Synthesis of diastereomeric anhydrides of (RS)-ketorolac and (RS)-etodolac, semi-preparative HPLC enantioseparation, establishment of molecular asymmetry and recovery of pure enantiomers

Article abstract of DOI:10.1039/c7nj02898a

Herein, enantioseparation of two anti-inflammatory drugs, namely, (RS)-ketorolac and (RS)-etodolac, commonly marketed and administered as racemates, was achieved by RP-HPLC. This method provided very low limit of detection values (3.69 and 3.02 ng mL^-1 for diastereomeric derivatives of (R)- and (S)-Ket, respectively) as compared to those reported in literature. (S)-Naproxen benzotriazole ester, which was used as a chiral reagent, was synthesized and characterized by UV, IR, and ¹H NMR spectroscopies, elemental analysis, and polarimetry. The diastereomeric derivatives were synthesized via microwave irradiation, separated on an analytical scale by RP-HPLC, and then isolated by preparative HPLC. The use of a mobile phase containing methanol and aqueous triethylamine phosphate (TEAP) in the isocratic mode was found to be successful for the separation of diastereomeric derivatives, and the separation conditions with respect to pH, flow rate, and buffer concentration were optimized. The diastereomeric derivatives were characterized, and their absolute configuration was established. Hydrolysis of the derivatives provided native enantiomers under mild reaction conditions. This study describes the successful enantioseparation of the above mentioned two analytes by semi-preparative HPLC with easy recovery of the native enantiomers without racemization and with the establishment of molecular asymmetry.

Full text of DOI:10.1039/c7nj02898a

NJC
View Article Online
View Journal
PAPER
Synthesis of diastereomeric anhydrides of
RS)-ketorolac and (RS)-etodolac, semi-preparative
(
Cite this:DOI: 10.1039/c7nj02898a
HPLC enantioseparation, establishment of molecular
asymmetry and recovery of pure enantiomers†
Poonam Malik and Ravi Bhushan
*
Herein, enantioseparation of two anti-inflammatory drugs, namely, (RS)-ketorolac and (RS)-etodolac,
commonly marketed and administered as racemates, was achieved by RP-HPLC. This method provided
ꢀ1
very low limit of detection values (3.69 and 3.02 ng mL for diastereomeric derivatives of (R)- and
(
S)-Ket, respectively) as compared to those reported in literature. (S)-Naproxen benzotriazole ester,
1
which was used as a chiral reagent, was synthesized and characterized by UV, IR, and H NMR
spectroscopies, elemental analysis, and polarimetry. The diastereomeric derivatives were synthesized via
microwave irradiation, separated on an analytical scale by RP-HPLC, and then isolated by preparative HPLC. The
use of a mobile phase containing methanol and aqueous triethylamine phosphate (TEAP) in the isocratic mode
was found to be successful for the separation of diastereomeric derivatives, and the separation conditions with
respect to pH, flow rate, and buffer concentration were optimized. The diastereomeric derivatives were
characterized, and their absolute configuration was established. Hydrolysis of the derivatives provided native
enantiomers under mild reaction conditions. This study describes the successful enantioseparation of the above
mentioned two analytes by semi-preparative HPLC with easy recovery of the native enantiomers without
racemization and with the establishment of molecular asymmetry.
Received 13th August 2017,
Accepted 5th October 2017
DOI: 10.1039/c7nj02898a
rsc.li/njc
enantiomers from diastereomeric derivatives after their separation,
with the exception of one report on the detagging of diastereomeric
1
. Introduction
Both ketorolac (Ket) and etodolac (Etd) are widely used as non- amides of (RS)-propranolol (Prl) that have been prepared using a
1
1
steroidal anti-inflammatory drugs (NSAIDs). The physiological (S)-levofloxacin (Lfx)-based chiral reagent. Although the H NMR
activity of both drugs is almost exclusively confined to their spectra of diastereomeric amides of (RS)-Etd prepared using
(
S)-(ꢀ)-enantiomers, whereas the drugs are marketed and (R)-(ꢀ)-1-cyclohexylethylamine and separated by preparative
administered as racemic mixtures. Without emphasizing the TLC were obtained to establish their configuration, the recovery
importance of enantiomeric resolution, the control of enantio- of native enantiomers was not achieved.
2
meric purity and the determination of individual enantiomeric
Ketorolac (Ket) has been introduced as a safer intravenous
3
drug molecules remain subjects of potential importance from alternative to opioid analgesics. It is chemically designated as
the clinical, analytical, and regulatory points of view. Enantio- (RS)-5-benzoyl-2,3-dihydro-1H-pyrrolizine-1-carboxylic acid (Fig. 1)
separation by an indirect approach of derivatization of a variety and has a stereogenic centre located within the pyrrolizine ring.
of racemates including several common pharmaceuticals has
Etodolac has the systematic chemical name (RS)-1,8-diethyl-
been limited to the synthesis, detection, and separation of 1,3,4,9-tetrahydropyrano[3,4-b]indole-1-acetic acid (Fig. 1). It is
diastereomeric derivatives, which are generally amides because used as an analgesic and for the treatment of rheumatoid
of the functional group compatibility of the chiral reagent and arthritis and osteoarthritis. It is considered relatively better as
this type of analytes. To the best of our knowledge, there is no a therapeutic agent as it produces less gastrointestinal toxicity
4
report in the literature that deals with the recovery of native in comparison with certain other NSAIDs.
Lal and Bhushan reported an efficient, simple, validated,
analytical, and semi-preparative HPLC method for the direct
Department of Chemistry, Indian Institute of Technology Roorkee, Roorkee 247667,
enantioresolution of (RS)-Ket using chiral columns of mono-
India. E-mail: rbushfcy54@gmail.com, rbushfcy@iitr.ac.in,
5
chloromethylated derivatives of cellulose and amylose. The
pmalik2289@gmail.com; Fax: +91-1332-286202; Tel: +91-1332-285795
enantioseparation of (RS)-Etd has also been reported by this group
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c7nj02898a
by different methods, which include (i) HPLC enantioseparation by
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017
New J. Chem.

View Article Online
Paper
NJC
discouraged in chemical methodology for a long time; (d) the
pre-column derivatization of (RS)-Ket reported by Vakily et al. in
16
1
995 has the drawback of rapid and complete racemization due
to the strongly basic conditions required for derivatization; and (e)
an indirect approach using (S)-(ꢀ)-a-(1-naphthyl)ethylamine as the
17
derivatizing reagent for the HPLC enantioseparation of (RS)-Etd
involves a very cumbersome derivatization reaction and multistep
sample preparation before the sample is loaded onto the column
although it has been claimed to be a simplified method with a
shortened derivatization time.
Taking into account the abovementioned literature, this
study was aimed at reporting a sensitive HPLC method for the
enantioseparation of (RS)-Ket and (RS)-Etd with the establishment
of the configuration of the diastereomeric derivatives, followed by
Fig. 1 Structures of (S)-naproxen and racemic drugs.
6a
pre-column formation of chiral ligand-exchange complexes;
(
ii) direct enantioresolution involving both achiral phases in TLC, easy recovery of the native enantiomers for their possible use.
6b
where a few L-amino acids are used as chiral inducing reagents;
iii) LC-MS separation of diastereomeric derivatives that have been chosen as the chiral moiety owing to its significant characteristics,
Therefore, (S)-(+)-naproxen (Npx, structure shown in Fig. 1) was
(
6c
synthesized using three different enantiomerically pure amines
helpful in enantioseparation by a derivatization approach, and a
[M + H] or [M] have been obtained for confirmation of the chiral reagent (CR) was synthesized and characterized.
formation and structures of the diastereomers); and (iv) TLC (S)-Naproxen (Npx) was chosen because of its large conjugated
+
+
(
enantioseparation of diastereomeric amide derivatives synthesized naphthyl ring having a higher molar absorptivity (e 4 100 000) as
2
using enantiomerically pure (R)-(ꢀ)-1-cyclohexylethylamine. A few compared to that of certain other chromophoric molecules such
other studies have been reported on the direct enantioseparation of as (S)-levofloxacin (e B 24 000), s-trichlorotriazine (e B 36 000),
7
,8
ꢀ1
ꢀ1
Ket using a polysaccharide-based CSP and a teicoplanin-based and the dinitrophenyl moiety (e B 30 000 mol cm ). Moreover,
9
stationary phase. Starek et al. reviewed TLC methods for the it is easily available as a pure (S)-enantiomer and does not require
determination of NSAIDs. The enantioseparation of Etd has been any other chiral auxiliary as is required by dinitrophenyl or
achieved on a chiral AGP (a1-acid glycoprotein) column and S-trichlorotriazine moieties to convert them into chiral reagents.
10
1
1
chiral stationary phases such as CHI-DMB [chemically known The carboxylic acid group of Npx is easily activated by the
0
as O,O -di(3,5-dimethylbenzoyl)-(2R,3R)-diallyltartardiamide] and introduction of different nucleophilic moieties, which serve as
(
R,R)-DNB-DPEDA [i.e., (R,R)-N-3,5-dinitrobenzoyl-1,2-diphenyl- effective leaving groups in a subsequent nucleophilic substitution
1
2
ethane-1,2-diamine], which are p-electron acceptor/donor reaction for the synthesis of diastereomeric derivatives. Its higher
stationary phases of broad applicability. molar absorptivity facilitates the sensitive detection of the deriva-
Literature related to the enantioseparation of Ket and Etd by tized analyte.
both direct and indirect approaches has been cited and discussed in
A (S)-naproxen-based CR was synthesized as its ester by a
the abovementioned reports; therefore, the same have not been simple straightforward reaction with 1H-benzotriazole and
1
mentioned herein. However, some of the drawbacks and limitations characterized by UV, IR, and H NMR spectroscopies, elemental
based on relevant literature reports have been briefly mentioned analysis, and polarimetry. It was used to synthesize diastereo-
hereinafter.
R)-(+)-l-(l-Naphthyl)ethylamine was used as a derivatizing The CR thus provided a pair of diastereomeric anhydrides for
reagent for (RS)-Ket, and the separation was carried out using each of the two analytes, and hydrolysis of the anhydrides
meric derivatives of the two NSAIDs via microwave irradiation.
(
1
3
an a-acid glycoprotein chiral column, but the experimental provided the native enantiomers under mild reaction condi-
procedure of sample preparation and analysis was very cumber- tions. The presence of carboxylic acid groups in both (RS)-Ket
some. Thus, there seems to be no reasonable justification for and (RS)-Etd is advantageous in the overall approach and in the
using a chiral derivatizing reagent (CDR) and then performing success of the experiments and results presented in this study.
separation on a chiral column. The other methods reported in The present study is novel and different from the existing
the literature suggest that there are several major limitations literature as it reports the semi-preparative HPLC enantio-
such as: (a) there have been a few reports on the direct enantio- separation of two analytes with easy recovery of the native
resolution of Ket and Etd using HPLC and chiral columns, but enantiomers without racemization and the establishment of
1
the latter are very expensive, have low durability and limited molecular asymmetry using H NMR and polarimetric studies.
applicability, and are very sensitive to the chromatographic
1
4
conditions; (b) the method described by Ing-Lorenzini et al.
for the determination of enantiomers of Ket involves a tedious
and time-consuming extraction procedure using liquid–liquid
extraction; (c) the HPLC separation of diastereomeric amides
involves the derivatization of (RS)-Ket with thionyl chloride/(S)-1- N-Hydroxybenzotriazole (OH-Btz), (S)-(+)-naproxen [[a]
phenylethylamine, but the use of thionyl chloride has been (c = 1, CHCl ), assay Z 98.0%], dicyclohexylcarbodiimide
2
. Materials and methods
1
5
2.1 Chemicals and equipment
2
5
D
= +661,
3
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017

View Article Online
NJC
Paper
(
(
DCC), and other reagents were purchased from Sigma-Aldrich The product in the filtrate was extracted with ethyl acetate and
St Louis, MO, USA). Pharmaceutical tablets of (RS)-ketorolac washed with brine, water, and saturated NaHCO separately.
3
marketed as Ketorol DT (Dr Reddy Laboratories Ltd, Hyderabad, Then, it was recrystallized from hot EtOH to obtain the reagent
India) containing 10 mg of an active pharmaceutical ingredient as an off-white solid. Characterization data of the CR are given in
and (RS)-etodolac marketed as Etova (Ipca Laboratories Ltd, the ESI.†
Mumbai, India) containing 400 mg of an active pharmaceutical
ingredient were purchased from a pharmacist’s shop in the local
market. All the reagents [concentrated HCl, sodium bicarbonate
2
.5 Microwave-irradiated synthesis of diastereomeric
derivatives
(
NaHCO
acetate (EtOAc), acetonitrile (MeCN), and methanol (MeOH)] of solutions of (RS)-Ket or (RS)-Etd (1 mM, 100 mL). To each
HPLC grade were purchased from E. Merck (Mumbai, India). mixture, a solution of the CR (1 mM, 200 mL) and triethylamine
3 3 4
), and phosphoric acid (H PO )] and solvents [ethyl Separate sets of reaction mixtures were prepared by taking
The HPLC system (LC-20 AD) was obtained from Shimadzu (20 mL) was added. The syntheses of the diastereomeric derivatives
Kyoto, Japan), the LiChrospher C18 column (length ꢁ i.d., were carried out independently. The experimental conditions were
(
2
5 cm ꢁ 4.6 mm, 5 mm particle size) was obtained from Merck optimized for best separation by varying the pH, ratio of racemic
(
Darmstadt, Germany); the preparative HPLC system (LC-20AP drug: CR (1:1, 1:1.5, 1:2, and 1:2.5), and microwave irradiation
Prominence preparative liquid chromatograph) was obtained (MWI) time (in the range from 60 s to 150 s at an interval of 30 s) at
from Shimadzu (Kyoto, Japan). 80% power (800 W). The diastereomeric mixture was subjected to
Other equipment/instruments used in the present investigations HPLC separation in each case.
were as follows: a microwave (Multiwave 3000, 800 W, PerkinElmer,
2.6 Preparative synthesis of diastereomeric derivatives
Shelton, CT, USA), a pH meter (CyberScan 510, Singapore), an FTIR
spectrometer (Nicolet 6700, Thermo Scientific, USA), an elemental The diastereomeric derivatives were synthesized on a preparative
analyzer (Vario EL III, Hanau, Germany), and a 400 MHz NMR level by scaling up the optimized synthesis conditions. (RS)-Ket
spectrometer (Jeol Inc., Peabody, USA). UV spectra were obtained in (150 mg, 59 mM) or (RS)-Etd (150 mg, 52 mM) was dissolved in
3
MeOH using the Shimadzu UV-2450 spectrophotometer. Double- 10 mL of 0.1 M NaHCO , and to each solution, a solution of the
distilled water was further purified using a Milli-Q system from CR (200 mg, 58 mM) in 20 mL acetonitrile was added. The yield
Millipore (Bedford, MA, USA) to obtain purified water (18.2 MO cm), was 320 mg (92%).
which was used throughout the experiments.
2.7 RPHPLC of diastereomeric derivatives
2
.2 Isolation and purification of the title compounds as
The diastereomeric mixture obtained from the reaction vial was
diluted 10 times using a solvent system consisting of MeOH
reference standards
–1
(RS)-Ket (M.W. 255 g mol ) was extracted, isolated, purified, and and TEAP buffer (10 mM, pH 3.5), and 20 mL of the above-
5
–1
characterized as reported earlier. (RS)-Etd (M.W. 287 g mol ) was mentioned mixture was injected onto the simple C18 column.
also extracted from the commercial tablets, isolated, purified, and The solvent system was degassed, sonicated, and filtered
characterized in a similar manner as described in an earlier through 0.45 mm filters before use.
6c
report. The lmax values, melting points, and characteristic IR
To optimize the separation conditions, the following varia-
peaks of the purified compounds were not included because the tions were carried out: the mobile phase was used in isocratic
focus of this study was on enantiomer resolution. The recovery was elution mode with (i) varying strengths of TEAP buffer (i.e.,
nearly 97% against the amounts reported on the commercial labels from 25 to 75% with a difference of 10% each time) for a period
in each case. The two compounds thus-obtained were used as of 5–25 minutes, (ii) buffer concentrations of 5 mM, 10 mM,
ꢀ
1
standard racemates. Calibration plots were developed using the and 15 mM, and (iii) flow rates of 0.5 and 1.0 mL min at each
standard references. stage of the experiment. The oven temperature was fixed at
40 1C. Detection was carried out at 320 nm using a PDA detector.
2.3 Standard solutions
2.8 Validation of the method
(a) (RS)-Ket and (RS)-Etd (1 mM each) were dissolved in 0.1 M
NaHCO , and the stock solutions were diluted with methanol to Validation studies were performed to assess recovery, linearity,
3
prepare the required working solutions.
accuracy, and precision using diastereomeric derivatives of
(
b) Triethylammonium phosphate (TEAP) buffer (10 mM) (RS)-Ket (as representatives) in accordance with the Interna-
1
8
was prepared, and its pH was adjusted to 3.5.
c) A solution of the CR (10 mM) was prepared in MeCN.
tional Council for Harmonisation (ICH) guidelines. Peak
(
areas (given by the system software) were used to determine
stability and recovery. The concentrations of the diastereomeric
2
.4 Synthesis and characterization of CR
ꢀ1
mixtures were in the range of 5–5000 ng mL . Linear plots of
Npx ester (Npx-OBtz). The CR was synthesized by the reaction of peak area vs. concentration were drawn, and the linear regres-
S)-(+)-Npx (5 mL, 0.2 M) with OH-Btz (5 mL, 0.2 M) in THF with sion equation was used to determine slopes and intercepts. The
(
the dropwise addition of DCC solution (3 mL, 0.36 M). The LOD and LOQ were calculated. To determine the precision, inter-
reaction mixture was stirred under a nitrogen atmosphere for day and intra-day assay studies have been carried out, and the
3
h. Dicyclohexylurea was precipitated and removed by filtration. results are presented as the relative standard deviation (RSD).
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017
New J. Chem.

View Article Online
Paper
.9 Isolation of diastereomeric derivatives by preparative
NJC
2
The methanolic solutions corresponding to the four enantiomers
were concentrated and allowed to crystallize after the addition of a
few drops of water. Crystals were obtained by decantation and then
air-dried and characterized.
HPLC
The separation conditions were optimized for preparative HPLC by
employing separation conditions similar to those used for analytical
RP-HPLC. Experiments were carried out using both isocratic and
gradient modes to achieve preparative enantioseparation using
MeOH and TEAP (pH 3.5) as a solvent system. Elution strength
was varied from 25% to 75% of TEAP buffer in a gradient over
3
. Results and discussion
The recovery of native enantiomers from diastereomeric derivatives
constitutes a difficult proposition because the experimental
conditions of hydrolysis may lead to either decomposition of
the enantiomers or racemization. Most of the time, the pure
enantiomer (and thus the corresponding diastereomer) of the
analyte is not available. Therefore, additional supporting inves-
tigations, such as isolation of diastereomers and recovery of
enantiomers (ensuring no racemization), would be required for
the determination of absolute configuration and confirmation
of the elution order; this has been achieved in the present case
ꢀ1
3
0 min. The flow rate was 5 mL min . The mixture of diastereomers
(as obtained from the preparative synthesis) was diluted with MeOH,
and 500 mL of the mixture was injected onto the C18 achiral column.
Then, two vials were placed in rack 1 in the sample collector block
for preparative HPLC. The time interval over which two peaks were
obtained was selected for the collection of separated fractions from
the fraction collector. A fraction of 2 mL was obtained for each
injection. Further injections were performed twenty times, and
fractions of nearly 40 mL were obtained during all the runs. The
fractions containing single and identical diastereomers were
combined and concentrated in vacuum. The diastereomeric
derivatives separated in this way were recovered, and their
purity was checked. These derivatives were designated as Ds-I
and Ds-II for the first and second eluted diastereomeric derivatives
of Ket and Ds-III and Ds-IV for the diastereomeric derivatives of
Etd, respectively. The characterization data of the two separated
diastereomers are given below.
1
by obtaining the H NMR spectra and specific rotation. We feel
that the success of the synthesis of diastereomeric derivatives
and their separation is also a subject to the recovery of native
enantiomers and verification of their configuration.
3.1 Chiral reagent
The reaction of organic amines with esters of carboxylic acids
has been the simplest method of acylation in organic synthesis.
However, with the use of DCC as a coupling reagent, both
2.10 Detagging of diastereomeric derivatives and isolation of activation and coupling proceed concurrently. The rate of
enantiomers
reaction of an amine with DCC is much lower in comparison
with the rate of addition of a carboxylic acid to one of the
double bonds of carbodiimide. Therefore, the synthesis of the
CR as an ester comprised the simple straightforward reaction of
Ds-I and Ds-II (100 mg each) were first dissolved individually in
a small amount of MeOH, and then, H SO (10 mL, 2 N) was
2 4
1
9
added. The solution(s) was irradiated under sealed vessel
conditions using MW for hydrolysis. The MWI system operated
at a frequency of 2.45 GHz with an 80% continuous microwave
irradiation power of 800 W. The experiment was performed in
the following manner.
(S)-(+)-Npx with OH-Btz in the presence of DCC. The CR has
been characterized (the relevant data are included in the manuscript)
and used for the synthesis of diastereomeric derivatives.
The sample of (S)-(+)-Npx obtained from Sigma-Aldrich was
first purified by recrystallization, and its enantiomeric purity
was ascertained polarimetrically. Because the reaction of
achiral OH-Btz took place only at the terminal carboxyl group
of (S)-(+)-Npx and not directly at the stereogenic centre, it was
considered that the CR synthesised in this way was enantio-
merically pure (supported by the characterization data). Since
no additional peaks were observed in the HPLC separation
chromatograms for the diastereomeric derivatives, it was further
inferred that the CR used for the synthesis of the diastereomeric
derivatives was enantiomerically pure. If there was any impurity in
the CR that corresponded to the (R)-(ꢀ)-isomer of naproxen,
additional peaks would have been obtained in the HPLC separation
chromatograms for the diastereomeric derivatives that would have
been formed from a small amount of the (R)-(ꢀ)-isomer of the CR.
(i) With each change (60, 90, 120, 150, and 180 s) in the MWI
time, samples were analyzed by HPLC as follows: 10 mL of each
solution was diluted ten times with the mobile phase, and 20 mL
of the diluted solution was applied to the simple C18 column.
Because the retention times of the diastereomeric derivatives of Ket
and Npx were different, two corresponding peaks were observed
until complete hydrolysis occurred. On the basis of the peak areas
(as given by the system software), it was observed that after a
few optimization runs, complete hydrolysis or conversion of the
individual diastereomeric anhydrides to their respective acids was
achieved within 6 min. (This was a simple achiral RPHPLC column;
therefore, one peak was obtained for an enantiomeric pair.)
(ii) The hydrolysates corresponding to Ds-I to Ds-IV were
freeze-dried, and the residues were extracted with water. It is
expected that only (R)- and (S)-Ket would go into the solution
because (S)-Npx is insoluble in water. The solution was centrifuged,
3.2 Diastereomeric derivatives
and (S)-Npx remained as a residue; the residue was further The optimized reaction conditions for the completion of the
shaken with water and centrifuged. The combined supernatant reaction in the synthesis of the diastereomeric derivatives were
was lyophilized, and the residue was dissolved in a small amount as follows: a 1 : 2 molar ratio of the analyte vs. CR and an MWI
of MeOH.
time of 150 s. It is known that the rates of reaction of
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017

View Article Online
NJC
Paper
enantiomers with chiral molecules are different. The two and substitutes the benzotriazole moiety. The overall reaction
enantiomers of the (RS)-analyte may react at different rates results in the formation of a pair of diastereomeric derivatives
with the CR, which is a chiral molecule. The derivatization containing an amide bond. However, in none of these reports,
reaction of (RS)-Ket or (RS)-Etd with the CR at different mole native enantiomers were obtained.
ratios exhibited slight kinetic resolution at a 1 : 1 ratio of
Moreover, there has been no report on the formation of
(RS)-Etd:CR. The completion of the derivatization reaction diastereoisomeric derivatives for any of the chosen analytes
was inferred from the identical peak areas (as provided by the [(RS)-Ket or (RS)-Etd] with a CR based on (S)-(+)-naproxen and
system software). The 94% recovery of the eluted diastereomers 1-hydroxybenzotriazole. It may be argued that the present CR,
provided an estimate of the yields. The effect of the MWI time being an ester with (S)-(+)-naproxen as the chiral moiety, is
on the completion of the derivatization reaction of (RS)-Ket is more reactive (in comparison with an amide) in forming the
shown in Fig. S1 in the ESI.†
diastereomeric derivatives of the chosen analytes under mild
A representative scheme for the synthesis of the CR and reaction conditions, and because of the mild conditions, the
diastereomeric derivatives of (RS)-Ket is shown in Fig. 2. Since formation of diastereomeric derivatives is faster and the chances
derivatization involved a reaction between the activated term- of racemization are reduced.
inal group of (S)-(+)-Npx and the carboxyl terminal group of the
A possible mode of bond formation leading to diastereo-
analyte [for example, (RS)-Etd], in neither of the two reactants, a meric derivatives in the form of anhydrides (as shown in Fig. 3)
direct attack at the stereogenic centre was involved, and no has been discussed hereinafter. In general, anhydrides and
racemization was detected throughout the study. Thus, the esters have intermediate reactivity.
following types of diastereomeric derivatives were considered
The electronegativities of N, C, and O atoms are 3.04, 2.55,
to have been formed: (a) [(R)-Ket-(S)-Npx], (b) [(S)-Ket-(S)-Npx], and 3.44, respectively, on the first scale developed by Linus
2
3,24
(c) [(R)-Etd-(S)-Npx], and (d) [(S)-Etd-(S)-Npx]. The first letter Pauling.
Moreover, the electronegativity of the carbonyl
[
capital (R) or (S)] represents the absolute configuration of the group ({CQO) calculated by the revised Lewis-Langmuir
2
5
enantiomer of the corresponding analyte, whereas the second equation is 2.706 ꢂ 0.01. The difference between the electro-
letter represents the absolute configuration of (S)-Npx.
negativities of carbon and oxygen is large enough to make the
have shown that a chiral CQO bond moderately polar. In the CR, there is a N atom (of the
2
0–22
Earlier reports in the literature
amide prepared from (S)-(+)-Npx and 1H-benzotriazole acts as a benzotriazole moiety) at the position of the hydroxyl hydrogen
CR for the formation of diastereoisomeric derivatives of various atom of the carboxyl group of the parent naproxen moiety.
pharmaceuticals and other compounds containing an amino Thus, the carbonyl carbon atom ({CQO) becomes more electron-
group; the amino group acts as a nucleophile, attacks the CR, deficient owing to the presence of highly electronegative elements
Fig. 2 Representative scheme of the synthesis of the CR and diastereomeric derivatives of (RS)-Ket.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017
New J. Chem.

View Article Online
Paper
NJC
Fig. 3 Plausible mechanistic route of anhydride synthesis.
(O and N) in its close vicinity (II in Fig. 3). The literature suggests that atoms); therefore, its formation is favored. The presence of a weak
the electronegativity depends on the environment of a particular base (triethylamine) during the derivatization reaction would
26–28
atom within a molecule as well as on its degree of hybridization.
facilitate the formation of the abovementioned diastereomeric
In the present situation of the synthesis of diastereomeric derivatives by making the analyte more nucleophilic in nature,
ꢀ
derivatives, the COO group of the analyte acts as a nucleophile and excess substrate molecules would not react with each other
and attacks the electrophilic carbon atom of the CR, and the to form naproxen anhydride.
pi-bond electrons undergo resonance to the more electronegative
3
.3 HPLC separation of diastereomeric mixtures
oxygen atom to form a tetrahedral adduct (as a transitional
intermediate III shown in Fig. 3) via a carbon-nucleophile bond It was observed that the same chromatographic conditions
2
3
with a change in hybridization from sp to sp . These inductive were successful for the separation of diastereomeric pairs of
and resonance effects result in the carbonyl carbon atom both analytes; these included a mobile phase that comprised a
having a large d charge and high reactivity, and it is thus MeOH–TEAP buffer (pH 3.5) at a flow rate of 1.0 mL min in
+
ꢀ1
ꢀ
easily attacked by a nucleophile (the COO group in the (R)- or the isocratic mode (from 75% to 25%) using a C18 column and
(S)-enantiomer of Ket or Etd in the racemic analyte). Moreover, detection at 320 nm.
the OH-benzotriazole moiety acts as a good leaving group and is
displaced very easily.
Actual chromatograms (obtained from the system) showing
the baseline separation of the pairs of diastereomeric deriva-
Thus, the overall reaction leads to the formation of the product tives of (RS)-Ket and (RS)-Etd are given in Fig. 4(i) and (ii),
(
IVa and b in Fig. 3) as a pair of diastereomeric derivatives, which respectively. The diastereomeric derivatives of both racemic
are structurally anhydrides. This structure is more stable owing to analytes displayed very good separation. Table 1 summarizes
stronger resonance or greater delocalization (with two oxygen the values of the retention factor k, separation factor a, and
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017

View Article Online
NJC
Paper
Fig. 4 Actual chromatograms (obtained from the system) showing the separation of the diastereomeric derivatives of (i) (RS)-Ket and (ii) (RS)-Etd.
Chromatographic conditions: column: LiChrospher C18 (250 mm ꢁ 4.6 mm I.D., particle size 5 mm); mobile phase: MeOH–TEAP buffer (pH 3.5), binary
ꢀ1
gradient (from 75% to 25%); flow rate: 1 mL min ; detection: 320 nm; and injection volume: 10 mL. The first elution peak corresponds to the (R,S)-
diastereomeric derivative in both the cases. The x-axis shows the time in min, and the y-axis shows the absorbance in mAU.
Table 1 HPLC separation data for diastereomeric derivatives of racemic
analytes
The peak areas, as obtained from the system software, were
similar, namely, 236 390 and 236 274 for the peaks of the two
diastereomeric derivatives of Ket and 288 726 and 288 364 for the
peaks of the two diastereomeric derivatives of Etd. An equimolar
mixture of the diastereomers, i.e., [(R)-Ket-(S)-Npx] and [(S)-Ket-(S)-
Npx], which was separated and isolated by preparative HPLC, was
run independently under the same optimized HPLC separation
conditions as the diastereomeric derivatives. The results were in
Characteristics of the separated diastereomers
Racemic analyte
k
1
k
2
a
Rs
(
(
RS)-Ketorolac
RS)-Etodolac
0.89
0.41
2.16
1.90
2.42
4.63
6.1
7.2
resolution Rs for the diastereomeric derivatives separated agreement in terms of the retention time and peak area.
under the optimized HPLC conditions. It was observed that A comparison of the present results with the literature shows
the resolution in the separation of the diastereomeric derivatives that the indirect approach is better in terms of selectivity,
of Etd was higher than that for the diastereomeric derivatives of sensitivity, lower costs, and easier experimental optimization.
Ket, whereas the retention time of the diastereomeric derivatives In contrast, CSPs are very expensive, the optimization of the
of Ket was higher.
separation conditions is cumbersome and time-consuming,
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017
New J. Chem.

View Article Online
Paper
NJC
1
and the detection of the underivatized enantiomers by UV spectro-
scopy is less sensitive in the absence of a strong chromophore.
3.5.1 H NMR spectra of the diastereomeric derivatives.
1
2
3
The chemical shifts of four (a and b) protons, i.e., H , H , H ,
The retention times were compared with those of the and H , in the two diastereomeric derivatives (as marked in
diastereomers Ds-I and Ds-II, which were separated by prepara- Fig. 2) were observed in their H NMR spectrum (Fig. 5). The
4
1
tive HPLC and run independently and as an equimolar mixture. relative orientation of the nucleus with respect to the external
On the basis of the data obtained from the system software, the field changes its chemical shift (d) and the ring current induced
resolution (Rs) of the diastereomeric derivatives was calculated by the delocalized p-electrons in the aromatic rings. It is clearly
1
2
as per the standard formula (as described in the ESI†).
seen that in the first eluted diastereomer (Ds-I), the H , H , and
3
4
Mobile phase: MeOH–TEAP buffer of a pH of 3.5 (75:25, v/v) H protons (shown in Fig. 2) are more deshielded, and the H
ꢀ
1
in the isocratic mode at a flow rate of 1 mL min ; k
1
=
proton is more shielded, as evidenced by the chemical shift
retention factor of the diastereomeric derivative (Ds-I); observed for the second diastereomer (Ds-II). The protons that
k2 = retention factor of the diastereomeric derivative (Ds-II); lie outside the ring experience deshielding because the induced
a = separation factor; and Rs = resolution.
magnetic field has the same direction outside the ring as the
external field. In contrast, the protons lying in the vicinity of the
aromatic ring experience shielding because both fields are in
opposite directions.
3.4 Isolation by preparative HPLC and characterization of
diastereomeric derivatives
The chromatographic separation conditions optimized for analytical
By observing the shielded and deshielded chemical shift
separation were adopted for preparative-scale separation and values, the conformations were drawn in a 3D view using
adjusted suitably. A mobile phase comprising MeOH-TEAP buffer the Chem3D Pro 12.0 software. Using these structures, all the
ꢀ1
in a ratio of 75 : 25 in the isocratic mode at a flow rate of 5 mL min
d-values could be explained well. Fig. 6 shows the 3D structures
was found to be optimal for the isolation of the derivatives of of different conformations of the (R,S)-diastereomeric deriva-
1
(RS)-Ket.
tive of Ket prepared with the CR. In Fig. 6a, the H proton (as
The diastereomeric derivatives isolated by preparative HPLC already marked in Fig. 2 and correlated with the NMR spectrum
1
were characterized by obtaining the H NMR, UV, and IR shown in Fig. 5) lies in the plane of p-electron-containing rings
spectra, specific rotation, and melting point, and CHN analysis (but outside the rings), and thus experiences a deshielding
2
4
results. The characterization data of the two diastereomeric effect, whereas the H proton in Ket (in Fig. 6b) and the H
derivatives of (RS)-Ket are given in the ESI.†
proton in the CR (in Fig. 6d) point towards the ring current of
the pyrrole moiety in Ket. The H proton on the asymmetric
carbon atom of the CR (Fig. 6c) points away from both the
3
3.5 Structures and configurations of diastereomeric derivatives
The three-dimensional chemical structures and configurations p-electron-containing rings.
of the diastereomeric derivatives of (RS)-Ket (as representa-
On the other hand, the 3D structures of different conformations
tives), together with the spatial orientation of groups with drawn for the (S,S)-diastereomeric derivative are shown in Fig. 7.
2
respect to the anhydride bond (as shown in Fig. 3), were The H proton (Fig. 7b) lies in the ring current of the naphthalene
3
examined and confirmed by developing molecular models ring, whereas the H proton (Fig. 7c) points towards the ring current
1
4
using the orbital molecular building system (Cochranes of of the pyrrole ring. The H and H protons (Fig. 7a and d,
Oxford Ltd, Leafield, Oxford OX8 5NT, England).
respectively) point away from all the rings.
1
Fig. 5 Sections of H NMR (400 MHz using CDCl
3
) spectra of diastereomeric derivatives, illustrating the differences in d values for a and b protons
corresponding to asymmetric carbon atoms.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017

View Article Online
NJC
Paper
Fig. 6 3D structures of different conformations (drawn using Chem3D
Pro 12.0 software) of the (R,S)-diastereomeric derivative of Ket prepared
1
with the CR. In (a) the H proton (as already marked in Fig. 2 and correlated
Fig. 7 3D structures of different conformations of the diastereomeric
derivative of Ket with the (S,S)-configuration prepared with the CR, as
with the NMR spectrum shown in Fig. 5) lies in the plane of the rings that
contain p-electrons, but outside the rings, and thus experiences
deshielding effect, whereas the H proton in Ket (in (b)) and the H proton
in the CR (in (d)) point toward the ring current of the pyrrole moiety in Ket.
2
a
drawn using Chem3D Pro 12.0 software. The H proton (b) lies in the ring
2
4
3
current of the naphthalene ring, whereas the H proton (c) points toward
1
4
the ring current of the pyrrole ring. The H and H protons (a and d,
respectively) point away from all the rings.
3
The H proton on the asymmetric carbon atom of the CR ((c)) points away
from both rings that contain p-electrons.
The recoveries of the enantiomers of both analytes [(RS)-Ket
Both pyrrole and naphthalene rings have strong anisotropic and (RS)-Etd] and that of (S)-(+)-Npx were in an appropriate
shielding effects in the presence of an external magnetic field molar ratio and in good yields.
1
that result in a chemical shift difference of 0.15 d for H protons
All the native enantiomers and (S)-(+)-Npx that were isolated
between the two diastereomeric derivatives. Similarly, the and purified were characterized by obtaining their specific
2
3
4
1
differences for H , H , and H were observed to be 0.16, 0.08, rotation and spectroscopic techniques such as H NMR, HRMS,
and ꢀ0.14 d, respectively (Fig. 5). Hence, the pyrrole ring and IR spectroscopies. The characterization data of the four
2
4
shields the H and H protons (as shown in Fig. 6b and d) purified enantiomers and (S)-Npx are presented in the ESI.†
3
2
and the H proton (in Fig. 7c), whereas the H proton (Fig. 7b)
experiences a shielding effect due to the naphthalene ring.
3.7 Elution order
Because the naphthalene ring contains more p-electrons, its The determination of the elution order of the species being
anisotropic effect will be greater, and it will shield to an even separated in a chromatographic experiment is an important
greater extent, as seen in Fig. 7b with respect to Fig. 6b.
criterion for establishing the success of the analytical procedure.
On the basis of these figures and values, it is inferred The diastereomeric derivatives were obtained by preparative
that the first eluted diastereomeric derivative corresponds to HPLC as the first and second eluted species. Furthermore, the
[(R)-Ket-(S)-Npx], and the second eluted diastereomeric derivative native enantiomers were isolated by the hydrolysis of each of
corresponds to [(S)-Ket-(S)-Npx]. These results are in agreement these derivatives and characterized; thus, the elution order was
with the calculated values of the specific rotation of the pure very clearly established.
enantiomers obtained by detagging of the corresponding dia-
3.8 Method validation
stereomeric derivatives.
Validation studies were carried out using ketorolac as a repre-
sentative. Sample solutions of both diastereomeric derivatives
3.6 Detagging of the diastereomeric derivatives
One of the objectives of the present investigations was to obtain of Ket were prepared in the concentration range from 45 to
ꢀ
1
the native enantiomers from the diastereomeric derivatives 5000 ng mL . Replicate samples were analyzed (n = 5) by HPLC
ꢀ
1
after their separation by HPLC. Therefore, experiments were using concentrations (45, 500, 2000, 3500, and 5000 ng mL of
designed to obtain the derivatives that could be detagged under the individual diastereomeric derivatives) of (RS)-Ket prepared
relatively mild conditions and thus anhydrides as these are reactive with the CR.
and undergo easy hydrolysis. Acid hydrolysis did not involve
Data for the accuracy and precision of the method were
any reaction at the stereogenic centre. Therefore, racemization determined according to ICH guidelines. The acceptance criterion
was neither theoretically expected nor experimentally observed. for accuracy is ꢂ15% of the reference value, and the found mean
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017
New J. Chem.

View Article Online
Paper
NJC
value is within 15% of the actual value taken. In addition, the diastereomeric derivatives that can be easily hydrolysed to the
precision determined at each concentration is within 15% of the native enantiomers without racemisation; (ii) for measuring the
coefficient of variation (CV). Regression equations were obtained by success of enantioseparation methods and the synthesis of
plotting graphs of the concentration of the diastereomeric deriva- diastereomers; (iii) for controlling the enantiomeric purity of
tive (x) against the peak area (y) and found to be y = 4.97x + 4721.4 chiral drugs in industrial and analytical laboratories (in particular,
2
2
(
r = 0.99) and y = 4.96x + 4723 (r = 0.99) for the first and second those associated with regulatory agencies); and (iv) for use in
eluted diastereomeric derivative, respectively. The slope, intercept, different stages of enantioselective synthesis as there is generally
relative standard deviation (RSD), recovery, and linearity were a chance of error in establishing the ee of a sample because a
determined using these regression equations.
purification step that employs conventional chromatography prior
The calculated RSD values for the (R,S)- and (S,S)-diastereomeric to loading the sample onto a chiral column may lead to unobserved
29
derivatives were 0.25–0.33% and 0.17–0.25%, respectively, for inter- enantiomeric enrichment.
day assay precision and 0.24–0.41% and 0.22–0.48%, respectively, Although the reported method represents an indirect
for intra-day assay precision. The solutions were considered stable approach to enantioseparation, it is simpler, as separation is
if the deviation from the nominal value was within ꢂ10.0%. The possible on an achiral C18 column with easy optimization of
calculated recoveries for the first and second eluted diastereomeric the chromatographic conditions, and the CR provides highly
derivatives were 99.4–100.6% and 99.2–100.1%, respectively, for the sensitive detector response in comparison with chiral columns,
inter-day assay and 99.4–102.3% and 99.3–102.7%, respectively, for which are highly expensive.
the intra-day assay, respectively (Table S2, given in the ESI†). The
method was inferred to be accurate owing to the very low RSD
values and high percentage recoveries.
Conflicts of interest
The limits of detection and quantification, i.e., the LOD and
LOQ, values were also determined to perform validation studies.
There are no conflicts to declare.
ꢀ
1
LOQs of 11.19 and 9.15 ng mL
3
and LODs of 3.69 and
ꢀ
1
.02 ng mL were found for the diastereomeric derivatives of
Acknowledgements
(R)- and (S)-Ket, respectively. Herein, the LODs are much lower
than the prescribed LOD (1%) in the pharmaceutical industry.
The authors gratefully acknowledge the financial assistance
received from the Council of Scientific and Industrial Research,
New Delhi, India, in the form of a senior research fellowship
4
. Novelty of the present work and
(to Poonam Malik).
conclusion
The enantioseparation of two commonly used drugs, namely,
References
(RS)-Ket and (RS)-Etd, has been achieved by both analytical and
preparative HPLC. Although a preparative HPLC setup may be
considered expensive in comparison with the preparative
separation by conventional open silica gel column chromato-
graphy, the latter is more cumbersome and time-consuming
because it requires verification/analysis of each fraction that is
collected manually.
The present method provides a successful approach for the
easy recovery of the native enantiomers of (RS)-Ket and (RS)-Etd
after their enantioseparation via derivatization. It does not
require a reference standard for comparison of the retention
times and is capable of establishing the true elution order. The
confirmation of the molecular asymmetry of the diastereomeric
derivatives, followed by the verification of the configuration of
the enantiomers recovered from the derivatives, makes this
method superior to those reported in literature in the area of
enantioseparation, where the use of a CR remains limited to
derivatization followed by separation of the diastereomeric
derivatives.
1 S. Alwera and R. Bhushan, Biomed. Chromatogr., 2016, 30,
1223–1233.
2 A. Dalal and R. Bhushan, J. Planar Chromatogr., 2016, 29,
366–371.
3 A. Macario and A. G. Lipman, Pain Med., 2001, 2, 336–351.
4 Modern Pharmacology with Clinical Applications, ed. R. Craig
and R. E. Stitzel, Lippincott Williams and Wilkins, Philadelphia,
USA, 6th edn, 2004.
5 M. Lal and R. Bhushan, Biomed. Chromatogr., 2016, 30,
1526–1534.
6 (a) M. Singh and R. Bhushan, Biomed. Chromatogr., 2016, 30,
1728–1732; (b) M. Singh and R. Bhushan, J. Planar Chromatogr.,
2016, 29, 184–189; (c) M. Singh and R. Bhushan, Biomed.
Chromatogr., 2015, 29, 1330–1337.
7 K. S. S. Dossou, E. Farcas, A.-C. Servais, P. Chiap,
B. Chankvetadze, J. Crommen and M. Fillet, J. Chromatogr. A,
2012, 1234, 56–63.
8 I. Matarashvili, D. Ghughunishvili, L. Chankvetadze,
N. Takaishvili, T. Khatiashvili, M. Tsintsadze, T. Farkas
and B. Chankvetadze, J. Chromatogr. A, 2017, 1483, 86–92.
9 O. H. Ismail, A. Ciogli, C. Villani, M. D. Martino, M. Pierini,
A. Cavazzini, D. S. Bell and F. Gasparrini, J. Chromatogr. A,
2016, 1427, 55–68.
This method and approach are useful for both academic and
industrial scientists in all areas of chemical science. The
method is self-contained and comprehensive and also endorses
the future usefulness of the reagents involved. It opens up
areas (i) for the development of new CRs that would yield
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017

View Article Online
NJC
Paper
1
1
1
1
1
0 M. Starek, M. Dabrowska and J. Skucinski, Recent Pat. Chem. 19 M. Bodanszky, Peptide Chemistry: A Practical Textbook,
Eng., 2010, 3, 49–73. Springer-Verlag, New York, 1993, 2nd edn, pp. 58–68.
1 H. P. Jadhav and D. B. Pathare, Int. J. Pharmacol. Pharm. Sci., 20 R. Bhushan and R. Dubey, J. Chromatogr. A, 2011, 1218,
015, 7, 77–80. 3648–3653.
2 X. Zhang, Z. Li, B. Shen, J. Chen and X. Xu, J. Anal. Sci., 21 R. Bhushan and H. Nagar, Biomed. Chromatogr., 2013, 27,
Methods Instrum., 2012, 2, 18–23. 750–756.
3 I. Tsina, Y. L. Tam, A. Boyd, C. Rocha, I. Massey and 22 S. Batra and R. Bhushan, Biomed. Chromatogr., 2014, 28,
T. Tarnowski, J. Pharm. Biomed. Anal., 1996, 15, 403–417. 815–825.
2
4 K. R. Ing-Lorenzini, J. A. Desmeules, M. Besson, J. L. Veuthey, 23 http://www.webelements.com.
P. Dayer and Y. Daali, J. Chromatogr. A, 2009, 1216, 24 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC
3
851–3856.
5 P. Hayball, J. Tamblyn, Y. Holden and J. Wrobel, Chirality,
993, 5, 31–35.
6 M. Vakily, B. Corrigan and F. Jamali, Pharmacol. Res., 1995,
2, 1652–1657.
Press, Boca Raton, Florida, 2003, 84th edn, Section 9,
Molecular Structure and Spectroscopy, Electronegativity.
25 D. Leah, G. O’Neale, F. B. Alcindor, T. L. Meek and
B. G. Patrick, THEOCHEM, 2003, 639, 151–156.
1
1
1
1
1
1
26 J. E. Huheey, J. Phys. Chem., 1965, 69, 3284.
7 Y. X. Jin, Y. H. Tang and S. Zeng, J. Pharm. Biomed. Anal., 27 A. D. Walsh, Discuss. Faraday Soc., 1947, 2, 18–24.
2008, 46, 953–958.
28 L. Pauling, Nature of the Chemical Bond, Cornell University
8 International Conference on Harmonization, Q2B Docu-
Press, 1960, pp. 88–107.
ment, Validation of Analytical Procedures, International 29 J. Martens and R. Bhushan, Helv. Chim. Acta, 2014, 97,
Conference of Harmonization, Geneva, 1996. 161–187.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2017
New J. Chem.