Root avoidance of toxic metals requires the GeBP‐LIKE 4 transcription factor in Arabidopsis thaliana

Summary Plants reorganize their root architecture to avoid growth into unfavorable regions of the rhizosphere. In a screen based on chimeric repressor gene‐silencing technology, we identified the Arabidopsis thaliana GeBP‐LIKE 4 (GPL4) transcription factor as an inhibitor of root growth that is induced rapidly in root tips in response to cadmium (Cd). We tested the hypothesis that GPL4 functions in the root avoidance of Cd by analyzing root proliferation in split medium, in which only half of the medium contained toxic concentrations of Cd. The wild‐type (WT) plants exhibited root avoidance by inhibiting root growth in the Cd side but increasing root biomass in the control side. By contrast, GPL4‐suppression lines exhibited nearly comparable root growth in the Cd and control sides and accumulated more Cd in the shoots than did the WT. GPL4 suppression also altered the root avoidance of toxic concentrations of other essential metals, modulated the expression of many genes related to oxidative stress, and consistently decreased reactive oxygen species concentrations. We suggest that GPL4 inhibits the growth of roots exposed to toxic metals by modulating reactive oxygen species concentrations, thereby allowing roots to colonize noncontaminated regions of the rhizosphere.


Fig. S1
Cd inhibits the growth of Arabidopsis thaliana plants (related to Figs 1, 3, and 4).           Fig. 3 Table S4 Growth of the wild-type and GPL4 CRES-T plants in the horizontal split media assay presented in Fig. 4 Table S5 Gene ontology biological processes represented by the genes in Clusters 1-5  Methods S1 Analysis of transcriptional activation activity and promoter binding assay.

Methods S2 Yeast one-hybrid (Y1H) analysis.
Notes S1 Accession numbers.    light blue and +Cd for '-/+Cd'; pink, increased color intensity denotes an increase in Cd concentration) are in the lower segments. Each experiment was performed with greater than or equal to six seedlings per genotype per treatment. Different letters indicate a significant difference in percentage allocation between genotypes or treatments (Tukey's HSD test, P < 0.01).   quencher KI (1 mM). Five seeds per genotype were sown and grown for 2 wk on a ½ MS agar plate without or with supplementation with 70 µM CdCl 2 (-Cd and +Cd, respectively, left) and with supplementation of KI in the presence or absence of Cd (right). The experiment was repeated three times, and combined results (mean ± SE) of total fresh weight (bottom left) and the longest root length (bottom right) are given. Each experiment was performed with three plates per treatment and supported the same conclusion. Shoot growth, and consequently the total fresh weight, much reduced by KI, both in the wild type and OX lines.
The extreme sensitivity of shoot growth to KI may be due to the critical role of ROS in cell expansion in the shoot (Lu et al., 2013)  µM CdCl 2 (+Cd; h), and medium containing both BSO (100 µg ml -1 ) and Cd (+BSO +Cd; i).
Bar, 1 cm. Mean values (± SE) of the longest root length (j) and fresh weight per plate (k) of three independent experiments. Each individual experiment was performed with three plates (15 seedlings) per genotype per treatment and supported the same conclusion.
Different letters indicate that means are significantly different between genotypes or treatments (Tukey's HSD (P < 0.05).

qRT-PCR
Root fresh weight (mg) WT 5.3 ± 0.7 a 5.4 ± 0.7 a 6.5 ± 0.5 b 2.5 ± 0.5 c 6.9 ± 0.7 f 0.6 ± 0. Absolute values (mean ± SE) of root fresh weight and the longest root length measured in the split media assays shown in Fig 3 and S4. In heterogeneous split media, root growth was dramatically reduced in the sides containing Cd (blue color), and root biomass was significantly increased in sides lacking Cd (red color) by the avoidance response in the wild type and OX, compared with the counterparts in control split media. Such root biomass increase in the side lacking Cd is likely due to development of more roots, and not to root length. By contrast, CRES-T and RNAi failed to exhibit such an avoidance response. Different characters in the superscript indicate significant difference in means between genotypes and treatment by Tukey's HSD test (P ≤ 0.05). Absolute values (mean ± SE) of the longest root length measured in the split soil assays shown in Fig. 3 are presented. Root biomass was not analyzed due to technical problem with soil experiment. In heterogeneous split media, root growth was dramatically reduced in the sides containing Cd (blue), and root biomass was significantly increased in sides lacking Cd (red color) by the avoidance response in the wild type and, compared with the counterparts in control split media. Different characters in the superscript indicate significant difference in means between genotypes and two sides of the medium by Tukey's HSD test, P ≤ 0.05. Absolute values (mean ± SE) of root fresh weight and lateral root number measured in the split root assays shown in Fig. 4. In heterogeneous split media, root biomass and lateral root number were dramatically reduced in the sides containing Cd (blue), whereas they were significantly increased in sides lacking Cd (red color) by the avoidance response in the wild type, compared with the counterparts in control split media. By contrast, such an avoidance response was significantly suppressed in CRES-T. Different characters in the superscript indicates significant difference in means between genotype and two sides of the medium analyzed by Tukey's HSD test, P ≤ 0.05.  Absolute values (mean ± SE) of root fresh weight and the longest root length measured from the split root assay in Fig. S7 are presented. In heterogeneous split medium, root biomass of wild-type plants in the side lacking PQ (red) was significantly increased compared to the counterpart in control split medium, and the root biomass increase is likely due to the development of more roots, because root length did not increase. By contrast, CRES-T lines failed to exhibit such an avoidance response.
Different characters in the superscript indicate significant difference between genotype and two sides of the medium analyzed by Tukey's test, P ≤ 0.01.

Table S8
Growth of wild-type and CRES-T plants in the vertical split media assays shown in Fig. 6 Absolute values (mean ± SE) of root fresh weight and the longest root length measured from the split root assay in Fig. 6

Root length (mm)
Method S1 Analysis of transcriptional activation activity and promoter binding assay.
Transcriptional activation activity of GPL4 was tested in two different ways. First, it was examined using a transient expression assay in Arabidopsis rosette leaves as described previously (Hiratsu et al., 2002). The effector (GPL4-CDS fused to the GAL4 DB domain), reporter (GAL4-LUC), and reference (luciferase from Renilla) plasmids were delivered into the rosette leaves of 4-wk-old Arabidopsis plants by particle bombardment (BIORAD).
Bombarded leaves were incubated for 12 h in darkness, and then luciferase activity was quantified using a Dual Luciferase Kit (Promega, Madison, Wisconsin, USA). For the promoter binding assay, Arabidopsis leaves were transformed with the effector (GPL4-CDS), reporter (promoter of candidate genes fused to the CDS of luciferase), and reference (luciferase from Renilla) plasmids, and luciferase activity was measured as above using a Tecan spectrophotometer (Infinite M200PRO).
Secondly, GPL4 transcriptional activation activity was tested in a yeast system. The Method S2 Yeast one-hybrid (Y1H) analysis.
A modified MATCHMAKER One-Hybrid System (Clontech Laboratories, Mountain View, CA, USA) was used to perform the Y1H to analyze the direct binding of promoters of selected genes and GPL4. pHisi-1 vectors harboring the promoters (1.5-2 kb) of each selected candidate gene were integrated into the genome of yeast strain YM4271, and transgenic strains were selected on synthetic dextrose medium agar plates without Ura and His (SD/-UH). To further verify the insertion of the promoters in the genome, yeast gDNA was extracted and PCR was performed with primers specific to the promoter region of each gene (Table S1). The transformed yeast cells harboring the target promoter were transformed again with a pGAD424 vector containing GPL4, and double transformed yeast cells were selected on synthetic dextrose medium lacking Ura and Leu (SD/-UL). The interaction between GPL4 and the target promoters was assessed by growth on SD medium lacking Ura, Leu, and His (SD/-UHL), but supplemented with 30 mM 3-amino-1,2,4-triazole (3-AT; Mitsuda et al., 2010).
Sequence data from this article can be found in the Arabidopsis Genome Initiative database under the following accession numbers: GPL4 (At1g44810), the closest GPL4-Homologue