F. He et al.
Table 1. Phenolic products of the autoxidation of flavan-3-ols detected by HPLC-MS/MS
Substrate
Peak No.
Product
Rt (min)
[M − H]− (fragment ions–MS2) (m/z)
(+)-Catechin
1
2
3
4
B-type dehydrodicatechin B11
Unknown
11.0
39.4
41.6
43.2
577 (533,439,425,393,269)
854 (792)
Unknown
628 (565,447)
741 (679)
Unknown
(−)-Epicatechin
5
6
A-type dehydrotricatechin
B-type dehydrotricatechin
(+)-Catechin
5.8
12.7
17.8
22.1
23.0
27.4
31.4
39.4
41.6
43.2
863 (725,711,699,575,425,411)
865 (727,713,681,577,425)
289 (245,205,151)
7
8
B-type dehydrodicatechin By
B-type dehydrodicatechin By
B-type dehydrodicatechin B22
A-type dehydrotricatechin
Unknown
577 (533,439,425,393,269)
577 (533,439,425,393,269)
577 (533,439,425,393,269)
863 (725,711,701,575,425,413)
854 (792)
9
10
11
2
3
Unknown
628 (565,447)
4
Unknown
741 (679)
(+)-Catechin and (−)-epicatechin
5
1
A-type dehydrotricatechin
B-type dehydrodicatechin B11
B-type dehydrodicatechin B12
B-type dehydrodicatechin By
B-type dehydrodicatechin B22
A-type dehydrotricatechin
Unknown
5.8
11.0
12.2
23.0
27.4
31.4
35.9
37.5
40.0
41.6
43.2
863 (725,711,699,575,425,411)
577 (533,439,425,393,269)
577 (533,439,425,393,269)
577 (533,439,425,393,269)
577 (533,439,425,393,269)
863 (725,711,701,575,425,413)
273 (229,187,139)
12
9
10
11
13
14
15
3
Unknown
591 (439,343,301,289,215)
591 (439,343,301,289,215)
628 (565,447)
Unknown
Unknown
4
Unknown
741 (679)
rings of the flavan-3-ol monomers.[14] The main fragment ions
of the potential autoxidation trimer also represented the similar
patterns of dissociation, because the m/z 727 ion to fragment
into m/z 439, 713 ion to generate into m/z 425, as well as m/z
681 ion into m/z 393 were all by losing a m/z 288 ion, which
corresponded to one monomeric flavan-3-ol by removing two
hydrogen atoms. These results indicated that flavan-3-ol units
in this trimer were linked with the same kind of IFL identified
in B-type dehydrodicatechin. Meanwhile, the appearance of the
fragmentionsatm/z 577and425suggestedthepossibilitythatthe
C6ꢁ-C8 IFL in the autoxidation trimer could be dissociated by the
striking effect (QM cleavage of IFL) to form its corresponding
dimers. Similar (+)-catechin trimers were synthesized in the
peroxidase-catalyzed and strawberry PPO catalyzed oxidation of
(+)-catechin, which were analyzed by reversed-phase (RP) HPLC
without coupling with MS.[5] In the present experiment, it was
the first time, to our knowledge, to synthesize and identify the
autoxidation trimeric flavan-3-ols of this unnatural type from the
autoxidationofflavan-3-ols,andthistrimerwasidentifiedasB-type
dehydrotricatechin. Scheme 2 displayed the detailed structure of
the [M − H]− molecular ion of the B-type dehydrotricatechin and
its partial fragment ions.
typical A-type procyanidin trimers, as shown in Fig. 2(b) and (c).
The appearance of the common fragment ions at m/z 845, 819,
801, 725, 711, 575 and 425 showed their similar but different
configurations with the identified B-type dehydrotricatechin,
suggesting the existence of additional IFL. According to the report
of Guyot et al., B-type dehydrodicatechin was easily oxidized into
A-type dehydrodicatechin by eliminating two hydrogen atoms
and forming an additional C–O–C IFL between two aromatic
rings, especially between the B-ring of the upper unit and
the A-ring of the lower unit.[12] On the other hand, if the
additional ether-type IFL (C–O–C bond) had no influence on
the C1ꢁ position on the B-ring of the first flavan-3-ol unit and
there was no more additional oxygen atom linked to the same
aromatic ring, the sixth main fragment ion should be at m/z
679 (681 minus 2), according to the dissociation rule we found
in the B-type dehydrotricatechin. However, we did not detect
this fragment ion in their – MS2 information, but obtained new
fragment ions at m/z 699 or 701. Therefore, according to the
basic structures of A-type dehydrodicatechins obtained by Guyot
et al., we hypothesized the detailed structures of the [M − H]−
molecular ions and partial fragment ions of the two trimeric
flavan-3-ols, and identified them as A-type dehydrotricatechins,
as shown in Scheme 3 (peak 5) and Scheme 4 (peak 11).[12] As
for peak 5, we concluded that one additional ether-type IFL
was formed between the C1ꢁ position on the B-ring of the first
unit and C7 position on the A-ring of the second unit, which
underwent benzopyran rearrangements during the autoxidation
and formed a quinone-methide structure of these rings. For peak
11, besides the ether-type IFL we mentioned before, in the first
flavan-3-ol unit an additional ether-type IFL was formed and
The other two trimers of flavan-3-ols, peaks 5 and 11 had
the same [M − H]− molecular ion at m/z 863 as those of the
natural A-type procyanidin trimer, which had been found in
various plants, for example, several natural A-type procyanidin
trimers isolated from plums and cinnamons and identified by
Gu et al.[10] Interestingly, the two new trimers almost had the
same fragment ions, at m/z 845, 819, 801, 725, 711, 699/701, 575,
425 and 411/413, which were totally different from those of the
c
Copyright ꢀ 2008 John Wiley & Sons, Ltd.
J. Mass. Spectrom. 2009, 44, 633–640