Downloaded from genesdev.cshlp.

org on December 11, 2014 - Published by Cold Spring Harbor Laboratory Press

RESEARCH COMMUNICATION

natural variation in vernalization (Nordborg and Bergelson

Multiple FLC haplotypes 1999; Shindo et al. 2006; Alonso-Blanco and Mendez-Vigo
defined by independent cis- 2014). Relatively rapid vernalization is thought to facili-
tate adaptation to habitats with high summer drought or
regulatory variation underpin other high mortality situations, whereas slow vernaliza-
life history diversity tion is thought to be an important adaptation to northern
latitudes (Stinchcombe et al. 2004). In some cases, this
in Arabidopsis thaliana variation has been mapped to the FLC locus itself (Coustham
et al. 2012), but the extent to which FLC variation accounts
Peijin Li,1 Daniele Filiault,2 Mathew S. Box,1,7 for natural variation in vernalization response and the
Envel Kerdaffrec,2 Cock van Oosterhout,3 importance of this natural variation to adaptation re-
Amity M. Wilczek,4 Johanna Schmitt,5 main important unanswered questions (Satake 2010). We
exploited the extensive collection of genotyped A. thaliana
Mark McMullan,3,8 Joy Bergelson,6 accessions to characterize the genetic architecture of the
Magnus Nordborg,2 and Caroline Dean1 FLC locus in worldwide accessions. Here, we describe the
1 phenotypic variation conferred by the different FLC haplo-
John Innes Centre, Norwich NR4 7UH, United Kingdom;
2 types, their geographical distribution, and the extent to
Gregor Mendel Institute, Austrian Academy of Sciences, 1030
which variation at FLC accounts for the phenotype. These
Vienna, Austria; 3Department of Environmental Sciences,
data reveal that the extensive allelic heterogeneity at this
University of East Anglia, Norwich NR4 7TJ, United Kingdom;
4 single floral repressor gene can account for a major fraction
Deep Spring College, Big Pine, California 93513, USA;
5 of the natural variation in vernalization rate that underpins
University of California at Davis, Davis, California 95616,
life history diversity in A. thaliana.
USA; 6Department of Ecology and Evolution, University
of Chicago, Chicago, Illinois 60637, USA
Results and Discussion
Relating molecular variation to phenotypic diversity is Single-nucleotide polymorphism (SNP) data from the
a central goal in evolutionary biology. In Arabidopsis Regional Mapping panel project (Horton et al. 2012)
thaliana, FLOWERING LOCUS C (FLC) is a major de- identified 20 FLC haplotypes across the 1307 accessions
terminant of variation in vernalization—the acceleration (Fig. 1A), with five major haplotypes predominating
of flowering by prolonged cold. Here, through analysis of (Hap1, Hap3, Hap5, Hap11, and Hap13), numbered accord-
1307 A. thaliana accessions, we identify five predomi- ing to sequence similarity. A Northern Swedish acces-
nant FLC haplotypes defined by noncoding sequence sion, Lov-1, that we studied previously fell into Hap9,
variation. Genetic and transgenic experiments show that which showed intermediate frequency (Coustham et al.
2012). In the worldwide samples, the FLC region is
they are functionally distinct, varying in FLC expression
characterized by high population structure and extensive
level and rate of epigenetic silencing. Allelic heteroge- haplotype sharing (Fig. 1B). Hap1, Hap3, and Hap13 have
neity at this single locus accounts for a large proportion broad geographical distributions with centroids in central
of natural variation in vernalization that contributes to Europe, while Hap5 is predominantly in the United
adaptation of A. thaliana. Kingdom, and Hap9 and Hap11 tend to be near coasts
(Fig. 1C–J).
Supplemental material is available for this article.
We selected 47 A. thaliana accessions that represented
Received May 22, 2014; revised version accepted July 1, the major haplotypes and generated a high-quality FLC
2014. sequence for each accession (Supplemental Table S1).
Remarkably, the predicted amino acid sequences from
these FLC alleles were identical, with very low poly-
morphism in the coding sequences but considerable
Reproductive timing is central to plant adaptation through nucleotide sequence polymorphism in noncoding regions
its major effect on fitness (Mitchell-Olds and Schmitt (Fig. 2A,B). The functional significance of this noncoding
2006). Many plants overwinter before flowering, thus polymorphism was assessed through analysis of the
aligning the transition to reproductive development with vernalization response of the 47 accessions, using a partial
spring. In Arabidopsis thaliana, this process, called vernalization treatment to enhance phenotypic diversity.
vernalization, involves expression and then subsequent The vernalization response—degree of reactivation of
epigenetic silencing of a floral repressor gene named FLC expression after 4 wk of cold—varied among acces-
FLOWERING LOCUS C (FLC) (Sheldon et al. 2000; sions and grouped with haplotype (Fig. 2). Two haplotypes
Michaels 2009). A. thaliana accessions show extensive (Hap3 and Hap13) showed relatively rapid vernalization
(RV) response and so were named RV1 and RV2. Four
[Keywords: allelic heterogeneity; FLOWERING LOCUS C; noncoding others (Hap9, Hap1, Hap5, and Hap11) showed a relatively
polymorphism; vernalization; adaptation] slow vernalization (SV) response and so were named SV1–4
Present addresses: 7Sainsbury Laboratory Cambridge, University of (Fig. 2B,C; Supplemental Fig. S1A,B). The reactivation of
Cambridge, Cambridge CB2 1LB, UK; 8The Genome Analysis Centre,
Norwich NR4 7UH, UK.
Corresponding author: [email protected]
Article published online ahead of print. Article and publication date are Ó 2014 Li et al. This article, published in Genes & Development, is
online at http://www.genesdev.org/cgi/doi/10.1101/gad.245993.114. Freely available under a Creative Commons License (Attribution 4.0 Interna-
available online through the Genes & Development Open Access option. tional), as described at http://creativecommons.org/licenses/by/4.0.

GENES & DEVELOPMENT 28:1635–1640 Published by Cold Spring Harbor Laboratory Press; ISSN 0890-9369/14; www.genesdev.org 1635
 Downloaded from genesdev.cshlp.org on December 11, 2014 - Published by Cold Spring Harbor Laboratory Press

Li et al.

locus (QTL) analyses (Supplemental Table S3; Alonso-

Blanco et al. 1998; Shindo et al. 2006; O’Neill et al. 2008;
Li et al. 2010; Salome et al. 2011). Strong-effect QTL for
vernalization response has been mapped over the FLC
genomic region in populations generated from parents
with different haplotypes: Col-0 (RV2) crossed to Lov-1
(SV1), Ull2–5 (SV2), or Var2–6 (SV4) (Shindo et al. 2006).
To further test the functional diversity of FLC haplotypes,
we undertook additional QTL analyses using F2 popula-
tions generated from crosses between accessions contain-
ing FLC alleles within and between haplotypes. In all
crosses between the RV 3 SV accessions, a QTL was
detected over the FLC genomic region, reinforcing the
view that the haplotypes confer functionally distinct
responses. No FLC QTL was detected in the cross be-
tween the RV1 3 RV2 accessions due to their very similar
vernalization phenotypes (Fig. 2D,E). This associa-
tion between FLC alleles and flowering time was signifi-
cantly reduced after 12 wk of vernalization compared

Figure 1. Different FLC haplotypes in 1307 worldwide A. thaliana

accessions. (A) Color-coded groups are indicated with numbers. The
rare Hap14 and Hap16–18 are not marked. (B) Haplotype structure of
the 650-kb FLC region in 1307 global A. thaliana accessions. Each
accession is represented in a column, and each SNP in the FLC 650-
kb window is represented in a row. At each SNP position, colors
indicate the most likely haplotype membership for each accession as
determined by fastPHASE analysis. Accessions are ordered by
haplotype at the FLC locus itself, which is delineated by solid black
horizontal lines. The prominent vertically oriented color blocks
indicate that haplotype sharing among accessions often extends
beyond the FLC locus itself. (C) Latitude and longitude of the
collection site corresponding to the accessions above (Horton
et al. 2012). (D–I) Geographical distribution of collection sites of Figure 2. Functional differentiation of FLC haplotypes. (A) Distri-
A. thaliana accessions carrying the five predominant and one bution of genetic polymorphism (the mean [6SEM] proportion of
intermediate frequency FLC haplotypes. (J) Centroid distribution of polymorphic sites) across the 10-kb FLC region. (B) Haplotype
the FLC haplotypes. analysis of 10-kb FLC genomic sequences from 47 A. thaliana
accessions using FLUXUS network analysis. Circle size illustrates
the frequency of the corresponding FLC haplotype. The number
along the branch shows the number of nucleotide differences.
FLC expression after transfer back to the warm differed Numbers in brackets indicate the corresponding haplotype numbers.
for each haplotype and correlated with flowering time (C) FLC expression of haplotypes containing FLC alleles with a slow
(Supplemental Fig. S1C,D). Additionally, 114 Swedish vernalization (SV, black symbols) or rapid vernalization (RV, white
A. thaliana accessions (Supplemental Table S2), which symbols) response. Expression values shown are mean (6SEM, n = 2–
13) from plants given 4 wk of cold followed by 10 d (T10) or 30 d
had been genotyped using the 250K SNP array (Horton (T30) of warm. Significant differences in vernalization response exist
et al. 2012), were also phenotyped for their vernalization between haplotypes at 10 d (T10) and 30 d (T30) (Kruskal-Wallis test:
response. Similar groupings of haplotypes with phenotypes H $ 18.42, d.f. = 5, P # 0.002 in all tests). (D) QTL analysis on
emerged from this analysis (Supplemental Fig. S2A,B). In populations generated from crosses between accessions containing
general, therefore, accessions in the different FLC haplotype different FLC haplotypes. Dashed horizontal lines show the signif-
icance thresholds. The FLC region is indicated by the vertical dashed
groups show distinct vernalization profiles that involve lines. (E) Summary of the presence of a QTL in the FLC region from
changes in silencing of FLC and altered flowering time. D and Supplemental Table S3. The ‘‘U’’ symbol indicates that there
The functional diversification of the different FLC is a QTL in the FLC genomic region in the F2 population; the ‘‘3’’
haplotypes is supported by previous quantitative trait indicates that there is no QTL.

1636 GENES & DEVELOPMENT

 Downloaded from genesdev.cshlp.org on December 11, 2014 - Published by Cold Spring Harbor Laboratory Press

FLC haplotypes underpin life history diversity

with 4 wk, suggesting that the functional difference be- conclude that noncoding cis variation within FLC is the
tween haplotypes is cold exposure-dependent (Supple- major factor determining differential vernalization re-
mental Fig. S2C,D). sponse in the characterized A. thaliana population.
To conclusively demonstrate that FLC variation is the Analysis of chimeric FLC fusions previously showed
major contributor to the QTL, we generated transgenic that cis polymorphism quantitatively modulates a chro-
lines differing only in their FLC allele. Twelve-kilobase matin silencing mechanism underpinning the epigenetic
FLC genomic clones from the A. thaliana accessions Edi-0, regulation of FLC in one Northern Swedish accession
Col-0, Lov-1, Ull2-5, and Var2-6, representing haplotypes (Coustham et al. 2012). To help define which of the many
RV1, RV2, SV1, SV2, and SV4, respectively, were trans- nucleotide differences between haplotypes determined
formed individually into FRI flc-2 (Michaels and Amasino variation in vernalization response, we first looked for
2001). FLC expression in the transgenic lines was com- natural recombination events across FLC genomic re-
pared with that from the natural accessions both with gions in accessions (Martin et al. 2010) and identified one
and without vernalization. The differential response to that we called R2 (Supplemental Fig. S3A). Polymorphism
vernalization shown by the FLC haplotypes was reca- in the R2 region showed the strongest association with
pitulated in the transgenic plants (R 2 = 0.80), whereas the maintenance of FLC silencing after vernalization
the correlation of prevernalization FLC expression between (Supplemental Fig. S3B). This region includes the nucle-
accessions and transgenic lines was weaker (R2 = 0.29) ation region important for FLC silencing; the VRE (ver-
(Fig. 3A,B). The transgenic analysis therefore shows that nalization response element), defined as important for
the SNPs defining the FLC alleles confer the vernaliza- maintenance of silencing (Sheldon et al. 2002; Finnegan
tion response characteristic of the parent A. thaliana and Dennis 2007; Angel et al. 2011); and a large number of
accession, with trans-regulation playing a greater role in the SNPs defining the haplotype RV1 (Fig. 3C). We isolated
prevernalization FLC expression. These data enable us to and swapped the FLC R2 region reciprocally between Lov-1
(SV1) and Edi-0 (RV1) FLC alleles, which show slow and
rapid silencing, respectively (Fig. 3C). The vernalization
response of multiple transgenic lines was then compared.
The SV1/RV1/SV1 (SRS) chimeric FLC shows stable FLC
silencing after 4 wk of cold, which results in a rapid ver-
nalization response compared with its SV1 FLC parent
(Fig. 3D). Thus, the SRS vernalization response matches
that of the RV1 accession rather than the SV1 accession.
Similarly, the RV1/SV1/RV1 (RSR) chimera shows a ver-
nalization response similar to the SV1 accession despite
containing only a small part of the SV1 FLC allele (Fig. 3D).
These results demonstrate that polymorphisms within
R2 from RV1 and SV1 contribute to the differences in FLC
silencing after 4 wk of vernalization, respectively. This
region also contains the four single-nucleotide changes
that cause the differential silencing of Lov-1 (SV1) and
Col-0 (RV2) FLC alleles (Coustham et al. 2012). Remark-
ably, the Edi-0 (RV1) and Col-0 (RV2) FLC alleles do not
share any polymorphisms that are not also found in other
FLC haplotypes (Figs. 2B, 3C), demonstrating that the
causative nucleotide changes that distinguish slow and
rapid vernalization have evolved multiple times. The re-
combination event characterizing the RV1 haplotypes, the
relative lack of variability of SNPs in the functional R2
region, and their relatively broad geographical distribution
(Hap3/RV1) (Fig. 1E) suggest a relatively recent spread of
Figure 3. Transgenic analysis shows that noncoding FLC poly- RV1 accessions, potentially indicating a strong advantage
morphism plays a central role in vernalization response of natural of rapid vernalization in some locations.
Arabidopsis accessions. (A) Prevernalization FLC expression (non- Consistent with this, faster vernalization has been
vernalizaion [NV]) in transgenic lines correlates weakly with natural
accessions. (B) Relative FLC expression [T30/(T30 + T10)] in trans-
predicted to confer a selective advantage to plants where
genic lines correlates significantly with expression in natural acces- avoidance of summer drought or other high risk is
sions. R2 shows the explained amount of variation; the corresponding beneficial (Caicedo et al. 2004; Stinchcombe et al. 2004,
P-values are 0.201 and 0.025 in A and B. (C) Schematic illustration of 2005; Scarcelli et al. 2007; Satake 2010). We analyzed
FLC polymorphisms and the constructs of R2 chimeras. The SNPs differences in fitness-related traits conferred by different
defining the different FLC haplotypes are shown by the vertical bars.
FLC alleles in F2 populations (Fig. 4; Supplemental Fig.
The bar transparency indicates the diversity of each SNP within
each haplotype. The bars shared between haplotypes signify shared S4). After a relatively short vernalization (4 wk), the slow
SNPs. The FLC gene structure is shown above: Filled black squares vernalization (Kulturen-1) FLC allele associated with lower
are exons, and horizontal lines are introns and untranslated regions. seed yield than the rapid vernalization (Rev-1) FLC allele.
RSR and SRS chimeras show the exchange of R2 regions between After longer vernalization (12 wk), both alleles produced
Edi-0 and Lov-1 FLC alleles. (D) FLC expression analysis of trans- a similar seed yield (Fig. 4A,B; Supplemental Fig. S4A).
genic FRI flc-2 plants containing chimeric FLC alleles. Plants were
given either no cold (NV) or 4 wk of cold followed by 10 d (T10) or Total seed weight represented an increased production of
30 d (T30) of warm. Values are means 6standard error from three seed number, as no significant difference was found in the
biological repeats. weight of each seed (Supplemental Fig. S4B). High fecundity

GENES & DEVELOPMENT 1637

 Downloaded from genesdev.cshlp.org on December 11, 2014 - Published by Cold Spring Harbor Laboratory Press

Li et al.

their contributions to life history diversity. Use of FLC

genomic sequence from a close relative, Arabidopsis
lyrata, did not help define which of the haplotypes are
ancestral because of the high sequence divergence and
large number of indels in the functionally important
intronic regions. However, rapid vernalization might have
arisen through selection pressure on the length of the
plant life cycle due to high mortality from drought
(Hoffmann 2005; Satake 2010), herbivory (Parachnowitsch
and Lajeunesse 2012), pathogens (Kemen et al. 2011),
agriculture, or human disturbance. The broader climate
envelope of A. thaliana compared with its close relatives,
which extends beyond the cool temperate regions into
more Mediterranean climates, argues that faster vernaliza-
tion could have increased the A. thaliana range (Hoffmann
2005). This and related studies should provide important
knowledge needed to assess plant response to climate
change.

Materials and methods

Figure 4. Comparison of total seed production between plants with
different vernalization responses. (A,B) Comparison of seed yield Plant materials and growth condition
from F2 plants homozygous or heterozygous for the different FLC
alleles after 4 wk (A; ANOVA, R2 = 0.41, F2227 = 80.43, P < 0.001) Transgenic lines were obtained as described previously (Coustham et al.
and 12 wk (B; ANOVA, R2 = 0.028, F2276 = 4.97, P = 0.008) of 2012). Plants were vernalized for 4 or 12 wk at 4°C and harvested for FLC
vernalization. Crosses indicate statistical outliers that fall more expression analysis 10 d (T10) and 30 d (T30) after transfer from cold, along
than three standard errors away from the mean. (C–E) Seed proxy with nonvernalized plants (NV). Flowering time was assayed as days after
values as a measure of fitness of field-grown accessions. The data set bolting when stems reached 3 cm in height.
included only accessions with active FRIGIDA alleles (16 accessions
from haplotypes RV1 and RV2 and 23 accessions from haplotype
SV1-4). FLC haplotypes with rapid and slower vernalization response Plant fitness estimation
phenotypes were analyzed. Differences in yield were measured in
Controlled environment conditions were as follows: Plants were vernal-
field experiments in Norwich as in Fournier-Level et al. (2011). P-
ized for 4 or 12 wk at 4°C under short-day conditions (8 h light) and moved
values are calculated from population structure-corrected ANOVA
to controlled environment conditions at 22°C under a long-day photope-
(see the Materials and Methods) (C; P = 0.0023), summer sowing
(D; P = 0.00017), and fall sowing (E; P = 0.99). riod (16 h light). Soil was from The Scotts Company Ltd., Sphagnum Moss
peat (pH 5.5–6.0), with fertilizer added (milligram/liter): N150, P200, and
K200. Twenty-five percent grit was added before use. Pot size was 5 3 5
cm. Plants were grown individually in each pot. In the greenhouse, total
has been associated with later flowering (Redei 1962); seed was collected from each plant after siliques had matured and dried
however, for these rapid vernalization and slow vernali- naturally and was weighed.
zation alleles, an inverse relationship was found between In the field, plants of 39 accessions were grown in common gardens at
flowering time and seed production after 4 wk of ver- the John Innes Center in spring, summer, and fall plantings as described
nalization (Supplemental Fig. S4C). A similar relationship previously (Fournier-Level et al. 2011). Fitness was estimated as the
product of silique number at senescence and length of a representative
between vernalization response and fitness was also found
silique (a proxy for seed number) for each individual (Fournier-Level et al.
for other rapid vernalization/slow vernalization allele 2011).
combinations (Supplemental Fig.S4D). These data sug-
gest that in some conditions, particularly after short cold
periods, faster vernalization response can lead to enhanced FLC haplotype analysis using FLUXUS network analysis
fitness. To further explore the association of FLC haplo- FLC genome sequences were aligned using ClustalW with manual
types with fitness, we analyzed the field data of A. thaliana corrections in Mega5 (Tamura et al. 2011) and analyzed in FLUXUS
accessions grown under different natural seasonal conditions network software using median joining setting (Bandelt et al. 1999).
in Norwich (United Kingdom), where spring-, summer-,
and fall-germinating cohorts are observed (Fournier-Level
FLC expression analysis
et al. 2011). Accessions with slow vernalization haplo-
types produced less seed compared with rapid vernal- FLC expression data were obtained as described previously (Box et al.
ization accessions in spring and summer sowings (which 2011). We generated 50 independent transgenic lines for each construct
experienced only brief exposure to vernalizing tempera- using the Agrobacterium tumefaciens floral dip technique. To overcome
tures), but fitness was similar between all groups in the variability in expression between transgenic lines, we pooled a random
10 seedlings from each of these 50 lines into a single sample. RNA was
fall sowings (with exposure to extended cold such that
extracted from this pool and analyzed for FLC expression using quantita-
all plants were fully vernalized) (Fig. 4C–E). These data tive RT–PCR as described in Box et al. (2011). For each time point, three
encourage further extensive transplantation experiments independent pools of the 50 transgenic line pools were assayed.
that thoroughly investigate fitness differences between the
haplotypes.
The emerging picture is that functional differences in Box plot and statistical analysis
vernalization have evolved predominantly through cis Box plot and statistical analysis of flowering time and seed production were
polymorphism at FLC. Multiple alleles have arisen and carried out in the GENESTAT package with the default setting (Ripatti
been maintained in the A. thaliana population due to et al. 2009).

1638 GENES & DEVELOPMENT

 Downloaded from genesdev.cshlp.org on December 11, 2014 - Published by Cold Spring Harbor Laboratory Press

FLC haplotypes underpin life history diversity

Genotyping and QTL analysis of F2 populations Bandelt HJ, Forster P, Rohl A. 1999. Median-joining networks for inferring
intraspecific phylogenies. Mol Biol Evol 16: 37–48.
F2 populations were genotyped using a 384-SNP Illumina GoldenGate Box MS, Coustham V, Dean C, Mylne JS. 2011. Protocol: a simple phenol-
genotyping assay (http://www.illumina.com/technology/goldengate_ based method for 96-well extraction of high quality RNA from
genotyping_assay.ilmn). Genome Studio software (http://www.illumina.com/ Arabidopsis. Plant Methods 7: 7.
software/genomestudio_software.ilmn) and custom-made scripts in R software Broman KW, Wu H, Sen S, Churchill GA. 2003. R/qtl: QTL mapping in
(http://www.R-project.org) were used for SNP calling. experimental crosses. Bioinformatics 19: 889–890.
Quantitative trait loci mapping was performed on raw phenotypic Caicedo A, Stinchcombe JR, Olsen KM, Schmitt J, Purugganan MD. 2004.
values without any transformation. QTL analyses were carried out using Epistatic interaction between Arabidopsis FRI and FLC flowering
the default parameters of the scanone function implemented in the R/qtl time genes generates a latitudinal cline in a life history trait. Proc
package (Broman et al. 2003). The LOD significance threshold was Natl Acad Sci 101: 15670–15675.
determined by permutation test (1000 permutations, significance level Coustham V, Li P, Strange A, Lister C, Song J, Dean C. 2012. Quantitative
of 5%) for each F2 population.
modulation of polycomb silencing underlies natural variation in
vernalization. Science 337: 584–587.
M&M test of the correlation between FLC polymorphism Finnegan EJ, Dennis ES. 2007. Vernalization-induced trimethylation of
histone H3 lysine 27 at FLC is not maintained in mitotically
and gene expression
quiescent cells. Curr Biol 17: 1978–1983.
A general linear model (GLM) with T30/(T30 + T10) as response variable Fournier-Level A, Korte A, Cooper MD, Nordborg M, Schmitt J, Wilczek
and SNP as the predictor was analyzed for each SNP separately. The AM. 2011. A map of local adaptation in Arabidopsis thaliana. Science
F coefficient (adjusted MS/error variance) was calculated and plotted 334: 86–89.
against the SNP position. The strength of the association of each SNP and Hoffmann MH. 2005. Evolution of the realized climatic niche in the
the response variable was expressed by calculating the effect size genus Arabidopsis (Brassicaceae). Evolution 59: 1425–1436.
estimator h2. This statistic expresses the ratio of variance explained in Horton MW, Hancock AM, Huang YS, Toomajian C, Atwell S, Auton A,
the dependent variable by a SNP. This analysis shows that SNPs in Muliyati NW, Platt A, Sperone FG, Vilhjalmsson BJ, et al. 2012.
the R2 region (Supplemental Fig. S5B) explain ;22.1% of the variation Genome-wide patterns of genetic variation in worldwide Arabidop-
inT30/(T10 +T30) (h2 = 0.221), which is considered a large effect size. sis thaliana accessions from the RegMap panel. Nat Genet 44: 212–
216.
Kang HM, Sul JH, Service SK, Zaitlen NA, Kong SY, Freimer NB,
fastPHASE haplotype analysis Sabatti C, Eskin E. 2010. Variance component model to account for
SNPs in a window 650 kb around FLC were extracted from a preimpu- sample structure in genome-wide association studies. Nat Genet
tation version of the Regional Mapping Project SNP panel described in 42: 348–354.
Horton et al. (2012). These SNPs were used as the input into fastPHASE Kemen E, Gardiner A, Schultz-Larsen T, Kemen AC, Balmuth AL,
version 1.4.0 (Scheet and Stephens 2006), which was run using the Robert-Seilaniantz A, Bailey K, Holub E, Studholme DJ, Maclean
default settings, with the exception of invoking the Pzp option to D, et al. 2011. Gene gain and loss during evolution of obligate
output raw likelihood values. For each SNP in the analysis, the cluster parasitism in the white rust pathogen of Arabidopsis thaliana. PLoS
membership with the highest likelihood was designated as the haplotype Biol 9: e1001094.
for that SNP. Li Y, Huang Y, Bergelson J, Nordborg M, Borevitz JO. 2010. Association
mapping of local climate-sensitive quantitative trait loci in Arabidopsis
thaliana. Proc Natl Acad Sci 107: 21199–21204.
Population structure-corrected ANOVAs for seed proxy Martin DP, Lemey P, Lott M, Moulton V, Posada D, Lefeuvre P. 2010.
analysis in accessions RDP3: a flexible and fast computer program for analyzing recombi-
nation. Bioinformatics 26: 2462–2463.
A standard approach for accounting for such genomic background effects Michaels SD. 2009. Flowering time regulation produces much fruit. Curr
is linear mixed models (Vilhjalmsson and Nordborg 2013). We assessed the Opin Plant Biol 12: 75–80.
significance of phenotypic differences between haplotypes by carrying out Michaels SD, Amasino RM. 2001. Loss of FLOWERING LOCUS C
an ANOVA on the residuals after subtracting the best linear unbiased activity eliminates the late-flowering phenotype of FRIGIDA and
predictors (BLUPs) of the effect explained by the genomic background. autonomous pathway mutations but not responsiveness to vernali-
The BLUPs were estimated using EMMAX (Kang et al. 2010) with zation. Plant Cell 13: 935–941.
a kinship matrix estimated from the Regional Mapping Project data Mitchell-Olds T, Schmitt J. 2006. Genetic mechanisms and evolution-
(Horton et al. 2012). ary significance of natural variation in Arabidopsis. Nature 441:
947–952.
Nordborg M, Bergelson J. 1999. The effect of seed and rosette cold
Acknowledgments treatment on germination and flowering time in some Arabidopsis
thaliana (Brassicaceae) ecotypes. Am J Bot 86: 470–475.
We thank all members of the Dean and Howard groups for discussions.
O’Neill CM, Morgan C, Kirby J, Tschoep H, Deng PX, Brennan M, Rosas U,
This research was supported by UK Biotechnology and Biological
Fraser F, Hall C, Gill S, et al. 2008. Six new recombinant inbred
Sciences Research Council (BBSRC) grant BB/I007857/1, Institute Stra-
populations for the study of quantitative traits in Arabidopsis thaliana.
tegic Programme grant BB/J004588/1, European Research Council Ad-
Theor Appl Genet 116: 623–634.
vanced Investigator grant 233039 ENVGENE, and National Science
Parachnowitsch AL, Lajeunesse MJ. 2012. Adapting with the enemy:
Foundation Frontiers in Integrative Biological Research grant NSF
local adaptation in plant-herbivore interactions. New Phytol 193:
EF-0425759.
294–296.
Redei GP. 1962. Supervital mutants of Arabidopsis. Genetics 47: 443–
460.
References
Ripatti S, Becker T, Bickeboller H, Dominicus A, Fischer C, Humphreys
Alonso-Blanco C, Mendez-Vigo B. 2014. Genetic architecture of naturally K, Jonasdottir G, Moreau Y, Olsson M, Ploner A, et al. 2009.
occurring quantitative traits in plants: an updated synthesis. Curr GENESTAT: an information portal for design and analysis of genetic
Opin Plant Biol 18C: 37–43. association studies. Eur J Hum Genet 17: 533–536.
Alonso-Blanco C, El-Assal SE, Coupland G, Koornneef M. 1998. Analysis Salome PA, Bomblies K, Laitinen RA, Yant L, Mott R, Weigel D. 2011.
of natural allelic variation at flowering time loci in the Landsberg Genetic architecture of flowering-time variation in Arabidopsis
erecta and Cape Verde Islands ecotypes of Arabidopsis thaliana. thaliana. Genetics 188: 421–433.
Genetics 149: 749–764. Satake A. 2010. Diversity of plant life cycles is generated by dynamic
Angel A, Song J, Dean C, Howard M. 2011. A Polycomb-based switch epigenetic regulation in response to vernalization. J Theor Biol 266:
underlying quantitative epigenetic memory. Nature 476: 105–108. 595–605.

GENES & DEVELOPMENT 1639

 Downloaded from genesdev.cshlp.org on December 11, 2014 - Published by Cold Spring Harbor Laboratory Press

Li et al.

Scarcelli N, Cheverud JM, Schaal BA, Kover PX. 2007. Antagonistic

pleiotropic effects reduce the potential adaptive value of the FRIGIDA
locus. Proc Natl Acad Sci 104: 16986–16991.
Scheet P, Stephens M. 2006. A fast and flexible statistical model for large-
scale population genotype data: applications to inferring missing
genotypes and haplotypic phase. Am J Hum Genet 78: 629–644.
Sheldon CC, Rouse DT, Finnegan EJ, Peacock WJ, Dennis ES. 2000. The
molecular basis of vernalization: the central role of FLOWERING
LOCUS C (FLC). Proc Natl Acad Sci 97: 3753–3758.
Sheldon CC, Conn AB, Dennis ES, Peacock WJ. 2002. 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.
Shindo C, Lister C, Crevillen P, Nordborg M, Dean C. 2006. Variation in
the epigenetic silencing of FLC contributes to natural variation in
Arabidopsis vernalization response. Genes Dev 20: 3079–3083.
Stinchcombe JR, Weinig C, Ungerer M, Olsen KM, Mays C, Halldorsdottir
SS, Purugganan MD, Schmitt J. 2004. A latitudinal cline in flowering
time in Arabidopsis thaliana modulated by the flowering time gene
FRIGIDA. Proc Natl Acad Sci 101: 4712–4717.
Stinchcombe JR, Caicedo AL, Hopkins R, Mays C, Boyd EW, Purugganan
MD, Schmitt J. 2005. Vernalization sensitivity in Arabidopsis thaliana
(Brassicaceae): the effects of latitude and FLC variation. Am J Bot 92:
1701–1707.
Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. 2011.
MEGA5: molecular evolutionary genetics analysis using maximum
likelihood, evolutionary distance, and maximum parsimony methods.
Mol Biol Evol 28: 2731–2739.
Vilhjalmsson BJ, Nordborg M. 2013. The nature of confounding in
genome-wide association studies. Nat Rev Genet 14: 1–2.

1640 GENES & DEVELOPMENT

 Downloaded from genesdev.cshlp.org on December 11, 2014 - Published by Cold Spring Harbor Laboratory Press

Multiple FLC haplotypes defined by independent cis-regulatory

variation underpin life history diversity in Arabidopsis thaliana
Peijin Li, Daniele Filiault, Mathew S. Box, et al.

Genes Dev. 2014 28: 1635-1640 originally published online July 17, 2014
Access the most recent version at doi:10.1101/gad.245993.114

Supplemental http://genesdev.cshlp.org/content/suppl/2014/07/11/gad.245993.114.DC1.html
Material

References This article cites 35 articles, 19 of which can be accessed free at:
http://genesdev.cshlp.org/content/28/15/1635.full.html#ref-list-1

Open Access Freely available online through the Genes & Development Open Access option.

Creative This article, published in Genes & Development, is available under a Creative Commons
Commons License (Attribution 4.0 International), as described at
License http://creativecommons.org/licenses/by/4.0.

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the top
Service right corner of the article or click here.

To subscribe to Genes & Development go to:

http://genesdev.cshlp.org/subscriptions

© 2014 Li et al.; Published by Cold Spring Harbor Laboratory Press