Absolute configuration and biological profile of two thiazinooxadiazol-3- ones with L-type calcium channel activity: A study of the structural effects

Article abstract of DOI:10.1039/c2ob25946j

In the framework of our interest in racemic thiazinooxadiazol-3-ones we determined the absolute configuration and the biological activity as L-type calcium channel blockers of two compounds that differ in the length of the acetal chain, which could affect the pharmacological profile. We observed an interesting inversion of the stereoselectivity, with the activity residing on the R-form for a short chain compound (n = 1) and on the S-form for a long chain one (n = 12). The length of the linear acetal chain appears to be able to invert the stereoselectivity of such a class of compounds, and in silico simulations suggested that this different behaviour might be explained by different hydrophilic and hydrophobic interactions with the binding site.

Full text of DOI:10.1039/c2ob25946j

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 8994
www.rsc.org/obc
PAPER
Absolute configuration and biological profile of two thiazinooxadiazol-3-ones
with L-type calcium channel activity: a study of the structural effects
Pierfranco Ioan,a Alessia Ciogli,b Francesco Sirci,c Roberta Budriesi,a Barbara Cosimelli,d Marco Pierini,b
Elda Severi,d Alberto Chiarini,a Gabriele Cruciani,c Francesco Gasparrini,b Domenico Spinellie and
Emanuele Carosati*c
Received 16th May 2012, Accepted 1st October 2012
DOI: 10.1039/c2ob25946j
In the framework of our interest in racemic thiazinooxadiazol-3-ones we determined the absolute
configuration and the biological activity as L-type calcium channel blockers of two compounds that differ
in the length of the acetal chain, which could affect the pharmacological profile. We observed an
interesting inversion of the stereoselectivity, with the activity residing on the R-form for a short chain
compound (n = 1) and on the S-form for a long chain one (n = 12). The length of the linear acetal chain
appears to be able to invert the stereoselectivity of such a class of compounds, and in silico simulations
suggested that this different behaviour might be explained by different hydrophilic and hydrophobic
interactions with the binding site.
The stereoselectivity of several ligand–channel interactions is
Introduction
well established, including DHPs, BTZs and PAAs interacting
with LTCCs;7,8 as an example, the activity of DTZ resides
L-Type calcium channel (LTCC) blockers are drugs approved for
the treatment of cardiovascular diseases in the 1980s.1 Their use
almost entirely in one (2S,3S) out of its four possible diastereo-
isomers.9,10 In this framework, our research group developed
represents a largely diffused system for monitoring the mamma-
lian cardiac activity,2 and has been extended to additional con-
myocardial calcium channel modulators starting from the mo-
ditions such as the Raynaud phenomenon, pulmonary
hypertension, diffuse esophageal spasm, and cerebral vaso-
spasm3 and as a coadjuvant in anticancer therapy, i.e. as inhibi-
tors of multi-drug resistance (MDR) activity.4,5
lecular resemblance of the chemical structure of 8-aryl-8H[1,4]-
thiazino[3,4-c][1,2,4]oxadiazol-3-one derivatives with that of
DTZ.11 Incidentally, some of these compounds were also prom-
ising inhibitors of MDR1 activity.12
The prototypical agents of LTCC blockers are: nifedipine
(NIF), a member of the 1,4-dihydropyridine class (DHP); diltia-
zem (DTZ), a 1,5-benzothiazepinone of the benzothiazepine
These molecules possess a chiral atom, central with respect to
the molecular structure, that might affect ligand–channel inter-
actions. The enantiomeric separation of the racemic mixture was
recently carried out,13 followed by absolute configuration deter-
mination and analysis of biological activity of the two enantio-
mers of the most potent compound of a series of 8-aryl-8-
hydroxy-5-methyl-8H[1,4]thiazino[3,4-c][1,2,4]oxadiazol-3-
ones and the corresponding 8-alkoxy derivatives, namely the
8-ethoxy derivative 1. The R(−)-enantiomer was proven to
be responsible for the decrease of the force of cardiac
muscle contraction, showing nanomolar negative inotropic
potency.13
class (BTZ); and verapamil (VER),
a member of the
phenylalkylamine class (PAA). They interact at discrete receptor
sites associated at the main α1 subunit of the channel and their
pharmacological effects are due to the modulation of calcium
entry through the LTCC.6
aDipartimento di Scienze Farmaceutiche, Università degli Studi di
Bologna, Via Belmeloro 6, 40126 Bologna, Italy
bDipartimento di Chimica e Tecnologie del Farmaco, Università
“La Sapienza”, Piazzale A. Moro 5, 00185 Roma, Italy
cDipartimento di Chimica, Università degli Studi di Perugia, via Elce di
Sotto 10, 06123 Perugia, Italy.
Binding experiments with [3H]diltiazem revealed a displace-
ment of the marked probe at significantly high concentrations of
1 (10−4 M) whereas it was potentiating DTZ binding at low con-
centrations (10−9–10−6 M). These findings indicated that these
two compounds share the same binding site even if limited to
high concentrations of 1.
E-mail: emanuele@chemiome.chm.unipg.it; Fax: +39 075 45646;
Tel: +39 075 585 5550
dDipartimento di Chimica Farmaceutica e Tossicologica, Università
degli Studi di Napoli “Federico II”, Via Montesano 49, 80131 Napoli,
Italy
eDipartimento di Chimica ‘G. Ciamician’, Università di Bologna, Via
Selmi 2, 40126 Bologna, Italy
One hypothesis of such behaviour might be the stereoselective
interaction with the channel, already known for DTZ.9 To gain
8994 | Org. Biomol. Chem., 2012, 10, 8994–9003
This journal is © The Royal Society of Chemistry 2012

View Article Online
further insight into this ligand–channel interaction we examined
the thiazinooxadiazol-3-ones 2 and 3 (Chart 1), already studied
as racemic mixtures.11,14
Results and discussion
For both compounds 2 and 3 we first carried out the enantio-
meric separation by using chiral chromatography, then assigned
the absolute configuration by simulation of CD spectra, and
finally determined the biological profile of the two enantiomers.
In silico molecular superimpositions of the studied compounds
over DTZ completed the stereoselective analysis of the ligand–
channel interaction.
These two compounds were selected among those previously
investigated as racemic mixtures,11,14 and chemical and biologi-
cal characteristics of the racemates guided the selection. Chemi-
cally, the nature of the substituent in the aryl moiety, the length
of the acetal chain, and the calculated log P (2.27, 1.61, and
7.18 for 1, 2, and 3, respectively) were considered: compounds 2
and 3 differ for the atom at the para-position of the aryl chain,
chlorine for 2 and bromine for 3; the acetal chain was of differ-
ent length, ranging from one carbon atom for 2 (R = CH3) to
twelve atoms for 3 (R = n-C12H25), thus affecting the log P as
well. Biologically, both chlorine for 2 and bromine for 3 (but
with a particular prevalence of bromine) favourably affect
the cardiovascular profile. In particular, 2 showed a negative ino-
tropic potency two-times lower than DTZ whereas 3 six-times
higher than DTZ (Table 1).14 Notably, 3 was the most
potent among the studied acetals with a long linear chain,14
thus raising the hypothesis of a key role played by the acetal
chain length.
Conformational analysis and calculations of the electronic
circular dichroism (ECD)
The pure enantiomers of 2 and 3 were obtained by semiprepara-
tive HPLC. Since the structures of 2 and 3 are strictly related to
that of 1, for which the absolute configuration was successfully
elucidated through suitable comparison between experimental
and theoretically calculated Electronic Circular Dichroism
(ECD) and Optical Rotatory Dispersion (ORD) spectra,13,15 we
expected a very similar trend for relevant pseudoscalar properties
such as optical rotation, ORD and ECD spectra. Nevertheless,
we could not exclude a priori conformational differences and, in
turn, different chiroptical properties when changing the 8-ethoxy
group (1) with 8-methoxy (2) or, especially, with the signifi-
cantly longer 8-dodecyloxy (3) ones.
Thus, in order to assign the absolute configuration to the first
and second chromatographically eluted enantiomers of 2 and 3
by an unambiguous procedure, we carried out a pertinent esti-
mation of their pseudoscalar properties without any aprioristic
assumption. Fig. 1 reports the resolved peaks of 2 (a) and 3 (b),
obtained by enantioselective HPLC.
Then, structures of 2 and 3 were modelled using a three-step
approach. First, exhaustive conformational searches of the rele-
vant geometries were performed through systematic (compound
2) and stochastic MonteCarlo (compound 3) algorithms (see the
Experimental section), whose structure optimization methods
Chart 1 Structures of diltiazem (DTZ), R-1, and the two new thiazino-
oxadiazol-3-ones 2 and 3.
Table 1 Cardiovascular activity of tested compounds
% Decrease (M SEM)
EC50 of inotropic negative activity
Vasorelaxant activity
Comp
Negative inotropic activitya
Negative chronotropic activityb
94 5.6f,g
EC50 c (μM)
95% conf lim (×10−6
)
Activityd (M SEM)
DTZ
1j
78 3.5e
77 1.7f
58 3.4f
47 1.4f
62 2.5
71 0.7
48 1.2
59 1.2e
43 2.1
67 0.8f
0.79
0.04
0.07
0.70–0.85
0.03–0.05
0.04–0.08
88 2.3h,i
19 0.9
5
0.2e
R(−)-1k
S(+)-1k
2l
16 0.8
15 0.9
22 1.1
15 1.3
12 0.9
16 0.9
11 0.7h
7
0.3h
1.54
0.80
1.21–1.88
0.53–1.01
14 0.9
35 2.3
21 1.3
R(−)-2
S(+)-2
3j
R(−)-3
S(+)-3
0.13
0.09–0.15
3
5
4
0.2
3
0.2
0.4m
0.1m
10 0.7
0.014
0.010–0.020
a Activity: decrease on developed tension in isolated guinea-pig left atrium at 5 × 10−5 M, expressed as percent changes from the control (n = 4–6).
The left atria were driven at 1 Hz. The 5 × 10−5 M concentration gave the maximum effect for most compounds. b Activity: decrease on atrial rate in
guinea-pig spontaneously beating isolated right atria at 5 × 10−5 M, expressed as percent changes from the control (n = 6–8). Pretreatment heart rate
ranged from 170 to 195 beats per min. The 5 × 10−5 M concentration gave the maximum effect for most compounds. c Calculated from log
concentration–response curves (Probit analysis according to Litchfield and Wilcoxon17 with n = 6–8). When the maximum effect was <50%, the EC50
inotropic, EC30 chronotropic and IC50 values were not calculated. d Activity: percent inhibition of calcium-induced contraction on K+-depolarized
f
guinea-pig aortic strips at 5 × 10−5 M. The 5 × 10−5 M concentration gave the maximum effect for most compounds. e At 10−5 M. At 10−6 M.
g EC30 = 0.07 μM (c.l. 0.064–0.075). h At 10−4 M. i IC50 = 2.6 μM (c.l. 2.2–3.1). j From ref. 14. k From ref. 13. l From ref. 11. m At 5 × 10−6 M.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 8994–9003 | 8995

View Article Online
Fig. 1 Enantioselective HPLC of compounds 2 (a) and 3 (b). Chiral-
pak-IB column (250 × 4.6 mm I.D.), n-hexane–isopropanol 95 : 5, flow
1.00 mL min−1, UV at 270 nm were used for 2 and (S,S)-Poly-brush-
DACH-ACR column (250 × 4.6 mm I.D.), n-hexane–dichloromethane
85 : 15, flow 1.00 mL min−1 UV at 254 nm were employed for 3. Dis-
played ECD spectra are recorded online.
Fig. 2 Optimized conformations of compounds 2 and 3. Structures
2_1–2_3 represent the only three geometries found by systematic confor-
mational search of 2, whereas structures 3_1–3_4 represent the most
stable geometries within each of the four ensembles of structures found
by stochastic conformational search of 3.
were based on molecular mechanics calculations (force field:
MMFF).
The comparison of experimental and calculated ECD spectra
allowed us to attribute the (S)-configuration to the first eluted
enantiomer of both 2 and 3. Similarly, in both cases the compari-
son of experimental and calculated ORD allowed us to assign
the (S)-configuration to the dextrorotatory enantiomers (Fig. 3).
Secondly, the obtained conformations were ranked according
to their Boltzmann population. Given the low flexibility of the
methoxy group, only three conformations were used for 2,
whereas thirty conformations, within an energy window of
2 kcal mol−1, were used for 3. These conformations were
grouped on the basis of the orientation assumed by the alkoxy
group at position 8 with respect to both the phenyl and bicyclic
ring moieties (Fig. 2).
According to the Boltzmann populations, three conformers of
2 and four among the thirty selected conformations of 3 rep-
resented a global population of 99% of the whole population of
2 and 3, respectively. However, in order to select three confor-
mations also for 3, the most representative geometries (confor-
mations 3_1, 3_2 and 3_3) were considered, which express 95%
of the whole population.
As a third step, for these six conformers first the geometry
was further optimized through the B3LYP/6-31G* DFT method,
then the relevant chiroptical properties were calculated. In par-
ticular, both ECD and ORD plots were derived for each of the
six optimized geometries and the final graphs, shown in Fig. 3,
were achieved by merging the data plot relevant to the single
conformations of 2 or 3 according to their corresponding Boltz-
mann populations.
Functional results
The functional behaviour of the two enantiomers of 2 and 3 was
evaluated according to a well-established protocol, already pub-
lished elsewhere16 and reported in the Experimental section.
Overall, negative inotropic activity and negative chronotropic
activity were evaluated on guinea-pig isolated left atria driven at
1 Hz and on spontaneously beating right atria, respectively
(Table 1). In addition, calcium antagonistic activity was tested on
K+-depolarized (80 mM) guinea-pig aortic strips (Table 1) and
guinea-pig ileum longitudinal smooth muscle (GPILSM)
(Table 2) to assess relaxant activity on vascular and non-vascular
smooth muscle, respectively.
The overall pharmacological profile of 2 and 3 was not sur-
prising since, in line with the previous experiments conducted
on the racemic mixtures, no significant vasorelaxant activity was
observed for the four enantiomers (two of 2 and two of 3)
whereas a weak relaxant activity on GPILSM was observed
8996 | Org. Biomol. Chem., 2012, 10, 8994–9003
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 2 Relaxant activity on K+-depolarized guinea-pig ileum
longitudinal smooth muscle
Comp
Activitya (M SEM) IC50 b (μM) 95% conf lim (×10−6
)
DTZ
1d
2
R(−)-2 68 2.3
S(+)-2
98 1.5c
0.2
61 1.3
0.11
0.085–0.13
3
15.68
13.53
16.32
11.81–19.83
10.07–15.88
14.03–19.24
65 1.4
22 0.3
3
a Percent inhibition of calcium-induced contraction on 80 mM K+-
depolarized guinea-pig longitudinal smooth muscle at 10−4 M. The
10−4 M concentration gave the maximum effect for most compounds.
b Calculated from log concentration–response curves (Probit analysis
according to Litchfield and Wilcoxon17 with
n = 6–8). When
the maximum effect was <50%, the IC50 values were not calculated.
c At 10−6 M. d From ref. 13.
difference was observed for the negative inotropic activity values
of the enantiomers of 2 as well as for those of 3. In the case of
the short methoxy group, at the same concentration (5 × 10−5 M)
the second eluted enantiomer, namely R-2, shows an intrinsic
negative inotropic activity of 71 0.7, higher than that of S-2,
which was 48 1.2. Even more marked was the difference in the
intrinsic activity of the two enantiomers of 3, for which the
value was 67 0.8 at 10−6 M for the first eluted enantiomer
(S-3), and only 43 2.1 at 5 × 10−5 M for the second eluted
enantiomer, namely R-3.
The most intriguing finding presented here is that the two mo-
lecules 2 and 3, when considering the negative inotropic activity
and potency, showed opposite stereoselective behaviour, with the
R-form of 2 and the S-form of 3 being the eutomers. The behav-
ior of 2, despite its lower potency, is in line with that of the
slightly flexible 1, for which the enantioselective activity has
been previously reported.13,15 In both cases the potency has
increased approximately 2-times in the more potent enantiomer.
Instead, an inversion of selectivity occurs when passing from the
short alkyl chains (methoxy or ethoxy) to the long, more
flexible, chain of 3 (dodecyloxy). In this case, besides the dis-
placement of activity there is also an increase of potency, with
the enantiomer about 10-times more potent than the racemic
mixture.
Thus, the use of computational methods, already used to
hypothesize the binding mode of other calcium channel antagon-
ists with similar or different scaffolds,18 was invoked to shed
some light on the hypothetical binding mode of these com-
pounds. As mentioned in the introduction, this family of thiazi-
nooxadiazolon-3-ones with calcium antagonist activity was
developed bearing in mind a structural similarity with DTZ,
which was first hypothesized14 and then proved with binding
experiments.13 The intriguing results reported here, i.e. the
inverted stereoselectivity, prompted us to look for alternative
methods to explain the reasons for such different behaviour. One
of the possibilities available so far is the use of a common refer-
ence framework, that would be a protein binding pocket in cases
where the target three-dimensional structure is available or, as in
the case of LTCC blockers, the molecular structure of a com-
pound that shares the binding site with the molecules to be
compared.
Fig. 3 Experimental and calculated ECD and ORD plots of first eluted
and (S) enantiomers of compounds 2 (a) and 3 (b).
limited to 2. Interestingly, for this compound activity and
potency of the two enantiomers were not significantly different
from those of the racemate, indicating that the non-vascular
smooth muscle activity and (weak) potency are not influenced
by the stereochemistry of the molecule. Conversely, a marked
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 8994–9003 | 8997

View Article Online
Table 3 Similarity scores obtained by FLAP when using DTZ as
the template and 1, 2, and 3 as ligand-molecules (a), or DTZ as the
ligand-molecule and 1, 2, and 3 as templates (b). The asymmetry
of the scores is due to the asymmetry of the FLAP scoring
functions. Symbols “∧” and “∨” are used to easily identify the
enantiomers with higher scores, the percentage values are the
difference between the two scores divided by the score with the
higher value, whereas colored cells are used to highlight the most
relevant scores: white = no relevance; light-gray = low relevance;
dark-gray = high relevance. In other words, light-gray is used to
mark those cells in which the difference between the scores of the
two enantiomers is larger than 10%, whereas dark-gray those for
which the difference is larger than 15%
Molecular superimpositions and binding mode hypotheses
Compounds 1, 2 and 3 were imported in the software FLAP19
and superimposed over DTZ, which was considered as a molecu-
lar template. In this phase the choice of the conformations is
aimed at determining a set of energetically stable solutions
which are also diverse from each other. With respect to the pre-
viously used method (geometry optimization for calculation of
ECD spectra), the method used here is based on an in-house
derivation of the MM3 force-field.20
This procedure, detailed in the Experimental section, was
applied to compounds 1, 2 and 3; thus, using 50 as the
maximum number of diverse conformers selectable by FLAP, 20
conformers were produced for R-1 as well as for S-1, 12 for 2
(both R- and S-forms), and 25 for 3 (both R- and S-forms). Mo-
lecular flexibility was considered also for the template, and for
DTZ the same analysis was carried out, and 10 conformations
were produced by FLAP.
Once imported, for each molecule the Molecular Interaction
Fields (MIF) were generated with the GRID probes DRY, OH2
and H.21 While the probe H simply refers to the molecular
shape, the DRY and OH2 probes refer to those hydrophobic and
hydrophilic regions around the molecule where favorable inter-
actions are likely to occur with the hydrophobic and hydrophilic
atoms and groups of the protein binding site, respectively. With
FLAP the molecular similarity is calculated, once two molecules
are superimposed, by assessing the degree of overlapping
between their corresponding MIF (details are given in the Exper-
imental section). The resulting similarity scores for the studied
molecules are reported in Table 3a; these scores are in the range
0–1, and the higher the score the more similar are the two mo-
lecules. The selected orientations (discussed below and reported
in Fig. 4) are those with the highest overall MIF-similarity
toward DTZ (named ‘Global’), obtained when superimposing
the two molecules (over DTZ) using FLAP.
Fig. 5 reports the MIF for DTZ (a) and the two enantiomers
of 1 (b,c), 2 (d,e) and 3 (f,g). Since the MIF of the three mo-
lecules have the same energy value, comparisons are allowed
in terms of extension and relative position of such MIF points.
DTZ has a wide hydrophilic region, which is extended from
the amine nitrogen to the carbonyl of the benzothiazepine ring,
till the ester group. Of course, part of this region (due to N) is
subjected to the flexibility of the ethylene chain, but a major
change would occur in the case of inversion of configuration of
one of the two carbon atoms. Nevertheless, since we used only
the (2S,3S) configuration of DTZ we focus here on the inter-
molecular comparisons, that is comparing DTZ with the two
enantiomers of 1, 2, or 3.
R-1 has a smaller (if compared to DTZ) blue-coded hydrophi-
lic region; S-1 has a hydrophilic region of the same volume of
R-1, but with the orientation as in Fig. 4 its MIF only partially
overlaps with those of DTZ. The DRY MIF of S-1 and R-1
overlap with those of DTZ to a similar extent. Thus, the most rel-
evant difference between the two enantiomers of 1 is their mo-
lecular shape (highlighted in Table 3 as dark-gray (high
relevance)), represented with the surface in Fig. 5. The S-form
has a better fit over DTZ than the R-form; however, such a better
fit has no direct relationship with the inotropic activity, which
resides only on the R-form.
a RMSD = 0.67. b RMSD = 0.48. c RMSD = 0.11. d RMSD = 1.20.
e RMSD = 1.54. f RMSD = 1.76.
A similar analysis can be derived for the two enantiomers of 2
(Table 3a, Fig. 4c, d and 5d, e). The R-form has a higher hydro-
philic similarity score, that could be responsible for the selective
negative inotropic activity, whereas the S-form has a better
shape-fit over DTZ, that nevertheless is not related to the activity.
Both scores are dark-gray colored in Table 3. These findings let
us suppose that, limited to 1 and 2, different driving forces
might guide the two enantiomers within the DTZ binding
pocket, mostly hydrophilic-based in the case of R and mostly
shape-based in the case of S, with therefore hydrophilicity as
responsible for the stereoselectivity.
When repeating the analysis for 3 (Table 3a, Fig. 4e, f and 5f,
g), the R-form reflects the same orientation of the previous two
molecules, with the alkyl chain partially superimposed on the
amine-ethylene chain of DTZ, but with a large part (at least 7 or
8 carbon atoms) completely free to move anywhere, and not
8998 | Org. Biomol. Chem., 2012, 10, 8994–9003
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 4 Molecular superimpositions of both the enantiomers of 1, 2,
and 3 over DTZ obtained using FLAP. Oxygen atoms are red, nitrogen
atoms are blue, bromine atoms (a, b, e, f) are purple, chlorine atoms
(c, d) are light-green. Carbon atoms of DTZ are green-colored, those of
the R-form of the three molecules (1, 2, and 3) are cyan-colored,
whereas those of the S-form are magenta-colored. Hydrogen atoms are
not reported for clarity.
superimposed at all over the template. Instead, the orientation of
S-3 is characterized by the folding of the alkyl chain in corres-
pondence of the p-methoxy-phenyl moiety of DTZ. These orien-
tations, in terms of scores, give rise to an even larger hydrophilic
score for R-3 (dark-gray), a higher hydrophobic score for S-3
(light-gray) and a reduced difference in terms of shape (H) for
the two enantiomers (white-colored, in contrast to 1 and 2, that
are dark-gray).
Some of these findings might be related to the observed
stereoselectivity; we believe that the different orientation of S-3
with respect to those of S-1 and S-2 might represent the reason
why the long chain causes the inversion of the stereoselectivity.
To check this hypothesis, the ligand-based superimposition pro-
cedure was repeated, but using the six orientations of the enan-
tiomers of 1, 2, and 3 as templates and the selected orientation
of DTZ as the ligand molecule (without conformers generation).
The collected similarity scores are reported in section b of
Table 3.
Interestingly, the orientations of the ligand–template pairs
obtained in this case were very close to those previously
obtained, with an average root-mean-square deviation (RMSD)
of 0.96, ranging from 0.11 of R-2 to 1.76 of S-3. Table 3b
reports the similarity scores calculated with this novel approach.
Most of the findings previously observed were still valid, with
Fig. 5 MIF obtained with the GRID probes H, OH2 and DRY for
DTZ (a) and the two enantiomers of 1 (b: R-1; c: S-1), 2 (d: R-2; e: S-2),
and 3 (f: R-3; g: S-3). Oxygen atoms are red, nitrogen atoms are blue,
bromine atoms are purple, chlorine atoms (c, d) are light-green. Carbon
atoms of DTZ are green-colored, those of R-1, R-2 and R-3 are cyan-
colored, whereas those of S-1, S-2 and S-3 are magenta-colored. Hydro-
gen atoms are not reported for clarity. The colored surface represents the
isocontour corresponding to the energy values of the MIF set to
+0.5 kcal mol−1 for the H probe. The yellow-coded spheres represent
grid points where the calculated energy with the DRY probe was
−0.8 kcal mol−1 or lower, whereas the blue spheres represent grid points
where the calculated energy with the OH2 probe was −4.5 kcal mol−1 or
lower. In the GRID force-field such a value represents, for example, the
energy of the interaction between a water molecule, donating a hydro-
gen, with a potent heteroatom, involved as a donor, for an optimally
oriented hydrogen bond interaction.
the hydrophilic fields likely responsible for
a
better
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 8994–9003 | 8999

View Article Online
superimposition of DTZ over the R-forms of 1 and 2 (with
respect to their corresponding S-forms). Concerning the simi-
larity scores of DTZ over the two enantiomers of 3, the major
finding was a higher value observed for DTZ with S-3 rather
than R-3 for both the hydrophilic and hydrophobic fields.
(flow rate 5.00 mL min−1 and T 25 °C), UV (λ = 270 nm) and
RI detections. Sample 2 was dissolved in the mobile phase
(35 mg, c ≈ 15 mg mL−1) and each injection was loaded with
200 μL (process yield ≈ 80%). The enantiomeric excesses were
determined using the analytical Chiralpak-IB column (250 ×
4.6 mm I.D.), the same mobile phase employed for the semipre-
parative mode at flow rate 1.00 mL min−1 and the UV/CD detec-
tion at 270 nm. The chromatographic steps afforded 2-first
eluted (13 mg, k = 1.78, e.e. 99.9%) and 2-second eluted
(14.5 mg, k = 2.10, α = 1.18, e.e. 96.0%) isolated samples.
The pure enantiomers of 3 were obtained by semipreparative
HPLC using an (S,S)-Poly-brush-DACH-ACR22 column (250 ×
10 mm I.D.), n-hexane–dichloromethane 90 : 10 as the eluent
(flow rate 5.00 mL min−1 and T 25 °C), UV (λ = 254 nm) and
RI detections. 100 μL multiple injections of racemate (18 mg,
c ≈ 30 mg mL−1 in the mobile phase) afforded pure enantio-
mers: 7.7 mg of 3-first eluted (k = 1.70, e.e. 97.0%) and 5.5 mg
of 3-second eluted (k = 1.99, α = 1.17, e.e. 99.0%). The enan-
tiomeric excesses were determined using the analytical (S,S)-
Poly-brush-DACH-ACR column (250 × 4.6 mm I.D.), n-
hexane–dichloromethane 85 : 15 at 1.00 mL min−1 and the UV/
CD detection at 254 nm.
Conclusions
In this study the stereoselectivity of two new thiazinooxadiazol-
3-ones with calcium antagonist activity has been studied. Enan-
tiomeric separation using chiral chromatography, ECD spectra to
assign the absolute configuration, functional studies on several
guinea-pig tissues, and molecular superimpositions over DTZ
have been combined to study the stereoselective behavior of the
two compounds.
Experimentally, an inversion on the stereoselectivity was
observed for compounds which mainly differ in the length of
their alkoxy chain: the activity was residing on the R-form for
the short chain and on the S-form for the long chain. Our
hypothesis, based also on the fact that the thiazinooxadiazol-3-
ones and DTZ share the same binding site at the LTCC, is that
the hydrophilic interactions are the main driving force of the
activity and, in turn, of the stereoselectivity.
The stereoselective behaviour of 1, already described else-
where,13,15 has been confirmed by a similar behaviour observed
for 2, that differs from 1 for a single carbon atom (methoxy/
ethoxy). The longer alkyl chain was supposed to be responsible
for the inversion of the selectivity, likely due to a better fit of the
chain of the S-form in a hypothetical hydrophobic subpocket of
the LTCC binding site where the p-methoxy-phenyl moiety of
DTZ is likely to bind.
The online ECD and UV spectra, ranging from 220–420 nm,
were recorded for each enantiomer of 2 and 3 samples.
ECD measurements. ECD measurements were taken on a
JASCO J-710 CD spectrometer, flushed with dry nitrogen using
10 mm path length quartz cuvettes. Measurements were taken on
both enantiomers of 2 in dichloromethane solutions (6.18 ×
10−5 M for 2-first eluted and 7.60 × 10−5 M for 2-second
eluted) and on 3-second eluted enantiomers in a 7.5 × 10−5
M
dichloromethane solution; spectra were solvent subtracted.
In summary, molecular shape and hydrophilic interactions are
the key factors for the inotropic activity and for the selectivity of
this set of thiazinooxadiazol-3-ones.
Optical rotations measurements. Optical rotations measure-
ments were taken on a JASCO P-1000 series polarimeter using
577, 546, 435, 405, 365 nm filters. The same enantiomeric solu-
tions for ECD offline measurements were used. The measured
values of optical rotation are given below.
Experimental
A. Chemistry
Compound 2, first eluted enantiomer: 577 nm, [α] = 645;
546 nm, [α] = 688; 435 nm, [α] = 844; 405 nm, [α] = 1207;
365 nm, [α] = 2050.
8-Aryl-8-alkoxy-5-methyl-8H[1,4]thiazino[3,4-c][1,2,4]oxadia-
zol-3-ones 2 and 3 were synthesized as previously reported.11,14
Compound 3, first eluted enantiomer: 577 nm, [α] = 105;
546 nm, [α] = 281; 435 nm, [α] = 551; 405 nm, [α] = 936;
365 nm, [α] = 1548.
B. Enantiomers separation
Apparatus. Analytical liquid chromatography was performed
on a JASCO chromatograph equipped with a Rheodyne model
7725i 20 μL loop injector, a PU-980 HPLC pump, a single
wavelength absorbance detector Mod Jasco-975 and a circular
dichroism detector Jasco Mod 995-CD. Chromatographic data
were collected and processed using Borwin software (Jasco
Europe, Italy). Semi-preparative liquid chromatography was per-
formed on a Waters chromatograph (Waters Associates) equipped
with a Rheodyne model 7012 500 μL loop injector, a spectro-
photometer UV SpectroMonitor 4100 and a refractive index
Waters R401 detector.
C. Molecular modelling calculations
Molecular modeling optimizations of structures were performed
by using the computer program SPARTAN 04 (Wavefunction
Inc., 18401 Von Karman Avenue, Suite 370, Irvine, CA 92612,
USA). Conformational searches of compounds 2 and 3 were per-
formed by molecular mechanic calculations based on the MMFF
force field, according to the systematic (compound 2) or stochas-
tic MonteCarlo (compound 3) algorithms implemented in
SPARTAN. All rotatable bonds were varied.
Chromatographic procedures. The enantiomers of 2 were col-
lected by semipreparative HPLC using a Chiralpak-IB column
(250 × 10 mm I.D.), n-hexane–isopropanol 95 : 5 as the eluent
Maximum energy gap from the lowest energy geometry for
kept conformations: 40 kJ mol−1; criterion adopted in the analy-
sis of similarity to define conformers as duplicates: R2 ≥ 0.9.
9000 | Org. Biomol. Chem., 2012, 10, 8994–9003
This journal is © The Royal Society of Chemistry 2012

View Article Online
Calculations supplied only three geometries for 2 (2_1, 2_2
and 2_3), while thirty conformations by an energy window of
2 kcal mol−1 for 3.
evaluate the percentage decrease of developed tension on an iso-
lated left atrium driven at 1 Hz (negative inotropic activity), the
percentage decrease in atrial rate on a spontaneously beating
right atrium (negative chronotropic activity) and the percentage
inhibition of calcium-induced contraction on K+-depolarized
aortic strips (vasorelaxant activity) and GPILSM (non-vascular
smooth muscle relaxant activity).
Data were analyzed using Student’s t-test. The potency of
drugs defined as EC50, EC30 and IC50 was evaluated from log
concentration–response curves (Probit analysis using Litchfield
and Wilcoxon or GraphPad Prism® software)17,23 in the appro-
priate pharmacological preparations. All data are presented as
mean SEM.
Due to the high number of conformers found for compound 3,
the most relevant structures were clustered. The grouping was
based on disposition assumed by the dodecyloxy-framework
with respect to the bicyclic and phenyl moieties of the molecule.
From this kind of analysis four typologies of conformations were
identified, corresponding to the following geometric properties:
type (1) conformers with the hydrocarbon chain folded on the
left of the middle position of the molecule, over the bicyclic
framework (3_1 the most stable among these structures); type (2)
conformers with the hydrocarbon chain folded on the right of
the middle position of the molecule, over the phenyl moiety
(3_2 the most stable among these structures); type (3) confor-
mers with a stretched hydrocarbon chain pointing away from
both bicyclic and phenyl frameworks (3_3 the most stable
among these structures); type (4) conformers with the hydro-
carbon chain folded between the bicyclic and phenyl frame-
works, from both the possible sides (3_4 the most stable among
these structures).
All the three geometries achieved for compounds 2 were
further optimized at the SCF level by the DFT B3LYP/6-31G*
method. Instead, due to the little abundance assessed for confor-
mations type 4 of compound 3 (4% their cumulative Boltzmann
Population, BP, from the molecular mechanics calculations),
only the most stable ones of type 1–3 of this molecule (i.e.,
structures 3_1, 3_2 and 3_3) were submitted to the same DFT
optimization performed on 2. The found relative energy amounts
(Erel, in kcal mol−1) and BP at 25 °C of the 2_x conformations
(with x varying from 1 to 3) were: Erel 0.00, BP 60.9%; Erel
0.31, BP 36.3%; Erel 1.81, BP 2.9%. For the 3_x conformations
(with x varying from 1 to 3) the same quantities were: Erel 2.85,
BP 0.8%; Erel 1.79, BP 4.6%; Erel 0.00, BP 94.6%, respectively.
After this refinement structures were subjected to the assessment
of the relevant chiroptical properties, carried out using the BLYP
method with the QZ4P large core basis set for 2 and with the
TZ2P large core basis set for 3, as implemented in the Amster-
dam Density Functional (ADF) package v. 2007.01. The set
options were: single point calculation; 40 singlet and triplet exci-
tations; diagonalization method: Davidson; optical rotation: 6
frequencies, from 4000.0 to 6000.0 Å. The calculated values of
optical rotation are given below.
Guinea-pig
atrial
preparations. Female
guinea-pigs
(300–400 g) were sacrificed by cervical dislocation. After thora-
cotomy the heart was immediately removed and washed by per-
fusion through the aorta with an oxygenated Tyrode solution of
the following composition (mM): 136.9 NaCl, 5.4 KCl, 2.5
CaCl2, 1.0 MgCl2, 0.4 NaH2PO4·H2O, 11.9 NaHCO3 and
5.5 glucose. The physiological salt solution (PSS) was buffered
at pH 7.4 by saturation with 95% O2–5% CO2 gas, and the temp-
erature was maintained at 35 °C. The following isolated guinea-
pig heart preparations were used: spontaneously beating right
atria and left atria driven at 1 Hz. For each preparation, the entire
left and right atria were dissected from the ventricles, cleaned of
excess tissue, hung vertically in a 15 mL organ bath containing
the PSS continuously bubbled with 95% O2–5% CO2 gas at
35 °C, pH 7.4. The contractile activity was recorded isometri-
cally by means of a force transducer (FT 0.3, Grass Instruments
Corporation, Quincy, MA, USA) using Power Lab® software
(AD-Instruments Pty Ltd, Castle Hill, Australia). The left atria
were stimulated by rectangular pulses of 0.6–0.8 ms duration
and about 50% threshold voltage through two platinum contact
electrodes in the lower holding clamp (Grass S88 Stimulator).
The right atrium was in spontaneous activity. After the tissue
was beating for several minutes, a length–tension curve was
obtained and the muscle was stretched to the length at which
90% of maximal force was developed. A stabilization period of
45–60 min was allowed before the atria were challenged by
various agents. During the equilibration period, the bathing sol-
ution was changed every 15 min and the threshold voltage was
ascertained for the left atria. Atrial muscle preparations were
used to examine the inotropic and chronotropic activity of the
compounds (0.01, 0.05, 0.1, 0.5, 1, 5, 10, 50 and 100 μM), first
dissolved in DMSO and then diluted with PSS. According to
this procedure, the concentration of DMSO in the bath solution
never exceeded 0.3%, a concentration which did not produce
appreciable inotropic and chronotropic effects. During the gener-
ation of cumulative concentration–response curves, the next
higher concentration of the compounds was added only after the
preparation reached a steady state.
Compound 2, first eluted enantiomer: 600 nm, [α] = 726;
560 nm, [α] = 900; 520 nm, [α] = 1155; 480 nm, [α] = 1564;
440 nm, [α] = 2307; 400 nm, [α] = 4041.
Compound 3, first eluted enantiomer: 600 nm, [α] = 67;
560 nm, [α] = 95; 520 nm, [α] = 149; 480 nm, [α] = 253;
440 nm, [α] = 485; 400 nm, [α] = 1188.
D. Functional assays
The pharmacological profile of all compounds was tested on
guinea-pig isolated left and right atria to evaluate their inotropic
and chronotropic effects, respectively, and on K+-depolarized
guinea-pig aortic strips and guinea-pig longitudinal smooth
muscle of ileum to assess calcium antagonist activity. Com-
pounds were checked at increasing concentrations in order to
Guinea-pig aortic strips and ileum longitudinal smooth
muscle (GPILSM). The thoracic aorta and ileum were removed
and placed in a Tyrode solution of the following composition
(mM): 118 NaCl, 4.75 KCl, 2.54 CaCl2, 1.20 MgSO4,
1.19 KH2PO4, 25 NaHCO3 and 11 glucose equilibrated with
95% O2–5% CO2 gas at pH 7.4. The vessel was cleaned of
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 8994–9003 | 9001

View Article Online
extraneous connective tissue. Two helicoidal strips (10 mm ×
1 mm) were cut from each aorta beginning from the end most
proximal to the heart. Vascular strips were then tied with surgical
thread (6-0) and suspended in a jacketed tissue bath (15 mL)
containing aerated PSS at 35 °C. Aortic strips were secured at
one end to plexiglass hooks and connected via the surgical
thread to a force displacement transducer (FT 0.3, Grass Instru-
ments Corporation) for monitoring changes in isometric contrac-
tion. Aortic strips were subjected to a resting force of 1 g. The
intestine was removed above the ileo-caecal junction. GPILSM
segments of 2 cm length were mounted under a resting tension
of 300–400 mg. Strips were secured at one end to a force displa-
cement transducer (FT 0.3, Grass Instruments Corporation) for
monitoring changes in isometric contraction and washed every
20 min with fresh PSS for 1 h. After the equilibration period,
guinea-pig aortic strips were contracted by washing in PSS con-
taining 80 mM KCl (equimolar substitution of K+ for Na+).
When the contraction reached a plateau (about 45 min) various
concentrations of the compounds (0.1, 0.5, 1, 5, 10, 50, 100 and
500 μM) were added cumulatively to the bath allowing for any
relaxation to obtain an equilibrated level of force. Addition of
the drug vehicle had no appreciable effect on K+-induced con-
traction (DMSO for all compounds).
ranked according to their similarity toward the template. In this
paper we observed a difference among the scores obtained for
corresponding template-candidate alignments due to the intrinsic
asymmetry of the FLAP scoring functions, as we briefly reported
in the caption of Table 3. This asymmetry is due to different
reasons. The first reason concerns the nature of the GRID Mo-
lecular Interaction Fields (MIF), that are a collection of float
numbers assigned to each node of the grid and reporting the
interaction energy between the molecule and the probe. In FLAP,
once two molecules are aligned, the MIF intersection is used for
the evaluation of molecular similarity. Thus, since it is highly
probable that when aligning the molecules their MIF points are
not aligned, a set of smoothing functions are applied to this dis-
crete system in order to allow the evaluation of the GRID energy
for the two aligned molecules in the same points of the space. It
implies that some small differences may occur because of this
discrete-to-continuous conversion. The second reason is the
speed of the FLAP screening calculation; we set up the speed
screening to 75%, that implies a sampling of the quadruplets that
are used for the molecular superpositions. In our case the
sampling was performed by a random selection of 1/6 of the
overall amount of quadruplets of both template and candidate.
However, even an exhaustive use of all the possible quadruplets
may not result in the equivalence of the scores, because of the
first reason.
Some details of the FLAP calculations are briefly listed: (1)
The conformers are produced by FLAP on-the-fly with a routine
that randomly modifies the molecule and adds a new confor-
mation to the set only when it differs from the ones already exist-
ing. First, a random search is performed using 40 samples for
each rotatable bond; this produces thousands of potential confor-
mers, which are minimized and retained only if within an energy
window of 20 kcal mol−1 from the minimum. Such conformers
are then ranked by energy and filtered according to their diversity
to the dataset, i.e. exclusion regards those conformers with low
calculated RMSD with respect to any of the other low-energy
conformers already selected. For the template DTZ the confor-
mational analysis produced 10 different conformations, whereas
20, 12, and 25 conformers were generated for the ligand mo-
lecules 1, 2, and 3, respectively. (2) GRID probes H, OH2 and
DRY were used. (3) The level of speed was set to 75%.
E. Molecular superimpositions
The program FLAP (Fingerprints for Ligands And Proteins) was
originally conceived as a fast algorithm for performing virtual
screening by means of a fingerprint-type description of both
ligands and protein binding sites according to their GRID Mo-
lecular Interaction Fields.19 It was designed to describe molecules
and protein structures in terms of a common reference frame-
work. In ligand-based virtual screening, compounds from virtual
libraries are ranked according to their similarity toward a given
reference molecule, called a template. Given the 3D structure of
a molecule, with the GRID force-field21 the interaction energies
between the molecule and a probe are calculated at different
points in space. As a result, one has the so-called Molecular
Interaction Fields (MIF): the hydrophobic and hydrogen bonding
interactions of the molecule with a virtual receptor are schema-
tised by the probes and encoded into energy values assigned to
the nodes of a grid built around each molecule. The core of the
FLAP process is finding the matches of the four hotspots of the
individual ligands to those of the template. When four hotspots
(quadruplets) of the ligand molecule are found to fit over four
hotspots of the template framework, a potentially favourable
superposition has been detected. The process is iterative and,
when executed exhaustively (maximum precision corresponds to
minimum speed), it continues until all the template quadruplets
are combined in all possible ways with all ligand quadruplets; in
cases of faster screening, a quadruplet sampling is carried out in
both the sets of the template and of each ligand molecule. For
each superposition, a score associated to the two molecules and
their common quadruplet of hotspots quantifies the overlap of
the two MIFs: a score is assigned to the superposition in order to
evaluate its “goodness” of MIF-overlapping. At the end of the
process, several superpositions of each ligand to each template
are memorized. Thus, compounds from virtual libraries are
Acknowledgements
Supported by grants from University of Bologna, and from
Istituto Nazionale per le Ricerche Cardiovascolari (INRC), Italy.
We thank Dr M. Baroni (Molecular Discovery Ltd, UK) for
valuable discussion on the FLAP algorithm, Dr F. Fusi and
Dr M. Frosini (University of Siena, Italy) for helpful discussion
on the pharmacological activity of this series of LTCC blockers,
and A. Casoni (University of Bologna, Italy) for animal care.
Notes and references
1 P. D. Henry, Am. J. Cardiol., 1980, 46, 1047–1058.
2 D. J. Triggle, Biochem. Pharmacol., 2007, 74, 1–9.
3 E. Grossman and F. H. Messerli, Prog. Cardiovasc. Dis., 2004, 47,
34–57.
4 M. Dean, T. Fojo and S. Bates, Nat. Rev. Cancer, 2005, 5, 275–284.
9002 | Org. Biomol. Chem., 2012, 10, 8994–9003
This journal is © The Royal Society of Chemistry 2012

View Article Online
5 A. Aszalos, Drug Discovery Today, 2007, 12, 838–843.
16 R. Budriesi, P. Ioan, A. Leoni, N. Pedemonte, A. Locatelli, M. Micucci,
A. Chiarini and L. J. Galietta, J. Med. Chem., 2011, 54, 3885–3094.
17 R. J. Tallarida and R. B. Murray, Manual of Pharmacologic
Calculations with Computer Programs, Springer-Verlag, New York, 2nd
edn, 1987.
18 (a) R. Budriesi, B. Cosimelli, P. Ioan, M. P. Ugenti, E. Carosati,
M. Frosini, F. Fusi, R. Spisani, S. Saponara, G. Cruciani, E. Novellino,
D. Spinelli and A. Chiarini, J. Med. Chem., 2009, 52, 2352–2362;
(b) E. Carosati, R. Budriesi, P. Ioan, G. Cruciani, F. Fusi, M. Frosini,
S. Saponara, F. Gasparrini, A. Ciogli, P. J. Stephens, F. J. Devlin,
D. Spinelli and A. Chiarini, J. Med. Chem., 2009, 52, 6637–6648.
19 (a) M. Baroni, G. Cruciani, S. Sciabola, F. Perruccio and J. S. Mason,
J. Chem. Inf. Model., 2007, 47, 279–294; (b) S. Cross, M. Baroni,
E. Carosati, P. Benedetti and S. Clementi, J. Chem. Inf. Model., 2010, 50,
1442–1450.
6 W. G. Nayler and P. Poole-Wilson, Basic Res. Cardiol., 1981, 76, 1–15.
7 Y. W. Kwon and D. J. Triggle, Chirality, 1991, 3, 393–404.
8 D. J. Triggle, Drug Discovery Today, 1997, 2, 138–147.
9 T. Tanaka, H. Inoue, T. Date, K. Okamura, K. Aoe, M. Takeda,
H. Kugita, S. Murata, T. Yamaguchi, K. Kikkawa, S. Nakajima and
T. Nagao, Chem. Pharm. Bull. (Tokyo), 1992, 40, 1476–1480.
10 R. Budriesi, B. Cosimelli, P. Ioan, E. Carosati, M. P. Ugenti and
R. Spisani, Curr. Med. Chem., 2007, 14, 279–287.
11 R. Budriesi, B. Cosimelli, P. Ioan, C. Z. Lanza, D. Spinelli and
A. Chiarini, J. Med. Chem., 2002, 45, 3475–3481.
12 (a) M. Viale, C. Cordazzo, B. Cosimelli, D. de Totero, P. Castagnola,
C. Aiello, E. Severi, G. Petrillo, M. Cianfriglia and D. Spinelli, J. Med.
Chem., 2009, 52, 259–266; M. Viale, et al., SciBX, 2(2), DOI: 10.1038/
scibx.2009.44; (b) M. Viale, C. Cordazzo, D. de Totero, R. Budriesi,
C. Rosano, A. Leoni, P. Ioan, C. Aiello, M. Croce, A. Andreani,
M. Rambaldi, P. Russo, A. Chiarini and D. Spinelli, Invest. New Drugs,
2011, 29, 98–109.
20 N. L. Allinger, Y. H. Yuh and J.-H. Lii, J. Am. Chem. Soc., 1989, 111,
8551–8566.
21 P. J. Goodford, J. Med. Chem., 1985, 28, 849–857.
13 E. Carosati, G. Cruciani, A. Chiarini, R. Budriesi, P. Ioan, R. Spisani,
D. Spinelli, B. Cosimelli, F. Fusi, M. Frosini, R. Matucci, F. Gasparrini,
A. Ciogli, P. J. Stephens and F. J. Devlin, J. Med. Chem., 2006, 49,
5206–5216.
14 R. Budriesi, E. Carosati, A. Chiarini, B. Cosimelli, G. Cruciani, P. Ioan,
D. Spinelli and R. Spisani, J. Med. Chem., 2005, 48, 2445–2456.
15 P. J. Stephens, F. J. Devlin, F. Gasparrini, A. Ciogli, D. Spinelli and
B. Cosimelli, J. Org. Chem., 2007, 72, 4707–4715.
22 (a) F. Gasparrini, D. Misiti, R. Rompietti and C. Villani, J. Chromatogr.,
A, 2005, 1064, 25–38; (b) A. Cavazzini, F. Dondi, S. Marmai,
E. Minghini, A. Massi, C. Villani, R. Rompietti and F. Gasparrini, Anal.
Chem., 2005, 77, 3113–3122; (c) A. Ciogli, I. D’Acquarica,
F. Gasparrini, C. Molinaro, R. Rompietti, P. Simone, C. Villani and
G. Zappia, J. Sep. Sci., 2010, 33, 3022–3032.
23 (a) H. Motulsky and A. Christopoulos, 2003; (b) H. Motulsky, 2003;
(c) http://www.GraphPad.com
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 8994–9003 | 9003