Breeding for color and lycopene content in adapted tomato germplasm

David M. Francis, Barb Franchino, Troy Aldrich, and Brenda Schult Dept. of Horticulture and Crop Science, Ohio Agricultural Research and Development Center, 1680 Madison Ave., Wooster, OH 44691.

Steven J. Schwartz, Minhthy Nguyen, and Charlotte Allen Dept. of Food Science and Technology, The Ohio State University, 2001 Fyffe Court, 059B Howlett Hall, Columbus, OH 43210.

Introduction

This project aimed to develop strategies and tools to improve the efficiency of breeding for color and lycopene content within breeding lines and varieties adapted to the Great Lakes region.Our approach of working within cultivated tomato is counter to conventional genetic mapping approaches which rely on wild species as a source of new traits and genetic variation. However the variation that occurs within adapted germplasm is more likely to be the useful for meeting short to medium range breeding goals. Therefore we hoped to develop techniques that would allow us to select for improved color and lycopene content within adapted breeding material and varieties. Our approach, sampling techniques, and varieties were chosen to maximize breeding applications for peeled product tomatoes.

Specific goals were:

To develop a sampling strategy to accurately measure the genetic contribution to fruit color in adapted varieties and breeding lines.
To measure the genetic contribution to lycopene content.
To test the hypothesis that genes other than “crimson” contributed to color and lycopene content in a breeding population.
Use genetic information to select parents for hybrids with improved color and lycopene content.

Materials and Methods

Plant Materials
A sample of 40 breeding lines and cultivars was chosen to represent the germplasm of processing tomatoes typically grown in the Great Lakes region. This breeding population was used to measure color and lycopene content and sampling strategies were developed based on the statistical analysis of quality data. Subsequent experiments were conducted with a smaller population of the best 19 lines and varieites and on segregating generations derived from crossing.

Trials were grown at The Ohio State University’s Vegetable Crops Branch in Fremont, OH, the Horticulture Farm in Wooster, OH, and in on-farm trials in grower’s fields. Production practices were as recommended for commercial growers. At the Fremont location, a Johnson mechanical harvester was used for once over harvest and fruit were sorted by hand into ripe, green, and cull categories. All other locations were harvested by hand. The ripe fruit were graded at 98% to 100% usable by United States Department of Agriculture inspectors and provided the raw material for color analysis. A randomized complete block design was used. Each block had one plot per genotype. Plots were 6.17 m in length consisting of 20 plants per plot and were harvested as a unit.

Color Measurement

Objective measurements in CIELAB color space were obtained with a CR-300 colorimeter (Minolta, Ramsey, N.J.). CIELAB is a three-dimensional color space, where L* represents white to black and the a*-b* plane may be visualized as a color wheel that is lighter or darker depending on the level of L*. For a typical observer, a ~1 unit difference in CIELAB space is slightly perceptible, and a ~2 unit difference is noticeable. A colorimeter measures the red, green, blue and total amount of light reflected from an object. The CR-300 colorimeter had an 8 mm diameter measuring area, a d/0- illuminating and viewing geometry, and used illuminant C. Chroma (C*) was calculated with the formula: (a*2 + b*2)1/2. Hue angle, in degrees, was calculated as: (180/PI) * cos-1 (a*/C*).

Fruit flesh was exposed for measurement by cutting the proximal end of the tomato transversely with a sharp knife, such that only the pericarp and the top of locular partitions were visible. The gelatinous placental tissue was not measured. Two measures were made on opposite sides of the exposed fleshy surface.

Statistical Methods

Main analysis: 40 genotypes. Mixed model analyses of variance were performed with the SAS procedure MIXED. Analyses were conducted for five dependent variables: L*, a*, b*, chroma, and hue. Genotype (G) was considered a fixed effect. The random effects were, year (Y), block within year (B\Y), G x Y, G x B\Y, and Fruit\GBY. In the statistical analysis plot is equivalent to genotype by block within year (GXB\Y); within plot variation equals Fruit\GBY; and within fruit variation is equivalent to error. Degrees of freedom were calculated via the Satterthwaite option. Standard errors for proportions of environmental variance were estimated by the method of Dickerson. Means were obtained with the lsmeans statement. Linear correlations of genotype means were calculated.

Sub-analysis: 19 genotypes. A data set that included a subset of 19 genotypes from the main analysis but additional years and locations was studied. Statistical procedures were analogous to the main analysis.
Marker analysis. The crimson gene, ogc , served as a control for marker analysis. The presence of the ogc allele was confirmed by pedigree and by inspection of flower color. An analysis of variance was conducted on the 40 genotypes with the SAS procedure GLM to study the effects of three levels of crimson (ogc/ogc, ogc/+, and +/+) on L*, a*, b*, chroma, and hue. The error term Gen(ogc) was used to test significance. The data were the means for each genotype calculated in the main analysis. Subsequent analysis were preformed using genotypic classes based on Amplified Fragment Length Polymorphisms (AFLPs).

Results

Table 1. Proportion of total variation for color. Data from sub-analysis of 19 genotypes, three years, and three locations each year.
Source	L	Hue	Chroma
Year	7.0	11.1	3.6
Location	24.9	28.9	5.0
Year X loc	0.0	3.6	19.5
Loc X rep(year)	1.5	3.6	0.0
Genotype	13.6	5.4	28.5
Genotype x Year	0.7	3.6	9.1
Genotype x loc	14.9	10.1	1.7
(G x loc x Y) (between fruit)	6.3	3.6	1.5
Error (within fruit)	31.2	30.1	31.1

A visual interpretation to Table 1

Genetic Variation

Between fruit variation.

Within fruit variation

Statistical analysis to partition variation allows us to determine which portion of the total variation is due to genetics and which is due to environment. The between fruit and within fruit components are magnified by the CIELAB scale where the difference between red and green are much larger than the differences between red and orange (see color wheel in “a visual interpretation to Table 1”).

Table 2. Correlation coefficients for lycopene content and color traits.
Source	L	a	b	Hue	Chroma
UV/vis	0.18 (P<0.001)	0.10 (P=0.006)	0.16 (P<0.001)	0.15 (P<0.001)	0.08 (P=0.01)
HPLC	0.21 (P<0.001)	0.11 (P=0.002)	0.16 (P<0.001)	0.17 (P<0.001)	0.07 (P=0.02)

Color is significantly correlated with lycopene content, though correlation coefficients are low to moderate. Lycopene content was measured using the a UV/vis ratio and using High Pressure Liquid Chromatography (HPLC). The shape of the curve in the graph suggests that linear correlation may not produce the best fit.

Lycopene content vs L value

Table 3. Analysis of variance to determine linkage between markers and color traits in a breeding population of 40 genotypes.
Marker	L P	R²	Hue P	R²	lycopene P	R²
og^c	0.01	0.27	0.01	0.26	0.02	0.11
C1	0.47		0.57		0.74
C2	0.01	0.28	0.05	0.10	0.05	0.09
C3	0.89		0.78		0.14	0.05
C4	0.64		0.67		nd
C5	0.66		0.54		nd
C6	0.01	0.29	0.04	0.11	nd
C7	0.11		0.50		nd
C8	0.85		0.25		nd

Table 4. Lycopene levels averaged over two years for selected genotypes

T Grouping

Mean

Gen

13.378

Ohio 9241

_{(ogc/ogc, C2/C2)}

13.033

Ohio 9242

_{(ogc/ogc, C2/C2)}

11.665

Heinz 9423

_{(ogc/ogc, +/+)}

11.595

Ohio 7983

_{(+/+, C2/C2)}

9.925

PS 696

_{(+/+, +/+)}

9.720

OX 38

_{(+/+, +/+)}

9.393

Ohio 8245

_{(+/+, +/+)}

Marker analysis.
By combining traditional breeding and the analysis of DNA landmarks called “molecular markers” we have identified a new locus (gene) that contributes to color and lycopene content. AFLPs, shown below, are a class of DNA based marker that can be used in the breeding and development of new tomato varieties. The markers allow us to group populations into genotypic classes for statistical analysis of triat data such as lycopene content and color (Table 3 and Table 5). Values of “P” below 0.05 are interpreted as evidence for linkage between the marker and the trait.

Table 5. Analysis of variance to determine linkage between markers and color traits in a segregating population. These analysis are used to confirm results from Table 3. F₂ population 98-8823:7814 X 1023 color data from 1998.
Marker¹	L* P	R²	a* P	R²	b* P	R²
C2	0.05*	0.30	0.17	0.16	0.04*	0.32
C3	0.34	0.88	0.99	-	0.22	0.14
C4	0.15	0.16	0.66	-	0.08	0.20

F₂ population 98-8823:7814 X 1023 lycopene data from 1999. Mean of marker class lycopene mg/100 gm
	P	R²	0	I
C2	0.047*	0.22	13.76	10.55
C3	0.94	-	11.12	11.04
C4	0.93	-	11.12	11.01

Conclusions

Sample size is important for accurately measuring the genetic contribution to color. We measure at least 24 fruit per plot, 2-4 replicate plots per location, and 3 locations each year.
Year, location, genotype, between fruit, and within fruit differences significantly contribute to variation for color. Genotype explains up to 28% of the variation, and it is this variation that we can use to improve the genetic potential of tomato varieties (Table 1).
Lycopene is correlated with color (Table 2). Genotype explains 29-43% of the variation for lycopene content. Unexplained (Error) variation accounts for the rest.

Ogc explains 18-36% of the genotypic variation for color and 11% of the variation for lycopene content (Table 3). There are significant differences within wt/wt and ogc/ogc categories which suggests that other genes may contribute to color.
We discovered a second locus linked to AFLP marker C2 which explains 10- 28% of the variation for color and 9% of the variation for lycopene content (Table 3 and Table 5). C2 is independent from Ogc.
In a segregating population used to confirm the results from the breeding population, C2 explains 16-32% of the variation for color and 22% of the variation for lycopene content (Table 5).
Marker-trait linkage was possible in a population of adapted varieties and breeding lines. The marker, C2, has potential to aid in selecting parents for hybrids with improved color and lycopene content.