FLOWERING LOCUS C influences the timing of shoot maturation in Arabidopsis thaliana

Many plant species undergo changes in leaf morphology as the vegetative shoot meristem matures (Poethig, 2003; Telfer et al., 1997). In Arabidopsis, juvenile leaves are produced before shoot maturation and are relatively round and lack trichomes on the abaxial side (underside) of the leaves. Adult leaves formed after the shoot meristem matures are more elliptical and have abaxial trichomes. An important difference between the juvenilevegetative and adult-vegetative phases is that only during the adult-vegetative phase can the meristem become competent to respond to environmental and endogenous signals that promote flowering (Telfer et al., 1997). Genetic backgrounds and environmental conditions that delay flowering also delay shoot maturation (Telfer et al., 1997). For example, both flowering and shoot maturation occur later in the Columbia accession (Col) than in the Landsberg errecta accession (Ler) (Lee et al., 1993; Telfer et al., 1997). An important difference between these two accessions occurs at the FLC locus. FLC levels are slightly elevated in Col compared to Ler, which has a weak allele of FLC (Michaels et al., 2003). FLC encodes a MADS-domain transcription factor that acts as a repressor of flowering (Michaels and Amasino, 1999a; Sheldon et al., 1999). Flowering is delayed under conditions or in genetic backgrounds in which FLC expression is elevated (Michaels and Amasino, 1999a). The observation that shoot maturation is delayed in Col relative to Ler suggests that FLC influences the rate of shoot maturation. However, there are many genetic differences between Col and Ler making it difficult to determine whether FLC expression is solely responsible for the delay in shoot maturation observed in Col. To further explore the potential role of FLC in the regulation of shoot maturation, we performed experiments in the Col background using mutations that eliminate FLC function, as well as mutations in known regulators of FLC. This allowed us to create a series of lines containing various levels of FLC expression that could be used to determine if there is a correlation between FLC expression and the timing of shoot maturation. The appearance of abaxial trichomes was used to determine when the transition from juvenile to adult leaves occurred (Telfer et al., 1997). The first leaf with four or more abaxial trichomes was scored as the first adult-vegetative leaf. Leaves were photographed at the time of trichome counting, and those images confirm that the appearance of abaxial trichomes corresponds to leaf shape changes associated with shoot maturation in Arabidopsis (Fig. 1) (Telfer et al., 1997). The number of leaves produced before flowering was also recorded for each genotype and treatment (Table 1). In the Col background, shoot maturation occurred earlier in an flc null mutant (flc-3) than in wild type (ttest, P < 0.0001) (Table 1, Fig. 2a). Thus, FLC acts to delay shoot maturation, and a reduction in FLC expression is sufficient to promote shoot maturation. To test whether increased FLC expression would further delay shoot maturation, we examined the timing of shoot maturation in late-flowering autonomous pathway mutants that have elevated levels of FLC expression (Michaels et al., 2003). Previous work in the Ler background has shown that shoot maturation is slightly delayed by autonomous pathway mutants (Telfer et al., 1997). In the Ler background, lacking a strong allele of FLC, autonomous-pathway mutants are known to have little effect on flowering time (Michaels et al., 2003). Therefore, we examined the effect of autonomous-pathway mutants in Col. FCA is part of the autonomous pathway, and loss of function of FCA delays flowering (Sheldon et al., 2000). Shoot maturation was delayed in the presence of an fca null allele (fca-9) relative to wild type (t-test, P < 0.0001) (Table 1, Fig. 2a). In the flc-3 background, fca-9 did not influence the timing of shoot maturation (flc-3, FCA was indistinguishable from flc-3, fca-9) (t-test, P 5 0.373) indicating that the delay induced by fca was FLC-dependent. Loss-of-function of a different