Biochemical investigation of pikromycin biosynthesis employing native penta- and hexaketide chain elongation intermediates

Article abstract of DOI:10.1021/ja042592h

The unique ability of the pikromycin (Pik) polyketide synthase to generate 12- and 14-membered ring macrolactones presents an opportunity to explore the fundamental processes underlying polyketide synthesis, specifically the mechanistic details of chain e

Full text of DOI:10.1021/ja042592h

Published on Web 05/19/2005
Biochemical Investigation of Pikromycin Biosynthesis
Employing Native Penta- and Hexaketide Chain Elongation
Intermediates
Courtney C. Aldrich,†,‡ Brian J. Beck, Robert A. Fecik, and
†,§
#
,
†,|,⊥
David H. Sherman*
Contribution from the Life Sciences Institute, Department of Medicinal Chemistry, Department of
Chemistry, and Department of Microbiology & Immunology, UniVersity of Michigan, Ann Arbor,
Michigan 48109-2216, and Department of Medicinal Chemistry, UniVersity of Minnesota,
Minneapolis, Minnesota 55455
Received December 9, 2004; E-mail: davidhs@umich.edu
Abstract: The unique ability of the pikromycin (Pik) polyketide synthase to generate 12- and 14-membered
ring macrolactones presents an opportunity to explore the fundamental processes underlying polyketide
synthesis, specifically the mechanistic details of chain extension, keto group processing, acyl chain release,
and macrocyclization. We have synthesized the natural pentaketide and hexaketide chain elongation
intermediates as N-acetyl cysteamine (NAC) thioesters and have used them as substrates for in vitro
conversions with engineered PikAIII+TE and in combination with native PikAIII (module 5) and PikAIV
(module 6) multifunctional proteins. This investigation demonstrates directly the remarkable ability of these
monomodules to catalyze one or two chain extension reactions, keto group processing steps, acyl-ACP
release, and cyclization to generate 10-deoxymethynolide and narbonolide. The results reveal the enormous
preference of Pik monomodules for their natural polyketide substrates and provide an important comparative
analysis with previous studies using unnatural diketide NAC thioester substrates.
Introduction
organized into modules each of which catalyzes one cycle of
elongation. A prototypical module minimally consists of three
The rapid rise in resistance to MLS (macrolide-lincomycin-
domains: an acyl carrier protein (ACP), an acyltransferase (AT)
that loads the ACP with a malonyl or methylmalonyl extender
unit, and a â-ketosynthase (KS) domain that catalyzes the
decarboxylative condensation between extender units (linked
to the ACP domain) and the growing polyketide chain (linked
via the KS active site Cys residue) to afford a â-ketoacyl-ACP.
Each elongation cycle results in a â-keto-group that can undergo
additional reductive processing steps. Catalytic domains that
perform these reactions include a ketoreductase (KR), dehy-
dratase (DH), and enoyl reductase (ER). The absence of â-keto
processing domains results in retention of a carbonyl group, a
KR alone gives rise to â-hydroxyl functionality, and a KR and
DH generate an alkene, while a KR, DH, and ER combination
leads to complete reduction to an alkane. Processing of the
growing polyketide through the PKS enzyme complex leads to
the fully elaborated chain elongation product that is released
either as a carboxylic acid or as macrolactone by the thioesterase
(TE) domain located on the C-terminus of the last module. The
collinearity between the genetic architecture and biochemical
organization of polyketide biosynthesis, coupled with increas-
ingly powerful tools to manipulate modular PKSs presents op-
streptomycin) antibiotics underscores an urgent need to develop
alternatives to current clinical antimicrobial agents. Semisyn-
1
thetic modifications of erythromycin have provided the ketolide
class of antibiotics, such as telithromycin (Ketek) and cethro-
mycin (ABT-773), which are characterized in part by a 3-keto
2
,3
group with activity against MLS-resistant pathogens. Con-
tinued advances toward effective redesign of polyketide bio-
synthesis suggest a growing role for metabolic engineering in
providing new ketolide templates for next-generation macrolide
antibiotics capable of overcoming antibiotic resistant microbial
pathogens.
Polyketide biosynthesis is catalyzed by modular multifunc-
tional proteins termed polyketide synthases (PKSs) that catalyze
the repetitive condensation of malonyl- or methylmalonyl
coenzyme-A thioester monomers. Bacterial type I PKSs are
†
Department of Medicinal Chemistry, University of Michigan.
Current Address: Center for Drug Design, University of Minnesota,
‡
Minneapolis, MN.
§
Current Address: American Type Culture Collection, Manassas, VA.
Department of Chemistry, University of Michigan.
Department of Microbiology & Immunology, University of Michigan.
Department of Medicinal Chemistry, University of Minnesota.
|
⊥
#
4,5
portunities to create novel structures based on a ketolide core.
(
(
(
1) Youngman, P.; Tepper, R.; Moore, J. Curr. Clin. Top. Infect. Dis. 2001,
2
1, 366-390.
Currently, our knowledge of type I PKSs is largely based on
the analysis of 6-deoxyerythronolide B polyketide synthase
2) Zhanel, G. G.; Wierzbowski, A. K.; Hisanaga, P.; Hoban, D. J. Curr. Infect.
Dis. Rep. 2004, 6, 191-199.
3) Reinert, R. R. J. Antimicrob. Chemother. 2004, 53, 918-927.
6,7
(DEBS), which appears to be unusually tolerant and can accept
10.1021/ja042592h CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 8441-8452
9
8441

A R T I C L E S
Aldrich et al.
a variety of nonnatural structural and stereochemical variations
to the polyketide chain. This tolerance has been partly inferred
lactonization and hydrolysis rates for the isolated TE domain
of epothilone PKS, a microbial secondary metabolite with
derivatives in clinical evaluation, have been examined using a
full-length NAC thioester substrate generated by direct ring
8
through in vivo experiments where N-acetylcysteamine (NAC)
thioester activated di- and triketide substrates supplemented in
the fermentation medium are incorporated onto early stage PKS
modules. These precursor directed biosynthetic approaches have
produced new 14- and 16-membered ring macrolides and pro-
18
opening of the natural product to the corresponding seco-acid.
19
A previously developed synthetic strategy provided the
advanced acyl chain (as the corresponding NAC thioester),
presumed to be a natural substrate in the epothilone pathway.
The goal of that in vivo study was to test the ability of the final
three modules of the epothilone PKS to extend and lactonize
vided latent functionality at the C-13 position of the macrolac-
tone ring, enabling further synthetic modifications.4
,9-13
More-
over, in vitro experiments using di- and triketide NAC thioester
substrates with purified modules from the DEBS and Pik PKS
systems have provided comparative information on the inherent
substrate specificity (kcat/KM) and channeling of these PKS
enzymes. Biochemical analyses of engineered monomodules
from DEBS3 (modules 5 and 6) revealed that the substrate
specificities were partly dependent on the stereochemistry and
functionality at the R and â position of the polyketide chain
19
the molecule in Escherichia coli cells. The need to employ
natural substrate advanced chain elongation intermediates is
critical for advancing our understanding of these systems. Thus,
combining versatile synthetic strategies toward studies of
isolated purified modular PKSs promises to unveil the complex
properties of these multifunctional enzymes. This concept has
motivated the current study, and its clear outcome demonstrates
the power of this approach for dissecting modular PKS function
and specificity and for expanding the potential for chemoen-
zymatic-based drug discovery approaches.
substrate and revealed that they have increased flexibility relative
to analogous Pik monomodular PKS enzymes.1
4-17
Specifically,
the calculated kcat values for extension and lactonization of dike-
tide substrates by the terminal Pik PKS monomodules were
approximately 3 orders of magnitude lower than the in vivo
rates of antibiotic production consistent with a strict substrate
specificity for these enzymes.17
The use of diketide and triketide NAC thioester substrates
has proven invaluable for investigating fundamental mechanisms
of polyketide biosynthesis. However, their relevance in deter-
mining the inherent selectivity of a non-native cognate module
remains unclear, as these substrates do not adequately represent
the natural chain elongation intermediates processed by non-
native modules. It is apparent that analysis of substrate specific-
ity of PKS modules with natural chain elongation substrates
would provide a more accurate measure of the catalytic and
kinetic capacity of these systems and accelerate acquisition of
knowledge required to design modular PKS systems with novel
specificities and processing characteristics.
Streptomyces Venezuelae ATCC 15439 produces the 12- and
14-membered ring macrolides, methymycin (3), and pikromycin
(4), respectively, through the activity of the pikromycin PKS
(Pik) (Figure 1). Initiation of polyketide biosynthesis by PikAI
and successive elongation through PikAIII provides an ACP5-
bound hexaketide thioester that can undergo two possible fates.
6
Transfer to PikAIV KS and subsequent elongation provides a
heptaketide-bound S-ACP intermediate that is cyclized by the
6
C-terminal thioesterase domain to provide the 14-membered ring
macrolactone narbonolide (2). Alternatively, the hexaketide can
be cyclized by the C-terminal thioesterase of PikAIV (by
skipping of the chain from ACP5 to ACP6 in the absence of a
20
final chain extension step) to provide the 12-membered ring
macrocycle 10-deoxymethynolide (1). The Pik PKS provides
an ideal starting point for chemoenzymatic synthesis of mac-
rocyclic lactones, since narbonolide is analogous to 6-deoxy-
erythronolide B but possesses the critical 3-keto group found
in ketolides, contains an alkene functionality at C10-C11
enabling subsequent chemical modifications, and is not limited
by protection of the C-6 hydroxyl group characteristic of the
Thus, generating fully elaborated polyketide chain elongation
intermediates is a fundamental requirement for elucidating the
complex biochemical processes involved in polyketide biosyn-
thesis. In some cases, previously developed synthetic method-
ologies can be applied directly or with little modification to
facilitate biosynthetic investigations. For example, in vitro
2
1
erythromycin-derived semisynthetic ketolides.
(
4) Carreras, C.; Frykman, S.; Ou, S.; Cadapan, L.; Zavala, S.; Woo, E.; Leaf,
T.; Carney, J.; Burlingame, M.; Patel, S.; Ashley, G.; Licari, P. J.
Biotechnol. 2002, 92, 217-228.
Herein we report the synthesis of the native pentaketide and
hexaketide chain elongation intermediates of pikromycin bio-
synthesis, including a kinetic analysis involving the final two
modules (Pik module 5 and Pik module 6) of the Pik PKS. This
work represents the first biochemical investigation of modular
PKS substrate specificity employing advanced polyketide chain
intermediates. This versatile synthetic approach to advanced
polyketide intermediates and their analysis as substrates for
modular PKSs represents a significant step toward exploring
mechanistically the basis for sequential chain extension, pro-
cessing, and macrolactone ring formation by the Pik PKS.
(
5) Yoon, Y. J.; Beck, B. J.; Kim, B. S.; Kang, H. Y.; Reynolds, K. A.;
Sherman, D. H. Chem. Biol. 2002, 9, 203-214.
(
6) Donadio, S.; Katz, L. Gene 1992, 111, 51-60.
(
7) Donadio, S.; Staver, M. J.; McAlpine, J. B.; Swanson, S. J.; Katz, L. Science
1
991, 252, 675-679.
(
(
8) Staunton, J.; Weissman, K. J. Nat. Prod. Rep. 2001, 18, 380-416.
9) Kosan Biosciences, Inc. Macrolide Antiinfective Agents. U.S. Patent 6,-
3
95,710, May 28, 2002.
(
(
(
(
(
(
(
(
10) Kosan Biosciences, Inc. Macrolide Antiinfective Agents. U.S. Patent 6,-
5
93,302, July 15, 2003.
11) Kosan Biosciences, Inc. and Ortho-McNeil Pharmaceutical, Inc. Ketolide
Antibacterials. U.S. Patent 6,458,771, Oct 1, 2002.
12) Kinoshita, K.; Williard, P. G.; Khosla, C.; Cane, D. E. J. Am. Chem. Soc.
2
001, 123, 2495-2502.
13) Frykman, S.; Leaf, T.; Carreras, C.; Licari, P. Biotechnol. Bioeng. 2001,
6, 303-310.
14) Wu, N.; Kudo, F.; Cane, D. E.; Khosla, C. J. Am. Chem. Soc. 2000, 122,
847-4852.
15) Yin, Y.; Lu, H.; Khosla, C.; Cane, D. E. J. Am. Chem. Soc. 2003, 125,
7
(18) Boddy, C. N.; Schneider, T. L.; Hotta, K.; Walsh, C. T.; Khosla, C. J. Am.
Chem. Soc. 2003, 125, 3428-3429.
4
(19) Boddy, C. N.; Hotta, K.; Tse, M. L.; Watts, R. E.; Khosla, C. J. Am. Chem.
Soc. 2004, 126, 7436-7437.
5
671-5676.
16) Watanabe, K.; Wang, C. C.; Boddy, C. N.; Cane, D. E.; Khosla, C. J.
Biol. Chem. 2003, 278, 42020-42026.
17) Beck, B. J.; Aldrich, C. C.; Fecik, R. A.; Reynolds, K. A.; Sherman, D. H.
J. Am. Chem. Soc. 2003, 125, 12551-12557.
(20) Beck, B. J.; Yoon, Y. J.; Reynolds, K. A.; Sherman, D. H. Chem. Biol.
2002, 9, 575-583.
(21) Omura, S., Ed. Macrolide Antibiotics Chemistry, Biology, and Practice,
2nd ed.; Academic Press: London, 2002.
8442 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Chemoenzymatic Investigation of Pikromycin Biosynthesis
A R T I C L E S
Figure 1. Modular organization of the pikromycin PKS. The extension of the polyketide chain through PikAI-II produces a pentaketide that is elongated
by PikAIII (module 5) to produce a hexaketide intermediate. PikAIV (module 6) lactonizes this intermediate to 10-deoxymethynolide (1) or catalyzes its
extension and subsequent cyclization to narbonolide (2). Both aglycones undergo further tailoring to afford methymycin (3) and pikromycin (4), respectively.
Scheme 1
Scheme 2
Results
Synthesis of Pentaketide and Hexaketide NAC Thioester
Substrates. The natural pentaketide and hexaketide substrates
for Pik modules 5 and 6 were obtained by total synthesis as
well as by degradation chemistry of 10-deoxymethynolide (1).
While the pentaketide may be obtained through degradation of
1
0-deoxymethynolide (1), a more versatile total synthesis
approach was developed on the basis of the potential to generate
analogues for future mechanistic studies. Retrosynthetically,
dissection of the desired pentaketide 5 across the olefin is via
a Horner-Wadsworth-Emmons olefination between â-keto
ester 6 and known aldehyde 722 (Scheme 1). The requisite
phosphonate building block was prepared by desymmetrization
of meso-diester 8 employing R-chymotrypsin to afford mo-
noester 9 in 94% ee (Scheme 2).23 The stereoselective hydrolysis
of the pro-S methyl ester was designed to allow protection of
the carboxylic acid during the subsequent phosphonate forma-
tion. Thus, addition of 3 equivalents of the lithium anion of
dimethyl methylphosphonate to 9 afforded â-ketophosphonate
6, wherein the first equivalent deprotonated the carboxy function
that served to protect it from over addition while the second
equivalent added to the methyl ester to provide the correspond-
(
22) Pilli, R. A.; de Andrade, C. K. Z.; Souto, C. R. O.; de Meijere, A. J. Org.
(23) Mohr, P.; Waespe-Sarcevic, N.; Tamm, C.; Gawronska, K.; Gawronski, J.
Chem. 1998, 63, 7811-7819.
K. HelV. Chim. Acta 1983, 66, 2501-2511.
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8443

A R T I C L E S
Scheme 3
Aldrich et al.
Scheme 4
ing â-ketophosphonate functionality. Next, Horner-Wadsworth-
Emmons olefination of 6 with aldehyde 7 was found to optimally
occur utilizing barium hydroxide as base to provide enone 10
2
4
(E:Z > 20:1). The acid was converted to the NAC thioester
1
1 employing EDC and catalytic DMAP without any observed
2
5
competitive Michael addition of the thiol to the enone.
Deprotection of the TBS ether with hydrofluoric acid in
acetonitrile provided the pentaketide NAC thioester 5 in only
five steps from 8. The convergent and efficient chemoenzymatic
synthesis of the pentaketide 5 also enables direct access to
additional analogue structures.
1
0-Deoxymethynolide (1) represents an excellent starting
material for degradation work, since direct hydrolysis yields
the natural hexaketide substrate. To obtain sufficient amounts
of 10-deoxymethynolide required for this approach, an engi-
neered mutant strain of S. Venezuelae that produces 10-
deoxymethynolide (25 mg/L) as the exclusive product was
used.2 The titer was further improved by incubating with XAD-
To confirm the stereochemistry of the C-2 stereocenter, the
mixture of epimeric hexaketide acids 13 was reduced to tetraols
1
9, which were selectively protected as p-methoxybenzylidene
acetals 20 and 21 (Scheme 4). The stereochemistry of the major
6
isomer was correlated with a related benzylidene acetal frag-
1
1
6 resin (500 g, 55 L incubation), which served to absorb the
0-deoxymethynolide and led to production levels (∼4-fold
29
ment. Conversion of the C-7 epimeric mixture of 13 to the
NAC thioesters 14 was achieved using standard conditions, and
the diastereomers were separated by preparative reverse-phase
HPLC (Scheme 3). Chemoselective oxidation of the allylic
improvement) of approximately 100 mg/L. Unfortunately, direct
hydrolysis of the 10-deoxymethynolide (LiOH, 50 °C, 5 h, 100%
consumption of 10-deoxymethynolide) was plagued by a number
of side reactions, which precluded isolation of the desired seco-
acid (not shown). Protection of the ketone prevented the
undesired epimerization of the C-6 stereocenter and hemiket-
alization with the C-3 alcohol. This was accomplished by Luche
reduction of 10-deoxymethynolide with NaBH4 and CeCl3 to
30
alcohol of 14 was accomplished with MnO2 to afford 15. The
native hexaketide 15 was in equilibrium with a mixture of both
R and â anomeric hemiacetals 16 (only the â-anomer is shown)
and was prepared directly before use due to its propensity to
undergo dehydration, leading to dihydropyran derivative 17.
Additionally, an authentic standard of 3-oxo-10-deoxymethyno-
lide 18 was required, since this is the putative product resulting
27
furnish allylic alcohol 12 (Scheme 3). Hydrolysis of 12
required significant forcing conditions28 (as compared to 10-
deoxymethynolide) to afford 13 as a 4:1 mixture of C-2 epimers.
(
28) Minimization of both 10-deoxymethynolide (1) and 12 at the AM1 level
does not reveal any structural change that could account for this great
difference in reactivity. The greater reactivity of 10-deoxymethynolide
versus 12 may be due to the enone function, which may undergo reversible
Michael addition of methoxide, providing a more conformationally flexible
(
24) Paterson, I.; Yeung, K.-P.; Smaill, J. B. Synlett 1983, 774-776.
25) Gilbert, I. H.; Ginty, M.; O’Neill, J. A.; Simpson, T. J.; Staunton, J.; Willis,
C. L. Bioorg. Med. Chem. Lett. 1995, 5, 1587-1590.
(
2
3
structure due to the reduced torsional strain as a result of the sp to sp
(26) Chen, S.; Xue, Y.; Sherman, D. H.; Reynolds, K. A. Chem. Biol. 2000, 7,
change in hybridization of the C-8 and C-9 carbons. However, the putative
1
9
07-918.
Michael adduct is not observable within the detection limit (<1%) of H
(
27) The resultant C-7 alcohol was assigned the 7S stereochemistry on the basis
of the most accessible face of the macrocycle. 10-Deoxymethynolide was
modeled using a Monte Carlo search in Macromodel, and the minimum
energy conformation was further optimized at the AM1 level with Gaussian
3 3
NMR when taken in CD ONa/CD OD.
(29) Evans, D. A.; Polniaszek, R. P.; DeVries, K. M.; Guinn, D. E.; Mathre, D.
J. J. Am. Chem. Soc. 1991, 113, 7613-7630.
(30) Aldrich, C. C.; Lakshmanan, V.; Sherman, D. H.; Fecik, R. A. J. Am. Chem.
Soc. 2005, in press.
9
8.
8444 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Chemoenzymatic Investigation of Pikromycin Biosynthesis
A R T I C L E S
Figure 2. Radio-TLC of in vitro reaction products produced using the
pentaketide and hexaketide NAC thioester substrates with recombinant
1
4
pikromycin modules. TLC autoradiography of C-lableled 10-deoxym-
ethynolide (1), narbonolide (2), and 3-oxo-10-deoxymethynolide (18).
14
2
-[ C]Methylmalonyl-CoA was used as the radiolabeled substrate; con-
centrations of all substrates and Pik monomodule concentrations are
described in the Experimental Section. Developed with 5% MeOH/CHCl3.
Lanes 1-4 were acquired together, while the data for lane 5 was obtained
separately. Lane 1, PikAIII+TE incubated with pentaketide 5 to afford 10-
deoxymethynolide (1); lane 2, PikAIII-TE incubated with 5 in the absence
of NADPH to afford 3-oxo-10-deoxymethynolide (18); lane 3, PikAIII
incubated with 5, no products were observed; lane 4, PikAIII and PikAIV
incubated with 5 to afford both narbonolide (2) as the upper spot and 10-
deoxymethynolide (1) as the lower spot; lane 5, PikAIV incubated with
hexaketide 15 to afford narbonolide (2).
from incubation of the native pentaketide with PikAIV or of
the pentaketide with PikAIII-TE in the absence of NADPH. This
authentic standard was prepared by Swern oxidation of 10-
deoxymethynolide (1) (Scheme 3).
Assay and Measurement of Kinetic Parameters. PikAIII,
PikAIV, and engineered PikAIII-TE holoenzymes were purified
as previously described1 and individually reacted with 2-[ C]-
methylmalonyl-CoA and each of the pentaketide 5 and hexa-
ketide 15 NAC thioesters. The resulting products were identified
using synthetic standards and quantified by radio-TLC. The
initial rates, V, at a given [S] were determined by single time
point stopped-time incubations at 20 minutes. All reactions were
run in triplicate employing a fresh enzyme preparation and
identified by comigration with authentic standards and HPLC
or preparative-TLC in conjunction with ESI(+) analysis to
rigorously confirm these assignments. Apparent steady-state
kinetic values were determined for enzyme-substrate pairings
that yielded a detectable product by fitting the normalized V vs
5,17
14
Figure 3. Normalized plots used to determine the steady-state kinetic
parameters of Pik monomodules with native penta- and hexaketide
substrates. (A) PikAIII-TE + pentaketide 5 as substrate to afford 10-
[S] plots to the Michaelis-Menten equation.
We first investigated the engineered PikAIII-TE fusion protein
-
1
-1
to accept and process pentaketide 5 to produce 10-deoxym-
ethynolide (1). In vitro studies of pentaketide 5 with PikAIII-
TE provided 10-deoxymethynolide as the sole product (Figure
deoxymethynolide (1), kcat/KM ) 0.55 ( 0.029 mM min ; (B) PikAIV
hexaketide 15 as substrate to afford narbonolide (2), kcat/KM ) 4.4 (
+
-
1
-1
0
.24 mM min ; (C) PikAIII‚PikAIV + pentaketide 5 to afford 10-
-1
deoxymethynolide (1), kcat(10-dml) ) 3.0 ( 0.33 min and KM(10-dml) )
-1
-1
-1
2, lane 1). Initial velocities as a function of substrate concentra-
0.25 ( 0.073 mM min , and narbonolide (2), k
cat(nbl)
) 3.3 ( 0.40 min
-
1
-1
and KM(nbl) ) 0.41 ( 0.094 mM min
.
tion were determined under saturating concentrations of meth-
ylmalonyl-CoA (800 µM) and NADPH (1 mM) and are plotted
in Figure 3A. The limited solubility of pentaketide substrate
pentaketide substrate 5 prevented determination of individual
kcat and KM parameters, but the KM is clearly greater than 1
mM. Likely, the increase in the specificity constant is due
exclusively to an enhancement of the catalytic efficiency (kcat),
since the KM (>1 mM) appears to be comparable to that of the
corresponding (2S,3R)-diketide substrate (KM ) 7.3 mM).1
When NADPH was excluded from the reaction mixture,
PikAIII-TE readily converted the pentaketide substrate 5 to
3-oxo-10-deoxymethynolide (18), indicating some flexibility in
[5] < 1 mM in aqueous buffer did not allow saturation to be
achieved; however, fitting of the data by linear regression
analysis to the reduced Michaelis-Menten equation, where [S]
<
KM, allowed determination of the specificity constant. The
5,17
specificity constant kcat/KM for this reaction is 0.55 ( 0.029
mM min , which is approximately 150-fold greater than the
-
1
-1
value obtained when utilizing the simple (2S,3R)-diketide NAC
thioester substrate analogue.1
5,17
The limited solubility of
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8445

A R T I C L E S
Aldrich et al.
Table 1. Steady-State Kinetic Parameters for the Extension and Lactonization of Penta- and Hexaketide Substrates by PikAIII, PikAIII-TE,
and PikAIV
pentaketide (5)
hexaketide (15)
k
cat/K
1
m
k
cat/K
1
m
module(s)
k
cat (min-1)
K
m
(mM)
(mM- min-1)
k
cat (min-1)
K
m
(mM)
(mM- min-1)
a
PikAIII
PikAIII-TE
PikAIV
PikAIII
+
-
nd
-
-
>1
-
-
nd
-
nd
-
nd
-
b
c
0.55 ( 0.029c
d
4.4 ( 0.24d
nd
-
nd
>1
3.0 ( 0.33c
3.3 ( 0.40d
0.25 ( 0.07c
0.41 ( 0.09d
12 ( 0.31c
8.1 ( 0.26d
nd
nd
PikAIV
a
No activity detected. b Not determined. c This value represents the steady-state kinetic parameter for 10-deoxymethynolide (1) formation. d This value
represents the steady-state kinetic parameter for narbonolide (2) formation.
the TE domain (Figure 2, lane 2). No 3-oxo-10-deoxymethyno-
lide (18) was observed in reactions with NADPH, indicating
rapid reduction of the â-keto group before TE catalyzed
lactonization. This is in contrast to earlier experiments with
PikAIII and PikAIV using diketide substrates, where an apparent
competition between ketoreduction by the KR5 domain of
PikAIII and lactonization by PikAIV resulted in a mixture of
triketide lactones with different oxidation states.17 Additionally,
we incubated hexaketide 15 with PikAIII-TE in both the
presence and absence of NADPH, but no product formation was
observed (see Supporting Information for radio-TLC). As
expected, incubation of the pentaketide 5 with native PikAIII
protein (e.g. lacking a TE domain) did not produce detectable
levels of 10-deoxymethynolide (1) (Figure 2, lane 3). In contrast,
incubation of simple diketides with the native PikAIII afforded
triketide lactone products due to spontaneous release. However,
in this case a hydroxyl group was appropriately located to
facilitate a slow intramolecular lactonization.17 The hexaketide-
bound S-ACP5 intermediate does not contain an analogous
hydroxyl group and thus is not released, and presumably remains
linked to the ACP5 domain.
sequential chain extension, processing, and release using in vitro
pairing of native PikAIII and PikAIV monomodules. Both 10-
deoxymethynolide (1) and narbonolide (2) were observed by
radio-TLC (Figure 2, lane 4). Thus, loading of the pentaketide
onto KS5 of PikAIII, extension, and keto group reduction
afforded an ACP bound hexaketide thioester. Intermodular
transfer to a full-length PikAIV and subsequent elongation,
release, and cyclization provided narbonolide (2), while TE-
mediated release in the absence of a second methylmalonate
extension afforded 10-deoxymethynolide (1).
A steady-state kinetic analysis of the sequential chain
elongation process employing the pentaketide 5 with both
PikAIII and PikAIV was performed. Thus, variation of the
pentaketide concentration utilizing 1 µM of PikAIII in the
presence of NADPH (1 mM), methylmalonyl-CoA (800 µM),
and PikAIV (1 µM), each experimentally determined to be at
saturating concentrations, provided the reaction rates shown in
Figure 3C that were fit by nonlinear regression analysis to the
Michaelis-Menten equation. The calculated kcat for 10-deoxym-
ethynolide (1) and narbonolide (2) synthesis were determined
-
1
to be 3.0 ( 0.33 and 3.3 ( 0.40 min respectively. Since
PikAIV produces both 10-deoxymethynolide (1) and narbono-
Next, we examined the ability of PikAIV to process the native
hexaketide substrate. Incubation of PikAIV with hexaketide 15
-
1
-1
lide (2), the composite kcat ) 6.3 min ( 0.52 min provides
a measure of the total flux through the PikAIII‚PikAIV complex.
This value is approximately 100-fold greater than the highest
value determined with diketide substrates and monomodular
1
4
and 2-[ C]methylmalonyl-CoA resulted in chain extension
to the final heptaketide, release, and cyclization to afford
narbonolide (2) as detected by radio-TLC (Figure 2, Lane 5).
Since 10-deoxymethynolide (1) does not incorporate radio-
labeled methylmalonyl-CoA, we were unable to observe it
through radio-TLC analysis and thus only report the kinetic
activity for the formation of narbonolide (2). In addition to a
low solubility value (<1 mM), the hexaketide substrate 15 is
in equilibrium with a mixture of anomeric hemiacetals 16 (only
the â-anomer shown). The apparent specificity constant for
hexaketide 15 with PikAIV as determined by analogy to the
pentaketide 5 with PikAIII-TE was found to be 4.4 ( 0.24
1
5,17
PikAIII-TE or PikAIV proteins.
for pentaketide 5 were found to be 0.25 ( 0.073 and 0.41 (
.094 mM for 10-deoxymethynolide (1) and narbonolide (2)
Similarly, the KM values
0
formation, respectively. The individual KM values are ap-
proximately 10-fold lower than the best diketide substrate with
1
5,17
Pik monomodules.
PikAIV modular pairing with the natural pentaketide substrate
using the kcat and KM values for 10-deoxymethynolide (1)
The specificity constant for the PikAIII‚
5
-
1
-1
-1
-1
mM min , which is approximately 3000-fold greater than
production is 12 ( 0.31 mM min , which exceeds by more
than 3000-fold the best reported value (kcat/KM ) 0.0038 mM
min ) obtained using a diketide analogue.
-1
the values obtained utilizing simple diketide substrates (Figure
1
5,17
-1
15,17
3
B).
pentaketide substrate 5 to afford 3-oxo-10-deoxymethynolide
18) was examined, but no product formation was observed, in
Additionally, the ability of PikAIV to accept the
A comparison
of these kinetic values for individual modules reacting with
pentaketide 5 or hexaketide 15 is shown in Table 1. Additionally,
hexaketide 15 was incubated with a mixture of PikAIII and
PikAIV to afford narbonolide as the exclusive product (see
Supporting Information for radio-TLC). The efficient processing
of the natural pentaketide 5 by Pik modules 5 and 6 in vitro
indicates that this chain elongation intermediate represents an
excellent reference point for comparative studies using modified
chain elongation substrates.
(
accordance with the apparent strict substrate specificity of the
Pik monomodules.
It is evident that the native pikromycin pathway pentaketide
and hexaketide chain elongation intermediates are processed
2-3 orders of magnitude more efficiently by both PikAIII-TE
and PikAIV relative to diketide substrates. We therefore
proceeded to examine the pentaketide 5 and its ability to undergo
8446 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Chemoenzymatic Investigation of Pikromycin Biosynthesis
A R T I C L E S
Discussion
ring structures may not be efficiently processed or may be
unstable. Although many polyketides undergo spontaneous
rearrangements, the timing of these events during polyketide
biosynthesis is not known. Recent crystallographic analysis of
a KS domain has suggested that PKS modules are passed down
a narrow protein channel. Such a channel may prevent
intramolecular reactions of the processed polyketide chain by
maintaining the elongation intermediate in an extended confor-
mation. In solution, the native hexaketide 15 investigated in
this study was found to exist in equilibrium with a mixture of
anomeric hemiacetals 16. Not only was this compound ex-
tremely labile, decomposing through dehydration to afford a
dihydropyran derivative 17, but PikAIV was found to possess
a high KM value (>1 mM) toward it, even though it is the native
substrate for this enzyme. By contrast, the KM for the pentaketide
Chiral Recognition of Diketide and Native Substrates. A
primary challenge of studying modular PKS systems has been
the availability of natural chain elongation intermediates for
investigation of substrate specificity in late-stage extension,
processing, and macrocyclization. To begin addressing the
specificity and catalytic efficiency of late-stage biosynthetic
modular PKS enzymes, short chain substrate mimics such as
diketide NAC thioesters were used to probe the specificity of
3
5
1
4,31-34
non-native monomodules of the DEBS PKS system.
Results from these model studies showed that all modules
efficiently processed the syn-diketide substrates with kcat values
-
1
on the order of 1 min , whereas anti-diketide substrates were
not accepted. Since the 6-deoxyerythronolide (DEBS) chain
elongation intermediates do not possess relative anti-stereo-
chemistry, these results supported the notion that such substrates
adequately mimicked the native advanced polyketide chain
intermediates. We have previously shown that PikAIII and
PikAIV process such diketide substrates almost 1000-fold less
efficiently than the corresponding DEBS modules, providing
the first comparative analysis of a PKS system against the DEBS
5
with PikAIII‚PikAIV pairing is 0.33 mM. The increased KM
value may instead reflect the lower concentration of the open
chain tautomer, suggesting that the hemiacetal tautomer is not
recognized. Thus, the apparent KM value for the open-chain
tautomer assuming this scenario is related as follows: KM )
KeqKM(obs), where Keq represents the equilibrium constant
between the open and closed chain tautomers of the hexaketide.
While the observed KM value of the pentaketide 5 with PikAIII‚
PikAIV pairing is certainly lower than hexaketide 15 with
PikAIV, this value still appears quite high for a native substrate.
However, it should be recognized that the present substrate relies
on simple diffusion to the KS domain, whereas under normal
in vivo conditions this substrate is channeled from the ACP5
domain of PikAIII. This channeling is kinetically much more
efficient, and a steady-state kinetic analysis using ACP-bound
diketide analogues and DEBS-TE monomodules quantified the
kinetic enhancement as being as much as 100 times faster than
1
7
PKS. Our current observations show clearly that the native
pentaketide 5 and hexaketide 15 NAC thioester substrates are
processed at 2-3 orders of magnitude more efficiently than
simple diketide analogues, suggesting that more remote struc-
tural features of the polyketide chain play an important role in
substrate recognition.
The pentaketide and hexaketide chain intermediates used in
this study serve as powerful molecular probes to investigate
the substrate recognition features employed by late stage PKS
modules; however, several potential limitations are evident. First
is their relatively poor aqueous solubility, which may be
4
7
33
problematic for even longer and more lipophilic substrates.
simple diffusion. Thus, the substrate affinity for the KS domain
A second important consideration regarding more highly func-
tionalized substrates is the propensity for polyketide chains to
undergo additional reorganizations through hemiacetalizations,
spiroacetalizations, and conjugate additions. These and other
potential intramolecular modifications leading to heterocyclic
is programmed in part through protein-protein interactions
between cognate KS and ACP domains.
Our investigation of the non-native fusion protein PikAIII-
TE wherein the C-terminal TE domain from PikAIV was
engineered onto the C-terminus of PikAIII employing the
pentaketide 5 led to the formation of 10-deoxymethynolide (1).
(
31) Weissman, K. J.; Bycroft, M.; Cutter, A. L.; Hanefeld, U.; Frost, E. J.;
Timoney, M. C.; Harris, R.; Handa, S.; Roddis, M.; Staunton, J.; Leadlay,
P. F. Chem. Biol. 1998, 5, 743-754.
-
1
-1
The specificity constant is 0.55 mM min , while the KM is
greater than 1 mM. For comparison, the calculated specificity
(
32) Chuck, J. A.; McPherson, M.; Huang, H.; Jacobsen, J. R.; Khosla, C.; Cane,
D. E. Chem. Biol. 1997, 4, 757-766.
-1
for the pentaketide with the PikAIII‚PikAIV pairing is 12 mM
(33) Wu, N.; Tsuji, S. Y.; Cane, D. E.; Khosla, C. J. Am. Chem. Soc. 2001,
-
1
min . Thus, the pentaketide is processed approximately 22-
fold less efficiently by the non-native fusion PikAIII-TE protein
versus the native proteins. This attenuation cannot be partitioned
into individual kinetic parameters, since only the specificity
constant of PikAIII-TE was measured. Likely, the decrease is
due to improper channeling of the elongated ACP5-hexaketide
substrate to the TE domain, which in turn may be caused by a
nonoptimal linker region between the ACP5 and TE domains
of the non-native fusion protein. We were intrigued about the
possibility of producing narbonolide by PikAIII-TE employing
the hexaketide as an unnatural substrate. Successful processing
of the hexaketide by PikAIII-TE in the absence of NADPH
could potentially generate narbonolide and provide another
mechanism for the formation of this polyketide. However,
hexaketide 15 was not elongated by PikAIII-TE either in the
presence or absence of NADPH, suggesting that PikAIII-TE is
highly evolved to process its native pentaketide substrate.
Further studies are required to establish whether failure of
1
23, 6465-6474.
(
34) Tsuji, S. Y.; Cane, D. E.; Khosla, C. Biochemistry 2001, 40, 2326-2331.
35) Keatinge-Clay, A. T.; Maltby, D. A.; Medzihradszky, K. F.; Khosla, C.;
Stroud, R. M. Nat. Struct. Mol. Biol. 2004, 11, 888-893.
36) Xue, Y.; Sherman, D. H. Nature 2000, 403, 571-574.
37) Thomas, I.; Martin, C. J.; Wilkinson, C. J.; Staunton, J.; Leadlay, P. F.
Chem. Biol. 2002, 9, 781-787.
(
(
(
th
(
(
(
38) Faber, K. Biotransformations in Organic Chemistry, 4 ed.; Springer-
Verlag: Berlin, 2000.
39) Kim, B. S.; Cropp, T. A.; Beck, B. J.; Sherman, D. H.; Reynolds, K. A. J.
Biol. Chem. 2002, 277, 48028-48034.
40) Stutzman-Engwall, K.; Conlon, S.; Fedechko, R.; Kaczmarek, F.; McArthur,
H.; Krebber, A.; Chen, Y.; Minshull, J.; Raillard, S. A.; Gustafsson, C.
Biotechnol. Bioeng. 2003, 82, 359-369.
(41) Arslarian, R. L.; Tang, L.; Blough, S.; Ma, W.; Qiu, R. G.; Katz, L.; Carney,
J. R. J. Nat. Prod. 2002, 65, 1061-1064.
(
(
42) Tang, L.; Qiu, R. G.; Li, Y.; Katz, L. J. Antibiotics 2003, 56, 16-23.
43) Neither 10-deoxymethynolide (1) nor narbonolide (2) was ionized by ESI-
(+), ESI(-), APCI(+), or APCI(-) either with or without formic acid,
thus precluding HPLC-MS analysis. However, the products readily formed
a sodium adduct when dissolved in methanol containing 200 µM NaCl.
44) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923-2925.
45) Lambalot, R. H.; Cane, D. E. J. Antibiot. 1992, 45, 1981-1982.
46) Chen, C.-S.; Fujimoto, Y.; Sih, C. J. J. Am. Chem. Soc. 1981, 103, 3580-
(
(
(
3
582.
(
47) The use of the corresponding CoA thioesters in such cases will serve to
increase the aqueous solubility.
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8447

A R T I C L E S
Aldrich et al.
reaction occurs at the substrate loading, chain extension, or
cyclization/release stage.
more complex heptaketide narbonolide (2) bearing three ad-
ditional chiral centers. By contrast, the chemical synthesis of
narbonolide from the pentaketide intermediate would require
approximately a dozen additional chemical operations. In order
for PKS modules to be truly useful as biocatalysts, however,
they must display a large total turnover number (TTON), which
is the total moles of substrate produced per mole of enzyme
PikAIII and PikAIV Interaction Leading to 10-Deoxym-
ethynolide. The formation of 10-deoxymethynolide (1) by
PikAIII and PikAIV is of great interest biochemically as the
elongated hexaketide chain bound to the ACP5 domain of
PikAIII must be channeled to the C-terminal TE domain of
38
2
0,36
over the lifetime of the biocatalyst. Additionally, the biocata-
lyst should be readily available and have good stability.
Currently, the optimal overexpression conditions afford ap-
proximately 10 mg/L of PikAIII and PikAIV, which translates
to about 100 nmol/L of culture. To be considered synthetically
PikAIV.
A full mechanistic understanding of this process
may allow the effective redesign of the Pik PKS assembly line
to afford exclusively the more valuable 14-membered ring
ketolide template narbonolide (2). On the other hand, it would
also provide a means to redesign modular PKSs for creation of
additional structural diversity in other complex macrolide
pathways. Although we have not systematically investigated
each potential model for generation of 10-deoxymethynolide
and narbonolide by the Pik PKS, we can already discount some
potential mechanisms from the data obtained in the current
analysis. First, direct spontaneous cyclization of the Pik ACP5-
bound elongated hexaketide does not occur, since PikAIII
lacking a TE domain did not produce 10-deoxymethynolide. In
a second model, an N-terminal truncated form of PikAIV lacking
its linker domain was proposed to exclusively catalyze the
synthesis of 10-deoxymethynolide.36 The observation that full-
length PikAIV could provide 10-deoxymethynolide when paired
with PikAIII is not fully consistent with the initial truncation
model, although it does not rigorously disprove it either. A third
model, termed the “skipping” model, proposes that the hexa-
ketide is channeled from the ACP5 domain of PikAIII to the
KS6 domain of PikAIV, but rather than undergoing a final
elongation, it can be transferred to the ACP6 domain, if this is
4
relevant, a TTON of at least 10 is required, since 100 nmol of
enzyme would produce at least 1 mmol of product, a syntheti-
cally tractable amount given the costs required to obtain the
required biocatalysts. Previously, we reported the use of simple
diketide NAC thioester substrates with PikAIII and PikAIV and
observed a maximal TTON of 10; however, this low value was
-
1
due to the extraordinarily low kcat values, ∼0.01 min . The
use of the native pentaketide substrate 5 with PikAIII and
PikAIV by contrast displays a TTON of approximately 300 (data
not shown), primarily as a result of the increased turnover
number of the native substrate. While this represents a significant
improvement over our earlier results, the TTON still requires
considerable optimization. Several factors may contribute to the
observed TTON, including product inhibition of the enzymes,
enzyme deactivation by aberrant decarboxylation of methyl-
malonyl S-ACP, and hydrolysis of substrate by the TE domain.
Enzyme deactivation through aberrant decarboxylation of ACP-
bound methylmalonyl leading to the formation of propionyl-
2
0
39
unoccupied by the usual methylmalonyl extender unit. Alter-
natively, direct transfer from ACP5 of PikAIII to ACP6 of
ACP is well-documented. The resulting stalled PKS assembly
line is unable to undergo subsequent Claisen condensation;
however, the rate of this spontaneous decarboxylation is not
known. A second mechanism that may cause poor TTON
involves hydrolysis of the substrate by the TE domain. In fact,
we observed significant hydrolysis of the substrate NAC
thioesters 5 during extended reaction times by the isolated Pik
TE domain (data not shown). Diffusive loading can occur to
either the KS or TE domain. Under normal physiological
conditions, this competitive mechanism is not operative, since
the TE receives its substrate by channeling from the antecedent
ACP domain; however, in vitro the substrates can diffusively
load onto either domain. Utilization of PKS modules to mediate
the in vitro chemoenzymatic synthesis of substrates represents
a powerful means to prepare ketolide analogues; however, the
current limitations identified here will need to be overcome to
apply PKS modules as synthetically useful biocatalysts.
3
7
PikAIV is possible, effectively bypassing the KS6 domain.
Subsequent channeling to the TE domain provides a mechanism
whereby the hexaketide effectively “skips” through PikAIV to
the TE domain. An in vivo study employing site-directed
mutants provided support for this model, since site-directed
inactivation of the KS6, AT6, or ACP6 domains of PikAIV
completely abolished 10-deoxymethynolide production, sug-
2
0
gesting that these active sites have a role in the process.
Currently, we are exploring the overexpression of site-directed
mutant forms of PikAIV to verify the skipping model using
the current in vitro system with pentaketide 5. A fourth model
suggests that the TE domain can directly accept the ACP-bound
hexaketide. While PikAIII and PikAIV interact through the
respective C- and N-terminal docking domains, the linker
regions between individual domains may impart flexibility to
the overall protein, allowing repositioning of the TE domain to
contact the ACP5 domain. Alternatively, the monomodular
arrangement would allow for direct interaction between ACP5
and the TE domain independent of this interaction. The
conformational flexibility and dynamics of intact PKS modules
are poorly understood at the present, and further investigation
will likely provide new insights into novel mechanisms of
polyketide assembly.
40
Although, polyketide derivatives (e.g. the avermectins and
41,42
epothilones ) with improved bioactivity have been developed
through genetic modifications of host strains and tailoring
enzymes, molecular genetic engineering of altered polyketide
biosynthetic pathways remains slow and tedious. Successful
manipulation of PKS modules to produce macrolactone deriva-
tives requires an understanding of the protein-protein interac-
tions between modules that allows channeling of substrates as
well as any substrate recognition features inherent to the
extending module. Alteration of the polyketide chain through
molecular genetic manipulation of the PKS (i.e. domain
exchanges or module fusions) can produce chains with varied
reduction or alkyl group functionalization, as well as the
Biocatalysis with PKS Modules. The use of recombinant
PKS modules as synthetically useful biocatalysts is an unex-
plored area for the chemoenzymatic synthesis of natural product
templates. In the present system, a pentaketide was converted
through the action of PikAIII and PikAIV into the significantly
8448 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Chemoenzymatic Investigation of Pikromycin Biosynthesis
A R T I C L E S
introduction of amino acids into the elongating chain. Efficient
processing of these derivatives requires the accommodation by
downstream modules. Prescreening of late-stage PKS modules
with synthetic substrates would likely identify limitations that
can be addressed by additional genetic modifications so that in
vivo generated intermediates are channeled and efficiently
processed.
EtOAc (250 µL), and the products were extracted (2 × 250 µL, EtOAc).
The extracts were dried using a Speed-vac and the resulting oil was
dissolved in CH
Merck, 20 × 20 cm with 2.5 cm concentration zone) and developed
using 5% MeOH/CHCl . The silica gel TLC plates were placed on a
2 2
Cl (30 µL) and spotted onto a silica gel TLC plate
(
3
phosphorimager screen and analyzed using a Typhoon (Molecular
Dynamics) phosphoimager. Product formation was quantified using
ImageQuant (Molecular Dynamics) using a standard curve generated
with 2-[14C]methylmalonyl-CoA. Apparent steady-state kinetic values
Significance. The highly favorable reactions between pen-
taketide 5 and hexaketide 15 NAC thioester substrates and the
final two modules [PikAIII (module 5) and PikAIV (module
(kcat and K ) were determined for enzyme-substrate pairings that
M
yielded a detectable product by fitting the normalized V vs [S] plots to
the Michaelis-Menten equation using GraphPad Prism software.
6
)] of the pikromycin PKS show that these enzymes have
evolved to optimally process their native substrates. Our results
also suggest that not only do PKS modules recognize vicinally
located structural features (R and â positions), but that more
remote structural features are decisively involved in proper
substrate recognition and efficient processing by the polyketide
synthases. A systematic investigation employing both structural
and stereochemical substrate analogues is necessary to under-
stand the recognition features and catalytic properties of these
modules and direct the development of bioactive structures.
Screening of downstream modules with substrate analogues may
identify specificity limitations that must be overcome through
protein reengineering before scaled synthesis. Eventually, these
downstream modules may be used biocatalytically or expressed
in a suitable host for biotransformation of synthetic substrates
to generate macrolactones that cannot be efficiently produced
by synthetic or biochemical methods alone.
Confirmation of Reaction Products. Reaction products were
identified by comigration with authentic standards and ESI(+)-MS
analysis was employed to rigorously confirm these assignments. A
large-scale enzymatic reaction (2 mL) with pentaketide 5 and both
PikAIII and PikAIV using the assay conditions described above (except
using cold MMCoA) was performed to provide sufficient 10-deoxym-
ethynolide (1) and narbonolide (2) for detection. The reaction was
quenched at 2 h and extracted with EtOAc (3 × 2 mL). The extracts
were dried using a Speed-vac, and the resulting oil was dissolved in
CH Cl (100 µL) and spotted onto a preparative silica gel TLC plate
2
2
3
and developed using 5% MeOH/CHCl . The overlapping band corre-
sponding to 10-deoxymethynolide (1) and narbonolide (2) was excised
and the semipurified reaction products recovered by extraction with
1
0% MeOH/CHCl
dissolved in 70:25:5 H
by analytical reverse-phase HPLC (Econosil C-18, 250 mm × 4.6 mm)
with a gradient from 30% to 70% CH CN:H O containing 0.1% TFA
over 50 min at a flow rate of 1 mL/min detecting at 233 nm.
0-Deoxymethynolide (1, t ) 32 min) and narbonolide (2, t ) 25
3
(15 mL). The extracts were concentrated and
2
O:CH CN:DMSO (200 µL) and then purified
3
3
2
1
R
R
Experimental Section
min) were manually collected and the fractions concentrated and then
redissolved in MeOH (1 mL) containing 200 µM NaCl and analyzed
Protein Overexpression and Purification. Cloning, overexpression,
43
by ESI(+). 10-Deoxymethynolide (1): (ESI+) m/z 319.1 (C17
H
28
O
4
and purification of PikAIII and PikAIV have been previously de-
+
1
7
+ Na requires 319.2). Narbonolide (2): (ESI+) m/z 375.1 (C H O
scribed. The PikAIII+TE protein was constructed as described by
20 32
5
+
1
5
+ Na requires 375.2). (See Supporting Information for HPLC traces.)
Cane and co-workers but expressed and purified using the same
conditions reported for PikAIII and PikAIV. For this work all proteins
were coexpressed in E. coli cotransformed with a compatible plasmid
that expresses Bacillus subtilus sfp under the control of the T7 promoter
to enhance the phosphopantetheine modification of these proteins.
Protein purity was confirmed to be greater than 95% by SDS-PAGE.
Proteins were stored at -80 °C in storage buffer (100 mM sodium
phosphate, pH 7.2, 1 mM EDTA, 1 mM TCEP, 20% glycerol) for
months without significant loss of activity.
Synthesis of Substrates and Standards. General Procedures. THF
and ether were distilled from Na/benzophenone, and CH
distilled from CaH . Flash chromatography was performed with Fisher
grade silica gel 60 (230-400 mesh) with the indicated solvent system.
2 2
Cl was
2
44
All reactions were performed under an inert atmosphere of dry argon
in oven-dried (150 °C) glassware. Optical rotations were determined
on an Rudolph Research Autopol III polarimeter using the sodium D
D
line (λ ) 589 nm) at 23 °C and are reported as follows: [R] ,
1
13
concentration (c ) g/100 mL), and solvent. H and C NMR spectra
were recorded on a Varian Mercury 300 spectrometer or Bruker 500
MHz spectrometer. Proton and carbon assignments when given are
based on COSY, HMQC, and TOCSY analysis. Proton chemical shifts
are reported in ppm from an internal standard of residual chloroform
In Vitro Polyketide Production. The reactivity of PikAIII, PikAIII-
TE, and PikAIV with SNAC pentaketide 5 and hexaketide 15 was
1
4
determined by measuring the incorporation of 2-[ C]methylmalonyl-
CoA into the resulting narbonolide and 10-deoxymethynolide products.
1
4
2
-[ C]Methylmalonyl-CoA (54.3 mCi/mmol) was purchased from
(7.26 ppm) or methanol (3.31 ppm), and carbon chemical shifts are
American Radiolabeled Chemicals, Inc., and all other chemicals were
purchased from Sigma. Enzymatic reactions were run in 400 mM
sodium phosphate buffer (pH 7.2) containing 5 mM NaCl, 1 mM
EDTA, 1 mM TCEP, 1 mM NADPH (for reactions with PikAIII and
PikAIII-TE), 20% glycerol, and 5% DMSO in a final volume of 50
µL. We have previously demonstrated that the enzyme activity rate
reported using an internal standard of residual chloroform (77.0 ppm)
or methanol (49.1 ppm). Proton chemical data are reported as follows:
chemical shift, multiplicity (ovlp ) overlapping, s ) singlet, d )
doublet, t ) triplet, q ) quartet, p ) pentet, m ) multiplet, br ) broad),
coupling constant, and integration. High-resolution mass spectra were
obtained at the University of Michigan Mass Spectrometry Laboratory
on a Waters Ultima magnetic sector mass spectrometer equipped with
an electrospray interface.
remained linear for up to 1 h using these conditions.17 In reactions
designed to determine the individual reactivity of each protein with
pentaketide 5 and hexaketide 15, protein was added to a concentration
of 1 µM, and pentaketide and hexaketide concentrations ranged from
10-Deoxymethynolide (1). A concentrated spore suspension (100
µL) of SC101626 was used to inoculate 1 L of SCM medium consisting
5
0 µM to 1 mM. In reactions designed to determine the reactivity of
PikAIII and PikAIV together with pentaketide 5, 1 µM of both PikAIII
and PikAIV were used. In all reactions, enzyme was equilibrated in
buffer for 5 min at 30 °C and the reaction was initiated with the addition
of 2-[ C]methylmalonyl-CoA (diluted for a specific activity ) 0.515
mCi/mmol) for a final saturating concentration of 800 µM. The reaction
was maintained at 30 °C for 20 min then quenched by vortexing with
2
of 15 g of soluble starch, 20 g of soytone (Difco), 0.1 g of CaCl , 1.5
g of yeast extract, 0.3 g of Mazu 204, and 10.5 g of MOPS (pH 7.2,
adjusted with 6 N KOH) in a 2 L baffled flask. This primary inoculum
was grown at 30 °C on a rotary shaker (200 rpm) for 24 h. The entire
volume was used to inoculate a 55 L of SCM media in a 75 L bioreactor
containing XAD-16 resin (10 g/L, 500 g, prewashed with MeOH prior
1
4
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8449

A R T I C L E S
Aldrich et al.
to addition to the reactor). During fermentor cultivation, the temperature
was maintained at 30 °C and the pH at 7.2 via addition of 1.0 N NaOH
or 1 N H SO . The aeration rate was set at 25 L/min and the agitation
2 4
was controlled at 400 rpm such that the dissolved oxygen tension was
maintained above 40%. The culture was grown for 66 h, and the XAD-
was filtered using a Centriprep YM-10 MW membrane filter to remove
the enzyme. The filtrate was then acidified to pH 2.0 by the addition
of concentrated HCl and extracted with EtOAc (3 × 150 mL). The
2
organic extracts were concentrated and chromatographed (SiO , 2%
MeOH/CH
2 2
Cl ) to afford (3.99 g, 83%) of the title compound as a
1
13
1
6 resin was collected by filtering through a 60-mesh screen. The resin
was washed with H O (500 mL) and then 10-deoxymethynolide was
eluted with MeOH (500 mL). The MeOH rinse was concentrated onto
Celite (20 g) to afford a brown cake. Flash chromatography (SiO , 200
g, 1% MeOH/CH Cl ) afforded 10-deoxymethynolide (5.17 g, 94 mg/
L) as a light yellow oil that slowly solidified at -20 °C, the H NMR
light yellow oil, the H and C NMR of which agreed with reported
values.23 [R]
46
2
D
3
-3.94 (c ) 7.18, CHCl ) corresponds to 94% ee.
(E)-(2S,4R,8R,9R)-9-(tert-Butyldimethylsilanyloxy)-2,4,8-trimeth-
2
yl-5-oxoundec-6-enoic Acid (10). To a solution of phosphonate 6 (194
mg, 0.73 mmol, 1.0 equiv) in THF (3.0 mL) was added finely ground
2
2
1
Ba(OH)
min. A solution of aldehyde 7 (202 mg, 0.88 mmol, 1.2 equiv) in
20:1 THF:H O (3 mL) was added and the reaction stirred for 16 h.
2
2
‚8H O (437 mg, 1.38 mmol, 1.9 equiv) and the reaction stirred
1
3
45
and C NMR of which matched literature data.
E)-(2S,4R,8R,9R)-S-2-Acetamidoethyl 9-Hydroxy-2,4,8-trimeth-
yl-5-oxoundec-6-enethioate (5). To a solution of TBS-pentaketide 11
44.5 mg, 0.0944 mmol) in CH CN (5 mL) at 23 °C in a 15 mL plastic
vial was added 48% aqueous HF (0.25 mL). After 16 h, the reaction
22
5
(
2
The reaction was partitioned between EtOAc (30 mL) and 1.0 N
aqueous HCl (30 mL). The aqueous layer was extracted with EtOAc
(
3
(2 × 20 mL), and the combined organic extracts were dried (Na
and concentrated under reduced pressure to a colorless oil. Flash
chromatography (SiO , 5 g, 2% MeOH/CHCl ) afforded the title
compound (203 mg, 75%) as a colorless oil: TLC R ) 0.18 (1% MeOH
+ 0.5% glacial AcOH/CH Cl ); [R] +27 (c ) 0.90, CH
(CDCl , 500 MHz) δ 6.94 (dd, J ) 15.9, 7.5 Hz, 1H), 6.14 (dd, J )
2 4
SO )
was quenched with saturated aqueous NaHCO
extracted with CH Cl SO
(3 × 30 mL), dried (Na
under reduced pressure. Flash chromatography (SiO
CH Cl ) afforded (28 mg, 84%) the title compound as a colorless oil:
TLC R ) 0.16 (3% MeOH/CHCl ); [R] +36.5 (c ) 0.203, MeOH);
H NMR (CDCl , 300 MHz) δ 6.90 (dd, J ) 15.9, 7.8 Hz, 1H), 6.15
dd, J ) 15.9, 1.2 Hz, 1H), 6.00 (br s, 1H), 3.51-3.57 (m, 1H), 3.35-
3
(5 mL) and then
2
2
2
4
), and concentrated
2
3
2
, 2 g, 3% MeOH/
f
1
2
2
2
2
D
2 2
Cl ); H NMR
f
3
D
3
1
3
15.9, 1.5 Hz, 1H), 3.55 (q, J ) 5.9 Hz, 1H), 2.86 (hext, J ) 7.0 Hz,
1H), 2.44-2.57 (m, 2H), 2.07-2.17 (m, 1H), 1.44-1.52 (m, 1H), 1.34-
1.43 (m, 2H), 1.18 (d, J ) 6.9 Hz, 3H), 1.11 (d, J ) 7.0 Hz, 3H), 1.02
(d, J ) 6.8 Hz, 3H), 0.88 (ovlp s, 9H), 0.86 (ovlp t, J ) 7.4 Hz, 3H),
(
3
2
1
6
2
2
.49 (m, 2H), 2.99-3.04 (m, 2H), 2.82 (hext, J ) 6.9 Hz, 1H), 2.70-
.79 (m, 1H), 2.42-2.52 (m, 1H), 2.10-2.20 (m, 1H), 1.98 (s, 3H),
.35-1.58 (m, 3H), 1.18 (d, J ) 6.9 Hz, 3H), 1.11 (app t, J ) 6.9 Hz,
0.03, (s, 3H), 0.02 (s, 3H); 13C NMR (CDCl
3
, 125 MHz) δ 203.0, 182.4,
1
3
3
H), 0.98 (t, J ) 7.5 Hz, 3H); C NMR (CDCl , 75 MHz) δ 203.8,
150.8, 127.9, 76.4, 41.5, 41.2, 37.0, 36.2, 26.8, 25.8, 18.1, 17.5, 16.6,
02.6, 170.2, 150.0, 128.1, 75.8, 46.7, 42.3, 41.5, 39.5, 37.2, 28.5, 27.3,
14.2, 9.6, -4.3, -4.5; HRMS (ESI+) m/z 393.2433 (C20
Na requires 393.2437).
38 4
H O Si +
+
3.3, 18.7, 17.0, 13.6, 10.5; HRMS (ESI+) m/z 380.1871 (C18
Na requires 380.1872).
4
H31NO S
+
+
(E)-(2S,4R,8R,9R)-S-2-Acetamidoethyl 9-(tert-Butyldimethylsila-
(
2S,4R)-6-(Dimethoxyphosphoryl)-2,4-dimethyl-5-oxohexanoic Acid
6). To a solution of dimethyl methylphosphonate (236 µL, 2.20 mmol,
.3 equiv) in THF (3 mL) at -78 °C was added a solution of n-BuLi
2.5 M in hexanes, 1.38 mL, 2.20 mmol, 3.3 equiv) to afford a white
nyloxy)-2,4,8-trimethyl-5-oxoundec-6-enethioate (11). To a stirring
solution of acid 10 (53.1 mg, 0.143 mmol, 1.0 equiv), EDC‚HCl (34
mg, 0.215 mmol, 1.5 equiv), and DMAP (1.7 mg, 0.014 mmol, 0.1
(
3
(
2 2
equiv) in CH Cl (1.5 mL) at 23 °C was added N-acetylcysteamine
heterogeneous mixture, which was stirred for an additional 15 min. A
solution of ester 9 (119 mg, 0.67 mmol, 1.0 equiv, azeotropically dried
with PhH, 2 × 5 mL) in THF (3 mL) was added slowly down the side
of the flask to the lithium phosphonate solution at -78 °C to afford a
clear solution. After stirring for 25 min, the cooling bath was removed
and the solution was allowed to warm to 0 °C. The reaction was
partitioned between EtOAc (25 mL) and 1 N aqueous HCl (10 mL)
and saturated aqueous NaCl (10 mL). The organic layer was separated
and the aqueous layer was extracted with EtOAc (2 × 25 mL). The
(22 µL, 0.20 mmol, 1.43 equiv). After 16 h, the reaction was diluted
with EtOAc (10 mL) and washed successively with 1 N aqueous HCl,
saturated aqueous NaHCO
(Na SO ) and concentrated under reduced pressure to a colorless oil.
Flash chromatography (SiO , 1% MeOH/CHCl ) afforded (60.3 mg,
90%) the title compound as a colorless oil. TLC R ) 0.15 (1% MeOH/
+ 36.3 (c ) 0.755, MeOH); H NMR (CDCl , 300 MHz)
3
, and saturated aqueous NaCl and then dried
2
4
2
3
f
1
CHCl
3
); [R]
D
3
δ 6.93 (dd, J ) 15.9, 7.4 Hz, 1H), 6.12 (dd, J ) 15.9, 1.0 Hz), 5.99
(br s, 1H), 3.56 (q, J ) 5.6 Hz, 1H), 3.35-3.47 (m, 2H), 3.01 (t, J )
6.2 Hz, 2H), 2.79 (hext, J ) 7.2 Hz, 1H), 2.69 (hext, J ) 6.9 Hz, 1H),
2.44-2.53 (m, 1H), 2.12-2.22 (m, 1H), 1.98 (s, 3H), 1.30-1.55 (m,
3H), 1.17 (d, J ) 6.9 Hz, 3H), 1.11 (d, J ) 7.0 Hz, 3H), 1.03 (d, J )
combined organic extracts were dried (Na
trated under reduced pressure to a colorless oil. Flash chromatography
SiO , 20 g, 5% MeOH/CH Cl ) afforded (142 mg, 80% yield) a
colorless oil: TLC R ) 0.20 (5% MeOH/CH Cl ); [R] -0.63 (c )
.2, MeOH); H NMR (CDCl , 500 MHz) δ 9.22 (br s, 1H, OH), 3.80
s, 3H, OMe), 3.77 (s, 3H, OMe), 3.21-3.26 (m, 1H, one H-6), 3.15-
.19 (m, 1H, one H-6), 2.80 (hext, J ) 6.9 Hz, 1H, H-4), 2.47-2.54
m, 1H, H-2), 2.10 (ddd, J ) 13.8, 8.9, 5.9 Hz, 1H, H-3), 1.32 (ddd,
J ) 13.8, 7.8, 5.7 Hz, 1H, H-3), 1.19 (d, J ) 6.9 Hz, 3H, C-4 Me),
2 4
SO ), filtered, and concen-
(
2
2
2
f
2
2
D
6.8 Hz, 3H), 0.88 (ovlp s, 9H), 0.87 (ovlp t, J ) 7.5 Hz, 3H), 0.04 (s,
1
13
2
3
3H), 0.03 (s, 3H); C NMR (CDCl
3
, 75 MHz) δ 203.5, 202.5, 170.1,
(
3
(
150.7, 127.6, 76.4, 46.5, 41.6, 39.5, 36.8, 29.8, 28.5, 26.8, 25.9, 23.2,
18.4, 18.2, 17.3, 14.3, 9.8, -4.2, -4.4; HRMS (ESI+) m/z 494.2743
+
(C24H45NO
4
SSi + Na requires 494.2736).
(
7S)-7-Dihydro-10-deoxymethynolide (12). To a stirring solution
of 10-deoxymethynolide (391 mg, 1.32 mmol, 1.0 equiv) and CeCl
O (492 mg, 1.32 mmol, 1.0 equiv) in MeOH (10 mL) at -10 °C
was added NaBH (50 mg, 1.32 mmol, 1.0 equiv) in three portions.
1
3
1
2
.12 (d, J ) 6.9 Hz, 3H, C-2 Me); C NMR (CDCl , 125 MHz) δ
3
3
3
‚
3
04.7 (d, JC-P ) 6.5 Hz, C-5), 179.9 (C-1), 53.2 (d, JC-P ) 3.8 Hz,
OMe), 53.1 (d, JC-P ) 3.8 Hz, OMe), 45.2 (C-4), 39.0 (d, JC-P
7H
2
3
2
)
4
1
31 Hz, C-6), 37.1 (C-2), 35.7 (C-3), 17.7 (C-4 Me), 15.8 (C-2 Me);
After 5 min, the reaction was diluted with EtOAc (20 mL) and poured
onto 1 N aqueous HCl (20 mL) and saturated aqueous NaCl (20 mL).
The aqueous layer was extracted with EtOAc (2 × 25 mL), and the
31
P NMR (CDCl
Na+ requires 289.0817).
2S,4R)-4-(Methoxycarbonyl)-2-methylpentanoic Acid (9). R-Chy-
3 19 6
) δ 24.8; HRMS (ESI+) m/z 289.0813 (C10H O P +
(
combined organic extracts were dried (Na
2
SO
4
) and concentrated under
, 40 g,
) afforded (275 mg, 70%) a white foam: TLC R
0.28 (5% MeOH/CHCl , vanillin stain); [R] ) + 84.0 (c ) 1.07,
MeOH); H NMR (CDCl , 300 MHz) δ 5.70 (ddd, J ) 15.9, 4.8, 1.8
motrypsin (EC 3.4.21.1, 500 mg, 25,500 units) was added to a
vigorously stirring suspension of 8 (5.16 g, 27.4 mmol, 1.0 equiv) in
reduced pressure to a yellow oil. Flash chromatography (SiO
2% MeOH/CHCl
2
3
f
)
100 mM sodium phosphate (pH 8.0, 150 mL) at 4 °C. A 1.0 N aqueous
3
D
1
NaOH solution (27.4 mL, 27.4 mmol, 1.0 equiv) was added dropwise
over 96 h to maintain the pH between 7.0 and 8.0. During this time
the substrate slowly dissolved to form a cloudy solution. Additional
R-chymotrypsin was added (500 mg) at 48 h. After 96 h, the reaction
3
Hz, 1H, H-9), 5.49 (ddd, J ) 15.9, 3.6, 1.8 Hz, 1H, H-8), 5.00 (ddd,
J ) 8.7, 5.1, 3.0 Hz, 1H, H-11), 4.09-4.14 (br m, 1H, H-7), 3.53 (dd,
J ) 10.2, 1.8 Hz, 1H, H-3), 2.48-2.62 (m, 2H, H-10, H-2), 1.84-
8450 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005

Chemoenzymatic Investigation of Pikromycin Biosynthesis
A R T I C L E S
1
1
.98 (m, 2H, H-4, H-6), 1.50-1.72 (m, 5H, one H-5, H-12, 2 OH),
.32-1.41 (m, 1H, one H-5), 1.27 (d, J ) 6.9 Hz, 3H, C-10 Me), 1.05
(C-9), 133.0 (C-8), 78.3 (C-3), 77.8 (C-11), 76.1 (C-7), 52.9 (C-2),
44.0 (C-10), 40.1 (AcNHCH
35.9 (C-4 or C-6), 29.1 (AcNHCH
(CH CdONHCH CH S), 18.0 (C-4 Me or C-6 Me), 16.4 (2C, C-10
2 2
CH S), 38.3 (C-4 or C-6), 36.3 (C-5),
(app t, J ) 6.6 Hz, 6H, C-2 Me, C-4 Me), 0.98 (d, J ) 6.9 Hz, 3H,
2
CH S), 28.4 (C-12), 22.5
2
1
3
C-6 Me), 0.91 (t, J ) 7.2 Hz, 3H, H-13); C NMR (CDCl
3
, 75 MHz)
3
2
2
δ 171.2, 130.8, 128.2, 78.7. 77.2, 75.9, 43.6, 37.8, 35.4, 32.6, 31.9,
and C-4 Me or C-6 Me), 13.1 (C-2 Me), 10.7 (C-13); HRMS (ESI+)
+
2
4.4, 20.7, 17.3, 16.6, 10.9, 10.5; HRMS (ESI+) m/z 321.2036
m/z 440.2451 (C21
H39NO
5
S + Na requires 440.2447).
+
1
(C H O
17 30 4
+ Na requires 321.2042).
Minor isomer C -epi-14: [R] ) +47 (c ) 0.095, MeOH); H NMR
2
D
(
E)-(2R,3S,4S,6R,7S,10R,11R)-3,7,11-Trihydroxy-2,4,6,10-tetra-
methyltridec-8-eneioic Acid (13) and (E)-(2S,3S,4S,6R,7S,10R,11R)-
,7,11-Trihydroxy-2,4,6,10-tetramethyltridec-8-eneoic Acid, Minor
Isomer (C -epi-13). To a stirring solution of diol 12 (184 mg, 0.62
3
(CDCl , 500 MHz) δ 5.52 (dd, J ) 15.6, 7.5 Hz, 1H, H-9), 5.45 (dd,
J ) 15.6, 5.9 Hz, 1H, H-8), 3.92 (dd, J ) 5.9, 3.6 Hz, 1H, H-7), 3.55
(dd, J ) 8.6, 3.4 Hz, 1H, H-3), 3.19-3.29 (part obsc m, 3H, H-11,
3
AcNHCH
8.6, 7.0 Hz, 1H, H-2), 2.13 (hext, J ) 6.9 Hz, 1H, H-10), 1.84 (s, 3H,
CH CdONHCH CH S), 1.70-1.78 (m, 1H, H-4), 1.59-1.64 (ovlp m,
2 2 2 2
CH S), 2.89-2.98 (m, 2H, AcNHCH CH S), 2.83 (dq, J )
2
mmol, 1.0 equiv) in MeOH (10 mL) at 23 °C was added a 0.62 M
aqueous LiOH solution (10 mL, 6.2 mmol, 10.0 equiv) to afford a
slightly cloudy solution. The resulting solution was refluxed for 9 d.
LiOH precipitated during the course of the reaction as a white solid;
thus an additional 1:1 MeOH:0.62 M aqueous LiOH solution (5 mL)
was added every 3 d. The reaction was cooled to 23 °C and quenched
with 1.0 N aqueous HCl (20 mL). This solution was partitioned between
EtOAc (30 mL) and saturated aqueous NaCl (30 mL). The aqueous
phase was extracted with EtOAc (2 × 30 mL), and the combined
3
2
2
1H, one H-5), 1.55-1.60 (ovlp m, 1H, H-6), 1.45-1.55 (ovlp m, 1H,
one H-12), 1.24-1.33 (m, 1H, one H-12), 1.07 (d, J ) 7.0 Hz, 3H,
C-2 Me), 0.98 (d, J ) 6.8 Hz, 3H, C-10 Me), 0.93 (d, J ) 6.8 Hz, 3H,
C-4 Me), 0.88 (ovlp t, J ) 7.4 Hz, 3H, H-13), 0.84-0.87 (ovlp m, 1H,
one H-5), 0.83 (d, J ) 6.7 Hz, 3H, C-6 Me); 13C NMR (CD
MHz) δ 204.2 (C-1), 173.5 (CH CdONHCH CH
3 2 2
OD, 125
S), 135.4 (C-9), 133.1
(C-8), 79.0 (C-3), 77.8 (C-11), 75.2 (C-7), 53.5 (C-2), 43.9 (C-10),
40.1 (AcNHCH CH S), 37.8 (C-6), 34.1 (C-4), 33.8 (C-5), 29.1
(AcNHCH CH S), 28.4 (C-12), 22.5 (CH CdONHCH CH S), 18.1
3
organic extracts were dried (Na
pressure to an oil. Flash chromatography (SiO
5% MeOH/CH Cl ) afforded (178 mg, 91%) the title compound as
an inseparable 4:1 mixture of 13 and C -epi-13: TLC R ) 0.09 (1%
glacial AcOH & 5% MeOH/CHCl , vanillin stain); H NMR (CD OD,
00 MHz) δ 5.47-5.61 (m, 2H, H-8, H-9), 3.95 (ovlp t, J ) 5.1 Hz,
.8H, H-7major), 3.95-4.01 (ovlp m, 0.2H, H-7minor), 3.61 (t, J ) 6.0
2
SO
4
) and concentrated under reduced
2
2
2
, 20 g, 1% glacial AcOH
2
2
3
2
2
+
2
2
(C-4 Me), 16.3 (2C, C-6 Me and C-10 Me), 15.6 (C-2 Me), 10.7 (C-
+
13); HRMS (ESI+) m/z 440.2451 (C21
H39NO
5
S + Na requires
2
f
1
440.2447).
3
3
3
0
3-Oxo-10-deoxymethynolide (18). To a stirring solution of (COCl)2
(90 µL, 1.03 mmol, 3.0 equiv) in CH Cl (3 mL) at -78 °C was added
2
2
Hz, 0.8H, H-3major), 3.47-3.52 (m, 0.2H, H-3minor), 3.25-3.29 (ovlp
m, 1H, H-11), 2.58-2.68 (m, 1H, H-2), 2.20 (hext, J ) 6.6 Hz, 1H,
H-10), 1.86 (ddd, J ) 13.5, 7.8, 3.6 Hz, 1H, one H-5), 1.61-1.73 (ovlp
m, 2H, H-4, H-6), 1.52-1.61 (ovlp m, 1H, one H-12), 1.28-1.42 (m,
a solution of DMSO (98 µL, 1.38 mmol, 4.0 equiv) in CH Cl (1 mL).
After 10 min, a solution of 10-deoxymethynolide (102 mg, 0.34 mmol,
2
2
1.0 equiv) in CH Cl (1.0 mL + 1.0 mL wash) was added dropwise
2
2
via cannula and the reaction stirred another 15 min at -78 °C. Et N
3
(237 µL, 1.70 mmol, 5.0 equiv) was added dropwise and the resulting
solution stirred 30 min at -78 °C and then the dry ice bath was removed
and the reaction warmed to 23 °C. The reaction solution was washed
1
3
H, one H-12), 1.16 (d, J ) 6.9 Hz, 3H, C-2 Me), 1.04 (d, J ) 6.6 Hz,
H, C-10 Me,), 0.89-0.98 (m, 10 H, C-4 Me, C-6 Me, H-13, one
1
3
3
H-5); C NMR (CD OD, 75 MHz) δ 156.8, 135.3, 133.0, 78.3, 78.0,
7
5.9, 44.2, 38.4, 36.9, 35.9 (2C), 28.5, 18.0, 16.6, 16.5, 12.2, 10.9;
successively with 1 N aqueous HCl (5 mL), H O (5 mL), and saturated
2
+
HRMS (ESI+) m/z 339.2144 (C17
E)-(2R,3S,4S,6R,7S,10R,11R)-S-2-Acetamidoethyl 3,7,11-Tri-
hydroxy-2,4,6,10-tetramethyltridec-8-enethioate (14) and (E)-
2S,3S,4S,6R,7S,10R,11R)-S-2-Acetamidoethyl 3,7,11-Trihydroxy-
,4,6,10-tetramethyltridec-8-enethioate, Minor Isomer (C -epi-14).
To a solution of 13 and C -epi-13 (64 mg, 0.20 mmol, 1 equiv), EDC‚
HCl (48 mg, 0.30 mmol, 1.5 equiv), and DMAP (2.5 mg, 0.1 mmol,
.1 equiv) in CH Cl (2 mL) at 23 °C was added N-acetylcysteamine
65 µL, 0.60 mmol, 3 equiv). After 16 h, the reaction was partitioned
H
32
O
5
+ Na requires 339.2147).
aqueous NaCl (5 mL) and then dried (Na SO ) and concentrated under
2
4
(
reduced pressure to an oil. Flash chromatography (SiO
2
, 10 g, 40%
EtOAc/hexanes) afforded (52 mg, 52%) a white crystalline solid: mp
2
(
2
97-98 °C; TLC R
f
) 0.70 (50% EtOAc/hexanes); [R]
D
+1.9 × 10 (c
1
) 1.4, MeOH); H NMR (CDCl , 500 MHz) δ 6.82 (dd, J ) 15.8, 5.2
3
2
Hz, 1H), 6.48 (dd, J ) 15.8, 1.2 Hz, 1H), 5.09 (ddd, J ) 8.3, 5.6, 2.3
Hz, 1H), 3.56 (q, J ) 7.1 Hz, 1H), 2.69-2.75 (m, 1H), 2.56-2.64 (m,
1H), 2.41-2.48 (m, 1H), 1.94 (ddd, J ) 14.1, 12.0, 2.0 Hz, 1H), 1.68-
1.78 (m, 1H), 1.55-1.63 (m, 1H), 1.28-1.36 (ovlp m, 1H), 1.31 (ovlp
d, J ) 7.1 Hz, 3H), 1.21 (d, J ) 7.0 Hz, 3H), 1.14 (d, J ) 6.8 Hz,
2
0
(
2
2
between 0.1 N aqueous HCl (5 mL) and EtOAc (5 mL). The aqueous
1
3
layer was extracted with EtOAc (2 × 5 mL), and the organic extracts
3H), 1.01 (d, J ) 6.4 Hz, 3H), 0.92 (t, J ) 7.4 Hz, 3H); C NMR
(CDCl , 75 MHz) δ 207.2, 203.9, 172.3, 147.7, 125.8, 75.4, 50.0, 45.1,
were washed with saturated aqueous NaCl (5 mL), dried (Na
2
4
SO ),
3
and concentrated under reduced pressure to a colorless oil. The resulting
oil was dissolved in 35% CH CN/H O (2 mL) and purified in two
portions by preparative reverse-phase HPLC (Econosil C18, 22 × 250
mm, Alltech). An isocratic elution of 65:35 H O:CH CN with 0.1%
TFA at a flow rate of 10 mL/min monitoring at 240 nm afforded (58
mg, 70%) 14 (t ) 28.0 min) as a white solid after lyophilization and
12.6 mg, 15%) of C -epi-14 (t ) 24.1 min) as a white solid. Major
isomer 14: [R] ) +7.60 (c ) 0.605, MeOH); H NMR (CD
00 MHz) δ 5.40-5.47 (m, 2H, H-8, H-9), 3.84 (t, J ) 4.4 Hz, 1H,
H-7), 3.56 (t, J ) 5.9 Hz, 1H, H-3), 3.22-3.26 (m, 2H, AcNHCH CH S),
.17-3.20 (ddd, J ) 9.3, 6.4, 3.4 Hz, 1H, H-11), 2.87-2.97 (m, 2H,
AcNHCH CH S), 2.80 (p, J ) 6.7 Hz, 1H, H-2), 2.12 (hext, J ) 6.6
Hz, 1H, H-10), 1.83 (s, 3H, CH CdONHCH CH S), 1.71-1.76 (m,
H, one H-5), 1.54-1.63 (m, 2H, H-4, H-6), 1.45-1.53 (m, 1H, one
41.6, 38.2, 38.1, 25.0, 17.3, 14.1, 13.5, 10.2, 9.4; HRMS (ESI+) m/z
+
317.1723 (C17
H O
26 4
+ Na requires 317.1729).
3
2
(E)-[10-(2R,4S,5R)-3R,4R,7S,8R,10S]-10-[2-(4-Methoxy-benzyl)-
5-methyl-[1,3]dioxan-4-yl]-4,8-dimethylundec-5-ene-3,7-diol (20). To
2
3
a 4:1 mixture of acids 13 and C
in MeOH (5 mL) was added TMSCH
2
-epi-13 (50 mg, 0.158 mmol, 1.0 equiv)
(2.0 M in Et O, 790 µL, 1.58
R
N
2 2
2
(
2
R
mmol, 10.0 equiv) at 0 °C. The reaction was stirred for 10 min at 0 °C
and then concentrated under reduced pressure. Flash chromatography
1
D
3
OD,
5
(SiO
ester as a colorless oil. The methyl ester was dissolved in THF (5 mL),
and LiAlH (10 mg, 0.25 mmol, 2.0 equiv) was added at 0 °C. The
2 2 2
, 10 g, 2% MeOH/CH Cl ) afforded 39 mg (75%) of the methyl
2
2
3
4
2
2
reaction was stirred for 16 h at 23 °C and then diluted with EtOAc (10
mL) and washed with 1.0 N aqueous sodium potassium tartrate solution
(10 mL). The aqueous layer was extracted with EtOAc (2 × 10 mL),
3
2
2
1
H-12), 1.21-1.31 (m, 1H, one H-12), 1.11 (d, J ) 6.9 Hz, 3H, C-2
Me), 0.95 (t, J ) 6.8 Hz, 3H, C-10 Me), 0.86 (ovlp t, J ) 7.5 Hz, 3H,
H-13), 0.85 (ovlp d, J ) 6.7 Hz, 3H, C-4 Me), 0.81 (ovlp d, J ) 6.8
2 4
and the combined organic extracts were dried (Na SO ), and concen-
trated under reduced pressure to a colorless oil (39.1 mg, 108%), which
was taken directly onto the next step. To a mixture of tetraols 19
prepared in the previous step (azeotropically dried with PhMe, 2 × 10
1
3
Hz, 3H, C-6 Me), 0.77-0.81 (ovlp m, 1H, one H-5); C NMR (CD
3
-
OD, 125 MHz) δ 203.7 (C-1), 173.4 (CH CdONHCH CH S), 135.5
3
2
2
6 4 2
mL) was added a solution of p-MeOC H CH(OMe) (30 µL, 0.17 mmol,
J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005 8451

A R T I C L E S
Aldrich et al.
1
.5 equiv) and CSA (2.7 mg, 0.012 mmol, 0.10 equiv) in DMF (1.0
mL), and the resulting solution was stirred for 16 h at 23 °C. The
reaction was diluted with EtOAc (50 mL), washed with H
O (3 × 25
mL), dried (Na SO ), and concentrated under reduced pressure to a
colorless oil. Flash chromatography (SiO , 5 g, 2% MeOH/CHCl
afforded the major isomer 20 (26 mg, 55%) and the minor isomer 21
7 mg, 15%). Data for major isomer 20. TLC R ) 0.38 (2% MeOH/
CHCl , vanillin stain); [R] ) -1.74 (c ) 0.575, MeOH); H NMR
CDCl , 500 MHz) δ 7.40 (d, J ) 8.7 Hz, 2H, Ar-ortho-H), 6.88 (d, J
8.7 Hz, 2H, Ar-meta-H), 5.52 (dd, J ) 15.7, 7.2 Hz, 1H, H-9), 5.46
125 MHz) δ 160.4, 134.0, 132.5, 132.2, 128.3, 113.9, 102.2, 85.3, 76.4,
74.8, 73.8, 54.8, 42.5, 38.0, 37.4, 32.7, 30.5, 27.5, 16.5, 15.7, 14.8,
+
2
11.1, 10.7; HRMS (ESI+) m/z 443.2777 (C H O + Na requires
2
5
40
5
2
4
443.2773).
2
3
)
Acknowledgment. We would like to thank Prof. Suzanne
Admiraal of the Department of Biological Chemistry, University
of Michigan Medical School for helpful comments on the
manuscript. This research was supported by grants from NIH
(GM48562) and NSF (NSF/BES-0118926) to D.H.S.
(
f
1
3
D
(
3
)
(dd, J ) 15.7, 5.5 Hz, 1H, H-8), 5.42 (s, 1H, acetal-H), 4.00-4.04 (m,
Supporting Information Available: HPLC traces of com-
pounds 1 and 2 isolated from in vitro reactions with PikAIII
and PikAIV, representative radio-TLC data used to determine
3
H, H-1, H-7), 3.79 (s, 3H, ArOCH ), 3.46 (dd, J ) 10.0, 2.1 Hz, 1H,
3
H-3), 3.33 (ddd, J ) 8.8, 5.1, 3.9 Hz, 1H, H-11), 2.20 (hext, J ) 6.8
Hz, 1H, H-10), 1.89 (ddd, J ) 11.9, 8.1, 3.7 Hz, 1H, one H-5), 1.70-
1
1
.80 (ovlp m, 2H, H-4, H-6), 1.62-1.70 (ovlp m, 1H, H-2), 1.54-
.62 (ovlp br s, 2H, C-7 OH, C-11 OH), 1.47-1.54 (ovlp m, 1H, one
steady-state kinetic parameters of Pik monomodules, and H
1
3
1
NMR and C NMR spectra of 5, 6, 9-14, 18, and 20. This
material is available free of charge via the Internet at
http://pubs.acs.org.
H-12), 1.28-1.37 (m, 1H, one H-12), 1.16 (d, J ) 6.9 Hz, 3H, C-2
Me), 0.98-0.90 (ovlp m, 1H, one H-5), 0.97 (ovlp d, J ) 6.8 Hz, 3H,
C10-Me), 0.93 (t, J ) 7.4 Hz, 3H, H-13), 0.87 (d, J ) 6.8 Hz, 3H, C-4
13
or C-6 Me), 0.85 (d, J ) 6.8 Hz, 3H, C-4 or C-6 Me); C NMR (C
6 6
D ,
JA042592H
8452 J. AM. CHEM. SOC.
9
VOL. 127, NO. 23, 2005