Plant Tandem CCCH Zinc Finger Proteins Interact with ABA, Drought, and Stress Response Regulators in Processing-Bodies and Stress Granules

Abstract

Although multiple lines of evidence have indicated that Arabidopsis thaliana Tandem CCCH Zinc Finger proteins, AtTZF4, 5 and 6 are involved in ABA, GA and phytochrome mediated seed germination responses, the interacting proteins involved in these processes are unknown. Using yeast two-hybrid screens, we have identified 35 putative AtTZF5 interacting protein partners. Among them, Mediator of ABA-Regulated Dormancy 1 (MARD1) is highly expressed in seeds and involved in ABA signal transduction, while Responsive to Dehydration 21A (RD21A) is a well-documented stress responsive protein. Co-immunoprecipitation (Co-IP) and bimolecular fluorescence complementation (BiFC) assays were used to confirm that AtTZF5 can interact with MARD1 and RD21A in plant cells, and the interaction is mediated through TZF motif. In addition, AtTZF4 and 6 could also interact with MARD1 and RD21A in Y-2-H and BiFC assay, respectively. The protein-protein interactions apparently take place in processing bodies (PBs) and stress granules (SGs), because AtTZF5, MARD1 and RD21A could interact and co-localize with each other and they all can co-localize with the same PB and SG markers in plant cells.

Fig 1. Putative AtTZF5 interacting partners identified by Y-2-H screens.

(A) The representation of 35 proteins identified by Y-2-H screens. (B) Nineteen out of 35 proteins are involved in stress responses. (C) Tissue expression patterns of 35 identified protein-coding genes.

Fig 2. AtTZF5 interacts with MARD1 and RD21A.

(A) AtTZF5 interacts with MARD1 and RD21A in a Y-2-H analysis. AtTZF5 was fused with GAL4 DNA binding domain (BD), whereas MARD1 and RD21A were fused with GAL4 activation domain (AD). The bZIP1+bZIP25 was used as a positive control pair. (B) Results of bimolecular fluorescence complementation (BiFC) analysis indicate that AtTZF5 interacts with MARD1 and RD21A in the cytoplasmic foci in Arabidopsis protoplasts. Interaction between bZIP1-CYFP and bZIP25-NYFP in the nucleus was used as a positive control. Images of cells with positive YFP signals were taken by exposing under green channel. Whereas images of cells without YFP signals were taken using all three channels (red, green, and blue) to show cell integrities (red fluorescence from chloroplasts). These experiments were repeated twice. Bar = 10μm.

Fig 3. MARD1 and RD21A can interact with AtTZF4 and AtTZF6.

(A, B) MARD1 and RD21A can interact with AtTZF4 and 6 but not AtTZF1 in Y-2-H analysis. AtTZF1, 4, and 6 were fused with GAL4 DNA binding domain (BD), whereas MARD1 and RD21A were fused with GAL4 activation domain (AD). (C) BiFC results indicate that MARD1 and RD21A can interact with AtTZF4 and AtTZF6 in cytolasmic foci in Arabidopsis protoplasts. Images of cells with positive YFP signals were taken by exposing under green channel. Whereas images of cells without YFP signals were taken using all three channels (red, green, and blue) to show cell integrities (red fluorescence from chloroplasts). These experiments were repeated twice. Bar = 10μm.

Fig 4. AtTZF1 (FL), RR-TZF and TZF of AtTZF5 were expressed in duplicate yeast cell lines.

Shown are results of Western blot analyses. Full-length AtTZF1 as well as RR-TZF and TZF of AtTZF5 were fused with GAL4 DNA binding domain in HA tagged pAS1 vector. HA-ZTL was used as a positive control for the expression in yeast cells.

Fig 5. TZF and RR-TZF fragments of AtTZF5 cannot interact with MARD1 and RD21A in Y-2-H analysis.

(A) AtTZF5 fragments used in this analysis. (B) Yeast cells fail to grow due to negative interaction between AtTZF5 truncated fragments (TZF and RR-TZF) and MARD1/RD21A. TZF, and RR-TZF of AtTZF5 were fused with GAL4 DNA binding domain (BD), whereas MARD1 and RD21A were fused with GAL4 activation domain (AD). This experiment was repeated three times.

Fig 6. TZF and RR-TZF fragments of AtTZF5 can interact with MARD1 or RD21A in cytoplasmic foci.

Shown are results of BiFC analysis using an Arabidopsis protoplast transient expression system. Images of cells with positive YFP signals were taken by exposing under green channel. Whereas images of cells without YFP signals were taken using all three channels (red, green, and blue) to show cell integrities (red fluorescence from chloroplasts). This experiment was repeated twice. Bar = 10μm.

Fig 7. TZF domain of AtTZF5 is sufficient for interaction with MARD1 in co-immunoprecipitation analysis.

AtTZF5 FL, TZF and RR-TZF fragments as well as GASA6 and TOC1 were tagged with GFP. MARD1 and ZTL were tagged with HA. GASA6 was used as a non-interacting control with MARD1. TOC1 and ZTL were used as a positive interacting pair. (A) Left panel shows signals of various input GFP tagged proteins (indicated by arrows). Anti-GFP antibody was used to pull down GFP-tagged proteins and revealed by Western blot analysis (arrow in right panel). (B) Left panel shows signals (indicated by arrows) of various input HA tagged proteins. Co-IP products were detected by anti-HA antibody as indicated by arrows in right panel.

Fig 8. MARD1 and RD21A can co-localize with AtTZF5 and PB (DCP2) and SG (UBP1b) markers in protoplast transient expression analyses.

(A) MARD1 and RD21A can co-localize with AtTZF5 in cytoplasmic foci. (B) MARD1 and RD21A can co-localize with PB marker DCP2. (C) MARD1 and RD21A can co-localize with SG marker UBP1b. Cellular images for GFP and mCherry signals were taken using green and red channel, respectively. Bright field images were also shown for cell integrities. Bar = 10μm.