Lessons from leaf epidermal patterning in plants - Intercellular Communication in Plants - page 231

2003). That the positive regulators interact also with the negative regulators was
shown by two types of experiments: Firstly, several mutant combinations in which
the gene dosage of one of the positive regulators and of TRY was reduced to 50%
(trans-heterozygous) exhibited patterning defects. Secondly, plants over-expressing
GL1 or GL3 showed a much stronger phenotype in the absence of TRY , indicating
that TRY represses them (Schnittger et al. , 1998; Szymanski & Marks, 1998).
The physical interactions between the transcription factors as revealed by yeast
two-hybrid studies provided a more detailed picture. Strong interactions were found
between GL1 and GL3 (Payne et al. , 2000). Using protein fragments, the interaction
sites were mapped to the N-terminal MYB repeats of GL1 and the first 96 amino
acids of GL3. GL3 can also homodimerize by its C-terminal region that contains
the bHLH domains. TTG1 also binds to GL3 but not to GL1 (Payne et al. , 2000).
The regions of GL3 responsible for GL1 and TTG1 binding are different. Given that
TTG1 can be substituted by over-expression of GL1 and GL3, it is likely that TTG1
has a more accessory function, for example in the stabilization of the complex.
The negative regulator TRY was shown to interact with GL3 (Esch et al. , 2003).
Interestingly, TRY binds to the same domain of GL3 as GL1, and competition studies
demonstrated that TRY can compete with GL1 for binding with GL3. This finding
strongly suggests that two alternative complexes can be formed, an active complex
consisting of GL1, TTG1 and GL3 and an inactive one containing TRY, TTG1 and
GL3 (Fig. 9.2A).
A potential target gene of the putative transcriptional complex consisting of
TTG1, GL3 and GL1 is the GL2 gene. On the one hand, it is possible to turn on
GL2 expression ectopically by over-expressing other trichome promoting factors
(Szymanski et al. , 1998). On the other hand, the GL2 phenotype combines several
aspects of other trichome morphogenesis mutants, suggesting that it acts as a master
gene coordinating the proper differentiation of the trichome (Rerie et al. , 1994).
Further studies are needed to clarify the above-mentioned additional role of GL2 in
trichome patterning.

9.3.5 Local cell-cell interactions leading to cell fate decisions: a model

As the position of trichome cells is determined neither by their relative position to
other cells nor by cell lineage, the trichome pattern appears to be created de novo .

Figure 9.2 A model for the selection of trichomes from the protodermal cell pool. (A) GL1, TTG1 and
GL3/EGL3 represent the active complex which turns on downstream target genes to specify trichome
cell identity. TRY/CPC compete with GL1 for binding with GL3/EGL3, forming an inactive complex
which fails to specify trichomes. (B) Initially, all epidermal cells are equivalent expressing the
activators GL1, GL3/EGL3 and TTG1 and begin to communicate with each other via TRY/CPC that
are believed to move from cell to cell (top diagram). A bias in the balance of the activator concentration
is postulated to increase the activity of the activators in one cell. The increased levels of the activator
leads to trichome cell fate determination and causes increased levels of the inhibitor, which in turn
laterally suppresses the neighbouring cells from gaining trichome fate (bottom diagram).