Download Light and an exogenous transcription factor

Journal of Experimental Botany, Vol. 56, No. 414, pp. 1093–1103, April 2005
doi:10.1093/jxb/eri101 Advance Access publication 14 February, 2005
This paper is available online free of all access charges (see http://jxb.oupjournals.org/open_access.html for further details)
RESEARCH PAPER
Light and an exogenous transcription factor qualitatively
and quantitatively affect the biosynthetic pathway of
condensed tannins in Lotus corniculatus leaves
Francesco Paolocci, Tessa Bovone, Nicola Tosti, Sergio Arcioni and Francesco Damiani*
Istituto di Genetica Vegetale sez. Perugia, CNR, Via Madonna Alta 130, I-06128 Perugia, Italy
Received 30 August 2004; Accepted 8 December 2004
Abstract
Introduction
The effects of increasing light and of a heterologous
bHLH transcription factor on the accumulation of condensed tannins (CT) were investigated in leaves of Lotus
corniculatus, a model legume species which accumulates these secondary metabolites in leaves as well as
reproductive tissues. Light and expression of the transgene increased the level of CT in a synergistic way. To
monitor how the changes in accumulation of condensed
tannins were achieved, the level of expression of four
key genes in the flavonoid pathway was estimated by
real-time RT-PCR analysis. Early genes of the pathway
(PAL and CHS) were affected less in their expression and
so appeared to be less involved in influencing the final
level of CT than later genes in the pathway (DFR and
ANS). Steady-state levels of DFR and ANS transcripts
showed a strong positive correlation with CT and these
genes might be considered the first rate-limiting steps in
CT biosynthesis in Lotus leaves. However, additional
factors mediated by light are limiting CT accumulation
once these genes are up-regulated by the transgene.
Therefore, the increment of the steady-state mRNA level
for DFR and ANS might not be sufficient to up-regulate
condensed tannins in leaves. The real-time RT-PCR
approach adopted showed that members within the
CHS and DFR gene families are differentially regulated
by the exogenous bHLH gene and light. This finding is
discussed in relation to the approaches for controlling
CT biosynthesis and for studying the expression profile
of multi-gene families.
Flavonoid biosynthesis supplies plants with a large variety
of secondary metabolites, such as flavonols, anthocyanins,
and condensed tannins (CT). They are polyphenolic compounds, synthesized by higher plants in response to both
internal metabolic cues and external signals.
CT act as protectants of plants against pathogens, pests,
and diseases, and they control seed permeability and
dormancy (Debeaujon et al., 2000). The capacity of CT
to interact strongly with amino acid residues in proteins has
led to their wide industrial use.
The palatability and nutritive value of forage legumes are
highly influenced by the presence and structure of CT
(Barry and McNabb, 1999). The presence of CT in leaves
of forage plants protects ruminant animals against pasture
bloat (Tanner et al., 1995). CT in fact make the conversion
of plant protein into animal protein more efficient by
increasing the amount of rumen bypass protein, thereby
reducing the need to feed supplemental protein. However,
CT only accumulate in the seed coats of the most valuable
forage species such as alfalfa and clovers, and are absent
from their leaves. The ability to manipulate CT synthesis
and to direct biosynthesis to the leaves of forage species has
been one of the most challenging goals for forage breeders
over the past few decades (Damiani et al., 2000).
CT are oligomers of flavan-3-ol monomers such as
catechin or epicatechin and flavon 3,4-diols (leucoanthocyanidins) that form the ‘starter unit’ on which tannin
condensation takes place. No information, however, is
available on the nature of condensing steps, and this
polymerization process is believed to occur by progressive
addition of flavon 3,4-diols to the starter unit (Stafford,
1989). Some researchers have proposed a series of enzymedriven reactions; others claim that, because condensation
Key words: ANS, CHS, condensed tannins, DFR, elongation
factor, gene expression, light intensity, Lotus corniculatus, PAL,
real-time RT-PCR.
* To whom correspondence should be addressed. Fax: +39 07550 14869. E-mail: francesco.damiani@igv.cnr.it.
ª The Author [2005]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved.
The online version of this article has been published under an Open Access model. Users are entitled to use, reproduce, disseminate, or display the Open Access version of this article
for non-commercial purposes provided that: the original authorship is properly and fully attributed; the Journal and the Society for Experimental Biology are attributed as the original
place of publication with the correct citation details given; if an article is subsequently reproduced or disseminated not in its entirety but only in part or as a derivative work this must be
clearly indicated. For commercial re-use, please contact: journals.permissions@oupjournals.org
1094 Paolocci et al.
can be achieved in vitro, additional enzymes and therefore
genes may not be necessary (Haslam, 1999). Until recently,
the biosynthesis of anthocyanidin and CT was believed to
share the same sequence of reactions, as the last common
intermediate, leucocyanidin, is formed with the enzyme
dihydroflavonol reductase (DFR) as the catalyst (Ray et al.,
2003). However, recent findings have shown that epicatechin formation occurs via anthocyanidin synthesis and
reduction, two steps catalysed by anthocyanidin synthase
(ANS) and anthocyanidin reductase (ANR), respectively
(Xie et al., 2003). Anthocyanidins are consequently essential for CT biosynthesis in species that accumulate only
epicatechin-based CT such as Arabidopsis thaliana
(Abrahams et al., 2003). A scheme of the CT biosynthetic
pathway is given in Fig. 1.
The regulation of CT synthesis and its tissue accumulation is complex, involving several regulatory genes acting
on different steps of the pathway. Because tannins are one of
the terminal products of a highly branched pathway, which
leads to the synthesis of several compounds with a broad
action spectrum, several steps of activation are required.
Fig. 1. Schematic representation of the CT pathway in Lotus corniculatus. Enzymatic activities are reported in italics. The question mark
indicates undetermined enzymatic activities.
Some of these steps control the initial phases of the
synthesis, the so-called ‘early genes’ (Martin et al., 1991;
Martin and Gerats, 1993; Pellettier et al., 1997); other
terminal steps of the pathway (‘late genes’) lead to the
synthesis of the final compounds.
The characterization of mutants defective in anthocyanin
accumulation in petunia flowers (an mutants) and in
pigmentation in the Arabidopsis seed coat (tt mutants)
has resulted in the identification of important regulatory
genes. Pigment biosynthesis seems to be mainly regulated
at the transcription level since most of the regulatory genes
identified encode transcription factors. The majority of
them belong to the R2R3-MYB and bHLH families, with
help from TTG1, a WD40 protein that may recruit these and
other transcription factors to the promoters of target genes
(Walker et al., 1999). Factors belonging to the homeodomain WRKY and WIP families have also been shown to
be involved in regulating certain aspects of flavonoid
accumulation in Arabidopsis (Kubo et al., 1999; Eulgem
et al., 2000; Sagasser et al., 2002).
The activation of flavonoid biosynthesis results from
a complex network into which multiple signals input to
regulate gene expression. There is a large body of factors
determining how external stimuli affect flavonoid synthesis.
High light and UV, pathogen attack, nitrogen, phosphorus,
and iron deficiencies, low temperature, and wounding, all
affect the synthesis of different compounds derived from the
flavonoid pathway (Dixon and Paiva, 1995), but in no case
was CT accumulation affected by such environmental
conditions.
Carter et al. (1999) showed that CT levels in Lotus
corniculatus plants were positively affected by CO2 concentration and negatively influenced by drought and temperature. However, a significant influence of tissue type and
genotype, as well as an interaction between the different
treatments, was observed. Preliminary observations also
showed that light intensity can modify CT accumulation in
a genotype-dependent manner in L. corniculatus leaves
(Paolocci et al., 1999).
Lotus corniculatus transgenic lines expressing Sn, a gene
encoding a bHLH protein from maize that promotes
anthocyanin accumulation, had been produced previously
(Paolocci et al., 1999). Lotus plants expressing Sn showed
either an enhanced or a decreased accumulation of CT in
their leaves, depending on whether the transgene generated
Sn protein or whether it was involved in gene silencing
(Robbins et al., 2003). These lines may help to scale-down
the order of complexity of the regulatory machine of
flavonoid biosynthesis. The aim of the present study was
to investigate the role of genetic and environmental factors
on condensed tannin biosynthesis. To achieve this, the
effects of Sn and light on the production of condensed
tannins, and on the steady-state transcript levels of four key
genes in the pathway, were estimated in Lotus corniculatus
leaves.
Gene expression and the synthesis of condensed tannins
Materials and methods
Plant material
Lotus corniculatus L. genotypes S50 and S41, isolated from the
cultivar ‘Leo’, and characterized for their different levels of CT
accumulation in leaves were kindly provided by Dr Mark Robbins
(IGER, Aberystwyth, UK). S50 plants transformed with the maize
bHLH regulatory gene Sn, showing either an enhanced (plants S5010 and S50-11) or a reduced (named S50-6, S50-9, S50-20) level of
CT in leaves, were obtained as described previously (Paolocci et al.,
1999; Robbins et al., 2003). All the plants were repetitively assessed
for their accumulation of CT by DMACA staining, as reported by Li
et al. (1996), and for anthocyanin-containing cells by monitoring the
leaves with the naked eye and under a stereomicroscope after an
overnight ethanol treatment to decolorize chlorophyll.
Light treatment
For each of the seven plants under investigation (S50, S41, S50-10,
S50-11, S50-6, S50-9, S50-20) cuttings were produced from stem
segments rooted in sand in a humid chamber. Once roots were
developed, the plantlets were transferred to pots. After 4 weeks in
a humid chamber (95% humidity, 23 8C), the plants were moved to
a growth chamber (12 h of light at 650 lmol photons mÿ2 sÿ1, 22 8C)
and 6 weeks later were split into three groups. All the groups were
maintained in the same growth chamber but they were differentially
shadowed to irradiate plants with 100, 400, or 1000 lmol photons
mÿ2 sÿ1 of light. Each genotype was represented by four cuttings per
light treatment.
Leaves were collected on all plants after 4 weeks of light treatment
and 4 h after the light was switched on. Leaves from each plant were
partly stained for CT, and partly freeze-dried in liquid nitrogen and
stored at ÿ80 8C for molecular analysis.
Quantification of CT was performed on five randomly selected
young leaves per each treated plant. After DMACA staining, the
leaves were washed four times with distilled water and observed
under a stereomicroscope (36 magnification) and the number of CTcontaining cells counted as a proportion of the total number of cells in
five optical fields per leaf.
RNA isolation and analysis
Total RNA from four young leaves was isolated using the Nucleo
Spin RNA Plant isolation kit (Macherey Nagel) according to the
supplier’s instructions, to which a further DNase treatment was added.
Spectrophotometric RNA determination was carried out three times
for each sample. Poly (A)+ RNA was recovered from total RNA by
using the poly (A)+ tract system II (Promega) and resuspended in 20 ll
of RNAsi-free water. Five microlitres of mRNA were reverse
transcribed at 37 8C for 60 min in the presence of 200 units of
SuperScript II H- Reverse Transcriptase (Invitrogen) and 100 pmol of
random hexamers (Pharmacia Biotech) according to the enzyme’s
supplier.
1095
Gene cloning
Extension of the partial DFR sequences of Lotus already published
(Paolocci et al., 1999) was performed using the 59-RACE and 39RACE System for Rapid Amplification cDNA Ends Version 2.0
(GIBCO-BRL) according to the supplier’s instructions, by using the
adaptor primers provided with the kits in combination with the DFR
gene-specific primers (59-TGGCATTGTCGGCATTAGAAAGG-39)
and (59-AGAAGGCCAAAACTGTACGGAGG-39) for the 59- and
39-RACE, respectively. Primers were then designed on the extreme 59
(59-CAAGCCAACCCAAACATAG-39) and 39 (59-ACAATGCAAACACAG-39) sequence ends, to amplify the full-length cDNA
as well as the DFR gene from L. corniculatus. The full-length cDNA
sequence and a genomic DFR sequence have been registered in the
NCBI’s sequence database under the accession numbers AF117263
and AY633707, respectively.
The amplification of elongation factor 1-a (EF) cDNA was
achieved using the primer pairs reported previously (Robbins et al.,
2003). The resulting partial cDNA fragment has been deposited in
NCBI’s sequence database under the accession number AY633710.
By aligning cDNA sequences encoding PAL and b-tubulin (Tub)
from plant species closely related to the Lotus genus, the primer pairs
PalFw (59-GGRCAATGYGGYAACCARATCGG-39) and PalRev
(59-GCATCYTGGTAYTGYTGGTAYTC-39), and TubFw (59AARCAYCAYCCNGGICARATIGARGC-39) and Tubrev (59TTNACRTCYTGRTTRTGYTGYTGYTCIGC-39) were designed
and used in a 50 ll volume of PCR reaction on cDNA from poly
(A)+ RNA (see above) in the presence of 10 pmol of each primer. The
resulting partial cDNA fragments encoding PAL and Tub were
deposited in Genbank under the accession numbers AY633709 and
AY633708, respectively.
Sequence analysis
Double-strand sequence analysis was carried out, either directly on
PCR products or on PCR fragments previously cloned in the pGEM-T
Easy Vector System I (Promega), using the Big Dye Terminator Cycle
Sequencing Kit and an ABI Prism 310 Sequence Analyser (Applied
Biosystems) according to the supplier’s instructions. The sequencing
primers were the Sp6 and T7 vector primers for cloned PCR fragments
and gene-specific primers for direct sequencing reactions.
Primer selection for real-time PCR analysis
Primers for quantitative RT-PCR analysis were designed using
OligoExpress Software (Applied Biosystems) on the PAL, Tub,
DFR, and EF cDNA sequences obtained from L. corniculatus
accession S50 and from the ANS and CHS gene sequences from L.
corniculatus deposited in the databases (accession numbers
AY028931 and AF308143, respectively). All primers used in realtime RT-PCR experiments are listed in Table 1.
Real-time RT-PCR analysis
For real-time RT- PCR analysis 18 preparations of first-strand
cDNA were made from total RNA from six different genotypes
Table 1. List of primer pairs used for real-time RT-PCR analysis
Target gene
Forward primer 59–39
Reverse primer 59–39
EF1-a
Tub
DFR
Sn
CHS
PAL
ANS
AAGGAGGCCGCTGAGATGA
GTGGTAACCAGATCGGTTCGA
CAATCATTCCACCCCTTGTTG
ACGGGAGCAGCACAGGAA
TTGATGGCCACCTTTGTGAA
CTCAGTGGCTTGGACCTCAGA
TCCACCTTATCTTTCCAGAAGACAA
CGCTCAGCCTTCAATTTGTCA
CTCCCCTGCGGATCTAAGC
GATCGAATAATGGGCCTCGTT
GTCGCTTTCGCTCCGACAT
TGAAACAATCCCAGGAACATCTT
CGATCATCTTGCTTGCATGC
TCGCTGGTAACTTCCGTATAATATGA
1096 Paolocci et al.
(S50, S50-10; S50-11, S50-6, S50-9, S50-20) grown under three
different light conditions. cDNAs were diluted to 100 ll, and 4 ll
used for each real-time PCR experiment. The PCR reaction was made
up using the SyBrGreen PCR core mix (Applied Biosystems)
according to the supplier’s instructions in 25 ll of final volume in
the presence of 2.5 pmol of each primer. Cycling parameters were:
two initial steps of 50 8C for 2 min and 95 8C for 10 min for UNG and
AmpliTaq Gold activation, respectively, a two-step cycle of 95 8C for
30 s and 60 8C for 1 min repeated 50 times, and a final step of 10 min
at 60 8C. Afterwards, the dissociation protocol was performed.
Amplifications were performed on ABI PRISM 5700 SDS apparatus
(Applied Biosystems). Four replicates for each RNA sample per
target gene were amplified and, for each transcript, the average Ct
(threshold cycle) was determined. Ct is defined as the point at which
fluorescence rises appreciably above the background.
Standard curves for target and housekeeping genes were obtained
by the amplification of a serially diluted mixture of cDNA samples,
with six dilution points, each one replicated four times. Different
strategies were investigated to search for the most reliable reference
for the data sets. Firstly, spectrophotometric quantification of RNAs
was used. However, accurate measurement of total RNA levels is not
a particularly reliable standard for analyses based on PCR. Consequently, a gene quantification method based on the relative expression
of the target gene versus a reference gene, that is supposed to be stably
expressed, was adopted according to the ABI PRISM 5700 User
Bulletin 2 (Applied Biosystems). Relative quantification of gene
expression was achieved using both EF and Tub as reference genes.
The amount of each endogenous reference per sample was determined
from the appropriate standard curve, then for each RNA sample the
amount of target transcript was divided by the amount of the
endogenous reference to obtain a normalized value of the relative
expression levels.
The final number was then treated as a quantitative trait and statistically analysed. Two-way analysis of variance (Model II, all effects
random) and the Duncan test for mean comparison were performed
using the GLM Procedure of the SAS program (SAS Institute).
Results
Effects of light intensity on levels of CT
It has been shown previously that in the Lotus lines under
investigation the increment/decrement of leaf CT quantity
is reflected by the increment/decrement of the number of
CT-containing leaf cells (Robbins et al., 2003) as determined by DMACA staining, which specifically binds to
CT-containing cells. Under natural outdoor growing conditions, S50 and S41 genotypes showed different levels of
CT in their leaves. The first genotype accumulated very few
leaf tannins, the second many more (Carron et al., 1994), to
levels similar to those of Lotus uliginosus Schk., the Lotus
species accumulating the greatest amounts of CT in leaves.
Interestingly, the S41 genotype accumulates CT in all three
mesophyll layers – vascular, palisade, and spongy (data not
shown) – unlike S50 which has detectable CT only in
vascular tissues. These two genotypes also behaved very
differently following transformation with Sn (Paolocci
et al., 1999; Robbins et al., 2003). Nevertheless, both
S50 and S41 showed a progressive increase in CT levels at
increasing levels of irradiation, although, irrespective of the
light treatment, the S41 accession always showed a larger
number of CT-containing cells than S50 (Fig. 2).
S50-10 and S50-11 transgenic lines, hereinafter termed
‘E-plants’ (enhanced) because of the expression of the maize
gene Sn which encodes a bHLH regulatory protein, had CT
levels intermediate between those of S50 and S41. They had
also accumulated CT in spongy and palisade mesophyll
cells, regardless of the light treatment, in contrast to control
S50 (data not shown). Conversely, the plants S50-6, S50-9,
and S50-20, hereinafter called ‘S-plants’ (suppressed), due
to the Sn silencing effect (Paolocci et al., 1999; Robbins
et al., 2003), showed very low CT accumulation and were
very depleted in CT compared with control plants (S50-C) at
medium and high light intensities. Light positively increased
CT levels in all the accessions, but its effect was greatest in
S50-C and E plants due to increases in the numbers of
CT-containing cells (Fig. 2). Conversely, within each plant
class the number of anthocyanin-containing cells was unaffected by the light treatment.
Selection of target genes and gene expression
quantification procedures
To investigate by real-time RT-PCR analysis the expression
profiles of the genes involved in CT biosynthesis on S50-C,
E, and S lines grown in the three light regimes, four genes
in the phenylpropanoid pathway were selected: PAL, CHS,
DFR, and ANS. While the first two structural genes code for
enzymes in the early steps of the phenylpropanoid pathway,
DFR and ANS encode for later steps and are the most
downstream genes that, although only partially, have been
cloned so far from Lotus.
To quantify and compare the expression of target genes
or, better, the steady-state levels of their mRNA under
specific treatments, normalization for the amount of RNA
used in the different samples was needed. This is generally
achieved by comparing the level of the target gene with
Fig. 2. CT levels in leaves of C, S, E, and S41 plants grown at low (L),
medium (M), and high (H) light conditions, estimated as a percentage of
CT-containing cells. Values are averages of three S plants, two E plants,
and one control replicated four times.
Gene expression and the synthesis of condensed tannins
those of a reference that is constitutively expressed. However, it is conceivable that reference genes are also affected
by the specific environmental conditions being examined,
by the possible effects resulting from the activity of the
exogenous transcription factor (Sn), or by the rol genes
delivered to the host genome during transformation using
Agrobacterium rhizogenes. There is increasing evidence
that generally used housekeeping genes (such as GAPDH,
actins, tubulins, etc.) may vary significantly under different experimental conditions (Pfaffl et al., 2001). The
three methods used for RNA standardization (total RNA,
steady-state of EF, and Tub mRNA) did not in fact yield
identical results. To determine which was the most accountable reference the correlation among all the references was
calculated (data not shown).
A positive and significant correlation was observed
between EF and RNA, while the Tub transcript was totally
unrelated to the amount of total RNA; this implies that by
using EF or Tub as a reference can give rise to differing
results. EF is reported to be constitutively expressed and has
been used widely to normalize gene expression of flavonol
genes in different plant tissues and at different developmental stages (Rosati et al., 1997; Debeaujon et al., 2003).
Because EF is positively correlated to total RNA levels and,
importantly, because its levels also reflect variation at the
RNA reverse transcription step, it was selected as the most
suitable reference.
Gene expression and light
As expected, Sn expression was not detectable after 50
cycles of amplification in control and S lines (Robbins
et al., 2003), and was low even in E lines, which, under low
light conditions, showed a Ct of 34.6 against a Ct of 26.9 for
DFR (which was the most highly expressed gene). Because
Sn in the transgenic lines is driven by a strong promoter
(CAMV 35S) it can be excluded that it is poorly transcribed; conversely, the low steady-state level of Sn mRNA
can be accounted for by a high rate of transcript degradation, suggesting that transcript decay is as important as
transcription rate in determining the net level of gene
expression (Khodursky and Bernstein, 2003). No differences in Sn transcript levels were observed between highand low-light-treated E plants, indicating that light had no
effect on the expression of this transgene or on the
degradation of its transcript.
From the real-time RT-PCR a quantitative analysis of the
results of these expression experiments was undertaken.
There were four replicates for each plant (six) under each
light treatment (three), giving a total of 72 observations for
each of the four genes analysed. Since the aim was to
compare the effect of light on different genetic backgrounds
(S, E, and C plants), data obtained from S50-10 and S50-11,
as well as from S50-6, S50-20, and S50-9, were pooled. A
two-way ANOVA was performed; light and plant type were
1097
the main sources of variation, and their interaction, if
significant, the error. Table 2 reports the mean square
values and their significance for every gene analysed.
The significance of the interactions indicated that different plant types responded in a different way to light. This
happened for PAL and CHS but not for DFR and ANS
which behaved similarly. For such genes, E plants showed
significantly higher values (1.4 for DFR, 1.2 for ANS) than
the other two classes which had the same result for ANS
(both 0.9). C plants showed a higher DFR level than S (1.0
and 0.8, respectively). Light had a similar effect on both
genes in C and S plants; the ANS and DFR values at high
light turned out to be significantly higher than values at low
light, whereas their values at intermediate light were not
significantly different from those under low and high light
conditions.
A more detailed picture of the effect of light on gene
expression is given in Fig. 3, where, for all genes, the level
of each transcript in each plant type under different light
treatments is recorded. The high regime did not have
a significant effect on PAL gene expression of S plants,
although in C plants medium irradiation reduced PAL
transcript levels relative to low irradiation. PAL gene
expression was reduced significantly by high light compared with low light in E plants. Generally higher light
decreased CHS transcript levels in the different lines as well,
although high light significantly increased CHS transcript
levels in E plants compared with low light conditions.
DFR and ANS show a similar pattern of RNA accumulation at different light levels; the higher light level produced no effect in E and C plants, though the level of RNA
in S plants grown in high light conditions was significantly
different with respect to S plants at low light levels.
From the quantitative analyses it is concluded that, at least
for the isoforms studied, transcript levels of PAL and CHS
are not correlated with CT accumulation. Conversely,
higher levels of DFR and ANS transcripts are found in
CT-enhanced genotypes. But, while increases in the expression of these genes could explain the higher CT
accumulation of E plants relative to C/S plants, they do
not explain the difference in CT accumulation shown by E
plants under different light treatments (Figs 2, 3). In fact, E
plants at low and high light intensity did not show
significant differences in DFR and ANS transcript levels,
Table 2. Scheme of ANOVA and s2 values for each source of
variation
d.f. PALa
Plant type (P)
2
Light treatment (L) 2
Interaction (P3L)
4
Error
63
CHSa
0.350 n.s. 0.198 n.s.
0.002 n.s. 0.562 n.s.
0.061 ** 0.319 **
0.013
0.041
DFRa
ANSa
1.871 **
0.155 **
0.004 n.s.
0.027
0.843 **
0.144 *
0.093 n.s.
0.043
a
n.s., Not significant; **, significant at P >0.01; *, significant at P
>0.05.
1098 Paolocci et al.
Qualitative analysis of gene expression
Fig. 3. cDNA levels of PAL, CHS, DFR, and ANS genes relative to the
EF cDNA measured through real-time RT-PCR. Values are averages of
three S plants, two E plants and one control replicated four times. Within
each graph independently, values with the same letter do not differ
significantly (P <0.05).
suggesting that once DFR and/or ANS are expressed above
a threshold level, other genes, whose expression is enhanced
by light, need to be activated to lead to higher CT biosynthesis and accumulation.
SYBR Green I (SG) is widely used in real-time PCR
applications as an intercalating dye. Binding of SG to
double-stranded DNA is non-specific, and additional testing such as DNA melting curve analysis is required to
confirm the generation of a specific amplicon. The use of
melting curve analysis eliminates the necessity for agarose
gel electrophoresis, providing an even more accurate and
informative tool for the qualitative control of PCR reactions, allowing even single nucleotide polymorphisms
(SNPs) to be revealed among the amplicons generated. A
dissociation protocol was therefore performed at the end of
each experiment as a quality control for the PCR, and to
check for the presence of different gene members potentially co-amplified by the primer pairs designed. In fact,
some of the genes under investigation are generally
arranged in multi-gene families in legumes, but compared
with these genes only a few or partial L. corniculatus
sequences are present in public databases. While dissociation curves showed only a single peak relative to Sn, EF,
Tub, PAL, and ANS in any treatment, two or multiple peaks
were revealed in cDNA samples amplified with CHS and
DFR primer pairs. Specifically, CHS showed a main peak at
76 8C and a less-abundant one at 73 8C. The dimension of
this secondary peak was variable between samples and
might explain the contrasting results between E and C
plants. C and S plants showed two peaks at low light and
one at high light; E plants, on the contrary, maintained the
polymorphic condition, even at high light (Fig. 4).
For DFR transcripts, a primer pair spanning the fourth
and fifth exons was specifically designed to amplify a region
of the transcript known to be polymorphic in Lotus (F
Paolocci et al., unpublished data). The melting curve
analysis showed the presence of a major peak, between
81 and 83 8C, and one or more secondary ones at 78 and
74 8C. Figure 5 shows the DFR dissociation curve resulting
from C, E, and S plants under high and low light conditions.
The products of the real-time RT-PCR were sequenced
from samples producing multiple peaks. While confirming
an amplicon size of 107 bp, as predicted for the amplification of DFR cDNA, sequencing revealed the presence of
several differences (SNPs) between the DFR cDNAs. After
cloning these cDNA fragments, three polymorphic sequences were obtained, named col2, col3, and col16 (Fig.
6). These cDNA transcripts potentially encode DFR proteins with distinctive amino acid composition and potentially different catalytic specificities.
To identify which of the three distinct DFR transcripts
was expressed under the condition tested the melting curve
analysis was performed on the three plasmids carrying the
polymorphic cDNA fragments.
When amplified in purity, the three plasmids showed
a Tm profile comprising a single dissociation peak of about
83, 82, and 81 8C for col16, col2, and col3, respectively, in
Gene expression and the synthesis of condensed tannins
1099
Fig. 4. Dissociation pattern of CHS amplicons from cDNA of C, E, and S plants grown under high and low light conditions.
accordance with the reduction in the pyrimidine number.
When a mix consisting of equal amounts of col2 and col3
was amplified, a major peak at 82 8C and a secondary one at
78 8C were observed, the secondary one likely representing
the dissociation of non-totally homologous sequence
strands. When the relative concentration of col3 increased,
the main peak shifted toward 81 8C. When only col2 and
col16 were present, the dissociation pattern showed only
one peak at 82 8C, regardless of their relative concentrations. When col3 and col16 were equally represented in the
mix, the dissociation pattern showed three peaks; at 83 8C
and 81 8C, each one specific for the pure cDNA fragments,
and one at 76 8C that was attributed to the dissociation of
the non-homologous annealed stretches.
Since, in the leaf cDNA analysed, a single peak was
never observed, then at least two different DFR forms must
always be co-expressed in Lotus leaves. At high light, all
plants showed two peaks at 81–82 8C and 78 8C indicating
that the forms col2 and col3 were present. At low light, E
plants showed a major peak at 81 8C along with a minor one
at 78 8C, indicating that the form equivalent to col3 was
predominant and a col2-like DFR form was also moderately
expressed. Conversely, in C and S genotypes, one large
peak at 82 8C and two more at 78 and 74 8C were present,
indicating that col2, col3, and col16 isoforms were simultaneously transcribed and accumulated. Of course, the
present analysis may not have identified all the different
DFR transcripts expressed in L. corniculatus leaves.
Discussion
The effect of environmental conditions on CT accumulation
is quite well documented, the effects of temperature and
1100 Paolocci et al.
Fig. 5. Dissociation pattern of DFR amplicons from cDNA of C, E, and S plants grown under high and low light conditions.
Fig. 6. Nucleotide and deduced amino acid sequences of three cloned
DFR amplicons (col2, accession no. AY633711; col3, no. AY633712;
col16, no. AY633713) resulting from real-time RT-PCR of control plants
under low light condition. Boxed areas indicate polymorphisms.
seasonality of plant growth on biosynthesis having been
analysed (Grebrehiwot et al., 2002). However, no information about the effect of light on their accumulation has yet
been reported. It is now reported that light and Sn expression
induce CT accumulation synergistically in L. corniculatus
leaves and that the light effect is more evident in plants
with low levels of CT (S50) compared with those producing
higher amounts of CT, such as S41 plants (Fig. 2).
Within each plant class, light seems to induce quantitative modulation on the target genes tested within quite
a narrow range (Fig. 3), but ANOVA shows a significant
and positive effect of light on DFR and ANS transcript
accumulation (Table 2). Conversely, no significant increases in PAL and CHS transcripts in response to light were
observed, although there was a significant interaction
between plant type and light. However, CHS transcript(s)
decreased in control and S plants when they are moved
from low to high light conditions (Fig. 3), and this trend
was paralleled by a reduction in the number of CHS
isoforms expressed (Fig. 4). On the contrary, E plants
showed a positive response of CHS transcript levels to
increased light, although isoform expression did not change
with varying light conditions. The present data contrast
with the observation reported by Ray et al. (2003) that
Gene expression and the synthesis of condensed tannins
transgenic alfalfa lines expressing Lc (a Sn paralogue) had
higher levels of CHS RNA, although Lc did not detectably
affect DFR transcript levels in alfalfa.
The role of CHS in tannin biosynthesis is somewhat
controversial. Colliver et al. (1997) obtained an increase of
CHS expression and CT levels in L. corniculatus hairy root
lines transformed with a heterologous CHS gene in antisense orientation. They proposed that the silencing of the
gene encoding one CHS isoform might induce the overexpression of the other gene members, thereby inducing
a possible re-routing of the phenylpropanoid biosynthetic
pathway In pea, CHS is encoded by a multi-gene family,
and the response of different members to elicitor and UV
irradiation varies during development and in response to
various external stimuli (Ito et al., 1997). Therefore, for
a finely tuned assessment of the effect of light and Sn on
CHS, the cloning and transcription analysis of each individual gene member is required. The finding that the
primer pair used can amplify two distinct CHS transcripts,
that have a differential expression pattern, confirms that the
regulation of CHS activity is complex and multi-genic.
Altogether, the present data stress the existence of a lightinsensitive phenylpropanoid pathway in Lotus leaves that
may ensure CT synthesis even when environmental conditions are not optimal. Interestingly, within each gene
family, the activity of one or only few gene members may be
dedicated to CT synthesis. This model is confirmed by the
expression of PAL. This gene was selected as a potential
endogenous control of the phenylpropanoid path, since it is
an early gene in this pathway and in maize is not transactivated by Sn (Tonelli et al., 1991). Unexpectedly,
a down-regulation of PAL transcripts by Sn was observed,
although this did not prevent the accumulation of higher
levels of CT in response to Sn. For this gene, extra care is
required for interpretation of the results because the pair of
primers selected amplify preferentially one particular member of the PAL gene family that is highly homologous to
PAL2 which, in Populus tremuloides, appeared to be less
associated with CT synthesis than other PAL genes (Kao
et al., 2002). Thus, the possibility cannot be excluded that
light and/or Sn stimulates expression of a highly expressed
PAL gene, that was not detected with the present primer pair,
which may significantly affect CT levels.
Expression of later steps in the pathway is more strictly
correlated with CT levels. In fact, at higher levels of DFR
and ANS there are correspondingly increased CT levels in S
and C plants. In E plants the effect of light on DFR and ANS
steady-state transcript levels is masked by Sn expression,
proving that the constitutive expression of the exogenous
regulator overrides the activity of the endogenous activator
of these genes. However, light affects CT levels in E plants,
probably through an effect on expression of endogenous
regulatory genes which may promote transcription/activity
of genes/enzymes downstream of DFR in the pathway.
Candidate genes are BAN (Devic et al., 1999), coding for
1101
anthocyanidin reductase (Xie et al., 2003), and LAR (Tanner
et al., 2003), whose cloning in Lotus is in progress in the
authors’ laboratory. It is worth mentioning that, in alfalfa,
the maize Lc gene does transactivate LAR transcript levels
(Ray et al., 2003), while in Arabidopsis TT8, a bHLH
protein, positively regulates DFR and BAN (Nesi et al.,
2000). Previous observations showed that Sn silencing
negatively affects LAR activity in transgenic Lotus (Damiani
et al., 1999). It can thus be speculated that, if LAR/BAN
expression is induced by Sn in the same way as DFR and
ANS, the effect of light would probably operate through later
genes in the CT pathway, such as those controlling polymerization and/or CT compartmentalization in vacuoles.
The use of the SG instead of an allele-specific probe in
real-time RT-PCR analysis was successful in monitoring
the expression of different members of a small DFR multigene family in Lotus (Robbins et al., 1998). Sn seems not
only to trigger the expression of col3-like DFR specifically,
but also to prevent the expression of those DFR forms
activated under low light conditions. In this scenario, the
col3 DFR gene should be the one strictly committed to CT
synthesis and Sn should show a differential activity on the
promoter regions of the different DFR genes. Preliminary
results have proved the presence of multiple DFR genes in
the S50 genome. Additionally, the predicted DFR proteins
encoded by col3, col2, and col16 clones showed two amino
acid substitutions (Fig. 6) downstream of the region that
determines the DFR substrate specificity (Johnson et al.,
2001). Extension of the cDNA sequences to the 59 end of
col16 and col3 did not reveal any amino acid substitutions
at the key positions responsible for substrate specificity of
DFR (F. Paolocci, unpublished results). Thus, it is likely
that all the DFR members identified as differentially expressed in response to light treatment and transgene
expression code for enzymes with similar substrate specificity; also it is likely that they may substitute functionally
for each other, as is also suggested for CHS. This could
explain the results obtained by Robbins et al. (1998) who
increased the level of CT through DFR antisense. However,
very recently Xie et al. (2004) identified two DFR genes in
Medicago truncatula, with remarkably similar patterns of
transcript accumulation in most tissues, coding for two
different and catalytically active DFR proteins. They proposed that, beside the single amino acid change in the
region identified by Johnson et al. (2001), other active site
residues also have an effect on substrate preferences of the
enzyme. Therefore, the possibility that the three cDNAs
identified in Lotus encode for functionally distinct DFR
proteins cannot be ruled out completely.
The expression analysis of ANS in a species accumulating both epicatechin- and catechin-based CT such as L.
corniculatus (Fig. 1) is of importance to both anthocyanin
and CT synthesis. In Arabidopsis, TT8 was shown to have
the dual metabolic function in boosting both CT and
anthocyanin routes (Nesi et al., 2000). Foo et al. (1996)
1102 Paolocci et al.
showed that the CT extender flavan units consist mostly of
epicatechin in L. corniculatus. A previous observation
showed that anthocyanins accumulated very poorly in L.
corniculatus leaves and that Sn expression increases
anthocyanin accumulation in only a few cells located in
the petiole, leaf base, and next to the central leaf midrib
region (Robbins et al., 2003). CT quantification carried out
through cell counts and HPLC analyses is highly concordant in leaves of Sn transgenic Lotus lines (Robbins et al.,
2003), while the leaf anthocyanin spectrophotometric
quantitation carried out according to Ray et al. (2003) did
not allow differences between leaves of Sn-induced and
-suppressed plants to be appreciated under the different
light conditions (data not shown). The observation that ANS
is clearly transactivated by Sn, irrespective of the light
treatment, and that this transactivation as well as the light
effect is largely devoted to accumulation of CT rather than
anthocyanins lead to the argument that both Sn and light
increase the flux of flavonoids in Lotus leaves mainly
towards the production of CT through the synthesis of both
catechin and epicatechin.
Despite the complexity of the genome, the organization
of most of the phenylpropanoid genes as small multi-gene
families in legumes, and the relatively few genes sequenced,
an experimental procedure has been set up to assay gene
modulation in L. corniculatus by real-time PCR analysis.
Real-time PCR and the use of SG were demonstrated to
be powerful, as well as inexpensive, methods of investigation to assess quantitative and qualitative modulation of
gene expression. The design of primer pairs spanning
polymorphic regions proved to be an efficient method to
detect the internal and external stimuli that specifically
trigger the expression/silencing of a specific gene, among
different members of a gene family. The present results also
provide a basis to support the proposal that silencing or
trans-activation of members of such structural gene families
do not necessarily have the desired effects if the action is not
targeted to specific isoforms (Colliver et al., 1997; Robbins
et al., 1998). Similarly, the present results suggest that
northern analysis alone cannot provide exhaustive or even
correct information for expression profiling of multi-gene
families (Gachon et al., 2004), particularly in polyploid
species.
Acknowledgements
Thanks are due to Dr Cathie Martin (John Innes Centre) for the
careful critical reading of the manuscript and for her useful comments and suggestions.
References
Abrahams S, Lee E, Walker AR, Tanner GJ, Larkin PJ, Ashton
AR. 2003. The Arabidopsis TDS4 gene encodes leucoanthocy-
anidin dioxygenase (LDOX) and is essential for proanthocyanidin
synthesis and vacuole development. The Plant Journal 35,
624–636.
Barry TN, McNabb WC. 1999. The implications of condensed
tannins on the nutritive value of temperate forages fed to
ruminants. British Journal of Nutrition 81, 263–272.
Carron TR, Robbins MP, Morris P. 1994. Genetic modification of
condensed tannin biosynthesis in Lotus corniculatus. 1. Heterologous antisense dihydroflavonol reductase down-regulates tannin
accumulation in ‘hairy root’ cultures. Theoretical and Applied
Genetics 8, 1006–1015.
Carter EB, Theodorou MK, Morris P. 1999. Responses of Lotus
corniculatus to environmental change. 2. Effect of elevated CO2,
temperature and drought on tissue digestion in relation to condensed tannin and carbohydrate accumulation. Journal of the
Science of Food and Agriculture 79, 1431–1440.
Colliver SP, Morris P, Robbins MP. 1997. Differential modification of flavonoid and isoflavonoid biosynthesis with an antisense
chalcone synthase construct in transgenic Lotus corniculatus.
Plant Molecular Biology 35, 509–522.
Damiani F, Paolocci F, Cluster PD, Arcioni S, Tanner GJ,
Joseph RG, Li YG, de Majnik J, Larkin PJ. 1999. The maize
Sn regulatory transgene alters proanthocyanidin production in
Lotus corniculatus in some tissues by transactivation and in others
by co-suppression of an endogenous regulator. Australian Journal
of Plant Physiology 26, 159–169.
Damiani F, Paolocci F, Turchetti V, Arcioni S. 2000. Toward the
identification of the genes for the synthesis of condensed tannins in
forage legumes. In: de Vries GE, Metzlaff K, eds. Phytospere ‘99
Highlights on European plant biotechnology research and technology transfer. Amsterdam: Elsevier, 257–263.
Debeaujon I, Léon-Kloosterziel KM, Koornneef M. 2000. Influence of the testa on seed dormancy, germination, and longevity
in Arabidopsis. Plant Physiology 122, 403–414.
Debeaujon I, Nesi N, Perez P, Devic M, Grandjean O, Caboche
M, Lepiniec L. 2003. Proanthocyanidin-accumulating cells in
Arabidopsis testa, regulation of differentiation and role in seed
development. The Plant Cell 15, 2514–2531.
Devic M, Guilleminot J, Debeaujon I, Bechtold N, Bensaude E,
Koornneef M, Pelletier G, Delseny M. 1999. The BANYULS
gene encodes a DFR-like protein and is a marker of early seed coat
development. The Plant Journal 19, 387–398.
Dixon RA, Paiva NL. 1995. Stress-induced phenylpropanoid metabolism. The Plant Cell 7, 1085–1097.
Eulgem T, Rushton PJ, Robatzek S, Somssich IE. 2000. The
WRKY superfamily of plant transcription factors. Trends in Plant
Science 5, 199–206.
Foo LY, Newman R, Waghorn G, McNabb WC, Ulyatt MJ. 1996.
Proanthocyanidins from Lotus corniculatus. Phytochemistry 41,
617–624.
Gachon C, Mingam A, Charrier B. 2004. Real-time PCR, what
relevance to plant studies? Journal of Experimental Botany 55,
1445–1454.
Grebrehiwot L, Beuselinck PR, Roberts CA. 2002. Seasonal
variations in condensed tannin concentration of three Lotus
species. Agronomy Journal 94, 1059–1065.
Haslam E. 1999. Che faro senza polifenoli? In: Groos GG,
Heminghway RW, Yoshida T, eds. Plant polyphenols. 2. Chemistry, biology, pharmacology, ecology. New York: Kluwer/Plenum
Press, 15–40.
Ito M, Ichinose Y, Kato H, Shiraishi T, Yamada T. 1997. Molecular evolution and functional relevance of the chalcone synthase
genes of pea. Molecular Genetics and Genomics 255, 28–37.
Gene expression and the synthesis of condensed tannins
Johnson ET, Ryu S, Yi H, Shin B, Cheong H, Choi G. 2001.
Alteration of a single amino acid changes the substrate specificity
of dihydroflavonol 4-reductase. The Plant Journal 25, 325–333.
Kao YY, Harding SA, Tsai CJ. 2002. Differential expression of
two distinct phenylanine ammonia-lyase genes in condensed
tannin-accumulating and lignifying cells of quaking aspen. Plant
Physiology 130, 796–807.
Khodursky AB, Bernstein JA. 2003. Life after transcription –
revisiting the fate of messenger RNA. Trends in Genetics 19,
113–115.
Kubo H, Peeters AJM, Aarts MGM, Pereira A, Koornneef M.
1999. ANTHOCYANINLESS2, a homeobox gene affecting anthocyanin distribution and root development in Arabidopsis. The
Plant Cell 11, 1217–1226.
Li YG, Tanner GJ, Larkin PJ. 1996. The DMACA-HCl protocol
and the threshold proanthocyanidin content for bloat safety in
forage legumes. Journal of the Science of Food and Agriculture
70, 89–101.
Martin C, Gerats T. 1993. Control of pigment biosynthesis genes
during petal development. The Plant Cell 5, 1253–1264.
Martin C, Prescott A, Mackay S, Bartlett J, Vrijlandt E. 1991.
Control of anthocyanin biosynthesis in flowers of Antirrhinum
majus. The Plant Journal 1, 37–49.
Nesi N, Debeaujon I, Jond C, Pelletier G, Caboche M, Lepiniec
L. 2000. The TT8 gene encodes a basis helix–loop–helix domain
protein required for expression of DFR and BAN genes in
Arabidopsis siliques. The Plant Cell 12, 1863–1878.
Paolocci F, Capucci R, Arcioni S, Damiani F. 1999. Birdsfoot
trefoil, a model for studying the synthesis of condensed tannins.
In: Groos GG, Heminghway RW, Yoshida T, eds. Plant polyphenols. 2. Chemistry, biology, pharmacology, ecology. New York:
Kluwer/Plenum Press, 343–356.
Pelletier MK, Murrell JR, Shirley BW. 1997. Characterization of
flavonol synthase and leucoanthocyanidin dioxygenase genes in
Arabidopsis (further evidence for differential regulation of ‘early’
and ‘late’ genes). Plant Physiology 113, 1437–1445.
Pfaffl MW. 2001. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acid Research 29, 2002–
2007.
Ray H, Yu M, Auser P, Blahut-Beatty L, McKersie B, Bowley S,
Westcott N, Coulman B, Lloyd A, Gruber MY. 2003. Expression of anthocyanins and proanthocyanidins after transformation of
alfalfa with maize Lc. Plant Physiology 132, 1448–1463.
Robbins MP, Bavage AD, Strudwicke C, Morris P. 1998. Genetic
manipulation of condensed tannins in higher plants. 2. Analysis of
1103
birdsfoot trefoil plants harboring antisense dihydroflavonol reductase constructs. Plant Physiology 116, 1133–1144.
Robbins MP, Paolocci F, Hughes JW, Turchetti V, Allison G,
Arcioni S, Morris P, Damiani F. 2003. Sn, a maize bHLH gene,
transactivates anthocyanin and condensed tannin pathways in
Lotus corniculatus. Journal of Experimental Botany 54, 239–248.
Rosati C, Cadic A, Duron M, Renou JP, Simoneau P. 1997.
Molecular cloning and expression analysis of dihydroflavanol 4reductase gene in flower organs of Forsythia3intermedia. Plant
Molecular Biology 35, 303–311.
Sagasser M, Lu GH, Hahlbrock K, Weisshaar B. 2002.
A. thaliana TRANSPARENT TESTA 1 is involved in seed
coat development and defines the WIP subfamily of plant zinc
finger proteins. Genes and Development 16, 138–149.
Stafford HA. 1989. The enzymology of proanthocyanidin biosynthesis. In: Hemingway RW, Karchesy JJ, eds. Chemistry and
significance of condensed tannins. New York: Plenum Press, 47–
70.
Tanner GJ, Francki KT, Abrahams S, Watson JM, Larkin PJ,
Ashton AR. 2003. Proanthocyanidin biosynthesis in plants.
Purification of legume leucoanthocyanidin reductase and molecular cloning of its cDNA. Journal of Biological Chemistry 278,
31647–31656.
Tanner GJ, Moate PJ, Davis LH, Laby RH, Li YG, Larkin PJ.
1995. Proanthocyanidins (condensed tannin) destabilize plant
protein foams in a dose dependent manner. Australian Journal of
Agricultural Research 46, 1101–1109.
Tonelli C, Consonni G, Dolfini SF, Dellaporta SL, Viotti A,
Gavazzi G. 1991. Genetic and molecular analysis of Sn, a lightinducible tissue-specific regulatory gene in maize. Molecular
Genetics and Genomics 225, 401–410.
Walker AR, Davison PA, Bolognesi-Winfield AC, James CM,
Srinivasan N, Blundell TL, Esch JJ, Marks MD, Gray JC.
1999. The TRANSPARENT TESTA GLABRA1 locus, which
regulates trichome differentiation and anthocyanin biosynthesis in
Arabidopsis, encodes a WD40 repeat protein. The Plant Cell 11,
1337–1350.
Xie DY, Dharma SB, Paiva NL, Ferreira D, Dixon RA. 2003. Role
of the anthocyanidin reductase, encoded by BANYULS in plant
flavonoid biosynthesis. Science 299, 396–399.
Xie DY, Jackson LA, Cooper JD, Ferreira D, Paiva NL. 2004.
Molecular and biochemical analysis of two cDNA clones encoding
dihydoflavonol-4-reductase from Medicago truncatula. Plant
Physiology 134, 979–994.