Uncovering the formation and selection of benzylmalonyl-CoA from the biosynthesis of splenocin and enterocin reveals a versatile way to introduce amino acids into polyketide carbon scaffolds

Article abstract of DOI:10.1021/jacs.5b00728

Selective modification of carbon scaffolds via biosynthetic engineering is important for polyketide structural diversification. Yet, this scope is currently restricted to simple aliphatic groups due to (1) limited variety of CoA-linked extender units, which lack aromatic structures and chemical reactivity, and (2) narrow acyltransferase (AT) specificity, which is limited to aliphatic CoA-linked extender units. In this report, we uncovered and characterized the first aromatic CoA-linked extender unit benzylmalonyl-CoA from the biosynthetic pathways of splenocin and enterocin in Streptomyces sp. CNQ431. Its synthesis employs a deamination/reductive carboxylation strategy to convert phenylalanine into benzylmalonyl-CoA, providing a link between amino acid and CoA-linked extender unit synthesis. By characterization of its selection, we further validated that AT domains of splenocin, and antimycin polyketide synthases are able to select this extender unit to introduce the phenyl group into their dilactone scaffolds. The biosynthetic machinery involved in the formation of this extender unit is highly versatile and can be potentially tailored for tyrosine, histidine and aspartic acid. The disclosed aromatic extender unit, amino acid-oriented synthetic pathway, and aromatic-selective AT domains provides a systematic breakthrough toward current knowledge of polyketide extender unit formation and selection, and also opens a route for further engineering of polyketide carbon scaffolds using amino acids.

Full text of DOI:10.1021/jacs.5b00728

Subscriber access provided by CARLETON UNIVERSITY
Article
Uncovering the Formation and Selection of Benzylmalonyl-CoA from
the Biosynthesis of Splenocin and Enterocin Reveals a Versatile
Way to Introduce Amino Acids into Polyketide Carbon Scaffolds
Chenchen Chang, Rong Huang, Yan Yan, Hongmin Ma, Zheng
Dai, Benying Zhang, Zixin Deng, Wen Liu, and Xudong Qu
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b00728 • Publication Date (Web): 12 Mar 2015
Downloaded from http://pubs.acs.org on March 16, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Uncovering the Formation and Selection of Benzylmalonyl‐CoA from
the Biosynthesis of Splenocin and Enterocin Reveals a Versatile Way
to Introduce Amino Acids into Polyketide Carbon Scaffolds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chenchen Chang,§,† Rong Huang,§,† Yan Yan,§,‡ Hongmin Ma,† Zheng Dai,† Benying Zhang,† Zixin
Deng,† Wen Liu,‡ and Xudong Qu*,†
†Key Laboratory of Combinatorial Biosynthesis and Drug Discovery (Wuhan University), Ministry of Education, Wuhan
University School of Pharmaceutical Sciences, 185 Donghu Rd., Wuhan 430071, China.
‡State Key Laboratory of Bioorganic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry, Chinese
Academy of Sciences, 345 Lingling Road, Shanghai 200032, China.
ABSTRACT: Selective modification of carbon scaffolds via biosynthetic engineering is important for polyketide structural diversi-
fication. Yet, this scope is currently restricted to simple aliphatic groups due to 1) limited variety of CoA-linked extender units,
which lack aromatic structures and chemical reactivity, and 2) narrow acyltransferase (AT) specificity, which is limited to aliphatic
CoA-linked extender units. In this report, we uncovered and characterized the first aromatic CoA-linked extender unit benzylmalo-
nyl-CoA from the biosynthetic pathways of splenocin and enterocin in Streptomyces sp. CNQ431. Its synthesis employs a deamina-
tion/reductive carboxylation strategy to convert phenylalanine into benzylmalonyl-CoA, providing a link between amino acid and
CoA-linked extender unit synthesis. By characterization of its selection, we further validated that AT domains of splenocin and
antimycin polyketide synthases are able to select this extender unit to introduce the phenyl group into their dilactone scaffolds. The
biosynthetic machinery involved in the formation of this extender unit is highly versatile and can be potentially tailored for tyrosine,
histidine and aspartic acid. The disclosed aromatic extender unit, amino acid-oriented synthetic pathway, and aromatic-selective AT
domains provides a systematic breakthrough towards current knowledge of polyketide extender unit formation and selection, and
also opens a route for further engineering of polyketide carbon scaffolds using amino acids.
Polyketides are a large family of structurally diverse natural
products possessing various important pharmacological activi-
ties.1 Indeed, many of these secondary metabolites have been
used for treatment of human, animal or plant diseases. How-
ever, the synthetic accessibility of polyketide scaffolds, in the
attempt to improve pharmacological performance, is made
difficult due to their inert reactivity and structural complexity.
Thus, alternative methods to modify polyketide structures,
such as through biosynthetic engineering, can be a more effi-
cient route. For example, the acyltransferase (AT) substrate
specificity of polyketide synthase (PKS) can by altered
through the use of domain swapping or site-directed mutagen-
esis that allows for the incorporation of different extender
units into the polyketide carbon scaffold.2 However, the ability
to make changes to the polyketide scaffold via this method is
currently limited by both the extender unit diversity and AT
specificity.
For extender units, there are two varieties that can be select-
Figure 1. Structures of natural polyketide extender units. Ami-
nomalonyl-ACP (1), dichloropyrrolypropylmalonyl-ACP (2), and
ed by PKS: the CoA-linked and acyl carrier protein (ACP)-
linked (Figure 1).3 The most commonly observed CoA-linked
benzylmalonyl-CoA (3) are derived from amino acids. Biosyn-
extender units, malonyl-CoA and alkylmalonyl-CoAs, are
thetic details of benzylmalonyl-CoA (3) are further elucidated in
synthesized by either α-carboxylation of acyl-CoA,3a ligation
this study.
of malonate with CoA4, or reductive carboxylation of α, β-
unsaturated acyl-CoA.3b,3e-k ACP-linked extender units, on the
Cα-side groups mostly contain simple aliphatics with limited
other hand, are less prevalent and only used by few PKS sys-
tems.3a,3c,d,3j Substrates for CoA-linked extender units mostly
originate from fatty acid metabolism. As a consequence, their
chemical reactivity and lack aromatic structures. In addition,
the specificity of AT domains employ extender units as sub-
strates is also limited in their specificity for aliphatic extender
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
1
2
3
units. Thus, due to limited extender unit diversity and narrow
retro biosynthetic analysis wherein we dissected each mole-
4
5
6
7
8
9
AT specificity, the current range of chemical space that can be
introduced into polyketide scaffolds through extender unit
selection is mostly restricted to simple aliphatics. Although in
vitro extender unit synthesis5, or utilizing synthetic acyl-N-
acetylcysteamine (acyl-NAC) thioesters as extender unit sur-
rogates6 can somewhat alleviate this predicament, their prepa-
ration cost, and low incorporation efficiency prevent their
widespread utility.
cule into its corresponding extender units to determine wheth-
er it originated from an amino acid. From this analysis, we
identified one group of natural products termed splenocins
(SPNs, Figure 2). SPNs, produced by the marine bacteria S. sp.
CNQ431, have potent anti-inflammatory activities.10 Interest-
ingly, they share an identical dilactone scaffold with antimy-
cins (ANTs, Figure 2), but have an unusual benzyl group in-
stead of alkyl groups at C7. In ANTs the C7-alkyls originate
from extender unit that consist of alkylmalonyl-CoAs and are
introduced by the AT domain of PKS AntD.11 The resem-
blance of SPNs and ANTs, indicated to us that they may de-
rive from similar biosynthetic pathways. Because a benzyl
group is part of the scaffold in SPNs, this further led us to
speculate that both a novel aromatic extender unit ben-
zylmalonyl-CoA (3) and an aromatic-selective AT domain
might be involved in the installation of a benzyl group into
SPNs. Additionally, in living organisms, benzyl compounds
are often derived from the aromatic amino acid phenylalanine,
suggesting that a novel route to convert phenylalanine into the
assumed CoA-linked extender unit is involved.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Like fatty acids, amino acids are the other major carbon
source that are widely used as biosynthetic building blocks for
peptides and alkaloids synthesis. In polyketide assembly their
involvement is limited, where only serine and proline are used
in the synthesis of aminomalonyl-ACP (1) and dichloropyrrol-
ypropylmalonyl-ACP (2) extender units (Figure 1).3c,3j Unlike
fatty acids, amino acids are rich in functional groups (e.g.
amino and thiol groups) that can further serve as chemical
handles for semi-synthesis. Moreover, aromatic structures
present in certain amino acids can be further diversified by
available biosynthetic machinery that includes tryptophan
halogenase7 and phenylalanine hydroxylase8. The ability to
biosynthetically incorporate and subsequently modify aro-
matic groups on polyketide scaffolds could help to enhance a
molecule’s bioavailability and bioactivity due to effectively
increasing lipophilicity and/or form π-π stacking interactions
with the binding site of a target protein.9. Therefore, finding a
versatile strategy to covert amino acids into the more com-
monly used CoA-extender units would be highly desirable.
Accordingly, based on in vivo and in vitro characterization, we
herein report the formation and selection of an aromatic ex-
tender unit benzylmalonyl-CoA (3) (Figure 1) from the bio-
synthetic pathways of splenocin (SPN) and enterocin (ENC)
(Figure 2). We found that this pathway employs a deamina-
tion/reductive carboxylation strategy to convert phenylalanine
into a CoA-linked extender unit, and then it utilizes an aro-
matic-selective AT domain to select benzylmalonyl-CoA (3)
for incorporation of a phenyl ring into the dilactone scaffold of
SPNs. This versatile extender unit synthetic pathway and the
broadly-selective AT described here, set the solid foundation
for further expansion of polyketide structural diversity through
the introduction of additional amino acids.
Cloning the Splenocin Biosynthetic Gene Cluster. To pro-
vide further detail into the biosynthetic mechanism of SPNs,
we first set out to clone the SPN biosynthetic gene cluster. A
fosmid library of CNQ431 was constructed and screened by
PCR amplification according to the nonribosomal peptide
synthase (NRPS)/PKS gene antC of the ANT biosynthetic
pathway from S. sp. NRRL 2288.11a Sequencing the positive
fosmid allowed us to identify a cluster of 16-genes (Figure 3A
and Table 1). With the exception of a redundant transposase
gene (orf1), this cluster is highly conserved in sequence (~95 %
overall identity) and organization to that of the S-form ants,11a
suggesting that they share a similar biosynthetic mechanism.
Further bioinformatic analysis of the genes within this locus
revealed three genes (spnC, D and M) that encode a hybrid
NRPS/PKS system putatively for dilactone scaffold formation,
one gene (spnE) that encodes a crotonyl-CoA reductive car-
boxylase (CCR) putatively for synthesizing extender units, and
eleven other genes that putatively encode the starter unit 3-
formamidosalicyl-SpnG (spnF-L, N and O), C8 acylation
(spnB) and pathway regulation (spnA). To correlate it with
SPN production, the nrps/pks spnC was further selected for
gene inactivation. LC-HRMS analysis of mutant strain fer-
mentation confirmed it completely lost the ability to produce
SPN-C (in our hands, we only observe SPN-C in the fermenta-
tion of CNQ431) as well as a few accompanied ANTs (A1-A5,
A10 and A21, Figure 4 trace II and Figure S1), hence estab-
lishing this gene cluster is indeed responsible for encoding
both of the SPNs and ANTs.
Based on the available model of ANT biosynthesis,11-12 we
next proposed the assembly logic of spn (Figure 3B). Trypto-
phan is first converted into the starter unit 3-formamidosalicyl-
SpnG by successive participation of SpnN, F, G, H-L and O,
and then sequentially condensed with L-threonine and py-
ruvate by the NRPS SpnC to form a peptidyl intermediate.
SpnD, a PKS containing ketosynthase (KS), AT, ACP, and a
thioesterase (TE) domains, selects benzylmalonyl-CoA (3) or
alkylmalonyl-CoAs as the extender unit and catalyzes the con-
densation with the peptidyl intermediate to afford a C2-
extension. The resulting polyketide-peptidyl intermediates are
further reduced by SpnM at C8, and then cyclized by SpnD-
TE to form the dilactone scaffold. Finally, the acyltransferase
Figure 2. Structures of SPNs, ANTs and ENC. R1, R3 are hy-
drogen or acyl groups, and R2 denotes alkyl groups. SPN-C
(R1= COCH(CH3)C2H5) and SPN-J (R1= H) are the focus of
this study. For details of ANTs concerned in this study please
see Figure S1.
RESULTS
Discovering Polyketides with Amino Acid Origin. We in-
itially performed an extensive literature search to find polyke-
tides of amino acid origin. We analyzed compounds using
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Biosynthetic gene clusters and proposed pathway. (A) Gene organization of the spn and ant (NRRL 2288) clusters, the de-
duced functions of the genes are labeled in color and summarized in Table 1. (B) Proposed biosynthesis of SPNs and ANTs in
CNQ431. Domain notation: ACP, acyl carrier protein; C, condensation; A, adenylation; PCP, peptidyl carrier protein; KR, ketoreduc-
tase; KS, ketosynthase; TE, thioesterase.
Table 1. Deduced Functions of ORFs in spn Biosynthetic Gene Cluster (accession number KP719128).
Gene
Size (a.a.)
Protein Homolog and Origin
Identity (%)
Proposed function
1a
2a
A
B
C
D
E
1880
145
173
433
2897
1273
406
113
428
82
CaiC (WP_018469583), S. sp. LaPpAH-202
(WP_026282526), S. sp. CNY228
AntA (AGG37747), S. sp. NRRL 2288
AntB (AGG37748), S. sp. NRRL 2288
AntC (AGG37749), S. sp. NRRL 2288
AntD (AGG37750), S. sp. NRRL 2288
AntE (AGG37751), S. sp. NRRL 2288
(WP_023422630); S. sp. GBA 94-10
AntF (AGG37752), S. sp. NRRL 2288
AntG (AGG37753), S. sp. NRRL 2288
AntH (AGG37754), S. sp. NRRL 2288
AntI (AGG37755), S. sp. NRRL 2288
AntJ (AGG37756), S. sp. NRRL 2288
AntK (AGG37757), S. sp. NRRL 2288
AntL (AGG37758), S. sp. NRRL 2288
AntM (AGG37759), S. sp. NRRL 2288
AntN (AGG37760), S. sp. NRRL 2288
AntO (AGG37761), S. sp. NRRL 2288
(WP_003946509), S. albus
92
83
100
92
95
96
99
93
97
96
97
98
98
98
96
97
96
97
97
98
97
NRPS (incomplete)
Hypothetical protein
Regulator
Acyltransferase
NRPS (C-A-PCP-C-A-KR-PCP)
PKS (KS-AT-ACP-TE)
Crotonyl-CoA reductive carboxylase
Transposase
3a
F
AMP dependent CoA-ligase
ACP
G
H
I
351
64
Phenylacetate-CoA oxygenase, PaaA
Phenylacetate-CoA oxygenase, PaaB
Phenylacetate-CoA oxygenase, PaaC
Phenylacetate-CoA oxygenase, PaaD
Phenylacetate-CoA oxygenase, PaaE
3-oxoacyl-ACP reductase
Tryptophan 2,3-dioxygenase
Esterase
J
262
140
367
258
282
225
798
134
399
K
L
M
N
O
4a
5a
6a
Histidine kinase
(WP_030696821), S. griseus
Dynein regulation protein LC7
ABC transporter
(WP_003946511), S. albus
a orfs beyond spn gene cluster.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
1
2
3
tion at the C7-position on the dilactone scaffold. Amino acid
4
5
6
7
8
9
sequence analysis of the specificity-conferring motifs within
SpnD-AT revealed the presence of smaller amino acid resi-
dues compared to ATs that typically use malonyl-CoA,
methylmalonyl-CoA or ethylmalonyl-CoA (Figure S3), and
suggests the presence of a larger protein cavity that can ac-
commodate a bulky benzyl group. To validate its function, we
overexpressed the entire protein of SpnD in an E. coli BAP-1
strain which contains a Bacillus phosphopantetheinyl transfer-
ase (Sfp) gene to covert the ACP domain of SpnD into its holo
form.14 benzylmalonyl-CoA (3) was further prepared by SpnE
and in situ incubated with the purified holo-SpnD to carry out
the transacylation reaction. After the reaction, SpnD was puri-
fied by SDS-PAGE and subsequently trypsin digested to de-
tect the presence of a benzylmalonyl moiety by LC-HRMS
analysis (Figure 5C). As expected, a 176.0516 Da shift corre-
sponding to the molecular weight of benzylmalonyl moiety
was clearly detected in the phosphopantetheinylated serine
peptide fragment (PVLVEIGPGDSLTK) of ACP. Meanwhile,
we evaluated the change of benzylmalonyl-CoA (3), confirm-
ing that its consumption was consistent with the uploading
activity of SpnD (Figure 5D). These results clearly validated
that SpnD-AT can select benzylmalonyl-CoA (3) as an ex-
tender unit for introducing the benzyl group into the SPN di-
lactone scaffold.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. HPLC analysis of SPN-C and ANTs production in S.
sp. CNQ431 and S. coelicolor recombinant strains. Peaks corre-
sponding to SPN-C (▼) and ANTs (●) are indicated. I) Wild-type
CNQ431; II) mWHU2450 (ΔspnC); III) mWHU2452 (ΔenccH);
IV) mWHU2451 (ΔenccP); V) mWHU2453 (S. coelicolor carry-
ing fosmid pWHU2405); and VI) wide-type S. coelicolor.
SpnB mediates the acylation of the C8 hydroxyl group to
complete the biosynthesis of SPNs and ANTs.
Characterization of Benzylmalonyl-CoA as an Extender
Unit and the selectivity of the AT Domain. In ANTs, the
C7-alkyl groups are introduced by the AntD-AT from extender
unit alkylmalonyl-CoAs which are synthesized through reduc-
tive carboxylation of α, β-unsaturated acyl-CoA precursors by
the CCR protein AntE. We previously established that AntE is
highly promiscuous and able to accept various substrates, in-
cluding cinnamoyl-CoA (4), for conversion into their reduc-
tive carboxylated products.13 In spn, we identified an AntE
homologous protein SpnE (95 % identity), which is speculated
to be responsible for producing benzylmalonyl-CoA (3) and
alkylmalonyl-CoAs. To verify this assumption, we overex-
pressed SpnE and purified it from E. coli to examine its sub-
strate specificity by LC-HRMS analysis. As anticipated, SpnE
efficiently accepts both cinnamoyl-CoA (4) and alkyl-α, β-
unsaturated acyl-CoAs including crotonyl-CoA (5) and 5-
methyl-2-hexenoyl-CoA (6) for conversion into corresponding
reductive carboxylated products (Figure 5AB and Figure S2A-
C). This provides direct evidence of benzylmalonyl-CoA (3)
participating in SPN biosynthesis and establishes the capabil-
ity of SpnE in providing a diverse array of extender units to
this pathway.
Figure 5. Characterization of SpnE and SpnD in vitro. (A) SpnE
and SpnD-catalyzed reactions. (B) HPLC analysis of SpnE cata-
lyzed reductive carboxylation of cinnamoyl-CoA (4) to produce
the benzylmalonyl-CoA (3) as well as the shunt product phe-
nylpropionyl-CoA (7) in the control (I, SpnE free) and reaction
system (II). (C) HRMS analysis of molecular weight of the phos-
phopantetheinylated
serine
peptide
fragment
(PVLVEIGPGDSLTK) in the SpnD reaction. (D) HPLC analysis
of the benzylmalonyl-CoA (3) consumption in the SpnD reaction
before (I) and after the reaction (II).
CCRs are known to have broad selectivity, but their prefer-
ence for aromatic substrates is not known. Kinetic analysis of
SpnE revealed it indeed has a 2.3 fold preference of cin-
namoyl-CoA (4) (kcat/Km =1.42± 0.008 min-1 mM-1) over the
typical aliphatic substrate crotonyl-CoA (5) (kcat/Km= 0.614±
0.041 min-1 mM-1) (Figure S2D). CCRs are typically con-
served among different species (normally over 75% identity),
which indicates that other CCRs may also have the ability to
accept aromatic substrates to generate aromatic CoA-linked
extender units.
Correlation of Amino acid Origin to Benzylmalonyl-CoA
Synthesis and Crosstalk of encc and spn. Cinnamoyl-CoA (4)
is a ubiquitous precursor for synthesis of phenylpropanoids in
plants,15 but its distribution in bacteria is rare and only known
in a few pathways.16,17 Both plant and bacteria systems utilize
a phenylalanine ammonia-lyase (PLA) and a CoA ligase (CL)
to sequentially convert phenylalanine to cinnamic acid (8) and
then to cinnamoyl-CoA (4). However, in spn, no obvious pal
and cl homologues are found. Flanking the spn locus are a few
genes encoding NRPS (orf1), putative protein (orf2) and a
Based on our assembly logic, SpnD-AT is assumed to select
benzylmalonyl-CoA (3) or alkylmalonyl-CoAs for introduc-
ACS Paragon Plus Environment

Page 5 of 9
Journal of the American Chemical Society
1
2
3
kinase (orf4) whose functions are not clear. To verify the po-
4
5
6
7
8
9
tential role of orf1, orf2, or orf4 in cinnamoyl-CoA (4) synthe-
sis, the fosmid containing the spn locus and flanking genes
(orf1-6) was chosen for heterologous expression. It was first
engineered to bear elements for conjugation and integration,
and then introduced into S. coelicolor. LC-HRMS analysis of
the culture extract of the recombinant strain showed that it
failed to produce SPNs while still being able to produce ANTs
(Figure 4, trace V). This result shows that the genes involved
in cinnamoyl-CoA (4) synthesis are not immediately flanking
spn but, instead, are further outside of this locus.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In order to get the entire genetic details, we further se-
quenced the CNQ431 genome.18 Upon careful analysis, we
identified a cluster of 19 genes which are highly identical
(~96 % identity) to the cluster of ENC from S. maritimus,16
(Figure S4, Table S2). Among the cluster, two genes (enccP,
and enccH) are homologues to encP and encH, whose function
are known as PLA and cinnamate CoA ligase, respectively.19
We therefore speculated that these two genes might be respon-
sible for supplying cinnamoyl-CoA (4) for benzylmalonyl-
CoA (3) synthesis. To test this assumption, we individually
inactivated the enccP and enccH genes, and analyzed their fer-
mentation. Gratifyingly, LC-HRMS analysis revealed that
inactivation of enccP completely abolishes the production of
SPN-C but not ANTs (except A-21) confirming the essential
role of this gene in the synthesis of benzylmalonyl-CoA (3)
(Figure 4 trace IV and Figure S1). The benzoyl group in the
C8-hydroxyl of A-21 might be derived from the degradation
of EnccP-catalyzed deamination as in ENC.19a Inactivation of
enccH shows no influence on the biosynthesis of SPN-C or
ANTs (Figure 4, trace III), suggesting that its function can be
compensated by other CoA ligases in the genome as observed
in the enc pathway.19a Indeed, CNQ431 has ten CoA-ligase
genes in the genome, including two involved in the fatty acid
degradation whose selectivity are very broad.13,20 To further
validate the function of EnccP, we carried out in vitro studies.
EnccP was initially overexpressed by pET and MBP-fusion
expression systems in E. coli. However, despite achieveing
soluble expressing of this protein in E. coli, no activity was
detected. We therefore turned to a Pseudomonas expression
system to successfully obtain active protein. Based on HPLC
analysis, EnccP was shown to convert phenylalanine into cin-
namic acid (8) (Figure 6). The use of other amino acids as
substrates, including tyrosine, histidine and tryptophan, did not
result in the deaminated products (data not shown), suggesting
EnccP is a phenylalanine-specific ammonia lyase.
Figure 6. HPLC analysis of EnccP catalyzed reaction. I)
Standard phenylalanine; II) Standard cinnamic acid (8); III)
EnccP reaction system.
en operons which controlled by promotors with different di-
rections (Figure S4). These genes might be not coordinately
expressed during SPN producing as similar to the biosynthesis
of rhodochelin and erythrochelin where one of the essential
genes is provided by an uncharacterized cluster.21-22
Figure 7. LC-HRMS analysis of ENC production in S. sp.
CNQ431 (I) and mWHU2451 (ΔenccP) (II). (A) HPLC trace of
ENC. (B) HRMS spectrum of ENC.
Engineering ant Pathway to Produce SPN. We were next
interested in whether this extender unit pathway can be ap-
plied to structural engineering. To simplify the process, we
chose to modify the structurally similar ANTs via engineer the
ant pathway from NRRL 2288.11a We first analyzed the AT
domain of AntD to see if it needed to be swapped for SpnD-
AT. Comparing the specificity-conferring motifs of SpnD-AT
and AntD-AT, we surprisingly found they are indeed the same
(Figure S3), suggesting AntD-AT is potentially able to accept
benzylmalonyl-CoA (3). In a previous study we validated that
AntE is able to catalyze the reductive carboxylation of cin-
namoyl-CoA (4).13 These results raise an interesting possibil-
ity that the non-production of SPNs by ants is probably due to
the lack of cinnamoyl-CoA (4). This assumption was further
supported by the fact that all of the ANT producers including
S. albus J1074, S. ambofaciens ATCC 23877 and S. sp. NRRL
2288 do not contain genes encoding for cinnamoyl-CoA (4).11a
Crosstalk between secondary metabolites is not common,
since it necessitates their expression coordinately which is
often fermentation and temporal condition dependent. To con-
firm the crosstalk of encc and spn, culture extract of SPNs
from CNQ431 was analyzed to see whether ENC is co-
produced. LC-HRMS analysis however showed no ENC was
produced, suggesting at least one gene except enccP in the encc
was not properly expressed under the SPN-producing condi-
tion. To further confirm this assumption, both CNQ431 and
enccP-inactivated strains were fermented under the ENC-
producing condition. Based on LC-HRMS analysis, a peak
with a molecular weight of ENC ([M+H]+ calcd. For
C22H21O10, 445.1120; observed, 445.1119) was clearly detect-
ed in the fermentation of CNQ431, while not in the enccP-
inactivated mutant (Figure 7), confirming encc indeed encodes
ENC and enccP is essential for its synthesis. enc contains sev
To test this assumption, we introduced the ermE*-enccP and
ermE*-enccP-enccH (ermE* is a constitutive promoter) cas-
settes respectively into NRRL 2288-AL2110, in which the
acyltransferase gene antB for C8-hydroxyl acylation was de-
leted.13 This strain only produces C8-hydroxyl deacyl-ANTs
that offers a cleaner metabolite background. Analysis of the
recombinant strain revealed both ermE*-enccP and ermE*-
enccP-enccH can lead to production of a new compound (Fig-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 9
1
2
3
ure 8), suggesting the function of enccH can be compensated
Interestingly, in spn Nature employs a more direct strategy
4
5
6
7
8
9
by an endogenous CoA ligase as we observed above. By scal-
ing up the fermentation, this compound was further purified in
enough quantity for spectrometric analysis. 1H-NMR, 13C-
NMR and HRMS analysis confirmed its identity as SPN-J
which contains a benzyl group in C7 (Figure 1).10 This result
demonstrates the feasibility of selectively engineering polyke-
tide carbon scaffolds through this benzylmalonyl-CoA extend-
er unit pathway. Moreover, it also demonstrates that the AT
domain of AntD is able to accept benzylmalonyl-CoA (3) as
an extender unit.
to produce benzylmalonyl-CoA (3). By employing an amino
acid ammonia lyase, it performs the transformation of the Cα
amino group of phenylalanine into a classic functional group
found in fatty acids-the α, β-double bond. The resulting deam-
inated product α, β-unsaturated acid can be recognized by a
promiscuous fatty acid CoA ligase to convert it into its CoA-
form, which can subsequently go into the CCR synthetic
pathway like other α, β-unsaturated acyl-CoAs from fatty acid
metabolism to generate a CoA-linked extender unit. Distinct
from the previously known ACP-linked models, this strategy
unifies amino acid degradation and fatty acid derived extender
unit synthesis, and is more energy-saving and concise. It is can
also be tailored to fit several other amino acids. Moreover,
compared with ACP-linked extender units which are used by
only a few PKS systems, CoA-linked extender units are more
useful for polyketide engineering. The benzylmalonyl-CoA (3)
pathway allows for the first time the ability to convert other
major carbon source-amino acids into CoA-linked extender
units, opening up a new avenues for polyketide extender unit
synthesis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Versatility for Converting Other Amino Acids into CoA-
linked Extender Units, and Significance for Expanding
Extender Unit Variety. CoA-linked extender units are major-
ly sourced from fatty acid metabolism and therefore mostly
contain aliphatics. Currently, there are no CoA-linked extend-
er units that have aromatic structures. Haloethylmalonyl-
CoA,3e,24 and allylmalonyl-CoA3f are extender units known to
bear active groups on their side chains (Figure 1). These func-
tionalities are introduced onto the extender units by unusual
biosynthetic machineries making them poorly amenable as
platforms for creating novel extender units. Although chemo-
enzymatic approaches result in the synthesis of unusual ex-
tender units or acyl-NAC surrogates, either their preparation
are costly or incorporation efficiency are low, making them
unsuitable for larger scale production. Hence, finding a more
versatile way to increase both the structural diversity and
chemical reactivity of CoA-linked extender units is a long-
term goal in the field of polyketide engineering.
Figure 8. HPLC analysis of the SPN-J production in AL2110
recombinant strains. Peaks corresponding to SPN-J (▼) is indi-
cated. I) Wide-type AL2110; II) mWHU2454 (AL2110 carry-
ing ermE*-enccP) and III) mWHU2455 (AL2110 carrying
ermE*-enccP-enccH).
DISCUSSION
Novel Strategy to Convert an Amino Acid into Polyketide
Extender Unit. Fatty acids and amino acids are the major
carbon sources in living organisms. Their utility in secondary
metabolite synthesis rarely overlap. Amino acids are mostly
used in peptide and alkaloid synthesis, while fatty acids are
used for polyketide synthesis. The major reason for this is due
to their distinct chemical properties. For instance, extender
units of polyketides contain a carboxylic group in the Cα of a
thioesterified carboxylate that serves as a driving force in the
Claisen condensation reaction when it is decarboxylated.23 In
contrast, this is not the case for amino acids due to the pres-
ence of an amino group in the Cα that decreases the acidity of
Cα proton and prevents it from undergoing the same chemistry
as with Cα carboxylates. Thus, to covert amino acids into
polyketide extender units, Nature utilizes a different approach.
For example, in the synthesis the aminomalonyl-ACP (1),
serine is first activated by loading onto an ACP, then two de-
hydrogenases are recruited to sequentially oxidize its Cβ-
hydroxyl group into a carboxyl group.3c In the synthesis of
dichloropyrrolypropylmalonyl-ACP (2), proline servers as a
starter unit and is assumed to be activated and modified to
form the dichloropyrroly-ACP, which further undergoes two
rounds of C-2 extension with two molecules of malonyl-CoA
before being converted into an extender unit.3j Since these
pathways require unique biosynthetic machinery to mediate
highly specific chemical reactions, it is challenging for both of
Nature and engineering efforts to tailor them to suit other ami-
no acids.
Amino acids, which are abundant in vivo and have plenty of
structural diversity and chemical reactivity, are an ideal source
for CoA-linked extender unit synthesis. In this study, we dis-
closed an unprecedented strategy to convert an amino acid into
a CoA-linked extender unit. The conversion is simple and
concise. More significantly, the bio-machinery involved in the
process are either readily available or highly promiscuous and
make this strategy highly versatile for incorporating additional
amino acids. Genes encoding amino acid deamination, such as
phenylalanine, tyrosine, histidine and aspartic acid are well
studied.25 Some of them are broadly-selective and can even
accept modified amino acids, such as m-tyrosine which can be
generated in vivo by introduction of a phenylalanine m-
hydroxylase.8,25d Bacterial CoA ligases, in particular those
involved in fatty acid degradation are very promiscuous, and
deaminated amino acids can be effectively converted by them
or their specific CoA-ligases, such as 4-hydroxycoumarate-
CoA ligase into α, β-unsaturated CoAs.13,20,26 We and others
have revealed several CCRs that are extremely promiscuous
and can tolerate diverse substitutions on the Cα, including
alkyl groups of up to ten carbons, halogenated alkyl-, cyclo-
hexenyl-, alkynyl-substituted alkyl- and even benzyl-
groups.3b,3i,13,26 Although AntE and SpnE are unable to turno-
ver 4-hydroxycoumarate-CoA and 3-indoleacrylic-CoA which
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
are deaminated products of tyrosine and tryptophan (data not
aspartic acid via exchanging individual biosynthetic machin-
ery. Additionally, their selection and incorporation into PKS
assembly can be performed by the AT domains of SpnD and
AntD, which overcome the current limitation of extender unit
selection. This versatile extender unit synthetic pathway
broadens the specificity of the AT domain and paves the way
for introducing additional amino acids and synthetic aromatics
with diverse structural diversity and chemical reactivity into
the polyketide carbon scaffold.
shown), sequence analysis with the well-studied CCR CinF
revealed that their binding pockets have the potential for en-
largement (Figure S5). By replacing the selectivity-conferring
residues (V163 and V362, numbers corresponding to CinF)
with smaller residues such as glycines, these CCRs may be
able to accept bulky aromatic substrates. Currently, several
structures of broad-selective CCRs have been made availa-
ble.3i,27 Based on the protein structures, directed evolution at
these sites could highly effective in expand CCR’s selectivity
scope. It is believed, the strategy disclosed here can be readily
tailored to convert other amino acids including tyrosine, histi-
dine and aspartic acid into CoA-linked extender units. Extend-
er units derived from amino acids, particularly those with ar-
omatic moieties and active groups will greatly expand their
chemical diversity, and facilitate polyketide structural engi-
neering.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT
Supporting Information. Experimental details and characteriza-
tion data. This material is available free of charge via the Internet
at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
quxd@whu.edu.cn
Broad Selectivity of the AT Domain and Potential to In-
troduce Structural Diversity into the Polyketide Scaffold.
With ATs serving as a gatekeeper for extender unit selection,
they are other critical factors for structural engineering. Cur-
rently introducing aromatic groups into polyketide scaffolds
through extender unit selection is not yet possible, mostly be-
cause ATs lack aromatic-selectivity.28 In this study, we charac-
terized the first example of an AT domain (SpnD-AT) that can
effectively select an aromatic extender unit to introduce a ben-
zyl group into the SPN scaffold. Through in vivo study we
further validated that the AT domain of AntD is also compe-
tent in accepting benzylmalonyl-CoA (3). These discoveries
overcome the current limitation of AT selection to now in-
clude aromatics.
Author Contributions
§These authors contributed equally.
Funding Sources
This work was supported by NSFC grants (31322002,
31270119, and 31400052), Program of NCET (NCET-12-0423),
the State Key Laboratory of Materials-Oriented Chemical Engi-
neering (KL13-19), and China Postdoctoral Science Foundation
(2014M560620 and 2014M552077).
ACKNOWLEDGMENT
We thank Prof. William Fenical at Scripps Research Institute
for provision of the strain CNQ431; Prof. Ikuro Abe and Dr. Li-
han Zhang at Tokyo University for gifting CoA substrates; Prof.
Yizheng Zhang at Shichuan University for provision of the plas-
mid pCIBhis; Prof. Chao Peng at National Center for Protein Sci-
ence Shanghai for performing the Mass analysis and Prof. Kui
Hong at Wuhan University for provision of the ENC fermentation
condition.
Previously, we demonstrated AntD can accept various types
and lengths of substituted extender units.13 Most of these sub-
strates are bulkier than extender units derived from amino
acids, suggesting AntD-AT is able to select other amino acid-
derived extender units. By coupling the versatile extender unit
synthetic strategy and AT swapping, a complete pathway to
introduce amino acid varieties into polyketide carbon scaffolds
can likely be constructed. In addition, besides aromatic groups
from amino acids, these biosynthetic proteins can also be tai-
lored to introduce synthetic aromatics into the polyketide scaf-
fold. CoA ligases converting cinnamate analogues into their
CoA forms are available.5e,20 These genes can be co-expressed
with selectivity-improved CCRs to constitute an artificial
pathway for converting synthetic cinnamates into CoA-linked
extender units. With the selection of SpnD-AT or AntE-AT,
introduction of synthetic cinnamate varieties into polyketide
scaffold is potentially feasible. Currently, elucidation of the
SpnE-AT structure is underway. With structural-based direct
evolution, it is believed this selectivity can be engineered to
allow for more diverse aromatic groups.
REFERENCES
1. Newman, D. J.; Cragg, G. M. J. Nat. Prod. 2012, 75, 311-335.
2. (a) Williams, G. J. Curr. Opin. Struct. Biol. 2013, 23, 603-612;
(b) Dunn, B. J.; Khosla, C. J. R. Soc. Interface 2013, 10, 20130297.
3. (a) Chan, Y. A.; Podevels, A. M.; Kevany, B. M.; Thomas, M. G.
Nat. Prod. Rep. 2009, 26, 90-114; (b) Wilson, M. C.; Moore, B. S.
Nat. Prod. Rep. 2012, 29, 72-86; (c) Chan, Y. A.; Boyne, M. T.; Po-
devels, A. M.; Klimowicz, A. K.; Handelsman, J.; Kelleher, N. L.;
Thomas, M. G. Proc. Natl. Acad. Sci. USA 2006, 103, 14349-14354;
(d) Jia, X.; Tian, Z.; Shao, L.; Qu, X.; Zhao, Q.; Tang, J.; Tang, G.;
Liu, W. Chem. Biol. 2006, 13, 575-585; (e) Liu, Y.; Hazzard, C.;
Eustáquio, A. S.; Reynolds, K. A.; Moore, B. S. J. Am. Chem. Soc.
2009, 131, 10376-10377; (f) Mo, S.; Kim, D. H.; Lee, J. H.; Park, J.
W.; Basnet, D. B.; Ban, Y. H.; Yoo, Y. J.; Chen, S.-w.; Park, S. R.;
Choi, E. A.; Kim, E.; Jin, Y.-Y.; Lee, S.-K.; Park, J. Y.; Liu, Y.; Lee,
M. O.; Lee, K. S.; Kim, S. J.; Kim, D.; Park, B. C.; Lee, S.-g.; Kwon,
H. J.; Suh, J.-W.; Moore, B. S.; Lim, S.-K.; Yoon, Y. J. J. Am. Chem.
Soc. 2010, 133, 976-985; (g) Wilson, M. C.; Nam, S.-J.; Gulder, T. A.
M.; Kauffman, C. A.; Jensen, P. R.; Fenical, W.; Moore, B. S. J. Am.
Chem. Soc. 2011, 133, 1971-1977; (h) Xu, Z.; Ding, L.; Hertweck, C.
Angew. Chem. Int. Ed. 2011, 50, 4667-4670; (i) Quade, N.; Huo, L.;
Rachid, S.; Heinz, D. W.; Müller, R. Nat. Chem. Biol. 2012, 8, 117-
124; (j) Mantovani, S. M.; Moore, B. S. J. Am. Chem. Soc. 2013, 135,
18032-18035.
CONCLUSION:
In summary, we have characterized the formation and selec-
tion of the aromatic extender unit benzylmalonyl-CoA (3)
from the biosynthesis of splenocin and enterocin. The synthe-
sis of this extender unit is distinct from the previous known
CoA-linked extender units. By employing
a deamina-
tion/reductive carboxylation strategy, phenylalanine is con-
verted into the CoA-linked extender unit. More significantly,
this pathway is concise and highly versatile which can be ap-
plied to a few amino acids including tyrosine, histidine and
4. An, J. H.; Kim, Y. S. Eur. J. Biochem. 1998, 257, 395-402.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
1
2
3
4
5
6
7
8
9
5. (a) Pohl, N. L.; Hans, M.; Lee, H. Y.; Kim, Y. S.; Cane, D. E.;
Khosla, C. J. Am. Chem. Soc. 2001, 123, 5822-5823; (b) Hughes,
Amanda J.; Keatinge-Clay, A. Chem. Biol. 2011, 18, 165-176; (c)
Koryakina, I.; Williams, G. J. ChemBioChem 2011, 12, 2289-2293; (d)
Koryakina, I.; McArthur, J.; Randall, S.; Draelos, M. M.; Musiol, E.
M.; Muddiman, D. C.; Weber, T.; Williams, G. J. ACS Chem. Biol.
2012, 8, 200-208; (e) Go, M. K.; Chow, J. Y.; Cheung, V. W. N.; Lim,
Y. P.; Yew, W. S. Biochemistry 2012, 51, 4568-4579.
6. Kennedy, J. Nat. Prod. Rep. 2008, 25, 25-34.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7. Runguphan, W.; Qu, X.; O'Connor, S. E. Nature 2010, 468, 461-
464.
8. Qu, X.; Jiang, N.; Xu, F.; Shao, L.; Tang, G.; Wilkinson, B.; Liu,
W. Mol. BioSyst. 2011, 7, 852-861.
9. (a) Ritchie, T. J.; Macdonald, S. J. Drug Discov. Today 2009, 14,
1101-1020; (b) Ward, S. E.; Beswick, P. Expert Opin. Drug Discov.
2014, 9, 995-1003.
10. Strangman, W. K.; Kwon, H. C.; Broide, D.; Jensen, P. R.;
Fenical, W. J. Med. Chem. 2009, 52, 2317-2327.
11. (a) Yan, Y.; Zhang, L.; Ito, T.; Qu, X.; Asakawa, Y.; Awakawa,
T.; Abe, I.; Liu, W. Org. Lett. 2012, 14, 4142-4145; (b) Sandy, M.;
Rui, Z.; Gallagher, J.; Zhang, W. ACS Chem. Biol. 2012, 7, 1956-
1961.
12. (a) Sandy, M.; Zhu, X.; Rui, Z.; Zhang, W. Org. Lett. 2013, 15,
3396-3399; (b) Liu, J.; Zhu, X.; Seipke, R. F.; Zhang, W. ACS Synth.
Biol. 2014; doi: 10.1021/sb5003136 (c) Seipke, R. F.; Patrick, E.;
Hutchings, M. I. Peer J. 2014, 2, e253.
13. Yan, Y.; Chen, J.; Zhang, L.; Zheng, Q.; Han, Y.; Zhang, H.;
Zhang, D.; Awakawa, T.; Abe, I.; Liu, W. Angew. Chem. Int. Ed.
2013, 52, 12308-12312.
14. Quadri, L. E. N.; Weinreb, P. H.; Lei, M.; Nakano, M. M.;
Zuber, P.; Walsh, C. T. Biochemistry 1998, 37, 1585-1595.
15. Hahlbrock, K.; Scheel, D. Annu. Rev. Plant Physiol. Plant Mol.
Biol. 1989, 40, 347-369.
16. Piel, J.; Hertweck, C.; Shipley, P. R.; Hunt, D. M.; Newman, M.
S.; Moore, B. S. Chem. Biol. 2000, 7, 943-955.
17. (a) Bezanson, G. S.; Desaty, D.; Emes, A. V.; Vining, L. C.
Can. J. Microbiol. 1970, 16, 147-151; (b) Hill, A. M.; Thompson, B.
L.; Harris, J. P.; Segret, R. Chem. Commun. 2003, 1358-1359.
18. Yu, M.; Dai, Z.; Qu, X.; Gao, X. Genome Announc. 2015, 3,
e01383-14.
19. (a) Xiang, L.; Moore, B. S. J. Bacteriol. 2003, 185, 399-404; (b)
Xiang, L.; Moore, B. S. J. Bacteriol. 2005, 187, 4286-4289.
20. Kaneko, M.; Ohnishi, Y.; Horinouchi, S. J. Bacteriol. 2003,
185, 20-27.
21. Bosello, M.; Robbel, L.; Linne, U.; Xie, X.; Marahiel, M. A. J.
Am. Chem. Soc. 2011, 133, 4587-4595.
22. Lazos, O.; Tosin, M.; Slusarczyk, A. L.; Boakes, S.; Cortés, J.;
Sidebottom, P. J.; Leadlay, P. F. Chem. Biol. 2010, 17, 160-173.
23. Hertweck, C. Angew. Chem. Int. Ed. 2009, 48, 4688-4716.
24. Eustáquio, A. S.; O’Hagan, D.; Moore, B. S. J. Nat. Prod. 2010,
73, 378-382.
25. (a) Fibriansah, G.; Veetil, V. P.; Poelarends, G. J.; Thunnissen,
A.-M. W. H. Biochemistry 2011, 50, 6053-6062; (b) Kyndt, J. A.;
Meyer, T. E.; Cusanovich, M. A.; Van Beeumen, J. J. FEBS Lett.
2002, 512, 240-244; (c) Schwede, T. F.; Rétey, J.; Schulz, G. E. Bio-
chemistry 1999, 38, 5355-5361; (d) Turner, N. J. Curr. Opin. Chem.
Biol. 2011, 15, 234-240; (e) Raj, H.; Szymański, W.; de Villiers, J.;
Rozeboom, H. J.; Veetil, V. P.; Reis, C. R.; de Villiers, M.; Dekker, F.
J.; de Wildeman, S.; Quax, W. J.; Thunnissen, A.-M. W. H.; Feringa,
B. L.; Janssen, D. B.; Poelarends, G. J. Nat. Chem. 2012, 4, 478-484.
26. Zhu, X.; Liu, J.; Zhang, W. Nat. Chem. Biol. 2015, 11, 115-120.
27. Zhang, L.; Chen, J.; Mori, T.; Yan, Y.; Liu, W.; Abe, I. Acta.
Cryst. Section F 2014, 70, 734-737.
28. Koryakina, I.; McArthur, J. B.; Draelos, M. M.; Williams, G. J.
Org. Biomol. Chem. 2013, 11, 4449-4458
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
38x17mm (300 x 300 DPI)
ACS Paragon Plus Environment