Session 5: Reproduction I

Chair: Detlef Weigel, Salk Institute, USA

email: detlef_weigel@qm.salk.edu
This session focused on flowering-time and meristem-identity genes.

Detlef Weigel (Salk Institute, La Jolla, USA) discussed preliminary results on the dissection of the promoter of the meristem-identity gene LEAFY (LFY) (Weigel et al., Cell 69: 843-859, 1992) using GUS fusion genes. The LFY promoter is quickly upregulated in response to long photoperiods, paralleling the rapid induction to flowering under such conditions. Surprisingly, there is also a slower upregulation in non-inductive short days, long before flower primordia are initiated. The expression is detected at the shoot apex, specifically in young leaf primordia, which is confirmed by in situ hybridization data that George Coupland showed in his presentation (see below). However, expression levels in young flowers are apparently much higher, and the LFY promoter can be used for ablation of flower primordia when fused to the gene encoding diphtheria toxin A chain. Weigel also discussed the UFO gene, whose cloning has previously been reported by the Haughn and Coen laboratories (Ingram et al., Plant Cell 7: 1501-1510, 1995) . He showed that UFO function requires LFY activity, and that simultaneous constitutive expression of LFY and UFO is sufficient to activate the promoter of the floral homeotic gene APETALA3 in a wide range of tissues outside of flowers.

George Coupland (John Innes Centre, Norwich, UK) discussed the flowering-time gene CONSTANS (CO) that his group has recently isolated (Putterill et al., Cell 80: 847-857, 1995) . CO appears to be ubiquitously expressed, with the level of expression being much higher in long days than in short days, suggesting that CO levels are directly limiting for flowering. This is consistent with the finding that transgenic plants in which CO is constitutively expressed at high levels (35S::CO) flower early, with three leaves in long days and four leaves in short days. Although the delay by short days is nowhere as strong as in wild type, the small delay indicates that other genes contribute to the delay in flowering time under short days. One such gene might be the EARLY IN SHORT DAYS 4 (ESD4) gene, since 35S::CO esd4 double mutants are day-neutral and flower with merely two leaves in both long and short days. Several other genes such as the late-flowering gene FCA, which is thought to act in a parallel pathway, are limiting for floral induction by 35S::CO, as an fca mutation attenuates the 35S::CO phenotype.

To understand CO function in more detail, Coupland and co-workers created a modified version of 35S::CO, in which CO is fused to the ligand binding domain of glucocorticoid receptor, and with which nuclear localization of CO protein can be regulated. Flowering of these plants can be induced at will with the glucocorticoid dexamethasone. Obvious targets of floral induction are the flower-meristem-identity genes LFY and APETALA1 (AP1), and Coupland and co-workers went on to determine how quickly these genes are activated upon induction of the CO gene compared to a shift from short to long days. It turns out that the transcriptional activation of LFY is very similar in both treatments, while the activation of AP1 is considerably slower with the CO fusion than with photoperiod, suggesting that CO defines a branch in the photoperiod-pathway that is LFY specific. The activation of AP1 in the CO transgenics might be mediated by LFY, which would fit with results that Marty Yanofsky (UC San Diego, USA) reported in the Reproduction II session.

Caroline Dean (John Innes Centre, Norwich, UK) reported on the third late-flowering gene that has been cloned, the FCA gene. FCA acts in the constitutively promoting, or autonomous, floral induction pathway. FCA is different from the two previously isolated late-flowering genes LD and CO, both of which appear to encode transcriptional regulators. The predicted FCA protein has two RNA binding domains of about 90 amino acids, and is similar to other developmental regulators such as the Sex-lethal gene of flies. Characteristic for these genes is that they regulate splicing, and that their own splicing is regulated as well. At least the latter part holds true for FCA, as several different, alternatively spliced transcripts can be detected. The gamma-B transcript, which contains the whole open reading frame, accounts for about 40% of FCA mRNAs. So far there is no evidence that the splicing of FCA itself is regulated by floral inductive conditions, but that might not be expected from the mutant phenotype, since fca mutants flower late under both long and short days. An exciting possibility is that FCA splicing is regulated by other flowering-time genes and/or that splicing of other flowering-time genes is regulated by FCA. Dean also reported that ap1 mutants flower earlier than wild-type, and that there is a complex interaction of ap1 mutations with different fca alleles.

Karen Hicks (University of Oregon, Eugene, USA) described an exciting story that might finally provide a molecular link between photoperiod, floral induction and circadian rhythm. Plant physiologists have for a long time invoked a circadian clock in the photoperiodic control of flowering, and Hicks and her colleagues in the Meeks-Wagner laboratory in collaboration with Steve Kay's group (University of Virginia, Charlottesville, USA), found now that the elf3 mutant, which was originally isolated because of its day-neutral early-flowering phenotype (Zagotta et al., Aust J Plant Physiol 19: 411-418, 1992) , is defective in circadian rhythm, as assayed by either leaf movement or activity of a cab2::luc transgene (Millar et al., Plant Cell 4: 1075-1087, 1992) . elf3 mutants are rhythmic in a light/dark cycle or in constant dark, but not in constant light. An additional defect is a failure to suppress hypocotyl elongation, especially in short days, which suggests a link with various photoreceptors, mutations in which cause a similar phenotype. The cloning of ELF3 as well as the search for genetic suppressors and enhancers are well under way.

Guillermo Cardon (Max Planck Institute, Cologne, Germany) discussed an Arabidopsis ortholog of the snapdragon gene encoding SQUAMOSA promoter binding protein 1 (SBP1) of snapdragon, which has been cloned in the Huijser laboratory (Klein et al., Mol Gen Genet 250: 7-16, 1996) . SBP genes in snapdragon as well as the SBP-like (SPL) genes in Arabidopsis constitute a new family of DNA binding proteins. The protein product of SPL3, which maps to chromosome 2 in a region where no floral mutant has been previously localized, binds to an element that is conserved between SQUAMOSA and its Arabidopsis ortholog AP1, and SPL3 RNA expression increases during the life cycle of long-day grown plants. SPL3 transcripts are more widely distributed than that of its putative target, AP1. Transgenic plants that express SPL3 constitutively (35S::SPL3) flower earlier than wild type, and the first flower is often subtended by a bract, which might be interpreted as a transformation of the last lateral shoot into a flower, akin to the shoot-to-flower conversion seen in 35S::AP1 plants. The plants also show other phenotypes, including curling of the leaf margins, carpelloid transformation of first whorl organs in the later formed flowers and a terminal flower. The relationship between the 35S::SPL3 effects and AP1 expression is not clear yet, as AP1 is not ectopically expressed in the transgenic plants, although an ap1 mutation appears to attenuate the 35S::SPL3 phenotype somewhat. Plants expressing antisense SPL3 RNA flower slightly later than wild type and show a modification of pedicels.

The final presentation was by Oliver Ratcliffe (John Innes Centre, Norwich, UK), who reported that TERMINAL FLOWER 1 (TFL1) is the Arabidopsis ortholog of the snapdragon gene CENTRORADIALIS (CEN), which has recently been cloned by Bradley, Carpenter, Coen and colleagues (Bradley et al., Nature 379: 791-797, 1996) . The expression pattern of CEN and TFL1 is similar in plants that have started to flower; both are expressed in subapical regions of the shoot meristem. In contrast to CEN, TFL1 expression is also detected in the vegetative shoot meristem, consistent with the early-flowering phenotype of tfl1, but not cen mutants. Ratcliffe proposed several models for the interaction of TFL1 with the LFY and AP1 genes, and announced that "Des Bradley is a top bloke."