Structure and formation of the fluorescent compound of lignum nephriticum

Article abstract of DOI:10.1021/ol901022g

The intense blue fluorescence of the infusion of Lignum nephriticum (Eysenhardtia polystachya), first observed in the sixteenth century, is due to a novel four-ring tetrahydromethanobenzofuro[2,3-c/]oxacine which is not present in the plant but is the end product of an unusual, very efficient iterative spontaneous oxidation of at least one of the tree's flavonoids.

Full text of DOI:10.1021/ol901022g

ORGANIC
LETTERS
2009
Vol. 11, No. 14
3020-3023
Structure and Formation of the
Fluorescent Compound of Lignum
nephriticum
A. Ulises Acun˜a,*,† Francisco Amat-Guerri,*,‡ Purificacio´n Morcillo,‡ Marta Liras,‡
and Benjam´ın Rodr´ıguez‡
Instituto de Qu´ımica-F´ısica “Rocasolano” (IQFR), CSIC, Serrano 119,
28006 Madrid, Spain, and Instituto de Qu´ımica Orga´nica General (IQOG), CSIC,
Juan de la CierVa 3, 28006 Madrid, Spain
roculises@iqfr.csic.es
Received May 11, 2009
ABSTRACT
The intense blue fluorescence of the infusion of Lignum nephriticum (Eysenhardtia polystachya), first observed in the sixteenth century, is
due to a novel four-ring tetrahydromethanobenzofuro[2,3-d]oxacine which is not present in the plant but is the end product of an unusual,
very efficient iterative spontaneous oxidation of at least one of the tree’s flavonoids.
Fluorescence from a liquid solution was first reported in 1565
by Monardes1 as the blue tinge of the infusion of a Mexican
medicinal wood, known in Europe as Lignum nephriticum.2,3
Throughout the centuries, Boyle, Newton, Herschel, and
many other scientists were intrigued by this unusual phe-
nomenon.2 However, by the time at which fluorescence could
be interpreted in electronic terms, the botanical origin of L.
nephriticum was uncertain, and hence, the molecular structure
of the earliest fluorophore remained unknown. From the
arduous search of the source of L. nephriticum,4-6 two genera
of trees appeared as plausible candidates: Eysenhardtia Kunht
and Pterocarpus Jacq. We found in sixteenth-century manu-
scripts of Sahagu´n7 that Aztec healers had already noticed
the “blue” color (“matlali” in Nahuatl) of the infusion of
coatli, a plant used to treat urinary disorders.2 These
observations, together with more detailed accounts from
Mexican botanists,4 leave little doubt that the source of coatli/
L. nephriticum is the Mexican tree Eysenhardtia polystachya
(Ort.) Sarg.2 However, this wood does not contain a large
amount of an easily water-soluble blue-fluorescent dye.8-10
Quite on the contrary, it is very rich (ca. 1% dry weight) in
a group of rare, nonfluorescing C- and O-ꢀ-glycosyl-R-
hydroxydihydrochalcones. To solve this paradox, we isolated
from E. polystachya two previously characterized,9-11 easily
† IQFR.
‡ IQOG.
(1) Monardes, N., Dos Libros/El Vno trata de todas las cosas que traen
de nuestras Indias occidentales que sirVen al uso de Medicina. . . , en casa
de Sebastian Trugillo, Seville, 1565.
(7) Sahagu´n, B. Matritensis Codex; Spanish Royal Academy of History,
ca. 1560-1564, f203v.
(8) (a) Dom´ınguez, X. A.; Franco, R.; D´ıaz Viveros, Y. ReV. Latinoamer.
Qu´ım. 1978, 9, 209–211. (b) Burns, D. T.; Dalgarno, B. G.; Gargan, P. E.;
(2) Acun˜a, A. U.; Amat-Guerri, F. Early History of Solution Fluores-
cence: The Lignum nephriticum of Nicola´s Monardes. Springer Ser.
´
Grimshaw, J. Phytochemistry 1984, 23, 167–169. (c) Alvarez, L.; Rios,
Fluoresc. 2008, 4, 3–20
(3) Partington, J. R. Ann. Sci. 1955, 11, 1–26
(4) Stapf, O. Bull. Miscell. Inform. Kew Gardens 1909, 293–302
(5) Mo¨ller, H. J. Ber. Deut. Pharm. Ges. 1913, 23, 88–154
(6) Safford, W. E. Ann. Rep. Smithsonian Inst. 1915, 271–298
.
M. Y.; Esquivel, C.; Cha´vez, M. I.; Delgado, G.; Aguilar, M. I.; Villarreal,
.
M. L.; Navarro, V. J. Nat. Prod. 1998, 61, 767–770
(9) Beltrami, E.; De Bernardi, M.; Fronza, G.; Mellerio, G.; Vidari, G.;
Vita-Finzi, P. Phytochemistry 1982, 21, 2931–2933
.
.
.
.
´
(10) Alvarez, L.; Delgado, G. Phytochemistry 1999, 50, 681–687.
.
10.1021/ol901022g CCC: $40.75
Published on Web 06/17/2009
 2009 American Chemical Society

water-soluble glucosyldihydrochalcones: coatline B (1) and
coatline A (3) (Figure 1). Unexpectedly, it was found that
addition of a water molecule to the spirocyclic moiety (step
4), yields compound IV, with an R-hydroxypropionic acid
substituent. A similar biphenyl derivative was isolated in the
phloridzin oxidation noted above,14 whereas dimerization of
a catechol via the corresponding o-quinone has been previ-
ously observed.15 The repetition of the oxidative step, now
at the 3,4-diphenol group of IV (step 5), yields V, as well
as of the Michael-type intramolecular addition, herein of the
2′-OH group to the o-quinone system (step 6), leads to VI
through a process already observed in some biphenol
compounds.15,16 The tautomerization of VI (step 7) produces
the keto form VII, that yields the final product 2 by R-OH
cis attack on C-3 (step 8). Under the same conditions,
coatline A (3) is unreactive, which is consistent with the
above reaction sequence. A somewhat similar reaction has
been postulated to explain the formation of a related azocine
with very low yield.17
Figure 1. Matlaline (2), the fluorophore of L. nephriticum, and
related compounds.
at room temperature and in slightly alkaline (pH ∼7.5) water
solution, 1 undergoes a fast, irreversible reaction giving rise
to a strongly blue-emitting compound with ca. 100% yield,
matlaline (2), the fluorophore of the long-sought L. nephriti-
cum. No other products were detected in the reaction, even
in trace quantities.
Matlaline has a molecular formula of C21H22O12, 14 atomic
mass units more than 1 (C21H24O11), and its structure was
established from NMR spectroscopic studies. The H and
1
13C NMR spectra of 2 show signals for a C-ꢀ-D-glucopyra-
nosyl group and two ortho aromatic protons almost identical
to those of 1. HSQC and HMBC experiments allow the
unambiguous assignment of all the protons and carbons of
the tetrasubstituted benzene and its C-sugar substituent (see
the Supporting Information). The remaining 1H and 13C NMR
signals of 2 are assigned to a trisubstituted olefin [δH 6.35
(s, 1H, H-5); δC 168.9 (C, C-6), 106.7 (CH, C-5)] conjugated
with a ketone [δC 191.5 (C, C-4)], a hemiacetal carbon [δC
94.9 (C, C-3)], a fully substituted sp3 carbon bearing an
oxygen [δC 89.0 (C, C-1)], a methylene group without vicinal
The formation of 2, with a new oxacine core, must result
from a sequence of highly directed repetitive reactions
(Scheme 1), triggered by the ionization of 1 at the 4′ position
in neutral or slightly alkaline medium (pK′a ) 6.5 ( 0.2).
The reactive anion can be easily detected as a transient
species absorbing at 338 nm if its conversion rate is retarded,
e.g., by deoxygenating the solution or lowering the pH. In
fact, conversion of 1 into 2 can be prevented in aerated
solutions at pH 4 or at pH 8 in the absence of oxygen. The
first step of the sequential reaction should be the oxidation
of the o-diphenol group of 1 by atmospheric dioxygen to
yield the highly reactive o-quinone I. Although in cathechols
this process occurs in the presence of oxidants or is catalyzed
by enzymes or transition metals,12,13 we observed full
reaction of 1 even in NH3(gas)-alkalinized ultrapure water
solution. In step 2, the C1′-spiro intermediate II is formed
from the Michael-type intramolecular addition of C-1′ on
C-6.13 Subsequent aromatization to the corresponding 3,4-
diphenol III (step 3), as that proposed in the enzymatic
oxidation of the dihydrochalcone phloridzin,14 followed by
2
protons [δH 2.54 and 2.47 (1H each, J ) 11.2 Hz, 2H-2);
δC 44.5 (CH2, C-2)], a (C)CH2CH(C)O- fragment [δH 4.22
3
(dd, 1H, J ) 12.1, 2.8 Hz, H-R), 2.12 and 1.98 (1H each,
2J ) 12.6 Hz, 2H-ꢀ); δC 70.0 (CH, C-R), 35.9 (CH2, C-ꢀ)],
and a carboxyl group [δC 177.4 (C)]. Moreover, the C-2
methylene proton at δΗ 2.47 shows a W-type coupling (4J
) 2.8 Hz) with one of the C-ꢀ methylene protons (δH 1.98),
whereas the observed vicinal couplings between the H-R and
2H-ꢀ protons (3J ) 12.1 and 2.8 Hz) indicate that H-R
possesses an axial orientation. In addition, the HMBC
spectrum of 2 displays correlations of H-5 with C-1′, C-1,
and C-3, between 2H-2 and C-1, C-3, C-4, C-6, and C-ꢀ,
Scheme 1. Plausible Sequential Reaction That Transforms Coatline B (1) into Matlaline (2) at Room Temperature in Water Solution
Org. Lett., Vol. 11, No. 14, 2009
3021

and between 2H-ꢀ and the carboxyl at δC 177.4, as well as
with C-1, C-2, C-6, and C-R. The significant HMBC
correlation between H-5 and C-1′ [δC 109.8 (C)] and all of
the above data are only compatible with a structure such as
2 for matlaline. This conclusion was also supported by the
synthesis and structural analysis of the aglucon 6 shown
below.
The total stereoselectivity shown in Scheme 1 for the
conversion of 1 into 2 may deserve further comment. First,
since none of the bonds to RC are broken along the
reaction sequence, the (RR) absolute configuration of 1 is
preserved in 2. However, diastereoisomers (RR,1R,3R)-2
and (RR,1S,3S)-4 (Figure 1), instead of the sole isomer 2,
may be expected as reaction products. The observed selectiv-
ity could be due to the presence of a C-ꢀ-D-glucopyranosyl
substituent at C3′ and/or to the (RR) chiral center in the
starting compound 1. This point was settled by the synthesis
of the (RR)/(RS) racemate 5, the coatline B aglucon which
is also present in E. polystachya in enantiopure (RR) form.10
The synthesis of 5 (Scheme 2) was carried out in three steps:
Figure 2. X-ray structure of a racemate 7 single crystal with four
molecules in the monoclinic cell unit. From left to right: enantiomers
(RR,1R,3R), (RS,1S,3S), (RR,1R,3R), and (RS,1S,3S).
reaction, based on the relative stability of diastereomers 2
and 4, with the CO2H substituent in an exo-equatorial (2) or
an endo-axial (4) orientation relative to the 2-oxabicyclo-
[3.3.1]non-6-ene system. From a steric point of view, 2 is
expected to be more stable than 4. In addition, a stabilizing
intramolecular H-bond can be formed in isomer 2 between
the cis-oriented CO2H and 3-OH groups. Should isomer 4
be formed via the (1S) isomer of VI, generated by attack of
the ionized 2′-OH of V from the Si-face of its o-quinone
ring, it could give rise to the more stable isomer 2 through
hemiacetal ring-opening, tautomerization and retro-Michael
reactions (back-steps 8, 7 and 6, respectively), subsequent
attack of the ionized 2′-OH from the Re-face of the o-quinone
ring, and final hemiacetal closure (Scheme 1).
Scheme 2
The most conspicuous property of 2 is, of course, the
intense fluorescence (λf ) 465 nm) of even much diluted
aqueous solutions. This is due to the combination of a large
absorption coefficient in the visible range (Table 1) and a
(1) condensation between 2′,4′-dihydroxyacetophenone and
3′,4′-dihydroxybenzaldehyde, both with their OH-groups
appropriately protected as benzyl ethers; (2) epoxide forma-
tion in the generated double bond; and (3) regioselective
cleavage to the corresponding R-alcohol by hydrogenation,
with simultaneous OH-deprotection. The matlaline aglucon
6 and its methyl ester 7 were also prepared from 5. Single-
crystal diffraction analysis of 7 indicated that the crystals
contain the racemic mixture of the (RR,1R,3R) and
(RS,1S,3S) forms, both presenting the CO2Me group in an
exo-equatorial orientation (Figure 2). Thus, the RC chiral
center drives the stereoselective transformation of 5 (a
racemic mixture) into 6.
Table 1. Absorption Properties and pK′a Values of Coatline B
(1) and Matlaline (2) in Water Solution
compd
pK′a
solution pH
λab (nm) (ε) (M-1 cm-1
)
1
6.5 ( 0.2
4-7
282 (23630 ( 300)
325 (15240 ( 200)
338 (21700 ( 600)
307 (5800 ( 600)
382 (14800 ( 200)
283 (4400 ( 300)
429.5 (33800 ( 800)
10a
4–5.5
2
5.5 ( 0.4
Since the same effect would operate in the transformation
of 1 into 2, total stereoselectivity can be expected for this
9
a Deoxygenated.
(11) At least two Eysenhardtia species produce strongly fluorescent
infusions: E. polystachya and E. officinalis. The work described here was
carried out with the first species because several of its components have
been previously characterized.
ca. 100% fluorescence quantum yield (Figure 3). The
fluorescence intensity of the aqueous solution depends
strongly on pH, as noticed also in historical times by Kircher
and Boyle,2 due to the protolytic equilibrium between an
emitting and a dark form (pK′a ≈ 5.5). In neutral or slightly
alkaline solution, the dominant species is the strongly
fluorescent dianion form of 2, in which both the 4′-OH and
the CO2H groups are ionized. Interestingly, two resonant
structures can be drawn for this species which remind those
(12) (a) Young, D. A.; Young, E.; Roux, D. G.; Brandt, E. V.; Ferreira,
D. J. Chem. Soc., Perkin Trans. 1 1987, 2345–2351. (b) Dorrestein, P.;
Begley, T. P. Bioorg. Chem. 2005, 33, 136–148.
(13) Guyot, S.; Cheynier, V.; Souquet, J.-M.; Moutounet, M. J. Agric.
Food Chem. 1995, 43, 2458–2462.
(14) Le Guerneve´, C.; Sanoner, P.; Drilleaub, J.-F.; Guyot, S. Tetrahe-
dron Lett. 2004, 45, 6673–6677.
(15) (a) Guyot, S.; Vercauteren, J.; Cheynier, V. Phytochemistry 1996,
42, 1279–1288. (b) Tanaka, T.; Mine, C.; Inoue, K.; Matsuda, M.; Kouno,
I. J. Agric. Food Chem. 2002, 50, 2142–2148.
(16) Matsuo, Y.; Tanaka, T.; Kouno, I. Tetrahedron 2006, 62, 4774–
4783.
3022
Org. Lett., Vol. 11, No. 14, 2009

It was known very early6,19 that the infusion of the
Philippine Pterocarpus indicus (narra tree), used traditionally
for kidney diseases, shows a striking fluorescence similar to
that of L. nephriticum. Moreover, it has been suggested5,6,20
that a Pterocarpus species was the source of, or at least a
medicinal substitute for, the Mexican wood. We investigated
samples of several Pterocarpus species collected in Mexico
(see the Supporting Information), but none of them contained
a substantial amount of any water-soluble blue-emitting
compound. We also analyzed samples of Pterocarpus indicus
Willd. from the Philippines to find out that its wood contains
large quantities of the strongly emitting compound 2 already
preformed, but only trace amounts of 1.21 Interestingly, the
presence of 2 in the wood of P. indicus has not been reported
in previous studies of the tree’s components.22
Figure 3. Spectral properties of matlaline (2) in water solution.
Absorption (red) and corrected fluorescence (blue) spectra, emission
yield (Φf), and lifetime (τf) of the dianion form at pH 9; absorption
spectrum of the nonemitting form (black) at pH 4.
Acknowledgment. We thank S. Castroviejo (Real Jard´ın
Bota´nico, CSIC, Madrid) for botanical advice, V. Hornillos
(Instituto de Qu´ımica Orga´nica General, CSIC) for helpful
discussions, and the Ministerio de Educacio´n y Ciencia
(Spain), and CSIC for financial support and a I3P fellowship
(P.M.).
of the strongly emitting fluorescein dianion. In contrast, the
-
monoanion form (CO2 ) of 2 is nonfluorescent and shows
UV-shifted weaker absorption (Figure 3). In addition to that,
the chemical conversion of 1 into 2 is abolished in water at
pH close to 5. All these pH-dependent effects may have
confounded the botanical search.18
Supporting Information Available: General experimental
procedures, isolation of compounds 1-3, preparation of 5-7,
characterization data, and X-ray crystallographic data of 7.
This material is available free of charge via the Internet at
http://pubs.acs.org.
By preparing 6, the aglucon form of 2, we also could
determine that the emitting properties of 6 are identical to
those of matlaline, within experimental error, and indepen-
dent of the 3′-glucosyl residue, which of course accounts
for the large water solubility of 2. Thus, the fluorescence of
the plant infusion originates not only from 2 but also from
compounds with the same chromophore but containing a
different glycoside group, which can also be formed from
the other two C- and O-ꢀ-xylopyranosyl R-hydroxydihydro-
chalcones present in E. polystachya.9,10
OL901022G
(19) Delgado, J. J. In Historia de Filipinas, written in 1754. Published
for the first time in Historia General Sacro-profana y Natural de las Islas
del Poniente, llamadas Filipinas; Mart´ınez de Zu´n˜iga, Ed.; Manila, 1892;
pp 415-416
.
(20) Muyskens, M. J. J. Chem. Educ. 2006, 83, 765–768
.
(21) Since the fluorescence of the infusion of both E. polystachya and
the Philippine P. indicus is identical, it is not unlikely that both trees were
interchanged as a commercial source of L. nephriticum, but at a much later
date than that of Monardes’ observation, when regular trade with the
Philippine Islands was eventually established.
(17) Crescenzi, O.; Napolitano, A.; Prota, G.; Peter, M. G. Tetrahedron
1991, 47, 6243–6250.
(18) In fact, Monardes already noticed that 0.5 h was needed for the L.
nephriticum infusion to develop its full blue “color”, indicating that he was
handling Eysenhardtia wood.
(22) (a) Cooke, R. G.; Rae, I. D. Aust. J. Chem. 1964, 17, 379–384. (b)
Gonzales, E. V. Philippine J. Sci. 1978, 105, 223–233. (c) Ragasa, C. Y.;
De Luna, R. D.; Hofilena, J. G. Nat. Prod. Res. 2005, 19, 305–309.
Org. Lett., Vol. 11, No. 14, 2009
3023