Homeodomain protein ATHB6 is a target of the protein phosphatase ABI1 and regulates hormone responses in Arabidopsis

Abstract

ABI1, a protein phosphatase 2C, is a key component of signal transduction in Arabidopsis. It regulates diverse responses to the phytohormone abscisic acid (ABA) such as stomatal closure, seed dormancy and inhibition of vegetative growth. By analysing proteins capable of interacting with ABI1, we have identified the homeodomain protein ATHB6 as a regulator of the ABA signal pathway. Critical for interaction between ATHB6 and ABI1 is an intact protein phosphatase domain and the N-terminal domain of ATHB6 containing the DNA-binding site. ATHB6 recognizes a cis-element present in its promoter, which encompasses the core motif (CAATTATTA) that mediated ATHB6- and ABA-dependent gene expression in protoplasts. In addition, transgenic plants containing a luciferase gene controlled by the ATHB6 promoter documented a strong ABA-inducible expression of the reporter which was abrogated in the ABA-insensitive abi1 mutant. Arabidopsis plants with constitutive expression of the transcriptional regulator revealed ABA insensitivity in a subset of ABI1-dependent responses. Thus, the homeodomain protein ATHB6 seems to represent a negative regulator of the ABA signal pathway and to act downstream of ABI1.

Fig. 1. Interaction of ATHB6 with ABI1 in the yeast two-hybrid system and in vitro. (A) ABI1 fused to the GAL4 DNA-binding domain (DB fusion) was analysed for interaction with fusions of ATHB6, ATHB5 and ATHB7 to the GAL4 activation domain (AD fusion), respectively. Cells transformed with empty vectors for AD and DB fusion (C) were used as control. The β-galactosidase activity and the standard deviation are presented as the relative values of three independent experiments. (B) The in vitro binding of ATHB6 and ABI1 protein was performed by affinity interaction. ATHB6 fused to MBP (MBP–ATHB6) or MBP as a control were immobilized on amylose resins. Binding of ABI1 was analysed by chromatography of radiolabelled ABI1 on a column containing the charged resins. Fractions were collected and quantified for radioactivity. Elution of the ATHB6–ABI1 complex was initiated by administration of maltose-containing solution indicated by an arrow, and 32% of applied ABI1 was recovered subsequently (continuous line). ABI1 chromatography on MBP-charged resin yielded a recovery of 3.8% of applied ABI1 in the elution fractions (dashed line). (C) Purified proteins used in the in vitro binding assay were separated by SDS–PAGE and visualized by silver staining. In addition, MBP and MBP–ATHB6 were identified by immunodetection using antibodies directed against MBP, and the radiolabelled ABI1 protein fraction was analysed by autoradiography. The positions and molecular weight of protein size markers are indicated on the left.

Fig. 2. Dependence of ABI1 and ATHB6 interaction on the functional catalytic domain of PP2C. (A) Specific protein phosphatase activity of purified ABI1, and both point-mutated forms abi1 (ABI1^Gly180Asp) and non-active ABI1^Asp177Ala (NAP) were measured and expressed as relative activities (n = 3, ± SD). (B) The homeodomain (HD) and leucine zipper (ZIP) regions of the AD fusion for ATHB6 (amino acids 1–311) and mutant versions thereof [Δ1 (amino acids 44–311), Δ2 (amino acids 44–217) and S67A (ATHB6^Ser67Ala)] are presented schematically (left panel). They were analysed for binding to DB fusions of ABI1 (A), abi1 (a) and NAP (N) in the yeast two-hybrid system. The β-galactosidase activities for combination with the empty AD vector were subtracted as background from the values presented.

Fig. 3. Specific binding of ATHB6 to a cis-element present in the ATHB6 promoter. The 30mer oligomeric DNA fragment (oligo α) contains a putative cis-element located between position –638 and –609 of the ATHB6 promoter. The core interaction motif (bold) is mutated in oligo β by the transition of A to G (underlined). ATHB6 fused to MBP and MBP as control were analysed for specific DNA binding to oligo α versus oligo β. Complexes of ATHB6 and DNA (arrowhead) were tested in the presence of increasing ionic strength (0, 50, 100 and 500 mM KCl). PKA-phosphorylated MBP–ATHB6 was tested for binding to oligo α in the absence of KCl. Complexes were separated from unbound radiolabelled DNA by EMSA and visualized by autoradiography.

Fig. 4. Requirement for a functional cis-element and ATHB6 for ABA-mediated promoter activation in Arabidopsis protoplasts. (A) The reporter constructs (wt and mutant) contain four tandemly orientated copies of the ATHB6-binding sequence (CAATTATTA, oligo A) or the point-mutated sequence (CAATTGTTA, oligo B), respectively, fused upstream to the CaMV 35S core promoter (35S core) that controls expression of firefly luciferase (LUC). They are represented schematically together with the effector constructs that allow constitutive expression of ATHB6 and ΔATHB6 under control of the 35S promoter (35S). In ΔATHB6, the α-helix 3 (amino acids 96–117) of the homeodomain (HD) essential for DNA binding was deleted. The positions of HD helices are indicated by shaded boxes. (B) The ATHB6-dependent activation of reporter constructs was tested in a transient gene expression system. Transfected Arabidopsis protoplasts were incubated for 24 h in the presence (+) or absence (–) of 30 µM ABA and analysed for expression of the LUC reporter. Co-transfection of a constitutively expressed aequorin gene allowed normalization of expression in independent experiments. The induction of LUC activity from wild-type and mutant reporter constructs by different effectors is shown relative to the expression from wild-type reporter in the absence of ectopically expressed effector. The data reflect the results of three independent transfections ± SD.

Fig. 5. ABA- and ABI1-dependent activation of the ATHB6 promoter. The regulation of the ATHB6 promoter was analysed by stable integration of an ATHB6 promoter driving LUC expression in Arabidopsis. (A) The specific LUC activity of seedlings homozygous for the ATHB6 reporter construct and the 35S LUC control was determined from extracts 24 h after ABA administration, respectively. Values are expressed as fold induction of specific LUC activity relative to the untreated control. (B) The time-dependent increase of specific LUC activity in the transgenic reporter line by 10 µM ABA. The arrow signifies the time point of challenge to exogenous ABA. (C) The ABI1 dependence of the ABA-mediated reporter activation was studied 24 h after ABA addition in bulked F₂ seedlings from crosses of the reporter line to the ABA-insensitive abi1 mutant (abi1) and to the wild-type (wt) of the same ecotype to avoid ecotype-specific differences in ABA sensitivity. The Mendelian segregation of the dominant mutant trait results in a quarter of the seedlings of the F₂ population expressing LUC in a wt ABI1 background. Specific activity is expressed as light units (LU) per µg of protein. The standard deviation among four independent experiments is indicated.

Fig. 6. Analysis of ABA responses in Arabidopsis plants with ectopic expression of ATHB6. (A) Northern blot analysis for ATHB6-specific RNA abundance in two transgenic Arabidopsis lines (sense 1 and sense 2) ectopically expressing ATHB6, and in control plants in which the ATHB6 structural gene had been replaced by the GUS coding sequence (control). Total RNA (20 µg) of seedlings treated with (+) or without (–) 10 µM ABA for 24 h was separated by electrophoresis and probed for the presence of ATHB6 transcripts (upper panel). The position of endogenous ATHB6 and ectopically expressed mRNA is indicated by the upper and lower arrowhead, respectively. The transcripts differ in size due to deletions in the untranslated leader sequences of the ATHB6 RNA constitutively expressed from the 35S promoter. The level of endogenous transcript induction by ABA is stated below and is corrected for equal RNA loading, which was determined by imaging and quantifying rRNA bands of the RNA samples stained with ethidium bromide (lower panel). (B) Fraction of seeds germinating in the presence of ABA after 5 days determined from a total of 80 seeds. (C) Five-day-old seedlings grown under sterile conditions were transferred on solid medium with various ABA concentrations. Root growth within 4 days of transfer was determined (n = 30, SD ± 14%). In the absence of ABA, root growth equalled 15 ± 1, 14 ± 2 and 14 ± 1 mm for the control, sense 1 and sense 2 lines, respectively. (D) Stomatal response measured by water loss of excised leaves. Leaves of a comparable developing stage (n = 5) from 4-week-old plants were excised and the loss of the fresh weight was measured at ambient conditions. The values are indicated for the control (open squares), as well as for the transgenic sense 1 (closed squares) and sense 2 (closed circles) lines.