priming acetyl

Accepted Manuscript
and elongating malonyl moieties onto FAS, but can also unload FAS by back transfer of the acyl
moieties onto free CoA. Loading and unloading of acyl moieties proceed in stochastic manner at high
-1
turnover rates of 100 s and without limiting the overall rate of fatty acid biosynthesis, as estimated
–1 [118]
from the reported production of one palmitate in less than one second (0.8 s ). It should be noted,
that fast transacylation kinetics may be a way to tolerate the polyspecific character of MAT. In the
equilibrium of substrate translocation between CoA and the ACP domain, CoA can function both as
activator and inhibitor, depending on its concentration. High concentrations of CoA shift the equilibrium
towards acyl-CoA, resulting in substrate unloading and inhibition of FAS. Low concentrations of CoA
[118-119]
shift the equilibrium towards acyl-ACP, resulting in substrate loading and activation of FAS.
Scavenging free CoA, for example by ATP:citrate lyase or phosphotransacetylase, leads to inhibition
[120]
of FAS, preventing a proof-reading function of the MAT domain. Besides the CoA level, also the
concentrations of acyl-CoAs play an important role in fatty acid biosynthesis. Due to the bifunctional
role of the MAT domain, the priming acetyl- and elongating malonyl moiety compete for FAS loading.
An imbalanced substrate pool, shifted to high concentrations of either one, can act inhibiting on
[121]
FAS. Relative substrate concentrations can also influence chain length. The impact of malonyl-CoA
concentration on chain length was first described in studies on mammalian type I FAS in crude
[122]
extracts of mammary glands. In general, high malonyl-CoA concentrations promote chain
elongation and result in longer chain lengths, whereas low concentrations (high acetyl-CoA
concentrations) favor priming of fatty acid synthesis, resulting in more short chain products. These
short chain products are not necessarily fatty acids, but may escape from FAS as acyl-CoA esters by
[97]
MAT-mediated off-loading. The MAT domain of murine FAS type I is particularly suited to transfer
acyl chains of up to octanoyl (C8) with similar efficiencies, while the transfer of longer chains, such as
palmitoyl (C16) leads to lower transfer efficiency. Thus, it seems the longer the acyl chain is grown,
[44b, 88]
the less likely it escapes from further elongation via the MAT-mediated off-loading.

Kinetic measurements on chicken type I FAS revealed turnover numbers for each catalytic step
[121]
(except chain termination). According to these measurements, the slowest step in the elongation
-1
cycle is ketoreduction catalyzed by the KR domain with a turnover number of 17 s . All other reactions
of the elongation cycle, catalyzed by KS, DH and ER, have similar turnover numbers in the range of
-1
30 to 40 s . Their rate constant was estimated to increase by a factor of two for each C2 unit added to
the growing fatty acid chain. Thus, once chain elongation has been initiated, it rapidly proceeds to

22

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

longer chain lengths, making termination of fatty acid biosynthesis at the stage of medium chain
lengths nearly impossible.

At longer acyl chain lengths, the competitive interplay between the KS and TE domains
becomes relevant for chain length regulation: The main product of fatty acid biosynthesis in animal
FAS type I is palmitic acid (C16). The KS domains of animal FAS type I show preformed binding
[27, 44a]
pockets, implying the regulation of chain length by restricted space as found in other FASs. After
rapidly elongating short and medium chain intermediates, the steric hindrance of the KS binding
channel restricts the elongation of chains longer than C16, so that they dwell on the ACP and become
[3c, 121]
targets for TE-mediated hydrolysis. Complete deletion of the TE domain leads to accumulation of

Accepted Manuscript
C18 and C20 acyl moieties on FAS under loss of FAS activity, giving insight into the native chain
[94]
length regulation of the KS domain. The TE domain is strongly substrate-specific and shows highest
activity in hydrolyzing palmitoyl moieties. This outstanding substrate specificity arises from the binding
channel of the TE, which serves as molecular ruler. The palmitoyl moiety fits perfectly into the binding
channel. Its hydrophilic head group is embedded in the active site and the long aliphatic tail is buried
in the hydrophobic groove. Binding affinities of shorter acyl chains are weakened as less hydrophobic
[92, 98b]
interactions between the aliphatic tail and the groove are formed (Fig. 11A).

3.2. Microbial type I FAS: Similar to animal FAS type I, substrate concentrations influence chain
length in microbial FAS type I. Lynen and coworkers discovered the influence of malonyl-CoA
concentration on product chain length already in the 1960s when analyzing purified baker’s yeast FAS
[123]
type I in vitro. Just a few years later, Sumper et al. explained chain length control as the output of
the competitive interplay between the elongating KS-mediated condensation and the terminating MPT-
mediated release of the acyl moiety. In this model, the MPT domain is able to release acyl chains at a
chain length longer than C13, and the transfer rates then increase with acyl chain length owing to
increasing hydrophobic interactions. The influence of the KS domain was termed as hydrophilic
repulsion. Treated in inverse manner to MPT, the repulsion increases with hydrophobicity of the acyl
chain. A simple mathematical model was able to describe chain length as a function of acetyl- to
malonyl-CoA concentration. It readily explained the influence of the malonyl-CoA concentration on the
chain lengths of the synthesized fatty acids. Studies during the following years and decades have
slightly revised this early model. First, Pirson et al. showed that also acyl-CoAs shorter than C13, as
decanoyl-CoA, are well accepted by the MPT of baker’s yeast FAS, pointing towards the broader
[82a]
substrate tolerance of MPT as defined by the model of Sumper et al. The broad substrate tolerance
of the MPT domain of baker’s yeast FAS has also been proven in the recent engineering approach for
[77b, 77c] [31c]
the production of hexanoyl- and octanoyl-CoA. Second, recent structural and functional
[77b, 77c]
data elucidated the role of the KS domain as strict regulator in chain length control with tight
substrate specificity. The KS-binding channel is of limited length, so that the growing acyl chain
encounters a sharp spatial barrier when reaching the end of the channel. In this property, the KS-
[4, 57]
domain is similar to well-studied KS enzymes of type II FAS.

The fate of an acyl-ACP can therefore be represented as follows: Subsequent to processing the
condensation product β-ketoacyl-ACP to yield the fully reduced acyl-ACP, the interplay of the domains

23

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

MPT and KS decides on termination (MPT) vs. further elongation (KS). In the case of a long acyl-ACP
(C16 or C18), the KS domain of baker’s yeast prevents the incorporation of the fatty acid into the
binding channel, while the MPT domain shows affinity for the hydrophobic tail and terminates
synthesis. A detailed kinetic analysis of the key domains MPT and KS has not been performed to date.
However, according to structural and functional data outlined above, termination of fatty acid synthesis
results from the strict decision of the KS to not accommodate the long acyl chain, which persuades the
more substrate tolerant MPT to release the chain. In addition to the specificities imposed by the
domains, substrate concentration is involved in the decision process, similar as outlined above for
animal type I FAS. An excess of the malonyl extender substrate shifts the fatty acid spectrum to longer
chain length. Since the MPT domain loads the malonyl moiety, but also releases the final product,

Accepted Manuscript
both compounds compete for the MPT active site. When the concentration of malonyl-CoA increases,
long chain acyl moiety can be repelled from MPT domain binding and forced to KS binding, where it is
eventually condensed to the C2-elongated acyl-ACP. Conversely, low relative concentrations of
malonyl-CoA results in the increase of shorter chain products (Fig. 12A). The impact of the malonyl-
CoA concentration on the chain length has also been shown in vivo. In living cells, malonyl- to acetyl-
CoA concentrations are largely influenced by the acetyl-CoA carboxylase (ACC) that catalyzes the
ATP-dependent carboxylation of acetyl-CoA. Chain length has been shown to respond to the ACC
[124]
activity with longer fatty acids being produced at high ACC activity , in line with in vitro data. Several
studies have also shown that other external factors can influence chain length. For yeast FAS, an
acyl-CoA binding protein was identified, that leads to a shift towards shorter acyl-CoA ester in vitro
[125]
and fatty acids in vivo. For M. smegmatis FAS, polysaccharides and small proteins have been
[126]
reported to modulate activity and to interfere in its bimodal product spectrum. The sensitivity of the
product chain length to external factors may originate from the complexation of acyl-CoA products in
the bulk solution that changes the active concentrations of substrates and products.

Apart from substrate concentrations and external factors, few examples have been reported
on microbial type I FAS with modulated chain length regulation. A secondary metabolite FAS type I in
Aspergillus species, originally identified in A. nidulans (Uniprot numbers: Q00681/α-chain, Q00706/β-
chain)[127], produces hexanoyl-CoA as precursor in the synthesis of the polyketide aflatoxin.[128] Another
secondary metabolite FAS type I, occurring in Ustilago maydis (Uniprot number: UM06460/single
chain FAS), produces a hexanoyl-CoA precursor for the glycolipid biosurfactant ustilagic acid.[129]
Mycobacterial type I synthesis is responsible for a bimodal product pattern with acyl chain lengths of
C16/C18 and C22-C26.[130] The structural basis for the inherent chain length regulation properties of
FASs is currently being disclosed. Best understood are structural features of the KS binding channel
that induce (at least partly) short chain acyl chain production, like in type I Aspergillus species and U.
maydis FASs.[131] In both cases, the gatekeeper effect of the KS-binding channel is involved in short
chain acyl-CoA synthesis (see Fig. 8A & B). In the Aspergillus species and U. maydis FASs, a glycine
located next to the gate keeping methionine is mutated to serine and alanine, respectively, which both
likely decrease the conformational mobility of methionine, so that the hexanoyl moiety is prevented
from binding and from further elongation (equivalent residues to baker’s yeast FAS Gly1250; see Fig.
8B & C). In the Aspergillus FAS, in addition, a tyrosine is protruding into the KS binding channel from

24

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

the opposite side of the gate keeping methionine that probably strengthens the steric barrier
(equivalent residues to baker’s yeast FAS Phe1279; see Fig. 8B & C). There is multiple evidence for
the relevance of these mutations in short chain acyl-CoA production. The mutation Gly1250Ser in
baker’s yeast FAS was identified in a self-cloning approach to induce production of the apple-like
flavor component ethyl caproate (C8 ethyl ester) in Japanese sake.[77a] The same Gly1250Ser
mutation, in isolation and in combination with other mutations (as for example Phe1279Tyr), was used
for the construction of short chain acyl-CoA producing baker’s yeast strains.[77c] An equivalent mutation
of glycine to serine in Corynebacterium ammoniagenes type I FAS similarly induces short acyl-CoA
production in vitro.[77b] Structural features of the MPT domain that collaborate in the release of the short
chain acyl-CoA are not known. A recent cryo-EM study on M. tuberculosis type I FAS reported

Accepted Manuscript
extended binding channels in the KS- and the MPT-domains, which can accommodate long acyl
chains and explain the property of this FAS to partly elongate fatty acids beyond palmitoyl-CoA to also
yield C24/26-CoA.[132] The structural features responsible for the bimodal spectrum of mycobacterial
FAS type I, i.e. producing acyl chain lengths of C16/C18 and C22-C26, remain elusive.

3.3. Regulation of fatty acid chain length in type II FAS

The product spectra of bacterial and mitochondrial FAS type II systems are generally dominated by
[51b, 57]
long chain (C16/C18) acyl-ACPs (and additionally octanoyl-ACP (C8) in the mitochondrial FAS).
In contrast to type I FAS systems, the designated ATs (MCATs) of type II FASs have no significant
influence on the product spectrum. Chain length control depends mainly on substrate specificity of the
respective KSs, which is essentially determined by their length of the substrate binding channels. All
three KSs of E. coli display different substrate specificities; FabH catalyzes the initiation of fatty acid
synthesis and shows highest activities for short chain substrates (<C4), while FabB and FabF catalyze
the elongation of C4-C14 acyl-ACP substrates. Overexpression of FabH leads to more frequent
[75]
initiation, and thus to a larger pool of short chain acyl-ACPs. Rapid elongation rates of FabB and
FabF usually prevent formation of medium chain products. Both enzymes can be inhibited by
cerulenin, while leaving FabH-mediated priming unaffected, which results in decreased elongation
rates and compromised feedback inhibition of FabH due to lower concentrations of long chain acyl-
[109b, 109d, 133]
ACPs, which further increase the initiation rate.

In E. coli FabF and mtKAS, the gatekeeper residues, Ile108 (E. coli FabF numbering) and Met154
(human mtKAS numbering) respectively, require reorientation during binding of long acyl chains. An
increased rotational barrier of the gatekeeper residue in mtKAS is considered as causing the bimodal
[47b, 57, 73b, 73e]
product spectrum of mitochondrial fatty acid synthesis.

Contrary to bacterial and mitochondrial FAS type II systems, plant FASs harness TEs to terminate
fatty acid biosynthesis. These TEs display tight substrate specificities and are the principal regulators
in chain length control. When plant TEs are overexpressed in E. coli, their substrate specificities

25

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[2a]
determine the product spectrum of fatty acid synthesis. As shown by mutagenesis studies, altering
[134]
TE activity relative to KAS activity can influence the product spectrum of fatty acid synthesis.

4. Directed engineering for modulation of fatty acid chain length

Fatty acids are considered to be strategically important platform chemicals accessible through
sustainable microbial approaches.[135] The recently acquired structural information on FASs provides
an excellent understanding of the molecular basis of fatty acid synthesis. The specific understanding

Accepted Manuscript
of chain length control, the characterization of a multitude of substrate-specific TEs, and the emerging
tools and means for metabolic engineering have fostered targeted approaches for modulating chain
length. There is large interest in short chain fatty acids, since these compounds are biotechnologically
valuable platform chemicals and biofuel precursors, and a few attempts on the synthesis of short chain
length fatty acids have been reported during recent years. We note that engineered FASs have
usually been embedded in microbial hosts, preferentially E. coli, S. cerevisiae and Yarrowia lipolytica,
in which the metabolisms were optimized towards high product titers. Metabolic engineering aims at
increasing the supply of acyl-CoA substrates and the reducing co-factor NADPH, or block competing
pathways, as β-oxidation. As fatty acids may also be toxic to the host cell, efflux pumps have been
inserted, too. Metabolic engineering approaches to increase product titers have been reviewed
[136]
elsewhere , and will not be considered here.

4.1. Chain length engineering of type I FAS: Type I FASs are ideal proteins in biotechnological
applications, because the compartmentalized reaction grants high catalytic efficiency and allows linear
[22, 137]
synthetic pathways with little branching. Estimated from the dimensions of the barrel-shaped
structure and assuming full loading of ACPs, fungal de novo fatty acid synthesis runs at effective
concentrations of active sites and covalently tethered substrates/intermediates of 1.8 mM, when
assuming full occupancy (see Fig. 3A). An analogous calculation on animal FAS, abstracted as
[37a]
cylindrical reaction spaces, results in effective concentration of 1.2 mM (see Fig. 4). For the E. coli
type II FAS system of separate enzymes, copy numbers range from 10,000 for malonyltransferase
[138] 3
FabD to 23,000 for dehydratase FabA. Considering an average volume of E. coli cell of 2.5 µm
[139]
, synthesis of fatty acids in E. coli proceeds at significantly lower molar concentrations of enzymes
of about 0.007 to 0.016 mM. Since in type I systems the ACP domain holds substrates and
intermediates within the reaction compartment, biosynthetic pathways are highly channeled by
[140]
proximity and do not suffer from side reactions.

Attempts to engineer the product spectrum of animal FAS type I have mainly harnessed the
pool of substrate-specific TEs. TEs were either applied as separate domain in trans or swapped with
the native TE domain. Animal FASs type I invest only 9% of the total sequence in linkers, which
[44b]
makes this fold highly susceptible to domain swaps (see Fig. 4). Just recently the modular nature
[141]
of the animal FAS fold was demonstrated in a protein engineering study. Following this concept,
the animal type I FAS was shown to produce short chain fatty acids when exchanging the cognate TE

26

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[142]
domains with the short chain acyl-ACP specific type II TEs. Specifically, the TE of rat FAS type I
[142a]
was swapped with human TE type II, shifting the product spectrum towards short chain fatty acids ,
and the TE of human FAS type I was swapped with human TE type II or FatB1 from Cuphea palustris
[142b]
inducing C6/C8 and C8/C10 production, respectively, in baker’s yeast. In the latter work, TE type
II or FatB1 were also co-expressed with human FAS truncated by the TE domain (Fig. 11B). Although
the interaction in trans revealed to be less efficient, data discloses that the ACP of type I animal FAS
fold is sufficiently exposed for such in trans engineering strategies. Intriguingly, simply the variation of
the TE domain linker length in animal FAS type I can influence product spectra. While addition of
residues or a deletion of nearly half the ACP-TE interdomain linker showed no effect, further
shortening resulted in reduced FAS activity and shifted the product spectrum towards longer chain

Accepted Manuscript
[142a]
lengths of C18 and C20. Changes in linker length likely affect the domain mobility and the
interplay of ACP and the terminating TE, so that acyl-ACP is steered towards the elongating KS-
mediated condensation. It should be noted that any domain swap approach is invasive, since it is
introducing non-cognate interactions between ACP and the enzymatic domain.

Figure 11: Chain length control in animal FAS type I. A) The MAT domain is responsible for loading and offloading
of the FAS with both the acetyl priming and malonyl elongating unit. The KS domain catalyzes chain elongation
and determines chain length through a limited-length channel. The TE domain terminates fatty acid synthesis and
exhibits tight substrate-specificity for long acyl chain lengths. B) Modulation of chain length. Flow scheme of
wildtype FAS and altered FAS, in which the reaction fluxes were changed. The fatty acid cycle is shown without
the processing sequence. The color of the arrows symbolizes the course of the reaction compared to wildtype
FAS; green and red mean that reactions are faster or slower, respectively. The new pathway through the type II
TE (TE2) is shown by a dashed arrow.

Microbial type I FASs have also been engineered for the production of short chain acyl-CoA.
The microbial type I FAS fold invests 50% of the total sequence in linkers and non-catalytic insertion
elements, leading to an overall intricate protein complex of the size of the ribosome (diameter of about
25 nm) (see Fig. 3A & B). Linkers and insertion elements are responsible for the assembly and the
[143]
stability of the barrel-shaped structure. As the enzymatic domains are tightly embedded in the
scaffolding matrix, engineering of the protein by domain swaps is difficult and may be envisioned just
highly homologous domains. The fungal FAS type I from baker’s yeast and from Y. lipolytica have

27

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[144]
been modified by inserting the TE domains into the multienzyme complex or by exchanging the
[145]
MPT with short chain acyl-ACP specific type II TEs , respectively. It should be noted that the
exchange of the bifunctional MPT domain by a TE domain is problematic, because such a swap does
not specifically target the termination function of MPT, as intended, but also impairs the malonyl
loading function (Fig. 12 A & B). The synthesis of short chain length acyl-CoAs and fatty acids has
recently also been achieved by mutating selected amino acids in active sites and binding channels of
[77c] [77b]
the domains KS, AT and MPT of baker’s yeast FAS and corynebacterial FAS. The KS binding
channel was mutated to increase the gatekeeper barrier by, for example, employing the glycine to
serine mutation that naturally occurs in Aspergillus FAS type I for producing hexanoyl-CoA as starter
[127-128]
unit in aflatoxin synthesis (Gly1250Ser, baker’s yeast FAS numbering; Table 1). Including

Accepted Manuscript
further mutations that support the gatekeeper barrier (Met1251Trp and Phe1279Tyr; Table 1),
selected by multiple sequence alignment and analyzed by molecular dynamics simulations, the KS
domain successfully prevented the accommodation of longer acyl chains and forced the acyl-ACP into
interaction with the MPT domain for premature acyl chain release (see Fig. 8B & C). A similar
approach has also succeeded in producing C14-CoA in Y. lipolytica FAS type I. Here, an isoleucine to
tryptophan mutation at the bottom of the KS-binding pocket spatially constrained the incorporation of
[146]
long acyl chains (Ile1251Trp, Y. lipolytica FAS numbering; Table 1). Engineering of the transferase
[77b, 77c]
domains can support KS-design. The active site arginine in the MPT domains of FASs from
baker’s yeast and C. ammoniagenes were mutated to lysine to omit the bidendate coordination of the
free carboxyl group of malonyl (Arg1834Lys, baker’s yeast FAS numbering; Table 1). The reduced
affinity for the malonyl moiety increased the export rate of acyl chain transfer and promoted pre-
mature termination of synthesis. The direct engineering of the MPT domain for improving the release
reaction, e.g. by increasing the hydrophobicity of the binding channel in order to favor short acyl chain
[77b]
export, failed in this study. The difficulty in engineering the MPT domain lies in the generally high
conformational variability of the AT binding site, and the missing structural information on MPT
domains in complex with acyl-chains. Finally, the acetyl-loading AT domains of FASs from baker’s
yeast and C. ammoniagenes were mutated to provide an alternative exit tunnel for medium length fatty
[88]
acids (Ile306Ala, baker’s yeast numbering; Table 1) (Fig. 12 C). ATs of PKS systems have
frequently been engineered. Here, approaches were particularly successful for ATs that bind small
[147]
acyl-CoA esters for priming synthesis. Just recently, it was shown for type I fungal and CMN-
bacterial FASs that single mutations in the domain-domain-interface of ACP and the elongating KS
[35a]
domain can severely shift chain length spectra. The product acyl chain length correlated with the
strength of the interactions between the ACP domain and enzymatic domains of the mutated FAS
determined by computational docking. The study demonstrates that targeted mutations distant from
the active site can be used to modulate the product spectrum (Fig. 12B). Direct engineering of type I
fungal and CMN-bacterial FASs is beneficial, because these FASs deliver short chain length products
as activated acyl-CoA esters that can directly be processed to alkyl derivatives. TE-based chain length
control requires reactivation of the free fatty acids to the acyl-CoA ester in an ATP-consuming
[109a]
process.

28

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Accepted Manuscript
Figure 12: Chain length control in yeast FAS type I. A) The KS domain determines chain length through a limited-
length channel, which accommodates acyl chain up to a certain chain length (C14 or C16). The MPT domain
incorporates both the malonyl substrate and the palmitoyl (C16) product. B) Chain length control in modified
baker’s yeast strains. Flow scheme of wildtype and mutated FAS, in which the reaction fluxes were changed. The
fatty acid cycle is shown without the processing sequence. The color of the arrows symbolizes the course of the
reaction compared to wildtype FAS; green and red mean that reactions are faster or slower, respectively. The
new pathway of product release through the AT domain is shown by a dashed arrow.

Organism S. cerevisiae Y. lipolytica R. toruloides C. ammon. E. coli M. musc.

A FAS α FAS α
Protein name FAS α subunit subunit subunit FAS fabF (KASII) Fasn
UniProt ID P19097 Q6CE16 M7XM89 D5NXL2 P0AAI7 P19096

KS mut./eq. G1250S G1216 G2301 G2599S G107 S112

position M1251W M1217 M2302 M2600W I108F G113

F1279Y F1245 F2330 A2637 I138 C135
F1335 F1301 F2386 L2684 A193M L191
V1254 I1220W I2305 M2603 L111 A116

B Protein name FAS β subunit FAS β subunit FAS β subunit FAS none Fasn
UniProt ID P07149 P34229 M7WSW5 Q04846 P19096
AT mut./eq.
position I306A I318 I349 I151A R606

C FAS α
Protein name FAS β subunit FAS β subunit subunit FAS fabD Fasn
UniProt ID P07149 P34229 M7XM89 Q04846 P0AAI9 P19096
MPT mut./eq.
position R1834K R1868 R667 R1408K R117 R606

Table 1: Mutations (mut.) for FAS chain length control within KS (A), AT (B) and MPT (C) published in literature
and their equivalent (eq.) positions in different model organisms relevant for genetic engineering

29

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

(C. ammon. = C. ammoniagenes; M. musc. = Mus musculus). The references are indicated by a color code
[77c] [146] [77b] [76]
(green: Gajewski et al. ; pink: Rigouin et al. ; yellow: Gajewski et al. ; blue: Val et al. ).

4.2. Chain length engineering of type II FAS: In spite of the successes in engineering type I
FASs for modulation of product chain length in fatty acid synthesis, most of the approaches have so
far focused on type II systems. Attempts to engineer the product spectrum of type II FAS systems
have mainly harnessed the pool of substrate-specific TEs from plants and bacteria. Examples of type
II FAS engineering approaches with heterologously expressed TEs are outlined below. For the
[2a, 148]
interested reader we further recommend reviews specialized on this topic.

Accepted Manuscript
The TE type II of rat mammary gland has been successfully utilized to engineer the product
spectrum of animal type I FAS towards shorter chain lengths, either as discrete protein or as C-
[142]
terminal fusion in exchange for the TE type I domain. Also type II TE of E. coli (TesA) has been
[109, 149]
engineered to alter the fatty acid product spectrum.
The substrate specificity of E. coli TesA was successfully switched towards medium chain
lengths. Since switch loop movement plays an important role in TE-mediated substrate specificity and
is highly sensitive to changes in hydrophobicity, structure-guided mutagenesis studies systematically
[111]
tailored the hydrophobic surface of the substrate binding channel of TesA. Mutation Leu109Pro
[114]
(PDB ID: 1JRL & 1V2G) was already shown to abolish switch loop movement due to lower
hydrophobicity, thereby decreasing catalytic efficiency for long acyl chains. From a computationally
guided library, mutation Ser122Lys was found to shift the product spectrum of TesA towards lauric
acid (C12) and mutation Met141Leu towards caprylic acid (C8). Most efficiently in producing caprylic
[111]
acid (C8) was the triple mutant Met141Leu/Tyr145Lys/Leu146Lys (PDB ID: 5TIF, 5TIE & 5TID) ,
which showed highly altered specificity while maintaining high activity. Furthermore, residues Leu11,
Gly72 and Ile156 were proposed as other potential targets for site-directed mutagenesis to alter the
[110]
substrate preference.
Since in plant type II FASs the substrate specificity of TEs can almost exclusively determine
fatty acid chain length, also plant type II TEs are suitable proteins for engineering. Efforts to engineer
fatty acid biosynthesis towards shorter chain lengths often utilize medium chain specific FatB
enzymes. In a proof of concept study UcFatB in Arabidopsis thaliana successfully increased levels of
lauric acid (C12) and myristic acid (C14), demonstrating that the native chain length control in fatty
[116]
acid synthesis can be overruled by heterologous proteins. This success was followed by
[150]
expression of UcFatB in E. coli. UcFatB is even able to interfere with the fungal FAS type I system
[151]
of Y. lipolytica. By now, a multitude of specific FatB enzymes from different plant species has been
[109c, 109d, 149, 150b, 152] [153] [154]
utilized not only in E. coli and higher plants , but also in algae and even
[155] [110-111, 114, 117, 134,
cyanobacteria . Efforts have also been made to tailor substrate specificities of TE
156] [76]
and KAS enzymes. For example, the FabF Ile108Phe mutant was successfully utilized in a
metabolic engineering study in E. coli. The FabF mutant was heterologously expressed in combination
with TesA and plant TEs. An induced FabB degradation prevented further elongation and accumulated
[133c]
octanoic acid (C8). In spite of engineering successes, the impact of mutations on plant TEs seems

30

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[134, 156a, 156b]

unpredictable. Although plant TEs overall share high sequence identity (35–60%), they
exhibit diverse substrate specificities and only limited structural information has been gathered thus
far. In general, approaches to produce short chain fatty acids in bacteria, plants and algae have been
marginally successful. Remarkable changes in the product spectrum and high quantities of the desired
products remain the exception.

Acknowledgements
M.G. thanks Dieter Oesterhelt (Martinsried, Germany) for many years of support. M.G. is also grateful
for valuable collaborations on FASs and related proteins. This work was supported by a Lichtenberg

Accepted Manuscript
grant of the Volkswagen Foundation to M.G. (grant number 85701). Further support was received from
the LOEWE program (Landes-Offensive zur Entwicklung wissenschaftlich-ökonomischer Exzellenz) of
the state of Hesse conducted within the framework of the MegaSyn and the DynaMem Research
Clusters.

Conflict of interest statement

Concerning competing financial interest, M.G. declares that he is a co-inventor of two patents
concerning the production of short fatty acids (EP patent application 15 162 192.7 filed on April 1st,
2015 and EP patent application 15 174 342.4 filed on June 29th, 2015).

31

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[1] M. Sumper, D. Oesterhelt, C. Riepertinger, F. Lynen, Eur. J. Biochem. 1969, 10, 377-387.
[2] aJ. Beld, D. J. Lee, M. D. Burkart, Mol. BioSyst. 2015, 11, 38-59; bD. A. Herbst, C. A.
Townsend, T. Maier, Nat. Prod. Rep. 2018, 35, 1046-1069.
[3] aO. Tehlivets, K. Scheuringer, S. D. Kohlwein, Biochim. Biophys. Acta 2007, 1771, 255-270;
bE. Schweizer, J. Hofmann, Microbiol. Mol. Biol. R. 2004, 68, 501-517; cS. Smith, FASEB J.
1994, 8, 1248-1259.
[4] S. W. White, J. Zheng, Y.-M. Zhang, C. O. Rock, Annu. Rev. Biochem. 2005, 74, 791-831.
[5] T. Maier, M. Leibundgut, D. Boehringer, N. Ban, Q. Rev. Biophys. 2010, 43, 373-422.
[6] aA. J. Kastaniotis, K. J. Autio, J. M. Kerätär, G. Monteuuis, A. M. Mäkelä, R. R. Nair, L. P.
Pietikäinen, A. Shvetsova, Z. Chen, J. K. Hiltunen, Biochim. Biophys. Acta 2017; bJ. K.
Hiltunen, M. S. Schonauer, K. J. Autio, T. M. Mittelmeier, A. J. Kastaniotis, C. L. Dieckmann,
J. Biol. Chem. 2009, 284, 9011-9015.
[7] Y.-J. Lu, Y.-M. Zhang, K. D. Grimes, J. Qi, R. E. Lee, C. O. Rock, Mol. Cell 2006, 23, 765-772.
[8] J. G. Van Vranken, S. M. Nowinski, K. J. Clowers, M. Y. Jeong, Y. Ouyang, J. A. Berg, J. P.

Accepted Manuscript
Gygi, S. P. Gygi, D. R. Winge, J. Rutter, Mol. Cell 2018, 71, 567-580.e564.
[9] M. R. Pollard, L. Anderson, C. Fan, D. J. Hawkins, H. M. Davies, Arch. Biochem. Biophys.
1991, 284, 306-312.
[10] aJ. Crosby, M. P. Crump, Nat. Prod. Rep. 2012, 29, 1111-1137; bR. N. Perham, Annu. Rev.
Biochem. 2000, 69, 961-1004.
[11] Y.-M. Zhang, M. S. Rao, R. J. Heath, A. C. Price, A. J. Olson, C. O. Rock, S. W. White, The
Journal of Biological Chemistry 2001, 276, 8231-8238.
[12] aS. E. Evans, C. Williams, C. J. Arthur, S. G. Burston, T. J. Simpson, J. Crosby, M. P. Crump,
ChemBioChem 2008, 9, 2424-2432; bS. E. Evans, C. Williams, C. J. Arthur, E. Płoskoń, P.
Wattana-amorn, R. J. Cox, J. Crosby, C. L. Willis, T. J. Simpson, M. P. Crump, J. Mol. Biol.
2009, 389, 511-528; cR. W. Haushalter, F. V. Filipp, K.-s. Ko, R. Yu, S. J. Opella, M. D.
Burkart, ACS Chem. Biol. 2011, 6, 413-418; dM. J. Jaremko, D. J. Lee, S. J. Opella, M. D.
Burkart, J. Am. Chem. Soc. 2015, 137, 11546-11549; eG. Shakya, H. Rivera, Jr., D. J. Lee, M.
J. Jaremko, J. J. La Clair, D. T. Fox, R. W. Haushalter, A. J. Schaub, J. Bruegger, J. F.
Barajas, A. R. White, P. Kaur, E. R. Gwozdziowski, F. Wong, S.-C. Tsai, M. D. Burkart, J. Am.
Chem. Soc. 2014, 136, 16792-16799; fG. A. Zornetzer, B. G. Fox, J. L. Markley, Biochemistry
2006, 45, 5217-5227.
[13] aD. R. Perez, M. Leibundgut, G. Wider, Biochemistry 2015, 54, 2205-2213; bE. Ploskon, C. J.
Arthur, S. E. Evans, C. Williams, J. Crosby, T. J. Simpson, M. P. Crump, J. Biol. Chem. 2008,
283, 518-528; cL. Tran, R. W. Broadhurst, M. Tosin, A. Cavalli, K. J. Weissman, Chem. Biol.
2010, 17, 705-716.
[14] A. Roujeinikova, W. J. Simon, J. Gilroy, D. W. Rice, J. B. Rafferty, A. R. Slabas, J. Mol. Biol.
2007, 365, 135-145.
[15] aJ. E. Cronan, Jr., Biochem. J. 2014, 460, 157-163; bM. Leibundgut, S. Jenni, C. Frick, N.
Ban, Science 2007, 316, 288-290.
[16] C. Nguyen, R. W. Haushalter, D. J. Lee, P. R. Markwick, J. Bruegger, G. Caldara-Festin, K.
Finzel, D. R. Jackson, F. Ishikawa, B. O'Dowd, J. A. McCammon, S. J. Opella, S.-C. Tsai, M.
D. Burkart, Nature 2014, 505, 427-431.
[17] S. Rafi, P. Novichenok, S. Kolappan, X. Zhang, C. F. Stratton, R. Rawat, C. Kisker, C.
Simmerling, P. J. Tonge, J. Biol. Chem. 2006, 281, 39285-39293.
[18] aA. Chen, R. N. Re, M. D. Burkart, Natural Product Reports 2018, 35, 1029-1045; bK. Finzel,
D. J. Lee, M. D. Burkart, ChemBioChem 2015, 16, 528-547; cD. M. Byers, H. Gong, Biochem.
Cell. Biol. 2007, 85, 649-662; dD. I. Chan, H. J. Vogel, Biochem. J. 2010, 430, 1-19.
[19] T. A. Ramelot, P. Rossi, F. Forouhar, H.-W. Lee, Y. Yang, S. Ni, S. Unser, S. Lew, J.
Seetharaman, R. Xiao, T. B. Acton, J. K. Everett, J. H. Prestegard, J. F. Hunt, G. T.
Montelione, M. A. Kennedy, Biochemistry 2012, 51, 7239-7249.
[20] aH. S. T. Bukhari, R. P. Jakob, T. Maier, Structure 2014, 22, 1775-1785; bH. Jenke-Kodama,
A. Sandmann, R. Müller, E. Dittmann, Mol. Biol. Evol. 2005, 22, 2027-2039.
[21] C. Vogel, M. Bashton, N. D. Kerrison, C. Chothia, S. A. Teichmann, Curr. Opin. Struct. Biol.
2004, 14, 208-216.
[22] L. J. Sweetlove, A. R. Fernie, Nat. Commun. 2018, 9, 2136-2148.
[23] M. Rawlings, J. E. Cronan, Jr., J. Biol. Chem. 1992, 267, 5751-5754.
[24] G. Petrovics, P. Putnoky, B. Reuhs, J. Kim, T. A. Thorp, K. D. Noel, R. W. Carlson, A.
Kondorosi, Mol. Microbiol. 1993, 8, 1083-1094.
[25] aD. A. Herbst, R. P. Jakob, F. Zähringer, T. Maier, Nature 2016, 531, 533-537; bM. Mathur, P.
E. Kolattukudy, J. Biol. Chem. 1992, 267, 19388-19395.
[26] K. J. Weissman, Natural Product Reports 2015, 32, 436-453.
[27] T. Maier, M. Leibundgut, N. Ban, Science 2008, 321, 1315-1322.

32

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[28] M. Grininger, Curr. Opin. Struct. Biol. 2014, 25, 49-56.

[29] aJ. K. Stoops, E. S. Awad, M. J. Arslanian, S. Gunsberg, S. J. Wakil, R. M. Oliver, J. Biol.
Chem. 1978, 253, 4464-4475; bS. J. Wakil, J. K. Stoops, V. C. Joshi, Annu. Rev. Biochem.
1983, 52, 537-579.
[30] D. Oesterhelt, H. Bauer, F. Lynen, P. Natl. Acad. Sci. USA 1969, 63, 1377-1382.
[31] aS. Jenni, M. Leibundgut, D. Boehringer, C. Frick, B. Mikolásek, N. Ban, Science 2007, 316,
254-261; bS. Jenni, M. Leibundgut, T. Maier, N. Ban, Science 2006, 311, 1263-1267; cP.
Johansson, B. Wiltschi, P. Kumari, B. Kessler, C. Vonrhein, J. Vonck, D. Oesterhelt, M.
Grininger, P. Natl. Acad. Sci. USA 2008, 105, 12803-12808; dI. B. Lomakin, Y. Xiong, T. A.
Steitz, Cell 2007, 129, 319-332.
[32] P. Johansson, B. Mulinacci, C. Koestler, R. Vollrath, D. Oesterhelt, M. Grininger, Structure
2009, 17, 1063-1074.
[33] P. Gipson, D. J. Mills, R. Wouts, M. Grininger, J. Vonck, W. Kühlbrandt, P. Natl. Acad. Sci.
USA 2010, 107, 9164-9169.

Accepted Manuscript
[34] E. D'Imprima, D. Floris, M. Joppe, R. Sánchez, M. Grininger, W. Kühlbrandt, bioRxiv 2018.
[35] aE. Rossini, J. Gajewski, M. Klaus, G. Hummer, M. Grininger, Chem. Commun. 2018, 54,
11606-11609; bC. Anselmi, M. Grininger, P. Gipson, J. D. Faraldo-Gómez, J. Am. Chem. Soc.
2010, 132, 12357-12364.
[36] M. Fischer, D. Rhinow, Z. Zhu, D. J. Mills, Z. K. Zhao, J. Vonck, M. Grininger, Protein Sci.
2015, 24, 987-995.
[37] aM. Fischer, M. Grininger, Beilstein J. Org. Chem. 2017, 13, 1204-1211; bA. Shiber, K.
Döring, U. Friedrich, K. Klann, D. Merker, M. Zedan, F. Tippmann, G. Kramer, B. Bukau,
Nature 2018, 561, 268-272.
[38] D. Boehringer, N. Ban, M. Leibundgut, J. Mol. Biol. 2013, 425, 841-849.
[39] aA. K. Joshi, V. S. Rangan, A. Witkowski, S. Smith, Chem. Biol. 2003, 10, 169-173; bS. Smith,
A. Witkowski, A. K. Joshi, Prog. Lipid Res. 2003, 42, 289-317.
[40] aJ. Brink, S. J. Ludtke, C. Y. Yang, Z.-W. Gu, S. J. Wakil, W. Chiu, P. Natl. Acad. Sci. USA
2002, 99, 138-143; bS. S. Chirala, P. Natl. Acad. Sci. USA 1992, 89, 10232-10236.
[41] T. Maier, S. Jenni, N. Ban, Science 2006, 311, 1258-1262.
[42] B. Chakravarty, Z. Gu, S. S. Chirala, S. J. Wakil, F. A. Quiocho, P. Natl. Acad. Sci. USA 2004,
101, 15567-15572.
[43] aF. M. C. Benning, Y. Sakiyama, A. Mazur, H. S. T. Bukhari, R. Y. H. Lim, T. Maier, ACS
Nano 2017, 11, 10852-10859; bE. J. Brignole, S. Smith, F. J. Asturias, Nat. Struct. Mol. Biol.
2009, 16, 190–197.
[44] aG. Pappenberger, J. Benz, B. Gsell, M. Hennig, A. Ruf, M. Stihle, R. Thoma, M. G. Rudolph,
J. Mol. Biol. 2010, 397, 508-519; bA. Rittner, K. S. Paithankar, K. V. Huu, M. Grininger, ACS
Chem. Biol. 2018, 13, 723-732.
[45] M. F. Viegas, R. P. P. Neves, M. J. Ramos, P. A. Fernandes, J. Phys. Chem. B 2017, 122, 77-
85.
[46] aC. Oefner, H. Schulz, A. D'Arcy, G. E. Dale, Acta Crystallogr. D 2006, 62, 613-618; bL.
Serre, E. C. Verbree, Z. Dauter, A. R. Stuitje, Z. S. Derewenda, J. Biol. Chem. 1995, 270,
12961-12964.
[47] aR. J. Heath, C. O. Rock, Natural Product Reports 2002, 19, 581-596; bW. Huang, J. Jia, P.
Edwards, K. Dehesh, G. Schneider, Y. Lindqvist, EMBO J. 1998, 17, 1183-1191; cJ. G. Olsen,
A. Kadziola, P. Von Wettstein-Knowles, M. Siggaard-Andersen, Y. Lindquist, S. Larsen, FEBS
Lett. 1999, 460, 46-52.
[48] J. L. Garwin, A. L. Klages, J. E. Cronan, Jr., J. Biol. Chem. 1980, 255, 11949-11956.
[49] J. K. Hiltunen, K. J. Autio, M. S. Schonauer, V. A. Kursu, C. L. Dieckmann, A. J. Kastaniotis,
Biochim. Biophys. Acta 2010, 1797, 1195-1202.
[50] aR. Schneider, B. Brors, F. Bürger, S. Camrath, H. Weiss, Curr. Genet. 1997, 32, 384-388;
bL. Zhang, A. K. Joshi, S. Smith, J. Biol. Chem. 2003, 278, 40067-40074.
[51] aG. Bunkoczi, S. Pasta, A. K. Joshi, X. Wu, K. L. Kavanagh, S. Smith, U. Oppermann, Chem.
Biol. 2007, 14, 1243-1253; bL. Zhang, A. K. Joshi, J. Hofmann, E. Schweizer, S. Smith, J.
Biol. Chem. 2005, 280, 12422-12429.
[52] A. Harington, C. J. Herbert, B. Tung, G. S. Getz, P. P. Slonimski, Mol. Microbiol. 1993, 9, 545-
555.
[53] R. Venkatesan, S. K. Sah-Teli, L. O. Awoniyi, G. Jiang, P. Prus, A. J. Kastaniotis, J. K.
Hiltunen, R. K. Wierenga, Z. Chen, Nat. Commun. 2014, 5, 4805-4816.
[54] aK. J. Autio, A. J. Kastaniotis, H. Pospiech, I. J. Miinalainen, M. S. Schonauer, C. L.
Dieckmann, J. K. Hiltunen, FASEB J. 2007, 22, 569-578; bA. J. Kastaniotis, K. J. Autio, R. T.
Sormunen, J. K. Hiltunen, Mol. Microbiol. 2004, 53, 1407-1421.

33

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[55] aZ.-J. Chen, R. Pudas, S. Sharma, O. S. Smart, A. H. Juffer, J. K. Hiltunen, R. K. Wierenga,

A. M. Haapalainen, J. Mol. Biol. 2008, 379, 830-844; bJ. M. Torkko, K. T. Koivuranta, I. J.
Miinalainen, A. I. Yagi, W. Schmitz, A. J. Kastaniotis, T. T. Airenne, A. Gurvitz, K. J. Hiltunen,
Mol. Cell. Biol. 2001, 21, 6243-6253.
[56] S. Smith, A. Witkowski, A. Moghul, Y. Yoshinaga, M. Nefedov, P. de Jong, D. Feng, L. Fong,
Y. Tu, Y. Hu, S. G. Young, T. Pham, C. Cheung, S. M. Katzman, M. D. Brand, C. L. Quinlan,
M. Fens, F. Kuypers, S. Misquitta, S. M. Griffey, S. Tran, A. Gharib, J. Knudsen, H. K.
Hannibal-Bach, G. Wang, S. Larkin, J. Thweatt, S. Pasta, PLoS ONE 2012, 7, e47196.
[57] C. E. Christensen, B. B. Kragelund, P. von Wettstein-Knowles, A. Henriksen, Protein Sci.
2007, 16, 261-272.
[58] S. Brody, C. Oh, U. Hoja, E. Schweizer, FEBS Lett. 1997, 408, 217-220.
[59] H. B. Clay, A. K. Parl, S. L. Mitchell, L. Singh, L. N. Bell, D. G. Murdock, PLoS ONE 2016, 11,
e0151171-0151193.
[60] aH. Angerer, S. Schönborn, J. Gorka, U. Bahr, M. Karas, I. Wittig, J. Heidler, J. Hoffmann, N.

Accepted Manuscript
Morgner, V. Zickermann, BBA - Mol. Cell Res. 2017, 1864, 1913-1920; bB. J. Floyd, E. M.
Wilkerson, M. T. Veling, C. E. Minogue, C. Xia, E. T. Beebe, R. L. Wrobel, H. Cho, L. S.
Kremer, C. L. Alston, K. A. Gromek, B. K. Dolan, A. Ulbrich, J. A. Stefely, S. L. Bohl, K. M.
Werner, A. Jochem, M. S. Westphall, J. W. Rensvold, R. W. Taylor, H. Prokisch, J.-J. P. Kim,
J. J. Coon, D. J. Pagliarini, Mol. Cell 2016, 63, 621-632.
[61] J. L. Harwood, Ann. Rev. Plant Physiol. Plant Mol. Biol. 1988, 39, 101-138.
[62] A. Witkowski, A. K. Joshi, S. Smith, Biochemistry 2002, 41, 10877-10887.
[63] A. Jones, H. M. Davies, T. A. Voelker, Plant Cell 1995, 7, 359-371.
[64] K.-I. Arnstadt, G. Schindlbeck, F. Lynen, Eur. J. Biochem. 1975, 55, 561-571.
[65] aG.-B. Kresze, L. Steber, D. Oesterhelt, F. Lynen, Eur. J. Biochem. 1977, 79, 181-190; bD.
Oesterhelt, H. Bauer, G.-B. Kresze, L. Steber, F. Lynen, Eur. J. Biochem. 1977, 79, 173-180.
[66] K. A. Abrahams, C. Chung, S. Ghidelli-Disse, J. Rullas, M. J. Rebollo-López, S. S. Gurcha, J.
A. G. Cox, A. Mendoza, E. Jiménez-Navarro, M. S. Martínez-Martínez, M. Neu, A. Shillings, P.
Homes, A. Argyrou, R. Casanueva, N. J. Loman, P. J. Moynihan, J. Leliévre, C. Selenski, M.
Axtman, L. S. Kremer, M. Bantscheff, I. Angulo-Barturen, M. C. Izquierdo, N. C. Cammack, G.
Drewes, L. Ballell, D. Barros, G. S. Besra, R. H. Bates, Nat. Commun. 2016, 7, 12581-12593.
[67] S. R. Luckner, C. A. Machutta, P. J. Tonge, C. Kisker, Structure 2009, 17, 1004-1013.
[68] A. Bhatt, V. Molle, G. S. Besra, W. R. Jacobs, Jr., L. Kremer, Mol. Microbiol. 2007, 64, 1442-
1454.
[69] W. Lee, B. Engels, J. Phys. Chem. B 2013, 117, 8095-8104.
[70] B. Sedgwick, J. W. Cornforth, Eur. J. Biochem. 1977, 75, 465-479.
[71] N. Blaquiere, D. G. Shore, S. Rousseaux, K. Fagnou, The Journal of Organic Chemistry 2009,
74, 6190-6198.
[72] W. Lee, S. R. Luckner, C. Kisker, P. J. Tonge, B. Engels, Biochemistry 2011, 50, 5743-5756.
[73] aK. S. Gajiwala, S. Margosiak, J. Lu, J. Cortez, Y. Su, Z. Nie, K. Appelt, FEBS Lett. 2009,
583, 2939-2946; bM. Moche, G. Schneider, P. Edwards, K. Dehesh, Y. Lindqvist, J. Biol.
Chem. 1999, 274, 6031-6034; cJ. G. Olsen, A. Kadziola, P. Von Wettstein-Knowles, M.
Siggaard-Andersen, S. Larsen, Structure 2001, 9, 233-243; dP. von Wettstein-Knowles, J. G.
Olsen, K. A. McGuire, A. Henriksen, FEBS Journal 2006, 273, 695-710; eJ. Wang, S. M.
Soisson, K. Young, W. Shoop, S. Kodali, A. Galgoci, R. Painter, G. Parthasarathy, Y. S. Tang,
R. Cummings, S. Ha, K. Dorso, M. Motyl, H. Jayasuriya, J. Ondeyka, K. Herath, C. Zhang, L.
Hernandez, J. Allocco, A. Basilio, J. R. Tormo, O. Genilloud, F. Vicente, F. Pelaez, L. Colwell,
S. H. Lee, B. Michael, T. Felcetto, C. Gill, L. L. Silver, J. D. Hermes, K. Bartizal, J. Barrett, D.
Schmatz, J. W. Becker, D. Cully, S. B. Singh, Nature 2006, 441, 358-361.
[74] V. S. Rangan, A. K. Joshi, S. Smith, Biochemistry 2001, 40, 10792-10799.
[75] J.-T. Tsay, W. Oh, T. J. Larson, S. Jackowski, C. O. Rock, J. Biol. Chem. 1992, 267, 6807-
6814.
[76] D. Val, G. Banu, K. Seshadri, Y. Lindqvist, K. Dehesh, Structure 2000, 8, 565-566.
[77] aK. Aritomi, I. Hirosawa, H. Hoshida, M. Shiigi, Y. Nishizawa, S. Kashiwagi, R. Akada, Biosci.
Biotech. Biochem. 2004, 68, 206-214; bJ. Gajewski, F. Buelens, S. Serdjukow, M. Janssen, N.
Cortina, H. Grubmüller, M. Grininger, Nat. Chem. Biol. 2017, 13, 363–365; cJ. Gajewski, R.
Pavlovic, M. Fischer, E. Boles, M. Grininger, Nat. Commun. 2017, 8, 14650.
[78] The PyMOL Molecular Graphics System, Version 2.0, Schrödinger, LLC.
[79] aM. Jenner, J. P. Afonso, H. R. Bailey, S. Frank, A. Kampa, J. Piel, N. J. Oldham, Angew.
Chem. Int. Edit. 2015, 54, 1817-1821; bM. Jenner, S. Frank, A. Kampa, C. Kohlhaas, P.
Pöplau, G. S. Briggs, J. Piel, N. J. Oldham, Angew. Chem. Int. Edit. 2013, 52, 1143-1147.
[80] M. Wallace, C. R. Green, L. S. Roberts, Y. M. Lee, J. L. McCarville, J. Sanchez-Gurmaches,
N. Meurs, J. M. Gengatharan, J. D. Hover, S. A. Phillips, T. P. Ciaraldi, D. A. Guertin, P.

34

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Cabrales, J. S. Ayres, D. K. Nomura, R. Loomba, C. M. Metallo, Nat. Chem. Biol. 2018, 14,
1021-1031.
[81] aR. Bressler, S. J. Wakil, J. Biol. Chem. 1962, 237, 1441-1448; bJ. S. Buckner, P. E.
Kolattukudy, L. Rogers, Arch. Biochem. Biophys. 1978, 186, 152-163; cY. Seyama, H.
Otsuka, A. Kawaguchi, T. Yamakawa, J. Biochem. 1981, 90, 789-797.
[82] aW. Pirson, L. Schuhmann, F. Lynen, Eur. J. Biochem. 1973, 36, 16-24; bS. Smith, A. Stern,
Arch. Biochem. Biophys. 1983, 222, 259-265.
[83] A. Kawaguchi, K. Arai, Y. Seyama, T. Yamakawa, S. Okuda, J. Biochem. 1980, 88, 303-306.
[84] H. Engeser, K. Hubner, J. Straub, F. Lynen, Eur. J. Biochem. 1979, 101, 413-422.
[85] V. C. Joshi, S. J. Wakil, Arch. Biochem. Biophys. 1971, 143, 493-505.
[86] J. Molnos, R. Gardiner, G. E. Dale, R. Lange, Anal. Biochem. 2003, 319, 171-176.
[87] C. W. Liew, M. Nilsson, M. W. Chen, H. Sun, T. Cornvik, Z.-X. Liang, J. Lescar, J. Biol. Chem.
2012, 287, 23203-23215.
[88] V. S. Rangan, S. Smith, Biol. Chem. 1997, 1-5.

Accepted Manuscript
[89] P. Paiva, S. r. F. Sousa, M. J. Ramos, P. A. Fernandes, ACS Catal. 2018, 8, 4860-4872.
[90] F. Lynen, Methods Enzymol. 1969, 14, 17-33.
[91] J. Ziegenhorn, R. Niedermeier, C. Nüssler, F. Lynen, Eur. J. Biochem. 1972, 30, 285-300.
[92] F. Cheng, Q. Wang, M. Chen, F. A. Quiocho, J. Ma, Proteins 2007, 70, 1228-1234.
[93] aC. Y. Lin, S. Smith, J. Biol. Chem. 1978, 253, 1954-1962; bM. Pazirandeh, S. S. Chirala, W.-
Y. Huang, S. J. Wakil, J. Biol. Chem. 1989, 264, 18195-18201.
[94] aN. Singh, S. J. Wakil, J. K. Stoops, J. Biol. Chem. 1984, 259, 3605-3611; bL. J. Libertini, S.
Smith, Arch. Biochem. Biophys. 1979, 192, 47-60; cJ. S. Mattick, J. Nickless, M. Mizugaki, C.-
Y. Yang, S. Uchiyama, S. J. Wakil, J. Biol. Chem. 1983, 258, 15300-15304; dS. Smith, E.
Agradi, L. J. Libertini, K. N. Dileepan, P. Natl. Acad. Sci. USA 1976, 73, 1184-1188.
[95] N. Nicolaides, Science 1974, 186, 19-26.
[96] aJ. Knudsen, S. Clark, R. Dils, Biochem. Bioph. Res. Co. 1975, 65, 921-926; bJ. Knudsen, R.
Dils, Biochem. Bioph. Res. Co. 1975, 63, 780-785.
[97] J. K. Hansen, J. Knudsen, Biochem. J. 1980, 186, 287-294.
[98] aL. J. Libertini, S. Smith, J. Biol. Chem. 1978, 253, 1393-1401; bM. K. Ritchie, L. C. Johnson,
J. E. Clodfelter, C. W. Pemble, IV, B. E. Fulp, C. M. Furdui, S. J. Kridel, W. T. Lowther, J. Biol.
Chem. 2016, 291, 3520-3530; cM. De Renobales, L. Rogers, P. E. Kolattukudy, Arch.
Biochem. Biophys. 1980, 205, 464-477.
[99] aR. J. Foster, R. F. Bonsall, A. J. Poulose, P. E. Kolattukudy, J. Biol. Chem. 1985, 260, 1386-
1389; bJ. Mikkelsen, A. Witkowski, S. Smith, J. Biol. Chem. 1987, 262, 1570-1574; cJ.
Naggert, A. Witkowski, B. Wessa, S. Smith, Biochem. J. 1991, 273, 787-790.
[100] W. Zhang, B. Chakravarty, F. Zheng, Z. Gu, H. Wu, J. Mao, S. J. Wakil, F. A. Quiocho, P.
Natl. Acad. Sci. USA 2011, 108, 15757-15762.
[101] I.-H. Park, J. D. Venable, C. Steckler, S. E. Cellitti, S. A. Lesley, G. Spraggon, A. Brock, J.
Chem. Inf. Model. 2015, 55, 1914-1925.
[102] M. Pazirandeh, S. S. Chirala, S. J. Wakil, J. Biol. Chem. 1991, 266, 1-7.
[103] H. B. Claxton, D. L. Akey, M. K. Silver, S. J. Admiraal, J. L. Smith, J. Biol. Chem. 2009, 284,
5021-5029.
[104] J. R. Whicher, G. Florova, P. K. Sydor, R. Singh, M. Alhamadsheh, G. L. Challis, K. A.
Reynolds, J. L. Smith, J. Biol. Chem. 2011, 286, 22558-22569.
[105] C. W. Pemble, IV, L. C. Johnson, S. J. Kridel, W. T. Lowther, Nat. Struct. Mol. Biol. 2007, 14,
704-709.
[106] E. M. Barnes, Jr., S. J. Wakil, J. Biol. Chem. 1968, 243, 2955-2962.
[107] aE. M. Barnes, Jr., A. C. Swindell, Jr., S. J. Wakil, J. Biol. Chem. 1970, 245, 3122-3128; bW.
M. Bonner, K. Bloch, J. Biol. Chem. 1972, 247, 3123-3133.
[108] A. K. Spencer, A. D. Greenspan, J. E. Cronan, Jr., J. Biol. Chem. 1978, 253, 5922-5926.
[109] aY. J. Choi, S. Y. Lee, Nature 2013, 502, 571-574; bP. Jiang, J. E. Cronan, Jr., J. Bacteriol.
1994, 176, 2814-2821; cX. Lu, H. Vora, C. Khosla, Metab. Eng. 2008, 10, 333-339; dE. J.
Steen, Y. Kang, G. Bokinsky, Z. Hu, A. Schirmer, A. McClure, S. B. del Cardayre, J. D.
Keasling, Nature 2010, 463, 559-562; eT. W. Tee, A. Chowdhury, C. D. Maranas, J. V.
Shanks, Biotechnol. Bioeng. 2014, 111, 849-857.
[110] Y.-C. Lo, S.-C. Lin, J.-F. Shaw, Y.-C. Liaw, J. Mol. Biol. 2003, 330, 539-551.
[111] M. J. Grisewood, N. J. Hernández Lozada, J. B. Thoden, N. P. Gifford, D. Mendez-Perez, H.
A. Schoenberger, M. F. Allan, M. E. Floy, R.-Y. Lai, H. M. Holden, B. F. Pfleger, C. D.
Maranas, ACS Catal. 2017, 7, 3837-3849.
[112] J. Li, U. Derewenda, Z. Dauter, S. Smith, Z. S. Derewenda, Nat. Struct. Biol. 2000, 7, 555-
559.

35

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[113] F. Kovačić, J. Granzin, S. Wilhelm, B. Kojić-Prodić, R. Batra-Safferling, K.-E. Jaeger, PLoS

ONE 2013, 8, e69125-69112.
[114] Y.-C. Lo, S.-C. Lin, J.-F. Shaw, Y.-C. Liaw, Biochemistry 2005, 44, 1971-1979.
[115] F. Jing, D. C. Cantu, J. Tvaruzkova, J. P. Chipman, B. J. Nikolau, M. D. Yandeau-Nelson, P.
J. Reilly, BMC Biochem 2011, 12, 1-16.
[116] T. A. Voelker, A. C. Worrell, L. Anderson, J. Bleibaum, C. Fan, D. J. Hawkins, S. E. Radke, H.
M. Davies, Science 1992, 257, 72-74.
[117] Y. Feng, Y. Wang, J. Liu, Y. Liu, X. Cao, S. Xue, ACS Chem. Biol. 2017, 12, 2830-2836.
[118] A. Stern, B. Sedgwick, S. Smith, J. Biol. Chem. 1982, 257, 799-803.
[119] J. D. Robinson, R. O. Brady, R. M. Bradley, J. Lipid Res. 1963, 4, 144-150.
[120] T. C. Linn, P. A. Srere, J. Biol. Chem. 1980, 255, 10676-10680.
[121] S.-I. Chang, G. G. Hammes, Acc. Chem. Res. 1990, 23, 363-369.
[122] J. C. Bartley, S. Abraham, I. L. Chaikoff, Biochim. Biophys. Acta 1967, 144, 51-60.
[123] F. Lynen, I. Hopper-Kessel, H. Eggerer, Biochem. Z. 1964, 340, 95-124.

Accepted Manuscript
[124] H. F. Hofbauer, F. H. Schopf, H. Schleifer, O. L. Knittelfelder, B. Pieber, G. N. Rechberger, H.
Wolinski, M. L. Gaspar, C. O. Kappe, J. Stadlmann, K. Mechtler, A. Zenz, K. Lohner, O.
Tehlivets, S. A. Henry, S. D. Kohlwein, Dev. Cell 2014, 29, 729-739.
[125] C. K. Schjerling, R. Hummel, J. K. Hansen, C. Borsting, J. M. Mikkelsen, K. Kristiansen, J.
Knudsen, J. Biol. Chem. 1996, 271, 22514-22521.
[126] aD. N. Brindley, S. Matsumura, K. Bloch, Nature 1969, 224, 666-669; bN. Papaioannou, H.-S.
Cheon, Y. Lian, Y. Kishi, ChemBioChem 2007, 8, 1775-1780.
[127] D. W. Brown, T. H. Adams, N. P. Keller, P. Natl. Acad. Sci. USA 1996, 93, 14873-14877.
[128] C. M. H. Watanabe, C. A. Townsend, Chem. Biol. 2002, 9, 981-988.
[129] B. Teichmann, U. Linne, S. Hewald, M. A. Marahiel, M. Bölker, Mol. Microbiol. 2007, 66, 525-
533.
[130] aS. Kikuchi, D. L. Rainwater, P. E. Kolattukudy, Arch. Biochem. Biophys. 1992, 295, 318-326;
bO. Zimhony, C. Vilchéze, W. R. Jacobs, Jr., Journal of Bacteriology 2004, 186, 4051-4055.
[131] aT. S. Hitchman, E. W. Schmidt, F. Trail, M. D. Rarick, J. E. Linz, C. A. Townsend, Bioorg.
Chem. 2001, 29, 293-307; bC. P. Woloshuk, R. Prieto, FEMS Microbiol. Lett. 1998, 160, 169-
176.
[132] N. Elad, S. Baron, Y. Peleg, S. Albeck, J. Grunwald, G. Raviv, Z. Shakked, O. Zimhony, R.
Diskin, Nat. Commun. 2018, 9, 3886.
[133] aR. J. Heath, C. O. Rock, J. Biol. Chem. 1996, 271, 10996-11000; bS. Jackowski, C. O. Rock,
J. Biol. Chem. 1987, 262, 7927-7931; cJ. P. Torella, T. J. Ford, S. N. Lim, A. M. Chen, J. C.
Way, P. A. Silver, P. Natl. Acad. Sci. USA 2013, 110, 11290-11295.
[134] Y. Feng, Y. Zhang, Y. Wang, J. Liu, Y. Liu, X. Cao, S. Xue, Appl. Microbiol. Biot. 2018, 102,
3173-3182.
[135] aP. P. Peralta-Yahya, F. Zhang, S. B. del Cardayre, J. D. Keasling, Nature 2012, 488, 320-
328; bA. J. J. Straathof, Chem. Rev. 2014, 114, 1871-1908.
[136] aM. K. Kang, J. Nielsen, J. Ind. Microbiol. Biot. 2017, 44, 613-622; bE. R. Marella, C.
Holkenbrink, V. Siewers, I. Borodina, Curr. Opin. Biotech. 2018, 50, 39-46; cW. Runguphan, J.
D. Keasling, Metab. Eng. 2014, 21, 103-113; dS. Sarria, N. S. Kruyer, P. P. Peralta-Yahya,
Nat. Biotechnol. 2017, 35, 1158-1166; eY. Zhang, J. Nielsen, Z. Liu, Biotechnol. Bioeng. 2018,
115, 2139-2147; fY. J. Zhou, N. A. Buijs, Z. Zhu, J. Qin, V. Siewers, J. Nielsen, Nat. Commun.
2016, 7, 11709.
[137] aM. Castellana, M. Z. Wilson, Y. Xu, P. Joshi, I. M. Cristea, J. D. Rabinowitz, Z. Gitai, N. S.
Wingreen, Nat. Biotechnol. 2014, 32, 1011-1018; bI. Wheeldon, S. D. Minteer, S. Banta, S. C.
Barton, P. Atanassov, M. Sigman, Nat. Chem. 2016, 8, 299-309.
[138] Y. Ishihama, T. Schmidt, J. Rappsilber, M. Mann, F. U. Hartl, M. J. Kerner, D. Frishman, BMC
Genomics 2008, 9, No pp given.
[139] B. Volkmer, M. Heinemann, PLoS ONE 2011, 6, e23126.
[140] T. Maier, Nat. Chem. Biol. 2017, 13, 344-345.
[141] A. Rittner, K. S. Paithankar, D. J. Drexler, A. Himmler, M. Grininger, Protein Sci. 2019, 28,
414-428.
[142] aA. K. Joshi, A. Witkowski, H. A. Berman, L. Zhang, S. Smith, Biochemistry 2005, 44, 4100-
4107; bC. Leber, N. A. Da Silva, Biotechnol. Bioeng. 2013, 111, 347-358.
[143] M. Fischer, B. Mulinacci, M. Joppe, R. Vollrath, K. Konstantinidis, P. Kötter, L. Ciccarelli, J.
Vonck, D. Oesterhelt, M. Grininger, bioRxiv 2018.
[144] Z. Zhu, Y. J. Zhou, A. Krivoruchko, M. Grininger, Z. K. Zhao, J. Nielsen, Nat. Chem. Biol.
2017, 13, 360-362.
[145] P. Xu, K. Qiao, W. S. Ahn, G. Stephanopoulos, P. Natl. Acad. Sci. USA 2016, 113, 10848-
10853.

36

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

[146] C. Rigouin, M. Gueroult, C. Croux, G. Dubois, V. Borsenberger, S. Barbe, A. Marty, F.

Daboussi, I. André, F. Bordes, ACS Synth. Biol. 2017, 6, 1870-1879.
[147] aK. Bravo-Rodriguez, A. F. Ismail-Ali, S. Klopries, S. Kushnir, S. Ismail, E. K. Fansa, A.
Wittinghofer, F. Schulz, E. Sanchez-Garcia, ChemBioChem 2014, 15, 1991-1997; bF. Wang,
Y. Wang, J. Ji, Z. Zhou, J. Yu, H. Zhu, Z. Su, L. Zhang, J. Zheng, ACS Chem. Biol. 2015, 10,
1017-1025.
[148] aJ. L. Blatti, J. Michaud, M. D. Burkart, Curr. Opin. Chem. Biol. 2013, 17, 496-505; bJ. J.
Thelen, J. B. Ohlrogge, Metab. Eng. 2002, 4, 12-21; cK. Dehesh, Eur. J. Lipid Sci. Tech.
2001, 103, 688-697; dT. P. Durrett, C. Benning, J. B. Ohlrogge, Plant J. 2008, 54, 593-607;
eR. Radakovits, R. E. Jinkerson, A. Darzins, M. C. Posewitz, Eukaryot. Cell 2010, 9, 486-501.
[149] T. Liu, H. Vora, C. Khosla, Metab. Eng. 2010, 12, 378-386.
[150] aT. A. Voelker, H. M. Davies, J. Bacteriol. 1994, 176, 7320-7327; bH. Yan, Z. Wang, F. Wang,
T. Tan, L. Liu, Eng. Life Sci. 2016, 16, 53-59.
[151] C. D. Rutter, S. Zhang, C. V. Rao, Appl. Microbiol. Biot. 2015, 99, 7359-7368.

Accepted Manuscript
[152] L. Fan, J. Liu, K. Nie, L. Liu, F. Wang, T. Tan, L. Deng, Enzyme Microb. Tech. 2013, 53, 128-
133.
[153] aK. Dehesh, A. Jones, D. S. Knutzon, T. A. Voelker, Plant J. 1996, 9, 167-172; bT. A. Voelker,
T. R. Hayes, A. M. Cranmer, J. C. Turner, H. M. Davies, Plant J. 1996, 9, 229-241; cD. S.
Knutzon, T. R. Hayes, A. Wyrick, H. Xiong, H. M. Davies, T. A. Voelker, Plant Physiol. 1999,
120, 739-746; dH. Tjellström, M. Strawsine, J. Silva, E. B. Cahoon, J. B. Ohlrogge, FEBS Lett.
2013, 587, 936-942; eT. A. Voelker, A. Jones, A. M. Cranmer, M. H. Davies, D. S. Knutzon,
Plant Physiol. 1997, 114, 669-677.
[154] aR. Radakovits, P. M. Eduafo, M. C. Posewitz, Metab. Eng. 2011, 13, 89-95; bJ. L. Blatti, J.
Beld, C. A. Behnke, M. Mendez, S. P. Mayfield, M. D. Burkart, PLoS ONE 2012, 7, e42949-
42912; cH. Lin, Y. K. Lee, J. Appl. Phycol. 2017, 29, 2811-2819; dY. Inaba, K. Nakahigashi, T.
Ito, M. Tomita, Phycol. Res. 2017, 65, 94-99; eK. W. Tan, Y. K. Lee, J. Biotechnol. 2017, 247,
60-67.
[155] aA. M. Ruffing, J. Appl. Phycol. 2013, 25, 1495-1507; bX. Liu, J. Sheng, R. Curtiss, III, P. Natl.
Acad. Sci. USA 2011, 108, 6899-6904.
[156] aL. Yuan, T. A. Voelker, D. J. Hawkins, P. Natl. Acad. Sci. USA 1995, 92, 10639-10643; bK.
M. Mayer, J. Shanklin, BMC Plant Biol. 2007, 7, 1-11; cM. T. Facciotti, L. Yuan, Fett-Lipid
1998, 100, 167-172.

37

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Biographical sketches

Christina Heil received her [Link]. degree in chemistry from the Goethe
University Frankfurt. She joined Martin Grininger’s laboratory in 2015 as a
doctoral student to pursue her interest in biomolecular chemistry. Her current
work is focused on the development of new tools for site-directed labeling of
multidomain protein complexes by exploiting genetic code expansion and
bioorthogonal reactions. She studies conformational dynamics and domain-
domain interactions within the mammalian fatty acid biosynthesis.

Accepted Manuscript
Sophia Wehrheim studied biochemistry and graduated from the Goethe
University Frankfurt in 2018. During her master’s studies she conducted
research in the field of protein diffusion dynamics. Recently she joined the lab
of Martin Grininger as Ph.D. student. Her current research is focused on
mitochondrial fatty acid synthesis.

Karthik Paithankar studied Physics in his [Link] and shifted to Biophysics during
his [Link]. After a PhD from University of Leipzig specializing on X-ray
crystallography, he performed studies on radiation damage at University of
Oxford as a research associate in the group of Prof. Elspeth Garman. He then
returned to Germany to join macromolecular beamlines at Helmholtz-Zentrum
Berlin für Materialien und Energie as a beamline scientist. From 2012 he is a
postdoctoral fellow at Goethe University Frankfurt in the group of Martin
Grininger.

Martin Grininger studied chemistry at the Universities Linz and Graz (Austria).
He then worked with Prof. Dieter Oesterhelt at the Max-Planck-Institute of
Biochemistry for his PhD, studying flavin sequestration and storage in Archaea
and Bacteria. After becoming a project group leader at the same institution, he
focused on multidomain fatty acid synthases and contributed to the current
structural understanding of these proteins. He also held a visiting professorship
at the University Vienna in Biological Chemistry. In 2012, he received a
Lichtenberg Professorship of the Volkswagen Foundation hosted by the University Frankfurt. He is co-
speaker of the research focus MegaSyn, which develops rules for engineering megasynthases for
custom compound synthesis.

38

This article is protected by copyright. All rights reserved.

ChemBioChem 10.1002/cbic.201800809

Table of Content
Fatty acid biosynthesis is an iterative process with strict chain length
control. Key players in chain length control are the three catalytic
enzymes acyltransferase (AT), ketosynthase (KS) and thioesterase
(TE), which decide upon chain elongation or termination.
Bioengineering of fatty acid synthases (FASs) can overwrite chain
length control in order to generate platform chemicals for biofuel
production.

Accepted Manuscript
Frontispiece

39

This article is protected by copyright. All rights reserved.