Synthesis and RNA-Binding Properties of Extended Nucleobases for Triplex-Forming Peptide Nucleic Acids

Article abstract of DOI:10.1021/acs.joc.9b01133

Triple-helix formation, using Hoogsteen hydrogen bonding of triplex-forming oligonucleotides, represents an attractive method for sequence-specific recognition of double-stranded nucleic acids. However, practical applications using triple-helix-forming oligonucleotides and their analogues are limited to long homopurine sequences. The key problem for recognition of pyrimidines is that they present only one hydrogen-bond acceptor or donor group in the major groove. Herein, we report our first attempt to overcome this problem by using peptide nucleic acids (PNAs) modified with extended nucleobases that form three hydrogen bonds along the entire Hoogsteen edge of the Watson-Crick base pair. New nucleobase triples (five) were designed, and their hydrogen bonding feasibility was confirmed by ab initio calculations. PNA monomers carrying the modified nucleobases were synthesized and incorporated in short model PNA sequences. Isothermal titration calorimetry showed that these nucleobases had a modest binding affinity for their double-stranded RNA (dsRNA) targets. Finally, molecular modeling of the modified triples in PNA-dsRNA helix suggested that the modest binding affinity was caused by subtle structural deviations from ideal hydrogen-bonding arrangements or disrupted π-stacking of the extended nucleobase scaffolds.

Full text of DOI:10.1021/acs.joc.9b01133

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE
Article
Synthesis and RNA binding properties of extended
nucleobases for triplex-forming peptide nucleic acids
Ilze Kumpina, Nikita Brodyagin, James Arthur MacKay,
Scott D Kennedy, Martins Katkevics, and Eriks Rozners
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.9b01133 • Publication Date (Web): 20 Sep 2019
Downloaded from pubs.acs.org on September 21, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Synthesis and RNA binding properties of extended nucleobases for triplex-
forming peptide nucleic acids
1
2
3
4
Ilze Kumpina, Nikita Brodyagin, James A. MacKay, Scott D. Kennedy, Martins
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2,
Katkevics, Eriks Rozners *
1Latvian Institute of Organic Synthesis, Aizkraukles 21, Riga, LV-1006, Latvia
2Department of Chemistry, Binghamton University, State University of New York,
Binghamton, NY 13902, USA
3Department of Chemistry and Biochemistry, Elizabethtown College, Elizabethtown, PA
1
7022, USA
4Department of Biochemistry and Biophysics, University of Rochester School of
Medicine and Dentistry, Rochester, NY 14642, USA
* e-mail erozners@binghamton.edu
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Abstract: Triple helix formation, using Hoogsteen hydrogen bonding of triplex-forming
oligonucleotides, represents an attractive method for sequence-specific recognition of
double-stranded nucleic acids. However, practical applications using triple helix-forming
oligonucleotides and their analogues are limited to long homopurine sequences. The
key problem for recognition of pyrimidines is that they present only one hydrogen bond
acceptor or donor group in the major groove. Herein, we report our first attempt to
overcome this problem by using peptide nucleic acids (PNAs) modified with extended
nucleobases that form three hydrogen bonds along the entire Hoogsteen edge of the
Watson-Crick base pair. Five new nucleobase triples were designed and their hydrogen
bonding feasibility was confirmed by ab initio calculations. PNA monomers carrying the
modified nucleobases were synthesized and incorporated in short model PNA
sequences. Isothermal titration calorimetry (ITC) showed that these nucleobases had
modest binding affinity for their double-stranded RNA (dsRNA) targets. Finally,
molecular modeling of the modified triples in PNA-dsRNA helix suggested that the
modest binding affinity was caused by subtle structural deviations from ideal hydrogen
bonding arrangements or disrupted pi-stacking of the extended nucleobase scaffolds.
ACS Paragon Plus Environment

Page 3 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
1
Triple helices are naturally occurring structures in RNA. Additionally, triple
helical binding of an oligonucleotide in the major groove of double-stranded RNA
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
dsRNA) may be used for sequence-selective molecular recognition of folded RNAs.1, 2
However, triple helices formed by negatively charged oligonucleotides are destabilized
3
by electrostatic repulsion. The Rozners group were the first to use peptide nucleic acid
(
PNA),4, 5 a DNA analogue built on neutral amide backbone (Figure 1), to overcome the
unfavorable electrostatic repulsion in triple helical recognition of dsRNA. Other groups
6
have also explored thermodynamics and kinetics of PNA-dsRNA triplex formation and
7
developed chemical modifications to improve PNA binding to dsRNA. Recently, the
Rozners group discovered that PNAs with cytosines (C) replaced by 2-aminipyridine (M)
nucleobases formed exceptionally stable and sequence-selective triple helices with
dsRNA under physiologically relevant conditions.8-10 The key innovative feature was the
+
+
increased basicity of M (pK of M ~6.5) compared to C (pK of C ~4.5) that enabled
a
a
formation of protonated M*G-C triple (Figure 1) at physiological pH. This discovery
opened the door for using PNA-dsRNA triplexes for developing novel fluorescent RNA
11
12
detection methods and modulating RNA function in biological systems. Taken
together, these studies identified M-modified PNA as a promising ligand for molecular
recognition of complex dsRNAs.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Structure of PNA and Hoogsteen base triples, U*A-U, C*G-C and M*G-C.
A remaining general problem for triple helical recognition of nucleic acids is that
the common Hoogsteen hydrogen-bonded triples (e.g., U*A-U and C*G-C or M*G-C,
Figure 1) limit the sequences that can be targeted to motifs where one of the strands of
double helix is formed by purine nucleosides only. Despite extensive effort in developing
modified nucleobases that would recognize pyrimidines in U-A (or T-A in DNA) and C-G
base pairs, triple helix formation is still limited to relatively long homopurine sequences
that may be interrupted by only a few pyrimidines.13-17 The problem of designing high
stability triples for recognition of U (or T) and C is that they present only one hydrogen
bond acceptor or donor group, respectively, in the major groove. To overcome this
problem, we envisioned designing extended nucleobases that would form three
hydrogen bonds along the entire Hoogsteen edge of the Watson-Crick base pair.
The idea of extending the Hoogsteen hydrogen bonding over both nucleosides of
a Watson-Crick base pair has been entertained since the early days of triple-helical
ACS Paragon Plus Environment

Page 5 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
18
DNA recognition. Dervan and co-workers pioneered the first design, where a
nucleobase analogue (D , Figure 2) spanned the major grove to form hydrogen bonds
3
with both C and G of the inverted C-G base pair. Other groups followed up with similar
designs.19-22 These early efforts shared common problems of reduced affinity
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
compared to the native triples) and lack of discrimination between the pyrimidines of C-
14
G and T-A base pairs. NMR structures later showed that D did not form the expected
3
hydrogen bonds, but instead intercalated in the double helix and stacked over the
pyrimidine-purine base pairs.22-24 Other studies25-27 used NMR in organic solvent to
evaluate hydrogen bonding potential and design novel extended nucleobases, e.g.,
ureido isoindolin-1-one (Ind, Figure 2). However, strong hydrogen bonding in organic
solvent did not necessarily translate into formation of strong and selective DNA triple
helices suggesting that other factors, such as, constrains on backbone conformation,
base triple isomorphism and nucleobase stacking, played subtle but important roles in
overall efficiency of triple helical binding.27, 28
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Selected examples of nucleobases that hydrogen bond to entire Hoogsteen
edge of Watson-Crick base pairs. In specific studies, the nucleobases were attached to
either native DNA, or various backbone modified analogues (e.g., LNA, PNA, etc).
4
In more recent studies, cytosine derivatives N -(2-guanidoethyl)-5-methylcytosine
29
30
Q, 4-[(3R,4R)-dihydroxypyrrolidino]-pyrimidin-2-one X and the related guanidine
31
32
derivative GP (Figure 2), and N-(4-(3-acetamidophenyl)thiazol-2-yl)acetamide (S,
Figure 2) showed promising binding to C-G and T-A base pairs, respectively. Overall,
while several nucleobases have given encouraging preliminary results, current designs
are far from optimal and typically suffer from either modest binding affinity (e.g., Q) or
modest sequence selectivity (e.g., S). In attempted triple helix formation involving all
four base pairs, Fox and co-workers33, 34 used S for T-A recognition, while Ohkubo et
ACS Paragon Plus Environment

Page 7 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
35
DA
al. used N-acetyl-2,7-diamino-1,8-naphtyridine( N , Figure 2) and Q for T-A and C-
Ac
36
G recognition, respectively. Chen and co-workers used Q-modified PNAs to form a
triple helix on a seven purine long stretch of dsRNA interrupted by one C, and showed
that the triplex formation stimulated ribosomal frameshifting in a cell-free translation
assay.37 Despite the promising results of these studies, a general solution to
recognition of any sequence of dsDNA or dsRNA is still elusive and design of novel
nucleobases remains an active area of research.17, 38
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The present manuscript describes our first efforts in designing extended
heterocyclic systems that would recognize the entire Hoogsteen edge of Watson-Crick
base pairs in the major groove of dsRNA. Electronic and geometric considerations
discussed above influenced our strategy for designing extended nucleobases. While we
used ab initio calculations of nucleobase triples as simple models for initial assessment
of the hydrogen bonding feasibility, the ultimate test was the experimentally determined
affinity of the modified PNAs for their matched target dsRNA hairpins.
We report syntheses of five new scaffolds of extended nucleobases (M1 through
M4 and T1, Figure 3). Isothermal titration calorimetry (ITC) showed that these
nucleobases had modest binding affinity for their dsRNA targets. Finally, molecular
modeling of the modified triples in a PNA-dsRNA helix provided additional insights into
structural reasons behind the observed affinities suggesting that the loss of binding
affinity compared to the canonical Hoogsteen triples was due to subtle structural
deviations from ideal hydrogen bonding arrangements or disrupted pi-stacking of the
extended nucleobase scaffolds.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Extended nucleobases and proposed hydrogen bonding schemes for
recognition of the entire Hoogsteen edge of Watson-Crick base pairs.
ACS Paragon Plus Environment

Page 9 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
Overall design. Our design of extended nucleobases (Figure 3) was to extend M
and T using an appropriate linker that would position a single hydrogen bond acceptor
or donor across the major grove for additional interaction with C and U, respectively. We
anticipated the identification of an ideal scaffold that could ultimately be used for either
G-C or C-G recognition depending on the attachment point to PNA backbone (c.f.,
M2*G-C and M2R*C-G in Figure 3). The idea was that a successful extended
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
nucleobase would retain and improve the excellent binding properties of M and T
observed in our previous studies.8-10
Toward this end, we first envisioned nucleobase M1, linking the M scaffold to a
pyridine through a simple ethylene linker (Figure 3). To test the geometrical viability of
our design, we used docking of the extended nucleobases on the G-C or A-U base pairs
followed by ab initio methods (HF-631G(d)) to confirm the intended H-bonding
geometries and conformational preferences of the triples. Models of the M1 nucleobase
containing a methyl group in the location of attachment to the PNA backbone were built
in silico. We performed a geometry optimization on the G-C base pair, docked it with a
geometry optimized M1 nucleobase, and finally re-ran the geometry optimization on the
base triple. We were pleased to find that the structure of the energy minimized triple
showed H-bonding distances in the typical range and agreement with our expectations
(Figure 4 M1). Furthermore, the minimized geometry of the M1 triple suggested that
planarity across the entire extended system was attainable and the triple energy was
lower than the sum of the M1 nucleobase and G-C energies. These considerations
together suggested that this strategy may prove useful in identifying privileged scaffolds.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Molecular modeling of triples (M1) M1*G-C geometry optimized using HF-
6
31g(d) and (M2-T1) views of initial docking of extended nucleobases on G-C and U-A
base pairs.
A similar strategy was used to identify scaffolds M2-M4 and T1. In each of these
examples the docking of the nucleobase with its respective Watson-Crick base pair
resulted in H-bonding geometries within expected distances of 2.5 Å or better (Figure
4). However, for each of these scaffolds, geometry optimized base triples led to
unexpected deviations of planarity of either the nucleobase itself (M2-M4) or the
nucleobase coming out of the plane of the Watson-Crick base pair (T1) (see the
Supporting Information, Figure SL for details). Returning to the docking geometries in
Figure 4, molecular energies for these arrangements predicted favorable interactions in
ACS Paragon Plus Environment

Page 11 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
all cases. We reasoned that the ultimate test would be in the experimental binding
affinity of the PNA for dsRNA given difficult to predict electronic effects such as H-
bonding in aqueous solution, the added complexity of the PNA triplex, and pi-stacking
interactions. Thus, we set out to prepare these extended nucleobases with the aim of
evaluating their binding properties in PNA-dsRNA triple helix.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of M1 PNA monomer. The synthesis of the M1 PNA monomer
started with nucleophilic aromatic substitution of commercially available 6-
chloronicotinonitrile 1 with 2-(pyridin-2-yl)ethan-1-amine 2 (Scheme 1). Protection of the
secondary amine with di-tert-butyl dicarbonate followed by saponification of the nitrile
gave the carboxylic acid 4. Arndt-Eistert homologation of 4 using thionyl chloride to
make the requisite acyl chloride gave a complex mixture that did not contain the desired
product; however, oxalyl chloride in the presence of catalytic amount of DMF yielded the
acyl chloride that was in situ converted to diazoketone 5 using TMS-diazomethane.
Completion of Arndt-Eistert homologation was achieved with concomitant esterification
with ethanol. The esterification was preferred over a more direct route to carboxylic
acid, because the ester was much easier to isolate and purify. Saponification of the
ethyl ester 6 gave the carboxylic acid, which was coupled to the PNA backbone 7 to
give after deallylation the final M1 PNA monomer 9.39
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 1. Synthesis of M1 PNA monomer
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of M2 PNA monomer. Nucleobases M2, M2R and MeOM2R were variations
of M1 where the one of the methylene groups was replaced by a carbonyl group. The
key consideration in their design was the synthetic accessibility of amide linkers that
would allow easier optimization of extended nucleobases. A convergent synthesis
(Scheme 2) started with coupling of halogenated heterocycles with malonate
40
derivatives. The copper-catalyzed coupling of 2-iodopyrimidine with diethylmalonate
gave a better yield of 10 than nucleophilic aromatic substitution used for the synthesis of
12. Basic hydrolysis of 10 afforded upon acidification 2-(pyrimidin-2-yl) acetic acid 11.
Using a similar approach, ethyl aminopyridyl acetate 13 was made by a nucleophilic
aromatic substitution reaction of 4-bromo-2-nitropyridine and tert-butyl ethyl malonate,
with subsequent cleavage of tert-butyl group, decarboxylation and reduction of the nitro
function. HATU mediated coupling of 11 with aminopyridyl derivative 13, followed by
ACS Paragon Plus Environment

Page 13 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
hydrolysis furnished the acetic acid derivative 15, which was converted to M2 PNA
monomer 17 using the same chemistry as in Scheme 1.
Scheme 2. Synthesis of M2 PNA monomer
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of M2R and MeOM2R. The synthesis of orthogonally protected
pyrimidyl diester 19, the key intermediate for acids 22a and 22b, started from 5-bromo-
3-chloropyrimidine (Scheme 3). The first acetate moiety was introduced following an
41
exact literature precedent, while for the second we adopted a variant of Reformatsky-
42
Negishi reaction. Attempts to use TFA in CH Cl for removal of the tert-butyl group in
2
2
19 resulted in decomposition of 20 after evaporation of the reaction mixture even at
4
3
room temperature. On the other hand, ZnBr mediated cleavage of the tert-butyl group
2
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
afforded pyrimidylacetic acid 20 in good yield. HATU mediated coupling of 20 with 2-
aminopyridine or 2-amino-4-methoxypyridine, followed by basic hydrolysis of ethyl
esters provided the acetic acid derivatives 21a and 21b, which were converted to M2R
and MeOM2R PNA monomers 24a and 24b, respectively, using the same chemistry as in
Schemes 1 and 2.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Synthesis of M2R and MeOM2R PNA monomers
ACS Paragon Plus Environment

Page 15 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Synthesis of M3. Inspired by the promising binding properties reported for Q29,
3
6, 44
(
Figure 2) and our own experience with M (Figure 1), we designed M3 and M3R
where the guanidine substituent of Q was replaced by the structurally isosteric M
Figure 3). Synthesis of M3 (Scheme 4) started with reductive alkylation of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
36
aminopyridine 13 followed by coupling with triazolyl uracil 26. After standard protection
and deprotection steps the carboxylic acid precursor 29 was coupled to the benzyl
45
protected PNA backbone 30. Hydrogenation gave the final M3 PNA monomer 32.
Scheme 4. Synthesis of M3 PNA monomer
H
CH3
N
N
CH3
N
CH3
N
O
O
N
O
N
N
NH
N
N
3
POCl , NEt3
N
N
MeCN/DCM (1:1)
O
7
1%
N
Boc
N
Boc
N
25
26
+
o
1. NEt , K CO , Py, 60 C
LiOH
3
2
3
2
. Boc O, DMAP, DCM
EtOH, H O
2
H N
2
3
41%
80%
Cl
NH2
NH
O
NHBoc
H
CO Et
N
2
N
28
1. AcOH, 3 Å mol. sieves
Cl
2
. NaBH CN, MeOH
3
3
. HCl, Et O
2
CO Et
2
CO Et
2
54%, two steps
13
27
CH3
N
CH3
O
N
N
O
N
CH3
N
N
N
N
O
N
N
Boc
Boc
Boc
Boc
N
N
Boc
30
H
Boc
N
N
CO Bn
2
N
Pd/C, H (1 atm)
N
FmocHN
2
MeOH
7%
DCC, HOOBt, DMF
76%
6
O
O
O
O
N
N
FmocHN
OBn
FmocHN
OH
CO H
2
31
32
29
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Synthesis of M3R. Synthesis of M3R (Scheme 5) followed a similar route as for
46
M3 (Scheme 4) starting from uracil ethyl acetate 33. Coupling of the triazolyl derivative
47
34 with amine 35 gave, after protection group manipulations, the acetic acid derivative
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
7, which was converted to M3R PNA monomer 39, using the same chemistry as in
Scheme 4.
Scheme 5. Synthesis of M3R PNA monomer
N
N
N
N
N
N
N
N
HN
N
Boc
O
N
3
Boc
N
N
N
N
35
Boc
Boc
N
N
NH
H
NH2
N
N
LiOH
EtOH, H O
N
1
2
. K CO , MeCN
. Boc O, DMAP, THF
2
O
POCl , NEt
3
MeCN DCM
81%
3
2
3
96%
2
N
O
O
CO Et
O
2
42% two steps
CO Et
2
COOH
3
34
36
CO Et
37
2
N
N
N
O
N
Boc
Boc
Boc
N
Boc
3
0
N
N
H
N
CO Bn
2
N
N
Pd/C, H (1 atm)
FmocHN
2
MeOH
88%
DCC, HOOBt, DMF
N
O
7
3%
O
O
O
O
N
N
FmocHN
OBn
FmocHN
OH
38
39
Synthesis of M4. In parallel to nucleobases connected with flexible alkyl linkers,
we pursued designs using heterocyclic linkers, such as oxazole in M4 (Figure 3). The
initial attempt at M4 used, as a precursor of oxazole formation, amide 42a, which was
prepared by successive couplings of 4-bromo-2-aminopyridine (40a) with Boc-glycine
ACS Paragon Plus Environment

Page 17 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
and, after deprotection, pyrimidine-2-carboxylic acid (Scheme 6). While trifluoroacetic
48
acid anhydride did not produce the desired ring closure reaction, treatment of 42a with
49
triphenylphosphinium iodide gave the target oxazole that precipitated from the reaction
mixture in a reasonably pure form. However, neither the previously used Cu catalyzed
arylation of malonates (Scheme 2), nor the Negishi reaction (Scheme 3), nor the Heck
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
50
reaction with tert-butyl((1-methoxyvinyl)oxy)dimethylsilane gave the desired ester 43b.
An alternative approach starting from 13 that had the acetic acid residue in place was
lower yielding but successful in providing the key precursor 44. In this case, oxazole did
not precipitate from the reaction mixture and could be purified only as 43b after
installation of the Boc group. Basic hydrolysis gave the carboxylic acid 44, which was
51
coupled with the unprotected version of PNA backbone 45 to give the M4 PNA
monomer 46. This one step procedure is preferable for compounds with limited solubility
that are difficult to separate from residual Ph PO after removal of the allyl protecting
3
group. However, the yields tend to be moderate due to an incomplete coupling step.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 6. Synthesis of M4 PNA monomer
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of T1. For recognition of A-U base pairs, we explored linking
additional functionality to U using the well-established substitution chemistry at the 5-
position. The T1 nucleobase (Figure 3) was envisioned as the product of a copper-
catalyzed click reaction between 3-azidophenol and a 5-alkynl uracil derivative. In our
initial synthetic studies, 5-ethynyl uracil was alkylated at N-1 with a bromoacetate ester.
Unfortunately, the desired product was accompanied by significant inseparable
dialkylation (N-1 and N-3) products. The copper catalyzed cycloaddition with 3-azido
phenol afforded the desired nucleobase, however, it was found that the phenolic residue
was problematic in the coupling to the PNA backbone. In a revised route (Scheme 7),
we chose to perform amide coupling to the PNA backbone prior to the key cycloaddition
step. This would in turn also allow for a late stage cycloaddition with various azide
ACS Paragon Plus Environment

Page 19 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
coupling partners. To this end, 5-iodouracil was treated with methylbromoacetate to
afford 47 in 75% yield. Notably, dialkylation was still observed (~5-10%), but a single
recrystallization removed the undesired impurity. Sonigashira coupling afforded alkynyl
uracil 48 in excellent yield, and treatment of 48 with aqueous sodium hydroxide
accomplished both hydrolysis of the methyl ester and concomitant removal of the TMS
group to afford 49. Notably, 49 could be easily prepared on a gram scale from readily
available starting materials. Finally coupling of 49 to the PNA backbone 7 occurred
uneventfully using HOBt and EDC.52
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 7. Synthesis of T1 and Tr PNA monomers
O
N
TMS
O
N
O
N
TMSA
O
I
I
PdCl (PPh )
2
3 2
Br
HN
NH
NH
NH
MeO
K2CO3, DMF
5%
CuI, NEt3
NaOH
O
N
H
O
O
EtOAc
O
65 °C
91%
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
91%
7
COOMe
O
COOMe
COOH
47
48
49
NH
N
O
O
HOBt, EDC, DMF
67%
49 + 7
O
N
FmocHN
O
50
Ot-Bu
t-Bu
O
CuSO (0.4 eq)
Na-ascorbate (2 eq)
4
51
O
54
THF/EtOH/H O
2
N3
N3
85%
H
77%
Ot-Bu
t-Bu
O
N
N
O
H (0.5 eq)
N
N
N
O
O
N
N
N
NH
N
NH
N
O
O
O
O
O
O
N
N
FmocHN
OR
FmocHN
Pd(PPh3)4
N-ethyl aniline
THF
OR
Pd(PPh3)4
N-ethyl aniline
THF
5
5
2 R = allyl
3 R = H
55 R = allyl
56 R = H
77%
91%
To avoid the previously encountered problems with the phenolic functional group
at the stage of PNA synthesis, protection as a t-butyl ether was proposed. While other
protecting groups may be more common on phenols, we chose t-butyl due to its general
stability and because it would be easily removed using TFA upon cleavage of the final
PNA from the solid support resin. To this end, attachment of the t-butyl group was
ACS Paragon Plus Environment

Page 21 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
accomplished using the method of Bartoli involving either Sc(OTf) or Mg(OCl ) as a
3
4 2
53
catalyst and di-tert-butyl dicarbonate as the t-butyl source. Butylation of 3-azido
phenol afforded similar yields of 51 using either catalyst, so the less expensive
Mg(ClO ) catalyst was chosen for larger scale reactions.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
2
With both the azide (51) and alkyne (50) cycloaddition partners available, the
[
3+2] cycloaddition step was explored. Initial studies utilized copper sulfate/sodium
ascorbate as the catalyst system in a 2:1 THF/water mixed solvent. To our delight, a
44% yield of the desired triazole 52 was obtained. Using N,N’-dimethylethylene diamine
as an additive increased the yields further, and in an effort to better dissolve the
reactants a 1:1:0.4 mixture of THF/EtOH/H O was employed. Under these optimized
2
conditions, 85% yield of triazole 52 resulted, which was converted to the T1 monomer
5
3 in 77% yield under standard deallylation conditions.39
An additional monomer, Tr, was made as a control to determine if the triazole
affects binding of the uracil derivative to A-U. The pivaloylmethyl (POM) group was
chosen to protect the triazole as it is known to be stable to the Fmoc chemistry used in
PNA synthesis and could be readily removed with aqueous ammonia from PNA to form
54
the free triazole. Tr monomer (56) was prepared by the [3+2] cycloaddtion of
54
azidomethyl pivalate 54 with 50 following the same conditions for cyclization as
described for 52. Removal of the allyl protecting group on the PNA backbone afforded
the POM-protected 5-triazolyl monomer 56.
Synthesis and RNA binding studies of PNAs modified with extended
nucleobases. We used the PNA monomers carrying the modified nucleobases M1-M4
and T1 along with the commercially available T and previously described M monomers
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to synthesize modified PNAs (Figure 5) on an Expedite 8909 DNA/RNA/PNA
synthesizer following our previously developed protocols.8, 10, 39 The binding of PNAs to
the complementary model hairpins was studied using isothermal titration calorimetry
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ITC), as in our previous studies that characterized M-modified PNAs.8, 10, 39 The
(
hairpins HRP1-HRP3 contain a variable base pair in the middle of the stem (colored
blue in Figure 5) that allows evaluation of modified nucleobases targeting U-A, C-G, and
G-C base pairs, respectively.
Figure 5. Sequences of dsRNA hairpins and modified PNAs. The modified nucleobase
is highlighted in red; the variable base pair is in blue.
To test the extended nucleobase M1 we used a slightly modified PNA sequence
10
(PNA1, Figure 5) that differs from our standard model 9-mer by being shifted two base
pairs up closer to the U loop that closes the hairpin stem. The reason for initially
4
choosing PNA1 was that it has three consecutive M nucleobases, which allows
checking for possible cooperative stabilization due to stacking of the extended
nucleobases. Substitution of one M with M1 resulted in significant loss of binding affinity
ACS Paragon Plus Environment

Page 23 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
(
~11 fold) of PNA2 (Table 1). Three consecutive M residues further decreased the
affinity of PNA3 ~2.5 fold compared to the singly modified PNA2. While the latter result
suggested that there might be some favorable contribution from stacking of the
extended nucleobases, clearly, the additional pyridine ring of M1 did not produce the
expected stabilizing effect. Comparison of ∆H and ∆S values for PNA1, PNA2 and
PNA3 (Table 1, columns 3 and 4) suggested that the decrease in binding affinity was
caused by unfavorable enthalpy, which was compensated by increase in entropy.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 1. Binding affinities of modified PNAs for dsRNA hairpins.a
6
-1
PNA (modified
base)
Target dsRNA K × 10 M
∆H (kcal/mol)
-T∆S (kcal/mol)
a
(variable base
pair)
PNA1 (M)
HRP1 (U-A)
HRP1 (U-A)
HRP1 (U-A)
HRP2 (C-G)
HRP2 (C-G)
HRP3 (G-C)
HRP3 (G-C)
170 ± 30
15 ± 1
6 ± 1
-71 ± 3
-44 ± 1
-18 ± 1
-103
60 ± 3
35 ± 1
9 ± 1
92
PNA2 (M1)
PNA3 (3 × M1)
PNA4 (M)
110b
PNA5 (M2)
PNA6 (P)
8 ± 1
-48 ± 3
NRd
38 ± 3
NRd
3c
PNA7 (M2R)
6.5 ± 0.5
4 ± 0.5
21 ± 3
-40 ± 1
-38 ± 0
-68 ± 4
31 ± 1
29 ± 0
58 ± 4
PNA8 (MeOM2R) HRP3 (G-C)
PNA9 (M3) HRP2 (C-G)
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
PNA10 (M3R)
PNA11 (M4)
PNA12 (T)
HRP3 (G-C)
7 ± 1
-41 ± 3
32 ± 3
27 ± 1
72b
HRP2 (C-G)
HRP3 (U-A)
HRP3 (U-A)
4 ± 0.5
94b
-36 ± 1
-82b
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
PNA13 (T1)
PNA14 (Tr)
16 ± 3
72 ± 7
-42 ± 6
-79 ± 3
32 ± 6
67 ± 3
HRP3 (U-A)
a
6
-1
Association constant K × 10 M , measured by ITC using a Malvern MicroCal iTC200
a
in 2 mM MgCl , 90 mM KCl, 10 mM NaCl, 50 mM potassium phosphate buffer pH 7.4 at
2
b
10 c
Data from reference.55 d Not reported in reference.55
2
5 °C. Data from reference.
Because M1 did not show the expected improvement in binding affinity or
significant cooperative stabilization, we decided to test other scaffolds using our
standard model system PNAX where we had previous data to benchmark the new
extended nucleobases.10, 55 Substitution of M with M2 resulted in significant loss of
binding affinity for PNA5, ~14 fold compared to PNA4. PNA7 and PNA8 having the
“
reversed” nucleobases, M2R and MeOM2R that target G-C interruptions in polypurine
tracts, showed similarly low binding affinity. The N-acyl group in M2 and M2R was
expected to decrease the pK of aminopyridine, which lowers the binding affinity. We
a
hypothesized that the electron donating methoxy group in MeOM2R would
counterbalance the unfavorable effect of the electron withdrawing N-acyl group,
however, we did not observe an increase in binding affinity of MeOM2R compared to
M2R. A concurrent study in our group supports the hypothesis, but also explains why
binding affinity is not restored upon the addition of 4-methoxy substituent on 2-
aminopyridine nucleobase.56
ACS Paragon Plus Environment

Page 25 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
A promising result was obtained with M3-modified PNA9, where the ~5 lower
affinity compared to PNA4 suggested that the one atom longer linker in M3 was better
tolerated than the shorter linker in M1. Interestingly, the “reversed” nucleobase, M3R
showed notably lower affinity, consistent with our previous observation that cytosine is
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
10
not a good match for recognition of G-C interruptions in polypurine tracts. Taken
together these results suggested that the M3 scaffold could be further improved by
replacing of cytosine with a better ligand for H-bonding to G-C.
Scaffolds built of more rigid, five membered linkers did not improve binding.
Affinity of M4 was weak and while T1 showed encouraging binding, the affinity was ~6
lower than that for the control PNA having unmodified T (cf., PNA12 and PNA13 in
Table 1). To understand the individual contributions of linker and second “nucleobase”
in T1, we synthesized and tested nucleobase Tr, which contained only the triazole linker
residue. Binding affinity of Tr-modified PNA14 was comparable with that of the control
PNA12 carrying unmodified T. This result suggested that the heterocyclic linker alone
did not interfere with triplex formation and that the decreased affinity of T1 was most
likely due to lack of favorable interactions by the phenol ring or steric clashes due to
non-planar conformations about the tricyclic system.
Taken together our results demonstrate the challenges in triple-helical molecular
recognition of pyrimidine interruptions in polypurine tracts of helical nucleic acids.
Interestingly, the binding of PNAs having our “reversed” nucleobases, M2R, MeOM2R
and M3R was comparable or better than that of PNA6 containing P, our current best
13
nucleobase for recognition of G-C interruptions in polypurine tracts. This result
emphasizes the most significant unsolved problem in triple-helical recognition – while it
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
is relatively easy to form the standard Hoogsteen triples, triples involving binding to
pyrimidine Hoogsteen edge are inherently difficult to attain. However, an encouraging
result was obtained with M3, which may be improved by future optimization.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Molecular modeling of PNA-dsRNA triple helices containing extended
nucleobases. To gain more insight into reasons behind the weak affinity of extended
nucleobases, we modeled M1 and T1 in a PNA-dsRNA triplex based on our recent
56
structural study. A potential reason for the weak RNA binding was that the hydrogen-
bonding conformations predicted by ab initio calculations were not favorable for
attachment of extended nucleobases to the PNA backbone. Bases M1 and T1, energy
minimized in the absence of RNA, were fit to M* and T* of PNA residues of M*G-C and
T*A-U triples, respectively, in the PNA-dsRNA structure. These models were subjected
to simulated-annealing calculations in which distance restraints derived from our recent
56
NMR study were applied to the modified residues while atom positions were fixed for
all other residues except the PNA residue preceding the modified residue. The base
triples resulting from these calculations are shown in Figures 5A and 5B (green
carbons). M1 in the M*-aligned M1*G-C triple (Figure 6A) is apparently well-positioned
to form a hydrogen-bond with the C amino group (2.1 Å for N-H distance), consistent
with the ab initio structure. Hydrogen-bonding angles and distances did not reveal any
distortions that could explain the low binding affinity of PNA2 and PNA3 observed in
ITC experiments (Table 1).
ACS Paragon Plus Environment

Page 27 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Simulated-annealing calculations of M1*G-C and T1*A-U triples. A and B,
56
green carbons show M1*G-C and T1*A-U calculated using NMR restraints applied in
structure calculations of M*G-C and T*A-U triples, that were used as starting
coordinates; A and B, thin and brown lines show M1 and T1 in the same calculations
with addition of a distance restraint (maximum 2.4 Å) for the expected third hydrogen-
bond (black, dashed line). C and D, space-filling view of models in A and B,
respectively, that include the third hydrogen-bond restraint. Note contact between O7’
(red) and H8’ (white) in D (T1).
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 85
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In contrast, T1 in the T*-aligned T1*A-U triple (Figure 6B) extends out away from
the A-U pair and is not well-positioned to form a hydrogen-bond with the U carbonyl
oxygen (4.8 Å for H-O distance). This contrasts with the ab initio structure. A second
simulated-annealing calculation was performed on each model with the addition of a
distance restraint for the expected third hydrogen-bond to C or U nucleobases in G-C
and A-U base pairs (dashed lines in Figures 5A and 5B) The resulting conformations
are shown as thin, brown lines in Figures 5A and 5B. In the M1*G-C triple, this restraint
had no effect as the N-H distance is already within the applied restraint (maximum 2.4
Å). In the T1*A-U triple, T1 is pulled back into the major groove (2.45 Å for H-O
distance) and apparently can form a triple much as observed in ab initio calculations.
However, there is a steric cost for this conformation. Introduction of the “hydrogen-bond”
distance restraint in T1*-AU forces the proximal end (H8’) of the nucleobase to swing
out towards the major groove resulting in contact with oxygen of the carbonyl group
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
O7’) of the previous PNA residue (Figure 6D, white and red). This contact, as well as
increased T1-A contact, is accompanied by an increase of van der Waals energy by
.6±0.6 kcal (average of 10 simulations). As suggested by Figure 6A, introduction of the
hydrogen-bond distance in M1*G-C results in no change in van der Waals energy (-
.2±0.5 kcal). The close tolerances surrounding the nucleobase-to-PNA linker is
4
0
demonstrated in Figure 6C. While steric factors near the PNA backbone demonstrate a
possible consideration in the design of extended nucleobases, they do not explain the
similarly poor affinities of M1 and T1. Similar docking of M2, M3, and M4 also do not
reveal obvious explanations for their affinities. Higher level modeling may be required to
explain the binding results.
ACS Paragon Plus Environment

Page 29 of 85
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
CONCLUSIONS
Our present study illustrates the challenges in designing extended heterocyclic
systems that hydrogen-bond the entire Hoogsteen edge of Watson-Crick base pairs.
Initial modeling and ab initio calculations suggested the feasibility of several hydrogen-
bonding scaffolds. The first challenge was to develop synthetic routes that link the
heterocyclic nucleobases and form the proposed extended scaffolds. To that end, we
successfully prepared five new extended nucleobases, M1-M4 and T1, incorporated
these at selected positions in PNA 9-mers, and measured their RNA binding affinity
using ITC. Somewhat surprisingly, the extended nucleobases displayed reduced
binding affinity compared to the original M and T nucleobases. Simulated-annealing
calculations of T1 identified another challenge of designing extended heterocyclic
systems. While T1 could achieve reasonable three hydrogen bond binding, it was
accompanied by significant distortion in PNA backbone leading to unfavorable
stereoelectronic interactions. In contrast, similar calculations on M1 did not reveal any
reasons for the decreased binding affinity, suggesting that other unidentified factors, for
example, subtle conformational preferences and distortions of hydration network, may
play significant role in overall binding energetics. Taken together, these results provide
important insights for development of next generation nucleobase analogues that may
address the long-standing limitation of triple helical recognition of nucleic acids – the
requirement for long polypurine tracts.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment