Synthesis of N3-substituted carboranyl thymidine bioconjugates and their evaluation as substrates of recombinant human thymidine kinase 1

Article abstract of DOI:10.1016/j.ejmech.2012.11.041

Four different libraries of overall twenty three N3-substituted thymidine (dThd) analogues, including eleven 3-carboranyl thymidine analogues (3CTAs), were synthesized. The latter are potential agents for Boron Neutron Capture Therapy (BNCT) of cancer. Li

Full text of DOI:10.1016/j.ejmech.2012.11.041

Accepted Manuscript
Synthesis of N3-Substituted Carboranyl Thymidine Bioconjugates and their
Evaluation as Substrates of Recombinant Human Thymidine Kinase 1
Hitesh K. Agarwal, Craig A. McElroy, Elena Sjuvarsson, Staffan Eriksson, Michael V.
Darby, Werner Tjarks
PII:
S0223-5234(12)00718-0
10.1016/j.ejmech.2012.11.041
EJMECH 5867
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 16 October 2012
Revised Date: 23 November 2012
Accepted Date: 29 November 2012
Please cite this article as: H.K. Agarwal, C.A. McElroy, E. Sjuvarsson, S. Eriksson, M.V. Darby, W.
Tjarks, Synthesis of N3-Substituted Carboranyl Thymidine Bioconjugates and their Evaluation as
Substrates of Recombinant Human Thymidine Kinase 1, European Journal of Medicinal Chemistry
(2013), doi: 10.1016/j.ejmech.2012.11.041.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Highlights
•
•
•
•
Four libraries of N3-carboranyl thymidine analogues (3CTAs) were synthesized.
3CTAs were evaluated for improved binding to human thymidine kinase 1.
Selected 3CTAs were subjected to in-depth enzyme kinetics studies (rk /K ).
cat
m
Solubility studies of 3CTAs were conducted in PBS at pH 7.4.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Synthesis of N3-Substituted Carboranyl Thymidine Bioconjugates and their
Evaluation as Substrates of Recombinant Human Thymidine Kinase 1
1
1
2
2
Hitesh K. Agarwal,* Craig A. McElroy, Elena Sjuvarsson, Staffan Eriksson, Michael
1
1
V. Darby, and Werner Tjarks*
1
Division of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, The Ohio
State University.
2
Department of Anatomy, Physiology and Biochemistry, Swedish University of Agricultural
Sciences, 751 23 Uppsala, Sweden.
*
Corresponding authors:
Hitesh K Agarwal: Division of Medicinal Chemistry & Pharmacognosy, The Ohio State
th
University, 500 West 12 Avenue, Room 422, Columbus, OH 43210, USA; Tel.: +1-
6
14-688-3149; E-mail address: agarwal.111@osu.edu
Werner Tjarks: Division of Medicinal Chemistry & Pharmacognosy, The Ohio State
th
University, 500 West 12 Avenue, Columbus, OH 43210, USA; Tel.: +1-614-292-7624;
Fax: +1-614 292 2435; E-mail address: tjarks.1@osu.edu
1

ACCEPTED MANUSCRIPT
Abbreviations: ATP, Adenosine triphosphate; 3CTAs, 3-Carboranyl Thymidine
1
0
Analogues; BNCT, Boron Neutron Capture Therapy; PBS, phosphate buffered saline; B,
boron-10; hTK1, human thymidine kinase 1; KMT, Kinase Mediated Trapping; DMSO,
dimethyl sulfoxide; dThd, thymidine; DMF, dimethyl formamide; TBDMS, tert.-
butyldimethylsilyl; TBAF, tetrabutylammonium fluoride; HR-MS, High Resolution Mass
Spectra; LR-MS, Low Resolution Mass Spectra: ESI-MS, Electrospray Ionization Mass
Spectra; AUCs, areas under the curves; PTA, Phosphorylation Transfer Assay; BSA,
Bovine serum albumin; dTMP, thymidine monophosphate.
2

ACCEPTED MANUSCRIPT
Abstract
Four different libraries of overall twenty three N3-substituted thymidine (dThd)
analogues, including eleven 3-carboranyl thymidine analogues (3CTAs), were
synthesized. The latter are potential agents for Boron Neutron Capture Therapy (BNCT)
of cancer. Linker between the dThd scaffold and the m-carborane cluster at the N3-
position of the 3CTAs contained amidinyl-(3e and 3f), guanidyl-(7e-7g),
tetrazolylmethyl-(9b1/2-9d1/2), or tetrazolyl groups (11b1/2-11d1/2) to improve human
thymidine kinase 1 (hTK1) substrate characteristics and water solubilities compared with
st
1
generation 3CTAs, such as N5 and N5-2OH. The amidinyl- and guanidyl-type N3-
substitued dThd analogues (3a-3f and 7a-7g) had hTK1 phosphorylation rates of <30%
relative to that of dThd, the endogenous hTK1 substrate, whereas the tetrazolyl-type N3-
substitued dThd analogues (9a, 9b1/2-9d1/2 and 11a, 11b1/2-11d1/2) had relative
phosphorylation rates (rPRs) of >40%. Compounds 9a, 9b1/2-9d1/2 and 11a, 11b1/2-
1
1d1/2 were subjected to in-depth enzyme kinetics studies and the obtained rk /Km
cat
(
k /K relative to that of dThd) ranged from 2.5-26%. The tetrazolyl-type N3-substitued
cat m
dThd analogues 9b1/2 and 11d1/2 were the best substrates of hTK1 with rPRs of 52.4%
and 42.5% and rk /K values of 14.9% and 19.7% respectively. In comparison, the rPR
cat
m
and rk /K values of N5-2OH in this specific study were 41.5% and 10.8%, respectively.
cat
m
Compounds 3e and 3f were >1,900 and >1,500 times, respectively, better soluble in PBS
pH 7.4) than N5-2OH whereas solubilities for 9b1/2-9d1/2 and 11b1/2-11d1/2 were
only 1.3 – 13 times better.
(
3

ACCEPTED MANUSCRIPT
Keywords:
Thymidine, m-Carborane, Boron Neutron Capture Therapy (BNCT), human Thymidine
Kinase 1 (hTK1), Thymidine conjugates.
1
. Introduction
Thymidine (dThd) bioconjugates with carborane clusters (3-carboranyl dThd
analogues or 3CTAs) have been studied extensively for Boron Neutron Capture Therapy
BNCT), which is a radio-chemotherapeutic modality for cancer treatment [1,2]. These
(
bioconjugates were designed for selective uptake and retention in tumor cells by kinase
mediated trapping (KMT) through 5′-monophosphorylation by human thymidine kinase 1
(hTK1), an enzyme of the salvage pathways of DNA precursor synthesis with high
activity levels only in rapidly proliferating cells [1,3]. So far, KMT has been employed
primarily in positron emission tomography (PET) of tumors using 3′-deoxy-3′-
18
18
[
F]fluorothymidine [ F-FLT] as the imaging agent [4].
First generation 3CTAs, such as N5 and N5-2OH (Fig. 1), were found to be
substrates of hTK1 in enzyme kinetic studies [5-9]. The in vitro uptake of N5 and N5-
OH in L929 TK (+) cells was 5-10 fold higher than in L929 TK (-) cells [6].
Biodistribution studies with N5-2OH in rats bearing intracranial RG2 glioma resulted in
0 fold higher boron concentrations in the tumor than in normal brain (27.6 µg B/g tumor
2
1
vs. 2.6 µg B/g tumor) [8]. BNCT with rats bearing intracranial RG2 glioma that had
received N5-2OH prior to irradiation increased survival times by a factor two compared
with BNCT of rats bearing RG2 glioma that had not received N5-2OH [8].
4

ACCEPTED MANUSCRIPT
These biological studies clearly established the therapeutic potential for 3CTAs. On
the other hand, the same studies [5-9] revealed two shortcomings of N5 and N5-2OH: 1)
The lipophilic nature of N5-2OH caused low water-solubility, which necessitated the use
of organic solvents, such as DMSO, to solubilize the agent for biological studies. 2)
Compared with endogenous dThd, N5 and N5-2OH show suboptimal hTK1 substrate
characteristics.
X-ray crystallographic data with various TK1-like enzymes indicated that hydrogen
bonds between the N3-H, C2=O, and C4=O of dThd and the amide backbone of the so-
called “lasso-loop” [10,11] of the enzymes are crucial for substrate binding [10-16]. In
the case of hTK1 co-crystallized with dThd triphosphate, N3-H and C2=O of the dThd
scaffold interacted with Val 172 and Val 174 of the lasso-loop [10,13]. Similar studies
with Bacillus anthracis TK (BaTK) indicated a hydrogen bond network between Tyr 179
and Arg 157 of the lasso-loop and N3-H and C4=O of dThd [1]. In the case of
Ureaplasma urealyticum TK (UuTK), hydrogen bonds occurred between N3-H and
C4=O of dThd and Ile 178 and Lys 180 of the lasso-loop, respectively [10,11]. In all of
these cases, N3-H and C2=O and/or C4=O acted as proton donors and acceptors,
respectively. Sequra-Peña et al. studied in depth the structural changes in Thermotoga
maritima thymidine kinase (TmTK) as a result of dThd and ATP binding [14].
Apparently, interactions similar to those discussed above contributed to dramatic
conformational changes in the lasso-loop resulting in the conversion of an open (apo)
form to a closed (holo) form of the enzyme, which locks dThd tightly into the substrate
binding pocket [12,14]. In the cases of N5 and N5-2OH, the N3-proton is replaced by a
bulky carboranyl alkyl chain. Thus, hydrogen bonds crucial for the induction of closed
5

ACCEPTED MANUSCRIPT
conformations of TK1-like enzymes are missing [1]. We hypothesize that this may limit
the abilities of N5 and N5-2OH to interact with the substrate-binding site of the enzyme.
In addition, computational studies indicated that N5-2OH cannot be docked effectively
into the substrate binding pockets of holo TK1-like enzymes [5] whereas docking was
possible in the case of an apo form. This suggests that the presence of the bulky
carborane cluster may also interfere with the induction of a completely closed enzyme
confirmation. Since the final binding confirmation of TK1-like enzymes with N5-2OH
and other 3CTAs is unknown, computer-aided structural design is not an effective tool to
support structure-activity-relationship (SAR) studies in pursuit of improved 3CTAs.
A bulky carboranyl substituent at N3 is the trademark of any 3CTA and this cannot
be changed. On the other hand, it is conceivable that the missing hydrogen bond
interaction between NH-3 and C=O, in part, can be compensated by introducing
functional groups in the vicinity of N3 of the 3CTA scaffold having the ability to re-
introduce hydrogen bonding. Thus, this paper discusses the synthesis, hTK1 substrate
characteristics, and the results of water solubility measurements of four novel classes of
N3-substituted dThd analogues: N3-amidinyl- (3a-3f), N3-guanidyl- (7a-7g), N3-(1-
tetrazolylmethyl)- (9a, 9b1/2-9d1/2), and N3-tetrazolyl-dThd analogues (11a, 11b1/2-
1
1d1/2) [Schemes 1-4]. The amidinyl- and guanidyl functions in the spacer between the
dThd scaffold and the m-carborane cage of target compounds 3a-3f and 7a-7g serve as
proton donors. They may restore hydrogen bond interactions with proton accepting atoms
in amino acid side chains as well as in the peptide back bone of the lasso loop of hTK1
thereby inducing a more closed enzyme conformation. This, in turn, may render these
types of dThd analogues more tightly bound to the active site. As a result, target
6

ACCEPTED MANUSCRIPT
compounds 3a-3f and 7a-7g may have better enzyme substrate characteristics in
comparison with N5-2OH. In addition, the basic nitrogen atoms in these structures
should be protonated under physiological conditions, which will improve their water-
solubilities. In contrast to the amidinyl- and guanidyl groups in target compounds 3a-3f
and 7a-7g, the tetrazolyl moieties in target compounds 9a, 9b1/2-9d1/2 and 11a, 11b1/2-
1
1d1/2 have proton accepting ability only. Nevertheless, they may still interact with
various proton donating atoms in amino acid side chains (e.g. Arg 165, Lys 170, Lys 180)
and in the peptide back bone of the lasso-loop thereby facilitating a closed enzyme
conformation. Lesnikowski et al. recently described the synthesis of several N3-
substituted triazolylcarboranyl dThd analogues [17,18] that may have similar TK1-
binding properties than the tetrazolyl-type N3-substitued dThd analogues (9a, 9b1/2-
9
d1/2 and 11a, 11b1/2-11d1/2) described in this paper.
2
. Results and discussion
2
.1. Chemistry
Four different libraries of N3-substituted dThd analogues were synthesized in two
general steps. In the first step, 3-(cyanomethyl)thymidine (2, Scheme 1) and 3',5'-(di-
tert.-butyldimethylsilyl)-3-cyanothymidine (5, Scheme 2) were synthesized by the
reaction of dThd and 3′,5′-di-TBDMS-protected dThd, with iodoactonitrile and cyanogen
bromide, respectively. In the second step, these reactive nitriles were treated with various
m-carboranylalkyl- [19], alkyl-, and arylamines to yield amidinyl- and guanidyl-type N3-
substitued dThd analogues (3a-3f and 7a-7g, Schemes 1 and 2). Compounds 2 and 5
were also treated with sodium azide to produce the tetrazolyl-type dThd analogues 8 and
7

ACCEPTED MANUSCRIPT
1
0 (Schemes 3 and 4), which were subsequently treated with various m-carboranylalkyl-
[
19] and alkyliodides to yield the tetrazolyl-type target compounds 9a, 9b1/2-9d1/2 and
1a, 11b1/2-11d1/2 (Schemes 3 and 4). m-closo-Carborane was used for the synthesis of
1
all carboranyl target compounds (3e-3f, 7e-7g, 9b1/2-9d1/2 and 11b1/2-11d1/2) because
it is more resistant to basic degradation than o-closo-carborane [19], which could occur
particularly during the synthesis and biological evaluation of the amidine- and guanidine
containing compounds 3e, 3f and 7e-7g.
3
-(Cyanomethyl)thymidine (2) was synthesized in 78% yield by selective N-
alkylation at the 3-position of dThd with iodoacetonitrile (Scheme 1) in presence of
potassium carbonate in DMF/acetone (1/1, v/v) at 60 °C for 6 h. Compound 2 was
purified by silica gel column chromatography. For the synthesis of the non-carboranyl
amidine-type dThd analogues 3a-3d, compound 2 and alkyl- or arylamines were heated
in a closed pressure reaction vessel (Quark Glass) in the presence of
hexafluoroisopropanol (HFIP) overnight at 95 °C. The use of HFIP for this type of
reaction has been previously reported by Snider and O’Hare [22]. The use of methanol,
tert.-butanol, tetrahydrofuran, and acetonitrile as solvents did not lead to product
formation in this study [22]. The reaction mixtures were added to water and extracted
with ethyl acetate. The aqueous layers were separated, evaporated, and purified by RP-18
HPLC. Yields ranged from 24% to 81%.
For the synthesis of the m-carboranylalkyl target compounds 3e and 3f, the same
method proved unsuccessful and a modified strategy was developed. Equimolar
quantities of 2, m-carboranyl alkyl amine [19], and 1 mL of HFIP were placed in a round
bottom flask and heated overnight at 120 °C without condenser. Heating resulted in
8

ACCEPTED MANUSCRIPT
complete solvent evaporation and formation of 3e and 3f in yields of 7% and 10%,
respectively, after purification by RP-18 HPLC.
The synthesis of N3-guanidine-type dThd analogues (7a-7g, Scheme 2) was carried
out with 3′,5′-di-TBDMS-protected dThd (4) [20] as the starting material, because initial
attempts to treat dThd with cyanogen bromide in presence of potassium carbonate
resulted in <5% yield of the desired product, 3-cyanothymidine. Reaction of 4 with
cyanogen bromide in presence of triethylamine (TEA) in acetone at 0 °C gave 3',5'-(di-
tert.-butyldimethylsilyl)-3-cyanothymidine (5) in 90% yield after column
chromatographic purification [21].
Compounds 7a-7g were synthesized by treating 5 with the amines shown in Scheme
2
in presence of HFIP in a closed pressure reaction vessel (Quark Glass) [22]. The
reactions of 5 with propyl- and butyl amine were carried out in an ice bath for 5 h to give
a and 6b, respectively, in > 95% yield whereas phenyl- and benzyl amines required
6
overnight reaction at room temperature to yield 6c and 6d, respectively, in > 90% yield.
The reaction of 1-(2-aminoethyl)-m-carborane [19], 1-(3-aminopropyl)-m-carborane [19],
and 1-(4-aminobutyl)-m-carborane [19] with 5 was carried out overnight at 95 °C to
afford 6e, 6f, and 6g, respectively, each in about 30-40% yield. Compounds 6a-6g were
dissolved in trifluoroacetic acid (TFA)/dichloromethane (10 %, v/v) and stirred at room
temperature for 1 h producing 7a-7g in 40-82% yield after purification by RP-18 HPLC.
Both N3-amidinyl- (3a-3f) and guanidyl-type (7a-7g) dThd analogues were found to be
unstable at room temperature (over several days) and were therefore stored at -5 °C.
However, even under these storage conditions partial to complete degradation was
observed after a ~ 1 year period.
9

ACCEPTED MANUSCRIPT
The synthesis of tetrazolylmethyl- and tetrazolyl-type target compounds 9a, 9b1/2-
d1/2 and 11a, 11b1/2-11d1/2 is described in Schemes 3 and 4, respectively. These
9
compounds were prepared using a click chemistry approach [23]. The reactive nitrile
group of 2 was treated with sodium azide in the presence of zinc bromide (ZnBr ) in a
2
mixture of water and isopropanol (2:1, v/v) at 90 °C for three days to produce 8 in 90%
yield (Scheme 3). In case of 5, the same reaction conditions led to a mixture of 3-
(tetrazol-5-yl)thymidine (10, Scheme 4) along with 3′- or 5′-mono-TBDMS and 3′,5′-(di-
TBDMS)-3-(tetrazol-5-yl)thymidine. Deprotection of remaining TBDMS-groups was
accomplished in situ by adding TFA in dichloromethane (10%, v/v) to the reaction
mixture for 1 h at room temperature to give 10 in 44% overall yield (Scheme 4). Both 8
and 10 were purified by RP-18 HPLC.
Compound 8 and 10 were N-alkylated at the tetrazolyl ring with 1-iodopropane, 1-(2-
iodoethyl)-m-carborane [19], 1-(3-iodopropyl)-m-carborane [19], and 1-(4-iodobutyl)-m-
carborane [19] in the presence of potassium carbonate in a mixture of acetone and
dimethylformamide (50%, v/v) at 60 °C overnight to produce target compounds 9a and
b
c
1
1a as well as 9b1/2-9d1/2 and 11b1/2-11d1/2 as isomeric (N -alkylated and N -
alkylated) mixtures (Schemes 3 and 4). In the case of the propyl analogues 9a and 11a it
c
was possible to isolate the N -isomers as pure compounds by RP-18 HPLC with 59 % to
7
1% yield, respectively. Unfortunately, it was not possible to separate the isomers of the
c
carboranylalkyl derivatives 9b1/2-9d1/2 and 11b1/2-11d1/2. Overall, the N -isomers
were the dominating species in these mixtures with 70-95% (Schemes 3 and 4). The
1
b
isomers were characterized by means of H-NMR. The chemical shifts for the N -CH
2
c
protons ranged from 4.44-4.62 ppm, whereas those for the N -CH protons ranged from
2
1
0

ACCEPTED MANUSCRIPT
c
b
4
.54-4.82 ppm. A downfield shift of 0.15-0.53 ppm for N -CH protons vs. the N -CH
2
2
protons was also reported by Sveshnikov and Nelson [24] for various alkyltetrazoles. The
c
b
observed N :N isomeric ratios were in accordance with those reported previously for
1
1
other alkylation reactions at tetrazole rings [25]. In addition, H- H NOESY NMR of the
b
mixture of 9b1 and 9b2 showed a NOE cross-peak between the triplet for N -CH (δ
2
4
.52) and the singlet for N3-CH (5.35 ppm). In contrast, no cross peak was observed
2
c
between the signals of N -CH and N3-CH (see Supplementary material). This result was
2
2
b
expected based on a shorter interproton distance between N -CH and N3-CH (2.5 Å)
2
2
c
than between N -CH and N3-CH (4.9 Å) (calculated with Chem3D Pro [Chambridge
2
2
Soft, Cambridge, MA, USA]). It unambiguously confirmed the isomer assignment.
2
.2. Enzymatic studies
Compounds 3a-3f, 7a-7g, 8, 9a, 9b1/2-9d1/2, 10, and 11a, 11b1/2-11d1/2 were
evaluated in phosphoryl transfer assays (PTAs) with recombinant hTK1 as described
previously with minor modifications [26-29]. Thymidine and N5-2OH were used as
reference compounds. The obtained phosphorylation rates are expressed as percent of the
phosphorylation rate of dThd, which was set to 100% (rPRs). Such comparison is used to
give an estimation of the structure activity relationships of substrate analogues compared
to the activity of the natural substrate, dThd. Analogues with relatively high PRs were
further selected for in depths kinetic studies (apparent k /K values) to determine the
cat
m
apparent catalytic efficiency of hTK1 for these substrates.
Most of the amidinyl-type dThd analogues (3a-3f) and all of the guanidyl-type dThd
analogues (7a-7g) proved to be unstable under phosphoryl transfer assay conditions, as
1
1

ACCEPTED MANUSCRIPT
evidenced by the production of phosphorylated side products. Fig. 2A shows the
monophosphate formation of the carboranyl compounds of both libraries (3e, 3f, 7e-7g).
For compounds 7e-7g, additional spots are visible at the level of dThd monophosphate
(dTMP). This indicates that these compounds may have been partially degraded to dThd
during the assay, which was then converted to dTMP. A Low Resolution-Electron Spray
Ionization (LR-ESI) mass spectra of an enzyme assay mixture containing compound 7f
3
1
+
and P-ATP showed a signal for the monophosphate of 7f (Calcd. for [M-PO H +H] :
3
2
+
5
3
49.3; Found: m/z 549.3) and a signal for dTMP (Calcd for [M+H] : 323.1; Found: m/z
23.1, see Supplementary material). The observed MS data support the formation of
dTMP during the enzyme assay as a result of degradation processes. Enzyme assay
mixtures containing compounds 3e and 3f did not appear to produce significant amounts
of dTMP as a side product. However, it is conceivable that in the case of both compounds,
dTMP may not be visible because of low concentrations
Compounds with alkylamidinyl- and alkylguanidyl substituents (3a, 3b, 7a, 7b) had
rPRs of 0-3%, whereas those with arylamidinyl- and arylguanidyl substituents (3c, 3d, 7c,
7
7
d) had slightly higher rPRs of 2-17% (Table 1). However, for compounds 3a-3d and
a-7d, 20-100% of the formed phosphorylated material appeared to be the dTMP side
product. Compounds with m-carboranylalkylamidinyl substituents 3e and 3f had rPRs of
% and 6% whereas the m-carboranylalkylguanidyl compounds 7e-7g had overall
3
slightly higher rPRs with 16-27% but also showed the formation of the dTMP side
product with rPRs of 6-12% (Table 1 and Fig. 2A). Overall, the rPRs of both the
amidinyl-type and the guanidyl-type N3-substitued dThd analogues were lower than that
of N5-2OH (rPR of 46%).
1
2

ACCEPTED MANUSCRIPT
In contrast to the amidinyl- (3a-3f) and guanidyl-type (7a-7g) dThd analogues, the
tetrazolylmethyl-type (8, 9a, 9b1/2-9d1/2) and the tetrazolyl-type dThd analogues (11a,
1
1b1/2-11d1/2) were stable under phosphoryl transfer assay conditions (Table 1, Fig. 2B
and Fig. 2C). With the exception of compound 10, all tetrazolyl-type dThd analogues
displayed relatively higher rPRs with 40-64%. The signals for monophosphate product of
1
0 may actually overlap with ATP related spots. The tetrazolylmethyl-type dThd
analogues with propyl- and butyl spacer between carborane and tetrazole (9c1/2, 9d1/2)
appeared to be somewhat better phosphorylated than their tetrazolyl-type counterparts
(11c1/2, 11d1/2) (58-64% vs. 42-43%).
The tetrazolyl-type dThd analogues 9a, 9b1/2-9d1/2 and 11a, 11b1/2-11d1/2 were
selected for in depths kinetic studies (apparent k /K values) based on their rPRs, which
cat
m
were higher than that of N5-2OH. Both dThd and N5-2OH were used as reference
compounds. These studies were conducted with varying substrate concentrations (0.4-300
µM) (see Experimental Section for details). The rk /K value for N5-2OH using the
cat
m
current recombinant hTK1 preparation with a specific activity of 441 nmol dThd
monophosphate formation per min and mg of enzyme was lower compared with the value
obtained with a different enzyme preparation used in a previous study (10.8% vs 35.8%)
[
9]. Narayanasamy et al. [26] reported a specific activity of 640 nmol dThd
monophosphate formation per min and mg of enzyme for a different recombinant hTK1
used in a previous study with 3CTAs. This suggests that specific activities between
enzyme preparations vary significantly and that this, in turn, may cause variations in the
rk /K values of 3CTAs. The rk /K values for compound 9a, 9c1/2, 9d1/2, and 11b1/2
cat
m
cat
m
were comparable with that of N5-2OH (7.6-12.9% vs. 10.8%). Compounds 9b1/2
1
3

ACCEPTED MANUSCRIPT
(
14.9%), 11a (26.0%) and 11d1/2 (19.7%) had notably higher rk /K values indicating
cat
m
that these compounds are better substrates of hTK1 than N5-2OH.
Based on the isomeric composition of the tetrazoyl target compounds 9a, 9b1/2-9d1/2
and 11a, 11b1/2-11d1/2 (Schemes and 4), it is reasonable to assume that their observed
c
enzymatic activities are primarily related to the N -isomers.
2
.3. Solubility studies
The solubilities of carboranyl compounds that appeared to be stable during enzyme
assays (3e, 3f, 9b1/2-9d1/2, and 11b1/2-11d1/2) were determined in phosphate-buffered
saline (PBS) at 25 °C in comparison with N5-2OH using thermodynamic solubility
measurement assay conditions (Table 3) [27,28]. The carboranylethyl dThd analogue 3e
and its propyl homologue 3f were > 1900 and >1500 times, respectively, better soluble in
PBS than N5-2OH. The solubility patterns for the tetrazolylmethyl-(9b1/2-9d1/2) and
tetrazolyl- (11b1/2-11d1/2) dThd analogues decreased with increasing chain length.
Compounds 9b1/2-9d1/2 were ~3-13 times better soluble in PBS than N5-2OH whereas
compounds 11b1/2 and 11c1/2 were ~2-4 times better soluble. The solubility profile of
1
1d1/2 was similar to that of N5-2OH. Overall, the tetrazolylmethyl derivatives were
slightly better soluble than their corresponding tetrazolyl counterparts.
3
. Summary and Conclusions
In summary, several N3-substituted dThd analogues containing amidinyl-(3a-3f),
guanidyl-(7a-7g), tetrazolylmethyl-(8, 9a, 9b1/2-9d1/2), tetrazolyl groups (10, 11a,
1b1/2-11d1/2) were successfully synthesized. The amidinyl- and guanidyl-type dThd
1
1
4

ACCEPTED MANUSCRIPT
analogues (3a-3f and 7a-7g) did not possess the necessary stability under aqueous
conditions. The excellent solubilities of 3e and 3f in PBS were expected because of the
presence of basic amidinyl groups in these structures, which should be protonated at
physiological pH (7.4). Unfortunately, the solubilities of the tetrazolyl-type dThd
analogues in PBS were only moderately better than that of N5-2OH. The
tetrazolylmethyl- (8, 9a, 9b1/2-9d1/2) and tetrazolyl-type dThd analogues (10, 11a,
1
2
1b1/2-11d1/2) had comparable or somewhat higher rPRs (40-64%) than that of N5-
OH whereas those of the amidine- (3a-3f) and guanidine-type (7a-7g) dThd analogues
had lower rPRs with 2-12% and 0-27%, respectively. The rk /K values for the
cat
m
tetrazolylcarboranyl dThd analogues with relatively high rPRs (9a, 9b1/2-9d1/2, 11a,
1
1
1b1/2-11d1/2) ranged from 7.6-26% in comparison to N5-2OH (10.8%). Compounds
1d1/2 (rk /K = 19.7%) and 9b1/2 (rk /K = 14.9%) were ~1.8 and ~1.4 times,
cat
m
cat
m
respectively, better substrates of hTK1 than N5-2OH.
The rk /K values obtained for compounds 9b1/2 and 11d1/2 indicate that proton
cat
m
accepting groups in the spacer between carborane cluster and dThd scaffold have the
potential to improve hTK1 substrate characteristics. In addition, basic functions in the
spacer seem to enhance aqueous solubility, as demonstrated at the example of compounds
3
e and 3f. Unfortunately, the validity of enzymatic data obtained for amidinyl- and
guanidyl-type dThd analogues (3a-3f and 7a-7g) is limited due to lack of stability under
aqueous and routine storage conditions. Preliminary studies in our laboratories indicated
that inserting more than two methylene groups between the dThd scaffold and basic
amino group containing functions in the spacer prevents decomposition under aqueous
conditions. As discussed in the introduction, a major shortcoming is lack of detailed
1
5

ACCEPTED MANUSCRIPT
structural information that can assist in delineating the basic molecular interactions
between hTK1 and 3CTAs.
4
. Experimental protocol
4
.1. Chemistry
1
13
H- and C NMR spectra were obtained on a Bruker DRX 400 at The Ohio State
1
13
University College of Pharmacy (400 MHz for H and 100 MHz for C and 128 MHz for
1
1
1
1
B). H- H NOESY NMR was obtained on a Bruker DRX 800 at The Ohio State
University Campus Chemical Instrument Center. Chemical shifts (δ) are reported in ppm
from internal deuterated chloroform, acetone, and methanol. Coupling constants are
1
3
reported in Hz. C NMR spectra are fully decoupled. Low- and High Resolution–
Electrospray ionization (LR/HR-ESI) mass spectra were obtained on a Micromass LCT
spectrometer at The Ohio State University Campus Chemical Instrumentation Center,
Columbus, OH. For all carborane-containing compounds, the mass of the most intensive
peak of the isotopic pattern was reported (80% boron-11 to 20% boron-10 distribution).
Measured patterns agreed with calculated patterns. Silica gel 60 (0.063-0.200 mm) used
for gravity column chromatography and silica gel 60 (0.015-0.049 mm) used for flash
column chromatography were purchased from Dynamic Adsorbents Inc. (3280 Peachtree
Corners Circle, E, Norcross, GA 30092, USA). Reagent-grade solvents were used for
column chromatography. Precoated glass-backed TLC plates with silica gel 60 F254
(0.25-mm layer thickness) from Dynamic Adsorbents (Norcross, GA) were used for TLC.
General compound visualization for TLC was achieved by UV light. Carborane-
containing compounds were selectively visualized by spraying the plate with a 0.06%
1
6

ACCEPTED MANUSCRIPT
PdCl /1% HCl solution and heating at 120 °C, which caused the slow (15-45 s) formation
2
2
+
of a gray spot due to the reduction of Pd to Pd0. Anhydrous solvents, such as
dichloromethane, were purchased directly either from Acros Organics (Morris Plains, NJ)
or from Sigma Aldrich (Milwaukee, WI). Pressure reaction vessels were purchased from
Quark Glass (NJ, USA). Other chemicals and solvents were purchased from standard
commercial suppliers. Tetrahydrofuran (THF) and benzene were distilled from sodium
and benzophenone indicator under argon. m-Carborane was purchased from Katchem Ltd.
(Prague, Czech Republic). Cyanogen bromide (purchased from Sigma Aldrich) is a
highly toxic compound. The Material Safety Data Sheet (MSDS) should be consulted
3
2
3
before usage. [γ- P]ATP (specific activity: 3000Ci/mmol) and [methyl- H]dThd
(specific activity: 83.7 Ci/mmol) were obtained from Perkin Elmer, Waltham, MA.
Unless specified otherwise, all reactions were carried out under argon atmosphere.
Preparative HPLC was performed with a Gemini 5µ C18 column (21.20 mm x 250 mm, 5
ꢀ
m particle size) supplied by Phenomenex Inc. CA, USA on a Hitachi HPLC system (L-
2
130) with a Windows based data acquisition and Hitachi Diode array detector (L-2455).
Purification was accomplished using a water (0.1% TFA)/methanol (0.1% TFA) gradient
at 7 mL/min flow rate (method: 100:0 to 50:50 over 70 min followed by 50:50 to 0:100
over 40 min). Analytical HPLC was carried out with a Gemini 5µ C18 110A Column
(250 x 4.6 mm) supplied by Phenomenex Inc. CA, USA using the HPLC system above.
Two different gradients [water (0.1% TFA)/acetonitrile (0.1% TFA) and water (0.1%
TFA)/methanol (0.1% TFA)] were used at 1 mL/min flow rate (method: 100:0 to 75:25
over 20 min followed by 75:25 to 100:0 over 10 min). See Supplementary material for
more details on compound purification and analysis.
1
7

ACCEPTED MANUSCRIPT
4
.1.1. 3-(Cyanomethyl)thymidine (2). Thymidine (10 g, 41.3 mmol) and potassium
carbonate (17 g, 123 mmol) were dissolved in acetone/DMF (100 mL, 1/1, v/v) and
iodoacetonitile (6.5 mL, 90 mmol) was added. The reaction mixture was refluxed for 6 h,
then added to water (750 mL), and extracted with ethyl acetate (250 mL × 4). The organic
layer was separated, dried using anhydrous magnesium sulfate (100 g), evaporated under
reduced pressure, and purified by silica gel column chromatography using
dichloromethane and methanol as the solvent system. The compound was eluted at 4%
methanol. R = 0.70 (10% methanol in dichloromethane), yield = 9.1 g, 78% as yellow
f
1
solid, (m.p. 52 °C); H NMR (400 MHz, CD OD, δ ppm): 7.86 (s, 1H, H-6), 6.25 (t, J =
3
6
3
2
.6 Hz, 1H, 1'H), 4.80 (s, 2H, CH -CN), 4.33-4.41 (m, 1H, 3'H), 3.87-3.95 (m, 1H, 4'H),
2
.78 (dd, J = 2.8 and 12.0 Hz, 2H, 5''H), 3.70 (dd, J = 3.1 and 12.0 Hz, 2H, 5'H), 2.23-
1
3
.33 (m, 1H, 2''H), 2.12-2.22 (m, 1H, 2'H), 1.89 (s, 3H, 5-CH ). C NMR (CD OD, 100
3
3
MHz, δ ppm): 163.29 (C-4 C=O), 150.62 (C-2 C=O), 136.77 (C-6), 115.46 (CN), 110.30
C-5), 88.27 (C-1'), 86.79 (C-4'), 71.40 (C-3'), 62.16 (C-5'), 41.09 (C-2'), 28.92 (N3-CH2-
(
CN), 13.05 (5-CH ); HR-MS (ESI-MS) for C H N NaO . Calcd: 304.0909. Found: m/z
3
12 15
3
5
+
3
04.0888 (M+Na) .
4
.1.2. General procedure for the synthesis of 3-[2-Imino-2-(propylamino)ethyl]thymidine
(3a), 3-[2-Imino-2-(butylamino)ethyl]thymidine (3b), 3-[2-Imino-2-
(phenylamino)ethyl]thymidine (3c), and 3-[2-Imino-2-(benzylamino)ethyl]thymidine (3d).
Compound 2 (280 mg, 1 mmol) and amine (1.1 mmol) were dissolved in
2
2
hexafluoroisopropanol (1 mL) in a closed pressure vessel (Quark Glass). The reaction
1
8

ACCEPTED MANUSCRIPT
o
mixture was heated at 95 C overnight, suspended in water (80 mL), and extracted with
ethyl acetate (50 mL × 3). The aqueous layer was evaporated under reduced pressure and
the residue purified by preparative HPLC. The percentage purities of the synthesized
compounds were > 95%.
4
.1.2.1. 3-[2-Imino-2-(propylamino)ethyl]thymidine (3a). Retention time (analytical
1
HPLC) = 9.1 min, yield = 80 mg, 24% (white amorphous solid); H NMR (400 MHz,
CD OD, δ ppm): 7.97 (s, 1H, H-6), 6.31 (t, J = 6.6 Hz, 1H, 1'H), 4.87 (s, 2H, N3-CH ),
3
2
4
.38-4.48 (m, 1H, 3'H), 3.95 (dd, J = 3.2 and 6.4 Hz, 1H, 4'H), 3.82 (dd, J = 3.0 and 12.0
Hz, 1H, 5''H), 3.75 (dd, J = 3.5 and 12.0 Hz, 1H, 5'H), 3.25 (t, J = 7.2 Hz, 2H, =NH-CH2),
2
0
.18-2.35 (m, 2H, 2'H and 2''H), 1.94 (s, 3H, 5-CH ), 1.60-1.71 (m, 2H, =NH-CH -CH ),
3 2 2
1
3
.99 (t, J = 7.4 Hz, 3H, propyl-CH ). C NMR (CD OD, 100 MHz, δ ppm): 165.18 (C-4
3
3
C=O), 165.08 (HN=C-NH), 152.35 (C-2 C=O), 137.57 (C-6), 110.77 (C-5), 89.18 (C-1'),
8
2
7.51 (C-4'), 72.28 (C-3'), 62.86 (C-5'), 45.33 (N3-CH ), 42.50 (=NH-CH ), 41.43 (C-2'),
2 2
2.04 (=NH-CH -CH ), 13.15 (5-CH ), 11.52 (propyl-CH ); HR-MS (ESI-MS) for
2
2
3
3
+
C H N O . Calcd: 341.1825. Found: m/z 341.1799 (M+H) .
15
25
4
5
4
.1.2.2. 3-[2-Imino-2-(butylamino)ethyl]thymidine (3b). Retention time (analytical
1
HPLC) = 10 min, yield = 160 mg, 45% (white amorphous solid); H NMR (400 MHz,
CD OD, δ ppm): 7.98 (s, 1H, H-6), 6.32 (t, J = 6.5 Hz, 1H, 1'H), 4.88 (s, 2H, N3-CH ),
3
2
4
3
2
.37-4.47 (m, 1H, 3'H), 3.90-4.00 (m, 1H, 4'H), 3.83 (dd, J = 2.5 and 11.9 Hz, 1H, 5''H),
.76 (dd, J = 3.4 and 11.9 Hz, 1H, 5'H), 3.25-3.35 (m, 2H, Methanol-CH and NH-CH ),
3
2
.20-2.38 (m, 2H, 2'H and 2''H), 1.96 (s, 3H, 5-CH ), 1.58-1.72 (m, 2H, NH-CH -CH ),
3
2
2
1
9

ACCEPTED MANUSCRIPT
1
3
1
.36-1.50 (m, 2H, butyl-CH -CH ), 0.99 (t, J = 7.2 Hz, 3H, butyl-CH ). C NMR
2 3 3
(
CD OD, 100 MHz, δ ppm): 165.00 (C-4 C=O), 152.27 (C-2 C=O), 137.50 (C-6), 110.70
3
(
C-5), 89.09 (C-1'), 87.44 (C-4'), 72.20 (C-3'), 62.79 (C-5'), 43.47 (N3-CH ), 42.42 (NH-
2
CH ), 41.35 (C-2'), 30.63 (NH-CH -CH ), 20.95 (butyl-CH -CH ), 13.98 (butyl-CH ),
2
2
2
2
3
3
1
3
3.08 (5-CH ); HR-MS (ESI-MS) for C H N O . Calcd: 355.1981. Found: m/z
3 16 27 4 5
+
55.1985 (M+H) .
4
.1.2.3. 3-[2-Imino-2-(phenylamino)ethyl]thymidine (3c). Retention time (analytical
1
HPLC) = 7.6 min, yield = 290 mg, 79% (white amorphous solid); H NMR (400 MHz,
CD OD, δ ppm): 8.00 (s, 1H, H-6), 7.45-7.60 (m, 3H, phenyl-Ar-H), 7.34 (d, J = 7.4 Hz,
3
2
H, phenyl-Ar-H), 6.35 (t, J = 6.5 Hz, 1H, 1'H), 5.07 (dd, J = 16.5 and 23.4 Hz, 2H, N3-
CH ), 4.38-4.48 (m, 1H, 3'H), 3.95 (dd, J = 3.2 and 6.4 Hz, 1H, 4'H), 3.83 (dd, J = 3.0
2
and 12.0 Hz, 1H, 5''H), 3.76 (dd, J = 3.4 and 12.0 Hz, 1H, 5'H), 2.20-2.38 (m, 2H, 2'H
13
and 2''H), 1.98 (s, 3H, 5-CH ). C NMR (CD OD, 100 MHz, δ ppm): 166.13 (C-4 C=O),
3
3
1
1
6
65.09 (HN=C-NH), 152.35 (C-2 C=O), 137.61 (phenyl-Ar-C), 135.04 (C-6), 131.56,
30.49, 126.83 (phenyl-Ar-C), 110.75 (C-5), 89.14 (C-1'), 87.52 (C-4'), 72.21 (C-3'),
2.80 (C-5'), 42.66 (N3-CH ), 41.38 (C-2'), 13.08 (5-CH ); HR-MS (ESI-MS) for
2
3
+
C H N O . Calcd: 375.1668. Found:. m/z 375.1677 (M+H) .
18
23
4
5
4
.1.2.4. 3-[2-Imino-2-(benzylamino)ethyl]thymidine (3d). Retention time (analytical
1
HPLC) = 11.8 min, yield = 310 mg, 81% (white amorphous solid); H NMR (400 MHz,
CD OD, δ ppm): 7.92 (s, 1H, H-6), 7.20-7.40 (m, 5H, benzyl-Ar-H), 6.31 (t, J = 6.5 Hz,
3
1
H, 1'H), 4.65 (s, 2H, N3-CH ), 4.42 (s, 3H, 3'H and NH-CH ), 3.90-3.97 (m, 1H, 4'H),
2 2
2
0

ACCEPTED MANUSCRIPT
3
2
1
1
.83 (dd, J = 3.0 and 12.0 Hz, 1H, 5''H), 3.75 (dd, J = 3.4 and 12.0 Hz, 1H, 5'H), 2.20-
1
3
.38 (m, 2H, 2'H and 2''H), 1.95 (s, 3H, 5-CH ). C NMR (CD OD, 100 MHz, δ ppm):
3
3
68.79 (HN=C-NH), 164.19 (C-4 C=O), 151.30 (C-2 C=O), 138.66 (benzyl-Ar-C),
35.90 (C-6), 128.53, 127.45, 127.20 (benzyl-Ar-C), 109.53 (C-5), 87.87 (C-1'), 86.28
(
C-4'), 71.02 (C-3'), 61.73 (C-5'), 43.35 (N3-CH ), 43.12 (benzyl-CH ), 40.38 (C-2'),
2 2
1
3
2.22 (5-CH ); HR-MS (ESI-MS) for C H N O . Calcd: 389.1825. Found: m/z
3 19 25 4 5
+
89.1808 (M+H) .
4
.1.3. General procedure for the synthesis of 3-[2-Imino-2-{2-(closo-1,7-
carboranyl)ethylamino}ethyl]thymidine (3e) and 3-[2-Imino-2-{2-(closo-1,7-
carboranyl)propylamino}ethyl]thymidine (3f). Carboranyl alkyl amine (1 mmol) and 2
(280 mg, 1 mmol) were dissolved in hexafluoroisopropanol (1 mL) in a round bottom
o
flask and heated at 120 C under argon atmosphere overnight to allow the solvent to
evaporate. The reaction mixture was suspended in water (80 mL) and extracted with ethyl
acetate (50 mL × 3). The aqueous layer was concentrated under reduced pressure and the
residue purified by preparative HPLC. The percentage purities of the synthesized
compounds were > 95%.
4
.1.3.1. 3-[2-Imino-2-{2-(closo-1,7-carboranyl)ethylamino}ethyl]thymidine (3e).
1
Retention time = (analytical HPLC) 14.8 min, yield = 32 mg, 7% (white wax); H NMR
(
400 MHz, CD OD, δ ppm): 7.96 (s, 1H, H-6), 6.30 (t, J = 6.6 Hz, 1H, 1'H), 4.84 (s, 2H,
3
N3-CH ), 4.37-4.45 (m, 1H, 3'H), 3.90-3.96 (m, 1H, 4'H), 3.81 (dd, J = 2.9 and 12.0 Hz,
2
1
H, 5''H), 3.75 (dd, J = 3.3 and 12.0 Hz, 1H, 5'H), 3.59 (s, 1H, Carboranyl H), 3.28-3.38
2
1

ACCEPTED MANUSCRIPT
(
m, 2H, CD -OD and NH-CH ), 2.35 (t, J = 7.7 Hz, 2H, Carborane-CH -), 2.20-2.30 (m,
3
2
2
1
3
2
4
1
4
H, 2'H and 2''H), 1.94 (s, 3H, 5-CH ). C NMR (CD OD, 100 MHz, δ ppm): 165.47 (C-
3 3
C=O), 165.05 (HN=C-NH), 152.33 (C-2 C=O), 137.61 (C-6), 110.83 (C-5), 89.18 (C-
'), 87.52 (C-4'), 74.00 (Carboranyl C), 72.29 (C-3'), 62.88 (C-5'), 57.50 (Carboranyl CH),
3.07 (N3-CH ), 42.60 (NH-CH ), 41.42 (C-2'), 34.98 (Carborane-CH -), 13.17 (5-CH );
2
2
2
3
+
HR-MS (ESI-MS) for C H B N O . Calcd: 469.3454. Found: m/z 469.3476 (M+H) .
1
6
33 10
4
5
4
.1.3.2. 3-[2-Imino-2-{3-(closo-1,7-carboranyl)propylamino}ethyl]thymidine (3f).
1
Retention time = (analytical HPLC) 15.2 min, yield = 48 mg, 10% (white wax); H NMR
(
400 MHz, CD OD, δ ppm): 7.98 (s, 1H, H-6), 6.34 (t, J = 6.6 Hz, 1H, 1'H), 4.88 (s, 2H,
3
N3-CH ), 4.44-4.52 (m, 1H, 3'H), 3.94-4.00 (m, 1H, 4'H), 3.84 (dd, J = 3.0 and 12.0 Hz,
2
1
6
H, 5''H), 3.77 (dd, J = 3.4 and 11.9 Hz, 1H, 5'H), 3.55 (s, 1H, Carboranyl H), 3.25 (t, J =
.7 Hz, 2H, NH-CH ), 2.22-2.36 (m, 2H, 2'H and 2''H), 2.04-2.16 (m, 2H, Carborane-
2
1
3
CH -), 1.97 (s, 3H, 5-CH ), 1.66-1.78 (m, 2H, Carborane-CH -CH -). C NMR (CD OD,
2
3
2
2
3
1
00 MHz, δ ppm): 164.32 (C-4 C=O), 163.88 (HN=C-NH), 151.20 (C-2 C=O), 136.44
(C-6), 109.69 (C-5), 88.03 (C-1'), 86.35 (C-4'), 72.26 (Carboranyl C), 71.13 (C-3'), 61.73
(
C-5'), 56.031 (Carboranyl CH), 48.86 (NH-CH -), 41.63 (N3-CH ), 41.45 (NH-CH ),
2
2
2
4
0.31 (C-2'), 33.70 (NH-CH -CH ), 28.07 (Carborane-CH -), 12.05 (5-CH ); HR-MS
2 2 2 3
+
(
ESI-MS) for C H B N O . Calcd: 483.3611. Found: m/z 483.3611 (M+H) .
17 35 10 4 5
4
.1.4. 3',5'-(Di-tert.-butyldimethylsilyl)-3-cyanothymidine (5). 3',5'-(Di-tert.-
butyldimethylsilyl)thymidine (4) [20] (10 g, 21.2 mmol) and cyanogen bromide (3 M
solution in dichloromethane; 14 mL, 42 mmol) were dissolved in acetone (50 mL).
2
2

ACCEPTED MANUSCRIPT
o
Triethylamine (2.75 mL, 20 mmol) was added at 0 C and the reaction mixture was
stirred for 2 h at this temperature [21]. Water (750 mL) was added and the resulting
mixture was extracted with ethyl acetate (250 mL × 4). The organic layer was separated,
dried using anhydrous magnesium sulfate (100 g), and evaporated. The residue was
purified by silica gel column chromatography using hexanes and dichloromethane as
solvent system. The compound was eluted at 50% dichloromethane in hexane. R : 0.63
f
1
(
100% dichloromethane) yield = 8.5 g, 81% (white solid, m.p. 93 °C); H NMR (400
MHz, CDCl , δ ppm): 7.63 (s, 1H, H-6), 6.27 (q, J = 5.7 and 7.8 Hz, 1'H, 1H), 4.35-4.45
3
(
m, 1H, 3'H), 3.96-4.06 (m, 1H, 4'H), 3.90 (dd, J = 2.4 and 11.5 Hz, 1H, 5''H), 3.78 (dd, J
2.2 and 11.5 Hz, 1H, 5'H), 2.32 (ddd, J = 2.4, 5.7 and 13.1 Hz, 1H, 2''H), 1.99-2.10 (m,
H, 2'H), 1.98 (s, 3H, 5-CH ), 0.90 and 0.93 (two s, 18H, ≡C-CH ), 0.08-0.14 (m, 12H,
=
1
3
3
13
Si-CH ). C NMR (CDCl , 100 MHz, δ ppm): 159.82 (C-4 C=O), 147.82 (C-2 C=O),
3
3
1
36.51 (C-6), 109.81 (C-5), 102.75 (C≡N), 88.99 (C-1'), 86.98 (C-3'), 72.73 (C-4'), 63.42
(
(
C-5'), 41.96 (C-2'), 26.32 and 26.11 (two ≡C-CH ), 18.80 and 18.38 (two Si-C≡), 13.32
3
-1
5-CH ), -4.44 and -4.99 (two Si-CH ). IR: (N3-CN) 2264.9 cm . HR-MS (ESI-MS) for
3
3
+
C H N NaO Si . Calcd: 518.2482. Found: m/z 518.2462 (M+Na) .
23
41
3
5
2
4
.1.5. General procedure for the synthesis of 3-(Propylcarbamimidoyl)-3',5'-(di-tert.-
butyldimethylsilyl)thymidine (6a), and 3-(Butylcarbamimidoyl)-3',5'-(di-tert.-
butyldimethylsilyl)thymidine (6b) [22]. Compound 5 (100 mg, 0.21 mmol) was dissolved
in hexafluoroisopropanol (200 µL) at 0 °C. n-Propyl or n-butyl amine (0.4 mmol) was
o
added and the reaction mixture was stirred for 5 h at 0 C. The aqueous layer was
evaporated under reduced pressure and the residue purified by preparative HPLC.
2
3

ACCEPTED MANUSCRIPT
4
0
1
.1.5.1. 3-(Propylcarbamimidoyl)-3',5'-(di-tert.-butyldimethylsilyl)thymidine (6a). Rf:
.40 (100% dichloromethane), Retention time (preparative HPLC) = 46-52 min, yield =
1
11 mg, 95% (white amorphous solid); H NMR (400 MHz, CDCl , δ ppm): 7.47 (s, 1H,
3
H-6), 6.28 (dd, J = 5.7 and 7.9 Hz, 1'H, 1H), 4.40-4.46 (m, 1H, 3'H), 3.91 (q, J = 2.3 Hz,
1
3
2
H, 4'H), 3.84 (dd, J = 11.4 and 2.6 Hz, 1H, 5''H), 3.73 (dd, J = 11.4 and 2.3 Hz, 1H, 5'H),
.16 (t, J = 7.0 Hz, 2H, NH-CH ), 2.23 (ddd, J = 13.1, 5.7 and 2.5 Hz, 1H, 2''H), 1.93-
2
.03 (m, 1H, 2'H), 1.89 (s, 3H, 5-CH ), 1.56-1.69 (m, 2H, NH-CH -CH ), 0.95 (t, J = 7.4
3
2
2
Hz, 2H, propyl-CH ), 0.85 and 0.90 (two s, 18H, ≡C-CH ), 0.08 and 0.04 (two s, 12H ,
3
3
13
Si-CH ). C NMR (CDCl , 100 MHz, δ ppm): 162.82 (C-4 C=O), 150.02 (C-2 C=O),
3
3
1
35.58 (C-6), 111.35 (C-5), 88.87 (C-1'), 86.41 (C-3'), 73.18 (C-4'), 63.88 (C-5'), 46.61
(
C-2'), 42.27 (NH-CH ), 26.82 and 26.60 (≡C-CH ), 23.26 (NH-CH -CH ), 19.29 and
2
3
2
2
1
8.86 (two Si-C≡), 13.83 (5-CH ), 12.49 (propyl-CH ), -3.96 and -4.47 (two Si-CH ).
3 3 3
+
HR-MS (ESI-MS) for C H N O Si . Calcd: 555.3398. Found: m/z 555.3376 (M+H) .
2
6
51
4
5
2
4
.1.5.2. 3-(Butylcarbamimidoyl)-3',5'-(di-tert.-butyldimethylsilyl)thymidine (6b). R : 0.45
f
(100% dichloromethane), Retention time (preparative HPLC) = 62-70 min, yield = 110
1
mg, 95% (white amorphous solid); H NMR (400 MHz, CD OD, δ pp0m): 7.66 (s, 1H,
3
H-6), 6.21 (dd, J = 5.8 and 8.0 Hz, 1'H, 1H), 4.40-4.54 (m, 1H, 3'H), 3.93 (dd, J = 3.1 and
5
2
1
.8 Hz, 1H, 4'H), 3.68-3.90 (m, 2H, 5''H and 5'H), 3.42 (t, J = 7.0 Hz, 2H, NH-CH2),
.10-2.58 (m, 2H, 2''H and 2'H), 1.91 (s, 3H, 5-CH ), 1.60-1.70 (m, 2H, NH-CH -CH -),
3
2
2
.39-1.40 (m, 2H, butyl-CH -CH ), 0.95 (t, J = 7.3 Hz, 2H, butyl-CH ), 0.89 and 0.92
2
3
3
1
3
(
two s, 18H, ≡C-CH ), 0.09 and 0.11 (two s, 12H , Si-CH ). C NMR (CDCl , 100 MHz,
3 3 3
2
4

ACCEPTED MANUSCRIPT
δ ppm): 162.48 (C-4 C=O), 153.79 (=N-C-NH), 149.68 (C-2 C=O), 137.66 (C-6), 111.12
(
C-5), 88.57 (C-1'), 87.17 (C-3'), 73.77 (C-4'), 64.22 (C-5'), 44.21 (NH-CH ), 41.72 (C-
2
2
1
'), 30.45(NH-CH -CH ), 26.52 and 26.29 (≡C-CH ), 20.81 (butyl-CH -CH ), 19.32 and
2 2 3 2 3
8.88 (two Si-C≡), 13.91 (5-CH ), 12.84 (butyl-CH ), -4.59 and -5.18 (two Si-CH ). HR-
3
3
3
+
MS (ESI-MS) for C H N O Si . Calcd: 569.3554. Found: m/z 569.3551 (M+H) and for
2
7
53
4
5
2
+
C H N NaO Si . Calcd: 591.3374. Found: m/z 591.3370 (M+Na) .
27
52
4
5
2
4
.1.6. General procedure for the synthesis of 3-(Phenylcarbamimidoyl)-3',5'-(di-tert.-
butyldimethylsilyl)thymidine (6c), and 3-(Benzylcarbamimidoyl)-3',5'-(di-tert.-
butyldimethylsilyl)thymidine (6d) [22]. Compound 5 (100 mg, 0.21 mmol) was dissolved
in hexafluoroisopropanol (200 µL), aniline or benzyl amine (0.4 mmol) was added, and
the reaction mixture was stirred for overnight at room temperature. The aqueous layer
was evaporated under reduced pressure and the residue purified by preparative HPLC.
4
.1.6.1. 3-(Phenylcarbamimidoyl)-3',5'-(di-tert.-butyldimethylsilyl)thymidine (6c). Rf:
0
.70 (1% methanol in dichloromethane), Retention time (preparative HPLC) = 62-72 min,
1
yield = 110 mg, 93% (white amorphous solid); H NMR (400 MHz, CD OD, δ ppm):
3
7
.61 (s, 1H, H-6), 7.20-7.40 (s, 2H, phenyl H), 6.90-7.20 (br s, 3H, phenyl H), 6.15-6.35
m, 1H, 1'H), 4.48 (m, 1H, 3'H), 3.93-3.4.00 (m, 1H, 4'H), 3.90 (dd, J = 2.5 and 11.5 Hz,
H, 5''H), 3.83 (dd, J = 1.8 and 11.5 Hz, 1H, 5'H), 2.10-2.30 (m, 2H, 2'H and 2''H), 1.92
s, 3H, 5-CH ), 0.96 and 0.95 (two s, 18H, ≡C-CH ) 0.15 and 0.16 (two s, 12H, Si-CH ).
(
1
(
3
3
3
1
3
C NMR (CD OD, 100 MHz, δ ppm): 163.69 (C-4 C=O), 150.69 (=N-C-NH), 147.64
3
(C-2 C=O), 136.55 (C-6), 130.30, 125.08, 123.14 (Aromatic-C), 111.16 (C-5), 89.36 (C-
2
5

ACCEPTED MANUSCRIPT
1
1
5
'), 86.94 (C-4'), 73.65 (C-3'), 64.15 (C-5'), 41.96 (C-2'), 26.55 and 26.35 (two ≡C-CH3),
9.33 and 18.91 (two Si-C≡), 13.15 (5-CH ), -4.40, -4.54, -5.14 (d, J = 9.8 Hz, Si-CH ), -
3
3
.18 (d, J = 3.0 Hz, Si-CH ); HR-MS (ESI-MS) for C H N O Si . Calcd: 589.3241.
3
29 49
4
5
2
+
Found: m/z 589.3251 (M+H) .
4
.1.6.2. 3-(Benzylcarbamimidoyl)-3',5'-(di-tert.-butyldimethylsilyl)thymidine (6d). Rf:
0
.75 (1% methanol in dichloromethane), Retention time (preparative HPLC) = 47-52 min,
1
yield = 120 mg, >99% (white amorphous solid); H NMR (400 MHz, CD OD, δ ppm):
3
7
1
1
2
.74 (d, J = 1.2 Hz, 1H, H-6), 7.35-7.53 (m, 5H, phenyl H), 6.29 (dd, J = 5.9 and 7.8 Hz,
H, 1'H), 4.73 (s, 2H, NH-CH -), 4.49-4.53 (m, 1H, 3'H), 4.00 (dd, J = 3.2 and 5.8 Hz,
2
H, 4'H), 3.92 (dd, J = 3.3 and 11.5 Hz, 1H, 5''H), 3.85 (dd, J = 3.2 and 11.5 Hz, 1H, 5'H),
.29 (ddd, J = 2.7, 5.9 and 13.3 Hz, 1H, 2'H), 2.16-2.24 (m, 1H, 2''H), 1.84 (s, 3H, 5-
13
CH ), 0.96 and 0.98 (two s, 18H, ≡C-CH ), 0.15 and 0.17 (two s, 12H, Si-CH ). C
3
3
3
NMR (CD OD, 100 MHz, δ ppm): 162.58 (C-4 C=O), 154.19 (=N-C-NH), 149.78 (C-2
3
C=O), 137.80 (C-6), 134.77, 130.15, 129.64, 128.86 (Aromatic-C), 111.19 (C-5), 89.67
(
C-1'), 87.26 (C-4'), 73.86 (C-3'), 64.28 (C-5'), 47.98 (NH-CH -), 41.78 (C-2'), 26.51 and
2
2
6.28 (two ≡C-CH ), 19.34 and 18.90 (two Si-C≡), 12.84 (5-CH ), -4.56 (d, J = 9.8 Hz,
3 3
Si-CH ), -5.18 (d, J = 3.0 Hz, Si-CH ); HR-MS (ESI-MS) for C H N O Si . Calcd:
3
3
30 51
4
5
2
+
6
03.3398. Found: m/z 603.3331 (M+H) and for C H N NaO Si . Calcd: 625.3217.
30 50 4 5 2
+
Found: m/z 625.3253 (M+Na) .
4
.1.7. General procedure for the synthesis of 3-[2-(closo-1,7-
Carboranyl)ethylcarbamimidoyl]-3',5'-(di-tert.-butyldimethylsilyl)thymidine (6e), 3-[3-
2
6

ACCEPTED MANUSCRIPT
(closo-1,7-Carboranyl)propylcarbamimidoyl]-3',5'-(di-tert.-butyldimethylsilyl)thymidine
(6f), and 3-[4-(closo-1,7-Carboranyl)butylcarbamimidoyl]-3',5'-(di-tert.-
butyldimethylsilyl)thymidine (6g) [22]. Compound 11 (100 mg, 0.21 mmol) was
dissolved in hexafluoroisopropanol (200 µL) and closo-1,7-carboranyl)alkyl amine (0.4
mmol) was added. The reaction mixture was stirred overnight at 90 °C. The aqueous
layer was evaporated under reduced pressure and the residue purified by preparative
HPLC.
4
.1.7.1. 3-[2-(closo-1,7-Carboranyl)ethylcarbamimidoyl]-3',5'-(di-tert.-
butyldimethylsilyl)thymidine (6e). R : 0.85 (1% methanol in dichloromethane), Retention
f
1
time (preparative HPLC) = 42-47 min, yield = 44 mg, >32% (white amorphous solid); H
NMR (400 MHz, CD OD, δ ppm): 7.70 (s, 1H, H-6), 6.24 (t, J = 6.6 Hz, 1'H, 1H), 4.45-
3
4
.55 (m, 1H, 3'H), 3.93-4.00 (m, 1H, 4'H), 3.89 (dd, J = 2.8 and 11.5 Hz, 1H, 5''H), 3.83
dd, J = 2.5 and 11.5 Hz, 1H, 5'H), 3.60 (Carbranyl-H), 3.50 (t, J = 7.6 Hz, 2H, NH-CH2),
.40 (t, J = 7.6 Hz, 2H, Carborane-CH ), 2.11-2.30 (m, 2H, 2'H and 2''H), 1.94 (s, 3H, 5-
(
2
2
13
CH ), 0.93 and 0.96 (two s, 18H, ≡C-CH ), 0.13 and 0.15 (two s, 12H, Si-CH ). C
3
3
3
NMR (CD OD, 100 MHz, δ ppm): 162.59 (C-4 C=O), 153.93 (=N-C-NH), 149.79 (C-2
3
C=O), 137.86 (C-6), 111.28 (C-5), 89.73 (C-1'), 87.34 (C-4'), 73.89 (Carboranyl-C/C-3'),
6
4.36 (C-5'), 57.54 (Carboranyl-CH), 43.96 (NH-CH ), 41.85 (C-2'), 35.11 (Carborane-
2
CH ), 26.64 and 26.41 (two ≡C-CH ), 19.45 and 19.01 (two Si-C≡), 12.95 (5-CH ), -4.46
2
3
3
and -5.04 (two Si-CH ); HR-MS (ESI-MS) for C H B N O Si . Calcd: 683.5027.
3
27 59 10
4
5
2
+
Found: m/z 683.5076 (M+H) .
2
7