• apomixis;
  • Boechera (Brassicaceae);
  • interspecific hybridization;
  • polyploidy;
  • genome painting;
  • heterochromatic chromosome


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • • 
    Understanding apomixis (asexual reproduction through seeds) is of great interest to both plant breeders and evolutionary biologists. The genus Boechera is an excellent system for studying apomixis because of its close relationship to Arabidopsis, the occurrence of apomixis at the diploid level, and its potentially simple inheritance by transmission of a heterochromatic (Het) chromosome.
  • • 
    Diploid sexual Boechera stricta and diploid apomictic Boechera divaricarpa (carrying a Het chromosome) were crossed. Flow cytometry, karyotype analysis, genomic in situ hybridization, pollen staining and seed-production measurements were used to analyse the parents and resulting F1, F2 and selected F3 and test-cross (TC) generations.
  • • 
    The F1 plant was a low-fertility triploid that produced a swarm of aneuploid and polyploid F2 progeny. Two of the F2 plants were fertile near-tetraploids, and analysis of their F3 and TC progeny revealed that they were sexual and genomically stabilized.
  • • 
    The apomictic phenotype was not transmitted by genetic crossing as a single dominant locus on the Het chromosome, suggesting a complex genetic control of apomixis that has implications for future genetic and evolutionary analyses in this group.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The study of apomixis (asexual reproduction through seeds) is of great interest for both practical and theoretical reasons. The harnessing of apomixis for agricultural purposes could be of great economic and humanitarian benefit, because it would enable the propagation of hybrid genotypes indefinitely (Spillane et al., 2004). For evolutionary biologists, apomictic systems allow for testing of hypotheses of the persistence of sexual reproduction despite its twofold cost (Charlesworth, 1990; West et al., 1999). Although biologists have long grappled with apomictic mechanisms, their control is still relatively poorly understood at the genetic and molecular levels (Bicknell & Koltunow, 2004). It has been hypothesized that apomixis is directly derived from normal sexual pathways (Nogler, 1984a; Holsinger, 2000; Koltunow & Grossniklaus, 2003). This bypassing of normal sexuality may have resulted from gene-expression changes during polyploidization and/or hybridization, as most apomictic species are allopolyploids (Carman, 1997). A number of model species have been well developed for the study of apomixis, each with their own advantages and disadvantages (Bicknell & Koltunow, 2004). Genetic studies with these systems suggest that the trait is under the control of either two or several genes (van Dijk et al., 1999; Noyes & Rieseberg, 2000; Albertini et al., 2001; Matzk et al., 2005), or by a single dominant locus (Leblanc et al., 1995; Bicknell et al., 2000), which can be located on supernumerary or hemizygous chromatin (Ozias-Akins et al., 1998; Roche et al., 2001; Sharbel et al., 2004).

In comparison with other apomictic complexes, apomictic reproduction in Boechera has several factors making it more akin to normal sexual reproduction. These factors combine to make Boechera a compelling and tractable system for genetic studies of the control of apomixis. Perhaps the most promising and rare characteristic of Boechera is that apomixis can occur at the diploid level (Böcher, 1951, 1969; Sharbel & Mitchell-Olds, 2001). Evidence of diploid apomictic Boechera lines has been gathered through cytological investigations (Böcher, 1951; Naumova et al., 2001); the occurrence of unreduced gametes and heterozygous genotypes (Dobešet al., 2004b; Sharbel et al., 2004, 2005); and the persistence of heterozygosity in progeny (Roy, 1995; Schranz et al., 2005). Additionally, these diploid lines often contain heterochromatic (Het) and/or supernumerary chromosomes that could be responsible for the apomictic phenotype (Böcher, 1954; Sharbel et al., 2004; Kantama, 2005; Sharbel et al., 2005). Another major advantage is the close relationship of Boechera to the model plant Arabidopsis thaliana (Koch et al., 2001). Boechera is the only documented case of natural apomixis in the Brassica family (Brassicaceae). In addition to A. thaliana having its complete genome sequenced, there is also an unrivalled understanding of normal sexual reproduction and a variety of apomixis-like mutations that have been identified (Koltunow & Grossniklaus, 2003), which should provide an excellent framework for comparison.

Several aspects of the Boechera breeding system simplify genetic and molecular investigations, compared with other apomictic lineages. First, apomictic Boechera accessions have probably evolved from self-compatible and highly self-fertilizing sexual types (Roy, 1995), whereas most other apomictic groups are derived from self-incompatible and out-crossing taxa (Asker & Jerling, 1992). Second, as in normal sexual taxa, the embryo is derived from the megaspore mother cell (MMC). The MMC generally enters meiosis I, but fails to complete the reductional phase (apomeiosis), and then undergoes normal meiosis II to form a nonreduced restitution nucleus (Taraxacum-type diplospory) (Böcher, 1951; Naumova et al., 2001; Taskin et al., 2004). In other forms of apomixis, the embryo either develops from the nucellus (apospory) or else forgoes meiosis (Antennaria-type diplospory) (Crane, 2001). Third, Boechera apomicts are pseudogamous, meaning that fertilization of the central cell is still required for normal endosperm development (Böcher, 1951; Naumova et al., 2001; Taskin et al., 2004). Pseudogamy is usually only found in conjunction with apospory (Richards, 1986; Asker & Jerling, 1992), and typically occurs in lineages that descend from outbreeding taxa (Mogie, 1992). Finally, apomixis in Boechera is incomplete (facultative apomixis), with sexual reproduction still occurring at an often high frequency (Böcher, 1951; Schranz et al., 2005). Both the formation of viable pollen and facultative apomixis allow for potential hybridization with sexual Boechera lineages.

The ability to cross sexual and asexual lineages of Boechera is critical for the establishment of a genetically tractable research system. Many of the major advances in our understanding of the genetic control of apomixis have come from analysis of the progeny from such crosses (Ozias-Akins et al., 1998; van Dijk et al., 1999; Noyes & Rieseberg, 2000). However, there are several important caveats for analysis of the inheritance of apomixis from such studies, including the problems of making interploidy and/or interspecific crosses (discussed by Bicknell & Koltunow, 2004; Matzk et al., 2005). In Boechera we can partially mitigate these complications. The presence of diploid apomicts means that we can avoid making interploidy crosses (Schranz et al., 2005). Also, a robust phylogenetic hypothesis exists for the relationship of sexual and apomictic lineages, with direct evidence of past hybridization events (Sharbel & Mitchell-Olds, 2001; Koch et al., 2003; Dobešet al., 2004a, 2004b; Schranz et al., 2005). Specifically, the sexual species Boechera stricta is almost exclusively diploid, sexual, and forms a well supported monophyletic clade (lineage II of Dobešet al., 2004b; Schranz et al., 2005). Boechera holboellii is diploid or triploid, often apomictic (Sharbel et al., 2004, 2005), paraphyletic (I. A. Al-Shehbaz, pers. comm.), and is scattered between two other lineages with other species of Boechera (lineages I and III of Dobešet al., 2004b; Schranz et al., 2005). The interspecific hybrid species Boechera divaricarpa has repeated independent origins by hybridization of B. stricta and B. holboellii-like plants (Koch et al., 2003; Dobešet al., 2004a, 2004b). These B. divaricarpa lineages are diploid or triploid, and are often apomictic (Schranz et al., 2005). Thus apomictic lineages of B. divaricarpa are known to have a mixed dosage of B. stricta and B. holboellii-like chromosomes. By making crosses between diploid B. stricta and B. divaricarpa, we are mirroring the evolutionary steps that have probably happened within the group.

We reported previously on reciprocal crosses between a sexual B. stricta line (SAD12 = ES6) and a number of diploid apomictic B. divaricarpa lines, including individuals from the Vipond Park site (VP9 = ES9) (Schranz et al., 2005). Analysis of the reciprocal crosses of B. stricta SAD12 and B. divaricarpa VP9 found that all F1 progeny were diploid and apomictically derived when VP9 was the maternal parent, whereas all F1 progeny were triploid and sexually derived when SAD12 was the maternal parent (Schranz et al., 2005).

In this work, we continue our genetic analysis of these genotypes. First, we provide additional information about the diploid apomictic parent (VP9), including analysis of male meiosis, pollen analysis showing nonreduced gamete production, cytology documenting the presence of a heterochromatic (Het) chromosome, microsatellite analysis of progeny derived from self-pollination demonstrating persistent heterozygosity, and genomic in situ hybridization (GISH) analysis showing its hybrid origin. From the cross we analysed the F1, F2 progeny, selected F3 and test-cross (TC) progeny using a combination of flow cytometry, karyotype analysis, genome painting (GISH), and seed and pollen grain analyses. In particular, we were interested in testing the hypothesis that the Het chromosome would act as a dominant locus and confer the apomictic phenotype on transmission. Considering Boechera's great potential for studies of apomixis, this study serves as a first genetic investigation into the control and transmission of apomixis in the group, and has important consequences for our understanding of the evolutionary history of apomixis and hybridization.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Plant materials

Analyses of reciprocal crosses made between a common sexual diploid B. stricta (Graham) Al-Shehbaz tester (SAD12) and a number of Boechera lines, including apomictic diploid B. divaricarpa (A. Nelson) A. Löve & D. Löve lines from the Vipond Park (VP9) site, have been described (Schranz et al., 2005). In this study we have characterized the SAD12 × VP9 cross in greater detail (Table 1; Fig. 1). We analysed the parental genotypes [SAD12 (ES 6) and VP9 (ES 9)]; two of the resulting F1 progeny (ES 116 and ES 117); 20 of their F2 progeny (including lines ES 136 and ES 137); 48 and 30 F3 progeny from lines ES 136 and ES 137, respectively; and test crosses involving ES 136 and ES 137 (Table 1; Fig. 1).

Table 1.  Information on plants from the Boechera SAD12 × VP9 population, including generation, line, chromosome number, genome size, days to flowering, seed set and pollen viability
GenerationLineChromosome numberGenome sizeaDays to floweringbSeeds per fruitcPollen viability (%)d
  • a

    Genome size presented as ratio compared with Brassica rapa standard.

  • b

    Days to flowering of nonvernalized plants measured from transfer of seedlings to soil until appearance of first open flower.

  • c

    Seed set per silique calculated by averaging number of seeds from five independent replicates of 10 siliques each.

  • d

    Pollen viability was assessed for selected lines by staining with Alexander's stain.

  • e The tester line did not flower unless vernalized, so days to flowering, seed set and pollen viability were not assessed.

  • f

    The test-cross lines did not flower following 300 d after transplant, so days to flowering, seed set and pollen viability were not assessed.

ParentalES9140.50122 77.5672.82
F1ES117210.74116  0.1424.52
F2ES116 #1nd0.67 75nd 
F2ES116 #2240.82199  0.02 
F2ES117 #1220.80 83  0.02 
F2ES117 #2230.80291  0.0040.00
F2ES117 #3 = ES136270.96102 10.6675.88
F2ES117 #4240.86129  0.16 
F2ES117 #5nd0.72113  0.6216.79
F2ES117 #6210.75135  0.04 3.43
F2ES117 #7281.02139  1.86 
F2ES117 #8230.79ndnd 
F2ES117 #9mosaic, 24, 271.12 94  0.02 
F2ES117 #10 = ES137270.98 91  9.5372.81
F2ES117 #11nd0.79182  0.14 
F2ES117 #12nd0.88206nd 
F2ES117 #13nd0.67 99  0.00 
F2ES117 #14220.76105  0.04 
F2ES117 #15210.74101  0.00 1.87
F2ES117 #16210.73200  2.6633.21
F2ES117 #17270.99 81  0.06 
F2ES117 #19nd0.91 92  0.02 
F3ES 136 #3nd1.00150  7.82 
F3ES 136 #9nd0.97113 17.0070.79
F3ES 136 #10nd0.97144  8.0276.41
F3ES 136 #16nd1.01 81 19.6267.23
F3ES 136 #24nd0.99 84 12.1464.28
F3ES 136 #30nd0.99140 46.8875.58
F3ES 136 #34nd0.99240 14.8480.51
F3ES 137 #9nd1.03145 14.4271.61
F3ES 137 #10nd0.98 98  5.6554.66
F3ES 137 #14nd1.04 73 10.1366.18
F3ES 137 #15nd0.99112 18.3068.76
F3ES 137 #16nd0.99 96  6.5573.09
TesterES 138nd1.01No floweringend 
Test crossES 136 × ES 138nd0.95No floweringfnd 
Test crossES 136 × ES 138nd0.97No floweringfnd 
Test crossES 136 × ES 138nd0.99No floweringfnd 
Test crossES 136 × ES 138nd1.01No floweringfnd 
Test crossES 136 × ES 138nd1.03No floweringfnd 
Test crossES 136 × ES 138nd0.99No floweringfnd 
Test crossES 136 × ES 138nd0.92No floweringfnd 
Test crossES 136 × ES 138nd1.04No floweringfnd 
Test crossES 137 × ES 138nd0.96No floweringfnd 
Test crossES 137 × ES 138nd0.97No floweringfnd 

Figure 1. Pedigree and development of the Boechera SAD 12 × VP9 population.

Download figure to PowerPoint

For the test crosses, the ES 136 and ES 137 lines were used as maternal parents, and a naturally occurring tetraploid B. stricta (ES 138) as the paternal pollen-donating parent. The ES 138 line of B. stricta was collected from the Quebec–Labrador border in Canada, on the north shore of the Strait of Belle Isle, by John E. Maunder and Nathalie Djan-Chékar of the Provincial Museum of Newfoundland (collection number NDC 99–348, Stn NDC Bor. 3; accession number NFM 3363).

Plant growth conditions

Seeds were germinated as described by Schranz et al. (2005) and transplanted into 11 × 11 × 13-cm pots. Plants were grown in a controlled growth room under long-day conditions (16 h light, 8 h dark). Days to flowering were measured as the number of days between seedling planting until the appearance of the first open flower. The ES138 line required a 6-wk vernalization treatment at 4°C to induce flowering. While we were successful in using ES 138 for genetic crossing, the plant perished before we collected measurements on the production of pollen or selfed seed.

Chromosome preparation

Fast-growing root tips were collected between 09:00 and 10:00 h and incubated in a 2-mm aqueous solution of 8-hydroxyquinoline at 15°C for 3 h, then fixed in ethanol/acetic acid (3 : 1) at 4°C. Meiotic preparations were made of spread meiotic cells from inflorescences with anthers containing pollen mother cells at meiosis. The antheres were collected and fixed in acetic acid : ethanol (1 : 3). To make the microscope slides, we rinsed the material three times in water and once in 10 mm citrate buffer pH 4.5, then transferred the material to a pectolytic enzyme mixture (0.3% pectolyase, 0.3% cellulase RS, 0.3% cytohelicase in 10 mm citrate buffer) for 2 h (meiotic material, 2.5 h) at 37°C. The root-tip or anther tissues were washed again, and with fine needles we dissected the root tip meristem in a small drop of 60% acetic acid on the slide, while heating carefully on a hot plate at 45°C for 60–90 s. The material was spread on the microscopic slide by dropping approx. 2 ml freshly prepared ethanol : acetic acid (3 : 1) onto the cells in the acetic acid solution, followed by a brief dehydration in ethanol 98% and air-drying. We selected the best chromosome spread preparations under the phase-contrast microscope.

All chromosome preparations were counterstained with 20 µl 0.2 µg ml−1 4′,6-diamidine-2-phenylindole (DAPI) in Vectashield Mounting Medium (Vector Laboratories, Burlingame, CA, USA). Selected mitotic cell complements were captured for quantification and karyotype analysis. Lengths of chromosome arms, heterochromatic regions and secondary constrictions (nucleolus organiser regions, NORs) were established with the measuring tool of Adobe photoshop, and statistical analyses were performed in Microsoft excel. For karyotype analysis, we cut out individual chromosomes digitally, ordering the short arm upwards and matching chromosomes where possible on the basis of length, centromere position, heterochromatin pattern and presence of satellites/NORs.

Genome painting

We isolated total genomic DNA from the diploid sexuals B. holboellii (Hornemann) A. Löve & D. Löve, BH208; B. stricta, BS2; and A. thaliana (accession Columbia) with the Nucleon Phytopure extraction kit (Amersham Biosciences, Amersham, UK). DNA was labelled with either digoxigenin-11-dUTP or biotin-16-dUTP using the nick-translation kit of Roche (Roche Diagnostics GmbH, Roche Applied Science, Mannheim, Germany). In the prehybridization step, we first dried the microscopic preparations at 67°C for 30 min before incubation with 1 µg ml−1 RNAse-A in 2 × SSC at 37°C for 1 h, and two wash steps in 2 × SSC for 5 min at 20°C. The preparations were rinsed in 10 mm HCl for 2 min and then treated with 100 µl pepsin (5 µg ml−1) in 10 mm HCl for 5 min at 37°C, washed three times in 2 × SSC for 5 min, and fixed in 10% formaldehyde for 10 min followed by two washes in 2 × SSC for 5 min (all steps at 20°C). The preparations were dehydrated in ethanol series (70%, 90% and absolute ethanol for 3 min each) and air-dried. For single-colour genome painting, we used a hybridization mixture of 50% formamide, 10% sodium dextrane sulphate, 2 × SSC, 0.25% SDS and 100 ng DNA probe (BH208 or BS2 genomic DNA) and 10 µg blocking DNA, in a total of 40 µl hybridization buffer. For the two-colour genome painting (GISH), the hybridization mixtures contained both BH208 and BS2 probes and blocking DNA (1 : 100 total genomic DNA of A. thaliana). The mixture was denatured at 100°C for 10 min and chilled immediately on ice for 10 min before use. For each preparation we added 40 µl of the hybridization mix and covered it with a 24 × 50 mm cover slide, followed by heating on an 80°C hot plate for 2.5 min before an overnight hybridization at 37°C in a humidified chamber. Post-hybridization washes involved three wash steps of 50% formamide in 2 × SSC (pH 7.0) at 42°C for 5 min, and two steps in 2 × SSC for 5 min at 20°C. The hybridization signals were detected with fluorescein isothiocyanate (FITC)-conjugated anti-DIG antibodies and amplified with FITC-conjugated rabbit anti-sheep antibodies for the digoxigenin-labelled probe, and with Avidin Texas-Red for the biotin-labelled probe that was amplified with biotinylated antiavidine and Avidin Texas-Red. The preparations were counterstained in 100 µl of a 2 µg ml−1 DAPI solution in 100 mm citrate buffer pH 6.0 for 10 min in the dark, and finally mounted in Vectashield (Vector Laboratories) under a 24 × 50-mm cover slip. Chromosomes were examined under a Zeiss Axioplan 2 photomicroscope equipped with epifluorescence illumination and filter sets for DAPI, FITC and Texas Red fluorescence. The images were captured with a Photometrics Sensys 1305 × 1024 image array CCD camera and analysed using genus ver. 2.7 image analysis workstation software (Applied Imaging International Ltd., Newcastle upon Tyne, UK). DAPI images were sharpened with a 7 × 7 Hi-Gauss high-pass spatial filter to accentuate minor details and heterochromatin differentiation of the chromosomes. We used the levels and curves tools in Adobe photoshop to improve DAPI heterochromatin differentiation banding, and used the saturation tool to enhance colour saturation of the green and red fluorescence signals.

DNA extraction, microsatellite amplification and analysis

The isolation of DNA, microsatellite amplification and analysis were carried out as described by Schranz et al. (2005). Amplifications were done with microsatellites (GC27, ICE3, Bf-3, ICE14, H23) that were identified as being heterozygous with two alleles each (bi-allelic) in the VP9 genotype (Schranz et al., 2005). Based on sequence similarity to Arabidopsis and preliminary genetic mapping results (M.E.S. and T.M.O., unpublished data) these loci are unlinked. Ten self-derived progeny from VP9, eight test-cross lines from ES136, and two test-cross lines from ES137 were analysed.

Genome-size measurements

Ploidy analyses were performed on a PARTEC (Münster, Germany) CCA-II flow cytometer using their CyStain UV precise P nuclei extract and staining kit (PARTEC GmbH) according to the manufacturer's protocol. Sample leaf material was measured in combination with an internal size standard (leaf material from Brassica rapa and/or Matthiola incana for larger genome individuals), and the ratio of the mean fluorescence intensity values for the 2c peaks for both sample and standards were calculated (Schranz et al., 2005). When Matthiola was used as standard, the ratio was converted to the Brassica units (by using the size ratio of Matthiolia compared with Brassica). A minimum of 10 000 fluorescence counts was collected for each sample run.

Seed collection and measurement

Seed collections were made from mature plants. Ten siliques from each of five mature reproductive axes per plant were collected (for a total of 50 siliques). The number of seeds from each group of 10 siliques was counted and the average number of seeds per fruit calculated (Table 1; see Fig. 6).


Figure 6. Comparison of genome size (measured by flow cytometry and given as ratio to reference Brassica rapa genome) and seed production of F2 lines of the Boechera SAD12 × VP9 population. Most lines are highly infertile except two tetraploids (ES 136 and ES 137) with greater seed production.

Download figure to PowerPoint

Pollen grain staining and measurement

Measurements of pollen counts and viability were estimated using a modified Alexander's stain (Alexander, 1980). Pollen grains were counted by collecting two to five freshly opened flowers from each plant. The Alexander stain that gave the best results contained the minimum amount of lactic acid (0.5 ml), suggested for thin-walled pollen (Alexander, 1980). A dilution of 1 : 20 Alexander stain stock solution to 50% glycerin was used. A total of 20 µl stain dilution was placed on an object slide. An individual stamen from each flower was put on a microscopic slide containing a drop of the stain dilution, and covered with a cover slip. We left the material for at least 10 min before analysing the pollen grains. The numbers of viable (purple) and nonviable (blue/green) pollen grains were recorded using an Axioskop 2 compound microscope (Carl Zeiss GmbH, Jena, Germany) at ×20 magnification. Between 400 and 2000 pollen grains for each plant were counted, from which the percentage of viable pollen was calculated (Table 1).

The sizes of both viable and nonviable grains were also measured using the stained pollen grains. Photos of the pollen grains were made with a Zeiss AxioCam camera and using the program AxioVision ver. (Carl Zeiss Vision GmbH, Hallbergmoos, Germany). The area of pollen grains was estimated from the photos using the ImageJ ver. 1.33µ program (Abramoff et al., 2004) from which the diameter of each grain was calculated. Between 20 and 44 individual photos, representing between 94 and 825 individual pollen grains per plant, were analysed (see Fig. 7).


Figure 7. Histograms of Boechera pollen grain size and viability. Alexander's stain was used to differentiate viable (purple) and nonviable (blue/green) pollen grains, and the diameter of each grain was measured by analysis of photos of pollen. Frequency based on total number of pollen grains counted. Size distribution and selected pollen grain images of (a) viable; (b) nonviable pollen grains of six selected lines used to illustrate the diversity of viability and size of the various lines. The lines used were SAD12 (diploid sexual); VP9 (diploid apomict); ES117 (triploid F1); ES137 (near-tetraploid F2); ES117 #15 (aneuploid F2); ES137 #15 (near-tetraploid F3). (c) Size distribution of all viable and nonviable pollen grains measured from aneuploid and polyploid lines (excluding the two diploid parental lines).

Download figure to PowerPoint


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Genome composition revealed by GISH

Genomic in situ hybridizations (GISH) were initially done with total genomic DNA from the diploid sexual B. stricta labelled as a probe (red) and blocked with genomic DNA from B. holboellii (Fig. 2a–d). Chromosome complements showed fluorescent signals on the pericentromere regions of only the seven B. stricta-derived chromosomes in the hybrid B. divaricarpa VP9 genome (Fig. 2c), confirming its homoploid origin (hybridization with no increase in ploidy) between a B. stricta- and B. holboellii-like plant. Additionally, the hybrid B. divaricarpa VP9 genome was found to contain a highly heterochromatic chromosome (Het) (indicated by arrows in Fig. 2a,c) that may have properties similar to B-chromosomes and/or Het chromosomes identified in 15-chromosome Boechera apomicts (Sharbel et al., 2004; Kantama, 2005). The fluorescent signal on the Het chromosome (Fig. 2c) suggests it is derived from the B. stricta ancestor of this hybrid line. Genome painting of ES 117 F1 line chromosomes (Fig. 2d) further revealed a total of 14 B. stricta chromosomes (seven derived from a reduced haploid B. stricta SAD12 gamete; the other seven from the nonreduced B. divaricarpa VP9 gamete) and seven B. holboellii-like chromosomes (seven derived from the nonreduced B. divaricarpa VP9 gamete).


Figure 2. Genome composition revealed by genomic in situ hybridization (GISH) of selected Boechera SAD12 × VP9 population lines. DAPI staining of chromosomes of: (a) paternal apomictic Boechera divaricarpa genotype (VP9); (b) F1 line (ES 117). Fluorescence in situ hybridizations done with total genomic DNA from the diploid sexual Boechera stricta labelled as a probe (red) and blocked with genomic DNA from Boechera holboellii staining the pericentromere regions of: (c) seven B. stricta-derived chromosomes in the hybrid B. divaricarpa (VP9) genome; (d) 14 B. stricta-like chromosomes in the F1 line (ES 117) genome (arrows in Fig. 5(a,c) denote the Het chromosome). Additional two-colour genome painting done by the simultaneous hybridization of both B. stricta (green) and B. holboellii (red) probes onto (e) one of the high-fertility F2 lines, ES 137, showing 14 B. stricta chromosomes (green) including a highly heterochromatic chromosome (arrow) and 13 B. holboellii-like chromosomes.

Download figure to PowerPoint

An additional two-colour genome painting was carried out with the simultaneous hybridization of both B. stricta (green) and B. holboellii (red) probes, and blocking with total genomic DNA of A. thaliana in order to further improve the discrimination of the parental species (Kantama, 2005). The painting of one of the highly fertile F2 lines, ES 137, revealed 14 B. stricta chromosomes (green), including the highly heterochromatic (Het) chromosome (arrow in Fig. 2e), and 13 B. holboellii-like chromosomes (Fig. 2e). Thus the two likely aneuploid gametes derived from the ES 117 line (with its 14 B. stricta and seven B. holboellii-like chromosomes) that united to form the ES 137 plant were effectively able to restore the chromosome balance of one B. stricta to one B. holboellii-like chromosome (Fig. 2e). The restored genomic ratio could explain the higher fertility of this line. However, we do not yet know if we have complete chromosome complements in these lines.

Analysis of pollen meiosis of VP9

Anther preparations of VP9 showed very few meiocytes. We found pachytene and metaphase I complements with fully paired bivalents. Metaphase I cells diplayed seven regular bivalents, showing that meiosis in this plant can be fully synaptic. At this stage we could identify the Het chromosome in a few cells forming a heteromorphic bivalent with one of the other chromosomes (arrow, Fig. 3). We did not find any anaphase I or anaphase II complement.


Figure 3. DAPI-stained chromosome spread of anther cells of VP9. Metaphase I complement with seven bivalents showing that meiosis in this plant is fully synaptic. Arrow, Het chromosome in the heteromorphic bivalent.

Download figure to PowerPoint

Microsatellite analysis of self-progeny of VP9 and test-cross lines

In our previous study (Schranz et al., 2005) we analysed microsatellites that were bi-allelic in VP9. In a sample of 10 progeny derived by self-pollination of VP9, heterozygosity was maintained (Fig. 4). The probability of obtaining all heterozygous genotypes by sexual reproduction with independent assortment is very unlikely (P < 0.001). This result strongly supports the conclusion that VP9 is an apomictic lineage. The microsatellites used were also highly polymorphic between VP9 and both SAD12 and ES138. The test-cross lines, where ES136 and ES137 were used as maternal plants and ES138 as a pollen donor, all contained alleles derived from ES138. This result, and the dramatic change in phenotype (exemplified by the shift to very late flowering time), both indicate that the near-tetraploid lines ES136 and ES137 reproduce sexually.


Figure 4. Microsatellite amplification of 10 progeny derived by self-pollination of VP9. At several unlinked microsatellite loci, including this locus (GC27), the initial VP9 plant was heterozygous. In a sample of 10 progeny derived by self-pollination of VP9, heterozygosity was maintained for all samples, suggesting that progeny were derived apomictically rather than sexually.

Download figure to PowerPoint

Genome size, chromosome numbers and karyotypes

Flow cytometry results of all lines were standardized to the diploid B. rapa genome (Table 1), and chromosome counts were done for most of the parental, F1 and F2 plants (Table 1). Comparison of flow cytometry values with chromosome counts (assuming all chromosomes are of approximately the same size) showed very good correspondence (R2 = 0.98) (Fig. 5a). Chromosome counts and karyotype analysis (Fig. 5b) confirmed that the parental genotypes were diploid (2n = 2x = 14). The F1 plants (ES 116 and ES 117) were also confirmed to be triploids (2n = 3x = 21) (Fig. 5b), as hypothesized from earlier flow cytometry results, probably caused by the union of a reduced and nonreduced gamete (Schranz et al., 2005)


Figure 5. Measures of genome size of lines in the Boechera SAD12 × VP9 population. (a) Correlation of genome size (measured by flow cytometry and given as ratio to reference Brassica rapa genome) and chromosome numbers. (b) Karyotypes and chromosome numbers of selected generations and lines. (c) Distribution in genome size (measured by flow cytometry and given as ratio to reference B. rapa genome) of the F3 lines derived from ES136 and ES137, demonstrating the genomic stability of these lines near tetraploidy.

Download figure to PowerPoint

The most striking result was the swarm of chromosome numbers of the F2 lines, ranging from triploid to tetraploid, with most being aneuploid (Table 1; Fig. 5a). While overall genome-size measurements agreed with the chromosome counts, there was one notable exception. The F2 line ES 117 #9 had the largest genome size measured by flow cytometry (1.12), but gave a mosaic of aneuploid chromosome numbers (24 and 27). Two F2 lines, ES 136 and ES 137, were found to be highly fertile (see below). Chromosome counts for both these lines showed that they were nearly tetraploid (2n ≈ 4x ≈ 27) (Table 1; Fig. 5b). Because of their fertility, we investigated the genome size of their F3 progeny. All F3 progeny had genome sizes of the same magnitude (ranging from 0.94 to 1.08) (Table 1; Fig. 5c), suggesting that the genomes of these two lines have stabilized near tetraploidy.

The strong correlation between the flow cytometry results and chromosome number allowed us to estimate the genome size of Boechera. A recent study (Johnston et al., 2005) estimated the haploid genome of B. rapa (n = 10) at 1C = 529 Mb. Using this estimate, the haploid Boechera genome (n = 7) would be approx. 264 Mb. However, their estimate for the size of the A. thaliana genome is 157 Mb (Johnston et al., 2005), which is larger than current estimates based on genome sequencing. Using the latter value, we infer that the haploid Boechera genome is roughly 1.7 times that of A. thaliana.

Seed production of aneuploid and polyploid F2 lines

Analysis of seed production of F1 and F2 plants found that most were highly infertile (Table 1; Fig. 6). However, two F2 lines, ES 136 and ES 137, had much greater seed set with c. 10 seeds per silique, compared with an average of only 0.36 seeds per silique for the other lines (Table 1; Fig. 6). Correlation between genome size and seed set (Fig. 6) was not significant, as most lines of varying genome sizes were infertile, but the two higher-fertility lines were nearly tetraploid. The F3 lines derived from ES 136 and ES 137 all maintained high levels of fertility, with an average of 15.11 seeds per fruit; however, there was still great variation in seed production (5.65–48.88 seeds per silique; Table 1).

Pollen grain viability and size

The viability of pollen grains differed between the two parental lines, with sexual B. stricta SAD12 producing almost all viable pollen, and the apomictic B. divaricarpa VP9 producing 74.88% viable pollen (Table 1). The triploid F1 plant (ES 117) had greatly reduced pollen viability of 24.52% (Table 1). The swarm of F2 plants displayed great variation in viability, from only 1.87% up to 75.88 and 72.81% for lines ES 136 and ES 137, respectively (Table 1). The F3 progeny of ES 136 and ES 137 maintained the higher-viability pollen, with an average of 69.92% (Table 1).

The analysis of pollen grain size of viable and nonviable pollen revealed several striking results (Fig. 7). The sexual B. stricta SAD12 produced almost exclusively reduced and viable pollen grains with an average diameter of 16.53 µm (Fig. 7a). The apomictic B. divaricarpa VP9 produced viable, nonreduced pollen grains with an average diameter of 22.32 µm. Interestingly, VP9 also produced some pollen grains of even larger diameter (Fig. 7a). The F1 line ES 117 produced viable pollen grains of varying size, from 17.66 to 25.82 µm (Fig. 7a), probably reflecting a wide range of aneuploid gametes. However, no viable gametes of haploid size were detected. The nonviable pollen of ES 117 (Fig. 7b) was found to be smaller than that of the viable pollen. The size range of the nonviable pollen was from 12.51 to 22.31 µm, corresponding to gametes that would be less than haploid to about diploid in size. The F2 line, ES 117 #15, was triploid and produced almost exclusively nonviable pollen (1.87% viability; Table 1). The size distribution of the nonviable grains for ES 117 #15 (Fig. 7b) ranged from 11.92 to 19.90 µm, and thus were smaller than diploid in size. Finally, one of the F3 lines derived from ES 137 (ES 137 #15) produced viable pollen grains from 18.89 to 25.09 µm, and nonviable pollen grains from 13.37 to 20.87 µm in diameter (Fig. 7a,b).

The six lines discussed above (Fig. 7a,b) illustrate the variability in pollen viability and size from the various generations. In addition, overall pollen grain viability and size for all nonparental lines was also determined (Fig. 7c). The average size of nonviable gametes was 15.88 µm, which is slightly less than that of haploid, and far less than the 21.91 µm of the diploid viable gametes (Fig. 7c). It should be noted that the ratio of nonviable to viable pollen grains measured for size was lower than that obtained by simply counting nonviable and viable pollen grains. This discrepancy was because nonviable pollen grains clumped together and thus it was difficult to take good photographs for size measurements.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Here we present the first genetic analysis of the mechanisms and inheritance of apomictic reproduction in Boechera. First, we characterized more fully the paternal diploid apomictic line VP9. Genome painting revealed its hybrid origin, whereas the analysis of male meiosis demonstrated that the chromosomes are fully synaptic, including a heterochromatic (Het) chromosome that fully pairs with one of the other chromosomes. Male meiosis produced both reduced and unreduced gametes, suggesting first- or second-division restitution. Heterozygosity was also maintained in self-pollinated progeny, thus supporting its apomictic nature. We wondered if this Het chromosome carries the genetic elements controlling apomixis in a manner similar to several other systems where supernumerary or hemizygous chromosomes carry an apomixis factor (Ozias-Akins et al., 1998; Bicknell et al., 2000; Roche et al., 2001; Labombarda et al., 2002; Akiyama et al., 2005). Therefore we analysed the progeny derived from a cross between a diploid sexual and this diploid apomictic line. The resulting triploid F1 was largely sterile, producing limited numbers of viable pollen and seed. When the few seeds were grown, they produced plants that were mostly aneuploid or near-tetraploid. Hence they were not apomictically derived replicates of the three maternal genomes, but rather the product of a reduced or unreduced meiosis. Two of the recovered lines were fertile, sexual and near-tetraploid. Genome painting (GISH) analysis of one of these fertile lines found that it had a nearly balanced chromosome complement of 14 B. stricta and 13 B. holboellii chromosomes, suggesting that genome balance may play an important role in establishing the sexual phenotype. However, future work will be needed to establish if these correspond to complete chromosomal complements. Below we discuss in greater detail some of the implications of our results for understanding the inheritance, control and evolution of apomixis in Boechera, and the potential importance of the generation of sexual tetraploid lineages.

Control of apomixis in Boechera

Asexual seed formation is a complex trait resulting from modification of the sexual life cycle (reviewed by Koltunow & Grossniklaus, 2003; Bicknell & Koltunow, 2004). The normal sexual pathway can be altered or deregulated by the inheritance of a single dominant gene, hemizygous genomic region, and/or supernumerary element (Nogler, 1984b; Sherwood et al., 1994; Bicknell et al., 2000; Roche et al., 2001). Our F1 plants were derived from the union of a reduced egg cell from the diploid sexual maternal plant and a nonreduced pollen grain from the apomictic paternal plant, as shown by GISH analysis. Hence the F1 progeny probably inherited the complete genome of the apomictic parent. However, there may have been cross-over events during the failed meioses of the paternal VP9 plant that could have produced recombinant pollen grains. Future segregation analyses of molecular markers will be necessary to examine this possibility.

Our chromosome analyses showed that the diploid apomictic paternal plant had an aberrant chromosome resembling the heterochromatic (Het) chromosome in other 14- and 15-chromosome Boechera apomicts (Kantama, 2005), but lacked the extra chromosome found in the 15-chromosome apomicts (Sharbel et al., 2004, 2005). The offspring individuals in the F2 progeny displayed varying aneuploid and polyploid chromosome numbers, suggesting that apomixis in Boechera is not regulated by the simple inheritance of a single dominant locus or factor (e.g. the Het chromosome). If it were, then the F1 line, which contains the Het chromosome, would be apomictic with all or most derived F2 progeny being identical to the maternal plant.

Various authors have pointed to polyploidy as being indispensible in the regulation and transmission of the apomixis trait (Quarin et al., 2001). But in Boechera, where naturally occurring triploids are indeed apomictic, they can be highly facultative with many progeny derived sexually (Schranz et al., 2005). The results of our crossing have also shown that an increase in ploidy from diploidy to triploidy is not necessarily associated with the transfer or expression of the apomictic phenotype.

It has also been postulated that the hybrid constitution of apomictic lineages, rather than polyploidy, will lead to expression of the apomictic phenotype (Bicknell & Koltunow, 2004). Specifically, the hybridization-derived floral asynchrony hypothesis postulates that apomixis is caused by the differential expression and temporal regulation of genes in normal sexual reproductive pathways (Carman, 1997; Koltunow & Grossniklaus, 2003). In agreement with this hypothesis is our demonstration by GISH that the diploid apomictic parent (VP9) is of hybrid origin, with seven B. stricta and seven B. holboellii-like chromosomes. In addition, all diploid and triploid lines that were apomictic and/or produced nonreduced gametes were also highly heterozygous, attesting to their likely hybrid origin (Schranz et al., 2005). Similarly, individuals found to be heterozygous at microsatellite loci (Dobešet al., 2004b) and at a sequenced molecular marker (Sharbel et al., 2004, 2005) tended to produce nonreduced gametes, and were concluded to be apomictic.

However, hybridization alone cannot explain apomixis in Boechera. Our triploid F1, which contained a complete haploid genome from the highly inbred maternal B. stricta plant and the potentially complete diploid B. divaricarpa genome from the apomictic paternal plant, is not apomictic. One possible explanation for this failure in transmission of the apomixis phenotype could be segregation of the apomixis gene(s) caused by recombination during the production of the 2n pollen grains in the apomictic paternal plant. Chromosome analysis, however, showed that all or part of the heterochromatic (Het) chromosome was transmitted to the F1 plant. Another explanation for the discrepancy may lie in the relative dosage of B. stricta and B. holboellii-like chromosomes. GISH analyses of the diploid apomictic showed there is a 1 : 1 genome ratio, but in the F1 triploid there is a ratio of 2 : 1 in favour of the sexual B. stricta genome. The higher dose of B. stricta might allow for the correct, and disruptive, expression of sexuality in the triploid. Naturally occurring triploid apomictic lineages may have two doses of B. holboellii-like chromosomes, or a dosage closer to 1 : 1 if the triploid apomictic line is derived from the union of two aneuploid gametes. However, one of the recovered F2 near-tetraploid lines has a nearly 1 : 1 genome ratio, yet is sexual. Sexuality at the tetraploid level may be feasible because of an interaction of dosage and polyploidization, or alternatively because of the segregation of a gamete-lethality factor.

The results of our pollen analysis may shed light on the possible segregation of a gamete-lethal factor or gene. In all samples, except the diploid sexual line, there a significant fraction of nonviable pollen grains were produced, which were of approximately the same size as viable haploid pollen. In contrast, viable pollen grains in all samples, except for the sexual diploid, were of approximately diploid size. No viable haploid-sized pollen was observed for any of these lines. Some of the nonviable pollen grains might be caused simply by chromosomal imbalances, particularly those generated from aneuploid F2 lines where the percentages of nonviable pollen grains were as high as 98%. However, both the apomictic diploid parent and the recovered sexual tetraploids produced, on average, 25% nonviable gametes and 75% viable diploid-sized pollen grains. This result would be consistent with the recessive lethal gametophytic selection hypothesis of (Nogler, 1984b), in which an apomixis gene or factor is lethal when it occurs in a haploid pollen grain in the absence of a wild-type allele. Future studies examining the patterns of inheritance of molecular and/or cytological markers will be critical for resolving models for the control of apomixis in Boechera.

Escape from apomixis, low fertility and return to sexuality

Traditionally, apomictic lineages were thought to have limited evolutionary potential and to be doomed to extinction (Darlington, 1939; Stebbins, 1950). Later, researchers realized that the maintenance of male function and facultative apomixis provided mechanisms for perpetuating apomictic lineages by transferring apomixis genes via hybridization with sexual relatives (van Dijk, 2003). The transfer of apomixis genes into new genetic backgrounds would allow the purging or masking of deleterious mutations that may have accumulated in the apomict (Muller's ratchet), and would generate new genetic diversity, allowing escape from parasitism. Another possibility is that the hybridization of asexuals with sexuals allows genes from the apomictic parent to re-enter the sexual gene pool (Chapman et al., 2003), undergo recombination, and potentially go on to form new apomictic lineages (de Wet, 1968).

We have detected such a shift in breeding system, from apomict diploid to low-fertility sexual triploid to fertile and sexual tetraploid, in our experimental cross of Boechera. In many ways our cross follows the classic model of the formation of autotetraploids via a triploid bridge, and the inherent problems of meiosis in triploids and the frequent endosperm developmental problem of the triploid block (Ramsey & Schemske, 1998; Husband, 2004; Henry et al., 2005).

There are two routes from apomictic genotypes to the formation of sexual tetraploids in Boechera. In this work we demonstrated that, by crossing the two diploids, we could generate sexual near-tetraploids in only two generations. In our earlier work, we frequently derived de novo tetraploids directly from the crossing of a sexual diploid and an apomictic triploid line. These sexual tetraploid lines can then undergo all the classic advantages of sex, such as recombination and independent assortment. There could also be chances for homologous pairing, translocations and epigenetic modifications often seen in new polyploid lines (Osborn et al., 2003), all generating new phenotypic diversity. Also, it is important to note that these tetraploid Boechera lines could potentially express apomixis, or some partial component of apomixis, at some frequency, caused by the penetration or expressiveness of the traits. The partial expression of apomixis, especially parthenogenesis, may be particularly important for the generation of new diploid apomictic lineages. The establishment of diploid hybrid apomictic haplotypes in Boechera could be caused by homoploid hybridization, or alternatively by base chromosome number reductions (Asker & Jerling, 1992). If reduced 2n egg cells from tetraploid Boechera lineages develop parthenogenically, they could establish new diploid lineages that could be either sexual or apomictic. Such cycles of ploidy have been described for other apomictic complexes, such as Potentilla argentea and the Bothriochloa–Dichanthium complex, and are known as the ‘diploid–tetraploid–dihaploid cycle’ (de Wet, 1968; Asker & Jerling, 1992). The independent assortment and recombination of chromosomes derived from the parental genotypes or species of the tetraploid means that there could be unusual constellations of B. stricta and B. holboellii chromosomes or chromosome regions in dihaploids produced parthenogenically.

Interestingly, few naturally occurring tetraploid Boechera lineages have been detected (reviewed by Dobešet al., 2006), but when they do occur, they are likely to be sexual (Böcher, 1969; Johnson, 1970; Schranz et al., 2005). The low frequency of tetraploids could be explained by a potential propensity to produce dihaploid offspring. Alternatively they may be selected against, possibly because of a fitness disadvantage. It is interesting to note that sexual tetraploids existing at low frequency in populations of dandelions have been postulated to be critical for the creation of new triploid apomictic cytotypes, by the production of reduced 2n gametes that recombine with reduced n gametes from sexual diploids (Verduijn et al., 2004). Overall, the historical occurrence of hybridization shifts in breeding systems and alterations of ploidy have profound implications for our understanding of the inheritance, transmission and evolution of genetic and quantitative variation for Boechera.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors thank Juliane Pfuetzenreuter, Silke Fuchs and Radim Vasut for technical assistance. Thanks to John E. Maunder and Nathalie Djan-Chékar of the Provincial Museum of Newfoundland and Labrador for providing seeds. The authors also thank Peter van Dijk for helpful comments. Support for this research was provided by the Max Planck Gesellschaft to M.E.S. and by the Thai Government to L.K.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Abramoff MD, Magelhaes PJ, Ram SJ. 2004. Image processing with ImageJ. Biophotonics International 11: 3642.
  • Akiyama Y, Hanna W, Ozias-Akins P. 2005. High-resolution physical mapping reveals that the apospory-specific genomic region (ASGR) in Cenchrus ciliaris is located on a heterochromatic and hemizygous region of a single chromosome. Theoretical and Applied Genetics 111: 10421051.
  • Albertini E, Porceddu A, Ferranti F, Reale L, Barcaccia G, Romano B, Falcinelli M. 2001. Apospory and parthenogenesis may be uncoupled in Poa pratensis: a cytological investigation. Sexual Plant Reproduction 14: 213217.
  • Alexander MP. 1980. A versatile stain for pollen, fungi, yeast and bacteria. Stain Technology 55: 1318.
  • Asker SE, Jerling L. 1992. Apomixis in Plants. Boca Raton, FL, USA: CRC Press.
  • Bicknell RA, Koltunow AM. 2004. Understanding apomixis: recent advances and remaining conundrums. Plant Cell 16: S228S245.
  • Bicknell RA, Borst NK, Koltunow AM. 2000. Monogenic inheritance of apomixis in two Hieracium species with distinct developmental mechanisms. Heredity 84: 228237.
  • Böcher T. 1951. Cytological and embryological studies in the amphil–apomictic Arabis holboellii complex. Biologiske Skrifter 6: 158.
  • Böcher T. 1954. Experimental taxonomical studies in the Arabis holboellii complex. Svensk Botanisk Tidskrift 48: 3144.
  • Böcher T. 1969. Further studies in Arabis holboellii and allied species. Saertryk af Tidsskrift 64: 141161.
  • Carman J. 1997. Asynchronous expression of duplicated genes in angiosperms may cause apomixis, bispory, tetraspory, and polyembryony. Biological Journal of the Linnean Society 61: 5194.
  • Chapman H, Houliston GJ, Robson B, Iline I. 2003. A case of reversal: the evolution and maintenance of sexuals from parthenogenetic clones in Hieracium pilosella. International Journal of Plant Sciences 164: 719728.
  • Charlesworth B. 1990. Mutation–selection balance and the evolutionary advantage of sex and recombination. Genetical Research 55: 199221.
  • Crane CF. 2001. Classification of apomictic mechanisms. In: SavidanY, CarmanJ, DresselhausT, eds. The Flowering of Apomixis: From Mechanisms to Genetic Engineering. Mexico: CIMMYT, IRD/European Commission DG VI, 2443.
  • Darlington C. 1939. The Evolution of Genetic Systems. Cambridge, UK: Cambridge University Press.
  • Van Dijk PJ. 2003. Ecological and evolutionary opportunities of apomixis: insights from Taraxacum and Chondrilla. Philosophical Transactions of the Royal Society of London, Series B – Biological Sciences 358: 11131120.
  • Van Dijk PJ, Tas ICQ, Falque M, Bakx-Schotman T. 1999. Crosses between sexual and apomictic dandelions (Taraxacum). II. The breakdown of apomixis. Heredity 83: 715721.
  • Dobeš CH, Koch MA, Sharbel TF. 2006. Embryology, karyology, and modes of reproduction in North American genus Boechera (Brassicaceae): a compilation of seven decades of research. Annals of the Missouri Botanical Garden. (In press.)
  • Dobeš C, Mitchell-Olds T, Koch MA. 2004a. Intraspecific diversification in North American Boechera stricta (=Arabis drummondii), Boechera×divaricarpa and Boechera holboellii (Brassicaceae) inferred from nuclear and chloroplast molecular markers – an integrative approach. American Journal of Botany 91: 20872101.
  • Dobeš CH, Mitchell-Olds T, Koch MA. 2004b. Extensive chloroplast haplotype variation indicates Pleistocene hybridization, radiation of North American Arabis drummondii Arabis drummondii, A.×divaricarpa, and A. holboellii (Brassicaceae). Molecular Ecology 13: 349370.
  • Henry I, Dilkes B, Young K, Watson B, Wu H, Comai L. 2005. Aneuploidy and genetic variation in the Arabidopsis thaliana triploid response. Genetics 170: 19791988.
  • Holsinger KE. 2000. Reproductive systems and evolution in vascular plants. Proceedings of the National Academy of Sciences, USA 97: 70377042.
  • Husband BC. 2004. The role of triploid hybrids in the evolutionary dynamics of mixed-ploidy populations. Biological Journal of the Linnean Society 82: 537546.
  • Johnston JS, Pepper AE, Hall AE, Chen ZJ, Hodnett G, Drabek J, Lopez R, Price HJ. 2005. Evolution of genome size in Brassicaceae. Annals of Botany 95: 229235.
  • Johnson TF. 1970. Investigations in the floral biology of the Arabis holboellii complex. MSc thesis. Seattle, WA, USA: University of Washington.
  • Kantama L. 2005. Chromosome studies and genetic analyses of natural and synthetic apomictic model species. PhD thesis. Wageningen, the Netherlands: Wageningen University.
  • Koch M, Haubold B, Mitchell-Olds T. 2001. Molecular systematics of the Brassicaceae: evidence from coding plastidic matK and nuclear Chs sequences. American Journal of Botany 88: 534544.
  • Koch MA, Dobeš C, Mitchell-Olds T. 2003. Multiple hybrid formation in natural populations: concerted evolution of the internal transcribed spacer of nuclear ribosomal DNA (ITS) in north American Arabis divaricarpa (Brassicaceae). Molecular Biology and Evolution 20: 338350.
  • Koltunow AM, Grossniklaus U. 2003. Apomixis: a developmental perspective. Annual Review of Plant Biology 54: 547574.
  • Labombarda P, Busti A, Caceres ME, Pupilli F, Arcioni S. 2002. An AFLP marker tightly linked to apomixis reveals hemizygosity in a portion of the apomixis-controlling locus in Paspalum simplex. Genome 45: 513519.
  • Leblanc O, Grimanelli D, Gonzalez-d, E-Leon D, Savidan Y. 1995. Detection of the apomictic mode of reproduction in maize–Tripsacum hybrids using maize RFLP markers. Theoretical and Applied Genetics 90: 11981203.
  • Matzk F, Prodanovic S, Baumlein H, Schubert I. 2005. The inheritance of apomixis in Poa pratensis confirms a five locus model with differences in gene expressivity and penetrance. Plant Cell 17: 1324.
  • Mogie M. 1992. The Evolution of Asexual Reproduction in Plants. New York, NY, USA: Chapman & Hall.
  • Naumova TN, Van Der Laak J, Osadtchiy J, Matzk F, Kravtchenko A, Bergervoet J, Ramulu KS, Boutilier K. 2001. Reproductive development in apomictic populations of Arabis holboellii (Brassicaceae). Sexual Plant Reproduction 14: 195200.
  • Nogler GA. 1984a. Gametophytic apomixis. In: JohriB, ed. Embryology of Angiosperms. Berlin: Springer, 475518.
  • Nogler GA. 1984b. Genetics of apospory in apomictic Ranunculus auricomus. 5. Conclusion. Botanica Helvetica 94: 411422.
  • Noyes RD, Rieseberg LH. 2000. Two independent loci control agamospermy (apomixis) in the triploid flowering plant Erigeron annuus. Genetics 155: 379390.
  • Osborn TC, Pires JC, Birchler JA, Auger DL, Chen ZJ, Lee HS, Comai L, Madlung A, Doerge RW, Colot V, Martienssen RA. 2003. Understanding mechanisms of novel gene expression in polyploids. Trends in Genetics 19: 141147.
  • Ozias-Akins P, Roche D, Hanna WW. 1998. Tight clustering and hemizygosity of apomixis-linked molecular markers in Pennisetum squamulatum genetic control of apospory by a divergent locus that may have no allelic form in sexual genotypes. Proceedings of the National Academy of Sciences, USA 95: 51275132.
  • Quarin CL, Espinoza F, Martinez EJ, Pessino SC, Bovo OA. 2001. A rise of ploidy level induces the expression of apomixis in Paspalum notatum. Sexual Plant Reproduction 13: 243249.
  • Ramsey J, Schemske DW. 1998. Pathways, mechanisms, and rates of polyploid formation in flowering plants. Annual Review of Ecology and Systematics 29: 467501.
  • Richards A. 1986. Plant Breeding Systems. London: George Allen & Unwin.
  • Roche D, Hanna WW, Ozias-Akins P. 2001. Is supernumerary chromatin involved in gametophytic apomixis of polyploid plants? Sexual Plant Reproduction 13: 343349.
  • Roy BA. 1995. The breeding systems of six species of Arabis (Brassicaceae). American Journal of Botany 82: 869877.
  • Schranz ME, Dobeš C, Koch MA, Mitchell-Olds T. 2005. Sexual reproduction, hybridization, apomixis, and polyploidization in the genus Boechera (Brassicaceae). American Journal of Botany 92: 17971810.
  • Sharbel TF, Mitchell-Olds T. 2001. Recurrent polyploid origins and chloroplast phylogeography in the Arabis holboellii complex (Brassicaceae). Heredity 87: 5968.
  • Sharbel TF, Voigt ML, Mitchell-Olds T, Kantama L, De Jong H. 2004. Is the aneuploid chromosome in an apomictic Boechera holboellii a genuine B chromosome? Cytogenetic and Genome Research 106: 173183.
  • Sharbel TF, Mitchell-Olds T, Dobeš C, Kantama L, De Jong H. 2005. Biogeographic distribution of polyploidy and B chromosomes in the apomictic Boechera holboellii complex. Cytogenetic and Genome Research 109: 283292.
  • Sherwood RT, Berg CC, Young BA. 1994. Inheritance of apospory in buffelgrass. Crop Science 34: 14901494.
  • Spillane C, Curtis MD, Grossniklaus U. 2004. Apomixis technology development – virgin births in farmers’ fields? Nature Biotechnology 22: 687691.
  • Stebbins G. 1950. Variation and Evolution in Higher Plants. New York: Columbia University Press.
  • Taskin KM, Turgut K, Scott RJ. 2004. Apomictic development in Arabis gunnisoniana. Israel Journal of Plant Sciences 52: 155160.
  • Verduijn MH, Van Dijk PJ, Van Damme JMM. 2004. The role of tetraploids in the sexual–asexual cycle in dandelions (Taraxacum). Heredity 93: 390398.
  • West S, Lively C, Read A. 1999. A pluralist approach to sex and recombination. Journal of Evolutionary Biology 12: 10031012.
  • De Wet J. 1968. Diploid–tetraploid–haploid cycles and the origin of variability in Dichanthium agamospecies. Evolution 22: 394397.