Heredity 84 (2000) 228±237 Received 25 June 1999, accepted 22 October 1999

Monogenic inheritance of apomixis in two Hieracium species with distinct developmental mechanisms

R. A. BICKNELL* , N. K. BORST & A. M. KOLTUNOWà Crop and Food Research, Private Bag 4704, Christchurch, New Zealand and àC.S.I.R.O. Division of Plant Industry, PO Box 350, Glen Osmond, South Australia 5064, Australia

Apomixis, the asexual formation of seed, has been known in angiosperms for more than a century yet the genetic mechanisms that control this trait remain poorly understood. Most members of the genus Hieracium are apomicts, forming predominantly asexual seed. Some purely sexual forms, however, also exist. In this paper we present a study of the inheritance of apomixis using two apomictic species of Hieracium which utilize very di€erent forms of megagametogenesis. In both cases the progeny inherited apomixis as a monogenic, dominant trait that could be transferred by both haploid and diploid male gametes. In common with observations from other systems, no diploid apomictic progeny were recovered from these crosses. This appears to have been caused by selection against the survival of diploid zygotes, rather than against the mediation of haploid gametes as has been noted in other systems. Crosses between the two apomicts showed that the dominant determinants in the two forms examined were closely linked, possibly allelic. The signi®cance of these data is discussed with respect to current theories on the associative link between gametophytic apomixis and polyploidy.

Keywords: apomixis, apospory, Hieracium, inheritance.

Introduction 1992) as ®rst described by Rosenberg (1906). Cytologi- cal comparisons were conducted between vegetatively Plants that reproduce by gametophytic apomixis form propagated isolates of three closely related species, an embryo directly from an egg cell in an unreduced the apomicts H. aurantiacum and H. piloselloides and a embryo sac. In contrast to sexual reproduction, the sexual accession of H. pilosella. Although development processes of meiosis and fertilization are avoided, in the sexual accession was found to be typical of the leading to the formation of genetically uniform progeny Polygonum-type, between the apomictic lines studied (Nogler, 1984a; Koltunow, 1993). Apomixis is therefore signi®cant di€erences were recorded in the number and typically de®ned by function as asexual reproduction precocity of aposporous initials formed, and in the through seed (syn. agamospermy) (Nogler, 1984a). The number, morphology, developmental cytology and fates developmental mechanisms employed by di€erent apo- of the resultant unreduced embryo sacs. It is apparent mictic species, however, are remarkably varied (Nogler, from these and other unreported data, that many of the 1984a; Koltunow et al., 1995), possibly re¯ecting the developmental details of apospory are quite variable in apparent polyphyletic origin of this trait among ¯ower- Hieracium, and that this variability is expressed between ing plants (see Carman, 1997). Known mechanisms of apospecies, between accessions of an apospecies and apomixis have been categorized in di€erent ways. even between ¯orets on an individual capitulum. Typically genera are grouped by similarities to a type- As a complement to this work, the inheritance of mechanism (Asker & Jerling, 1992). From our own apomixis was studied using these three accessions. observations in Hieracium subgenus Pilosella it is Apomixis has been described as a heritable trait in apparent that this form of characterization can be several species (Asker & Jerling, 1992). Dominant misleading (Koltunow et al., 1998). The members of this inheritance for apospory has been described in the grass taxon are reported to form asexual seed by apospory of genera Panicum (Savidan, 1981), Brachiaria (Borges do the Hieracium-type (Nogler, 1984a; Asker & Jerling, Valle et al., 1994) and Pennisetum (Dujardin & Hanna, 1985). The results of Christo€ (1942) indicate that *Correspondence. E-mail: [email protected] dominant inheritance may also apply in Hieracium

228 Ó 2000 The Genetical Society of Great Britain. APOMIXIS IN HIERACIUM 229 aurantiacum (reviewed by Nogler, 1994). Similarly, P4 and P2 appear to be self-incompatible under cool Gadella (1991) noted that apospory may be inherited winter conditions. Pollinations were conducted without as a dominant factor in H. pilosella. Nogler (1984b) prior emasculation. conducted a detailed study of the inheritance of To ensure the genetic integrity of stocks used, the apospory in Ranunculus auricomus, also concluding that plants were maintained vegetatively, either by divi- the trait was conferred by the inheritance of a single sion of mature plants or through micropropagation dominant determinant. That determinant, however, (Bicknell, 1994). The inheritance studies were conducted could only be transferred in a heterozygotic state in a in a glasshouse at Lincoln, New Zealand, maintained diploid or polyploid gamete. Nogler noted that this with a minimum night temperature of 12°C, minimum mechanism favoured the involvement of a polyploid day temperature of 18°C and maximum day tempera- parent and ensured the formation of a polyploid, ture of 25°C. Flowering was promoted by day-length apomictic zygote. extension lighting, using high pressure sodium vapour There is a very close association between gametophy- lamps with a photoperiod of 16 h (Yeung, 1989). tic apomixis and polyploidy (reviewed in Asker & Jerling, 1992). In common with most other taxa Breeding scheme for the inheritance of apomixis containing gametophytic apomicts, all naturally occur- ring apomictic forms of Hieracium have been recorded The breeding scheme used is detailed in Fig. 1. Inherit- as polyploid (Skalinska, 1970; Gadella, 1991) yet several ance studies with apomictic species are subject to some diploid sexual species are known. The recent recovery of unique constraints. As the trait results in the formation a diploid, apomictic plant of H. aurantiacum (Bicknell & of predominantly maternal seed, apomixis restricts Borst, 1997) indicates, however, that polyploidy is not access to the female gamete, the egg cell. Furthermore, an obligate requirement for the expression of apomixis when a rare hybridization event does occur in these in this taxon. plants either a reduced or an unreduced egg may be The aims of the current study were to determine the used, resulting in the formation of di€erent progeny patterns of inheritance in our experimental accessions classes and frustrating the analysis of segregation. To of H. aurantiacum and H. piloselloides, to test for any avoid these diculties crosses were designed using an prejudice against the mediation of haploid gametes in apomict as the staminate parent and a sexual plant as the transmission of dominant alleles conferring apomixis, the recipient. to test for allelism between the two apomicts tested and Initially it was hoped that by using a triploid apomict to karyotype the chromosomes of H. pilosella. as the staminate parent in a cross with a diploid, sexual recipient, the inheritance of apomixis could be studied at the diploid level. Following a cross between the Materials and methods apomictic triploid D3 and the sexual diploid P2, how- ever, most of the progeny were found to be either Plant materials triploid or aneuploid (Table 1, cross 1). The expected Two apomictic biotypes of Hieracium were used, a diploid class was only represented by three weak triploid accession (2n ˆ 27, x ˆ 9) of H. piloselloides individuals. The triploid progeny of this cross were (designated `D3') and an aneuploid accession (2n ˆ therefore chosen for further analysis. Representative 3x + 4 ˆ 31) of H. aurantiacum (designated `A3.4'). sexual and apomictic hybrid triploid progeny were

Preliminary studies using test-crosses and pollen via- backcrossed to the sexual recurrent parent and the BC1 bility stains revealed that both had a high degree of progeny analysed for ploidy and reproductive mode. pollen fertility and that both frequently produced Again, as triploids were the most common progeny class haploid pollen. They were also selected for their in the BC1 populations, they were analysed for the distinctive morphologies. The bright orange ¯owers of inheritance of apomixis. H. aurantiacum and the upright form and strap-shaped Both the apomictic triploid D3 and aneuploid A3.4 leaves of H. piloselloides provided useful marker were then crossed to the sexual tetraploid P4 and the characteristics for con®rming hybridity in the progeny. progeny analysed for ploidy (Table 1, crosses 4 and 7). A3.4 was originally obtained from an adventive pop- Triploids were recovered from both crosses in sucient ulation in Central Otago, South Island, New Zealand, numbers to permit an assessment of the inheritance of and D3 was obtained from a wild population in apomixis. Backcrosses were conducted for both the

Steiermark, Austria. Two sexual biotypes were used; a sexual and apomictic triploid progeny classes of the F1, tetraploid accession of H. pilosella from Caen, France using the sexual accession P4 as the recurrent, maternal

(designated `P4') and an anther culture-derived diploid parent. The triploid BC1 progeny were analysed for (Bicknell & Borst, 1996) of P4 (designated `P2'). Both the inheritance of apomixis. Linkage between the

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. 230 R. A. BICKNELL ET AL.

of the marker in the resistant seedlings was con®rmed by Southern analysis (data not shown). The hybrid progeny were also clearly morphologically intermediate between their dissimilar parents.

Scoring progeny for ploidy and mode of reproduction The ploidy of progeny plants was determined by chro- mosome counting using aceto-orcein staining of root-tip squash preparations as previously described (Bicknell & Borst, 1997). The karyotype of P2 is illustrated in Fig. 2. For scoring, apomixis was de®ned as the formation of an embryo(s) at petal senescence following the decapitation of an immature capitulum. Ovule structure was visual- ized by ovule clearing (see below). In apomictic biotypes of Hieracium maternal embryos arise spontaneously before the ®nal senescence of the ¯ower (stage 10 in Koltunow et al., 1998). In sexual types, prevention of fertilization by bud decapitation (Koltunow et al., 1995) leads to the retention of a quiescent mature embryo sac up until this developmental stage, so that di€erences between sexual and apomictic types could be readily determined. It should be noted that this measure is a score of parthenogenesis rather than apomixis as a whole as it does not include a measure of meiosis. In a separate study the rate of female meiosis was determined to be less than 2% in this material (to be reported separately). Three cut capitula were scored from each plant. Immature capitula were decapitated just prior to ¯oret opening (stages 5 or 6 in Koltunow et al., 1998) then allowed to progress to stage 10. The outer bracts were removed from the harvested heads and the tissue ®xed overnight in FAA (Stelly et al., 1984). The ovaries, attached to the receptacle of the capitulum, were then treated to remove oxalate inclusions and impregnated Fig. 1 The breeding scheme used in the study. (a) Utilization with methyl-salicylate as previously described (Koltunow of the sexual diploid `P2' as the pistillate parent. (b) Utilization et al., 1998). The cleared ovaries were then examined by of the sexual parent `P4' as the pistillate parent. (c) Test for di€erential interference contrast on an Olympus BH2 linkage/allelism between the AD allele of D3 and the AA allele or Leica DM R microscope and scored for the presence of A3.4. *A similar scheme was used with A3.4 as the of embryo sacs and/or embryos. At least six ovaries apomictic, staminate parent. **Proposed genotype. were examined from each capitulum. If embryos were ***Transgenic for a simplex, hemizygous copy of p35S-NPTII. detected the plant was scored as a putative apomict. If embryos were absent but normally developed embryo dominant alleles from H. piloselloides and H. aurantia- sacs were present, the plant was scored as sexual. cum (`AD' and `AA', respectively) was tested by Finally, if only degenerate structures were seen the plant hybridizing A3.4 and D3, followed by a test-cross was scored as sterile. In a small number of individuals, using P2 as the pistillate parent. To facilitate the ¯owers either failed to form or developed abnormally. recovery of rare hybrids from among the predomi- These plants were also scored as sterile. nantly maternal progeny of A3.4, the D3 parent used was ®rst transformed with a single copy of a chimeric Statistical analysis pnos-NPTII-3¢nos sequence (Bicknell & Borst, 1994). The progeny were then screened on a medium supple- For each cross, progeny were assigned to a progeny class mented with 100 mg/L kanamycin and the inheritance (apomictic, sexual or sterile) and the frequency of each

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. APOMIXIS IN HIERACIUM 231

Table 1 Crosses conducted to determine the mode of inheritance of apomixis in Hieracium

Progeny typeà Expected ratio§ Chi-squared± Cross number F1 F1 and details ploidy n Apo Sex Ster ploidy Apo Sex No Ster Sex + Ster Apo + Ster 1. P2 ´ D3 2x 3 0 2 1 2x 1/3 2/3 3x 94 57 33 4 3x 2/3 1/3 0.45 1.54 0.72 2x + i 88 2. P2 ´ PD3.apo 2x 2 0 0 2 2x 1/3 2/3 3x 30 15 9 6 3x 2/3 1/3 0.19 3.80 0.70 2x + i 16 3. P2 ´ PD3.sex 2x 0 0 0 0 2x 0/3 3/3 0.0 0.0 0.0 3x 26 0 19 7 3x 0/3 3/3 0.0 0.0 0.0 2x + i 11 4. P4 ´ D3 3x 62 18 40 4 3x 1/3 2/3 0.14 0.52 0.72 4x 15 7 6 2 4x 2/3 1/3 0.96 2.70 0.58 3x + i 22 5. P4 ´ PD3.apo 3x 48 16 28 4 3x 1/3 2/3 0.18 0.0 0.22 4x 25 14 10 1 4x 2/3 1/3 0.75 1.28 0.48 3x + i 7 6. P4 ´ PD3.sex 3x 59 0 54 5 3x 0/3 3/3 0.0 0.0 0.0 4x 12 0 7 5 4x 0/3 3/3 0.0 0.0 0.0 3x + i 7 7. P4 ´ A3.4 3x 102 29 65 8 3x 1/3 2/3 0.26 1.10 0.53 4x 13 7 4 2 4x 2/3 1/3 0.05 0.96 0.84 3x + i 20 8. P4 ´ PA3.apo 3x 27 7 17 3 3x 1/3 2/3 0.19 0.67 0.68 4x 14 7 6 1 4x 2/3 1/3 0.96 1.75 0.45 3x + i 13 9. P4 ´ PA3.sex 3x 63 0 53 10 3x 0/3 3/3 0.0 0.0 0.0 4x 18 0 12 6 4x 0/3 3/3 3x + i 31

Number of progeny at that ploidy level as determined by chromosome counts. à Apo, apomictic; sex, sexual; ster, sterile. § Expected progeny ratio if the apomictic parent had the genotype Aaa. ± Calculated for progeny groups of greater than 10 individuals. No Ster, steriles not used in analysis; Sex + Ster, steriles counted as sexuals; Ster + Apo, steriles counted as apomicts. class determined. Steriles were separately considered apomictic donor had the genotype Aaa with respect to either as having arisen independently of apomixis/ apomixis. Chi-squared values were computed for the sexuality or as representing either dysfunctional sexuals three di€erent assignments of the sterile class described or dysfunctional apomicts. In the former case they were above. omitted from the analysis entirely whereas in the latter cases they were either assigned to the sexual or to the Results apomictic progeny class, respectively. Segregation for apomixis among the F and backcross 1 Apomixis was inherited as a monogenic, (BC) progeny was examined by the chi-squared test for a dominant trait in H. piloselloides ®xed ratio hypothesis with application of the Yates's continuity correction factor to compensate for the low One hundred and eighty-®ve hybrids were recovered number of observations in some data sets. The test was when the triploid apomict D3 (2n ˆ 27) was crossed to applied against a null hypothesis that apomixis was the diploid sexual accession P2 (2n ˆ 18). Of these, only a dominant trait and that in each case the original three were diploids (2n ˆ 18) (Table 1, cross 1), two

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. 232 R. A. BICKNELL ET AL.

triploid sexual F1 plant to the sexual parent yielded only sexual progeny (Table 1, cross 3). Sexuality therefore appears to be the phenotype of the homozygous recessive genotype, and apomixis appears to be con- ferred by the inheritance of a single dominant factor.

The transfer of the dominant allele was mediated by both haploid and diploid male gametes The poor recovery of diploids from the cross between P2 and D3 clearly indicated a prejudice against the survival Fig. 2 The ideotype of the nine chromosomes of Hieracium of products following the involvement of haploid, male pilosella, compiled from more than 20 independent observa- gametes. It remained uncertain, however, whether this tions of root-tip mitotic ®gures of the sexual diploid P2. selection had occurred at the level of haploid gamete Chromosome 1 is a large submetacentric chromosome bearing viability/ecacy, or of diploid zygote survival. To a conspicuous single satellite on the short arm. This was the investigate this, a sexual tetraploid P4 (2n ˆ 36) was only satellite identi®ed in the karyotype of P2; no other used as a pistillate parent in a cross with the apomictic secondary constrictions were evident, indicating that chromo- triploid D3 (Table 1, cross 4). The diploid P2 was some 1 bears the sole nucleolus organizer region in this plant. originally derived by anther culture from P4 (Bicknell & Chromosome 2 is a large metacentric and chromosome 3 is subtelocentric. Chromosomes 4 and 5 are submetacentric, Borst, 1996), so the P2 genome represented a subset of whereas chromosomes 6, 7, 8 and 9 are small metacentrics. P4. Triploid F1 progeny recovered from this cross No indications of B chromosomes were observed. (P4 ´ D3) were expected to have arisen from the union of a diploid egg with a haploid sperm nucleus, whereas tetraploids would have arisen from the involvement of of which were sexual and one sterile. Eighty-eight a diploid sperm nucleus. A total of 99 progeny were aneuploid progeny were recovered, ranging in chromo- recovered, ranging in ploidy from 3x to 4x (Table 1, some number from 19 (2x + 1) to 26 (2x + 8), with cross 4). Sixty-two of the F1 plants were triploids the most common classes being 23 (2x + 5) and 24 (2n ˆ 27) and 15 were tetraploids (2n ˆ 36). Apomixis

(2x + 6) chromosomes (data not shown). Ninety-four segregated among both the triploid and tetraploid F1 triploids (2n ˆ 27) were recovered. For the statistical progeny. Segregation, tested by the chi-squared test, analysis it was assumed that the triploid progeny was found to be consistent with the expectations of resulted from the union of a haploid gamete from the a monogenic, dominant model of inheritance. It was pistillate parent, with a diploid gamete from the concluded that haploid gametes could transmit the staminate parent. As no plants were obtained with a dominant allele conferring apomixis in this system. chromosome number greater than 27, it appears that Backcrosses were conducted to test for dominant completely unreduced gametes did not contribute to the inheritance. Triploid apomictic and sexual F1 segregants formation of this population. The statistical analysis were crossed to the sexual tetraploid parent P4 (Table 1, indicates the data are consistent with the hypothesized crosses 5 and 6). The apomictic F1 backcross (Table 1, genotype. Assignment of the steriles into either the cross 5) resulted in the formation of a segregating sexual or apomictic classes did not lead to the rejection population, with ratios of apomictic and sexual biotypes of the null hypothesis. similar to the F1 (Table 1, cross 4). Conversely, back- To determine which F1 segregant class represented the crossing a sexual F1 plant to the sexual parent resulted homozygous recessive phenotype, backcrosses were in the formation of only sexual and sterile progeny conducted using representative triploid apomictic and (Table 1, cross 6). sexual F1 segregants, crossed back to the sexual parent P2 (Table 1, crosses 2 and 3). Because the apomictic Apomixis was also inherited as a monogenic, parent D3 formed almost entirely maternal seed, it was dominant trait in H. aurantiacum not practical to conduct the reciprocal backcrosses between it and the F1 segregants. Backcrossing one of Our cytological investigations of the apomicts, the apomictic F1 triploids to its sexual parent yielded 48 H. piloselloides (D3) and H. aurantiacum (A3.4) revealed progeny. The two diploids recovered from this cross that in these species distinct di€erences were apparent were both sterile, whereas the 30 triploid BC1 plants in the timing and degree of aposporous embryo sac recovered segregated for apomixis in a manner similar to formation. It was therefore of interest to test whether the F1 population (Table 1, cross 2). Backcrossing a the inheritance of apomixis from A3.4 was similar to

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. APOMIXIS IN HIERACIUM 233 that recorded for D3. One hundred and thirty-®ve plants this plant system. To test for allelism it was necessary were recovered from a cross between the sexual to combine the two dominant alleles in a single plant tetraploid P4 and the apomict A3.4 (Table 1, cross 7). through hybridization, requiring that one of the Of these, 102 were recorded as triploids. Segregation for apomicts be used as a recipient parent. To facilitate apomixis among the triploid progeny was consistent the recovery of hybrids a hemizygous, transgenic D3 with a monogenic, dominant pattern of inheritance, parent was used (see Methods). Ten transgenic seedlings assuming the union of a diploid egg with a haploid were recovered after the pollination of 30 capitula of sperm nucleus. Results from the backcrosses con®rmed A3.4 with transgenic pollen of D3. The number of previous ®ndings. As observed with D3, apomixis seedlings screened was estimated at 1550, indicating a segregated when a triploid apomictic F1 was back- recovery rate of 0.64% and a fertilization rate of 1.3%. crossed to the sexual parent (Table 1, cross 8), but not Of the 10 hybrids, four were tetraploids, four were when a triploid sexual F1 was backcrossed (Table 1, triploids and two were aneuploids (3x + 4 and 3x + 5). cross 9). The results indicate that the apomict A3.4 also All were either apomictic or sterile. Most of the carries a single dominant allele for apomixis and that the apomicts expressed the trait at only a low level, forming allele could be transmitted in either a haploid or diploid only a small number of asexual seeds on each capitulum. male gamete. A3.4 is an aneuploid carrying four copies Two of the tetraploids, however, set abundant germin- of four chromosomes and three of the remaining ®ve. able seed after the decapitation of immature buds. The Chi-squared tests indicated that the trait was monogenic best of these plants, designated `AD4', was test-crossed in this plant; however, sample sizes are inadequate to to the sexual diploid P2 and the triploid progeny scored determine between trisomic (Aaa) and tetrasomic (Aaaa) for apomixis. One hundred and sixty progeny were inheritance. It is therefore not possible to use these data recovered. Table 2 summarizes the results of chi- to assign the Apo locus to a given chromosome, or squared tests conducted for three proposed models of chromosome set in this plant. the genotype of AD4, with the three possible assign- With the exceptions of the triploid progeny of ments of the sterile class, assuming monogenic domi-

P4 ´ PD3.apo (Table 1, cross 5) and the tetraploid nance for both the AA and AD alleles. Given the ®nding progeny of P4 ´ PA3.apo (Table 1, cross 8) the treat- described above, that sterility does not appear to be ment of sterility as an event independent of apomixis/ preferentially associated with either apomixis or sexual- sexuality provided the best goodness of ®t for the data ity, close linkage, possibly allelism, is indicated as the under the experimental assumptions. It was therefore most probable genetic model. concluded that sterility segregated as an independent trait in these experimental populations. Discussion

The dominant factors of A3.4 and D3 Apospory in Hieracium is inherited as a are closely linked dominant monogenic trait As both A3.4 and D3 could be used to transmit a Apospory in H. piloselloides and H. aurantiacum was dominant factor conferring apomixis, we wished to test inherited as a monogenic, dominant trait. This agrees whether they acted at the same, or at di€erent loci in with the ®ndings of Gadella (1991) with H. pilosella and

Table 2 Allelism test between Hieracium aurantiacum and H. piloselloides

Progeny observed Predicted Chi-squared Models for progeny ratio Test-cross Apo Sex Ster AD4 genotype Apo:Sex No Ster Sex + Ster Apo + Ster P2 ´ AD4.apo 123 28 9 Single allele 1:1 58.5 46.2 67.6 (ADaaa or AAaaa) Unlinked alleles 3:1 3.02 0.30 4.80 (ADaaa or BAbbb) Close linkage/allelism 5:1 0.085 4.76 0.76 (ADAAaa) Apo, apomictic; Sex, sexual; Ster, sterile. No Ster, steriles not used in analysis; Sex + Ster, steriles counted as sexuals; Ster + Apo, steriles counted as apomicts.

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. 234 R. A. BICKNELL ET AL. with similar ®ndings reported for the aposporic taxa tions, despite occasional hybridization events. More Panicum (Savidan 1981), Pennisetum (Ozias-Akins recently, Grimanelli et al. (1998) recorded a similar et al., 1993), Ranunculus (Nogler, 1984b) and Brachiaria pattern of inheritance for the transfer of diplospory (Borges do Valle et al., 1994), and for the diplosporous from Tripsacum dactyloides into Tripsacum/maize species Tripsacum dactyloides (Leblanc et al., 1995). It is hybrids. They further noted, however, that although a also consistent with the ®nding that sexual biotypes profound restriction against the simplex transmission of could be recovered following the anther culture of an the dominant factor for apomixis was observed in the apomictic pentaploid of H. pilosella (Bicknell & Borst, BC1 generation it was not apparent in the subsequent 1996). Both D3 and A3.4 were simplex for the dominant BC3 generation. allele at the Apo locus. Typically, seed formation in ®eld- By contrast, Mogie (1988, 1992) has proposed that adapted facultative apomicts is largely maternal; very apomixis is conferred by the inheritance of incompletely little meiotic seed is produced. As apomixis signi®cantly penetrant alleles. In polyploid biotypes increased allelic reduces the availability of female gametes for hybrid- dosage would permit the expression of the trait, whereas ization, hybridization events in mixed populations of the presence of those alleles would be masked in such apospecies are expected usually to involve a sexual diploids. Carman (1997) conducted a comprehensive biotype as the pistillate parent. Sexuality is the homo- assessment of associations between apomixis, polyspory, zygous, recessive phenotype, so this mechanism will lead polyembryony, base chromosome number and several to the dispersion of dominant alleles. The identi®cation forms of polyploidy. He concluded that apomixis arises of two polyploids each bearing a single dominant allele from interactions between asynchronously controlled at Apo therefore is consistent with the expectation that regulatory elements inherited from dissimilar parents. genotypes bearing more than one dominant allele should This mechanism would require the inheritance of intact, arise only rarely in this taxon. or near-intact, regulatory cascades located throughout the genomes of both parents. Carman hypothesized that this would probably act against the mediation of Diploid apomixis was prevented by selection haploid gametes, and therefore ensure that apomictic against diploid zygotes progeny were polyploid. The proposals of Nogler and Gametophytic apomixis has been recorded in more than Carman, outlined above, predict that selection will 400 species of ¯owering plant, from 35 diversely aligned occur at the level of the gamete. Haploid gametes families (Carman, 1997). Despite this taxonomic diver- carrying either a dominant determinant for apomixis or sity, however, with very few exceptions gametophytic only a partial series of regulatory elements are expected apomicts have been recorded as polyploids (Asker & either to fail to survive or to fail to transfer the entire Jerling, 1992; Carman, 1997). In the small number of trait. In Hieracium, however, this interpretation does cases when diploid apomicts have been recorded they not apply. In the current study both haploid and have typically been weak experimentally derived anom- diploid male gametes were successfully used to mediate alies (Nogler, 1982; Leblanc et al., 1996; Bicknell & the transfer of the AD allele for apospory from Borst, 1997) in species predominantly composed of H. piloselloides and the AA allele from H. aurantiacum. polyploid races. The introduction of apomixis into crops Furthermore, no evidence of segregation distortion was is an important goal in apomixis research (Hanna, seen among the progeny that would implicate the action 1995), yet the majority of crops are diploids or stable of gamete-level selection. allopolyploids, the types of species in which gameto- Crosses between the diploid sexual plant P2 and three phytic apomixis is almost unknown. Before the objective di€erent triploid pollen donors (D3, PD3apo and PD3sex) of transferring apomixis can be achieved, therefore, the yielded only ®ve diploids among the 270 progeny causes and consequences of this fundamental associa- assessed. Of those ®ve, three were clearly sterile and all tion need to be more fully understood. grew very poorly. Furthermore, as the criterion used for During his pioneering work with the aposporous assigning sexuality was based on the presence of a apomict Ranunculus auricomus, Nogler (1984b) noted typical quiescent gametophyte at ¯oral senescence, it that A+, the dominant determinant that conferred was unclear whether the two `sexual' diploids identi®ed apomixis, could not be transferred in a homozygous would have been capable of completing sexual seed state. Homozygosity of A+ was associated with gamete formation. By contrast, 115 of the 270 progeny assessed lethality, so A+ could only be transmitted by a diploid were aneuploids and 150 were triploids. When a or polyploid gamete. Nogler noted that similar mecha- tetraploid sexual parent was used in crosses with the nisms of gamete-level selection against homozygous apomicts D3 and A3.4, triploids were frequently recov- dominant alleles could explain both the establishment of ered among the progeny. This indicates that many of polyploidy and also its retention in apomictic popula- the haploid male gametes of both D3 and A3.4 were

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. APOMIXIS IN HIERACIUM 235 functional and their products viable. As apomicts were was signi®cantly enhanced by fertilization with a diploid well represented among the triploid progeny of these gamete from the other parent. This requirement for crosses there appears to have been no diculty in fertilization by a diploid gamete is consistent with transferring the trait via these gametes. It is unlikely that the dominance model of heterosis (Ledig, 1986) which the poor recovery of diploids was caused by an predicts that a high load in a haploid gamete would be inviability of reduced P2 egg cells. Their reduced state better masked by fertilization with a diploid, than with a and ecacy were clearly demonstrated by the recovery haploid gamete from a highly heterozygotic partner. of other ploidy classes from these crosses, from crosses The use of an anther-culture derived sexual diploid with AD4, and also by the segregation patterns for (Bicknell & Borst, 1996) may have further ampli®ed ploidy and apomixis seen in the resulting hybrids. this e€ect. P2 was derived from P4, a tetraploid, In Hieracium, therefore, the formation of apomictic self-incompatible, obligate outcrossing accession of diploids was discouraged through selection against H. pilosella. Both polyploidy and outcrossing can mask diploid hybrids, acting after fertilization, rather than deleterious, recessive mutations (Richards, 1997). A through gamete-level selection acting prior to this event. further indication of the deleterious genetic load carried Furthermore, the nonrecovery of apomictic diploids by these plants was demonstrated by an apparent appears to be attributable to the poor viability of all intolerance to inbreeding. Lower seed-set was routinely the diploid products of the polyploid apomictic parents, seen after the backcrossing of F1 segregants to their rather than to any speci®c prejudice against the sexual parent (data not shown), and the percentage of apomictic progeny class. sterile BC1 individuals was always greater than that seen Mogie's model (1988) predicts selection following among the corresponding F1 population (Table 1). fertilization and would therefore be consistent with the Conversely, F1 hybrids were typically more vigorous observations of inheritance in Hieracium. It also pre- than either of their parents. dicts, however, that the diploid genotype Aa would not As a form of asexual reproduction, apomixis leads to express apomixis at an identi®able level. In a previous the sequential accumulation of deleterious mutations, a study (Bicknell & Borst, 1997) a diploid, apomictic form phenomenon often referred to as `Mullers Ratchet' of H. aurantiacum was recovered from an experimental (Muller, 1964). In sexual diploid species, deleterious population. That plant, designated A2, was derived recessive mutations may be exposed to selective pres- from A3.4, a genotype used in this study. As A2 appears sures as a result of recombination at fertilization, but to have formed by dihaploidy it is expected to have the probably more critically, also during haploid gameto- genotype Aa at the Apo locus. Aposporous initials in A2 phyte formation and function (the `Clean Egg Hypothe- were shown to have been even more numerous than sis'; see Richards, 1997). In gametophytic apomicts, in A3.4, often causing ovule abortion because of the however, the megagametophyte is unreduced. Only unrestrained competitive proliferation of multiple asex- microgametogenesis and pollen function, when these ual structures. In contrast to the predictions of Mogie's events are still operative, are subject to this form of model, the genotype Aa appears to express apomictic gamete-level selection, which may be why reduced pollen tendencies at least as strongly as the genotype from continues to be formed by many of the autonomous which it was derived by reduction. apomicts of Hieracium.

Reduced diploid hybrid viability may be primarily The establishment of polyploid apomixis caused by a high genetic load in apomictic Because polyploidy increases the number of alleles Hieracium present in a genotype it provides a mechanism for The poor recovery of diploids in this study may simply improving the tolerance of a ®xed genotype to the be a re¯ection of the deleterious genetic loads carried by accumulation of recessive load. The masking of del- the accessions of Hieracium studied. When a diploid eterious recessives will be particularly pronounced where sexual plant was used as the pistillate parent, vigorous polyploidy is established through cross-hybridization hybrid progeny were recovered following fertilization rather than by endoreduplication or self-fertilization. It with a diploid gamete from an apomict, but not is interesting to note that hybridization, particularly BIII following fertilization with a haploid gamete. In con- hybridization, the fertilization of an unreduced egg cell trast, when a tetraploid sexual was used as the pistillate by pollen of another genotype, has been documented in parent, seedlings arising from the mediation of a haploid several apomictic plants, including Hieracium (Skalinska, gamete from an apomict were commonly recovered. 1973; Gadella, 1991). An apparent hybrid origin is also a With respect to both the sexual parent P2 and the feature of many apomictic groups (Asker & Jerling, apomicts D3 and A3.4, therefore, haploid gamete use 1992) indicating that it has often been an important

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. 236 R. A. BICKNELL ET AL. in¯uence in determining the composition of apomictic many of the di€erences observed in the timing and taxa. By providing a mechanism for increasing the frequency of initial cell di€erentiation, and in the tolerance of a genotype to mutation, however, poly- general variety of subsequent developmental events ploidy can also be expected to lead to the accumulation (Koltunow et al., 1998). The inheritance data do not of a greater mutational load in these plants. Song et al. exclude, however, the potential in¯uence of physiolog- (1995) demonstrated that a genetic distance of between ical variation, environmental e€ects or other genetic loci

2.5% and 9.6% developed between a population of F5 modifying the impact of a dominant determinant at the plants and their F2 homozygous ancestors following Apo locus. The modulating e€ects of alleles at modi®er four generations of sel®ng. In the absence of the gamete- loci were indicated throughout this study by clear level selection mechanism discussed above, autonomous di€erences in the number and morphologies of embryos apomicts like Hieracium may accumulate mutations at observed amongst the sibling progeny of the crosses an even greater rate. Furthermore, Song et al. (1995) described. Furthermore, none of the apomictic progeny noted that wide hybridization appeared to lead to a examined in these experiments exhibited the high level of prejudice against paternally inherited alleles, possibly apomixis observed in the parents (to be published because of incompatibility between paternal nuclear and elsewhere). In a successful apomict, it is necessary but maternally inherited cytoplasmic elements. Wide hy- not sucient that the events of meiosis, fertilization and bridization in apomicts would create similar circum- parthenogenesis be uniquely regulated. They must also stances and can therefore be expected to exert similar be co-ordinated in association with an array of other selection pressures for gene silencing. developmental functions. This was clearly demonstrated In summary, the transition from diploidy to poly- in the diploid apomict A2 (Bicknell & Borst, 1997) in ploidy in apomictic Hieracium would confer advantages which aposporous initial di€erentiation was proli®c. associated with masking mutations accumulated by Seed germination rates, however, were very low because the action of Mullers ratchet. Conversely, however, the of the frequent, premature degeneration of the ovule institution of polyploidy would encourage the continued caused by the actions of multiple, competing asexual accumulation of mutations and reduce the competitive- structures within. An important remaining question ness of any diploids that may subsequently arise. This therefore is what in¯uence do the modi®er loci have on process would rapidly favour the establishment of the expression of the apomictic locus in this plant? For polyploid apomixis yet discourage any reversion to example, do they in¯uence spatial and temporal events diploidy. As an apomictic taxon ages, the degree of of apomixis in these plants, or the frequency and deleterious genomic modi®cation can be expected to penetrance of the trait in ovules, and is their presence increase and the impacts of this e€ect would become essential to promote high yields of viable seed? In more pronounced. Haploid gamete lethality would agricultural systems, seed number and viability are follow, because of the exposure of recessive lethal alleles frequently important determinants of overall yield. The located throughout the genome during periods of transfer of apomixis in an economic format is therefore haploid gametophyte formation and function. When likely to require changes in the expression of several coupling phase linkage becomes established between a genes. The genetic information and genotyped stocks lethal recessive allele and an allele for apomixis, derived from this study are now being used to conduct Nogler's hypothesis of linked lethality would apply. genetic and developmental analyses of the impacts of Grimanelli et al. (1998) noted that the prevention of apomixis-modi®er expression in Hieracium. simplex transmission for a dominant apomixis-factor from alloploid BC Tripsacum/maize plants was not 1 Acknowledgements apparent by the subsequent BC4 generation. They attributed this transmission ratio distortion to the operation The authors wish to thank the sta€ of the Jardin of a trans-acting incompletely penetrant system. From Botanique, Caen, France; Jardin Alpin Florealpe, our observations with Hieracium it appears that the Aubert, Switzerland; and Dr M. Douglas, Alexandra, phenomenon of transmission distortion should be more New Zealand for providing the plant material used. We generally de®ned as resulting from multiple cis- and also wish to acknowledge the e€orts of R. Braun and trans-acting lethal factors of varying penetrance, located S. Lambie in assessing the ploidy and breeding type of throughout the ®xed genome of an apomict. the plants studied. Our special thanks to Gian Nogler and Michiel Willemse for their interest and discussions during the course of this work. This work was supported Modulators of the apomictic locus by grants from the New Zealand Foundation for Molecular di€erences in the Apo locus between Research Science and Technology, the Australian H. piloselloides and H. aurantiacum may account for Research Council and CSIRO Australia.

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237. APOMIXIS IN HIERACIUM 237

References LEBLANC, O., GRIMANELLI, D., ISAM-ISAM-FARIDIFARIDI, N., BERTHAUD, J. ANDSAVIDAN , Y. 1996. Reproductive behavior in maize± ASKER, S. E.AND JERLING , L. 1992. Apomixis in Plants. CRC Tripsacum polyhaploid plants: Implications for the transfer Press, Boca Raton, FL. of apomixis into maize. J. Hered., 87, 108±111. BICKNELL, R. A. 1994. Micropropagation of Hieracium LEDIG, F. T. 1986. Heterozygosity, heterosis, and ®tness in aurantiacum. Plant Cell Tissue Organ Cult., 37, 197±199. outbreeding plants. In: Soule , M. E. (ed.) Conservation BICKNELL, R. A.AND BORST , N. K. 1994. Agrobacterium-mediated Biology: The Science of Scarcity and Diversity, pp. 77±104. transformation of Hieracium aurantiacum. Int. J. Plant Sci., Sinauer Associates, Sunderland, MA. 155, 467±470. MOGIE, M. 1988. A model for the evolution and control of BICKNELL, R. A.AND BORST , N. K. 1996. Isolation of reduced generative apomixis. Biol. J. Linn. Soc., 35, 127±153. genotypes of Hieracium pilosella using anther culture. Plant MOGIE, M. 1992. The Evolution of Asexual Reproduction in Cell Tissue Organ Cult., 45, 37±41. Plants. Chapman & Hall, London. BICKNELL, R. A.AND BORST , N. K. 1997. Isolation of a diploid, MULLER,MULLER,H.J.H. J. 1964. The relation of recombination to muta- apomictic plant of Hieracium aurantiacum. Sex. Plant tional advance. Mutat. Res., 1, 2±9. Repro., 10, 168±172. NOGLER, G. A. 1982. How to obtain diploid apomictic BORGESDO VALLE , C., GLIENKE, C.AND LEGUIZAMON , G. O. C. Ranunculus auricomus plants not found in the wild state. 1994. Inheritance of apomixis in Brachiaria, a tropical Bot. Helvet., 92, 13±22. forage grass. Apomixis Newsl., 7, 42±43. NOGLER, G. A. 1984a. Gametophytic apomixis. In: Johri, B. M. CARMAN, J. G. 1997. Asynchronous expression of duplicate (ed.) Embryology of Angiosperms, pp. 475±518. Springer, genes in angiosperms may cause apomixis, bispory, Berlin. tetraspory, and polyembryony. Biol. J. Linn. Soc., 61, NOGLER, G. A. 1984b. Genetics of apospory in apomictic 51±94. Ranunculus auricomus. V. Conclusion. Bot. Helvet., 94, CHRISTOFF, M. 1942. Die genetische Grundlage der apomik- 411±422. tischen Fortp¯anzung bei Hieracium aurantiacum L. Z. NOGLER, G. A. 1994. Genetics of gametophytic apomixis Ð Indukt. Abstamm. Vererb. Lehre, 80, 103±125. a historical sketch. Polish Bot. Stud., 8, 5±11. DUJARDIN, M.AND HANNA , W. W. 1985. Cytology and repro- OZIAS-OZIAS-AKINSAKINS, P., LUBBERS, E. L., HANNA, W. W.AND M cNAY, J. W. duction of reciprocal backcrosses between pearl millet and 1993. Transmission of the apomictic mode of reproduction sexual and apomictic hybrids of pearl millet ´ Pennisetum in Pennisetum: co-inheritance of the trait and molecular squamulatum. Crop Sci., 25, 55±62. markers. Theor. Appl. Genet., 85, 632±638. GADELLA, T. W. J. 1991. Variation, hybridization and reproduc- RICHARDS, A. J. 1997. Plant Breeding Systems, 2nd edn. Stanley tive biology of Hieracium pilosella L. Proc. Kon. Ned. Akad. Thornes, Cheltenham, UK. V. Wetensch., 94, 455±488. ROSENBERG, O. 1906. UÈ ber die Embryobildung in der Gattung GRIMANELLI, D., LEBLANC, O., ESPINOSA, E., PEROTTI, E., Hieracium. Ber. Dtsch. Bot. Ges., 24, 157±161. GONZALEZDE LEON, D.AND SAVIDAN , Y. 1998. Non-Mende- SAVIDAN, Y. 1981. Genetics and utilization of apomixis for the lian transmission of apomixis in maize±Tripsacum hybrids improvement of Guinea grass (Panicum maximum Jacq.). caused by a transmission ratio distortion. Heredity, 80, Proc. XIV. Int. Grassl. Cong., Lexington, KY, p. 182. 40±47. SKALINSKA, M. 1970. Further cytological studies in natural HANNA, W. W. 1995. Use of apomixis in cultivar development. populations of Hieracium aurantiacum L. Acta. Biol. Crac. Adv. Agron., 54, 333±350. Ser. Bot., 13, 111±118. KOLTUNOW, A. M. 1993. Apomixis: Embryo sacs and embryos SKALINSKA, M. 1973. Further studies in facultative apomixis of formed without meiosis or fertilization in ovules. Pl. Cell., 5, Hieracium aurantiacum L. Acta Biol. Crac. Ser. Bot., 16, 1425±1437. 121±137. KOLTUNOW, A. M., BICKNELL, R. A.AND CHAUDHURY , A. M. 1995. SONG, K., LU, P., TANG, K.AND OSBORN , T. 1995. Rapid genome Apomixis: Molecular strategies for the generation of change in synthetic polyploids of Brassica and its implica- genetically identical seeds without fertilization. Pl. Physiol., tions for polyploid evolution. Proc. Natl. Acad. Sci. U.S.A., 108, 1345±1352. 92, 7719±7723. KOLTUNOW, A. M., JOHNSON, S. D.AND BICKNELL , R. A. 1998. STELLY, D. M. P., PELOQUIN, S. J., PALMER, R. G.AND CRANE , C. F. Sexual and apomictic development in Hieracium. Sex. Plant 1984. Mayer's hemalum-methyl salicylate: a stain-clearing Repro., 11, 213±230. technique for observations within whole ovules. Stain LEBLANC, O., GRIMANELLI, D., GONZALEZDE LEON, D.AND Technol., 59, 155±161. SAVIDAN, Y. 1995. Detection of the apomictic mode of YEUNG, E. C. 1989. Hieracium. In: Helevy, A. (ed.) CRC reproduction in maize-Tripsacum hybrids using maize RFLP Handbook of Flowering, vol. 6, pp. 361±364. CRC Press, markers. Theor. Appl. Genet., 90, 1198±1203. Boca Raton, FL.

Ó The Genetical Society of Great Britain, Heredity, 84, 228±237.