A protein kinase target of a PDK1 signalling pathway is involved in root hair growth in Arabidopsis

Abstract

Here we report on a lipid-signalling pathway in plants that is downstream of phosphatidic acid and involves the Arabidopsis protein kinase, AGC2-1, regulated by the 3'-phosphoinositide-dependent kinase-1 (AtPDK1). AGC2-1 specifically interacts with AtPDK1 through a conserved C-terminal hydrophobic motif that leads to its phosphorylation and activation, whereas inhibition of AtPDK1 expression by RNA interference abolishes AGC2-1 activity. Phosphatidic acid specifically binds to AtPDK1 and stimulates AGC2-1 in an AtPDK1-dependent manner. AtPDK1 is ubiquitously expressed in all plant tissues, whereas expression of AGC2-1 is abundant in fast-growing organs and dividing cells, and activated during re-entry of cells into the cell cycle after sugar starvation-induced G1-phase arrest. Plant hormones, auxin and cytokinin, synergistically activate the AtPDK1-regulated AGC2-1 kinase, indicative of a role in growth and cell division. Cellular localisation of GFP-AGC2-1 fusion protein is highly dynamic in root hairs and at some stages confined to root hair tips and to nuclei. The agc2-1 knockout mutation results in a reduction of root hair length, suggesting a role for AGC2-1 in root hair growth and development.

Figure 1

Interaction of AtPDK1 with AGC2-1 and AGC1-1. (A) Yeast two-hybrid interaction assay with wild-type AtPDK1-1 or AtPDK1-2 and with the pleckstrin homology (PH) domain of AtPDK1-1 fused to the Gal4 DNA-binding domain, and the wild-type AGC2-1 or AGC1-1 or their indicated C-terminal mutations fused to the Gal4 activation domain. (B) Schematic diagram showing the domain structure of AGC2-1 with alignments to AGC1-1, to Atp70S6K and to AtPDK1-1, and the site-directed mutations. Hatched box=catalytic domain; black box=ser/thr protein kinase active site signature. Conserved PDK1 phosphorylation site is underlined. (C) AGC1-1 and AGC2-1 substrate specificities against the indicated synthetic peptides in the presence or absence of the PKI inhibitory peptide. Activities are expressed as a percentage of the kemptide phosphorylation. In all cases, the AGC1-1 and AGC2-1 protein levels were determined by Western blotting using the HA antibody. (D) Interaction and activity of wild-type and mutant versions of GST-AGC2-1 when coexpressed with Myc-AtPDK1 in Arabidopsis protoplasts. Two independent experiments are shown. GST-AGC2-1 activity is expressed as a percentage of the wild type. Myc-AtPDK1 levels are shown in crude extracts or in GST pull-downs.

Figure 2

AGC2-1 but not AGC1-1 activity is dependent on AtPDK1 levels. (A) Schematic showing the AtPDK1-RNAi construct design. (B) AGC1-1 and AGC2-1 activities with and without the cotransfection of the AtPDK1-RNAi construct (PDKi) or the PDKi construct alone as a control. (C) Protein levels of AGC1-1 and AGC2-1 detected through the HA tag in the same samples as in (B). (D) Efficiency of the AtPDK1 RNA interference as determined by the protein levels of the Myc-AtPDK1 coexpressed within the same cells in the first four samples indicated in (B). (E) Phosphorylation of AGC1-1 and AGC2-1 by AtPDK1 in vitro. GST-tagged AGC1-1 and AGC2-1 were purified from transfected Arabidopsis cells under a noninduced inactive state and its phosphorylation was tested in a kinase reaction in the presence or absence of AtPDK1 and visualised by autoradiography. (F) Coimmunoprecipitation of endogenous AtPDK1 and AGC2-1. Arabidopsis cell lysate was tested for endogenous levels of AtPDK1 (lane 1) or AGC2-1 (lane 2) and for AtPDK1 in the AGC2-1 immunoprecipitate (lane 3) using AtPDK1- and AGC2-1-specific antibodies for immunoprecipitation (IP) and for Western blots (Blot) as indicated.

Figure 3

Binding specificity and activation of AtPDK1 by lipids. (A) N-terminally GST-tagged proteins of various lipid-binding domains were purified from E. coli and incubated with either PI3P or PA beads. The specifically bound proteins were eluted and detected by a GST antibody on Western blots. (B) AtPDK1 activity was determined against the synthetic peptide PIFtide. Phospholipids were added to cells for the indicated times and treatments labelled as follows: PA (solid line, filled square), PI(4,5)P₂ (dotted bold line, filled triangle), PI3P (solid line, filled circle), PI4P (solid line, cross), no lipid (dashed line, diamond). The activities of purified HA-AtPDK1, expressed in transfected protoplasts, were determined in triplicate.

Figure 4

Regulation of AGC2-1 activity by phosphatidic acid and the plant hormones auxin and cytokinin. (A) Activity of HA-AGC2-1 that was expressed and then purified from Arabidopsis protoplasts treated with various lipids for the indicated times as follows: PA (solid line, filled square), PI(4,5)P₂ (solid line, filled circle), PI(3,4,5)P₃ (solid line, open circle), no lipid (solid line, filled triangle). (B) Activity of HA-AGC1-1 in an experiment as described in (A). (C) Phosphatidic acid specifically activates AGC2-1 through AtPDK1. The activity of HA-AGC2-1 in cells treated with PA (solid line, filled square), with 0.01% n-butanol to inhibit phosphatidic acid production (dotted line, open square), with 0.01% sec-butanol, as an inactive analogue (solid line, open square), or the control without treatment (solid line, filled square). AtPDK1 levels were ablated by expression of the PDKi construct as in Figure 2 in cells treated with PA (solid line, cross). (D) Arabidopsis cells were transfected with HA-AGC-2-1 and cotransfected either with the putative PA-binding domain of TOR fused to GFP (TOR-PA) or the PH domain of AtPDK1 fused to GFP (PDK1-PH), incubated for 24 or 48 h and the AGC2-1 activities were determined in triplicate. (E) Effects of the elevation of endogenous PA levels through activation of the heterotrimeric G-protein by mastoporan on AGC2-1 activity. Cells were treated for the indicated times with PA (solid line, filled square), 4 μM mastoporan analogue MAS7 (solid line, open square) or its inactive analogue Mas17 and with PA in cells where AtPDK1 levels were ablated by the cotransformation of the PDKi construct (solid line, cross) and compared to control without treatment (solid line, filled triangle. (F) Auxin and cytokinin synergistically elevate AGC2-1 activity. The hormones were added either singly or together (NAA=0.5 mg/l, Kin=0.05 mg/l), and HA-AGC2-1 activities were determined over a 24 h period. To control the expression of transfected constructs, in all experiments the HA-AGC1-1, HA-AGC2-1 or Myc-AtPDK1 levels were monitored by Western blot analysis of total cell lysates (data not shown).

Figure 5

Expression of AtPDK1 and AGC2-1 mRNA. (A) Northern blot with 3′ gene-specific probes of AtPDK1, AGC2-1, cyclin D3 (Riou Khamlichi et al, 1999) and the actin depolymerising factor 2 as a loading control (ADF2) (Allwood et al, 2002). Samples were chosen to represent various growth and cell division rates: cell suspension in logarithmic growth (CS-log), cell suspension in stationary phase (CS-Stat), roots taken from Arabidopsis plants grown in liquid with 3% sugar. (Root SB) or root from Agar plate with 0.5% sugar (Root agar). (B) Northern blot of RNA samples from sucrose starvation-induced cell cycle synchronisation experiment. Asynchronous culture (As) and time points of samples taken after re-addition of sugar to the medium are indicated. (C) Flow cytometry analysis of the number of cells with G1 and G2 DNA content and the percentages of cells in various cycle stages from selected samples used to prepare the Northern blot in (B).

Figure 6

Role of AGC2-1 in root hair growth. (A) Root hair phenotype of 7-day-old roots of wild-type and (B) the agc2-1/agc2-1 T-DNA insertional knockout mutant, imaged 10 mm from the primary root tip. Scale bar=100 μM. (C) GFP-AGC2-1 localisation in the primary root tip. (D) Localisation of GFP control in the primary root tip. (E, F) GFP-AGC2-1 localisation pattern in immature root hair located 5 mm from the root tip. (G) GFP-AGC2-1 localisation pattern 10 mm from the root tip. (H) Localisation of GFP control in root hair 10 mm from the root tip. Scale bar=100 μM.