IN Arabidopsis, the reappearance of HTH (HOTHEAD) alleles from the grandparents in the offspring of homozygous hth mutants reported by Lolle et al. (2005) is not consistent with classical Mendelian genetics. Several explanations have been put forward to explain this unexpected reversion of single nucleotide polymorphisms. Lolle et al. (2005) suggested that RNAs synthesized by the parent may be stored in plants of subsequent generations that do not carry the corresponding genomic information and that these RNAs may be used as a template to restore the genomic information carried by the previous generation. Four alternative hypotheses have been proposed to account for this non-Mendelian behavior: two of them are also based on template-directed gene conversion (Chaudhury 2005; Ray 2005), the third one appeals to a process of mutation accumulation followed by selection (Comai and Cartwright 2005; Henikoff 2005), and the fourth one involves chimerism (Krishnaswamy and Peterson 2007). However, Peng et al. (2006) reported that hth mutant shows a tendency toward outcrossing and recover a normal genetic behavior when grown in isolation. The authors suggested that this propensity to outcross may provide an alternative explanation for the apparent genetic instability of hothead mutants. Nevertheless, Lolle et al. (2006) argued that outcrossing, while possible, could not be the sole source of “reversion” (Lolle et al. 2006), leaving open the debate whether hothead is a true genetic outlaw (Pennisi 2006; Gawrylewski 2008).

To measure the level of outcrossing in hth mutant compared to wild type, we grew each of them next to (10 cm) a plant homozygote for a transgene conferring resistance to hygromycine. In the offspring of wild type, no hygromycine-resistant plant was found among 2980 (MS media, hygromycine 30 mg/liter). Thus Arabidopsis wild-type plants are highly resistant to cross-pollination, at least under our growth conditions. On the contrary, hth plants grown in the same context produced 12.1% of hygromycine-resistant plantlets (hth-4: 60/377; hth-8: 36/224; hth-10: 51/611), unambiguously demonstrating the unusual high susceptibility of the hth mutant to cross-pollination.

To test if this characteristic of the hth mutant could be the cause of its apparent genetic instability, we grew hth mutants either in isolation or at various distances from the hygromycine-resistant plants and quantified the appearance in the hth progeny of “revertant” plants (i.e., showing a wild-type phenotype and not the typical abnormal flower of hth mutants). When grown in complete isolation, no revertant was found in the offspring of the mutants among 9944 plants (Table 1). At 10 and 20 cm, hth plants produced 12.4 and 6.9% of revertants, respectively (Table 1). At 50 and 150 cm, this proportion dropped drastically to <0.1% (Table1). The proportion of revertant is thus dependent on the availability and distance to the HTH wild-type pollen grain source. Strikingly, all 430 revertants found were carrying the transgene conferring the resistance to hygromycine that can be originated only from the pollen grain donor, demonstrating that all these “reversion” events are caused by outcrossing.

Lolle et al. (2005) detected two HTH/HTH embryos among 141 dissected from fruits developing on selfed hth plants, suggesting that hth plants can produce HTH ovules, which is not consistent with the outcrossing explanation. However, among 92 plantlets obtained from the same self-cross, none were HTH/HTH. Moreover, when hth plants were crossed as female with wild-type pollen, no HTH/HTH plant was found among 230 tested (Lolle et al. 2005), thus not confirming the ability of hth plants to produce HTH ovules. To test this ability again, we used previously isolated “revertants” carrying the transgene conferring the resistance to hygromycine. These plants are the results of a cross between the hth-8 mutant as female and hygromycine-resistant plants (HTH/HTH) as male. The genotype of these plants, determined by PCR-based assay (Lolle et al. 2005), was systematically HTH/hth (337/337) and never HTH/HTH. Thus the two crossing experiments using hth plants as female did not reveal any ability of hth mutant to produce HTH ovule (0/230 and 0/337).

The transmission through pollen grains of a wild-type HTH allele from a homozygous mutant (hth/hth) described previously (Lolle et al. 2005) is also not consistent with the outcrossing explanation. We tested this transmission through pollen grains by crossing hth mutants as male with HTH plants as female. To get rid of contamination or self-pollination of the female plant, crosses were done in isolation and a male sterile plant (bm3) (Gaillard et al. 1998) was used as female. The genotype of the progeny was determined by PCR-based assay (Lolle et al. 2005). Under these conditions, 100% (hth-4: 168/168; hth-8: 92/92 ; hth-10: 92/92) of the F₁ progeny was heterozygous for the hth allele. Thus we did not reveal any ability of the hth mutant to produce HTH pollen grains.

In summary,

1. hth plants are highly susceptible to outcrossing.

2. Reversion is dependent on the presence of a wild-type allele donor plant in the vicinity and its frequency is correlated to the distance between hth and donor plants.

3. A revertant systematically carries genetic information that can be provided only by outcrossing.

4. As expected under genetic laws (Mendel 1866; Druery and Bateson 1901), we did not confirmed that hth plants can be a source of pollen or ovule bearing an HTH allele.

Altogether, these results strongly argue for outcrossing as the reason for the apparent genetic instability in hth.

Arabidopsis plants were cultivated in a greenhouse with a photoperiod of 16 hr/day and 8 hr/night, a temperature of 20°, and humidity at 70%. The hth-4, hth-8, and hth-10 mutants were generously provided by R. Pruitt (Purdue University) and are in the Landsberg accession. The hygromycine-resistant plant carried the selection marker hpt (Granger and Cyr 2001) at position 1,095,540 bp on chromosome 5. To measure the frequency of reversion in hth mutants, we grew a F₂ population derived by selfing of hth heterozygous plants and selected hth mutants according to the shape of the first flowers. The first inflorescence was then cut to eliminate any possible prior pollen contamination, notably from heterozygous or wild-type sister plants. hth mutants were then placed either in isolation in a Plexiglas cabinet or at various distances from a line of hygromycine-resistant plants. The offspring of these plants were scored for the wild-type phenotype (“revertant”). The presence of the hygromycine transgene in the “revertant” plants was tested by PCR (5′-TTCCTAAAACCAAAATCCAG-3′ and 5′-ATCAATTGTAGATCCGGCAAACA-3′) with appropriate controls. Genotyping of the HTH locus was performed with specific primers and digestion to reveal the presence or absence of a polymorphic restriction site: hth-8, 5′-TTGGAGAAACTTGCTTACCCGATCT-3′ and 5′-TTGTTTCCAAGTCTCTCCCGAAGAA-3′; ScrF1; hth-4, 5′-CGAAGCTGGTGAAGGAGTCGT-3′ and 5′-GTGACCCAATAGCTCCACTAGATAA-3′; Hpy99I; hth-10, 5′-ATCAACACCAGGCTTTACTCTCGAA-3′ and 5′-TAGAGATCGGGTAAGCAAGTT-3′; MmeI.

• Received August 13, 2008.
• Accepted October 6, 2008.

• Copyright © 2008 by the Genetics Society of America