Control Mechanism for Carbon-Chain Length in Polyunsaturated Fatty-Acid Synthases

Article abstract of DOI:10.1002/anie.201900771

Polyunsaturated fatty acids (PUFAs) such as docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) are essential fatty acids. PUFA synthases are composed of three to four subunits and each create a specific PUFA without undesirable byproducts. However, detailed biosynthetic mechanisms for controlling final product profiles have been obscure. Here, the bacterial DHA and EPA synthases were carefully dissected by in vivo and in vitro experiments. In vitro analysis with two KS domains (KS_A and KS_C) and acyl-acyl carrier protein (ACP) substrates showed that KS_A accepted short- to medium-chain substrates while KS_C accepted medium- to long-chain substrates. Unexpectedly, condensation from C₁₈ to C₂₀, the last elongation step in EPA biosynthesis, was catalyzed by KS_A domains in both EPA and DHA synthases. Conversely, condensation from C₂₀ to C₂₂, the last elongation step for DHA biosynthesis, was catalyzed by the KS_C domain in DHA synthase. KS_C domains therefore determine the chain lengths.

Full text of DOI:10.1002/anie.201900771

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Control mechanism for carbon chain length in polyunsaturated
fatty acid synthases
Authors: Tohru Dairi, Shohei Hayashi, Mai Naka, Kenshin Ikeuchi,
Makoto Ohtsuka, Kota Kobayashi, Yasuharu Satoh, Yasushi
Ogasawara, Chitose Maruyama, Yoshimitsu Hamano, and
Tetsuro Ujihara
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: Angew. Chem. Int. Ed. 10.1002/anie.201900771
Angew. Chem. 10.1002/ange.201900771
Link to VoR: http://dx.doi.org/10.1002/anie.201900771
http://dx.doi.org/10.1002/ange.201900771

10.1002/anie.201900771
Angewandte Chemie International Edition
COMMUNICATION
Control mechanism for carbon chain length in polyunsaturated
fatty acid synthases
Shohei Hayashi[b], Mai Naka[b], Kenshin Ikeuchi[b], Makoto Ohtsuka[b], Kota Kobayashi[b], Yasuharu
Satoh[a], Yasushi Ogasawara[a], Chitose Maruyama[c], Yoshimitsu Hamano[c], Tetsuro Ujihara[d], and
Tohru Dairi*[a]
Abstract: Polyunsaturated fatty acids (PUFAs) such as
docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) are
essential fatty acids for humans. Microalgae and some marine
bacteria synthesize these PUFAs with PUFA synthase from acetyl
units. The PUFA synthases are composed of three to four subunits
and each create a specific PUFA without undesirable byproducts even
though the multiple catalytic domains in each huge subunit are very
similar. However, detailed biosynthetic mechanisms for controlling
final product profiles have been obscure. Here, we carefully dissected
bacterial DHA and EPA synthases by in vivo and in vitro experiments.
In vivo analysis showed that β-ketoacyl synthase (KS)/chain length
factor-like domains in “C”-subunits of EPA and DHA synthases
controlled DHA or EPA production. Furthermore, in vitro analysis with
two KS domains in the “A” and “C” subunits (KSA and KSC) and acyl-
ACP substrates showed that KSA accepted short to medium chain
substrates while KSC accepted medium to long chain substrates.
Unexpectedly, condensation from C18 to C20, the last elongation step
in EPA biosynthesis, was catalyzed by KSA domains in both EPA and
DHA synthases. Conversely, condensation from C20 to C22, the last
elongation step for DHA biosynthesis, was catalyzed by the KSC
domain in DHA synthase. Based on our results, we converted a
microalgal DHA synthase into an EPA synthase.
(EPA; C20:5 ω3), docosapentaenoic acid (DPAω6 or DPAω3;
C22:5 ω6 or ω3), and arachidonic acid (ARA; C20:4 ω6) are
essential fatty acids for humans and we ingest them from fish oils.
However, because of the increased demand, commercial
fermentative processes using microalgae, yeasts, and fungi have
been developed to produce DHA, EPA, and ARA[1], respectively.
Polyunsaturated fatty acids (PUFAs, Fig. 1A) such as
docosahexaenoic acid (DHA; C22:6 ω3), eicosapentaenoic acid
[a]
[b]
Dr. Y. Satoh, Dr. Y. Ogasawara, Prof. Dr. T. Dairi*.
Graduate School of Engineering
Hokkaido University,
N13-W8, Kita-ku, Sapporo, 060-8628, Japan.
E-mail: dairi@eng.hokudai.ac.jp
Mr. S. Hayashi, Ms. M. Naka, Mr. K. Ikeuchi, Mr. M. Ohtsuka, Mr. K.
Kobayashi.
Graduate School of Chemical Sciences and Engineering
Hokkaido University,
N13-W8, Kita-ku, Sapporo, 060-8628, Japan.
Dr. C. Maruyama, Prof. Dr. Y. Hamano.
Department of Bioscience
Fukui Prefectural University
Fukui, 910-1195, Japan.
Dr. T. Ujihara.
Kyowa Hakko Bio Co. Ltd.
1-6-1, Ohtemachi, Chiyoda-ku, Tokyo, 100-8185, Japan.
Figure 1. (A): Chemical structures of docosahexaenoic acid (DHA;
[c]
[d]
C22:6
ω3),
eicosapentaenoic
acid
(EPA;
C20:5
ω3),
docosapentaenoic acid (DPAω6 and DPAω3; C22:5 ω6 or ω3),
arachidonic acid (ARA; C20:4 ω6), and eicosatetraenoic acid (ETA;
C20:4 ω3). (B): Domain organizations of PUFA synthases. KS: β-
ketoacyl synthase, MAT: malonyl CoA transacylase, ACP: acyl carrier
protein, KR: β-ketoacyl reductase, DHPKS: polyketide synthase-type
dehydratase, AT: acyltrasferase, CLF: chain length factor-like
domain, DHFabA: FabA-type dehydratase, ER: enoyl reductase.
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201900771
Angewandte Chemie International Edition
COMMUNICATION
Figure 2. Proposed biosynthetic pathway of EPA and DHA synthases. The catalytic cycle in PUFA biosynthesis is schematically shown (gray
circle).
In microorganisms, PUFAs are biosynthesized by two
pathways, the aerobic desaturase/elongase pathway and the
anaerobic PUFA synthase pathway. In the former pathway,
specific desaturases and elongases catalyze individual
desaturation and elongation steps to synthesize PUFAs from oleic
acid (C18:1 ω9)[1c]. In the latter pathway, PUFA synthases
composed of huge enzyme complexes with multiple catalytic
domains synthesize PUFAs[2].
The biosynthetic genes for DHA, EPA, and ARA have been
identified in marine bacteria; Moritella marina, Photobacterium
profundum, and Aureispira marina, respectively (Fig. 1B)[3]. In
contrast, only DHA producers have been isolated from microalgae
This article is protected by copyright. All rights reserved.

10.1002/anie.201900771
Angewandte Chemie International Edition
COMMUNICATION
such as Schizochytrium sp. and Aurantiochytrium sp.[4]. The
PUFA synthases identified in these organisms are composed of
three or four subunits and possess similar domain structures
including acyltransferase (AT), acyl carrier protein (ACP), malonyl
CoA transacylase (MAT), β-ketoacyl synthase (KS), ketoacyl
reductase (KR), dehydratase (DH), and enoyl reductase (ER)
domains in a manner similar to fatty acid synthases (FASs) (Fig.
2). However, PUFA synthases have unique features distinct from
FASs. First, PUFA synthases have multi-tandem ACP domains
and we previously showed that PUFA productivities increased
depending on the number of ACPs without profile changes[5].
Second, PUFA synthases possess two-types of dehydratase,
polyketide synthase (PKS)-type dehydratase (DHPKS) and FabA-
type dehydratase (DHFabA). We recently showed that ARA and
EPA synthases utilized DHPKS and DHFabA depending on the
carbon chain length for introduction of saturation or cis double
bonds on growing acyl chains[6]. Third, PUFA synthases contain
two ketoacyl synthase domains, KSA and KSC, in subunit “A” and
“C”, respectively, and a chain length factor (CLF)-like domain,
which has similarity to the KS domain but has no conserved active
residues (Fig. S1), exists next to the KSC domain. To date,
however, there are few reports on the function of these KS
domains. Here, we dissected the PUFA biosynthetic machineries
by in vivo and in vitro experiments using the EPA and DHA
synthases. Moreover, we applied the obtained results to a
microalgal DHA synthase and succeeded in conversion into an
EPA synthase.
To find clues to how the enzymes control the carbon chain
length of the final products, we employed in vivo gene exchange
assays. The EPA synthase genes (epa) of P. profundum and DHA
synthase genes (dha) of M. marina, both of which comprise four
genes (A to D), were cloned into pDuet vectors with T7 promoters
and introduced into E. coli for evaluation of PUFA profiles. The
transformant harboring epa-ABCD genes produced EPA as the
main product concomitant with significant amounts of DPAω3.
The transformant harboring dha-ABCD genes also produced DHA
as the main product with small amounts of eicosatetraenoic acids
(ETA) (Figs. 3A and 3B). Because P. profundum and M. marina
were shown to produce only EPA and DHA[7], DPAω3 and ETA
were suggested to be by-products produced by the heterologous
expression system. As shown in Fig. 1B, Epa-ABCD and Dha-
ABCD have very similar domain structures, each EPA synthase
gene was replaced with the corresponding DHA synthase gene.
When epa-A, -B or -D was replaced with the corresponding DHA
gene, the PUFA profiles (EPA production) were almost the same
as that of the transformant expressing epa-ABCD, although the
PUFA productivities tended to decrease. By replacing epa-C with
dha-C, however, the major product was changed to DHA (Fig. 3A).
Similarly, EPA was produced by replacing dha-C with epa-C (Fig.
3B). The same result was also obtained with the EPA synthase
Figure 3. GC-MS analysis traced at m/z 79[5] of products produced in the gene replacement assay. (A) Gene replacement of epa genes with
dha genes (from 2nd to 5th). Standards (top) and epa-ABCD (bottom). (B) Gene replacement of dha genes with epa genes (from 2nd to 5th).
Standards (top) and dha-ABCD (bottom). (C) Co-expression of epa-C-dha-C-chimera1 with epa-ABD (middle) or dha-ABD (bottom). Standards
(top). (D) Co-expression of dha-C-epa-C-chimera2 with dha-ABD (middle) or epa-ABD (bottom). Standards of methyl esters of ARA,
eicosatetraenoic acid (ETA, C20:4 ω3), EPA (0.16 mM (A), 0.16 mM (B), 0.03 mM (C), 0.3 mM (D)), DPAω3, and DHA (top) were injected.
This article is protected by copyright. All rights reserved.

10.1002/anie.201900771
Angewandte Chemie International Edition
COMMUNICATION
gene (epas-C) of Shewanella oneidensis MR-1 (Fig. S2). These
results suggested that the “C” gene is the key factor for controlling
the carbon chain length of final products.
KSA), Epa-KSC-CLF (Epa-KSC), Dha-KSA-MAT (Dha-KSA), and
Dha-KSC-CLF (Dha-KSC), were successfully obtained as soluble
enzymes (Fig. S5). We prepared acyl-ACP substrates using the
previously constructed single apo-ACP of EPA synthase of
Shewanella oneidensis MR-1 and phosphopantetheinyl
transferase Sfp[6]. The first elongation reaction in EPA and DHA
biosynthesis is perhaps the condensation of malonyl-ACP (1-
ACP) and acetyl-ACP (2-ACP). However, it was recently reported
that malonyl-CoA, but not acetyl-CoA, could be loaded on the five
tandem ACP domains of M. marina[8]. We therefore examined
whether acetyl-ACP (2-ACP) could be accepted by the KS
domains. When acetyl-ACP (2-ACP) and malonyl-ACP (1-ACP)
were incubated with Epa-KSA or Dha-KSA, 3-oxobutyryl-ACP (3-
ACP) was clearly detected by LC-MS, while no product was
formed with the KSC enzymes (Figs. 4A and S6). We then used
only malonyl-ACP (1-ACP) as the substrate and incubated it with
the KSA domain. As shown in Fig. S7, 3-oxobutyryl-ACP (3-ACP)
was clearly detected, suggesting that acetyl-ACP (2-ACP) formed
from malonyl-ACP (1-ACP) via a decarboxylation reaction would
be used for the condensation reaction.
We next constructed chimeric “C” genes to narrow down the
chain length control domain. The dha-C gene has an EcoRI site
in its center region (Fig. S3) that divides the gene into two regions,
the upstream KSC/CLF-like domains and the downstream DHFabA
domains. Therefore, we artificially introduced an EcoRI site into
the corresponding locus of the epa-C gene without amino acid
changes in the translated enzymes. In the case of epa-C-dha-C-
chimera1, EPA was produced by co-expression with dha-ABD
(Fig. 3C). The transformant expressing dha-C-epa-C-chimera2
together with epa-ABD produced both EPA and DHA (Fig. 3D).
These results clearly indicated that the KSC or CLF-like domain
controlled the carbon chain length. To examine which domain is
responsible for this control, we constructed dozens of chimeric
genes in which the dha-C gene and epa-C gene were fused at
different points between the KSC domain and CLF-like domain.
However, all the constructs lost PUFA productivity, suggesting
that the quaternary structures of these regions are important.
Because the functions of the KSA domain in the “A” subunit
and the KSc domain in the “C” subunit are still unclear, we carried
out in vitro experiments with truncated KS enzymes and acyl-ACP
substrates (Fig. S4). We tried to express a truncated enzyme
containing only the KSA or the KSC domain, but no recombinant
enzyme was obtained. Therefore, the former and the latter
domain were co-expressed with a MAT and CLF-like domain,
respectively, and four truncated enzymes, Epa-KSA-MAT (Epa-
We then examined the subsequent elongation steps. When
malonyl-ACP (1-ACP) and butyryl-ACP (4-ACP) or 3-cis
hexenoyl-ACP (6-ACP) were used as substrates, the KSA
enzymes showed high activities to form 3-oxohexanoyl-ACP (5-
ACP) and 5-cis 3-oxooctenoyl-ACP (7-ACP, Figs. 4B and S8)
while the KSC enzymes weakly catalyzed these reactions. We
next examined the functions of the two KS domains during middle
to late biosynthetic stages. However, none of the predicted
Figure 4. Deconvoluted MS spectra obtained by HPLC-ESI-TOF-MS analysis of in vitro KS
reaction mixtures. (A) Acetyl-ACP (2-ACP)/malonyl-ACP (1-ACP) with Epa-KSA (middle), with
EPA-KSC (bottom), or without enzyme (top). (B) Butyryl-ACP (4-ACP)/1-ACP with Epa-KSA
(middle), with EPA-KSC (bottom), or without enzyme (top). (C) 4,7-cis Decadienoyl-ACP (8-ACP)/1-
ACP with Epa-KSA (2nd), with Epa-KSC (3rd), with Dha-KSA (4th), with Dha-KSC (bottom), or without
enzyme (top). (D) 3,6,9,12,15-cis Octadecapentaenoyl-ACP (10-ACP)/1-ACP, with Epa-KSA (2nd),
with Epa-KSC (3rd), with Dha-KSA (4th), with Dha-KSC (bottom), or without enzyme (top).
This article is protected by copyright. All rights reserved.

10.1002/anie.201900771
Angewandte Chemie International Edition
COMMUNICATION
substrates were commercially available and their chemical
synthesis was also difficult. Because 4,7-cis decadienoyl-ACP (8-
ACP) and 3,6,9,12,15-cis octadecapentaenoyl-ACP (10-ACP)
were the sole substrates that we could chemically synthesize (Fig.
S4), we also performed in vivo analysis besides the in vitro assay.
When 4,7-cis decadienoyl-ACP (8-ACP) and malonyl-ACP (1-
ACP) were used as substrates, Epa-KSA and Epa-KSC showed
we randomly introduced a mutation into the region corresponding
to the KSB or CLF-like domain in the AsorfB gene by error-prone
PCR. In random screening, we isolated no EPA producer with the
CLF-like domain as a template even though more than 3,000
mutants were screened. In contrast, we successfully obtained a
KSB domain mutant designated K01 that produced EPA as a
minor product (EPA/DHA = 0.18) and possessed one mutation
(F230L), after screening 2,000 mutants. Therefore, we used the
K01 gene as a template and obtained a mutant (K02) producing
increased EPA (EPA/DHA = 0.30) with an additional mutation
(N65L). Using the same strategy, a mutant (K03) that had an
additional mutation (I231T) and produced EPA as the major
product (EPA/DHA = 1.07) was successfully isolated (Fig. S11).
Thus, we were able to alter the product profiles through three
amino acid substitutions, N65L, F230L, and I231T.
almost the same activities and formed
a
6,9-cis 3-
oxododecadienoyl-ACP (9-ACP, Fig. 4C). Similarly, both Dha-
KSA and Dha-KSC showed the same activities (Fig. 4C). To get
more information about the role of the KSC domain, we then
constructed two mutated enzymes, Epa-C-KS0 and Dha-C-KS0, in
which the catalytic Cys residues in the KSC domains were mutated
to Ala (Fig. S1), and co-expressed each with epa-ABD. In the
case of epa-C-KS0 expression, α-linoleic acid (ALA; C18:3 ω3)
and 7,10,13-cis hexadecatrienoic acid (C16:3 ω3) were produced
as major and minor products, respectively (Figs. S9 and S10).
Similarly, DHA production was completely abolished and ALA
was produced as the major product with dha-C-KS0 (Fig. S9).
Considering that ALA was produced as the major product in both
cases, the products including ALA were shunt products probably
formed by chain elongation from C12:3 ω3 by the KSA domain.
Therefore, the KSC domain was suggested to catalyze the intrinsic
chain elongation during the middle biosynthetic stage.
We next carried out an in vitro experiment with 3,6,9,12,15-
cis octadecapentaenoyl-ACP (10-ACP) as the substrate to
investigate the final biosynthetic reaction of EPA. When Epa-KSA
and Dha-KSA were used as catalysts, the estimated 5,8,11,14,17-
cis 3-oxoeicosapentaenoyl-ACP (11-ACP) was detected by LC-
MS while Epa-KSC and Dha-KSC showed no activity (Fig. 4D),
indicating that the KSA domain again participated in the last chain
elongation in EPA biosynthesis. To investigate the final chain
elongation in DHA biosynthesis, we employed a combination
enzyme assay (Fig. 5A) because preparation of the intrinsic
substrate, 2,5,8,11,14,17-cis eicosahexaenoyl-ACP (12-ACP),
was difficult. After addition of Epa-KR-DHPKS into the
abovementioned reaction mixture using 3,6,9,12,15-cis
octadecapentaenoyl-ACP (10-ACP) and Dha-KSA, a product
whose molecular weight was identical to the estimated product,
3-hydroxy-5,8,11,14,17-cis eicosapentaenoyl-ACP (13-ACP),
was detected. By further addition of Epa-DHFabA and Dha-KSC, a
plausible 2,4,7,10,13,16,19-docosaheptaenoyl-ACP (14-ACP,
Fig. 5B) was also detected, suggesting that the KSC domain in
Dha-C catalyzed the chain elongation from C20 to C22.
As mentioned above, the difference between DHA and EPA
biosynthesis was brought about by the substrate specificity of the
KSC/CLF-like domain against C20-ACP substrates. Only the KSC
domain of DHA synthase catalyzed the chain elongation, although
the corresponding amino acid sequence in EPA synthase is very
similar. This fact prompted us to try to convert a microalgal DHA
biosynthetic enzyme into an EPA biosynthetic enzyme. We used
the DHA biosynthetic enzymes AsOrfABC of Aurantiochytrium sp.
OH4, which has been used as a practical DHA producer and
produced 44 g/L DHA[4b]. For AsorfA, we utilized SsorfA of
Schizochytrium sp. because we were unable to clone AsorfA,
probably due to its large size, and because AsorfA and SsorfA
have the same domain structures. Three DHA biosynthetic genes
were heterologously expressed in E. coli and co-production of
DHA (67%) and DPAω6 (33%) was confirmed (Fig. S11). Then,
Figure 5. In vitro combination reactions. (A) Reaction scheme of in
vitro combination reactions. (B) HPLC analysis (UV 210 nm) of in vitro
combination reactions using 3,6,9,12,15-cis octadecapentaenoyl-
ACP (10-ACP), Dha-KSA, and Epa-KR-DHPKS (top), plus Epa-DHFabA
and Dha-KSC (bottom).
We recently showed that ARA and EPA synthases utilized
DHPKS and DHFabA depending on the carbon chain length for
introduction of saturation or cis double bonds on growing acyl
chains[6]. As for the control mechanism of the carbon chain length,
Orikasa et al., reported that “B” subunits, Dha-B and Epa-B, were
key enzymes for determining the final product, DHA or EPA[9].
However, in this study, we unveiled the control mechanism of the
carbon chain length, especially the mechanism involved in
creating DHA and EPA.
The condensation of 3,6,9,12,15-cis octadecapentaenoyl-
ACP (10-ACP), the last condensation in EPA biosynthesis, was
catalyzed by the KSA domain in both DHA and EPA biosynthesis.
In EPA biosynthesis, a 5,8,11,14,17-cis 3-oxoeicosapentaenoyl-
ACP (11-ACP) intermediate would be released after the reactions
catalyzed by the KR, DHPKS, and ER domains. Conversely,
2,5,8,11,14,17-cis eicosahexaenoyl-ACP (12-ACP) formed by KR
and DHFabA domains would be used as the substrate of the KSC
domain to form 4,7,10,13,16,19-cis docosahexaenoyl-ACP (15-
ACP) in DHA biosynthesis (Fig. 2).
We also converted the microalgal DHA synthase to EPA
synthase based on the obtained results. The K03 mutant
possessing triple mutations, N65L, F230L, and I231T, produced
EPA as the major product. Although we cannot provide a
mechanism for substrate recognition, the F230 and I231 residues
This article is protected by copyright. All rights reserved.

10.1002/anie.201900771
Angewandte Chemie International Edition
COMMUNICATION
F. McCord, B. D. Tyreus, E. N. Jackson, Q. Zhu, Nat. Biotechnol. 2013,
31, 734–740; c) E. Sakuradani, A. Ando, S. Shimizu, J. Ogawa, J. Biosci.
Bioeng. 2013, 116, 417–422.
J. G. Metz, P. Roessler, D. Facciotti, C. Levering, F. Dittrich, M. Lassner,
R. Valentine, K. Lardizabal, F. Domergue, A. Yamada, K. Yazawa, V.
Knauf, J. Browse, Science 2001, 293, 290–293.
a) N. Morita, M. Tanaka, H. Okuyama, Biochem. Soc. Trans. 2000, 28,
943–945; b) E. E. Allen, D. H. Bartlett, Microbiology 2002, 148, 1903–
1913; c) T. Ujihara, M. Nagano, H. Wada, S. Mitsuhashi, FEBS Lett.
2014, 588, 4032–4036.
in the mutated enzyme were suggested to be located in an α-helix
forming a cavity with the catalytic Cys residue by model building
studies[10] with the crystal structure of mammalian fatty acid
synthase[11] (PDB, 2VZ8), which has 25% identity with our
enzyme (Fig. S12). Therefore, the helix region would be essential
for strict substrate recognition.
[2]
[3]
Acknowledgements
[4]
a) X. Song, X. Zang, X. Zhang, J. Oleo Sci. 2015, 64, 197–204; b)
WO2014112574: Docosahexaenoic acid producing microorganism and
use thereof.
This study was supported in part by Grants-in-Aid for Research
on Innovative Areas from MEXT, Japan (JSPS KAKENHI Grant
Number 16H06452 to T.D.) and Grants-in-Aid for Scientific
Research from JSPS (15H03110 and 18H03937) to T.D. We
thank Robbie Leiws, MSc, from Edanz Group for editing a draft
of this manuscript.
[5]
[6]
[7]
S. Hayashi, Y. Satoh, T. Ujihara, Y. Takata, T. Dairi, Sci. Rep. 2016, 6,
35441.
S. Hayashi, Y. Satoh, Y. Ogasawara, C. Maruyama, Y. Hamano, T.
Ujihara, T. Dairi. Angew. Chem. Int. Ed. DOI: 10.1002/anie.201812623.
a) E. E. Allen, D. Facciotti, D. H. Bartlett. Applied. Environ. Microbiol.
1999, 65, 1710–1720; b) K. B. Kautharapu, J. Rathmacher, L. R. Jarboe.
Appl. Microbiol. Biotechnol. 2013, 97, 2859–2866.
[8]
[9]
O. Santin, G. Moncalian, J. Biol. Chem. 2018, 293, 12491–12501.
Y. Orikasa, M. Tanaka, S. Sugihara, R. Hori, T. Nishida, A. Ueno, N.
Morita, Y. Yano, K. Yamamoto, A. Shibahara, H. Hayashi, Y. Yamada,
R. Yu, K. Watanabe, H. Okuyama, FEMS Microbiol. Lett. 2009, 295,
170–176.
Keywords: polyunsaturated fatty acid synthase • β-ketoacyl
synthase • biosynthesis • carbon chain length
[1]
a) R. J. Winwood, OCL. 2013, 20, D604; b) Z. Xue, P. L. Sharpe, S.
Hong, N. S. Yadav, D. Xie, D. R. Short, H. G. Damude, R. A. Rupert, J.
E. Seip, J. Wang, D. W. Pollak, M. W. Bostick, M. D. Bosak, D. J. Macool,
D. H. Hollerbach, H. Zhang, D. M. Arcilla, S. A. Bledsoe, K. Croker, E.
[10]
[11]
L. A. Kelley, S. Mezulis, C. M. Yates, M. N. Wass, M. J. E. Sternberg,
Nat. Protoc. 2015, 10, 845–858.
T. Maier, M. Leibundgut, N. Ban, Science 2008, 321, 1315–1322.
This article is protected by copyright. All rights reserved.

10.1002/anie.201900771
Angewandte Chemie International Edition
COMMUNICATION
Entry for the Table of Contents
COMMUNICATION
Eicosapentaenoic acid (EPA; C20) and
docosahexaenoic acid (DHA; C22)
synthase complexes possess two β-
ketoacyl synthase domains (KSA and
KSC). We revealed that KSA and KSC
function at early and late biosynthetic
steps, respectively. In the last steps,
however, KSA catalyzed C18 to C20
elongation in both EPA and DHA
biosynthesis while KSC catalyzed C20
to C22 elongation in DHA biosynthesis,
showing that KSC domains determine
the chain length.
Shohei Hayashi, Mai Naka, Kenshin
Ikeuchi, Makoto Ohtsuka, Kota
Kobayashi, Yasuharu Satoh, Yasushi
Ogasawara, Chitose Maruyama,
Yoshimitsu Hamano, Tetsuro Ujihara,
and Tohru Dairi*
Page No. – Page No.
Control mechanism for carbon chain
length in polyunsaturated fatty acid
synthases
This article is protected by copyright. All rights reserved.