Enantioselective synthesis and physicochemical properties of libraries of 3-amino- and 3-amidofluoropiperidines

Article abstract of DOI:10.1002/chem.201302423

The enantioselective syntheses of 3-amino-5-fluoropiperidines and 3-amino-5,5-difluoropiperidines were developed using the ring enlargement of prolinols to access libraries of 3-amino- and 3-amidofluoropiperidines. The study of the physicochemical properties revealed that fluorine atom(s) decrease(s) the pK_a and modulate(s) the lipophilicity of 3-aminopiperidines. The relative stereochemistry of the fluorine atoms with the amino groups at C3 on the piperidine core has a small effect on the pK _a due to conformationnal modifications induced by fluorine atom(s). In the protonated forms, the C-F bond is in an axial position due to a dipole-dipole interaction between the N-H⁺ and C-F bonds. Predictions of the physicochemical properties using common software appeared to be limited to determine correct values of pK_a and/or differences of pK _a between cis- and trans-3-amino-5-fluoropiperidines.

Full text of DOI:10.1002/chem.201302423

DOI: 10.1002/chem.201302423
Full Paper
&
Enantioselectivity
Enantioselective Synthesis and Physicochemical Properties of
Libraries of 3-Amino- and 3-Amidofluoropiperidines
Aurꢀlie Orliac,[a] Julie Routier,[b] Fabienne Burgat Charvillon,[b] Wolfgang H. B. Sauer,[b]
Agnes Bombrun,*[b] Santosh S. Kulkarni,[c] Domingo Gomez Pardo,[a] and Janine Cossy*[a]
Abstract: The enantioselective syntheses of 3-amino-5-fluo-
ropiperidines and 3-amino-5,5-difluoropiperidines were de-
veloped using the ring enlargement of prolinols to access li-
braries of 3-amino- and 3-amidofluoropiperidines. The study
of the physicochemical properties revealed that fluorine
atom(s) decrease(s) the pKa and modulate(s) the lipophilicity
of 3-aminopiperidines. The relative stereochemistry of the
fluorine atoms with the amino groups at C3 on the piperi-
dine core has a small effect on the pKa due to conformation-
nal modifications induced by fluorine atom(s). In the proton-
ated forms, the CÀF bond is in an axial position due to
a dipole–dipole interaction between the NÀH+ and CÀF
bonds. Predictions of the physicochemical properties using
common software appeared to be limited to determine cor-
rect values of pKa and/or differences of pKa between cis- and
trans-3-amino-5-fluoropiperidines.
Introduction
selective NMDA antagonist for Parkinson’s disease as well as in
neceprevir (5,5-difluoropiperidine), a NS3 serine protease-inhib-
itor for the treatment of hepatitis C. It is worth noting that
a combination of an amino group at C3 and one or two fluo-
rine atom(s) at C5 on a piperidine core has never been exploit-
ed so far.
The aminopiperidine motif is present in a variety of natural oc-
curring alkaloids and is extensively used in medicinal chemis-
try. Due to their easy access, 4-aminopiperidine derivatives are
mainly synthesized and studied. However, piperidines with an
amino group at C3 are the most commonly encountered motif
in natural products such as in pseudodistomin,[1] aeruginosin
89A and B,[2] or pseudobactin B,[3] as well as in recent drugs
like tofacitinib,[4] alogliptin,[5] nemonoxacin,[6] balofloxacin,[7]
and troxipide.[8]
Because medicinal chemists often require fine-tuning of the
desired physicochemical and pharmaceutical properties of the
designed compounds, we would like to report herein the syn-
thesis of libraries of optically active 5-fluoropiperidines A2, B2,
and 5,5-difluoropiperidines A3 and B3. The influence of one or
two fluorine atom(s) on the physicochemical properties (pKa,
log D, and log P) of fluoropiperidines A2, B2 and A3, B3 com-
pared with non-fluorinated piperidines A1 and B1, were stud-
ied (Figure 1). Furthermore, predictions using commercially
available programs were tested to predict the pKa and log D of
piperidines A and B.
3-Aminopiperidines are associated with a large number of
biological activities and in consequence they are not selec-
tive.[9] One proven strategy to decrease unwanted properties is
to modulate the pKa of basic amines, which can be obtained
by the introduction of one or more fluorine atoms.[10,11] Inter-
estingly, 5-fluoropiperidines and 5,5-difluoropiperidines are
present in drugs such as MK-0657 (5-fluoropiperidine), a NR2B-
[a] Dr. A. Orliac, Dr. D. Gomez Pardo, Prof. J. Cossy
Laboratoire de Chimie Organique
ESPCI ParisTech, 10 Rue Vauquelin
75231 Paris Cedex 05 (France)
Fax: (+33)1-40-79-46-60
E-mail: janine.cossy@espci.fr
[b] Dr. J. Routier, Dr. F. Burgat Charvillon, Dr. W. H. B. Sauer, Dr. A. Bombrun
Merck Serono S.A. Geneva
9 Chemin des mines
1202 Geneva (Switzerland)
E-mail: abombrun@bluemail.ch
[c] Dr. S. S. Kulkarni
Syngene International Ltd
Plot Nos. 2 and 3, Bommasandra IV Phase
Jigani Link Road, Bangalore, 560 100 (India)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201302423.
Figure 1. Piperidines A and B.
Chem. Eur. J. 2014, 20, 3813 – 3824
3813
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Results and Discussion
Table 1. Synthesis of piperidines A1.
Synthesis of piperidines A and B
The synthesis of piperidines A2, B2 and A3, B3, has been en-
visaged from 4-aminoprolinols C through aziridinium inter-
mediates D, which would lead, after a nucleophilic attack, to
the desired piperidines E or F (Scheme 1).[12] To be successful
in synthesizing
3-substituted piperidines by ring enlargement of prolinols, the
nitrogen of these latter has to be substituted by an alkyl
group (PG1). We have to point out that the synthesis of
difluoropiperidines A3 and B3 has been envisaged from ami-
noprolinols C by ring expansion through aziridinium intermedi-
ates D, oxidation of the obtained piperidines F followed by di-
fluorination of ketopiperidines G (Scheme 1).
[a] Yield from 1.
Scheme 1. Retrosynthetic analysis of piperidines A2, A3, B2, and B3.
DAST=diethylaminosulfur trifluoride; TFAA=trifluoroacetic anhydride.
Synthesis of piperidines A1
3-Aminopiperidines A1 were obtained in two steps from the
commercially available protected 3-aminopiperidines 1 by
treatment with various benzoic acids (ArCOOH) [Et3N, propyl-
phosphonic anhydride solution (T3P), CH2Cl2] followed by the
cleavage of the N-tert-butylcarbamate protecting group (HCl,
dioxane). The synthesized piperidines A1a–A1g are reported
in Table 1.
Scheme 2. Access to piperidines A2.
At first, 4-hydroxyprolines 2 were transformed to 4-amino-
prolinols 5 (Table 2) by treatment with SOCl2/MeOH to form
the corresponding methyl esters, which were then N-protected
(PG1ÀCl, in which PG1 =trityl (Tr) or 4-methoxybenzyl (PMB)) to
produce the corresponding N-alkylated proline esters 3. The
hydroxy group at C4 was then transformed to an amino group
in three steps. After mesylation (MsCl, Et3N, CH2Cl2), followed
by the addition of an azide derivative (NaN3 or n-Bu4NN3), the
Synthesis of piperidines A2
The access to aminopiperidines A2 was realized from 4-amino-
prolinols 5 synthesized from optically active cis- and trans-4-hy-
droxyprolines 2a–2d, which are either naturally occurring and/
or commercially available (Scheme 2).
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3814
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 2. Synthesis of 4-aminoprolinols 5.
Table 3. Synthesis of piperidines 7 from prolinols 5.
Entry
Prolinols 5
Piperidines 6
Piperidines 7
(yield)
(cond, yield)
1
2
3
4
5
6
[a] Overall yield from 2.
azido compounds 4 were isolated and reduced by LiAlH4 (THF,
08C) to afford the corresponding amino derivatives, which
were then transformed to the corresponding N-tert-butylcarba-
mate derivatives 5 after treatment with Boc2O (dioxane, RT;
Table 2).
Prolinols 5 were then transformed to the N-substituted
3-amino-5-fluoropiperidines A2 through piperidines 7. To
obtain piperidines 7, prolinols 5 were treated with diethylami-
nosulfur trifluoride (DAST; 1.4 equiv) in THF and the corre-
sponding 5-fluoropiperidines
6 were obtained in good
yields.[13] We have to point out that N-trityl prolinols 5 were ex-
clusively converted to 6 and N-p-methoxybenzyl prolinols 5
were transformed to a mixture of piperidines 6 and pyrroli-
dines 6’, which were then separated. Cleavage of the N-trityl
protecting group (HCl in EtOH) and the N-PMB protecting
group (Pd/C, HCOOH) in piperidines 6 led to 3-N-Boc-5-fluoro-
piperidines 7 (Table 3).
Having piperidines 7 in hand, they were transformed to a va-
riety of N-benzoyl piperidines A2 by treatment with various
benzoyl chlorides (ArCOCl, Et3N, CH2Cl2) followed by the cleav-
age of the N-Boc protecting group [HCl (4m), dioxane]
(Table 4). Different benzoyl groups, substituted by an electron-
donating group (p-OMe, m-OMe, o-OMe) or an electron-with-
drawing group (p-NO2, m-NO2, o-NO2), were introduced at N1
to study the electronic effect of the benzoyl substituents on
the physicochemical properties of the 3-aminopiperidines
A2[14] (Table 4).
Synthesis of piperidines A3
The synthesis of 5,5-difluoropiperidines A3 was realized from
piperidin-3-ols 11, which were obtained in seven steps from
4-hydroxyprolinols 2 (Scheme 3). Thus, 4-hydroxyprolines 2
were transformed to prolinol esters 8 (SOCl2, MeOH then BnBr,
Et3N, CH2Cl2). After mesylation followed by the addition of
nBu4NN3 (CH3CN), the 4-azidoproline esters 9 were isolated
(73–87%). The reduction of 9 followed by treatment with
Boc2O led to 10 (73–91%), which were transformed to protect-
ed 3-aminopiperidinols 11 (trifluoroacetic anhydride (TFAA),
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3815
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Et3N, THF, MW, 8 h, 1008C then NaOH, 77–88%),[15] precursors
of piperidines A3 (Scheme 3).
Table 4. Synthesis of piperidines A2.
The formation of 5,5-difluoropiperidines A3 was achieved in
five steps from 11 (Scheme 3). Piperidines 11 were then depro-
tected (H2, Pd/C, MeOH) to produce 12, which was then treat-
ed with various benzoyl chlorides (Et3N, CH2Cl2) to produce N-
benzoylamidopiperidines 13 (Table 5). After oxidation of 13
(2-iodoxybenzoic acid (IBX), DMSO), the resulting ketones were
treated with DAST to afford 5,5-difluoropiperidines and, after
cleavage of the N-tert-butylcarbamate group (HCl, dioxane), pi-
peridines A3 were isolated (Table 5).
Table 5. Synthesis of difluoropiperidines A3.
[a] Yield from 7.
[a] Yield from 12.
Synthesis of piperidines B1
3-Amidopiperidines B1 were synthesized from 14, which are
Scheme 3. Synthesis of 5-amino-3-hydroxypiperidines 11, precursors of A3.
commercially available, by using the same synthetic pathway
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3816
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
mate group (HCl, dioxane, MeOH), a benzoylation was realized
(ArCOCl, Et3N, CH2Cl2) and, after the N1 deprotection, piperi-
dines B2c–B2h[17] were isolated (Table 8).
Table 6. Synthesis of piperidines B1.
Table 7. Synthesis of fluoropiperidines B2a, B2b, ent-B2a, and ent-B2b.
[a] Yield from 14.
than for synthesizing A1 from 1 (benzoylation and N1-depro-
tection) (Table 6).
Synthesis of piperidines B2
[a] Yield from 15.
5-Fluoro-3-amidopiperidines B2 were prepared either from
prolinols 15,[16] piperidines
(Scheme 4).
6 and from piperidines 7
Table 8. Synthesis of fluoropiperidines B2c–B2h.
Scheme 4. Retrosynthetic analysis of fluoropiperidines B2.
The synthesis of piperidines B2 possessing a free amino
group at N1 and a benzoyl group at N3 (B2a and B2b) was re-
alized from prolinols 15 (Table 7). After introduction of the
benzoyl group at N4 (PhCOCl, Et3N, CH2Cl2), the resulting proli-
nols 16 were treated with DAST to produce piperidines 17
and, after treatment with HCl, piperidines B2 were obtained
(Table 7).
The access to piperidines B2c–B2h was realized from pro-
tected 3-aminopiperidines 6. After cleavage of the N-carba-
[a] Yield from 18.
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3817
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The transformation of 7 to piperidines B2i–B2n was also re-
alized. After a protection/deprotection sequence involving
a N1-Fmoc protection [N-(9-fluorenylmethoxycarbonyloxy)suc-
cinimide (Fmoc-OSu), dioxane/water], a N3-Boc cleavage (HCl,
dioxane/MeOH), a N3-benzoylation (ArCOCl, Et3N, CH2Cl2) and
Table 10. Synthesis of difluoropiperidines B3.
a
N1-Fmoc cleavage, piperidines B2i–B2n were isolated
(Table 9).
Table 9. Synthesis of fluoropiperidines B2i–B2n.
[a] Yield from 19.
[a] Yield from 21.
Synthesis of piperidines B3
5,5-Difluoropiperidines B3 were synthesized from 12. After pro-
tection of 12 (Fmoc-Cl, NaHCO3, dioxane/water), the resulting
piperidines 20 were oxidized (using IBX), treated with DAST
and, after cleavage of the N-tert-butylcarbamate, 5,5-difluoropi-
peridines 21 were produced. Treatment of 21 with various
benzoyl chlorides and cleavage of the N-Fmoc group (piperi-
dine, DMF) produced piperidines B3[18] (Table 10).
pH 3 to 12 in a MeOH/H2O solution with an ionic strength ad-
justed to 0.15m KCl.[21]
The pKa of 28 amidopiperidines B have been measured and
the values are reported in Figure 2. It can be noticed that an
electron-withdrawing or electron-donating substituent on the
benzoyl group at N3 as well as its position on the aromatic
ring has no effect on the pKa. Piperidines B1 exhibit a pKa
around 9.1 and the pKa is decreased to 7.3 for piperidines
trans-B2 and to 7.1 for piperidines cis-B2. In addition, when
two fluorine atoms are present at C5, the pKa of B3 is de-
creased to 5.5. A DpKa of 1.6–2.1 is observed in between piper-
idines B1 and B2 and a DpKa of 3.5–3.6 in between non-fluori-
nated piperidines B1 and difluoropiperidines B3.
Based on the data reported in the literature,[10] one would
expect a decrease in basicity of 2.0–2.3 of pKa units for piperi-
dines possessing a fluorine atom at C5 in an equatorial posi-
tion due to its electron-withdrawing properties and a decrease
of 1.4 of pKa units is expected when a fluorine atom at C5 oc-
cupies an axial position due to a 1,3-syn dipolar stabilization of
the protonated species. To establish the axial or equatorial po-
sition of the fluorine atom in trans-B2 and cis-B2, NMR spectro-
scopic studies in CD3OD and [D6]DMSO of B2d and B2g were
Physicochemical properties of piperidines A and B
The modulation of the physicochemical properties of piperi-
dines substituted by fluorine atom(s) has been previously re-
ported in the literature.[10,11,19,20] In the present study, the physi-
cochemical properties of piperidines A and B, specifically the
acidity/basicity properties (pKa) and the lipophilicity (log D7.4
and log P) of 64 piperidines have been measured and the
values are reported in the Supporting Information.
Effect of fluorine atom(s) on the pKa
The pKa values of piperidines A and B were measured by po-
tentiometric titration at a concentration of 0.1 to 0.2 mm from
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3818
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The dipole of the CÀF bond in the axial position is favorably
aligned antiparallel to the dipole of the HÀN+ bond. This ar-
rangement stabilizes the protonated form and explains the de-
crease of the pKa of 1.8 units instead of the predicted 2.3 units.
For the cis-B2 compounds, the difference of pKa in between
cis-B2 and B1 is around 2.0 similar to the one observed in be-
tween trans-B2 and B1. This value is related also to the axial
orientation of the fluorine atom. Thus, conformational studies
1
were achieved on compound B2g by H NMR spectroscopy in
CD3OD on the non-protonated and protonated form. The
values of the coupling constants used to determine the confor-
1
mations are reported in Figure 4. The H NMR analysis of the
^
Figure 2. pKa values of piperidines B. Data points are given for B1 ( ),
~
&
*
trans-B2 ( ), cis-B2 ( ), and B3 ( ).
performed. According to the coupling constants for H4 in
[D6]DMSO (JH4–F =40.0, JH4–H4’ =13.4, JH4–H3 =11.3, JH4–H5 =2.1 Hz)
for trans-B2d, the C3-amido group occupies an equatorial po-
sition and the fluorine atom at C5 occupies an axial position in
the non-protonated form as well as in the protonated form
trans-B2dÀH+ (Figure 3). Protonated 3-fluoropiperidines trans-
B2dÀH+ exist exclusively as a single conformer with the fluo-
rine atom in an axial position.[22]
Figure 4. Acid/base equilibrium for B2g in CD3OD.
non-protonated B2g form in CD3OD (JH5–F =47.9, JH5–H4 =7.8,
J
H5–H6 =7.8, JH5–H4’ =3.8, JH5–H6’ =3.8 Hz) reveals that the substitu-
ents at C3 and C5 are in an equatorial position.[23] On the con-
trary, for the protonated form, the C3- and C5-substituents are
in an axial position (JH3–H2 =4.7, JH3–H2’ =4.7, JH3–H4 =4.7, JH3–H4’
=
4.7 Hz). The conformer in which the C3- and C5-substituents
are in an axial position is stabilized by a 1,3-syn CF···+NH dipo-
lar stabilization,[10] which explains the similar pKa of trans-B2
and cis-B2.
It is worth pointing out that the analysis of the 1H NMR
spectroscopic data of B2g in CDCl3 (JH6–F =37.3, JH6–H6’ =14.0,
JH6–H5 =1.4 Hz) shows that in the non-protonated form, the
amido group occupies the axial position (Figure 5). The stabili-
ty of this conformation can be explained by an intramolecular
hydrogen-bonding between the oxygen of the carbonyl and
the N1-hydrogen atom.
Figure 3. Acid/base equilibrium for B2d in [D6]DMSO.
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3819
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 5. Conformation of B2g in CDCl3.
The introduction of two fluorine atoms at C5 (compound
B3) implies a decrease of 3.6 units of pKa compared with the
pKa of B1. This decrease corresponds to the expected values
due to the polar CÀF bond (axial), which is 1,3-syn to the
N+ÀH bond (1.4 pKa units) of the protonated piperidine and
the second fluorine atom is in an equatorial position (2.3 pKa
units) (Figure 6).
Figure 7. Measured pKa values of piperidines A. Data points are given for A1
~
&
^
*
(
), cis-A2 ( ), trans-A2 ( ), and A3 ( ).
around 8.2, for the cis-A2 series around 7.9, for the trans-A2
series around 7.6 and for the A3 series around 7.0.
The comparison between the pKa of A1 and cis-A2 shows
a decrease of 0.3 of pKa units when the fluorine at C5 and the
amino group at C3 are cis. For the trans-A2 series, a decrease
of 0.6 units of pKa is noticed and corresponds to a fluorine
atom at C5 trans to the amino group at C3. When two fluorine
atoms at C5 are present (compounds A3), a decrease of 1.2
units of pKa has been measured.
The pKa difference between the cis-A2 and trans-A2 (0.3 pKa
units) is larger than for the cis-B2 and trans-B2 (0.2 pKa units)
and the pKa for the cis-A2 is somewhat higher than expect-
ed.[10,24] Quantum mechanical calculations in vacuo clearly indi-
cate a favored axial position for the ammonium group for
both cis- and trans-A2 in its protonated form, due to hydrogen
bonding between the oxygen of the amide and the ammo-
nium groups (Figure 8). In aqueous solution, however, the
energy differences is reduced and, for trans-A2 even reversed.
Thus, one can assume that conformational equilibrium be-
tween the axial and equatorial positions of the fluorine atom
can explain the reduced electron-withdrawing effect of this
atom in cis-A2’’–H+.
Figure 6. Acid/base equilibrium for B3c in CD3OD.
The pKa values of 28 piperidines A have been determined.
The results are reported in Figure 7. As previously observed for
piperidines B, it can be noticed that the presence of an elec-
tron-withdrawing or electron-donating substituent as well as
its position on the benzoyl group at N1 have no effect on the
pKa of A1, A2, and A3. For the A1 series, the pKa values are
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3820
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 8. Conformational studies of cis-A2 and trans-A2 in vacuum.
For the difluoropiperidines A3, the increase of the acidity of
these compounds, compared with A1, is due to the electrone-
gativity of the two fluorine atoms.
The introduction of one or two fluorine atom(s) on the pi-
peridines A and B decreases their pKa; this effect is more im-
portant in piperidines B than in piperidines A, which is due to
the proximity of the fluorine atom to the basic center. In both
series A and B, it is important to note that the relative stereo-
chemistry of the substituents influences the pKa but not with
the same impact.[10,11,25]
Figure 9. Log P values of piperidines A with an o-OMe-Ph substituent. Data
~
&
*
^
points are given for A3 ( ), A1 ( ), cis-A2 ( ), and trans-A2 ( ).
atom is present at C3, as in compound A2, the log P values are
smaller than for A1. Replacement of one hydrogen atom by
a fluorine atom in 3-aminopiperidines A had the propensity to
decrease the log P value (À0.3 for A2e and À0.2 for A2h com-
pared with 0.0 for A1d). Furthermore, when two fluorine
atoms are present, such as in compound A3, the log P values
are higher than for A1 (e.g., Dlog P=0.5 between A1d and
A3d; Figure 9).
Interestingly, all o-NO2-substituted compounds exhibited the
lowest lipophilicity in the A2 series with log P=À1.1 for A2k
(see the Supporting Information, Table S1, entry 16). In summa-
ry, for 3-aminopiperidines A the introduction of two fluorine
atoms is necessary to significantly increase the lipophilicity of
the piperidines (0.3–0.8 log P units).
Modulation of the lipophilicity
The distribution coefficient (log D) is the logarithmic coefficient
of the distribution of a molecule between octanol and an
aqueous media at a given pH. Log D is the parameter of
choice to determine the lipophilicity of ionizable molecules.
In the present study, log D of piperidines A and B were
measured at pH 7.4 by using the micro-shake flask technic.[26,27]
To correct the ionization state of piperidines A and B, log P
values were calculated from log D values at pH 7.4 and pKa
values according to the following equation log D (pH)=
log PÀlog (1+10pK –pH). All the data are listed in the Supporting
a
Information.
The study of the difference of log D between A1, A2, and
A3 can be of interest as it is related to the ionization potential
of piperidines A. The introduction of one fluorine atom did not
affect significantly the log D despite the correction due to the
ionized species induced by the decrease of pKa between A1
and A2. However, the introduction of two fluorine atoms sig-
nificantly impacts log D. The difluorinated amines A3a–A3 f ex-
hibited a log D between 0.3 and 0.6 (see the Supporting Infor-
mation, Table S1, entries 26–31). This important increase in the
log D of A3 compared with the log D of A1 is related to the
decrease of the pKa (pKa below 7.4). The introduction of two
fluorine atoms is necessary to significantly modulate the lipo-
philicity of the aminopiperidine with 1.2 of log D units as an
average gain.
The log P values of non-fluorinated 3-amidopiperidines B1
were calculated and are comprised in between 0.7 and 1.0 (see
the Supporting Information, Table S1, entries 34–39). In this
series, the effect of the substituents on the aryl group has no
significant effect except for B1g which possesses a o-nitro sub-
stituent on the benzoyl group at N3 (log P=0.1, see the Sup-
porting Information, Table S1, entry 40). In this latter case, the
amido group is not in the same plane as the phenyl moiety
which increases the accessible polar surface of the amido
group and therefore decreases log P.
Introduction of one fluorine atom usually raised the log P
values but exceptions were observed wherein the introduction
of a fluorine atom on the piperidine core decreases the lipo-
philicity. In piperidine cis-B2, the presence of the fluorine atom
at C5, cis to the amido group at C3, has no effect on the lipo-
philicity. In piperidine trans-B2, the presence of a fluorine atom
at C5, trans to the amido group at C3, slightly decreases the
lipophilicity. For example, the log P value for trans-3-amido-5-
The inherent lipophilicity log P values range between À0.4
to 0.4 for non-fluorinated 3-aminopiperidines A1 with the
smallest value for the o-nitro compound A1g (see the Sup-
porting Information, Table S1, entries 1–8). When a fluorine
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3821
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
directly related to the increase of the concentration of the un-
charged species in the octanol phase.
Predictions
As many important biophysical parameters like solubility, oral
bioavailability, and CNS penetration are intimately related to
log D, predictions are often used to screen out compounds
that are considered unsuitable[29] because they would not fit
a required property profile. Several programs are available for
the calculation of log D, pKa, and log P, which are in general ad-
equate but, their performance on pharmaceutical data sets
have been challenged and the shortcomings were attributed
to an insufficient coverage of the concerned chemical space in
training the models.[30]
Before using predictions to decide if compounds should be
synthesized or not, we were therefore interested in checking
whether the available programs are appropriate for the piperi-
dines under investigation. To this end, we have compared our
experimental values (see the Supporting Information, Table S1)
with the predictions of several commercially available
programs, namely ACD,[31] Discovery Studio/Pipeline Pilot,[32]
Epik,[33] GastroPlus,[34] Marvin,[35] pKcalc[36] and StarDrop[37] (see
the Supporting Information, Table S2).[38] Typically, log D predic-
tions are performed by predicting the log P value for the neu-
tral species at first and then correcting the result for an ioniza-
ble compound based on the prediction of its pKa according to
the Scherrer–Howard Equation.[39]
To estimate the quality of the results, correlation coefficients
r2 between predictions and experiments and the Root Mean
Square Errors (RMSEs) have been calculated for each of the
subseries and for the whole data set (Tables 11 and 12).
Figure 10. Log P values of piperidines B with a N3-benzoyl substituent. Data
~
&
*
^
points are given for B3 ( ), B1 ( ), cis-B2 ( ), and trans-B2 ( ).
fluoropiperidine B2a is 0.5, whereas for the cis isomer ent-B2b
the log P value is 0.8 (Figure 10). We can deduce from these
values that for 3-amido-5-fluoropiperidines B, the relative ste-
reochemistry between the fluorine atom at C5 and the amido
group at C3 has no significant impact on the pKa and on the
lipophilicity.[28]
Table 11. Predictions of pKa.
Entry
Predictor
A and B
A
B
r2
RMSE
r2
RMSE
r2
RMSE
1
2
3
4
5
6
DS/PiPi
pKcalc
Epik
Marvin
GastroPlus
ACD
0.03
0.97
0.91
0.73
0.94
0.77
2.49
0.94
0.77
1.28
0.43
1.09
0.42
0.86
0.87
0.88
0.88
0.89
0.98
0.97
0.69
1.67
0.38
1.52
0.00
0.99
0.95
0.99
0.97
0.99
3.38
0.92
0.84
0.72
0.47
0.22
As for piperidines A, the presence of two fluorine atoms on
the piperidine core of B increases the lipophilicity. The log P
values of B3 ranged from 1.1 to 1.3 (see the Supporting Infor-
mation, Table S1, entries 58–63). As noticed previously, the
presence of a substituent on the N3-benzoyl group has no in-
fluence on the log P values except for piperidine B3g having
a o-NO2 substituent on the benzoyl group (see the Supporting
Information, Table S1, entry 64). As for compound A3g, piperi-
dine B3g shows the lowest log P value (log P=0.3).
Table 12. Predictions of log D7.4
.
Entry
Predictors
A and B
A
B
r2
RMSE
r2
RMSE
r2
RMSE
1
2
3
4
5
DS/PiPi
Marvin
GastroPlus
StarDrop
ACD
0.01
0.70
0.77
0.59
0.51
0.77
0.57
0.70
0.50
0.98
0.68
0.64
0.60
0.75
0.55
0.66
0.67
0.70
0.33
0.69
0.04
0.74
0.77
0.58
0.69
0.87
0.44
0.71
0.62
1.20
We have to point out that, in considering the log D value at
pH 7.4, the introduction of a fluorine atom at C5 has an impact
on the lipophilicity. All piperidines B1 are poorly lipophilic
molecules with log D ranging from À1.1 to À0.7, whereas
log D for cis-B2 ranged from 0.6 to 0.7 and for trans-B2 be-
tween 0.1 and 0.4 (except o-nitro-substituted compounds).
This increase of log D values between B1 and B2 can be ex-
plained by the decrease of the pKa between 3-amido-5-fluoro-
piperidines B2 and B1 as described above. This observation is
In general, the predicted pKa values are too high, but repro-
duce the overall trend of the decreasing acidity of the substi-
tuted piperidines related to the addition of fluorine atom(s).
The influence of aryl substituents is typically underestimated
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3822
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
and, in addition, none of the programs is able to differentiate
between cis- and trans-isomers of monofluorinated piperidines.
Discovery Studio/Pipeline Pilot predicts a pKa of 10.4 for all
3-aminopiperidines B, irrespective of the presence of fluorine
atom(s) at C5 or substituents on the N-benzoyl group, which
led to poor correlations (Table 11, entry 1). The seemingly ac-
ceptable performance in the prediction of log D for 3-aminopi-
peridines A can be attributed to the partial cancellation of
errors (Table 12, entry 1). The dedicated pKa predictors, pKcalc
and Epik, show good global correlations, albeit with rather
large RMSEs (Table 11, entries 2 and 3). Although these pro-
grams take into account the effect of fluorine atom(s), the in-
fluence of the substituents on the N-benzoyl group is largely
ignored. The high r2 values must therefore be regarded as an
artifact and not as an indication of good reliability. A similar
observation was made for the pKa predictions utilizing Marvin,
which benefits from error cancellation to give acceptable,
though under-predicted log D values (Table 11, entries 2 and
4). GastroPlus exhibits the lowest RMSEs for the pKa and pre-
dicts log D consistently well (Table 11, entry 5 and Table 12,
entry 3); the r2 values are also acceptable. However, its fails to
differentiate the pKa value of the non-fluorinated 3-aminopi-
peridines A and B. StarDrop does not provide a model for pKa,
but predicts log D values directly from the structure. The esti-
mates show the lowest RMSEs, but are still too noisy to be
truly predictive (Table 12, entry 4). ACD provides the best pKa
predictions for the 3-amidopiperidines B, whereas the 3-amino-
piperidines A are consistently predicted about 1.5 pKa units
too high, which could be attributed to a systematic error in
the handling of this scaffold (Table 11, entry 6). The extension
of chemical space coverage with additional data is a recom-
mended procedure to improve the physicochemical predic-
tions of commercially available programs[40] and has been illus-
trated for substituted pyrrolidines and 4-amino-piperidin-1-yl
amides.[41] Applying a simple subtractive correction term to
aminopiperidines A reduces the RMSE to 0.18 (series A) and
0.20 (entire set), respectively, improving the overall correlation
significantly to 0.97. Whether the corrected values would lead
to a similar improvement for log D predictions it remains
doubtful, as 3-amidopiperidines A are already better described
than 3-aminopiperidines B.
dition of fluorine atom(s) allows the fine-tuning of the physico-
chemical properties of the resulting compounds, making
3-amino-5-fluoropiperidines and 3-amino-5,5-difluoropiperi-
dines useful building blocks for medicinal chemists. For 3-
amidopiperidines B, one fluorination at C5 reduces the pKa by
about 1.8 units, with a difference of 0.25 between the cis and
the trans forms, but has only little impact on log P (0 for cis-
monofluoro, À0.2 for trans-monofluoro, +0.4 for difluoro).[42]
The basicity of the primary amino group in 3-amino-piperidine
amides A loses 0.3 units on the addition of a fluorine atom cis
to the amino group at C3, 0.6 units if the fluorine atom is in
trans to the amino group at C3, and 1.2 units for two fluorine
atoms, with only moderate variation of log P (À0.3 for the first
F, +0.7 for the second fluorine atom).[42] We have also noticed
that regardless of the series, A or B, compounds with ortho-
nitro group have a different behavior as a result of their partic-
ular conformational preferences. All these effects were not ad-
equately predicted by in silico methods, so these methods
should be used with caution when contemplating the poten-
tial use of such compounds and efforts have to be made to
develop programs to better predict the properties of 3-amido-
and 3-aminopiperidines.
Experimental Section
Representative procedure
DAST (0.02 mL, 0.16 mmol, 1.4 equiv) was added dropwise to
a stirred solution of 5a (53 mg, 0.12 mmol, 1 equiv) in THF (1 mL)
at 08C. The reaction was stirred 1 h at 08C and 1 h at RT. The reac-
tion mixture was cooled to 08C and a saturated aqueous solution
of Na2CO3 (2 mL) was added to adjust to pH 12. The phases were
separated and the organic phase was extracted with EtOAc (2ꢂ
5 mL). The organic fractions were gathered, dried over Na2SO4, fil-
tered, and concentrated under reduced pressure. The crude materi-
al was purified by flash chromatography (silica gel, PE/EtOAc=
95:5) to afford 6a as a white solid (38 mg, 0.08 mmol, 71%). M.p.
1
1208C; [a]2D0 =À52.5 (c=0.5, CHCl3); H NMR ([D6]DMSO, 400 MHz,
1208C): d=7.50–7.44 (m, 6H), 7.34–7.26 (m, 6H), 7.21–7.14 (m,
3H), 6.43 (d, J=7.7 Hz, 1H), 4.96 (dm, J=47.9 Hz, 1H), 4.04 (m,
1H), 2.82 (m, 1H), 2.74 (d, J=11 Hz, 1H), 2.03–1.83 (m, 1H), 1.64
(dd, J=9.9 and 9.9 Hz, 1H), 1.57–1.37 ppm (m, 11H); IR (neat): n˜ =
3437, 2361, 1711, 1169, 747, 709, 629 cmÀ1; HRMS (ESI): m/z calcd
for C29H33O2N2FNa: 483.24183 [M+Na+]; found: 483.24174.
In summary, none of the tested programs showed sufficient
accuracy and precision to warrant their use as a filter before
the synthesis of substituted piperidines. Although improve-
ments could certainly be achieved by the extension of the cov-
ered chemical space by using experimental values of con-
cerned structures, the inherent incapability of the models to
differentiate diastereomers will be difficult to overcome right
now.
Acknowledgement
A.O. thanks Merck Serono S. A. for a grant, and Merck Serono
S. A. is gratefully acknowledged for financial support.
Keywords: basicity · enantioselectivity · fluorine · lipophilicity ·
stereoelectronic effects
Conclusion
Robust synthetic routes to optically active 3-amino-5-fluoropi-
peridines and 3-amino-5,5-difluoropiperidines have been de-
veloped. Diverse substituted analogues were synthesized and
the influence of fluorine atom(s) on the basicity and lipophilici-
ty of 3-aminopiperidines was evaluated. The stereospecific ad-
[1] M. Ishibashi, Y. Ohizumi, T. Sasaki, H. Nakamura, Y. Hirata, J. Kobayashi, J.
Org. Chem. 1987, 52, 450–453.
[2] K. Ersmark, J. R. Del Valle, S. Hanessian, Angew. Chem. 2008, 120, 1220–
1242; Angew. Chem. Int. Ed. 2008, 47, 1202–1223.
[3] N. Andriollo, A. Guarini, G. Cassani, J. Agric. Food Chem. 1992, 40, 1245–
1248.
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3823
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[4] M. Soth, J. C. Hermann, C. Yee, M. Alam, J. W. Barnett, P. Berry, M. F.
Browner, K. Frank, S. Frauchiger, S. Harris, Y. He, M. Hekmat-Nejad, T.
Hendricks, R. Henningsen, R. Hilgenkamp, H. Ho, A. Hoffman, P.-Y. Hsu,
D.-Q. Hu, A. Itano, S. Jaime-Figueroa, A. Jahangir, S. Jin, A. Kuglstatter,
A. K. Kutach, C. Liao, S. Lynch, J. Menke, L. Niu, V. Patel, A. Railkar, D.
Roy, A. Shao, D. Shaw, S. Steiner, Y. Sun, S.-L. Tan, S. Wang, M. D. Vu, J.
Med. Chem. 2013, 56, 345–356.
[5] J. Feng, Z. Zhang, M. B. Wallace, J. A. Stafford, S. W. Kaldor, D. B. Kassel,
M. Navre, L. Shi, R. J. Skene, T. Asakawa, K. Takeuchi, R. Xu, D. R. Webb,
S. L. Gwaltney, J. Med. Chem. 2007, 50, 2297–2300.
[21] A. Avdeef, B. Testa, Cell. Mol. Life Sci. 2002, 59, 1681–1689.
[22] a) D. C. Lankin, N. S. Chandrakumar, S. N. Rao, D. P. Spangler, J. P. Snyder,
J. Am. Chem. Soc. 1993, 115, 3356–3357; b) J. P. Snyder, N. S. Chandraku-
mar, H. Sato, D. C. Lankin, J. Am. Chem. Soc. 2000, 122, 544–545; c) A.
Sun, D. C. Lankin, K. Hardcastle, J. P. Snyder, Chem. Eur. J. 2005, 11,
1579–1591.
[23] This conformation is validated by X-ray diffraction analysis.
[24] 1H NMR studies cannot be realized in these series due to the presence
of rotamers and heated 1H NMR spectroscopic studies would change
the equilibrium between the different conformations.
[6] C.-R. Li, Y. Li, G.-Q. Li, X.-Y. Yang, W.-X. Zhang, R.-H. Lou, J.-F. Liu, M.
Yuan, P. Huang, S. Cen, L.-Y. Yu, L.-X. Zhao, J.-D. Jiang, X.-F. You, J. Anti-
microb. Chemother. 2010, 65, 2411–2415.
[25] H. H. Jensen, L. Lyngbye, A. Jensen, M. Bols, Chem. Eur. J. 2002, 8,
1218–1226.
[26] E. H. Kerns, J. Pharm. Sci. 2001, 90, 1838–1858.
[7] L.-S. Feng, M.-L. Liu, B. Wang, Y. Chai, X.-Q. Hao, S. Meng, H.-Y. Guo, Eur.
J. Med. Chem. 2010, 45, 3407–3412.
[8] K. Kusugami, K. Ina, T. Hosoyashi, F. Kobayashi, H. Kusajima, K. Momo, Y.
Nishio, Digestion Digest. Liver Dis. 2000, 32, 305–311.
[27] L. Hitzel, A. P. Watt, K. L. Locker, Pharm. Res. 2000, 17, 1389–1395.
[28] G. Vistoli, V. Straniero, A. Pedretti, L. Fumagalli, C. Bolchi, M. Pallavicini,
E. Valoti, B. Testa, Chirality 2012, 24, 566–576.
[29] a) A. Avdeef, Curr. Top. Med. Chem. 2001, 1, 277–351; b) D. F. Veber, S. R.
Johnson, H.-Y. Cheng, B. R. Smith, K. W. Ward, K. D. Kopple, J. Med.
Chem. 2002, 45, 2615–2623; c) H. E. Selick, A. P. Beresford, M. H. Tarbit,
Drug Discovery Today 2002, 7, 109–116; d) H. Pajouhesh, G. R. Lenz,
NeuroRx 2005, 2, 541–553; e) S. G. Summerfield, A. J. Stevens, L. Cutler,
M. del Carmen Osuna, B. Hammond, S. P. Tang, A. Hersey, D. J. Spalding,
P. Jeffrey, J. Pharmacol. Exp. Ther. 2006, 316, 1282–1290; f) S. K. Bhal, K.
Kassam, I. G. Peirson, G. M. Pearl, Mol. Pharm. 2007, 4, 556–560.
[30] a) R. Mannhold, G. I. Poda, C. Ostermann, I. V. Tetko, J. Pharm. Sci. 2009,
98, 861–893; b) S. Martel, F. Gillerat, E. Carosati, D. Maiarelli, I. V. Tetko,
R. Mannhold, P. A. Carrupt, Eur. J. Pharm. Sci. 2013, 48, 21–29.
[31] ACD/PhysChem, version 12.01, Advanced Chemistry Development, Inc.,
Toronto, ON, Canada, www.acdlabs.com, 2009.
[9] a) M. Burger, M. Lindvall, W. Han, J. Lan, G. Nishiguchi, C. Shafer, C. Bella-
macina, K. Huh, G. Atallah, C. McBride, W. Antonios-McCrea, Jr., T. Zavo-
rotinskaya, A. Walter, P. Garcia (Novartis Vaccines and Diagnostics, Inc.,
USA), WO 2008/106692, 2008 [Chem. Abstr. 2008, 149, 332352]; b) M.
Burger, J. Lan, M. Lindvall, G. Nishiguchi, M. Tetalman (Novartis AG,
Switz), WO 2009/109576, 2009 [Chem. Abstr. 2009, 151, 358734]; c) M.
Burger, J. Lan, M. Lindvall, G. Nishiguchi, M. Tetalman (Novartis AG,
Switz), US 2010/0216839, 2009 [Chem. Abstr. 2010, 153, 334028];
d) D. A. Erlanson, D. Marcotte, G. Kumaravel, J. Fan, D. Wang, J. H.
Cuervo, L. Silvian, N. Powell, M. Bui, B. T. Hopkins, A. Taveras, B. Guan, P.
Conlon, M. Zhong, T. J. Jenkins, D. Scott, A. A. Lugovskoy (Biogen Idec
Ma Inc.),WO 2011/029046, 2010 [Chem. Abstr. 2011, 154, 361044]; e) X.
Wang, A. J. Ebens Jr., (Genentech, Inc., USA, F. Hoffman-La Roche), US
2011/0059961, 2010 [Chem. Abstr. 2011, 154, 344337].
[32] Discovery Studio version 3.1, Accelrys Inc.
[33] Epik version 2.2 Schrçdinger LLC 2011, J. C. Shelley, A. Cholleti, L. L.
Frye, J. R. Greenwood, M. R. Timlin, M. Uchimaya, J. Comput.-Aided Mol.
Des. 2007, 21, 681–691.
[34] GastroPlus version 8.0.0002.
[35] Marvin pKa, version 5.7.0.
[10] M. Morgenthaler, E. Schweizer, A. Hoffmann-Rçder, F. Benini, R. E.
Martin, G. Jaeschke, B. Wagner, H. Fischer, S. Bendels, D. Zimmerli, J.
Schneider, F. Diederich, M. Kansy, K. Mꢃller, ChemMedChem 2007, 2,
1100–1115.
[11] H. Hilpert, W. Guba, T. J. Woltering, W. Wostl, E. Pinard, H. Mauser, A. V.
Mayweg, M. Rogers-Evans, R. Humm, D. Krummenacher, T. Muser, C.
Schnider, H. Jacobsen, L. Ozmen, A. Bergadano, D. W. Banner, R. Hoch-
strasser, A. Kuglstatter, P. David-Pierson, H. Fisher, A. Polara, R. Narqui-
zian, J. Med. Chem. 2013, 56, 3980–3995.
[36] pKcalc 4.1.0; CompuDrug. F. Csizmadia, J. Szegezdi, F. Darvas, Expert
System Approach for Predicting pKa in Trends in QSAR and Modeling, Pro-
ceedings of the 9th European Symposium on Structure-Activity Relation-
ships: QSAR and Molecular Modeling, Sep 7–11, 1992, Strasbourg,
(Ed.: C. G. Wermuth), ESCOM, Leiden, 1993, 507–510.
[12] a) L. Somekh, A. Shanzer, J. Am. Chem. Soc. 1982, 104, 5836–5837; b) C.
Ye, J. M. Shreeve, J. Fluorine Chem. 2004, 125, 1869–1872; For reviews
see: c) J. Cossy, D. Gomez Pardo, Chemtracts 2002, 15, 579–605; d) H.
Ferret, I. Dꢀchamps, D. Gomez Pardo, L. Van Hijfte, J. Cossy, ARKIVOC
(Gainesville, FL, U.S.) 2010, 8, 126–159; e) D. D. Tynoshenko, ARKIVOC
(Gainesville, FL, U.S.) 2011, 1, 329–345;.
[13] a) I. Dꢀchamps, D. Gomez Pardo, J. Cossy, Synlett 2007, 263–267; b) I.
Dꢀchamps, D. Gomez Pardo, J. Cossy, Eur. J. Org. Chem. 2007, 4224–
4234; c) B. Anxionnat, B. Robert, P. George, G. Ricci, M.-A. Perrin, D. Go-
mez Pardo, J. Cossy, J. Org. Chem. 2012, 77, 6087–6099.
[14] The absolute and relative stereochemistry in compounds A2i and A2l
were confirmed by X-ray diffraction analysis. CCDC-928423 (A2i), CCDC-
928422 (A2l), CCDC-928419 (B2d), CCDC-928420 (B2g), CCDC-928421
(B2k), and CCDC-928418 (ent-B3b) contain the supplementary crystal-
lographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[37] StarDrop version 5.1 Optibrium 2011.
[38] It should be noted that due to the homogenicity of the sample and the
fact that some of the used versions are not the latest ones, this exercise
is not to be construed as a comparative evaluation of the programs in
general. For this purpose, see, for example, a) J. C. Dearden, M. T. D.
Cronin, D. C. Lappin, J. Pharm. Pharmacol. 2007, 59 (Suppl. 1) A-7; b) M.
Meloun, S. Bordovskꢄ, Anal. Bioanal. Chem. 2007, 389, 1267–1281;
c) G. T. Balogh, B. Gyarmati, B. Nagy, L. Molnꢄr, G. M. Keseru˝, QSAR
Comb. Sci. 2009, 28, 1148–1155; d) C. Liao, M. C. Nicklaus, J. Chem. Inf.
Model. 2009, 49, 2801–2812; e) A. C. Lee, G. M. Crippen, J. Chem. Inf.
Model. 2009, 49, 2013–2033; f) J. Manchester, G. Walkup, O. Rivin, Z.
You, J. Chem. Inf. Model. 2010, 50, 565–571; g) M. Rupp, R. Kçrner, I. V.
Tetko, Comb. Chem. High Throughput Screening 2011, 14, 307–327.
[39] R. A. Scherrer, S. M. Howard, J. Med. Chem. 1977, 20, 53–58.
[40] S. K. Bhal: “Importance of Training Predictors”, ACD/Labs Application
Note, to be found at http://www.acdlabs.com/download/app/admet/
appnote_trainingpred.pdf 2010.
[15] A. Cochi, B. Burger, C. Navarro, D. Gomez Pardo, J. Cossy, Y. Zhao, T.
Cohen, Synlett 2009, 2157–2161, and references therein.
[16] Prolinols 15 were previously prepared from 4 after reduction by using
LiAlH4.
[41] J. Munoz-Muriedas: “logD and pKa Enhancements: Applications in
ADME Profiling”, 5th Annual ACD/Labs UK Users’ Meeting, Old Windsor
(UK), May 23, 2007; http://www.acdlabs.com/download/publ/2007/
ukum2007_adme.pdf.
[17] The absolute and relative stereochemistry in compounds B2d, B2g,
and B2k were confirmed by using X-ray diffraction analysis.[14b]
[18] The absolute configuration in the enantiomer of compound B3b was
confirmed by using X-ray diffraction analysis.[14b]
[42] The average values of log P do not take into account all o-NO2-substi-
tuted compounds due to their atypical behavior.
Received: June 24, 2013
Revised: November 6, 2013
Published online on February 13, 2014
[19] S. Purser, P. R. Moore, S. Swallow, V. Gouverneur, Chem. Soc. Rev. 2008,
37, 320–330.
[20] W. K. Hagmann, J. Med. Chem. 2008, 51, 4359–4369.
Chem. Eur. J. 2014, 20, 3813 – 3824
www.chemeurj.org
3824
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim