// Twitter Cards // Prexisting Head The Biologist Is In: Making a new "Blue" tomato

Wednesday, October 15, 2014

Making a new "Blue" tomato

1. Tomato "Indigo Rose".
"Blue" is the color label applied to the new breed of anthocyanin rich tomatoes. "Indigo Rose" (image #1 at left) is the first officially available variety with the trait. The variety was bred at OSU, using two genes from wild relatives of tomatoes. The atroviolaceum ('atv') gene was introgressed from Solanum cheesemanii. The anthocyanin fruit ('Aft') gene is a transcription factor introgressed from S. chilense. The two genes combine to result in a tomato with dark purple anthocyanin pigment production when exposed to sunlight.

The high-anthocyanin traits managed to escape from the OSU breeding program before the official release, under the names "OSU Blue" or "P20". This variety was not yet stable and didn't taste very good to most people, but it did successfully introduce tomato breeders to the interesting traits a few years early. Breeders quickly took to trying to incorporate anthocyanin expression into better tasting types of tomatoes.

2. F2 tomatoes, showing pigment on fruit and calyx.
I've been growing a miniature tomato variety called "Tiny Tim" for the last several years. I saved a batch of open-pollinated seeds two years ago, as my previous batch was running out. Last year, one of the seedlings turned out to grow much faster and larger than all the others. It was the result of a cross to one of the other tomatoes growing the previous year. I grew several F2s this year, allowing me to identify the other parent as a "Roma" tomato.

Among the F2s, I noted a range of anthocyanin phenotypes in the fruit and leaves/stems. The anthocyanin pigment produced on the fruit when sun-exposed came in three levels (none, middle, and high in image #2.)

3. F2s (top two) & "Indigo Rose" (bottom).
The anthocyanin pigment produced in the calyxes also came in three levels (high, none, and middle in image #2), but independent from the fruit pigment. The pigment produced in the rest of the plant wasn't as obvious. The no-pigment plants were entirely green. the medium-pigment plants had the anthocyanin highlights on the calyxes and leaf edges. The high-pigment plants showed increased pigment over the entire plant where sunlight hit, at a level about half of that seen in "Indigo Rose" tomato plants (image #3).

4. Original; color-enhanced; postureized.
The high-pigment plants also appeared more of a red/brown color rather than the purple of "Indigo Rose" plants. Image #4 shows a section of the image #3 after using the color-enhance filter in GIMP (center) and then the posterize filter in GIMP (right). The enhanced images more clearly convey the difference in color which is visually seen on examining the plants. This either indicates a different mix of anthocyanin pigments, or is a visual artifact caused by the blending of green chlorophyll with the anthocyanin purple. I would need to do some chromatography or micro-dissection experiments to discriminate between these possibilities.

5. Anthocyanins on unripe "Tiny Tim".
The level of pigment in the F2s was a surprise as I hadn't noted any anthocyanin expression in the parent varieties. After seeing the F2s, I re-examined some "Tiny Tim" plants grown this year and found they did have anthocyanin expression. The fruit show a level of anthocyanin production comparable to the high-fruit-pigment F2s (image #5), but the small size of the fruit made it hard to notice. The F1 showed a pigment level like the medium-fruit-pigment F2s, suggesting it is a single trait with partial dominance. The color of the calyx and leaf/stem was comparable to the F2 plants with the middle level of pigment for each feature. Looking into the lineage of "Tiny Tim" suggests the middle-pigment trait was contributed from S. pimpinellifolium, used in the breeding to contribute small fruit size to "Tiny Tim". As anthocyanin pigments on the shoulder is common in many wild tomato relatives, I suspect the fruit trait also came from S. pimpinellifolium.

"Tiny Tim" is an open-pollenated variety, so it should be homozygous for any alleles impacting pigment production. The increased calyx/leaf/stem pigment intensity in the F2s over what is seen in "Tiny Tim" suggests the involvement of a second gene from the "Roma" parent that enhances the expression of the first gene. This second gene would have been hidden in "Roma" because that variety doesn't have any anthocyanin pigment production.

What are the expected genetics for this cross?

The fruit pigment appears driven by one gene. Under the model of partial dominance, the cross ...

1tt1tt x 1R1R

… produces an F1 …


… that shows a low level of anthocyanins in the fruit. Low amounts of anthocyanin pigment was noted in the fruit of the real F1. Selfing the F1 produces F2s …


… where 1/4 have high-pigment on the fruit (1tt1tt) and another 1/2 have low pigment on the fruit (1tt1R). I only grew 10 F2s this year, so it is hard to estimate real ratios, but all three color classes were observed.

The calyx/leaf/stem pigment appears to involve two genes. If we assume both involved alleles are recessive, the cross …

2tt2tt3TT3TT x 2R2R3r3r

… produces an F1 …


… that shows no anthocyanins in the calyx/leaf/stem. No anthocyanin pigments were observed in the calyx/leaf/stem of the real F1. Selfing the F1 produces F2s …


… where 1/16 are expected to express the recessive alleles from both parents and thus show the high-pigment trait. Another 3/16 are expected to express the recessive allele from "Tiny Tim" and show the medium-pigment trait. The remaining 12/16 should only have green chlorophyll evident in the unripe fruit. This year I grew 10 F2s and only one shows the high-pigment trait. 1/10 approximates 1/16 reasonably well for the numbers I grew. Some, but not all showed the middle-pigment trait. I didn't note exactly how many F2s showed the middle-pigment trait and they've begun dying back from the cold, so I will have to screen more F2s next year at an earlier stage to better estimate the true ratios of the different color classes.

The dark pigment of "Indigo Rose" fruit is due to the interaction of two traits, the anthocyanin fruit ('Aft') trait combined with the atroviolaceum ('atv') trait. The 'Aft' trait by itself only produces a small amount of pigment on the fruit shoulder. The 'atv' trait by itself only produces dark pigment on the calyx/leaf/stem of the plant.

If the fruit pigment in the F2s is driven by a single gene, as it appears, and two genes are responsible for the calyx/leaf/stem pigment, then 1/64 of the F2s will contain both high-anthocyanin traits.

6. Derived from S. hirsutum.

There are several anthocyanin traits floating around that have been introgressed from different wild tomato relatives.
  1. S. cheesemanii
    • "atv" gene: pigment throughout plant. Seen in variety "Indigo Rose" (image #1).
  2. S. chilense
    • "Aft" gene: pigment on fruit shoulder. Seen in variety "Indigo Rose". (image #1)
  3. S. hirsutum
    • [unnamed] gene: pigment on fruit shoulder, similar to "Aft". Described at maprc.blogspot.com. (image #6)
  4. S. peruvianum

  5. 7. Derived from S. peruvianum.
    • [unnamed] gene: pigment on fruit shoulder, similar to "Aft". Seen in variety "Purple Smudge". (image #7)
  6. S. pimpinellifolium
    • gene #1: pigment on fruit shoulder, similar to "Aft". Described here.
    • gene #2: pigment throughout plant, similar to "atv". Described here.
  7. Conventional tomatoes
The four fruit pigment traits and the two plant pigment traits seem to behave similar to the others in each category. Because the species are so closely related, the traits may represent different alleles of the same genes. If so, combinations of a trait from each category (like the "atv" and "Aft" in "Indigo Rose") should result in a strong increase in the total pigment produced relative to either trait alone, especially when the modifier trait (gene #3) is also present.

I've isolated a line that appears homozygous for gene #1 and one that appears homozygous for both genes #2 and #3. Unfortunately, crossing these two lines would simply recreate the F1 (heterozygous for all three traits) rather than help me generate a triple-homozyous line.

Comparing the lightly-pigmented fruit in images #6 and #7 to my pigmented F2s suggests they are showing a different mix of anthocyanins from the other "blue" lineages. I look forward to finding one of the rare segregants which contains all three genes, so I can find out!

  1. "Indigo Rose" tomato: http://extension.oregonstate.edu/gardening/purple-tomato-debuts-indigo-rose
  2. 'Aft' gene: http://www.esalq.usp.br/tomato/Aft.pdf
  3. 'atv' gene: http://www.esalq.usp.br/tomato/atv.pdf
  4. Escape of "P20" tomato : http://www.tomatoville.com/showthread.php?t=16989
  5. "Tiny Tim" tomato: http://tatianastomatobase.com/wiki/Tiny_Tim
  6. "Roma" tomato: http://tatianastomatobase.com/wiki/Roma
  7. "Orange Smudge" tomato: http://tatianastomatobase.com/wiki/Purple_Smudge