The biosynthesis of hydroxycinnamoyl quinate esters and their role in the storage of cocaine in Erythroxylum coca

Article abstract of DOI:10.1016/j.phytochem.2012.09.009

Complexation of alkaloids is an important strategy plants utilize to facilitate storage in vacuoles and avoid autotoxicity. Previous studies have implicated hydroxycinnamoyl quinate esters in the complexation of purine alkaloids in Coffea arabica. The goal of this study was to determine if Erythroxylum coca uses similar complexation agents to store abundant tropane alkaloids, such as cocaine and cinnamoyl cocaine. Metabolite analysis of various E. coca organs established a close correlation between levels of coca alkaloids and those of two hydroxycinnamoyl esters of quinic acid, chlorogenic acid and 4-coumaroyl quinate. The BAHD acyltransferase catalyzing the final step in hydroxycinnamoyl quinate biosynthesis was isolated and characterized, and its gene expression found to correlate with tropane alkaloid accumulation. A physical interaction between chlorogenic acid and cocaine was observed and quantified in vitro using UV and NMR spectroscopic methods yielding similar values to those reported for a caffeine chlorogenate complex in C. arabica. These results suggest that storage of cocaine and other coca alkaloids in large quantities in E. coca involves hydroxycinnamoyl quinate esters as complexation partners.

Full text of DOI:10.1016/j.phytochem.2012.09.009

Phytochemistry 91 (2013) 177–186
Contents lists available at SciVerse ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
The biosynthesis of hydroxycinnamoyl quinate esters and their role in the storage
of cocaine in Erythroxylum coca
José Carlos Pardo Torre a, Gregor W. Schmidt a, Christian Paetz b, Michael Reichelt a, Bernd Schneider b,
Jonathan Gershenzon a, John C. D’Auria a,
⇑
a Department of Biochemistry, Max Planck Institute for Chemical Ecology, Beutenberg Campus, Hans-Knoell-Strasse 8, D-07745 Jena, Germany
b NMR Research Group, Max Planck Institute for Chemical Ecology, Beutenberg Campus, Hans-Knoell-Strasse 8, D-07745 Jena, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 22 October 2012
Complexation of alkaloids is an important strategy plants utilize to facilitate storage in vacuoles and
avoid autotoxicity. Previous studies have implicated hydroxycinnamoyl quinate esters in the complexa-
tion of purine alkaloids in Coffea arabica. The goal of this study was to determine if Erythroxylum coca uses
similar complexation agents to store abundant tropane alkaloids, such as cocaine and cinnamoyl cocaine.
Metabolite analysis of various E. coca organs established a close correlation between levels of coca
alkaloids and those of two hydroxycinnamoyl esters of quinic acid, chlorogenic acid and 4-coumaroyl
quinate. The BAHD acyltransferase catalyzing the final step in hydroxycinnamoyl quinate biosynthesis
was isolated and characterized, and its gene expression found to correlate with tropane alkaloid accumu-
lation. A physical interaction between chlorogenic acid and cocaine was observed and quantified in vitro
using UV and NMR spectroscopic methods yielding similar values to those reported for a caffeine chloro-
genate complex in C. arabica. These results suggest that storage of cocaine and other coca alkaloids in
large quantities in E. coca involves hydroxycinnamoyl quinate esters as complexation partners.
Ó 2012 Elsevier Ltd. All rights reserved.
Dedicated to the memory of Meinhart H.
Zenk.
Keywords:
Cocaine
Hydroxycinnamoyl quinate esters
Tropane alkaloid
Erythroxylum coca
Vacuole
Chlorogenic acid
1. Introduction
throughout human history. It is found in the genus Erythroxylum,
accumulating in quantities of up to 1% dry weight in the
Alkaloids are one of the largest groups of secondary metabolites
in plants (Otani et al., 2005). Their abundance, together with their
important biological activities, has stimulated great interest in the
way alkaloids are transported and stored in planta (Ferreira et al.,
1998; Shitan et al., 2009; Sirikantaramas et al., 2008; Yazaki et al.,
2008). They are often stored in specialized compartments, such as
the vacuole which can take up to 90% of the total cell volume
(Kutchan, 2005; Roytrakul and Verpoorte, 2007; Yazaki et al.,
2008). The late Meinhart H. Zenk devoted considerable effort to
investigating alkaloid transport into vacuoles, evaluating the valid-
ity of different transport system models, such as the ion-trap
mechanism and specific carrier-mediated transport (Deus-
Neumann and Zenk, 1984, 1986).
Tropane alkaloids represent a medicinally important class of
alkaloids most commonly found in the Solanaceae, although they
also appear in the Convolvulaceae, Erythroxylaceae, Proteaceae,
and Rhizophoraceae (Asano et al., 2001). Of the various tropane
alkaloids, cocaine (1) (methylecgonine benzoyl ester) (Fig. 1) is
the most prominent due to its social and economical importance
mature leaves of two of the 230 known species, Erythroxylum coca
and Erythroxylum novogranatense (Bieri et al., 2006; Plowman and
Hensold, 2004). Cocaine (1) appears to serve as a toxin deterring
the attack of herbivores and pathogens (Nathanson et al., 1993). In
addition, it has been shown to be a genotoxic compound inducing
micronuclei formation in mammalian cells (Harborne and
Khan, 1993; Sonnante et al., 2010). Plants that synthesize cocaine
(1) may avoid autotoxicity by sequestering this alkaloid in
their vacuoles. Such localization has been shown via immunogold
labeling in the mature leaves (Ferreira et al., 1998).
To keep alkaloids trapped within the vacuoles, some research-
ers have proposed that they are complexed with polyphenols.
The most intensively studied interaction to date is that of the
purine alkaloid caffeine with chlorogenic acid (3) (caffeoyl
quinate) (Baumann et al., 1991; Waldhauser and Baumann,
1996; Zeller and Saleeb, 1997). In addition, hydroxycinnamoyl
quinate esters have also been implicated in complex formation
with other purine alkaloids (Baumann et al., 1991; Wink and
Roberts, 1998).
Chlorogenic acid (3) is the most abundant hydroxycinnamoyl
quinate ester found in plants (Sondheimer et al., 1961). Found in
high concentrations in many plant species, it acts as a defense
⇑
Corresponding author. Tel.: +49 03641 571335; fax: +49 03641 571302.
E-mail address: dauria@ice.mpg.de (J.C. D’Auria).
0031-9422/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.phytochem.2012.09.009

178
J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
against pathogens as well as serving as a UV protectant (Rhodes
and Wooltorton, 1976; Sander and Petersen, 2011). Chlorogenic
acid (3) and other hydroxycinnamoyl quinate esters are found in
the vacuoles and apoplasts where they aid peroxidases in hydro-
gen peroxide scavenging (Sakihama et al., 2002; Takahama et al.,
1999). These substances are also thought to participate in lignin
biosynthesis by serving as a source of caffeoyl and coumaroyl sub-
units that can be converted to one of the three phenylpropanoids
that comprise the lignin biopolymer (Chabannes et al., 2001;
Riboulet et al., 2009).
Chlorogenic acid (3) may be directly synthesized by the acyla-
tion of quinic acid (4) with caffeoyl-CoA. An indirect route pro-
ceeds via 4-coumaroyl quinate (4-2) ester which subsequently
becomes hydroxylated via the action of a P450 monooxygenase
(Lotfy et al., 1992; Niggeweg et al., 2004; Ulbrich and Zenk,
1979). In the direct route, the acylation reaction can be catalyzed
by a hydroxycinnamoyl-CoA:quinate hydroxycinnamoyl transfer-
ase (HQT) (Ulbrich and Zenk, 1979). This type of acyltransferase
belongs to the BAHD superfamily, a class of enzymes involved in
the modification of a large variety of plant secondary metabolites
(D’Auria, 2006).
In this study, the role of polyphenolic compounds in mediating
the storage of tropane alkaloids in E. coca was investigated. First, a
close correlation between tropane alkaloid and hydroxycinnamoyl
quinate ester levels was established in various organs of the plant.
Next, the BAHD acyltransferase responsible for producing
hydroxycinnamoyl quinate esters was characterized and expres-
sion localized for the corresponding gene. Finally, experiments
were conducted to demonstrate the association of tropane alka-
loids and hydroxycinnamoyl quinate esters in vitro which suggests
the likelihood of complexation in vivo.
2. Results and discussion
2.1. Content of tropane alkaloids and hydroxycinnamoyl quinate esters
are correlated throughout leaf and organ development in coca
There are two major tropane alkaloids in coca (E. coca), cocaine
(1) and cinnamoylcocaine (2). Previous work showed that these are
present together in high concentrations in leaves ranging from
0.5% of dry weight in mature leaves to over 10% of dry weight in
the youngest leaves (Docimo et al., 2012). Any compound that is
involved in the complexation of these alkaloids in storage should
be present in comparably high concentrations in the same organs
and growth stages. Since alkaloids have been previously reported
to be complexed with polyphenolic compounds, several develop-
mental stages of E. coca leaves were screened, as well as the stems
and the roots for the presence of polyphenolic compounds using
LC–MS. The only compounds that had a significant correlation with
cocaine (1) and cinnamoylcocaine (2) were the hydroxycinnamoyl
esters of quinic acid (HQA) (Fig. 2). The major HQA constituents in
E. coca were found to be chlorogenic acid (3) (caffeoyl quinate) and
4-coumaroyl quinate (4-3) which contribute 63% and 18% of the to-
tal HQAs present in stage one (young rolled) leaves, respectively.
Similar to previous reports, the highest levels of cocaine (1) and
cinnamoylcocaine (2) were found in the youngest leaves (stage
1) with an average quantity of 349
et al., 2012). The total amounts of HQAs in the youngest leaves
were 243
mol gꢀ1 dry wt. In mature (stage 3) leaves, the average
quantity of tropane alkaloids dropped to 50
mol gꢀ1 dry wt while
the total HQAs were down to 27
mol gꢀ1 dry wt with the ratio of
l
mol gꢀ1 dry wt (Docimo
l
l
l
4-coumaroyl quinate (4-3) to chlorogenic acid (3) at an average of
2:1. The roots, which do not contain tropane alkaloids at all, were
also devoid of HQAs. Such close correlation of alkaloid and
O
O
O
H
O
N
N
O
O
O
O
OH
O
O
O
O
H
O
H
O
H
O
Cocaine (1)
Cinnamoylcocaine (2)
Chlorogenic acid (3)
H
O
O
H
O
O
O
O
H
O
H
O
R
1
H
O
6
2
5
3
H
O
4
H
O
R3
R5
H
O
H
O
4-coumaric acid R = OH (5)
4-coumaroyl CoA R = S-CoA (5-1)
R4
H
O
Shikimic acid (6)
phenyllactic acid (7)
Name
R3 R4 R5 Number
OH OH OH (4)
Quinic Acid
3-O-4-coumaroylquinic acid
4-O-4-coumaroylquinic acid
5-O-4-coumaroylquinic acid
(5) OH OH (4-1)
OH (5) OH (4-2)
OH OH (5) (4-3)
Fig. 1. Structures of the main compounds utilized in this study.

J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
179
hydroxycinnamate ester content has also been reported for both
Coffea arabica (coffee) and Atropa acuminata plants (Aerts and
Baumann, 1994; Harborne and Khan, 1993). In coffee plants, the
correlation of alkaloids and HQAs can also be observed in the roots.
using Ni-chelate chromatography and subjected to enzyme assay.
Only one of the candidates, EcBAHD4, was capable of converting
4-coumaroyl Coenzyme A (5-1) and quinic acid (4) to form the
product 4-coumaroyl quinic acid (4-3) (Fig. 4), and was named
EcHQT (hydroxycinnamoyl quinate transferase). A total of three
peaks corresponding to the mass of 4-coumaroyl quinic acid
([M–H]ꢀ: m/z 337) were visible in the assays with EcHQT. Using
LC–MSn techniques (Clifford et al., 2003, 2008), it was possible to
ascertain that the minor peak 1 at a retention time of 10.4 min
was the 3-O-ester of trans-4-coumaroyl quinic acid (4-1) (MS2 of
m/z 337: base peak m/z 163) and the large peak at a retention time
of 14.0 min (peak 2 in Fig. 3) contained a mixture of both 5-O (4-3)
(MS2 of m/z 337: base peak m/z 191) and 4-O (4-2) (MS2 of m/z
337: base peak m/z 173) esters of trans 4-coumaroyl quinic acid
respectively. Peak 3 (retention time 15.5 min) was determined to
be the 5-O-ester (MS2 of m/z 337: base peak m/z 191) of cis-4-
coumaroyl quinic acid (Clifford et al., 2008). None of the other
EcBAHD proteins could carry out this reaction including EcBAHD3
(Fig. 4) which shares 48% identity with EcBAHD4 (EcHQT) on the
amino acid level.
2.2. Isolation and characterization of E. coca hydroxycinnamoyl-
CoA:quinate hydroxycinnamoyl transferase (HQT)
Since tropane alkaloid levels correlate with HQA content in E.
coca plants, the enzyme(s) responsible for HQA biosynthesis were
examined. Several plant HQT enzymes have been previously iso-
lated and characterized, and all shown to belong to the BAHD class
of acyltransferases (Lepelley et al., 2007; Lotfy et al., 1992;
Niggeweg et al., 2004). Since the highest levels of HQAs were found
in the young leaves, an E. coca young leaf kZAPII cDNA library was
utilized to screen for candidate hydroxycinnamoyl quinate trans-
ferase enzymes. Out of the original 8000 sequences obtained from
the library, 6 BAHD-like sequences were identified based upon
their gene ontology and similarity with other previously identified
BAHD acyltransferases.
In order to understand how EcHQT contributes to HQA levels
and a possible interaction between HQAs and tropane alkaloids
in coca, its biochemical properties were further investigated using
larger amounts of active enzyme obtained by heterologous expres-
sion in the yeast Pichia pastoris. To facilitate purification, the EcHQT
recombinant protein used for characterization experiments con-
tained an extra 8 amino acids at the N-terminus corresponding
to the introduction of a Strep II tag. The pH optimum of the EcHQT
enzyme using phosphate buffer was found to be pH 7.5 with activ-
ity reduced to 50% and 57% at pH 5.7 and 8.3, respectively. The use
of bis-tris propane and Tris–HCl buffers at pH 7.5 reduced the
activity to 78% and 50%, respectively, compared to that with phos-
phate buffer at pH 7.5. Such a reduction of activity in Tris based
buffers has been reported for other HQT enzymes (Lotfy et al.,
1992; Ulbrich and Zenk, 1979). The addition of various monovalent
or divalent metal ions did not enhance the activity significantly,
but Cu2+ as well as Fe2+ almost completely inhibited the enzyme
activity at 5 mM. Zn2+ also inhibited the activity to 78% of the max-
imal activity when used at a concentration of 5 mM. To assess tem-
perature stability, EcHQT was incubated for 30 min at 4, 10, 25, 35,
45 and 55 °C before being added to the enzyme assay. The enzyme
was stable at 4 and 10 °C, but activity was reduced to 65% and 31%
of maximal activity when incubated at 25 and 35 °C, respectively.
No activity was observed after incubation at higher temperatures.
Since E. coca contains a mixture of hydroxycinnamate quinate
esters, the ability of EcHQT to utilize a wide range of CoA thioesters
as well as acyl acceptors was investigated. When quinic acid (4) is
substituted with shikimic (6) or phenyllactic (7) acid, the relative
activities with 4-coumaroyl-CoA (5-1) are reduced to 3.8% and
0.7%, respectively (Table 1). Substitution of 4-coumaroyl-CoA (5-
1) with caffeoyl-CoA led to a reduction in relative activity to 22%.
No activity was detected using acetyl-CoA, cinnamoyl-CoA, hexa-
noyl-CoA or malonyl-CoA. Kinetic parameters were determined
for the best substrates, quinic acid (4) and 4-coumaroyl-CoA (5-
Although it has proven to be very difficult to assign function of
BAHD acyltransferases based on primary structure alone, the phy-
logenetic placement of related sequences provide hints as to their
possible biochemical activities. All six EcBAHD protein sequences
were thus aligned with 56 previously characterized BAHDs from
other plant species using the ClustalX program. The corresponding
alignment was subjected to Bayesian analysis which resulted in a
phylogenetic tree (Fig. 3). Of the six EcBAHD candidates, only Ec-
BAHD3 and EcBAHD4 cluster within the vicinity of several other
biochemically characterized hydroxycinnamoyl transferases which
utilize either shikimate and/ or quinate as the acyl acceptor. The
most closely related sequences include enzymes from Cynara
cardunculus, Nicotiana tabacum and Coffea canephora with amino
acid identities of 60%, 59% and 58% respectively.
To determine which if any of the six EcBAHD candidate genes
encoded an active hydroxycinnamoyl quinate transferase, the open
reading frames of all six genes were heterologously expressed in
Escherichia coli BL21 cells, and the recombinant proteins purified
500
R² = 0.92
400
300
200
100
0
Leaf Stage 1
Leaf Stage 2
Leaf Stage 3
Stem
1) (Table 2). The Km value for quinic acid (4) was 537.4 lM with
a turnover number of 1.69 sꢀ1. For 4-coumaroyl-CoA (5-1), sub-
0
50
100
150
200
250
300
350
strate inhibition was observed with an inhibition constant of
Hydroxycinnamoyl quinate esters [µmol g-1 dry wt]
3.28 mM. The Km value for 4-coumaroyl-CoA (5-1) was 18.5
lM
with a turnover number of 1.73 sꢀ1
.
Fig. 2. Correlation of tropane alkaloid content in E. coca with that of hydroxycin-
namoyl quinate esters (HQAs) in different organs and developmental stages. The
two major tropane alkaloids are cocaine and cinnamoylcocaine (2), while the two
major HQAs are caffeoyl quinate (3) (chlorogenic acid) and 4-coumaroyl quinate (4-
2). Leaf stage 1 (}) represents the youngest leaves which are still rolled. Leaf stage 2
(s) includes slightly older leaves that are almost completely unrolled but still
expanding. Leaf stage 3 (+) are mature and fully expanded. Stems are represented
by (ꢀ). Each dot represents one biological replicate. The relationship between
tropane alkaloids and HQAs is linear with an r2 of 0.92 and a P < 0.001.
Previously characterized HQT enzymes show similar kinetic
traits (Table 2). For example, HQT from N. tabacum has a Km value
of 727 lM for quinate (4) using 4-coumaroyl-CoA (5-1) as a sub-
strate (Niggeweg et al., 2004). Other HQT enzymes biochemically
characterized include those from C. cardunculus and Arabidopsis
thaliana which have similar substrate affinities to those reported
here for EcHQT (Hoffmann et al., 2003; Sonnante et al., 2010). It

180
J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
V
II
III
IV
0.5
I
Fig. 3. A phylogenetic tree of the BAHD family of acyltransferases updated from D’Auria (2006) illustrating the affinities of the six E. coca sequences isolated in this study.
Enzymes included in this tree have been functionally characterized by either genetic mutant screening or biochemical assay (accession numbers are listed in Section 4.6 and
in more detail in Supplementary Table 1). Two of the E. coca sequences, EcBAHD3 and EcBAHD4 (later characterized as EcHQT and highlighted in red) cluster within the
vicinity of other hydroxycinnamoyl transferases. The color scheme used represents clades I through V according to the scheme reported by D’Auria (2006) with the addition of
a lighter shade in clade V representing HQT-like enzymes. The distance bar represents the number of amino acid changes per site.
is not possible to predict whether an HQT-like enzyme is capable of
using either shikimate (6) or quinate (4) based on sequence simi-
larity alone. In fact, in some plant species separate enzymes carry
out the two reactions separately, while in other plants the reac-
tions can be catalyzed by the same enzyme (Petersen et al.,
2009). In coca however, no shikimate esters have been found to
date. Additionally, the relative activity of EcHQT with shikimate
(6) suggests that this enzyme is most likely a true HQT.
Substrate inhibition via hydroxycinnamoyl-CoA has been re-
ported once before for a member of the BAHD family, a rosmarinic
acid synthase from Coleus blumei (Sander and Petersen, 2011). This
property may play a role in regulating the amount and proportions
of HQAs in plants, as well as the rate of lignin formation and its
composition, since HQAs are associated with lignin deposition
and their biosynthesis consumes hydroxycinnamoyl-CoAs which
are intermediates to the major lignin monomers. Measurements
of the internal pool sizes of hydroxycinnamoyl-CoA esters in differ-
ent organs and tissues should help to understand the significance
of substrate inhibition. Although three possible pathways for the
biosynthesis of chlorogenic acid (3) have been proposed, the
substrate specificity of E. coca HQT supports the idea that it is
synthesized via a coumaroyl quinic acid intermediate in this spe-
cies (Niggeweg et al., 2004; Stöckigt and Zenk, 1974; Ulbrich and
Zenk, 1979).
2.3. EcHQT gene expression
Given the possible role of HQAs in tropane alkaloid complexa-
tion, the localization of EcHQT transcript in E. coca organs was
investigated (Fig. 5). The highest EcHQT mRNA levels were found
in stage 1 leaves with a 13-fold higher expression compared to
stage 3 leaves. This pattern correlated well with metabolite analy-
sis in which hydroxycinnamate quinate esters in the stage 1 leaves
were found to be approximately 10 times more abundant than in
stage 3 leaves. A close correlation between EcHQT transcripts and
HQA metabolite levels was not seen in stems. Maturing stems
are known to be active in lignin deposition (Grabber, 2005) which
consumes large amounts of chlorogenic acid (3) and its precursor
4-coumaroyl quinate (4-3). Neither HQA metabolites nor EcHQT
transcripts could be detected in coca root tissue, which coincides
with the absence of cocaine and other tropane alkaloids in coca
roots. Interestingly, plants producing and storing alkaloids in their

J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
181
Table 2
1.5
1.0
0.5
0.0
EcHQT
Kinetic parameters of EcHQT and related enzymes from Nicotiana tabacum and Cynara
cardunculus. The substrate used in
parentheses.
a constant concentration is indicated in
KM
(lM)
kcat
kcat/KM
Mꢀ1 sꢀ1
)
KI (mM)
(sꢀ1
)
(
l
EcHQT
EcHQT
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
Quinic acid (4) (4-
coumaroyl-CoA)
4-Coumaroyl-CoA (5-1)
(quinic acid)
537.4 122
18.5 1.9
1.69
1.73
0.003
0.093
–
Boiled Control
3.28 1.01
NtHQT
Quinic acid (4) (4-
coumaroyl-CoA)
CcHQT1
Quinic acid (4) (4-
coumaroyl-CoA)
4-Coumaroyl-CoA (5-1)
(quinic acid)
727 150
4.62a 0.006a
–
EcBAHD3
216 76
16.93a 0.078a
11.68a 0.011a
–
–
1034 77
0
5
10
15
20
25
30
CcHQT2
Time [min]
Quinic acid (4) (4-
coumaroyl-CoA)
4-Coumaroyl-CoA (5-1)
(quinic acid)
59
4
21.82a 0.370a
5.73a 0.005a
–
–
Fig. 4. HPLC-MS analysis of acyltransferase assays with recombinant protein using
coumaroyl-CoA and quinic acid as substrates. EcHQT (EcBAHD4) was demonstrated
to catalyze this reaction, but assays of boiled EcHQT and EcBAHD3 protein showed
no activity. Total ion chromatograms obtained in the negative mode are depicted.
Peak 1: 3-O-ester of trans-4-coumaroyl quinic acid (4-1), peak 2: a mixture of both
5-O and 4-O esters of trans 4-coumaroyl quinic acid (4-2) and (4-3), peak 3: 5-O-
ester of cis-4-coumaroyl quinic acid.
1065 70
a
Calculated from data in the original publications: NtHQT (Niggeweg et al.,
2004), CcHQT1 and CcHQT2 (Sonnante et al., 2010).
By measuring the spectral shift of chlorogenic acid (3) with
increasing concentrations of cocaine (1), it is possible to calculate
the strength and stoichiometry of the interaction from the Beer–
Lambert Law and the equilibrium constant (Kappeler et al., 1987;
Sondheimer et al., 1961). Using six different concentrations of
cocaine (1) and a wavelength of 364 nm, the molar absorptivity
Table 1
Catalytic activity of EcHQT with various substrates relative to activity with quinic acid
and 4-coumaroyl-CoA, which had a specific activity of 8.54 nkat (mg protein)ꢀ1
.
Substrate 1
Substrate 2
Relative activity (%)
Quinic acid (4)
Shikimic acid (6)
Phenyllactic acid (7)
Quinic acid (4)
Quinic acid (4)
Quinic acid (4)
Quinic acid (4)
Quinic acid (4)
4-Coumaroyl-CoA (5-1)
4-Coumaroyl-CoA (5-1)
4-Coumaroyl-CoA (5-1)
Acetyl-CoA
Cinnamoyl-CoA
Caffeoyl-CoA
100
3.8
0.7
0
0
22
0
(the molar extinction coefficient, eC) was calculated and plotted
vs. the inverse of the equilibrium constant (KꢀC 1) (Fig. 7B). Based
on the mean of the intersection points, the equilibrium constant
and the molar absorptivity were determined to be 9.1 1.4 mol Lꢀ1
and 1901 87 L molꢀ1 cmꢀ1, respectively. The identification of 15
valid intersection points suggests that the model used is accurate
and that cocaine (1) and chlorogenic acid (3) associate in a 1:1 ratio
(Fig. 7B). The equilibrium constant determined previously for the
caffeine-chlorogenate complex, which was also shown to associate
Hexanoyl-CoA
Malonyl-CoA
0
root tissues, such as N. tabacum and C. arabica, are reported to
accumulate both HQAs and transcripts encoding their respective
hydroxycinnamoyl transferases in the roots (Lepelley et al., 2007;
Niggeweg et al., 2004).
600
500
400
300
200
100
0
25
20
15
10
5
Hydroxycinnamoyl quinate esters
EcHQT transcript
2.4. UV–Vis and NMR spectral studies demonstrate the interaction of
cocaine and chlorogenic acid
Earlier work had reported evidence for a physical interaction
between caffeine and chlorogenic acid (3) which was thought to
be important for the storage of caffeine in the vacuole (Kappeler
et al., 1987). A similar approach was used herein in E. coca to look
for a physical interaction between cocaine (1) and chlorogenic acid
(3) starting by measuring the UV spectrum of chlorogenic acid (3)
when incubated in the presence of cocaine (1). Fig. 6A shows the
absorbance shift induced by incubating both cocaine (1) and
chlorogenic acid (3) together. To understand which function on
the cocaine (1) molecule has the greatest ability to interact with
chlorogenic acid (3), tropine and benzoic acid were tested sepa-
rately. While the addition of tropine has a negligible effect on the
spectrum of chlorogenic acid (3) (Fig. 6B), the addition of benzoic
acid induces a strong absorbance shift (Fig. 6C). These results
strongly suggest that the interaction between cocaine (1) and
chlorogenic acid (3) is based on the benzoic acid moiety of cocaine
(1).
ND
0
1
2
3
Stem
Root
Leaf Stage
Fig. 5. Transcript levels of EcHQT correlate with levels of hydroxycinnamoyl
quinate esters in most organs of four month old E. coca plants. The values represent
the average of three biological replicates. Transcript measurements are normalized
to leaf stage 3. Error bars represent standard deviations. Neither EcHQT transcript
nor hydroxycinnamoyl quinate esters were detected in the roots.

182
J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
in a 1:1 ratio, was 47.3 9.5 L molꢀ1 (Kappeler et al., 1987). This
CGA : Cocaine
3.5
3
A
corresponds to
a DG value for the association reaction of
HO
COOH
O
ꢀ9.56 kJ molꢀ1 for the caffeine-chlorogenate complex and
CO2Me
O
N
O
O
ꢀ5.47 kJ molꢀ1 for the cocaine-chlorogenate complex.
2.5
2
O
H
Further evidence for the interaction between cocaine (1) and
chlorogenic acid (3) was obtained from 1H NMR experiments.
OH
OH
OH
Chemical shift differences (Dd) were measured in the proton spec-
1.5
1
trum of chlorogenic acid (3) in the presence of excess amounts of
cocaine (1) and in the proton spectrum of cocaine (1) in the pres-
ence of excess amounts of chlorogenic acid (3) (Fig. 8). 1H,1H COSY
experiments were performed to accurately assign all of the over-
CGA : Cocaine
CGA
Cocaine
0.5
0
lapping signals (see Supplemental Fig. 1). The highest
Dd values
were generally observed for those protons near to or attached to
the aromatic rings of both molecules. However, the two highest
shifts corresponded to protons at C2 (80.4 Hz) and C3 (85.3 Hz)
of the tropane ring. All chemical shift differences detected were to-
wards higher field, which means greater nuclear shielding caused
by increased magnetic anisotropy due to the induced field from
350
355
360
365
370
375
380
Wavelength [nm]
CGA : Tropine
3.5
3
B
HO
COOH
H3C
N
O
the overlap of the
p orbitals of the two compounds, as proposed
OH
O
H
O
2.5
2
for the caffeine-chlorogenate complex (D’Amelio et al., 2009;
Horman and Viani, 1972). The structural similarity of 4-coumaroyl
quinate (4-3) to chlorogenic acid (3) should allow the complexa-
tion of 4-coumaroyl quinate (4-3) with cocaine (1) in a correspond-
ing manner, especially since the hydroxyl groups responsible for
the difference between the two HQAs were not found to be
significantly involved in these interactions. Similarly, complexa-
tion between cinnamoylcocaine (2) and chlorogenic acid (3) would
also be likely.
OH
OH
OH
1.5
1
CGA : Tropine
CGA
Tropine
0.5
0
350
355
360
365
370
375
380
Wavelength [nm]
3. Conclusions
CGA : Benzoic Acid
In the present study, three lines of evidence were obtained
suggesting that hydroxycinnamate esters of quinate complex
with tropane alkaloids in E. coca and thus facilitate storage in
vacuoles. First, there was a strong correlation between the con-
tent of the hydroxycinnamate quinate esters (HQAs), chlorogenic
acid (3) and 4-coumaroyl quinate (4-3), and the tropane
alkaloids, cocaine (1) and cinnamoylcocaine (2), in various or-
gans of E. coca. In addition, in vitro complexation of chlorogenic
acid and cocaine was demonstrated independently by shifts in
UV absorption maxima and 1H NMR signals. The enzyme respon-
sible for quinate ester formation from hydroxycinnamoyl-CoA
and quinic acid, EcHQT, was also isolated and shown to belong
to the BAHD family of acyltransferases (D’Auria, 2006). The
expression of the gene encoding EcHQT was found to correlate
with tropane alkaloid accumulation.
3.5
3
C
HO
COOH
HO
O
O
O
O
H
2.5
2
OH
OH
OH
CGA : Benzoic Ac.
CGA
Benzoic Ac.
1.5
1
0.5
0
350
355
360
365
370
375
380
Wavelength [nm]
Fig. 6. UV–Vis spectrophotometry: absorbance shift measurements in the spectrum
of chlorogenic acid (0.2 mM) in the presence of (A) cocaine (1) (4 mM), (B) tropine
(4 mM) and (C) benzoic acid (4 mM). Solid lines represent measurements obtained
from mixed samples, whereas dotted and dashed lines represent single compound
spectra. No absorbance shift was detected when chlorogenic acid was incubated
with tropine.
The biological roles of HQA-tropane alkaloid complex formation
remain to be ascertained. However, previous work on coffee plants
indicated that complexation between chlorogenic acid (3) and
caffeine increases the apparent solubility of caffeine (Zeller and
Saleeb, 1997). The highest levels of caffeine in coffee plants are
approximately 2% dry weight in the cotyledons (Zheng and
Ashihara, 2004). Since the levels of tropane alkaloids in E. coca
can be even higher, over 10% dry weight, a similar mechanism
may be required to achieve vacuolar storage. The formation of a
complex between tropane alkaloids and HQAs may aid in vacuolar
storage by withdrawing alkaloids from the transport equilibrium.
Another role for complex formation may be to prevent the leakage
of tropane alkaloids from the vacuole by changing their partition
coefficients in a similar fashion to what was shown in studies using
caffeine (Waldhauser and Baumann, 1996). In fact, the authors of
this study suggest that complex formation may be the driving force
behind transporting purine alkaloids into the vacuole. Testing the
biological importance of HQA-tropane alkaloid complexation
in vivo will require independent manipulation of tropane alkaloid
or HQA content.
4. Experimental
4.1. General experimental procedures
LC–MS was performed using a HPLC 1100 series chromatograph
system (Agilent Technologies, Böblingen, Germany) equipped with
a diode array detector and coupled to an Esquire 6000 ESI-Ion elec-
trospray ion trap mass spectrometer (Bruker Daltonics, Bremen,
Germany) for cocaine (1) and chlorogenic acid derivatives. For
the EcHQT assay, an API 3200 tandem mass spectrometer (Applied
Biosystems) equipped with a turbo spray ion source was used. 1H
NMR spectra were measured on a DRX 500 NMR spectrometer
(Bruker-BioSpin, Rheinstetten, Germany) equipped with an inverse
probe (5 mm).

J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
183
6 5 4 3 2
1
0.6
0.5
0.4
0.3
0.2
0.1
0
300
A
B
5
6
1 (25mM)
4
3
250
200
150
100
50
2 (100mM)
3 (150mM)
4 (200mM)
5 (250mM)
6 (300mM)
2
1 (25mM)
2 (100mM)
3 (150mM)
4 (200mM)
5 (250mM)
6 (300mM)
1
0
0
1
2
3
-1
4
5
350
355
360
365
370
375
380
-3
-1
εC 10 [L cm mol ]
Wavelength [nm]
Fig. 7. (A) Bathochromic shift measurements in the spectrum of chlorogenic acid (3) (0.2 mM) in the presence of increasing concentrations of cocaine (1) (lines 1–6).
Quantification of the interaction between cocaine (1) and chlorogenic acid (3) was performed at 364 nm, represented by the vertical line in panel A. (B) Curve molar
absorptivity (
e
C) vs. inverse equilibrium constant (Kꢀ1) plots. The data was based on the values recorded from panel A (lines 1–6).
e
C and KC were calculated from the mean of
C
the intersection points (s) of lines 1 through 6 using the equations described by Kappeler et al. (1987). eC and KC were determined to be 1901 87 L molꢀ1 cmꢀ1 and
9.1 1.4 L molꢀ1, respectively (dotted lines).
O
-26.3
-8.5
-80.4
N
-40.0
O
O
-19.0
O
O
HO
OH
-22
-85.3
-68.9
-35.8
-22
O
-62.6
-68.9
-35.8
-60.1
-37.5
-25.0
O
HO
-25.9
-65.9
-34.4
OH
OH
-62.6
-100
OH
Fig. 8. Complex formation between chlorogenic acid (3) and cocaine (1) is indicated by chemical shift differences (
in the presence of the other as compared to a reference compound measured in pure form (left structure: chlorogenic acid (3); right structure: cocaine (1)). Proton shifts are
represented by circles at the attached carbon atom that differ in size according to the magnitude of the shift. The exact values are shown in Hz next to the circle. All d values
D
d) shifts observed in the 1H NMR spectra of each molecule
D
are given in Hz and are negative because the shifts were always towards high field (lower d values). Some protons could not be assigned due to overlap. A circle representing
100 Hz is included for scale.
4.2. Plant growth conditions and extraction
volume was 20 lL. Quantitative chemical analysis of plant tissue
extracts was carried out on a Agilent 1100 HPLC with diode array
detector (Agilent Technologies, Böblingen, Germany) with UV
detection at 235, 280 and 320 nm for cocaine (1), cinnamoylco-
caine and HQAs, respectively. The column and chromatographic
conditions were the same as in the qualitative analysis, using
0.05% CF3CO2H instead of HCO2H, and an injection volume of
E. coca plants were grown as reported previously (Docimo et al.,
2012). Leaves, roots and stems from E. coca plants were frozen in
N2, ground and extracted in 1 mL of 5% MeOH, 0.1% HCO2H in H2O.
4.3. Chromatographic analysis of plant extracts
40 lL. Compounds were quantified through external standard cal-
Qualitative LC–MS analysis was carried out on a Agilent 1100
HPLC (Agilent Technologies, Böblingen, Germany) equipped with
ibration curves recorded with authentic standards for cocaine and
chlorogenic acid. 4-Coumaroyl quinate (4-(1-3)) was quantified
using chlorogenic acid (3) as an external standard using a molar
response factor of 0.79, which was experimentally determined as
follows: standard curves were created for caffeic acid and chloro-
genic acid (3) on a molar basis at 320 nm resulting in the same
slope for both regression lines. Thus it was concluded that a
response factor of 1 can be used between the hydroxycinnamic
acid and its quinate ester. In a second step, standard curves on a
molar basis were created for caffeic acid and 4-coumaric acid (5)
at 320 nm. The slopes for the two compounds were different (slope
a Nucleodur Sphinx RP 5
lm column (25 cm ꢁ 4.6 mm, 5
lm,
Macherey–Nagel, Düren, Germany) coupled to an Esquire 6000
mass spectrometer, operated in alternating ion polarity mode in
the range m/z 60–1000 with a target mass m/z 300; nebulizer pres-
sure, 40 psi; drying gas flow, 11 L/min; gas temperature, 330 °C;
capillary voltage, 113.5 V). The mobile phase gradient used
consisted of 0.2% HCO2H (A) and CH3CN (B) at a flow rate of
1 mL/min as follows: 90–69% A (21 min), 69–45% A (6 min),
45–0% A (0.1 min), 0–90% A (2 min), 90% A (3.9 min), the injection

184
J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
caffeic acid: 4523, slope 4-coumaric acid: 5736), resulting in a
response factor of 0.79. As the response factor between the
hydroxycinnamic acid and its quinate ester is 1, it was deduced a
response factor at 320 nm of 0.79 for 4-coumaroylquinic acid
(4-(1-3)) in relation to chlorogenic acid (3). Cinnamoylcocaine (2)
was quantified using cinnamic acid as external standard using a
molar response factor of 1.
AAV66311, DBNTBT-AAM75818, BAPT-AAL92459, DBBT-Q9FPW3,
TAT-AAF34254, DBAT-AAF27621, Cer2-AAM64817, Glossy2-
CAA61258, AMAT-AAW22989, NtMAT1-BAD93691, MpAAT1-
AAU14879,
VAAT-AF193790_1,
HMT/HLT-BAD89275,
BanAAT-AX025506,
RhAAT1-AAW31948,
At5MAT-NM_113880,
AtSHT-NP_179497, CsHCT-ACF37072, AtHHT1-NP_851111, Os-
MAT1-NP_001046857, TpHCT2-ACI16631, AtACT-NP_200924,
AtSDT-NP_179932, AtSCT-AAP81804, GmIF7MaT-NP_001237760,
4.4. Chemicals
CbRAS-CAK55166,
EcBAHD2-JQ413185,
CcHQT-ABO77956,
EcBAHD3-JQ413186,
EcBAHD1-JQ413184,
EcHQT-JQ413187,
4-Coumaroyl-CoA (5-1) was synthesized as previously de-
scribed (Wang et al., 2000), with the following exceptions. The
activated succinimide ester of 4-coumaric acid was first purified
using preparative TLC, employing CHCl3/MeOH (20:1) as the mo-
bile phase before proceeding to the reaction with Coenzyme A. Fol-
lowing synthesis, the 4-coumaroyl-CoA (5-1) was purified using
solid-phase extraction cartridges (500 mg Chromabond C 18 ec,
Macherey–Nagel) preconditioned with consecutive washes of
methanol, dH2O, and 20 mM (NH4)OAc solution (5 column vol
each). After evaporation of the organic solvent, (NH4)OAc was
added to the water phase to a final concentration of 20 mM, and
the mixture was loaded onto the SPE cartridge. The column was
rinsed with 5 column volumes of 20 mM (NH4)OAc solution, to
elute unreacted Coenzyme A. The 4-coumaroyl-CoA (5-1) was
recovered by elution with distilled H2O. Pure fractions containing
only 4-coumaroyl-CoA (5-1) were identified by HPLC-MS and
HPLC-UV, and lyophilized overnight.
EcBAHD5-JQ413188, and EcBAHD6-JQ413189.
4.7. Expression and purification of EcHQT
The EcHQT open reading frame was ligated into a pPICHOLI vec-
tor (MoBiTec, Göttingen, Germany) modified with a Gateway-
compatible N-terminal Strep Tag II. Introduction of the plasmid into
P. pastoris strain KM71 via electroporation was performed according
to the pPICHOLI manual. Expression, lysis, purification and
SDS–PAGE electrophoresis analysis were performed as previously
described (Jirschitzka et al., 2012) with the following exceptions.
The forward primer GAC TGG TTC CAA TTG ACA AGC and the reverse
primer GCA AAT GGC ATT CTG ACA TCC were used for cloning. The
purified fraction of the StrepTrap HP column was run through a
Resource Q 1 ml column (GE Healthcare, Freiburg, Germany) and
was eluted in 20 column volumes of a NaCl (0–500 mM) linear
gradient in Tris–HCl (100 mM), 10% (v/v) glycerol, pH 8.0 buffer.
Quinic acid (4) was purchased from Alfa Aesar (Karlsruhe,
Germany). The rest of the chemicals and solvents were purchased
from Sigma–Aldrich (Munich, Germany), Carl Roth (Karlsruhe, Ger-
many), Merck (Darmstadt, Germany) or VWR (Leuven, Belgium).
4.5. Identification of EcHQT
4.8. Characterization and kinetics of EcHQT, chemical analysis of
enzyme assays
Six sequences from an E. coca young leaf cDNA library (Docimo
et al., 2012), identified as putative BAHD acyltransferases genes,
were cloned into the Gateway expression vector system and ex-
pressed in E. coli BL21 (DE3) cells (Invitrogen), using the pH9GW
vector, a modified pET-T7 (28a) vector (Merck) with a N-terminal
His-Tag which is Gateway compatible (Yu and Liu, 2006). The for-
ward primer TAA TAC GAC TCA CTA TAG GG and the reverse primer
CCC AAG GGG TTA TGC TAG TT were used for the cloning. Heterol-
ogous expression of the proteins and purification via Ni-chelate
chromatography was performed as previously described (D’Auria
et al., 2007).
Biochemical characterization and determination of the kinetic
parameters was achieved as described (D’Auria et al., 2002) with
the following exceptions. K2HPO4/KH2PO4 buffer, bis-tris propane
buffer and Tris–HCl buffer were tested in the pH ranges 5.7–8.3,
6.7–9.5 and 7.3–9.1, respectively. The assays were stopped adding
HCl (1 M) to pH 1.0 and the analysis was carried out on a Agilent
1200 series equipment (Agilent Technologies, Böblingen, Germany)
coupled to an API 3200 tandem mass spectrometer (Applied Bio-
systems, Darmstadt, Germany) equipped with a turbo spray ion
source, using
(5 cm ꢁ 4.6 mm, 1.8
many). Injection volume was 5
8 min, at 20 °C and a flow rate of 800
a
ZORBAX RRHT Eclipse XDB-C18 column
m) (Agilent Technologies, Böblingen, Ger-
L, separation was achieved within
L/min, using HCO2H 0.05%
l
4.6. Phylogenetic analysis of BAHD family members
l
l
Analysis of protein sequences (see Supplementary Table 1 for
accession numbers) were aligned using ClustalX, and the resulting
alignment was submitted to Bayesian analysis using Mr. Bayes
v3.1.2 with two MCMC runs each of 10,14,311 generations
(Huelsenbeck and Ronquist, 2001). The 50% consensus tree was
visualized as an unrooted tree using FigTree v1.3.1 (http://tree.
bio.ed.ac.uk/). The following enzymes and their corresponding
Genbank accession numbers were used for the analysis
(enzyme-accession number): DAT-AAC99311, MAT-AAO13736,
CbBEAT-AAC18062, CbBEBT-AAN09796, NtBEBT-AAN09798, CHAT-
AAN09797, BPBT-AAU06226, HCBT-CAB06430, SAAT-AAG13130,
Vinorine synthase-CAD89104, SalAT-AAK73661, Ss5MaT2-AA
R26385, Gt5AT-BAA74428, Dm3MAT2-AAQ63616, Dm3MAT1-
AAQ63615, Dv3MAT-AAO12206, Vh3MAT1-AAS77402, Lp3MAT1-
AAS77404, Ss5MaT1-AAL50566, Sc3MaT-AAO38058, Pf5MaT-AA
L50565, Pf3AT-BAA93475, ACT-AAO73071, NtHCT-CAD47830, AtH-
CT-NP_199704, AsHHT1-BAC78633, NtHQT-CAE46932, CmAAT
(1-4)-CAA94432; AAL77060; AAW51125; AAW51126, Pun1-
(A) and CH3CN (B) as mobile phases as follows: 95% A (0.5 min),
95–40% A (3.5 min), 40–0% A (0.1 min), 0% A (0.9 min), 0–95% A
(0.1 min), 95% A (2.9 min); injection volume was 5 lL. The mass
spectrometer was operated in negative ionization mode; curtain
gas, 25 psi; turbo heater temperature, 700 °C; nebulising gas,
70 psi; heating gas, 60 psi; collision gas, 6 psi; ion spray voltage,
ꢀ4500 eV. Analytes were monitored by scheduled multiple reac-
tion monitoring (MRM): chlorogenic acid (3) m/z 353.13 ?
190.88; 4-coumaroyl quinic acid m/z 337.13 ? 190.88; decluster-
ing potential (DP), ꢀ35 V; entry potential (EP), ꢀ4 V; collision cell
entry potential (CEP), ꢀ36 V; collision energy (CE), ꢀ22 V; collision
cell exit potential (CXP), ꢀ4 V. Quantification was based on a stan-
dard curve of authentic chlorogenic acid (3) (Sigma–Aldrich).
Enzyme assays where quinic acid (4) was substituted with shi-
kimic (6) or phenyllactic acid (7) were analyzed according to Sup-
plemental Table 2, while assays with caffeoyl-CoA, acetyl-CoA,
cinnamoyl-CoA, hexanoyl-CoA, and malonyl-CoA were analyzed
according to Supplemental Table 3.

J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
185
Asano, N., Yokoyama, K., Sakurai, M., Ikeda, K., Kizu, H., Kato, A., Arisawa, M., Höke,
D., Dräger, B., Watson, A.A., Nash, R.J., 2001. Dihydroxynortropane alkaloids
from calystegine-producing plants. Phytochemistry 57, 721–726.
Baumann, T.W., Rodriguez, M.F., Kappeler, A.W., 1991. Chlorogenic acid in leaf discs,
suspension-cultured cells, and protoplasts of coffee (Coffea arabica l.).
Physiological role and subcellular localization. ASIC, 14e Colloque Scientifique
International sur le Café, San Francisco, CA, USA.
4.9. Quantitative real-time PCR analysis
Experiments were performed as described (Docimo et al., 2012).
Primers targeting the EcHQT transcript were designed (Fwd: 50-CG
ATGTGGAGGAGTCTGTCTTGG-30, Rev: 50-CCGTCTTGTTGTTTGGAGG
TTGC-30) and standard curve analysis showed a PCR efficiency of
77% and R2 of 0.996. Expression of EcHQT was normalized to gene
6409 (Genbank Accession No. JN020150) and gene 10,131 (Gen-
bank Accession No. JN020153) (Docimo et al., 2012) expression
using qBASE v1.3.5 (Hellemans et al., 2007).
Bieri, S., Brachet, A., Veuthey, J.-L., Christen, P., 2006. Cocaine distribution in wild
Erythroxylum species. J. Ethnopharmacol. 103, 439–447.
Chabannes, M., Ruel, K., Yoshinaga, A., Chabbert, B., Jauneau, A., Joseleau, J.-P.,
Boudet, A.-M., 2001. In situ analysis of lignins in transgenic tobacco reveals a
differential impact of individual transformations on the spatial patterns of
lignin deposition at the cellular and subcellular levels. Plant J. 28, 271–282.
Clifford, M.N., Johnston, K.L., Knight, S., Kuhnert, N., 2003. Hierarchical scheme for
LC–MSn identification of chlorogenic acids. J. Agric. Food Chem. 51, 2900–2911.
Clifford, M.N., Kirkpatrick, J., Kuhnert, N., Roozendaal, H., Salgado, P.R., 2008. LC–
MSn analysis of the cis isomers of chlorogenic acids. Food Chem. 106, 379–385.
D’Amelio, N., Fontanive, L., Uggeri, F., Suggi-Liverani, F., Navarini, L., 2009. NNMR
reinvestigation of the caffeine-chlorogenate complex in aqueous solution and in
coffee brews. Food Biophys. 4, 321–330.
4.10. UV and NMR spectroscopy
The UV spectra of cocaine hydrochloride (1) (4 mM) and chloro-
genic acid (3) (0.2 mM) were measured separately as well as
together in solution in a Shimadzu UV-2501PC spectrometer
(Duisburg, Germany) in quartz cuvettes QS 104.002B-QS (Hellma
Analytics, Müllheim, Germany) in the range 350–380 nm. The same
procedure was followed using tropine (4 mM) and benzoic acid
(4 mM) instead of cocaine (1). The equilibrium constant, KC, and
D’Auria, J.C., 2006. Acyltransferases in plants: a good time to be BAHD. Curr. Opin.
Plant Biol. 9, 331–340.
D’Auria, J.C., Chen, F., Pichersky, E., 2002. Characterization of an acyltransferase
capable of synthesizing benzylbenzoate and other volatile esters in flowers and
damaged leaves of Clarkia breweri. Plant Physiol. 130, 466–476.
D’Auria, J.C., Reichelt, M., Luck, K., Svatoš, A., Gershenzon, J., 2007. Identification and
characterization of the bahd acyltransferase malonyl coa: anthocyanidin 5-O-
glucoside-600-O-malonyltransferase (At5MAT) in Arabidopsis thaliana. FEBS Lett.
581, 872–878.
Deus-Neumann, B., Zenk, M.H., 1984. A highly selective alkaloid uptake system in
vacuoles of higher plants. Planta 162, 250–260.
Deus-Neumann, B., Zenk, M.H., 1986. Accumulation of alkaloids in plant vacuoles
does not involve an ion-trap mechanism. Planta 167, 44–53.
the extinction coefficient (eC, molar absorptivity) at a chosen wave-
length (364 nm) were obtained as a mean of the intersection points
of different
e
C vs Kꢀ1 curves as described (Kappeler et al., 1987).
C
KC
CGA þ TA ꢀ C
ð1Þ
ð2Þ
Docimo, T., Reichelt, M., Schneider, B., Kai, M., Kunert, G., Gershenzon, J., D’Auria, J.,
2012. The first step in the biosynthesis of cocaine in Erythroxylum coca: the
characterization of arginine and ornithine decarboxylases. Plant Mol. Biol. 78,
599–615.
1
Kc
½TAꢂ0½CGAꢂ0
AðkÞ
AðkÞ
¼
ꢃ e0 ꢀ ½TAꢂ0 ꢀ ½GCAꢂ0 þ
e0
Ferreira, J.F.S., Duke, S.O., Vaughn, K.C., 1998. Histochemical and immuno-
cytochemical localization of tropane alkaloids in Erythroxylum coca var. Coca
and e. Novogranatense var. Novogranatense. Int. J. Plant Sci. 159, 492–503.
Grabber, J.H., 2005. How do lignin composition, structure, and cross-linking affect
degradability? A review of cell wall model studies. Crop Sci. 45, 820–831.
Harborne, J.B., Khan, M.B., 1993. Variations in the alkaloidal and phenolic profiles in
the genus Atropa (Solanaceae). Bot. J. Linn. Soc. 111, 47–53.
Hellemans, J., Mortier, G., De Paepe, A., Speleman, F., Vandesompele, J., 2007. Qbase
relative quantification framework and software for management and
automated analysis of real-time quantitative pcr data. Genome Biol. 8, R19.
Hoffmann, L., Maury, S., Martz, F., Geoffroy, P., Legrand, M., 2003. Purification,
cloning, and properties of an acyltransferase controlling shikimate and quinate
ester intermediates in phenylpropanoid metabolism. J. Biol. Chem. 278, 95–
103.
Horman, I., Viani, R., 1972. The nature and conformation of the caffeine-
chlorogenate complex of coffee. J. Food Sci. 37, 925–927.
Huelsenbeck, J.P., Ronquist, F., 2001. Mrbayes: Bayesian inference of phylogenetic
trees. Bioinformatics 17, 754–755.
Jirschitzka, J., Schmidt, G.W., Reichelt, M., Schneider, B., Gershenzon, J., D’Auria, J.C.,
2012. Plant tropane alkaloid biosynthesis evolved independently in the
Solanaceae and Erythroxylaceae. Proc. Natl. Acad. Sci. USA 109, 10304–
10309.
where CGA, chlorogenic acid (3); TA, tropane alkaloid (namely
cocaine (1)); C, complex of both compounds.
Reference spectra of chlorogenic acid (3) and cocaine (1)
(0.2 mm each) were measured as references against the external
standard TMSP-d4 prior to the complexation experiments.
The shift differences
(D
d) in the 1H NMR spectrum of
chlorogenic acid (3) (0.2 mM) in the presence of cocaine (1) hydro-
chloride (50 mM) and vice versa were measured in deuterated
phosphate buffer (10 mM, pH 7.5) at 27 °C. Chemical shift differ-
ences (D
d) in the 1H NMR spectra were determined by comparing
the chemical shifts of the non-complexed reference and the result-
ing chemical shifts after complexation (see Supplementary Mate-
rial). Water suppression was achieved using the PURGE sequence
as described (Simpson and Brown, 2005).
Acknowledgements
Kappeler, A.W., Baumann, T.W., Greutert, H., 1987. Complexation of purine alkaloids
in the coffee plant. ASIC, 12e Colloque Scientifique International sur le Café,
Montreux, Switzerland.
Kutchan, T.M., 2005. A role for intra- and intercellular translocation in natural
product biosynthesis. Curr. Opin. Plant Biol. 8, 292–300.
Lepelley, M., Cheminade, G., Tremillon, N., Simkin, A., Caillet, V., McCarthy, J., 2007.
Chlorogenic acid synthesis in coffee: an analysis of CGA content and real-time
rt-pcr expression of HCT, HQT, C3H1, and CCOAMT1 genes during grain
development in c. Canephora. Plant Sci. 172, 978–996.
Lotfy, S., Fleuriet, A., Macheix, J.-J., 1992. Partial purification and characterization of
hydroxycinnamoyl CoA: transferases from apple and date fruits.
Phytochemistry 31, 767–772.
We thank Dr. Michael A. Phillips for his help with the E. coca
cDNA library, Katrin Luck, Dr. Teresa Docimo and Jan Jirschitzka
for technical assistance. In addition, we thank Franziska Dolke for
synthesizing 4-coumaroyl-CoA, Derek Mattern for helping with
the editing of the manuscript, and Dr. Tamara Krügel, Andreas
Weber and the rest of the gardening staff of the MPI-ICE for their
help in plant maintenance.
Appendix A. Supplementary data
Nathanson, J.A., Hunnicutt, E.J., Kantham, L., Scavone, C., 1993. Cocaine as
naturally occurring insecticide. Proc. Natl. Acad. Sci. USA 90, 9645–9648.
a
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.phytochem.2012.
09.009.
Niggeweg, R., Michael, A.J., Martin, C., 2004. Engineering plants with increased
levels of the antioxidant chlorogenic acid. Nat. Biotechnol. 22, 746–754.
Otani, M., Shitan, N., Sakai, K., Martinoia, E., Sato, F., Yazaki, K., 2005.
Characterization of vacuolar transport of the endogenous alkaloid berberine
in Coptis japonica. Plant Physiol. 138, 1939–1946.
Petersen, M., Abdullah, Y., Benner, J., Eberle, D., Gehlen, K., Hücherig, S., Janiak, V.,
Kim, K.H., Sander, M., Weitzel, C., Wolters, S., 2009. Evolution of rosmarinic acid
biosynthesis. Phytochemistry 70, 1663–1679.
Plowman, T., Hensold, N., 2004. Names, types, and distribution of neotropical
species of Erythroxylum (Erythroxylaceae). Brittonia 56, 1–53.
References
Rhodes, M.J.C., Wooltorton, L.S.C., 1976. The enzymic conversion of hydro-
xycinnamic acids to p-coumarylquinic and chlorogenic acids in tomato fruits.
Phytochemistry 15, 947–951.
Aerts, R.J., Baumann, T.W., 1994. Distribution and utilization of chlorogenic acid in
coffea seedlings. J. Exp. Bot. 45, 497–503.

186
J.C. Pardo Torre et al. / Phytochemistry 91 (2013) 177–186
Riboulet, C., Guillaumie, S., Méchin, V., Bosio, M., Pichon, M., Goffner, D., Lapierre, C.,
Pollet, B., Lefevre, B., Martinant, J.P., Barrière, Y., 2009. Kinetics of
phenylpropanoid gene expression in maize growing internodes: relationships
with cell wall deposition. Crop Sci. 49, 211–223.
Roytrakul, S., Verpoorte, R., 2007. Role of vacuolar transporter proteins in plant
secondary metabolism: Catharanthus roseus cell culture. Phytochem. Rev. 6,
383–396.
Takahama, U., Hirotsu, M., Oniki, T., 1999. Age-dependent changes in levels of
ascorbic acid and chlorogenic acid, and activities of peroxidase and superoxide
dismutase in the apoplast of tobacco leaves: mechanism of the oxidation of
chlorogenic acid in the apoplast. Plant Cell Physiol. 40, 716–724.
Ulbrich, B., Zenk, M.H., 1979. Partial purification and properties of
hydroxycinnamoyl-CoA: quinate hydroxycinnamoyl transferase from higher
plants. Phytochemistry 18, 929–933.
Sakihama, Y., Cohen, M.F., Grace, S.C., Yamasaki, H., 2002. Plant phenolic antioxidant
and prooxidant activities: phenolics-induced oxidative damage mediated by
metals in plants. Toxicology 177, 67–80.
Sander, M., Petersen, M., 2011. Distinct substrate specificities and unusual substrate
flexibilities of two hydroxycinnamoyltransferases, rosmarinic acid synthase and
hydroxycinnamoyl-CoA:shikimate hydroxycinnamoyl-transferase, from Coleus
blumei benth. Planta 233, 1157–1171.
Shitan, N., Morita, M., Yazaki, K., 2009. Identification of a nicotine transporter in leaf
vacuoles of Nicotiana tabacum. Plant Signal. Behav. 4, 530–532.
Simpson, A.J., Brown, S.A., 2005. Purge NMR: effective and easy solvent suppression.
J. Magn. Reson. 175, 340–346.
Waldhauser, S.S.M., Baumann, T.W., 1996. Compartmentation of caffeine and
related purine alkaloids depends exclusively on the physical chemistry of their
vacuolar complex formation with chlorogenic acids. Phytochemistry 42, 985–
996.
Wang, C.-Z., Maier, U.H., Zenk, M.H., 2000. Synthesis of 3,30,5-trihydroxybiphenyl-2-
carboxylic acid, a component of the bitterest natural product amarogentin and
its Coenzyme A and n-acetyl cysteamine thiol esters. J. Nat. Prod. 63, 371–374.
Wink, M., Roberts, M.F., 1998. Compartmentation of alkaloid synthesis, transport
and storage. In: Roberts, M.F., Wink, M. (Eds.), Alkaloids. Biochemistry, Ecology
and Medicinal Applications. Plenum Publishing Corporation, New York, pp.
239–262.
Sirikantaramas, S., Yamazaki, M., Saito, K., 2008. Mechanisms of resistance to self-
produced toxic secondary metabolites in plants. Phytochem. Rev. 7, 467–477.
Sondheimer, E., Covitz, F., Marquisee, M.J., 1961. Association of naturally occurring
compounds, the chlorogenic acid-caffeine complex. Arch. Biochem. Biophys. 93,
63–71.
Sonnante, G., D’Amore, R., Blanco, E., Pierri, C.L., De Palma, M., Luo, J., Tucci, M.,
Martin, C., 2010. Novel hydroxycinnamoyl-Coenzyme A quinate transferase
genes from artichoke are involved in the synthesis of chlorogenic acid. Plant
Physiol. 153, 1224–1238.
Yazaki, K., Sugiyama, A., Morita, M., Shitan, N., 2008. Secondary transport as an
efficient membrane transport mechanism for plant secondary metabolites.
Phytochem. Rev. 7, 513–524.
Yu, X.H., Liu, C.J., 2006. Development of an analytical method for genome-wide
functional identification of plant acyl-coenzyme a-dependent acyltransferases.
Anal. Biochem. 358, 146–148.
Zeller, B.L., Saleeb, F.Z., 1997. Caffeine-caffeic acid complex formation in coffee
extracts. ASIC, 17e Colloque Scientifique International sur le Café, Nairobi,
Kenya.
Stöckigt, J., Zenk, M.H., 1974. Enzymatic synthesis of chlorogenic acid from caffeoyl
coenzyme a and quinic acid. FEBS Lett. 42, 131–134.
Zheng, X.-q., Ashihara, H., 2004. Distribution, biosynthesis and function of purine
and pyridine alkaloids in Coffea arabica seedlings. Plant Sci. 166, 807–813.