Synthesis and Pharmacological Evaluation of σ2 Receptor Ligands Based on a 3-Alkoxyisoxazole Scaffold: Potential Antitumor Effects against Osteosarcoma

Article abstract of DOI:10.1002/cmdc.202000461

Since its initial discovery as the basis for nicotinic acetylcholine receptor ligands, the 3-alkoxyisoxazole scaffold has been shown to be a versatile platform for the development of potent σ1 and σ2 receptor ligands. Herein we report a further SAR exploration of the 3-alkoxyisoxazole scaffold with the aim of obtaining potent σ2 receptor ligands. Various substitutions on the benzene ring and at the basic amino regions resulted in a total of 21 compounds that were tested for their binding affinities for the σ2 receptor. In particular, compound 51 [(2S)-1-(4-ammoniobutyl)-2-(((5-((3,4-dichlorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyrrolidin-1-ium chloride] was identified as one of the most potent σ2 ligands within the series, with a K_i value of 7.9 nM. It demonstrated potent antiproliferative effects on both osteosarcoma cell lines 143B and MOS?J (IC₅₀ values of 0.89 and 0.71 μM, respectively), relative to siramesine (IC₅₀ values of 1.81 and 2.01 μM). Moreover, compound 51 inhibited clonal formation of osteosarcoma 143B cells at 1 μM, corresponding to half the dose required of siramesine for similar effects. The general cytotoxicity profile of compound 51 was assessed in a number of normal cell lines, including HaCaT, HAF, and LO2 cells. Furthermore, FACS analysis showed that compound 51 likely inhibits osteosarcoma cell growth by disruption of the cell cycle and promotion of apoptosis.

Full text of DOI:10.1002/cmdc.202000461

Accepted Article
Title: Synthesis and pharmacological evaluation of σ2 receptor ligands
based on a 3-alkoxyisoxazole scaffold: potential antitumor effects
against osteosarcoma
Authors: Jun-Jie Shi, Kun-Hang Jia, Hao Sun, Hendra Gunosewoyo,
Fan Yang, Jie Tang, Jian Luo, and Lifang Yu
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemMedChem 10.1002/cmdc.202000461
Link to VoR: https://doi.org/10.1002/cmdc.202000461
01/2020

10.1002/cmdc.202000461
ChemMedChem
Synthesis and pharmacological evaluation of σ2 receptor
ligands based on a 3-alkoxyisoxazole scaffold: potential
antitumor effects against osteosarcoma
Jun-Jie Shi,† , # Kun-Hang Jia,§, # Hao Sun,† Hendra Gunosewoyo,┴ Fan Yang,† Jie Tang, Jian Luo,§
and Li-Fang Yu*, †
†Shanghai Engineering Research Center of Molecular Therapeutics and New Drug
Development, School of Chemistry and Molecular Engineering, East China Normal University,
3663 North Zhongshan Road, Shanghai 200062, China
§Shanghai Key Laboratory of Regulatory Biology, Institute of Biomedical Sciences and School
of Life Sciences East China Normal University, 500 Dongchuan Road, Shanghai 200241,
China
┴School of Pharmacy and Biomedical Sciences, Faculty of Health Sciences, Curtin University,
Bentley, Perth, WA 6102, Australia
Shanghai Key Laboratory of Green Chemistry and Chemical Process, School of Chemistry and
Molecular Engineering, East China Normal University, 3663 North Zhongshan Road, Shanghai
200062, China
Graphical Abstract
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Abstract
Since the initial discovery as nicotinic acetylcholine receptor ligands, the 3-alkoxyisoxazole scaffold
has now been shown as a versatile platform for the development of potent σ1 and σ2 receptor ligands.
Herein we report a further SAR exploration on the 3-alkoxyisoxazole scaffold with the aim of
obtaining potent σ2 receptor ligands. Various substitutions on the benzene ring and the basic amino
regions resulted in a total of 21 compounds that were tested for their binding affinities to σ2 receptor.
In particular, compound 51 was identified as one of the most potent σ2 ligands within the series with
a Ki value of 7.9 nM, and demonstrated potent anti-proliferative effects on the both osteosarcoma cell
lines 143B and MOS-J (IC50 values of 0.89 and 0.71 μM, respectively), compared to siramesine (IC50
values of 1.81 and 2.01 μM). Moreover, compound 51 inhibited the clonal formation of the
osteosarcoma 143B cells at 1 μM concentration, corresponding to half the dose required of
siramesine for similar effects. The general cytotoxicity profile of compound 51 was assessed in a
number of normal cell lines, including HaCaT, HAF, and LO2 cells. Furthermore, FACS analysis
showed that compound 51 likely inhibits the osteosarcoma cell growth by disruption of the cell cycle
and promotion of cell apoptosis.
Keywords: σ2 receptor, 3-alkoxyisoxazoles, osteosarcoma, antiproliferative effects, colony
formation, cytotoxicity, siramesine
Introduction
The sigma (σ) receptors were originally classified by Martin et al. in 1976 as an additional subtype
of opioid receptors family [1]. However, later it was found that the pharmacological effects produced
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
by SKF-10047 failed to be antagonized by naloxone [2]. Since then there has been subsequent
confusion in the literature, associating the σ receptors with the PCP binding site of NMDA receptor.
Currently, σ receptors are considered a unique entity with no homology to opioid receptors or other
mammalian proteins [3], and two distinct subtypes termed σ1 and σ2 have been identified [4]. Both of
these receptors have been implicated in the pathophysiology of various neurological disorders and
[6]
cancer [5]. The σ1 receptor contains 223 amino acids with two transmembrane domains
and the
crystal structure of the human σ1 receptor in complex with two chemically divergent ligands was
first reported in 2016 [7]. The σ1 receptor regulates the activity of a large number of cellular proteins
[8]
including ion channels, G protein-coupled receptors (GPCRs), transcription factors, endoplasmic
reticulum (ER) chaperone protein BiP, as well as acting as a molecular chaperone to repair misfolded
proteins [9]. Activation of the σ1 receptor results in the transport of cholesterol, ceramides, and
essential amino acids that are necessary for cell growth and reproduction, along with the synthesis of
various growth factors involved in neuronal growth [10]. Very recently, choline has been identified as
the endogenous ligand for σ1 receptor [11]
.
In 1990, Hellewell et al. first reported a distinct population of  receptors that were different to those
from guinea pig brain, which was then termed the σ2 receptor [12]. It comprises of 18-21 kDa
membrane protein [12], which are highly expressed in the liver, kidney, and proliferating tumor tissues
[5a, 13]. Similar to the 1 story, the identification of 2 receptor has taken many twists and turns. At
the outset, σ2 receptor was considered as a histone binding protein [14]. In 2011, progesterone
receptor membrane component 1 (PGRMC1) protein complex, a potential therapeutic target for
inhibiting tumorigenesis [15], was thought to be the binding site of σ2 receptor as overexpression and
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
knockdown of the PGRMC1 protein results in a subsequent increase and decrease, respectively, of
the σ2 selective radioligand binding [16]. However, recent studies have suggested that PGRMC1 and
σ2 receptor are two different molecular entities as evidenced by fluorescent labeling experiments [17]
.
In 2017, Alon et al. identified TMEM97, which is an endoplasmic reticulum-resident transmembrane
protein that regulates the sterol transporter NPC1. It was revealed that Asp 29 and Asp 56 are
essential amino acid residues for ligand recognition [18]
.
Even before its identification and molecular cloning, the σ2 receptor has been long considered to be a
potential therapeutic target for cancer. From the initial proposal as reliable biomarkers of tumors [13b]
,
a variety of structurally diverse σ2 receptor ligands have been studied in different tumors, including
pancreatic cancer, breast cancer, metastatic melanoma, and neuroblastoma, with siramesine (1, Ki, σ1
= 17.0 nM, Ki, σ2 = 0.12 nM), SV-119 (2, Ki, σ1 = 1417 nM, Ki, σ2 = 5.2 nM ), WC-26 (4, Ki, σ1 =
1436 nM, Ki, σ2 = 2.6 nM), and RHM-138 (5, Ki, σ1 = 544 nM, Ki, σ2 = 12.3 nM) utilized in
preclinical trials (Figure 1) [19]. Although the exact roles of σ2 receptor in these tumors remain
unclear, development of selective 2 radioligands in tandem with fluorescent labeling have led to a
better understanding of the cellular actions and biological pathways associated with the 2 receptor.
For example, SW-43 (3, Ki, σ2 = 18 nM), a fluorescently labeled σ2 receptor ligand, mediated
pancreatic cancer cells death by lysosomal membrane permeabilization, while PB-28 (6, Ki, σ1 =
13.0 nM, Ki, σ2 = 0.28 nM) could induce cell death via mitochondrial superoxide production and
caspase activation in pancreatic cancer (Figure 1) [19a, 20]. For human SK-N-SH neuroblastoma cells,
CM-572 (7, Ki, σ1 > 10 μM, Ki, σ2 = 14.6 nM) induced an immediate dose-dependent increase in
cytosolic calcium concentration，and the newer benzoxazole and benzothiazole derivatives were
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
reported to inhibit the viability of breast cancer cells [19h, 21]. A study with fluorescent probe K05-138
(8, Ki, σ1 = 1100 nM, Ki, σ2 = 45 nM) suggests that σ2 ligands are internalized by an endocytotic
pathway in breast cancer cells [22]. Other potential antitumour mechanisms have also been reported in
conjunction with these σ2 receptor ligands, including reactive oxygen species (ROS), P-glycoprotein,
[19c, 23]
DNA fragmentation, proline rich acidic protein 1 (PRAP-1) cleavage, among others.
.
Moreover, σ2 ligands have also been proposed to be effective for imaging the proliferation status in
tumors as potential diagnostic tools for cancer [24], with [18F] ISO-1 (9，Ki, σ1 = 330 nM, Ki, σ2 = 7.0
nM) currently being evaluated in the phase I clinical trial as a potential positron emission
tomography (PET) marker [24d]. Synergistic or potentiating effects of σ2 receptor ligand with an
anticancer agent, as exemplified by SV-119 combined with gemcitabine or paclitaxel, was also
demonstrated to stabilize pancreatic cancer progression in rodents [19b, 25]. Other studies also showed
that σ2 receptor ligands could be used in combination with other antitumor agents to selectively
target the cancer cells [26]. In addition, σ2 receptor ligands have also been associated with the
modulation of metabolic stimulation, neuralgia, and anticonvulsant effects [11, 27]
.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Figure 1. Structures of selected σ2 receptor ligands.
Our previous work in the development of  receptor ligands have identified 3-alkoxyisoxazole 10
scaffold as σ1 receptor ligands with high affinity and selectivity (Ki, σ1 = 4.1 nM, Ki, σ2 = 1312 nM)
[28]. This compound was originally obtained during our investigation of the 4β2 nicotinic receptor
ligands through a comparative pharmacophore analysis between the nicotinic receptor and the σ1
receptor. Since then, we have identified σ1 ligands with good pharmacokinetic properties and
excellent brain/plasma ratio, one of which has been shown to possess significant analgesic effects in
formalin-induced mouse inflammatory pain model [29]. Subsequently, incorporation of a
cyclobutylaminoethoxy side-chain on the isoxazole afforded highly potent and selective σ1 receptor
ligands that demonstrate neurite outgrowth activity on the N1E-115 neurite cell model [30]. In the
present study, we report our structure-activity relationship (SAR) findings on the 3-alkoxyisoxazole
derivatives that possess moderate to good σ2 affinities (Ki values in the range of 1.8 nM to 1008 nM)
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
and their potential antitumor activities in osteosarcoma cells.
Results and Discussion
Chemistry
As described previously [29], the key intermediate 12 was synthesized starting from the commercially
available dimethyl 2-butynedioate 11 in 4 steps, involving sequential [2+3] cycloaddition, Mitsunobu
reaction, reduction, and iodination (Scheme 1). Compounds 13-25 were synthesized via substitution
reaction of iodide 12 with various phenols under basic condition, followed by deprotection of the
Boc group in the presence of either hydrochloric acid or trifluoroacetic acid (TFA). Compounds
23-25 underwent reductive amination with formaldehyde to give the tertiary amines 26-28.
Compound 30, which differs in the position of the secondary amine, was synthesized in a similar
manner to compound 13 using compound 29 as the starting material.
Scheme 1.
Reagents and conditions: (a) substituted phenol, K2CO3, DMF, rt; (b) HCl, EtOAc, rt or TFA, CH2Cl2,
rt; (c) HCHO, NaBH4, CF3CH2OH, 40 ºC .
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
A series of N-cycloalkyl substituents were introduced instead of the original pyrrolidine ring to afford
the N-cycloalkylaminoethoxyisoxazole derivatives 35-41 (Scheme 2). In a similar manner to the
synthesis of compounds 13-25 (Scheme 1), ethanolamine 31 and N-methylethanolamine 32
underwent sequential Mitsunobu reaction with 3-hydroxyisoxazole-5-carboxylic acid methyl ester,
lithium borohydride reduction, and iodination to give the iodides 33 and 34 [30]. Substitution reaction
with the appropriate phenols under basic conditions, Boc deprotection with HCl, and reductive
amination with the corresponding cyclic ketones provided compounds 35-41. Alkaline hydrolysis of
ester 41 followed by amide coupling with dimethylamine using HATU in DMF as a solvent afforded
amide 42. Reduction with lithium aluminum hydride formed the final diamine 43.
In turn, a series of N-(aminoalkyl) substituents were attached to the pyrrolidine site to give the
isoxazoles 48-51 as described in Scheme 2. Iodoaminoalkanes 46 and 47 were first prepared from the
corresponding aminoalcohols 44 and 45, respectively, via Boc protection and iodination using Appel
condition. Substitution reaction of iodoaminoalkanes 46-47 with compounds 10, 16, and 21 gave the
desired final compounds 48-51. All the final products 10, 13-28, 30, 35-43, and 48-51 were obtained
either as hydrochloride or trifluoroacetate salts with the exception of free bases 26-28 and 42.
Scheme 2.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Reagents and conditions: (a) phenol or substituted phenol, K2CO3, DMF, rt; (b) HCl, EtOAc, rt, then
aqueous NaOH solution; (c) cyclic ketone, NaBH3CN, CH3CO2H, CH3OH, rt; (d) NaOH, THF/H2O,
rt;
(e)
dimethylamine
hydrochloride,
1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate
(HATU), N, N-diisopropylethylamine, DMF, rt; (f) LiAlH4, THF, 0 ºC to rt; (g) di-tert butyl
dicarbonate, K2CO3, THF/H2O, rt; (h) I2, PPh3, imidazole, CH2Cl2, 0 ºC to rt; (i) K2CO3, DMF, rt or
heat to 60 ºC ; (j) HCl/EtOAc, rt.
Radioligand Binding Studies
In vitro binding affinities of compounds 10, 13-28, 30, 35-40, 42, 43, and 48-51 at σ1 andσ2 receptor
were determined by the National Institute of Mental Health’s Psychoactive Drug Screening Program
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
(NIMH PDSP) using the standard [3H]-(+)-pentazocine and [3H] DTG binding assay with rat brain
homogenate and PC12 cells respectively(Tables 1 and 2). Firstly, keeping the 2-pyrrolidine moiety
intact, the effects of monosubstitution on the phenolic ether region were explored including fluoro
(13-15), trifluoromethyl (16-17), chloro (18) and methoxy (19) groups, to investigate the optimal
electronic requirements for enhanced σ2 affinity. The o-fluoro substituted compound 13 had
approximately 3.3-fold increase on the σ2 receptor affinity vs the parent compound 10, while the
m-fluoro substituted (14) and p-fluoro substituted (15) analogues had 5-fold and 3-fold increase
respectively, indicating that m-substitution was preferred for optimal σ2 receptor affinity.
Introduction of a trifluoromethyl group at the p-position (17) resulted in a 14-fold increase in σ2
receptor affinity (Ki = 91.7 nM) compared to 10. The p-chloro (18) and p-methoxy (19) analogues
showed 9-fold and 2-fold increase, with σ2 Ki values of 139 and 675 nM, respectively. Compared to
17, the trifluoromethyl substitution at the m-position resulted in a slightly more potent compound 16
with a Ki value of 88 nM. Within the tested analogues 13-19, affinity at σ2 receptor was improved
upon monosubstitution with an electron withdrawing group (EWG). Following this observation,
various EWG disubstitutions (fluoro and chloro groups) were introduced to the phenolic ether region
to further probe the electronic requirements on the benzene ring. The m, p-difluoro substituted
compound 20 retained a similar affinity at σ2 receptor (Ki, σ2 = 373 nM) compared to the
monofluoro analogues 13-15. On the other hand, the m, p-dichloro (21) and m, m-dichloro (22)
analogues showed higher σ2 receptor binding affinities than the monochloro analogue 18, with
approximately 29-fold and 25-fold increase on the σ2 affinity relative to parent compound 10.
Simultaneous introduction of fluorine and chlorine at the m- and p-positions (23-25; Ki, σ2 = 128 nM,
202 nM, and 226 nM, respectively) resulted in compounds with higher affinity than the difluoro
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
analogue 20 but lower affinity than the dichloro substituted compounds 21 (Ki, σ2 = 45 nM) and 22
(Ki, σ2 = 52 nM), therefore dichloro substitution was proved to be superior for 2 binding affinity
compared to the monochloro substitution. We also examined the effect of alkyl substitution on the
free nitrogen atom of pyrrolidine ring to probe its ionic/hydrogen bonding requirements. Methylation
of the nitrogen atom of the pyrrolidine (compounds 26-28) generally maintained the σ2 binding
affinites, with Ki values ranging from 71.3 to 324 nM. A further heterocycle alteration from
2-pyrrolidine (10; Ki, σ2 = 1312 nM) to 3-pyrrolidine (30; Ki, σ2 = 685 nM) resulted in a 2-fold
improvement for the σ2 receptor binding, accompanied with approximately 45-fold decrease in 1
affinity. In addition, the σ1 binding assay for compounds 10, 13-28 and 30 showed that the
introduction of substituents on the benzene ring is generally tolerated, with the exception of 28 (Ki,
σ1 = 0.72 nM) being 6-fold more potent than the parent compound 10 and the methoxy-substituted
19 (Ki, σ1 = 210 nM) being the least potent 1 binder in this series, suggesting the unfavorable effect
of monosubstitution with an EDG.
Table 1. The σ2 receptor binding profiles for analogues 10, 13-28, and 30.
Analogue
R1
H
H
H
H
R2
H
Ki σ1a (nM)
4.1 ± 0.7
6.0 ± 0.8
1.4 ± 0.2
4.4 ± 0.6
Ki σ2a (nM)
1312c
10b
13
14
15
2-F
3-F
4-F
396 ± 72
277 ± 32
471 ± 54
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
H
H
3-CF3
4-CF3
6.6 ± 0.6
13.3 ± 1.2
13.0 ± 1.5
210 ± 19
88 ± 12
91.7 ± 10
139 ± 16
675 ± 138
373 ± 44
46.5 ± 8.5
52.2 ± 9.6
129 ± 29
205 ± 47
258 ± 58
258 ± 58
324 ±74
16
17
18
19
20
21
22
23
24
25
26
27
28
30
H
4-Cl
H
4-OCH3
3-F, 4-F
3-Cl, 4-Cl
3-Cl, 5-Cl
3-F, 4-Cl
3-Cl, 4-F
3-Cl, 5-F
3-F, 4-Cl
3-Cl, 4-F
3-Cl, 5-F
-
H
4.71 ± 0.54
12.1 ± 1.4
4.59 ± 0.42
10.8 ± 1.0
3.93 ± 0.54
2.78 ± 0.38
3.34 ± 0.54
3.73 ± 0.34
0.72 ± 0.10
184 ± 34
H
H
H
H
H
CH3
CH3
CH3
-
71.3 ± 7.8
685 ± 109
aSee experimental section. Radioligands: σ1, [3H]-(+)-pentazocine; σ2, [3H] DTG (ditolylguanidine).
Receptor material: σ1, Guinea pig brain homogenate; σ2, PC12 cells. bThe Ki values for compounds
c
10 and 13-19 were cited from the literature.[28-29]. Data shown as average of Ki values from two
measurements: 665 ± 29 nM and 1958 ± 392 nM.
Subsequent SAR studies were focused on the effects of: i) replacing the 2-pyrrolidine ring of
compound 10 with N-cycloalkylaminoethoxy derivatives to probe the steric bulk at the N-terminal
site (35-40; Table 2) and ii) extension of aminoalkyl groups to the N-terminal atom (42, 43, 48-51;
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Table 2), which has proven beneficial for the 2 receptor binding affinities in other reported
scaffolds [23, 24]. The N-cyclopentyl derivative 36 had similar σ1 and σ2 receptor affinities with the
N-cyclobutyl parent compound 35 [30], while the N-cyclohexyl analogue 37 showed an improvement
for σ2 receptor affinity with a Ki value of 23 nM, which is a 9-fold increase relative to the
N-cyclobutyl analogue 35. However, the affinity of compound 37 for σ1 decreased by an order of
magnitude compared to the ring-contracted analogs, indicating that at this N-terminal site, the larger
steric hindrance confers a preference for σ2 receptor affinity over the σ1. Keeping the cyclohexyl
substituent intact, further substitution at the benzene ring region was introduced, such as m,
p-dichloro (38) and m-trifluoromethyl (39), to give single digit nanomolar σ2 receptor ligands with
Ki values of 1.8 nM and 2.5 nM, respectively. As seen in compounds 23-28, N-methylation of
compound 39 resulted in compound 40 (Ki = 4.4 nM) which maintained the binding affinity at σ2
receptor. However, concurrent introduction of N-methyl and dimethylamide group at the cyclohexyl
ring (42) resulted in a significant reduction for σ2 affinity (Ki = 1008 nM). Reduction of the
dimethylamide compound 42 resulted in the dimethylamino analogue 43 with Ki value of 76.6 nM at
the 2 receptor. In order to examine further effects of aminoalkyl substituents around the N-terminal
region, installation of aminoalkyl groups with varying length to the original 2-pyrrolidine group
(48-51 vs 10) resulted in highly potent 2 receptor ligands with binding affinities ranging from 7.9
nM to 74.7 nM. Compound 48 (Ki = 74.7 nM) with 6C-aminoalkyl substitution had 18-fold increase
on σ2 affinity when compared to the parent compound 10. Introduction of a 3-trifluoromethyl group
(48 vs 49) on the phenolic ether region resulted in compound 49 with a Ki (2) value of 8.5 nM. The
3,4-dichlorosubstituted analogue 50 displayed a very similar 2 affinity to compound 49 (Ki = 10.3
nM) and chain contraction from a 6C to 4C (compound 51; Ki = 7.9 nM) was tolerated. In summary,
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
although introduction of methyl groups to the nitrogen atom may not be conducive for σ2 receptor
affinity, incorporation of the alkyl amino groups (48-51) proved to be beneficial for optimal σ2
binding.
Table 2. σ2 receptor binding profiles for analogues 35-40, 42, 43, and 48-51.
Analogue
35b
36
R1
R2
n
0
1
2
2
2
2
2
2
-
R
Ki σ1a (nM)
0.2 ± 0.0
0.48 ± 0.044
2.83 ± 0.26
NTc
Ki σ2a (nM)
198 ± 64
H
H
H
H
H
H
H
268 ± 48
H
H
22.8 ± 4.2
1.81 ± 0.33
2.53 ± 0.35
4.44 ± 0.81
1008 ± 138
76.6 ± 10
74.7 ± 8.6
8.46 ± 1.5
10.3 ± 1.4
7.92 ± 1.1
37
H
H
3-Cl, 4-Cl
3-CF3
3-CF3
H
38
H
H
NT
39
CH3
H
NT
40
CH3
-CON(CH3)2
NT
42
CH3
-CH2N(CH3)2
H
NT
43
d
-(CH2)6NH2
-(CH2)6NH2
-(CH2)6NH2
-(CH2)4NH2
-
-
-
-
H
/
48
-
3-CF3
3-Cl, 4-Cl
3-Cl, 4-Cl
/
/
/
49
-
50
-
51
a See experimental section. Radioligands: σ1, [3H]-(+)-pentazocine; σ2, [3H] DTG (ditolylguanidine).
Receptor material: σ1, Guinea pig brain homogenate; σ2, PC12 cells. b The Ki value for compound
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
35 are cited from the literature. [30] c not tested; d See supporting information Table S1.
In order to obtain a preliminary binding data of the newly synthesized analogues on the 1 receptor,
selected compounds 16, 21, 30, 36, 37, 48-51 were tested at a single concentration on guinea pig 1
receptor. The percentage inhibition data of compounds at 10 M against [3H]-(+)-pentazocine are
listed in Supporting Information Table S1. Compounds 16, 21, 36, and 37 were identified to be the
most potent σ1 binders with 93-97% inhibitory activities, correlating to the observed Ki values of
0.48 to 12.1 nM. On the other hand, compound 30 had 85.9% inhibition on the guinea pig 1 assay,
corresponding to an observed Ki value of 184 nM. Compounds 48-51 all had 16.1-77.8% inhibition
on the guinea pig 1 assay, indicative of triple digit nanomolar Ki values, although further
experiments are still needed. In order to further obtain a rough predictive correlation for these
preliminary 1 binding results, we conducted molecular docking studies on compounds 21 (Ki, σ1 =
12.1 ± 1.4 nM – this value obtained from Psychoactive Drug Screening Program) and 51 (%
inhibition of 77.8% at the guinea pig 1 receptor) utilizing the solved crystal structure of σ1 (PDB
ID 5HK1). The protein structure was prepared by Protein Preparation Wizard, and missing atoms and
all hydrogen atoms were added to the protein according to their local environment in the step. All
compounds were flexibly docked into the binding pocket defined by residues W89, M93, L95, Y103,
L105, F107, I124, D126, W164, and E172 using Glide with XP scoring function. Most reasonable
poses were selected for visualization analysis. The σ1 interactions with each compound were
pictured using Pymol. All of the molecular docking work was performed using the Maestro software.
The proposed binding modes are shown in Figure S2, demonstrating the formation of hydrogen bond
between Y103, E172 and basic nitrogen atom of the pyrrolidine (compound 21; Figure S2A; Glide
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
score of -9.0) which is crucial for 1 receptor binding affinity, consistent with a reported literature
[8]. On the other hand, introduction of an aminoalkyl group on the pyrrolidine nitrogen (compound
51; Glide score of -6.1) resulted in its retroflexed conformation in the receptor pocket and therefore
no longer capable of forming an effective H-bond, consistent with the weak σ1 affinity (Figure S2B).
Evaluation of Alkoxyisoxazole-based 2 Receptor Ligands against Osteosarcoma Cells
Although the role of σ2 receptor has been implicated in various tumor progression, thus far there
have been very limited number of reports with respect to its role in osteosarcoma. Osteosarcoma is
the most common bone tumor and one of the most harmful tumors, which often occurs in young
children and the elderly [31]. Although the 5-year survival rate of osteosarcoma is 60%, it is highly
prone to metastasis, including primary and secondary metastasis [31c, 32]. Once metastasis of
osteosarcoma occurs, its 5-year survival rate drops to 19% [32b, 33], while the survival rates of
pulmonary metastasis and bone metastases were 23% and 0% respectively [31c]. The pathogenesis of
osteosarcoma is complex and remains to be well understood, with some studies demonstrating the
role of mutations in the retinoblastoma gene and p53 proteins [34]. To date, there is still no ideal
therapeutic for osteosarcoma, which is generally treated with two or more anticancer drugs in
combination, including doxorubicin, methotrexate, and cisplatin [35]. With these considerations in
mind, selected alkoxyisoxazoles analogues with potent σ2 receptor affinities (compounds 21, 37,
38-40, 49-51) along with compound 10 were tested for their antiproliferative activities against the
human osteosarcoma cell 143B (Table 3). The N-aminoalkyl chain compounds 49-51 displayed
respectable antiproliferative activities with IC50 values ranging from 0.89-2.81 M while the other
tested compounds were not active (IC50 > 10 M). Siramesine was included as a positive control in
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
this assay and found to also be effective with an IC50 value of 1.81 M. Although it remains to be
seen whether or not inhibition of both 1 and 2 receptors are required, it is clear for the
alkoxylisoxazole scaffold that the N-aminoalkyl chain extension from the pyrrolidine ring represents
an important structural element for the observed antiproliferative activity in osteosarcoma cells.
Table 3. Antiproliferative activities of selected compounds against osteosarcoma cell line 143B.
Compound
Ki σ2 (nM)a
1312
IC50 (μM)b
> 10
10
46.5 ± 8.5
22.8 ± 4.2
1.81 ± 0.33
2.53 ± 0.35
4.44 ± 0.81
8.46 ± 1.5
10.3 ± 1.4
8.02 ± 1.1
0.12
> 10
21
> 10
37
> 10
38
> 10
39
> 10
40
2.81 ± 0.12
3.02 ± 0.12
0.89 ± 0.02
1.81 ± 0.05
49
50
51
Siramesine
a See experimental section. Radioligands: σ2, [3H] DTG (ditolylguanidine); The Ki values listed in
this table are the average ± SEM from a minimum of 3 saturation binding assays. b Osteosarcoma cell
lines were treated with σ2 ligands (gradient concentration) or DMSO vehicle for 48 h, and cell
viability relative to vehicle was determined by crystal violet assay method (Mean ± SEM, n ≥ 3).
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Cytotoxicity against Osteosarcoma Cells and Normal Cells
As a preclinical experimental antitumor drug, siramesine is widely regarded as the compound of
choice for comparison studies in the evaluation of novel σ2 receptor ligands against various tumors
[19e, 19i, 23b]. Cytotoxic activities of the ligands against two osteosarcoma cell lines 143B and MOS-J
(Figure 2A and 2B) were expressed as concentration-dependent inhibition curves. The results
showed that the inhibitory effect of alkoxyisoxazole 51 on the proliferation of both osteosarcoma cell
lines was significantly stronger than that of the positive control siramesine with IC50 values of 0.89
and 0.71 μM vs. 1.81 and 2.01 μM for 143B and MOS-J cells respectively. As seen in Figure 2C,
compound 51 also displayed reduced cytotoxicity to the tested normal cells, including human
immortalized keratinocytes HaCaT (IC50 = 6.47 μM), human embryonic villous cells HAF (IC50 =
3.35 μM) and human normal embryonic liver cells LO2 (IC50 > 10 μM).
Figure 2. A-B. Effects of 51 and siramesine on cell viability was concentration-dependent in
different osteosarcoma cell lines. After seeding for 12 h, 143B and MOS-J cells were treated in the
presence of DMSO vehicle or indicated concentrations of 51 and siramesine. Cells were incubated
for 48 h and cell viability relative to vehicle was determined by crystal violet assay method (n ≥ 3;
bars, SEM); C. Osteosarcoma cells (143B and MOS-J) and normal cells (HaCaT, HAF, and LO2)
were treated in the presence of DMSO vehicle or different concentrations of the σ2 ligands for 48 h.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Then cell viability relative to vehicle was determined by crystal violet assay method (n ≥ 3; bars,
SEM).
Inhibition of the Colony Formation Ability of Osteosarcoma Cells
To further gain insights into the antiproliferative activity, the abilities of the alkoxyisoxazole-based
2 ligands to inhibit colony formation was assessed [36]. The clonogenic assay is principally a
measure of the cell abilities to undergo unlimited division and often used as an assay of choice for
determining the effectiveness of cytotoxic agents. As can be seen in Figure 3, head-to-head
comparison of compound 51 vs siramesine was conducted at 1 and 2 M concentrations. The number
of cells forming colony under co-cultivation with siramesine and compound 51 were significantly
reduced compared to the DMSO control group suggesting that both 51 and siramesine were able to
suppress long-term clonogenic survival and trigger apoptosis.
Figure 3. σ2 ligands inhibited the colony-formation ability of osteosarcoma cell. 143B cells were
plated at a density of 1000 cells per well in six-well plates. After 24 h of adhesion, cells were treated
with compound 51 or DMSO at appropriate density for the experimental goals, and fixed after
culture for another 10 days.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Apoptosis and Cell Cycle Analysis of 3-Alkoxyisoxazole 51 against Osteosarcoma Cells
Cell cycle analysis was performed in the osteosarcoma cell line (Figure 4A,B) as there have been
reports that σ2 receptor ligands may impair the cell cycle [23b]. When compared to the vehicle control
group whereby the proportion of cells was 47.5%, 40.6%, and 11.9% in the G0G1 phase, S phase and
G2M phase respectively, the experimental group treated with compound 51 showed the proportion of
cells being 56.6%, 37.8%, and 5.6% in the G0G1 phase, S phase and G2M phase respectively. These
results suggest that compound 51 could induce cell-cycle arrest in the G0G1 phase. Furthermore,
some σ2 receptor ligands are reported in the literature to inhibit tumor growth in various animal
models by promoting apoptosis [18, 19c, 23]. Within this context, the osteosarcoma cells 143B was
treated with compound 51 in order to confirm the ability of this compound to induce apoptosis
(Figure 4C). Through fluorescence activated cell sorting (FACS) analysis of apoptosis by Annexin V
assay, up to 33.4% osteosarcoma cells were in the process of apoptosis (Q2 and Q3) upon treatment
with 5 μM of compound 51. In comparison, less than 10% of the osteosarcoma cells were in the
process of apoptosis in the vehicle control group. These results demonstrated that the σ2 receptor
ligand 51 could inhibit the osteosarcoma cell growth by inducing apoptosis. Taken together, the
observed antiproliferative activity of compound 51 in osteosarcoma cells could be mediated in part
via disruption of the cell cycle and induction of cell apoptosis.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Figure 4. Apoptosis and cell cycle analysis with 143B cells grown in 10% FBS complete medium.
(A, B) Effect of compound 51 (5 μM) on cell cycle of 143B cells. Cells were treated with ligand for
24 h, then harvested, fixed in ethanol (80%) and stained with propidium iodide. Results were
expressed as percentage of cells residing in different phase of cell cycle. (C) FACS analysis of
apoptosis by Annexin V assay. Cells were exposed for 12 h to compound 51 (5 μM). Q1 area
represents necrotic cells; Q2 late-apoptotic cells; Q3 early-apoptotic cells; Q4 viable cells. The
images are representative of three experiments.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
Conclusions
This present study represents a continuation of our SAR work on the 3-alkoxyisoxazole scaffold as a
versatile ligands for σ receptors. Various EWGs are tolerated in the phenolic ether region
(compounds 13-30; Table 1), however these compounds are mostly potent at the 1 receptor. In this
work, replacement of the 2-pyrrolidine ring with the N-cycloalkylaminoethoxy derivatives (35-40, 42,
43; Table 2) and further extension of aminoalkyl chains to the pyrrolidine nitrogen atom (48-51;
Table 2) resulted in a number of 3-alkoxyisoxazoles derivatives with potent 2 receptor binding
affinities, i.e. those with Ki values of less than 100 nM (compounds 21, 22, 28, 37-40, 43, 48-51).
Evaluation of selected most potent 2 ligands to inhibit the proliferation of osteosarcoma cells
identified compound 51 (Ki σ2 = 7.9 nM) as the most active analogue, with IC50 values of 0.89 μM
and 0.71 μM on 143B and MOS-J cells, respectively. Furthermore, compound 51 was demonstrated
to inhibit clone formation, induce cell-cycle arrest and apoptosis of osteosarcoma cells. While it
remains to be seen whether or not inhibition of both 1 and 2 receptors are necessary for
antiproliferative activity of the alkoxyisoxazoles in osteosarcoma cells, it is clear that the extra
aminoalkyl chain on the pyrrolidine ring is needed. Overall, this study represents the potential use of
3-alkoxyisoxazole-based σ2 receptor ligand 51 for the treatment of osteosarcoma. In vivo studies on
osteosarcoma mice models will be performed in the near future.
Experimental section
Chemistry
General methods
All starting reagents and solvents were obtained commercially and used without further purification,
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
except that anhydrous THF and CH2Cl2 were acquired by distillation over sodium or CaH2,
respectively. All non-aqueous reactions were run under an inert atmosphere without moisture. The
progress of reactions was monitored by TLC performed on silica gel GF254 or dipping into KMnO4
solution or phosphomolybdic acid solution followed by heating. Silica gel for column
chromatography (CC) was of 200-300 mesh particle size. 1H and 13C NMR spectra were recorded on
a Bruker Advance III 400 spectrometer at 400 MHz (1H) and 101 MHz (13C). Specific rotation was
performed on Rudolph Research Analytical Autopol VI Polarimeter (λ = 589 nm, T = 20 ºC). High
resolution mass spectra were performed using a Bruker ESI-TOF high-resolution mass spectrometer.
Purity of all final compounds (> 95%) was detected by analytical HPLC, which was carried out on
an Agilent 1200 HPLC system with a ZORBAX Eclipse XDB-C18 column, with detection at 220,
254, or 270 nm on a variable wavelength detector G1365D; flow rate = 1.4 mL/min; gradient of 0−
100% methanol in water (both containing 0.05 vol% of TFA) in 25 min.
General procedures
The procedure for N-tert-butoxycarbonyl protection reaction (A). To a stirred suspension of
amine (6.0 mmol) and K2CO3 (12 mmol) in THF/H2O (20 mL, 1:1) was added di-tert butyl
bicarbonate (6.3 mmol) slowly at 0 ºC under the nitrogen. After stirring 3h at rt, the suspension was
diluted with water (50 mL) and extracted by EtOAc (30 mL × 3), the combined organic layer was
washed by brine, dried over Na2SO4 and evaporated in vacuo. The residue was purified by flash
chromatography to give the product.
The procedure for the preparation of iodide (B). To a stirred solution of alcohol (1.4 mmol),
imidazole (4.2 mmol), and PPh3 (2.1 mmol) in anhydrous CH2Cl2 (20 mL) was added I2 (2.1 mmol)
with ice cooling under nitrogen. After stirring 2h at rt, the solvent was evaporated and the residue
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
was purified by flash chromatography to give the iodide product.
The procedure for the preparation of phenyl ethers (C). To a stirred solution of an iodide (0.9
mmol) and substituted phenol (1.8 mmol) in anhydrous DMF (4 mL) was added K2CO3 (2.7 mmol)
under N2. After stirring overnight at rt, the mixture was diluted with water (50 mL) and extracted
with EtOAc (2×30 mL), then the combined organic layer were washed with brine (30 mL), dried
over Na2SO4, concentrated in vacuum and the residue was purified by flash chromatography to give
the product and advanced to the next step without further purification.
The procedure for N-tert-butoxycarbonyl deprotection reaction (D). D1：To a solution of the
N-Boc protected compound (0.39 mmol) in EtOAc (2 mL) was added HCl/EtOAc (2 mol/L, 2 mL)
under nitrogen. After stirring overnight, the solvent was evaporated and the residue was triturated
with diethyl ether (10 mL). The precipitate was filtered to give the HCl salt. D2：To a solution of the
N-Boc protected compound (0.39 mmol) in CH2Cl2 (10 mL) was added TFA (1 mL). After stirring
overnight the solvent was removed and the residue was purified by preparation HPLC followed by
lyophilized to obtain the TFA salt.
The procedure for N-methylation reaction (E). To a solution of amine (1.0 mmol) and HCHO (1.0
mL) in CF3CH2OH (4 mL) was added NaBH4 (1.2 mmol) at 0°C under nitrogen. After stirring for 2
h at 40 ºC , the solvent was diluted with EtOAc (40 mL), washed with water (40 mL) and brine (30
mL), dried over Na2SO4 and concentrated under vacuum. The residue was purified by flash
chromatography to give N-methylation product.
The procedure for N-cycloalkylation reaction (F). To a solution of amine (1.0 mmol), ketone (1.2
mmol) and acetic acid (0.1 mL) in CH3OH (20 mL) was added Na(CN)BH3 (1.4 mmol) at rt under
nitrogen. After stirring overnight, the solvent was diluted with EtOAc (40 mL), washed with water
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
(40 mL) and brine (30 mL), dried over Na2SO4 and concentrated under vacuum. The residue was
purified by flash chromatography to give N-cycloalkylation product.
The procedure for N-aminoalkylation reaction (G). A solution of reactant (0.55 mmol) and iodide
(0.77 mmol) dissolved in anhydrous DMF (4 mL) was added K2CO3 (2.2 mmol) under N2. After
stirring for 12 h at rt, the mixture was diluted with water (50 mL) and extracted with EtOAc (3×30
mL), then the combined organic layers were washed with water (80×2 mL) and brine (50 mL), dried
over Na2SO4, concentrated in vacuo and the residue was purified by flash chromatography to give
the product.
(S)-2-(((5-((2-Fluorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyrrolidin-1-ium chloride (13).
This compound was obtained from 12 and 2-fluorophenol employing methods C and D1. White solid;
total yield 55% for 2 steps; purity 98.0%; [α]2D0 = + 13 (c = 1 mg/mL, MeOH). 1H NMR (400 MHz,
DMSO-d6) δ 9.90 (br s, 1H), 9.39 (br s, 1H), 7.37–7.19 (m, 2H, phenol), 7.18–7.11 (m, 1H, phenol),
7.00 (s, 1H, phenol), 6.47 (s, 1H, isoxazol), 5.27 (s, 2H, CH2-O-phenol), 4.44–4.38 (m, 2H,
CH2-O-isoxazol), 3.91–3.87 (m, 1H, pyrrolidin), 3.18–3.16 (m, 2H, pyrrolidin), 2.22–1.99 (m, 1H,
pyrrolidin ), 1.97–1.78 (m, 2H, pyrrolidin ), 1.73–1.54 (m, 1H, pyrrolidin). 13C NMR (101 MHz,
DMSO-d6) δ 170.7 (isoxazol), 168.5 (isoxazol), 151.7 (d, JC-F = 247.0 Hz, phenol), 145.2 (d, JC-F
10.3 Hz, phenol), 124.8 (d, JC-F = 3.4 Hz, phenol ), 122.2 (d, JC-F = 6.9 Hz, phenol), 116.2 (d, JC-F
=
=
18.3 Hz, phenol), 115.6 (phenol), 95.8 (isoxazol), 68.4 (C-O-phenol), 61.3 (C-O-isoxazol), 57.3
(pyrrolidin), 44.9 (pyrrolidin), 26.1(pyrrolidin) , 23.2 (pyrrolidin). HRMS (ESI): calcd for
C15H18FN2O3 [M+H]+, 293.1296; found, 293.1293. The interpretation of 1H and 13C NMR signals of
14-30 are similar with 13.
(S)-2-(((5-((3-Fluorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyrrolidin-1-ium chloride (14).
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
This compound was obtained from 12 and 3-fluorophenol employing methods C and D1. White solid;
total yield 51% for 2 steps; purity 97.9%; [α]2D0 = + 9.3 (c = 1 mg/mL, MeOH). 1H NMR (400 MHz,
DMSO-d6) δ 9.29 (br s, 2H), 7.35 (q, J = 8.3 Hz, 1H), 6.95 (dt, J = 11.2, 2.4 Hz, 1H), 6.89 (dd, J =
8.3, 2.4 Hz, 1H), 6.83 (td, J = 8.6, 2.4 Hz, 1H), 6.47 (s, 1H), 5.23 (s, 2H), 4.44 (dd, J = 11.2, 3.9 Hz,
1H), 4.36 (dd, J = 11.2, 8.2 Hz, 1H), 3.91 (qd, J = 8.1, 3.8 Hz, 1H), 3.25–3.12 (m, 2H), 2.16–2.05 (m,
13
1H), 2.01–1.83 (m, 2H), 1.77–1.65 (m, 1H). C NMR (101 MHz, DMSO-d6) δ 170.8, 168.6, 162.9
(d, JC-F = 244.1 Hz), 158.8 (d, JC-F = 11.1 Hz), 130.8 (d, JC-F = 10.1 Hz), 111.1, 108.1 (d, JC-F = 21.3
Hz), 102.4 (d, JC-F = 25.4 Hz), 95.7, 68.4, 60.8, 57.3, 44.9, 26.2, 23.2. HRMS (ESI): calcd for
C15H18FN2O3 [M+H]+, 293.1296; found, 293.1293.
(S)-2-(((5-((4-Fluorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyrrolidin-1-ium chloride (15).
This compound was obtained from 12 and 4-fluorophenol employing methods C and D1. White solid;
total yield 53% for 2 steps; purity 96.1%, [α]2D0 = + 11 (c = 1 mg/mL, MeOH). H NMR (400 MHz,
1
DMSO-d6) δ 9.77 (br s, 1H), 9.26 (br s, 1H), 7.19–7.11 (m, 2H), 7.10–7.02 (m, 2H), 6.44 (s, 1H),
5.18 (s, 2H), 4.49–4.33 (m, 2H), 3.97–3.83 (m, 1H), 3.25–3.12 (m, 2H), 2.15–2.04 (m, 1H), 2.01–
13
1.81 (m, 2H), 1.74–1.65 (m, 1H). C NMR (101 MHz, DMSO-d6) δ 170.8, 169.1, 157.0 (d, JC-F
=
237.8 Hz), 153.7 (d, JC-F = 1.9 Hz), 116.3 (d, JC-F = 8.2 Hz), 116.0 (d, JC-F = 23.2 Hz), 95.6, 68.5, 61.1,
57.4, 45.1, 26.2, 23.3. HRMS (ESI): calcd for C15H18FN2O3 [M+H]+, 293.1296; found, 293.1293.
(S)-2-(((5-((3-(Trifluoromethyl)methyl)isoxazol-3-yl)oxy)methyl)pyrrolidin-1-ium-2,2,2-trifluor
oacetate (16). This compound was obtained from 12 and 3-trifluoromethylphenol employing method
C and D2. Yellow oil; total yield 34% for 2 steps; purity 98.5%; [α]2D0 = + 22 (c = 1 mg/mL, MeOH).
1H NMR (400 MHz, DMSO-d6) δ 9.59 (br s, 1H), 9.06 (br s, 1H), 7.56 (t, J = 8.0 Hz, 1H), 7.37 (m,
2H), 6.47 (s, 1H), 5.32 (s, 2H), 4.45 (dd, J = 8.0, 4.0 Hz, 1H), 4.34 (dd, J = 10.0, 4.0 Hz, 1H), 3.96
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
(m, 1H), 3.23 (t, J = 6.0 Hz, 2H), 2.18-2.06 (m, 1H), 2.04–1.81 (m, 2H), 1.81–1.61 (m, 1H). 13C
NMR (101 MHz, DMSO-d6) δ 170.8, 168.6, 158.5 (TFA), 157.7, 130.8, 130.2 (q, JC−F = 31.6 Hz),
123.8 (q, JC−F = 271.0 Hz), 119.0, 118.0 (q, JC−F = 4.0 Hz), 115.6 (TFA), 111.4 (q, JC−F = 4.0 Hz),
95.6, 68.4, 60.8, 57.5, 45.2, 26.0, 23.3. HRMS (ESI): calcd for C16H18F3N2O3 [M+H]+, 343.1264;
found, 343.1281.
(S)-2-(((5-((4-(Trifluoromethyl)phenoxy)methyl)isoxazol-3-yl)oxy)methyl)
pyrrolidin-1-ium
chloride (17). This compound was obtained from 12 and 4-trifluoromethylphenol employing
methods C and D1. White solid; total yield 36% for 2 steps; purity 98.1%; [α]2D0 = + 17 (c = 1
mg/mL, MeOH). 1H NMR (400 MHz, DMSO-d6) δ 9.40 (br s, 1H), 8.96 (br s, 1H), 7.69 (d, J = 8.6
Hz, 2H), 7.23 (d, J = 8.6 Hz, 2H), 6.48 (s, 1H), 5.32 (s, 2H), 4.45 (dd, J = 11.2, 3.6 Hz, 1H), 4.40–
4.27 (m, 1H), 3.94 (dd, J = 7.9, 3.3 Hz, 1H), 3.22 (t, J = 6.6 Hz, 2H), 2.19–2.04 (m, 1H), 2.04–1.84
(m, 2H), 1.80-1.64 (m, 1H). 13C NMR (101 MHz, DMSO-d6) δ 170.8, 168.6, 160.2, 127.04 (q, JC-F
=
3.7 Hz), 124.4 (q, JC-F = 271.9 Hz), 122.0 (q, JC-F = 32.2 Hz), 115.3, 95.7, 68.5, 60.7, 57.6, 45.3, 26.0,
23.3. HRMS (ESI): calcd for C16H18F3N2O3 [M+H]+, 343.1264; found, 343.1296.
(S)-2-(((5-((4-Chlorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyrrolidin-1-ium-2,2,2-trifluoro
acetate (18). This compound was obtained from 12 and 4-chlorophenol employing methods C and
D2. White solid; total yield 46% for 2 steps; purity 97.1%; [α]2D0 = + 8.3 (c = 1 mg/mL, MeOH). 1H
NMR (400 MHz, DMSO-d6) δ 9.44 (br s, 1H), 8.93 (br s, 1H), 7.36 (d, J = 9.0 Hz, 2H), 7.07 (d, J =
8.9 Hz, 2H), 6.44 (s, 1H), 5.21 (s, 2H), 4.44 (dd, J = 11.2, 3.6 Hz, 1H), 4.33 (dd, J = 11.0, 8.6 Hz,
1H), 3.94 (s, 1H), 3.21–3.17 (m, 2H), 2.17-2.06 (m, 1H), 2.02-1.84 (m, 2H), 1.77–1.64 (m, 1H). 13C
NMR (101 MHz, DMSO-d6) δ 170.8, 168.9, 158.1 (TFA), 156.2, 129.3, 125.2, 116.9 (TFA), 116.6,
95.5, 68.5, 60.8, 57.6, 45.3, 26.0, 23.3. HRMS (ESI): calcd for C15H18ClN2O3 [M+H]+, 309.1000;
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
found, 309.0996.
(S)-2-(((5-((4-Methoxyphenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyrolidin-1-ium-2,2,2-trifluoro
acetate (19). This compound was obtained from 12 and 4-methoxyphenol employing methods C and
D2. White solid; total yield 57% for 2 steps; purity 98.9%; [α]2D0 = + 16 (c = 1 mg/mL, MeOH). 1H
NMR (400 MHz, DMSO-d6) δ 9.40 (br s, 1H), 8.94 (br s, 1H), 6.97 (d, J = 9.0 Hz, 2H), 6.87 (d, J =
9.0 Hz, 2H), 6.41 (s, 1H), 5.12 (s, 2H), 4.44 (dd, J = 11.2, 3.5 Hz, 1H), 4.36–4.24 (m, 1H), 3.94 (d, J
= 7.7 Hz, 1H), 3.70 (s, 3H), 3.21 (t, J = 6.8 Hz, 2H), 2.22–2.01 (m, 1H), 2.03–1.77 (m, 2H), 1.81–
1.58 (m, 1H). 13C NMR (101 MHz, DMSO-d6) δ 170.7, 169.5, 157.1 (TFA), 153.9, 151.3, 117.2
(TFA), 115.8, 114.6, 95.2, 68.5, 61.1, 57.6, 55.3, 45.3, 26.0, 23.3. HRMS (ESI): calcd for
C16H21N2O4 [M+H]+, 305.1496; found, 305.1517.
(S)-2-(((5-((3,4-Difluorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyrro-ledin-1-ium chloride
(20). This compound was obtained from 12 and 3,4-difluorophenol employing methods C and D1.
White solid; total yield 30% for 2 steps; purity 98.9%; [α]2D0 = + 17 (c = 1 mg/mL, MeOH). 1H NMR
(400 MHz, D2O) δ 7.13 (q, J = 9.9 Hz, 1H), 6.91–6.86 (m, 1H), 6.78–6.67 (m, 1H), 6.21 (s, 1H),
5.06 (s, 2H), 4.48 (dd, J = 11.5, 3.2 Hz, 1H), 4.31 (dd, J = 11.4, 7.8 Hz, 1H), 4.03 (qd, J = 8.1, 3.3 Hz,
13
1H), 3.33 (t, J = 7.2 Hz, 2H), 2.25–2.12 (m, 1H), 2.13–1.92 (m, 2H), 1.88–1.78 (m, 1H). C NMR
(101 MHz, D2O) δ 171.2, 169.1, 153.4 (dd, JC-F = 9.0, 2.3 Hz), 151.0 (dd, JC-F = 247.7, 13.9 Hz),
145.5 (dd, JC-F = 239.8, 12.7 Hz), 117.0 (dd, JC-F = 18.8, 1.2 Hz), 111.0 (dd, JC-F = 6.3, 3.4 Hz), 105.0
(d, JC-F = 20.7 Hz), 95.5, 68.1, 62.0, 58.5, 45.9, 25.7, 23.3. HRMS (ESI): calcd for C15H17F2N2O3
[M+H]+, 311.1202; found, 312.1206.
(S)-2-(((5-((3,4-Dichlorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyro-lidin-1-ium
chloride
(21). This compound was obtained from 12 and 3,4-dichlorophenol employing methods C and D1.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000461
ChemMedChem
White solid; total yield 38% for 2 steps; purity 97.0%; [α]2D0 = + 13 (c = 1 mg/mL, MeOH). 1H NMR
(400 MHz, DMSO-d6) δ 9.60 (br s, 2H), 7.57 (d, J = 8.9 Hz, 1H), 7.40 (s, 1H), 7.08 (d, J = 8.5 Hz,
1H), 6.48 (s, 1H), 5.27 (s, 2H), 4.42 (d, J = 7.4 Hz, 2H), 4.02–3.79 (m, 1H), 3.18 (d, J = 6.5 Hz, 2H),
2.24–2.01 (m, 1H), 2.01–1.78 (m, 2H), 1.77–1.65 (m, 1H). 13C NMR (101 MHz, DMSO-d6) δ 170.8,
168.4, 156.8, 131.7, 131.1, 123.4, 116.8, 115.8, 95.8, 68.5, 61.1, 57.3, 44.9, 26.2, 23.2. HRMS (ESI):
calcd for C15H17Cl2N2O3 [M+H]+, 343.0611; found, 343.0631.
(S)-2-(((5-((3,5-Dichlorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)pyro-lidin-1-ium
chloride
(22). This compound was obtained from 12 and 3,5-dichlorophenol employing methods C and D1.
White solid; total yield 46% for 2 steps; purity 95.3%; [α]2D0 = + 14 (c = 1 mg/mL, MeOH). 1H NMR
(400 MHz, DMSO-d6) δ 9.61 (br s, 1H), 9.13 (br s, 1H), 7.23 (t, J = 1.6 Hz, 1H), 7.19 (d, J = 1.6 Hz,
2H), 6.48 (s, 1H), 5.29 (s, 2H), 4.50–4.30 (m, 2H), 3.98–3.85 (m, 1H), 3.24–3.11 (m, 2H), 2.14–2.05
13
(m, 1H), 2.00–1.84 (m, 2H), 1.77–1.65 (m, 1H). C NMR (101 MHz, DMSO-d6) δ 170.8, 168.3,
158.8, 134.7, 121.3, 114.3, 95.8, 68.5, 61.1, 57.4, 45.1, 26.1, 23.2. HRMS (ESI): calcd for
C15H17Cl2N2O3 [M+H]+, 343.0611; found, 343.0608.
(S)-2-(((5-((4-Chloro-3-fluorophenoxy)methyl)isoxazol-3-yl)oxy)methyl)
pyrrolidin-1-ium
chloride (23). This compound was obtained from 12 and 3-fluoro-4-chlorophenol employing
methods C and D1. White solid; total yield 40% for 2 steps; purity 99.0%; [α]2D0 = + 17 (c = 1
1
mg/mL, MeOH). H NMR (400 MHz, DMSO-d6) δ 9.73 (br s, 1H), 9.23 (br s, 1H), 7.52 (t, J = 8.8
Hz, 1H), 7.22 (dd, J = 11.3, 2.8 Hz, 1H), 6.91 (dd, J = 8.8, 1.8 Hz, 1H), 6.48 (s, 1H), 5.25 (s, 2H),
4.59–4.22 (m, 2H), 3.91 (s, 1H), 3.20 (s, 2H), 2.20–2.04 (m, 1H), 2.01–1.83 (m, 2H), 1.77–1.61 (m,
1H). 13C NMR (101 MHz, D2O) δ 171.2, 168.8, 157.4 (d, JC-F = 247.1 Hz), 157.0 (d, JC-F = 9.9 Hz),
130.8, 113.0 (d, JC-F = 17.8 Hz), 111.8 (d, JC-F = 3.2 Hz), 104.0 (d, JC-F = 24.8 Hz), 95.5, 68.1, 61.6,
This article is protected by copyright. All rights reserved.