Prevention of early flowering by expression of FLOWERING LOCUS C requires methylation of histone H3 K36

Abstract

Flowering represents a crucial transition from a vegetative to a reproductive phase of the plant life cycle. Despite extensive studies, the molecular mechanisms controlling flowering remain elusive. Although the enzymes involved are unknown, methylation of histone H3 K9 and K27 correlates with repression of FLOWERING LOCUS C (FLC), an essential transcriptional repressor involved in flowering time control in Arabidopsis thaliana; in contrast, methylation of H3K4 correlates with FLC activation. Here we show that loss-of-function of SET DOMAIN GROUP 8 (SDG 8), which encodes a homologue of the yeast SET2 histone methyltransferase, results in reduced dimethylation of histone H3K36, particularly in chromatin associated with the FLC promoter and the first intron, regions that contain essential cis-elements for transcription. sdg8 mutants display reduced FLC expression and flower early, establishing SDG8-mediated H3K36 methylation as a novel epigenetic memory code required for FLC expression in preventing early flowering. This is the first demonstrated role of H3K36 methylation in eukaryote development.

Full text

LETTERS
1256 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 12 | DECEMBER 2005
Prevention of early flowering by expression of
FLOWERING LOCUS C requires methylation
of histone H3 K36
Zhong Zhao1, Yu Yu1, Denise Meyer1, Chengjun Wu1,2 and Wen-Hui Shen1,3
Flowering represents a crucial transition from a vegetative
to a reproductive phase of the plant life cycle. Despite
extensive studies, the molecular mechanisms controlling
flowering remain elusive. Although the enzymes involved are
unknown, methylation of histone H3 K9 and K27 correlates
with repression of FLOWERING LOCUS C (FLC), an essential
transcriptional repressor involved in flowering time control
in Arabidopsis thaliana; in contrast, methylation of H3K4
correlates with FLC activation1–4. Here we show that loss-of-
function of SET DOMAIN GROUP 8 (SDG 8), which encodes a
homologue of the yeast SET2 histone methyltransferase, results
in reduced dimethylation of histone H3K36, particularly in
chromatin associated with the FLC promoter and the first intron,
regions that contain essential cis-elements for transcription.
sdg8 mutants display reduced FLC expression and flower
early, establishing SDG8-mediated H3K36 methylation as a
novel epigenetic memory code required for FLC expression in
preventing early flowering. This is the first demonstrated role of
H3K36 methylation in eukaryote development.
In eukaryotes, genomic DNA is wrapped around histone octomers
(composed of two molecules each of histones H2A, H2B, H3 and H4),
resulting in highly packaged chromatin in the nucleus5. Current models
suggest that changes in chromatin structure positively or negatively effect
the accessibility of DNA to transcription factors, resulting in temporal
and spatial variation in gene expression. Histones, particularly their tails,
are subject to multiple types of covalent modification, including acetyla-
tion, methylation, phosphorylation and ubiquitination. Combinations
of such modifications have been proposed to serve as a ‘histone code’,
specifying a chromatin state that can determine the transcriptional activ-
ity of the embedded gene6.
To date, four lysines within histone H3 (K4, K9, K27 and K36) and
one lysine within histone H4 (K20) have been identified as being meth-
ylated by SET (Su(var)3-9, E(z) and Trithorax-conserved)-domain
1Institut de Biologie Moléculaire des Plantes (IBMP), Centre National de la Recherche Scientifique (CNRS), Université Louis Pasteur de Strasbourg (ULP), 12 rue
du Général Zimmer, 67084 Strasbourg Cédex, France. 2Present address: Biotechnology Research Institute of Yunnan Academy of Agriculture Sciences, Kunming
650223, Yunnan, People’s Republic of China.
3Correspondence should be addressed to W.-H. S. (e-mail: Wen-Hui.Shen@ibmp-ulp.u-strasbg.fr)
Received 27 May 2005; accepted 28 September 2005; published online: 20 November 2005; DOI: 10.1038/ncb1329
histone-lysine-methyltransferases (HKMTs), and one lysine within H3
(K79) as being methylated by a non-SET-domain HKMT7. Compared
to the well-established function of methylation at other lysines, little
is known about H3K36 methylation. H3K36 di- and tri-methylation is
enriched in actively transcribed genomic regions in yeast8,9 and meta-
zoa10, but the Saccharomyces cerevisiae set2 mutant lacking H3K36
methylation is viable and shows no marked growth defects11. The
mouse nuclear receptor binding protein, NSD1, which is involved in
acute myeloid leukaemia, was shown to methylate H3K36, as well as
H4K20, in vitro12. However, the in vivo function of H3K36 methylation
by NSD1 has not been established.
The Arabidopsis genome encodes 39 SET-domain proteins — named
SDG1 to SDG39 in the Plant Chromatin Database (www.chromdb.org).
SDG8, a 1759-amino-acid protein encoded by a gene (At1g77300) con-
taining 15 exons (Fig. 1a), shows the highest homology with SET2, the
sole H3K36 HKMT of S. cerevisiae11. The sequence homology between
the two proteins is limited to the region spanning the SET domain
and its surrounding cysteine-rich AWS (Associated With SET) and C
(Cysteine-rich) domains (see Supplementary Information, Fig. S1). The
plant protein also contains two motifs that are conserved within the
RPB1 subunits of RNA polymerase II, a CW (cysteine and tryptophan
conserved) domain that is also present in some other chromatin fac-
tors13, and a long amino-terminal extension (Fig. 1b).
To characterize the function of SDG8, we identified two mutants
(sdg8-1 and sdg8-2) from the Salk collection of Arabidopsis T-DNA inser-
tion lines14. Polymerase chain reaction (PCR) analysis confirmed that
sdg8-1 and sdg8-2 contain a T-DNA inserted in exon 7 and 2, respectively
(Fig. 1a). Both the sdg8-1 and sdg8-2 mutations are recessive and only
homozygous (hereafter called mutant) plants had a phenotype. At early
stages, the sdg8-1 and sdg8-2 mutants grew similarly to wild-type plants
(Fig. 1c). However, both sdg8-1 and sdg8-2 mutants flowered earlier,
slowing down rosette growth (Fig. 1d). Flowers of the mutant plants were
smaller (Fig. 1e) but contained morphologically normal flower organs
(four sepals, four petals, five–six stamens and two fused carpels per
ncb1329.indd 1256ncb1329.indd 1256 1/12/05 3:26:33 pm1/12/05 3:26:33 pm
Nature Publishing Group
© 2005
© 2005 Nature Publishing Group

NATURE CELL BIOLOGY VOLUME 7 | NUMBER 12 | DECEMBER 2005 1257
LETTERS
flower). Reverse transcriptase (RT)–PCR analysis revealed that SDG8
transcripts were not detectable in either the sdg8-1 or the sdg8-2 mutant
(Fig. 1f). More recently, a third mutant allele (sdg8-3) containing a T-
DNA inserted in intron 14 (Fig. 1a) was obtained from the Syngenta
collection, and this mutant showed a phenotype similar to sdg8-1 and
sdg8-2. Furthermore, a genomic fragment containing the SDG8 coding
region together with the upstream 3,023 bp and the downstream 308
bp non-coding regions could rescue both sdg8-1 and sdg8-2 mutants,
resulting in plants with a wild-type phenotype (Table 1). Together, these
data establish that SDG8 prevents early flowering.
Flowering represents the transition from vegetative to reproductive
development, a shift that is temporally controlled by developmental state
(for example, size and age) and regulated by environmental cues, such as
light and temperature1. Similarly to wild-type plants, all three (sdg8-1,
sdg8-2 and sdg8-3) mutants flowered earlier under inductive long-day
photoperiods than under short-day photoperiods (Table 1), indicating
that SDG8 is dispensable for the photoperiod response. Nevertheless,
when compared to wild-type plants, mutant plants flowered earlier
under all studied photoperiods. The strongest lag occurred under short-
day photoperiods, indicating a positive effect of light quantity on early
flowering. Exposure to a prolonged period of cold (vernalization) pro-
moted flowering of the wild-type plants but had no significant effect on
the sdg8-1 and sdg8-2 mutants (Table 1). Expression of SDG8 in wild-type
plants, however, was unchanged by vernalization (see Supplementary
Information, Fig. S2a). This latter result suggests that SDG8 does not
act as an initiator of the vernalization response pathway.
To investigate the molecular mechanism underlying early flowering of
the mutant plants, the expression of floral pathway genes was analysed.
Consistent with their normal photoperiod-responsive phenotype, the
mutants showed a normal expression of the CONSTANS ( CO) gene (data
not shown), a key regulator specific of the photoperiod pathway1. The
vernalization pathway and the autonomous (photoperiod-independent;
thought to be regulated in response to developmental cues) pathways
converge on FLC1, which encodes a MADS-box transcription factor. FLC
represses flowering in a dosage-dependent manner by negatively regulat-
ing the expression of several genes that promote flowering; for example,
FLOWERING LOCUS T (FT) and SUPPRESSOR OF OVEREXPRESSION
OF CO 1 (SOC1) (refs 15–18). In both sdg8-1 and sdg8-2 mutant plants,
FLC expression was markedly repressed whereas the downstream genes
FT and SOC1 were activated (Fig. 2a).
a
b
f
ce
d
sdg8-2
T-DNA
sdg8-1
T-DNA
sdg8-3
T-DNA
8651 bp
CW AWS SET C
1759 aa
RPB1
Col
Col
sdg8-1
Col sdg8-1 sdg8-2
Col
sdg8-1
sdg8-1
SDG8
ACTIN
Figure 1 Loss of SDG8 function causes early flowering. (a) Exon–intron
structure of the SDG8 gene. Boxes represent exons; lines represent introns;
triangles indicate T-DNA insertions. The T-DNA was located in the second
exon in sdg8-2, in the seventh exon in sdg8-1 and in the fourteenth intron
in sdg8-3. (b) Schematic diagram of the SDG8 protein. The CW, AWS, SET
and C boxes represent conserved domains. Two regions with homology to the
consensus sequence of the RPB1 motif are presented in sequence alignment.
(c) Wild-type Col and mutant sdg8-1 plants 29 days after germination (DAG).
(d) Wild-type Col and mutant sdg8-1 plants 46 DAG. (e) Wild-type Col and
mutant sdg8-1 flowers. (f) RT–PCR analysis of SDG8 expression in wild-type
Col and mutant sdg8-1 and sdg8-2 plants 20 DAG. ACTIN serves as an
internal control. Scale bars represent 10 mm in c and d, and 1 mm in e.
ncb1329.indd 1257ncb1329.indd 1257 1/12/05 3:26:38 pm1/12/05 3:26:38 pm
Nature Publishing Group
© 2005
© 2005 Nature Publishing Group

1258 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 12 | DECEMBER 2005
LETTERS
The methylation level of different histone H3 lysine residues in mutant
and wild-type plants was analysed by western blotting with specific anti-
bodies2 to investigate whether SDG8 regulated FLC expression through
histone methylation. Although the levels of di-methylated H3K9 and
H3K27 were unchanged, both sdg8-1 and sdg8-2 mutants contained a
reduced level of di-methylated H3K36 (Fig. 2b), indicating that SDG8
encodes a major enzyme controlling H3K36 methylation in Arabidopsis.
A significant increase in di-methylated H3K4 was also observed in the
mutants (Fig. 2b), suggesting that decreased H3K36 methylation favours
H3K4 methylation. This finding is contrary to the previously described
cross-talk among histone modifications19.
We next examined histone H3 methylation at FLC by chromatin
immunoprecipitation (CHIP). Similar sets of PCR primers to those pre-
viously reported2 were used to cover the promoter (region A), the first
exon (region B), the first intron (regions C, D, E and F) and some down-
stream exons and introns (region G) of FLC (Fig. 3a). Consistent with a
previous report on non-vernalized plants2, FLC chromatin was found to
be rich in acetylated H3 and di-methylated H3K4, whereas di-methyl-
ated H3K9 and H3K27 were undetectable in wild-type plants (Fig. 3b).
Control experiments demonstrated that the actively expressed ACTIN20,21
was embedded in chromatin that was rich in acetylated H3 and di-meth-
ylated H3K4, whereas the silenced transposon Ta3 was primarily found in
chromatin rich in di-methylated H3K9 and H3K27. Furthermore our data
show that H3K36 di-methylation was strongly associated with FLC and
weakly with ACTIN, but was undetectable with Ta3 chromatin (Fig. 3b).
Compared with wild-type plants, sdg8-1 plants showed a marked reduc-
tion in H3K36 di-methylation in regions A, D, E and F, and a slightly
decreased level of di-methylated H3K36 in region B. There were slightly
increased levels of di-methylated H3K4 and actetylated H3 in region
B, but no changes in other regions or in H3K9 and H3K27 di-methyla-
tion in sdg8-1 plants (Fig. 3b). These results were reproducible in three
independent experiments. Although only a slight increase in di-methyl-
ated H3K4 was detected in region B of the FLC locus (Fig. 3b), a more
pronounced increase of di-methylated H3K4 was observed with whole
nuclear extracts of sdg8 mutants (Fig. 2b), indicating that sdg8 does not
uniformly affect H3K4 di-methylation over the entire genome.
The regions showing the most pronounced changes of H3K36 di-
methylation by sdg8-1 are located within the promoter and the first
intron of FLC, which correspond to regions with essential cis-acting
functions22. These regions have also been reported to be regulated
by H3K9 and H3K27 di-methylation2,4, H3K4 di- and tri-methyla-
tion2,3, and histone acetylation20,23. Interestingly, whereas the H3K36
di-methylation of FLC region B changed little in sdg8-1 mutants, this
region was highly responsive to H3K9 and H3K27 di-methylation in
vernalization2,4, and to H3K4 tri-methylation in early flowering elf7
and elf8 mutants3. In contrast, the regions D, E and F were markedly
reduced for H3K36 di-methylation in sdg8-1 but H3K9 di-methylation
was persistently undetectable in these regions upon vernalization2. Thus,
it seems t hat SDG8 targets H3K36 methylation in specific regions of FLC
chromatin that are distinct from the regions regulated by H3K4, H3K9
and H3K27 methylation in flowering time control.
Taken together, our genetic, physiological and molecular data indi-
cate that SDG8 methylates H3K36 of FLC chromatin, which positively
regulates FLC transcription to prevent early flowering. To test whether
ectopic ex pression of FLC could inhibit early flowering in an sdg8 mutant
background, sdg8-1 and sdg8-2 mutants, and wild-type plants were
transformed with a 35S:FLC construct that expresses FLC cDNA under
the control of the constitutive 35S promoter. 35S:FLC transformation
resulted in transgenic plants with different flowering times, including
early, unchanged and late, in either wild-type Col or mutants sdg8-1
and sdg8-2 backgrounds (see Supplementary Information, Table S1), as
previously reported for Ler and C24 ecotypes15,16. Specifically, 6 of 14
Col, 4 of 26 sdg8-1 and 6 of 42 sdg8-2 transgenic lines showed a very late
flowering phenotype (more than 18 rosette leaves under long-day pho-
toperiods). RT–PCR analysis demonstrated that the late-flowering phe-
notype correlates with high levels of FLC transcripts (see Supplementary
Information, Fig. S2b). Together, these data show that overexpression
of FLC alone is sufficient to cause a significant delay in flowering time
in sdg8 mutants and in wild-type plants. Nevertheless, 35S:FLC was
expressed at lower levels in the mutant than in the wild-type plants and
correlates with the reduced delay in flowering. This suggests that sdg8
also affects 35S:FLC expression.
Although the enzymes involved in H3K4, H3K9 and H3K27 methyla-
tion of FLC chromatin are not currently known, there is high sequence
homology of the ELF7 and ELF8 mutants to subunits of the yeast PAF1
transcriptional complex3. This suggests that, as in yeast, a SET1-like
HKMT may be recruited by the PAF1 complex during H3K4 methylation
of FLC chromatin. In yeast, the PAF1 complex also interacts with SET2,
consistent with an additional role in H3K36 methylation8,9. Functional
interaction between SDG8 and VERNALIZATION INDEPENDENCE
4 (VIP4; which encodes a homologue of LEO1, a subunit of the PAF1
complex8) was genetically tested. The vip4 mutant has been previously
shown to have reduced FLC expression and exhibit early flowering24,25.
The double mutant vip4–sdg8-1 flowered earlier than either the vip4 or
Table 1 Rosette leaf number at fl owering of wild-type (Col) and mutant plants in different growth conditions
SD–V MD–V LD–V SD + V MD + V LD + V
Col 63.0 ± 3.5 (12) 15.7 ± 0.3 (15) 11.1 ± 0.2 (18) 33.5 ± 0.6 (15) 12.8 ± 0.2 (15) 10.8 ± 0.3 (15)a
sdg8-1 22.5 ± 4.2 (12) 8.5 ± 0.2 (20) 8.0 ± 0.1 (21) 21.9 ± 1.0 (15) 8.8 ± 0.2 (15) 8.1 ± 0.2 (15)a
sdg8-2 23.4 ± 3.8 (12) 9.1 ± 0.2 (25) 7.9 ± 0.1 (21) 21.7 ± 0.7 (15) 9.2 ± 0.2 (15) 8.2 ± 0.2 (15)a
sdg8-3 21.4 ± 2.3 (20) 8.9 ± 0.5 (17) 8.1 ± 0.4 (20) n.t. n.t. n.t.
sdg8-1 pCSDG8b65.0 ± 4.4 (9) n.t. 11.0 ± 0.3 (11) n.t. n.t. n.t.
sdg8-2 pCSDG8b63.7 ± 3.0 (11) n.t. 11.3 ± 0.2 (27) n.t. n.t. n.t.
vip4–sdg8-1 18.3 ± 0.8 (15) 7.3 ± 0.4 (16) 6.0 ± 0.3 (15) n.t. n.t. n.t.
vip4 25.1 ± 1.2 (15) 12.1 ± 0.3 (24) 9.8 ± 0.4 (15) n.t. n.t. n.t.
Plants without (–) and with (+) vernalization (V) were tested in different photoperiods: long day (LD) was 16 h light and 8 h dark, medium day (MD) was 12 h light and 12 h dark, and short day
(SD) was 8 h light and 16 h dark. Values shown are mean number of rosette leaves ± standard deviation. Numbers in parentheses are the total numbers of plants evaluated. n.t., not tested. aThe
differences between wild-type and mutants are smallest at LD + V among all tested growth conditions; nevertheless these differences are highly signifi cant (P = 0.0001) according to a t-test.
bFrom 33 and 41 transformants, 11 and 13 independent transgenic lines (showing wild-type phenotype), were obtained for sdg8-1 pCSDG8 and sdg8-2 pCSDG8, respectively. The statistical
data shown were for progeny of a homozygous line of sdg8-1 pCSDG8 or sdg8-2 pCSDG8.
ncb1329.indd 1258ncb1329.indd 1258 1/12/05 3:26:40 pm1/12/05 3:26:40 pm
Nature Publishing Group
© 2005
© 2005 Nature Publishing Group

NATURE CELL BIOLOGY VOLUME 7 | NUMBER 12 | DECEMBER 2005 1259
LETTERS
sdg8-1 single mutants (Table 1), sug gesting that H3K4 and H3K36 meth-
ylation function synergistically to regulate FLC expression. Interestingly,
sdg8 mutants flowered earlier than the vip4 mutant (Table 1), indicating
that SDG8 does not function in an exclusively VIP4-dependent manner
and that additional VIP4-independent pathways exist in SDG8-mediated
flowering-time control.
RT–PCR analysis demonstrated that SDG8 is ubiquitously expressed
in different organs of Arabidopsis plants (data not shown). However, in
situ hybridization showed that SDG8 transcripts were more abundant
in young tissues but were distributed in a non-uniform manner in clus-
ters of cells (Fig. 4), implying that SDG8 expression is highly regulated.
The presence of abundant SDG8 transcripts in the inflorescence (ovules
as shown in Fig. 4c), where FLC expression is undetectable15, further
suggests that SDG8 may have additional roles that are independent of
FLC activation. Further exploration will define these additional roles of
SDG8 in plant growth and development and determine whether H3K36
methylation also has a critical function in other eukaryotes.
METHODS
Plant materials and growth conditions. All Arabidopsis alleles were derived
from the Columbia (Col) ecotype. sdg8-1, sdg8-2 and sdg8-3 alleles corre-
spond to Salk_065480, Salk_026442 and SAIL_880_B03, respectively, of the
T-DNA insertion strains from the Arabidopsis Biological Resource Center
(ABRC) and Nottingham Arabidopsis Stock Centre (NASC) (http://arabi-
dopsis.org). The vip4 mutant used in this study corresponds to Salk_122755,
which contains a T-DNA inserted in exon 4 of the VIP4 (At5g61150) gene.
Seeds were obtained from ABRC and NASC. The studied photoperiods were
8 h light and 16 h dark for short-day, 12 h light and 12 h dark for medium-
day, and 16 h light and 8 h dark for long-day. For vernalization tests, four-
day-old seedlings on MS media were maintained under short-day conditions
at 4 °C for 40 days then transferred to soil and grown under long-day at
22 °C. Flowering time was measured from a developmental perspective, as
the total number of rosette leaves15.
RT –PCR analysis. Total RNA was extracted with TRI Reagent (Molecular Research
Center Inc., Cincinnati, OH) from seedlings grown under medium-day conditions.
Contaminating DNA was eliminated by treatment with RNase-free RQ1-DNase
(Promega, Charbonnières, France). RT–PCR was performed as described17, using
gene specific primers (see Supplementary Information, Table S2).
Histone extraction and western-blot analysis. Histone-enriched protein extrac-
tion from 20-day-old seedlings and western blot analysis were performed as
described26. Upstate (Euromedex, Mundolshein, France) antibodies against his-
tone H3 dimethyl-Lys 36 (Catalogue no. 07-369), H3 dimethyl-Lys 4 (Catalogue
no. 07-030), H3 dimethyl-Lys 9 (Catalogue no. 07-212), H3 dimethyl-Lys 27
(Catalogue no. 07-322) and H3 (Catalogue no. 05-499) were used.
Chromatin immunoprecipitation. 20-day-old seedlings grown under medium-
day conditions were used for chromatin immunoprecipitation experiments as
previously des cribed21. Anti-acetyl-Histone H3-Lys 9 (Upstate Catalogue no. 06-
942) was used to detect acetylation. PCR detection of FLC regions2, ACTIN and
Ta3 (Ref. 21) was performed as described.
In sit u hybridization. Digoxigenin labelling of RNA probes, tissue preparation
and in situ hybridization were performed as described27. Tissue sections were
8 µm thick. A plasmid containing the 260 bp 3′-untranslated region (UTR) of
SDG8 was used to prepare sense (negative control) and antisense probes.
Gene constructs and plant transformation. A 12 kb Nhe I restriction frag-
ment containing the SDG8 gene together with the upstream 3023 bp and
downstream 308 bp non-coding regions was cloned from the BAC clone
H3K36m >
H3K27m >
H3 >
H3K4m >
H3K9m >
FLC
FT
SOC1
A
CTIN
sdg8-1
sdg8-1
sdg8-2
Col
Col
sdg8-1
Col
sdg8-1
Col
sdg8-1
Col
DAG 4 8 12 16 20 24
ab
Figure 2 Effects of sdg8 on gene expression and histone H3 methylation.
(a) RT–PCR analysis of floral-pathway gene expression. Wild-type Col and
mutant sdg8-1 plants grown under 12-h light and 12-h dark photoperiods
were analysed for FLC, SOC1 and FT expression at 4, 8, 12, 16, 20 and 24
DAG. ACTIN served as an internal control. (b) Western blot analysis of lysine
methylation of histone H3. Histone-enriched protein extracts from 20-day-
old wild-type Col, and sdg8-1 and sdg8-2 mutant plants grown under a 12-h
light and 12-h dark photoperiod were probed with antibodies that specifically
recognize the indicated forms of histone H3.
a
b
PromoterExon 1 Exon 7
1 kb
Intron 1
AB C D E F G
Col
sdg8-1
Col
sdg8-1
Col
sdg8-1
Col
sdg8-1
Col
sdg8-1
Col
sdg8-1
Col
sdg8-1
Input H3K36mH3K4mH3K9mH3K27mH3K9a
−
FLC
A
B
C
D
E
F
G
ACTIN
Ta3
Figure 3 ChIP analysis of histone H3 modifications at the FLC locus.
(a) A schematic representation of FLC gene structure indicating the regions
examined by ChIP. Black boxes represent exons; lines represent promoter
and introns; bars labelled A–G represent regions amplified by PCR.
(b) ChIP analysis with antibodies against modified histone H3. Twenty-day-
old wild-type Col and mutant sdg8-1 plants grown under 12-h light and 12-h
dark photoperiods were analysed with antibodies that specifically recognize
modified histone H3, as indicated at the top of each panel. Input and no
antibody (–) are shown as positive and negative controls, respectively. FLC
regions and controls (ACTIN for an actively expressed gene and Ta3 for a
repressed transposon) were amplified by PCR.
ncb1329.indd 1259ncb1329.indd 1259 1/12/05 3:26:44 pm1/12/05 3:26:44 pm
Nature Publishing Group
© 2005
© 2005 Nature Publishing Group

1260 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 12 | DECEMBER 2005
LETTERS
T14N5 (ABRC) into pCambia1300 (CAMBIATM, Canberra, NSW) resulting
in the binary vector pCSDG8. FLC cDNA containing the 88 bp 5′-UTR, the
591 bp open reading frame (ORF) and the 238 bp 3′-UTR was amplified by
RT–PCR and cloned behind the CaMV 35S promoter in pBinHyg-TX28, result-
ing in the binary vector p35S:FLC. The binary vectors were introduced into
Agrobacterium tumefaciens and the resulting strains were used to transform
Arabidopsis plants29.
Note: Supplementary Information is available on the Nature Cell Biology website.
ACKNOWLEDGEMENTS
The authors thank the Nottingham Arabidopsis Stock Center and the Arabidopsis
Biological Resource Center for seeds and BAC clone, and K. Richards and
H. Huang for helpful comments on the manuscript. Z.Z. is supported by a
postdoctoral fellowship from the French Ministère de la Recherche et des Nouvelles
Technologies, Y.Y. by a research training fellowship from the Association Franco-
Chinoise pour la Recherche Scientifique & Technique (AFCRST), and C.W. by a
four-month-visiting fellowship from the United Nations Educational, Scientific and
Cultural Organization (UNESCO). Research in the W.-H.S. laboratory is supported
by the Centre National de la Recherche Scientifique (CNRS).
COMPETING INTERESTS STATEMENT
The authors declare that they have no competing financial interests.
Published online at http://www.nature.com/naturecellbiology
Reprints and permissions information is available online at http://npg.nature.com/
reprintsandpermissions
1. Boss, P. K., Bastow, R. M., Mylne, J. S. & Dean, C. Multiple pathways in the decision to
flower: enabling, promoting, and resetting. Plant Cell 16, S18–S31 (2004).
2. Bastow, R. et al. Vernalization requires epigenetic silencing of FLC by histone methyla-
tion. Nature 427, 164–167 (2004).
3. He, Y., Doyle, M. R. & Amasino, R. M. PAF1-complex-mediated histone methylation of
FLOWERING LOCUS C chromatin is required for the vernalization-responsive, winter-
annual habit in Arabidopsis. Genes Dev. 18, 2774–2784 (2004).
4. Sung, S. & Amasino, R. M. Vernalization in Arabidopsis thaliana is mediated by the PHD
finger protein VIN3. Nature 427, 159–164 (2004).
5. Luger, K., et al. Crystal structure of the nucleosome core particle at 2.8 A resolution.
Nature 389, 251–260 (1997).
6. Strahl, B. D. & Allis, C. D. The language of covalent histone modifications. Nature 403,
41–45 (2000).
7. Sims III, R. J., Nishioka, K. & Reinberg, D. Histone lysine methylation: a signature for
chromatin function. Trends Genet. 19, 629–639 (2003).
8. Krogan, N. J. et al. Methylation of histone H3 by Set2 in Saccharomyces cerevisiae
is linked to transcriptional elongation by RNA polymerase II. Mol. Cell. Biol. 23,
4207–4218 (2003).
9. Xiao, T. et al. Phosphorylation of RNA polymerase II CTD regulates H3 methylation in
yeast. Genes Dev. 17, 654–663 (2003).
10. Bannister, A. J. et al. Spatial distribution of di- and tri-methyl lysine 36 of histone H3
at active genes. J. Biol. Chem. 280, 17732–17736.
11. Strahl, B. D. et al. Set2 is a nucleosomal histone H3-selective methyltransferase that
mediates transcriptional repression. Mol. Cell. Biol. 22, 1298–1306 (2002).
12. Rayasam, G. V. et al. NSD1 is essential for early post-implantation development and
has a catalytically active SET domain. EMBO J. 22, 3153–3163 (2003).
13. Perry, J. & Zhao, Y. The CW domain, a structural module shared amongst vertebrates, verte-
brate-infecting parasites and higher plants. Trends Biochem. Sci. 28, 576–580 (2003).
14. Alonso, J. M. et al. Genome-wide insertional mutagenesis of Arabidopsis thaliana.
Science 301, 653–657 (2003).
15. Michaels, S. D. & Amasino, R. M. FLOWERING LOCUS C encodes a novel MADS domain
protein that acts as a repressor of flowering. Plant Cell 11, 949–956 (1999).
16. Sheldon, C. C. et al. The FLF MADS box gene: a repressor of flowering in Arabidopsis
regulated by vernalization and methylation. Plant Cell 11, 445–458 (1999).
17. Lee et al. The AGAMOUS-LIKE 20 MADS domain protein integrates floral inductive
pathways in Arabidopsis. Genes Dev. 14, 2366–2376 (2000).
18. Samach, A. et al. Distinct roles of CONSTANS target genes in reproductive development
of Arabidopsis. Science 288, 1613–1616 (2000).
19. Fischle, W., Wang, Y. & Allis, C. D. Histone and chromatin cross-talk. Curr. Opin. Cell
Biol. 15, 172–183 (2003).
20. He, Y., Michaels, S. D. & Amasino, R. M. Regulation of flowering time by histone
acetylation in Arabidopsis. Science 302, 1751–1754 (2003).
21. Johnson, L. M., Cao, X. & Jacobsen, S. E. Interplay between two epigenetic marks: DNA
methylation and histone H3 lysine 9 methylation. Curr. Biol. 12, 1360–1367.
22. Sheldon, C. C., Conn, A. B., Dennis, E. S. & Peacock, W. J. Different regulatory regions
are required for the vernalization-induced repression of FLOWERING LOCUS C and for
the epigenetic maintenance of repression. Plant Cell 14, 2527–2537 (2002).
23. Ausin, I., et al. Regulation of flowering time by FVE, a retinoblastoma-associated
protein. Nature Genet. 36, 162–166 (2004).
24. Zhang, H. & van Nocker, S. The VERNALIZATION INDEPENDENCE 4 gene encodes a
novel regulator of FLOWERING LOCUS C. Plant J. 31, 663–673 (2002).
25. Oh, S., Zhang, H., Ludwig, P. & van Nocker, S. A mechanism related to the yeast
transcriptional regulator Paf1c is required for expression of the Arabidopsis FLC/MAF
MADS box gene family. Plant Cell 16, 2940–2953 (2004).
26. Yu, Y., Dong, A. & Shen, W.-H. Structure-function characterization of the tobacco SET-
domain protein NtSET1 reveals its role in histone methylation, chromatin binding and
segregation. Plant J. 40, 699–711 (2004).
27. Jackson, D. in Molecular Plant Pathology: A Practical Approach (eds Bowles, D. J., Gurr,
S. J. & McPherson, M.) 163–174 (Oxford University Press, Oxford, UK 1991).
28. Gatz, C. Novel inducible/repressible gene expression systems. Methods Cell Biol. 50,
411–424 (1995).
29. Clough, S. J. & Bent, A. F. Floral dip: a simplified method for Agrobacterium-mediated
transformation of Arabidopsis thaliana. Plant J. 16, 735–743 (1998).
a
b
c
SAM
L2
ads
abs
ms
ov ov
ov
ov
Placenta
ov
gp
ms
L3
L1
Figure 4 In situ hybridization analysis of SDG8 expression. (a) A longitudinal
section through a vegetative apex. The shoot apical meristem (SAM) and
leaves (L1 to L3) are indicated. (b) A longitudinal section through a floral
apex. The adaxial sepal (ads), the medial stamen (ms), the gynoecial
primordium (gp) and the abaxial sepal (abs) are indicated. (c) A longitudinal
section through part of a gynoecium. The ovule primodium (ov) and the
placenta are indicated. Hybridization signals are dark brown areas. A negative
control using a sense probe did not generate detectable hybridization signals
(data not shown). Wild-type Col plants grown under a 12-h light and 12-h
dark photoperiod were analysed. Scale bars represent 10 µm in all panels.
ncb1329.indd 1260ncb1329.indd 1260 1/12/05 3:26:48 pm1/12/05 3:26:48 pm
Nature Publishing Group
© 2005
© 2005 Nature Publishing Group

NATURE CELL BIOLOGY ADVANCE ONLINE PUBLICATION 1
ERRATUM
Owing to a technical error, the pages of this manuscript were origi-
nally mis-numbered by a 100 pages.  is has now been corrected on-
line.  e corrected online manuscript is numbered 100 pages higher
than the mis-numbered version.
© 2005 Nature Publishing Group

SUPPLEMENTARY INFORMATION
WWW.NATURE.COM/NATURECELLBIOLOGY 1
Figure S1 Alignment of SDG8 and SET2 amino acid sequences.
© 2005 Nature Publishing Group
© 2005 Nature Publishing Group

SUPPLEMENTARY INFORMATION
2 WWW.NATURE.COM/NATURECELLBIOLOGY
Figure S2 Effect of vernalization on SDG8 and FLC expression and correlation of late flowering with 35S:FLC overexpression.
© 2005 Nature Publishing Group
© 2005 Nature Publishing Group

SUPPLEMENTARY INFORMATION
WWW.NATURE.COM/NATURECELLBIOLOGY 3
Table S1 Flowering time of 35S:FLC transgenic T1 wild-type (Col) and mutant (sdg8-1 and sdg8-2) plants at long-day photoperiods
Phenotype No. of rosette leaves No. of plants Mean rosette leaf number per plant
Col control
93
10.1 ± 0.9 (10)
10 3
11 4
Col
35S:FLC
early 8 2 8.0 ± 0 (2)
unchanged
10 4
10.5 ± 0.8 (6)
11 1
12 1
very late
41 1
69.8 ± 17.2 (6)
57 1
76 1
80 2
85 1
sdg8-1 control
72
8.1 ± 0.7 (12)
87
93
sdg8-1
35S:FLC
early 51 5.8 ± 0.4 (6)
65
unchanged
74
7.8 ± 0.8 (9)
83
92
late 10 2
11 1
13 3 12.0 ± 1.5 (7)
14 1
23 1
very late 25 1 35.0 ± 12.1 (4)
39 1
53 1
sdg8-2 control
72
8.0 ± 0.6 (12)
88
92
sdg8-2
35S:FLC
early 51 5.9 ± 0.3 (8)
67
unchanged
77
8.1 ± 0.9 (17)
82
98
late
10 5
11.5 ± 1.4 (11)
11 1
12 3
13 2
14 2
18 1
26 1
very late 40 2 38.3 ± 13.9 (6)
44 1
62 1
© 2005 Nature Publishing Group
© 2005 Nature Publishing Group

SUPPLEMENTARY INFORMATION
4 WWW.NATURE.COM/NATURECELLBIOLOGY
Table S2 Gene-specifi c primers used in RT–PCR analysis
Gene Forward Primer Reverse Primer Corresponding region (nucleotide after stop codon = +1)
SDG8 AAACAGAGTGTTTCCTCCATGG TAAAGAAAGGAGAGGGATAGG –101 to +163
FLC AAGCTGAGATGGAGATGTC TCGATGCAATTCTCACACG –73 to +238
SOC1 CGAGCAAGAAAGACTCAAGTGTTTAAGG GAAGTGACTGAGAGAGAGAGAGTGAG –230 to +149
FT TACGAAAATCCAAGTCCCACTG AAACTCGCGAGTGTTGAAGTTC –206 to -87
ACTIN AAGTCATAACCATCGGAGCTG ACCAGATAAGACAAGACACAC –392 to +21
© 2005 Nature Publishing Group
© 2005 Nature Publishing Group