ABA and the ubiquitin E3 ligase KEEP ON GOING affect proteolysis of the Arabidopsis thaliana transcription factors ABF1 and ABF3

Abstract

The ABA Binding Factor/ABA-Responsive Element Binding Proteins (ABF/AREB) subfamily of bZIP-type transcription factors are positive effectors of ABA responses. Here, we examine the proteolytic regulation of two members: Arabidopsis thaliana ABF1 and ABF3. Both transcription factors are unstable in seedlings, and their degradation is sensitive to proteasome inhibition. ABA treatment of seedlings leads to their rapid accumulation, the result of slowed proteolysis. Deletion of the conserved C-terminal region required for 14-3-3 interaction destabilizes the proteins. The degradation of ABF1 and ABF3 are slower in vivo in seedlings lacking the ubiquitin E3 ligase KEEP ON GOING (KEG), and in vitro in extracts from keg seedlings, implicating KEG in their degradation. ABF1 and ABF3 are ubiquitylation substrates of KEG in vitro, and in vitro pull-down assays document their direct interaction. In contrast to ABI5, another KEG substrate, the degradation of ABFs and proteolytic regulation of ABFs by ABA still occurs in keg seedlings, suggesting that additional E3s participate in ABF1 and ABF3 proteolysis. Loss of ABF1 or ABF3 in the keg background has a phenotypic effect similar to the loss of ABI5, and there is no additional rescue of the keg phenotype in abf1 abf3 abi5 keg seedlings. This result suggests that the abundance of other substrates is altered in keg seedlings, affecting growth. In conclusion, ABF1 and ABF3 abundance is affected by ABA and KEG, and the conserved C4 region serves as a stabilizing element.

Keywords: ABF; Arabidopsis thaliana; KEG; abscisic acid; proteolysis; ubiquitin; ubiquitin E3 ligase.

Figure 1

Degradation of ABF1 and ABF3 can be detected by CHX chase assay, and their levels are sensitive to MG132.(a) Seedlings from independent transgenic lines (designated A–C) expressing Myc-ABF1 (top) or Myc-ABF3 (bottom) under the control of the 35S promoter were grown for 6 days and treated with CHX for 3 or 6 h. CHX 0 represents mock treatment with solvent alone for 6 h. Proteins were extracted and immunoprecipitated by anti-Myc beads. The results were visualized after SDS-PAGE by anti-Myc immunoblotting. The number below represents signal intensity quantified by ImageJ, normalized to CHX 0.(b) Seedlings grown as described in (a) were treated with MG132 or solvent control for 24 h, then proteins were extracted and 100 μg visualized by SDS-PAGE and anti-Myc immunoblotting. Ponceau S staining was used as the loading control.

Figure 2

ABF1 and ABF3 proteins rapidly accumulate in response to ABA.Seedlings from three independent transgenic lines expressing either Myc-ABF1 (left) or Myc-ABF3 (right) were grown for 6 days and treated with ABA for 0.5, 1, 2 or 6 h. Time 0 represents mock treatment with solvent alone for the length of 6 h. Proteins were extracted and the results were visualized after SDS-PAGE and anti-Myc immunoblotting. Ponceau S staining was used as the loading control. Col represents non-transgenic control seedlings.

Figure 3

ABA slows down ABF1 and ABF3 protein degradation.Seedlings from independent transgenic lines expressing Myc-ABF1 (a) or Myc-ABF3 (b) were grown for 6 days and treated with ABA or with ethanol as mock treatment for 6 h, then treated with CHX for the indicated times. Proteins were extracted (a), or were extracted and immunoprecipitated by anti-Myc beads (b). Results were visualized after SDS-PAGE by anti-Myc immunoblotting. The percentages represent signal intensities quantified by ImageJ normalized to CHX 0. Ponceau S staining was used as the loading control.(c) Data from independent experiments were plotted normalized to 0 CHX (ABF1, n = 6; ABF3, n = 8). The percentage remaining is statistically different in ABA samples than the control by a Student's t–test, with P < 0.0001 for both ABF1 and ABF3.

Figure 4

ABF1 and ABF3 protein degradation is slowed in keg.Seedlings from independent transgenic lines in +/keg background expressing Myc-ABF1 (a) or Myc-ABF3 (b) under the control of the 35S promoter were grown for 6 days [wild-type (WT) siblings] or 14 days (keg siblings) and treated with cycloheximide for the indicated times. CHX 0 represents mock treatment with solvent alone. Proteins were extracted and immunoprecipitated by anti-Myc beads. Results were visualized by anti-Myc immunoblotting. The percentages represent western blot signal intensity quantified by ImageJ and compared with CHX 0.(c) Data from three independent experiments were plotted normalized to CHX 0 (ABF1, n = 6; ABF3, n = 8). The percentage remaining is statistically different in keg samples than in the WT by a Student's t–test: P < 0.0031 and P < 0.0002, for ABF1 and ABF3, respectively.

Figure 5

In vitro degradation of ABF1 and ABF3 is slowed in keg.Bacterially expressed recombinant GST-FUS3 (a), His-Flag-ABF1 (b), His-Flag-ABF3 (c), His-Flag-ABF1^ΔC4 (d) or His-Flag-ABF3^ΔC4 (e) was incubated with a 7–day-old Col or a 14–day-old keg seedling extract over the indicated time courses [note the shorter times in (d) and (e) compared with (b) and (c)]. His-Flag-tagged protein levels were visualized by anti-Flag immunoblotting. Ponceau S staining was used as the loading control. (f) Data from independent experiments comparing the in vitro degradation of full-length proteins and their respective ΔC4 forms were plotted normalized to time 0 (ABF1, n = 3; ABF3, n = 2).(g) Protein at 90 or 10 min for full-length and ΔC4 forms, respectively, were normalized to time 0. Student's t–tests indicate that the loss of the same protein was significantly slower in keg extracts compared with the wild type (WT; n = 3; P < 0.03, P < 0.02 and P < 0.03 for ABF1, ABF1^ΔC4 and ABF3 ^ΔC4, respectively). (Test not performed for ABF3, this specific time course was repeated twice, and a different time course, with the same results was performed once.) Note: cannot compare full-length and ΔC4 forms here because time points are different (see Figures 5f and 6 for these comparisons).

Figure 6

The deletion of nine C–terminal amino acids in the C4 domain destabilizes ABF1 and ABF3. Bacterially expressed recombinant His-Flag-ABF1 or His-Flag-ABF3 was incubated with 7–day-old Col seedling protein extract over the indicated time course. His-Flag-tagged protein levels were visualized by anti-Flag immunoblotting. Ponceau S staining was used as the loading control.

Figure 7

MG132 treatments result in ABF accumulation in keg. Seedlings from independent transgenic lines in the +/keg background expressing Myc-ABF1 (a) or Myc-ABF3 (b) in the KEG/keg background were grown under constant light for 14 days and treated with MG132, or DMSO as a mock treatment, for 6 h. Proteins were extracted and the results were visualized by anti-Myc and anti-ABI5 immunoblotting (b). Ponceau S staining was used as the loading control.

Figure 8

ABF1 and ABF3 protein levels are affected by ABA in the wild type (WT) but not in keg seedlings. Seedlings from independent transgenic lines in the +/keg background expressing Myc-ABF1 (a) or Myc-ABF3 (b) were grown under constant light for 14 days and treated with ABA, or ethanol as a mock treatment, for 6 h. Proteins were extracted and the results were visualized by anti-Myc or anti-ABI5 immunoblotting (b). Ponceau S staining was used as the loading control.

Figure 9

KEG ubiquitylates ABF1 and ABF3 and their C4 deletion forms in vitro. In vitro ubiquitylation assays with yeast E1, Arabidopsis His-UBC8 (E2), GST-KEG-RK (E3), ubiquitin (Ub), and His-HA-ABF1, His-HA-ABF3 or their C4 deletion forms (substrates). Substrate proteins visualized by anti-HA immunoblot (top). GST-KEG self-ubiquitylating E3 activity indicated by higher migrating forms in the anti-GST immunoblot (bottom).

Figure 10

The loss of ABF1 or ABF3 only marginally rescues the keg phenotype.(a) Chlorophyll content of 10–day-old light-grown keg seedlings compared with keg seedlings with loss-of-function mutations in one or more ABF gene or ABI5.(b) Chlorophyll content of 10–day-old light-grown KEG/KEG or KEG/keg seedlings with loss-of-function mutations in one or more ABF gene or ABI5.Comparisons were made using anova with Tukey–Kramer post-hoc tests: *P < 0.007, compared with keg alone. Results are shown as means ± SDs for three biological replicates (n = 9). All KEG/KEG or KEG/keg seedlings had significantly more chlorophyll than their corresponding keg siblings (anova, P < 0.05).