Enzymatic epoxide-opening cascades catalyzed by a pair of epoxide hydrolases in the ionophore polyether biosynthesis

Article abstract of DOI:10.1021/ol200100e

Our recent findings of the first epoxide hydrolase Lsd19, involved in lasalocid A biosynthesis, led us to investigate a long-standing controversial issue on the mechanism of enzymatic epoxide-opening cascades. The site-directed mutagenesis and domain diss

Full text of DOI:10.1021/ol200100e

ORGANIC
LETTERS
2011
Vol. 13, No. 7
1638–1641
Enzymatic Epoxide-Opening Cascades
Catalyzed by a Pair of Epoxide Hydrolases
in the Ionophore Polyether Biosynthesis
Atsushi Minami,† Akira Migita,† Daiki Inada,† Kinya Hotta,‡ Kenji Watanabe,§
Hiroki Oguri,† and Hideaki Oikawa*,†
Division of Chemistry, Graduate School of Science, Hokkaido University, Sapporo 060-
0810, Japan, Department of Biological Sciences, National University of Singapore,
Singapore 117543, Republic of Singapore, and School of Pharmaceutical Sciences,
University of Shizuoka, Shizuoka, 422-8526, Japan
hoik@sci.hokudai.ac.jp
Received January 13, 2011
ABSTRACT
Our recent findings of the first epoxide hydrolase Lsd19, involved in lasalocid A biosynthesis, led us to investigate a long-standing controversial
issue on the mechanism of enzymatic epoxide-opening cascades. The site-directed mutagenesis and domain dissection analysis to reveal the
mechanism of the reaction catalyzed by Lsd19 is examined, especially in the role of acidic amino acid pair and catalytic domains.
Natural polyether metabolites show significant diversity
in molecular skeletons found in ionophore antibiotics,1
marine ladder toxins,2 polyether triterpenes,3 and anno-
naceous antitumor acetogenins4 (Figure S1, Supporting
Information). Asrepresentedbythe Cane-Celmer-West-
ley unified biogenetic scheme for ionophore antibiotics5
and the Nakanishi-Shimizu biogenetic scheme for marine
ladder toxins,6,7 it has been suggested that these polyether
skeletons are constructed by nucleophilic epoxide-opening
cascades of the corresponding polyepoxides after installing
a number of stereogenic centers on the simple polyolefins
in enzymatic epoxidation. This simple yet versatile strategy
for the biosynthesis of complex polyether metabolites
prompted chemists to synthesize ionophore antibiotics
and marine dinoflagellate toxins using epoxide-opening
cascades.8-10 The successful biomimetic synthesis of var-
ious polyether systems supports these biosynthetic pro-
posals. In parallel with these synthetic studies, extensive
biosynthetic studies of polyethers, especially ionophore
antibiotics,11,12 have also been carried out and a number
of experimental results, including identification of the gene
cluster for monensin biosynthesis13 and several gene
disruption experiments,14,15 strongly supported the poly-
epoxide model of ionophore polyether biosynthesis.
† Hokkaido University.
‡ National University of Singapore.
§ University of Shizuoka.
(1) Dutton, C. J.; Banks, B. J.; Cooper, C. B. Nat. Prod. Rep. 1995,
12, 165–181.
(2) Yasumoto, T.; Murata, M. Chem. Rev. 1993, 93, 1897–1909.
(8) Faul, M. M.; Huff, B. E. Chem. Rev. 2000, 100, 2407–2473.
(9) Nicolaou, K. C.; Frederick, M. O.; Aversa, R. J. Angew. Chem.,
Int. Ed. 2008, 47, 7182–7225.
(10) Vilotijevic, I.; Jamison, T. F. Science 2007, 317, 1189–1192.
(11) Hutchinson, C. R.; Sherman, M. M.; Vederas, J. C.; Nakashima,
T. T. J. Am. Chem. Soc. 1981, 103, 5953–5956.
ꢀ
(3) Fernandez, J. J.; Souto, M. L.; Norte, M. Nat. Prod. Rep. 2000,
17, 235–246.
ꢁ
(4) Bermejo, A.; Figadere, B.; Zafra-Polo, M.-C.; Barrachina, I.;
Estornell, E.; Cortes, D. Nat. Prod. Rep. 2005, 22, 269–303.
(5) Cane, D. E.; Celmer, W. D.; Westley, J. W. J. Am. Chem. Soc.
1983, 105, 3594–3600.
(6) Shimizu, Y. Chem. Rev. 1993, 93, 1685–1698.
(7) Nakanishi, K. Toxicon 1985, 23, 473–479.
(12) Hutchinson, C. R.; Sherman, M. M.; McInnes, A. G.; Walter,
J. A.; Vederas, J. C. J. Am. Chem. Soc. 1981, 103, 5956–5959.
(13) Oliynyk, M.; Stark, C. B. W.; Bhatt, A.; Jones, M. A.; Hughes-
Thomas, Z. A.; Wilkinson, C.; Oliynyk, Z.; Demydchuk, Y.; Staunton,
J.; Leadlay, P. F. Mol. Microbiol. 2003, 49, 1179–1190.
r
10.1021/ol200100e
Published on Web 03/04/2011
2011 American Chemical Society

Until recently, however, no experimental evidence in sup-
port of the proposals at the enzymatic level has been
provided.
oxazolidinone skeleton (Figure 1A, B).20,21 The relaxed
substrate tolerance indicates that Lsd19 provides an intri-
guing model system in which to investigate the mechanism
of polyether formation using bisepoxide analogs. Here, we
report site-directed mutagenesis and domain dissection
analysis to investigate the catalytic mechanism of poly-
ether ring formation catalyzed by Lsd19.
First, we examined the kinetic parameters of the second
epoxide-opening reaction using monocyclic ether inter-
mediates 5a and 5b to simplify analysis by excluding the
sequential epoxide-opening reaction. The Michae-
lis-Menten equation was fitted to plots of initial velocity
vs substrate concentration to yield kcat = 1.6 ( 0.4 s-1
,
Km = 88 ( 12 μM, and kcat/Km = 1.8 ( 0.2 Â 10-2
s
-1 μM
-1 for 5a and kcat = 2.1 ( 1.1 s-1, Km = 370 ( 42 μM, and
kcat/Km = 0.57( 0.2 Â 10-2
s
-1 μM -1 for 5b, respectively.
Thus, Lsd19 is 3.2-fold more specific for 5a than 5b. These
observations indicated that simplified analog 5a is a more
suitable substrate for Lsd19 reaction. In addition, as these
results correspond approximately to our previous observa-
tion that the reactivity of epoxide-opening reactions
decreased in the order 3a, 3b, 4a, 4b, and C15-C24-
bisepoxide analog,20 it is anticipated that putative biosyn-
thetic intermediate 3a may show the lowest Km value
among these compounds.
Figure 1. (A) Lsd19 catalyzing sequential epoxide-opening re-
action from plausible intermediate 3 to 1. (B) Lsd19 reaction
with substrate analogs.
To date, epoxide-opening reactions are divided into two
distinct types according to the reaction mechanism: step-
wise hydrolysis and direct hydrolysis. In the reaction
catalyzed by the former type of EHs, aspartic acid acts as
a nucleophile to afford covalent enzyme-substrate com-
plex and then hydrolysis of the intermediate results in diol
formation (Figure S2A, Supporting Information).22,23 On
the other hand, in the reaction catalyzed by the latter type
Smith et al. and we independently identified the gene
cluster of the polyether antibiotic lasalocid A (1),16,17
which has a tetrahydrofuran (THF) ring and tetrahydro-
pyran (THP) ring interconnected via a carbon-carbon
bond.18 Furthermore, we identified the function of epoxide
hydrolase (EH) Lsd19 in catalysis of the epoxide-opening
cascades with a energetically favored 5-exo cyclization and
disfavored 6-endo cyclization from the putative biosyn-
thetic intermediate, bisepoxyprelasalocid (3a), to 1 (Figure
1A).19 This is the first experimental evidence for the
enzymatic epoxide-opening cascades. Gene disruption ex-
periments of epoxide hydrolase also demonstrated the
function as a catalysis of the epoxide-opening cascades.16
To characterize this unique enzyme, we performed an in
vitro study using structurally diverse bisepoxides. The
results showed that Lsd19 can catalyze the epoxide-open-
ing reaction with substrate analogs, such as benzyl-pro-
tected analog 3b, C12-C24-bisepoxide 4a lacking the
salicylate moiety of 3a, and C13-C24-bisepoxide 4b in
which the left-hand segment of 3a is replaced with an
of EHs such as limonene 1,2-epoxide hydrolase (LEH),24
a
pair of acidic amino acid residues was revealed to play key
roles in the epoxide-opening reaction with activation of the
nucleophilic water and epoxide (Figure S2B, Supporting
Information).25,26 Because only the latter mechanism can
be applied to epoxide-opening cascades involved in iono-
phore antibiotics biosynthesis, we expected Lsd19 to have
a similar acidic amino acid pair for construction of the
polyether system.
Recent genetic analysis of polyether antibiotics identi-
fied several putative EHs. Two fused N- and C-terminal
(20) Matsuura, Y.; Shichijo, Y.; Minami, A.; Migita, A.; Oguri, H.;
Watanabe, M.; Tokiwano, T.; Watanabe, K.; Oikawa, H. Org. Lett.
2010, 12, 2226–2229.
(14) Bhatt, A.; Stark, C. B. W.; Harvey, B. M.; Gallimore, A. R.;
Demydchuk, Y. A.; Spencer, J. B.; Staunton, J.; Leadlay, P. F. Angew.
Chem., Int. Ed. 2005, 44, 7075–7078.
(15) Gallimore, A. R.; Stark, C. B. W.; Bhatt, A.; Harvey, B. M.;
Demydchuk, Y.; Bolanos-Garcia, V.; Fowler, D. J.; Staunton, J.;
Leadlay, P. F.; Spencer, J, B. Chem. Biol. 2006, 13, 453–460.
(16) Smith, L.; Hong, H.; Spencer, J, B.; Leadlay, P, F. ChemBio-
Chem 2008, 9, 2967–2975.
(17) Migita, A.; Watanabe, M.; Hirose, Y.; Watanabe, K.; Tokiwano,
T.; Kinashi, H.; Oikawa, H. Biosci. Biotechnol. Biochem. 2009, 73, 169–176.
(18) Westley, J. W.; Blount, J. F.; Evans, R. H., Jr.; Stempel, A.;
Berger, J. J. Antibiot. 1974, 27, 597–603.
(19) Shichijo, Y.; Migita, A.; Oguri, H.; Watanabe, M.; Tokiwano,
T.; Watanabe, K.; Oikawa, H. J. Am. Chem. Soc. 2008, 130, 12230–
12231.
(21) The spectral data of the compounds appeared in this paper were
found in the Supporting Information of refs 19 and 20.
(22) Arand, M.; Wagner, H.; Oesch, F. J. Biol. Chem. 1996, 271,
4223–4229.
(23) Yamada, T.; Morisseau, C.; Maxwell, J. E.; Argiriadi, M. A.;
Christianson, D. W.; Hammock, B. D. J. Biol. Chem. 2000, 275, 23082–
23088.
(24) van der Werf, M. J.; Orru, R. V. A.; Overkamp, K. M.; Swarts,
H. J.; Osprian, I.; Steinreiber, A.; de Bont, J. A. M.; Faber, K. Appl.
Microbiol. Biotechnol. 1999, 52, 380–385.
(25) Arand, M.; Hallberg, B. M.; Zou, J.; Bergfors, T.; Oesch, F.; van
der Werf, M. J.; de Bont, J. A. M.; Jones, T. A.; Mowbray, S. L. EMBO
J. 2003, 22, 2583–2592.
(26) Johansson, P.; Unge, T.; Cronin, A.; Arand, M.; Bergfors, T.;
Jones, T. A.; Mowbray, S. L. J. Mol. Biol. 2005, 351, 1048–1056.
Org. Lett., Vol. 13, No. 7, 2011
1639

domains of Lsd19 (Lsd19A: 133 aa, Lsd19B: 149 aa
(Figure S3A, Supporting Information) resemble the
known EH genes, such as monBII (140 aa) and monBI
(144 aa) involved in monensin biosynthesis,10 nigBI (145
aa), nigBII (155 aa) involved in nigericin biosynthesis,27
and tmnB (141 aa) required for the single epoxide-opening
reaction intetronomycinbiosynthesis.28 Inparticular, nanI
(313 aa) involved in nanchangmycin biosynthesis29 shows
the same fused domain organization as lsd19.16,17 These
observations indicate that a pair of EHs, which exists
as either individual or fused subunits, is important
for polyether formation. Based on these findings, multi-
ple sequence alignment of these EHs and EH-domains
was examined (Figure S3B, Supporting Information).
Although key acidic residues of LEH are located at
different positions to those of target EHs, we found that
acidic amino acid residues, possible candidates for cataly-
tic residues, are highly conserved among these putative
polyether EHs. Therefore, these conserved acidic residues
located in Lsd19A (Glu22, Asp33, Asp38, and Glu65) and
Lsd19B domains (Asp154, Arg165, Glu169, Asp170,
Glu181, Glu197, and Asp207) were mutated to alanine
to investigate the roles for epoxide opening reaction.
We first examined the unique enzyme-required process
toaffordenergetically disfavored 6-endoproduct 6a(THF-
THP).30 In the enzyme reactions with cell-free extracts of
eachmutant and substrate analog 4afor 60min (FigureS4,
Supporting Information), formation of 6a was predomi-
nantly observed in the case of nine mutants, similar to the
wild-type reaction. On the other hand, in the reaction with
D170A and E197A mutants, production of 6a was com-
pletely abolished and monocyclic ether 5a was significantly
accumulated (88, 82%) accompanying slow nonenzymatic
formation of THF-THF product 7a (5%, 10%). These
results indicated that only D170A and E197A mutants lost
the catalytic activity of the 6-endo cyclization reaction,
while the other nine mutants retained essentially the same
activity as the wild-type control.
Figure 2. LC-MS profiles of the reaction products with various
mutants from (A) 5 min reaction and (B) 60 min reaction.
As expected from the significant homology between
Lsd19A and Lsd19B (Figure S3A, Supporting Infor-
mation), the corresponding acidic amino acids to Asp170
and Glu197 are conserved in Lsd19A (Asp38 and Glu65).
To focus on the first cyclization, double mutants (D38A/
E197A and E65A/E197A) were constructed and the reac-
tion time was set to 5 min. These mutants showed the same
LC-MS profiles compared with control reaction with-
out enzyme (Figure 2A, B and Figure S7, Supporting
Information). Loss of the cyclization activity established
that Asp38 and Glu65 located in the Lsd19A play key roles
in the 5-exo epoxide-opening reaction.
Because each epoxide opening reactions are indepen-
dently catalyzed by the individual domains (Lsd19A and
Lsd19B), we predicted that discrete Lsd19A and Lsd19B
may catalyze the THF and THP formation, respectively.
To investigate this hypothesis, Lsd19A and Lsd19B were
expressed as thioredoxin fusion proteins (Figure S5, Sup-
porting Information), and in vitro analysis using equimolar
amounts of each domain was carried out with 4a.
Although cyclization activity was not detected in Lsd19A
reaction within the experimental error caused by competi-
tion with the nonenzymatic reaction, very low-level but
reproducible formation of 6a (25%) was observed in
Lsd19B reaction (Figure S8, Supporting Information).
Based on these findings and additional data with the
substrate analog 4b, it was indicated that Lsd19B catalyzes
6-endo cyclization of 5ab to afford 6ab. All our efforts to
restore the enzymatic activity for sequential cyclization,
For detailed functional analysis of D170A and E197A
mutants, 4a was incubated with purified mutants (Figure
S5, S6, Supporting Information) for 5 min. LC-MS anal-
ysis of the reaction products showed a clear increase in 5a
production (43% for D170A, 52% for E197A) compared
with nonenzymatic reaction (22%) (Figure 2A), but none
of 6a. Formation of 6a was not observed even in prolonged
reaction time (60 min) (Figure 2B). These results indicated
that both D170A and E197A mutants had retained the
5-exo epoxide-opening activity but lost the second 6-endo
epoxide-opening activity.
(27) Harvey, B. M.; Mironenko, T.; Sun, Y.; Hong, H.; Deng, Z.;
Leadlay, P. F.; Weissman, K. J.; Haydock, S. F. Chem. Biol. 2007, 14,
703–714.
(28) Demydchuk, Y.; Sun, Y.; Hong, H.; Staunton, J.; Spencer, J. B.;
Leadlay, P. F. ChemBioChem 2008, 9, 1136–1145.
(29) Sun, Y.; Zhou, X.; Dong, H.; Tu, G.; Wang, M.; Wang, B.;
Deng, Z. Chem. Biol. 2003, 10, 431–441.
(30) Without purification of the Lsd19 mutants, enzymatic activity
was securely detected by the formation of 6a which was not observed in
the nonenzymatic reaction. Therefore, initial characterization of each
mutants was performed with cell-free extracts.
1640
Org. Lett., Vol. 13, No. 7, 2011

such as cleavage of thioredoxin-tag resion and admixing of
each dissected domains, were unsuccessful.
proposed mechanism catalyzed by a pair of EH may
explain other ionophore polyether biosyntheses with triep-
oxide intermediates such as monensin and nigericin. Re-
cent findings that most of the ionophore antibiotic gene
clusters including monensin and nigericin have only two
EHs or fused two EH domains suggest that a single EH or
EH domain catalyzes two rounds of epoxide-opening
reactions (Figure S9, Supporting Information). Further-
more, identification of the catalytic domain Lsd19B in
energetically disfavored 6-endo cyclization set the stage to
analyze an important factor to control the regioselectivity
in an epoxide-opening reaction. The stereostructures of
each domain are essential for this purpose. Thus, we are
currently working to solve the X-ray crystal structure of
Lsd19 complexed with a substrate analog.
The above mutational analyses and domain dissection
analysis of Lsd19 demonstrated that individual domains of
Lsd19 catalyze epoxide-opening cascades independently in
lasalocid biosynthesis. In addition, loss of the activity with
mutants in Lsd19A (D38/E197A or E65/E197A) and
Lsd19B (D170 or E197) suggested that acidic amino acid
pair play critical roles in these reactions. Based on these
findings, the catalytic mechanism of Lsd19 is proposed as
follows (Figures 3, 4). First, Lsd19A recognizes the inter-
nal C18-C19 epoxide of the substrate and catalyzes
energetically favored 5-exo cyclization. Either D38 or
E65 directly acts as an base for deprotonation of the
hydroxyl group. The other acidic amino acid may be
involved in recognition and/or activation of the epoxide
in a similar manner to LEH reaction. Then, energetically
disfavored 6-endo cyclization is catalyzed by Lsd19B with
another acidic amino acid pair (D170-E197) in a similar
manner. The highly conserved acidic amino acid pair in
various EHs involved in polyether biosynthesis (Figure S3,
Supporting Information) suggests that the mechanism
shown above for polyether construction is widely applic-
able to the biosynthesis of ionophore polyethers.
Figure 4. Epoxide-opening cascades catalyzed by Lsd19 in
lasalocid A biosynthesis.
Acknowledgment. This work was supported by the
Grants-in-Aid for Scientific Research [22108002 to H.
Oikawa and 21710213 to A. Minami] from Japan Society
for the Promotion of Science (JSPS), and partially sup-
ported by the Global COE Program (Project No. B01:
Catalysis as the Basis for Innovation in Material Science)
from Monbukagakusho and the Naito Foundation. We
expressed sincere thanks to late Prof. C. R. Hutchinson
(University of Wisconsin) for his encouragement and
stimulating discussion on the lasalocid biosynthesis.
Figure 3. Catalytic mechanism of epoxide-opening reaction.
In summary, biochemical and mutational analyses of
epoxide hydrolase Lsd19 significantly advance our under-
standing of the enzymatic epoxide-opening reaction in
lasalocid A biosynthesis, especially in the catalytic mech-
anism including the role of each catalytic domain. The
Supporting Information Available. Experimental de-
tails and LC-MS data. This material is available free of
charge via the Internet at http://pubs.acs.org.
Org. Lett., Vol. 13, No. 7, 2011
1641