A comparative study of enol aldehyde formation from betamethasone, dexamethasone, beclomethasone and related compounds under acidic and alkaline conditions

Article abstract of DOI:10.1016/j.steroids.2008.09.009

Enol aldehydes are one type of key degradation and metabolic intermediates from a group of corticosteroids containing the 1,3-dihydroxyacetone side chain on their D-rings, such as betamethasone, dexamethasone, beclomethasone, and related compounds. The formation of enol aldehydes from these corticosteroids is via acid-catalyzed β-elimination of water from the side chain, a process known as Mattox rearrangement. It was recently reported by our group that enol aldehydes could also be formed directly from the corresponding 17,21-diesters of these corticosteroids but only under alkaline condition, which was proposed to follow a variation pathway of the original Mattox rearrangement. In this paper, we report the results of a comparative study of enol aldehyde formation from these structurally similar corticosteroids (under the original acidic Mattox condition) and their 17,21-diesters (under the alkaline Mattox variation condition), respectively. In general, enol aldehydes were found to be formed under both conditions; however, the ratios of the E- and Z-isomers of the enol aldehyde were different in each case. The only exception was beclomethasone 17,21-diester under the alkaline condition, where a competing elimination of HCl from the 9,11-positions became predominant. These results can be explained by their structural differences with regard to the Mattox mechanism and its variation pathway. Lastly, solvent effect under acidic condition was studied between an aprotic and a protic solvent and the result suggests that enol aldehyde formation is greatly favored in an aprotic environment.

Full text of DOI:10.1016/j.steroids.2008.09.009

steroids 7 4 ( 2 0 0 9 ) 30–41
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/steroids
A comparative study of enol aldehyde formation from
betamethasone, dexamethasone, beclomethasone and
related compounds under acidic and alkaline conditions
Bin Chen, Min Li∗, Mingxiang Lin, Gilbert Tumambac, Abu Rustum
Global Quality Services – Analytical Sciences, Schering-Plough Corporation, 1011 Morris Avenue, Union, NJ 07083, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Enol aldehydes are one type of key degradation and metabolic intermediates from a group
of corticosteroids containing the 1,3-dihydroxyacetone side chain on their D-rings, such as
betamethasone, dexamethasone, beclomethasone, and related compounds. The formation
of enol aldehydes from these corticosteroids is via acid-catalyzed ␤-elimination of water
from the side chain, a process known as Mattox rearrangement. It was recently reported by
our group that enol aldehydes could also be formed directly from the corresponding 17,21-
diesters of these corticosteroids but only under alkaline condition, which was proposed to
follow a variation pathway of the original Mattox rearrangement. In this paper, we report the
results of a comparative study of enol aldehyde formation from these structurally similar
corticosteroids (under the original acidic Mattox condition) and their 17,21-diesters (under
the alkaline Mattox variation condition), respectively. In general, enol aldehydes were found
to be formed under both conditions; however, the ratios of the E- and Z-isomers of the enol
aldehyde were different in each case. The only exception was beclomethasone 17,21-diester
under the alkaline condition, where a competing elimination of HCl from the 9,11-positions
became predominant. These results can be explained by their structural differences with
regard to the Mattox mechanism and its variation pathway. Lastly, solvent effect under acidic
condition was studied between an aprotic and a protic solvent and the result suggests that
enol aldehyde formation is greatly favored in an aprotic environment.
Received 3 April 2008
Received in revised form
21 August 2008
Accepted 4 September 2008
Published on line 30 September 2008
Keywords:
Corticosteroids
Betamethasone
Dexamethasone
Beclomethasone
Enol aldehyde
© 2008 Elsevier Inc. All rights reserved.
1.
Introduction
dihydroxyacetone side chain, a process known as Mattox
rearrangement (Scheme 1) [4]. During the Mattox process,
Corticosteroids containing the 1,3-dihydroxyacetone side
chain on their D-rings, such as betamethasone, dexametha-
sone, cortisol and related compounds (Fig. 1), are potent
anti-inflammatory agents that have been formulated into var-
ious pharmaceutical dosage forms [1–3]. One of the main
degradation pathways of these corticosteroids in the formu-
lated drug products is the formation of enol aldehyde via
␤-elimination of water from the D-rings containing the 1,3-
using betamethasone (1) as an example, the side chain of
betamethasone D-ring would undergo dehydration to form
the corresponding betamethasone enol aldehyde, which has
E- and Z-isomers (3 and 4), through a presumed enol inter-
mediate (2) [5]. The process is catalyzed by acid which can be
either strong acids such as sulfuric acid [6] and methanolic
HCl [7] or weak acids such as acetic acid and certain Lewis
acids [8,9]. Recently we reported that betamethasone enol
∗
Corresponding author. Tel.: +1 9088203391; fax: +1 9088206359.
E-mail address: min.li@spcorp.com (M. Li).
0039-128X/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2008.09.009

steroids 7 4 ( 2 0 0 9 ) 30–41
31
Fig. 1 – Structures of the molecules mentioned in this study.

32
steroids 7 4 ( 2 0 0 9 ) 30–41
Scheme 1 – Formation of betamethasone enol aldehyde (3 and 4) from betamethasone (1) via Mattox rearrangement under
the usual acidic condition.
Scheme 2 – Formation of betamethasone enol aldehyde (3 and 4) from betamethasone dipropionate (5) via the revised
Mattox rearrangement under alkaline condition.
aldehyde can also be generated from betamethasone dipro-
pionate (5) under alkaline condition, which was proposed to
follow a revised Mattox rearrangement mechanism (Scheme 2)
[10]. In this variation of Mattox rearrangement, both the 17-
and 21-acyl groups are required to effect a concerted elimi-
nation leading to the formation of enol aldehyde; none of the
mono esters (betamethasone 17-propionate and betametha-
sone 21-propionate) produces enol aldehyde under this
condition.
secondary degradants dependent upon the formula and/or
storage condition of a particular drug product [11]. Enol
aldehydes are also important metabolic intermediates; enol
aldehydes and some of its further degradants were reported
as metabolites for some of the corticosteroids mentioned
above [12]. In this paper, we report the results of a com-
parative study on the formation of enol aldehydes from
several structurally similar corticosteroids under acidic con-
dition (the original acidic Mattox condition) and from their
17,21-dipropionate esters under alkaline condition (the alka-
line Mattox variation condition), respectively. In addition,
solvent effect on the formation of enol aldehydes from
Enol aldehydes are one type of the key degradants of
betamethasone and other corticosteroids with similar D-
ring structures, which can further degrade into a number of
Table 1 – Summary of the HPLC and/or LC–MS analysis conditions used to generate data in this study.
No.
1
Instrumentation
HPLC or LC–MS analysis conditionsa
Data generated
Fig. 7A
A Waters HPLC system equipped
with a Model 2695 separation
module and a Model 2996 photo
diode array (PDA) detector
Column: YMC J’sphere ODS-H80, 150 × 4.6 mm, 3 ␮m. Mobile phase A:
acetonitrile/water, 25/75, v/v. Mobile phase B: acetonitrile/water, 70/30,
v/v; Gradient: linear, 0% B to 100% B in 25 min. Flow rate: 1.5 mL/min.
2
3
4
A Thermo Electron Surveyor HPLC
system equipped with a PDA
detector, coupled with a linear ion
trap mass spectrometer
A Thermo Electron Surveyor HPLC
system equipped with a PDA
detector, coupled with a linear ion
trap mass spectrometer
A Waters HPLC system equipped
with a Model 2695 separation
module and a Model 2996 photo
diode array (PDA) detector
Column: Supercosil ABZ-Plus C18, 250 × 4.6 mm, 5 ␮m. Isocratic
elution: 1: 1 acetonitrile: 10 mM ammonium acetate at a flow rate of
2 mL/min.
Fig. 3A and B; Fig. 5A
and B; Fig. 7B.
Column: YMC Basic, 150 × 4.6 mm, 3 ␮m. Mobile phase A:
acetonitrile/water, 35/65, v/v. Mobile phase B: acetonitrile/water,
90/10, v/v; Gradient: 0–5 min, 0% B; 5–20 min, 0% B to 100% B. Flow
rate: 1.5 mL/min.
Column: YMC Hydrosphere C-18, 50 × 4.6 mm, 3 ␮m. Mobile phase A:
water. Mobile phase B: acetonitrile/methanol, 8/25, v/v; Gradient:
0–21 min, 33% B; 21–35 min, 33% B to 63% B. Flow rate: 2.0 mL/min.
Column temp: 35 ◦C.
Fig. 7C and D
Fig. 5D
a
Unless otherwise noted, all gradients are linear and column temp is ambient.

steroids 7 4 ( 2 0 0 9 ) 30–41
33
Fig. 2 – UV spectra of the peaks of interest. (A) Betamethasone enol aldehyde, Z-isomer; (B) Betamethasone enol aldehyde,
E-isomer; (C) Dexamethasone enol aldehyde; (D) Beclomethasone enol aldehyde, E-isomer; (E) Beclomethasone enol
aldehyde, Z-isomer; (F) Betamethasone 9,11-epoxide enol aldehyde (mixture of the E- and Z-isomers).
betamethasone under acidic condition was also investi-
gated.
2.2.
Instrumentation
A variety of HPLC and LC–MS instruments were used under
various conditions for the HPLC and LC-PDA/UV-MSn (n = 1 to
3) analyses performed in this study, which are summarized in
Table 1. Various 1D and 2D NMR experiments were performed
either on a Varian Inova 500 NMR spectrometer operating at a
proton frequency of 500 MHz (carbon frequency of 125 MHz) or
on a Varian Inova 600 NMR spectrometer operating at a pro-
ton frequency of 600 MHz (carbon frequency of 150 MHz with
a 3 mm carbon–proton dual probe.
2.
Experimental
2.1.
Materials
All the steroids and related compounds used as starting mate-
rials for the degradation studies (1, 5, 7 to 10, and 11) were
produced in house with certified purities of greater than 99%;
all the starting materials except for 11 are also commercially
available. Solvents (methanol, acetonitrile, and water) were of
HPLC grade and purchased from Fisher Scientific. Sulfuric acid
(certified ACS plus) and 1 N NaOH (certified) were also supplied
by Fisher Scientific.
2.3.
Structure identification/characterization
LC-PDA/UV-MSn (n = 1 to 3) was the major analytical tool used
in this study. NMR was also used in cases where LC-PDA/UV-

34
steroids 7 4 ( 2 0 0 9 ) 30–41
Table 2 – Reaction conditions used to generate enol aldehyde in this studya.
No.
1
Reaction conditions
Results
Fig. 7A
1/10 volume of conc. sulfuric acid was added to betamethasone solution in 1:1 acetonitrile:water (v/v). The resulting
solution was allowed to stand for ca 20 h.
2
3
4
5
6
7
8
9
1/20 volume of conc. sulfuric acid was added to betamethasone solution in acetonitrile. The resulting solution was
allowed to stand for ca 50 min.
1/10 volume of conc. sulfuric acid was added to betamethasone solution in methanol. The resulting solution was
allowed to stand for ca 30 min and 3 d.
1/20 volume of 1 N NaOH aqueous solution was added to betamethasone dipropionate solution in acetonitrile. The
resulting solution was allowed to stand for ca 20 min.
1/20 volume of conc. sulfuric acid was added to dexamethasone solution in acetonitrile. The resulting solution was
allowed to stand for ca 30 min.
1/20 volume of 1 N NaOH aq solution was added to dexamethasone dipropionate solution in acetonitrile. The
resulting solution was allowed to stand for ca 40 min.
1/20 volume of conc. sulfuric acid was added to beclomethasone solution in acetonitrile. The resulting solution was
allowed to stand for ca 40 min.
1/20 volume of 1 N NaOH aq. solution was added to beclomethasone dipropionate solution in acetonitrile. The
resulting solution was allowed to stand for ca 20 min.
1/20 volume of 1 N NaOH aq. solution was added to beclomethasone 9,11-epoxide dipropionate solution in
acetonitrile. The resulting solution was allowed to stand for ca 15 min.
Fig. 3A
Fig. 7C and D
Fig. 5A
Fig. 3B
Fig. 5B
Fig. 3C
Fig. 5C
Fig. 5D
a
Unless otherwise noted, concentration of the starting material is ca 1 mg/mL. All reactions were carried out at room temp.
MSn could not provide unambiguous structure assignment.
Peak identification of the reaction products was based on pre-
vious studies by our group [10] and comparison of their LC–MSn
fingerprinting data (retention times, molecular weights and
LC–MSn fragmentation patterns) [13] and UV spectra with
those of the corresponding authentic samples.
Enol aldehydes displayed characteristic UV absorption pro-
files (Fig. 2) as compared to the corresponding steroids. All the
Fig. 3 – The study for generation of enol aldehyde in acetonitrile under acidic condition from 1 (A), 7 (B), and 9 (C). The
reactions were carried out at room temp and were monitored by LC–MS. For detailed analysis and reaction conditions, refer
to Tables 1 and 2, respectively.
Peak identities: (1) 1; (2) 4; (3) 3; (4) Betamethasone sulfates; (5) 7; (6) Dexamethasone enol aldehyde; (7) 9; (8) Beclomethasone
enol aldehyde, Z-isomer; (9) Beclomethasone enol aldehyde, E-isomer.

steroids 7 4 ( 2 0 0 9 ) 30–41
35
Fig. 4 – Comparison of the LC–MS2 fragmentation pattern of dexamethasone enol aldehyde generated from the acidic stress
study with those of the Z- and E-isomers of betamethasone enol aldehyde.
enol aldehydes generated in this study exhibit very similar
UV absorption profiles (Figs. 2A through 2E), i.e., two absorp-
tion maxima at approximately 240 and 270 nm, respectively,
except for the enol aldehydes of betamethasone 9,11-epoxide
(12) in which case the two absorption maxima appear to have
merged together (Fig. 2F). The exception observed in the latter
case is apparently due to the fact that the UV absorption max-
imum for betamethasone 9,11-epoxide is at 250 nm, which is
closer to the UV absorption maximum at ∼270 nm for the enol
aldehyde chromophore. On the other hand, betamethasone
(1), dexamethasone (7), and beclomethasone (9) all have a UV
absorption maximum at ∼240 nm.
first part of the current study, we compared the formation of
enol aldehydes from dexamethasone (7) and beclomethasone
(9), respectively, with that of betamethasone (1) under acidic
condition where the original Mattox mechanism should be fol-
lowed. In the second part of the current study, we compared
the formation of enol aldehydes from the 17,21-dipropionate
esters of dexamethasone and beclomethasone (8 and 10),
respectively, with that of betamethasone 17,21-dipropionate
(5) under alkaline condition where the Mattox variation mech-
anism should be operative as we reported recently [10].
3.1.
Comparative study of enol aldehyde formation
from betamethasone (1), dexamethasone (7), and
beclomethasone (9) under acidic condition
2.4.
Reaction conditions
The structures of all the relevant compounds are shown in
Fig. 1. Of these compounds, dexamethasone (7) is the epimer
of betamethasone (1). The only structural difference between
the two molecules lies in the orientation of the methyl group
bonded to C-16 in the D-ring (␤-position for betamethasone
and ␣-position for dexamethasone). The only structural dif-
ference between betamethasone and beclomethasone (9) is
that 9-fluoro in betamethasone is replaced by 9-chloro in
beclomethasone.
All reaction conditions used to generate enol aldehydes from
the compounds of interest in this study are listed in Table 2.
3.
Results and discussion
There appear to be very few detailed studies in the literature
regarding the formation of enol aldehydes from betametha-
sone and its closely related analogs such as dexamethasone
and beclomethasone. Other than the study results published
by our group recently [10,14], the only in-depth study regard-
ing the formation of betamethasone enol aldehydes (3 and 4)
under acidic condition appears to be the degradation study
of betamethasone reported by Hidaka et al. [15]. Although the
acidic conditions used in the two studies were somewhat dif-
ferent, both 3 and 4 were isolated by the two groups. In the
3.1.1. Dexamethasone (7) vs. betamethasone (1) under
acidic condition
Formation of enol aldehydes from betamethasone (1) under
acidic condition is known to proceed via the Mattox mecha-
nism [4]. In this study, when dexamethasone (7) was treated
under the same acidic condition as that used for betametha-

36
steroids 7 4 ( 2 0 0 9 ) 30–41
Table 3 – 1H and 13C NMR signal assignment for the Z-isomer of dexamethasone enol aldehyde (13a)a
.
No.b
Proton ı (ppm)
Carbon ı (ppm)
1
2
3
7.30, d, 10.1 Hz
152.83
129.05
185.32
124.23
6.22, dd, 10.1, 1.9 Hz
6.01, dd, 1.9, 1.2 Hz
4
5
166.94
6
2.34, 2.65, m, m
30.20
7
8
9
1.41, 1.86, m, m
27.30
2.48, m, 2JHF = 34.4 Hz
32.71, 2JCF = 19.2 Hz
101.65 1JCF = 174.6 Hz
48.02, 2JCF = 22.4 Hz
71.20, 2JCF = 37.9 Hz
41.15
10
11
12
13
14
15
16
17
18
19
20
21
22
4.09, d, 3JHF = 12.2 Hz, 5.44 (OH), d, 4.0 Hz
1.68, 2.62, dt 13.9, 2.2, 3.8 Hz; dd 13.9, 2.2 Hz
45.55
44.18, 3JCF = 1.5 Hz
33.24
31.52
152.69
17.38
22.80, 3JCF = 5.3 Hz
143.60
187.37
26.54
1.72, m
1.34, 1.72, m, m
3.32, m
1.17, s
1.51, s
7.81 (OH), s
9.59, s
1.12, d 7.2 Hz
a
1H and 13C NMR spectra were obtained on a Varian Inova 600 spectrometer operating at a proton frequency of 600 MHz (carbon frequency of
150 MHz) with a 3 mm carbon–proton dual probe. All spectra were taken at 25 ◦C on DMSO-d6 solutions of the compounds. The 2D experiments,
HSQC, HSQCTOXY, HMBC and NOESY were used to establish the connectivity of the proton and carbon nuclei. 2D NOESY experiments were
used to establish the stereochemistry of the groups around the C-17–C-20 double bond.
b
The numbering of the steroid rings follows the usual convention.
to C-16 and the ␤-oriented methyl group attached to C-
13. Consequently, both the Z-isomer (4) and E-isomer (3)
of betamethasone enol aldehyde are formed in comparable
yields. In the case of dexamethasone (7), since the methyl
group attached to C-16 is now ␣-oriented (downward), the eno-
lized side chain would preferably take an orientation that is
less sterically hindered, which is the orientation that should
lead to the formation of the Z-isomer of the enol aldehyde as
shown in Scheme 3. Therefore, the Z-isomer of dexametha-
sone enol aldehyde (13a) was generated as the predominant
product from 7 under acidic condition.
sone (Table 2), dexamethasone enol aldehyde was generated
(Fig. 3B). The dexamethasone enol aldehyde (13) formed was
found to be mostly the Z-isomer based on the comparison of
its LC–MS2 fragmentation pattern [13] with those of the Z- and
E-isomers of betamethasone, respectively (Fig. 4). Upon isola-
tion of the dexamethasone enol aldehyde peak, NMR analysis
(Table 3) showed that the Z-isomer was formed predominantly
(∼90%), while the E-isomer was formed as a minor product
(∼10%) [16]. This product distribution is quite different from
the Mattox rearrangement of 1 in which the ratio of the Z-
isomer (4) vs. the E-isomer (3) formed from 1 is approximately
40% vs. 60% (Fig. 3A). The high preference in the formation
of the Z-isomer of dexamethasone enol aldehyde (13a) over
the E-isomer (13b) can be well explained stereochemically
(Scheme 3). The enolized side chain [20,21-dihydroxy-20(21)-
en] attached to the D-ring could have two orientations which
would lead to the formation of the Z-isomer and E-isomer of
the corresponding enol aldehydes, respectively (Scheme 3). In
the case of betamethasone (1), the two orientations of the
enolized side chain appear to encounter similar steric hin-
drance from the ␤-oriented (upward) methyl group attached
3.1.2. Beclomethasone (9) vs. betamethasone (1) under
acidic condition
Beclomethasone (9) is structurally identical to betametha-
sone (1) except that 9-fluoro in 1 is replaced by 9-chloro in 9.
From the above stereochemical discussion, it can be predicted
that both the Z-isomer and the E-isomer of beclomethasone
enol aldehyde should be generated in comparable yields from
beclomethasone (9) under acidic condition. Indeed, after 9 was
stressed under the same acidic condition as used for stress-

steroids 7 4 ( 2 0 0 9 ) 30–41
37
Scheme 3 – Proposed stereochemical explanation for the formation of enol aldehydes under acidic condition:
betamethasone (1) vs. dexamethasone (7) (only the D-ring is shown for each molecule). In the case of betamethasone, the
two conformations (A and B) of the presumed intermediate (after enolization) leading to the formation of the enol aldehyde
E- and Z-isomers appear comparable with regard to steric hindrance. Therefore, both the E- and Z-isomers of
betamethasone enol aldehyde (3 and 4) were generated in comparable yields from betamethasone under acidic condition.
Whereas in the case of dexamethasone, the conformation (C) of the intermediate leading to the formation of enol aldehyde
Z-isomer is highly favored over the other conformation (D) due to the much less steric hindrance encountered by
conformation (C). Therefore, the Z-isomer of dexamethoasone enol aldehyde (13a) was generated predominantly. Similar
explanations can also be applied to the generation of enol aldehyde from betamethasone 17,21-dipropionate (5) and its
dexamethasone counterpart (8) under alkaline condition.

38
steroids 7 4 ( 2 0 0 9 ) 30–41
Fig. 5 – The study for generation of enol aldehyde in acetonitrile under alkaline condition from 5 (A), 8 (B), 10 (C), and 11 (D).
The reactions were carried out at room temp and were monitored by LC–MS; the chromatograms were obtained under
different conditions. For detailed analysis and reaction conditions, refer to Tables 1 and 2, respectively. Peak identities: (1) 5;
(2) 4; (3) 3; (4) 8; (5) Dexamethasone enol aldehyde (13); (6) Dexamethasone 21-monopropionate; (7) 10; (8) 11; (9) 12 (both Z-
and E-isomers).
while the 17-propionyl provides a good leaving group for the
17-oxygen/hydroxyl. Approximately 20–30% of 5 was con-
verted to both isomers of betamethasone enol aldehyde (3
and 4) within 20 min at room temperature (Fig. 5A). Inter-
estingly, the ratios of the two betamethasone enol aldehyde
isomers formed under the alkaline condition are different
from that under the acidic condition. While the E- and Z- iso-
mers of betamethasone enol aldehyde (3 and 4) were formed
in comparable yields from betamethasone (1) under the acidic
condition (Fig. 3A), the Z-isomer of betamethasone enol alde-
hyde (4) was more favored from 5 under the alkaline condition,
with the ratio of the Z-isomer (4) vs. the E-isomer (3) being ∼3:1
(Fig. 5A).
ing betamethasone (1) and dexamethasone (7), the Z- and
E-isomers of beclomethasone enol aldehyde were generated
in a ratio very similar to what has been observed for 1 under
the same reaction condition (Fig. 3C and A).
3.2.
Comparative study of enol aldehyde formation
from betamethasone dipropionate (5), dexamethasone
dipropionate (8), and beclomethasone dipropionate (10)
under alkaline condition
It has been recently reported by our group that betametha-
sone enol aldehyde can be readily generated directly from
betamethasone 17,21-dipropionate (5) at room temperature
under alkaline condition [10]. In this revised version of Mat-
tox rearrangement, the process appears to be concerted: the
21-propionyl apparently activates the molecule, which makes
it susceptible to attack by a nucleophile at the carbonyl group,
3.2.1. Dexamethasone dipropionate (8) vs. betamethasone
dipropionate (5) under alkaline condition
It appears that the same stereochemical consideration illus-
trated in Scheme 3 can also be applied to the generation of

steroids 7 4 ( 2 0 0 9 ) 30–41
39
Fig. 6 – 1H NMR spectrum of betamethasone 9,11-epoxide enol aldehyde (both Z- and E-isomers).
the enol aldehydes from 8 and 5 under the alkaline condition.
Not surprisingly, the Z-isomer of dexamethasone enol alde-
hyde (13a) was generated from 8 (Fig. 5B) predominantly, while
the E-isomer of betamethasone enol aldehyde (3) formed
accounted for a more significant portion in the two isomers
of betamethasone enol aldehyde (3 and 4) generated from 5
(Fig. 5A).
results confirmed that the single peak contains both the Z-
and E-isomers in a ratio of ∼55:45 as shown in Fig. 6 [17].
3.3.
Solvent effect on the reaction rate of enol aldehyde
formation from betamethasone under acidic condition
The solvent effect on enol aldehyde formation under acidic
condition was also investigated (Fig. 7). The reaction rates
varied significantly in 1:1 acetonitrile:water, acetonitrile, and
methanol. At room temperature, the reaction was very slow
when betamethasone was dissolved in an acidified 1:1 (v/v)
mixture of acetonitrile and water. Less than 5% of betametha-
sone enol aldehyde was generated after 20 h of reaction
(Fig. 7A). When betamethasone was dissolved in an acidified
acetonitrile solution, betamethasone enol aldehyde was gen-
erated at a much faster rate even though a much less amount
of sulfuric acid was added to the reaction solution. The yield of
betamethasone enol aldehyde reached ∼15% within an hour
at room temperature (Fig. 7B). On the other hand, when ace-
tonitrile was replaced with methanol, there was essentially no
reaction at room temperature, even after 3 days (Fig. 7C and
D). These results suggest that an aprotic environment favors
the formation of betamethasone enol aldehyde.
In summary, the comparative study of enol alde-
hyde formation from betamethasone, dexamethasone, and
beclomethasone under the original acidic Mattox rearrange-
ment condition and from their 17,21-diesters under the
alkaline variation of Mattox condition has shown that the
two mechanisms (the original Mattox and its variation)
should be generally applicable for corticosteroids contain-
ing the 1,3-dihydroxyacetone side chain on their D-rings and
their 17,21-diesters, respectively. Certain deviation and vari-
ation from the mechanisms can be explained by the subtle
structural differences between these steroids. A clear under-
standing of the degradation pathways of these steroids and
3.2.2. Beclomethasone dipropionate (10) vs.
betamethasone dipropionate (5) under alkaline condition
When beclomethasone dipropionate (10) was treated under
the same alkaline condition, nonetheless, no beclomethasone
enol aldehyde was formed. Instead, only a trace amount of
betamethasone 9,11-epoxide enol aldehyde (12) appeared to
be produced (Peak 9 in Fig. 5C). The initial structure assign-
ment was based on the molecular weight (Mw 354) expected
of 12 and its characteristic UV profile. The predominant
degradation product of 10 under the alkaline condition was
betamethasone 9,11-epoxide 17,21-dipropionate (11, Peak 8
in Fig. 5C), which was formed from the 9,11-elimination of
HCl from 10. It became evident that 11 would further degrade
via the variation of the Mattox rearrangement to produce 12.
Therefore, when a pure sample of 11 was subjected to the
same alkaline condition, 12 was formed in ∼13% yield within
15 min (Fig. 5D). Based on the stereochemistry of betametha-
sone 9,11-epoxide with regard to the ␤-orientation of its two
methyl groups attached to the C-13 and C-16 positions, respec-
tively (refer to Scheme 3 for similar stereochemistry), both
the Z- and E-isomers of the enol aldehyde are expected to
form. Nonetheless, only a single peak was observed which
could be attributed to 12 based on its molecular weight and
UV profile (Fig. 5D). One possibility would be that the Z- and
E-isomers of 12 co-elute under the LC–MS condition used in
this study. Thus, the enol aldehyde peak was isolated via semi-
preparative HPLC and sent for NMR analysis. Indeed, the NMR

40
steroids 7 4 ( 2 0 0 9 ) 30–41
Fig. 7 – The study of solvent effect on generation of betamethasone enol aldehyde (3 and 4) from betamethasone (1) under
acidic condition at room temp. (A) in 1:1 acetonitrile:water, 20 h; (B) in acetonitrile, 30 min; (C) in methanol, 30 min; (D) in
methanol, 3 days. Chromatograms were obtained under different conditions. For detailed analysis and reaction conditions,
refer to Tables 1 and 2, respectively. Peak identities: (1) Betamethasone (1); (2) Betamethasone enol aldehyde, Z-isomer (4); (3)
Betamethasone enol aldehyde, E-isomer (3); (4) Betamethasone sulfates.
[2] Hogg JA. Steroids, the steroid community, and Upjohn in
perspective: a profile of innovation. Steroids 1992;57:593–616.
[3] Hirschmann R. The cortisone era: aspects of its impact.
Some contributions of the Merck Laboratories. Steroids
1992;57:579–92.
their 17,21-diesters is very important in maintaining and
improving the quality of the drug products formulated with
these drug substances.
[4] Mattox VR. Steroids derived from bile acids. XV. The
formation of a glyoxal side chain at C-17 from steroids with
dihydroxyacetone and ꢀ16-keto side chains. J Am Chem Soc
1952;74:4340–7.
[5] Weiss G, Monder C, Bradlow L. 17-Deoxygenation: a new
pathway of Cotisol metabolism, isolation of
Acknowledgment
We would like to thank Dr. T.-M. Chan’s group at Schering-
Plough Research Institute for performing the NMR experi-
ments.
17-deoxycortolonic acids. J Clin Endocrinol Metab
1976;43:696–9.
r e f e r e n c e s
[6] Lewbart ML, Mattox VR. The mechanism of the Porter-Silber
reaction. I. Rearrangement of the dihydroxyacetone group of
steroids. J Org Chem 1964;29:513–21.
[7] You Z, Khalil MA, Ko DH, Lee HJ. Suppression of the Mattox
rearrangement of 16␣-cyanoprednisolones in acid: synthesis
[1] Herzog H, Oliveto EP. A history of significant steroid
discoveries and developments originating at the Schering
Corporation (USA) since 1948. Steroids 1992;57:617–23.

steroids 7 4 ( 2 0 0 9 ) 30–41
41
of methyl 16␣-prednisolonecarboxylates. Tetrahedron Lett
1995;36:3303–6.
and from betamethasone 17,21-dipropionate (under
alkaline condition) were reported by us previously in
Ref. 10.
[8] Herzog HL, Gentles MJ, Marshall H, Hershberg EB. Weak
acid-catelyzed rearrangement of the dihydroxyacetone side
chain in steroids. J Am Chem Soc 1961;83:4073–6.
[9] Lewbart ML, Monder C, Boyko WJ, Singer CJ, Iohan F.
Synthesis, isolation, and characterization of the Cis and
Trans isomers of steroidal
[15] Hidaka T, Huruumi S, Tamaki S, Shiraishi M, Ninato H.
Studies on betamethasone: behavior of betamethasone in
acid or alkaline medium, photolysis, and oxidation.
Yakugaku Zasshi 1980;100:72–80.
[16] Dexamethasone enol aldehyde was reported previously by
Hofmeister et al. However, it was not specified as to which
isomer was isolated and characterized: Hofmeister H,
Laurent H, Wiechert R Eine neue synthese von
20-hydroxy-17(20)-en-21-aldehydes. J Org Chem
1989;54:1332–8.
[10] Li M, Chen B, Lin M, Chan TM, Rustum A. A variation of
Mattox rearrangement mechanism under alkaline
condition. Tetrahedron Lett 2007;48:3901–5.
[11] The knowledge has been obtained through the studies of
in-house steroid family drug products in our group. Some of
the results from the studies are being published.
[12] Singer CJ, Iohan F, Monder C. 11ß,
11␤-hydroxy-ꢀ16-corticosteroiden. Chem. Ber. 1973; 106:
2263–2267. In the current study, the NMR data indicate the
predominant formation of the Z-isomer and its 1H and 13
NMR signals are assigned (Table 3) with the help of relevant
2D NMR data. On the other hand, the NMR signals of the
E-isomer could not be assigned in their entirety due to its
low amount and signal overlap with the signals of the
Z-isomer. The determination of the ratio between the
two isomers is based on the peak area of the Z- aldehyde
signal (s, 9.59 ppm) vs. that of the E-aldehyde signal.
(s, 9.68 ppm).
C
20-Dihydroxy-3-oxopregna-4, 17(20)-dien-21-al: an
intermediate in the biological 17-dehydroxylation of
Cortisol. Endocrinol 1986;119:1356–61.
[13] We have found that LC–MSn (n is usually 2 to 3)
fragmentation patterns can be used as a very effective
molecular fingerprinting tool for structural elucidation
during the identification of steroid impurities in our
laboratory. The details of the work will be published
elsewhere.
[17] The structures of both isomers were confirmed by 1H NMR,
13C NMR, 2D HSQC, and 1D NOE experiments. The
determination of the ratio between the two isomers is based
on the peak area of the Z- aldehyde signal (s, 9.62 ppm) vs.
that of the E-aldehyde signal. (s, 9.70 ppm).
[14] Some of the early results for the formation of enol
aldehydes from betamethasone (under acidic condition)