Investigation of fungal iterative polyketide synthase functions using partially assembled intermediates

Article abstract of DOI:10.1021/ja4001823

Iterative polyketide synthases (PKSs) are large, multifunctional enzymes that resemble eukaryotic fatty acid synthases but can make highly functionalized secondary metabolites using complex and unresolved programming rules. During biosynthesis of the kinase inhibitor hypothemycin by Hypomyces subiculosus, a highly reducing iterative PKS, Hpm8, cooperates with a nonreducing iterative PKS, Hpm3, to construct the advanced intermediate dehydrozearalenol (DHZ). The identity of putative intermediates in the formation of the highly reduced hexaketide portion of DHZ were confirmed by incorporation of ¹³C-labeled N-acetylcysteamine (SNAC) thioesters using the purified enzymes. The results show that Hpm8 can accept SNAC thioesters of intermediates that are ready for transfer from its acyl carrier protein domain to its ketosynthase domain and assemble them into DHZ in cooperation with Hpm3. Addition of certain structurally modified analogues of intermediates to Hpm8 and Hpm3 can produce DHZ derivatives.

Full text of DOI:10.1021/ja4001823

Communication
pubs.acs.org/JACS
Investigation of Fungal Iterative Polyketide Synthase Functions
Using Partially Assembled Intermediates
Zhizeng Gao,†,∥ Jingjing Wang,‡,∥ Amy K. Norquay,† Kangjian Qiao,‡ Yi Tang,*,‡,§ and John C. Vederas*,†
†Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
‡Department of Chemical and Biomolecular Engineering and §Department of Chemistry and Biochemistry, University of California,
Los Angeles, California 90095, United States
S
* Supporting Information
probe the mechanism of modular bacterial PKS systems relies
ABSTRACT: Iterative polyketide synthases (PKSs) are
large, multifunctional enzymes that resemble eukaryotic
fatty acid synthases but can make highly functionalized
secondary metabolites using complex and unresolved
programming rules. During biosynthesis of the kinase
inhibitor hypothemycin by Hypomyces subiculosus, a highly
reducing iterative PKS, Hpm8, cooperates with a non-
reducing iterative PKS, Hpm3, to construct the advanced
intermediate dehydrozearalenol (DHZ). The identity of
putative intermediates in the formation of the highly
reduced hexaketide portion of DHZ were confirmed by
incorporation of 13C-labeled N-acetylcysteamine (SNAC)
thioesters using the purified enzymes. The results show
that Hpm8 can accept SNAC thioesters of intermediates
that are ready for transfer from its acyl carrier protein
domain to its ketosynthase domain and assemble them
into DHZ in cooperation with Hpm3. Addition of certain
structurally modified analogues of intermediates to Hpm8
and Hpm3 can produce DHZ derivatives.
on the use of N-acetylcysteamine (SNAC) thioesters of
partially assembled precursors.15−17 Using PKS domain
inactivation followed by addition of putative intermediate
SNAC thioesters also allows bacterial polyketides to be
produced in cell-free systems.18,19 However, it is challenging
to apply this advanced precursor feeding approach to iterative
PKS systems because domain inactivation abolishes the
production of polyketides. In addition, this approach often
fails with whole or disrupted fungal cells because of catabolism
of the precursors, except in some isolated cases.20−22 Therefore,
the development of new methods to probe fungal PKS function
using partially assembled intermediates could aid the
elucidation of these complex machineries.
Hypothemycin biosynthesis has been studied in considerable
detail.23−25 Two iterative PKS proteins, Hpm8 and Hpm3,
build the polyketide backbone of hypothemycin. Hpm8, a
highly reducing PKS (HRPKS), first assembles a reduced
hexaketide intermediate. With the assistance of a starter unit
acyl-carrier protein transacylase (SAT) domain, this newly
formed hexaketide is transferred to Hpm3, a nonreducing PKS
(NRPKS), where it is further extended to a nonaketide, which
then undergoes regioselective cyclization and macrolactoniza-
tion to afford (6′S,10′S)-7′,8′-dehydrozearalenol (DHZ).24
Subsequent post-PKS modifications of DHZ by other enzymes
afford hypothemycin. Although the general functions of the two
PKSs have been assigned, it remains unresolved how Hpm8
controls the tailoring of the intermediates en route to its
hexaketide product using its reductive domains (KR, DH, and
ER) in a permutative fashion. To probe the programing rules of
HRPKSs, confirmation of the structures of enzyme-bound
intermediates and the way they interact with HRPKSs would be
desirable. Here we report efforts to confirm such intermediate
structures by in vitro incorporation of partly assembled
precursors into Hpm8.
olyketides display a wide spectrum of biological activities
Pthat make them an important resource for drug
discovery.1−4 Iterative polyketide synthases (PKSs), which
often occur in fungi, have only a single copy of each domain
(e.g., keto reduction, dehydration, and enoyl reduction) that
can be utilized repeatedly for multiple cycles of chain
elongation and tailoring of functionality (see Figure 1).2,5−7
The control of product functionality by an iterative PKS
depends on the structure of the growing chain covalently
attached to the enzyme as a thioester as well as the exact
protein sequence. Although understanding of the biosynthesis
of polyketides is advancing rapidly, knowledge of the detailed
programming by iterative PKSs is still limited.8,9 A key
requirement for understanding the mechanisms of PKS
enzymes is the determination of the structures of intermediates
that remain enzyme-bound during numerous successive steps of
elongation and modification. Chemical termination at inter-
mediate steps in conjunction with mass spectrometry (MS) has
been effective for some modular bacterial PKS systems.10−12 In
fungal systems, direct FT-ICR-MS has been used to examine
potential intermediates loaded on the PKS.13,14 However, MS
has limitations for elucidation of the stereochemistry of
intermediates and analysis of possible isomers. An alternative
approach pioneered by Cane, Hutchinson, and co-workers to
On the basis of the accepted mechanism for polyketide
biosynthesis,1 we propose that there are 14 acyl carrier protein
(ACP)-tethered intermediates en route to the hexaketide that is
eventually transferred to Hpm3 [Figure 1; also see the
Supporting Information (SI)].24 Among these 14 compounds,
four are classified as “ready” precursors because they have the
correct functionality to proceed to the next round of chain
elongation by Claisen condensation. The other 10 putative
Received: January 7, 2013
Published: January 28, 2013
© 2013 American Chemical Society
1735
dx.doi.org/10.1021/ja4001823 | J. Am. Chem. Soc. 2013, 135, 1735−1738

Journal of the American Chemical Society
Communication
Figure 2. (a) Incorporation of triketide 4 into DHZ. (b) Mass spectra
of (left) DHZ incorporating 4 and (right) the unlabeled DHZ
standard. (c) 13C NMR spectra of (top) DHZ incorporating 4 and
(bottom) the unlabeled DHZ standard.
Figure 1. Biosynthesis of hypothemycin. Hpm8 assembles a
hexaketide intermediate, which is then transferred to Hpm3. Hpm3
extends the hexaketide intermediate to DHZ. Intermediates shown in
red were synthesized as 13C-labeled SNAC thioesters.
for 15 min (preload); (2) addition of 10 μM Hpm3, 4 mM
NADPH, 0.4 mM malonyl-CoA, incubated for 1 h; and (3)
subsequent addition of 0.4 mM malonyl-CoA at 1 h intervals
(nine times) and addition of 0.2 mM labeled substrate every 2 h
(three times). Analysis of the reaction mixture by LC−MS after
incorporation of 4 generated a compound identical to DHZ in
terms of LC retention time and UV absorbance. This
compound showed a major peak in the mass spectrum at m/
z 318 ([M − H]−) (Figure 2b), indicating that it consists
primarily of a singly 13C-labeled DHZ. The incorporation ratio
was calculated to be 78% on the basis of this mass spectrum.27
To provide further confirmation of the location of the 13C
enrichment, we isolated ∼20 μg of 13C-labeled DHZ (based on
UV absorbance). The 13C NMR spectrum (Figure 2c) showed
only a single resonance at 72.8 ppm, corresponding to the C6′
position of DHZ. This demonstrates that C6′ of DHZ is 13C-
enriched, consistent with the specific incorporation of the
triketide portion of 4 and its subsequent conversion into the
final product. These data illustrate that the ketosynthase (KS)
domain of Hpm8 can be primed by trans-thioesterification with
the “ready” precursor 4, which can then be elongated to the
hexaketide moiety that is subsequently transformed by the
NRPKS Hpm3 into DHZ. This represents the first example of
intermediates are “unready” precursors because they must first
be modified by KR, DH, or ER domains prior to chain
extension.
Our idea was to synthesize of all of the “ready” precursors (2,
4, 6, and 8) and several “unready” precursors (1, 3, 5, and 7) as
13C-labeled SNAC thioesters (see Table 1) and determine
whether these compounds could be linked to Hpm8 to give
corresponding enzyme-bound intermediates that could be
elaborated to the correct hexaketide and ultimately to 13C-
labeled DHZ. We selected the “unready” precursors 1, 3, 5 and
7 because they cover the scope of possible oxidation states at
the β-position of the intermediates. The stereochemistry of the
β-hydroxyl of “unready” precursor 3 is based on the
stereoselectivity of the KR domain of Hpm8.25
The first substrate tested was “ready” triketide 4, which was
labeled with 13C at the carbonyl that is proposed to become
C6′ in DHZ (Figure 2a; for syntheses, see the SI). To avoid a
large background of unlabeled DHZ that could be formed by
initial self-loading of unlabeled malonyl-CoA,26 we designed a
“preloading” assay consisting of (1) incubation of Hpm8 (10
μM) with a large excess of labeled substrate (e.g., 4) (0.2 mM)
1736
dx.doi.org/10.1021/ja4001823 | J. Am. Chem. Soc. 2013, 135, 1735−1738

Journal of the American Chemical Society
Communication
the incorporation of a labeled advanced intermediate by a
purified iterative HRPKS.
precursor addition, ranged from 1.0 to 2.6 based on UV
absorbance. Interestingly, the more advanced the precursor, the
more DHZ was produced, possibly because advanced
precursors require fewer additional reactions to reach the
final product. Alternatively, the higher yields may also be due to
higher rates of priming (trans-thioesterification) of the more
advanced precursors onto the cysteine thiol of the KS domain.
In contrast, with exception of diketide 1, which may also
mimic malonyl-CoA, conversions of the “unready” precursors 1,
3, 5, and 7 gave relatively lower ratios of incorporation into
DHZ (41, 15, 19, and 8%, respectively) and poorer relative
RAL production (0.4, 0.6, 0.4, and 0.9, respectively). For
example, the lack of α,β-dehydration in 3 compared with 4 and
the unrealized enoyl reduction of 7 compared with 8 both
resulted in a >5-fold decrease in the incorporation of the
labeled precursor. Therefore, the Hpm8 machinery clearly
prefers precursor analogues that are structurally comparable to
the correctly tailored intermediates ready for the next round of
chain extension. The differences between the incorporation
efficiencies obtained in the “ready” and “unready” precursor
feeding experiments provide insights into the mechanisms of
the Hpm8 multidomain enzyme. For the “ready” precursors,
the subsequent enzymatic steps can be initiated immediately
after priming (trans-thioesterification) of the KS domain.28 The
carbon chains of “ready” precursors are correctly functionalized,
so they can easily be primed on KS domains and serve as the
starting material for next round of chain elongation. For
incorporation of “unready” precursors, several alternative
mechanisms can be considered. One possibility is that the
“unready” precursors load directly onto the ACP by trans-
esterification and are then further converted into the “ready”
forms in the normal fashion. However, as free ACPs are
normally loaded with malonyl-CoA by the MAT domain,
competition for the vacant ACP thiol is likely to be very
inefficient. An alternative path would involve loading on the
active-site thiol of the KS domain with subsequent trans-
esterification to the ACP domain followed by modification into
the “ready” precursors. This “KS domain to ACP domain”
transfer of incompletely tailored fragments does not occur
naturally, so this process may be very slow or impossible. The
most likely pathway may be direct action of reducing/tailoring
domains on the “unready” SNAC esters to transform them into
“ready” SNAC derivatives. In our previous studies,25 we found
that “unready” SNAC esters such as 1 and 5 can be reduced by
Hpm8 to give the “ready” precursors 2 and 6, respectively. The
KR domain can recognize the SNAC moiety, and this
recognition can correctly direct the acyl chain to the active
site for reduction. The analogous concept may apply for the
direct transformation of 7 to 8 by the ER domain or of 3 to 4
by the DH domain. The DH domain has been shown to be
inefficient at utilizing SNAC precursors,29 which may account
for the significantly lower incorporation of 3.
We next examined the incorporation of the putative
intermediates shown in Table 1: diketides 1 and 2, triketide
3, tetraketides 5 and 6, and pentaketides 7 and 8. The “ready”
precursors 2, 6, and 8 were incorporated efficiently, with 13C
enrichments of 37, 27, and 39%, respectively, based on mass
spectra (Table 1). The relative amounts of the resorcylic acid
lactone (RAL) produced, compared with that obtained with no
Table 1. Incorporation of Partially Assembled Precursor
Analogues
We also observed that the “unready” β-keto analogues 1 and
5 produced much less DHZ, perhaps because they inhibit the
natural chain extension. With the unlabeled DHZ production
inhibited, the incorporation ratio (13C enrichment) of 1 and 5
would also be relatively high, as observed.
To probe further the substrate tolerance of the KS domain of
Hpm8, the unnatural precursors 9−12 were chemically
synthesized for use in the assay to see whether they would
produce DHZ analogues. Compounds 9 and 10 have an
epimeric distal alcohol, whereas 11 lacks the conjugated double
bond. As the functions of the tailoring domains of PKS
a
b
* indicates the location of the 13C label. Structures of “ready”
c
precursors are shown in red. The relative resorcylic acid lactone
(RAL) yields include DHZ and, for unnatural precursor analogues, the
amount of the corresponding DHZ analogues. Compounds 9−12
have incorrect stereochemistry or functionality for conversion to DHZ
and (except for 12) were therefore transformed to the corresponding
DHZ analogues. Production is the amount of DHZ analogue formed,
expressed as a percentage of total RAL production.
d
e
1737
dx.doi.org/10.1021/ja4001823 | J. Am. Chem. Soc. 2013, 135, 1735−1738

Journal of the American Chemical Society
Communication
enzymes are processive and limited to the two carbons adjacent
to the thioester carbonyl,1−7 such “incorrect” functionality
cannot be modified to give the natural DHZ product. Hpm8 is
able to accept 9 or 10 to form 10′-epi-DHZ, but with much
lower efficiency (the relative ratio of epi-DHZ to DHZ formed
directly from malonate was 25:75; Table 1). This is consistent
with our in vivo observation that Hpm8 showed discrimination
toward the stereochemistry of that hydroxyl group.25
Compound 11, which represents an over-reduced precursor,
was similarly incorporated by this system to generate β-
zearalenol (the relative ratio of β-zearalenol to DHZ was 3:7).
Compound 12 is a diastereomer of the “unready” precursor 3.
Our previous work25 showed that the KR domain of Hpm8
displays strict stereospecificity toward β-keto intermediates that
is dependent on the chain length. We selected 12, which is not
an expected intermediate, to test whether the DH domain is
able to eliminate a triketide alcohol with unnatural stereo-
chemistry. The results showed that very little if any 12 was
incorporated into DHZ. This demonstrates that the (S)-
hydroxyl group is not a good substrate in the DH active site,
which is consistent with results for other PKS and FAS
systems.30 The overall decrease in RAL yield suggests that the
system is inhibited by this unnatural precursor analogue.
In summary, a series of 13C-labeled intermediate SNAC
thioesters were chemically synthesized and incorporated into
DHZ using purified Hpm8 and Hpm3. Our results show an
interesting pattern of the incorporation of partially assembled
precursor analogues at the HRPKS stage: (1) “ready”
precursors are easily recognized and taken up by Hpm8; (2)
“unready” precursors are incorporated less effectively by Hpm8,
but some incorporation is still observed, albeit with lower yields
of DHZ; (3) unnatural precursor analogues can be incorpo-
rated, but the efficiency is dependent on the nature of the
structural changes. Our findings not only further support the
processive nature of polyketide biosynthesis but also provide
guidelines for precursor-directed biosynthesis to generate novel
polyketides with improved biological profiles.
REFERENCES
■
(1) Staunton, J.; Weissman, K. J. Nat. Prod. Rep. 2001, 18, 380.
(2) Cox, R. J. Org. Biomol. Chem. 2007, 5, 2010.
(3) Li, J.-H.; Vederas, J. C. Science 2009, 325, 165.
(4) Zabala, A. O.; Cacho, R. A.; Tang, Y. J. Ind. Microbiol. Biotechnol.
2012, 39, 227.
(5) (a) Kennedy, J.; Auclair, K.; Kendrew, S. G.; Park, C.; Vederas, J.
C.; Hutchinson, C. R. Science 1999, 284, 1368. (b) Hutchinson, C. R.;
Kennedy, J.; Park, C.; Auclair, K.; Vederas, J. C. In Handbook of
Industrial Mycology; An, Z., Ed.; Marcel Dekker: New York, 2005;
Chapter 17, pp 479−492.
(6) Khosla, C.; Tang, Y.; Chen, A. Y.; Schnarr, N. A.; Cane, D. E.
Annu. Rev. Biochem. 2007, 76, 195.
(7) Crawford, J. M.; Townsend, C. A. Nat. Rev. Microbiol. 2010, 8,
879.
(8) Ma, S. M.; Li, J. W.-H.; Choi, J. W.; Zhou, H.; Lee, M. K. K.;
Moorthie, V. A.; Xie, X.; Kealey, J. T.; Da Silva, N. A.; Vederas, J. C.;
Tang, Y. Science 2009, 326, 589.
(9) Fisch, K. M.; Bakeer, W.; Yakasai, A. A.; Song, Z.; Pedrick, J.;
Wasil, Z.; Bailey, A. M.; Lazarus, C. M.; Simpson, T. J.; Cox, R. J. J.
Am. Chem. Soc. 2011, 133, 16635.
(10) Tosin, M.; Spiteller, D.; Spencer, J. B. ChemBioChem 2009, 10,
1714.
(11) Tosin, M.; Betancor, L.; Stephens, E.; Li, W. M. A.; Spencer, J.
B.; Leadlay, P. F. ChemBioChem 2010, 11, 539.
(12) Tosin, M.; Demydchuk, Y.; Parascandolo, J. S.; Per, C. B.;
Leeper, F. J.; Leadlay, P. F. Chem. Commun. 2011, 47, 3460.
(13) Meehan, M. J.; Xie, X.; Zhao, X.; Xu, W.; Tang, Y.; Dorrestein,
P. C. Biochemistry 2011, 50, 287.
(14) Vagstad, A. L.; Bumpus, S. B.; Belecki, K.; Kelleher, N. L.;
Townsend, C. A. J. Am. Chem. Soc. 2012, 134, 6865.
(15) (a) Cane, D. E.; Yang, C. J. Am. Chem. Soc. 1987, 109, 1255.
(b) Yue, S.; Duncan, J. S.; Yamamoto, Y.; Hutchinson, C. R. J. Am.
Chem. Soc. 1987, 109, 1253.
(16) Cane, D. E.; Lambalot, R. H.; Prabhakaran, P. C.; Ott, W. R. J.
Am. Chem. Soc. 1993, 115, 522.
(17) (a) Cane, D. E.; Tan, W. T.; Ott, W. R. J. Am. Chem. Soc. 1993,
115, 527. (b) Cane, D. E.; Luo, G. L. J. Am. Chem. Soc. 1995, 117,
6633.
(18) Kim, C.-Y.; Alekseyev, V. Y.; Chen, A. Y.; Tang, Y.; Cane, D. E.;
Khosla, C. Biochemistry 2004, 43, 13892.
(19) Yuzawa, S.; Kapur, S.; Cane, D. E.; Khosla, C. Biochemistry 2012,
51, 3708 and references therein.
ASSOCIATED CONTENT
■
(20) (a) Yoshizawa, Y.; Li, Z.; Reese, P. B.; Vederas, J. C. J. Am.
Chem. Soc. 1990, 112, 3212. (b) Li, Z.; Martin, F. M.; Vederas, J. C. J.
Am. Chem. Soc. 1992, 114, 1531. (c) Harrison, P. H.; Noguchi, H.;
Vederas, J. C. J. Am. Chem. Soc. 1986, 108, 3833. (d) Liu, Y.; Li, Z.;
Vederas, J. C. Tetrahedron 1998, 54, 15937.
S
* Supporting Information
Synthetic procedures for labeled precursors and experimental
details for incorporations by Hpm8 and Hpm3. This material is
available free of charge via the Internet at http://pubs.acs.org.
(21) Brobst, S.; Townsend, C. A. Can. J. Chem. 1994, 72, 200.
(22) Tsantrizos, Y. S.; Zhou, F.; Famili, P.; Yang, X. J. Org. Chem.
1995, 60, 6922.
(23) Reeves, C. D.; Hu, Z. H.; Reid, R.; Kealey, J. T. Appl. Environ.
Microbiol. 2008, 74, 5121.
(24) Zhou, H.; Qiao, K.; Gao, Z.; Meehan, M. J.; Li, J. W.-H.; Zhao,
X.; Dorrestein, P. C.; Vederas, J. C.; Tang, Y. J. Am. Chem. Soc. 2010,
132, 4530.
AUTHOR INFORMATION
■
Corresponding Author
yitang@ucla.edu; john.vederas@ualberta.ca
Author Contributions
∥Z.G. and J.W. contributed equally.
(25) Zhou, H.; Gao, Z.; Qiao, K.; Wang, J.; Vederas, J. C.; Tang, Y.
Nat. Chem. Biol. 2012, 8, 331.
(26) Ma, S. M.; Tang, Y. FEBS J. 2007, 274, 2854.
(27) Biemann, K. Mass Spectrometry: Organic Chemical Applications;
McGraw-Hill: New York, 1962.
(28) Tsukamoto, N.; Chuck, J. A.; Luo, G.; Kao, C. M.; Khosla, C.;
Cane, D. E. Biochemistry 1996, 35, 15244.
(29) Valenzano, C. R.; You, Y. O.; Garg, A.; Keatinge-Clay, A.;
Khosla, C.; Cane, D. E. J. Am. Chem. Soc. 2010, 132, 14697.
(30) Kwan, D. H.; Schulz, F. Molecules 2011, 16, 6092.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Mark Miskolzie for assistance with NMR experi-
ments and the Natural Sciences and Engineering Research
Council of Canada, the Canada Research Chair in Bioorganic
and Medicinal Chemistry, and the U.S. National Institutes of
Health (1R01GM085128 and 1DP1GM106413) for financial
support. Dedicated to Prof. Christoph Tamm (Univ. Basel) on
the occasion of his 90th birthday.
1738
dx.doi.org/10.1021/ja4001823 | J. Am. Chem. Soc. 2013, 135, 1735−1738