Diastereomeric 2-aminomethyl-1,4-benzodioxane mandelates: Phase diagrams and resolution

Article abstract of DOI:10.1016/j.tetasy.2013.05.010

The diastereomeric salts of (R)- and (S)-2-aminomethyl-1,4-benzodioxane with unichiral mandelic acid form a simple eutectic, whose binary phase melting point diagram shows the unique eutectic at 0.35 M ratio of the less soluble diastereomer. Such an eutectic composition, near to 0.5, is consistent with the modest efficiency previously reported for their separation via crystallization from ethanol/ethyl acetate. However, the ternary solubility phase diagram, obtained from solubility measurements in methanol, shifts the eutectic to a lower molar ratio (0.10) of the less soluble diastereomer, thus indicating an optimal resolvability of the diastereomeric mandelates. This was confirmed by the highly efficient resolution of racemic 2-aminomethyl-1,4-benzodioxane with (R)-mandelic acid via a single crystallization from methanol. The ready availability of both the racemic substrate and the resolving acid makes this simple and efficient resolution procedure very attractive to obtain the enantiomers of 2-aminomethyl-1,4-benzodioxane, which are important synthetic intermediates.

Full text of DOI:10.1016/j.tetasy.2013.05.010

Tetrahedron: Asymmetry 24 (2013) 796–800
Contents lists available at SciVerse ScienceDirect
Tetrahedron: Asymmetry
journal homepage: www.elsevier.com/locate/tetasy
Diastereomeric 2-aminomethyl-1,4-benzodioxane mandelates: phase
diagrams and resolution
⇑
Cristiano Bolchi, Marco Pallavicini , Laura Fumagalli, Paola Ruggeri, Ermanno Valoti
Dipartimento di Scienze Farmaceutiche, Università degli Studi di Milano, via Mangiagalli 25, I-20133 Milano, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 20 March 2013
Accepted 17 May 2013
The diastereomeric salts of (R)- and (S)-2-aminomethyl-1,4-benzodioxane with unichiral mandelic acid
form a simple eutectic, whose binary phase melting point diagram shows the unique eutectic at
0.35 M ratio of the less soluble diastereomer. Such an eutectic composition, near to 0.5, is consistent with
the modest efficiency previously reported for their separation via crystallization from ethanol/ethyl ace-
tate. However, the ternary solubility phase diagram, obtained from solubility measurements in methanol,
shifts the eutectic to a lower molar ratio (0.10) of the less soluble diastereomer, thus indicating an opti-
mal resolvability of the diastereomeric mandelates. This was confirmed by the highly efficient resolution
of racemic 2-aminomethyl-1,4-benzodioxane with (R)-mandelic acid via a single crystallization from
methanol. The ready availability of both the racemic substrate and the resolving acid makes this simple
and efficient resolution procedure very attractive to obtain the enantiomers of 2-aminomethyl-1,4-ben-
zodioxane, which are important synthetic intermediates.
Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction
benzodioxane-2-carboxylate or of 1,4-benzodioxane-2-methyl
mesylate;20 resolution of 1,4-benzodioxane-2-carboxylic acid 2
A variety of pharmacologically active compounds1–6 have 2-
aminomethyl-1,4-benzodioxane 1 as a key chiral substructure,
with dehydroabietylamine or para-substituted 1-phenylethylam-
ines (Chart 1).21,22
including many
a1-adrenergic receptor (a1-AR) antagonists. The
first studies on the marked difference in antihypertensive action
between the enantiomers of 2-diethylaminomethyl-1,4-benzodi-
oxane (Prosympal) were conducted by Bovet in the 1930s.7 Succes-
sively, in the sixties and in the seventies, the pharmacological
effects of the enantiomers of other antiadrenergic amin-
omethylbenzodioxanes, such as guanoxan and WB4101, were
investigated.8,9 In 1979, the (S)-enantiomer of WB4101 was proven
O
O
COOH
NH2
*
*
O
O
1
2
to be 50 times as potent as the (R)-enantiomer in blocking
a-
OH
COOH
adrenergic receptor activity.9 More recently, we have reported that
the (S)-enantiomers of WB4101 and of a wide number of amin-
omethylbenzodioxane derivatives are more potent antagonists
X
NH2
than the (R)-enantiomers at the three a1-adrenergic receptor (a1-
p-substituted
1-phenylethylamine
AR) subtypes.10–15 Hence, much effort has been devoted to obtain-
ing, as synthesis intermediates, both enantiomers of 1,4-benzodi-
3
oxanes bearing, at the stereogenic C2 center,
a suitably
Chart 1. 2-Aminomethylbenzodioxane and benzodioxane-2-carboxylic acid and
their respective resolving agents.
substituted carbon, that is, CH2NH2, –CH2OH, –CH2OMs, –CH2OTs,
–CH2Hal, –COOH, or –COOMe. To this end, we have applied differ-
ent strategies: the use of building blocks from the chiral pool, such
as (S)-solketal,16,17 resolution of solketal hydrogen esters with 1-
phenylethylamine,10,18,19 entrainment resolution of methyl 1,4-
Each approach has its own drawbacks. The use of only one glyc-
erol acetonide enantiomer as a building block of the chiral dioxane
portion implies that two different multi-step synthetic strategies
have to be adopted to obtain both 2-substituted 1,4-benzodioxane
enantiomers. On the other hand, starting from the two solketal
enantiomers, obtained via resolution of the racemate, allows a
⇑
Corresponding author. Tel.: +39 2 50319336; fax: +39 2 50319359.
E-mail address: marco.pallavicini@unimi.it (M. Pallavicini).
0957-4166/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetasy.2013.05.010

C. Bolchi et al. / Tetrahedron: Asymmetry 24 (2013) 796–800
797
unique strategy to be applied. However, this is still laborious and
requires that the hydroxyl groups of both building blocks, glycerol
and catechol, are, at some time, protected, deprotected or made
more reactive.
unique literature example of rac-1 resolution via diastereomeric
salt formation was not encouraging: rac-1 was resolved by (S)-3
to give, after four crystallizations (R)-1 in 31% yield, relative to
the theoretical amount, that is half of the starting racemate, while
(ꢀ)-dibenzoyltartaric acid precipitates, from all of the collected
mother liquors, (S)-1, which was isolated in 31% yield after three
recrystallizations.8 In summary, the whole procedure allows 15
mol of both (R)- and (S)-1 to be obtained from 100 mol of rac-1.
Such modest yields prompted us to characterize two diastereo-
meric mandelates, in particular the (R)-mandelates of (S)- and (R)-
1 [(S)-1ꢁ(R)-3 and (R)-1ꢁ(R)-3] and to investigate the solid phase
behaviour of their mixtures. It is known that the separation via
crystallization process depends on differences in the individual salt
properties and also on the binary melt phase diagram. Three main
types of phase diagrams occur: simple eutectic; two or more eutec-
tics with double salt formation, and no eutectic (mixed crystals).
Simple eutectic behavior is the most suitable one for the separation
of diastereomeric salts via crystallization, because it makes the res-
olution achievable directly from the 1:1 starting mixture. The
melting temperature of (S)-1ꢁ(R)-3 (167.4 °C) exceeded that of
(R)-1ꢁ(R)-3 (155.3 °C) by 12 degrees; that is a moderate difference
portending not a big difference in solubility and, in the case of sim-
ple eutectic behavior, an eutectic composition not far from 1:1. The
DSC analysis of a series of differently proportioned (S)-1ꢁ(R)-3/(R)-
1ꢁ(R)-3 mixtures revealed that the two diastereomeric mandelates
produce a simple eutectic with two clearly detectable melting pro-
files, the former of the eutectic and the latter of the excess pure
diastereomer. Figure 1 shows the DSC traces of the two pure salts
and of a number of their mixtures. The resultant experimental bin-
ary diagram, depicted in Figure 2, is typical of a simple eutectic
without detectable partial solid solution.
The classical resolution approach remains the most attractive
one, if an easily accessible racemate can be efficiently resolved
with a readily available resolving agent in an early step of a short-
ened synthesis. Indeed, rac-2 can be easily prepared by condensa-
tion of catechol with ethyl 2,3-dibromopropionate and successive
saponification;23 its resolution with dehydroabietylamine or para
substituted 1-phenylethylamines is very efficient, but such resolv-
ing agents are not ready available and their use is rather uncom-
mon. Moreover, the conversion of the enantiomers of 2 into (R)-
and (S)-1 requires additional steps, for instance esterification,
reduction to an alcohol, mesylation or tosylation, amination via
azide formation and reduction or Gabriel synthesis. Therefore,
since rac-1 is as easily prepared as rac-2 by condensation of cate-
chol with 2,3-dibromopropionitrile and successive reduction,3,24
we decided to investigate its resolvability by unichiral acids. An
efficient and easy resolution of rac-1 would mean a radically short-
ened access to (R)- and (S)-1 and would undoubtedly be preferable
over the resolution of the acid.
2. Results and discussion
Much work has been done in the field of racemate resolutions
via diastereomeric salt formation. This notwithstanding, the
resolvability of a racemate by a particular resolving agent remains
mostly unpredictable and the selection of the latter largely based
on empirical criteria though the correlation between resolution
efficiency and differences in crystal structures between diastereo-
meric salts have been investigated.25–29 Studies on such correla-
tions have mainly focused on the importance of hydrogen bond
patterns and aromatic moieties packing in determining different
supramolecular assembling in diastereomeric salt pairs and thus
different solubilities. Previously, we have found that rac-2 is effi-
ciently resolved by p-nitro and p-methyl substituted 1-phenyleth-
ylamine but not by unsubstituted 1-phenylethylamine.22
Consistently, the diastereomeric benzodioxanecarboxylates of the
two para-substituted 1-phenylethylamines have markedly differ-
ent solubilities and different crystal structures, in particular at
the boundaries of the chains formed by the ionic pairs, while the
1-phenylethylamine benzodioxanecarboxylates have similar solu-
bilities and no significant differences can be found between their
structures.
As can be seen in Figure 2, the experimental values fit accept-
ably within the theoretical ones (solid curve) calculated on the ba-
sis of the melting points of (S)-1ꢁ(R)-3 and (R)-1ꢁ(R)-3 and of the
respective fusion enthalpies (55.19 and 39.16 kJ molꢀ1) using the
These observations on the resolvability of phenylethylamine
benzodioxanecarboxylates led us to suppose that diastereomeric
aminomethylbenzodioxane mandelates, which can be seen as
analogous salts with exchanged amino and carboxy moieties,
would be also resolvable by selective crystallization. Analogously
to p-substituted 1-phenylethylamines, p-substituted mandelic
acids have also been found to sometimes have a higher resolving
ability than the unsubstituted parent compound.30 Furthermore,
according to the so-called ‘family approach’ or ‘Dutch resolution’,
both mixtures of p-substituted 1-phenylethylamines and mixtures
of p-substituted mandelic acids sometimes have the capability to
resolve better than individual pure 1-phenylethylamine and man-
delic acid respectively.31 Nevertheless, we decided to first try the
resolution with ready available mandelic acid 3 hoping that it
might efficiently resolve rac-1 itself. In mandelic acid, the presence
of the polar OH function, instead of a methyl, as in 1-phenylethyl-
amine, at the stereogenic carbon bearing the group responsible for
salt formation, should positively contribute to the ability of resolv-
ing rac-1, thus making the para-substitution unnecessary, which is
essential to 1-phenylethylamine for resolving rac-2. However, the
Figure 1. Upper curves: superimposed DSC traces of pure (S)-1/(R)-3 (black trace),
eutectic mixture [35/65 (S)-1ꢁ(R)-3/(R)-1ꢁ(R)-3 (red trace)] and mixtures with
intermediate compositions (blue traces); lower curves: superimposed DSC traces of
pure (R)-1/(R)-3 (black trace), eutectic mixture (65/35 (R)-1ꢁ(R)-3/(S)-1ꢁ(R)-3 (red
trace) and mixtures with intermediate compositions (blue traces).

798
C. Bolchi et al. / Tetrahedron: Asymmetry 24 (2013) 796–800
180
170
160
150
140
130
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
R
R
molar ratio of ( )- .( )-
1
3
Figure 2. Binary melting-point phase diagram for the diastereomeric system (S)-
1ꢁ(R)-3/(R)-1ꢁ(R)-3. The solid curve represents the values calculated on the basis of
the Schröder–van Laar equation.
Schröder–van Laar equation. The two branches of the theoretical
curve intersect at 0.35 veu [molar fraction of the higher melting
diastereomer (S)-1ꢁ(R)-3] and such an eutectic composition is con-
sistent with the trend of the experimental values and with the DSC
trace of the 0.35/0.65 mixture, which exhibits only the fusion of the
eutectic at 138.8 °C (melting peak maximum).
Figure 3. Ternary phase diagram for (S)-1ꢁ(R)-3 and (R)-1ꢁ(R)-3 salts in methanol at
23 °C. The solid red line represents the solubility isotherm. The concentrations of
the components are expressed as weight percentages.
The resolving ability S of (R)-3 can range from 0 (no resolution)
to 1, when 100% of (S)-1 is obtained with 100% ee. If the composi-
tion of the eutectic of the binary phase diagram (0.35 veu) is inde-
pendent of the nature of the crystallization solvent and the
crystallization is stopped when the mother liquor reaches the eu-
tectic composition, then pure (S)-1ꢁ(R)-3 has to be obtained with
represents the eutectic composition in the diastereomeric system.
Only one eutectic was observed in the ternary diagram, which
again proves the simple eutectic behavior.
The knowledge of the solubility diagram in methanol and of the
eutectic composition (0.10 veu of (S)-1ꢁ(R)-3) allowed a reliable
0.89 S value to be calculated for the diastereoselective crystalliza-
S = (1 ꢀ 2veu)/(1 ꢀ
veu) = 0.46. This means that, under the best
crystallization conditions, one can obtain 23 mol of diastereomer-
ically pure (S)-1ꢁ(R)-3 from 100 mol of 1:1 (S)-1ꢁ(R)-3/(R)-1ꢁ(R)-3
mixture, that is 23 mol of enantiomerically pure (S)-1 from
100 mol of rac-1. On the basis of the calculated 0.46 S value, the
0.31 efficiency of the literature resolution with mandelic acid in
3:1 ethyl acetate/ethanol seemed to have little room for improve-
ment. This notwithstanding, the use of such an unusual crystalliza-
tion solvent prompted us to measure the solubility of (S)-1ꢁ(R)-3
and (R)-1ꢁ(R)-3 in alcohols, namely methanol and ethanol. The re-
sults were noteworthy: the lower melting (R)-1ꢁ(R)-3 diastereomer
was 8.7 and 5.7 times more soluble than the higher melting (S)-
1ꢁ(R)-3 diastereomer in methanol and in ethanol, respectively
(Table 1).
tion from methanol [S = (1 – 2
veu)/(1 ꢀ
veu) = 0.89] and the opti-
mal initial concentration to be selected. As shown in the ternary
diagram, the resolution of the racemate starts at some point lying
on the straight line AB and only the ternary (R,R) salt/(S,R) salt/
methanol mixtures having compositions lying between C (92.7%
methanol) and D (58.2% methanol) produce, under equilibrium
conditions at 23 °C, pure (S)-1ꢁ(R)-3 as a precipitate with increasing
yield from C to D. Therefore, the optimal crystallization should be
accomplished from solutions obtained by dissolving 41.8 g of a 1:1
(R)-1ꢁ(R)-3/(S)-1ꢁ(R)-3 mixture in 58.2 g of methanol (ternary com-
position D) so as to precipitate the maximum amount of diastereo-
merically pure (S)-1ꢁ(R)-3. However, such a high concentration
makes the complete removal of the mother liquor from the precip-
itate by filtering and washing difficult. Hence we used a more di-
lute solution of the two salts (80.5% methanol; point F in the
ternary diagram). At this concentration at 23 °C, the solubility dia-
gram (see the red dashed line) indicates that the system is com-
posed of mother liquor having 86/10.5/3.5 methanol/(R,R)/(S,R)
composition (point H) in equilibrium with precipitated (S)-1ꢁ(R)-
3 containing enantiomerically pure (S)-1 (point I). Indeed, under
such conditions, a precipitate of (S)-1ꢁ(R)-3 was formed and col-
lected in 65% yield [32.5% of the starting rac-1] and with near to
100% ee (97.2%) of (S)-1 (S = 0.63), while the diastereomeric com-
position of unprecipitated salt, 73.47% (R)-1ꢁ(R)-3 and 26.53% (S)-
1ꢁ(R)-3, was in full agreement with the above reported 10.5% and
3.5% that is the diastereomeric ratio corresponding to point G.
The amine (S)-1 was liberated from precipitated (S)-1ꢁ(R)-3 in
quantitative yield.
On the basis of such solubility differences, a maximum value of
S can be calculated to range from 0.89 to 0.82 for crystallization
from saturated methanolic and ethanolic (R)-1ꢁ(R)-3 saturated
solutions respectively. The modest 0.46 S value calculated from
0.35 veu, compared to the high 0.82–0.89 S value calculated from
the solubility differences, indicated that although solubility and
fusibility of the two diastereomeric mandelates are related, the
location of the eutectic in the ternary solubility diagram has to
be more eccentric than in the binary one, that is more (R)-1ꢁ(R)-3
enriched. Therefore, we constructed the ternary solubility diagram
for (S)-1ꢁ(R)-3 and (R)-1ꢁ(R)-3 in methanol at 23 °C (Fig. 3). The
maximum solubility (point E on the red solubility curve) was ob-
served at approximately 80% diastereomeric excess (d.e.), at a com-
position of 90:10 (R)-1ꢁ(R)-3/(S)-1ꢁ(R)-3 (point E0), which hereby
Table 1
The resolution of rac-1 with (R)-3 was also carried out in etha-
nol, where the solubility ratio between (S)-1ꢁ(R)-3 and (R)-1ꢁ(R)-3
was lower than in methanol, 5.7 vs. 8.7. This resulted in a lower,
but still acceptable, 0.52 resolving ability, that is precipitation of
(S)-1ꢁ(R)-3 in 53% yield (27.5% of the starting 1:1 diastereomer
mixture) with 98.3% d.e.
Solubility (mg/mL) of the diastereomeric salts (R)-1ꢁ(R)-3 and (S)-1ꢁ(R)-3 in methanol
and in ethanol at 23 °C
Solvent
(R)-1ꢁ(R)-3
(S)-1ꢁ(R)-3
Solubility ratio
MeOH
EtOH
261.0
22.9
29.9
4.0
8.7
5.7

C. Bolchi et al. / Tetrahedron: Asymmetry 24 (2013) 796–800
799
Finally, we prepared two additional pairs of diastereomeric salts
of 1: the monotartrates with
-(ꢀ)-tartaric acid and the ketogulo-
nates with (ꢀ)-2,3:4,6-di-O-isopropylidene-2-keto- -gulonic acid.
4.2. DSC analyses
D
L
Samples of 2–3 mg were run in sealed aluminium pans with a
heating rate of 5 K minꢀ1. The pure diastereomeric salts (S)-1ꢁ(R)-
3 and (R)-1ꢁ(R)-3 were prepared by combining equimolar quanti-
ties of (R)-3 and (S)- or (R)-1 in ethanol, recovering the precipitate
and recrystallizing it from ethanol. The 1:1 (S)-1ꢁ(R)-3/(R)-1ꢁ(R)-3
mixture was prepared by concentration of a methanol solution
containing equivalent amounts of rac-1 and (R)-3. The DSC trace
of such a mixture exhibited the melting peak of the eutectic, fol-
lowed by the anticipated fusion of the exceeding (S)-1ꢁ(R)-3 diaste-
reomer. The other non equimolar diastereomeric mixtures were
prepared by mixing (S)-1ꢁ(R)-3 with increasing quantities of (R)-
1ꢁ(R)-3.
The solubilities in methanol were identical for the two ketogulo-
nates and similar for the two monotartrates thus precluding any
chance of diastereoselective crystallization from this solvent and
discouraging the use of such resolving agents. Comparison of the
structures of these two acids, both unable to resolve rac-1, with
that of mandelic acid suggests that the latter owes its resolving
ability of rac-1 to a beneficial combination of structural require-
ments, such as rigidity and the presence of both phenyl and hydro-
xy group at the stereogenic carbon.
3. Conclusion
4.3. Solubility measurements
2-Aminomethyl-1,4-benzodioxane rac-1 was efficiently resolved
with (R)-mandelic acid (R)-3, in methanol after constructing the bin-
ary phase diagram of the two diastereomeric mandelates and their
solubility curve in methanol. Although the crystallization could
not be accomplished at the ternary composition (point D, Fig. 3)
assuring the highest resolving ability (0.89), nevertheless, under
the selected conditions, an experimental 0.63 resolution efficiency
(97.2% ee ꢂ 65% yield) was obtained, which is markedly higher than
Samples of 1 g of pure (S)-1ꢁ(R)-3 and (R)-1ꢁ(R)-3 and of their
mixtures, prepared as described for DSC analyses, were introduced
into a glass vessel and kept at a constant temperature of 23 °C.
Methanol was gradually added by a burette with 0.05 mL accu-
racy waiting for attainment of equilibrium conditions after each
addition. The homogeneity of the suspensions was maintained
with a magnetic stirrer. Methanol was added until no crystals were
visible, at which point the achieved ternary solutes/solvent compo-
sitions, expressed as weight percentages, were plotted in a ternary
solubility phase diagram obtaining the solubility isotherm.
both the 0.46 value predicted on the basis of the 0.35 veu of the bin-
ary phase diagram and the 0.31 value reported in the literature. The
results of this resolution via diastereoselective crystallization dem-
onstrate that due to the different location of the eutectics, the solu-
bility diagram, not the binary diagram, allows the resolution
efficiency to be correctly predicted and the proper initial salt con-
centration to be selected in order to precipitate the less soluble salt
diastereomerically pure in a predetermined yield. There is a large
discrepancy between the binary phase diagram, which shows a near
4.4. Resolution of rac-1 in methanol
(R)-Mandelic acid (1.842 g, 12.11 mmol) was added to a stirred
solution of rac-1 (2 g, 12.11 mmol) in methanol (20 mL) at 23 °C. A
white solid precipitated after 10 min. The suspension was stirred
overnight at 23 °C and then filtered at this temperature to give
(S)-1ꢁ(R)-3 (1.25 g, 65% of the theoretical amount) as a white crys-
0.5 eutectic composition (0.35
exhibits a much more eccentric eutectic (0.10
v
eu), and the solubility curve, which
eu). This is due to
v
the alcoholic solvent, which is influential upon the eutectic location.
It is reasonable to assume that the OH function, linked to the stere-
ogenic carbon of the resolving agent, is important in making (S)-
1ꢁ(R)-3 have a higher melting and be less soluble than (R)-1ꢁ(R)-3;
for this reason, a hydroxylic solvent such as methanol is able to con-
talline solid: ½a D25
¼ ꢀ94:7 (c 1, MeOH); mp 164.6 °C; ee of (S)-1
ꢃ
97.2% [determined by HPLC analysis of the acetamide of the amine
liberated from a sample of the salt on a Chiralcel OJ column; hex-
ane/2-propanol 90/10; 0.8 mL/min; 280 nm; (S)-1 acetamide: tR
ꢄ15 min; (R)-1 acetamide: tR ꢄ18 min]; 1H NMR (DMSO-d6): d
2.94 (dd, 1H, J = 7.2, 13.4 Hz), 3.05 (dd, 1H, J = 3.9, 13.4 Hz), 3.97
(dd, 1H, J = 7.4, 12.1 Hz), 4.28–4.32 (m, 2H), 4.61 (s, 1H), 6.81–
6.90 (m, 4H), 7.12–7.25 (m, 3H and NH3+), 7.35 (d, 2H, J = 7.7 Hz).
The salt was decomposed by treatment with 1 M NaOH and dichlo-
romethane. The organic phase was separated, dried over Na2SO4
and concentrated to give (S)-1 (651 mg, 65% of the theoretical
vert a moderate difference in fusibility, that is a 12 °C Dmp, to a large
solubility difference shifting the eutectic towards diastereomerical-
ly pure (R)-1ꢁ(R)-3, and thus allowing a high diastereoselection in
the crystallization.
4. Experimental
4.1. General
amount, that is, half of the starting rac-1) as
a light oil:
½
a 2D5
ꢃ
¼ ꢀ56:9 (c 1, CHCl3); ee identical to that previously deter-
mined for the amine liberated from a sample of the precipitated
salt; 1H NMR (CDCl3): d 1.49 (br s, 2H), 2.98 (m, 2H), 4.01 (dd,
1H, J = 7.4, 11.3 Hz), 4.12 (m, 1H), 4.28 (dd, 1H, J = 2.2, 11.3 Hz),
6.81–6.90 (m, 4H).
Proton nuclear magnetic resonance (1H NMR) spectra were re-
corded using a Varian Gemini 300 (300 MHz). Chemical shifts are
reported in delta (d) units. Coupling constants are reported in Hertz
(Hz). The melting points were determined by DSC analysis and cor-
respond to the extrapolated onset temperature, while the final
melting temperatures were utilized to construct the binary phase
diagram. The DSC curves were recorded and integrated with the
aid of a TA Instruments DSC Q2010 apparatus. Optical rotations
were determined in a 1 dm cell with a 1 mL capacity using a Per-
kin–Elmer 241 polarimeter. HPLC analyses of 1 were performed
after conversion into the corresponding acetamide using a Chiral-
cel OJ column (250 mm ꢂ 4.6 mm i.d.). The racemate of 1 was syn-
thesized from catechol and 2,3-dibromopropionitrile,3,24 while (R)-
and (S)-1, which were necessary to prepare the samples of diaste-
reomeric mandelates, were obtained by resolution of rac-2 and
successive multi-step transformation of the carboxy group into
an amine function.21,10
4.5. Resolution of rac-1 in ethanol
(R)-Mandelic acid (1.842 g, 12.11 mmol) was added to a stirred
solution of rac-1 (2 g, 12.11 mmol) in ethanol (95 mL) at 25 °C. A
white solid precipitated after 10 min. The suspension was stirred
overnight at 23 °C and then filtered at this temperature to give
(S)-1ꢁ(R)-3 (1.38 g, 72% of the theoretical amount) as a white crys-
talline solid: mp 160.6 °C; ee of (S)-1 82.16% (determined by HPLC
of the acetamide of the amine liberated from a sample of the salt
under the above reported analysis conditions). The solid was sus-
pended in ethanol (13 mL), heated at 60 °C for 1 h, cooled to
25 °C and stirred overnight. The suspension was filtered to give
(S)-1ꢁ(R)-3 (1.02 g, 53% of the theoretical amount) as a white crys-

800
C. Bolchi et al. / Tetrahedron: Asymmetry 24 (2013) 796–800
13. Pallavicini, M.; Budriesi, R.; Fumagalli, L.; Ioan, P.; Chiarini, A.; Bolchi, C.;
talline solid: ½a D25
ꢃ
¼ ꢀ94:9 (c 1, MeOH); mp 165.9 °C; ee of (S)-1
Ugenti, M. P.; Colleoni, S.; Gobbi, M.; Valoti, E. J. Med. Chem. 2006, 49, 7140–
7149.
98.3%.
14. Valoti, E.; Pallavicini, M.; Villa, L.; Pezzetta, D. J. Org. Chem. 2001, 66, 1018–
1025.
15. Fumagalli, L.; Pallavicini, M.; Budriesi, R.; Gobbi, M.; Straniero, V.; Zagami, M.;
Chiodini, G.; Bolchi, C.; Chiarini, A.; Micucci, M.; Valoti, E. E. J. Med. Chem. 2012,
58, 184–191.
Acknowledgment
We thank the Italian Ministry of University and Research for
financial support.
16. Ferri, V.; Pallavicini, M.; Piccini, D.; Valoti, E.; Villa, L. Farmaco 1988, 43, 1153–
1163.
17. Villa, L.; Valoti, E.; Villa, A. M.; Pallavicini, M.; Ferri, V.; Iuliano, E.; Brunello, N.
Farmaco 1994, 49, 587–606.
References and notes
18. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahedron: Asymmetry 1994, 5,
5–8.
19. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O. Tetrahedron: Asymmetry 2001, 12,
2489–2495.
20. Bolchi, C.; Pallavicini, M.; Fumagalli, L.; Rusconi, C.; Binda, M.; Valoti, E.
Tetrahedron: Asymmetry 2007, 18, 1038–1041.
21. Bolchi, C.; Fumagalli, L.; Moroni, B.; Pallavicini, M.; Valoti, E. Tetrahedron:
Asymmetry 2003, 14, 3779–3785.
22. Bolchi, C.; Pallavicini, M.; Fumagalli, L.; Marchini, N.; Moroni, B.; Rusconi, C.;
Valoti, E. Tetrahedron: Asymmetry 2005, 16, 1639–1643.
23. Koo, J.; Avakian, S.; Martin, G. J. J. Am. Chem. Soc. 1955, 77, 5373–5375.
24. Tomiyama, T.; Wakabayashi, S.; Yokota, M. J. Med. Chem. 1989, 32, 1988–1996.
25. Fogassy, E.; Faigl, F.; Ács, M.; Simon, K.; Kozsda, É.; Podányi, B.; Czugler, M.;
Reck, G. J. Chem. Soc., Perkin Trans. 2 1988, 1385–1392.
26. Kinbara, K.; Sakai, K.; Hashimoto, Y.; Nohira, H.; Saigo, K. J. Chem. Soc., Perkin
Trans. 2 1996, 2615–2622.
27. Pallavicini, M.; Valoti, E.; Villa, L.; Piccolo, O.; Marchetti, E. Tetrahedron:
Asymmetry 2000, 11, 1957–1964.
28. Bálint, J.; Egri, G.; Czugler, M.; Schindler, J.; Kiss, V.; Juvancz, Z.; Fogassy, E.
Tetrahedron: Asymmetry 2001, 12, 1511–1518.
1. Braeuer, N.; Buchmann, B.; Klar, U.; Ter Laak, A.; Lindenthal, B.; Langer, G.;
Rotgeri, A. 2,5-Disubstituted 2H-indazole as EP2 receptor antagonists and their
preparation and use in the treatment of diseases associated with EP2 receptor
activity. DE 102009049662 A1, 2011.
2. Bolchi, C.; Pallavicini, M.; Fumagalli, L.; Ferri, N.; Corsini, A.; Rusconi, C.; Valoti,
E. Bioorg. Med. Chem. Lett. 2009, 19, 5500–5504.
3. Birch, A. M.; Bradley, P. A.; Gill, J. C.; Kerrigan, F.; Needham, P. L. J. Med. Chem.
1999, 42, 3342–3355.
4. Catt, J. D.; Mattson, R. J. (Aminomethyl)benzodioxanes and –benzopyrans as
serotonergic receptor agonists. US 5391570 A, 1995.
5. Xu, W.; Zheng, W.; Baly, D. L.; Galan, A. A.; Ibrahim, M. A.; Jaeger, C.; Kearney,
P.; Leahy, J. W.; Lewis, G. L.; McMillan, K. et al. TAO kinase inhibitors for
pharmaceutical use and for screening for kinase modulators. WO 2005040355
A2, 2005.
6. Mayer, P.; Imbert, T.; Marien, M. New derivatives of 8-([1,4]-benzodioxan-2-
ylmethyl)-8-azabicyclo[3.2.1]octane-3-alkyl ureas or imidazolidinones,
methods for their preparation, and their therapeutic applications for treating
neurodegenerative diseases. FR 2789681 A1, 2000.
7. Bovet, D.; Simon, A. Bull. Sci. Pharm. 1935, 42, 466–473.
8. Stenlake, J. B.; Breimer, D. D.; Smail, G. A.; Stafford, A.; Bowman, W. C. J. Pharm.
Pharm. 1968, 20, 82S–93S.
29. Marchini, N.; Bombieri, G.; Artali, R.; Bolchi, C.; Pallavicini, M.; Valoti, E.
Tetrahedron: Asymmetry 2005, 16, 2099–2106.
9. Nelson, W. L.; Powell, M. L. J. Med. Chem. 1979, 22, 1125–1127.
10. Bolchi, C.; Catalano, P.; Fumagalli, L.; Gobbi, M.; Pallavicini, M.; Pedretti, A.;
Villa, L.; Vistoli, G.; Valoti, E. Bioorg. Med. Chem. 2004, 12, 4937–4951.
11. Fumagalli, L.; Bolchi, C.; Colleoni, S.; Gobbi, M.; Moroni, B.; Pallavicini, M.;
Pedretti, A.; Villa, L.; Vistoli, G.; Valoti, E. Bioorg. Med. Chem. 2005, 13, 2547–
2559.
30. Kinbara, K.; Sakai, K.; Hashimoto, Y.; Nohira, H.; Saigo, K. Tetrahedron:
Asymmetry 1996, 7, 1539–1542.
31. Vries, T.; Wynberg, H.; van Echten, E.; Koek, J.; ten Hoeve, W.; Kellogg, R. M.;
Broxterman, Q. B.; Minnaard, A.; Kaptein, B.; van der Sluis, S.; Hulshof, L.;
Kooistra, J. Angew. Chem., Int. Ed. 1998, 37, 2349–2354.
12. Pallavicini, M.; Fumagalli, L.; Gobbi, M.; Bolchi, C.; Colleoni, S.; Moroni, B.;
Pedretti, A.; Rusconi, C.; Vistoli, G.; Valoti, E. Eur. J. Med. Chem. 2006, 41, 1025–
1040.