Content uploaded by Wayne Parrott
Author content
All content in this area was uploaded by Wayne Parrott on Jan 06, 2016
Content may be subject to copyright.
1260 WWW.CROPS.ORG CROP SCIENCE, VOL. 50, JULY–AUGUST 2010
RESEARCH
W , Trifolium rep ens L., is an allotetraploid
(2n=4x = 32) legume that is believed to have resulted
from the hybridization of T. occidentale Coombe (Ellison et al.,
2006) and a second, currently unknown Tr ifolium species (Hand
et al., 2008). Pollination in this species is controlled by a game-
tophytic self-incompatibility system that Atwood (1942) deter-
mined was regulated by a single locus with many di erent alleles,
including a rare allele that confers self-compatibility (Sf). Due
to the tetraploid genome and outcrossing nature of the species,
white clover is highly heterozygous.
There are many di erent leaf marks and other morphological
traits found within white clover, many of which have been the
subject of genetic studies. For consistency, the original genetic
nomenclature has been maintained for each trait, but the notation
has been modernized to that of Quesenberry et al. (1991). The
lack of leaf mark (Fig. 1a) is recessive to the presence of all leaf
marks (Brewbaker, 1955; Carnahan et al., 1955). The most com-
mon leaf mark is the multiallelic white V mark (gene symbol V;
Fig. 1b) on the upper epidermis of each lea et (Brewbaker, 1955).
This trait is highly variable, with a range of marks from a single V
mark to a V mark with a yellow tip (Vby; Fig. 1c). The marginal
Leaf Trait Coloration in White Clover
and Molecular Mapping of the Red Midrib
and Lea et Number Traits
Rebecca M. Tashiro, Yuanhong Han, María J. Monteros, Joseph H. Bouton, and Wayne A. Parrott*
ABSTRACT
White clover (Trifolium repens L.) is a highly out-
crossing heterozygous allotetraploid species,
for which classic inheritance studies have been
inconclusive. With the aid of molecular markers,
it is now possible to study the genes control-
ling morphological traits. The objectives of this
study were to catalog the leaf marks in white
clover and map the location of leaf morphologi-
cal traits based on cosegregation with molecu-
lar markers. A mapping population segregating
for eight morphological traits consisting of leaf
marks and number of lea ets was developed
and phenotyped at two different locations dur-
ing the summer and winter seasons. A con r-
mation population, derived by sel ng one of the
mapping population parents, was produced
and phenotyped in one location at two different
times of the year. Through the use of previously
published simple sequence repeat (SSR) marker
maps, linkages between the mapped molecular
markers and genes for three different morpho-
logical traits was identi ed. The red midrib and
red eck traits were found to be controlled by
two closely linked dominant genes on linkage
group (LG) B1. The trifoliolate trait is controlled
by at least one gene on LG H1. The identi ca-
tion of molecular markers linked to loci affect-
ing leaf morphology traits resolves con icting
hypotheses on the genetics of these complex
traits and has potential for molecular breeding
improvement of white clover.
R.M. Tashiro and W.A. Parrott, Institute of Plant Breeding, Genet-
ics, and Genomics, Univ. of Georgia, 111 Riverbend Rd., Athens,
GA 30602; Y. Han, M.J. Monteros, and J.H. Bouton, Samuel Rob-
erts Noble Foundation, Forage Improvement Division, Ardmore, OK
73401. Received 20 Aug. 2009. *Corresponding author (wparrott@
uga.edu).
Abbreviations: LG, linkage group; LOD, logarithm of the odds; PCR,
polymerase chain reaction; SSR, simple sequence repeat.
Published in Crop Sci. 50:1260–1268 (2010).
doi: 10.2135/cropsci2009.08.0457
Published online 20 Apr. 2010.
© Crop Science Society of America | 5585 Guilford Rd., Madison, WI 53711 USA
All rights reserved. No part of this periodical may be reproduced or transmitted in any
form or by any means, electronic or mechanical, including photocopying, recording,
or any information storage and retrieval system, without permission in writing from
the publisher. Permission for printing and for reprinting the material contained herein
has been obtained by the publisher.
CROP SCIENCE, VOL. 50, JULY–AUGUST 2010 WWW.CROPS.ORG 1261
mark (Vm; Fig. 1d) (Lenoble and Papineau, 1970) is rarely
seen in naturalized populations.
Other leaf marks found in white clover contain antho-
cyanin pigments. Some examples are the redspot leaf mark
(Vr2; Fig. 1e) (Hovin and Gibson, 1961), red lea et mark
(Vrl; Fig. 1f) (Corkill, 1971), red midrib mark (Rm; Fig.
1g) (Carnahan et al., 1955; Corkill, 1971), red leaf mark
(Rl; Fig. 1h) (Carnahan et al., 1955; Corkill, 1971), and red
eck mark (Rf; Fig. 1i) (Carnahan et al., 1955). The iden-
ti cation of these leaf marks is based on somewhat vague
written descriptions in the literature and on a black and
white photograph published by Corkill (1971). Both the
descriptions and the photograph are inadequate to prop-
erly identify these marks.
Although it is agreed that all leaf marks described
above are dominant traits, there is disagreement regarding
the genetic control of these traits. Carnahan et al. (1955)
and Brewbaker (1955) concluded that the presence of the
various leaf marks is controlled by one of two di erent
genes (V and R) that each contain multiple alleles, such
that Rm, Rl, and Rf would all be di erent alleles of the
R locus. In contrast, Corkill (1971) observed low recom-
bination frequencies between the leaf marks within each
locus and concluded accordingly that the R and V loci
each consist of a series of tightly linked genes. Thus, under
the Corkill hypothesis, Rm, Rl, and Rf represent linked
(but di erent) loci, collectively known as the R locus.
Clover leaves di er not only in their leaf marks but
also in their number of lea ets. White clover typically
has trifoliolate leaves (Fig. 2a), but multifoliolate (greater
than three lea ets per leaf ) genotypes exist within natural-
ized populations. The multifoliolate leaves of white clover
are traditionally collected as good luck charms, with the
four-leaf clover (Fig. 2b) recognized worldwide as a sym-
bol of good fortune. When this trait is combined with a
mutant elongated petiolule, the leaf morphology is altered
from palmate to pinnate (Fig. 2c). Despite the popularity
of the four-lea et trait in clover, it has not been possible
to determine its genetic control. Ford and Claydon (1996)
determined that the trait was mostly recessive but were not
able to observe any Mendelian segregation in the progeny.
Thus, the information available on the genetic control of
the multifoliolate trait is limited to that available from other
Trifolium species. Knight (1969) studied the multifoliolate
trait in crimson clover (T. incar natum L.) and found that
there were two types of multifoliolate leaves: one that was
strongly in uenced by environmental variation and one
that was not. The environmentally conditioned multifo-
liolate trait inheritance could not be determined, but the
non-environmentally controlled multifoliolate trait was
found to be a single gene recessive trait (Knight, 1969). In
red clover (T. pratense L.), the multifoliolate trait was rst
studied by Simon (1962), who determined it was condi-
tioned by homozygous recessive alleles at one of two loci.
Also studying red clover, Jaranowski and Broda (1978)
determined that the multifoliolate trait was controlled by
homozygous recessive alleles at three loci, and Taylor (1982)
determined it was a quantitative recessive trait.
White clover genetics are complicated by allotetraploidy,
extensive heterozygosity, and a highly outcrossing reproduc-
tive system. Therefore, homozygous lines are not available for
inheritance studies. Furthermore, many of the morphologi-
cal traits under study are highly in uenced by environment.
For example, many of the traits containing anthocyanin are
best observed at temperatures below 10°C (Carnahan et al.,
1955). As a result, some traits such as the red lea et (Vrl) trait
are not visible in the summer (Davies, 1963). In addition, the
multifoliolate trait in white clover was found to be environ-
mentally conditioned in a germplasm source registered by
Baltensperger et al. (1991), which supports Knight’s (1969)
observations in crimson clover. By looking at these traits at
the molecular level, the environmental e ects on each trait
can be separated from the gene itself. As such, mapping mor-
phological traits found in white clover with molecular mark-
ers may be a more e ective way to determine the inheritance
of these traits, many of which have been studied for nearly a
century without satisfactory conclusions.
The development of white clover genetic maps based
on molecular markers (Barrett et al., 2004; Jones et al.,
2003; Zhang et al., 2007) has allowed some important
agronomic traits to be mapped (Barrett et al., 2005a, b;
Cogan et al., 2006) and macrosynteny between white clo-
ver linkage groups (LGs) and chromosomes of the model
legume Medicago truncatula Gaertn. to be determined
(George et al., 2008). In the 2004 white clover simple
sequence repeat (SSR) map, the red eck mark (Rf) was
mapped as the R locus onto LG B1 (Barrett et al., 2004).
The parents used to create the mapping population in that
study were forage genotypes and, as such, had limited
morphological diversity for additional leaf marks.
The objectives of the research described here were to
inventory the leaf marks expressed in white clover (many of
which are shown in Fig. 1) and map the location of genes for
leaf morphological traits based on cosegregation with molec-
ular markers from an existing white clover linkage map.
MATERIALS AND METHODS
Plant Materials
Two phenotypically distinct white clover genotypes were used
as parents to develop a mapping population (Fig. 3). One par-
ent, GA02-56 (hereafter referred to as GA43 in keeping with
its name in the literature), is an agronomic genotype out of the
cultivar ‘Durana’ (Bouton et al., 2005) that was also used as a
parent for construction of the genetic map of Zhang et al. (2007).
This genotype has trifoliolate green leaves, the intermediate
white V mark (Vi), and the red eck leaf mark (Rf ). The second
genotype, 05-O-34, contains several traits of ornamental value
(Tash iro et al., 2009). This genotype has multifoliolate leaves, the
1262 WWW.CROPS.ORG CROP SCIENCE, VOL. 50, JULY–AUGUST 2010
Fig. 1. Different leaf marks found in white clover. The gene symbols are as originally proposed by the authors that described them, but
the notation has been modernized as described by Quesenberry et al. (1991).
Fig. 2. Different leaf morphologies found in white clover. The gene symbols for the leaf marks present on e ach leaf are indicated in parentheses.
Fig. 3. The white clover genotypes used as parents to create the mapping population. a) Ornamental-type parent 05-O-34; b) Agronomic-
type parent GA43.
CROP SCIENCE, VOL. 50, JULY–AUGUST 2010 WWW.CROPS.ORG 1263
marginal mark (Vm), red lea et (Vrl), red midrib (Rm), and red
eck (Rf ) leaf marks. This genotype is also self-compatible (Sf).
A reciprocal pseudo-testcross mapping population (Grattapaglia
and Sedero , 1994) was made consisting of 178 F1’s resulting
from reciprocal crosses between the two parents. Due to the self-
compatibility present in 05-O-34, the F1 progeny derived from
its seed were tested for hybridity by using 40 SSR marker prim-
ers. Only those individuals that had markers derived from both
parents were used in the mapping population. Individuals that
were the result of sel ng were incorporated into a con rmation
population that was developed by sel ng the 05-O-34 parent,
resulting in a total of 141 individuals.
Morphological Trait Evaluation
The 178 individuals in the mapping population were grown in
12-cm pots using potting mix made up of equal parts Fafard #3
potting soil (Conrad Fafard, Inc., Agawam, MA), river sand,
and farm soil [Cecil sandy clay loam (clayey, kaolinitic, thermic,
Typic Kanhapludults)] and grown in a University of Georgia
greenhouse until the foliage from each individual had lled out
the pot. Each individual was scored for the presence and/or
absence of each morphological trait in the greenhouse on 22
Aug. 2007 and 29 Mar. 2008 (Table 1). For lea et number,
individuals were scored as either being trifoliolate or multi-
foliolate. An individual with at least one multifoliolate leaf on
the evaluation date was scored as multifoliolate. The 141 indi-
viduals that were obtained by sel ng the 05-O-34 parent were
grown as described above, scored for each trait in the green-
house on 12 Aug. 2008 and 31 Mar. 2009 (Table 1), and used
for con rmation of mapped traits and hypothetical genotypes.
Cuttings of both parents and each individual in the map-
ping population were obtained and used for replicated eld
trials. Rooted cuttings of both parents and 140 individuals in
which 05-O-34 was the maternal parent were planted at the
University of Georgia Plant Sciences Farm (Oconee County,
GA) in Cecil sandy clay loam (clayey, kaolinitic, thermic, Typic
Kanhapludults) soil with a pH of 5.9. Of these 140 individuals,
89 were used for trait mapping. The ramets were planted on
75-cm centers in a randomized complete block design, with
four blocks of each genotype, on 6 Dec. 2007. Rooted cut-
tings of both parents and 89 individuals in which GA43 was
the maternal parent were planted at the Plant Sciences Farm on
18 Apr. 2008 with the same experimental design as described
above. Morphological data for each individual in each block
were scored for each trait on 2 July 2008 (Table 1). Individu-
als for which 05-O-34 was the maternal parent (89) were also
scored on 26 Mar. 2008, and those for which GA43 was the
maternal parent (89) were scored on 19 Mar. 2009 (Table 1).
SSR Amplifi cation and Amplicon Detection
DNA was extracted from young leaves of both parents and each
genotype in the mapping population using the Plant DNeasy
Mini Kit (Qiagen, Valencia, CA). DNA quanti cation for
each sample was performed using a TBS-100 mini- uorometer
(Turner Biosystems, Sunnyvale, CA). After quanti cation, each
sample was diluted to 10 ng L–l and treated with 0.05 U Longlife
RNase (G Biosciences, Maryland Heights, MO). From the origi-
nal 343 primer pairs used by Zhang et al. (2007) to create their
linkage map, 96 were selected based on their even distribution
in the di erent LGs and screened for polymorphism between the
two parents of the current mapping population.
A total of 78 primer pairs (81%) were polymorphic, which
translates to a marker spacing of around 20 cM, with between
3 and 6 SSR markers per LG. Fluorescently labeled SSR frag-
ments were ampli ed as described by Zhang et al. (2007) using
either 96- or 384-well polymerase chain reaction (PCR) plates,
with the exception of the source of the PCR reagents, which
were obtained from Promega (Madison, WI). After PCR,
plates with di erent uorescent tags were pooled together for
fragment analysis using the ABI PRISM 3730 Genetic Ana-
lyzer (Applied Biosystems, Foster City, CA) also as described
by Zhang et al. (2007). Simple sequence repeat fragments were
visually scored with GeneMapper 3.7 or 4.0 software (Applied
Biosystems, Foster City, CA) as dominant markers as described
by Zhang et al. (2007). Once the initial LGs were drawn as
described below, an additional 48 primer pairs selected from
the LGs of interest were screened for polymorphisms between
the mapping population parents. From the 48 additional primer
pairs evaluated, 41 (85%) of them were polymorphic. Twenty-
four of the 41 polymorphic primer pairs were selected based on
the number and quality of alleles scored in the parents and used
to screen the mapping population as described above. The addi-
tional scored alleles were then added to the previously screened
marker and phenotypic data to develop LGs with enhanced
marker saturation. Markers found to be linked to phenotypic
data in the mapping population were screened as described
above in the con rmation population.
Linkage Map Development
Linkage maps were developed using the Kosambi mapping func-
tion of JoinMap 3.0 (Van Ooijen and Voorrips, 2001) and drawn
using MapChart 2.2 (Voorrips, 2002) with each locus coordinate
rounded to the nearest whole number. Simple sequence repeat
fragments that segregated in a 1:1 ratio in the mapping popula-
tion were used to create single-parent LGs for each parent as
described by Zhang et al. (2007). Consensus maps for LG B1
were developed using bridging loci as described by Barrett et
al. (2004). The leaf morphology traits in white clover leaves are
either present or absent and therefore were mapped as qualitative
traits. All the traits except red eck leaf mark (Rf) are present
only in the ornamental parent and were coded as np × nn as per
software instructions for markers segregating in the rst parent.
Because the red eck leaf mark (Rf) is present in both mapping
parents and segregating in the mapping population, it was coded
as hk × hk as per software instructions for markers segregating
in both parents. In an e ort to reduce the e ects of genotype
× environment interaction on morphological trait expression in
white clover, data for the traits were collected and analyzed by
the di erent date and location combinations separately. When
mapping the morphological traits, those showing no obvious
environmental in uence were pooled and the data were mapped
as a single dominant trait. The traits showing strong environ-
mental e ects were mapped separately based on each individual
date (summer and winter) and location ( eld or greenhouse)
combination. After initial linkage map development, the addi-
tional marker data were added to the original data and linkage
maps were created as described above with a logarithm of the
odds (LOD) score ≥ 5.0.
1264 WWW.CROPS.ORG CROP SCIENCE, VOL. 50, JULY–AUGUST 2010
In the con rmation population that was developed by self-
ing of parent 05-O-34, segregating markers were used to cre-
ate LGs using the F2 population type in JoinMap 3.0. As in
the mapping population, each phenotypic marker was either
pooled or mapped individually based on evaluation date and
location. Linkage maps were created as described above based
on an LOD score ≥ 5.0.
RESULTS
During phenotypic evaluations of the mapping and con r-
mation populations it was noted that, while the red eck
(Rf) trait was observed by itself, whenever the red midrib
(Rm) trait was present, the red eck trait was always also
present. Since the inheritance of R locus traits in white clo-
ver was unclear, the two R locus traits evaluated in this
study were tested using both the Carnahan et al. (1955)
single-gene hypothesis and the Corkill (1971) linked-gene
hypothesis. In both populations, the inheritance data for
the two traits controlled by the R locus, red midrib (Rm)
and red eck (Rf), failed to conform to a model in which
a single gene controls both phenotypic characters (data not
shown). Instead, the two traits were found to be controlled
by dominant alleles at two di erent genes which seem to
be simply inherited and tightly linked (r = 0.23%) based on
frequency of recombination (Fehr, 1987) for the two genes
in the con rmation population (Table 1). Expression of the
red midrib (Rm) trait was stable across all environments,
with the exception of two genotypes in the mapping popu-
lation (Table 2). Since this was most likely due to scoring
error, the two genotypes were scored as missing data for
this trait. Expression of the red eck (Rf) trait was also rela-
tively stable across most environments, except during the
summer greenhouse data collection period, when expres-
sion was signi cantly lower (Table 3). Data for this trait t
a single gene model when p ≥ 0.05 across all environments,
except in the case of summer evaluations of the mapping
population in the greenhouse but did not t a single gene
model when p ≥ 0.1 for half of the environments. There-
fore, there is less con dence about a single gene inheritance
model for the red eck (Rf) trait than there is for the red
midrib (Rm) trait. Future studies in controlled environ-
ments should determine whether the growth temperature
is causing the variation in segregation between the crossed
population and the selfed population. In an outcrossing spe-
cies such as white clover, it is common to see skewed seg-
regation patterns (Barrett et al., 2005a). The outcrossing
nature of white clover could also explain the fewer-than-
expected recessive genotypes and the increase in heterozy-
gotes observed in the selfed population.
The gene conditioning the red midrib (Rm) trait is
linked to markers segregating in the ornamental parent on
what corresponds to LG B1 (Fig. 4a) of the map described
by Zhang et al. (2007). Phenotypic data for this trait were
pooled for mapping due to their high penetrance within the
population. Since the red eck (Rf) trait is segregating in
the mapping population but inherited from both the orna-
mental and forage parents, this trait was mapped by creating
a consensus linkage map with markers segregating in both
parents (Fig. 4b). The gene conditioning the red eck (Rf)
trait is also linked to markers segregating on LG B1.
The data for both red midrib (Rm) and red eck (Rf)
traits were each originally mapped separately based on
date and location. The separate data for each trait mapped
to a similar location on the same LG (data not shown).
Since the phenotypic data for the red eck trait mapped
so closely on the same LG, the data were pooled and the
linkage map recreated. The pooled red eck (Rf) trait data
mapped 1 cM above the red midrib (Rm) trait data on
LG B1 of the consensus map (Fig. 4b). The two traits are
anked on either side by molecular markers ats041 and
RCS3084. In the con rmation population, both traits
also mapped to LG B1, with both traits mapping to the
same location on the LG (Fig. 4c). In the con rmation
population, the two traits are anked on each side by
molecular markers BG232 and ats099. In both cases, the
morphological traits were mapped to the interval anked
by the common markers ats075 and ats099.
The phenotypic data for lea et number in both popu-
lations showed strong environmental in uence. Therefore,
Table 1. Frequencies of the white clover leaf marks observed within the F1 mapping population (reciprocal pseudo-testcross
between 05-O-34 × GA43) and the S1 confi rmation population (selfi ng of ornamental parent 05-O-34).
F1 mapping population† S
1 confi rmation population
Tra it
Summer
fi eld
Summer
greenhouse
Winter
fi eld
Winter
greenhouse
Summer
greenhouse
Winter
greenhouse
Intermediate white V (Vi)9396 8998 3941
Red fl eck (Rf) 122 107 113 123 113 116
Red midrib (Rm) 89 87 81 87 111 110
Red leafl et (Rl)304031143
Marginal mark (Vm)81806880101100
Tri foliolate leaf 44 71 96 101 26‡55
Multifoliolate leaf 134 107 62 77 114 85
†Mapping population size consisted of 178 individuals, except during the winter fi eld phenotyping, in which 20 individuals had died in all four blocks, so the population con-
sisted of 158 individuals at this date.
‡In the confi rmation population, leafl et number data failed to be recorded for one individual during each phenotyping date.
CROP SCIENCE, VOL. 50, JULY–AUGUST 2010 WWW.CROPS.ORG 1265
Table 2. Segregation of the white clover red midrib (Rm; Fig. 1g) trait within the F1 mapping population (reciprocal pseudo-
testcross between 05-O-34 × GA43) and the S1 confi rmation population (selfi ng of ornamental parent 05-O-34). Chi-squared
(χ2) goodness-of fi t test for separate gene hypothesis based on assumed genotypes of the mapping population parents.
Population Environment Assumed genotype†Red midrib Green midrib Expected ratio χ2p value for χ2
F1Summer greenhouse Rmrm x rmrm 87 91 1:1 0.06 0.8065
F1Summer fi eld Rmrm x rmrm 89 89 1:1 0.00 1.0000
F1Winter greenhouse Rmrm x rmrm 87 91 1:1 0.06 0.8065
F1Winter fi eld‡Rmrm x rmrm 81 77 1:1 0.06 0.8065
S1Summer greenhouse Rmrm ⊗111 30 3:1 0.77 0.38 02
S1Winter greenhouse Rmrm ⊗110 31 3:1 0.47 0.4930
†Assumed genotype based on the separate gene hypothesis of Corkill (1971).
‡During the winter fi eld phenotyping, 20 plants had died in all blocks in the fi eld, so the chi-square values were tested against a population of 158 instead of 178.
Table 3. Segregation of the white clover red fl eck (Rf; Fig. 1i) trait within the F1 mapping population (reciprocal pseudo-test-
cross between 05-O-34 × GA43) and the S1 confi rmation population (selfi ng of ornamental parent 05-O-34). Chi-squared (χ2)
goodness-of fi t test for separate gene hypothesis based on assumed genotypes of the mapping population parents.
Population Environment Assumed genotype†Red fl eck No red mark Expected ratio χ2p value for χ2
F1Summer greenhouse Rfrf x Rfrf 107 71 3:1 19.34 < .0001
F1Summer fi eld Rfrf x Rfrf 122 56 3:1 3.28 0.0701
F1Winter greenhouse Rfrf x Rfrf 123 55 3:1 2.69 0.1010
F1Winter fi eld‡Rfrf x Rfrf 113 45 3:1 0.68 0.4096
S1Summer greenhouse Rfrf ⊗113 28 3 :1 1.61 0 .2045
S1Winter greenhouse Rfrf ⊗116 25 3 :1 3. 43 0.06 40
†Assumed genotype based on separate gene hypothesis of Corkill (1971).
‡During the winter fi eld phenotyping, 20 plants had died in all blocks in the fi eld, so the chi-square values were tested against a population of 158 instead of 178.
Fig. 4. Linkage maps indicating the location of the loci conditioning the red midrib and red fl eck traits on linkage group (LG) B1 of the map
described by Zhang et al. (2007). Linkages are based on segregation in the mapping population (reciprocal pseudo-testcross between
05-O-34 × GA43) and confi rmation population (selfi ng of ornamental parent 05-O-34). Rm = Red midrib trait; Rf = Red fl eck trait. a)
Single parent linkage map of LG B1 showing the location of the red midrib locus based on segregation in the ornamental parent within the
mapping population. b) Consensus map of LG B1 showing the location of the red midrib locus and red fl eck locus based on segregation
in both the ornamental-type parent and agronomic-type parent within the mapping population. c) Confi rmation linkage map of LG B1
showing the location of the red midrib locus and red fl eck locus based on segregation within the confi rmation population.
1266 WWW.CROPS.ORG CROP SCIENCE, VOL. 50, JULY–AUGUST 2010
the dat a f rom ea ch co lle ct io n d ate an d l oc at ion we re map pe d
separately. While attempting to map the gene(s) controlling
multifoliolate expression, which is believed to be a recessive
trait, the data did not segregate with any markers from the
ornamental parent. In contrast, when the dominant trifolio-
late trait data were used to map the gene(s), they segregated
with markers inherited from the agronomic parent. The
gene(s) responsible for trifoliolate leaves (Fig. 5) is linked to
molecular markers segregating in the GA43 parent on what
corresponds to LG H1 of the map described by Zhang et
al. (2007). Although the gene(s) controlling the trifoliolate
trait always maps to the same LG, the data collected in the
winter and in the summer map to locations 30 cM apart
(Fig. 5). Within the winter phenotypes, the data collected
in both the greenhouse and the eld map to locations 19 cM
apart. Molecular marker RCS2681 maps to the same loca-
tion as the winter eld data. Simple sequence repeat marker
TRSSRA02C02 maps 2 cM above the winter greenhouse
data. Summer data from the greenhouse maps 28 cM away
from data collected in the eld.
Although molecular markers distributed throughout
all white clover LGs were evaluated, the density of mark-
ers in some LGs was too low and therefore insu cient
to detect marker-trait linkages for all of the traits segre-
gating in the mapping population. The red lea et (Vrl)
trait, which was only visible during the winter evalua-
tions, failed to segregate with any of the molecular mark-
ers used in this study. Likewise, the marginal mark (Vm)
and the intermediate white V mark (Vi), although visible
during all collection dates, did not segregate with any of
the molecular markers evaluated.
DISCUSSION
Before this study, only one white clover morphological
trait had been placed on the white clover linkage map,
namely the R locus on LG B1 (Barrett et al., 2004). The
results from this study made it possible to re ne the loca-
tion reported by Barrett et al. (2004), to map two addi-
tional traits, and to conclusively determine the nature of
the R locus. The use of molecular markers permitted the
development of both single-parent and consensus maps for
assignment of these trait-speci c genes and to overcome
the historic di culties associated with mapping traits in
heterozygous genotypes.
Mapping of separate genes for red eck (Rf) and red
midrib (Rm) in the same population made it possible to
clarify the confusion that has been associated with the
genetic control of these traits. Barrett et al. (2004) fol-
lowed the Carnahan et al. (1955) premise that red leaf
marks are conditioned by a single gene, named R, with
all of the di erent morphotypes due to di erent alleles of
that gene. The R locus mapped by Barrett et al. (2004)
is actually the red eck (Rf) trait, as determined by their
description of this phenotypic trait.
In this study, the two dominant alleles of the genes
that control red eck and red midrib expression, Rf and
Rm, are linked in coupling phase in the ornamental par-
ent, which explains the observation that all individuals
with red midrib leaf mark also possess the red eck leaf
mark. In the forage parent, the dominant red eck leaf
mark allele (Rf) and the recessive red midrib leaf mark
allele (rm) are also in coupling linkage, resulting in indi-
viduals with the red eck trait but without red midrib
leaf mark in the mapping population. The di erences in
linkage between the two traits between the two popula-
tions are therefore a result of the di erence in the paren-
tal genotypes used to develop the two populations. The
distinct mapping of both traits, in conjunction with the
inheritance data, con rms Corkill’s (1971) assertion that
the R locus actually comprises a series of linked genes.
With the mapping of the trifoliolate leaf trait, the
location of at least one gene responsible for lea et num-
ber in the species has been discovered and highlights the
complexity of this trait. The popular multifoliolate trait
in white clover is controlled by a recessive gene(s), and its
expression is also strongly in uenced by the environment.
As such, the data for the trifoliolate trait were mapped
separately. The trifoliolate trait segregated with molecular
Fig. 5. Single parent linkage map of linkage group (LG) H1 showing
the location of the trifoliolate leaf trait based on segregation in
the GA43 parental map from the mapping population (reciprocal
pseudo-testcross between 05-O-34 × GA43). Sum = summer;
Win = winter, GH = greenhouse; Fld = Field; 3Leaf = trifoliolate
leaf trait.
CROP SCIENCE, VOL. 50, JULY–AUGUST 2010 WWW.CROPS.ORG 1267
markers located on LG H1 of the map described by Zhang
et al. (2007). The di erence in mapping location of this
trait between summer and winter evaluation dates is likely
due to the environmental e ect on multifoliolate expres-
sion but may also be an artifact of the population size and
the low marker density or it might mean that di erent loci
control lea et number in the summer and winter.
Because the basal species of Tr ifolium often have pen-
tafoliolate leaves (Ellison et al., 2006; Zohary and Heller,
1984), it is believed that the genus Trifolium originated
from multifoliolate ancestors and that the number of leaf-
lets was reduced during evolutionary time (Eames, 1961;
Jaranowski and Broda, 1978; Zohary and Heller, 1984).
The presence of a dominant locus that inhibits the expres-
sion of multifoliolate leaves, leading to trifoliolate leaves
in white clover, supports the premise that lea et number
suppressors in the Fabaceae in general, and Trifolium in
particular, resulted in lower lea et number (Eames, 1961;
Zohary and Heller, 1984). There is another trait which
sometimes appears in white clover populations in which
the petiolule of the middle lea et is elongated. It was
noted in this study that whenever the multifoliolate trait
and the elongated petiolule were expressed together, the
resulting leaves were frequently pinnately compound (Fig.
2c) rather than palmately compound (Fig. 2b). These mul-
tifoliolate pinnate leaves bear an even greater resemblance
to the typical leaf morphology of legumes (Eames, 1961).
As more molecular markers become available in white
clover and the LGs become more saturated, it should be
possible to map the other morphological traits described
here and further clarify the inheritance of these traits.
During the development of their molecular map, Zhang
et al. (2007) showed that SSR markers from related
legume species could be successfully utilized as markers
in white clover. Accordingly, the complete sequencing of
the closely related (George et al., 2008) reference species
Medicago truncatula Gaertn. genome (www.medicago.org;
veri ed 24 Mar. 2010) will facilitate the development of
additional molecular marker resources and comparative
mapping e orts that may be useful to identify the gene(s)
responsible for the observed variation in many ornamental
and agronomic traits within white clover populations.
In summary, the advent of molecular marker-based
maps for white clover means that the tools are nally in
place to start addressing long-standing questions on clover
genetics and evolution. It was possible to identify the loca-
tion of two new morphological traits utilizing molecular
markers from previously published linkage maps in white
clover. The successful mapping of the red midrib trait in
a population that also contains the previously mapped red
eck trait resolves the con icting hypotheses of earlier
researchers studying R locus inheritance in white clover.
The successful mapping of at least one gene responsible
for trifoliolate lea et number in the species highlights its
complexity and brings white clover breeders and research-
ers one step closer to unlocking the genetic mechanisms
behind multiple lea et expression in white clover and x-
ing this trait for breeding purposes.
Acknowledgments
The authors are grateful to the following people at the Sam-
uel Roberts Noble Foundation: Dr. Yan Zhang for the initial
screening of SSR markers used in this study and advice on map-
ping; Ann Harris and Jarrod Steele for running the molecular
markers through the sequencer; and Dr. Brindha Narasimham-
oorthy for her advice on using the mapping software. From the
Univ. of Georgia, Kevin Michael Payne, Jr., Jace Morgan, and
Alexandra Kerr are gratefully appreciated for maintenance of
the mapping populations and their assistance in phenotyping
and eld planting. This work was partially funded by the Sam-
uel Roberts Noble Foundation and the Georgia Agricultural
Experiment Stations.
References
Atwood, S.S. 1942. Oppositional alleles causing cross-incompati-
bility in Trifolium repens. Genetics 27:333–338.
Baltensperger, D.D., D.S. Wo ord, and W.H. Anderson. 1991.
Registration of Fl-Ml white clover germplasm. Crop Sci.
31:853.
Barrett, B., A. Gri ths, M. Schreiber, N. Ellison, C. Mercer, J.
Bouton, B. Ong, J. Forster, T. Sawbridge, G. Spangenberg,
G. Bryan, and D. Wood eld. 2004. A microsatellite map of
white clover. Theor. Appl. Genet. 109:596–608.
Barrett, B., C. Mercer, and D. Wood eld. 2005a. Genetic map-
ping of a root-knot nematode resistance locus in Trifolium.
Euphytica 143:85–92.
Barrett, B.A., I.J. Baird, and D.R. Wood eld. 2005b. A QTL anal-
ysis of white clover seed production. Crop Sci. 45:1844–1850.
Bouton, J.H., D.R. Wood eld, J.R. Caradus, and D.T. Wood.
2005. Registration of ‘Durana’ white clover. Crop Sci. 45:797.
Brewbaker, J.L. 1955. V-leaf markings of white clover. J. Hered.
46:115–123.
Carnahan, H.L., H.D. Hill, A.A. Hanson, and K.G. Brown. 1955.
Inheritance and frequencies of leaf markings in white clover.
J. Hered. 46:109–114.
Cogan, N.O.I., M.T. Abberton, K.F. Smith, G. Kearney, A.H.
Marshall, A. Williams, T.P.T. Michaelson-Yeates, C. Bowen,
E.S. Jones, A.C. Vecchies, and J.W. Forster. 2006. Individual
and multi-environment combined analyses identify QTLs for
morphogenetic and reproductive development traits in white
clover (Trifolium repe ns L.). T he or. A ppl. Ge ne t. 112 :1401– 1415 .
Corkill, L. 1971. Leaf markings in white clover. J. Hered. 62:307–310.
Davies, W.E. 1963 Leaf markings in Trifolium repens. p. 94–98. In
C.D. Darlington and A.D. Bradshaw (ed.) Teaching genetics
in school and university. Philosophical Library, New York.
Eames, A.J. 1961 Morphology of the angiosperms. McGraw-Hill
Book Company, New York, NY.
Ellison, N.W., A. Liston, J.J. Steiner, W.M. Williams, and N.L.
Taylor. 2006. Molecular phylogenetics of the clover genus
(Trifo li um- Leguminosae). Mol. Phylogenet. Evol. 39:688–705.
Fehr, W.R. 1987 Genetic principles. p. 26–46. In W.R. Fehr (ed.)
Pr inc iples of cultivar development. McGraw-H ill, New York.
Ford, J.L., and R.B. Claydon. 1996. Inheritance of multifoliolate
1268 WWW.CROPS.ORG CROP SCIENCE, VOL. 50, JULY–AUGUST 2010
leaves in white clover. Agron. Soc. New Zealand Spec. Publ.
11:167–170.
George, J., T.I. Sawbridge, N.O.I. Cogan, A.R. Gendall, K.F.
Smith, G.C. Spangenberg, and J.W. Forster. 2008. Compari-
son of genome structure between white clover and Medicago
truncatula supports homoeologous group nomenclature based
on conserved synteny. Genome 51:905–911.
Grattapaglia, D., and R. Sedero . 1994. Genetic-linkage maps
of Eucalyptus-grandis and Eucalyptus-urophylla using a pseudo-
testcross- mapping strategy and RAPD markers. Genetics
137:1121–1137.
Hand, M.L., R.C. Ponting, M.C. Drayton, K.A. Lawless, N.O.I.
Cogan, E.C. Brummer, T.I. Sawbridge, G.C. Spangenberg,
K.F. Smith, and J.W. Forster. 2008. Identi cation of homolo-
gous, homoeologous and paralogous sequence variants in an
outbreeding allopolyploid species based on comparison with
progenitor taxa. Mol. Genet. Genomics 280:293–304.
Hovin, A.W., and P.B. Gibson. 1961. A red leaf marking in white
clover. J. Hered. 52:295–296.
Jaranowski, J.K., and Z. Broda. 1978. Leaf mutants in diploid red-
clover (Trifolium pratense L). Theor. Appl. Genet. 53:97–103.
Jones, E.S., L.J. Hughes, M.C. Drayton, M.T. Abberton, T.P.T.
Michaelson-Yeates, C. Bowen, and J.W. Forster. 2003. An
SSR and AFLP molecular marker-based genetic map of white
clover (Trifolium repens L.). Plant Sci. 165:531–539.
Knight, W.E. 1969. Inheritance of multifoliolate leaves glabrous
leaves and petiolulate lea et attachment in crimson clover
Trifolium incar natum L. Crop Sci. 9:232–235.
Lenoble, M., and J. Papineau. 1970. Note sur une nouvelle marque
foliaire chez Trifoliom repens. (In French.) Ann Amelior Des
Plant 20:485–487.
Quesenberr y, K.H., R.R. Smith, N.L . Taylor, D.D. Baltensperger,
and W.A. Parrott. 1991. Genetic nomenclature in clovers and
special purpose legumes.1. Red and white clover. Crop Sci.
31:861–867.
Simon, U. 1962. Inheritance of polyphylly in red clover (Trifolium
pratense, L.). Crop Sci. 2:258.
Tashiro, R.M., J.H. Bouton, and W.A. Parrott. 2009. ‘Frosty
Morning,’ ’Patchwork Quilt,’ ’Irish Mist’ and ‘Pistachio Ice
Cream’ ornamental white clover, Trifolium rep ens L. Hort-
Science 44:1779–1782.
Taylor, N.L. 1982. Registration of gene marker germplasm for
red-clover. Crop Sci. 22:1269.
Van Ooijen J.W., R.E. Voorrips. 2001 JoinMap 3.0, software for
the calculation of genetic linkage maps. Plant Research Int.,
Wageningen, the Netherlands.
Voorrips, R.E. 2002. MapChart: Software for the graphical pre-
sentation of linkage maps and QTLs. J. Hered. 93:77–78.
Zhang, Y., M.K. Sledge, and J.H. Bouton. 2007. Genome mapping
of white clover (Tr ifoliu m repens L.) and comparative analysis
within the Trifolieae using cross-species SSR markers. Theor.
Appl. Genet. 114:1367–1378.
Zohary, M., and D. Heller. 1984 The genus Tr ifolium. Israel Acad.
of Sci. and Humanities, Jerusalem, Israel.