Zinc Finger Protein 6 (ZFP6) regulates trichome initiation by integrating gibberellin and cytokinin signaling in Arabidopsis thaliana

Authors


Author for correspondence:

Yinbo Gan

Tel: +86 571 88982204

Email: ygan@zju.edu.cn

Summary

  • The Arabidopsis trichome is a model system for studying cell development, cell differentiation and the cell cycle in plants. Our previous studies have shown that the ZINC FINGER PROTEIN5 (ZFP5) controls shoot maturation and epidermal cell fate through GA signaling in Arabidopsis.
  • We have identified a novel C2H2 zinc finger protein ZINC FINGER PROTEIN 6 (ZFP6) which plays a key role in regulating trichome development in Arabidopsis.
  • Overexpression of ZFP6 results in ectopic trichomes on carpels and other inflorescence organs. Gain- and loss-of-function analyses have shown that the zfp6 mutant exhibits a reduced number of trichomes in sepals of flowers, cauline leaves, lateral branch and main inflorescence stems in comparison to wild-type plants.
  • Molecular and genetic analyses suggest that ZFP6 functions upstream of GIS, GIS2, ZFP8, ZFP5 and key trichome initiation regulators GL1 and GL3.We reveal that ZFP6 and ZFP5 mediate the regulation of trichome initiation by integrating GA and cytokinin signaling in Arabidopsis. These findings provide new insights into the molecular mechanism of plant hormone control of epidermal trichome patterning through C2H2 transcriptional factors.

Introduction

Trichomes are specialized structures that develop from epidermal cells in the aerial parts of plants (Marks, 1997; Szymanski et al., 2000). The development of the trichome results from an integration of both external signals and endogenous developmental programs (Hülskamp et al., 1994; Pesch & Hülskamp, 2009; Maes et al., 2011). Trichome cell differentiation in Arabidopsis is controlled by several different types of transcription factors (Szymanski et al., 2000; Ishida et al., 2008). Trichome cell patterns in Arabidopsis have been studied by molecular genetic methods for several decades, and regulatory gene networks involved in epidermal trichome development have been identified (Marks, 1997; Pesch & Hülskamp, 2009; Schiefelbein et al., 2009; Balkunde et al., 2010; Grebe, 2012). At the core of these networks are positive and negative regulators of trichome and nonroot-hair cell fate (Szymanski et al., 2000; Ishida et al., 2008). The positive regulators controlling trichome cell fate are an R2R3 MYB-type transcription factor GLABRA1 (GL1; Oppenheimer et al., 1991), which acts, in part redundantly, with MYB23 (Kirik et al., 2005), a basic helix–loop–helix transcription factor GLABRA3 (GL3; Payne et al., 2000), its homolog ENHANCER OF GL3 (EGL3; Bernhardt et al., 2003; Zhang et al., 2003) and a WD40-repeat protein TRANSPARENT TESTA GLABRA1 (TTG1; Galway et al., 1994; Walker et al., 1999; Grebe, 2012). These positive regulators form a complex that activates the expression of a homeodomain protein GLABRA2 (GL2), which promotes trichome formation (Rerie et al., 1994; Pesch & Hülskamp, 2004, 2009). On the other hand, the negative regulators of trichome cell development are small R3 single-repeat MYB transcriptional regulators that can bind to the TTG1/GL1/MYB23/GL3/EGL3 activation complex and may inhibit the formation of activating regulatory complexes (Kirik et al., 2004a,b; Wang et al., 2007, 2008; Wester et al., 2009). These negative regulator genes are TRIPTYCHON (TRY), CAPRICE (CPC; Schellmann et al., 2002) ENHANCER OF TRY AND CPC (ETC) 1, 2, 3 (Kirik et al., 2004a,b) and TRICHOME-LESS (TCL; Wang et al., 2007).

It is known that phytohormones play distinct roles in the regulation of trichome cell differentiation in plants (Traw & Bergelson, 2003; Maes et al., 2011). GA is the first plant hormone molecule reported to regulate the initiation and morphogenesis of trichome development in Arabidopsis (Chien & Sussex, 1996; Telfer et al., 1997; Perazza et al., 1998; Gan et al., 2006, 2007a; Zhou et al., 2011; An et al., 2012). Our previous studies have shown that the C2H2 zinc finger protein coding transcription factor GLABAROUS INFLORESCENCE STEMS (GIS) acts in a GA-responsive pathway to regulate trichome initiation in inflorescence organs (Gan et al., 2006). Further, molecular and genetic analyses demonstrate that GIS acts upstream of GL1 and downstream of SPY (Gan et al., 2006). In addition, like GIS, two GIS homologous genes, GIS2 and ZFP8, mediate the regulation of trichome initiation by a GA signal pathway. Furthermore, GIS, GIS2 and ZFP8, which encode proteins that are largely equivalent in function, play partially redundant and essential roles in inflorescence trichome initiation and in its regulation by GA (Gan et al., 2007a; Zhou et al., 2012). Recently, we reported that a ZINC FINGER PROTEIN, ZFP5, also plays an important role in controlling trichome cell development through GA signaling (Zhou et al., 2011). The molecular analyses suggest that ZFP5 functions upstream of GIS, GIS2, ZFP8, and ZFP8 has been identified as the direct target of ZFP5 in controlling trichome initiation in Arabidospsis (Zhou et al., 2011).

Jasmonic acid (JA) and salicylic acid are also reported to be involved in trichome development (Traw & Bergelson, 2003; An et al., 2011). The application of JA caused significant increases in the number of trichomes on leaves (Lee & Lee, 2001; Traw & Bergelson, 2003). Salicylic acid decreases trichome number (Bowling et al., 1997; Traw & Bergelson, 2003). Cytokinins also stimulate trichome development on the inflorescence stem (Greenboim-Wainberg et al., 2005; Gan et al., 2007a). The control trichome production by cytokinins requires two genes expressed in the late development of inflorescence organs, ZFP8 and GIS2. GIS2 and ZFP8 also interacted with GIS to integrated GA and cytokinin signals in the regulation of trichome cell differentiation (Gan et al., 2006, 2007a). By contrast, GIS does not play a significant role in the cytokinin pathway (Gan et al., 2007a). Here we report the role of a novel C2H2 zinc finger protein, ZINC FINGER PROTEIN 6 (ZFP6), which belongs to the GIS subfamily and plays a key role in trichome cell differentiation in Arabidospsis by integrating GA and cytokinin signaling.

Materials and Methods

Plant materials and growth condition

Arabidopsis thaliana (L.) Heynh. ecotype Col-0 was used as a wild-type (WT) control for all experiments in this study. The plants were grown in a controlled growth room with the following growth conditions: 20–23°C, 95–120 μmol m−2 s−1, 16 : 8 h photoperiod and 70–81% humidity. For phenotype analysis experiments, seeds were surface-sterilized with 5% (v/v) NaClO solution for 7 min and washed at least five times in sterile distilled water, then plated on MS medium and vernalized at 4°C for 3 d in the dark and then placed in a growth room as previously described (Zhou et al., 2011). Trichome production analysis experiments on different organs were carried out as previously described (Gan et al., 2006, 2007a).

Isolation of a ZFP6 loss-of-function mutant and zfp5zfp6 double mutant

A ZFP6 mutant with a T-DNA insertion in the promoter region was identified in the Nottingham Arabidopsis Stock Centre (NASC) line collection (catalog number N118188). Homozygous mutants were screened by ensuring that all of their progenies were resistant to Basta (35 mg l−1). The presence of the T-DNA insertion at the ZFP6 promoter was confirmed by PCR using gene-specific primers (5′ primer, 5′-TTCCGATGACTGGTTTTGAAC-3′; 3′ primer, 5′-AAGGTCATGCATGAGTTCAGG-3′; and a T-DNA insertion primer, Spm1: 5′-CTT ATT TCA GTA AGA GTG TGG GGT TTT GG-3′. The zfp5zfp6 double mutant was selected from F2 populations by double selection on both kanamycin (50 mg l−1) and Basta (35 mg l−1), and further PCR experiments verified the homozygous zfp5zfp6 double mutant.

Cloning constructs

For all cloning constructs of overexpression and RNAi inhibition, expression (35S::ZFP6, ZFP6-RNAi,) sequences were first amplified and inserted into the pENTR-1A vector (Invitrogen) before being recombined into an appropriate destination vector using the Gateway LR reaction system (Invitrogen). All destination vectors were obtained from the Flanders Interuniversity Institute for Biotechnology (VIB) as previously described (Gan et al., 2006; Zhou et al., 2011). The destination vector pH2GW7 was used to prepare the 35S::ZFP6 construct and the destination vector pK7GWIWG2(II) was used for the ZFP6-RNAi construct. For all cloning experiments, gene-specific fragments were first amplified by PCR from cDNA (35S::ZFP6, ZFP6-RNAi) using primers containing SalI and NotI restriction sites, then purified using a gel extraction kit (Qiagen) before restriction digestion and cloning. The following primers were used: ZFP6 overexpression, 5′-GGTCGACGCTCACATATATGGCGACTGAA-3′ and 5′-AAGCGGCCGCTAATCATTCATGGCCCAAGG-3′; ZFP6-RNAi, 5′-TGGTCGACGGTTTGCCACGTCATCATAA-3′ and 5′-TGATGTCGACTGGTTTGCCACGTCATCATA-3′. For the production of pZFP6 promoter fusion constructs, a 1.8 kb promoter fragment was amplified using the primers 5′-AGTCGACTGCAACGTAAAAGGTCCTGA-3′ and 5′-AAGCGGCCGCCAAAGAGAGAGAGAGAGACAGATTT-3′ and substituted for the 35S promoter in pH2GW7 after restriction with SacI and SpeI. The coding sequences of ZFP6 were then recombined into a modification of the same entry vector that was used for the production of overexpression constructs. For all of the plant-transforming experiments, all binary vectors were transformed into the Agrobacterium GV3101 strain, which was then used to transform WT A. thaliana using the floral-dip method as described by Clough & Bent (1998). Transgenic seedlings were first selected on MS agar plates with appropriate antibiotics and then verified and confirmed by PCR using the corresponding primers.

GA and cytokinin treatments

Plant hormones GA3 (Sigma) and 6-benzylamino-purine (BA, Sigma) were used in all experiments that involved external GA and cytokinin treatments, as previously described (Gan et al., 2006, 2007a; Zhou et al., 2011). For induction experiments, the plants were then sprayed with either 100 μM GA3 or 100 μM 6BA (for cytokinin induction experiments) or a corresponding negative control (mock solution). The inflorescence organs of sprayed plants were harvested for RNA extraction at 2 and 4 h (GA or cytokinin) after treatment. For hormone-treated trichome phenotype experiments, control, ZFP6-RNAi and mutant plants were grown in soil, and after the first three to four leaves had emerged, the plants were sprayed twice a week with GA3, BA solutions or a negative control until they were harvested for analysis as described (Gan et al., 2007b; Zhou et al., 2011).

RNA extraction and real-time PCR

Total RNA was extracted from Arabidopsis organs with TRIzol Reagent (Invitrogen), and cDNA was synthesized from 3 μg of total RNA using M-MLV transcriptase (Promega) and oligold (T18) primers in a 25 μl reaction according to the manufacturer's instructions. For real-time PCR, the cDNAs were diluted to 100 μl, and 2 μl was added to 12.5 μl of SYBR-green PCR mix (Takara, Tokyo, Japan) and 0.5 μl of each primer (200 nM final concentration) in 25 μl reactions. PCR was performed using a Stratagene M×3005P thermal cycler (California, USA) as previously described (Zhou et al., 2011). UBQ10 was used as an endogenous control gene to normalize expression of the other genes according to Gan et al. (2006). Relative expression levels were calculated by subtracting the threshold cycle (Ct) values for UBQ10 from those of the target gene (to give △Ct) and then calculating math formula as previously described (Gan et al., 2005). Reverse transcription PCR (RT-PCR) experiments were performed at least once on three independent samples.

Statistical analyses

All data shown in the table and figures were tested by means of ANOVA for significance using the Statistix program Version 3.5 (Analytical Software, St Paul, MN, USA). A Student's t-test was calculated at the probability of either 5% (*, < 0.05) or 1% (**, < 0.01) as previously described (Gan et al., 2010, 2011).

Results

Overexpression of ZFP6 increases trichome production in inflorescence stems and flowers

Sequence analysis has shown ZFP6 is most related to ZFP5 (Gan et al., 2006). In order to investigate whether C2H2 zinc finger protein ZFP6 plays a role in trichome initiation, we created 35S::ZFP6 transgenic lines and found that overexpression of ZFP6 resulted in an abnormally high density of trichomes on inflorescence organs and flowers (Fig. 1a,b). In addition, ZFP6 overexpression caused the formation of ectopic trichomes on carpels and petals (Fig. 1b). We selected two representative overexpressed transgenic lines, line 8 and line 14, which exhibited much higher levels of ZFP6 overexpression thanthe WT plants (Supporting Information, Fig. S1A). Both lines had significantly increased trichome numbers in second cauline leaves, second lateral branches and main inflorescence stem compared with WT plants (Fig. 2a).

Figure 1.

Phenotypes of 35S::ZFP6 overexpressors. (a) Main inflorescence stems of the Arabidopsis thaliana wild-type (WT) and 35S::ZFP6 overexpression transgenic line. (b) Floral organs from the wild-type and 35S::ZFP6 transgenic line.

Figure 2.

Trichome initiation in inflorescence organs of Arabidopsis thaliana zfp5, zfp6, ZFP6-Ri-13, and 35S::ZFP6 lines, and the zfp5 zfp6 double mutant. (a) Trichome density on inflorescence organs and rosette leaves in two 35S::ZFP6 lines, O-ZFP6-8 and O-ZFP6-14. Trichome density on the second and third branches (b), the second and third cauline leaves (c), and sepals and main stem internodes (d) in wild-type (WT) plants, zfp5, zfp6, ZFP6-Ri-13, and zfp5zfp6 double mutant. O, overexpresser; control, the transformed empty vector; lat. br., lateral branch; caul., cauline leaves. Trichome number was calculated from at least 20 plants. Error bars represent + SE. A Student's t-test was calculated at the probability of 1% (**, < 0.01).

Loss-of-function of ZFP6 inhibits the inflorescence trichome initiation

In order to explore the function of ZFP6 in trichome initiation, we selected a T-DNA insertion ZFP6 mutant (catalog number N118188) from the NASC. Genomic PCR was performed to verify the presence of the T-DNA at the expected location. We also constructed a vector for silencing the expression of ZFP6 using an RNAi strategy. We generated eight ZFP6-RNAi lines and all of them showed similar phenotypes. Quantitative PCR results showed that ZFP6 expression was significantly suppressed in the ZFP6 mutant and ZFP6-RNAi-13 line (Fig. S1B). To study trichome initiation patterns in the ZFP6 mutant further, we first examined trichome distribution on the main stems, lateral branch, cauline leaves and flowers in the zfp6 mutant, ZFP6-Ri-2 and zfp5zfp6 double mutant, and compared these with the WT plants. The results showed that in the zfp6 mutant and the ZFP6-Ri-2 and zfp5zfp6 double mutants, trichome production was not affected in the first lateral branch, the first cauline leaf or the first internode, but was significantly decreased in successive stem internodes, lateral branches, cauline leaves and sepals, compared with WT plants. (Fig. 2b–d).

ZFP6 controls trichome cell differentiation during inflorescence development through GA signaling

In order to study whether ZFP6 controls the trichome initiation through GA signaling in the same manner as ZFP5, we first checked whether the expression of ZFP6 in inflorescence organs of the WT could be induced by the external application of GA to the WT plants in the first instance, and also to the GA deficiency mutant ga1-3. WT and ga1-3 mutant plants were sprayed with 100 μM GA and harvested 4 and 6 h after GA application. As shown in Fig. 3(a), the expression of ZFP6 was significantly higher in GA-treated plants than the negative controls in both WT and gal-3 backgrounds. This result suggests that ZFP6 expression is induced by GA. As we have shown before, external GA application could promote trichome production on leaves and inflorescence organs (Chien & Sussex, 1996; Gan et al., 2006; Zhou et al., 2011). In order to check whether zfp6 responds to external GA application, WT, zfp5, zfp6, and zfp5zfp6 plants were treated with 10 and 100 μM GA as described previously (Gan et al., 2006; Zhou et al., 2011). The results indicate that trichome initiations in the lateral branch (the second and third) and cauline leaves (the second and third) are less sensitive than WT to GA application in zfp5, zfp6, and zfp5zfp6 (Fig. 4a–d; Table S1). This suggests that ZFP6, just like ZFP5, is required for the GA signaling pathway that stimulates trichome initiation in inflorescence organs (Fig. 4a–d).

Figure 3.

Expression of ZFP6 in response to GA and cytokinin applications. (a) The relative expression levels of ZFP6 in inflorescence organs of Arabidopsis thaliana wild-type (WT) and gal-3 after GA treatment (100 μM) for 4 and 6 h. (b) The relative expression levels of ZFP6 and ZFP5 in inflorescence organs of the WT after 6-benzylamino-purine (BA) treatment (100 nM) for 2 h. (c) The relative expression levels of ZFP8, GIS, GIS2 and GL1 in inflorescence organs of zfp6 mutant after GA treatment (100 μM) for 4 h. (d) The relative expression levels of ZFP8, GIS, GIS2 and GL1 in inflorescence organs of the zfp6 mutant after BA treatment (100 nM) for 2 h. ARR5 was used as the indicator gene for cytokinin response. Transcript abundance was measured by real-time PCR and the values were normalized against the levels of UBQ10 as a housekeeping gene. At least eight plants were sampled for gene expression analyses. Error bars represent + SE. A Student's t-test was calculated at the probability of 1% (**, < 0.01).

Figure 4.

The effect of GA application on trichome initiations in Arabidopsis thaliana wild-type (WT), zfp5, zfp6 and zfp5 zfp6 double mutants. Trichome initiation on the inflorescence organs of WT, zfp5, zfp6 and zfp5 zfp6 double mutants after treatment with varying concentrations of GA. Trichome initiation is shown on the second lateral branch (a); the third lateral branch (b); the second cauline leaf (c) and the third cauline leaf (d) of wt, zfp5, zfp6 and zfp5 zfp6 double mutants. The error bars indicate standard error (SE). The second and third lateral branches and the second and third cauline leaves of 20 plants were examined.

ZFP5 and ZFP6 are required to control trichome cell differentiation during inflorescence development through cytokinin signaling

Cytokinins are also reported to be able to promote trichome initiation in flowers and inflorescence organs (Greenboim-Wainberg et al., 2005; Gan et al., 2007a). In order to study whether ZFP6 and ZFP5 control trichome initiation through cytokinin signaling, we first checked whether the expressions of ZFP6 and ZFP5 in inflorescence organs of the WT were induced by external BA application. Wild-type plants were sprayed with 100 μM BA and plants were harvested at 2 h after BA application. As shown in Fig. 3(b), the expressions of ZFP6 and ZFP5 are both significantly higher in BA-treated plants than in the negative control in the WT plant. These results indicate that the expressions of ZFP5 and ZFP6 are induced by cytokinin. To explore the roles of ZFP5 and ZFP6 in cytokinin signaling, we further characterized trichome phenotypes of zfp5 and zfp6 mutant in response to external 6BA application. The external cytokinin application experiments showed that trichome production in zfp5, zfp6 and zfp5zfp6 plants on the lateral branch (the second and third) and cauline leaves (the second and third) were less sensitive than WT to 6BA application (Fig. 5a–d; Table S2). This indicates that both ZFP5 and ZFP6 are required for cytokinin signaling to regulate trichome cell differentiation in inflorescence organs in Arabidopsis.

Figure 5.

The effect of 6-benzylamino-purine (BA) application on trichome initiations in Arabidopsis thaliana wild-type (WT), zfp5, zfp6 and zfp5 zfp6 double mutants. Trichome initiation on the inflorescence organs of the WT, zfp5, zfp6 and zfp5 zfp6 double mutants after treatment with varying concentrations of GA. Trichome initiation is shown on the second lateral branch (a); the third lateral branch (b); the second cauline leaf (c) and the third cauline leaf (d) of the WT, zfp5, zfp6 and zfp5 zfp6 double mutants. The error bars indicate standard error (SE). The second and third lateral branches, and the second and third cauline leaves from 20 plants were examined.

ZFP6 acts upstream of the ZFP5 and GIS

Trichome initiation is believed to be controlled by the TTG1/GL1/MYB23/GL3/EGL3 complex (Szymanski et al., 2000; Ishida et al., 2008). Our previous results have showed that GIS and its subfamily genes, GIS2, ZFP8 and ZFP5, act upstream of GL1 and GL3. In order to study the genetic position of ZFP6 in the trichome initiation pathway, we first examined the relative expression levels of GL1 and GL3 in developing inflorescence shoots of the zfp6 mutant and 35S::ZFP6 plants. The results showed that expressions of GL1 and GL3 were significantly down-regulated in the zfp6 mutant, but significantly up-regulated in 35S::ZFP6 (Fig. 6a). These results suggest that ZFP6 acts upstream of the trichome initiation complex. In order to confirm this result, we overexpressed the ZFP6 gene in the gl1, gl3, and ttg-1 mutant backgrounds. Trichome initiation in these mutants was not rescued (Fig. 6b), confirming that ZFP6 functions upstream of the trichome initiation complex.

Figure 6.

Genetic interaction between ZFP6 and trichome initiation complex GL1/GL3/TTG1. (a) Relative expressions of GL1 and GL3 in developing inflorescence shoots of Arabidopsis thaliana zfp6 and 35S::ZFP6. Values represent the ratios of gene expression in a particular genotype to the corresponding wild-type or transgenic control. O-ZFP6-8, 35S::ZFP6-overexpressing line 8. Error bars represent SE. A Student's t-test was calculated at the probability of 1% (**, < 0.01). (b) Trichome production on main stem of gl1, gl3, ttg1, 35S::ZFP6::gl1, 35S::ZFP6::gl3 and 35S::ZFP6::ttg1.

Our previous results have shown that GIS and GIS subfamily gene GIS2 and ZFP8 also act upstream of GL1 and GL3, and that ZFP5 functions upstream of GIS, GIS2 and ZFP8. Therefore, we proceeded to examine the genetic relationship between ZFP6 and GIS, GIS2, ZFP8 and ZFP5. We examined the expression of ZFP5, GIS, ZFP8 and GIS2 in the zfp6 mutant and in 35S::ZFP6, and found that the relative expression of ZFP5, GIS, ZFP8 and GIS2 in the developing inflorescence shoots of zfp6 mutants was significantly lower than that of WT plants (Fig. 7a), whereas the expression of ZFP5 and GIS was significantly higher in 35S::ZFP6 than in control plants (Fig. 7b). In a complementary experiment, we checked whether a fusion of the ZFP6 promoter with the ZFP6 coding sequence could complement the zfp6 mutant phenotype. The insertion of the ZFP6 promoter driving the ZFP6 coding sequence into the zfp6 mutant resulted in the restoration of a WT phenotype in the flower and lateral branches (Fig. 8a). To determine whether the trichome defect of zfp6 mutants could be overcome by increased activity of ZFP5, we overexpressed ZFP5 in the zfp6 mutant background. This resulted in the restoration of WT degrees of trichome initiation in cauline leaves and lateral branches (Fig. 8b,c). Furthermore, the insertion of the ZFP5 promoter driving the ZFP5 in the zfp6 mutant resulted in a WT phenotype in the cauline leaves and lateral branches (Fig. 8b,c). Taken together, these results suggested that ZFP6 acts upstream of ZFP5.

Figure 7.

Relative expression of ZFP5, ZFP8, GIS and GIS2 in developing inflorescence shoots of Arabidopsis thaliana zfp6 and 35S::ZFP6 (O-ZFP6-8). (a) The relative expression levels of ZFP5, ZFP8, GIS and GIS2 in developing inflorescence shoots of zfp6. (b) The relative expression levels of ZFP5 and GIS in developing inflorescence shoots of 35S::ZFP6-8. Error bars represent SE. A Student's t-test was calculated at the probability of either 5% (*, < 0.05) or 0.1% (**, < 0.01).

Figure 8.

Rescue of the zfp5 and zfp6 mutant phenotype by the ZFP6 coding sequence. The ZFP6 promoter driving the ZFP6 coding sequence rescues the Arabidopsis thaliana zfp6 mutant phenotype on the flower and the second lateral branch (a). Trichome initiation on the flower. (b, c) Trichome initiation on the second cauline leaf (b) and the second lateral branch (c) of the wild-type (WT), zfp6, 35S::ZFP5::zfp6 and pZFP5::ZFP5::zfp6.

Expression pattern of ZFP6 gene

Like GIS, GIS2, ZFP8 and ZFP6, sequence analysis showed that ZFP6 contains a conserved C2H2 domain and is most similar to the ZFP5 family of transcription factors (Meissner & Michael, 1997; Payne et al., 2004; Gan et al., 2006, 2007a; Zhou et al., 2011). Quantitative RT-PCR results showed that ZFP6 was highly expressed in roots, mature stem and lateral branches, but was low in rosette leaves and siliques (Fig. 9).

Figure 9.

Expression pattern of the ZFP6 gene. Quantitative reverse transcription polymerase chain reaction analysis of ZFP6 expression in different tissues of Arabidopsis thaliana wild-type plants. All values were normalized using an internal standard (UBQ10). Error bars represent SE. Ros., rosette leaf; caul., cauline leaf; Fl., flower; Mat., fully elongated first internode of the main stem; dev. 1st br., the first developing branch (lateral branch); dev. 2nd br., the second developing branch (lateral branch); sil., silique.

Discussion

Zinc Finger Protein 6 (ZFP6) regulates the trichome initiation by integrating GA and cytokinin signaling

We have identified five C2H2 transcriptional factors that regulate trichome cell differentiation in inflorescence organs through the GA and cytokinin pathways (Gan et al., 2006, 2007a; Zhou et al., 2011). These transcriptional factors belong to GIS clade (GIS, GIS2 and ZFP8) and ZFP5 clade (ZFP5 and ZFP6; Gan et al., 2006). Molecular and genetic analyses demonstrate that GIS acts upstream of the key trichome initiation complex regulators GL1 and GL3, and downstream of the GA signaling repressor SPINDLY (SPY) in a GA signaling pathway (Gan et al., 2006). The control of trichome production by cytokinins requires two GIS-related genes: ZFP8 and GIS2 (Gan et al., 2007a). Cytokinin-inducible genes GIS2 and ZFP8 play a prominent role in the cytokinin response, where they act downstream of SPINDLY and upstream of GLABROUS1 (Gan et al., 2007a). In addition, like GIS, GIS2 and ZFP8 are also required to control trichome production by GAs. By contrast, GIS does not play a direct role in the cytokinin response. Recently, a ZINC FINGER PROTEIN, ZFP5, was reported to play an important role in controlling trichome cell development also through GA signaling (Zhou et al., 2011, 2012). The molecular analyses suggest that ZFP5 functions upstream of GIS, GIS2 and ZFP8 by directly regulating ZFP8 expression to control trichome cell differentiation. In this study, we have investigated the molecular mechanisms by which ZFP6 integrates cytokinin and GA signaling in the regulation of trichome cell differentiation in Arabidopsis. We found that the influence of GA and cytokinin on trichome production is modulated by the combined actions of ZFP6 and the most closely related transcription factor ZFP5, which was reported previously to regulate trichome cell differentiation by GAs (Zhou et al., 2011). Mutation of ZFP6 suppressed the expression of GIS, GIS2, ZFP8, GL1 and GL3, whereas overexpression of ZFP6 promotes expression of GL1 and GL3. The overexpression of ZFP6 could not restore gl1 and gl3 glabrous mutant phenotypes (Figs 6, 7). These results indicate that ZFP5 controls shoot maturation by functioning upstream of GIS, GIS2, ZFP8, GL1 and GL3 (Figs 6, 7). Further molecular and genetic analyses showed that ZFP6 functions upstream of ZFP5 to control trichome cell differentiation by integrating GA and cytokinin signaling (Figs 7, 8). The mutation of ZFP6 precluded the expression of ZFP5, GIS, GIS2, ZFP8 and GL1 in response to external GA and cytokinin application (Fig. 3c,d). These results led us to propose a model whereby GA and cytokinin control trichome initiation through five C2H2 transcriptional factors (Fig. 10).

Figure 10.

Model of ZFP6 integrating GA and cytokinin signals to control trichome initiation on Arabidopsis inflorescence organs.

Functional specialization of ZFP5 and ZFP6

We have identified five C2H2 transcriptional factors that regulate trichome cell differentiation in inflorescence organs (Gan et al., 2006, 2007a; Zhou et al., 2011). These transcriptional factors belong to the GIS clade (GIS, GIS2 and ZFP8) and the ZFP5 clade (ZFP5 and ZFP6; Gan et al., 2006). The functional specialization of genes encoding paralogous transcription factors may represent an important mechanism through which plants regulate similar cell differentiation programs at different development stages (Lee & Schiefelbein, 2001). For the GIS clade, GIS, ZFP8 and GIS2 genes encode functionally equivalent proteins that have diverged in their response to phytohormones and in their roles in the control of trichome cell development in inflorescence organs (Gan et al., 2006, 2007a). GIS is mainly involved in regulating trichome cell initiation on lateral branches, while GIS2 specifically regulates the initiation of trichomes in the sepals of flowers, and ZFP8 specifically controls trichome cell differentiation on cauline leaves (Gan et al., 2006, 2007a).

Although the expression patterns of GIS, ZFP8 and GIS2 have diverged in ways that reflect their function in different inflorescence organs, they are all expressed mainly in the shoot. ZFP5 encodes a protein functionally equivalent to GIS (Zhou et al., 2012), but is expressed mainly in the root (Zhou et al., 2011). Similarly, although ZFP6 is expressed mainly in the root, it also plays a key role in trichome cell differentiation in inflorescence organs. A similar situation has been reported for the MYB transcription factor WEREWOLF 1 (WER) and GL1. WER is an activator of trichome initiation functionally equivalent to GL1, but is expressed predominantly in the root, whereas GL1 regulates trichome development in the shoot (Choi et al., 2001).

As mentioned earlier, GIS, ZFP8 and GIS2 appear to function cumulatively in the modulation of specific trichome cell differentiation, as double and triple mutants have smaller numbers of trichomes in the inflorescence organs than the single mutants. In this respect, the ZFP5 clade differs from the GIS clade, which plays a general role rather than a specific role in regulating trichome cell differentiation in inflorescence organs. As reported here, the mutant phenotype of zfp5 is very similar to the mutant phenotype of zfp6, which may indicate that ZFP5 and ZFP6 are largely equivalent in function. However, it seems that they function separately, as the double mutant zfp5zfp6 did not show a significantly stronger phenotype than either single mutant. By contrast, double mutants gisgis2, giszfp8 and gis2zfp8 of the GIS clade all showed additive effects and significantly stronger phenotypes than the single mutants (Gan et al., 2007a).

In summary, we have identified and characterized a new C2H2 zinc finger protein, ZFP6, which plays a role in regulating trichome cell initiation in inflorescence organs. Our results demonstrate that ZFP6 controls inflorescence trichome development by regulating the expressions of GIS, GIS2, ZFP8 and ZFP5 through the integration of GA and cytokinin signaling. Our results demonstrate the importance of functional specialization within the C2H2 transcription factor family in the control of trichome cell differentiation and in response to various hormonal and developmental cues. This represents an increase in our understanding of how developmental and hormonal signals are integrated in plants. Further system biology approaches, which integrate the analysis of C2H2 gene networks with the precise mapping of hormone influences within plants, may reveal a fuller picture of how transcriptional factors integrate hormonal signals and environmental cues to control trichome cell differentiation.

Acknowledgements

We thank Prof. Simon McQueen-Mason and Dr Clare Steele-King from the University of York, UK, for critical reading of the manuscript. The research was supported by the National Natural Science Foundation of China (grant nos. 30970167, 31228002, 31000093); Zhejiang Provincial Natural Science Foundation of China (grant no. Z31100041); PhD Programs Foundation of Ministry of Education of China (grant no. 20120101110079) and the International Scientific and Technological Cooperation Project of Ministry of Science and Technology of China (grant no. 2010DFA34430).

Ancillary