Brief Communications

Brief Communications

Inheritance of Resistance in Smooth Bromegrass to the Crown Rust Fungus N. J. Delgado, M. D. Casler, and C. R. Grau Common smooth bromegrass (Bromus inermis Leyss.) is octoploid, 2n ⫽ 8x ⫽ 56, with a genome structure of AAAAB1B1B2B2. Tetrasomic inheritance patterns have been observed in smooth bromegrass, but disomic inheritance is also expected from cytologic observations. Smooth bromegrass is susceptible to the crown rust fungus (Puccinia coronata Corda.). The objective of this study was to determine the inheritance of smooth bromegrass resistance to P. coronata. Seven smooth bromegrass clones, three susceptible and four resistant, were selfed and crossed in a diallel with bulked reciprocals. Inoculations were made with a population of P. coronata from PL-BDR1 smooth bromegrass. Resistance of smooth bromegrass to this population of P. coronata is complex. At least three genes appear to be involved in this host-pathogen interaction, one tetrasomic dominant gene which determines susceptibility (S) and two dominant genes (R1 and R2) that may be complementary and could be inherited either tetrasomically or disomically. Other genes may be involved in the smooth bromegrass–P. coronata interaction, possibly accounting for the lack of fit to expected ratios of some progeny. Heterogeneity for avirulence phenotype in the pathogen population may also have contributed to lack of fit of some progeny. Multiple resistance genes were detected because a pathogen population, likely consisting of genotypes with different genes for virulence, was used to challenge the host. Smooth bromegrass (Bromus inermis Leyss.) is a perennial cool season forage grass that belongs to the subfamily Festucoideae (Casler and Carlson 1995). Com-

480

monly grown smooth bromegrass is octoploid, 2n ⫽ 8x ⫽ 56, with a genome structure of AAAAB1B1B2B2 (Armstrong 1980; Ghosh and Knowles 1964). Meiosis is irregular with formation of bivalents and quadrivalents. The total bivalent number ranges from a low of 5 to a maximum of 28 with a predominance of 18 to 23. B. inermis is best classified as a partial allopolyploid species in view of the octoploid make-up of its chromosome complement and the chromosome configurations during meiosis (Armstrong 1973; Elliot and Love 1948; Hill and Carnahan 1957). Homologous chromosome pairing may be controlled genetically in tetraploid and octoploid smooth bromegrass; the high bivalent frequency might also be based on low chiasma frequencies (Armstrong 1991). Relatively few genetic studies have been conducted with the perennial grasses because most are polyploids, cross-pollinated and largely self-sterile with small floret size ( Dunn and Lea 1978). Tetrasomic inheritance patterns have been observed for seedling pubescence and a virescent trait in smooth bromegrass ( Dunn and Lea 1978; Ghosh and Knowles 1964). For seedling pubescence, the progeny showed partial dominance for pubescence and little differences between reciprocals ( Knowles 1980). The inheritance of resistance to the crown rust fungus has been most extensively studied in oat (Avena sativa L.). At least 68 genes for resistance to P. coronata have been reported (Simons et al. 1978). At least 40 of these genes derive from A. sterilis L., a wild relative of cultivated oats (Martens and Dyck 1989). Dominance, partial dominance, and complementary gene action are the most mentioned forms of gene action, although lack of dominance is not rare (Simons 1970). A combination of resistance and inhibitor genes has also been cited to explain some genetic data obtained in oats ( Dietz and Murphy 1930;

Wilson and McMullen 1997). The degree of suppression was dosage dependent; an increasing dosage of the resistance gene Pc62, relative to the suppressor, decreased the degree of suppression (Wilson and McMullen 1997). The degree of dominance for some crown rust resistance genes appear to be influenced by the genetic background upon which the gene is functioning (Simons and Murphy 1954). Crown rust resistance genes have been mapped to both the A and C genomes of oat ( Bush and Wise 1996), suggesting a potential genomewide distribution of crown rust resistance genes. Numerous crown rust resistance genes appear to be present in the diploid ryegrasses Lolium perenne L. and L. multiflorum Lam. While resistance generally appears to be dominant and of moderate to high heritability (Reheul and Ghesquiere 1996; Wilkins 1975), individual resistance genes have not been described in this genus. Crown rust resistance in Lolium appears to be controlled by a large number of genes with small effects and a few genes with large effects (Wilkins 1975; Wilkins 1978a,b). Resistance is highly race specific, with P. coronata populations showing rapid genetic changes on resistant Lolium clones (Wilkins 1978a,b). Coevolution of a gene-for-gene system is the most likely explanation for the observed race specificity and phenotypic plasticity of P. coronata on Lolium. Gene-for-gene specificity of the P. coronata–oat interaction is well documented (Simons 1970). A rust fungus was observed on smooth bromegrass leaves growing in the fields of the University of Wisconsin Agricultural Research Station (ARS) at Arlington, WI, and was classified as Puccinia coronata f. sp. bromi ( Delgado et al. 2001). The objective of this study was to determine the inheritance of crown rust resistance in smooth bromegrass. Relatively few genetic studies have been conducted in smooth bromegrass, generally indicating a tetra-

Table 1. Chi-squared tests of goodness-of-fit for a three-gene model (two dominant resistant genes, either disomic or tetrasomic in inheritance, and one monogenic dominant gene that conditions susceptibility and is tetrasomic in inheritance) Observeda

a b

Expecteda

Parent clonesa S

R

Total

S

R

Ratio

P

Type of progeny

2 (S) 5 (R) 6 (R) 13 (S) 15 (R) 22 (S) 28 (R) 2⫻5 2⫻6 2 ⫻ 15 2 ⫻ 28 13 ⫻ 5 13 ⫻ 6 13 ⫻ 15 13 ⫻ 28 22 ⫻ 5 22 ⫻ 6 22 ⫻ 15 22 ⫻ 28 2 ⫻ 13 2 ⫻ 22 13 ⫻ 22 5⫻6 5 ⫻ 15 5 ⫻ 28 6 ⫻ 15 6 ⫻ 28 15 ⫻ 28

4 2 93 0 96 1 8 55 100 85 61 76 23 78 36 12 52 3 86 0 0 0 131 50 46 95 56 112

19 8 132 115 128 2 16 212 260 194 208 248 161 208 210 60 192 12 158 184 136 250 238 63 80 180 84 203

— — 58 112 56 — — 159 195 146 156 207 134 173 175 45 144 — 118 184 136 250 104 28 35 79 37 89

— — 74 3 72 — — 53 65 49 52 41 27 35 35 15 48 — 40 0 0 0 134 35 45 101 47 114

1S:0Rb 7S:9Rb 7S:9R 35S:1R 7S:9R 1S:0Rb 7S:9Rb 3S:1R 3S:1R 3S:1R 3S:1R 5S:1R 5S:1R 5S:1R 5S:1R 3S:1R 3S:1R 3S:1Rb 3S:1R 1S:0R 1S:0R 1S:0R 7S:9R 7S:9R 7S:9R 7S:9R 7S:9R 7S:9R

— — 0.00 0.08 0.00 — — 0.75 0.00 0.00 0.15 0.00 0.40 0.00 0.85 0.37 0.50 — 0.00 1.00 1.00 1.00 0.70 0.00 0.82 0.37 0.05 0.78

Self Self Self Self Self Self Self S⫻R S⫻R S⫻R S⫻R S⫻R S⫻R S⫻R S⫻R S⫻R S⫻R S⫻R S⫻R S⫻S S⫻S S⫻S R⫻R R⫻R R⫻R R⫻R R⫻R R⫻R

15 6 39 115 32 1 8 157 160 109 147 172 138 130 174 48 140 9 72 184 136 250 107 13 34 85 28 91

S ⫽ susceptible, R ⫽ resistant. Chi-squared not computed due to insufficient number of progeny.

somic mode of inheritance, but cytologic evidence suggests that a disomic behavior might be expected for some genes, particularly those from the B1 or B2 genomes. The large number of resistance genes for P. coronata in oat suggests a genomewide distribution. In a complex polyploid such as smooth bromegrass, individual resistance genes may show either disomic or tetrasomic inheritance. Thus the present study may contribute to a better understanding of the breeding behavior of this species.

Materials and Methods Plant Materials Thirty-two smooth bromegrass clones, eight from each of four different populations (Alpha, WB19e, Lincoln, and WB88S) were screened for reaction to the crown rust causal organism P. coronata. Clones were screened in a randomized complete block design with four blocks, for two consecutive cuttings, with the entire experiment repeated twice in time. Seven clones were selected for their susceptible or resistant reaction to P. coronata, three susceptible and four resistant, to investigate the genetics of the resistance to this pathogen ( Table 1). Clones classified as resis-

tant showed no crown rust symptoms throughout the screening. A diallel with bulked reciprocals was made among the seven clones by planting paired-clone crossing blocks at the Arlington ARS in April 1997. Each crossing block consisted of two adjacent squares of 18 plants each, on a 60 cm spacing, and planted in a 3 ⫻ 3 configuration. The plants on the periphery of the square were from one clone but the plant in the center was from the opposite clone and was the only clone for which seeds were harvested. Adjacent squares represented reciprocals of a cross. The crossing blocks were isolated by at least 100 m from other smooth bromegrass to avoid pollen contamination ( Hittle 1954). The progeny of these crosses were evaluated for crown rust reaction during late summer and late fall 1997. The number of progeny evaluated per cross ranged from 12 to 243. All seven clones were also selfed in the greenhouse during the winter of 1997– 1998. Smooth bromegrass is a self-incompatible species with different degrees of incompatibility, so 12–18 plants per clone were used for selfing. The clones were transplanted from 30 cm3 plastic pots to 225 cm3 plastic pots in November 1997, allowed to grow for 2–3 weeks, and then tak-

en to cold frames during the first week of December. The temperature in the cold frames ranged from ⫺6⬚C to 4⬚C during unusually warm winter days and the plants were kept in the cold frames for 4 weeks. After the vernalization period, the plants were taken into the greenhouse and were isolated from one another by cheesecloth cages. Every morning during flowering time the plants were shaken to allow the pollen to flow among panicles within the cages. Flowering occurred from mid-February to mid-March. Seeds were harvested between April and May 1998 and were stored in the refrigerator until the end of May, when they were planted for their crown rust evaluation. Pathogen An isolate of crown rust was collected from infected smooth bromegrass leaves at the Arlington ARS. The population was classified as P. coronata f. sp. bromi and maintained with repeated inoculations on PL-BDR1 smooth bromegrass, a population developed at the U.S. Regional Pasture Research Laboratory ( University Park, PA) for resistance to Pyrenophora bromi ( Died.) Dresch. ( Berg et al. 1990). Urediospores of P. coronata were collected from infected PL-BDR1 bromegrass plants that were inoculated with the field population. The urediospores were harvested with a vacuum pump every 2 days until the fungus stopped sporulating. The inoculum was prepared by suspending 7to 10-day-old spores, stored at about 4⬚C, in deionized water containing 2 drops Tween 20 surfactant per 100 ml and agitated manually. The spore concentration used was approximately of 2–4 ⫻ 105/ml. The plants were inoculated 1 month after planting. An air sprayer was used to evenly apply about 25 ml of inoculum suspension per flat (72 seedlings), after which plants were placed in a dew chamber for 48 h and returned to the greenhouse benches under fluorescent light supplementation (14-h daylight) and a temperature range of 21⬚C–28⬚C. Symptoms appeared approximately 12–15 days after inoculation. Data Collection and Analysis The progeny were evaluated for infection type using a scale from zero to four where 0 ⫽ no uredinia, 1 ⫽ chlorotic flecks, 2 ⫽ small uredinia surrounded by necrosis with limited urediospore production, 3 ⫽ medium size uredinia with moderate sporulation, and 4 ⫽ large uredinia and abundant sporulation (Welty and Barker 1993).

Brief Communications 481

Plants with infection type 0 or 1 were classified as resistant and all others as susceptible. The chi-squared test was used to test for goodness-of-fit of tetrasomic inheritance models, beginning with a single gene. Dominance of susceptibility or resistance was hypothesized based on observed distributions of both selfed and crossed progeny. Because single-gene models fit poorly to most of the observed segregation ratios, the model was progressively expanded to hypotheses including multiple genes, epistasis, and multiple modes of gene action. Chi-squared tests were applied to each progeny family and a model was selected based on adequate fit to a majority of the progeny families.

Results and Discussion The results obtained for the progeny from the diallel crosses and selfs indicated that the inheritance of resistance of smooth bromegrass to P. coronata is complex, likely governed by multiple genes having different modes of action and possibly spread across both A and B genomes. In the segregation of progeny originated from the cross of susceptible (S) ⫻ resistant (R) clones, the number of susceptible plants is higher than the number of resistant plants, suggesting that susceptibility to P. coronata may be dominant over resistance ( Table 1). Two of the S ⫻ R crosses to clone 13 segregated in a 5 S:1 R ratio, suggesting tetrasomic inheritance. This indicated that clone 13 may be duplex for one susceptibility gene. The segregation obtained from the self of clone 13 supported this hypothesis. Progeny from other susceptible clones did not apparently segregate for this gene, suggesting a triplex or quadruplex condition in susceptible clones 2 and 22. Conversely 6 of 11 S ⫻ R progeny appeared to segregate for two dominant resistance genes ( Table 1). This hypothesis is supported by the segregation observed for most of the progeny from R ⫻ R crosses (five of six crosses), which fit a 7 S:9 R ratio, indicating the possible involvement of two dominant epistatic resistance genes. Only the segregation of the progeny from cross 5 ⫻ 15 does not conform with the expectations under this hypothesis. Therefore it is hypothesized that at least three genes are involved in determining resistance of the four clones that are resistant to the population of P. coronata under study: one tetrasomic dominant gene for susceptibility and two dominant resistance genes ( Table 2). From these results,

482 The Journal of Heredity 2000:91(6)

Table 2. Hypothesized genotypes for all clones under a three-gene model (two dominant resistant genes, either disomic or tetrasomic in inheritance, and one monogenic dominant gene that conditions susceptibility and is tetrasomic in inheritance) Genetic model Clone

Phenotype

All tetrasomic

5 6 15 28 2 13 22

Resistant Resistant Resistant Resistant Susceptible Susceptible Susceptible

R1r1r1r1 R1r1r1r1 R1r1r1r1 R1r1r1r1 r1r1r1r1 R1R1R1㛮㛮 r1r1r1r1

R2r2r2r2 R2r2r2r2 R2r2r2r2 R2r2r2r2 r2r2r2r2 R2R2R2㛮㛮 r2r2r2r2

it cannot be known unequivocally if these two dominant resistance genes segregate disomically or tetrasomically, because simplex genotypes have segregation ratios equal to diploid heterozygotes. For the interaction P. coronata–oat, most resistance genes are dominant, but resistance is occasionally recessive ( Harder et al. 1984; Simons 1970; Simons et al. 1978). Two of the four S ⫻ R crosses to clone 13 did not fit expected ratios under the proposed model ( Tables 1 and 2), caused by an excess of resistant phenotypes within each cross. It has been found that the degree of dominance varies according to the genetic background on which some crown rust genes function in oat (Simons and Murphy 1954). Thus background or modifier genes may affect the expression of resistance and observed S:R segregation ratios, particularly for this group of clones which was derived from diverse sources. Additive or partially dominant resistance genes, which are present in oat, plus duplex dominant resistance genes in an autotetraploid genome could all cause an excess of resistant phenotypes, yet would be difficult to detect in a multiplegene chi-squared test. Each of these modes of gene action could create the appearance of minor genes with smaller effects than a completely dominant gene, creating an apparent excess of resistant phenotypes. Given the observations for Avena and Lolium, genotype variation for additional resistance genes among these clones is highly probable. Similarly meiotic irregularities, common in smooth bromegrass (Armstrong 1991), may have caused distorted segregation ratios in the progeny of clone 13. Both of these explanations may also apply to the progeny from the 5 ⫻ 15 cross. Verification that additional resistance genes are segregating in these progeny will require additional experiments based on single-urediospore isolates of the pathogen. The results indicate that use of a population of the pathogen, instead of a single

ssss ssss ssss ssss SSS㛮 SSss SSS㛮

Disomic/tetrasomic

Source population

R1r1 R1r1 R1r1 R1r1 r1r1 R1㛮㛮 r1r1

Alpha Alpha WB19e WB88S Alpha WB19e Lincoln

R2r2 R2r2 R2r2 R2r2 r2r2 R2㛮㛮 r2r2

ssss ssss ssss ssss SSS㛮 SSss SSS㛮

urediospore isolate, possibly allowed detection of a larger number of genes determining susceptibility or resistance. The pathogen population may consist of different races and could be viewed as a superrace, having potentially many different avirulence genes that would challenge the resistance genes of the plant. However, possible epistatic relationships among genes were difficult to confirm, because avirulence genes do not act together to confer one phenotype of the pathogen, but a range of heterogeneous virulence phenotypes. For example, a 7 S:9 R ratio was observed in the progeny of most R ⫻ R crosses, but this does not indicate, unequivocally, the existence of complementary epistasis. At a minimum, this ratio only indicates that the plants would be susceptible to the population of the pathogen if they are homozygous recessive for either of the two dominant resistance genes detected. With heterogeneous virulence phenotypes, some plants would be susceptible to some races that comprise the pathogen population, whereas other plants would be susceptible to different races within the same pathogen population. Alternatively the relatively consistent fit of the 7 S:9 R ratio in the R ⫻ R crosses suggests that all four resistant clones have the same diheterozygote genotype. This was surprising given the extremely diverse pedigrees of clones 5, 6, and 15 (Alpha and WB19e, which has Alpha as a parent) versus clone 28 (WB88S, an uncultivated population from southern Siberia). The coincidence of these hypothesized genotypes in clones of diverse origin suggests that these two resistance genes may be complementary. Evidence of complementary gene action in the P. coronata–oat interaction have been reported by different authors (Gregory and Wise 1994; Hayes et al. 1939; Weetman 1942). Segregation of the progeny from selfs of clones 6 and 15, and from cross 5 ⫻ 15, fit to a 3 R:1 S ratio (P ⫽ .23–1.00), further

supporting the hypothesis of genotypic variation for virulence in the pathogen population. Only one dominant gene for resistance appeared to be segregating in these selfs and crosses, instead of the two hypothesized dominant resistance genes ( Table 2). This could occur if most of the inoculum used for these selfed and crossed progeny were collected from plants of PL-BDR1 that were susceptible to races carrying only one of the two corresponding avirulence genes. Because PLBDR1 is itself a heterogeneous population, variability in avirulence genotypes and phenotypes is possible. The P. coronata–oat interaction is controlled by a gene-for-gene system. Existence of resistance genes on both the A and C genomes of oat suggest that the system evolved in diploid wild relatives of oat, prior to polyploidization and speciation. Racial specialization and segregation patterns also suggest coevolution of a gene-for-gene system in the P. coronata– Lolium interaction. Our research on the P. coronata–Bromus interaction has uncovered potentially numerous crown rust resistance genes, three of which could be identified and described. A molecular linkage map will be required to locate these and other resistance genes to either or both the A and B genomes. Nevertheless the likely existence of numerous resistance genes and genotypic variation for pathogen virulence phenotype suggests strong parallels of the inheritance and control of crown rust resistance in the three diverse grass genera Bromus, Avena, and Lolium.

Conclusions The resistance of smooth bromegrass to P. coronata is complex. The results indicate that at least three genes are involved in this host-pathogen interaction, one tetrasomic dominant gene which determines susceptibility (S) and two dominant resistance genes (R1 and R2) that could be inherited tetrasomically or disomically and may be complementary to each other. The susceptibility gene appears to reside in the A genome, which segregates tetrasomically, while the two putative resistance genes may be in either the A or B genomes. Other factors, such as modifier genes or meiotic irregularities, may account for the lack of fit to expected ratios of some progeny. Heterogeneity in the pathogen population may have been responsible for the lack of fit of progeny which occasionally showed monogenic in-

stead of digenic segregation, apparently due to variation in avirulence phenotypes within the pathogen population. However, the use of a population of the pathogen likely allowed us to detect a larger number of genes controlling resistance to P. coronata than would have been possible with a single urediospore isolate. From the Departments of Agronomy ( Delgado and Casler) and Plant Pathology (Grau), University of Wisconsin, Madison, Madison, WI 53706. Address correspondence to M. D. Casler at the address above or e-mail: [email protected]. 䉷 2000 The American Genetic Association

References Armstrong, KC, 1973. Chromosome pairing in hexaploid hybrids from B. erectus (2n ⫽ 28) ⫻ Bromus inermis (2n ⫽ 56). Can J Genet Cytol 15:427–436. Armstrong KC, 1980. The cytology of tetraploid ‘‘Bromus inermis’’ and the Co colchicine-induced octoploid. Can J Bot 58:582–587. Armstrong KC, 1991. Chromosome evolution of Bromus. In: Chromosome engineering in plants: genetics, breeding and evolution. Part B (Gupta PK and Tsuchiya T, eds). Amsterdam: Elsevier.

Martens JW and Dyck PL, 1989. Genetics of resistance to rust in cereals from a Canadian perspective. Can J Plant Pathol 11:78–85. Reheul D and Ghesquiere A, 1996. Breeding perennial ryegrass with better crown rust resistance. Plant Breed 115:465–469. Simons MD, 1970. Crown rust of oats and grasses. Monograph no. 5. St. Paul, MN: American Phytopathological Society. Simons MD and Murphy HC, 1954. Inheritance of resistance to two races of Puccinia coronata Cda. var. avenae Fraser and Led. Proc Iowa Acad Sci 61:170–176. Simons MD, Martens JW, McKenzie RIH, Nishiyama I, Sadanaga K, Sebesta J, and Thomas H, 1978. Oats: a standardized system of nomenclature for genes and chromosomes and catalogue of genes governing characters. USDA agriculture handbook no. 509. Washington, DC: USDA. Weetman LM, 1942. Genetic studies in oats of resistance to two physiologic races of crown rust. Phytopathology 32:19. Welty RE and Barker RE, 1993. Reaction of twenty cultivars of tall fescue to stem rust in controlled and field environments. Crop Sci 33:963–967. Wilkins PW, 1975. Inheritance of resistance to Puccinia coronata Corda and Rhynchosporium orthosporum Caldwell in Italian ryegrass. Euphytica 24:191–196. Wilkins PW, 1978a. Specialisation of crown rust on highly and moderately resistant plants of perennial ryegrass. Ann Appl Biol 88:179–184.

Berg CC, Zeiders KE, and Sherwood RT, 1989. Registration of PL-BDR1 smooth bromegrass germplasm. Crop Sci 29:1578.

Wilkins PW, 1978b. Specialisation of crown rust (Puccinia coronata Corda) on clones of Italian ryegrass (Lolium multiflorum Lam.). Euphytica 27:837–841.

Bush AL and Wise RP, 1996. Crown rust resistance loci on linkage groups 4 and 13 in cultivated oat. J Hered 87:427–432.

Wilson WA and McMullen MS, 1997. Dosage dependent genetic suppression of oat crown rust resistance gene Pc-62. Crop Sci 37:1699–1705.

Casler MD and Carlson IT, 1995. Smooth bromegrass. In: Forages. An Introduction to grassland agriculture, vol. 1, 5th ed ( Barnes RF, et al., eds). Ames, IA: Iowa State University Press; 313–324.

Received July 12, 1999 Accepted July 9, 2000 Corresponding Editor: James L. Hamrick

Delgado NJ, Grau CR, and Casler MD, 2001. Host range and alternate host of a Puccinia coronata population from smooth bromegrass. Plant Dis (in press). Dietz SM and Murphy HC, 1930. Inheritance of resistance to Puccinia coronata avenae, p.f. III. Phytopathology 20:120. Dunn GM and Lea HZ, 1978. Inheritance of a virescent trait in bromus inermis. Can J Genet Cytol 20:499–503. Elliot FC and Love RM, 1948. The significance of meiotic chromosome behavior in breeding smooth bromegrass, Bromus inermis Leyss. J Am Soc Agron 40:335– 341. Ghosh AN and Knowles RP, 1964. Cytogenetic investigations of a chlorophyll mutant in bromegrass, Bromus inermis Leyss. Can J Genet Cytol 6:221–231. Gregory JW and Wise RP, 1994. Linkage of genes conferring specific resistance to oat crown rust in diploid Avena. Genome 37:92–96. Harder DE, McKenzie RIH, and Martens JW, 1984. Inheritance of adult plant resistance to crown rust in an accession of Avena sterilis. Phytopathology 74:352–353. Hayes HK, Moore MB, and Stakman EC, 1939. Studies of inheritance in crosses between Bond, Avena byzantina, and varieties of A. sativa. Minn Agric Exp Stat Tech Bull 137:1–38. Hill HD and Carnahan HL, 1957. Karyology of natural 4x, 6x, and 8x progenies of a tetraploid (4x) clone of Bromus inermis Leyss. Agron J 49:449–452. Hittle CN, 1954. A study of the polycross progeny testing technique as used in the breeding of smooth bromegrass. Agron J 46:521–523 Knowles RP, 1980. Seedling pubescence as a genetic marker in smooth bromegrass (Bromus inermis Leyss.). Can J Plant Sci 60:1163–1170.

Isolation, Mapping, and Characterization of Two Barley Multiovary Mutants J. D. Soule, D. A. Kudrna, and A. Kleinhofs Mutations in homeotic genes disturb the spatial and temporal patterns of development, often leading to the appearance of tissues in abnormal locations. Many homeotic genes, involved in flower development, code for proteins with a highly conserved domain called the MADS box, which acts as a sequence-specific DNA binding protein. Two floral development mutants were isolated from a fast neutron irradiated M2 barley population. The phenotypes are multiovary, that is, stamens replaced with carpels, designated mo7a, and stamens replaced with carpels and lodicules converted to leaflike structures, designated mo6b. These phenotypes resemble the Arabidopsis mutants APETALA3 (AP3) and PISTILATA (PI). The mo6b


and mo7a mutants were mapped to the centromeric region of chromosome 1 (7H) and to the telomeric region of chromosome 3 (3H), respectively. Homeotic genes comprise a family of genes that are involved in developmental pathways of higher organisms. Mutations in these genes disturb the spatial and temporal patterns of development, often leading to the appearance of body parts in abnormal locations. Homeotic mutants have been identified in many plant species, but are best characterized in Arabidopsis thaliana and Antirrhinum majus. Flower development has been intensely studied in dicots, where four distinct differentiation zones or whorls have been described. The first, or outermost, whorl leads to sepal, second to petal, third to stamen, and fourth to carpel development. The model of floral development where specific genes act in two adjacent whorls is referred to as the ABC model ( Bowman et al. 1993; Coen and Meyerowitz 1991; Meyerowitz et al. 1991). In flowering mutants, homeotic genes fall into three separate classes—A, B, and C—each controlling organ identity in two adjacent whorls: A, whorls 1 and 2; B, whorls 2 and 3; and C, whorls 3 and 4. In A. thaliana and A. majus, several mutants altering this developmental pattern have been identified and their genes cloned. These genes encode proteins with a similar structure and contain a highly conserved motif called the MADS box (for homologies found in yeast MCM1, Arabidopsis AGAMOUS, snapdragon DEFICIENS, and mammals SRF) (Schwarz-Sommer et al. 1990). Numerous mutants altering barley flower development are known, but the molecular characterization of the genes is lacking. For example, five mutants, designated multiovary (mo1–mo5), have been previously reported and partially described ( Kamra 1966; Kamra and Nilan 1959; Moh and Nilan 1953; Tazhin 1971). The mo1 and mo3 phenotypes were reported to have a varying number of carpels, no anthers, and normal-appearing lodicules. The mo2 phenotype was one normal appearing stamen and three carpels. The mo4 phenotype was three carpels and three stamens. The mo5 phenotype was four normal-appearing carpels with the lodicules missing and two leaflike structures where the lodicules should be ( Tazhin 1971). Allelism tests among these mutants have not been reported and only the mo5 locus has been mapped, and is located on chromosome 1 (7H) near Amy2 ( Tazhin 1980). These mu-


tants are similar to PI and AP3 mutants in Arabidopsis (Jack et al. 1992) and presumably act on the B region. Here we describe the isolation, characterization, and mapping of two multiovary mutants, mo6b and mo7a, representing two different loci on chromosomes 1 (7H) and 3 (3H), respectively. These mutants will contribute to characterization of the floral development in barley and other monocots.

Methods Mutant Development and Selection Barley (Hordeum vulgare cv. Steptoe) seeds were exposed to 4.0 Gy of fast neutrons at the International Energy Laboratory in Vienna, Austria. The M1 and M2 were grown in the field and multiovary mutants were selected at Pullman, WA, during the summers of 1992 (mo7a) and 1995 (mo6b). An apparently sterile plant having multiple carpel structures within the flower was identified, crossed with Steptoe pollen, and three seeds harvested. This mutant was designated mo7a. The F1 were grown in the greenhouse and F2 seeds collected. The F2 seeds were planted in the field at Pullman, WA, and segregation was scored. A second multiovary mutant, designated mo6b, was selected from mutant head rows grown in the field at Pullman, WA. Wild-type plants from a row containing mutants were harvested to isolate a heterozygote and recover the sterile mutation. Allelism Testing To determine allelism, heterozygotes for mo6b and mo7a were intercrossed. The mo7a parents were F1 plants from a mutant mo7a ⫻ Steptoe cross. Using these plants ensured that all mo7a parents were heterozygotes. Heterozygotes for mo6b were wild-type plants from rows segregating for the multiovary trait. Two-thirds of these wild-type plants were expected to carry the mutant allele in the heterozygous state. Crosses were made using 8–10 putative mo6b parents to ensure at least one heterozygote was used. The plants used as parents were selfed and the F3 seeds were grown to determine the F2 genotype. One hundred fourteen F1 mo6b ⫻ mo7a plants were grown in the field, phenotyped, and F2 seed collected. The F2 seeds were grown in the greenhouse to verify segregation of the mo7a and mo6b phenotypes. The mo6b and mo7a mutants were also tested for allelism to mutants 5102 and 5103 as described above. Seed of the mu-

tants 5102 and 5103 were obtained from the Institut fur Pflanzengenetik und Kulturpflanzenforschung, Gatersleben, Germany, via Dr. Kunzel. The phenotype of these mutants is four normal-appearing carpels, no stamens, and the lodicules appear to be converted to leaflike structures. Thus they resemble the phenotype of mutant mo6b. Genetic Mapping Heterozygous mo6b and mo7a plants were crossed to Morex to map the genes with respect to molecular markers using the bulk segregant analysis method (Michelmore et al. 1991). Crosses were made using eight individual wild-type plants from a segregating mutant family to ensure that a heterozygote was used at least for one cross for each mutant. The mo6b ⫻ Morex and mo7a ⫻ Morex F1 plants were allowed to self-pollinate and the resulting F2 seeds were grown in the field. Leaf tissue samples were collected from all plants approximately 1 month after germination and the plants were phenotyped at flowering for the multiovary trait. DNA samples from 15 wild-type and 15 mutant plants were bulked. The bulk DNA samples were digested with restriction enzymes EcoRI, HindIII, and XbaI (Gibco BRL) and Southern blotted on positively charged nylon membranes (Genescreen Plus, NEN Life Sciences) as described in Kleinhofs et al. (1993). Informative RFLP probes covering the entire genome in approximately 20 cM increments were hybridized to identify genetic linkage for each mutant. Once the approximate gene locations were determined, individual mutant F2 plants from the mo6b (35 plants) and mo7a (29 plants) by Morex crosses were used to create a detailed RFLP map. DNA was isolated from the individual mutant plants and digested with DraI, EcoRI, EcoRV, HindIII, and XbaI, and Southern blotted. Linked probes, identified in the bulk segregant analysis, were used for mapping. Since only the homozygous mo mutants were analyzed, the recombination frequencies were calculated within the recessive class of a 1:3 segregating F2 population according to the formula: p ⫽ (h ⫹ 2b)/2n, where p is the recombination frequency, n is the number of homozygous recessive individuals in F2, h is the number of heterozygous recombinant individuals in F2, and b is the number of homozygous recombinant individuals in F2 (Allard 1956) as described by Hinze et al. (1991).

Figure 1. Wild-type and mutant barley flowers. The lemma and palea have been removed to reveal the internal structures. (A) Wild-type flower consists of one carpel (center), three stamens (top and sides), and two lodicules ( bottom). (B) Mutant mo6b flower has four carpels (top and middle) and two leaflike structures ( bottom). Stamens and lodicules are absent. It is presumed that the three missing stamens give rise to the extra carpels and that the two missing lodicules give rise to the leaflike structures. (C) Mutant mo7a flower has between five and seven abnormal-appearing carpel-like structures (top and sides) and normal-appearing lodicules ( bottom). Stamens are not observed. Color photographs may be viewed at http://barleygenomics.wsu.edu, click on ‘‘developmental mutants.’’

Genomic Plant DNA Isolation, Southern Blotting, and Hybridization Genomic DNA was isolated and Southern blotted as previously described ( Kleinhofs et al. 1993). Barley and rice DNA probes were hybridized at 65⬚C and 62⬚C, respectively, for 16–18 h. The membranes were washed sequentially with 2⫻ SSC and 1% SDS at 65⬚C for 25 min, 1⫻ SSC and 1% SDS at 65⬚C for 25 min, and a final wash using 0.5⫻ SSC and 1% SDS at 65⬚C for 15– 20 min. In some cases, an additional wash with 0.25⫻ SSC and 1% SDS at 65⬚C for 30 min was included. Low copy number informative probes used in this study were identified from previously mapped populations ( Kleinhofs et al. 1993; Graingenes: http://

probe.nal.usda.gov:8300/cgi-bin/browse/ graingenes; http://barleygenomics.wsu.edu). Barley MADS box clones were isolated from a barley genomic DNA lambda library using the rice MADS box clone OsMAD3 (previously Rag6) ( Kang et al. 1998) as a probe. The lambda clone 22 was subcloned to yield the MADS box-specific fragment JS192.5. The MADS box cDNA clone JS001 was isolated by a nested PCR approach from cDNA libraries made from barley flower tissue 0–3 days after pollination. The clones were shown to be MADS box related by sequencing (information about these clones is available at http://barleygenomics.wsu.edu and their characterization will be described in detail elsewhere).

Results and Discussion Phenotype The wild-type barley flower consists of two lodicules, three stamens, and a carpel ( Figure 1A), all enclosed by a lemma and palea. The mo6b mutant phenotype has four carpels, no stamens, and two sepallike structures where the lodicules would be located in the wild-type flower ( Figure 1B). All other visual characteristics of the mutant appear normal except that it is completely sterile. The mutant phenotype segregated 3.2:1 wild-type to mutant in a population of 706 F2 plants with a chisquared of 0.71 (3:1 ratio tested) and a P of .4, suggesting a single, recessive, Mendelian gene. This mutant appears to be


tants 5102 and 5103, respectively. When mo6c and mo6d were crossed with mo7a, the F1 progeny were all wild-type, demonstrating complementation, as expected.

Figure 2. Genetic map of barley chromosome 1 (7H) showing location of mo6b. (1) A barley chromosome 1 skeletal map with the centromere (C) indicated. (2) Genetic linkage map generated using 35 F2 mo6b mutant plants. (3) The Steptoe ⫻ Morex map for the mo6b linkage region ( Kleinhofs et al. 1993). The genetic distances are shown in centiMorgans. The RFLP markers are from the North American Barley Genome Mapping Project except JS192.5, which was generated in this study; RZ242, a rice cDNA from M. Sorrels, Cornell University, NY; and KFP190, a clone obtained from B. Tibbot, University of Wisconsin. Detailed description of all RFLP markers can be found at http:/ /barleygenomics.wsu.edu, click on ‘‘barley molecular marker database.’’

Genetic Mapping Preliminary mapping using the bulked segregant analysis method (Michelmore et al. 1991) located mo6b to the centromeric region of barley chromosome 1 (7H) near Amy2 and mo7a to the chromosome 3 (3H) short-arm telomeric region (data not shown). Mapping of the mo6b trait with 35 mo6b plants from an mo6b ⫻ Morex segregating population located the gene between markers ABG011 and RZ242 and cosegregating with Hsp17 and the MADS box gene probe JS192.5 ( Figure 2). The mo6b locus appears to be in a similar location as the mo5 locus ( Tazhin 1980). The mo5 gene was mapped using only two-point data, thus the exact correspondence with the mo6b locus could not be determined. Twenty-nine mo7a plants from a population of 196 mo7a ⫻ Morex F2 individuals were used to map the mo7a locus, demonstrating that it cosegregated with ABC171A and the MADS box clone JS001B ( Figure 3). JS001B is an RFLP marker from a MADS box containing cDNA clone from barley. We expect that the mo7a mutant represents a MADS box-containing gene. The distorted segregation ratios (discussed above) probably contributed to the differences in map distances compared to the Steptoe ⫻ Morex map ( Figure 3).

similar to mo5 and maps to the same chromosome region. We could not confirm identity due to lack of mo5 seed for allelism testing. The phenotype of the mo7a mutant was variable in the number of carpel-like structures, lacked stamens, and the lodicules appeared to be normal ( Figure 1C). The carpel-like structures appeared abnormal, sometimes terminating in stamenlike structures. Pollination with wild-type pollen produced a few seeds. The phenotype of the mutant plant is distinct from the cv. Steptoe parent. The head is twisted and compacted with shortened awns. Only three mutants were identified in the initial segregating population of 56 plants, a ratio of 18.6:1 wild-type to mutant with a chisquared value of 12.5 (3:1 ratio tested). In the next generation, the segregation ratio was 4.9:1 wild-type to mutant in a population of 1137 progeny with a chi-squared value of 39.9 (3:1) and a very low P. The poor segregation ratios suggest that additional mutations and/or chromosomal rearrangements were present in the original selection and contributed to the poor survival or transmission of the mo7a gametes. This mutant appears to be similar to mo1 and mo3, but the seeds of these mutants were not available for allelism testing.

notype, indicating that the two mutants are not allelic. Two other mutants obtained from Dr. Kunzel, designated 5102 and 5103, were tested for allelism with mo6b. The mutants 5102 and 5103 resemble the mo6b phenotype with four carpels and two leaflike structures. The F1 progeny from the mo6b ⫻ 5102 and mo6b ⫻ 5103 crosses segregated for the multiovary phenotype in both crosses, indicating allelism to mo6b (data not shown). We suggest the gene symbols mo6c and mo6d for the mu-

Allelism Testing The mo6b and mo7a mutants were crossed to determine allelism. All of the 114 F1 plants exhibited a wild-type phe-

Figure 3. Genetic map of barley chromosome 3 (3H) showing location of mo7a. (1) A skeletal map for barley chromosome 3 (3H) with the centromere (C) indicated. (2) Genetic linkage map generated using 29 F2 mutant mo7a plants. (3) The Steptoe ⫻ Morex map for the mo7a linkage region ( Kleinhofs et al. 1993). The genetic distances are shown in centiMorgans. The JS001B marker was generated in this study. See Figure 2 for other RFLP markers.


From the Departments of Genetics and Cell Biology (Soule and Kleinhofs) and Crop and Soil Sciences ( Kudrna and Kleinhofs), Washington State University, Pullman, WA 99165-6420. This is scientific paper 0700-03 from the College of Agriculture and Home Economics Research Center, Washington State University, Pullman, WA, project 0196. This study was supported by USDANRI grant 9600794 and the North American Barley Genome Mapping Project. We thank Dr. G. Kunzel for providing the mutants 5102 and 5103, Dr. H. Brunner for the fast neutron irradiation of seeds, Dr. E. Meyerowitz for Arabidopsis AG, PI, and AP3 clones, Dr. A. Graner for the MWG barley RFLP probes, Dr. G. An for the RAG6 probe, and Dr. P. Hayes for the flowering cDNA libraries. Address correspondence to J. D. Soule at the address above. 䉷 2000 The American Genetic Association

References Allard RW, 1956. Formulas and tables to facilitate the calculation of recombination values in heredity. Hilgardia 24:235–278. Bowman JL, Alvarez J, Weigel D, Meyerowitz EM, and Smyth DR, 1993. Control of flower development in Arabidopsis thaliana by APETALA1 and interacting genes. Development 118:721–743. Coen ES and Meyerowitz EM, 1991. The war of the whorls: genetic interactions controlling flower development. Nature 353:31–37. Hinze K, Thompson RD, Ritter E, Salamini F, and Schulze-Lefert P, 1991. Restriction fragment length polymorphism-mediated targeting of the mlo-o resistance locus in barley (Hordeum vulgare). Proc Natl Acad Sci USA 88:3691–3695. Jack T, Brockman L, and Meyerowitz E, 1992. The homeotic gene APETALA3 of Arabidopsis thaliana encodes a MADS box and is expressed in petals and stamens. Cell 68:683–697. Kamra P, 1966. Genetic control of the development of floral organs in Hordeum vulgare. In: Mechanism of mutation and inducing factors ( Lengerova, A, ed.). Prague: Academia; 213–215. Kamra P and Nilan R, 1959. Multi-ovary in barley. J Hered 50:159–165. Kang HG, Jeon JS, Lee S, and An G, 1998. Identification of class B and class C floral organ identity genes from rice plants. Plant Mol Biol 38:1021–1029. Kleinhofs A, Kilian A, Maroof MAS, Biyashev RM, Hayes PM, Chen FQ, Lapitan N, Fenwich A, Blake TK, Kanazin V, Ananiev E, Dahleen L, Kudrna DA, Bollinger J, Knapp SJ, Liu B, Sorrells M, Huen M, Franckowiak JD, Hoffman D, Skadsen R, and Steffenson BJ, 1993. A molecular, isozyme, and morphological map of barley (Hordeum vulgare) genome. Theor Appl Genet 86:705–712. Meyerowitz EM, Bowman JL, Brockman LL, Drews GN, Jack T, Sieburth LE, and Weigel D, 1991. A genetic and molecular model for flower development in Arabidopsis thaliana. Development 1(suppl):157–167. Michelmore RW, Paran I, and Kessell RV, 1991. Identification of markers linked to disease resistance genes by bulked segregant analysis: a method to detect markers in specific genomic regions using segregating populations. Proc Natl Acad Sci USA 88:9828–9832. Moh C and Nilan R, 1953. Multi-ovary in barley. J Hered 44:183–184. Schwarz-Sommer Z, Huijser P, Nacken W, Saedler H, and Sommer H, 1990. Genetic control of flower development by homeotic genes in Antirrhinum majus. Science 250:931–936. Tazhin OT, 1971. Inheritance of the multi-ovarian characteristic by barley. In: Biological sciences. Alma-Ata: Kazakh State University Press; 69–72 (as cited in Tazhin 1980).

Tazhin OT, 1980. The linkage of the genes mo5 and n in barley. Barley Genet Newslett 10:69–72. Received August 24, 1999 Accepted June 15, 2000 Corresponding editor: Kendall R. Lamkey

The Evergreen Gene is Essential for Flower Initiation in Carnation G. Scovel, T. Altshuler, Z. Liu, and A. Vainstein One of the leading cut-flower crops in the world, the greenhouse carnation (Dianthus caryophyllus), has been subjected to intense breeding efforts for the past few hundred years. As an ornamental crop, flowering and flower architecture are major breeding targets that are constantly in demand. In an ongoing breeding program aimed at improving these characteristics, two mutants heterozygous for a mutation in a gene termed evergreen (e) were identified. In these mutants, spike-like clusters of bracteoles subtend each flower. Genetic analysis of the mutants confirmed the semidominant nature of this nuclear mutation and that the two original mutants were allelic at the evergreen locus. In homozygous mutant plants, a more severe phenotype was observed. Flower formation was completely blocked and spikelike clusters of bracteoles did not subtend any flowers. Morphological characterization of mutant plants revealed that vegetative growth and inflorescence structure are not affected by the mutant allele. In plants heterozygous for the evergreen mutation, fertility, petal and pistil length, calyx diameter, and stamen number were not affected. However, flowers from these heterozygous plants had a reduced number of petals, suggesting an intriguing link between evergreen and the double flower (d) gene that determines petal number in carnation. The control by evergreen of bracteole formation, floral meristem initiation, and petal number in carnation is discussed in comparison to the recessive leafy (lfy) and floricaula (flo) mutants of Arabidopsis and Antirrhinum, respectively. Mutations are an important source of new genetic variation in nature, as well as in breeding populations. The exploitation of mutants often provides insight into genetic mechanisms that were previously not understood. This has especially been the

case in the field of developmental biology, in which most of the genetics of flower formation has been unraveled with the aid of mutants (Coen and Meyerowitz 1991; Irish and Kramer 1998; Weigel and Meyerowitz 1994). Aside from the indisputable importance of this type of research in completing our understanding of plant development, there may also be some side benefits along the way. Some of the developmental mutations appearing in commercially important plants could be interesting as new breeding characteristics. This approach may be particularly relevant to ornamentals, where novelty is a driving force. Furthermore, genes previously identified and characterized via analyses of mutants have already been used to develop new strategies for reducing flowering time (Weigel and Nilsson 1995). Early steps toward flower formation involve the transition from inflorescence to floral meristem. This transition has been found to be regulated by one key regulatory gene—floricaula (flo) in Antirrhinum and leafy (lfy) in Arabidopsis (Coen et al. 1990; Schultz and Haughn 1991; Weigel et al. 1992). FLO and LFY share significant sequence and functional homology. Floral meristem identity mutants homozygous for a recessive allele in flo or lfy develop leafy inflorescent shoots in place of flowers. In Arabidopsis, based on the phenotypic effect of the mutation, a number of lfy mutants have been described. Weak mutant alleles cause a reduction in the number of petals and stamens, and strong mutant alleles cause their absence (Schultz and Haughn 1991; Weigel et al. 1992). Since floral characteristics remain, even in complete loss-of-function lfy mutants, the involvement of redundant pathways for determining flower identity in Arabidopsis was suggested. The APETALA1 and APETALA2 homeotic genes were shown to represent this alternative floral pathway. In double mutants containing strong alleles of both genes (lfy, ap1 or lfy, ap2), there was strong enhancement of the lfy single-mutant phenotype (Weigel et al. 1992). This combination of mutant genes in lfy and ap1 seems to cause complete shutoff of flower formation, although some of the leaf-like structures formed in place of flowers have morphological characteristics resembling carpels ( Irish and Kramer 1998; Weigel et al. 1992). In contrast to Arabidopsis lfy mutants, in homozygous flo mutants of Antirrhinum, the transformation from inflorescence to floral meristem is almost completely blocked and floral organs are not com-


monly produced (Coen et al. 1990). Nevertheless, an AP1-type pathway was also characterized in Antirrhinum. This pathway is manifested in the squamosa mutant ( Huijser et al. 1992; Irish and Kramer 1998). Similar types of mutants have also been reported in petunia and pea ( Hofer et al. 1997; Souer et al. 1998) but have not been reported in carnation. In carnation and other economically established ornamentals, flower architecture is of the utmost importance for breeders; nevertheless very little is known about the mechanism(s) controlling flower development ( Yu et al. 1999). Only one locus of carnation has been identified and genetically characterized to date (Scovel et al. 1998). This single locus, double flower (d), was shown to determine flower type: the recessive combination of wild-type alleles dd being responsible for the single-flower phenotype, while Dd and DD plants generate semi-double and double flowers, respectively. In this study we present a phenotypic and genetic analysis of another flower mutant—evergreen (e)—which was isolated from a breeding population of greenhouse carnation (Dianthus caryophyllus, 2n ⫽ 30). In plants homozygous for this mutation, flower formation is completely blocked, while inflorescence development is normal. Instead of the flower, the mutant generates spike-like clusters of bracteoles. Moreover, clusters of bracteoles are also formed in plants heterozygous for the mutation, but heterozygous plants eventually develop flowers. The involvement of evergreen in the control of bracteole formation, the initiation of flowers, and the determination of petal number revealed by the carnation mutant is discussed in comparison to lfy and flo mutants of Arabidopsis and Antirrhinum, respectively.

Materials and Methods As an integral part of an ongoing breeding program, a large F1 population of carnation (D. caryophyllus) consisting of 10,000 offspring from a cross between two breeding lines of the standard type (male b.l. 1684 and female b.l. 1635) was used to screen for flower mutants. The seedlings were germinated in paper plugs and then transferred to the selection greenhouse and planted at a density of 54 plants/m2 ( Holley and Baker 1991). The plants were allowed to flower without removing the apical shoot, a practice known as stopping (Galbally and Galbally 1997), and as a result selection was focused on the crown


flowering stem. Flowering began 2 months later and the population was screened regularly (over a period of 5 months) for the appearance of mutants. Two mutants, 7689 and 96689, were discovered in which spikes of bracteoles subtended each flower. To assess the mode of inheritance of the new trait, self-pollinations of the flowering mutant lines (7689 and 96689), open directed pollination between two flowering mutant lines, and reciprocal crosses between a wild-type breeding line ( b.l. 1326-1) and flowering mutant line 7689 were performed. All crosses were carried out in the fall/winter and seeds were collected and sown in the summer. Two-month-old seedlings were transplanted to the trial greenhouse in September and the data were recorded during the following winter and spring. The appearance of nonflowering segregants was verified by leaving the crown flower stems on the plants for an additional 1.5 months. To characterize the effects of the mutation in the evergreen (e) gene on flower organogenesis, petals, stamens, and bracteoles were counted on random samples (16–30 segregants per phenotypic class) of wild-type and mutant offspring segregating in F1 and F2 families. Likewise, segregating families were used to measure pistil length, petal length, calyx diameter, and pollen viability. For each of these measurements, the sample size ranged from 70 to 96 segregants per phenotypic class. To allow comparison between flowers of mutant and wild-type segregants, only segregants with single or semidouble flower types were analyzed. Pistil length was measured after removing the petals. Petal length and calyx diameter were measured on flowers in the brush stage, when the petals are standing straight and have not yet begun to curl outward. Petal length refers to the length of the petals from the point of the junction between adjacent calyx leaves to the tip of the petals. Calyx diameter was measured on the same flowers at its widest point just below the junctions between calyx leaves. The percentage of viable pollen was determined by squashing mature anthers in a drop of 2% acetocarmine solution [2% (w/v) carmine in 45% (v/v) acetic acid] and counting the number of stained pollen grains out of the total amount of analyzed pollen using a light microscope. A minimum of 1000 pollen grains were scored in each sample.

Results Isolation and Genetic Characterization of the Mutants A carnation family consisting of 10,000 F1 offspring obtained from a cross between

germinally and somatically stable breeding lines ( b.l. 1684 and 1635) was screened for spontaneous mutants with unusual floral morphology. Two plants (7689 and 96689) formed spike-like clusters of bracteoles, at the end of which a flower eventually formed ( Figure 1A,B). Both mutants grew vigorously and could be easily propagated. To determine the genetic basis of the mutant phenotype, a number of crosses were performed ( Table 1). Reciprocal crosses of the mutant and wild-type plants yielded wild-type and mutant progeny in a 1:1 ratio. When these mutant plants were self-pollinated, three phenotypic classes were obtained in their progeny: wild-type, flowering mutants (parental mutant type) and novel non-flowering mutants ( Table 1). The non-flowering mutants (ca. ¼ of the siblings) were essentially identical to the flowering mutants except that they did not develop any flowers ( Figure 1C). The segregation ratio indicated that the flowering mutants are heterozygous for a semidominant mutant allele and the non-flowering mutants are homozygous for the same mutation. The wild-type allele is henceforth termed evergreen (e) and the mutant allele EVERGREEN (E), due to the semi-dominant nature of the mutation. To genetically analyze the two original flowering mutants (7689 and 96689), an F1 population derived from a cross between these two lines was established. The frequency of the nonflowering mutant progeny (approximately 25%) in the F1 population ( Table 1) suggested that the two original mutants were allelic at the e locus. Since the nonflowering mutant phenotypes do not generate flowers and are 100% sterile, we could not perform additional crosses to further verify this conclusion. The EVERGREEN Mutant Morphology In D. caryophyllus, during the transition from the vegetative to the inflorescence phase, the internodes lengthen and two pairs of bracts are formed. During the transition from the inflorescence to the floral phase, there is a dramatic shortening of the internodes and the formation of two pairs of bracteoles, which are collectively designated the epicalyx (Phillips 1968; Williams 1893). In the Ee mutants 7689 and 96689 the transition from the vegetative to the inflorescence phase was normal with respect to internode length and bract formation. However, the transition to the floral phase was disturbed as a result of repeated bracteole formation. Spike-like clusters of bracteoles were formed prior

to flower development, resulting in delayed flower initiation. The homozygous non-flowering EE mutants generated only spikes of bracteoles and never generated flowers ( Figure 1C,E). It is important to note that the mutation is uniformly expressed, resulting in the blockage of flower formation at all positions along the stem. Within the axils of some bracteoles in a spike, a secondary shoot formed ( Figure 1E). Similar secondary branching of spikes was also observed in flowering Ee mutants, suggesting that the spikes are somewhat indeterminate and shoot-like in nature. The number of bracteoles varied in the range of 6 to 22 ( Figure 1B), both in the original vegetatively propagated Ee (96689 or 7689) mutants and in the heterozygous (Ee) and homozygous (EE) segregants derived from them. However, up to 54 bracteoles developing in the apical position were scored ( Figure 1C). Heterozygous Ee plants strongly expressing the mutant phenotype developed flowers in reverse order relative to wild-type carnation plants. The terminal bud of these Ee mutant plants flowered last or not at all. In contrast, the terminal bud of wild-type (ee) plants flowers first and the lateral buds flower later ( Figure 1F). Similar to flowering Ee mutants, inflorescence formation in non-flowering EE mutants was unaltered by the mutation. Internode elongation and bract formation along the inflorescence branches was the same in mutants as in the wild-type plants. Crosses were made between an Ee line ( b.l. 11692) with alternate branching (one branch per node) and an ee line ( b.l. 2217) characterized by dichotomic inflorescence branching (two branches, at 180 degrees from one another, are formed at each node). Selfing of dichotomic F1 flowering Ee mutants yielded dichotomic nonflowering EE mutant segregants ( Figure 1D) indicating that control of inflorescence branching pattern is independent of that of flower formation by the e locus. In wild-type carnations, locus d, which is not linked to the e locus, controls the number of petals (Scovel et al. 1998). The difference between single (dd), semi-double (Dd), or double flowers (DD) is in the ←

Figure 1. (A) Wild-type (ee) (left) and flowering mutant (Ee) (right) carnation segregants showing the formation of spikes of bracteoles in the mutant prior to flower organogenesis. ( B) Two Ee flowering mutant segregants showing the variation in spike size. (C) A large spike of bracteoles subtended by a pair of bracts without flower organogenesis. ( D) Alternate and dichotomic inflorescence branching in wild-type (ee) and non-flowering mutant (EE) carnations. From left to right: wild type alternate, wild type dichotomic, non-flowering mutant alternate, non-

flowering mutant-dichotomic. ( E) Non-flowering mutant with secondary inflorescences formed in the axils of bracteoles. ( F) The effect of gene E on apical dominance, from left to right: wild type with apical flowering, weak flowering mutant segregant with apical flowering, strong flowering mutant segregant showing flower bud formation at the axial positions and no flower organogenesis at the apical position.


Table 1. Genetic analysis of gene e in directed self and open-pollinated populations Number of Plants

a

b

Cross

Generation

ee ⫹ Ee

EE

Ratio

␹2

P

96689 ⫻ 96689 7689 ⫻ 7689 96689 ⫻ 7689 7689 ⫻ WT WT ⫻ 7689

F1 F2 F1 F1 F1

79 98 155 30,29a 24,28a

26 34 68 0 0

3:1 3:1 3:1 1:1b 1:1b

0.004 0.040 3.590 0.018 0.300

0.90–0.98 0.80–0.90 0.05–0.10 0.90 0.50–0.70

On the left is the number of wild-type (WT, ee) segregants and on the right is the number of flowering mutant (Ee) segregants. The ratio between the number of wild-type and flowering mutant segregants.

amount of petals per flower ( Figure 2). The original mutants 7689 and 96689 are also heterozygous at the d locus and are thus EeDd in genotype. These EeDd mutant plants generated flowers with fewer petals than the semi-double type (eeDd), but more petals than the single-flower type (eedd) ( Figure 2). Analysis of segregating populations revealed that on average, approximately 30% fewer petals developed in flowers of EeDd mutants in comparison to their eeDd siblings ( Table 2). In an Eedd⫻EeDd cross, none of the flowering Ee segregants exhibited petal numbers characteristic of semi-double flowers. Instead, only single flowers and flowers with fewer petals than semi-double flowers were generated by these flowering Ee mutant plants. Moreover, in an EeDd⫻EeDd cross segregating for all three (single, semi-double, double) flower types, there were no flowering Ee mutant segregants with a petal number similar to that of double flowers, the latter developing only in ee wild-type segregants. It is worth noting that the number of petals in single dd flowers was not affected by the E allele and both eedd and Eedd plants developed flowers with five petals. These genetic data indicate that the E mutation reduces the number of petals only in the presence of the D allele, suggesting an interesting genetic interaction between these dominant E and D mutations.

Analyses of other flower parameters in flowering Ee mutants, such as stamen number, pistil and petal length, and calyx diameter, revealed that these are not affected by the presence of the mutant E allele and that the flower structure remains intact ( Table 3). Likewise, there was no detrimental effect of the mutant allele on male ( Table 3) or female (data not shown) fertility.

Discussion The greenhouse carnation (D. caryophyllus) is known to mutate spontaneously at a high frequency ( Holley and Baker 1991; Mehlquist 1941). Many color variants are the result of mutations in commercially successful varieties, and these have been extensively exploited in the case of carnation. In an ongoing breeding program in which we are constantly seeking out novel flower variants in terms of color, shape, size, and type, we isolated two phenotypically similar mutants, 7689 and 96689. Based on the genetic analysis ( Table 1) they appeared to be the result of a mutation at the same locus, which we termed evergreen (e). The phenotype of these mutants is expressed as an increase in the number of bracteoles formed prior to flower formation. This repetitive bracteole formation results in a spike-like structure subtending each flower. Spike formation

Figure 2. The effect of mutant alleles E and D on petal number. From left to right: wild-type single flower (eedd), flowering mutant segregant with intermediate flower type (EeDd), and wild-type semi-double (eeDd) flower.


includes internode shortening and continued bracteole formation beyond the normal two pairs in wild-type carnations. These spikes possess characteristics of both flowers and shoots. They resemble shoots in that they repeatedly generate bracteoles and some of them form secondary shoots in the axils of the bracteoles. However, the lack of internode elongation between bracteoles and the formation of bracteoles instead of leaves/bracts suggests that the spikes are partially floral in nature. Genetic analysis revealed that these two mutants are caused by semidominant mutations in the e locus. The semi-dominant nature was manifested in the appearance of three phenotypical classes in offspring from self-pollinated heterozygous mutants: wild type (ee), flowering (Ee), and non-flowering mutants (EE). The semi-dominant nature of mutation E is unique in comparison to the recessive lfy (Arabidopsis), flo (Antirrhinum), alf (petunia), and uni (pea) mutants (Carpenter and Coen 1990; Hofer et al. 1997; Schultz and Haughn 1991; Souer et al. 1998) and thereby permitted an analysis in a situation where both wild-type and mutant alleles are expressed. An interesting attribute of the heterozygous Ee flowering mutant was that the apical bud position is often the last to flower and frequently does not flower at all ( Figure 1F). In contrast, the apical bud tends to flower first in wild-type plants. Therefore it would seem that the e allele plays a role in determining apical dominance. Moreover, as opposed to flo and lfy, inflorescence development was normal in non-flowering mutants and the inhibition of flower formation was complete regardless of bud position along the inflorescence stem. In the cases of lfy and flo (Carpenter et al. 1995; Coen et al. 1990; Irish and Kramer 1998; Weigel et al. 1992), complete blockage of flower initiation occurs only on the lower part of the inflorescence; the appearance of carpelloid tissues is frequent on the upper part. The absence of an effect of allele E on plant development through inflorescence formation indicates that this gene is involved in the regulation of the initiation/ execution of flower organogenesis. In wildtype (ee) plants, these steps—which take place sequentially or in parallel—include suppression of both bracteole formation and secondary meristems in the axils of bracteoles, and initiation/differentiation of floral organs. In contrast, mutant plants (EE) develop spike-like clusters of bracteoles with occasional secondary shoots

(PhD dissertation). Fort Collins, CO: Colorado State University.

Table 2. Analysis of the effect of mutant allele E on petal number

a

Bracteole number

Petal number

Cross

Generation

eeDd

EeDd

eeDd

EeDd

Eedd⫻EeDd EeDd⫻EeDd

F1 F2

4.0⫾0.1 4.0⫾0.1

11.1⫾1.1 11.2⫾0.8

31.3⫾1.0 30.4⫾2.4

21.3⫾1.3 20.4⫾1.6

a

Petal number (per semi-double (Dd) flower) of wild-type (ee) and flowering mutant (Ee) segregants in each family, was found to be significantly different (p ⬍ 0.0001) using the two-sample t-test assuming equal variances.

Table 3. Analysis of the wild type and flowering mutants with regard to various flower characteristics

Phenotype

Stamen number

Viable pollen Pistil length (%) (mm)

Petal length (mm)

Calyx diameter (mm)

Wild Type Mutant

4.2⫾0.8 3.1⫾0.7

57 61

23.6⫾0.8 25.0⫾0.6

12.4⫾0.3 13.2⫾0.3

41.5⫾0.8 42.7⫾0.6

All the data for both phenotypes were collected from segregants of the F2 cross as in Table 2.

and are unable to develop flowers. The effect of the mutant allele E on petal number, as manifested in flowering Ee plants with reduced petal number ( Figure 2), points to an intriguing link between genes E and D. For example, E may affect the functional level of D. Alternatively, E and D may regulate a common target gene for petal formation. This role of gene e in regulating petal formation is intriguing in analogy to the reduction of petals and stamens in weak lfy mutants (Schultz and Haughn 1991). Note also that direct transcriptional activation by LFY of floral organ identity genes AGAMOUS and APETALA1 was recently demonstrated ( Busch et al. 1999; Wagner et al. 1999). The involvement of gene e in regulating the latter steps in flower development may be particularly important in carnations, which are perpetually flowering and neutral to day length and temperature, as compared to long-day plants such as Arabidopsis and Antirrhinum where induction to flowering, regulated by LFY/FLO, is probably a key step. Future studies targeted at cloning gene e and characterizing its expression pattern may help elucidate this gene’s precise mode of action and determine whether it is indeed a master switch turning on the floral program in a continuous state of flower induction. Furthermore, characterization of gene e may also lead to a new breeding trait. A non-flowering carnation could be of interest as a decorative green branch to be mixed with flowering bouquets, since today the cutfoliage industry has become an important part of the market for ornamentals. From the Kennedy-Leigh Centre for Horticultural Research and the Otto Warburg Center for Biotechnology in Agriculture, The Hebrew University of Jerusalem, Rehovot 76-100, Israel (Scovel and Vainstein), R. Shemi

Ltd., Shedema, Israel (Altshuler), and Department of Cell Biology and Molecular Genetics, University of Maryland, College Park, Maryland ( Liu). Address correspondence to Alexander Vainstein at the address above or e-mail: [email protected]. We thank Dr. Hilary Voet for her help with the statistical analysis and R. Shemi Ltd. for their support and the use of their facilities. This research was also supported by the Israeli Ministry of Agriculture, the Israeli Ministry of Science, and the Association of Israeli Flower Growers. 䉷 2000 The American Genetic Association

Schultz EA and Haughn GW, 1991. Leafy, a homeotic gene that regulates inflorescence development in Arabidopsis. Plant Cell 3:771–781. Scovel G, Ben-Meir H, Ovadis M, Itzhaki H, and Vainstein A, 1998. RAPD and RFLP markers tightly linked to the locus controlling carnation (Dianthus caryophyllus) flower type. Theor Appl Genet 96:117–122. Souer E, van de Krol A, Kloos D, Spelt C, Bliek M, Mol J, and Koes R, 1998. Genetic control of branching pattern and floral identity during Petunia inflorescence development. Development 125:733–742. Wagner D, Sablowski RWM, and Meyerowitz EM, 1999. Transcriptional activation of APETALA1 by LEAFY. Science 285:582–584. Weigel D, Alvarez J, Smyth DR, Yanofsky MF, and Meyerowitz EM, 1992. LEAFY controls floral meristem identity in Arabidopsis. Cell 69:843–859. Weigel D and Meyerowitz EM, 1994. The ABC’s of floral homeotic genes. Cell 78:203–209. Weigel D and Nilsson O, 1995. A developmental switch sufficient for flower initiation in diverse plants. Nature 377:495–500. Williams FN, 1893. A monograph of the genus Dianthus. J Linn Soc 29:346–478. Yu D, Kotilainen M, Pollanen E, Mehto M, Elomaa P, Helariutta Y, Albert VA, and Teeri TH, 1999. Organ identity genes and modified patterns of flower development in Gerbera hybrida (Asteraceae). Plant J 17:51–62. Received February 9, 2000 Accepted May 16, 2000 Corresponding Editor: Brandon Gaut

References Busch MA, Bomblies K, and Weigel D, 1999. Activation of a floral homeotic gene in Arabidopsis. Science 285: 585–587. Carpenter R and Coen ES, 1990. Floral homeotic mutations produced by transposon mutagenesis in Antirrhinum majus. Genes Dev 4:1483–1493. Carpenter R, Copsey L, Vincent C, Doyle S, Magrath R, and Coen E, 1995. Control of flower development and phyllotaxy by meristem identity genes in Antirrhinum. Plant Cell 7:2001–2011. Coen ES, Romero JM, Doyle S, Elliot R, Murphy G, and Carpenter R, 1990. FLORICAULA: a homeotic gene required for flower development in Antirrhinum majus. Cell 63:1311–1322. Coen ES and Meyerowitz EM, 1991. The war of the whorls: genetic interactions controlling flower development. Nature 353:31–37. Galbally J and Galbally E, 1997. Carnations and pinks for garden and greenhouse. Portland, OR: Timber Press. Hofer J, Turner L, Hellens R, Ambrose M, Matthews P, Micheal A, and Ellis N, 1997. UNIFOLIATA regulates leaf and flower morphogenesis in pea. Curr Biol 7:581–587. Holley WD and Baker R, 1991. Carnation production 2. Dubuque, IA: Kendall/Hunt. Huijser P, Klein J, Lonnig WE, Meijer H, Saedler H, and Sommer H, 1992. Bracteomania, an inflorescence anomaly, is caused by the loss of function of the MADS-box gene squamosa in Antirrhinum majus. EMBO 11:1239– 1249. Irish VF and Kramer EM, 1998. Genetic and molecular analysis of angiosperm flower development. Adv Bot Res 28:197–230. Mehlquist GAL, 1941. Inheritance in the carnation, Dianthus caryophyllus: inheritance of nine abnormal types. Proc Am Soc Hort Sci 38:699–704. Phillips D, 1968. Floral morphogenesis of carnation

Floral-Color Polymorphism in Ipomoea purpurea: Biased Inheritance of the Dark Allele is not a General Explanation for its Maintenance S. Paulsen and M. D. Rausher A previous investigation reported the existence in a single population of the morning glory (Ipomoea purpurea) of non-Mendelian inheritance at the W locus influencing flower color. In addition, it was shown that the magnitude of biased inheritance in that population was sufficient to maintain a floral-color polymorphism at that locus at frequencies approximating those observed in natural populations. The current investigation was undertaken to determine whether this biased inheritance was characteristic of other I. purpurea populations, and thus whether it provides a general explanation for maintenance of the polymorphism. The current study found no evidence for biased inheritance in two additional polymorphic populations examined. Non-Mendelian inheritance thus seems unlikely to constitute a general explanation for the maintenance of this floral-color polymorphism in I. purpurea.


Understanding the maintenance of genetic variation remains one of the challenges of evolutionary biology. Although theoretical investigations have identified several mechanisms that can maintain a polymorphism at a single locus [for a summary see Hartl (1980)], there are few unequivocal demonstrations that such mechanisms operate in nature ( Endler 1986; Hartl 1980), nor is it clear what mechanisms most commonly operate to maintain variation in nature. Flower color variation displayed by natural populations of the tall morning glory (Ipomoea purpurea) has been intensively studied to determine the processes responsible for its maintenance. Both the intensity and the hue of pigmentation are highly variable and appear to be under the control of four major, unlinked loci ( Ennos 1981; Ennos and Clegg 1983; Epperson and Clegg 1988) that exhibit variation in natural populations. The W locus, which determines whether the corolla is white (ww), lightly pigmented (Ww), or darkly pigmented (WW), has been the major focus of investigations attempting to elucidate mechanisms responsible for maintaining this variation. Typically, the white allele occurs at low frequencies, ranging from zero to approximately 0.4, with an average of 0.1 ( Epperson and Clegg 1986). The W locus polymorphism appears to be protected, as judged by perturbation experiments. Subramaniam and Rausher (2000) established populations with either low or atypically high frequencies of the w allele. After one generation gene frequencies converged toward the more typical, intermediate frequencies found in natural populations. Moreover, a series of investigations has revealed the operation of a mechanism that can account for protection of the white allele. In particular, white-flowered plants, when rare, appear to enjoy a transmission advantage that arises because of increased selfing. When white flowers are in the minority, bumblebees fail to visit them as frequently as they do lightly or darkly pigmented flowers ( Brown and Clegg 1984; Epperson and Clegg 1987; Fry and Rausher 1997; Rausher et al. 1993). This undervisitation appears to cause increased selfing without any detectable pollen discounting (Rausher et al. 1993). In addition, inbreeding depression is minimal in this species (Chang and Rausher 2000). Because the additional transmission pathway through self-pollination is not compensated for by inbreeding depression or pollen discounting, white-flowered plants are expected to en-


joy a greater overall success in passing copies of their genes to offspring compared to the pigmented genotypes. The frequency-dependence of this transmission advantage—visitation and selfing rates are equal for the three genotypes when whites are common ( Epperson and Clegg 1987; Rausher et al. 1993)—implies that any additional selective advantage of appropriate magnitude associated with the dark allele can balance this transmission advantage and yield a stable polymorphism ( Fry and Rausher 1997). The nature of such a dark-allele advantage has been more difficult to identify. For example, in an experimental population in which the frequency of the dark allele was lower than normal (0.5), Rausher and Fry (1993) failed to find an advantage for the dark allele in either viability or seed production. Although there was some evidence for overdominance in seed size, subsequent experiments failed to reveal offspring fitness effects substantial enough to account for protection of the dark allele (Mojonnier and Rausher 1997). Analysis of reproductive success in the same experimental population also failed to find any increased success of light or dark plants over whites as pollen donors ( Fry and Rausher 1997). One potential mechanism for protecting the dark allele has been identified. Fry and Rausher (1997) found evidence for nonMendelian inheritance favoring the dark allele in pollen produced by heterozygous plants in their experimental population. In addition, controlled crosses indicated variation among pollen parents in the proportions of pollen carrying the light and dark alleles. Finally, using a simple genetic model with parameters estimated from experimental populations, Fry and Rausher showed that jointly the two factors—biased inheritance for the dark allele from heterozygous pollen parents and frequency-dependent transmission bias favoring whites when rare—can produce a stable polymorphism at the W locus. When 60% of the pollen produced by heterozygous plants carries the W allele, the model produced equilibrium frequencies approximating those seen in natural populations. The objective of the investigation reported here was to determine whether this explanation holds generally for most populations of I. purpurea, or is specific to just some, such as the one examined by Fry and Rausher (1997). Specifically we used controlled crosses involving a large number of replicated genotypes from two different populations to determine wheth-

er non-Mendelian inheritance of paternal alleles at the W locus is characteristic of other natural populations of this species, and thus provides a general explanation for why the dark allele is maintained in polymorphic populations.

Materials and Methods Biased inheritance acting through the pollen parent can be detected readily as a deviation from expected Mendelian ratios in the offspring produced by crossing a heterozygous pollen parent to a homozygous seed parent. We adopted this approach by performing experimental crosses using plants of each genotype collected in the field as newly germinated seedlings. Each heterozygous plant chosen for testing as a pollen parent was crossed to three seedparent plants from the same population, one each of the WW, Ww, and ww genotypes. Different seed parents were used for each pollen parent. The offspring from the Ww ⫻ ww crosses, which are either white or light in color, were scored first in both populations sampled because determination of their offspring genotype is unambiguous. The crosses to Ww and WW seed parents, whose offspring are more difficult to score, were included for one of the two populations sampled to determine whether any biased inheritance that existed depended on the maternal genotype at this locus. We collected newly germinated seedlings from two populations in Durham County, North Carolina. Approximately 1000 plants were sampled from the Duke University Field Station (‘‘Field Station’’ population), and another 1000 from an agricultural field in northern Durham County (the ‘‘Riverlea’’ population). Seedlings were raised to first flowering to identify their genotypes. The ww genotype was rare, and its number limited the number of crosses performed. Thus, in total, 27 heterozygotes were tested as pollen parents from the Field Station population and 30 from the Riverlea population. We performed all crosses in the greenhouse. The flowers on plants serving as seed parents were emasculated as buds in the afternoon prior to their opening the next morning. We then hand-pollinated each seed parent by rubbing the stigma with a dehisced anther collected from its paired heterozygous pollen parent. We strove to perform enough pollinations to produce 120 seeds from each cross; however, a number of plants did not survive long enough or bloom sufficiently to attain

this number. Each seed produced was scarified to promote germination and the seedlings were raised in flats in the greenhouse to minimize differential survival as a source of bias. We scored flower color when the plants first bloomed. While the progeny from white seed parents were easily distinguished as either white or light, the light and dark offspring produced from the crosses to light and dark seed parents can be difficult to distinguish. Consequently we assigned light and dark phenotypes to progeny only after we had scored each plant’s flower at least twice on separate days, and the scorings agreed. In addition, the identity of the capsule of each seed was recorded so that if bias was observed it could be linked to differences in embryo survival among capsules. To test for biased inheritance we applied an exact binomial test (Sokal and Rohlf 1969) to the progeny of Ww ⫻ ww and Ww ⫻ WW crosses. For the cross to a heterozygous seed parent we tested the progeny using a Chi-square goodness-of-fit test (Sokal and Rohlf 1969) to the Mendelian expectations. In addition, a G test (Sokal and Rohlf 1969) was applied to each population to test for heterogeneity among pollen parents in the probability of transmitting the dark allele.

Results We did not detect biased inheritance in either population sampled. For the Field Station population, the progeny of 27 heterozygous pollen parents crossed only to white seed parents were scored. Offspring germination and survival to first flowering was high, 99.34% (3019 of 3039 seeds planted germinated), eliminating the possibility that differential survival obscured the true inheritance rates of each allele. The proportion of offspring inheriting the dark allele exhibits a unimodal distribution centered on 0.5 ( Figure 1), as would be expected if there were no average deviation from Mendelian inheritance. Individually, only two sets of progeny differed significantly from Mendelian expectation using the exact binomial test. One cross produced a ratio of 49 whites:71 lights (P ⫽ .055) while the other produced a 43:70 ratio (P ⫽ .014). Given the number of tests performed, two significant outcomes would be expected by chance with reasonable probability. Because excess, unscored seeds were available for these two pollen parents, we repeated the test for biased inheritance. For the first pollen par-

Figure 1. Frequency distribution of proportion of offspring from a particular cross that inherited the dark allele from the Ww pollen parent. All plants crossed were collected from the Field Station population. A total of 27 crosses between Ww pollen parent and ww seed parent were conducted. The number of offspring tested per cross ranged from 89 to 123, with a median of 113.

ent, extra seeds from the cross to the white seed parent were planted and scored. These seeds produced a ratio of 50 whites:48 lights, suggesting that the earlier ‘‘significant’’ result was spurious. For the second pollen parent with apparent significant inheritance bias, there were no additional seeds from the cross with the white seed parent to score. Instead we planted the seeds produced from this pollen parent crossed to the Ww seed parent. A ratio of 37 ww:69 Ww:27 WW was produced, which does not differ significantly from Mendelian expectations (P ⫽ 0.43). It thus appears unlikely that any of the 27 crosses truly exhibited biased inheritance. This pattern is confirmed by the absence of detectable heterogeneity among pollen parents in the proportion of transmitted alleles carrying the dark allele (G test: P ⫽ .209), and the absence of a significant deviation from Mendelian expectation for the entire data set, pooled over pollen parents (1555 ww:1464 Ww; P ⫽ .101, twotailed test). Overall offspring survival for the Riverlea population was 98.02% (3576 of 3648 seeds planted), again virtually precluding the possibility of differential survival among genotypes influencing apparent inheritance rates. For this population we tested 30 Ww plants as pollen parents crossed to white-flowered seed parents. The number of offspring tested per cross ranged from 10 to 120, with a median of 50. Once again, the overall distribution of inheritance frequencies for crosses to ww seed parents exhibited little evidence of

deviation from equal inheritance of dark and white alleles ( Figure 2). Moreover, none of the 30 crosses to white seed parents individually exhibited a significant deviation from the Mendelian expectations (P ⬎ .05 in all cases). There was no detectable heterogeneity in the transmission ratio among crosses (G test: P ⫽ .57), and pooling over the crosses did not reveal any significant deviation from the expected 1:1 ratio of whites:lights (811:783; P ⫽ .499). For 13 of the Riverlea pollen parents tested above we also scored the progeny from their crosses to heterozygous light (Ww) and homozygous dark (WW) seed parents, and again failed to detect biased inheritance. The number of offspring tested per cross ranged from 18 to 127, with a median of 71 for the crosses to light seed parents and ranged from 40 to 122 with a median of 75 for the crosses to the dark seed parent. Among all these crosses, only one produced progeny that differed significantly from Mendelian expectation: a cross to a heterozygous seed parent produced progeny in a 21 ww:66 Ww:40 WW ratio (P ⫽ .05). As before, one anomalous significant result would be expected by chance. The same pollen parent, when crossed to a white seed parent, produced a ratio of 54 ww:43 Ww, and when crossed with a dark seed parent produced a ratio of 50 Ww:32 WW, neither result being significant. ( The latter result produced P ⫽ .06, but the bias is in the direction opposite that anticipated.) It thus seems that the nominally significant ratio when crossed with a heterozygous seed parent is probably an artifact of chance. This inference is confirmed by the absence of heterogeneity in inheritance ratios among crosses (P ⫽ .65 and P ⫽ .13 for light and dark seed parents, respectively), and the failure of the pooled samples to deviate significantly from Mendelian expectations (225 ww:481 Ww:246 WW, P ⫽ .597 and 523 Ww:507 WW, P ⫽ .640).

Discussion The principle implication of our results is that non-Mendelian inheritance of paternal alleles at the W locus in seeds sired by heterozygote pollen parents is not a general explanation for the widespread maintenance of the W-locus polymorphism in I. purpurea. While this explanation may apply to the population studied by Fry and Rausher (1997), it clearly does not apply to the two populations examined in this study. Moreover, our failure to detect bias


Figure 2. Frequency distributions, by seed parent genotype, of proportion of offspring that inherited the dark allele in crosses from Ww pollen parents. All plants crossed were collected from the Riverlea population. See text for variation in number of offspring tested per cross. (A) A total of 30 Ww pollen parents each crossed to a different ww seed parent. (B) A total of 13 of the Ww pollen parents used in (A) were also each crossed to 13 Ww seed parents. The frequency of the dark allele in the offspring is reported rather than the proportion of offspring inheriting the dark allele from the pollen parent. (C) The same 13 Ww pollen parents used in ( B) were each crossed to a different WW seed parent.


is not due to lack of statistical power: if anything, the trend in both populations was for the white allele to be inherited more frequently than the dark allele. The difference between our results and those of Fry and Rausher (1997) may be explained in two reasonable ways: (1) populations differ in whether they contain alleles modifying inheritance bias, either at ‘‘restorer’’ loci or at the W locus itself (the ‘‘modifier-variation’’ hypothesis); and (2) the biased inheritance detected by Rausher and Fry was artifactual (‘‘artifact’’ hypothesis). Although the modifier-variation hypothesis is made plausible by the apparent existence in other plant species of variation among populations in the presence or absence of restorer genes for characters such as male sterility and sex ratio ( Boutin-Stadler et al. 1990; Koelewijn and Van Damme 1995; Taylor 1994), we at this point have no direct evidence that it is applicable to the W locus in I. purpurea, or even that restorer or modifier alleles exist. It seems to us unlikely that Fry and Rausher’s results were completely artifactual because replicate crosses from their population exhibited significant heterogeneity in the proportion of dark alleles inherited from the heterozygous pollen parent and because two of six crosses exhibited significant deviations from the expected 1:1 ratio of whites to lights. There thus appear to be real deviations from Mendelian inheritance in that population. However, the evidence that there is a net bias favoring the dark allele in that population is less compelling. The crossing experiments reported suggest no net bias: of the two crosses exhibiting significant bias, one favored the dark allele while the other favored the white allele. This pattern suggests that there may be segregating in this population a locus linked to the W locus that exhibits biased inheritance. If true, the small number of replicate crosses used in the field experiment (the 2155 experimental plants were derived from just 28 wild-collected progenitor plants) means that it is quite possible that the dark allele was accidentally associated with the biasing allele at the linked locus in the experimental plants, producing artifactual evidence of nonMendelian inheritance at the W locus. While these considerations suggest to us that the existence of inheritance biases are unlikely, we cannot completely rule out the possibility that they occur in some populations and contribute to active maintenance of the W locus polymorphism, as

envisioned by Fry and Rausher (1997). This possibility raises the question of whether the dark allele can be maintained in populations without bias by gene flow from populations with bias. Such a phenomenon is conceivable, since at intermediate to high frequencies there is no difference in selfing rates among the W locus genotypes, and thus no selection favoring the white allele ( Brown and Clegg 1984; Rausher et al. 1993). If W locus variation were truly neutral at these frequencies, very little gene flow would be required to maintain similar gene frequencies in populations with and without biased inheritance as ( Hartl and Clark 1989). It seems unlikely, however, that in the face of even small amounts of gene flow some populations would be biased while others remained bias free. Thus the results of this study indicate that either (1) all populations, including that studied by Fry and Rausher (1997), truly lack biased inheritance, or (2) some populations have biased inheritance and others lack it, implying that gene flow between populations is minimal. An additional possibility is that we may have inadvertently biased our study against finding very strong biased inheritance: in order to test heterozygotes collected from natural populations for biased inheritance, we sampled populations known to be polymorphic and disregarded populations which appeared to be fixed for the dark allele. Populations of the latter type may be expected to harbor any alleles with the strongest inheritance bias, alleles for which the bias is strong enough to overcome the transmission advantage enjoyed by the white allele because of increased selfing. However, alleles with very strong bias are not relevant to understanding the maintenance of variation in polymorphic populations, the goal of our investigation, because by definition they lead to fixation of the dark allele. In conclusion, our results appear to rule out biased inheritance of alleles from heterozygous pollen parents as a universal mechanism for protecting the dark allele in polymorphic populations of I. purpurea. Similarly, other investigations have failed to detect a dark-allele advantage in postgermination viability or in either the male or female components of fitness ( Fry and Rausher 1997; Mojonnier and Rausher 1997; Rausher and Fry 1993). The one stage of the life cycle in which selection on the W locus has not yet been examined is the seed stage, in which seeds lay buried in the soil for at least 7 months. It is conceivable that pleiotropic effects of the

W locus acting during this period may generate a selective advantage that protects the dark allele. This possibility deserves examination. From the Department of Biology, Box 90338, Duke University, Durham, North Carolina 27708-0325. Address correspondence to Susan Paulsen at the address above or e-mail: [email protected]. NSF Grant DEB 9318919 supported this study. 䉷 2000 The American Genetic Association

Taylor DR, 1994. The genetic basis of sex ratio in Silene alba (⫽ S. latifolia). Genetics 136:641–651. Received February 9 2000 Accepted August 30, 2000 Corresponding Editor: Brandon Gaut

Genetic Variability in the Iberian Imperial Eagle (Aquila adalberti) Demonstrated by RAPD Analysis

References Boutin-Stadler V, Saumitou-Laprade P, Valero M, Jean R, and Vernet P, 1990. Spatio-temporal variation of male sterility frequencies in two natural populations of Beta maritima. Heredity 63:395–400. Brown BA and Clegg MT, 1984. Influence of flower color polymorphism on genetic transmission in a natural population of the common morning glory, Ipomoea purpurea. Evolution 38:796–803. Chang S-M and Rausher MD, 2000. The role of inbreeding depression in maintaining the mixed mating system of the common morning glory, Ipomoea purpurea. Evolution 53:1366–1376. Endler JA, 1986. Natural selection in the wild. Princeton, NJ: Princeton University Press. Ennos RA, 1981. Quantitative studies of the mating system in two sympatric species of Ipomoea (Convolvulaceae). Genetica 57:93–98. Ennos RA and Clegg MT, 1983. Flower color variation in the morning glory, Ipomoea purpurea. Am Nat 128: 840–858. Epperson BK and Clegg MT, 1986. Spatial-autocorrelation analysis of flower color polymorphisms within substructured populations of morning glory (Ipomoea purpurea). Am Nat 128:840–858. Epperson BK and Clegg MT, 1987. Frequency-dependent variation for outcrossing rate among flower-color morphs of Ipomoea purpurea. Evolution 41:1302–1311. Epperson BK and Clegg MT, 1988. Genetics of flower color polymorphism in the common morning glory (Ipomoea purpurea). J Hered 79:64–68. Fry JD and Rausher MD, 1997. Selection on a floral color polymorphism in the tall morning glory (Ipomoea purpurea): transmission success of the alleles through pollen. Evolution 51:66–78. Hartl DL, 1980. Principles of population genetics. Sunderland, MA: Sinauer. Hartl DL and Clark AG, 1989. Principles of population genetics. Sunderland, MA: Sinauer. Koelewijn HP and Van Damme JMM, 1995. Genetics of male sterility in gynodioecious Plantago coronopus. Genetics 139:1759–1775. Mojonnier LE and Rausher MD, 1997. Selection on a floral color polymorphism in the common morning glory (Ipomoea purpurea): the effects of overdominance in seed size. Evolution 51:608–614. Rausher MD, Augustine D, and VanderKool A, 1993. Absence of pollen discounting in genotypes of Ipomoea purpurea exhibiting increased selfing. Evolution 47: 1688–1695. Rausher MD and Fry JD, 1993. Effects of a locus affecting floral pigmentation in Ipomoea purpurea on female fitness components. Genetics 134:1237–1247. Sokal RR and Rohlf FJ, 1969. Biometry. San Francisco: W. H. Freeman. Subramaniam B and Rausher MD, 2000. Balancing selection on a floral polymorphism. Evolution 54:691–695.

J. A. Padilla, M. Martıńez-Trancoń, A. Rabasco, J. C. Parejo, M. E. Sansinforiano, and M. I. Guijo RAPD analysis was used to estimate the genetic diversity in an Iberian imperial eagle (Aquila adalberti) population, one of the most threatened bird species in the world. Forty-five of 60 arbitrarily designed primers amplified 614 loci in 25 individual eagles, 59.7% of which were polymorphic. In contrast to the traditional allozyme analysis performed in a previous study, the RAPD method has revealed a high level of heterozygosity in this species (H ⫽ 0.267 ⫾ 0.008). The genetic distances estimated between 25 eagles can serve to establish more adequate mating in order to preserve genetic variability. Conservation efforts being carried out in Spain in this species might be successful based on the results obtained in the present work. The Iberian imperial eagle (Aquila adalberti) is a diurnal species of prey, living almost exclusively in the southwest of the Iberian peninsula (Gonza´lez and Gonza´lez 1991). Despite being one of the most threatened bird species in the world (Collar and Andrew 1988) with a decreasing census of less than 126 pairs, until recently genetic studies had not been carried out on this species. Recently, molecular and cytogenetic studies have shown a clear separation between A. adalberti and eastern imperial eagle (A. heliaca) (Padilla et al. 1999; Seibold 1994; Seibold et al. 1996). A possible consequence of the small number of breeding animals is the loss in the amount of genetic variation present in this species. This loss of variability can result in a significant decrease in fitness (inbreeding depression) and consequently a high risk of extinction ( Bijlsma et al. 1994; Frankel and Soule´ 1981; Hedrick and Miller 1992; Schonewald-Cox et al. 1983; Soule´ 1987). Knowledge of genetic variation is considered to

be of major importance for the adaptability of populations and species to changing and deteriorating environments ( Lynch and Lande 1993) and constitutes an important step in conservation of endangered species ( Tamate et al. 1995). A previous study carried out in order to estimate the polymorphism levels of the Iberian imperial eagle by allozyme electrophoresis, revealed the complete absence of genetic variation in 22 loci tested ( Negro and Hiraldo 1994). However, the allozyme variation represents only a small portion of the total genetic variation ( Barrowclough 1983). DNA molecular techniques have been shown to be more efficient in demonstrating genetic relationships and genetic diversity, since they detect the variability of DNA in coding and noncoding regions. One means of investigating DNA molecular variation is the random amplified polymorphic DNA (RAPD) method (Welsh and McClelland 1990; Williams et al. 1990, 1993). RAPD polymorphisms are generated by applying polymerase chain reactions (PCRs) to genomic DNA samples, using random short oligonucleotides as primers, which do not discriminate between coding and noncoding regions. Therefore it is reasonable to expect that the technique samples the genome more randomly than other methods based on the coding regions variability ( Lynch and Milligan 1994). The RAPD technique has been extensively used to analyze genetic variation in populations of plants and invertebrate animals. The use of RAPD methodology for genetic characterization of vertebrate populations is growing rapidly. In birds, RAPD has been used for several purposes such as identification of geographic populations of migratory species ( Haig et al. 1997), the study of genetic variation within and among populations ( Nusser et al. 1996; Horn et al. 1996; Sharma et al. 1998), sex determination ( Lessells and Mateman 1998), and to distinguish segregation of alleles in crosses between inbred lines (Wei et al. 1997). RAPD analysis is especially useful for conservation biology because it requires minimum sample material, no prior sequence knowledge, no radioactive probes, and enhances the feasibility of molecular studies when access to specimens is restricted ( Kozol et al. 1994). Besides, this technique is particularly valuable for revealing variation in species with low genetic variability when other techniques fail to reveal differences among individuals ( Bowditch et al. 1994; Dawson et al. 1993). Both facts


determined that the RAPD technique was employed to quantify the genetic variability within the Iberian imperial eagle population.

Materials and Methods Samples One milliliter of whole blood was taken from each of the 25 Iberian imperial eagle specimens (12 females, 11 males, and 2 of unknown sex). Seventeen of the samples were collected from animals kept in captivity for breeding purposes and eight were from different nests in the wild (six juvenile and two adult birds). The blood was drawn from the brachial vein and transferred to a tube containing EDTA as an anticoagulant and submitted to our laboratory by members of the Environmental Agency of the Extremadura Regional Government (Spain). DNA Extraction Genomic DNA was extracted from 75 ␮l of blood as described by Sambrook et al. (1989). DNA concentrations were measured with GeneQuant spectrophotometer (Pharmacia, Uppsala, Sweden).

Table 1. Summary of data obtained by RAPD analysis for 45 primers with the 25 Eagles

Name

Primer sequence (5⬘-3⬘)

GC content (%)

OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA OPA

GCGCACGG GCCGTCCGAG CAGCCTCGGC GGGACGTCTC CCGCGCCGGT GTCCGAGGCC CGAGGCCGTC TCGGCGAGCC GCCATCGGGC AAACGGGCGG AGGGCTCGGC GTCCACGCCG GACCTCGCCG GTGCCTGCCG TTTGCCCGCC TCCCGAGCCG CAGGTGCGGC CTGGAGCGGC CACGGACGGC CGCCATCCGC TGGGCACGGC CGGCTCGGGT GCCGCTCCTG GTGCCAGCCG CGTGCGTGGC GCGGTAGGCG ACCCGTGCCG CGGCGAGGAC CACGGTCGGC GCACGCACCG CTGGTGCGGC CAGGTGCGGCGG CAGGTGCGGCAA AAAGCCCGCG GGGATTGTCA GGCATCATAC CCTAGCTCAC GCTTAATCCG CTGTGGACGG ACGTCGAGCA AACCGCGGTCT CAAGGGAGGT AGCAGCGTGG AATCGGGTCG CAGCTATGACCAG

87.5 80 80 70 90 80 80 80 80 70 80 80 80 80 70 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 80 83.3 66.6 70 50 50 60 50 70 60 63.7 60 70 60 53.8

001 002 003 004 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 030 031 032 034 035 037 040 044 045 046 048 049 052 053 056 057 065 068

RAPD-PCR Amplification RAPD reactions were performed following the method of Williams et al. (1990) in a total volume of 50 ␮l containing 100 ng of genomic DNA, 2 ␮M primer, 200 ␮M (each) dNTPs (Pharmacia), 2.5 mM MgCl2, and 1 U of DyNAcyme II DNA polymerase ( Finnzymes Oy). Sixty different 8- to 13-mer primers (Amersham-Pharmacia) were used. Each amplification was performed on a DNA thermocycler (Perkin-Elmer) programmed with an initial denaturation step at 94⬚C for 3 min followed by 40 cycles each at 94⬚C for 1 min, 40⬚C for 1 min 45 s, and 72⬚C for 2 min 30 s, and a final extension step at 72⬚C for 10 min. A negative control reaction was prepared for each primer employed in all PCR amplifications to identify contamination of reactions with nontarget template DNA. All amplification reactions were repeated twice in order to test reproducibility.

weight marker (100 bp ladder) included on each gel ( Boehringer Mannheim).

Fragment Visualisation Amplified DNA fragments were resolved by electrophoresis in 2% agarose gel, run at 120 V for 180 min in 0.5⫻ TBE buffer and stained with ethidium bromide. The gels were photographed under ultraviolet ( UV) light. The size of the RAPD bands was estimated with a DNA molecular

Statistical Analysis To determine the genetic variability in the eagle population, we employed the asymptotically unbiased estimator of gene diversity given by Lynch and Milligan (1994). The following assumptions were made in the data analysis: (1) all RAPD loci show complete dominance; (2) all loci have two


Data Analysis The RAPD profiles were analyzed manually. About one-fourth of the primers tested did not produce interpretable data. Only the primers that showed clear amplification patterns were considered. Amplified DNA fragments ( bands) were scored as 1 (fragment present) or 0 (fragment absent) for individual birds.

Average no. scored bands/eagle (range)

Total number of scored bands

Approx. bands size range ( bp)

15.8 (12–17) 10.7 (8–12) 9.6 (7–12) 11.2 (5–14) 18.2 (8–15) 11.4 (5–18) 7.7 (6–10) 12.2 (10–14) 11.3 (7–14) 14.1 (11–16) 9.7 (7–12) 6.9 (6–10) 12.6 (8–18) 11.0 (6–13) 6.6 (5–8) 9.3 (8–10) 12.0 12.9 (8–17) 9.0 (5–11) 7.9 (4–11) 10.0 (6–14) 13.6 (9–17) 9.0 (7–12) 10.6 (7–14) 7.2 (3–10) 9.2 (6–11) 9.1 (4–11) 7.6 (3–10) 11.0 (4–14) 15.8 (13–18) 6.5 (4–8) 7.9 (5–9) 10.4 (8–12) 11.4 (8–13) 7.4 (6–12) 5.0 (4–6) 7.5 (7–9) 7.4 (6–9) 10.5 (6–14) 8.5 (5–11) 7.3 (5–10) 4.5 (3–6) 9.2 (6–14) 8.1 (4–13) 7.9 (6–10)

21 13 16 17 17 19 10 17 17 19 13 11 22 16 8 10 12 19 12 11 17 17 12 15 11 12 11 10 19 20 8 10 12 14 12 6 10 12 14 11 12 7 15 16 11

234–980 370–980 410–1230 280–1200 350–1230 270–850 260–780 500–350 160–980 220–1000 300–700 400–1100 240–1150 290–800 210–650 300–700 250–700 200–1000 375–900 300–1130 280–850 290–800 260–875 250–650 280–820 350–1033 220–1500 340–890 350–1200 200–1000 480–720 300–800 300–900 420–2000 320–900 480–900 330–700 300–1175 290–700 220–850 300–1000 300–750 200–1100 300–2000 250–680

alleles; (3) the marker alleles from different loci do not comigrate to the same position on a gel; and (4) the population is in Hardy–Weinberg equilibrium ( Lynch and Milligan 1994). The percentage of bands absent at a given locus gives the frequency of the recessive homozygote (q2), from which the frequency of the two alleles can be calculated. Hardy–Weinberg equilibrium cannot be determined empirically using dominant markers. Thus the calculated heterozygosity represents the heterozygosity expected given Hardy– Weinberg equilibrium. To reduce the bias associated with analysis of dominant markers, no loci with extremely low recessive allele frequencies (q2 ⬍ 3/n) were included in the population analysis ( Lynch and Milligan 1994). The estimates of genetic relationships among eagles were performed using clus-

ter analysis with the NTSYS-pc package, version 2.00 (Applied Biostatistics, 1997). An UPGMA analysis was carried out using the matrix of genetic distances previously estimated by Nei’s method ( Nei 1972) and a dendrogram representing the relationship between the 25 imperial eagles was obtained.

Results Sixty primers were used to amplify the template DNA of 25 imperial eagles. Table 1 summarizes the banding patterns obtained with the 45 primers (75%) that produced high-intensity amplification products and minimal smearing. A typical scorable gel is shown in Figure 1, where polymorphic as well as monomorphic bands can be seen. Overall 614 RAPD loci were amplified. The total number of fragments per primer ranged from 6 to 22, with an average of 13.6. Only the primer OPA 018 produced complete monomorphic amplification products. The rest of the primers yielded 367 polymorphic bands (59.7%). The size of DNA fragments ranged from 0.16 to 2 kb. Individual RAPD profiles produced from 3 to 18 fragments/ primer with an average of 9.8 fragments. We have found a positive and significant correlation coefficient (0.38; P ⬍ .001) between the G⫹C content in the primers and the total number of amplified fragments. The expected heterozygosity calculated on the assumption of Hardy–Weinberg equilibrium was 0.267 (SE ⫾ 0.008) considering all 614 loci analyzed. This value represents the gene diversity in the population of the 25 analyzed eagles. The genetic distance values computed from observed marker frequencies ( Nei 1972) for the 25 eagles analyzed ranged from 0.076 to 0.308. The UPGMA dendrogram generated is presented in Figure 2.

Figure 1. RAPD banding pattern produced with primer OPA 065 generated from 25 Iberian imperial eagles (Aquila adalberti). Assigned numbers of each eagle (1–25) are indicated at the bottom of the photograph. M: denotes the lanes with molecular weight marker (100 bp ladder).

polymorphism and heterozygosity observed in this study (P ⫽ .597, H ⫽ 0.267, n ⫽ 25) are higher than those observed by RAPD studies in other threatened avian species ( Bowditch et al. 1994; Haig et al. 1994; Nusser et al. 1996). The high correlation found between GC content in the primers and the total number of bands generated from it might also be a factor contributing to the high band polymorphism observed in this species.

It is generally considered that in the past century the Iberian imperial eagle has been represented by a very small population, experiencing a continuous drift since then [see Negro and Hiraldo (1994) for a review]. In recent years, several causes such as shooting, poisoning, trapping, electrocution, and breeding failure caused by agricultural chemicals and disturbance when nesting, have caused the population size to decrease (Collar et al. 1994). A low

Discussion The use of the RAPD technique has revealed the existence of genetic variation in the Iberian imperial eagle population, in contrast to the allozyme analysis of Negro and Hiraldo (1994) which did not detect variation in this species. Although traditionally allozymes have been used to assess population structure, this technique often detects little variability, particularly in birds ( Barrowclough 1983). The RAPD protocol applied to the eagle DNA produced a high number of bands and proved to be a useful method for evaluating polymorphisms in this species. The levels of

Figure 2. UPGMA dendrogram constructed from Nei’s genetic distances among 25 Iberian imperial eagles. Lines are marked with the assigned number of each eagle and the corresponding sex (m: male; f: female; ?: unknown sex). An asterisk indicates samples collected from nests in the wild.


genetic variability is expected for small populations that have suffered a bottleneck [see Cornuet and Luikart (1996) for a review]. Thus the high value of heterozygosity found in this work on the Imperial eagle population is surprising. This fact could be explained by the differential contribution to the estimated heterozygosity value of polymorphic bands amplified from noncoding DNA sequences. It is well known that the RAPD technique detects coding as well as noncoding DNA sequences, and many of the most informative polymorphic sequences are those derived from repetitive (noncoding) DNA sequences in the genome ( Haymer 1994). The term ‘‘repetitive DNA’’ describes a class of DNA sequences found in multiple copies in a genome. This is a class of DNA found to some extent in all eukaryotes. It is not uncommon to find that up to 50% of the DNA in the genome of a species consists of sequences of this type ( Lohe and Roberts 1988). Repetitive DNA sequences are a major structural feature of heterochromatic regions of the genome and it has been generally established that they are rich in GC base pairs (Pardue and Gall 1970). Several chromosome regions of the Imperial eagle’s complement contain this class of GC-rich heterochromatin (Padilla et al. 1999). Moreover, noncoding regions have higher mutation rates than the coding sequences (Jeffreys et al. 1988), which are several orders of magnitude higher than the typical forward mutation rates for genes (Ayala 1976). In this sense, high values of heterozygosity have been reported in endangered species by minisatellite analysis, which detect noncoding DNA sequences exclusively (Stephens et al. 1992). Repetitive DNA sequences constitute substantial portions of most eukaryotic genomes and they represent an enormous, largely untapped reservoir of genetic variation ( Haymer 1994). The topology of the dendrogram represents relative genetic relationships among eagles and may reflect the population structure of this species. The genetic heterogeneity of the population structure observed here provides guidelines for management of the species gene pool in captivity, allowing us to plan mating in order to maintain genetic diversity. Thus monitoring genetic diversity through molecular analysis can facilitate conservation efforts for the Iberian imperial eagle being carried out in Spain, such as improvements in habitats and development of a program of breeding in captivity.


From the Department of Genetic and Animal Breeding, Faculty of Veterinary Studies, University of Extremadura, Spain. This research is part of a project on conservation of the Iberian Imperial Eagle supported by Junta de Extremadura Government. We are especially grateful to D. Javier Caldera and D. Ańgel Sańchez ( Enviromental Agency of the Junta de Extremadura) for supplying the samples used in this work. Address correspondence to J. A. Padilla, Gene´tica y Mejora Animal, Departamento de Zootecnia, Facultad de Veterinaria, Avda. Universidad s/n. 10071, Caćeres, Spain, or e-mail: [email protected].

americanus (Coleoptera: Silphidae). Ann Entomol Soc Am 87:928–935. Lessells CM and Mateman AC, 1998. Sexing birds using random amplified polymorphic DNA (RAPD) markers. Mol Ecol 7:187–195. Lohe A and Roberts P, 1988. Evolution of satellite DNA sequences in Drosophila. In: Heterochromatin ( Verma R, ed). Cambridge: Cambridge University Press; 148– 186.

䉷 2000 The American Genetic Association

Lynch M and Lande R, 1993. Evolution and extinction in response to environmental change. In: Biotic interactions and global change ( Kareiva PM, Kingsolver JG, and Huey RB, eds). Sunderland, MA: Sinauer; 234–250.

References

Lynch M and Milligan BG, 1994. Analysis of population genetic structure with RAPD markers. Mol Ecol 3:91– 99.

Ayala FJ, 1976. Molecular genetics and evolution. In: Molecular evolution (Ayala FJ, ed). Sunderland, MA: Sinauer; 1–20.

Negro JJ and Hiraldo F, 1994. Lack of allozyme variation in the Spanish imperial eagle Aquila adalberti. Ibis 136: 87–90.

Barrowclough CF, 1983. Biochemical studies of microevolutionary processes. In: Perspectives in ornithology ( Brush AH and Clark CG, eds). New York: Cambridge University Press; 223–261.

Nei M, 1972. Genetic distance between populations. Am Nat 106:283–292.

Bijlsma R, Ouborg NJ, and Van Treuren R, 1994. On genetic erosion and population extinction in plants: a case study in Scabiosa columbaria and Salvia pratensis. In: Conservation biology ( Loeschcke V, Tomiuk J, and Jain S, eds). Basel, Switzerland: Birkhauser Verlag; 255– 271. Bowditch BM, Albright DG, Willians JGK, and Braun MJ, 1994. Use of randomly amplified polymorphic DNA markers in comparative genome studies. Meth Enzymol 224:294–309. Collar NJ and Andrew P, 1988. The ICBP world checklist of threatened birds. ICBP technical publication no. 8. Cambridge: ICBP.

Nusser JA, Goto RM, Ledic DB, Fleischer RC, and Miller MM, 1996. RAPD analysis reveals low genetic variability in the endangered light-footed clapper rail. Mol Ecol 5: 463–472. Padilla JA, Martıńez-Trancoń M, Rabasco A, and Fernańdez-Garcıá JL, 1999. The karyotype of the Iberian imperial eagle (Aquila adalberti) analyzed by classical and DNA replication banding. Cytogenet Cell Genet 84: 61–66. Pardue MH and Gall J, 1970. Chromosomal localization of mouse satellite DNA. Science 168:1356–1358. Sambrook J, Fritsch EF, and Maniatis T, 1989. Molecular cloning. A laboratory manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

Collar NJ, Crosby MJ, and Stattersfiled AJ, 1994. Birds to watch 2: the world list of threatened birds. Cambridge: Birdlife International.

Schonewald-Cox CM, Chambers SM, Macbryde B, and Thomas L, 1983. Genetics and conservation: a reference for managing wild animal and plant populations. London: Benjamin-Cumming.

Cornuet JM and Luikart G, 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001–2014.

Seibold I, 1994. Untersuchungen zur molekularen Phylogenie der Greifvo¨gel anhand von DNA-sequenzen des mitochondriellen Cytochrom b-Gens (PhD dissertation). Heidelberg: University of Heidelberg.

Dawson IK, Chalmers KJ, Waugh R, and Powell W, 1993. Detection and analysis of genetic variation in Hordeum spontaneum populations from Israel using RAPD markers. Mol Ecol 2:151–159. Frankel OH and Soule´ ME, 1981. Conservation and evolution. Cambridge: Cambridge University Press. Gonza´lez LM and Gonza´lez JL, 1991. El mundo del aguila imperial Ibe´rica. Madrid: Fondnat. Haig SM, Rhymer JM, and Heckel DG, 1994. Population differentiation in randomly amplified polymorphic DNA of red-cockaded woodpeckers Picoides borealis. Mol Ecol 3:581–595. Haig SM, Gratto-Trevor CL, Mullins TD, and Colwell MA, 1997. Population identification of Western Hemisphere shorebirds throughout the annual cycle. Mol Ecol 6: 413–427. Haymer DS, 1994. Random amplified polymorphic DNAs and microsatellites: What are they, and can they tell us anything we don’t already know? Ann Entomol Soc Am 87:717–722. Hedrick PW and Miller PS, 1992. Conservation genetics: techniques and fundamentals. Ecol Appl 2:30–46. Horn PL, Rafalski A, and Whitehead PJ, 1996. Molecular genetic (RAPD) analysis of breeding Magpie Geese. Auk 113:552–557. Jeffreys AJ, Royle NJ, Wilson V, and Wong Z, 1988. Spontaneous mutation rates to new length alleles at tandemrepetitive hypervariable loci in human DNA. Nature 332:278–281. Kozol AJ, Traniello JFA, and Williams SM, 1994. Genetic variation in the endangered burying beetle Nicrophorus

Seibold I, Helbig AJ, Meyburg BU, Negro JJ, and Wink M, 1996. Genetic differentiation and molecular phylogeny of European Aquila eagles according to cytochrome b nucleotide sequences. In: Eagle studies (Meyburg BU and Chancellor RD, eds). Berlin: WWGBP; 1– 15. Sharma D, Rao KB, Singh HP, and Totey SM, 1998. Randomly amplified polymorphic DNA (RAPD) for evaluating relationships among varieties of guinea fowl. Genet Anal 14:125–128. Soule´ ME, 1987. Viable populations for conservation. Cambridge: Cambridge University Press. Stephens JC, Gilbert DA, Yuhki N, and O’Brien SJ, 1992. Estimation of heterozygosity for single-probe multilocus DNA fingerprints. Mol Biol Evol 9:729–743. Tamate HB, Shibata K, Tsuchiya T, and Ohtaishi N, 1995. Assessment of genetic variation within populations of sika deer in Japan by analysis of randomly amplified polymorphic DNA (RAPD). Zool Sci 12:669–673. Wei R, Dentine MR, and Bitgood JJ, 1997. Random amplified polymorphic DNA markers in crosses between inbred lines of Rhode Island Red and White Leghorn chickens. Anim Genet 28:291–294. Welsh J and McClelland M, 1990. Fingerprinting genomes using PCR with arbitrary primers. Nucleic Acids Res 18:7213–7218. Williams JGK, Kublelik AR, Livak KJ, Rafalski JA, and Tingey SV, 1990. DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Res 18:6531–6535. Williams JGK, Hanafey MK, Rafalski JA, and Tingey SV,

1993. Genetic analysis using random amplified polymorphic DNA markers. Method Enzymol 218:704–740. Received September 28, 1999 Accepted July 9, 2000 Corresponding Editor: Oliver A. Ryder

Genetic Analyses of Plumage Color Mutations on the Z Chromosome of Japanese Quail F. Minvielle, S. Ito, M. InoueMurayama, M. Mizutani, and N. Wakasugi Genetic analyses were performed with four sex-linked plumage color mutations (roux, brown, imperfect albino, and cinnamon) in Japanese quail (Coturnix japonica). Roux and brown quail have similar plumage color, but plumage of roux quail is paler. Pure, F1 and F2 matings were carried out with roux and brown stocks, and 357, 338, and 273 progeny with either roux or brown plumage color were obtained from each mating type, respectively. These allelism tests showed that mutations for roux and brown colors were alleles (*R and *B) from the same locus BR, and that BR*B was dominant over BR*R. Two alleles at the AL locus, AL*A (imperfect albino) and AL*C (cinnamon) were used to estimate the recombination frequency between the BR and AL loci on the Z chromosome. It was estimated to be 38.1 ⫾ 1.0% based on 4615 chicks from the test crosses. Three plumage color loci have been reported on the Z chromosome of the Japanese quail, formerly named Coturnix coturnix japonica (Wakasugi 1984) and now renamed Coturnix japonica (Crawford 1990). The first locus (AL) has two kinds of mutations: imperfect albinism ( Lauber 1964; Sittmann et al. 1966; Somes 1988) or sex-linked white ( Homma et al. 1968; Wakasugi and Kondo 1973), and red-eyed brown (Wakasugi and Kondo 1973) or cinnamon ( Homma and Jinno 1969; Somes 1988; Truax and Johnson 1979). These mutations are caused by the recessive alleles AL*A and AL*C, respectively, and *C is dominant over *A. The second locus has only one mutation, brown ( Homma 1968; Somes 1988; Wakasugi and Kondo 1973), which is caused by a recessive allele. Another sex-linked plumage color mutation roux has been independently discovered in France and is reported to be controlled

by a recessive gene (Somes 1988). Roux quail have similar but paler plumage color than brown ones. The recombination frequency between loci for imperfect albinism and brown and that between loci for imperfect albinism and roux were reported to be 35% (Wakasugi and Kondo 1973) and 30% (Perramon A, personal communication cited by Bitgood and Somes 1990), respectively. The two objectives of this work were to carry out for the first time a joint analysis of the inheritance of the roux and brown mutations, and to obtain a reliable estimation of the linkage relationship between genes for those mutations and AL.

Materials and Methods Quail The roux and imperfect albino quail from the Institut National de la Recherche Agronomique ( INRA), France, the brown quail from the Graduate School of Bioagricultural Sciences, Nagoya University, Japan, and the cinnamon quail from the Nippon Institute for Biological Science, Japan, were used for the allelism test and linkage analyses described below. The roux color mutation appeared spontaneously more than 15 years ago and was introduced in a wild-type laboratory INRA quail colony. A roux line was constituted and maintained since then as a closed population. On the other hand, the brown color mutation was found among newly hatched chicks at a quail hatchery in Toyohashi city, Aichi Prefecture, Japan, more than 30 years ago and maintained at Nagoya University. These two mutants appear to be less common than the imperfect albino and have been described thus far only in France and Japan, respectively. Cross experiments for genetic analyses were performed by both the Institut National de la Recherche Agronomique, Jouy-en-Josas, France, and the Faculty of Agriculture, Gifu University, Japan. Allelism Test Between roux and brown Plumage Under the hypothesis that the brown and roux mutations were caused by two different alleles at the same locus, and because the brown mutation was described first, hypothetical alleles for the brown and roux mutations were named BR*B and BR*R, respectively, according to the gene nomenclature of Crittenden et al. (1996). Pure line ( F0) matings (BR*B/BR*B with BR*B/W, and BR*R/BR*R with BR*R/W) and reciprocal ( F1) crosses (BR*B/BR*B

with BR*R/W, and BR*R/BR*R with BR*B/ W) were first performed with the two lines. The down color of the resulting progeny was visually examined at hatching, and sex was recorded later. Then crossbred males (BR*B/BR*R and BR*R/ BR*B) were mated with pure-line females (BR*B/W and BR*R/W) and the down color of their descendants ( F2) was recorded as before. Frequencies of the two observed phenotypes (roux and brown) were compared to expectations under a one-locus, two-allele, genetic model of inheritance by chi-squared analysis. Linkage Analysis of BR*R and AL*A Wild-type alleles were named BR*N and AL*N (Crittenden et al. 1996). Two test crosses were carried out. Double heterozygous wild-type males (BR*R AL*N/BR*N AL*A) were mated to single recessive hemizygous roux (BR*R AL*N/W) and imperfect albino (BR*N AL*A/W) females. In the course of the cross experiment, BR*R was found to be hypostatic to AL*A. Therefore newly hatched chicks were classified into three types according to their down colors as follows: wild-type (BR*N AL*A/BR*R AL*N or BR*N AL*N/BR*R AL*N or BR*N AL*N/BR*N AL*A or BR*N AL*N/W), roux (BR*R AL*N/BR*R AL*N or BR*R AL*A/BR*R AL*N or BR*R AL*N/W), and imperfect albino (BR*N AL*A/BR*N AL*A or BR*R AL*A/BR*N AL*A or BR*N AL*A/W or BR*R AL*A/W). Linkage Analysis of BR*B and AL*C Test crosses of double heterozygous wildtype males (BR*B AL*N/BR*N AL*C or BR*N AL*N/BR*B AL*C) with double recessive hemizygous golden females (BR*B AL*C/W) were performed. The golden quail could be easily identified from either the brown or cinnamon quail because they showed the lightest color among the three phenotypes. Chicks were divided into parental and recombinant types according to the resulting four down colors: wild-type (BR*N AL*N/BR*B AL*C or BR*N AL*N/W), brown (BR*B AL*N/BR*B AL*C or BR*B AL*N/W), cinnamon (BR*N AL*C/BR*B AL*C or BR*N AL*C/W), and golden (BR*B AL*C/BR*B AL*C or BR*B AL*C/W). Estimation of Recombination Frequency Between BR and AL The maximum likelihood method described by Green (1963) was used to estimate the recombination frequency (␪) and its standard error (SE).


Results and Discussion

Table 1. Allelism test for roux and brown plumage color in Japanese quail

Table 1 shows the results of the allelism test between *B and *R. From the crosses of brown males and roux females, a total of 144 F1 chicks hatched, and all of them were brown. However, the reciprocal cross between roux cocks and brown hens produced 91 brown and 103 roux F1 chicks. No wild-type or other down colors appeared in those crosses. As shown in Figures 1 and 2, the roux is paler than the brown in both chicks and adults. These results were in agreement with our hypothesis that *R and *B were alleles, and they indicated that BR*R was recessive to BR*B. Also, when the sex of the 52 tenweek-old progeny from the crosses between roux males and brown females was examined, the 27 brown birds and the 25 roux birds were found to be all male or female, respectively. This result was in agreement with that expected for two sexlinked alleles. For further confirmation ( Table 2), crosses of BR*B/BR*R crossbred males with BR*B/W and BR*R/W pure-line females produced 125 brown and 33 roux, and 52 brown and 63 roux F2 chicks, respectively. Corresponding segregation ratios did not deviate significantly from expected values 3:1 and 1:1, respectively. Therefore all results were compatible with the hypothesis that a single sex-linked gene BR with two alleles, *B and *R, determined brown and roux plumage color in Japanese quail. The results of the linkage analysis between BR*R and AL*A are shown in Table 3. The first test cross (BR*R AL*N/BR*N AL*A ⫻ BR*R AL*N/W) yielded 508 wildtype (BR*N AL*A/BR*R AL*N or BR*N AL*N/BR*R AL*N or BR*N AL*N/W), 620 roux (BR*R AL*N/BR*R AL*N or BR*R AL*A/BR*R AL*N or BR*R AL*N/W), and 347 imperfect albino (BR*N AL*A/W or BR*R AL*A/W). The second test cross (BR*R AL*N/BR*N AL*A ⫻ BR*N AL*A/W) produced 364 wild-type (BR*R AL*N/BR*N AL*A or BR*N AL*N/BR*N AL*A or BR*N AL*N/W), 161 roux (BR*R AL*N/W), and 501 imperfect albino (BR*N AL*A/BR*N AL*A or BR*R AL*A/BR*N AL*A or BR*N AL*A/W or BR*R AL*A/W). The results of linkage analysis for BR*B and AL*C are presented in Table 4. In the progeny of the first cross (BR*B AL*N/ BR*N AL*C ⫻ BR*B AL*C/W), brown and cinnamon were parental type, and golden and wild-type were recombinant type, and it was the opposite in the progeny of the second cross (BR*B AL*C/BR*N AL*N ⫻ BR*B AL*C/W). The total offspring from

Mating typea ( F0 cock ⫻ F0 hen)

Total

Brown

Roux

Expected ratiob

␹2

Brown ⫻ brown Brown ⫻ roux Roux ⫻ brown Roux ⫻ roux

162 144 194 195

162 144 91 0

0 0 103 195

1:0 1:0 1:1 0:1

— — 0.742 —


a b

Number of newly hatched F1 chicks

Four single-pair matings per mating type. Based on the assumption that roux and brown are controlled by sex-linked alleles *R and *B, with *B being dominant over *R.

Table 2. Segregation of plumage color in the progeny of heterozygous F1 males and pure-line F 1 females

a b

Number of newly hatched F2 chicks

Mating typea ( F1 cock ⫻ F1 hen)

Total

Brown

Roux

Expected ratiob

␹2

Brown ⫻ brown Brown ⫻ roux

158 115

125 52

33 63

3:1 1:1

1.426 1.052

Eight single-pair matings per mating type. Based on the assumption that roux and brown are controlled by sex-linked alleles *R and *B, with *B being dominant over *R.

Table 3. Linkage analysis of the locia for roux and imperfect albino plumage color in Japanese quail Number of offspring Mating type (cock ⫻ hen) Wild-type ⫻ roux (BR*R AL*N/BR*N AL*A ⫻ BR*R AL*N/W) Expected frequencyb Expected number Wild-type ⫻ imperfect albino (BR*R AL*N/BR*N AL*A ⫻ BR*N AL*A/W) Expected frequencyb Expected number

Wild-type

Roux

Imperfect albino

Total

508 (1 ⫹ ␪)/4 509.2

620 (2 ⫺ ␪)/4 597.0

347 1/4 368.8

1475 1 1475.0

364 (1 ⫹ ␪)/4 354.2

161 (1 ⫺ ␪)/4 158.8

501 1/2 513.0

1026 1 1026.0

Locus BR has three alleles: *N (wild-type), *B ( brown), *R (roux); locus AL has three alleles: *N (wild-type), *C (cinnamon), *A (imperfect albino). b ␪ is the recombination frequency.

a

Table 4. Linkage analysis of the locia for brown and cinnamon plumage color in Japanese quail Number of offspring

a

b

Mating type (cock ⫻ hen)

Parental type

Recombinant type

Total

Wild-type ⫻ golden (BR*B AL*N/BR*N AL*C ⫻ BR*B AL*C/W) and (BR*N AL*N/BR*B AL*C ⫻ BR*B AL*C/W) Expected frequencyb Expected number

1307 1⫺␪ 1308.6

807 ␪ 805.4

2114 1 2114.0

Locus BR has three alleles: *N (wild-type), *B ( brown), *R (roux); locus AL has three alleles: *N (wild-type), *C (cinnamon), *A (imperfect albino). ␪ is the recombination frequency.

mique, Laboratoire de Geńe´tique Factorielle, Jouy-enJosas, France (Minvielle), Faculty of Agriculture, Gifu University, Gifu 501-1193, Japan ( Ito and Inoue-Murayama), Nippon Institute for Biological Science, Kobuchizawa, Japan (Mizutani), and Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya, Japan (Wakasugi). We would like to thank Dr. J.-L. Gueńet at Institut Pasteur, Paris, Dr. M. Niwa-Kawakita at Institut Curie, Paris, J.-L. Monvoisin at INRA, Jouy-en-Josas, and Dr. Y. Maeda, Kagoshima University, Kagoshima, for their cooperation. The final version of this article was prepared when Francis Minvielle was at Kagoshima University under a long-term Invitation Research Fellowship from the Japan Society for the Promotion of Science (JSPS). Address correspondence to Shin’ichi Ito at the address above. 䉷 2000 The American Genetic Association

References Bitgood JJ and Somes RG Jr, 1990. Linkage relationships and gene mapping. In: Poultry breeding and genetics (Crawford RD, ed). Amsterdam: Elsevier; 469– 495. Figure 1. Wild-type ( left), brown (center), and roux (right) Japanese quail chicks.

both crosses segregated into 1307 parental type and 807 recombinant type birds. Because *B and *R were shown to be at the same locus, breeding results from Tables 3 and 4 could be merged in order to obtain optimal maximum likelihood estimation of the recombination frequency ␪ between BR and AL. The estimation of ␪ ⫾ SE was 0.381 ⫾ 0.010. Observed numbers in phenotypic classes were in good agreement with expected ones obtained by using the estimate of ␪ (␹2 ⫽ 2.758, df ⫽ 4, .60 ⬎ P ⬎ .50). The recombination frequency estimated in the present work was higher than that reported by Wakasugi and Kondo (1973), and the difference may be attributed to the sample size (4615 versus about 400). To conclude, Japanese

quail have two different sex-linked plumage color loci, AL and BR, each one with three known alleles. The first locus, already known, has alleles *N, *C, and *A, and alleles at the second one are *N, *B, and *R. These mutant genes could be used for sexing newly hatched chicks in commercial and research fields. Indeed, the roux allele has been introgressed in commercial stock in France (Minvielle et al. 1999, 2000). These quail data contribute also to intercomparisons of avian species, as already noted for the Japanese quail and the turkey ( Bitgood and Somes 1990), and for the Japanese quail and the chicken (Silversides and Me´rat 1991). From the Institut National de la Recherche Agrono-

Crawford RD, 1990. Origin and history of poultry species. In: Poultry breeding and genetics (Crawford RD, ed). Amsterdam: Elsevier; 1–41. Crittenden LB, Bitgood JJ, Burt DW, Ponce de Leon FA, and Tixier-Boichard M, 1996. Nomenclature for naming loci, alleles, linkage groups and chromosomes to be used in poultry genome publications and databases. Genet Sel Evol 28:289–297. Green MC, 1963. Methods for testing linkage. In: Methodology in mammalian genetics ( Burdette WJ, ed). San Francisco: Holden-Day; 56–82. Homma K, 1968. Sex-linked dilute character in the Japanese quail. In: Proceedings of the 55th meeting of the Japanese Society of Zootechnical Science, Obihiro. Tokyo: Japanese Society of Zootechnical Science; 47. Homma K and Jinno M, 1969. Sex-linked plumage characters in the Japanese quail. In: Proceedings of the 57th meeting of the Japanese Society of Zootechnical Science, Kagoshima. Tokyo: Japanese Society of Zootechnical Science; 82. Homma K, Jinno M, Sato K, and Ando A, 1968. Studies on perfect and imperfect albinism in the Japanese quail (Coturnix coturnix japonica). Jpn J Zootech Sci 39:348– 352. Lauber JK, 1964. Sex-linked albinism in the Japanese quail. Science 146:948–950. Minvielle F, Gandemer G, Maeda Y, Leborgne C, Hirigoyen E, and Boulay M, 2000. Carcass characteristics of a heavy Japanese quail line under introgression with the roux gene. Br Poult Sci 41:41–45. Minvielle F, Hirigoyen E, and Boulay M, 1999. Associated effects of the roux plumage color mutation on growth, carcass traits, egg production, and reproduction of Japanese quail. Poult Sci 78:1479–1484. Silversides FG and Me´rat P, 1991. Homology of the s⫹ locus in the chicken with Al⫹ in the Japanese quail. J Hered 82:245–247. Sittmann K, Wilson WO, and McFarland LZ, 1966. Buff and albino Japanese quail. J Hered 57:119–124. Somes RG Jr, 1988. International registry of poultry genetic stocks. Storrs, CT: Storrs Agricultural Experiment Station. Truax RE and Johnson WA, 1979. Genetics of plumage color mutants in Japanese quail. Poult Sci 58:1–9. Wakasugi N, 1984. Japanese quail. In: Evolution of domesticated animals (Mason IL, ed). New York: Longman; 319–321. Wakasugi N and Kondo K, 1973. Breeding methods for maintenance of mutant genes and establishment of strains in the Japanese quail. Exp Anim 22(suppl):151– 159. Received December 1, 1999 Accepted July 17, 2000

Figure 2. Wild-type ( left), brown (center), and roux (right) Japanese quail adults.

Corresponding Editor: Susan J. Lamont


Fifty Microsatellite Markers for Japanese Quail B. B. Kayang, M. Inoue-Murayama, A. Nomura, K. Kimura, H. Takahashi, M. Mizutani, and S. Ito A Japanese quail genomic library enriched for (CA/GT)n simple sequence repeats was screened and positive clones were sequenced. Fifty original microsatellite sequences were isolated that consisted mainly of perfect repeats of the dinucleotide (CA/ GT)n motif and a corresponding number of polymerase chain reaction (PCR) primer pairs complementary to unique DNA sequences flanking the microsatellite repeats were designed to detect the repeats. Fortysix percent (23 of 50) of the markers revealed polymorphism in two unrelated quail individuals (one male and one female) randomly sampled from a population of wild quail origin. All 50 primer pairs were tested in the PCR for their ability to amplify chicken genomic DNA. Amplification products were obtained for 14 (28.0%) of the markers at the annealing temperature optimized for quail. These results provide an opportunity to begin characterizing the quail genome for the development of a genetic map for this economically valuable species and the eventual construction of a comparative genetic map in Phasianidae, which comprises a number of agriculturally important species of poultry. The construction of genomic maps for farm animals has become an indispensable tool in the study of the genetics underlying the control of economically important traits. In recent years, partial genetic maps have been reported for farm animals such as cattle ( Bishop et al. 1994), sheep (Crawford et al. 1995), goat ( Vaiman et al. 1996), pig (Rohrer et al. 1994), and chicken (Groenen et al. 2000). The construction of these maps has been greatly boosted by the development of highly polymorphic DNA markers, particularly microsatellite markers (Georges and Andersson 1996). Microsatellites, also known as simple sequence repeats (SSRs), are tandem repeats of 1–6 bases in length that occur abundantly and at random in most eukaryotic genomes ( Hamada et al. 1982; Stallings et al. 1991; Tautz and Renz 1984). Their high level of polymorphism ( Tautz 1989) and codominance mode of inheritance, coupled with the ease with which


they can be typed using the polymerase chain reaction (PCR), have made them the markers of choice in genome mapping and linkage analysis (Cheng et al. 1995) as well as analyses of genetic diversity and evolution of animal genomes (MacHugh et al. 1997; O’Brien et al. 1993; Takahashi et al. 1998). Linkage maps of polymorphic markers are being developed in agriculturally important species. In poultry, mapping efforts have so far concentrated on the chicken (Gallus gallus), with more than 800 microsatellite markers isolated (Groenen et al. 2000) while much remains to be done on other economically important species. Japanese quail (Coturnix japonica) is one such species for which a genetic map is yet to be constructed. However, not only is it a valuable egg and meat producer (Minvielle 1998), but it has long been recognized and used as an excellent laboratory research animal (Padgett and Ivey 1959) and has also been recommended as a pilot animal for poultry (Wilson et al. 1961) because of its small body size, rapid generation turnover, and high egg production (Wakasugi and Kondo 1973). In view of this, mapping efforts in this species would not only be essential in the genetic improvement of quail per se but could also be of economic benefit to other closely related species, especially if Japanese quail is to used as a model for poultry. At present, marker information in Japanese quail is very scanty. Only two autosomal linkage groups based on plumage color and blood protein markers have been reported ( Ito et al. 1988a,b; Shibata and Abe 1996), while DNA markers are yet to be developed. Recently attempts have been made to isolate microsatellites in Japanese quail using chicken-specific primers. While one report indicated specific amplification products in 26.7% (32 of 120) of chicken primers tested ( Inoue-Murayama et al. 1998), another study showed that 22.9% (11 of 48) of chicken primers amplified a locus in Japanese quail (Pang et al. 1999). This confirms that some marker information is shared by these species and would thus be useful for comparative mapping that is needed for the exchange of genetic information essential in promoting the genetic improvement of their economic traits. However, the limited rate of success achieved so far indicates that this screening procedure using heterologous primers designed for chicken may not yield sufficient quail markers for genome mapping or quantitative trait loci (QTL) analysis (Pang et al. 1999), thus un-

derscoring the need to isolate original markers in Japanese quail. As a preliminary step in the construction of a quail genetic map and the construction of a comparative genetic map in Phasianidae, which includes several agriculturally important species of poultry, this study was conducted to isolate original polymorphic microsatellite markers in Japanese quail.

Materials and Methods A quail colony maintained at Gifu University was used in this study. This colony, which originated from wild quail initially captured and domesticated at the National Institute of Genetics, Mishima, in 1968– 1970 ( Kawahara 1973), was introduced from the National Institute of Animal Industry, Tsukuba. Approximately 1 ml of blood was drawn from the jugular vein of one male and one female randomly sampled from the colony and DNA extracted using the QIAamp blood kit (Qiagen, Valencia, CA). DNA was similarly obtained from blood collected by wing venipuncture from one female White Leghorn and one female Fayoumi chicken and later used for tests of quail primers in chickens. Both breeds of chicken were sampled from populations maintained at the Nippon Institute for Biological Science. A DNA library enriched for the dinucleotide repeat motif (CA/GT )n was constructed using genomic DNA isolated from the blood of a female Japanese quail, which is the heterogametic sex, following the method of Takahashi et al. (1996). The enriched library was transformed into LX2-Blue MRF’ Ultracompetent Cells (Stratagene, La Jolla, CA). Clones having microsatellite sequences were detected by hybridization to horse radish peroxidaselabeled (CA/GT )n probe using the ECL direct nucleic acid labeling and detection systems (Amersham Pharmacia Biotech, Buckinghamshire, England). Plasmids were extracted from positive clones using the GFX Micro Plasmid Prep Kit (Amersham Pharmacia Biotech), electrophoresed on 1% agarose gel to confirm extraction as well as reveal DNA density, and then sequenced by the dye termination method employing an ABI Prism 377 DNA sequencer (Perkin-Elmer, Foster City, CA). The resulting sequences allowed for the verification of microsatellite-containing clones and the designing of PCR primers flanking the microsatellite repeat. Primers for PCR typing were 18–22 bp long and were designed with the assistance of GENETYX-Homology version 2.2.2

Table 1. Characterization of simple sequence repeats in Japanese quail

Locus name

Accession number

Repeat array

Forward primer (5’–3’)

Reverse primer (5’–3’)

GUJ0001 GUJ0002 GUJ0003 GUJ0004 GUJ0005 GUJ0006 GUJ0007 GUJ0008 GUJ0009 GUJ0010 GUJ0011 GUJ0012 GUJ0013 GUJ0014 GUJ0015 GUJ0016 GUJ0017 GUJ0018 GUJ0019 GUJ0020 GUJ0021 GUJ0022 GUJ0023 GUJ0024 GUJ0025 GUJ0026 GUJ0027 GUJ0028 GUJ0029 GUJ0030 GUJ0031 GUJ0032 GUJ0033 GUJ0034 GUJ0035 GUJ0036 GUJ0037 GUJ0038 GUJ0039 GUJ0040 GUJ0041 GUJ0042 GUJ0043 GUJ0044 GUJ0045 GUJ0046 GUJ0047 GUJ0048 GUJ0049 GUJ0050

AB035652 AB035813 AB035814 AB037157 AB035815 AB035816 AB035817 AB035818 AB035819 AB035820 AB035821 AB035822 AB035823 AB035824 AB035825 AB035826 AB035827 AB035828 AB035829 AB035830 AB035831 AB035832 AB035833 AB035834 AB035835 AB035836 AB035837 AB035838 AB035839 AB035840 AB035841 AB035842 AB035843 AB035844 AB035845 AB035846 AB035847 AB035848 AB035849 AB035850 AB035851 AB035852 AB035853 AB035854 AB035855 AB035856 AB035857 AB035858 AB035859 AB035860

(CA)7TG(CA)13 (CA)13 (CA)9 (CA)10 (CT )11CG(CA)13 (CA)14 (CA)15 (CA)10 (CA)14 (CA)15 (CA)13 (CA)6TA(CA)6 (CA)10 (CA)9 (CA)9 (CA)9 (CA)14 (CA)10 (CA)21 (CA)8 (CA)11 (CA)15 (CA)7TA(CA)11 (CA)13AA(CA)3 (CA)9 (CA)16 (CA)15 (CA)9 (CA)11CT(CA)2 (CA)31 (CA)9 (CA)5CTG(CA)9 (CA)13 (CA)9CG(CA)2 (CA)14 (CA)9TA(CA)4 (CA)10C(CA)2 (CA)19 (CA)19 (CA)12 (CA)11 (CA)8 (CA)9TGTG(CA)2 (CA)16 (CA)18 (CA)9 (CA)23 (CA)14 (CA)11 (CA)8

GAAGCGAAAGCCGAGCCA AGGTTGTGCTTTGCTTGTAT AGGGAAGAAGCAACTGTTC AGCTCTCCTATGGGGCAAC GCTCTGCTCTCACAGCAGT TGGGATGATAATGAGGTACGG TGACTGCTTTCCACACACA CATGGTTATCAACCTGCAGA CACGCTTGCTTCTTGCTTCA TTCCTTCTGGGTGCTGCTCA TACTTGATACACCAGCTGTC TTTATGTACTGTTTGGGCGC ACCAAACCCGAGATCCGACA TGCTGGGGTTGCTTTCTCCA AGGTGGTCCCCAATGCCCTT AATGAATGTCTGGGTGGTGC AGAGAGATTAGAGGAGCTGC ATCCCGCGCCGTCCTTTGTT GGGGGCTGTAGGTCTGGATC AATGTCCTTGTGCAGCTCCA GAGCATTTCTAGTCTGTCTC AAACTTATTCTCGCGCTCCC GAGAGGTACAGCAACACTTT TCACACCTTCGGGCTGATCT CCTGAGCGAATACACAACTG CATGAACATCTCTCTTCATG TTCACAGATGACAATCTAGC TGAACAAAGCAGAAAGGAGC GAGCATTTCTAGTCTGTCTC TGCACCAATCCCAGCTGTTT AAGGGCAGGGGCTGGGAACA GAGGCTGCGAACAACACACA TCTGCTCTCACAGCAGTGCA CGTAACGGTCCAATATGGAT AATACTGGTTTTGTGATGGC CTTTCACATTGCTTTTGCCT CCATTCCTCCATCGTTCTGA TACATCCAGCAATCGCCCAC CAAAGAGCAGAGGGAATGGA GTTGAAGCTCCCATCCCTCC AAAATGTCTGCAAAATGGGC TCAGTGCCTTTGTGTTGTCC GAGACCAGGTGGTCCCCAAT GCCTTGAAACCTGAGTGATC ACATGCACCACCATTCTTGC GCCATGTTTGTCACCTTGCA GAGATAAGACTGGCTGGGGC AACGCATACAACTGACTGGG GAAGCAGTGACAGCAGAATG CTGCCATGTTACTAATCTAG

CAGCACTTCGGAGCACAGGA GAGCATGTTGCACATTTCTT ATTCCAGAATCTGGACTGG CTGAGCACGAGGACTGGGAA TGGATCTGGAGCTGCAACGC AGGATAGCATTTCAGTCACGG CAGAAGGTAAAAGGACGGA ACATGCCAGTCCTTCACAAT TATGTTTGGTGCCCTGCTAG CATAGACACATCCCTCCCTC CACCCTATACCAATGAAAGG CTTGGACATAGAGTAAGCCA AGCGTTCGCGTTCCTCTTTC TCTCGGTGGTTTGCTCTGAC GGAAGCAGAGCATCGTTCCC CATGGAGTGTTGGGTATTGC GGCACTAAAACCATCGAGAG CGGCACCACGAAGTACTCCA ATCGGGCACGCGAGGACCAT CAGCATTGTGCAAAGCAGTG GATCAATACACAGGCTAAGG TAAGCAAGGAAGAGGTGGCA CGTTTCTTTCTGGAGTGTCT ATGCGACGGGGTGCCTTAAA AGTGTTAGGTGAGGACTGCT GTGTTCTGCATCACAAACAT CTGCAAGTAACAGAAGGTAA CCTTACCTACATGAAACGTC ATACACAGGCTAAGGAAACC AACGCACAATGGAAAGTGGG CGCCTCTGCGGTGTGCAACT GCTAAGACGAGGTGAAGGCT GCATAGAGCCCAGCAGTGTT TCCACGATGCAGAGGTATTT GGGCAATAAAAGAAAGACTG CACTAAAGATTGGCTAACAG GGGAAGGAGTGTAGGAAAGA CACGGGTGAGTCCATTAGTG CCGAGAGATGGGTTTTTTCC ACACCCCCACGGTCTTTGCA TGAAACATACCTGAGTGCTA ACAGCCTTCCCCAAATTCCT GGAAGCAGAGCATCGTTCCC TGCATTTCAGCAGCTCTCAG CATGCACAAATGAGCGTGCA ACTGGTTGGGACTGAAGGAT TCACCGTGGCTGGCCAACTT GGATAGCATTTCAGTCACGG CGGTAGCATTTCTGACTCCA TGGTTTCTTTACACTTGACA

a

Ta quail

AmplificaLength tion in chickenc ( bp)

Polymorphismd

56 50 48 59 59 55 51 58 60 62 58 58 55 60 60 55 60 55 50 64 62 69 55 55 60 60 55 55 55 64 55 55 55 55 55 55 55 60 60 55 55 55 55 55 60 55 55 55 55 55

⫹ 0 ⫹ 0 0 0 0 ⫹ 0 ⫹ 0 0 ⫹ 0 0 0 ⫹ 0 0 0 ⫹ 0 ⫹ 0 0 0 ⫹ 0 ⫹ 0 0 0 0 0 0 0 0 0 0 0 0 ⫹ 0 ⫹ 0 ⫹ 0 0 ⫹ 0

⫹ ⫺ ⫹ ⫺ ⫺ ⫹ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫺

b

235 149 146 185 135 125 90 174 205 156 164 144 139 147 135 249 165 245 191 209 157 136 237 170 245 120 176 150 152 181 164 199 199 247 152 151 182 262 188 180 130 191 141 206 251 236 284 140 243 147

The locus code GUJ stands for Gifu University Japanese quail and is in accordance with the standardized nomenclature procedure proposed by Crittenden et al. (1996). Annealing temperature. c The ⫹ sign indicates that amplification products were obtained using the same annealing temperature as in quail. The 0 indicates that amplification products were not obtained using the same annealing temperature as in quail. d The ⫹ and ⫺ signs indicate polymorphic and nonpolymorphic in two unrelated quail individuals (one male and one female) randomly sampled from the population.

a

b

(Software Development, Tokyo, Japan) so as to amplify DNA fragments in the range of 100–300bp. PCR was performed on a Takara PCR Thermal Cycler ( Takara Biomedicals, Tokyo, Japan) in 15 ␮l reaction mixtures containing 20 ng of the DNA template, 0.3 ␮M of forward and reverse primers, 130 ␮M of each dNTP, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.7 mM MgCl2 and 0.6 U AmpliTaq Gold (Perkin-Elmer). After initial incubation at 95⬚C for 9 min, PCR amplification was performed for 30 cycles each with denaturing at 95⬚C for 30 sec, annealing for 1 min at 48⬚C–69⬚C (as opti-

mized per marker), and extension at 72⬚C for 1 min. This was followed by a final cycle at 72⬚C for 5 min. To determine their suitability for PCR, primer pairs were first tested in one male and one female randomly sampled from the quail colony. Five microliter aliquots of PCR products were analyzed on 2.5% agarose gels containing 0.05 ␮g/ml of ethidium bromide in 1⫻ TBE buffer using a Mupid Minigel electrophoresis system (Cosmo Bio, Tokyo, Japan), at 100 V for 40–45 min. Initially two annealing temperatures, 55⬚C and 60⬚C, were used and if no

detectable PCR product bands were observable after electrophoresis, the PCR was repeated using a one-degree incremental series (⫾1) employing a gradient thermal cycler (Corbett Research, Mortlake, Australia) until an optimum temperature (Ta) was obtained. Using the annealing temperature optimized for quail, primer pairs were tested in chicken, employing DNA templates obtained from White Leghorn and Fayoumi breeds, to determine their applicability as markers for chickens. Finally, a homology search with se-


quences registered in GenBank was carried out using the DBGET database system ( http://www.genome.ad.jp, visited June 8, 2000).

Results and Discussion Out of a total of 1273 clones containing inserts, 372 (29.2%) gave a positive signal after hybridization to the (CA/GT )n probe. Of the positive clones, 368 were sequenced and characterized. The remaining four samples were not sequenced because prior screening on 1% agarose gel revealed double or weak bands. Of the sequenced clones, 248 (67.4%) contained (CA/GT )n microsatellites with six or more repeat numbers and were further characterized, while the remainder were considered uninformative due to absent, small, or interrupted microsatellites. Of the clones that had a large enough repeat array (n ⬎ 6), 176 (71.0%) were duplicated sequences previously isolated from the same library. Takahashi et al. (1996) also found 50.8% duplicates in their (CA/GT )nenriched chicken library. This is in contrast to Gibbs et al. (1997) who only found 19.4% duplicates from their chicken library using a different enrichment protocol (Armour et al. 1994). Takahashi et al. (1996) recognized that conditions used in the construction of their library (Ostrander et al. 1992) could result in a high frequency of clone duplication and have since improved the efficiency of the library by the use of sonicated genomic DNA fragments and the use of more efficient cloning vectors and enzymes ( Takahashi H, personal communication). Seventy-two (19.6%) of the clones sequenced were found to contain unique microsatellites of the (CA/GT )n-repeat type. However, only 50 of them could be developed as PCR primer pairs to detect (CA/ GT )n repeats. The rest were unsuitable either due to insufficient sequence information or short sequences flanking the repeat. The sequences of the 50 unique clones have been submitted to GenBank. The locus name, GenBank accession number, microsatellite repeat array, as well as the primer pairs designed for regions flanking the microsatellite repeats are shown in Table 1. The number of (CA/ GT )n repeats in these clones varied between 6 and 31, which is close to the range of 6–28 repeats reported by Takahashi et al. (1996) for chicken. Seventy-eight percent of the microsatellites were perfect repeats, 20.0% were imperfect repeats, and 2.0% had a compound repeat of (CT )11


and (CA)13. The allele size of the unique clones ranged from 90 to 284 bp, while the optimum annealing temperature in quail was from 48⬚C to 69⬚C (see Table 1). Fortysix percent (23 of 50) of the markers revealed polymorphism in two unrelated quail individuals (one male and one female) randomly sampled from the population ( Table 1). It is thus very likely that more of the markers would be polymorphic if more individuals are screened. An homology search conducted to find out whether any of the 50 sequences matched the sequences registered in GenBank showed no significant homologies. As shown in Table 1, 14 (28.0%) of the quail-specific primers successfully amplified loci in chicken at the annealing temperature optimized for quail. These results were the same when both White Leghorn and Fayoumi DNA were used as templates. Chicken-specific primers have been successfully tested in other Phasianidae on the assumption that since they are related species their SSR loci could be similar. For example, in turkeys 92.0% of 48 chicken primer pairs yielded PCR products at 4 mM MgCl2 concentration ( Levin et al. 1995), whereas 69.1% of 88 primer pairs gave amplification products at 1.5 mM MgCl2 concentration ( Liu et al. 1996). Similarly, studies in Japanese quail have also indicated that 26.7% of 120 primers ( Inoue-Murayama et al. 1998) and 22.9% of 48 chicken primers (Pang et al. 1999) amplified loci (1.7 mM and 2–2.5 mM MgCl2, respectively). While the application of chicken markers to other Phasianidae is useful in view of the large number of chicken microsatellites available, it is also important, in comparative studies, to determine the applicability of markers isolated from other Phasianidae to chicken. Our study is the first attempt to detect microsatellite loci in chicken using quail-specific primers. A similar study involving turkey ( Huang et al. 1999) showed that 80.0% of 60 turkey-specific primer pairs amplified products in chicken at 2 mM MgCl2 concentration. Our low success rate (28.0%) could be attributed to the relatively high stringency that was employed in our amplification reactions (1.7 mM MgCl2) and our rigid enforcement of the optimum annealing temperatures determined for quail. It is possible that a higher success rate could have been achieved by adjusting the PCR conditions for amplification in chickens. In this study, the quail markers that successfully amplified loci in chickens suggest that quail markers may also be useful in chickens and that such mark-

ers would be important in the construction of a comparative genetic map in these species. However, the products amplified from chickens need to be sequenced to confirm their homology with the quail sequences. In summary, we have used a library enriched for SSRs to successfully isolate microsatellite loci from Japanese quail. This demonstrates that original SSR loci in the quail genome may be targeted for use in marker development for this species. The present work indicates that the markers isolated would not only provide a useful base for quail genome mapping, but also for comparative mapping especially with chickens. However, many more markers must be isolated to provide the resources necessary for the construction of a comprehensive quail genetic map and eventually the construction of a comparative genetic map in Phasianidae. From the United Graduate School of Agricultural Science ( Kayang) and the Faculty of Agriculture ( InoueMurayama, Nomura, Kimura, and Ito), Gifu University, Gifu 501-1193, Japan, National Institute of Agrobiological Resources, Tsukuba, Japan ( Takahashi), and Nippon Institute for Biological Science, Kobuchizawa, Yamanashi, Japan (Mizutani). We gratefully acknowledge the dedicated technical assistance of Y. Ueda, whose efforts greatly aided this work. This research was financially supported by the Japan Livestock Technology Association. Address correspondence to M. InoueMurayama at the address above or e-mail: [email protected]. 䉷 2000 The American Genetic Association

References Armour JAL, Neumann R, Gobert S, and Jeffreys AJ, 1994. Isolation of human simple repeat loci by hybridization selection. Hum Mol Genet 3:599–605. Bishop MD, Kappes SM, Keele JW, Stone RT, Sunden SL, Hawkins GA, Solinas TS, Fries R, Grosz MD, Yoo J, and Beattie CW, 1994. A genetic linkage map for cattle. Genetics 136:619–639. Cheng HH, Levin I, Vallejo RL, Khatib H, Dodgson JB, Crittenden LB, and Hillel J, 1995. Development of a genetic map of the chicken with markers of high utility. Poult Sci 74:1855–1874. Crawford AM, Dodds KG, Ede AJ, Pierson CA, Montgomery GW, Garmonsway HG, Beattie AE, Davies K, Maddox JF, Kappes SW, Stone RT, Nguyen TC, Penty JM, Lord AE, Broom JE, Buitkamp J, Schwaiger W, Epplen JT, Matthew P, Matthews ME, Hulme DJ, Beh KJ, McGraw RA, and Beattie CW, 1995. An autosomal genetic linkage map of the sheep genome. Genetics 140:703– 724. Crittenden LB, Bitgood JJ, Burt DW, Ponce de Leon FA, and Tixier-Boichard M, 1996. Nomenclature for naming loci, alleles, linkage groups and chromosomes to be used in poultry genome publications and data bases. Genet Sel Evol 28:289–297. Georges M and Andersson L, 1996. Livestock genomics comes of age. Genome Res 6:907–921. Gibbs M, Dawson DA, McCamley C, Wardle AF, Armour JAL, and Burke T, 1997. Chicken microsatellites isolated from libraries enriched for simple tandem repeats. Anim Genet 28:401–417. Groenen MA, Cheng HH, Bumstead N, Benkel BF, Briles

WE, Burke T, Burt DW, Crittenden LB, Dodgson J, Hillel J, Lamont S, de Leon AP, Soller M, Takahashi H, Vignal A, 2000. A consensus linkage map of the chicken genome. Genome Res 10:137–147. Hamada H, Petrino MG, and Kakunaga T, 1982. A novel repeated element with Z-DNA-forming potential is widely found in evolutionary diverse eukaryotic genomes. Proc Natl Acad Sci USA 79:6465–6469. Huang HB, Song YQ, Hsei M, Zahorchak R, Chiu J, Teuscher C, and Smith JE, 1999. Development and characterization of genetic mapping resources for the turkey (Melleagris gallopavo). J Hered 90:240–242. Inoue-Murayama M, Nomura A, Ide H, Kimura K, Tatsuda K, Takahashi H, Nagamine Y, Hanada H, Takeda T, Tsudzuki M, and Ito S, 1998. Application of chicken microsatellite markers to Japanese quail and Chinese painted quail. Anim Genet 29(suppl 1):46. Ito S, Kimura M, and Isogai I, 1988a. Linkage between panda plumage and albumin loci in Japanese quail. Jpn J Zootech Sci 59:822–824. Ito S, Kimura M, and Isogai I, 1988b. A sex difference in recombination values between extended brown and phosphoglucose isomerase loci in Japanese quail. Jpn J Zootech Sci 59:801–805. Kawahara T. 1973. Comparative study of quantitative traits between wild and domestic Japanese quail (Coturnix coturnix japonica). Exp Anim 22(suppl):139–150. Levin I, Cheng HH, Baxter-Jones C, and Hillel J, 1995. Turkey microsatellite loci amplified by chicken-specific primers. Anim Genet 26:107–110. Liu Z, Crooijmans RPMA, van der Poel JJ, and Groenen

MAM, 1996. Use of chicken microsatellite markers in turkey: a pessimistic view. Anim Genet 27:191–193. MacHugh DE, Shriver MD, Loftus RT, Cunningham P, and Bradley DG, 1997. Microsatellite DNA variation and the evolution, domestication and phylogeography of taurine and zebu cattle (Bos taurus and Bos indicus). Genetics 146:1071–1086. Minvielle F, 1998. Genetics and breeding of Japanese quail for production around the world. In: Proceedings of the 6th Asian Pacific Poultry Congress, Nagoya, Japan, June 4–7, 1998. Tsukuba: Japan Poultry Science Association; 122–127. O’Brien SL, Womack J, Lyons LA, Moore KL, Jenkins NA, and Copeland NG, 1993. Anchored reference loci for comparative genome mapping in mammals. Nat Genet 3:103–112. Ostrander EA, Jong PM, Rine J, and Duyk G, 1992. Construction of small insert genomic DNA libraries highly enriched for microsatellite repeat sequences. Proc Natl Acad Sci USA 89:3419–3423. Padgett CA and Ivey WD, 1959. Coturnix quail as a laboratory research animal. Science 129: 267–268. Pang SWY, Ritland C, Carlson JE, and Cheng KM, 1999. Japanese quail microsatellite loci amplified with chicken-specific primers. Anim Genet 30:195–199. Rohrer GA, Alexander LJ, Keele JW, Smith TP, and Beattie CW, 1994. A microsatellite map of the porcine genome. Genetics 136:231–245. Shibata T and Abe T, 1996. Linkage between the loci for serum albumin and vitamin D binding protein (GC) in the Japanese quail. Anim Genet 27:195–197.

Stallings RL, Ford AF, Nelson D, Torney DC, Hildebrand CE, and Moyzis RK, 1991. Evolution and distribution of (GT )n repetitive sequences in mammalian genomes. Genomics 10:807–815. Takahashi H, Nirasawa K, and Furukawa T, 1996. An efficient method to clone chicken microsatellite repeat sequences. Jpn Poult Sci 33:292–299. Takahashi H, Nirasawa K, Nagamine Y, Tsudzuki M, and Yamamoto Y, 1998. Genetic relationships among Japanese native breeds of chicken based on microsatellite DNA polymorphisms. J Hered 89:543–546. Tautz D and Renz M, 1984. Simple sequences are ubiquitous repetitive components of eukaryotic genomes. Nucleic Acids Res 12:4127–4138. Tautz D, 1989. Hypervariability of simple sequences as a general source for polymorphic DNA markers. Nucleic Acids Res 17:6563–6571. Vaiman D, Scibler L, Bourgeois F, Oustry A, Amigues Y, and Cribu EP, 1996. A genetic linkage map of the male goat genome. Genetics 144:279–305. Wakasugi N and Kondo K, 1973. Breeding methods for maintenance of mutant genes and establishment of strains in the Japanese quail. Exp Anim 22(suppl):151– 159. Wilson WO, Abbott UK, and Abplanalp H, 1961. Evaluation of Coturnix (Japanese quail) as pilot animal for poultry. Poult Sci 40:651–657. Received March 14, 2000 Accepted July 17, 2000 Corresponding editor: Lyman Crittenden
