Previous studies have shown that TZF1 transcription was up-regulated by high salt and that TZF1 overexpression plants enhanced salt stress tolerance by regulating cellular ion balance and limiting oxidative and osmotic stress (Han et al., 2014). To further investigate the underlying mechanism of TZF1 in enhancing plant salt stress tolerance and determine whether the RNA-binding ability of TZF1 is required for enhancing salt stress tolerance, Arabidopsis plants overexpressing wild-type TZF1 (TZF1 OE; CaMV35S:TZF1-GFP) (Lin et al., 2011) and TZF1 mutant containing a histidine to tyrosine substitution in its second zinc finger motif (TZF1^H186Y OE; CaMV35S:TZF1^H186Y-GFP) (Li et al., 2019) were employed in this study. The TZF1^H186Y OE mutant allele was obtained from a genetic screen of ethyl methanesulfonate (EMS)-mutagenized TZF1 OE homozygous population and was isolated as a TZF1 OE intragenic mutant of the transgene CaMV35S:TZF1-GFP. Notably, TZF1 OE plants exhibit dwarf and late flowering while TZF1^H186Y OE plants are morphologically similar to wild-type (Col-0) plants with normal flowering time. The reversion of the overexpression phenotypes is likely due to loss-of-function of the mutated transgene CaMV35S:TZF1^H186Y-GFP, as evidenced by a previous study (Li et al., 2019).

Wild-type, TZF1 OE, TZF1^H186Y OE, and tzf1 T-DNA knockout seedlings were transferred to MS medium with or without NaCl or sorbitol treatment. Under salt stress (200 mM NaCl), the survival rate, fresh weight, and chlorophyll content of the wild-type, tzf1, and TZF1^H186Y OE seedlings were significantly reduced, and most seedling leaves showed bleached phenotypes. In contrast, the survival rate and chlorophyll content of TZF1 OE were significantly higher than those of the other plants ( Figures 1A-D ; Supplementary Figure 1 ). No significant differences in the leaf bleach phenotypes were observed among plants treated with 300 mM or 400 mM sorbitol ( Supplementary Figure 2 ), indicating that TZF1 is more likely to be involved specifically in salt stress response but not general osmotic stress response. To further confirm that TZF1^H186Y OE conferred a revertant phenotype of TZF1 OE, we recapitulated the mutagenesis event by overexpressing TZF1^H186Y directly in the wild-type background under the control of the CaMV 35S promoter. Upon treatment with 200 mM NaCl, TZF1^H186Y OE L8 seedlings exhibited a high mortality rate ( Supplementary Figure 3 ) as was found in the original EMS allele of TZF1^H186Y OE.

Under salt stress, plants utilize specific mechanisms to modulate Na⁺/K⁺ homeostasis. Thus, a low cytosolic Na⁺/K⁺ ratio is a key indication of salt stress tolerance (Zhao et al., 2021). Na⁺ and K⁺ concentrations and the Na⁺/K⁺ ratio were measured in leaves of Arabidopsis plants subjected to 400 mM NaCl in soil for 14 days. Under such treatment, growth of the wild-type, tzf1, and TZF1^H186Y OE plants was inhibited, while TZF1 OE had fewer withered yellow leaves compared to the other plants ( Figure 1E ). Moreover, TZF1 OE plants showed a reduced accumulation of Na⁺ ions and exhibited a lower intracellular Na⁺/K⁺ ratio during salt stress compared to wild-type plants ( Figure 1F ). These results indicate that TZF1 overexpression results in high salt-stress tolerance, suggesting that TZF1 functions as a positive regulator of salt-stress tolerance in Arabidopsis.

The production of hydrogen peroxide (H₂O₂), a major reactive oxygen species (ROS) induced by salt stress, is an indicator of oxidative damage. To further assess the role of TZF1 in salt stress response, TZF1 OE, TZF1^H186Y OE, and wild-type seedlings as well as adult plant leaves were stained with 3,3’-diaminobenzidine (DAB) to determine the H₂O₂ levels under normal and stress conditions. Under normal growth conditions, all plants leaves showed no differences in the degree of DAB staining. However, in the presence of 175 mM NaCl, the staining intensity of the TZF1 OE plants was significantly lower than that of the wild-type and TZF1^H186Y OE plants ( Figure 2 ). These results indicate that H₂O₂ accumulation in the wild-type and TZF1^H186Y OE plant leaves was greater than that in the TZF1 OE plants during salt stress and suggest that TZF1 overexpression enhanced salt tolerance by reducing ROS accumulation.

Previous studies showed that TZF1 could colocalize with both PB and SG markers in Arabidopsis protoplasts. TZF1 localized to cytoplasmic foci in intact plants and these cytoplasmic granules can be induced by either wounding or MeJA treatment (Pomeranz et al., 2010). Moreover, OsTZF1 cytoplasmic foci could be induced by NaCl treatment (Jan et al., 2013). To investigate if TZF1 cytoplasmic foci could be induced by high salt as well, we examined seven-day-old seedlings of TZF1 OE and TZF1^H186Y OE root cells with or without 200 mM NaCl treatment at 4 h and 24 h. Results showed that NaCl treatment could induce the formation of cytoplasmic foci in both TZF1 OE and TZF1^H186Y OE plants ( Figure 3 ). Although there was a significant induction in number, but the size of the cytoplasmic foci were similar ( Supplementary Figure 4 ). Interestingly, salt-induced TZF1-GFP foci appeared to be more abundant at 24 h in TZF1^H186Y OE plants than TZF1 OE plants, suggesting that TZF1^H186Y-GFP protein might be more stable, albeit non-functional, in salt stress tolerance. As the green auto-fluorescence from the cytoplasm of the light-grown seedlings interfered with the observation of GFP foci, etiolated seedlings were further examined. Again, the results confirmed that TZF1-GFP foci were salt-inducible. Consistently, both diffuse GFP and cytoplasmic granule signals were stronger in TZF1^H186Y OE plants than that in TZF1 OE plants before and after NaCl treatment ( Supplementary Figure 5 ).

Although OsTZF1 cytoplasmic foci were induced by salt stress, the identities of the foci were unknown (Jan et al., 2013). We have shown previously that TZF1 could completely co-localize with both PB (DCP2) and SG (PABP8) markers (Pomeranz et al., 2010). It was therefore confusing whether TZF proteins are components of PBs, SGs, or both. Subsequently, it was found that although plant DCP2 is a major component of mRNA decapping complex, it was not a PB-specific marker (Motomura et al., 2014), raising the possibility that TZF1 might be mainly localized in SGs and could be recruited or exchanged to PBs in response to specific cues. To test this hypothesis, TZF1 sub-cellular localization was re-examined using a set of different markers in Arabidopsis protoplast transient expression analyses (Sheen, 2001). Results showed that TZF1 did not, or only very partially, co-localize with authentic PB markers DCP1 and DCP5, respectively, but completely co-localize with SG markers G3BP and UBP1b (Sorenson and Bailey-Serres, 2014; Jang et al., 2020; Solis-Miranda et al., 2023). Importantly, both TZF1-GFP and TZF1^H186Y-GFP were completely co-localized with SG marker UBP1b ( Figure 4A ). To determine if the difference in protein abundance in intact plants was specific to fusion proteins with GFP tag, TZF1-mCherry and TZF1^H186Y-mCherry were independently expressed in Arabidopsis protoplasts. Results showed that the signal of TZF1^H186Y-mCherry was still much stronger than TZF1-mCherry ( Figure 4B ).

To further determine if the salt induction of TZF1 SG assembly was a result of change in protein abundance, immunoblot analysis using eleven-day-old seedlings was conducted. Results showed that the protein level of either TZF1-GFP or TZF1^H186Y-GFP was unaffected by salt treatment and remained nearly the same ( Figure 5A ). Therefore, salt-induced TZF1 SG assembly is likely mediated by unknown post-translational regulatory mechanisms. Noticeably, TZF1^H186Y-GFP accumulated at a higher level than TZF1-GFP before and after salt treatment ( Figure 5B ). To determine protein stability, time-course analysis using seven-day-old seedlings treated with protein synthesis inhibitor cycloheximide (CHX), proteosome inhibitor MG115/132, and a combination of CHX and MG115/132 was conducted. Results showed that TZF1 protein was extremely unstable—it almost completely disappeared after seedlings were treated with CHX for just 1 h. In contrast, TZF1 accumulation was enhanced by MG115/132. Additionally, TZF1 protein even at 1 h, had already been degraded nearly 50%, suggesting that the slow action of MG132/115 treatment could not prevent the fast turnover of TZF1. This fate of TZF1 was evidenced by the combined CHX and MG115/132 treatment in which TZF1 had already been degraded at 1 h before MG115/132 could protect it from being degraded ( Figures 5C, D ). Compared to TZF1-GFP, TZF1^H186Y-GFP was remarkably stable ( Figure 5D ), confirming that the higher steady-state accumulation of TZF1^H186Y-GFP and TZF1^H186Y-mCherry was due to enhanced protein stability. Together, these results suggest that although TZF1^H186Y-GFP is more stable, accumulates at higher levels, and localizes to SGs, it is non-functional in enhancing plant salt stress tolerance compared to wild-type TZF1.

To uncover the role of TZF1 in modulating transcript levels of salt stress-related genes, we performed deep sequencing of RNA extracted from the wild-type, TZF1 OE, and TZF1^H186Y OE plants treated with 150 mM NaCl. In total, there were nearly 1,000 up-regulated and 1176 down-regulated differentially expressed genes (DEGs) present in the TZF1 OE plants ( Figure 6A ; Supplementary Figure 6A ; Supplementary Dataset 1 ). Gene ontology (GO) analysis based on biological process clustering of DEGs in TZF1 OE plants revealed that up-regulated DEGs are over-represented by genes that are subject to ubiquitin transferase activity ( Supplementary Figure 6B ) while down-regulated DEGs preferentially associated with response to stress and response to stimulus ( Figure 6B ). Although the analysis of down-regulated DEGs in TZF1 OE is the focus of this report, we also highlight below the nexus between the outstanding up-regulated DEGs ( Supplementary Figure 6C ) and salt stress tolerance (see Discussion). For example, Na⁺/H⁺ antiporters (NHXs) (Bassil et al., 2011), CBL-interacting protein kinase 8 (CIPK8) (Yin et al., 2020), RGA-LIKE3 (RGL3) (Shi et al., 2017), SCF E3 ligase (PP2-B11) (Jia et al., 2015), annexin (AnnAt1) (Jia et al., 2015), and vacuolar protein sorting 23A (VPS23A) (Lou et al., 2020) are all implicated as positive regulators in salt stress tolerance response.

It has been shown that the TZF motif of TZF1 is required for both RNA targeting and turnover (Qu et al., 2014). Additionally, TZF1^H186Y could not trigger the decay of a reporter gene containing AU-rich elements (Li et al., 2019). Since the H186Y mutation is in the second zinc finger region, we tested whether the TZF region is important for TZF1 targeting and subsequent degradation of mRNAs in response to salt stress. RNA immunoprecipitation coupled with qPCR (RIP-qPCR) was performed to determine if any of the down-regulated genes were potential direct targets of TZF1 under salt stress. Eight down-regulated DEGs in TZF1 OE plants were chosen from the set of salt or oxidative stress-related mRNAs containing ARE-like motifs in 3’-UTR ( Figure 6C ). Among these genes, the autoinhibited Ca²⁺-ATPase genes ACA11 and ACA9 encoding calcium pumps (Li Z. et al., 2023), KAT3 encoding a subunit of potassium channel (Sun et al., 2015), ARP1 encoding an ABA-regulated RNA-binding protein 1 (Jung et al., 2013), and Hsp90.5 encoding a heat shock protein (Song et al., 2009) are all implicated in negative regulation of the salt stress tolerance response. Although ACA4 was also strongly down-regulated by salt stress in TZF1 OE plant in the RNA-seq analysis ( Supplementary Figure 7 ), ACA4 mRNA does not contain typical AREs or UREs in its 3’-UTR, and thus was not selected for further analysis. RIP-qPCR results showed that TZF1, but not the TZF^H186Y, could bind the F1 region of ACA11 mRNA ( Figures 6D, E ). No significant TZF1 binding signals were detected from the rest of the seven potential target genes ( Supplementary Figure 8 ; Supplementary Datasets 2 , 3 ). These results indicate that ACA11 down-regulation is a potential cause of enhanced salt stress tolerance in TZF1 OE plants.

To further examine whether TZF1 binds ACA11 mRNA directly, RNA electrophoretic mobility shift assays (EMSAs) were performed. The migration of an F4 probe, an extended derivative of the F1 region bound by TZF1 in the RIP-qPCR analysis ( Figures 6D, E ), was retarded by GST-TZF1 ( Figure 7A ). However, most of the bound complex was unable to enter the gel, so the binding affinity could not be accurately determined. Previous studies showed that both the arginine-rich (RR) and TZF domains of TZF1 were required for high-affinity ARE RNA binding (Qu et al., 2014). In addition, ARE₁₉ was shown to be a conserved TZF1 bound RNA motif in 3’-UTR of both plant (Qu et al., 2014) and human (Brooks and Blackshear, 2013) genes. Accordingly, recombinant MBP-TZF1 (RR-TZF) was overexpressed and purified for further EMSAs. Our results show that ARE₁₉ shifted readily upon incubation with increasing concentrations of MBP-RR-TZF ( Supplementary Figure 9A ). This clear interaction between ARE₁₉ and MBP-RR-TZF was leveraged in competition experiments, which were used to estimate the binding affinity of ACA11 F1, F4, and F4 deletion derivatives for TZF1. In these assays, unlabeled RNA probes were introduced as competitors to the MBP-RR-TZF–ARE₁₉ protein–RNA complex. The resulting gel images showed that F4 could disrupt MBP-RR-TZF binding to ARE₁₉ ( Figure 7B ). Interestingly, F4 exerted stronger competition than the F1, F4Δ84, and F4Δ134 derivatives ( Figure 7C ). Closer examination of the F4 nucleotide sequence revealed two U-rich segments; each of the deletion derivatives contains only one of these segments. Taken together, these results indicated that TZF1 binds ACA11 mRNA at the 3’-UTR and that recognition of this region is dependent upon two U-rich segments in the 3’-UTR. To directly assess if TZF1^H186Y was defective in mRNA binding, additional EMSAs were conducted. Results showed that MBP-RR-TZF^H186Y was unable to bind the consensus ARE₁₉ motif ( Supplementary Figure 9A ) and that binding between MBP-RR-TZF^H186Y and F4 motif was strongly impaired ( Supplementary Figure 9B ), suggesting that TZF1^H186Y might be defective in mRNA binding in general.

To investigate if reduced ACA11 mRNA levels in TZF1 OE lines were due to TZF1 targeting ACA11 mRNA for degradation, mRNA half-life analysis was conducted. An Arabidopsis protoplast transient expression system was used for this assay. In brief, TZF1-mCherry and TZF1^H186Y-mCherry were used as effectors and GFP-ACA11-F3 (does not bind TZF1) and GFP-ACA11-F4 (binds TZF1) were used as reporters ( Figure 8A ). Protoplast samples were co-transformed with a distinct effector and reporter pair and incubated for 10 h before mRNA half-life time-course experiments were conducted using actinomycin D at 0 h to block further transcription. Results showed that degradation of GFP-ACA11-F4 mRNA was faster in the presence of TZF1 compared to TZF1^H186Y ( Figures 8B, C ). In contrast, a negligible difference in GFP-ACA11-F3 half-life was observed when it was co-expressed with either TZF1 or TZF1^H186Y, in agreement with our observation of specific binding of TZF1 to the F1/F4 region of ACA11 in RIP-qPCR experiments ( Figures 6D, E ). To determine if the effectors were expressed at similar levels and the expression of the reporters correlated with corresponding mRNA half-life, we employed fluorescence microscopy. We found that the accumulation of GFP-ACA11-F4, but not GFP-ACA11-F3, was significantly reduced in the presence of TZF1 ( Figure 8D ), consistent with the results of mRNA half-life assays ( Figures 8B, C ). Immunoblot analysis further confirmed that TZF1 could specifically reduce the accumulation of GFP-ACA11-F4 protein. Note that the decay of GFP-ACA11-F4 protein was not obvious after 4 h, likely due to the accumulation and stability of the GFP protein ( Figure 8E ; Supplementary Figure 10 ). Together, these results suggest that TZF1 binds and enhances turnover of ACA11 mRNA.

To determine the role of ACA11 in salt stress tolerance, we compared the salt stress tolerance of wild-type and aca11 T-DNA insertional mutant under salinity conditions. The aca11 has the T-DNA inserted into the sixth exon of the ACA11 ( Figure 9A ). RT-PCR results revealed that ACA11 transcript was absent in homozygous aca11 line ( Figures 9B, C ). In the presence of 200 mM NaCl, aca11 seedlings showed a significantly higher survival rate than wild-type seedlings ( Figures 9D, E ). These results indicate that ACA11 plays a negative role in salt stress tolerance and that TZF1 targets ACA11 mRNA for degradation to enhance salt stress tolerance.

Wild-type Arabidopsis thaliana Columbia-0 (Col-0), TZF1 OE, TZF1^H186YOE, tzf1 (SALK_143721), and aca11 (SALK_060930) plants were used in this study. The plant seeds were sterilized and plated on Murashige and Skoog (MS) medium (Phytotech) for 2 days at 4°C, and then transferred to a growth chamber with 16/8 h light/dark cycles at 22°C. The light intensity was 50 µmol m^{− 1} s^{− 1}.

Seven-day-old seedlings grown on MS plates were transferred to new MS plates containing different concentrations of NaCl for another eight days under 16/8 h light/dark cycles. The seedlings with completely bleached cotyledons were counted as dead plants; all others were counted as viable plants. At least 35 seedlings were used for each treatment.

Sixteen seedlings were placed in a 2-ml centrifuge tube and the plant fresh weight was recorded as W. To each tube, 1 ml aliquots of 80% (v/v) acetone were added to immerse the plants. The samples were then placed in the dark for 48 h at 22°C. The acetone extract was isolated and its volume was recorded as V. For each extract, absorbance at 645 nm and 663 nm was measured using a spectrophotometer (TECAN Infinite M200 Pro) and chlorophyll content was calculated using the formula [(8.02 × Abs ₆₆₃) +(20.21 × Abs ₆₄₅)]×V/W as previously described (Li et al., 2021).

Accumulation of Na⁺ and K⁺ ions was measured according to a previously described (Jiang et al., 2019). Plants grown in the presence or absence of NaCl treatment were separately harvested and dried for 48 h at 65°C before their weight was measured. All samples were then digested with 75% (v/v) nitric acid and 25% (v/v) hydrogen peroxide at 180°C for 3 h. The samples were then diluted to 30 ml with ddH₂O and filtered. The Na⁺ and K⁺ contents in each solution were measured using an inductively coupled plasma optical emission spectrometer (ICAP6300).

Seedlings and leaves were incubated in DAB staining solution (1 mg/ml DAB dissolved in ddH₂O, pH 3.8) in the dark at 22°C for 8 h. The samples were then destained with acetic acid:glycerin:ethanol (1:1:4 by volume) at 80°C for 25 min before imaging. Images of DAB staining results were quantified by using Image J (NIH).

Eleven-day-old seedlings grown on MS plates under 12/12 h light/dark cycles were treated with 150 mM NaCl at Zeitgeber time (ZT)2 for 4 h before sample collection and total RNA extraction. Three biological replicates were prepared for each sample. Total RNA was extracted using Trizol reagent (Invitrogen). The RNA was then treated with DNase I (DNA-free™ DNA Removal Kit, Invitrogen) to remove genomic DNA. RNA quality was evaluated on a Bioanalyzer 2100 instrument (Agilent). Sequencing libraries were prepared using the Directional RNA Library Prep Kit (E7760S, New England Biolabs) according to manufacturer’s instructions. The 150-nucleotide (nt) paired-end high-throughput sequencing was performed using an Illumina Hiseq X TEN platform. After removing the low-quality sequencing reads, the remaining clean reads were mapped to the Arabidopsis reference genome (TAIR10) using Tophat2 software. DEGs were analyzed using edgeR software. Genes with q < 0.05 and |log 2 _ratio| ≥ 1.5 were identified as DEGs. The enrichment of DEGs in different functional categories was performed by using agriGO (GeneOntology) V2.0. Venn diagram analysis was conducted using VENNY 2.1. Three biological replicates were prepared for each sample. To select candidate DEGs for further analysis, there were 339 genes in the GO “response to stimulus” and 210 genes in the GO “response to stress” category. Salt stress is commonly caused by high concentrations of Na⁺. The salt stress induced not only ionic stress but also other secondary stresses, which include the accumulation of toxic compounds such as ROS. The plant cell minimizes cellular damage by reducing cytoplasmic Na⁺ and limiting the ROS concentration in the cell (Yang and Guo, 2018). Therefore, the genes in the GOs “response to stimulus” and “response to stress” were used for BLAST search and yielded 31 genes responding to salt stress, 12 genes involved in hydrogen peroxide catabolic process and ROS metabolic process, and 10 genes involved in inorganic cation transmembrane transport. Among these 53 genes, 8 genes containing ARE-like motifs in 3’-UTR of the mRNA were selected for the subsequent RIP-seq analysis.

Eleven-day-old seedlings grown on MS plates under 12/12-h light/dark cycles were treated with 150 mM NaCl at ZT2 for 4 h before samples were collected for crosslinking. RIP was performed as previously described (Koster et al., 2014). Briefly, the cross-linked tissues (1.5 g) were ground with liquid nitrogen and resuspended in 750 µl prewarmed (60°C) RIP lysis buffer to make viscous homogenates. After centrifugation, the cell extract supernatant was filtered through a 0.45 µm filter. The extract was then pre-cleared twice with 50 µl Sepharose beads before mixing with 15 µl washed Sepharose beads coated with GFP antibody (Invitrogen). The beads were incubated with extract for 2 h at 4°C and then washed three times with RIP washing buffer for 10 min at 4°C. The beads were then washed with RIP lysis buffer for 5 min at 4°C. The RNA was purified from the immunoprecipitated RNPs and 100 µl input, respectively, using the Trizol reagent (Invitrogen). For quantitative analysis of TZF1 binding to RNA, reverse transcription was performed using SuperScript™ IV First Strand Synthesis System. After DNase treatment, RNA samples were reverse transcribed with random hexamer primers. Quantitative PCR was performed to determine the level of TZF1-bound RNA, and the 2^−ΔCT method was used to calculate the ratio of RIP to the input.

Seven-day-old TZF1 OE and TZF1^H186Y OE seedlings were incubated with 30 μM cycloheximide (CHX), 50 μM MG115/132, or both CHX and MG115/132. Seedlings were collected at different time points, and then total proteins were extracted and protein levels were determined by immunoblot analysis with anti‐GFP antibody (Roche).

Recombinant GST-TZF1 and MBP-TZF1 (RR-TZF) proteins were overexpressed in Escherichia coli BL21 (DE3) cells. Bacterial cultures were grown to Abs ₆₀₀ = 0.6, at which point 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was added for induction of protein production. ZnCl₂ (0.1 mM) and glucose [0.2% (w/v); MBP-RR-TZF only] were also added at the time of addition induction. Cultures were then grown for an additional 12-16 h at 18°C before harvesting. GST-TZF1 was purified using glutathione-Sepharose 4B resin (GE Healthcare Bio Sciences) as previously described (Qu et al., 2014). For purification of MBP-RR-TZF, cell pellets from a 1 L overexpression culture were resuspended in 20 ml extraction buffer [50 mM Tris-HCl, pH 8.5; 100 mM NaCl; 1 mM EDTA, 1 mM DTT, 1 mM PMSF, 0.5X EDTA-free protease inhibitor cocktail (APExBIO)]. The resuspended cells were lysed at 15,000 psi using a French pressure cell press, and the resulting whole-cell lysate was supplemented with 0.2% (v/v) Triton X-100 and incubated at 4°C for 10 min with gentle nutation. The lysate was then supplemented with 10 mM MgCl₂ and 10 mM CaCl₂ and 5 µg DNase I (Sigma-Aldrich) and incubated at 20°C for an additional 10 min. After incubation, the lysate was clarified by centrifugation at 12,000 x g and 4°C for 15 min and then filtered using a 0.45-µm syringe filter (Roche). The filtered lysate was passed over a gravity-flow column packed with 0.5 ml amylose resin (New England Biolabs) that had been pre-equilibrated with extraction buffer. After collecting the flowthrough, the amylose resin was washed with 20 ml extraction buffer to remove weakly bound proteins. Bound MBP-RR-TZF was eluted with 0.5 ml additions of elution buffer (50 mM Tris-HCl, pH 8.5, 10 mM amylose, 1 mM DTT) and individual elution fractions were collected and assessed by SDS-PAGE. The fractions containing near-homogeneous MBP-RR-TZF were pooled and dialyzed twice against 500 ml storage buffer (20 mM Tris-HCl, pH 8.0; 50 mM NaCl; 1 mM DTT) at 4°C. The final protein concentration was determined by measuring Abs₂₈₀ (ϵ = 81,820 M^-1 cm^-1). Prior to use in competition assays, aliquots of MBP-RR-TZF were concentrated to ~80 μM using a Vivapsin 500 10,000 MWCO centrifugal concentrator (Sartorius). All protein aliquots were stored at -80°C.

The F4, F4Δ84, F4Δ134, and F1 RNAs were generated by run-off in vitro transcription (IVT) using T7 RNA polymerase. The DNA templates for IVT were prepared by PCR using cDNA as the template. The F4, F4Δ84, and F4Δ134 IVT templates were amplified using F4-F, F4Δ84-F, and F4Δ134-F forward primers, respectively, and F4-R as the common reverse primer. Due to poor amplification of the F4Δ134 template after the first PCR, the amplicon from the first round was used as the template in a second PCR that used the same set of primers as in the first round. IVT reactions contained 1x IVT buffer [40 mM Tris-HCl, pH 7.6; 24 mM MgCl₂; 2 mM spermidine; 0.01% (v/v) Triton X-100]; 10 mM DTT; 5 mM each of ATP, CTP, GTP, and UTP; 0.002 U thermostable inorganic pyrophosphatase (NEB); 0.15-0.3 µg template, and T7 RNA polymerase (purified in-house)]. After 4 h at 37°C, the reactions were treated with DNase I (Roche), extracted with phenol-chloroform, and dialyzed using 3,500-MWCO tubing (BioDesign Inc. New York) against 3.5 L ddH₂O three times over 20 h at 4°C and once over 1.5 h at 20°C. Following dialysis, the RNAs were precipitated with 0.3 M sodium acetate and 2.5 volumes of ethanol, resuspended in autoclaved ddH₂O, and quantitated by measuring the absorbance at Abs ₂₆₀ and using their respective extinction coefficients (OligoAnalyzer, IDT).

To prepare 5′-[³²P]-labeled F4 (5′-[³²P]-F4), F4 was first dephosphorylated by incubating with calf intestinal phosphatase (NEB) for 2 h at 37°C. Dephosphorylated F4 was then extracted with phenol-chloroform, precipitated with sodium acetate and ethanol, and resuspended in ddH₂O as described above. Following clean-up, dephosphorylated F4 (5 µM) was incubated with γ-[³²P]-ATP and T4 polynucleotide kinase (NEB) for 45 min at 37°C. The labeling reaction was quenched with the addition of urea dye [7 M urea, 1 mM EDTA, 0.05% (w/v) xylene cyanol, 0.05% (w/v) bromophenol blue, 10% (v/v) phenol], loaded onto a denaturing 8% (w/v) polyacrylamide/7 M urea/1X Tris-Borate-EDTA (TBE) gel, and electrophoresed in 1X TBE buffer for 75 min. Full-length 5′-[³²P]-F4 was identified in the gel by autoradiography, excised from the gel, and eluted using the crush-and-soak method. Eluted 5′-[³²P]-F4 was precipitated with sodium acetate and ethanol and resuspended to a final specific activity of 200,000 dpm/µl in ddH₂O.

For TZF binding to ARE_19, a pre-mixed solution of 100 nM 6-FAM-labeled ARE₁₉ (6-FAM-ARE₁₉, IDT) and 40 µM 20-nt decoy oligo (MilliporeSigma) in binding buffer [20 mM HEPES-KOH, pH 7.5; 50 mM NaCl; 5 µM ZnCl₂, 4 mM MgCl₂] was aliquoted into multiple tubes. Importantly, inclusion of the 20-nt decoy oligo minimized apparent nucleolytic decay of the target RNA during the binding reaction and improved the resolution of free 6-FAM-ARE₁₉ bands in the gel image. MBP-RR-TZF (or H186Y mutant) was serially diluted from 40 μM to 156 nM by diluting with equal volumes of storage buffer [20 mM Tris-HCl, pH 8; 50 mM NaCl; 1 mM DTT]. To initiate the binding reaction, equal volumes of MBP-RR-TZF were combined with the pre-mixed aliquots of 6-FAM-ARE₁₉ and decoy oligo. As a negative control, a single aliquot of 6-FAM-ARE₁₉ and decoy oligo was diluted with an equal volume of storage buffer. The binding reactions were incubated at 23°C for 10 min. Loading dye containing xylene cyanol, bromophenol blue, and 50% (v/v) glycerol was added to each tube, and each reaction mix was loaded onto a native 8% (w/v) polyacrylamide (29:1 acrylamide:bisacrylamide)/1X Tris-Borate (TB) gel. The gel was electrophoresed in 1X TB buffer at 100 V for 1 h at 4°C and then scanned using an Amersham Typhoon Biomolecular Imager (Cytiva) on the Cy2 setting.

For TZF binding to ACA11-F4, a pre-mixed solution of 1 nM 5′-[³²P]-F4 and 40 µM 20-nt decoy oligo in binding buffer was aliquoted into multiple tubes. MBP-RR-TZF (or H186Y mutant) was serially diluted from 40 μM to 156 nM by diluting with equal volumes of storage buffer. To initiate the binding reaction, equal volumes of MBP-RR-TZF were combined with the pre-mixed aliquots of 5′-[³²P]-F4 and decoy oligo. The binding reactions were incubated and resolved by native polyacrylamide gel electrophoresis as described for 6-FAM-ARE₁₉, except the binding reactions were resolved on a 6% (w/v) polyacrylamide/1X TB gel, and the gel was exposed to storage phosphor screen which was then scanned using an Amersham Typhoon Biomolecular Imager (Cytiva) on the phosphorimager setting.

The binding affinity of MBP-RR-TZF for ACA11 F4 and deletion derivatives (F1, F4Δ84, F4Δ134) was measured in a gel-shift assay by monitoring dissociation of MBP-RR-TZF from 6-FAM-ARE₁₉, which was previously established as a minimal RNA ligand for MBP-RR-TZF (Qu et al., 2014). For all competition assays, a fixed concentration of 20 µM MBP-RR-TZF was used to ensure near-complete binding to ARE₁₉ in the absence of competitor RNA, and then subject to a competition with increasing concentration of either F4 or one of the F4 deletion derivatives. All reported RNA and protein concentrations reflect the final concentrations in the binding assay. A pre-mixed solution of 50 nM 6-FAM-ARE₁₉ and 20 µM 20-nt decoy oligo in binding buffer (20 mM HEPES-KOH, pH 7.5; 50 mM NaCl; 5 µM ZnCl₂, 4 mM MgCl₂) was aliquoted into multiple tubes and supplemented with increasing concentrations of competitor RNA. Two aliquots that were designated as positive- and negative-binding controls were supplemented with ddH₂O instead of competitor RNA. To the binding reaction and positive-control tubes, 20 µM MBP-RR-TZF was added, whereas storage buffer was added to the negative control tube. The binding reactions were incubated and resolved by native polyacrylamide gel electrophoresis as described for 6-FAM-ARE₁₉. Following electrophoresis, gels were scanned using an Amersham Typhoon imager (Cytiva) on the Cy2 setting, and the intensity of bands corresponding to bound and unbound 6-FAM-ARE₁₉ were quantified using ImageQuant software (Molecular Dynamics). The fraction of 6-FAM-ARE₁₉ bound at each competitor concentration was normalized using the fraction of ARE₁₉ bound in the positive control sample and plotted as % bound ARE₁₉. Data were fit to a hyperbolic binding isotherm ([ΔARE₁₉ bound]*K_comp )/([competitor] + K _comp) to yield the K _comp value for binding of MBP-RR-TZF to a competitor RNA.

The TZF1, TZF1^H186Y, ACA11-F3, and ACA11-F4 were cloned into the pENTR^™/D-TOPO^® vectors. All constructs were subcloned into the Gateway^® destination vectors with N-terminal GFP or C-terminal mCherry tag by using the LR recombination reaction. Arabidopsis protoplast samples were co-transformed with a distinct effector and reporter pair (TZF1+ACA11-F3, TZF1+ACA11-F4, TZF1^H186Y+ACA11-F3, TZF1^H186Y+ACA11-F4) and incubated for 10 h before mRNA half-life time-course experiments were conducted. Actinomycin D (Sigma) at 100 μg/ml was used to block the transcription at the beginning of the time-course experiments. Total RNA was extracted by using a RNeasy plant mini kit (Qiagen), then treated with TURBO DNA-free™ Kit (Ambion) to remove contaminating genomic and plasmid DNA. First-strand cDNA was synthesized by using the SuperScript III reverse transcriptase (Invitrogen) according to the supplier’s instructions. ACTIN1 was used as an internal control for RT-PCR analysis.