Biomimetic synthesis of natural and "unnatural" lignans by oxidative coupling of caffeic esters

Article abstract of DOI:10.1002/ejoc.200900804

The metal-mediated oxidative coupling of caffeic acid esters has been employed in the biomimetic synthesis of dimeric lignans and neolignans. Phenethyl and methyl caffeate esters were used as substrates and MnO ₂, Mn(OAc)₃ and Ag

Full text of DOI:10.1002/ejoc.200900804

FULL PAPER
DOI: 10.1002/ejoc.200900804
Biomimetic Synthesis of Natural and “Unnatural” Lignans by Oxidative
Coupling of Caffeic Esters
Carmelo Daquino,[a] Antonio Rescifina,[a] Carmela Spatafora,[a] and Corrado Tringali*[a]
Keywords: Oxidation / Lignans / Biomimetic synthesis / Reaction mechanisms / Manganese
The metal-mediated oxidative coupling of caffeic acid esters
has been employed in the biomimetic synthesis of dimeric
lignans and neolignans. Phenethyl and methyl caffeate es-
ters were used as substrates and MnO2, Mn(OAc)3 and Ag2O
as oxidative coupling agents. The manganese-mediated re-
actions afforded in good yields the unusual benzo[kl]xan-
thene lignans 6 and 15 as the major products accompanied
by minor amounts of the aryldihydronaphthalene lignans
(Ϯ)-7 and (Ϯ)-16. When Ag2O was employed, the neolignan
(Ϯ)-17 was obtained as the major product. This biomimetic
route was also used to obtain the natural benzo[kl]xanthene
lignans rufescidride (9) and mongolicumin A (10). A compu-
tational study of the coupling reactions was also carried out.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
“natural” biosynthetic process. Thus, it is possible, in prin-
ciple, to obtain compounds unprecedented in the literature
that maintain a basic “natural” skeleton and possibly offer
a bioactivity profile similar to, or better than, that of a nat-
ural analogue. Owing to the lack of stereocontrol both in
metal- or enzyme-mediated phenolic radical coupling reac-
tions, racemic mixtures are frequently obtained.[9] Never-
theless, a number of interesting products have been ob-
tained in such a way and in some cases the bioactive race-
mate has been resolved to obtain the most active enanti-
omer. In this regard, one interesting example is the reported
dimerization of methyl caffeate (1) with Ag2O,[10] which af-
fords a neolignan racemate related to the natural dimer
3Ј,4-di-O-methylcedrusin (2). This latter was isolated as one
of the active compounds in “dragon’s blood”, the blood-
red latex produced by some Croton spp. growing in South
America and employed in traditional medicine for its
wound-healing and anticancer properties. The racemate was
then resolved to give the (2R,3R) enantiomer (3), which
shows promising antitumour properties against breast can-
cer cell lines and is more potent than its (2S,3S) enantiomer,
not only as an antiproliferative but also as an antiangiog-
enic agent.[5] In addition, the related neolignan racemate
(Ϯ)-4 [only the (2R,3R) enantiomer is reported] has been
found to be highly active against chloroquine-resistant Plas-
modium falciparum and Leishmania donovani.[4a]
Introduction
Lignans and related compounds (neolignans, oxyneolig-
nans and mixed lignans) are widely distributed within the
plant kingdom. In these plant secondary metabolites car-
bon skeletons are normally formed of two phenylpropanoid
(C6C3) units. Notwithstanding this relatively simple basic
skeleton, their structural diversity is really remarkable.
Their biosynthetic origin is the shikimate pathway: the term
“lignan”, originally introduced by Haworth[1] and related to
the woody tissue from which many of the first specimens
were obtained, refers to dimers generated by β–βЈ (8–8Ј)
oxidative coupling of two cinnamic acid residues. Accord-
ing to IUPAC recommendations,[2] “neolignans” are dimers
that originate from coupling other than 8–8Ј coupling, al-
though the term “lignan” is frequently employed in a
broader sense (including neolignan), as in the title of this
article. Lignans, neolignans and other related compounds
are accumulated in vascular plants as a chemical defence
system,[3] and this may explain their diverse biological ac-
tivities, including cytotoxic,[4] antimitotic, antileishman-
ial,[5] antiangiogenic,[6] cardiovascular[7] and antiviral ac-
tivity.[8] The biological activities and structural variety of
lignans and related compounds make them an attractive
target for chemical synthesis or modification. In particular,
biomimetic coupling reactions carried out on natural pre-
cursors may afford “unnatural” products through a radical
phenolic oxidative coupling mechanism that mimic the
The interesting biological properties of these natural and
synthetic neolignans, together with the increasing attention
devoted to antiangiogenic agents in cancer therapy,[11a,11b]
prompted us to carry out a biomimetic synthesis of new
“unnatural” neolignans as a continuation of our search for
bioactive analogues of natural products.[12] The results of
this work are reported below.
[a] Dipartimento di Scienze Chimiche, Università degli Studi di
Catania,
Viale A. Doria 6, 95125 Catania, Italy
Fax: +39-095-580138
E-mail: ctringali@unict.it
Eur. J. Org. Chem. 2009, 6289–6300
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6289

C. Daquino, A. Rescifina, C. Spatafora, C. Tringali
FULL PAPER
Scheme 1. Oxidative coupling of compounds 1 and 5.
Results and Discussion
to new cycles. The UV spectrum of 6 [λmax = 386 nm (ε =
11000 –1 cm–1)] was indicative of an extensively conjugated
system and the product exhibited intense fluorescence un-
der UV light (366 nm) with emission at λmax = 513 nm and
ΦF = 0.7 (referenced to anthracene). The IR spectrum
showed, as expected, intense absorption bands for the hy-
First, we selected CAPE (caffeic acid phenethyl ester, 5)
as the substrate for the oxidative coupling reaction. This
natural product is a component of propolis and it is re-
ported to be an antiinflammatory, antioxidant and antitu-
mour agent.[13a–13d] To the best of our knowledge, CAPE
had never been employed previously in phenolic oxidative
coupling reactions, so we expected unreported dimerization
products from this biomimetic coupling process. The reac-
tion was carried out initially on an analytical scale at room
temperature, employing MnO2 in dichloromethane as the
oxidative agent (Scheme 1). After 1 h, monitoring of the re-
action mixture by HPLC–UV showed the presence of two
main products 6 and 7, together with a large amount of
unreacted 5. By prolonging the reaction time up to 4 hours,
a higher product/substrate ratio was observed. Preparative
coupling afforded a green crude reaction mixture. To avoid
further oxidation of the products, the reaction was
quenched by the careful addition of ascorbic acid. After
DIOL silica gel purification two products, 6 (48%) and 7
(16.5%), were obtained, together with residual 5.
droxy and conjugated ester groups (ν
= 3532 and
˜
max
1716 cm–1, respectively). Acetylation of 6 afforded a perace-
tate 8, clearly demonstrated by the presence of three ace-
1
toxy groups in the ESI-MS and H NMR spectra, which
indicates that three acetylable hydroxy groups are present
in the structure of 6. Further data were obtained by NMR
experiments, namely NOE, HSQC and HMBC data (Fig-
ure 1).
Compound 6 showed a negative ESI-MS peak at m/z =
561 ([M – H]–), which suggested the formation of a tetra-
hydro dimer of 5. This was confirmed by elemental analysis,
which established the molecular formula C34H26O8, lacking
four hydrogen atoms with respect to a simple dimerization
product. The 1H and 13C NMR spectroscopic data for 6 are
reported in the Exp. Sect.; these data appeared immediately
not to be consistent with those of the most frequently re-
Figure 1. Selected COSY, NOE and HMBC correlations for com-
pound 6.
On the basis of these data, a careful literature search al-
ported caffeic dimers, such as the dihydrobenzofuran neo- lowed us to identify the conjugated nucleus of the dimer 6
1
lignans 3 and 4.[14]
as a benzo[kl]xanthene structure. In fact, the H and 13C
Both spectra showed some signals attributable to two NMR spectroscopic data of the natural benzoxanthene lig-
non-equivalent phenethyl moieties. In addition, further low- nans rufescidride (9), mongolicumin A (10) and yunnaneic
1
field signals were observed in the H NMR spectrum indi- acid H (11), isolated, respectively, from Cordia rufescens,[15]
cating a significant modification of the caffeic acid sub- Taraxacum mongolicum[16] and Salvia yunnanensis,[17] al-
structures. The 13C NMR spectroscopic data suggested that most superimpose those of the dimer 6, the structure of
two degrees of unsaturation could reasonably be attributed which was consequently established. Benzo[kl]xanthene lig-
6290
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 6289–6300

Biomimetic Synthesis of Natural and “Unnatural” Lignans
nans are rarely reported either as natural products or prod- aryldihydronaphthalene lignans.[9,18,22] The trans stereo-
ucts of oxidative coupling. Maeda et al. obtained benzo[kl]- chemistry is consistent with the diastereoselectivity ob-
xanthenes such as 12 and 13[18] in low yields by employing served in previously reported syntheses of aryldihydronaph-
caffeic acid derivatives as substrates and Ag2O as the oxi- thalene lignans by oxidative coupling in which the main
dative agent followed by acetylation of the dimeric product. product is normally a trans racemic mixture.[17,18]
The NMR spectroscopic data of the core structures of these
benzoxanthenes are also in perfect agreement with those
reported herein for compounds 6 and 8.
Figure 2. COSY and NOE correlations for compound (Ϯ)-7.
Once the structures of the products 6 and (Ϯ)-7 had been
established, we decided to check the general applicability of
this oxidative coupling reaction and at the same time to
verify whether the bulk phenethyl pendant could have a role
in the formation of the products. Thus, we carried out the
reaction under analogous conditions (MnO2 in CHCl3 at
room temperature) by employing methyl caffeate (1) instead
of CAPE (5) as the substrate. Purification and spectral
analysis of the products confirmed the expected formation
of the benzo[kl]xanthene lignan 15 as the main product
(isolated yield 51%) and the dihydronaphthalene lignan
(Ϯ)-16 as the minor product (isolated yield 7%; Scheme 1).
At this point we examined in detail the mechanism for
the formation of these products. Although the reaction
mechanisms previously reported for phenolic oxidative cou-
The minor product 7 showed a negative ESI-MS peak at pling reactions have only rarely been supported by experi-
m/z = 565 ([M – H]–), which suggests a dehydro dimer of mental data or calculations, the formation of aryldihydro-
5. The molecular formula C34H30O8 was determined by ele- naphthalene lignan analogues of 7 and 16 has been de-
mental analysis. The H and 13C NMR spectroscopic data scribed and the mechanism is shown in ref.[16], Scheme 2,
1
of 7 are reported in the Exp. Sect. It was evident that, in where only the (1R,2S) enantiomers are again reported.
addition to the two couples of phenethyl pendants and phe-
In this mechanism, after the formal removal of a hydro-
nolic rings, two new high-field signals were present [δ = gen atom from the p-phenolic group of 1 (or 5), a phenoxy
4.36/46.0 (1H/13C) and 3.86/48.3 ppm (1H/13C)] as part of radical with the unpaired electron at the β position may
the four-spin system C-1/C-4 established through a COSY couple with another β radical (8–8Ј coupling) to generate a
NMR experiment (Figure 2). A literature search rapidly reactive bis-quinonemethide A. Its tautomer B undergoes
suggested that 7 could have the structure of an aryldihyd- an intramolecular cyclization (6–7Ј), that is, an electrophilic
ronaphthalene lignan. This was confirmed by comparison aromatic substitution at the 6-position of the activated aro-
with the data reported for the natural lignan rabdosiin (14) matic ring,[23] to give the product (Ϯ)-7 [or (Ϯ)-16].
isolated from Rabdosia japonica.[19] A number of aryldihyd-
The formation of the benzo[kl]xanthene lignans 6 and 15
ronaphthalene lignans have previously been reported in the is less obvious. Indeed, only Maeda et al.[18] have previously
literature derived either from natural sources[20] or as prod- proposed possible mechanisms for the formation of benzo-
ucts of the phenolic oxidative coupling of phenylpropan- xanthene lignans, but these mechanisms cannot reasonably
oids.[21] A chiral HPLC analysis confirmed that 7 was, as explain the formation of 6 and 15. We examined various
expected, a racemic mixture. The trans stereochemistry of alternative hypotheses and we came to the conclusion illus-
the substituents at C-1 and C-2 in (Ϯ)-7 [only the (1R,2S) trated in Scheme 3, which has been corroborated by the cal-
enantiomer is reported in Scheme 1] was established on the culations discussed below. In the mechanism proposed by
basis of the J1,2 value (2.5 Hz), corroborated by NOE data us, both the dimeric products 6/15 and (Ϯ)-7/(Ϯ)-16 could
and in agreement with J values reported for other trans- originate by the same 8–8Ј coupling reaction, generating the
Eur. J. Org. Chem. 2009, 6289–6300
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6291

C. Daquino, A. Rescifina, C. Spatafora, C. Tringali
FULL PAPER
mediate EЈ undergoes further cyclization through nucleo-
philic attack of the 3-OH on the 5Ј-position of the qui-
nonemethide moiety, giving rise to the intermediate FЈ,
which then tautomerizes into the final products 6/15. To
support this proposed mechanism an in silico study at the
AM1 level was conducted starting from the intermediates
B and BЈ. The best results, in terms of activation energy
and product ratio, are in good agreement with the experi-
mental data. In Table 1 we report the formation enthalpies
of the optimized geometries of the eight possible transition
states (TSs) that originate from the 2–7Ј intramolecular
SEAr reaction of B and the analogous 6–7Ј SEAr reaction
of BЈ (Figure 3a and 3b). Of course, in these transition
states all E/Z geometries get lost. Only two among the eight
calculated TSs show a marked stability, namely 6–7Ј-ct (cis-
trans referring to the relative positions of 6-H, 7Ј-H and
8Ј-H), leading to compounds (Ϯ)-7/(Ϯ)-16 with 7Ј–8Ј trans
fusion, and 2–7Ј-ct (cis-trans referring to the relative posi-
tions of 2-H, 7Ј-H and 8Ј-H), affording the intermediate CЈ,
which evolves to the benzo[kl]xanthene products 6/15. The
optimized geometries of the most stable TSs for the pro-
posed SEAr mechanism are reported in Figure 4.
Scheme 2. Mechanism for the formation of compounds (Ϯ)-7 and
(Ϯ)-16.
quinonemethide intermediate AЈ. This is followed by a step
in which the only difference between the two mechanisms
is the position of the hydroxy group meta to the propenyl
chain, that is, “exo” in the formation of 7/16 (intermediate
B, see Scheme 2), whereas it is “endo” (intermediate BЈ) in
the formation of 6/15. Note here that A and AЈ are geomet-
rical isomers that could be present in solution together with
all the other possible E/Z stereoisomers. Intermediates B
and BЈ, from which the transition states originate, are con-
formational isomers and may be easily interchanged.
A 2–7Ј intramolecular cyclization step, very similar to
that reported in Scheme 2, gives rise to the aryldihydron-
aphthalene intermediate CЈ. Two further oxidative steps on
CЈ lead sequentially to the intermediates DЈ and EЈ. Inter-
Table 1. ∆Hf and Boltzmann distribution for the optimized geome-
tries of the eight possible transition states (TSs) originating from
the 6–7Ј and 2–7Ј intramolecular SEAr reaction of B.
TS
∆Hf [kcalmol–1]
Boltzmann distribution [%]
6–7Ј-ct
6–7Ј-tc
6–7Ј-tt
6–7Ј-cc
2–7Ј-ct
2–7Ј-tc
2–7Ј-tt
2–7Ј-cc
–202.97
–200.42
–200.31
–198.97
–203.29
–199.03
–198.98
–188.13
36.54
0.51
0.42
0.04
62.40
0.05
0.05
0.00
Scheme 3. Mechanism for the formation of compounds 6 and 15.
6292
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 6289–6300

Biomimetic Synthesis of Natural and “Unnatural” Lignans
The observation that dihydrobenzofuran neolignan ana-
logues of compounds 3 and 4 were not obtained, at least as
the main products, in the MnO2-mediated coupling reac-
tions with both caffeic esters 1 and 5 prompted us to carry
out the coupling reaction in the presence of Ag2O, which is
frequently reported in the literature as the oxidative agent
in the dimerization of phenylpropanoids.[24]
It was immediately apparent that the treatment of 5 with
Ag2O in CH2Cl2 at room temperature gives rise to a more
complex reaction mixture. A study of the dependence of the
conversion rate and yield on the time and oxidant/substrate
molar ratio (see Exp. Sect.) showed that with a molar ratio
of 1:2, the main product, different to 6, is formed after
90 min (yield 58%; conversion 68%).
Preparative coupling allowed the isolation and charac-
terization of this product as the dihydrobenzofuran neolig-
nan (Ϯ)-17 [Scheme 4; only the (2R,3R) enantiomer is re-
ported], obtained as a racemate. The spectroscopic data for
(Ϯ)-17 are in agreement with those of related dihydrobenzo-
furan neolignans previously reported in the literature.[25]
The expected trans stereochemistry of the substituents at C-
2 and C-3 was established on the basis of the J1,2 value
(7.0 Hz).[25]
Figure 3. a) 6–7Ј and b) 2–7Ј intramolecular SEAr reactions.
The result of the Ag2O-mediated coupling reaction
clearly indicates that the oxidative agent may play a major
role in directing the formation of either the dihydrobenzofu-
ran neolignans or the benzoxanthene lignans, yielding Ͻ1%
products in the presence of Ag2O. To establish whether the
MnO2-mediated reaction could be influenced by the pres-
ence of Mn2+ ions formed by the reduction of MnO2, we
carried out the coupling reaction of 5 with Ag2O in the
presence of Mn(ACAC)2. Under these conditions, a 54%
conversion of CAPE was observed and the main product
was again the benzoxanthene lignan 6 (51%). The aryldihy-
dronaphthalene lignan (Ϯ)-7 was detected in trace amounts,
whereas the neolignan (Ϯ)-17 was undetectable (Scheme 4).
These data strongly suggest that Mn2+ ions can stabilize
radicals derived from caffeic esters by complexation of the
o-hydroxyquinone form, which favours 8–8Ј coupling in-
stead of 5–8Ј coupling. The latter leads to dihydrobenzofu-
ran neolignans, according to a previously reported mecha-
nism,[18a] as detailed in Scheme 5.
Figure 4. Optimized geometries of the most stable transition states
for the proposed SEAr mechanism.
The relative percentages from the Boltzmannn popula-
tions of the more stable TS (62.40% for 2–7Ј-ct and 36.54%
for 6–7Ј-ct) are quite in agreement with the higher yields
observed for the benzo[kl]xanthene lignans 6 and 15. Note
here also that the 6–7Ј-ct TS, by far the most stable among
those leading to aryldihydronaphthalenes, produces the ste-
reoisomers (Ϯ)-7 and (Ϯ)-16 with trans stereochemistry of
the substituents at C-1 and C-2. This result is in perfect
agreement with all the literature data on the syntheses of
To corroborate this hypothesis, we performed a DFT
aryldihydronaphthalene lignans by oxidative coupling, computational study by utilizing methyl caffeate as the
which invariably report the diastereomer with trans substi- model compound starting from its most stable conforma-
tution as the main product.[22]
tion.[26] Methyl caffeate radicalization yields the 3-OH and
Scheme 4. Oxidative coupling of compound 5 with Ag2O and Ag2O/MnII(ACAC)2.
Eur. J. Org. Chem. 2009, 6289–6300
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6293

C. Daquino, A. Rescifina, C. Spatafora, C. Tringali
FULL PAPER
Scheme 5. Mechanism for the formation of compounds (Ϯ)-17.
Figure 5. Optimized geometries and spin density distributions for the methyl caffeate radical and its complexes with Ag+ and Mn2+. ∆E
in kcalmol–1.
4-OH radicals, the latter being the more stable (about
2.37 kcalmol–1) due to the more extended electronic delo- ergies for the radical coupling reactions are nearly zero,[27]
calization in the presence of the p-CH=CHCO2Me group. we compared the stability of all the possible diastereomeric
Then, taking into account the fact that the activation en-
Accordingly, complexes of the methyl caffeate radical at products arising from the 5–5Ј, 5–8Ј and 8–8Ј coupling reac-
the 4-position with both Mn2+ and Ag+ were optimized and tions of the methyl caffeate 4-OH radical to establish the
the spin density distributions of these radical cations are expected
percentage
distribution.
A
preliminary
reported in Figure 5 together with that of the parent caffe- Monte Carlo[28]-simulated annealing[29] was used to sample
ate radical. The values of ∆E for complex formation are geometries from a Boltzmannn-weighted distribution. The
also reported. In the case of the manganese complexes, the heating time was set to 10000 steps followed by a constant
high-spin complex is more stable than the low-spin one.
temperature simulation, that is, 400 K, with 1ϫ105 steps
Figure 5 clearly shows for the MnII complex II a dra- and finally a cooling time of 2ϫ104 steps. The geometries
matic variation of the spin density at C-5 (0.03) and C-8 obtained were selected within a range of 10.0 kcalmol–1 in
(0.42) in comparison with the uncomplexed radical I (0.15 terms of the energy difference between the isomers and in
and 0.25, respectively): the spin density at C-5 is reduced to the next step a full optimization of the selected geometries
almost zero whereas at C-8 it is strongly enhanced. This was carried out by using the semiempirical AM1 Hamilto-
implies a tendency towards 8–8Ј coupling. Conversely, in nian. The lowest ∆Hf values obtained for each series are
the AgI complex III, the densities at C-5 (0.11) and C-8 reported in Table 2. On the basis of these results, and con-
(0.33) are not significantly changed and, in principle, a mix- sidering the sum of each diastereoisomeric pair derived
ture of 5–5Ј, 5–8Ј and 8–8Ј coupling could occur. The for- from the same type of coupling, the 5–5Ј coupling should
mation of one main product by 5–8Ј coupling can be ex- be the predominant process, furnishing 64.65% of the ad-
plained on the basis of other arguments.
duct, followed by the 5–8Ј reaction with a yield of 31.87%
6294
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 6289–6300

Biomimetic Synthesis of Natural and “Unnatural” Lignans
and only traces of the 8–8Ј coupling (3.48%). Nevertheless, product 6. The highest yield of the minor product 7 (23.7%)
in the case of the 5–5Ј coupling of Ag+-complexed caffeate was obtained with an n-hexane/dichloromethane (1:1) mix-
radicals the two molecules are constrained to approach ture.
each other in a head-to-head fashion, and it is reasonable
that the repulsion between the two positively charged re-
gions could preclude the incoming reaction, thus causing
the formation, through 5–8Ј coupling, of the dihydrobenzo-
furan neolignan as the largely predominant product.
In the attempt to improve the conversion rate, we exam-
ined a further Mn-based oxidative agent, namely Mn-
(OAc)3. We studied the conversion and yield with respect
to the substrate. The oxidative agent ratio showed that the
benzoxanthene lignan 6 is obtained in 71% yield and 89%
conversion in 6 h (CHCl3, room temperature) with a 1:4
CAPE/Mn(AcO)3 ratio (Table 4). Under these conditions
the minor product 7 was obtained in 22% yield. Thus, this
appears the most convenient way to obtain 6, with the fur-
ther advantage of the simple use of the reagent and an eas-
ier work-up of the reaction mixture.
We applied this methodology to the synthesis of the nat-
ural benzoxanthene lignans mongolicumin A (10) and ru-
fescidride (9). First, we tried to obtain mongolicumin A by
direct oxidative coupling of caffeic acid in various solvent
systems, but the reaction was unsatisfactory due to the for-
mation of a complex mixture with low yields (Ͻ10%) of the
Table 2. ∆Hf and Boltzmann distribution for all the possible dia-
stereomeric products arising from 5–5Ј, 5–8Ј and 8–8Ј coupling re-
actions of the methyl caffeate radical.
Coupling product
∆Hf [kcalmol–1] Boltzmann distribution [%]
5–5Ј-(R,R)
5–5Ј-(R,S)
5–8Ј-(R,R)
5–8Ј-(R,S)
8–8Ј-(R,R)
8–8Ј-(R,S)
–231.709
–230.125
–231.104
–230.635
–229.557
–229.631
60.41
4.24
21.90
9.97
1.63
1.85
Finally, we tried to obtain a higher CAPE conversion benzo[kl]xanthene product. Thus we turned to a different
and a better yield of the unusual benzo[kl]xanthene lignan strategy (Scheme 6). In the first step, the benzo[kl]xanthene
6. First, we tested various solvent systems. In Table 3 we lignan 15 was subjected to alkaline hydrolysis to give 18.
report the results of this screening. Although toluene al- By subsequent cyclization under acidic conditions, this af-
lowed a higher conversion of CAPE, the highest yield of 6 forded rufescidride (9). A mild basic hydrolysis of 9 allowed
(72%) was obtained in chloroform. The more oxygenated mongolicumin A (10) to be obtained in a good overall yield
solvents gave both low conversion and low yield of the main (40% from caffeic acid).
Table 3. Conversions and yields of 6 and 7 in various solvent systems.[a]
Entry
Solvent
Conversion [%][b]
Yield [%] of 6[c]
Yield [%] of (Ϯ)-7[d]
1
2
3
4
5
6
7
8
CHCl3
n-hexane/CH2Cl2 (1:1)
CCl4
toluene
CH2Cl2
tBuOH
Et2O
dioxane
EtOAc
MeOH
THF
MeCN
Me2(CO)
67Ϯ2
79Ϯ2
82Ϯ5
92Ϯ3
68Ϯ1
25Ϯ7
38Ϯ6
34Ϯ4
45Ϯ6
39Ϯ2
42Ϯ2
45Ϯ4
32Ϯ1
72Ϯ2
57Ϯ3
55Ϯ6
53Ϯ5
48Ϯ4
20Ϯ5
11Ϯ1
10Ϯ1
7Ϯ2
7Ϯ3
4Ϯ1
4Ϯ3
2Ϯ1
2.8Ϯ0.3
23.7Ϯ2.1
21.8Ϯ2.4
3.2Ϯ0.6
16.5Ϯ0.9
3.6Ϯ0.7
3.3Ϯ0.6
nd[e]
9
1.5Ϯ0.3
1.2Ϯ0.2
n.d.[e]
10
11
12
13
n.d.[e]
n.d.[e]
[a] Reactions were performed at 298 K on the 7 m scale using a 1:3.2 oxidant (MnO2)/substrate ratio. The data were determined by
HPLC analysis (in triplicate) of the crude reaction mixture. [b] The data are referenced to 5 as the standard. [c] The data are referenced
to 6 as the standard. [d] The data are referenced to 7 as the standard. [e] n.d.: not detected.
Table 4. Conversion and yield with respect to the substrate/oxidant ratio.[a]
Entry
Substrate/oxidant
Conversion [%][b]
Yield [%] of 6[c]
Yield [%] of (Ϯ)-7[d]
1
2
3
4
5
6
1:0.5
1:1
1:2
1:3
1:4
56Ϯ3
61Ϯ6
72Ϯ7
83Ϯ6
89Ϯ4
97Ϯ4
30.6Ϯ2
33.8Ϯ2
57.4Ϯ3
63.5Ϯ3
70.9Ϯ4
48.1Ϯ2
20.0Ϯ2
23.3Ϯ2
28.9Ϯ1
26.5Ϯ4
21.9Ϯ3
8.3Ϯ1
1:8
[a] Reactions were performed in CHCl3 at 298 K using Mn(OAc)3 as the oxidative agent. The data were determined by HPLC analysis
(in triplicate) of the crude reaction mixture. [b] The data are referenced to 5 as the standard. [c] The data are referenced to 6 as the
standard. [d] The data are referenced to 7 as the standard.
Eur. J. Org. Chem. 2009, 6289–6300
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6295

C. Daquino, A. Rescifina, C. Spatafora, C. Tringali
FULL PAPER
antioxidant, antiallergic[30] and an inhibitor of Type I DNA
topoisomerase.[31] The dihydrobenzofuran (Ϯ)-17 is related
to bioactive neolignans such as compounds 2–4. Of these,
3 has been very recently reported as a promising cell death
inducer by modulating the mitochondrial pathway and G2/
M cell cycle arrest.[4] The benzoxanthenes 6 and 15 are re-
lated to compounds reported as potential antitumour
agents.[32] In addition, the benzoxanthene lignans are
strongly fluorescent both in solution and in the solid state
under UV light at 366 nm (Figure 6), and this property may
be of considerable interest in the design of new fluorescent
probes for biomedical applications.[33] In conclusion, we
think that these biomimetic syntheses will be useful for the
future exploitation and study of the above cited lignans and
neolignans or their analogues.
Scheme 6. Synthesis of compounds 18, 9 and 10.
Conclusions
The phenolic oxidative coupling reactions of CAPE (caf-
feic acid phenethyl ester, 5) and methyl caffeate (1) have
been carried out under various conditions. Among them
Mn(OAc)3 in CHCl3 afforded the new benzo[kl]xanthene
lignans 6 and 15 in good yields and allowed the synthesis
of the natural lignans mongolicumin A (10) and rufescid-
ride (9). Further, the previously unreported racemic trans-
dihydronaphthalene lignans (Ϯ)-7 and (Ϯ)-16 were ob-
tained from Mn-mediated caffeic ester dimerization. In the
presence of Ag2O the main product was the previously un-
reported racemic trans-dihydrobenzofuran neolignan (Ϯ)-
17. A mechanism for the formation of these dimers has
been proposed and corroborated by experimental data and
calculations. Note here that an “orientation” effect towards
8–8Ј coupling was observed in the presence of MnO2 and
Mn(OAc)3.
Figure 6. A reversed round-bottomed flask containing a solid sam-
ple of the benzo[kl]xanthene lignan 6, photographed under UV
light (366 nm).
In contrast to aryldihydronaphthalene lignans, benzo[kl]-
xanthene lignans are rarely encountered in nature. To the
best of our knowledge, this is the second reported synthesis
of a benzo[kl]xanthene lignan by oxidative coupling and the
first one to be synthesized in considerable yield. A further
point worth noting is that the benzo[kl]xanthene lignan
yunnaneic acid H (11) and the aryldihydronaphthalene lig-
nan rabdosiin (14), both dimers of rosmarinic acid and
structurally related to compounds 6 and (Ϯ)-7, respectively,
have previously been isolated by the same plant, Salvia yun-
Experimental Section
General: NMR spectra were recorded with a Varian Unity Inova
spectrometer operating at 499.86 (1H) and 125.70 MHz (13C) and
equipped with
a gradient-enhanced, reverse-detection probe.
Chemical shifts (δ) are referenced to TMS solvent signals. All
NMR experiments, including two-dimensional spectra, that is,
nanensis.[16] This is a strong indication of the biomimetic COSY, HSQC, HMBC and NOEDS, were performed by using soft-
ware supplied by the manufacturers and acquired at constant tem-
perature (298 K). UV/Vis spectra were recorded using a Perkin–
Elmer Lambda 25 spectrophotometer. The fluorescence spectrum
was recorded with a Spex Fluorolog-2 (mod. F-111) spectrofluo-
rimeter; the fluorescence quantum yield was obtained by using an-
thracene in CH2Cl2 as the standard (ΦF = 0.27). IR spectra were
recorded with a Perkin–Elmer Spectrum BX FT-IR System spec-
trometer. High-performance liquid chromatography (HPLC) was
carried out using an Agilent Series G1354A pump and an Agilent
UV G1314A detector. An auto-sampler Agilent Series 1100
nature of the synthesis reported herein and supports the
hypothesis that the natural products 10 and 14 are formed
in nature through a mechanism very similar to that pro-
posed here. The “unnatural” lignans 6 and (Ϯ)-7 are closely
related to known natural products and, in principle, may be
“natural” products biogenetically related to CAPE and not
yet discovered in nature.
The biomedical importance of the aryldihydronaph-
thalene lignans and dihydrobenzofuran neolignans has al-
ready been cited above. In particular, rabdosiin (14) is an G1313A was used for sample injection.
6296 Eur. J. Org. Chem. 2009, 6289–6300
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Biomimetic Synthesis of Natural and “Unnatural” Lignans
The qualitative and quantitative analyses of the reaction mixtures
were carried out by HPLC–UV using a reversed-phase column
(Luna® C18 column, 5 µm; 4.6ϫ250 mm; Phenomenex) and the
following solvent system: eluent A: H2O; eluent B: CH3CN. The
elution gradient had the following profile: t0min = 100% A, t40min
= 100% B, t50min = 100% B; the flow rate was 1 mLmin–1. The
chiral HPLC analyses of the racemates [(Ϯ)-7, (Ϯ)-16 and (Ϯ)-17]
were carried out by HPLC–UV using a Chiralpak® IA column
(5 µm; 4.6ϫ250 mm). Liquid chromatography–mass spectrometry
(LC–ESI-MS) was carried out with an HPLC Waters 1525 instru-
ment equipped with a Waters Micromass ZQ2000 mass spectrome-
ter. LiChroprep Si-60 and LiChroprep DIOL 25–40 (Merck) were
used as stationary phases for column chromatography. TLC was
carried out by using precoated silica gel F254 plates (Merck). Ce-
rium sulfate and phosphomolybdic acid were used as spray rea-
gents. All chemicals except MnO2 were of reagent grade and used
as purchased from Sigma–Aldrich. MnO2 was prepared as de-
scribed previously.[34] Elemental analyses were performed with a
Perkin–Elmer 240B microanalyser.
1716, 1599, 1453, 1192, 1144, 1086 cm–1. 1H NMR [500 MHz,
(CD3)2CO, 298 K]: δ = 8.08 (s, 1 H, 3-H), 7.47 [d, 3J(H,H) =
8.5 Hz, 1 H, 4-H], 7.4–7.2 (m, 10 H, 2ЈЈЈЈ-H, 3ЈЈЈЈ-H, 4ЈЈЈЈ-H, 5ЈЈЈЈ-
H, 6ЈЈЈЈ-H), 7.36 (overlapped with other signals, 1 H, 5-H), 7.28
(overlapped with other signals, 1 H, 11-H), 6.72 (s, 1 H, 8-H), 4.62
3
3
(t, JH,H = 9.0 Hz, 2 H, 1ЈЈЈ-H), 4.47 (t, JH,H = 9.0 Hz, 2 H, 1ЈЈ-
3
H), 3.10 (t, JH,H = 7.0 Hz, 2 H, 2ЈЈ-H, 2ЈЈЈ-H) ppm. 13C NMR
[125 MHz, (CD3)2CO, 298 K]: δ = 170.8 (2Ј-CO), 166.7 (3Ј-CO),
149.0 (C-7a), 148.3 (C-9), 142.6 (C-6), 142.1 (C-10), 138.9 (C-1ЈЈЈЈ),
137.5 (C-6a), 129.8 (C-2ЈЈЈЈ, C-6ЈЈЈЈ), 129.6 (C-3), 129.2 (C-3ЈЈЈЈ, C-
5ЈЈЈЈ), 127.5 (C-3a), 127.4 (C-4ЈЈЈЈ), 126.3 (C-11b), 125.1 (C-2),
123.9 (C-11c), 122.6 (C-1), 122.1 (C-4), 120.6 (C-5), 110.6 (C-11a),
112.1 (C-11), 104.4 (C-8), 66.6 (C-1ЈЈ, C-1ЈЈЈ), 35.7 (C-2ЈЈ), 34.9 (C-
2ЈЈЈ) ppm. For selected COSY, NOEDS and HMBC correlations
see Figure 1. MS (ESI): m/z = 561 [M – H]–. C34H26O8: calcd. C
72.59, H 4.66; found C 72.23, H 4.68.
Bis(2-phenylethyl) 1-(3,4-Dihydroxyphenyl)-6,7-dihydroxy-1,2-dihy-
dronaphthalene-2,3-dicarboxylate [(؎)-7]: Yellow-brown amorph-
ous powder. Rf (TLC) = 0.21 (8% MeOH/CH2Cl2). UV (EtOH):
λmax (ε) = 248 (15100), 330 (8620 –1 cm–1), nm. IR (CHCl ): ν =
˜
Theoretical Approach: Semiempirical and quantum mechanical cal-
culations were carried out at the AM1[35] and DFT[36] levels of
theory, respectively, in the gas phase. For DFT calculations the
B3LYP functional was employed which combines the three-coeffi-
cient-dependent hybrid functional for the exchange energy pro-
posed by Becke (B3)[37] with the correlation functional proposed
by Lee, Yang and Parr (LYP).[38] Pople’s 6-311G(d) basis set was
used for all elements except silver and manganese, for which the
effective core potential LANL2DZ was used.[39] Moreover, the un-
restricted open-shell approach in conjunction with the 6-311G(d)
basis set was used for radical species, allowing the best results to
be obtained both with regard to calculation time and correlation
with experimental data in the study of aroxyl radicals.[40] The doub-
let nature of the radicals was confirmed by the values of the S2
operator. Before annihilation of the first spin contaminant, the S2
values were found to be about 0.75 in all cases. The spin densities
for each atom were obtained by summing the diagonal terms of
the spin density matrix. For manganese complexes both high- and
low-spin states were considered. All calculations were carried out
with the Gaussian03 suite of programs.[41] For each molecule
studied, a full geometry optimization was performed with no con-
straints and the harmonic vibrational frequencies were calculated
at the same level of theory with the aim to characterize all struc-
tures as minima or saddle points. For the minima, all the wave-
numbers obtained were positive, whereas in the case of the transi-
tion states (TSs) only one wavenumber was imaginary. This unique
imaginary frequency, which is associated with the transition vec-
tor,[42] describes the atomic motion at the TS and can be used to
trace the intrinsic reaction coordinate[43] (IRC analysis) pathway
that connects the reactants and products.
3
3524, 2920, 1702, 1603, 1512, 1453, 1262, 1190 cm–1. 1H NMR
[500 MHz, (CD3)2CO, 298 K]: δ = 7.55 (s, 1 H, 4-H), 7.19–7.30 (m,
3
10 H, aromatic rings), 6.94 (s, 1 H, 5-H), 6.69 (d, JH,H = 8.5 Hz,
4
2 H, 5Ј-H), 6.59 (s, 1 H, 8-H), 6.45 (d, JH,H = 1.5 Hz, 1 H, 2Ј-H),
3
4
3
6.40 (dd, JH,H = 8.5, JH,H = 1.5 Hz, 1 H, 6Ј-H), 4.36 (d, JH,H
=
2.5 Hz, 1 H, 1-H), 4.28 (m, 2 H, 1ЈЈ-H)*, 4.19 (m, 2 H, 1ЈЈЈ-H)*,
3
3
3.86 (d, JH,H = 2.5 Hz, 1 H, 2-H), 2.94 (t, JH,H = 7.0 Hz, 2 H,
2ЈЈ-H)#, 2.83 (t, JH,H = 7.0 Hz, 2 H, 2ЈЈЈ-H)# ppm. 13C NMR
3
[125 MHz, (CD3)2CO, 298 K]: δ = 172.5 (3-CO2R)§, 166.9 (2-
CO2R)§, 147.9 (C-6), 145.4 (C-7), 144.8 (C-4Ј), 144.4 (C-3Ј), 138.9
(C-1ЈЈ), 138.3 (C-4), 136.1 (C-1Ј), 130.7 (C-8a), 129.7 (C-2ЈЈЈЈ, C-
6ЈЈЈЈ), 129.1 (C-3ЈЈЈЈ, C-5ЈЈЈЈ), 127.0 (C-4ЈЈЈЈ), 124.7 (C-4a), 122.9
(C-3), 119.5 (C-5), 116.9 (C-8), 116.6 (C-6Ј), 115.8 (C-5Ј), 115.2 (C-
2Ј), 65.9 (C-1ЈЈЈ)*, 65.6 (C-1ЈЈЈ)*, 48.3 (C-2), 46.0 (C-1), 35.6 (C-
2ЈЈЈ)#, 35.5 (C-2ЈЈЈЈ)# ppm. Values with identical superscripts may
be interchanged. MS (ESI): m/z = 565.4 [M – H]–. C34H30O8: calcd.
C72.07, H 5.34; found C 72.01, H 5.28. For NOEDS and COSY
correlations see Figure 2. Chiral HPLC analysis at 330 nm: Chi-
ralpak® IA (5 µm; 4.6ϫ250 mm), 2-propanol/hexane (30:60), flow
rate of 0.5 mLmin–1, retention time: 26.31, 30.45 min.
Acetylation of 6: Compound 6 (20 mg, 0.035 mmol) was dissolved
in pyridine (2 mL) and acetic anhydride (2 mL). The solution was
stirred at room temperature for 3 h and, after standard work-up,
the peracetate 8 was obtained in 95% yield.
Bis(2-phenylethyl) 6,9,10-Tris(acetyloxy)benzo[kl]xanthene-1,2-di-
carboxylate (8): Yellow amorphous powder. Rf (TLC) = 0.18 (100%
CH2Cl2). UV (CH2Cl2): λmax (ε) = 258 (18254), 376 (7689 –1
cm–1) nm. IR (CHCl ): ν = 2914, 1770, 1724, 1587, 1413, 1257,
˜
3
1172, 1016 cm–1. H NMR [500 MHz, (CD3)2CO, 298 K]: δ = 8.36
1
3
3
Reaction of CAPE with MnO2: A solution of CAPE (100 mg,
0.35 mmol) in CH2Cl2 (25 mL) was added to a stirred suspension
of MnO2 (300 mg, 3.5 mmol) in CH2Cl2 (25 mL). The reaction
mixture was then stirred at room temperature for 4 h and treated
with a saturated solution of ascorbic acid in methanol. After fil-
tration, the solvent was removed under reduced pressure and the
yellow-brown residue obtained was purified by column chromatog-
raphy over silica gel using a gradient of CH2Cl2/MeOH (from 0 to
4%) to provide 6 (42 mg, 42.7%) and 7 (12 mg, 12.1%).
(s, 1 H, 3-H), 7.71 (d, JH,H = 8.5 Hz, 1 H, 5-H), 7.53 (d, JH,H
=
8.5 Hz, 1 H, 4-H), 7.39–7.20 (m, 10 H, aromatic rings), 4.58 (t,
3
3JH,H = 7.5 Hz, 2 H, 1ЈЈ-H), 4.51 (t, JH,H = 7.5 Hz, 2 H, 2ЈЈЈ-H),
3
3
3.12 (t, JH,H = 7.0 Hz, 2 H, 2Ј-H), 3.09 (t, JH,H = 7.0 Hz, 2 H,
1Ј-H), 2.41, 2.33 and 2.31 (each s, 3 H, COCH3) ppm. 13C NMR
[125 MHz, (CD3)2CO, 298 K]: δ = 172.8, 169.7, 168.1, 167.8, 167.3,
165.1, 149.7, 143.5, 141.2, 138.1, 137.4, 134.9, 134.2, 130.8, 130.1,
128.8, 128.7, 128.4, 128.2, 127.2, 126.5, 126.3, 124.8, 124.3, 123.6,
123.4, 121.3, 120.2, 116.7, 112.4, 66.7, 65.9, 34.8, 34.1, 26.2,
22.5 ppm. MS (ESI): m/z = 687 [M – H]–.
Bis(2-phenylethyl) 6,9,10-Trihydroxybenzo[kl]xanthene-1,2-dicarb-
oxylate (6): Yellow amorphous powder. Rf (TLC) = 0.36 (8%
MeOH/CH2Cl2). UV (CH2Cl2): λmax (ε) = 271 (27700), 386
Synthesis of Methyl Caffeate (1): Concd. H2SO4 (2 mL) was added
to a solution of caffeic acid (1 g, 5.5 mmol) in MeOH (70 mL). The
resulting mixture was stirred for 1 h at reflux. After cooling at room
(11000 –1 cm–1) nm. IR (CHCl ): ν = 3522, 3346, 2952, 2925, 2852,
˜
3
Eur. J. Org. Chem. 2009, 6289–6300
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6297

C. Daquino, A. Rescifina, C. Spatafora, C. Tringali
FULL PAPER
temp., ethyl acetate (200 mL) was added to the mixture. The or-
ganic phase was washed with a NaHCO3 solution and saturated
brine, dried with anhydrous Na2SO4 and the solvent evaporated
5ЈЈ), 129.1 (C-3ЈЈЈ, C-5ЈЈЈ), 127.3 (C-5Ј), 127.1 (C-4ЈЈЈ), 118.6 (C-
6Ј), 117.8 (C-4), 117.0 (C-5), 116.1 (C-5Ј), 116.0 (C-8Ј), 113.8 (C-
2Ј), 87.8 (C-2), 66.6 (C-10Ј), 65.3 (C-10ЈЈ), 56.3 (C-3), 35.7 (C-7ЈЈ),
35.4 (C-7ЈЈЈ) ppm. MS (ESI): m/z = 589 [M + Na]+. C34H30O8:
calcd. C72.07, H 5.34; found C 75.25, H 5.30. Chiral HPLC analy-
sis at 330 nm: Chiralpak® IA (5 µm; 4.6ϫ250 mm), ethanol/hex-
ane (80:20), flow rate of 0.6 mLmin–1, retention time: 8.86,
10.96 min.
1
under vacuum to yield 1 (yield 99%). The H NMR spectroscopic
data for the product are in perfect agreement with literature
data.[44]
Reaction of Methyl Caffeate with MnO2: A solution of methyl caf-
feate (75 mg, 0.36 mmol) in CH2Cl2 (25 mL) was added to a stirred
suspension of MnO2 (300 mg, 3.5 mmol) in CH2Cl2 (25 mL). The
reaction mixture was then stirred at room temperature for 4 h and
then treated with a saturated solution of ascorbic acid in methanol.
After filtration, the solvent was removed under reduced pressure
and the yellow-brown residue obtained was purified by column
chromatography over silica gel using a gradient of CH2Cl2/MeOH
(from 0 to 4%) to give 15 (37 mg, 51%) and (Ϯ)-16 (5 mg, 7%).
Reaction of CAPE with Ag2O in the Presence of Mn(ACAC)2: A
solution of CAPE (4 mg, 0.014 mmol) in CHCl3 (10 mL) was
added to a stirred suspension of Ag2O (13 mg, 0.056 mmol) and
Mn(ACAC)2 (14 mg, 0.140 mmol) in CHCl3 (2 mL). The reaction
mixture was then stirred at room temperature for 5 h. Then, a satu-
rated solution of ascorbic acid in MeOH was added. The solvent
was evaporated under vacuum to yield a residue which was ana-
lysed by HPLC.
Dimethyl
6,9,10-Trihydroxybenzo[kl]xanthene-1,2-dicarboxylate
(15): Yellow amorphous powder. Rf (TLC) = 0.57 (8% MeOH/
Reactions of CAPE with Mn(OAc)3: Mn(OAc)3 (substrate/oxidative
CH2Cl2). UV (MeOH): λmax (ε) = 270 (24254), 393 (11051 agent ratio 1:0.5, 1:1, 1:2, 1:3, 1:4, 1:8) was added to six aliquots
–1 cm–1) nm. IR (Nujol): ν = 3288, 2715, 1712, 1674, 1598, 1461,
(2 mL) of a solution of CAPE (40 mg, 0.07 mmol) dissolved in
˜
1377, 1274, 1153, 1087 cm–1. 1H NMR [500 MHz, (CD3)2CO, CHCl3 (20 mL). For each aliquot the reaction was replicated three
3
298 K]: δ = 8.17 (s, 1 H, 3-H), 7.53 (d, JH,H = 8.5 Hz, 1 H, 4-H), times. The reaction mixtures were stirred for 6 h at room tempera-
3
7.33 (d, JH,H = 8.5 Hz, 1 H, 5-H), 7.19 (s, 1 H, 11-H), 6.72 (s, 1
ture (the reactions were monitored at regular time intervals by
TLC). Then a saturated solution of ascorbic acid in MeOH was
H, 8-H), 3.97 (s, 3 H, OCH3), 3.88 (s, 3 H, OCH3) ppm. 13C NMR
[125 MHz, (CD3)2CO, 298 K]: δ = 171.2 (2-CO), 166.9 (1-CO), added. After filtration the solvent was evaporated in vacuo and the
149.0 (C-7a), 147.7 (C-9), 142.7 (C-6), 142.5 (C-10), 137.5 (C-6a), residue was dissolved in MeOH (10 mL)/H2O (200 mL). Sub-
129.6 (C-3), 127.4 (C-3a), 125.6 (C-11b), 124.9 (C-2), 124.1 (C-11c), sequently the mixture was extracted twice with EtOAc. The organic
122.1 (C-4), 121.6 (C-2), 120.6 (C-5), 110.7 (C-11a), 112.5 (C-11), phase, after washing with a solution of NaHCO3 and saturated
104.6 (C-8), 52.8 and 52.5 (OCH3) ppm. MS (ESI): m/z = 381 [M – brine, was dried with anhydrous Na2SO4 and the solvent evapo-
H]–. C20H14O8: calcd. C 62.83, H 3.69; found C 62.75, H 3.72.
rated under vacuum to yield a residue which was analysed by
HPLC. The results are reported in Table 4.
Dimethyl 1-(3,4-Dihydroxyphenyl)-6,7-dihydroxy-1,2-dihydronaph-
thalene-2,3-dicarboxylate [(؎)-16]: Yellow amorphous powder. Rf
(TLC) = 0.47 (8% MeOH/CH2Cl2). IR, UV, MS and NMR spec-
troscopic data are in agreement with literature data.[19] Chiral
Reaction of Caffeic Acid with Mn(OAc)3: Mn(OAc)3 (23 mg,
0.086 mmol, substrate/oxidative agent ratio 1:4) was added to nine
stirred aliquots of caffeic acid (4 mg, 0.021 mmol) in chloroform
HPLC analysis at 250 nm: Chiralpak® IA (5 µm; 4.6ϫ250 mm), 2- (2 mL), acetone or ethyl acetate (for each solvent the reaction was
propanol/hexane (30:60), flow rate of 0.5 mLmin–1, retention time:
12.61, 25.26 min.
replicated three times). The reaction mixtures were stirred for 8 h
at room temperature, after which the solvent was removed in vacuo
and a saturated solution of ascorbic acid in MeOH (2 mL) was
added to all samples. The filtrate solutions were analysed by HPLC.
The averages and standard deviations of the caffeic acid conver-
sions and mongolicumin A (10) yields are as follows: chloroform:
conversion 48Ϯ5%, yield 3Ϯ2%; acetone: conversion 93Ϯ3%,
yield 6Ϯ2%; ethyl acetate: conversion 74Ϯ5%, yield 9Ϯ3%.
Reaction of CAPE with Ag2O: A solution of CAPE (200 mg,
0.704 mmol) in CHCl3 (10 mL) was added to a stirred suspension
of Ag2O (163 mg, 0.703 mmol) in CHCl3 (10 mL). The reaction
mixture was then stirred at room temperature for 2 h. After fil-
tration, the solvent was removed under reduced pressure and the
dark-brown residue obtained was purified by column chromatog-
raphy over silica gel using a gradient of CH2Cl2/MeOH (from 0 to
8%) to provide (Ϯ)-17 (35 mg, 17.6%).
Synthesis of 18: Compound 15 (50 mg, 0.13 mmol) was dissolved
under nitrogen in 2  aqueous potassium hydroxide (10 mL). The
solution was stirred for 2 h at room temperature. Then the solution
was acidified with 2  HCl and extracted with ethyl acetate, the
organic layer was washed with water and saturated brine and dried
with Na2SO4. The solvent was removed under reduced pressure to
obtain 18 (48 mg, 99%) as a brown powder. Rf (TLC-RP18) = 0.18
(50% MeOH/H2O). UV (MeOH): λmax (ε) = 263 (18100), 392
2-Phenylethyl 2-(3,4-Dihydroxyphenyl)-7-hydroxy-5-[(1E)-3-oxo-3-
(2-phenylethoxy)prop-1-enyl]-2,3-dihydro-1-benzofuran-3-carboxyl-
ate [(؎)-17]: Yellow-brown amorphous powder. Rf (TLC) = 0.55
(8% MeOH/CH2Cl2). UV (MeOH): λmax (ε)
=
330
(17705 –1 cm–1) nm. IR (CHCl ): ν = 3536, 3030, 2955, 1733, 1702,
˜
3
1619, 1607, 1496, 1263, 1170, 1135 cm–1. 1H NMR [500 MHz,
(CD3)2CO, 298 K]: δ = 8.20 (br. s, 2 H, 3-OH, 4-OH), 7.52 (d, 3JH,H
= 16.0 Hz, 7Ј-H), 7.35–7.20 (m, 10 H, aromatic rings), 7.14 (d,
3JH,H = 1.5 Hz, 1 H, 6-H), 6.92 (d, 3JH,H = 1.5 Hz, 1 H, 4-H), 6.89
(d, 4JH,H = 2.0 Hz, 1 H, 2Ј-H), 6.87 (d, 3JH,H = 8.0 Hz, 1 H, 5Ј-H),
(6301 –1 cm–1) nm. IR (Nujol): ν = 3468, 3389, 1707, 1681, 1258,
˜
1190 cm–1. 1H NMR (500 MHz, CD3OD, 298 K): δ = 8.14 (s, 1 H,
3
3
3-H), 7.38 (d, JH,H = 8.5 Hz, 1 H, 4-H), 7.23 (d, JH,H = 8.5 Hz,
1 H, 5-H), 7.09 (s, 1 H, 11-H), 6.70 (s, 1 H, 8-H), 4.02 (s, 3 H,
OCH3) ppm. 13C NMR (125 MHz, CD3OD, 298 K): δ = 172.3,
168.0, 148.2, 146.9, 141.6, 141.5, 136.9, 128.8, 128.6, 126.2, 125.2,
123.2, 124.2, 121.3, 120.2, 111.6, 109.7, 103.9, 51.8 ppm. Selected
NOESY correlation: 4.02 and 7.09. MS (ESI): m/z = 367 [M –
H]–. C19H12O8: calcd. C 61.96, H 3.28; found C 61.75, H 3.32.
3
4
3
6.76 (dd, JH,H = 8.0, JH,H = 2.0 Hz, 1 H, 6Ј-H), 6.26 (d, JH,H
=
=
3
3
16.0 Hz, 8Ј-H), 5.95 (d, JH,H = 7.0 Hz, 2-H), 4.46 (t, JH,H
3
7.0 Hz, 2 H, 10Ј-H), 4.41 (t, JH,H = 7.0 Hz, 2 H, 10-H), 4.31 (d,
3JH,H = 7.0 Hz, 3-H), 3.03 (t, JH,H = 7.0 Hz, 4 H, 11-H, 11Ј-
3
H) ppm. 13C NMR [125 MHz, (CD3)2CO, 298 K]: δ = 171.0 (C-8),
167.1 (C-9Ј), 149.9 (C-7), 146.3 (C-4Ј), 146.1 (C-3Ј), 145.4 (C-7Ј), Synthesis of Rufescidride (9): Compound 18 (33 mg, 0.09 mmol)
142.4 (C-7a), 139.2 (C-1ЈЈ), 138.9 (C-1ЈЈЈ), 132.6 (C-1Ј), 129.7 (C- was dissolved in 2  sulfuric acid solution (50 mL) and warmed at
2ЈЈ, C-6ЈЈ), 129.68 (C-5), 129.6 (C-2ЈЈЈ, C-6ЈЈЈ), 129.2 (C-3ЈЈ, C-
70 °C for 4 h. The aqueous phase was extracted with ethyl acetate
6298
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 6289–6300

Biomimetic Synthesis of Natural and “Unnatural” Lignans
P. Eun-Jung, H. Ji-Young, L. S. Kook, J. Nutr. Biochem. 2006,
17, 356–362.
(4ϫ50 mL) and then the organic layer was washed with water and
saturated brine and dried with Na2SO4. The solvent was removed
under reduced pressure to give 9 (30 mg) as a red powder in quanti-
tative yield. The ESI-MS spectrum and NMR spectroscopic data
are in perfect agreement with literature data.[45]
[14] C. Spatafora, C. Tringali, Targets in Heterocyclic Systems.
Chemistry and Properties, 2007, vol. 11, pp. 284–312.
[15] S. A. Souza da Silva, A. L. Souto, M. de Fatima Agra, E. V.
Leitao da-Cunha, J. M. Barbosa-Filho, M. Sobral da Silva, R.
Braz-Filho, ARKIVOC 2004, 6, 54–58.
Synthesis of Mongolicumin
A (10): Compound 9 (23 mg,
[16] S. Shi, Y. Zhang, K. Huang, S. Liu, Y. Zhao, Food Chem. 2008,
0.06 mmol) was dissolved in ethyl acetate (50 mL) and added under
nitrogen to an aqueous solution of 2  sodium hydrogen carbonate
(100 mL). The solution was vigorously stirred for 2 h. Then the
organic layer, containing 10, after standard work-up, was recuper-
ated and furnished 9 (5 mg, conversion 80%). The aqueous phase
was acidified with 2  HCl until pH ≈ 2 and extracted with ethyl
acetate (3ϫ100 mL). The organic layer was washed with a solution
of NaHCO3 and saturated brine and dried with Na2SO4. The sol-
vent was removed under reduced pressure to give 10 (18 mg, 95%)
as a yellow-brown powder. The ESI-MS spectrum and NMR spec-
troscopic data are in perfect agreement with literature data.[46]
108, 402–406.
[17] T. Tanaka, A. Nishimura, Y. Kouno, G. Nonaka, C.-R. Yang,
Chem. Pharm. Bull. 1997, 45, 1596–1600.
[18] a) S. Maeda, H. Masuda, T. Tokoroyama, Chem. Pharm. Bull.
1994, 42, 2506–2513; b) S. Maeda, H. Masuda, T. Tokoroyama,
Chem. Pharm. Bull. 1995, 43, 935–940.
[19] I. Agata, T. Hatano, S. Nishibe, T. Okuda, Chem. Pharm. Bull.
1988, 36, 3223–3225.
[20] I. Agata, T. Hatano, S. Nishibe, T. Okuda, Phytochemistry
1989, 28, 2447–2450.
[21] R. S. Ward, Chem. Soc. Rev. 1982, 11, 75–125.
[22] M. Orlandi, B. Rindone, G. Molteni, P. Rummakko, G. Bru-
now, Tetrahedron 2001, 57, 371–378.
[23] A. Pelter, R. S. Ward, R. R. Rao, Tetrahedron 1985, 41, 2933–
Acknowledgments
2938.
[24] a) M. Bruschi, M. Orlandi, B. Rindone, P. Rummakko, L.
Zoia, J. Phys. Org. Chem. 2006, 19, 592–596; b) G. Lemière,
M. Gao, A. De Groot, R. Dommisse, J. Lepoivre, L. Pieters,
V. Buss, J. Chem. Soc. Perkin Trans. 1 1995, 1775–1779.
[25] S. M. O. Van Dyck, G. L. F. Lemière, T. H. M. Jonckers, R.
Dommisse, Molecules 2000, 5, 153–161.
[26] N. F. L. Moncado, R. Calheiros, S. M. Fiuza, F. Borges, A. Ga-
spar, J. Garrido, M. P. Marques, J. Mol. Model. 2007, 13, 865–
877.
This research was supported by a grant from the Università degli
Studi di Catania (Progetti di Ricerca di Ateneo, Catania, Italy) and
by Ministero dell’Università e della Ricerca (MIUR) (PRIN 2007,
Rome, Italy). The authors acknowledge Prof. Salvatore Sortino
(Università degli Studi di Catania) for determination of the fluores-
cence data. CPU time from the Italian supercomputing center CI-
NECA is also acknowledged.
[27] C. J. Houtman, Holzforschung 1999, 53, 585–589.
[28] N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. N.
Teller, E. Teller, J. Chem. Phys. 1953, 21, 1087–1092.
[29] S. Kirkpatrick, C. D. Gelatt, M. P. Vecchi, Science 1983, 220,
671–680.
[1] R. D. Haworth, J. Chem. Soc. 1942, 448–456.
[2] a) G. P. Moss, Pure Appl. Chem. 2000, 72, 1493–1523; b) http://
www.chem.qmul.ac.uk/iupac/lignan/.
[3] D. R. Gang, A. T. Dinkova-Kostova, L. B. Davin, N. G. Lewis,
Phylogenetic links in plant defense systems: lignans, isoflavon-
oids, and their reductases, in: Phytochemicals for Pest Control,
ACS Symposium Series, 1997, vol. 658, pp. 58–89.
[4] J. S. Bose, V. Gangan, R. Prakash, S. K. Jain, S. K. Manna, J.
Med. Chem. 2009, 52, 3184–3190.
[5] a) S. Apers, A. Vlietinck, L. Pieters, Phytochem. Rev. 2004, 2,
201–207; b) S. Van Miert, S. Van Dyck, T. J. Schmidt, R. Brun,
A. Vlietinck, G. Lemiere, L. Pieters, Bioorg. Med. Chem. 2005,
13, 661–669; c) A. K. Prasad, V. Kumar, P. Arya, S. Kumar, R.
Dabur, N. Singh, A. K. Chhillar, G. L. Sharma, B. Ghosh, J.
Wengel, C. E. Olsen, V. S. Parmar, Pure Appl. Chem. 2005, 77,
25–40.
[6] S. Apers, D. Paper, J. Buergermeister, S. Baronikova, S.
Van Dyck, G. Lemiere, A. Vlietinck, L. Pieters, J. Nat. Prod.
2002, 65, 718–720.
[7] E. L. Ghisalberti, Phytomedicine 1997, 4, 151–166.
[8] J. L. Charlton, J. Nat. Prod. 1998, 61, 1447–1451.
[9] L. B. Davin, H.-B. Wang, A. L. Crowell, D. L. Bedgar, D. M.
Martin, S. Sarkanen, N. G. Lewis, Science 1997, 275, 362–366.
[10] L. Pieters, S. Van Dyck, M. Gao, R. Bai, E. Hamel, A. Vliet-
inck, G. Lemiere, J. Med. Chem. 1999, 42, 5475–5481.
[11] a) R. Kerbel, J. Folkman, Nat. Rev. Cancer 2002, 2, 727–739;
b) Y. Crawford, N. Ferrara, Cell Tissue Res. 2009, 335, 261–
269.
[12] a) V. Cardile, L. Lombardo, C. Spatafora, C. Tringali, Bioorg.
Chem. 2005, 33, 22–33; b) S. Grasso, L. Siracusa, C. Spatafora,
M. Renis, C. Tringali, Bioorg. Chem. 2007, 35, 137–152.
[13] a) F. M. Da Cunha, D. Duma, J. Assreuy, F. C. Buzzi, R. Ni-
ero, M. M. Campos, J. B. Calixto, Free Radical Res. 2004, 38,
1241–1253; b) Y.-T. Lee, M.-J. Don, P.-S. Hung, Y.-C. Shen, Y.-
S. Lo, K.-W. Chang, C.-F. Chen, L.-K. Ho, Cancer Lett. 2005,
223, 19–25; c) X. Debing, W. Dong, H. Yujun, X. Jiayin, Z.
Zhaoyang, L. Zengpeng, X. Jiang, Anti-Cancer Drugs 2006, 17,
753–762; d) H. H. Jin, P. H. Joo, C. Hwa-Jin, M. Hye-Young,
[30]
[31]
[32]
[33]
H. Ito, T. Miyazaki, M. Ono, H. Sakurai, Bioorg. Med. Chem.
1998, 6, 1051–1056.
Y. Kashiwada, K. F. Bastow, K.-H. Lee, Bioorg. Med. Chem.
Lett. 1995, 5, 905–908.
N. C. Birnberg, Q. Weng, H. Liu, J. Avruch, J. Kyriakis, PCT
Int. Appl. 2004, 36 pages; [Chem. Abstr. 2004, 141, 1222–1258].
L. Patsenker, A. Tatarets, O. Kolosova, O. Obukhova, Y. Povro-
zin, I. Fedyunyayeva, I. Yermolenko, E. Terpetschnig, Ann. N.
Y. Acad. Sci. 2008, 1130, 179–187.
[34]
[35]
C. Daquino, M. C. Foti, Tetrahedron 2006, 62, 1536–1547.
a) M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P. Stewart,
J. Am. Chem. Soc. 1985, 107, 3902–3909; b) M. J. S. Dewar, C.
Jie, J. Yu, Tetrahedron 1993, 49, 5003–5038.
[36]
[37]
P. Hohenberg, W. Kohn, Phys. Rev. B 1964, 136, 864–871.
a) A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100; b) A. D.
Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[38]
[39]
C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283; P. J.
Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 284–298; P. J. Hay,
W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[40]
[41]
W. Bors, C. Michel, K. Stettmaier, S. P. Kazazic´, L. Klasinc,
Croat. Chem. Acta 2002, 75, 957–964.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Mont-
gomery Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas,
J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson,
P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rab-
uck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Or-
tiz, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komar-
omi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A.
Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M.
Eur. J. Org. Chem. 2009, 6289–6300
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6299

C. Daquino, A. Rescifina, C. Spatafora, C. Tringali
FULL PAPER
Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Re-
plogle, J. A. Pople, Gaussian03, Revision D.02, Gaussian, Inc.,
Wallingford, CT, 2004.
[45]
S. A. Souza da Silva, A. Lopes Souto, M. de Fátima Agra,
E. V. Leitão da-Cunha, J. M. Barbosa-Filho, M. Sob-
ral da Silva, R. Braz-Filhoc, ARKIVOC 2004, 6, 54–58.
[46] S. Shi, Y. Zhao, H. Zhou, Y. Zhang, X. Jiang, K. Huang, J.
Chrom. A 2008, 1209, 145–152.
[42] J. W. Mciver Jr., Acc. Chem. Res. 1974, 7, 72–77.
[43] a) K. Fukui, Acc. Chem. Res. 1981, 14, 363–368; b) M. Head-
Gordon, J. A. Pople, J. Chem. Phys. 1988, 89, 5777–5786.
[44] M. C. Foti, C. Daquino, C. Geraci, J. Org. Chem. 2004, 69,
2309–2314.
Received: July 17, 2009
Published Online: November 6, 2009
6300
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 6289–6300