Outstanding effects on antithrombin activity of modified TBA diastereomers containing an optically pure acyclic nucleotide analogue

Article abstract of DOI:10.1039/c4ob00149d

Herein, we report optically pure modified acyclic nucleosides as ideal probes for aptamer modification. These new monomers offer unique advantages in exploring the role played in thrombin inhibition by a single residue modification at key positions of the TBA structure.

Full text of DOI:10.1039/c4ob00149d

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Outstanding effects on antithrombin activity of
modified TBA diastereomers containing an
optically pure acyclic nucleotide analogue†
Cite this: Org. Biomol. Chem., 2014,
12, 5235
M. Scuotto,‡ M. Persico,‡ M. Bucci, V. Vellecco, N. Borbone, E. Morelli, G. Oliviero,
E. Novellino, G. Piccialli, G. Cirino, M. Varra,* C. Fattorusso* and L. Mayol
Received 20th January 2014,
Accepted 16th May 2014
Herein, we report optically pure modified acyclic nucleosides as ideal probes for aptamer modification.
These new monomers offer unique advantages in exploring the role played in thrombin inhibition by a
single residue modification at key positions of the TBA structure.
DOI: 10.1039/c4ob00149d
www.rsc.org/obc
The investigation of the complex and intriguing molecular
Introduction
bases of TBA–thrombin interactions has lasted now for more
Aptamers are short, structured, single-stranded RNA or DNA than two decades.11–14 There is
a certain amount of
ligands characterized by high affinity and specificity for their knowledge4,8,9,11–15 on the fact that TBA loops are fundamental
target molecules. Aptamers can be considered as oligonucleo- for both G-quadruplex folding and thrombin interaction. In
tide analogs of antibodies,1,2 but, unlike antibodies, they can TBA experimentally determined structures, either alone8,9 or
be chemically derivatized to extend their lifetimes and/or their in complex with thrombin,11–14 the TT loops always preserved
bioavailability3,4 and, particularly, to explore the molecular the same conformation, having T4 and T13 stacked on the
basis of their biological properties.5,6 The thrombin binding underlying G quartet and T3 and T12 available for thrombin
aptamer (TBA) is a 15-mer oligonucleotide (ON) which folds interaction. In contrast, the TGT loop showed higher confor-
into a typical chair-like G-quadruplex structure containing one mational flexibility, especially concerning the T7 residue,
TGT and two TT loops (Fig. 1).7–9 It binds thrombin acting as which is reported to interact with thrombin in some X-ray
an anticoagulant agent.10
complexes.11,12 Previous structure–activity relationship (SAR)
studies on TBA analogues containing an acyclic nucleotide,
performed by us16,17 and others,18,19 confirmed that some
loop residues, specifically T4, T13, G8, and T9, are crucial to
preserve the G-quadruplex folding typology, whereas the
others, T3, T12, and T7, are uniquely involved in thrombin
inhibition. Indeed, ONs modified at T3, T12, and T7 showed
significantly different biological behaviours with respect to
TBA, not related to the stability of their G-quadruplex struc-
ture. We rationalized17 the SARs obtained by substituting a or
b (Fig. 2) at T3, T12, and T7, through enzyme-aptamer multiple
binding modes (Fig. 3). Disadvantageously, since the intro-
duced residues a or b presented a pro-chiral carbon (Fig. 2),
each obtained modified sequence existed as a mixture of two
diastereomers.
Fig. 1 Schematic representation of the NMR derived structure of TBA.
White and black rectangles indicate G bases in anti and syn confor-
mation, respectively. Loop residues not stacked on the G tetrads are
highlighted by red circles.
On this basis, the substitution of the identified key residues
with a rationally designed and optically pure acyclic nucleoside
analogue would allow the investigation of the molecular bases
of TBA–thrombin interaction. Here, we successfully approach
this issue by the substitution of T3 or T7 or T12 with each
pure stereoisomer of the acyclic nucleoside c (Fig. 2).
Dipartimento di Farmacia, via D. Montesano 49, 80131 Napoli, Italy.
E-mail: varra@unina.it, cfattoru@unina.it; Fax: +39 081678552; Tel: +39 081678540,
+39 081678544
†Electronic supplementary information (ESI) available: Fig. S1–S3 and Tables S1
and S2. See DOI: 10.1039/c4ob00149d
‡These authors equally contributed to this work.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 5235–5242 | 5235

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 2 Acyclic nucleosides mimicking thymine residues. The pro-chiral
carbon of a and b is labelled with #.
Fig. 4 CD spectra of modified acyclic nucleosides (3.6 × 10−4 M in
CH3OH) R-c (positive spectrum) and S-c (negative spectrum).
Table 1 Modified sequences containing the c derivatives instead of a T
residue. The incorporation of R-c or S-c stereoisomer was specified sub-
stituting, at each position, the letter T of the unmodified sequence for
the letter R or S, respectively
Fig. 3 Schematic representation of the four possible orientations of
TBA with respect to thrombin assuming the same strand polarity of the
X-ray complex. White and black rectangles indicate G bases in anti and
syn conformation, respectively.
ON
Sequence
TBA
GGTTGGTGTGGTTGG
GGTTGGTGTGGR12TGG
GGR3TGGTGTGGTTGG
GGTTGGR7GTGGTTGG
GGTTGGS7GTGGTTGG
GGTTGGTGTGGS12TGG
GGS3TGGTGTGGTTGG
Results and discussion
Synthesis
TBA-R12
TBA-R3
TBA-R7
TBA-S7
TBA-S12
TBA-S3
We adopted the Horhota method20 to obtain the stereo-
selective synthesis of R and S enantiomers of the acyclic
nucleoside c (R-c and S-c in Scheme 1, Fig. 2 and Fig. 4). The R
or S stereochemistry was attained by ring opening of the enan-
tiomerically pure chiral epoxides 2 by reaction with the modi-
fied nucleobase 1 (Scheme 1). Each achieved molecule was
converted into the corresponding phosphoramidite building
block (5a,b) and incorporated into the TBA sequence to obtain
a pool of six ONs (Table 1).
As expected, in PBS and K+ (90 mM KCl, 10 mM K2HPO4,
pH = 7.3) buffers, the modified sequences fold into G-quadru-
plexes having CD profiles (Fig. 5) and apparent melting temp-
eratures (Table 2 and Fig. S1 in ESI†) comparable to that of
TBA. Particularly, a stabilizing effect on the G-quadruplex
structure (ΔT = 4–5 °C) was attained by replacing the T residue
Structural characterization
To probe whether the introduction of R or S acyclic nucleoside
c at TBA position 3, 12 or 7 affected the overall G-quadruplex
topology formed in solution, we characterized all modified
TBAs by circular dichroism (CD) and CD melting experiments.
Scheme 1 Synthesis of monomers 5a,b. (i) 1 1.0 g (6.6 mmol), 2 (pure
enantiomer R or S) 0.490 g (6.6 mmol), potassium carbonate 0.152 g
(1.1 mmol), dry DMF (24 mL), 12–18 h, 80 °C, yields R-c 60% and S-c
45%; (ii) R-c or S-c 560 mg (2.5 mmol), 4,4’-dimethoxytritylchloride
850 mg (2.5 mmol), 4-dimethylaminopyridine 15 mg (0.12 mmol), pyri-
dine (20 mL), 2.0 h, r.t., yields 85%; (iii) 4a,b 859 mg, (1.6 mmol),
2-cyanoethyl diisopropylchlorophosphoramidite 536 µL (2.4 mmol),
DIPEA 1.7 mL (10 mmol), DCM (9 mL), 40 min, r.t., yields 90%.
Fig. 5 CD spectra of TBA and modified ONs in PBS (A) and K+ phos-
phate buffer (B).
5236 | Org. Biomol. Chem., 2014, 12, 5235–5242
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 2 Tm values of TBA and modified ONs
Tm °C ( 1.0 °C)
K+
ON
PBS
TBA
50
55
55
54
56
51
51
34
39
39
38
39
35
35
TBA-R12
TBA-R3
TBA-R7
TBA-S7
TBA-S12
TBA-S3
at positions 3 and 12 with the R stereoisomer of c or at posi-
tion 7 with both the R and S stereoisomers.
Fig. 7 Prolonged fibrinogen clotting time (s) in the presence of TBA,
TBA-S3, or TBA-S12 at different concentrations (30–200 nM). The
graphics were drawn from scattering curve data.
Fibrinogen assay
The new ONs were then evaluated for their ability to inhibit
fibrinogen hydrolysis by a fibrinogen clotting assay. The clot-
ting of fibrinogen induced by human α-thrombin was
measured spectrophotometrically at different aptamer concen-
trations, following the increase in absorbance at 380 nm as a
function of time (Fig. 6 and 7, Table 3, Fig. S2 in ESI,† and
Experimental section).
At 20 nM concentration, TBA and TBA-R12 prolonged the
basal clotting time of 56 and 199 s, respectively (Fig. 6 and
Table 3).
Table 3 Prolonged fibrinogen clotting times of TBA and modified ONs.
Each value was calculated subtracting the clotting time produced by
thrombin in the absence of any inhibitor (i.e., 25.6
1.5 s) from that
measured in the presence of the aptamer
Prolonged clotting time (s)
ON
15 nM
20 nM
30 nM
100 nM
200 nM
TBA
24.6 2.7 56.4 5.0
159.4 11.3 256.4 19.7 N.T.a
Thus, TBA-R12 was 3.5 fold more active than TBA. At the
same concentration, TBA-S12 was completely inactive (Fig. 7
and Table 3). A similar activity trend was also observed
between TBA-R3 and TBA-S3, the former being 1.9 fold more
active than TBA, and the latter completely inactive (Fig. 6 and
7 and Table 3). These activity trends were maintained at all
tested concentrations, thus evidencing a stereoselective inter-
action with the target, involving the modified T12 or T3 resi-
dues, which is crucial for thrombin inhibition.
TBA-R12 105.4 5.7 198.9 10.1 463.4 13.6 759.6 23.8 N.T.
TBA-R3 50.4 4.0 108.4 5.1 323.4 12.5 576.4 22.8 N.T.
TBA-R7 49.4 3.0 89.4 3.8
TBA-S7 39.5 3.0 67.4 4.0
167.4 6.0 304.4 19.9 N.T.
181.4 9.7 355.4 18.0 N.T.
TBA-S12 N.A.b
TBA-S3 N.A.
N.A.
N.A.
10.9 5.7
4.9 3.0
16.4 2.8
18.4 5.1
24.4 3.7
26.4 4.6
a Not tested. b Not active.
mental section for details). As observed by Feigon et al.8 and
confirmed by X-ray data,12,14 due to the TBA symmetry, resi-
dues T3 and T12 occupy equivalent positions with respect to
the G quartets and, by consequence, their binding clefts at
thrombin Exosite I (ABE I) can be “exchanged” by rotating the
structure 180° around the Y-axis (binding modes III and IV in
Fig. 3 and 8). Conversely, the TBA symmetry is lost when either
T3 or T12 is modified and, in particular, TT loops are no
longer equivalent. In our case, the monomer c is characterized
by (i) higher conformational flexibility and (ii) greater steric
hindrance with respect to thymidine. Our computational
results indicated that TBA-R3 and TBA-R12, which preserved
the stereochemistry of the thymidine residue, could easily
adopt bioactive conformation of the TT loops (Fig. 9a). The sig-
nificant increase in inhibitory activity of TBA-R3 and TBA-R12
with respect to TBA indicates the presence of additional
favourable hydrophobic contacts with the protein in, at least,
one of the TT loop binding clefts. In contrast, when the stereo-
chemistry of c is inverted with respect to the C-4′ carbon atom
of the natural counterpart (i.e., TBA-S3 and TBA-S12), it is not
possible for both binding sites and TT loops to properly place
at their thrombin whatever binding mode is assumed. Indeed,
Molecular modelling studies
The resulting SARs of the new modified ONs were rationalized
with the aid of molecular modelling studies (see the Experi-
Fig. 6 Prolonged fibrinogen clotting time (s) in the presence of TBA or
each active aptamer at different concentrations (15–100 nM). The
graphics were drawn from scattering curve data.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 5235–5242 | 5237

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 10 (a) TBA-R7 (carbons = orange) and (b) TBA-S7 (carbons = pink)
superimposed on the bioactive conformation of TBA in complex with
thrombin (carbons = green; PDB ID: 1HAP). Heteroatoms are coloured
as follows: O = red; N = blue; P = magenta. Thrombin ABE I is shown as
a green ribbon. Hydrogens are omitted with the exception of those
present in the modified residue chiral carbon (R-c12 and S-c12) and
in TBA T7 C4’.
Fig. 8 X-ray structures of the complex between TBA and human
α-thrombin presenting aptamer TT loops interacting with thrombin ABE
I. (a) and (c) Binding mode III; PDB structure 1HAO (TBA carbons = cyan).
(b) and (d) Binding mode IV; PDB structure 4DIH (TBA carbons = yellow).
Heteroatoms are coloured as follows: O = red; N = blue; P = magenta.
Thrombin ABE I is shown as a green ribbon.
factor able to modulate the inhibition of thrombin. Further
considerations can be done taking into account the results
obtained with aptamers modified at position 7 of the TGT
loop. The substitution of T7 with the R or S enantiomer of c
produced almost equally active aptamers (Fig. 2 and Table 3).
On the one hand, the lack of significant stereoselectivity of
action between the two diastereomers would suggest that the
TGT loop is not directly involved in thrombin binding. On the
other hand, TBA-R7 and TBA-S7 showed increased inhibitory
activity (either with respect to TBA and to previously developed
analogues bearing the monomer b)17 which cannot be
ascribed to the increased stability of the quadruplex structure
(Tm (PBS) TBA = TBA-T7b17 = 37.5 °C; TBA-R7 = 38 °C; TBA-S7 =
39 °C), thus supporting the possible direct involvement of the
TGT loop in thrombin inhibition.11,12,17,18 Our molecular mod-
elling studies showed that, contrary to what was observed for
the more rigid TT loops, the higher conformational flexibility
of the TGT loop allows both c enantiomers to adopt the bio-
active conformation11,12 of the residue at position 7, without
affecting the height of P7–P9 phosphate groups from the
underlying G tetrad plane (Fig. 10). In particular, most TBA-R7
conformers presented, similarly to TBA bioactive confor-
mation, G8 stacked on the G1–G6–G10–G15 quartet and T9
Fig. 9 (a) TBA-R12 (carbons = orange) and (b) TBA-S12 (carbons = pink)
superimposed on the bioactive conformation of TBA in complex with
thrombin (carbons = cyan; PDB ID: 1HAO). Heteroatoms are coloured as
follows: O = red; N = blue; P = magenta. Thrombin ABE I is shown as a
green ribbon. Hydrogens are omitted with the exception of those
present in the modified residue chiral carbon (R-c12 and S-c12) and
in TBA T12 C4’. The position of P13 is highlighted by a dashed circle and
an arrow.
although monomers c are characterized by higher structural pointing outside (Fig. 10a and Table S1 in ESI†).
flexibility with respect to thymine, TT loop conformation is
The opposite trend was observed for TBA-S7 (Fig. 10b and
highly limited by the stacking of T4 and T13 on the G-quartet Table S1 in ESI†). Analyzing our results in the light of reported
(Table S1 in ESI†). Therefore, when the nucleobase analogue TBA–thrombin binding modes and X-ray structures,11–14 it
moiety of c in TBA-S3 and TBA-S12 projects in the same direc- resulted that the two different TGT loop conformations can
tion as the nucleobase of T3 and T12 in the bioactive confor- assume the binding modes I (Fig. S3a and S3c in ESI†) and II
mation of TBA,11,12,14 the succeeding phosphate group (i.e., P4 (Fig. S3b and S3d in ESI†) represented in Fig. 3.
or P13) is forced to distance itself from the G-quartet plane,
Taken together, our results indicate that both TT loops and
likely causing steric clashes with the protein backbone (Fig. 9b). the TGT loop are involved in thrombin inhibition, but play
Another important outcome which emerged from our different roles. The TT loops are necessary for high affinity
results is the higher inhibitory activity of TBA-R12 compared thrombin inhibition by establishing a stereoselective inter-
to TBA-R3.
action with the protein through residues at positions 3 and 12.
In this case, the lack of structural symmetry does not In contrast, the TGT loop seems to be involved in less specific
explain the difference in activities of TBA-R12 and TBA-R3, interactions with thrombin. Indeed, the TGT loop is not
since the same TT loop binding mode at thrombin ABE I could sufficient for significant thrombin binding when the binding
be achieved for both analogues by rotating the structure by
ability of the TT loops is impaired (TBA-S3 and TBA-S12). More-
180° (Fig. 8). This indicates that, although TT loops–ABE I over, no significant stereoselectivity of action was observed
interaction is necessary for biological activity, it is not the only between TBA-R7 and TBA-S7. The results of our molecular
5238 | Org. Biomol. Chem., 2014, 12, 5235–5242
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
modelling studies suggest that the flexibility of the TGT loop, 64.0 mmol), urea (5.76 g, 96 mmol) and HCl (37% w/w,
together with the availability of multiple aptamer-protein 0.096 mL) in EtOH (20 mL) was refluxed for 3 h. After fil-
binding modes, likely contributes to the observed lack of
stereoselectivity between TBA-S7 and TBA-R7. Anyway, a single
modification of the TGT loop at position 7 can enhance throm-
bin inhibition, through a still undefined molecular mechan-
ism. A fascinating hypothesis, supported by some available
X-ray data,11,12 is that the high affinity binding of the TT loops
could drive the interaction of the TGT loop with a second unit
of thrombin, and this, in turn, due to thrombin allostery,
somehow favours the binding of another TBA monomer in a
cooperative mechanism.
tration, the collected white solid was suspended in 5% NaOH
(24 mL) and refluxed for 1 h. The reaction was cooled to room
temperature and the solid 1 collected by filtration and dried
(yields 60%; Rf 0.42 in 9 : 1 CH2Cl2–CH3OH, v/v).
1H (400 MHz, mixture of CD3OD and CDCl3): δ 4.00 (bs,
2H), 3.62 (m, 4H), 3.03 (t, J = 7.3 Hz, 2H), 2.67 (t, J = 7.3 Hz,
2H), 2.10 (q, J = 7.3 Hz, 2H).
13C (400 MHz, mixture of CD3OD and CDCl3): δ 166.3,
155.2, 147.3, 113.4, 34.8, 30.5, 18.0.
ESI mass (positive mode): calculated 152.1; found 153.1
[M + H]+, 175.1 [M + Na]+.
1-(2,3-Dihydroxypropyl)-6,7-dihydro-1H-cyclopenta[d]pyrimidine-
2,4(3H,5H)-dione [S-c or R-c].20 A mixture of compound 1 (1.0 g,
6.6 mmol) and anhydrous potassium carbonate (0.152 g,
1.1 mmol) in dry DMF (24 mL) was heated at 80 °C for 5 min.
Compound 2 (pure enantiomer R or S) (0.490 g, 6.6 mmol) was
then added. The reaction mixture was stirred at 80 °C for 18 h
under argon. The solution was concentrated under reduced
pressure and the residue purified by column chromatography
on silica gel eluted with 90 : 10 CH2Cl2–CH3OH to give S-c or
R-c as white solids (yields S-c 45%, R-c 60%, each of them
calculated as the average of two different reactions). Fig. S1†
reports the CD profiles of each pure enantiomer S-c or R-c.
Experimental
Materials and methods
Chemicals and anhydrous solvents were purchased from
Fluka-Sigma-Aldrich. TLCs were run on Merck silica gel 60
F254 plates. Silica gel chromatography was performed using
Merck silica gel 60 (0.063–0.200 mm). The API 2000 (Applied
Biosystems) mass spectrometer was used to perform the ana-
lyses of the intermediates and the monomer. Melting points of
intermediates R-c and S-c were measured using a Buchi
Melting Point B-540. NMR experiments were recorded using
Varian Mercury Plus 400 MHz and Unity Inova 500 MHz spec-
trometers and processed using the Varian VNMR software
package. NMR spectra were calibrated using the solvents’
residual proton or carbon signals as internal standards.
31P-NMR spectra were calibrated using triphenylphosphine as
an external standard (δ −6 ppm). Reagents and phosphorami-
dites for DNA syntheses were purchased from Glenn Research.
ON syntheses were performed on a PerSeptive Biosystem Expe-
dite DNA synthesizer. HPLC analyses and purifications were
carried out using a JASCO PU-2089 Plus HPLC pump equipped
with a JASCO BS-997-01 UV detector. CD experiments were per-
formed on a JASCO 715 spectropolarimeter equipped with a
PTC-348 temperature controller. The fibrinogen assay was per-
formed using a JASCO 530 UV spectrophotometer equipped
with the PTC-348 temperature controller. [α]2D0 values for R-c
and S-c were determined using a JASCO P-2000 polarimeter.
Molecular modelling calculations were performed on SGI
Origin 200 8XR12000 and E4 Server Twin 2× Dual Xeon 5520,
equipped with two nodes. Each node was 2× Intel Xeon Quad-
Core E5520, 2.26 GHz, 36 GB RAM. The molecular modelling
graphics were carried out on SGI Octane 2 workstations.
MALDI-ToF measurements were performed using a Voyager
DE-STR instrument (Applied Biosystems). High resolution
mass spectra of 5a and 5b were performed on a Thermo LTQ
Orbitrap XL mass spectrometer (ESI positive mode).
1
S-c H NMR (400 MHz, pyridine-d5): δ 4.66 (m, 1H, CHOH),
4.45 (dd, 1H, J1 = 14.0 Hz, J2 = 3.6 Hz, CHaHbOH), 4.08 (d, 2H,
J = 5.2 Hz, CH2N), 3.93 (dd, 1H, J1 = 14.0 Hz, J2 = 8.6 Hz
CHaHbOH), 3.17 (m, 1H, CHaHb), 2.83 (m, 1H, CHaHb), 2.69
(m, 2H, CH2), 1.82 (m, 2H) (Fig. S4†).
S-c 13C NMR (100 MHz, CD3OD): δ 163.0, 160.4, 153.2,
111.6, 69.4, 63.9, 32.6, 27.0, 21.0 (Fig. S5†).
ESI mass (positive mode): calculated 226.1; found 227.1
[M + H]+, 249.1 [M + Na]+. [α]2D0 = –51.4.
S-c in the crystalline state was obtained from CH3COCH3–
CH3OH (99 : 1, v : v). The measured melting point was in the
range 141–143 °C.
R-c 1H NMR (400 MHz, pyridine-d5): δ 4.66 (m, 1H, CHOH),
4.45 (dd, 1H, J1 = 14.0 Hz, J2 = 3.6 Hz, CHaHbOH), 4.08 (d, 2H,
J = 5.2 Hz, CH2N), 3.93 (dd, 1H, J1 = 14.0 Hz, J2 = 8.6 Hz
CHaHbOH), 3.17 (m, 1H, CHaHb), 2.83 (m, 1H, CHaHb), 2.69
(m, 2H, CH2), 1.82 (m, 2H) (Fig. S6†).
R-c 13C NMR (100 MHz, CD3OD) 163.0, 160.4, 153.2, 111.6,
69.4, 63.9, 32.6, 27.0, 21.0 (Fig. S7†).
ESI mass (positive mode): calculated 226.1; found 227.1
[M + H]+, 249.1 [M + Na]+. [α]2D0 = +54.3.
R-c in the crystalline state was obtained from CH3COCH3–
CH3OH (99 : 1, v : v). The measured melting point was in the
range 141–143 °C.
1-(2,4-Dioxo-3,4,6,7-tetrahydro-2H-cyclopenta[d]pyrimidin-
1(5H)-yl)-3-((3-methoxyphenyl)(4-methoxyphenyl)(phenyl)methoxy)-
propan-2-yl-P-2-cyanoethyl-N,N-cianopropylphosphonamidate 5a
and 5b. Compounds R-c or S-c (560 mg, 2.5 mmol), 4,4′-
dimethoxytrityl chloride (850 mg, 2.5 mmol) and 4-dimethyl-
Synthesis procedure
Synthesis of monomers 5a and 5b
6,7-Dihydro-1H-cyclopenta[d]pyrimidine-2,4(3H,5H)-dione
(1). 1 was obtained as previously reported by Takaya et al.21 aminopyridine (15.0 mg, 0.12 mmol) were dissolved in dry
Briefly, a solution of ethyl-2-oxocyclopentanecarboxylate (10 g, pyridine (20 mL). The resulting solution was stirred at room
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 5235–5242 | 5239

View Article Online
Paper
Organic & Biomolecular Chemistry
temperature under argon for 1.5 h. Dry methanol (200 μL) was 20 mM NaH2PO4 aqueous solution containing 20% CH3CN; B:
then added to quench the reaction. After 30 min under stirring 1.0 M NaCl, 20 mM NaH2PO4 aqueous solution containing
the solution was concentrated under reduced pressure and the 20% CH3CN, elution time 15.3 min). The oligomers were suc-
residue purified by column chromatography on silica gel cessively desalted by molecular exclusion chromatography on
(eluted with 95 : 5 : 1 CH2Cl2–CH3OH–Et3N) to give mono- Biogel P-2 Fine. The purity was checked on HPLC using an
dimethoxytritylated derivative 4a or 4b as a white foam (1.16 g, analytical reverse phase column (Purosphere RP-18, Merck).
85% yield from each c stereoisomer; Rf 0.7 eluted with Finally, the sequences were analyzed in negative mode by MAL-
CH2Cl2–CH3OH 95 : 5 v/v). The solid (859 mg, 1.6 mmol) was DI-TOF experiments. Calculated molecular weight: [4710.25];
dried in a vacuum overnight before being dissolved in measured molecular weight: TBA-R3 [4710.72]; TBA-R7 [4710.94];
anhydrous DCM (9 mL) and diisopropylethylamine (1.7 mL, TBA-R12 [4710.54]; TBA-S3 [4710.11]; TBA-S7 [4711.57]; TBA-S12
10 mmol) under argon. 536 μL of β-cyanoethyl diisopropyl- [4710.72].
chlorophosphoramidite was then added (2.4 mmol). After
The concentrations of the samples used in CD and UV
40 min the reaction was quenched by addition of dry methanol experiments were determined by measuring the absorbance at
(100 μL), diluted with ethyl acetate (15 mL) and finally washed 260 nm at 80 °C and using the open access program available
with 10% sodium carbonate solution (15 mL) and brine on http://basic.northwestern.edu/biotools/OligoCalc.html.22
(15 mL). The organic layer was dried on magnesium sulphate
CD experiments
and concentrated in a vacuum. The residue was purified by
silica gel chromatography eluted with n-hexane, ethyl acetate, To perform the CD experiments on modified nucleosides R-c
and triethylamine (80 : 10 : 10). The fractions containing the and S-c each compound was dissolved in CH3OH at the
product were collected and concentrated under vacuum yield- final concentration of 3.6 × 10−4 M. The two solutions were
ing the phosphoramidite building block 5a [from R-c] or 5b equilibrated for 10 min at r.t. and then the CD spectra were
[from S-c] as a white foam (1.05 g, 90% Rf 0.75 in CHCl3– registered.
MeOH–TEA 97 : 3 : 0.05 v/v). HPLC chromatograms (Fig. S14
To perform the CD experiments on modified TBA
and S15†) were used to confirm yields of the conversions of 4a sequences, each ON was dissolved in the potassium (90 mM
and 4b to 5a and 5b. After the washing procedure, the concen- KCl, 10 mM KH2PO4, pH 7.4) or PBS (90 mM KCl, 10 mM
trated organic layers were dissolved in DCM and analyzed on a KH2PO4, pH 7.4) phosphate buffer at the final ON concen-
Hypersil silica gel column, particle size 5 μm, eluted with tration of 2.0 × 10−5 M and submitted to the annealing pro-
50 : 50 n-hexane–ethyl acetate. Rt: 5a 15.1 and 5b 14.7 [min]. cedure (heating at 90 °C and slowly cooling at r.t.). Before each
The yields of reactions, calculated as the ratio between the area experiment the samples were equilibrated at 10 °C for 30 min.
of the peak produced by 5a or 5b and the total area of peaks, CD spectra were recorded from 200 to 400 nm at 100 nm
were 92 and 89%, respectively.
min−1 scanning rate, 16 s response, 1.0 nm bandwidth. Each
1H NMR (500 MHz, DMSO-D6): δ 11.05 (1H), 7.41–7.24 CD profile was obtained by taking the average of three scans
(9H), 6.87 (4H), 4.18 (m, 1H), 3.93 (m, 1H), 3.78 (s, 6H), from which the spectrum of the background buffer was
3.78–3.40 (m, 3H), 3.14–2.85 (m, 3H), 2.65 (m, 1H), 2.49 (m, subtracted.
4H), 1.92 (m, 2H), 1.08–0.97 (m, 12H) (Fig. S8 and S9†).
CD melting–folding curves were obtained by monitoring
13C-NMR (100 MHz, DMSO-D6): δ 162.0, 159.0, 158.6, 153.0, the variation of absorbance at 295 nm from 10 to 90 °C and
145.5, 136.2, 130.5, 128.7, 128.5, 127.5, 119.5, 114.0, 111.5, vice versa in K+ phosphate buffer (80 mM KCl, 20 mM KH2PO4
86.4, 70,6, 65.2, 59,0, 55.9, 49.0, 43.3, 32.9, 28.0, 25.2–24.9, pH = 7.4) and PBS (147 mM NaCl, 20 mM NaH2PO4, 23 mM
21.2, 20.5 (Fig. S10 and S11†).
KCl, pH = 7.4) at [ON] = 2.0 × 10−5 M using a 1.0 cm cuvette.
31P-NMR (202 MHz, DMSO): δ 148.9, 148.2 (Fig. S12 and Two melting–folding experiments for each ON were recorded
S13†).
at 0.5 °C min−1 and 1.0 °C min−1 heating–cooling rate. No sub-
HR-ESI mass (positive mode): calculated 728.3339; found stantial differences between melting and cooling curves were
751.3216 [M + Na]+. Results of mass spectra are shown in detected. Both types of curves were independent of tempera-
Fig. S16 and S17.†
ture scanning rate (0.5 °C min−1 or 1.0 °C min−1). Fig. S2†
5a [α]2D0 = +55.3 (solvent CH3CN); 5b [α]2D0 = –45.7 (solvent shows the melting curves registered at the temperature
CH3CN).
Synthesis of oligomers. TBA and analogues were syn-
thesized using standard solid phase DNA chemistry on a
scanning rate of 1.0 °C min−1
.
Fibrinogen clotting assay
controlled pore glass (CPG) support following the β-cyanoethyl The fibrinogen clotting times were measured spectrophoto-
phosphoramidite method. The oligomers were detached from metrically.23 ONs were incubated for 1 min at 37 °C in 1.0 mL
the support and deprotected by treatment with an aqueous of PBS containing 2.0 mg mL−1 of fibrinogen (fibrinogen from
ammonia solution (33%) at 55 °C overnight. The combined fil- human plasma, F 3879, Sigma-Aldrich) in a PMMA cuvette
trates and washings were concentrated under reduced (vol. 1.5 mL, c.o. 1 cm, Brand). 100 μL of human thrombin
pressure, dissolved in H2O, and purified by HPLC using an (10 NIH per mL; Sigma-Aldrich, T8885, human thrombin suit-
anionic exchange column eluted with a linear gradient (from able for the thrombin time test) was then added to the solu-
0% to 100% B in 30 min) of phosphate buffer at pH 7.0 (A: tion containing the fibrinogen and the ON. The time required
5240 | Org. Biomol. Chem., 2014, 12, 5235–5242
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
for fibrin polymerization was determined from the UV scatter- SA/MM calculations, we applied a tether force of 100 kcal Å−2
ing curve, registered as a function of time (wavelength fixed at to the guanine bases of the two quartets.
380 nm) in the presence of each ON. Each curve was deter-
Resulting conformers were analyzed and loop nucleotides
mined in triplicate at different concentrations. The clotting were classified on the basis of: (i) glycosidic bond χ values, i.e.,
time value reported as M SE was derived as the maximum of 0° < χ < 90° = syn; −60° < χ < −180° = anti;26 90° < χ < 180° and
the second derivative of each scattering curve. The basal clot- −60° < χ < 0° = s/a; (ii) the interatomic distance between the
ting time was determined by measuring the UV scattering as a centroid of the ring atoms of the nucleobase of each loop
function of time produced in the absence of any ONs. The nucleotide and the centroid of the ring atoms of the nucleo-
15-mer GTGTGTGTGTTGTGT was used as the negative control. bases of the two G-tetrads (Pseudo_Atom Define command,
In the absence of any inhibitor, the clotting time value was Biopolymer Module, Insight 2005). When the distance was
25.6 1.5 s. The prolonged fibrinogen clotting times of TBA <8 Å, the loop nucleotide was classified as “stacked”; when the
and modified ONs were calculated by subtracting the clotting distance was >12 Å it was classified as “not-stacked”. A 3D
time produced by thrombin in the absence of any inhibitor visual inspection was used to classify the nucleotide when the
from that measured in the presence of the aptamer.
distance was included between 8 Å and 12 Å. In any case, a
“stacked” nucleotide presented at least one nucleobase atom
under the G-tetrads. Occurrence rates were calculated for all
conformers (Tables S1 and S2 in ESI†).
Molecular modelling
Experimentally determined structures of TBA alone (PDB ID:
New modified TBA conformers, resulting from SA/MM cal-
148D) and in complex with thrombin (PDB IDs: 1HAO, 1HAP, culations, were superimposed on the experimentally deter-
1HUT, 4DIH, and 4DII) were downloaded from the Protein mined structures of TBA in complex with thrombin (PDB IDs:
Data Bank (PDB, http://www.rcsb.org/pdb/) and analyzed using 1HAO, 1HAP, 1HUT, 4DIH and 4DII) by fitting heavy atoms of
the Homology module of Insight 2005 (Accelrys Software Inc., the guanine bases of two quartets, and the overlap with the
San Diego, CA). Hydrogens were added to all these structures residues interacting with thrombin ABE I was evaluated. All
considering a pH value of 7.4 (Biopolymer Module, Insight possible TBA binding modes (Fig. 3) were considered. For each
2005).
aptamer/enzyme complex a subset around the ligand includ-
Since the replacement of T3, T7 and T12 residues with ing all protein and water molecules having at least one atom
nucleoside c produced diastereomers characterized by S or R within a 6 Å radius from any given aptamer atom was defined.
configuration at the acyclic linker, for each new TBA analogue, The created subsets were displayed and analyzed using Insight
the two diastereomers were built by modifying the experi- 2005 Biopolymer and Homology modules (Accelrys Software
mentally determined structure of TBA in complex with throm- Inc., San Diego, CA).
bin (PDB ID: 1HAO; Insight 2005 Builder module). Atomic
potentials and charges were assigned using the CVFF force
field.24
Conclusions
The conformational space of the new modified analogues
The results presented herein clearly establish the role of loop
conformational preferences on the overall TBA–thrombin com-
plexes and illustrate how a stereochemical modification can
impact on aptamer-protein recognition. Furthermore, since T
residues are often loop components of aptamers, our analyses
can be widely reproduced in other cases, using the R/S nucleo-
side derivatives c as two simple alternatives for the T residue,
to acquire and/or refine the structure–activity relationship
between aptamers and their targets.
was sampled through 200 cycles of simulated annealing (SA)
followed by molecular mechanics (MM) energy minimization.
During the SA procedure, the temperature is altered in time
increments from an initial temperature to a final temperature
by adjusting the kinetic energy of the structure (by rescaling
the velocities of the atoms). The following protocol was
applied: the system was heated to 1000 K over 2000 fs (time
step of 1.0 fs); a temperature of 1000 K was applied to the
system for 2000 fs (time step of 1.0 fs) to surmount torsional
barriers; successively, the temperature was linearly reduced to
300 K in 1000 fs (time step of 1.0 fs). Resulting conformations
were then subjected to MM energy minimization within
Insight 2005 Discover_3 module (CVFF force field) until the
maximum rms derivative was less than 0.001 kcal Å−1, using a
conjugate gradient25 as the minimization algorithm. In order
to reproduce the physiological environment where these mole-
cules act and to evaluate the effects of the implicit solvent, we
sampled the conformational space through the combined pro-
cedure of SA/MM calculations, using the dielectric constant of
water (ε = 80r). Moreover, in order to allow a complete relax-
ation of the structures preserving the monomolecular chair-
like G-quadruplex folding topology, during the entire course of
Acknowledgements
This work was funded by the Italian MIUR grant FIRB
RBFR12WB3W_003 (M.P.).
Notes and references
1 A. D. Keefe, S. Pai and A. Ellington, Nat. Rev. Drug Discovery,
2010, 9, 537.
2 P. S. Pendergrast, H. N. Marsh, D. Grate, J. M. Healy and
M. Stanton, J. Biomol. Technol., 2005, 16, 224.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 5235–5242 | 5241

View Article Online
Paper
Organic & Biomolecular Chemistry
3 D. Musumeci, G. Oliviero, G. N. Roviello, E. M. Bucci and 16 T. Coppola, M. Varra, G. Oliviero, A. Galeone, G. D’Isa,
G. Piccialli, Bioconjugate Chem., 2012, 23, 382.
4 D. Musumeci and D. Montesarchio, Pharmacol. Ther., 2012,
136, 202.
5 R. E. Wang, H. Wu, Y. Niu and J. Cai, Curr. Med. Chem.,
2011, 18, 4126.
6 C. L. Esposito, S. Catuogno, V. de Franciscis and
L. Cerchia, Discovery Med., 2011, 11, 487.
7 L. C. Bock, L. C. Griffin, J. A. Latham, E. H. Vermaas and
J. J. Toole, Nature, 1992, 355, 564.
8 P. Schultze, R. F. Macaya and J. Feigon, J. Mol. Biol., 1994,
235, 1532.
L. Mayol, E. Morelli, M. R. Bucci, V. Vellecco, G. Cirino and
N. Borbone, Bioorg. Med. Chem., 2008, 16, 8244.
17 N. Borbone, M. Bucci, G. Oliviero, E. Morelli, J. Amato,
V. D’Atri, S. D’Errico, V. Vellecco, G. Cirino, G. Piccialli,
C. Fattorusso, M. Varra, L. Mayol, M. Persico and
M. Scuotto, J. Med. Chem., 2012, 55, 10716.
18 A. Pasternak, F. J. Hernandez, L. M. Rasmussen, B. Vester
and J. Wengel, Nucleic Acids Res., 2011, 39, 1155.
19 T. B. Jensen, J. R. Henriksen, B. E. Rasmussen,
L. M. Rasmussen, T. L. Andresen, J. Wengel and
A. Pasternak, Bioorg. Med. Chem., 2011, 19, 4739.
9 K. Y. Wang, S. H. Krawczyk, N. Bischofberger, S. Swaminathan 20 T. Horhota, J. W. Szostak and L. W. McLaughlin, Org. Lett.,
and P. H. Bolton, Biochemistry, 1993, 32, 11285. 2006, 8, 5345.
10 W.-X. Li, A. V. Kaplan, G. W. Grant, J. J. Toole and 21 T. Takaya, H. Yoshimoto and E. Imoto, Bull. Chem. Soc.
L. L. K. Leun, Blood, 1994, 83, 677. Jpn., 1967, 40, 2844.
11 K. Padmanabhan, K. P. Padmanabhan, J. D. Ferrara, 22 W. A. Kibbe, Nucleic Acids Res., 2007, 35(suppl.2), W43.
J. E. Sadler and A. Tulinsky, J. Biol. Chem., 1993, 268, 17651.
12 K. Padmanabhan and A. Tulinsky, Acta Crystallogr., Sect. D:
Biol. Crystallogr., 1996, 52, 272.
13 J. A. Kelly, J. Feigon and T. O. Yeates, J. Mol. Biol., 1996,
256, 417.
23 Y. R. De Cristofaro and E. Di Cera, J. Protein Chem., 1991,
10, 455.
24 P. Dauber-Osguthorpe, V. A. Roberts, D. J. Osguthorpe,
J. Wolff, M. Genest and A. T. Hagler, Proteins, 1988, 4, 31.
25 R. Fletcher, in Practical Methods of Optimization, John Wiley
& Sons, New York, 1980, vol. 1.
14 I. Russo Krauss, A. Merlino, A. Randazzo, E. Novellino,
L. Mazzarella and F. Sica, Nucleic Acids Res., 2012, 40, 8119. 26 (a) S. Neidle, Principles of nucleic acid structures, Academic
15 A. Avino, C. Fabrega, M. Tintore and R. Eritja, Curr. Pharm.
Des., 2012, 18, 2036 and references cited therein.
Press, London, 2008; (b) J. Reichert and J. Sühnel, Nucleic
Acids Res., 2002, 30, 253.
5242 | Org. Biomol. Chem., 2014, 12, 5235–5242
This journal is © The Royal Society of Chemistry 2014