<%BANNER%>

Color and Carotenoid Content in Squash (Cucurbita spp.)

Permanent Link: http://ufdc.ufl.edu/UFE0042511/00001

Material Information

Title: Color and Carotenoid Content in Squash (Cucurbita spp.)
Physical Description: 1 online resource (160 p.)
Language: english
Creator: Itle, Rachel
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: carotenoids, color, correlation, cucurbita, genetic, heritability, hrm, qtl, rapds, snps, squash, ssrs
Horticultural Science -- Dissertations, Academic -- UF
Genre: Horticultural Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Carotenoids serve as a cancer preventative and as antioxidants that protect against degenerative health problems such as cardiovascular disease, cataract formation, and macular degeneration in older adults. Carotenoids with high importance in human health are those that serve as precursors to vitamin A (pro-vitamin A carotenoids), and function in normal vision, bone growth, cell division and differentiation, and reproduction. Squash (Cucurbita spp.) is an excellent dietary source of carotenoids. The diversity and range of carotenoid types and concentrations within squash provide a means to increase the nutritional value through plant breeding. The objectives of this research were to: 1) determine if the carotenoids of squash could be correlated with colorimetric analysis using the CIE L*a*b* color value system, 2) determine heritability and gene action of flesh color in nine Cucurbita populations derived from three crosses, 3) create a genetic linkage map in C. moschata (2n=2x=40) with sufficient marker density for quantitative trait loci (QTL) studies to map genomic regions associated with color, and 4) design a protocol to recover monomorphic simple sequence repeat (SSR) markers using high resolution melting (HRM) curve analysis. Strong correlations were found between color value a* and total carotenoids (r = 0.91) and color value b* and chroma with lutein (r = 0.87). Broad-sense heritabilities ranged from 0.19 to 0.82 for L*, 0.12 to 0.32 for a*, 0.40 to 0.93 for b*, 0.36 to 0.92 for chroma, and 0.14 to 0.15 for hue across all three crosses. Significant combinations of additive and dominance gene effects were identified for color space values L* and hue (P ? 0.05). A linkage map was constructed using random amplified polymorphic DNA (RAPD) and SSR markers. QTL associations were examined using composite interval mapping in MapQTL 5.0. Twenty-one linkage groups were obtained with QTLs detected on LG7 and 18 for L*, LG3, LG10, LG18 for a*, LG 5 and LG17 for b* and chroma, and LG 3, LG13, LG14, LG15, and LG19 for hue. Four of the ten monomorphic markers tested, M042, M120, M009, P098 were able to be recovered using HRM for addition to the linkage map data set.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Rachel Itle.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Olmstead, James.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042511:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042511/00001

Material Information

Title: Color and Carotenoid Content in Squash (Cucurbita spp.)
Physical Description: 1 online resource (160 p.)
Language: english
Creator: Itle, Rachel
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: carotenoids, color, correlation, cucurbita, genetic, heritability, hrm, qtl, rapds, snps, squash, ssrs
Horticultural Science -- Dissertations, Academic -- UF
Genre: Horticultural Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Carotenoids serve as a cancer preventative and as antioxidants that protect against degenerative health problems such as cardiovascular disease, cataract formation, and macular degeneration in older adults. Carotenoids with high importance in human health are those that serve as precursors to vitamin A (pro-vitamin A carotenoids), and function in normal vision, bone growth, cell division and differentiation, and reproduction. Squash (Cucurbita spp.) is an excellent dietary source of carotenoids. The diversity and range of carotenoid types and concentrations within squash provide a means to increase the nutritional value through plant breeding. The objectives of this research were to: 1) determine if the carotenoids of squash could be correlated with colorimetric analysis using the CIE L*a*b* color value system, 2) determine heritability and gene action of flesh color in nine Cucurbita populations derived from three crosses, 3) create a genetic linkage map in C. moschata (2n=2x=40) with sufficient marker density for quantitative trait loci (QTL) studies to map genomic regions associated with color, and 4) design a protocol to recover monomorphic simple sequence repeat (SSR) markers using high resolution melting (HRM) curve analysis. Strong correlations were found between color value a* and total carotenoids (r = 0.91) and color value b* and chroma with lutein (r = 0.87). Broad-sense heritabilities ranged from 0.19 to 0.82 for L*, 0.12 to 0.32 for a*, 0.40 to 0.93 for b*, 0.36 to 0.92 for chroma, and 0.14 to 0.15 for hue across all three crosses. Significant combinations of additive and dominance gene effects were identified for color space values L* and hue (P ? 0.05). A linkage map was constructed using random amplified polymorphic DNA (RAPD) and SSR markers. QTL associations were examined using composite interval mapping in MapQTL 5.0. Twenty-one linkage groups were obtained with QTLs detected on LG7 and 18 for L*, LG3, LG10, LG18 for a*, LG 5 and LG17 for b* and chroma, and LG 3, LG13, LG14, LG15, and LG19 for hue. Four of the ten monomorphic markers tested, M042, M120, M009, P098 were able to be recovered using HRM for addition to the linkage map data set.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Rachel Itle.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Olmstead, James.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042511:00001


This item has the following downloads:


Full Text

PAGE 1

1 COLOR AND CAROTENOID CONTENT IN SQUASH ( C ucurbita spp ) By RACHEL ANN ITLE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Rachel Ann Itle

PAGE 3

3 To my grandparents, Edwin and Marie Paulone, wh o were unable to pursue advanced degrees and sacrificed in countless ways to give future generations of their family the opportunity To my parents, Gary and Rebecca Itle, who never stopped believing in me and who never let me stop believing in myself

PAGE 4

4 ACKNOWLEDGMENTS I thank my past committee chair, Dr. Eileen Kabelka for her guidance throughout my degree program I thank my curren t committee chair, Dr. James Olmstead for his patience, advice and chapter revisions throughout the last stage of this process I thank my committee members Dr s Paul Lyrene, Gloria Moore, and Amy Simonne for their thoughtful commentaries and mentorship t hroughout the course of my degree. I also thank Dr s C. Eduardo Vallejos and Jos Chaparro for taking time to give direction with data analysis It has been a privilege to learn from them. I also thank Dr Les Padley, Kristen Young, Lauren Coleman and H Patricia Rodr guez for their assistance in field, greenhouse and laboratory work and making the world of cucurbits enjoyable. Without the unconditional love and relentless encouragement of those closest to me, I simply would not have completed this jou rney. I thank my parents, Gary and Rebecca Itle, who helped me find strength and motivation when I had none. They sacrificed to provide me with an education and encouraged me find a career for which I am passionate. I thank my sister, Rene Minick, for her constant support and uplifting words. I thank my beautiful nieces, Grace and Elise Minick, who have given me more simple joys than they know. I thank Daro Chvez for his love, friendship, and unending support. He shared this experience with me and hel d my hand along the way. The love and gratitude I have for all of them extend far past the words on this page. Above all others, I am most thankful to my Lord Jesus Christ. He blessed me with the ability to complete this degree and carried me through this process He surrounded me with an academic support system, friends and family while teaching me strength endurance and humility. Without Him, none of this would have been possible. Glory to God alone!

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LI ST OF ABBREVIATIONS ................................ ................................ ........................... 13 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 The Cucurbitaceae Family ................................ ................................ ...................... 16 Economic Importance of Pumpkins, Squash and Gourds ................................ ....... 17 Nutrition of Squash ................................ ................................ ................................ 18 Importance of Carotenoids and Vitamin A in Human Health ............................ 19 Recommended Dietary Allowance (RDA) for Carotenoids and Vitamin A ........ 20 Increasing Pro Vitamin A Rather than Pre formed Vitamin A Content ............. 21 Analysis of Squash Fruit ................................ ................................ ......................... 22 Carotenoid Analysis ................................ ................................ ......................... 22 Color Analysis ................................ ................................ ................................ .. 23 Gene Number and Inheritance of Color in Cucurbitaceae ................................ ...... 25 Molecular Tools for Increasing Color Content of Squash ................................ ........ 26 Molecular Markers Used in Cucurbitaceae ................................ ....................... 26 Genetic Linkage Maps Available in the Cucurbita genus ................................ 28 Quantitative Trait Loci (QTL) Detection in Cucurbitaceae ................................ 29 Research Rationale ................................ ................................ ................................ 30 2 CORRELATION BETWEEN L* A* B* COLOR SPACE VALUES AND CAROTENOID CONTENT IN PUMPKINS AND SQUASH ( Cucurbita spp.) .......... 31 Introduction ................................ ................................ ................................ ............. 31 Materials and Methods ................................ ................................ ............................ 34 Plant Material ................................ ................................ ................................ ... 34 Field Trials ................................ ................................ ................................ ........ 34 Colorimetric and HPLC analysis ................................ ................................ 35 Data analysis ................................ ................................ ............................. 36 Results and Discussion ................................ ................................ ........................... 37 Colorimetric Evaluation ................................ ................................ ..................... 37 Carotenoid Content ................................ ................................ .......................... 39 Correlations between Color Space Values and Carotenoids ............................ 42

PAGE 6

6 Conclusion ................................ ................................ ................................ .............. 44 3 HERITABILITY ESTIMATES OF L* A* B* COLOR SPACE VALUES IN WINTER SQUASH ( Cucurbita spp.) ................................ ................................ ....... 52 Introduction ................................ ................................ ................................ ............. 52 Materials and Methods ................................ ................................ ............................ 54 Plant Material ................................ ................................ ................................ ... 54 Field Trials ................................ ................................ ................................ ........ 54 Colorimetric Analysis ................................ ................................ ........................ 55 Variance, Heritability and Standard Error Estimations by Generation Means Analysis ................................ ................................ ................................ ......... 56 Epistatic Interaction Estimates ................................ ................................ ......... 57 Data Analysis ................................ ................................ ................................ ... 58 Results and Discussion ................................ ................................ ........................... 59 Analysis of variance ................................ ................................ ......................... 59 C. pepo ) populations ............................ 59 C. pepo ) populations ............................. 60 C. moschata ) populations ................... 60 L* a* b* Color Space Means and Ranges ................................ ........................ 60 C. pepo ) populations ............................ 60 C. pepo ) populations ............................. 61 C. moschata ) populations ................... 62 Rationale for Selecting Single Fruit Color Space Value Measurements ........... 63 Broad and Narr ow Sense Heritability Estimates ................................ ............... 64 C. pepo ) populations ............................ 65 C. pepo ) populations ............................. 65 C. moschata ) populations ................... 66 Epistatic Gene Effect Estimates ................................ ................................ ....... 67 C. pepo ) populations ............................ 67 C. pepo ) populations ............................. 68 C. moschata ) populations ................... 69 Breeding for Increased Flesh Color in Winter Squash ................................ ...... 71 4 CONSTRUCTION OF A RAPD AND SSR BASED GENETIC LINKAGE MAP AND QTL MAPPING OF L*A*B* COLOR SPACE VALUES IN WINTER SQUASH ( Cucurbita moschata ) ................................ ................................ ............. 99 Introduction ................................ ................................ ................................ ............. 99 Materials and Methods ................................ ................................ .......................... 101 Plant Material ................................ ................................ ................................ 101 DNA Extraction ................................ ................................ ............................... 101 Marker Selection, Polymorphism Screening, and Genotyping ........................ 102 Linkage Map Construction and QTL Mapping Analyses ................................ 103 Results and Discussion ................................ ................................ ......................... 104 Polymorphic Marker Detection ................................ ................................ ....... 104 Linkage Map C omposition ................................ ................................ .............. 105

PAGE 7

7 Cucurbita Map Comparisons ................................ ................................ .......... 105 QTL Analyses ................................ ................................ ................................ 107 QT L detection of L* a* b* chroma and hue color space values ................ 107 Implications for population selection on QTL detection ............................ 108 Implica tions and Practical Applications of QTL Analysis in Breeding for Color and Carotenoid Content in Cucurbita ................................ ................. 110 5 DEVELOPMENT AND UTILIZATION OF A PROTOCOL FOR SMALL DNA SEQUENCE DIFFERENCE DET ECTION IN MONOMORPHIC SIMPLE SEQUENCE REPEAT (SSR) MARKERS IN WINTER SQUASH ( Cucurbita moschata ) USING HIGH RESOLUTION MELTING CURVE ANALYSIS .............. 128 Introduction ................................ ................................ ................................ ........... 128 Materials and Methods ................................ ................................ .......................... 131 Plant Material and DNA Extraction ................................ ................................ 131 Marker Selection and Polymorp hism Screening ................................ ............. 131 DNA Sequencing, Alignment and Comparison ................................ ............... 132 High Resolution Melt Curve Analysis Protocol Development ......................... 132 Results and Discussion ................................ ................................ ......................... 134 DNA Sequence Differences ................................ ................................ ............ 134 Genotyping DNA Sequence Differences Using the LightCycler 480 .............. 135 Evaluating DNA Concentration Effects on Genotyping With High Resolution Melt Curve Analysis ................................ ................................ .................... 136 6 CONCLUSIONS ................................ ................................ ................................ ... 147 LIST OF REFERENCES ................................ ................................ ............................. 150 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 160

PAGE 8

8 LIST OF TABLES Table page 2 1 Means and broad sense heritability ( H ) estimates of color space values, L*, a*, b*, c hroma and h ue a measured in fruit flesh of 11 Cucurbita cultigens. ....... 45 2 2 Means and broad sense heritability ( H ) estimates of carotenoids (g/g FW) measured in 11 Cucurbita cultigens. ................................ ................................ .. 46 2 3 Pearson correlation coefficients (r) between color space values (L*, a*, b*, chroma and hue ) and carotenoids calculated from the means of 11 Cucurbita cultigens. ................................ ................................ ............................ 47 2 4 color space values. ................................ ................................ ............................. 47 3 1 Harvest outline for all populations in two Cucurbita pepo and one Cucurbita moschata families in 2008 and 2009. ................................ ................................ 74 3 2 Examination of fixed effects due to genotypes and expected mean squares due to random effects in three populations of Cucurbita pepo derived from ................................ .............................. 75 3 3 Examination of fixed effects due to genotypes and expected mean squares due to random effects in three populations of Cucurbita pepo derived from ................................ ................................ 76 3 4 Examination of fixed effects due to genotypes and expected mean squares due to random effects in three populations of Cucurbita pepo derived from ................................ .................... 77 3 5 Means, standard deviations (SD), and ranges of colorimetric values (L*, a*, b*, chroma and hue ) of individuals from measured in two Cucurbita pepo an d one Cucurbita moschata families. ................................ ................................ ....... 78 3 6 Means, standard deviations (SD), and ranges of colorimetric values (L*, a*, b*, chroma and hue ) a ( Cu curbita pepo ) BC populations. ................................ ................................ ....... 80 3 7 Means, standard deviations (SD), and ranges of colorimetric values (L*, a*, b*, chroma and hue ) a ( Cucurbi ta pepo ) BC populations. ................................ ................................ ....... 82 3 8 Means, standard deviations (SD), and ranges of colorimetric values (L*, a*, b*, chroma and hue ) a (SDub) ( Cuc urbita moschata ) BC populations. ................................ ................... 84

PAGE 9

9 3 9 Expected mean squares due to random effects in respective parent 1, parent 2, and F 1 of three Cucurbita populations ................................ ........................... 86 3 10 Genetic means estimates for heritability of colorimetric values (L*, a*, b*, chroma and hue ) a ( Cucurbita pepo ) populations with one fruit per genotype b ................................ 87 3 11 Variance estimates for calculating heritability of colorimetric values (L*, a*, b*, chroma and hue ) a ( Cucurbita pepo ) populations with one fruit pe r genotype b ................................ 88 3 12 Genetic means estimates for heritability of colorimetric values (L*, a*, b*, chroma and hue ) a ( Cucurbita pepo ) populatio ns with one fruit per genotype b ................................ 89 3 13 Variance estimates for calculating heritability of colorimetric values (L*, a*, b*, chroma and hue ) a ( Cucurbita pepo ) populations with one fruit per genotype b ................................ 90 3 14 Genetic means estimates for heritability of colorimetric values (L*, a*, b*, chroma and hue ) a ( Cucurbita moschata ) populations with one fruit per genotype b .......................... 91 3 15 Variance estimates for calculating heritability of colorimetric values (L*, a*, b*, chroma and hue ) a (SDub) ( Cucurbita moschata ) populations with one fruit per genotype b ............ 92 3 16 Epistatic gene effect estimates for colorimetric v alues (L*, a*, b*, chroma and hue ) a Cucurbita pepo ) populations with one fruit per genotype b ................................ ............................ 93 3 17 Epistatic gene effect estimates for co lorimetric values (L*, a*, b*, chroma and hue ) a Cucurbita pepo ) populations with one fruit per genotype b ................................ ............................ 95 3 18 Epistatic gene effect estima tes for colorimetric values (L*, a*, b*, chroma and hue ) a ( Cucurbita moschata ) populations with one fruit per genotype b ......................... 97 4 1 L oci summary in C. moschata linkage map ................................ ..................... 111 4 2 Individual linkage group summaries for marker composition, length, density of markers and average marker distance in C. moschata ............................... 112 4 3 Linkage map summary in C. moschata. ................................ ........................... 114

PAGE 10

10 4 4 Common simple sequence repeat (SSR) markers on linkage groups between the C. moschata F 2 ................................ ......... 115 4 5 QTL detected for flesh color using L*a*b* color space values in a Cucurbita moschata F 2 ............................ 116 5 1 epeat (SSR) markers in winter squash, Cucurbita moschata ................................ ............................. 140 5 2 Summary of melting curve genotyping of scorable simple sequence repeat (SSR) markers, prior to DNA concentratio n adjustments based on Cp value ... 141 5 3 Estimates of DNA concentration in winter squash, Cucurbita moschata and F 2 rossing point (Cp) value ................................ ......... 142

PAGE 11

11 LIST OF FIGURES Figure page 2 1 A sampling of the range of fruit shapes and rind colors present in C. moschata fruit from the initial screen in 2007. ................................ .................... 48 2 2 A sampling of the range of fruit shapes and rind colors present in C. pepo fruit from the initial screen in 2007. ................................ ................................ ..... 49 2 3 Cross sections of the mesocarp of six C. moschata cultigens selected from the pre screen in 2007 for the 2008 correlation study. ................................ ....... 50 2 4 Cross sections of the mesocarp of five C. pepo cultigens selected from the pre screen in 2007 for the 2008 correlation study. ................................ ............. 51 4 1 Example gel images imaged under UV ligh t with ethidium bromide stain. ........ 117 4 2 Linkage map of Cucurbita moschata using both random amplified polymorphic DNA (RAPD) and simple sequence repeat (SSR) markers. ......... 118 4 3 Color space value L* quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and individual linkage group LOD scores ( P ................................ ................................ ................................ ......... 120 4 4 Color space value a* quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and individual linkage group LOD scores ( P ................................ ................................ ................................ ......... 121 4 5 Color space value b* quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and individual linkage group LOD scores ( P ................................ ................................ ................................ ......... 123 4 6 Color space value chroma quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and individual linkage group LOD scores ( P ................................ ................................ ..................... 124 4 7 Color space value hue angle quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and individual linkage group LOD scores ( P ................................ ................................ ..................... 125 5 1 Normalized and temperature shifted melting curves for simple sequence repeat (SSR) marker M042 showing three genotypic classes at sensitivity of 0.30 for 90 F 2 progeny ................................ ................................ ...................... 143 5 2 Normali zed and temperature shifted difference plot of melt curve shape showing three genotypic classes at sensitivity of 0.30 for 90 F 2 progeny ......... 144

PAGE 12

12 5 3 rial dilutions ............................. 145 5 4 Normalized and shifted melting curves for simple sequence repeat (SSR) marker M120 at sensitivity of 0.30 for 90 F 2 progeny ................................ ....... 146

PAGE 13

13 LIST OF ABBREVIATIONS a* CIE L*a*b* numeric color space value measuring color direction in red (a* > 0) or green (a* < 0) b* CIE L*a*b* numeric color space value measuring color direction in yellow (b* > 0) or blue (b* < 0) BB U S winter squash butternut cul Cucurbita moschata .) C CIE L*a*b* numeric color space value measuring chroma (saturation or vividness). A s chromaticity increases a color becomes more vivid; as it decreases a color becomes more dull. H CIE L*a*b* numeric color space v alue measuring (tint of color). This is an angular measurement where 0 = red, 90 = yellow, 180 = green, and 270 = blue. H Broad sense heritability measuring the ratio of genotypic variation in relation to the phenotypic variation present among individu als. h 2 Narrow sense heritability measuring the ratio of additive variation in relation to the phenotypic variation present among individuals. HPLC High Performance Liquid Chromatography HRM High Resolution Melting. L* CIE L*a*b* numeric color space valu e measuring lightness or darkness. The measurement ranges from black (0) to white (100) PI Plant Introduction TGA U.S. winter squash acorn Cucurbita pepo .) TKB U.S. winter squash acorn Cucurbita pe po .) RAPD Marker Random Amplified Polymorphic DNA Marker. SDub French ( Cucurbita moschata .) SNP Marker Single Nucleotide Polymorphic Marker. SSR Marker Simple Sequence Repeat Marker

PAGE 14

14 Abstract of Di ssertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COLOR AND CAROTENOID CONTENT IN SQUASH ( Cucurbita spp.) By Rachel Ann Itle December 2010 Chair: James W. Olmstead Major: Horticultural Science Carotenoids serve as a cancer preventative and as antioxidants that protect against degenerative health problems such as cardiovascular disease, cataract formation, and macular degeneration in o lder adults. Carotenoids with high importance in human health are those that serve as precursors to vitamin A (pro vitamin A carotenoids), and function in normal vision, bone growth, cell division and differentiation, and reproduction Squash ( Cucurbita spp.) is an excellent dietary source of carotenoids. The diversity and range of carotenoid types and concentrations within squash provide a means to increase the nutritional value through plant breeding. The objectives of this research were to: 1) determin e if the carotenoids of squash could be correlated with colorimetric analysis using th e CIE L*a*b* color value system, 2) determine heritability and gene action of flesh color in nine Cucurbita populations derived from three crosses, 3) create a genetic linkage map in C. moschata (2n=2x=40) with sufficient marker density for quantitative trait loci (QTL) studies to map genomic regions asso ciated with color, and 4) design a protocol to recover monomorphic simple sequence repeat (SSR) markers using high re solution melting (HRM) curve analysis.

PAGE 15

15 Strong correlations were found between color value a* and total carotenoids ( r = 0.91) and color value b* a nd chroma with lutein ( r = 0.87). Broad sense heritabilities ranged from 0.19 to 0.82 for L*, 0.12 to 0.32 fo r a*, 0.40 to 0.93 for b*, 0.36 to 0.92 for chroma, and 0.14 to 0.15 for hue across all three crosses Significant combinations of additive and dominance gene effects were identified for color space values L* and hue ( P A linkage map was construct ed usin g random amplified polymorphic DNA (RAPD) and SSR markers. QTL associations were examined using composite interval mapping in MapQTL 5.0. Twenty one linkage groups were obtained with QTLs detected on LG7 and 18 for L*, LG3, LG10, LG18 for a*, LG 5 and LG17 for b* and chroma and LG 3, LG13, LG14, LG15, and LG19 for hue Four of the ten monomorphic markers tested, M042, M120, M009, P098 were able to be recovered using HRM for addition to the linkage map data set.

PAGE 16

16 CHAPTER 1 INTRODUCTION The Cuc urbitaceae Family Cucurbitaceae, the gourd family, is a large plant family comprised of 825 species within 118 genera (Wang et al., 2007). Of the 118 genera, Citrullus, Cucumis, and Cucurbita are the most economically important, with the Cucurbita genus containing pumpkins, squash and gourds (Lebeda et al., 2007). Members of the Cucurbitaceae (Wang et al., 2007), with origins in India (cucumber, Cucumis sativus ) Africa ( Cucumis melo melon and Citrullus lanatus watermelon), and Central and South America ( Cucurbita spp., squash and pumpkin) (Wehner and Maynard, 2003). Yet, cucurbit crops can be adapted to grow in many agricultural environments, such as monocultu res in developed countries and in small garden settings with low input levels in non developed countries (Lebeda et al., 2007). Members of this family have been used by humans for both cultural and nutritional value for over 12,000 years (Brothwell and Bro thwell, 1969) and are associated with the origins of present day agriculture and the development of human civilizations. Additionally, although an Old World plant, cucurbit crops were one of the first to be domesticated in the New World (Bisognin, 2002). C ucurbits are also noted for a variety of shapes, sizes and colors of plants and fruit (Bates, Robinson, and Jeffrey, 199 0). In addition to food usage fruit from Cucurbitaceae have been used as storage containers and as a fiber source for sponges and sho e production (Moravec, Lebeda, and Kristkova, 2004).

PAGE 17

17 Taxonomy of the Cucurbitaceae The Cucurbitaceae family is in the phylum Magnoliophyta, class Magnoliopsida, subclass Dileniidae, order Violales (Walters and Keil, 1996). Additionally, all species comm only cultivated in the Cucurbitaceae family belong to the Cucurbitoideae subfamily (Wehner and Maynard, 2003). Plants of this subfamily are trailing herbs with climbing tendrils (Walters and Keil, 1996) that are sensitive to frost and complete their lifec ycle within one year (Wehner and Maynard, 2003). Leaves are palmately veined and alternate, and can be deep to simple lo bed and estipulate. Cucurbits are monoecious, with separate male (staminate) and female (pistillate) flowers both borne along the vine at the base of the leaf petiole (Wehner and Maynard, 2003) and rarely are perfect (Walters and Keil, 1996). Flowers can have five distinct sepals, or sepals can be varying degrees of connate to absent entirely, with an inferior ovary. Frui t is characteri zed by two, five connate carpels and one locule (Walters and Keil, 1996). Flowers are insect pollinated by bees since the pollen is too large and sticky to travel by wind (Wehner and Maynard, 2003). Additionally, there are several parietal ovules that co ntain up to five locules with an additional axile ovule present. Fruit of this family has been classified as a berry, pepo, capsule or achene (Walters and Keil, 1996). Economic Importance of Pumpkins, Squash and Gourds In 2007 the top five producers o f pumpkins, squash and gourds were China, India, Russia, the U.S. and Egypt and in 2009, 43,900 acres of both fresh market and processing squash were harvested in the U.S. totaling a production value of $203,464,000 (FAOSTAT, 2010). In 2009, Florida harvest ed 8,800 acres totaling $51,480,000 gross revenue a nd currently ranks 1st in U.S followed by California

PAGE 18

18 ($32,160,000), Georgia ($28,892,000), New York ($23,004,000), and Michigan ($11,739,000). The average price for U.S. grown squash was $28.20 / hundred weight (cwt) i n 2009. The price for Florida squash sold in 2009 was $45.00 / cwt., which decreased from $53.00 in 2008 and $52.80 in 2007. In 2009, Florida accounted for 20% of total U.S. harvested squash acreage and 25% of total U.S. cash receipts (USD A NASS, 2010). In the most recent census of agriculture ( 2007 ) Miami Dade County was the harvested acres, produced on only 18 farms. Hillsborough County wa s second i n the state for acreage with 623 harvested acres and 33 farms, followed by Hardee County with 133 harvested acres and 9 farms. Of the 188 farms producing 7,349 acres of squash in 2007, only 10 farms were larger than 100 acres, yet they accounted for 5,915 (80%) of the total squash acreage in Florida (USDA NASS, 2009). Florida produces a large amount of the U.S. summer squash ( C.pepo ) crop, but also produces winter squash ( C. moschata and C. pepo ), calabaza ( C. moschata ), banana squash ( C. maxima ), gourds ( Lagenaria spp. and Luffa spp.) and tropical squash ( Cucurbita spp.) (Swaider and Ware, 2002). The Florida squash crop is produced almost entirely for fresh market sale and is among a limited number of crops that is shipped from Florida 12 months of the ye ar (Mossler and Nesheim 2003) with s hipments at their peak from November to April ( Pollack, 1996 ). Nutrition of Squash An important component of the nutritional value of squash, Cucurbita spp., is its carotenoid content which serves as one source of pro v itamin A in plant tissues (Gross, 1991). Winter squash ( Cucurbita spp.) is also noted as a good source of dietary iron

PAGE 19

19 (Garrison and Somer, 1995) Developing countries produce roughly four times the total number of Cucurbita spp., pumpkins, squashes, and gourds, as do developed countries. The number of Cucurbita species that are produced in developing countries as compared to that of developed countries is reflective of Cucurbita spp. being a staple food crop in the developing world (Lebeda et al., 2007). In light of this, an increase in the nutritional content of squash would benefit both developed and developing nations; however it may have a greater benefit on the diets of people in developing nations. Importance of Carotenoids and Vitamin A in Human Health Roughly 60 different carotenoids occur naturally in plant tissue s (Seroczynska et al., 2006) and serve as a cancer preventative as well as antioxidants that protect the body against degenerative diseases (Murkovic et al., 2002) such as cardiovascula r disease, the formation of cataracts, and macular degeneration in older adults (Tan g, 2010). More specifically, carotenoid s of high importance in human health are pro vitamin A carotenoids which function to promote normal vision, bone growth, cell divis ion and differentiation, and reproduction (ODS/NIH, 2006) Although p lant sources do not provide vitamin A directly as animal sources do, converted into the bioavaliable form of retinol in the human body. Over 40 carotenoids have been identified as vitamin A precursors in plants (Santos and Simon, 2006). Carotenoids from plants that are common sources of pro vitamin A include beta carotene, alpha carotene, and beta cryptoxanthin; beta car otene is made into retinol twice a s efficiently as the latter two (ODS/NIH, 2006 ; Olson, 1996; Pavia and Russell, 199 9). Further, since plants supply the precursors to vitamin A, the amount of pro vitamin A carotenoids within plant tissues is converted int o retinol activity units, RAE. RAEs can then be used

PAGE 20

20 to compare vitamin A biological activity among the differing forms of vitamin A (Trumbo et al., 2001). The active form of vitamin A is r etinol which can be made into both retinal and retinoic acid (Tru mbo et al., 2001) Retinal is used as a pigment which is linked to a protein, opsin, to compose the chromoprotein rhodopsin. This protein is found in light sensitive rod cells in the eye and aid s the eye in adaptation to dim light. In bright light, rhodop sin breaks down and in low light the process is reversed (Alpern, 1971; Zile, 2001). Retinoic acid functions to aid in vertebrate embryonic development (Zile, 2001), as well as maintaining epithelial tissues such as skin, hair and mucous membranes, reprodu ctive health in sperm formation, and immune system functioning by aiding some white blood cell s in efficiently fighting infection ( ODS/NIH, 2006 ). Bo th the pre formed animal sources, vitamin A such as liver, and pro vitamin A carotenoid sources, such a s red, orange and yellow pigmented fruit, are absorbed in the intestine, transported in the blood as vitamin A esters, and stored in the liver when the amount ingested is <1mg ( ODS/NIH, 2006; Olson, 199 9 ; Pavia and Russell, 1999). With larger amounts, th e absorbed carotenoids are stored in the blood plasma and other tissue (IVACG, 1999). Recommended Dietary Allowance (RDA) for Carotenoids and Vitamin A To date, no official RDAs h ave been issued for carotenoid levels ( Murkovic et al., 2002) However, ther e are established RDA levels for Vitamin A released by the Institute of Medicine (Trumbo et al., 2001) Currently, for infants less than one year of age, the RDA range is from 400 500 g RAE. RDAs for children are 300, 400 and 600 g RAE for age ranges o f 1 3, 4 8 and 14 18 years, respectively. Adult males require 900 g RAE and adult females 700 g RAE. Pregnancy alters these levels by

PAGE 21

21 increasing the RDA by 70 g RAE. Additionally, lactating women have nearly double the requirement of non lactating wome n with 1,300 g RAE. carotene ha d the same vit amin A activity as 1 g retinol when converting values into RAEs. In 2001, the U S Institute of Medicine ( IOM ) set the new RAE standard as 1 carotene to 1 g retinol. The conversion rate of pro vitamin A carotenoids was reevaluated after observing individuals who had adequate intake of plant sources of pro vitamin A, but blood serum levels still showed deficiency. Questions then arose regarding the bioav ailability of pro vitamin A in plant sources. Further studies also suggest that the conversion rate may be closer to 1 g retinol : 21 carotene Under the previous conversion rate of 1 g retinol : 6 g carotene all populations have to ability to m eet vitamin A daily requirements from existing food sources. However, the new ratios make it increasingly difficult for those who obtain their primary source of vitamin A from pro vitamin A sources in vegetable crops to consume recommended levels of vitami n A, thus increasing the probability of vitamin A deficiency (West et al., 2002) This new standard shows an increased need to select for and increase carotenoid content in plants, such as squash, through plant breeding. Increasing Pro Vitamin A Rather t han Pre formed Vitamin A Content As reported by the World Health Organization (WHO) in 2001, 118 countries have noted vitamin A deficiencies as a common health problem, primarily in Africa and South East Asia. There are an estimated 100 to 140 million Vit amin A deficient children resulting in 250,000 to 500,000 cases of childhood blind ness annually Unfortunately, half of these children die within 12 months of lost sight. Child birth related deaths in

PAGE 22

22 women are estimated at 600,000 annually, with the maj ority being caused by Vitamin A defi ciency (Santos and Simon, 2006). Pre formed vitamin A sources are found in animal products such as meat, liver and eggs although these products are typically more expensive t han plant pro vitamin A sources. In develo ping countries, it is often easier to obtain sources rich in pro vitamin A carotenoids instead of pre formed vitamin A (Ong and Tee, 1992) Paradoxically, a lthough many developing nations have ample vegetables rich in pro vitamin A carotenoids vitamin A deficiencies still occur The cause of this is likely due to diets which are low in fat and protein, both of which aid in a b sorption of vitamin A precursors (Ameny and Wilson, 1997). However, vitamin A deficiency may also present itself if the individual h as a lower capacity to convert pro vitamin A forms into usable vitamin A within the body or if he or she has a high body mass index (Tang, 2010). These results suggest that increasing pro vitamin A forms may be more beneficial in terms of lower cost for t he consumer, and increases in pro vitamin A carotenoid levels in plants could beneficially impact both people in the developed and developing world. Analysis of Squash Fruit Carotenoid Analysis The predominant carotenoids found within squash ( Cucurbita s pp.) include lutein, carotene, and zeaxanthin (Gross, 1991). The most accurate way to analyze carotenoid content in plant tissues is through high performa nce liquid chromatography (HPLC ) Although highly precise and repeatable, this extraction method requires specialized l ab training This method also forces the r esearcher to examine the costs and benefits to using this method by considering the number of samples to be tested against the operation cost of running the equipment and hazardous solvents used

PAGE 23

23 (Rodreiguez Amaya and Kimura, 2004). In addition, c arotenoid analysis is complicated by the variable composition of vegetables throughout the year, differences in what is actually eaten and what is sampled, and different cooking method s that influence carotenoid bioavailab ility (Heinonen et al. 1989) To decrease time and increase efficiency in plant breeding selection, an alternative to chemical carotenoid analysis would be useful. Color Analysis Since carotenoids are a class of plant pigments, thei r levels have a direct impact on the color of plant tissues. For example, tomato carotenoid biosynthesis has been studied thoroughly as it pertains to fruit ripening. As the fruit matures, color development proceeds through stages of green, orange, pink and red; this is due to accumulation of the carotenoid lycopene within the chromoplasts of the fruit (Hirschberg, 2001). Within squash, a range of white, cream, yellow, and orange colors in the mesocarp of squash are also conditioned by carotenoids (Gross, 1991). A way to indir ectly and objectively measure the carotenoid content in squash by its color would be beneficial for manipulating squash color in a breeding program. The human eye is able to discern different colors and intensities. In the human eye, there are three types of cones used to detect the primary colors of light red, green and blue and their interaction allows the perception of color. However, due to subjectivity of the human eye from factors such as lighting on the object, background color behind the object and age of the eye, an objective method of color observation is desirable. The tristimulus color measurement system has three color sensors which have spectral sensitivity curves similar to that of the human eye and are referred to as color matching funct ions. These three functions are detected as x, y, and z coordinates

PAGE 24

24 and can be converted into a numerical value system (Konica Minolta Sensing, Inc., 1998) These coordinates can then be converted into one of many color systems The most preferred systems to measure color of fruit and vegetables are the Hunter tristimulus system and the CIE L*a* b* system (Seroczynska et al. 2006 ) Many studies have been conducted that correlate the measurement of color with pigment concentration in other agricultural pro ducts such as paprika (Ramakrishnan and Francis, 1973), leaves of parsley (Berset and Caniaux, 1983), blueberry (Francis, 1985), salmon (Ando et al., 1992), red pepper (Reeves, 1987), grapes (Watada and Abbott, 1975), peaches (Morrison, 1990), sweet potato (Ameny and Wilson, 1997; Takahata et al., 1993; Simonne et al. 1993), carrots (Ling et al., 1996), Swiss chard (Ihl, 1994), squash (Seroczynska et al. 2006; Francis, 1962), cranberries and wines (Francis, 1969). Due to the correlation of color intensi ty with nutrients such as carotene in sweet potato (Simonne et al. 1993; Ameny and Wilson, 1997), color intensity may be considered for use as a reliable indication of nutrient value, such as vitamin A. Currently, v isual observations of pumpkin varieti es ( C.pepo, C. maxima, and a cross between C. maxima and C. moschata ) by Murkovic et al. (2002) indicated that flesh with a higher carotene level had an orange appearance while flesh with higher lutein content and lower carotene content had bright yellow c oloration. The benefit to correlation of plant color with pigment concentration is both the rapidity of the method as well as the quantitative and descriptive evaluation of color in a single determination (Ameny and Wilson, 1997). Pigment concentration c orrelated with color values would also enable the plant breeder to indirectly select for increased

PAGE 25

25 nutrient content with a faster turn around in the selection process. Additionally, it would open the opportunity to breed for changing squash color to fit sp ecialty markets. Gene N umber and Inheritance of Color in Cucurbitaceae Flesh color in squash spans a wide range of color and intensity, indicating that expression of flesh color is likely a quantitative trait, controlled by many genes and influenced by env ironmental factors, rather than a qualitative trait controlled by few genes. Knowledge of the gene number and heritability of flesh color would allow plant breeders to estimate the population size and time needed to manipulate flesh color in squash. Fact ors affecting rind and flesh color in various plants in Cucurbitaceae have been studied. In C. pepo thirteen loci have been identified that control rind color (Paris et al., 2003). However, in C. maxima only three loci determining light and dark green rind color were identified (Lopez et al., 2003). In watermelon ( Citrullus lanatus ) canary yellow to scarlet red flesh color was found to be controlled by a single, dominant gene, while salmon yellow, white and red fleshed cultivars deviated from the ratio of a one gene model. Deviated ratios in salmon yellow, red and white flesh were suggestive of quantitative traits controlling flesh color (Gusmini and Wehner, 2006). Flesh color in melon ( Cucumis melo) has been found to be controlled by both major and mi nor genes (Cuevas et al. 2010). In melon, flesh color in both the F 2 and BC 1 populations fit an epistatic two carotene content were moderate, 0.55 in the F 1 F 2 and BC populations and 0.68 in the F 3 populations. To date, gene number and the inheritance of flesh color have not been studied in Cucurbita moschata or Cucurbita pepo

PAGE 26

26 Molecular T ools for Increasing Color Content of Squash Marker assisted selection (MAS) is a tool used by p lant breeders to make found to have significant associations with markers, MAS can be used to more effectively and efficiently select for desired traits in a population with a faster turn around time than by examination of phenotypic information alone. Marker types that have been used in the cucurbits include protein, morphological, and nuclear and plastid markers (Lebeda et al., 2007) However, a s a whole, little geno mic information is available for most cucurbit species. At present, t he largest amount of information in the Cucurbitaceae family exists in three principle economic crops: cucumber watermelon and melon ( Wang et al., 2007). Molecular Markers Used in Cu curbitaceae Morphological markers are typically used for cucurbit germplasm characterization aimed at screening for important traits (Lebeda et al., 2007). However, t he use of these phenotypic markers has been proven insufficient in mapping genomes within Cucurbitaceae due to the limited number available (Wang et al., 2007). Variations in isozyme profiles have assay ed diversity in Cucumis, Citrullus, Momordica and Cucurbita T he most developed and comprehensive information is available for cucumber (Knerr et al., 1989; Meglic et al., 1996). However, isozyme markers used in Cucurbita have produced varied, and at times, conflicting results in defining taxonomic relationships (Puchalski and Robinson, 1978; Decker Walters et al., 1990). Protein markers have a lso been used to evaluate enzyme s involved in seed storage in both melon, and cucumber (Bretting and Wid rlechner, 1995). T he most thorough isozyme marker analyses have been examining relationships between wild

PAGE 27

27 and domesticated individuals in C. pepo (Ignar t and Weeden, 1984; Decker and Wilson, 1987; and Decker W alters et al., 1993). Plastid markers, both chloroplast and mitochondrial, have been useful in determining phylogen e tic relatedness due to the ir high level of conservation (Decker Walters, et al., 2004a). In Cucurbita mitochondrial sequences have been used to evaluate different accessions suggest ing a minimum of six independent domestication events (Sanjur, et al., 20 02). Relationships between domesticated Cucurbita and its wild relatives ha ve al so been examined using chloroplast DNA (Wilson et al., 1992). Nuclear markers provide higher levels of variation and are more useful in determining relationships within and below the level of plant species than morphological, enzyme, or plastid markers (Le beda et al., 2007). Random amplified polymorphic DNA (RAPD) marker s are a dominant marker system with primers composed of randomly assembled 10 base pair sequences. This marker system is commonly used as a starting point in crops that do not have a high l evel of advancement in molecular marker systems such as members of the Cucurbitaceae family. RAPD and simple sequence repeat (SSR) markers have been most widely used across cucurbit crops, including Lagenaria, Citrullus, Cucumis and Cucurbit a. RAPDs have been used at the intraspecific level for cucurbits (Lebeda et al., 2007). However, due low variation in and low repeatability of RAPD markers, cucurbit breeders have sought other more stable and codominant marker systems where heterozygotes can unambiguo usly be determined ( Dijkhuizen et al., 1996; Garcia Mas et al., 2000). Restriction fragment length polymorphisms (RFLPs) were among the first types of codominant marker systems used in cucurbit crops With the establishment and

PAGE 28

28 availability of SSR markers SSRs quickly became the most used codominant marker system in cucurbit crops PCR based molecular marker systems that have been utilized less frequently in cucurbit crops include amplified fragment length polymorphisms (AFLPs), inter simple sequence repe ats (ISSRs), cleaved amplified polymorphic sequences (CAPS), restriction satellites, and sequence based amplified polymorphisms (SBAPs) (Lebeda et al., 2007). Genetic Linkage Maps Available in the Cucurbita genus Molecular markers can also be used to map t he chromosomes in the haploid genome for use in studying regions of the genome associated with complex phenotypic traits controlled by many genes in quantitative trait loci (QTL) studies. Knowledge of both desirable and undesirable QTL alleles in relati trait of interest is equally useful. Knowing the location of genes affecting different phenotypic traits in relation to one another provides information on genetic linkage and the likelihood of recombination during meiosis. This inform ation will aid the plant breeder in decisions such as best crosses to make, population size, and number of selfing generations needed until homozygosity is attained. L inkage map development within Cucurbita is limited, and only five genetic maps for the Curcubita now exist The first two maps constructed were of interspecific crosses between C. moschata and C. pepo and were composed solely of RAPD markers. The first, published by Lee et al. ( 1995 ) was a RAPD map of a n F 2 population. The second map, cr eated from a BC 1 population, was published by Brown and Myers ( 2002 ) and was composed of both RAPD and morphological markers. A third map in Cucurbita was created in C. pepo using an F 2 population with RAPD SSR, and morphological, markers (Zraidi and Lel ley, 2004)

PAGE 29

29 This third map was recently expanded and compared to a second C.pepo map. Zraidi et al. (2007) used two F 2 populations, created from two intraspecific crosses to create the first consensus map for C. pepo using RAPD, AFLP, SSR, and SCAR mark ers. Because of t he limited number of SSRs available for Cucurbita Gong et al. ( 2008a) developed over 500 SSR primers which were polymorphic in C. pepo C. moschata and C. ecuadorensis The researchers used these SSRs to update their first C. pepo map (Zraidi et al., 2007) and constructed a fourth published map in Cucurbita and the first SSR based map in C. pepo in 2008. The fifth and most recent Cucurbita map was also published in 2008. Gong et al. (2008b) constructed the first SSR based map of C. m oschata using an F 2 population, and examined its synteny with their most recent C. pepo map. The limited availability of genetic maps within Cucurbita indicates a need for genotypic information to be acquired to aid in breeding for desirable traits in sq uash, such as increasing color and carotenoid content. Quantitative Trait Loci (QTL) Detection in Cucurbitaceae Within Cucurbitaceae most of QTL identification has been with Cucumis sativus and Cucumis melo. Major areas of QTL identification in Cucumis sativus, include disease resistance, fruit quality traits and components of yield (sex expression, earliness, and lateral b ranching) (Staub et al., 2008). In Cuc u mis melo, QTL identification similarly includes fruit quality and disease resistance, in add ition to floral biology QTLs. Furthermore, Cucumis melo, melon, is considered the model plant within Cucurbitaceae ; most genomic work has been done in this species. Within melon, QTL regions and identified genes are sufficient in number to produce a synth etic genetic map (Pitrat, 2008).

PAGE 30

30 Unlike cucumber and melon, no QTLs have been identified for squash (Cucurbita spp.) or within Cucurbita Research Rationale Knowledge of color and carotenoid content in squash, Cucurbita spp., is currently limited. A bet ter understanding of the carotenoid content within squash flesh as it relates to color might expedite selection for color and nutrient content in squash. Additional knowledge of gene number and inheritance of flesh color would enable plant breeders to mor e effectively design populations to manipulate the desired trait. Finally, availability of markers associated with flesh color and a genetic linkage map identifying regions of the genome linked with flesh color would help breeders who wish to increase fle sh color and study its relation to other phenotypic traits. The objectives of this research project were to: 1. Attempt to correlate flesh color in Cucurbita with carotenoid content as measured by HPLC, 2. Measure flesh color segregation in multiple Cucurbita p opulations using L*a*b* color space values and determine the heritability of these components of color, and 3. Develop a linkage map for C. moschata and perform QTL analyses for L*a*b* color space values in the population.

PAGE 31

31 CHAPTER 2 CORRELATION BETWEEN L* A* B* COLOR SPACE VALUES AND CAROTENOID CONTENT IN PUMPKINS AND SQUASH ( Cucurbita spp .) Introduction Carotenoids are the principle pigments responsible for the many colors of leaves, fruits, and flowers in plants (Gross, 1991). They act as photoprotective a gents and accessory light harvesting complexes. Carotenoids also play an important role in human health by acting as sources of pro vitamin A or by acting as protective antioxidants required for proper reproduction, growth and development, a normal functio ning ocular system, epithelial cell integrity, and immune system functionality (FAO/WHO, 2002; Murkovic et al., 2002). In vegetables, common pro vitamin A carotenoids include cryptoxanthin (ODS/NIH, 2006). Other common caroten oids such as lycopene, lutein and zeaxanthin do not have vitamin A activity but serve as antioxidants. Pumpkins and squash ( Cucurbita spp.) are excellent dietary sources of carotenoids (Gross, 1991) and, in 2007, ranked 11 th in amount produced among othe r vegetables produced around the world (FAOSTAT, 2008). The predominant carotenoids found in pumpkins and squash i carotene (Gross, 1991). Based on reports by Holden et al. (1999) and Murkovic et al. (2002), the carotene found within Curcurbita species and their various fruit types can vary dr amatically. In their studies, the fresh weight (FW) carotene in summer type squash ( C. pepo ) were 0.0 Article reproduced with permission from: Itle, R.A. and E.A. Kabelka. 2009. Correlation betw een L* a* b* color space values and carotenoid content in pumpkins and squash ( Cucurbita spp.). HortSci., 44(3):633 637

PAGE 32

32 to 21.3 g/g, 0.3 to 1.7 g/g, and 0.6 to 23.0 g/g, respectively. Within winter type squash ( C. moschata and C. maxi ma carotene from 7.1 to 74.0 g/g FW. The mesocarp ( flesh ) colors of pumpkins and squash generally include a wide range of whites, yellows, and oranges (Gross, 1991). Th is color is based on the particular carotenoid types and concentrations which are influenced by genetic and environmental factors. Over a dozen genes that affect the rind and flesh color of squash have been described ( Paris and Brown, 2005) and include D ( Dark ), l 1 ( light coloration 1), l 2 ( light coloration 2) and B ( Bicolor ). Tadmor et al. (2005) studied the effects of these genes in different combinations within near isogenic lines (NILs) of C. pepo In genetic backgrounds that lacked either the dominan t D or dominant L 2 alleles, a yellow flesh color developed. In genetic backgrounds with either dominant D or L 2, a yellow orange flesh color developed and when the dominant allele of B interacts with the dominant allele of L 2, an intense orange flesh c olor will o ccur. One additional gene that a ffects squash flesh color is the dominant Wf ( White flesh ) which confers a white flesh color by preventing yellow pigment accumulation (Paris and Brown, 2005). The broad range in carotenoid types and concentratio ns among and within Cucurbita species indicates the potential for genetic improvement of these compounds through plant breeding. Accuracy in breeding will require estimates of carotenoid types and their concentrations that are precise enough to distinguish genotypic differences among breeding material. One obstacle is that the extraction and analysis of carotenoid content is time consuming and expensive. In practical breeding programs, it is not realistic to analyze the carotenoid content of even a small se gregating population for

PAGE 33

33 selection of genotypes with high levels of carotenoids. An alternative reliable method to estimate carotenoid content and concentration would be beneficial. High performance liquid chromatography (HPLC) is used to chemically analy ze tissues for carotenoid types and concentrations (Gross, 1991). It is labor intensive and expensive but is a reproducible and highly sensitive process that can separate, identify, purify, and quantify carotenoid levels. In contrast, colorimeters, which o bjectively measure and describe visible color, are relatively inexpensive and easy to use. The preferred methods to objectively measure color are the tristimulus Hunter and the CIE L*a*b* systems (Seroczynska et al., 2006). Tristimulus color measurement sy stems have three color sensors with spectral sensitivity curves similar to that of the human eye and are referred to as color matching functions (Konica Minolta Photo Imaging, U.S.A., Inc., Mahwah, NJ). These three functions are detected as x, y, and z coo rdinates which can be converted into the desired color measurement value system. Previous studies have correlated color measurement systems with carotenoid content in vegetable crops such as sweet potato (Ameny and Wilson, 1997 ; Simonne et al., 1993), pepper (Reeves, 1987), and winter type squash (Francis, 1962; Seroczynska et al., 2006). Interestingly, the authors of these studies differ in their opinions as to how reliable colorimetric analysis are at es timating carotenoid content and concentrations. If a fair estimate of carotenoid content and concentration could be obtained, this rapid and inexpensive method could be very useful in breeding pumpkins and squash for enhanced carotenoid levels. The objecti ve of this research was to determine if the carotenoid content and concentration

PAGE 34

34 of pumpkin and squash ( C. moschata and C. pepo ) can be correlated with colorimetric analysis using the CIE L*a*b* color value system. Materials and Methods Plant M aterial An initial selection of 15 C. moschata (Figure 2 1) and 15 C. pepo (Figure 2 2) cultigens [cultivars, heirlooms, and plant introductions (PI)] was made based on subjective descriptions of flesh color with samples representing a range of whites, yellows, an d oranges. Seeds of c ultivars and heirlooms were purchased from commercial seed producers while PIs were obtained from the USDA ARS North Central (Ames, IA) and Southern (Athens, GA) Regional Plant Introduction Stations. Field Trials A preliminary field study was conducted 6 April 15 June 2007 at the Plant Science Research and Education Center (PSREC) located in Citra, FL to evaluate and select from the 15 C. moschata and 15 C. pepo cultigens those that would provide a range of white, yellow, and orange flesh color based on colorimetric analysis using the method described below. A randomized complete block design with two replications per cultigen, eight plants per plot, was used. Temperatures recorded by the Florida Automated Weather Network (FAWN) reco rded 22C avg., 38C max., and 5C min. over the course of the growing season. From the preliminary study, six C. moschata (Figure 2 3 ) and five C. pepo (Figure 2 4 ) cultigens were selected for both colorimetric and HPLC analysis following plantings at tw o locations, PSREC and the Gulf Coast Research and Education Center (GCREC), Wimauma, FL, during 11 March 27 June 2008. Average shelter air t emperatures at 60cm recorded by FAWN were 23 C avg., 39 C max., and 2C min. and 23C avg.,

PAGE 35

35 36 C max., and 4 C mi n. for PSREC and GCREC, respectively. A randomized complete block design with two replications per cultigen, eight plants per plot, was used at each location. Recommended conventional cultural practices and fertility rates for Florida squash were followed for both 2007 and 2008 (Olson and Simonne, 2007). Colorimetric and HPLC a nalysis A total of 220 mature fruit (11 cultigens x 5 fruit x 2 replications per location x 2 locations) were harvested for colorimetric and HPLC analysis. Fruit were considered m ature at the time they were ready for harvest based on common commercial practices for each squash type. Color was recorded using a Minolta CR 400 Colorimeter (Minolta Camera Co., Ltd., Ramsey, NJ) tristimulus color analyzer, equipped with an 8 mm diameter measuring area and diffuse illumination of a 2 Standard Observer. The L* coordinate indicates darkness or lightness of color and ranges from black (0) to white (100). Coordinates, a* and b*, indicate color directions: +a* is the red direction, a* is th e green direction, +b* is the yellow direction and b* is the blue direction. Chroma is the saturation or vividness of color. As chromaticity increases, a color becomes more intense, as it decreases a color becomes more dull. Hue angle is the basic unit of color and can be interpreted, for example, as 0 = red and 90 = yellow. Both chroma and hue are derived from a* and b* using the followi ng equations: Metric chroma: C = [ (a*) 2 + (b*) 2 ] and metric hue a ngle: h = tan 1 (b*/a*) [degrees]. Each fruit was sliced transversely, and L*a*b* color space measurements from the edible flesh (mesocarp) of each fruit were recorded within five minutes to avoid discoloration. Our pr eliminary study revealed r eplicate measurements of the mesocarp of each fruit per cultigen to be not significantly ( P from one another. T herefore, avoiding the seed cavity and surrounding tissue (approx. 10 mm), three

PAGE 36

36 random measurements per fruit were recorded. A total of 660 (11 cultigens x 5 fruit x 3 colorimeter measurements per fruit x 2 replications per location x 2 locations) colorimetric values were recorded. The separation and quantification of carotenoids were accomplished by H PLC. A total of 22 samples (11 cultigens per location) were prepared for HPLC analysis, directly following colorimetric measurements. From each of five fruit harvested per replication, a 10.0 +/ 0.1 g FW cubed flesh sample was cut and combined to make a 1 00.0 g +/ 0.5 g sample per cultigen per location. Samples were vacuum sealed in plastic storage bags within 50 minutes of initial fruit slicing, wrapped with aluminum foil, labeled and held at 20C. Samples were then sent frozen to Craft Technologies, In c. (Wilson, NC) for saponification and HPLC processing. Carotenoids measured included lutein, zeaxanthin, cis lutein zeaxanthin, carotene, cis carotene, and total carotenoids. All fruit was processed with the colorimeter and prepared for HPLC within three days of field harvest and underwent HPLC analysis within four weeks of harvest. Data a nalysis Color space values and carotenoid content of the Cucurbita cultigens, across locations, were subjected to analysis of variance by the GLM procedure of SAS ( Statistical Analysis System version 9.2, SAS Institute, Cary, NC). Cultigens, locations, replications, fruit, and repl icate measurements within fruit were considered as random effects. Least significant differences (LSD) among cultigens were determined at a 5% significance level. Cultigen means across locations were calculated for each trait. Components of variance were e stimated by using the VARCOMP procedure of SAS. Broad sense heritabilities, based on cultigen means, were calculated for each trait s ) of rank correlations (Steel, Torrie

PAGE 37

3 7 and Dickey, 1997) were calculated to test differences in rank order among the cultigens between the two locations. Pearson correlation coefficients (r) and linear regression (R 2 ) between color space values and carotenoid content were calculated from the means of the cultigens across locati ons using the CORR and REG procedure of SAS. Scatter plots of the data indicated that all relationships between color space values and carotenoid content were linear. Results and Discussion Colorimetric E valuation Significant ( P 0.0001) differences were observed among cultigens for flesh color represented by the five color space values L*, a*, b*, chroma, and hue. This demonstrates that genetic variation for flesh color is present among the cultigens tested. Locations were not significantly ( P fferent for any of the color space value s, but there was a significant ( P < 0.0001) cultigen by location interaction. Based on s cultigens between locations was noted. This indicates that the flesh color of each cultigen was consistent despite different environments. Replicate measurements within fruit of each cultigen were not significantly ( P P 0.02) differences were observed for t he color space values among fruit within cultigens. Variance component analysis, however, revealed that estimates of variance among fruit within cultigen s was less than 1.5% of the phenotypic variance for each color space value. With such a small effect, l ittle would be gained by evaluating more fruit per cultigen. Broad sense heritability ( H ) estimates of the five color space values (Table 2 1 ) ranged from 0.81 to 0.93 indicating that genetic improvement and effectiveness of selection for color would be mo derate to high.

PAGE 38

38 Subjective observations of flesh color were consistent with the color space value hue in that lower hue angles corresponded to more orange red flesh and higher hue angles corresponded to more yellow flesh (Table 2 C. mo schata ), with a mean hue angle of 77.5, represented orange red flesh color in this study and was significantly ( P 0.05) C. pepo ) reflected a whitish yellow flesh color with a mean hue angle of 102.9. Differences in hue of the cultigens tested may be due to the various types and ratios of pigments present. Mean color space values a* (+ a* red direction; a* green direction) ranged from C. moschata C. pepo ) to 14.8 C. moschata ) (Table 2 1). The lowest mean color space value b* (+b* y ellow direction; b* blue direction) was 9.8 found within PI 314806 ( C. pepo ) and the C. moschata ). Cultigens with hue 4.9, and mean b* orange flesh, mean a* values ranging from 6.6 to 14.8, and mean b* values ranging from 47.3 to 71.4. A range of L*, lightness (83.8) to darkness (70.5) of color, and chroma, dullness (10.0) to vividness (72.2) of color was present within the cultigens evaluated (Table 2 1) and may reflect different concentrations of pigments. Three yellow orange cultigens with hue values not significantly ( P C. moschata C. moschata C. pepo

PAGE 39

39 (mean chroma = 72.2) compare (mean L* = 70.5) and duller (mean chroma = 47.8). The yellow orange flesh and rind B ( Bicolor ) gene. Typically, acorn type squash have light angle = 97.7). Several genes tha t affect rind as well as the flesh color of squash have been described ( Paris and Brown, 2005; Tadmor et al., 2005) including D ( Dark ), l 1 ( light coloration 1), l 2 ( light coloration 2) and B ( Bicolor ). Tadmor et al. (2005) studied the effects of these pa rticular genes in different combinations within near isogenic lines (NILs) of C. pepo and reported that when the dominant allele of B interacts with the dominant allele of L 2 an intense orange rind and flesh color will develop. The flesh ividness (mean chroma = 65.3). Carotenoid Content Among cultigens tested, significant ( P 0.03) differences were ob carotene, cis carotene, and total carotenoids suggesting genetic variation in carotenoid accumulation N o significant differences were observed for lutein ( P = 0.08), zeaxanthin ( P = 0.07 ) or cis lutein zeaxanthin ( P = 0.06 ) but this may be a reflection of restricted sampling for analysis. In this study, locations were not significantly ( P > 0.05) different for any carotenoid measured. Broad sense heritability ( H ) estimates of the carotenoids measured ranged from 0.37 to 0.85 (Ta ble 2 2)

PAGE 40

40 implying that genetic improvement and effectiveness of selection would be most effective for carotene (H 2 = 0.85) and carotene (H 2 = 0.74). On average, orange red and yellow orange flesh colored cultigens contained 21.5 g/g FW total carotenoids while yellow flesh colored cultigens contained 2.4 g/g FW total carotenoids (Table 2 2). This re presented a 9 fold increase in total carotenoids in orange red and yellow orange flesh colored cultigens versus yellow flesh colored cultigens. The most abundant carotenoids within the orange red and yellow orange flesh colored cultigens were lutein (avera ge = 9.8 g/g FW carotene (average = 4.5 carotene (average = 5.4 g/g FW). The most abundant carotenoid within the yellow flesh colored cultigens was lutein (average = 1.7 g/g FW). Zeaxanthin (range = 0.1 0.2 g/g FW), cis lutein zeaxanthin (range = 0.1 1.2 g/g FW), and cis carotene (range = 0.1 1.7 g/g FW), were measured but not found in all cultigens evaluated. Tadmor et al. (2005) evaluated the carotenoid content of NILs of C. pepo differing in fruit pigmentation loci B ( Bicolor ), D ( Dar k ), l 1 ( light coloration 1), and l 2 ( light coloration 2). In their study, genetic backgrounds that lacked either the dominant D or dominant L 2 alleles had yellow flesh color On average, these NILs contained 1.2 g/g FW lutein, 0.6 g/g FW carotene an carotene, and 1.8 g/g FW total carotenoids. Within our yellow flesh colored C. pepo cultigens, similar carotenoid concentrations were obtained with an average lutein concentration of 1.8 g/g FW, an average carotene concentration of 0 .6 g/g FW, and an average total carotenoid concentration of 2.7 g/g FW (Table 2 2). In our study, no significant ( P

PAGE 41

41 0.05) differences among the yellow flesh colored C. pepo cultigens for any of the carotenoids were identified. In this study we evaluat orange flesh colored C. pepo cultigen. The yellow orange flesh color of this cultigen is due to the presence of the B pigmentation allele. In our study, this cultigen had at least a 3 fold increase in total carotenoid conten t compared to other C. pepo cultigens (Table 2 2). In the Tadmor et al. (2005) study, the effect of B on carotenoid content within their C. pepo NILs was dependent on the genetic combinations of pigmentation loci D l 1 and l 2. In genetic backgrounds with either dominant D or dominant L 1, the B / B genotype had an approximate 2 fold increase in total carotenoid content. In NILs with B / B and L 2/ L 2 genotypes total carotenoid content increased 10 further investigated to determine its allelic state at the pigmentation loci D l 1 and l 2 and the cause of its increased carotenoid content. The C. moschata content (42.3 g/g FW) and signif icantly ( P carotene (14.9 g/g FW) and carotene (15.3 g/g FW) than to other C. moschata cultigens evaluated in this study (Table 2 2). This suggests genetic variability is present within C. moschata that leads to elevated carotenoid conte nt and concentrations. Additional evidence for variability in carotenoid content and concentration within C. moschata is provided by Murkovic et al. (2002). The C. moschata cultigens in their study averaged 86.7 g/g FW total carotenoids (range = 41.6 to 1 43.0 g/g FW), carotene (range = 9.8 to carotene (range = 31.0 to 70.0 g/g FW). The carotenoid averages and their ranges were higher than those found in our study. It

PAGE 42

42 would be interesting to investigate the genetic basis that leads to elevated carotenoids within C. moschata and to compare this with the C. pepo findings of Tadmor et al. (2005). Correlations between Color Space Values and C arotenoids Pearson correlations coefficients (r) between color space values, L*, a*, b*, chroma, and hue, with lutein, carotene and total carotenoids, the four most prominent carotenoid s measured in this study, were calculated (Table 2 3). The color value L* (lightness or darkness) correlated ( P < 0.03) negatively with lutein (r = 0.68) and total carotenoids (r = 0.66). A negative correlation between L* and certain carotenoids would be expected since any increase in pigment would increase the darkness and thereby decrease L*. The strength of our two linear relationships, however, would be considered weak (r 0.70). The color value a* (color direction in carotene (r = carotene (r = 0.70). The color value b* (color direction in yellow or blue) and chroma (saturation or vividness of color) correlated strongly with lutein (r = 0.87 for both) and moderately with total carotenoids (r = 0.75 and r = 0.76). Hue (tint of color; an angula r measure) was weakly correlated carotene (r = carotene (r = 0.69), but moderately correlated with lutein (r = 0.80) and total carotenoids (r = 0.83). The negative correlation observed between hue and the carotenoids measured in this study suggests that as hue angle s decrease, carotenoid concentrations increase. carotene in winter type squash have previously been reported. Francis (1962) evaluated cultigens of

PAGE 43

43 C. maxima and C. moschata and identifi ed moderate to strong correlations between L* (r = 0.78), a* (r = 0.83), b* (r = 0.79), chroma (r = 0.93), and hue (r = 0.96) with total carotenoids. In a study evaluating C. maxima germplasm, Serocz y nska et al. (2006) reported poor to fair correlations between L* (r = 0.53), a* (r = 0.77), b* (r = 0.76), and chroma (r = 0.77) with total carotenoids and poor to fair correlations between L* (r = 0.54), a* (r = 0.74), b* (r = 0.66), and chroma (r = 0.67) with carotene Comparing our findings with those of Francis (1962) and Serocz y nska et al. (2006) revealed similar correlations and strengths except our correlation between a* (r = 0.91) and total carotenoids was stronger. In other vegetable crops, s tudies ha ve reported correlations between colorimetric values al. (2000) found L* (r = 0.91 and r = 0.92) and a* (r = 0.87 for both) to be strongly correlated with lycopene content. In sweet potato, Simonne et al. (1993) reported a carotene (r = 0.99) whereas Ameny and Wilson (1997) r carotene (r = 0.74). All four reports viewed colorimetric analysis as an appropriate estimator of carotenoid concentrations. A report by Reeves (1987), however, questioned the wisdom of assessing carotenoid concentrations by tristimulus colorimetry. He stated that while correlations may show statistical significance, the variation explained by the correlation ma y be too low to be of practical use. In his evaluation of pepper color and carotenoid data, the best parameter to predict total carotenoid content was a negative correlation (r = 0.713) with L* from pureed peppers. In his view, a mathematical equation bas ed on

PAGE 44

44 this correlation would, at best, explain only 51% (R 2 ) of the observed variation between L* and total pigment content. The strongest linear relationship, in our study, was found between a* and total carotenoids (r = 0.91; R 2 = 0.83). The R 2 value in dicates that 83% of the variation in total carotenoids can be accounted for by the change in a*. The color values b* and chroma were also strongly correlated with lutein (r = 0.87; R 2 = 0.76 for both). The R 2 values indicate that 76% of the variation in lu tein can be accounted for by the change in b* or chroma. While regression equations based on these correlations may account for only 76 83% of the variation for lutein and total carotenoids, respectively, they may still be useful for estimating these conce ntrations. Regression equations for the prediction of lutein and total carotenoids based on a*, b* and chroma, based on our findings, are provided in Table 2 4 Co nclusion In this study, we found a range of color and carotenoid types and concentrations wit hin pumpkins and squash. This genetic variation should make it possible to increase the nutritional value through crossing and selection from within and among the different types with high levels of carotenoids. Based on this study, strong correlations bet ween colorimetric values and carotenoid content were identified. These close associations will assure that indirect selection for high carotenoid content within pumpkin and squash breeding material will be successful, easy to implement, and inexpensive.

PAGE 45

45 Table 2 1. Means and broad sense heritability ( H ) estimates of color space values, L*, a*, b*, chroma and hue a measured in fruit flesh of 11 Cucurbita cultigens. a Numeric description of color using L*a*b* CIELAB color space. L* (lightness or darkness) ranges from black (0) to white (100); a* color direction in red (a* > 0) or green (a* < 0); b* color direction in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more vivid; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red and 90 = yellow. b Cultigens include cultivars, heirlooms, and plant introductions (PI) previously selected to represent a range of color from white to orange based on L*a*b* color spaces values. c Values shown are based on 5 fruit x 3 colorimetric measurements per fruit x 2 replications per location x 2 locations per cultigen. Cultigen b Species L* c a* b* Chroma Hue Color (subjective) Butterbush C. moschata 73.7 14. 8 65.9 67.7 77.5 orange red Waltham Butternut C. moschata 73.4 10.4 71.4 72.2 81.8 vivid yellow orange Sucrine DuBerry C. moschata 70.5 6.6 47.3 47.8 82.2 dull yellow orange Ponca Butternut C. moschata 74.8 7.3 68.1 68.5 83.9 yellow orange PI 458 728 C. moschata 75.4 0.2 47.8 47.8 90.0 dull yellow Tennessee Sweet Potato C. moschata 83.2 4.9 31.5 31.9 99.0 light yellow Table Gold Acorn C. pepo 76.5 9.4 64.5 65.3 81.9 yellow orange Table King Bush C. pepo 83.8 2.4 43.4 43.5 93.3 light yell ow C. pepo 81.6 2.5 40.8 40.9 93.8 light yellow Fordhook Acorn C. pepo 83.4 4.9 36.3 36.6 97.7 light yellow PI 314806 C. pepo 82.2 2.2 9.8 10.0 102.9 light whitish yellow LSD 0.05 0.8 0.8 1.5 1.6 0.8 H 0.81 0.87 0.93 0.93 0.91

PAGE 46

46 Table 2 2. Means and broad sense heritability ( H ) estimates of carotenoids (g/g FW) measured in 11 Cucurbita cultigens. Cultigen a Lutein b Zeaxanthin cis Lutein/ Zeaxanthin Carotene C arotene Cis Carotene Total Carotenoids Color (subjective) Butterbush 9.2 0.2 0.9 14.9 15.3 1.7 42.3 orange red Ponca Butternut 8.7 0.1 0.6 1.8 2.1 0.3 13.9 vivid yellow orange Waltham Butternut 17.3 0.1 1.2 3.0 3.8 0.9 26.8 dull yellow orange Sucrine DuBerry 4.3 0.0 0.3 2.6 2.1 0.4 9.7 yellow orange PI 458728 3.9 0.0 0.3 0.4 0.6 0.1 5.3 dull yellow Tennessee Sweet Potato 0.6 0.0 0.3 0.0 0.2 0.0 1.1 light yellow Table Gold Acorn 9.3 0.2 0.9 0.0 3.6 0.6 14.7 yellow orange Table Ki ng Bush 2.6 0.0 0.2 0.0 0.9 0.1 3.9 light yellow Sweet Potato 1.4 0.0 0.1 0.0 0.6 0.1 2.2 light yellow Fordhook Acorn 1.4 0.0 0.1 0.0 0.4 0.0 1.9 light yellow PI 314806 0.1 0.0 0.0 0.0 0.0 0.0 0.1 light whitish yellow LSD 0.05 10.4 0.2 0.8 4.1 5.6 0.9 20.4 H 0.37 0.49 0.42 0.85 0.74 0.59 0.59 a Cultigens include cultivars, heirlooms, and plant introductions (PI) previously selected to represent a range of color from white to orange based C. moschata. C. pepo b Separation and quantification of carotenoids were accomplished by HPLC. Values shown based on 100 gram fresh weight samples per cu ltigen per location.

PAGE 47

47 Table 2 3. Pearson correlation coefficients (r) between color space values (L*, a*, b*, chroma and hue ) and carotenoids calculated from the means of 11 Cucurbita cultigens. Traits a L* a* b* Chroma Hue r P value r P value r P valu e r P value r P value Lutein 0.68 0.0209 0.84 0.0013 0.87 0.0006 0.87 0.0006 0.80 0.0033 Carotene 0.49 0.1246 0.70 0.0155 0.45 0.1609 0.47 0.1438 0.62 0.0433 Carotene 0.49 0.1303 0.77 0.0052 0.55 0.0809 0.56 0.0703 0.69 0.0198 Total Carotenoids 0.66 0.0285 0.91 0.0001 0.75 0.0076 0.76 0.0062 0.83 0.0017 a Numeric description of col or using L*a*b* CIELAB color space. L*, lightness or darkness; a* color direction in red or green; b* color direction in yellow or blue; chroma saturation or vividness; hue tint of color. Table 2 irs of related carotenoids with color space values. Pairs of relationships a Regression equations r R 2 Total Carotenoids and a* Total Carotenoids = 6.1226 + 1.7106*a* 0.91 0.83 Lutein and b* Lutein = 6.3881 + 0.2446*b* 0.87 0.76 Lutein and Chroma Lute in = 6.3743 + 0.2418*Chroma 0.87 0.76 a Lutein and total carotenoids in g/g FW.

PAGE 48

48 Figure 2 1. A sampling of the range of fruit shapes and rind colors present in C. moschata fruit from the initial screen in 2007.

PAGE 49

49 Figure 2 2 A sampling of the range of fruit shapes and rind colors present in C. pepo fruit from the initial screen in 2007.

PAGE 50

50 A B C D E F Figure 2 3 Cross sections of the mesocarp of s ix C. moschata cul tigens selected from the pre screen in 2007 for the 2008 correlation study

PAGE 51

51 A B C D E Figure 2 4 Cross sections of the mesocarp of five C. pepo cultigens selected from the pre screen in 2007 for the 2008 correlation study. A) and E) PI 314806

PAGE 52

52 CHAPTER 3 HERITABILITY ESTIMATES OF L* A* B* COLOR SPACE VALUES IN WINTER SQUASH ( Cucurbita spp. ) Introduction The mesocarp tissue of squash ( Cucurbita spp.) has a wide range of color, from white to orange, which is conditioned by differen t carotenoid types and concentrations. In addition, carotenoids contribute to the nutritional value of squash (Gross, 1991), as an antioxidant source that protect s the body against cancer and degenerative diseases (Murkovic et al., 2002). Certain caroteno ids within squash can also serve as a source of pro vitamin A, a precursor of vitamin A available in plant tissues, which functions to promote normal vision, bone growth, cell division and differentiation, and reproduction (ODS/NIH, 2006). In chapter 2, correlations between L* a* b* color space values and carotenoid content within flesh color in squash ( Cucurbita spp.) were identified (Table 2 3). Increase in color space values a* (+ a* red direction; a* green direction) b* (+b* yellow direction; b* b lue direction) and chroma (vividness or saturation of color) were correlated with an increase in total carotenoid content and lutein, respectively. Collectively, these results indicate that L* a* b* color space values can be used as indirect estimates of carotenoid content. These relationships are beneficial due to the reduced time and cost associated with characterization of nutrient content by color analysis as compared with chemical analysis. Correlation results from this study indicate that squash wi th dark (lower L*) and more vivid (higher chroma ) flesh color has more nutritional value than lighter colored varieties. Previous studies have determin ed the genetic control of rind and flesh color in members of the Cucurbitaceae. In C. pepo (Paris, 2000 ) and C. maxima (Lopez Anido

PAGE 53

53 et al., 2003 ) thirteen and three loci determining light and dark green rind color were identified, respectively. In other cucurbit crops, only flesh color has been examined. In watermelon, Citrullus lanatus both canary yello w and scarlet red flesh color were found to be contr olled by a single, dominant gene dominant to white flesh and coral red, respectively. However, populations segregating for salmon yellow and red flesh lacked progeny with clear parental phenotypes and ha d a range of intermediate colors. This deviation from a one gene model was suggestive of quantitative trai t s controlling flesh color (Gusmini and Wehner, 2006). Flesh color in melon ( Cucumis melo ) is controlled by both major and minor genes ( Cuevas et al. 2010). In melon the flesh color in both F 2 and BC 1 populatio ns fit an epistatic two gene model. Narrow sense heritabilities of flesh carotene content in these populations were moderate; 0.55 in the F 1 F 2 and BC populations and 0.68 in the F 3 populations. Previous observations suggest in heritance of flesh color of squash is quantitative, influenced by genetic and environmental factors. Thus, knowledge of the heritability and gene action of flesh color genes would give the plant breeder an idea of the population size and time frame needed to manipulate flesh color and thus, nutritional content in squash. To date, there are no available population studies examining the gene number or the inheritance of genes controlling flesh color in squash ( C. moschata and C. pepo ). The objective of this study was to determine heritability and gene action of flesh color measured by L*a*b* color space values in one C. moschata and two C. pepo populations

PAGE 54

54 Materials and Methods Plant M aterial Results from an initial screening of C. moschata and C. pepo cul tigens (Chapter 2) were used to identify parents to cross for population development. Within each species in the preliminary screen, cultigens were selected based on flesh color extremes and similar fruit shape for crossing One cross was made per plant. U nsuccessful crosses were removed before new crosses were attempted. Fruit was harvested 45 days after successful cross completion, and seed was extracted two weeks after the fruit was removed from the plant. Reciprocal crosses were also attempted for al l cross combinations. From successful crosses, two C. pepo [ Table Gold Acorn, TGA, ( yellow orange ) x PI 314806 ( light whitish yellow ) ] and [ Table King Bush, TKB, ( light yellow ) x PI 314806], and one C. moschata [ Butterbush, BB, ( orange red ) x Suc rine DuBerry, SDub, ( dull yellow orange ) ] F 1 individuals were randomly selected and selfed to create three F 2 populations. Additionally, F 1 individuals were reciprocally backcrossed to each parent. All cr osses were made in a greenhouse. Field Trials The field trial used for selecting parents for crossing was described in Chapter 2. F 2 and backcross (BC) screens were conducted 9 April 3 July 2008 and 6 April 4 July 2009, respectively. One hundred seeds (genotypes) for each F 2 and BC population descri bed previously were planted in a completely random design at PSREC Eight seeds and six each of parent 1, parent 2 and the F 1 were also included within each F 2 and BC population, respectively. A maximum of three mature fruit were harvested from each plant for colorimetric readings. F ruit of the F 2 and related BC populations were harvested along the same timeline in 2008 and 2009 (Table 3 1). During the 2008 field

PAGE 55

55 trials, t emperatures recorded by the Florida Automated We ather Network (FAWN) were 24C avg., 39C max., and 2 C min. In 2009, t emperatures recorded by FAWN were 24C avg., 38C max., and 1 C min. over the course of the growing season. Recommended conventional cultural practices and fertility rates for Florida squash were followed for all field tr ials for both 2008 and 2009 (Olson and Simonne, 2007). Colorimetric A nalysis Color was recorded using a Minolta CR 400 Colorimeter (Minolta Camera Co., Ltd., Ramsey, NJ) tristimulus color analyzer, equipped with an 8 mm diameter measuring area and diff use illumination of a 2 Standard Observer. Color measurements include L* a* b*, chroma and hue The L* coordinate indicates darkness or lightness of color and ranges from black (0) to white (100). Coordinates, a* and b*, indicate color directions: +a* i s the red direction, a* is the green direction, +b* is the yellow direction and b* is the blue direction. Chroma is the saturation or vividness of color. As chromaticity increases, a color becomes more intense, and as it dec reases a color becomes duller Hue angle is the basic unit of color and can be interpreted, for example, as 0 = red and 90 = yellow. Both chroma and hue are derived from a* and b* using the follow ing equations: Metric chroma: C [ (a*) 2 + (b*) 2 ] and metric hue a ngle: h = tan 1 (b */a*) [degrees]. Using fruit that had been sliced transversely, L*a*b* color space measurements from the edible flesh (mesocarp) of each fruit were recorded in succession to avoid discoloration. The preliminary study revealed r eplicate measurements of the mesocarp of each fruit per cultigen were not significantly ( P herefore, avoiding the seed cavity and surrounding tissue (approx. 10 mm), three random measurements per fruit were recorded.

PAGE 56

56 Variance, Heritability and Standard Error Estimations by Generation Means Analysis Variances were calculated for each color space value within a generation for each of the nine populations. Additive variances were calculated as ( Warner 1952 ; Wright 1968): 2 A = 2 2 F2 ( 2 BC1 P1 + 2 BC 1P 2 ) (3 1) where 2 F2 equals the total variance present in the F 2 popul ation, and 2 BC1 P1 and 2 BC 1P 2 are the variances in the backcross populations, with the F 1 crossed back to parent 1 and parent 2 of the original population cross, respectively. Environmental variances were calculated as ( Warner 1952 ; Wright 1968): 2 E = ( 1/2) 2 F1 +(1/4) 2 P1 + (1/4) 2 P2 (3 2 ) Phenotypic variances equal the variance among plants of each F 2 population, for each trait (Mahmud and Kramer, 1951): 2 P = 2 F2 (3 3) Genetic variances were calculated as ( Wright 1968): 2 G = 2 P 2 (3 4) where 2 P equals phenotypic variance among the F 2 plants. Broad sense heritabilities were calculated using genetic variance and phenotypic variance: H = 2 G / 2 P (3 5) Narrow sense heritabilities were calculated as (Sleper an d Poehlman 2006): h 2 = 2 A / 2 P (3 6) Standard errors were calculated as (Lynch and Walsh, 1998): SE = Sqrt[Var(V x )], with Var(V x ) (3 7 )

PAGE 57

57 where Var(Vx) equals the variance of the desired parameter. Estimates of Var(V x ) were calculated as: [2[Var( k ) 2 ] / (n+2) ] (3 8) where Var ( k ) is each of the variance components used to calculate V x Standard error calculations for broad sense heritability weres calculated as (Hallauer and Miranda, 1988): SE( H ) = [SE( 2 G )] / 2 P ( 3 9) Estimates of standard error for narrow sense heritability estimates were calculated as (Hallauer and Miranda, 1988): SE( h 2 ) = [SE( 2 A )] / 2 P (3 10 ) Epistatic Interaction E stimat es Epistatic interactions were calculated for each color space wit hin a generation for each of the three crosses and were estimated based on a six parameter non weighted model as described by Hayman (1958) and Gamble (1962). Means ( m ) and additive ( a ), dominance ( d ), additive x additive ( aa ), additive x dominance ( ad ) an d dominance x dominance ( dd ) effects were calculated as: m = F2 (3 11) a = BC1P1 BC1P2 (3 12) d = (1/2) P1 ( 1 /2) P 2 + F1 4 F2 + 2( BC1P1 + BC1P2 ) (3 13) aa = 4 F2 + 2( BC1P1 + BC1P2 ) (3 14) ad = (1/2) P1 + (1/2) P2 + BC1 P1 BC 1P 2 (3 15) dd = P1 + P 2 + 2 F 1 + 4 F2 4( BC1P1 + BC1P2 ) (3 16) All calculations including estimates derived from parent 1 parent 2 and F 1 were calculated using 2008 data after ANOVA revealed significant genotype by yea r interactions present for nearly all color space values all three populations. Variances were calculated as (Gamble, 1962):

PAGE 58

58 m = 2 F2 (3 17) a = 2 BC1P1 2 BC1P2 (3 18) d = (1/4) 2 P1 ( 1 /4) 2 P 2 + 2 F1 16 2 F2 + 4( 2 BC1P1 + 2 BC1P2 ) (3 19) aa = 16 2 F2 + 4( 2 BC1P1 + 2 BC1P2 ) (3 20) ad = (1/4) 2 P1 + (1/4) 2 P2 + 2 BC1 P1 2 BC 1P 2 (3 21) dd = 2 P 1 + 2 P2 + 4 2 F 1 + 16 2 F2 16( 2 BC1P1 + 2 BC1P2 ) (3 22) Standard errors were calculated for each of the six parameters by the sum of the square root of the variances for e ach factor within the equation (Gusmini et al., 2007). Data Analysis Color space values for all F 2 and BC populations were subjected to analysis of variance by the GLM procedure of SAS ( Statistical Analysis System version 9.2, SAS Institute, Cary, NC). Fruit and replicate measurements within a fruit were considered as random eff ects. Least significant differences (LSD) among genotypes were determined at a 5% significance level. Additionally, color space values for parent 1, parent 2, and F 1 used to create each population were analyzed by SAS GLM to assess differences between yea rs. examine all parent 1, parent 2 and F 1 individuals from all three crosses over two years. Generation means standard deviations, and ranges were calculated for each trait using the MEANS proce dure of SAS. Least significant differences among generations within each year were determined at a 5% significance level. All heritability estimates were initially calculated based on a maximum of three fruit per genotype with three color measurements p er fruit within the F 2 BC 1 P1 and BC 1P2 generations. However, ANOVA indicated highly significant differences between fruit within a genotype for nearly all color space values for all populations. Therefore, estimates were also calculated using one fruit pe r genotype due to variable fruit maturity within a genotype. Fruit within a genotype were selected based upon highest (a*, b*,

PAGE 59

59 and chroma ) or lowest (L* and hue ) average replicate measurements within a fruit. Selection of fruit with highest or lowest aver age fruit replicate measurements for each color space value was based upon the direction of each respective color space value correlated with carotenoid levels (Table 2 3) and represent the genetic potential of the genotype examined. Selected single fruit measures of heritability are thus reported in this chapter. Epistatic interactions were also initially calculated using both three and one fruit, with one fruit selection estimates repor ted here Estimates were then tested to determine if values obtaine test. Color space values with the highest correlations to carotenoid levels, a*, b* and Chroma, were also examined for the p resence of transgressive segregants using LSD at a 5% significance level. Additionally, all hue a ngle values were converted from degrees to radians for mean and variance calculation used for heritability and epistatic interaction estimates. Results and Discussion Analysis of variance ( Table Gold Acorn x PI 314806 ) ( C. pepo ) populations For the parents and F 1 significant interactions existed between year and genotype for all color space values. Genotypes were significantly different from one another for all color space values, with the exception of L* ( P = 0.0523). Fruit wi thin a genotype were significantly different from one another for all color space values. Replications within a fruit for all color space values were not significantly different from one another, with P values ranging from 0.9208 (L*) to 0.9990 (a*). Anal yses of variance for the F 2 and the BC 1 populations developed from this cross are summarized in table 3 2.

PAGE 60

60 ( Table King Bush x PI 314806 ) ( C. pepo ) populations For the parents and F 1 significant interactions existed between year and genotype for L*, a*, and b* ( P = 0.0031, P = 0.0039, and P = 0.0489, respectively), while color space values chroma and hue showed no interaction between year and genotype. Genotypes were significantly different for all color space values except L* ( P = 0.1338). Fruit within a genotype was significantly different for all color space values. Replications with in a fruit were not found to be significantly different from one another for all color space values. Analyses of variance for the F 2 and the BC 1 populations developed fro m this cross are summarized in table 3 3. ( Butterbush x Sucrine DuBerry ) ( C. moschata ) populations For the parents and F 1 significant interactions existed between year and genotype for all color space values except for L* ( P = 0 .2214). Genotypes were significantly different from one another for color space values L*, b* and chroma and showed no significant differences for a* ( P = 0.1947) and hue ( P = 0.4544). Fruit within a genotype were significantly different for all color space values ( P < 0.0001 ), however, replications within a fruit were not significantly different for all color space values. Analyses of variance for the F 2 and the BC 1 populations developed from this cross are summarized in table 3 4. L* a* b* Color Space Means and Ranges ( Tab le Gold Acorn x PI 314806 ) ( C. pepo ) populations For color space value L*, PI 314806 had the lightest flesh in both 2008 and 2009, while TGA had the darkest flesh in both years ; the F 1 L* values fell in betwe en both parents (Table 3 5 ). Means of L* for e ach backcross were between the F 1 and the respective backcrossed parent meas ured in the same year (Table 3 6 ). Means of color

PAGE 61

61 space value a* were highest in TGA in both 2008 and 2009, and were higher by over 10 units for all a* mean values. In 2008, mean a* values of the F 1 and F 2 population were lower than PI 314806 by 0.8 units, and in 2009 the mean of the F 1 was lower than PI 314806 by 1.0 unit. This indicates that TGA had more red color than did any subsequent generation or population developed using TGA. Similar to L* mean values, both BC 1 P1 and BC 1P2 a* mean values were between their respective parents. For color space value b*, TGA had the highest values in 2008 and 2009, indicating greater yellow coloration (Table 3 5). TGA was notably higher in chroma than all other s, indicating a more saturated and vivid color. Mean color space values for chroma and b* were nearly identical for all generations, with the largest difference being 0.9 units between the mean of b* for TGA in 2009 and the mean of chroma for TGA in 2009. All generation mean hue angles, with the exception of TGA values in both years were over 90. Hue angle means were highest for F 1 plants in 2008 and 2009, indicating a greener hue. L owest hue angle means were recorded for TGA p lants in both years, indicating a yellow orange hue. ( Table King Bush x PI 314806 ) ( C. pepo ) populations Mean L* values of F 1 and F 2 generations from crosses with TKB were very similar to crosses with TGA, with the exception of backcross populations (T ables 3 5, 3 7). Both backcross means fell outside of the means of their respective parents. Mean a* color space values of F 1 and F 2 were lower than PI 314806 in 2008, and the F 1 generation mean was lower than PI 314806 in 2009. Similar to the F 1 means the BC 1P2 mean a* value was lower than PI 314806 a* mean. This indicates that both the F 1 and the BC 1P2 population had the greenest color.

PAGE 62

62 TKB in both 2008 and 2009 had the highest mean values of b*, indicating the greatest yellow color, while PI 314806 P 2 had the low e st b* mea n values in both years. The highest chroma value was in TKB, while PI 314806 had the lowest value. ( Butterbush x Sucrine DuBerry ) ( C. moschata ) populations L* mean color space values for all C. moschata populations were lower than all generation means for both the TGA and TKB populations, indicating the BB x SDub populations had the darkest flesh (Tables 3 5, 3 6, 3 7, and 3 8). In 2008, mean L* values of the F 2 and the F 1 generation were higher than BB. In 2009, mean L* valu es of the F 1 and the BC 1 generations were also higher than mean values of BB. Mean color space a* values were highest in BB in both years, indicating it had the greatest red color. Means of color space value b* were highest for BB and lowest for SDub in both years, indicating that BB had yellower flesh than SDub. For chroma BB had the highest value, while SDub had the lowest value. Results from b* and chroma mean values indicate that BB was more yellow and more vivid in color than SDub for both years. For both color space values b* and chroma both TGA and TKB were the most yellow and most vivid of the generations in their respective populations. This could be due to the genetic composition of each genotype. Over a dozen genes that affect the rind an d flesh color of squash have been described ( Paris and Brown, 2005) including D ( Dark ), l 1 ( light coloration 1), l 2 ( light coloration 2) and B ( Bicolor ). The effects of these genes were studied wi thin near isogenic lines (NILs) in C. pepo by Tadmor et a l. (2005) In genetic backgrounds that lacked either the dominant D or dominant L 2 alleles, a yellow flesh color developed. In genetic backgrounds with either dominant D or L 2, a yellow orange flesh color developed and when the dominant allele of B

PAGE 63

63 inte racts with the dominant allele of L 2, an intense orange flesh color will occur. The interaction of these genes may be responsible for the increased yellow orange and yellow pigmentation, respectively, of both TGA and TKB as they relate to the other gener ations within their populations. In both years, TGA had a lower hue angle indicating a more orange flesh color, and TKB had a higher hue angle, indicating a more yellow flesh color. White flesh, ( Wf ) in C. pepo is documented as dominant to colored flesh a nd prevents accumulation of yellow in mesocarp tissues (Paris and Brown, 2005). This is evident in the distribution of mean hue angles for F 1 and BC 1P 2 generations in both the TGA and TKB populations. For the BB populations, mean hue angle values were hig hest in the F 2 and F 1 generations in 2008, and BC 1P1 had the highest hue angle in 2009, indicating that these generations had the most yellow orange hue within the population. Additionally, these three generation means had higher hue angle values than di d BB for each respective year. All generation mean color space a*, b*, and hue angles were also lower than both the TGA and TKB populations. This indicates that the C. moschata populations had the most orange flesh. Rationale f or Selecting Single Fruit Color Space Value Measurements Variation for fruit within a genotype is problematic for obtaining accurate heritability and epistatic interaction estimates. This is due to the error variance being confounded within the genotypic variance estimate, and wil l result in inaccurate estimates of heritability and epistatic interactions. Nearly all color space values within each population of each cross were found to be significantly different for fruit within a genotype. This is likely due to the difficulty in obtaining three fruit from an individual plant that were at a similar maturity, as squash has indeterminate flowers that can vary

PAGE 64

64 by several days in maturity along the vine. As flesh color development continues until maturity, this non genetic variation i s problematic for our analyses. Due to this variation, single fruit selections for each color space value within a genotype were made and used for heritability and epistatic interaction analysis. In addition, ANOVA analyses revealed significant genotype year interactions for nearly all color space values for the respective parent 1, parent 2, and F 1 in all crosses when examined separately. Exceptions were chroma ( P = 0.0534) and hue ( P = 0.1910) in the TKB x PI 314806 cross and L* ( P = 0.2214) in the B B x SDub cross. Parent 1, parent 2 and F 1 data for each cross were therefore not able to be combined for 2008 and 2009 and only data from 2008 for P 1 P 2 and F 1 for each cross were used for heritability calculations. When these crosses were examined colle ctively, all color space values had a highly significant genotype year interaction (Table 3 9). To further examine the genotype year interactions, the year variation was tested by using ere grown in both 2008 and 2009 (TGA, TKB, PI 314806, BB, SDub, and the F 1 created from each cross of the former) genotypes for L* (r = 0.97, P = <0.0001), a* (r = 0.97, P = <0.0001), b (r = 0.98, P = <0.0001), chroma (r = 0.97, P = <0.0001), and hue (r = 0.93, P = 0.0002) between years, These results indicate that although the variation in fruit measurements within a genotype were significant, the relative differences between genotype s were not altered when examined in multiple growing seasons. Broad and Narrow Sense Heritability Estimates Estimates of negative broad and narrow sense heritabilities were obtained for some populations. Negative estimates of narrow sense heritability a rose from negative

PAGE 65

65 additive variance estimates, and are to be regarded as zero (Robinson et al. 1955) and all negative estimates are reported as such herein Obtaining negative variance components for heritability calculations may occur due to failure to meet assumptions of the model, the presence of additional data correlations present, or due to sampling error (Thompson and Moore, 1963). In addition, negative heritability estimates can occur if true heritability is low to moderate (0.10 0.25) and the number of observations is not large enough (Gill and Jensen, 1968). All variance estimates used to calculate broad and narrow sense heritability are reported to aid explaining the source(s) of variation causing the negative estimates and prevent reporting bias. ( Table Gold Acorn x PI 314806 ) ( C. pepo ) populations Broad sense heritability for c olor space value L* was good (0.82 ) and similar to the broad sense heritability estimate obtained for hue angle (0.86) (Table 3 10). Narrow sense heritability esti mates were negative for L* and all other color space values in the population with the exception of hue due to large variation in backcross populations to parent 1 for all color space values (Table 3 11) Broad sense heritability was low for color space value a*(0.32) and was the lowest for the TGA populations Color space values b* and chroma had similar broad se nse heritability estimates (0.93 and 0.92 respectivel y) and were the highest broad sense heritability estimates for the TGA populations. Addi tionally, color space value b* had the highest broad sense heritability for all populations from all three crosses (Tables 3 10, 3 12, and 3 14). ( Table King Bush x PI 314806 ) ( C. pepo ) populations Color space value L* had a broad sense h eritability esti mate of 0.59 and a narrow sense heritability estimate of 1.00 (Table 3 12). This was the highest narrow sense heritability estimate for all color space values for all three crosses (Tables 3 10, 3 12,

PAGE 66

66 and 3 14). This indicates that in the TKB cross, cont rol of color space value L* has little environmental control. Color space value a* had the lowest broad sense heritability estimate for the TKB populations and is the lowest reported broad sense heritability estimate for all populations derived from TGA, TKB, and BB crosses (Tables 3 10, 3 12, and 3 14). Color space values b* and chroma had the same broad sense heritability estimate (0.88) and were the highest broad sense heritabilities obtained for the TKB populations. Hue angle had a moderate broad se nse heritab ility (0.67) and a high narrow sense heritability (0.92), which was t he highest estimate for the TKB. All other narrow sense heritability estimates for the TKB populations were considered zero due to negative variance estimates (Table 3 13). ( B utterbush x Sucrine DuBerry ) ( C. moschata ) populations Color space value L* had a broad sense heritability estimate of 0.19 and was the lowest positive broad sense estimate for the BB populations (Table 3 14). Additionally, color space value L* had the highest narrow sense heritability estimate for the BB populations. Color space values a* and hue were the only color space value to have both negative broad and narrow sense heritability estimate for the BB populations due to negative variance additive a nd genetic variance estimates calculated from large variance estimates (Table 3 15) Color space values b* and chroma had similar broad sense heritabilities (0.40 and 0.36, respectively) and moderate narrow sense heritabilities (0.76 and 0.60 respectivel y). Color space value chroma also had the lowest positive narrow sense heritability for all three crosses (Tables 3 10, 3 12, and 3 14). Additionally, narrow sense heritability estimates obtained for b* and chroma were the most similar from all crosses t o narrow sense estimates obtained for beta carotene

PAGE 67

67 associated fruit color in melon ( Cucumis melo ) with F 2 and BC 1 generations, 0.55, and F 3 families, 0.68 (Cuevas et al. 2010). F lesh color heritability estimates obtained in other cucurbit crops, namely cucumber ( Cucumis sativus) (Cuevas et al. 2009), and watermelon ( Citullus lanatus ) (Bang et al. 2010) were able to be evaluated using chi square goodness of fit test for gene segregation ratios. Flesh color in cucumber associated with beta carotene indica ted a two recessive gene model and flesh color in watermelon also indicated control by a two gene model. Data obtained in this study were not able to be easily classified into scorable groups and this calculation method was therefore not an option. Epista tic Gene Effect Estimates ( Table Gold Acorn x PI 314806 ) ( C. pepo ) populations All color space value means (m) were found to be significantly different than zero for all ep istatic gene effects ( P 0.04) (Table 3 16 ) with the exception of hue angle For hue this may be due in part to values being on a smaller scale due to conversion to radians. No significant gene effects were identified for color space value s L* or a*. For color space value b*, the a dditive x additive interaction was significant at a 10% level ( P = 0.09 ). The gene effect was in the direction of TGA, the more yellow parent. This epistatic effect was observed as a larger difference from the F 1 to BC 1P1 means (21.8 units) as compa red to 0.8 units between the F 1 and the BC 1P2 means (Table 3 6). In addition, LSD analysis revealed the F 1 mean was significantly different from the BC 1P1 mean, and was not different from the BC 1P2 mean. Color space value chroma had t he same significant gene effect at the same significance level as did color space value b*. The ef fect was also noted in the direction of TGA, th e more vivid parent (Tables 3 16 ). There was also a larger difference present

PAGE 68

68 between the F 1 mean and the BC 1P1 mean (21.8 and 21.7 units for b* and chroma, respectively) than the F 1 mean and the BC 1P2 mean (0.8 and 0.9 unitis for b* and chroma, respectively). Significant differences were also found between F 1 means and BC 1P1 means (Table 3 6). Hue angle had significant gene e ffects for additive ( P = 0.00), additive x additive ( P = 0.00) and additive x dominan ce ( P = 0.00) interactions. All gene effects were toward the parent with the lower hue angle, TGA D ifferences were smaller between F 1 and BC 1P2 (1.5 units) than F 1 and BC 1P1 (7.4 units) (Table 3 6). Additionally, t he mean of PI 314806 was not significantly different than F 1 or BC 1P2 when examined by LSD analysis Among these genotypes, additive effects were involved in control of hue angle color space values (Tables 3 16). ( Table King Bush x PI 314806 ) ( C. pepo ) populations All color space value means (m) were found to be significantly different than zero, for all epistatic gene effect estimate s ( P 0.02) (Table 3 17 ) with the exception of hue angle No significant gene effects were noted for color space values L* and a*. Additive and a dditive x additive gene effects were also significant for color space values b* and chroma at the 10% significance level ( P = 0.09 and 0.07 for each effect, respectively, for both b* and chroma ), both in the direction of TKB the more yellow and vivid parent (Table 3 17 ). Differences between F 1 and BC generations were similar for both b* and chroma with differences of 1.7 and 1.8, respectively, between F 1 and BC 1P2 and 17.8 and 17.9, respectively, between F 1 and BC 1 (Table 3 7 ). In addition, both color space values showed significant differences between the F 1 and the BC 1 P1 means and not between the F 1 and BC 1P2 means, displaying the effect of additive gene effects for both b* and ch roma color space values.

PAGE 69

69 Significant dominance gene effects ( P = 0.00) in the direction of the parent with the higher hue angle, PI 314806 and additive x dominance ( P = 0.00) and dominance x dominance ( P = 0.00) gene effects in the direction of parent with the lower hue angle, TKB, were identified in hue angle (Table 3 1 7 ). Similar to the distribution of generation means for F 1 BC 1 P1 and BC 1P2 populations in th e TGA x PI 314806 cross the TKB population s also had a smaller difference between the F 1 a nd the BC 1P2 population than between the F 1 and the BC 1 P1 population However unlike the TGA cross the mean of F 1 was significantly different than the mean of P 2 but not different than BC 1P2 in LSD analysis Therefore dominance effect s in the directio n of PI 314806 are not observed with the present data. However, the F 1 is significantly different than BC 1P1 suggesting the presence of both additive x dominance and dominance x dominance effects. Additionally, the only gene effects found to be sign ificant at the 5% significance level within thi s population were for hue angle. ( Butterbush x Sucrine DuBerry ) ( C. moschata ) populations All color space value means (m) were found to be significantly different than zero, for all epistatic gene effect estimate calculations ( P = 0.00) (Table 3 1 5 ) with the exception of hue angle Dominance x dominance interaction was found to be significant ( P = 0.00) for color space value L* in the direction of the darker parent, SDub Differences from F 1 to BC 1P2 and from F 1 to BC 1 P1 means were almost equal (1.1 and 1.0 units, respectively) (Table 3 8), and the F 1 mean was significantly different than both BC population means. However, the mean of the F 1 is higher than BB in both 2008 and 2009, and the means of the F 2 and BC 1 P1 generations, are higher th an BB in each respective year. Therefore the present data do not provide evidence of a dominance effect in the direction of SDub.

PAGE 70

70 Color space value b had significant dominance x dominance gene effects at the 10% sig nificance level ( P = 0.08) (Table 3 15). The gene effect in b* was in the direction of the less yellow and less vivid p arent, SDub. Distance from the F 1 to BC 1P1 mean in b* was 2.1 units, and distance from the F 1 to BC 1P2 mean was 4.3 units (Table 3 8). However, all genotypic means were significantly different from one another in LSD analysis wit hin the backcross populations, and BC1P1 was closer to BB (4.5 units) than BC1P2 was to SDub (5.9 units). The data does not support dominance x dominance intera ction. No significant gene interactions were noted for color space value chroma Hue angle had significant dominance x dominanc e gene effects ( P = 0.00) in the direction of the parent with the lower hue angle, SDub (Table 3 18 ). A greater difference was observed between means of the F 1 and BC 1 P1 population (4.0 units) than between the F 1 and BC 1P2 population (0.3 units). Additionally, the mean of the F 1 was significantly different from BC 1 P1 but not different from BC 1P2 or parent 2, SDub, as determined by LSD analysis. Th ese results support a dominance gene effect in the direction of SDub. Overall, only dominance x dominace gene effects were found to be significant within this population. Additionally, the t ype of epistatic interaction could not be de t ermined for the BB cross, but could be determined for Hue angle in both the TGA and TKB crosses Interaction type is determined by comparing signs of dominance (d) and dominance x dominance (dd) interactions; like signs indicate complementary epistasis wh ile unlike signs indicate duplicate epistasis (Kearsey and Pooni, 1996). It is also a requirement that both dominance (d) and dominance x dominance (dd) interactions must be significant to determine epistasis type which was only met in the

PAGE 71

71 TGA and TKB cro sses For both crosses, duplicate epista sis was observed for hue angle, at the 10% significance level in the TGA cross and at the 5% significance level in the TKB cross (Tables 3 14 and 3 16). Breeding for Increased Flesh Color in Winter Squash Transgres sive segregation was examined using LSD analysis for fruit(s) with the highest average replica te measurements for a*, b*, and chroma color space values for each plant of F 2 and BC 1 generati ons of all three populations. Transgressive segregation was examine d only for these color space values due to the strength of correlations observed between these color space values and carotenoid content (Chapter 2, Table 2 3). No transgressive segregation was observed for color space values a*, b*, and chroma for the TGA populations examined. TGA had the highest of the aforementioned color space values among both F 2 and BC 1 populations derived from the genotype. This lack of transgressive segregation could be a reflection of limited homogenous data collected for each geno type, the necessity for a larger number of individuals within the population, or the relatedness of the parents. These data suggest that individuals within these populations cannot be used to exceed the values of their respective parental generations, in either color or carotenoid content. However, reported correlations between L* a* b* color space values and carotenoid levels could be strengthened by keeping replications per location separate and examining more samples with HPLC analysis. Correlation be tween color and carotenoid content could then be re examined. Strengthening correlations may provide additional color space values to examine for transgressive segregants as they relate to carotenoid content. Transgressive segregation was observed for co lor space values a* in the BC 1P1 population of the TKB cross. Additionally, transgressive segregation was observed in

PAGE 72

72 the BB cross for color space values a* and b* in the F 2 population, and for color space values a* and b* in the BC 1P1 population. By sel ecting these individuals that are significantly higher than the parent 1 means (TKB and BB, respectively), it is possible to rapidly improve color and thereby carotenoid content in squash through plant breeding. Although there was an absence of transgressi ve genotypes in the TGA and TKB populations for color space values b* and chroma there is the potential to shift population means in the direction of the highest value parent ( TG A and TKB, respectively) in all populations for these c olor space values. Th is is illustrated by the additive gene effects of these traits and distribution of generation means. In addition, high broad sense heritabilities for these color space values for both populations suggest breeding for color as it relates to nutritional cont ent can be done successfully. Transgressive segregation was obs erved for color space values b* and chroma for the F 2 population and a and b* for the BC 1P1 population of the BB cross. Data suggest that populations within this cross have the highest poten tial for genetic improvement for these three color space values. However, of the three crosses, the BB cross has the lowest heritabilities for these three color space values. This indicates that while the potential for improvement is there, it would be th e most difficult of the three crosses to breed. T o obtain more accurate heritability estimates, more measurements per genotype could be used if fruit maturity was controlled. Additionally, beta carotene levels in some muskmelon cultivars have been shown to be influenced by soil type in different locations (Lester and Eischen, 1995). This suggests that other carotenoids controlling color in cucurbit crops may also be affected by different environments. Planting F 2 and BC

PAGE 73

73 generations in either multiple y ears and/or multiple locations may enable partitioning out the environmental variation and aid in providing better estimates of heritability and epistatic gene effects.

PAGE 74

74 Table 3 1. Harvest outline for all populations in two Cucurbita pepo and one Cucurbi ta moschata families in 2008 and 2009 a Number proceeding parentheses indicates P 1 P 2 and F 1 plants from which f ruit were harvested in the 2008 season. Number in parentheses indicates P 1 P 2 and F 1 plants from which fruit were harvested in the 2009 season ; this n umber is a combination of plants within both BC 1 P1 and BC 1 P2 populations. b Number proceeding parenthese s indicates P 1 P 2 and F 1 fruit harvested from respective plants in the 2008 season. Number in parentheses indicates P 1 P 2 and F 1 fruit harvested from respective plants in the 2009 season; this number is from the combination of plants within both BC 1 P1 and BC 1 P2 populations. Genotype Plant Number a Fruit Number b Harvest after Planting (weeks) Table Gold Acorn (TGA) 8 (9) 24 (27) 10 PI 314806 5(7) 15 (19) 10 F 1 (TGA x PI 314806) 8 (10) 24 (30) 10 F 2 (TGA x PI 314806) 96 276 10 BC 1 P1 (TGA x F 1 ) 74 210 10 BC 1 P2 (F 1 x PI 314806) 91 260 10 Table King Bush (TKB) 8(10) 22 (26) 11 PI 314806 6 (7) 16 (18) 11 F 1 (TKB x PI 314806) 8 (11) 22 (32) 11 F 2 (TKB x PI 314806) 91 260 11 BC 1 P1 ( F 1 x TKB ) 75 219 11 BC 1 P2 ( F 1 x PI 314806 ) 84 246 11 Butterb ush (BB) 8 (12) 21 (36) 12 Sucrine DuBerry (SDub) 8 (11) 23 (29) 12 F 1 (BB x SDub) 8 (10) 23 (30) 12 F 2 (BB x SDub) 86 249 12 BC 1 P1 ( F 1 x BB ) 67 201 12 BC 1 P2 ( F 1 x SDub ) 88 232 12

PAGE 75

75 Table 3 2 Examination of fixed effects due to genotypes and expected mean squares due to random effects in three populations of Cucurbita pepo derived from Table Gold Acorn (TGA) x PI 314806. Sources Df L* a* b* Chroma Hue F 2 ( TGA x PI 314806) Test of fixed effects (F values) Genotype a 95 3.08 *** 14.26 *** 49.76 *** 50.35 *** 12.46 *** Expected mean squares Fruit(Genotype) b 178 15.51 *** 0.63 *** 11.63 *** 11.50 *** 6.27 *** Rep(Fruit) b 6 3.87 0.45 2.86 3.12 2.89 Residual 546 2.55 0.31 2.5 3 2.65 1.20 BC 1 P1 (TGA x F 1 ) Test of fixed effects (F values) Genotype a 70 6.34 *** 24.31 *** 42.57 *** 42.50 *** 33.25 *** Expected mean squares Fruit(Genotype) b 137 17.13*** 6.49*** 52.13*** 52.17*** 11.15*** Rep(Fruit) b 6 2. 34* 3.72 70.53* 69.19* 9.58 Residual 414 3.91 3.20 25.21 25.18 4.86 BC 1P2 (F 1 x PI 314806) Test of fixed effects (F values) Genotype a 90 1.84*** 5.15 *** 4.56 *** 4.57 *** 4.93 *** Expected mean squares Fruit(Genotype) b 167 11.44 0.24*** 5.39* ** 5.37*** 3.95*** Rep(Fruit) b 6 21.82 0.2183* 3.88** 3.98** 1.23* Residual 514 3.44 43.96 1.28 1.32 0.62 a Residual fix ed effects: MS(Fruit(G eno type ) ). b Residual: MS(Residual). *Significant at the 0.05 probability level. **Significant at the 0.01 probab ility level. ***Significant at the 0.001 probability level.

PAGE 76

76 Table 3 3 Examination of fixed effects due to genotypes and expected mean squares due to random effects in three populations of Cucurbita pepo derived from Table King Bush (TKB ) x PI 314806. Sources Df L* a* b* Chroma Hue F 2 ( TKB x PI 314806) Test of fixed effects (F values) Genotype a 90 7.47*** 11.38*** 39.77*** 39.68*** 18.50*** Expected mean squares Fruit(Genotype) b 167 10.40*** 0.66 11.36*** 11.26*** 5.15*** Rep(Fruit) b 6 2.52 0.20 2.19 2.29 0.86 Residual 514 3.44 0.13 2.60 2.63 0.88 BC 1 P1 ( F 1 x TKB ) Test of fixed effects (F values) Genotype a 74 4.47*** 6.17*** 63.31*** 63.27*** 16.40*** Expected mean squares Fruit(Genotype) b 142 10.50*** 1.98*** 19.03*** 18.79 *** 5.60*** Rep(Fruit) b 6 3.54 0.17 3.48 3.52 0.39 Residual 432 4.14 0.39 5.95 5.78 1.01 BC 1P2 (F 1 x PI 314806) Test of fixed effects (F values) Genotype a 83 21.33*** 4.09*** 3.98*** 3.96*** 4.48*** Expected mean squares Fruit(Genotype) b 160 10.75*** 0.28*** 3.25*** 3.31*** 2.81*** Rep(Fruit) b 6 9.89* 0.54*** 4.10* 3.93* 12.55*** Residual 486 3.64 0.11 1.80 1.81 1.87 a Residual fixed effects: MS(Fruit(G eno type ) ). b Residual: MS(Residual). *Significant at the 0.05 probability level. **Sign ificant at the 0.01 probability level. ***Significant at the 0.001 probability level.

PAGE 77

77 Table 3 4 Examination of fixed effects due to genotypes and expected mean squares due to random effects in three populations of Cucurbita pepo derived from Butterbus h (B B ) x Sucrine DuBerry (SDub) Sources Df L* a* b* Chroma Hue F 2 ( BB x SDub ) Test of fixed effects (F values) Genotype a 85 10.43*** 11.48*** 9.07*** 9.07*** 11.10*** Expected mean squares Fruit(Genotype) b 161 4.42*** 9.39*** 16.78*** 17.23* ** 7.58*** Rep(Fruit) b 6 1.94 5.88*** 5.77 3.48 5.90*** Residual 492 1.63 1.31 3.00 2.75 1.17 BC 1 P1 ( F 1 x BB ) Test of fixed effects (F values) Genotype a 67 4.56*** 9.68*** 10.65*** 11.50*** 9.41*** Expected mean squares 15.19*** Fruit(Geno type) b 134 6.22*** 14.84*** 15.05*** 15.19*** 11.52*** Rep(Fruit) b 28 1.57* 2.61** 2.94 3.40 1.99* Residual 402 1.02 1.42 2.56 2.52 1.13 BC 1P2 (F 1 x SDub ) Test of fixed effects (F values) Genotype a 4.91*** 6.79*** 10.12*** 11.40*** 5.46*** Expected mean squares Fruit(Genotype) b 5.99*** 18.42*** 18.27*** 6.27*** 15.44 Rep(Fruit) b 1.59 3.16*** 4.34 1.28 2.90** Residual 1.52 1.57 3.31 3.04 1.58 a Residual fixed effects: MS(Fruit(G eno type ) ). b Residual: MS(Residual). *Significant at the 0 .05 probability level. **Significant at the 0.01 probability level. ***Significant at the 0.001 probability level.

PAGE 78

78 Table 3 5 Mean s, standard deviations (SD) and ranges of colorimetric values (L*, a*, b*, chroma and hue ) a of individuals from measured in two Cucurbita pepo and one Cucurbita moschata families L a* b* Genotype b n Mean c (SD) Range Mean (SD) Range Mean (SD) Range Table Gold Acorn (TGA) 8(24) d 76.7 (2.7) 66.0 82.0 8.3 (2.7) 0.9 15.7 61.8 (7.3) 44.4 74.6 PI 314806 5 (15) 82.7 (2.3) 77.8 87.7 2.0 (0.3) 2.4 1.4 8.9 (1.3) 6.7 13.2 F 1 (TGA x PI 314806) 8(24) 80.5 (1.8) 76.9 84.5 2.8 (0.5) 4.3 1.6 12.3 (2.1) 9.3 17.0 F 2 (TGA x PI 314806) 96(276) 81.3 (3.3) 69.4 88.8 2.8 (1.2) 8.7 0.4 17 .1 (8.4) 5.1 44.7 Table King Bush (TKB) 8(22) 80.4 (2.3) 75.8 85.1 0.6 ( 1.7) 3.0 3.6 49.4 (5.1) 40.3 60.1 PI 314806 6(16) 83.7 (2.0) 78.1 87.4 2.1 (0.3) 2.8 1.5 9.1 (1.7) 6.8 13.9 F 1 (TKB x PI 314806) 8(22) 81.3 (3.0) 73.9 87.2 2.4 (0.4) 3.4 1.7 12.1 (2.1) 7.4 18.0 F 2 (TKB x PI 314806) 91(260) 81.4 (3.7) 68.6 90.6 2.7 (1.1) 7.1 0.9 18.9 (7.5) 7.1 43.6 Butterbush (BB) 8(21) 67.4(3.3) 59.2 73.2 22.4(3.6) 15.0 30.9 66.9(3.4) 58.4 75.2 Sucrine Du Berry (SDub) 8(23) 65.1(2.4) 61.3 72.3 18.6(4.3) 6.5 24.2 54.2(3.4) 47.1 64.5 F 1 (BB x SDub) 8(23) 68.9(2.1) 65.4 75.0 18.4(4.3) 2.7 22.3 61.5(3.8) 50.5 72.3 F 2 (BB x SDub) 86 (249) 68.3(2.7) 60.3 78.2 18.1(3.9) 1.2 28.4 62.1(4. 8) 49.7 77.8

PAGE 79

79 Table 3 5 Continued Chroma Hue Genotype b n Mean (SD) Range Mean (SD) Range Table Gold Acorn (TGA) 8(24) d 62.4 (7.5) 44.4 75.6 82.6 (2.4) 76.5 88.8 PI 314806 5(15) 9.1 (1.3) 6.9 13.4 102.7 (1.5) 99.9 105.1 F 1 (TGA x PI 314806) 8(24) 12.7 (2.1) 9.5 17.5 103.1 (1.7) 96.7 106.7 F 2 (TGA x PI 314806) 96(276) 17.4 (8.4) 5.2 44.7 100.2 (3.3) 90.7 110.7 Table King Bush (TKB) 8(22) 49.4 (5.1) 40.3 60.2 90.9 (1.9) 86.5 93.9 PI 314806 6(16 ) 9.3 (1.7) 7.0 14.1 103.4 (1.8) 99.6 106.6 F 1 (TKB x PI 314806) 8(22) 12.4 (2.0) 7.6 18.1 101.6 (2.3) 95.4 104.9 F 2 (TKB x PI 314806) 91(260) 19.0 (7.5) 7.3 43.6 99.1 (3.6) 88.7 108.5 Butterbush (BB) 8(21) 70.1(2.6) 63.2 76.8 71.5 (3.4) 64.2 78.7 Sucrine DuBerry (SDub) 8(23) 57.5(3.6) 50.5 67.2 71.1(4.2) 66.7 83.1 F 1 (BB x SDub) 8(23) 64.3(4.3) 50.7 75.1 73.5(3.6) 70.5 86.8 F 2 (BB x SDub) 86 (249) 64.8(4.8) 50.2 79.8 73.8(3.5) 66.3 91.3 a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a c olor becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b C. pepo families include TGA x PI 314806, and TKB x PI 314806. BB x SDub is the C. moschata family. c Values shown are based on 3 colorimetric measurements per fruit. d plants.

PAGE 80

80 Table 3 6 Mean s, standard deviations (SD ), and ranges of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table Gold Acorn (TGA) x PI 314806 ( Cucurbita pepo ) BC populations L a* b* Genotype n Mean b (SD) Range Mean (SD) Range Mean (SD) Range Table Gold Acorn (TGA) 9(27 ) c 72.9(3.6) 61.4 80.5 11.1(4.0) 0.2 19.5 68.5(5.4) 54.7 79.1 PI 314806 7(19 ) 81.9(2.8) 74.0 87.2 2.5(0.5) 4.2 1.9 11.3( 1.9) 7.8 16.7 F 1 (TGA x PI 314806) 10(30) 80.3(2.4) 73.2 86.9 3.5(0.6) 5.8 2.2 15.4(2.0) 11.6 21 .0 BC 1 P1 (TGA x F 1 ) 74(210) 79.4(4.3) 66.2 87.7 1.9(4.6) 11.7 15.2 37.2(16.6) 9.8 77.0 BC 1 P2 (F 1 x PI 314806) 91(260) 80.9(2.7) 7.7 87.2 2.9(0.5) 5.0 1.5 14.6(2.2) 8.2 21.3

PAGE 81

81 Table 3 6 Continued a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of co lor) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Values shown are based on 3 colorimetric measurements per fruit. c plants. Chroma Hue Genotype n Mean b (SD) Range Mean (SD) Range Table Gold Acorn (TGA) 9(27) c 69.4(5.6) 53.7 80.8 80.9(3.1) 74.5 90.2 PI 314806 7(19) 11.6(1.9) 8.1 17.2 102.5(1.4) 99.2 105.1 F 1 (TGA x PI 314806) 10(30) 15.8(2.0) 11.9 21.8 102.8(2.1) 98.0 107.1 BC 1 P1 ( TGA x F 1 ) 74(210) 37.5(16.6) 10.1 78.4 95.4(6.9) 77.9 109.5 BC 1 P2 (F 1 x PI 314806) 91(260) 14.9(2.2) 8.5 21.5 101.3(1.9) 96.0 108.0

PAGE 82

82 Table 3 7 Mean s, standard deviations (SD), and ranges of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table King Bush (TKB ) x PI 314806 ( Cucurbita pepo ) BC populations L a* b* Genotype n Mean b (SD) Ra nge Mean (SD) Range Mean (SD) Range Table King Bush ( TKB ) 10(26) 80.5(2.3) 72.9 83.8 0.4(1.6) 3.5 6.3 53.1(3.6) 44.8 61.9 PI 314806 7(18) 81.5(2.6) 3.7 1.6 2.4(0.5) 3.8 1.6 11.2(1.4) 8.6 14.7 F 1 (TKB x PI 314806) 11(32) 80 .7(2.8) 73.3 86.9 2.9(0.5) 5.2 1.7 14.9(1.6) 12.0 18.8 BC 1 P1 ( F 1 x TKB ) 75(219) 81.2(3.2) 68.5 89.5 3.2(1.4) 7.6 3.8 32.7(12) 11.1 59.7 BC 1 P2 ( F 1 x PI 314806 ) 84(246) 80.3(2.7) 70.6 88.2 2.5(0.5) 5.9 1.0 13.2(1.8) 8.8 20 .5

PAGE 83

83 Table 3 7 Continued a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Values shown are based on 3 colorimetric measurements per fruit. c plants. Chroma Hue Genotype n Mean b (SD) Range Mean (SD) Range Table King Bush ( TKB ) 10(26) 53.1(3.6) 44.9 61.7 90.5(1.7) 84.1 94.0 PI 314806 7(18) 11.4(1.5) 8.8 15.0 102.3(1.3) 99.4 105.2 F 1 (TKB x PI 314806) 11( 32) 15.2(1.6) 12.2 19.5 100.9(1.6) 95.8 105.3 BC 1 P1 ( F 1 x TKB ) 75(219) 33.1(11.9) 11.3 59.7 96.6(3.5) 85.5 105.4 BC 1 P2 ( F 1 x PI 314806 ) 84(246) 13.4(1.9) 9.1 21.3 100.8(1.8) 93.7 106.0

PAGE 84

84 Table 3 8 Mean s, standard deviations (SD), and ranges of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Butterbush (BB) x Sucrine DuBerry (SDub) ( Cucurb ita moschata ) BC populations L a* b* Genotype n Mean b (SD) Range Mean (SD) Range Mean (SD) Range Butterbush ( BB ) 12(36) 69.6(2.2) 64.9 76.0 16.8(4.5) 2.8 26.4 69.0(3.4) 59.1 77.7 Sucrine DuBerry (SDub) 11(29) 66.9(1.9) 63.6 72 .4 14.5(4.5) 5.0 21.8 53.2(4.1) 38.7 60.4 F 1 (BB x SDub ) 10(30) 69.8(1.9) 65.4 74.0 16.6(3.2 ) 9.3 25.2 63.4(3.1) 56.7 69.8 BC 1 P1 ( F 1 x BB ) 67(201) 70 .8 (2.3) 64.6 80.0 12.5(4.5) 5.5 22.9 65.5(4.4) 43.3 76.1 BC 1 P2 ( F 1 x SDub ) 88(2 32) 68.7(2.4) 62.0 77.1 15.2(4.3) 1.2 24.1 59.1(4.6) 42.8 71.0

PAGE 85

85 Table 3 8 Continued a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividnes s) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Values shown are based on 3 colorimetric measuremen ts per fruit. c plants. Chroma Hue Genotype n Mean b (SD) Range Mean (SD) Range Butterbush ( BB ) 12(36) 71.2(3.4) 59.2 79.8 76.3(3.7) 68.4 87.3 Sucrine DuBerry (SDub) 1 1(29) 55.3(4.6) 41.6 63.4 74.9(4.2) 68.4 84.1 F 1 (BB x SDub ) 10(30) 65.6(3.3) 58.1 72.5 75.4(2.5) 69.0 81.5 BC 1 P1 ( F 1 x BB ) 67(201) 66.8(4.7) 43.4 77.0 79.4(3.9) 70.3 96.7 BC 1 P2 ( F 1 x SDub ) 88(232) 61.2(5.0) 43.4 72.4 75.7(3.7) 68. 0 88.6

PAGE 86

86 Table 3 9 E xpected mean squares due to ran dom effects in respective parent 1, parent 2, and F1 of three Cucurbit a populations resulting from C. pepo Table Gold Acorn (TGA) x PI 314806, C. pepo Table King Bush (TKB) x PI 314806, and C. moschata Butterbush (BB) x Sucrine DuBerry (SDub) over 2008 and 2009. Sources Df L* a* b* Chroma Hue Expected mean squares Year a 1 34.57 420.79 1733.45** 1320.57* 172.73 Genotype b 8 5501.41*** 12274*** 81560 *** 86997*** 22719*** Genotype x Year c 8 116.77*** 189.77*** 132.44*** 215.85*** 164.88*** Fruit(Genotype) c 203 13.52*** 22.04*** 24.62*** 26.61*** 21.33*** Rep(Fru it) c 72 2.40 2.00 5.50 5.73 1.62 Residual 983 4.89 5.04 9.98 10.51 4.51 a Residual(Year): 0.9761*MS(Year*G eno type)+ 0.0239*MS(Residual ) b Residual(Genotype): 0.9964*MS( Year*G eno type )+ 0.9012*MS(Fruit(G eno type)) 0.8976*MS( Residual ) c Residual : MS( Residual ). *Significant at the 0.05 probability level. **Significant at the 0.01 probability level. ***Significant at the 0.001 probability level.

PAGE 87

87 Table 3 10 Genetic means estimates for heritability of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table Gold Acorn (TGA) x PI 314806 ( Cucurbita pepo ) populations with one fruit per genotype b a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) range s from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes m ore dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respecti vely. Frui t with lowest L* and hue due correlation direction of all measured carotenoids. c 2 A 2 P 2 E 2 G H h 2 are additive variance = 2 2 F2 ( 2 BC1 P1 + 2 BC 1P 2 ), phenotypic variance = 2 F2 environmental variance = (1/2) 2 F1 +(1/4) 2 P1 + (1/4) 2 P2 genotypic variance broad sense heritability narrow sense heritability estimate = 2 A / 2 P respectively. d Values converted from degrees to radians for calculation Genetic para meter c L SE a* SE b* SE Chroma SE Hue d SE 2 A 0.00 --0.00 --0.00 --0.00 --0.0000 --2 P 7.93 2.82 0.98 0.99 77.29 8.79 77.11 8.78 0.0036 0.0596 2 E 1.45 0.53 0.67 0.29 5.69 1.82 5.79 1.86 0.0005 0.0001 2 G 6.48 1.25 0.32 0.32 71.60 11.19 71.32 11.17 0.0031 0.0005 H 0.82 0.16 0.32 0.32 0.93 0.14 0.92 0.14 0.86 0.15 h 2 0.00 --0.00 --0.00 --0.00 --0.00 --

PAGE 88

88 Table 3 11. Variance estimates for calculating heritability of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table Gold Acorn (TGA) x PI 314806 ( Cucurbita pepo ) populations with one fruit per genotype b a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intens ity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fruit with lowest L* and hue due correlation direction of all measured carotenoids. c Values converted from degrees to radians for calculation. Variance estimate L* a* b* Chroma Hue c 2 P1 4.56 2.55 14.51 14.74 0.0005 2 P2 0.13 0.01 0.90 0.84 0.0003 2 F1 0.56 0.05 3.68 3.79 0.0006 2 F2 7.93 0.98 77.29 77.11 0.0036 2 BC1P1 13.55 22.88 285.79 285.62 0.0869 2 BC1P2 3.55 0.17 3.46 3.42 0.0009

PAGE 89

89 Table 3 12 Genetic means estimates for heritability of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table King Bush (TKB) x PI 314806 ( Cucurbita pepo ) populations with one fruit per genotype b a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fruit with lowest L* and hue due correlation dire ction of all measured carotenoids. c 2 A 2 P 2 E 2 G H h 2 are additive variance = 2 2 F2 ( 2 BC1P1 + 2 BC1P2 ), phenotypic variance = 2 F2 environmental variance = (1/2) 2 F1 +(1/4) 2 P1 + (1/4) 2 P2 genotypic variance, broad sense heritability, narrow sense heritability estimate = 2 A / 2 P respectively. d Values converted from degrees to radians for calculation. Genetic parameter c L SE a* SE b* SE Chroma SE Hue d SE 2 A 13.21 3.50 0.00 --0.00 --0.00 --0.0036 0.0013 2 P 11.33 3.37 1.04 1.02 60.97 7.81 60.14 7.75 0.0039 0.0628 2 E 4.59 1.44 0.92 0.39 7.15 2.56 7.11 2.57 0.0014 0.0004 2 G 6.74 2.20 0.12 0.42 53.82 9.30 53.03 9.19 0.0026 0.000 7 H 0.59 0.19 0.12 0.40 0.88 0.15 0.88 0.15 0.66 0.18 h 2 1 .00 --0.00 --0.00 --0.00 --0.92 0.00

PAGE 90

90 Table 3 1 3 Variance estimates for calculating heritability of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table King Bush (TKB) x PI 314806 ( Cucurbita pepo ) populations with one fruit per genotype b a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intens ity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fruit with lowest L* and hue due correlation direction of all measured carotenoids. c Values converted from degrees to radians for calculation. Variance estimate L* a* b* Chroma Hue c 2 P1 3.83 3.49 22.34 22.50 0.0012 2 P2 2.57 0.03 2.01 2.01 0.0006 2 F1 5.98 0.08 2.13 1.96 0.0018 2 F2 11.33 1.04 60.97 60.14 0.0039 2 BC1 P1 6.00 1.99 148.60 146.15 0.0036 2 BC1P2 3.45 0.11 2.00 2.05 0.0006

PAGE 91

91 Table 3 14 Genetic means estimates for heritability of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Butterbush (BB) x Sucrine DuBerry (SDub) ( Cucurbita moschata ) populations with one fruit per genotype b a Numeric description of color using L* a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes m ore intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fruit with lowest L* and hue due correlation direction of all measured carotenoids. c 2 A 2 P 2 E 2 G H h 2 are additive variance = 2 2 F2 ( 2 BC1P1 + 2 BC1P2 ), phenotypic variance = 2 F2 environmental variance = (1/2) 2 F1 +(1/4) 2 P1 + (1/4) 2 P2 genotypic variance, broad sense heritability, narrow sense heritability estimate = 2 A / 2 P respectively. Genetic parameter c L* SE a* SE b* SE Chroma SE Hue SE 2 A 4.83 1.92 0.00 --17.09 7.47 13.46 7.64 0.0000 --2 P 5.84 2.42 13.06 3.61 22.48 4.74 22.48 4.74 0.0029 0.0540 2 E 4.76 1.39 20.40 6.15 13.44 3.80 14.37 4.75 0.0048 0.0014 2 G 1.08 1.64 0.00 --9.05 5.09 8.12 5.84 0.0000 --H 0.19 0.28 0.00 --0.40 0.23 0.36 0.26 0.00 --h 2 0.83 0.33 0.00 --0.76 0.33 0.60 0.34 0.00 --

PAGE 92

92 Table 3 15. Variance estimates for calculating heritability of colorimetric values (L*, a*, b*, chroma and hue ) a measured in Butterbush (BB) x Sucrine DuBerry (SDub) ( Cucurbita moschata ) populations with one fruit per genotype b a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0 ). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overa ll highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fruit with lowest L* and hue due correlation direction of all measured carotenoids. c Values converted from degrees to rad ians for calculation. Variance estimate L* a* b* Chroma Hue c 2 P1 11.04 12.13 12.94 4.93 0.0027 2 P2 3.99 19.53 11.57 12.08 0.0060 2 F1 2.00 24.97 14.62 20.23 0.0053 2 F2 5.84 13.06 22.48 22.48 0.0029 2 BC1P1 2.96 16.29 12.48 13.34 0.0035 2 BC1P2 3.89 13.82 15.40 18.16 0.0029

PAGE 93

93 Table 3 16 Epistatic gene effect estimates for colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table Gold Acorn (TGA) x PI 314806 ( Cucurbita pepo ) populations with one fruit per genotype b Gene effect c L* SE P v alue a* SE P value b* SE P value m 79.54 2.82 0.00 2.48 0.99 0.01 18.47 8.79 0.04 a 1.88 4.14 0.65 1.91 4.80 0.69 24.78 17.01 0.15 d 3.85 14.04 0.78 3.36 10.42 0.75 13.73 49.00 0.78 aa 3.06 8.80 0.73 3.51 9.64 0.72 59.77 34.87 0.09 ad 0.79 4.27 0.85 3.91 4.87 0.42 2.68 17.12 0.88 dd 5.96 20.19 0.77 6.44 19.68 0.74 49.17 76.78 0.52

PAGE 94

94 Table 3 1 6 Continued a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dul l. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fr uit with lowest L* and hue due correlation direction of all measured carotenoids. c Gene effects were calculated using calculations described by Hayman (1958) and Gamble (1962). Gene effect c Chroma SE P value Hue SE P value m 18.74 8.78 0.04 1.74 3.42 0.61 a 24.81 17.00 0.15 1.04 0.30 0.00 d 13.77 48.97 0.78 1.05 0.64 0.10 aa 59.58 34.88 0.09 2.21 0.60 0.00 ad 2.87 17.12 0.87 1.66 0.30 0.00 dd 48.90 76.74 0.53 2.27 1.21 0.06

PAGE 95

95 Table 3 17 Epistatic gene effect estimates for colorimetric values (L*, a*, b*, chroma and hue ) a measured in Table King Bush (TKB) x PI 314806 ( Cucurbita pepo ) populations with one fruit per genotype b Gene effect c L* SE P value a* SE P value b* SE P value m 79.87 3.37 0.00 2.39 1.02 0.02 20.08 7.81 0.01 a 0.84 3.07 0.79 0.39 1.45 0.79 21.13 12.27 0.09 d 4.00 15.06 0.79 1.55 5.10 0.76 0.67 39.83 0.99 aa 0.49 11.55 0.97 0.95 3.11 0.76 45.90 25.23 0.07 ad 2.39 3.32 0.47 1.39 1.73 0.43 0.40 12.52 0.97 dd 5.04 19.05 0.79 3.74 7.36 0.61 29 .35 58.46 0.62

PAGE 96

96 Table 3 17 Continued a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) as chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fruit with lowest L* and hue due correlation direction of all measured carotenoids. c Gene effects we re calculated using calculations described by Hayman (1958) and Gamble (1962). Gene effect c Chroma SE P value Hue SE P value m 20.32 7.75 0.01 1.71 3.60 0.63 a 21.09 12.17 0.09 0.08 0.07 0.24 d 0.70 39.54 0.99 0.86 0.29 0.00 aa 45.75 24.98 0.07 0.24 0.21 0.26 ad 0.47 12.42 0.97 0.78 0.07 0.00 dd 29.35 58.02 0.61 1.52 0.37 0.00

PAGE 97

97 Table 3 1 8 Epistatic gene effect estimates for colorimetric values (L*, a*, b*, chroma and hue ) a measured in measured in Butterbush (BB) x Sucrine DuBerry ( Cucurbita moschata ) populations with one fruit per genotype b Gene effect c L* SE P value a* SE P value b* SE P value m 67.41 2.42 0.00 19.12 3.61 0.00 63.68 4.74 0.00 a 1.93 2.62 0.46 2.40 5.49 0.66 6.11 5.28 0.25 d 7. 54 11.25 0.50 17.29 19.03 0.37 2.24 22.18 0.92 aa 3.55 7.71 0.65 12.52 22.80 0.58 4.84 18.59 0.80 ad 1.06 3.26 0.75 4.89 6.17 0.43 0.10 5.83 0.99 dd 73.99 15.04 0.00 33.18 28.67 0.25 52.17 29.81 0.08

PAGE 98

98 Table 3 1 8 Continu ed a Numeric description of color using L*a*b* CIELAB color space. L* (lightness) ranges from black (0) to white (100); a* intensity in red (a* > 0) or green (a* < 0); b* intensity in yellow (b* > 0) or blue (b* < 0). Chroma (saturation or vividness) a s chromaticity increases a color becomes more intense; as it decreases a color becomes more dull. Hue (tint of color) an angular measurement where 0 = red, 45 = orange red, and 90 = yellow. b Fruit with overall highest a*, b* and chroma values were selected due to strong correlations with total carotenoids, lutein and lutein, respectively. Fruit with lowest L* and hue due correlation direction of all measured carotenoids. c Gene effects were calculated using calculations described by Hayman (1958) and Gamble (1962). Gene effect c Chroma SE P value Hue SE P value m 66.37 4 .74 0.00 1.2704 3.0929 0.68 a 5.32 5.61 0.35 0.0603 0.0804 0.46 d 2.52 22.59 0.91 0.2708 0.2827 0.34 aa 0.23 21.20 0.99 0.1913 0.3328 0.57 ad 1.14 5.98 0.85 0.0671 0.0930 0.47 dd 44.87 31.01 0.15 1.7409 0.4242 0.00

PAGE 99

99 CHAPTER 4 CONSTRUCTION OF A RAPD AND SSR BASED GENETIC LINKAGE MAP AND QTL MAPPING OF L*A*B* COLOR SPACE VALUES IN WINTER SQUASH ( Cucurbita moschata ) Introduction Carote carotene, serve as vitamin A (retinol) precursors that function in many body processes including vision, bone growth, cell division and differentiation, and reproduction (ODS/NIH, 2006). Carotenoid content in squash fruit, a main component of its nutritional value (Gross, 1991) can be indirectly increased through phenotypic selection for fruit with increased flesh pigmentation. The range of color in squash flesh is indicative of a quantitative trait, controlled by many gene s, rather than a qualitative tra it controlled by few genes ( Chapter 3, Table 3 5). By marking the genomic regions involved in flesh color of squash fruit, plant breeders may use the information as a tool to more effectively select for an increase in color and thereby nutritional content. Currently, linkage map development within the Cucurbita genus is limited; there exists only five genetic maps for the Curcubita genus to date The first two maps constructed were of interspecific crosses between C. moscha ta and C. pepo and composed of random amplified polymorphic DNA ( RAPD ) markers. The first, used an F 2 population (Lee et al., 1995 ), while a BC 1 population was used for t he second map that included morphological markers ( Brown and Myers 2002 ). The third map in Cucurbita was created in 2004 by Zraidi and Lelley in C. pepo using an F 2 population. It was created primarily with RAPD markers but included simple sequence repeat ( SSR ) markers morphological, and phenotypic characters. This map was expanded upo n and compared to a second C .pepo map reported in 2007 Zraidi et al. (2007) used two F 2

PAGE 100

100 populations, created from two intraspecific crosses to create the first consensus map for C. pepo using RAPD, amplified fragment length polymorphism ( AFLP ) SSR, and sequence characterized amplified region (SCAR) markers. More recently due to the utility of SSR markers in mapping and diversity studies, and the limited number of SSRs available for Cucurbita over 500 SSR primers polymorphic in C. pepo C. moschata and C. ecuadorensis were developed (Gong et al ., 2008a) This group used these SSRs to update their first C. pepo map (Zraidi et al .,2007) and constructed a fourth published map in Cucurbita and the first SSR based map in C. pepo in 2008. The fifth and m ost recent Cucurbita map was the first SSR based map of C. moschata using an F 2 population, and examined its synteny with the most recent C. pepo map (Gong et al, 2008b) The current limited availability of genetic maps within Cucurbita is due in large part t o the limited funding available for research since crops within this genus are minor and do not rank among the most economically important vege table crops. Currently within Cucurbitaceae genetic mapping is most developed in cucumber ( Cucumis sativu s ) and melon ( Cucumis melo ) (Paris, 2008). Of the two, Cucumis sativus has had more marker development including single nucleotide polymorphism (SNP) and expressed sequence tag (EST) marker systems (Staub, et al., 2008). Yet Cucumis melo using SSR marker s, is the most advanced genetic map and species in the Cucurbitaceae Cucumis sativus and Cucumis melo have had numerous quantitative trait loci (QTLs) identified. Major areas of QTL identi fication in Cucumis sativus include fruit quality traits, disease resistance, and components of yield (Staub et al., 2008), while Cucmis melo QTL identification

PAGE 101

101 similarly includes fruit quality and disease resistance, in addition to flower biology QTLs (Pi trat, 2008). To date, no QTLs have been identified within the Cucurbita genus. The objectives of this experiment were to create a genetic linkage map in C. moschata that could be compared with existing C. moschata and C. pepo linkage maps (Gong et al, 2 008b) and would have sufficient marker density to perform quantitative trait loci (QTL) studies to map regions of the genome associated with mesocarp coloration. Materials and Methods Plant Material The U S (BB) and French heirloom (SDub) were selected from germplasm screen ing for color and carotenoid correlation conducted 6 April 15 June 2007 at the University of Florida Plant Science Research and Education Center (PSREC) in Citra, Fl. Selections were b ased on differences in subjective flesh color observations (see Chapter 2) BB was crossed with SDub in a greenhouse to produce F 1 seedlings, one of which was selected to self to create an F 2 population segregating for flesh color. Ninety F 2 individuals w ere grown in a completely randomized design with eight individuals each of BB SDub and F 1 plants at the PSREC during 9 April 3 July 2008. The plot was grown under fertility rates and cultural practices as outlined by Olson and Simonne (2007) for both f ield experiments conducted in 2007 and 2008. DNA Extraction Newly emerging disease and insect free true leaves from three week old plants were collected for DNA extrac tion from the field planting at the PSREC. Leaves were frozen and held at 80 C prior to lyophilization using a Labconco freeze dryer (Kansas

PAGE 102

102 City, M O ) for 48 hours prior to extraction, for which a modified CTAB DNA extraction protocol was used (Kabelka and Young, 2010). Marker Selection, Polymorphism Screening and Genotyping Molecular m arker advancement within C.moschata has been limited, and randomly amplified polymorphic DNA (RAPD) markers have been the primary marker system used for detecting genetic differences BB and SDub were screened against 1,200 RAPD primer combinations ( Eurof ins MWG Operon Technologies, U.S. ) to identify those that exhibited polymorphism between the parents. Simple sequence repeat (SSR) markers are currently the most use ful marker system available for identifying polymorphism below the species level in Cucurbi taceae (Lebeda et al., 2007) However, SSR markers identified from genomic sequences of C. moschata, and C. pepo have only recently become available (Gong et al., 2008 a ). As these SSR markers became available BB and SDub were screened for polymorphism wi th 262 derived from C. moschata and 193 from C. pepo (455 total) For RAPD PCR the final reaction volume of the parental screens was 26 l, and contained 5 l of template DNA, 13.44 l of DNA grade water (Fisher, Pittsburgh, PA), 2.5 l of 10x PCR buffer 2.0 l of 25 mM MgCl 2 (Promega, Madison, WI), 2.0 l of 2.5 M dNTPs (from where?) 0.06 l of Taq polymerase, and 1 l 20 pm of forward and reverse primer s PCR amplification parameters used an initial denaturation step at 95 C for 5 min, followed by 40 cycles of : denaturation at 94 C for 1 min, 43 C annealing for 1 min, and extension at 72 C for 2 min, with a final extension step at 72 C for 5 min All amplifications were run on an Eppendorf Mastercycler (Hauppauge, NY). For amplification of SSR marke rs, the PCR protocol for parental screens was changed to 33 cycles, and primer specific annealing temperatures (Gong et al., 2008a) were used.

PAGE 103

103 Additionally, for both the RAPD and SSR PCR protocols, the final PCR reaction volume was 25 l for the F 2 progen y screens, with the amount of DNA grade water per reaction reduced from 13.44 to 12.44 l RAPD PCR products were separated on a 1.5% agarose gel at 100 volts for 5 hours; SSR PCR products used a six percent poly acry lamide gel at 260 volts for two hours Both gel types were run in 0.5 x (tris base, boric acid, EDTA) TBE buffer For RAPD PCR products, 15uL of ethidum bromide (EtBr) was added to the agarose gel, while the acrylamide gels for SSR PCR products were stained with 50 l ethidium bromide (EtBr) diluted in 200mL 0.5x TBE buffer for 25 minutes Gels of both marker types were viewed under UV light and images of each gel were recorded in digital format. Additionally, a subset of four SSR markers initially determined to be monomorphic by polyacrylamide gel electrophoresis, M120, M042, M009, and P098, were genotyped using high resolution melting (HRM) (see Chapter 5). Linkage Map C onstruction and QTL Mapping Analyses P olymorphic markers identified for both marker types in BB and SDub were used to genoty pe the entire F 2 population of 90 individuals, (Figure 4 1). All marker data were analyzed in JoinMap 3.0 (Ooijen and Voorrips, 2001) function with a LOD score threshold of 3.0 and RE C threshold of 0.45. Skewed markers determined b ased on the Chi square test in JoinMap, evaluated at a significance level of P ulations and are indicated on the map (Figure 4 2.) Phenotypes for the color space values L* a* b* chroma and hue to be used for QTL mapping were obtained as outlined in Chapter 3. Due to highly significant variation

PAGE 104

104 betw een fruit from the same genotype, the highest value measurement (genetic potential) for each of the color space values were used for QTL analysis. MapQTL 5.0 (Ooijen, 2004) was used to perform interval mapping (IM) and composite interval mapping (MQM), w ith a mapping step size of 1 cM, for each of the five color space values. Only MQM results are reported. Cofactors for MQM analyses were selected manually based upon both genome wide and linkage group specific LOD score thresholds as determined by a 1,000 iteration permutation test, at P were used to reduce residual variance of multiple segregating QTLs, thereby increasing the genetic effect of the associated map position and increasing the significance of the test. QTLs were determined significant when LOD values exceeded either the genome wide or linkage group specific LOD score threshold. Results and Discussion Polymorphic Marker Detection BB and SDub were screened against 1,200 RAPD primer combinations, 700 of which were found to ampl ify at least one polymorphic locus (Table 4 1). Both parents were subsequently re screened for the 700 primer combinations for repeatability. Five hundred thirty eight of the 700 primers (77%) were confirmed polymorphic and used for genotyping 90 F 2 indiv iduals. SSR marker p arental screens revealed a high level of monomorphism with 177 identified from markers derived from C. moschata (68%) and 121 from markers derived from C. pepo (63% ) being unusable in the mapping population Of the 262 SSR markers de rived from C. moschata genomic DNA, only 49 (19 %) were polymorphic, with 36 (14 %) that did not amplify, and 177 ( 68 %) that were monomorphic. Of the 193 SSR markers derived from C pepo g enomic DNA, only 12 (6 %) were polymorphic, with 60 (31 %) that did not amplify, and 121 (6 3 %) that were

PAGE 105

105 monomorphic. In total, only 61 SSR markers had polymorphisms between the two parents and were able to be used for genotyping. The low level of polymorphism illustrates the high level of inbreeding present among C. moschat a germplasm. Additionally, 46 of the 298 markers that were found to b e monomorphic in the parents were also found on the SSR based synteny map of C. moschata and C. pepo (Gong et al., 2008b). Linkage Map Composition A linkage map was constructed in C mos chata (2n=2x=40) from an F 2 population of 90 individuals derived from an intraspecific cross between BB and SDub (Figure 4 2). Three hundred nineteen RAPD and 39 SSR loci were used to construct the linkage map. The total SSR loci include four monomorphi c C. moschata SSR loci that were recovered using HRM, and are included in the total number of SSR markers. Of these, 269 loci (235 RAPD and 34 SSR), were placed on the linkage map after analysis (Table s 4 1 and 4 2 ). A total of 21 linkage groups were obt ained (Table 4 2) with a total map length of 1086 9 cM, and averages of 51.8 cM per linkage group and 4.0 cM between markers (Table 4 3). The number of mapped linkage groups should equal the base chromosome number of 20 for C. moschata It is likely that insufficient marker density in our population reduced our ability to identify the correct number of linkage groups. Only 75% of the marker loci genotyped could be placed on linkage groups (Table 4 1). Cucurbita Map Comparisons The first SSR based lin kage map for C. moschata was published by Gong et al. (2008a), and is currently the most current published map for the species. Within the same publication, Gong et al. (2008a) also examined the synteny of the SSR based C.

PAGE 106

106 moschata map with the most curre nt and singular SSR based C. pepo linkage map (Gong et al., 2008b). Comparing the linkage map produced in this study with both the previously reported C. moschata and C. moschata C. pepo synteny map revealed both similarities and differences in linkage g roup homology between the three maps (Table 4 4). Linkage groups in this study that have homology with only one linkage group of the previous C. moschata map are LG3, LG5, LG6, LG11, LG13, LG14, and LG 17. Of these, LG 11, LG 13 and LG 14 have only one shared SSR between the two groups. Linkage groups LG3, LG5, LG6, and LG17 have two shared SSRs, which increases our confidence in the homology between the linkage groups examined. However, when more than two SSR markers were shared between the linkage ma p produced in this study and the current SSR based C. moschata linkage map, homology was often found to multiple linkage groups. This is displayed most notably in LG1, the largest linkage group. The same splitting of linkage group homology also occurred f or shared SSRs greater than two for the C ucurbita synteny map in LG1. Identification of SSR loci on different linkage groups compared to previously published maps is likely due to additional SSR loci located in the C. moschata genome. Alternatively, chrom osomal translocations may have occurred. Additional comparative mapping studies with different C. moschata genotypes will be necessary to elucidate these differences. Due to multiple linkage groups having more than one homologous group in the two previous ly reported maps, the numbering of linkage groups for the present map were arbitrarily assigned. It was not possible to demonstrate f urther map linkage group homology with other published Cucurbita maps. Just three maps were constructed prior to the work s by

PAGE 107

107 Gong et al. (2008a) and these maps only contain RAPD and AFLP markers (Lee et al., 1995; Brown and Myers, 2002; and Zraidi et al., 2007). No AFLPs are on the map created in this study, and RAPDs are not suitable markers to serve as anchor points for m ap alignment. QTL Analyses QTL detection of L* a* b* chroma and hue color space values All QTL analyses were performed using phenotypes from only 1 fruit per genotype due to highly significant amounts of variation for fruit within a genotype for all genera tions in all color space values ( P < 0.0001), likely due to variable fruit maturity within a genotype. Fruit within a genotype were selected based upon highest (a*, b*, and chroma ) or lowest (L* and hue ) average replicate measurements within a fruit. Sel ection of fruit with highest or lowest average fruit replicate measurements for each color space value was based upon the direction of each respective color space value correla ted with carotenoid levels ( Chapter 2, Table 2 3) and represent the genetic pote ntial of the genotype examined. Analyses of QTLs were performed separately for each color space value. All associations reported were identified above the individual linkage group LOD threshold for each color space value, unless otherwise noted. QTLs we re identified on linkage groups 7 and 18 for color space value L* (Figure 4 3, Table 4 5). QTLs associated with color space a* value were identified on linkage groups 3, 10, and 18 (Figure 4 4, Table 4 5). Linkage groups 5 and 17 had QTLs associated with both color space values b* and chroma The LOD score distribution of markers was nearly identical when comparing b* to chroma on the same linkage group (Figures 4 5 and 4 6, Table 4 5). Color space value hue angle had the most QTL identified, on linkage groups 3, 13, 14,

PAGE 108

108 15, and 19. Additionally, genome wide LOD score significance was detected on linkage groups 3, 13, and 15 (Figure 4 7, Table 4 5). Implications for population selection on QTL detection Of the three populations used to evaluate herita bility in chapter three, the Butterbush (BB) population was selected due to the presence of transgressive F 2 segregants for color space value, hue the overall angular measurement of color. Unfortunately, two drawbacks to using this rationale mapping an d QTL population selection were subsequently identified. First, populations with transgressive segregation were identified prior to correlation between color space values and carotenoid content. This was necessary to allow sufficient time for genotyping and linkage mapping. Therefore, it remained unknown which color space value(s) would give the best estimate of the relationship between color and carotenoid content of squash flesh. After the correlations were calculated, it was found that color space val ues a*, b* and chroma would have been more appropriate to examine transgressive segregation ( Chapter 2, Table 2 3). Although evidence of transgressive segregation for a quantitative trait is valuable from a breeding standpoint, the identification of geno mic regions contributing to the quantitative trait is often better accomplished by selecting parents with a large amount of variation for the trait. Therefore, use of transgressive segregation as selection criteria for which population to map and perform QTL analyses is not as important as selecting parents with the largest differences for the trait in question; the wider phenotypic range for the trait of interest increases power of QTL in progeny. Unfortunately, the only intraspecific C. moschata cross we performed was between parents with little variation for flesh color. Based on variation for flesh color, the C. pepo Table Gold Acorn (TGA)

PAGE 109

109 would have most likely increased o ur probability of QTL detection ( Chapter 3, Table 3 1). However, these QTL assi gnments should be viewed with caution, as heritability estimates for hue angle in this population indicated very little genetic variance present for this trait (Table 3 14). Therefore, validation needs to be performed using different genetic backgrounds t o confirm these QTL locations. Additionally, nutritional carotenoid levels do not need to be increased within butternut squash ( C. moschata ), but could be increased in acorn squash ( C. pepo ). No official recommended dietary intake levels for carotenoids, other than vitamin A precursors, have been reported (Murkovic et al., 2002) Therefore, the only carotenoid within squash with a recommended dietary intake is pro vitamin A, which is converted int o retinol activity units, RAE (Trumbo et al., 2001) In 1. 00 cup (205g) of baked butternut squash ( C. moschata ) served without salt, there are 1,144 RAE. For adult females, over the age of 19, this serving accounts for 163% of the minimum daily value. For adult males, this is 127%. In 1.00 cup (205g) of baked acorn squash ( C. pepo ) served without salt, there are 43 RAE. For adult females, over the age of 19, this serving accounts for 6% of the percent daily value. For adult males, this is 5% (USDA ARS, 2010). Although butternut squash vitamin A levels do not ne ed to be increased, this work may be used to identify genomic flesh color associations in C. pepo Evidence of linkage group homology between C. moschata and C. pepo (Gong et al., 2008b) suggests that C. moschata QTLs may be used to identify syntenic geno mic regions in C. pepo controlling flesh color and be used to increase its color and nutrient content through plant breeding.

PAGE 110

110 Implications and Practical Applications of QTL Analysis in Breeding for Color and Carotenoid Content in Cucurbita Alignment of hom ologous groups between the C. moschata map produced in this study with SSR based maps produced in C. moschata and the synteny map of C. moschata and C.pepo (Table 4 4) show promise. However, due to the high level of monomorphic SSR markers identified in th e C. moschata linkage map presented in this chapter, complete alignment other C. moschata and C. pepo populations would not be possible. With increased recovery of monomorphic SSR markers using high resolution melting (HRM) real time PCR analysis (see Cha pter 5), QTL information obtained here could be compared with other Cucurbita populations through map alignment. This would enable similar regions associated with color to be identified in other populations. In addition, as more QTL studies and map align ments become available within the Cucurbita genus, multiple traits can be used in squash breeding efforts.

PAGE 111

111 Table 4 1. Loci summary in C. moschata linkage map Marker Type RAPD SSR Total primer combinations 1200 455 Total polymorphic prime r combinations 700 61 a Total rescreened 538 --Total polymorphic primer combinations in F 2 222 39 Total polymorphic loci b in F 2 319 39 Total loci on linkage map 235 34 a Count excludes polymorphic markers identified using real time PCR high resolu tion melting. b The number of polymorphic loci differs from the number of markers because some markers had multiple polymorphic loci on a gel.

PAGE 112

112 Table 4 2. Indiv idual linkage group summaries for marker composition, length, density of markers and average marker distance in C. moschata Marker type LG 1 LG 2 LG 3 LG 4 LG 5 LG 6 LG 7 LG 8 LG 9 LG 10 LG 11 LG 12 LG 13 RAPD 77 35 19 1 4 7 9 12 11 8 6 6 7 3 SSR 12 4 2 0 2 3 1 1 1 0 2 0 2 Total 89 39 21 14 9 12 13 12 9 6 8 7 5 Map statistics Length (cM) 162.2 105.1 63.9 37.5 76.8 75.6 49.1 108.7 48.8 47.2 53.6 42.8 59.7 Marker density (markers/cM) 0.55 0.37 0.33 0.37 0.12 0.16 0.26 0.11 0.18 0.13 0.15 0.16 0.08 Average marker distance (cM) 1.82 2.69 3.04 2.68 8.53 6.30 3.78 9.06 5.42 7.87 6.70 6.11 11.94 Largest gap (cM) 23.0 14.6 16.7 6.1 21.6 36.3 15.1 21.8 14.2 16.1 21.3 20.3 23.6 Gaps >15cM (n) 2 0 1 0 2 1 1 3 0 1 1 1 2

PAGE 113

113 Table 4 2. Continued Marker type LG 14 LG 15 LG 16 LG17 LG 18 LG 19 LG 20 LG 21 RAPD 4 5 3 0 3 2 2 2 SSR 1 0 1 2 0 0 0 0 Total 5 5 4 2 3 2 2 2 Map statistics Length (cM) 33.8 18.1 27.5 17.4 4.5 13.4 22.5 18.7 Marker density (markers/cM) 0.15 0.28 0.15 0.11 0.67 0.15 0.09 0.11 Average marker distance (cM) 6.76 3. 62 6.88 8.70 1.50 6.70 11.25 9.35 Largest gap (cM) 24.2 6.0 12.5 17.4 2.5 13.4 22.5 18.7 Gaps >15cM (n) 1 0 0 1 0 0 1 1

PAGE 114

114 Table 4 3. Linkage map summary in C. moschata. Entire Map Total map length (cM) 1086.9 Total markers 269.0 Average mark er density ( Marker/ cM) 4.0 Largest gap (cM) 36.3 Total no. linkage groups 21 Average cM/linkage group 51.8

PAGE 115

115 Table 4 4 Common simple sequ ence repeat (SSR) markers on linkage groups between the C. moschata F 2 population of and ZHOU x WB integrated F 2 populations, and the C. moschata and C. pepo synteny map a a Population and map information for integrated F 2 C. moschata and C. moschata and C. pepo synteny maps were taken from Gong et al., 2008 b C. moschata Linkage group homology in C. moschata Common SSRs Linkage group homology in Cucurbita synteny map Comm on SSRs LG1 LGm7, LGm2, LGm20, LGm1b, LGm25, LGm4, LGm22, LGm1a 13 LG7, LG11, LG20 5 LG2 LGm14, LGm19 5 LG 13 2 LG3 LGm3 2 LG 16 1 LG5 LGm1b 2 LG10 1 LG6 LGm15 2 ----LG7 LGm10,LGm13 2 ----LG11 LGm8 1 ----LG13 LGm5b 1 ----LG14 LGm5 1 ----LG17 LG m21a 2 LG 14 1

PAGE 116

116 Table 4 5 QTL detected for flesh co lor using L*a*b* color space values in a Cucurbita moschata F 2 population of QTL a Linkage group Interval length (cM) b LOD peak position (cM) Maximum LOD Nearest marker R 2 (%) c Genetic effect: a c d L* lightness1 7 1.90 29 .95 2.43 OPAB03A 12.3 2.52 l ightness2 7 3.00 36.96 2.06 OPAA09 16.9 1.03 l ightness 3 18 2.00 1.00 3.38 OPL11 17.9 1.33 a* red/green1 3 2.37 55.40 3.55 OPT20 14.3 1.80 red/green2 10 8.10 27.12 2.50 OPN11 10.7 1.62 red/green3 18 4.57 3. 03 2.58 OPAD10A 10.3 2.21 b* yellow/blue1 5 6.00 72.94 3.35 OPI17 28.9 3.77 yellow/blue2 17 5.00 0.00 1.77 M235 8.2 1.73 Chroma colorsaturation 1 5 6.00 72.94 3.27 OPI17 27.8 3.79 colorsaturation2 17 11.00 3.00 5.00 1.97 M235 9.8 1.94 Hue hueangle1 3 3.77 63.18 5.85 OPAF14A 28.5 1.80 hueangle2 13 16.58 4.00 4.49 OPY13A 15.5 1.63 hueangle3 14 16.00 4.00 2.43 OPAI13 21.9 1.32 hueangle4 15 9.66 14.55 3.44 OPM03 14.7 1.02 hueangle5 19 3.39 13.39 1.68 OPBG18B 21.2 2.39 a QTLs reported are based on LOD scores above linkage group specific LOD score threshold. b Interval length based upon linkage group region with LOD scores above linkage group specific LOD score threshold c Percent explained variance and genetic e ffect reported at interval with the highest LOD score. d Additive genetic effect. Denotes QTL associations with genome wide LOD score threshold.

PAGE 117

117 A B Figure 4 1. Example gel images imaged under UV ligh t with ethidiu m bromide stain. A) A RAPD marker (OPQ15) with multiple polymorphic loci, both dominant and codo minant, on a 1.5% agarose gel, B) A SSR marker (M005) with a codominant locus on a 6% polyacrylamide gel In both gel images, lane 1 is 1 and the following lanes are the F 2 progeny.

PAGE 118

118 LG1 LG2 LG3 LG4 Figure 4 2 Linkage map of Cucurbita moschata using both random amplified polymorphic DNA (RAPD) and s imp le sequence repeat (SSR) markers. desig nation indicate RAPD markers. Markers ending with a letter designation such d by an individual marker. Markers beginnin C. moschata e derived from the C. pepo genome. Asteriks (*) indicate s ignificantly skewed at the 0.05 probability level.

PAGE 119

119 LG5 LG6 LG7 LG8 LG9 LG10 LG11 LG12 LG13 LG14 LG15 LG16 LG17 LG18 LG19 LG 20 LG2 1 Figure 4 2 Continued

PAGE 120

120 0.0 0.5 1.0 1.5 2.0 2.5OPBB14B OPAU03B OPAU03C OPY04 OPA16 OPAB03B OPAK16 OPBA10A OPAB03AOPAP04BOPAT19D OPAA09 M204LOD Score Markers 0.0 0.5 1.0 1.5 2.0 2.5 LOD Score Group 7 Legend LOD 2.0 LOD 3.7 0.0 1.0 2.0 3.0 4.0OPL11 OPAD10A OPBA06CLOD Score Markers 0.0 1.0 2.0 3.0 4.0 LOD ScoreGroup 18 Legend LOD 1.5 LOD 3.7 Figure 4-3. Color space value L* quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and in dividual linkage group LOD scores ( P 0.05). OPAB03A OPAP04B OPAT19D

PAGE 121

121 0 1 2 3 4P125 OPP17 OPAG09B OPAU03D OPX05B OPW02A P206 OPU02A OPAR19 OPQ09A OPE18 OPH05A OPQ12 OPG09C OPAI10 OPX19B OPS01C OPU01 OPT20 OPAQ20 OPAF14ALOD Score Markers 0 1 2 3 4 LOD Score Group 3 Legend LOD 2.7 LOD 3.9 0.0 0.5 1.0 1.5 2.0 2.5OPAT19A OPR08 OPN11 OPAZ14 OPY08A OPH05BLOD Score Markers 0.0 0.5 1.0 1.5 2.0 2.5 LOD Score Group 10 Legend LOD 2.1 LOD 3.9 Figure 4-4. Color space value a* quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and in dividual linkage group LOD scores ( P 0.05). OPP17 OPAG09B OPAU03D OPX05B OPW02A P206 OPQ09A OPE18 OPH05A OPQ12 OPG09C OPAI10 OPU01 OPT20

PAGE 122

122 2.40 2.42 2.44 2.46 2.48 2.50 2.52 2.54 2.56 2.58OPL11 OPAD10A OPBA06CLOD Score Markers 2.40 2.42 2.44 2.46 2.48 2.50 2.52 2.54 2.56 2.58 LOD Score Group 18 Legend LOD 1.6 LOD 3.9 Figure 4-4. Continued

PAGE 123

123 0 1 2 3 4M232 OPAB18 OPAD19C OPG15B OPAC03B OPAZ03 OPX03B M229 OPI17LOD Score Markers 0 1 2 3 4 LOD ScoreGroup 5 Legend LOD 2.3 LOD 3.8 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7M235 M119LOD Score Markers 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 LOD ScoreGroup 17 Legend LOD 1.7 LOD 3.8 Figure 4-5. Color space value b* quantitative trait loci (QTL) linkage group associations evaluated using both genome wide and in dividual linkage group LOD scores ( P 0.05). OPX03B M229

PAGE 124

124 Legend LOD 2.3 LOD 3.8 0 1 2 3 4M232 OPAB18 OPAD19C OPG15B OPAC03B OPAZ03 OPX03B M229 OPI17LOD Score Markers 0 1 2 3 4 LOD ScoreGroup 5 1.3 1.4 1.5 1.6 1.7 1.8 1.9M235 M119LOD Score Markers 1.3 1.4 1.5 1.6 1.7 1.8 1.9 LOD ScoreGroup 17 Legend LOD 1.8 LOD 3.8 Figure 4-6. Color space value chroma quant itative trait loci (QTL) linkage group associations evaluated using both genom e wide and individual linkage group LOD scores ( P 0.05). OPX03B M229

PAGE 125

125 Legend LOD 2.5 LOD 3.4 0 1 2 3 4 5P171 OPY13A OPAH18 M145 OPAM11LOD Score Markers 0 1 2 3 4 5 LOD ScoreGroup 13 Legend LOD 2.0 LOD 3.4 0 1 2 3 4P125 OPP17 OPAG09B OPAU03D OPX05B OPW02A P206 OPU02A OPAR19 OPQ09A OPE18 OPH05A OPQ12 OPG09C OPAI10 OPX19B OPS01C OPU01 OPT20 OPAQ20 OPAF14ALOD Score Markers 0 1 2 3 4 LOD ScoreGroup 3 Figure 4-7. Color space value hue angle qu antitative trait loci (QTL) linkage group associations evaluated using both genome wide and individual linkage group LOD scores ( P 0.05). OPP17 OP AG09B OPAU03D OPX05B OPW02A P206 OPQ09A OPE18 OPH05A OPQ12 OPG09C OPAI10 OPU01 OPT20

PAGE 126

126 0.0 0.5 1.0 1.5 2.0 2.5OPAI13 MC60 OPQ09B OPZ08 OPAY10LOD Score Markers 0.0 0.5 1.0 1.5 2.0 2.5 LOD ScoreGroup 14 Legend LOD 1.6 LOD 3.4 0 1 2 3 4OPAF07 OPBC04A OPA09B OPM03 OPA09ALOD Score Markers 0 1 2 3 4 LOD Score Group 15 Legend LOD 1.9 LOD 3.4 Figure 4-7. Continued

PAGE 127

127 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6OPY06B OPBG18BLOD Score Markers 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 LOD Score Group 21 Legend LOD 1.5 LOD 3.4 Figure 4-7. Continued Group 19

PAGE 128

128 CHAPTER 5 DEVELOPMENT AND UTILIZATION OF A PROTOCOL FOR SMALL DNA SEQUENCE DIFFERENCE DETECTION IN MONOMORPHIC SIM PLE SEQUENCE REPEAT ( SSR ) MARKERS IN WINTER SQUASH ( Cucurbita moschata ) USING HIGH RESOLUTION MELTING CURVE ANALYSIS Introduction DNA markers have proven to be valuable tools in plant breeding and genetics. The utility of molecular markers depends on the presence of polymorphisms in target material, and the method of polymorphism detection. Detection methods can vary in complexity, and the nature of polymorphism under scrutiny. For example, restriction fragment length polymorphism (RFLP) markers are a hy bridization based method that requires the availability of a previously sequenced clone, and detect s polymorphisms due to variable distances between adjacent restriction sites. Variable distances can be due to base substitutions in a restriction site, ins ertion and/or deletion of DNA sequence (indels), or other chromosomal rearrangements (Xu, 2010). In contrast, polymerase chain reaction (PCR) based methods are more expedient than hybridization based methods, and do not require previously sequenced clone s or large amounts of DNA. PCR based markers utilizing random oligonucleotide primers such as amplified fragment length polymorphism (AFLP) and random amplified polymorphic DNA (RAPD) are often used in minor plant species with a lack of prior genome level information (Schlotterer, 2004). These markers do not require any sequence information prior to analysis, and for this reason they lack locus specificity. However, PCR based markers derived from sequence tagged sites have the advantage of targeting spec ific genome regions. Polymorphism resulting from these markers takes advantage of indels, either in the polymerase chain reaction amplified region between two molecular marker primer binding sites, or in the primer binding sites themselves

PAGE 129

129 resulting in a loss of amplification products (Xu, 2010). For example, allele variation at simple sequence repeat (SSR) loci results from variable numbers of repeated DNA units, typically in di or tri nucleotide motifs (Ellegren, 2004). As the wealth of DNA sequence information for plant species continues to grow, through expressed sequence tag library development, transcriptional profiling, and whole genome sequencing, the availability of markers based on single nucleotide polymorphisms (SNP) is increasing. SNP site s are the most abundant molecular marker s in plant genomes, widely and evenly distributed throughout the genome, and occurring as frequently as every 60 120 base pairs in some species (Agarwal et al., 2008). SNP alleles result from a single base pair chan ge rather than large indels or repeat sequence variation, thus polymorphisms are not typically differentiated through traditional visualization methods such as agarose gel electrophoresis. Specialized detection methods such as DNA sequencing, primer exten sion, and enzymatic cleavage are often employed in high throughput assays (Chagne et al., 2007). An alternative to these approaches is the analysis of the melting behavior of double stranded PCR amplified DNA fragments at high temperatures, which depends very much on the sequence and the sequence composition. The thermal stability of PCR product results from its nucleotide sequence (Montgomery et al., 2007). Small differences in sequence composition, such as SNP allele differences leading to variation in GC nucleotide content, result in differing stability of double stranded DNA (Reed et al., 2007). High resolution melting analysis utilizes dyes that fluoresce in the presence of double stranded DNA coupled with instruments that monitor the fluorescence during heating of the PCR amplification product

PAGE 130

130 (Montgomery et al., 2007). As temperature increases and melting of double stranded DNA occurs, and fluorescence signal decreases. This leads to a characteristic melting profile that has been used to genoty pe diverse plant species such as grape (Mackay et al., 2008), apple (Chagne et al., 2008), barley (Lehmensiek et al., 2008), lupin (Croxford et al., 2008), almond (Wu et al., 2008), and potato (De Koeyer et al., 2010). Breeding populations resulting fr om crosses between parents with a high level of inbreeding often result in a low level of allele polymorphism at a given locus. Although a marker locus may be present and amplifies during PCR, parents with the same alleles at the locus will not produce se gregating progeny, resulting in monomorphic loci. The high frequency of SNP markers in plant genomes will help in situations such as this by increasing the pool of available markers. However, in minor agricultural species such as Cucurbita moschata the l ack of available genome sequence limits the use of SNP of a SNP in the amplicon that results in the addition or deletion of a restriction enzyme cleavage site. However, th ese cleaved amplified polymorphic sequence (CAPS) markers require an additional restriction digestion step prior to genotyping, and require sequencing of the original amplification product to determine the appropriate enzyme to use (Agarwal et al., 2008). Similar to CAPS markers, high resolution melting curve analysis may be used to recover monomorphic marker loci by allowing genotyping of sequence variation that does not result in the loss of primer binding sites or presence of indels large enough for tra ditional electrophoretic analysis The objective of these experiments was to develop a protocol to genotype the BB x SDub F 2 population for 10

PAGE 131

131 SSR markers previously discarded due to lack of segregation between the parents and progeny using high resolutio n melting curve analysis. Materials and Methods Plant M aterial and DNA Extraction For these experiments, an F 2 population of 90 individuals was developed from a cross between the U.S. DuBer (see C hapters 3 and 4). The F 2 population, and eight individuals each of BB, SDub and F 1 plants were grown during 9 April 3 July 2008 in a completely randomized design at the University of Florida Plant Science Research and Education Center (PSREC) in Citra, Fl. Newly emerging disease and insect free true leaves from three week old plants were collected for DNA extraction using the same protocol described in Chapter 4. Marker Selection and Polymorphism Screening Molecul ar marker development within C.moschata has been limited, and randomly amplified polymorphic DNA (RAPD) markers have been the primary marker system used for detecting genetic differences. Simple sequence repeat (SSR) markers are currently the most widely used marker system in Cucurbitaceae, but only recently have become available for the Cucurbita genus in 2008 (Gong et al., 2008). Initially, the parents BB and SDub were screened for polymorphism with 455 SSR markers, 262 derived from C. moschata and 193 from C. pepo genomes, respectively. The PCR reaction total volume was 26 l, and contained 5 l of template DNA 13.44 l DNA grade water (Fisher, Pittsburgh, PA), 2.5 l 10 x PCR buffer, 2.0 l 25 mM MgCl 2 (Promega, Madison, WI), 2.0 l of 2.5 M dNTPs, 0.06 l Taq polymerase, and 1 l of 20 pm F/R primer. PCR amplification parameters used an initial denaturation step of

PAGE 132

132 95 C for 5 min, followed by 33 cycles of denaturation at 94 C for 1 min, primer specific annealing temperature for 1 min, and DNA exte nsion at 72 C for 2 min, with a final extension step of 72 C for 5 min. All PCR reactions were run on an Eppendorf Mastercycler (Hauppauge, NY). PCR products were subjected to electrophoresis on a six percent polyacryalmide gel at 260 volts for two hou rs in 0.5 x TBE buffer and stained with 50 l ehidium bromide diluted in 200mL 0.5x TBE buffer for 25 minutes and imaged under an ultraviolet light source. DNA Sequencing, Alignment and Comparison A set of ten SSR markers all determined to be monomorphic i n BB and SDub by acrylamide gel electrophoresis, were selected for sequencing of the amplification products of both parents. DNA samples consisting of 45 l of PCR amplification product were purified using the QIAquick PCR Purification Kit from Qiagen (Va lencia, CA) and Biotechnology Research (ICBR). Sequencing was performed using the ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction kit, v. 3.1 (Perkin Elmer/ Applied Biosyste ms, Foster City, CA), analyzed on an ABI 3130 DNA sequencer (Applied Biosystems, Inc., Fullerton, CA) and edited with Sequencher 3.1.1. (Gene Codes, 1998) Sequence reads of both parents were aligned using the complete alignment function in ClustalX (Conway Institute, Dublin, Ireland) and quality values determining the probability of the correct base call were calculated using FinchTV software (Seattle, WA). High Resolution Melt Curve Analysis Protocol Development All melt curve analyses we re performed using a LightCycler 480 (Roche, Indianapolis, IN). The instrument utilizes real time PCR technology to quantify

PAGE 133

133 fluorescence signal changes resulting from double stranded DNA denaturation during a post amplification temperature gradient appli ed to the total PCR reaction volume. The PCR final reaction volume was 10 l, contained 2 l of template DNA, 1.4 l PCR grade water (Roche, Indianapolis, IN), 5 l 2 x Master Mix (Roche, Indianapolis, IN), 0.8 l 25 mM MgCl 2 (Roche, Indianapolis, IN), 2.0 l 2.5 M dNTPs, and 0.8 l 20 pm F/R protocol, which was used as the starting point for determining small genotypic differences with the initial ten sequenced mark ers. Protocol parameters were altered to align with the current PCR protocol in squash, and included a decrease in amplification cycle number from 45 to 35, altering denaturing, annealing and extension times to 1, 1, and 2 min respectively. Further modifi cations to the standard protocol made as a result of experimentation, are outlined in the results and discussion section. Markers with the largest deletions were screened first, and markers with only SNP differences screened last. In addition, as recommen ded by Roche, a large range of 65 95 C was initially used for the high resolution melting step. Markers were screened with BB, SDub, F 1 and the first 5 F 2 individuals to determine an optimized melting curve temperature (Tm) for detecting genotypic dif ference based on actual PCR product melting. Melting curve program temperatures were optimized for each marker with a 20 C range centered on the expected Tm value for each marker. However, plates that were prepared for troubleshooting with multiple marke rs maintained a large range of temperatures for melting conditions. All optimization experiments were run in an effort to design protocol parameters to detect polymrophisms between parental genotypes and the expected 1:2:1 co dominant allele segregation ratio in the F 2 progeny. All

PAGE 134

134 primers were analyzed using the gene scanning genotyping analysis option in the LightCycler 480 software, based on melting curve at a baseline sensitivity of 0.30. Higher sensitivity values were examined on an individual mark er basis in an attempt to match the expected segregation ratio for an F 2 population. Results and Discussion Parental screens revealed a high level of monomorphic SSR markers, with 190 identified from markers derived from C. moschata (73 percent) and 125 fr om markers derived from C. pepo (65 percent). The degree of genetic similarity between BB and SDub for available SSR markers necessitated the use of RAPD markers to develop a linkage map suitable for quantitative trait locus (QTL) analysis. However, RAPD markers are not useful for comparing synteny between previous C. moschata and C. pepo linkage maps. In an effort to align our linkage map with previously developed maps, we identified 46 markers found to be monomorphic in BB and SDub, but present on the SSR based synteny map of C. moschata and C. pepo (Gong et al., 2008). For these experiments, a total of 10 of these 46 markers were used to develop a protocol to DNA Sequence Differences Sequence an alysis of alleles of monomorphic SSR markers revealed small sequence differences in nine of the ten initial markers sent for sequencing. Both deletions and SNPs were identified (Table 5 1). The largest sequence difference was observed in marker M042 with a six base pair deletion in BB. Markers M120, P098, and M066 also had deletions over one base pair in size with four, three, and two base pair deletions identified, respectively. Four potential SNPs were identified in maker M120, but in each case the bas e calling software was unable to accurately assign a nucleotide

PAGE 135

135 difference at the SNP site. However, seven out of eight quality values determining accurate readings of base calls were over the Q=20 threshold, perhaps indicating that these SNPs are wobble locations. Only one marker, M261, had no sequence differences between BB and SDub. All other markers had at least one SNP present. Genotyping DNA Sequence Differences Using the LightCycler 480 Of the nine markers exhibiting sequence variation between B B and SDub, three (33%) were able to be genotyped utilizing high resolution melt curve analysis. Marker M042, with the largest sequence difference between the two parents (a six base pair deletion), showed three distinct melting curve groups (Figure 5 1), with genotypic classes segregating in the expected 1:2:1 ratio for a codominant marker in an F 2 population (Table 5 2 ). Given the relatively large deletion present in BB, this marker may have been scorable utilizing higher resolution genotyping and/or st aining methods (i.e., higher percentage gels, silver staining, or capillary electrophoresis). Four distinct melting curve groupings were identified for marker M009 (Figure 5 2). Progeny scored as a fourth genotypic class were evaluated using melting curve temperature shifted difference plot graphs (data not shown) and determined to have a similar curve to those progeny homozygous for the allele inherited from SDub (Table 5 2). Interestingly, M009 contained a single SNP between the two parents, a genotypic difference that would not be scorable using higher resolution genotyping methods due to a lack of size polymorphism. Progeny did not segregate as expected for marker P098 (3 base pair deletion in SDub), Evaluation of P098 at 0.40 sensitivity revealed clo ser adherence to a 1:2:1 segregation pattern than did progeny evaluation at 0.30 ( = 16.667, 26.689, respect ively) (Table 5 2 ). Of the 269 loci on the C. moschata linkage map reported in

PAGE 136

136 Chapter 4, 24 (9%) showed significant deviation from the expected segregation ratio. P098 is likely an example of distorted segregation due to facto rs such as gametophytic selection. The six remaining markers did not display genotypic differences that were initially scorable using high resolution melting curve analysis. Evaluating DNA Concentration Effects on Genotyping With High Resolution Melt Cu rve Analysis High resolution melting curve analysis is particularly sensitive to variation in template DNA concentration (N. Bassil, personal communication). With a thermocycler capable of real time PCR quantification, we were able to develop a method to experimentally define the working concentration of the template DNA samples using crossing point (Cp) values. The Cp value indicates the cycle during PCR where the sample fluorescence enters an exponential growth phase. Thus, difference in the Cp value o f one sample compared to another can be used as an indication of the starting DNA concentration of the reaction mixture. Since the amplification product grows exponentially during PCR, relatively small differences in Cp values may indicate large difference s in template DNA concentration. For example, using the SSR marker M066, the Cp values for BB and SDub varied by over four cycles (19.13 and 23.71, respectively), indicating the starting DNA concentration for BB was 16x more concentrated than SDub. Cp values of BB, an F 1 individual, and three F 2 progeny were selected for comparison between DNA concentration as measured by absorbance at 260 nm using two instruments; an Eppendorf BioPhotometer 6131 v.35 ( Hauppauge, NY ) and a Genesys 10 spectrophotometer v .2 (Thermo Scientific, Waltham, MA) (Table 5 3). Using the SSR marker M066, average Cp values from replicate experiments indicated

PAGE 137

137 that BB and the F 1 individual should have significantly higher DNA concentrations than the F 2 progeny. However, the relatio nship between Cp value and DNA concentration measured by absorbance was not linear. This may be an indication of contaminants in the DNA, such as RNA, proteins, polysaccharides, and/or polyphenols that may be inhibiting a true measure of DNA concentration by absorbance. To further examine the effect of DNA concentration based on dilutions on relative Cp values, four markers ( M120, M256, M261, and P224 ) were selected to create serial dilutions for concentration differences. None of these markers showed a d ifference between BB and SDub at sensitivity of 0.30 for melting curve analysis, although each contained a deletion and/or SNP between the two parents. Serial dilutions, ranging from 1:10 to 1:320, were prepared for each parent for each marker in two repl icates. PCR amplification was allowed to continue until all samples reached saturation, as indicated by cessation of exponential growth of amplification product, For each marker both parents showed a distinct progression from most concentrated to most dil ute in terms of number of cycles required to reach Cp. There was approximately a once cycle difference to Cp in neighboring serial dilutions for each parent replicate for each marker. Analysis of variance indicated the only significant difference was amon g dilutions for each parent ( P <0.001). No significant differences were identified between markers ( P = 0.9941), replicates within a maker ( P = 0.9523), or between parents ( P = 0.0767), suggesting that the only factor affecting Cp cycle number difference w as dilution, representative of DNA concentration. The one cycle difference in Cp value between serial dilutions is due to the sample having half the amount of DNA as compared to its

PAGE 138

138 earlier counterpart. This is the basis for using Cp value as a measure ment of relative DNA concentration. To examine the effect of DNA concentration on genotyping by high resolution melting, the SSR marker M120 was used. Serial dilutions of each parent were prepared in replicate, and amplification cycle number was selected based upon the point where all dilutions had the largest separations between one another. This cycle was expected to display the highest detection of polymorphic differences in comparison between the melt curves produced. Cycle 29 was selected for ending amplification in BB, followed with genotyping by high resolution melting. Two distinct groups were seen in replicates of BB (Figure 5 3). Dilutions of 1:10, 1:20, 1:40, and 1:80 all genotyped the same, while further dilutions were genotyped together or f ell into unknown genotypic classes. A similar trend for Cp values was seen in SDub (data not shown). These data suggest a point where samples with the same genotype are genotyped differently based upon differences in DNA concentration; it appears to be a range of approximately 4 cycles of difference. This indicates that Cp values based on DNA concentration for all samples need to be relative to one another to minimize DNA concentration effect and maximize the potential to identify the correct genotypic cl ass. Using the SSR marker M120, which we were initially unable to genotype using melting curve analysis, we attempted to genotype the population again after adjusting the DNA concentration of each F 2 individual based on the Cp values from the initial PCR. Amplification was terminated at 29 cycles while all samples still displayed exponential growth. At a sensitivity of 0.30, differences were observed in Cp values of the samples, reflective of genotypic differences rather than concentration differences

PAGE 139

139 (Fi gure 5 4). Although skewed for the expected segregation ratio ( 28.311), the marker still presented three distinct genotypic classes. To test for the artificial creation of genotypic classes by preventing PCR from running to saturation, marker M261 (no sequence difference) was run for the population and amplifica tion was stopped at 30 cycles during exponential growth. At a sensitivity of 0.30, the parents and 89 of 90 F 2 progeny were placed in the same genotypic class, indicating artificial differences were not being created.

PAGE 140

140 Table 5 1 DNA sequence difference s detected between for initial ten monomorphic simple sequence repeat (SSR) markers in winter squash, Cucurbita moschata Quality Value Q a SSR p rimer Polymorphism detected Base pair difference DNA s equence B ase pai r n umber P 1 P 2 M009 b SNP 1 CTG CCG 24 11 57 M042 Deletion, P 1 c 6 TCGTCG P 2 86 91 M066 Deletion, P 2 d 2 CT P 1 21 22 M120 Deletion, P 2 4 TCTC P 1 42 45 SNP 1 ANT e ATT 60 49 49 SNP 1 TNT TTT 73 52 59 SNP 1 TNT TTT 77 35 10 SNP 1 GNG GTG 80 36 45 M256 SNP 1 TCT TGC 23 8 9 M261 none P039 Deletion, P 2 1 A P 1 4 P098 Deletion, P 2 3 AAG P 1 39 41 P224 SNP 1 CACT CCAT 14 37 18 SNP 1 CACT CCAT 15 14 17 P235 SNP 1 TGT GTA 3 4 1 SNP 1 TGT GTA 4 4 1 SNP 1 TGT GTA 5 7 4 a Quality values exceeding the baseline value of Q=20 indicate that correct base pair identification has a probability greater than 99%. b C. moschata genome. Markers beginning with C. pepo genome. c 1 d 2 e N indicates base calling software could not assign a base at this location. The site is a putative SNP.

PAGE 141

14 1 Table 5 2 Summary of melt ing curve ge notyping of scorable simple sequence repeat (SSR) marker s, prior to DNA concentration adjustments based on Cp value, in winter squash ( Cucurbita moschata ) with 90 F 2 progeny resulting from a cross F2 individuals SSR primer a P1 allele b value Adjusted c F1 allele value Adjusted P2 allele value Adjusted value Adjusted Total F2's cored d M042 25 0.278 44 0.022 2 1 0.100 0.400 89 M009 16 1.878 50 0.556 4(20) e 15.211 0.100 17.64 4 2.533* 90 P098 3(1) 16.900 15.211 65(1) 8.889 9.800 18 0.900 26.689 27.600 88 7(1) 0.4 10.678 9.344 56 (6 ) 2.689 6.422 17(1) 1.344 0.900 14.711 16.667 88 No significant difference of expected segregation ratio of 1:2:1, a All markers ru n under standard LightCycler 480 protocol with modifications made to align protocol to standard PCR protocol for squash, C. moschata and C. pepo. b Sensitivity level of 0.30 was used to determine number of individuals for genotyping in each class, unless otherwise marked. c Adjusted chi square value with different genotypes with similar melt curve shape included. d Total F 2 individuals scored e Values in parentheses indicate number of indivi genotyped into one of the three main genotypic classes using difference plot analysis.

PAGE 142

142 Table 5 3 Estimates of DNA concentration in winter squash, Cucurbita moschata and F 2 progeny resulting from a cross between as measured by crossing point (Cp) value in the LightCycler 480 software using simple sequence repeat (SSR) marker M066 and compared to absorbance readings from the Eppendorf Biophotometer and the Thermo Scientifc Genesys spectrophotometer Sample Average Cp Value a Difference in Cp value from P 1 Concentration difference estimate based on CP value BioPhotometer Concentration b Genesys Concentration b P 1 18.675 0 0 302.5 517.5 F 1 19.7 1.02 5 2x 523.1 625.0 F 2 (95) 23.005 4.33 16x 61.8 77.5 F 2 (80) 22.52 3.845 16x 213.1 652.5 F 2 (11) 22.83 4.155 16x 467.8 605.0 a Average based on Cp values determined from M066 A and M066 B. b Concentration readings (ng/l) were determined from absorbance of 260mn.

PAGE 143

143 Figure 5 1. Normalized and temperature shifted melting curves for simple sequence repeat (SSR) marker M042 showing three genotypic classes at sensitivity of 0.30 for 90 F 2 Cucurbita moschata F 2 individuals that inherited the BB allele are denoted by red, the heterozygotes by blue, and the SDub alleles by green.

PAGE 144

144 Figure 5 2. Normalized and temperature shifted difference plot of melt curve shape showin g three genotypic classes at sensitivity of 0.30 for 90 F 2 progeny (SDub) for simple sequence repeat (SSR) marker M009 in Cucurbita moschata F 2 individuals that inherited the BB allele are denoted by green, the heterozogotes by blue, and the SDub alleles by red. Genotypes, grouped separately, indicated by the purple curves, aligned with the shape of individuals with the SDub allele and were added to that genotypic class.

PAGE 145

145 1 4 2 5 3 6 1 4 2 5 3 6 1 4 2 5 3 6 1 4 2 5 3 6 1 4 2 5 3 6 7 8 7 8 Figure 5 1:20, (3) 1:40, (4) 1:80, (5) 1:160, (6) 1:320, (7) 1:640, (8)1:1280 for simple sequence repeat (SSR) marker M120. Amplification was st opped during mid exponential growth. A) BB replicate 1, B) BB replicate 2. B A D C B A

PAGE 146

146 Figure 5 4 Normalized and shifted melting curves for simple sequence repeat (SSR) marker M120 at sensitivity of 0.30 for 90 F 2 progeny resulting from a cross (BB) (SDub). A) Parents, F 1 and F 2 progeny prior to adjusting relative DNA concentrations based on Cp value and ending amplification during mid exponential amplification and B) Parents, F 1 and F 2 progeny after adjusting rel ative DNA concentrations based on Cp value and ending amplification during mid exponential amplification. F 2 individuals that inherited the BB allele are denoted by blue, the F 1 alleles by red, and the SDub alleles by green. Unknown individuals, indicated by maroon curves, were genotyped based on similarity of melt curve to the above known genotypes. A B B A A

PAGE 147

147 CHAPTER 6 CONCLUSIONS A range of color within pumpkins and squash was identified, using L*a*b* color space values, and these color space values were correlat ed with different carotenoid types and concentrations. Strong correlations were found between color value a* and total carotenoids ( r = 0.91) and color value b* a nd chroma with lutein ( r = 0.87). G enetic variation should make it possible to increase the nut ritional value through crossing and selection from within and among the different types with high levels of carotenoids. These close associations will assure that indirect selection for high carotenoid content within pumpkin and squash breeding material wi ll be successful, easy to implement, and inexpensive. Heritability and gene action was determined for flesh color in winter squash in both Cucurbita moschata and Cucurbita pepo F 2 BC 1P1 and BC 1P2 segregating populations. Broad sense heritabilities rang ed from 0.19 to 0.82 for L*, 0.12 to 0.32 for a*, 0.40 to 0.93 for b*, 0.36 to 0.92 for chroma and 0.14 to 0.15 for hue across all three crosses. Additionally, transgressive segregation for color space values a*, b* and chroma was examined and was identif ied in one C. pepo and the C. moschata population. Color space values that do not have transgressive segregation in crosses suggest that the population means could be shifted and color within squash flesh can be increased over time. Color space values th at do have transgressive segregation in crosses suggest that transgressive individuals can be used to increase flesh color through breeding with fewer generations. However, a breeding challenge is presented due lower broad sense heritabilities for color s pace values and the presence of narrow sense heritabilities with negative, essentially zero, values. This indicates that flesh color has a large non

PAGE 148

148 genetic variance which needs to be more adequately examined and controlled. Planting F 2 and backcross pop ulations within the same year, in the same location, or over multiple years and locations may allow better estimates of heritability. A linkage map of a F 2 population segregating for flesh color in C. moschata (2n=2x=40) was created from 235 RAPD and 42 SS R markers. A total of 21 linkage groups were obtained at a LOD score threshold of 3.0. Creation of this linkage map was then used to identify regions of the genome associated with squash flesh color. QTLs were detected on LG7 and 18 for L*, LG3, LG10, LG1 8 for a*, LG 5 and LG17 for b* and chroma and LG 3, LG13, LG14, LG15, and LG19 for hue However, given the low heritability estimates for some of the traits, these QTL assignments are tentative and await validation in different populations and environmen ts. Additionally, one way to strengthen the linkage map and thereby strengthen the associations identified in the QTL analysis would be to use the candidate gene approach. Sequences identified in the carotenoid pathway in other plant crops and/or in rela ted species within Cucubitaceae such as Cucumis sativus (cucumber) or Cucumis melo (melon), could be used to genotype the current F 2 population. Homology between the linkage groups obtained in this study was also examined with the most recently publishe d SSR based map in Cucurbita which examined the synteny of the C. moschata with the C. pepo genome (Gong, et al., 2008b). Homology was identified for several linkage groups. However, due to a large number of SSR markers on the Gong et al. C. moschata map identified as monomorphic in the Recovery of these monomorphic SSR markers is needed before sufficient alignment

PAGE 149

149 can occur with the published reference linkage map. With i ncreased recovery of monomorphic SSR markers using high resolution melting (HRM) real time PCR analysis, QTL information obtained here could be compared with other Cucurbita populations through map alignment. This would enable similar regions associated w ith color to be identified in other populations. In addition, as more QTL studies and map alignments become available within the Cucurbita genus, multiple traits can be used in squash breeding efforts.

PAGE 150

150 LIST OF REFERENCES Agarwal, M, N. Shrivastava, an d H. Padh. 2008. Advances in molecular marker techniques and their applications in plant species. Plant Cell Rep 27:617 631. Alpern, M. 1971. Rhodopsin kinetics in the human eye. J. Physiol. 217:447 471. Ameny, M.A. and P.W. Wilson. 1997. Relationsh carotene contents in white fleshed African sweetpotatoes ( Ipomoea batatas Lam). J. Sci. Food. Agric. 73: 301 306. Ando, M., H. Toyohara, and M. Sakaguchi 1992. Post mortem tenderization of rainbow trout muscle caused b y disintegration of collagen fibres in the pericellular connective tissue. Nippon Suisan Gakkaishi 58: 567 570. Arias, R., T. C. Lee, L. Logendra, and H. Janes. 2000. Correlation of Lycopene measured by HPLC with the L*, a*, b* color readings of a hydro po nic tomato and the relationshi p of maturity with color and lycopene content. J. Agric. Food. Chem. 48:1697 1702. Bang, H., A.R. Davis, K. Sunggil, D.L. Leskovar, and S.R. King. 2010. Flesh color inheritance and gene interactions among canary yellow, pale yellow, and red watermelon. J. amer. Soc. Hort. Sci. 135: 362 368. Bates, D.M., R.W. Robinson, and C. Jeffrey. Eds. 1990. Biology and untilization of the Cucurbitaceae. Comstock. Ithaca, NY. Berset, C., and P. Caniaux. 1983. Relationship between colou r evaluation and chlorophyllian pigment content in dried parsley leaves. Journ. Food Sci. 48 : 1854 1857. Bisognon, D.A. 2002. Origin and evolutiona of cultivated cucurbits. Cincia Rural. 32:715 723. Bretting, P.K. and M.P. Wildrlechner. 1995. Genetic m arkers and plant genetic resource management. Plant Bredd. Rev. 13:11 86. Brown, J.S., R.J. Schnell, T. Ayala Silva, J.M. Moore, C.T. Tondo, and M. C. Winterstein. 2009. Broad sense heritability estimates for fruit color and morphological traits from op en pollinated half sib mango famil ies. HortSci. 44:1552 1556. Brown, R.N., and J.R. Myers. 2002. A genetic map of squash ( Cucurbita sp p.) with randomly amplified polymorphic DNA markers and morphological markers. J. Am er. Soc. Hort Sci. 127 :568 575.

PAGE 151

151 Brothwell, D. and P. Brothwell. 1969. Food in Antiquity: A survey of the diet of early peoples. Fredrick A. Praeger Publishers. New York, NY. Chagne, D., J. Batley, D. Edwards, and J.W. Forster. 2007. Single nucleotide polymorphisms genotyping in plants In: Association Mapping in Plants, N.C. Oraguzie, E.H.A. Rikkerink, S.E. Gardiner, and H.N. De Silva, eds. Springer, New York. Chagne, D., K. Gasic, R.N. Crowhurst, Y. Han, H.C. Bassett, D.R. Bowatte, T.J. Lawrence, E.H.A. Rikkerink, S.E. Gardiner, and S.S. Korban. 2008. Development of a set of SNP markers present in expressed genes of the apple. Genomics 92:353 358. Cuevas, H.E., H. Song, J.E. Staub, and P.W. Simon. 2009. Inheritance of beta carotene associa ted flesh color in cucumber ( Cucumis sativu s L) fruit. Euphytica 171:301 311. Cuevas, H.E., J.E. Staub, and P.W. Simon. 2010. Inheritance of beta carotene associated mesocarp color and fruit maturity or melon ( Cucumis melo L.) Euphytica. 173:129 140. Croxford, A.E., T. Rogers, P.D.S. Caligari, and M.J. Wilkinson. 2008. High resolution melt analysis to identify and map sequence tagged site anchor points onto linkage maps: a white lupin (Lupinus albus ) map as an exemplar. New Phytol 180:594 607. pene concentration of tomato fruit can be estimated from chromat icity values. HortScience 27:465 466. Decker Walers, D. S., and H.D. Wilson. 1987. Allozyme variation in the Cucurbita pepo complex: C. pepo var. ovifera vs. C. texana. Syst. Bot. 12:263 273 Decker Walers, D.S., T.W. Walters, U Psluszny, and P.G. Kevan. 1990. Genealogy and gene flow among annual domesticated species of Cucurbita Can. J. Bot. 68:782 789. Decker Walters, D.S., T.W. Walters, C.W. Cowan, and B.D. Smith. 1993. Isozymic charact erization of wild populations of Cucurbita pepo. J. Ethnobiol. 13:55 72. Decker, D.S., S.M. Chung, and J.E. Staub. 2004. Plastic sequence evolution: a new pattern of nucleotide substitutions in the Cucurbitaceae. J. Molec. Evol. 58:606 614. De Koeyer, D. K. Douglass, A. Murphy, S. Whitney, L. Nolan, Y. Song, and W. De Jong. 2010. Application of high resolution DNA melting for genotyping and variant scanning of diploid and autotetraploid potato. Mol Breeding 25:67 90.

PAGE 152

152 Dijkhuizen, A. W.C. Kennard, M.J. H avey, and J.E. Staub. 1996. RFLP variation and genetic relationships in cultivated cucumber. Euphytica. 90:79 87. Ellergren, H. 2004. Microsatellites: Simple sequences with complex evolution. Nature Reviews Genetics 5:435 445. FAOSTAT. 2008. Crop produc tion stats. November 24, 2008. http://faostat.fao.org FAOSTAT. 2010. Countries by commodity stats. July 15, 2010. http://faostat.fao.org FAO/WHO. 2002. Human Vitamin and Mineral Requirements, Report of a joint Food and Agricultural Organization of the United States (FAO) and World Health Organization (WHO) http://faostat.fao.org Fehr, W.R. 1987. Heritability, p. 95 105. In: W.R. Fe hr (ed.) Principles of cultivar development, Volume 1, Theory and technique. Macmillian Publishing Company, New York. Francis, F.J. 1962. Relationship between flesh color and pigment content in squash. J. Amer. Soc. Hort. Sci. 81:408 414. Francis, F. 19 69. Pigment content and color in fruits and vegetables.Food Technol. 23 : 32 36. Francis, F.J. 1985. Blueberries as a colorant in gredient in food products. J Food. Sci. 50:754 756. Gamble E E 1962. Gene effects in corn (Zea mays L.): I. separation and relative importance of gene effects for yield. Can J Plant Sci 42:339 348. Garcia Mas, J., M. Oliver, H. Gomez Paniagua, and M.C. de Vicente. 2000. Comparing AFLP, RAPD and RFLP markers for measuring gengetic diversity in melon. Theor. Appl. Genet. 101 :860 864. Garrison, R. and E. Somer. 1995. Chapter 6, Trace Minerals: Chromimum, cobalt, copper, fluoride, selenium, zinc, and other minerals. In: Garrison, R. and E. Somer (eds.)The nutrition desk reference, 3 rd ed. Keats Publishing, New Canann, CT. G ill, J. T. and E. L. Jensen. 1968. Probability of obtaining negative estimates of heritability. Biometrics. 24: 517 526. Gong, L., G. Stift, R. Kofler, M. Pachner, and T. Lelley. 2008 a Microsatellites for the genus Cucurbita and an SSR based genetic link age map of Cucurbita pepo L. Theor. Appl. Genet. 117 :37 48.

PAGE 153

153 Gong, L., M. Pachner, K. Kalai, and T. Lelley. 2008 b SSR based genetic linkage map of Cucurbita moschata and its synteny with Cucurbita pepo. Genome. 51:878 887. Gross, J. 1991. Pigments in veg etables: Chlorophylls and carotenoids. VanNostrand Reinhold, New York, New York. Gu smini, G. and T. C. Wehner. 2006 Qualitative inheritance of rind pattern and flesh color in watermelon. J. Hered. 97: 177 185. Hallauer A R and J. B. M iranda. 1988. Quantitative genetics and maize breeding. Iowa State University Press, Ames IA. Hayman, B.I. 1958. The separation of epistatic from additive and dominance variation in generation means. Genetica. 31:133 146. Heinonen, M.I.,V. Ollilainen, E. K. Linkola, P.T. Varo, and P.E. Koivistoinen, 1989.Carotenoids in Finnish foods: vegetables, fruits and berries. J. Agric. Food Chem.37:655 659 Holden, J.M, A.L. Eldridge, G.R. Beecher, I.M. Buzzard, S.A. Bhagwat, C.S. Davis, L.W. Douglass, S.E. Gebhardt, D.B. Haytw itz, and S. Schakel. 1999. Carotenoid content of U.S. Foods: An update of the database. J. Food Comp. Anal. 12:169 196. Hirschberg, J. 2001. Carotenoid biosynthesis in flowering plants. Curr. Opin. Plant Biol. 4:210 218. Ignart, F. and N.F. Weeden. 19 84. Allozyme variation in cultivars of Cucurbita peop L. Euphtica. 33:779 785. Ihl, M. 1994. Correlation for pigment content through colour determination using tristimulus values in a green leafy vegetable, Swiss chard. J. Sci. Food Agric. 66 : 527 531. I VAC G 2004. IVAC G Report 1999. International Vitamin A Consultative Group, Washington D.C. Kabelka, E. A. and K. Young. 2010. Identification of molecular markers associated with resistance to squash silverleaf disorder in summer squash ( Cucurbita pepo ). Euphytica. 173:49 54. Kearsey M J and H.S. Pooni. 1996. In: M.J. Kersey and H.S. Pooni (eds.) The genetic analysis of quantitative traits. Chapman and Hall, London

PAGE 154

154 Knerr, L.D. and J.E. Staub, D.J. Holder and B.P.May. 1989. Genetic diversity in Cucumis sativus L. assessed by variation at 18 allozyme coding loci, Theor. Appl. Genet. 78:119 128. Konica Minolta Sensing, Inc., 1998. Konica Minolta Photo Imaging U.S.A. Inco. Mahwah NJ. Lebeda, A., M.P. Widrlenchner, J. Staub, H. Ezura, J. Zalapa, and E. K Cucurbits (Cucurbitaceae; Cucumis spp., Cucurbita spp., Citrullus spp.), p. 271 376. In: R.J. Singh (ed.) Genetic Resources, Chromosome Engineering, and Crop Improvement, Vol. 3 Vegetable Crops. CRC Press, Taylor & Francis Group, Boca Raton FL. Lee, Y.H., H.J. Jeon, K.H. Hong, and B.D. Kim. 1995. Use of random amplified polymorphic DNA for linkage group analysis in an interspecific cross hybrid F 2 generation of Cucurbita J. Korean Soc. Hort. Sci. 36 : 323 330. Lehmensiek, A., M. Sutherlan d, and R. McNamara. 2008. The use of high resolution melting (HRM) to map single nucleotide polymorphism markers linked to a covered smut resistance gene in barley. Theor Appl Genet 117:721 728. Lester, G.E. and F. Eischen. 1995. Beta carotene content of postharvest orange fleshed muskmelon fruit: Effect of cultivar, growing location and fruit size. Plant Foods Hum. Nutri. 49:191 197. Ling, P.; V. Ruzhitsky ; A. Kapanidis, and T. C. Lee. 1996. Correlation between color machine vision and colorimeter fo r food applications. Chemical markers for processed and stored f oods. ACS Symp. 631: 253 278. Lopez Anido, F., V. Cravero, P. Asprelli, E. Cointry, I. Firpo, and S.M. Garcia. 2003. Inheritance of immature fruit color in Cucurbita maxima var. Zapallito (Ca rrire) Millan. Cucurbit Genet. Coop. Rpt. 26:48 50. Lynch M, and B. Walsh. 1998 Genet ic and analysis of quantitative traits. Sinauer association, Inc, Sunderland, MA Mackay, J., C. Wright, and R. Bonfiglioli. 2008. A new approach to varietal identifi cation in plants by microsatellite high resolution melting analysis: application to the verification of grapevine and olive cultivars. Plant Methods 4:8. Meglic, V. and J.E. Staub. 1996. Inheritance and linkage relationships of allozyme and morphological loci in cucumber Cucumis sativus L.) Theor. Appl. Genet. 92:865 872.

PAGE 155

155 Montgomery, J., C.T. Wittwer, R. Palais, and L. Zhou. 2007. Simultaneous mutation scanning and genotyping by high resolution DNA melting analysis. Nat Protoc 2:59 66. Moravec, J., A. Lebeda and E. Kristkova. 2004. History of growing and breeding cucurbitaceous vegetables in Czech Lands, Progress in cucurbit genetics and breeding research, p. 21 38. In: Lebeda, A., and H.S. Paris (eds.) Proceedings of Cucurbitaceae 2004, the 8 th EUCAR PIA Meeting on cucurbit genet ics and breeding. Olomouc, Czech Republic. Palacky Univeristy in Olomouc. Morrison, D. 1990. Color and California State University. Masters Abstr.Int. 30 (02), 78. Mossle r, M. A., and O. N. Nesheim. 2003. Florida crop/pest management profile: squash. Univeristy of Florida IFAS Extension. CIR 1265. varieties of pumpkins. J. Food. Comp. Anal. 15:633 638. ODS/NIH. 2006. Dietary supplement fact sheet: Vitamin A and carotenoids. November 24, 2008. http://ods.od/nih.gov/factsheets/vitamina.asp Olson, J.A. 1999. Carotenoids and human health. Archivos Lation Americanos de Nutricion. 4 9: 7 11S. Olson, S.M., and E. Simonne. 2007. Vegetable Production Handbook for Florida. University of Florida. IFAS Extension. Ong S.H. and E.S.Tee. 1992 Natura l sources of carotenoids plants a nd o ils. In: Parker, L. (ed.) Methods in Enzymology. Vol 213: Caroteoids Chemistry, Quantity and Antioxidation. Academic Press, New Yor k, NY. Ooigen, J.W. and R.E. Voorrips. 2001. JoinMap 3.0, Software for the calculation of genetic linkage maps. Plant Research International, Wageningen, Netherlands. Ooigen, J. W. 2004. MapQTL 5, Software for the mapping of quantitative trait loci in experimental populations. Kyazma B.V., Wageningen, Netherlands. Paris, H. S., F. Baumkoler and A. Hanan. 2003. Fruit color inheritance in a cross of a dark colored accession with a light colored accession in Cucurbita pepo Cucurbit Genetics Cooperative Report 26:44 45. Paris, H.S. and R.N. Brown. 2005. The genes of pumpkin and squash. HortScience 40:1620 1630.

PAGE 156

156 Paris, H. S. 2008. Summer squash, p.351 379. In: Prohens, J. and F. Nuez (eds.) Vegetables I, Asteraceae, Brassicaceae, Chenopodicaceae, and Cucurbitaceae. In: Prohens, J., Nuez and M.J. Carena (eds.) Handbook of Plant Breeding. Springer, New York, NY. Pavia, S.A. and R.M. Russell. 1999. Beta carotene and other carotenoids as antioxidants. J. Am. Coll. Nutr. 18:426 33. Pitrat, M. 2008. Melo n, p.283 315. In: Prohens, J. and F. Nuez (eds.) Vegetables I, Asteraceae, Brassicaceae, Chenopodicaceae, and Cucurbitaceae. In: Prohens, J., Nuez and M.J. Carena (eds.) Handbook of Plant Breeding. Springer, New York, NY. Pollack, S. 1996. Squash and pum pkins: An economic assessment of the feasibility of providing multiple peril crop insurance. Economic Research Service. (202) 219 0505. Puchalski, J.T. and R.W.Robinson. 1978. Comparative electrophoretic analysis of isozymes in Cucurbita species Cucurbit Genet. Coop. Rep. 1 28. USDA. 2009. 2007 Census of Agriculture. Florida State and County Data. National Agricultural Statistics Service. Vol. 1. Geographic Area Series. Part 9. USDA. 2010. Vegetables 2009 Summary. National Agricultural Statistics Servi ce. Vg 1 2(10). USDA ARS. 2010. Nutrient Data Laboratory. USDA National Nutrient Database for Standard Reference. SR 23. October, 27, 2010. http://www.nal.usda.gov/fnic/foodcomp/search/ Ramak rishnan, T.V., and F. J. Francis. 1973. Color and carotenoid changes in heated paprika. J Food. Sci. 38: 25 28. Reed, G.H., J.O. Kent, and C.T. Wittwer. 2007. High resolution DNA melting analysis for simple and efficient molecular diagnostics. Pharmacog enomics 8:597 608. Reeves, M.J. 1987. Re evaluation of Capsicum color data. J. of Food Sci. 52:1047 1049. Robinson, H. F. R.E. Comstock, and P.H. Harvey. 1955. Genetic variances in open pollinated varieties of corn. Genetics. 40:45 60. Rodreiguez Ama ya, D.B. and M.Kimura. 2004. HarvestPlus handbook for carotenoid analysis. HarvestPlus Technical Monograph Series 2. Washington, D.C.

PAGE 157

157 Sacks E.J. and D.M. Francis. 2001. Genetic and environmental variation for tomato flesh color in a population of modern breeding lines. J Amer Soc Hort Sci. 126:221 226. Sanjur, O.I., D.R. Piperno, T.C. Andres, and L. Wessel Beaver. 2002. Phylogenetic relationships among domesticated and wilde species of Cucurbita (Cucurbitaceae) inferred from a mitochondrial gene: implic ations for crop plant evolution and areas of origin. Proc. Nat. Acad. Sci. 99:535 540. Santos, C.A. F. and P.W. Simon. 2006. Heritabilities and minimum gene number estimates of carrot carotenoids. Euphytica. 151:79 86. Seroczynska, A., A. Korzeniewska, J. Sztangret Wisniewska, K. Niemirowicz Szczytt and M. Gajewski. 2006. Relationship between carotenoid content and flower or fruit flesh colour of winter squash ( Cucurbita maxima Duch.). Horticulturae. 18:51 61. Schlotterer, C. 2004. The evolution of molecular markers just a matter of fashion? Nature Reviews Genetics 5:63 69. Simonne, A.H., S.J. Kays, P.E. Koehler, and R.R. Eitenmiller. 1993. Assessment of beta carotene content in sweet potato ( Ipomoea batatas Lam.) breeding lines in relation to die tary requirements. J. Food Compos. Anal 6:336 345. Staub, J.E., M.D. Robbins, and T.C. Wehner. 2008. Melon, p.241 282. In: Prohens, J. and F. Nuez (eds.) Vegetables I, Asteraceae, Brassicaceae, Chenopodicaceae, and Cucurbitaceae. In: Prohens, J., Nuez a nd M.J. Carena (eds.) Handbook of Plant Breeding. Springer, New York, NY. Steel, G.D., J.H. Torrie, and D.A. Dickey. 1997. Principles and Procedures of Statistics: A biometrical approach. McGraw Hill, New York. Swaider, J.M., and G.W. Ware. 2002. Other Vegetables Eggplants, Pumpkins and Squashes, Okra, Salsify, Celeriac, Florence Fennel, Orach, Parsley, p. 565 582. In: Swaider, J.M and G.W. Ware (eds.) Producing Vegetable Crops. 5th ed. Interstate Publishers, Inc., Danville, IL. Tadmor, Y., H.S. Par is, A. Meir, A.A. Schaffer and E. Lewinsohn. 2005. Dual role of the pigmentation gene B in affecting carotenoid and vitamin E content in squash ( Cucurbita pepo ) mesocarp. J. Agric. Food. Chem. 53:9759 9763. Takahata, Y.; T. Noda, and T. Nagata, T. 1993. HPLC determination of carotene content of sweet potato cultivars and its relationship with color values. Jpn. J. Breed 43: 421 427.

PAGE 158

158 Tang, G. 2010. Biconversion of dietary provitamin A carotenoids to vitamin A in humans. Amer. Journ. of Clin. Nutr. 91(suppl):1468S 73S. Tho m pson, W.A. and J.R. Moore. 1963. Non negative estimates of variance components. Technometrics 5: 441 449. Trumbo P, A.A. Yates S. Sch licker, and M. Poos. 2001. Dietary reference intakes: vitamin A, vitamin K, arsenic, boron,chromium, copper, iodine, i ron, manganese, molybdenum,nickel, silicon, vanadium, and zinc. J Am DietAssoc 101:294 301. Van Ooijen, J.W. & R.E. Voorrips, 2001. JoinMap 3.0, Software for the calculation of genetic linkage maps.Plant Research International, Wageningen, the Netherland s. West C E A. Eilander and M. van Lies hout 2002. Consequences of revised estimates of carotenoid bioefficacy for dietary control of vitamin A deficiency in developing countries. J Nutr. 132: Suppl 9:2920S 6S. Walters, D.R. and D.J. Keil. 1997. Cucu rbitaceae the cucumber family, p.334 225. In: Walters, D.R. and D.J. Keil (eds.) Vascular Plant Taxonomy, 4 th ed. Kendall/Hunt Publishing Co. Dubuque, IO. Wang, Y. H., T. Joobeur, R.A. Dean, and J.E. Staub. 2007. Cucurbits, p. 315 329. In: Kole, C. (ed. ) Genome Mapping and Molcular Breeding in Plants, Vol. 5 Vegetables. Springer Verlag Berlin Heidelberg. Warner J N 1952. A method for estimating heritability. Agron J 44:427 430. Watada, A.; K. Norris, J. Worthington, and D. Massie 1976. Estimation o f chlorophyll and carotenoid contents of whole tomato by light absorbance technique. J. Food Sci. 41 : 329 332. Wehner, T. C. and D.N. Maynard. 2003. Cucurbitaceae (V ine crops). In: Encyclopedia of Life Sciences. Nature Publishing. WHO, 2001.World Health Organizatio n: Global prevalence of vitamin A deficiency. Micronutrient Deficiency Information System, MDIS Working Paper No. 2. Wilson, D.H. J. Doebley, and M. Duvall. 1992. Chloroplast DNA diversity among wild and cultivated members of Cucurbita (Cucurb itaceae). Theor. Appl.Genet. 84:859 865. Wright S, 1968. The genetics of qualitative variability. In: Evolution and genetics of populations genetic and biometric foundations (Wright S, ed). Chicago IL: University of Chicago Press.

PAGE 159

159 Wu, S B., M. Wirthensoh n, P. Hunt, J. Gibson, and M. Sedgley. 2008. High resolution melting analysis of almond SNPs derived from ESTs. Theor Appl Genet 118:1 14. Xu, Y. 2010. Molecular Plant Breeding. CABI Publishing, Oxfordshire, United Kingdom. Zile, M.H. 2001. Function of Vitamin A in vertebrate embryonic development. Journ. Of Nutr. 131:705 708. Zraidi, A. and T. Lelley. 2004. Genetic map for pumpkin Cucurbita pepo using random amplified polymorphic DNA markers. p. 507 514. In: Lebeda A., Paris H.S. (eds.) Progress in cuc urbit genetics and breeding research. Proceedings of the Cucurbitaceae 2004, 8th EUCARPIA meeting on cucurbit genetics and breeding, Olomouc. Zraidi, A., G. Stift, M. Pachner, A. Shojaeiyan, L. Gong, and T. Lelley. 2007. A consensus map for Cucurbita pepo Mol.Breed. 20 : 375 388.

PAGE 160

160 BIOGRAPHICAL SKETCH Rachel obtained her Bachelor of Science degree in h orticulture with dual minors in biology and plant pathology from The Pennsylvania State University in May 2006. While pursuing her undergraduate degree, Ra chel had the opportunity to have an undergraduate research experiences and two internship positions in horticulture. She served as a summer intern at Phipps Conservatory and Botanical Gardens in Pittsburgh, PA (2004) worked as a research assistant in Dr. Breeding and Genetics Program at Penn State (2004 2005), and interned at Ball Helix, a genetic research division on Ball Horticultural Company in W est Chicago, IL (2005). She also studied under Dr. Paul Backman in the Plant Patholog y Department and Penn State where she examined the effect of biological and cultural management strategies in orchard soils with apple replant disease (2005 2006). Rachel was awarded the Alumni Fellowship and began her Doctor of Philosophy degree in plant breeding and genetics in the Horticultural Sciences Department at the University of Florida in August 2006 in Genetics Program While studying for her doctorate she had the privilege of teaching the Vegetable Ga rdening class for non major undergraduate students and found a passion for teaching in addition to plant breeding. After graduation, Rachel will continue her career in plant breeding and genetics with a postdoctoral research position in the Blueberry Bree ding and Genetics Program at the University of Florida under the direction of Dr. James Olmstead. in plant breeding and genetics.