Synthesis and study of cyclic pronucleotides of 5-fluoro-2′- deoxyuridine

Article abstract of DOI:10.1016/j.bmcl.2012.06.011

A one-step method for the synthesis of cyclic pronucleotide (cProTide) derivatives of 5-fluoro-2′-deoxyuridine (FdUrd), utilizing a novel phosphoramidating reagent, is described. Stereochemistry at phosphorus was established by NMR studies and modeling. C

Full text of DOI:10.1016/j.bmcl.2012.06.011

Bioorganic & Medicinal Chemistry Letters 22 (2012) 4497–4501
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and study of cyclic pronucleotides of 5-fluoro-20-deoxyuridine
⇑
⇑
Harsh V. Jain , Thomas I. Kalman
Department of Chemistry, University at Buffalo, Buffalo, NY 14260, United States
a r t i c l e i n f o
a b s t r a c t
A one-step method for the synthesis of cyclic pronucleotide (cProTide) derivatives of 5-fluoro-20-deoxy-
uridine (FdUrd), utilizing a novel phosphoramidating reagent, is described. Stereochemistry at phospho-
rus was established by NMR studies and modeling. Cytotoxicity data of representative cProTide
derivatives of FdUrd are presented. The observed cell-to-cell variations in activity suggests that it is fea-
sible to screen for structural variations in the cProTide moiety favoring metabolic activation in cancer
cells, which may lead to an increase in the therapeutic effectiveness of FdUrd. The method described is
applicable to all anticancer and antiviral nucleoside analogs having both the 50- and the 30-OH groups
available for modification, forming cProTide derivatives capable of delivering the 50-monophosphates
to cells.
Article history:
Received 6 May 2012
Accepted 4 June 2012
Available online 9 June 2012
Keywords:
Prodrug
Nucleoside
Anticancer
Pronucleotide
Ó 2012 Elsevier Ltd. All rights reserved.
Most anticancer and antiviral nucleoside analogs require meta-
bolic activation in the target cells by initial phosphorylation to the
50-monophosphates by nucleoside kinases.1,2 Pre-formed 50-mono-
phosphates cannot be used as drugs, because their negative charge
impedes crossing of the cell membrane. In addition, they are
susceptible to hydrolytic dephosphorylation outside the cells to
the parent nucleosides by enzymes, such as phosphatases and 50-
nucleotidases. Inefficient metabolic activation by nucleoside ki-
nases can be the cause of the lack of therapeutic effectiveness of
many nucleoside analogs, and represents an important mechanism
of drug resistance. Bioreversible nucleoside monophosphate pro-
drugs (ProTides), which can penetrate the cell membrane, can
overcome these limitations. To this end, a wide variety of ProTide
approaches have been developed, for example, aryloxy phosphor-
amidate diesters, bisPOM, cycloSAL derivatives, SATE, and oth-
ers.2–8
sed in cancer cells.11 Since FdUrd has both 50- and 30-OH groups
available for prodrug modification, it permits the design and devel-
opment of a variety of 30,50-cyclic pronucleotide derivatives (cPro-
Tides) for the intracellular delivery of the corresponding 50-
monophosphates, providing an alternative to acyclic ProTides,
which upon activation release toxic aromatic alcohols, like phenol
or naphtol.10 Cyclic phosphate12 and cyclic phosphoramidate13
prodrugs of the anti-HCV 20-C-methyl-cytidine were recently re-
ported, the synthesis of the latter requiring several steps with
low overall yields.13 In this Letter we describe the synthesis of rep-
resentative 30,50-cyclic phosphor-amidate derivatives of FdUrd,
using a simple method applicable for all nucleoside analogs with
free 30- and 50-OH groups.
The synthesis of methoxyalanyl and methoxyglycinyl cPro-
Tides of FdUrd was accomplished in a single step by reacting
the unprotected nucleoside with the easily accessible reagents
5-Fluoro-20-deoxyuridine (FdUrd) is a potent anticancer drug
with serious toxic side effects, which restrict its therapeutic utility.
FdUrd requires intracellular phosphorylation to its 50-monophos-
phate, FdUMP, in order to exert its effect. A variety of FdUMP
ProTides have been reported, including acyclic amino acid phos-
phoramidates of FdUrd.9,10 In addition to overcoming drug resis-
tance due to inefficient phosphorylation, ProTide modifications
may also increase tumor selectivity of FdUrd, if the enzymes in-
volved in the metabolic activation of the prodrugs are overexpres-
bis(4-nitrophenyl)-methoxy-L-alaninyl and bis(4-nitrophenyl)-
methoxyglycinyl phosphoramidate, respectively. These phosphor-
amidate reagents, which are stable at room temperature, were
synthesized from phosphorus oxychloride in two steps with over-
all 80–85% yield after purification (Scheme 1). 1,8-Diazabicy-
clo[5.4.0]undec-7-ene (DBU) was chosen as the base, based on
the work of Dabkowski et al. who used DBU for activating their
P(III) reagents containing p-nitrophenol as leaving group for the
phosphitylation of the 50-hydroxyl of nucleosides.14 FdUrd was
reacted for 2 h at room temperature with the appropriate phos-
phoramidating reagent in the presence of DBU as the activator
base yielding 30–34% of product after column chromatography
(Scheme 2).
⇑
Corresponding authors at present address: Laboratory of Chemistry, CDER/OPS/
OBP/DTP, U.S. Food and Drug Administration, 8800 Rockville Pike, Building 29A,
Room 3B08, Bethesda, MD 20892, United States. Tel.: +1 530 378 4362 (H.V.J.); tel.:
+1 716 688 2800 (T.I.K.).
Using DBU as activator, increasing time and temperature did
not lead to improved yields. Various other bases (Et3N, DMAP,
tBuOK) and other solvents (ACN, pyridine, DMF) were tried, but
E-mail addresses: hvjain@buffalo.edu (H.V. Jain), tkalman@buffalo.edu (T.I.
Kalman).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.06.011

4498
H. V. Jain, T. I. Kalman / Bioorg. Med. Chem. Lett. 22 (2012) 4497–4501
O2N
OH
R1
NH
O
O
O2N
NO2
O
P
a
b
R2
O
O
O
P
O
O
1
NO2
Cl
O2N
Scheme 1. Outline of the synthesis of bis(p-nitrophenyl) amino acid alkylester phosphoramidate. (a) POCl3 (0.5 equiv), Et3N (1 equiv) Et2O, 0 °C to rt, 10 h; (b) amino acid
alkylester hydrochloride (0.5 equiv), Et3N (1 equiv), 0 °C to rt, 2 h.
O
F
O
N
HN
O2N
F
HN
O
N
R1
O
O
P
O
O
DBU (2.4 eq.),
O
O
NH O R2
+
O
P
O
ACN/DCM,
O
O
HO
0oC to rt, 2 h
HN
O2N
HO
1.0 eq
R1
O
1.2 eq
R2
O
1a R1 = CH3; R2 = CH3
1b R1 = H; R2 = CH3
2a R1 = CH3; R2 = CH3
2b R1 = H;
R2 = CH3
Scheme 2. Outline of the synthesis of cProTides.
either complicated reaction mixture was obtained or no desired
product could be isolated.
In Figure 4, irradiation of the NH of the alanine moiety of SE
lead to the appearance of peaks for 30-H and 50-H in the NOESY1D
spectrum. In contrast, FE lacked peaks of 30-H and 50-H protons
(Fig. S4). Therefore, it was concluded that SE is the CC (and/or
CT) diastereomer and FE is the TC diastereomer. Hence, configura-
tion at phosphorus in SE is Rp and FE is Sp respectively. These re-
sults are in agreement with the recently published study by
Meppen et al. where they observed similar NOE peaks between
protons of their cyclic phosphoramidate, corresponding to CC.13
All pronucleotides of FdUrd prepared in this study showed
modest activity in A2780 ovarian cancer cell line and mouse fibro-
blast LM/TKꢀ cell line lacking thymidine kinase (Table 2).
It was of interest to compare the cytotoxicities of the cProTides
with a corresponding acyclic 50-phosphoramidate pronucleotide of
FdUrd. For this reason, the recently reported15 50-phosphoramidate
prodrug 3 (see below), was synthesized and evaluated for cytotox-
icity in the same cell lines. As shown in Table 2, relative to 3 in the
A2780 cell line, 2b was four-fold more active, whereas both 2a dia-
stereomers showed activities similar to 3. In the LM/TKꢀ cell line,
2a-cis was found to be 3.6-fold more active than 3, whereas 2a-
trans and 2b were less active.
Phosphoramidation of FdUrd leads to the formation of a mix-
ture of diastereomeric products, which were separated by silica
gel column chromatography into fast eluting (FE) and slow eluting
(SE) fractions. Using 31P NMR, the ratio of the diastereomers was
calculated to be 5:1 (FE:SE), a magnitude of stereoselectivity that
is intriguing. The ratio is 1:1 in the case of acyclic phosphorami-
dates,15 such as 3 (see structure below). Stereochemistry can play
important role in the enzyme-mediated metabolic activation of the
cProTides. The chirality at phosphorus is established when pseudo-
rotation16 places the p-NO2-phenolate leaving group from an equa-
torial to apical position of the pentacovalent trigonal bipyramidal
intermediate (see Fig. 1). It is conceivable that the energy barrier
of rotation maybe higher for the formation of the SE diastereomer,
but it is unclear why this should be so.17
In the absence of defractable crystals, computational and NMR
studies were undertaken to establish the chirality of the two iso-
mers. The energies of the diastereomers of 2a and their conformers
trans-chair (TC), cis-chair (CC), trans-twist (TT) and cis-twist (CT)
(see Fig. 2) were calculated by Hartree-Fock and B3LYP ab initio
methods. The results are shown in Table 1. The relatively high
D
E-value of 24.9 kcal/mol obtained for TT indicated that it is an
O
energetically unfavorable form to accommodate by the strained
bicyclic fused system (Table 1).
F
HN
In order to investigate the stereochemistry at phosphorus, a
NOESY1D NMR study was conducted. In the minimized struc-
tures TC, CC and CT, the distances between the hydrogens of
the NH of alanine and the 30-H and 50-H of the sugar were cal-
culated. Figure 3 shows that in TC the NH of alanine and 30-H of
the nucleoside is 4.5 Å apart, whereas in CC the distance be-
tween the NH of alanine, 30-H and 50-H on nucleoside is 2.2
and 3.4 Å, respectively. The possibility was considered that CC
and CT may coexist in the solution as a mixed population due
to the small energy difference (1.7–3.6 kcal/mol, Table 1) which
is less than the energy required for the inversion of trimethyla-
mine (approx. 7 kcal/mol).18
O
N
O
O
H
N
O
P
O
O
O
HO
CH3
3
Cell-to-cell variations in activity of different amino acid derivatives
is not unique to cProTides. Recent studies of McGuigan et al.10

H. V. Jain, T. I. Kalman / Bioorg. Med. Chem. Lett. 22 (2012) 4497–4501
4499
FU
O
FU
O
O
O
Me
Me
O
O
O
P
O
P
FU
Me
N
H
COOMe
MeOOC
N
H
O
O
SE
O2N
1x
NO2
O
P
O
O
O
FU
N
H
COOMe
FU
O
O
O
5x
Me
O
P
O
Me
O
O
MeOOC
N
H
P
O
MeOOC
N
H
O
FE
NO2
Figure 1. Pseudorotation of pentacovalent trigonal bipyramidal intermediates leading to the formation of FE and SE in a 5:1 ratio. Geometry around the phosphorus in the
trigonal bipyramids is arbitrary.
Figure 3. Energy minimized conformations of FE (TC) and SE (CC and CT)
Figure 2. Possible diastereomers and preferred conformations of 2a. Protons have
diastereomers. Distances between protons of the amino acid NH and the 30-H and
been omitted for the purpose of clarity. In this image trans-isomers are with the
50-H of the sugar are indicated.
amino acid side chain below, and the pyrimidine above the plane of the sugar ring;
in the cis-isomers both are above the plane.
showed that varying the amino acid and aromatic components of
acyclic 50-phosphoramidates of FdUrd greatly affects their cytotox-
Table 1
icty. They found that the simple replacement of the methyl group of
Energies of diastereomers and their conformers calculated by ab initio methods
3 with cyclohexyl resulted in an approx. 300-fold increase in cyto-
Entry HF/6-311+G(d,p)
D
Ea (kcal/
mol)
B3LYP/6-
311+G(d,p)
D
Ea (kcal/
mol)
toxicty in L1210/TKꢀ cells in culture, but only approx. 30-fold in-
crease in CEM/TKꢀ cells, and a slight decrease in HeLa/TKꢀ cells.
Thus, changing the various substituents on the cyclic phosphoram-
idate moiety may increase the potency and tumor selectivity of the
FdUrd cProTides described in this study increasing the therapeutic
index and overall effectiveness of FdUrd.
TT
CT
TC
CC
ꢀ1705.04157655 29.9
ꢀ1713.319054
24.9
2.9
0.0
ꢀ1705.08383224
ꢀ1705.08925091
ꢀ1705.08647668
3.4
0.0
1.7
ꢀ1713.35410151
ꢀ1713.35873921
ꢀ1713.35748406
0.8
a
D
E is energy difference obtained from HF and B3LYP calculations, where energy
A hypothetical pathway for the intracellular formation of
FdUMP, an inhibitor of the thymidylate synthase, the active form
of TC was kept as reference.

4500
H. V. Jain, T. I. Kalman / Bioorg. Med. Chem. Lett. 22 (2012) 4497–4501
Figure 4. NOESY1D overlaid with 1D NMR spectra of SE diastereomer. The 30-H and 50-H protons of the sugar (red arrows) appeared after irradiation of alanine NH (black
arrow).
Table 2
FU
FU
FU
Cytotoxicity studies of FdUrd and its pronucleotides
Carboxyl
esterase
O
O
O
a
Entry
Compound
IC50 (lM)
5'
3'
H2O
A2780
1.4 1.2
LM/TKꢀ
O
O
O
O
O
O
1
2
3
4
5
2a-trans(FE)
2a-cis (SE)
2b
3
60.7 8.0
12.7 2.3
141 31
45.6 2.2
4.5 1.1
P
P
O
P
O
O
O
1.4 0.7
0.55 0.12
2.4 1.0
O
O
HN
HN
HN
O
O
O
Phospho-
diesterase
H2O
R
R
?
FdUrd
0.026 0.008
R
2a R = CH3
2b R = H
?
H
a
FU
FU
FU
Results are mean standard deviation of at least 3 independent experiments
Phosphor-
amidase
performed in triplicate.
O
O
O
H2O
O
O
?
H2O
P
O
O
HO
O
O
O
OH
O
OH
of FdUrd cProTides, is outlined in Figure 5, involving three sequen-
tial enzyme catalyzed reactions: carboxylesterase(s), phosphodies-
terase(s) and phosphoramidase(s). Stereoselectivity, if present, like
in the case of LM/TK- cells (see Table 2), may reside in the carboxy-
lesterase or phophodiesterase step. Phospho-diesterase cleavage of
the high energy P-30O bond19 is followed by the phosphoramidase
step. An alternative non-enzymatic cleavage of the P-30O bond is
also possible. This activation mechanism is different from that of
the acyclic phosphoramidate pronucleotides, such as 3: it may in-
volve a phosphodiesterase step and it does not require the initial
release of a potentially toxic aromatic residue, such as phenol or
naphtol.10
FdUMP
P
P
O
O
N
H
O
O
HN
O
R
R
Figure 5. Hypothetical mechanism of the intracellular activation of cProTides 2a
and 2b to the active form FdUMP.
form, FdUMP, in tumor tissues, and at the same time minimizing
the relatively higher toxicity to normal tissues, due to the differen-
tial expression of TP.
In conclusion, this Letter describes a simple method for the syn-
thesis of cyclic phophoramidate pronucleotides (cProTides) in one
step under mild conditions. Results from cytotoxicity studies indi-
cated that cProTides with appropriate structural features may be
used to deliver nucleoside monophosphates into cells, providing
an alternative approach to the more conventional acyclic pronucle-
otides, in cases where both the 30- and 50-OH groups of the nucle-
oside are available for modification. It would be of interest to study
the stereochemical requirements and the detailed mechanisms of
the intracellular metabolic activation of cProTides. The produg
technology described in this communication is adaptable to paral-
lel synthesis of a large number of cProTides of dFUrd, as well as
other nucleoside analogs. A variety of amino acid esters differing
in both the amino acid and the ester function can be activated by
It should be noted that the fluoropyrimide derivative most
widely used in cancer chemotherapy today is capecitabine, an or-
ally bioavailable prodrug of 5-fluorouracil (FU).20 Its metabolic
activation to form FU is carried out by three consecutive reactions
catalyzed by carboxylesterase(s), cytidine deaminase and thymi-
dine phosphorylase (TP), respectively, which must be followed by
metabolic activation of FU to FdUMP carried out by multiple en-
zymes. TP is frequently overexpressed in tumor tissues, which
may endow capecitabine with a degree of tumor selectivity.20 In
the case of FdUrd, the high levels of TP in tumor tissues is a disad-
vantage. FdUrd serves as a good substrate for TP, which breaks it
down to FU, reducing its availability to be phosphorylated by thy-
midine kinase. The cProTide approach described in this report pre-
vents such degradation, enhancing the formation of the active

H. V. Jain, T. I. Kalman / Bioorg. Med. Chem. Lett. 22 (2012) 4497–4501
4501
7. Drontle, D. P.; Wagner, C. R. Mini-Rev. Med. Chem. 2004, 4, 409.
8. Cahard, D.; McGuigan, C.; Balzarini, J. Mini-Rev. Med. Chem. 2004, 4, 371.
9. Abraham, T. W.; Kalman, T. I.; McIntee, E. J.; Wagner, C. R. J. Med. Chem. 1996,
39, 4569.
10. Mc Guigan, C.; Murziani, P.; Slusarczyk, M.; Gonczy, B.; Vande Voorde, J.;
Liekens, S.; Balzarini, J. J. Med. Chem. 2011, 54, 7247.
11. Dubowchik, G. M.; Walker, M. A. Pharmacol. Ther. 1999, 83, 67.
12. Gunic, E.; Girardet, J.-L.; Ramasamy, K.; Stojsavljevic-Petkov, V.; Chow, S.; Yeh,
L.-T.; Hamatake, R. K.; Raney, A.; Hong, Z. Bioorg. Med. Chem. Lett. 2007, 17,
2452.
the simple method described here to furnish the corresponding re-
agents 1 (see Scheme 1). Subsequent reaction with the unprotected
nucleosides can yield a diverse library of cProTides, which can be
subjected to high throughput screening in cellular systems
in vitro for the desired biological activity.
Supplementary data
13. Meppen, M.; Pacini, B.; Bazzo, R.; Koch, U.; Leone, J. F.; Koeplinger, K. A.;
Rowley, M.; Altamura, S.; Di Marco, A.; Fiore, F.; Giuliano, C.; Gonzalez-Paz, O.;
Laufer, R.; Pucci, V.; Narjes, F.; Gardelli, C. Eur. J. Med. Chem. 2009, 44, 3765.
14. Dabkowski, W.; Ozarek, A.; Tworowska, I. New J. Chem. 2005, 29, 1396.
15. Congiatu, C.; Brancale, A.; Mason, M. D.; Jiang, W. G.; McGuigan, C. J. Med.
Chem. 2006, 49, 452.
Supplementary data (synthetic details, analytical data and pro-
cedures for biological and computational studies) associated with
this article can be found, in the online version, at http://
dx.doi.org/10.1016/j.bmcl.2012.06.011.
16. Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70.
[17]. The strain caused by the fused bicyclic ring system of 2a may be a factor, in
light of previous observations regarding the effects of strain on the
pseudorotation of cyclic alkyloxyphophoranes (Gorenstein, D.; Westheimer,
F. H. J. Amer. Chem. Soc. 1970, 92, 634).
18. Koeppl, G. W.; Sagatys, D. S.; Krishnamurthy, G. S.; Miller, S. I. J. Amer. Chem.
Soc. 1967, 89, 3396.
19. Greengard, P.; Rudolph, S. A.; Sturtevant, J. M. J. Biol. Chem. 1969, 244, 4798.
20. Samid, D. In Fluoropyrimidines in Cancer Therapy; Rustum, Y. M., Ed.; Humana:
Totowa, 2003; p 261.
References and notes
1. Advances in Antiviral Drug Design; Hostetler, K. Y., Declercq, E., Eds.; Elsevier:
New York, 2007; Vol. 5, p 167.
2. Ariza, M. E. Drug Design Rev. - Online 2005, 2, 373.
3. Peterson, L. W.; McKenna, C. E. Expert Opin. Drug Deliv. 2009, 6, 405.
4. Mehellou, Y.; Balzarini, J.; McGuigan, C. Chem. Med. Chem. 2009, 4, 1779.
5. Hecker, S. J.; Erion, M. D. J. Med. Chem. 2008, 51, 2328.
6. Poijarvi-Virta, P. Curr. Med. Chem. 2006, 13, 3441.