American Rose Annual, 1983
Breeding For Red Colors in Roses
H. H. Marshall and L. M. Collicutt

Agriculture Canada Research Station, Box 3001
Morden, Manitoba, Canada r0g 1j0

The Rose breeding program at the Agriculture Canada Research Station in Morden, Manitoba, Canada has the objective of producing hardy, attractive roses for the Canadian Prairies. A range of pleasing colors is sought, as most existing hardy shrub roses bear pink flowers.

There is little doubt that flower color in roses has been studied for hundreds of years, but genetic results have been hard to interpret. Two publications by Dr. Walter E. Lammerts describing the inheritance of color based on visual identification of colors are among the best we have read. The development of paper and thin layer chromatography has provided a fairly simple means of separating mixed pigments. This separation of pigments permits a better understanding of the complex interactions between the different red pigments in roses.

The red pigments in roses all belong to the large family of anthocyanin pigments which provide most of the red colors in flowers, fruits and leaves. Tomatoes and poppies are among those bearing unrelated red pigments. Only three of the six or seven groups of anthocyanins occur in roses, but since each of these can occur in at least two forms and in varying amounts, a great range of pink and red colors is possible, as any rose lover knows.

Anthocyanidins are anthocyanins without the sugar or glucoside portion of the molecule. The three anthocyanidins found in roses are cyanidin, peonidin and pelargonidin. These closely related pigments differ at only one location on the molecule but each provides a distinctive color range. In newly opened roses each occurs mainly in the diglucoside form called cyanin, peonin and pelargonin. With age and under genetic control, each can change to the monoglucoside form which is less intensely colored. Anthocyanins are also pH indicators, being more red and fairly stable in acid conditions but more blue and fading in alkaline media.

Cyanin, the only red pigment in old garden roses, is the most common red pigment in roses. It occurs alone in many modern roses and together with one or both of the other red pigments in all other red or pink roses. In highly pigmented roses, it gives the classical blood red rose color and when diluted produces varying rose shades. There is a shift towards a more blue hue due to dilution because cyanin absorbs light less strongly in the blue than in the green and yellow wavelengths. The light not absorbed is reflected, thus giving the red or slightly blued pink shades. This should not be confused with blueing due to chemical change with age. A modified form of cyanin, again under genetic control, gives the lavender shades of the so-called blue roses.

Peonin, as suggested by the name, is the common red pigment found in peonies. It occurs frequently in the Rosa sections Cinnamomeae, Carolinae and Minutifoliae. Although there have been no reports of a rose containing only peonin, it is frequently the most prominent anthocyanin pigment present. It occurs in R. rugosa and many of its hybrids, where it is responsible for the pinkish or purplish shades of red. It also contributes to the cardinal red colors of cultivars such as Europeana and Adelaide Hoodless.

Pelargonidin, the pigment named from the common garden geranium of the genus Pelargonium, is responsible for the scarlet and shrimp pink shades found in both pelargoniums and roses. Strangely enough it has not been reported in any wild rose and is known to occur only in recently originated roses such as Independence and Tropicana. Pelargonin also seems to occur only in the presence of cyanin where it may or may not be associated with peonin.

In our study at Morden, pigment extracts from about 3000 roses were analyzed from 1973 to 1981. These samples were taken from native species, hardy and nonhardy cultivars and a large number of seedlings. Analyses of the results from 45 families of seedlings has given some insight into the color relationships found in roses. Paper chromatography was used to separate pigments. Amounts of pigments after separation were rated in daylight. These ratings were compared to those rated under long wave ultra violet (UV) light where the pink fluorescence of peonin and the yellow fluorescence of pelargonin served to confirm their identification and rating. Light absorption curves for the three pigments from roses and from other sources were developed using a spectrophotometer as a further check.

Our study showed that each of the three red pigments was highly heritable and seemed to be inherited quantitatively. Cyanin and peonin ratings showed some dominant genetic characteristics, while pelargonin was in part recessive. The expression of all three appeared to be controlled by few to many genes with the result that each occurred at several levels of pigmentation.

There were also interactions among the three pigments. A positive correlation between cyanin and either or both of peonin and pelargonin occurred and may have been due to genes that control the amount of any form of pigment. Such a gene could complicate attempts to eliminate cyanin and concentrate peonin or pelargonin. There was a small negative correlation between peonin and pelargonin suggesting that the two pigments may compete in some way and therefore there would be difficulty in breeding a rose strongly pigmented with both peonin and pelargonin. We will return to this point later.

Although segregation ratios were of limited value because of the small families tested and the large numbers required for tetraploid analysis, valuable observations were noted. Cyanin apeared in over 99 percent of all seedlings but never in a full range from white to deep red in one family. The usual range was about one-third to one-half of the full range, usually near the mean of both parents. Segregation for pelargonin seemed to be the least complicated behaving as a recessive gene in most instances, but also appearing from several crosses where one or both parents were known to carry genes for peonin but not for pelargonin. Segregation for peonin was more complex but suggested that it was not present in a homozygous form in any parent. Peonin segregated in each of the three native species tested, suggesting that some factor maintains a segregating condition. There is the possibility that peonin is a sub-lethal gene as four plants selected for unusually high peonin grew poorly and were sterile.

The relation of pelargonin to peonin appears to be one of competition as mentioned previously. The fact that there was only a small negative correlation between the two precludes these pigments being alleles. However, some of the genes appear to be the same for the two pigments as pelargonin will segregate from parents containing only peonin. If pelargonin was recessive it could not be expressed by the progeny of parents having genes for peonin unless part of the genes for peonin were identical to that for pelargonin. This could also explain the occurrence of pelargonin only in recent cultivated roses. The old roses of Europe and Asia carried genes for cyanin only. Peonin was possibly introduced from Austrian Copper when yellow was introduced into hybrid roses. Peonin has been reported in this old cultivar. R. rugosa and R. roxburghii are among other peonin bearing roses that have been used as parents. Segregation could have separated genetic factors for peonin which, when combined with other genetic factors resulted in the pelargonin pigment appearing where no scarlet color had been known previously. [according to Eugster and Märki-Fischer, 'Hansa' is colored by almost pure peonin.]

The pathway to color in roses seems to lead from early precursor chemicals through several steps of flavonoids to one, two or three anthocyanins. As found in snapdragons, the pathway may be blocked early, giving the white of R. rugosa alba or later, giving the white of most hybrid tea roses. The pathway then branches leading to cyanin, and then to pelargonin and/or peonin.

Rose breeders have known for some time that it is possible to breed for red color in roses, however this study pinpoints certain areas where difficulty can be expected and identifies a source of pelargonin among those roses which usually carry this chemical path through to peonin. There seems litle reason to expect a blue rose as the locus on the anthocyanin which gives blue color is never involved in Rosa or seemingly in the whole Rosaceae family.

CybeRose notes:

1) Pelargonin was first observed in the dwarf polyanthas which had no influence from R. rugosa, Austrian Copper or R. roxburghii. The relevant gene was apparently derived from R. multiflora which reportedly produces the related flavonol, kaempferol. However, since peonin does not "blue" as does cyanin, it is less likely to mask the orange color of pelargonin.

2) The appearance of pelargonin from parents which lack the pigment would not be surprising if one or both produced kaempferol, a possibility not explored in this study.

3) Pelargonin is apparently derived from dihydrokaempferol. In the cyanin pathway, dihydrokaempferol is converted to dihydroquercitin, which is then converted to cyanin. Peonin is produced from cyanin. Thus, there is no specific connection between peonin and pelargonin. However, if all three pigments are produced, the cyanin could be converted to peonin, which apparently does not "blue" when combined with co-pigments. Thus, pelargonin+peonin gives an orange shade, whereas pelargonin+cyanin(+co-pigments) has enough "blue" to counteract the orange tendency.

4) Dephinidin has been identified in the fruit of a peach cultivar Cresthaven and some blackberries, all members of the Rosaceae, though not in the flowers.

5) Pelargonin and kaempferol may be due to a less active form of the enzyme, flavonoid 3' hydroxylase (F3'H) which succeeds in converting only part of the dihydrokaempferol to dihydroquercitin.

6) In tetraploid species the chromosomes normally form pairs. Sometimes quadrivalents (groups or four) may form and DNA segments exchanged between chromosomes that do not normally meet. Among the progeny one may find specimens with extra copies of some genes, but a lack of others. Such modified chromosomes will not pair so readily with their "normal" counterparts. Thus, plants with extra peonin genes could be expected to be weak and have reduced fertility. The same phenomenon may be responsible for Lammerts' observation that many pelargonidin bearing roses are less vigorous than their siblings. Also, the observation that Tropicana gives only about 39% viable pollen.

7) Delphinin might be produced by a hyperactive form of F3'H (F3'5'H) which would hydroxylize both the 3' and 5' positions giving dihydromyricetin. This would then be converted by the existing enzymes to delphinin. The enzyme that converts cyanin to peonin, if present, might then convert delphinin (blue-violet) to petunidin (purple) and even malvidin (mauve). White and yellow flowered species should be examined for the presence of the flavonol, myricetin, which would imply the presence of F3'5'H.

8) It might also be useful to study the inheritance in reciprocal crosses to see if there is a cytoplasmic component in the inheritance of peonin. The fact that the high-peonin segregates tend to be sterile and weak hints that the cytoplasm may be involved.