Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
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