M. Grignon-Dubois, et al.
Phytochemistry 174 (2020) 112312
1
8–19, which is in agreement with the observed difference in molecular
precursors and coniferyl alcohol, followed by coupling of the two ra-
weight (162), due to loss of a glucose group (Table 2). These data
suggested that compounds 16 and 17 are O-glucoside of compounds 18
and 19, while 15 was assigned to tricin 7-O-glucoside (Kwon et al.,
dicals.
2
.3. Quantification of individual flavonoid compounds within the three
2
002).
species
Sub-fractions containing mixture of these products were isolated
1
and analysed by NMR. The H NMR spectra (DMSO‑d
6
) were very
The individual phenolics were quantified in the four crude extracts
S1–S3 and G) obtained from each species. Comparison of the results
complex due to overlapping signals, especially between 5.5 and
(
3
.0 ppm. However, two sets of four signals (intensity 2:1:1:1) and two
obtained with samples S1–S3 and G within a species showed the same
HPLC profile and a very weak variability in the phenolic concentra-
tions. In contrast, HPLC profiles, phenolic content and concentrations
varied substantially between the species (Fig. 4 and Table S5). Con-
sidering the low variability between samples S1–S3 and G of a given
species, the comparison between daughter and parental specimens has
been discussed only for samples G, which are consisting of 70 in-
dividuals and should ensure the best representation of their respective
meadow. However, the full set of data is available in supplementary
material (Table S5). Results obtained from samples G are summarized
in Table 1 (peaks given in order of elution on the HTec column).
sets of three multiplets (intensity 1:1:1) were observed in the aromatic
region, which suggested for 16–19 a tricin-type structure linked to a
phenolic ring. Additionally, the spectra of fractions dominated by
1
6–17 (F2) and 18–19 (F4) showed the presence of two sets of methoxy
group which differed in their intensity and chemical shift with the most
intense at about 3.82–3.85 ppm and the less at about 3.73–3.74 ppm.
This suggests that both flavone fragment and aromatic ring were sub-
stituted by methoxy groups.
Fraction F2 contains a mixture of 16 and 17 (ratio 36:64) along with
smaller amount of 15. A symmetrical tri-substituted B-ring was evident
from the two-proton singlets for H-2′ and H-6′ respectively at 7.35
(
compound 16) and 7.37 ppm (compound 17) in combination with
methoxyl resonances (s, 6 H) at 3.88 (16) and 3.85 ppm (17). Two sets
of coupled doublets (J = 2 Hz) at respectively 6.48 and 6.94 ppm
2.3.1. Comparison of the parental species
The two parental species were found to differ dramatically in their
foliar phenolic chemistry. Visual inspection of the HPLC profiles re-
vealed clear qualitative and quantitative differences in composition for
the two parent lines (Fig. 2, Table 1). A total of 15 individual flavonoids
(peaks 1–14 and 16) were quantified in the crude extract of S. alterni-
flora, whereas only seven (peaks 3–4 and 7–11) were found in S.
maritima. Six of the seven common products are flavone 6-C-mono and
diglucosides of luteolin (3, 4), apigenin (7, 10) and chrysoeriol (8, 11),
and one flavone O-glycoside of tricin (9), but their concentrations differ
considerably between the two parents. A large predominance of com-
pounds 7–8 and 10–11 was observed in the paternal parent S. maritima
(
compound 16) and 6.46 and 6.93 ppm (compound 17) were typical for
meta coupled proton H-8 and H-6, while the one-proton singlets at 7.16
compound 16) and 7.03 (compound 17) were assignable to H-3. Two
(
sets of three multiplets overlapped between 6.69 and 6.98 ppm and
singlets at 3.73 (16) and 3.75 ppm (17) were in agreement with a
guaiacyl ring. The aliphatic regions showed the presence of glucose-
and glycerol-type moieties. These data suggested that 16 and 17 were
composed of tricin linked at O-4’ by an identical guaiacylglycerol
fragment and differed only by the stereochemistry of the aliphatic part
of the molecules. The downfield shifts of the H-6 and H-8 signals in
compound 16 (respectively 6.48 and 6.94 ppm) and compound 17
−1
(respectively 2.265, 1.279, 1.363 and 0.695 mg g ), while compound
−1
(
6.46 and 6.93 ppm) suggest that the glucosyl fragment was attached to
9 was only present in low amount (0.104 mg g ). In contrast, com-
pounds 7–8 and 10–11 were found in lower amounts in the maternal S.
O-7. It was confirmed by the presence of resonances at 12.87 and
−
1
8
.78 ppm for compound 16, and 12.96 ppm and 8.75 for compound 17,
alterniflora (respectively 0.170, 0.540, 0.145, and 0.109 mg g ), while
compound 3 was about four time more abundant in S. alterniflora
which eliminate the possibilities of glycosylation at C-5 (flavone frag-
ment) or at C-4'' (guaiacyl fragment). In addition, since the ring B has to
be symmetrically substituted, the ether between tricin and the guaia-
cylglyceryl fragment can only be at C (4′)-O.
−
1
−1
(0.695 mg g ) than in S. maritima (0.157 mg g ). However, whatever
the parental species, flavone 6-C-monoglucosides (4, 10, and 11) were
found in lower concentrations than their respective 2″-O-glucosides (3,
7, and 8). Tricin 7-O-diglucuronic acid (9) was also common to the
parental species, but while it is the most abundant phenolic in S. al-
By comparing these data with literature reports, the diaster-
eoisomeric pair 16/17 was identified as the two epimers of tricin 4′-O-
−1
(
β-guaiacylglyceryl) ether 7-O-β-glucopyranoside (Bouaziz et al., 2002;
terniflora (3.942 mg g , 49% of the total), it is also the least abundant
−1
Bukhari et al., 2016; Colombo et al., 2005; Lee et al., 2015a), and
consequently 18–19 as the two epimers of tricin 4′-O-(β-guaiacylgly-
ceryl) ether. Compounds 18–19 were obtained in mixture with com-
pounds 17 and 20. Insufficient sample quantitites did not allow further
purification and the identity of the isomeric pair 18/19 could not be
fully confirmed by NMR. However, it was possible to analyze the aro-
in S. maritima (0.104 mg g , 2% of the total).
Eight flavonoids (1–2, 5–6, 12–14 and 16) were only found in the
maternal parent S. alterniflora. They were identified as schaftoside (1:
−1
−1
0.293 mg g ), isoschaftoside (2: 0.240 mg g ), isoorientin 2″-O- caf-
−1
feoylglucoside (5: 0.371 mg g ), isoorientin 2″-O- feruloylglucoside
−1
(6:
0.431 mg g ),
isoscoparin
2″-O-caffeoylglucoside
(12:
1
−1
−1
matic and methoxy regions in the H NMR spectra, and the results
0.253 mg g ), isoscoparin 2″-O-feruloylglucoside (13: 0.533 mg g ),
−1
confirmed assignment to the tricin 4′-O-(β-guaiacylglyceryl) ether pair
of isomers (Bouaziz et al., 2002; Huang et al., 2010; Lee et al., 2015b;
Nakajima et al., 2003; Syrchina et al., 1992; Wenzig et al., 2005) (Table
S3 in supplement materials).
isovitexin 2″-O-feruloylglucoside (14: 0.087 mg g ), and tricin 4′-O-
−1
guaiacylglyceryl ether 7-O-glucoside (16: 0.084 mg g ). The greater
flavonoid chemical diversity observed in S. alterniflora is mainly due to
the presence of the acylated forms of the flavone C-glucosides by caffeic
and/or ferulic acid (5–6 and 12–14). All together, they represent 21%
of the total flavonoid. Acylation of flavone is known to modulate the
physiological activity of the resulting flavonoid ester by changing so-
lubility, stability, reactivity and interaction with cellular targets
(Viskupicova et al., 2012 and references therein). It has been reported
to play a role in UV-B protection mechanisms (Kowalska et al., 2007).
In addition, five tricin derivatives (15, 17–20), were below the
detection level in the crude extracts, and their concentrations were not
determined in this study. Compounds 18–20 were found in both S.
maritima and S. alterniflora, whereas compounds 15–17 were only
found in S. alterniflora.
Due to overlap of signals in the 4–5 ppm region, it has been im-
3
possible to determine the J (H-7″, H-8″) coupling values to assign the
erythro/threo stereochemistry. However, compounds 16–19 were re-
spectively eluted at 17.7, 18.5, 24.6 and 24.9 min. Comparison of these
retention values with those reported by Bouaziz et al. (2002) in similar
analytical conditions (Rt for M 688: erythro 15.1 min, threo 15.6 min; Rt
for M 526: erythro 19.3 min, threo 19.8 min) suggest that compounds 16
and 18 might be more likely the erythro forms and compounds 17/19
the threo forms. NMR data are available in Table S4 (Supplementary
materials). According to the most accepted hypothesis, flavonolignans
are likely biosynthesized by oxidative radicalization of their flavonoid
7