Tagging and mapping candidate loci for vernalizationand flower initiation in hexaploid oat

Itamar C. Nava • Charlene P. Wight •

Marcelo T. Pacheco • Luiz C. Federizzi •

Nicholas A. Tinker

Received: 21 September 2011 / Accepted: 11 February 2012

� Her Majesty the Queen in Rights of Canada 2012

Abstract Flowering time is a decisive factor in the

adaptation of oat. Some oat varieties require low

temperatures for floral initiation, a process called

vernalization. The objectives of this study were to

clone, characterize, and map genes associated with

vernalization in oat, and to identify markers linked to

quantitative trait loci (QTL) that affect vernalization

response. Genetic linkage maps were developed using

Diversity Arrays Technology markers in recombinant

inbred lines from the oat populations UFRGS 8 9

UFRGS 930605 and UFRGS 881971 9 Pc68/5*Star-

ter. Flowering time and response to vernalization were

characterized using field trials and controlled green-

house experiments, and QTL were identified in two

genetic regions on each of the two maps. PCR primer

pairs anchored in the conserved coding regions of the

Vrn1, Vrn2, and Vrn3 genes from wheat, barley, and

Lolium were used to amplify and clone corresponding

oat sequences. Cloned sequences corresponding to the

targeted genes were recovered for both Vrn1 and Vrn3.

A copy of the Vrn3 gene was mapped using a PCR

amplicon, and an oat Vrn1 fragment was mapped by

restriction fragment length polymorphism analysis.

The location of the mapped Vrn1 locus was homol-

ogous to major QTL affecting flowering time in other

work, and homoeologous to major QTL affecting

response to vernalization in this study.

Keywords Days to heading � Flowering �Vernalization � Hexaploid oat (Avena sativa) �DArT markers � Quantitative trait loci (QTL)

Introduction

Hexaploid oat (Avena sativa L.) is an important cereal

crop worldwide. It is adapted to a wide range of

environments and is cultivated predominantly in

temperate regions or in winter seasons. Flowering

Electronic supplementary material The online version ofthis article (doi:10.1007/s11032-012-9715-x) containssupplementary material, which is available to authorized users.

Present Address:I. C. Nava

Department of Agronomy and Plant Genetics,

University of Minnesota, 411 Borlaug Hall,

1991 Upper Buford Circle, Saint Paul,

MN 55108-6026, USA

C. P. Wight � N. A. Tinker (&)

Agriculture and Agri-Food Canada, ECORC,

K.W. Neatby Bldg, 960 Carling Ave., C.E. Farm, Ottawa,

ON K1A 0C6, Canada

e-mail: nick.tinker@agr.gc.ca

M. T. Pacheco � L. C. Federizzi

Department of Crop Science, Agronomy School,

Federal University of Rio Grande do Sul (UFRGS),

Av. Bento Goncalves, 7712, Porto Alegre,

RS 91501-970, Brazil

123

Mol Breeding

DOI 10.1007/s11032-012-9715-x

(heading1) time is a decisive factor in the adaptation of

oat, and its transition from vegetative shoot apical

meristem to reproductive phase is mainly regulated by

photoperiod and vernalization. Hexaploid oat is a

long-day plant, with longer photoperiods promoting

earlier flowering in most genotypes (Sorrells and

Simmons 1992). Vernalization is the requirement of a

long exposure to low temperatures to induce flower-

ing, an adaptation to protect sensitive floral organs

against the cold (Chouard 1960). Oat genotypes vary

in their response to vernalization, and both spring and

winter forms of A. sativa exist. However, the charac-

teristics that define a winter oat (vernalization require-

ment and tolerance to freezing) are expressed to

varying degrees. Many winter oat varieties from

around the globe can be increased as spring-seeded

annuals, while others are not capable of producing

seed under these conditions (Tinker et al. 2009).

A pre-requisite to the detailed genetic analysis of

flowering time or any other trait is the existence of a

molecular linkage map. It is also important that

markers on the map be transferable between different

laboratories and different genetic populations. Oat has

benefited from the availability of mapped restriction

fragment length polymorphism (RFLP) markers that

have allowed detailed mapping (e.g., Wight et al.

2003) and cross-species comparisons (e.g., Moore

et al. 1995). However, RFLP markers are costly and

labour-intensive. Relatively few simple sequence

repeat (SSR) markers are available in oat (Wight

et al. 2010), single nucleotide polymorphism (SNP)

markers are only now being developed (Oliver et al.

2011), and other marker types such as amplified

fragment length polymorphism (AFLP) and rapid

amplification of polymorphic DNA (RAPD) often

provide inconsistent results for comparative mapping.

Recently, Diversity Arrays Technology (DArT) mark-

ers have been developed in oat, and these were used to

enhance the map of the hexaploid oat reference

population Kanota 9 Ogle (KO) (Tinker et al.

2009). These new markers have improved the ability

to conduct comparative genomics in oat, and addi-

tional maps based on DArT markers are under

development. These markers will, therefore, be useful

for refining information about genes and quantitative

trait loci (QTL) that affect important agronomic traits

such as flowering time.

QTL associated with early flowering and response

to vernalization have been identified in the Kano-

ta 9 Ogle and Ogle 9 TAMO-301 oat mapping pop-

ulations (Holland et al. 1997; Holland et al. 2002).

Recently, QTL affecting flowering time were identi-

fied in three Brazilian oat populations: UFRGS

8 9 Pc68/5*Starter, UFRGS 881971 9 Pc68/5*Star-

ter, and UFRGS 8 9 UFRGS 930605 (Locatelli et al.

2006). One QTL with a major effect on flowering time

was identified in each cross. Comparative mapping

showed that this QTL, with earliness alleles originat-

ing from UFRGS 8 and UFRGS 881971, is located in

the region homologous to KO linkage group 17. It was

putatively identified as the Di1 gene, a major dominant

gene for day length insensitivity (Burrows 1986). The

same region containing the QTL and Di1 also showed

homology to the rice gene Hd1 and the Arabidopsis

gene CONSTANS (CO) (Locatelli et al. 2006). Hd1

and CO are flowering promoter genes that integrate

metabolic pathways in response to photoperiod (Hay-

ama et al. 2003; Yan et al. 2006).

The synchronization of flowering time is a complex

process that is most studied and best understood in the

model system Arabidopsis thaliana. The vernalization

requirement of Arabidopsis is mainly controlled by the

FRIGIDA (FRI) and FLOWERING LOCUS C (FLC)

genes (Michaels and Amasino 1999). FLC acts as a

repressor of flowering and encodes a MADS-box

transcriptional regulator that delays both the transition

to reproductive apex development and the promotion

of flowering by long days until the plants have been

vernalized (Sheldon et al. 1999; Michaels and Ama-

sino 1999). In plants that have not been vernalized,

FLC is expressed at high levels and represses the

transcription of two floral promoters: FLOWERING

LOCUS T (FT) and SUPPRESSOR OF OVER-

EXPRESSION OF CONSTANS 1 (SOC1) (Helliwell

et al. 2006). The expression of FLC is down-regulated

by vernalization, thereby allowing the long day

induction of FT and SOC1 (Sheldon et al. 1999;

Michaels and Amasino 1999). The natural allelic

variation at the FRI and FLC loci among different

accessions of Arabidopsis accounts for most of the

differences in flowering time between spring (early)

and winter (late) ecotypes (Michaels et al. 2003).

1 Flowering time is often estimated based on observations of

heading date in cereals because floral initiation happens within

heads or panicles where it is difficult to observe. Although we

recorded observations for heading, we use the term flowering

hereafter for consistency with other literature in cereal crops.

Mol Breeding

123

Genetic analyses have identified at least three genes

that affect the vernalization requirement in wheat and

barley: Vrn1, Vrn2, and Ft (Vrn3). These genes are

implicated in both the vernalization response and the

response to day length (Trevaskis et al. 2007). Vrn1

promotes the transition to reproductive development

and encodes a MADS-box transcription factor with

high similarity to the Arabidopsis meristem identity

genes APETALA1 (AP1), CAULIFLOWER (CAU), and

FRUITFULL (FUL) (Yan et al. 2003; Trevaskis et al.

2003). In varieties that require vernalization to flower,

the expression of Vrn1 is induced by vernalization

(Yan et al. 2003). In varieties without this requirement,

expression of Vrn1 increases at floral initiation and

remains high throughout subsequent stages of the

reproductive phase (Trevaskis et al. 2007). This

suggests that Vrn1 is essential in regulating meristem

identity in a way that is not limited to the vernalization

response. The Vrn1 gene has been mapped in colinear

regions of the long arms of chromosomes 5A, 5B, and

5D in wheat and of chromosome 5H in barley (Iwaki

et al. 2002; Yan et al. 2003). Allelic variation at the

Vrn1 locus is mainly associated with natural mutations

in regulatory regions of the Vrn1 promoter or first

intron (Fu et al. 2005). This allelic variation leads to the

elimination or reduction of the vernalization require-

ment and, consequently, divides varieties into the

winter and spring growth habits.

In vernalization-responsive wheat and barley vari-

eties, the delay of flowering prior to winter is mediated

by the Vrn2 gene (Yan et al. 2004). The Vrn2 locus

includes two duplicated genes: Zcct1 and Zcct2. The

proteins encoded by these genes each contain a

putative zinc-finger motif (which might mediate

DNA binding) and a CCT domain. There are no clear

homologues for these genes in Arabidopsis (Yan et al.

2004). Vrn2 is expressed with a diurnal pattern under

long days (Dubcovsky et al. 2006; Trevaskis et al.

2006), which is similar to Ghd7, the gene with the

most homology in rice (Xue et al. 2008). Once plants

have experienced vernalization, the expression of

Vrn2 decreases rapidly (Yan et al. 2004; Dubcovsky

et al. 2006). Non-functional mutations or the complete

deletion of both Zcct1 and Zcct2 genes are associated

with spring growth habit in wheat and barley (Yan

et al. 2004; Dubcovsky et al. 2005; Distelfeld et al.

2009). Diploid wheat and barley accessions with a

winter growth habit have at least one functional Zcct

gene (Yan et al. 2004; Karsai et al. 2005).

The acceleration of flowering under long days is

mediated by the Vrn3 gene, a homologue of the

Arabidopsis photoperiod gene FLOWERING LOCUS

T (FT). Vrn3 encodes an RAF kinase inhibitor-like

protein involved in cellular signalling (Yan et al. 2006;

Distelfeld et al. 2009). The induction of FT (Vrn3) in

the leaves results in the production of a long-distance

flowering signal (florigen) that travels through the

leaves to the shoot apex and promotes flowering

(Huang et al. 2005; Corbesier et al. 2007; Tamaki et al.

2007). In barley, the induction of Vrn3 by long days

requires the PHOTOPERIOD 1 (Ppd-H1) gene, which

is a pseudo-response regulator gene (Turner et al.

2005). The Vrn3 gene is responsible for natural allelic

variation in vernalization requirement and day length

response in wheat and barley, providing additional

sources of adaptive diversity for these economically

important crops.

Like wheat and barley, oat varieties grown in

southern Brazil show various responses to low tem-

perature-dependent floral initiation, but the genetic

and molecular factors affecting this response have not

been elucidated. The high level of similarity between

the Vrn1, Vrn2, and Vrn3 genes of wheat, barley, and

other temperate cereals suggested that these genes

might also have similar structures in hexaploid oat.

The availability of vernalization gene sequences from

other temperate cereals provided an opportunity to

explore this.

The objectives of this study were: (1) to clone and

characterize regions of the genes associated with

vernalization in oat using sequence information from

other grass species, (2) to enhance and update the

linkage maps of Brazilian oat populations using DArT

markers, (3) to locate the vernalization genes on the

oat maps, and (4) to identify markers linked to QTL

that affect the response to vernalization in oat.

Materials and methods

Genetic populations

Two genetic populations of recombinant inbred lines

(RILs) developed from the crosses UFRGS 8 9

UFRGS 930605 (U86) and UFRGS 881971 9 Pc68/

5*Starter (U71P) were analyzed in this study. From

the two crosses, a total of 154 and 142 RILs were

developed by single-seed descent to the F5 generation.

Mol Breeding

123

These populations were selected based on the differ-

ential response to vernalization of the parental lines

observed in previously reported studies (Locatelli

et al. 2006; Locatelli et al. 2008). The oat breeding line

Pc68/5*Starter was developed at the University of

Minnesota, USA, whereas all other parental lines were

developed at the Federal University of Rio Grande do

Sul (UFRGS), Brazil. An additional set of 80 RILs

from the Kanota 9 Ogle (KO) mapping population

(Tinker et al. 2009) was also used in this study.

Kanota, UFRGS 930605, and UFRGS 881971 are

known to show greater response to vernalization than

the other parental lines.

Field tests of early and late plantings

Parental lines and RILs from the crosses U86 and

U71P were included as entries in an experiment

designed to measure differences in flowering time for

early and late plantings in the field. For the U86

population, 154 RILs were evaluated in the early

planting and 149 RILs in the late planting. For the

U71P population, 138 and 135 RILs were evaluated in

the early and late plantings, respectively. The exper-

iment was conducted at the Eastern Cereal and Oilseed

Research Centre, Agriculture and Agri-Food Canada,

Ottawa, Canada, in 2007 (latitude 45� 220 N, longitude

75� 430 W). The first planting was seeded on May 3

and the second on May 25. Twenty seeds for each

individual RIL or parental line were sown in non-

randomized, unreplicated hill plots. The hill plots were

spaced 0.50 m apart in each row, with the rows 1.0 m

apart. Flowering time, estimated as the number of days

to heading, was recorded as the approximate time of

panicle appearance at the 55th stage of Zadoks’ scale

(Zadoks et al. 1974).

Greenhouse tests of vernalized and non-vernalized

plants

Parental lines and RILs from each of the crosses U86

and U71P were grown under controlled conditions in

an experiment designed to measure the response to

vernalization in oat seedlings. Parental lines and RILs

were subjected to vernalized and non-vernalized

treatments in a randomized complete block design,

replicated twice. A total of 138 and 142 RILs from the

cross U86 was evaluated for the vernalized and non-

vernalized treatments, respectively. For the U71

population, 129 RILs were evaluated for both vernal-

ized and non-vernalized treatments.

For the vernalized treatment, four seeds of each RIL

and parental line were germinated in plastic trays

containing sterilized soil at room temperature. When

seedling coleoptiles were approximately 20 mm tall,

the trays were transferred to a cold chamber with the

temperature adjusted to 2�C and a 14-h photoperiod.

Seedlings were kept under these conditions for three

weeks with no water or fertilizer supplementation.

The non-vernalized treatment followed the same

procedures described for the vernalized treatment with

the exception of the cold treatment. The seeds for this

treatment were germinated 4 days before the end of

the third week of the vernalization treatment. When

the coleoptiles in both treatments reached approxi-

mately the same size, the trays were transferred to a

greenhouse with supplemental lighting at 24�C with a

14-h photoperiod. The days to flowering were scored

from the day the seedlings were transferred to the

greenhouse to the beginning of panicle emergence

(stage 50 of Zadoks’ scale). The response to vernal-

ization was calculated as the difference in days to

flowering between the plants in the non-vernalized and

vernalized treatments.

DNA isolation and purification

Ten random seeds of each parental line and RIL from

the crosses U86, U71P, and KO were grown in the

greenhouse. Tissue was harvested when plants were

approximately 20 cm tall. Immediately after harvest-

ing, 20 ml tissue samples were frozen in liquid

nitrogen, ground, and stored at -70�C. Large-scale

DNA extraction was performed using the protocol

described by Wight et al. (2003).

Cloning of candidate gene sequences from oat

Sequences related to the Vrn1, Vrn2, and Vrn3 genes of

various grass species were obtained from GenBank

[NCBI (National Center for Biotechnology Informa-

tion); http://www.ncbi.nlm.nih.gov/]. Heterologous

alignments containing genomic sequences from wheat,

barley, and Lolium were produced for each candidate

gene using the CLUSTAL W (Thompson et al. 1994)

add-in module within the BioEdit Sequence Alignment

Editor (Hall 1999). A consensus template sequence

was derived from each alignment. To locate any introns

Mol Breeding

123

present, the genomic and cDNA sequences were

aligned using the program est2genome (http://mobyle.

pasteur.fr/cgi-bin/portal.py?#forms::est2genome). Prim-

ers anchored in conserved coding regions were then

designed using the program Primer3 (Rozen and Skaletsky

2000).

Twenty-two primer pairs representing the Vrn1,

Vrn2, and Vrn3 genes were used to amplify sequences

from the oat varieties UFRGS 8, UFRGS 930605,

UFRGS 881971, Pc68/5*Starter, Kanota, Ogle, and

TAMO-301. Primer pairs producing successful ampli-

cons are shown in Electronic Supplementary Material

Fig. S1. PCR (polymerase chain reaction) performed

in an Eppendorf Mastercycler machine (Eppendorf,

Hamburg, Germany) was used to determine the

optimal annealing temperature for each primer pair

using the following conditions: 3 min at 94�C, 40

cycles of 30 s at 94�C, 45 s at 53–65�C, and 2 min at

72�C, then 10 min at 72�C, and hold at 4�C. PCR

reactions were conducted in a volume of 25 ll

containing 1–2 mM genomic DNA, 200 nM of each

primer, 1.5 mM MgCl2, 200 lM dNTPs, 2.5 U of Taq

DNA polymerase (Invitrogen Canada, Inc., Burling-

ton, ON, Canada), and 1 9 PCR buffer (as supplied by

the manufacturer). Amplified fragments were sepa-

rated in 1.5% agarose gels and visualized using

ethidium bromide staining.

PCR products of the expected sizes were extracted

from agarose gels using the Wizard� SV Gel and PCR

Clean-Up System (Promega Corp., Madison, WI,

USA). The purified DNAs were ligated into the vector

pCR�4-TOPO (Invitrogen Canada, Inc.) and trans-

formed into E. coli following the manufacturer’s

instructions. Eight to twelve random clones from each

ligation were inoculated into liquid LB medium ?

kanamycin and incubated at 37�C under shaking

conditions for 12–14 h. To isolate DNA for amplifi-

cation, tubes containing LB medium and grown cells

were centrifuged for 10 min at 4,000 rpm. The pellets

were suspended in 500 ll of 10 mM MgSO4, then

boiled for 10 min at 100�C, centrifuged for 5 min at

12,500 rpm, and stored at 4�C. Inserts were amplified

using M13 universal primers and ‘‘heat soak’’ PCR;

i.e., tubes containing water, Taq buffer, MgCl2, and

template DNA (from the MgSO4 stocks) were held at

94�C for 30 min before the dNTPs, primers, and Taq

DNA polymerase were added for PCR. Sequencing

was performed in an ABI 3130xl Genetic Analyzer

(Applied Biosystems, Inc., Carlsbad, CA, USA).

The similarities of the cloned oat sequences to

known gene sequences from GenBank were deter-

mined using BLAST (Basic Linear Alignment

Sequence Tool; Altschul et al. (1990)) from NCBI.

Only sequences with BLAST similarity to known

genes were considered further in this study. The

ClustalW algorithm within the program Megalign

(DNASTAR, Inc., Madison, WI, USA) was used to

align and compare sequences. The sequence variabil-

ity in oat was estimated by the frequency of SNPs

found among the parental lines using the program

SeqMan (DNASTAR, Inc., Madison, WI, USA).

RFLP analysis of a Vrn1 candidate gene clone

Restriction enzyme digestion and Southern blotting

were performed as described by Wight et al. (2003)

with some modification. The insert from the clone

Vrn1-T4 (As_ITA_21, isolated as described above;

Genbank accession HQ910536) was PCR-amplified,

then labeled with 32P using the Strip-EZ kit from

Ambion, Inc. (Austin, Texas, USA). Hybridization to

blots containing DNA from lines in the KO population

was performed using ULTRAhyb buffer (Ambion,

Inc.), according to the manufacturer’s instructions.

After washing, the blots were exposed to film at –70�C

for 7 days.

Mapping of Vrn3 candidate gene loci

The As-Vrn3 gene was mapped in the U86, U71P, and

KO mapping populations using the primer pair Ver-23

(50-GCAATGAGATGAGGACCTTCT-30 (forward)

and 50-CGCTGGCAGTTGAAGTAGAC-30 (reverse))

(Fig. S1). PCR reactions were performed in an

Eppendorf Mastercycler Gradient machine under the

following conditions: 3 min at 94�C, 40 cycles of 30 s

at 94�C, 45 s at 61�C, and 2 min at 72�C, then 10 min

at 72�C, and hold at 4�C.

SSR marker mapping

The oat SSR marker AM87 (primers: 50-GAGCAA

GCTCTGGATGGAAA-30 (forward) and 50-CCCGTT

TATGTGGTTGTTAGC-30 (reverse)) identified by

Pal et al. (2002) was mapped in the U86, U71P, and

KO populations in the manner described by Wight

et al. (2010).

Mol Breeding

123

DArT marker analysis

Protocols for DArT marker analysis were identical to

those described by Tinker et al. (2009) and were

performed by Diversity Arrays Technology Pty Ltd

(DArT P/L) (http://www.diversityarrays.com). Since

the populations used in this study were assayed during

the development of the oat DArT arrays, genotyping

was performed using the complete set of DArT dis-

covery arrays containing approximately 21,000 clones

rather than the reduced typing-arrays (approx. 2,500

clones) that are currently being used commercially by

DArT P/L.

Molecular marker scores were generated using the

software DArTsoft (DArT P/L). Briefly, the relative

hybridization intensity of each clone on each slide was

determined by dividing the hybridization signal in the

target channel (genomic representation) by the hybrid-

ization signal in the reference channel (polylinker).

Clones with variable relative hybridization intensity

across the slides were subjected to fuzzy k-means

clustering to convert relative hybridization intensities

into binary scores (presence vs. absence). Clones that

did not fit an expected bimodal (two-cluster) distribu-

tion were discarded from further analysis. Standard

methods of marker discovery were employed using a

combination of parameters automatically extracted

from the array data using the DArTsoft program: (1)

marker quality (Q), which measures between-cluster

variance as a percentage of total variance in fluores-

cent signal distribution among tested samples, (2)

marker call rate (percentage of effective scores), and

(3) polymorphism information content (PIC). The

markers reported in this paper were selected with

Q [ 73, call rate[80%, and PIC [ 0.1.

Linkage mapping

Redundancy of co-segregating markers within the

mapping data was removed as described in Tinker

et al. (2009). Markers were merged when they

segregated with [99% identity, or when they were

derived from clones with identical sequence and

segregated with[95% identity. Consensus scores for

merged markers were generated for single ‘surrogate’

markers which represented non-missing and non-

conflicting data for all component markers within a

cluster. Conflicting scores for markers with common

sequences were resolved by majority rules. If it was

not possible to generate a consensus score (e.g., if

conflicting scores were split equally among markers

with common sequences), then a missing value was

assigned.

Linkage maps were developed for each mapping

population using the program JoinMap 4.0 (Kyazma,

the Netherlands; van Ooijen and Voorips 2001).

Linkage groups were formed at LOD 5 and 40%

maximum recombination frequency, unless indicated

otherwise. Marker order was estimated using linear

regression, taking the shortest map as the best. These

maps were compared to each other and to the most

recent KO reference map (Tinker et al. 2009) using

C2maps, an enhancement of the software program M5

(Tinker 1999), available from the author.

QTL analysis

QTL were detected by simple interval mapping using

the program MQTL (Tinker and Mather 1995). The

experiment-wide false-positive rate for QTL main

effects was estimated based on 10,000 random

permutations, and QTL effects were estimated based

on a partial regression coefficient in a multi-locus

linear model as described by Tinker et al. (1996). The

proportions of phenotypic variance explained by each

QTL and by the final multi-locus model were

estimated using the R2 statistics in the individual and

multi-locus linear regressions.

Comparative mapping of identified QTL with

flowering time-related QTL from other studies in oat

was done using the Oatgenes database (Wight et al.

2006; http://avena.agr.gc.ca/oatgenes/). As the Oatg-

enes framework map is based on the KO map of Wight

et al. (2003), C2Maps was used to bridge this infor-

mation to the most recent version of the KO map

(Tinker et al. 2009).

Results

Field tests of early and late plantings

The number of days to flowering for the early and late

field plantings of the U86 and U71P populations is

presented in Fig. 1a, b. Days to flowering varied

between populations and planting dates. For the U86

population, days to flowering ranged from 33 to

48 days in the early planting and from 29 to 54 days in

Mol Breeding

123

the late planting, with small differences observed

between the parental lines UFRGS8 and UFRGS

930605. The average of the number of days to

flowering among RILs was 37 days for the early

planting and 34 days for the late planting. The

standard deviation of the number of days to flowering

among RILs was 2.3 for the early planting and

3.9 days for the late plantings. For the U71P popula-

tion, the number of days to flowering ranged from 34

to 48 days in the early planting and from 30 to 54 days

in the late planting. The parental lines UFRGS 881971

and Pc68/5*Starter showed a greater divergence in the

number of days to flowering compared to the parental

lines UFRGS8 and UFRGS 930605 (Fig. 1a, b). In this

population, the average of the number of days to

flowering among RILs was 37 and 36 days and the

standard deviation was 2.1 and 4.3 days for the early

and late plantings, respectively.

The response to early planting, estimated as the

difference in the number of days to flowering for the late

versus early plantings, varied among RILs for both

populations. For the U86 population, 85% of the RILs

showed a longer vegetative period after early planting.

For the U71P population, 57% of the RILs showed a

longer vegetative period after early planting, whereas

17% did not show differences between the planting

dates and 26% showed a shorter vegetative period after

early planting. The parental lines UFRGS 8 and UFRGS

881971 showed shorter vegetative periods after early

planting, UFRGS 930605 did not show a difference

0

10

20

30

40

50

60

70

80

29 -30

31 -32

33 -34

35 -36

37 -38

39 -40

41 -42

43 -44

45 -46

47 -48

49 -50

51 -52

53 -54

Nu

mb

er o

f R

ILs

Planted Early

Planted LateU605

(39 days)

U8(38 days)

A

U605(39 days)

U8(40 days)

Days to flowering

0

10

20

30

40

50

60

70

29 -30

31 -32

33 -34

35 -36

37 -38

39 -40

41 -42

43 -44

45 -46

47 -48

49 -50

51 -52

53 -54

Planted Early

Planted Late

Days to flowering

Nu

mb

er o

f R

ILs

B

U71(40 days)

U71(42 days)

Pc68(37 days)

Pc68(34 days)

0

10

20

30

40

50

6025 -

27

28 -30

31 -33

34 -36

37 -39

40 -42

43 -45

46 -48

49 -51

52 -54

55 -57

58 -60

U71(33 days)

Pc68(32 days)

Pc68(32 days)

U71(49 days)

Vernalized

Non-vernalized

Days to flowering

Nu

mb

er o

f R

ILs

D

0

10

20

30

40

50

60

70

25 -27

28 -30

31 -33

34 -36

37 -39

40 -42

43 -45

46 -48

49 -51

52 -54

55 -57

58 -60

Nu

mb

er o

f R

ILs

U8(26 days)

U605(30 days)

U8(28 days)

U605(45 days)

Vernalized

Non-vernalized

C

Days to flowering

Fig. 1 Distribution of days to flowering from the oat mapping

population U86 (a, c) and U71P b, d) in field tests (a, b) with

early (May 3) and late (May 25) plantings and grown in

greenhouse tests (c, d) from vernalized versus non-vernalized

plants. Means of parental genotypes grown adjacent to tests

containing progenies are indicated with arrows

Mol Breeding

123

between planting dates, and Pc68/5*Starter showed a

longer vegetative period after early planting (Fig. 1a, b).

Greenhouse tests of vernalized and non-vernalized

plants

Variation in the number of days to flowering under

vernalized versus non-vernalized treatments was

observed for both populations (Fig. 1c, d). For the

U86 population, the number of days to flowering

among RILs varied from 25 to 42 days for the

vernalized treatment and from 25 to 58 days for the

non-vernalized treatment. The average of the number

of days to flowering among RILs was 31 days for the

vernalized treatment and 35 days for the non-vernal-

ized treatment. The standard deviation was higher for

the non-vernalized treatment (7.5 days) than the

vernalized treatment (2.8 days). For the U71 popula-

tion, the number of days to flowering among RILs

varied from 28 to 39 days for the vernalized treatment

and from 27 to 57 days for the non-vernalized

treatment. The average of the number of days to

flowering was 33 and 37 days for the vernalized and

non-vernalized treatments, respectively. Similar to the

U86 population, the standard deviation in this popu-

lation was higher for the non-vernalized treatment

(6.8 days) in comparison to the vernalized treatment

(2.3 days).The parental lines UFRGS 8 and Pc68/

5*Starter did not show any substantial difference in

vegetative period under vernalized versus non-vernal-

ized treatments. However, the observed vegetative

period for the parental lines UFRGS 930605 and

UFRGS 881971 was 15 and 16 days shorter, respec-

tively, when vernalized. Response to vernalization

(non-vernalized minus vernalized) among the RILs

varied from -5 to 21 days for both populations.

Cloning of candidate gene sequences from oat

Oat genomic sequences associated with the vernali-

zation genes Vrn1, Vrn2, and Vrn3 were isolated based

on orthology with other grass species. To enhance the

probability of amplifying the target genes in oat, only

the coding regions of the consensus sequences for each

gene were used for primer design (Electronic Supple-

mentary Material Fig. S1). Of those that were tested,

seven primer pairs amplified oat sequences with

putative homology to the Vrn1, Vrn2, or Vrn3 genes

(Fig. S1).

A total of 21 oat amplicons targeting Vrn1,

generated using the primer pairs Ver1-1, Ver1-6,

Ver1-18, and Ver1-19 (Fig. S1), were cloned and

sequenced. These sequences ranged in length from

190 to 620 nucleotides (GenBank accessions

HQ910516 to HQ910536). When BLASTed against

the NCBI nr/nt database (http://blast.ncbi.nlm.nih.

gov/Blast.cgi, accessed 25 November 2011), the

HQ910516 sequence (derived using the primer pair

Ver1-1) showed 85–87% similarity to Vrn1 genes

from Lolium multiflorum, Festuca arundinacea, Loli-

um perenne, Festuca pratensis, Hordeum vulgare, and

Secale cereale. Similarly, the HQ910517 sequence

(derived using the primer pair Ver1-6), showed 78%

similarity to Vrn1 genes from Triticum timopheevii

and Triticum turgidum.

Sequences derived using the primer pair Ver1-18

(HQ910518–HQ910520) showed 98–100% similarity

to cDNA clones for the ‘Fruitful-like’ MADS-box

transcription factors isolated from Avena sativa and

A. strigosa (a diploid oat). The overlapping, longer

sequences derived using the primer pair Ver1-19

(HQ910521–HQ910536) showed 78–79% similarity

to Vrn-A1 genomic sequences from Triticum aestivum

and T. turgidum, Vrn1 genomic sequences from

T. monococcum and T. urartu, and Vrn-D1 sequences

from T. monococcum and Aegilops tauschii.

Two oat sequences derived from the Vrn2-based

primers were amplified using the Ver2-8 primer pair

(Fig. S1). BLAST analyses of these sequences

(HQ910537 and HQ910538) showed 89–90% similar-

ity to the Avena sativa LTR-retrotransposon OARE-1.

However, the ends of the sequences (ranging from 14 to

52 bases) showed no similarity to the retrotransposon.

Cloned oat sequences corresponding to the Vrn3

gene were isolated using the primer pairs Ver3-23 and

Ver3-24 (Fig. S1). For the 12 sequences that matched

the target gene, the trimmed sequence alignments

varied in length from 390 to 600 nucleotides (Gen-

Bank accessions HQ910539 to HQ910550). These

sequences showed 91–96% similarity to flowering

locus T (Ft3) genes from Lolium perenne, L. temulen-

tum, L. multiflorum, and Festuca pratensis, as well as

Vrn3 genes from Triticum turgidum (Vrn-B3), T. aes-

tivum (Vrn-B3), and T. monococcum (Vrn-Am3), and a

flowering time locus T-like protein 1 (Ft1) gene from

Hordeum vulgare.

The sequence variability in oat, estimated by the

frequency of SNPs among parents, was highest for the

Mol Breeding

123

Vrn3 gene. The highest number of SNPs was detected

between the parents UFRGS 881971 and Pc68/

5*Starter for both Vrn1 and Vrn3.

DArT marker analysis and linkage maps

A combined total of 464 DArT markers were scored in

one or both of the Brazilian populations. After

removing redundant and closely-linked markers, there

were 364 DArT markers scored in one or both

populations. Of these, 202 were also mapped in KO.

This included 171 markers scored in U86, 193 scored

in U71P, and 79 scored in both populations. In

addition to these DArT markers, the SSR AM87 was

scored in both populations. Partial scores from 20

AFLP markers from a previous study (Locatelli et al.

2006) were also included. The resulting linkage maps

for the U86 and U71P mapping populations are

presented as supplementary data (Figs. S2 and S3).

Molecular mapping of Vrn1 in hexaploid oat

None of the primer pairs designed to amplify the Vrn1

gene produced fragments that were polymorphic

amongst the parents used in this study. Therefore,

the clone Vrn1-T4 was used to probe Southern blots

representing lines from the KO population. This clone

contains a sequence amplified from the cultivar

TAMO-301 using the primer pair Ver1-19. The first

hit from BLAST analysis of the entire sequence was a

Vrn-A1 sequence from wheat (T. aestivum) (Fig. S4).

When the same sequence was trimmed to remove the

introns, the first hit was to the cDNA clone from A.

strigosa representing the ‘Fruitful-like’ MADS-box

transcription factor, and when the trimmed sequence

was translated, the first hit was to a K-box motif from

MADS-box transcription factor 15 in Zea mays (Fig.

S4).

When the Vrn1-T4 clone was used as an RFLP

probe, a dominant EcoRV polymorphism was

revealed, as was a second monomorphic band. The

polymorphic locus was found to be located on KO

group 22_44_18 (Fig. S5), which is homologous to

U86 group 1 and U71P group 2. Comparative mapping

with wheat and barley supports this as being the

location of a Vrn1 gene in oat (Fig. S5).

Since an EcoRV cut site was found in the oat Vrn1

sequences (Fig. S6), the possibility exists that both

RFLP bands represent only one locus. However, the

sequence alignments suggest that there is more than

one Vrn1 locus in hexaploid oat (Fig. S6). The most

likely known location for the second locus would be

KO group 24_26_34, which is homoeologous to KO

group 22_44_18 (Fig. S5). A third homoeologous

group has not yet been consistently identified in oat.

Molecular mapping of Vrn3 in hexaploid oat

As-Vrn3 was mapped in populations from the crosses

U86, U71P, and KO. Figure S7 shows the amplifica-

tion pattern of fragments generated using the Ver3-23

primer pair with the parental lines UFRGS 881971 and

Pc68/5*Starter, as well as eight RILs derived from this

cross. The same amplification pattern was observed in

U86 and KO. Mendelian segregation analysis showed

no deviation from the expected 1:1 segregation pattern

(v2 = 2.70, P = 0.1003) in the KO population, which

contains a mixture of F7 and F10-derived lines (Tinker

et al. 2009). Since the other two populations were F5-

derived, there was an expected segregation of 56:44 in

favour of the dominant locus. For the U86 mapping

population, this ratio was not rejected (v2 = 1.94,

P = 0.1637), while for U71P mapping population this

ratio was potentially rejected (v2 = 5.52, P = 0.0188)

with a slight upward bias in the frequency of the

dominant locus derived from UFGRS 881971. The

Ver3-23 primer pair was used to map As-Vrn3 as a

single locus in all three mapping populations (Fig. 2).

The labelling of three linkage groups in U71P as ‘‘4a,

4b, and 4c’’ (Fig. S3) reflects an initial uncertainty

about the joining of markers in these three groups.

When mapped at LOD 6, the three groups were

separated. When mapped at LOD 5, markers from all

three groups formed a single linear order that was very

consistent with the three separate groups placed end-

to-end. However, U71P-4a showed homology to KO

group 17 while U71P-4b matched with KO group 6.

Since these two KO groups are probably homoeolo-

gous (Tinker et al. 2009), the separate grouping of 4a

and 4b was considered correct. The small 4c linkage

group could be joined to 4b, but we chose to present it

separately for consistency with the modified LOD 6

threshold chosen for this node.

Locations and estimates of QTL

Several regions of the maps contained QTL affecting

flowering time in one or both Brazilian populations

Mol Breeding

123

(Table 1). The region on linkage group 6 of both maps

(Fig. 3) is highly significant in both populations. The

region represented by marker U86PM2, an AFLP

marker reported in a previous work as being highly

related to photoperiod response in the U86 population

(Locatelli et al. 2006) also affected flowering time in

U86. A QTL was also identified on U71P linkage

group 1, which is homologous to U86 groups 2, 3, and

4 and KO group 11_41_20_45 (Fig. 4).

In the U86 mapping population, both the QTL on

linkage group 6 as well as the QTL at U86PM2

(linkage group unknown) affect the number of days to

flowering for traits measured in the field and green-

house tests. Phenotypic variation observed for the

number of days to flowering for early (E) and late

(L) field-planted plots was associated with the molec-

ular markers opt-10121, AM87, opt-14149 (linkage

group 6), and U86PM2 (Table 1). For the QTL on

Fig. 2 Molecular mapping of the Vrn3 gene in oat on linkage

groups 10, 4, and 6 from the crosses UFRGS 8 9 UFRGS

930605 (U86), UFRGS 881971 9 Pc68/5*Starter (U71P), and

Kanota 9 Ogle (KO), respectively. Linkage groups 4a, 4b and

4c were separated for consistency with mapping methods, but

this figure shows that 4b and 4c could be joined based on

comparative mapping. Group map 4a is not shown because it

contains no common markers and may belong to a different

chromosome. Stacked horizontal lines show the locations of

additional markers placed on the KO map framework. Solidlines are used to join the Vrn3 loci from the three populations,

while dotted lines join markers found to be in common between

the linkage groups. A list of the markers joined by lines is

provided in Table S8. Boxes show the approximate locations of

QTL identified in previous studies, some of which are placed on

the KO map using comparative mapping (Wight et al. 2006).

The letters in each box indicate which reference should be

consulted for further information concerning the particular

QTL: H, Holland et al.1997; h, Holland et al. 2002; S,

Siripoonwiwat et al.1996; B, Beer et al. 1997; P, Portyanko

et al. 2005. The letter in bold text in a grey box highlights a QTL

for which the comparison-wide type I error rate was P \ 0.001.

The remaining markers had comparison-wide type I error rates

of p \ 0.05

Mol Breeding

123

linkage group 6, the allele from the parental line

UFRGS 8 increased the number of days to flowering

for both early (1.5 days) and late (3.0 days) planting,

while for the QTL at U86PM2, the allele from UFRGS

8 reduced the number of days to flowering for both

early (3.6 days) and late (5.3 days) plantings

(Table 1). However, the results of QTL analysis did

not show any significant genomic region controlling

the response to early planting (L - E) in the U86

population (Table 1).

The QTL on linkage group 6 and the one associated

with the marker U86PM2 were also associated with

phenotypic variation observed for the vernalized (V),

non-vernalized (NV) greenhouse-grown plants, and

for the response to vernalization (NV - V) in the U86

mapping population. Phenotypic variation for the

vernalized treatment was associated with the molec-

ular marker U86PM2, where the allele from the

parental line UFRGS 8 reduced the number of days to

flowering by 3.8 days (Table 1). Phenotypic variation

observed among the RILs for the non-vernalized

treatment was associated with the molecular markers

opt-10121, AM87, opt-11392, opt-12736, opt-4341,

opt-14149 (linkage group 6), and U86PM2 (Table 1).

Table 1 Summary of QTL affecting flowering time and vernalization response in the Avena sativa population UFRGS 8 9 UFRGS

930605 and UFRGS 881971 9 Pc68/5*Starter

Locus Linkage group Map position (cM) UFRGS 8 9 UFRGS 930605a

Flowering time E/Lb Flowering time V/NVc

E L L - E V NV NV - V

opt-10121 6 16 2.0* 3.0* -1.4 1.6 6.9* -5.4*

AM87 6 16.3 2.0* 3.0* -1.4 1.6 6.9* -5.4*

opt-11392 6 17.6 1.2ns 2.3 1.1 1.4 6.3* -5.0*

opt-12736 6 19.3 1.2 2.2 -1.1 1.6 5.8* -4.3*

opt-4341 6 21.6 0.9 1.9 -1.0 1.3 4.9* -3.8*

opt-14149 6 23.9 1.4 3.0* -1.6 1.5 6.5* -5.3*

U86PM2 Unknown -3.6* -5.3* 1.9 -3.8* -10.9* 8.0*

UFRGS 881971 9 Pc68/5*Startera

opt-7308 6 39.1 1.6* 5.3* -3.6* 0.9 0.1 0.8

opt-2102 6 39.2 1.6* 5.3* -3.6* 0.9 0.1 0.8

opt-1472 6 44.1 1.7* 4.6* -2.9* 1.2 2.5 -1.4

U71PM16 6 48.2 1.2 4.0* -2.8* 0.8 1.9 -1.3

AM87 6 52.8 0.9 3.6* -2.6* 0.9 2.2 -1.4

opt-15383 6 55.1 1.2 3.8* -2.6* 0.9 0.8 0.09

opt-18007 6 58.3 0.9 3.2* -2.3* 0.3 1.1 -1.1

opt-11295 6 60.3 0.9 3.7* -2.7* 0.5 0.3 0.02

opt-10163 1 47.9 -0.9 -1.4 0.5 -1.5* 1.2 -2.8

opt-13153 1 49.4 -0.9 -1.3 0.4 -1.6* 1.9 -3.8

opt-0096 1 52.6 -1.1 -1.6 0.4 -1.5* -0.02 -1.6

* Significant with experiment-wide type I error rate P \ 0.05 (significance of QTL was tested using 10,000 permutations of the data,

each giving maximum test statistics (TS) for a complete genome-wide scan using simple interval mapping); nsNot significant at the

0.05 probability levela Additive main effect (number of days increase caused by parent 1 relative to parent 2)b Flowering time evaluated for early (E) and late (L) field-planted plots. The column L—E represents the response to early plantingc Flowering time evaluated for vernalized (V) and non-vernalized (NV) greenhouse-grown plants. The column NV—V represents the

response to vernalizationd In both populations, the position of the effect on group 6 corresponds to a position near locus opt-18063 on Kanota 9 Ogle group

24. The locus opt-10121 showed the largest effects on UFRGS 8 9 UFRGS 930605 and the loci opt-7308 and opt-2102 showed the

largest effects on UFRGS 881971 9 Pc68/5*Starter

Mol Breeding

123

For the QTL on linkage group 6, the allele from the

parental line UFRGS 8 increased the number of days

to flowering on average by 6.2 days, whereas for the

QTL at U86PM2, the allele from UFRGS 8 reduced

the number of days to flowering by 10.9 days

(Table 1). Response to vernalization was associated

with the same markers detected for the non-vernalized

treatment, but with effects in the opposite direction.

In the U71P population, a QTL detected on linkage

group 6 affects the number of days to flowering for

early (E) and late (L) field-planted plots, and the

response to early planting (L - E). Variation

observed among RILs for early planting was associ-

ated with the molecular markers opt-7308, opt-2102,

and opt-1472 (Table 1). The number of days to

flowering in the early planting was increased on

average by 1.6 days when the RILs carry the allele

from the parental line UFRGS 881971 (Table 1). In

the late planting, phenotypic variation was associated

with molecular markers opt-7308, opt-2102, opt-1472,

U71PM16, AM87, opt-15383, opt-18007 and opt-

11295 (Table 1). The allele from the parent UFRGS

881971 increased the number of days to flowering on

average by 4.2 days, when plants were evaluated in

Fig. 3 Comparative map analysis of QTLs located on linkage

group 6 in the crosses UFRGS 8 9 UFRGS 930605 (U86) and

UFRGS 881971 9 Pc68/5*Starter (U71P). Markers linked to

QTL in these populations are shown in bold. Stacked horizontallines show the locations of additional markers placed on the map

framework. Solid lines are used to join markers that were linked

to QTL in this study and are found in common between the

linkage groups. Dotted lines join all other common markers. A

list of the markers joined by lines is provided in Table S8. The

three AM87 loci are highlighted using rectangles. The markers

in italics are found on KO groups 5_30 and 7_10_28 instead of

group 24_26_34. Square boxes show the approximate locations

of QTL identified in previous studies, some of which are placed

on the KO map using comparative mapping (Wight et al. 2006).

The letters in each box indicate which reference should be

consulted for further information concerning the particular

QTL: H, Holland et al. 1997; h, Holland et al. 2002; S,

Siripoonwiwat et al. 1996. The letters in bold text in grey boxesindicate QTL for which the experiment-wide type I error rate

was P \ 0.05 or the comparison-wide type I error rate was

P \ 0.001. The remaining markers had comparison-wide type I

error rates of P \ 0.05

Mol Breeding

123

the late planting (Table 1). The response to early

planting was associated with the same markers

affecting the number of days to flowering in the late

planting. However, the allele from the parental line

UFRGS 881971 reduced the number of days to

flowering on average by 2.9 days (Table 1).

Interestingly, the QTL on linkage group 6 did not

affect the number of days to flowering for the

vernalized (V) and non-vernalized (NV) treatments,

and the response to vernalization (NV - V) in the

U71P mapping population (Table 1). The phenotypic

variation observed for the vernalized treatment was

associated with the molecular markers opt-10163, opt-

13153, and opt-0096 on linkage group 1 (Table 1). For

this treatment, the allele from the parental line UFRGS

881971 reduced the number of days to flowering on

average by 1.5 days (Table 1).

The results from the current study were compared

to those from previous work on flowering time in oat.

Figure 3 illustrates how the QTL found on linkage

group 6 in both U86 and U71P are found in regions

homologous to KO group 24_26_34, which was

Fig. 4 Comparative map analysis of QTL located on linkage

group 1 in the cross UFRGS 881971 9 Pc68/5*Starter (U71P).

Stacked horizontal lines show the location of additional markers

placed on the map framework. Solid lines are used to join

markers that were linked to QTL in this study and are found in

common between the linkage groups. Dotted lines join all other

common markers. A list of the markers joined by lines is

provided in Table S8. Comparative mapping demonstrates that

QTL main effects identified and characterized on linkage group

1 of the U71P population (marker names in bold) showed

homology to linkage group 11_41_20_45 of the Kanota 9 Ogle

(KO) population. No QTL were found on U86 group 4. Boxesshow the approximate locations of QTL identified in previous

studies, some of which are placed on the KO map using

comparative mapping (Wight et al. 2006). The letters in eachbox indicate which reference should be consulted for further

information concerning the particular QTL: H, Holland et al.

1997; h, Holland et al. 2002; S, Siripoonwiwat et al. 1996. The

comparison-wide type I error rates for these QTLs were

P \ 0.05

Mol Breeding

123

previously found to contain many QTL associated

with flowering time. This is the same region found to

be homoeologous to the As-Vrn1 gene locus on KO

group 22_44_18 (Fig. S5). Because of the apparent

importance of the region on U86 and U71P groups 6

and KO group 24_26_34, a search was made for an

easily applied marker for this region. The SSR marker

AM87 was found to mark this region in all three

populations (Fig. 3), and this marker could be suitable

for marker-assisted breeding.

Figure 4 illustrates how the QTL found on U71P

group 1 map to the same region as the flowering time

QTL found in the homologous region on KO group

11_41_20_45 by Holland et al. (2002) and Siri-

poonwiwat et al. (1996). No QTL were found in the

corresponding region in the U86 population.

Discussion

Oat varieties differ in their patterns of growth and

development in response to vernalization. In a tem-

perate climate such as Ottawa, Canada, some winter

oat varieties will not flower at all when planted in the

spring, whereas other varieties are delayed to varying

degrees, depending on how early they are planted and

on how cold the soil is during germination. This

suggests that vernalization response is a quantitative

trait with multiple loci, multiple alleles, and consid-

erable environmental influence. None of the germ-

plasm used in this study showed obligate vernalization

dependence, yet it was known from previous studies

(Locatelli et al. 2008) that some exposure to cold

temperatures would accelerate flowering in lines such

as UFRGS 930605 and UFRGS 881971.

Natural variation in vernalization requirement in

wheat, barley, and other temperate cereals is mainly

associated with allelic differences in the Vrn1, Vrn2,

and Vrn3 vernalization genes. These genes are also

implicated in the response to photoperiod (Trevaskis

et al. 2007). The response to vernalization observed in

UFRGS 930605 and UFRGS 881971 in the green-

house tests of vernalized and non-vernalized plants

supports the proposition that these parental lines have

vernalization genes similar to those identified in

wheat, barley, and rye (Dubcovsky et al. 1998) but

with effects that are more facultative. The genetic

basis of vernalization in wheat and barley is very

similar, with Vrn-1 and Vrn-3 alleles conferring spring

growth habits (Yan et al. 2006), and dominant or

functional alleles of Vrn-2 conferring a winter growth

habit (Yan et al. 2004). Although some oat cultivars

grown in southern Brazil show a strong vernalization

requirement, the genetic control of the vernalization

response could be the result of different allele

combinations at any of the Vrn1, Vrn2, and Vrn3 loci.

Vrn1, Vrn2, and Vrn3 play important roles in key

regulatory pathways in response to vernalization and

photoperiod in cereal crops. In this study, the identi-

fication of oat sequences with high similarity to Vrn1

and Vrn3 sequences from wheat, barley and other

grass species has provided good evidence that these

genes are conserved in the oat genome. Oat sequences

derived from the Vrn2-based primers did not show

direct similarity to grass Vrn2 sequences, instead

showing maximum similarity to a retrotransposon

(OARE-1) derived from Avena sativa. This suggests

that a non-functional version of the Vrn2 gene,

disrupted by a retrotransposon insertion in the first

intron, may have been cloned.

The analysis of DArT markers allowed the gener-

ation of a large number of polymorphic markers in the

U86 and U71P mapping populations. These markers

were used to develop linkage maps, providing an

opportunity for the molecular mapping of the Vrn1 and

Vrn3 vernalization genes in oat and the identification

of markers linked to genomic regions (QTL) that

affect vernalization response and flower initiation in

oat. Furthermore, these markers allowed us to conduct

comparative genomics in oat and will be useful in

future studies for refining information about genes and

QTL that affect important agronomic traits.

The data from the field tests raises questions as to

whether or not we have detected any vernalization

response, or whether these data provide only infor-

mation about other genetic factors affecting flowering

time. For approximately 200 of the 296 RILs from the

two mapping populations U86 and U71, the number of

days to flowering was reduced when plants were

grown in the late Ottawa planting versus the early

Ottawa planting. This response can largely be attrib-

uted to environmental factors such as temperature and

photoperiod. Day length was longer and temperature

was higher for the later plantings, and both of these

factors can accelerate the rate of development and

flowering. However, minimum grass temperatures

ranging from 1 to 10�C were recorded on 19 nights

following the early planting at a nearby weather

Mol Breeding

123

station, and only one of these events coincided with

the late-seeded experiment. These temperatures may

have been marginal in their effect on vernalization,

and/or they may have affected each population

differently. In the U86 population, there was little

difference between parental phenotypes in early

versus late planting, and it seems possible that

vernalization differences were not a factor in this

population under field conditions; i.e., most genotypes

received adequate vernalization to overcome their

marginal requirements. This is supported by the lack

of QTL identified for the early-minus-late response in

this population (Table 1). In the U71P population,

there was also little difference between parental values

for early versus late response, but the direction of

effect differed between parents. The Pc68 parent

flowered faster under late-planted conditions (Fig. 1b)

while the U71 parent flowered slower under late-

planted conditions (Fig. 1b). Segregation of progenies

in the U71P population also revealed a QTL for early-

minus-late response, suggesting that vernalization

could have been a genetic factor affecting flowering

time under these conditions (Table 1). In summary,

the field tests have provided information about a

variety of QTL, one of which could be vernalization

dependent. However, this work illustrates that some

oat genotypes have a marginal requirement for

vernalization that is manifested only under complete

lack of cool temperatures.

In an attempt to eliminate the influence of temper-

ature fluctuations and photoperiod, tests were con-

ducted under controlled greenhouse conditions. Under

these conditions, there remained only subtle differ-

ences in experimental conditions, caused by the fact

that the vernalized plants were started earlier under

short days, and the non-vernalized plants were staged

to catch up with them. However, both treatments were

exposed to the same long day-lengths and temperature

regimes throughout the remaining stages of develop-

ment. The distribution of number of days to flowering

was similar in both the U86 and U71P populations in

the greenhouse tests (Fig. 1c, d). However, the higher

phenotypic variation among RILs for the non-vernal-

ized treatment suggests that vernalization played an

important role in the transition from vegetative to

reproductive development in these populations. The

effects of vernalization in the promotion of flowering

are also demonstrated by the shorter vegetative period

after vernalization for the parental lines UFRGS

930605 and UFRGS 881971 (Fig. 1c, d). Under the

controlled conditions, two QTL affecting vernaliza-

tion responses were identified in the U86 population

(linkage group 6 and the unlinked marker U86PM2)

and one QTL in the U71P population (linkage group

1). However, the QTL on group 1 of U71P was only

present under vernalized conditions, and did not affect

vernalization response, while the QTL on group 6 of

U71P (found under field conditions) was not a

significant genetic factor in the greenhouse. These

results suggest different genetic mechanisms affecting

vernalization response and flowering time in each of

these populations.

Although the QTL effects and expression patterns

differed between populations, the most consistent

QTL position was that on linkage group 6 of both

maps, corresponding to KO group 24_26_34 (Tinker

et al. 2009). This group is also homoeologous to KO

group 22_44_18, where the clone Vrn1-T4 was

mapped in this study. While this clone contains the

K-box motif found in all MIKC-type MADS-box

genes (Preston and Kellogg, 2006), the entire sequence

(including the introns) best matched the DNA

sequence from a wheat (T. aestivum) Vrn-A1 gene

(Fig. S4). This, plus the comparative mapping evi-

dence from wheat and barley (Fig. S5), indicates that

we have identified one of the Vrn1 loci in hexaploid

oat. Because of the homoeology between KO groups

22_44_18 and 24_26_34, it seems plausible that a

second, unmapped Vrn1 locus would be at the same

location as the QTL reported in this study, and this is

supported by the numerous other flowering-related

QTL that have been mapped to both of these positions

(Fig. S5). It is, therefore, possible that Vrn1 in oat

functions at both locations, and that QTL in oat are

located at either or both loci depending on the parental

alleles. It is not yet possible to identify which region of

the oat map contains the third homoeologous region,

which may also contain Vrn1-related QTL. It is also

possible that the QTL on group 6 of each population

are attributable to different loci, or to different alleles

of the same locus. This could explain why the effects

at this QTL were stronger in the greenhouse experi-

ments for U86 and in the field experiments for U71P.

The other two QTL (marker U86PM2 in U86, and

linkage group 1 in U71P) were not directly associated

with any candidate genes identified in this study.

However, the strong effects of U86PM2 on the

greenhouse-measured vernalization response suggest

Mol Breeding

123

that this may also be the location of a major

vernalization-related gene. The same locus was also

detected in earlier studies (Locatelli et al. 2006),

where it conferred earliness in both Brazil and Ottawa

in three populations. In that study, it was postulated

that the effect was related to photoperiod, since the

effect was stronger under Brazilian winter conditions

where photoperiod was short. However, current evi-

dence from the greenhouse tests where photoperiod

was not a factor suggests that the effect may be

primarily through differences in vernalization require-

ment. It is unfortunate that U86PM2 remained

unlinked to all of the DArT markers in U86. However,

the U86PM2 marker was mapped to linkage group 4c

in U71P (Fig. S3), and since none of the DArT markers

on U71P-4c segregated in U86, this suggests that the

unlinked status of U86PM2 in U86 is valid. Moreover,

since U71P-4c shows strong homology to linkage

group 17 in KO, this further validates the position of

the QTL reported by Locatelli et al. (2006). Linkage

groups KO-6 and KO-17 are suspected to be homo-

eologues, providing indirect evidence that a second

Vrn3 locus may exist on KO-17, where it may be

responsible for some or all of the reported QTL for day

length response and vernalization.

There is still considerable research needed to fully

understand genetic variability and mechanisms affect-

ing flowering time in oat. We have provided indirect

evidence in oat for the co-location of Vrn1 with

vernalization requirement, as well as for Vrn3 with

photoperiod sensitivity, and it is a logical proposition

that these mechanisms are conserved in cereals. While

there remain questions about why this evidence is

inconsistent among different populations, the

sequences and QTL mapping reported here will

provide information to assist in the refinement of this

knowledge. There is now an opportunity to use simple

PCR-based markers to assist in the selection of

germplasm with specific adaptations to flowering

time, and this can facilitate the use of more diverse

germplasm in oat breeding programs.

Acknowledgments We are grateful to Andrzej Kilian for

providing customized pre-commercial services for the oat DArT

marker assay used in this work, and for his excellent advice in

interpretation of results. We thank Steve Harrison and Paul

Murphy for useful advice on winter oats and vernalization

conditions. We thank Julie Chapados for the sequencing support

and Dr. Weikai Yan and Annick Gauthier for assistance with

planning and execution of fieldwork. A portion of this research

was made possible by generous support from Quaker Oats

Company (USA), Quaker Tropicana Gatorade (Canada), and the

Agriculture and Agri-Food Canada Matching Investment

Initiative. The authors also thank the Brazilian Council of

Scientific and Technological Development (CNPq) for the

financial support of this research and for the scholarship

provided for the first author.

References

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990)

Basic local alignment search tool. J Mol Biol 215:403–410

Beer SC, Siripoonwiwat W, O’Donoughue LS, Souza E, Mat-

thews D, Sorrells ME (1997) Associations between

molecular markers and quantitative traits in an oat germ-

plasm pool: can we infer linkages? J Ag Genomics (for-

merly JQTL) vol. 3, article 1 http://wheat.pw.usda.gov/

jag/papers97/paper197/indexp197.html. Published with

permission from CAB Intl. Accessed 05 Dec 2011

Burrows VD (1986) Breeding oats for food and feed: conven-

tional and new techniques and materials. In: Webster FH

(eds) Oats: chemistry and technology. American Associa-

tion of Cereal Chemists, Saint Paul, Minnesota, pp 13–46

Chouard P (1960) Variation and its relations to dormancy. Annu

Rev Plant Physiol 11:191–238

Corbesier L, Vicent C, Jang S, Fornara F, Fan Q, Searle I,

Giakountis A, Farrona S, Gissot L, Turnbull C, Coupland G

(2007) FT protein movement contributes to long-distance

signaling in floral induction of Arabidopsis. Science

316:1030–1033

Distelfeld A, Li C, Dubcovsky J (2009) Regulation of flowering

in temperate cereals. Curr Opin Plant Biol 12:178–184

Dubcovsky J, Lijavetzky D, Appendino L, Tranquilli G (1998)

Comparative RFLP mapping of Triticum monococcumgenes controlling vernalization requirement. Theor Appl

Genet 97:968–975

Dubcovsky J, Chen C, Yan L (2005) Molecular characterization

of the allelic variation at the Vrn-H2 vernalization locus in

barley. Mol Breed 15:395–407

Dubcovsky J, Loukoianov A, Fu D, Valarik M, Sanchez A, Yan L

(2006) Effect of photoperiod on the regulation of wheat ver-

nalization genes Vrn1 and Vrn2. Plant Mol Biol 60:469–480

Fu D, Szucs P, Yan L (2005) Large deletions within the first

intron in Vrn-1 are associated with spring growth habit in

barley and wheat. Mol Genet Genomics 273:54–65

Hall TA (1999) BioEdit: a user-friendly biological sequence

alignment editor and analysis program for Windows 95/98/

NT. Nucleic Acids Symp Ser 41:94–98

Hayama R, Yokoi S, Tamaki S, Yano M, Shimamoto K (2003)

Adaptation of photoperiodic control pathways produces

short-day flowering in rice. Nature 422:719–722

Helliwell CA, Wood CC, Robertson M, Peacock WJ, Dennis ES

(2006) The Arabidopsis FLC protein interacts directly in

vivo with SOC1 and FT chromatin and is part of a high-

molecular-weight protein complex. Plant J 46:183–192

Holland JB, Moser HS, O’Donoughue LS, Lee M (1997) QTLs

and epistasis associated with vernalization response in oat.

Crop Sci 37:1306–1316

Mol Breeding

123

Holland JB, Portyanko VA, Hoffman DL, Lee M (2002)

Genomic regions controlling vernalization and photope-

riod responses in oat. Theor Appl Genet 105:113–126

Huang T, Bohlenius H, Eriksson S, Parcy F, Nilsson O (2005)

The mRNA of the Arabidopsis gene FT moves from leaf to

shoot apex and induces flowering. Science 309:1694–1696

Iwaki K, Nishida J, Yanagisawa T, Yoshida H, Kato K (2002)

Genetic analysis of Vrn-B1 for vernalization requirement

by using linked dCAPS markers in bread wheat (Triticumaestivum L.). Theor Appl Genet 104:571–576

Karsai I, Szucs P, Meszaros K, Filichkina T, Hayes PM, Skinner

JS, Lang L, Bedo Z (2005) The Vrn-H2 locus is a major

determinant of flowering time in a facultative 9 winter

growth habit barley (Hordeum vulgare L.) mapping pop-

ulation. Theor Appl Genet 110:1458–1466

Kleinhofs A, Kilian A, Saghai Maroof MA, Biyashev RM,

Hayes PM, Chen FQ, Lapitan N, Fenwick A, Blake TK,

Kanazin V, Ananiev E, Dahleen L, Kudrna D, Bollinger J,

Knapp SJ, Liu B, Sorrells M, Heun M, Franckowiak JD,

Hoffman D, Skadsen R, Steffenson BJ (1993) A molecular,

isozyme and morphological map of the barley (Hordeumvulgare) genome. Theor App Genet 86:705–712

Locatelli AB, Federizzi LC, Milach SCK, Wight CP, Molnar SJ,

Chapados JT, Tinker NA (2006) Loci affecting flowering time

in oat under short-day conditions. Genome 49:1528–1538

Locatelli AB, Federizzi LC, Milach SCK, McElroy AR (2008)

Flowering time in oat: genotype characterization for photope-

riod and vernalization response. Field Crops Res 106:242–247

Michaels SD, Amasino RM (1999) FLOWERING LOCUS Cencodes a novel MADS domain protein that acts as a

repressor of flowering. Plant Cell 11:949–956

Michaels SD, Ditta G, Gustafson-Brown C, Pelaz S, Yanofsky

M, Amasino RM (2003) AGL24 acts as a promoter of

flowering in Arabidopsis and is positively regulated by

vernalization. Plant J 33:867–874

Moore G, Devos KM, Wang Z, Gale MD (1995) Grasses, line up

and form a circle. Curr Biol 5:737–739

Oliver RE, Lazo GR, Lutz JD, Rubenfield MJ, Tinker NA,

Wisniewski Morehead NH, Adhikary D, Jellen EN, Mau-

ghan PJ, Anderson JM, Brown Guedira GL, Chao S, Beattie

AD, Carson ML, Rines HW, Obert DE, Bonman JM,

Jackson EW (2011) Model SNP development for complex

genomes based on hexaploid oat using high-throughput 454

sequencing technology. BMC Genomics 12:77

Pal N, Sandhu JS, Domier LL, Kolb FL (2002) Development

and characterization of microsatellite and RFLP-derived

PCR markers in oat. Crop Sci 42:912–918

Portyanko VA, Hoffman DL, Lee M, Holland JB (2001) A

linkage map of hexaploid oat based on grass anchor DNA

clones and its relationship to other oat maps. Genome

44:249–265

Portyanko VA, Chen G, Rines HW, Phillips RL, Leonard KJ,

Ochocki GE, Stuthman DD (2005) Quantitative trait loci for

partial resistance to crown rust, Puccinia coronata, in cul-

tivated oat, Avena sativa L. Theor Appl Genet 111:313–324

Preston JC, Kellogg EA (2006) Reconstructing the evolutionary

history of paralogous APETALA1/FRUITFULL-like genes

in grasses (Poaceae). Genetics 174:421–437

Rozen S, Skaletsky HJ (2000) Primer3 on the WWW for general

user and for biologist programmers. Methods Mol Biol

132:365–386

Sheldon CC, Burn JE, Perez PP, Metzger J, Edwards JA,

Peacock WJ, Dennis ES (1999) The FLF MADS

box gene: a repressor of flowering in Arabidopsis reg-

ulated by vernalization and methylation. Plant Cell

11:445–458

Siripoonwiwat W, O’Donoughue LS, Wesenberg D, Hoffman

DL, Barbosa-Neto JF, Sorrells ME (1996) Chromosomal

regions associated with quantitative traits in oat. J Ag

Genomics (formerly JQTL) Vol. 2, article 3 http://wheat.pw.

usda.gov/jag/papers96/paper396/indexp396.html Published

with permission from CAB Intl. Accessed 05 Dec 2011

Sorrells ME, Simmons SR (1992) Influence of environment on

the development and adaptation of oat. In: Marshal HG,

Sorrells ME (eds) Oat science and technology. Madison,

Wisconsin, pp 115–163

Tamaki S, Matsuo S, Wong HL, Yokoi S, Shimamoto K (2007)

Hd3a protein is a mobile flowering signal in rice. Science

316:1033–1036

Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W:

improving the sensitivity of progressive multiple sequence

alignment through sequence weighting, position-specific

gap penalties and weight matrix choice. Nucleic Acids Res

22:4673–4680

Tinker NA (1999) Management of multiple molecular marker

maps with multiple molecular marker map manager

(Mmmmm!). J Agric Genomics 4. http://wheat.pw.usda.

gov/jag/papers99/paper499/indexp499.html Published with

permission from CAB Intl. Accessed 05 Dec 2011

Tinker NA, Mather DE (1995) MQTL: software for simplified

composite interval mapping of QTL in multiple environ-

ments. J Quant Trait Loci (now J Agric Genomics) 1.

http://wheat.pw.usda.gov/jag/papers95/paper295/

indexp295.html. Published with permission from CAB Intl.

Accessed 05 Dec 2011

Tinker NA, Mather DE, Rossnagel BG, Kasha KJ, Kleinhofs A,

Hayes PM, Falk DE, Ferguson T, Shugar LP, Legge WG,

Irvine RB, Choo TM, Briggs KG, Ullrich SE, Franckowiak

JD, Blake TK, Graf RJ, Dofing SM, Saghai Maroof MA,

Scoles GJ, Hoffman D, Dahleen LS, Kilian A, Chen F,

Biyashev RM, Kudrna DA, Steffenson BJ (1996) Regions

of the genome that affect agronomic performance in two-

row barley. Crop Sci 36:1053–1062

Tinker NA, Kilian A, Wight CP, Heller-Uszynska K, Wenzl P,

Rines HW, Bjørnstad A, Jannink J-L, Anderson JM,

Rossnagel BG, Stuthman DD, Sorrells ME, Jackson EW,

Tuvesson S, Kolb FL, Olsson O, Federizzi LC, Carson ML,

Ohm HH, Molnar SJ, Scoles GJ, Eckstein PE, Bonman JM,

Ceplitis A, Langdon T (2009) New DArT markers for oat

provide enhanced map coverage and global germplasm

characterization. BMC Genet 10:39–61

Trevaskis B, Bagnall DJ, Ellis MH, Peacock WJ, Dennis ES

(2003) MADS box genes control vernalization-induced

flowering in cereals. Proc Natl Acad Sci USA 100:

13099–13104

Trevaskis B, Hemming MN, Peacock WJ, Dennis ES (2006)

HvVrn2 responds to daylenght, whereas HvVrn1 is regu-

lated by vernalization and developmental status. Plant

Physiol 140:1397–1405

Trevaskis B, Hemming MN, Dennis ES, Peacock WJ (2007)

The molecular basis of vernalization-induced flowering in

cereals. Trends Plant Sci 12:352–357

Mol Breeding

123

Turner A, Beales J, Faure S, Dunford RP, Laurie DA (2005) The

pseudo-response regulator Ppd-H1 provides adaptation to

photoperiod in barley. Science 310:1031–1034

van Ooijen JW, Voorrips RE (2001) JoinMap. Software for the

calculation of genetic linkage maps 3.0. http://www.kyazma.

nl/index.php/mc.JoinMap. Accessed 12 Dec 2010

von Zitzewitz J, Szucs P, Dubcovsky J, Yan L, Francia E,

Pecchioni N, Casas A, Chen TH, Hayes PM, Skinner JS

(2005) Molecular and structural characterization of barley

vernalization genes. Plant Mol Biol 59:449–467

Wight CP, Tinker NA, Kianian SF, Sorrells ME, O’Donoughue

LS, Hoffman DL, Groh S, Scoles GJ, Li CD, Webster FH,

Phillips RL, Rines HW, Livingston SM, Armstrong KC,

Fedak G, Molnar SJ (2003) A molecular marker map in

‘Kanota’ 9 ‘Ogle’ hexaploid oat (Avena spp.) enhanced

by additional markers and a robust framework. Genome

46:28–47

Wight CP, Bergeron D, Tinker NA (2006) Oatgenes: a com-

prehensive database of oat markers and QTLs. Proc. 2006

American Oat Workers’ Conference, July 23-26, Wil-

mington, NC, USA. http://wheat.pw.usda.gov/ggpages/

oatnewsletter/v50/AOWC/Oat_book5.pdf. Accessed 05

Dec 2011

Wight CP, Yan W, Mitchell Fetch J, Deyl J, Tinker NA (2010) A

set of new simple sequence repeat and avenin DNA

markers suitable for mapping and fingerprinting studies in

oat (Avena spp.). Crop Sci 50:1207–1218

Xue WY, Xing YZ, Weng XY, Zhao Y, Tang W, Wang L, Zhou

H, Yu S, Xu C, Li X, Zhang Q (2008) Natural variation in

Ghd7 is an important regulator of heading date and yield

potential in rice. Nat Genet 40:761–767

Yan L, Loukoianov A, Tranquilli G, Helguera M, Hahima T,

Dubcovsky J (2003) Positional cloning of the wheat ver-

nalization gene Vrn1. Proc Natl Acad Sci USA 100:

6263–6268

Yan L, Loukoianov A, Blechl A, Tranquilli G, Ramakrishna W,

SanMiguel P, Bennetzen JL, Echenique V, Dubcovsky J

(2004) The wheat Vrn2 gene is a flowering repressor down-

regulated by vernalization. Science 303:1640–1644

Yan L, Fu D, Li C, Blechl A, Tranquilli G, Bonafede M, Sanchez

A, Valarik M, Dubcovsky J (2006) The wheat and barley

vernalization gene Vrn3 is an orthologue of FT. Proc Natl

Acad Sci USA 103:19581–19586

Zadoks JC, Chang TT, Zonzak CF (1974) A decimal code for the

growth stages of cereals. Weed Res 14:415–421

Mol Breeding

123