Genetic Analysis of Flower Colour Variation in A/Hum Schoenoprasum L. (Wild Chives)

Heredity 60 (1988) 253-256 The Genetical Society of Great Britain Received 4 June 1987

Genetic analysis of flower colour variation in A/hum schoenoprasum L. (wild chives)

J. P. StevensK and Department of Biology, University of York, S. M. Bougourd York YOl 5DD, U.K.

Genetic analysis of three naturally occurring flower colour variants of Allium schoenoprasum revealed the involvement of three diallelic loci in the control of flower colouration: Wiw controls the presence/absence of pigment, Did controls the distribution of pigment in the tepals (perianth segments), and P/p controls pigment hue. There is no evidence of linkage betveen the three loci.

INTRODUCTION Flower colour variants that show simple Men- delian inheritance provide genetic markers which Flowercolour variants are found relatively are useful in studies of the breeding system. The frequently in species that have anthocyanin pig- following is a report of the genetic analysis of mentation (Kay, 1982). Allium schoenoprasum L. purple-tip, pink and white flowered morphs of A. (wild chives) is no exception. The flowers of this schoenoprasum. species are normally purple, but occasional white flowered plants have been reported from populations in Europe (Steam, 1980) and Canada (Scog- MATERIALS AND METHODS gan, 1978). During studies of the population cytogenetics Description of flower colour morphs and breeding behaviour of A. shoenoprasum in Europe, four white flowered plants and one of Purple Tepals and tepal veins purple (occasionally each of two further flower colour variants (desig- tepal veins green in plants from the population nated pink and purple-tip) were found. The white near Tintagel); anthers pink or purple (except and pink morphs occurred in a total sample of for occasional male sterile plants which have 1477 adult plants that were randomly collected, yellow anthers). This is the standard flower outside the flowering season, from 8 sites in Britain colour morph. and 10 in continental Europe; the purple-tip morph Purple-tip Apical half to third of each tepal and was found among a large number of plants that tepal vein purple, basal parts white and green were raised by D. S. Holmes from seeds and seed- respectively; anthers pink or purple. lings collected from the Wye valley (Powys, Pink Tepals light pink; tepal veins green or purple; Wales). The low frequency of flower colour anthers white. variants in the samples suggests that, in these popu- White Tepals and anthers white; tepal veins pale lations at least, such forms are the result of recur- green (occasionally greenish-purple in plants rent mutation and are probably not maintained by from Tintagel). natural selection. The only population that may There is a subtle, more or less continuous variation be an exception is from near Tintagel (Cornwall, in the intensity of tepal colouration within each of England), because four of the 96 plants sampled the three pigmented morphs. In addition, the distri- from this site were white flowered. bution of pigment in the tepals of pinks is occasion- *Presentaddress: Department of Botany and Microbiology, ally patchy. Patchiness may be environmentally University College of Swansea, Singleton Park, Swansea induced, because two plants scored as uniform SA2 8PP, U.K. pink in 1984 were rescored as patchy pink in 1985. 254 J. P. STEVENS AND S. M. BOUGOURD

Breeding programme ing pigment distribution is termed "D". Morphs in which pigment is restricted to the apical part of Fiveexperimental cross-pollinations were carried each tepal are assigned the genotype dd, and out: crosses 1—3 were between different pairs of purple and white flowered plants, cross 4 was morphs with uniformly coloured tepals D-. between a purple and a purple-tip morph, and cross 5 was between a purple-tip and a pink morph. Purple-tipxpink All but one of the parental plants were also self- Thepurple-tip parent used in this cross bred true pollinated. The pink, purple-tip and three white when selfed. The pink parent, however, segregated flowered parents originated from Carreg Cennen pinks and whites in a 3: 1 ratio (table 1). This (Dyfed, Wales), the Wye valley (Powys, Wales) suggests that the pink parent was heterozygous at and Tintagel (Cornwall, England) respectively. the locus controlling pigment production (Ww), The purple flowered parents were randomly selec- and that it was homozygous at another locus caus- ted from the Tintagel and River Wye population ing pink rather than purple pigmentation. This samples. locus is designated "P". Complementation occur- The flowers on female parents were emascu- lated when in bud and isolated in porous cel- red in cross 5, all the Fl plants being purple (table 1). Therefore, pink is recessive to purple, and pink lophane bags prior to cross-pollination, which was morphs are assigned the genotype pp and purple achieved by gently rubbing two freshly dehisced anthers on to each receptive stigma. Self-pollina- morphs PP. A new phenotype, which had tinges tions were carried out in a similar manner, but of pink restricted to the tips of the tepals and white omitting bud emasculation. F2 families were gener- anthers, appeared in the F2 generation of cross 5. ated by either selfing or sib-mating the Fl plants. This phenotype is designated pink-tip, and assigned the genotype W-ddpp. The Fl generation of cross 5 was expected to contain both WW homozygotes and Ww heterozy- RESULTS AND CONCLUSIONS gotes. All but one of the F2 families segregated white flowered plants, and therefore clearly arose Purple Xw/iite from heterozygous Ww parents. The expected number of each flower colour morph in this group The six parental plants used in crosses 1-3 all bred of F2 families was calculated using a trihybrid true when selfed (table!), which suggests that they ratio which was modified to account for the epis- were homozygous at the locus or loci controlling tatic effect of the w allele on the other colour genes. flower colour, although admittedly three of the self The one F2 family that did not segregate whites families were small. The Fl plants from all three arose from a cross between two Fl plants and crosses were purple flowered, and in the F2 gener- contained 41 progeny. The probability that the ation, purples and whites segregated in proportions parents of this family were both Ww heterozygotes that are not significantly different from 3: 1 ratios is extremely low (P =075'=7.54x 1O). There- (tablel). It is concluded that pigment production fore, it was assumed that at least one of the parents is controlled by a single locus showing normal was a WW homozygote, and the expected number Mendelian inheritance, and that white is recessive of each flower colour morph in the F2 was calcu- to purple. This locus is designated "W", in accord- lated using a standard Mendelian dihybrid ratio. ance with the symbolism recommended by Grant In neither group did the expected frequencies differ (1975, 166). White flowered morphs are assigned significantly from the observed, although the the genotype ww,andcoloured morphs W-. expected number of pink-tips in the group that did not segregate whites was rather low (2.6). Purplexpurple-tip Thepurple-tip parent, which was used in crosses DISCUSSION 4 and 5, bred true when selfed (table 1). The Fl plants of cross 4 were all purple, and in the F2 Biosyntheticpathways of pigment production in generation, purples and purple-tips segregated in flowers often involve several steps that are con- a 3: 1 ratio (table 1). Thus, it appears that the trolled by different genes. In such cases, pigment restriction of pigmentation to the tips of the tepals production may be blocked by mutation occurring is controlled by a single recessive allele showing at one of several loci. For example, five different normal Mendelian inheritance. The locus controll- loci are involved in the control of flower = = = = Table 1 Crosses to elucidate the mode of inheritance of pigment presence, hue and distribution. PP purple, PPT purple-tip, PK pink, PKT pink-tip, WH white Parental Fl F2 Cross Expected Chi- no. Phenotypes Genotypes PP PPT PK WH PP PPT PK PKT WH ratio square df P

I purple x white WWxww 3 0 0 0 20 0 0 0 4 3:1 089 1 >03 purple self ww 2 0 0 0 white self ww 0 0 0 10

2 purple x white WW x ww 28 0 0 0 130 0 0 0 39 3:1 033 1 >0.5 purple self WW 10 0 0 0 white self ww 0 0 0 45

0 0 5 3:1 061 1 3 purple x white WW x ww 13 0 0 0 22 0 >03 purple self WW 6 0 0 0 white self ww 0 0 0 2

4 purple x purple-tip DD x dd 26 0 0 0 106 35 0 0 0 3:1 0002 1 >0.9 purple-tip self dd 0 24 0 0 4 5 purple-tipx pink WWddPPx WwDDpp 33 0 0 0 83 21 17 10 35 27:16:9:9:3* 601 >0.1 21 8 6 6 0 9:3:3:1 519 3 >0.1 purple-tip self WWddPP (as for cross 4) pink self WwDDpp 0 0 20 8 3:1 019 1 >0.5

* 27 16 white:9 and those for crosses Expected numbers for crosses involving Ww El plants were calculated from the modified trihybrid ratio, purple: purple-tip:9 pink:3 pink-tip, involving WW Fl plants from the dihybrid ratio, 9 purple :3 purple-tip:3 pink:1 pink-tip 256 J. P. STEVENS AND S. M. BOUGOURD pigmentation in Antirrhinum majus (Harrison and REFERENCES Stickland, 1974; Stickland and Harrison, 1974), ENNOS,B.. A. AND CLEGG, M. T. 1983. Flower color variation three genotypically different, but phenotypically in the morning glory, ipomoea purpurea. J. Hered., 74, white, morphs occur in Hyacinthoides non-scripta -247-250. (Endymion non-scriptus) (Stickland and Harrison, GRANT,V. 1975. Genetics of Flowering Plants. Columbia Univ. 1977), and segregation at three different lqci Press, New York. accounts for the 8 different flower colour HARRISON, B. J. AND STICKLAND, R. G. 1974. Precursors and genetic control of pigmentation. II. Genotype analysis of phenotypes of Ipomoea purpurea (Ennos and pigment controlling genes in acyanic phenotypes in Clegg, 1983). In Allium schoenoprasum, three dial- Antirrhinum majus. Heredity, 33, 112-115. lelic loci involved in the control of flower colour KAY, Q. 0. N. 1982. Intraspecific discrimination by pollinators have been identified (W/w, P/p and D/d). The and its role in evolution. In Armstrong, J. A., Powell, J. possibilities that more loci, or additional alleles at M. and Richards, A. J. (eds.) Pollination and Evolution, Royal Botanic Gardens, Sydney, pp. 9-28. these known loci, exist in natural populations can- SCOGGAN, H. j1978.The Flora of Canada. NatI. Mus. Nat. not be ruled out. There is no evidence (from cross Sci. Canada, Publ. Bot., 7 Ottawa. 5) of linkage between "W", "D" and "P", and all STEARN, W. T. 1980. Allium L. In Tutin, T. G., Heywood, V. three loci are suitable for use as genetic markers. H., Burges, N. A., Moore, D. M., Valentine, D. H., Walters, S. M. and Webb, D. A. (eds.) Flora Europaea, Vol. 5, Cambridge Univ. Pres, Cambridge, pp. 49-69. STICKLAND, R. G. AND HARRISON, B. .i. 1974. Precursors and genetic control of pigmentation. I. Induced biosynthesis Acknowledgements We are grateful to Mrs S. Lewis for scoring of pelargonidin, cyanidin and delphinidin in Antirrhinum flower colours in some experimental progenies, to Mr K. Part- majus. Heredity, 33, 108—112. ridge for help with maintaining the plants, and to Dr D. P. STICKLAND, R. G. AND HARRISON, B. .i. 1977. Precursors and Stevens and Dr T. J. Crawford for their critical discussion and genetic control of pigmentation. III. Detection and distri- comments. The financial assistance of S.E.R.C. is gratefully bution of different white genotypes of bluebells (Endymion acknowledged. species). Heredity, 39, 327—333.