ABF transcription factors of Thellungiella salsuginea

Abstract

ABF transcription factors are the key regulators of ABA signaling. Using RACE-PCR, we identified and sequenced the coding regions of four genes that encode ABF transcription factors in the extremophile plant Thellungiella salsuginea, a close relative of Arabidopsis thaliana that possesses high tolerance to abiotic stresses. An analysis of the deduced amino acid sequences revealed that the similarity between Thellungiella and Arabidopsis ABFs ranged from 71% to 88%. Similar to their Arabidopsis counterparts, Thellungiella ABFs share a bZIP domain and four conservative domains, including a highly conservative motif at the C-terminal tail, which was reported to be a canonical site for binding by 14-3-3 regulatory proteins. Gene expression analysis by real-time PCR revealed a rapid transcript induction of three of the ABF genes in response to salt stress. To check whether Thellungiella ABF transcription factors can interact with abundant 14-3-3 proteins, multiple constructs were designed, and yeast two-hybrid experiments were conducted. Six of the eight tested Ts14-3-3 proteins were able to bind the TsABFs in an isoform-specific manner. A serine-to-alanine substitution in the putative 14-3-3 binding motif resulted in the complete loss of interaction between the 14-3-3 proteins and the ABFs. The role of 14-3-3 interaction with ABFs in the salt and ABA signaling pathways is discussed in the context of Thellungiella survivability.

Introduction

The phytohormone abscisic acid (ABA) is one of the key factors that govern diverse developmental and adaptation processes in the plant. This molecule regulates the synthesis of seed storage proteins, seed dormancy, maturation, seed germination and root growth.¹ It is extensively involved in adaptation to biotic and abiotic stresses, in particular, drought, salinity, heat and low temperatures.^2-4 Due to intensive agriculture practices and global climatic changes, drought and high salinity have become two of the most important factors that negatively affect crop productivity worldwide. Under such conditions, plant cells accumulate ABA, which in turn triggers multiple physiological and biochemical changes, including stomatal closure and the reprogramming of stress-responsive gene expression.^4-6

Numerous stress-responsive genes were found to be up- or downregulated under the action of ABA.^7-9 Promoter analysis of ABA-inducible genes revealed a conserved cis-acting element, designated as ABA-responsive element (ABRE). A family of ABRE-binding transcription factors was isolated from Arabidopsis, using the yeast one-hybrid system and ABRE sequences as a bait. These transcription factors were named ABA-responsive element binding factors (AREBs) or ABRE binding factors (ABFs).^10,11 The subfamily of ABF/AREB proteins belongs to the family of basic-region/leucine zipper (bZIP) transcription factors.¹² Members of this subfamily of transcriptional regulators are key factors determining ABA-mediated transcriptional regulation of stress-related genes. ABF/AREB family members comprise a highly conserved bZIP domain in their C-terminal tails; three N-terminal conserved domains, designated as C1, C2 and C3; and a C4 domain, located in the C-terminus.¹¹ To bind DNA and regulate the expression of downstream genes, ABFs must form dimers.¹² In addition, ABFs/AREBs themselves demonstrate responsive expression patterns to various abiotic stresses.^10,11

Little was known about the mechanism of ABF activation until recently. ABA-dependent activation of ABF2 was found to be repressed by the addition of protein kinase inhibitors, indicating that ABA-dependent phosphorylation may be essential for the functional activation of ABFs. In fact, the conserved N- and C-terminal domains of the ABF proteins contain invariable serine/threonine residues that can be phosphorylated by different kinases.^13,14 It was shown recently that several protein kinases belonging to the CDPK and SnRK2 families phosphorylate specific motifs containing Ser/Thr residues and thereby influence the transactivation activity of ABFs in transient assays.^14-16

In addition, a number of studies revealed that multiple members of the 14-3-3 regulatory protein family act as mediators in ABA signaling through direct interaction with ABFs from barley, Arabidopsis and rice.^16-18 However, little is known about the specificity and the roles of such interactions. Acting as dimers, 14-3-3s are known to be able to form a variety of heterodimeric complexes with the interacting proteins; thus, the specificity of binding may play an important role in the formation of such regulatory structures. Furthermore, the regulatory proteins of the transcriptional complex may stabilize the dimeric structure of the transcription factors. Indirect evidence that 14-3-3 adaptor proteins can play such a role in ABA signaling arose once these proteins were identified for the first time in plants as belonging to DNA G-box binding complexes.¹⁹ Later, 14-3-3s were shown to interact with the ABA signaling effector viviparous 1 (VP1) in a yeast two-hybrid assay.²⁰

Because Thellungiella, a close relative of Arabidopsis, is known for its greater tolerance to various stresses, especially high salinity, our aims were to analyze the structure of ABF genes in this extremophile species and to measure the effect of high salinity on the expression of ABFs. As 14-3-3 proteins were proposed to be a part of the ABA signaling pathway, we were interested in whether the multiplicity of 14-3-3 isoforms plays a role in ABF regulation.

Results

Identification of novel ABF genes

Using degenerate primers, four central fragments of the putative ABF genes were amplified by PCR (Fig. S1). The 5′ and 3′ end sequences were determined using the RACE technique. The newly identified genes were named TsABF1–4 according to their most similar Arabidopsis counterparts (GenBank accession numbers JQ971971, JQ971972, JQ971973 and JQ971974, respectively). The similarities of the deduced amino acid sequences of Thellungiella ABF1, ABF2, ABF3 and ABF4 to their Arabidopsis homologs were 71, 77, 85 and 88%, respectively. The alignments of the Thellungiella isoforms with their homologs from Arabidopsis are given in Figure S2. All of the identified isoforms share the bZIP domain with four leucine repeats as well as 3 N-terminal (C1, C2 and C3) and one C-terminal (C4) conserved domains (Fig. 1a). Similarly to those from Arabidopsis, the ABFs from Thellungiella contain invariable Ser/Thr residues in their conserved domains (Fig. 1b). These residues are the putative targets for Ser/Thr kinases. Moreover, the C4 conserved domain contains the canonical mode II 14-3-3 interaction motif (RRTLT/SGP) (Fig. 1b), shown to be involved in the interaction with 14-3-3 proteins in barley, Arabidopsis and rice.^16,18,21 Phylogenetic analysis shows that the Thellungiella ABFs, together with those of Arabidopsis, belong to the same clade in the phylogenetic tree of the nine subfamily members (Fig. 2).

Figure 1. Sequence analysis of Thellungiella (ecotype Yakutsk) ABF transcription factors. A - Structure of the ABF family proteins. bZIP, basic-region/leucine zipper domain. C1-C4, conserved domains within the family members. B - Alignment of the derived amino acid sequences of the Thellungiella ABFs. The conserved domains are framed. Residues in black are 75% conserved within the domains. The Ser/Thr residue, which is essential for phosphorylation-dependent interaction with 14–3-3s, is denoted by an arrowhead. A multiple sequence alignment was performed using the ClustalW program.

Figure 2. A non-rooted phylogenetic tree, showing the relationship between the identified ABF isoforms of Thellungiella and known proteins from the ABF family of Arabidopsis.

ABF gene expression in Thellungiella plants, exposed to stress

Gene expression changes were monitored via real-time RT-PCR. To determine whether the expression of ABFs depends on photoperiodicity, we performed measurements at different time points over 24 h, using the ABF2 expression changes as an example (Fig. S3). A slight increase in the expression of ABF2 (10%) was detected in the shoots 8 h after the light went on, which may have been due to the transpirational water loss. As shown in Figure 3, the expression of ABF2 and ABF3 genes was strongly induced by salinity. The induction of the ABF4 gene was moderate. The expression of ABF1 was not significantly affected by the stress (data not shown). In the shoot and roots, ABF2 showed the strongest induction (Fig. 3a and b). Generally, salt stress resulted in a rapid induction of the ABF genes in the shoot with peaks between 2 and 8 h after the onset of stress, whereas the expression returned to the near basal level after 24 h. As ABFs are known for their great importance in ABA-mediated stomatal closure, such a quick expression change in the ABF genes might play a crucial role in plant adaptation by reducing transpiration levels. The stress-induced upregulation of ABF gene expression lasted longer in the roots than in the shoots.

Figure 3. Salt stress induced ABF gene expression in the shoots (A) and roots (B) of Thellungiella plants, ecotype Yakutsk (control = 1). Six-week-old plants, grown hydroponically, were treated with 200 mM NaCl. The roots and shoots were harvested at 0, 2, 8 and 24 h after the plants were exposed to stress. Total RNA was isolated, converted to cDNA and subjected to comparative real-time RT–PCR quantification. The relative expression levels (relative units, reu) were normalized using the level of Actin gene expression and calculated with respect to the gene expression of the control plants, considered as 1. The bars show the means ± 95% C.I. (n = 3).

ABF transcription factors interact with 14-3-3 regulatory proteins in a yeast two-hybrid screen

In our yeast two-hybrid screen, we used eight isoforms of Thellungiella 14-3-3 as bait. ABF2, with the RRTESGP motif in its C-terminal tail, and ABF4, with the motif RRTLTGP, were used as prey. Yeast cells, transformed with the different 14-3-3 isoforms identified by our group earlier,²² ABF2 and ABF4, were examined for their growth on selective medium (SD-LWHA) due to the activation of the reporter HIS3 and ADE2 genes (Fig. 4a). Six of the tested 14-3-3 proteins were able to interact with both ABF transcription factors. Remarkably, yeast cells transformed with the 14-3-3 ε (Epsilon), ο (Omicron) and φ (Phi) isoforms either did not grow or grew poorly. To quantitatively evaluate the interaction, β-galactosidase activity was determined using activation of the reporter LacZ gene (Fig. 4b). The quantitative assay revealed substantial differences in the interaction affinities for the different 14-3-3 isoforms (Fig. 4b). The 14-3-3 ξ (chi), λ (lambda), ω (omega), ψ (psi) and υ (upsilon) isoforms showed high affinity for both ABF2 and ABF4 in all replicates. To demonstrate that the above described C-terminal motif of the ABF transcription factors was the site of interaction with the 14-3-3 proteins, we introduced a point mutation in the ABF2 protein that replaced S³⁹³ with an alanine: ABF2^S393A (Fig. 4). The introduction of this point mutation resulted in the complete loss of interaction with 14-3-3s, thereby confirming that this C-terminal motif is the actual 14-3-3 binding site.

Figure 4. The 14–3-3 proteins of Thellungiella interact with ABF transcription factors in yeast two-hybrid assays. A - Interaction of 14–3-3s and ABFs, represented qualitatively. TsABF2^S393A, mutant form of the ABF2 protein. Full-grown yeast cultures were spotted on selective media (SD –LWHA for ABF2 and SD-LWH for ABF4) and were grown for an additional 3–5 d. B - Interactions are shown quantitatively by measuring β-galactosidase activity (n = 3; ± SE).

Discussion

High salinity or drought results in increases in plant ABA levels, which in turn affect the expression of a vast number of stress-related genes. In the salt- and ABA-mediated regulation of gene expression, ABF transcription factors play a pivotal role.²³ In our research, we identified and characterized four members of the ABF family in Thellungiella salsuginea, an extremophile plant that has emerged as a new model species that can be easily compared with Arabidopsis thaliana; thus, T. salsuginea is an Arabidopsis Related Model Species (ARMS).^24-27 In addition to the bZIP domain, the identified proteins contain four highly conserved regions that are almost invariable between these two species. The ABF transcription factors of Thellungiella share up to 88% identity in the amino acid sequence with their Arabidopsis counterparts. The differences in amino acid sequences between the Thellungiella and Arabidopsis ABFs were observed mainly in interdomain regions. The highest diversity was observed for TsABF1, whose Arabidopsis homolog was reported to be induced by ABA and cold but not by salt and drought.¹⁰ Although the identity between TsABF1 and AtABF1 was not high, the expression pattern of TsABF1 (data not shown) was similar to that observed for AtABF1, which is salt-insensitive.¹⁰ ABF2, ABF3 and ABF4 were characterized as the master transcription factors that regulate the salt and ABA-induced ABRE-dependent gene expression based on the observed phenotype of the abf2 abf3 abf4 triple mutant plants.^23,28

The quantitative RT-PCR analysis revealed some clear differences in the salt-induced TsABF expression patterns in the root and shoot tissues. Of the ABF genes, ABF2 had the strongest response in both the shoots and roots. In addition to the evidence in favor of ABF redundancy, data concerning the different expression patterns of ABFs as reported here and in a number of other studies indicate that there are both similarities and specificities in their functions under various conditions.^10,28 In addition to the above-mentioned, the short-term upregulation of ABF gene expression in response to stress is typical for Thellungiella. This plant has a rapid response to stress, implicated in the expression rate changes of the important stress-related genes. In fact, the highest changes in the expression rates of the Arabidopsis ABF genes were commonly observed after 18–24 h of stress exposure.^11,29

In addition to regulation at the level of transcription, ABF transcription factors are also subject to post-translational regulation, notably phosphorylation and proteasome-mediated degradation.^15,16,30,31 Although members of the SNRK2 protein kinase family have been shown to phosphorylate and activate ABF transcription factors, the precise mechanism of activation remains unclear.^30,32 Interaction with members of the 14-3-3 protein family was shown to form another level of control over the activity of a number of ABF transcription factors.^17,33 Recently, it has become clear that ABA-induced phosphorylation, 14-3-3 interaction and the prevention of degradation of ABFs are related events that control the amount of available ABF protein and thereby the associated transcriptional activity.¹⁶

A yeast-two hybrid screen with ABF2 and ABF4 as prey and the eight 14-3-3s as bait shows that the 14-3-3 proteins interact in an isoform-specific manner. The diversity of the 14-3-3 isoforms interacting with ABF transcription factors may result in the formation of a variety of homo- or heterodimeric complexes with different partners, either ABFs or other components of the transcriptional machinery. From the other side, some specificity, observed in the yeast two-hybrid assay, and different binding affinities may serve as prerequisites to the formation of certain complexes and thus certain types of transcriptional regulation. The isoform specificity of the highly conserved members of the 14-3-3 protein family with respect to their interactions with specific target proteins may reside in the highly variable C-terminal end of the proteins.^34,35 Some of the Thellungiella 14-3-3 isoforms differ markedly at their C-terminus from their Arabidopsis counterparts (Fig. S4).

A possible reason for the 14-3-3 and ABF interaction is that 14-3-3 proteins may regulate the DNA-binding activity of transcription factors. Recently, Rajagopalan and co-authors demonstrated that the human p53 transcription factor forms tetramers more efficiently at lower concentrations when the dimers are bound together by 14-3-3 proteins.³⁶ Such an interaction was shown to increase the DNA-binding activity of p53. In the case of plant transcription factors, ABF proteins must form a dimer for effective DNA binding and transcriptional regulation of the downstream gene. Taking into account that 14-3-3 proteins naturally act as dimers, it becomes an attractive proposition that the interaction with 14-3-3 proteins may promote the formation of functionally active ABF dimers.

As the multiple 14-3-3 isoforms were found to be expressed in different organs and tissues of the plant,^37-40 the observed and as yet undiscovered interaction with as many isoforms ensures that the transcriptional function of the ABFs will be provided in every cell where it is needed.

In conclusion, our data indicate that while there were similarities with those of Arabidopsis, the Thellungiella ABFs were distinct, especially in terms of the level of regulation of gene expression. In addition, the observed specificity of interaction with different Thellungiella 14-3-3 isoforms, together with the difference in certain 14-3-3 isoforms from their Arabidopsis counterparts, may be a prerequisite for the formation of new transcriptional complexes and, thus, alterations in the regulation of the expression of stress-related genes.

Materials and Methods

Plant material and growth conditions

The experiments described in this work were performed using the plants of the ecotype Yakutsk, belonging to the species T. salsuginea. The plants were grown hydroponically in a growth chamber in black 900 ml pots with half-strength Hoagland solution. Four plants were grown in each pot at 22°C day time and 10°C night time temperatures under long day conditions with 14 h-light/10 h-dark cycles. The Hoagland solution was replaced weekly. When the plants reached the age of six weeks, we started the salt treatment. On the third day after the scheduled replacement of the Hoagland solution, the concentration of salt was raised until it reached 200 mM within one hour by four supplementations of 5 M NaCl (after vigorous stirring of the medium, plants were placed back in the pots). The short-term effect of salt stress on ABF expression was monitored. Pools of four plants were harvested at 0 (control group), 2, 8 and 24 h after the plants were exposed to stress. The roots and shoots were frozen separately in liquid nitrogen and stored at -70°C until needed.

RNA extraction and RACE-PCR

Total RNA was extracted from the pools of four plants using the TRIzol^® reagent (Invitrogen) according to the manufacturer’s protocol. The RNA samples were then treated with DNases to prevent contamination by genomic DNA. The extractions were repeated at least twice for each time point in the case of the salt treatment. Two micrograms of RNA was used for cDNA synthesis by SuperScript II reverse transcriptase in accordance with the manufacturer’s protocol (Invitrogen). To identify the full length RNA transcripts of the ABF genes, cDNA was synthesized using SMART first strand synthesis technology from Clontech.

Identification of Thellungiella ABF genes

Based on the homology of the ABF family genes among different species, we designed degenerate primers (see Table S1) that bordered the C3 conservative region and the bZIP domain using the Arabidopsis genes as a template. Four central fragments of the ABF coding regions were amplified using degenerate primers and sequenced. Full-length cDNA sequences were determined by the 5′- and 3′-RACE methods according to the Clontech SMART techniques. The primers that were used for the sequencing are listed in Table S1 of the supplementary materials. The phylogenetic tree was built using MUSCLE for multiple alignment, PhyML for tree building and TreeDyn for tree rendering.⁴¹

Quantitative RT-PCR

Real-time RT-PCR was performed with oligonucleotides that were specific for the 3′ non-translated regions of the different ABF genes. The primers were designed so that the size of the PCR product ranged between 110 and 150 bp. Each amplicon was cloned and sequenced to verify the specificity of the primers. The amplification of the Actin gene fragment was used as an internal control. The real-time RT-PCR was performed on the Chromo4™ system (BioRad) using gene-specific primers and a SYBR-Green PCR Master kit (Syntol, Russia), containing a Hot Start Taq polymerase. The amplification of the PCR products was monitored via the intercalation of SYBR-Green. The melting curves (from 50°C to 96°C) were analyzed to monitor the specificity of the PCR amplification. The gene expression levels were measured three times for each cDNA sample. At the same time, two or three independent cDNA samples were taken for each time point of treatment. The data obtained were analyzed with Opticon Monitor 3 software. The calculations were performed using the REST 2005 program.⁴² The Delta-delta-Ct method was used to quantify the fold change in ABF gene expression during stress exposure with respect to expression levels under the control (no stress) conditions. The primers that were used in the real-time RT-PCR are presented in Table S1.

Yeast two-hybrid assay

The improved yeast strain PJ69–4A, which contains an extra reporter gene selecting for adenine auxotrophy (ADE2), was used in the assay.⁴³ The pGADT7 AD vector (Clontech) and the pBD-GAL4 Cam (Stratagene) vector were used. The coding regions of the eight 14–3-3 isoforms from Thellungiella were cloned into the pBD-GAL4 Cam vector using the appropriate primers and restriction enzymes, listed in Table S1. The coding regions of the ABF2 and ABF4 genes were cloned into the pGADT7 AD vector. A point mutation was introduced into the putative 14–3-3 binding site of the ABF2 gene by PCR using the reverse primer with the S³⁹³ substitution so that the serine residue was replaced by an alanine. The primers and the enzymes used for preparing the constructs are shown in Table S1. The transformation of the yeast cells was performed by the lithium acetate method.⁴⁴ Briefly, the yeast cells were transformed with different combinations of the AD-BD constructs and grown up on the selective SD medium, which lacked the amino acids leucine and tryptophan (-LW). The double-transformed cells were additionally grown up in 1 ml liquid SD medium (-LW). Then, 7 µl aliquots of the yeast culture were spotted on selective SD plates lacking the amino acids leucine, tryptophan, histidine and adenine (-LWHA) and grown for 4–5 d at 30°C (at least three individual replicates). For the quantitative β-galactosidase assays, 1 ml aliquots of yeast culture were spun down and used for activity assays. To confirm whether the clones contained the correct DNA constructs, DNA was isolated from 3 ml samples of the full-grown cultures. PCR was performed using vector- and gene-specific primers. In the quantitative assay, the release of o-nitrophenol from the substrate ONPG was recorded at 420 nm and normalized to the cell density measured photometrically at 600 nm; the β-galactosidase activity was presented in Miller units. The detailed β-galactosidase activity protocol is described at http://labs.fhcrc.org/gottschling/Yeast%20Protocols/Bgal.html.