1382 Journal of Natural Products, 2010, Vol. 73, No. 8
Schuehly et al.
ipso position and a low-field shift of ∼6 ppm for the ortho and
para positions. Such changes found in carbon NMR shift values
indicated the acetylation at C-6 in subsystem D together with C-1
and C-6 in subsystem E (see Table 1). Selective inversion of the
proton resonance H-2/6 in subunit C led to an observed NOE
interaction with H-3 of subunit D, therefore supporting the
attachment of subunit C to subunit D at C-2 via an aryl ether
linkage. Similarly, selective inversion of the H-3 resonance in
subunit F led to an NOE for H-2/6 of subunit A, thus supporting
the attachment of subunit A to C-2 in subunit F. The experimental
results from acetylation as well as from NOE observations narrowed
the number of possible structures down to two, i.e., a structure with
an attachment of subunit E to C-1 or C-6 of subunit F, respectively.
The remaining C-1 or C-6 positions of subunit F, respectively, are
linked further to subunit B. Taking into account that subunit F must
be substituted nonsymmetrically, the only possible attachment point
for subunit E is that to C-6 in subunit F. This assignment was further
corroborated by a weak NOE signal observed between H-5 of
subunit F and H-3 of subunit E. Therefore, the structure of
garrettilignan A (1) was assigned as shown.
The molecular formula and the number of phenylpropanoids units
indicate the presence of a trimeric neolignan. From a biosynthetic
point of view, however, only subunits C and D, as well as E and
F, respectively, each resemble the known neolignan obovatol due
to the oxidation pattern of the phenolic ring and the ether linkage.
Hence, garrettilignan A (1) is most adequately described as a
substituted dineolignan bearing two additional 4-allylphenol moi-
eties (A and B) attached to subunit F. The relative configuration of
C-7 and C-8 of the propyl chain of subsystem E could not be
assigned in compound 1. The 1H and 13C NMR chemical shifts for
compounds 1 and 1a are presented in Table 1.
Di- and trimeric lignans and neolignans are rare natural
compounds. Besides occurring in gymnosperms, dilignans have
been occasionally found, for example, in Aizoaceae, Asteraceae,
Leguminosae, Myristicaceae, Rubiaceae, and Saururaceae,15-19
whereas dineolignans have been found in Magnoliaceae and
Saururaceae.20 In Magnoliaceae, which is known to be a rich source
of lignans of manifold structures as well as compounds of mixed
biosynthetic origin such as monoterpenyl or sesquiterpenyl lignans,
dineolignans have been reported from Magnolia officinalis Rehder
& Wilson and M. oboVata Thunb. () M. hypoleuca Sieb. &
Zucc.).21,22 A trineolignan, i.e., magnolianin, has been reported only
from the bark of M. oboVata.23 The occurrence of dineolignans
bearing additional 4-allylphenyl moieties as described herein for
garrettilignans A and B has not been reported previously.
Experimental Section
General Experimental Procedures. Optical rotations were measured
in MeOH on a Perkin-Elmer 341 polarimeter. UV-vis spectra were
recorded on a UV-160A spectrophotometer (Shimadzu). IR spectra were
taken as KBr pellets on a Perkin-Elmer 281 spectrophotometer. All
1D (1H and 13C) and 2D (COSY, HMBC, and HSQC) NMR spectra
were recorded at 298 K on a Varian Unity Inova 600 MHz spectrometer
using CDCl3 as solvent and referenced to TMS as internal standard.
EIMS were recorded on a Hewlett-Packard HP 6890 instrument fitted
with a HP 7890 detector. ESIMS for compounds 1, 2, and 6 were
measured in positive and negative mode on a Thermo Finnigan LQ
Deca XPPLUS mass spectrometer with autosampler using a SB-C18
Zorbax column (3.5 µm; 150 × 2.1 mm; Agilent Technologies) with
a guard column at a flow rate of 300 µL/min using an acetonitrile
gradient in water. ESI-MS spectra for compound 1a were recorded on
a MALDI Synapt HDMS System (Waters, Milford, MA) in positive
ion V time-of-flight mode using a LockSpray dual electrospray ion
source. Leu-enkephalin was used for lock-mass correction.
Compound 2 showed an ion peak of m/z 865.3418 [M - H2O
+ Na]+ using positive-mode HR-ESIMS, indicating a molecular
formula of C54H52O10 (calcd m/z 860.3560). As in compound 1, 1H
and 13C NMR spectra indicated the presence of six phenylpropanoid
units, suggesting an isomer of 1. Spin system assignments based
on DQF-COSY, HSQC, and HMBC experiments revealed the
presence of three 4-allylphenol subunits (A-C), two 4-allyl-1,2,6-
trihydroxyphenol subunits (D and F), and one 4-trihydroxypropyl-
1,2,6-trihydroxyphenol subunit (E). As with compound 1, nonsym-
metric substitution patterns in rings D-F of 2 were concluded from
High-Resolution LC-MS Analysis. High-resolution mass spectra
were obtained using an Agilent 1100 HPLC coupled to a JEOL
AccuTOF (JMS-T100LC) (Peabody, MA). All isolated compounds were
prepared in MeOH and injected directly into a 0.3 mL/min stream of
either MeOH or 80% MeOH/20% deionized H2O. A 20 µL sample
(approximately 0.1 mg/mL) was injected manually at 0.5 min, while
mass drift compensation standards [L-tryptophan (negative ion), PEG
(positive ion)] were injected at 1.5 min over the course of a 2 min run.
Semipreparative and analytical HPLC separations were performed
using an Agilent 1100 Series instrument equipped with a diode-array
detector. Compound mixtures were separated on an HPLC preparative
column packed with LiChrosorb RP-18 (7 µm, 250 × 10 mm, Merck,
Darmstadt). Analytical HPLC-DAD analysis was performed using a
SB-C18 Zorbax column (3.5 µm; 150 × 2.1 mm; Agilent Technologies)
equipped with a guard column at a flow rate of 300 µL/min and a
gradient elution program. Preparative HPLC was performed on a Varian
R PrepStar SD-1 with a Dynamax R solvent delivery system and UV
detector. For TLC analysis, precoated Si60 F254 plates (Merck) were
used. Detection was performed under UV light at 254 and 366 nm,
and visualization with spraying with vanillin-sulfuric acid reagent and
heating.
Acetylation of 1 or 2, respectively, was achieved by dissolving 20
mg of each of the compounds in 1 mL of absolute pyridine and adding
200 µL of acetic anhydride. The mixture was stirred at room temperature
overnight, poured into 8 mL of H2O, and then extracted with 2 mL of
CH2Cl2 (3×). The combined and dried CH2Cl2 layers were evaporated
to yield ca. 17 and 20 mg of crude 1a and 2a, respectively. The crude
compounds were then purified by semipreparative HPLC using CH3CN
(90 f 100% in H2O) to yield 7 and 8 mg of 1a and 2a, respectively.
Plant Material. Leaves of Magnolia garrettii Craib were collected
in August 2009 from a specimen growing in the temperate house of
the Botanical Garden in Graz. A voucher specimen is deposited at the
Herbarium of the Institute of Plant Sciences at the University of Graz.
Extraction and Isolation. The dried and powdered leaves (1280 g)
of M. garrettii were extracted with CH2Cl2 by percolation to yield a
residue of 34 g. About 40 g of silica gel (40-63 µm) was coated with
a portion (20 g) of the crude CH2Cl2 extract for fractionation (VLC)
using a gradient of n-hexane/EtOAc from 100% n-hexane within six
gradient steps of 5-10-15-20-50% f 100% EtOAc each using 500
mL of eluent, resulting in 15 fractions, V1-V15. Fractions V4-V14
1
the observation of six 13C and two H NMR shift values. HMBC
correlations between H-9 in subunit E and C-6 of subunit D proved
the linkage of these subunits. Selective inversion of proton
resonance H-3 in subunit D led to an observed NOE for H-2/6 of
subunit C, therefore indicating an attachment of subunit C to subunit
D at C-2 via an aryl ether linkage. Selective inversion of H-3 of
subunit F led to NOE signals for H-2/6 in subunit A, thus indicating
the attachment of subunit A to C-2 in subunit F. In addition, an
NOE signal was observed between H-5 in subunit F and H-3 in
subunit E. To further elucidate the linkage position for the different
subunits, the phenolic hydroxy positions were acetylated, resulting
in triacetylated compound 2a, as could be verified by the presence
1
of three acetyl signals in H and 13C NMR spectra. Differences
between the triacetylated compounds 1a and 2a were only seen in
the 13C shift values of rings D and E (Table 1). In comparison to
the signals found in compound 1a, differences in acetylation shifts
were only found in ring D, corroborating also the HMBC correlation
between subunits E and D. In conclusion, the difference between
the two isomeric compounds 1 and 2 was found to lie in the different
linkage between subunits E and D; that is, the obovatol moiety
comprising subunits D and C is attached via position C-6 to the
trihydroxypropyl chain of subunit E in compound 2, whereas it is
attached via position C-1 in compound 1. Taking these informations
1
together, garrettilignan B (2) was assigned as shown. The H and
13C NMR chemical shift data for compounds 2 and 2a are presented
in Table 1.