DNG cytidine: Synthesis and binding properties of octameric guanidinium-linked deoxycytidine oligomer

Article abstract of DOI:10.1016/j.bmc.2004.05.010

The synthesis of guanidinium-linked cytidyl oligomer (DNG-C₈), a cationic DNA analog, and the corresponding cytidine monomers is described. The DNG monomer synthesis was streamlined to produce a shorter route to the final monomer than previously reported for thymidine and subsequent solid-phase synthesis produced an octameric cytidyl DNG strand. Because octameric deoxyguanosine would be used as the complementary strand in our studies, it was necessary to investigate guanosine self-association. Singular value decomposition was used to mathematically deconvolve the spectral data and confirm the presence of transitions due to DNA-G₈ self-association. Job plots show the binding stoichiometry of DNG-C₈ with DNA-G ₈ to be 1:1. Thermal denaturation studies of the DNG-C ₈·DNA-G₈ duplex established a T _m≥90°C and a ΔG°=-13.3kcalmol^-1, indicating the DNG-C₈·DNA-G₈ duplex is over 1000 times more stable than that of DNA-C₈·DNA-G₈.

Full text of DOI:10.1016/j.bmc.2004.05.010

Bioorganic & Medicinal Chemistry 12 (2004) 4233–4244
DNG cytidine: synthesis and binding properties of octameric
guanidinium-linked deoxycytidine oligomer
Istvan E. Szabo and Thomas C. Bruice*
Department of Chemistry and Biochemistry, University of California, UCSB, Santa Barbara, CA 93106, USA
Received 27 February 2004; revised 7 May 2004; accepted 7 May 2004
Available online 19 June 2004
Abstract—The synthesis of guanidinium-linked cytidyl oligomer (DNG-C8), a cationic DNA analog, and the corresponding cytidine
monomers is described. The DNG monomer synthesis was streamlined to produce a shorter route to the final monomer than
previously reported for thymidine and subsequent solid-phase synthesis produced an octameric cytidyl DNG strand. Because
octameric deoxyguanosine would be used as the complementary strand in our studies, it was necessary to investigate guanosine self-
association. Singular value decomposition was used to mathematically deconvolve the spectral data and confirm the presence of
transitions due to DNA-G8 self-association. Job plots show the binding stoichiometry of DNG-C8 with DNA-G8 to be 1:1. Thermal
denaturation studies of the DNG-C8ÆDNA-G8 duplex established a Tm P 90 °C and a DGꢀ ¼ ꢁ13:3 kcal molꢁ1, indicating the DNG-
C8ÆDNA-G8 duplex is over 1000 times more stable than that of DNA-C8ÆDNA-G8.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
methylimino,9 carbamates,10 acetals,11 heterocycles,12
and locked nucleic acids.13
Oligonucleotide analogs capable of recognizing and
binding complementary RNA and DNA with the in-
tended purpose of arresting cellular processes at the
translational and transcriptional level are known as
antisense and antigene agents,1;2 respectively. Key goals
in designing antisense/antigene agents are the following:
high specificity to and affinity for the target sequence,
stability in the presence of cellular nucleases and mem-
brane permeability. In the specific case of RNA targeted
antisense therapy, RNase H activation3 is a desirable
feature. Since antisense agents were first introduced, a
diverse range of analogs have been developed. Substi-
tution of the phosphodiester oxygens while preserving
the phosphate core results in the formation of such
analogs as phosphorothioates, phosphoroamidates, and
methylphosphonates,4 while total replacement of the
furanose–phosphate moiety with peptide linkages results
in peptide nucleic acids (PNA),5 PHONA,6 and PNAA.7
These neutral linkages eliminate the mutual repulsion
seen in the anionic backbones of duplex DNA and are
resistant to nucleases. Other variations on the internu-
cleoside linker theme include morpholino,8 methylene-
In recent years, approaches involving the incorporation
of positive charges in antisense oligomers have been
developed. Positive charges can be added to the bases14
or the sugar rings to give zwitterionic DNA.15 The
phosphate backbone and phosphoramidate linkages can
be alkylated with alkylamines to produce positively
charged phosphate triester linkages16;17 and cationic
phosphoramidate linkages.18 A promising approach
designed by this laboratory is the replacement of inter-
nucleoside phosphate linkage with an achiral, positively
charged guanidinium group19 (Fig. 1). This species,
designated as DNG, for deoxy-nucleic-guanidine,
exhibits high affinity and specificity for DNA as well as
resistance to nucleases.20 Thymidyl DNG oligomers
have been demonstrated to have a high affinity for, and
bind in a 2:1 stoichiometry with, complementary adenyl
DNA oligomers,21 while DNG-A oligomers form 1:2
complexes with thymidyl DNA oligomers.22 A guanidi-
nium linked RNA analog is currently in develop-
ment.23;24
The guanidinium chemistry described here proceeds in
the 30 ! 50 direction and is compatible to standard solid-
phase DNA synthesis, making it possible for the design
of DNG/DNA chimera. To this date, DNG oligonu-
cleotides have been synthesized with only adenyl and
Keywords: Oligomer; Monomer; DNA; Antisense/antigene.
* Corresponding author. Tel.: +1-80-5893-2234; fax: +1-80-5893-2229;
e-mail: tcbruice@chem.ucsb.edu
0968-0896/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.05.010

4234
I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
Figure 1. Structure of phosphodiester and guanidinium linkage.
thymidyl moieties. We present in this report, for the first
time, the total synthesis of the cytidine DNG monomer
and solid-phase synthesis of octameric cytidyl DNG, as
well as a preliminary investigation into DNG-C8 bind-
ing properties with natural DNA-G8.
2. Results and discussion
Since their inception, the synthesis of DNG oligomers
has evolved dramatically. Initially, the synthesis of the
oligomers occurred in the 50 ! 30 direction, resulting in a
backbone comprised of intermediate thiourea linkages
which were subsequently oxidized and amidated to
provide the desired guanidinium linkages.25 Ultimately
the direction of oligomer synthesis was reversed to run
30 ! 50 and protected guanido groups were formed
during monomer coupling reactions. During oligomer
cleavage from the solid support, the guanido groups
were deprotected to provide the guanidinium linkages.22
Synthesis in the 30 ! 50 direction has the added advan-
tage of being compatible with standard DNA technol-
ogies, making DNA/DNG chimera feasible.
Scheme 2. 50-Trityl is removed to expose amine for the next coupling
reaction.
This loading monomer is a N4-benzoyl-50-MMTr-NH-
30-O-succinate derivative of 8, which was prepared from
the 20-deoxycytidine starting material (Scheme 4). For-
mation of Hg2S from HgS often results during the
course of the coupling reaction, precipitating out and
turning the solution and glass support a dark gray color.
This Hg2S precipitate was removed by the addition of
20% thiophenol in DMF.
Acetic anhydride was used to acylate any unreacted
amines, thereby capping incompletely coupled oligo-
mers. The 50-trityl group of the terminal monomer was
removed with dichloroacetic acid in order to regenerate
the 50-amino moiety necessary for the coupling reaction
to follow (Scheme 2). The process was repeated until the
oligomer of desired length was achieved. The 50-trityl
protecting group was allowed to remain after the last
coupling reaction in order to facilitate purification.
The coupling reaction (Scheme 1) proceeded with the
abstraction of the sulfur atom from an Fmoc-protected
thiourea using mercury(II) and transformed the thiourea
into an activated carbodiimide, which reacted readily
with a free amino group to give the protected guanido
linkage. The use of long-chain-alkylamine controlled
pore glass is now preferred to polystyrene resin as a solid
support because the oligomer can be removed easily
from the glass using methanolic ammonia, a step which
also removes all base labile protecting groups from the
oligomer. The controlled pore glass (CPG) was first
loaded with a ‘primer’ monomer using a base labile
succinyl-ester linkage following a standard procedure.26
Cleavage and partial deprotection of the oligomer was
performed in one step, which consists of soaking the
CPG in methanolic ammonia overnight at 60 °C
(Scheme 3). The supernatant was concentrated to a
white powder and re-dissolved in 10% acetonitrile in
Scheme 1. First step in monomer coupling is the formation of the carbodiimide and subsequent addition to form the protected guanido group.

I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
4235
Scheme 3. Cleavage from CPG, deprotection of guanido groups and nucleobases occurs in one step using methanolic ammonia at 60 °C.
triethylammonium acetate buffer for reverse-phase (RP)
HPLC purification. Because acetic anhydride was used
to cap all uncoupled amines after each coupling reaction
and the 50-MMTr protecting group was not removed
after addition of the last monomer, purification of the
final oligomer was straightforward.
The synthesis of the final cytidine coupling monomer
(Scheme 4) began with benzoylating commercially
available 20-deoxycytidine, using an established transient
protection method,26 in order to protect the exo-cyclic
amine on the nucleobase. The resultant N4-benzoyl-20-
deoxycytidine 1 was treated with excess methanesulfonyl
chloride to provide 30,50-dimesyl deoxycytidine 2. Re-
fluxing a DMF solution of 2 with potassium phthal-
imide displaced the 30-mesyl and formed a 2,30-anhydro
species 3, preserving the 50-mesyl moiety.27 Treatment of
3 with LiN3 in DMF displaced the 50-mesyl group and
broke the 2,30-anhydro bond resulting in 30,50-diazido-
N4-benzoyl cytidine 4. Catalytic hydrogenation of the
azides to amines using activated carbon and hydrogen
Using a balance of aqueous buffer and organic solvent
as eluents, ‘trityl-on’ purification easily separated
incompletely coupled oligomers from the trityl-pro-
tected final oligomer. Following purification, the trityl
group was removed with 2% dichloroacetic acid in
methylene chloride and the oligomer was again purified
by RP-HPLC.
Scheme 4. Reagents and conditions: (a) BzCl/pyridine; (b) MsCl/pyridine; (c) KPhTh/DMF; (d) LiN3/DMF; (e) H2ðgÞ/Pd–C/EtOH; (f) MMTrCl/
pyridine; (g) Fmoc-NCS/pyridine; (h) TsCl/pyridine; (d) LiN3/DMF; (e) H2ðgÞ/Pd–C/EtOH; (f) MMTrCl/pyridine.

4236
I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
gas gave 30,50-diamino-N4-benzoyl cytidine 5, which was
reacted with 0.9 equiv of MMTrCl in pyridine at ꢁ70 °C
to yield the 50-MMTr-NH derivative 6. Although a
small amount of ditritylated product formed, it was
easily separated by silica gel chromatography. Lastly, 6
was reacted with 9-fluorenylmethoxy-carbonylisothio-
cyanate23;28 in pyridine to give N4-benzoyl-30-Fmoc-
NCS-50-MMTr-NH-20deoxycytidine 7.
DNA-C8ÆDNA-G8 also contained DNA-G8 self-associ-
ated structures, the transitions for all of the species
would be included the spectrum. We employed singular
value decomposition (SVD), a linear algebra based com-
putational38 method, to analyze the melting data. SVD
factors the input matrix, A (representing spectral absorp-
tion values), into three matrices, U, V T and S, where
superscript T indicates V is transposed and S is a diag-
onal matrix. Much of the insight into applying SVD to
deconvolve spectra has been provided by work of Henry
and Hofrichter,39 Hendler and Shrager,40 and Haq et al.41
2.1. Physical properties
It is well known that DNA rich with deoxyguanosine
can form complex noncovalent secondary structures
with itself, namely G-quadruplexes and tetraplexes.29;30
Given that octameric deoxyguanosine (DNA-G8) would
be needed to complement our DNG-C8, it was necessary
to determine whether octameric guanosine self-associa-
tion would present a problem. Although deoxycytosine
can also self-associate to form double helices31 and
i-motifs,32 this generally requires a low pH to protonate
the cytosine N3 (pKa ¼ 4:17).33 DNA-C8, in our working
buffer ([KCl] ¼ 100 mM; [KHPO4] ¼ 10 mM; pH 7), was
heated to 90 °C and showed no increase in absorbance
with respect to temperature. Additionally, the overall
positive charge of DNG-C8 strongly disfavors self-
association.34 Binding stoichiometry was determined by
performing Job plots35 on DNA and DNG hexomeric
oligomers (see Methods). For both the DNA-C6ÆDNA-
G6 and DNG-C6ÆDNA-G6 mixtures, the binding was
determined to be 1:1 (Fig. 2). This binding stoichiometry
suggests the formation of a duplex DNGÆDNA strand
with Watson–Crick base pairing. Although the possi-
bility of other base pairing arrangements exist,36 they are
unlikely to be formed in a an aqueous solution buffered
to pH 7. Furthermore, previously published computa-
tional37 and circular dichroism21 studies provide evi-
dence that the base pairing is indeed Watson–Crick and
the bound DNGÆDNA strands are helical in structure.
The data set is created by scanning the sample from 300
to 220 nm at every degree as the sample is heated from 5
to 90 °C at a rate of 0.5 °C/min. The result can be
visualized as a 3-D surface (Fig. 3). This absorbance
data is arranged into a matrix, A, containing absorbance
values whose columns represent temperature and rows
represent wavelengths. SVD expresses matrix A as the
product of three matrices, USV T. The diagonal matrix S
contains the square roots, or singular values, of the
eigenvalues for the vectors specified in U and V . The
singular values of S are a measure of weight, in terms of
value, for those columns in U and V , and indicate the
number of significant spectral species.
A plot of the (diagonal entry number in matrix S) versus
(10 ꢂ log of that value) yields a curve which begins to
deviate sharply for entry numbers 1–5 (Fig. 4) for DNA-
G8 and DNA-C8ÆDNA-G8, suggesting that there are at
most five statistically significant singular values. Auto-
correlation calculations were performed in order to
measure the ‘signal-to-noise’ ratio in the column vectors
of V and determine which columns contained valid
information and, therefore, the number of significant
singular values. Eq. 1 was used, along with the estab-
lished criterion of CðViÞ P 0:8 for vectors with a strong
‘signal’,39 to confirm that there were at most five sig-
In order to confirm that the observed transitions were
indeed from deoxyguanosine and deoxycytidine base-
pairing, rather than deoxyguanosine self-association, we
deconvolved the spectra to determine the number of
species undergoing transition. If a heated solution of
Figure 3. 3-D mesh plot of DNA-C8ÆDNA-G8 melting. Wavelength
scans (300–220 nm) are taken every degree as sample is heated from 5
to 90 °C at a rate of 0.5 °C/min. ([DNA] ¼ 15 lM; [KCl] ¼ 100 mM;
[KHPO4] ¼ 10 mM; pH 7).
Figure 2. Job plot illustrating 1:1 binding of DNA-C6, and DNG-C6,
to DNA-G6. Total oligomer concentration was 12 lM and buffer
contained 100 mM [KCl], 10 mM [KHPO4], adjusted to pH 7.

I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
4237
A difference spectrum involves creating a new D matrix,
from matrices V and U using a minimum number of S
values, then subtracting D from the original matrix A
and plotting the difference. Selecting the correct number
of S values will produce a difference plot containing only
noise. For the sample containing only DNA-G8, the
difference spectra computed with three singular values
gave rise to randomness in the plot (Fig. 5). Note
that four singular values did not increase this random-
ness. This indicates that for DNA-G8, three singular
values are statistically significant. Difference spectra
for DNA-C8ÆDNA-G8 indicates there are four signifi-
cant singular values, one more than for DNA-G8
(Fig. 6).
Figure 4. Plot of 10 ꢂ log (singular value) versus position in matrix S
reveals that 1–5 appear significant for the samples DNA-G8 and DNA-
C8ÆDNA-G8.
Although the DNA-G8 sample contained multiple
transitions, we cannot say exactly what structures they
arise from. However, we can say that these transitions
also occur in the DNA-C8ÆDNA-G8 sample, indicating
that deoxyguanosine self-association occurs even in
presence of deoxycytosine. SVD clearly indicates that
the sample containing DNA-C8ÆDNA-G8 has one more
transition than the sample containing only DNA-G8,
due to the interaction of deoxyguanosine and deoxy-
cytidine residues.
nificant singular values for both DNA-G8 and DNA-
C8ÆDNA-G8 (data not shown).
nꢁ1
X
CðViÞ ¼
Vj;iVjþ1;i
ð1Þ
j¼1
Vj;i is the jth element in the ith column of matrix V .
Figure 5. Plots of difference matrix, D, for the first four singular values of the diagonal matrix S for DNA-G8. The plot generated using three singular
values (lower left) shows a complete subtraction of the spectrum, leaving only noise.

4238
I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
Figure 6. Plots of difference matrix, D, for the first four singular values of the diagonal matrix S for DNA-C8ÆDNA-G8. The plot generated using four
singular values (lower right) shows a complete subtraction of the spectrum, leaving only noise.
2.2. Thermodynamic calculations
While repetitive scanning over multiple wavelengths is
used for SVD, thermal denaturation studies monitor the
absorbance at a constant wavelength. Although deoxy-
guanosine has a kmax at 252 nm, it shows the largest
temperature-correlated absorbance between 270 and
275 nm. Given that deoxycytidine has a kmax at 272 nm,
we chose that as the wavelength to observe.
At 272 nm, DNA-G8 has a clear transition at ꢃ50 °C,
due to the thermal disassociation of a secondary struc-
ture (Fig. 7).
The melting curves for DNA-C8ÆDNA-G8 and DNG-
C8ÆDNA-G8, at 272 nm, indicate that two very different
Figure 7. The melting curve of octameric deoxyguanosine at 272 nm;
heated from
5
to 90 °C ([DNA] ¼ 3 lM; [KCl] ¼ 100 mM;
transitions are occurring (Fig. 8). DNA-C8ÆDNA-G8
shows a typical melting curve with a Tm at 38 °C, while
DNG-C8ÆDNA-G8 shows a long upward inflection,
suggestive of the lower half of a sigmoid curve. The
published melting temperatures for DNG-T8ÆDNA-A8
and DNG-TA5ÆDNA-T8 are 63 and 79 °C, respectively,
while the Tm of their respective DNAÆDNA counterparts
is below 10 °C.21;22 Assuming a similar increase in Tm,
and given that the Tm of DNA-C8ÆDNA-G8 was exper-
[KHPO4] ¼ 10 mM; pH 7).
imentally determined to be 38 °C, the melting tempera-
ture of our DNG-C8 with complementary DNA-G8 is
expected to be above 90 °C.
In a binding study of DNG-T with DNA-A, we found
that an increase in ionic strength (l) results in a decrease

I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
4239
Figure 8. The melting curves of DNA-C8ÆDNA-G8 and DNA-
C8ÆDNG-G8 heated from
[DNA:DNG] ¼ 7.5 lM; [KCl] ¼ 100 mM; [KHPO4] ¼ 10 mM; pH 7).
Figure 9. Plot of ‘a versus temperature’ for DNA-C8ÆDNA-G8.
5
to 90 °C ([DNA:DNA] ¼ 7.5 lM;
in the DNGÆDNA melting temperature. This l affect is
opposite to that seen with DNAÆDNA interactions and
is attributed to salt ions disrupting the favored electro-
static attraction between the negative phosphates and
the positive guanidiniums. Increasing monovalent cat-
ion concentration, specifically sodium or potassium,
stabilizes the formation of guanosine quartets, which are
unable to hydrogen-bond with cytosine residues.42;43
Thus, when l of solutions containing either DNA-
C8ÆDNA-G8 or DNG-C8ÆDNA-G8 oligomers was in-
creased, from 0.13 to 0.33 and 0.53, we observed no
hyperchromic shift.
Following the work of Marky and Breslauer,44 we cal-
culated the thermodynamic data for the melting transi-
tions of DNA-C8ÆDNA-G8 and DNG-C8ÆDNA-G8. The
first step is to convert the experimental ‘absorbance
versus temperature’ curve into an ‘a versus temperature’
profile (Figs. 9 and 10). Defining a as being equal to the
fraction of single strands in the duplex state, we can
state that if x is the height between the absorbance curve
and upper baseline, or last absorbance value, and (x þ y)
is the height between the upper and lower baselines, or
first and last absorbance values, then a ¼ x=ðx þ yÞ.
Figure 10. Plot of ‘a versus temperature’ for DNG-C8ÆDNA-G8.
1
KT
¼
ð3Þ
m
nꢁ1
ðCT=2nÞ
DHVH
ꢀ
ꢁ
1
1
ln KðTÞ ¼ ln KðTmÞ ꢁ
ꢁ
ð4Þ
ð5Þ
R
T
Tm
DGꢀ ¼ ꢁRT ln KðT Þ
The goal is to use this new ‘a versus temperature’ plot to
Table 1 summarizes the thermodynamic results and
parameters used for DNA-C8ÆDNA-G8 and DNG-
C8ÆDNA-G8, clearly demonstrating the tighter binding
of DNG to its DNA template. Because DNG-C8ÆDNA-
G8 appears to have a Tm at or above the temperature
limit of the experiment, we used a conservative estimate
of Tm ¼ 90 °C for our calculations and defined a ¼ 1=2
at that temperature. The DDGꢀ of ꢁ4.8 kcal molꢁ1
translates into over three orders (3.4) of magnitude in-
crease in binding. This means that DNG-C8 binds
DNA-G8 over 1000 times more tightly than DNA-C8.
This increase is attributed to the powerful attraction
between the positively charged DNG guanidinium
groups and negatively charged phosphates of DNA.
find the value for the term ðoa=oTÞT¼T in the van’t Hoff
m
equation (Eq. 2), where the n ¼ 2 (strand molecularity),
Tm ¼ 311:15 K, and R ¼ 1:9872 cal molꢁ1 Kꢁ1. Because
ðoa=oTÞT¼T is the slope of the curve for ‘a versus tem-
perature’ amt a ¼ 1=2 and T ¼ Tm, computational meth-
ods can be used to get an equation f ðxÞ, which fits the a
versus temperature’ curve, and calculate its first deriv-
ative f 0ðxÞ, evaluating it at Tm. At the melting temper-
ature Tm where a ¼ 1=2, the equilibrium constant for
non-self-complementary sequences is given by (Eq. 3),
where CT is the total strand concentration (15 lM) and
n ¼ 2. ln KðTÞ is given by (Eq. 4), where T ¼ 298:15 K,
and DGꢀ is given by (Eq. 5).
ꢀ
ꢁ
The strength and fidelity of the GÆC base pair, due to
three hydrogen bonds, in combination with the elec-
trostatic attraction of DNG for DNA suggests that
oa
oT
DHVH ¼ ð2 þ 2nÞRTm2
ð2Þ
T¼Tm

4240
I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
Table 1. Thermodynamic data and parameters for DNA-C8ÆDNA-G8
and DNG-C8ÆDNA-G8
4 °C refrigerator overnight. The samples were placed in
the spectrophotometer’s Peltier heating block, the tem-
perature was ramped from 5 to 90 °C at a rate of
0.5 °C minꢁ1 under an atmosphere of nitrogen gas, and
data was collected at 1 °C intervals.
DNA-C8ÆDNA-G8
y0 þ a
DNG-C8ÆDNA-G8
f ¼ y0 þ ax þ bx2 þ cx3
Regression
equation
ꢀ
ꢁ
f ¼
ꢂ
ꢃ
ꢁðxꢁx
Þ
0
b
1 þ e
Parameters
a ¼ 1:0762
y0 ¼ 0:9989
b ¼ ꢁ11:5479
x0 ¼ 36:5959
y0 ¼ ꢁ0:0102
38 °C
a ¼ ꢁ5:7815 ꢂ 10ꢁ5
b ¼ 5:2645 ꢂ 10ꢁ5
c ¼ ꢁ1:2457 ꢂ 10ꢁ6
90 °C
3.4. Job plots
Five 1 mL samples were prepared identically to the
thermal denaturation procedure, using the same buffer
and annealing conditions, each with a different mole
fraction of deoxycytidine and deoxyguanosine oligo-
mers. The mole fraction of each strand was varied from
0.1 to 0.9, in 0.2 increments, while the total strand
concentration was kept at 12 lM. The absorption of
each sample was recorded at room temperature and the
results were plotted against mole fraction of oligomer in
order to determine the binding stoichiometry.
Tm
oa=oT
DHVH
ln KðTÞ
DGꢀ
ꢁ0.0230
ꢁ0.0209
ꢁ26.5 kcal molꢁ1
14.3660
ꢁ32.9 kcal molꢁ1
22.4218
ꢁ8.5 kcal molꢁ1
ꢁ13.3 kcal molꢁ1
strategically placed guanidinium-linked cytidines in a
DNA/DNG chimera may yield a powerful antisense
combination.
3. Experimental
3.1. Materials
3.5. Solid-phase synthesis of octameric cytidyl DNG
3.5.1. Loading. Commercially available long-chain-
ꢀ
alkylamine controlled pore glass (CPG), with 500 A pore
All anhydrous solvents, triethylamine, methane sulfonyl
chloride, monomethoxy trityl chloride, 9-fluorenyl-
methylchloroformate, mercury(II) chloride, and potas-
sium isothiocyanate were purchased through Aldrich
and used without further purification. HPLC grade
triethylammonium acetate buffer was purchased from
Fluka, HPLC grade acetonitrile was purchased from
Fisher, long-chain-alkylamine controlled pore glass was
purchased from Sigma, ammonia gas was purchased
from Matheson, and the 20-deoxycytidine monohydrate
starting material was purchased from Acros.
size and 80–120 mesh size, was used as the solid support
and the N4-benzoyl-30-hydroxy-50-NH-MMTr-20,50-
deoxycytidine monomer 8 was loaded onto the support
as the 30-succinyl ester using standard methods. UV
analysis of the trityl cation released upon treatment with
2% dichloroacetic acid (DCA) in dichloromethane
(DCM) revealed that the loading was 31 lmol/g. The
cytidine-loaded CPG was dried under vacuum and
stored in a desiccator at room temperature.
3.5.2. Coupling. Trityl-protected, cytidine-loaded CPG
(60 mg) was placed in a screw cap vial with a coarse
fritted bottom and stopcock and the trityl protecting
group was removed, exposing the 50-amine, with the
addition of 10 mL 2% DCA in DCM. The CPG was
washed 8 mL of DCM, followed by 10 mL DMF, then
soaked in DMF (2 mL for 2 h prior to initial coupling.
The 25 mg (20 equiv) of N4-benzoyl-30-NH-Fmoc-50-
NH-MMTr-20,30,50-deoxycytidine monomer 13 was dis-
solved in 1 mL DMF and 5 mg HgCl2 (2 equiv) and 7 lL
TEA were separately dissolved in 0.5 mL DMF each.
The DMF was removed from the pre-soaked cytidine-
loaded CPG and the 1 mL coupling monomer in DMF
solution was added, followed by the simultaneous
addition of the HgCl2 and TEA in DMF solutions. A
fine yellow-white precipitate immediately formed and
the vial was fitted with a Teflon gasket and capped
tightly, then agitated gently for 1.5 h at room tempera-
ture.
3.2. General
Hydrogenations were carried out with a Parr hydroge-
nator equipped with a 500 mL hydrogenation vessel.
TLC was carried out on silica gel (Kieselger 60 F254
)
glass-backed commercial plates and visualized by UV
light. 1H and 13C NMR spectra were obtained on a
Varian Unity 500 MHz. Analytical reverse phase HPLC
was performed on a Hewlett–Packard 1050 system
equipped with a quaternary solvent delivery system and
UV detector set at 260 nm and a 10 ꢂ 250 mm Macro-
sphere C8 reverse-phase column purchased from Altech.
UV spectra, thermal studies and Job plots were obtained
on a Cary 100 Bio UV/vis spectrophotometer equipped
with a temperature programmable cell block and a PC
interface running Cary WinUV software from Varian,
Inc.
3.3. Thermal denaturation experiments
DNA-C8ÆDNA-G8 and DNG-C8ÆDNA-G8 samples were
dissolved in a 1:1 stoichiometry, with a total oligomer
concentration of 15 lM, in buffer containing 100 mM
[KCl], 10 mM [KHPO4], adjusted to pH 7. Samples were
covered with a layer of mineral oil and heated to 90 °C,
allowed to cool to room temperature, then stored in a
3.5.3. Capping. After removing the coupling solution and
washing with DMF, the CPG retained a slight gray color
due to remaining mercury(I) sulfide precipitate. This was
removed by washing with 10 mL of 20% thiolphenol in
DMF, followed by copious amounts of DMF. Any un-
reacted amines were capped by the addition of 1 mL of

I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
4241
100 mM acetic anhydride in DMF and 1 mL of 200 mM
TEA in DMF. The reaction mixture was sealed and
agitated for 10 min at room temperature then filtered off
and washed with 10 mL of DMF, 10 mL MeOH, and
finally an additional 10 mL of DMF.
(RNG) elucidated a quick path to the final coupling
monomer beginning with a 30,50-diamino ribonucleoside
and this route has now been applied to the deoxynucleic
monomers as well.
3.7. N4-Benzoyl-20-deoxycytidine (1)26
3.5.4. Trityl-deprotection and assay. To prepare for the
next coupling and to assay the yield of the last coupling,
the 50-trityl moiety was removed with 10 mL 2% DCA in
DCM. The bright yellow deprotection solution was col-
lected and a 20 lL aliquot was diluted into 980 lL of 60%
perchloric acid in absolute ethanol. This sample was UV-
analyzed to determine the total monomethoxy-trityl cat-
ion released, and thus the molarity of monomer coupled.
20-Deoxycytidine (3.31 g; 14.6 mmol) was suspended in
anhydrous pyridine and cooled to 0 °C in an ice bath
while stirring under argon. Chlorotrimethylsilane
(9.24 mL; 5 equiv) was added dropwise with a syringe
and the mixture was stirred for 30 min. Benzoyl chloride
(8.45 mL; 5 equiv) was added dropwise with a syringe
and the mixture was removed from the ice bath and
allowed to stir at room temperature for 2 h. After again
cooling to 0 °C while stirring, ice cold distilled water
(20 mL) was slowly added and followed by the slow
addition of concentrated aqueous ammonia (20 mL),
15 min later. After 30 min the reaction mixture was
rotovapped to a minimum volume and distilled water
(100 mL) added, then filtered through filter paper. The
aqueous filtrate was extracted with diethyl ether
(2 ꢂ 100 mL), causing crystals to immediately form, and
was placed in the refrigerator overnight. Crystals were
filtered off and vacuum dried yielding 4.155 g (86%) of
pure material. TLC (1:5, MeOH–EtOAc) Rf ¼ 0:20;
LRMS (ESI) m=z calculated for C16H17N3O5 (MþH)þ
3.5.5. Cycle repeated. The coupling/capping/deblocking
cycle was repeated six times, yielding a heptamer. After
the seventh coupling reaction, the capping and deblock-
ing steps were skipped to allow the monomethoxytrityl
protecting group to remain on the final octamer, facili-
tating HPLC purification. The coupling yields were
estimated by trityl-UV analysis to be as follows: first
coupling (dimer) was 99%; second coupling (trimer) was
98%; third coupling was 96%; fourth coupling was 92%;
fifth coupling was 91%, sixth 86%. Because the trityl
protecting group was not removed after the seventh
coupling, no estimation could be made.
1
332.1168, found 332.1120; H NMR (400 MHz DMSO-
d6) d (ppm) 2.00 (d, 1H, 30-OH), 2.15 (dd, 1H, 50-OH),
2.28 (m, 2H, 20-H2), 3.68 (m, 1H, 30-H), 3.76 (m, 2H, 50-
H2), 3.95 (m, 1H, 40-H), 4.66 (dd, 1H, H5), 4.74 (dd, 1H,
10-H), 7.32 (d, 1H, H6), 7.40–7.85 (m, 5H, N4-Bz), 8.25
(s, 1H, N4-H).
3.5.6. Cleavage and base-labile deprotection. After the
seventh coupling, the solid support was washed with the
thiophenol/DMF solution followed by copious amounts
of DMF and finally MeOH. The reaction vial was dried
under vacuum overnight to remove all traces of solvent.
The CPG was transferred to a pressure resistant vial, a
4 mL solution of methanolic ammonia was added, and
the vial was sealed and placed in a 60 °C oven for 20 h to
cleave the oligomer from the solid support and remove
base-labile protecting groups. After cooling, the depro-
tection solution was removed by vacuum centrifugation,
yielding a white residue containing the crude trityl-
protected octameric cytidine.
3.8. N4-Benzoyl-30,50-O-dimesyl-20-deoxycytidine (2)
N4-Benzoyl-20-deoxycytidne 1 (3.72 g; 11.2 mmol) was
dissolved in anhydrous pyridine (60 mL) and stirred
under argon at 0 °C while methanesulfonyl chloride
(1.74 mL; 2 equiv) was added dropwise with a syringe.
The reaction mixture was allowed to come to room
temperature over 2 h and the product was then precipi-
tated by pouring into vigorously stirring ice water
(350 mL) and maximized by refrigerating overnight. The
solid was filtered, washed with cold water, and dried
under vacuum to give 5.32 g of white powder (97%).
3.5.7. HPLC purification. A gradient from 100% eluent
A (100 mM triethylammonium acetate, pH 7.0) to 90%
eluent B (acetonitrile) over 40 min with a flow rate of
1.5 mL/min was used. Fractions containing the purified
oligomer were collected and lyophilized in a Labconcoâ
CentriVapâ concentrator. The DNG-C8 oligomer was
characterized by ESI high-resolution mass spectrometry;
(MþH)þ 2256.1062, found 2256.1034. After 2% TFA in
DCM treatment, to remove the 50-trityl protecting
group, the HPLC purification was repeated. The fully
deprotected oligomer was characterized by ESI high-
resolution mass spectrometry; (MþH)þ 1983.9861,
found 1983.9888.
TLC (1:5, MeOH–methylene chloride) Rf
¼
0.55;
LRMS (ESI) m=z calculated for C17H19N3O7S (MþH)þ
1
488.0719, found 488.0705. H NMR (400 MHz DMSO-
d6) d (ppm) 2.27 (m, 2H, 20-H2), 3.14 (s, 3H, Ms-CH3),
3.18 (s, 3H, Ms-CH3), 3.69 (m, 1H, 30-H), 3.77 (m, 2H,
50-H2), 3.95 (m, 1H, 40-H), 4.67 (dd, 1H, H5), 4.73 (dd,
1H, 10-H), 7.31 (d, 1H, H6), 7.39–7.85 (m, 5H, N4-Bz),
8.24 (s, 1H, N4-H).
3.9. N4-Benzoyl-50-O-mesyl-2,30-anhydro-20-deoxycyt-
idine (3)
3.6. Synthesis of cytidine monomers
The monomers used in the solid-phase synthesis
(Schemes 1 and 2) were prepared according to Scheme 4.
Recent work in synthesizing ribonucleic guanidine
N4-Benzoyl-30,50-O-dimesyl-20-deoxycytidine 2 (5.32 g;
10.9 mmol) was dissolved in anhydrous N,N-dimethyl-
formamide (80 mL), potassium phthalimide (4.0 g;

4242
I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
2 equiv) was added, and the mixture was heated to 90 °C
for 15 min while stirring under argon then allowed to
cool to room temperature. The solid was removed by
filtration, washed with 40 mL of DMF and the filtrates
were concentrated under vacuum to a yellow solid, and
dried under vacuum overnight. The crude product was
dissolved acetone/MeOH (2:1) and adsorbed on silica
gel then loaded onto a pre-existing silica column and
eluted with acetone. Pure fractions collected were con-
centrated under vacuum to yield 3.0 g (60%) of white
foam. TLC (acetone) Rf ¼ 0:50; LRMS (ESI) m=z cal-
culated for C16H16N6O4 (MþH)þ 392.0838, found
room temperature overnight. The reaction mixture was
cooled again to ꢁ70 °C and anhydrous CH2Cl2 (20 mL)
was added drop wise and the mixture was allowed to
warm to room temperature. H2O (20 mL) was added and
the mixture was vigorously shaken, then centrifuged. The
top organic layer was removed and concentrated to a
solid under vacuum, then re-dissolved in 5% MeOH/1%
TEA in CH2Cl2 and eluted with the same mixture on a
silica column to afford the pure product as 0.320 g (50%)
of white solid. TLC (EtOAc) Rf ¼ 0:45; LRMS (ESI) m=z
calculated for C21H26N4O5 (MþH)þ 602.2689, found
(m, 2H, 50-NH), 2.15 (dd, 2H, 30-NH2), 2.34 (m, 2H, 20-
H2), 2.67 (m, 2H, 50-H2), 2.97 (m, 1H, 30-H), 3.75 (s, 3H,
MMTr-p-OMe), 4.22 (m, 1H, 40-H), 4.57 (dd, 1H, H5),
5.82 (dd, 1H, 10-H), 6.63–7.17 (m, 15H, MMTr-H), 7.24
(d, 1H, H6), 7.42–7.92 (m, 5H, N4-Bz), 8.16 (s, 1H, N4-H).
1
602.2696. H NMR (400 MHz DMSO-d6) d (ppm) 1.94
1
392.0850. H NMR (400 MHz DMSO-d6) d (ppm) 2.25
(m, 2H, 20-H2), 3.05 (s, 3H, Ms-CH3), 3.52 (m, 1H, 30-
H), 3.62 (m, 2H, 50-H2), 3.94 (m, 1H, 40-H), 4.79 (dd,
1H, 10-H), 5.67 (dd, 1H, H5), 7.21 (d, 1H, H6), 7.38–7.84
(m, 5H, N4-Bz).
3.10. 30,50-Diazido-N4-benzoyl-20,30,50-deoxycytidine (4)
3.13. N4-Benzoyl-30-NH-FmocNCS-50-NH-MMTr-
20,30,50-deoxycytidine (7)
N4-Benzoyl-50-O-mesyl-2,30-anhydro-20-deoxycytidine 3
(1.79 g; 4.58 mmol) and LiN3 (4.48 g; 2 equiv) were dis-
solved in anhydrous DMF (40 mL) at 90 °C while stir-
ring under argon for 4 h, then concentrated to a solid
under vacuum. CH3Cl (50 mL) was added and the
mixture was sonicated, filtered, and the filtrate reduced
to a minimum volume. The filtrate was eluted through a
silica column with CH3Cl and pure fractions were con-
centrated under vacuum to yield 0.680 g (70%) of white
foam. TLC (1:5, MeOH–methylene chloride) Rf ¼ 0:60;
LRMS (ESI) m=z calculated for C22H28N4O5 (MþH)þ
30-Amino-N4-benzoyl-50-NH-MMTr-20,30,50-deoxycyti-
dine 6 (0.320 g; 0.53 mmol) was dissolved in DCM
(20 mL), Fmoc-NCS (0.031 g; 1.5 equiv) was added as a
solid, and the mixture was allowed to stir under argon at
room temperature for 1 h. H2O (20 mL) was added; the
mixture was shaken vigorously and then centrifuged.
The top organic layer was removed and concentrated to
a solid under vacuum, then re-dissolved in EtOAc and
eluted with the same mixture on a silica column, which
had previously been eluted with 5% TEA in EtOAc, then
washed with only EtOAc, to afford the pure product as
0.140 g (30%) of white solid. TLC (1:1, hexanes–EtOAc)
Rf ¼ 0:60; LRMS (ESI) m=z calculated for C52H46N6O6S
(MþH)þ 883.3200, found 883.3243. 1H NMR (400 MHz
DMSO-d6) d (ppm) 2.01 (m, 2H, 50-NH), 2.16 (dd, 1H,
30-NH), 2.30 (m, 2H, 20-H2), 2.67 (m, 2H, 50-H2), 2.97
(m, 1H, 30-H), 3.74 (s, 3H, MMTr-p-OMe), 4.33 (m, 2H,
Fmoc-CH2), 4.20 (m, 1H, 40-H), 4.54 (dd, 1H, H5), 4.72
(m, 2H, Fmoc-CH2), 5.82 (dd, 1H, 10-H), 6.63–7.17 (m,
15H, MMTr-H), 7.24 (d, 1H, H6), 7.20–7.92 (m, 13H,
N4-Bz þ Fmoc-CH), 8.13 (s, 1H, N4-H), 8.22 (d, 1H,
Fmoc-a-H).
1
382.1298, found 382.1278. H NMR (400 MHz DMSO-
d6) d (ppm) 1.68 (m, 2H, 50-H2), 1.98 (m, 1H, 30-H), 2.19
(m, 2H, 20-H2), 3.92 (m, 1H, 40-H), 4.57 (dd, 1H, H5),
5.79 (dd, 1H, 10-H), 7.20 (d, 1H, H6), 7.37–7.83 (m, 5H,
N4-Bz), 8.05 (s, 1H, N4-H).
3.11. 30,50-Diamino-N4-benzoyl-20,30,50-deoxycytidine (5)
30,50-Diazido-N4-benzoyl-20,30,50-deoxycytidine 4 (0.600
g; 1.57 mmol) was dissolved in anhydrous ethanol, a
catalytic amount of activated palladium on carbon was
added, and the mixture hydrogenated (50 psi) for 4 h
then filtered through Celite, and concentrated under
vacuum to yield 0.517 g (100%) of white foam. TLC (1:1,
MeOH–EtOAc) Rf ¼ 0:25; LRMS (ESI) m=z calculated
for C23H30N4O7S (MþH)þ 330.1488, found 330.1479.
1H NMR (400 MHz DMSO-d6) d (ppm) 2.01 (dd, 2H,
30-NH2), 2.26 (m, 2H, 50-NH2), 2.41 (m, 2H, 20-H2), 2.68
(m, 2H, 50-H2), 2.98 (m, 1H, 30-H), 4.22 (m, 1H, 40-H),
4.57 (dd, 1H, H5), 5.82 (dd, 1H, 10-H), 7.22 (d, 1H, H6),
7.42–7.92 (m, 5H, N4-Bz), 8.17 (s, 1H, N4-H).
Acknowledgements
This work was supported by the National Institutes of
Health Grant 5R37DK09171-39. We thank Dr. Renee
Szabo for conversations and advice.
References and notes
3.12. 30-Amino-N4-benzoyl-50-NH-MMTr-20,30,50-deoxy-
cytidine (6)
1. Agrawal, S.; Zhao, Q. Y. Antisense therapeutics. Curr.
Opin. Chem. Biol. 1998, 2(4), 519–528.
2. Demesmaeker, A.; Haner, R.; Martin, P.; Moser, H. E.
Antisense Oligonucleotides. Acc. Chem. Res. 1995, 28(9),
366–374.
3. Zamaratski, E.; Pradeepkumar, P. I.; Chattopadhyaya, J.
A critical survey of the structure-function of the antisense
30,50-Diamino-N4-benzoyl-20,30,50-deoxycytidine 5 (0.350
g; 1.06 mmol) and MMTrCl (0.029 g; 0.9 equiv) were
cooled to ꢁ70 °C under argon then suspended in anhy-
drous TEA (10 mL) with stirring and allowed to come to

I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
4243
oligo/RNA heteroduplex as substrate for RNase H. J.
Biochem. Biophys. Meth. 2001, 48(3), 189–208.
4. Demesmaeker, A.; Altmann, K. H.; Waldner, A.; Wende-
born, S. Backbone modifications in oligonucleotides and
peptide nucleic-acid systems. Curr. Opin. Struct. Biol.
1995, 5(3), 343–355.
5. Nielsen, P. E. Structural and biological properties of
peptide nucleic acid (PNA). Pure Appl. Chem. 1998, 70(1),
105–110.
6. Peyman, A.; Uhlmann, E.; Wagner, K.; Augustin, S.;
Breipohl, G.; Will, D. W.; Schafer, A.; Wallmeier, H.
Phosphonic ester nucleic acids (PHONAs): oligonucleo-
tide analogues with an achiral phosphonic acid ester
backbone. Angew. Chem., Int. Ed. Engl. 1996, 35(22),
2636–2638.
7. Fujii, M.; Yoshida, K.; Hidaka, J.; Ohtsu, T. Nucleic acid
analog peptide (NAAP). 2. Syntheses and properties of
novel DNA analog peptides containing nucleobase linked
beta-aminoalanine. Bioorg. Med. Chem. Lett. 1997, 7(5),
637–640.
8. Summerton, J.; Weller, D. Antisense properties of mor-
pholino oligomers. Nucleosides Nucleotides 1997, 16(7–9),
889–898.
9. Morvan, F.; Sanghvi, Y. S.; Perbost, M.; Vasseur, J. J.;
Bellon, L. Oligonucleotide mimics for antisense therapeu-
tics: Solution phase and automated solid-support synthesis
of MMI linked oligomers. J. Am. Chem. Soc. 1996, 118(1),
255–256.
10. Waldner, A.; Demesmaeker, A.; Lebreton, J.; Fritsch, V.;
Wolf, R. M. Ureas as backbone replacements for the
phosphodiester linkage in oligonucleotides. Synlett 1994,
(1), 57–61.
11. Wang, J. Y.; Matteucci, M. D. The synthesis and binding
properties of oligonucleotide analogs containing diaste-
reomerically pure conformationally restricted acetal link-
ages. Bioorg. Med. Chem. Lett. 1997, 7(2), 229–232.
12. vonMatt, P.; Altmann, K. H. Replacement of the phos-
phodiester linkage in oligonucleotides by heterocycles: The
effect of triazole- and imidazole-modified backbones on
DNA/RNA duplex stability. Bioorg. Med. Chem. Lett.
1997, 7(12), 1553–1556.
of DNA. Proc. Natl. Acad. Sci. U.S.A. 1994, 91(17), 7864–
7868.
20. Barawkar, D. A.; Bruice, T. C. Synthesis, biophysical
properties, and nuclease resistance properties of mired
backbone oligodeoxynucleotides containing cationic
internucleoside guanidinium linkages: deoxynucleic gua-
nidine/DNA chimeras. Proc. Natl. Acad. Sci. U.S.A. 1998,
95(19), 11047–11052.
21. Linkletter, B. A.; Szabo, I. E.; Bruice, T. C. Solid-phase
synthesis of deoxynucleic guanidine (DNG) oligomers and
melting point and circular dichroism analysis of binding
fidelity of octameric thymidyl oligomers with DNA
oligomers. J. Am. Chem. Soc. 1999, 121(16), 3888–3896.
22. Linkletter, B. A.; Szabo, I. E.; Bruice, T. C. Solid-phase
synthesis of oligopurine deoxynucleic guanidine (DNG)
and analysis of binding with DNA oligomers. Nucl. Acids
Res. 2001, 29(11), 2370–2376.
23. Kojima, N.; Bruice, T. C. Replacement of the phospho-
rodiester linkages of RNA with guanidinium linkages: the
solid-phase synthesis of ribonucleic guanidine. Org. Lett.
2000, 2(1), 81–84.
24. Kojima, N.; Szabo, I. E.; Bruice, T. C. Synthesis of
ribonucleic guanidine: replacement of the negative phos-
phodiester linkages of RNA with positive guanidinium
linkages. Tetrahedron 2002, 58(5), 867–879.
25. Dempcy, R. O.; Browne, K. A.; Bruice, T. C. Synthesis of
the polycation thymidyl DNG, its fidelity in binding
polyanionic DNA/RNA, and the stability and nature of
the hybrid complexes. J. Am. Chem. Soc. 1995, 117(22),
6140–6141.
26. Gait, M. J. In Oligonucleotide Synthesis: A Practical
Approach; IRL: Oxford Oxfordshire; Washington, DC,
1984; Vol. p 13, p 217.
27. Horwitz, J. P.; Chua, J.; Noel, M.; Donatti, J. T.
Nucleosides. 11. 20,30-dideoxycytidin. J. Org. Chem.
1967, 32(3), 817.
28. Kearney, P. C.; Fernandez, M.; Flygare, J. A. Solid-phase
synthesis of 2-aminothiazoles. J. Org. Chem. 1998, 63(1),
196–200.
29. Simonsson, T. G-quadruplex DNA structures––variations
on a theme. Biolog. Chem. 2001, 382(4), 621–628.
30. Gilbert, D. E.; Feigon, J. Multistranded DNA structures.
Curr. Opin. Struct. Biol. 1999, 9(3), 305–314.
31. Zarudnaya, M. I.; Samijlenko, S. P.; Potyahaylo, A. L.;
Hovorun, D. M. Structural transitions in polycytidylic
acid: proton buffer capacity data. Nucleosides Nucleotides
Nucl. Acids 2002, 21(2), 125–137.
32. Gueron, M.; Leroy, J.-L. The i-motif in nucleic acids.
Curr. Opin. Struct. Biol. 2000, 10(3), 326–331.
33. Saenger, W.. In: Principles of Nucleic Acid Structure;
Springer: New York, 1984; Vol. p 20, p 556.
34. Mergny, J.; Lacroix, L. Kinetics and thermodynamics of i-
DNA formation: phosphodiester versus modified oligode-
oxynucleotides. Nucl. Acids Res. 1998, 26(21), 4797–4803.
35. Job, P. Formation and stability of inorganic complexes in
solution. Ann. Chim. 1928, 9, 113–203.
36. Tinoco, I. Nucleic acid structures, energetics, and dynam-
ics. J. Phys. Chem. 1996, 100(31), 13311–13322.
37. Luo, J.; Bruice, T. C. Nanosecond molecular dynamics of
hybrid triplex and duplex of polycation deoxyribonucleic
guanidine strands with a complimentary DNA strand. J.
Am. Chem. Soc. 1998, 120(6), 1115–1123.
13. Kurreck, J.; Wyszko, E.; Gillen, C.; Erdmann, V. A.
Design of antisense oligonucleotides stabilized by locked
nucleic acids. Nucl. Acids Res. 2002, 30(9), 1911–1918.
14. Ueno, Y.; Mikawa, M.; Matsuda, A. Nucleosides and
nucleotides. 170. Synthesis and properties of oligodeoxy-
nucleotides
containing
5-N-2-N,N-bis(2-amino-
ethyl)amino ethyl carbamoyl-20-deoxyuridine and 5-N-3-
N,N-bis(3-aminopropyl)amino propyl carbamoyl-20-deoxy-
uridine. Bioconjug. Chem. 1998, 9(1), 33–39.
15. Manoharan, Muthiah; Guinosso, C. J.; Dan Cook, P.
Novel functionalization of the sugar moiety of nucleic
acids for multiple labeling in the minor groove. Tetrahe-
dron Lett. 1991, 32(49), 7171–7174.
16. Skibo, E. B.; Xing, C. G. Chemistry and DNA alkylation
reactions of aziridinyl quinones: Development of an
efficient alkylating agent of the phosphate backbone.
Biochemistry 1998, 37(43), 15199–15213.
17. Jung, P. M.; Histand, G.; Letsinger, R. L. Hybridization
of alternating cationic/anionic oligonucleotides to RNA
segments. Nucleosides Nucleotides 1994, 13(6–7), 1597–
1605.
18. Chaturvedi, S.; Horn, T.; Letsinger, R. L. Stabilization of
triple-stranded oligonucleotide complexes: use of probes
containing alternating phosphodiester and stereo-uniform
cationic phosphoramidate linkages. Nucl. Acids Res. 1996,
24(12), 2318–2323.
38. MathWorks Matlab, MATLAB Version 6.5 Release 13; The
MathWorks, Inc.: 2002.
39. Henry, E. R.; Hofrichter, J. Singular value decomposi-
tion––application to analysis of experimental-data. Meth.
Enzymol. 1992, 210, 129–192.
19. Dempcy, R. O.; Almarsson, O.; Bruice, T. C. Design and
synthesis of deoxynucleic guanidine––a polycation analog
40. Hendler, R. W.; Shrager, R. I. Deconvolutions based on
singular-value decomposition and the pseudoinverse––a

4244
I. E. Szabo, T. C. Bruice / Bioorg. Med. Chem. 12 (2004) 4233–4244
guide for beginners. J. Biochem. Biophys. Meth. 1994,
28(1), 1–33.
41. Haq, I.; Chowdhury, B. Z.; Chaires, J. B. Singular value
decomposition of 3-D DNA melting curves reveals com-
plexity in the melting process. Eur. Biophys. J. Biophys.
Lett. 1997, 26(6), 419–426.
42. Murchie, A. I. H.; Aboul-ela, F.; Laughlan, G.; Luisi, B.;
Lilley, D. M. J. Structure of parallel-stranded guanine
tetraplexes. Nucl. Acids Mol. Biol. 1995, 9, 143–164.
43. Pinnavaia, T. J.; Marshall, C. L.; Mettler, C. M.; Fisk, C.
I.; Miles, H. T.; Becker, E. D. Alkali-metal ion specificity
in solution ordering of a nucleotide, 50-guanosine mono-
phosphate. J. Am. Chem. Soc. 1978, 100(11), 3625–
3627.
44. Marky, L. A.; Breslauer, K. J. Calculating thermodynamic
data for transitions of any molecularity from equilib-
rium melting curves. Biopolymers 1987, 26, 1601–
1620.