Enantioseparation of chiral pharmaceuticals by vancomycin-bonded stationary phase and analysis of chiral recognition mechanism

Article abstract of DOI:10.1002/chir.23052

The drug chirality is attracting increasing attention because of different biological activities, metabolic pathways, and toxicities of chiral enantiomers. The chiral separation has been a great challenge. Optimized high-performance liquid chromatography (HPLC) methods based on vancomycin chiral stationary phase (CSP) were developed for the enantioseparation of propranolol, atenolol, metoprolol, venlafaxine, fluoxetine, and amlodipine. The retention and enantioseparation properties of these analytes were investigated in the variety of mobile phase additives, flow rate, and column temperature. As a result, the optimal chromatographic condition was achieved using methanol as a main mobile phase with triethylamine (TEA) and glacial acetic acid (HOAc) added as modifiers in a volume ratio of 0.01% at a flow rate of 0.3?mL/minute and at a column temperature of 5°C. The thermodynamic parameters (eg, ΔH, ΔΔH, and ΔΔS) from linear van 't Hoff plots revealed that the retention of investigated pharmaceuticals on vancomycin CSP was an exothermic process. The nonlinear behavior of lnk′ against 1/T for propranolol, atenolol, and metoprolol suggested the presence of multiple binding mechanisms for these analytes on CSP with variation of temperature. The simulated interaction processes between vancomycin and pharmaceutical enantiomers using molecular docking technique and binding energy calculations indicated that the calculated magnitudes of steady combination energy (ΔG) coincided with experimental elution order for most of these enantiomers.

Full text of DOI:10.1002/chir.23052

Received: 30 September 2018
DOI: 10.1002/chir.23052
Revised: 25 December 2018
Accepted: 27 December 2018
R E G U L A R A R T I C L E
Enantioseparation of chiral pharmaceuticals by
vancomycin‐bonded stationary phase and analysis of chiral
recognition mechanism
Jiaxi Li1,2 | Ruixia Liu1,2
| Liyang Wang1,2 | Xiaoling Liu1,2 | Hongjie Gao1,2
1 State Key Laboratory of Environmental
Criteria and Risk Assessment, Chinese
Research Academy of Environmental
Science, Beijing, People's Republic of
China
2 Basin Research Center for Water
Pollution Control, Chinese Research
Academy of Environmental Science,
Beijing, People's Republic of China
Abstract
The drug chirality is attracting increasing attention because of different biologi-
cal activities, metabolic pathways, and toxicities of chiral enantiomers.
The chiral separation has been a great challenge. Optimized high‐performance
liquid chromatography (HPLC) methods based on vancomycin chiral stationary
phase (CSP) were developed for the enantioseparation of propranolol, atenolol,
metoprolol, venlafaxine, fluoxetine, and amlodipine. The retention and
enantioseparation properties of these analytes were investigated in the variety
of mobile phase additives, flow rate, and column temperature. As a result, the
optimal chromatographic condition was achieved using methanol as a main
mobile phase with triethylamine (TEA) and glacial acetic acid (HOAc) added
as modifiers in a volume ratio of 0.01% at a flow rate of 0.3 mL/minute and at
a column temperature of 5°C. The thermodynamic parameters (eg, ΔH, ΔΔH,
and ΔΔS) from linear van 't Hoff plots revealed that the retention of investigated
pharmaceuticals on vancomycin CSP was an exothermic process. The nonlinear
behavior of lnk′ against 1/T for propranolol, atenolol, and metoprolol suggested
the presence of multiple binding mechanisms for these analytes on CSP
with variation of temperature. The simulated interaction processes between
vancomycin and pharmaceutical enantiomers using molecular docking
technique and binding energy calculations indicated that the calculated
magnitudes of steady combination energy (ΔG) coincided with experimental
elution order for most of these enantiomers.
Correspondence
Ruixia Liu, State Key Laboratory of
Environmental Criteria and Risk
Assessment, Chinese Research Academy
of Environmental Science, Beijing 100012,
People's Republic of China.
Email: liurx@craes.org.cn
Funding information
Basic Scientific and Research Program for
Central Nonprofit Research Institutes,
Grant/Award Number: 2017‐YSKY‐10;
National Natural Science Foundation of
China, Grant/Award Number: 21277133
KEYWORDS
chiral pharmaceutical, chiral recognition mechanism, enantioseparation, HPLC‐vancomycin CSP,
molecular docking, thermodynamic analysis
1 | INTRODUCTION
metabolic pathways, and toxicities of chiral enantiomers,
the chiral separation has been a great challenge and
Pharmaceuticals are a large class of compounds with
diverse properties and applications. It has been reported
that more than 50% of these pharmaceuticals in current
use are chiral compounds.1,2 Because of the identical phys-
icochemical properties but different biological activities,
still remains a hot issue, which has attracted attention
of agricultural, pharmaceutical, and environmental com-
munities.2-4
Of the enantioseparation techniques,
a
high‐
performance liquid chromatography (HPLC) with chiral
Chirality. 2019;1–12.
wileyonlinelibrary.com/journal/chir
© 2019 Wiley Periodicals, Inc.
1

2
LI ET AL.
stationary phases (CSPs) has been recognized to be the
most predominant method for chiral separation.5-7 Since
the first introduction of macrocyclic antibiotics as chiral
selector in 1994 by Armstrong et al,8 the vancomycin‐
based CSPs have been widely applied for the separation
of various chiral pharmaceuticals, pesticides and indus-
trial materials with liquid chromatography (LC) in differ-
ent modes. An efficient separation of a number of acidic
drugs was obtained using vancomycin immobilized on
silica as CSP in reversed phase LC.9 Two vancomycin‐
based CSPs with chiral selectors of Chirobiotic V and
Chirobiotic V2 were evaluated for the enantioseparation
of β‐blockers and profens in reversed‐phase and polar‐
organic mode, where it indicated that the better
enantioseparation of the selected drugs was achieved on
the Chirobiotic V2 column in polar‐organic mode.10 The
vancomycin immobilized onto diol hydride‐based silica
particles as CSP in nano‐LC was used to separate the
enantiomers of various nonsteroidal anti‐inflammatory
drugs and β‐blockers, showing a better chromatographic
separation performance for chiral acidic compounds.11
Specifically regarding basic pharmaceuticals, R‐ and S‐
enantiomers of cinacalcet were effectively separated on
a Chirobiotic V column packed with vancomycin as a
CSP in polar ionic mode LC, and successively determined
by tandem mass spectrometric method.12 The enantio-
meric separation and quantification of metoprolol, pro-
pranolol, fluoxetine, and venlafaxine were conducted on
commercialized vancomycin CSP in LC, coupled with
tandem mass spectrometry.13 Although some work has
been conducted on the enantiomeric separation of several
basic drugs based on vancomycin CSP in LC, to the
best of our knowledge, the enantiomers amlodipine
separation using vancomycin‐based CSP has not been
previously reported, thus chromatographic conditions of
extensive basic drug enantiomers and separation of
structurally similar chiral pharmaceuticals should be
systematically optimized and evaluated.
and steric interactions were more dominant in reversed‐
phase mode.18-20 Unequivocally, the acidic and basic
additives in mobile phase impacted chiral recognition
mechanism by regulating ion‐pairing interaction and
altering proton activity of the mobile phase medium,
further influencing electrostatic interactions between
analyte and the CSP.21-23 The liophilic ions, as the mobile
phase modifiers, have also been used to control retention
and selectivity of analytes,24,25 where enantiorecognition
mechanisms of basic pharmaceuticals on a vancomycin‐
based column in polar ionic mode involved chaotropic
interaction between the liophilic ions and the analytes
on the CSP.26 Fortunately, computer simulation methods
have been proposed as useful approaches to gain insight
into chiral recognition mechanism. For instance, the
molecular docking has been employed to simulate the
bonding between pharmaceutical enantiomers and CSPs
and to predict the order of isomer peaks in chromato-
graphic separation.11,12,17 Nevertheless, because of the
complex interactions between multiple pharmaceuticals
and CSPs, there is still a need for comprehensive and
extensive understanding on chiral recognition mecha-
nisms and thermodynamics of chiral basic pharmaceuti-
cals separations in vancomycin‐based CSPs.
Therefore, the aims of this work, using vancomycin‐
based CSPs, are to systematically and extensively
optimize enantioseparation conditions of amine group‐
contained basic chiral pharmaceuticals including pro-
pranolol, atenolol, metoprolol, venlafaxine, fluoxetine,
and amlodipine; to investigate their thermodynamic
properties of the chiral separation process so as to provide
an extra evidence for the chiral recognition mechanism;
and to further explain bonding strength of isomers
and possible chiral recognition mechanisms of pharma-
ceutical racemates based on molecular docking approach.
In addition, to help improve enantioseparation
efficiency of chiral compounds or develop new chiral
separation methods, there are critical demands to explore
chiral recognition mechanisms. Enantiorecognition
mechanisms involve a variety of molecular interactions.
With multifunctional groups, cavities, and numerous ste-
reogenic centers, vancomycin was able to provide active
sites for hydrogen bond, dipole‐dipole interaction, π‐π
interaction, and electrostatic interactions.14,15 The inter-
action between the analyte and the CSP totally depended
on chromatographic modes and mobile phase addi-
tives.16,17 It was found that enantiorecognition mecha-
nisms on vancomycin in polar ionic mode was mainly
derived from hydrogen bonding, electrostatic forces, and
steric hindrance, whereas hydrophobic, electrostatic,
2 | MATERIALS AND METHODS
2.1 | Instrumentation and chemicals
All the enantiomeric separation experiments were per-
formed on an HPLC system from Agilent Technologies
(Palo Alto, CA, USA) with an ultraviolet detector (1260
VWD VL, Agilent Technologies, Germany). S‐(−)‐pro-
pranolol (PHO) and R‐(+)‐propranolol were purchased
from TLC Pharmaceutical Standards Ltd (Canada). Pure
R‐ and S‐enantiomers of atenolol (ATL), metoprolol
(MTP), venlafaxine (VLX), fluoxetine (FLX), and
amlodipine (ADP) were obtained from Toronto Research
Chemicals Inc. (Canada). The chemical structures of chi-
ral compounds were shown in Figure 1.

LI ET AL.
3
to 25°C, controlled with a precise temperature controller
(1260TCC‐G1316A/Agilent Technologies, USA). The
chromatographic system was equilibrated at each temper-
ature for at least 1 hour before each experiment. Metha-
nol, with small amount of TEA and HOAc as modifiers
in various volume ratios of 0% to 0.03% (v/v), was used
as mobile phase. The separation of chiral pharmaceuti-
cals was performed at the mobile phase flow rate of 0.2
to 0.4 mL/minute in the reversed‐phase mode. Each
solution was injected in triplicate with the injection
volume of 5 μL.
The resolution (Rs) was main parameter to evaluate
the separation efficiency of enantiomers.27 By measure-
ment of the retention time and peak width, the Rs value
for enantiomers could be calculated using the following
Equation 1:
FIGURE 1 Structures of six pharmaceuticals
Methanol (MeOH) of HPLC grade was purchased
from Mreda (USA). Triethylamine (TEA) and glacial
acetic acid (HOAc) were of analytical grade. The mobile
phase was filtered through CNW 0.22‐μm filter (CNW,
GER) and degassed before use. Milli‐Q ultrapure water
was used throughout the experiment.
Rs ¼ 2ðtR2 − tR1Þ=ðw2 − w1Þ;
(1)
where tR1 and tR2 are the retention time and w1 and w2
are base‐peak width of the first and second eluted enan-
tiomer, respectively.
Separate stock solutions of individual enantiomer
were prepared by dissolving the substances in methanol
at a concentration of 100 mg/L and stored in glass stan-
dard bottles (CNW, GER). The working standards of each
chiral drug were made up at concentrations of 1 mg/L by
mixing appropriate volumes of individual enantiomer
solutions and diluting with the same solvent. All solu-
tions were stored at −4°C when not in use. All glassware
was deactivated by methanol and dichloromethane to
prevent the sorption of analytes to the hydroxyl sites on
the glass surface.
To ensure reliable results in the separation process, a
strict quality control procedure was followed. The glass-
ware was carefully washed with ultrapure water and
methanol to avoid contamination prior to use. To further
validate the reproducibility of enantioseparation method,
the relative standard deviations (RSD) for retention time
and Rs of enantiomers were calculated with 6 parallel
injections at 1 mg/L of concentration under the optimized
chromatographic conditions.
2.3 | Thermodynamic calculations of
enantioseparation
2.2 | Chromatographic condition
The thermodynamic parameters can be accessible by
retention factor (k′) and the selectivity factor (α), which
are related to ΔG, the Gibbs energy, and Δ (ΔG), the dif-
ference in the molar Gibbs energy of two enantiomers,
respectively28:
Enantioselective separations were achieved using a
vancomycin‐bonded column (5 μm, 150 × 2.1 mm id;
Astec Chirobiotic V, Sigma‐Aldrich) with an ultraviolet
detector (1260 VWD VL, Agilent Technologies, Germany)
and an automatic injector (1260 ALS‐G1329B, Agilent
Technologies, Germany). All the samples of 5 μL at
1 mg/L of concentration were injected into the HPLC
for analysis. A computer and an Agilent ChemStation
data processing workstation were used to collect and ana-
lyze the data. The detection wavelength was set to 226 nm
for metoprolol and fluoxetine, 290 nm for propranolol,
237 nm for amlodipine, and 230 nm for atenolol and
venlafaxine. Before first use, the column was equilibrated
with the prescribed mobile phase until a stable baseline
was achieved. Between runs, methanol as a blank sample
was injected to avoid interference of the prior sample.
The column was operated at temperature range of 5°C
Δ ðΔGÞ ¼ −RTln α;
(2)
(3)
(4)
′
′
α ¼ k2 =k1 ;
ΔG ¼ −RTlnk′;
where R is the gas constant and T is the absolute temper-
ature. From Gibbs free energy equation, Δ (ΔG) can be
deduced as follows:
Δ ðΔGÞ ¼ Δ ðΔHÞ–TΔ ðΔSÞ:
(5)

4
LI ET AL.
In general, the plots of the natural logarithms of chro-
matographic k′ value and α value against the reciprocal
of absolute temperature are termed as van 't Hoff plots
and are linear.
the RCSB Protein Data Bank database (http://www.rcsb.
org/)29 and assumed as a macromolecular receptor. The
charges of vancomycin in its protonated state were calcu-
lated by the Gasteiger‐Marsili method.34 The grid maps
were calculated by AutoGrid, and the dimensions of the
grid box was set to 40 × 40 × 40 Å, with the grid spacing
set to 0.375 Å. The docking conformation possessed the
lowest binding free energy because of its being thermody-
namically favorable, which was the most stable confor-
mations of the models.
lnk′ ¼ −ΔH=RT þ ΔS=R þ lnφ;
(6)
(7)
lnα ¼ −Δ ðΔHÞ=RT þ Δ ðΔSÞ=R;
where ΔH and ΔS are the enthalpy and entropy
changes, respectively, when the analytes transfer from
the mobile phase to the stationary phase. Correspond-
ingly, Δ (ΔH) and Δ (ΔS) are the difference in enthalpy
and entropy changes between two isomers, respectively.
The intent of φ is the phase ratio of the column (station-
ary/mobile).
3 | RESULTS AND DISCUSSION
3.1 | Effect of flow rate
The effect of the flow rate of the mobile phase on the
enantiomer separation was investigated in the range of
0.2 to 0.4 mL/minute, where a mixture of MeOH, HOAc,
and TEA in the volume ratio of 100:0.01:0.02 was used as
a mobile phase, and the separation was performed at
25°C. As shown in Figure 2, with the increase of the
flow rate, the α values of the six pharmaceuticals were
nearly unchanged, whereas their Rs values slightly
decreased except for that of VLX, indicating that
lower flow rate was preferred for enantioseparation.
This observation was likely because of reduced resistance
to mass transfer at lower flow rate, which was also
obtained for enantioseparation of dansyl‐amino acids on
norvancomycin‐bonded CSP and amino acids or analogues
onto cinchona alkaloid‐derived zwitterionic CSP.30 The
Van Deemter equation, which described the relationship
between height equivalent to a theoretical plate of a
chromatographic column (HETP) and the flow rate of
mobile phase (ν presented as linear velocity), could
2.4 | Molecule simulation method
To understand binding mechanism of enantiomers onto
CSPs behind the observed results of enantioresolution,
the molecular docking was performed for mechanics sim-
ulation between molecules using Autodock (version 4.0)
software.29,30 In a molecular mechanics simulation, the
primary mechanisms contained H‐bonding, van der
Waals interaction, steric repulsion, and electrostatic inter-
action. The Lamarckian genetic algorithm (LGA) was
applied to search the conformational and orientational
space by iterations of 10. The binding free energy (ΔG)
of pharmaceutical onto the vancomycin CSPs was
obtained based on the most stable docking conformation
by semi‐empirical binding‐free energy function31:
ΔG ¼ ΔGvdw þ ΔGelec þ ΔGH−bond þ ΔGsol
þ ΔGconfom þ ΔGtor
(8)
provide further explanation for this result31,32
:
where ΔGvdw, ΔGelec, and ΔGH‐bond represent the energies
of van der Waals' interaction, electrostatic interaction,
and H‐bonding, respectively, and ΔGsol is the desolvation
upon binding and hydrophobic interaction, ΔGconfom
models deviation of bond length and bond angle, and
ΔGtor represents restriction of internal rotors and global
rotation and translation. It should be noticed that there
were usually some assumptions in modeling chiral sepa-
rations of chromatographic techniques using docking
method, such as not taking solvation effect, buffer effect
and entropy difference into consideration.30,32
The molecular three‐dimensional structures of chiral
pharmaceuticals containing S‐ and R‐isomers were
obtained from the Pubchem database (https://pubchem.
ncbi.nlm.nih.gov/). They were then optimized using the
molecular mechanics Powell method.33 The three‐
dimensional structure of vancomycin was obtained from
HETP ¼ A þ B=ν þ Cν
(9)
where A is the coefficient for eddy diffusion, which
relates to particle diameter of stationary phase stuffing;
B is the coefficient for longitudinal diffusion, which is
proportional to the diffusion coefficient of the mobile
phase; and C is the coefficient for resistance to mass
transfer. As for liquid chromatography, the longitudinal
diffusion coefficient of liquid is significantly smaller than
that of gas, which results in the neglection of the second
term of Equation 9. Thus, reducing ν values produced the
lower HETP value and hence the higher resolution of
two enantiomers.33 In addition, a lower flow rate would
undoubtedly cause retention time extending from 8 to
17 minutes at a flow rate of 0.4 mL/minute to 20 to
37 minutes at a flow rate of 0.2 mL/minute, and slight

LI ET AL.
5
3.2 | Effect of acid and base additives
Our preliminary experiment indicated that all six phar-
maceuticals were unable to be eluted and separated
within 30 minutes at a flow rate of 0.5 mL/minute using
pure methanol as mobile phase (data not presented here).
The additives in mobile phase are critical for the
enantioseparation, and the effect of acid and base
additives has been previously reported by a number of
studies.9,35,36 Various volume ratios of HOAc and TEA
added to methanol as mobile phase were selected for
assessing effect on enantioseparation at 25°C and a flow
rate of 0.3 mL/minute. As shown in Figure 3, at a certain
percent of HOAc (0.01%, v/v), the addition of TEA had a
positive effect on the enantioseperation for most of
selected pharmaceuticals, even more for amlodipine that
was not separated at all without TEA additives, suggest-
ing that TEA played a crucial role on enantioseparation
of these pharmaceuticals. Specifically, increasing TEA
percentage from 0.01% to 0.03% led to a decrease in the
items of k′, Rs, and α for most of the investigated analytes,
apart from amlodipine that there was no constant trend
in the Rs and α values, and propranolol that its α values
slightly raised with increased TEA addition. In the case
of certain percent of TEA (0.01%, v/v) (as observed in
Figure 4), decreased k′ values were observed for five of
the investigated pharmaceuticals, with increasing HOAc
percentage from 0% to 0.03%, while the k′ values of
venlafaxine achieved the maximum at HOAc percentage
of 0.01%. The profiles of selectivity and resolution were
found to be highly analyte‐dependent. For propranolol,
atenolol, and metoprolol, the α and Rs values were
decreasing with increasing HOAc percentage, and the
overall increased trends were observed in α values for
venlafaxine, fluoxetine, and amlodipine, and Rs values
for venlafaxine and fluoxetine.
The above observations were associated with the
chemical structures of target pharmaceuticals and addi-
tives in the mobile phase and stationary phase. The acid
basic additives would lead to the alteration of mobile
phase pH, further influencing the dissociation and
protonation of both analytes and vancomycin that has
several ionization constants (pKa: 2.9, 7.2, 8.6, 9.6, 10.4,
and 11.7) in vancomycin. In the present study, the pH
ranges of mobile phase with different ratios of acid basic
additives varied from 6.1 to 7.6, which was lower than
pKa of target analytes ranged from 8.6 to 10.1 and close
to isoelectric point of vancomycin (7.2), thus, almost
keeping the selected pharmaceuticals and vancomycin
in the neutral form.
FIGURE
2
Effect of flow rate of mobile phase on the
enantioseparation of six pharmaceuticals (chromatographic
conditions: Methanol containing 0.01% HOAc and 0.02% TEA,
column temperature for 25°C)
peak broadening was also observed. Considering both
of the resolution and retention time, a flow rate of
0.3 mL/minute was selected in this study.
Additionally, between selected pharmaceuticals, best
selectivity and resolution for vancomycin‐based CSPs
and target analytes were observed for fluoxetine with

6
LI ET AL.
FIGURE 3 Effect of TEA additives on the enantioseparation of
six pharmaceuticals under certain percent of HOAc (0.01%, v/v)
(chromatographic conditions: column temperature 25°C and flow
rate of 0.3 mL/min)
FIGURE 4 Effect of HOAc additives on the enantioseparation of
six pharmaceuticals under certain percent of TEA (0.01%, v/v)
(chromatographic conditions: column temperature 25°C and flow
rate of 0.3 mL/min)
highest pKa value of 10.1 despite the insignificant linear
relationship between pKa values of analytes and their
enantioseparation efficiency. Those results indicated a
vital role of the dipole–dipole interaction between
vancomycin‐based CSPs and target analytes. In common
case, the hydrogen bonding between analytes and
CPS should be of less importance in methanol mobile
phase system than in nonalcoholic type solvents, eg,
acetonitrile‐based mobile phase20 because of the stronger
hydrogen bonding between methanol and CSP, hence,
dipole–dipole or electrostatic interactions might be a
dominated interaction for chiral recognition because of
the increased selectivity.9,35

LI ET AL.
7
It was reported that the acid and base additives could
also possibly adjust both polarity and ionic strength of
mobile phase.37 Chemically, both of TEA and basic
pharmaceuticals would combine with dissociative silanol
groups of the stationary phase.38 On the one hand, TEA
as a masking agent was highly beneficial to improve
tailing of the peak shape, but on the other hand, an
excess of TEA could compete interaction sites with basic
pharmaceuticals on the stationary phase, resulting in
negative impact on enantioseparation of pharmaceuticals.
The excess of HOAc in mobile phase could cause an
increased electrostatic repulsion to target analytes by
the protonation of functional groups in vancomycin.35
Of these pharmaceuticals in the current study,
propranolol, atenolol, and metoprolol maintained a good
enantioseparation performance at lower HOAc percent-
age, while venlafaxine, fluoxetine, and amlodipine
showed an improved enantioselectivety at higher HOAc
percentage, which was not consistent with previous
studies where it was confirmed that vancomycin‐based
CPS gave an increased enantioselectivity with the
increase of pH value. The inconsistent result might be
somewhat related to diverse ionization constants of
investigated pharmaceuticals.39
3.3 | Temperature effects and
thermomechanical analysis
Because of the instability of vancomycin at the tempera-
ture higher than 37°C,40 the chromatographic column
was performed in the temperature range of 5°C to 25°C.
The effects of temperature on the enantioseparation of
the six chiral pharmaceuticals were shown in Figure 5.
As observed for all six pharmaceuticals, the Rs and α
values decreased by increasing the column temperature,
apart from fluoxetine that had lower Rs value at 10°C.
The consistent conclusion was also obtained by Ding
et al, who reported that vancomycin CSP performed
better enantioselectivity for amino acid derivatives at
lower temperature.41 Additionally, in the tested tempera-
ture ranges, the reduced k′ values was observed at higher
temperature for venlafaxine, fluoxetine, and amlodipine,
while there was no significant variation in k′ values for
propranolol, atenolol, and metoprolol.
FIGURE 5 Effect of temperature on enantioseparation of six
pharmaceuticals (chromatographic conditions: methanol
containing 0.01% HOAc and 0.02% TEA, flow rate of 0.3 mL/min)
The lnk′ of the six pharmaceuticals was plotted
against 1/T, and the details of the linear regression
equations were given in Table 1. It was clear that higher
linearity was obtained for venlafaxine, fluoxetine, and
amlodipine with correlation coefficients of 0.85 to 0.96.
Correspondingly, the ΔH and ΔS terms were derived
from the slope (−ΔH/R) and intercept (ΔS/R + lnφ) of
the linear plot. The negative values of ΔH were observed
for venlafaxine, fluoxetine, and amlodipine, indicating
that energy was released during transferring of these
solutes from mobile phase to stationary phase. For each
pharmaceutical, the ΔH or ΔS/R + lnφ terms of S‐type
enantiomer were less negative or higher than those of
R‐type enantiomer, which was in accordance to the

8
LI ET AL.
elution order of enantiomers (S‐type first eluted) during
the chromatographic separation. Regarding propranolol,
atenolol, and metoprolol, the nonlinear relationships of
lnk′ versus 1/T were presented in Table 1, which could
be related to multiple adsorption mechanisms or
change in adsorption mechanism within the temperature
range of 5°C to 25°C. The similar nonlinear behavior
was previously reported for the enantioseparation of
metoprolol enantiomers on vancomycin‐immobilized
CSP9 and conformationally rigid spirolactam on (R)‐N‐
(3,5‐dinitrobenzoyl)phenylglycine‐derived CSPs,42 where
this unusual behavior could be ascribed to the
temperature‐dependent interaction of additives with the
stationary phase and/or the target analytes.
The plots of lnα versus 1/T showed a good linearity
for all analytes, with correlation coefficients above 0.92.
The positive slopes (−Δ (ΔH)/R) and negative intercepts
(Δ (ΔS)/R) of linear plots were obtained for all pharma-
ceuticals, giving the negative values of Δ (ΔH) from −0.3
to −1.6 kJ/mol and Δ (ΔS) from −0.2 to −4.5 J/mol.
This result indicated that the chiral separations of these
pharmaceuticals were dominantly enthalpically driven.
Furthermore, the more negative value of Δ (ΔH) and
Δ (ΔS) represented a good enantioseparation perfor-
mance.28 Of these pharmaceuticals, fluoxetine and
venlafaxine enantiomers with highest values of |Δ (ΔH)|
(1.0 and 1.6 kJ/mol) and |Δ (ΔS)| (2.5 and 4.5 J/mol)
indeed attained excellent resolutions (RsFLX = 2.8,
RsVLX = 2.52) and enantioselectivities (αFLX = 1.12,
αVLX = 1.12) in the process of seperation.
The iso‐enantioselective temperatures (Tiso), where
the elution order of the enantiomers was reversed, could
be calculated as the ratio between Δ (ΔH) and Δ (ΔS).
As presented in Table 1, the Tiso values of these
pharmaceuticals were pretty high, making the reversion
of the enantioelution order impossible under ambient
experimental conditions.
3.4 | Method reproducibility and
adaptability
Given baseline separation for enantiomers of six pharma-
ceuticals, the optimal chromatographic condition was
obtained with methanol as a main mobile phase
containing TEA and HOAc in volume ratio of 0.01%,
respectively, a flow rate of 0.3 mL/minute and column
temperature of 5°C. Under optimal chiral separation
conditions, the chromatographic separation results of
each enantiomer were shown in Figure 6, and the RSD
for retention time and Rs of enantiomers were below
0.92% and 1.80%. The results showed that the method
had good reproducibility and separation efficiency.

LI ET AL.
9
To further understand the micro‐mechanisms of
pharmaceutical enantiomers binding onto the vancomy-
cin, the molecular docking were performed by Autodock.
The calculated, most steady combination energy (ΔG) of
pharmaceutical enantiomers with two domains of vanco-
mycin and the differences of binding free energy (ΔΔG)
for S‐ and R‐type of enantiomers were presented in
Table 2. All analyte‐vancomycin complexes showed bind-
ing free energies in the range of −4.35 to −7.39 kcal/mol,
indicating the spontaneity of the binding between analyte
molecules and vancomycin. By comparison of the ΔΔG
term between two domains of vancomycin, the ΔΔG
values of all target analytes on the Domain II were above
0, and their absolute values (|ΔΔG|) were greater than
those of the Domain I, except for that of venlafaxine,
indicating a good corresponding to the experimental
elution order of S‐ and R‐type of enantiomers and better
enantiomeric separation of analytes on the Domain II.
These results suggested that Domain II was the main
active site for enantiomeric separation of investigated
pharmaceuticals.
3.5 | Analysis and prediction of
chromatographic retention
The multiple adsorption mechanisms, such as hydrogen‐
bond, dipole–dipole interaction, π–π interaction, electro-
static interaction, and steric hindrance, have been
reported for the enantioseparation of chiral compounds
on vancomycin stationary phase.43,44 The complex struc-
ture of macromolecule would provide diverse binding
sites, resulting in multiple retention mechanisms. In the
view of the three‐dimensional structure of vancomycin
obtained from the RCSB Protein Data Bank (http://
www.rcsb.org/) database, the molecule of vancomycin
was assumed to comprise two adsorption active domains
for the interaction with analyes,45 as shown in Figure 7.
Regarding the Domain II, the lowest ΔG value was
observed for R‐amlodipine (−7.05 kcal/mol), and the
highest one was found for S‐metoprolol (−4.35 kcal/mol),
demonstrating that a stronger affinity between amlodipine
and vancomycin, and weaker affinity between metoprolol
and vancomycin, respectively. Additionally, the calculated
ΔG magnitudes for all enantiomers were in the order of
S‐metoprolol > R‐metoprolol > S‐venlafaxine > S‐ateno-
lol > R‐venlafaxine > S‐fluoxetine > R‐atenolol > R‐
fluoxetine > S‐propranolol > S‐amlodipine > R‐proprano-
lol > R‐amlodipine. This sequence was somewhat consis-
tent with experimental elution order for the enantiomers
of target pharmaceuticals (as shown in Figure 6), except
for S‐ and R‐ propranolol. Nevertheless, the significant
differences were obtained for propranolol between pre-
dicted elution order and experimental profile. This deviant
result might be because of the limitation of Autodock
software, where there was neither the π–π interaction
between naphthalene nucleus of propranolol and vanco-
mycin involved nor the solvent effect included in the
calculation process of molecular docking.10,29 Unequivo-
cally, the retention arrangement and mechanism of
analyes were strongly controlled by molecular structures.
As indicated in Figure 1, the molecules of all analytes
comprised benzene rings, which could undoubtedly
contribute to binding on the packing of vancomycin by
face‐to‐face π–π interaction. Therefore, the more benzene
rings in the structure of propranolol, fluoxetine, and
amlodipine, providing stronger π–π interaction, would be
responsible for longer retention time of these pharmaceuti-
cals. Similarly, the characteristic side chains of analyes
likewise played important roles in retention behavior.
FIGURE 6 Chromatogram for the enantioseparation of six
pharmaceuticals under optimized chromatographic conditions
(methanol containing 0.01% HOAc and 0.01% TEA, flow rate of
0.3 mL/min, column temperature 5°C)
FIGURE 7 Two structural domains of vancomycin

10
LI ET AL.
TABLE 2 Most steady combination energy of vancomycin inter-
action with drug enantiomers (kcal/mol
The peak order of all the enantiomers somewhat
corresponded to binding free energy except for proprano-
lol, which was likely due to the neglect of π–π interaction
and solvent effect in the calculation process of molecule
docking. These results showed that the molecular
modeling technique provided a good perspective to fur-
ther understand the chiral recognition mechanism, which
should be further investigated.
Domain I
Domain II
Compounds
ΔG
ΔΔG
ΔG
ΔΔG
S‐ propranolol
R‐ propranolol
−7.39
−7.10
−0.29
−6.04
−7.02
0.98
S‐ atenolol
R‐ atenolol
−6.09
−5.88
−0.21
−0.13
0.80
−5.17
−5.95
0.78
0.47
0.46
0.33
0.92
S‐ metoprolol
R‐ metoprolol
−4.79
−4.66
−4.35
−4.82
ACKNOWLEDGMENTS
This work was supported by Basic Scientific and Research
Program for Central Nonprofit Research Institutes
(No.2017‐YSKY‐10) and National Natural Science
Foundation of China (No.21277133).
S‐ venlafaxine
R‐ venlafaxine
−4.94
−5.74
−5.05
−5.51
S‐ fluoxetine
R‐ fluoxetine
−6.22
−5.95
−0.27
0.49
−5.67
−6.00
S‐ amlodipine
R‐ amlodipine
−5.75
−6.24
−6.13
−7.05
ORCID
Ruixia Liu https://orcid.org/0000-0001-7547-5471
The presence of amino, acylamino, and hydroxyl groups in
the molecules were also beneficial to form hydrogen bond
and dipolar interaction with vancomycin. For instance,
metoprolol with less acylamino group, compared with
atenolol, would form weaker hydrogen bond onto vanco-
mycin, resulting in shorter retention time on vancomycin.
In addition to hydrogen bond, steric hindrance of alkoxy
chain in metoprolol was also a reasonable explanation for
the lower retention time of metoprolol enantiomers.
REFERENCES
1. Agranat I, Wainschtein SR, Zusman EZ. The predicated demise
of racemic new molecular entities is an exaggeration. Nat Rev
Drug Discov. 2012;11(12):972‐973.
2. Sanganyado E, Lu Z, Fu Q, Schlenk D, Gan J. Chiral pharma-
ceuticals: a review on their environmental occurrence and fate
processes. Water Res. 2017;124:527‐542.
3. Ribeiro AR, Castro PML, Tiritan ME. Chiral pharmaceuticals in
the environment. Environ Chem Lett. 2012;10(3):239‐253.
4. Ribeiro C, Santos C, Gonã§Alves V, Ramos A, Afonso C, Tiritan
ME. Chiral drug analysis in forensic chemistry: an overview.
Molecules. 2018;23:1‐47.
4 | CONCLUSION
Methanol as a main mobile phase with addition of TEA
and HOAc as modifiers could be successfully used
for the enantiomeric separation of six pharmaceuticals
on a vancomycin CPS. An increased retention time,
enantioselectivity, and enantioresolution at lower per-
centage of HOAc and TEA were observed for all the
analytes. The reduced flow rate of mobile phase and col-
umn temperature were beneficial to efficient enantiomers
separation. The thermodynamic parameters revealed that
the retention of investigated pharmaceuticals on vanco-
mycin CSP was an exothermic process. The nonlinear
van 't Hoff plots of propranolol, atenolol, and metoprolol
indicated multiple binding sites and adsorption
mechanisms of these anlytes on CSP. In comparing the
thermodynamic parameters of six pharmaceuticals, the
racemates with higher |Δ (ΔH)| and |Δ (ΔS)| had excellent
resolutions and enantioselectivities of the enantiomer
separation. The molecular docking results showed that
the calculated differences in binding free energy between
S‐ and R‐type enantiomer were above 0, indicating a
good agreement with experimentally eluted peak order.
5. Kasprzykhordern B. Pharmacologically active compounds in the
environment and their chirality. Chem Soc Rev. 2010;39(11):
4466‐4503.
6. Lara‐Martã PA, Gonzã¡Lez‐Mazo E, Pintado‐Herrera MG,
Baena‐Nogueras RMA. Determination of pharmaceuticals in
coastal systems using solid phase extraction (SPE) followed by
ultra performance liquid chromatography‐tandem mass spec-
trometry (UPLC‐MS/MS). Curr Anal Chem. 2016;12:1‐10.
7. Kadkhodaei K, Forcher L, Schmid MG. Separation of enantio-
mers of new psychoactive substances by HPLC. J Sep Sci.
2017;41:1‐35.
8. Armstrong DW, Tang Y, Chen S, Zhou Y, Bagwill C, Chen JR.
Macrocyclic antibiotics as a new class of chiral selectors for
liquid chromatography. Anal Chem. 1994;66(9):1473‐1484.
9. Svensson LA, Dönnecke J, Karlsson KE, Karlsson A, Vessman J.
Polar organic phase liquid chromatography with packed
capillary columns using a vancomycin chiral stationary phase.
Chirality. 2000;12(8):606‐613.
10. Bosáková Z, Curínová E, Tesarová E. Comparison of
vancomycin‐based stationary phases with different chiral selec-
tor coverage for enantioselective separation of selected drugs

LI ET AL.
11
in high‐performance liquid chromatography. J Chromatogr a.
2005;1088(1‐2):94‐103.
25. Flieger J, Siwek A, Pizoń M. Usefulness of chaotropic salt
additive in RP‐HPLC of organic nonionized compounds. J Sep
Sci. 2013;36(3):469‐476.
11. Rocchi S, Rocco A, Pesek JJ, Matyska MT, Capitani D, Fanali S.
Enantiomers separation by nano‐liquid chromatography: use of
a novel sub‐2 μm vancomycin silica hydride stationary phase.
J Chromatogr a. 2015;1381:149‐159.
26. Mehdinia A, Baradaran KT, Jabbari A, Azizzanjani MO, Ziaei E.
Magnetic molecularly imprinted nanoparticles based on grafting
polymerization for selective detection of 4‐nitrophenol in aque-
ous samples. J Chromatogr a. 2013;1283:82‐88.
12. Wang D, Li F, Jiang Z, Yu L, Guo X. Chiral recognition
mechanisms of four β‐blockers by HPLC with amylose chiral
stationary phase. Iran J Pharm Res. 2014;13(2):449‐457.
27. Evans SE, Davies P, Lubben A, Kasprzyk‐Hordern B. Determi-
nation of chiral pharmaceuticals and illicit drugs in
wastewater and sludge using microwave assisted extraction,
solid‐phase extraction and chiral liquid chromatography
coupled with tandem mass spectrometry. Anal Chim Acta.
2015;882:112‐126.
13. Ribeiro AR, Santos LHMLM, Maia AS, Delerue‐Matos C, Castro
PML, Tiritan ME. Enantiomeric fraction evaluation of pharma-
ceuticals in environmental matrices by liquid chromatography‐
tandem mass spectrometry. J Chromatogr A. 2014;1363:226‐235.
28. Berkecz R, Ilisz I, Forró E, Fülöp F, Armstrong DW, Péter A.
14. Xiao TL, Armstrong DW. Enantiomeric separations by HPLC
using macrocyclic glycopeptide‐based chiral stationary phases:
an overview. Methods Mol Biol. 2004;243:113‐171.
LC enantioseparation of aryl‐substituted
β ‐lactams using
variable‐temperature conditions. Chromatographia. 2006;63
(S13):S29‐S35.
15. Svensson LA, Karlsson KE, Karlsson A, Vessman J. Immobilized
vancomycin as chiral stationary phase in packed capillary liquid
chromatography. Chirality. 2015;10:273‐280.
29. Tanchuk VY, Tanin VO, Vovk AI, Poda G. A new, improved
hybrid scoring function for molecular docking and scoring based
on AutoDock and AutoDock Vina. Chem Biol Drug Des.
2016;87(4):618‐625.
16. Lajkó G, Grecsó N, Tóth G, et al. A comparative study of
enantioseparations of N(α)‐fmoc proteinogenic amino acids on
quinine‐based zwitterionic and anion exchanger‐type chiral
stationary phases under hydro‐organic liquid and subcritical
fluid chromatographic conditions. Molecules. 2016;21:1‐19.
30. Li W, Liu C, Tan G, Zhang X, Zhu Z, Chai Y. Molecular
modeling study of chiral separation and recognition mechanism
of β‐adrenergic antagonists by capillary electrophoresis. Int J
Mol Sci. 2012;13(1):710‐725.
17. Phyo YZ, Cravo S, Palmeira A, et al. Enantiomeric resolution
and docking studies of chiral xanthonic derivatives on
chirobiotic columns. Molecules. 2018;23:1‐22.
31. Morris GM, Goodsell DS, Halliday RS, et al. Automated docking
using a Lamarckian genetic algorithm and an empirical binding
free energy function. J Comput Chem. 2015;19:1639‐1662.
18. Berthod A. Chiral recognition mechanisms with macrocyclic
32. Luque FJ, Orozco M. Polarization effects in generalized molecu-
lar interaction potential: new Hamiltonian for reactivity studies
and mixed QM/MM calculations. J Comput Chem. 1998;19(8):
866‐881.
glycopeptide selectors. Chirality. 2010;21:167‐175.
19. Zhou T. Separation of carboxylates by hydrophilic interaction
liquid chromatography on titania. J Chromatogr A. 2010;1217
(1):82‐88.
33. Powell MJD. An efficient method for finding the minimum of a
function of several variables without calculating derivatives.
Comput J. 1964;7(2):155‐162.
20. Szabó ZI, Foroughbakhshfasaei M, Noszál B, Tóth G.
Enantioseparation of racecadotril using polysaccharide‐type chi-
ral stationary phases in polar organic mode. Chirality. 2018;
30(1):95‐105.
34. Osterberg F, Morris GM, Sanner MF, Olson AJ, Goodsell DS.
Automated docking to multiple target structures: incorporation
of protein mobility and structural water heterogeneity in
AutoDock. Proteins. 2002;46(1):34‐40.
21. Hoffmann CV, Reischl R, Maier NM, Lämmerhofer M,
Lindner W. Investigations of mobile phase contributions to
enantioselective anion‐ and zwitterion‐exchange modes on
35. Zhang D, Xu X, Ma S. The study of enantioseparation of
zolmitriptan on vancomycin‐bonded chiral stationary phase. J
Sep Sci. 2005;28(18):2501‐2504.
quinine‐based zwitterionic chiral stationary phases.
Chromatogr a. 2009;1216(7):1157‐1166.
J
22. Zhang T, Holder E, Franco P, Lindner W. Method development
and optimization on cinchona and chiral sulfonic acid‐based
zwitterionic stationary phases for enantiomer separations of
free amino acids by high‐performance liquid chromatography.
J Chromatogr a. 2014;1363:191‐199.
36. Hsieh ML, Li GY, Chau LK, Hon YS. Single‐step approach to
beta‐cyclodextrin‐bonded silica as monolithic stationary phases
for CEC. J Sep Sci. 2015;31:1819‐1827.
37. Kirkland KM, Neilson KL, Mccombs DA. Comparison of a new
ovomucoid and a second‐generation α 1 ‐acid glycoprotein‐
based chiral column for the direct high‐performance liquid
chromatography resolution of drug enantiomers. J Chromatogr
a. 1991;545(1):43‐58.
23. Yong G, Yang W, Xi H, et al. Determination of hydroxysafflor
yellow a in biological fluids of patients with traumatic brain
injury by UPLC‐ESI‐MS/MS after injection of Xuebijing.
Biomed Chromatogr. 2014;28:1090‐1095.
24. Smuts JP, Hao XQ, Han Z, Parpia C, Krische MJ, Armstrong
DW. Enantiomeric separations of chiral sulfonic and phosphoric
acids with barium‐doped cyclofructan selectors via an ion
interaction mechanism. Anal Chem. 2014;86(2):1282‐1290.
38. Mao Y, Carr PW. Separation of selected basic pharmaceuticals
by reversed‐phase and ion‐exchange chromatography using
thermally tuned tandem columns. Anal Chem. 2015;73:
4478‐4485.

12
LI ET AL.
39. Gasper MP, Berthod A, Nair UB, Armstrong DW. Comparison
and modeling study of vancomycin, ristocetin a, and teicoplanin
for CE enantioseparations. Anal Chem. 1996;68(15):2501‐2514.
44. Aboul‐Enein HY, Serignese V. Enantiomeric separation of
several cyclic imides on a macrocyclic antibiotic (vancomycin)
chiral stationary phase under normal and reversed phase
conditions. Chirality. 2015;10:358‐361.
40. Raverdy V, Ampe E, Hecq JD, Tulkens PM. Stability and com-
patibility of vancomycin for administration by continuous
infusion. J Antimicrob Chemoth. 2013;68(5):1179‐1182.
45. Hu CQ, Hong JW. Investigation on the chiral recognition site of
enantioselective separation of ofloxacin by capillary zone
electrophoresis using vancomycin as a chiral selector. Acta
Pharm Sin. 2009;44:905‐910.
41. Ding GS, Liu Y, Cong RZ, Wang JD. Chiral separation of
enantiomers of amino acid derivatives by high‐performance
liquid chromatography on a norvancomycin‐bonded chiral
stationary phase. Talanta. 2004;62(5):997‐1003.
How to cite this article: Li J, Liu R, Wang L, Liu
X, Gao H. Enantioseparation of chiral
pharmaceuticals by vancomycin‐bonded stationary
phase and analysis of chiral recognition
mechanism. Chirality. 2019;1–12. https://doi.org/
10.1002/chir.23052
42. Pirkle WH. Unusual effect of temperature on the retention of
enantiomers on a chiral column. J Chromatogr A. 1991;
558(1):1‐6.
43. Boronovã K, Lehotay J, Hroboňová K, Armstrong DW. Study
of physicochemical interaction of aryloxyaminopropanol
derivatives with teicoplanin and vancomycin phases in view
of quantitative structure‐property relationship studies.
Chromatogr A. 2013;1301:38‐47.
J