Identification of lignans by liquid chromatography-electrospray ionization ion-trap mass spectrometry

Article abstract of DOI:10.1002/jms.1276

The fragmentation pattern of 30 compounds belonging to different classes of the lignan family was studied by liquid chromatography-electrospray ionization ion-trap mass spectrometry. On the basis of the observed fragmentation patterns, identification of d

Full text of DOI:10.1002/jms.1276

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2008; 43: 97–107
Published online 30 August 2007 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/jms.1276
Identification of lignans by liquid
chromatography-electrospray ionization ion-trap mass
spectrometry
∗
˚
Patrik C. Eklund, M. Josefin Backman, Leif A. Kronberg, Annika I. Smeds and
¨
Rainer E. Sjoholm
˚
˚
Laboratory of Organic Chemistry, Process Chemistry Centre, Abo Akademi University, Biskopsgatan 8, FIN-20500 Abo, Finland
Received 22 December 2006; Accepted 2 July 2007
The fragmentation pattern of 30 compounds belonging to different classes of the lignan family was
studied by liquid chromatography-electrospray ionization ion-trap mass spectrometry. On the basis of the
observed fragmentation patterns, identification of different types of lignans was achieved. For example,
dibenzylbutyrolactone lignans showed a characteristic fragmentation pathway by the loss of 44 Da (CO2)
from the lactone moiety, whereas dibenzylbutanediols showed a loss of 48 Da by a combined loss
of formaldehyde and water from the 1,4-butanediol moiety. Lignan glycosides readily lost the sugar
residue to give the parent lignan as their primary product ion. In addition, several compound-specific
fragmentations were observed and used for identification of individual compounds.
A versatile method for analyses of lignans was developed using LC separation on a C8 column followed
by fragmentation and detection of ions produced in the ion trap. Copyright  2007 John Wiley & Sons,
Ltd.
KEYWORDS: lignan; HPLC; ESI; MS; ion trap; fragmentation
by their protective effects against hormone-related cancers
and due to their antioxidant effects.10–12 Therefore, new
and improved methods for analyses of lignans in different
matrices are needed and, in fact, regularly reported.13,14
The analysis of lignans in plant material as well as in
mammalian body fluids has traditionally employed GC
and GC/MS techniques of derivatized samples. The use
of LC/ESI/MS offers an alternative and sensitive method
for the analyses of underivatized lignans and lignans in
glycosidic and oligomeric forms. However, compared to
traditional electron ionization (EI), the fragmentation behav-
iors of organic compounds in collision induced dissociation
(CID) employed in MS/MS techniques are still not fully
understood. The use of fragmentation patterns for struc-
tural analysis may therefore not always be straightforward,
especially when unexpected fragmentations are observed.
In order to allow the identification of lignans by examining
their fragmentation behavior, we here report the fragmen-
tation data of different types of lignans using ESI-ion-trap
mass spectrometry.
INTRODUCTION
The combination of liquid chromatography and electrospray
ionization mass spectrometry (LC/ESI/MS) has become a
routine method in many areas of analytical organic chem-
istry. The provided mass spectrometric data as well as the
attainable high sensitivity has made these methods well
suited for the identification of specific compounds in mix-
tures of natural products. For example, successful qualitative
and quantitative analyses of lignans, flavonoids, alkaloids,
and lignin-based polyphenolic mixtures by LC/ESI/MS have
been reported previously.1–8 For qualitative purposes, the
use of electrospray ionization in combination with an ion
trap is superior, as the isolation of molecular ion species,
followed by fragmentation in the ion trap, provides useful
structural information on the analytes. Moreover, an unam-
biguous identification of compounds can be achieved by
comparison of known fragmentation patterns or by using
authentic reference compounds.
Lignans are plant constituents, which are defined by
a ˇ-ˇ-linkage of two phenyl propane units. These natural
productshaveattractedmuchinterestovertheyearsowingto
their widespread occurrence in the plant kingdom and their
broad range of biological activity.9 A dietary intake of lignans
has been shown to have beneficial effects on human health
EXPERIMENTAL
Reference compounds and chemicals
The structures of the analyzed compounds are presented in
Fig. 1.
Hydroxymatairesinol (5) was isolated from knotwood
of Norway spruce (Pica abies) as previously described.15
ŁCorrespondence to: Patrik C. Eklund, Laboratory of Organic
Chemistry, Process Chemistry Centre, Abo Akademi University,
Biskopsgatan 8, FIN-20500 Abo, Finland. E-mail: paeklund@abo.fi
˚
˚
Copyright  2007 John Wiley & Sons, Ltd.

98
P. C. Eklund et al.
Figure 1. Chemical structures of the analyzed lignans.
Compounds 1–3, 14, 15, and 22–25 were synthetically pre-
pared from hydroxymatairesinol according to our previously
published methods.16–19 Compound 13 was isolated from
Araucaria angustifolia knotwood and compound 4 from Pinus
sylvestris knotwood as previously reported.15 Compound 27
was isolated from Picea pungens.20 Compounds 29 and 30
were obtained by rhamnosylation of 1 and 13 by a strepto-
myces strain21 and compound 26 was obtained by cyclization
of 13 under acidic conditions. The methylated derivatives
9 and 16 were prepared in our laboratory by methylation
of the parent lignans (1 and 13) by MeI/K2CO3 in dry ace-
tone. Compound 10 was prepared by demethylation of 1 in
BBr3/CH2Cl2, and a subsequent etherification of the prod-
uct with diiodomethane afforded compound 8. Compounds
11 and 12 were prepared by total synthesis essentially as
described by Ma¨kela¨ et al.22; however, the desulfurization
of the intermediate was performed with AgNO3 and N-
chlorosuccinimide23 to afford the oxo-functionalized product
(12), which was reduced by NaBH4 to the corresponding
alcohol (11). Compound 17 was isolated from Picea abies
according to Erdtman.24
Compounds 6, 7 and 18–21 were provided by Prof. S.
Nishibe (University of Hokkaido, Japan). Podophyllotoxin
(28) was purchased from Fluka. The identity and purity of
the reference compounds was determined by GC, GC-MS
and NMR spectroscopy (and in this work by LC-MS).
Stock solutions (1mg/ml) of the reference samples
were prepared in MeOH (HPLC grade, Rathburn). The
dilutions for both LC and direct infusion analyses were
done with equal amounts of MeOH/ACN (1 : 1, v/v) and
0.1% acetic acid. Water was purified with a Simplicity 185
purifying system (Millipore, Bedford, MA, USA). The final
concentrations of analytes were 10–50 µg/ml. All solutions
°
were stored at ꢀ20 C and protected from light.
Sesame seeds were purchased from a local store. The
seeds were milled using a Polymix analytical mill A
10 (Kinematica AG, Switzerland). Approximately 0.5 g of
the milled seeds was weighed, thoroughly mixed with
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

ESI-ion-trap analysis of lignans
99
Figure 1. (Continued).
quartz sand (granularity 1–2 mm, Mallinckrodt Baker B.V.)
and poured into extraction tubes for accelerated solvent
extraction (ASE). The ASE extraction was performed using
an apparatus from Dionex Corp. (Sunnyvale, CA, USA)
as described previously25 with slight modifications. The
material was extracted first with hexane, then with acetone
and finally with acetone–water (70 : 30, v/v). The three
fractions were collected in separate glass tubes. The acetone
and acetone–water fractions were combined, and the solvent
was evaporated to dryness using a rotary evaporator. The
extract was redissolved in 9 ml of acetone–water (70 : 30,
v/v). An aliquot of 3 ml was transferred into a test tube and
the solvent was evaporated using a stream of nitrogen gas.
Enzymatic hydrolysis was performed by adding 1 mg of an
enzyme preparation containing 492 units of ˇ-glucuronidase
and 10 units of sulfatase (ˇ-glucuronidase/sulfatase, type
H-1, from Helix pomatia (Sigma-Aldrich Co., St Louis, MO,
USA)) dissolved in 1 ml of 0.01 M acetate buffer pH 5.0
added. The sample was placed in an ultrasonic bath for
5 min and then centrifuged for 10 min at 4000 rpm.
Liquid chromatography
The liquid chromatography/UV absorbance/electrospray
ionization ion-trap mass spectrometry (LC/UV/ESI/MSn)
analyses were performed on an Agilent 1100 Series HPLC
instrument (Agilent Technologies) consisting of a binary
pump, a vacuum degasser, an autosampler, a thermostatted
column compartment and a variable wavelength detec-
tor (VWD-UV). The analyzed compounds were chro-
matographed on a 3.5 µm, 2.1 ð 150 mm reversed-phase C8
analytical column (XTerra MS C8, Waters). Eluent A was
0.1% acetic acid with 1% 2-propanol (v/v) and eluent B was
acetonitrile–methanol (50 : 50). The gradient was as follows:
10–30% B from 0 to 20 min, 30–50% from 20 to 45 min,
50–95% from 45 to 46 min, 95% from 46 to 50 min, 95–10%
from 50 to 51 min and isocratic at 10% to equilibrate the
column from 51–55 min. The sample injection volume was
20 µl and the column flow rate was 0.3 ml/min. UV detection
was set at 280 nm.
°
and keeping at 37 C for 19 h, i.e. according to a slight
modification of the method optimized by Milder et al.13
The enzymatically hydrolyzed solution was liquid–liquid
extracted with 2 ð 0.75 ml of ethyl acetate. The ethyl acetate
phase was evaporated to dryness under nitrogen gas and
0.5 ml of MeOH/ACN/0.1% HAc (15 : 15 : 70 v/v/v) was
Mass spectrometry
For mass spectral analyses an Agilent 1100 Series LC/MSD
SL Trap ion-trap mass spectrometer was coupled to the
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

100
P. C. Eklund et al.
O
O
R2
H3CO
-O
H3CO
-O
H3CO
-O
Pathway I
-CO2
Pathway IIIa
R1 = H,
O
O
R1
R2 = OH (4)
R
-H2O
OCH3
OCH3
OCH3
Pathway IV
OH
m/z = 355
OH
OH
Pathway II
R1 = H,
R2 = OH (4)
- m/z 46
R = H, m/z 313 (1)
R = O, m/z 327 (2)
.
R1 = OH,
R2 = H (5)
-H2O
-CH3
O
Path-
way IIIb
H3CO
-O
.O
O
O
O
O
-O
R
H3CO
-O
OCH3
OCH3
OH
OH
R = H, m/z 342 (1)
R = O, m/z 356 (2)
m/z 327
OCH3
O
m/z 355
Scheme 1. Some commonly observed fragmentation patterns of dibenzylbutyrolactone lignans of the guaiacyl family.
Scheme 2. Fragmentation of dibenzylbutanediol lignans, exemplified by compound 13.
°
Agilent 1100 Series LC system via an ESI interface. The
operation parameters of the ESI ion source were as follows:
as follows: (1) drying gas temperature: 325 C, (2) drying gas
flow: 5 l/min, and (3) nebulizer gas pressure: 15 psi.
°
(1) drying gas (N2) temperature, 350 C, (2) drying gas flow,
12 l/min, (3) nebulizer gas (N2) pressure, 40 psi, (4) end plate
voltage, 3.5 kV, (5) end plate offset, ꢀ500 V, (6) capillary exit,
ca ꢀ116 V, and (7) skim 1 set at ꢀ40 V. Ion-trap parameters
were as follows: (1) accumulation time, 20 ms, (2) averages 5
and rolling averaging on and ion charge control on, target
set at 20 000. CID experiments coupled with multiple mass
spectrometry (MSn) employed helium as the collision gas. For
full-scan MS analysis, the spectra were recorded in the mass
range m/z 100–600 and target mass was set at m/z 350. The
fragmentation amplitude was 1.5 V and compound stability
100%. The data acquisition (Auto MSn) was set to subject
the two most abundant ions to multiple mass spectrometry
(MSn, n D 3).
RESULTS AND DISCUSSION
Although lignans have previously been analyzed by LC-
ESI-ion-trap mass spectrometry, no systematic study of
characteristic fragmentation patterns of different types of
lignans has been reported. Recently, the fragmentation
patterns of furofuran-type lignans and their glucosides were
reported by Ye et al.5 In accordance with this study, we have
noticed that lignan glycosides initially lose the sugar unit. In
our case, where lignan rhamnosides were used as reference
compounds (29 and 30), the corresponding fragmentation
was the loss of the deoxyhexose residue to give an ion at
[M ꢀ H ꢀ 146]ꢀ. Upon further fragmentation, the aglycones
behaved as the parent lignans.
In additional studies of fragmentation patterns, the pure
compounds were directly infused into the ESI source by a
syringe pump at a rate of 5 µl/min and at a concentration
of 200 µg/ml. The solvent was a 50/50 v/v mixture of 0.1%
acetic acid and acetonitrile. The gas flows to the source were
All lignans with a free phenolic group could be analyzed
in the negative ion mode, and were detected as deprotonated
molecules, although the eluent was acidic (Table 1). This
‘wrong-way-round’ deprotonation mechanism is probably
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

ESI-ion-trap analysis of lignans
101
Scheme 3. Fragmentation of furofurano lignans, exemplified by compound 18.
Scheme 4. Proposed fragmentation pattern of non-phenolic butyrolactone lignans demonstrated by compound 9.
abundance, and further fragmentation to MS3 was in most
due to the fact that the acetate anion is able to abstract
the proton from the phenol in gas-phase reactions, i.e.
the following ionization mechanism is plausible: M C
CH3COOꢀ ! [M ꢀ H]ꢀ C CH3COOH. This mechanism is
supported by the results of Znamenskiy et al.26 from their
studies on the mechanism of ion formation in ESI. They
found that evaporation of hydronium ions from the charged
droplets could result in drastic increase in the pH followed
by the deprotonation of the analyte. Lignans lacking phenolic
groups cannot be deprotonated and need to be analyzed in
positive ion mode either by protonation or by formation
of adducts (Table 2). All lignans formed protonated species
[M C H]C in positive ion mode, but NaC or KC adducts
were also detected at almost equal intensities. Some of the
non-phenolic lignans could also be analyzed as their acetate
adducts in negative ion mode. Comparison of negative and
positive ion mode detection, however, showed that adduct
formation was detrimental to further fragmentation. The
first fragmentation of adducts usually gave [M C H]C of low
cases possible but with a serious decrease in sensitivity. In
negative ion mode, the deprotonated molecule [M ꢀ H]ꢀ
was the only ion detected. Compared with positive ion
mode, in which one fragmentation step was needed to give
the parent lignan, the lignan moiety could directly be further
fragmented in negative ion mode by MS2 ꢀ MS4 experiments.
In general, clearer and more useful information was obtained
in negative ion mode detection.
When different classes of lignans: dibenzylbutyrolac-
tones, dibenzylbutanediols, aryltetralines, furano- and furo-
furanolignans (Fig. 1), were compared, some characteristic
fragmentations of the different classes, as well as of the
individual lignans, could be observed.
Dibenzylbutyrolactone lignans
Most of the dibenzylbutyrolactones (1–5, 7, 10, and 12)
gave the characteristic primary product ion [M ꢀ H ꢀ 44]ꢀ
presumably by the loss of CO2 from the lactone ring
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

102
P. C. Eklund et al.
Scheme 5. Possible positive ion mode fragmentation of non-phenolic dibenzylbutanediol lignans demonstrated by compound 16.
Scheme 6. Positive ion mode fragmentation of compound 21.
(Scheme 1, Pathway I). The lactone lignans 2 and 7 gave
[M ꢀ H ꢀ 15]ꢀ as the major product ion (Pathway II), but
it was accompanied by the loss of CO2 at almost the same
intensity. Compounds 4, 5, and 25 gave [M ꢀ H ꢀ 18]ꢀ as
their primary product ion (MS2), which can be explained by
the loss of water from the aliphatic alcohol groups (Pathway
IIIa and IIIb). Compound 4 also lost 46 Da almost to the same
extent as it lost water. This may be explained by a concerted
loss of CO2 and H2 or alternatively CO and H2O (Scheme 1,
Pathway IV). The loss of CO (28 Da) was, however, not a
generally observed fragmentation reaction for the analyzed
lignans.
The aryltetralin lactone lignan 22 did not lose CO2 as a pri-
mary fragmentation; instead the primary fragmentation was
found to be the loss of a methyl radical. It therefore seemed
that the more flexible structure of the dibenzylbutyrolac-
tones (compared with aryltetralines) facilitated and favored
the cleavage of CO2. The loss of a methyl radical was the most
commonly observed fragmentation behavior (MS1 –MS4) of
all lignans containing methoxyl groups. This result has also
been obtained from methoxylated flavonoids.27 It is worth
mentioning that 22 and 25 also gave the lactone characteristic
loss of 44 Da (CO2) upon further fragmentations.
was the loss of 48 Da. This loss could be explained
by the combined loss of formaldehyde and water from
the diol structure (Scheme 2, Pathway I). Compared to
the dibenzylbutyrolactone lignans, the dibenzylbutanediol
lignans were also more readily cleaved at the ˇ-position
(Pathway II). On the basis of our observations, it can
be concluded that most lignans containing hydroxymethyl
groups (13–15, 24) lose formaldehyde easily in the negative
ion mode (Scheme 2, Pathway III). Lariciresinol (24), a
tetrahydrofuranolignan, which like the butanediol lignans
contains a hydroxymethyl group, also gave [M ꢀ H ꢀ 30]ꢀ
as its primary product ion. This fragmentation was observed
already in MS1 and seemed to take place in the ionization
source. The lactol lignan 27 gave, as most butanediol
lignans, a primary product ion [M ꢀ H ꢀ 48]ꢀ. In this
case the product ion is most probably formed by the
loss of formaldehyde from the lactol moiety and water
from the benzylic alcohol group. The loss of formaldehyde
from lactol lignans has been suggested previously.5 For
the aryltetralin diol lignan 23, the loss of 48 Da was not
observed, but [M ꢀ H ꢀ 46]ꢀ was detected as a minor ion
in addition to the base peak [M ꢀ H ꢀ 15]ꢀ in MS2. The
loss of 46 in the case of lignan diols could be attributed
to a combined loss of a methoxyl (31 Da) and a methyl
radical (15 Da) or by the loss of CH2OH and a methyl
radical. In accordance with the fragmentation patterns of
the corresponding butyrolactone lignan 22, the aryltetralin
Butanediol lignans
The most generally observed fragmentation pattern (in
negative ion mode) for the butanediol lignans (13–15)
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

ESI-ion-trap analysis of lignans
103
Table 1. Negative ion mode fragmentations (phenolic lignans)
Retention time [M ꢀ H]ꢀ
Compound
(min)
m/z
ESI-MSn m/z (Relative intensity %)a
Matairesinol (1)
29.6
357
MS2 [357]: 342 (60), 313 (100), 298 (30) 209 (80)
MS3 [357 ! 313]: 298 (100), 161 (40)
MS3 [357 ! 342]: 313 (85), 298 (100), 178 (70), 122 (80)
MS3 [357 ! 209]: 191 (100), 165 (40)
7-Oxomatairesinol (2)
25.3
371
MS2 [371]: 356 (100), 327 (50), 205 (60)
MS3 [371 ! 356]: 327 (100), 136 (25)
MS3 [371 ! 327]: 312 (100), 175 (50)
MS4 [371 ! 356 ! 327]: 297 (25), 283 (100), 191 (40), 136 (40)
Enterolactone (3)
32.7
23.8
297
373
MS2 [297]: 253 (100), 189 (17)
MS3 [297 ! 253]: 235 (20), 195 (100), 143 (47)
MS2 [373]: 355 (97), 327 (100), 235 (28), 223 (40)
MS3 [373 ! 355]: 340 (69), 311 (100)
Nortrachelogenin (4)
MS3 [373 ! 327]: 312 (100), 147 (41)
7-Hydroxymatairesinol (5)
18.4
373
MS2 [373]: 355 (100), 311 (13)
MS3 [373 ! 355]: 340 (30), 311 (100), 296 (59), 231 (37), 160 (58)
MS4 [373 ! 355 ! 311]: 296 (100), 281 (17), 160 (30)
MS2 [387]: 357 (17), 339 (48), 329 (100), 249 (15) 193 (13)
MS3 [387 ! 329]: 314 (100), 193 (8), 135 (3)
MS2 [371]: 356 (100), 312 (21), 295 (12), 209 (5)
MS3 [371 ! 356]: 341 (15), 327 (33), 297 (29) 205 (29), 147 (37), 121 (100)
MS2 [329]: 285 (60), 207 (100), 189 (20)
MS3 [329 ! 207]: 161 (100)
Trachelogenin (6)
29.7
34.7
20.1
22.7
27.9
21.3
387
371
329
313
311
361
Arctigenin (7)
4, 40-Dihydroxyenterolactone (10)
7-Hydroxyenterolactone (11)
7-Oxoenterolactone (12)
Secoisolariciresinol (13)
MS2 [313]: 251 (18), 239 (100), 189 (35), 147 (22)
MS3 [313 ! 239]: 131 (100)
MS2 [311]: 267 (100), 265 (12), 249 (10), 159 (75)
MS3 [311 ! 267]: 249 (7), 159 (100)
MS2 [361]: 346 (90), 331 (14), 313 (30), 165 (100), 179 (30)
MS3 [361 ! 346]: 223 (30), 179 (80), 165 (100), 122 (45)
MS3 [361 ! 165]: 147 (100), 129 (8)
7-Hydroxysecoisolariciresinol (14)
Enterodiol (15)
13.2
27.3
26.3
34.6
25.9
25.7
26.7
17.2
22.3
19.1
35.1
17.3
377
301
357
371
387
417
355
359
MS2 [377]: 329(100)
MS3 [377 ! 329]: 299 (49), 284 (24), 193 (31), 178(100)
MS2 [301]: 271 (15), 253(100)
MS3 [301 ! 253]: 146(100)
Pinoresinol (17)
MS2 [357]: 342 (7), 311 (6), 151 (100), 136 (30)
MS3 [357 ! 151]: 136 (100)
Phillygenin (18)
MS2 [371]: 356 (100), 326 (2)
MS3 [371 ! 356]: 341(9), 163 (26), 151 (18), 136(13), 121 (100)
MS2 [387]: 181 (100), 372 (40), 166 (34), 151 (95) 136 (37)
MS3 [387 ! 181]: 136 (100), 123 (35)
Medioresinol (19)
Syringaresinol (20)
MS2 [417]: 402 (33), 181 (100) 166 (33), 151 (21)
MS3 [417 ! 181]: 166 (100), 151 (66)
Conidendrin (22)
MS2 [355]: 340(100)
MS3 [355 ! 340]: 325(100), 296 (70), 278 (30)
Cyclolariciresinol (23)
Lariciresinol (24)
MS2 [359]: 344(100), 313 (10)
MS3 [359 ! 344]: 329 (30), 313(100), 159 (60)
359/329 MS2 [329]: 299 (35), 284 (20), 192 (30), 178 (100), 160 (30)
MS3 [329 ! 178]: 160 (100), 147 (45), 122 (20)
Iso-hydroxymatairesinol (25)
Anhydro-secoisolariciresinol (26)
Todolactol A (27)
373
343
375
MS2 [373]: 355 (100), 311 (10)
MS3 [373 ! 355]: 311(100), 281 (50), 216 (30), 159 (65)
MS2 [343]: 328 (100), 313 (2)
MS3 [343 ! 328]: 313 (100), 297 (26), 203 (43), 163 (32), 161(24)
MS2 [375]: 327(100), 191 (30)
MS3 [375 ! 327]: 191 (100), 176 (10)
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

104
P. C. Eklund et al.
Table 1. (Continued)
Retention time
(min)
[M ꢀ H]ꢀ
m/z
Compound
ESI-MSn m/z (Relative intensity %)a
Matairesinol-rhamnoside (29)
26.7
19.5
503
507
MS2 [503]: 357 (100)
MS3 [503 ! 357]: 313 (100), 298 (30), 209 (60), 147 (35)
Secoisolariciresinol-rhamnoside (30)
MS2 [507]: 361 (100)
MS3 [507 ! 361]: 346 (90), 165 (100)
a Compound-specific fragment ions are in boldface.
Table 2. Positive ion mode fragmentations
Retention time
[M C H]C
Compound
(min)
m/z
[M C KC]C
ESI-MSn m/z (Relative intensity %)a
Hinokinin (8)
49.3
355
393
453
425
429
–
MS2 [355]: 337 (100), 319 (30), 261 (12), 135 (24)
MS3 [355 ! 337]: 319 (100), 289 (75),
261 (18), 161 (25)
Podophyllotoxin (28)
Dimethylmatairesinol (9)
Dimethylsecoisolariciresinol (16)
Sesamin (21)
33.3
39.3
31.5
49.4
415
MS2 [415]: 397 (100), 313 (4), 229 (6)
MS3 [415 ! 397]: 379 (58), 351 (89),
313 (94), 282 (100)
387
MS2 [387]: 369 (100), 319 (12), 249 (4),
204 (5), 151 (80)
MS3 [387 ! 369]: 338 (80), 351 (50), 151 (18)
MS2 [373]: 355 (100), 340 (6), 151 (8)
MS3 [373 ! 355]: 340 (100), 291 (2),
177 (70), 147 (20)
MS2 [337]: 319 (100), 289 (89), 279 (18),
267 (34) 173 (41)
MS3 [337 ! 319]: 289 (100), 261 (36)
373 [M C H ꢀ H20]C
337 [M C H ꢀ H2O]C
a Compound specific fragment ions are in boldface.
structure seemed to stabilize the aliphatic part of the molecule
and the loss of a methyl radical was the major fragmentation
pathway observed.
addition to the ˛,ˇ-cleavage, the loss of a methyl radical was
the second most abundant fragmentation pathway observed
for the furofurano lignans 17–20.
Furofuranolignans
Non-phenolic lignans
Phenolic lignans belonging to the furofurano class (17, 19–20)
were generally cleaved between the ˛- and ˇ-position in the
side chain to give the product ion m/z 151 (guaiacyl) or
m/z 181 (syringyl) (negative ion mode). Compound 17 gave
m/z 151 as the major primary product ion (MS2), which
in turn gave the base peak m/z 136 in MS3 (Scheme 3,
Pathway Ia) in accordance with results obtained for the
corresponding glucoside by Ye et al.5 This fragmentation
pattern can certainly be considered as the most characteristic
fragmentation for this class of lignans (in negative ion
mode). However, compound 18 gave [M ꢀ H ꢀ 15]ꢀ as the
primary product ion (Scheme 3, Pathway II), and product
ions from the ˛,ˇ-cleavage were in fact observed only in
MS3. Fragmentation of [M ꢀ H ꢀ CH3]ꢀ gave m/z 121 and
163 as the most abundant ions. These product ions are
most probably formed from the dimethoxyphenyl moiety as
proposed in Scheme 3, Pathway III. Product ions from the
guaiacyl moiety (m/z 151 and 136) were observed only in low
abundances (Scheme 3, Pathway Ib), which indicated that
the dimethoxyphenyl unit is more readily cleaved from the
lignan skeleton and rearranged to form m/z 121 and 163. In
In positive ion mode, the most commonly observed frag-
mentation of the non-phenolic lignans was the loss of water
to give the ion [M C H ꢀ 18]C (Table 2). The butyrolactone
lignans 8, 9, and 28 could lose water from the protonated
lactone to yield an acylium ion (Scheme 4) as described in
detail for butyrolactones previously.28 Further fragmenta-
tion of [M C H ꢀ H2O]C resulted in a second loss of water.
For compounds 8 and 9 this fragmentation must involve the
acylium intermediate, whereas compound 28 could elimi-
nate water also from the aliphatic alcohol group. A possible
mechanism for the elimination of water from the acylium
ion is presented in Scheme 4. The diol lignan 16 also readily
loses two molecules of water; in fact, the first water molecule
is lost already in the ionization source and [M C H ꢀ H2O]C
was detected as the base peak in MS1. A proposed frag-
mentation pattern for compound 16 is shown in Scheme 5.
The corresponding phenolic compound 13 showed a similar
fragmentation pattern in positive ion mode.
Also, the non-phenolic furofurano lignan 21 lost water
from the substituted tetrahydrofuran ring in positive ion
mode detection (Scheme 6).
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

ESI-ion-trap analysis of lignans
105
Figure 2. HPLC chromatograms of 28 lignan reference compounds. From top to bottom: (a) UV chromatogram. Extracted ion
chromatograms: selected ions (b) 371 (MS) (c) 417 (MS) and (d) 343 (MS) (e) 205 of 371 ! 356 ꢀMS3ꢁ showing the identification of
the overlapping compounds 7, 18 and 26.
The protonated molecule [M C H]C was detected only in
molecular ion species and the fragmentation patterns of the
compounds studied in this work are presented in Table 1.
Some specific product ions suitable for identification pur-
poses are displayed in boldface.
low abundance for compound 21, and as for the diol lignans,
[M C H ꢀ H2O]C was the major ion detected in MS1. Further
fragmentation led to a second loss of water (Scheme 6). The
loss of water already in the ionization source prior to the
detection in the ion trap seems to be specific (under these
conditions) for non-phenolic furofurano- and butanediol
lignans in positive ion mode. Moreover, compound 21
showed a poor response in positive ion mode, which must
be due to a low proton affinity. Therefore there seems to
be variation in the proton affinity of structurally related
lignans, which leads to variation in mass spectrometric
response. When non-phenolic lignans are analyzed under
these conditions, attention should be paid to this problem.
In addition to the generally observed fragmentation pat-
terns discussed above, compound-specific fragmentations
were observed both in the negative and in the positive ion
mode detection. These specific fragmentations in combi-
nation with the observation of molecular ion species can
be used for the identification of individual compounds.
Moreover, generally observed fragmentations and specific
fragmentation may provide a means for structural eluci-
dation of hitherto unknown lignans. Retention times, the
LC/ESI/MS analyses of the reference lignans and
hydrophilic lignans in a sesame seed extract
An LC/ESI/MS method was established with the aim of
analyses of lignans in plant extracts. Initially, the method
was optimized by the use of a mixture of 28 reference
compounds. Auto-MS detection was performed in such a
way that the two most abundant ions were subjected to
MS2 –MS3 experiments.
The chromatographic separation was performed on a
C8 reversed-phase column and the column was eluted with
a gradient consisting of (1) methanol : acetonitrile (50 : 50)
and (2) 0.1% acetic acid (see ‘Experimental’). The UV
chromatogram in Fig. 2(a) was obtained from the analysis of
the mixture of a reference compounds, showing that some
of the compounds coeluted. Each compound could easily be
identified in the chromatogram by the knowledge of their
specific ion-trap fragmentation behavior and according to
their retention times. For example, by extracting from the MS1
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

106
P. C. Eklund et al.
Figure 3. HPLC chromatogram of a sesame seed extract. Extracted ion chromatograms of [M ꢀ H]ꢀ at m/z 373 and 375.
data the ion m/z 371, which corresponds to the deprotonated
molecule of lignans 2, 7, and 18, it could be shown that at
least one lignan detected at this m/z eluted with the peaks at
approximately 25 and 35 min (Fig. 2(b)). By comparing the
retention times and the extracted ion chromatogram with
that obtained of the ions m/z 417 and 343 (deprotonated
molecules of 20 and 26, respectively), it was evident that the
peak at 25 min was due to lignans 2 and 20 and that the
lignan 26 was one of the components in the peak at 35 min
(Fig. 2(c) and (d)). To be able to separately observe lignan 7
and 18, the specific product ions generated in MS3 had to
be used since both compounds generate almost exclusively
m/z 356 in MS2 (loss of 15 Da from m/z 371). In MS3, lignan
7 produces an abundant product ion at m/z 205 (Fig. 2(e)),
while lignan 18 forms a product ion at m/z 151 (not shown).
The usefulness of this method was further demonstrated
by the analysis of an acetone/water extract (i.e. hydrophilic
lignans) of sesame seeds, which are known to be rich in
dietary lignans.3,29,30 In Fig. 3 the HPLC chromatogram of
the acetone/water extract is shown. The identification of
specific lignans in the extract was based on the retention
times, the observation of the deprotonated molecules and
fragmentation patterns in the ion trap. Several known lignan
constituents of sesame seeds24 (1, 5, 17, 21, 23, 24) were
identified along with the lignans 25 and 27, which have
been recently reported by Smeds et al.3 Sesaminol, a major
lignan in sesame seeds, was tentatively identified by its
fragmentation pattern ([M ꢀ H]ꢀ m/z 369 ! 219 [M ꢀ
150]ꢀ). In addition, isomeric structures of 27 and 25 were
detected as shown by the extracted ion chromatograms of
m/z 373 and 375 (Fig. 3). The chromatogram also contained
several compounds with the [M ꢀ H]ꢀ at m/z 357.
CONCLUSIONS
On the basis of the fragmentation patterns of reference
compounds, a method for the identification of different types
of lignans was developed. The CID fragmentation patterns
showed some general trends for different types of lignans. In
negative ion mode all butyrolactone lignans lost CO2 from the
lactone moiety, and the diols or lignans containing aliphatic
hydroxyl groups generally lost water. Lignans containing
hydroxymethyl groups lost formaldehyde easily. The most
generally observed fragmentation for all lignans containing
guaiacyl (3-hydroxy-4-methoxyphenyl) moieties was the loss
of a methyl radical. In positive ion mode, most lignans lost
water. Owing to extensive adduct formation and loss of
sensitivity upon further fragmentation, positive ion mode
detection was not as informative as negative ion detection.
Although several general fragmentation pathways for
different types of lignans were observed, there was large
variation in the fragmentation behavior between compounds
of the same class. Small changes in the structure, i.e. methy-
lation of a phenolic group or even a different diastereomer,
could lead to totally different fragmentation behavior. There-
fore, identification of a compound should not be based on the
observed fragmentation pattern alone. In combination with
retention times, molecular ion species (protonated or depro-
tonated) and compound-specific product ions produced in
the ion trap, the reference compounds used here, can be
identified unambiguously.
One should also bear in mind that the fragmentation
patterns and results reported here are specific for these
conditions. Changes in ionization, fragmentation or detection
conditions will certainly affect the results obtained.
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms

ESI-ion-trap analysis of lignans
107
14. Pen˜alvo JL, Haajanen KM, Botting N, Adlercreutz H. Quantifi-
cation of lignans in food using isotope dilution gas chro-
matography/mass spectrometry. Journal of Agricultural and Food
Chemistry 2005; 53: 9342.
Acknowledgements
Prof. Sansei Nishibe at the Department of Pharmacognosy, Univer-
sity of Hokkaido, Japan, and Dr. Stefan Willfo¨r at the Laboratory of
˚
Wood and Paper Chemistry, Abo Akademi University, are gratefully
15. Willfo¨r S, Ahotupa M, Hemming J, Reunanen M, Eklund P,
Sjo¨holm R, Eckerman C, Pohjamo S, Holmbom B. Antioxidant
activity of knotwood extractives and phenolic compounds of
selected tree species. Journal of Agricultural and Food Chemistry
2003; 51: 7600.
acknowledged for supplying reference compounds. The authors
also thank Ms Tove Sla¨tis, Laboratory of Organic Chemistry, Abo
˚
Akademi University, for running part of the LC-MS experiments.
This work was part of the activities at the Process Chemistry Cen-
˚
tre at Abo Akademi University within the Centre of Excellence
16. Eklund P, Sillanpa¨a¨ R, Sjo¨holm R. Synthetic transformation
of hydroxymatairesinol from Norway pruce (Picea abies)
to 7-hydroxysecoisolariciresinol, (C)-lariciresinol and (C)-
cyclolariciresinol. Journal of the Chemical Society, Perkin
Transactions 1 2002; 19: 1906.
17. Eklund P, Lindholm A, Mikkola J-P, Smeds A, Lehtila¨ R,
Sjo¨holm R. Synthesis of (ꢀ)-matairesinol, (ꢀ)-enterolactone, and
(ꢀ)-enterodiol from the natural lignan hydroxymatairesinol.
Organic Letters 2003; 5: 491.
Programme by the Academy of Finland.
REFERENCES
1. Smeds A, Hakala K. Liquid chromatographic-tandem mass
spectrometric methodforthe plantlignan7-hydroxymatairesinol
and its potential metabolites in human plasma. Journal of Chro-
matography B 2003; 793: 297.
2. Zhao Y, Chen B, Yao S. Simultaneous determination of abietine-
type diterpenes, flavonolignans and phenolic compounds
in compound preparations of Silybum marianum and Salvia
miltiorrhiza by HPLC-DAD-ESI MS. Journal of Pharmaceutical and
Biomedical Analysis 2005; 38: 564.
18. Eklund P, Sjo¨holm R. Oxidative transformation of the natural
lignan hydroxymatairesinol with 2,3-dichloro-5,6-dicyano-1,4-
benzoquinone. Tetrahedron 2003; 59: 4515.
19. Eklund P, Willfo¨r S, Smeds A, Sundell F, Sjo¨holm R.
A
new lariciresinol-type butyrolactone lignan derived from
hydroxymatairesinol and its identification in spruce wood.
Journal of Natural Products 2004; 67: 927.
3. Smeds AI, Eklund PC, Sjo¨holm RE, Willfo¨r SM, Nishibe S,
Deyama T, Holmbom BR. Quantification of a broad spectrum
of lignans in cereals, oilseeds, and nuts. Journal of Agricultural
and Food Chemistry 2007; 55: 1337.
4. Wang D, Liu Z, Gou M, Liu S. Structural elucidation and
identification of alkaloids in Rhizoma Coptidis by electrospray
ionization tandem mass spectrometry. Journal of Mass
Spectrometry 2004; 39: 1356.
5. Ye M, Yan Y, Guo D-A. Characterization of phenolic compounds
in the chinese herbal drug Tu-Si-Zi by liquid chromatography
coupled to electrospray ionization mass spectrometry. Rapid
Communications in Mass Spectrometry 2005; 19: 1469.
20. Willfo¨r S, Eklund P, Sjo¨holm R, Reunanen M, Sillanpa¨a¨ R, von
Shoultz S, Hemming J, Nisula L, Holmbom B. Bioactive phenolic
substances in industrially important tree species. Part 4:
identification of two new 7-hydroxy divanillyl bytyrolactol
lignans in some spruce, fir and pine species. Holzforschung 2005;
59: 413.
21. Eklund P, Holmstro¨m T, Al-Ubaydy L, Sjo¨holm R, Hakala J.
Rhamnosylation of lignans by a Streptomyces strain. Tetrahedron
Letters 2006; 47: 1645.
22. Ma¨kela¨ TH, Wa¨ha¨la¨ KT, Hase TA. ˛,ˇ-Dibenzyl-ꢂ-butyrolactone
lignan alcohols: total synthesis of (š)-70-hydroxyenterolactone,
(š)-70-hydroxymatairesinol and (š)-8-hydroxyenterolactone.
Steroids 2000; 65: 437.
6. Jungbluth G, Ternes W. HPLC separation of flavonols, flavones
and oxidized flavonols with UV-DAD-, electrochemical and ESI-
Ion trap MS detection. Fresenius Journal of Analytical Chemistry
2000; 367: 661.
7. Syrja¨nen K, Sipila¨ J, Bjo¨rk H, Brunow G. The identification of
dilignols from dehydrogenation mixtures of coniferyl alcohol
and apocynol [4-(1-Hydroxyethyl)-2-methoxyphenol] by LC-
MS/MS. Journal of Agricultural and Food Chemistry 2000; 48:
5211.
8. He X-G, Lian L-Z, Lin L-Z. Analysis of lignan constituents from
Schisandra chinensis by liquid chromatography-electrospray
mass spectrometry. Journal of Chromatography A 1997; 756: 81.
9. Ayres DC, Loike JD. Lignans: Chemical, Biological and Clinical
Properties. Cambridge University Press: Cambridge, USA; 1990.
10. Saarinen NM, Wa¨rri A, Ma¨kela¨ SI, Eckerman C, Reunanen M,
Ahotupa M, Salmi SM, Franke AA, Kangas L, Santti R.
Hydroxymatairesinol, a novel enterolactone precursor with
antitumor properties from coniferous tree (Picea abies). Nutrition
and Cancer 2000; 36: 207.
11. Kangas L, Saarinen N, Mutanen M, Ahotupa M, Hirsinummi R,
Unkila M, Pera¨la¨ M, Soininen P, Laatikainen R, Korte H,
Santti R. Antioxidant and antitumor effect of hydroxy-
matairesinol (HM-3000, HMR), a lignan isolated from the knots
of spruce. European Journal of Cancer Prevention 2002; 11:Suppl. 2
48.
12. Lee K-H, Xiao Z. Lignans in treatment of cancer and other
diseases. Phytochemistry Reviews 2003; 2: 341.
13. Milder IEJ, Arts ICW, Venema DP, Lasaroms JJP, Wa¨ha¨la¨ K,
23. Asano Y, Kamikawa T, Tororoyama T. A simple method for the
synthesis of lignan skeletons. Syntheses of (š) -parabenzlactone
and (š) -hinokinin. Bulletin of the Chemical Society of Japan 1976;
49: 3232.
24. Erdtman H. Constitution of resin phenols and their biogenetic
connections. I. Pinoresinol. Preliminary communication. Svensk
Kemisk Tidskrift 1934; 46: 229.
25. Willfo¨r S, Hemming J, Reunanen M, Eckerman C, Holmbom B.
Lignans and lipophilic extractives in Norway spruce knots and
stemwood. Holzforschung 2003; 57: 27.
26. Znamenskiy V,
Mariginean I,
Vertes A.
Solvated
ion
evaporation from charged water nanodroplets. Journal of Physical
Chemistry A 2003; 107: 7406.
27. Cuyckens F, Claeys M. Mass spectrometry in the structural
analysis of flavonoids. Journal of Mass Spectrometry 2004; 39:
1.
28. Crotti AEM, Fonseca T, Hong H, Staunton J, Galembeck SE,
Lopes NP, Gates PJ. The fragmentation mechanism of five-
membered lactones by electrospray ionization tandem mass
spectrometry. International Journal of Mass Spectrometry 2004;
232: 271.
29. Pen˜alvo JL, Heinonen SM, Aura A-M, Adlercreutz H. Dietary
sesamin is converted to enterolactone in humans. The Journal of
Nutrition 2005; 135: 1056.
Hollman PCH. Optimization of
a liquid chromatography-
30. Moazzami A, Kamal-Eldin A. Sesame seed is a rich source of
dietary lignans. Journal of American Oil Chemist Society 2006; 83:
719.
tandem mass spectrometry method for quantification of the
plant lignans secoisolariciresinol, matairesinol, lariciresinol, and
pinoresinol in food. Journal of Agricultural and Food Chemistry
2004; 52: 4643.
Copyright  2007 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2008; 43: 97–107
DOI: 10.1002/jms