Unprecedented Cyclization Catalyzed by a Cytochrome P450 in Benzastatin Biosynthesis

Article abstract of DOI:10.1021/jacs.8b02769

Benzastatins have unique structures probably derived from geranylated p-aminobenzoic acids. The indoline and tetrahydroquinoline scaffolds are presumably formed by cyclization of the geranyl moiety, but the cyclization mechanism was unknown. We studied the benzastatin biosynthetic gene cluster of Streptomyces sp. RI18; functions of the six enzymes encoded by it were analyzed by gene disruption in a heterologous host and in vitro enzyme assays. We propose the biosynthetic pathway for benzastatins in which a cytochrome P450 (BezE) is responsible for the cyclization of geranylated p-acetoxyaminobenzoic acids; BezE catalyzes elimination of acetic acid to form an iron nitrenoid, nitrene transfer to form an aziridine ring, and nucleophilic addition of hydroxide ion to C-10 and chloride ion to C-9 to generate the indoline and tetrahydroquinoline scaffolds, respectively. Discovery of this enzyme, which should be termed cytochrome P450 nitrene transferase, provides an important insight into the functional diversity of cytochrome P450.

Full text of DOI:10.1021/jacs.8b02769

Subscriber access provided by Kaohsiung Medical University
Article
Unprecedented cyclization catalyzed by a
cytochrome P450 in benzastatin biosynthesis
Hayama Tsutsumi, Yohei Katsuyama, Miho Izumikawa, Motoki Takagi,
Manabu Fujie, Noriyuki Satoh, Kazuo Shin-ya, and Yasuo Ohnishi
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b02769 • Publication Date (Web): 02 May 2018
Downloaded from http://pubs.acs.org on May 2, 2018
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 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 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.
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 10
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Unprecedented cyclization catalyzed by a cytochrome P450 in benꢀ
zastatin biosynthesis
Hayama Tsutsumi1, Yohei Katsuyama1,2,*, Miho Izumikawa3,+, Motoki Takagi3, Manabu Fujie4, Noriꢀ
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
yuki Satoh4, Kazuo Shinꢀya5 and Yasuo Ohnishi1,2,*
1Department of Biotechnology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1ꢀ1ꢀ1 Yayoi,
Bunkyoꢀku, Tokyo 113ꢀ8657, Japan
2Collaborative Research Institute for Innovative Microbiology, The University of Tokyo, 1ꢀ1ꢀ1 Yayoi, Bunkyoꢀku, Tokyo
113ꢀ8657, Japan
3Japan Biological Informatics Consortium (JBIC), 2ꢀ4ꢀ7 Aomi, Kotoꢀku, Tokyo 135ꢀ0064, Japan
4Okinawa Institute of Science and Technology Graduate University, 1919ꢀ1 Tancha, Onnaꢀson, Kunigamiꢀgun, Okinawa
904ꢀ0495, Japan
5National Institute of Advanced Industrial Science and Technology (AIST), 2ꢀ4ꢀ7 Aomi, Kotoꢀku, Tokyo 135ꢀ0064, Japan
+Deceased on 23 December 2015
Supporting Information Placeholder
ABSTRACT: Benzastatins have unique structures probably derived from geranylated pꢀaminobenzoic acids. The indoline
and tetrahydroquinoline scaffolds are presumably formed by cyclization of the geranyl moiety, but the cyclization mechaꢀ
nism was unknown. We studied the benzastatin biosynthetic gene cluster of Streptomyces sp. RI18; functions of the six enꢀ
zymes encoded by it were analyzed by gene disruption in a heterologous host and in vitro enzyme assays. We propose the
biosynthetic pathway for benzastatins in which a cytochrome P450 (BezE) is responsible for the cyclization of geranylated pꢀ
acetoxyaminobenzoic acids; BezE catalyzes elimination of acetic acid to form an iron nitrenoid, nitrene transfer to form an
aziridine ring, and nucleophilic addition of hydroxide ion to Cꢀ10 and chloride ion to Cꢀ9 to generate the indoline and tetraꢀ
hydroquinoline scaffolds, respectively. Discovery of this enzyme, which should be termed cytochrome P450 nitrene transferꢀ
ase, provides an important insight into the functional diversity of cytochrome P450.
icatumtoxin, griseofulvin, and fumitremorgin C).2 Recently,
Introduction
several researchers reported that cytochrome P450 mutants
Cyclization is one of the most important reactions
in the biosynthesis of natural products; it produces characꢀ
teristic molecular scaffolds, which are often important for
the bioactivity of natural products as represented by the βꢀ
lactam ring. Various cyclization reactions, such as the
Diels–Alder reaction, oxidative cyclization, and nucleoꢀ
philic addition, are used for natural product biosynthesis.1,2
However, some natural products seem to be synthesized by
still unknown cyclization reactions.
are capable of catalyzing carbene and nitrene transfer reacꢀ
tions as artificial reactions (Scheme 1b).16 For instance,
engineered P450BM3 was used for the olefin cyclopropanaꢀ
tion via the carbene transfer reaction without any oxidative
step.17 In this reaction, engineered P450BM3 forms an ironꢀ
carbenoid species for the carbene transfer reaction.17 Thus,
cytochrome P450s are potentially capable of catalyzing
these transfer reactions, but such cytochrome P450s have
so far not been found in nature.
Cytochrome P450s are useful biocatalyst catalyzꢀ
ing various oxidation reactions including hydroxylation,3–6
epoxidation,7,8 and biaryl coupling9–11 (Scheme 1a).12 Most
cytochrome P450s require redox partners (e.g., ferredoxin
and ferredoxin reductase) to reduce the ferric heme iron
during their catalytic cycle. Some can use peroxides to byꢀ
pass the reduction of the heme iron. In the biosynthesis of
natural products, cytochrome P450s play important roles to
generate structural diversity13–15 and some are responsible
for oxidative cyclization (e.g., in the biosynthesis of viridꢀ
Benzastatin derivatives are natural products isolatꢀ
ed from Streptomyces species (Figure 1).18–25 For example,
JBIRꢀ67 (5a), 7ꢀhydroxyl benzastatin D (6′d), virantmycin
(6d), and JBIRꢀ73 (6′e) were isolated from Streptomyces
sp. RI18. Benzastatin derivatives show interesting bioactivꢀ
ities, such as neuronal cell protection19 and antiviral activiꢀ
ties,22 and have diverse structures, such as indoline and
tetrahydroquinoline scaffolds, that are probably derived
from geranylated pꢀaminobenzoic acid (PABA) or PABA
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 10
OH
b
a
R
Fe4+
N
hydroxylation
N2
1
2
3
N
S
O
N
O-
S
O
O
S
O
epoxidation
O
O
O
O
O
4
5
6
R
Fe3+
Fe4+
or
Fe2+
R
biaryl coupling
R
2
7
8
9
OH
c
H
H
N
10
NH2
NH2
9
R1
R1
HO
HO
R1
oxidase
HO
cyclase
HO
5a and 5b
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
O
O
O
O
O
OH
R2
R2
H
N
R1 = H, R2 = H: 2a
R1 = Me, R2 = H: 2b
1 = Me, R2 = OH: 2c
10
9
OH
R
6'c
OH
d
H
H
N
Fe2+/3+
10
O
Fe4+/5+
N
9
O
N
R2
HO
HO
R1
AH
R2
H
R1
10
N
O
O
5a and 5b
OH
HO
9
HO
-OH
Cl-
HO
R1
R1
H
O
AcOH
O
R2
O
N
10
9
R1 = H, R2 = H: 4a
R1 = Me, R2 = H: 4b
R1 = Me, R2 = OH: 4c
Cl
6c
Scheme 1. Reactions catalyzed by cytochrome P450 and proposed cyclization reactions catalyzed by BezE. (a) Oxidaꢀ
tive reactions catalyzed by cytochrome P450. (b) Mechanisms of aziridine ring formation catalyzed by engineered cytoꢀ
chrome P450. (c) Putative mechanism for the cyclization of geranyl moiety to generate indoline and tetrahydroquinoline
scaffolds, which is based on the previously predicted cyclization reaction via epoxidation and epoxideꢀopening for benꢀ
zastatins D and E.16 (d) Possible cyclization reactions catalyzed by BezE. Red and blue arrows indicate the reactions to synꢀ
thesize indoline and tetrahydroquinoline scaffolds, respectively. Oxygen derived from H2O is shown in green.
derivatives, such as pꢀaminobenzamide. The biosynthesis
of benꢀzastatins D and E was proposed to be initiated from
geranylation of pꢀaminobenzamide followed by epoxidaꢀ
tion and cyclization in Streptomyces nitrosporeus 3064
(Scheme 1c).26 However, neither the enzymes involved in
these reactions nor any gene clusters for the biosynthesis of
benzastatin derivatives were known for a long time. Wehad
determined a draft sequence of the Streptomyces sp. RI18
genome, which led the identification of the biosynthetic
gene cluster (bez gene cluster).27
duced by Streptomyces sp. RI18 (Figure 2). Three (bezD,
bezH, and bezI) of the 10 genes are probably involved in
the supply of the precursors, isoprenyl pyrophosphate and
PABA. BezF, a putative UbiAꢀtype polyprenyltransferase,28
appears to be responsible for the geranylation of PABA.
Here we analyzed the bez gene cluster by a heterꢀ
ologous gene expression system and in vitro assays of reꢀ
combinant enzymes. We propose a probable benzastatin
biosynthesis pathway involving an unprecedented cyclizaꢀ
tion reaction catalyzed by a cytochrome P450 (Scheme
1d).
Results
In silico analysis and heterologous expression of bez
gene cluster
Biosynthetic gene clusters similar to the bez clusꢀ
ter were discovered in Streptomyces olivaceus, Streptomy-
ces niveus, Streptomyces ipomoeae, Streptomyces
acidiscabies, and Saccharothrix espanaensis (Figure 2).
Sequence comparison between the bez cluster and these
gene clusters indicated that 10 genes (bezA-J) are responsiꢀ
ble for the biosynthesis of the benzastatin derivatives proꢀ
*
†
‡
Figure 1. Structures of benzastatin derivatives. , , ,
and indicate benzastatin derivatives isolated from Strep-
tomyces sp. RI18, Streptomyces nitroporeus 3064, S. nitro-
poreus AMꢀ2722, and Streptomyces sp. SANK 60101, reꢀ
spectively.
§
ACS Paragon Plus Environment

Page 3 of 10
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 2. Benzastatin biosynthetic gene cluster from Streptomyces sp. RI18 and homologous clusters from other actiꢀ
nomycetes. bezD encodes a putative transꢀisoprenyl pyrophosphate synthase.31 bezH and bezI encode a PabAꢀPabB fusion
protein and PabC homologues, respectively. PabA, PabB, and PabC are responsible for the biosynthesis of PABA.32 Thus,
BezD and BezHꢀBezI are probably the enzymes for the supply of the precursors, isoprenyl pyrophosphate and PABA, reꢀ
spectively, to enhance the production of benzastatins.
Tables 2ꢀ7 and Supplementary Figures 2, 3 and 18ꢀ54;
see Scheme 2 for their structures). JBIRꢀ73 (6′e) was not
produced in this system. Compounds 7a and 7b are a novel
BezA and BezB are putative methyltransferases, showing
homologies to a γꢀtocopherol methyltransferases, which
catalyzes benzene ring methylation,29 and RebM, which
class of benzastatin derivatives, which possess
a
catalyzes the Oꢀmethylation of a sugar,30 respectively. Two
cytochrome P450s BezC and BezE were predicted to be
responsible for incorporation of a hydroxy group or epoxiꢀ
dation, which presumably occurs prior to the cyclization.
BezG is a putative Nꢀacetyltransferase and BezJ is a protein
homologous to the Nꢀoxygenase AurF, which catalyzes
oxidation of the arylamine of PABA to synthesize pꢀ
nitrobenzoate.33–35 The bez cluster was analyzed by heterꢀ
ologous gene expression and several putative benzastatin
biosynthesis pathways were predicted.27 However, the cyꢀ
clization mechanism of benzastatin remained elusive beꢀ
cause the productivity of some benzastatins was unstable.
benz[b]azepine scaffold. This result demonstrates that
bezA-J are responsible for benzastatin biosynthesis. Howꢀ
ever, the profile of benzastatin derivatives produced by the
recombinant S. lividans strain was different from that proꢀ
duced by Streptomyces sp. RI18 (Figure 3a and b). The
recombinant strain produced much more 2a, 5a, and 7a and
much less 2c, 6′d, and 6d than Streptomyces sp. RI18.
Moreover, only the recombinant strain produced 6′c and 6c.
These differences are probably caused by differences in
gene expression; bezA, bezB, bezC, and bezD-J are under
the control of different native promoters in Streptomyces
sp. RI18, while all of them were forcibly expressed using
the tipA promoter in the heterologous expression system.
To express the bez cluster more stably than what
was achieved in the previous work,27 a chromosomeꢀ
integrative expression plasmid, pTYMꢀbezA-J, harboring
bezA, bezB, bezC, and the bezD-J operon, was constructed.
Each gene or gene set was placed under the control of the
tipA promoter (Supplementary Figure 1). Streptomyces
lividans was transformed with this plasmid and metabolites
produced by the recombinant strain were analyzed by liqꢀ
uid chromatographyꢀtandem mass spectrometry (LCꢀ
MS/MS). As a result, production of the following benꢀ
zastatin derivatives was reproducibly observed; 7ꢀhydroxyl
benzastatin J (2a), 7ꢀhydroxyl benzastatin B (2b), 7ꢀ
hydroxyl Oꢀdemethylbenzastatin A (2c), JBIRꢀ67 (5a), 7ꢀ
In vivo analysis of the function of bez genes
To analyze the function of each biosynthesis enꢀ
zyme gene, a series of heterologous expression plasmids,
each of which lacks one of the biosynthesis enzyme genes
(bezA, bezB, bezC, bezE, bezG, and bezJ), was constructed
and transferred into S. lividans (Supplementary Table 1).
Deletion of bezG or bezJ abolished the ability to
produce compounds that possess bicyclic scaffolds (5a, 5b,
6c, 6d, 6′c, 6′d, 7a, and 7b); the ꢀbezG and ꢀbezJ strains
produced only “linear” benzastatins (2a, 2b, and 2c) (Figꢀ
ure 3g and h). The production of benzastatin derivatives
with bicyclic scaffolds was restored by introducing the deꢀ
leted gene (bezG or bezJ) into each mutant using another
chromosome integration plasmid (pTYM2k) (Supplemenꢀ
tary Figure 4). These results indicate that bezG and bezJ
are essential for the second ring formation. Inactivation of
hydroxyl
benzastatin
F
(5b),
7ꢀhydroxyl
Oꢀ
demethylbenzastatin D (6′c), Oꢀdemethylvirantmycin (6c),
7ꢀhydroxyl benzastatin D (6′d), virantmycin (6d), benꢀ
zastatin K (7a), and methylbenzastatin K (7b) (Figure 3b,
characterization of these compounds is described in Supꢀ
plementary Information, with reference to Supplementary
ACS Paragon Plus Environment

Journal of the American Chemical Society
bezE abolished the production of three of the four benꢀ
Page 4 of 10
1
zastatins with a tetrahydroquinoline scaffold (6c, 6′c, and
6d), and decreased the yield of the two benzastatins with an
indoline scaffold (5a and 5b) and one of the four benꢀ
zastatins with a tetrahydroquinoline scaffold (6′d, 6′d was
only detectable in MS analysis) (Figure 3f). These defects
of the ꢀbezE strain were restored by introducing bezE into
the mutant using pTYM2k (Supplementary Figure 4).
This result suggests that BezE is not necessarily required
for the formation of indoline and tetrahydroquinoline scafꢀ
folds, but has an important role in this step. Taken together
with the results of bezE, bezG, and bezJ deletion, we specuꢀ
lated that biosynthetic intermediates that are amenable to
the second ring cyclization should be produced by BezG
and BezJ, and that BezE should facilitate and control the
cyclization of these intermediates. Because formation of
the benz[b]azepine scaffold (7a and 7b) was enhanced in
the ꢀbezE strain (Figure 3f), the cyclization generating the
benz[b]azepine scaffold seems to occur nonꢀenzymatically.
Formation of the indoline and tetrahydroquinoline scaffolds
also seems to occur nonꢀenzymatically with a much lower
efficiency compared with the benz[b]azepine scaffold forꢀ
mation. Meanwhile, BezE appears to be essential for the
introduction of a chloro group to produce 6c and 6d. Thereꢀ
fore, we speculated that the reaction catalyzed by BezE
should involve the introduction of a chloro group.
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
The ꢀbezC strain could not produce the linear
benzastatin 2c and any of the four benzastatins having a
tetrahydroquinoline scaffold (6c, 6′c, 6d, and 6′d) (Figure
3e). This suggests that BezC catalyzes the hydroxylation of
2b to produce 2c, and that the hydroxyl group at Cꢀ17 is
essential for the formation of a tetrahydroquinoline scafꢀ
fold. The ꢀbezA strain produced only 2a, 5a, and 7a (Figꢀ
ure 3c). This result clearly shows that BezA catalyzes
methylation at Cꢀ13, and indicates that the methyl group at
Cꢀ13 is essential for the hydroxylation at Cꢀ17 by BezC.
Figure 3. Heterologous expression of the benzastatin
biosynthetic gene cluster. LCꢀMS/MS analysis of benꢀ
zastatins extracted from (a) Streptomyces sp. RI18, S. li-
vidans strains harboring (b) pTYMꢀbezA-J, (c) pTYMꢀ
bezA-JꢀbezA, (d) pTYMꢀbezA-JꢀbezB, (e) pTYMꢀbezA-
JꢀbezC, (f) pTYMꢀbezA-JꢀbezE, (g) pTYMꢀbezA-JꢀbezG,
(h) pTYMꢀbezA-JꢀbezJ, and (i) pTYM19gt (negative conꢀ
trol). Compound 6′d* was predicted to be an isomer of 6′d
by comparing its tandem mass spectrum with that of 6′d
(Supplementary Figure 6). The compounds observed only
by MS analysis are depicted in gray.
The ꢀbezB strain lost the ability to produce 6d and
6′d, and accumulated 6c and 6′c (Figure 3d), suggesting
that the Oꢀmethylation reaction catalyzed by BezB is the
final step of 6d and 6′d biosynthesis, and that chlorination
occurs prior to this reaction (Scheme 2).
In vitro analysis of the enzymes for the modification of
the geranyl moiety
To obtain further insights into the biosynthetic
pathway for benzastatin derivatives, we first analyzed modꢀ
ification of the geranyl moiety by BezA and BezC in vitro.
Recombinant BezA and BezC were prepared as Nꢀ
terminally His6ꢀtagged proteins (Supplementary Figure
7). The recombinant BezC showed a Soret band shift from
420 to 450 nm as a typical cytochrome P450, when CO was
bubbled into the BezC solution, indicating that BezC was
produced as an active form (Supplementary Figure 8d).
As described above, BezA and BezC were indicated to catꢀ
alyze the methylation at Cꢀ13 and the following hydroxylaꢀ
tion at Cꢀ17, respectively. However, against our expectaꢀ
tion, 2a and 2b were not served as substrates of BezA and
BezC, respectively (Supplementary Figure 8a and b).
Therefore, we hypothesized that the modification of the
geranyl moiety should occur at an earlier step, namely on
geranyl pyrophosphate (GPP, 1a). When BezA was incuꢀ
Nonꢀenzymatic degradation of virantmycin (6d)
Because the chloro group of virantmycin (6d) is a
good leaving group and the thiol group of ergothioneine
has a high nucleophilicity,36 we assumed that JBIRꢀ73 (6′e)
is nonꢀenzymatically synthesized from 6d by the nucleoꢀ
philic attack of the thioketone of ergothioneine (Scheme 2
and Supplementary Figure 5b). To prove this assumption,
a physiological amount of ergothioneine was incubated
with 6d in HEPES buffer (pH 7.4) for 1 h. As expected,
formation of 6′e was observed (Supplementary Figure
5a). This result suggests that 6′e is synthesized nonꢀ
enzymatically from 6d and ergothioneine. In addition, conꢀ
version of 6d into 6′d and its isomer 6′d* was also obꢀ
served, suggesting that 6′d and 6′d* are also synthesized
by the nonꢀenzymatic hydrolysis of 6d.
ACS Paragon Plus Environment

Page 5 of 10
Journal of the American Chemical Society
bated with 1a and Sꢀadenosylmethionine (SAM), Sꢀ
adenosylꢀ ꢀhomocysteine was formed, suggesting that
product of BezG reaction is pꢀacetoxyaminobenzoic acid
1
2
3
4
5
6
7
8
L
(PAcABA). It should be noted that in this experiment we
used BezG of Streptomyces niveus, which shows 87% amiꢀ
no acid sequence identity with BezG of Streptomyces sp.
RI18, because we could not produce the Streptomyces sp.
RI18 BezG in an active form in Escherichia coli.
BezA catalyzes methylation of 1a (Supplementary Figure
8c). To identify the product of BezA reaction, we aimed to
transfer it onto PABA using recombinant BezF. Because
BezF appears to be a membraneꢀbound prenyltransferase,
we prepared crude BezF protein solubilized from the memꢀ
brane fraction of the recombinant S. lividans expressing
bezF. We confirmed that the crude BezF solution possessed
a distinct geranylation activity; 2a was produced when
PABA and 1a were incubated with the crude BezF solution
(Figure 4d and Supplementary Figure 9). BezA was reꢀ
moved from the BezA reaction mixture with ultrafiltration
and the reaction product was transferred onto PABA by the
crude BezF solution. As a result, we observed the producꢀ
tion of 2b (Figure 4b). This result clearly shows that BezA
catalyzes Cꢀmethylation of 1a to synthesize methylgeranyl
pyrophosphate (MGPP, 1b).
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
Next, we analyzed the function of BezC using a
BezAꢀBezC oneꢀpot reaction. BezA and BezC were incuꢀ
bated with 1a, SAM, NADH, ferredoxin (CamA), and ferꢀ
redoxin reductase (CamB), and the product was transferred
onto PABA as described above. As a result, we observed
the formation of 2c (Figure 4a). When BezA was removed
from the reaction mixture, only 2a was detected (Figure
4c), indicating that hydroxylation of GPP could not occur
prior to methylation also in vitro. Taken together, we conꢀ
cluded that BezA and BezC catalyze sequential methylation
and hydroxylation of 1a to synthesize hydroxymethylgeraꢀ
nyl pyrophosphate (HMGPP, 1c). In addition, BezF was
shown to have a relaxed donor substrate specificity; BezF
can catalyze the geranylation of PABA using any of 1a, 1b,
and 1c.
Figure 4. In vitro analysis of BezA and BezC. BezA,
BezC, CamA, and CamB were incubated with 1a, SAM,
and NADH. After the reaction, products were transferred
onto PABA by BezF. (a) In the presence of all the enꢀ
zymes, 2c was detected. (b) When BezC, CamA, CamB,
and NADH were removed from the reaction mixture, 2b
was detected. (c) When BezA was removed from the reacꢀ
tion mixture, 2a was detected. (d) When BezF was incubatꢀ
ed with 1a and PABA, 2a was detected.
In vitro analysis of BezG and BezJ
To investigate the functions of BezG and BezJ, we
prepared recombinant enzymes (Supplementary Figure
7). Regardless of our extensive attempts to observe BezJ
activity using various possible substrates and electron proꢀ
viders, we could not detect any oxidation activity of BezJ.
Therefore, we decided to deduce the function of BezJ with
In vitro analysis of BezE
According to the obtained results, we predicted
that BezE, annotated as a cytochrome P450, should cataꢀ
lyze the cyclization of geranylated PAcABA derivatives
(4a, 4b, and 4c) by inducing the acetate elimination couꢀ
pled with the cyclization reaction. To confirm this predicꢀ
tion, recombinant BezE was prepared; it showed a Soret
band shifting from 420 to 450 nm, when CO was bubbled
into the BezE solution (Supplementary Figures 7 and
12a). Because the possible substrates (4a, 4b, and 4c) of
BezE are highly unstable, we tried oneꢀpot reaction of
BezE, BezF, and BezG using 1a and chemically syntheꢀ
sized PHABA as substrates. When these three enzymes
were incubated with 1a and PHABA, 5a was produced as
expected (Figure 5a). When BezE was removed from the
reaction mixture, 7a was produced (Figure 5b), which was
consistent with the in vivo experiment (Figure 3f). In addiꢀ
tion, a small amount of 5a was observed in the absence of
BezE, probably due to the nonenzymatic cyclization of 4a.
a
feeding
experiment.
We
synthesized
pꢀ
hydroxyaminobenzoic acid (PHABA) and fed it to the
ꢀbezJ strain. As a result, the ꢀbezJ strain restored the proꢀ
duction of benzastatin derivatives with indoline and tetraꢀ
hydroquinoline scaffolds (Supplementary Figure 10).
Therefore, BezJ was suggested to catalyze the oxidation of
PABA to synthesize PHABA. According to this result,
PHABA was predicted to be a substrate of BezG. When
PHABA was incubated with BezG and acetylꢀCoA, we
observed the formation of a compound whose m/z correꢀ
sponded to acetylated PHABA ([M+H]+ = m/z 196) (Supꢀ
plementary Figure 11a). The product was highly unstable
and could not be isolated, but it showed a characteristic
fragment due to the loss of acetate ([M]+ = m/z 136) in
MS/MS analysis (Supplementary Figure 11b). BezG did
not show any acetylation activity against PABA (Suppleꢀ
mentary Figure 11c), indicating that the acetylation occurs
on the hydroxyl group. Therefore, we concluded that the
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 10
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 5. In vitro analysis of BezE. (a) When BezE, BezF, and BezG were incubated with PHABA, 1a, and acetylꢀCoA,
formation of 5a was observed. (b) When BezE was removed from the reaction mixture of (a), 7a and a small amount of 5a
were observed. (c) First, 1b was synthesized from GPP (1a) by BezA. After this reaction, BezE, BezF, BezG, and PHABA
were added into the reaction mixture. As a result, formation of 5b was observed. (d) When BezE was removed from the seꢀ
cond reaction of (c), formation of 7b was observed. (e) First, 1c was synthesized from 1a by BezA and BezC. After this reacꢀ
tion, BezE, BezF, BezG, and PHABA were added into the reaction mixture, resulting in the formation of 6c and 6′c. (f) When
BezE was removed from the second reaction of (e), 6c and 6′c were not formed. (gꢀh) BezEFG reaction using stable isotopeꢀ
labeled compounds. Mass spectra of 5a produced by the BezEFG reaction are shown: (g) normal reaction; (h) in the presence
of 50% H218O; (i) in the presence of 18O2.
ment. When BezE was removed from the reaction mixture,
In an independent timeꢀcourse experiment, a small amount
7b was formed (Figure 5d), which was also consistent with
of 7a, as well as a large amount of 5a, was detected in the
the in vivo experiment (Figure 3f). Next, we analyzed the
BezEFG reaction, while a large amount of 7a and a trace
role of BezE in tetrahydroquinoline synthesis. Compound
1c, which had been synthesized from 1a by BezA and BezC
amount of 5a were detected in the BezFG reaction at 120ꢀ
min incubation (5a was only detectable in MS analysis)
in vitro, was added to the reaction mixture containing
(Supplementary Figure 13a). These results indicate that
BezE, BezF, BezG, and PHABA. As a result, we observed
BezE enormously facilitate 5a formation. Interestingly, the
addition of a redox partner did not affect the rate of 5a forꢀ
the formation of 6c and 6′c (Figure 5e). Compounds 5a
and 5b were also detected because 1a and 1b should be
mation in the BezEFG reaction (Supplementary Figure
included in the reaction mixture. When BezE was removed
13b). To obtain further insights into the cyclization mechaꢀ
nism, we conducted in vitro analysis of BezE using H218O
from the reaction mixture, no formation of these comꢀ
pounds was observed (Figure 5f). From these results, we
concluded that BezE is responsible for the second ring cyꢀ
clization of benzastatins and the introduction of a Cl group
and 18O2. As a result, in the presence of 50% H218O, apꢀ
proximately 50% of 5a was labeled with 18O in the
BezEFG reaction (Figure 5g and h). In contrast, in the
presence of 18O2, 18O was not incorporated into 5a (Figure
at Cꢀ9. Because 6d was easily hydrolyzed to 6′d in the reꢀ
action buffer (Supplementary Figure 5), we think that 6′c
5i). These results strongly indicate that cyclization never
observed in the reaction is a shunt product synthesized by
occurs via epoxidation and that BezE is not a monooxygenꢀ
ase. In addition, we showed that hemin (an ironꢀcontaining
the hydrolysis of 6c.
porphyrin) could not catalyze the formation of 5a, indicatꢀ
ing that the protein scaffold of BezE is required for the cyꢀ
clization reaction (Supplementary Figure 13c).
Next, we analyzed the substrateꢀbinding spectrum
of BezE. Because 4a, 4b, and 4c (actual substrates of
BezE) were unstable, 2a, 2b, and 2c were used for the
analysis as the substrate analogs. The addition of these
compounds to BezE resulted in type II shift, which is obꢀ
served when nitrogen binds to the heme iron (Supplemenꢀ
tary Figure 12b). The dissociation constants were deterꢀ
mined to be 3.45 ± 0.86 µM (2a), 2.01 ± 0.29 µM (2b), and
BezE also catalyzed the formation of 5b when 1b,
instead of 1a, was used as a substrate (Figure 5c). Note
that 1a should be included in the reaction mixture because
we prepared 1b with the enzymatic conversion of 1a by
BezA. This is why 5a was also produced in this experiꢀ
ACS Paragon Plus Environment

Page 7 of 10
Journal of the American Chemical Society
1.56 ± 0.60 µM (2c) (Supplementary Figure 12c). This
result indicates that the heme of BezE binds with the amino
group of 4a, 4b, and 4c during the catalysis.
introduced into 5a, and how a benz[b]azepine scaffold is
1
2
3
4
5
6
7
8
synthesized. Instead of the oxidation of a double bond, our
studies indicate that this cyclization should be initiated by
activation of aromatic amine by the successive Nꢀoxidation
and Oꢀacetylation as depicted in Schemes 1d and 2. Then,
the activated aromatic amine is subjected to the cyclization
reaction. The cytochrome P450 BezE promotes and conꢀ
trols the cyclization reaction to form indoline and tetrahyꢀ
droquinoline scaffolds. This reaction probably proceeds via
(i) elimination of the acetoxy group resulting in iron
nitrenoid formation, (ii) aziridine ring formation by the
nitrene transfer to the nearest double bond of the geranyl
moiety, and (iii) aziridine ring opening by the nucleophilic
attack of HO− to Cꢀ10 and Cl− to Cꢀ9 to generate the indoꢀ
line and tetrahydroquinoline scaffolds, respectively
(Scheme 1d). Importantly, BezE does not catalyze any
oxidative reaction in this step. This is consistent with the
observation that BezE catalyzed indoline scaffold forꢀ
mation in the absence of any redox partners (ferredoxin,
ferredoxin reductase, and NAD(P)H), which are usually
required for the reaction catalyzed by cytochrome P450s.
The feeding experiment using 18O also supports this hyꢀ
pothesis; 18O of H218O, not 18O of 18O2, was incorporated
into 5a.
Discussion
In this study, we propose the following biosynthetꢀ
ic pathways for benzastatin derivatives (Scheme 2). BezA
and BezC are involved in the modification of 1a. BezA
methylates 1a to produce 1b, and BezC hydroxylates 1b to
produce 1c. Meanwhile, BezJ and BezG catalyze the seꢀ
quential hydroxylation and Oꢀacetylation of PABA to proꢀ
duce PAcABA. Then, BezF assembles these building
blocks to produce three geranylated PAcABA derivatives
(4a, 4b, and 4c). These derivatives are cyclized by BezE to
produce benzastatins with indoline and tetrahydroquinoline
structures. OꢀMethylation of 6c catalyzed by BezB results
in 6d. Compound 6d is further converted to 6′e by the nonꢀ
enzymatic nucleophilic attack of ergothioneine. Comꢀ
pounds 6′c and 6′d are probably synthesized mainly from
6c and 6d, respectively, by nonꢀenzymatic hydrolysis.
Compound 6′c also might be produced from nonꢀenzymatic
cyclization of 4c because a trace amount of 6′d was detectꢀ
ed in the ꢀbezE strain. Feeding of the “linear” benzastatin
2a to a recombinant S. lividans strain expressing bezE,
bezG, and bezJ did not lead to the formation of 5a (Supꢀ
plementary Figure 14a). Feeding of the “linear” benꢀ
zastatins 2b and 2c to the ꢀbezA and ꢀbezC mutants, reꢀ
spectively, also did not lead to the formation of cyclized
benzastatins (Supplementary Figure 14b and c). These
results strongly indicate that the “linear” benzastatins (2a,
2b, and 2c) are shunt products synthesized by geranylation
of PABA. Potentially, geranylated PHABA derivatives can
also be synthesized by BezF. However, these compounds
were not detected in the ꢀbezG strain, and therefore they
seem to be unstable. This notion was also supported by the
observation that the yield of “linear” benzastatins (2a, 2b,
and 2c) produced by the ꢀbezG strain was significantly
lower than that produced by the ꢀbezJ strain (Figure 3g
and h). Similarly, the Nꢀacetoxy intermediates (4a, 4b, and
4c) seem to be unstable because they were not detected in
the bezE strain and in vitro experiments using BezG and
BezF. When we analyzed current genome database, we
found that many actinobacteria possess putative benzastatin
biosynthetic gene clusters, suggesting that this system is
widely distributed among actinobacteria (Supplementary
Figure 15).
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
The BezEꢀcatalyzing reaction resembles the
nitrene transfer reactions catalyzed by some cytochrome
P450 mutants.38 In this reaction, engineered cytochrome
P450s eliminate N2 from an azide substrate and form a puꢀ
tative iron nitrenoid reactive intermediate. Iron nitrenoid is
used for the nitrene transfer reaction to produce an aziridine
derivative. Although an azide is commonly used as a
source of nitrene in these artificial reactions, it is reported
that an Nꢀacetoxy group also can be the source of nitrene.39
Therefore, the aboveꢀproposed reaction mechanism of
BezE is convincing. Furthermore, type II substrate binding
spectra of BezE with 2a, 2b, and 2c strongly support the
proposed reaction mechanism. Type II shift is usually obꢀ
served when an amine of an inhibitor binds to the heme to
inhibit the oxidative reaction. Because 2a, 2b, and 2c are
analogs of 4a, 4b, and 4c, respectively, the observed shifts
strongly indicate that the amines of 4a, 4b, and 4c also
should coordinate with the heme iron during catalysis. This
binding state is suitable for iron nitrenoid formation. Furꢀ
thermore, hemin did not catalyze 5a formation, suggesting
that protein scaffold of BezE is critical for this reaction.
Taken together, our results demonstrate that the nitrene
transfer reaction used by artificial cytochrome P450s is also
used in nature to biosynthesize natural products. Phylogeꢀ
netic analysis showed that BezE and BezE homologues
encoded by the putative benzastatin biosynthetic gene clusꢀ
ters are evolutionally related to AmphL,40 NysL,41 PimD,42
PenM,43 and PntM,43 all of which catalyze monooxygenaꢀ
tion with reducing partners (Supplementary Figure 16b).
Thus, BezE seems to have evolved from a P450 catalyzing
monooxygenation to develop its unique nitrene transfer
activity. Among BezE and BezE homologues, the heme
binding region, including the Cys residue that coordinates
with heme iron, is well conserved (Supplementary Figure
16a).
The formation of indoline and tetrahydroquinoline
structures is the most interesting reaction in benzastatin
biosynthesis. We initially expected that these compounds
were synthesized by the cyclization of “linear” benzastatins
(2a, 2b, and 2c) as shown in Scheme 1c. This reaction is
most likely to proceed via epoxidation of the double bond
followed by nucleophilic attack of the aromatic amine to
the epoxide, in an analogous manner to polyether biosynꢀ
thesis as proposed by Yoo et al.26,37 However, this mechaꢀ
nism could not explain why two putative oxygenases (BezJ
and BezE) and a putative Nꢀacetyltransferase (BezG) are
required for the correct cyclization, why 18O of H218O was
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 10
NH2
NH2
BezF
R1
PPO
PPO
1
2
3
4
5
6
7
8
HO
13
HO
O
O
17
GPPs
GPP (1a
)
PABA
R2
R1=R2=H: 7-hydroxyl benzastatin J (2a
R1=Me, R2=H: 7-hydroxyl benzastatin B (2b
BezA
)
BezJ
)
H
N
R1=Me, R2=OH: 7-hydroxyl O-demethylbenzastatin A (2c
)
OH
HO
H
N
R1
MGPP (1b
)
O
PHABA
HO
BezC
9
O
O
If R2=H
BezG
O
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
NE
O
NH
O
R1=H: benzastain K (7a)
R1=Me: methylbenzastain K (7b
PPO
)
NH
BezF
R1
OH
HMGPP (1c
HO
O
HO
OH
)
H
H
GPPs
O
N
R2
10
If R2=H
BezE
or
9
PAcABA
R1=R2=H: (4a
R1=Me, R2=H: (4b
R1=Me, R2=OH: (4c
)
HO
R1
)
GPP modification
PABA modification
)
O
NE (weak)
R1=H: JBIR-67 (5a
R1=Me: 7-hydroxyl benzastain F (5b
)
If R2= OH
NE
)
If R2= OH
BezE
OH
OH
OH
Cl
H
H
N
N
NE
HO
HO
HO
O
O
cyclization
7-hydroxyl O-demethylbenzastatin D (6'c
)
O-demethylvirantmycin (6c
)
O
S
BezB
BezB
H
N
O
H
O
H
N
N
HO
H
NE
10
NE
N
9
HO
O
N+
OH
Cl
N
O
O
HO
JBIR-73 (6'e
)
7-hydroxyl benzastatin D (6'd
)
virantmycin (6d)
O
Scheme 2. Proposed biosynthetic pathway for benzastatin derivatives. NE, nonenzymatic; PABA, pꢀaminobenzoic acid;
GPP, geranyl pyrophosphate; MGPP, methylgeranyl pyrophosphate; HMGPP, hydroxymethylgeranyl pyrophosphate. Meanꢀ
ing of compound number and suffix: 1, GPP derivatives; 2, “linear” benzastatins derived from PABA; 4, biosynthesis interꢀ
mediates derived from GPP derivatives and PAcABA; 5, benzastatins with an indoline scaffold; 6, benzastatins with a tetraꢀ
hydroquinoline scaffold; 6′, degradation products of 6; 7, benzastatins with a benz[b]azepine scaffold; a, no modification on
the geranyl moiety; b, a methyl group on the geranyl moiety; c, methyl and hydroxy groups on the geranyl moiety; d, methyl
and methoxy groups on the geranyl moiety; e, substitution of the chloro group of 6d with ergothioneine.
The acetoxy group formed on the aromatic amine
is unstable and subjected to hydrolysis or photolysis to
generate an active phenylnitrenium ion.44 Takeuchi et al.
reported that phenylnitrenium ion can react with a double
bond to synthesize an aziridinium ion and that the
aziridinium ion can be hydrolyzed easily.45 Thus, even in
the absence of BezE, nonꢀenzymatic cyclization can proꢀ
ceed to synthesize not only 5a and 5b, but also 6′c. Howꢀ
ever, in this nonꢀenzymatic cyclization, 7a and 7b are synꢀ
thesized in higher amounts. We speculate that the pheꢀ
nylnitrenium ion should directly react with the methyl
group (Cꢀ17) to generate a sevenꢀmembered ring because
the allylic C−H bond is relatively weak. Because 7a and 7b
are synthesized in larger amounts in the absence of BezE
than in the presence of BezE, we consider that these comꢀ
pounds are shunt products. Therefore, we propose that
BezE is required for the proper secondꢀring cyclization of
benzastatin derivatives.
Intriguingly, the hydroxyl group of the geranyl
moiety seems to determine the cyclization specificity.
When the geranyl moiety has been hydroxylated at Cꢀ17,
the cyclization occurs to generate a tetrahydroquinoline
scaffold, while an indoline scaffold is generated in the abꢀ
sence of this hydroxyl group. According to the synthetic
study on benzastatin E, the absolute stereochemistry of 5a
and 5b should be (9S,10R) (Supplementary Table 8).23,46
In contrast, the absolute stereochemistry of 6d was reported
to be (9R,10R), indicating that 6c also should have the
same stereochemistry, (9R,10R).47 These indoline and tetꢀ
rahydroquinoline compounds can be synthesized from the
similar aziridine intermediates with (9S,10S) stereochemisꢀ
try by SN2 reaction (Scheme 1d and Supplementary Figꢀ
ure 17). We speculate that the steric hindrance caused by
the presence of the hydroxyl group at Cꢀ17 inhibits the nuꢀ
cleophilic attack of HO− by shutting out HO− from the acꢀ
tive site. Thus, a Cl− ion, probably captured near the active
site, can efficiently react with Cꢀ9 to synthesize the tetraꢀ
ACS Paragon Plus Environment

Page 9 of 10
Journal of the American Chemical Society
We thank H. Onaka (pTYM19gt and pTYM2k), Y. Zhang
hydroquinoline scaffold. In the absence of this hydroxyl
group at Cꢀ17, HO− can enter into the cavity that would be
occupied by the hydroxyl group. We speculate that this
cavity is closer to Cꢀ10 than Cꢀ9, so that the nucleophilic
attack of HO− can occur to Cꢀ10 to generate the indoline
scaffold.
(GZB05ꢀdir), and A. Miyanaga and A. Arisawa (pET28bꢀ
camA and pET28bꢀcamB) for providing experimental maꢀ
terials. This research was supported in part by a funding
program for next generation worldꢀleading researchers
from the Bureau of Science, Technology, and Innovation
Policy, Cabinet Office, Government of Japan (to Y.O.), a
GrantꢀinꢀAid for Scientific Research on Innovative Areas
from the Ministry of Education, Culture, Sports, Science
and Technology of Japan (MEXT) (to Y.K.), a grant from
the New Energy and Industrial Technology Development
Organization (NEDO) of Japan (to K.S. and N.S.), and Jaꢀ
pan Society for the Promotion of Science (JSPS) A3 Foreꢀ
sight Program.
1
2
3
4
5
6
7
8
9
Another interesting point in the reaction catalyzed
by BezE is how the chloro group is introduced. Most chloro
groups of secondary metabolites are introduced by oxidasꢀ
es, such as flavinꢀdependent halogenases.48–50 There are
also enzymes that catalyze a nucleophilic halogenation of
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
SAM by using L
ꢀmethionine as a leaving group.51,52 Haloꢀ
genation in the benzastatin biosynthetic pathway uses neiꢀ
ther of these mechanisms and probably occurs nucleophilꢀ
ically, coupled with aziridine ring opening. We speculate
that the active site of BezE has a specific site to accommoꢀ
date Cl− and a catalytic acid to enhance the aziridine ring
opening, so that BezE can specifically catalyze nucleoꢀ
philic addition of Cl− at Cꢀ9 when 4c was used as a subꢀ
strate (Scheme 1d). However, further structural and bioꢀ
chemical studies on BezE are required for proposing enzyꢀ
matic mechanism for halogenation.
REFERENCES
(1) Pang, B.; Wang, M.; Liu, W. Nat. Prod. Rep. 2016, 33,
162.
(2) Tang, M. C.; Zou, Y.; Watanabe, K.; Walsh, C. T.;
Tang, Y. Chem. Rev. 2017, 117, 5226.
(3) Xu, L. H.; Fushinobu, S.; Takamatsu, S.; Wakagi, T.;
Ikeda, H.; Shoun, H. J. Biol. Chem. 2010, 285, 16844.
(4) Shrestha, P.; Oh, T. J.; Liou, K.; Sohng, J. K. Appl. Mi-
crobiol. Biotechnol. 2008, 79, 555.
(5) Takahashi, S.; Nagano, S.; Nogawa, T.; Kanoh, N.;
Uramoto, M.; Kawatani, M.; Shimizu, T.; Miyazawa, T.;
Shiro, Y.; Osada, H. J. Biol. Chem. 2014, 289, 32446.
(6) Pohle, S.; Appelt, C.; Roux, M.; Fiedler, H. P.; Süssꢀ
muth, R. D. J. Am. Chem. Soc. 2011, 133, 6194.
(7) Gaisser, S.; Lill, R.; Staunton, J.; Méndez, C.; Salas, J.;
Leadlay, P. F. Mol. Microbiol. 2002, 44, 771.
(8) Malla, S.; Thuy, T. T. T.; Oh, T. J.; Sohng, J. K. Arch.
Microbiol. 2011, 193, 95.
In summary, analysis of the benzastatin biosynꢀ
thetic gene cluster enabled us to propose unprecedented
cyclization and halogenation reactions catalyzed by BezE.
To the best of our knowledge, BezE is the first native cytoꢀ
chrome P450 that was reported to catalyze nitrene transfer.
We propose “P450 nitrene transferase” for BezE.
Accession codes
The bez gene cluster was deposited in the DNA Data Bank
(9) Funa, N.; Funabashi, M.; Ohnishi, Y.; Horinouchi, S. J.
Bacteriol. 2005, 187, 8149.
of Japan (DDBJ) under the accession number LC177364.
ASSOCIATED CONTENT
(10) Zhao, B.; Lamb, D. C.; Lei, L.; Kelly, S. L.; Yuan, H.;
Hachey, D. L.; Waterman, M. R. Biochemistry. 2007, 46,
8725.
(11) Zhao, B.; Guengerich, F. P.; Bellamine, A.; Lamb, D.
C.; Izumikawa, M.; Lei, L.; Podust, L. M.; Sundaramoorꢀ
thy, M.; Kalaitzis, J. A.; Reddy, L. M.; Kelly, S. L.; Moore,
B. S.; Stec, D.; Voehler, M.; Falck, J. R.; Shimada, T.; Waꢀ
terman, M. R. J. Biol. Chem. 2005, 280, 11599.
(12) Girvan, H. M.; Munro, A. W. Curr. Opin. Chem. Biol.
2016, 31, 136.
Supporting Information.
The supporting information is available free of charge on
the ACS publication website.
Materials and methods, Supplementary tables, and suppleꢀ
mentary figures about experimental details, structural eluꢀ
cidation (tandem mass spectra and NMR spectra), funcꢀ
tional analysis of bez genes in vivo and in vitro, wide distriꢀ
bution of bez cluster homologues, and sequence alignment
and phylogenetic analysis of BezE and P450s (PDF).
(13) Podust, L. M.; Sherman, D. H. Nat. Prod. Rep. 2012,
29, 1251.
(14) Zhang, X.; Li, S. Nat Prod Rep. 2017, 34, 1061.
(15) Rudolf, J. D.; Chang, C. Y.; Ma, M.; Shen, B. Nat
Prod Rep. 2017, 34, 1141.
(16) Brandenberg, O. F.; Fasan, R.; Arnold, F. H. Curr.
Opin. Biotechnol. 2017, 47, 102.
AUTHOR INFORMATION
Corresponding Authors
*Y.O.; ayasuo@mail.ecc.uꢀtokyo.ac.jp
*Y. K.; aykatsu@mail.ecc.uꢀtokyo.ac.jp
(17) Coelho, P. S.; Brustad, E. M.; Kannan, A.; Arnold, F.
H. Science. 2013, 339, 307.
(18) Kim, W. G.; Kim, J. P.; Kim, C. J.; Lee, K. H.; Yoo, I.
D. J. Antibiot. 1996, 49, 20.
Notes
The authors declare no competing financial interests.
(19) Kim, W. G.; Kim, J. P.; Koshino, H.; ShinꢀYa, K.;
Seto, H.; Yoo, I. D. Tetrahedron. 1997, 53, 4309.
(20) Kim, W. G.; Ryoo, I. J.; Park, J. S.; Yoo, I. D. J. Anti-
biot. 2001, 54, 513.
Acknowledgements
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 10
(21) Lee, J. G.; Yoo, I. D.; Kim, W. G. Biol. Pharm. Bull.
2007, 30, 795.
(37) Hotta, K.; Chen, X.; Paton, R. S.; Minami, A.; Li, H.;
Swaminathan, K.; Mathews, I. I.; Watanabe, K.; Oikawa,
H.; Houk, K. N.; Kim, C. Y. Nature. 2012, 483, 355.
(38) Farwell, C. C.; Zhang, R. K.; McIntosh, J. A.; Hyster,
T. K.; Arnold, F. H. ACS Cent. Sci. 2015, 1, 89–93.
(39) Singh, R.; Kolev, J. N.; Sutera, P. A.; Fasan, R. ACS
Catal. 2015, 5, 1685.
1
2
3
4
5
6
(22) Nakagawa, A.; Iwai, Y.; Hashimoto, H.; Miyazaki, N.;
Oiwa, R.; Takahashi, Y.; Hirano, A.; Shibukawa, N.;
Kojima, Y.; Omura, S. J. Antibiot. 1981, 34, 1408.
(23) Motohashi, K.; Nagai, A.; Takagi, M.; Shinꢀya, K. J.
Antibiot. 2011, 64, 281.
(24) Kim, W. G.; Kim, J. P.; Yoo, I. D. J. Antibiot. 1996,
49, 26.
(40) Byrne, B.; Carmody, M.; Gibson, E.; Rawlings, B.;
Caffrey, P. Chem. Biol. 2003, 10, 1215.
7
8
(25) HirotaꢀTakahata, Y.; Kobayashi, H.; Kizuka, M.;
Ohyama, T.; KitamuraꢀMiyazaki, M.; Suzuki, Y.; Fujiwara,
M.; Nakajima, M.; Ando, O. J. Antibiot. 2016, 69, 747.
(26) Yoo, S.; Kim, J.; Yi, K. Y. Bull Korean Chem Soc.
1999, 20, 139.
(27) Hayashi, T. Study on the biosynthesis of benzastatin
derivatives from Streptomyces. Ph. D. Thesis. The Univerꢀ
sity of Tokyo, Tokyo, Japan, March, 2013.
(28) Young, I. G.; Leppik, R. A.; Hamilton, J. A.; Gibson,
F. J. Bacteriol. 1972, 110, 18.
(29) Koch, M.; Lemke, R.; Heise, K. P.; Mock, H. P. Eur.
J. Biochem. 2003, 270, 84.
(30) Singh, S.; McCoy, J. G.; Zhang, C.; Bingman, C. A.;
Phillips, G. N.; Thorson, J. S. J. Biol. Chem. 2008, 283,
22628.
(31) Wang, K. C.; Ohnuma, S. Biochim. Biophys. Acta.
2000, 1529, 33.
(32) Rossi, M.; Amaretti, A.; Raimondi, S. Nutrients. 2011,
3, 118.
(33) He, J.; Hertweck, C. J. Am. Chem. Soc. 2004, 126,
3694.
(34) Chanco, E.; Choi, Y. S.; Sun, N.; Vu, M.; Zhao, H.
Bioorg. Med. Chem. 2014, 22, 5569.
(41) Brautaset, T.; Sletta, H.; Degnes, K. F.; Sekurova, O.
N.; Bakke, I.; Volokhan, O.; Andreassen, T.; Ellingsen, T.
E.; Zotchev, S. B. Appl. Environ. Microbiol. 2011, 77,
6636.
(42) Kells, P. M.; Ouellet, H.; SantosꢀAberturas, J.; Apariꢀ
cio, J. F.; Podust, L. M. Chem. Biol. 2010, 17, 841.
(43) Zhu, D.; Seo, M. J.; Ikeda, H.; Cane, D. E. J. Am.
Chem. Soc. 2011, 133, 2128.
(44) Chakraborty, M.; Jin, K. J.; Glover, S. A.; Novak, M.
J. Org. Chem. 2010, 75, 5296.
(45) Takeuchi, H.; Koyama, K.; Mitani, M.; Ihara, R.; Uno,
T.; Okazaki, Y.; Kai, Y.; Kasai, N. J. Chem. Soc. Perkin
Trans. 1. 1985, 1985, 677.
(46) Toda, N.; Ori, M.; Takami, K.; Tago, K.; Kogen, H.
Org. Lett. 2003, 5, 269.
(47) Ori, M.; Toda, N.; Takami, K.; Tago, K.; Kogen, H.
Angew. Chem. Int. Ed Engl. 2003, 42, 2540.
(48) Blasiak, L. C.; Drennan, C. L. Acc. Chem. Res. 2009,
42, 147.
(49) van Pée, K. H.; Patallo, E. P. Appl. Microbiol. Bio-
technol. 2006, 70, 631.
(50) Nakamura, H.; Schultz, E. E.; Balskus, E. P. Nat.
Chem. Biol. 2017, 13, 916.
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
(35) Choi, Y. S.; Zhang, H.; Brunzelle, J. S.; Nair, S. K.;
Zhao, H. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 6858.
(36) Nakajima, S.; Satoh, Y.; Yanashima, K.; Matsui, T.;
Dairi, T. J. Biosci. Bioeng. 2015, 120, 294.
(51) Eustáquio, A. S.; Pojer, F.; Noel, J. P.; Moore, B. S.
Nat. Chem. Biol. 2008, 4, 69.
(52) Deng, H.; O’Hagan, D.; Schaffrath, C. Nat. Prod. Rep.
2004, 21, 773.
Table of Contents Artwork
ACS Paragon Plus Environment