Chemoenzymatic route to macrocyclic hybrid peptide/polyketide-like molecules

Article abstract of DOI:10.1021/ja0352202

Hybrid peptide-polyketides are a class of medically and biologically important natural products characterized by stereochemical and functional diversity. In their biosynthesis, hybrids are often macrocyclized to achieve rigid structures that populate bioa

Full text of DOI:10.1021/ja0352202

Published on Web 05/20/2003
Chemoenzymatic Route to Macrocyclic Hybrid Peptide/Polyketide-like
Molecules
Rahul M. Kohli,† Martin D. Burke,‡ Junhua Tao,† and Christopher T. Walsh*,†
Department of Biological Chemistry and Molecular Pharmacology, HarVard Medical School,
Boston, Massachusetts 02115, and Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
Received March 19, 2003; E-mail: christopher_walsh@hms.harvard.edu
A variety of natural products with therapeutic activities are
enabled to populate biologically favorable conformers by macro-
cyclization.1 As with many polyketides (PK) and nonribosomal
peptides (NRP), macrocyclic hybrid PK/NRP natural products, such
as the immunosupressant rapamycin and the anticancer agent
epothilone, have the ability to interact with biological targets or
modulate protein-protein interactions.2 In the biosynthesis of these
ring-containing natural products, linear chains constructed by a
similar logic are subjected to essentially identical macrocyclization
machinery, a thioesterase (TE) domain at the C-terminal end of
the multimodular enzymatic assembly lines.1,2a,3 We have demon-
strated that the TE domains excised from the NRPS assembly lines
for tyrocidine 1, gramicidin, and surfactin, as well as the epothilone
hybrid synthetase, retain macrocyclization activity when excised
and studied as autonomous catalytic domains.4 We have shown that
Figure 1. Macrocyclic natural products and chemoenzymatically synthe-
the excised tyrocidine TE domain, TycC TE, is permissive for
alterations in most side chains or the length of a peptidyl thioester
substrate.4,5 Further, the TE will recognize peptidyl oxoesters
attached to PEGA beads, so that solid-phase peptide synthesis
(SPPS) can be used in a chemoenzymatic approach to production
of libraries of cyclic peptides, although previously the functional
diversity of PK and hybrid molecules remained inaccessible by this
method.6
sized hybrid-like products. Polyketide epitopes are shown in red, with bonds
formed on macrocyclization highlighted by shading.
Scheme 1a
We present here a chemoenzymatic strategy to access the
stereochemical and functional diversity of macrocyclic PK/NRP
natural products in a manner amenable to efficient library synthesis.
Specifically, we have made use of small building blocks presenting
embedded PK units in the form of ꢀ-amino acids that allow for the
construction of linear hybrid peptide/polyketide-like chains through
iterative amide bond formation. Further, to increase the potential
for attaining the biologically active structures seen in many
macrocyclic natural products, we demonstrate the use of the NRPS-
derived enzyme, TycC TE, to effect macrocyclization of these
linear, hybrid-like substrates.
In PK biosynthesis, following chain elongation by the formation
of tethered â-keto-acyl intermediates by ketosynthase domains,
coordinated action of ketoreductase domains within the same PKS
module typically convert the initial 3-keto condensation product
to the 3-OH form. A subsequent iteration would then yield a 3,5-
diol with 2,4-dimethyl, monomethyl, or methylene groups, depend-
ing on the selection of methylmalonyl or malonyl building blocks.
The erythromycin skeleton contains such four-carbon signature
epitopes, for example, in the part of structure 2 spanning C2-C5
(Figure 1), where glycosylation of the 5-OH in this system is crucial
for the gain of antibiotic activity.7 Using a laboratory analogue of
this biosynthetic pathway, developed by Evans and co-workers for
a Conditions: (a) TiCl4, DIPEA, N-Fmoc-glycinal, then H2O2-MeOH,
81%, dr 7.5:1; (b) Zn(BH4)2; (c) 2,2-dimethoxypropane, TsOH, 61% over
two steps; (d) Me4N(OAc)3BH; (e) 2,2-dimethoxypropane, PPTS, 42% over
two steps; (f) LiOOH, THF-H2O, 60-61%.
the efficient construction of polypropionate systems,8 we developed
a four-step route to ꢀ-amino acids compatible with SPPS with PK
functionality spanning Câ-Cꢀ (Scheme 1A). Specifically, condensa-
tion of the Z trichlorotitanium enolate derived from the â-keto imide
5 with Fmoc-glycinal9 provided the all-syn aldol adduct 6.10
Subsequent syn or anti diastereoselective ketone reduction, diol
ketalization, and lithium hydroperoxide-mediated hydrolysis of the
chiral auxiliary afforded the desired Fmoc-protected ꢀ-amino acids
8a and 8b.11 Modifications of this synthetic scheme may provide
access to significant diversity in the PK building block, in the form
of other diastereomers, altered oxidation states, and non-hydrogen
substituents at CR.
† Department of Biological Chemistry and Molecular Pharmacology, Harvard
Medical School.
‡ Department of Chemistry and Chemical Biology, Harvard University.
9
7160
J. AM. CHEM. SOC. 2003, 125, 7160-7161
10.1021/ja0352202 CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
macrocycle, we next carried out tandem insertion of our ꢀ-amino
acid building blocks in all four possible permutations, yielding 11a-
d, replacing residues 5-8 in tyrocidine A. Remarkably, upon
incubation with TycC TE, the hybrid-like macrocycle was the
predominant product observed in each reaction.12 Macrocycles
4a-d contain eight stereogenic carbons in an eleven-atom stretch.
That all four combinations were cyclized with similar efficiency
suggests our building block approach should allow for rapid
introduction of multiple PK epitopes into a diversified library.
Further, that PK functionality can be introduced into a macrocycle
via an NRPS-derived TE domain suggests that TE tolerance could
have facilitated the evolution of hybrid synthetases in nature.
We have demonstrated a method for accessing the diversity of
PK/NRP natural products via a remarkably versatile macrocycliza-
tion catalyst and a building-block approach to introducing diversity.
It remains to be seen whether any of these polyketide epitopes will
suffice on their own for binding to biological targets of polyketides.
It will be intriguing to investigate if the advantages endowed to
hybrid natural products in their functional diversity, improved cell
permeability, or modification by tailoring enzymes (as in the
glycosylation of erythromycin C5-hydroxyl) will translate to our
hybrid-like molecules. We anticipate the synthesis of diverse natural
product-like libraries tailored for such particular assays or for
screening for novel biological activities.
Figure 2. HPLC traces showing chemoenzymatic synthesis products. (A)
Soluble thioesters 9a and 9b are converted to cyclic products 3a and 3b
with minor flux to hydrolysis product (denoted by *). (B) Enzymatic reaction
products released into solution upon incubation of substrates tethered to
the solid phase. Single (3a and 3b) and tandem (4a-d) insertion products
are observed. Products were confirmed by MS/MS fragmentation. Minor
products eluting from 13 to 15 min are hydrolysis product (*) and truncated
products from SPPS. Peaks eluting at 20 min correspond to enzyme, with
variable amounts of catalyst observed.
Acknowledgment. Funding was provided by NIH GM20011
(C.T.W.), the Medical Scientist Training Program (R.M.K.), and
the HHMI (M.D.B.). M.D.B. gratefully acknowledges S. L.
Schreiber for materials and helpful discussions and M. Movassaghi
for helpful discussions.
It was unclear at the outset if a TE domain derived from a strictly
NRP synthetase would be capable of macrocyclizing a hybrid-like
substrate. As a first test of the ability of TE domains to cyclize
peptide backbones with polyketide substructures, we constructed
peptide chains by SPPS containing the building blocks 8a and 8b,
which were subsequently converted to soluble peptide thioester
substrate 9a and 9b (Scheme 1B). The building blocks were viewed
as replacements for a dipeptide in 1, the natural tyrocidine A peptide
scaffold, and targeted to a region known from prior studies to be
tolerant to alteration.5b Both diastereomeric substrates were smoothly
macrocyclized upon reaction with TycC TE to generate the
macrolactams 3a and 3b, with minor flux to hydrolysis (Figure
2A).12 Mass spectroscopy fragmentation confirmed the expected
head-to-tail cyclization linkage. Cyclization kinetics of the soluble
substrates could be measured (9a, kcat 37 min-1, KM 68 µM; 9b,
36 min-1, 77 µM), revealing that catalytic efficiency is only
modestly decreased with these substrates relative to that of the
natural substrate (60 min-1, 3 µM).4a
Using ꢀ-amino acid building blocks to introduce polyketide units
into polypeptide structures offers potential for synthesizing diverse
libraries of hybrid PK/NRP-like macrocycles. As a demonstration
of the feasibility of this aim, we next turned to substrates tethered
to a solid-phase resin (Scheme 1B). This scheme is made to
resemble the physiologic situation as linear substrates are con-
structed while tethered to PEGA resin via a linker mimicking
pantetheine. TycC TE subsequently catalyzes both release of the
substrate from its oxoester linkage to the solid-phase as well as
macrocyclization.6 To verify our methodology, the linear solid-
phase substrates 10a and 10b corresponding to soluble 9a and 9b
were synthesized. Incubation with the purified TE domain resulted
in the generation of cyclization products 3a and 3b (Figure 2B).
To generate diversity and increase the PK functionality in the
Supporting Information Available: Experimental procedures and
characterization of compounds (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
References
(1) Kohli, R. M.; Walsh, C. T. Chem. Commun. 2003, 297-307.
(2) (a) Du, L.; Sanchez, C.; Shen, B. Metab. Eng. 2001, 3, 78-95. (b)
Schreiber, S. L. Science 1991, 251, 283-287. (c) Nicolaou, K. C.; Ritzen,
A.; Namoto, K. Chem. Commun. 2001, 1523-1535.
(3) (a) Cane, D. E.; Walsh, C. T. Chem. Biol. 1999, 6, R319-25. (b) Cane,
D. E.; Walsh, C. T.; Khosla, C. Science 1998, 282, 63-68. (c) Marahiel,
M. A.; Stachelhaus, T.; Mootz, H. D. Chem. ReV. 1997, 97, 2651-2674.
(4) (a) Trauger, J. W.; Kohli, R. M.; Mootz, H. D.; Marahiel, M. A.; Walsh,
C. T. Nature 2000, 407, 215-218. (b) Kohli, R. M.; Trauger, J. W.;
Schwarzer, D.; Marahiel, M. A.; Walsh, C. T. Biochemistry 2001, 40,
7099-7108. (c) Boddy, C. N.; Schneider, T. L.; Hotta, K.; Walsh, C. T.;
Khosla, C. J. Am. Chem. Soc. 2003, 125, 3428-3429.
(5) (a) Kohli, R. M.; Takagi, J.; Walsh, C. T. Proc. Natl. Acad. Sci. U.S.A.
2002, 99, 1247-1252. (b) Trauger, J. W.; Kohli, R. M.; Walsh, C. T.
Biochemistry 2001, 40, 7092-7098.
(6) Kohli, R. M.; Walsh, C. T.; Burkart, M. D. Nature 2002, 418, 658-661.
(7) Hansen, J. L.; Ippolito, J. A.; Ban, N.; Nissen, P.; Moore, P. B.; Steitz, T.
A. Mol. Cell 2002, 10, 117-128.
(8) Evans, D. A.; Clark, J. S.; Metternich, R.; Novack, V. J.; Sheppard, G. S.
J. Am. Chem. Soc. 1990, 112, 866-868.
(9) Myers, A. G.; Zhong, B.; Movassaghi, M.; Kund, D. W.; Lanman, B. A.;
Kwon, S. Tetrahedron Lett. 2000, 41, 1359.
(10) To confirm this stereochemical assignment, 6 was subjected to stereose-
lective ketone reduction followed by δ-lactonization, and TBDPS protec-
tion. The lactone was characterized by NOESY 1H NMR analysis. See
Supporting Information.
(11) In the corresponding acetonides, 1,3-syn- and 1,3-anti-diol relationships
were confirmed via diagnostic 13C NMR resonances. (a) Rychnovsky, S.
D.; Skalitzky, D. Tetrahedron Lett. 1990, 31, 945-948. (b) Evans, D.
A.; Rieger, D. L.; Gage, J. R. Tetrahedron Lett. 1990, 31, 7099-7100.
(12) Products of reactions with soluble substrates and solid-phase substrates
all yield cyclization-to-hydrolysis ratios >1. See Supporting Information.
JA0352202
9
J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003 7161