Functional characterization of two p-coumaroyl ester 3'-hydroxylase genes from coffee tree: evidence of a candidate for chlorogenic acid biosynthesis.

Abstract

Chlorogenic acid (5-CQA) is one of the major soluble phenolic compounds that is accumulated in coffee green beans. With other hydroxycinnamoyl quinic acids (HQAs), this compound is accumulated in particular in green beans of the cultivated species Coffea canephora. Recent work has indicated that the biosynthesis of 5-CQA can be catalyzed by a cytochrome P450 enzyme, CYP98A3 from Arabidopsis. Two full-length cDNA clones (CYP98A35 and CYP98A36) that encode putative p-coumaroylester 3'-hydroxylases (C3'H) were isolated from C. canephora cDNA libraries. Recombinant protein expression in yeast showed that both metabolized p-coumaroyl shikimate at similar rates, but that only one hydroxylates the chlorogenic acid precursor p-coumaroyl quinate. CYP98A35 appears to be the first C3'H capable of metabolising p-coumaroyl quinate and p-coumaroyl shikimate with the same efficiency. We studied the expression patterns of both genes on 4-month old C. canephora plants and found higher transcript levels in young and in highly vascularized organs for both genes. Gene expression and HQA content seemed to be correlated in these organs. Histolocalization and immunolocalization studies revealed similar tissue localization for caffeoyl quinic acids and p-coumaroylester 3'-hydroxylases. The results indicated that HQA biosynthesis and accumulation occurred mainly in the shoot tip and in the phloem of the vascular bundles. The lack of correlation between gene expression and HQA content observed in some organs is discussed in terms of transport and accumulation mechanisms.

Full text

Page 1

Functional characterization of two p-coumaroyl ester
3¢-hydroxylase genes from coffee tree: evidence of a candidate
for chlorogenic acid biosynthesis
Venkataramaiah Mahesh Æ Æ Rachel Million-Rousseau Æ Æ Pascaline Ullmann Æ Æ
Nathalie Chabrillange Æ Æ Jose ´ Bustamante Æ Æ Laurence Mondolot Æ Æ
Marc Morant Æ Æ Michel Noirot Æ Æ Serge Hamon Æ Æ Alexandre de Kochko Æ Æ
Danie `le Werck-Reichhart Æ Æ Claudine Campa
Received: 6 June 2006/Accepted: 24 January 2007/Published online: 27 February 2007
? Springer Science+Business Media B.V. 2007
Abstract
jor soluble phenolic compounds that is accumulated in
coffee green beans. With other hydroxycinnamoyl qui-
nic acids (HQAs), this compound is accumulated in
particularingreenbeansofthecultivatedspeciesCoffea
canephora. Recent work has indicated that the biosyn-
thesis of 5-CQA can be catalyzed by a cytochrome P450
enzyme, CYP98A3 from Arabidopsis. Two full-length
cDNA clones (CYP98A35 and CYP98A36) that encode
putative p-coumaroylester 3¢-hydroxylases (C3¢H) were
isolated from C. canephora cDNA libraries. Re-
combinant protein expression in yeast showed that both
Chlorogenic acid (5-CQA) is one of the ma-
metabolized p-coumaroyl shikimate at similar rates, but
that only one hydroxylates the chlorogenic acid pre-
cursor p-coumaroyl quinate. CYP98A35 appears to be
the first C3¢H capable of metabolising p-coumaroyl
quinate and p-coumaroyl shikimate with the same effi-
ciency.Westudiedtheexpressionpatternsofbothgenes
on 4-month old C. canephora plants and found higher
transcript levels in young and in highly vascularized or-
gans for both genes. Gene expression andHQA content
seemed to be correlated in these organs. Histolocaliza-
tion and immunolocalization studies revealed similar
tissue localization for caffeoyl quinic acids and p-cou-
maroylester 3¢-hydroxylases. The results indicated that
HQAbiosynthesisandaccumulationoccurredmainlyin
the shoot tip and in the phloem of the vascular bundles.
The lack of correlation between gene expression and
HQA content observed in some organs is discussed in
terms of transport and accumulation mechanisms.
Keywords
Coffea canephora ? Cytochrome P450 hydroxylase
Caffeoyl quinic acids ? Chlorogenic acid ?
Abbreviations
CQA
5-CQA
HQA
C3¢H
HQT
caffeoyl quinic acid
chlorogenic acid
hydroxycinnamoyl quinic acid
p-coumaroyl ester 3¢-hydroxylases
hydroxycinnamoyl-CoA: quinate
hydroxycinnamoyl transferase
hydroxycinnamoyl-CoA: shikimate/
quinate hydroxycinnamoyl transferase
consensus degenerate hybrid
oligonucleotide primers
cytochrome P450 monooxygenases
HCT
CODEHOP
CYP
V. Mahesh ? N. Chabrillange ? J. Bustamante ?
M. Noirot ? S. Hamon ? A. de Kochko ?
C. Campa (&)
Laboratoire de Ge ´nomique et Qualite ´ du cafe ´, IRD, UMR
1097 DGPC, 911 Avenue Agropolis, BP 64501, 34394
Montpellier cedex 5, France
e-mail: campa@mpl.ird.fr
R. Million-Rousseau ? P. Ullmann ? D. Werck-Reichhart
Department of Plant Stress Response, Institute of Plant
Molecular Biology, CNRS-UPR 2357, Universite ´ Louis
Pasteur, 28 rue Goethe, 67083 Strasbourg, France
L. Mondolot
Laboratoire de Botanique, Phytochimie et Mycologie,
UMR 5175 CEFE-CNRS, Faculte ´ de Pharmacie, 15 av.
Charles Flahault, BP 14491, 34093 Montpellier cedex 5,
France
V. Mahesh
Avesthagen graine, A Plant Genome Biology Laboratory,
9th floor, Discoverer, ITPL, Bangalore, India
M. Morant
Department of Plant Biology, Thorvaldsensvej 40, 1871
Fredericksberg C, Copenhagen, Denmark
123
Plant Mol Biol (2007) 64:145–159
DOI 10.1007/s11103-007-9141-3

Page 2

Introduction
Chlorogenic acid, or 5-caffeoyl quinic acid (5-CQA), is
one of the most widespread soluble phenolics in plants.
In coffee trees, it is the major hydroxycinnamoyl quinic
acid (HQA) accumulated in green beans. HQA con-
tent has been widely studied in coffee beans as these
compounds play an important role in coffee cup quality
(Clifford, 1985; Leloup et al., 1995). Extensive bio-
chemical analyses have shown that beans of some
Coffea species contain high levels of 5-CQA, but also
dicaffeoyl and feruloyl quinic acids (diCQA and
FQA). In green beans of Coffea canephora, HQAs can
represent more than 11% of the total dry matter, and
5-CQA alone represents about 68% of the total HQA
content (Anthony et al., 1993; Ky et al., 2001; Campa
et al., 2005). Studies carried out on other plant species
suggested physiological roles for HQAs, including
response to oxidative stress (Grace et al., 1998) or
resistancetophytopathogens
Matsuda et al., 2003). Recent studies underlined their
potent antioxidant activity, which is not only useful for
the plant but also of interest for human health (Zang
et al., 2003; Niggeweg et al., 2004; Jin et al., 2005).
5-CQA biosynthesis in Solanaceae (tomato, tobacco
and potato) and coffee trees was initially thought to
occur via transesterification from caffeoyl-CoA and
quinic acid (pathway 1a in Fig. 1) by the hydroxycin-
namoyl-CoA: quinate hydroxycinnamoyl transferase
HQT (Sto ¨ckigt and Zenk, 1974; Ulbrich and Zenk,
1979). The existence of another route involving direct
3¢-hydroxylation of p-coumaryol quinic acid (pathway
2) was first suggested by Ku ¨hnl et al. (1987) in carrot
cell cultures, and studies of the impact of the level of
expression of the HQT gene in tobacco and tomato
plants recently demonstrated that this route might be
predominant in Solanaceae (Niggeweg et al., 2004).
This pathway could be used by plants which are able to
accumulate 5-CQA, and may be particularly relevant
for diCQA synthesis, compounds that accumulate at
high levels in some coffee species. However, another
pathway may coexist and has to be taken into consid-
eration (part 1b of the pathway 1). It implies hydrox-
ylation of p-coumaroyl shikimate to caffeoyl shikimic
acid, which is then converted to caffeoyl-CoA, a
substrateofa hydroxycinnamoyl-CoA:
hydroxycinnamoyl transferase HCT (Hoffmann et al.,
2003). Whether synthesized directly from a p-couma-
royl ester precursor or from caffeoyl CoA, a critical
step in the synthesis of HQAs is the 3-hydroxylation of
the phenolic ring. It was recently demonstrated that
this hydroxylation step is catalyzed by the cytochrome
(Takahama, 1998;
shikimate
P450 monooxygenases belonging to the CYP98 family.
These enzymes do not use the free p-coumaric acid as
substrate as mentioned in initial metabolic schemes,
but instead use both shikimate and quinate esters of
p-coumaric acid. The enzymes are consequently re-
ferred to as p-coumaroyl ester 3¢-hydroxylases (C3¢H)
(Schoch et al., 2001; Gang et al., 2002). Arabidopsis
EMS or TDNA insertion mutants in the C3¢H
(CYP98A3) gene are characterized by reduced growth
andreduced epidermal
(Franke et al., 2002b; Abdulrazzak et al., 2006). They
accumulate p-coumarate esters and are affected in the
biosynthesis of lignin, thus providing direct evidence
that p-coumaroyl shikimate and/or p-coumaroyl qui-
nate are probably important intermediates in the lignin
pathway (Franke et al., 2002a). In sweet potato, a
fourth route has been described (pathway 3) which
involves caffeoyl glycoside as the activated intermedi-
ate (Villegas and Kojima, 1986). In coffee plants,
experiments using radiolabeled substrates indicated
that leaf 5-CQA biosynthesis may occur via p-couma-
royl quinic acid synthesis (Colonna, 1986).
The aim of this work was to clarify the biosynthetic
route of CQA and more generally of HQAs in coffee
plants. We describe the isolation and characterization
of two different CYP98 cDNAs and their correspond-
ing genes in C. canephora plants. Functional analysis
of the resulting proteins expressed in yeast showed
that although both catalyze the 3¢-hydroxylation of
p-coumaroyl shikimate, only one of these enzymes is
involved in 5-CQA biosynthesis. We compared the two
gene expression patterns to the HQA content in
different tissues from C. canephora seedlings, to the
histolocalization of the caffeoyl quinic acids and to the
immunolocalization of the CYP98 proteins. The sum of
these results is discussed in terms of possible routes for
caffeoyl quinic acid biosynthesis in C. canephora
plants.
fluorescencephenotype
Materials and methods
Plant materials and growth conditions
Fruits of Coffea canephora Pierre were collected in La
Re ´union Island. Seeds were placed on vermiculite
imbibed with sterile water and maintained in dark at
27?C. When roots developed, the seedlings were
transferred to pots and cultured in tropical green-
houses. Five 4-month old seedlings were harvested
during spring, at midday, and bulk samples were made
with shoot tips, leaves and petioles from node 1 to node
4, apical stems (from apex to node 1), medium stems
146Plant Mol Biol (2007) 64:145–159
123

Page 3

(from node 1 to node 2), basal stems (from node 2 to
node 4), cotyledon leaves, hypocotyls and roots. The
bulk samples were divided in two batches and imme-
diately frozen in liquid nitrogen. One batch was di-
rectly used for RT-PCR experiments and the other was
lyophilized before extraction for analyses of hydroxy-
cinnamoyl ester content. For histolocalization and im-
munolocalization, fresh organs were collected at the
same period on remaining seedlings and immediately
prepared for analyses.
Isolation of full length cDNA
The CODEHOP primer design was done as described
by Morant et al. (2002) after multiple alignment
(clustalW, Thompson et al., 1994) of CYP98As
peptidic sequences available in Genebank: Sorghum
bicolor(AAC39316),Glycine
Arabidopsis thaliana (NP850337), Triticum aestivum
(CAE47489, CAE47490, and CAE47491), Ocimum
basilicum (AAL99200, and AAL99201), Pinus taeda
(AAL47685), Sesamum indicum (AAL47545). The
primers were designed using the default parameters of
theCODEHOPalgorithm
codehop.html) using the codon usage of coffee trees.
Four CYP98A primers were obtained, two forward 5¢-
CTACTGCTATTACTGTTGAATGGGCNRTNGC-
3¢ (98f2) and 5¢-CCAAGAGTTCAACAAAAGGTT
CAAGARGARHTNGA-3¢ (98f3) and two reverse,
max(AAB94587),
(http://blocks.fhcrc.org/
5¢-TGTGGAAGCATAAGTGGAGTNGGNGGRTG-
3¢ (98r1) and 5¢-GGATCTCTAGCAACAGCCCANA
CRTTNAC-3¢ (98r2) positioned outside the structural
motifs. The expected amplified fragments were about
200 nucleotides long.
PCR was performed using the High Fidelity PCR
Master kit (Roche, Switzerland) on fruit or leaf C.
canephora cDNA libraries (Mahesh et al., 2006). The
PCR program was established according to the
CODEHOP web page tips, including successively a
touch down and a classical PCR: 3 min initial dena-
turation at 94?C, then 20 touch down cycles (1 min
94?C, 2 min 70?C (–1?C/cycle), 2 min 72?C), followed
by 29 cycles of 1 min at 94?C, 1 min of individual pri-
mer pairs annealing temperatures and 1 min at 72?C,
and finally a 10 min extension at 72?C. When no
amplified fragment was obtained with a primer pair,
the touch down starting annealing temperature was
decreased by 5?C increments until successful amplifi-
cation.
PCR products were checked on 1% (w/v) agarose
gels, and cloned into TOPO-TA vector (Invitrogen,
The Nederlands). Plasmids from 25 colonies from each
library were sent for sequencing (MWG-Biotech,
France). To assign a function to the sequences, a Blast
search was conducted on GenBank accessions (http://
www.ncbi.nlm.nih.gov/BLAST/).
To obtain the complete coding sequence of each
gene, oneforward primerforeach sequence
Fig. 1 Proposed pathways for the biosynthesis of chlorogenic
acid (5-caffeoyl quinic acid) in plants. The three different routes
in caffeoyl quinic acid pathway are labeled 1(a and b), 2 and 3.
PAL, phenylalanine ammonia-lyase; HQT, hydroxycinnamoyl
CoA quinate hydroxycinnamoyltransferase; HCT, hydroxycin-
namoyl CoA shikimate/quinate hydroxycinnamoyltransferase;
C3H (and C3¢H), p-coumarate 3¢-hydroxylases; UGCT, UDP
glucose:cinnamate glucosyl transferase; HCGQT, hydroxycinna-
moyl D-glucose: quinate hydroxycinnamoyl transferase
Plant Mol Biol (2007) 64:145–159 147
123

Page 4

5¢-CGAGTCATTGGCTCCGATAGAA-3¢ (fc1) and
5¢-GGTACGAGCGTGTTATGATCGAGAC-3¢ (fc2),
and one reverse also for each sequence 5¢-TTCTA
TCGGAGCCAATGACTCG-3¢ (rc1) and 5¢-AGTCC
GTCTCGATCATAACACGCTC-3¢ (rc2) were de-
signed. The amplified fragments obtained with these
primers and the T3 and T7 universal ones, from the
cDNA libraries, overlapped about 200 bp the already
known CcCYP98A partial sequences. The fragments
were cloned into TOPO-TA, sequenced, and the
resulting sequences were assembled using the SeqMan
software of the Lasergene package (DNASTAR Inc.,
USA).
Gene structure
Coffea canephora genomic DNA (~100 ng) extracted
from leaves using the DNAeasy plant mini kit (Qiagen,
Germany) was used as template to amplify the corre-
sponding genomicsequences
CYP98A36. The forward primers included the trans-
lational initiation codon (5¢-AACCAATGGCTCTGC
TTCTGATCC-3¢ for CYP98A35 and 5¢-GATCATG
GCACTTTTTCTACTG-3¢ for CYP98A36). The re-
verse primers were designed downstrean the previously
described position of a CYP98A3 intron (Accession
No NP850337) (5¢-TTCTATCGGAGCCAATGACT
CG-3¢ for CYP98A35 and 5¢-AGTCCGTCTCGATCA
TAACACGCTC-3¢ for CYP98A36). A long PCR
amplification was carried out using the Arrow Taq
DNA polymerase (Qbiogene, USA) with the following
conditions: 94?C for 5 min, followed by 24 amplifica-
tion cycles (30 s at 94?C, 50 s at 56?C and 5 min at
72?C), and a final extension reaction at 72?C for
10 min. The resulting fragments were cloned in TOPO-
TA and double strand sequenced.
ofCYP98A35 and
Southern-blot hybridization
A 349 bp probe, recognizing the first exon of both
C. canephora CYP98A, was prepared by PCR ampli-
fication using total DNA as template and as primers 5¢-
GGTGAGGTTCAAGTGCTATGC-3¢ and 5¢-TCAA
GGACTGCACCATACCTG-3¢. Ten micrograms of
C. canephora total DNA were digested with EcoRV,
DraI, or NsiI. Fragments were separated on a 0.8% (w/
v) agarose gel at 80 V for 4 h. DNA was transferred
onto a nylon membrane (GE Healthcare Bio-Sci-
ences) and hybridized with the32P radiolabeled probe
using a standard protocol (Sambrook and Russell,
2001).
Phylogeny
The full length amino acid sequences from the P450
family CYP98A available at £http://drnelson.utmem.
edu/CytochromeP450.html‡ and GeneBank’s were
aligned using ClustalW (Thompson et al., 1994). The
following accessions were used: AAC39316 (Sorghum
bicolor), AAB94587 (Glycine max), NP850337 (Ara-
bidopsisthaliana), AAU44038
BAC44836 (Lithospermum erythrorhizon), AAG52369
and AAM67314 (A. thaliana), CAE47489, CAE47490
and CAE47491 (Triticum aestivum), AAL99200 (Oci-
mum basilicum), CAD20576 (Solenostemon scutella-
rioides), AAL47685(Pinus
(Sesamumindicum), AAT06912
(Populus trichocarpa), AAS57921 (Camptotheca ac-
uminata). A rooted phylogenetic tree was constructed
with Mega 3.1 (Kumar et al., 2004) from these align-
ments using CYP51G2 from A. thaliana (AC002329) as
outgroup.
(Oryzasativa),
taeda), AAL47545
majus),(Ammi
Protein overexpression in Yeast (Saccharomyces
cerevisiae)
The C. canephora fruit cDNA library was used as
template for PCR amplification using 5¢-GCGGAT
CCATGGCTCTGCTTCTGA-3¢
CCTTACAGCTCTACTGGCAC-3¢ as forward and
reverse primers for CYP98A35 and 5¢-GCGGATCC
ATGGCACTTTTTCTAC-3¢ and 5¢-CGGGTACCT-
TACATATCCACAGCCAC-3¢
introduce BamHI and KpnI restriction sites, respec-
tively at the 5¢ and 3¢ ends of the two coffee
CYP98A cDNA. The PCR was carried out using the
High Fidelity PCR Master Kit (Roche) under the
following conditions: 5 min of initial heating at 94?C
and 24 amplification cycles (30 s at 94?C, 60 s at
60?C, 90 s at 72?C). The reaction was completed by a
10-min extension at 72?C. The amplicons were cloned
into TOPO-TA and sequenced. After BamHI and
KpnI digestion, the fragments were directionally
subcloned into the expression cassette of the plasmid
pYeDP60, and the cloning sites were confirmed by
sequencing.TheS. cerevisiae
derivative of the W303-B strain that expresses the
Arabidopsis cytochrome P450 reductase ATR1 upon
galactose induction, was then transformed and grown
as described by Pompon et al. (1996). Yeast micro-
somes were isolated after 24 h of induction on 20 g/l
galactose at 30?C.
and5¢-CGGGTA
for CYP98A36to
strainWAT11,a
148Plant Mol Biol (2007) 64:145–159
123

Page 5

Assay conditions for 3¢-hydroxylase activities
Total P450 content was evaluated in each microsome
preparationby spectrophotometric
(Omura and Sato, 1964). Assays for p-coumaroyl-
shikimate/quinate 3¢-hydroxylase (C3¢H) were per-
formed in a total volume of 100 ll, as already de-
scribed by Schoch et al. (2001). The final concentration
was 0.2 lM for P450 and 30 to 50 lM for the substrate
(coumaric acid, coumaroylshikimate or coumaroylqui-
nate). The reaction was incubated under shaking in the
dark at 27?C for 30 min and then stopped by addition
of 5 ll of 4 M HCl. After extraction with ethyl acetate,
the organic phase was evaporated under argon and the
reaction products dissolved in 150 ll of 10% acetoni-
trile, 90% water, and 0.2% acetic acid (v/v/v) for re-
verse-phaseHPLC analysis
100RP-18 column, 4 x 15 mm, 5 lm; flow rate of 1 ml/
min; 5 min of isocratic 10% acetonitrile, and then a
linear gradient from 10% to 48% acetonitrile in water
containing 0.2% acetic acid). Absorbance was moni-
tored at 320 nm with a diode array detector. Substrate
amounts and products were calculated according to
Ku ¨hnl et al. (1987). Retention times and UV spectra
confirmed the nature of the products.
measurement
(Merck LiChrospher
Semi-quantitative RT-PCR
The primers 5¢-CCATCTTCAAGGACTGCACCA-
TACC-3¢
and5¢-TTCTATCGGAGCCAATGACT
CG-3¢ for CYP98A35 and 5¢-CAGGGCAGAGTCT
ACTAGTGAAG-3¢
and
CATAACACGCTC-3¢ for CYP98A36 were designed
so as to frame the intron in both sequences. Total RNA
was extracted from about 100 mg of plant tissues using
the RNeasy Plant Mini Kit (Qiagen). RNA was
quantified by spectrophotometry and the cDNA was
synthesized from 2 lg of total RNA using the kit
Superscript II cDNA synthesis (Invitrogen). The PCR
reactions were conducted in a final volume of 50 ll
under the following conditions: 94?C for 5 min, fol-
lowed by 23, 26, 29, 32 or 35 amplification cycles (30 s
at 94?C, 50 s at 58?C and 60 s at 72?C) and a final
extension at 72?C for 10 min.
5¢-AGTCCGTCTCGAT
Tissue print hybridization
Transversal hand cuts of petioles from nodes 1 and 2 of
young shoot tip leaves, and stems, were pressed firmly
for 2 s three times onto a nitrocellulose membrane
(0.2 lm) and blotted dry. The membranes were incu-
bated 1 h in PBS containing Tween 20 (1% v/v) and
then in a blocking buffer (buffer 1 with 5% (w/v)
blocking agent (Biorad, USA) for 2 h). After rinsing
(buffer 1), they were incubated with anti-CYP98A3
polyclonal serum diluted 1: 1000 in buffer 2 (buffer 1
with 1% blocking agent) for 2 h. After 3 washes with
buffer 1, the membranes were incubated with 1: 7,500
diluted secondary antibody (goat) anti-rabbit IgG
conjugated with alkaline phosphatase (Promega, USA)
in buffer 2. After rinsing (buffer 1), protein-antibody
complexes were detected with 5-bromo-4-chloro-3-
indolyl phosphate and nitro blue tetrazolium as sub-
strates (Sigma, USA). Additional Western-blot analy-
ses were performed with CYP98A3 antibodies to
confirm the recognition of CYP98A in a protein extract
from C. canephora leaves and of the recombinant
proteins.
Histochemical detection of caffeoylquinic acids
and lignin
Coffea canephora seedlings freshly collected samples
were embedded in 3% (w/v) agarose (type II EEO,
Panreac). Cross-sections (40 lm, Leica VT 1000S
microtome), were immersed (30 s) in Neu’s reagent
(1% (w/v) 2-amino-ethyldiphenylborinate (Fluka) in
absolute methanol) and mounted in glycerine: water
(10:90, v/v) solution (Neu, 1957). Microscope obser-
vations (Nikon Optiphot) under UV light (filter
UV-1A: 365 nm excitation filter, 400 nm barrier filter)
allowed to identify caffeoylquinic acids by a specific
greenish–white fluorescence (Mondolot-Cosson et al.,
1997) while feruloyl derivatives were bright blue. The
sections were also stained with Mirande’s reagent for
lignin and cellulose detection (green and red-pink
coloration respectively) (Mondolot et al., 2001). Pho-
tographs were taken with a digital Nikon coolpix 4500
camera.
HPLC analysis of hydroxycinnamoyl ester content
Hydroxycinnamoyl ester extraction was carried out
following the method described by Ky et al. (2001).
The different compounds were identified on the basis
of their retention time and UV spectra using the HPLC
analytical procedure described by Bertrand et al.
(2003). Retention times were those described by Ky
et al. (1997) for 3-caffeoyl quinic acid (3-CQA,
6.1 min), 3-feruloyl quinic acid (3-FQA, 9.2 min), 4-
and 5-CQA (10.9 min), 4-FQA (14.3 min), 5-FQA
(15.2 min), 3,4-dicaffeoyl quinic acid (3,4-DiCQA,
18.8 min), 3,5-DiCQA (19.5 min) and 4,5-DiCQA
(22.1 min). Coumaroyl quinic, coumaroyl shikimic and
caffeoyl shikimic acids were used as control. Quantifi-
cation was performed by comparison to a 5-CQA
Plant Mol Biol (2007) 64:145–159 149
123

Page 6

standard (Sigma-Aldrich, USA). The hydroxycinna-
moyl quinic acid content of each analyzed sample was
expressed as mean percentage of dry weight (% DW)
from three different extractions. Data were analyzed
using Statistica 5.1. Between-organ differences were
tested using a one-way ANOVA (Newman-Keuls).
Results
Isolation of full length C. canephora CYP98A
cDNAs
To specifically isolate the CYP98A homologs from
the large CYP gene family, PCR primers were con-
structed according to the CODEHOP primer design
strategy avoiding the strong consensus regions com-
mon to other P450 families (Fig. 2). Their selectivity
was increased conducting a touch-down PCR starting
at a high annealing temperature (70?C). A single
band of the expected size was obtained with each
primer pair 98f2-98r1 and 98f3-98r2. After sequenc-
ing and blast analysis, 17 of the fifty clones analyzed
displayed significant homology with genes from the
CYP98A family. Even though all of these partial
sequences had 93% similarity with the partial de-
duced protein sequence of Ocimum basilicum p-
coumaryl3¢-shikimate
(AAL99200), it was possible to divide them into two
groupsaccording to their
Among each of the two groups, the sequences were
almost identical with the exception of very few nu-
cleotides, which might be due to Taq misamplifica-
tion, and the between group homology didn’t exceed
70.3%.Thefull-length
sponding to each gene were obtained by PCR
amplificationusingspecific
(Accession No DQ269126) and CYP98A36 (Acces-
sion No DQ269127) differed by their 5¢- and 3¢-UTR
(75 and 28 bp for the 5¢-end, and 156 and 194 for the
3¢-end respectively), but both open reading frames
had the same length and encoded a protein of 508
amino acids (Fig. 2). Comparison of the coding se-
quences showed that they share 72.4% identity at the
nucleotide level, and 88% similarity at the amino
acid level. When the deduced sequences were com-
pared to 13 CYP98As from other plant species, both
showedthebestsimilarity
trichocarpa sequences (CYP98A27 and CYP98A23):
89% similarity was observed between the CYP98A35
and CYP98A27 and 98% between the CYP98A36
and CYP98A23. For both genes, the lowest score was
obtained with CYP98A1 from S. bicolor (82%).
hydroxylase isoform1
sequencehomology.
cDNAsequencescorre-
primers.CYP98A35
scorewith Populus
Alignment with CYP98A3 from A. thaliana (Fig. 2)
showed that both deduced protein sequences contain
the classically conserved domains of the P450 pro-
teins such as the ERR triad, cysteine in the heme-
binding domain and the six probable substrate rec-
ognition sites specific to CYP98A (SRS). However,
there were significant differences in the SRS1, SRS2,
and SRS6 domains, which may be indicative of dif-
ferences in substrate specificity.
Southern blotting, using a C. canephora CYP98As
specific probe, showed that each gene was present as
a single copy in the C. canephora genome (Fig. 3).
Digestion by EcoRV, whose site is supposed absent
from both sequences, gave only one signal of high
molecular weight. DraI cuts outside CYP98A36 gene
(upper band) and inside CYP98A35 (lower band) and
NsiI, which cut differently inside both genes, showed
a three-band pattern (the third band may be ex-
plained by a different allelic form of one of the two
genes).
A phylogenetic tree constructed using 16 other
CYP98A sequences and the CYP51G2 sequence as
outgroup showed that CYP98A35 and CYP98A36
were grouped with other dicot CYP98A sequences
(Fig. 4). Nevertheless, CYP98A36 seemed to be
more closely related to CYP98A3 from A. thaliana
than CYP98A35. Genes were clustered in two dis-
tinct groups indicating that the encoding genes re-
sulted from an early duplication of the ancestral
gene.
Gene structure
Sequence analysis of C. canephora genomic DNA
amplified with specific primers designed to match the
respective5¢
and3¢-end
CYP98A36 genes revealed the presence of introns in
both gene sequences (Fig. 5). The two amplified
fragments differed by their length (4 kb and 1.5 kb
for CYP98A35 and CYP98A36, respectively). Like
most A-type P450s, both genes shared a phase 0
intron at position 882 on the nucleotide sequence,
corresponding to the intron labeled M by Paquette
et al. (2000). However, the two introns differed in
length (106 nucleotides for CYP98A35 and 490 for
CYP98A36) and in sequences.
CYP98A35 genomic sequence showed an additional
116 bp phase 1 intron at position 484. In this respect
CYP98A35 differs from the prototype A-clade plant
P450s and CYP98As (e.g. CYP98A3 from A. thaliana
or CYP98A36). These features indicate that the
CYP98A35 additional intron was most likely gained
after the gene duplication.
of CYP98A35and
Moreover, the
150 Plant Mol Biol (2007) 64:145–159
123

Page 7

Enzymatic properties of recombinant CcCYP98A
proteins
For functional analysis, the two CcCYP98As were
co-expressed in yeast with a plant cytochrome P450
reductase. The microsomal fractions, which were ex-
tracted from the recombinant yeasts expressing each
of the CcCYP98As, were tested for their enzymatic
activity and substrate specificity with p-coumaric acid,
p-coumaroyl shikimic or p-coumaroyl quinic acids as
substrates. A negative control consisted in a micro-
somal fraction from yeast transformed with a void
plasmid; microsomes from yeast expressing CYP98A3
from A. thaliana were used as positive control. When
assays were performed in presence of p-coumaric
acid, no conversion into more oxygenated compound
wasdetectedwithrecombinant
CYP98A36 and CYP98A3 incubated in the presence
CYP98A35,
of NADPH (data not shown). This indicates that, as
already observed for A. thaliana CYP98A3, p-coum-
aric acid is not a relevant substrate for CYP98A35
and CYP98A36. However, in the presence of p-cou-
maroylshikimate, both recombinant proteins were
able to form caffeoylshikimate (Fig. 6). Despite very
similar affinities, CYP98A35 appeared slightly more
efficient than CYP98A36 with this substrate, with a
Kcat/Kmof 0.25 instead of 0.15 min–1lM– 1(Table 1).
Whenincubatedwithp-coumaroylquinate,
CYP98A35 was able to form caffeoylquinate. The
enzyme affinity for p-coumaroylquinate was two
fold lower than for p-coumaroylshikimate, but its
efficiency was identical with shikimate and quinate
esters. The conversion of p-coumaroylquinate by
CYP98A36 was too slow to allow reliable determi-
nation of kinetic constants. It is worth mentioning
that the catalytic turnover of both the CcCYP98As
only
Fig. 2 Comparison of the
amino acid sequences
encoded by the two
C. canephora CYP98As
aligned with CYP98A3 from
A. thaliana. Sequence
alignments were performed
with the Clustal W method.
Black boxes enclose amino
acids that are identical in the
two Coffea proteins and
common with CYP98A3. The
P450 conserved domains are
underlined: (I): Proline rich
membrane hinge (PPGP),
(II): I-helix involved in
oxygen binding and activation
(A/G-G-X-E/D-T-T/S), (III):
ERR triade (E-X-X-R-R),
(IV): Clade signature (PERF)
and (V): Heme binding region
(F-X-X-G-X-F-X-C-X-G).
The putative substrate
recognition sites are indicated
by the double arrow (as SRS1
to SRS6). The primer position
is indicated by arrows
Plant Mol Biol (2007) 64:145–159151
123

Page 8

was about two orders of magnitude lower than that
of CYP98A3 of A. thaliana. Thus it cannot be
entirely excluded that the shikimate and quinate
esters of p-coumaric acid are not their preferred
substrates.
Gene-specific transcript analysis of the two
CcCYP98As
The different gene structures and protein activities of
CYP98A35 and CYP98A36 suggest that they have
evolved to fulfil specific functions in the phenylpropa-
noid grid pathway and genes would be expected to be
differentially expressed in the plant. In order to further
clarify their respective functions, accumulation of their
transcripts was analyzed by semi-quantitative RT-PCR
in different organs of four month-old C. canephora
plants. Except in old leaves (i.e. leaves from nodes 3
and 4 and in cotyledon leaves), CYP98A35 and
CYP98A36 appeared to be expressed in all the plant
organs (Fig. 7). Transcripts from both genes were
particularly abundant in shoot tip and petioles. What-
ever the insertion level of the leaf on the stem, petioles
showed a higher level of transcripts compared to the
corresponding leaves or stems. A slight decrease was
observed in older tissues.
Hydroxycinnamoyl ester content, histolocalization
of caffeoyl quinic acids and tissue print
hybridization
As already reported for green coffee beans (Anthony
et al., 1993), HPLC analyses showed that caffeoyl- and
feruloyl quinic acids were the major hydroxycinnamoyl
esters accumulated in all organs of the 4-month-old
plants (Table 2). Highest HQA concentrations were
measured in young tissues (such as shoot tips and
young stems), and in cotyledon leaves. Among these
compounds, a monocaffeoylquinate, the 5-CQA (or
chlorogenic acid), and a dicaffeoylquinate, 3,5-diCQA,
were particularlyabundant
ffeoylquinate was the most abundant compound in the
shoot tip and young stem, representing more than 50%
of the total hydrocycinnamoyl esters present in these
organs (Table 2). The cotyledon predominantly accu-
mulated monocaffeoylquinate, presumably chlorogenic
acid.
Histolocalization of these caffeoyl quinic acids was
performed by staining cross sections from the dif-
ferent organs of the plant with the Neu reagent
(Mondolot-Cosson et al., 1997). Under UV light
(365 nm excitation filter, 400 nm barrier filter), caf-
feoyl quinic acids (mono and dicaffeoyl quinic acids)
were identified by a specific greenish–white fluores-
cence. Figure 10 shows that in young organs from the
shoot tip, such as developing leaf (Fig. 10A1), the
fluorescence was widely distributed in all the tissues.
In petioles (Fig. 10B1) or stems (Fig. 10C2) from
node 2, the fluorescence was more associated with
the vascular bundles (Fig. 10C1), intensively labelling
the cell layers surrounding the vascular bundles,
phloem cells (primary or secondary phloem from
the stems), and some cells of the medullar paren-
chyma.
Transversal sections of the same organs were
printed onto nitrocellulose membranes and incu-
batedwith polyclonal antibodies
recombinant CYP98A3 (Schoch et al., 2001). As
shown in Fig. 9 (A and B) these antibodies give a
positive, apparently specific but relatively low inten-
sity signal with both recombinant coffee CYP98As,
and specifically recognize proteins of the same
molecular mass in a crude protein extract from
C. canephora leaves. The same antibodies were pre-
viously shown to be very specific of the C3¢H type of
CYP98s and not to recognize the closely related
CYP73 or CYP98A8 and CYP98A9 (not metaboliz-
ing p-coumaryol esters of quinate and shikimate)
from A. thaliana (Schoch et al., 2001). In the tissue
prints, high protein accumulation was detected in all
(Fig. 8).The dica-
raised against
Fig. 3 Southern blot for estimation of Coffea canephora
CYP98A gene (CYP98A35 and CYP98A36) copy number. Ten
micrograms of total DNA digested by the corresponding EcoRV,
DraI or NcoI and hybridized with the corresponding32P-labeled
cDNA probe
152 Plant Mol Biol (2007) 64:145–159
123

Page 9

parts of very young organs, such as shoot tips
(Fig. 10A2). In petioles and stems from node 2, the
proteins were particularly visible around the vascular
bundles, especially in non-lignified tissues (Fig. 10B2
and C3).
Discussion
Two coding sequences, CYP98A35 and CYP98A36,
were isolated from C. canephora fruit and leaf EST
libraries, using a degenerate PCR-based strategy tar-
geted to the CYP98 family of cytochrome P450
monooxygenases. As previously reported for wheat
(Morant et al., 2002), this strategy proved to be
effective for isolating cDNAs that encoded putative
p-coumaroylester3¢-hydroxylases,
showing a high degree of similarity to other CYP98A
sequences such as those from Camptotheca acuminata
or Ocimum basilicum. These genes seemed to be
present as a single copy in C. canephora genome, and
the encoded proteins had the same length (508 aa) but
displayed only88%similarity.
CYP98A36 share a common intron (Fig. 5), though
different in length, at the expected position for P450
bothsequences
CYP98A35and
Fig. 4 Phylogenetic tree of the CYP98A family. The phylogenetic
representation of the CYP98 family built by the neighbour-joining
method and bootstrapped 1,000 times in Mega3.1. CYP51G2
involved in sterol biosynthesis was used to root the tree. Branch
values under 50% were omitted from the figure. • CYP98s from
Coffea h CYP98 from Gymnosperms, D CYP98s from Monocots.
The sequences used for alignment were those from Sorghum
bicolor (CYP98A1,AAC39316),
AAB94587), Arabidopsis thaliana (CYP98A3, NP850337), Oryza
sativa (CYP98A4, AAU44038), Lithospermum erythrorhizon
(CYP98A6, BAC44836), A. thaliana (CYP98A8, AAG52369),
Glycine max (CYP98A2,
A.
(CYP98A10 CAE47489), T. aestivum (CYP98A11, CAE47490),
T. aestivum (CYP98A12, CAE47491), Ocimum basilicum (CY-
P98A13v1, AAL99200), Solenostemon scutellarioides (CYP98A14,
CAD20576), Pinus taeda (CYP98A19, AAL47685), Sesamum
indicum (CYP98A20, AAL47545), Ammi majus (CYP98A21,
AAT06912), Camptotheca acuminata (CYP98A28, AAS57921),
C. canephora (CYP98A35, DQ269126), C. canephora (CYP98A36,
DQ269127). The sterol 14-a-demethylase from A. thaliana
(CYP51G2) was used as outgroup
thaliana(CYP98A9AAM67314), Triticum aestivum
Fig. 5 Genomic
CYP98A35 (A) and CYP98A36 (B) genes
structure of thetwo Coffeacanephora
Plant Mol Biol (2007) 64:145–159 153
123

Page 10

genes of the A-clade (Paquette et al., 2000), but
CYP98A35 acquired an additional intron located at
nucleotide 484 from the ATG (between amino acids
Pro and Glu). This significant difference in gene
structure and the only moderate similarity between the
two paralogs indicate early duplication of the ancestral
gene and, very likely, acquisition of specific functions
and regulations (Jeong et al., 2006).
This hypothesis was supported by the functional
expression of the two CYP98A35 and CYP98A36
proteins in yeast. While both recombinant proteins
metabolized p-coumaroyl shikimate at similar rates,
only CYP98A35 was able to hydroxylate p-coumaroyl
quinate to form chlorogenic acid. All the members of
the CYP98 family described so far (Schoch et al., 2001;
Gang et al., 2002; Morant et al., 2007), metabolized
shikimate esters of p-coumaric acid more efficiently
than quinate esters. CYP98A35 is the first to catalyze
with equal efficiency the hydroxylation of quinate and
shikimate esters of p-coumaric acid. It is thus very
likely that this enzyme significantly contributes to the
biosynthesis of chlorogenic acid and other CQAs in the
coffee plant via the direct route involving hydroxyl-
ation of p-coumaroylquinate (Fig. 1, pathway 2).
Nevertheless, the low catalytic turnover of both pro-
teins (compared to CYP98A3 but not to other plant
P450 enzymes) may also indicate that the shikimate
and quinate esters of p-coumaric acid are not their
Fig. 6 HPLC analysis of the
products of p-coumaroyl-
shikimate (A) and
p-coumaroyl-quinate
(B) metabolism by
recombinant CYP98A35 and
CYP98A36. Absorbance was
monitored at 320 nm.
Conversion is shown after
30 min incubation. Controls
(1A and 1B) were performed
using microsomes of yeast
transformed with a void
plasmid. Peak X is the
substrate, peak Y is the
product as indicated by
retention times and UV
spectra by comparison with
authentic samples. As
substrates were obtained by a
two steps enzymatic reaction,
traces of the isomers 3- and
4-coumaroyl-shikimates (X¢)
and of the precursor,
p-coumaroylCoA (X¢¢) were
also detected. Assays with
CYP98A35 (2A and 2B) and
CYP98A36 (3A and 3B) were
carried out using 0.2 lM of
each recombinant P450
Table 1 Substrate specificity of the recombinant CYP98As from C. canephora in comparison with the recombinant CYP98A3 from
A. thaliana
Yeast expressed enzyme SubstrateKm(lM)Kcat(min–1)Kcat/Km(min–1M–1)
CYP98A35p-Coumaroylshikimate
p-Coumaroylquinate
p-Coumaroylshikimate
p-Coumaroylquinate
p-Coumaroylshikimate
p-Coumaroylquinate
4.1 ± 0.4
10.3 ± 2.6
3.5 ± 0.7
nd
7.0 ± 1.0
18.0 ± 2.0
1.0 ± 0.0
2.3 ± 0.1
0.5 ± 0.03
nd
612 ± 30
399 ± 22
0.25
0.23
0.15
nd
87
22
CYP98A36
CYP98A3
Data for CYP98A3 are those determined by Schoch et al. (2001)
nd: not determined. The conversion of p-coumaroylquinate by CYP98A36 was too slow to allow determination of catalytic parameters
154Plant Mol Biol (2007) 64:145–159
123

Page 11

preferred substrates. This does not exclude a contri-
bution of the shikimate ester route (pathway 1b, either
via CYP98A35 or CYP98A36), in particular for the
synthesis of the diCQAs.
The occurrence of two different C3¢H in coffee tree,
one able to efficiently hydroxylate both p-coumaroyl
quinate and p-coumaroyl shikimate, and the other
capable of hydroxylating only the shikimate ester,
raises the interesting possibility that CYP98A35 and
CYP98A36 may have acquired very specific functions.
The former could be dedicated to the biosynthesis of
CQAs, the latter to the formation of the lignin pre-
cursors. In vitro, CYP98A35 still metabolises the
shikimate ester, but in vivo its substrate specificity
might be restricted to the metabolism of quinate esters
via interactions with- or tissue-specific expression of- a
specific hydroxycinnamoyl CoA: quinate hydroxinna-
moyl transferase such as those recently described in
Solanaceae (Niggeweg et al., 2004). In Solanaceae, two
hydroxycinnamoyl transferases with a strong substrate
preference for shikimate and quinate, respectively,
have been described (Hoffmann et al., 2003; Niggeweg
et al., 2004). A low hydroxycinnamoyl CoA: shikimate
hydroxycinnamoyl transferase activity, which leads to
the biosynthesis of p-coumaroyl shikimic acid, was
previously identified in cell suspensions of Coffea
arabica (Ulbrich and Zenk, 1980). In the same cultures
hydroxycinnamoyl CoA: quinate hydroxycinnamoyl
transferase activity was also detected (Ulbrich and
Zenk, 1979).
In an attempt to further clarify the respective roles
of CYP98A35 and CYP98A36, we performed semi-
quantitative RT-PCR experiments to determine if they
differ in their tissue-specific expressions. These exper-
iments did not provide a clear-cut answer, since both
genes showed very similar expression patterns in all
organs, including roots, of the 4-month-old C. cane-
phora seedlings (Fig. 7). For both genes, the transcript
levels were low in the leaves, particularly in older
leaves, and almost undetectable in cotyledons. Sur-
prisingly, at each node, expression appeared higher in
the petiole than in the corresponding leaf. This
observation suggests higher CcCYP98As expression in
the plant organs where vascular tissues are propor-
tionally more abundant, such as roots, stems and pet-
ioles, similarly to what was observed in Arabidopsis
inflorescence stems and roots for CYP98A3 (Schoch
et al., 2001; Franke et al., 2002a).
We also attempted to correlate CYP98A35 and
CYP98A36expressionwiththeaccumulationof5-CQA
and other hydroxycinnamoyl esters. As was usually
observed, evaluation of the hydroxycinnamoyl deriva-
tivescontentinyoungseedlingsofC.canephorashowed
that neither coumaroyl nor caffeoyl shikimate, nor
coumaroyl quinate are accumulated in the plant tissues
(Table 2).Thisresulthasbeenconfirmedongreenbeans
from several Coffea species (Anthony et al., 1993;
Clifford, 2000; Campa et al., 2005). These compounds,
especially shikimate esters, could be considered as
transient intermediates which are rapidly converted to
downstream compounds. It is noteworthy that even in a
A.thalianaT-DNAinsertionalmutantfortheCYP98A3
gene, no accumulation of p-coumaroyl shikimate has
been detected (Abdulrazzak et al., 2006). However,
downstream quinate derivatives, particularly mono- or
dicaffeoylquinic acids, accumulate mainly in young or-
gans. These compounds, successively described as
growth regulators, disease resistance factors, antioxi-
dants, or as affecting the organoleptic quality of fruits
(Molgaard and Ravn, 1988; Maher et al., 1994; Macheix
and Fleuriet,1998) may constitutean antioxidative pool
for the cells.
Fig. 7 (A) Expression levels of CYP98A35 and CYP98A36 in
different organs of 4-month-old Coffea canephora plants. mRNA
expression levels were determined by semi-quantitative RT-PCR
after 23, 26 or 29 cycles of amplification. St: shoot tip ; L1-3:
leaves from the node 1 to 3; S1-3: stems from the node 1 to 3; P1-
3: petioles from the node 1 to 3; Hy: hypocotyls ; R: roots. (B) b-
actin was used as reference for 23 cycles of amplification
Plant Mol Biol (2007) 64:145–159155
123

Page 12

In some organs, HQAs accumulation did not cor-
relate to expression of CcCYP98s. For example in
cotyledons, the ester content was very high, particu-
larly of 5-CQA, whereas the level of expression of both
genes was very low (Table 2). Another example is
provided by leaves: despite a decrease in gene
expression in aging leaves, the HQA level was identical
to that detected in younger leaves. However, in very
young organs, such as shoot tips or young stems, the
high level of gene expression appeared to match the
HQA content. The high level of 3,5-dicaffeoylquinic
acid, the main compound accumulated, characterized
these organs, as well as the mesophyll localization of
the p-coumaroyl 3¢-hydroxylases, and of the caffeoyl
quinic acids. Interestingly, in older organs both HQAs
Table 2 Evaluation of the content in the major hydroxycinnamoyl esters present in the different organs of 4-month-old seedlings of C.
canephora
OrganCQA DiCQA5-FQA HQA
3– 4/5– 3,4–3,5– 4,5–
Shoot tip
Stem
0.02e
0.43ab
0.39bc
0.20d
0.04e
4.90c
5.78b
4.80c
1.59f
1.04f
0.13d
0.22c
0.30b
0.15d
0.09de
7.73b
9.23a
2.01c
0.50de
0.35de
0.19bc
0.24b
0.20bc
0.15bc
0.06c
0.10cd
0.10cd
0.19b
0.05de
0.04de
13.11b
16.02a
7.89c
2.62ef
1.61f
Apical
Medium
Basal
Hypocotyl
Petiole
Node 1
Node 2
Node 3
Young
Medium
Old
0.48ab
0.55a
0.37bc
4.16c
3.27d
2.24ef
0.15cd
0.22cd
0.18cd
1.52cd
1.09cde
0.51d
0.09bc
0.15bc
0.16bc
0.18b
0.13bc
0.08cde
6.59cd
5.42d
3.54def
Leaf
Node 1
Node 2
Node 3
Cotyledon
Root
Young
Medium
Old
0.12de
0.14de
0.27cd
0.55a
0.02e
2.59de
2.53de
2.92de
7.63a
1.09f
0.12d
0.12d
0.17cd
0.70a
0.03e
0.97de
0.55de
0.40de
1.51cd
0.16d
0.07bc
0.07bc
0.14bc
1.08a
0.04c
0.08cde
0.08cde
0.09cd
0.47a
0.02e
3.94de
3.49def
3.99de
11.94b
1.36f
Values are expressed in percentage of the dry mass (% DW). For the same class of compounds, values followed by the same letter
indicate no significant between-organ difference at P £ 0.05 according to one-way ANOVA. As shown in Fig. 8, 4-CQA and 5-CQA
cannot be separated in our analytic system
Fig. 8 HPLC analysis of hydrocynammoyl esters accumulated in
node 1- leaves of Coffea canephora 4-month-old plants. 3-CQA:
3-caffeoylquinic acid; 4/5-CQA: 4- and 5-caffeoylquinic acid; 5-
FQA: 5-feruloylquinic acid; 3,4-diCQA: 3,4-dicaffeoylquinic
acid; 3,5-diCQA: 3,5-dicaffeoylquinic acid; 4,5-diCQA: 4,5-
dicaffeoylquinic acid. Absorbance was monitored at 325 nm.
Compounds eluted in each peak were assumed from retention
times by comparison with authentic samples
Fig. 9 (A) Immunoblot analysis of a Coffea canephora leaf
crude protein extract with polyclonal antiserum (1/1,000) raised
against 4-His-tagged CYP98A3. (B) Immunoblot analysis of the
recombinant microsomes with polyclonal antiserum (1/10,000)
raised against 4-His-tagged CYP98A3. Twelve micrograms of
protein were loaded in each lane. A35, recombinant CYP98A35;
A36, recombinant CYP98A36; A3, recombinant CYP98A3; C-,
microsomes of yeast transformed with a void plasmid; M,
molecular mass markers (kDa).
156Plant Mol Biol (2007) 64:145–159
123

Page 13

and the proteins were concentrated in the vascular
bundles. Previous studies have shown that caffeoyl
quinic acids are localized in some cells forming the leaf
bundle sheath, and in the phloem cells and lignifying
tissues from petioles and stems (Mondolot et al., 2006).
As the polyclonal antibodies used, prepared against
CYP98A3, recognize with the same efficiency both
CYP98A35 and CYP98A36, we were unable to dif-
ferentiate their respective distributions. Our semi-
quantitative PCR experiments indicate that their genes
have very similar spatiotemporal expression patterns
at the level of the organ (except in old leaves), but
Fig. 10 Histolocalization of the caffeoylquinic acids, lignin, and
CYP98A35 and CYP98A36 proteins. Neu’s reagent was used to
visualize caffeoylquinic acids under UV light (A1, B1, C2) and
Mirande’s reagent for lignin under visible light (C1) in cross
sections of different organs of Coffea canephora 4-month-old
plants. Tissue print hybridization was performed on the same
organs (A2, B2, C3) by printing transversal hand cuts onto
nitrocellulose membranes. (A): (A1) Developing leaves from the
shoot tip; the greenish–white fluorescence (arrows), characteris-
tic of caffeoylquinic derivatives, was observed in all the cells, and
was particularly abundant in the blade mesophyll (m), and in the
medullar parenchyma (mp); (A2) identical distribution of
CYP98A35 and CYP98A36 was revealed using polyclonal
antibodies; (B): (B1) Petiole from the leaf of the node 2; the
greenish–white fluorescence (arrows) is restricted to the vascular
bundle, in cell layers surrounding the phloem and in the phloem
cells (ph). The blue fluorescence observed in the cell wall of the
xylem vessels (x) is due to hydroxycinnamic compounds bound
to lignin; (B2) In the same way, proteins are abundant in the
vascular bundle, particularly in the phloem compared to xylem
and cortical parenchyma (cp); (C): Stem 2, situated between the
node 2 and 3; (C1) vascular formations are clearly visible using
Mirande’s reagent i.e. primary phloem (pph), secondary phloem
(sph), xylem (x) and medullar parenchyma (mp); (C2) the level
of greenish–white fluorescence indicates that caffeoyl esters were
concentrated around the vascular bundle, principally in the cells
from the pph and sph, and in the medullar parenchyma; (C3)
proteins seem to be identically localized. Scale bars = 200 lm
Plant Mol Biol (2007) 64:145–159157
123

Page 14

provided
expression, which will have to be further investigated.
Taken together, our results, and particularly the lack
of correlation between CYP98A35 and CYP98A36
gene expression and the HQAs content in some organs
of the plant, indicate that the level of HQAs in plant
tissues resulted not only from biosynthesis, but also
from catabolism and transport/storage mechanisms. In
cotyledon leaves, where CYP98A35 and CYP98A36
expression is barely detectable, the high HQAs level
may be attributed to a great capacity of these old or-
gans to store compounds formed during the early
stages of plant development. This hypothesis is
supported by the results obtained by Aerts and
Baumann (1994) with C. arabica seedlings showing that
the cotyledonary 5-CQA was biosynthesized during
embryogenesis. It would support the proposed role of
5-CQA as a carbon reservoir and/or as a protective
anti-oxidant in young tissues. Identical roles can be
attributed to the HQAs biosynthesized in the plant
apex, even if their accumulation and/or storage mech-
anisms are still unknown. In aging organs, both genes
are expressed and proteins and HQAs are localized in
the vascular tissues, close to phloem cells. Protein
localization restricted to vascular bundles may indicate
that the pathway is mainly devoted to the synthesis of
lignin at this stage of plant development. This tissue-
specific localization may also be an effective plant
adaptation to enable rapid response to oxidative
stresses, either by synthesizing antioxidative com-
pounds in the vicinity of the tissues ensuring their
distribution in the plant, presumably in the form of
CQA, or by installing an antioxidant barrier near the
bundles to restrict systemic stress response.
A better understanding of the mechanisms leading
to the chlorogenic acid biosynthesis in coffee plants
will also be acquired by studying the hydroxycinna-
moyl-CoA: shikimate/quinate
transferases existing in these plants, other enzymes
which are involved in the last steps of the chlorogenic
acid biosynthetic pathway.
noinformation ontheir tissue-specific
hydroxycinnamoyl
Acknowledgements
graine and J. Bustamante by INIA from Venezuela
Mahesh V. was granted by Avesthagen
References
Abdulrazzak N, Pollet B, Ehlting J et al (2006) A coumaroyl-
3¢-ester hydroxylase insertion mutant reveals the existence
of nonredundant meta-hydroxylation pathways and essential
roles for phenolic precursors in cell expansion and plant
growth. Plant Physiol 140:30–48
Aerts RJ, Baumann TW (1994) Distribution and utilization of
chlorogenic acid in Coffea seedlings. J Exp Bot 45:497–503
Anthony F, Clifford MN, Noirot M (1993) Biochemical diversity
in the genus Coffea L.: chlorogenic acids, caffeine and
mozambioside contents. Genet Resour Crop Evol 40:61–70
Bertrand C, Noirot M, Doulbeau S et al (2003) Chlorogenic acid
content swap during fruit maturation in Coffea pseud-
ozanguebariae. Qualitative comparison with leaves. Plant
Sci 165:1355–1361
Campa C, Doulbeau S, Dussert S et al (2005) Qualitative
relationship between caffeine and chlorogenic acid contents
among wild Coffea species. Food Chem 93:135–139
Clifford MN (1985) Chlorogenic acids. In: Clarke RJ, Macrae R
(eds) Coffee, I: chemistry. Elsevier Applied Science,
Amsterdam, pp 153–202
Clifford MN (2000) Chlorogenic acids and other cinna-
mates—nature, occurence, dietary burden, absorption and
metabolism. J Sci Food Agric 80:1033–1043
Colonna JP (1986) Biosynthe `se et renouvellement de l’acide
chloroge ´nique et des depsides voisins dans le genre Coffea.
Cafe ´ Cacao The ´ 30:247–258
Franke R, Humphreys JM, Hemm MR et al (2002a) The
Arabidopsis REF8 gene encodes the 3-hydroxylase of
phenylpropanoid metabolism. Plant J 30:33–45
Franke R, Hemm MR, Denault JW et al (2002b) Changes in
secondary metabolism and deposition of an unusual lignin in
the ref8 mutant of Arabidopsis. Plant J 30:47–59
Gang DR, Beuerle T, Ullmann P et al (2002) Differential
production of meta hydroxylated phenylpropanoids in sweet
basil peltate glandular trichomes and leaves is controlled by
the activities of specific acyltransferases and hydroxylases.
Plant Physiol 130:1536–1544
Grace SC, Logan BA, Adams WW III (1998) Seasonal differ-
ences in foliar content of chlorogenic acid, a phenylpropa-
noid antioxidant, in Mahonia repens. Plant Cell Environ
21:513–521
Hoffmann L, Maury S, Martz F et al (2003) Purification, cloning,
and properties of an acyltransferase controlling shikimate
and quinate ester intermediates in phenylpropanoid metab-
olism. J Biol Chem 278:95–103
Jeong Y-M, Mun J-H, Lee I et al (2006) Distinct roles of the first
introns on the expression of Arabidopsis profiling gene
family members. Plant Physiol 140:196–209
Jin U-H, Lee J-Y, Kang S-K et al (2005) A phenolic compound,
5-caffeoylquinic acid (chlorogenic acid), is a new type and
strong matrix metalloproteinase-9 inhibitor: Isolation and
identification from methanol extract of Euonymus alatus.
Life Sci 77:2760–2769
Ku ¨hnl T, Koch U, Heller W et al (1987) Chlorogenic acid
biosynthesis: characterization of a light-induced microsomal
5-O-(4-coumaroyl)-D-quinate/shikimate 3¢-hydroxylase from
carrot (Daucus carota L.) cell suspension cultures. Arch
Biochem Biophys 258:226–232
Kumar S, Tamura K, Nei M (2004) MEGA3: integrated software
for molecular evolutionary genetics analysis and sequence
alignment. Brief Bioinform 5:150–163
Ky CL, Louarn J, Dussert S et al (2001) Caffeine, trigonelline,
chlorogenic acids and sucrose diversity in wild Coffea arabica
L. and canephora P. accessions. Food Chem 75:223–230
Ky CL, Noirot M, Hamon S (1997) Comparison of five
purification methods for chlorogenic acids in green coffee
beans (Coffea sp.) J Agric Food Chem 45:786–790
Leloup V, Louvrier A, Liardon R (1995) Degradation mecha-
nisms of chlorogenic acids during roasting. In: Proceedings
of the 16th International Congress ASIC, Kyoto, pp 9–14
Macheix J-J, Fleuriet A (1998) Phenolic acids in fruits. In: Rice-
Evans CA, Packer L (eds) Flavonoids in health and disease.
Marcel Decker Inc, New York, pp 35–59
158 Plant Mol Biol (2007) 64:145–159
123

Page 15

Maher EA, Bate NJ, Ni W et al (1994) Increase disease
suceptibility of transgenic tobacco plants with suppressed
levels of preformed phenylpropanoid products. Proc Natl
Acad Sci USA 91:7802–7806
Mahesh V, Rakotomalala JJ, Le Gal L et al (2006). Isolation and
genetic mapping of a Coffea canephora phenylalanine
ammonia-lyase gene (CcPAL1) and its involvement in the
accumulation of caffeoyl quinic acids. Plant Cell Rep
25:986–992
Matsuda F, Morino K, Miyashita M et al (2003) Metabolic flux
analysis of the phenylpropanoid pathway in wound-healing
potato tuber tissue using stable isotope-labeled tracer and
LC-MS spectroscopy. Plant Cell Physiol 44:510–517
Molgaard P, Ravn H (1988) Evolutionary aspects of caffeoyl
ester distribution in Dicotyledons. Phytochem 8:2411–2421
Mondolot L, Roussel J-L, Andary C (2001) New applications for
an old lignified element staining reagent. Histochem J
33:379–385
Mondolot L, La Fisca P, Buatois B et al (2006) Evolution in
Caffeoylquinic acid content and histolocalization during
Coffea canephora leaf development. Annals Bot 98:33–40
Mondolot-Cosson L, Andary C, Guang-Hui D et al (1997)
Histolocalization de substances phe ´noliques intervenant lors
d’interactions plante-pathoge `ne chez le tournesol et la
vigne. Acta Botanica Gallica 144:353–362
Morant M, Hehn A, Werck-Reichhart D (2002) Conservation
and diversity of gene families explored using the CODE-
HOP strategy in higher plants. BMC Plant Biol 2:7
Morant M, Schoch GA, Ullmann P et al (2007) Catalytic activity,
duplication and evolution of the CYP98 cytochrome P450
family in wheat. Plant Mol Biol 63:1–19
Neu R (1957) A new reagent for differentiating and determining
flavones on paper chromatograms. Naturwissenschaften
43:82
Niggeweg R, Michael AJ, Martin C (2004) Engineering plants
with increased levels of the antioxidant chlorogenic acid.
Nature Biotech 22:746–754
Omura T, Sato R (1964) The carbon monoxide binding pigment
of liver microsomes. I. Evidence for its hemoprotein nature.
J Biol Chem 239:2370–2378
Paquette SM, Bak S, Feyereisen R (2000) Intron-exon organi-
zation and phylogeny in a large superfamily, the paralogous
Cytochrome P450 genes of Arabidopsis thaliana. DNA Cell
Biol 19:307–317
Pompon D, Louerat B, Bronine A et al (1996) Yeast expression
of animal and plant P450s in optimized redox environments.
Methods Enzymol 272:51–64
Sambrook J, Russell DW (2001) Molecular cloning: A laboratory
manual. Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, New York
Schoch G, Goepfert S, Morant M et al (2001) CYP98A3 from
Arabidopsis thaliana is a 3¢-hydroxylase of phenolic esters, a
missing link in the phenylpropanoid pathway. J Biol Chem
276:36566–36574
Sto ¨ckigt J, Zenk H (1974) Enzymatic synthesis of chlorogenic
acid from caffeoyl coenzyme A and quinic acid. Phytochem
42:131–134
Takahama U (1998) Ascorbic acid-dependant regulation of
redox levels of chlorogenic acid and its isomers in the
apoplast of leaves of Nicotiana tabacum L. Plant Cell
Physiol 39:681–689
Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W:
improving the sensitivity of progressive multiple sequence
alignment through sequence weighting, position-specific gap
penalties and weight matrix choice. Nucleic Acids Res
22:4673–4680
Ulbrich B, Zenk M (1979) Partial purification and properties of
p-hydroxycinnamoyl-CoA:
transferase from higher plants. Phytochem 18:929–933
Ulbrich B, Zenk M (1980) Partial purification and properties of
p-hydroxycinnamoyl-CoA:
transferase from higher plants. Phytochem 19:1625–1629
Villegas RJA, Kojima M (1986) Purification and characterization
of Hydroxycinnamoyl
D-Glucose/Quinate hydroxycinna-
moyl transferase in the root of sweet potato, Ipomoea
batatas Lam. J Biol Chem 261:8729–8733
Zang LY, Cosma G, Gardner H et al (2003) Effect of
chlorogenic acid on hydroxyl radical. Mol Cell Biochem
247:205–210
quinatehydroxycinnamoyl
shikimate-phydroxycinnamoyl-
Plant Mol Biol (2007) 64:145–159159
123