Unique proline-benzoquinone pigment from the colored nectar of bird's coca cola tree functions in bird attractions

Article abstract of DOI:10.1021/ol3017879

The major pigment responsible for the dark brown nectar of the bird's Coca cola tree , Leucosceptrum canum (Labiatae), was isolated and identified as a unique symmetric proline-quinone conjugate, 2,5-di-(N-(-)- prolyl)-para-benzoquinone (DPBQ). Behavioral experiments with both isolated and synthetic authentic samples indicated that DPBQ functions mainly as a color attractant to bird pollinators.

Full text of DOI:10.1021/ol3017879

ORGANIC
LETTERS
2012
Vol. 14, No. 16
4146–4149
Unique ProlineꢀBenzoquinone Pigment
from the Colored Nectar of “Bird’s Coca
Cola Tree” Functions in Bird Attractions
Shi-Hong Luo,†,‡ Yan Liu,† Juan Hua,† Xue-Mei Niu,§ Shu-Xi Jing,† Xu Zhao,†
Bernd Schneider,^ Jonathan Gershenzon,^ and Sheng-Hong Li*,†
State Key Laboratory of Phytochemistry and Plant Resources in West China,
Kunming Institute of Botany, Chinese Academy of Sciences, Kunming 650201, P. R. China,
Key Laboratory for Conservation and Utilization of Bio-resources, Yunnan University,
Cui Hubei Road 2, Kunming 650091, P. R. China, Max Planck Institute for Chemical
€
Ecology, Beutenberg Campus, Hans-Knoll-Strasse 8, D-07745 Jena, Germany, and
Graduate University of Chinese Academy of Sciences, Beijing 100039, P. R. China
shli@mail.kib.ac.cn
Received June 29, 2012
ABSTRACT
The major pigment responsible for the dark brown nectar of the “bird’s Coca cola tree”, Leucosceptrum canum (Labiatae), was isolated and
identified as a unique symmetric proline-quinone conjugate, 2,5-di-(N-(ꢀ)-prolyl)-para-benzoquinone (DPBQ). Behavioral experiments with both
isolated and synthetic authentic samples indicated that DPBQ functions mainly as a color attractant to bird pollinators.
The chemistry and ecological function of floral colored
nectar have recently attracted much research interest.1 In
nature, flowernectarisusuallya colorless liquid becauseits
contentsincluding nutrientsand secondarymetabolites are
mostly colorless. However, at least 68 taxa from 20 genera
belonging to 15 families have been documented to possess
colored nectar, covering different broad categories of
colors including yellow, amber-orange, red, brown, green,
blue, and black,2 and probably many other colored nectar
plants are still to be discovered. Although a few such plants
have been phytochemically and ecologically investigated,
pigmented compounds in the colored nectar of most plants
and their ecological functions are still largely unknown.
Leucosceptrum canum Smith, a large tree (up to 10 m
high3) belonging to the family Labiatae (= Lamiaceae),
has an unusual dark brown nectar and is the only colored
nectar plant so far discovered in the Labiatae family. It is a
favorite species of birds, and field observations have
reported that it can attract over 40 kinds of birds for
feeding within 40 min.4 For this reason, L. canum is also
called “bird’s Coca cola tree” by ornithologists.4 Most
recently, Zhang et al.5 reported that a purple anthocyani-
din, 5-hydroxyflavylium, was the cause of the nectar
coloration of L. canum and functioned as a foraging signal
in bird pollination. Unfortunately, this compound was
only characterized by UV and mass spectral methods,
† Kunming Institute of Botany.
‡ Graduate University of Chinese Academy of Sciences.
§ Yunnan University.
^ Max Planck Institute for Chemical Ecology.
(1) (a) Olesen, J. M.; Ronsted, N.; Tolderlund, U.; Cornett, C.;
Molgaard, P.; Madsen, J.; Jones, C. G.; Olsen, C. E. Nature 1998, 393,
529. (b) Hansen, D. M.; Beer, K.; Mueller, C. B. Biol. Lett. 2006, 2,
165–168.
(3) Peng, H.; Wu, Z. Y. Acta Bot. Yunnan. 2001, 23, 278–286.
(4) Han, L. X. Nat. Watch 2009, 1, 17–18.
(5) Zhang, F. P.; Cai, X. H.; Ren, Z. X.; Zachary, L. R.; Li, D. Z.
(2) Hansen, D. M.; Olesen, J. M.; Mione, T.; Johnson, S. D.; Mueller,
C. B. Biol. Rev. 2007, 82, 83–111.
New Phytol. 2012, 193, 188–195.
r
10.1021/ol3017879
Published on Web 08/02/2012
2012 American Chemical Society

and therefore its identification must be considered tenta-
tive. We have been interested in the secondary metabolites
of L. canum and their ecological functions because of its
uniqueness as a large woody Labiatae with colored nectar
and have found that the glandular trichomes of L. canum
harbor defensive sesterterpenoids with a novel C25 carbon
framework.6 To discover the pigmented compounds re-
sponsible for the colored nectar of L. canum and their
ecological significance, we initiated an investigation three
years ago and recently purified the major pigment and
unambiguously determined its chemical structure with
comprehensive spectroscopic analyses, especially 1D and
2D NMR and HRMS, and chemical synthesis. With the
purified and synthetic authentic samples, we were also able
to conduct behavioral experiments on a bird pollinator of
L. canum.
L. canum blooms from early December to late March of
the next year, and each plant can produce hundreds of
inflorescences and thousands of white tubular flowers
(Figure 1aꢀd). The tubular flowers are filled with dark
browncolored nectar(Figure 1d) inthe morning. The most
common visitors are Japanese White-eye (Zosterops
japonicus) (Figure 1a), Blue Winged Minla (Minla
cyanouroptera) (Figure 1b), and Black-headed Sibia
(Heterophasia melanoleuca) (Figure 1c) based on our
own field observations. Large scale HPLC analysis of
109 nectar samples collected from the Botanical Garden
of Kunming Institute of Botany and Dehong with a
detection wavelength of 400 nm indicated that all sam-
ples contained a predominant peak with maximum UV
absorptions at 215, 369, and 525 nm (Figure S1), which
was accordingly targeted as the responsible pigmented
compound of the dark brown nectar of L. canum. How-
ever, when we started to isolate this compound for
spectroscopic identification and subsequent behavioral
experiments, it was found to be so unstable that isola-
tion and identification were very difficult. After two
years of effort, we were finally able to obtain a 27.5 mg
sample of this compound with sufficient purity for
identification as a dark brown solid from 645 mL of
L. canum nectar.
Figure 1. Leucosceptrum canum and its flower visitors. (a)
Japanese White-eye (Zosterops japonicus); (b) Minla cyanour-
optera; (c) Heterophasia melanoleuca; (d) Dark brown nectar
of L. canum; (e) Chemical structure of 2,5-di-(N-prolyl)-para-
benzoquinone (DPBQ); (f) Japanese White-eye probes
DPBQ (500 μg/mL) in the behavioral experiments.
(Figure S4), which were further classified by DEPT experi-
ments as three methylenes, two methines including an
olefinic one (δC 100.1), and three quaternary carbons
including an olefinic one (δC 148.6) and two carbonyl
1
carbons (δC 179.9, and 173.4). Its H NMR spectrum
(Figure S3) displayed seven groups of signals, including
an olefinic singlet at δH 5.23, a broad doublet at δH 4.99, a
two-proton multiplet at δH 3.35, and four multiplets at δH
1.78ꢀ2.21, which were readily assigned to the correspond-
ing methylene and methine carbons through a hetero-
nuclear single-quantum correlation (HSQC) experiment
1
1
(Figure S6). The Hꢀ H COSY correlations (Figure S5)
established a fragment of ꢀCHCH2CH2CH2ꢀ with the
terminal protons and carbons occurring relatively down-
field, suggesting the existence of a tetrahydropyrrole moi-
ety, which was substituted by a carboxylic group at C-20
The pigmented compound was shown to have a mole-
cular formula of C16H18N2O6 by positive and negative ESI
mass spectrometry and high-resolution ESI-MS.7 Eight
carbon resonances were found in its 13C NMR spectrum
1
13
owing to the Hꢀ C long-range correlations from H-20
and H2-30 to the carbonyl carbon at δC 173.4 in the
heteronuclear multiple bond coherence (HMBC) spectrum
(Figure S7). Thus a proline unit was evident in the struc-
ture. The remaining signals in the NMR spectra were
ascribable to an R,β-unsaturated keto residue, which was
attached to the nitrogen of proline because of the HMBC
correlations from H-20 and H2-50 to the olefinic quaternary
carbon and ROESY correlation between H-3/H-6 and
H-20 (Figure S8). In the HMBC spectrum (Figure S7), an
(6) Luo, S. H.; Luo, Q.; Niu, X. M.; Xie, M. J.; Zhao, X.; Schneider,
B.; Gershenzon, J.; Li, S. H. Angew. Chem., Int. Ed. 2010, 49, 4471–4475.
(7) Natural DPBQ:Dark brown solid;[R]D25 =ꢀ46.4 (c = 0.1, water);
UV/vis (water): λmax (log ε) 215 (3.55), 369 (3.37), 525 (2.10) nm; IR (KBr):
ν
max 3426, 2978, 1624, 1540, 1411, 1385, 1290, 1187, 1009 cmꢀ1; Positive
ESI-MS: m/z (%) 357 (7) [MþNa]þ, 335 (5) [MþH]þ, 317 (9), 301 (40),
219 (31), 203 (100); Neꢀgative ESI-MS: m/z (%) 355ꢀ(42) [MþNaꢀ2H]ꢀ,
289 (100) [MꢀCOOH] , 243 (41) [Mꢀ2COOHꢀH] , 217 (29), 215 (77);
HR-ESI-MS: m/z 335.1169 [MþH]þ (m/zcalcd [C16H19N2O6]þ
=
4
unusual J coupling from the proton at δH 5.23 to the
335.1238); 357.1060 [MþNa]þ (m/zcalcd [C16H18N2O6Na]þ = 357.1057);
1H NMR (600 MHz, DMSO-d6): δ 5.23 (s, 2H, H-3 and H-6), 4.99 (brd,
J=5.4Hz,2H,H2-20), 2.20(m, 2H,H2-30a), 2.00 (m, 2H, H2-30b), 1.89 (m,
2H, H2-40a), 1.78 (m, 2H, H2-40b), 3.35 (m, 4H, H4-50); 13C NMR (150
MHz, DMSO-d6): δ 179.9 (s, 2C, C-1 and C-4), 173.4 (s, 2C, C2-60), 148.6
(s, 2C, C-2 and C-5), 100.1 (d, 2C, C-3 and C-6), 62.6 (d, 2C, C2-20), 51.1 (t,
2C, C2-50), 31.1 (t, 2C, C2-30), 21.6 (t, 2C, C2-40). Synthetic (ꢀ)-DPBQ:
Dark brown solid; [R]D19 = ꢀ41.1 (c = 0.1, water). Synthetic (þ)-DPBQ:
Dark brown solid; [R]D19 = þ58.3 (c = 0.1, water).
carbon at δC 100.1 (from H-3 to C-6 and from H-6 to C-3)
was also observed, indicating a symmetric structure for the
compound. Considering its above molecular formula, the
compound was straightforwardly established to be either
2,5-di-(N-prolyl)-para-benzoquinone or 2,5-di-(N-prolyl)-
ortho-benzoquinone. However, it is difficult to distinguish
Org. Lett., Vol. 14, No. 16, 2012
4147

para-benzoquinone and ortho-benzoquinone and deter-
mine the absolute configuration of the proline in the
structure by an NMR method, and an attempt to crystal-
lize the compound for X-ray analysis also failed; therefore
synthesis of this compound was necessary. By condensa-
tion of para-benzoquinone under alkaline conditions with
L-(ꢀ)-proline and D-(þ)-proline respectively, 2,5-di-(N-(ꢀ)-
prolyl)-para-benzoquinone ((ꢀ)-DPBQ) and 2,5-di-(N-
(þ)-prolyl)-para-benzoquinone ((þ)-DPBQ) were synthe-
sized (Figure 2). Both products were found to have identical
retention times and 1H and 13C NMR spectra to the isolated
pigmented compound (Figures S9ꢀS11). Therefore,
the planar structure of the nectar pigment should be 2,5-
di-(N-prolyl)-para-benzoquinone (DPBQ hereafter). Its
CD curve was consistent with that of (ꢀ)-DPBQ but
opposite to that of (þ)-DPBQ (Figure 3), indicating that
the proline in its structure was the L-(ꢀ)-form (Figure 1e),
which was also supported by their optical rotations (DPBQ:
[R]D25 = ꢀ46.4; (ꢀ)-DPBQ: [R]D19 = ꢀ41.1; (þ)-DPBQ:
Figure 4. Color comparison of DPBQ in water at different
concentrations (aꢀd: 100, 500, 1000, and 1500 μg/mL of DPBQ
respectively) with L. canum nectar in deepest color (e).
As Figure 4 shows, DPBQ is the major cause of the dark
brown color of L. canum nectar. Although a number of
aminoquinones such as nakijiquinones have been reported
from marine sponges,8 quinoneꢀanimo acid conjugates
are rarely found in the plant kingdom. DPBQ represents a
novel type of natural pigment in floral nectar. DPBQ is
generally unstable in organic solvents such as DMSO
(degradation and presumably polymerization may occur)
but is relatively stable in water especially under alkaline
conditions such as in L. canum nectar (pH = 8.40 ( 0.17).
We have found that the pH of L. canum nectar could
fluctuate at least between 7.93 and 8.57. It can be therefore
postulated that the color difference of L. canum nectar in
different flower stages as pointed out by Zhang et al5 could
have been caused by different concentrations of DPBQ
and pH changes in the nectar. Alternatively, it is possible
that other amino acids besides proline could also conjugate
with para-benzoquinone to result in other different colors.
Such compounds were actually obtained as minor consti-
tuents of the nectar of L. canum but unfortunately were not
fully identified due to their instability and paucity of
sample amounts.
19
[R]D = þ58.3). Consequently, the nectar pigment was
determined as 2,5-di-(N-(ꢀ)-prolyl)-para-benzoquinone.
Figure 2. Synthetic routes of (ꢀ)-DPBQ and (þ)-DPBQ.
Before we could test DPBQ on the behavior of bird
visitors, it was important to know the concentration of this
pigment in the nectar of L. canum. However, when nectar
was removed on the morning of the first day that the flower
was open, we found that new nectar was secreted on the
next day but the brown color became lighter. After nectar
removal on successive days, the same thing occurred until
the nectar finally had a purple-like color on the fifth
sampling day when the flower senesced. These observa-
tions implied that DPBQ was regularly synthesized and
secreted but that its concentration changed with flower
stage and feeding by visitors. Therefore, HPLC was em-
ployed to quantify DPBQ in the nectar sampled from five
inflorescences (each 10 flowers) over five consecutive days.
It was found that the concentration of DPBQ was extre-
mely high (up to 1.39 mg/mL) in nectar on the first two
days of blooming, but the level gradually decreased with
the flowering time and removal of old nectar. On the
fifth day of blooming, as the flowers became senesced,
the concentration of DPBQ dropped to a minimum
Figure 3. CD spectra of DPBQ, (ꢀ)-DPBQ, and (þ)-DPBQ.
(8) (a) Shigemori, H.; Madono, T.; Sasaki, T.; Mikami, Y.; Kobayashi,
J. Tetrahedron 1994, 50, 8347–8354. (b) Kobayashi, J.; Madono, T.;
Shigemori, H. Tetrahedron 1995, 51, 10867–10874. (c) Takahashi, Y.;
Kubota, T.; Ito, J.; Mikami, Y.; Fromont, J.; Kobayashi, J. Bioorg. Med.
Chem. 2008, 16, 7561–7564. (d) Takahashi, Y.; Kubota, T.; Kobayashi, J.
Bioorg. Med. Chem. 2009, 17, 2185–2188. (e) Takahashi, Y.; Ushio, M.;
Kubota, T.; Yamamoto, S.; Fromont, J.; Kobayashi, J. J. Nat. Prod. 2010,
73, 467–471.
4148
Org. Lett., Vol. 14, No. 16, 2012

(Figure S2). The result was in good agreement with our
above observations of changes in nectar color.
Naıve Japanese White-eyes (Z. japonicus) were used for
¨
testing their behavioral responses to nectar, nectar sugar,
and DPBQ, with a method described in the literature9
with suitable modifications (Methods in the Supporting
Information). As shown in Figure 5, Japanese White-eyes
probed the colored nectar of L. canum in an incredibly
higher proportion (%) (85.40 ( 5.94) than the clear nectar
sugar (p < 0.001) and water (p < 0.001) (Figure 5a). Each
time the nectar was drunk up by birds in just one probe.
After the exteriors of the glass dishes of nectar sugar and
water were painted dark brown, this proportion dramati-
cally decreased (56.67 ( 4.08) and proportions for nectar
sugar and water obviously increased (Figure 5b). Under
this circumstance, birds consumed much higher amounts
(μL) of L. canum nectar (95.34 ( 2.44) and nectar sugar
(85.63 ( 4.12) than water (34.02 ( 5.32) (Figure 5c). In
parallel experiments, natural DPBQ and the synthetic (ꢀ)-
DPBQ and (þ)-DPBQ were also tested at a medium
natural concentration (500 μg/mL). It was found that, in
the first 20 trials, Japanese White-eyes absolutely preferred
the pigmented compounds (DPBQ, (ꢀ)-DPBQ, and (þ)-
DPBQ) over water. However, with increased trials, this
preference decreased dramatically. It was found that birds
probed DPBQ, (ꢀ)-DPBQ, and (þ)-DPBQ in proportions
of 71.8 ( 4.8, 71.0 ( 2.6, and 74.0 ( 6.0 respectively at 50
trials, which were still significantly higher than that of water
(p < 0.001) (Figure 5dꢀf). When the glass dish of water was
painted outside to match the color of DPBQ solution, birds
probed DPBQ and water in almost equal proportions
(p = 0.70) (Figure 5g). Moreover, the corresponding
amounts consumed per trial also showed no significant
difference between DPBQ and water (Figure 5h). The
above results clearly indicatedthat DPBQ functions mainly
as a color attractant to bird pollinators in L. canum nectar.
Insummary, wehavechemicallycharacterizedthemajor
pigment of the dark brown nectar of a woody Labiatae as a
unique symmetric prolineꢀquinone conjugate, 2,5-di-(N-
(ꢀ)-prolyl)-para-benzoquinone, and demonstrated its func-
tion in attracting birds as potential pollinators. This is the first
time that both the elusive chemistry and ecological function
of pigment in colored nectar have been unambiguously
disclosed, which should shed more light on the evolution
and ecological significance of this enigmatic floral trait.
Figure 5. Preference and consumption of nectar (N), nectar
sugar (Ns) DPBQ, (ꢀ)-DPBQ, (þ)-DPBQ (each 500 μg/mL),
and water (W) by the bird Z. japonicus. (a, dꢀf) Distribution of
first choice; (b, g) Distribution of first choice (outside painted
dark brown to match nectar); (c, h) Consumption per trial
(outside painted to match color of nectar or DPBQ solutions).
Acknowledgment. This research was supported finan-
cially by the “Hundred Talents Program” of the Chinese
Academy of Sciences (CAS), the National Natural Science
Foundation of China (31070320), the National Basic
Research Program of China (973 Program) on Biological
Control of Key Crop Pathogenic Nematodes, and the
State Key Laboratory of Phytochemistry and Plant Re-
sources in West China (P2009-ZZ01).
Supporting Information Available. Experimental pro-
cedures regarding the general experimental details, nectar
materials, representative HPLC profile of nectar, isola-
tion of DPBQ, quantification of DPBQ and results,
behavioral experiment, statistical analysis, synthesis of
(ꢀ)-DPBQ and (þ)-DPBQ, 1D and 2D NMR spectra of
DPBQ, and overlays of HPLC chromatograms and 1D
NMR spectra of DPBQ, (ꢀ)-DPBQ, and (þ)-DPBQ.
This material is available free of charge via the Internet
at http://pubs.acs.org.
(9) Johnson, S. D.; Hargreaves, A. L.; Brown, M. Ecology 2006, 87,
2709–2716.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 16, 2012
4149