Dramatic changes in gene expression occur at the apex of the plant on the transition to flowering. Recently, these changes were analyzed using microarrays containing oligonucleotides derived from almost all Arabidopsis genes, and within 7 days of a shift from short to long day lengths, the expression of 332 genes was induced or repressed at least twofold in both the Landsberg erecta and Columbia accessions (Schmid et al., 2003). Perhaps surprisingly, of these 332 genes, the majority (231) showed reduced expression during floral induction, whereas the remainder were increased. At least one of the repressed genes, encoding an AP2-like transcription factor, was shown to actively repress flowering when overexpressed, suggesting that its downregulation during flowering may play an active role in floral induction (Schmid et al., 2003). Similarly, some of the earliest genes that show increased expression during floral induction appear to promote flowering. For example, the SOC1 gene of Arabidopsis is induced at the meristem within a few hours of shifting plants from short to long days (Borner et al., 2000; Samach et al., 2000), and the orthol-ogous gene in Sinapis alba behaves similarly (Bonhomme et al., 2000). Mutations in SOC1 delay flowering, but do not have an effect on floral development (Borner et al., 2000; Lee et al., 2000; Samach et al., 2000). In contrast, AP1 and LFY, two genes that confer floral meristem identity on the developing floral primoridum, are among the first genes with established roles in floral development whose transcripts increase in abundance at the meristem (Mandel et al., 1992; Weigel et al., 1992). Mutations in these genes cause shoots or flowers with shoot characters to develop from primordia that would normally form flowers, and the proteins encoded by these genes are involved in activation of downstream transcription factors that confer organ identity on floral organs (Weigel & Meyerowitz, 1993). Recent work indicates that LFY and orthologous genes in other species, as well as some of the transcription factors involved in the development of floral organs, can act non-cell autonomously to influence floral organ development (Wu et al., 2003). In this section, I review these data and describe the mechanisms thought to underlie the non-cell autonomy.
7.3.1 Some of the transcription factors that control floral meristem or organ identity act non-cell autonomously in the developing flower
The first molecular evidence for non-cell autonomy of genes involved in floral development came from studying floricaula (flo) mutants in Antirrhinum majus. The FLO and LFY genes are orthologues and have similar functions in Antirrhinum and Arabidopsis, respectively (Coen et al., 1990; Weigel et al., 1992). FLO was originally cloned based on the insertion of a transposon so that excision of the transposon restores FLO activity (Coen et al., 1990). Excision of the transposon can generate somatic sectors so that a single plant contains both mutant and wild-type cells. The shoot meristem and flowers that form on the flanks of the meristem consist of three cell layers, the L1, L2 and L3, which represent separate clones formed during embryogenesis (Huala & Sussex, 1993; Wu et al., 2003). FLO mRNA is expressed in all three layers, but a transposon excision in the flo mutant can generate a wild-type clone of cells that restores gene function in a single layer. Restoration of FLO gene function in the L1 allowed the formation of phenotypically wild-type flowers, whereas FLO function in the L2 or L3 caused the development of flowers that still showed some aspects of the mutant phenotype (Carpenter & Coen, 1995; Hantke et al., 1995). In flo mutants, the expression of the MADS box transcription factors PLE, which is involved in stamen and carpel development, and DEF, which is involved in petal stamen and development, is reduced. Analysis of expression of these genes in flowers in which FLO mRNA is expressed in only one layer showed that FLO can induce signaling events that affect other layers. For example, in plants expressing FLO mRNA only in the L1, DEF mRNA expression occurred in all layers in a similar pattern to wild-type plants, although its expression occurred later and at lower levels (Carpenter & Coen, 1995; Hantke et al., 1995).
In Arabidopsis, LFY acts non-cell autonomously in a similar way to FLO in Antirrhinum, but in contrast another floral meristem identity gene, AP1, acts almost exclusively cell autonomously. Expression of LFY in the L1 layer using the ML1 epidermal-specific promoter was sufficient to fully complement the lfy mutant phenotype (Sessions et al., 2000), and some plants showed a phenotype similar to that of plants overexpressing LFY from the CaMV 35S promoter (Weigel & Nilsson, 1995). Also, as shown in Antirrhinum for FLO, expression of LFY in the L1 layer was sufficient to activate expression of the gene encoding the MADS box transcription factors APETALA3 (AP3), required for petal and stamen development, and AGAMOUS (AG), required for stamen and carpel development, in all three layers. However, a quite different effect was observed when AP1 was expressed in the L1 layer in ap1 mutants (Sessions et al., 2000). In these plants rescue of sepal and petal identity occurred in the first and second whorls, but only in the L1 layer. The L2 and L3 layers of the sepals of these plants were more typical of bracts. Similarly, the activation of AP3 and AG in ap1 ML1::AP1 plants was restricted to the L1 layer, indicating that AP1 acts cell autonomously. These observations suggest that LFY, but not AP1, triggers signaling between cell layers in the developing flower.
Non-cell autonomy in floral development is not restricted to floral meristem identity genes that act early in floral development, such as AP1 and LFY, but is also shown by genes that specify the identity of individual organs. An example of this comes from studies with the DEF and GLOBOSA (GLO) genes of Antirrhinum, which are required for normal petal and stamen development (Sommer et al., 1990; Troebner et al., 1992). Transposon-induced reversion of def mutations generated somatic sectors in which activity of the gene occurred only in the L1 (Perbal et al., 1996). This was sufficient to cause expansion of second whorl organs, similar to the expansion shown by petal lobes, and accumulation of pigment in L1 cells similar to the epidermis of petals. However, L2 and L3 cells accumulated chlorophyll and did not show petal identity. Expression of DEF mRNA in the L1 is therefore not sufficient to confer full petal identity on inner layers. These observations were confirmed using transgenic Antirrhinum plants expressing DEF only in the L1 from the ANTIRRHINUM FIDDLEHEAD (AFI) promoter (Efremova et al., 2001). In contrast, sectors in which DEF mRNA was expressed in the L2 and L3 were sufficient to confer petal identity on the L1 layer, but not to promote petal expansion. This observation indicates that DEF can act non-cell autonomously in the inner layers to confer petal identity on the epidermis. Further examples of the analysis of cell autonomy in floral organ identity gene function are those performed in Arabidopsis on AG, which is required for stamen and carpel expression, and on the PISTILLATA (PI) and APETALA 3 (AP3) genes that are required for petal and stamen identity (Bouhidel & Irish, 1996; Sieburth et al., 1998; Jenik & Irish, 2001).
7.3.2 Movement of transcription factors between cells defines one mechanism for short-distance signaling in the developing flower
Signaling between animal cells typically involves activation of a receptor located in the membrane of one cell by a ligand formed in a nearby cell (Pires-daSilva & Sommer, 2003). Until recently short-distance signaling between plant cells during flower development was assumed to be exclusively based on similar processes. However, analysis of KNOTTED (KN), a homeobox transcription factor of maize, demonstrated that transcription factors can move between plant cells (Jackson et al., 1994) and that this may provide an alternative to membrane-bound receptor-based systems in plant cells. The original observation was that although KN mRNA is detected only in the L2 and L3 layers of maize SAM, the protein is found in the L1 (Jackson et al., 1994). The authors proposed that although KN is a nuclear transcription factor, it may nevertheless move between plant cells. This was later confirmed using fluorescently labeled forms of the KN protein (Lucas et al., 1995; Kim etal., 2002, 2003).
The mechanism by which KN moves between plant cells has not been established, but is likely to involve movement through plasmodesmata. These are channels that connect neighboring cells, and consist of an outer membrane derived from plasma membrane and a segment of the endoplasmic reticulum that is continuous between the two cells (Lucas et al., 2001; see Chapter 5). Plasmodesmata are either made at the cell plate during cytokinesis (primary plasmodesmata) or are formed between existing cells (secondary plasmodesmata) (Lucas et al., 2001; Wu etal., 2002). The types of molecule that move through plasmodesmata appear to be tightly regulated by the developmental stage of the plant and the environmental signals to which it is exposed (Ding et al., 1992; Oparka et al., 1999; Kim et al., 2003). Also, the size exclusion limit of plasmodesmata, which defines the size of the molecules that can move through particular plasmodesmata, appears to change depending on the cell type in which they are present and the stage in development at which they are studied. Evidence that KN moves through plasmodesmata is indirect and is based on the observation that expression of the KN protein increases the size exclusion limit of plasmodesmata, allowing the KN mRNA to move to neighboring cells (Lucas et al., 1995). Also, RNAs and proteins encoded by other genes have been shown to move between plant cells through plasmodesmata (Lucas et al., 2001; Wu et al., 2002; see also Chapter 5).
DEF was the first protein involved in flower development that was shown to move between cells (Perbal et al., 1996). In chimeric Antirrhinum plants that expressed the DEF gene only in the L2 and L3 layers of petals, the mRNA was detected specifically in these layers, whilst the DEF protein was present in all three layers. This observation demonstrated that the DEF protein moves from inner layers into the L1 and probably explains the wild-type L1 phenotype shown by plants expressing DEF mRNA only in the L2 and L3. In contrast, in sectored plants expressing the DEF mRNA only in the L1, DEF protein was detected in the L1, but not the L2 or L3 layers (Perbal et al., 1996). This indicated that the DEF protein could not move through plasmodesmata from the L1 to underlying layers, consistent with the cell autonomy of the complementation of the def mutant phenotype. Taken together, these data indicate that trafficking of the DEF MADS box transcription factor occurs between layers but that this movement is polar; it occurs from the L2 or L3 layers to the L1, but not from the L1 to inner layers. Movement of DEF from the L2 to the L1 is likely to occur through secondary plasmodesmata, since primary plasmodesmata, which are formed during cell division, cannot be formed between cells that are not clonally related. However, DEF is also unlikely to be trafficked through primary plasmodesmata, since small, cell autonomous revertant sectors were previously described on Antirrhinum def mutant petals (Carpenter & Coen, 1990).
Trafficking of the LFY transcription factor was tested in transgenic plants in which expression of LFY mRNA from the epidermal ML1 promoter complemented the lfy mutation (Sessions et al., 2000). Although in these plants LFY mRNA was detected only in the L1 layer, the protein was detected in all layers. LFY protein is therefore able to traffick from the L1 to inner layers. Movement of this protein was tested more extensively using fusions with GFP (Wu et al., 2003). Expression of several LFY-GFP fusion proteins from the ML1 promoter complemented the lfy mutation, as shown for the wild-type LFY protein. GFP fluorescence was detected strongly in the L1 but also up to four cell layers deeper into the meristem. A gradient of fluorescence was detected from the L1 into the deeper layers. Previously, proteins were proposed to move through plasmodesmata either using a specific mechanism that utilizes signals within the protein or through nontargeted movement that involves only diffusion (Crawford & Zambryski, 1999). The gradient of fluorescence observed with LFY-GFP fusion proteins suggests that the movement may be non-targeted because proteins that move through a targeted mechanism, such as viral movement proteins, are able to move further and do not show a gradient in abundance in deeper layers. LFY-GFP also did not appear to move laterally within a layer, but moved effectively between layers, suggesting that movement might be restricted to secondary plasmodesmata and not occur between primary plasmodesmata.
During flower development, not all transcription factors move between cells. Expression of API mRNA from the ML1 promoter did not rescue the apl mutant phenotype in underlying layers, and AP1:GFP fusion proteins remained in the L1 cells when expressed from the ML1 promoter (Wu et al., 2003). If LFY protein moves passively between cells through secondary plasmodesmata, then why does AP1, which is a smaller protein than LFY, not move between cells? Several possibilities have been suggested (Wu et al., 2003). Efficient nuclear localization of transcription factors may reduce the possibilities for movement, and there is evidence that AP1:GFP is more efficiently targeted to the nucleus than LFY:GFP fusions. Alternatively, the formation of higher order complexes, similar to those formed by AP1 and other MADS box transcription factors during floral development, might effectively increase the size of the protein and thereby restrict its movement by diffusion.
Short-range signaling between plant cells during floral development may not only involve trafficking of organ and meristem identity proteins. Expression of DEF in the L1 of Antirrhinum partially restores the petal identity of underlying cells, but the protein itself apparently does not move to the L2 (Efremova et al., 2001). This suggests that DEF activates signaling processes in the L1 that influence the identity of the subepidermal cells, but the nature of these signals is not known. These signals may be even more effective at promoting petal and stamen identity in Arabidopsis since epidermal expression of DEF in an ap3 mutant almost completely rescued the effect of the ap3 mutation on petal and stamen identity, but also in this species the AP3 protein was not able to move between cell layers.
How important is transcription factor movement between cells in the development of a wild-type flower? During root development, the SHORTROOT (SHR) mRNA, which encodes a transcription factor of the GRAS family, is expressed only in the stele, but is required for the development of the adjacent endodermis (Helariutta et al., 2000; see also Chapter 8). Analysis of the SHR protein showed that it is present both in the stele and endodermis, indicating that the protein must move from the stele to the endodermis, where it is required for the development of this cell layer and for the activation of the SCARECROW gene (Nakajima et al., 2001).
In contrast, LFY mRNA is expressed in all layers of the flower primordium and within their domains of expression the floral organ identity genes are also expressed in all layers, therefore there is no obvious requirement for protein movement. One possibility is that the induction of transcription of floral meristem identity and floral organ identity genes does not occur reliably within their domain of expression and that short-distance transcription factor movement ensures that gene activity occurs reliably in every cell.
Was this article helpful?