Characterization of Photodegradation Products of Bepotastine Besilate and In Silico Evaluation of Their Physicochemical, Absorption, Distribution, Metabolism, Excretion and Toxicity Properties

Article abstract of DOI:10.1016/j.xphs.2020.03.004

Bepotastine (BPT) is a H₁-receptor antagonist. It is used as a besilate salt in ophthalmic solution for allergic conjunctivitis and orally for the treatment of allergic rhinitis and urticaria/pruritus. Its systematic forced degradation study is unreported. The same was carried out in different conditions prescribed by International Conference on Harmonisation. The stressed solutions were subjected to reversed phase liquid chromatographic analysis, and BPT was observed to be labile under photobasic condition only, yielding 5 photodegradation products. The structures of the latter were elucidated from data generated by liquid chromatography–high-resolution mass spectrometry and multistage mass spectrometry. Of the 5, 4 products were further isolated and subjected to nuclear magnetic resonance spectroscopy to justify the proposed structures. Two of them, with similar accurate mass, were additionally and unambiguously characterized from their heteronuclear multiple bond correlation data, hydrogen deuterium exchange mass data, and quantum chemical analysis using density functional theory calculations. One degradation product had a structure that could only be explained by unusual rearrangement involving conversions of N-oxide into hydroxylamine, similar to Meisenheimer rearrangement. The physicochemical, as well as absorption, distribution, metabolism, excretion, and toxicity properties of BPT and its characterized photodegradation products were evaluated in silico by ADMET Predictor software.

Full text of DOI:10.1016/j.xphs.2020.03.004

Journal Pre-proof
Characterization of photodegradation products of bepotastine besilate and in silico
evaluation of their ADMET properties
Dilip Kumar Singh, Archana Sahu, Aabid Abdullah Wani, Prasad V. Bharatam, Asit K.
Chakraborti, Sanjeev Giri, Saranjit Singh
PII:
S0022-3549(20)30128-3
https://doi.org/10.1016/j.xphs.2020.03.004
XPHS 1894
DOI:
Reference:
To appear in: Journal of Pharmaceutical Sciences
Received Date: 19 September 2019
Revised Date: 13 February 2020
Accepted Date: 6 March 2020
Please cite this article as: Singh DK, Sahu A, Wani AA, Bharatam PV, Chakraborti AK, Giri S, Singh
S, Characterization of photodegradation products of bepotastine besilate and in silico evaluation of
their ADMET properties, Journal of Pharmaceutical Sciences (2020), doi: https://doi.org/10.1016/
j.xphs.2020.03.004.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier Inc. on behalf of the American Pharmacists Association.

Characterization of photodegradation products of bepotastine besilate and in
silico evaluation of their ADMET properties
Dilip Kumar Singha, Archana Sahua, Aabid Abdullah Wanib, Prasad V. Bharatamb, Asit K.
Chakrabortib,c, Sanjeev Girid, Saranjit Singha*
aDepartment of Pharmaceutical Analysis, National Institute of Pharmaceutical Education and
Research (NIPER), Sector 67, S.A.S. Nagar 160 062, Punjab, India
bDepartment of Medicinal Chemistry, National Institute of Pharmaceutical Education and
Research (NIPER), Sector 67, S.A.S. Nagar 160 062, Punjab, India
cPresent address: Department of Chemistry, Indian Institute of Technology Ropar, Rupnagar 140
001, Punjab, India
dDMPK and Pharmaceutical Development, Aurigene Discovery Technologies Limited,
Hyderabad 500 049, Telangana, India
*Corresponding Author: ssingh@niper.ac.in

Highlights
•
•
•
Bepotastine (BPT) besilate was found to degrade to five products in photobasic stress
conditions.
Structure characterization of the photodegradation products was done using LC-HRMS,
LC-MSn and NMR.
Comprehensive mass fragmentation profiles were established for BPT and its
photodegradation products.
•
•
The photodegradation pathway of BPT besilate was postulated.
On in silico evaluation of ADMET properties, two photodegradation products showed
mutagenicity potential as compared to BPT.

Abstract
Bepotastine (BPT) is a H1-receptor antagonist. It is used as a besilate salt in ophthalmic solution
for allergic conjunctivitis, and orally for the treatment of allergic rhinitis and utricaria/pruritus.
Its systematic forced degradation study is unreported. The same was carried out in different
conditions prescribed by International Conference on Harmonisation. The stressed solutions
were subjected to reversed phase liquid chromatographic analysis, and BPT was observed to be
labile under photobasic condition only, yielding five photodegradation products. The structures
of the latter were elucidated from data generated using liquid chromatography-high resolution
mass spectrometry and multi-stage mass spectrometry. Of the five, four products were further
isolated and subjected to nuclear magnetic resonance spectroscopy to justify the proposed
structures. Two of them, with similar accurate mass, were additionally and unambiguously
characterized from their heteronuclear multiple bond correlation data, hydrogen deuterium
exchange mass data, and quantum chemical analysis using density function theory calculations.
One degradation product had a structure that could only be explained by unusual rearrangement
involving conversions of N-oxide into hydroxylamine, similar to Meisenheimer rearrangement.
The physicochemical, as well as absorption, distribution, metabolism, excretion and toxicity
properties of BPT and its characterized photodegradation products were evaluated in silico by
ADMET Predictor™ software.
Keywords: Bepotastine, Bepotastine besilate, Chemical stability, Photodegradation, Degradation
products(s), Liquid Chromatography-Mass Spectrometry (LC-MS), Nuclear Magnetic
Resonance (NMR) spectroscopy

1. Introduction
Forced degradation studies (stress studies) on drugs have multiple benefits and are an essential
regulatory requirement [1]. These play a vital role in all stages of drug development [2].
According to ICH Q1A(R2) guideline, stress testing should explore the effect on the drug
substance of temperature above that recommended for accelerated testing, humidity (e.g., 75%
RH or greater) where appropriate, hydrolysis across a wide range of pH values, and oxidation
[3]. Additionally, photostability testing is also recommended, as it reveals intrinsic stability of
the drug upon exposure to light. It helps further in separation, identification and characterization
of photodegradation products. A systematic approach to photostability testing, which acts as a
prognosticator of prospective photostability concerns, is proposed separately in ICH Q1B
guideline [4].
Bepotastine (BPT) is a selective second-generation histamine H1-receptor antagonist, which is
marketed as an ophthalmic solution (BepreveTM) in the United States for the treatment of allergic
conjunctivitis. It is also approved in Japan, where it is marketed as tablets (Talion®) for the
treatment of allergic rhinitis and utricaria/pruritus [5-7]. The marketed formulations contain the
BPT as a besilate salt, which is chemically (+)-(S)-4-{4-[(4-chlorophenyl)(2-pyridyl)methoxy]
piperidino} butyric acid monobenzenesulfonate. Only a few reports exist on the development of
stability-indicating methods for separation of BPT from its known impurities and degradation
products (DPs) [8-11]. Reports on systematic stress degradation studies and structural
characterization of DPs of BPT are lacking.
Therefore, we carried out systematic forced degradation studies by following the protocols
established earlier in our laboratory [12]. The primarily focus was to: i) separate and identify
DPs of BPT formed under forced degradation conditions; ii) characterize their structure using
high-resolution mass spectrometry (HRMS), multi-stage mass spectrometry (MSn) and nuclear
magnetic resonance (NMR) spectroscopy; iii) provide mechanistic explanation to their
formation, and iv) predict in silico their physicochemical as well as absorption, distribution,
metabolism and toxicity (ADMET) properties using ADMET PredictorTM software. The results
are reported herein.
2. Experimental
2.1. Drug and Reagents

Pure BPT besilate was provided as a gratis sample by Intera Healthcare (Hyderabad, India).
Hydrochloric acid (HCl), sodium hydroxide (NaOH), hydrogen peroxide (H2O2) and buffer salts
were of analytical grade and purchased from local manufacturers. TEMPO ((2,2,6,6-
Tetramethylpiperidin-1-yl)oxyl) was sourced from Spectrochem Pvt. Ltd. (Mumbai, India).
HPLC grade methanol was procured from J.T. Baker, Phillipsburg, NJ, USA. Deuterated water
(D2O) and deuterated dimethyl sulfoxide (DMSO-d6) were procured from Sigma-Aldrich (St.
Louis, Missouri, USA). A bench-top purification system (PURELAB®, ELGA LabWater, High
Wycombe, UK) was employed to obtain ultra-pure water.
2.2. Apparatus and Instruments
General apparatus (like oil bath, Dri-Bath, pH analyzer, sonicator, precision analytical balance,
autopipettes, etc.), stability and photostability chambers, and instruments (like HPLC, UPLC-
HRMS (OrbitrapTM), LC-ion trap MS (MSn), and NMR spectrometer) were of same model,
make, and even of same specifications as outlined in our recent publication [13].
2.3. Forced degradation studies on BPT besilate and sample preparation for HPLC
analysis
2.3.1. Hydrolytic studies
For hydrolytic studies, stock solution (4 mg/mL) of BPT besilate was prepared in H2O. Acidic
and alkaline hydrolyses were carried out at 75 °C in 1N HCl and 1N NaOH, while neutral
hydrolysis was performed in H2O. The concentration in all the stressed solutions was 2 mg/mL.
The samples were withdrawn at suitable time intervals. Before injecting into HPLC, acidic and
basic stressed samples were neutralized by adding opposite solution of base and acid,
respectively, while neutral condition samples were simply diluted with H2O. The concentration
at the time of injection was 1 mg/mL.
2.3.2. Oxidative studies
Oxidative degradation study was carried out in 3% v/v H2O2 at room temperature for 5 d. The
stressed samples were diluted with water to obtain BPT besilate concentration of 1 mg/mL
before HPLC analysis.
2.3.3. Photodegradation studies
Photodegradation studies were carried out in solution as well as in solid state. For preparation of
acid, base and neutral solution samples, 2.5 mL stock solution of BPT besilate (4 mg/mL) was

added to separate 5 mL quartz volumetric flasks, each containing 2.5 mL of 0.2 N HCl, 0.2 N
NaOH and water, respectively. Accordingly, the final molar concentration of HCl and NaOH
was 0.1 N, and BPT besilate concentration was 2 mg/mL in each sample. To evaluate the
involvement of radical pathway in photodegradation of BPT [14, 15], the above studies in
solution state were carried out with added 2.5 mM TEMPO. The flasks were exposed to 4500 lx
cool white and 0.8 W/m2 UV-A radiations in the photostability chamber complying to Option 2
of the ICH guideline Q1B [4]. The samples were withdrawn from time to time. Solid BPT
besilate was exposed in thin layer in petridish to 6 million lx cool white and 1000 W/m2 UV-A
radiations. In all the cases, dark controls wrapped in aluminium foil were exposed parallely. The
solutions for HPLC analysis were prepared by dilution to BPT besilate concentration of 1
mg/mL.
2.3.4. Thermal studies
For thermal stress study, solid BPT besilate in sealed glass ampoules was heated in a Dri-Bath
(Thermolyne, Dubuque, IA, USA) at 60 °C for 21 d.
2.4. LC method development and optimization
An acceptable separation of the drug and its DPs was achieved on the reversed phase column by
optimizing various LC parameters, viz., stationary phase, mobile phase composition, pH, buffer
concentrations, flow rate and column temperature. The optimized conditions for HPLC
(analytical and semi-preparative) and LC-MS (HRMS and MSn) analyses are given in Table 2.
2.5. Enrichment and isolation of photobasic degradation products (DPs 2-5)
For enrichment of photobasic DPs 2-5, solutions containing 200 mg/mL drug were prepared in
0.2 N NaOH. These were exposed to cool white and UV-A radiations, as mentioned above, for
10 d in the photostability chamber. The extent of formation of the DPs was monitored by HPLC
(Table S1). Once sufficient conversion of BPT into DPs was achieved, isolation of the latter was
attempted by multiple injections on a semi-preparative column (Table S1). The collected
fractions were dried on a rotavapor (BUCHI R-114, Flawil, Switzerland).
2.6. Preparation of BPT as free base
To avoid overlapping of the proton and carbon signals of besilate with signals of BPT during
NMR acquisition, desalting was done using a semi-preparative method (Table S1).

2.7. Mass studies (HRMS, MSn, LC-HRMS and LC-MSn) on BPT and its degradation
products (DPs 1-5)
HRMS and MSn mass studies were carried out on the drug in ESI positive mode for the
establishment of comprehensive fragmentation pathway. Five ppm BPT besilate solution in
methanol:water (50:50 v/v) was directly infused into respective mass spectrometers at a flow rate
of 10 µl/min. The optimized mass parameters are outlined in Tables S2 and S3, respectively.
These were used for acquisition of data for protonated molecular ions of BPT, DPs and their
fragments. The stressed samples were subjected to LC-HRMS and LC-MSn studies in the same
ionization mode to establish fragmentation pathway of each DP. The hydrogen deuterium
exchange (HDX) mass studies were carried out by preparing separate solutions of isolated BPT,
DP 3 and DP 4 fractions in deuterated water (D2O). These solutions were directly infused into
the MSn system.
2.8. NMR studies on BPT and its isolated degradation products (DPs 2-5)
For collection of NMR data, BPT (free base) and isolated DPs 2-5 were dissolved in DMSO-d6.
1D (1H, 13C and DEPT-135) and 2D (COSY, TOCSY, HSQC and HMBC) NMR spectra were
1
acquired on a 500 MHz NMR spectrometer equipped with PFG 5 mm BBO probe at 298 K. H
NMR spectra of the BPT and DPs were recorded using the double presaturation pulse sequence
for the suppression of moisture and residual signals. 1H NMR chemical shifts were referenced to
tetramethylsilane (δ 0.00 ppm), which was added as an internal standard in DMSO-d6 at a
concentration of 0.03% v/v. DMSO-d6 solvent signal was referenced at δ 39.50 ppm for 13C
chemical shifts.
2.9. Density functional theory (DFT) calculations
The quantum chemical calculations were carried out on BPT and DPs 3-4 using GAUSSIAN09
suite of programs and by following the same procedural details as mentioned in our previous
publication [13].
2.10. In silico prediction of ADMET properties of BPT and its degradation products (DPs
1-5)
Physicochemical as well as absorption, distribution, metabolism, excretion and toxicity
(ADMET) of BPT and its five DPs were predicted using ADMET Predictor™ software (version
9.5.0.16, Simulation Plus, Lancaster, CA, USA).

3. Results and discussion
3.1. HPLC analysis of stressed samples and degradation behaviour of BPT
Evaluation of HPLC chromatographic behaviour of all the stressed samples (Fig. S1) indicated
that BPT degraded only under basic photolytic condition (on exposure of drug solution in 0.1N
NaOH to UV/visible light for 5 d), leading to formation of five DPs (Fig. 1). The drug
decomposed to an extent of 23.93%, forming 2.64% DP 1; 4.67% DP 2; 1.93% DP 3; 1.83% DP
4, and 12.86% DP 5.
3.2. Mass studies on BPT
The mass fragmentation pathway of BPT was established by systematic pooling of HRMS and
MSn experimental data, by following the previously established protocol from our laboratory
[16]. The HRMS line spectrum of the BPT is shown in Fig. S2. The MSn data were generated up
to MS6, and the same are listed in Table S4. Elemental composition calculator provided
molecular formulae, error (mmu), ring plus double bonds (RDB) and losses for each fragment,
and the resultant values are listed in Table 1. The postulated mass fragmentation pathway of BPT
is given in Fig. 2.
As shown in Table S4 and Fig. 2, protonated BPT (m/z 389) fragmented in MS2 step into five
product ions of m/z 371, 220, 218, 188 and 186 on the loss of H2O, C9H15NO2, C9H17NO2,
C12H8NCl and C12H10NCl, respectively. In MS3 step, the product ions of m/z 371 and 220 lost
C9H15NO2 and H2O, respectively, resulting in a common ion of m/z 202. Additionally, the
precursor of m/z 371, on loss of C12H10NOCl, generated a fragment of m/z 152. The common
product ion of m/z 202 dissociated further in MS4 step into ion of m/z 167 upon loss of chlorine
radical. The product ion of m/z 218 had low intensity, hence its further dissociation was not
observed. The precursor of m/z 188 resulted into fragments of m/z 170 (loss of H2O), 102 (loss
of C4H6O2) and 100 (loss of C4H8O2). Of these, the ion of m/z 102 further fragmented into
product ion of m/z 84 upon neutral loss of water, while no further dissociation was observed for
the ion of m/z 100. The precursor of m/z 170 further fragmented into multiple ions, viz., m/z 152,
142, 116 and 87 upon loss of H2O, C2H4, C4H6, and C5H9N, respectively. The ion of m/z 152,
which was also formed from precursor ion of m/z 371, lost H2O, CO, C4H6O and C5H7N to yield
fragments of m/z 134, 124, 82 and 71, respectively, of which only the ion of m/z 124 dissociated
further into m/z 109 and 96 upon loss of methyl radical and C2H4, respectively. The product ion
of m/z 142, which was formed from precursor of m/z 170, underwent fragmentation into ion of

m/z 98 upon loss of CO2, while the product ion of m/z 116, which was generated from the same
precursor ion, dissociated further into ions of m/z 98 and 86 upon loss of H2O and CH4N radical,
respectively.
The MS2 precursor ion of m/z 186 dissociated into product ions of m/z 168, 142 and 116 upon
loss of H2O, CO2 and C4H6O, respectively. The ion of m/z 168 further dissociated into two
fragments of m/z 150 (loss of H2O) and 140 (loss of CO), while another fragment of m/z 142
was converted to ion of 112 upon neutral loss of C2H6. The product ion of m/z 150, which was
formed from the precursor of m/z 168, lost CO to result in ion of m/z 122.
3.3. NMR studies on BPT
1D (1H, 13C and DEPT-135) and 2D NMR (COSY, TOCSY, HSQC and HMBC) spectra of BPT
(free base) are shown in Figs. S3-S9. The proton and carbon chemical shift values, along with
their multiplicities and proton-proton/proton-carbon connectivities, are compiled in Tables 2 and
S5. The proton NMR signals of BPT (Table 2 and Fig. S3) included: i) four aromatic methines of
pyridine ring observed at δ 7.19 ppm, δ 7.73 ppm, δ 7.51 ppm and δ 8.40 ppm corresponding to
H-1, H-2, H-3 and H-6, respectively; ii) two doublets of positions H-9/13 and H-10/12 at δ 7.35
ppm and δ 7.30 ppm, respectively; iii) two aliphatic methines of H-7 and H-14 at δ 5.61 ppm and
δ 3.45 ppm, respectively; iv) four non-equivalent methylene protons (H-15, H-16, H-18 and H-
19) of piperidine ring that were also confirmed by 2D experiments (COSY (Fig. S4) and HSQC
(Fig. S8); v) protons of H-15a,19a and H-15b,19b between δ 1.62-1.68 ppm and δ 1.83-1.90
ppm, respectively; vi) non-equivalent H-16a,18a and H-16b,18b protons observed between δ
2.46-2.52 ppm and δ 2.86-2.91 ppm, respectively, and vii) three methylenes of H-20, H-21 and
H-22 of butanoic acid at δ 2.55 ppm, δ 1.64 ppm and δ 2.12 ppm, respectively. All proton
positions and their connectivities were further confirmed by COSY (Fig. S4) and TOCSY (Fig.
S5) spectra.
13C NMR spectrum (Fig. S6) of BPT showed 17 signals relating to 21 carbons present in the
structure. All methine and methylene carbons were distinguished with the help of DEPT-135
(Fig. S7), while carbon-proton connectivities were verified by 2D HSQC (Fig. S8) and HMBC
studies (Fig. S9). The carbon spectrum of BPT comprised of: i) two signals for aromatic methine
carbons C-9/13 and C-10/12 at δ 129.22 ppm and δ 128.76 ppm, respectively; iii) two quaternary
carbons C-8 and C-11 of the same aromatic ring at δ 141.17 ppm and δ 132.57 ppm,
respectively; iv) five signals (C-1 to C4, and C-6) of pyridine ring at δ 123.21 ppm, δ 137.70

ppm, δ 120.91 ppm, δ 161.71 ppm and δ 149.32 ppm, respectively; v) two carbon signals of
methylenes of piperidine ring at δ 29.83 ppm and δ 49.72 ppm, corresponding to the positions C-
15/19 and C-16/18, respectively; vi) four signals of C-20 to C-23 of butanoic acid moiety at δ
56.86 ppm, δ 21.38 ppm, δ 34.06 ppm and δ 176.28 ppm, respectively, and vii) two aliphatic
methine carbon signals for C-7 and C-14 at δ 80.47 ppm and δ 71.37 ppm, respectively.
3.4. Characterization of degradation products (DPs 1-5) based on LC-HRMS, LC-MSn and
NMR data
Structural characterization of the DPs was done systematically using LC-HRMS, LC-MSn and/or
NMR studies. LC-HRMS line spectra of all five DPs are shown in Figs. S10-S14. The accurate
and exact mass values of each DP, along with their molecular formulae, errors (mmu) and RDB
values, are listed in Table 3. The fragmentation pathway of each DP was established through LC-
HRMS and LC-MSn data, as shown in Tables 3 and S6, respectively. HDX mass studies were
performed for BPT, DP 3 and DP 4 using D2O, and their comparative mass line spectra are
shown in Fig. S15. The isotopic abundance [M+H+2]+ pattern due to chlorine was observed in
case of BPT and DPs 2-5, while it was absent for DP 1 (Fig. S16). The amalgamated mass and
NMR data were used to elucidate the structure of each DP, as discussed below.
3.4.1. DP 1
This DP could not be isolated in pure form as it resolved very close to besilate salt peak.
Therefore, its structure characterization was done based on mass data. Its protonated form
([M+H]+) had an accurate m/z of 188.1280 (Fig. S10), which was 201 Da lower than BPT.
+
Elemental composition calculator suggested its molecular formula as C9H18NO3 . The error
between exact and accurate masses was found to be -0.1 mmu. Further, its exact mass was
similar to the fragment of m/z 188 formed in case of BPT (Fig. 2). Also, the fragment ions
generated from m/z 188 for both BPT and DP 1 matched with each other, including absence of
the chlorine atom, which was additionally confirmed by isotopic abundance [M+H+2]+ analyses
of BPT and all its five DPs (Fig. S16). Based on these observations, the product was
characterized as 4-(4-hydroxypiperidin-1-yl)butanoic acid, and it is proposed to be formed upon
cleavage of ether bond of BPT.
3.4.2. DP 2

DP 2 had an experimental mass ([M+H]+) of m/z 303.1252 (Fig. S11). The difference of 86 Da
between BPT and DP 2 and corresponding molecular formula for the loss, i.e., C4H6O2,
highlighted that butyric acid moiety was lost from BPT by N-dealkylation. Hence DP 2 was
elucidated to be 2-((4-chlorophenyl)(piperidin-4-yloxy)methyl)pyridine. Its structure was even
justified by the fragmentation pathway, which is included in Fig. 2. DP 2 essentially followed the
same dissociation sequence m/z 303→220→202→167 as for BPT. The [M+H+2]+ isotopic peak
appeared at m/z 305 and was 32.65% of the DP, which highlighted the presence of chlorine atom
in its structure (Fig. S16). The same was also confirmed through loss of chlorine on conversion
of precursor ion of m/z 202 to product ion of m/z 167 (Table S6, and Fig. 2). Additionally, the
1
structure of DP 2 was supported by NMR results. The H NMR spectrum of DP 2 (Fig. S17)
showed the presence of all the protons of BPT, except for H-20 to H-22 for the butyric acid
moiety. Additionally, the chemical shift values of H-16a,b and H-18a,b were deshielded as
compared to BPT (Table 2), which could be attributed to the altered neighbouring environment.
The absence of butyric acid moiety was also confirmed by changes in 13C, DEPT-135 NMR,
COSY, TOCSY, HSQC and HMBC spectra (Figs. S18-S23 and Table S7), as compared to those
for BPT (Figs. S4-S9).
3.4.3. DP 3
The accurate m/z of protonated DP 3 ([M+H]+) was 405.1569 Da (Fig. S12), which was 16 Da
+
higher than BPT. Elemental composition calculator suggested C21H26ClN2O4 as its most
probable molecular formula with exact mass of 405.1576 Da, highlighting addition of oxygen to
+
BPT (C21H26ClN2O3 ). The RDB value of DP 3 was 9.5, similar to BPT, suggesting that an
oxygen addition did not generate any new double bond. Additionally, HDX studies indicated
same two labile hydrogens in the DP as to BPT (Fig. S15), suggesting that added oxygen did not
exist in hydroxyl form, but there was rearrangement in some manner in the molecule.
Considering unusual nature of reaction involved in the formation of DP 3, the pure product was
subjected to 1D (1H, 13C and DEPT-135) and 2D (COSY, TOCSY, HSQC and HMBC) NMR
studies. The spectra are shown in Figs. S24-S30 and the data are compiled in Table S8. The
salient NMR results of DP 3, as compared to BPT, are discussed below:
1. 1H, 13C and DEPT-135 spectra (Figs. S24-S26) revealed that number and type of
protons/carbons in DP 3 were same as BPT (see Table S8). The proton and carbon δ

values for the positions 1 to 13 were also similar to BPT. The major changes in the proton
spectra were with respect to positions 16, 18 and 20. The proton pair H-16a,18a was
deshielded from δ 2.46-2.52 ppm (BPT) to δ 3.30-3.35 ppm (DP 3); the pair H-16b,18b
was deshielded from δ 2.86-2.91 ppm to δ 3.43-3.46 ppm, and H-20 was deshielded from
δ 2.55 ppm to δ 3.45 ppm. The deshielding was even observed for C-16, C-18 and C-20
(Fig. S25). This highlighted insertion of oxygen atom in-between the positions 17 and 20.
2. COSY, TOCSY and HSQC spectra (Figs. S27-S29) showed the same correlations as
BPT, supporting the fact that no new correlations were generated on addition of oxygen
at the said positions.
3. A differentiating but still supporting observation was seen in HMBC spectrum of DP 3
(Fig. S30), where positions C-15 and C-19 did not give correlation with H-20, whereas
the same existed in HMBC spectrum of BPT (Fig. S9).
The data highlighted insertion of oxygen between nitrogen (N17) of piperidine ring and carbon
(C20) of butyric acid moiety in structure of BPT to generate DP 3. This structure was thence
sought to be justified through mass fragmentation pathway, which is proposed in Fig. 3.
As shown in Fig. 3, on mass fragmentation, DP 3 gave characteristic product ions of m/z 319,
204 and 186 upon neutral loss of C4H6O2, C12H8ClN and C12H10ClNO, respectively. The
precursor of m/z 319 was further reduced to fragments of m/z 301 (loss of H2O), 220 (loss of
C5H9NO) and 218 (loss of C5H11NO). The intensity of product ion of m/z 301 was too less for
further fragmentation on MSn, but based on feasibility, it could be hypothesized that it converted
into m/z 220 upon loss of C5H7N (dotted arrow on left side in Fig. 3). The fragments of m/z 220
and 218 were found similar to those of BPT (Fig. 2). The m/z 220 fragment further dissociated
sequentially into ions of m/z 202 and 167, which were also seen in BPT fragmentation pathway.
Like fragment of m/z 301, the intensity of m/z 204 was also too low, and it is purported to
convert into m/z 186 upon loss of water (shown by dotted arrow). The latter ion of m/z 186 was
observed to further dissociate into product ion of m/z 168 (-H2O), which also fragmented
additionally into ions of m/z 150 (loss of H2O) and 140 (loss of CO). The product ion of m/z 150
was yet further reduced into m/z 122 upon loss of CO.
The presence of parent fragment of m/z 319 and its dissociation to product ion of m/z 301 upon
loss of water strongly supported the addition of oxygen between nitrogen of piperidine ring and

butyric acid moiety during formation of DP 3. The two other characteristic fragments of m/z 168
and 140 also supported this change.
Hence, mass and NMR data collectively reinforced the proposed structure of DP 3, which can be
named
chemically
as
(S)-4-(4-((4-chlorophenyl)(pyridin-2-yl)methoxy)piperidin-1-
yloxy)butanoic acid.
3.4.4. DP 4
Similar to DP 3, the experimental accurate m/z of 405.1568 of protonated ([M+H]+) DP 4 (Fig.
S13) was 16 Da higher than BPT. Therefore, the best possible molecular formula and exact mass
+
were also same as DP 3, viz., C21H26ClN2O4 and 405.1576 Da, respectively. Nitrogen rule
indicated the presence of same number of nitrogens in DP 4, as BPT and DP 3, and even same
RDB values. These observations, along with similarity in results of HDX study (Fig. S15),
highlighted that the added oxygen even in this case, like DP 3, was not in either keto or hydroxyl
form. However, as the mass fragmentation behaviour was different than BPT and DP 3,
therefore, plausibility was considered of addition of an oxygen atom at a different location than
in DP 3 and perhaps in the form of N-oxide at any one of the two tertiary nitrogens of BPT. To
delineate the exact location, the mass fragmentation profile of DP 4 was followed critically,
which provided support to the formation of N-oxide at piperidine nitrogen in BPT molecule.
Figure 4 shows that in the first step of mass dissociation, DP 4 lost H2O to yield a fragment of
m/z 387, which further formed product ions of m/z 315, 286, 220, 218, 186, and 184 upon
release of C3H4O2, C4H7NO2, C9H13NO3, C9H15NO3, C11H8ClN and C11H10ClN, respectively.
The fragment ion of m/z 315 sequentially dissociated into ions of m/z 202 and 167 upon the loss
of C6H11NO and chlorine radical, respectively. The ion of m/z 202 was additionally generated
from two different precursors of m/z 286 and 220 with associated loss of C5H8O and H2O,
respectively. The product ion of m/z 186 was further reduced by following the sequence: m/z
168 (loss of H2O)→m/z 150 (loss of H2O)→m/z122 (loss of CO). No further fragmentation of
the ions of m/z 218 and 184 was observed due to their low ion intensity. The first-step generation
of the fragment of m/z 387, highlighted that the N-oxide was formed at the piperidine ring, as
otherwise the fragment of m/z 389 would have resulted from N-oxidation of the pyridine
nitrogen in BPT. Even the observed fragments of m/z 186 and 184, which were 2 Da lower than
the respective BPT product ions of m/z 188 and 186, gave credence to supposition of N-oxide
formation at the piperidine nitrogen.

The N-oxide formation and its location were also confirmed through 1D and 2D NMR studies.
1
13
Similar to DP 3, H and C NMR spectra (Figs. S31 and S32) of DP 4 showed the presence of
same number of protons and carbons, as in BPT. However, DEPT-135 spectrum (Fig. S33)
indicated absence of any changes in the types of carbons, viz., CH and CH2. A few salient
observations with respect to the changes in chemical shift values of protons and carbons in DP 4
(Table 2) are as follows:
1. Deshielding of protons H-16a,18a (δ 3.29-3.34 ppm) and H-16b,18b (δ 3.39-3.42 ppm)
against their appearance at δ 2.46-2.52 ppm and δ 2.86-2.91 ppm, respectively, in case of
BPT.
2. Deshielding of H-20 proton (δ 3.31 ppm) compared to BPT (δ 2.55 ppm). However, the
change in chemical shift value (δ 3.31 ppm) for this proton was lesser than in DP 3 (δ
3.45 ppm). This observation is aligned to the above-discussed oxidation of tertiary
nitrogen of the piperidine ring.
3. The deshielding also happened in case of carbons C-16/18 and C-20 of DP 4, where
peaks appeared at δ 61.34 ppm and δ 67.62 ppm, respectively, against δ 49.73 ppm and δ
56.86 ppm in case of BPT (see Table 2). The corresponding value for C-20 of DP 3 was δ
68.59 ppm, higher even than DP 4, highlighting differential structure of the two.
4. Homo-nuclear (COSY and TOCSY) and hetero-nuclear (HSQC and HMBC) NMR
studies on DP 4 showed similar correlations, as observed in case of BPT (Figs. S34-S37),
unlike HMBC changes in DP 3.
Hence the proposed structure of DP 4 (Fig. 4 and Tables 2 and S9) was well justified to be (S)-1-
(3-carboxypropyl)-4-((4-chlorophenyl)(pyridin-2-yl)methoxy)piperidine 1-oxide.
3.4.5. DP 5
The HRMS line spectrum of protonated DP 5 ([M+H]+) is shown in Fig. S14. Its experimental
accurate mass was 220.0519 Da, with the exact mass value being 220.0524 Da. These were
similar to a mass fragment of BPT, with even the same subsequent fragmention. The structure
was hence postulated to be 4-chlorophenyl)(pyridin-2-yl)methanol, which was also confirmed
through 1D and 2D NMR data (Table S10). Its 1H NMR spectrum (Fig. S38) showed the absence
of aliphatic protons H-14 to H-16 and H-18 to H-22, which indicated absence of the 4-(piperidin-
1-yl)butanoic acid moiety. Even 13C NMR and DEPT-135 NMR spectra (Figs. S39 and S40)
highlighted the absence of one methine (C-14) and seven methylene (C-15, C-16, and C-18 to C-

22) carbons. Moreover, carbonyl carbon C-23 was also found to be absent in 13C NMR spectrum.
Additionally, 2D (COSY, TOCSY, HSQC and HMBC) NMR spectra (Figs. S41-S44) supported
the proposed structure.
3.5. Consideration of 1,2-Meisenheimer type rearrangement in generation of DP 3
As discussed above, the structure of DP 3, which was characterized based on mass, HDX and
NMR data, was established to have an oxygen incorporated between piperidine nitrogen (N17)
and butyric acid (C20). This structure was unusual as for as the possibility of its formation
directly from BPT under the studied stress condition was concerned. Thereupon, it was
considered that DP 3 was perhaps generated under the reaction conditions from DP 4, through
1,2-Meisenheimer type rearrangement. A critical review of the literature for 1,2-Meisenheimer
rearrangement [17, 18] brought another possible structural variant for DP 3 (Fig. S45), the one
with oxygen embedded in the piperidine ring between N17 and C18. The consideration of the
latter was based on pathway for rearrangement of certain piperidine N-oxides to hexahydro-1,2-
oxazepines, as reported by Quin et al. [19]. Similar rearrangement has also been reported by
different researchers in the presence of light [20, 21]. Many other literature reports discussed
intra-molecular rearrangements of tertiary amine oxides, wherein the reaction schemes indicated
the conversion of oxides to hydroxylamines [22, 23]. To explore as to why in DP 4, the
rearrangement happened to incorporate oxygen between N17 and C20 and not N17 and C18, we
resorted to seek answer from the study of energetics of these two variants by employing quantum
chemical DFT study.
Initially, calculations were performed to identify the thermodynamic stable conformer of BPT
among many possible by rotation at the stereo-centre. The simplest two, viz., BPT I (axial) and
BPT II (equatorial) were subjected to investigation. Their energetically optimized 3D structures
are given in Fig. S46. BPT II (equatorial) was found to be more stable by 4.01 kcal/mol as
compared to BPT I conformer. Similarly, conformers of DP 4 were also optimized and it was
found that equatorial geometry structure was more stable, as shown from structures included in
Fig. S46. Subsequently, DFT studies were focused on two possible variants that could form via
1,2-Meisenheimer rearrangement or its type, as discussed above. Their potential energy surface
diagrams (PES) are displayed in Fig. 5. The energy value comparison revealed that formation of
both variants was thermodynamically favorable, while the variant with oxygen added between
N17 and C18 (i.e. N17-O-C18) was exergonic by 6.83 kcal/mol, the other (N17-O-C20) was

exergonic by 14.71 kcal/mol. The activation barrier for the formation of these two variants was
62.30 kcal/mol and 55.96 kcal/mol, respectively. The activation energy of 55.96 kcal/mol for
variant (N17-O-C20) was close to the reported value of 56.2 kcal/mole for 1,2-Meisenheimer
rearrangement in N,N-dimethylethanamine oxide by Zhang [24]. As the value was very high for
any acceptable chemical reaction pathway, the author had proposed that an alternative path via
radical mechanism instead of concerted sigmatropic rearrangement was favorable. Accordingly,
the energy required for N17-C20 bond breaking in DP 4 was calculated and the same was
estimated to be 15.98 kcal/mol. The O-C20 bond formation energy was -29.82 kcal/mol. Both
were found to be much lower than activation energy of 55.96 kcal/mol for the rearrangement.
Hence, we suggest radical pathway to be involved in the observed rearrangement and generation
of DP 3.
3.6. Postulated mechanisms for degradation of BPT
The mechanistic pathways involved in the formation of photo DPs 1-5 are shown in Figs. 6-8.
The same are discussed below.
3.6.1. DP 1 and DP 5
The pathway involved in photodegradation of BPT besilate to DPs 1 and 5 is shown in Fig. 6. It
is proposed that light energy initially caused homolytic cleavage at C7-O bond of the drug to
result in benzylic (I) and alkoxy (II) radicals. The latter, through water-assisted intra-molecular
1,2-hydrogen shift, is converted into another isomeric carbon centered radical (III), which yields
DP 1 on interaction with free hydrogen radical available in the process of simultaneous
formation of DP 2 from BPT (Fig. 7, Pathway B). This supposition is well supported by the fact
that water offers specific assistance through cooperative hydrogen bond formation to promote
various organic reactions in the absence [25-27] and presence [28, 29] of a metal-based catalyst.
The water is also known to assist in radical molecule interaction [30].
In view of these and a few literature reports on water-assisted rearrangement of radicals [31-34],
we further planned to explore the specific assistance provided by water by resorting to DFT
studies. The pathways followed in the absence (A) and the presence (B) of water are shown in
Fig. 6. The free energy difference of 3.97 kcal/mol revealed that III was rather more favourable
than II. However, in the absence of water, the activation energy required in the process of
conversion was 30.90 kcal/mol, which was seemingly higher for the reaction conditions

employed. As photodegradation study was performed in aqueous BPT solution, DFT study was
repeated with sequential addition of water molecules. The transition state for II→III conversion
was obtained in the presence of two water molecules, and the energy barrier got reduced to 21.89
kcal/mol (Pathway B, Fig. 6). This supported the existence of carbon-centered radical III in the
aqueous condition of the reaction and upon exposure to light.
It is proposed that oxidation of benzylic radical (I) to an unstable peroxide species (IV) by
involvement of Russell termination mechanism results in DP 5 and an intermediate keto product
(V) [35]. The latter, owing to its photo-reactivity, also reduces to DP 5 in the alkaline
environment of the reaction [36-38].
3.6.2. DP 2
As shown in Fig. 7, BPT can also be converted under light to unstable iminium intermediate by
two different pathways (A and B) involving electron removal, followed by hydrogen abstraction
and vice versa [39, 40]. The base hydrolysis of the common intermediate further yields DP 2.
3.6.3. DP 3 and DP 4
In case of these DPs, tertiary nitrogen (piperidine ring) of BPT shall convert into aminium
radical cation in the presence of light, which subsequently would react with molecular oxygen to
first yield DP 4 [41], which can then undergo 1,2-Miesenheimer type rearrangement via radical
pathway to form DP 3 (Fig. 8). The absence of concerted sigmatropic rearrangement in this case
was justified by the DFT study, as discussed in Section 3.5.
3.7. Study in the presence of TEMPO
The involvement of radicals in photo base degradation of BPT to DPs 1-5 was justified by results
of a quencher study in 0.1 N NaOH carried out for a total period of 8 d. Figure 9 shows that in
the presence of TEMPO, degradation of BPT was inhibited significantly.
3.8. Prediction of physicochemical and ADMET properties of degradation products and
comparison with those of BPT
The predicted physicochemical and ADME properties of BPT and its five DPs are summarized
in Table S11. Their toxicity profiles are separately listed in Table 4.
The characterized DPs 1-5 had higher predicted solubility than BPT. Also, DPs 1 and 4 had
relatively high predicted solubility than DPs 2, 3 and 5. The lipophilicity of the latter three DPs

was higher than BPT and DPs 1 and 4. The values of volume of distribution in humans at steady
state (Vd), permeability through rabbit cornea (Perm_Cornea) and qualitative likelihood
(high/low) of crossing the blood-brain barrier (BBB_Filter) for DPs 2 and 5 were found to be
higher than BPT and rest of the DPs. BPT and DPs 2, 3 and 4 were predicted as substrates for P-
glycoprotein (P-gp), an observation also reported by Ohashi et al. [42]. Neither BPT nor the DPs
were predicted as inhibitors of P-gp. Rather, DPs 3 and 4 were predicted as inhibitors of
OATP1B. The substrate and inhibitor properties for different CYPs varied among BPT and its
DPs, as highlighted in Table S11.
As shown from the data in Table 4, DP 2 was predicted to be a skin sensitizer, while DP 5
showed susceptibility towards skin and respiratory sensitization. pIC 50 value for DP 2 was
higher than 5.5 mol/L, which indicated potential for cardiac toxicity, as compared to BPT and
other DPs. BPT and DPs 1 and 3-5 were predicted to be negative for phospholipidosis, while DP
2 was susceptible. BPT and DP 5 were highlighted to be associated with reproductive toxicity.
DP 5 showed potential for producing chromosomal aberrations, with possibility of genotoxicity.
DPs 1 and 3 showed positive results for mutagenicity, while other DPs, including BPT, were
found to be non-mutagenic in ten in silico models.
In addition to the above, the ecotoxicity of BPT and its DPs was also predicted through the same
software keeping in view the environmental concern. Bioconcentration factor (BCF) was
estimated higher for DPs 2 and 5, as both were envisaged to be non-biodegrading. Only DP 1
was postulated to be biodegradable. BPT, DP 2 and DP 3 showed major susceptibility for fathead
minnow (Pimephales promelas) toxicity. The computed Daphnia_LC50 values of drug and its
DPs were higher than 0.22 mg/L, highlighting less probability of being toxic to Daphnia magna.
4. Conclusion
This study explored degradation behaviour of BPT under various stress conditions. BPT showed
photolability in basic medium, resulting in previously unreported five DPs. The same were
characterized by using LC-HRMS, LC-MSn and NMR tools. Additionally, in silico
physicochemical and ADMET profile was assessed using ADMET PredictorTM software, in
which two DPs were predicted to be associated with mutagenic risk. The finding that BPT
decomposes through a photodegradation pathway in basic conditions may be important in
understanding its fate in aquatic environment on exposure to sunlight in alkaline sludge, or
perhaps upon exposure to UV during treatment of waste water [43, 44].

Acknowledgement
The authors gratefully acknowledge Simulation Plus (Lancaster CA, USA) for providing license
assess to ADMET Predictor™ (version 9.5.0.16). We are thankful to A. Y. Deshpande, Aurigene
Discovery Technologies Limited for providing incessant use of the HRMS facility.
References
1. Singh S, Junwal M, Modhe G, Tiwari H, Kurmi, M, Parashar N, Sidduri P. Forced
Degradation Studies to Assess the Stability of Drugs and Products. TrAC Trends in Anal
Chem 2013; 49:71-88.
2. Alsante KM, Ando A, Brown R, Ensing J, Hatajik TD, Kong W, Tsuda Y. The Role of
Degradant Profiling in Active Pharmaceutical Ingredients and Drug Products. Adv Drug
Deliv Rev 2007; 59(1):29-37.
3. ICH, Q1A(R2) Stability Testing of New Drug Substances and Products, Geneva,
Switzerland (2003).
4. ICH, Q1B Stability Testing: Photostability Testing of New Drug Substances and
Products, Geneva, Switzerland (1996).
5. Available at:
https://www.accessdata.fda.gov/drugsatfda_docs/nda/2009/022288s000_PharmR.pdf.
Accessed August 25, 2019.
6. Available at:
https://www.accessdata.fda.gov/drugsatfda_docs/nda/2009/022288s000_ClinPharm.pdf.
Accessed August 25, 2019.
7. Lyseng-Williamson KA. Oral Bepotastine. Drugs 2010; 70(12):1579-1591.
8. Phadke RK, Gaitonde VD. A Novel Validated Method for Determination of Genotoxic
Impurites in Bepotastine Besilate (API) with Reserve Phase HPLC Technics. World J
Pharm Res 2016; 5(5):808-821.
9. Damle MC, Ghode SP. Stability Indicating Chromatographic Method for Estimation of
Bepotastine Besilate. Indo Am J Pharm 2018; 5(6):5222-5228.
10. Kotla NR, Nagaraju CVS, Rajan ST, Eshwaraiah S, Kumar Reddy G, Rakesh M,
Kishore M, Chakravarthy IE. Stability Indicating HPLC Method for the Quantification
of Bepotastine Besilate and its Related Substances. Der Pharma Chem 2014; 6(2):343-
351.

11. Soliman MM, Darwish MK, Abdel-Razeq SAM. Validated Stability Indicating HPTLC,
UHPLC and UV-Spectrophotometric Techniques for the Determination of Bepotastine
Besilate in Presence of its Oxidative Degradate. Asian J Appl Chem Res 2019; 3(1):1-14.
12. Singh S, Bakshi M. Guidance on Conduct of Stress Test to Determine Inherent Stability
of Drugs. Pharm Technol On-Line 2000; 4:1-14.
13. Singh DK, Sahu A, Wani AA, Bharatam PV, Kotimoole CN, Batkulwar KB, Deshpande
AY, Giri S, Singh S. Stability Behaviour of Antiretroviral Drugs and their
Combinations. 10: LC-HRMS, LC-MSn, LC-NMR and NMR Characterization of
Fosamprenavir Degradation Products and In Silico Determination of Their ADMET
Properties. Eur J Pharm Biopharm 2019; 142:165-178
.
14. Gualandi A, Mazzarella D, Ortega-Martínez A, Mengozzi L, Calcinelli F, Matteucci E,
Monti F, Armaroli N, Sambri L, Cozzi PG. Photocatalytic Radical Alkylation of
Electrophilic Olefins by Benzylic and Alkylic Zinc-Sulfinates. ACS Catal 2017;
7(8):5357-5362.
15. Shan XH, Zheng HX, Yang B, Tie L, Fu JL, Qu JP, Kang YB. Copper-Catalyzed
Oxidative Benzylic C-H Cyclization via Iminyl Radical from Intermolecular Anion-
Radical Redox Relay. Nat Commun 2019; 10(1):908.
16. Narayanam M, Handa T, Sharma P, Jhajra S, Muthe PK, Dappili PK, Shah RP, Singh S.
Critical Practical Aspects in the Application of Liquid Chromatography-Mass
Spectrometric Studies for the Characterization of Impurities and Degradation Products. J
Pharm Biomed Anal 2014; 87:191-217.
17. Al-Iraqi MA, Al-Rawi JM, Khuthier AH. Studies in Tertiary Amine Oxides Part II-
Carbon-13 Nuclear Magnetic Resonance Spectra of Selected Acetylenic Amines, their
N-Oxides and the Meisenheimer Rearrangement Products. Org Magn Reson 1980;
14(3):161-165.
18. Castagnoli Jr N, Craig JC, Melikian AP, Roy SK. Amine-N-Oxide Rearrangements:
Mechanism and Products of Thermolysis. Tetrahedron 1970; 26(18):4319-4327.
19. Quin LD, Shelburne FA. Rearrangement of Some Piperidine N-Oxides to Hexahydro-1,
2-oxazepines. J Org Chem 1965; 30(9):3135-3138.

20. Poole JS. Recent Advances in the Photochemistry of Heterocyclic N-Oxides and their
Derivatives. In: Larionov O. (ed.) Heterocyclic N-Oxides. Topics in Heterocyclic
Chemistry, Vol. 53, Springer, Cham, 2017; pp. 111-151.
21. Sharp JT. Seven-Membered Rings with Two or More Heteroatoms. In: Katritzky AR,
Charles WR. (eds.) Comprehensive Heterocyclic Chemistry, Vol. 7, Elsevier, Pergamon,
1984; pp. 593-651.
22. Carlson RM, Heinis LJ. Intramolecular Rearrangements of Tertiary Amine Oxides. J
Org Chem 1979; 44(14):2530-2533.
23. Lepley AR, Cook, PM, Willard GF. Proton Magnetic Resonance Emission in the
Intramolecular Rearrangement of a Tertiary Amine Oxide. J Am Chem Soc 1970;
92(4):1101-1102.
24. Zhang X. DFT Studies on Meisenheimer Rearrangement. Comput Theor Chem 2011;
966(1-3):383-390.
25. Khatik GL, Kumar R, Chakraborti AK. Catalyst-free Conjugated Addition of Thiols to
α,β-Unsaturated Carbonyl Compounds in Water. Org Lett 2006; 8(11):2433-2436.
26. Chankeshwara SV, Chakraborti AK. Catalyst-free Chemoselective N-tert-
butyloxycarbonylation of Amines in Water. Org Lett 2006; 8(15):3259-3262.
27. Chakraborti AK, Rudrawar S, Jadhav KB, Kaur G, Chankeshwara, SV. “On Water”
Organic Synthesis: A Highly Efficient and Clean Synthesis of 2-Aryl/Heteroaryl/Styryl
Benzothiazoles and 2-Alkyl/Aryl Alkyl Benzothiazolines. Green Chem 2007; 9:1335-
1340.
28. Parikh N, Kumar D, Raha Roy S, Chakraborti AK. Surfactant Mediated Oxygen
Reuptake in Water for Green Aerobic Oxidation: Mass-Spectrometric Determination of
Discrete Intermediates to Correlate Oxygen Uptake with Oxidation Efficiency. Chem
Commun 2011; 47:1797-1799.
29. Seth K, Roy SR, Pipaliya BV, Chakraborti AK. Synergistic Dual Activation Catalysis by
Palladium Nanoparticles for Epoxide Ring Opening with Phenols. Chem Commun 2013;
49(52):5886-5888.
30. Vöhringer-Martinez E, Hansmann B, Hernandez H, Francisco JS, Troe J, Abel1 B.
Water Catalysis of a Radical-Molecule Gas-Phase Reaction. Science 2007; 315:497-501.

31. Van Hoomissen DJ, Vyas S. 1,2-H Atom Rearrangements in Benzyloxyl Radicals. J
Phys Chem A 2018; 123(2):492-504.
32. Fernández-Ramos A, Zgierski MZ. Theoretical Study of the Rate Constants and Kinetic
Isotope Effects of the 1,2-hydrogen-atom Shift of Methoxyl and Benzyloxyl Radicals
Assisted by Water. J Phys Chem A 2002; 106(44):10578-10583.
33. Teixeira-Dias JJ, Furlani TR, Shores KS, Garvey JF. Transition States for H Atom
Transfer Reactions in the CH2CH2OH Radical: The Effect of a Water Molecule. Phys.
Chem. Chem. Phys. 2003; 5(22):5063-5069.
34. Dhameliya TM, Chourasiya SS, Mishra E, Jadhavar PS, Bharatam PV, Chakraborti AK.
Rationalisation of Benzazole-2-carboxylate vs Benzazine-3-one/Benzazine-2,3-dione
Selectivity Switch during Cyclocondensation of 2-Amino thiophenols/phenols/anilines
with 1,2-Biselectrophiles in Aqueous Medium. J Org Chem 2017; 82(19):10077-10091.
35. Russell GA. Deuterium-Isotope Effects in the Autoxidation of Aralkyl Hydrocarbons.
Mechanism of the Interaction of Peroxy Radicals. J Am Chem Soc 1957; 79(14):3871-
3877.
36. Fasani E, Albini A, 2011 Photostability Stress Testing. In: Baertschi SW, Alsante KM,
Reed RA (Eds.) Pharmaceutical Stress Testing, CRC Press, p. 195.
37. Bachmann, WE. The Photochemical Reduction of Ketones to Hydrols. J Am Chem Soc
1933; 55(1):391-395.
38. Cohen SG, Sherman WV. Inhibition and Quenching of the Light-induced Reductions of
Benzophenone to Benzpinacol and to Benzhydrol. J Am Chem Soc 1963; 85(11):1642-
1647.
39. Beatty JW, Stephenson CR. Amine Functionalization via Oxidative Photoredox
Catalysis: Methodology Development and Complex Molecule Synthesis. Acc Chem Res
2015; 48(5):1474-1484.
40. Condie AG, González-Gómez JC, Stephenson CR. Visible-light photoredox catalysis:
Aza-Henry reactions via C−H functionalization. J Am Chem Soc 2010; 132(5):1464-
1465.
41. Baertschi SW, Alsante KM, Santafianos D. 2011 Stress Testing: The Chemistry of Drug
Degradation. In: Baertschi SW, Alsante KM, Reed RA (Eds.) Pharmaceutical Stress
Testing, CRC Press, pp. 49-154.

42. Ohashi R, Kamikozawa Y, Sugiura M, Fukuda H, Yabuuchi H, Tamai I. Effect of P-
glycoprotein on Intestinal Absorption and Brain Penetration of Antiallergic Agent
Bepotastine Besilate. Drug Metab Dispos 2006; 34(5):793-799.
43. Fatta-Kassinos D, Vasquez MI, Kümmerer K. Transformation Products of
Pharmaceuticals in Surface Waters and Wastewater Formed During Photolysis and
Advanced Oxidation Processes-Degradation, Elucidation of Byproducts and Assessment
of their Biological Potency. Chemosphere 2011; 85(5):693-709.
44. Kim I, Yamashita N, Tanaka H. Photodegradation of Pharmaceuticals and Personal Care
Products During UV and UV/H2O2 Treatments. Chemosphere 2009; 77(4):518-525.

Figure captions
Fig. 1. HPLC chromatograms showing separation of degradation products (DPs 1-5) of BPT
besilate formed under photo basic stress condition and separation of besylic acid. Key: d, day(s).
Fig. 2. Mass fragmentation pathways of BPT, DP 1, DP 2 and DP 5. Structures of DPs are shown
in solid boxes. The fragments of m/z 371, 220, 218, 152, 150, 140, 134, 122, 109, 112, 102, 100,
96, 87, 86, 84, 82 and 71 were observed only during MSn studies. Exact mass and molecular
formula (in square brackets) are given below each structure.
Fig. 3. Mass fragmentation pathway of DP 3. See Fig. 2 for the structures of fragment ions not
shown in this figure. Exact mass and molecular formula (in square brackets) are given below
each structure. The conversion of fragments of m/z 301 to m/z 220 and m/z 204 to m/z 186 is
postulated on logical consideration.
Fig. 4. Mass fragmentation pathway of DP 4. The structures of fragment ions not shown here are
given in Fig. 2. Exact mass and molecular formula (in square brackets) are given below each
structure.
Fig. 5. Potential energy surface for the two possible variants of DP 3.
Fig. 6. Photodegradation pathway of BPT besilate yielding DP 1 and DP 5. The structure of salt
moiety is not shown in the figure. DFT results of 1,2-hydrogen shift without (Pathway A) and
with (Pathway B) mediation of water are shown in dashed boxes. Also, the possibility of
hydrogen abstraction on conversion of BPT to BPT radical during formation of DP 1 from (III) is
explained by Pathway B shown in Figure 7. Square brackets enclose cited references.
Fig. 7. Photodegradation pathway of BPT besilate resulting in DP 2. The structure of salt moiety
is not shown in the figure.
Fig. 8. Photodegradation pathway of BPT besilate forming DP 3 and DP 4. The structure of salt
moiety is not shown in the figure.
Fig. 9. Photo base degradation behaviour of BPT besilate in 0.1 N NaOH without and with added
2.5 mM TEMPO (radical quencher).

Table 1. Interpretation of mass fragmentation data obtained for bepotastine.
Exact
Mass
difference
from
Accurate/nominal Best possible
mass of
the most
probable
structure
Possible
parent
fragment
Molecular
formula for
the losses
Error
(mmu)
Peak No
mass of the
molecular
formula
RDB
observed ion*
parent ion
+
[M+H]+
389.1617
371
220
218
202.0417
C21H26ClN2O3 389.1626
-0.9
-
-
9.5
10.5
7.5
8.5
8.5
-
-
-
+
a
b
c
d
C21H24ClN2O2 371.1521
[M+H]+
18.0105
H2O
C12H11ClNO+ 220.0524
C12H9ClNO+ 218.0367
[M+H]+ 169.1102
C9H15NO2
C9H17NO2
C9H15NO2
H2O
-
[M+H]+ 171.1259
C12H9ClN+
202.0418
-0.1
a
b
169.1103
18.0106
+
e
f
g
h
i
188.1275
186.1124
170.1174
168.1007
167.0728
152
C9H18NO3
188.1281
186.1125
170.1176
168.1019
167.0730
152.1070
-0.6
-0.1
-0.2
-1.2
-0.2
-
1.5
2.5
2.5
3.5
9.0
3.5
[M+H]+ 201.0345
C12H8ClN
+
C9H16NO3
[M+H]+ 203.0501 C12H10ClN
+
C9H16NO2
e
f
18.0105
18.0106
34.9688
H2O
H2O
Cl∙
+
C9H14NO2
C12H9N+∙
d
a
g
h
f
g
h
j
j
k
f
g
o
l
e
e
j
C9H14NO+
219.0451 C12H10ClNO
18.0106
18.0106
43.9899
28.0313
27.9949
18.0106
27.9949
27.9949
70.0419
54.0470
15.0235
30.0469
86.0368
88.0524
H2O
H2O
CO2
C2H4
CO
H2O
CO
CO
C4H6O
C4H6
CH3
C2H6
C4H6O2
C4H8O2
k
l
150
142.1221
142.0862
140
134
124.1119
122
C9H12NO+
C8H16NO+
150.0913
142.1226
142.0863
140.1070
134.0964
124.1121
122.0964
116.0706
-
-0.5
-0.1
-
4.5
1.5
2.5
2.5
4.5
2.5
3.5
1.5
+
l’
m
n
o
p
q
C7H12NO2
C8H14NO+
C9H12N+
C8H14N+
C8H12N+
-
-0.2
-
0.1
+
116.0707
C5H10NO2
∙
r
s
t
109
112
102
100
C7H11N+∙
C6H10NO+
C5H12NO+
C5H10NO+
109.0886
112.0757
102.0913
100.0757
-
-
-
-
3.0
2.5
0.5
1.5
u

v
v’
w
x
98.0966
98.0603
96
C6H12N+
C5H8NO+
C6H10N+
98.0964
98.0600
96.0808
87.0441
86.0362
84.0808
82.0651
71.0491
0.2
0.3
1.5
2.5
2.5
1.5
2.0
1.5
2.5
1.5
l’
q
o
g
q
t
43.9899
18.0106
28.0313
83.0735
30.0344
18.0105
70.0419
81.0579
CO2
H2O
C2H4
C5H9N
CH4N∙
H2O
-
-
-
-
-
-
+
87
C4H7O2
+∙
y
z
aa
bb
86
84
82
71
C4H6O2
C5H10N+
C5H8N+
C4H7O+
j
j
C4H6O
C5H7N
*Ions with nominal mass were only observed in MSn studies.

Table 2. 1H and 13C NMR chemical shifts of isolated products in DMSO-d6.
6
6
6
3
6
6
3
1
2
1
2
1
2
1
2
N 5
N 5
N 5
N 5
N 5
4
1
2
19
16
19
19
19
16
4
4
4
4
O
O
O
O
7
O
OH
7
7
14
7
14
7
14
14
18
18
18
18
O
O
21
Chemical
strcuture
3
3
3
O
O
21
21
8
8
8
8
17N
17N
N 17
20
8
23
OH
15
15
15
15
NH
17
9
9
9
9
9
23
20
22
13
12
23
OH 24
13
12
13
12
24
13
12
13
12
OH 24
22
16
16
20
22
10
10
10
10
10
11
11
11
11
11
Cl
Cl
Cl
Cl
Cl
DP 5
Bepotastine (free base)
DP 2
DP 3
DP 4
1H
(δ ppm)
13C
(δ ppm)
1H
(δ ppm)
13C
(δ ppm)
1H
(δ ppm)
13C
(δ ppm)
1H
(δ ppm)
13C
(δ ppm)
1H
(δ ppm)
13C
(δ ppm)
Position
1
2
3
4
5
7.19 (t)
7.73 (t)
7.51 (d)
-
123.21
137.70
120.91
161.71
-
7.22 (t)
7.77 (t)
7.52 (d)
-
123.29
137.78
120.93
161.57
-
7.21 (t)
7.76 (t)
7.58 (d)
-
123.31
137.83
121.08
161.45
-
7.19 (br)###
7.75 (t)
7.51 (d)
-
123.34
137.83
121. 03
160.41
-
7.18 (t)
7.77 (t)
7.51 (d)
-
122.80
137.50
120.51
164.09
-
-
-
-
-
-
6
7
8
9
10
11
12
13
14
8.40 (d)
5.61 (s)
-
7.35 (d)
7.30 (d)
-
7.30 (d)
7.35 (d)
3.45 (br)
149.32
80.47
8.41 (d)
5.64 (s)
-
7.37 (d)
7.32 (d)
-
7.32 (d)
7.37 (d)
3.58 (sep)
149.40
80.45
8.40 (d)
5.61 (s)
-
7.38 (d)
7.31 (d)
-
7.31 (d)
7.38 (d)
3.64 (br)
149.32
80.70
8.40 (d)
5.61 (s)
-
7.36 (d)
7.30 (d)
-
7.30 (d)
7.36 (d)
3.55 (sep)
149.34
80.49
8.38 (d)
5.66(s)
-
7.36 (d)
7.29 (d)
-
7.29 (d)
7.36 (d)
-
148.91
75.28
143.93
128.80
128.50
131.94
128.50
128.80
-
141.17
129.22
128.76
132.57
128.76
129.22
71.37
141.07
129.26
128.81
132.59
128.81
129.26
70.31
140.92
129.33
128.80
132.65
128.80
129.33
67.31
140.94
129.24
128.81
132.65
128.81
129.24
70.75
1.62-1.68 (br),
1.83-1.90 (br)
2.46-2.52 (br)#,
2.86-2.91 (br)
-
2.46-2.52 (br)#,
2.86-2.91 (br)
1.62-1.68 (br)
1.83-1.90 (br)
2.55 (t)
1.65-1.72 (m),
1.88- 1.94 (m)
2.80-2.86 (m),
3.04-3.10 (m)
1.74-1.83 (br),
2.08-2.12 (br)
3.30-3.35 (br),
3.43-3.46 (br)
-
3.30-3.35 (br),
3.43-3.46 (br)
1.74-1.83 (br),
2.08-2.12 (br)
3.45 (br)##
1.88 (sep)
2.06 (t)##
-
1.92-1.97 (br),
2.00-2.04 (br)
3.29-3.34 (br),
3.39-3.42 (br)
-
3.29-3.34 (br),
3.39-3.42 (br)
1.92-1.97 (br),
2.00-2.04 (br)
3.31**
15a,b
29.83
28.37
25.02
26.78
-
-
16a,b
17
49.72
-
40.68
-
58.79
-
61.34
-
-
-
-
-
-
-
*
2.80-2.86 (m),
3.04-3.10 (m
1.65-1.72 (m),
1.88- 1.94 (m)
18a,b
49.72
40.68
58.79
61.34
19a,b
29.83
28.37
25.02
26.78
-
-
20
21
22
23
24
56.86
21.38
34.06
176.28
-
-
-
-
-
-
-
-
-
-
-
68.59
18.62
33.60
176.44
-
67.62
18.92
33.59
176.55
-
-
-
-
-
-
-
-
-
-
-
1.64 (br)
2.12 (t)
1.83 (sep)
2.02 (t)
-
*
-
*
*
*Not observed; #merge with solvent peak; ##merge with 16b and 18b protons; ###suppressed due to presaturation pulse sequence during acquisition; **multiplicity not clear; s
(singlet); d (doublet); t (triplet); sep (septet); br (broad), and m (multiplets).

Table 3. LC-HRMS data of degradation products (DP 1 to DP 5).
Probable
molecular
formula
(Exact mass;
error in mmu; (Isotopic
RDB;
Experimental
mass
Experimental
mass
Accurate mass of the fragment ions, best possible molecular formula
(Exact mass; error in mmu; RDB)
DPs
for [M+H+2]+
for
[M+H]+
abundance %)
nitrogen rule)
+
170.1175, C9H16NO2 (170.1176; -0.1; 2.5); 152.1070; C9H14NO+
(152.1070; 0.0; 3.5); 124.1122, C8H14N+ (124.1121; 0.1; 2.5);
+
C9H18NO3
+
190.1320
(0.52%)
116.0708, C5H10NO2 (116.0706; 0.2; 1.5); 102.0916, C5H12NO+
DP 1
DP 2
DP 3
188.1280
303.1252
405.1569
(188.1281; -0.1;
1.5; odd)
(102.0913; 0.3; 0.5); 100.0759, C5H10NO+ (100.0757; 0.2; 1.5);
98.0603, C5H8NO+ (98.0600; 0.3; 2.5); 96.0812, C6H10N+ (96.0808;
0.4; 2.5)
C17H20N2OCl+
(303.1259; -0.7; (32.65%)
8.5; even)
305.1219
220.0520, C12H11NOCl+ (220.0524; -0.4; 7.5); 202.0416, C12H9NCl+
(202.0418; -0.2; 8.5); 167.0727, C12H9N+∙ (167.0730; -0.3; 9.0)
+
319.1202, C17H20ClN2O2
(319.1208; -0.6; 8.5); 220.0522,
C12H11ClNO+ (220.0524; -0.2; 7.5); 218.0364, C12H9ClNO+ (218.0367;
+
C21H26ClN2O4
407.1534
+
-0.3; 8.5); 204.1231, C9H18NO4 (204.1230; 0.1; 1.5); 202.0417,
(405.1576; -0.7;
(31.31%)
C12H9ClN+ (202.0418; -0.1; 8.5); 186.1123, C9H16NO3 (186.1125; -
+
9.5; even)
+
0.2; 2.5); 168.1018, C9H14NO2 (168.1019; -0.1; 3.5); 167.0727,
C12H9N∙+ (167.0730; -0.3; 9.0)
+
387.1463, C21H24ClN2O3 (387.1470; -0.7; 10.5); 315.1237,
C18H20ClN2O+ (315.1259; -2.2; 9.5); 286.0984, C17H17ClNO+
(286.0993; -0.9; 9.5); 220.0515, C12H11ClNO+ (220.0524; -0.9; 7.5);
218.0363, C12H9ClNO+ (218.0367; -0.4; 8.5); 202.0416, C12H9ClN+
+
C21H26ClN2O4
407.1533
DP 4
DP 5
405.1568
220.0519
(405.1576; -0.8;
(31.04%)
+
9.5; even)
(202.0418; -0.2; 8.5); 186.1123, C9H16NO3 (186.1125; -0.2; 2.5);
+
+
184.0969, C9H14NO3 (184.0968; 0.1; 3.5); 168.1017, C9H14NO2
(168.1019; -0.2; 3.5); 167.0727, C12H9N∙+ (167.0730; -0.3; 9.0)
C12H11NOCl+
222.0484
202.0414, C12H9NCl+ (202.0418; -0.4; 8.5); 167.0725, C12H9N+∙
(167.0730; -0.5; 9.0)
(220.0524; -0.5;
(33.29%)
7.5; odd)