Novel chiral selector based on mefloquine - A comparative NMR study to elucidate intermolecular interactions with acidic chiral selectands

Article abstract of DOI:10.1002/chir.22029

The synthesis, ab initio calculations, and a comparative nuclear magnetic resonance study of a novel chiral mefloquine-based selector (SO) are presented. On a series of variously N-acyl protected leucine selectands (SAs), a feasibility study of mefloquine carbamate as a basic chiral solvating agent, and potential fluorophilic high-performance liquid chromatography selector has been undertaken and evaluated. An analogy is drawn between the new SO and tert-butylcarbamoyl quinidine as a reference. Chirality 24:936-943, 2012. 2012 Wiley Periodicals, Inc. Copyright

Full text of DOI:10.1002/chir.22029

CHIRALITY (2012)
Novel Chiral Selector Based on Mefloquine – A Comparative NMR
Study to Elucidate Intermolecular Interactions with Acidic
Chiral Selectands
2
1
3
MICHAL KOHOUT,1, HANSPETER KÄHLIG, DENISE WOLRAB, ALEXANDER ROLLER, AND WOLFGANG LINDNER1
*
1Institute of Analytical Chemistry, University of Vienna, 1090 Vienna, Austria
2Institute of Organic Chemistry, University of Vienna, 1090 Vienna, Austria
3Institute of Inorganic Chemistry, University of Vienna, 1090 Vienna, Austria
ABSTRACT
The synthesis, ab initio calculations, and a comparative nuclear magnetic reso-
nance study of a novel chiral mefloquine-based selector (SO) are presented. On a series of
variously N-acyl protected leucine selectands (SAs), a feasibility study of mefloquine carbamate
as a basic chiral solvating agent, and potential fluorophilic high-performance liquid chromatogra-
phy selector has been undertaken and evaluated. An analogy is drawn between the new SO and
tert-butylcarbamoyl quinidine as a reference. Chirality 00:000-000, 2012. © 2012 Wiley Periodicals, Inc.
KEY WORDS: NMR; enantiodiscrimination; chiral solvating agents; chiral recognition;
carbamoylmefloquine; cinchona alkaloids
INTRODUCTION
p-basic QN- or QD-based SO in high-performance liquid
chromatography (HPLC) was achieved for strongly p-acidic
(S)-N-3,5-dinitrobenzoylleucine and (R)-N-3,5-dinitrobenzoyl-
leucine, respectively. Following the chemical and functional
reciprocity principle, we envisioned that our new selector
(9 S;10 R)-tert-butylcarbamoyl-N-allyl mefloquine (tBuCMFQ)
should follow similar intermolecular recognition pathway, as
the tert-butylcarbamoyl quinidine (tBuCQD) type chiral SO
but being driven by an ambivalent p–p stacking ability with a
p-basic or p-acidic counterpart of the SA. The concept of
ambivalence was previously described for nucleophilic–
electrophilic potential of organic molecules.14 Here we refer to
the “p-neutral” character of the quinoline core of MFQ, which
can thus interact with both p-acidic and p-basic types of SAs.
Using a set of acidic SAs with different electron density
distribution on the aromatic moiety, we aimed to prove the
validity of this concept. We synthesized a series of reference
Chromatographic separation of racemic mixtures on chiral
stationary phases (CSPs) has developed from purely analytical
into a widely accepted industrial production tool.1 The
systematic evolution of novel chiral selectors has provided a
wide range of CSPs that enable to date the separation of a great
variety of racemic mixtures of interest. In the last years, our
group has introduced several chiral anion exchanger type
SOs and CSPs based on cinchona alkaloids, which show
enantioseparation capabilities for a broad range of chiral acidic
compounds.2–4 This successful pattern prompted us to
synthesize a novel selector based on mefloquine (MFQ), which
is structurally related to the cinchona motif, in particular to
quinine (QN) and quinidine (QD). Mefloquine is a well known
and most widely used antimalarial drug,5 and because of its
aminoethanol-like constitution, it has been screened for various
pharmacological activity.6,7 As can be seen in Figure 1, the
structure of MFQ is derived from cinchona alkaloids. The
quinoline ring of MFQ is substituted by two electron-
withdrawing groups instead of the electron-donating methoxy
group, and the quinuclidine moiety of cinchonans is replaced
by a less bulky piperidine ring. Although the strongly
deactivating effect of trifluormethyl group is well recognized,
in aromatic systems, a strong positive resonance effect was
observed.8 In consequence, this +R effect reduces the p-acidity
of the quinoline core of MFQ, which is therefore only weakly
p-acidic or “p-neutral”. Thus, we assumed that MFQ would
sufficiently interact with chiral p-acidic and p-basic analytes
(SAs), and the trifluoromethyl groups would provide additional
fluorophilic potential in the course of diastereoselective SO–SA
intermediate associates primarily driven by ion-pair formation.
Recently, several detailed nuclear magnetic resonance
(NMR) studies concerning the specific interactions of p-basic
QN- or QD-based chiral selectors (SOs) and negatively charged
chiral selectands (SAs) have been published.9–13 Conceptually,
the intermolecular interactions and the molecular recognition
principle of the diastereomeric (R)-SO-(R)-SA and (R)-SO-
(S)-SA associates are based on concerted electrostatic, p–p,
steric, and hydrogen bonding interactions leading to
well-distinguishable entities. The best separation using a
© 2012 Wiley Periodicals, Inc
1
compounds for the H NMR investigations of diastereomer
SO–SA mixtures and compared the nonequivalences
(differences between the chemical shifts of two enantiomers
in the presence of a chiral auxiliary) induced by tBuCMFQ
and tBuCQD. We also investigated the complexation-induced
chemical shifts (CISs) of respective SOs in the presence of
enantiomerically pure SAs and compared the strength of
complexation with results obtained in HPLC measurements.
A similar methodology has already been used for comparison
of the chiral discriminating properties of cyclodextrine-based
selectors in reversed phase (RP) HPLC and NMR.15
MATERIALS AND METHODS
General Methods
Nuclear magnetic resonance measurements were performed on Avance
DRX NMR spectrometers (Bruker BioSpin, Rheinstetten, Germany)
*Correspondence to: Michal Kohout, Institute of Analytical Chemistry,
University of Vienna, 1090 Vienna, Austria. E-mail: Michal.Kohout@univie.ac.at
Received for publication 17 November 2011; Accepted 3 February 2011
DOI: 10.1002/chir.22029
Published online in Wiley Online Library
(wileyonlinelibrary.com).

KOHOUT ET AL.
performed with graphite-monochromated Mo Ka radiation, l = 0.71073Å
at 100(2)K. The single crystal was positioned at 35mm from the detector,
and 691 frames were measured, each for 30 sec over 1ꢀ scan width. The data
were processed using Bruker’s software package.16 The structure was
solved by direct methods and refined by full-matrix least-squares techni-
ques. Nonhydrogen atoms were refined with anisotropic displacement para-
meters. Hydrogen atoms were placed in geometrically calculated positions
and refined as riding atoms in the subsequent least-squares model refine-
ments. The isotropic thermal parameters were estimated to be 1.2 times
the values of the equivalent isotropic thermal parameters of the atoms to
which hydrogens were bonded. For structure solution SHELXS-97,17 for
refinement SHELXL-97,18 and for molecular diagram, Crystals (University
of Oxford) were used.
Fig. 1. Selectors and chiral solvating agents, respectively, used in this study.
Solvents and Reagents
Racemic MFQ hydrochloride was obtained from Kreamer&Martin
Pharma Handels – GmbH. Leucine was purchased from Bachem.
Dimethylformamide, allyl bromide, tert-butylisocyanate, tartaric acid, and
substituted benzoic acids were recieved from Sigma-Aldrich. Variously
protected leucine derivatives were prepared according to procedures
described in the literature.19,20 Solvents were purchased from VWR Austria;
the solvents labeled as dry were distilled from P2O5 before use.
operating at 600.13 and 400.13MHz for 1H, 150.92 and 100.62MHz for 13C,
and 564.62 and 376.45MHz for 19F. Chemical shifts are referenced internal
to the residual, nondeuterated solvent signal for H (CDCl3: d = 7.26 ppm;
1
CD3OD: d = 3.31 ppm), or to the carbon signal of the solvent for 13C (CDCl3:
d = 77.00ppm; CD3OD: d = 49.00ppm), or external to CFCl3 (d = 0 ppm) for
19F. The two-dimensional NMR spectra were obtained by using standard
sequences as supplied by the manufacturer. The gradient-enhanced
correlation spectroscopy (ge-COSY) experiments were carried out with the
minimal spectral width required; 128 increments of eight scans and 2 K data
points were obtained. The two-dimensional nuclear Overhauser effect
spectroscopy (NOESY) spectra were acquired in 2 K data points by using
eight scans of 128 increments each, with a mixing time of 0.8 sec. The
gradient-enhanced heteronuclear single quantum correlation (ge-HSQC)
and gradient-enhanced heteronuclear multiple bond correlation (ge-HMBC)
experiments were performed with the minimum spectral width required
in F1 and F2 domain in 2 K data points by using 16 and 64 scans of 128
increments, respectively.
Synthetic Procedures
The synthetic route, numbering of prepared compounds and assign-
ment of atoms are depicted in Figures 2 and 3.
rac-N-Allyl mefloquine (1).
A suspension of MFQ hydrochloride
(10.0 g, 24.1 mmol) and K2CO3 (3.33 g, 24.1 mmol) in dry dimethylformamide
(60 ml) was stirred at room temperature for 15min. Allyl bromide
(2.09ml, 24.1mmol) was added and the reaction mixture was stirred at
room temperature for 17 h, diluted with water (150 ml) and stirred further
for 30min. The precipitate was filtered, washed with methanol, and dried
in vacuum. The crude product was purified by crystallization from metha-
nol. It was obtained 6.94 g (69%) of racemic 1, mp 125–126 ꢀC, mp lit.6
121–123 ꢀC.
High-performance liquid chromatography measurements were per-
formed on an 1100 Series HPLC system (Agilent Technologies, Waldbronn,
Germany) consisting of a solvent degasser, a pump, an autosampler, a
column thermostat, and a UV–vis detector. Data acquisition and analysis
were accomplished with ChemStation chromatographic data software from
Agilent Technologies. Liquid chromatography–mass spectrometry analysis
(9 R;10 S)-N-Allyl mefloquine (+)-O,O-diacetyl-L-tartaric acid
monoester (3).
In an oven-dried and nitrogen-flushed flask, N-allyl
mefloquine (2.50 g, 5.98 mmol) was dissolved in dry dichloromethane
(60 ml). Freshly prepared and dried (+)-O,O-diacetyl-L-tartaric acid
anhydride (2.00 g, 9.34 mmol) was added in the nitrogen counter stream,
and the reaction mixture was stirred overnight. It was decomposed with
5% aq. NaHCO3 (100 ml), stirred for 5 min, and acidified with 2 M aq.
HCl to pH = 2.5. The layers were separated, and the aqueous layer was
extracted with dichloromethane (2 ꢁ 25 ml). The combined organic
was performed on
a 4000 Q TRAP LC/MS/MS system (Applied
Biosystems/MDS Sciex, Foster City, USA) equipped with a standard
electrospray source and coupled with a 1200 HPLC system (Agilent
Technologies). Chromatographic purifications were performed on silica
gel Kieselgel 60 (Merck, Darmstadt, Germany). Melting points were mea-
sured using a Leica Gallen III system (Reichert). X-ray analysis was
Fig. 2. The assignment of atoms of tBuCMFQ and analytes used for NMR measurements.
Chirality DOI 10.1002/chir

NOVEL CHIRAL SELECTOR BASED ON MEFLOQUINE
NMR (400.13MHz, CDCl3) d 1.10 (1H, m); 1.23 (1H, m); 1.80 (2H, m);
2.02 (4H, bs); 2.12 (4H, bs); 2.72 (1H, m); 3.07 (1H, m); 3.76 (1H, m); 4.06
(2H, m); 5.75 (2H, CH2CH, m); 5.85 (1H, d, J = 2.0 Hz); 6.18 (1H, d,
J = 2.0 Hz); 6.30 (1H, CHCH2, m); 7.01 (1H, H9, d, J= 2.0Hz); 7.57 (1H, H3,
s); 7.81 (1H, H6, J = 8.0 Hz); 8.22 (1H, H7, s); 8.24 (1H, H5, s). 13C NMR
(100.62 MHz, CDCl3) d 170.8 (CO), 170.3 (CO), 169.1 (CO), 166.6 (CO),
148.3 (Cq), 148.0 (Cq), 144.8 (Cq), 144.0 (Cq), 130.3 (Cq), 130.0 (Cq), 129.5
(CH), 128.2 (CHCH2), 128.1 (CH), 126.3 (CH), 125.6 (Cq), 124.5 (CH2CH),
115.0 (CH), 71.9 (CH), 71.6 (CH), 69.7 (CH), 62.9 (CH), 54.8 (CH2), 54.2
(CH2), 24.1 (CH2), 23.1 (CH2), 23.0 (CH2), 20.5 (CH3), 20.4 (CH3). 19F
NMR (376.45 MHz, CDCl3) d ꢂ60.8 (CF3), ꢂ68.3 (CF3). MS: calculated
C28H28F6N2O8 (634.2), found [M + H]+ = 635.1.
(9 S;10 R)-N-Allyl mefloquine (4).
To a solution of 2 (0.90 g,
1.42 mmol) in dichloromethane (20 ml), an aqueous solution of K2CO3
(15 ml, pH = 10) was added, and the reaction mixture was vigorously
stirred at room temperature for 18 h. The separated aqueous solution
was extracted with dichloromethane (2 ꢁ 40 ml), and the combined
organic solution was dried with anhydrous MgSO4. The solvent was
evaporated, and the crude product was purified by column chromatogra-
phy (eluent dichloromethane/methanol, 10/1) to yield 0.53 g (91%) of
enantiomerically pure 4, mp 38–40 ꢀC. 1H NMR (400.13 MHz, CD3OD,
CD3COOD) d 1.03 (1H, m); 1.26 (1H, m); 1.68–1.1.88 (4H, m); 3.16
(1H, dt, 2J = 12.4 Hz, 3J = 3.8 Hz); 3.58 (1H, m); 3.71 (1H, dt, 3J = 12.1 Hz,
3J = 3.1 Hz); 4.17 (1H, dd, 2J = 14.1 Hz, 3J = 9.0 Hz); 4.33 (1H, dd,
2J = 13.8 Hz, 3J = 5.5 Hz); 5.79 (2H, CH2 = CH, m); 6.25 (1 H, CH = CH2,
m); 6.26 (1H, H9, d, 3J = 3.4 Hz); 7.94 (1H, H6, t, 3J = 8.2 Hz); 8.22 (1H,
H3, s); 8.32 (1H, H7, d, 3J = 7.2 Hz); 8.39 (1H, H5, d, 3J = 8.6 Hz). 13C
NMR (100.62 MHz, CD3OD, CD3COOD) d 151.5 (Cq), 149.4 (Cq), 149.1
(Cq), 144.9 (Cq), 130.6 (CH), 130.2 (Cq), 129.3 (CH), 128.8 (CH), 128.5
(CH), 127.4 (Cq), 126.3 (CH2CH), 123.6 (Cq), 117.1 (CH), 66.5 (CH),
65.9 (CH), 56.4 (CH2), 54.0 (CH2), 24.2 (CH2), 23.9 (CH2), 22.7 (CH2).
19F NMR (376.45 MHz, CD3OD, CD3COOD) d ꢂ62.0 (CF3), ꢂ69.8
(CF3). MS: calculated for C20H20F6N2O (418.2), found [M + H]+ = 419.4.
Analogously, (9 R;10 S)-N-allyl mefloquine (5) was prepared and 0.50 g
(85%) of 5 was obtained, mp 39–40 ꢀC. 1H NMR (400.13 MHz, CD3OD,
CD3COOD) d 0.92 (1H, m); 1.15 (1H, m); 1.64–1.76 (4H, m); 2.82
(1H, dt, 2J = 11.6 Hz, 3J = 3.9 Hz); 3.27 (1H, m); 3.35 (1H, m); 3.86 (1H,
dd, 2J = 14.2 Hz, 3J = 8.7 Hz); 4.08 (1H, dd, 2J = 14.2 Hz, 3J = 5.5 Hz); 5.61
(2H, CH2CH, dd, JcisH = 10.1 Hz, JtransH = 17.0 Hz); 6.16 (1H, H9, d,
3J = 3.3 Hz); 6.23 (1H, CHCH2, m); 7.90 (1H, H3, t, 3J = 8.1 Hz); 8.21 (1H.
H3, s); 8.27 (1H, H7, d, 3J = 7.3 Hz); 8.42 (1H, H5, d, 3J = 8.5 Hz). 13C
NMR (100.62 MHz, CD3OD, CD3COOD) d 153.0 (Cq), 149.3 (Cq), 148.9
(Cq), 144.8 (Cq), 132.0 (CHCH2), 130.4 (CH), 129.1 (CH), 128.9 (CH),
127.7 (Cq), 126.4 (Cq), 123.7 (Cq); 123.0 (CH2CH), 117.0 (CH), 67.4
(CH), 65.3 (CH), 56.9 (CH2), 54.0 (CH2), 24.7 (CH2), 24.5 (CH2), 23.6
(CH2). 19F NMR (376.45 MHz, CD3OD, CD3COOD) d ꢂ62.0 (CF3),
ꢂ69.8 (CF3). MS: calculated for C20H20F6N2O (418.2), found
[M + H]+ = 419.3.
Fig. 3. The synthetic route for the preparation of (9 S;10 R)-tert-butylcarbamoyl-
N-allyl mefloquine (6); (a), (b) (R,R)-DATAAN, CH2Cl2, r.t.; (c) K2CO3, CH2Cl2/
H2O; (d) tert-butylisocyanate, dibutyltin dilaurate, toluene.
solution was washed with water (2 ꢁ 25 ml), brine (25 ml), and dried with
anhydrous MgSO4. The solvent was evaporated; the crude product was
placed in a Soxhlet extractor and extracted with hot toluene (50 ml) for
1 h. The resulting solid was dissolved in dichloromethane, the solvent
was evaporated, and the extraction procedure was repeated. The
product 2 was harvested as a white crystalline powder, 1.48 g (78%), mp
195.5–197 ꢀC. 1H NMR (400.13 MHz, CDCl3) d 1.12 (1H, m); 1.33 (1H,
m); 1.77 (3H, m); 2.08 (3H, CH3, s); 2.15 (4H, bs); 2.80 (1H, m); 3.12
(1H, m); 3.96 (1H, bs); 4.24 (1H, m); 4.36 (1H, bs); 5.83 (3H, m); 6.01
(1H, bs); 6.27 (1H, CHCH2, m); 6.67 (1H, m); 7.78 (1H, H6, t, J = 7.8 Hz);
8.12 (1H, H3, s); 8.21 (1H, s); 8.23 (1H, s). 13C NMR (100.62 MHz, CDCl3)
d 171.4 (CO), 170.8 (CO), 166.6 (CO), 165.9 (CO), 144.4 (Cq), 142.4 (Cq),
129.7 (CH), 128.8 (CH), 126.1 (CH), 125.8 (CH), 125.4 (Cq), 117.0 (CH),
71.7 (CH), 70.9 (CH), 69.4 (CH), 63.4 (CH), 56.2 (CH2), 55.6 (CH2),
24.4 (CH2), 22.9 (CH2), 22.6 (CH2), 21.1 (CH3), 20.4 (CH3). 19F NMR
(376.45 MHz, CDCl3) d ꢂ61.5 (CF3), ꢂ69.3 (CF3). MS: calculated for
C28H28F6N2O8 (634.2), found [M + H+] = 635.3.
(9 S;10 R)-tert-Butylcarbamoyl-N-allyl mefloquine (6).
To a
solution of (0.50 g, 1.20 mmol) and tert-butylisocyanate (0.34 ml,
4
2.99 mmol) in dry toluene (25 ml), dibutyltin dilaurate (0.02 ml) was
added and the reaction mixture was heated to 90 ꢀC for 24 h. The solvent
was removed, the crude product was dried in vacuum at 50 ꢀC, and the
reaction procedure was repeated under the conditions given previously.
The solvent was removed under reduced pressure, and the crude product
was purified by column chromatography (eluent CH2Cl2/MeOH, 20/1).
It was obtained 0.52 g (84%) of compound 6, mp 49.5–50.5 ꢀC. 1H NMR
(600.13 MHz, CD3OD) d 1.10 (1H, H15a, m) 1.12 (1H, H14b, m); 1.29
(9H, (CH3)3, s); 1.59 (2H, H13a,b, m); 1.66 (1H, H15a, m); 1.71 (1H, H14a
,
m) 2,50 (1H, H12a, dt, 2J = 12.1 Hz, 3J = 13.1 Hz); 2.89 (1H, H10, m); 3.10
(1H, H12b, dt, 2J = 12.1 Hz, 3J = 3.9 Hz); 3.44 (1H, H18b, dd, 2J = 14.2 Hz,
3J = 8.2 Hz); 3.69 (1H, H18a dd, 2J = 14.2 Hz, 3J = 4.6 Hz); 5.38 (1H,
,
(9 S;10 R)-N-Allyl mefloquine (+)-O,O-diacetyl-L-tartaric acid
H20b, d, 3J = 10.1 Hz); 5.44 (1H, H20a, d, 3J = 17.2 Hz); 6.09 (1H, H19, dddd,
3
3
3
monoester (2).
The combined organic solution from the extraction
3J19-20a = 17.2, J19-20b = 10.1, J19-18b = 8.2, J19-18a = 4.6); 6.86 (1H, H9, d,
3
3
was evaporated under reduced pressure, and the crude product was
3J = 3.8 Hz); 7.90 (1H, H6, t, J6-5 = 8.5 Hz, J6-7 = 7.3 Hz); 7.94 (1H, H3, s);
8.29 (1H, H7, d, 3J = 7.3 Hz); 8.52 (1H, H5, d, 3J = 8.5 Hz). 13C NMR
(150.92 MHz, CD3OD) d 155.6 (CO); 151.3 (C4); 148.8 (C2); 144.9 (C8a);
purified by column chromatography (eluent dichloromethane/2-propanol,
1
5/1). It was obtained 1.23 g (65%) of monoester 3, mp 125.5–127.5ꢀC. H
Chirality DOI 10.1002/chir

KOHOUT ET AL.
135.3 (C19); 130.7 (C7); 130.2 (C8); 129.4 (C5); 129.0 (C6); 127.7 (C4a);
Assignment of Absolute Configuration of Mefloquine
125.0 (C17); 122.7 (C16); 120.0 (C20); 116.4 (C3); 70.4 (C9); 63.5 (C10);
57.8 (C18); 54.2 (C12); 51.3 (C25); 28.9 (C26); 25.8 (C15); 25.6 (C13); 24.6
(C14). 19F NMR (564.62 MHz, CD3OD): ꢂ62.1(CF3–C17); ꢂ70.0(CF3–
C16). MS: calculated for C25H29F6N3O2 (517.2), found [M + H]+ = 518.0.
Derivatives
Although the absolute configuration of MFQ is described in
the literature,22,23 the data correlating the elution order of the
respective enantiomers with their absolute configuration are
not available for the N-allyl derivative. Therefore, we decided
to use the similarity between cinchona alkaloids and
MFQ. We employed (+)-O,O-dibenzoyl-L-tartaric anhydride
derivatized QN and QD as reference compounds in an RP
HPLC measurement (Fig. 4). With these results, we were able
to assign indirectly via a chromatographic elution order of the
diastereomers the absolute configuration of the less retained
diastereomer as (9 S;10 R)-N-allyl mefloquine (+)-O,O-
dibenzoyl-L-tartric acid monoester and of the more retained
diastereomer as (9 R;10 S)-N-allyl mefloquine (+)-O,O-
dibenzoyl-L-tartric acid monoester. We also prepared a
cocrystal of (9 R;10 S)-N-allyl mefloquine with monomethyl
(+)-O,O-diacetyl-L-tartrate and performed an X-ray analysis
(Fig. 5) that confirmed the result obtained by the HPLC
measurement.
(R)-N-3,5-Bis(trifluormethyl)benzoylleucine (R-BTFMB-Leu).
To a solution of (R)-leucine (0.13 g; 0.97mmol) and (iPr)2NEt (0.17ml;
0.97mmol) in 50% aqueous acetonitrile (20ml), O-succinimidyl-3,5-bis
(trifluoromethyl)benzoate (0.30 g; 0.97mmol) was added portionwise. The
reaction mixture was stirred and heated to 40ꢀC over night, acidified with
2 M aq. HCl to pH= 2, diluted with a saturated solution of NaCl (100ml)
and extracted with ethyl acetate (10ꢁ 20ml). The combined organic
solution was dried with anhydrous MgSO4. The solvent was evaporated,
and the crude product (R-BTFMB-Leu) was purified by column
chromatography (eluent dichloromethane/methanol, 5/1). It was obtained
0.21g (66%) of white crystalline product, m.p. 169.5–172.5 ꢀC. 1H NMR
(400.13 MHz, CD3OD) d 1.00 (6H, 2 ꢁ CH3, m); 1.79 (3H, CH2, CH, m);
4.68 (1H, H101, m), 8.16 (1H, CH, s), 8.48 (2H, 2 ꢁ CH2, s). 13 C NMR
(100.62 MHz, CD3OD) d 176.2 (C= O); 166.9 (C= O); 137.9 (Cq); 129.2
(CH); 126.0 (CH); 123.3 (Cq); 53.3 (CH); 41.4 (CH2); 26.4 (CH); 23.4
(CH3); 21.7 (CH3). 19 F NMR (376.45MHz, CD3OD) d ꢂ64.9 (2ꢁ CF3).
MS: calculated for C15H15F6NO3 (371.1), found [M + H]+ = 372.2.
Analogously, (S)-N-3,5-bis(trifluormethyl)benzoylleucine (S-BTFMB-Leu)
was prepared, mp 168–171 ꢀC. 1H NMR (400.13 MHz, CD3OD) d 0.96
(6H, 2 ꢁ CH3, m); 1.80 (3H, m); 4.68 (1H, H101, m); 8.15 (1H, CH, s);
8.46 (2H, 2 ꢁ CH, s). 13C NMR (100.62 MHz, CD3OD) d 176.0 (CO);
166.8 (CO); 137.8 (Cq); 129.2 (CH); 126.0 (CH); 123.3 (Cq); 53.9 (CH);
41.4 (CH2); 26.3 (CH); 23.4 (CH3); 21.6 (CH3). 19F NMR (376.45 MHz,
CD3OD) d ꢂ64.8 (2 ꢁ CF3). MS: calculated for C15H15F6NO3 (371.1),
found [M + H]+ = 372.0.
Ab initio Calculations
Having the confirmed starting structure of (9 R;10 S)-N-allyl
mefloquine, we decided to perform the ab initio calculations
for (9 R;10 S)-tert-butylcarbamoyl-N-allyl mefloquine. The
structure of 9-O-tert-butylcarbamoyl QN was confirmed in a
previous study,9 in which it was shown that QN-based selector
exhibited remarkable complexation properties to (S)-N-3,5-
dinitrobenzoylleucine. This behavior is strongly influenced
by pꢂp stacking of the aromatic moieties of respective
compounds. To support this assumption and the reciprocity
principle with MFQ-based selector, we calculated electron
density of selected SO–SA couples (Fig. 6). Optimization to
minimum energy was carried out in Gaussian 03 on DFT
B3LYP 6–31 g(d) level.24 Because protonation realizes
selectively on the piperidine or quinuclidine ring of the
RESULTS AND DISCUSSION
Synthesis
The separation of (rac)-1 by using (+)-O,O-diacetyl-L-tartaric
acid anhydride was performed using a method described
in the literature.21 It was monitored with RP HPLC, using a
Kinetex RP C-18 (150 ꢁ 4.6 mm, 5 mm) column. Each
diastereomer was obtained in diastereomeric excess > 99%.
Fig. 4. Reversed phase high-performance liquid chromatography measurement of (+)-O,O-dibenzoyl-L-tartaric monoesters: hydrolyzed (+)-O,O-dibenzoyl-L-
tartaric anhydride (1) and respective monoesters of quinine (2), quinidine (3), and rac-N-allyl-mefloquine (4).
Chirality DOI 10.1002/chir

NOVEL CHIRAL SELECTOR BASED ON MEFLOQUINE
also spatial interactions play an important part in the molecular
recognition process. This is clearly documented by the so-
called “Linker-Leu” derivative. It is not as p-acidic as DNB-Leu
or DCB-Leu, but it is still very well enantioseparated on a
p-basic QD-AX. Surprisingly high a-value was obtained for
TMB-Leu, although it was less retained, which we designed
as a strongly p-basic SA for the MFQ selector. As shown in
Figure 6, TMB-Leu is a very spatially demanding molecule,
and thus, p–p interactions are suppressed; however, no
change of elution order was noticed. The stericaly preferred
enantiomer can form with the selector a strong complex based
on London dispersion forces, which leads to an effective
enantioseparation with a-value of 5.5. Such interactions are
not affordable to Bz-Leu, which can be, in addition, considered
as a p-neutral molecule. Although it possesses a highly p-acidic
aromatic core, the lowest a-value was found for the resolution of
PFB-Leu. Overall, it was also the weakly retained compound,
similar to (S)-BTFMB-Leu, (S)-Bz-Leu and (S)-TMB-Leu.
Fig. 5. X-ray structure of a cocrystal of (9 R;10 S)-N-allyl mefloquine with
(+)-O,O-diacetyl-L-tartaric acid monomethyl ester (CCDC 850684).
NMR Spectroscopy
A full characterization of the MFQ selector 6 (Fig. 3) as well
as the selectands (Fig. 7) was performed by comparative
analysis of two-dimensional spectra of homonuclear and
heteronuclear correlations. On the basis of a previous study,9
the conformation of tBuCMFQ can be described as anti-open-
like. Analogously to anti-open conformation of cinchonans,
the most pronounced nuclear Overhauser effects were found
between protons H9–H10 and H5–H10. Because of higher
flexibility and a lower space filling ability of the piperidine ring,
no other correlation between an aromatic proton and the
aliphatic ring was obtained.
Similarly to cinchona alkaloids, tert-butylcarbamoyl-N-allyl
mefloquine (SO1) is a multifunctional compound able to form
higher order aggregates.26 To eliminate this phenomenon, we
performed all our measurements with diluted samples
(20 mM) in methanol-d4. This solvent should simulate the HPLC
measurement conditions in a polar organic mode. The binding
stoichiometry between tBuCMFQ and both enantiomers of
TMB-Leu as the potentially best p-basic selectand was
determined by a continuous variation titration protocol.27,28
The most pronounced CISs were found for the benzylic proton
H9 with the maximum value at TMB-Leu molfraction of 0.5
for the more stable tBuCMFQ-(R)-TMB-Leu and the less
stable tBuCMFQ-(S)-TMB-Leu complex. The corresponding
Job plot is depicted in Figure 8.
respective selectors (vide infra), we assume that the aromatic
core would be affected by this N-protonation to a much lesser
extent than the aliphatic part of the molecule, and thus, free-
base forms of the selectors were used for calculations.
As can be seen in Figure 6, the p–p stacking between QN
selector and (S)-DNB-Leu is obviously favored because a
strong p-base (SO) interacts with a strong p-acid (SA). As
envisioned, the situation is different for MFQ selector, which
should be a weak p-acid. The two trifluoromethyl electron-
withdrawing groups represent the electron-rich part of the
molecule, whereas the aromatic system is neither electron-rich
nor electron-deficient. The selectand, which we designed as
the strong p-base for effective p–p interaction with MFQ, is
essentially p-basic but also spatially demanding. Consequently,
this feature can negatively influence its p–p stacking abilities.
HPLC Determination of the Differential Free Binding
Energies ΔR;S (ΔG)
In a preceding study performed in our work group,25 we
demonstrated that the enantioselectivity of cinchona carbamates
is not significantly influenced by immobilization on modified
silica. Therefore, nonselective retention increments or adverse
effects caused by immobilization can be neglected, and the
calculated ΔR;S ðΔGÞ data can be considered as good estimates
for ΔR;S ðΔGÞ values. Using the QD-AX column (commercially
available tBuCQD), we measured different SAs (Fig. 7) in
racemic as well as enantiomerically pure forms. From the
respective a-values, the corresponding differential free binding
energies were calculated (Table 1).
Such behavior is typical for a 1:1 complexation and indicates
that, in the first place, the piperidine nitrogen is involved in the
formation of a salt bridge (ion pair). This hypothesis is further
supported by pronounced CISs of the protons H10, H18, and
H12 that are located in the vicinity of piperidine nitrogen.
As expected, p-acidic SAs were better resolved than the
p-basic SAs. It is also obvious (Table 1) that not only p–p but
A series of equimolar mixtures of stronger complexed pure
(R)-enantiomers of SAs with each SO were prepared. The
Fig. 6. Electrostatic potential-derived charges – total electron density of (a) tBuCQN selector, (b) (S)-DNB-Leu, (c) (9 R;10 S)-tBuCMFQ, (d) (S)-TMB-Leu. Color
code: red, the highest electron density; blue, the lowest electron density. [Color figure can be viewed in the online version, which is available at wileyonlinelibrary.com.]
Chirality DOI 10.1002/chir

KOHOUT ET AL.
Fig. 7. The list of analytes used in the study.
TABLE 1. Differential free binding energies. Chiral stationary
phase: QD-AX. Mobile phase: methanol 98%, acetic acid 2%,
ammonium acetate 0.5%. All measurements were carried out
at 23 ꢀC
t0
tR
k
a
ΔG[kJ/mol]
ꢂ5.90
(R)-DNB-Leu
(S)-DNB-Leu
(R)-LINKER-Leu
(S)-LINKER-Leu
(R)-BTFMB-Leu
(S)-BTFMB-Leu
(R)-DCB-Leu
(S)-DCB-Leu
(R)-TMB-Leu
(S)-TMB-Leu
(R)-Bz-Leu
1.86
1.86
1.86
1.86
1.86
1.86
1.86
1.86
1.86
1.86
1.86
1.86
1.86
1.86
51.77
6.40
27.64
4.62
12.46
3.03
39.30
7.62
12.20
3.74
6.54
3.73
4.50
26.85
2.44
13.87
1.48
5.70
0.63
20.14
3.01
5.56
1.01
2.52
1.00
1.69
0.96
11.00
9.35
9.04
6.69
5.49
2.52
1.76
ꢂ5.50
ꢂ5.42
ꢂ4.68
ꢂ4.19
ꢂ2.28
ꢂ1.39
(S)-Bz-Leu
(R)-PFB-Leu
(S)-PFB-Leu
3.64
Fig. 8. Job plot for the diastereomeric complexes tBuCMFQ-(R)-TMB-Leu
and tBuCMFQ-(S)-TMB-Leu. The total concentration was 20 mM.
1
recorded H NMR spectra were compared with that of pure
10mM chiral substrates. The diagnostic protons H9 of
tBuCMFQ and tBuCQD exhibited the most pronounced shifts
and also clearly showed the difference between the studied
selectors.
We found a similar trend for the concentration-related
CISs of tBuCQD with (R)-SAs as we observed in HPLC
experiments (Fig. 9). With the exception of (R)-Bz-Leu and
(R)-PFB-Leu, all the SAs were complexed with a similar
strength. In this case, the complex of (R)-TMB-Leu with the
selector appears to be the strongest one. This again indicates
that spatial arrangement of SA–SO pair may be even more
important than reciprocal p–p interactions.
p-acid ((R)-Linker-Leu) as well as for the pair of p-acidic
(R)-BTFMB-Leu and p-basic (R)-TMB-Leu. (R)-PFB-Leu was
again the least complexed SA. Besides the homologues
trends of the two selectors for the given SAs, it should be
mentioned that the performance in complexation of the
strongly bound (R)-SAs was one order of magnitude different
for each other. This is remarkable and may be reflected by
the chromatographic behavior as well.
To directly compare the performance of the novel MFQ
selector, we prepared a series of racemic as well as
enantiomer-enriched mixtures of SAs with tBuCQD and
tBuCMFQ (Table 2). The establishment of binding isotherms
would be necessary for calculation of relative free energies of
binding from NMR measurements. As an approximation, we
used differences in CISs obtained at a single concentration
A similar trend was found also for the complexation of
tBuCMFQ with the SAs (Fig. 10). Analogous complexation
was found for very strong p-acid ((R)-DNB-Leu) and weaker
Chirality DOI 10.1002/chir

NOVEL CHIRAL SELECTOR BASED ON MEFLOQUINE
1
TABLE 2. H NMR (600 MHz, CD3OD, 25 ꢀC) CISs and
nonequivalences of SAs (10 mM) in the presence of equimolar
amounts of tBuCMFQ and tBuCQD
MFQ-AX
H101
QD-AX
H101
SA
d
d
DNB-Leu
ꢂ0.01(R)
ꢂ0.03(S)
0.10(R)
0.09(S)
ꢂ0.01(R)
ꢂ0.03(S)
0(R)
0.017
0.10(R)
ꢂ0.05(S)
0.20(R)
0.07(S)
0.09(R)
ꢂ0.04(S)
0.08(R)
ꢂ0.04(S)
0.09(R)
ꢂ0.01(S)
0.01(R)
ꢂ0.04(S)
0(R)
0.146
Linker-Leu
BTFMB-Leu
DCB-Leu
TMB-Leu
Bz-Leu
0.07
0.011
0
0.132
0.129
0.128
0.103
0.048
0
0(S)
0(R)
0(S)
0(R)
0
0
0(S)
PFB-Leu
0.01(R)
0(S)
0.017
0(S)
Fig. 9. Concentration dependent CISs of H9 of tBuCQD (SO2) with N-acyl
protected (R)-leucines.
Fig. 11. 1H NMR resonances of H101 of PFB-Leu: (a) without a selector, (b)
with tBuCQD, (c) with tBuCMFQ.
Fig. 10. Concentration dependent CISs of H9 of tBuCMFQ (SO1) with
N-acyl protected (R)-leucines.
For the MFQ selector, we observed by NMR a satisfactory
distinction factor of 0.017 with small low-frequency shifts for
DNB-Leu. Similar signal separation with d of 0.011 was found
for BTFMB-Leu. It has to be noted that a distinction value
of 0.017 was found for PFB-Leu; while using tBuCQD, no
enantiodiscrimination was observed (Fig. 11). That means
that the trifluoromethyl groups of MFQ are capable to support
the recognition of PFB-Leu, which is most probably realized by
SO–SA fluorophilic interactions. Fluorophilicity is defined as
the natural logarithm of the perfluoro(methylcyclohexane)/
toluene partition coefficient (P).29,30 We use this term to point
out that a molecule with a certain amount of fluorine atoms
may express enhanced affinity to another fluorine-containing
molecule.
corresponding to 1:1 SO–SA complex and relate them to
a-values obtained in HPLC. The difference of CISs of
respective enantiomers found for the racemic SAs with
tBuCQD corresponded very well to the data obtained
with HPLC (see Table 1). Such behavior is in accordance with
results reported in the literature.15
The nonequivalence values obtained for tBuCMFQ are, in
general, much lower than for tBuCQD. In addition, the
enantioseparation was observed mainly for p-acidic SAs. It
can be seen (Table 2) that there was no distinction for p-basic
TMB-Leu, although the pure enantiomers were complexed
with different strength. As we stated previously, TMB-Leu is
most probably spatially demanding, and any typical p–p type
of interaction is therefore problematic.
As it has been shown, the HPLC separation of PFB-Leu with
QD-AX is satisfactory but not excellent. However, under the
Chirality DOI 10.1002/chir

KOHOUT ET AL.
hydroxyl compounds by 19F NMR spectroscopy. Tetrahedron: Asymmetry
conditions of the NMR experiment, we did not observe any
distinction. The resolution found in NMR measurements with
tBuCMFQ as the chiral discriminating agent supports our
assumption that fluorophilic interactions of the trifluoromethyl
groups of MFQ enter the process of chiral recognition.
Assuming that we would obtain analogous selector coverage
of tBuCMFQ as for tBuCQD on the respective CSPs, we should
be able to relate the nonequivalencies obtained by NMR
measurements to the performance of the MFQ selector in
HPLC. In such a case, we can expect separation of PFB-Leu
with a > 1.76. That means that tBuCMFQ has a potential as a
selector in HPLC for separation of polyfluorinated compounds.
2005;16:1547–1555.
12. Rudziska E, Berlicki Ł, Kafarski P, Lämmerhofer M, Mucha A. Cinchona
alkaloids as privileged chiral solvating agents for the enantiodiscrimination
of N-protected aminoalkanephosphonates—a comparative NMR study.
Tetrahedron: Asymmetry 2009;20:2709–2714.
13. Uccello-Barretta G, Vanni L, Balzano F. NMR enantiodiscrimination
phenomena by quinine C9-carbamates. Eur J Org Chem 2009;2009:860–869.
14. Volovik SV. Problem of ambivalence in organic reactions. Theor Exp
Chem 1988;24:268–274.
15. Laverde A, da Conceição GJA, Queiroz SCN, Fujiwara FY, Marsaioli AJ.
An NMR tool for cyclodextrin selection in enantiomeric resolution by
high-performance liquid chromatography. Magn Reson Chem
2002;40:433–442.
16. Bruker programs SAINT-NT, version 7.68A; SADABS-2008/1; XPREP-
2008/2. Madison, WI, USA: Bruker AXS; 2011.
17. Sheldrick GM. Phase annealing in SHELX-90: direct methods for larger
structures. Acta Crystallogr A 1990;A46:467–473.
18. Sheldrick GM. A short history of SHELX. Acta Crystallogr A 2008;
A64:112–122.
19. Gavioli E, Maier NM, Haupt K, Mosbach K, Lindner W. Analyte
templating: enhancing the enantioselectivity of chiral selectors upon
incorporation into organic polymer environments. Anal Chem
2005;77:5009–5018.
20. Kurosu M, Li K. New chiral derivatizing agents: convenient determination
of absolute configurations of free amino acids by 1H NMR. Org Lett
2009;11:911–914.
21. Lindner W, Leitner C, Uray G. Liquid chromatographic separation of
enantiomeric alkanolamines via diastereomeric tartaric acid monoesters.
J Chromatogr A 1984;316:605–616.
CONCLUSION
We synthesized a novel chiral selector based on MFQ
and evaluated its properties as a chiral discriminating agent
for N-acyl-protected amino acids. Although this selector does
not exhibit expected ambivalent complexation and separation
abilities, it showed a certain affinity to polyfluorinated com-
pounds. We documented this behavior on a sufficient NMR res-
olution of rac-N-pentafluorobenzoylleucine. This feature
requires further investigations with a polyfluorinated version
of the N-allyl mefloquine carbamate type selector, which
could provide improved enantiodiscrimination of chiral
polyfluorinated and perfluorinated molecules.
22. Carroll FI, Blackwell JT. Optical isomers of aryl-2-piperidylmethanol
antimalarial agents. Preparation, optical purity, and absolute stereochemistry.
J Med Chem 1974;17:210–219.
ACKNOWLEDGMENTS
The authors thank Prof. J. Svoboda (Institute of Chemical
Technology, Prague, Czech Republic) for critical comments
on the manuscript.
23. Karle JM, Karle IL. Structure of the antimalarial (ꢃ)-mefloquine
hydrochloride. Acta Crystallogr C 1991;47:2391–2395.
24. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
Cheeseman JR, Montgomery Jr. JA, Vreven T, Kudin KN, Burant JC,
Millam JM, Iyengar SS, Tomasi J, Barone V, Mennucci B, Cossi M,
Scalmani G, Rega N, Petersson GA, Nakatsuji H, Hada M, Ehara M,
Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y,
Kitao O, Nakai H, Klene M, Li X, Knox JE, Hratchian HP, Cross JB,
Bakken V, Adamo C, Jaramillo J, Gomperts R, Stratmann RE, Yazyev O,
Austin AJ, Cammi R, Pomelli C, Ochterski JW, Ayala PY, Morokuma K,
Voth GA, Salvador P, Dannenberg JJ, Zakrzewski VG, Dapprich S,
Daniels AD, Strain MC, Farkas O, Malick DK, Rabuck AD, Raghavachari K,
Foresman JB, Ortiz JV, Cui Q, Baboul AG, Clifford S, Cioslowski J,
Stefanov BB, Liu G, Liashenko A, Piskorz P, Komaromi I, Martin RL,
Fox DJ, Keith T, Al-Laham MA, Peng CY, Nanayakkara A, Challacombe M,
Gill PMW, Johnson B, Chen W, Wong MW, Gonzalez C, Pople JA. Gaussian
03, Revision C.02. Gaussian, Inc., Wallingford CT, 2004.
LITERATURE CITED
1. Francote ER. Enantioselective chromatography as a powerful alternative for
the preparation of drug enantiomers. J Chromatogr A 2001;906:379–397.
2. Lämmerhofer M, Lindner W. Quinine and quinidine derivatives as chiral
selectors I. Brush type chiral stationary phases for high-performance liquid
chromatography based on cinchonan carbamates and their application as
chiral anion exchangers. J Chromatogr A 1996;741:33–48.
3. Lämmerhofer M, Lindner W. Liquid chromatographic enantiomer
separation and chiral recognition by cinchona alkaloid-derived
enantioselective separation materials. Adv Chromatogr 2008;46:1–107.
4. Hoffmann CV, Pell R, Lämmerhofer M, Lindner W. Synergistic effects on
enantioselectivity of zwitterionic chiral stationary phases for separations
of chiral acids, bases, and amino acids by HPLC. Anal Chem
2008;80:8780–8789.
5. Schlagenhauf P, Adamcova M, Regep L, Schaerer M, Rhein H-G. The
position of mefloquine as a 21st century malaria chemoprophylaxis.
Malaria J 2010;9:357.
25. Lah J, Maier NM, Lindner W, Vesnaver G. Thermodynamics of binding of
(R)- and (S)-dinitrobenzoyl leucine to cinchona alkaloids and their tert-
butylcarbamate derivatives in methanol: evaluation of enantioselectivity
by spectroscopic (CD, UV) and microcalorimetric (ITC) titrations. J Phys
Chem B 2001;105:1670–1678.
26. Uccello-Barretta G, Balzano F, Quintavalli C, Salvadori P. Different
enantioselective interaction pathways induced by derivatized quinines. J
Org Chem 2000;65:3596–3602.
6. Jayaprakash S, Iso Y, Wan B, Franzblau SG, Kozikowski AP. Design,
synthesis, and SAR studies of mefloquine-based ligands as potential
antituberculosis agents. ChemMedChem 2006;1:593–597.
7. Kunin CM, Ellis WY. Antimicrobial activities of mefloquine and a series of
related compounds. Antimicrob Agents Chemother 2000;44:848–852.
27. Job P. Formation and stability of inorganic complexes in solution. Ann
8. Sheppard WA. The electronic properties of fluoroalkyl groups. Fluorine
p-p interaction1. J Am Chem Soc 1965;87:2410–2420.
9. Maier NM, Schefzick S, Lombardo GM, Feliz M, Rissanen K, Lindner W,
Lipkowitz KB. Elucidation of the chiral recognition mechanism of
cinchona alkaloid carbamate-type receptors for 3,5-dinitrobenzoyl amino
acids. J Am Chem Soc 2002;124:8611–8629.
10. Uccello-Barretta G, Balzano F, Salvadori P. Rationalization of the
multireceptorial character of chiral solvating agents based on quinine
and its derivatives: Overview of selected NMR investigations. Chirality
2005;17:S243–S248.
Chim 1928;9:113–203.
28. Homer J, Perry MC. Molecular complexes. Part 18.-A nuclear magnetic
resonance adaptation of the continuous variation (job) method of
stoicheiometry determination.
1986;82:533–543.
29. Kiss LE, Kövesdi I, Rábai J. An improved design of fluorophilic molecules:
prediction of the ln P fluorous partition coefficient, fluorophilicity, using
3D QSAR descriptors and neural networks.
2001;108:95–109.
30. Mercader AG, Duchowicz PR, Sanservino MA, Fernández FM, Castro EA.
QSPR analysis of fluorophilicity for organic compounds. J Fluorine Chem
2007;128:484–492.
J
Chem Soc, Faraday Trans
1
J Fluorine Chem
11. Abid M, Török B. Cinchona alkaloid induced chiral discrimination for the
determination of the enantiomeric composition of a-trifluoromethylated-
Chirality DOI 10.1002/chir