Transcriptional Activation of APETALA1 by LEAFY

See allHide authors and affiliations

Science  23 Jul 1999:
Vol. 285, Issue 5427, pp. 582-584
DOI: 10.1126/science.285.5427.582


Plants produce new appendages reiteratively from groups of stem cells called shoot apical meristems. LEAFY (LFY) and APETALA1 (AP1) are pivotal for the switch to the reproductive phase, where instead of leaves the shoot apical meristem produces flowers. Use of steroid-inducible activation of LFY demonstrated that early expression of AP1 is a result of transcriptional induction by LFY. This AP1 induction is independent of protein synthesis and occurs specifically in the tissues and at the developmental stage in which floral fate is assumed. Later expression of AP1 appears to be only indirectly affected by LFY.

The above ground body plan in higher plants is generated postembryonically by a group of undifferentiated stem cells, the shoot apical meristem (SAM). Initially, the Arabidopsis thaliana SAM produces leaves with axillary, second-order shoot meristems. At the transition to the reproductive phase, the primary shoot switches to the production of flowers. Two meristem identity factors, LEAFY (LFY) and APETALA1 (AP1), are necessary and sufficient for this transition (1–4). Severe disruption of the onset of reproduction is observed in the loss-of-function lfy-6mutant; most flowers are replaced by leaves and second-order shoots (3). In the strong ap1-1 mutant, flowers have partial shoot character (1). The gain-of-function phenotype produced by constitutive expression of either LFY or AP1 results in formation of flowers or leaves and flowers in positions normally occupied by leaves and second-order meristems (2,4). The LFY protein localizes to the nucleus, and LFY binds to a putative AP1 promoter element in vitro (5). AP1 is a potential transcriptional target of LFY because it acts, in part, downstream of LFY (1–7). Moreover, AP1 expression is delayed and reduced in lfymutants [our data and (2, 8–10)], whereas constitutive ectopic expression of LFY results in precociousAP1 expression (5). However, these data do not allow separation of direct transcriptional activation by LFY from downstream effects that influence gene expression. To test whether LFY acts as a transcriptional activator in vivo, we constructed a steroid hormone–inducible, posttranslational LFY switch (11). This switch allows us to monitor the immediate effect of LFY activation on transcription, in the presence of translational inhibitors.

We transformed plants segregating for the sterilelfy-6–null mutation with a constitutively expressed translational fusion of LFY to the rat glucocorticoid receptor hormone binding domain [35S::LFY-GR (12)]. Using antiserum to LFY (13), we identified several lines expressing full-length fusion protein (Fig. 1A). The amount of LFY-GR protein detected in the nuclei increased after dexamethasone hormone treatment (14) (compare Fig. 1C with Fig. 1B). Thus, activation of the fusion protein results in proper subcellular localization of LFY-GR (Fig. 1D) (5, 15).

Figure 1

A biologically active LFY switch. Immunochemical detection of the LFY-GR protein on immunoblot (A) and in fixed tissue (B to D) with polyclonal antibody to LFY (13). (A) The full-length fusion protein is overexpressed in inflorescence extracts. The genotypes tested were lfy-6 35S::LFY-GR (lanes 1 and 2) and 35S::LFY (lane 3). Equal volumes of crude extract were loaded. Plants were untreated. For immunolocalization, 17-day-oldlfy-6 35S::LFY-GR seedlings were mock-treated (0.1% ethanol) (B) or treated with 5 μM dexamethasone (Sigma) in 0.1% ethanol (14) (C). Insets show cells of (B) and (C) at higher magnification (1400-fold). (D) Immunolocalization of untreated 17-day-old 35S::LFY seedlings. Arrowheads point to cells showing background (B) or increased levels (C and D) of staining for the LFY protein in nuclei. Original magnification, ×320 (B) and ×480 (C and D). (E) A mature flower in untreated lfy-635S::LFY-GR plants consists of sepals and carpels, lacks petals and stamens, and is infertile. (F and G)lfy-6 35S::LFY-GR inflorescences were treated with 5 μM dexamethasone after bolting as described (20), and flowers were observed 7 days after treatment. (F) Partially rescuedlfy-6 35S::LFY-GR flowers have sepal-stamen mosaic organs (asterisk) as well as some stamenoid organs (arrowheads). (G) Fully rescued lfy-6 35S::LFY-GR flowers have a normal complement of all floral organ types: four sepals, four petals, six stamens, and two fused fertile carpels. (H andI) Seedlings of the genotype lfy-6/+ 35S::LFY-GR or +/+ 35S::LFY-GR were treated at day 7 (H) and day 10 (I) by a single, 2-hour submersion in 5 μM dexamethasone, and the phenotypes were observed after bolting. (H) An axillary inflorescence meristem acquired floral fate and generated a single flower (arrow). (I) A secondary inflorescence meristem adopted floral fate and generated a single flower, which is subtended by a specialized leaf (bract). Mock-treated inflorescences and seedlings, as well as dexamethasone-treated plants not expressing the LFY-GR fusion protein, were indistinguishable from untreated siblings.

In lfy-6–null mutants, floral organs that require the expression of the class B homeotic genes (petals and stamens) are absent (3, 16). To test whether the fusion protein is biologically active, we followed the development oflfy-6 35S::LFY-GR flowers after dexamethasone treatment. As expected, petals and stamens were also absent in untreated lfy-6 35S::LFY-GR flowers (Fig. 1E). This defect was partially or fully rescued after hormone treatment (Fig. 1, F and G, respectively). In addition, hormone treatment of seedlings resulting from an outcross of lfy-635S::LFY-GR to the Ler wild type reproduced characteristic LFY gain-of-function phenotypes (4) in that second-order shoots were converted to flowers (Fig. 1, H and I). Similarly, treatment of lfy-6 35S::LFY-GR seedlings caused conversion of secondary shoots to flowers (17). These data demonstrate that the LFY switch we constructed is functional.

To test whether LFY acts as a transcriptional activator in vivo, we monitored AP1 expression inlfy-6 inflorescences after LFY-GR activation using in situ hybridization (14). Early AP1 expression in the wild type is first observed in young stage 1 (18) flowers immediately after the transition to flowering (Fig. 2H) (19). By contrast,AP1 is absent from stage 1 flowers in lfy-6inflorescences immediately after the transition to flowering (compareFig. 2, C and H) (2, 9, 10). After hormone induction in lfy-6 35S::LFY-GR plants at the floral transition (14), we detectedAP1 RNA in early stage 1 primordia (Fig. 2D) as well as in stage 1 primordia that are formed in the axils of bracts (Fig. 2, B and E to G). This expression was not observed in mock-treated siblings (compare Figs. 2, A and B, and 2, C and D). Thus, activation of LFY-GR in lfy-6 mutant plants causes rapid changes inAP1 mRNA toward the expected wild-type expression pattern. These changes occur in the tissues and at the developmental stage when floral fate is first assumed, suggesting that LFY might act as a transcriptional activator of AP1.

Figure 2

Early AP1 expression in response to LFY activation in lfy-6 35S::LFY-GR. Presented are representative results of eight independent induction experiments analyzed by four separate in situ hybridization experiments. Tissues were fixed, sectioned, and hybridized as described (29). Developmental stages of plant and specifics of the hormone treatment are described (14). (A and B) Longitudinal sections of lfy-6 35S::LFY-GR seedlings. (C and D) Longitudinal sections oflfy-6 35S::LFY-GR inflorescences. (E toG) Serial transverse sections of lfy-635S::LFY-GR inflorescences. Young axillary inflorescences and seedlings were treated with solution without hormone (mock) (A and C) and 5 μM dexamethasone (dex) (B, D, and E to G). Plants were treated twice, at time 0 and again after 4 hours. They were harvested 6 or 8 hours after the first treatment and processed immediately. Sections in (A) and (B) as well as those in (C) and (D) and those in (E) to (G) are from the same in situ hybridization slides and are therefore fixed, probed, exposed, and developed identically. (H) Wild type was not treated (−). Early stage 1 (e1) and stage 1 (1) flowers are indicated throughout. Arrowheads point to sites of early AP1 expression. The inflorescence architecture is different in young lfy-6 meristems compared with wild-type meristems as all floral primordia in the lfy-6 mutants arise in the axils of bracts. Dexamethasone treatments of seedlings without the LFY-GR fusion protein showed AP1 expression indistinguishable from mock-treated seedlings. Original magnification, ×160.

To be able to determine whether AP1 induction occurs independently of protein synthesis, we first retested our procedure for cycloheximide inhibition of protein synthesis (20) in the 35S::LFY-GR genotype. Protein synthesis–dependent GUS staining was analyzed in an AP3::GUS transgenic line, where the LFY-regulated AP3 promoter (21) controls the production of the RNA for a bacterial beta-glucuronidase (GUS). GUS staining increased when lfy-6 35S::LFY-GR inflorescences (14) were treated with dexamethasone alone but not when cycloheximide was applied together with dexamethasone (compareFig. 3B with Fig. 3, A, C, and D), indicating that protein synthesis is abolished efficiently. This finding is further supported by experiments with a heat shock–inducible promoter fused to the GUS reporter in seedlings and inflorescences (20) (see supplementary figure, available

Figure 3

Increased GUS activation in AP3::GUS plants in response to dexamethasone only in the absence of protein synthesis inhibitors. Young axillary inflorescences of lfy-635S::LFY-GR AP3::GUS plants (14) were mock-treated (A) or treated with 5 μM dexamethasone (dex) (B), 10 μM cycloheximide (cyc) (C), or 5 μM dexamethasone plus 10 μM cycloheximide (dex cyc) (D). GUS assays were performed as described (30). Original magnification, ×20.

We next tested whether AP1 is an immediate target of LFY in vivo by analysis of AP1 expression after LFY activation in the absence of protein synthesis. We observedAP1 induction in young lfy-635S::LFY-GR flower primordia (Fig. 4, C and D) in the presence of hormone and cycloheximide. This induction was not observed when only cycloheximide was used (compare Figs. 4, A and B, with Fig. 4, C and D). Thus, LFY activation in lfy-6 mutants results in early expression of AP1 independent of new protein synthesis, indicating that AP1 is a direct transcriptional target of LFY.

Figure 4

AP1 is an immediate target of LFY. Experimental setup is as described in Fig. 2. Longitudinal sections of seedlings that were treated with (A and C) 10 μM cycloheximide (cyc) or (B and D) 5 μM dexamethasone plus 10 μM cycloheximide (dex cyc). Sections in (A) and (B) as well as those in (C) and (D) are from the same in situ hybridization slides. Original magnification (A to D), ×160. (E) Autoradiograph of two representative in situ slides after overnight exposure. (F to H) RT-PCR analysis of five axillary inflorescences treated with solution without hormone (M), 5 μM dexamethasone (D), 5 μM dexamethasone plus 10 μM cycloheximide (DC), and 10 μM cycloheximide (C). (F) RT-PCR with AP1 primers. (G) RT-PCR with ubiquitin primers (14). PCR products were subjected to electrophoresis on agarose gels and after transfer to nylon membranes were probed with the labeled PCR product. Triplicate experiments were quantitated with a phosphorimager (H). The mean of the AP1 values corrected by normalization with the ubiquitin (Ub) lanes and their standard error are shown.

After the transition to flowering, AP1 was expressed in lfy-null mutants in more mature stage 3 flowers (Figs. 2C and 4C), although the expression level was reduced compared with the wild type (2, 9, 10). This laterAP1 expression is thus quantitatively dependent on LFY but does not require LFY activity. We measured AP1 expression in entire inflorescences, which monitor primarily the more abundant laterAP1 expression, using both semiquantitative in situ hybridization and reverse transcription followed by polymerase chain reaction (RT-PCR) (Fig. 4, E to H). We detected a general increase ofAP1 mRNA levels after LFY activation in hormone-treated compared with mock-treated lfy-6 35S::LFY-GR plants. Unlike the AP1 induction in early flower primordia, this general increase was only observed when no protein synthesis inhibitor was present (Fig. 4, E to H), suggesting that this response is controlled indirectly by LFY (22).

Using inducible LFY activity, we were thus able to show that AP1 is an immediate target of transcriptional activation by LFY. This finding is consistent with the observed temporal and spatial expression pattern of both genes. AP1 is expressed later and within theLFY expression domain until stage 3 of flower development (3, 5, 19,23–25). However, after photoinduction and activation of upstream regulators, lag times of up to 56 hours have been observed between the onset of LFY and AP1expression (23, 24). These lag times could be explained by postulating that a critical threshold level of LFY is required for AP1 induction. Consistent with this notion, LFY expression increases during vegetative development (26). A second possibility, not mutually exclusive with the first, is that a function independent of LFY marks time or developmental stage, ensuring that AP1 induction by LFY occurs at the right time and place in development. A similar model appears to be true for targets of the floral homeotic geneAPETALA3 (20). In support of this second hypothesis, we did not detect AP1 induction using in situ hybridization when we activated LFY in lfy-635S::LFY-GR plants before the floral transition (14, 17).

We conclude that LFY acts as a transcriptional activator and that it exerts its meristem identity activity in part by direct transcriptional activation of AP1.

  • * Present address: Molecular Genetics Department, John Innes Centre, Norwich NR4 7UH, UK.

  • To whom correspondence should be addressed. E-mail: meyerow{at}


View Abstract

Navigate This Article