Dinucleosides with non-natural backbones: A new class of ribonuclease A and angiogenin inhibitors

Article abstract of DOI:10.1002/chem.201102816

Ribonuclease A (RNase A) serves as a convenient model enzyme in the identification and development of inhibitors of proteins that are members of the ribonuclease superfamily. This is principally because the biological activity of these proteins, such as a

Full text of DOI:10.1002/chem.201102816

DOI: 10.1002/chem.201102816
Dinucleosides with Non-Natural Backbones: A New Class of Ribonuclease A
and Angiogenin Inhibitors
Joy Debnath,[b] Swagata Dasgupta,*[a] and Tanmaya Pathak*[a]
Abstract: Ribonuclease A (RNase A)
serves as a convenient model enzyme
in the identification and development
of inhibitors of proteins that are mem-
bers of the ribonuclease superfamily.
This is principally because the biologi-
cal activity of these proteins, such as
angiogenin, is linked to their catalytic
ribonucleolytic activity. In an attempt
to inhibit the biological activity of an-
giogenin, which involves new blood
vessel formation, we employed differ-
ent dinucleosides with varied non-natu-
ral backbones. These compounds were
synthesized by coupling aminonucleo-
sides with dicarboxylic acids and
amino- and carboxynucleosides with an
amino acid. These molecules show
competitive inhibition with inhibition
constant (Ki) values of (59ꢁ3) and
(155ꢁ5) mm for RNase A. The com-
pounds were also found to inhibit an-
giogenin in a competitive fashion with
corresponding Ki values in the micro-
molar range. The presence of an addi-
tional polar group attached to the
backbone of dinucleosides was found
to be responsible for the tight binding
with both proteins. The specificity of
different ribonucleolytic subsites were
found to be altered because of the in-
corporation of a non-natural backbone
in between the two nucleosidic moiet-
ies. In spite of the replacement of the
phosphate group by non-natural link-
ers, these molecules were found to se-
lectively interact with the ribonucleo-
lytic site residues of angiogenin, where-
as the cell binding site and nuclear
translocation site residues remain un-
perturbed. Docked conformations of
the synthesized compounds with
RNase A and angiogenin suggest
a
Keywords: angiogenin · cell bind-
ing site · inhibitors · nucleosides · ri-
bonuclease A
binding preference for the thymine–ad-
enine pair over the thymine–thymine
pair.
Introduction
ic, and antiviral activities.[11] RNase activity is also required
for the antiviral and neurotoxic activity of eosinophil cation-
ic protein (ECP).[12] Thus the design of inhibitors of the cat-
alytic activity of ribonucleolytic proteins is expected to have
an indirect effect on their purported biological activity. The
homology of these proteins with ribonuclease A (RNase A)
permits its use as a model system in the development of in-
hibitors.
The catalytic site of RNase A contains various different
subsites. Hydrolysis of the phosphodiester bond occurs at
the P1 subsite that is composed of His-12, Lys-41, and His-
119 residues. Similarly, His-13, Lys-40, and His-114 residues
constitute the P1 subsite of angiogenin (Figure 1), which
bears about 65% identity with RNase A. Apart from the P1
subsite, these two proteins have two other important sub-
sites for the recognition of nucleobases of their substrate
(RNA), namely B1 and B2 subsites. A comparison of the ri-
bonucleolytic sites of these two proteins is given in Figure 1.
The B1 and B2 subsites are known for their pyrimidine and
purine specificity, respectively.[13] The ribonucleolytic activity
of angiogenin, though 10ꢀ5–10ꢀ6 weaker[14,15] than RNase A,
is nevertheless essential for its biological activity.[16,17] The
designed low-molecular-weight inhibitors for the biologically
active RNases have been tested on bovine pancreatic
RNase A as a model protein of this superfamily.
Several members of the ribonuclease (RNase) superfamily
display a multitude of activities in addition to the catalytic
activity normally associated with these proteins. Angiogenin
is one such member that is a potent inducer of angiogenesis,
the process of new blood vessel formation.[1,2] The processes
of tumor growth, invasion, and metastasis are dependent
upon angiogenesis and angiogenin is known to induce many
steps of this complex cascade.[3–8] Tumors and cancer cells
are known to use this process for ready access to nutrients.[9]
The biological activity of angiogenin is dependent on its ri-
bonucleolytic activity.[10] Eosinophil-derived neurotoxin
(EDN) is a catalytically efficient RNase with its ribonucleo-
lytic activity being essential for its neurotoxic, helminthotox-
[a] Prof. S. Dasgupta, Prof. T. Pathak
Department of Chemistry
Indian Institute of Technology Kharagpur
Kharagpur-721302 (India)
Fax : (+91)3222-255303
E-mail: swagata@chem.iitkgp.ernet.in
tpathak@chem.iitkgp.ernet.in
[b] Dr. J. Debnath
Department of Pharmaceutical Sciences
University of Tennessee
847 Monroe, Johnson Bldg, Room 331
Memphis, Tennessee 38163 (USA)
Reports of many nucleotidic inhibitors for RNase A with
very low Ki values do exist.[18,19] These nucleotidic inhibitors
have limited success for angiogenin. Despite the fact that
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102816.
1618
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1618 – 1627

FULL PAPER
Two sets of backbone-modified dinucleosides, one with a
nonpolar backbone and the other set of compounds with
polar backbones have been synthesized. In order to prepare
these backbone-modified dinucleosides, the nonparticipating
functional groups of nucleosides were masked with appro-
priate protecting groups. The protecting groups were select-
ed in a manner such that the number of deprotection steps
was minimized for regenerating the functional groups of nu-
cleosides.
For the synthesis of these dinucleosides, two monomeric
nucleosidic moieties were prepared individually, one with
carboxylic and the other with an amino group. To generate
the thymidine monomeric unit with one carboxylic group we
started with compound 1.[25] Thus, the amino group of par-
tially protected aminothymidine 1 was treated with succinic
and glutaric anhydride to produce compounds 2 and 3, re-
spectively (Scheme 1).
Figure 1. Key residues of ribonucleolytic site (A) RNase A and (B) an-
giogenin.
these molecules completely mimic the substrate of RNases
and bind strongly with these enzymes, there appear to be
two basic problems regarding their administration in the
body. First, there is the difficulty associated with the trans-
port of these compounds through the biological membrane.
The high negative charge on the phosphate group makes
them difficult to use in vivo.[20–24] Second, the possible hy-
drolysis of the polyphosphate group by different enzymes
present
in
body
fluid
namely
Ap3A hydrolase,
Ap4A hydrolase and diphosphoinositol polyphosphate phos-
phohydrolases (DIPP proteins) are likely to render them in-
effective. Apart from these features, their possible toxicity
and likely accumulation of degraded products in the body,
which can inhibit essential system pathways, cannot be over-
looked.[21]
Our attempt to develop nucleoside-based inhibitors with
carboxylic and amino functionalities showed successful in-
hibition of the ribonucleolytic activity of RNase A.[25–27] In
this study, a non-natural backbone was used to link two nu-
cleosidic moieties in lieu of the phosphate or pyrophosphate
groups.[18–24] This linkage minimizes the ionic character of
the compounds. Two sets of backbone-modified dinucleo-
sides have been designed, one with a nonpolar backbone
and the other set of compounds with a polar backbone. The
amide bond was selected to connect the modified nucleosi-
dic units, because of their expected stability towards phos-
phohydrolases. The nature of the linkers between the nucle-
osidic moieties has also been varied by altering the spacer
length and polar groups. The variant nature of the linkers is
expected to produce valuable information about their mode
of binding with the proteins.
Scheme 1. Synthesis of aminonucleoside-dicarboxylic acid hybrids 2 and
3.
5’-Carboxy thymidine 6 was prepared from partially pro-
tected thymidine 4[28] by oxidation of the 5’-OH group using
K2S2O8 as the oxidizing agent and RuCl3 as a catalyst
(Scheme 2).[29] Compound 5, with an amino group at the 5’-
position, was obtained from thymidine using a reported
method.[30] 2’,3’-Protected 5’-carboxyadenosine 9 was synthe-
sized from partially protected adenosine 7 (Scheme 3) and
Results and Discussion
The basic skeleton chosen for the dinucleoside inhibitors de-
veloped in this study are based on thymidine and adenosine.
Scheme 2. Synthesis of 5’-carboxythymidine 6 from compound 4.
Chem. Eur. J. 2012, 18, 1618 – 1627
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1619

T. Pathak, S. Dasgupta and J. Debnath
All acid-labile trityl, isopropylidene and tert-butoxycarbonyl
masking groups were removed in one step with the help of
50% TFA treatment to afford the heterodimer 15
(Scheme 5).
Scheme 3. Synthesis of 5’-carboxyadenosine analogue 9 from compound
7.
the 5’-amino-5’-deoxyadenosine derivative 8 was prepared
from adenosine using a literature procedure.[31]
The thymidine monomeric units 2 and 3 were coupled
separately with 5 through amide bonds to produce modified
dinucleosides with nonpolar backbones (Scheme 4) by using
Scheme 5. Synthesis of heterodimer 15 with neutral backbone.
For the introduction of a carboxylic group in the internu-
cleoside bond of a dinucleoside, a partially protected aspart-
ic acid derivative l-Boc-AspACTHNUGRTNEUNG(OBzl)OH was selected. To
reduce the number of deprotection steps, a benzyl group
was selected for the protection of g-carboxylic group of as-
partic acid and 3’/5’-group of thymidine. Although in general
it was difficult to synthesize a 5’-O-benzyl-protected amino-
nucleoside from the corresponding azidonucleoside keeping
the benzyl group intact, the use of SnCl2-PhSH-Et3N[33] led
to the reduction of azidothymidine 16[34] to give the benzyl-
protected aminothymidine 17 in moderate yield (Scheme 6).
Compound 17 was coupled with l-Boc-Asp
dicyclohexylcarbodiimide (DCC) and N-hydroxysuccin-
amide (NHS) to afford fully protected nucleoside amino
ACHTUNGTREN(NUNG OBzl)OH using
AHCTUNGTRENNUNG
acid conjugate 18. Selective deprotection of the Boc group
of 18 with a mixture of 4m TMSCl and 4m PhOH afforded
compound 19 with one free amino group (Scheme 6). Modi-
fied dinucleosides with polar backbones were then prepared
by separately coupling 19 with carboxynucleosides 6 and 9
to afford fully protected dinucleosides 20 and 22, respective-
ly. In case of 20 all the three benzyl groups were removed
by using Pd/C under H2 atmosphere to afford the homodi-
nucleoside 21. For the synthesis of the heterodimer 23, the
benzyl groups of 22 were removed by using the same proce-
dure mentioned above and the remaining Boc and the iso-
propylidene groups were removed by 50% TFA in dichloro-
methane (Scheme 6).
Scheme 4. Synthesis of dinucleosides 11 and 13 with neutral backbones.
a combination of 1-[3-(dimethylamino)propyl]-3-ethylcarbo-
diimide (EDC) and 1-hydroxybenzotriazole (HOBT) as cou-
pling reagent and diisopropylethylamine (DIEA) as the
base to obtain fully protected dimers 10 and 12, respective-
ly.[32] Since compound 12 was always contaminated with an
inseparable impurity the compound was directly taken to
the deprotection step. All acid-labile groups of 10 and 12
were removed in one step by 50% TFA treatment to afford
the required dimeric nucleosides 11 and 13, respectively
(Scheme 4). A similar strategy was used for the synthesis of
a dinucleoside 15, in which thymidine and adenosine ana-
logues were linked with a nonpolar backbone. Thus amino-
nucleoside/succinic acid hybrid 2 was coupled with the ade-
nosine analogue 8 to obtain the fully protected dimer 14.
The inhibition of the ribonucleolytic activity of RNase A
was checked by agarose-gel-based assay, for which the deg-
radation of tRNA by RNase A was monitored (Figure 2).
The most intense band observed in lane 1 was due to the
1620
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1618 – 1627

Dinucleosides with Non-Natural Backbones
FULL PAPER
Figure 2. Agarose-gel-based assay. RNase concentration: 0.5 mm and con-
centration of all compounds 0.92 mm. Lane1: tRNA; Lane 2: tRNA and
RNase A; Lanes 3, 4 and 5: tRNA and RNase A with increasing concen-
tration of inhibitors.
Scheme 6. Synthesis of homodimer 21 and heterodimer 23 with polar
backbone.
Figure 3. Ribonucleolytic inhibition of RNase A by backbone modified
dinucleosides (23, 21, 15, 11, and 13). RNase A and compound concentra-
tion are 0.27 mm and 0.7 mm respectively.
presence of the control, tRNA. The maximum possible deg-
radation of tRNA by RNase A resulted in the faint intensity
of the band observed in lane 2. The differential intensity of
bands in lanes 3 to 5 indicate the extent of RNase A inhibi-
tion by the compounds at three different concentrations
(lower to higher concentrations, respectively). These results
qualitatively showed a better inhibition potency of 15, 21,
and 23 compared to the other compounds.
The histogram of relative ribonucleolytic activities ob-
tained from the precipitation assay at a particular inhibitor
concentration (0.7 mm) further confirmed the comparative
inhibitory efficacy of the backbone modified nucleosides
(Figure 3). It was observed that compounds with a carboxyl-
ic functionality between two nucleoside moieties (21 and 23)
showed better inhibition. It was also observed that the thy-
mine–adenine pair showed better inhibitory potential over
the thymine–thymine pair. This was probably because of the
better anchoring of the thymine–adenine base pair at B1 and
B2 sites respectively.
To determine the nature of inhibition and the inhibition
constants, kinetic experiments were conducted with 23, 21,
15, and 11. The inhibition constant values (Ki) obtained for
23, 21, 15, and 11 were (59ꢁ3), (155ꢁ5), (443ꢁ2) and
(544ꢁ2) mm, respectively. From the nature of the Dixon
plots of the four compounds (Figure 4A–D) it was con-
firmed that the mode of inhibition was competitive in
nature. The order of values obtained for the inhibition con-
stants correlated well with the previous agarose gel and pre-
cipitation assay.
After the kinetic experiments, we proceeded with docking
studies to visualize the conformation of the synthesized
Chem. Eur. J. 2012, 18, 1618 – 1627
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1621

T. Pathak, S. Dasgupta and J. Debnath
the Nd1 atom of His-119 by means of a possible hydrogen-
bonding network (Figure 5A). It was found that in spite of
the presence of the thymine nucleobase the carboxylic
moiety occupied the B1 site of the protein.
Figure 5. Docked conformation of A) 23 (yellowish-green) and B) 21
(green) with RNase A (1FS3).
The docked conformation of 21 also showed the same
mode of interaction, but in this case one of the thymine
moieties interacted with the B1 site (Val-43) residue (Fig-
ure 5B). The carboxylic group of this compound interacted
only with the His-12 residue, whereas for 23 the carboxylic
group interacted with both His-12 and Lys-41. For molecule
21, it was observed that the nitrogen atoms of the linker
amide bond were in a position to form a five-membered
ring through a hydrogen-bonding network with the Nd1 atom
of His-119 similar to 23.
Figure 4. Dixon plots for RNase A inhibition by A) Compound 23: sub-
~
&
^
strate concentrations: 580 mm ( ), 320 mm ( ) and 275 mm ( ); B) Com-
&
^
pound 21: substrate concentrations: 650 mm ( ), 490 mm ( ) and 320 mm
~
^
(
); and C) Compound 15: substrate concentrations: 500 mm ( ), 350 mm
~
&
( ) and 220 mm ( ). D) Compound 11: substrate concentrations: 750 mm
~
&
^
(
), 500 mm ( ) and 350 mm ( ). The RNase A concentration: 7.5 mm.
compounds in the protein. From the docked conformation
of compound 23, we found that its carboxylic group was
within hydrogen-bonding distance of both His-12 and Lys-
41. On the other hand the 2’- and 3’-OH groups of adeno-
sine were in a position to form a five-membered ring with
In the docked conformation of compound 15, it was found
that the adenine moiety was in close proximity to His-119,
His-12, and Lys-41 (Figure 6A). Here we also find that the
1622
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1618 – 1627

Dinucleosides with Non-Natural Backbones
FULL PAPER
Like RNase A the precipitation assay was also conducted
in case of angiogenin. The concentration of the five com-
pounds (23, 21, 15, 11, and 13) was fixed at 25 mm with the
angiogenin concentration at 0.25 mm. Compounds of this
series showed the same trend of ribonucleolytic inhibition of
angiogenin as in the case of RNase A (Figure 7). Within the
series of compounds, 23 and 21 showed 21 and 17% ribonu-
cleolytic inhibition, respectively, but compound 13 showed
only 2% inhibition.
Figure 7. Inhibition of ribonucleolytic activity of RNase A (gray) and an-
giogenin (black) by backbone modified dinucleosides.
From the precipitation assay it was found that 23 and 21
showed better inhibitory activity compared to the other
compounds. Therefore these two compounds were selected
for the kinetic study with 6-FAM-dArUdAdA-6-TAMRA as
the substrate. The reciprocal of the increase in fluorescence
intensity of the fluorescein moiety was plotted against the
reciprocal of substrate concentration (Lineweaver–Burk
plot), and from the secondary plot we found that the inhibi-
tion constant values for 23 and 21 as (670ꢁ5) and (729ꢁ
2) mm, respectively. The nature of the Lineweaver–Burk
plots (Figure 8) also indicated the competitive mode of in-
hibition of the ribonucleolytic activity of angiogenin. The in-
hibition constant values for compounds 11, 15, 21, and 23
are summarized in Table 1.
Specific protein–ligand interactions for compounds 23 and
21 were also investigated by docking studies. From the
docked structure of 23 it was found that the carboxylic
group interacted with His-114. On the other hand, three ni-
trogen atoms of the adenine nucleobase were in a position
to form hydrogen bonds with His-13 and Lys-40. The ade-
nine moiety was also found to interact with the B1 site resi-
due (Leu-115) rather than B2 site residues. In contrast, the
thymine moiety interacted with the B2 site residue (Gln-
108) (Figure 9). The reverse binding observed may be a
result of the incorporation of a non-natural backbone be-
tween the two nucleosidic moieties. Similarly for 21
(Figure 10) we found the carboxylic group did not interact
with any of the His residues of the catalytic site, but was
found to be close to Gln-12. Only His-114 was within hydro-
Figure 6. Docked conformation of A) 15 (purple) and B) 11 (yellowish-
green) with RNase A (1FS3).
backbone oxygen atom of Val-43 (member of B1 site) was
within a hydrogen-bonding network with the 2’- and 3’-OH
groups of the adenine moiety. The amide nitrogen and
oxygen atoms (of the linker) are in a position to form a hy-
drogen-bonding network with Arg-10, Gln-11, and Arg-39
residues. However, for compound 11, among the three main
P1 site residues, only HIs-12 was found to be within hydro-
gen-bonding distance (Figure 6B).
Apparently the presence of a carboxylic group (in the
linker) for 23 and 21 results in a tighter interaction with the
active site residues resulting in favorable ionic interactions
with the active site residues. In contrast, there is no such
polar ionizable group for 15 and 11, which have instead a
linker that is hydrophobic in nature. This increase in hydro-
phobicity is manifested by a decrease in inhibition efficacy
as reflected from their inhibition constant (Ki) values. Thus
23 and 21 showed better inhibition than 15 and 11. The suc-
cessful inhibitory activity of the backbone-modified dinu-
cleosides with polar and nonpolar backbones on RNase A
encouraged us to extend this study to angiogenin, the homo-
logue of RNase A that is a potent blood vessel inducer.
Chem. Eur. J. 2012, 18, 1618 – 1627
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1623

T. Pathak, S. Dasgupta and J. Debnath
Figure 10. Docked conformation of 21 (cyan) with angiogenin (1B1I).
Conclusion
Figure 8. Lineweaver–Burk plots for the inhibition of angiogenin by A)
~
~
)
&
^
^
23: 0 mm ( ), 0.4 mm ( ) and 1.1 mm ( ); B) 21: 0 mm ( ), 0.6 mm (
Backbone-modified non-phosphate-linked dinucleosides
have been employed to inhibit the ribonucleolytic activity of
RNase A and angiogenin.[35] Among the five members of
this series we find that compounds 23 and 21 with a carbox-
ylic group present in the linker region were able to inhibit
the ribonucleolytic activity of RNase A and angiogenin
more effectively. For both hydrophobic and hydrophilic link-
ers we find that the thymine–adenine pair was more effec-
tive compared to the thymine–thymine pair. Comparing the
Ki value of 23 with TpdA-3’-p and TpdA (Ki values are 137
and 1200 mm, respectively),[36,37] we observe that the incorpo-
ration of the non-natural backbone in place of phosphate
groups effectively enhanced their inhibitory activity towards
RNase A. The Ki value of 23 ((59ꢁ3) mm) is higher than
pTpdA-3’-p (14.2 mm),[36] which is likely due to presence of
three highly ionizable phosphate groups. However, for an-
giogenin the enhancement was more pronounced. The incor-
poration of this type of non-natural backbone was probably
the reason behind their better inhibitory activity. The dock-
ing poses of these compounds with RNase A and angiogenin
revealed that these compounds interacted typically with the
ribonucleolytic site residues. However, it also shows how
synthetic modifications of the inhibitors can generate a pro-
found alteration in its mode of interaction with the target
protein.
&
and 2.3 mm ( ). Angiogenin concentration: 0.34 mm.
Table 1. Inhibition constant (Ki) values for compounds 23, 21, 15, and 11.
Ki [mm] for
RNase A
Ki [mm] for
angiogenin
23
21
15
11
G
(670ꢁ5)
E
A
R
–
–
AHCTUNGTRENNUNG
Thus the introduction of a non-natural backbone had
been proven to be a good venture for the designing of inhib-
itors with lower ionic character. We detect that the presence
of the non-natural backbone with one carboxylic group was
effective in increasing the inhibitory activity of the protein
compared to the phosphate analogue. Therefore it is possi-
ble that this kind of modification can be utilized for the
identification of potent inhibitors of ribonucleases.
Figure 9. Docked conformation of 23 with angiogenin (1B1I).
gen-bonding distance of a nitrogen atom of the amide bond
that serves as the linker.
1624
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1618 – 1627

Dinucleosides with Non-Natural Backbones
FULL PAPER
hygroscopic white solid. 1H NMR: [D6]DMSO: d=1.77 (s, 3H),2.05–2.10
(m, 1H), 2.16–2.21 (m, 1H), 2.36 (m, 3H), 3.28–3.43 (m, 2H), 3.52 (dd,
J=4, 12 Hz, 1H), 3.60 (m, 1H), 3.74 (m, 1H), 3.95 (m, 1H), 4.02 (m,
1H), 4.29 (m, 1H), 4.67 (t, J=5.6 Hz, 1H), 5.83 (d, J=6 Hz, 1H), 6.17 (t,
J=6.8 Hz, 1H), 7.77 (s, 1H), 8.18 (s, 1H), 8.35 ppm (s, 1H); 13C NMR
[D6]DMSO: d=12.6, 30.9 (CH2), 37.3 (CH2), 41.3 (CH2), 49.3, 61.6
(CH2), 71.5, 72.9, 83.9, 84.0, 85.4, 88.2, 109.8 (C), 119.8 (C), 136.7, 140.9,
149.5 (C), 150.8 (C), 153.0, 156.5 (C), 164.3, 172.1 (C), 172.2 ppm (C);
HRMS (ESI+): m/z calcd for C24H32N9O9 [M+H]+: 590.2322; found:
590.2315.
Experimental Section
Materials: Bovine pancreatic ribonuclease A (RNase A), recombinant
human angiogenin, human serum albumin (HSA), yeast tRNA, 2’,3’-
cCMP, D2O, [D6]DMSO, CDCl3were from Sigma–Aldrich. 6-FAM-dArU-
dAdA-6-TAMRA was purchased from Integrated DNA Technologies
(Coralville, IA) and all other reagents (analytical grade) were from SRL
(India) and also from Sigma–Aldrich. Column chromatographic separa-
tions were performed using silica gel (60–120 and 230–400 mesh). HPLC
separations were performed by Shimadru Prominence Series HPLC
system. Luna 10 mm C18 column (100 ꢁ) was used for the separation.
Solvents were dried and distilled following standard procedures. TLC
was carried out on precoated plates (Merck silica gel 60, f254), and the
spots visualized with UV light or by charring the plates dipped in 5%
H2SO4/MeOH solution or 5% H2SO4/vanillin/EtOH or 5% ninhydrin in
MeOH. 1H NMR (400 MHz) and 13C NMR (100 MHz) spectra were re-
corded on a Bruker NMR spectrometer (d scale). UV/Vis measurements
were made using a Perkin–Elmer UV/Vis spectrophotometer (Model
Lambda 25). Fluorescence measurements were recorded on a Spex Fluo-
rolog-3 machine. Compounds 1,[38] 5,[30] 6,[28] 7, 8,[31] and 16[34] and l-Boc-
Compound 21: 10% Pd/C (0.1 g) was added to a solution of compound
20 (0.09 g, 0.1 mmol) in dry MeOH (10 mL) and the mixture was stirred
at 608C under H2 atmosphere. After 8 h (TLC), the mixture was diluted
with MeOH and filtered through a bed of Celite, which was further
washed several times with MeOH. Combined organic layers were evapo-
rated under reduced pressure. The crude residue was purified with the
help of HPLC to afford compound 21 (0.04 g, 68%); hygroscopic white
solid.1H NMR: [D4]MeOH: d=1.88–1.90 (m, 6H), 2.19 (m, 1H), 2.36 (m,
2H), 2.68 (m, 2H), 3.67–3.89 (m, 4H), 4.36 (brs, 1H), 4.49 (m, 1H), 4.61
(m, 2H), 6.22 (m, 1H), 6.33 (t, J=6.4 Hz, 1H), 7.87 (s, 1H), 7.92 ppm (s,
1H); 13C NMR [D6]DMSO: d=12.0, 12.1, 35.1 (CH2), 38.0 (CH2), 50.2,
52.2, 66.8 (CH2), 73.7 (CH2), 77.3, 81.2, 84.4, 84.5, 91.9, 111.2 (C), 111.6
(C), 135.5, 138.9, 150.4 (C), 151.1 (C), 163.6 (C), 163.7 (C), 170.1 (C),
170.2 (C), 175.0 (C); HRMS (ESI+): m/z calcd for C24H30N6O12Na
[M+Na]+: 617.1819; found: 617.1822.
AspACHTUNGTRENNUNG(OBzl)OH were prepared following literature procedures. Concen-
trations of the solutions were determined spectrophotometrically using
the following data: for RNase A e278.5 =9800mꢀ1 cmꢀ1
,
for 2’,3’-cCMP
[39]
[40]
[41]
e268 =8500mꢀ1 cmꢀ1
,
for angiogenin e278 =12500mꢀ1 cmꢀ1
,
for 6-FAM-
dArUdAdA-6-TAMRA e260 =102400mꢀ1 cmꢀ1
.
[41]
Compound 23: 10% Pd/C (0.1 g) was added to a solution of compound
22 (0.1 g, 0.1 mmol) in dry MeOH (10 mL) and the mixture was stirred at
608C under H2 atmosphere. After 8 h (TLC), the mixture was diluted
with MeOH and filtered through a bed of Celite, which was further
washed several times with MeOH. Combined organic layers were evapo-
rated under reduced pressure. The crude residue was stirred with 50%
TFA in CH2Cl2 (3 mL) at room temperature. After 4 h (TLC), all volatile
materials were removed under reduced pressure. The crude residue was
purified with the help of HPLC to afford compound 23 (0.03 g, 62%);
Compound 11: A solution of compound 10 (0.2 g, 0.2 mmol) and 50%
TFA in CH2Cl2 (10 mL) was stirred for 4 h at room temperature. All vol-
atile matters were removed under reduced pressure. The crude residue
was purified over silica gel to afford compound 11 (0.1 g, 79%); hygro-
scopic white solid. 1H NMR: [D6]DMSO: d=1.78 (s, 3H), 1.81 (s, 3H),
1.99–2.21 (m, 4H), 2.34 (brs, 4H), 3.27–3.31 (m, 2H), 3.53 (dd, J=3.6,
12 Hz, 1H), 3.61 (m, 1H), 3.72 (m, 2H), 4.13, (m, 1H), 4.28 (m, 1H),
6.11–6.20 (m, 2H), 7.50 (s, 1H), 7.77 ppm (s, 1H); 13C NMR [D6]DMSO:
d=12.4, 12.6, 30.9 (CH2), 37.3 (CH2), 38.7 (CH2), 41.2 (CH2), 49.4, 61.7
(CH2), 71.5, 84.0, 84.2, 85.3, 85.4, 109.8 (C), 110.2 (C), 136.6, 150.8 (C),
150.9 (C), 164.2 (C), 172.0 (C), 172.2 ppm (C); HRMS (ESI+): m/z calcd
for C24H32N6O10Na [M+Na]+: 587.2072; found: 587.2054.
1
hygroscopic white solid. H NMR: [D4]MeOH: d=1.64 (s, 3H), 1.80 (dd,
J=5.2, 16.4 Hz, 1H), 2.20–2.28 (m, 2H), 2.50 (m, 1H), 3.66 (m, 2H), 3.79
(m, 1H), 3.91 (brs, 1H), 4.41 (m, 1H), 4.53–4.62 (m, 2H), 4.83 (d, J=
2 Hz, 1H), 6.22 (d, J=1.6 Hz, 1H), 6.32 (t, J=6 Hz, 1H), 7.80 (s, 1H),
8.65 (s, 1H), 8.82 ppm (s, 1H); 13C NMR [D6]DMSO: d=12.0, 30.9
(CH2), 37.3 (CH2), 48.9, 50.9, 66.9, 83.3, 83.4, 84.5, 84.8, 86.9, 91.3, 111.4
(C), 123.8 (C), 136.2, 137.5 (C), 144.4, 149.5 (C), 150.6 (C), 150.9 (C),
152.1, 152.2 (C), 163.9 (C), 173.2 ppm (C). HRMS (ESI+): m/z calcd for
C24H29N9O11Na [M+Na]+: 642.1884; found: 642.1887.
Compound 13: A solution of compound 3 (0.4 g, 0.67 mmol) in dry THF
(15 mL) was cooled to ꢀ158C. EDC·HCl (0.18 g, 0. 94 mmol) and HOBT
(0.13 g, 0.94 mmol) were added to this solution and the mixture was
stirred at 08C. After 1 h a solution of compound 5 (0.24 g, 0.67 mmol) in
DIEA (0.5 mL, 2.68 mmol), kept at 08C, was added to the reaction mix-
ture and the mixture was stirred further for 24 h. After completion of the
reaction (TLC), volatile materials were removed under reduced pressure.
The residual material was diluted with CH2Cl2 and the organic solution
was washed with 5% HCl, 10% NaHCO3, and finally with brine. Organic
layers were combined, dried over anhyd. Na2SO4, filtered and the filtrate
was evaporated under reduced pressure. The crude residue was purified
over silica gel to afford compound 12 as a gummy solid (41% yield;
HRMS (ESI+): m/z calcd for C50H63N6O10Si [M+H]+: 935.4369; found:
935.4377), which was taken directly for the deprotection step. Thus, a so-
lution of compound 12 (0.16 g, 0.17 mmol) and 50% TFA in CH2Cl2
(5 mL) was stirred with for 4 h at room temperature. All volatile matters
were removed under reduced pressure. The crude residue was purified
over silica gel to afford the compound 13 (0.09 g, 85%); hygroscopic
white solid. 1H NMR: [D6]DMSO: d=1.68–1.75 (m, 2H), 1.79 (d, J=
8 Hz, 6H), 2.01–2.24 (m, 8H), 3.24 (m, 1H), 3.32–3.35 (m, 1H), 3.54 (dd,
J=3.6, 12 Hz, 1H), 3.62 (dd, J=2, 11.6 Hz, 1H), 3.75 (m, 2H), 4.13 (t,
J=2.8 Hz, 1H), 4.30 (m, 1H), 6.12–6.20 (m, 2H), 7.48 (s, 1H), 7.78 ppm
(s, 1H); 13C NMR [D6]DMSO: d=12.5, 12.7, 21.8 (CH2), 35.0 (CH2), 37.5
(CH2), 38.7 (CH2), 41.3 (CH2), 49.4, 61.8 (CH2), 71.6, 84.0, 84.3, 85.4,
85.6, 109.9 (C), 110.2 (C), 136.6, 136.6, 150.8 (C), 150.8 (C), 164.1 (C),
164.2 (C), 172.2 (C), 172.5 ppm (C); HRMS (ESI+): m/z calcd for
C25H35N6O10 [M+H]+: 579.2414; found: 579.2406.
Agarose-gel-based assay: Inhibition of RNase A by all dinucleosides was
checked qualitatively by the degradation of tRNA in an agarose gel. In
this method, of RNase A (20 mL; 0.66 mm) was mixed with the com-
pounds (10, 15, and 20 mL; 0.92 mm) to a final volume of 50 mL and the
resulting solutions were incubated for 6 h at 378C. Aliquots (20 mL) from
incubated mixtures were then mixed with of tRNA solution (20 mL;
5.0 mgmLꢀ1) and sample buffer (10 mL; containing 10% glycerol and
0.025% bromophenol blue). The mixture was then incubated for another
30 min. A volume of 15 mL from each solution were extracted and loaded
onto a 1.1% agarose gel. The gel was run in 0.04m Tris/acetic acid/EDTA
(TAE) buffer (pH 8.0). The undegraded tRNA was visualized by ethid-
ium bromide staining under UV light.
Precipitation assay: Inhibition of the ribonucleolytic activity of RNase A
and angiogenin were quantified by the precipitation assay as described
by Bond.[42] In this method RNase A (10 mL; 2 mm) was mixed with each
dinucleoside (50 mL; 1 mm) to a final volume of 100 mL and incubated for
2 h at 378C. An aliquot (20 mL) of the resulting solutions from the incu-
bated mixtures was then mixed with tRNA (40 mL; 5 mgmLꢀ1) and Tris/
HCl buffer (40 mL; pH 7.5 containing 5 mm EDTA and 0.5 mgmLꢀ1
HSA). After incubation of the reaction mixture at 258C for 30 min, of
ice-cold perchloric acid (200 mL, 1.14n; containing 6 mm uranyl acetate)
was added to quench the reaction. The solution was then kept in ice for
another 30 min and centrifuged at 48C at 12000 rpm for 5 min. An ali-
quot (100 mL) of the supernatant was taken and diluted to 1 mL. The
change in absorbance at 260 nm was measured and compared to a control
Compound 15: A solution of compound 14 (0.24 g, 0.2 mmol) and 50%
TFA in CH2Cl2 (7 mL) was stirred for 4 h at room temperature. All vola-
tile materials were removed under reduced pressure. The crude residue
was purified over silica gel to afford the compound 15 (0.093 g, 71%);
Chem. Eur. J. 2012, 18, 1618 – 1627
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1625

T. Pathak, S. Dasgupta and J. Debnath
set. For angiogenin the same method was followed. The concentration of
angiogenin and the inhibitors were 0.25 and 25 mm respectively.
[5] F. Soncin, Proc. Natl. Acad. Sci. USA 1992, 89, 2232–2236.
[6] G. F. Hu, J. F. Riordan, Biochem. Biophys. Res. Commun. 1993, 197,
682–687.
Inhibition kinetics: The inhibition of RNase A by 23, 21, 15, 11 was as-
sessed individually by a spectrophotometric method as described by An-
derson and co-workers.[40] The assay was performed in oligovinylsulfonic
acid free[43] Mes-NaOH buffer (0.1m, pH 6.0 containing 0.1m NaCl) using
2’, 3’-cCMP as the substrate. For 23, 21, 15, 11 the inhibitor concentra-
tions were ranged from 0–0.18, 0–0.18, 0–0.39 and 0–0.42 mm, respectively,
and the substrate concentrations were used 275–580, 320–650, 220–500
and 350–750 mm, respectively. RNase A concentration was 7.5 mm for all
the four compounds. The inhibition constants (Ki) were determined from
initial velocity data. The reciprocal of initial velocity was plotted against
the inhibitor concentration (Dixon Plot) according to the Equation (1) in
which v is the initial velocity, [S] the substrate concentration, [I] the in-
hibitor concentration, Km the Michaelis constant, Ki the inhibition con-
stant and Vmax the maximum velocity.
[7] G. F. Hu, J. F. Riordan, b. L. Vallee, Proc. Natl. Acad. Sci. USA
1994, 91, 12096–12100.
[8] S. I. Jimi, K. I. Ito, K. Kohno, M. Ono, M. Kuwano, Y. Itagaki, H.
Isikawa, Biochem. Biophys. Res. Commun. 1995, 211, 476–483.
[9] J. Folkman, M. Klagsbrun, Science 1987, 235, 442–447.
[10] P. A. Leland, K. E. Staniszewski, C. Park, B. R. Kelemen, R. T.
Raines, Biochemistry 2002, 41, 1343–1350.
[11] a) G. J. Gleich, D. A. Loegering, M. P. Bell, J. L. Checkel, S. J. Ac-
kerman, D. J. McKean, Proc. Natl. Acad. Sci. USA 1986, 83, 3146–
3150; b) D. D. Leonidas, E. Boix, R. Prill, M. Suzuki, R. Turton, K.
Minson, G. J. Swaminathan, R. J. Youle, K. R. Acharya, J. Biol.
Chem. 2001, 276, 15009–15017.
[12] C. G. Mohan, E. Boix, H. R. Evans, Z. Nikolovski, M. V. Nogueꢂs,
C. M. Cuchillo, K. Ravi Acharya, Biochemistry 2002, 41, 12100–
12106.
[13] B. M. Fisher, J. E. Grilley, R. T. Raines, J. Biol. Chem. 1998, 273,
34134–34138.
ꢀ
ꢁ
1
v
Km
Vmax½SꢂKi
1
Vmax
Km
½Sꢂ
¼
½Iꢂ þ
1 þ
ð1Þ
[14] R. Shapiro, J. F. Riordan, B. L. Vallee, Biochemistry 1986, 25, 3527–
3532.
[15] R. Shapiro, S. Weremowicz, J. F. Riordan, B. L. Vallee, Proc. Natl.
Acad. Sci. USA 1987, 84, 8783–8787.
[16] R. Shapiro, B. L. Vallee, Biochemistry 1989, 28, 7401–7408.
[17] T. P. Curran, R. Shapiro, J. F. Riordan, Biochemistry 1993, 32, 2307–
2313.
[18] A. Russo, K. R. Acharya, R. Shapiro, Methods Enzymol. 2001, 341,
629–648.
[19] K. Kumar, J. L. Jenkins, A. M. Jardine, R. Shapiro, Biochem. Bio-
phys. Res. Commun. 2003, 300, 81–86.
[20] G. I. Yakovlev, V. A. Mitkevich, A. A. Makarov, Mol. Biol. 2006, 40,
867.
[21] A. G. McLennan, Pharmacol. Ther. 2000, 87, 73–89.
[22] D. D. Leonidas, R. Shapiro, L. I. Irons, N. Russo, K. R. Acharya,
Biochemistry 1999, 38, 10287–10297.
[23] A. M. Jardine, D. D. Leonidas, J. L. Jenkins, C. Park, R. T. Rains,
K. R. Acharya, R. Shapiro, Biochemistry 2001, 40, 10262–10272.
[24] K. Tonan, P. Xu, J. L. Jenkins, A. Russo, R. Shapiro, F. Ni, Biochem-
istry 2003, 42, 11137–11149.
Purchased recombinant human angiogenin contains a substantial amount
of bovine serum albumin (BSA). Pure angiogenin was obtained after re-
moval of BSA using Amicon Ultra centrifugal filter devices (Millipore)
with a 30 kD molecular weight cutoff. Purified angiogenin was checked
to be free of contaminating ribonucleases by zymogram electrophore-
sis.[44,45] 6-FAM-dArUdAdA-6-TAMRA was used as a substrate for the
kinetic experiment of human angiogenin. Kinetic experiments were car-
ried out in oligovinylsulfonic acid free[43] Mes-NaOH buffer (0.1m, pH 6.0
containing 0.1m NaCl), using substrate concentrations from 20 to 122 nm.
For 23 and 21, the concentration ranged from 0 to 1.1 and 0 to 2.3 mm,
respectively. The angiogenin concentration was 0.23 mm. An increase in
fluorescence emission intensity at 515 nm, upon excitation at 490 nm, in-
dicates the progress of the reaction.[46] The inhibition constants (Ki) were
determined from initial velocity data using Lineweaver–Burk plots—
Equation (2), in which v is the initial velocity, [S] the substrate concentra-
tion, [I] the inhibitor concentration, Km the Michaelis constant, Ki the in-
hibition constant and Vmax the maximum velocity.
ꢂ
ꢃ
ꢂ
ꢃ
1
v
Km
Vmax
½Iꢂ
1
½Sꢂ
1
Vmax
½Iꢂ
¼
1 þ
þ
1 þ
ð2Þ
Ki
Ki
[25] J. Debnath, S. Dasgupta, T. Pathak, Bioorg. Med. Chem. 2009, 17,
4921–4927.
Docking studies: The crystal structure of RNase A and human angioge-
nin (PDB entry 1FS3 and 1B1I, respectively) was downloaded from the
Protein Data Bank.[47] The 3D structures of the backbone modified dinu-
cleosides were generated by Sybyl6.92 (Tripos Inc., St. Louis, USA) and
their energy-minimized conformations were obtained with the help of the
MMFF96 force field using MMFF96 charges with a gradient of 0.005 kcal
molꢀ1. The FlexX software as part of the Sybyl suite was used for docking
of the dinucleosides with RNase A. PyMol[48] was used for visualization
of the docked conformations.
[26] J. Debnath, S. Dasgupta, T. Pathak, Bioorg. Med. Chem. 2009, 17,
6491–6495.
[27] A. Samanta, D. D. Leonidas, S. Dasgupta, T. Pathak, S. E. Zogra-
phos, N. G. Oikonomakos, J. Med. Chem. 2009, 52, 932–942.
[28] G. H. Jones, M. Taniguchi, D. Tegg, J. G. Moffatt, J. Org. Chem.
1979, 44, 1309–1317.
[29] R. S. Varma, M. E. Hogan, Tetrahedron Lett. 1992, 33, 7719–7720.
[30] I. Kers, J. Stawinski, A. Kraszewski, Tetrahedron 1999, 55, 11579–
11588.
[31] H. Ikeuchi, M. E. Meyer, Y. Ding, J. Hiratake, N. G. J. Richards,
Bioorg. Med. Chem. 2009, 17, 6641–6650.
[32] M. Albrecht, M. Napp, M. Schneider, Synthesis 2001, 0468–0472.
[33] a) M. Bartra, P. Romea, F. Urpꢃ, J. Vilarrrasa, Tetrahedron 1990, 46,
587–594; b) R. C. Claussen, B. M. Rabatic, S. I. Stupp, J. Am. Chem.
Soc. 2003, 125, 12680–12681.
[34] H. B. Lazrek, J. W. Engels, W. Pfleiderer, Nucleoside Nucleotide
1998, 17, 1851–1856.
[35] There is only one report so far on the use of backbone-modified
non-phosphate linked dinucleosides as RNase A inhibitors, see: N.
Thiyagarajan, B. D. Smith, R. T. Raines, K. R. Acharya, FEBS J.
2011, 278, 541–549.
[36] M. Irie, H. Watanabe, K. Ohgi, M. Tobe, G. Matsumura, Y. Arata,
T. Hirose, S. Inayama, J. Biochem. 1984, 95, 751–759.
[37] N. Russo, R. Shapiro, J. Biol. Chem. 1999, 274, 14902–14908.
[38] J. Debnath, S. Dasgupta, T. Pathak, Bioorg. Med. Chem. 2010, 18,
8257–8263.
Acknowledgements
J.D. thanks CSIR, India for a fellowship. S.D. and T.P. are grateful to the
Department of Science and Technology (DST), New Delhi, India, for fi-
nancial support. J.D. thanks Dr. Kalyan S. Ghosh for his unforgettable
help in HPLC purification.
[1] J. W. Fett, D. J. Strydom, R. R. . Lobb, E. M. Alderman, J. L. Be-
thune, J. F. Riordan, B. L. Vallee, Biochemistry 1985, 24, 5480–5486.
[2] D. J. Strydom, Cell. Mol. Life Sci. 1998, 54, 811–824.
[3] J. Badet, F. Soncin, J. D. Guitton, O. Lamare, T. Cartwright, D. Bar-
ritault, Proc. Natl. Acad. Sci. USA 1989, 86, 8427–8431.
[4] R. Bicknell, B. L. Vallee, Proc. Natl. Acad. Sci. USA 1988, 85, 5961–
5965.
[39] M. Sela, C. B. Anfinsen, Biochim. Biophys. Acta. 1957, 24, 229–235.
1626
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1618 – 1627

Dinucleosides with Non-Natural Backbones
FULL PAPER
[40] D. G. Anderson, G. G. Hammes, F. G. Walz, Biochemistry 1968, 7,
1637–1645.
[46] B. R. Kelemen, T. A. Klink, M. A. Behlke, S. R. Eubanks, P. A.
Leland, R. T. Raines, Nucleic Acids Res. 1999, 27, 3696–3701.
[47] H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H.
Weissig, I. N. Shindyalov, P. E. Bourne, Nucleic Acids Res. 2000, 28,
235–242.
[41] R. T. Raines, M. P. Toscano, D. M. Nierengarten, J. H. Ha, R. Auer-
bach, J. Biol. Chem. 1995, 270, 17180–17184.
[42] M. D. Bond, Anal. Biochem. 1988, 173, 166–173.
[43] B. D. Smith, M. B. Soellner, R. T. Raines, J. Biol. Chem. 2003, 278,
20934–20938.
[48] W. L. DeLano, The PyMOL Molecular Graphics System, DeLano
Scientific, San Carlos, CA, 2004. USA. http://pymol.sourceforge.net/.
[44] A. Blank, R. H. Sugiyama, C. A. Dekker, Anal. Biochem. 1982, 120,
267–275.
Received: September 8, 2011
[45] J.-S. Kim, R. T. Raines, Protein Sci. 1993, 2, 348–356.
Published online: January 19, 2012
Chem. Eur. J. 2012, 18, 1618 – 1627
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1627