Mass spectrometric studies of trimethylsilylpantothenic acid and related substances

Article abstract of DOI:10.1002/jms.157

The characteristic fragment of trimethylsilylated pantothenic acid (TMS-PA) at m/z 291 upon electron ionization was shown to originate from the molecular ion by a McLafferty rearrangement instead of by ejection of 1,1,3,3-tetramethyl-1,3-disilacyclobutane. The verification consisted of labelling experiments and high-resolution mass spectrometry of the fragment and studies on its isotopic distribution. The remaining fragmentation pathways of TMS-PA were clarified by B/E-linked scans and collision-induced dissociation. Copyright

Full text of DOI:10.1002/jms.157

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2001; 36: 555–562
Mass spectrometric studies of trimethylsilylpantothenic
acid and related substances
Michael Rychlik∗
Institut fu¨ r Lebensmittelchemie der Technischen Universita¨ t Mu¨ nchen, Lichtenbergstr. 4, D-85748 Garching, Germany
Received 15 December 2000; Revised 2 February 2001; Accepted 15 February 2001
The characteristic fragment of trimethylsilylated pantothenic acid (TMS-PA) at m/z 291 upon electron
ionization was shown to originate from the molecular ion by a McLafferty rearrangement instead of by
ejection of 1,1,3,3-tetramethyl-1,3-disilacyclobutane. The verification consisted of labelling experiments
and high-resolution mass spectrometry of the fragment and studies on its isotopic distribution. The
remaining fragmentation pathways of TMS-PA were clarified by B/E-linked scans and collision-induced
dissociation. Copyright  2001 John Wiley & Sons, Ltd.
KEYWORDS: pantothenic acid; trimethylsilylpantothenic acid; electron ionization; collision-induced dissociation
(FinniganMAT, Bremen, Germany)oranMD800quadrupole
mass spectrometer (Fisons Instruments, Manchester, UK).
The ionization energy was 70 eV.
INTRODUCTION
The vitamin (R)-pantothenic acid (PA) is involved in
steroid, fatty acid and phosphatide metabolism as part of
physiologically active conjugates such as coenzyme A (CoA)
and acyl carrier protein (ACP).
Multiple-stage mass spectrometry upon CID was per-
formed on a Saturn 2000 ion trap mass spectrometer
(Varian, Walnut Creek, CA, USA) in the non-resonant
mode. The further settings of the CID processes were as
follows: r.f. storage level, m/z 48; CID time, 20 ms; ion-
ization r.f., m/z 48; ejection amplitude, 20 V; low digital
to analog conversion (DAC) offset, 6 DACs; high DAC
offset, 2 DACs; isolation time, 5 ms; and broadband ampli-
tude, 30 V. The optimum CID voltage for dissociation of
the isolated precursor ion was determined by using the
MS/MS Toolkit software and is detailed in Tables 3–6.
The MS/MS Toolkit was also applied to perform triple-
stage mass spectrometry in which the voltage for the
second CID was adopted from the preceding MS/MS
experiments.
The determination of PA is commonly performed by
microbiological or chromatographic assays which are either
tedious or lack sensitivity and specificity. Therefore, enzyme-
linked immunosorbent assays (ELISAs) were established
as fast and sensitive alternatives.1 Recently, a further
improvement was achieved by the development of a stable
isotope dilution assay (SIDA) using an isotopomer of PA
as the internal standard and detecting the isotopomers
by gas chromatography/mass spectrometry (GC/MS) after
trimethylsilylation.2
The derivative of PA was found to be tris(trimethylsilyl)-
PA showing a characteristic fragment [M ꢀ 144]C upon
electron ionization (EI). Quantitation of the isotopomers was
based on the detection of this ion, but its composition and
formation are controversial. Whereas Prosser and Sheppard3
attributed the generation of this ion to a McLafferty
rearrangement, Banno et al.4 postulated the loss of two
trimethylsilyl group accompanied by the migration of two
hydrogens. Hence the purpose of this study was to resolve
this contradiction.
Mass spectrometry of metastable ions and in the
high-resolution mode
Spectra were recorded by means of a MAT 95 S double-
focusing mass spectrometer (Finnigan MAT). High-resolu-
tion mass spectrometry (HRMS) of the McLafferty ion of
TMS-PA was performed in the multiple ion electroscan
detection mode using perfluorokerosene for calibration;
the lock mass was m/z 280.9824, the calibration mass was
m/z 292.9824 and the calculated resolution was 14 400.
HRMS of other ions were recorded in the full-scan mode.
Product ions of metastable ion transitions were detected
by B/E-linked scans.
EXPERIMENTAL
Single-stage mass spectrometry and multiple-stage
mass spectrometry upon collision-induced
dissociation (CID)
Single-stage mass spectra in the EI mode were recorded by
means of either a MAT 95 S sector-field mass spectrometer
Gas chromatography/mass spectrometry (GC/MS)
High-resolution (HR) GC/MS was performed by means of
a model 5300 gas chromatograph (Carlo Erba, Hofheim,
Germany) or a model 3800 gas chromatograph (Varian)
ŁCorrespondence to: M. Rychlik, Institut fu¨r Lebensmittelchemie
der Technischen Universita¨t Mu¨nchen, Lichtenbergstr. 4, D-85748
Garching, Germany. E-mail: rychlik@leb.chemie.tu-muenchen.de
DOI: 10.1002/jms.157
Copyright  2001 John Wiley & Sons, Ltd.

556
M. Rychlik
using a DB-5 capillary column (30 m ð 0.32 mm fused-silica
capillary, film thickness of the stationary phase df D 0.25 µm)
(Fisons Instruments, Mainz, Germany) which was coupled
to the mass spectrometers mentioned above.
from Fluka (Buchs, Switzerland), acetone, calcium oxide,
dichloromethane, diethylamine, ethyl acetate, hydrochlo-
ric acid, methanol, pyridine and Na2SO4 from Merck
(Darmstadt, Germany), sodium 3-(trimethylsilyl)[2,2,3,3-
d4]propionate from Isotec (Miamisburg, OH, USA) and
[2H9]trimethylchlorosilane from Transol (Essen, Germany).
The samples were applied by the cold on-column
°
technique at 60 C. One minute after injecting the sample,
°
the temperature of the oven was raised to 280 C at a rate
Syntheses
of 15 C minꢀ1. The flow-rate of the carrier gas (helium) was
°
Calcium [15N, 13C3]-(R)-pantothenate
2 ml minꢀ1
.
[15N, 13C3]-ˇ-Alanine (100 mg, 1.07 mmol) was dissolved in
water (2 ml) and calcium oxide (29.1 mg, 0.52 mmol) was
added to the solution. The mixture was stirred until the
calcium oxide was dissolved, then the water was evaporated
in vacuo. Methanol (5 ml), diethylamine (2 µl) and (R)-
(ꢀ)-pantolactone (161.2 mg, 1.24 mmol) were added to the
residue and the mixture was heated in an atmosphere of
Nuclear magnetic resonance (NMR) spectroscopy
NMR spectra were recorded with an AM 360 instrument
(Bruker, Karlsruhe, Germany). 1H NMR: transmitter fre-
quency, 360.13 MHz; spectral width, 7246.377 Hz, repetition
time, 3.2 s; 16 scans; 32K data set. 13C NMR: transmit-
ter frequency, 90.56 MHz; spectral width, 23 809 Hz; rep-
etition time, 2.5 s; 256 scans; 64K data set, 1 Hz line
broadening. Processing was performed by multiplication
with a Lorentz—Gaussian function prior to transforma-
tion. Shifts are expressed in ppm downfield from sodium
3-(trimethylsilyl)[2,2,3,3-d4]propionate in D2O, and tetra-
methylsilane in CDCl3, respectively. Coupling constants J
are expressed in hertz.
°
nitrogen at 70 C for 8 h. The target compound (180.6 mg) was
precipitated by pouring the cold mixture into cold acetone
and purified by recrystallization from methanol–acetone.
The numbering of calcium [15N,13C3]-(R)-pantothenate is
detailed in Fig. 1.
1H NMR (D2O), υ D 0.90 (s, Me), 0.94 (s, Me), 2.43 (dm,
2
2
2JHC 118, H-2), 3.40 (d, JHH 11, H-40), 3.45 (dm, JHC 138,
H-3), 3.52 (d, JHH 11, H-40), 3.99 (s, H-20). 13C NMR (D2O),
2
1
1
2
υ D 38.9 (ddd, JC2,C3 36.3, JCN 10.9 and JCC 3.6, C-3), 39.4
Chemicals
1
1
2
(dd, JCC 36.3 and 45.1, C-2), 183.2 (dd, JC1,C2 45.1 and JCC
The following chemicals were obtained commercially:
[15N, 13C3]-ˇ-alanine, N,O-bis(trimethylsilyl)trifluoroaceta-
mide (BSTFA), D2O, (S)-(ꢀ)-3,3-dimethyl-2-hydroxybutyric
acid, ˛-hydroxy-ꢀ-butyrolactone, (S)-(ꢀ)-ˇ-hydroxy-ꢀ-
butyrolactone, imidazole, neopentyl-glycol, panthenol, (R)-
(ꢀ)-pantolactone, (R)-pantothenic acid hemicalcium salt and
(R)-pantothenic acid sodium salt from Aldrich (Steinheim,
4.4, COOH).
(R)-Pantoylpropylamide
(R)-(ꢀ)-Pantolactone (162 mg, 1.24 mmol) was added to
a mixture of propylamine (63 mg), methanol (5 ml) and
diethylamine (2 µl), and the suspension was heated in an
°
atmosphere of nitrogen at 70 C for 8 h. Pouring the cold
Germany),
N-methyl-N-tert-butyldimethylsilyltrifluoroa-
mixture into cold acetone gave the target compound (222 mg)
as a brown oil.
cetamide (MBSTFA) and trimethylchlorosilane (TMCS)
Figure 1. Mass spectrum of tris(trimethylsilyl)pantothenic acid in the EI mode.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 555–562

MS of TMS-pantothenic acid and related substances
557
1H NMR (CDCl3), υ D 0.94 (t, H-1, 3J7.5), 1.00 (s, Me), 1.08
(s, Me), 1.55 (tq, 3J7.5, 7.1 H-2), 3.27 (dt, 3J7.1, 6.4, H-3), 3.95
Mass spectrum of TMS-neopentyl-glycol (relative inten-
sities (%) in parentheses) in the EI mode: m/z 143 (100), 147
(97), 73 (66), 158 (24), 103 (19), 75 (18), 144 (16), 148 (15), 133
(11), 145 (10), 149 (10), 45 (8), 88 (7), 191 (4), 233 (3). Mass
spectrum of TMS-panthenol in the EI mode: m/z 73 (100),
103 (69), 117 (46), 277 (43), 157 (41), 187 (37), 247 (29), 75 (21),
147 (13), 158 (12) 278 (10), 406 (7). Mass spectrum of TMS-
pantolactone in the EI mode: m/z 75 (100), 143 (55), 73 (37),
187 (26), 45 (15), 144 (12). Mass spectrum of TMS-pantoinic
acid in the EI mode: m/z 73 (100), 103 (75), 147 (58), 157 (53),
117 (48), 247 (47), 75 (29), 220 (25), 133 (13), 74 (13) 231 (10),
349 (3).
2
2
(d, J8.8, H-40), 4.03 (d, J8.8, H-40), 4.13 (s, H-20). 13C NMR
(CDCl3), υ D 14.5 (C-1), 21.9 (C-2), 23.2 (Me), 24.4 (Me), 42.4
(C-3), 43.9 (C-30), 74.2 (C-40), 78.8 (C-20), 176.4 (CONH).
Pantoinic acid and 2,4-dihydroxybutyric acid
(R)-(ꢀ)-Pantolactone (1.30 g, 10 mmol) or ˛-hydroxy-ꢀ-
butyrolactone (951 mg, 9.3 mmol), respectively, was dis-
solved in an aqueous solution of potassium hydroxide
(11 mmol lꢀ1, 10 ml) and refluxed for 5 h. After addition
of hydrochloric acid (1.5 ml, 37%), the solution was extracted
with ethyl acetate (2 ð 20 ml). Removal of the solvent
gave pantoinic acid (709 mg) and 2,4-dihydroxybutyric acid
(650 mg), respectively, as grey powders.
tert-Butyldimethylsilyl derivative of pantothenic acid
(R)-Pantothenic acid sodium salt (10 mg) was reacted with a
mixture of MBSTFA (50 µl) and pyridine (50 µl) for 60 min at
1
NMR data for pantoinic acid: H NMR (D2O), υ D 1.03
(s, Me), 1.20 (s, Me), 4.10 (d, 2J9.0, H-40), 4.15 (d, 2J9.0, H-40),
80 C. After cooling to room temperature, the solution was
°
4.39 (s, H-20).
evaporated to dryness in a stream of nitrogen and hexane
(500 µl) was added to the residue.
Trimethylsilyl derivatives of pantothenic acid
(R)-Pantothenic acid hemicalcium or sodium salt (200 mg)
was reacted with a mixture of BSTFA (2 ml) and TMCS (1 ml)
[2H]9TMS derivative of pantothenic acid
(R)-Pantothenic acid sodium salt (240 mg) was reacted with
a mixture of [2H]9TMCS (400 µl), imidazole (200 mg) and
dichloromethane (10 ml) for 8 h at ambient temperature.
The suspension was filtered and the filtrate evaporated to
dryness. Addition of hexane (10 ml) to the residue gave the
title compound.
°
for 10 min at 60 C. The suspension was then evaporated to
dryness in a stream of nitrogen and hexane (10 ml) was
added to the residue. After heating for 5 min at 60 C, the
suspension was filtered.
Trimethylsilylation of pantothenic acid in aqueous solu-
tion was achieved as follows: after acidifying the solution to
a pH of 3 (pK D 4.4), PA was extracted in ethyl acetate. The
extracts were dried over anhydrous Na2SO4 and evaporated
to dryness in a stream of nitrogen. Then, BSTFA (100 µl) and
pyridine (100 µl) were added and the mixture was heated for
°
RESULTS AND DISCUSSION
Mass spectrometry of trimethylsilyl derivatives of
pantothenic acid and of related compounds upon
electron ionization
°
60 min at 80 C in a closed vial. After cooling to room tem-
perature, the solution was evaporated to dryness in a stream
of nitrogen and hexane (100 µl) was added to the residue.
1H NMR (CDCl3), υ D ꢀ0.07 (s, TMS), ꢀ0.04 (s, TMS),
For quantitation of PA via the recently developed SIDA
method,2 the vitamin has to be volatilized by trimethylsilyla-
tion. In order to elucidate the mass spectrometric behaviour
of its TMS derivative, the latter was generated by react-
ing the sodium salt of PA with a mixture of BSTFA and
TMCS. GC/MS and NMR experiments revealed the result-
ing tris-TMS-PA (1) to be sufficiently pure and to show data
consistent with earlier studies.3,4
3
0.13 (s, TMS), 0.62 (s, Me), 0.75 (s, Me), 2.37 (t, J5.5, H-2),
2
2
3
3.06 (d, J9.5, H-40), 3.15 (d, J9.5, H-40), 3.34 (dt, J6.1, 6.0,
H-3), 3.88 (s, H-20). 13C NMR (CDCl3), υ D 0.27 (TMS), 0.30
(TMS), 1.9 (TMS), 20.1 (Me), 21.6 (Me), 34.7 (C-3), 36.2 (C-2),
40.2 (C-30), 69.0 (C-40), 76.9 (C-20), 173.42 (CONH), 173.47
(COOTMS).
The mass spectrum of 1 in the EI mode exhibits
one prominent fragment at m/z 291 (see Fig. 1). In an
earlier study, Prosser and Sheppard3 proposed a hydrogen
migration on cleavage of the bond between C-20 and C-30
(for assignment of carbon atoms, see Fig. 1) to form this
fragment, a prominent path known as the McLafferty-like
rearrangement. By contrast, Banno et al.4 assumed this ion to
originate from the molecular ion by loss of two TMS groups
associated with a double hydrogen transfer. To date, this kind
of fragmentation has not been reported in the literature, but
the route shown in Fig. 2 could be postulated. In a concerted
or sequential reaction the two proximate TMS groups may
be expelled with a concomitant two-hydrogen shift from the
ejected TMS groups to the remaining TMS-PA molecule. The
loss of the TMS groups might be energetically favoured by
formation of stable 1,1,3,3-tetramethyl-1,3-disilacyclobutane,
which reportedly is formed in a radical mechanism upon
pyrolysis of tetramethylsilane.5
HRMS of tris(trimethylsilyl)-PA, (relative intensities
(%) in parentheses) in the EI mode (resolution 14 400):
m/z 103.05537 (100), 73.0460 (98), 291.1322 (75), 247.15161
(57), 157.1059 (56), 117.07347 (51), 75.0250 (28), 201.07669
(25), 147.0711 (23), 292.1420 (23), 420.2018 (14). HRMS of
tris(trimethylsilyl)-[15N,13C3]PA (relative intensities (%) in
parentheses) in the EI mode (resolution 10 000): m/z 103.0565
(100), 73.0463 (95), 295.1363 (80), 157.1029 (57), 247.1485
(56), 117.0720 (51), 205.0912 (38), 75.0248 (28), 147.0649 (26),
424.2024 (23), 296.1436 (21).
Trimethylsilyl derivatives of other compounds
The respective compound (10 mg) was reacted with a
mixture of BSTFA (50 µl) and pyridine (50 µl) for 60 min
°
at 80 C. After cooling to room temperature, the solution was
evaporated to dryness in a stream of nitrogen and hexane
(500 µl) was added to the residue.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 555–562

558
M. Rychlik
Figure 2. Hypothetical fragmentation route of tris(trimethylsilyl)pantothenic acid by ejection of
1,1,3,3-tetramethyl-1,3-disilacyclobutane.
In order to verify either the McLafferty-like mechanism or
the disilacyclobutane route, two groups of experiments were
performed. The first group involved syntheses of substances
related to PA with subsequent mass spectral investigation
of their TMS derivatives, and the second group was focused
on analysing the elemental composition of the respective
fragment. The latter task was solved by measuring its isotopic
distribution and by HRMS.
To study the influence of the ˇ-alanine moiety on the
decomposition of TMS-PA, the former was substituted
(i) with its analogue labelled with carbon-13 and nitrogen-15
and (ii) with propylamine. The target compounds were easily
accessible by reacting pantolactone with the hemicalcium salt
of [15N, 13C3]-ˇ-alanine and 1-propylamine, respectively, in
the presence of dimethylamine. As expected, the TMS deriva-
tives of [15N, 13C3]PA and pantoylpropylamide provided
the corresponding fragment of considerable abundance at
[M ꢀ 144]C to that of TMS-PA, which can be seen in Figs 3
and 4. These findings corroborated the ions origin to be
the bis-TMS-pantoyl moiety. Even if the amide group is
absent, as evidenced by the mass spectra of TMS-pantoinic
acid, the same fragmentation was observed. In contrast,
the TMS derivative of pantolactone did not reveal such a
decomposition, which could be explained, supposing the
disilacyclobutane mechanism, by the absence of two TMS
groups. In a like manner, assuming the McLafferty rear-
rangement, the distance between the carbonyl group and
the ꢀ-hydrogen may be too large for the necessary 1,5-H
shift.
In a further series of experiments, the pantoyl group
was structurally varied in order to unravel the true
reaction sequence. On the one hand, TMS-2-hydroxy-
3,3-dimethylbutyric acid, with only one trimethylsilylated
hydroxyl, was prepared. Regarding its mass spectrum, no
[M ꢀ 2TMS C 2H]C ion arising from the disilacyclobutane
route was observed, whereas the McLafferty fragment of
m/z 220 was found in low abundance. On the other hand, for
TMS-2,4-dihydroxybutyric acid, differing from the pantoyl
Figure 3. Mass spectrum of tris(trimethylsilyl)-[15N, 13C3]pantothenic acid in the EI mode.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 555–562

MS of TMS-pantothenic acid and related substances
559
Figure 4. Mass spectrum of bis(trimethylsilyl) pantoylpropylamide in the EI mode.
group only by the absence of the two methyl groups at C-30,
the disilacyclobutane fragment [M ꢀ 2TMS C 2H]C might be
expected, but was absent. Likewise, the McLafferty fragment
of m/z 220 was not found, obviously owing to the steric
hindrance of a ꢀ-hydrogen migration to the carbonyl moiety.
In particular, the latter study indicated that the McLafferty
mechanism was the more probable.
The most straightforward test for distinguishing between
the two hypotheses was the variation of the trialkylsilyl
groups. First, the tert-butylsilyldimethyl (TBDMS) deriva-
tives were prepared, and, following the disilacyclobutane
mechanism, an [M ꢀ 2TBDMS C 2H]C ion at m/z 333 would
be expected. However, in the spectrum (Fig. 5) this ion
was not observed, which might be attributed to the steric
Figure 5. Mass spectrum of tris(tert-butyldimethylsilyl) pantothenic acid in the EI mode.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 555–562

560
M. Rychlik
Figure 6. Mass spectrum of tris([2H9]trimethylsilyl)pantothenic acid in the EI mode.
hindrance of this process by the bulky TBDMS groups.
In contrast, the McLafferty fragment of m/z 375 was
formed in 20% abundance. However, the most striking
argument supporting the McLafferty mechanism was found
by analysing tris-[2H9]TMS-PA. From Fig. 6 it is seen that
the McLafferty ion of m/z 309 was found in 15% abun-
dance, whereas the ion of m/z 302 that would be formed
by the disilacyclobutane mechanism could not be detected.
Moreover, the improbability of the disilacyclobutane mech-
anism is evidenced by the following considerations. The
[M ꢀ 2TMS C 2H]C ion should occur in substances includ-
ing two TMS groups at a similar distance as in TMS-PA.
This is the case for TMS-neopentyl glycol, but this com-
pound yielded no disilacyclobutane fragment of m/z 204
upon EI.
The McLafferty rearrangement was not easy to confirm
by HRMS. The two alternative fragments of elemental com-
position C11H25O4N1Si2 (m/z 291.13220) for the McLafferty
ion and C12H25O5N1Si1 (m/z 291.15021) for the disilacyclobu-
tane mechanism, differ by only 18 mu. Hence a resolution
high as 16 200 was necessary to differentiate them. As direct
inlet of TMS-PA resulted in decomposition, HRMS had
to be run in the GC/MS mode. In consequence, accurate
mass measurements were hampered by the high source
pressure caused by the GC carrier gas. Unfortunately, a
signal of the respective ion was not detectable beyond a
maximum resolution of 14 400 even in the selected ion mon-
itoring mode. Thus the mass determination resulted in a
signal maximum at m/z 291.13220 and a minor signal at
m/z 291.15021 showing an intensity of 30%. Although the
10% valley definition of resolution could not be achieved,
HRMS reveals the ion very likely to be the McLafferty frag-
ment.
Further support for the above considerations comes from
the isotopic distribution of the ion of m/z 291. Comparing
the abundances of the ions of m/z 292 to 295 (see Table 1),
the measured data matched the predicted abundances
for the McLafferty fragment. In contrast, especially the
abundances of the ions of m/z 293 and 294 show considerable
discrepancies from the calculated data for the fragment
resulting from the disilacyclobutane path.
Fragmentation routes of trimethylsilyl derivatives
of pantothenic acid and related compounds by
B/E-linked scan techniques and collision-induced
dissociation
After having clarified the nature of the m/z 291 ion, the
other fragmentation processes of TMS-PA were examined
by B/E-linked scanning. Owing to its low abundance, the
metastable ion analysis of the molecular ion could not be
performed in the GC/MS mode. On the other hand, in direct
Table 1. Isotopic distribution of the ion of m/z 291
Isotopic distribution (%)
Calculated for the ion having the composition
a
b
m/z Measured C12H25O5N1Si1
C11H25O4N1Si2
291
292
293
294
295
100.0
22.7
9.95
100.0
19.3
5.5
100.0
23.2
8.5
1.5
0.3
0.5
0.1
1.7
0.1
a Fragment resulting from the disilacyclobutane route.
b Fragment resulting from the McLafferty rearrangement.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 555–562

MS of TMS-pantothenic acid and related substances
561
inlet experiments TMS-PA could not be volatilized without
decomposition. Therefore, only the product ions of MC at
m/z 291, 247 and 201 were subjected to linked scanning.
In Table 2 the product ions of the respective precursor ions
are listed. In consequence, a fragmentation scheme in the EI
mode can be proposed as displayed in Fig. 7.
The ion [M ꢀ CH3]C and those of m/z 73, 75, 103 and 117
are characteristic fragments of trimethylsilyl compounds.
Moreover, if two TMS groups are present in a molecule, the
ion of m/z 147 is well known. For all compounds including
the pantoyl moiety, such as TMS-PA, TMS-pantoylamide
and TMS-panthenol, the fragment of m/z 247 was found,
indicating that it originates from MC by ˛-cleavage between
C-10 and C-20. As evidenced by the B/E-linked scan, this
fragment gives rise to m/z 157. The McLafferty fragment
at m/z 291 decomposes to m/z 201 by expulsion of
trimethylsilanol and subsequently yields an ion of m/z 159.
The EI fragmentation was then compared with the CID
route by applying a supplementary a.c. voltage on the
endcaps of an ion trap mass spectrometer.
Interestingly, MC decomposes differently upon CID
to EI, in that only trimethylsilanol is expelled to give
m/z 345 as detailed in Table 3. In order to compare the
further route of fragmentation using the two techniques,
the major EI ions of m/z 420, 291 and 247 were also
subjected to CID. All these ions eject TMSOH to give
the ions m/z 330, 201 and 157, respectively, as was
already found in the metastable decomposition experiments.
The structures of the fragmentation products were then
clarified by CID of compounds analogous to TMS-PA. TMS-
pantoylpropylamide, containing TMS groups only at C-20
and C-40, in the same way expels TMSOH from MC and
Table 3. CID of TMS-pantothenic acid
Table 2. Metastable transitions in
the first field-free region in
TMS-pantothenic acid
Product ion (CID voltage) at m/z upon
Precursor ion
at m/z
MS/MS
MS3
435 (MC)
420
291
247
201
345 (19 V)
330 (18 V)
201 (24 V)
157 (27 V)
159 (30 V)
n.d.a
n.d.
159 (30 V)
n.d.
Precursor ion
at m/z
Product ion
at m/z
291
201
247
201
159
157
n.d.
a n.d., not determined.
Figure 7. Fragmentation scheme of tris(trimethylsilyl)pantothenic acid in the EI mode. Solid arrow, fragmentation upon EI; dashed
arrow, fragmentation upon CID.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 555–562

562
M. Rychlik
Table 4. CID of TMS-pantoylpropylamide
by CID of TMS-2,4-dihydroxybutyric acid and TMS-2-
hydroxy-3,3-dimethylbutyric acid (TMS-HDMB). For both
substances, the products of ˛-cleavage between C-10 and C-
20, the ions of m/z 219 and 159, respectively, were subjected
to CID as detailed in Table 5. The former ion, including
two TMS groups, ejects TMSOH in a similar way to 2. The
ion of m/z 159 of TMS-HDMB, however, lacks the TMS
group at C-40 and decomposes without loss of TMSOH
at a high dissociation voltage of 33 V directly to m/z 73.
Therefore, it can be concluded that in 2 TMSOH is ejected
from C-40.
Precursor ion
Product ion (CID Voltage)
at m/z
at m/z
333 (MC)
318
189
243 (21 V)
228 (23 V)
173 (30 V)
Table 5. CID of
(A) TMS-2-hydroxy-3,3-dimethylbutyric acid
and (B) TMS-2,4-dihydroxybutyric acid
The ion 4 of m/z 291 expels TMSOH from C-1 to give
ion 5 of m/z 201. Upon further CID, this ion fragments to
ion 6 of m/z 159 by ejection of ketene. The first reaction path
is established by CID of TMS-pantoylpropylamide (Table 5),
theMcLaffertyionofwhichatm/z189doesnotejectTMSOH.
As TMS-pantoylpropylamide does not include a TMS group
at C-1, this result implies that in 4 the TMSOH is expelled
from that carbon atom. The loss of ketene in the second
reaction is evidenced by CID of TMS-[15N, 13C3] PA, which
gives the respective ion of m/z 161, as shown in Table 6.
The mass shift of 2 u can be attributed to the ejection of a
species containing two labelled atoms, which is assumed to
be [13C2]-ketene.
Precursor ion
at m/z
Product ion (CID voltage)
at m/z
(A) 159
(B) 219
73 (33 V)
129 (18 V)
Table 6. CID of TMS-[15N, 13C3]pantothenic acid
Product ion (CID voltage) at m/z upon
Precursor ion
at m/z
MS/MS
MS3
439 (MC)
424
349 (19 V)
334 (18 V)
205 (24 V)
157 (27 V)
161 (30 V)
n.d.a
n.d.
161 (30 V)
n.d.
Acknowledgements
I am grateful to Mrs D. Fottner for her excellent technical assistance.
Appreciation is also due to Mr J. Scho¨nauer for skillful tuning of the
double-focusing Mass spectrometer.
295
247
205
n.d.
a n.d., not determined.
REFERENCES
[M ꢀ CH3]C as outlined in Table 4. Therefore, it can be
concluded that in TMS-PA the trimethylsilanols are ejected
from C-20 and C-40.
For the ion 2 at m/z 247 sketched in Fig. 7, however, the
TMS moiety at C-40 was found to be lost, thus producing the
cyclopropane derivative 3. This mechanism was evidenced
1. Song WO, Smith A, Wittwer C, Wyse BW, Hansen RG. Nutr. Res.
1990; 10: 439.
2. Rychlik M. J. Agric. Food Chem. 2000; 48: 1175.
3. Prosser AR, Sheppard AJ. J. Pharm. Sci. 1971; 60: 909.
4. Banno K, Matsuoka M, Horimoto S, Kato J. J. Chromatogr. 1990;
525: 255.
5. Fritz G, Grobe J. Z. Anorg. Allg. Chem. 1962; 315: 157.
Copyright  2001 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2001; 36: 555–562