Novel antibiotics: macrocyclic peptides designed to trap Holliday junctions.

Article abstract of DOI:10.1021/ol020204f

[reaction: see text] Described are the syntheses of eight macrocyclic peptides designed to trap Holliday junctions in bacteria, thereby inhibiting bacterial growth. These macrocycles were designed from linear dimerized hexapeptides that bind to the C-2 sy

Full text of DOI:10.1021/ol020204f

ORGANIC
LETTERS
2003
Vol. 5, No. 2
109-112
Novel Antibiotics: Macrocyclic Peptides
Designed to Trap Holliday Junctions
Megan L. Bolla, Enrique V. Azevedo, Jason M. Smith, Rachel E. Taylor,
Dev K. Ranjit,† Anca M. Segall,† and Shelli R. McAlpine*
Department of Chemistry, Molecular Biology Institute, and Center for Applied and
Experimental Genomics, 5500 Campanile DriVe, 208 CSL, San Diego State UniVersity,
San Diego, California 92182-1030
mcalpine@chemistry.sdsu.edu
Received October 10, 2002
ABSTRACT
Described are the syntheses of eight macrocyclic peptides designed to trap Holliday junctions in bacteria, thereby inhibiting bacterial growth.
These macrocycles were designed from linear dimerized hexapeptides that bind to the C-2 symmetrical Holliday junction. They were synthesized
from three monomers using a combinatorial-like strategy that permits elucidation of the monomer role in accumulation of Holliday junctions
and antibiotic activity. These macrocycles are an important step in designing and synthesizing a new class of antibiotics.
Although peptides rarely function well as drugs due to their
low bioavailability and rapid degradation within cells,1 they
make convenient initial synthetic targets due to their ease
of assembly and modification. Using peptides as initial leads
allows rapid identification of the structural requirements of
an active biological inhibitor. There are a large number of
natural product cyclic peptides that have interesting biological
activity.2,3 Often the conversion of these active peptides into
peptidomimetics has been the most successful approach for
making new biologically active compounds.4
Antibiotic resistance is an extreme public health concern.5
As more pathogenic bacteria become resistant to first- and
second-line antibiotics, easily treatable infectious diseases
are becoming life-threatening infections. To keep up with
the evolutionary pressure from pathogenic bacteria during
the infectious disease process we must continually develop
new antibiotics and inhibit new biological targets in bacteria.
One such new target is the Holliday junction (HJ) intermedi-
ate generated by the XerC/D site-specific recombinase.6 Site-
† Department of Biology.
(1) Hirschmann, R.; Smith, A. B.; Sprengeler, P. A. New PerspectiVes
in Drug Design; Dean, M., Newton, C. G., Eds.; Academic Press, Inc.:
New York, 1995; p 1. (b) West, M. L.; Fairlie, P. D. Trends Pharm. Sci.
1995, 16, 67.
(4) (a) Ward, D. E.; Gai, Y.; Lazny, R.; Pedras, M. S. C. J. Org. Chem.
2001, 66, 7832. (b) Caba, J. M.; Rodriguez, I. M.; Manzanares, I.; Giralt,
E.; Albericio, F. J. Org. Chem. 2001, 66, 7568. (c) Didemnin, B.; Tarver,
J.; Pfizenmayer, A. J.; Joullie´, M. M. J. Org. Chem. 2001, 66, 7575. (d)
Bursavich, M. G.; West, C. W.; Rich, D. H. Org. Lett. 2001, 3, 2317.
(5) (a) Neu, H. C. Science 1992, 257, 1064. (b) Kaufman, M.; Friday,
M. Worries Rise Over Effect of Antibiotics in Animal Feed. Washington
Post, March 17th, 2000, p A1. (c) Shortridge, V. D.; Doern, G. V.;
Brueggemann, A. B.; Beyer, J. M.; Flamm, R. K. Clin. Infect. Dis. 1999,
29, 1186-1188.
(2) (a) For a comprehensive discussion on peptide synthesis, see: Lloyd-
Williams, P.; Albericio, F.; Giralt, E. Chemical Approaches to the Synthesis
of Peptides and Proteins; CRC Press: Boca Raton, FL, 1997. (b) Albericio,
F.; Carpino, L. A. In Methods in Enzymology; Fields, G. B., Ed.; Academic
Press: New York, 1997; pp 104-126. (c) For a review on biologically
active peptides, see: Izumiya, N.; Kato, T.; Aoyagi, H.; Waki, M.; Kondo,
M. Synthetic Aspects of Giologically Important ActiVe Cyclic Peptides;
Halsted Press: New York, 1979. (d) For a review of peptide synthesis,
see: Humphrey, J. M.; Chamberlin, A. R. Chem. ReV. 1997, 97, 2243.
(3) Bodanszky, M.; Bodanszky, A. The Practice of Peptide Synthesis,
2nd ed.; Springer: New York, 1994.
(6) Blakely, G.; May, G.; McCulloch, R.; Arciszewska, L. K.; Burke,
M.; Lovett, S. T.; Sherratt, D. J. Cell 1993, 75, 351-361.
10.1021/ol020204f CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/01/2003

specific recombination can control gene expression, amplify
episome copy number, create genetic diversity, and separate
chromosomes at bacterial cell division. By trapping the HJ
intermediate, we may inhibit site-specific recombination.7
Blocking this recombination reaction should eventually lead
to bacterial death. Thus, compounds that successfully trap
the HJ will lead to new antibiotics. Such inhibitors are
reminiscent of the quinolone/fluoroquinolone class of anti-
biotics, which stabilize a normally transient intermediate.
Herein we describe the synthesis of eight potential
antibiotics. We based our design on the cocrystal structure8
of the hexapeptide that dimerizes to give the symmetrical
active lead structure9 and the C-2 symmetrical Holliday
junction (Figure 1). The previously identified peptide leads
fit the approximate size of the HJ binding site, which is
estimated to be ∼25 Å by 10 Å.8 Because symmetrical linear
peptides are known to trap the HJ intermediate, we designed
C-2 symmetric cyclic peptides that mimic the symmetry of
the HJ binding site and contain residues found in these lead
compounds (Figure 2).8,9
Figure 2. Synthesis strategy.
Our synthetic approach was chosen to simplify the
synthesis of C-2 symmetrical macrocyclic hexapeptides for
rapid biological assessment. By cyclizing the peptides, we
aimed to increase their rigidity10 so that we could identify
the exact nature of contacts between the compounds and the
HJ. At the same time this should decrease their degradation
rate inside cells.
The minimal number of residues involved in binding to
the Holliday junction is not known, but several important
interactions have been identified.9 Hydrophobic residues such
as tryptophan and phenylalanine are known to be important
for binding to the DNA because they appear in all of the
lead compounds9 and are thought to stack with the DNA
nucleotides that surround the center of the HJ. Hydrophilic
residues such as arginine and lysine may form hydrogen
bonds either with the protein that assembles the HJ or with
the DNA substrate itself, or both. Starting from commercially
available natural and unnatural amino acids, we have
synthesized cyclic hexameric peptides. Initially, we chose
to use a number of hydrophobic residues in the macrocyclic
peptides and sequentially exchange hydrophobic residues
with hydrophilic residues.
Figure 1. Cocrystal structure was obtained by mixing a mutant
Cre protein with loxS DNA substrates. The crystal was dependent
on blocking catalysis.8 A similar cocrystal structure was obtained
between wild-type Cre protein and lox substrates, but only in the
presence of the linear lead peptides. This latter crystal contains
additional electron density in the HJ center, consistent with bound
peptide in the central “hole”.8 Note the C-2 symmetry of the
structure. One lead linear hexapeptide structure is Lys-Trp-Trp-
Cys-Arg-Trp, where the active peptide is a dimer of this linear
peptide.
are linear dodecapeptides, and their flexibility presumably
prevents accurate pinpointing of the specific residues in-
volved in the binding event. However, electron density
attributed to the peptides binding to the HJ is detected in
the center of the crystal structure.
With the goal of gaining insight into the biological
mechanism of action and finding new antibiotics for this
unique target, we describe the synthesis of macrocycles that
were designed from the C-2 symmetric HJ binding site and
Segall’s discovery of lead compounds.9 These macrocycles
should be more rigid than the linear lead peptides, and they
A great deal of literature has documented the various
advantages and disadvantages associated with peptide chem-
istry.3 The eight compounds described represent an initial
set of potential antibiotics that lay the groundwork for our
(7) Nash, H. A. Cellular and Molecular Biology, 2nd ed.; Neidhardt, F.
H. C., Ed.; ASM Press: Washington, DC, 1996; pp 2363-2376.
(8) (a) Ghosh, K.; Cassell, G. C.; Segall, A.; Van Duyne, G. D.
Unpublished results with the dodecapeptide binding in the center of the
C-2 symmetric Holliday junction site. (b) Gopaul, D. N.; Guo, F.; Duyne,
G. D. V. EMBO J. 1998, 17, 4175-4187.
(9) (a)Cassell, G.; Klemm, M.; Pinilla, C.; Segall, A. M. J. Mol. Bio.
2000, 299, 1193-1202. (b) Klemm, M.; Cheng, C.; Cassell, G.; Shuman,
S.; Segall, A. M. J. Mol. Bio. 2000, 299, 1203-1216.
(10) (a) Tyndall, J. D. A.; Fairlie, D. P. Curr. Med. Chem. 2001, 8, 893.
(b) Fairlie, D. P.; Tyndall, J. D. A.; Reid, R. C.; Wong, A. K.; Abbenante,
G.; Scanlon, M. J.; March, D. R.; Bergmann, D. A.; Chai, C. L. L.; Burkett,
B. A. J. Med. Chem. 2000, 43, 1271.
110
Org. Lett., Vol. 5, No. 2, 2003

future combinatorial approach. This approach provides the
framework from which small, targeted libraries of peptides
will be developed, with the eventual goal of converting active
peptides into peptidomimetics.
Using O-(7-azabenzatrizol-1-yl)-N,N,N′,N′-tetramethyl-
uronium hexafluorophosphate (HATU), Hunig’s base, and
(dimethylamino)pyridine (DMAP) as coupling reagents,
L-phenylalanine methyl ester and N-Boc-protected residue
2 (Figure 3) were coupled to give the dipeptide 1-2-Boc (80-
Figure 4. Synthesis of macrocycles. Reagents and conditions: (a)
HATU (1.2 equiv; TBTU, PyBOP, and/or DEPBT used as coupling
agents), DMAP (0.2 equiv), Hunig’s base (3 equiv), CH2Cl2; (b)
TFA (20%), CH2Cl2, anisole (2 equiv); (c) Ba(OH)2 (4 equiv),
MeOH. An asterisk (*) indicates a linear hexamer precursor: 8
examples, yields 31-94%.
Figure 3. Synthesis of Fragment 1. Reagents and conditions: (a)
HATU (1.2 equiv), DMAP (0.2 equiv), Hunig’s base (3 equiv),
CH2Cl2; (b) TFA (20%), CH2Cl2, anisole (2 equiv).
the yield. The one-pot ring-closing yields varied from 21 to
48%.13 The peptides were then purified using reverse-phase
HPLC and confirmed using LCMS and high-resolution mass
spectroscopy.13
These eight macrocyclic compounds have been tested for
their ability to bind to the Holliday junction in an in vitro
assay (Figure 5). This assay involves a recombination
reaction between one radiolabeled double-stranded DNA
94% yield). Deprotection of the amine on residue 2 using
20% TFA and 2 equiv of anisole in methylene chloride gave
the free amine 1-2 (∼quantitative yields). Coupling of this
dipeptide to monomer 3a or 3b gave the desired tripeptide
(Fragment 1) in high yields (65-94%).11
Fragment 1 is then separated into two equal aliquots
(Figure 4). The acid is deprotected in the first aliquot using
4 equiv of barium hydroxide, while the amine is deprotected
in the second aliquot using TFA/anisole. These two trimer
peptides are coupled using multiple coupling agents to give
the linear hexapeptides.
The linear hexapeptides were cyclized by deprotecting the
acid using 4 equiv of barium hydroxide. Upon workup of
the free acid, we subjected the compounds to 20% TFA in
methylene chloride with 2 equiv of anisole. Following
deprotection of the free amine, we subjected the crude, dry
product to HATU, TBTU, PyBop, and/or DEPBT coupling
reagents (1.2 equiv each), DMAP (0.5 equiv) and Hunig’s
base (3 equiv) in methylene chloride.12 The final macro-
cyclizations took approximately 4 days due to the low
concentration (0.005-0.01 M) that was required to maximize
(11) All dipeptide and tripeptide structures were confirmed using 1H
NMR. All linear hexameric peptides were confirmed using LCMS and 1H
NMR, and cyclized peptides were all confirmed using LCMS, 1H NMR,
and high-resolution mass spectrometry. See Supporting Information for
spectra.
(12) R. K. Guy (UCSF) communicated his experience where the use of
at least three of these coupling reagents facilitates efficient ring-closing
reactions by providing a choice of reagents for the specific substrate.
Figure 5. In vitro assay results. (A) Holliday junction intermediate;
(B) recombination products; (C) free DNA.
Org. Lett., Vol. 5, No. 2, 2003
111

substrate and an unlabeled partner DNA substrate. Successful
recombination gives recombination products and very small
amounts of HJ intermediate (lane 2). A previously identified
lead dimer of Lys-Trp-Trp-Cys-Arg-Trp is used as a positive
control because it is known to accumulate HJs8 (lane 3).
During the recombination reaction, two initial DNA cleavage
and ligation reactions must take place to form the junction,
and two subsequent DNA cleavage and ligation reactions
resolve the HJs into double-stranded recombination products.7
At 0.01 mg/ml concentrations of peptides (Figure 5), a
significant amount of HJ intermediates (A) accumulate in
the presence of four of these macrocyclic hexamers (1-2a-
3a, 1-2a-3b, 1-2a-3a, and 1-2c-3b seen in lanes 5-8) when
compared to a control reaction not treated with peptide (lane
2). These four compounds appear to be as potent as the lead
peptide (lane 3) in trapping the HJ intermediate. Tryptophan
and phenylalanine are residues within the previously known
peptide leads,8 and it is thought that they stack with the DNA
bases in the HJ.7 The four structures that contain phenyl-
alanine coupled to the 1,2,3,4-tetrahydroisoquinoline-3-
carboxylic acid moiety (lanes 5-8) appear to stabilize the
HJ better than the other four compounds (lanes 4 and 9-11).
This binding effect may be the result of the aromatic group
in the tetrahydroisoquinoline π-stacking with DNA bases.
The relatively low binding of the compounds containing
glycine or the piperdine moiety in position 2 (lanes 9-11)
may be explained by their inability to π-stack with the DNA
bases. The flexibility of the amino octanoic chain may
explain the lack of binding for the phenyl and indole moieties
present in macrocyclic hexamer 1-2b-3b (lane 4).
ously known lead peptides contain hydrophilic elements not
present in the cyclic hexamers described here, we plan to
incorporate these into the next generation of macrocyclic
peptides. The syntheses of these new compounds are
underway.
In summary, we describe a unique approach to building a
potential new class of antibiotics. Our straightforward
synthetic route provides novel macrocyclic compounds that
stabilize the Holliday junction. The in vitro assays of these
compounds suggest that this is a viable approach to designing
a new class of antibiotics.
Acknowledgment. We thank Pfizer, La Jolla, for equip-
ment and financial donations as well as their fellowships to
M.L.B. (summer 2002) and E.V.A. (2001-2003). We thank
San Diego State University for financial contributions. We
thank Dr. Indrawan McAlpine at Pfizer, La Jolla, for helpful
discussions of this work. We also thank Professor Rodney
Guy and Ms. Sarah Galicia at UCSF for helpful discussions
and the use of their equipment. We thank Dr. Greg Roth at
Boehringer-Ingelheim for helpful discussions and Boeh-
ringer-Ingelheim Pharmaceuticals for financial support. We
thank Ms. Anne Warren for her contribution of synthetic
material. Anca Segall and Dev Ranjit were supported by NIH
R01 GM52847 to A.M.S.
Supporting Information Available: General experimen-
1
tal procedures, H NMR available for dipeptide, tripeptide,
and linear hexapeptide intermediates, and high-resolution
mass spectra and two wavelength HPLC traces (224 nm and
254 nm; given as evidence of the purity of final macrocyclic
peptides). This material is available free of charge via the
Internet at http://pubs.acs.org.
The results from the in vitro assay indicate that our design
was an effective initial effort in synthesizing cyclic com-
pounds that trap the Holliday junction. Because the previ-
(13) R. K. Guy (UCSF) and S. Galicia (UCSF) kindly allowed us to use
their preparatory HPLC system for purification of cyclized product.
OL020204F
112
Org. Lett., Vol. 5, No. 2, 2003