Chain elongation, macrolactonization, and hydrolysis of natural and reduced hexaketide substrates by the picromycin/methymycin polyketide synthase

Article abstract of DOI:10.1002/anie.200502246

(Chemical Equation Presented) Dual action: An analogue of a natural hexaketide thioester substrate is incubated with recombinant picromycin/methymycin synthase (PICS) module 6 with its attached thioesterase (TE). A mixture of 12- and 14-membered-ring macrolactones 10-deoxymethynolide (1) and narbonolide (2) are generated (see scheme) by competing chain elongation and direct lactonization of the substrate. KS = ketosynthase, AT = methylmalonyl transferase, ACP = acyl carrier protein, metmal CoA = methylmalonyl coenzyme A.

Full text of DOI:10.1002/anie.200502246

Angewandte
Chemie
PICS module 6, a homodimer of subunit Mw = 142 kDa
encoded by pikAIV, is organizationally one of the simplest
known PKS modules, as it only consists of a ketosynthase
(KS), a methylmalonyl transferase (AT), and an acyl carrier
protein (ACP) domain as well as a C-terminal thioesterase
(TE; Scheme 1). In spite of this apparent structural simplicity,
PICS module 6(+TE) is one of the catalytically most intrigu-
ing PKS modules, on the basis of its demonstrated role in the
formation of both 10-deoxymethynolide (1)[7] and narbono-
lide (2), the parent macrolide agylcones of methymycin and
picromycin, respectively.[4,6,8–10]
Biosynthesis
DOI: 10.1002/anie.200502246
Chain Elongation, Macrolactonization, and
Hydrolysis of Natural and Reduced Hexaketide
Substrates by the Picromycin/Methymycin
Polyketide Synthase**
Scheme 1. Conversion of hexaketide into 10-deoxymethynolide (1) and
narbonolide (2) by PICS module 6+TE.
We have reported the expression and purification of
recombinant PICS module 6 + TE from Escherichia coli and
have investigated its relative substrate specificity by using N-
acetylcysteamine (SNAC) thioesters of 2-methyl-3-hydroxy-
pentanoic acid as diketide analogues of the natural hexake-
tide chain-elongation substrate.[11] It was found that PICS
module 6 + TE processed both syn-diketide diastereomers
(2S,3R)-3 and (2R,3S)-4 with a modest 2.5:1 preference in
kcat/Km for 4, but did not process either of the two anti dike-
tides (Scheme 2). Similar findings were reported independ-
ently by Sherman and co-workers,[9] who also described
extensive in vivo and in vitro investigations of the biochem-
ical basis for the competition between elongation of the
natural hexaketide substrate by PICS module 6 to give 2 and
direct cyclization to 1, with the final macrolactonization being
controlled in both cases by the C-terminal TE domain.[8,12]
Although active KS6, AT6, and ACP6 domains were found to
be essential for the formation of 1, the precise role of these
Jiaquan Wu, Weiguo He, Chaitan Khosla, and
David E. Cane*
Macrolide polyketides are a structurally diverse group of
natural products, widely used for their pharmaceutically
important antimicrobial, antitumor, and immunosuppressant
properties.[1] The picromycin/methymycin synthase (PICS) of
Streptomyces venezuelae, which is responsible for the forma-
tion of the 12- and 14-membered-ring antibiotics methymycin
and picromycin,[2,3] is a typical modular polyketide synthase
(PKS), in which each module is responsible for a single round
of polyketide-chain elongation and functional-group modifi-
cation.[4–6] Each module is in turn made up of a specific
combination of fatty acid synthase-like domains that catalyze
the individual biochemical steps of polyketide biosynthesis.
[*] Dr. J. Wu, W. He, Prof. Dr. D. E. Cane
Department of Chemistry
Box H, Brown University
Providence, RI 02912-9108 (USA)
Fax: (+1)401-863-9368
E-mail: david_cane@brown.edu
Prof. Dr. C. Khosla
Departments of Chemical Engineering
Chemistry and Biochemistry
Stanford University
Stanford, CA 94305 (USA)
[**] This research was supported by NIH Grants GM22172 (D.E.C) and
CA66736 (C.K.).
Scheme 2. Conversion of diketide-SNAC analogues into triketide keto-
lactones by PICS module 6+TE. metmal CoA=methylmalonyl coen-
zyme A.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 7557 –7560
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7557

Communications
domains in the delivery of the hexaketide substrate to the
in CDCl3.[13,15] Incubation of 5 with recombinant PICS
module 6 + TE and [2-14C]methylmalonyl CoA produced a
4:1 mixture of 1 and 2, as monitored by RP-HPLC with
concatenated inline UV/Vis and flow-scintillation detection
(Scheme 3). The identities of the enzymatic reaction products
were confirmed by carrying out a preparative-scale reaction
over 0.5 hours with 5 (2 mm) and methylmalonyl CoA (2 mm)
in the presence of PICS module 6 + TE (13.2 mm).[16] The
resulting macrolide aglycones, purified by preparative RP-
active site of the TE domain remains to be established.
Investigations, both by us and Sherman and co-workers, have
recently established that recombinant PICS TE alone can
efficiently catalyze the macrolactonization of synthetic seco-
hexaketide acyl SNAC 5 to 1 while catalyzing the exclusive
hydrolysis of the corresponding 7-dihydro-seco-SNAC thio-
ester 6.[13,14] Most recently, Sherman also reported that PICS
module 6 + TE can convert the seco-SNAC thioester 5 and
methylmalonyl coenzyme A (CoA) into 2, although they did
not monitor the competing direct lactonization to 1 and
reported only a
kcat/Km value for the chain elongation and the lactonization
reaction.[15] We now report that in the presence of the
cosubstrate methylmalonyl CoA PICS module 6 + TE cata-
lyzes competing macrolactonization and chain elongation of
5, which is converted into a 4:1 mixture of 1 and 2 (Scheme 3).
1
HPLC, displayed H NMR and HR-ESI mass spectra that
were identical to authentic samples of 1 as well as literature
values for 1 and 2.[17] The steady-state kinetic parameters for
the formation of both 1 and 2 were determined by carrying
out a series of incubations at concentrations of 5 that ranged
from 0.5 to 4.0 mm, with the yields of 1 and 2 determined by
HPLC/diode-array UV (230 nm) and phosphorimaging with
TLC, respectively. The corresponding kcat and Km values were
calculated by direct fitting of the plots of v versus [S] to the
Michaelis–Menten equation (Table 1).[17] The relative values
Table 1: Steady-state kinetic parameters for PICS module 6+TE.
Substrate
Product
k
cat [minÀ1
]
KM [mm]
kcat/KM [mÀ1 sÀ1
]
5
1
17.9Æ1.0
3.3Æ0.3
3.2Æ0.7[a]
5.2Æ1.5
91.5Æ10.6
223.3Æ56.7[a]
14.1Æ4.7
43.0Æ5.2[a]
4.4Æ0.7
2
6
7
8
9
6.9Æ0.7
1.2Æ0.1
6.8Æ1.2
2.1Æ0.5
2.1Æ0.4
2.6Æ0.9
64.7Æ14.3
9.5Æ2.0
43.6Æ17.3
[a] Formation of 1 in the absence of methylmalonyl CoA.
Scheme 3. Conversion of hexaketide seco-SNAC thioester 5 into 1 and
2 by PICS module 6+TE.
of kcat showed the expected 4:1 preference for the formation
of the macrolactonization product 1 over chain extension and
cyclization to 2. The observed Km values were not corrected
for the presence of the corresponding hemiacetal form of 5, as
the concentration of this species in the reaction buffer was not
determined and all of the substrate was consumed in the
preparative-scale reaction. Interestingly, incubation of 5 with
PICS module 6 + TE under identical conditions but in the
absence of methylmalonyl CoA gave exclusively 1, but with
approximate 2.2-fold increases in both kcat and kcat/Km values
relative to the cyclization to 1 in the presence of the
methylmalonyl CoA cosubstrate.
PICS module 6 + TE is also shown to convert 6 and methyl-
malonyl CoA into a mixture of the direct hydrolysis product 7
and the isomeric chain-extended heptaketide lactones 8 and 9
(Scheme 4).
The requisite substrates 5 and 6 were each prepared by
degradation and semisynthesis from 1, as previously de-
scribed.[13] The seco-SNAC thioester 5 is in equilibrium with
approximately 50–60% of the corresponding 3,7-hemiacetal
We previously reported that incubation of 6 (7S/7R 3:1)
with the recombinant PICS TE alone results exclusively in
hydrolysis to the corresponding diastereomeric mixture of
seco acids 7.[13,14] By contrast, preparative-scale incubation of
6 (2 mm) and methylmalonyl CoA (1 mm) with PICS
module 6 + TE (10 mm) gave 7, 8, and 9 in a 6:3:4 mixture
that could be readily separated by flash chromatography
(Scheme 4). The NMR spectroscopic and chromatographic
properties of 7 were identical to those previously deter-
mined.[13] The structure and stereochemistry of 8 and 9 were
determined by 1H and 13C NMR spectroscopy, including
HMQC and NOESY analysis.[17] Both compounds exhibited
HR-ESI mass spectra consistent with the molecular compo-
sition C20H34O5 expected for the isomeric lactones. The
Scheme 4. Hydrolysis and lactonization of 7-dihydro-seco-SNAC-thio-
ester 6 by PICS module 6+TE.
7558
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7557 –7560

Angewandte
Chemie
1H NMR spectrum of 8 closely resembled that of 2, with
several key differences. Thus, the two olefinic protons
exhibited the expected upfield shift and change in multiplicity
relative to the enone double bond of 2: H10 (d = 5.34 ppm,
dd, J = 15.4, 7.8 Hz) coupled to H11 (d = 5.42 ppm, dd, J =
15.4, 7.7 Hz), with similar changes in the 13C NMR spectrum
(C10 d = 134.1; C11 133.6 ppm). The presence of only a single
resonance for allylic H9 in both the 1H NMR (d = 3.47 ppm, t,
(2R,3S)-4
(kcat = 0.057 Æ 0.003 minÀ1
and
kcat/Km =
0.027mÀ1 sÀ1); some of the difference in the kcat/Km values
rises from an approximate tenfold decrease in the Km value
relative to (2R,3S)-4 (35 Æ 4 mm).[9,11] It would be expected
that the natural hexaketide intermediate, which is delivered
to PICS module 6 as the corresponding acyl thioester bound
to the ACPof PICS module 5, would have an even lower
native Km value.[18] Although both the KS6 or ACP6 domains
have been reported to be essential for in vivo macrolactoni-
zation of the hexaketide produced by PICS module 5,[8] it is
not clear whether the observed dominant macrolactonization
of 5 by PICS module 6 + TE involves mandatory transfer to
the TE domain by way of transient covalent attachment to
either the KS6 or ACP6 domains. Irreversible formation of
the acyl-KS species has been shown to be the sole determi-
nant of the kcat/Km value for the processing of substrates by
PKS modules.[19] The fact that the kcat/Km values measured for
lactonization and for chain-elongation of 5 are different may
suggest that these two reactions do not involve partitioning of
a common hexaketide KS intermediate. Indeed, recombinant
PICS TE alone catalyzes the efficient conversion of 5 into
1.[13,14] Interestingly, the observed steady-state kinetic param-
eters for the conversion of 5 into 1 by PICS module 6 + TE in
the absence of methylmalonyl CoA are effectively the same
as those that we previously determined for the macrolacto-
nization of 5 by recombinant PICS TE (kcat = 54.4 Æ 5.8 minÀ1,
Km = 4.1 Æ 0.8 mm, kcat/Km = 221 Æ 50mÀ1 sÀ1).[13] The kcat value
for the formation of 1 by PICS module 6 + TE in the absence
of methylmalonyl CoA is larger than the combined kcat value
for both the lactonization and chain elongation by the PICS
module 6 + TE in the presence of methylmalonyl CoA, thus
suggesting that the presence of a methylmalonyl thioester
attached to the ACPof module 6 may inhibit access of the
hexaketide substrate to the active site of the downstream TE
domain. Alternatively, it may reflect slower cyclization of the
heptaketide substrate than of the hexaketide substrate by the
TE domain.
1
J = 8.5 Hz—correlated to both H10 and H8 in the H COSY
spectrum) and 13C NMR spectra (d = 78.8 ppm) is consistent
with the formation of a single diastereomer of 8. The
configuration of 8 was readily assigned as the 9S isomer, as
based on the NOESY spectrum, which showed crosspeaks
between H9 and the H19 methyl protons (d = 1.03 ppm) as
well as between H9 and both the olefinic H11 proton (d =
5.42 ppm) and one of the H7 methylene protons (d =
0.74 ppm), but not with either H8 or H10 (Figure 1). Com-
Figure 1. Key NOESY correlations for (9S)-9-dihydronarbonolide (8).
(Energy-minimized conformation calculated by using the MM2 module
of Chem3D.)
The chain elongation/lactonization of 6 into 8 and 9 by
PICS module 6 + TE also takes place at rates comparable to
those observed for the processing of 5, the analogue of the
natural hexaketide substrate. Interestingly, although a mix-
ture of isomeric lactones is observed, chain elongation
appears to be diastereoselective. The formation of the two
lactones from 6 is in contrast to the exclusive hydrolysis of the
same substrate by recombinant PICS TE alone,[13,14] thus
indicating that both the PICS KS6 domain and the TE domain
can process reduced hexaketide and heptaketide subtrates,
respectively.
In summary, we have shown that recombinant PICS
module 6 + TE is capable of mediating both chain elongation
and lactonization as well as competing lactonization or
hydrolysis of both the natural hexaketide substrate 5 and
the reduced 7-dihydro analogue 6 and have established the
full set of steady-state kinetic parameters for both sets of
reactions. The observed 4:1 ratio of lactonization to chain
elongation for the processing of 5 may represent the intrinsic
ratio of these two processes that leads to the characteristic
formation of 12- and 14-membered-ring macrolides, thereby
definitively excluding the necessity for N-terminal truncation
pound 9 was assigned the isomeric d-lactone structure based
on comparison with the spectra of 8. Thus, the 13C NMR
resonance for the lactonic C5 (d = 82.4 ppm) appeared down-
field of the corresponding signal for C5 in 8 (d = 71.7 ppm),
with the expected differences in the attached H5 protons as
well (9: d = 4.32 ppm, dd, J = 9.8, 2.0 Hz; 8: d = 4.02 ppm, dd,
J = 5.4, 2.4 Hz). The signals for C13 (9: d = 76.8; 8: 80.1 ppm)
and H13 (9: d = 3.44 ppm, td, J = 8.4, 3.5 Hz; 8: d = 4.77 ppm,
td, J = 10.3, 3.1 Hz) also showed the expected complementary
changes in chemical shift. Although 9 is most likely a single
diastereomer (H9 d = 3.44 ppm, t, J = 5.1 Hz; C9 d =
75.9 ppm), the configuration at C9 was not assigned. The
steady-state kinetic parameters for the formation of 8 and 9
(lactonization) as well as 7 (hydrolysis) were also determined
(Table 1).
The observed kcat and kcat/Km values for chain-elongation/
cyclization of 5 to 2 (Table 1) are nearly two orders of
magnitude greater than the corresponding steady-state
parameters previously determined for the diketide analogue
Angew. Chem. Int. Ed. 2005, 44, 7557 –7560
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7559

Communications
of PICS module 6 for formation of the hexaketide macro-
lactonization product 10-deoxymethynolide (1), as has been
suggested at one time.[12]
Received: June 26, 2005
Published online: October 25, 2005
Keywords: antibiotics · biosynthesis · lyases · polyketides ·
.
synthases
[1] D. OꢀHagan, The Polyketide Metabolites, E. Norwood, New
York, 1991.
[2] S. Omura, H. Takeshima, A. Nakagawa, J. Miyazawa, J. Antibiot.
1976, 29, 316 – 317.
[3] T. Hori, I. Maezawa, N. Nagahama, M. Suzuki, J. Chem. Soc.
Chem. Commun. 1971, 304 – 305.
[4] Y. Xue, L. Zhao, H.-W. Liu, D. H. Sherman, Proc. Natl. Acad.
Sci. USA 1998, 95, 12111 – 12116.
[5] Y. Xue, D. Wilson, D. H. Sherman, Gene 2000, 245, 203 – 211.
[6] L. Tang, H. Fu, M. C. Betlach, R. McDaniel, Chem. Biol. 1999, 6,
553 – 558.
[7] R. H. Lambalot, D. E. Cane, J. Antibiot. 1992, 45, 1981 – 1982.
[8] B. J. Beck, Y. J. Yoon, K. A. Reynolds, D. H. Sherman, Chem.
Biol. 2002, 9, 575 – 583.
[9] B. J. Beck, C. C. Aldrich, R. A. Fecik, K. A. Reynolds, D. H.
Sherman, J. Am. Chem. Soc. 2003, 125, 12551 – 12557.
[10] B. J. Beck, C. C. Aldrich, R. A. Fecik, K. A. Reynolds, D. H.
Sherman, J. Am. Chem. Soc. 2003, 125, 4682 – 4683.
[11] Y. Yin, H. Lu, C. Khosla, D. E. Cane, J. Am. Chem. Soc. 2003,
125, 5671 – 5676.
[12] Y. Xue, D. H. Sherman, Nature 2000, 403, 571 – 575.
[13] W. He, J. Wu, C. Khosla, D. E. Cane, Bioorg. Med. Chem. Lett.,
DOI:10.1016/j.bmcl.2005.09.077.
[14] C. C. Aldrich, L. Venkatraman, D. H. Sherman, R. A. Fecik, J.
Am. Chem. Soc. 2005, 127, 8910 – 8911.
[15] C. C. Aldrich, B. J. Beck, R. A. Fecik, D. H. Sherman, J. Am.
Chem. Soc. 2005, 127, 8441 – 8452.
[16] Incubation for more than 1 h with > 15 mm PICS module 6 + TE
resulted in competing hydrolysis of the accumulated narbonolide
(2) apparently to the corresponding seco acid.
[17] See the Supporting Information for details of preparative and
kinetic experiments, as well as complete spectroscopic data for
new compounds.
[18] N. Wu, S. Y. Tsuji, D. E. Cane, C. Khosla, J. Am. Chem. Soc.
2001, 123, 6465 – 6474.
[19] J. Wu, K. Kinoshita, C. Khosla, D. E. Cane, Biochemistry 2004,
43, 16301 – 16310.
[20] H. Lu, S. C. Tsai, C. Khosla, D. E. Cane, Biochemistry 2002, 41,
12590 – 12597.
7560 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7557 –7560