Stability studies of oxazolidine-based compounds using 1H NMR spectroscopy

Article abstract of DOI:10.1002/jps.22108

A series of oxazolidine-based compounds with a variety of substituents in positions 2 and 3 was synthesized and their stability studied. Ring opened intermediates formed on addition of limiting amounts of D₂O to oxazolidine solutions, as observ

Full text of DOI:10.1002/jps.22108

DRUG DISCOVERY INTERFACE
Stability Studies of Oxazolidine-Based Compounds Using
1H NMR Spectroscopy
GERARD P. MOLONEY,1 MAGDY N. ISKANDER,1 DAVID J. CRAIK2
1Department of Medicinal Chemistry, Victorian College of Pharmacy, (Monash University), 381 Royal Parade, Parkville,
Victoria 3052, Australia
2Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland 4072, Australia
Received 23 February 2009; revised 24 December 2009; accepted 18 January 2010
Published online 12 March 2010 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.22108
ABSTRACT: A series of oxazolidine-based compounds with a variety of substituents in positions
2 and 3 was synthesized and their stability studied. Ring opened intermediates formed
on addition of limiting amounts of D2O to oxazolidine solutions, as observed by NMR. As the
hydrolysis reactions proceeded, a series of novel dimeric b-amino alcohol compounds formed via
an internal reaction between ephedrine and the ring opened intermediates. 2-Phenyl substi-
tuted oxazolidine compounds containing electron withdrawing nitro substituents were more
rapidly hydrolyzed than the unsubstituted derivative and methoxy substituted compounds, with
the nitro substituents appearing to stabilize the ring opened intermediates. Two oxazolidine
derivatives, with a methyl and proton at position 2, were found to be more stable to oxazolidine
hydrolysis than the 2-phenyl substituted compounds. Oxazolidines incorporating phenyl sub-
stituents at position 3 were synthesized and found to be less stable than those incorporating a
methyl substituent at position 3. These fundamental structure–activity relationships may be
useful when choosing oxazolidine derivatives as synthetic intermediates and as prodrugs for
the delivery of compounds containing either b-amino alcohol or aldehyde components. ß 2010
Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 99:3362–3371, 2010
Keywords: NMR spectroscopy; prodrugs; stabilization; stability; synthesis; structure
INTRODUCTION
b-amino alcohol group. It has also been utilized in
the synthesis of dipeptide compounds as transition
state mimics for proteolytic enzymes.9 Oxazolidines
can be used as labile transient intermediates for
synthesis, with an ability to direct stereochemical
preference. Previous studies have also assessed
the usefulness of oxazolidines as a drug delivery
system10–12 and have shown that oxazolidines
undergo facile hydrolysis in the pH range 1–11 at
378C. It was found that by variation of substitution on
the 2-aryl ring of the oxazolidine derivatives it
was possible to control oxazolidine hydrolysis. The
reaction rates in neutral and basic solutions were
found to decrease with both increasing steric effects of
the substituents derived from the carbonyl compo-
nent, and increasing basicity of the oxazolidines.12
Therefore, it would seem likely that upon adminis-
tration of many of the biologically active oxazolidines
previously described, hydrolysis of the unstable
ring system may occur, resulting in the release of
the b-amino alcohol derivative and the aldehyde or
Oxazolidines 1 (Fig. 1) have been synthesized for
a variety of applications, including bacteriocides,
fungicides, herbicides,1 and mosquito repellents.2
Several members of this class have been applied as
chiral derivatizing or resolving agents,3 and used as
chiral auxiliaries to effect asymmetric inductions in a
variety of reactions.4–6 Ephedrine 2 and pseudoephe-
drine 3 react quantitatively with aldehydes to yield
chiral oxazolidine derivatives with high diastereo-
meric purity.7 Oxazolidines are used widely as optical
inductors due to their easy hydrolysis.8
There have been a number of reports in which the
labile nature of the oxazolidine system has been
exploited, including its use for protection of the
Correspondence to: David J. Craik (Telephone: 61-7-3346-2019;
Fax: 61-7-3346-2101; E-mail: d.craik@imb.uq.edu.au)
Journal of Pharmaceutical Sciences, Vol. 99, 3362–3371 (2010)
ß 2010 Wiley-Liss, Inc. and the American Pharmacists Association
3362
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010

STABILITY STUDIES OF OXAZOLIDINE-BASED COMPOUNDS
3363
Figure 1. Structures of oxazolidines and derivatives.
ketone compound. In the absence of evidence for
modes of action for oxazolidines themselves, it is
possible that the activity of these compounds may be
due to the hydrolysis products.
There are numerous drugs containing a b-amino
alcohol moiety, including sympathomimetic drugs
such as ephedrine, salbutamol, terbutaline, and
ethambutol, as well as b-adrenergic blocking agents;
for example, propanolol and alprenolol. Drugs of the
sympathomimetic class have difficulty transporting
across cell membranes to their site of actions due
to pKa values >8 and low lipophilicity. Furthermore,
drugs of this type cannot be administered dermally
because, in the pH range compatible with skin (pH 3–
8), they largely exist as cations and cannot be
efficiently absorbed. Several of these drugs also go
through pronounced first-pass metabolism; for exam-
ple, in the gastrointestinal tract, phenylephrine,
isoprenaline, and terbutaline undergo significant
first-pass metabolism and thus the ability to admin-
ister these drugs orally is restricted.
Oxazolidines appear to be suitable prodrugs for
molecules containing the b-amino alcohol moiety.
Oxazolidines are much weaker bases (pKa 6–7) than
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010

3364
MOLONEY, ISKANDER, AND CRAIK
the parent b-amino alcohols, giving the cyclic systems
higher lipophilicity at physiological pH.13 This
increased lipophilicity may become advantageous
when b-amino alcohol type drug delivery problems
are due to low lipophilicity, as is the case for dermal
absorption.13
500 mL deuterated solvent) was prepared and 700mL
of D2O was added. Spectra were recorded at various
intervals for up to several weeks. Oxazolidine hydro-
lysis was monitored by the relative intensities of peaks
corresponding to the oxazolidine and the decomposi-
tion products.
There are a number of biologically important
molecules that incorporate a carbonyl group; for
example, steroids. Moreover, side effects due to
higher than desired in vivo levels of steroids resulting
from topical application, may be avoided by using an
oxazolidine prodrug that hydrolyzes slowly to allow
controlled release of the parent drug. In considering
oxazolidines as prodrug candidates for carbonyl
containing substances, their weakly basic character
may be advantageous in that the transformation of
such substances into oxazolidines introduces a read-
ily ionizable moiety, which in turn changes the
solubility characteristics of the parent compound.
The aim of this study was to examine the stability of
a series of compounds incorporating substituents at
positions 2 and 3 of the oxazolidine ring, as
summarized in Figure 1. Compounds 4–8 explore
the influence of aromatic substitutions at position 2;
compound 9 explores alkyl substituents at both
positions 2 and 3; compounds 10–13 explore the
influence of varying aromatic substitutions at posi-
tion 3; and the dimeric oxazolidine 14 was known
from earlier work14 and was studied to assess the
reactivity of the N,N’-methylene bridge. 1H NMR
spectroscopy was used to follow hydrolysis, as it has
the advantage of providing information on the
chemical structure of intermediates as well as their
relative concentrations during the reactions. These
studies should further enhance the understanding of
factors that affect oxazolidine hydrolysis and may be
useful for selecting promoieties for the delivery of
b-amino alcohols or carbonyl containing compounds.
A solution of the oxazolidine compounds (0.035 M)
in the appropriate solvent (1.75 ꢀ 10ꢁ5 mol of the
oxazolidines 4–8 in deuterated solvent, 500 mL) was
prepared and 700 mL of D2O was added. Most
oxazolidine hydrolysis occurred in the first few hours,
so spectra were recorded frequently over this period.
Further spectra were recorded after several days and
again after several weeks.
Synthetic Procedures
3,4S-Dimethyl-2S,5R-diphenyl-oxazolidine
4.15–20
(1R,2S)-(ꢁ)-2-Methylamino-1-phenyl-1-propanol (2.06 g,
12.5 mmol) and freshly distilled benzaldehyde (1.32 g,
12.5 mmol) were dissolved in anhydrous toluene
(30 mL). Anhydrous sodium sulphate (500 mg) was
added and the solution was heated to reflux for 8 h.
The solution was then filtered hot and the toluene
evaporated under reduced pressure to yield a
white solid, which was recrystallized from ethanol
to give 2.68 g (85%) of 4 as fine white crystals. A
similar approach was used for compounds 5–9.
3,4S-Dimethyl-5R-phenyloxazolidine 10.14,16,20–23
A stirred solution of paraformaldehyde (750 mg,
25 mmol) and potassium hydroxide (10 mg) in super
dry methanol (40 mL) was treated with (1R, 2S)-(ꢁ)-
ephedrine (2.08 g, 12.5 mmol). Sodium sulphate
(500 mg) was added and the suspension was heated
to reflux for 6 h. The drying agent was filtered off, the
methanol evaporated under reduced pressure and the
resulting crude oil purified by distillation to afford
1.88 g (85%) of compound 10 as a clear oil.
3,5R-Diphenyl-4S-methyloxazolidine 11. A cataly-
tic amount of potassium hydroxide was added to
a suspension of formaldehyde (50 mg, 1.67 mmol)
in anhydrous methanol (5 mL). When the formalde-
hyde had dissolved, the solution was treated with
(1R,2S)-2-phenylamino-1-phenyl-1-propanol (93 mg,
0.41 mmol) and refluxed for 4 h. The methanol
evaporated under reduced pressure and the resulting
residue was filtered through a plug of silica to remove
decomposition products. Purification was by radial
chromatography of the dark yellow residue eluting
with chloroform/petroleum ether. A similar approach
was used for 12 and 13.
EXPERIMENTAL
NMR Methods
1H NMR data were recorded on a Bruker AM-300
NMR spectrometer operating at 300.13 MHz. 1H NMR
spectra were recorded in d6(CH3)2CO and d4CH3
OH for 4–8 and in d4CH3OH for 9–13. The choice of
solvent was determined by the solubility of the
oxazolidine compounds and the need to avoid an
overlap of the NMR signals of the compounds and
those of the solvent. Chemical shifts are expressed in
ppm relative to tetramethylsilane. Conditions for
Fourier transform measurements were spectral width
4500 Hz, pulse width 5.0ms, acquisition time 2.0 s,
repetition time 2.0 s, number of data points 16,384 and
number of transients 32. A solution of each oxazolidine
compound (1.75 ꢀ 10ꢁ5 mol of the oxazolidines in
Melting points were determined on a Mettler FP2
melting point apparatus and are uncorrected. Infra-
red spectra were recorded on a Hitachi 270-30 grating
spectrophotometer. Mass spectra were recorded on a
Jeol JMX DX-300 double-focusing instrument at
70 eV. Fast atom bombardment (FAB) mass spectra
were recorded on the same instrument with a Jeol-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010
DOI 10.1002/jps

STABILITY STUDIES OF OXAZOLIDINE-BASED COMPOUNDS
3365
FABMS source and argon as incident particle.
Flash chromatography was performed on Merck
Kieselgel 60 F254 (230–400 mesh) silica. Preparative
thin layer chromatography was performed on glass
plates coated with 2 and 4 mm layers of silica gel
(Merck Kieselgel 60 F254). Analytical thin-layer
chromatography (TLC) was performed on Merck
Kieselgel 60 F254 precoated aluminium backed plates.
Radial chromatography was performed using a Model
7924T Harrison Research Chromatotron instrument
on Merck Kieselgel 60 PF254. Elemental analysis was
carried out by the Australian Microanalytical Service,
AMDEL, Melbourne, and the Chemical and Micro
Analytical Service, Pty. Ltd, Melbourne.
the electronic effects of the phenyl substituent at
position 2 on oxazolidine ring stability. The influence
of the placement of functional groups at positions 200
and 400 was also observed.
The stability studies on the 2-phenyl substituted
compounds 4–8 indicated that, under the experi-
mental conditions used, oxazolidine hydrolysis did
not go to completion. In previous experiments under
different conditions,24 conversion of the oxazolidine to
the corresponding b-amino alcohol and aldehyde
component occurred quickly. In the current study,
the addition of a limited amount of water to the
oxazolidine compounds in an organic solvent solution
resulted in slower oxazolidine hydrolysis.
LC/MS was performed by injection of samples (5mL)
onto a Phenomenex Luna C8(2) column (50 mmꢀ
An example of the trends observed in 1H NMR
spectra following addition of D2O to the 2-phenyl
oxazolidine/d6-acetone solutions for the hydrolysis of
4 is shown in Figure 2. However, the 1H NMR spectra
of the major intermediate are not consistent with
the formation of a cationic Schiff base 15a16 that was
proposed based on absorption spectroscopy studies. A
model cationic Schiff base system characterized by
NMR indicated an imine proton shift of d 9–10 ppm
and NCH3 shift of around d 4 ppm25 whereas the
observed major intermediate had shifts of ꢂd 5 and d
2 ppm, respectively. The observed chemical shifts are
instead consistent with a ring opened intermediate
15b that will be in equilibrium with 15a. Rate
determining hydrolysis of the intermediates then
occurs to give the final hydrolysis product 2 and
benzaldehyde. Subsequently, the dimeric species, 16,
may be formed via reaction of 15a with the hydrolysis
product 2, as illustrated in Figure 3. LC/MS analysis
supported the interpretation of the dimer, with a
peak at 459.1 Da, which corresponds in mass to a
potassium adduct of the dimer (418.26 Da).
˚
4.6 mm, 5 mm, 100 A). Samples were eluted at 0.5 mL/
min using a gradient of 5% solvent B (80% MeOH):
95% solvent D (water/0.1% formic acid) to 100%
solvent B over 4 min. Solvent delivery was achieved
using an Agilent 1200 Series HPLC pump system
(Agilent Technologies, Santa Clara, CA). Multimode
electrospray mass spectra were acquired in positive
ion mode on an Agilent 6120 Quadrupole mass
spectrometer (Agilent Technologies). Scan data were
acquired at a fragmentor voltage of 130 V over a mass
range of m/z 100–1000. Instrument control and data
analysis was performed using Agilent MSD Chem-
Station Rev. B.04.01 software (Agilent Technologies).
RESULTS AND DISCUSSION
The primary aim was to use 1H NMR spectroscopy to
obtain information on the factors affecting oxazoli-
dine ring stability. The main interest centered on
the effects of substituents at positions 2 and 3 of the
oxazolidine ring. (ꢁ)-Ephedrine 2 and (ꢁ)-norephe-
drine 3 were chosen as the b-amino alcohol precursors
for several reasons. Firstly, 2 and related compounds
are important as sympathomimetic drugs; secondly,
2 and 3 were readily available and the chemistry
involved in oxazolidine formation, using these start-
ing materials, is straightforward; and finally, ephe-
drine-derived oxazolidines have relatively simple
1H NMR spectra. As oxazolidine hydrolysis occurs,
the appearance of extra peaks in the spectra resulting
from oxazolidine degradation products should be easy
to monitor and identify. A diverse range of oxazolidine
derivatives incorporating substituents of differing
steric and electronic nature at positions 2 and 3 was
synthesized, as shown in Figure 1. The first series of
compounds had varying aromatic substituents at
position 2, namely compounds 4–8. The phenyl-
substituted derivatives incorporated both methoxy
and nitro functions at either positions 200 or 400. A
comparison of compounds 4–8 allowed examination of
To investigate substituent effects on the reaction
kinetics the experiment was repeated in d4CH3OH for
the five oxazolidine compounds 4–8, and intermedi-
ates were observed immediately following addition
of D2O. Tables 1 and 2 show 1H NMR parameters for
the parent oxazolidines 4–8 and the corresponding
hydrolysis products 15b and 17–20 in d6(CH3) CO/
2
D2O and d4CH3OH/D2O. Figure 4 shows the degree of
intermediate formation as a function of time for 4–8.
The least stable oxazolidine derivatives are the nitro-
substituted oxazolidines 7 and 8. Under the condi-
tions examined, the p-nitrophenyl substituted
derivative
8 underwent 89% decomposition to
form the corresponding intermediate, whereas the
o-nitrophenyl derivative 7 experienced 62% decom-
position. This result was expected since the nitro
groups draw electron density from carbon C-2,
encouraging electron donation from the oxazolidine
nitrogen to form the imine bond with resultant ring-
opening. The unsubstituted oxazolidine derivative
4 underwent 34% oxazolidine decomposition to
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010

3366
MOLONEY, ISKANDER, AND CRAIK
Figure 2. Spectra used to monitor the hydrolysis of 4 following addition of D2O to an
oxazolidine/d6(CH3)2CO solution. Signal assignments for 4 are shown in spectrum 1, the
ring opened and cationic Schiff base intermediates 15a and 15b in spectrum 3 and the
dimeric species 16 and benzaldehyde in spectrum 6.
form the corresponding intermediate, whereas the p-
methoxyphenyl substituted derivative 6 underwent
15% decomposition. Surprisingly, the o-methoxyphe-
nyl substituted compound 5 experienced 40% hydro-
lysis to form the intermediate 17. The greater
reactivity of 5, compared with 4 and 6, may be
attributed to intramolecular catalysis by the
ortho-situated methoxy group. A similar enhanced
reactivity was observed for the related oxazo-
lidine, 2S-(o-hydroxyphenyl)-3,4S-dimethyl-5R-phe-
nyloxazolidine relative to the unsubstituted
2-phenyloxazolidine 4.24
The intermediates observed in the hydrolysis of
oxazolidines 4–8 persisted for varying lengths of time,
depending on the substituent in the aromatic group
attached to the imine bond. However, in all cases the
intermediates eventually underwent further reaction
to form a second compound. To investigate the
identity of the second compound, ephedrine (2) was
added to the solution from an experiment carried out
on 4 in d6(CH3) 2CO/D2O after incubation for 4 days.
Examination of spectrum 6 in Figure 2 showed that a
peak for the aldehydic proton of benzaldehyde was
present at d 9.63 ppm at T ¼ 113 h. Clearly, oxazoli-
dine hydrolysis must be progressing to release
ephedrine and benzaldehyde. However, ‘‘spiking’’ of
the reaction mixture with ephedrine at this time
showed that no ephedrine was present in the reaction
mixture. Evidently, any ephedrine formed undergoes
further reaction to form the identified product.
Figure 3. Mechanism for formation of cationic Schiff
base intermediate, ring hydrolysis intermediate and dimer
product.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010
DOI 10.1002/jps

STABILITY STUDIES OF OXAZOLIDINE-BASED COMPOUNDS
3367
Table 1. 1H NMR Spectral Data for Oxazolidines 4–8, in d6(CH3)2CO/D2O (A) and d4CH3OH/D2O (B)
Chemical Shift (ppm)
4
5
6
7
8
Proton
A
B
A
B
A
B
A
B
A
B
CH3-4a, d
NCH3, s
H-4 dq
1.07 (5.7)a 1.19 (6.0) 1.05 (6.0) 1.22 (6.1) 1.06 (5.9) 1.16 (6.0) 1.07 (6.1) 1.19 (6.0) 1.07 (6.0) 1.19 (5.8)
2.48
2.48
2.20
2.72
2.43
2.43
2.66
2.66
2.44
2.44
2.65
2.65
2.49
2.49
2.73
2.73
2.57
2.57
2.24
2.7
H-5, d
4.68 (8.5) 4.84 (9.4) 4.61 (8.8) 4.78 (8.8) 4.64 (8.8) 4.73 (8.6) 4.59 (8.7) 4.60 (9.0) 4.74 (8.8) 4.87 (9.0)
H-2, s
OCH3, s
Arom Protons, m
4.87
5.01
3.72
7.3–7.5
5.36
3.87
6.8–7.5
5.54
3.70
7.0–7.7
4.81
3.85
6.7–7.3
4.99
5.36
5.57
5.16
5.19
7.2–7.5
7.0–7.5
7.1–7.6
7.1–7.7
7.3–8.2
7.3–8.3
aCoupling constants in Hz in parentheses.
Examination of the 1H NMR spectra resulting from
addition of D2O to the d4CH3OH and d6(CH3)2CO/
oxazolidine solutions after several days suggested
that ephedrine 2 was reacting with the cationic Schiff
base intermediate 15a to form the dimeric species 16,
as outlined in Figure 3. Spectrum 6 in Figure 2 shows
that most of the signals corresponding to the
dimeric intermediate are identifiable. A doublet for
the CH3-3a protons at d 0.95 ppm and a singlet for the
NCH3 protons at d 2.14 ppm integrate to six protons.
The H-3 doublet of quartets of 16 at d 2.58 (which
overlaps the corresponding proton H-4 of the parent
oxazolidine 4 at d 2.48 ppm) and the doublet for H-2
of 16 at d 4.34 ppm both integrate to two protons.
The bridging methine proton, designated H-5,
was assigned as a singlet at d 4.92 ppm and the
corresponding proton, H-2 of the parent oxazolidine 4,
as a singlet at d 4.87 ppm. The aromatic signals
integrate correctly to the required number of
protons corresponding to the parent oxazolidine 4,
the b-amino alcohol dimer 16 and benzaldehyde.
Interesting to note is the benzaldehyde signal at d
9.63 ppm, which integrates in a ratio of 1:3 with the
methyl protons of the second formed product, con-
sistent with the formation of a b-amino alcohol dimer.
Detailed 1H NMR data for the proposed b-amino
alcohol dimers 16 and 21–24 resulting from the
hydrolysis of the oxazolidines 4–8 are given in
Table 3.
Figure 4 also shows that the nitro-substituted
intermediates from 7 and 8 are the most stable and
can be detected for up to 5 days. Figure 5 shows the
complementary formation of the b-amino alcohol
dimers, showing that the methoxy substituted
analogs dimerize most rapidly. Thus, it seems that
under the conditions examined, the nitro substituents
stabilize the intermediates or render them less
reactive to ephedrine, and thus less likely to undergo
dimerization. The enhanced electron density at C-5
of the nitro-substituted intermediates relative
to the methoxy analogs probably contributes to a
diminished rate of nucleophilic attack upon this
position in the intermediate compounds.
Under the experimental conditions used here, the
dimeric compounds were stable and present for
several weeks. However, in most cases, spectra
recorded after 4 weeks showed the gradual appear-
ance of peaks corresponding to ephedrine, though
only in small amounts (<5%). Although control
over the extent of initial oxazolidine hydrolysis
Table 2. 1H NMR Spectral Data for the Intermediates 15 and 17–20, in d6(CH3) CO/D2O (A) and d4CH3OH/D2O (B)
2
Chemical Shift (ppm)
15
17
18
19
20
Proton
A
B
A
B
A
B
A
B
A
B
CH3-3a, d
NCH3, s
H-3 dq
0.7 (5.7)a 0.62 (5.7) 0.71 (6.1) 0.67 (6.3) 0.74 (6.7) 0.69 (6.6)
0.68 (6.1) 0.64 (6.0) 0.65 (6.1) 0.66 (6.1)
1.64
2.07
1.56
1.97
1.72
2.07
1.72
2.06
1.67
2.07
1.62
1.98
1.78
1.73
2.04
1.64
2.04
1.62
1.95
—
H-2, d
CH¼N, s
OCH3, s
4.38 (8.3) 4.33 (8.8) 4.30 (9.7) 4.33 (8.6) 4.37 (8.5) 4.36 (9.8) 4.22 (10.5)
4.2 (9.8)
5.25
4.3 (8.9)
4.48
4.31 (9.0)
4.49
4.49
—
4.47
—
5.25
3.24
5.31
3.13
4.47
3.22
4.47
3.09
5.22
—
—
—
—
Arom Protons, m
6.8–7.3
6.78–7.16 6.95–7.5
6.32–7.6
6.52–7.8
6.5–7.7
6.8–7.8
6.7–7.7
6.9–7.2
6.8–7.6
aCoupling constants in Hz in parentheses.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010

3368
MOLONEY, ISKANDER, AND CRAIK
Figure 4. Plot of percentage formation of the hydrolysis intermediates against
time following hydrolysis of compounds 4–8. Times are in units of seconds and are
represented on a log scale. Composite exponential functions were used to fit the data (i.e.,
ꢀ
ꢁ
k
t
t
1
1
2
I ¼ I0
ðeꢁk ꢁ eꢁk Þ where I is the observed intensity at time t, I0 is the initial
k
ꢁk
2
1
intensity (arbitrary units) and k1 and k2 are rate constants (sꢁ1) for the formation and
degradation, respectively, of the intermediate). The values of k1 and k2 for the various
intermediates are as follows: 4: k1 ¼ 0.21 sꢁ1, k2 ¼ 9.0 ꢀ 10ꢁ5 sꢁ1; 5: k1 ¼ 0.025 sꢁ1, k2 ¼
2.0 ꢀ 10ꢁ5 sꢁ1; 6: k1 ¼ 0.02 sꢁ1, k2 ¼ 3.0 ꢀ 10ꢁ5 sꢁ1; 7: k1 ¼ 0.05 sꢁ1, k2 ¼ 5.0 ꢀ 10ꢁ7 sꢁ1; 8:
k1 ¼ 0.05 sꢁ1, k2 ¼ 2.5 ꢀ 10ꢁ7 sꢁ1
.
can be exerted by substituent variation, eventually all
2-phenyloxazolidine derivatives almost completely
hydrolyzed to give mainly the dimeric product with
small amounts of the oxazolidine starting material.
The formation of the dimeric compounds reported
here is consistent with the previous report of a
structurally related dimer 14 incorporating oxazoli-
dine rings instead of the b-amino alcohol groups.24,26
None of the proposed b-amino alcohol dimer deriva-
tives 21–24 have been previously reported. However,
related compounds of the general structure 25,
differing only in the incorporation of a carbonyl
function at the bridging carbon C-5, have been
synthesized.27 The existence of these structurally
related compounds seems to indicate that the
formation of the proposed dimeric compounds is
possible and that they should be reasonably stable.
The additional reactivity of 5 can be attributed to
direct catalytic involvement of the o-methoxy group.
Under the experimental conditions used here,
the dimeric compounds were stable and present for
several weeks. However, in most cases, spectra
Table 3. 1H NMR Spectral Data for the b-Amino Alcohol Dimers 16, and 21–24, in d6(CH3)2CO/D2O (A) and d4CH3OH/
D2O (B)
Chemical Shift (ppm)
16
21
22
23
24
Proton
A
B
A
B
A
B
A
B
A
B
CH3-3a, d
NCH3, s
H-3 dq
0.95 (6.1)a 0.82 (6.5) 0.95 (6.2) 0.85 (6.5) 0.95 (6.1) 0.81 (6.2) 0.96 (6.6)
1.0 (6.7)
2.64
3.27
0.95 (6.7)
2.14
2.66
1.1 (6.8)
2.24
3.13
2.14
2.58
2.44
2.93
2.14
2.50
2.48
3.0
2.14
2.49
2.42
2.92
2.14
2.66
H-2, d
4.34 (9.2) 4.42 (9.0) 4.35 (9.2) 4.42 (8.0) 4.35 (9.0) 4.41 (8.8) 4.45 (9.4) 4.96 (9.2) 4.35 (9.0) 4.52 (8.8)
H-5, s
OCH3, s
Arom Protons, m
4.92
_
7.0–7.81
5.01
_
7.0–7.8
4.29
3.69
6.8–7.6
4.87
3.94
6.9–7.8
4.96
3.78
6.86–7.9
4.67
3.91
7.0–7.6
5.44
_
6.8–7.7
5.57
_
6.7–7.7
5.04
_
7.3–8.2
5.01
_
7.3–8.3
aCoupling constants in Hz in parentheses.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010
DOI 10.1002/jps

STABILITY STUDIES OF OXAZOLIDINE-BASED COMPOUNDS
3369
Figure 5. Plot of percentage formation of the b-amino alcohol dimers 16 and 21–24
against time. Times are in units of seconds and are represented on a log scale. An
t
exponential function was used to fit the data (i.e., I ¼ I0eꢁk where I is the observed
intensity at time t, I0 is the initial intensity (arbitrary units) and k1 is the rate constant
(sꢁ1) for the formation of the dimer). The values of k1 for the various dimers are as
follows: 16: k1 ¼ 1.3 ꢀ 10ꢁ5 sꢁ1; 21: k1 ¼ 6.0 ꢀ 10ꢁ6 sꢁ1; 22: k1 ¼ 4.5 ꢀ 10ꢁ6 sꢁ1; 23:
1
k1 ¼ 4.0 ꢀ 10ꢁ6 sꢁ1; 24: k1 ¼ 5.0 ꢀ 10ꢁ7 sꢁ1
.
recorded after 4 weeks showed the gradual appear-
ance of peaks corresponding to ephedrine, though
only in small amounts (<5%). Although control
over the extent of initial oxazolidine hydrolysis
can be exerted by substituent variation, eventually
all 2-phenyloxazolidine derivatives almost completely
hydrolyzed to give mainly the dimeric product with
small amounts of the oxazolidine starting material.
To compare the effect of alkyl substituents at
position 2 relative to aryl or unsubstituted deriva-
tives, two further oxazolidines 9 and 10 were
examined. An experiment was performed using
similar conditions to those used to study the 2-phenyl
substituted oxazolidines 4–8. Compounds 9 and 10
were found to be more stable to hydrolysis than the 2-
phenyl substituted oxazolidines 4–8. Only one major
decomposition product was evident in both cases.
Minor amounts (<5%) of unidentified decomposition
shown to be unstable compared with the dimeric
compounds 16 and 21–24, respectively, it is reason-
able to assume that the intermediate formed from the
hydrolysis of the oxazolidines 9 and 10 coincides with
related dimeric products. Table 4 shows 1H NMR
parameters for the parent oxazolidines 9 and 10 and
Table 5 shows the 1H NMR data for the corresponding
b-amino alcohol dimers 26 and 27.
Three oxazolidine derivatives 11–13, which incor-
porate 2 or 4-substituted phenyl substituents at
position 3, were tested to compare the effect of an aryl
function relative to an alkyl substituent on the
Table 4. 1H NMR Data for Oxazolidine Derivatives 9 and
10 in d4CH3OH/D2O
Chemical
Shift (ppm)
Coupling (Hz)
1
products were also evident in the H NMR spectra.
Proton
9
10
Multiplicity
9
10
Spiking the reaction mixtures with ephedrine con-
firmed that the newly formed peaks did not corre-
spond to ephedrine. The intermediates resulting from
the hydrolysis of the oxazolidines 9 and 10 were not
stable and underwent rapid hydrolysis to form the
dimeric compounds 26 and 27. The intermediates
were not detected in the 1H NMR spectra, which is not
surprising given that the imine bonds would be
relatively unstable without a phenyl substituent.
Given that the intermediates 15 and 17–20 were
CH3-4a
NCH3
H-4
H-5
H-2
CH3-2a
H-2B
H-2A
Aromatic
Protons
1.12
2.28
2.57
4.55
4.36
1.37
—
—
7.35-
7.45
1.15
2.38
2.66-
4.54
—
—
4.36
4.7
d
s
dq
d
q
d
d
d
m
6.1
—
6.1, 9.0
9.0
5.3
5.3
—
—
—
6.3
—
6.2, 8.7
8.6
—
—
4.0
3.9
—
7.35-
7.45
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010

3370
MOLONEY, ISKANDER, AND CRAIK
Table 5. 1H NMR Spectral Data for the b-Amino Alcohol
Dimers 26 and 27 in d4CH3OH/D2O
The N,N0-methylene bridge of 14 is very reactive
and prone to hydrolysis. Previous studies29,30
have utilized the unstable and reactive nature of
this compound in the synthesis of a number of
N-substituted norephedrine derivatives by reacting
the oxazolidine dimer 14 with various acetate
derivatives. It appears that the basicity of the
oxazolidine nitrogen is the driving force causing the
preferential hydrolysis of 31. The oxazolidine nitro-
gen of 31 is more basic than the secondary amine 30,
firstly because it is a tertiary nitrogen and secondly
because the hydroxymethyl group is an electron
donating substituent, driving electron density toward
the nitrogen. These factors assist in the oxazolidine
nitrogen of 31 by donating electron density to form
the cationic imine intermediate.
Chemical
Shift (ppm)
Coupling (Hz)
Proton
26
27
Multiplicity
26
27
CH3-3a
NCH3
H-3
H-2
H-5
CH3-5a
CH2
Aromatic
Protons
0.89
2.49
3.08
4.5
3.32
1.17
—
0.84
2.44
2.98
4.46
—
d
s
dq
d
q
d
s
m
6.6
—
6.4, 8.7
8.6
5.3
7.0
6.6
—
6.4, 8.8
8.6
—
—
—
—
—
2.86
7.35-
7.45
—
—
7.35-
7.45
oxazolidine nitrogen. Several experiments were
attempted but the progress of the hydrolysis reactions
could not be monitored until solutions became clear
enough for reasonable 1H NMR spectra to be
recorded; in a number of cases this took several
weeks. For example, 700 mL of D2O was added to a
In summary, a number of oxazolidine-based com-
pounds derived from ephedrine and various aldehyde
compounds were synthesized and their stability
studied using 1H NMR spectroscopy. Oxazolidine
hydrolysis did not yield ephedrine and the aldehyde
component as expected, but on addition of
D2O to d4CH3OH or d6(CH3)2CO/oxazolidine solu-
tions the formation of intermediate ring opened
and cationic Schiff base derivatives was detected.
As the hydrolysis reactions proceeded a series of
novel dimeric b-amino alcohol compounds formed
via a reaction between ephedrine and the cationic
Schiff base intermediate. The relative stability of
the oxazolidine compounds was examined using
1H NMR spectroscopy. The resultant structure–
reactivity relationships may be useful when deciding
on oxazolidine derivatives as prodrugs for the
delivery of compounds containing either a b-amino
alcohol and or aldehyde component.
1
0.04 M solution of 11 in d4CH3OH and the H NMR
spectrum recorded after 48 days indicated that 70%
oxazolidine decomposition had occurred. Two decom-
position products were detected, as defined by two
new doublets at d 1.04 and 1.16 ppm assigned to either
the b-amino alcohol 28 or the b-amino alcohol dimer
29. The doublet for the CH3-4a protons of the parent
oxazolidine derivative 11 appears at d 0.89 ppm.
NMR stability studies of the o- and p-methoxy
substituted derivatives 12 and 13 were hampered by
solubility problems and no accurate measure of
oxazolidine hydrolysis could be determined. Future
work on the stability of these compounds will require
the use of alternative solvent systems.
Previously we studied the stability of the dimeric
oxazolidine derivative 14 obtained from the conden-
sation of 3 and formaldehyde.14 An attempt was
made to synthesize 4-methyl-5-phenyloxazolidine
30 by the condensation of 3 with formaldehyde,
as previously reported.27 The reaction appeared to
proceed smoothly and yielded a single product, but on
ACKNOWLEDGMENTS
We thank Dr. Shane Simonsen and Dr. David Wilson
(University of Queensland) for valuable assistance in
the revision of the manuscript. D.J.C. is supported by
a National Health and Medical Research Council
Fellowship.
1
examination of 13C and H NMR spectra it became
apparent that the product was not 30 as expected, but
the dimeric species 14 (structure shown in Fig. 1).
Further investigation revealed that a dimeric species
would be expected if a molar excess of formalde-
hyde26,28 was used in the reaction, compared with the
single molar equivalent for production of the mono-
mer. We previously reported14 that dimer 14 is stable
in chloroform; however, in methanol it spontaneously
hydrolyzes to yield two compounds, the oxazolidine 30
and the N-hydroxymethyl oxazolidine derivative 31,
which, on addition of water, break down at substan-
tially different rates; 30 is essentially stable whereas
31 breaks down rapidly.
REFERENCES
1. Pardasani P, Pardasani RT, Chaturvedi V, Saxena A. 2003.
Stereoisomeric synthesis of oxazolidinone and fused pyrroli-
dine derivatives via azomethine ylides and their antimicrobial
activity. Indian J Chem B 42B:412–415.
2. Taylor W, Hall T, Schreck C. 1996. Synthesis and mosquito
repellent properties of 2,2-dialkyl- and 2-dialkyl-4,4-dimethyl-
N-acetyloxazolidines. Pestic Sci 46:307–314.
3. Chavda S, Coumbarides GS, Dingjan M, Eames J, Flinn A,
Northen J. 2007. Mutual kinetic separation of isotopomers
of pentafluorophenyl 2-phenyl propionate using quasi-
enantiomeric oxazolidinones. Chirality 19:313–320.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010
DOI 10.1002/jps

STABILITY STUDIES OF OXAZOLIDINE-BASED COMPOUNDS
3371
4. O’Brien P, Warren S. 1996. Asymmetric synthesis with diphe-
nylphosphine oxides: Bicyclic aminals and oxazolidines as
chiral auxiliaries. Tetrahedron: Asymmetry 7:3431–3444.
5. Kim S-M, Jun J-G. 2002. Diastereoselective alkylation and
aldol reactions mediated by the d-mannitol-derived oxazoli-
din-2-one as a chiral auxiliary. Synth Comm 32:3851–3863.
6. Armstrong D, He L, Yu T, Lee J, Liu Y-S. 1999. Enantiomeric
impurities in chiral catalysts, auxiliaries, synthons and resolv-
ing agents. Part 2. Tetrahedron: Asymmetry 10:37–60.
7. Agami C, Meynier F, Rizk T. 1987. Enantioselective alkylations
of g-oxo esters. Synth Comm 17:241–250.
8. Walker R, Huang M-J, Leszczynski J. 2001. An investigation of
isomeric differences in hydrolytic rates of oxazolidines using
computational methods. J Mol Struct (THEOCHEM) 549:137–146.
9. Bundgaard H, Rasmussen GJ. 1991. Prodrugs of peptides. 9.
Bioreversible N-alpha-hydroxyalkylation of the peptide bond to
effect protection against carboxypeptidase or other proteolytic
enzymes. Pharm Res 8:313–322.
10. Tomioka H, Tatano Y, Yasumoto K, Shimizu T. 2008. Recent
advances in antituberculous drug development and novel drug
targets. Expert Rev Resp Med 2:455–471.
11. Asbill C, El-Kattan A, Michniak B. 2000. Enhancement of
transdermal drug delivery: Chemical and physical approaches.
Crit Rev Ther Drug Carrier Syst 17:621–658.
12. Neelakatan L. 1971. Asymmetric synthesis. II. Synthesis and
absolute configuration of oxazolidines derived from (ꢁ)-ephedrine
and aromatic aldehydes. J Org Chem 36:2256–2260.
18. Fife T, Hagopian L. 1968. Oxazolidine hydrolysis. Participation
of solvent and buffer in ring opening. J Am Chem Soc 90:1007–
1014.
19. Engel J. 1982. Great Britain Patent. GB2096132-A. 3-Phenyl-
hydroxy-propylamino-1-m-methoxy-phenyl-propan-1-one prodn.-
in improved yield from phenyl-oxazolidine deriv., ketone and
acid. C07C-097/10 ed., U.K.: Degussa Ag(de). p 5.
20. Beckett AH, Jones GR, Hollingsbee DA. 1978. Degradation of
(ꢁ)-ephedrine in solution and during extraction with diethyl
ether. J Pharm Pharmacol 30:15–19.
21. Pfanz H, Kirchner G. 1958. Untersuchungen zum Konstella-
tionsproblem in der Ephedrinreihe. Justus Leibigs Anal Chem
614:149–158.
22. Young-Harvery J. 1985. M Pharm Thesis. ed., Victorian College
of Pharmacy Ltd, Melbourne.
23. Page P, Heaney H, Reignier S, Rassias G. 2003. 1,2-Ami-
nothioethers derived from ephedrine and pseudoephedrine:
Heterobidentate ligands for the palladium-catalysed asym-
metric allylic substitution reaction. Synlett 1:22–28.
24. Johansen M, Bundgaard H. 1983. Prodrugs as drug delivery
systems XXV: Hydrolysis of oxazolidines—A potential new
prodrug type. J Pharm Sci 72:1294–1298.
25. Mayr H, Ofial AR, Wurthwein E-U, Aust NC. 1997. NMR
spectroscopic evidence for the structure of iminium ion pairs.
J Am Chem Soc 119:12727–12733.
26. Bundgaard H, Buur A, Chang S, Lee V. 1986. Prodrugs of
timolol for improved ocular delivery: Synthesis, hydrolysis
kinetics and lipophilicity of various timolol esters. Int J Pharm
33:15.
27. Salas-Coronado R, Sanchez-Ruiz S, Galvez-Ruiz J, Flores-
Parra A. 2003. Nucleophilic reactions at optically active
bis(4-methyl-5-phenyl-1,3-oxazolidin-3-yl)-methane. Hetero-
cycles 60:1111–1122.
13. Baudet M, Gelbcke M, Lagrange G. 1979. Spectres RMN-13C
DE b-Aminoalcools Derives DE L’Ephedrine ET De Leurs
Oxazolidines. Anal Lett 12:641–655.
14. Moloney G, Iskander M, Craik D. 1997. 1H NMR spectroscopic
studies of the stability of an oxazolidine condensation product.
Magn Reson Chem 35:337–341.
15. Kuhnert K, Danks T. 2001. Highly diastereoselective synthesis
of homochiral 1,3-oxazolidines under thermodynamic control
using focused microwave irradiation under solvent free condi-
tions. Green Chem 3:68–70.
16. Soliman S, Abdine H, El-Nenaey E. 1975. Physicochemical
study on some synthesized oxazolidine derivatives: Differentia-
tion of Diastereomers. Aust J Chem 28:49–56.
17. Fife T, Hutchins J. 1980. General-acid-catalyzed ring opening
of oxazolidines. Hydrolysis of 2-[4-(dimethylamino)styryl]-N-
phenyl-1,3-oxazolidine. J Org Chem 45:2099–2104.
28. Takeuchi S, Kochi M, Sakaguchi K, Nakagawa T, Mizutani T.
1978. Benzaldehyde as a carcinostatic principle in figs. Agric
Biol Chem 42:1449.
29. Soliman S, El-Nenaey E. 1978. Comparative study of the con-
densation of (ꢁ)-ephedrine and benzaldehyde. Pharmazie 33:
235.
30. Mu¨ller H, Schuart J. 1975. Zum sterischen Verlauf asymme-
trischer Induktionen. VI. Synthese und Reduktion von rac.
alpha-Aminopropiophenon-Derivaten mit symmetrischer
Harnstoffstruktur. J Prakt Chem 317:431–4434.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 8, AUGUST 2010