Collisionally-induced dissociation mass spectra of organic sulfate anions

Article abstract of DOI:10.1039/b009019k

The collisionally-induced dissociation mass spectra of a variety of organic sulfate ester anions are described and mechanistically rationalized. A cyclic syn-elimination pathway, analogous to that of the Cope elimination, is postulated for the commonly observed formation of bisulfate anion (HSO₄^-, m/z 97). Deuterium labeling experiments confirm that the proton transferred to oxygen during bisulfate elimination normally originates from the C-2 position, although examination of the spectra of polyfunctional steroids reveals that the proton abstracted may originate from more distant sites as well. Adamantyl, phenyl, and vinyl sulfate anions, which do not readily lend themselves to a cyclic syn-elimination, do not give rise to an m/z 97 ion. Instead, these sulfates undergo both heterolytic and homolytic S-O bond cleavages to yield an m/z M - 80 anion, resulting from loss of neutral SO₃, as well as an ion at m/z 80, corresponding to SO₃^-·, respectively. Sulfates that can give rise to a resonance stabilized radical by homolytic C-O bond fission, as exemplified by benzyl and linalyl sulfates, can be recognized by the formation of an m/z 96 (SO₄^-·) ion.

Full text of DOI:10.1039/b009019k

View Article Online / Journal Homepage / Table of Contents for this issue
Collisionally-induced dissociation mass spectra of organic sulfate
anions
Athula B. Attygalle,* Silvina García-Rubio, Jennifer Ta and Jerrold Meinwald
2
Baker Laboratory, Department of Chemistry and Chemical Biology, Cornell University, Ithaca,
New York 14853, USA
Received (in Cambridge, UK) 8th November 2000, Accepted 18th January 2001
First published as an Advance Article on the web 1st March 2001
The collisionally-induced dissociation mass spectra of a variety of organic sulfate ester anions are described
and mechanistically rationalized. A cyclic syn-elimination pathway, analogous to that of the Cope elimination,
Ϫ
is postulated for the commonly observed formation of bisulfate anion (HSO4 , m/z 97). Deuterium labeling
experiments confirm that the proton transferred to oxygen during bisulfate elimination normally originates from
the C-2 position, although examination of the spectra of polyfunctional steroids reveals that the proton abstracted
may originate from more distant sites as well. Adamantyl, phenyl, and vinyl sulfate anions, which do not readily
lend themselves to a cyclic syn-elimination, do not give rise to an m/z 97 ion. Instead, these sulfates undergo both
heterolytic and homolytic S–O bond cleavages to yield an m/z M Ϫ 80 anion, resulting from loss of neutral SO3,
Ϫ
ؒ
as well as an ion at m/z 80, corresponding to SO3 , respectively. Sulfates that can give rise to a resonance stabilized
radical by homolytic C–O bond fission, as exemplified by benzyl and linalyl sulfates, can be recognized by the
Ϫ
ؒ
formation of an m/z 96 (SO4 ) ion.
Introduction
Sulfate esters constitute a widespread and highly diverse group
of natural and non-natural compounds. The resonance-
stabilized nature of sulfate mono-ester anions underlies their
unique chemistry, providing an electrostatic component to
specific interactions without giving rise to significantly basic or
nucleophilic behavior. The industrial importance of anionic
surfactants such as sodium dodecyl sulfate (SDS) has tended to
overshadow the significance of the increasing recognized num-
ber of biologically active natural products that contain a sulfate
moiety. While once believed to occur exclusively as metabolites
of marine organisms1 (i.e. iejimalide D, 1, from the tunicate
Eudistoma cf. rigida,2 and squalamine, 2, from the dogfish
shark3), sulfate esters are now known from many other sources
(e.g. uzarigenin 3-sulfate, 3, from the Ranunculaceae plant
Adonis aleppica4).5 In addition, knowledge of the enzymes that
catalyze sulfate ester synthesis has grown exponentially during
the last decade.6 The formation of sulfate esters7 is an import-
ant step in the elimination8 or bioactivation9 of xenobiotics
and drugs, and in the biotransformations of many hormones
and neurotransmitters.6a,b,c,e Recent studies of sulfated carbo-
hydrates have revealed that many of these compounds populate
extracellular spaces, and mediate a diverse range of events that
contribute to intercellular recognition in both normal and
pathological processes.10
Since encountering the first glycosylated nucleoside sulfate
ester HF-7 (4) in the venom of a funnel-web spider,11 we
became interested in the possibility that similar anionic sulfate
esters might have gone undetected in neurotoxic venoms. The
ideal technique for searching for such compounds is clearly
mass spectrometry. The facile ionization of sulfates obviously
renders them amenable to negative-ion mass spectrometry. Par-
ticularly in conjunction with electrospray ionization (ESI) and
liquid chromatography (LC), mass spectrometry (MS) has
already provided a powerful approach to the characterization
of sulfated compounds in biological fluids.12–16 Nevertheless,
although several interesting features have been noted in the
mass spectra of these sulfate esters,12,14,15,17 little attention has
been paid to the details of their fragmentation mechanisms. A
Ϫ
prominent product ion peak clearly attributable to HSO4 has
been observed at m/z 97 in most spectra of organic sulfates.17
The formation of this anion requires the transfer of a proton to
oxygen, in addition to the cleavage of a C–O bond. The origin
of this transferred proton, however, has not been established,
498
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b009019k

View Article Online
spectrum of [1,1-2H2]hexadecyl sulfate. A comparison of the
CID product-ion spectra derived from m/z 323 and 321, from
the quasi-molecular ions of [1,1-2H2]hexadecyl sulfate and
hexadecyl sulfate, respectively, showed the formation of an m/z
97 ion in both cases (Fig. 1A and 1B). Since an m/z 98 ion was
not observed for [1,1-2H2]hexadecyl sulfate, it is evident that the
C-1 deuterium atoms were not abstracted. On the other hand,
the product ion spectrum of m/z 323 from [2,2-2H2]hexadecyl
sulfate does show a peak at m/z 98 (Fig. 1C), indicating that
a deuteron from the C-2 position is transferred specifically.
Consequently, the fragmentation process is best represented as
a concerted syn-elimination (Scheme 1) similar to the mechan-
Scheme 1
ism for the Cope amine N-oxide elimination and other pyrolytic
eliminations that occur in the gas phase.20
A literature survey of sulfate ester mass spectra supports this
conclusion. The negative-ion ESI mass spectra of sulfates of
2-phenylethanol (5), α-ionol (6), and vomifoliol (7), all of which
Fig. 1 Product ion spectra of molecular anions of hexadecyl sulfate
(A), [1,1-2H2]hexadecyl sulfate (B), and [2,2-2H2]hexadecyl sulfate (C)
(collision energy 45 eV; cone voltage 56 V).
although the hydrogen atom geminal to the sulfate group has
been suggested as its source.12 We here report results obtained
with a number of natural and synthetic sulfates subjected
to electrospray ionization and MS-MS experiments under
collisionally-induced dissociation (CID) conditions. These
results clarify the mechanism of bisulfate† elimination as well
as the other fragmentation patterns characteristic of this
important group of compounds.
possess protons in the C-2 position, show a peak for the bisulfate
anion, whereas those of benzyl sulfate, 2,5-dimethyl-3-oxo-2,3-
dihydrofuran-4-yl sulfate (8),15 the radio-labeled bromasil
metabolite 9,16 and 3β-hydroxy-17-oxoandrost-5-en-19-yl sul-
fate (10),14 lacking C-2 protons, do not. Our experimental
results with 2-methyl-2-pentyl- and 3-methyl-3-pentyl sulfates
showed the formation of the m/z 97 peak as the major frag-
Results and discussion
Compared to high-energy negative-ion CID mass spectra of
organic sulfate esters,18 those spectra generated by low-energy
collisions are relatively simple.19 For example, the negative-ion
CID spectrum of dodecyl sulfate anion shows only the parent
ion at m/z 265 (100%), a prominent product ion at m/z 97 (17%),
Ϫ
ؒ
and a very weak signal at m/z 80 (SO3 , <1%). The m/z 97 peak
is generally understood to represent the HSO4 anion. In
ment, confirming that the protons in the position C-1 are not
Ϫ
Ϫ
involved in the formation of HSO4
.
accord with this assignment, we have observed the correspond-
ing formation of an m/z 99 (H34SO4Ϫ) ion from the 34S-
isotopomer of dodecyl sulfate.
To determine whether a geminal C-1 proton is utilized in the
formation of bisulfate anion, we synthesized and recorded the
An important requirement for facile HSO4Ϫ elimination is the
ability of the resulting neutral product to accommodate a
double bond between C-1 and C-2. Thus, the spectrum of
isopinocamphyl sulfate (11) shows the anticipated m/z 97
peak, whereas that of adamantyl sulfate (12) does not, since the
adamantyl ring system cannot accommodate a double bond in
its rigid skeleton21 (Fig. 2). Instead, an ion at m/z 151 is formed
† The IUPAC name for bisulfate is hydrogen sulfate.
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506
499

View Article Online
Fig. 2 Product ion spectrum of quasi-molecular anion of adamantyl
sulfate 12 (collision energy 66 eV; cone voltage 60 V).
by heterolytic cleavage of an S–O bond resulting in loss of SO3
Ϫ
ؒ
from the m/z 231 ion, while an m/z 80 ion (SO3 ) originates
from a homolytic cleavage of the same bond (Scheme 2).
Scheme 2
This homolytic process is not unique to 12. The low intensity
peak observed at m/z 80 in many sulfate ester spectra (Fig. 1)
originates from homolytic fission of an S–O bond, either directly
from the quasi-molecular anion, or indirectly from the bisulfate
anion itself as depicted in Scheme 3. Evidence for this fragmen-
Fig. 3 Product ion spectra of bisulfate anions derived from hexadecyl
sulfate (m/z 97) (A), [1,1-2H2]hexadecyl sulfate (m/z 97) (B), and
[2,2-2H2]hexadecyl sulfate (m/z 98) (C) (collision energy 45 eV; cone
voltage 56 V).
Scheme 3
tation mechanism was obtained by recording product ion
spectra of bisulfate anions originating from hexadecyl sulfate,
[1,1-2H2]hexadecyl sulfate, and [2,2-2H2]hexadecyl sulfate (Fig.
Ϫ
Scheme 4
3). The formation of an m/z 80 ion from both HSO4 and
DSO4Ϫ by loss of a hydroxyl or deuteroxyl radical, respectively,
confirms this proposal (Scheme 3).
An interesting conformational effect is revealed in the CID
spectra derived from the m/z 467 ions of the epimeric 5α-
cholestan-3β-yl sulfate (14a) and 5α-cholestan-3α-yl sulfate
(15a). Both 14a and 15a show the formation of the bisulfate
anion; however, the intensity of the m/z 97 peak from the 3α-
sulfate is considerably greater than that from the 3β-epimer
(Fig. 4A). The observed differences appear to reflect the greater
relief of conformational strain upon the elimination of the
axial 3α-substituent. Similar intensity differences have been
observed in the fast atom bombardment ionization spectra of
14a and 15a.18 Based on these observations, it is likely that the
stereochemistry of an unknown axial or equatorial steroidal
sulfate can be determined if both epimers are available.
Steroidal sulfates, which are not able to activate the
appropriate steroid receptors, are of particular interest since
they represent an inactive form of transport from which the
hormone can be regenerated in the target tissue.6c We have
therefore paid particular attention to the CID spectra of this
group of biologically important sulfates. The formation of
Ϫ
HSO4 has been observed in the spectra of many steroidal
sulfates.12,19,22,23 For example, the negative ion ESI spectrum
obtained from dehydroepiandrosterone sulfate (13) shows two
major peaks: the most abundant peak at m/z 97, resulting from
the loss of a bisulfate anion from ring A, and the molecular ion
at m/z 367 (Scheme 4).
500
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506

View Article Online
Ϫ
cholestan-3α-ol sulfate (15c) give rise to DSO4 (m/z 98)
(Fig. 4C). In this case, the relative intensities of the m/z 98 ion
are much lower than those recorded for the m/z 97 ion from 14a,
15a, 14b and 15b due to the anticipated primary kinetic isotope
effect.
Not surprisingly, the formation of the bisulfate anion is
observed also in the spectra of compounds with a sulfate group
attached to a five-membered ring. Reported spectra for the sul-
fate of testosterone (16),22 and 3β-hydroxy-5α-androst-5-en-17β-
yl sulfate14 (17) show the expected m/z 97 peak. Mass spectra of
the sulfates of 5α-androst-2-en-17β-ol (18a), 5α-[17α-2H1]-
androst-2-en-17β-ol (18b), and 5α-[16,16-2H2]androst-2-en-
17β-ol (18c) (Fig. 5) confirm that the proton transfer takes
place from C-16.
Neither aryl nor vinyl sulfate esters, in which a C-2 proton is
attached to an sp2 carbon, undergo bisulfate elimination.12,17,24
Thus, the spectrum of phenyl sulfate fails to exhibit a signal at
m/z 97 (Fig. 6A). Instead, a peak at m/z 93 for a loss of SO3, and
Ϫ
ؒ
at m/z 80 for SO3 are both observed (Scheme 5).
Fig. 4 Product ion spectra recorded by repeated experiments under
identical conditions of m/z 467 of 5α-cholestan-3β-yl sulfate (14a,
lower six traces) and 5α-cholestan-3α-yl sulfate (15a, upper six traces)
(A), m/z 468 of 5α-[3α-2H1]cholestan-3β-yl sulfate (14b, lower six
traces) and 5α-[3β-2H1]cholestan-3α-yl sulfate (15b, upper six traces)
(B), and m/z 471 of 5α-[2,2,4,4-2H4]cholestan-3β-yl sulfate (14c, lower
six traces), and 5α-[2,2,4,4-2H4]cholestan-3α-yl sulfate (15c, upper six
traces) (C). See text for experimental details (collision energy 40 eV). All
spectra were normalized to the intensity of the parent ion (not shown).
Scheme 5
Spectra published for 2,5-dimethyl-3-oxo-2,3-dihydrofuran-
4-yl sulfate (8),15 3,7-dioxoandrost-4-en-4-yl sulfate (19),25 and
sulfated tyrosine derivatives of peptides26 suggest that the loss
of SO3 from the quasi-molecular anion is a common feature
observed in the spectra of vinyl and aryl sulfates.
The product ion spectra derived from m/z 468 of 5α-[3β-
2H1]cholestan-3β-yl sulfate (14b) and 5α-[3α-2H1]cholestan-3α-
yl sulfate (15b) show similar intensity differences for the m/z 97
peak (Fig. 4B); moreover, a secondary kinetic isotope effect is
observed, since the relative intensities of the m/z 97 ion
obtained from this experiment are lower than those recorded
from their non-deuteriated analogs. In accordance with
expectation, the product ion spectra of m/z 471 for both 5α-
[2,2,4,4-2H4]cholestan-3β-yl sulfate (14c) and 5α-[2,2,4,4-2H4]-
Estron-3-yl 3-sulfate (20) behaves analogously, losing SO3 to
produce a phenoxyl anion at m/z 269 (21). While the observ-
ation that spectra of aryl sulfates fail to show a m/z 97 peak
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506
501

View Article Online
Fig. 6 Product ion spectra of quasi-molecular anions of phenyl
Fig. 5 Product ion spectra of quasi-molecular ions of sulfates of
5α-androst-2-en-17β-ol, 18a (A) (collision energy 56 eV; cone voltage
39 V), 5α-[17α-2H1]androst-2-en-17β-ol, 18b (B) (collision energy 52 eV;
cone voltage 56 V) (B), and 5α-[16,16-2H2]androst-2-en-17β-ol, 18c (C)
(collision energy 45 eV; cone voltage 56 V).
sulfate (A) (collision energy 26 eV; cone voltage 42 V), benzyl sulfate
(B) (collision energy 23 eV; cone voltage 24 V), and 2-phenylethyl
sulfate (C) (collision energy 26 eV; cone voltage 25 V) anions.
has been attributed to the lack of a geminal proton,12,14 this
behavior appears rather to reflect the difficulty in forming
a benzyne. Interestingly, at higher collision energies, the
m/z 269 (21) ion derived from 20 undergoes a cycloreversion of
the C ring to yield a bicyclic ion at m/z 145 (22), which is
diagnostic for a sulfate group attached to a steroidal aromatic
A ring (Scheme 6) as noted previously for 17β-estradiol
3-sulfate,12 and estriol 3-sulfate.12b
Similarly, a sulfate group attached to the D ring (Scheme 7)
can be recognized by the characteristic ion m/z 177 (24), which
is observed in the spectra of the sulfate derivative of boldenone
(17β-hydroxyandrosta-1,4-dien-3-one)12a (23) and the sulfate of
testosterone22 (16). While the formation of 24 requires cleavage
of ring C, we do not have sufficient data to justify a charge-
remote or a charge-mediated mechanism analogous to that
proposed in Scheme 6.
Scheme 6
These generalizations can now be applied to interpret the
spectrum of 17β-estradiol 3,17-disulfate,12 which shows signals
at m/z 215 for the doubly-charged anion (25a), and m/z 431 for
the singly-charged anion (25b) (Fig. 7A). The product ion
spectrum of 25a shows a loss of SO3 to produce the doubly-
charged ion 26 at m/z 175 (Fig. 7C) (rather than the previously
proposed C6H7O4S)12b whose nature is confirmed by the obser-
vation that its isotopic peaks are separated by 0.5 mass units
(Fig. 7C inset). A precursor ion experiment confirmed that 26
originated from the m/z 215 ion (25a). On the other hand, the
S–O bond can also be cleaved homolytically, to form two
502
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506

View Article Online
Scheme 7
Scheme 8
Scheme 9
an m/z 253 ion (28), as well as a charge-directed cycloreversion
of the C ring, yielding m/z 145 (22) and m/z 177 (24) ions
(Scheme 9, Fig. 7D). A product ion scan of 24 showed that it
fragments further to form the m/z 97 anion, whereas a precur-
sor ion scan confirmed that it can originate from both the m/z
215 (25a) and m/z 350 (27) ions.
The CID spectrum of the singly-charged anion of 17β-
estradiol 3,17-disulfate (25a, m/z 431), showed a loss of SO3
and the elements of H2SO4, suggesting that a non-ionized
monosulfate ester can also undergo a cyclic syn-elimination
(Fig. 7B).
Fig. 7 Negative-ion ESI spectrum of 17β-estradiol 3,17-disulfate
(cone voltage 40 V) (A). Product ion spectrum of 25b, m/z 431 (collision
energy 44 eV; cone voltage 17 V) (B). Product ion spectra of doubly-
charged ions 25a, m/z 215 (collision energy 44 eV; cone voltage 29 V)
(C) and 26, m/z 175 (collision energy 10 eV; cone voltage 46 V) (D). The
inset in C shows an expansion of the m/z 175 region, recorded under
source CID conditions, depicting the resolution of the isotopic peaks.
(Note: the sample used was prepared from estradiol and may contain
monosulfates as impurities.)
The fragmentation of benzyl sulfate is particularly interest-
ing (Fig. 6B). The main fragment ion is observed at m/z 96,
corresponding to the loss of a resonance-stabilized benzyl rad-
Ϫ
Ϫ
ؒ
ؒ
ical to give SO4 . In fact, the formation of SO4 also appears
to be characteristic for allylic sulfates that can yield stabilized
radicals as products, since the major peak in the product ion
spectrum of linalyl sulfate (29, Scheme 10) is also at m/z 96.
Pathways for the formation of minor ions from benzyl sulfate
are rationalized in Scheme 11. For example, the elimination of a
separate anion radicals at m/z 350 (27) and m/z 80, respectively
(Scheme 8).
At higher collision energies, the doubly-charged fragment
ion of m/z 175 (26) undergoes an HSO4Ϫ elimination, producing
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506
503

View Article Online
Scheme 10
Scheme 11
Ϫ
C-1 benzylic proton gives rise to HSO3 (m/z 81) by loss of
a neutral molecule of benzaldehyde.
Recording mass spectra of sulfate esters with hydroxy groups
proximal to the sulfate moiety revealed that alternative mechan-
Ϫ
isms for transfer of a proton to produce HSO4 are possible.
The CID spectra of the m/z 439 ion of cortison-21-yl sulfate
(30) and the m/z 441 ion of hydrocortison-21-yl sulfate (31)
showed the m/z 97 product ion in spite of the absence of a C-2
proton (Fig. 8A and 8C). The proton transfer mechanism
depicted in pathway a (Scheme 12) with the elimination of two
Fig.
8
Negative-ion CID spectra of cortison-21-yl sulfate (30)
recorded from solutions in H2O–CH3CN (A), and D2O–CH3CN (B),
and hydrocortison-21-yl sulfate anion (31) recorded from solutions in
H2O–CH3CN (C), and D2O–CH3CN (D); (collision energy 42 eV; cone
voltage 46 V).
that CID spectra obtained from solutions of both 30 and 31 in
D2O show a signal at m/z 98, indicating that an exchangeable
proton is transferred (Fig. 8B and 8D). These spectra also show
a low intensity peak at m/z 97 in addition to that of the
more intense m/z 98 ion, indicating that pathway a competes
favorably with another mechanism (b) in which a carbon-bound
Scheme 12
neutral molecules, a ketene and a ketone, entails properties of
heterolytic fragmentations described by Grob and Schiess.27
The mechanism we propose is supported by the observation
504
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506

View Article Online
proton is transferred from the C-16 position. Moreover, the
formation of the m/z 81 ion appears to follow a pathway similar
to that illustrated in Scheme 11(a) since this signal was not
affected by the deuterium exchange.
35.5, 35.8, 35.9, 36.1, 36.2, 39.1, 39.5, 40.0, 42.6, 54.3, 56.2,
56.5, 66.2 (t, J 20 Hz).
5ꢀ-[2,2,4,4-2H4]Cholestan-3ꢁ-ol and 5ꢀ-[2,2,4,4-2H4]choles-
tan-3ꢀ-ol. Both these compounds were prepared by the deuter-
ium exchange of 5α-cholestan-3-one with Na–D2O–
dioxane,31,32 followed by the reduction of the carbonyl group.31
The diastereomeric mixture (ratio 10 : 1) of deuteriated
cholestanols was separated by flash chromatography (hexane–
Et2O, 7 : 3). Isomer 3ꢁ: δC (100 MHz, CDCl3) 12.1, 12.3, 18.7,
21.2, 22.6, 22.8, 23.8, 24.2, 28.0, 28.2, 28.7, 31.4 (m), 32.1, 35.4
(br s), 35.5, 35.8, 36.2, 36.8 (br s), 38.1 (m), 39.5, 40.0, 42.6, 44.8
(br s), 54.3, 56.3, 56.5, 71.1 (br s). Isomer 3α: δC (100 MHz,
CDCl3) 11.2, 12.1, 18.7, 20.8, 22.6, 22.8, 23.8, 24.2, 28.0,
28.2, 28.8 (m), 29.6 (br s), 32.0 (br s), 32.1, 35.5, 35.7, 35.8 (m),
36.0, 36.2, 39.0 (br s), 39.5, 40.0, 42.6, 54.3, 56.2, 56.5, 66.4
(br s).
Based on the results of the current study, and on the inform-
ation from previous investigations on the mass spectrometric
fragmentation of sulfate anions,12,14,15,17 the following general-
izations can be made. (1) When protons are available at the C-2
position, and C-1 and C-2 are sp3 hybridized and able to
accommodate a double bond, a proton from the C-2 position is
transferred to the sulfate moiety and the C–O bond is broken to
form a bisulfate anion (m/z 97). (The CID MS of cortison-21-yl
sulfate suggests that more distant protons can also be elimin-
ated.) (2) Vinyl and phenyl sulfates do not undergo the afore-
mentioned bisulfate elimination to form the m/z 97 ion. The
spectra of these anions show a peak for the loss SO3 and a peak
Ϫ
ؒ
at m/z 80 for the SO3 . (3) Benzylic and allylic sulfates, which
can give rise to a resonance-stabilized radical by the homolytic
Ϫ
fission of the C–O bond, yield SO4 anion as the major
fragment observed at m/z 96.
5ꢀ-Cholestan-3ꢀ-ol. 5α-Cholestan-3α-ol was obtained as a
minor stereoisomer by the reduction of cholestan-3-one with
LiAlH4 (ratio 8 : 1). The product was isolated by flash
chromatography.
᝽
An understanding of these collisionally-induced dissociation
patterns should contribute significantly to the recognition and
characterization of organic sulfates.
5ꢀ-Androst-2-en-17ꢁ-ol. Reduction of 5α-androst-2-en-17-
one with LiAlH4 provided the desired β-alcohol as a single
diastereomer. δC (100 MHz, CDCl3) 11.1, 11.7, 20.5, 23.4, 28.6,
30.3, 30.4, 31.4, 34.7, 35.6, 36.7, 39.8, 41.5, 42.9, 51.0, 54.2.
81.9, 125.8, 125.9.
Experimental
General procedures
All reactions were carried out under argon using dry and
freshly distilled solvents unless otherwise noted. [2,2-2H2]-
Hexadecanoic acid was purchased from Cambridge Isotopes.
Sodium dodecyl sulfate, hexadecan-1-ol, 2-methylpentan-2-ol,
5ꢀ-[17-2H1]Androst-2-en-17ꢁ-ol. Reduction of 5α-androst-2-
en-17-one with LiAlD4 yielded the product alcohol as a single
diastereomer. The 13C NMR spectrum of the product showed
that the C-17 signal at δC 81.9 had changed to a triplet (t, J 21
Hz).
3-methylpentan-3-ol,
(Ϫ)-isopinocamphol,
1-adamantol,
dehydroepiandrosteronyl sulfate, dihydrocholesterol, estrone,
estradiol, (Ϫ)-linalool, phenol, 2-phenylethanol, and benzyl
alcohol were Aldrich reagents. 5α-Androst-2-en-17-one,
cortison-21-yl sulfate, and hydrocortison-21-yl sulfate were
5ꢀ-[16,16-2H2]Androst-2-en-17ꢁ-ol. Following the procedure
described above, the title dideuteriated alcohol was obtained
starting from 5α-androst-2-en-17-one. An examination of the
13C NMR spectrum of the product showed the absence of a
C-16 signal, which is observed at δC 30.4 in that of its non-
deuteriated analog.
obtained from Steraloids Inc. (RI). H NMR and 13C NMR
1
spectra were recorded on a Varian Unity 500 instrument operat-
ing at 499.93 MHz and a Varian XL 400 spectrometer operating
at 100.58 MHz, using Me4Si as the internal standard. Chemical
shifts are reported in ppm and J values are given in Hz.
[1,1-2H2]Hexadecan-1-ol.28 [1,1-2H2]Hexadecan-1-ol was
prepared by a described procedure reducing methyl hexa-
decanoate with LiAlD4. δC (100 MHz, CDCl3) 14.1, 22.7, 25.7,
29.4, 29.4, 29.6, 29.6, 29.6, 29.7, 31.9, 32.6, 62.2 (quintet, J 20
Hz).
Sulfation of alcohols
Sulfates of primary and secondary alcohols were prepared
using DMF–SO3 complex (1.1 eq.) in DMF in the presence of
pyridine at 40 ЊC for 1 h.22,33 Tertiary alcohols were sulfated
using pyridine–SO3 complex (2 eq.) in CCl4 at 0 ЊC for 16 h.15
[2,2-2H2]Hexadecan-1-ol.29 [2,2-2H2]Hexadecan-1-ol was pre-
pared by a described procedure reducing [2,2-2H2]hexadecanoic
acid with LiAlH4. δH (500 MHz, CDCl3) 0.88 (3 H, t, J 7.0
Hz), 1.26 (26 H, m) 3.64 (2 H, s). δC (100 MHz, CDCl3) 14.1,
22.7, 25.5, 29.3, 29.4, 29.6, 29.7, 31.9, 31.9 (quintet, J 22 Hz),
62.9.
Mass spectrometry
The CID mass spectra were recorded using a Micromass
(Beverly, MA) Quattro I triple quadrupole mass spectrometer
equipped with an electrospray ion source. The source temper-
ature was held at 85 ЊC. The argon gas pressure in the collision
cell was adjusted to attenuate precursor ion transmission by
50%. Collision energy was optimized for each experiment and
laboratory frame values are given in each corresponding figure
caption. Samples were infused into the ESI source as aceto-
nitrile–water solutions at a rate of 5 µl minϪ1. To compare the
relative intensities of product ions derived from 3α and 3β
epimers of 5α-cholestan-3-yl sulfate, 5α-[3-2H]cholestan-3-yl
sulfate and 5α-[2,2,4,4-2H4]cholestan-3-yl sulfate, CID spectra
of parent anions were recorded under identical conditions
(laboratory frame of reference collision energy: 40.0 eV). For
each recording 10 acquisition scans were co-added by multiple-
channel analysis (MCA). After recording three profiles in this
way, the sample was changed to its epimer and three profiles
were recorded. The overall experiment was repeated immedi-
ately thereafter by switching samples. Thus, for each set of 3α
and 3β epimers, twelve repetitive CID experiments were
5ꢀ-[3ꢀ-2H1]Cholestan-3ꢁ-ol and 5ꢀ-[3ꢁ-2H1]cholestan-3ꢀ-ol.
Treatment of cholestan-3-one with LiAlD4 according to the
procedure described above,28 provided the title cholestanols as a
mixture of two diastereomers (ratio 7 : 1, 67% overall yield),
which was separated by flash chromatography (hexane–Et2O,
7 : 3). The signal for the H-3 proton at δH 3.55 observed in the
1H NMR spectrum of 5α-cholestan-3β-ol30 was absent in that
of the deuteriated β-epimer. Isomer 3ꢁ: δC (100 MHz, CDCl3)
12.0, 12.2, 18.3, 21.2, 22.5, 22.8, 23.8, 24.2, 28.0, 28.2, 28.7,
31.4, 32.0, 35.4, 35.5, 35.8, 36.1, 37.0, 38.1, 39.5, 40.0, 42.5,
1
44.8, 54.3, 56.2, 56.4, 70.8 (t, J 20 Hz). Similarly, in the H
NMR spectrum of the deuteriated β-epimer, a signal for an H-3
proton, observed at δH 4.0 (br s) for its non-deuteriated analog,
was absent. Isomer 3α: δC (100 MHz, CDCl3) 11.2, 12.1, 18.7,
20.8, 22.6, 22.8, 23.8, 24.2, 28.0, 28.2, 28.6, 28.9, 32.0, 32.1,
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506
505

View Article Online
T. Yanagihara, N. Murao, H. Watanabe, J. Goto and T. Niwa,
conducted. For deuterium exchange experiments, 30 and 31
were dissolved in D2O and investigated by negative-ion ESI.
J. Mass Spectrom., 1997, 32, 401; (d) B. L. Ackermann, T. A.
Gillespie, B. T. Regg, K. F. Austin and J. E. Coutant, J. Mass
Spectrom., 1996, 31, 681.
14 S. Murray, N. B. Rendell and G. W. Taylor, J. Chromatogr., 1996,
738, 191.
15 B. Boss, E. Richling, M. Herderich and P. Schreier, Phytochemistry,
1999, 50, 219.
16 R. W. Reiser and A. J. Fogiel, Rapid Commun. Mass Spectrom.,
1994, 8, 252.
17 P. Rudewicz and K. M. Straub, Anal. Chem., 1986, 58, 2928.
18 K. B. Tomer and M. L. Gross, Biomed. Environ. Mass Spectrom.,
1988, 15, 89.
Acknowledgements
This research was supported in part by grant GM53830 from
the National Institutes of Health and award DS0260 from the
Research Corporation. SGR gratefully acknowledges a post-
doctoral fellowship from the Pew Latin American Fellows
Program in the Biomedical Sciences (P0236SC).
19 S. J. Gaskell and M. H. Reilly, Rapid Commun. Mass Spectrom.,
1988, 2, 139.
References
1 (a) X. Fu, M. L. G Ferreira, F. J. Schmitz and M. Kelly, J. Org.
Chem., 1999, 64, 6706; (b) J. A. Kalaitzis, P. de A. Leone, L. Harris,
M. S. Butler, A. Ngo, J. N. A. Hooper and R. J. Quinn, J. Org.
Chem., 1999, 64, 5571; (c) T. Wakimoto, A. Maruyama,
S. Matsunaga and N. Fusetani, Bioorg. Med. Chem. Lett., 1999, 9,
727; (d) F. D. Horgan, B. Sakamoto and P. J. Scheuer. J. Nat. Prod.,
2000, 63, 210; (e) Y. Nakao, S. Matsunaga and N. Fusetani,
Tetrahedron Lett., 1993, 34, 1511.
2 Y. Kikuchi, M. Ishibashi, T. Sasaki and J. Kobayashi, Tetrahedron
Lett., 1991, 32, 797.
3 M. N. Rao, A. E. Shinnar, L. A. Noecker, T. L. Chao, B. Feibush,
B. Snyder, I. Sharkansky, A. Sarkahian, X. Zhang, S. R. Jones,
W. A. Kinney and M. Zasloff, J. Nat. Prod., 2000, 63, 631.
4 G. F. Pauli, U. Matthiesen and F. R. Fronczek, Phytochemistry,
1999, 52, 1075.
20 J. March, Advanced Organic Chemistry: Reactions, Mechanisms and
Structure, Wiley, New York, 1992, p. 896; several examples of base-
catalyzed β-elimination of sulfates have been described (J. H.
Thomas and N. Tudball, Biochem. J., 1970, 119, 57P; J. H. Thomas,
K. S. Dogson and N. Tudball, Biochem. J., 1968, 110, 687; R.
Contestabile, T. Jenn, M. Akhtar, D. Gani and R. A. John,
Biochemistry, 2000, 39, 3091; H. Ueno, J. J. Likos and D. E. Metzler,
Biochemistry, 1982, 18, 4387; R. A. John and P. Fasella,
Biochemistry, 1969, 11, 4477; L. J. Fowler and R. A. John, Biochem.
J., 1972, 130, 569; A. G. Brown, D. F. Corbett, A. J. Eglington and
T. T. Howarth, Tetrahedron, 1983, 39, 2551). To the best of our
knowledge there are no reports of the gas phase cyclic mechanism
that we propose here for the elimination of bisulfate. However,
analogous oxy-Cope rearrangements of deprotonated diallyl ethers
have been studied by Bowie and co-workers: P. C. H. Eichinger and
J. H. Bowie, J. Chem. Soc., Perkin Trans. 2, 1987, 1499; P. C. H.
Eichinger, S. Dua and J. H. Bowie, Int. J. Mass Spectrom. Ion
Processes, 1994, 133, 1.
5 K. Hostettmann and A. Marston, Chemistry and Pharmacology
of Natural Products. Saponins, Cambridge University Press,
Cambridge, 1995.
6 (a) C. N. Falany, FASEB J., 1997, 11, 1; (b) R. M. Weinshilboum,
D. M. Otterness, I. A. Aksoy, T. C. Wood, C. Her and R. B.
Raftogianis, FASEB J., 1997, 11, 3; (c) C. N. Falany, FASEB J.,
1997, 11, 206; (d) K. G. Bowman and C. R. Bertozzi, Chem. Biol.,
1999, 6, R9; (e) J. W. Kehoe and C. R. Bertozzi, Chem. Biol., 2000, 7,
R57.
7 C. D. Klaassen and J. W. Boles. FASEB J., 1997, 11, 404.
8 (a) E. Baumann, Ber. Dtsch. Chem. Ges., 1876, 54; (b) W. B. Jakoby,
R. D. Sekura, E. S. Lyon, C. J. Marcus and J. L. Wang, in Enzymatic
Basis of Detoxification, ed. W. B. Jacoby, Academic Press,
New York, 1980, Vol II, p. 199; (c) G. J. Mulder and W. B. Jakoby, in
Conjugation Reactions in Drug Metabolism: An Integrated Approach,
ed. G. J. Mulder, Taylor and Francis, Philadelphia, 1990, p. 107.
9 (a) C. C. Lai, J. A. Miller, E. C. Miller and A. Liem, Carcinogenesis,
1985, 6, 1037; (b) H. Glatt, FASEB J., 1997, 11, 314; (c) A. E. Buhl,
D. J. Waldon, C. A. Baker and G. A. Johnson, J. Invest. Dermatol.,
1990, 95, 553.
10 K. G. Bowman, S. Hemmerich, S. Bhakta, M. S. Singer, A. Bistrup,
S. D. Rosen and C. R. Bertozzi, Chem. Biol., 1998, 5, 447.
11 J. McCormick, Y. Li, K. McCormick, H. I. Duynstee, A. K. van
Engen, G. A. van der Marel, B. Ganem, J. H. van Boon and
J. Meinwald, J. Am. Chem. Soc., 1999, 121, 5661.
12 (a) L. O. G. Weidolf, E. D. Lee and J. D. Henion, Biomed. Environ.
Mass Spectrom., 1988, 15, 283; (b) H. Zhang and J. Henion,
Anal. Chem., 1999, 71, 3955.
21 (a) J. Bredt, H. Thouet and J. Schmitz, Liebigs Ann. Chem., 1924,
437; (b) P. M. Warner, Chem. Rev., 1989, 89, 1067; (c) W. Maier and
P. v. R. Schleyer, J. Am. Chem. Soc., 1981, 103, 1891.
22 K. Chatman, T. Hollenbeck, L. Hagey, M. Vallee, R. Purdy,
F. Weiss and G. Siuzdak, Anal. Chem., 1999, 71, 2358.
23 S. J. Gaskell, Biomed. Environ. Mass Spectrom., 1988, 15, 99.
24 J. Yinon, L. D. Betowski and R. D. Voyksner, in Applications of
LC-MS in Environmental Chemistry, ed. D. Barceló, Elsevier,
Amsterdam, 1996, p. 187.
25 G. K. Poon, G. M. F. Bisset and P. Mistry, J. Am. Soc. Mass
Spectrom., 1993, 4, 588.
26 P. A. Wabnitz, J. H. Bowie and M. J. Tyler, Rapid Commun. Mass.
Spectrom., 1999, 13, 2498.
27 C. A. Grob and P. W. Schiess, Angew. Chem., Int. Ed. Engl., 1967, 6,
1.
28 (a) H. Taguchi, B. Paal and W. L. F. Armarego, J. Chem. Soc.,
Perkin Trans. 1, 1997, 3, 303; (b) A. Firouzi, F. Atef, A. G. Oertli,
G. D. Stucky and B. F. Chmelka, J. Am. Chem. Soc., 1997, 119, 3596.
29 L. Debrauwer, A. Paris, D. Rao, F. Fournier and J. C. Tabet,
Org. Mass Spectrom., 1992, 27, 709.
30 C. J. Pouchert, Aldrich Library of 13C and 1H FT NMR Spectra,
Aldrich Chemical Company, Milwaukee, 1993, vol. 3, 566C, 567A.
31 J. Karliner, H. Budzikiewicz and C. Djerassi, J. Org. Chem., 1966,
31, 710.
32 R. C. Cookson, D. P. G. Hamon and R. E. Parker, J. Chem. Soc.,
1962, 5014.
33 S. Futaki, T. Taike, T. Yagami, T. Ogawa, T. Akita and T. Kitagawa,
J. Chem. Soc., Perkin Trans. 1, 1990, 1739.
13 (a) K. A. Thomsson, N. G. Karlsson and G. C. Hansson,
J. Chromatogr., A, 1999, 854, 131; (b) A. Tjernberg, P. O. Edlund
and B. Noren, J. Chromatogr., B, 1998, 715, 395; (c) S. Ikegawa,
506
J. Chem. Soc., Perkin Trans. 2, 2001, 498–506