A cyclopentane conformational restraint for a peptide nucleic acid: Design, asymmetric synthesis, and improved binding affinity to DNA and RNA

Article abstract of DOI:10.1021/ol0348811

(Matrix presented) A strategy to restrict the highly flexible backbone conformation of a peptide nucleic acid (PNA) by incorporation of a cyclopentane ring is proposed. An asymmetric synthesis of cyclopentane-modified PNA is reported, and its binding prop

Full text of DOI:10.1021/ol0348811

ORGANIC
LETTERS
2003
Vol. 5, No. 15
2695-2698
A Cyclopentane Conformational
Restraint for a Peptide Nucleic Acid:
Design, Asymmetric Synthesis, and
Improved Binding Affinity to DNA and
RNA
Michael C. Myers, Mark A. Witschi, Nataliya V. Larionova, John M. Franck,
Russell D. Haynes, Toshiaki Hara,† Andrzej Grajkowski,‡ and Daniel H. Appella*
Department of Chemistry, Northwestern UniVersity, EVanston, Illinois 60208
dappella@chem.northwestern.edu
Received May 20, 2003
ABSTRACT
A strategy to restrict the highly flexible backbone conformation of a peptide nucleic acid (PNA) by incorporation of a cyclopentane ring is
proposed. An asymmetric synthesis of cyclopentane-modified PNA is reported, and its binding properties were determined. The cyclopentane
ring leads to a significant improvement in the binding properties of the resulting PNA to DNA and RNA.
The high binding affinity of peptide nucleic acids to
complementary DNA and RNA has spurred development of
numerous biochemical, biomedical, and medicinal applica-
tions for this class of molecules.1 The most well-known class
of peptide nucleic acids (PNAs) consists of nucleic acid bases
attached to an achiral peptide backbone that is made up of
N-(2-aminoethyl)glycine units (Figure 1).2 These oligonucle-
otide mimics bind sequence specifically to DNA and RNA
with higher affinity than complementary oligonucleotides.
Structures of PNA-DNA and PNA-RNA duplexes have
been well characterized, in addition to PNA2DNA triplexes
consisting of two polypyrimidine PNAs bound to one
polypurine DNA.3 Numerous backbone modifications to
PNAs have been explored with the intention of preorganizing
the highly flexible PNA backbone into the requisite confor-
mations necessary for binding to DNA and RNA.4 Proper
preorganization could significantly increase the PNA binding
affinity for oligonucleotides. Such increases in binding
affinity could lower the detection limits for PNA-based
diagnostic applications and improve the antisense properties
for this class of molecules. Currently, there are relatively
few modifications that have successfully yielded PNAs with
† Laboratory of Cell Biology, National Cancer Institute, National Institutes
of Health, 37 Convent Dr., Bethesda, MD 20892.
‡ Division of Therapeutic Proteins, Center for Biologics Evaluation and
Research, Food and Drug Administration, 8800 Rockville Pike, Bethesda,
MD 20892.
(1) (a) Nielsen, P. E. Curr. Opin. Biotechnol. 2001, 12, 16. (b) Doyle,
D. F.; Braasch, D. A.; Simmons, C. G.; Janowski, B. A.; Corey, D. A.
Biochemistry 2001, 40, 53. (c) Kuhn, H.; Demidov, V. V.; Coull, J. M.;
Fiandaca, M. J.; Gildea, B. D.; Frank-Kamenetskii, M. D. J. Am. Chem.
Soc. 2002, 124, 1097. (d) Okamoto, A.; Tanabe, K.; Saito, I. J. Am. Chem.
Soc. 2002, 124, 10262.
Figure 1. Structure of a peptide nucleic acid.
(2) Nielsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O. Science 1991,
254, 1497.
10.1021/ol0348811 CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/03/2003

improved binding affinity over the original, unmodified
backbone. The most successful modifications described to
date involve incorporating proline and derivatives of 4-hy-
droxyproline into a PNA backbone.4 We describe in this
communication a novel conformational restraint for a PNA
backbone that leads to increased binding affinity to comple-
mentary DNA and RNA.
Yamamoto’s diasteroselective alkylation of dimenthylsuc-
cinate (1) would be the most direct route.9 We successfully
scaled up (by 3 times) Yamamoto’s asymmetric alkylation
of (-)-dimenthylsuccinate with 1,3-propanediol ditosylate
to a 70 g scale (Scheme 1). Purification of the resulting
Scheme 1. Asymmetric Alkylation, Removal of Chiral
A simple strategy for rigidifying a PNA is to incorporate
a cyclic ring into the C2-C3 carbon-carbon bond of the
PNA backbone (Figure 1). Nielsen and co-workers initially
examined this strategy by incorporating a cyclohexane ring
at this position.5 The facile synthesis of these PNAs started
from the well-known and commercially available trans-1,2-
diaminocyclohexane (available as a single enantiomer in
either the (R,R) or (S,S) form). The results of this study
demonstrated that the PNAs derived from the (R,R) enanti-
omer bound very weakly to DNA and RNA, while the (S,S)-
cyclohexyl PNAs bound with slightly weaker affinity
compared to the unmodified PNA.
Auxiliaries, and Curtius Rearrangement
Analysis of the preferred PNA backbone dihedral angles
when bound to DNA or RNA, in addition to molecular
modeling studies (using molecular mechanics calculations
with the MM3 force field), indicated to us that the cyclo-
hexane ring does not possess the optimal dihedral angles to
promote PNA binding to DNA or RNA.6 From NMR
structures of PNA-DNA and PNA-RNA duplexes, the
preferred dihedral angle about C2-C3 is 130-165° and 60-
80°, respectively.3 Our molecular modeling studies indicate
that the minimum energy conformation of a trans-diequatorial
cyclohexane ring is about 50-65°. The conformational
constraints of the cyclohexane ring prevent this dihedral angle
from attaining values that are considerably outside this range.
Modeling studies of a corresponding cyclopentane indicated
that this ring would be better suited to adopt the requisite
C2-C3 dihedral angles in PNAs. Energy-minimized con-
formations of a trans-diequatorial cyclopentane ring pos-
sessed dihedral angles of 70-90°. In addition, the broad
potential energy well suggested that the cyclopentane could
adopt dihedral angles up to 160°. Based on these predictions,
we embarked on a synthesis of (S,S)-cyclopentanediamine
to determine if it was a suitable conformational restraint for
a PNA.
cyclopentane dimenthyl ester was enhanced by quenching
the unreacted dianion of dimenthylsuccinate with tereph-
thaldicarboxaldehyde, resulting in easily separable impurities.
In contrast to the reported dr of 24:1, we routinely obtained
2 as a 9:1 mixture of diastereomers (based on GC analysis).
The (1S,2S)-cyclopentane diastereomer 2 was highly crystal-
line and was separated from the (1R,2R)-cyclopentane
diastereomer by recrystallization. The material that was
obtained in this fashion had >99% de (based on GC
analysis), and the absolute stereochemistry at the substituted
cyclopentane carbons was confirmed as (S,S) on the basis
of a crystal structure of 2.
Removal of the (-)-menthyl chiral auxiliaries was prob-
lematic because 2 was resistant to hydrolysis using Yama-
moto’s alkaline conditions that were published for the
hydrolysis of the corresponding cyclopropane dimenthyl ester
(10% KOH in 9:1 MeOH/H2O, 60 °C).9b Acidic conditions
were also unsuccessful at hydrolyzing the menthyl esters of
2, as were reactions using peroxy anion. These findings are
in line with related hydrolyses of 1,2-disubstituted cyclo-
hexane menthyl esters where aqueous tetrabutylammonium
hydroxide (TBAH) gave low yields of mono- and dicar-
boxylic acids.10 Under more vigorous conditions (1.5 N KOH
in 5:1 MeOH/H2O, reflux) hydrolysis occurred, but the
product obtained was partially racemized based on a chiral
HPLC analysis of a derivative of the isolated diacid (details
concerning evaluation of ee are given in the Supporting
Information).
To test the effects of cyclopentane modification in a PNA,
we required multigram quantities of enantiomerically pure
(S,S)-trans-1,2-cyclopentanediamine. Only a few syntheses
of this diamine have appeared in the literature, and most rely
on resolutions of racemic trans-1,2-cyclopentanediamine to
obtain enantiomerically enriched material.7 An enzymatic
resolution of this diamine has recently been published,8 but
in our work, we felt that an asymmetric synthesis based on
Since 2 could not be hydrolyzed without racemization,
alternate conditions were developed to remove the chiral
(3) (a) Brown, S. C.; Thomson, S. A.; Veal, J. M.; Davis, D. G. Science
1994, 265, 777. (b) Betts, L.; Josey, J. A.; Veal, J. M.; Jordan, S. R. Science
1995, 270, 1838. (c) Eriksson, M.; Nielsen, P. E. Nature Struct. Biol. 1996,
3, 410.
(4) (a) Kumar, V. Eur. J. Org. Chem. 2002, 2021 and references within.
(b) Vilaivan, T.; Lowe, G. J. Am. Chem. Soc. 2002, 124, 9326. (c) Slaitas,
A.; Yeheskiely, E. Eur. J. Org. Chem. 2002, 2391.
(5) Lagriffoule, P.; Wittung, P.; Eriksson, M.; Jensen, K. K.; Norde´n,
B.; Buchardt, O.; Nielsen, P. E. Chem. Eur. J. 1997, 3, 912.
(6) See the Supporting Information for more details.
(7) (a) Jaeger, V. F. M.; Blumendal, H. B. Zeit. Anorgan. All. Chem.
1928, 161. (b) Toftland, H.; Pedersen, E. Acta Chim. Scand. 1972, 26, 4019.
(c) Goto, M.; Takeshita, M.; Sakai, T. Bull. Chem. Soc. Jpn. 1979, 52,
2589.
(8) Luna, A.; Alfonso, I.; Gotor, V. Org. Lett. 2002, 4, 3627.
(9) (a) Misumi, A.; Iwanaga, K.; Furuta, K.; Tamamoto, H. J. Am. Chem.
Soc. 1985, 107, 3343. (b) Furuta, K.; Iwanaga, K.; Yamamoto, H. Org.
Synth. 1989, 67, 76.
2696
Org. Lett., Vol. 5, No. 15, 2003

Figure 2. PNAs to explore effects of cyclopentane.
auxiliaries. Reaction with LiAlH4 reduced the menthyl esters
to a diol that was subsequently oxidized with catalytic
TEMPO, NaOCl2, and NaOCl to give 3.11 Overall yields for
this two-step reduction-oxidation were good, and no ad-
ditional purification steps were necessary. Chiral HPLC
analysis showed an ee >99% for 3. With enantiomerically
pure 1,2-dicarboxylic acid in hand, the next steps focused
on installing the nitrogens onto the cyclopentane ring utilizing
a double Curtius rearrangement. After rearrangement to a
bis-isocyanate, reaction with neat tert-butyl alcohol afforded
di-tert-butyl carbamate 4 as a stable solid.12
mentation, it was found that a solution of trifluoroacetic acid
in anhydrous ethyl ether would effect a monodeprotection
to give 5 in modest yields. Fortunately, 60% of unreacted
starting material 4 could be recovered after the reaction by
an acid-base aqueous workup and recycled for use in
subsequent mono-deprotections. Repeated monodeprotections
gave sufficient quantities of 5 to make the necessary PNA
monomer.
The final steps in the synthesis of the cyclopentane PNA
monomer involved attaching a methylenecarboxy group and
a thymine base to the amine. Reaction between monoamine
5 and ethylglyoxalate afforded an intermediate imine that
was immediately hydrogenated with a palladium catalyst to
produce 6. Alternative alkylation procedures using methyl
bromoacetate routinely gave dialkylated material. The reduc-
tive amination was advantageous because no dialkylated
byproduct was formed.
To make a PNA monomer, the next step focused on
differentiating the two nitrogens of 4 so that a single tert-
butyl carbamate protecting group was present (Scheme 2).
Scheme 2. Monodeprotection and Synthesis of Cyclopentane
PNA Monomer
Combination of thymine carboxylic acid with 6 and
HATU15 formed the fully protected cyclopentyl-PNA mono-
mer 7. Hydrolysis of the ethyl ester cleanly afforded 8, which
was used directly in the solid-phase synthesis of PNAs. We
have used this synthetic route to produce 8, in eight steps
with a 4% overall yield.
Using standard solid-phase peptide synthesis procedures,16
two PNA heptamers were made to test the ability of the
cyclopentane ring to promote PNA binding to DNA and
RNA (Figure 2). In each PNA, a lysine was present as the
first residue and the N-terminal was left unprotected in order
to promote aqueous solubility and prevent aggregation.
The effects of cyclopentane substitution in PNAs were
determined by examining the melting temperature for each
PNA bound to a hepta-adenine oligonucleotide. PNAs 9 and
10 were compared for their ability to form stable PNA2DNA
triplexes.17 The effects on DNA binding are dramatic. A
This type of monoprotected diamine is frequently a synthetic
intermediate in the synthesis of PNA monomers.13 In the
synthesis of a regular PNA monomer, monoprotected eth-
ylenediamine is obtained by combining a 5:1 ratio of
ethylenediamine to di-tert-butyl dicarbonate (Boc2O).14 The
excess ethylenediamine is not recovered. While this proce-
dure is acceptable for ethylenediamine, it was extremely low
yielding when applied to cyclopentane diamine. Furthermore,
the extensive loss of diamine under these conditions forced
us to devise a more efficient method for obtaining a
monoprotected cyclopentane diamine. After much experi-
(10) Hasegawa, T.; Yamamoto, H. Synlett 1999, 1, 84.
(11) Zhao, M.; Li, J.; Mano, E.; Song, Z.; Tschaen, D. M.; Grabowski,
E. J.; Reider, P. J. J. Org. Chem. 1999, 64 2564.
(12) Ongeri, S.; Aitken, D. J.; Husson, H. P. Synth. Commun. 2000, 30,
2593.
(13) Breipohl, G.; Will, D. W.; Peyman, A.; Uhlmann, E. Tetrahedron
1997, 53, 14671.
(14) Krapcho, A. P.; Kuell, C. S. Synth. Commun. 1990, 20, 2559.
(15) Carpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397.
(16) Egholm, M.; Casale, R. A. The Chemistry of Peptide Nucleic Acids.
In Solid-Phase Synthesis, a Practical Guide; Kates, S. A., Albericio, F.,
Eds.; Marcel Dekker: New York, 2000; Chapter 13.
Org. Lett., Vol. 5, No. 15, 2003
2697

single cyclopentane residue in the middle of PNA 10
increases the melting temperature by almost 6 °C relative to
control PNA 9 (Table 1). In the PNA-RNA duplex, the
in order to incorporate cyclopentane into a PNA. The
cyclopentane modification improves the stability of PNA-
DNA triplexes and PNA-RNA duplexes for a poly-T PNA.
We are currently examining the incorporation of cyclopen-
tane into additional PNA sequences to determine whether
this modification is compatible with other bases, and we are
examining whether multiple cyclopentanes will further
stabilize duplexes and triplexes. We are also exploring the
use of the cyclopentane ring as a scaffold for attachment of
functional groups that will result in new PNAs with useful
chemical and biological properties.
Table 1. Tm Data for PNA/Oligonucleotide Complexesa
PNA
Tm PNA2DNA
Tm PNA/RNA
9
10
44.4
50.3
48.3
51.4
a All samples were prepared in 10 mM phosphate buffer (pH 7) with
100 mM NaCl. Each strand concentration is 5 µM for RNA and 7 µM for
DNA. Each melting temperature is the average of a forward and backward
run from 20 to 80 °C. Data were obtained using d(A)7 and r(A)7 as the
oligonucleotides.
Acknowledgment. We gratefully acknowledge financial
support from Northwestern University’s Weinberg College
of Arts and Sciences, Department of Chemistry, VP of
Research, Institute of Biotechnology and Nanoscience in
Medicine, and an American Cancer Society Institutional
Research Grant. We also would like to acknowledge Dr.
Serge Beaucage and Dr. Andrzej Wilk for helpful comments.
cyclopentane ring also confers increased melting tempera-
tures to the duplex.
In conclusion, the cyclopentane ring can be a useful
conformational restraint for the C2-C3 dihedral angle of
the PNA backbone. Molecular modeling and computational
analyses were useful guides for designing this conformational
restraint. An asymmetric synthesis of cyclopentane diamine,
and a monodeprotection were important synthetic develop-
ments that allowed access to enough cyclopentane monomer
Supporting Information Available: Molecular modeling
information for cyclohexane vs cyclopentane PNA mono-
mers, all synthetic procedures for preparation of cyclopentane
PNA monomers, procedure for determining % ee of 3,
MALDI mass spectra for PNAs 9 and 10, and melting curves
for all duplexes and triplexes. This material is available free
of charge via the Internet at http://pubs.acs.org.
(17) Kim, S. K.; Nielsen, P. E.; Egholm, M.; Buchardt, O.; Berg, R. H.;
Norde´n, B. J. Am. Chem. Soc. 1993, 115, 6477.
OL0348811
2698
Org. Lett., Vol. 5, No. 15, 2003