Reports

The molecular basis of natural variation has been defined for relatively few traits. We have focused on natural variation in vernalization, the acceleration of flowering by prolonged cold that aligns flower and seed development with favorable environmental conditions. Vernalization is an inherently quantitative process, with increasing periods of cold giving progressive acceleration of flowering. In A. thaliana this is the result of cell-autonomous stochastic switching of the FLC locus to an epigenet-ically silenced state, with longer cold increasing the probability of that switch occurring (1). Extensive natural variation exists in A. thaliana accessions for how much cold is required to give complete epigenetic silencing of FLC (2, 3). In some accessions, a vernalization period of 4 to 6 weeks of cold gives incomplete FLC silencing; the gene then reac-tivates after transfer back to warm conditions, and flowering is delayed. This variation has been mapped to several major loci including FLC itself (3, 4). To better understand the molecular basis of this natural variation, we focused on two vernalization-requiring A. thaliana genotypes: Lov-1, an accession collected from the northern range of the species in northern Sweden (62.5°N) (5), and FRI Col, the reference Columbia accession containing an introgressed active FRIGIDA allele (6). Lov-1 requires 9 weeks of cold to fully saturate the vernalization requirement, thereby epigenetically silencing FLC and enabling activation of the A. thaliana floral promoter FT (Fig. 1A). Partial vernalization of Lov-1 (4 weeks of cold) led to incomplete epigenetic silencing and reactivation of the Lov-1 FLC allele upon transfer of plants back to warm conditions, no activa-tion of FT, and thus late flowering. In contrast, 4 weeks of cold was sufficient to fully saturate the vernalization requirement in FRI Col, enabling FT activation and early flowering in favorable conditions (Fig