|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
MOLECULAR CHARACTERIZATION OF PHOTOPERIODIC FLOWERING IN
STRAWBERRY (Fragaria SP.)
PHILIP JACOB STEWART
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA INT PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
02007 Philip J. Stewart
To my parents,
who always believed in me, and taught me to love knowledge and the natural world.
To the memory of my grandfather, Jackson Lee Stewart, Jr.,
who introduced me to agriculture, and who will forever be my image of the American farmer.
To my children, Zea and Adina,
who have been a constant inspiration and joy to me,
and reminders of the things in life that are truly important.
And, most of all, to my wife, Cynthia,
who has been unwavering in her faith in my abilities,
in her willingness to sacrifice for my academic and professional pursuits,
and, most importantly, in her love.
I would like to take this opportunity to thank at least a few of the many people who have
have assisted me along the path to this degree.
First, I would like to offer my love and gratitude to my parents, Lee and Faith Stewart,
and my in-laws, Barton and Teri Cowan, all of whom have provided continuous support, both
emotional and Einancial, throughout the course of my graduate studies. None of this would have
been possible without them.
Secondly, I would like to thank those that who have provided invaluable assistance in my
academic and professional development, in particular my graduate advisors, Kevin Folta and
Craig Chandler. Their help was critical to the work described here. I am also grateful for the
assistance and suggestions offered by the other members of my committee: Paul Lyrene, Natalia
Peres, and Jose Chaparro, all of whom helped to shape this study, and Daniel Sargent, of East
Malling Research, U.K., for the use of his diploid mapping population. Much thanks also go to
Dawn Bies and Maureen Clancy, for technical assistance, and also to my fellow students in the
Folta lab: Denise Tombolato, Stefanie Maruhnich, Thelma Madzima, and Wendy Gonzalez. I've
enjoyed working with all of you and wish you much success as you Einish your studies and move
into professional life.
And, Einally, the greatest debt of gratitude is owed my beautiful wife, Cynthia, who has
not only been tolerant but encouraging as I have dragged her from place to place across the
country in pursuit of this dream. Thank you so much--I hope it was worth it.
TABLE OF CONTENTS
ACKNOWLEDGMENT S ................ ................. 4...............
LI ST OF T ABLE S................................ 8
LI ST OF FIGURE S ................. ................. 9......... ...
AB STRAC T ................. ................. 1......... 1....
1 INTRODUCTION ................. ................. 12.............
2 LITERATURE REVIEW.............. ................. 14
The genus Fragaria .............. .............. 14..._ _. ....
Polyploidy in Strawberries .........._.... ............... 15...__ .. ....
Flowering Habits in Strawberry ........._..... ............... 16....._. ...
Short-Day Strawberries .........._.... ............... 16..._._. .....
Remontant Strawberries .........._.... .... .. ... ............... 18....
Factors Affecting Expression Of Photoperiodic Flowering ................ ................. ...._23
Temperature Effects On Flowering ................. ................. 23.............
Vernalization Requirements .........._.._ ......... ............... 24....
Juvenility and Plant Age Effects. .........._.._......... ...._ .... ....._._ .............2
Growth Pattern Differences Between Plants Of Differing Flowering Habits .........._.._..........26
Inheritance Of Flowering Habit In Strawberry .........._.._. .....__ ........__..........2
Molecular Markers For Flowering Habit ......_.._._ .... ..._.. .....___ ..........3
Molecular Control Of Flowering In Arabidopsis ................ ..............33. .............
Components Upstream of CO. ................ ................. 34.............
Components Downstream of CO ................ ..............35. ..............
The CONSTANS-LIKE Gene Family ..........._. .. .. ....... .. ...... ...........3
Components of the Flowering Pathways are Conserved Among Species .........._.........38
Using Model Systems To Understand Development In Rosaceae .........._... .............40
3 CHARACTERIZATION OF CONSTANS-LIKE GENES IN Fragaria .............. .............44
Introduction ............_... ............... 44..._._. ....
R results ........._... ... ......... ._... .... ... ._ _... .................. .. .......4
Cloning And Identification Of Four CONSTANS-LIKE Genes Of Fragaria.............__46
FrCO/FrCOL Gene Structure ........._._._....... .... ............... 48...
Southern blot analysis of FrCO Copy Number ..................... .._.. ..... ......... .... 50
Mapping Of Fragaria CO And COL Genes In A Diploid Mapping Population ........... 50
Comparison Of Strawberry Genotypes Under Varying Photoperiods ................... ........ 51
Allele-Specific Expression Patterns Differ Betwen Short-Day And Day-Neutral
Genotypes ................ ................. 52..............
Expression Of Other Genes Under Different Photoperiods ................. ............... .... 53
Some FraCOL2 Transcripts Contain An Unspliced Intron.............. ................. 54
Discussion .............. ... ....... _. .... .. .............. 54
The Co Gene Family In Fragaria (Rosaceae) ......._.__ ..... ...._ .. ....._.._.......5
Evolutionary Relationships Among FrCO Sequences.............. ............... 58
FrCO Expression ........._ ....... .... ..............61....
Expression of Other COL Genes ........._.. ....___......... ...........6
Materials and Methods ........._ ........_. ..............68....
Plant M material ............_._ ..............68._.._......
Plant Growth Conditions ........._ ........_. ..............69....
Cloning of FraCO .........._... ... ..... .. ..............70...
Identification of COL Genes in Genebank.............. ................70
Gene Structure Characterization.............. ............ 70
Southern Blot Analysis ........._.._.._ ......_. ..............71.....
Genetic Linkage Mapping COL genes in Fragaria ......____ .......___ .............71
Extraction of Nucleic Acids ........._.__........_. ..............72....
Reverse Transcription and RT- PCR ............_. .....___ .....__ ...........7
4 IDENTIFICATION AND CHARAC TERIZATION OF FLOWERING-RELATED
GENES IN Fragaria AND OTHER ROSACEAE ................. ................. ......... ..93
Introduction ................. ................. 93.............
Results ................. ...... ...... .... ......... .. ... ..........9
Identification of Rosaceae Flowering Gene Homologs ................. ........... .......... 94
FrSOC1 Expression ................. ................. 97.............
FrTFL1 and FrLFY Expression ........._.. ....___........_. ...........9
Discussion ........._...... .... .... .............. 98....
Rosaceae MADS-Box Genes ............_._ ..............98..._. ....
The FT/TFL1 Gene Family in Rosaceae.............. ............... 101
Other Flowering Genes ........._ ........_. ............._ 105..
FrSOC1, FrLFY, and FrTFL1 Expression ..............._ ................ 105......._...
Methods and Materials................ ...... ......... 109
Identification of Rosaceae Flowering Genes ..............._ .............. ........_.... 109
Phylogenetic Tree Construction ................. ......... ......... .. ..... ....10
RNA Extraction and RT-PCR Experiments ................ .............................. 109
5 Fragaria ALLELES OF POLYGALACTURONASE INHIBITOR PROTEIN GENES... 118
Introduction ............. ...... __ ............._ 118...
Results .............. ........__ ........ ........ ..........12
Identification of Fragaria PGIP Alleles .................._ ....__ ..........12
Amino Acid Similarities Among Alleles ........._ ....... __ ....._ .......... 2
Discussion ............. ..... .. ............. 123...
Methods and Materials.............. ............... 126
Plant and Genetic Material ............. ............._ ....._ .......... 12
Cloning and Sequencing of Fragaria PGIP Alleles ............_. ....... ............ 127
Sequence Analysis ........._ ........_. ............._ 127...
LIST OF REFERENCES .........._.._ ......... ............._. 132...
B IO GRAPHIC AL SKET CH ........._.._.. ...._... ............... 151
LIST OF TABLES
3-1 Comparison of amino acid identity (%) of the predicted protein encoded by FraCO to
those of Group la CONSTANS-LIKE genes in various species. .........._. .........._.....75
3-2 RT-PCR primers for strawberry CONSTANS-like genes and controls, sequence
source, Thl, and approximate size in cDNA (as calculated from sequence). ................... .76
3-3. GenBank accession numbers for Fragaria COL genes and other Rosaceae orthologs....77
3-4 Mean inflorescence and runner production of. xanana~ssa and E. vesca genotypes
under LD (16 h) and SD (8 h) photoperiods, at 18/160C day/night temperature. ............ 86
3-5. Sources of plants used in this chapter. ................ .......... .. .............. 90.
4-1 Identified Fragaria genes related to flowering and photoperiodic response, with the
closest match at the amino acid level among Arabidopsis and all annotated
transcripts. ................ ................ 112........ .....
4-2 RT-PCR primers for strawberry flowering genes and controls, sequence source, Thl,
and approximate size in cDNA (as calculated from sequence). ................ ............... 114
5-1 FaPGIP alleles identified in various octoploid strawberry cultivars. ............. ............ 129
5-2 Percentage of identity between PGIP isoforms at the mRNA and amino acid levels..... 130
LIST OF FIGURES
2-1 Simplified diagram showing interactions of genes and environmental factors
governing flowering in Arabidopsis thalianaiii~~~~~iiiii~~~~ ....._.. ............_. ..........__....... 4
3-1 Cladogram of CO-like genes, including all Arabidopsis and full-length Rosaceae
COL genes as well as all those from other species Group la genes demonstrated to
functionally complement Arabidopsis co mutants. ................ ................ ......... 78
3-2 Alignment of the predicted amino acid sequence for FraCO with those of AtCO and
functionally confirmed dicot homologs, with maj or conserved domains noted. ........._.....79
3-3 Alignment of the predicted amino acid sequence for FrvCOL1 with possible apple
and Arabidopsis homologs, with maj or conserved domains noted. ................ .............. 80
3-4 Structures of COL alleles from Fragaria species: xanana~ssa Strawberry Festival'
(Fra), E. vesca FDP815 (Fry), E. nubicola FDP 601 (Frn), and E. iinumae FRA377
(Fri) ................. ................. 8......... 1....
3-5 NJ phylogram tree of intron sequence divergence among FrCO alleles..............._.._. ..... 82
3-6 Comparisons of hypervariable section of coding region in FrCO. ........._.._.. ........._.... 83
3-7 Southern blot of genomic DNA from E. xanana~ssa 'Strawberry Festival' and F.
vesca Hawaii 4, cut with six different enzymes and probed with a portion ofFrvCO ..... 84
3-8 Mapping of FrCO in a reference diploid population. ........._._......___ ............. 85
3-9 Northern blot of RNA from Strawberry Festival', a SD genotype, and 'Diamante', a
DN genotype, collected every 4 h, showing expression of FraCO under short and
long day conditions. ........... _.. .............. 87..._ _. ....
3-10 RT-PCR assay of expression ofFrCO transcript under short (8 h) and long (16 h)
days in three genotypes: E. vesca Hawaii-4 (SD, diploid), xanana~ssa 'Camarosa'
(SD, octoploid), and xanana~ssa 'Diamante' (DN, octoploid). ............. .................. 88
3-11 Expression of 35 and C9 alleles ofFraCO at dawn under SD, using RT-PCR at 33
and 40 cycles, in 'Strawberry Festival' (F) and 'Diamante' (D). ................ ................ 89
3-12 RT-PCR assay of FrCOL2 expression in E. vesca Hawaii-4, showing variation in
splicing efficiency at several points under short days. ......... ................ ..............91
3-13 Ubiquitin controls after equalization of RT-PCR samples using RN-Ubiql primer set
to standardize template amounts. ........._.. ...._._ ..............92...
4-1 NJ tree showing similarity between full-length amino acid sequences of Rosaceae
and Arabidopsis MADS-Box genes ................ ................ 111........ ....
4-2 NJ phylogeny tree showing relationships between Arabidopsis and Rosaceae
members of the FT/TFL1 gene family, based on full predicted protein sequence. ........ 1 13
4-3 RT-PCR assay of expression of FrSOC1 under short (8 h) and long (16 h)
photoperiods in 'Camarosa' (SD, octoploid) and 'Diamante' (DN, octoploid). ........... 115
4-4 RT-PCR assay of expression of FrLFY under short (8 h) and long (16 h)
photoperiods in Hawaii-4 (SD, diploid) and 'Camarosa' (SD, octoploid). ................... 116
4-5 RT-PCR assay of expression ofFrTFL~ under short (8 h) and long (16 h)
photperiods in Hawaii-4 (SD, diploid). ................ ......... ......... ..........17
5-1 Amino acid alignment of the polymorphic portion of a number of Fragaria PGIP
genes. Shaded areas indicate the xxLxLxx beta-sheet regions. ........._.__... ......_...... 131
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
MOLECULAR CHARACTERIZATION OF PHOTOPERIOD FLOWERING IN
Philip J. Stewart
Chair: Craig Chandler
Cochair: Kevin Folta
Major: Horticultural Science
The initiation of flowering is a critical developmental transition for most plant species,
affecting reproductive potential and evolutionary success. The timing of this event in strawberry
(Fragaria sp.) is conditioned by a number of factors, including photoperiod. Sensitivity to
photoperiod varies among strawberry genotypes, which are generally divided into three
categories: short-day, everbearing, and day-neutral. The genetic basis for these variations is not
known, but existing model plant systems such Arabidopsis thaliana and rice reveal a well-
conserved network of genes and proteins governing the perception of photoperiod and the
regulation of floral initiation. A number of these genes were identified and characterized in
strawberry and other relatives in the family Rosaceae, and their expression in strawberry under
long and short photoperiods assayed. Expression profiles suggest that while many of the critical
genes governing photoperiodic flowering are conserved between strawberry and model species,
their expression and relationships to one another are unlike those of any previously characterized
The initiation of flowering is among the most important developmental transitions a plant
makes, second perhaps only to germination. Mistimed flowering endangers the reproductive
potential of plant in a number of ways. A plant that flowers too early or too late may find the
seasonal conditions not conducive to seed development, or yield seeds that germinate at points in
the season that make seedling survival difficult or impossible. In cross-pollinated species even a
difference of a few days may be adequate to render reproduction impossible, putting a plant out
of syncronization with others of its species. Any of these difficulties has the potential to render
an individual an evolutionary dead end.
Because of this necessity, plants have developed an extraordinary regulatory network
allowing the Eine tuning of the transition to flowering by a range of environmental and internal
factors. In the model eudicot Arabidopsis, these factors have been shown to include the
environmental inputs of photoperiod, light quality, and temperature, and internal pathways
controlled by the hormone gibberellic acid and autonomous factors such as developmental age.
At the molecular level, each of these inputs represents separate elements of genetic
pathways that converge at a few floral integrators to regulate flowering through a common set of
meristem identity genes. These integrators convert the complex and heterogeneous inputs from
the various environmental and internal cues into a decision to flower, influencing a choice made
in a small cluster of meristmatic cells, and driving development from a vegetative to
reproductive state through the activation of genes that remodel cell fate. Key among these floral
integrators is CONSTANS (CO), which serves to tie the perception of photoperiod with that of
the spectral quality of light received.
Flowering in cultivated strawberry (Fragaria xanana~ssa Duch.) has historically received
a great deal of attention from researchers, for two primary reasons. First, strawberry is an
attractive platform for flowering research because despite a narrow genetic base (Sjulin & Dale,
1987) it exhibits at least three distinct and ranging flowering habits: junebearing (short day), day-
neutral (photoperiod insensitive), and everbearing (often referred to as "long day" but more
accurately another distinct form of photoperiod insensitivity). These clear phenotypic
delineations in such a genetically homogeneous background suggests fairly simple genetic
difference between markedly contrasting flowering habits.
Secondly, in an increasingly global market where fruit is readily transported not only
across countries but around the world, farmers depend on a precise understanding of flowering
behavior to time harvest to coincide with the best market windows in order to obtain the best
price for their fruit. In an era of narrowing profit margins, this timing may be the difference
between a profitable crop and a loss.
The genus Fragania
All strawberry species belong to the genus Fragaria, and are members of Rosaceae, a
family that contains a large number of economically significant crops, primarily fruits such as
apple (Malus domestic Borkh.), pear (Pyrus sp.), stone fruits (Prunus sp.), and brambles (Rubus
sp.), but also ornamentals such as rose (Rosa sp.). Strawberries belong to the Rosoideae
subfamily, which includes the genera Rubus, Rosa, Potentilla, and Duchesnea. Authorities vary
somewhat on the exact number of strawberry species, but most recent authors list about twenty
species (Hancock and Luby, 1993; Folta and Davis, 2006). Native strawberries occur in much of
North America, Europe, and Asia, as well as parts of South America, and Hawaii (Darrow, 1966;
Hancock and Luby, 1993).
Cultivated strawberry, xanana~ssa Duch., is a species of relatively recent hybrid origin,
being derived from the accidental hybridization of two New World octoploid species, F.
virginiana Duch. and E. chiloensis (L.) P. Mill., beginning in France some time in the late 1700s
(Darrow, 1966). Although the polyploid nature of the crop and its hybrid origin have assured that
the species contains some genetic diversity, the actual germplasm base used is relatively small,
with only 53 founding clones (Sjulin and Dale, 1987) and 17 initial cytoplasm donors (Dale and
Sjulin, 1990) contributing to the cultivated varieties. Some recent introgressions from the wild
have been made, most notably from a selection of E virginiana ssp. Duch. var. glauca (S. Wats)
Staudt, but in general very little new genetic material has been added since the early years of
strawberry breeding (Darrow, 1966).
Polyploidy in Strawberries
As has been previously mentioned, the cultivated strawberry is an octoploid (2n=8x=56).
The maj ority of strawberry species, however, are diploids, and there exist tetraploid and
hexaploid taxa as well (Folta and Davis, 2006). Considerable evidence suggests that the
cultivated strawberry is an alloploid, with either three or four distinct subgenome types. The
earliest model, that of Federova (1946), proposed a composition of AABBBBCC, but this was
replaced with AAA'A'BBBB when Senanayake and Bringhurst (1967) found evidence of partial
homology between the A and C genomes. Mounting evidence of diploidization (Bymne and
Jelenkovic, 1976; Arulsekar, et al. 1981) prompted Bringhurst (1990) to propose a fully
diploidized model: AAA'A'BBB'B'.
The origins of the component genomes have long been a subj ect of debate. As early as
1926, Ichijima suggested that one of these might be derived from E. vesca, based on cytogenetic
observations ofE. virginiana (8x) x E. vesca L. (2x) hybrids, in which the formation of seven
bivalents was noted. Further data, including cytology (Senanayake and Bringhurst, 1967) and
sequence analysis of the chloroplast trnL-tmnF region (Potter et al., 2000), the nuclear sequences
ITS (Potter et al., 2000) and polygalacturonase inhibitor protein (PGIP) (Chapter 5 of this work),
have strengthened this view, and the A genome is now widely believed to be derived from F.
vesca (Folta and Davis, 2006).
The origins of the other genomes have been less clear. The study of ITS and tmnL-trnF
sequences implicated E. nubicola as a possible A-type genome donor, in addition to E. vesca
(Potter et al., 2000). Recent work by DiMeglio and Davis (unpublished but cited in Folta and
Davis, 2006) examining the alcohol dehydrogenase locus implicates two other species, F.
mandshurica Staudt and E. iinumae Makino. Potter et al. (2000) and Harrison et al. (1997) found
E. iinumae to be the most distantly related of all other Fragaria species examined, and as such
may represent the ancestor of one or both of B and B' genomes. Unlike most diploid species,
which possess at least a degree of compatibility with other diploid Fragaria (Dowrick and
Williams 1959), E. iinumae is not compatible with E. vesca (Bors, 2000; Sargent et al., 2004b),
and displays a karyotype distinct from the other diploid species (Iwatsubo and Naruhashi, 1989,
1991), which may account for the failure of B genome chromosomes to pair with those of the A
genomes. E. mandshurica, on the other hand, is a newly described diploid species (Staudt, 2003)
not included in previous studies. It is however thought to be related to the tetraploid species F.
orientalis Losinsk., to which it bears a strong resemblance. Potter et al. (2000) had placed F.
orientalis in the A clade, and thus E. mandshurica may represent A' genome in the octoploid
Flowering Habits in Strawberry
Although many authors (Fuller, 1897; Fletcher, 1917) had suggested the presence of
environmental influences on flowering in strawberry, Sudds (1928) was perhaps the first to note
the specific effects of photoperiod and temperature on flower bud induction in strawberry. Soon
after, Darrow and Waldo (1929, 1930, 1933) conducted an exhaustive series of experiments,
classifying over a hundred octoploid strawberry genotypes as flowering under either short or
long days, finding the groups to generally coincide with what were already termed "June
bearers" and "ever bearers".
It is believed that the natural state of most strawberry species is as a short day (SD) plant
(Darrow, 1966), though there is considerable variation in sensitivity. It appears that the earliest
cultivated octoploids all flowered under short days (Darrow, 1966), although the limits of the
required photoperiod vary considerably between genotypes. In general, most SD genotypes form
flower buds under photoperiods of less than 14 h, which was first suggested by Darrow (1936).
Three maj or factors seem to dictate response to photoperiod: genotype, temperature, and chilling.
Genotype effects on photoperiodic flowering in SD strawberries
Darrow concluded that photoperiods of 9.5 to 13.5 h induced the greatest number of
flowers in SD cultivars, with an optimum around 12 h (1936). While still generally true, further
studies and perhaps continued progress in breeding have widened this range somewhat. On one
end of the range, some authors (Izhar, 1997; Faedi et al., 2002) have delineated a class of
strawberries they refer to as "infra short-day." These plants flower under longer day lengths
(typically 13.5-14 h photoperiod) and with little chilling requirement, but seem to represent
merely an extension of the existing spectrum of sensitivities already known in SD plants, rather
than a distinct flowering habit. In other genotypes, photoperiods of considerably less than 12 h
are required for optimal flowering. In a study by Heide (1977) 'Abundance' demonstrated peak
flowering under a 10 h photoperiod, the shortest tested in the study. A number of studies have
shown robust flowering under 8 h photoperiods in some cultivars (Sonsteby, 1997; Leshem and
Koller, 1964; Barger et al. 1997; Moore and Hough, 1962). Many early researchers compared
short and long day lengths in terms of natural light versus natural light plus an artificial extension
of day length, making it difficult to assign specific photoperiods.
It is interesting to note that while SD photoperiods are needed for optimal floral
induction, they may actually delay floral development. Several studies of SD cultivars have
shown that continuing SD photoperiods may actually slow the development of previously
initiated buds compared to longer photoperiods (Moore and Hough, 1962; Durner and Poling,
1987). Consequently, Salisbury and Ross (1992) consider strawberries to be SD for purposes of
flower induction, but LD plants for flower development. A similar relationship for temperature
has also been noted (see below).
The first maj or exception to this flowering pattern noted was the alpine strawberry, F.
vesca var. semperflorens Duch. Found at several places in the European Alps, these represent a
mutation of the common diploid European wood strawberry, and were among the earliest types
to be cultivated. Though first mentioned as early as 1553, they gained prominence in the mid
1700's when they were introduced into France and England, possibly from Turin, Italy
(Duchesne, 1766). In 1811, a runnerless form was obtained by Labaute (Darrow, 1966). Now
only occasionally cultivated, these varieties are of only minor commercial importance (Darrow,
1966), although in the early nineteenth century a white-fruited variety was extensively cultivated
in Quebec (Fletcher, 1917). Fairly little work has been done to characterize the floral physiology
of these varieties, and the fact that many workers have studied unnamed genotypes that are either
seed-propagated or even collected directly from the wild (Chabot, 1978) makes it difficult to
consistently track genotype-specific differences. Sironval and El Tannir-Lomba (1960) found
that their selection ofF. vesca var. semperflorens required days longer than 12 h to form flower
buds, and that SD treatments inhibited flowering in plants that had previously reached the
flowering stage under LD conditions, though this appears to be the exception, not the rule
(Darrow, 1966). As in other strawberries, temperature has a significant effect, with moderately
low temperatures favoring both reproductive development and biomass production (Chabot,
Relatively quickly after the introduction of octoploid cultivars, however, so-called
everbearingg" (EB) strawberries were identified (Darrow, 1966). Such genotypes initiate
flowering most heavily under the long days of summer, and often on unrooted or newly rooted
runners, resulting in what is in most locations primarily a fall harvest, rather than an early
summer one, or in some cases two distinct crops.
These genotypes appear to trace to a few distinct sources within cultivated types.
suggesting spontaneous mutation of a single gene. Development of such types took place
independently in both North American and Europe, but very little crossover has ever occurred
between these two sources of the trait. The first North American EB cultivar described was
possibly 'Oregon Everbearing,' obtained in 1882, although Fletcher (1917) describes an
everbearing selection of what was likely E. virginiana, found in Ohio in 1852. The first
successful everbearer, however, was likely 'Pan-American,' an apparent sport of the June bearer
'Bismarck' found in 1898 (Fletcher, 1917). 'Pan-American,' in turn, became the source of the
trait for a number of other successful everbearing cultivars, including 'Progressive' and
'Rockhill' (Darrow, 1966). This last selection is the source of the trait in the proprietary
everbearing cultivars developed by Driscoll Strawberry Associates (Darrow, 1966). Other
sources of the trait in North America include 'Gem,' a sport of 'Champion' introduced in 1933
that proved quite popular for several decades. That the trait is the result of instability at a single
locus is supported by documented reversion to short day flowering in the case of 'June Rockhill',
a short day mutant of the everbearer 'Rockhill' (Darrow, 1966).
European octoploid everbearers, called "remontant" by the French, may have had their
origin somewhat earlier. Unlike the American sources of the trait, which can largely be traced to
mutations identified in cultivated plants, the source of the European everbearers is unclear.
Although claimed at the time to be hybrids of large-fruited octoploids with the everbearing
Alpine diploids (Fletcher, 1917), it seems doubtful that this was truly the case. First, such
hybrids of E vesca with the octoploids are difficult to make (Yarnell, 1 93 1; Mangelsdorf and
East, 1927) and even more difficult to restore to adequate fertility, and secondly the inheritance
of the trait in the octoploids seems to be different than in Alpine strawberries (Brown and
Wareing, 1965; Clark, 1937). More likely, accidental cross-pollination resulted in seedlings of
pure F. xanana~ssa pedigree, among which was selected the original everbearer. Some, such as
'Louis Gauthier,' a double-cropping French cultivar, may have been simply SD genotypes with
unusually permissive photoperiod or chilling requirements (Fletcher, 1917). Richardson (1914)
suggests that the first true European octoploid everbearer was 'Gloede's Seedling,' introduced in
1866, and this may be the source for most or all European everbearers that followed (Ahmadi et
al., 1990). 'Mabille' and 'l'Inepuisable,' considered by some (Fletcher, 1917) to be pure Alpine
strawberry, followed this, but none were commercially successful. A French priest, Abbe
Thivolet, was among the first to breed for this trait, and after more than a decade of breeding
both octoploids and diploids (Fletcher, 1917) introduced a series of everbearers, culminating in
' St. Joseph', which Darrow (1966) refers to as the "first true large-fruited everbearer." Although
a considerable improvement over the preceding everbearers, 'St. Joseph' and Thivolet' s later 'St.
Antoine de Padoue' were not especially productive, but spurred the development of later
improved types. While EB plants derived from this source remain in production today, this
source has contributed very little to American breeding.
Although they have been called "long day" cultivars, these varieties lack the characteristic
behaviors associated with true long day plants, in the sense that they do not appear to be
regulated by the length of the dark period, and do not react identically to a short day with a brief
night break of light as they do to long photoperiods (Dennis et al., 1970; Durner et al. 1984), nor
is flowering inhibited by SD (Durner et al., 1984). Rather, these genotypes seem more dependent
on the total amount of light received. Dennis et al. (1970) found no difference in flower
formation in the everbearing cultivar Geneva grown under 12 h photoperiods versus those grown
under a 10 h photoperiod with a 2 h night break, a treatment which often inhibits flowering in SD
plants (Downs and Piringer, 1955; Piringer and Scott, 1964). Marked increases, however, were
seen with 18 and 24 h photoperiods. Additionally, they saw a significant effect of light intensity
when increased from 1200 to 2400 f-c on the 18 and 24 h photoperiod treatments (increases were
also seen on the 12 h and 10 + 2 h treatments, but these were not statistically significant). This
study and the later study by Durner et al. (1984) both noted similar levels of flowering for both
12 h photoperiods and 10 + 2 h night break treatments, suggesting that amount of light, not
duration of either darkness or light, is the critical factor.
The third distinct flowering habit is day-neutral (DN). Observers of wild American
octoploids had noted that many plants of E virginiana subsp. glauca (then called E. ovalis)
flowered in the summer and fall when most wild and commercial short day cultivars had stopped
(Darrow, 1966). Researchers with the USDA evaluated a large number of E. virginiana subsp.
glauca selections in the 1930s and 1940s, and a many were included in breeding efforts at the
time. A number of everbearing varieties resulted from these efforts, including two, 'Arapahoe'
and 'Ogallala,' (Hildreth and Powers, 1941) that are still in existence today. Whether the trait
expressed in these cultivars was the older EB trait or the DN trait derived from glauca is
unknown, as sources of both are present in the pedigrees, the strong tendency of most remontant
seedlings of 'Arapahoe' and 'Ogallala' to fruit in their first year is distinct from those other EB
parents (Ourecky and Slate, 1967), and more typical of DN genotypes (Ahmadi et al., 1991).
Regardless, the introduction of the DN trait that has had the largest impact came in the 1970's,
when the University of California breeding program, under Royce Bringhurst and Victor Voth,
utilized a single selection ofE. virginiana subsp. glauca from the Wasatch Mountains in Utah to
introduce this habit into commercial strawberries (Bringhurst and Voth, 1984). Strawberries
derived from this source constituted a maj ority of the cultivars planted in California in 1999
These were distinct from the everbearing habit in that the plants were truly insensitive to
the length of the photoperiod received, and fruit at nearly the same rate over a broad range of
photoperiods (Durner et al., 1984; Ahmadi et al., 1991). Day-neutral cultivars offered improved
heat tolerance and longer harvest seasons compared with the earlier everbearers, but runner
poorly and are difficult to propagate (Durner et al., 1984). Although day-neutrals appear to differ
from older everbearing types, many researchers have failed to distinguish between them, and in
some cases use the terms interchangeably, resulting in confusion. For the purposes of this work,
"day-neutral" (DN) will refer to the trait derived from E. virginiana ssp. glauca, while
everbearingg" will refer to that derived from 'Pan-American', 'Gloede's Seedling,' or other older
sources, and the generic "remontant" will be used to refer to cases where both are implied or the
author has not been clear in distinguishing.
Other flowering habits
Arguably a fourth flowering habit, poorly characterized at this time, is the
amphiphotoperiodic behavior exemplified by the E. chiloensis selection CHI24-1 (Yanagi et al.
2006). This genotype behaves as a standard SD plant under most conditions, with long
photoperiods inhibiting flowering, but initiates flowers under continuous lighting as well.
Although many will cease flowering altogether (Thompson and Guttridge, 1960), a few
cultivated SD strawberry cultivars may fall into this category, as both Sparkle' (Collins and
Barker, 1964) and 'Sweet Charlie' (unpublished observations) have been observed to flower
under such conditions as well, but thus far little research has been done on this phenomenon.
There has been much interest in recent years in finding novel sources of remontancy, and
workers have identified a number of wild accessions that seem to display varying degrees of
photoperiod-insensitivity or continual flowering, mostly among F. virginiana (Sakin et al., 1997;
Hancock et al., 2001, 2002; Serge and Hancock, 2005a, b). These new sources have not yet been
adequately characterized to determine whether they fall neatly into the DN or EB categories, or
whether they represent truly novel mechanisms. Inheritance studies by Serge and Hancock
(2005a) suggest that there may be multiple genetic mechanisms at work.
Factors Affecting Expression Of Photoperiodic Flowering
Temperature Effects On Flowering
A relationship between temperature and flower initiation was established by early workers,
and was highlighted in the large studies of flowering by Darrow and Waldo (1929, 1930, 1933).
Research has also shown that lower temperatures may permit flowering under longer daylengths
even in normally SD cultivars (Darrow and Waldo, 1934; Darrow, 1936). Hartmann (1947b)
concluded that this was the reason that cultivars grown in the coastal district of California near
Watsonville, where the mean temperature is 170C, were able to flower and fruit through the long
days of summer, while the same cultivars planted inland near Sacramento, where the mean
temperature was 230C, behaved as strict short day plants and produced only a spring crop. In
greenhouse studies he found that plants of four cultivars important to the region all flowered
under long days at 170C, but none flowered at 230C. Three of the four same genotypes flowered
freely under both temperatures under short day conditions, while a fourth, 'Fairfax,' failed to
flower at all under the higher temperature, even under 10 h days. Even everbearing cultivars
under long days are inhibited by high temperatures, as are day-neutral types, though in general
the critical temperature is higher for DN cultivars (Durner et al., 1984; Manakasem and
Goodwin, 2001). Heide (1977) found two of four Scandinavian SD cultivars still flowered under
photoperiods of 14, 16, and even 24 h when grown at 120C or 180C, though flowering was
substantially decreased under the longer photoperiods. The effects were highly dependent on
genotype, however, particularly in the middle of the temperature range, a consistent finding of
such studies. In some cases cultivars that are considered to be short day in some locations may
behave as remontant at other locations 'Climax', a British selection, behaved like a SD plant
when grown in the United States, but as a two-crop variety in England where temperatures are
cooler and summer days are longer (Downs & Piringer, 1955).
Much like SD photoperiods, which are necessary to develop flower buds, but inhibit the
development of flower trusses, low temperatures allow for flower induction under long
photoperiods, but may not be optimal for fruit production, because lower temperatures slow
development of the flower trusses (Darrow, 1966). Le Miere et al., (1996) found optimum
temperatures for truss development to be about 190C in 'Elsanta'. Temperatures that fluctuate
diurnally seem to be more effective in promoting flower development than continuous
temperatures, even at the same average temperature (Hartmann, 1947a). Such thermoperiodic
rhythms have been shown to have similar effects in other species (Seneca, 1974; Alvarenga and
Valio, 1989) and may allow for more robust entrainment of the circadian clock (Salome and
Temperature also plays a second role in the expression of flowering, by conditioning the
plant for flowering before the season begins. A certain amount of chilling, generally defined as
time between 00C and 70C, may be required to break bud dormancy and proceed with the normal
cycle of development, though the extent of chilling required is highly cultivar-dependent
(Piringer and Scott, 1964; Voth and Bringhurst, 1970; Durner and Poling, 1986; Darnell and
Hancock, 1996). Increased chilling has been shown to correlate with increases in leaf area and
number, petiole length, and runner initiation (Guttridge 1969; Bringhurst et al., 1960; Piringer
and Scott, 1964; Braun and Kender, 1985; Lieten, 1997). Chilling has been shown to promote
both vegetative and reproductive development (Darnell and Hancock, 1996). Durner and Poling
(1987) found chilling to enhance vegetative growth and reduce flower induction while promoting
floral differentiation. Non-chilled plants produce fruit of smaller size (Harmann and Poling,
1997) and lower quality (Bringhurst et al., 1960) than do the identical genotypes with adequate
Because of the related decrease in flower induction, however, chilling beyond the required
amount is of no advantage and can even markedly reduce the number of flowers. Piringer and
Scott (1964), for example, saw substantial decreases in flower number on 'Marshall' with
Juvenility and Plant Age Effects
Juvenility may also play a role in the expression of photoperiod-sensitivity In general,
young plants devote more resources to vegetative than reproductive growth, and may not flower
at all during this portion of their life cycle. This allows a plant to attain sufficient size to support
fruit and seed development before flowering (Thomas and Vince-Prue, 1984) Ourecky and Slate
(1967), working with the EB trait, noted that when populations were retained for a year and
scored for a second season, more plants were determined to be everbearing (a finding supported
by unpublished worked cited in Scott & Lawrence (1975)), casting a degree of doubt on earlier
work that utilized only first year data. However, Ahmadi et al. (1990) found that DN progeny
generally flowered within four months of germination. While not an issue for plants in
cultivation, this effect may have serious impacts on the ability of breeders to make selections
based on flowering habit. Ito and Saito (1962) found that older plants responded more robustly to
both photoperiod and temperature. This may have been in part a function of plant size, although
Hartmann (1947a) found that a single intact leaf was enough to perceive inductive photoperiods.
Growth Pattern Differences Between Plants Of Differing Flowering Habits
In addition to the differences in flower production, strawberries of differing flowering
habits also differ in terms of plant architecture and growth patterns. The production of runners
varies considerably among SD, DN, and EB genotypes. Runner initiation is affected by
photoperiod and temperature as well. In SD cultivars, there appears to be a balance between
runners and inflorescences ranging from almost no runners and many flowers under optimal SD
conditions, through an intermediate area where flowers and runners coexist, and then solely
runnering under the longest days. SD plants will runner in response to short days with a night
break, and they tend to shift from reproductive to vegetative growth under high temperatures
(Piringer and Scott, 1964; Durner et al., 1984). Interestingly, DN cultivars, while flowering in a
photoperiod insensitive manner, seem to retain a sensitivity to photoperiod in respect to
runnering, with the number of runners increasing under long days, night break, or high
temperature conditions. This is in contrast to EB types, which do not show photoperiodic or
temperature changes in runner initiation (Durner et al., 1984). In many, though not all cases
(Durner et al., 1984) EB genotypes runner very little, which may be a barrier to efficient
Nicoll and Galletta (1987) also noted differences in plant architecture between DN, EB,
and SD plants. Under long days, the main axis of SD and weakly DN plants remains vegetative
and most axillary buds develop as runners, while under short days, the main axis terminates in an
inflorescence and an upper axillary bud develops as a branch crown. In EB plants, under all
photoperiods, most or all axillary buds that develop do so as branch crowns, rather than runners,
and it is the termination of these crowns with inflorescences that results in most flowering,
though the main axis will occasionally terminate in a flower and continue growth from a side
shoot as in SD plants. In DN plants, by contrast, the growth is characterized by very low rates of
branch crown formation, with upward growth quickly terminated by a terminal inflorescence and
continued from an upper bud, with reduced development of leaves, runners, and branch crowns.
Interestingly, these growth habits closely parallel those seen in E. vesca by Brown and Wareing
(1965), but in that case were found to associate closely with the runnering locus, or a gene close
by it, rather than the seasonality locus, described below. The runnerless gene from 'Baron
Solemacher' conveyed a growth pattern like that observed in DN plants, while runnerless plants
descended from 'Bush White' had a many-crowned habit like that of EB octoploids, and the wild
type E. vesca was similar to the SD octoploids.
Inheritance Of Flowering Habit In Strawberry
Inheritance of photoperiod insensitivity has been described in a number of species. Most
commonly, it is a single gene trait, conferred by either a dominant allele, as in rice (Oryza sativa
L,) (Chandrartna, 1953), sweet pea (Lathyrus odoratus L.) (Ross and Murfet, 1985), jute
(Corchorus sp.) (Joshua and Thakare, 1986), and tetraploid Sea Island cotton (Gossypium
barbadense) (Lewis and Richmond, 1960), or a recessive allele, as in diploid upland cotton
(Gossypium hirsutum L.) (Lewis and Richmond, 1957), cucumber (Cucumis sativus L.)
(Dellavecchia and Peterson, 1984), or okra (Abelmoschus esculentus (L.) Moench) (Wyatt,
1985). In a few instances, the inheritance has been shown to be a more complicated mechanism,
such as the two dominant alleles at separate loci responsible in some hexaploid wheat (Tritcum
aestivum L.) (Maystrenko and Aliev, 1986) or the three genes believed to be involved in sesame
(Sesamnum indicum L.) (Kotecha et al. 1975).
Brown and Wareing (1965) demonstrated that seasonality in E. vesca is conveyed by a
single gene, which they designated S (referred to as SFL by some later authors (Albani et al.,
2004)), with the perpetual flowering habit displayed by the homozygous recessive. They further
demonstrated that the trait was independently inherited from the non-runnering habit, which
proved to be another recessive trait at a different locus. Federova (1937) may have identified
seasonality in E. vesca as a recessive trait earlier, but it is unclear with which species he was
working. Ahmadi et al. (1990) expanded on this by observing that the California E. vesca, which
differs significantly from the European E. vesca morphologically, appears to have two additional
genes conveying photoperiod sensitivity, yielding a 1:63 ratio in the F2 Of a cross between Alpine
E. vesca and local California plants.
Richardson (1914) conducted some of the first studies of the inheritance of the EB trait.
Using 'St. Antoine de Padoue' and 'Laxton's Perpetual', both derived from 'St. Joseph', as
sources of the trait, he found ratios of 1:1 and 3:1, suggesting a simply inherited character.
A considerably different model was that of Clark (1937), who investigated the
expression of the EB trait from 'Mastodon', an everbearer derived from 'Pan-American', in
breeding populations grown in New Jersey. In general, there was a strong trend towards
producing approximately one third everbearers in crosses of EB with non-everbearers
(presumably SD), with an average of two-thirds in crosses of EB x EB, whereas Mastodon selfed
yielded 80% EB, though in a small population. Clark also cites an earlier EB x EB cross,
reported by Macoun (1924), which yielded slightly less, 56.29% EB. In nearly all cases, crosses
of SD x SD, in which one of the parents had one EB parent, yielded no EB progeny. There were,
however, exceptions. One SD plant with an EB parent, N.J. 220, did give some EB offspring in
one cross. Additionally, two everbearers, N.J. 1 and N.J. 8, produced very low numbers of
everbearers in their progeny in crosses with non-everbearers (0% and 8.8%, respectively) and
when selfed (0% and 11.9%, respectively). Clark concluded that the results suggested a complex,
polygenic inheritance, but with evidence that the character is largely, though not wholly,
conveyed by a maj or, dominant allele, possibly modified by other genes.
Later work attempted to clarify this model, although with limited success. Powers (1954)
performed a partial diallel crossing of three EB and seven SD genotypes and developed a three
locus model with four dominant alleles, A ', A, B, and C. conveying the everbearing trait with
varying strength, ranging from A >A>B>C. Any of these alone was theorized to be inadequate
for expression of the trait, as was aaBC_, but A or A plus a dominant allele at either of the
other loci resulted in the everbearing character. He also theorized that as many as four recessive
genes account for the presence of EB plants in the SD x SD progenies.
Although mostly adequate to explain the observed data (though 5 of his 49 families do
exhibit high chi-square values with this model), Powers' model assumed that all seven of the SD
parents in his study were of the same genotype, aaBbCc. Unless there is a significant unknown
selective value to the heterozygote at the B and C loci, it seems exceedingly unlikely that all
three unrelated short day genotypes would be double heterozygotes.
Two confounding factors may have been at work here. First, the everbearing parents used
are primarily breeding selections, so the source of the trait cannot be readily ascertained. Since
the Cheyenne USDA breeding program was utilizing F. virginiana ssp. glauca in addition to EB
cultivars as parents, it is possible that Powers' observations are the result of a mingling of DN
and EB sources. Secondly, Powers used the production of flowers in July, August, or September
as criteria for EB. In light of the fact that Powers' experiments were conducted in the field at
Cheyenne, Wyoming, it may be that low temperatures allowed flowering to continue even during
the long days of summer. Mid-summer day length in Cheyenne is slightly less than 15 h, whereas
the 1971-2000 mean monthly temperatures for Cheyenne in June, July, August, and September
were 16.4, 19.8, 18.8, and 13.60C, respectively (National Climatic Data Center, 2001). Heide
(1977) found that at 180C, three of four SD cultivars still initiated at least some flowers under 14
h days, and thus Powers may have, in fact, been observing everbearing behavior in non-
everbearing genotypes. Later studies may have been affected similarly.
Ourecky & Slate (1967) examined 46 progenies and suggested complementary dominant
genes segregating in an octoploid manner. This is somewhat at odds with later work suggesting
that octoploid strawberries are diploidized (Arulsekar et al. 1981), despite recent hybrid origin.
More recent work by Sugimoto et al. (2005) found a 1:1 ratio of EB : SD in a cross of 'Ever
Berry' (EB) x 'Toyonaka' (SD), as well as in its reciprocal, whereas 'Ever Berry' selfed gave 3:1
and 'Toyonaka' selfed gave all SD progeny. This is in line with earlier work using 'Ever Berry'
as a parent (Monma et al., 1990; Igrashi et al., 1994), and implies a simple monogenic dominant
Although most authors have considered all sources of the EB trait to be genetically
identical, it may be worth noting that these Japanese studies, as well as Richardson's early
studies (1914), the only studies to show clear monogenic ratios, have used cultivars that probably
carry the everbearing trait derived from the European source (or sources). The studies which
have obtained multigene or confusing inheritance ratios have all used American sources of the
trait: Clark (1937), used 'Mastodon', derived from the 'Pan-American' source; Ourecky and
Slate (1967) used a wide range of cultivars derived from either 'Pan-American' or 'Streamliner'
(a seedling of unknown pedigree); and Powers (1954), as previously mentioned, may have been
using a mix of EB and DN parents. Thus the traits, while similar, may have slightly different
genetic mechanisms, or the American sources possess contain modifying genes not found in the
Characterization of the inheritance of the day-neutrality trait has been clouded by the fact
that studies have indiscriminately included "everbearers" from both the 'Pan-American' source
and true day-neutrals from the E. virginiana ssp. glauca source. At first, day-neutrality was
thought to be the result of a single dominant gene. Ahmadi et al. (1990) found this in an
inspection of nearly 30,000 progeny of crosses between day-neutral and short day plants over the
course of Hyve years. These results firmly agree with a single-gene hypothesis, showing
significant deviation from the model in only one family during one year. Importantly, unlike
early work on everbearers, there were no seedlings that expressed the trait among progenies from
SD x SD. Although the study clearly deals primarily with the glauca source, it is not clear which
remontant cultivars were used and whether all derive from glauca. However, their results differ
significantly from later studies that did include EB from the 'Pan-American' source, which
suggests that all four of the DN cultivars used are derived from glauca.
Serge and Hancock (2005a) also studied the inheritance of day-neutrality, looking at both
cultivated DN and EB cultivars, and apparent novel sources of remontancy found in wild E.
virginiana. Segregation ratios varied widely, apparently suggesting inheritance more complex
than a single gene model. A similar study by Shaw and Famula (2005) using 45 cultivated
genotypes thought to derive their photoperiod insensitivity from the E. virginiana ssp. glauca
source, though not identified in the paper, found strong evidence for the presence of a maj or
dominant locus for day-neutrality, with putative homozygotes flowering more robustly under LD
conditions than heterozygotes. However, it was noted that even homozygous DN plants were not
wholly true-breeding for the trait, suggesting either the effects of other minor loci or deviations
from diploid inheritance.
The system used for scoring progenies may have had a significant impact on the results of
such studies. A primary difference between Richardson' s (1914) studies and many of those that
followed is that Richardson considered everbearers to be individuals that continued flowering
through October, rather than stopping at the end of the summer. Ahmadi et al. (1990) considered
this a more accurate identification of remontant genotypes, but theirs appears to be the only other
study to have done this, as all others reviewed here continued ratings only into August or mid-
September. Serge and Hancock (2003) evaluated five methods of scoring populations for
remontancy, including flowering within 100 days of germination, flowering before a specific
date in the field, flowering under both long and short days in greenhouse or field, and flowering
on newly formed runners in the field. Scoring by flowering within 100 days of germination was
not a good predictor of remontant flowering in the field, but greenhouse observations, if
conducted over the course of an entire season, were well correlated to field performance.
Molecular Markers For Flowering Habit
The ability to screen seedlings for flowering habit at a very early age would be of great
benefit to breeders, saving the time and resources required to plant out and evaluate seedlings
lacking the desired habit. A number of attempts have been made to develop such a system, but
none has seen wide application in breeding.
Albani et al. (2004) used inter-simple sequence repeat (ISSR) markers to identify three
DNA products associated with seasonal flowering in 1,049 plants of a E vesca ssp. vesca
(SFL/SFL) x E. vesca ssp. semperf lorens (sfl/sfl) BC1 population, and successfully converted
these to sequence characterized amplified region (SCAR) markers. Two of these markers were
linked to the seasonal flowering locus at 1.7 and 3.0 cM, while a third, SCAR2, was mapped to
the same location as SFL. This represents the most tightly associated marker for flowering habit
yet developed in Fragaria, but while this work constitutes an important research tool, the impact
of these markers on practical breeding has been limited because of the differences between this
trait and the DN and EB traits of the cultivated octoploid, as well as the lack of commercial
importance of F. vesca.
Kaczmarska and Hortynski (2002) identified a single randomly amplified polymorphic
DNA (RAPD) marker through bulk segregant analysis of a small Fl population, segregating 1:1
for remontancy; however, no attempt was recorded to discern how closely linked this marker was
with the trait or how reliable it was across other progenies.
The previously cited study by Sugimoto et al. (2005) attempted to identify RAPD
markers linked to the EB trait. Five were identified; however the linkages, ranging from 11.8 to
24.3 cM, were rather weak. These weak linkages, coupled with the difficulties sometimes
encountered when trying to reproduce RAPD markers (Paran and Michelmore, 1993), may limit
the practical use of these markers.
Molecular Control Of Flowering In Arabidopsis
Although the effects of genetic variation on flowering, even allelic variation at a single
locus, could be clearly seen in many species, the underlying system often proved complex,
making it difficult to elucidate the roles of individual genes clearly. Among the first steps
towards such an understanding was made through the development of mutant lines of
Arabidopsis thaliana, (a long day (LD) annual species), displaying aberrant flowering patterns.
Reinholz (1945) noted differences in flowering among irradiated Arabidopsis seedlings, and
Redei (1962) followed up on this work by treating imbibed Arabidopsis seeds with X-rays. He
selected four mutants that exhibited altered timing of flowering, designating them constans (co),
luminidependens (ld), and gigan2tea (gi-1, gi-2). All of these lines flowered significantly later
than the wild type control under long days, although the co plants flowered slightly faster under
Subsequent work has shown co to be an important element of a complex network of
regulatory pathways governing flowering which are linked specifically to the perception of
photoperiod. There are three distinct parts to photoperiod perception in Arabidopsis:
photoreceptors that perceive light, an internal oscillator that approximates a 24 h cycle, and the
output path to the meristem identity genes involved in flower initiation (Simpson, 2003). CO
itself lays at the junction between the inputs of light quality and photoperiod on one side, and on
the other side a series of genes, such as FT and SOC1, directly upstream of the meristem identity
genes. A simplified diagram of this network is shown as Figure 2-1.
Components Upstream of CO
The perception of photoperiod begins with photoreceptors. Two groups, the
phytochromes (PhyA, B, D, and E) and the cryptochromes (Cryl and 2), react to red/far-red and
blue light, respectively, to entrain the complex feedback loop of the plant' s central oscillator
(Millar, 2003). The central oscillator, in tumn, produces a number of rhythmic outputs, including
CO expression (Suarez-L6pez et al., 2001). This rhythm regulates the base expression level of
CO; however, CO protein abundance is greatly influenced by a number of factors. Key among
these are the further effects of photoreceptors on transcript and protein stability. PhyB has been
shown to promote the degradation of CO protein under red light early in the day, while in the
evening phyA counteracts the effects of phyB and stabilizes CO protein when activated under
far-red light, as do the cryptochromes under blue (Valverde et al. 2004). Another apparent blue
light receptor, FKF l, has been shown to be required to produce the peak in CO transcript level at
the end of the day required to trigger flowering in Arabidopsis (Imaizumi, et al. 2003). A
member of the same family of proteins as CO, CONSTANS-LIKE9, also appears to repress
expression of CO, though the mechanism is currently unknown (Cheng and Wang, 2005), as
does the very similar CONSTANS-LIKE10 (Cheng and Wang, 2006).
Components Downstream of CO
CONSTANS acts directly on two maj or downstream flowering components, Flowering
Locus T (FT) and SUPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1) (Samach,
2000). Both ultimately act to trigger a suite of meristem identity genes, (in fact many of the same
genes), leading to flowering, and either may be adequate to trigger flowering (Samach, 2000),
though this is disputed (Yoo et al., 2005). In addition to photoperiodic control through CO,
SOC1 also responds to signals from the other pathways regulating flowering in Arabidopsis, the
autonomous, temperature, and gibberellin pathways (Samach, 2000; Moon et al., 2003), and acts
as an integrator of these independent inputs with the common set of meristem genes downstream.
While FT and SOC1 act largely in parallel in inducing flowering (Moon et al., 2005), FT has
been shown able to activate SOC1 to an extent itself (Yoo et al., 2005). Of the two, FT appears
more strongly induced by the photoperiod pathway, while SOC1 is under greater control of the
autonomous pathway and regulated primarily by the gene Flowering Locus C (FLC) rather than
by CO (Moon et al., 2005). In Arabidopsis, another gene, TWIN SISTER OF FT(TSF), which is
very similar to FT has also been shown to perform similar functions (Yamaguchi et al., 2005)
although there is evidence of temperature effects on both genes' activities, suggesting that TSF is
more active at low temperatures whereas FTis more active at higher temperatures (Blazquez et
Primary among downstream genes is LEAFY (LFY), which is activated by SOC1 and
triggers a number of genes controlling floral development, including API (Nilsson et al., 1998).
FT, however, bypasses LFY and activates API directly (Nilsson et al., 1998), as well as other
meristem identity genes such as SEPALLATA3 (SEP3) and FRUITFULL (FUL) (Teper-
Bamnolker and Samach, 2005).
Acting as a counterbalance to FT is TERMIINAL FLO WER1 (TFL1). TFL1 acts to
suppress the reproductive transition in the meristem, repressing expression of APl. TFL1 is a
member of the same family of genes as FT and TSF, and Hanzawa et al. (2005) demonstrated
that a change of a single amino acid residue in TFL1 is adequate to shift its function to that of a
promoter like FT, though the resulting protein is a weaker promoter of flowering than FT. TFL1
expression is elevated in CO-overexpressor lines, and may in fact mirror CO's promotion ofFT
expression to an extent (Simon et al., 1996). Loss of function t/71 mutants are early flowering
and the inflorescences, unlike the indeterminate growth pattern of wild type Arabidopsis, end in
a terminal flower which is often characterized by abnormal floral organs (Bradley, 1997).
The CONSTANS-LIKE Gene Family
CO is the best understood of a large family of genes, all of which share some basic
characteristics. Specifically, all encode proteins which possess a highly conserved region of 129
amino acids called the CCT ("CO, COL, and TOC", after the three classes of genes that possess
it) domain near the C-terminal end, as well as two zinc-finger type domains side by side at the
other end, though the exact amino acid composition of these domains varies. There are also four
small internal domains present in the canonical CO protein that are retained to some extent in
various other members of the COL family. All are transcription factors, though their exact mode
of action is still unknown.
Griffiths et al., (2003) characterized this family in Arabidopsis, rice, and barley, and
found that three general groups were discernable in all three species, along with a fourth in the
two monocot species. They were able to further subdivide this first group into three subgroups.
The first of these, Group la, contains CO itself and a few closely related genes, specifically
COLl and COL2 in the case of Arabidopsis. These Arabidopsis genes have been shown to be
unable to complement co mutants, and thus are not functionally interchangeable with CO, but do
have effects on circadian rhymicity of leaf movements and expression of cab2 transcript (Ledger
et al., 2001). Many species, including Populus deltoides and P. trichocarpa (Yuceer, 2002) and
Bra~ssica nigra (Lagercrantz et al., 2002), seem to possess multiple members of this subgroup,
but it is not known whether these other members are functionally equivalent to CO. This study
represents the first published documentation of Group la COL genes within the Rosaceae.
Group Ib COL genes are known only in monocots (Griffiths et al., 2003), but Group Ic
genes, represented by COL3 -15 in Arabidopsis, are represented in nearly all species investigated
(Jeong et al., 1999; Griffiths et al., 2003; Hecht et al., 2005). They differ from the Group la
genes primarily in the middle region of the gene, where they lack the characteristic M2 and M3
domains (Griffiths et al., 2003). The role of these genes is poorly understood, but at least one,
COL3, has been shown to have effects on photomorphogenesis, flowering, and inflorescence
development (Datta et al. 2006). Zobell et al. (2005) found members of this family to be the only
COL genes present in the primitive plant Physcomitrella patens, suggesting that they may be the
most ancient representatives of the family. Two Group Ic genes from apple, possibly paralogs
resulting from alloploidy in the species' distant past, represent the only COL proteins previously
characterized in the Rosaceae (Jeong et al., 1999). Unfortunately, the authors did not take into
account the diurnal expression pattern of many COL genes when documenting the expression of
the genes, so it is difficult to derive much meaningful expression information from their work.
Another Rosaceae Group Ic gene, designated PrpCo, was mapped in a peach-almond mapping
population (Silva et al., 2005), but did not correspond to any known quantitative trait loci (QTLs)
for flowering time in almond and was not further characterized.
Group Ib, Id, le, and IV COL genes appear to occur only monocots, but Group II and III
genes are present in Arabidopsis and other dicots. The functions of these groups of genes are still
imperfectly known. Very little information on Group II genes appears to be available. A group
III gene, AtCOL9, has been shown to play a role in flowering, apparently through the suppression
of CO, and, as a result, FT (Cheng & Wang, 2005). Early work presented by the same authors
also showed a related gene, AtCOL10, seems to have similar effects(Cheng & Wang, 2006), and
both appear to be expressed in a circadian fashion. A Group III gene from perennial ryegrass,
LpCOL1, has also been shown to be under oscillator control and is implicated in vernalization
(Ciannamea et al., 2006).
Components of the Flowering Pathways are Conserved Among Species
In some respects, Arabidopsis, a long-day annual, might seem a poor model for flowering
in strawberry, a normally short-day perennial in an entirely different taxonomic family.
However, while there are no doubt critical differences in the physiology of these two species,
mounting evidence suggests that the pathways regulating flowering are remarkably well
conserved between taxa, even those separated from Arabidopsis by considerable evolutionary
distance. The components of the photoperiod and floral pathways in rice, a short day monocot
species, have proven nearly identical to and interchangeable with those of Arabidopsis (Izawa et
CO has been identified in a number of other species. Besides the model species of
Arabidopsis and rice, a number of studies have suggested that homologs of CO fulfill
comparable roles in other species. These include the dicots Bra~ssica nigra L. (Robert et al.,
1998), Japanese morning glory (Pharbitis nil (L.) Choisy) (Liu et al., 2001), and potato (Solan2um
tuberosum L.) (Martinez-Garcia et al., 2002; Beketova et al., 2006) as well as monocots such as
rice (Yano et al., 2000), common wheat (Nemoto et al., 2003), perennial ryegrass (Lolium
perenne L.) (Armstead et al., 2005; Martin et al., 2005), and barley (Hordeum vulgare L.)
(Griffiths et al., 2003). In some, although the behavior of CO itself appears to be the same,
completely different photoperiod-dependent processes are being regulated. In potato, tuber
development and other associated morphological changes rather than flowering (Martinez-Garcia
et al., 2002; Rodriguez-Falcon et al., 2006). Although modern potato cultivars have been selected
for relative insensitivity to day length, some wild species such as Solan2um demissum and S.
tuberosum ssp. andigena are strictly dependent on short days for tuber formation (Ewing and
Struik, 1992). whereas in aspen (Populus trichocarpa) CO regulates both flowering and the
processes of bud set and growth cessation in response to shortening days in the fall (Boihlenius et
al., 2006). Expressed homologs of CO are also known in species with no known
photoperiodicity, such as apple (for example, GenBank accession EBl141172 and others) and
tomato (Drobyazina and Khavkin, 2006), though their function, if any, remains unclear.
Not only is CO well conserved among species, but so are most of the other components
of the system regulating flowering, both upstream and downstream. Even rice, a short-day
monocot, contains homologs of nearly all the maj or components described in Arabidopsis (Izawa
et al. 2003). Similar pictures are emerging in poplar (Boihlenius et al., 2006; Hsu et al., 2006;
Mohamed, 2006), potato (Rodriguez-Falcon et al., 2006), and apple (Wada et al., 2002; Kotada
and Wada, 2005; Kotada et al., 2006) and this seems likely to be the case in most plant species.
Although the molecular basis for the variation in flowering habit seen in Fragaria is still
unknown, such variations in other crops have been shown to related directly to alterations in this
small group of critical genes. Researchers in Japan have demonstrated that two maj or QTL
affecting heading date, Hd1 and Hd3a, correspond to orthologs of CO and FT, respectively
(Yano et al., 2000; Kojima et al. 2002), whereas a third heading-date QTL, Hd9, has been found
to be tightly linked to the rice homolog of SOC1 (Tadege et al., 2003). Photoperiod insensitivity
in barley has been shown to result from disruptions in the circadian oscillation of CO caused by
mutations in Ppd-H1, a member of the pseudo-response regulator class of genes involved in
circadian clock function (Turner et al., 2005)
Using Model Systems To Understand Development In Rosaceae
Research to date has provided a detailed understanding of the gross physiology of
flowering and development in fruit crops such as strawberry, while work in models such as
Arabidopsis and rice have provided most of what is known about the mechanisms at work at a
molecular level. To truly capitalize on these bodies of work it is necessary to find ways of
integrating information from both.
This is already beginning to occur in apple. Studies have identified and characterized
apple orthologs of many of the components known in Arabidopsis, including AFL1 and AFL2
(apparent equivalents of LFY), MdAP1, MdTFL1, MdIFT, and over a dozen MADS-Box
transcription factors involved in reproductive development (Sung et al., 1999; Yao et al., 1999;
Kotoda et al, 2000, 2002; Wada et al., 2002; Kotoda and Wada, 2005). In general, such research
has shown the components in apple to play roles similar to the homologous Arabidopsis genes.
MdAPI and the pair of LEAFY homologs, MdLFY-1 and M~dLFY-2, have been found to promote
flowering in transgenic Arabidopsis, and appear to be involved in the promotion of flowering in
apple as well (Kotoda et al., 2000; Wada et al., 2002). Similarly, MdTFL1 has been shown to
retard the transition to flowering in transgenic Arabidopsis (Kotoda and Wada, 2005) and
suppression of2~dTFL1 through expression of antisense construct of the gene has been
demonstrated to shorten the vegetative juvenile phase of 'Orin' apple trees from more than six
years to as few as eight months (Kotoda et al., 2006). Some studies, however, have shown that
mechanisms in apple behave differently than would be predicted by the Arabidopsis model. For
example, overexpression of AFL1 in apple, unlike overexpression of the similar LFY in
Arabidopsis, does not result in precocious flowering (N. Kotoda, unpublished research cited in
Kotoda et al., 2006). Regulation of LFY in Arabidopsis is primarily controlled by photoperiod
(Samach, 2000) and apple, an apparently day-neutral species, may regulate flowering through a
pathway not primarily dependent on LFY orthologs.
Classical physiology has already offered numerous hints of what may be occurring in
strawberry at the molecular level. Several researchers (Vince-Prue and Guttridge, 1973;
Thompson and Guttridge, 1960) have suggested that the evidence supports the existence of a
phloem-mobile inhibitor of flowering, synthesized in the leaves. This is reminiscent of the
inhibitory effects of CO in rice (Yano et al., 2002), on FT, which along with TFL1, has been
shown to move from the leaves to the meristem in Arabidopsis (An et al., 2004). Similarly, the
effects of light quality on flowering as shown by Vince-Prue and Guttridge (1973) suggest the
possibility of phytochrome-mediated regulation of a hypothetical CO homolog. To truly
understand what is occurring within the plant in each of these cases, however, will require the
development of the needed molecular tools.
The well-documented complex network of factors affecting flowering in strawberry:
photoperiod, light quality, temperature, vernalization, all have corresponding mechanisms
documented in the molecular models. These mechanisms may not prove identical in strawberry,
and in some cases may prove radically different, yet this knowledge can act as a foundation for
investigations into the interactions between genes and environment that control vital aspects of
the crop's development.
To date little such work has been conducted in strawberry. Yet in many ways strawberry
represents an attractive target for such studies--it has clearly delineated flowering phenotypes, a
short life cycle, ease of hybridization, and convenient plant size. The octoploid genome number
and alloploid nature of the cultivated species represent challenges, but the many diploid relatives
have much simpler and smaller genomes. Like apple, strawberry is an important crop species,
and a better understanding of the genes critical to its development might have economic impacts.
Such knowledge might lead to improved cultivars and more efficient cultural practices,
benefiting the farmer, consumer, and environment.
Photo rece pto rs
Te mpe ratu re
Circadian oscillator CONSTANS
transcription (CO) s
TFL 1 ~ FT -
Meristem Identity Genes
F LOWVERIN G
Figure 2-1. Simplified diagram showing interactions of genes and environmental factors
governing flowering in Arabidopsis thalianaiiii~~~~~~iiiiii (arrows represent promotive effects,
bars represent repressive effects).
CHARACTERIZATION OF CONSTANS-LIKE GENES IN Fragaria
The initiation of flowering is a critical developmental milestone in most plant species and
represents a vital process for both natural evolution as well as crop production and development.
The mechanisms involved in switching between vegetative and reproductive development have
been shown in model species to be precisely regulated by a number of pathways, governed by
both internal and environmental factors. Among these is photoperiod, which is critical to the
growth and development of many crop species, including the cultivated strawberry. As in many
other species, the timing of flowering in strawberry is an important factor with major impacts not
only on the adaptation of cultivars to particular locales or cultural methods, but also, because of
the relationship with the timing of the subsequent harvest, plays a vital role in determining the
market value of the crop as well.
Despite a narrow genetic base (Sjulin & Dale, 1987), commercial strawberries display a
wide spectrum of sensitivities to photoperiod, ranging in phenotype from strictly short-day (SD)
and "long-day" everbearing (EB) cultivars to day-neutral (DN) types that flower regardless of
photoperiod. Because these different types share overwhelmingly similar genetic backgrounds
and may even co-exist in the progeny from a single cross (Ahmadi et al., 1991), the differences
are likely attributable to the qualitative and quantitative attributes of a small number of critical
genes, thus strawberry provides a unique model system to investigate this important pathway.
The ability to characterize the molecular basis for these differing phenotypes could have
important implications for cultural techniques and cultivar development not only in Fragaria,
but also potentially in the other cultivated members of the Rosaceae, for which timing of
flowering is also a critical factor.
Studies in strawberry allow direct translation of molecular paradigms devised in long-day
and short-day model systems to a single species with ranging photoperiodic flowering habits.
Studies in Arabidopsis have shown that the gene CONSTANS (CO) plays a critical central role in
the regulatory pathway responsible for photoperiodic flowering (Simpson and Dean, 2002;
Corbesier and Coupland, 2005), acting as a dynamic link between photoreceptors, the circadian
oscillator, and the genes governing meristem-identity. Mutations in the CO gene cause delayed
flowering in Arabidopsis, a LD plant (Putterill et al., 1995), and early flowering in rice, a SD
plant (Yano et al. 2000). Its function as a regulator of flowering is well-conserved across species,
having been characterized in both other eudicots (Robert et al., 1998; Liu et al., 2001) and
monocots (Yano et al., 2000; Martin et al., 2004; Nemoto et al., 2003), though the downstream
effect varies between species, either suppressing or promoting flowering. CO is one member of a
large family of CONSTANS-like genes, numbering 17 in Arabidopsis, 16 in rice (Oryza sativa) ,
and at least nine in barley (Hordeum vulgare) (Griffiths et al. 2003). No accounting of these
genes appears to have been previously published in strawberry.
The high degree of conservation in both structure and function, even between monocots
and eudicots, suggests that COL genes likely have important roles in controlling flowering time
across species. A similar situation is likely the case within Fragaria, yet its diverse flowering
habits suggest that regulation may be imparted through novel alleles of known genes,
uncharacterized patterns of expression, or pathways that circumvent photoperiodic influence by
acting on elements downstream from the photoperiodic regulators. This study proposes to
identify, catalog and characterize probable COL genes in Fragaria, describe tissue specific,
temporal and photophysiological expression patterns. The results of this study provide additional
information about into the mechanisms that regulate photoperiodic development. These results
have implications for the evolution of the gene family in Rosaceae.
Because of the difficulties in coordinating the numbering of family members across
species that may possess different numbers of such genes, a decision was made to simply number
the COL genes in the order they were found, rather than attempting to match numbering of
Arabidopsis genes. In this report, FrCOL is used to refer to genes at a particular locus across
members of the genus Fragaria, whereas FraCOL, FrvCOL, FrnCOL, and FriCOL are used to
designate genes of E xanana~ssa, E. vesca, E. nubicola, and E. iinumae, respectively. Aside from
numbering, genes were named for their apparent Arabidopsis homologs.
Cloning And Identification Of Four CONSTANS-LIKE Genes Of Fragania
A series of EST libraries have been generated by this laboratory, representing transcripts
accumulating in mature xanana~ssa plants (Folta et al., 2005), developing flower buds, various
root tissues, 12 fruit stages (K. Folta et al., unpublished) and E. vesca seedlings (J. Slovin et al.,
unpublished). These EST collections served as a basis for gene discovery based on direct
sequencing, colony hybridization and PCR with degenerate primers.
Simple homology comparisons to sequences within public databases identified
convincing homologs of at least Hyve distinct Constans-like (COL) genes. Four were designated
FrCOL1, FrCOL2, FrCOL3, and FrCOL4, whereas the fifth, which shared the greatest
homology with AtCO and the rice Hd1, was presumed to be the Fragaria ortholog of CO and
designated FrCO. These genes represent the canonical three distinct subgroups of the COL
family as outlined by Griffiths et al. (2003) (Figure 3-1). Close matches to each of these genes
were found among publicly available ESTs in other Rosaceae species as well (Table 3-1).
A full-length cDNA clone ofFraCO was obtained from a flower tissue library derived
from E. xanana~ssa 'Strawberry Festival'. Seven individual clones were identified and
sequenced, and all contained transcripts completely identical except that two contained a
somewhat truncated version of the 3' UTR. The 1,197 bp coding region shows considerable
similarity to AtCO and other previously characterized homologs, with the most similar known
transcript being PdCO1, from Populus (Table 3-1). Comparison of the amino acid sequence to
those proteins designated as Group I suggested that FrCO belongs in Group la (Figure 3-2), a
division that includes all genes experimentally demonstrated to complement the Arabidopsis co
mutant and those believed most likely to play an important role in flowering behavior. The
predicted protein contains a CCT domain and two intact B-Box regions, more closely resembling
similar proteins in other eudicots, rather than the less conserved second B-Box present in some
monocots (Martin et al., 2004). All four middle domains common to Group la COL proteins
were also identified.
Full-length coding sequences for three other FrCOL genes identified could be deduced
from overlapping E. vesca ESTs, as well as part of a fourth. Although these genes contained the
basic structures of COL genes, they bore less similarity to AtCO than did FrCO and appear
unlikely to be the Fragaria ortholog of CO. These were designated FrCOL1, FrCOL2, FrCOL3,
FrCOL1 and FrCOL2 both encode predicted proteins consistent with Group Ic COL
genes, with the maj or difference from the Group la genes being the presence of the M2c (Zobell
et al., 2005) rather than the M2 domain (Figure 3-3). Several copies of both were identified
among E. vesca EST sequences, as well as a single, slightly different, E. xanana~ssa 'Strawberry
Festival' FrCOL1 sequence. FrCOL1 is roughly equal in similarity to both AtCOL3 and AtCOL4
at the amino acid level (55%), with AtCOL3 having slightly higher similarity in the maj or
functional domains. It is most similar to M~dCOL1 and M~dCOL2 from apple, the only previously
characterized COL genes in the Rosaceae (Jeong et al. 1999), and to PrpCO, a COL gene from
peach mapped by Silva et al. (2005). FrCOL2, while retaining a similar structure at the amino
acid level, was not closely related to FrCOL1 at the nucleotide level, suggesting the two genes
diverged some time ago. Among Arabidopsis COL genes, the predicted protein most closely
resembled that of AtCOL5. Based on their relative frequency in BLAST-X searches of Rosaceae
ESTs, FrCOL1 and its orthologs are possibly among the most commonly expressed COL genes
in the family, whereas FrCOL2 and its homologs are expressed at a somewhat lesser level (data
FrCOL3 was the least CO-like of the transcripts identified, with a structure consistent
with a classification in Group II. FrCOL3 was found only in F. vesca libraries derived from
young seedlings, and although matches were found among Rosaceae ESTs (primarily in an apple
seedling library), it appears to be a much rarer transcript than the FrCOL1 and FrCOL2 genes.
Only a partial sequence for FrvCOL4 was found among the sequenced ESTs, comprised
of the B-Box region and much of the 5' UTR. Although the full sequence was not available, this
clearly represents a gene distinct from those previously described. The closest match in
Arabidopsis was AtCOL10, indicating this is a likely member of Group III, and if so, all four
groups present in Arabidopsis have also been identified in Fragaria (Figure 3-2). No members of
Ib, Id, le, or IV were found, none of which have thus far been identified in eudicots (Griffiths et
FrCO/FrCOL Gene Structure
Group I and II COL genes in Arabidopsis possess a single intron located about a third of
the length from the 3' end of the coding region, between the M3 and M4 regions (Fig. 3-4).
Using internal primers designed to include ~400 bp of the coding region and to flank the putative
intron site (Table 3-2), based on the previously identified cDNA, partial genomic versions of
FrCO, FrCOL1, and FrCOL2 were amplified and sequenced from genomic DNA of E
xanana~ssa 'Strawberry Festival', E. vesca FDP815, and E. nubicola FDP60 1 (Figure 5-4.)
Additionally, a product was also amplified in xanana~ssa flower cDNA, using the same primer
In all cases, an intron was shown to be present at the predicted site in the amplified
genomic product (Figure 3-4). FrCO intron length ranged from 704 to 746 bp. Introns in F.
nubicola, E vesca, and three nearly identical alleles from E. xanana~ssa (FaCO-35) were highly
similar in both sequence and length (704, 707, and 706 bp, respectively), whereas a second F.
xanana~ssa allele (FraCO-C9) was considerably different, containing a 28 bp insertion and
numerous other polymorphisms, for a total length of 746 bp. Somewhat similar to this allele
were those of E iinumae and E. nubicola (Figure 3-5). Alleles also differed significantly in
sequence at one site in the coding region, between the Ml and M2 domains. This region appears
to be characterized by highly repetitive sequence containing multiple CAA codons, and is poorly
conserved between and even within species (Figure 3-5A,B). FraCO-C9 and E. iinumae were
considerably divergent in this region from the other alleles sequenced (Figure 3-5). FrCOL1 was
shown to contain a 82 bp intron in both E.x anana~ssa and E. vesca, similar in length to the
related AtCOL3 (101 bp), but considerably smaller than that ofAtCOL4 (403 bp). Additionally, a
12-bp insertion was identified in a single 'Strawberry Festival' EST, designated 7C. A primer
landing in this insertion amplified a product in all eight octoploid cultivars tested, but not in any
of the diploids, suggesting that this particular insertion may be unique to the octoploids.
FrvCOL2 contained a 118 bp intron, whereas FrnCOL2 a 119 bp intron, both similar in
length to the 88 bp intron of AtCOL5. The cDNA product for FrCOL1 was the same size as that
predicted from the sequenced transcript. However two product sizes were observed for FrCOL2,
consistent with the presence of an unspliced transcript variant, and a partial EST transcript
containing nearly all of the intron sequence was identified among E. vesca sequence data.
Southern blot analysis of FrCO Copy Number
Southern blot analysis displayed a banding pattern consistent with a single FrCO locus in
E. vesca, as in other diploid organisms, showing two bands in those lanes where the probe
fragment was cut once by the chosen enzyme, and one in those where it was not cut. The
cultivated octoploid, E. xanana~ssa, however, showed multiple bands, frequently as many as six
for enzymes that cut with in the probe region (Figure 3-7).
Mapping Of Fragaria CO And COL Genes In A Diploid Mapping Population
The fragment of FrnCO generated by the RN-FaCO primers was found to lack one of the
two HindlII enzyme restriction sites possessed by the corresponding fragment of FrvCO, and this
was used to generate a CAPS marker for mapping (Figures 3-8A, 3-8B). This marker segregated
in a 14:24: 18 (aa:ab:bb) ratio in a subset of the reference population developed by Sargent et al.
(2004), and was found to map to linkage group VI, with linkages to SSR markers E1VFnl53,
UJDFO25, FAC-005, and EMFv160AD, with LOD values of 15.48, 7.79, 11.15, and 13.71
respectively (Figure 3-8C).
No polymorphisms were found between E. vesca FDP815 and E. nubicola FDP60 1 in the
intron and flanking coding region ofFrCOL1. Two polymorphic restriction sites were identified
in FrCOL2, and mapping was attempted first using EcoRV. Because of suspect segregation
ratios and banding patterns that were difficult to resolve, a decision was made to score the
marker as a dominant marker, but the incongruous segregation ratio persisted, yielding a 40:41
(aa:b_) ratio. The first 24 individuals were also scored using the second enzyme, AflHII, yielding
the same pattern seen among these plants with the first enzyme. The resulting mapping data did
not convincingly place the gene on any of the seven linkage groups or in a strong linkage
relationship with any of the established markers.
Comparison Of Strawberry Genotypes Under Varying Photoperiods
In order to test the expression of FrCO and other genes under various photoperiods, a pair of
growth chamber experiments were performed. The first growth chamber experiment consisted of
a comparison of the SD cultivar Strawberry Festival with the DN cultivar Diamante under 8 and
16 h photoperiods. After several weeks under these treatments, these genotypes reacted as
predicted, with Strawberry Festival' flowering heavily under SD and producing numerous
runners under LD, and 'Diamante' flowering vigorously under both photoperiods. The DN plants
produced runners only under the 16 h photoperiod, and even then only produced a very few
runners. Although observations were made, no flower or runner data were collected in this
The second growth chamber experiment, under cooler temperatures, compared
'Diamante' to the SD cultivar Camarosa, and to a selection of F. vesca, Hawaii-4. 'Diamante'
again flowered under both conditions, more vigorously under LD than SD. 'Camarosa' did not
runner and flowered only under SD. Unexpectedly, Hawaii-4, a yellow-fruited diploid selection
presumed to be closely related to F. vesca ssp. semperflorens and previously classified as not
photoperiod-sensitive (Oosumi et al., 2006), behaved in a clearly SD fashion, flowering
profusely under an 8 h photoperiod and runnering primarily under LD (Table 3-4).
The expression pattern ofFraCO in the first study was studied by RNA-gel (northern)
blot and subsequent hybridization to a radio-labeled CO probe. Transcript levels varied both
between plants grown under SD and LD conditions, as well as between SD and DN genotypes
(Figure 3-9). Under LD, transcript level in 'Strawberry Festival' peaks rather weakly before
dawn, whereas under SD a strong peak appears in the early part of the day. Transcript levels in
'Diamante', however, are lower under both conditions, with a broader peak in the afternoon,
later than in 'Strawberry Festival', and almost no transcript present under long days.
One concern was that RNA-gel blot experiments may not give precise measurement of
CO transcripts because of potential cross-hybridization with other family members. To verify the
results, gene-specific primers were developed and used to evaluate steady-state RNA levels in a
similar experiment, this time using semi-quantitative RT-PCR. Similar results were seen when
the growth chamber experiment was repeated with 'Diamante', 'Camarosa', and Hawaii-4
(Figure 3-10) using these methods. All cDNA samples were tested with primers for ubiquitin and
the ubiquitin control was used to balance template amount across samples within each set.
Allele-Specific Expression Patterns Differ Betwen Short-Day And Day-Neutral Genotypes
Although allele-specific primers for the FraCO allele types 35 and C9 amplified products
from genomic DNA of 'Strawberry Festival' and 'Diamante', 33 cycles of RT-PCR showed that
while both alleles are actively expressed in 'Diamante', only the FraCO-35 allele appears to be
significantly expressed in 'Festival' at dawn under short days, although a trace of FraCO-C9 is
visible. A more visible band appeared at 40 cycles, suggesting that trace amounts may still be
expressed, although this may be the effect of buildup from slightly non-specific priming (Figure
3-11). Additionally, though the amount of FraCO-35 and FraCO-C9 product present at 33
cycles in 'Diamante' seems similar, the amount of FraCO-35 product present in 'Strawberry
Festival' appears higher than either allele in 'Diamante'. No differences were noted in the
expression of the 7C allele of FraCOL1, which appeared to be actively and equally expressed in
both cultivars (not shown). As before, template levels were adjusted by equalizing with ubiquitin
Expression Of Other Genes Under Different Photoperiods
Expression of other COL genes in these samples was also been investigated using RNA
from the preceding experiments to see if their expression correlated in some way with relevant
phenomena. From the first experiment, cDNA was generated from RNA taken at the 5 AM time
point (the point of peak FraCO expression under SD conditions, corresponding with hour 8 in
Figure 3-9) and from the low point of FraCO expression, 5 PM, in the SD and LD 'Strawberry
Festival' and 'Diamante' sample sets. These cDNA samples were used as template for RT- PCR
to compare quantitative differences in expression of the CO and COL genes, as well ubiquitin
and actin as controls (Figure 3-12). The previously noted difference in FraCO level was
observed again, and was accompanied by higher levels of FraCOL2 under short days, but
FraCOL1 appears more highly expressed under long day conditions. No measurable levels of
FraCOL3 transcript were observed under either condition in either cultivar. All transcripts
except FraCOL1 were undetectable at the 5 pm time point under both SD and LD conditions,
coinciding with the low point in FraCO expression seen in the previous experiment, suggesting
that they too may cycle diurnally. Control genes were uniform between treatments, suggesting
that the differences observed in specific targets were reflective of actual variations in transcript
The same process was repeated for the entire diurnal time course ofF. vesca Hawaii-4
from the second photoperiod experiment. In this case, very little expression of FrvCOL1 and
FrvCOL3 were seen. FrCOL2 showed some evidence of circadian cycling, but not the large
difference in amplitude between SD and LD seen with FrCO, and in fact seemed to show
evidence of a 12 hour period of variation (Figure 3-13).
Some FraCOL2 Transcripts Contain An Unspliced Intron
Under some conditions, a second product of higher molecular weight was visible for
FraCOL2. The predicted size of this fragment was consistent with the retention of the 1 18-1 19
bp intron shown in the diploid FrCOL2 genes. Additionally, a partial FrvCOL2 EST, DY674801,
was identified in Genbank with more than 95% of the intron, the sequence ending only a few
bases short of the end. Conditions under which the unspliced transcript was detected varied
between genotypes. In Strawberry Festival', it was present only under LD conditions, at the 5
am time point, whereas in 'Diamante' it was visible only under SD conditions, but at both 5 am
and 5 pm (not shown). In F. vesca, in the second trial, however, it was clearly visible at two
points of the day under short photoperiods (Figure 3-12).
The Co Gene Family In Fragaria (Rosaceae)
The results of these studies clearly demonstrate the presence of a family of COL proteins
in Fragaria similar to those previously demonstrated in a number of other species, accounting
for all subgroups demonstrated in Arabidopsis and including a likely homolog of CO.
The five members of the COL family identified in this work cover the full range of COL
diversity described in the Arabidopsis, the only eudicot in which the entire family is known to be
documented, with representatives in all subgroups. Unlike in Arabidopsis, only a single member
oflIa, FrCO, was identified. Unless there is significant deviation at the nucleotide level, the
possibility of other la COL genes in Fragaria seems remote. Both the results of the F. vesca
Southern blot, and the fact that no others were isolated in two hybridization experiments, one
probed with AtCO and the other with FraCO, suggest that other such genes may not exist in
strawberry. No evidence of multiple group la loci was seen among ESTs from other diploid
Rosaceae species, although two distinct but closely-related classes of CO were found in apple.
Multiple group la genes are common, even in diploid species, however, and have been
demonstrated in poplar, tomato (Lycopersicon esculentum Mill.), barley, and oilseed rape
(Bra~ssica napus L.) as well as in Arabidopsis (Fig. 3-1), but have been shown not to be the case
in rice (Yuceer et al., 2002; Ben-Naim et al. 2006; Griffiths et al., 2003; Robert et al., 1998). It is
possible that rather than possessing distinct roles, the non-CO group la genes in these species are
merely the result of duplication, coupled perhaps with at least partial loss of function within the
genome, although AtCOL1 has been shown to shorten the circadian cycle when overexpressed in
wild type plants (Ledger et al. 2001). AtCO and AtCOL1 are very similar, even at the nucleotide
level, but AtCO overexpressors show no change in circadian cycle. However it is possible that
reciprocal mutations have separated functions of the original gene into the two genes, with the
current AtCO retaining the ability regulate flowering and AtCOL1 possessing a role in circadian
Although FrCO mapped to the same linkage group in the diploid population as the
seasonal flowering locus (SFL), there was no significant linkage with this locus, as 38cM
separate the two, casting doubt on the idea that non-seasonal F. vesca may be FrCO mutants.
Neither of the two QTLs for date of first flower in the diploid mapping population mapped to the
location of FrCO (Sargent et al., 2006a). This suggests that variation in FrCO is not a significant
source of variation in flowering time in this population, and that unlike in other species, such as
rice (Yano et al., 2002), the lack of seasonality is not the result of mutations in the CO gene.
Homologs of all Fragaria COL genes were found in ESTs from other Rosaceae species,
though in no case were homologs of all Hyve genes found in a single species (Table 3-3). This is
most likely due to the limited number ESTs available from a given species, as COL proteins are
generally relatively low in abundance and only apple begins to approach the large numbers of
ESTs need to reliably locate rare transcripts.
Searches of publicly available ESTs identified group la genes in both Malus and Prunus
species, and it appears that there may be two distinct members of the group in apple. Unlike
strawberry, apple does not appear to be photoperiodic (Carew and Battey, 2005). However,
group la COL genes nonetheless appear to be present in all higher plant species thus far
investigated, including some that lack photoperiodic sensitivity, such as tomato (Lycopersicon
esculentum) (Ben-Naim et al. 2006).
Group Ic, homologs of AtCOL3 -1 5 represented in Fragaria by FrCOL1 and FrCOL2,
appear to be among the most highly expressed COL genes in the Rosaceae, constituting 6 of 19
COL ESTs identified in Fragaria, 13 of 20 in Prunus, and 87 of 106 in Malus, of which a large
maj ority are homologs of FrCOL1. Prior to this study, the only Rosaceous genes yet
characterized were members of this group, a single pair from Malus domestic, M~dCOL1 and
M~dCOL2 (Jeong et al, 1999). A very similar gene designated PrpCO was also mapped to Prunus
linkage group G1 on P. dulcis x P. persica F2 map by Silva et al. (2005), but was not further
The group Ic genes may be the most ancient members of the family, being the only COL
genes identified in the primitive plant P. patens (Zobell et al. 2005). As with all non-CO
members of the family, the function of these genes is still unclear. Zobell (2006) found no
discernible phenotype in single mutant knockouts for any of the three members in P. patens.
However, work by Datta et al. (2006) has shown early flowering and changes in lateral
branching and root development in Arabidopsis COL3 knockout mutants, and have suggested a
role as a positive regulator of red light signaling. FrCOL1 and its homologs may have similar
functions in the Rosaceae.
Although two polymorphisms suitable for mapping were detected, mapping FrCOL2 in
the diploid reference population proved difficult, and it was not possible to identify linkages with
any published markers. Scored as a dominant marker, one would anticipate a segregation ratio of
3:1, but in this case the pattern was close to 1:1, skewing towards the E. vesca parent. Some parts
of this map have shown considerable distortion (Sargent et al., 2004a, 2006b), including one
marker, BFACT-010, that segregated 79:5 rather than 3:1 (Sargent et al., 2006b). This distortion
of segregation ratios may be due to the parents chosen as parents of the reference population.
Although Sargent et al. (2004a) produced their population from two selfed Fl E. vesca x F.
nubicola hybrids, previous researchers found Fl plants to be self-incompatible, like the F.
nubicola parent (Evans and Jones 1967) or sterile (Dowrick and Williams, 1959). Although the
mechanics of self-incompatibility have yet to be elucidated in strawberry, it may be that this
system remains functional to an extent, resulting in bias in what pollen may successfully fertilize
the Fl plants. Additionally, although sexually compatible, E. vesca and E. nubicola do not
overlap in their ranges and are distinct on both a morphological and molecular level (Sargent et
al. 2005, 2006b; Potter et al., 2000) suggesting that there has been time for divergence between
these species. Such divergence might result in imperfect pairing of chromosomes, and if large-
scale rearrangements exist between the species, may even be reflected in the formation of
multivalents and gametes with chromosomal abnormalities.
FrCOL3 is the sole representative of group II COL genes identified. While maintaining
the general COL structure, group II genes have only one B-Box and the four conserved middle
regions. Though common to both monocots and dicots, comprising 4 of the 17 Arabidopsis COL
genes and 3 of the 16 in rice, their function remains unknown, and this may represent the first
report of such a gene outside of these two model systems. Apparent homologs of FrCOL3 were
identified among ESTs of only one other Rosaceae species, peach, although a distinct transcript
matching the criteria of a group II COL gene was found among apple ESTs as well.
Mining of EST libraries revealed only one partial transcript of FrvCOL4, consisting of
the 5' end the gene. Although we can only guess at the rest of the gene' s structure, this end is
clearly characteristic of a class of COL gene not otherwise represented by the other FrCOL
genes, group III. The B-Box region of group III genes is distinctive because instead of a pair of
B-Boxes, these genes have a single B-Box followed by a modified zinc-finger region. Although
impossible to observe in an EST, these genes are also distinguished from the others in the family
because of their differing intron structure, with two splice locations in the middle of the gene and
a third within the CCT region, none of which correspond to the intron location in the other
family members (Griffiths et al. 2003). This suggests that the differentiation from the other COL
genes is very ancient.
The function of this class of genes is still imperfectly known, but AtCOL9 has been
shown to play a role in flowering, apparently through the suppression of CO, and, as a result, FT
(Cheng & Wang, 2005). A related gene, AtCOL10, was also shown to have similar role in early
work presented by the same authors (Cheng & Wang, 2006), and both appear to be expressed in
a diurnal fashion.
Evolutionary Relationships Among FrCO Sequences
The relationships between the octoploid FrCO alleles and those of the diploid species
reinforce some aspects of the established views of the origins of the diploid genomes of F.
xanana~ssa, while perhaps casting doubt on others.
One allele class, FraCO-35, bears striking resemblance to that of E vesca. Within the
coding region, only the previously mentioned variable site between the Ml and M2 regions
differs, and in fact this region differs considerably even between two selections ofE. vesca,
Hawaii-4 and FDP815. Even within the intron, the nucleotide sequence is 98% identical between
FraCO-35 and E. vesca (Figure 3-5).
The idea that E. vesca may represent a diploid progenitor of the octoploids is an old one,
suggested first by Ichijima (1926) and reinforced by further work, both cytogenetic (Senanayake
and Bringhurst, 1967) and molecular (Potter et al, 2000; Folta and Davis, 2006; also Chapter 5 of
this dissertation) and this result seems to strengthen this view.
The origin of the other component genomes has never been as clear, though a number of
species have been suggested. The sequence of the other type of octoploid allele, FraCO-C9,
does not seem to clearly implicate any of these, but may indirectly point to two species, F.
nubicola and E. iinumae. E. nubicola has long been discussed as a progenitor of the octoploid
species, and was suggested as a possible source of the A or A' genomes by both Senanayake and
Bringhurst (1967) and Potter et al. (2000). Staudt (1989) had proposed E. nubicola, along with E.
vesca, as the ancestors of the hexaploid species E. moschata, but sequence data analyzed by
Potter et al. (2000) implicates E. orientalis, and Lin and Davis (2000) showed that E. viridis was
the likely cytoplasm donor for E. moschata. The range ofE. nubicola, today confined to a
relatively small area of South Asia, does not overlap the ranges ofE. moschata, E viridis, or E.
vesca. If E nubicola is in fact an ancestor of both E. moschata and the octoploid species, this
may suggest that it or a closely related species once existed across a much broader geographic
E. iinuntae has primarily been implicated in unpublished work by DiMeglio and Davis
(cited in Folta and Davis, 2006). Currently, populations ofE. iinuntae are confined primarily to
the island of Hokkaido and the Kuril Islands. The species appears to be an outlier from the rest of
the genus both at the morphological level (Sargent et al., 2004b), and at the molecular level
(Harrison et al, 1997; Potter et al., 2000, Sargent, 2005). E. iinuntae has proven incapable of
producing hybrids with other diploid species (Bors, 2000; Sargent et al., 2004b) but polyploids
derived from the species do not appear to have been investigated and may in fact be capable of
producing allopolyploid offspring in combination with other species.
No diploid species contained the 28bp insertion in the intron of the FraCO-C9 allele, and
E. iinuntae contained 5bp and 7bp intron deletions not seen in FraCO-C9 or any other allele.
However, an inspection of SNPs within the intron on either side of the 28 bp insertion reveals an
interesting pattern--on the upstream side of the insertion, C9 and E. nubicola share the same
base at 7 of 8 SNP sites, compared to only 1 between C9 and E. iinuntae. Downstream of this
insertion, however, more SNPs match E. iinua~ne, with 9 of 18, compared to 6 matches between
C9 and nubicola and 3 sites where C9 did not match either species. Small microsatellite repeats
in the intron, all downstream of the insertion, also seem to more closely resemble E. iinuntae,
with two of the three identical between the two alleles (though the difference between E.
nubicola and the other two alleles is only one repeat in both cases), and the third SSR containing
five repeats of AT in FraCO-C9, four in E. iinuntae, and three in all other sequenced alleles. The
coding regions of the gene are nearly identical among all alleles, with the exception of the
variable region between the Ml and M2 domains, which matched exactly between FraCO-C9
and E. nubicola. Given the level of polymorphism in this region, with even members of the same
species showing variability, this evidence would seem to make a strong case that at least this
portion of the gene originated in E. nubicola.
The division seemingly indicated by the 28 bp insertion may represent a slightly uneven
crossing over event in the species' past, creating a chimeric gene with an upstream end derived
from E. nubicola or a near relative, and a downstream end derived from E. iinumae or an ancient
related species. If this locus lies in a region characterized by frequent crossover or disruption by
interchanges between genomes, this may explain the failure to express of the FraCO-C9 allele
seen in 'Strawberry Festival', for example because of damage to the promoter. Nemoto et al.
(2003) found that the CO locus in one of the three genome pairs of hexaploid was not expressed,
despite an intact coding region, and traced this to an 63 bp deletion in the promoter upstream
from the gene.
It seems unlikely that these two classes represent the entire octoploid complement of
FraCO allele types. Only FraCO-35-type sequences were identified in sequencing of cDNA
from 'Strawberry Festival', and only two identical FraCO-C9 sequences were identified (by the
size difference caused by the insertion and by failure to cut with Xbal) among 16 clones
amplified from genomic DNA of 'Strawberry Festival'. It is possible that alleles are not being
identified by the PCR-based methods used in this study due to mismatches in primer sequence,
or that the FrCO locus has been eliminated from one or more of the component genomes of the
In previously characterized SD plants, such as rice (Yano et al., 2000) and perennial
ryegrass (Martin et al., 2004), CO expression patterns remained about the same as that seen in
Arabidopsis. The differences in flowering behavior come as a result not of differences in CO
expression pattern, but from differences in the output, with CO acting as an inhibitor of FT and
SOC1, and hence flowering, rather than promoting them as in Arabidopsis.
In strawberry, it appears that something quite different is occurring. Under short days,
FrCO transcript levels conform to a robust diurnal cycle, though the peak is shifted so that unlike
under SD in Arabidopsis, it occurs during the day, rather than after dusk. Under long days,
however, there is little FrCO expression at all, though there is a perceptible peak during daylight
hours as well. Based on transcript accumulation alone, this suggests that FrCO, like AtCO, is a
promoter of flowering, and that the lack of flowering under long days does not necessarily derive
from an inhibition of flowering, but instead from a lack of promotion, and possibly inhibition
through other mechanisms. Interestingly, even the DN cultivar, Diamante, displayed this pattern,
suggesting that the critical difference between SD and DN cultivars is not one of CO expression.
If so, then flowering under LD conditions in Diamante must be the result of elements further
down the pathway than FrCO. Expression of FrCO in 'Diamante' was slightly different than the
other cultivars, with a somewhat flatter, wider peak. In light of the evidence that one of the two
FraCO alleles in 'Strawberry Festival' is being expressed at greatly reduced levels, while
'Diamante' expresses both alleles robustly, this broad peak may be the result of two or more
alleles expressed somewhat out of phase with each other, with the overlapping curves giving the
impression of a wider peak.
Some caution is warranted in the interpretation of these results, as this data reflects only
mRNA transcript levels. Analysis of transcript levels does not take into consideration the
mechanisms that regulate post-translational regulatory mechanisms. Elegant studies by Valverde
et al. (2004) demonstrate that light-quality driven alterations in CO stability and localization are
critical components of CO action.
The expression pattern seen in strawberry does not appear to have been previously
described, and may represent a fundamental change in the perceived function of CO. In
Arabidopsis and rice, according to the external coincidence model, flowers are initiated or
inhibited in response to whether the peak of CO occurs in daylight or darkness. Without knowing
the dynamics of protein level, it is difficult to say for certain what the mechanism is here, but it
appears that the critical factor is not one of when the peak in CO occurs, but how high it is.
Whether this trend seen at the transcript level is continued at the protein level, where further
regulation may occur (Valverde et al., 2004), remains to be seen. If it is, then some factor other
than the integration of photoperiod and circadian cycle provided by CO is likely responsible for
the perception of photoperiod in strawberry, at least with respect to flowering.
CO transcription is regulated by a number of factors, and a switch between SD and LD
flowering in eudicots may be a result of a change in one of these elements, rather than a change
in the functioning of CO itself. The evidence available does not seem to clearly indicate a
mechanism for the reduced FrCO expression seen under LD, but several mechanisms that either
reduce CO transcription or encourage degradation of the transcript have been identified. A
probable blue-light photoreceptor, FKF l, appears to regulate CO transcript level by repressing
CDF l, a repressor of CO transcription (Imaizumi et al., 2003). Other possibilities include the
Group III COL proteins, as both AtCOL9 (Cheng and Wang, 2005) and AtCOL10 (Cheng and
Wang, 2006) have been shown to decrease CO transcript level in Arabidopsis overexpressing
The new evidence also calls into question the role of phytochrome in the regulation of
flowering, and the degree to which it is maintained across species. In Arabidopsis and other LD
plants, far red-enriched extensions of day length at the end of the day accelerate flowering,
whereas red light extensions delay flowering (Kadman-Zahavi and Ephrat, 1974). This has been
explained by Valverde et al. (2004), who found that phytochromes mediated the stability of the
CO protein, with phyB, activated under red light, encouraging the degradation of the protein in
the morning, and phyA, activated under far-red, stabilizing the protein. The phyB receptor also
regulates flowering through the upregulation of PFT1, a repressor of FT (Halliday et al., 2003) as
well as possibly promoting the expression of LFY (Blazquez and Weigel, 1 999).
While this model explains the flowering responses to light quality observed in LD plants, it
becomes clear that something different is happening in SD plants. Vince-Prue and Guttridge
(1973) found that just the opposite occurred in strawberry, with far red extensions inhibiting
flowering if applied at the beginning of the day, and red light extensions inhibiting if applied at
the end of the day. Kadman-Zahavi and Ephrat (1974) grew plants under blue light supplemented
with far red, and found flowering delayed compared to blue light alone or shaded conditions.
While this might seem consistent if FrCO was a repressor of flowering, the expression data seen
in our study would seem to suggest that it is not. In fact, rice, in which CO actually has been
determined to be a repressor of flowering, does not display this pattern, but instead reacts in the
same manner as LD plants (Kadman-Zahavi et al, 1976). Because most of this regulation is
occurring at the level of the protein, it is difficult to gauge what is occurring in strawberry based
solely on our transcript level data.
Kadman-Zahavi and Ephrat (1974) described two clear groups of SD plants, based on this
response. With strawberry, they place Chil pailrl l~lnhean morifolium Ramat. and P. nil, while in the
group of SD plants that respond like LD plants to red and far-red light they place M~imosapudica
L, Amaranthus sp., and Cosmos bipinnatus Cay.. Descriptions from earlier work also suggest
that Salvia occidentalis Sw. is also similar to strawberry (Meij er, 1959), whereas corn (Zea mays
L,), rice, and sorghum (Sorghum vulgare L.) belong to the other group (Kadman-Zahavi et al,
1976; Lane, 1962). Roses, close relatives of strawberries, have been shown to have an increase in
flowering in response to red light extensions at the end of the day (Maas and Bakx, 1995, 1998),
so it appears both groups may exist even within the Roisoideae. If so, this may imply a relatively
simple difference in the mechanisms controlling such responses.
Expression of Other COL Genes
Despite belonging to a common subgroup, FraCOL1 and FraCOL2 displayed rather
different expression profiles. FraCOL2 appears to be expressed in a pattern similar to FraCO, a
pattern previously shown to be common to other COL group Ic genes in Arabidopsis. Microarray
data by Smith et al. (2004) show that AtCOL3 and AtCOL4 follow similar diurnal patterns, with
peaks around dawn under long days. Most other COL genes that have been investigated,
including AtCOL1, AtCOL2, and AtCOL9 (Ledger et al. 2001, Cheng & Wang, 2005) also show
The variation in splicing efficiency of FraCOL2 does not seem to follow a clear pattern,
however it may serve as a means of negative regulation of its function, as the introns in E.
nubicola FDP60 1, E. vesca FDP815, and E. vesca Hawaii 4 all contain multiple stop codons that
would result in a truncated protein. Such a truncated protein would lack the M4 and CCT
domains. Little is known about the functions of these elements. The CCT domain has been
proposed to be involved in the localization of the protein; however the co mutants co-5 and co-7,
with mutations in the CCT domain, localize correctly but do not function properly (Robson et al.,
200 1). Given that the function of FrCOL2 is unknown, it is difficult to speculate on the role of
this truncated protein, if any, but it seems conceivable that it might compete with the functional
version of the protein in its interactions with either DNA or protein.
Splice variants of the CO ortholog PnCO of Japanese morning glory were documented by
Liu et al. (2001), who noted both unspliced and alternatively spliced transcripts in addition to the
correctly spliced version. A maj ority of transcripts contained the full intron, whereas smaller
numbers were properly spliced or retained a small 26 bp segment of intron sequence. The partial
retention of the intron resulted in a frameshift and premature stop codon, but the presence of the
entire intron caused the coding sequence to remain in the correct reading frame. Despite this,
only the properly spliced transcript was able to complement Arabidopsis co mutation in
transgenic plants. The flowering-related gene Proliferating Inflorescence M~eristem (PIM) from
pea, a homolog of the Arabidopsis meristem identity gene AP1, has also been shown to
commonly exist as an unspliced transcript (Taylor et al., 2002).
Transcripts with unspliced introns are moderately common in plants (Ner-Gaon et al.,
2004). Alexandrov et al. (2006) found that about 7% of transcripts demonstrated some sort of
alternative splicing, of which 27% (Alexandrov et al., 2006) to 3 8% (lida et al., 2004) contain
one or more unspliced introns. Ner-Gaon et al. (2004) found that transcripts associated with
physiological flux were more prone to intron retention than those involved in functions such as
metabolism and housekeeping. lida et al. (2004) found that retained introns were more common
in plants that have received a recent stress, such as cold, heat, ultraviolet light, dehydration, and
various hormone treatments. By producing non-functional truncated products or, alternatively,
products that are degraded by nonsense-mediated decay (NMD) caused by frameshifts (Nar-
Gaon et al., 2004), an extra level of regulation in genes that are critically sensitive to changes in
transcript level may be provided.
Unlike FraCO and FraCOL2, FraCOL1 transcript appears to remain at relatively
constant levels through the course of the day. It is the only FraCOL gene expressed in all
cultivars at the 5 pm time point under short days, the apparent low point of the FraCO cycle.
In contrast, the closest Arabidopsis homologs, AtCOL3 and AtCOL4 are both expressed
in a diurnal fashion (Smith et al., 2002), although Jeong et al. (1999) did not investigate diurnal
expression of the apple homologs of the FraCOL1 gene that they characterized, nor is it clear at
what point or points in the diurnal cycle tissue was collected. While both genes were expressed
at the same relative level in all tissues, M~dCOL2 was consistently expressed at a higher levels
than M~dCOL1. And while most tissues showed a uniform but low level of expression, higher
expression of both genes was seen in developing fruit and flower tissue, which may partly
explain the great number identified in apple and peach EST libraries, because many of these
were derived from fruit tissue. In strawberry, however, the maj ority of ESTs are from libraries
derived from developing seedlings not yet producing flowers or fruit, so bias in tissue of origin
cannot entirely explain the preponderance of this group among Rosaceae ESTs in GenBank.
Datta et al. (2006) oberseved decreases in the lateral branching of inflorescences in
Atcol3 mutants, and found evidence that AtCOL3 promotes the formation of branches and
inhibits the growth of the primary shoot during short days. If FraCOL1 has a similar function in
strawberry, it may be that selection for increased numbers of inflorescences or proliferation of
branch crowns has selected for a mutation that results in continual expression of this gene, rather
than diurnal fluctuation.
Similarly, it is also possible that the elevated expression levels for FraCOL1 and the two
Malus genes seen in fruit and flower tissue derives from a role within infloresence development,
and that selection for increased flower number might have encouraged selection for relatively
continuous expression. It would be interesting to examine the expression of this gene in E.
daltoniana, a strawberry species characterized by single flowers (Sargent et al. 2004b). Datta et
al. (2006) demonstrated the necessity of a pair of amino-acids, VP, near the C-terminal, for
binding with COP1 and conferring normal red-light phenotype. FrCO, FrCOL1, and FrCOL2 all
possess this motif, but it is lacking in FrCOL3 (the C-terminus ofFrCOL4 is currently
uncharacterized, although group III genes in rice and Arabidopsis lack this pair), suggesting
possible roles for these first three, at least, in photomorphogenic development.
FraCOL3 was not expressed at perceptible levels in any of the RT-PCR experiments,
though we were able to amplify a weak band from Strawberry Festival' flower cDNA.
FrvCOL3 occurs several times among Fragaria EST sequences, and it is worth noting that all of
these are from developing seedlings, although there are only a relatively small number of ESTs
available from mature plants. In Malus and Prunus, FrCOL3-type transcripts occur several times
in mature and seedling material. Two Arabidopsis orthologs, AtCOL6 and AtCOL7, coincide
with QTLs for rosette leaves at flowering, days to budding, and days to flowering, suggesting
possible roles in flowering, although they share this interval of the chromosome with other
flowering and development related genes such as MADS-Box transcription factors and genes
involved in meristem development and hormone synthesis (Bandaranayake et al., 2004).
Materials and Methods
Six octoploid cultivars of xanana~ssa were used during the course of the work. These
cultivars and the sources from which they were obtained are given in Table 3-5. All were
obtained as rooted plants but were propagated as needed throughout the course of the
experiments. Only the second group of 'Diamante' was used directly from the nursery; all others
were multiple by runners in the growth chamber prior to experiments. E. vesca Hawaii-4 was
obtained as runner plants from the collection of Dr. Thomas Davis at the University of New
Hampshire and propagated by runners in the laboratory for use in the experiments.
Plant Growth Conditions
Plants were maintained on shelves in the laboratory at room temperature (230C), under
cool white fluorescent lighting when not in use in diurnal experiments. The first diurnal
experiment, utilizing Strawberry Festival' and 'Diamante', was conducted in the growth
chamber in enclosed compartments, under either 8 h or 16 h photoperiods under cool white
fluorescent lighting. Each treatment consisted of three 25 cm pots per genotype, containing four
mature plants in ProMix BX soilless medium, and the pots randomized within the compartment.
Temperature was maintained at approximately 230C.
The second diurnal experiment was conducted in growth chambers, with six plants in each
treatment of the octoploid cultivars Earliglow and Diamante, and two each of Camarosa. Ten
plants of E. vesca Hawaii-4 were also grown in each chamber. Plants were grown in 10O-cm
square pots in ProMix BX soilless medium and watered and fertilized as needed. Once again, 8 h
and 16 h photoperiods were used, with a photon flux of approximately 300 Clmol~s *~m-2 at the
level of the soil surface. Photon flux was found to be fairly uniform through the chamber. Three
flats containing two of each E.x a72nanassa cultivar were arranged in the center of the chamber,
and the arrangements of the pots within each flat randomized. E. vesca plants were placed
together in a fourth flat. Temperature was set at 180C day / 160C night. Plants in both
experiments were watered and fertilized as needed. Total number of runners and inflorescences
were counted on each plant at 4, 6, and 8 weeks from the beginning of the light treatments,
except for the E. vesca, for which data was collected only at 6 and 8 weeks.
Cloning of FraCO
A partial EST sequence (CO3 80854) resembling CO was identified among sequences
derived from the octoploid 'Queen Elisa'. A single primer based on this sequence, FaCOtop (5'-
TGGATGTTGGAGTTGTACCAG-3 ') was used along with the M13 reverse primer to amplify a
product from a mixed tissue 'Strawberry Festival' cDNA library by PCR. The product generated
was cloned and sequenced, and a second primer, FaCO-R (5'-
CGGCATTGTTCCTTCATACTAA-3 ) was developed. FaCOtop and FaCO-R were used to
amplify a 344 bp product for use as a probe, using Touchdown PCR as described in Sargent et al.
(2004). This probe was hybridized against approximately 20,000 colonies of a F. x anana~ssa
' Strawberry Festival' flower tissue library in E. coli /Gateway vector.
Identification of COL Genes in Genebank
One Fragaria COL gene, FraCOL1, had previously been identified in the description of
an earlier EST library (Folta et al., 2005). All others were identified by TBLAST-N searches
(Altschul et al., 1997) of Fragaria ESTs in GeneBank, using each Arabidopsis COL gene as the
query, as well as searches using only AtCO CCT domain and B-Box region. Identified Fragaria
sequences were also searched against Malus, Prunus, and Rosa ESTs to identify other Rosaceae
Gene Structure Characterization
Primers flanking the full coding region for each of the FraCOL genes were designed, as
well as a primer set (RN) enclosing a roughly 400 bp segment of the coding region believed to
include the putative intron site (Table 3-2). This segment of each gene was amplified from
genomic sequence using Touchdown PCR (as Sargent et al. 2004, except extension time was
increased to 1 min) from 'Strawberry Festival', FDP601, and FDP815, and from 'Strawberry
Festival' flower cDNA. The presence ofintron sequence was verified by comparing the fragment
size generated from genomic sequence to that generated from cDNA and the size predicted from
the EST or ESTs that it was identified from. Combinations of the flanking and RN primers were
used to confirm the absence of other introns.
Southern Blot Analysis
Total DNA was extracted from of E xanana~ssa 'Strawberry Festival' and E. vesca
Hawaii-4 using a modified cold CTAB method (Tombolato et al., in preparation). Ten Clg of each
were digested with four restriction enzymes: EcoRV, HindlII, BamnHI, and EcoRI. Digested
samples were separated by size using gel electrophoresis on 0.8% agarose gels, blotted onto
nylon membranes, and UV cross-linked. A FraCO probe was developed using the RN-FrCO-F
and FrCO-full-R (Table 3-2) with 'Strawberry Festival' genomic DNA in a touchdown PCR
reaction as described above. The probe was labeled by random priming as per manufacturer' s
instructions, hybridized overnight at 600C, and given three washes of 20 minutes each with with
a wash buffer comprised of lx SSC and 0.1% SDS.
Genetic Linkage Mapping COL genes in Fragaria
FrCO was mapped in a diploid reference population developed by Sargent et al. (2004,
2006). This is a F2 population from the cross of E. vesca FDP8 15 x E. nubicola FDP601i, and
consists of 94 individuals. The RN-FrCO primer set (Table 3-2) was used to amplify a product
by PCR from each individuals DNA (30 cycles, 520C annealing, 1 min extension). Ten Cll of
each PCR product was digested in a reaction with 0.3 Cll of HindlII enzyme, 1.3 Cll 10x bovine
serum albumin, 1.3 Cll 10x Promega Buffer B, and 0.3 Cll dH20. Digested product was separated
by gel electrophoresis on 1% agarose gels, stained with ethidium bromide, and photographed
under UV light. FDP601 alleles displayed three bands when digested, whereas FDP815
displayed only two. Size differences allowed the trait to be scored as a co-dominant marker.
Attempts to map the FrCOL2 locus were performed as above, with the following
exceptions: the RN-FrCOL2 primer pair (Table 3-2) was used, extension time was set at 45 s per
cycle, and PCR products were digested with Af71II enzyme. When this showed an unusual
segregation pattern, the process was repeated with the EcoRV enzyme. In both cases, the
FDP815 allele was not cut by the enzyme, whereas the FDP601 allele was. However, due to
incomplete digestion, it was scored as a dominant marker.
Marker data was entered into the JoinMap 3.0 software program and mapped relative to
marker data provided by Dan Sargent, consisting of most of the markers appearing in Sargent et
al., 2004 and Sargent et al., 2006.
Extraction of Nucleic Acids
Except as noted above for use with the Southern blot, DNA was extracted using the
Qiagen DNEasy Plant DNA extraction kit, according to the manufacturer's instructions. Some
genomic DNA was provided by others, namely that of E vesca Hawaii-4, E. iinumae FRA377,
E. bucharica FRA520, and E. mandshurica FME, from Dr. Thomas Davis at the University of
New Hampshire, and E. vesca FDP815 and E. nubicola FDP60 1 and that of the F2 mapping
population from these parents from Dr. Daniel Sargent of East Malling Research.
RNA was extracted using a modification of the pine cone method of Chang et al. (1993) as
described in Folta et al. (2005). Briefly, 1 g of tissue was frozen with liquid nitrogen and ground
with a mortar and pestle, then incubated at 650C for 10 min in a CTAB-based extraction buffer
(2% CTAB, 2% polyvinylpyrrolidone, 100 mM Tris-HCI (pH 8.0), 25 mM EDTA, 2.0 M NaC1,
0.5 g/ml spermidine, and 2.0% beta-mercaptoethanol). Samples were then allowed to cool to
room temperature and equal volumes of chloroform:0ctanol were added and the samples
homogenized using a Polytron T10-3 5 tissue homogenizer at 90% of full speed. The organic and
aqueous phases were separated via centrifugation at 8,000 x g. The supernatant was then
removed, mixed with equal volumes chloroform:0ctanol, and then separated via centrifuge again.
LiCl was added to the resulting supernatant to a concentration of 2.5 M and allowed to
precipitate overnight at 40C, after which it was centrifuged again at 10,000 x g. The pellet was
resuspended in 500 Cll SSTE (1 M NaC1, 0.5% SDS, 10 mM Tris-HCI (pH 8.0), 1 mM EDTA)
and again purified with chloroform:0ctanol, then precipitated with two volumes of 100% ethanol.
The pellet was then washed with 76% ethanol, 0.3M sodium acetate, dried using a SpeedVac,
and resuspended in 50C1l of 10 mM Tris-HCL (ph 8.0), 2.5 mM EDTA. Samples were quantified
Reverse Transcription and RT- PCR
For the first photoperiod experiment 1 Clg of RNA from each condition was reverse
transcribed using AMV reverse transcriptase (Promega Inc., Madison, WI) as per the
manufacturer' s protocol. The resulting cDNA was diluted 1:10 with TE buffer and used as
template for RT-PCR reactions.
Initial experiments were conducted with the 'Camarosa' SD cDNA set to determine the
likely linear range for each pair of primers. For RN-FrCO, RN-Ubiq, and RN-Actin primers sets,
reactions were conducted at 24, 28, 32, and 36 cycles, and the highest number of cycles which
retained the relative band intensities seen at lower numbers of cycles was selected. RT-PCR
reactions used the RN primer sets described in Table 3-2 using standard 3 step PCR (520C
annealing temperature, 30 seconds extension, for 32 cycles for RN-FrCO and RN-Actin, 28
cycles for RN-Ubiq), separated by electrophoresis on 1% agarose gel, then stained with ethidium
bromide and photographed under UV. Template amounts were adjusted using ubiquitin and actin
RN primer sets (Table 3-2) to assure similar amounts of template DNA between cultivars and
treatments and the PCR repeated and gels photographed (Figure 3-13). Allele-specific RT-PCR
trials used the same PCR conditions, but 33 cycles was used because slightly less product was
anticipated than for with the RN-FrCO primer pair. A second round was conducted at 40 cycles,
with the idea of reach saturation and testing the whether any FrCO-C9 product was present in
' Strawberry Festival'.
The second experiment used a similar procedure, but reverse transcription was done using
ImProm II reverse transcriptase (Promega Inc., Madison, WI) as per manufacturer' s instructions.
The resulting 20 Cll reaction was not diluted but used directly as template in a PCR reaction
identical to above, except that 35 cycles was used for FrCOL2. Template volume used in each
reaction was adjusted using ubiquitin and actin RN primers to equalize template amounts. The
products were again separated by electrophoresis (on a 1.5% agarose gel), stained and
Table 3-1. Comparison of amino acid identity (%) of the predicted protein encoded by FraCO to
those of Group la CONSTANS-LIKE genes in apple (Mahts dontestica), castor bean
(Ricinus conanunis), Cottonwood (Popubts deltoides), thale cress (Arabidopsis
thaliana) rape (Bra~ssica napus), Japanese morning glory (Pharbitis nil), rice (Oryza
sativa), wheat (Triticunt aestivunt), and perennial ryegrass (Lolium perenne).
Acc. No. B-Box 1 B-Box 2 M1 M2 M3 M4 CCT Overall
z Aalus xdomestica CO gene assembled from ESTs EBl130204, EBl141442, and DR996733
SR'Rcinus conununis CO gene assembled from ESTs EG683156 and EG686997
Table 3-2. RT-PCR primers for strawberry CONSTANS-like genes and controls, sequence
source, Thl, and approximate size in cDNA (as calculated from sequence).
z Sequence obtained in the course of this study
Y Approximate size in transcripts containing unspliced intron
x 5' portion of transcript sequenced from library using universal primers
Table 3-3. GenBank accession numbers for Fragaria COL genes and other Rosaceae orthologs.
COL gene orthologs
and 66 ESTs
and 20 other
Figure 3-1. Cladogram of CO-like genes, including all Arabidopsis and full-length
Rosaceae COL genes as well as all those from other species Group la genes
demonstrated to functionally complement Arabidopsis co mutants
(underlined). NJ tree was constructed based on the full-length predicted amino
acid sequence of each gene. Because the full sequence was not available,
FvCOL4 is not included in the analysis, but appears most similar to AtCOL9
and AtCOL10. Strawberry genes are shown in bold. Groups follow the criteria
of Griffiths et al., 2003.
AtCO MLKQE-------------SNDIGSGENNR-ARPOlDTCRSNATYHDALMCA 46
BnCOal MFKQE-------------SNNIGSEENNTGPRAOCDTCGSITYHDALNCQ 47
PnCO MLKEE SCEVLDLDVTIGSSSGSRSGNKQNWARVCDICRSAAC SVYCRADLAYLCGGCDAR 60
FaCO MLKEE-------------SNGAAAAN--SWARVCDTCRSAPCVCADALSGCDAT 45
Consensus M:K:E S.. : : .R CD C S C:VYC:AD AYLC .CDAT
AtCO VHSANRVASRHKRVRBCESCERAPAAFLCEADDASLCTACDEHAPRHQVL 106
BnCOal VHSANRVASRHKRVRBCESCERAPAAFMCEADDVSLCTACDEHAPRHQVV 107
PnCO VHGANTVAGRHERVLT BCEACESAPATVICKADAASLCAACDDHAPARHV 120
FaCO IHAANRVASRHERVWFCEACERAPAALLCKADAASLCTACDIHAPRHQVL 105
Consensus :H.AN VA.RH:RV BCE:CE APA:.:C:AD .SLC:ACD ::HSANPLARRH:RVP :
AtCO PI SG--------NSFSSMTTTHHQSEKTMTDPEKRLVVDQEEGDAEEVSLP 157
BnCOal PITG--------NSCSSLATANHT---TVTEPEKRVVLVQE---Dg TSLP 149
PnCO PI SGTLYGPPTSNPCRESSMMVGLTGDAAEEDNGFLTQDAEETMEDEASLN 179
FaCO PI SG-------------GQIVVGSTPADTTED-GFLSQEGDEEMEDEASLN 151
Consensus PI:G : : : D OT
AtCO ---NSD- KNNNNQNNG----------------LLFEDEYLNVYSM KF E -- 194
BnCOal ---NSDNHNNNNQNNE----------------LLFQDDYLDAYSMKFQNP 190
PnCO PNPNPHPNPVKSNNSTNMCKGGNNNNNEMSCAVEAI~DAYLAESCNFEKY- 237
Consensus N.I : :.N.:D YL:L.::.S. : F
AtCO SQHQQNCsPQ-S gcRVPQT4----S'GGDVPKE--SRHCNQNQNKGSGHN 247
BnCOal TQHKQDCTVP EK-- NJGGDRVVP LQL~EE-----TRGNLHHKH--IYSGHN 239
PnCO INQQQNYSVPQRNMS'TRGDSIVP-NIGKNQFHYTQGLQQNH-AFCEWMIT 294
FaCO TNDQHSYGVPHK-I SfGGDSVVPVQTGEGKVTQMQMQQKHNFH--QLGMEYESSKAG 259
Consensus :..VP .' GD :VP:: : : H:
AtCO NG SINHNAY I S SMTGVVM ESTADIVTTASHPRTPKGTVEQQgPDPASQM ITVTQLPE 307
BnCOal NGSINHNAYNPSTDVEA 'DTSHPKTHKGKIEKLPEPLIQIL-----SMR 294
PnCO D-----MVSI SSMOVGVVPE STLBDTSI SHSRASKGTIDLFSGPPIQMPPQLQLQMR 349
FaCO DGSI SHTVSVSSMOVGVVPDSTMfiEMSVCHPRTPKGTIDLFNGLTQP-LSMR 317
Consensus :: .SM:...VP:.* : H: G: Q: :SMR
AtCO AIRVLRYREKRKTRKFEKTRYASRAYAEiRPRVNGiRFAEEEAEGNTLY- 365
BnCOal ARVLRYREKKKRRKFEKTIRYASRKAYAERRPRINGRFAKIEEEQYTLYD 354
PnCO ARVLRYREKKKTRKFEKTIRYASRKAYAETRPRIKGRFAKEDDEDIYPME 409
FaCO ARVLRYREKKKTRKFEKTIRYASRKAYAEARPRIKGRFAKEDDEDMSSME 377
Consensus ARVLRYREK:K RKFEKTIRYASRKAYAE RPR::GRFAKE : :.V :
AtCO GYGIVPSF---- 373
BnCOal GYGIVPSFYGQK 366
PnCO GYGIVPSF---- 417
FaCO GYGIVPSY---- 385
Figure 3-2. Alignment of the predicted amino acid sequence for FraCO with those of
AtCO and functionally confirmed dicot homologs, with major conserved domains noted.
HSANPLSRRHERVPITPFYDAVGPAKSAS----SSVNFVDE---DG--- V 121
HSANPL: .:RVP: PFYD::... : FV:::: DG
ASWLLP-NPPAM------------------ENPDLNSGQ- YFEDLLYGVP 176
ASWLLA--- ----------------------KEGIEITN--F---DY--P 144
ASWLL ::: :F DLDY
LEDAQEQNsSCTDGVVP EQ KNMQPQLVNDHSFEIDFSAA SPVGHACRVS 236
LEEAQEQNSCHADGVVPV Q KNMQP LLVNDQSFELDFSAG SPVGHACQVS 236
KTEAQEQNSS!.TDGVVPV Q KSAQP-----QSFEMELPG- SPY---LQVS 208
KVESLEQNSSMTDGVVPgJTVP VERVITNENCFEMDFTG GSGTGGNCSSS 295
IEVTSEENSSONDGVVPVQ:4KLFLN----EDYFNFDLSA- SIQGN-FQTT 196
:E:NS.I DGVVP : : F::::.. SK : ::VS:
SDISVVPDGtAVT-----------AAVET-SQPAVQLSS VOVRVRRKRNK 284
SSHGVPGSVADVS YP YGG PAT SGADPG TQRAVPL T SAREARVMRYRE KRKNRKF 35 5
.::: VP:..: :V ::..:R ARV:RYREKRKNRKF
EKTIRYASRKAYAE RPRIKGRFAK1.T: :G:VP:F
.Alignment of the predicted amino acid sequence for FrvCOL 1 with possible
apple and Arabidopsis homologs, with major conserved domains noted.
B-Box 1 B-Box 2
FCRADSAFLCVNC DSKI HAANKLASRHARVWIC EVC EQAPAHVTC KADDAALCVTC DRD I
FCRADSAFLCVNC DSKI HAANKLASRHPRVWJICEVC EQAPAHVTC KADDAALCVTC DRD I
FC RADSAFLC INC DTK IHAANKLASRHARVWI4CEVC EQAPAHVTC KADDATLCVTC DRE I
FC RADAAFLCG DC DGK IHTANKLASRHE RVWI~CEVC EQAPAHVTC KADAAALCVTC DRD I
:CR.D:AFLC .CD K:H:ANKLASRH RVW:CEVCEQAPAHVTCKAD A:LCVTCDR:I
MadCOL1 ------------------------------------------MLCDKSTL 16
MadCOL2 ------------------------------------------MSCDQATL 16
FaCOL1 -------------------------------------------AK3DCSTT 16
AtCOL4 MDP TWI DSL TRSC EANSNTNHKRKRE RE TLKHRE KKKKRFRE RKMASKL CDSC KSATAAL 60O
AtCOL3 -----------------------------------------MASR3DCSAT 18
791 bp 746 bp 362 bp
79 bp 70 bp 36 bp
796 bp 704 bp 362 bp
784 bp 703 bp 362 bp
688 bp 119 bp 383 bp
82 bp 310 bp
118 bp 383 bp
Figure 3-4. Structures of COL alleles from Fragaria species: xanana~ssa Strawberry
Festival' (Fra), E. vesca FDP815 (Fry), E. nubicola FDP 601 (Frn), and F.
iinumae FRA377 (Fri). Thinner black lines indicate introns, dark grey
indicates conserved domains, and crosshatching indicates maj or indel
82 bp 322 bp
F. x ananassa A (35)
F. x ananassa B (C9)
'i I F. iinumae
0.0 0.01 0.02 0.03 0.04 0.05
Figure 3-5. NJ phylogram tree of intron sequence divergence among FrCO alleles.
F, nubicola FDP601
F. xananassa Festival-C9
---. x ananassa F~estival-35
F. vesca Hawani-4
F, vesca FDP81~5
F. bucharica FDP520
F, iinumae FDP377
0.0 0.1 0.2 0.3 0.4
F. vesca Hawaii-4
F vesca FDP815
F. nubicola FDP601
F x ananassa 'Festival' C9
F x ananassa 'Festival' 35
F. bucluvica FDP520
F. iinumae FRA377
F. vesca Hawaii-4
F vesca FDP815
F. nubicola FDP601
F x ananassa 'Festival' C9
F x ananassa 'Festival' 35
F. bucluvica FDP520
F. iinumae FRA377
ATG TGAAC GGATAACAAT TTTCACACAGGAAACAGC TAT TG -- --AC CCATGAT -TAC
AAC AAC GGATTC TTC TTC GGAGTGGAGGT TGATGAGTAC TTGGAC CTTG TGGAG TACAAC
AAC AAC GGATTC TTC TTT GGAGTGGAGG TTGATGAGTAC TTGGACTCTTG TGGAG TACAAC
Figure 3-6. Comparisons of hypervariable section of coding region in FrCO. (A) Alignment of
variable region of seven FrCO alleles, (B) NJ phylogram comparing nucleotide
sequence within the variable region.
1 2 3 4 5 6
Figure 3-7. Southern blot of genomic DNA from E. xanana~ssa 'Strawberry Festival' and
E. vesca Hawaii 4, cut with six different enzymes and probed with a portion of
FrvCO. Lane 1=XholI, 2=XbalI, 3=HindlII, 4=EcoR V, 5=EcoR I, 6=BamnHI.
136 bp 403 bp 199 b:
nd Ill EcoR I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
aa bb ab ab ab bb ab bb aa ab bb bb ab ab h
nd Ill Hind Ill
72.5-U UFFxa01 EO3
~~ ~J 1,1)
Figure 3-8. (A) Map of the portion of the FrCO product amplified for use in mapping, showing the
locations of distinguishing restriction sites and lengths of the resulting fragments. (B) Gel
showing characteristic band patterns after digestion with Hind III for the F. vesca parent
(lane 1, genotype aa), the F. nubicola parent (lane 2, genotype bb), twelve members of the
F2 (lanOS 3-14), and the 564 bp band of the hHind III size standard (lane 15). (C) Partial
map of linkage group VI (after Sargent et al. 2005), showing the predicted location of
Table 3-4. Mean inflorescence and runner production ofE. xanana~ssa and E. vesca genotypes
under LD (16 h) and SD (8 h) photoperiods, at 18/160C day/night temperature.
4 weeks 6 weeks 8 weeks
Genotype Inflorescences Runners Inflorescences Runners Inflorescences Runners
Diamante 1.2 0.0 3.2 0.0 3.8 0.2
Camarosa 0.0 0.0 0.0 0.0 0.0 0.0
Earliglow 0.0 2.2 0.0 3.0 0.0 3.8
Hawaii 4 --0.1 2.8 0.2 3.8
Man t~l Day
Diamante 0.4 0.0 0.6 0.0 1.0 0.0
Camarosa 0.0 0.0 0.3 0.0 1.0 0.0
Earliglow 0.0 2.4 0.0 2.8 0.0 3.0
Hawaii 4 --3.0 1.8 3.3 1.8
Short Day (8 h)
Long Day (16 h)
Figure 3-9. Northern blot of RNA from Strawberry Festival', a SD genotype, and 'Diamante', a
DN genotype, collected every 4 h, showing expression of FraCO under short and
long day conditions.
Long Day (16 h L)
~ ~I~ ~
Figure 3-10. RT-PCR assay of expression ofFrCO transcript under short (8 h) and long (16 h)
days in three genotypes: E. vesca Hawaii-4 (SD, diploid), xanana~ssa 'Camarosa'
(SD, octoploid), and xanana~ssa 'Diamante' (DN, octoploid). Dawn occurred at 8
a.m. in both chambers, and dusk was at 4 p.m. in the short day chamber and at
midnight in the long day chamber.
Fra CO- C9
Figure 3-11. Expression of 35 and C9 alleles ofFraCO at dawn under SD, using RT-PCR at 33
and 40 cycles, in 'Strawberry Festival' (F) and 'Diamante' (D).
Table 3-5. Sources of plants used in this chapter.
Garden Gate Nursery
University of N.H.
Figure 3-12. RT-PCR assay of FrCOL2 expression in F. vesca Hawaii-4, showing variation in
splicing efficiency at several points under short days.
Short Day (8 h L)
Short Day (8 h L)
as~~ al memn
Long Day (16 h L)
Inr rrII (rr
Hawaii-4 (SD) agg ga
11pm 3am 7am 11am 3pm 7pm 11pm 3am 7am 11am 3pm 7pm
Figure 3-13. Ubiquitin controls after equalization of RT-PCR samples using RN-Ubiql primer
set to standardize template amounts.
IDENTIFICATION AND CHARACTERIZATION OF FLOWERING-RELATED GENES INT
Fragaria AND OTHER ROSACEAE
Although CONSTANS plays a central role in the integration of photoperiod and the
circadian clock into the regulation of flowering, CO itself does not directly control the
reproductive transition. Instead, CO, along with other inputs, regulates a network of proteins
which through a cascade of interactions eventually trigger the activation or repression of a suite
of meristem identity genes that ultimately govern the developmental fate of the meristem.
In Arabidopsis, the two most important regulators of flowering directly downstream from
CO are FLOWERING LOCUS T (FT) and SUPPRESSOR OF OVEREXPRESSION OF;
CONSTANS 1 (SOC1) (Samach, 2000). These two proteins act in parallel to promote flowering,
and are affected by different environmental inputs. A third protein, TERM7INAL FLOWER 1
(TFL1) is closely related to FT, but has an opposing role. Rather than promoting a transition to
flowering in the meristem, TFL1 delays it, causing growth to remain vegetative (Kobayashi et
al., 1999). FT and TFL seem to have fairly direct impacts on the meristem, but SOC1 acts
primarily through the activation of protein called LEAFY (LFY) to promote flowering. FT
appears to be under greater photoperiodic control than SOC1 (Moon et al., 2005). Although
promoted by CO, SOC1 is also under regulation by the other pathways regulating flowering:
vernalization, autonomous, and gibberellin, and serves as the point at which these regulatory
networks integrate into the general flowering pathway (Izawa et al., 2003).
In addition to these components, a number of other genes affect flowering as well. These
include components of the circadian clock (Turner et al., 2005) and photoreceptors (Izawa et al.,
2003), which affect CO transcript and protein levels, as well as components of the other floral
Mutations in these genes cause significant changes in flowering time and sensitivity to
environmental inputs in Arabidopsis, as well as in crop plants such as rice (Izawa et al., 2003).
Although the genetic basis for variation in flowering habit in strawberry is currently unknown, it
would seem quite possible that differences in the expression of this small suite of genes lie at the
core of it. By identifying and characterizing the orthologous genes in Fragaria, it may be
possible to pinpoint the location of the mutation or mutations involved in the shift from short day
to everbearing or day-neutral flowering. This could in turn be the basis for molecular markers to
assist in breeding, a target for genetic engineering, or even shape cultural practices by clarifying
the interaction of the different physiological mechanisms. Such knowledge may have impacts
even outside of strawberry, particularly for the other important crops within the Rosaceae.
To that end, efforts were made to identify the genes comprising this network in Fragaria,
and the expression of some of these genes was examined in strawberry plants under long and
short day conditions.
Identification of Rosaceae Flowering Gene Homologs
Screening of Fragaria EST libraries developed in this laboratory (Folta et al., 2005, as
well as others as outlined in Chapter 3) revealed a number of sequences related to transcripts
known to be involved in flowering in Arabidopsis, rice, or apple (Tables 4-1 & 4-2). In addition
to Fragaria transcripts, ESTs from other Rosaceae species were also searched for MADS-box
and FT/TFL1i-like genes, in order to gain insight into the nature of these genes and gene families
within the larger taxon, and to help fill in gaps where no strawberry genes are known but may
Six strawberry MADS-Box genes were identified in ESTs, adding to the three already
documented in GenBank. Most had fairly clear correspondence with one of the fifteen
documented MADS-Box genes in apple, although not all had clear homologs in Arabidopsis.
These included genes related to the Arabidopsis transcription factors SOC1, SEPALLA TA3,
PISTILLA TA, SHORT VEGETA TIVE PHASE, SHATTERPROOFl/2, and AGAM~OUS-LIKE14,
as well as one transcript not closely related to any Arabidopsis gene but somewhat similar to the
apple MdM~ADS11. These were designated, in order, Fra;SOC1, FrvSEP3, FrvPI, FrvSVP,
FrvM~ADS1, FrvM4~rDS2, and FrvM~ADS3. These genes are listed in Table 1, and their implied
relationship to the other known Rosaceae MADS-Box genes and a selection of Arabidopsis
MADS-Box genes are shown in Figure 1.
Fra;SOC1 was first identified as a partial transcript (GenBank Accession CO817776)
resembling a MADS box transcription factor among a small EST library developed from whole
plant tissues of F. xanana~ssa Strawberry Festival' (Folta et al., 2005). Because the glycerol
stock of this clone was available, we were able to sequence the gene from the other end,
obtaining the entire sequence. The 642 bp coding region encoded a predicted MADS Box
transcription factor, with high similarity to the Arabidopsis SOC1, with 63% identity at the
amino acid level. The closest overall match annotated in the public databases was VvM4ADS8,
again from grapevine (Vitis vinifera L.), to which it shared 68% identity at the amino acid level.
Other than this strawberry gene and the previously identified relative in apple (Valazideh et al.,
2006), no apparent SOC1 homologs were found among Rosaceae ESTs. A second Fragaria
gene, FrvM4rDS3 (GenBank Accession DY675364), also grouped within the SOC clade, but
more closely resembled AGL14 than SOC1.
An EST (DY674124) bearing considerable similarity to the floral repressors TFL1 and
A TC was also identified among sequences derived from cold-stressed seedlings of the F. vesca
selection Hawaii-4. This transcript was initially passed over because BLAST-X searches showed
only partial similarity to known genes. Upon closer inspection, it became clear that the transcript
retained 147 bp of intron sequence at a site near the middle of the gene, which resulted in a
frameshift, causing multiple stop codons and corrupting more than half of the predicted protein.
This likely represents only a partial failure to complete splicing, as the apple (DQ535888) and
Arabidopsis (NM_120465) genes each contain three introns. With the exon sites identified by
comparison to these likely homologs, the spliced transcript yielded a coding region of 666 bp and
encoded a protein with 83% identity to ARABIDOPSIS THALIANA CENTRORADIALIS (A TC),
and a lower 63% identity to the Arabidopsis TFL1 (Figure 4-2), although the proportion of
positive matches, 91% and 85%, respectively, were much closer. The most similar annotated
transcript was from Populus, a "Flowering Locus T-like protein" (BAD22601), with which the
protein shared 85% identity. At the nucleotide level, the most similar annotated sequence was
again from Poplar, PnTFL1 (ABl181 184), with an e-value of e-60, and once again this transcript
was significantly more similar than the nearest Rosaceae gene, MdTFL1-1 (AB052994), with an
e-value of 9e-28.
Efforts to amplify a strawberry ortholog of FT from genomic or cDNA using degenerate
primers based on apple and Arabidopsis sequence were unsuccessful, as was colony
hybridization with Arabidopsis FT as a probe, and no likely candidates were identified in any of
the EST libraries. No Fragaria sequences were identified with significant similarity to the other
members of the Arabidopsis family: TSF, MFT (M~OTHER OF FT), and BFT (BROTHER OF
F7), however apple ESTs resembling M~FT and BFT were identified (GenBank Accessions
EBl134193 and EB13 8045, respectively), as well as a partial transcript of aM2FT-like gene from
almond (GenBank accession BU57441 1). Relationships of apple and strawberry amino acid
sequences with members of the FT TFL1 family in Arabidopsis are shown in Figure 4-2.
Other genes thought to be involved in flowering
Also identified were a number of elements of the circadian clock, including apparent
orthologs of several response and pseudo-response regulators, TOC1, FKFl, ZTL, as well as
several genes with roles in photomorphogenesis, including genes similar to HY5 and
hemeoxygenase (Table 1). Attempts to use degenerate primers based on consensus sequence
from apple, pear, and Arabidopsis LEAFY genes to amplify the strawberry ortholog from either
cDNA library or genomic DNA were not successful. However, a fosmid from F. vesca,
containing the genomic sequence of LEAFY was obtained from Dr. Thomas Davis of the
University of New Hampshire.
Surprisingly, RT-PCR assays showed FrSOC1 expression did not parallel flowering as
expected. Rather, in 'Camarosa', FrSOC1 generally followed FrCO expression under short days,
but with a lag of several hours, with a peak near the end of the day and tapering off into the
night. Under long days, however, the expression levels in all but 'Diamante' were extremely
high, and remained so through out the course of the day (Figure 4-3). In 'Diamante', the day-
neutral genotype, however, FrSOC1 expression followed similar patterns under both conditions.
FrTFL1 and FrLFYExpression
Because of the unexpected results seen with FrSOC1, RT-PCR assays of FrLFY
expression were also conducted on the short day octoploid cultivar, Camarosa. Although
expression levels were low, an increase to 36 cycles revealed expression of FrLFY (Figure 4-4)
and Fr TFL1 (Figure 4-5) to be the opposite of that anticipated by the Arabidopsis model, just as
with FrSOC1. Under short days, when 'Camarosa' was initiating flowers, neither was detectable
in leaf tissues. Under long days, when growth was primarily vegetative, however, both were
expressed at all time points, though levels were low.
Rosaceae MADS-Box Genes
MADS-box genes play critical roles in the a number of plants processes, especially in the
development of fruits and flowers (Becker and Theissen, 2003), and as such are important targets
of study in crop plants. The Arabidopsis MADS-box gene family comprises roughly a hundred
members, though only one group, the so-called MIKC class which contains 39 Arabidopsis
genes, has been functionally characterized to a significant extent (Becker and Theissen, 2003).
Becker and Theissen (2003) break this class into fourteen groups, one of which, TM8, does not
exist in Arabidopsis. The eight Fragaria MADS-box transcription factors identified in this study
along with those identified by previous researchers represent seven of these groups. An eighth
group, APl-like MADS-box factors, is represented in several other species within the family,
including peach (PrpM~ADS6), loquat (EjAPl), almond (PrdM~ADS2), and several genes in apple
(MdAP1, MdM~ADS5, M~dMADS2, M~dMADS12, and M~dMADS12a), but none are yet known in
The six groups not yet attested in any of the Rosaceae are the AGL12, AGL15, TM~8, TT16,
ANR1, and FLC groups. Of these, FLC is the best characterized, playing a critical role in the
regulation of flowering by non-photoperiodic factors, especially vernalization (Boss et al., 2004).
It is worth noting that FLC orthologs have been identified only within the Brassicaceae, and are
not functional within all ecotypes of Arabidopsis, nor have AGL15 genes (Becker and Theissen,
2003). With the exception of the ANR1 group, none of these groups were attested in the legume
species Medicago, soybean, and lotus (Hecht et al., 2005). The absence of AGL12, AGL15, and
ANR1-type genes may be explained by the absence of root-derived Rosaceae ESTs, as these
genes appear to be primarily active in root development (Becker and Theissen, 2003). The
remaining groups are mostly poorly characterized in Arabidopsis.
Among the most numerous Rosaceae MAD S-box genes are those of the SEPALLA TA
group. The SEP group in Arabidopsis consists of four members that interact with other MADS-
box proteins to shape organ identity in the four whorls of flower development (Ditta et al., 2004).
Two such genes were identified in strawberry, Frak4~rDS-RIN (Vrebalov and Giovannoni, 2002),
with greatest similarity to AtSEP1 and AtSEP2, and a F. vesca EST we designated FrvSEP3.
AtSEP1 and AtSEP2 are very similar to each other, and likely represent recent duplications, and
this may account for the seven apple genes in this group, although in the case of apple, these
duplications may also be the result of polyploidization in the evolution of this species'. The bulk
of the Rosaceae SEP genes most closely resemble the SEP1/SEP2 genes of Arabidopsis,
including FraAL4~DS-RIN, peach and almond genes, and all but one of the apple genes in the
group. FrvSEP3 and the pear gene PbM4~rDS2 (GenBank Accession AB265800) most closely
resemble SEP3, whereas M~dMADS4 does not bear clear relation to a specific Arabidopsis
member, though it most closely resembles SEP4. In Arabidopsis, the SEPALLA TA genes appear
to be largely redundant in function, and plants lacking all three produce flowers composed solely
of sepals (Pelaz et al., 2000). Sung et al. (2000) investigated the expression of two such genes
from apple, M~dnif4DS3 and M~dnB4rDS4, and found very different expression patterns, with the
first expressed only in the floral primordium, whereas the second was expressed throughout
floral and early fluit development, particularly in vascular bundles. Yao et al. (1999) found seven
MADS box genes were expressed in distinct patterns during flower and fruit development. It
seems likely that Fragaria genes within this family also play roles in reproductive development,
although the developmental differences between Arabidopsis, apple, and strawberry fruit require
some caution in predicting function based on these other species.
One new strawberry MADS-box gene, designated FrvM~ADS3, shares some similarities
with the SEP subfamily, but appears most closely related to the AGL6 group, which it shares
with the apple gene MdM~ADS11. The function of these genes is poorly understood. In
Arabidopsis, their expression seems to be confined to floral organs (Becker and Theissen, 2003)
There is evidence that OM4rDS1, an AGL6-type gene from orchid, interacts with OM~ADS3, a
factor involved in flower formation, and promotes flowering when expressed ectopically through
the upregulation of SOC1 and FT in Arabidopsis (Hsu et al., 2003). Similar results were achieved
by Tian et al. (2005) in their study of another AGL6-like gene, DIM~ADS18, from bamboo
(Dendrocalamnus latiflorus Munro).
The AP3 and PISTILLATA class genes are class B organ identity genes, and combine with
class A and C genes to specify petals and stamens, respectively (Krizek and Meyerowitz, 1996).
Both have been previously identified in apple (Yao et al., 2001) and rose (Kitahara et al., 2001).
The three apple mutants--'Rae Ime', 'Spencer Seedless', and 'Wellington Bloomless'--which
produce apetalous flowers and parthenocarpic fluit, were found by Yao et al. (2001) to contain
retrotransposon insertions in introns of2~dPl, abolishing expression and suggesting a critical role
for the gene in both fluit and flower development. The strawberry PI gene identified here might
thus be an interesting target for RNAi- or antisense-based suppression of expression.
Parthenocarpy might be a means to limit malformation of finit, which can result from incomplete
pollination, and might also increase fluit size.
The gene designated FrvSVP most closely resembles the SVP group of MADS-Box genes.
This group appears to play a role in the control flowering in Arabidopsis, with both SVP