Download HIGH FRUIT SUGAR CHARACTERIZATION, INHERITANCE AND

HIGH FRUIT SUGAR CHARACTERIZATION, INHERITANCE AND LINKAGE OF
MOLECULAR MARKERS IN TOMATO
By
NIKOLAOS GEORGELIS
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
2002
Copyright 2002
by
Nikolaos Georgelis
I dedicate this thesis to my parents Peter and Caterine.
ACKNOWLEDGMENTS
I want to thank my advisor, Jay W. Scott, for his patience and advice throughout my
research. I also appreciate the help that I received from the members of my committee,
Harry J. Klee and Elizabeth A. Baldwin, as well as their willingness to provide me with
space in their labs for some periods. Finally, I would like to thank Karen Pearce, Hesham
Agrama, Holly Sisson and Denise Tieman for contributing to my research.
iv
TABLE OF CONTENTS
page
ACKNOWLEDGMENTS ................................................................................................. iv
LIST OF TABLES............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
ABSTRACT....................................................................................................................... ix
CHAPTER
1
INTRODUCTION .........................................................................................................1
2 CHARACTERIZATION OF HIGH SUGARS FROM PI270248 ...............................11
Introduction................................................................................................................. 11
Material and Methods ................................................................................................. 14
Fall 2001 ................................................................................................................ 14
Spring 2002 ............................................................................................................ 15
Both Seasons .......................................................................................................... 15
Results......................................................................................................................... 16
Discussion ................................................................................................................... 18
3 RELATIONSHIP OF HIGH FRUIT SUGARS WITH OTHER TRAITS...................20
Introduction.................................................................................................................. 20
Materials and Methods................................................................................................. 24
Fall 2001 ................................................................................................................ 24
Spring 2002 ............................................................................................................ 25
Both Seasons .......................................................................................................... 25
Results.......................................................................................................................... 27
Discussion .................................................................................................................... 28
4 INHERITANCE OF HIGH SUGARS IN FRUIT DERIVED FROM PI270248 AND
SEASONAL EFFECTS ON FRUIT SUGAR ...................................................................33
Introduction.................................................................................................................. 33
Materials and Methods................................................................................................. 35
Environmental Effects among Seasons.................................................................. 35
Fall 2001 ..........................................................................................................35
v
Spring 2002......................................................................................................36
Both Seasons....................................................................................................37
Genetic and Environmental Effects within Season................................................ 37
Fall 2001 ..........................................................................................................38
Spring 2002......................................................................................................38
Results.......................................................................................................................... 38
Inheritance of High Sugars derived from PI270248 .............................................. 38
Fall 2001 ..........................................................................................................38
Spring 2002......................................................................................................39
Environmental Effects between Seasons ............................................................... 39
Discussion .................................................................................................................... 42
Inheritance of High Sugars derived from PI270248 .............................................. 42
Environmental Effects between Seasons ............................................................... 46
5 RAPD MARKERS LINKED TO HIGH SUGARS FROM ACCESSION PI270248 .49
Introduction.................................................................................................................. 49
Material and Methods .................................................................................................. 55
Fall 2001 ................................................................................................................ 55
Spring 2002 ............................................................................................................ 56
Both Seasons .......................................................................................................... 56
Results.......................................................................................................................... 59
Discussion .................................................................................................................... 62
6 PEDIGREE OF TOMATO LINES USED IN EXPERIMENTS .................................70
7 KLEE’S PROTOCOL FOR DNA EXTRACTION FROM TOMATO LEAVES.......71
LIST OF REFERENCES...................................................................................................72
BIOGRAPHICAL SKETCH .............................................................................................81
vi
LIST OF TABLES
page
Table
2-1. Total sugar means and fructose/glucose ratios of PI270248 and 7833 in fall 2001 and
spring 2002 at Bradenton, Florida..................................................................................17
3-1. Comparison of PI270248 and 7833 for physical and chemical traits in fall 2001 and
spring 2002 at Bradenton, Florida..................................................................................28
3-2. Correlation coefficients (r) of total sugars with glucose, fructose, fructose/glucose,
titratable acidity, pH, yield, earliness, and fruit size from F2 in fall 2001 and spring
2002 at Bradenton, Florida.............................................................................................29
3-3. Sugar levels for F2 plants grouped by plant habit and pedicel type ...............................29
4-1. Output of Joint Scaling Test Worksheet showing the failure of the
additive/dominance model .............................................................................................42
4-2. Estimates of the interaction parameters z .......................................................................42
4-3. The effect of Season on the sugar level of PI270248, 7833 and F1 separately where
p-value indicates the significance of the season effect for each line..............................43
4-4. Average monthly temperature (oC), rainfall (cm) and solar radiation (W/m2) in fall
and springz ......................................................................................................................44
4-5. Average daily rainfall for the 10 days before harvest for fall 2001 and spring 2002
seasons............................................................................................................................44
5-1. Primers, band size, and Chi-square goodness of fit test to a 1:2:1 or 3:1 single gene
ratio at p≤0.05.................................................................................................................60
5-2. Markers that were significantly linked to sugar, pH, TA, yield, SS, plant habit and
fruit sizez .........................................................................................................................63
5-3. Sugar means of F2 individuals with or without each marker z ........................................64
5-4. Combinations of markers that allow for groups of plants with a high sugar meanz .......65
A-1. Source and pedigree of breeding lines used in experiments..........................................70
vii
LIST OF FIGURES
page
Figure
1-1. Selected morphological traits of tomato: a) fruit of PI270248 (left) and Fla.7833
(right), b) jointed pedicel showing adsission zone, c) jointless pedicel, d)
indeterminate plant habit with vegetative apices, e) determinate plant habit with
stems terminating in two flower clusters........................................................................5
2-1. Sucrose, glucose and fructose means for PI270248 and 7833 in fall 2001 and spring
2002 at Bradenton, Florida.............................................................................................16
2-2. Frequency distribution for total sugars of PI270248 and 7833 plants in fall 2001 and
spring 2002 at Bradenton, Florida..................................................................................17
4-1. Frequency distribution of sugar level for PI270248, 7833, F1 and F2 generations in
fall 2001. The sugar mean and standard deviation for each generation is shown
above their respective diagram.......................................................................................40
4-2. Frequency distribution of sugar values for PI270248, 7833, F1, F2, BC1 and BC2
generations in spring 2002. The sugar mean and standard deviation for each
generation is shown above their respective diagram......................................................41
5-1. RAPD markers linked to high sugars (c=coupling, r=repulsion, P1=PI270248,
P2=7833). Arrows point to the polymorphic bands .......................................................61
5-2. Linkage groups determined by MAPMANAGER (QTXb15) and confirmed by
MAPMAKER 3.0. Kosambi map function was used.....................................................62
viii
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
HIGH FRUIT SUGAR CHARACTERIZATION, INHERITANCE AND LINKAGE OF
MOLECULAR MARKERS IN TOMATO
By
Nikolaos Georgelis
December 2002
Chair: J.W. Scott
Major Department: Horticultural Sciences
Consumers often complain about the overall flavor of fresh market tomatoes.
Tomato flavor is controlled by sugars, acids and volatiles, and one way to improve it is to
increase the sugar level within a certain acid range.
The small-fruited cherry accession PI270248 (L. esculentum var. cerasiforme) was
tested as a source of high sugars. It was shown that fructose and glucose were the major
fruit sugars present in almost equal quantities, while sucrose was present only in trace
amounts. PI270248 was crossed to the large-fruited inbred line Fla.7833-1-1-1 (7833)
that had significantly lower sugars in the ripe fruit. Sugars in the F2 were positively
correlated with soluble solids, glucose, fructose, pH and titratable acidity, and inversely
correlated with fruit size. Indeterminate plants had more sugars than determinate plants.
Yield, earliness and pedicel type were not significantly correlated with sugars at p≤0.05.
ix
The Broad-sense heritability of sugars was 0.72 in fall 2001 and 0.86 in spring
2002. Additive and heterozygote X heterozygote epistatic effects were significant at
p≤0.05. Epistasis skewed the F1 sugar mean toward PI270248. Seasonal effects on sugar
level were significant at p≤0.05 for PI270248 and the F1 that had more sugars in spring
2002 than in fall 2001, while 7833 was not affected by season. The general increase of
sugars in spring was possibly due to higher solar radiation.
Six random amplified polymorphic DNA (RAPD) markers linked to high sugars were
found, five dominant and one co-dominant. Five of the markers were also linked to small
fruit size, one to low yield, and one to indeterminate plant habit. Combinations of
markers for breeding purposes were tested. All the markers together explained 35% of the
sugar variation in the F2 of spring 2002.
x
CHAPTER 1
INTRODUCTION
The organoleptic quality of fresh market tomato (Lycopersicon esculentum Mill.) is
affected by fruit appearance, flavor (taste and aroma)(Shewfelt, 1993) and texture
(Causse et al., 2001; Vickens, 1977). Although the perception of flavor is influenced by
many factors, taste (sweetness, sourness) is one of its most important components and it
is determined basically by sugars and acids (Kader et al., 1977; Malundo et al., 1995;
Stevens et al., 1977a; Stevens et al., 1977b). Sugars in L. esculentum are mostly
comprised of glucose and fructose with trace amounts of sucrose (Davies and Hobson,
1981; Stevens, 1972), and Petro-Turza (1987) showed that the sweet taste of tomato was
attributed to reducing sugars. Fructose and glucose are found in almost equal quantities in
tomato fruit with fructose being a little higher, while sucrose usually does not exceed
0.1% (Davies and Hobson, 1981; Davies and Kempton, 1975; Petro-Turza, 1987). It has
also been shown that fructose is twice as sweet as glucose (Biester, 1925; Stevens et al.,
1977a) and it could be expected that modification of the fructose to glucose ratio without
changing the overall sugar level could give rise to sweeter fruits with better flavor.
However, accessions of other species like L. chmielewskii (Chetelat et al., 1995) and L.
hirsutum (Hadas et al.1995; Schaffer et al., 1998) accumulate sucrose instead of reducing
sugars. Citric and malic acids are the primary acids in tomato fruit with citric
predominating and determining sourness (Davies and Hobson, 1981). Sugar and organic
acids comprise the majority of the total dry matter content of tomato fruit. It has been
documented that the reducing sugars (fructose and glucose) generally correlate with
1
2
soluble solids (SS) content (Jones and Scott 1984; Kader et al., 1977; Malundo et al.,
1995; Stevens 1972, 1977a). Hence, taking soluble solids (SS) measurements provides a
good estimate of the sugar level.
Another important flavor component is aroma, which is determined by volatiles
(Baldwin et al., 1998; 2000; Kazeniac and Hall, 1970; Krumbein and Auerswald, 1998).
It has also been shown that volatiles can influence the perception of sweetness (Baldwin
et al., 1998). Volatiles can be divided into primary odorants such as methyl salicylate,
which bind just one olfactory receptor and secondary odorants such as Safrole, that bind
more than one (Amoore, 1952). They can also be divided into top notes (De Rovira,
1997) and background notes. The top notes are more noticeable in perception than the
background notes. In tomato, more than 400 volatiles have been reported, but only 30 are
present in concentrations >1nL.L-1 (Buttery 1993; Buttery and Ling, 1993a, 1993b;
Buttery et al. 1989). Buttery et al. (1971) determined odor thresholds for these
compounds (present in concentrations of 1nL.L-1 or more). No single compound has been
found to be reminiscent of a ripe tomato. Buttery (1993) has suggested that a combination
of cis-3-hexenal, cis-3-hexenol, hexenal, 1-penten-3-one, 3-methylbutanal, trans-2hexenal, 6-methyl-5-hepten-2-one, methyl salicylate, 2-isobutylthiazole and B-ionone, at
appropriate concentrations, produces the aroma of a fresh, ripe tomato. However, the
odor thresholds determined by Buttery (1993) may be inflated according to Tandon et al.
(2000). In the latter research, tomato fruit homogenate instead of water was used to dilute
volatiles, which was closer to reality. In conclusion, fewer than 30 volatiles may be
perceived in tomato fruit out of the 400 reported.
3
Finally, texture is another factor that plays an important role in the overall flavor
perception. More specifically, the state of the cell wall (the condition of the middle
lamella) and the mechanism of tissue disruption, i.e., whether fruit cells break across cell
walls or between cells, seem to influence juiciness and flavor impact (Vickers, 1977).
The relative contribution of taste and aroma to the overall flavor has not been
elucidated yet. Some investigators have emphasized the importance of volatiles to the
overall tomato-like flavor (Baldwin et al., 1998; Brauss et al., 1998; Krumbein and
Auerswald, 1998). However, the perception of aroma is not independent of the taste
(Voirol and Daget, 1987). It has been documented that aroma is perceived with greater
intensity when the level of reducing sugars is raised within a given range of acids
(Malundo et al., 1995), and that, generally, cultivars with higher sugars have a more
acceptable flavor (Baldwin et al., 1998). Thus, reducing sugars influence overall flavor in
two ways: they have a direct impact on the sweetness and also affect the intensity of
aroma perception. The above evidence shows that the level of sugars is a trait of
particular interest if a change in flavor is desired.
Consumers often complain about the flavor of fresh market tomatoes (Bruhn et al.,
1991). Indeed, most breeding efforts have been dedicated to other traits such as yield,
disease resistance, and traits that influence the post-harvest life and handling of the fruit.
Most commercial cultivars of fresh market tomatoes have a low level of sugars with total
soluble solids ranging from 4 to 5% (Kavanagh and McGlasson, 1983; McGlasson et al.,
1983). Malundo et al. (1995) found that increasing sugar to levels higher than that found
in most large fruited commercial cultivars enhanced the aroma intensity and made the
overall flavor more acceptable within a certain acid range.
4
There are several potential sources of high sugar among the wild species L.
chmielewskii, L. hirsutum, L. pimpinellifolium and some sweet cherry accessions
(‘Sugar’).Theoretically, these lines could be crossed to large fruited tomatoes that have a
low sugar level and, through modified backcrossing, the sugar level could be increased in
large fruit. In this study, accession PI270248 (‘Sugar’) (L. esculentum var. cerasiforme)
was used as a high sugar source. PI270248 is a small-fruited cherry tomato with a tall,
indeterminate vine, jointed pedicel, moderate yield and small fruit size (Figure 1-1). The
reason that accession PI270248 was preferred, apart from the sweet taste, was that its
fruit was resistant to bacterial spot [incited by Xanthomonas campestris pv. Vesicatoria
(Doidge) Dye](Scott et al., 1989), which is a major problem in Florida. In our
experiment, it was crossed to a large-fruited tomato line Fla.7833-1-1-1 (7833) that
produced low sugar fruits. Additionally, 7833 had a short, determinate vine, jointless
pedicel and high yield (Figure 1-1).
Since the sugar level of PI270248 has not been published yet, it would be useful to
determine the amounts of fructose, glucose and sucrose that account for its sweet taste
and compare its sugar level to that of Fla.7833 (7833) which had a sweetness level
similar to most commercial, large fruited cultivars. It would also be of interest to
determine the relationship of sugars coming from PI270248 with physical traits (like fruit
size, plant habit and yield) and chemical traits (such as pH, soluble solids content and
titratable acidity), because it would elucidate if sugars could be transferred to largefruited tomatoes without affecting other traits.
Sugar level in tomato fruit seems to be a polygenic trait. There has been research in
which genetic maps were constructed and Quantitative Trait Loci (QTL) analyses were
5
a)
b)
c)
d)
e)
Figure 1-1. Selected morphological traits of tomato: a) fruit of PI270248 (left) and
Fla.7833 (right), b) jointed pedicel showing adsission zone, c) jointless pedicel, d)
indeterminate plant habit with vegetative apices, e) determinate plant habit with stems
terminating in two flower clusters
conducted (Bernacchi et al., 1998; Causse et al., 2001; Fulton et al., unpublished;
Tanksley et al., 1996). Those attempts have shown that there is more than one gene
controlling the level of total sugars. However, in all cases, different crosses were used
from the one used in this research. Hence, an estimate of the heritability of high sugars,
and the number of genes controlling high sugars, would give an idea of how many genes
6
need to be transferred to low sugar tomatoes to raise their sugar level, and the possible
gain from such a transfer.
However, incorporation of high sugar level from a small-fruited source into largefruited tomatoes is likely to be extremely difficult. Selection in F2 and possible F3
generations would be required between backcrosses to incorporate these genes into large
fruited tomatoes. Moreover, difficulties would be faced in phenotypic selection as sugar
level is influenced by the environment (Husain et al. 2001; Saliba-Colombani et al.,
2001), which means that genotypes with lower potential for high sugars might be selected
at the expense of genotypes with higher potential that were not favored by the
environmental conditions. Additionally, in order to get a fair estimate of sugar level, one
has to resort to laborious and time-consuming methods, like High Performance Liquid
Chromatography (HPLC) analysis. Another problem being faced during such attempts is
the negative effects on phenotype that are carried along with the genes of interest by
linkage drag. Indeed, most potential high sugar sources (wild species or cherry tomatoes)
have traits such as small fruit size and low yield that may be transferred along with high
sugars to large fruited tomatoes. McGillinary and Clemente (1956) showed that soluble
solids were inversely correlated to fruit size in the same plant, while Emery and Munger
(1970) reported that indeterminate plants had more soluble solids than determinate forms
of almost isogenic plants. These studies show that there are undesirable physiological
relationships between soluble solids (and probably sugars) and fruit size and plant habit.
Finally, Stevens and Rudish (1978) reported that yield was inversely correlated to soluble
solids.
7
The use of molecular markers for Marker Assisted Selection (MAS) would facilitate
the incorporation of genes contributing to high sugars into commercial cultivars. Markers
could reduce the plant generations required between crosses to incorporate genes into a
cultivar and are not subject to environmental variation. Moreover, markers linked to the
trait of interest, but not linked to undesirable traits, could be used to transfer a significant
part of the trait without negative effects caused by linkage drag.
There are many kinds of markers that have already been used in breeding programs,
including isozymes (Feuerstein et al., 1990; Summers et al., 1988;), restriction fragment
length polymorphisms (RFLP) (Osborn et al., 1987), random amplified polymorphic
DNAs (RAPD) (Williams et al., 1990), microsatellites (Akagi et al., 1996), sequence
characterized amplified regions (SCAR) (Barret et al., 1998), cleaved amplified
polymorphic sequences (CAPS) (Caranta et al., 1999), and amplified fragment length
polymorphisms (AFLP) (Caranta et al., 1999).
Isozymes (Market and Miller, 1959) were among the first group of molecular
markers used for genetic diversity assessment and genetic linkage map development.
They are cost effective and co-dominant. Their basic limitations are that much of the
genome does not code for proteins, different biochemical procedures are required to
visualise allelic differences for enzymes having different functions, and many proteins
take their final form after post-transcriptional steps that remove parts of the DNA
sequence and thus can mask variation present at the DNA level.
Restriction Fragment Length Polymorphisms (RFLPs)(Botstein et al.,1980) are codominant, repeatable, and have a known map position in the tomato genome. On the
other hand, they are expensive and laborious. They have been used extensively in tomato
8
for the construction of genetic maps (Tanksley et al., 1992) and for linkage to agronomic
traits (Osborn et al., 1987).
Random amplified polymorphic DNAs (RAPDs) (Williams et al., 1990) are
sometimes unreliable, show bands of low clarity and are usually not co-dominant, but
they are cost effective and easy to use, especially with large populations used in breeding
programs. An additional disadvantage is that they appear to be in clusters and not evenly
distributed in the tomato genome either in interspecific (Grandillo and Tanksley, 1996) or
intraspecific crosses (Saliba-Colombani et al., 2000). However, there are some examples
of RAPDs linked to polygenic agronomic traits (Doganlar et al., 2000; Foolad and Chen,
1998; Wing et al., 1994).
Microsatellites (Morgante and Oliveri, 1993) can require considerable investment to
generate, but are then highly polymorphic and inexpensive to use in mapping and MAS.
They are highly repeatable and target hypervariable regions of the genome.
Polymorphism is usually due to differences in length of the amplified product. They can
be cost effective, but the start-up costs are large. These costs should be justifiable for
crops where large-scale mapping and MAS are a practical necessity.
Sequence Characterized Amplified Regions (SCARs)(Paran and Michelmore, 1993)
are PCR-based secondary markers that come from RAPD polymorphisms. They amplify
a single fragment with high reproducibility. Many are co-dominant and their
polymorphism can often be increased by digesting the PCR product with restriction
enzymes.
Cleaved Amplified Polymorphic Sequences (CAPSs)(secondary markers)
(Lyamichev et al., 1993) are identified with two oligonucleotide primers synthesised on
9
the basis of known DNA sequences. Like SCARs, they specifically amplify single
fragments. However, polymorphism of CAPSs is revealed by digestion of the amplified
DNA with several restriction endonucleases.
Finally, Amplified Fragment Length Polymorphisms (AFLPs)(Vos et al., 1995) are
usually dominant markers, need radioactive labeling, and are more laborious to work with
than RAPDs. Some of their advantages are that they need a smaller quantity of DNA than
RFLPs, are more reliable and reproducible than RAPDs, and give more bands per
reaction than RAPDs and RFLPs.
Several screening techniques have been used to find linkage between a trait and a
marker, such as single plant comparisons, comparisons of near isogenic lines (Martin et
al., 1991; Young et al., 1988), F2 tail end comparisons (Darvasi and Soller, 1994), bulked
segregant comparisons (Michelmore et al., 1991), and recombinant inbred lines (Mohan
et al., 1994).
In this research, RAPD markers were used mainly because it was easier and faster to
screen large populations. Additionally, they were cheaper than RFLPs or AFLPs. Single
plant comparisons were used to determine which markers were linked to high sugars.
In summary, the first goal of the present research was to characterize the sugars from
PI270248 (fructose, glucose and sucrose levels) compared to that of the low sugar parent.
Then starting from the cross PI 270248 X 7833, the relationships of sugars with physical
traits (such as plant habit, yield and fruit size) and chemical traits (such as soluble solids
content, pH and titratable acidity) were determined. The inheritance of high sugars and
the effect of different seasons (spring, fall) on the total sugar level were also determined.
Finally, attempts were made to find RAPD polymorphisms linked to the genes
10
controlling total sugar level in tomato. These polymorphisms could be a helpful tool in
MAS in order to incorporate high sugar level from accession PI270248 into large fruited
tomatoes.
CHAPTER 2
CHARACTERIZATION OF HIGH SUGARS FROM PI270248
Introduction
Sugars constitute an important component of tomato (Lycopersicon esculentum) fruit
as they determine sweetness and influence the overall tomato flavor (Baldwin et al. 1998;
Stevens et al., 1979). The major sugars in tomato fruit are glucose, fructose and sucrose
(Davies and Hobson, 1981; Stevens, 1972). Sucrose is the form in which sugars are
transferred from the sites of production (leaves) to fruit. Then, sucrose is broken down to
glucose and fructose by the action of multiple forms of Acid invertase (cell bound,
soluble Acid invertase and Alkaline invertase) and Sucrose synthase (Yelle et al., 1991).
The action of those enzymes reduces the concentration of sucrose in the fruit and
increases the gradient between leaves and fruit. It has been shown that Sucrose synthase
is more important at the early stages of fruit development and determines the sink
strength of the fruit, while Acid invertase was active in fruit ripening (D’Aoust et al.,
1999).
Generally, tomato fruits contain predominantly the reducing sugars, glucose and
fructose, with fructose being found at slightly higher concentrations than glucose, while
sucrose does not exceed concentrations more than 0.1% (Davies and Hobson, 1981;
Davies and Kempton 1975; Petro-Turza, 1987). However, wild species accessions of L.
peruvianum, L .chmielewskii, and L. hirsutum have sucrose as the predominant sugar. It
has been shown, in all cases, that sucrose accumulation is a recessive monogenic trait
(Chetelat et al., 1995; Hadas et al., 1995; Schaffer et al., 1998; Stommel, 1992). The gene
11
12
controlling this trait encodes for an inactive form of invertase, which does not convert
sucrose to glucose and fructose efficiently. The reducing sugars, glucose and fructose are
found in very small quantities in these fruits. The sugar level of these wild species
accessions is elevated compared to L. esculentum (Stevens and Kader, 1976). In some
experiments, the gene responsible for high sucrose level was introgressed into L.
esculentum lines and apart from changing the sugar composition it reduced fruit size
significantly (Klann et al., 1996). A possible reason could be that fruit absorbs less water
due to reduced osmotic potential. The total sugar level remained at the same levels as
before the introgression of the gene, which showed that the gene conferring high sucrose
level did not affect the total amount of sugar in tomato fruit. However, sucrose is an
inactive form of sugar and unlike fructose and glucose, it is not consumed by fruit during
respiration. Thus, the sugar level can remain high and stable through time.
Apart from the high sucrose level, L. hirsutum has been used as a source of the gene
Fgr, which modulates the ratio of fructose to glucose (Levin et al., 2000). More
specifically, it has been shown that L. hirsutum has a Fru:Glu ratio more than 1.5:1 which
contrasts with L. esculentum where glucose and fructose are almost equimolar. This
increased ratio can be transferred to L. esculentum without changing the overall sugar
level, which means that the amount of fructose is increased and that of glucose is equally
reduced. Biester (1925) showed that fructose is almost twice as sweet as glucose, which
means that tomatoes with the Fgr gene compared to tomatoes with an equal sugar level
should taste sweeter.
In this research, the small cherry accession PI270248 was examined as a potential
source of high sugars. More sweetness was perceived in this line (J.W.Scott, personal
13
communication) than in large fruited lines with sugar levels similar to those grown
commercially. Moreover its fruit are resistant to bacterial spot [incited by Xanthomonas
campestris pv. vesicatoria (Doidge) Dye](Scott et al., 1989), which is a major problem in
Florida. Accession PI270248 belongs to species L. esculentum var. cerasiforme (National
Plant Genetic Resources Center). However, its fruit (diameter=2cm) is smaller than the
average size of cherry tomatoes (~3cm). It can be assumed that one or more of the wild
species is in its ancestry and thus, it is possible that sucrose is the predominant fruit sugar
or there could be an increased fructose to glucose ratio. PI270248 could be used by
breeders, as a source of high sugars, in order to transfer genes conferring sweetness to
large-fruited tomatoes, that otherwise have a low sugar level. Hence, it would be useful if
the relative levels of glucose, fructose and sucrose were known. A high level of sugars
compared to the average, generally low, sugar level of large fruited tomatoes would
verify its usefulness as a source of sweetness.
If sucrose was found in traces and fructose almost equal to glucose, the accession
could be used as a source of sweetness attributed to glucose and fructose, which is also
the type of sweetness found in commercial large fruited cultivars. However, if sucrose is
found at significantly higher concentrations than 0.1%, it could be considered that some
or most of the sweetness could be attributed to sucrose. Sucrose may provide a type of
sweetness other than the one provided by glucose and fructose (Karen Koch, personal
communication). This may be considered seriously by breeders, as potential sources of
sweetness are sought.
14
Thus, the goal of this research was to determine the levels of glucose, fructose and
sucrose and the ratio fructose:glucose of accession PI270248 compared to a large fruited
tomato to understand what sugars will be transferred from this accession.
Material and Methods
Accession PI270248 and inbred Fla.7833-1-1-1 (7833), which had a low sugar level
similar to many commercial cultivars, were grown in fall 2001 and spring 2002 at Gulf
Coast Research and Education Center, in Bradenton.
Fall 2001
Seed were sown in the greenhouse in Black Beauty spent coal (Reed Minerals Div.,
Highland, Ind.) medium on 27 Jul., 2001 and transplanted into Todd planter flats (3.8cm3 cell size)(Speedling, Sun city, Fla.) on 6 Aug., 2001. They were transplanted to the
field on 30 Aug., 2001, on 20-cm-high, 81-cm-wide beds of EauGallie fine sand that had
been fumigated with 67% methyl bromide: 33% chroropicrin at 392 kg.ha-1 and covered
with white polyethylene mulch 2 weeks before transplanting. The entries were arranged
in a randomized complete block design with three blocks and ten plants per plot. Plants
were spaced 46cm apart within plots that were 91cm apart in rows, with 152cm between
rows. Recommended fertilizer and insecticide programs were followed (Hochmuth et al.,
1988). Plants were grown with stake culture and irrigated by seepage from ditches
adjacent to the six experimental beds.
On 10 Dec., 2001, approximately 15 table-ripe fruits per plant from PI270248 and 3-4
table-ripe fruits per plant from 7833 were harvested. Fruits from each plant were ground
using a Waring blender and the fruit homogenate was stored in plastic bags at –20oC until
use. Fruit homogenate (40g) was added to 70 ml of 80% ethanol, boiled for 15min (with a
loose-fitting cover), cooled, and vacuum-filtered through Whatman #4 filter paper. The
15
resulting extract was brought up to 100ml with 80% ethanol; and 12ml then was passed
through a C-18 Sep Pak (Waters/Millipore, Milford, Mass.) and a 0.45-µm Millipore
filter. The filtered extract was injected by a Bio-Rad AS-100 HPLC autosampler (BioRad, Richmond, Calif.), fitted with a 20-µl sample loop, into a Perkin Elmer Series 410
HPLC system. Sugars were analyzed using a Waters Sugar Pak column at 90C with a
mobile phase of 100µM ethylenediamine-tetraacetic acid disodium-calcium salt
(CaEDTA) and a flow rate of 0.5 ml.min-1. A Perkin Elmer LC-25 Refractive Index
detector was used to measure sugars. Filtered analytical grade reagents were used for
standard preparation to establish high performance liquid chromatography (HPLC)
retention times and calibration. Determination of purity of individual peaks was
accomplished by absorbance index (all wavelengths monitored simultaneously) on a
Perkin Elmer LC-235 Diode Array detector.
Spring 2002
Seed were sown in the greenhouse in Black Beauty spent coal medium on11 Jan., 2002
and transplanted into Todd planter flats on 25 Jan., 2002. They were transplanted to the
field on 6 Mar., 2002 on beds of EauGallie fine sand covered with black polyethylene
mulch 2 weeks before transplanting. The entries were arranged in a randomized complete
block design with four blocks and ten plants per plot. Other growing procedures and
techniques were the same as in fall 2001. Fruit was harvested on 23 May, 2002, was
ground and stored at –20oC. Values of sugars were estimated using the same proceduces
as in fall 2001.
Both Seasons
The estimates of means and variances were made using SAS (SAS Institute, Cary,
N.C.). The same software and more specifically, General Linear Model and Duncan’s
16
multiple range test were also used to compare total sugar level means and
fructose/glucose ratio between PI270248 and 7833, while the frequency distributions
were made using Microsoft Excel 2000.
Results
Samples of 21 PI270248 and 30 7833 plants in fall, and 30 PI270248 and 38 7833
plants in spring were taken. The means of sucrose, glucose and fructose, and their
standard deviations are summarized in Figure 2-1.
SUGARS (FALL, SPRING)
3
g/100g of f.w.
2.5
2
SUCROSE (FALL)
GLUCOSE (FALL)
FRUCTOSE (FALL)
1.5
SUCROSE (SPRING)
GLUCOSE (SPRING)
1
FRUCTOSE (SPRING)
0.5
0
PI270248
7833
Figure 2-1. Sucrose, glucose and fructose means for PI270248 and 7833 in fall 2001 and
spring 2002 at Bradenton, Florida
The frequency distributions show that the total sugar values of PI270248 lie between
4-6g/100g of f.w. while those of 7833 between 1-3g/100g of f.w. with no overlapping in
either season (Figure 2-2).
17
PI270248 and 7833
% of individuals
100
80
PI270248 in fall
60
PI270248 in spring
7833 in fall
40
7833 in spring
20
0
0-1
1-2
2-3
3-4
4-5
5-6
g/100g of f.w.
Figure 2-2. Frequency distribution for total sugars of PI270248 and 7833 plants in fall
2001 and spring 2002 at Bradenton, Florida
Glucose and fructose levels were similar to each other and much higher than sucrose
for both genotypes, each season. Glucose and fructose were significantly higher for
PI270248 than 7833 both seasons as expected.
Comparisons of total sugar means and fructose/glucose ratios between PI270248 and
7833 were made (Table 2-1).
Table 2-1. Total sugar means and fructose/glucose ratios of PI270248 and 7833 in fall
2001 and spring 2002 at Bradenton, Florida.
Total Sugars(g/100g of f.w.)
Fructose/Glucose
Line
Fall
Spring
Fall
Spring
PI270248
4.66az
5.00a
1.032b
0.99b
7833
2.21b
2.17b
1.125a
1.09a
z
mean separation within column by Duncan’s multiple range test at p≤0.05.
18
The analysis of variance (p<0.001) along with Duncan multiple range test showed that
the total sugar level of PI270248 was significantly higher than that of 7833 in both
seasons. It also showed that the ratio Fru/Glu of 7833 was significantly higher than that
of PI270248 (p<0.001) in both seasons.
Discussion
The results clearly show that glucose and fructose are the predominant sugars in
accession PI270248. Both of them account for more than 99% of total sugars in its fruit.
On the other hand, sucrose is found at concentrations lower than 0.1g/100g f.w., which is
consistent with what is found in most lines of the species L. esculentum (Davies and
Hobson, 1981). Fructose is found in almost equal concentrations with glucose, which is
also consistent with previous results for this species. The difference between fructose to
glucose ratios of 7833 (1.10) and PI270248 (1.01) was not large (0.09) even though it
was statistically significant. It could be deduced that PI270248 is not among the lines that
have a high Fru/Glu ratio, as was found in L. hirsutum (Fru/Glu>1.5). On the contrary,
fructose seemed to be almost equal with glucose. Since the total sugars of PI270248 were
significantly higher than those of line 7833 for both seasons it confirms the potential of
PI270248 as a source of high sugar for breeding purposes. The sugar level of 7833 was
expected to be similar to average level of commercial cultivars used in Florida, like
‘Florida 47’ and ‘Sanibel’ (Soluble Solids: 4 to 5%, (Georgelis, unpublished data)).
Furthermore, recent experiments showed that addition of reducing sugars
(Frucose:Glucose=14:11) to large fruited fresh market tomatoes with a low sugar level
enhanced not only their sweetness but also their aroma intensity and made their overall
flavor more acceptable (Malundo et al., 1995). These results, combined with ours, show
19
that accession PI270248 can be used in breeding programs as a source of the reducing
sugars, glucose and fructose that may improve tomato flavor. Additionally, since sucrose
was not found in concentrations adequate to explain part of the sweetness of PI270248
and Fru:Glu ratio was not very different from most commercial cultivars, a change in the
type of sweetness of large-fruited tomatoes, by genes from that accession, is not
expected.
CHAPTER 3
RELATIONSHIP OF HIGH FRUIT SUGARS WITH OTHER TRAITS
Introduction
There has been considerable research about the relationship of soluble solids with other
traits using crosses different from the one used in this research (Emery and Munger,
1970; McGillinary and Clemente, 1956; Stevens and Rudich, 1978). In some cases, these
relationships were estimated in attempts to construct genetic maps in tomato (Bernacchi
et al., 1998; Tanksley et al., 1996). Many times, soluble solids content was used instead
of sugar level, because the former can be measured in a much easier and faster way than
the latter. Additionally, soluble solids content is a more important trait in processing
tomatoes. A fair estimate of soluble solids can be made with handheld or digital
refractometers, while sugar determination requires more laborious and time-consuming
methods, such as High Performance Liquid Chromatography (HPLC) analysis. Sugars
constitute a large part of total soluble solids (~50%), while the remainder is made up of
acids, vitamins and polyphenols (Davies and Hobson, 1981; Davies and Kempton, 1975).
It has been shown that sugar content is positively correlated with total soluble solids
content in tomato fruit and in most cases this correlation is high (Jones and Scott, 1984;
Kader et al., 1977; Malundo et al., 1995; Saliba-Colombani et al., 2001; Stevens 1972;
Stevens et al., 1977a). Hence, generally, soluble solids content measurements can give a
fair estimate of the sugar level in tomato fruit.
In an attempt to transfer high sugars from a wild species or a cherry tomato to large
fruited tomatoes it is useful to know how other important traits are affected by such a
20
21
change in sugar content. One of the traits of this kind is plant habit, which is an important
trait for commercial fresh market tomatoes. Indeterminate plants grow longer and they do
not have a concentrated fruit set. They are predominant in greenhouses where the support
of the long vines is easier. On the other hand, determinate tomatoes have a more compact
plant size and they are ideal for field crops, where the means of plant support are
restricted. Moreover, they have a concentrated fruit set which results in an easier and
cheaper harvest. Emery and Munger (1970) showed that indeterminate plants had higher
soluble solids than determinate plants using nearly isogenic lines that differred only in
plant habit.
Another important trait is the type of pedicel. Tomatoes without an abscission zone in
the pedicel are called jointless tomatoes. Jointlessness is controlled by recessive genes,
with j2 being the one used commercially. Jointless tomatoes are generally preferred
because they do not have stems that puncture each other after harvest as do jointed when
they are not de-stemmed. There has not been research up to now that shows the
relationship of sugar with type of pedicel. However, field testing of jointless tomatoes
indicated a trend for them to be bland and not very sweet (J.W. Scott, personal
communication).
Earliness is another important trait of commercial tomatoes since early cultivars
sometimes produce ripe fruits at a high-price period. Cultivars with later maturity often
have larger vines and perhaps a larger leaf:fruit ratio. In addition, ripening times may be
longer than those of early ripening plants. It could be expected that the later maturing
plants have more time to provide fruits with carbohydrates. Thus, it would be interesting
to see if earliness could affect sugar level.
22
High yield is always desirable and a major factor in why growers select a cultivar to
grow. Historically, major breeding efforts have been dedicated to maximize this trait.
Nowadays, there is also pressure by consumers for better organoleptic traits. However,
yield is always a trait of importance in any breeding effort and it needs to be as high as
possible. Unfortunately, it has been shown that yield is inversely correlated with soluble
solids (Stevens and Rudich, 1978), a fact that creates problems in efforts for flavor
improvement via increased soluble solids or sugars.
Fruit size is also among the important traits for fresh market tomatoes where large size
is often preferable. However, there are several cherry and grape tomatoes in the market
with intense flavor characteristics. McGillinary and Clemente (1956) showed that larger
fruits had slightly lower soluble solid values than smaller fruits within the same line.
Additionally, there has been research where plants were irrigated with saline water
(Adams, 1991; Gough and Hobson, 1990) and this resulted in tomatoes with higher
soluble solids content but smaller fruit size. Thus, it might be difficult to develop large
fruited tomatoes with high sugars.
Finally, two chemical traits of particular interest are pH and titratable acidity (TA).
Paulson and Stevens (1974) showed that pH was highly correlated with [H+] and TA
with citric acid and malic acid, while the correlation between pH and TA was very low in
some cases. Values of pH are crucial for processing tomatoes since values higher than 4.4
mean susceptibility of the pulp to thermophilic pathogens (Paulson and Stevens, 1974).
Thus, pH values as low as possible (up to the point that it does not adversely affect taste)
should be bred into tomato cultivars for industrial use. Both pH and TA can give an
estimate of how sour the fruit tastes. Harvey (1920) reported that the sourness of a
23
solution is related to its total free acids and hydrogen ion concentration. In some cases,
TA showed higher correlation with sourness than did pH (Jones and Scott, 1984), while
in another report the opposite was observed (Baldwin et al., 1998). In fresh market
tomatoes, it has been shown that a good balance between sugars and acids (sugar:TA or
SS:TA) was very important for a good overall tomato flavor (Baldwin et al., 1998;
Stevens, 1972; Stevens et al., 1977a), while elevated TA values increased the overall
flavor acceptability (Baldwin et al., 1998). Saliba-Colombani et al. (2001), used a cherry
accession for crossing and estimated a positive correlation between sugar and TA. The
relationship of sugars to pH has been variable as it has been positive, negative and
insignificant (Bernacchi et al., 1998; Saliba-Colombani et al., 2001; Tanksley et al.,
1996).
The objective of this research was to evaluate the relationship of high sugars derived
from PI270248 with other traits to discern which ones may be important to consider in
the transfer of high high sugars into large fruited tomato lines. The physical traits,
measured in this experiment, were plant habit (indeterminate, determinate), pedicel type
(jointed, jointless), yield, earliness and fruit size. The chemical traits were pH, titratable
acidity (TA) and soluble solids content (SS). Positive or negative correlations of sugar
level with another trait of interest could mean that one or more genes controlling sugar
level are linked to one or more genes controlling the other trait. The other possibility is
that there are pleiotropic genes controlling both traits. In the latter, trying to transfer
genes controlling high sugar level from accession PI270248 into another line would
inevitably affect other traits. It would not be desirable to transfer small fruit size or low
yield or, in some cases, indeterminate growth to lines destined for the market.
24
Hence, a positive or negative relationship of sugar level with another trait would mean
that it might be very difficult to pass an intact sugar level from PI270248 on to another
line without affecting the other trait. However, it might be possible to pass a part of
sweetness without affecting the other trait significantly. For example, it may be extremely
difficult to pass an intact level of sweetness from the indeterminate PI270248 onto the
determinate 7833 without affecting plant habit, but it may be possible to pass a part of
this sweetness without affecting plant habit. On the other hand, if sugar level was weakly
or not correlated with another trait then it probably could be transferred to other lines
without affecting that particular trait.
Materials and Methods
Fall 2001
Accession PI270248 was crossed to Fla.7833-1-1-1 (7833) and subsequently F2 seed
were obtained. The parents and F2 were grown in a Randomized Complete Block Design
(RCBD) with 3 blocks and 10 plants per plot. However, samples were taken from 21
plants of PI270248, as some plants were lost. Seed were sown in the greenhouse in Black
Beauty spent coal (Reed Minerals Div., Highland, Ind.) medium on 27 Jul., 2001, and
transplanted into Todd planter flats (3.8-cm3 cell size)(Speedling, Sun city, Fla.) on 6
Aug., 2001. They were transplanted to the field on 30 Aug., 2001 on 20-cm-high, 81-cmwide beds of EauGallie fine sand that had been fumigated with 67% methyl bromide:
33% chroropicrin at 392 kg.ha-1 and covered with white polyethylene mulch 2 weeks
before transplanting. Plants were spaced 46cm apart within plots that were 91cm apart in
rows, with 152cm between rows. Recommended fertilizer and insecticide programs were
followed (Hochmuth et al., 1988). Plants were grown with stake culture and irrigated by
seepage from ditches adjacent to the six experimental beds.
25
Spring 2002
The parents and F2 were grown in a Randomized Complete Block Design (RCBD)
with 4 blocks with 10 plants per plot for the parents and 50 plants per plot for the F2.
However, samples were taken from 30 plants of PI270248, 38 of 7833 and 198 of F2, as
some plants were lost. Seed were sown in the greenhouse in Black Beauty spent coal
medium on 11 Jan., 2002 and transplanted into Todd planter flats on 25 Jan., 2002. They
were transplanted to the field on 6 Mar., 2002 on beds of EauGallie fine sand covered
with black polyethylene mulch 2 weeks before transplanting. Other growing procedures
and techniques were the same as in fall 2001.
Both Seasons
Plant habit was determined, in the field, visually and plants were separated into 2
groups (indeterminate growth, determinate growth). Earliness and yield were determined,
in the field, subjectively when the earliest plants had almost 100% table-ripe fruit. For
earliness, plants were separated into 3 groups (1-3 scale, 1=very late fruit ripening,
3=very early fruit ripening). Yield was determined visually, by taking into account the
number and the size of fruits per plant. Plants were rated on a scale 1-5 (1=very low
yield, 5=very high yield). The pedicel type and fruit size were also determined in the field
when some fruit were ripe in each plant. The type of pedicel was determined visually and
plants were separated into 2 groups (jointed, jointless). For fruit size, a Craftsman caliper
was used to measure the average diameter of 3-5 table-ripe fruits per plant.
On 10 Dec., 2001 and 23 May, 2002, approximately 3 –15 table-ripe fruits per plant
were harvested depending on the fruit size. Fruits from each plant were ground using a
Waring blender and the fruit homogenate was stored in bags at –20oC prior to analysis.
26
Part of that pulp was used to measure sugars and part of it was used for measuring pH,
titratable acidity (TA) and soluble solids content (SS).
Sugars and acids were extracted using the following procedure: Fruit homogenate
(40g) was added to 70ml of 80% ethanol, boiled for 15 min (with a loose-fitting cover),
cooled, and vacuum-filtered through Whatman #4 filter paper. The resulting extract was
brought up to 100ml with 80% ethanol; and 12ml was then passed through a C-18 Sep
Pak (Waters/Millipore, Milford, Mass.) and a 0.45-µm Millipore filter. There was 5-8ml
of filtered extract per plant available for HPLC analysis.
Sugar measurement was conducted, at the USDA Citrus and Subtropical Products Lab
in Winterhaven, FL. The filtered extract was injected using a Bio-Rad AS-100 HPLC
autosampler (Bio-Rad, Richmond, Calif.), fitted with a 20-µl sample loop, into a Perkin
Elmer Series 410 HPLC system. Sugars were analyzed using a Waters Sugar Pak column
at 90C with a mobile phase of 100µM ethylenediamine-tetraacetic acid disodium-calcium
salt (CaEDTA) and a flow rate of 0.5 ml.min-1. A Perkin Elmer LC-25 Refractive Index
detector was used to measure sugars. Filtered analytical grade reagents were used for
standard preparation to establish High Performance Liquid Chromatography (HPLC)
retention times and calibration. Determination of purity of individual peaks was
accomplished by absorbance index (all wavelengths monitored simultaneously) on a
Perkin Elmer LC-235 Diode Array detector.
The pH, titratable acidity, and soluble solids content were measured using the
following proceduce: Tubes (50ml) were filled up with tomato pulp from each plant and
were centrifuged with a Lounges Instrument Corp. centrifuge at 12000rpm for 5min. The
aliquot was filtered through filter paper (Fisher Brand) into 50ml beakers and was used to
27
measure pH with a Corning pH meter 340. Two to three drops of filtered aliquot were
also used to measure soluble solids content with an Atago-101 digital refractometer.
Then, 10ml of filtered aliquot were pipetted into 50ml flasks and 5-6 drops of
phenolphthalein were added. 0.1N NaOH was slowly poured into the flasks until
phenolphthalein changed from transparent to slightly purple and then the amount of
NaOH added was determined. Titratable acidity was measured using the formula:
T.A.(%Citric Acid)=(ml 0.1N NaOH) X 0.064 X 100
Pearson correlation coefficients (SAS Institute, Cary, N.C.) were determined for total
sugars with earliness, yield, fruit size, pH, titratable acidity, acids and soluble solid
content. Thirty F2 plants from fall 2001 and 198 F2 plants from spring 2002 were used for
analysis. The relationship of total sugars with plant habit and type of pedicel was
determined with Duncan multiple range test (SAS). This procedure provides the
advantage that F2 lines from different seasons can be used in the same analysis. However,
the F2 of each season was treated as a different line. So, in both cases there was a factorial
experiment with two factors (line and plant habit or line and pedicel type). Duncan’s
multiple range test was also used in order to compare the means of all the traits between
line PI270248 and 7833.
Results
For both seasons, PI270248 was significantly earlier and had significantly higher
sugars, soluble solids, TA, and pH than 7833 (Table 3-1). Additionally, PI270248 had
lower yield and fruit size than 7833.
28
Table 3-1. Comparison of PI270248 and 7833 for physical and chemical traits in fall
2001 and spring 2002 at Bradenton, Florida
Season Genotype Sugar(g/100g f.w.) SS(%) TA(%Citric)
pH
PI270248
7833
4.66az
2.21b
7.98a
3.81b
0.54a
0.29b
4.33a
4.25b
3a
1b
Fruit
size
2.0a 0.8a
4.5b 3.8b
Spring PI270248
2002
7833
5.00a
2.17b
8.03a
3.80b
0.60a
0.31b
4.36a
4.22b
3a
1b
2.0a 0.8a
4.5b 3.95b
Fall
2001
z
Earliness Yield
Mean separation in columns by Duncan’s multiple range test at p≤0.05
Abbreviations: SS= Soluble solids content, TA= Titratable acidity
Since PI270248 and 7833 were significantly different for all traits measured, it would
be expected that all of them would segregate in the F2 generation.
The relationship of sugar level with these traits and some information about
relationships between other important traits in the F2 population is in Table 3-2.
Indeterminate plants had significantly higher sugars (p<0.001) than determinate
plants, while the effect of type of pedicel on sugars was not significant (p=0.258) (Table
3-3).
Discussion
The data taken in spring 2002 are much more powerful than in fall 2001, because the
sample contained 198 plants instead of 30 in the second case. Data from fall were used to
give an insight about the correlation of sugar with other traits and give more support to
data obtained in spring.
Sugars were expected to correlate positively with soluble solids since they constitute a
significant part of them (Davies and Hobson, 1981). Indeed, the correlation was very
strong (~0.9) in both seasons, which suggests that soluble solids measurement is a very
good estimate of sugars.
29
Table 3-2. Correlation coefficients (r) of total sugars with glucose, fructose,
fructose/glucose, titratable acidity, pH, yield, earliness, and fruit size from F2 in fall 2001
and spring 2002 at Bradenton, Florida
Sugar
Fall
SS
0.898***z
Glucose (Glu) 0.990***
Fructose (Fru) 0.990***
Fru/Glu
-0.096
pH
0.457*
TA
0.317
Yield
-0.050
Earliness
0.160
Fruit size
-0.457*
Trait
z
Sugar
pH
pH
TA
TA
Spring
Fall
Spring
Fall Spring
0.923*** 0.415* 0.263*** 0.385* 0.577***
0.995***
0.995***
-0.387***
0.306***
0.395*** -0.500** -0.445***
-0.043
-0.124
-0.468***
the *, **, and *** indicate significance at p<0.05,0.01, and 0.001, respectively
Abbreviations: TA= Titratable acidity (% citric acid), SS= Soluble solid content (%)
Units: Glucose (g/100g f.w.), Fructose (g/100g f.w.)
Table 3-3. Sugar levels for F2 plants grouped by plant habit and pedicel type
z
Variable
Plant habit
Indeterminate
Determinate
Pedicel type
Jointed
Jointless
Plant #
Sugar (g/100g f.w.)
166
62
3.68az
2.98b
185
43
3.51a
3.42a
Mean separation within variable determined by Duncan’s multiple range test.
In this experiment, it was shown that total sugars (primarily reducing sugars) were
positively correlated to pH (0.306) and TA (0.395) and both correlations were highly
significant (p<0.001) in spring. In fall, there was also a positive correlation between
sugars and pH (0.457) and and between sugars and TA (0.317). However, the latter
correlation was not significant, possibly due to the small sample size. The correlation of
pH and TA was negative (-0.500 in fall, -0.445 in spring). The positive correlation
30
between sugar and TA confirms data from previous research where a cherry accession
was used (Cervil) (Saliba-Colombani et al., 2001). Reports on correlations between sugar
and pH have varied widely and they have been positive, negative, or insignificant
(Bernacchi et al., 1998; Saliba-Colombani et al., 2001; Tanksley et al., 1996). A positive
correlation of tomato fruit sugar with both TA and pH has not been reported yet.
However, Anderson (1957) suggested that there were lines with both pH and TA high,
and cases where pH and TA were not always inversely correlated. Stevens (1986)
suggested that both of them should be measured in tomato to determine the organoleptic
quality of the fruit. The positive correlation of sugar with pH in this experiment may be
an obstacle in an effort to transfer sugars from PI270248 to cultivars suitable for
processing. The elevation of sugar level may be accompanied by the increase of pH to
levels that are not desired by the processing industry. The positive correlation between
sugars and pH, and between sugars and TA means that plants with high sugars generally
have more free organic acids (citric and malic acid) and less hydrogen ion concentration
than plants with low sugars. Thus, whether plants with high sugars have more sourness
than plants with low sugars cannot be answered. In the future, a taste panel may be
helpful to answer this question.
It was also shown that indeterminate plants had a higher sugar mean than determinate.
This is in agreement with what Emery and Munger (1970) have shown. They suggested
that their results may be explained by the fact that indeterminate plants have more leaves
per fruit than do determinate plants. In Florida, determinate plants are the predominant
type grown. So, breeders may face an additional problem trying to transfer high sugars
from the indeterminate PI270248 to determinate large fruited tomatoes.
31
The correlation between sugars and yield was insignificant in both seasons. This
contradicts the negative correlation that has been documented between yield with sugars
or soluble solids in previous literature (Bernacchi et al., 1998; Saliba-Colombani et al.,
2001; Stevens and Rudich, 1978). The negative correlation, shown in previous research,
may be explained by the fact that plants with high yield may also have a high ratio of
fruit/leaf tissue and thus, they cannot provide fruit with as much carbohydrates as plants
with a lower ratio. However, in this research the ratio of fruit/leaf tissue did not seem to
be a problem. This is encouraging to breeders as it shows that it may be possible to
uncouple high sugars from low yield to some extent. However, once high sugars are
moved into more advanced high yielding backgrounds, a negative association of high
sugars and yield might yet be encountered.
Fruit size was negatively correlated to sugars (-0.468, p<0.001) and this confirmed
what other researchers have shown (McGillinary and Clemente, 1956; Goldman et al.
1995; Saliba-Colombani et al., 2001). However, it has not been delineated if this is
mostly due to linked genes or genes with pleiotropic effects. There is a tendency to accept
the second theory as the negative correlation has been observed in many cases. Stevens
(1986) suggested that large fruits tend to have more water in the cells, which dilutes
sugars and lowers their concentration. The negative correlation was quite strong in this
experiment and most probably it will be a problem to transfer high sugars from PI270248
to large fruited inbreds.
Finally, earliness and type of pedicel did not show any significant correlation with
sugars. It might have been expected that plants that ripen their fruits earlier have less time
to provide them with carbohydrates than plants that ripen later. Indeed, in spring, there
32
was slight negative correlation of sugars with earliness (-0.124), but it was not significant
at the 5% level (p=0.08).
To summarize, fruit size seemed to be the parameter that was inversely correlated to
sugars, starting from the cross PI270248 X 7833, and there may be an serious problem in
transferring sugars to large fruited tomatoes. However, Stoner and Thompson (1966)
showed that it is possible to select for fruits with high sugars and good size in their
research. An additional problem may also occur when sugars are transferred to
determinate plants, as high sugars seemed to be linked to indeterminate plant habit.
One cannot tell for sure if these parameters (plant habit and fruit size) will be
insurmountable obstacles to transferring high sugars to determinate large fruited
tomatoes. However, there is no need to transfer the high level of sugars from PI270248
(Brix: 8-8.5, Sugars: 5gr/100gr f.w.) intact. For instance, soluble solids of 6% (sugars of
3.5 g/100g of f.w.) in tomatoes with good fruit size and yield may be sufficient to
improve the overall tomato flavor significantly and keep growers satisfied.
CHAPTER 4
INHERITANCE OF HIGH SUGARS IN FRUIT DERIVED FROM PI270248 AND
SEASONAL EFFECTS ON FRUIT SUGAR
Introduction
Sugars constitute an important part of total soluble solids in tomato (Lycopersicon
esculentum) fruit (Kader et al., 1977; Malundo et al., 1995; Stevens et al., 1977a, 1977b)
and account for the sweet taste (Stevens et al. 1979). Taste is a crucial part of the overall
tomato flavor along with aroma. Taste in tomatoes is determined, primarily, by sugars
(glucose, fluctose) and acids (citric, malic) while aroma is determined by volatiles.
Consumers often complain about the flavor quality of large-fruited fresh market
tomatoes (Bruhn et al., 1991). One possible reason for this is that growers around the
world harvest tomatoes before they reach the desirable table-ripe (TR) stage, in order to
lengthen their shelf-life. However, tomatoes harvested before TR seem to have lower
amounts of soluble solids (Betancourt et al. 1977; Kader et al., 1977) and consequently,
they do not taste as good as a tomato harvested at the TR stage. Another reason that
accounts for less desirable tomato flavor is that commercial cultivars have a low potential
for producing good flavor. In breeding, there are other priorities such as high yield,
disease resistance and shipping ability, and it is difficult to breed for improved flavor
(Scott, 2002). Moreover, organoleptic traits, such as sugars seem more complicated and
there is evidence that they are under polygenic control (Bernacchi et al., 1998; Fulton et
al., unpublished; Saliba-Colombani et al., 2001; Tanksley et al. 1996). Additionally, the
33
34
importance of aroma components and the pathways leading to their formation, have
started to be delineated in the last few years (Tandon et al., 2000).
Sugars, apart from giving sweetness to tomato fruit, influence the intensity of aroma
perception (Malundo et al., 1995; Stevens et al., 1979). It has been shown that a sugar
level higher than most large-fruited cultivars, makes the fruits taste sweeter, increases the
aroma intensity and makes the overall flavor more acceptable. Hence, it seems that
transferring genes conferring sweet taste in tomato from a source of high sugars to large
fruited fresh market tomatoes, would have a great impact on the overall flavor.
In this project, accession PI270248 was studied as a source of high sugars. The
ultimate goal was to transfer the high sugar level to large-fruited cultivars to be grown
primarily in Florida. Fresh market tomatoes are among the most economically important
crops in the state with area around 17000ha (32.6% of US) and income more than $500
million (43.7% of US) per year (USDA Ag Statistics Board Vg 1-2 (01), Jan.01).
Currently, the predominant cultivars used by tomato growers in Florida are ‘Florida 47’
and ‘Sanibel’, both having a low sugar level (4-5% SS).
Knowledge about the inheritance of a trait of interest generally improves breeding
efficiency. Thus, one of the goals of this research was to determine genetic and
environmental effects on high sugars derived from accession PI270248. Genetic effects,
like heritability, would show how much of the sugar level variation could be attributed to
genetic factors and would give an estimate of the expected selection gain when
transferring genes conferring sweetness from PI270248 to large-fruited tomatoes. If
environmental effects accounted for most of the sugar variation, then not much gain
would be expected. An estimate of the number of genes accounting for the sugar level
35
difference between PI270248 and Fla.7833-1-1-1 (7833) (a line with a sugar level similar
to most commercial cultivars in Florida) was also attempted. Environmental effects
among seasons were also estimated to give an idea of how different seasons (fall, spring)
in Florida, could influence sugars coming from PI270248 in Florida.
Materials and Methods
Environmental Effects among Seasons
The small-fruited inbred accession PI270248 (L. esculentum var. cerasiforme) was
crossed with the large-fruited inbred 7833, which had a low sugar level similar to many
commercial cultivars. PI270248, 7833 and the F1 were grown in fall 2001 and spring
2002 at Gulf Coast Research and Education Center, Bradenton.
Fall 2001
Seed were sown in the greenhouse in Black Beauty spent coal (Reed Minerals Div.,
Highland, Ind.) medium on 27 Jul., 2001 and transplanted into Todd planter flats (3.8cm3 cell size)(Speedling, Sun City, Fla.) on 6 Aug., 2001. They were transplanted to the
field on 30 Aug., 2001 on 20-cm-high, 81-cm-wide beds of EauGallie fine sand that had
been fumigated with 67% methyl bromide: 33% chroropicrin at 392 kg.ha-1 and covered
with white polyethylene mulch 2 weeks before transplanting. The entries were arranged
in a randomized complete block design with three blocks and ten plants per plot. Plants
were spaced 46cm apart within plots that were 91cm apart in rows, with 152cm between
rows. Recommended fertilizer and insecticide programs were followed (Hochmuth et al.
1988). Plants were grown with stake culture and irrigated by seepage from ditches
adjacent to the six experimental beds. On 10 Dec., 2001, approximately 15 table-ripe
fruits per plant from PI270248, 10 from F1 and 3-4 from 7833 were harvested. Fruits
from each plant were ground using a Waring blender and the fruit homogenate was stored
36
in bags at –20oC until use. Fruit homogenate (40g) was added to 70ml of 80% ethanol,
boiled for 15 min (with a loose-fitting cover), cooled, and vacuum-filtered through
Whatman #4 filter paper. The resulting extract was brought up to 100ml with 80%
ethanol; and 12ml then was passed through a C-18 Sep Pak (Waters/Millipore, Milford,
Mass.) and a 0.45-µm Millipore filter. The filtered extract was injected using a Bio-Rad
AS-100 High Performance Liquid Chromatography (HPLC) autosampler (Bio-Rad,
Richmond, Calif.), fitted with a 20-µl sample loop, into a Perkin Elmer Series 410 HPLC
system. Sugars were analyzed using a Waters Sugar Pak column at 90C with a mobile
phase of 100µM ethylenediamine-tetraacetic acid disodium-calcium salt (CaEDTA) and a
flow rate of 2 ml.min-1. A Perkin Elmer LC-25 Refractive Index detector was used to
measure sugars. Filtered analytical grade reagents were used for standard preparation to
establish HPLC retention times and calibration. Determination of purity of individual
peaks was accomplished by absorbance index (all wavelengths monitored
simultaneously) on a Perkin Elmer LC-235 Diode Array detector.
Spring 2002
Seed were sown in the greenhouse on 11 Jan., 2002 and transplanted into Todd planter
flats on 25 Jan., 2002. They were transplanted to the field on 6 Mar., 2002. Beds were
covered with black polyethylene mulch 2 weeks before transplanting. The rest cultural
practices followed were the same as in fall 2001. The entries were arranged in a
randomized complete block design with four blocks and ten plants per plot. Fruit were
harvested on 23 May, 2002, ground and stored at –20oC. Using the same proceduces as in
fall 2001, values of sugars were estimated.
37
Both Seasons
Total sugar values were taken from 21 plants of PI270248, 30 of 7833 and 25 of F1 in
fall and 30 plants of PI270248, 38 of 7833 and 35 of F1 in spring. The experiment was a
factorial, with 2 factors. Season (fall, spring) and Line (PI270248, 7833, F1). The
significance of each effect along with their interaction was analysed by SAS (SAS
Institute, Cary, N.C.), using the General Linear Model procedure PROC GLM. The
effects of season on each line, were also estimated by PROC GLM (SAS). The sugar
means for each season were compared using Duncan’s multiple range test.
The average monthly temperature (oC), rainfall (cm) and solar radiation (W/m2) were
obtained from the internet site http://fawn.ifas.ufl.edu. Average daily rainfall (cm) was
also obtained from the same site.
Genetic and Environmental Effects within Season
PI270248 was crossed to 7833 and F1 was obtained. The F1 was self-pollinated to
obtain F2 seed. The F1 was also backcrossed to PI270248 and 7833 to obtain BC1 and
BC2 generations. In fall 2001, 21 PI270248, 30 7833, 25 F1 and 30 F2 plants were grown
at Gulf Coast Research and Education Center, in Bradenton, FL. In spring 2002, 30 plants
of PI270248, 38 of 7833, 35 of F1, 35 of BC1, 93 of BC2 and 198 of F2 were grown.
Plants were grown as described previously. Additionally, in spring 2002, BC1, BC2 and
F2 were arranged in a randomized complete block design with four blocks and ten plants
per plot for BC1, twenty-four plant per plot for BC2, and fifty plants per plot for the F2.
Fruit was harvested on 23 May, 2002 was ground and stored at –20oC. Using the same
procedures, as described previously, values of sugars were estimated.
38
Fall 2001
The broad sense heritability was calculated by hand using a method based on
subtracting the average variation of PI270248, 7833 and F1 (environmental variation)
from the phenotypic variation of F2 (Allard, 1960). The frequency distribution of plants,
according to their sugar level, within each generation, was done using Microsoft Excel
2000.
Spring 2002
By comparing the generation means and variances from populations P1, P2, F1, F2, BC1
and BC2, it could be possible to determine estimates of heritability, effective factor
number and components of variance involved in the sugar level (Mather and Jinks, 1982).
The adequacy of the additive-dominance model was tested using the joint scaling test
(Cavalli, 1952), which uses weighted least square estimates based on the generation
means. The generation means analysis was based on the mean and standard deviation of
PI 270248 (P1), 7833 (P2) and the respectives F1, F2, BC1 and BC2 according to Mather
and Jinks (1982). The analysis was calculated using a spreadsheet program (Ng, 1990).
This program calculates the number of effective factors (k)(Wright, 1934), and estimates
the broad-sense heritability (Mahmud and Kramer, 1951) and the narrow-sense
heritability (Warner, 1952). Finally, the frequency distribution of plants, according to
their sugar level, within each generation, was done using Microsoft Excel 2000.
Results
Inheritance of High Sugars derived from PI270248
Fall 2001
The frequency distribution of the plants of each generation is given in Figure 4-1.
Sugar values from PI270248 were higher than 7833 and did not overlap. The F1 was
39
highly skewed toward to the high sugar parent PI270248 and their values overlapped to a
considerable extent. The F2 mean was lower than that of the F1 and close to the midparent value. Broad-sense heritability (H2) was found to be approximately 0.72.
Spring 2002
The frequency distribution of the plants of each generation is given in Figure 4-2.
Sugar values from PI270248 were higher than 7833 and did not overlap. The F1 was
highly skewed toward the high sugar parent PI270248 and their values overlapped to a
considerable extent. The F2 mean was lower than that of the F1 and close to the midparent value. The BC2 mean was between that of the F1 and 7833, while the BC1 mean
was lower than that of both PI270248 and the F1.
The Joint Scaling Test Worksheet using the Chi-square goodness of fit test, showed
that the additive/dominance model was inadequate (Table 4-1). The software using the ttest, tested additive and dominance effects, and homogygote X homozygote, homozygote
X heterozygote, and heterozygote X heterozygote interactions. Additive effects and the
heterozygote X heterozygote interaction were found significant at the 5% level (Table 42). The results for narrow-sense heritability and effective factor number could not be
estimated since there was epistasis. Broad-sense heritability was 0.86.
Environmental Effects between Seasons
Analysis of variance indicated significance for season, line and their interaction at the
5% level (data not shown). Therefore, the effect of Season was analysed separately for
each Line (Table 4-3).
The sugar mean of PI270248 and F1 in spring, was significantly higher than the one in
fall. However, the sugar mean of 7833 did not differ significantly between the two
seasons.
40
PI270248
4.66(0.33)
7833
2.21(0.33)
100%
100%
80%
80%
60%
60%
40%
40%
20%
20%
0%
0%
2
2.5
3
3.5
4
4.5
5
5.5
2
2.5
sugar (g/100g f.w.)
3
3.5
4
4.5
sugar (g/100g f.w.)
X
F1
4.43(0.37)
100%
80%
60%
40%
20%
0%
2
2.5
3
3.5
4
4.5
5
5.5
5
5.5
sugar (g/100g f.w .)
F2
3.4(0.79)
100%
80%
60%
40%
20%
0%
2
2.5
3
3.5
4
4.5
sugar (g/100g f.w .)
Figure 4-1. Frequency distribution of sugar level for PI270248, 7833, F1 and F2
generations in fall 2001. The sugar mean and standard deviation for each generation is
shown above their respective diagram
5
5.5
41
PI270248
5.00 (0.25)
7833
2.17 (0.20)
100%
100%
80%
80%
60%
60%
40%
40%
20%
20%
0%
0%
2
2.5
3
3.5
4
4.5
5
5.5
2
2.5
sugars (gr/100gr f.w.)
3
3.5
4
4.5
5
5.5
sugar (gr/100gr f.w.)
X
F1
4.67 (0.36)
100%
80%
60%
40%
20%
0%
2
X
2.5
3
3.5
4
4.5
5
5.5
X
sugar (gr/100gr f.w.)
BC2
2.9 (0.55)
BC1
4.34 (0.55)
100%
100%
80%
80%
60%
60%
40%
40%
20%
20%
0%
0%
2
2.5
3
3.5
4
4.5
5
2
5.5
2.5
3
3.5
4
4.5
5
5.5
sugar (gr/100gr f.w.)
sugar (gr/100gr f.w.)
F2
3.5 (0.73)
100%
80%
60%
40%
20%
0%
2
2.5
3
3.5
4
4.5
5
5.5
sugar (gr/100gr f.w.)
Figure 4-2. Frequency distribution of sugar values for PI270248, 7833, F1, F2, BC1 and
BC2 generations in spring 2002. The sugar mean and standard deviation for each
generation is shown above their respective diagram
42
Table 4-1. Output of Joint Scaling Test Worksheet showing the failure of the
additive/dominance model
Generation Obs. Xz Expt. Xz Squared Deviation Goodness of Fit
PI270248 5.00
4.90
0.01
4.44
BC1
4.34
4.57
0.05
5.87
F1
4.67
4.23
0.20
53.29
F2
3.50
3.86
0.13
48.98
BC2
2.90
3.16
0.07
20.76
7833
2.17
2.09
0.01
5.72
Sum of Contributions
139.07
X23,0.05=7.81
z
Obs. X= Observed sugar mean, Expt. X= Expected sugar mean.
Table 4-2. Estimates of the interaction parameters z
St.error
t
Additive effect
1.415 0.028 50.92
Dominance effect 0.015 0.783 0.02
hom.X hom.
0.48 0.301 1.59
hom.X het.
0.05 0.226 0.22
het.X het.
1.55 0.502 3.19
t3,0.05=3.18
Abbreviations: St. error= Standard error, hom.X hom.= homozygote X homozygote
interaction, hom. X het.= homozygote X heterozygote interaction, het.X het.=
heterozygote X heterozygote interaction.
z
The average monthly temperature, rainfall and solar radiation for fall and spring are
shown in Table 4-4. The average daily rainfall up to 10 days before fruit harvest was
potentially important and it is summarized for both seasons in Table 4-5.
Discussion
Inheritance of High Sugars derived from PI270248
Since the F1 was strongly skewed towards the high sugar parent it appeared that there
43
Table 4-3. The effect of Season on the sugar level of PI270248, 7833 and F1 separately
where p-value indicates the significance of the season effect for each line
Sugar (g/100g
p-value
Line
Season Plant #
f.w.)
Spring
30
5.00
PI270248
0.0002
Fall
21
4.66
Spring
38
2.17
7833
0.5437
Fall
30
2.21
Spring
35
4.68
F1
0.0295
Fall
25
4.43
were largely dominance effects. However, dominance effects would also be expected to
skew the F2 sugar mean in the same direction to some extent. On the contrary, the F2
mean was almost equal to the mid-parent value in both seasons. The BC2 mean was
between the F1 and 7833 as expected. Surprisingly, the BC1 mean was lower than both
PI270248 and the F1. Dominance effects by themselves could not account for these
results, as it would be expected that BC1 was between PI270248 and the F1. Analysis
using methods described by Mather and Jinks (1982) showed that additive-dominance
model did not explain the results from PI270248, 7833, F1, F2, BC1 and BC2. Epistasis
was detected and specifically, the heterozygote X heterozygote (het X het) type of
interaction was significant. This means that there were nonallelic genes, affecting sugars
that were interacting with each other when they were found in a heterozygous condition.
The number of the interacting genes remains unknown. The significance of the het X het
interaction may help explain the facts discussed above that could not be explained by
simple dominance. This type of interaction was expected to exert its maximum effects in
the F1, where all the genes were in a heterozygous condition. The high sugar values in the
F1 may mean that the epistatic effects boost sugars well above the midparent value.
Additionally, heterozygosity is greatly reduced in BC1, BC2 and F2, especially if the
44
Table 4-4. Average monthly temperature (oC), rainfall (cm) and solar radiation (W/m2) in
fall and springz
Spring
Month
March
April
May
Avg.
Temp.
21.00
23.00
26.00
Max.
Temp.
31.71
32.11
34.33
Min.
Temp.
9.44
12.42
16.72
Tot.Rainfall
0.35
6.30
6.72
Avg. Solar
Radiation
242.91
258.14
271.32
Fall
Month
September
October
November
December
Avg.
Temp.
25.00
23.00
20.00
21.00
Max.
Temp.
34.70
32.89
29.99
28.49
Min.
Temp.
17.69
8.94
10.65
11.64
Tot.Rainfall
25.45
6.72
0
0.12
Avg. Solar
Radiation
180.11
181.01
161.43
136.51
z
Abbreviations: Avg. Temp.= Average monthly temperature, Max. Temp.= Maximum
monthly temperature, Min. Temp.= Minimum monthly temperature, Tot. Rainfall= Total
rainfall, Avg. Solar Radiation= Average solar radiation.
Table 4-5. Average daily rainfall for the 10 days before harvest for fall 2001 and spring
2002 seasons
Rainfall (cm)
Days
Fall
Spring
before
harvest
1
0
0
2
0
0
3
0
0
4
0
0
5
0.125 3.475
6
0
0.025
7
0
0.025
8
0
3.200
9
0
0
10
0
0
45
number of interacting genes is higher than two and maybe this is why the epistatic effects
are reduced in these generations.
Partition of variance into dominant, additive and all the interaction effects showed that
along with epistasis, additive effects were significant at the 5% level. This may also
provide some explanation about the fact that the F2 was very close to the mid-parent
value. The significance of additive effects also means that selection in the F2 or
subsequent generations may be quite effective, since homozygotes for genes increasing
sugars can be distinguished from heterozygotes to some extent.
The results above are in agreement with what Lower and Thompson (1967) showed
with soluble solids. Crossing two small-fruited tomato lines, they also found epistasis that
enhanced the F1 close to the high sugar parent, and significant additive effects.
Interestingly, Lower and Thompson (1967) also reported that the backcross of the high
soluble solids parent with the F1 had lower soluble solids than both of them, like in this
research. Their results could be used as a comparison to this research as it has been
shown that soluble solids have a high positive correlation with sugars (Jones and Scott,
1984; Kader et al., 1977; Malundo et al., 1995; Saliba-Colombani et al., 2001; Stevens
1972; Stevens et al., 1977a).
Unfortunately, due to epistasis, an estimate of the number of genes affecting sugars
could not be obtained. However, epistasis did not influence the estimation of broad-sense
heritability. Previous research on the broad-sense heritability of sugars or soluble solids
has shown that it can be as low as 0.13 (Conti et al., 1988). Saliba-Colombani et al.
(2001) used the cherry tomato ‘Cervil’ as a source of high sugars and showed that the
broad-sense heritability was 0.61, while Lower and Thompson (1967) showed that the
46
narrow-sense heritability of soluble solids content was 0.75 indicating that the broadsense heritability would be higher. In the present research, broad-sense heritability was
0.72 in fall and 0.86 in spring. These values show that sugar variation can be mostly
explained by genetic variation. So a successful transfer of genes affecting sugar to
tomatoes that have a low sugar level can have a significant effect on that level. The
environmental variation explained as much as 30% of the phenotypic variation (fall).
This may be enough to make selection in a breeding program difficult. Hence, there is
still need for molecular markers that, among other things, facilitate the selection.
Environmental Effects between Seasons
Seasonal effects were evident for PI270248 and the F1 genotypes expressing increased
sugars in spring (Table 4-3).
The cultural practices (irrigation, fertilization, disease and pest control) were the same
in both seasons. However, temperature, rainfall and solar radiation were factors not
controlled. Rainfall can affect sugar measurements, especially if it happens close to the
date of fruit harvest. It seems that after a heavy rainfall (~3.25cm or more) tomato fruits
absorb water, their volume increases and sugars get diluted (J.W. Scott, personal
communication). In fall, there was no rainfall for at least 40 days prior to harvest. In
spring, there were two heavy rainfalls 4 and 7 days prior to fruit harvest and they could
have lowered the sugar measurements. However, sugars in spring were higher and
rainfall cannot explain this increase.
The average monthly temperatures in both seasons were favorable for growth of
tomato plants. In fall, these temperatures ranged from 20-23oC and in spring, from 2326oC during most of the period of the fruit development. The higher temperature in
spring might have resulted in higher photosynthetic rates (increasing carbohydrate
47
production), but also in higher respiration rates (increasing carbohydrate consumption).
Taking these into account along with the fact that the temperature difference between
seasons was not great, it can be deduced that the bulk of sugar difference cannot be
explained by temperatures.
Solar radiation, in spring, was much higher than in fall, and this could account for the
sugar difference, since higher solar radiation means higher photosynthetic rates and more
carbohydrates. A positive correlation between sugar and solar radiation has already been
documented (Davies and Hobson, 1981; Forshey and Alban, 1954).
As mentioned above, the environment did not influence the sugar level of line 7833.
One possible explanation is that line 7833 reached its maximum genetic potential to
produce sugars in fall. Hence, solar radiation was not the limiting factor and its increase
in spring did not affect sugars in 7833. Another possible reason could be that, compared
to PI270248 and F1, line 7833 had larger fruits, much higher yield and lower leaf:fruit
ratio. So, even if more carbohydrates were produced in spring, they would have to get
distributed to much more fruit mass. Additionally, in PI270248 and the F1 (both
indeterminate) more carbohydrates were expected to be produced in spring since they had
higher leaf:fruit tissue ratio than the determinate 7833.
To summarize, it can be concluded that the environmental differences between
seasons in Florida can influence the concentration of sugars in tomato fruits especially for
genotypes with genetically increased sugars. Factors like temperature, rainfall and
extreme environmental events cannot be controlled and can affect sugar level either in
spring or in fall. However, solar radiation will almost always be higher in spring and this
48
will favor this season against fall in the production of more sugars, especially with high
sugar plants.
.
CHAPTER 5
RAPD MARKERS LINKED TO HIGH SUGARS FROM ACCESSION PI270248
Introduction
Consumers often are dissatisfied with the flavor of fresh market tomatoes
(Lycopersicon esculentum)(Bruhn et al., 1991). Indeed, most breeding efforts have been
dedicated to other traits such as yield, disease resistance, and traits that influence the
post-harvest life and handling of the fruit. Most commercial cultivars of fresh market
tomatoes have a low level of sugars (Total soluble solids (SS) from 4 to 5%, (Kavanagh
and McGlasson, 1983; McGlasson et al., 1983), including the two most predominant
cultivars in Florida, ‘Florida 47’ and ‘Sanibel’ (SS: 4-5 %) (Georgelis, unpublished data).
Malundo et al. (1995) found that an increase of sugar level (glucose and fructose) higher
than most large fruited commercial cultivars enhanced the aroma intensity and made the
overall flavor more acceptable within some limits of the acid level. Baldwin et al. (1998)
showed that sugars were positively correlated with overall flavor acceptability. Hence, if
the sugar level of the existing large-fruited fresh market cultivars was increased it is
possible that the overall flavor would be improved.
There are some wild species tomatoes with high fruit sugar levels, like L. chmielewskii,
L. hirsutum and L. pimpinelifolium, which could be used as potential sources of sugars.
However, some of them have sucrose as the predominant sugar with just small amounts
of glucose and fructose. So, a possible transfer of genes from these species would
probably change the type of sweetness in large-fruited tomatoes that usually have
fructose and glucose as predominant sugars. Additionally, interspecific crosses usually
49
50
entail a host of prebreeding tasks (Jones and Qualset, 1984). Another source of high
sugars could be some cherry tomatoes with elevated sugar levels. Their high sugar level
is primarily determined by glucose and fructose and a transfer of these to large fruited
tomatoes would not be expected to change the type of sweetness perceived by consumers.
In this project, the small cherry accession PI270248 was studied as a potential source of
high sugars.
Before molecular markers were developed, incorporation of high sugar levels from
cherry tomatoes into large-fruited tomatoes has been extremely difficult. High sugars
from sources other than PI270248 are under polygenic control (Causse et al., 2001;
Fulton et al., unpublished paper) and sugars also seem to be influenced by the
environment (Saliba-Colombani et al., 2001). Backcrossing these genes into large fruited
tomatoes that have a low sugar content would be difficult, as many generations would be
required and errors would be encountered in phenotypic selection. Another hindrance to
breeding programs would be the negative effects of linkage drag (McGillivary and
Clemente, 1956).
The use of molecular markers would facilitate the incorporation of genes controlling
high sugars into commercial cultivars. Molecular markers could reduce the time required
to incorporate genes into a cultivar, could improve the selection target(s), while markers
linked to high sugars, but not linked to other undesirable traits, could help with transfer of
sweetness without significantly affecting other traits. Additionally, with the use of
molecular markers, selection could be done without measuring sugars. Sugars are usually
measured by High Performance Liquid Chromatography (HPLC) analysis after they have
been extracted from tomato pulp. This is a laborious and time-consuming procedure,
51
especially if hundreds of plants have to be measured. Soluble solids are highly correlated
with sugars (Jones and Scott, 1984; Kader et al., 1977a; Malundo et al., 1995; SalibaColombani et al., 2001; Stevens, 1972; Stevens et al., 1977a) and a fair estimate of sugar
level can be obtained if soluble solids are measured. However, measuring soluble solids
of one or two fruits per plant in the field, using a hand refractometer, does not give a
reliable estimate of the actual level of soluble solids per plant (Georgelis, personal
experience). Instead, 5-10 table-ripe fruits per plant should be taken to a laboratory and
ground into pulp. Measurements based on the pulp of many fruits give a more reliable
estimate of soluble solids and consequently of sugars. Thus, measuring either sugars or
soluble solids is laborious, and molecular markers could circumvent these measurements.
There have been several cases where molecular markers (almost entirely RFLPs)
linked to sugars were found. However, most of these studies used a wild species in the
initial cross (Bernacchi et al., 1998; Tanksley et al., 1996). These markers may be useful
in an attempt to transfer sugar from a wild species to large-fruited tomatoes, but they are
not useful when the high sugar plant is the same species. Many of these markers do not
even show a single polymorphism in an intraspecific cross (Foolad et al., 1993). So, it
was decided to try to find other markers linked to high sugars from PI270248.
There are many kinds of markers one can choose from. Many of these markers have
already been used in breeding programs, including: isozymes (Feuerstein et al., 1990;
Summers et al., 1988), restriction fragment length polymorphisms (RFLP) (Osborn et al.,
1987), random amplified polymorphisms (RAPD)(Williams et al., 1990), microsatellites
(Akagi et al., 1996), sequence characterized amplified regions (SCAR) (Barret et al.,
52
1998), cleaved amplified polymorphic sequences (CAPS) (Caranta et al., 1999), and
amplified fragment length polymorphisms (AFLP) (Caranta et al., 1999).
Isozymes (Market and Miller, 1959) were among the first group of molecular
markers used for genetic diversity assessment and genetic linkage map development.
They are cost effective and co-dominant. Their basic limitations are, that much of the
genome does not code for proteins, different biochemical procedures are required to
visualise allelic differences for enzymes having different functions, and many proteins
take their final form after post-transcriptional steps that remove parts of the DNA
sequence and thus can mask variation present at the DNA level.
Restriction Fragment Length Polymorphisms (RFLPs) (Botstein et al.,1980) are codominant and quite reliable markers, and their position is known in the tomato genome.
On the other hand, they are expensive and laborious. They have been used extensively in
tomato for the construction of genetic maps (Tanksley et al., 1992) and linkage to
agronomic traits (Osborn et al., 1987).
Random amplified polymorphic DNAs (RAPDs) (Williams et al., 1990) are
sometimes unreliable, show bands of low clarity and are not co-dominant, but they are
cost effective and easy to use, especially with large populations used in breeding
programs. An additional disadvantage is that they appear to be in clusters and not evenly
distributed in the tomato genome either in interspecific (Grandillo and Tanksley, 1996) or
intraspecific crosses (Saliba-Colombani et al., 2000). However, there are some examples
of RAPDs linked to polygenic agronomic traits (Doganlar et al., 2000; Foolad and Chen,
1998; Wing et al., 1994).
53
Microsatellites (Morgante and Oliveri, 1993) require considerable investment to
generate, but are then highly polymorphic and inexpensive to use in mapping and Marker
Assisted Selection (MAS). They are highly repeatable and target hypervariable regions of
the genome. Polymorphism is usually due to differences in length of the amplified
product. They can be cost effective, but the start-up costs are large. These costs should be
justifiable for crops where large-scale mapping and MAS are a practical necessity.
Sequence Characterized Amplified Regions (SCARs)(Paran and Michelmore, 1993)
are PCR-based secondary markers that come from RAPD polymorphisms. They amplify
a single fragment with high reproducibility. Many are co-dominant and their
polymorphism can often be increased by digesting the PCR product with restriction
enzymes.
Cleaved Amplified Polymorphic Sequences (CAPSs) (secondary markers)
(Lyamichev et al., 1993) are identified with two oligonucleotide primers synthesised on
the basis of known DNA sequences. Like SCARs, they specifically amplify single
fragments. However, polymorphism of CAPSs is revealed by digestion of the amplified
DNA with several restriction endonucleases.
Finally, Amplified Fragment Length Polymorphisms (AFLPs) (Vos et al., 1995) are
usually dominant markers, need radioactive labeling, and are more laborious to work with
than RAPDs. Some of their advantages are that they need a smaller quantity of DNA than
RFLPs, they are more reliable and reproducible than RAPDs, and give more bands per
reaction than RAPDs and RFLPs.
In this research, RAPD markers were chosen because they are inexpensive and fast.
RAPDs linked to high sugars would be very convenient to use in an attempt to select for
54
high sugar phenotypes, especially when hundreds of individuals would have to be
screened.
Several screening techniques have been used to find linkage between a trait and a
marker, such as single plant comparisons, comparisons of near isogenic lines (Martin et
al., 1991; Young et al., 1988), F2 tail end comparisons (Darvasi and Soller, 1994), bulked
segregant comparisons (Michelmore et al., 1991) and recombinant inbred lines (Mohan et
al., 1994).
One major problem with RAPDs linked to high sugars, could be that some
polymorphisms might also be linked to genes controlling undesirable traits. This could be
due to tight linkage of genes controlling these traits with genes controlling sugar level, or
due to pleiotropic effects. These kinds of polymorphisms may cause problems in a MAS
used to incorporate only the trait of high sugars from PI 270248 into large-fruited
tomatoes. Traits such as plant habit (sp, sp+), pedicel type, yield, earliness and fruit size
were measured along with sugars in this research. Thus, polymorphisms linked to high
sugars and also linked to traits such as sp+, or low yield or small fruit size should be
excluded or used with caution.
Polymorphisms linked to sugar level and at least some of the traits above are likely,
because there has already been evidence about correlations between sugar level or soluble
solids with traits such as fruit size and plant habit. Emery and Munger (1970) showed that
indeterminate tomato plants produced more soluble solids in fruit than isogenic
determinate plants. MacGillinary and Clemente (1956) showed that smaller fruits had
more solids content than larger fruits in ‘San Marzano’. Also, Stevens and Rudich (1978)
showed that there was an inverse relationship between yield and soluble solids. However,
55
in this research, it is was shown that there was no significant correlation between yield
and sugars (Chapter 3).
Screening techniques that make use of bulked DNA from many plants could help us
find directly polymorphisms not linked to the traits mentioned above, by combining DNA
from plants with uniform high sugars which would be divergent in all the other traits
mentioned above. The same could be done with low sugar plants. But, up to now, it has
been shown that sugar level trait is polygenic and pooling of DNA from different plants
even with uniform high sugar level and uniform low sugar level separately, along with
the fact that RAPDs are not co-dominant could result in a loss of polymorphisms linked
to sugar level. These were the main reasons that only single plant comparisons were used,
in order to find the maximum possible number of polymorphic bands linked to sugar
level. Later, the ones that were also linked to traits such as fruit set or fruit size were
marked.
Material and Methods
Fall 2001
Accession PI270248 was crossed to Fla.7833-1-1-1 (7833) and subsequently F2 seed
were obtained. The parents,F2 , F3 and F4,were grown in a Randomized Complete Block
Design (RCBD) with 3 blocks and 10 plants per plot. However, samples were taken from
21 plants of PI270248, as some plants were lost. Seed were sown in the greenhouse in
Black Beauty spent coal (Reed Minerals Div., Highland, Ind.) medium on 27 Jul., 2001,
and transplanted into Todd planter flats (3.8-cm3 cell size)(Speedling, Sun city, Fla.) on 6
Aug., 2001. They were transplanted to the field on 30 Aug., 2001 on 20-cm-high, 81-cmwide beds of EauGallie fine sand that had been fumigated with 67% methyl bromide:
33% chroropicrin at 392 kg.ha-1 and covered with white polyethylene mulch 2 weeks
56
before transplanting. Plants were spaced 46cm apart within plots that were 91cm apart in
rows, with 152cm between rows. Recommended fertilizer and insecticide programs were
followed (Hochmuth et al., 1988). Plants were grown with stake culture and irrigated by
seepage from ditches adjacent to the six experimental beds.
Spring 2002
The parents and F2 were grown in a Randomized Complete Block Design (RCBD)
with 4 blocks with 10 plants per plot for the parents and 50 plants per plot for the F2.
However, samples were taken from 30 plants of PI270248, 38 of 7833 and 198 of F2, as
some plants were lost. Seed were sown in the greenhouse in Black Beauty spent coal
medium on 11 Jan., 2002 and transplanted into Todd planter flats on 25 Jan., 2002. They
were transplanted to the field on 6 Mar., 2002 on beds of EauGallie fine sand covered
with black polyethylene mulch 2 weeks before transplanting. Other growing procedures
and techniques were the same as in fall 2001.
Both Seasons
Plant habit was determined in the field visually and plants were separated into 2 groups
(indeterminate growth, determinate growth). Earliness and yield were subjectively
determined in the field when the earliest plants had almost 100% table-ripe fruit. For
earliness, plants were separated into 3 groups (1-3 scale, 1=very late fruit ripening,
3=very early fruit ripening). Yield was determined visually, by taking into account the
number and the size of fruits per plant. Plants were rated on a scale 1-5 (1=very low
yield, 5=very high yield). The pedicel type and fruit size were also determined in the field
when some fruit were ripe in each plant. The type of pedicel was determined visually and
plants were separated into 2 groups (jointed, jointless). For fruit size, a Craftsman caliper
was used to measure the average diameter of 3-5 table-ripe fruits per plant.
57
On 10 Dec., 2001 and 23 May, 2002, approximately 3 –15 table-ripe fruits per plant
were harvested depending on the fruit size. Fruits from each plant were ground using a
Waring blender and the fruit homogenate was stored in bags at –20oC until use. Part of
that pulp was used to measure sugars and part of it was used for measuring pH, titratable
acidity (TA) and soluble solids content (SS).
Sugar and acid were extracted using the following procedure: Fruit homogenate (40g)
was added to 70ml of 80% ethanol, boiled for 15 min (with a loose-fitting cover), cooled,
and vacuum-filtered through Whatman #4 filter paper. The resulting extract was brought
up to 100ml with 80% ethanol; and 12ml was then passed through a C-18 Sep Pak
(Waters/Millipore, Milford, Mass.) and a 0.45-µm Millipore filter. There was 5-8ml of
filtered extract per plant available for HPLC analysis.
Sugar measurement was conducted, at the USDA Citrus and Subtropical Products Lab
in Winterhaven, FL. The filtered extract was injected using a Bio-Rad AS-100 HPLC
autosampler (Bio-Rad, Richmond, Calif.), fitted with a 20-µl sample loop, into a Perkin
Elmer Series 410 HPLC system. Sugars were analyzed using a Waters Sugar Pak column
at 90C with a mobile phase of 100µM ethylenediamine-tetraacetic acid disodium-calcium
salt (CaEDTA) and a flow rate of 0.5 ml.min-1. A Perkin Elmer LC-25 Refractive Index
detector was used to measure sugars. Filtered analytical grade reagents were used for
standard preparation to establish High Performance Liquid Chromatography (HPLC)
retention times and calibration. Determination of purity of individual peaks was
accomplished by absorbance index (all wavelengths monitored simultaneously) on a
Perkin Elmer LC-235 Diode Array detector.
58
The pH, titratable acidity, and soluble solids content were measured using the
following proceduce: Tubes (50ml) were filled up with tomato pulp from each plant and
centrifuged with a Lounges Instrument Corp. centrifuge at 12000rpm for 5min. The
aliquot was filtered through filter paper (Fisher Brand) into 50ml beakers and was used to
measure pH with a Corning pH meter 340. Two to three drops of filtered aliquot were
also used to measure soluble solids content with an Atago-101 digital refractometer.
Then, 10ml of filtered aliquot were pipetted into 50ml flasks and 5-6 drops of
phenolphthalein were added. 0.1N NaOH was slowly poured into the flasks until
phenolphthalein changed from transparent to slightly purple and then the amount of
NaOH added was determined. Titratable acidity was measured using the formula:
T.A.(%Citric Acid)=(ml 0.1N NaOH) X 0.064 X 100
RAPD marker analysis. In fall 2001, DNA was taken from the parents PI270248 and
7833, and F2, F3 and F4 plants with high and low sugars. Half a rosette leaf was used in
each case and DNA was extracted according to the Klee’s protocol (see Appendix C) for
DNA extraction. PCR was conducted at Gulf Coast Research and Education Center,
Bradenton, FL. A Stratagene Robocycler 96 was used, and the conditions were the
following: 1 cycle of 94oC for 3min, 40 cycles of 94oC for 1min, 35oC for 1min, 72oC for
1.5min and 1 cycle of 72oC for 7min. The materials used per reaction (Total
volume=25µl) were: 50mM KCl, 10mM Tris-Cl (pH 9.0), 0.1% Triton-X, 2mM MgCl2,
200µM dNTPs (each), 0.2µM primer, 5-30ng DNA and 1u Taq polymerase.
Screening technique. There were 200 Operon and 800 UBC (University of British
Columbia) primers tested for polymorphisms between PI270248 and 7833. The ones that
did not yield any products or have any polymorphic bands, were excluded from further
59
analysis. The polymorphic markers were tested with five F3 plants that had very high
sugar level versus three F3 and two F4 plants that had a very low sugar levels. The low
sugar plants belonged to F3 or F4 lines that did not segregate for sugar level and their
homozygosity was expected and needed to distinguish high from low sugar plants using
dominant RAPD markers. Primers that did not seem to correlate with sugar level were
excluded from further analysis. These promising primers were tested with all 198 F2
plants from spring 2002.
Data analysis. The genetic mapping of the markers was conducted by
MAPMANAGER (QTXb15) software (Manly et al., 2001) and confirmed by
MAPMAKER 3.0 (Lander et al., 1987). QTL analysis was conducted by
MAPMANAGER (QTXb15) software. The purpose of using those programs was to find
linkages among the markers, determine which of them were linked to high sugars (in
coupling or repulsion), find how much they contribute to the phenotype, and determine
which of them were linked to other physical or chemical traits. Linkage of each marker to
high sugars was also tested using SAS (SAS Institute, Cary, N.C.). Markers linked to
high sugars were tested whether they fit a 3:1 ratio (dominant markers) in the F2 using
Chi-square Goodness of Fit test. The size of the markers was approximated using Alpha
Imager 2000 v4.03 software. Finally, the percentage of sugar variation explained by all
the markers found was estimated using SAS.
Results
Of 800 UBC primers (University of British Columbia) and 200 OPERON primers
tested with PI270248 and 7833, 148 of them did not produce any visible band. The rest
gave a total of 5964 bands (7 bands per primer) while 235 primers gave 303 polymorphic
bands between the parents. Finally, 6 primers produced 7 bands linked to high sugars in
60
the F2 population, 5 in coupling and 2 in repulsion (Table 5-1). It was observed that 2
bands given by UBC269 primer segregated as allelic in the F2 generation in spring 2002
and they fit 1:2:1 ratio (Table 5-1). Thus, the two bands were considered as one codominant marker. The rest of the markers gave bands that segregated in the F2 population
of spring 2002 fit the 3:1 ratio as expected (Table 5-1). Table (5-1) also indicates the size
of each marker. Figure (5-1) depicts the bands linked to high sugars.
Table 5-1. Primers, band size, and Chi-square goodness of fit test to a 1:2:1 or 3:1 single
gene ratio at p≤0.05
Primer
Band size
Observed
Expected
X2
(Marker)
(kbp)
No of plants No of plants statistic
OPAE4
144
148.5
0.90
1.515
(OP4)
54
49.5
52
49.5
UBC269
1.76, 1.65
0.693
48
99
(269)
98
49.5
UBC731
50
148.5
1.07
0.0067
(731)
148
49.5
UBC744
153
148.5
1.20
0.545
(744)
45
49.5
UBC489
152
148.5
0.59
2.969
(489)
46
49.5
UBC290
159
148.5
0.83
2.969
(290)
39
49.5
X21,.05=3.84, X22,.05=5.99
Three markers (OP4, 731, 489) were in one linkage group, two (744, 290) were in
another linkage group, while one (269) was unlinked (Figure 5-2).
The linkages between 744 and 290, and between 731 and 489 were significant at
p<0.001, while the linkage between OP4 and 731 was significant at p<0.01 but not at
p<0.001. The p<0.001 indicates linkage while p<0.01 is just suggestive. Hence, OP4 may
not be linked to 731 and 489.
61
OP4[c]
P1
P2
731[r]
P1
P2
269[c]-269[r]
P1
P2
744[c]
P1 P2
290[c]
P1 P2
489[c]
P1 P2
Figure 5-1. RAPD markers linked to high sugars (c=coupling, r=repulsion, P1=PI270248,
P2=7833). Arrows point to the polymorphic bands
Single marker regression (MAPMANAGER (QTXb15)) rather than interval mapping
was used to estimate the significance of the linkage between markers and other traits
measured (Table 5-2). The significance of every linkage was confirmed by SAS using
general linear model procedure (p<0.05).
62
744
OP4
26.5cM
45.3cM
290
731
34.2cM
489
Figure 5-2. Linkage groups determined by MAPMANAGER (QTXb15) and confirmed
by MAPMAKER 3.0. Kosambi map function was used
The sugar means of the plants with and without the markers are shown in Table 5-3.
The sugar mean of heterozygotes for marker 269 was almost equal to the average of the
homozygotes and this was in agreement with the additive nature of sugars shown in
Chapter 4.
Markers were grouped to find combinations that increased the sugar mean of
selected plants. Those with adequate numbers of plants to provide a reliable sample are
summarized in Table 5-4. The percentage of sugar variation in the F2 of spring 2002 that
was explained by all the markers together was 35% (data not shown).
Discussion
Six primers out of approximately 1000 that were tested, yielded six markers linked to
high sugars. four in coupling, one in repulsion and one co-dominant . Although, five of
them were found in 2 linkage groups, a QTL analysis using interval mapping was not
performed since the markers were very few and the distances among them were large
enough not to allow for a reliable placement of potential QTLs. Identification of markers
in at least 3 regions suggests polygenic control of sugars as has been reported with other
high sugar sources (Fulton et al., unpublished; Saliba-Colombani et al., 2001). However,
63
Table 5-2. Markers that were significantly linked to sugar, pH, TA, yield, SS, plant habit
and fruit sizez
Trait
Sugar
SS
TA
pH
Yield
Plant
Habit
Fruit
size
z
Marker
269
OP4
744
290
731
489
744
290
OP4
731
489
744
290
OP4
731
489
269
744
489
290
St.
6.8
42.8
37.7
6
14.7
7
54.9
7.3
77.9
28.7
14
32.6
12.3
34.9
38
50.3
20.6
9.6
16.6
10.1
%
3
18
16
4
6
3
22
5
30
12
6
14
8
15
16
20
9
4
7
7
p-value
0.0331
<0.0001
<0.0001
0.04874
0.00064
0.03039
<0.0001
0.026
<0.0001
<0.0001
0.0009
<0.0001
0.002
<0.0001
<0.0001
<0.0001
<0.0001
0.008
0.0002
0.0063
Effect
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
269
20.6
9
<0.0001
Indet.
744
290
OP4
731
489
48.3
28.5
44.7
15.4
8.1
22
18
21
8
4
<0.0001
<0.0001
<0.0001
0.00045
0.0175
+
-
Abbreviations: SS= Soluble solids content, TA= Titratable acidity, St.=
MAPMANAGER statistic, %= percentage of phenotypic variation explained by the
marker, p-value= significance of linkage of the marker to the trait, Effect= positive or
negative effect on the phenotype of a trait, Indet.= Indeterminate plant habit.
an estimate of the number of these genes would be very speculative at this point. It is
possible that there were more genes affecting sugars, but markers linked to them were not
found in our research. In a recent study, Saliba-Colombani et al. (2001) used a cross
between the cherry tomato ‘Cervil’ and the large fruited tomato ‘Levovil’. They
constructed a genetic map using mainly RFLPs with some RAPDs and approximately 4
64
Table 5-3. Sugar means of F2 individuals with or without each marker z
Marker Group Sugars (g/100g f.w.) Plants (No)
(+)
3.60
153
744(c)
(-)
3.15
45
(-)
3.70
50
731(r)
(+)
3.40
148
(A)
3.69
52
269
(B)
3.32
45
(H)
3.49
101
(+)
3.63
144
OP4(c)
(-)
3.13
54
(+)
3.56
159
290(c)
(-)
3.32
39
(+)
3.56
151
489(c)
(-)
3.33
44
z
Abbreviations: (c)= coupling, (r)= repulsion, (+)= plants with the band, (-)= plants
without the band, A= Homozygotes for the large band, B= Homozygotes for the small
band, H= Heterozygotes.
QTLs for sugar were found. One of the markers significantly linked to high sugars in that
study was OPERON primer OPAE4. This primer was also linked to high sugars in the
present experiment, where it explained 18% of the phenotypic variation of 198 F2 plants.
Another marker that explained a fair amount of variation (16%) was 744. The rest of the
markers, although significantly linked to high sugars, explained no more than 4-5% of the
phenotypic variation. Unfortunately, time did not allow for locating the chromosomal
location of the markers. Saliba-Colombani et al. (2001) showed that OP4 marker was
found on chromosome 2. In this experiment OP4 was located 45.3cM away from 731,
which was, in turn, linked to 489 (34.3cM)(Figure 5-2). Since 45.3cM is very close to
independent assortment and the linkage between OP4 and 731 was only significant at
p<0.01, it cannot be definitely concluded that 731 and 489 were located on chromosome
2. Further study is necessary. One way to find the location of the markers would be to
link the RAPD markers to RFLPs whose position is already known in tomato genome.
65
Table 5-4. Combinations of markers that allow for groups of plants with a high sugar
meanz
Sugar (g/100g
Plant
Combination
%
f.w.)
(No)
744(c)
3.850
21.00%
42
731(r)
744(c)
269
3.995
10.60%
21
OP4(c)
489(c)
744(c)
3.900
18.50%
37
731(r)
OP4(c)
269
3.970
14.50%
29
OP4(c)
489(c)
z
Abbreviations: (c)= coupling, (r)= repulsion, %= percentage of plants selected out of the
F2 population.
The markers identified could also be used in fine mapping programs in tomato. All of
them were linked apart from sugars to fruit size, yield, soluble solids, pH and titratable
acidity (Table 5-2). So, they can contribute along with other markers to find a more
precise location of QTLs affecting all these traits. This would show if some of the traits
that are correlated to each other are closely linked genes or genes with pleiotropic effects.
One of the goals of this research was to find markers useful to a Marker Assisted
Selection (MAS) program trying to transfer high sugar from the small-fruited PI270248
to large fruited tomatoes that have a low sugar level. It would be desirable that these
markers were closely linked to genes controlling high sugars and not linked to genes
controlling other undesirable traits such as small fruit size, low yield and indeterminate
plant habit. Unfortunately, two markers (OP4, 744) explaining 18% and 16%,
respectively, of the sugar variation in the F2 population were also linked to small fruit
size and explained 21% and 22% of the fruit size variation. This means that plants with
66
the markers had generally high sugars, but also small fruit size. Marker 290 explained
just 4% of sugar variation, but it was linked to small fruit size (18% of variation) and low
yield (7 %). Markers 731 and 489 were also linked to small fruit size, while the codominant marker 269, that was not linked to small fruit size, was linked to indeterminate
habit (9%). The question remains whether all these markers can be used in a MAS
program without transferring low yield, small fruit size and indeterminate habit to
cultivars destined to the market. It is recommended that 290 should be avoided. It was
either closely linked to a weak gene or loosely linked to a strong gene for high sugars
because it could not explain more than 4% of the phenotypic variation. Additionally, it
seemed to have a closer linkage to small fruit size (18%) and low yield (7%). Finally, all
the markers were linked to pH or TA or both (Table 5-2). Selecting for high sugars using
these markers would likely increase TA that is highly correlated to free acids (Paulson
and Stevens, 1974). A change in pH is not clear but it also seems to be increased by most
markers. A balance between sweetness and sourness has been shown to be important for
a good overall flavor (Jones and Scott, 1984; Malundo et al., 1995). The simultaneous
transfer of the sourness along with sweetness to a large fruited tomato may not disturb the
balance and promote, in this way, a more acceptable flavor. Sourness depends on both pH
and TA (Harvey, 1920). Thus, it is not predictable if a simultaneous transfer of sweetness
and sourness could be achieved using the markers found in this research.
All the markers together explained 35% of sugar variation in the F2. Earlier, in this
research it was shown that 86% of the sugar variation was explained by genetic variation
in spring 2002 (Chapter 4). Thus, 51% of the sugar variation in the F2 remained
unexplained by the markers. There are several possible explanations for this. First, there
67
were more genes affecting sugars, but no markers linked to them were found. Second,
some of the markers were not very closely linked to genes affecting sugars. Thirdly, there
was epistasis that enhanced sugars in the F1 (Chapter 4) that complicated the analysis of
sugar variation, and it might be a reason for part of the variation not explained by the
markers herein. Finally, 5 of 6 markers linked to high sugars were not co-dominant, thus
heterozygous plants, with non-dominant sugar control are grouped with one of the
homozygous classes increasing experimental variability.
None of the markers by itself could explain much sugar variation. Thus, combinations
of markers may be of more use in modified backcrossing in breeding programs. All the
possible combinations of markers were tested, and the ones with a high sugar average and
a reasonable population size were selected. The population size could give to breeders the
opportunity to select against small fruit size, low yield and indeterminate plant habit.
High sugars could be confirmed by measuring soluble solids, as the correlation between
the two was very high (0.92). In almost half of the best combinations of markers, 731 is
found (linked to sugars in repulsion). This marker may be more useful in breeding than
anticipated from the single marker regression results where it explains 5% of the sugar
variation. A major disadvantage of RAPDs is the fact that they are generally not codominant markers in most cases. For instance, using RFLPs (co-dominant) one can
separate a population into three discrete classes since heterozygotes can be distinguished
from the homozygotes. In a case of RAPD markers the heterozygous class is
incorporated into one of the two homozygous classes. Specifically for 731, the
heterozygotes that may have relatively high sugars were incorporated into the
homozygotes with low sugars, raising the sugar mean. Thus, the sugar mean difference
68
between the classes with and without the marker could be greatly reduced and the marker
would not be able to explain as much phenotypic variation as it could, if it was codominant. However, by selecting against these markers, one may reject some
heterozygotes with high sugars, but the plants that will remain, will be homozygous for
high sugars. The sugar mean of plants remaining after selecting against 731 was
3.7g/100g f.w. This mean is even higher than that of plants selected for OP4 and 744
(3.63, 3.6).
Another disadvantage of many RAPDs was their poor reproducibility. Often, the
results were extremely sensitive to changes in factors affecting PCR. It would be much
easier to use the markers found in this experiment in a breeding program if they were
converted to SCARs, which would be fairly reproducible. Unfortunately, time did not
allow for the conversion of the markers to SCARs, but it could be done in the future.
SCARs have much better reproducibility as longer primers are used and in many cases
they are co-dominant. They could also verify that 269 is a co-dominant marker. Our
markers could also be tested with different low sugar parents. In this research, they were
tested with 7833, but they would only be useful to breeding efforts if they worked with
different low sugar lines.
To summarize, it seems that most of the markers linked to high sugars were not linked
to low yield or indeterminate habit and this means that sugars can be uncoupled to some
extent from these traits. However, the question about whether high sugars could be
uncoupled from small fruit size remains unanswered since five out of six markers of this
research with linkage to high sugars were also linked to small fruit size. Our results do
not contradict previous research that has shown a negative correlation between sugars and
69
fruit size (McGillinary and Clemente, 1956). This correlation has also been illustrated by
co-localization of QTLs of these traits on chromosomes 2, 3, and 11 (Goldman et al.,
1995; Paterson et al., 1991; Saliba-Colombani et al., 2001). Since these co-localizations
have been found in many cases, it may be assumed that part the negative correlation
between sugars and fruit size it is due to pleiotropic effects of some genes (physiological
relationship), since other research supports pleiotropy (McGillinary and Clemente, 1956).
However, there may be genes like the one linked to the co-dominant marker 269 found in
this research, that affect sugars without an effect on fruit size. These genes may be
suitable for transferring a significant part of sweetness from small-fruited to large-fruited
tomatoes and markers like 269 may be useful. However, markers like OP4, 744, 489 and
731 should not be excluded from MAS programs, because they may help to transfer a
part of sweetness enough to enhance the overall flavor without reducing the fruit size
significantly.
APPENDIX A
PEDIGREE OF TOMATO LINES USED IN EXPERIMENTS
Table A-1. Source and pedigree of breeding lines used in experiments
Seed
Designation
Pedigree
Source
PI270248
SP 99
PI270248 'SUGAR'
[7647B X (7546B X NC13G-1)-BK-2-1]-BK-17833
990036-1
1-1-1-1
70
APPENDIX B
KLEE’S PROTOCOL FOR DNA EXTRACTION FROM TOMATO LEAVES
1. Grind up leaf tissue in eppendorf tube.
2. Add 500ul Extraction Buffer.
3. Spin 5 min. at high speed in micro-centrifuge.
4. Transfer 350ul of supernatant to a fresh eppendorf containing 350ul isopropanol.
5. Mix by inversion and spin 10 min. at top speed in micro-centrifuge.
6. Pour off liquid and dry pellet by letting it sit upside down on a paper towel.
7. Once the tube is dry add 400ul TE and resuspend by vortexing and shaking at
room temperature for 30 min.
8. Use 2ul of prep per PCR reaction.
Extraction Buffer
0.2 M Tris-HCl, pH 9.0
0.4 M LiCl
25mM EDTA
1% SDS
71
LIST OF REFERENCES
Adams, P., 1991. Effects of increasing the salinity of the nutrient solution with major
nutrients or sodium chloride on the yield, quality and composition of tomatoes grown in
rockwool. Journal of Horticultural Science 66: 201-207.
Akagi, H., Yokozeki, Y., Inagaki, A., Nakamura, A., and Fujimura, T., 1996. A
codominant DNA marker closely linked to the rice nuclear restorer gene, Rf-1, identified
with inter-SSR fingerprinting. Genome 39: 1205-1209.
Allard, R.W., 1960. Principles of plant breeding. J. Wiley, New York.
Amoore, J.E., 1952. The stereochemical specificities of human oldfactory receptors.
Perfumer Essential Oils 43: 321-323.
Anderson, R.E., 1957. Factors affecting the acidic constituents of the tomato. Ph.D.
dissertation, University of Illinois.
Baldwin, E.A., Scott, J.W., Einstein, M.A., Malundo, T.M.M., Carr, B.T., Shewfelt, R.L.,
and Tandon, K.S., 1998. Relationship between sensory and instumental analysis for
tomato flavor. J. Am. Soc. Hort. Sci. 125: 906-915.
Baldwin, E.A., Scott, J.W., Shewmaker, C.K., and Schuch, W., 2000. Flavor trivia and
tomato aroma: biochemistry and possible mechanisms for control of important aroma
components. HortScience. 35: 1013-1022.
Barret, P., Deloume, R., Foisset, N., and Renard, M., 1998. Development of a SCAR
(sequence characterized amplified region) marker for molecular tagging of the dwarf
BREIZH (Bzh) gene in Brassica napus L. Theor. Appl. Genet. 97: 828-833.
Bernacchi, D., Beck-Bunn, T., Eshed, Y., Lopez, J., Petiard, V., Uhlig, J., Zamir, D., and
Tanksley, S., 1998. Advanced backcross QTL analysis in tomato. I. Identification of
QTLs for traits of agronomic importance from Lycopersicon hirsutum. Theor. Appl.
Genet. 97: 381-397.
Betancourt, L.A., Stevens, M.A., and Kader, A.A., 1977. Accumulation and loss of
sugars and reduced ascorbic acid in attached and detached tomato fruits. J. Amer. Soc.
Hort. Sci. 102: 721-723.
Biester, A.M., 1925. Carbohydrate studies. I. Relative sweetness of pure sugars. Am. J.
Physiol. 73: 387-400.
72
73
Botstein, D., White, R., Skolnick, M., and Davis, R., 1980. Construction of a genetic
linkage map in man using restriction fragment length polymorphisms. American Journal
of Human Genetics. 32: 314-331.
Brauss, M.S., Linforth, R.S., and Taylor, A.J., 1998. Effect of variety, time of eating and
fruit-to-fruit variation on volatile release during eating of tomato fruits (Lycopersicon
esculentum). J. Agric. Food Chem. 46: 2287-2292.
Bruhn, C.M., Feldman, N., Garlitz, C., Harwood, J., Ivans, E., Marshall, M., Riley, A.,
Thurber, D., and Williamson, E., 1991. Consumer perceptions of quality: apricots,
cantaloupes, peaches, pears, strawberries, and tomatoes. J. Food Qual. 14: 187-195.
Buttery, R.G., 1993. Quantitative and sensory aspects of flavor of tomato and other
vegetables and fruits, p.259-286. In: T.E. Acree and R. Teranishi (eds.). Flavor science:
sensible principles and techniques. Amer. Chem. Soc., Washington, D.C.
Buttery, R.G., and Ling, L.C., 1993a. Enzymatic production of volatiles in tomatoes, p.
137-146. In: P. Scheier and P. Winterhalter (eds.). Flavor precursors. Allured Publ.,
Wheaton, III.
Buttery, R.G., and Ling, L.C., 1993b. Volatile components of tomato fruit and plant
parts: Relationship and biogenesis, p. 23-34. In: Teranishi, R., Buttery, R.G., Sugisawa,
H. (Eds.), Bioactive volatile compounds from plants: ACS Symposium Series No. 525.
American Chemical Society, Washington, DC.
Buttery, R.G., Seifert, R.M., Guadagni, D.G., and Ling, L.C., 1971. Characterization of
additional volatile components of tomato. J. Agr. Food Chem. 19: 524-529.
Buttery, R.G., Teranishi, R., Flath, R.A., and Ling, L.C., 1989. Fresh tomato aroma
volatiles-Composition and sensory studies, p. 213-222. In: Teranishi, R., Buttery, R.G.,
Shahidi, F. (Eds.), Flavor chemistry: trends and developments: ACS Symposium Series
No. 388. American Chemical Society, Washington, DC.
Caranta, C., Thabuis, A., and Palloix, A., 1999. Development of a CAPS marker for the
Pvr4 locus: a tool for pyramiding potyvirus resistance genes in pepper. Genome 42:
1111-1116.
Causse, M., Saliba-Colombani, V., Lesschaeve, I., and Buret, M., 2001. Genetic analysis
of organoleptic quality in fresh market tomato. 2. Mapping QTLs for sensory attributes.
Theor. Appl. Genet. 102: 273-283.
Cavalli, L.L., 1952. An analysis of linkage of quantitative inheritance, p. 135-144. In
Quantitative inheritance, E.C.R. Reeve and C.H. Waddington (eds.), HMSO, London.
74
Chetelat, R.T., De Verna, J.W., and Bennett, A.B., 1995. Introgression into tomato
(Lycopersicon esculentum) of the Lycopersicon chmielewskii sucrose accumulator gene
(sucr) controlling fruit sugar composition. Theor. Appl. Genet. 91: 327-33.
Conti, S., Sanguineti M.C., Toni, B., and Azzoni, A., 1988. Inheritance of quality traits in
processing tomato (Lycopersicon esculentum Mill.). Euphytica 37: 121-127.
Darvasi, A., and Soller, M., 1994. Selective DNA pooling for determination of linkage
between a molecular marker and a quantitative trait locus. Genetics 138: 1365-1373.
Davies, J.N., and Hobson, G.E., 1981. The constituents of tomato fruit-The influence of
environment, nutrition, and genotype.CRC Critical Reviews in Food Science and
Nutrition 15: 205-280.
Davies, J.N., and Kempton, R.J., 1975. Changes in the individual sugars of tomato fruit
during ripening. J. Sci. Fd. Agric. 26: 1103-1110.
De Rovira, D., 1997. Manual-Flavor nomenclature workshop: An odor description and
sensory evaluation workshop. Flavor Dynamics, Somerset, N.J.
Doganlar, S., Tanksley, S.D., and Mutschler, M.A., 2000. Identification and molecular
mapping of loci controlling fruit ripening time in tomato. Theoretical and applied
genetics 100: 249-255.
Emery, G.C., and Munger, H.M., 1970. Effects of inherited differences in growth habit
on fruit size and soluble solids in tomato. J. Amer. Soc. Hort. Sci. 95: 410-412.
Feuerstein, U., Brown, A.H.D., and Burdon, J.J., 1990. Linkage of rust resistance genes
from wild barley (Hordeum spontaneum) with isozyme markers. Plant Breeding 104:
318-324.
Foolad, M.R., and Chen, F.Q., 1998. RAPD markers associated with salt tolerance in an
interspecific cross of tomato (Lycopersicon esculentum X L. pennellii). Plant Cell
Reports 17: 306-312.
Foolad, M.R., Jones, R.A., Rodriguez, R.L., 1993. RAPD markers for constructing
intraspecific tomato genetic maps. Plant Cell Reports 12: 293-297.
Forshey, C.G. and Alban, E.K., 1954. Seasonal quality changes in greenhouse tomatoes.
Proc. Am. Soc. Hortic. Sci. 64: 372.
Fulton, T.M., Bucheli, P., Voirol, E., Lopez, J., Petiard, V., and Tanksley, S.D., 2001.
Quantitative Trait Loci (QTL) affecting sugars, organic acids and other biochemical
properties possibly contributing to flavor, identified in four advanced backcross
populations of tomato (unpublished), Dept. of Plant Breeding, Cornell University, Ithaca,
NY.
75
Goldman, I.L., Paran, I., and Zamir, D., 1995. Quantitative trait locus analysis of a
recombinant inbred line population derived from a Lycopersicon esculentum X
Lycopersicon cheesmanii cross. Theor. Appl. Genet. 90: 25-932.
Gough, C., and Hobson, G.E., 1990. A comparison of productivity, quality, shelf-life
characteristics and consumer reaction to the crop from cherry tomato plants grown at
different levels of salinity. Journal of Horticultural Science. 65: 431-439.
Grandillo, S., and Tanksley, S.D., 1996. Genetic analysis of RFLPs, GATA
microsatellites and RAPDs in a cross between L. esculentum and L. pimpinellifolium.
Theor. Appl. Genet. 92: 957-965.
Hadas, R., Schaffer, A.A., Miron, D., Fogelman, M., and Granot, D., 1995. PCRgenerated molecular markers for the invertase gene and sucrose accumulation in tomato.
Theor. Appl. Genet. 90: 1142-1148.
Harvey, R.B., 1920. The relation between the total acidity, the concentration of the
hydrogen ion and the taste of the acid solutions. J. Am. Chem. Soc. 42: 712-714.
Hochmuth, G.J., Maynard, D.N., and Sherman, M., 1988. Tomato production guide for
Florida, p.22. Florida Cooperative Extension Circular 98C.
Husain, S.E., James, C., Shields, R., and Foyer, C.H., 2001. Manipulation of fruit sugar
composition but not content in Lycopersicon esculentum fruit by introgression of an acid
invertase gene from Lycopersicon pimpinellifolium. New Phytologist 150: 65-72.
Jones, R.A., and Qualset, C.O., 1984. Breeding crops for environmental stress tolerance,
p. 305-340. In: Collins, G.B., Petolino, J.F. (eds) Application of genetic engineering to
crop improvement, Nijihoff/Junk, Holland.
Jones, R.A. and Scott, S.J., 1984. Genetic potential to improve tomato flavor in
commercial F1 hybrids. J. Amer. Soc. Hort. Sci. 109: 318-321.
Kader, A.A., Stevens, M.A., Albright-Holton, M., Morris L.L., and Agazi, M., 1977.
Effect of fruit ripeness when picked on flavor and composition in fresh market tomatoes.
J. Amer. Soc. Hort. Sci. 102: 724-731.
Kavanagh, E.E., and McGlasson, W.B., 1983. Determination of sensory quality in fresh
market tomatoes. CSIRO Fd. Res. Q. 43: 81-89.
Kazeniac, S.J., and Hall, R.M., 1970. Flavor chemistry of tomato volatiles. J. Food. Sci.
35: 519-530.
76
Klann, E.M., Hall, B., and Bennett, A.B., 1996. Antisense Acid invertase (TIV1) gene
alters soluble sugar composition and size in transgenic tomato fruit. Plant Physiol. 112:
1321-1330.
Krumbein, A., and Auerswald, H., 1998. Characterization of aroma volatiles in tomatoes
by sensory analysis. Nahrung 42: 395-399.
Lander, E.S., Green, P., Abrahamson, J., Barlow, A., Daly, M.J., Lincoln, S.E., and
Newburg, L., 1987. MAPMAKER: an interactive computer package for constructing
primary genetic linkage maps of experimental and natural populations. Genomics 1: 17481.
Levin, I., Gilboa, N., Yeselson, E., Shen, S., and Schaffer, A.A., 2000. Fgr, a major locus
that modulates the fructose to glucose ratio in mature tomato fruits. Theor. Appl. Genet.
100: 256-262.
Lower, R.L. and Thompson, A.E., 1967. Inheritance of acidity and solids content of
small-fruited tomatoes. Proc. Am. Soc. Hortic. Sci. 91: 486-494.
Lyamichev, V., Brow M.A.D., and Dahlberg, J.E., 1993. Structure-specific
endonucleolytic cleavage of nucleic acids by eubacterial DNA polymerases. Science
260: 778-783.
Mahmud, I., and Kramer, H.H., 1951. Segregation for yield, height and maturity
following a soybean cross. Agron. J. 44: 605-608.
Malundo, T.M.M., Shewfelt, R.L., and Scott, J.W., 1995. Flavor quality of fresh market
tomato (Lycopersicon esculentumMill.) as affected by sugar and acid levels. Postharv.
Biol. Technol. 6: 103-110.
Marc-Andre D’Aoust, Serge Yelle, and Binh Nguyen-Quoc, 1999. Antisense inhibition
of tomato fruit Sucrose synthase decreases fruit setting and the sucrose unloading
capacity of young fruit. The Plant Cell 11: 2407-2418.
Manly, K.F., Cudmore, Jr, R.H., and Meer J.M., 2001.Map Manager QTX, crossplatform software for genetic mapping. Mamm. Genome 12: 930-932.
Market, C.L, and Miller, F., 1959. Multiple forms of enzymes: Tissue ontogenic and
species specific patterns. Proc. Natl. Acad. Sci. 45: 753-763.
Martin, G.B., Williams, J.G., and Tanksley, S.D., 1991. Rapid identification of markers
linked to a Pseudomonas resistance gene in tomato by using random primers and near
isogenic lines. Proc. Natl. Acad. Sci. 88: 2336-2340.
Mather, K., and Jinks, J.L., 1982. Biometrical Genetics. Third Ed. Chapman and Hall,
London.
77
McGillivary, J.H., and Clemente, L.J., 1956. Effect of tomato size on solids content.
Proc. Am. Soc. Hprt. Sci. 68: 466-469.
McGlasson, W.B., Sumeghy, J.B., Morris, L.L., McBride, R.L., Best, D.J., and
Tigchelaar, E.C., 1983. Yield and evaluation of F1 tomato hybrids incorporating the nonripening nor gene. Aust. J. Exp. Agric. Anim. Husb. 23: 106-112.
Michelmore, R.W., Paran, I., and Kesseli, R.V., 1991. Identification of markers linked to
disease resistance genes by bulked segregant analysis: a rapid method to detect markers
in specific genomic regions by using segregating populations. Proc. Natl. Acad. Sci. 88:
9828-9832.
Mohan M., Nair, S., Bentur, J.S., Prasada Rao, U., and Bennett, J., 1994. RFLP and
RAPD mapping of the rice Gm2 gene that confers resistance to biotype 1 of gall midge
(Orseolia oryzae). Theor. Appl. Genet. 87: 782-788.
Morgante, M., and Oliveri, M., 1993. PCR amplified microsatellites as markers in plant
genetics. Plant Journal 3: 175-182.
Ng, T.J., 1990. Generation means analysis by microcomputer. Hortscience 25: 363.
Osborn, T.C., Alexander, D.C., and Fobes, J.F., 1987. Identification of restriction
fragment length polymorphisms linked to genes controlling soluble solids content in
tomato fruit. Theor. Appl. Genet. 73: 350-356.
Paran, I., and Michelmore, R.W., 1993. Development of reliable PCR-based markers
linked to downy mildew resistance genes in lettuce. Theor. Appl. Genet. 85: 985-993.
Paterson, A.H., Damon, S., Hewitt, J.D., Zamir, D., Rabinowitch, H.D., Lincoln, S.E.,
Lander, E.S., and Tanksley, S.D., 1991. Mendelian factors underlying quantitative traits
in tomato: comparison across species, generations, and environments. Genetics 127: 181197.
Paulson, K.N., and Stevens, M.A., 1974. Relationships among titratable acidity, pH and
buffer composition of tomato fruits. Journal of Food Science 39: 254-357.
Petro-Turza, M., 1987. Flavor of tomato and tomato products. Food Rev. Int. 2: 309-351.
Picha, D.H., 1986. Effect of harvest maturity on the final fruit composition of cherry and
large-fruited tomato cultivars. J. Amer. Soc. Hort. Sci. 111: 723-727.
Saliba-Colombani, V., Causse, M., Gervais, L., and Philouze, J., 2000. Efficiency of
RFLP, RAPD, and AFLP markers for the construction of an intraspecific map of the
tomato genome. Genome 43: 29-40.
78
Saliba-Colombani, V., Causse, M., Langlois, D., Philouze, J., and Buret, M., 2001.
Genetic analysis of organoleptic quality in fresh market tomato. 1. Mapping QTLs
physical and chemical traits. Theor. Appl. Genet. 102: 259-272.
Schaffer, A.A., Petreikov, M., Miron, D., Fogelman, M., Spiegelman, M., Bnei-Moshe,
Z., Shen, S., Granot, D., Hadas, R., Dai, N., Levin, I., Bar M., Friedman, M., Pilowsky,
M., Gilboa, N., and Chen, L., 1998. Modification of carbohydrate content in developing
tomato fruit. Hortscience 34: 12-15.
Scott, J.W., 2002. A breeder’s perspective on the use of molecular techniques for
improving fruit quality. HortScience 37: 464-467.
Scott, J.W., Somodi, G.C., and Jones, J.B., 1989. Resistance to bacterial spot fruit
infection in tomato. Hortscience 24: 825-827.
Shewfelt, R.L., 1993. Measuring quality and maturity, p. 100-124. In: R.L. Shewfelt and
S.E. Prussia (eds.). Postharvest handling: A systems approach, Academic Press, New
York.
Stevens, M.A., 1972. Relationships between components contributing to quality variation
among tomato lines. J. Am. Soc. Hortic. Sci. 97: 70-73.
Stevens, M.A., and Kader, A.A., 1976. Varietal and postharvest effects on tomato fruit
composition and flavor. Proc. Second Tomato Quality Workshop, July 1976, Univ. of
Calif., Davis, Veg. Crops Series, No. 178, 108.
Stevens, M.A., Kader, A.A., Albright-Holton, M., Algazi, M., 1977a. Genotypic variation
for flavor and composition in fresh market tomatoes. J. Amer. Soc. Hort. Sci. 102, 680689.
Stevens, M.A., Kader, A.A., Albright-Holton, M., 1977b. Intercultivar variation in
composition of locular and pericarp portions of fresh market tomatoes. J. Amer. Soc.
Hort. Sci., 102,689-692.
Stevens, M.A., Kader, A.A., Albright-Holton, M., 1979. Potential for increasing tomato
flavor via increased sugar and acid content. J. Amer. Soc. Hort. Sci., 104, 40-42.
Stevens, M.A., Rudich, J., 1978. Genetic potential for overcoming physiological
limitations on adaptability, yield, and quality in the tomato. Hortscience, 13, 673-678.
Stommel, J.R., 1992. Enzymatic components of sucrose accumulation in the wild tomato
species Lycopersicon peruvianum. Plant Physiol., 99, 324-328.
79
Summers, R.W., Koebner, R.M.D., Liollins, T.W., Forster, J., MaCartney, D.P., 1988.
The use of an isozyme marker in breeding wheat (Triticum aestivum) resistant to the
eyespot pathogen (Pseudocercosporella herpotrichoides), p. 1195-1197. Proceedings of
the Seventh International Wheat Genetics Symposium/ edited by T.E. Miller and R.M.D.
Koebner.
Tandon, K.S., Baldwin, E.A., Shewfelt, R.L., 2000. Aroma perception of individual
volatile compounds in fresh tomatoes (Lycopersicon esculentum, Mill) as affected by the
medium of evaluation. Postharvest Biology and Technology, 20, 261-268.
Tanksley, S.D., Ganal, M.W., Prince, J.P., de Vicente, M.C., Bonierbale, M.W., Broun,
P., Fulton, T.M., Giovannoni, J.J., Grandillo, S., Martin, G.B., Messeguer, R., Miller,
J.C., Miller, L., Paterson, A.H., Pineda,O., Roder,M.S., Wing, R.A., Wu, W., Young,
N.D., 1992. High density molecular linkage maps of the tomato and potato genomes.
Genetics 132, 1141-1160.
Tanksley, S.D., Grandillo, S., Fulton, T.M., Zamir, D., Eshed, Y., Petiard, V., Lopez, J.,
and Beck-Bunn, T., 1996. Advanced backcross QTL analysis in a cross between an elite
processing line of tomato and its wild relative L. pimpinellifolium. Theor. Appl. Genet.,
1996, 213-224.
United States Department of Agriculture (USDA) Ag Statistics Board Vg 1-2 (01). Jan.
2001.
Vickers, A., 1977. Structural and mechanical indicators of flavor quality, p. 45-50. In: R.
Scanlan (ed.). Flavor quality: objective measurement. ACS Symp. Ser. 51. Amer. Chem.
Soc., Washington, D.C.
Voirol, E., Daget, N., 1987. Nasal and retronasal oldfactory perception of a meat aroma.
P. 309-316. In: M. Martens, G.A. Dalen, and H. Russwurm (eds.). Flavour science and
technology. Wiley, New York.
Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T., Hornes, M., Frijters, A.,
Pot, J., Peleman, J., Kuiper, M., 1995. AFLP: a new technique for DNA fingerprinting.
Nucleic Acids Res., 23, 4407-4414.
Warner, J.N., 1952. A method for estimating narrow-sense heritability. Agron. J., 44,
427-430.
Warner, J.N., 1952. A method for estimating narrow-sense heritability. Agron. J., 44,
427-430.
Wehrhahn, C., Allard, R.W., 1965. The detection and measurement of the effects of
individual genes involved in the inheritance of quantitative character in wheat. Genetics,
51, 109-119.
80
Williams, S., Kubelik, A., Livak, K., Rafalski, J., Tingey, S., 1990. DNA polymorphisms
amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Research.,
18, 6531-6535.
Wing, R.A., Zhang, H.B., Tanksley, S.D., 1994. Map-based cloning in crop plants.
Tomato as a model system. I. Genetic and physical mapping of jointless. Mol. Gen.
Genet., 242, 681-688.
Wright, S., 1934. The results of crosses between inbred strains of guinea pigs differing in
number of digits. Genetics, 19, 537-551.
Yelle, S., Chetelat, R.T., Dorais, M., DeVerna, J.W., and Bennett, A.B., 1991. Sink
metabolism in tomato fruit. IV. Genetic and biochemical analysis of sucrose
accumulation. Plant Physiol., 95, 1026-1035.
Young, N.D., Zamir, D., Ganal, M.W., Tanksley, S.D., 1988. Use of isogenic lines and
simultaneous probing to identify DNA markers tightly linked to the Tm-2 gene in tomato.
Genetics, 120, 579-585.
BIOGRAPHICAL SKETCH
I was born in Lesvos, a Greek island, in 1977. Influenced by my father’s occupation
(tomato grower) I decided to pursue a B.S. degree in horticultural sciences, at Aristotle
University of Thessaloniki, in 1995. Being thrilled by recent advances in genetics, I
decided to pursue an M.S. degree in plant genetics in the US. I started my graduate
studies in August 2000, one month after I graduated from Aristotle University.
81