Time course and amplitude of DNA methylation in the shoot apical meristem are critical points for bolting induction in sugar beet and bolting tolerance between genotypes.

Abstract

An epigenetic control of vernalization has been demonstrated in annual plants such as Arabidopsis and cereals, but the situation remains unclear in biennial plants such as sugar beet that has an absolute requirement for vernalization. The role of DNA methylation in flowering induction and the identification of corresponding target loci also need to be clarified. In this context, sugar beet (Beta vulgaris altissima) genotypes differing in bolting tolerance were submitted to various bolting conditions such as different temperatures and/or methylating drugs. DNA hypomethylating treatment was not sufficient to induce bolting while DNA hypermethylation treatment inhibits and delays bolting. Vernalizing and devernalizing temperatures were shown to affect bolting as well as DNA methylation levels in the shoot apical meristem. In addition, a negative correlation was established between bolting and DNA methylation. Genotypes considered as resistant or sensitive to bolting could also be distinguished by their DNA methylation levels. Finally, sugar beet homologues of the Arabidopsis vernalization genes FLC and VIN3 exhibited distinct DNA methylation marks during vernalization independently to the variations of global DNA methylation. These vernalization genes also displayed differences in mRNA accumulation and methylation profiles between genotypes resistant or sensitive to bolting. Taken together, the data suggest that the time course and amplitude of DNA methylation variations are critical points for the induction of sugar beet bolting and represent an epigenetic component of the genotypic bolting tolerance, opening up new perspectives for sugar beet breeding.

Full text

Page 1

Journal of Experimental Botany, Page 1 of 13
doi:10.1093/jxb/erq433
RESEARCH PAPER
Time course and amplitude of DNA methylation in the shoot
apical meristem are critical points for bolting induction in
sugar beet and bolting tolerance between genotypes
Marie-Ve ´ronique Trap-Gentil1,2,*, Claire He ´brard1,2,3, Cle ´ment Lafon-Placette1,2, Alain Delaunay1,2,
Daniel Hage `ge1, Claude Joseph1,2, Franck Brignolas1,2, Marc Lefebvre3, Steve Barnes3and Ste ´phane Maury1,2,†
1Universite ´ d’Orle ´ans, UFR-Faculte ´ des Sciences, UPRES EA 1207 ‘Laboratoire de Biologie des Ligneux et des Grandes Cultures’
(LBLGC), rue de Chartres, BP 6759, F-45067 Orle ´ans, France
2INRA, USC1328 Arbres et Re ´ponses aux Contraintes Hydriques et Environnementales (ARCHE), F-45067 Orle ´ans, France
3SESVanderHave NV/SA, Soldatenplein Z2 nr15, Industriepark, B-3300 Tienen, Belgium
* Present adress: INRA Rennes, UMR 1099 BiO3P INRA-Agrocampus Rennes-Universite ´ Rennes 1, BP 35327, F-35653 Le Rheu
Cedex, France.
yTo whom correspondence should be addressed. E-mail: stephane.maury@univ-orleans.fr
Received 3 September 2010; Revised 25 November 2010; Accepted 3 December 2010
Abstract
An epigenetic control of vernalization has been demonstrated in annual plants such as Arabidopsis and cereals, but
the situation remains unclear in biennial plants such as sugar beet that has an absolute requirement for
vernalization. The role of DNA methylation in flowering induction and the identification of corresponding target loci
also need to be clarified. In this context, sugar beet (Beta vulgaris altissima) genotypes differing in bolting tolerance
were submitted to various bolting conditions such as different temperatures and/or methylating drugs. DNA
hypomethylating treatment was not sufficient to induce bolting while DNA hypermethylation treatment inhibits and
delays bolting. Vernalizing and devernalizing temperatures were shown to affect bolting as well as DNA methylation
levels in the shoot apical meristem. In addition, a negative correlation was established between bolting and DNA
methylation. Genotypes considered as resistant or sensitive to bolting could also be distinguished by their DNA
methylation levels. Finally, sugar beet homologues of the Arabidopsis vernalization genes FLC and VIN3 exhibited
distinct DNA methylation marks during vernalization independently to the variations of global DNA methylation.
These vernalization genes also displayed differences in mRNA accumulation and methylation profiles between
genotypes resistant or sensitive to bolting. Taken together, the data suggest that the time course and amplitude of
DNA methylation variations are critical points for the induction of sugar beet bolting and represent an epigenetic
component of the genotypic bolting tolerance, opening up new perspectives for sugar beet breeding.
Key words: Beta vulgaris altissima (sugar beet), bisulfite, bolting, devernalization, DNA methylation, DNMT, FLOWERING
LOCUS C (FLC), McrBC, vernalization, VERNALIZATION INSENSITIVE 3 (VIN3).
Introduction
Vernalization is the acquisition or acceleration of compe-
tence to bolt and flower resulting from exposure to extended
periods at low temperatures that mimic winter conditions.
This process typically relieves a block to the photoperiod
pathway (Chouard, 1960; Schmitz and Amasino, 2007). In
some species, vernalization can be reversed by devernaliza-
tion, an exposure to a few days of temperatures ranging
from 25 ?C to 35 ?C (Purvis and Gregory, 1952). Mitotic
Abbreviations: BD, bolting delay; BI, bolting index; DBI, differential bolting index between vernalization and devernalization; DNMT, DNA methyltransferase; FLC,
FLOWERING LOCUS C; HPLC, high-performance liquid chromatography; MS-PCR, methylation-sensitive PCR; %mC, percentage of methylcytosine; VIN3,
VERNALIZATION INSENSITIVE 3.
ª The Author [2011]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved.
For Permissions, please e-mail: journals.permissions@oup.com

Journal of Experimental Botany Advance Access published January 12, 2011
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 2

activity in meristematic cells is required during vernaliza-
tion for the subsequent formation of floral organs. Early
studies associated vernalization with a form of cell memory
susceptible to being maintained for weeks until optimal
light conditions occurred (Lang, 1965). This phenomenon
is transmissible through mitosis, therefore pointing to
an epigenetic mode of regulation (Metzger, 1988). Two
main findings suggested the involvement of DNA methyla-
tion in the vernalization process: hypomethylating treat-
ments could mimic the effects of vernalization on flowering
(Burn et al., 1993) and plants displaying genome-wide
defects in DNA methylation exhibited an early flowering
phenotype (Finnegan et al., 1998). Other studies have
strengthened the hypothesis for epigenetic regulation of
vernalization (Brock and Davidson, 1994; Fieldes et al.,
2005; Dennis and Peacock, 2007; Schmitz and Amasino,
2007). FLOWERING LOCUS C (FLC), the central gene of
vernalization in winter-annual Arabidopsis thaliana, is re-
pressed both in vernalized plants and in plants with low
levels of genomic methylation, which in turn de-represses
flowering. Nevertheless, FLC is not regulated directly by
modifications of DNA methylation (Finnegan et al., 2005).
It has been proposed that the reduction of FLC activity by
low methylation is likely to depend on a change of methyla-
tion status at another locus, resulting in low FLC expression
and formation of inactive chromatin structures in the FLC
gene region (Genger et al., 2003; Finnegan et al., 2005; Dennis
and Peacock, 2007; Schmitz and Amasino, 2007). The mole-
cular mechanism supporting cell memory of vernalization was
then elucidated. The induction of the VERNALIZATION
INSENSITIVE 3 (VIN3) gene by cold promotes histone H3
deacetylation at the FLC locus. This epigenetic modification
relieves FLC-mediated repression of flowering (Bastow et al.,
2004; Sung and Amasino, 2004). Moreover, prolonged cold
conditions promote histone H3 methylation at both Lys9 and
Lys27 (H3K9 and H3K27) through the recruitment of a
complex containing the VERNALIZATION 2 (VRN2) pro-
tein. These epigenetic modifications lead to a stable repression
of FLC that will be reset in the progeny.
Orthologues of genes involved in the vernalization
pathway were identified in several plant species (Dennis
and Peacock, 2007; Reeves et al., 2007), revealing interspe-
cific differences in the control of the flowering process. For
instance, whereas the sugar beet FLC-like gene (BvFLC)
behaves as a repressor of flowering in transgenic Arabidop-
sis plants, its transcriptional activity is restored upon return
to warmer temperatures (Reeves et al., 2007). Sugar beet
(Beta vulgaris altissima) is a biennial root crop that provides
25% of the world’s sugar. This long-day species exhibits an
absolute requirement for vernalization, between 2 ?C and
10 ?C, depending on the genotype, whereas Arabidopsis will
ultimately flower under most environmental conditions due
to promotion by autonomous pathway genes (Koornneef
et al., 1998). In vernalized sugar beet, a rapid elongation of
the stem (bolting) associated with the use of stored sucrose
is followed by the development of an indeterminate in-
florescence. This process can be reversed at temperatures of
;25 ?C (devernalization)dependingon the genotype
(Lexander, 1980; Smit, 1983; Perarnaud et al., 2001). It
is widely accepted that sugar beet plants can bolt with-
out flowering, but they rarely flower without bolting
(Mutasa-Gottgens et al., 2009). Cold spring temperatures
and predicted global climate changes increase the risk of
exposing sugar beet crops to vernalizing temperatures
during the early part of the season (Perarnaud et al., 2001;
Porter and Semenov, 2005; Thomas et al., 2008). Bolting
resistance is an essential trait for beet crops, although it
needs to be reversible in order to enable seed production
(Reeves et al., 2007; Mutasa-Gottgens et al., 2009). In this
context, the objectives of the present study were to de-
termine (i) the variations of global DNA methylation in
relation to bolting induction in sugar beet; (ii) the genotypic
variations of global DNA methylation in relation to bolting
tolerance; and (iii) the DNA methylation marks on
vernalization genes. This study was realized on the shoot
apical meristem of biennial sugar beet genotypes submitted
to various bolting conditions.
Materials and methods
Plant material, growth conditions, and treatments
Eighteen genotypes (named G1–G18) of biennial sugar beet
hybrids (B. vulgaris altissima), provided by SESVanderHave
(Tienen, Belgium), were selected for their distinct bolting charac-
teristics. After seed germination, growth was conducted in a room
for 8 weeks at 22 ?C under a 16 h photoperiod (700 lmol m?2s?1).
Plants were submitted for various durations (0, 3, 6, 9, 12, 15, or 18
weeks) to different temperatures (22, 7, or 4 ?C) with or without
addition of 250 lM 5-azacytidine, a DNA hypomethylating drug,
or hydroxyurea, a DNA hypermethylating drug (Sigma-Aldrich,
Saint-Quentin Fallavier, France) according to previous work on
sugar beet cell lines (Causevic et al., 2005, 2006). The chemicals
were added to 8-week-old plants and the treatment was repeated
every 6 weeks for a period of 18 weeks (four treatments in total).
The effects of these methylating treatments on the shoot apical
meristem (Supplementary Fig. S1 available at JXB online) were
confirmed by high-performance liquid chromatography (HPLC)
(Supplementary Fig. S2). All treatments were performed on six
genotypes (G1–G6). Cold treatment at 4 ?C (vernalization) was
repeated on 18 genotypes (G1–G18) in an independent set of
experiments. For each genotype and treatment duration, 24 plants
were sampled as follows: shoot apical meristems of 14 plants were
immediately collected, dissected away from differentiated tissues
(Supplementary Fig. S1), and frozen in liquid nitrogen. Shoot
apical meristems (;8000 in all) were stored at –80 ?C until used.
The 10 remaining plants were placed for 6 weeks at 22 ?C under
light conditions optimal for bolting (1000 lmol m?2s?1). During
this period, two bolting parameters—that is, the bolting index (BI)
which is the percentage of bolting plants, and bolting delay (BD)
which is the average number of days required for a visible bolting
initiation—were monitored daily. Usually when bolting was
noticed, a flower developed afterwards. DBI was defined by the
difference between the BI of cold treatment at 4 ?C (vernalization)
and that of cold treatment at 4 ?C (vernalization) followed by
1 week at 26 ?C (devernalization).
Determination of DNA methylation percentages by HPLC
Genomic DNA was purified from sugar beet shoot apical
meristems using an already published RNase A digestion (com-
plete removal of RNAs), phenol/chloroform extraction, and
ethanol precipitation (Causevic et al., 2005). The global methyla-
tion percentage of genomic DNA was determined by HPLC after
2 of 13 | Trap-Gentil et al.
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 3

the enzymatic hydrolysis of DNA into nucleosides, with the use of
commercial standards (Sigma-Aldrich) (Causevic et al., 2005). In
order to test for RNA contamination, the corresponding nucleoside
standards (with distinct retention times) were also injected. Further-
more, total RNA extracts were also shown to be hypomethylated.
The methylcytosine percentages were calculated using the following
formula: %mC¼[mC/(C+mC)]3100, where ‘C’ represents 2#-deoxy-
cytidine content and ‘mC’ 5-methyl-2’-deoxycytidine content. Two
to three independent analyses and three replicates were performed
for each measurement. To calibrate the method, the methylcytosine
percentages of genomic DNA extracted from A. thaliana plantlets
vernalized (23 d at 4 ?C) or not, according to Finnegan et al. (1998),
were calculated (4.660.2% and 6.360.7%, respectively) in agree-
ment with published data on this plant.
Bioinformatics identification of sugar beet homologous genes
Using TBLASTX, two homologous sequences (best hits) to the
Arabidopsis VIN3 genes were identified in the sugar beet expressed
sequence tag (EST) databases (http://compbio.dfci.harvard.edu/tgi/
cgi-bin/tgi/Blast/index.cgi). These two sequences (BQ593505 and
BQ584308, with e-values of 9.1310?27and 1.1310?21), were
named BvVIN3 and BvVIN3-like, respectively. The BvFLC
sequence (DQ189215) corresponding to an FLC-like gene in sugar
beet has already been reported in Reeves et al. (2007). Sugar beet
sequences homologous (best hits) to a member of the transposon
CACTA family and the 5S rRNA genes in Arabidopsis were
identified in databases and named BvCACTA (RL1AR_A06) and
Bv5S (Z25804). As the sugar beet mitochondrial genome is
available (http://www.ncbi.nlm.nih.gov/nuccore/BA000009), a mi-
tochondrial gene encoding a cytochrome oxidase subunit 1
(DQ381450) and named BvCYT was selected. Mitochondrial
DNA is hypomethylated and is used as a control for DNA
methylation analysis.
The determination of transposable elements (TEs) in the BvFLC
and BvVIN3 genomic sequences was realized by the identification of
potential coding sequences (Fgenesh on http://linux1.softberry.com/
berry.phtml) followed by similarity analysis using BLASTP and
analysis of dispersed genomic repeats using PLOTREP (http://
repeats.abc.hu/cgi-bin/plotrep.pl). Computational analysis of mito-
chondrial RNA (miRNA) targets was also done on BvFLC and
BvVIN3 genomic sequences using psRNATarget (http://bioinfo3.
noble.org/psRNATarget/index.php?function¼function2).
Restriction analysis and methyl-sensitive PCR semi-quantitative
amplification
Methyl-sensitive PCR (MS-PCR) was performed using different
restriction enzymes. A 500 ng aliquot of genomic DNA (from the
shoot apical meristem of sugar beet or poplar) was digested by
overnight incubation at 37 ?C with 35 U of HpaII (Qbiogene,
Illkirch Graffenstaden, France), 50 U of MspI (Promega, Madi-
son, WI, USA), or McrBC (New England BioLabs, Ipswich, UK),
or at 25 ?C with 35 U of Sau3AI (Qbiogene). MspI and HpaII
recognize the sequence -C1C2GG-: HpaII is inhibited when a single
C or both are methylated, whereas MspI activity is blocked only
when C1is methylated, providing an evaluation of CG methylation.
McrBC recognizes DNA containing two methylated cytosine res-
idues separated by 30–2000 bp and cleaves the DNA at multiple sites
close to one of the methylated sites. Sau3AI recognizes the -GATC-
sequence but is completely blocked when C is methylated. After
ethanol precipitation and resuspension in 25 ll of ultrapure water,
digested or undigested genomic DNA was analysed by electro-
phoresis on a 0.8% ethidium bromide-stained agarose gel. Using
imaging software (ImageTool for Windows version 3.00), the
quantity of McrBC-digested genomic DNA corresponding to
hypermethylated sequences is expressed as a percentage of DNA
that migrates under the 5.0 kb region (Causevic et al., 2005).
Poplar genomic DNA that is hypomethylated was used as
a control (Gourcilleau et al., 2010). Amplification by PCR using
digested or undigested genomic DNA was done using specific
primers without an M13 tail and HotStar Taq?enzyme (Qiagen)
(95 ?C for 5 min, 25–40 cycles of 94 ?C for 30 s, annealing
temperature for 1 min, 72 ?C for 1 min, then 72 ?C for 10 min).
Sequences of primers were: BvVIN3 (5#-CCAGAATGAGCAAG-
GAGGAC-3# and 5#-TTCCCACAAGAATGCGATA-3#), BvFLC
(5#-CGGGAAAGCTTCAATAGCTG-3# and 5#-CGTTGCTAA-
CTGGTACACAAA-3#)
BvCAC
GAATC-3# and 5#-TTGTCGAGGTACAAGCGTGA-3#), Bv5S
(5#-AGGTCTGAGCGCGAAGTTAC-3# and 5#-TACTACTCT-
CGCCCAAGCAC-3#),
BvCYT
GAGGTC-3# and 5#-TGAGCCCAAACAAGAAATCC-3#), and
PopTUB (5#-AGGTGGAACTGGGTCTGGAA-3# and 5#-CAC-
TCATCAGCATTCTCAACAA-3#). The number of PCR cycles
was adjusted to avoid reaching saturation. PCR products were
analysed by electrophoresis. Two biological repeats (shoot apical
meristem) for each of the two genotypes were analysed in triplicate.
(5#-GTTTTCCACTCAGGC-
(5#-TTTTCGGTCATCCA-
Enzymatic activities and immunodetection of DNMT
The extraction and measurement of enzymatic activity or immu-
nodetection of DNA methyltransferase (DNMT, EC 2.1.37) were
performed as described previously by Causevic et al. (2005) for
sugar beet cell lines. The protein concentration was determined
using Protein Assay reagent (Bio-Rad, Marnes-la-Coquette,
France) and a standard curve established with different solutions
of bovine serum albumin (from 0 to 33 lg ll?1). Similar amounts
of sugar beet protein extracts were immunodetected using a 1:1000
dilution of antibodies raised against carrot DNA methyltransferase
1 (MET-1; Bernacchia et al., 1998; Causevic et al., 2005). A carrot
MET-1 recombinant protein, along with a molecular weight
marker (Prestained SDS–PAGE Standards, Low Range, Bio-
Rad), was used to identify the corresponding isoforms. After
detection, blots were scanned and the intensity of the immunode-
tected bands quantified in arbitrary units using imaging software
(ImageTool for Windows version 3.00). Two independent shoot
apical meristem extracts were analysed in duplicate for all
genotypes and treatments. In order to confirm that immunode-
tected bands correspond to MET-1 isoforms, an immunoprecipita-
tion experiment using protein G immobilized on Sepharose 4B
(Sigma-Aldrich) according to Causevic et al. (2005) and the
measurement of the remaining soluble DNMT enzymatic activity
was performed and showed a decrease of ;80% of the activity in
the supernatant after immunoprecipitation.
Bisulfite treatment, conversion efficiency, and sequencing
Bisulfite treatment, which results in the conversion of unmethy-
lated cytosines to uracils but does not affect methylated cytosines,
was used to determine the cytosine methylation status in CG,
CHG, or asymmetric CHH (H could be A, T, or C) contexts. An
Epitect Bisulfite Kit (Qiagen, Courtaboeuf, France) and 500 ng of
genomic DNA were used according to the manufacturer’s recom-
mendations. Two controls were performed to test the bisulfite
conversion efficiency. A Universal Methylated DNA Standard
(Zymo Research, California, USA) was used. This standard is
a DNA sequence of pUC19, which contains enzymatically meth-
ylated C in the CG context, whereas all the other Cs are
unmethylated. After sequencing, optimal bisulfite conversion
(;99%) was confirmed by substitution of all Cs at non-CG sites
by T. Then, the conversion of cytosine in sugar beet genomic DNA
extracted from shoot apical meristem was evaluated using HPLC
analysis as previously mentioned. The selection of regions for
methylation analysis was directed by the search of C-rich
sequences using a CpG island (http://www.cpgislands.com/). The
primers were designed using the method of Gruntman et al.
(2008) adapted to plant methylation and the Kismeth soft-
ware (http://katahdin.mssm.edu/kismeth; see results in Fig. 6).
Due to a possible bias using the bisulfite approach, these data were
Trap-Gentil et al. | 3 of 13
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 4

confirmed using another set of primers (data not shown) designed
according to the recommendations of Henderson et al. (2010) and
verifiedusing Net Primer
premierbiosoft.com/netprimer/netprlaunch/netprlaunch.html#). Each
primer was tailed at the 5# end with the M13 forward 5#-TGTAA-
AACGACGGCCAGT-3# or M13 reverse 5#-CAGGAAACAGC-
TATGACC-3# sequence to allow the direct sequencing in triplicate
of the PCR product without any cloning step, according to Oetting
et al. (1995), Leakey et al. (2008), and the manufacturer’s recom-
mendations (http://www3.appliedbiosystems.com/cms/groups/mcb_
marketing/documents/generaldocuments/cms_039258.pdf). Classi-
cal subcloning in the TOPO TA cloning vector (InVitrogen,
Cergy-Pontoise, France) and sequencing (10–20 clones) using an
automated DNA sequencer ABI 3100 (Applied Biosystems) using
the BigDye?Terminator v3.1 kit protocol (Applied Biosystems,
Courtaboeuf, France) was also done in parallel as previously
described (Causevic et al., 2006) and has confirmed the strategy.
When >50% of the clones still exhibited a cytosine in any position,
it was considered as hypermethylated. Sequences of primers for
BvVIN3, BvFLC, and BvCYT were, respectively: 5#-CCAGAAT-
GAGCAAGGAGGAY-3# and 5#-TTCCCACAAGAATGCGA-
TA-3#, 5#-CGGGAAAGCTTCAATAGCTG-3# and 5#-CGTTG-
CTAACTGGTACACAAA-3#, or 5#-TTTTYGGTYATYYAGA-
GGTT-3# and 5#-TRARCCCAAACAAAAAATCC-3# (Y¼C,
T and R¼G, A). PCR amplification used HotStar Taq?enzyme
(Qiagen): 95 ?C for 5 min, 40 cycles of 94 ?C for 30 s, annealing
temperature for 1 min, 72 ?C for 1 min, then 72 ?C for 10 min.
PCR products were then purified using a MiniElute PCR purifi-
cation kit (Qiagen). Methylation profile analyses were performed
on two sensitive (G7 and G8) and two resistant genotypes (G12
and G13). Genotypes G7, G8, G12, and G13 display a BI of 92,
64, 3, and 8%, and a BD of 29, 30, 32, and 36 d, respectively. A
hypomethylated genomic sequence from poplar, that is a low
methylated plant like Arabidopsis (Gourcilleau et al., 2010), encoding
PopTUB, a TUBULIN beta-2 (GRAIL3.0068008502; C. Lafon-
Placette, personal communication), confirmed the optimal bisulfite
conversion in the present conditions using both direct sequencing
and 20 subclones (<5% of methylation at each cytosine position)
using the following primers: 5#-AGGTGGAAYTGGGTYTGGA-3#
and 5#-CACTCATCARCATTCTCAACAA-3#.
software(http://www.
Semi-quantitative RT-PCR analysis
Total RNAs were isolated using Nucleospin?
(Macherey-Nagel, Hoerdt, France) and reverse transcribed using
a high capacity cDNA reverse transcription kit (Applied Bio-
systems). Two biological and three technical replicates were
performed for each sequence, genotype (G7, G8, G12, and G13),
and duration of treament. Primers for BvVIN3 and BvFLC
are identical to those designed for MS-PCR or bisulfite sequencing
but without the M13 tails. For BvVIN3-like, primers 5#-GAGCT-
GAGGTCCAGAAGCTG-3#
GAACTCCTC-3# were used for amplification. Constitutively
expressed BvTUBULIN, BvUBIQUITIN, and BvACTIN genes
were used as internal standards to normalize the amount of
mRNA in the PCR. BvTUBULIN amplification was also used to
show that there was no genomic DNA contamination since it
contains a short intron (118 bp). Furthermore, a control without
reverse transcriptase during the cDNA synthesis was performed to
confirm the absence of genomic DNA in all the total RNAs
preparations. PCR conditions and number of cycles was adjusted
to avoid reaching saturation as presented above for MS-PCR.
PCR products were revealed on an 8% polyacrylamide gel after
ethidium bromide staining and quantified with imaging software
(ImageTool for Windows version 3.00).
RNA Plant
and5#-CGACAACCCAA-
Statistical analysis
Statistical analyses were performed using the SPSS statistical
software package (SPSS version 11.0.1 PC, Chicago, IL, USA).
Means are expressed with their standard error and compared by
analysis of variance [ANOVA; general linear model (GLM)
procedure]. Data were found to meet the assumptions of homo-
scedasticity and normality distribution of residuals. Genotype
and duration of treatments effects were evaluated by two-way
ANOVA(GLMprocedure)
Yijk¼l+Gi+Tj+(Gi3Tj)+eijk; where Yijk are individual values,
l the general mean, Githe effect of the genotype j considered as
random, Tjthe effect of treatment j considered as fixed, Gi3Tjthe
genotype by duration of treatment interactions, and eijk the
standard error. Tjis noted Djwhen it corresponds to the effect of
the duration of treatment. Statistical tests were considered
significant at *P <0.05, **P <0.01, or ***P <0.001.
usingthefollowingmodel:
Results
Effect of temperature on sugar beet bolting
A prolonged cold period was necessary to observe induction
of bolting in the six biennial genotypes (named G1–G6)
of sugar beet (Fig. 1). Earlier and stronger bolting was
observed at 4 ?C than at 7 ?C (Fig. 1B, C, H, I). The
percentage of bolting plants (BI) increased with the duration
of cold treatment, while the mean number of days required
for bolting initiation (BD) decreased; negative correlations
were established between BI and BD (Table 1). When
vernalized plants were submitted to one additional week at
26 ?C (devernalization treatment), a decrease of BI was
observed mainly for G4 and G6 at the end of the treatment
(Fig. 1C, D, I, J). Genotypes G1, G2, and G3 started to
bolt after 9 weeks at 4 ?C and reached 100% BI at the end
of the treatment with a BD value under 25 d; therefore,
these genotypes were considered as sensitive (S) to bolting
compared with G4, G5, and G6 which were classified as
resistant (R).
Effect of DNA methylating treatments on sugar beet
bolting
Shoot apical meristems (Supplementary Fig. S1b at JXB
online) of plants treated with DNA hypo- and hyper-
methylating drugs showed significant variations in the ratios
of methylation measured between treated and non-treated
plants (Supplementary Fig. S2). DNA hypomethylating
treatment was not sufficient to induce any bolting in plants
cultivated at 22 ?C even after 18 weeks (Fig. 1E, K) or
longer (data not shown). In contrast, vernalized plants at
7 ?C treated with a DNA hypermethylating agent exhibited
lower BI and higher BD in the genotypes sensitive to
bolting compared with vernalized plants (Fig. 1B–F, H–L;
see G1 at 12 weeks, and G2 and G3 at 18 weeks).
A negative correlation was observed between DNA methyl-
ation of treated plants and BI (Table 1). These results imply
that artificial DNA hypermethylation affects bolting in-
duction in sugar beet genotypes, suggesting a possible role
in the control of this phenomenon.
Effect of temperature on DNA methylation
A significant temperature effect only was detected for the
cytosine methylation percentages (%mC) in shoot apical
4 of 13 | Trap-Gentil et al.
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 5

meristems (Fig. 2A). Plants grown at 22 ?C exhibited
a significant effect of the treatment duration on %mC,
with a sharp fall at 18 weeks (Fig. 2A). Significant
genotype and treatment duration effects were found for
treatment at 4 ?C with a strong DNA hypermethylation
phase at 9 weeks followed by a hypomethylation phase at
Fig. 1. Characterization of the bolting index (BI) corresponding to the percentage of bolting plants and of the bolting delay (BD)
corresponding to the average number of days required for bolting, for six sugar beet genotypes (G1, G2, and G3 were considered as
sensitive to bolting, while G4, G5, and G6 were considered resistant) after 0, 3, 6, 9, 12, 15, or 18 weeks: at 22 ?C for control treatment
(A and G), at 7 ?C (B and H), or at 4 ?C (C and I) for vernalization treatment, at 4 ?C followed by 1 week at 26 ?C (D and J) for
devernalization treatment, at 22 ?C with addition of 5-azacytidine for DNA hypomethylating treatment (E and K), and at 7 ?C with addition
of hydroxurea for DNA hypermethylating treatment (F and L). nd, not determined.
Trap-Gentil et al. | 5 of 13
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 6

15 weeks. These %mC variations in shoot apical meris-
tems of the six genotypes treated at 4 ?C were confirmed
on 18 genotypes (see open circles in Fig. 2A) and showed
that all the tested biennial genotypes exhibited DNA
hyper- and hypomethylation phases. A significant geno-
type by treatment duration interaction was observed for
%mC in shoot apical meristems of devernalized plants,
indicating that the effect of the duration of the treatment
depended on the genotype. Devernalization (one addi-
tional week at 26?C after vernalization) was associated
with a lower %mC at 0 weeks (Fig. 2A) and a later
hypermethylation phase at 12 weeks compared with
vernalization. Altogether, these data showed that global
genomic DNA methylation in the shoot apical meristem
is remodelled in response to temperature and duration of
the treatment.
Genomic DNA methylation levels were confirmed using
McrBC that cuts methylated DNA (Fig. 2B). In verna-
lized sugar beet plants, genomic DNA was sensitive to
McrBC digestion, especially at 9 weeks at 4 ?C, compared
with the hypomethylated poplar DNA (Fig. 2B). MS-PCR
analysis using McrBC restriction was performed on two
genomic loci, a CACTA transposon and a 5S rDNA gene.
The weakest intensity of bands was found for both genes
at 9 weeks, suggesting a hypermethylation state in good
agreement with the global variations of DNA methyla-
tion (Fig. 2A, C). The amplification of the hypomethylated
BvCYT (a mitochondrial gene encoding a cytochrome
oxidase subunit1) and PopTUB (a TUBULIN beta-2) were
unaffected by McrBC restriction.
The DNA methylation percentage was negatively corre-
lated with BI during vernalization and positively with DBI,
the difference between BI of vernalized plants and devernal-
ized plants (Table 1). Significant differences for %mC were
reported between resistant (R) and sensitive (S) genotypes in
vernalized plants (Fig. 3A), R genotypes being hypomethy-
lated compared to S genotypes. In devernalized plants, DNA
methylation differed between R and S genotypes only at 9
weeks (Fig. 3B). Vernalized and devernalized plants exhibited
significant differences for %mC except for S genotypes at the
end of the treatment.
Effect of temperature on DNMT activity and
accumulation of MET-1 isoforms
DNMT (EC 2.1.37) activity exhibited distinct variations
between vernalized and devernalized plants (Fig. 4A).
Immunodetection of proteins extracted from shoot apical
meristems at 15 weeks (using antibodies raised against
MET-1, a member of the DNMT family) revealed two main
bands: one at >100 kDa (the size of the recombinant MET-
1 protein), another at 90 kDa and a few smaller bands (Fig.
4B). These isoforms were shown to be associated with an
enzymaticDNMT activity
experiments (see details in the Materials and methods).
MET-1 isoforms accumulated in higher amounts in vernal-
ized than in devernalized plants, but no differences were
observed between R and S genotypes. A positive correlation
was detected between the accumulation of the 100 kDa
isoform and the DNMT activity (r¼0.71 at P <0.001).
These data show that DNMT activity and MET-1 isoforms
are affected differently in response to vernalization and
devernalization.
usingimmunoprecipitation
Relative mRNA abundance of BvFLC and BvVIN3
during vernalization
Homologues of genes from the A. thaliana vernalization
pathway were identified in sugar beet EST databases for
VIN3 (BvVIN3 and BvVIN3-like) or using the available
genomic sequence for the BvFLC gene (Reeves et al., 2007).
Relative mRNA abundance for these sequences was studied
during cold exposure at 4 ?C in four genotypes: G7, a very
sensitive (high BI) early bolting genotype (low BD); G8, a
sensitive early bolting genotype; G12, a resistant early
bolting genotype and G13, a resistant late bolting genotype.
BvFLC mRNA accumulation displayed differences between
the duration of cold and genotypes (Fig. 5A, B) for the two
detected bands (BvFLCa at 400 bp and BvFLCb at 300 bp)
but not for plants kept at 22 ?C (Supplementary Fig. S3A at
JXB online). The sensitive genotypes (G7 and G8) showed
a transient decrease in mRNA levels at 3 weeks for both
bands, while resistant genotypes showed a constant increase
Table 1. Linear correlations (Pearson’s coefficients, r) between
mean values (all genotypes and durations, n¼42) of cytosine
methylation percentage of the global genome (%mC), DNA
methyltransferase (DNMT) activity, bolting index (BI), bolting delay
(BD), DBI (difference between the BI of cold treatment at 4 ?C and
the BI of cold treatment at 4 ?C followed by 1 week at 26 ?C), and
the methylation percentage of specific cytosines in the BvVIN3
sequence analysed by bisulfite sequencing. Correlations were
computed for treatments at 4 ?C (vernalized plants), at 4 ?C
followed by 1 week at 26 ?C (devernalized plants), and at 7 ?C
with hypermethylating treatment (vernalized 7 ?C/
hypermethylated).
Variables TreatmentsBIBD DBI
Bolting delayVernalized
Devernalized
Vernalized
Devernalized
Vernalized
Devernalized
Vernalized 7 ?C/
hypermethylated
Vernalized
Devernalized
Vernalized
–0.81***
–0.60*
–0.33* DBI–0.44*
–0.56*
%mC global genome–0.51***
0.34*
–0.55*
DNMT activity
–0.39*
% methylation of specific
cytosines in the
BvVIN3 sequence
C1
C4
C10
C26
0.85*
0.94**
0.95***
–0.89*
Only significant values are indicated by an asterisk: *P <0.05;
**P <0.01; ***P <0.001.
6 of 13 | Trap-Gentil et al.
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 7

between 0 and 15 weeks; a low level of accumulation was
found in the late bolting genotype (G13). BvVIN3 displayed
an increase of mRNA accumulation between the beginning
and the end of the cold exposure (Fig. 5A), but not at 22 ?C
(Supplementary Fig. S3A), as expected, while BvVIN3-like
was unchanged (Fig. 5A). For BvVIN3, slight differences
were observed between genotypes such as a low level of
accumulation for the late bolting genotype (G13; Fig. 5B).
DNA methylation pattern of BvFLC and BvVIN3 during
vernalization
The BvFLC region analysed by bisulfite sequencing was
hypermethylated with a decrease during vernalization in S
genotypes, while only a transient decrease at 3 weeks was
observed in R genotypes (Fig. 6A and see region of
cytosines 44–55 in Fig. 6B). Computational analysis of TEs
and miRNA (see Materials and methods) in the BvFLC
gene proposed 15 potential miRNAs in the coding sequence
(data not shown) and the occurrence of two putative
retrotransposons: one similar to a Hypericum perforatum
retrotransposon protein (gb|ADK92871.1, 593 bp with an
e-value of 2310?16) at 29 000 bp before the transcription
site and another similar to a Hypomoea batatas retrotrans-
poson protein (gb|AAV88076.1, 1358 bp with an e-value of
1310?33) between exons 1 and 2. The BvVIN3 region
analysed was hypomethylated, with a progressive hyper-
methylation in S genotypes or hypomethylation in R
genotypes during vernalization (Fig. 6A). Computational
analysis revealed no TEs or miRNA for BvVIN3 (data not
shown). The region of cytosines 1–10 was the most poly-
morphic while cytosine 26 and region 33–37 displayed
a similar hypomethylation during vernalization in both R
and S genotypes (Fig. 6B). Significant correlations were
established between the methylation levels of the BvVIN3
cytosines C1, C10 and the BI, or between C4, C26 meth-
ylation levels and the BD (Table 1). bisulfite conversion
efficiency was confirmed by sequencing known hypomethy-
lated sequences such as the mitochondrial BvCYT and
PopTUB (Fig. 6B). MS-PCR also showed the hypermethy-
lation state of BvFLC compared with BvVIN3 using
McrBC and Sau3AI enzymes (Fig. 6B). The use of the
Fig. 2. Variations in the global DNA methylation rate in the shoot
apical meristem of sugar beet genotypes during different temper-
ature treatments. (A) Each bar corresponds to the mean of the six
genotypes with their standard errors (n¼24) for global genomic
DNA methylation. DNA methylation percentage measured with
the HPLC method in the shoot apical meristem after 0, 3, 6, 9,
12, 15, and 18 weeks at 22 ?C for control treatment (white bars),
at 4 ?C (black bars) for vernalization treatment, and at 4 ?C
followed by 1 week at 26 ?C (grey bars) for devernalization
treatment. nd, not determined. Open circles correspond to the
mean of the 18 sugar beet genotypes with their standard errors
(n¼24) for global genomic DNA methylation. For each graph, G
indicates the genotype effect (G1–G6), T the treatment effect
(22 ?C, 4 ?C, and 4 ?C with 1 week at 26 ?C), D the treatment
duration effect (0, 3, 6, 9, 12, 15, and 18 weeks), and (G3T) or
(G3D) the genotype by treatment or treatment duration inter-
actions, respectively, determined by two-way ANOVA for the six
sugar beet genotypes (see Materials and methods). Significant
differences between control plants (22 ?C) and vernalized or
devernalized plants are indicated by an asterisk (at P <0.05). (B)
Global DNA methylation was also measured with the McrBC
restriction method (see Materials and methods) on non-digested
(control) DNA or McrBC-digested genomic DNA from the shoot
apical meristem after 0, 9, or 18 weeks at 4 ?C and DNA from the
poplar shoot apical meristem grown at 22 ?C used as a control
(hypomethylated). An image of a typical agarose gel is shown
(top) with the corresponding quantification (bottom) of the
digested fraction. (C) MS-PCR (see Materials and methods for
details) using the same digestions to determine the methylation
status of specific loci for BvCAC, a transposon of the CACTA
family, Bv5S, an rRNA gene, BvCYT, a mitochondrial cytochrome
oxidase subunit 1, and PopTUB encoding a tubulin gene in
poplar as a control.
Trap-Gentil et al. | 7 of 13
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 8

isoschizomers MspI and HpaII confirmed the changes
reported using bisulfite for the methylation on BvVIN3
cytosine 22 between genotypes and/or treatment durations.
MS-PCR performed on plants treated at 22 ?C (Supplemen-
tary Fig. S3B at JXB online) showed no variation in the
methylation status of BvFLC and BvVIN3 during the
duration of the treatment, in contrast to cold-treated plants
(Fig. 6B).
Discussion
Global DNA methylation kinetics and induction of sugar
beet bolting
DNA hypomethylation was not sufficient to promote
bolting in biennial sugar beet, which is consistent with what
was observed in Xanthium strumarium and Lemna paucicos-
tata (Kondo et al., 2007) but opposite to what was obtained
in Arabidopsis, wheat, and flax (Burn et al., 1993; Brock and
Davidson, 1994; Finnegan et al., 1998; Fieldes et al., 2005).
DNA hypermethylation treatment was shown for the first
time to decrease the percentage of bolting plants and
increase the number of days required for bolting in sugar
beet. In agreement with this, DNA hypermethylating treat-
ments that could induce gene silencing have already been
shown to induce hypermethylation in sugar beet cell lines
with a similar range of values and to inhibit morphogenesis
with a limited impact on cell growth (Causevic et al., 2005,
2006). DNA hypomethylation was not sufficient to induce
de novo organogenesis.
Ageing of the shoot apical meristem at 22 ?C was
associated with a late and limited DNA hypomethylation
and no bolting. The effect of ageing on DNA methylation has
already been reported, but with contradictory trends (Fraga
et al., 2002; Sha et al., 2005). In all the tested genotypes of
sugar beet, DNA methylation in the shoot apical meristem
exhibited dynamic and reversible hyper- and hypomethylation
Fig. 3. Variations in global DNA methylation during
vernalization and devernalization in genotypes resistant to bolting
(white bars) and sensitive to bolting (black bars). Each bar
corresponds to the mean with their standard errors (n¼12) of the
three resistant genotypes (G1, G2, and G3) or the three
sensitive genotypes (G4, G5, and G6) for DNA methylation
percentage in the shoot apical meristem after 0, 9, and
18 weeks at 4 ?C (A) for vernalization treatment or at 4 ?C followed
by 1 week at 26 ?C (B) for devernalization treatment.
Significant differences (at P <0.05) between resistant and sensitive
genotypes for a given treatment are indicated by an asterisk, while
significant differences (at P <0.05) between treatments for a given
group of genotypes (resistant or sensitive) are indicated by a hash
sign in B.
Fig. 4. Variations in total DNMT activity (A) and MET-1 isoforms
(B) during vernalization (4 ?C; white bars) and devernalization (4 ?C
and 26 ?C; black bars) in the shoot apical meristem of sugar
beet genotypes after 0, 3, 6, 9, 12, 15, and 18 weeks.
Significant differences (at P <0.05) between vernalized and
devernalized mean values of DNMT activity are indicated by an
asterisk. Western blots with MET-1 antibodies for proteins
extracted from the sugar beet shoot apical meristem of
genotypes sensitive (S) or resistant (R) to bolting after 15 weeks at
4 ?C for vernalization (V4?C) and 15 weeks at 4 ?C plus 1 week at
26 ?C for devernalization (V4?C/DV26?C) and a MET-1
recombinant protein (RP). The sizes of detected bands are
indicated in kDa.
8 of 13 | Trap-Gentil et al.
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 9

changes during cold treatment. In Arabidopsis, vernalization
was only reported to cause a decrease in DNA methylation in
seedlings (Finnegan et al., 1998; Genger et al., 2003), but no
time course study on meristematic cells is available. The
dynamics of DNA methylation have already been reported
and is a way to actively reprogram the epigenome in both
plants and animals (Feng et al., 2010; Zhang et al., 2010).
Furthermore, the amplitude and time course of changes
in DNA methylation, as well as bolting induction, were
dependent on temperature, treatment duration, and sugar
beet genotype. Indeed, significant correlations were es-
tablished between bolting induction (BI and DBI) and
DNA methylation in vernalized, vernalized/devernalized, and
vernalized/hypermethylated plants. In these last two types of
plants, the increase or delay of the hypermethylation phase
was followed by a decrease of bolting induction. A decrease
in bolting after devernalization has already been reported
in sugar beet (Perarnaud et al., 2001) but the present data
showed, for the first time, that this effect is associated with
variations in DNA methylation kinetics, DNMT activity,
and MET-1 isoform accumulation. Epigenetic patterns
participate in the control of gene expression (Zhang et al.,
2006; Zilberman et al., 2007) and are transmitted by mitosis.
Thus, changes that arise in dividing cells in the shoot apical
meristem during environmental stress could be propagated to
daughter cells and explain the cell memory of vernalization
(Lang, 1965; Metzger, 1988; Boyko and Kovalchuck, 2008).
The kinetics of DNA methylation could indicate successive
determination steps of the shoot apical meristem state
leading to bolting.
The range of values in the shoot apical meristem cells
obtained for global DNA methylation is consistent with
those reported for similar organs in Pinus radiata (Fraga
et al., 2002). The high amplitude of DNA methylation
variations was confirmed by several approaches and showed
that loci such as CACTA transposons or 5S rRNA genes in
sugar beet undergo methylation with kinetics similar to
those of the genome as a whole. In Arabidopsis, CACTA
elements were shown to transpose and increase in copy
number in hypomethylated ddm1 mutants (Miura et al.,
2001), confirming the role of DNA methylation in genome
protection. 5S rDNA units are differentially affected by
hyper- and hypomethylation participating in their differen-
tial chromatin compaction and expression during develop-
ment and in response to environmental variations in
Arabidopsis (Vaillant et al., 2008).
The variations in DNA methylation could arise from the
enzymatic activity of chromatin-modifying enzymes. Several
DNMT bands were assessed in sugar beet. This is certainly
due to a process of limited proteolysis of the MET-1 protein
(>100 kDa) as already shown in several plant DNMTs
(Bernacchia et al., 1998; Causevic et al., 2005; Gourcilleau
et al., 2010). DNMT activity was correlated with the
amounts of the 100 kDa MET-1 isoform, as already
reported (Causevic et al., 2005; Gourcilleau et al., 2010),
but not with the global DNA methylation kinetics. In oil
palm, DNA hypomethylation is not associated with de-
creased expression of genes encoding DNMTs (Rival et al.,
2008), suggesting the involvement of reverse enzyme activ-
ities, a distinct mitotic index, or post-translational modifi-
cations. Removal of cytosine methylation patterns can be
realized by incorporation of unmethylated cytosine during
replication or by active demethylation catalysed by the
four bifunctional helix–hairpin–helix glycosylases and AP
lyases ROS1 (REPRESSOR OF SILENCING 1), DML2
Fig. 5. mRNA relative abundance during vernalization at 4 ?C of
sugar beet homologous genes (BvFLC, BvVIN3, and BvVIN3-
like) of the Arabidopsis vernalization pathway. RT-PCR analyses
were performed with reverse-transcribed total RNA isolated from
the shoot apical meristem of plants treated at 4 ?C for 0, 3, or 15
weeks. BvTUBULIN, BvUBIQUITIN, and BvACTIN were used as
internal controls. Results are shown for a very sensitive early
bolting (G7), a sensitive early bolting (G8), a resistant early bolting
(G12) and a resistant late bolting genotype (G13). (A) Images of
electrophoretic analysis of mRNA relative abundance. Technical
controls are presented in the
Materials and methods. (B) mRNA relative abundance of BvFLC
a and b transcripts and BvVIN3 is represented by a bar with its
corresponding standard error (n¼2 biological33 technical repli-
cates). The mRNA relative abundance of BvVIN3-like, being
constitutive and stable, is not detailed. Quantifications were
performed with Image Tool software and gene values were
corrected by the expression of BvTUBULIN, BvUBIQUITIN, and
BvACTIN.
Trap-Gentil et al. | 9 of 13
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 10

Fig. 6. Analysis of cytosine methylation status using bisulfite sequencing for BvFLC and BvVIN3 in the shoot apical meristem of
genotypes sensitive (G7) and resistant (G12) to bolting treated at 4 ?C for 0, 3, or 15 weeks. (A) The proportion of methylcytosine for
BvFLC and BvVIN3 during vernalization treatment is indicated with white bars for resistant and black bars for sensitive genotypes. (B)
Detailed methylation profiles for BvFLC (399 bp), BvVIN3 (241 bp), BvCYT (157 bp), a mitochondrial cytochrome oxidase subunit 1, and
PopTUB (177 bp) encoding a tubulin gene in poplar used as controls for bisulfite conversion. CG, CHG, and asymmetric CHH (H¼A, T,
or C) sites are shown by circles, triangles, and squares, respectively. According to their methylation percentage (%mC), cytosine sites are
considered as hypermethylated (%mC >50%, filled circles, triangles, or squares), or hypomethylated (%mC <50%, open circles, triangles,
or squares). Arrows and black frames show cytosine sites and domains, respectively, with methylation polymorphism or with restriction
sites that are discussed in the text. Images correspond to examples of methyl-sensitive PCR (MS-PCR) analyses for each gene. Genomic
DNA was digested or not with enzymes (C). McrBC cuts in hypermethylated regions while Sau3AI does not digest DNA if the recognition
site is methylated. MspI and HpaII recognize the sequence -C1C2GG-: HpaII is inhibited when a single C or both are methylated,
whereas MspI activity is blocked only when C1is methylated, providing an evaluation of CG methylation.
10 of 13 | Trap-Gentil et al.
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 11

(DEMETER-LIKE 2), and DML3 (DEMETER-LIKE 3)
in vegetative tissues, and DME (DEMETER) in endosperm
(Feng et al., 2010). The dynamics of DNA methylation/
demethylation are certainly important for the epigenomic
plasticity that could participate in the plant response to
environmental variations in a timely manner such as for the
vernalization process (Zhang et al., 2010). Future works will
have to clarify the relative importance of methylation and
active demethylation at key stages of the vernalization
process in sugar beet.
DNA methylation as a component of genotypic
tolerance to bolting
A genotype effect was detected for DNA methylation in
vernalized plants. Furthermore, significant quantitative dif-
ferences in DNA methylation levels were reported between
R and S genotypes. These latter displayed the highest hyper-
methylated phase (acquisition of competence) in agreement
with the highest number of bolting plants. In addition,
devernalization delayed the hypermethylation phase partic-
ularly on R genotypes, explaining the significant genotype
by treatment effect observed for global DNA methylation
variations. This fits well with the bolting decrease observed
in R genotypes at 18 weeks during devernalization. The
sugar beet bolting tolerance could involve fine quantitative
differences during bolting in the global kinetics of DNA
methylation in relation to the relative importance of methy-
lation and active demethylation as previously mentioned.
The analysis of epigenetic recombinant inbred lines (epi-
RILs) in Arabidopsis with distinct DNA methylation marks
showed high heritability for flowering time and inheritance
over at least eight generations (Johannes et al., 2009),
encouraging the identification of causative epigenetic quan-
titative trait loci in plants.
Vernalization genes are targeted by DNA methylation
during bolting
Two differential mRNA splicing transcripts were detected
for BvFLC (Reeves et al., 2007). BvFLC undergoes a cold-
induced down-regulation which is reversed in leaves at
normal temperatures, while no variation is observed in the
shoot apical meristem (Reeves et al., 2007). In contrast,
AtFLC expression in Arabidopsis is irreversibly inhibited by
vernalization in both leaves and shoot apical meristems
(Sheldon et al., 2000; Searle et al., 2006). In agreement with
this, it was found here that both transcripts were still
expressed after long exposure to cold in the sugar beet shoot
apical meristem. In addition, the present data showed
differences of expression between S and R genotypes. The
early transient repression of BvFLC which is a flowering
repressor (Reeves et al., 2007) in S genotypes found only
during cold exposure could explain their sensitivity to
bolting. The BvFLC locus undergoes different modifications
of DNA methylation between genotypes and during vernal-
ization. This is in contrast to AtFLC in which only one
methylated cytosine was found in both vernalized and
control plants, offering no explanation for the changes in
expression profiles (Finnegan et al., 2005).
BvVIN3 expression was shown to increase only during cold
exposure in the same way as AtVIN3 (Sung and Amasino,
2004). BvVIN3 methylation profiles exhibited differences
between genotypes and in relation to the duration of the cold
exposure, while AtVIN3 expression is not regulated by DNA
methylation in Arabidopsis mutants for DNA methyltransfer-
ase (Finnegan et al., 2005). Furthermore, DNA methylation
variations of both BvFLC and BvVIN3 did not follow the
kinetics of methylation of the global genome as did BvCAC
and Bv5S. BvFLC was highly methylated, particularly in non-
CG contexts, compared with BvVIN3. Because bisulfite
sequencing could contain bias (incomplete unmethylated
cytosine conversion and primer hybridization bias), several
technical controls were carried out (see details in the Materials
and methods) to confirm the methylation data (Universal
Methylated DNA Standard and HPLC analysis for cytosine
conversion, distinct sets of primers, unmethylated control
sequences, and MS-PCR for amplification bias). In addition,
the hypermethylation of BvFLC compared with BvVIN3 was
in good agreement with the finding of TEs and potential
miRNAs in BvFLC only (Zhang et al., 2006; Feng et al.,
2010). Finally, the methylation states of specific cytosines in
BvVIN3 were correlated with bolting variables. Altogether,
these data lead to the proposal that vernalization genes are
targeted by DNA methylation in a genotype-dependent way
only after cold exposure. In sugar beet, differences between
genotypes that are resistant or sensitive to bolting should
lead to the extensive characterization of loci affected by
DNA methylation during vernalization and devernalization.
This will clarify the role of DNA methylation and will open
up perspectives for the use of epialleles in marker-assisted
selection strategies for plant breeding (Gentil and Maury,
2007).
Supplementary data
Supplementary data are available at JXB online.
Figure S1. Plant material before and after bolting. (A) An
8-week-old sugar beet plantlet before temperature treat-
ment. (B) Longitudinal section of a 26-week-old sugar beet
shoot apex. The plant material collected during first sam-
pling is delimited by a dark dotted line and corresponds to
the shoot apical meristem. (C) A 32-week-old plant after
optimal bolting conditions. An elongating stem is indicated
by white arrows.
Figure S2. Effects of DNA hypomethylating and hyper-
methylating treatments (see Materials and methods) on the
global genomic DNA cytosine methylation percentage
(%mC) of six sugar beet genotypes (G1–G6) sensitive or
resistant to bolting. The methylation ratio corresponds to
the ratio between %mC values between treated and control
plants. The corresponding standard errors, for each geno-
type and treatment, are indicated. The dotted line indicate
a theoretical ratio value equal to 1 that corresponds to
identical %mC values between treated and control plants.
Trap-Gentil et al. | 11 of 13
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 12

Figure S3. (A) mRNA relative abundance during vernaliza-
tion at 22 ?C of sugar beet homologous genes (BvFLC and
BvVIN3) of the Arabidopsis vernalization pathway. RT-PCR
analyses were performed with reverse-transcribed total RNA
isolated from the shoot apical meristem of plants treated at
22 ?C for 0, 3, or 15 weeks. BvTUBULIN and BvUBIQUITIN
were used as internal controls. Results are shown for a very
sensitive early bolting genotype (G7). (B) Pictures correspond
to examples of methyl-sensitive PCR (MS-PCR) analyses for
BvFLC and BvVIN3. Genomic DNA extracted from sensi-
tive (G7) or resistant (G12) genotypes to bolting exposed at
22 ?C for 0, 3, or 15 weeks. DNA was digested with McrBC
or not (C) and amplified by semi-quantitative PCR.
Acknowledgements
Two PhD grants (M-VT-G and CH) were supported by
SES-VanderHave, Belgium and the ‘Conseil Re ´gional de la
Re ´gion Centre’, France or the ‘Ministe `re de l’Enseignement
Supe ´rieur et de la Recherche’ and ANRT (convention
CIFRE no. 0379/2008), France, respectively. The authors
thank Gilles Moreau, Carole Latruwe, Adisa Causevic,
Mourad Kamiri, Jean-Paul Charpentier, and Marie-Claude
Lesage-Descauses for technical assistance. The authors are
grateful to Rino Cella for providing MET-1 antibodies and
recombinant protein. We are grateful to Pierre Devaux for
the careful reading of the manuscript.
References
Bastow R, Mylne JS, Lister C, Lippman Z, Martienssen RS,
Dean C. 2004. Vernalization requires epigenetic silencing of FLC by
histone methylation. Nature 427, 164–167.
Bernacchia G, Primo A, Giorgetti L, Pitto L, Cella R. 1998. Carrot
DNA-methyltransferase is encoded by two classes of genes with
differing patterns of expression. The Plant Journal 13, 317–329.
Boyko A, Kovalchuck I. 2008. Epigenetic control of plant stress
response. Environmental and Molecular Mutagenesis 49, 61–72.
Brock R, Davidson J. 1994. 5-Azacytidine and gamma rays partially
substitute for cold treatment in vernalizing winter wheat. Environmental
and Experimental Botany 34, 195–199.
Burn JE, Bagnall DJ, Metzger JD, Dennis ES, Peacock WJ.
1993. DNA methylation, vernalization, and the initiation of flowering.
Proceedings of the National Academy of Sciences, USA 90, 287–291.
Causevic A, Delaunay A, Ounnar S, Righezza M, Delmotte F,
Brignolas F, Hage `ge D, Maury S. 2005. DNA methylating and
demethylating treatments modify phenotype and cell wall
differentiation state in sugarbeet cell lines. Plant Physiology and
Biochemistry 43, 681–691.
Causevic A, Gentil M-V, Delaunay A, El-Soud WA, Garcia Z,
Pannetier C, Brignolas F, Hage `ge D, Maury S. 2006. Relationship
between DNA methylation and histone acetylation levels, cell redox
and cell differentiation states in sugarbeet lines. Planta 224, 812–827.
Chouard P. 1960. Vernalization and its relations to dormancy. Annual
Review of Plant Physiology 11, 191–238.
Dennis ES, Peacock WJ. 2007. Epigenetic regulation of flowering.
Current Opinion in Plant Biology 10, 520–527.
Feng S, Jacobsen SE, Reik W. 2010. Epigenetic reprogramming in
plant and animal development. Science 330, 622–627.
Fieldes MA, Schaeffer SM, Krech MJ, Brown JCL. 2005. DNA
hypomethylation in 5- azacytidine-induced early-flowering lines of flax.
Theoretical and Applied Genetics 111, 136–149.
Finnegan EJ, Genger RK, Kovac K, Peacock WJ, Dennis ES.
1998. DNA methylation and the promotion of flowering by
vernalization. Proceedings of the National Academy of Sciences, USA
95, 5824–5829.
Finnegan EJ, Kovac KA, Jaligot E, Sheldon CC, Peacock WJ,
Dennis ES. 2005. The downregulation of FLOWERING LOCUS C
(FLC) expression in plants with low levels of DNA methylation and by
vernalization occurs by distinct mechanisms. The Plant Journal 44,
420–432.
Fraga MF, Rodriguez R, Canal MJ. 2002. Genomic DNA
methylation–demethylation during aging and reinvigoration of Pinus
radiata. Tree Physiology 22, 813–816.
Genger RK, Peacock WJ, Dennis ES, Finnegan EJ. 2003.
Opposing effects of reduced DNA methylation on flowering time in
Arabidopsis thaliana. Planta 216, 461–466.
Gentil M-V, Maury S. 2007. Characterization of epigenetic
biomarkers using new molecular approaches. In: Varshney R,
Tuberosa R, eds. Genomic assisted crop improvement, Vol. 1. Berlin:
Springer, 351–370.
Gourcilleau D, Bogeat-Triboulot M-B, Le Thiec D, Lafon-
Placette C, Delaunay A, El-Soud WA, Brignolas F, Maury S.
2010. DNA methylation and histone acetylation: genetic variations in
hybrid poplars, impact of water deficit and relationships with
productivity. Annals of Forest Science 67, 208 1–10.
Gruntman E, Qi Y, Slotkin RK, Roeder T, Martienssen RA,
Sachidanandam R. 2008. Kismeth: analyzer of plant methylation
states through bisulfite sequencing. BioMed Central Bioinformatics 9,
371–384.
Henderson IR, Chan SR, Cao X, Johnson L, Jacobsen SE. 2010.
Accurate sodium bisulfite sequencing in plants. Epigenetics 5, 1–3.
Johannes F, Porcher E, Teixeira FK, et al. 2009. Assessing the
impact of transgenerational epigenetic variation on complex traits.
PLoS Genetics 5, e1000530.
Kondo H, Miura T, Wada KC, Takeno K. 2007. Induction of
flowering by 5-azacytidine in some plant species: relationship between
the stability of photoperiodically induced flowering and flower-inducing
effect of DNA demethylation. Physiologia Plantarum 131, 462–469.
Koornneef M, Alonso-Blanco C, Peeters AJ, Soppe W. 1998.
Genetic control of flowering time in Arabidopsis. Annual Review of
Plant Physiology and Plant Molecular Biology 49, 345–370.
Lang A. 1965. Physiology of flower initiation. In: Lang A, ed.
Encyclopedia of plant physiology, Vol. XV/1. Berlin: Springer Verlag,
1380–1536.
Leakey TI, Zielinski J, Siegfried RN, Siegel ER, Fan CY,
Cooney CA. 2008. A simple algorithm for quantifying DNA methylation
levels on multiple independent CpG sites in bisulfite genomic
sequencing electropherograms. Nucleic Acids Research 36, e64.
12 of 13 | Trap-Gentil et al.
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from

Page 13

Lexander K. 1980. Present knowledge of sugar bolting mechanisms.
43rd Winter Congress. Bruxelles: Institut International de Recherches
Betteravie `res, 245–258.
Metzger JD. 1988. Localization of the site of perception of
thermoinductive temperatures in Thlaspi arvense L. Plant Physiology
88, 424–428.
Miura A, Yonebayashi S, Watanabe K, Toyama T, Shimada H,
Kakutani T. 2001. Mobilization of transposons by a mutation
abolishing full DNA methylation in Arabidopsis. Nature 411, 212–214.
Mutasa-Gottgens E, Qi A, Mathews A, Thomas S, Phillips A,
Hedden P. 2009. Modification of gibberellin signalling (metabolism &
signal transduction) in sugar beet: analysis of potential targets for crop
improvement. Transgenic Research 18, 301–308.
Oetting WS, Lee HK, Flanders DJ, Wiesner GL, Sellers TA,
King RA. 1995. Linkage analysis with multiplexed short tandem
repeat polymorphisms using infrared fluorescence and M13 tailed
primers. Genomics 30, 450–458.
Perarnaud V, Souverain F, Prats S, Dequiedt B, Fauche `re J,
Richard-Molard M. 2001. Influence du climat sur le phe ´nome `ne de
monte ´e a ` graine de la betterave: synthe `se. ITB-Me ´te ´o-France, http://
www.institut-betterave.asso.fr.
Porter JR, Semenov MA. 2005. Crop responses to climatic variation.
Philosophical Transactions of the Royal Society B: Biological Sciences
360, 2021–2035.
Purvis ON, Gregory FG. 1952. Studies in vernalization of cereals. XII.
The reversibility by high temperature of the vernalized condition in
Petkus winter rye. Annals of Botany 16, 1–21.
Reeves PA, He Y, Schmitz RJ, Amasino RM, Panella LW,
Richards CM. 2007. Evolutionary conservation of the FLOWERING
LOCUS C-mediated vernalization response: evidence from the
sugarbeet (Beta vulgaris). Genetics 176, 295–307.
Rival A, Jaligot E, Beule ´ T, Finnegan EJ. 2008. Isolation and
expression analysis of genes encoding MET, CMT and DRM
methyltransferases in oil palm (Elaesis Guineensis Jacq.) in relation to
the ‘mantled’ somaclonal variation. Journal of Experimental Botany 59,
3271–3281.
Schmitz RJ, Amasino RM. 2007. Vernalization: a model for
investigating epigenetics and eukaryotic gene regulation in plants.
Biochimica et Biophysica Acta 1769, 269–275.
Searle I, He Y, Turck F, Vincent C, Fornara F, Kro ¨ber S,
Amasino RA, Coupland G. 2006. The transcription factor FLC
confers a flowering response to vernalization by repressing meristem
competence and systemic signaling in Arabidopsis. Genes and
Development 20, 898–912.
Sha AH, Lin XH, Huang JB, Zhang DP. 2005. Analysis of DNA
methylation related to rice adult plant resistance to bacterial blight
based on methylation-sensitive AFLP (MSAP) analysis. Molecular
Genetics and Genomics 273, 484–490.
Sheldon CC, Rouse DT, Finnegan EJ, Peacock WJ, Dennis ES.
2000. The molecular basis of vernalization: the central role of
FLOWERING LOCUS C (FLC). Proceedings of the National Academy
of Sciences, USA 97, 3753–3758.
Smit A. 1983. Influence of external factors on growth and
development of sugar-beet (Beta vulgaris L.). Agricultural Research
Reports, PUDOC.
Sung S, Amasino RM. 2004. Vernalization and epigenetics:
how plants remember winter. Current Opinion in Plant Biology
7, 4–10.
Thomas B, Adams S, Collier R, Fellows J, Jenner C,
Jaggard KW, Qi A, Semenov MA, Wossink A. 2008. AC0301:
vulnerability of UK agriculture to extreme events. Interim report, UK
Department for Environment Food and Rural Affairs (DEFRA), 12.
http://www.defra.gov.uk/science/project_data/documentlibrary/
AC0301/AC0301_6900_IR.doc.
Vaillant I, Tutois S, Jasencakova Z, Douet J, Schubert I,
Tourmente S. 2008. Hypomethylation and hypermethylation of the
tandem repetitive 5S rRNA genes in Arabidopsis. The Plant Journal
54, 299–309.
Zhang M, Kimatu JN, Xu K, Liu B. 2010. DNA cytosine methylation
in plant development. Journal of Genetics and Genomics 37,
1–12.
Zhang X, Yazaki J, Sundaresan A, et al. 2006. Genome-wide
high-resolution mapping and functional analysis of DNA methylation
in Arabidopsis. Cell 126, 1189–1201.
Zilberman D, Gehring M, Tran RK, Ballinger T, Henikoff S. 2007.
Genome-wide analysis of Arabidopsis thaliana DNA methylation
uncovers an interdependence between methylation and transcription.
Nature Genetics 39, 61–69.
Trap-Gentil et al. | 13 of 13
by guest on January 13, 2011
jxb.oxfordjournals.org
Downloaded from