TRICHOMELESS2 is closely related to TCL1
In Arabidopsis, a total of six single-repeat MYB transcription factors have been characterized so far [13]. In order to examine whether there is any uncharacterized single-repeat MYB transcription factors, the entire amino acid sequence of TCL1 was used as a template to BLAST Arabidopsis protein database (http://www.ncbi.nlm.nih.gov). In addition to TRY, CPC, TCL1, ETC1, ETC2 and ETC3/CPL3 that have been reported previously, our analysis revealed that this is another single-repeat MYB transcription factor encoded by At2g30424 in the Arabidopsis genome. The gene encoding this uncharacterized single-repeat MYB was designated TRICHOMELESS2 (TCL2). TCL2 is nearby two other single-repeat MYB genes, ETC2 and TCL1, on chromosome II (Figure 1A), with ETC2 is a tandem repeat gene with TCL2, while TCL2 and TCL1 is separated by another gene At2g30430. Phylogenic analysis using full length protein sequences showed that TCL2 is closely related to TCL1 (Figure 1B). TCL2 shares 80% identify with TCL1 at the amino acid level. As shown in Figure 1C, like TCL1 and other single-repeat MYBs, TCL2 also contains the conserved amino acid signature [D/E]L × 2 [R/K] × 3 L × 6 L × 3R that is required for the interaction with R/B-like bHLH transcription factors[22] and the conserved amino acids that have been shown to be required for the cell-to-cell movement of the single-repeat MYB transcription factor CPC [23].
Expression and subcellular localization of TCL2
To characterize the function of TCL2, we first examined the expression of TCL2 across various tissues and organs. We also compared the expression pattern of TCL2 with that of TCL1 because TCL2 is more closely related to TCL1 than other single-repeat MYBs at the amino acid level. Various tissues and organs of wild-type plants were harvested, and RT-PCR was used to examine the expression of TCL2 and TCL1. We found that TCL2 is expressed in all tissues/organs tested except young siliques, where TCL1 is highly expressed (Figure 2). TCL2 is also highly expressed in cotyledons, organs that normally do not produce any trichomes and where the transcript of TCL1 was at undetectable level (Figure 2). To get more details about when, where and in response to what stimulus TCL2 is expressed, we then searched public available gene expression database genevestigator (http://www.genevestigator.com/gv) using At2g30424 as gene ID, however, results showed that there is no probeset could be found for the gene ID provided, possibly because TCL2 is previously unidentified gene, and no probes for this gene have been printed on microarrays used for all those experiments that provided data for the database.
To examine TCL2's subcellular localization, we generated transgenic plants expressing TCL2-GFP fusion protein under the control of TCL2's own promoter, a genomic DNA fragment that covers the region -1502 to +1 of the start codon of TCL2. The transgenic plant produces reduced number of trichomes (Figure 3A), a phenotype similar to that of PTCL1:TCL1-GFP transgenic plant [13],indicating that the reduction in trichome number is due to an increase on the copy number of the TCL2 gene, as overexpression of TCL2 showed a greatly reduced trichome phenotype (Figure 4). These results also indicate that both the TCL2 promoter and TCL2-GFP fusion protein are likely functional. By using the transgenic plants obtained, we showed that TCL2-GFP proteins are mainly observed in the nucleus of epidermal cells (Figure 3B).
TCL2 is a negative regulator of trichome formation
To analyze the function of TCL2 in trichome formation, we generated transgenic plants overexpressing HA tagged TCL2 under the control of 35S promoter (35S:HA-TCL2). Transcript levels of TCL2 were examined by RT-PCR using HA-specific and TCL2 specific primers (Figure 4A). As observed previously for all other single-repeat R3 MYB transcription factors, transgenic plants overexpressing TCL2 resulted in glabrous phenotypes, with no trichome formation on rosette leaves, inflorescence stems, cauline leaves or floral organs (Figure 4B).
To further analyze the function of TCL2 in trichome formation, we sought loss-of-function alleles of TCL2. By searching the SALK T-DNA Express Database, we found there are four T-DNA insertion mutants that are related to TCL2. However, in all those four lines, the T-DNA was inserted in the promoter regions, and expression of TCL2 is largely unaffected as examined by RT-PCR (data not shown). As a result, all the mutants have wild-type trichome phenotypes.
Therefore, we took RNAi approach to knock down the expression of TCL2. Since TCL2's ORF fragment is relative small, only 303 bp in length, we used the full-length ORF to create the RNAi construct. We selected three independent transgenic lines displaying ectopic trichome formation on inflorescence stems and pedicels (Figure 4B), similar to that of tcl1 mutant [13]. RT-PCR results showed that expression level of TCL2 in the transgenic lines was dramatically reduced (Figure 4A), while expression of other single-repeat MYB genes was largely unaffected (see Additional file 1), indicating that the phenotype observed is caused by down-expression of TCL2.
When we quantified the number of trichomes on internodes and pedicels, we found that compared with tcl1 mutant, there are fewer internodes and pedicels bearing ectopic trichomes in TCL2RNAi mutant (Figure 5).
TCL2 partially rescued tcl1mutant trichome phenotypes
Previously, we identified TCL1 as a major single-repeat MYB transcription factor that negatively regulates trichome formation on the inflorescences and pedicels [13]. The tcl1 mutant confers ectopic trichome formation on the inflorescence stems and pedicels. Here we showed that transgenic plants with reduced expression of TCL2 have similar phenotype as that of tcl1 (Figure 4B, Figure 5). Loss-of-function mutations in any other single-repeat MYBs including TRY, CPC, ETC1, ETC2, and ETC3 do not result in similar inflorescence and pedicel trichome phenotypes as shown in tcl1 mutant or TCL2RNAi lines. Therefore, we wanted to further examine the relationship between TCL2 and TCL1 in trichome formation on the inflorescence stems and pedicels. Double mutant was generated between tcl1 and TCL2RNAi mutants, however, the double mutant was indistinguishable from tcl1 single mutant (data not shown), possibly because higher order redundancy function among single repeat genes [19]. Therefore we turned to test if TCL2 could rescue tcl1 mutant phenotype. Transgenic plants were generated to express TCL2 under the control of TCL1 promoter in a tcl1 background (P
TCL1
:TCL2/tcl1). Our previous results have showed that expression of TCL1-GFP fusion protein under the control of TCL1 promoter fully rescued tcl1 trichome phenotype [13], indicating that the TCL1 promoter used is fully functionally. As shown in Figure 5, expression of TCL2 protein under the TCL1 promoter only partially rescued tcl1 mutant phenotype, indicating that TCL2 is functionally similar but not identical to TCL1 in controlling trichome patterning on the inflorescence stems and pedicels.
TCL2 interacts with GL3
It has been proposed that single-repeat MYB transcription factors control trichome formation by competing with GL1 for binding GL3, thus blocking the formation of TTG1-GL3-GL1 activator complex. We have previously demonstrated that TRY, CPC, TCL1, ETC1, ETC2, and ETC3, interact with GL3 using a protoplast transfection system [19]. To test if TCL2 controls trichome formation using a similar mechanism, we tested if TCL2 interacts with GL3 in plant cells.
A protoplast transient expression system was used to test the interaction between TCL2 and GL3. A Gal4-GUS reporter, together with the effectors GL3 and a Gal4 DNA binding domain (GD) fused TCL2 (GD-TCL2) were co-transfected into Arabidopsis protoplasts. GD alone and GD-TCL1 were used as negative and positive control, respectively. As shown in Figure 6A, with or without GL3, GD alone could not activate the expression of the reporter gene. In the absence of GL3, neither GD-TCL1 nor GD-TCL2 activated the expression of the reporter gene presumably due to the fact that single-repeat R3 transcription factors do not have activation domains. However, in the presence of GL3, both GD-TCL2 and GD-TCL1 activated the reporter gene, indicating that TCL2 interacts with GL3 in plant cells. It should be noted that GL3 can function as a transcriptional activator, however, alone it cannot be recruited to the Gal4 DNA binding elements in the reporter gene [20]. It was also observed that the expression of the reporter gene activated by GD-TCL2 in the presence of GL3 is much higher than that of GD-TCL1 (Figure 6A), implying that the binding affinity between TCL2 and GL3 may be higher than that between TCL1 and GL3.
TCL2 suppresses the expression of GL1
In addition to competing with GL1 for binding GL3, we also reported previously that TCL1 can directly suppress the expression of GL1 [13]. Because TCL2 is more closely related to TCL1 than any other single-repeat MYBs at the amino acid level (Figure 1) and TCL2RNAi lines displayed similar inflorescence and pedicels trichome phenotypes as tcl1 mutants, we then tested if TCL2 also affects the expression of GL1. As shown in Figure 6B, expression level of GL1 is dramatically reduced in 35S:HA-TCL2 seedlings, similar to that in 35S:HA-TCL1 seedlings, suggesting that TCL2 can also suppress the expression of GL1.
Neither GL1-GL3 activator complex nor SPLs alone activates TCL2
Available evidence suggested that TTG1-GL3-GL1 activator complex activates both GL2 and some single-repeat MYB genes. Our previous reporter showed that GL1 and GL3 are required and sufficient to activate GL2 and a subset of single-repeat MYB genes including TRY, CPC, ETC1 and ETC3, but not ETC2 and TCL1[19, 20]. To test if TCL2 is regulated by GL1-GL3 complex, GL1 and GL3 were co-transfected into protoplasts and RT-PCR was used to examine the expression of TCL2. Expression of CPC was used as a positive control. As expected, CPC was strongly induced by GL1 + GL3. However, such activation was not observed for TCL2 (Figure 7A).
We wanted to further investigate the molecular mechanism controlling the activation of TCL2 transcription. Recently, it has been showed that TCL1 is directly activated by MIR156 directed SPLs [21]. In order to explore the possibility that expression of TCL2 may also be controlled by MIR156 directed SPLs, we tested the expression of TCL2 in 35S:MIR156 plant. Indeed, we found that the expression level of TCL2 in 35S:MIR156 plant was dramatically reduced (Figure 7B), similar to that of TCL1 [21]. To further test if SPLs can directly active TCL2, we cloned five SPL genes including SPL3, SPL9, SPL10, SPL13 and SPL15, transiently expressed them in protoplasts, and then used RT-PCR to examine the expression of TCL2. The five SPLs were chosen because all of them but SPL15 have been shown to regulate trichome formation [21], while SPL15 is closely related to SPL9 [24]. To our surprise, no significant changes were observed (data not shown). Since over-expression of SPLs in plants has already been shown to be able to induce the expression of TCL1, and SPL9 has been shown to directly bind to the promoter region of TCL1 [21], we test if the expression of TCL1 in protoplasts transfected with SPLs is affected. However, we did not see any significant changes (data not shown). These results suggested that SPLs alone may not be sufficient to activate the expression of TCL1. To explore this possibility further, all five SPLs cloned were fused with GD and co-transfected with Gal4:GUS reporter gene into protoplasts, then GUS activities were assays. Indeed, none of the five SPLs tested could activate the reporter gene (data not shown). These results indicated that although SPLs can bind directly to the promoter region of TCL1, they might require co-activators to regulate the expression of their target genes, including TCL1, TRY and possibly, TCL2.