Medium benzene-fused oxacycles with the 5-fluorouracil moiety: Synthesis, antiproliferative activities and apoptosis induction in breast cancer cells

Article abstract of DOI:10.1016/S0040-4020(03)00871-8

On the basis of the increase in lipophilicity, novel benzannelated six- and seven-membered derivatives have been synthesized, starting from 2-hydroxybenzyl alcohols, 2-hydroxybenzaldehydes, and catechol. The X-ray structure of (RS)-1-(2,3-dihydro-5H-1,4-benzodioxepin-3-yl)-5-fluorouracil (5) is presented and compared with its conformational analysis at the semiempirical (AM1) and ab initio (6-31G^*) levels and NOE effects. A good agreement between both experimental and theoretical data was found showing a chair conformation for the 2,3-dihydro-5H-1,4-dioxepin ring and an axial orientation of the 5-FU moiety on the C3 position. Compounds 5 and (RS)-1-(7-methoxy-2,3-dihydro-5H-1,4-benzodioxepin-3-yl)-5-fluorouracil (6) were found to be the most potent inhibitor of MCF-7 cells growth. (RS)-1-(2,3-Dihydrobenzoxepin-2-yl)-5-fluorouracil (8) induced apoptosis up to 57.33% of cell population after 24 h.

Full text of DOI:10.1016/S0040-4020(03)00871-8

Tetrahedron 59 (2003) 5457–5467
Medium benzene-fused oxacycles with the 5-fluorouracil moiety:
synthesis, antiproliferative activities and apoptosis induction in
breast cancer cells
a
Estrella Saniger, Joaquın M. Campos, Antonio Entrena, Juan A. Marchal, Ines Suarez,
a
a
b
c
´
Antonia Aranega, Duane Choquesillo, Juan Niclos, Miguel A. Gallo and Antonio Espinosa
´
´
c
d
d
a
a
,
´
*
´
´
´
a
´
´
´
Departamento de Quımica Farmaceutica y Organica, Facultad de Farmacia, c/ Campus de Cartuja s/n, 18071 Granada, Spain
´
Departamento de Ciencias de la Salud, Facultad de Ciencias Experimentales y de la Salud, Paraje de las Lagunillas s/n, 23071 Jaen, Spain
b
c
´
Departamento de Ciencias Morfologicas, Facultad de Medicina, Avenida de Madrid s/n, 18071 Granada, Spain
Departamento de Quımica Inorganica, Facultad de Farmacia, c/ Campus de Cartuja s/n, 18071 Granada, Spain
d
´
´
Received 25 March 2003; revised 15 May 2003; accepted 28 May 2003
Abstract—On the basis of the increase in lipophilicity, novel benzannelated six- and seven-membered derivatives have been synthesized,
starting from 2-hydroxybenzyl alcohols, 2-hydroxybenzaldehydes, and catechol. The X-ray structure of (RS)-1-(2,3-dihydro-5H-1,4-
benzodioxepin-3-yl)-5-fluorouracil (5) is presented and compared with its conformational analysis at the semiempirical (AM1) and ab initio
(6-31Gp) levels and NOE effects. A good agreement between both experimental and theoretical data was found showing a chair conformation
for the 2,3-dihydro-5H-1,4-dioxepin ring and an axial orientation of the 5-FU moiety on the C3 position. Compounds 5 and (RS)-1-(7-
methoxy-2,3-dihydro-5H-1,4-benzodioxepin-3-yl)-5-fluorouracil (6) were found to be the most potent inhibitor of MCF-7 cells growth. (RS)-
1-(2,3-Dihydrobenzoxepin-2-yl)-5-fluorouracil (8) induced apoptosis up to 57.33% of cell population after 24 h. q 2003 Elsevier Science
Ltd. All rights reserved.
1. Introduction
effects than 5-FU, have been diligently sought in a number
of laboratories. seco-Nucleosides are non-classical nucleo-
sides in which the ‘sugar’ is linear instead of cyclic as is
usual. Among them, the isopropoxy 2 and cyclopentoxy 3
analogues were reported to exhibit antitumor activity
against the HEp human cells.4 The most lipophilic structure
proved to be the most active compound of all those tested,
being 4-fold more active than 5-FU. Afterwards, we
embarked on a program to synthesize a wide range of
5-FU derivatives linked to saturated annelated heptatomic
moieties.5,6 Nevertheless, since systematic study of struc-
ture–activity relationship in seven-membered 5-FU O,N-
acetals has not been realized so far, it is thought that much
effort should be made in this class of derivatives to search
for new anticancer agents. Compound 4 has been previously
described by us.7 On one hand, with the idea of increasing
the lipophilicity of 4, in this paper, we propose a series of
bioisosteric benzannelated seven-membered O,N-acetals
such as 5–8. On the other, the preparation of a six-
membered benzannelated derivative such as 9 could
emphasize the significance of the biological activity of the
seven-membered 5-FU O,N-acetals. In all the cases, the
attachment of the 5-FU moiety occurs at the N 1-position of
the pyrimidine ring. To our knowledge, up to now, no
attempt has been made to synthesize benzannelated 5-FU
O,N-acetals. Our aim was to fill this gap and in this paper we
report the synthesis and antitumor activity of the title
Breast cancer is a common and often fatal disease.
Excluding cancers of the skin, that of the breast is the
most common cancer among women, accounting for nearly
one out of every three cancers diagnosed in American
women. Each year, over 186,000 new cases and 46,000
deaths are reported in the United States alone.1
As part of their action on neoplastic cells, many anticancer
drugs activate apoptosis (programmed cell death). Apop-
tosis may be a primary mechanism of antineoplastic agents.2
Although breast cancer is most often treated with conven-
tional cytotoxic agents it has proven difficult to induce
apoptosis in breast cancer cells using these drugs.3
Improved clinical response may be obtained by identifying
therapies that are particularly effective in activating
apoptosis and determining how those therapies may be
modified to effect maximum apoptosis induction.
Novel derivatives of 5-fluorouracil (5-FU, 1) possessing a
broader spectrum of antitumor activity and fewer side
Keywords: acetals; benzodioxepins; X-ray crystal structure; antitumor
compounds.
*
Corresponding author. Tel.: þ34-958-243850; fax: þ34-958-243845;
e-mail: aespinos@ugr.es
0040–4020/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4020(03)00871-8

5458
E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
Chart 1.
compounds, starting from salicyl alcohol or salicyl aldehyde
(Chart 1).
days). All condensations were carried out in a one-pot
reaction under our standard conditions.4–7
The preparation of 8 is shown in Scheme 2. Starting from
acrolein the corresponding non-isolated phosphonium salt
1710 reacted via a Wittig reaction with 2-hydroxybenzalde-
hyde and formed (Z)-18 (15%). The cyclization reaction
using boron trifluoride diethyl etherate in diethyl ether as a
solvent during six days gave 2-methoxy-2,3-dihydro-
benzoxepin 19 (34%). Formation of (Z)-18 must be justified
by means of two factors that have an influence on the
stereoselectivity of the Wittig reaction: the ylide structure
and the reaction conditions. As a rule of thumb the non-
stabilized ylides predominantly give (Z) alquenes.11 In our
case, the reaction conditions have been very mild and 17 is a
non-stabilized ylide that gives only isomer (Z) in a very fast
process. The experimental evidence that 18 is a Z isomer is
2. Results and discussion
2.1. Chemistry
At this stage of our research we have carried out the work
with racemic mixtures. Our synthetic strategy was to
synthesize the seven-membered cycloacetal and then
condense it with the 5-FU base. The synthesis of compounds
5–7 is depicted in Scheme 1. Reduction of 2-hydroxy-5-
methoxybenzaldehyde with sodium borohydride gave the
2-hydroxy-5-methoxybenzyl alcohol8 10 whilst 2-hydroxy-
benzyl and 2-hydroxy-3-methoxybenzyl alcohols were
commercially available. Alkylation of the phenolic group
with bromoacetaldehyde dimethyl acetal using sodium
hydride as a base in dimethylformamide (DMF) gave the
hydroxyacetals 11–13. Compound 11 has been previously
reported9 and we have used the same protocol to synthesize
12 and 13. The cyclization reaction under conditions
previously reported by us6 produced the 3-methoxy-2,3-
dihydro-5H-1,4-benzodioxepins 14–16. Subsequent substi-
tution of the acetalic methoxy groups by the 5-FU moiety in
the presence of the silylating agents, 1,1,1,3,3,3-hexa-
methyldisilazane (HMDS) and trimethylchlorosilane (TCS),
under acid catalysis (SnCl4) in acetonitrile during 24 h
yielded 5 (45%), 7 (20%) with a reaction time of 24 h, and 6
(40%) when the reaction time was increased up to 144 h (6
0
0
the following: for this compound J3 ,4 ¼11.3 Hz and this
value is very close to J2,3 in cis-2-butene (J2,3¼10.9 Hz),12
and very far from the corresponding J2,3 value in trans-2-
butene (J2,3¼15.1 Hz).12 Besides this, in compound 19 the
double bond configuration has to be Z and in this case,
J4,5¼11.9 Hz. The substitution of the methoxy group was
carried out as described previously for 24 h to yield 8 (10%).
The synthesis of 9 followed a rather unexpected route in
which the key step was the formation of 4-methoxy-8-
chromanol 20.
The alkylation of catechol to give the o-hydroxyphenoxy-
propanal dimethyl acetal 21 was carried out in two different
Scheme 1. Reagents: (i) BrCH2CH(OMe)2, HNa, anhydrous DMF; (ii) BF3·Et2O in anhydrous Et2O; (iii) 5-FU, HMDS, TCS, SnCl4/CH2Cl2, CH3CN.
Scheme 2. Reagents: (i) See Ref. 10; (ii) 2-hydroxybenzaldehyde; (iii) BF3·Et2O; (iv) 5-FU, HMDS, TCS, SnCl4/CH2Cl2, CH3CN.

E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
5459
Scheme 3. Reagents: (i) K2CO3, anhydrous acetonel, BrCH2CH2CH(OMe)2; (ii) K2CO3, anhydrous acetone, BrCH2CH2CH(OCH2)2; (iii) anhydrous MeOH,
1% H2SO4; (iv) BF3·Et2O in anhydrous Et2O.
ways, as is depicted in Scheme 3:
acetalic (for compounds 14–16 and 19 the range for acetalic
carbons is d 100–101 ppm).
(a) directly by treatment with potassium carbonate in
anhydrous acetone, using 3-bromopropanal dimethyl
acetal as a reactant (21, 23%),
One interpretation of the unexpected formation of 20 could
be the following (Scheme 4).
(b) through 22 by alkylation with 2-(2-bromoethyl)-1,3-
dioxolane in anhydrous acetone and potassium
carbonate as base (21, 21%) and subsequent
transacetalization reaction with anhydrous methanol
in 1% sulfuric acid (21, 60%).
Boron trifluoride coordinated one of the methoxy groups of
21 to give the oxocarbenium ion 24, which resulted from the
loss of the methoxy group, and which could progress in the
two following manners:
(a) Intramolecular cyclization with the phenolic hydroxyl
group with formation of 23. This process was highly
improbable due to the poor nucleophilicity of the
hydroxy phenolic group on being complexed to the
Lewis acid BF3.
(b) Intramolecular cyclization with the aromatic ring,
whose non-substituted position adjacent to the ethereal
oxygen atom was exceedingly nucleophilic so as to
form the non-acetalic chromane 20.
Finally, the intramolecular cyclization reaction catalyzed by
boron trifluoride diethyl etherate gave 4-methoxy-8-chro-
manol (20, 51%) instead of the expected 2-methoxy-3,4-
dihydro-2H-1,5-benzodioxepin 23. The spectroscopic
characterization of the supposed compound (23) did not
result concordant with the proposed structure in spite of the
fact that the HR LSIMS gave a compatible molecular ion.
The chemical shift of H-4, the presence of three aromatic
protons (H-5, H-6 and H-7), the proton of the phenolic
1
group in H NMR, and also the corresponding 13C NMR
The substitution of the methoxy moiety was carried out by
the reaction of 20 with 5-FU under the above-mentioned
conditions to yield the chromane analogue 9 (44%). The
substitution of the benzylic methoxy group, although non-
acetalic, was possible under the conditions described for the
reaction due to the formation of the stable carbenium ion 25.
In this sense, the reactivity of the benzylic methoxy group of
20 is similar to that of an acetalic one.
signals, were over-riding arguments to reject structure 23 as
the correct one. The most important signal of 20 is the H-4a
(d, 4.33), which appears as a double doublet with coupling
constants of 3.3 and 6.7 Hz with H-3a and H-3b,
respectively. The 1H NMR chemical shift of H-4a confirms
that this proton is not acetalic (for compounds 14–16 and 19
the range for acetalic hydrogen atoms is <d 4.8 ppm). The
proton of the phenoxy group appears at d 5.56 ppm. After a
thorough study of the 1H NMR spectrum, it has been
possible to determine the coupling constants of H-4a with
H-3a and H-3b being 3.3 Hz and 6.7 Hz, respectively. J2a,3b
is 9.1 Hz and J2b,3a and J2b,3b are 6.6 and 3.5 Hz,
respectively. Both H-2a and H-2b resonate at the same
field as a multiplet, which eliminates the geminal constant.
The determination of the other coupling constants has been
possible on the basis of the coupling constants of H-3.
Regarding H-3a and H-3b, the former appears at d 2.16 ppm
as a ddd and the latter at d 2.04 ppm as a dddd with the
following coupling constants: J3a,4a¼3.3 Hz, J3a,2b¼6.6 Hz,
J3a,3b¼14.2 Hz and J3b,2b¼3.5 Hz, J3b,4a¼6.7 Hz,
J3b,2a¼9.1 Hz, J3b,3a¼14.2 Hz. It is worth mentioning that
proton H-3a forms a 908 angle with H-2a and, accordingly,
J3a,2a¼0 Hz. Finally, the 13C NMR chemical shift of C-4 is d
71.23 ppm, which confirms that this carbon atom is not
It is worth mentioning that the signal of H-6 for compound 9
appears as a sharp doublet (J¼6.1 Hz) at d 6.96 ppm, due to
the coupling with the adjacent F atom. This fact, together
with the elemental analysis data, demonstrates
unambiguously that the 5-FU fragment is linked to the
aminalic carbon through N-1 and not through N-3.13
Regarding the 13C NMR spectrum, the C-4, C-5 and C-6
signals (of the pyrimidine base) show couplings with the F
atom of 25.7 for C-4, 229.1 for C-5 and 33.4 Hz for C-6, all
appearing as doublets, whilst C-2 appears as a singlet (d
149.89 ppm).
2.2. Molecular structure and inter-molecular
interactions of 5
Single crystals of 5 were obtained by the slow diffusion of

5460
E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
Scheme 4. Reagents: (i) BF3·Et2O in anhydrous Et2O; (ii) –MeO·BF23 1; (iii) H2O; (iv) 5-FU, HMDS, TCS, SnCl4/CH2Cl2, MeCN.
ether in its methanol solution. The crystal structure reveals
the following interesting facts: (a) the seven-membered ring
of the 2,3-dihydro-1,4-benzodioxepin moiety shows a chair
conformation, and (b) the 5-FU fragment is in an equatorial
orientation (Fig. 1). In the crystal, pairs of molecules related
by the symmetry transformation (1¼2xþ1, 2yþ2, 2zþ2
are linked by two symmetrical and rather linear hydrogen
bonding interactions of the type N(30)–H(30)· · ·O(20)(1
centroid distance dc–c and the inter-planar dp–p distance are
˚
˚
3.77 and 3.53 A or 3.84 and 3.64 A, respectively. These data
are consistent with slipping angles b¼g¼20.72 or 19.028,
respectively. Such ring–ring stacking interactions co-
operate with the referred hydrogen bonds to give a corrugate
2D molecular framework (Fig. 2) where the C6-benzene
ring and the 5-fluoruracil moiety of the existing conformer
define an open dihedral angle of 64.18. These layered
structures interact by Van der Waals forces to build up the
crystal.
˚
(d(N· · ·O) 2.84 A, /(N–H· · ·O) 1738). It is noteworthy
that the O atoms of the dioxepin ring are not involved in this
kind of intermolecular interactions. In addition, pairs of
5-fluorouracil rings (symmetry code (2¼2x, 2y, 2z) and of
benzene rings (symmetry transformation (3¼12x, 12y,
12z) from adjacent molecules build up centro-symmetric
p,p-stacking interactions (parallel rings, with dihedral
inter-planar angles a¼0.008). In these stacks, the centroid–
2.3. Conformational analysis
We have used theoretical methods at semiempirical and ab
initio levels to compare the results with the data obtained by
X-ray crystallography. In view of the biological importance
Figure 1. ORTEP drawing of 5 at 50% probability.

E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
5461
Figure 2. Packing of the molecules of 5 in the unit cell showing their hydrogen bond network and p,p-stacking interactions.
of compounds 5–7 (see below), we have carried out a
detailed conformational analysis of the simplest compound
of the series (5) to gain more insight into the molecular
structure prior to exploring further structure–activity
relationships. The conformational analysis has been carried
out by means of the Tripos force field,14 implemented in the
Sybyl calculation package.15 The initial geometries have
been generated from standard fragments of the program
libraries and their optimization have been carried out by
means of the Powell method,16 up to the point where the
energy gradient is less than 0.01 kcal/mol. The electrostatic
component has been calculated by means of the atomic
charges generated according to the Gasteiger method,17
using a distance-dependent dielectric constant equal to the
unity (1¼1). The conformational analysis has been carried
out in two phases: in the first one, the possible confor-
mations of the base ring 5 have been identified and used as
starting points, and in the second one, the hydrogens of the
C-3 atom of the ring have been substituted by the 5-FU
moiety and the possible rotamers around the exocyclic C–N
bond have been investigated.
the 1,4-dioxepane ring bonds, except the one that is
common to the benzene system. The conformations thus
obtained have been subsequently minimized once the
imposed restrictions to the torsional angles were removed,
and have been compared. The final result has led to only six
conformations: two chair, two boat and two twist-boat
conformations.
The most stable conformation is the chair, followed by the
twist-boat and the boat, the conformations of the same
family being isoenergetic. The relative energy values of the
chair and twist-boat conformations coincide with the
reported values9 for such a compound, calculated from
13C NMR data at a very low temperature (21208C). At such
a temperature, the observed conformational population for
the chair conformation is higher than 98%, other confor-
mations not being detectable. Moreover, an ab initio
calculation at the 6-31Gp level (Gaussian9818) of such
conformations has been undertaken. The same order of
stability has been found, although there exists a slightly
higher energetic difference between the three studied
conformations. Figure 3 shows the conformers and their
described energy values.
2.3.1. Conformational analysis of 2,3-dihydro-5H-1,4-
benzodioxepin. A chair form has been defined as a starting
conformation for the seven-membered ring and, once
optimized, conformational searches have been carried out
by means of the Sybyl gridsearch utility by rotating each of
2.3.2. Conformational analysis of 5. In the second phase,
the six conformations found for the 2,3-dihydro-1,4-
benzodioxepin have been used as starting conformations,
Figure 3. The only three conformations of 1,4-benzodioxepin, together with their relative energy values (kcal/mol) calculated by means of molecular
mechanics (Tripos) and ab initio (Gaussian). The calculated conformational populations at 258C are shown in parentheses. The three remaining conformations
are the inverted ones of these.

5462
E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
substituting the 3-b hydrogen atom for a 5-fluorouracil-1-yl
fragment. Each of the new molecules were minimized as
described previously and, once optimized, a conformational
search by means of the gridsearch Sybyl utility has been
carried out, rotating the exocyclic C–N bond at 108
intervals. Subsequently, the restrictions of the different
geometries obtained in each conformational search were
released and a new minimization of all the conformations
was carried out. After such a minimization, the confor-
mations obtained in all the conformational searches were
compared with each other to identify those which are
different from a geometric and energetic point of view.
phase that agrees perfectly with the one calculated as the
most stable in vacuum.
The 1H NMR studies in solution show the existence of NOE
effects between protons 3a and 5a on one hand, and
between 2b and 60 on the other (Fig. 5). Both NOE effects
are compatible with the calculated distances for the most
˚
stable conformation of this compound: 2.29 and 2.87 A,
respectively. If the twist-boat conformation were more
important than the calculated one, a NOE effect would have
been observed between the 2a and 5a hydrogen atoms,
˚
because they are located at a 2.37 A distance; nevertheless,
such a NOE effect is not experimentally observable, which
confirms that the chair is the most important conformation,
practically the only one for compound 5.
Twenty conformations for the compound 5 were obtained,
relative energies being lower than 10 kcal/mol in all cases.
The 10 most stable conformations have been optimized by
means of ab initio calculations at level 6-31Gp, reducing
them to eight energetically and geometrically distinct
conformations. Figure 4 shows the four most stable
conformations found for this compound.
2.4. Antiproliferative activities and apoptosis induction
The IC50 values20 of compounds are shown in Table 1. The
most active compounds are 5 and 6. Compound 4 shows less
antiproliferative activity (IC50¼23^0.88 mM). The lipo-
philicity in this structure has been increased by means of a
fused benzene ring, and an unsaturation has been introduced
to give 8. An increase has been obtained in its anti-
proliferative activity (IC50¼14^1.02 mM). On comparing
structures 8 and 5, it is worth emphasizing that the
bioisosteric change of carbon for oxygen and the saturation
of the double bond in compound 5 increases the anti-
proliferative activity twice (IC50¼7^0.61 mM). The intro-
duction of a methoxy group into the benzene ring of 5
provokes different influences on the antiproliferative
activities. Thus, the C-7 substitution produces an increase
of the antiproliferative activity (6, IC50¼4.5^0.33 mM),
whilst if C-9 is the substituted position it gives rise to a
The most stable conformation is a chair for the 1,4-
benzodioxepin fragment with the 5-FU group in an
equatorial orientation, while the second one is a twist-boat
with 5-FU in an isoclinal orientation.19 The third one in
importance is a boat with the substituent equatorial and,
finally, the fourth one is a chair with the substituent in an
axial orientation.
If we pay attention to the calculated conformational
populations, the first of the above-mentioned is the only
significant conformation because it represents more than
93% of the equilibrium of such a compound. Actually,
X-ray crystallography shows a conformation in the solid
Figure 4. Representation of the four most stable conformations of compound 5. The relative energy values calculated by means of molecular mechanics
(Tripos) and ab initio (Gaussian) are expressed in kcal/mol. The four remaining conformations are the inverted ones of these, and the calculated conformational
populations at 258C are shown in parentheses considering the eight more stable conformations of this compound.

E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
5463
˚
Figure 5. Calculated distances in A in the chair and twist-chair conformations of 5. The distances of 2.41 and 3.00 A are compatible with the experimentally
˚
˚
observed NOE effects. The 2.57 A distance of the twist-boat conformation is compatible with a NOE effect that is not experimentally observed, and it may thus
be concluded that the chair conformation is the only existing one in solution.
decrease in the antiproliferative activity (7,
IC50¼22^0.93 mM).
programmed cell death up to 43% of cell population.24
Accordingly, the early apoptotic inductions and the low IC50
values give rise to a significant antitumoral activity. Our
recent findings suggest that compounds 5–9 provoke a
G0/G1-phase cell cycle arrest, whereas Ftorafur [1-(2-
tetrahydrofuranyl)-5-fluorouracil], a known prodrug† of
5-FU, induces a S-phase cell cycle arrest and, accordingly,
the title compounds are drugs per se. Also compounds 5–9
administered intravenously twice a week (with a 50 mg/kg
dose each time) induce neither toxicity nor death in mice
after a one-month treatment (results not shown).
Apoptosis has been studied in terms of cancer development
and treatment, with attempts made to identify its role in
chemotherapeutic agent-induced cytotoxicity. Cytotoxic
agents often induce only a fraction of the cells to become
apoptotic. To fully exploit apoptosis as a mechanism of
antineoplastic agent response, a larger proportion of cells
needs to be recruited into apoptosis. Paclitaxel (Taxolw),
cyclophosphamide and cytosine arabinoside are the only
commonly used cytotoxic agents shown to elicit apoptosis
in breast cancer cells.21,22 Quantitation of apoptotic cells
was done by monitoring the binding of fluorescein
isothiocyanate (FITC)-labelled annexin V (a phosphatidyl-
serine-binding protein) to cells in response to our title
compounds as described.23 The apoptosis study shows that
compounds 4, 7, 8 and 9, at the IC50 concentration, provoke
early apoptosis in the cells treated for 24 and 48 h. It is
worth pointing out that compound 7 induces greater
apoptosis at 48 h (46.73%) than at 24 h (40.08%) and so
does compound 4 [48 h (53.92%) and 24 h (46.63%)]. The
compounds that show the most important apoptotic indexes
at 24 h are 8 (57.33%) and 9 (54.33%), whereas at 48 h are 4
(53.92%) and 8 (51.37%). These compounds are more
potent as apoptosis inductors against the MCF-7 human
breast cancer cells than paclitaxel (Taxolw), which induced
3. Conclusion
We have completed the synthesis and biological evaluation
of novel benzannelated medium oxacyclic 5-FU O,N-
acetals. The synthesized compounds exhibited good anti-
proliferative activities against the breast cancer cell line
MCF-7, among which derivative 6 was found to be the most
cytotoxic. Our investigation suggest that 8 may be
particularly useful in stimulating the apoptotic process in
breast cancers. At present, studies are being carried out to
determine the mechanism of action at the molecular level of
the title compounds.
4. Experimental
4.1. Chemistry
Table 1. Antiproliferative activities, and apoptosis induction in the MCF-7
human breast cancer cell line after treatment for 24 and 48 h for the
compounds
Comp.
IC50 (mM)a
% Apoptosisb
Melting points (mp) were taken in open capillaries on an
Electrothermal melting point apparatus and are uncorrected.
Nuclear magnetic resonance (NMR) spectra were recorded
on a 400.1 MHz 1H and 100.03 MHz 13C NMR Bruker
24 h
48 h
Control
1.24
46.63
8.45
1.24
53.92
12.17
3.50
46.73
51.37
35.49
4
5
6
7
8
9
23^0.88
7^0.61
4.5^0.33
22^0.93
14^1.02
69^2.31
1
ARX 400 or 300.13 MHz H and 75.78 MHz 13C NMR
Bruker AMX-300 spectrometers, and chemical shifts (ppm)
are reported relative to the solvent peak (CHCl3 in CDCl3 at
d 7.24 and 77.1 ppm; DMSO in DMSO-d6 at d 2.50 and
39.5 ppm). Signals are designated as follows: s, singlet; bs,
1.50
40.08
57.33
54.33
a
See Ref. 20. The data are means^SEM of three independent
determinations.
Apoptosis was determined using an annexin V-based assay (Ref. 23). The
data indicate the percentage of cells undergoing apoptosis in each sample.
All experiments were conducted in duplicate and gave similar results.
†
Prodrugs have been described as the chemical modification of a
b
biologically active compound to form a new structure, which will
liberate the parent drug upon either in vivo enzymatic or non-enzymatic
attack.

5464
E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
broad singlet; d, doublet; dd, doublet of doublet; ddd, double
doublet of doublet; dddd, double double doublet of doublet;
t, triplet; q, quadruplet; m, multiplet. Coupling constants (J)
are expressed in hertz. The NMR, MS, and IR spectral data
of all compounds were consistent with the assigned
structures. All final products had satisfactory (within
^0.4%) C, H, and N analyses. Liquid secondary ion mass
spectra (LSIMS), electron impact (EI) and high-resolution
mass spectra (HR) were carried out on a VG AutoSpec Q
high-resolution mass spectrometer (Fisons Instruments).
Analytical thin-layer chromatography (TLC) was per-
formed using Merck Kieselgel 60F-254 plates as the solvent
system. Chromatograms were visualized under a UV lamp
(254 nm), by placing the air-dried plates in a tank of I2
vapour, or charring with a 20% sulfuric acid/methanol
solution. Preparative separations were performed by flash
chromatography on silica gel (Merck; 230–400 mesh) using
mixtures of diethyl ether/hexane or CH2Cl2/MeOH as
eluents. Acetonitrile was dried by refluxing and distilling
from calcium hydride and anhydrous methanol was
prepared by the magnesium drying procedure. Dimethyl-
formamide was distilled from phosphorus pentoxide and
54.00 (OMe)2). HR LSIMS calcd for C12H18O5Na (MþNa)þ
265.1051, found: 265.1052. Anal. for C12H18O5: calcd: C
59.49; H 7.49. Found: C 59.53; H 7.45.
4.1.1.2. Synthesis of (RS)-3-methoxy-2,3-dihydro-5H-
1,4-benzodioxepins 14–16. The preparation of 14 was
previously reported9 but we used the Lewis acid BF3·Et2O
in anhydrous diethyl ether, formerly established by us,6 and
we proceeded with the same method to obtain 15 and 16 (as
liquids), starting from the acyclic acetals 11–13, respect-
ively. In the case of 14, we found that the distillation of 11 in
vacuo on a Kugelrohr apparatus (1508C/5 mm Hg)‡ pro-
duced the required cyclization to yield the final cycloacetal
14 in an one-pot process (40%).§
(RS)-3-Methoxy-2,3-dihydro-5H-1,4-benzodioxepin 14.
Yield: 40%.
(RS)-3,7-Dimethoxy-2,3-dihydro-5H-1,4-benzodioxepin 15.
1
Yield: 45%. Rf (2/1, diethyl ether/hexane): 0.37. H NMR
(300 MHz, CDCl3) d 6.91 (d, 1H, J¼8.7 Hz, H-9), 6.70 (dd,
1H, J¼3.0, 8.7 Hz, H-8), 6.63 (d, 1H, J¼3.0 Hz, H-6), 5.12
(d, 1H, Jgem¼14.1 Hz, H-5b), 4.81 (dd, 1H, J¼3.3, 5.6 Hz,
H-3a), 4.36 (d, 1H, Jgem¼14.1 Hz, H-5a), 4.01 (m, 2H,
H-2a, H-2b), 3.74 (s, 3H, 7-OMe), 3.49 (s, 3H, 3-OMe). 13C
NMR (75 MHz, CDCl3) d 155.10 (C-7), 152.91 (C-10),
131.15 (C-11), 120.86 (C-8), 113.82 (C-9), 113.62 (C-6),
101.54 (C-3), 73.00 (C-2), 63.23 (C-5), 55.71 (7-OMe),
55.63 (3-OMe). HR LSIMS calcd for C11H14O4Na
(MþNa)þ 233.0789, found: 233.0788. Anal. for
C11H14O4: calcd: C 62.85; H 6.71. Found: C 62.48; H 6.89.
˚
stored over molecular sieves (4 A). All reactions were
carried out in dry glassware and under dry argon.
Evaporations were carried out in vacuo with a rotary
evaporator.
4.1.1. Starting materials
4.1.1.1. Synthesis of 1-(hydroxymethyl)-2-phenoxy-
acetaldehyde dimethyl acetals 11–13. The preparation
of 11 was previously reported9 and we proceeded with the
same method to obtain 12 and 13 but changing the solvent of
the reaction DMSO by DMF, starting from 2-hydroxy-5-
methoxybenzaldehyde8 and the commercially available
2-hydroxy-3-methoxybenzaldehyde, respectively. The
purification of the liquid acyclic acetals 12 and 13 was
performed by flash chromatography using a diethyl ether/
hexane (1/2) mixture as eluant.
(RS)-3,9-Dimethoxy-2,3-dihydro-5H-1,4-benzodioxepin 16.
1
Yield: 80%. Rf (2/1, diethyl ether/hexane): 0.3 H NMR
(300 MHz, CDCl3) d 6.88 (pt, 1H, J¼7.6 Hz, H-7), 6.79 (dd,
1H, J¼1.6 Hz, 7.6, H-6 Or H-8), 6.65 (dd, 1H, J¼1.6,
7.5 Hz, H-8 Or H-6), 5.22 (D, 1H, J¼14.4 Hz; H-5b), 4.87
(dd, 1H, J¼3.3, 7.0 Hz; H-3a), 4.36 (D, 1H, J¼14.4 Hz,
H-5a), 4.20 (dd, 1H, J¼3.3 Hz, 13.0, H-2a), 4.1 (dd, 1H,
J¼7.0, 13.0 Hz, H-2b), 3.82 (S, 3H, 9-OMe), 3.46 (S, 3H, 3-
OMe). 13C NMR (75 MHz, CDCl3) d 150.50 (C-9), 147.90
(C-10), 130.49 (C-11), 122.61 (C-7), 120.20 (C-6), 111.44
(C-8), 101.25 (C-3), 72.37 (C-2), 62.85 (C-5), 56.08 (9-OMe),
55.48 (3-OMe). HR LSIMS calcd for C11H14O4Na (MþNa)þ
233.0789, found 233.0789. Anal. for C11H14O4: calcd: C
62.85; H 6.71. Found: C, 62.88; H, 6.68.
4.1.1.3. (Z)-4-(2-Hydroxyphenyl)-3-butenal dimethyl
acetal 18. Potassium t-butoxide (39.2 g, 0.35 mol) was
added in three portions to a mechanically stirred mixture of
crude phosphonium chloride 1710 (from 0.3 mol of acrolein)
in THF (400 mL) at 08C. The orange mixture was stirred at
room temperature during 2 h and cooled in ice. 2-Hydroxy-
benzaldehyde (16.5 mL, 0.31 mol) was added in 2 min and
the reaction mixture was stirred for 16 h at room
temperature. Extractive workup gave a syrup from which
most of the Ph3PO was removed by crystallization from
diethyl ether. The filtrate was concentrated and purified by
flash chromatography using a diethyl ether/hexane (1/9)
mixture as eluant, giving (Z)-18 (9.36 g, 15%) as a
1-(Hydroxymethyl)-5-methoxy-2-phenoxyacetaldehyde
dimethyl acetal 12. Yield: 72%. Rf (4/1, diethyl ether/
1
hexane): 0.4. Rf (3/1, diethyl ether/hexane): 0.2. H NMR
(300 MHz, CDCl3) d 6.82 (d, 1H, J¼2.9 Hz, H-6), 6.78 (d,
1H, J¼8.8 Hz, H-3), 6.72 (dd, 1H, J¼2.9, 8.8 Hz, H-4), 4.66
(t, 1H, J¼5.2 Hz, H-10), 4.60 (s, 2H, CH2OH), 3.98 (d, 2H,
J¼5.2 Hz, H-20), 3.72 (s, 3H, 5-OMe), 3.40 (s, 6H, (OMe)2).
13C NMR (75 MHz, CDCl3) d 154.13 (C-5), 150.62 (C-2),
131.34 (C-1), 114.78 (C-4), 113.55 (C-3), 113.19 (C-6),
101.80 (C-10), 66.42 (C-20), 61.83 (CH2OH), 55.64 (5-OMe),
53.94 (OMe)2). HR LSIMS calcd for C12H18O5Na (M þ Na)þ
265.1051, found 265.1050. Anal. for C12H18O5: calcd: C
59.49; H 7.49. Found: C 59.35;H 7.22.
1-(Hydroxymethyl)-3-methoxy-2-phenoxyacetaldehyde
dimethyl acetal 13. Yield: 64%. Rf (4/1, diethyl ether/
1
hexane): 0.4. H NMR (300 MHz, CDCl3) d 6.99 (t, 1H,
J¼7.9 Hz, H-5), 6.87 (d, 1H, J¼2.3 Hz, H-6 or H-4), 6.84
(dd,01H, J¼2.3 Hz, 4.0, H-4 or H-6), 4.69 (t, 1H, J¼5.1 Hz,
H-1 ), 4.60 (d, 2H, J¼7 Hz, CH2OH), 4.14 (d, 2H,
J¼5.1 Hz, H-20), 3.83 (s, 3H, 3-OMe), 3.42 (s, 6H, OMe2).
13C NMR (75 MHz, CDCl3) d 152.05 (C-2), 146.10 (C-3),
134.77 (C-1), 123.98 (C-5), 121.52 (C-6), 112.28 (C-4),
102.51 (C-10), 71.53 (C-20), 61.50 (CH2OH), 55.75 (3-OMe),
‡
According to Ref. 9 the distillation under a better vacuum, and therefore a
lower temperature of distillation (1108C/1 mm Hg), yielded the acyclic
acetal (62% yield).
The combined yield of 14 is <35% for the two-step process (see Ref. 9).
§

E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
5465
colourless oil. Rf (2/1, diethyl ether/hexane): 0.66. 1H NMR
(300 MHz, CDCl3) d 7.15 (dt, 1H, J¼1.7, 7.6 Hz, HAr), 7.04
(dd, 1H, J¼1.7, 7.5 Hz, HAr), 6.87 (m, 2H, HAr), 6.47 (d,
1H, J¼11.3 Hz, H-040), 6.11 (s, 1H, OH), 5.87 (dt, 1H,
J¼7.6, 11.3 Hz, H-3 ), 4.47 (t, 1H, J¼5.3 Hz, H-10), 3.35 (s,
6H, OMe), 2.42 (ddd, 2H, J¼1.6, 5.3, 7.6 Hz, H-20). 130C
NMR (75 MHz, CDCl3) d 152.77 (C-1), 129.91 (C-4 ),
129.33 (C-5), 128.76 (C-3), 126.71 (C-300), 123.33 (C-2),
120.18 (C-4), 116.22 (C-6), 103.83 (C-1 ), 53.91 (OMe),
32.93 (C-20). HR LSIMS calcd for C12H16O3Na (MþNa)þ
231.0997, found 231.0996. Anal. for C12H16O3: calcd: C,
69.21; H, 7.74. Found: C, 68.99; H, 7.82.
Method B. Anhydrous MeOH (27 mL) containing H2SO4
(0.13 mL) was added to 21 (3.84 g, 18.5 mmol) to give a
final 1% concentration of the acid. The solution was kept at
room temperature for 72 h, neutralized with a methanolic
solution of KOH, filtered and concentrated to dryness. After
˚
adding CHCl3 (35 mL) and washing with H2O (2 A 25 mL),
the organic layer was dried (MgSO4), filtered and
concentrated in vacuo, and the residue was purified by
flash chromatography using a diethyl ether/hexane (1/6)
mixture, yielding 21 (4.19 g, 60%) as a colourless liquid. Rf
(2/1, diethyl ether/hexane): 0.5. 1H NMR (300 MHz,
CDCl3) d 6.87 (m, 4H, HAr), 6.37 (bs, 1H, OH), 4.65 (t,
2H, J¼5.5 Hz, H-10), 4.10 (t, 2H, J¼6 Hz, H-30), 3.38 (s,
6H, (OMe)2), 2.10 (q, 2H, J¼5.8 Hz, H-20). 13C NMR
(75 MHz, CDCl3) d 146.64 (C-1), 145.92 (C-2), 122.35
(C-5), 120.10 (C-4), 115.07 (C-6), 113.87 (C-3), 102.47
(C-10), 66.15 (C-30), 53.23 (OMe2), 32.37 (C-20). HR LSIMS
calcd for C11H16O4Na (MþNa)þ 235.0946, found
235.0945. Anal. for C11H16O4: calcd: C, 62.25; H, 7.60.
Found: C, 62.28; H, 7.65.
4.1.1.7. (RS)-4-Methoxychroman-8-ol 20. The pro-
cedure was the same applied for the preparation of 19, but
using 21 (4.19 g, 19.7 mmol), anhydrous diethyl ether
(85 mL) containing BF3·Et2O (1.42 mL). The residue was
purified by flash chromatography with diethyl ether/hexane
(1/6), yielding 1.81 g (51%) of 20 as a colourless liquid. Rf
(2/1, diethyl ether/hexane): 0.49. 1H NMR (300 MHz,
CDCl3) d 6.85 (q, 2H, J¼4.8, 9.6 Hz, HAr), 6.79 (d, 1H,
J¼4.8 Hz, HAr), 5.06 (m, 1H, HAr), 5.56 (bs, 1H, OH), 4.33
(dd, 1H, J¼3.3, 6.7 Hz, H-4a), 4.31 (m, 2H, J¼3.5, 6.6,
9.1 Hz, H-2a, H-2b), 3.42 (s, 3H, OMe), 2.16 (dddd, 1H,
J¼3.3, 6.6, 14.2 Hz, H-3a), 2.04 (dddd, 1H, J¼3.5, 6.7, 9.1,
14.2 Hz, H-3b). 13C NMR (75 MHz, CDCl3) d 145.92 (C-9),
144.39 (C-8), 120.42 (C-10), 119.95 (C-6), 119.95 (C-5),
114.48 (C-7), 71.23 (C-4), 62.57 (C-2), 55.95 (OMe), 27.43
(C-3). HR LSIMS calcd for C10H12O3Na (M þ Na)þ
203.0694. found 203.0694. Anal. for C10H12O3: calcd: C,
66.60; H, 6.71. Found: C, 66.30; H, 6.68.
4.1.1.4. (RS)-2-Methoxy-2,3-dihydrobenzoxepin 19.
Compound 18 (2.3 g, 7.5 mmol) was dissolved in dry
diethyl ether (30 mL), and a few drops of BF3·OEt2 were
added. The solution was kept at room temperature for 7
days, washed with an aqueous solution of K2CO3 (10%),
and the organic layer was dried (MgSO4), filtered and
concentrated. The residue was purified by flash chromato-
graphy with diethyl ether/hexane (1/9), yielding 1.32 g
(34%) of 19 as a colourless oil. Rf (2/1, diethyl ether/
1
hexane): 0.7. H NMR (300 MHz, CDCl3) d 7.1 (m, 4H,
HAr), 6.32 (d, 1H, J¼11.9 Hz, H-5), 5.74 (ddd, 1H, J¼3.7,
5.5, 11.9 Hz, H-4), 4.94 (dd, 1H, J¼2.3, 7.4 Hz, H-2a), 3.55
(s, 3H, OMe), 2.80 (ddd, 1H, 13.6, H-3a), 2.72 (dddd, 1H,
J¼1.8, 5.5, 7.4, 13.6 Hz, H-3b). 13C NMR (75 MHz,
CDCl3) d 153.67 (C-10), 131.86 (C-5), 128.45 (C-8),
128.24 (C-6), 128.33 (C-11), 128.24 (C-6), 125.21 (C-4),
123.33 (C-7), 120.52 (C-9), 104.6 (C-2), 56.23 (OMe),
38.65 (C-3). HR LSIMS calcd for C11H12O2Na (M þ Na)þ
199.0837, found 199.0837. Anal. for C11H12O2: calcd: C,
74.98; H, 6.86. Found: C, 75.23; H 7.12.
4.1.1.5. 2-(2-Hydroxyphenoxy)ethyl-1,3-dioxolane 22.
2-(2-Bromoethyl)-1,3-dioxolane (16.44 g, 90.8 mmol) was
added in small portions to a suspension of catechol (10 g,
90.8 mmol) and K2CO3 (7.53 g, 54.5 mmol) in anhydrous
acetone (250 mL) after stirring at room temperature for
30 min under argon. The resulting suspension was refluxed
during 6 h, then brought back to room temperature, filtered
and diethyl ether (220 mL) was added and the resulting
solution washed with a NH4Cl saturated solution. The
organic phase was dried (MgSO4), filtered and evaporated
under reduced pressure to yield a colourless oil which, upon
flash chromatography using diethyl ether/hexane (1/6) as
eluant, gave pure 22 (3.90 g, 21%). Rf (4/1, diethyl ether/
4.1.2. Final products
4.1.2.1. Reaction between the benzoannelated seven-
or six-membered compounds 14–16, 19–20 and
5-fluorouracil. General procedure. A 1.0 M solution of
SnCl4/CH2Cl2 (1.25 mmol) was added dropwise with stirring
under argon at rt to a suspension of 14–16, 19–20 (1 mmol),
5-fluorouracil (1 mmol), which contains trimethylchloro-
silane (TCS, 0.8 mmol) and 1,1,1,3,3,3-hexamethyldisilazane
(HMDS, 0.8 mmol) in dry acetonitrile (10 mL/mmol of
14–16, 19–20). After 24–144 h of stirring, the reaction was
quenched by the addition of a concentrated aqueous solution
ofNa2CO3. Thesolventwasremovedwitharotaryevaporator.
The sticky residue was dissolved in MeOH, and silica gel was
added. The solvent was removed with a rotary evaporator, and
the residue was applied to the top of a flash chromatography
column packed with CH2Cl2/MeOH (100/1). Elution with
mixtures of CH2Cl2/MeOH by gradient elution (100/1!100/
5) and concentration with a rotary evaporator gave the target
molecules 5–9.
1
hexane): 0.57. H NMR (300 MHz, CDCl3) d 6.84 (m, 4H,
HAr), 5.06 (t, 1H, J¼4.5 Hz, H-30), 4.20 (t, 2H, 5.8, H-10),
4.00 (m, 2H, (OCH2)2); 3.90 (m, 2H, (OCH2)2), 2.15 (q, 2H,
J¼5.8 Hz, H-20). 13C NMR (75 MHz, CDCl3) d 147.21
(C-1), 146.10 (C-2), 122.70 (C-5)0, 120.07 (C-4), 115.12
(C-6), 114.62 (C-3), 0 103.07 (C-1 ), 66.36 (C-30), 65.04
(OCH2)2), 33.49 (C-2 ). HR (LSIMS) calcd for C11H14O4-
Na (MþNa)þ 233.0789, found: 233.0789. Anal. for
C11H14O4: calcd.: C, 62.85; H, 6.71. Found: C, 62.88; H,
6.60.
4.1.1.6. 2-(2-Hydroxyphenoxy)acetaldehyde dimethyl
acetal 21. The following two methods have been carried
out.
Method A. Following the procedure described for the
obtention of 22, but changing 2-(2-bromoethyl)-1,3-dioxo-
lane by 3-bromopropanal dimethyl acetal, gave pure 21
(1.2 g, 23%).
(RS)-1-(2,3-Dihydro-5H-1,4-benzodioxepin-3-yl)-5-fluoro-
uracil 5. Reaction time: 24 h. Yield: 43%. White solid, mp
196–1988C (decomp.). Rf (9.5/0.5, CH2Cl2/MeOH): 0.61;
Rf (9.9/0.1, CH2Cl2/MeOH): 0.32. 1H NMR (400 MHz,

5466
E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
DMSO-d6) d 11.96 (s, 1H, NH), 7.95 (d, 1H, JH–F¼7 Hz,
H5-FU), 7.27 (m, 2H, AA0 part of an AA0BB0 system, H-6 and
H-7 or H-8 and H-9), 7.06 (m, 2H, BB0 part of an AA0BB0
system, H-8 and H-9 or H-6 and H-7), 5.91 (dd, 1H, J¼1.6,
7.2 Hz, H-3a), 4.92 (d, 1H, Jgem¼14.2 Hz, H-5a), 4.87 (d,
1H, Jgem¼14.2 Hz, H-5b), 4.46 (dd, 1H, J¼2.2, 12.9 Hz,
H-2a), 4.13 (dd, 1H, J¼7.4, 12.9 Hz, H-2b). 13C NMR
(100 MHz, DMSO-d6) d 158.26 (C-10), 156.96 (d,
J¼25.81 Hz, C-45-FU), 148.71 (s, C-25-FU), 139.96 (d,
J¼229.57 Hz, C-55-FU), 130.47 (C-11), 129.41 and 129.23
(C-6 and C-8), 125.78 (d, J¼34.21 Hz, C-65-FU), 123.45
(C-9), 120.19 (C-7), 83.99 (C-3), 72.56 (C-2), 68.65 (C-5).
HR LSIMS calcd for C13H11N2O4FNa (MþNa)þ 301.0600,
found 301.0600. Anal. for C13H11N2O4F: calcd: C, 56.33;
H, 3.79; N, 10.13. Found: C, 56.35; H, 3.87; N, 10.01.
18.1 Hz, H-3a). 13C NMR (100 MHz, CDCl3) d 156.88 (d,
J¼24.7 Hz, C-45-FU), 155.47 (C-25-FU), 150.86 (C-10),
140.41 (d, J¼233.9 Hz, C-55-FU), 132.47 (C-5), 132.47 (d,
J¼29 Hz, C-65-FU), 126.92 (C-11), 126.63 (C-8), 123.89
(C-6), 123.64 (C-4), 123.32 (C-7), 120.26 (C-9), 83.68
(C-2), 35.85 (C-3). HR (EI) calcd for C14H11N2O3FNa
274.0753, found 274.0753. Anal. for C14H11N2O3F: calcd:
C, 61.31; H, 4.04; N, 6.93. Found: C, 61.68; H, 5.32; N,
7.24.
(RS)-1-(8-Hydroxychroman-4-yl)-5-fluorouracil 9. Reac-
tion time: 24 h. Yield: 44%. White solid, mp 138–1408C.
1
Rf (9/1, CHCl3/MeOH): 0.62. H NMR (400 MHz, CDCl3)
d 9.08 (s, NH), 6.96 (d, 1H, JH–F¼6.1 Hz, H5-FU), 5.08 (s,
1H, OH), 5.7 (t, 1H, J¼5.6 Hz, H-4a), 4.28 (ddd, 1H, J¼3.3,
7.2, 11.4 Hz, H-2a), 4.16 (ddd, 1H, J¼3.2, 8.0, 11.4 Hz,
H-2b), 2.29 (dddd, 1H, J¼3.3, 5.6, 8.0, 11.3 Hz, H-3a), 2.37
(dddd, 1H, J¼3.2, 7.2, 11.3 Hz, H-3b). 13C NMR
(100 MHz, CDCl3) d 157.10 (d, J¼25.7 Hz, C-45-FU),
149.89 (s, C-25-FU), 145.92 (C-11), 144.39 (C-10), 139.81
(d, J¼229.1 Hz, C-55-FU), 127.09 (d, J¼33.4 Hz, C-65-FU),
120.43 (C-6), 120.33 (C-8), 117.80 (C-7), 115.10 (C-9),
63.65 (C-4), 50.02 (C-2), 27.16 (C-3). HR LSIMS calcd for
C13H11N2O4FNa 278.0702, found: 278.0703. Anal. for
C13H11N2O4F: calcd: C, 56.12; H, 3.98; N, 10.07. Found:
C, 55.99; H, 4.12; N, 10.02.
(RS)-1-(7-Methoxy-2,3-dihydro-5H-1,4-benzodioxepin-3-
yl)-5-fluorouracil 6. Reaction time: 144 h. Yield: 40%.
White solid, mp 220–2228C. Rf (9/1, CH2Cl2/MeOH): 0.49.
1H NMR (400 MHz, DMSO-d6) d 11.16 (s, 1H, NH), 7.86
(d, 1H, JH–F¼5.4 Hz, H5-FU), 6.97 (d, 1H, J¼8.7 Hz, H-9),
6.88 (d, 1H, J¼3.1 Hz, H-6), 6.80 (dd, 1H, J¼3.1, 8.7 Hz,
H-8), 6.05 (d, 1H, J¼7.9 Hz, H-3a), 4.81 (d, 1H,
J¼13.7 Hz, H-5a), 4.65 (d, 1H, J¼13.7 Hz, H-5b), 4.59
(dd, 1H, J¼3.4, 12.3 Hz, H-2b), 4.35 (dd, 1H, J¼1.5,
12.3 Hz, H-2a), 3.7 (s, 3H, 7-OMe). 13C NMR (100 MHz,
DMSO-d6) d 157.06 (d, J¼25.4 Hz, C-45-FU), 154.96 (s,
C-25-FU), 152.71 (C-7), 149.32 (C-10), 139.15 (d,
J¼224.5 Hz, C-55-FU), 132.41 (C-11), 125.88 (d,
J¼31.5 Hz, C-65-FU), 121.15 (C-8), 114.48 (C-9), 114.14
(C-6), 84.88 (C-3), 73.49 (C-2), 70.28 (C-5), 55.41 (7-
OMe). HR LSIMS calcd for C14H13N2O5FNa (MþNa)þ
331.0706, found 331.0706. Anal. for C14H13N2O5F: C,
54.55; H, 4.25; N, 9.09. Found: C, 54.87; H, 4.33; N, 8.79.
4.1.3. X-Ray crystallographic study. Crystallographic data
were collected at 293 K using graphite monochromated
˚
Mo Ka radiation (l¼0.71073 A) on a Siemens P4 diffract-
ometer. The structure was solved by direct methods using
the program SHELXS9725 and all non-hydrogen atoms
were refined with anisotropic thermal parameters by full-
matrix least squares techniques of F 2 using the program
SHELXL97.26 Hydrogen atoms were inserted in calculated
positions and refined isotropically. Molecular graphics and
geometric calculations were obtained from SHELXTL27
and PLATON.28 Relevant crystal data: formula
C13H11FN2O4, formula weight 278.24, T¼293(2) K, crystal
system triclinic, space group P-1, unit cell dimensions
(RS)-1-(9-Methoxy-2,3-dihydro-5H-1,4-benzodioxepin-3-
yl)-5-fluorouracil 7. Reaction time: 24 h. Yield: 16%. White
1
solid, mp 130–1328C. Rf (9/1, CH2Cl2/MeOH): 0.5. H
NMR (400 MHz, CDCl3) d 9.25 (s, 1H, NH), 7.59 (d, 1H,
JH–F¼6 Hz, H5-FU), 7.04 (d, 1H, J¼8.0 Hz, H-5), 6.92 (dd,
1H, J¼1.4, 8.0 Hz, H-6), 6.76 (dd, 1H, J¼1.4, 7.6 Hz, H-4),
6.12 (d, 1H, J¼5.9 Hz, H-3a), 5.07 (d, 1H, J¼14.2 Hz,
H-5a), 4.79 (d, 1H, J¼14.2 Hz, H-5b), 4.59 (dd, 1H, J¼2.3,
13 Hz, H-2a), 3.97 (dd, 1H, J¼5.9, 13 Hz, H-2b), 3.88 (s,
3H, 9-OMe). 13C NMR (100 MHz, CDCl3) d 156.82 (d,
J¼26.5 Hz, C-45-FU), 151.04 (s, C-25-FU), 148.72 (C-8),
147.39 (C-10), 140.61 (d, J¼236.7 Hz, C-55-FU), 130.72
(C-11), 125.03 (d, J¼34.2 Hz, C-65-FU), 124.08 (C-7),
120.42 (C-6), 112.59 (C-8), 84.20 (C-3), 73.32 (C-2), 69.43
(C-5), 56.20 (9-OMe). HR (LSIMS) calcd for C14H13N2O5-
FNa (MþNa)þ 331.0706, found: 331.0706. Anal. for
C14H13N2O5F: calcd: C, 54.55; H, 4.25; N, 9.09. Found:
C, 54.88; H, 4.37; N, 9.45.
˚
a¼6.548(2), b¼7.794(2) and c¼12.504(3) A, and a¼
90.33(2), b¼92.24(2) and g¼107.55(2)8, Z¼2, D¼
1.520 mg m3, m(Mo Ka)¼0.124 mm21, measured/unique
reflections 4395/3535 (R(int) 0.051), refined parameters
225, refinement method¼full-matrix lest-squares on F 2,
final R1 (I.2s(I))¼0.066 and wR2¼0.176, and GOF¼
1.051. CCDC reference number 197927. Copies of the data
can be obtained, free of charge, on application to CCDC, 12
Union Road, Cambridge, CB2 1EZ, UK [fax: þ44(0)-1223-
336033 or e-mail: deposit@ccd.cam.acuk].
4.1.4. Conformational analysis. The study of the possible
conformations has been carried out by the Sybyl15 program
on a Silicon Graphics workstation. The different molecules
have been constructed from standard fragments of the
libraries of the programme. The force field Amber 4.1,29
implemented in the Sybyl programme, has been used in the
energy calculations. Once the initial geometries generated,
we proceeded to their optimization using the Powell16
method. The atomic charges were calculated by means of the
AM130 hamiltonian implemented in the MOPAC 6.031
programme. A distance-dependent dielectric constant with a
value of 1¼1 was used and the optimization was continued
(RS)-1-(2,3-Dihydrobenzoxepin-2-yl)-5-fluorouracil 8.
Reaction time: 24 h. Yield: 10%. White solid, mp 215–
2178C. Rf (9/0.5, CH2Cl2/MeOH): 0.60. 1H NMR
(400 MHz, CDCl3) d 9.81 (d, 1H, J¼5.2 Hz, NH), 7.20 (d,
1H, J¼5.6 Hz, H5-FU), 7.11 (m, 1H, H-8 or H-7), 7.06 (m,
1H, H-7 or H-8), 6.96 (d, 1H, J¼7.4 Hz, H-9 or H-6), 6.5 (d,
1H, J¼9.9 Hz, H-6 or H-9), 6.4 (d, 1H, J¼11.9 Hz, H-5),
6.3 (ddd, 1H, J¼11.9, 0 Hz, H-4), 5.88 (ddd, 1H, J¼2.7, 6.4,
9.1 Hz, H-2a), 4.14 (m, 1H, H-3b), 2.61 (dd, 1H, J¼6.4 Hz,

E. Saniger et al. / Tetrahedron 59 (2003) 5457–5467
5467
2
˚
until the energy gradient was less than 0.01 kcal/mol A .
12. Pretsch, E.; Bu¨hlmann, P.; Affolter, C. Structure
Determination of Organic Compounds; Springer: Berlin,
2000.
Conformational searches were carried out by means of
molecular dynamics, using the ‘simulated annealing’ tech-
nique, heating the molecule up to 1000 K for 1000 ps and
coolingitdownlaterexponentiallyto200 K,maintainingit for
another 1000 ps. Five hundred heating–cooling cycles were
carriedoutoneachmoleculeandthegeometriesobtainedatthe
end of each cooling period were kept. These 500 confor-
mations were optimized under the same conditions described
before and were compared with each other to remove those
which were geometrically and energetically equal. Such
conformations were optimized using ab initio (3-21G)
calculations using the Gaussian9818 program. After this,
those conformationsthatwereenergeticallyandgeometrically
different were selected for their subsequent study.
13. Osaki, S.; Watanabe, Y.; Hoshiko, T.; Nagase, T.;
Ogawawara, T.; Furukawa, H.; Uemura, A.; Ishikawi, K.;
Mori, A.; Iigok, M.; Tokuzen, R. Chem. Pharm. Bull. 1986,
34, 150–157.
14. Clark, M.; Cramer, R. D. J. Comp. Chem. 1989, 10, 982–1012.
15. SYBYL Molecular Modeling Software, Tripos Inc. 1699
S. Hanley Road, St. Louis MO 63144-2913, www.tripos.com.
16. Powell, M. J. D. Math. Program. 1997, 12, 241–254.
17. (a) Gasteiger, J.; Marsili, M. Tetrahedron 1980, 36,
3219–3228. (b) Marsili, M.; Gasteiger, J. Croat. Chem. Acta
1980, 53, 601–614. (c) Gasteiger, J.; Marsili, M. Org. Magn.
Reson. 1981, 15, 353–360.
18. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.;
Montgomery, J. A., Jr.; Stratmann, R. E., Burant, J. C.;
Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.;
Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.;
Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford,
S.; Ochterski, J.; Peterson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari,
K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Lomaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.;
Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.;
Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian, Inc.:
Pittsburgh, PA, 1998.
4.2. Biological activity
The antiproliferative activities20 and apoptosis induction23
in the MCF-7 human breast cancer cell line were followed
in accordance with the protocols previously reported. The
results are recorded in Table 1.
Acknowledgements
The authors are grateful to the Instituto Carlos III for its
financial support (Fondo de Investigaciones Sanitarias,
01/0928). The award of a grant from the Ministerio de
´
Educacion, Cultura y Deporte to E. S. is gratefully
acknowledged.
´
19. Entrena Guadix, A.; Gallo Mezo, M. A.; Campos Rosa, J.;
´
Gomez Vidal, J. A.; Espinosa Ubeda, A. J. Org. Chem. 1997,
´
References
62, 337–349.
´
´
´
20. Campos, J.; Nun˜ez, M. C.; Rodrıguez, V.; Gallo, M.A.;
Espinosa, A. Bioorg. Med. Chem. Lett. 2000, 10, 767–770.
21. Meyn, R. E.; Stephens, L. C.; Hunter, N. R.; Milas, L. Int.
J. Cancer 1995, 60, 725–729.
1. Harris, J. R.; Lippman, M. E.; Morrow, M.; Hellman, S.
Diseases of the Breast; Lippincott-Raven: New York, 1996;
p 159.
2. Hickman, J. A. Cancer Metastasis Rev. 1992, 11, 121–139.
3. Rasbridge, S. A.; Gillet, C. E.; Seymour, A. M.; Patel, K.;
Richards, M. A.; Rubens, R. D.; Millis, R. R. R. Brit. J. Cancer
1994, 70, 335–341.
22. Milas, L.; Hunter, N. R.; Kurdoglu, B.; Mason, K. A.; Meyn,
R. E.; Stephens, L. C.; Peters, L. J. Cancer Chemother. Pharm.
1995, 35, 297–303.
23. Chadderton, A.; Villeneuve, D. J.; Gluck, S.; Kirwan-Rhude,
A. F.; Gannon, B. R.; Blais, D. E.; Parissenti, A. M. Breast
Cancer Res. Tr. 2000, 59, 231–244.
´
4. Campos, J.; Pineda, M. J.; Gomez, J. A.; Entrena, A.; Trujillo,
´
M.A.; Gallo, M.A.; Espinosa, A. Tetrahedron 1996, 52,
´
8907–8924.
24. Saunders, D. E.; Lawrence, W. D.; Christensen, C.; Wappler,
N. L.; Ruan, H.; Deppe, G. Int. J. Cancer 1997, 70, 214–220.
25. Sheldrick, G. M. SHELXS97: Program for the solution of
´
5. Gomez, J. A.; Trujillo, M.A.; Campos, J.; Gallo, M.A.;
´
´
Espinosa, A. Tetrahedron 1998, 54, 13295–13312.
´
6. Trujillo, M.A.; Gomez, J. A.; Campos, J.; Espinosa, A.; Gallo,
´
¨
crystal structures; University of Gottingen: Germany, 1997.
´
M.A. Tetrahedron 2001, 57, 3951–3961.
26. Sheldrick, G. M. SHELXL97: Program for the refinement of
¨
crystal structures; University of Gottingen: Germany, 1997.
´
´
´
7. Campos, J.; Gomez, J. A.; Trujillo, M.A.; Gallo, M.A.;
Espinosa, A. Il Farmaco 1997, 52, 263–269.
8. Hammond, M. L.; Zambias, R. A.; Chang, M. N.; Jensen, N. P.;
McDonald, J.; Thompson, K.; Boulton, D. A.; Kopka, I. E.;
Hand, K. M.; Opas, E. E.; Luell, S.; Bach, T.; Davies, P.;
MacIntyre, D. E.; Bonney, R. J.; Humes, J. L. J. Med. Chem.
1990, 33, 908–918.
27. Bruker, SHELXTL: Integrated system for the determination of
crystal structures; Bruker AXS Inc.: Madison, WI, USA,
2000.
28. Spek, A. L. PLATON: A Multipurpose Crystallographic Tool;
Utrecht University: The Netherlands, 2002.
29. Weiner, P. K.; Kollman, P. A. J. Comput. Chem. 1981, 2,
287–303.
9. Lachapelle, A.; St-Jacques, M. Tetrahedron 1988, 44,
5033–5044.
30. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.
J. Am. Chem. Soc. 1985, 107, 3902–3909.
31. Stewart, J. J. P. MOPAC 6.0 QCPE No. 455, 1990.
10. Wanner, M. J.; Koomen, G.-J. Eur. J. Org. Chem. 1998,
889–895.
11. Schlosser, M. Top. Stereochem. 1970, 5, 1–30.