Introduction

Plants are continuously exposed to various abiotic and biotic stresses in natural environment, and have evolved many adaptive mechanisms, which are dependent on a variety of signaling pathways, in order to react to altered conditions and combat pathogen attacks for survival. These stress-related signaling pathways are individually regulated by phytohormones including abscisic acid (ABA), salicylic acid (SA), jasmonic acid (JA), and ethylene (ET) 1, 2, 3, 4, and can result in various physiological responses such as osmotic adjustment, stomatal movement, and oxygen bursts. The cellular processes of plant-stress responses are further complicated by web-type crosstalks among different signaling pathways. The identification and characterization of important signaling components have provided insights into these specific pathways; however, gaps still exist in our understanding of plant-stress response mechanisms. Transcriptome analyses have generated considerable data, which show extensive overlapping on gene expression between biotic and abiotic stresses. Many of these overlapped genes encode signaling components, including transcription factors and protein kinases 5, 6, 7, 8, 9.

Transcription factors are promising candidates of common players involved in crosstalk between abiotic and biotic stress signaling 4. For instance, AtMYC2/JIN1, encoding the basic helix-loop-helix transcription factor, is not only involved in the ABA-mediated drought stress-signaling pathway, but also involved in JA-mediated wound response and JA/ET-mediated pathogen defense responses. Overexpression of AtMYC2 resulted in improved osmotic stress tolerance. AtMYC2 deficiency resulted in elevated expression of JA/ET-regulated defense genes and increased plant resistance to the necrotrophic fungal pathogen Fusarium oxysporum 10, 11, 12. Additionally, BOS1, an R2R3MYB transcription factor, was shown to be required for biotic and abiotic stress responses in Arabidopsis. bos1 mutant plants showed increased susceptibility to necrotrophic pathogens, and impaired tolerance to drought, salinity, and oxidative stress 13.

Many transcription factors involved in signaling crosstalk are multi-functional proteins. As a large multi-functional gene family, the Arabidopsis NAC family contains more than one hundred members. NAC proteins are plant-specific transcription factors characterized by a conserved N-terminal region known as the NAC domain. NAC proteins have crucial functions in plant developmental processes. Several NAC members were found to be determinative regulators in diverse plant developmental programs, including formation of apical shoots 14, lateral roots, and floral organs 15, 16, 17; and leaf senescence 18. Additionally, NAC proteins have been implicated in disease defense and abiotic stress signaling. Several NAC proteins have been shown to interact with virus proteins 19, 20, 21, or to play direct roles in regulating pathogen defense 22. Some NAC proteins were recently identified as abiotic stress regulators. The NAC cDNAs ANAC019, ANAC055, and ANAC072/RD26 were isolated as promoter-binding proteins of ERD1 (early-responsive to dehydration 1). Their expression was induced by drought, high salinity, ABA, and methyl jasmonate (MeJA). Plants individually overexpressing these genes showed a significant increase in drought tolerance or salt tolerance 23, 24, 25, 26. The anac019anac055 double mutant showed increased resistance to B. cinerea compared with the wild-type plants, and transgenic plants overexpressing ANAC019 or ANAC055 showed decreased resistance to this pathogen 27. Moreover, NAC transcription factors such as ATAF2 and OsNAC6 have been shown to be involved in both biotic and abiotic stress responses. Overexpression of the ATAF2 gene, which is a wound-, SA-, JA-, and salinity-responsive gene, resulted in increased susceptibility to fungal pathogens and decreased expression of pathogenesis-related genes 28. Additionally, the rice gene OsNAC6 was found to be an abiotic and biotic stress-related gene. OsNAC6-overexpression plants were tolerant to dehydration/high-salinity stresses and showed slightly increased tolerance to blast disease 29, 30.

ATAF1 (At1g01720), encoding a NAC protein, was previously reported to be a wounding-inducible gene 19. The ataf1-1 mutant line showed decreased resistance to a biotrophic fungus as a consequence of increased ABA level 22, 31. ATAF1 was also reported to be a negative regulator in drought response based on analysis of ataf1 mutants 32. Here, we show that ATAF1 plays important roles in both abiotic and biotic stress responses. We found that ATAF1 overexpression lines, but not its null mutants, showed enhanced plant drought tolerance, revealing a positive function in plant drought response. Additionally, ATAF1 overexpression lines were hypersensitive to high salinity, ABA, oxidative stress, and necrotrophic-pathogen infection. Our data indicate that ATAF1 mediates Arabidopsis response to both abiotic and biotic stresses.

Results

Arabidopsis ATAF1 is a potential stress-related NAC gene

In order to systematically survey the function of Arabidopsis NAC members in stress signaling, we conducted an in silico analysis by mining public microarray data (Arabidopsis Biological Resource Center) for all NAC family genes under different stress treatments. As summarized in Supplementary information, Table S1, members of NAC family genes were found to be induced by multiple stresses, suggesting functional redundancy as well as specificities. We are in the process of applying forward and reverse genetic approaches to functionally study these stress-related NAC members. The following experimental results are based on our detailed investigation of one stress-related NAC representative, ATAF1 (At1g01720, ANAC002).

Upregulation of ATAF1 in response to multiple abiotic stresses and Botrytis cinerea infection

To confirm the microarray data, detailed expression patterns of ATAF1 were analyzed by RNA gel blot using an ATAF1-specific probe. ATAF1 transcripts rapidly accumulated upon 250 mM NaCl, dehydration, or 50 μM ABA, and were maintained at a high level during the entire treatment period. Furthermore, ATAF1 was found to be upregulated by mechanical wounding and 50 μM MeJA. In addition to induction of ATAF1 by these abiotic stimuli or stress-related phytohormones, accumulation of ATAF1 transcripts was also observed in seedlings infected with B. cinerea (Botrytis cinerea) (Figure 1). ABA is generally considered to be an important biomolecule regulating plant responses to drought and high salinity, and in particular it plays an essential role in some ABA-dependent stress response processes. To determine whether ATAF1 accumulation during drought and salinity stresses was dependent on ABA, aba2 (an Arabidopsis ABA-deficient mutant) was used for analyzing ATAF1 gene expression. The RNA gel blot revealed that ATAF1 was still evidently induced in aba2 plants subjected to drought or NaCl treatment (Supplementary information, Figure S1). This finding indicates that induction of ATAF1 in response to drought or high salinity can be achieved through an ABA-independent pathway, although ABA alone is able to induce ATAF1 expression.

Figure 1
figure 1

Northern blot analysis of ATAF1 transcript. Expression of ATAF1 in response to high salinity (250 mM NaCl), drought, ABA (50 μM), wound, MeJA (50 μM) treatments, and B. cinerea infection. A: ATAF1 probe. R: rRNA: methylene blue-stained ribosome RNA used as loading controls.

Full size image

Genetic analysis of ATAF1 gene

To investigate the functions of ATAF1 in vivo, we generated transgenic Arabidopsis plants overexpressing this gene. RNA gel blot revealed that all of the transgenic lines expressed ATAF1 at higher levels compared with wild-type plants (data not shown). The highest level of ATAF1 expression was found in three transgenic lines, designated as ATAF1-O1, ATAF1-O2, and ATAF1-O3. In western blot assays, these three lines exhibited high expression of the Myc-ATAF1 protein (Figure 2A). In addition, two T-DNA insertion lines with different target sites in ATAF1 gene exons were obtained from the Salk mutant collection: Salk-067648 (ataf1-1) and Salk-057618 (ataf1-2). RT-PCR results showed that the expression of ATAF1 was completely abolished in the ataf1-1 and ataf1-2 mutant lines, but was higher in the overexpressing lines (Figure 2B).

Figure 2
figure 2

Genetic analysis of ATAF1. (A) ATAF1 transcript and protein levels in ATAF1-overexpression lines. Western blot was performed with anti-Myc monoclonal antibody. (B) Comparison of ATAF1 transcript levels in wild-type, ataf1-1, ataf1-2, overexpression plants ATAF1-O1 and ATAF1-O2. (C) Growth phenotype of 35-day-old ATAF1-overexpression plants in comparison with ataf1 mutants and wild-type plants. WT: wild-type plants. Bar = 1 cm.

Full size image

The ATAF1-transgenic lines showed altered phenotypes, such as relative dwarfism with small size, light green and round-shaped leaves, short petioles, sterility at the early flowering stage, and short siliques. Although ataf1 mutants exhibited long primary roots (data not shown), their phenotypes bear a close resemblance to those of wild-type plants (Figure 2C).

Overexpression of ATAF1 confers drought tolerance in Arabidopsis

Since ATAF1 is a drought-inducible gene, it likely regulates drought signaling. To characterize such a function of ATAF1, a whole-plant drought assay was performed in soil. Seedlings (10-day old) were transferred from MS plates to water-saturated soil for 7 days, and then water was withheld to cause a severe water deficit in soil. After 21 days, most of the control and mutant plants were wilted due to the extreme water deprivation (Figure 3A). After rewatering, 94% of overexpression plants continued to grow and successfully produced seeds, whereas most of the other plants did not recover and died from drought, showing a lower survival ratio (32% for wild-type, 28% for ataf1-1, and 30% for ataf1-2) (Figure 3A). This experiment was repeated several times with similar results. Our results indicate that high levels of ATAF1 can enhance plant drought tolerance.

Figure 3
figure 3

Drought-tolerance assays of wild-type, ataf1-1, ataf1-2, ATAF1-O1, and ATAF1-O2 plants. (A) Photographs were taken when the drought-stressed phenotypes appeared and 6 h after being rewatered. Bar = 1 cm. (B) Comparison of phenotypes of excised leaves under the air-dry treatment for 0 h, 6 h, and 10 h. Bar = 1 cm. (C) Leaf water loss assay. Leaf weight was measured at the indicated time point (n = 18). Curves were drawn based on the data from three independent experiments, and bars indicate standard errors.

Full size image

Regulation of transpiration plays a vital role in plant response to drought stress. To address whether the ATAF1 function in enhancing plant drought tolerance was associated with transpiration, we compared the wilting phenotype of detached leaves from different plants that were subjected to air-dry treatment at room temperature with 65% humidity. After 6 h, obvious phenotypic changes were observed in these leaves (Figure 3B). At this time point, only the leaves from ATAF1-O1 and ATAF1-O2 plants retained normal shape and they only showed a slightly wilting phenotype, whereas the leaves of other plants were severely wilted. Similar phenomena were observed at 10 h of the treatment (Figure 3B). We further addressed the reasons for these phenotypic alterations by measuring water loss ratios in detached leaves from these plants (Figure 3C). In this assay, we found that the leaf weight of ATAF1-O1 and ATAF1-O2 plants decreased at the slowest rate of all analyzed plants, whereas other plants showed comparable water loss rates. Therefore, this quantitative result is consistent with the phenotypic observations in the leaf-wilting assay.

To reveal how ATAF1 controls drought-responsive genes, we examined the expression of a group of marker genes in the drought-stressed plants (Supplementary information, Figure S2). ATAF1-O1 plants produced dramatically high basal levels of ADH1 (encoding an alcohol dehydrogenase) transcripts (0 h) and further increased its expression to reach the highest level compared with other plants. The expression patterns of RD29A and COR47 were very similar. They showed low and comparable basal expression among different plants at 0 h, and then divergently increased in the assayed drought-stressed plants. After 2 h of treatment, transcripts of these two markers accumulated to a higher level in ataf1-1 plants compared with wild-type or overexpression lines. However, transcript levels decreased in ataf1-1 plants thereafter (10 vs 2 h). In contrast, ATAF1-O1 plants induced small amounts of transcripts of the two marker genes at the 2-h time point, but accumulated the highest level of their transcripts at late stages of stress (10 h) in comparison with the other assayed plants. The drought-induction of RD22 was markedly repressed in ATAF1-O1 plants, but slightly enhanced in the mutant plants. These results indicate that ATAF1 may control different downstream ABA- and stress-responsive genes through specific manners that are different from other transcription factors.

Altered ATAF1 expression affects plant sensitivity to ABA

The transcript level of ATAF1 was dramatically induced by ABA in RNA gel blot assay (Figure 1), suggesting the potential function of ATAF1 in ABA response. To elucidate the role of ATAF1 in ABA signaling, ATAF1-O1, ataf1-1 mutant, and control plants were germinated on MS plates with different amounts of ABA (0 , 0.25, 0.5, and 1 μM). No difference in the growth efficiencies (green cotyledons) and germination rates was observed among the untreated plants (Figure 4A). However, in the presence of 0.25 μM ABA for 3 days (the lowest concentration tested), the growth efficiencies decreased to 51.3% for ATAF1-O1, 78.5% for wild-type plants, 72.8% for vector plants, and 87.2% for ataf1-1 plants. Similar inhibition effects were also observed for these plants grown on 0.5 μM ABA plates. ATAF1-O1 plants exhibited the most severe growth inhibition by ABA among the assayed plants, indicating that high levels of ATAF1 transcripts rendered plants more sensitive to ABA. The highest concentration of ABA (1 μM) used in this study led to extreme growth inhibition for all of the plants (Figure 4A); however, only ATAF1-O1 plants displayed the most severely arrested growth under this condition, as reflected by retardation of cotyledon greening (Figure 4B). On the contrary, the germination rate of ataf1-1 plants was the lowest and the germination rate of ATAF1-O1 was the highest compared with that of the control plants (data not shown). To further correlate ATAF1 function with plant sensitivity to ABA during the post-germination stage, root elongation inhibition was analyzed for these plants. Seedlings (4-day old) were transferred to MS medium with 10 μM ABA. After 7 days, root growth inhibition by ABA was the least severe for the ataf1-1 mutant, but the most severe for ATAF1-O1 plants among the assayed plants (Figures 4C, 4D). Taken together, these results indicate that ATAF1 modulates plant ABA signaling and high ATAF1 expression contributes to ABA hypersensitivity in Arabidopsis.

Figure 4
figure 4

Sensitivity of ATAF1 overexpression lines to ABA. (A) Effects of ABA on growth efficiencies. Growth efficiencies (green cotyledons) represented by the chart were based on three repeated experiments, in each of which an average of 90 seeds were counted. (B) Seedlings were grown on MS medium with or without 1 μM ABA and their phenotypes were scored on the 12th day after sowing. (C) Phenotypic comparison of root length in the presence of ABA. (D) Quantitative analysis of root growth. The experiment was performed in triplicate (n = 20) and repeated thrice. Bars indicate standard errors. (E) Effects of ABA on stomatal aperture of plants. Epidermal peels from plants were kept for 12 h in the dark, incubated under light in buffer for 3 h, and then treated with 0 and 50 μM ABA for 2 h before aperture measurements. Data are mean ratios of width to length ± SE of three independent experiments (n = 30 to 40). WT: wild type, Vec: vector control plants. (F) Stomatal aperture ratio, data are ratios of stomatal aperture of plants treated with 50 μM ABA to that of untreated plants.

Full size image

Stomata numbers on the leaf and aperture of stomata are major factors to determinate water evaporation. Under microscope, stomata numbers in identical area were counted for ATAF1-O1, ataf1-1, and control plants, and no stomata number changes were observed (data not shown). Overexpression of some ABA signaling components, such as SDIR1, ABF3, and ABF4, can reduce leaf water loss, because these genes have regulatory functions in controlling stomatal movement 33, 34. To determine whether ATAF1 can act on the adjustment of stomatal openings in a similar manner, we treated leaves of four genotypes with ABA to analyze stomatal aperture. Treating the leaf epidermis of ATAF1-O1 plants with ABA caused nearly complete closure of stomata; the effect was not so pronounced in the control and ataf1-1 plants (Figure 4E). Indeed, treating the leaf epidermis of ATAF1-O1 plants with ABA caused the most severe closure of stomata (57.9% of untreated stomata aperture) compared with other plants (63.4% for wild-type plants, 64.2% for Vec control plants, and 69.5% for ataf1-1 plants) (Figure 4F). Obviously, ATAF1 overexpression elevated the sensitivity of stomatal closure mediated by ABA, whereas ataf1-1 mutants showed slightly opposite phenotypes. These results indicate that ATAF1 is involved in ABA-dependent stomatal closure.

To further substantiate this finding, the expression of several ABA/stress-responsive marker genes was analyzed in ATAF1-O1, ataf1-1, and wild-type plants (Supplementary information, Figure S3). For ADH1 gene, wild-type and ataf1-1 plants showed similar expression patterns in response to ABA treatment. However, ADH1 was highly expressed in ATAF1-O1 plants at the 0 h time point and its transcripts were further increased by ABA treatment. Differently, RD22 showed higher induction by ABA in the ataf1-1 mutant than ATAF1-overexpression and wild-type plants (Supplementary information, Figure S3). The results suggest that ATAF1 is involved in plant response to ABA by differentially regulating ABA/stress-responsive genes.

Transgenic plants overexpressing ATAF1 are hypersensitive to high salinity

ATAF1 was found to be induced by salt; therefore, we asked whether this induction was correlated with plant-salinity response. As shown in Figure 5A, the germination rates (radicle emergence) of ATAF1-O1 and ATAF1-O2 were dramatically influenced by 100 mM NaCl compared with those of other plants. There was no difference of germination rates among all untreated plants (data not shown). Moreover, the greatest growth retardation and lesion were also found for ATAF1-O1 and ATAF1-O2 in the presence of 100 mM NaCl (Figure 5B). The percent of green seedlings was 32% and 35% for ATAF1-O1 and ATAF1-O2, and nearly 90% for other plants (Figure 5B). Similar phenomena were also observed under a lower NaCl concentration (50 mM) (data not shown). These experiments indicate that high level of ATAF1 transcripts sensitizes plants to high salinity.

Figure 5
figure 5

High-salinity assay of plants. (A) Germination rates (radical emergence) of wild type, ataf1-1, ataf1-2, ATAF1-O1, and ATAF1-O2 plants on 1/2 MS medium containing 100 mM NaCl for 5 days. Percentages are means (n = 60 to 90 each) of three repeats ±SE. (B) Growth phenotype of transgenic and mutant plants on 1/2 MS medium (top panel) or containing 100 mM NaCl (bottom panel). Seeds were germinated and grown for 7 days. Numbers below the panel indicate the percentage of seedlings with green cotyledons in total seedlings grown on 100 mM NaCl medium (n = 60 to 90). Bar = 1 cm.

Full size image

To further address the regulatory function of ATAF1 in salt response, expression patterns were checked for stress-responsive marker genes under NaCl treatment. Similar to ABA, NaCl induction led to high ADH1 expression in ATAF1-O1 plants, but did not notably change its expression in the ataf1-1 mutant. Additionally, three downstream stress-signaling genes, RD22, RD29A, and COR47, showed similar expression patterns. Their expression, which was unaffected in the ataf1-1 mutant, was reduced in ATAF1-O1 plants at the early stage of salt stress (2 h). Interestingly, the transcripts of these three genes were slightly increased in ATAF1-O1 plants at the late treatment stage (10 h). These results may suggest a feedback regulation of those genes affected by ATAF1 under salt stress conditions (Supplementary information, Figure S4).

Altered ATAF1 expression levels affect plant susceptibility to the fungal pathogen B. cinerea

ATAF1 has been shown to be induced by JA, which is an important defense-related phytohormone implicated in the regulation of plant resistance to pathogens. In order to determine the function of ATAF1 in biotic stress reactions, we checked the disease susceptibility of the different groups of plants inoculated with B. cinerea, a non-specific necrotrophic pathogen. Leaves from 4-week-old ATAF1-O1 and ATAF1-O2 plants, ataf1-1 and ataf1-2 mutants, and wild-type plants were drop-inoculated with the B. cinerea spore suspension, and disease progression was observed 7 days after inoculation. The examined disease symptoms included chlorosis, necrosis expansion, and tissue maceration surrounding the primary infection sites. As shown in Figure 6A, ATAF1-O plants were more sensitive to B. cinerea infection in detached leaves than in ataf1 mutants and wild-type plants, in which B.cinerea infection resulted in milder disease symptoms. Furthermore, we found that the leaf lesion size of ATAF1-O plants was larger than that of other plants (by 9- to 16-fold) (Figure 6B).

Figure 6
figure 6

Responses of plants to B. cinerea infection. (A) Disease symptoms of detached leaves 5 days after inoculation with a 3-μl droplet of B. cinerea spores (105 spores/ml). (B) Lesion size was measured 5 days after inoculation. Date points represent average lesion size of measurements from 30 lesions. The experiments were repeated thrice. (C) Expression of PR-1 and PDF1.2 in response to B. cinerea infection. Each lane contains 15 μg total RNA.

Full size image

To define whether the role of ATAF1 in biotic responses of plants was restricted in their reaction to a necrotrophic pathogen, we also tested its function by using the hemi-biotrophic pathogen Pseudomonas syringae. Both ataf1-1 mutant and ATAF1-O1 plants showed comparable disease symptoms in response to the avirulent bacterial strain (P. syringae pv tomato avirulent strain DC3000 carrying avrRpm1) and the virulent strain DC3000 in comparison with wild-type plants (Supplementary information, Figure S5).

We examined the expression of the defense-response markers PDF1.2 (linked to JA and ET signalings), and PR1 (linked to systemic acquired resistance (SAR)). As shown in Figure 6C, PDF1.2 expression pattern was unaffected in the ataf1-1 mutant and ATAF1-O1 plants in response to B. cinerea, suggesting that the ATAF1 gene functions via a resistance/response pathway that may be independent of PDF1.2. However, ATAF1-O1 plants induced PR1 expression more strongly than other plants, and the lowest expression of PR1 was found in ataf1-1 mutants (Figure 6C), revealing a negative correlation between ATAF1 and plant disease resistance.

Transgenic plants overexpressing ATAF1 are hypersensitive to oxidative stress

Reactive oxygen intermediates (ROIs) have been implicated in signaling in response to both pathogens and abiotic stresses. In plant abiotic stresses, ROIs likely act as signals to induce reactive oxygen scavengers and other protective mechanisms, and as damaging agents that may contribute to stress injury in plants. Additionally, ROIs play important roles in plant-pathogen interactions such as cell wall strengthening and rapid host cell death. To test whether ATAF1 was involved in ROI signaling, the sensitivity of plants to oxidative stress was assayed using rose bengal that generates ROIs in medium when exposed to light, and paraquat (methyl viologen) that increases the ROI formation by inhibiting electron transport in the reduction of NADP to NADPH during photosynthesis 13. In the presence of 6 μM rose bengal, a complete block was found in the growth of ATAF1-O1 and ATAF1-O2 seedlings that were etiolated without any expanded green leaves (Figure 7). However, ataf1-1, ataf1-2, and control plants were minimally affected by this treatment, as these plants developed normally. Likewise, ATAF1-O1 and ATAF1-O2 seedlings showed more sensitivity than other seedlings when germinated on the 1/2 MS medium supplemented with 1 μM paraquat (Figure 7). These results suggest that ATAF1-mediated stress responses are likely related to ROI signaling pathways.

Figure 7
figure 7

ATAF1-overexpressing plants are hypersensitive to oxidative stresses. Seedlings (5-day old) germinated on 1/2 MS medium were transferred to 1/2 MS medium with or without 6 μM rose bengal. Photographs were taken 5 days after treatment (middle panel). Young seedlings were compared between wild-type, ataf1-1, ataf1-2, ATAF1-O1, and ATAF1-O2 plants in the presence of 1 μM paraquat (lower panel). The pictures were taken after 3 weeks. Bar = 1 cm.

Full size image

Discussion

ATAF1 functions in ABA signaling and abiotic stress responses

ABA is an essential mediator in triggering plant responses to most of the common abiotic stresses, including drought, salinity, high temperature, oxidative stress, and cold 35, 36. Nearly all ABA-deficient (aba) and ABA-insensitive (abi) mutants exhibit salt insensitivity during germination and are susceptible to drought because of the impaired stomatal aperture regulation (e.g., abi1 and abi2), while plants overexpressing the ABA-response pathway genes, such as SDIR1, ABF3, and ABF4, are tolerant to drought and hypersensitive to salt 33, 34, 37, 38. ATAF1 overexpression plants showed remarkably enhanced tolerance to drought but were hypersensitive to salt stress.

Stomatal closure is a key ABA-controlled process in dealing with water deficit conditions. Our results indicate that ATAF1 is involved in stomatal regulation. Its overexpression resulted in lower transpiration and enhanced drought tolerance. The stomatal openings of ATAF1-O1 transgenic plants were smaller than those of wild-type plants under normal light growth conditions in soil. Furthermore, stomatal closure in the epidermis of ATAF1-O1 plants was more sensitive than that of wild-type and ataf1-1 plants after ABA treatment. ATAF1-O1 plants also exhibited the most severe growth inhibition by ABA. Thus, our results suggest that ATAF1 may play a positive role in ABA-mediated guard cell control and regulation of drought response.

Recently, ATAF1 was described as a negative regulator in plant drought tolerance by phenotypically comparing wild-type plants and ataf1 mutants 32. However, here we found the opposite phenotypes of drought response. We have carefully read the previous paper and compared our experiments. We used the same Salk mutants to validate the documented function of ATAF1. Very surprisingly, we could not observe the same phenotype for these mutant lines as what was previously reported. In fact, we repeatedly observed that ATAF1 overexpression lines, but not mutant plants, were more tolerant to soil-drought stress than controls. Such findings were also supported by other physiological experiments, including leaf water loss measurement and stomatal closure assay. Moreover, we found more rapid induction of some marker genes in the mutant plants compared with other assayed plants under drought stress, but the induction was dramatically diminished at late stage of the treatment. In contrast, ATAF1-O1 plants had significant transcript accumulations of some stress-responsive markers at the late time point, demonstrating a positive correlation with the enhanced drought tolerance of these plants. Considering that the plant response to drought stress spans a relatively long time, high accumulation of stress-responsive genes at late stage of the stress may be a major contributor to plant drought tolerance.

ATAF1 functions in biotic stress

In this study, ATAF1 enhanced susceptibility to infection by B. cinerea, but did not influence the plant response to P. syringae pv tomato. SAR is a mechanism of induced defense that confers long-lasting protection against a broad spectrum of microorganisms. Pathogen-triggered SAR process is dependent on the phytohormone SA. In contrast, plant pathogenesis caused by necrotrophic pathogens is closely related to JA and ET defense pathways 39. ATAF1 was found to rapidly respond to induction by MeJA, but not by SA (data not shown). To date, molecular and cellular mechanisms in plant-necrotrophic pathogen interactions remain largely elusive. However, ROIs are considered to be important molecules involved in plant disease reaction to necrophrophic pathogens such as B. cinerea. Elevated ROI levels can promote hypersensitive cell death. In one previous study, ROIs were reported to be induced by B. cinerea to accumulate to a high level, which may be toxic to host plants and consequently facilitate infection 40. Conversely, reciprocal experiments indicated that chemical-mediated blockade of ROI generation decreased cell death in different plants challenged by pathogens 41, 42 and restricted necrosis development in B. cinerea-infected Arabidopsis 43. The disease assay in our study showed that ATAF1-O plants displayed enhanced lesion formation after B. cinerea infection, indicating the negative role of ATAF1 in plant defense to this necrotrophic fungus. Alternatively, ATAF1 overexpression rendered plants hypersensitive to both the rose bengal and paraquat compounds, which promote the generation of reactive oxygen species. Considering the important roles of ROIs in necrotrophic disease processes, these findings suggest that ATAF1 may sensitize the reaction of plant cells to ROIs and accordingly contribute to the enhancement of plant susceptibility after B. cinerea infection. In our study, more induction of PR1 transcripts was found in ATAF1-O1 plants infected with B. cinerea, implying that there was a B. cinerea-triggered hypersensitive response that was thought to positively correlate with B. cinerea growth 43. In addition to pathogen challenge, abiotic stresses such as high salinity can elevate ROI levels in plants. The salt-hypersensitivity of ATAF1-O plants may also be a consequence of their enhanced sensitivity to ROIs.

Functional overlap between ATAF1 and other stress-regulatory NAC genes

In this study, mutant lines did not show distinct morphological variations as well as stress-related phenotypic changes when compared with control plants. Therefore, we speculate that there is likely functional redundancy among the NAC genes. Many Arabidopsis NAC members have been predicted to be stress-signaling related (Table S1). Recent studies have demonstrated that several NAC genes were involved in plant abiotic stress responses, including RD26 and its homologs ANAC019 and ANAC055. These NAC genes are responsive to drought and high salinity, as well as plant hormones ABA and JA, with varied expression patterns. Overexpression of these NAC genes in plants increased the survival rate for plants grown in water-deficient soil 25. In this study, ATAF1 overexpression resulted in many stress-related physiological effects in plants. These effects were similar to the observations described for RD26, ANAC019, and ANAC055. ATAF1 functionally resembles these characterized NAC genes, especially RD26. For example, both ATAF1 and RD26 were induced by exogenous ABA and by drought or high salinity in an ABA-independent manner, and their overexpression transgenic plants showed similar hypersensitivities to ABA. Obviously, functional overlap in stress responses exists among these NAC genes. The functional redundancy is likely responsible for the almost unchanged phenotype in the ataf1 mutant lines.

In spite of the extensive functional similarities between ATAF1 and other NAC homologs, the function of ATAF1 is not completely identical to that of other NAC genes, including RD26. For example, transcript accumulation was observed for RD26 in response to cold stress 44, but this was not found with ATAF1 (data not shown). Recently, OsNAC6, a rice ortholog of ATAF1, was reported to be a positive contributor to disease resistance and salinity tolerance 29, 30. This finding may indicate the functional discrepancies of NAC genes between monocotyledon and dicotyledon plants. In the biotic stress assay, we showed that ATAF1 played a negative role in plant response to B. cinerea, resembling the function of ATAF2 in plant pathogenesis. ATAF2, the closest Arabidopsis homolog of ATAF1, has been described as a repressor, which downregulated many pathogenesis-related genes, and consequently enhanced plant susceptibility to Fusarium oxysporum (a soil-borne fungal pathogen) in ATAF2 overexpression lines. Both ATAF1 and ATAF2 can be induced by the wounding/pathogenesis-related phytohormone MeJA. Moreover, we repeatedly isolated ATAF2 as a protein partner of ATAF1 in yeast two-hybrid experiments (unpublished data). These findings suggest that these two proteins are functionally associated and redundant. Nevertheless, it seems that both NAC genes also maintain their unique features. For example, unlike ATAF1, ATAF2 cannot be induced by ABA 28. Taken together, many other NAC homologs could contribute to the plant abiotic and biotic stress responses. Plants have evolved multiple functionally related NAC proteins that may be beneficial for their survival, given the fact that they are frequently challenged by extreme circumstances.

ATAF1 is an important gene with diverse biological functions

Based on our experimental data, ATAF1-O transgenic lines showed many evident growth phenotypic changes such as delayed flowering time and sterility at early flowering stage. As a regulator in ABA signaling, ATAF1 can activate ABA-responsive gene expression and result in ABA sensitivity. Similarly, other ABA signaling-related transcription factors such as ABF4 and ABI3 were also described to produce the phenotype of delayed flowering time in their overexpression lines 34, 45. Thus, we hypothesize that the ABA signaling pathway transduced by ATAF1 is likely involved in the flowering process. However, the intrinsic molecular link between ATAF1 and flowering control requires further elucidation.

In conclusion, ATAF1 appears to be a multiple-function transcription factor not only participating in abiotic and biotic stress responses, but also potentially regulating plant developmental processes.

Materials and Methods

Plant material and growth conditions

Arabidopsis accession Columbia (Col-0) was used for ATAF1 cDNA cloning, gene transformation, and phenotypic control in this study. Salk067648 and Salk057618 (ATAF1 T-DNA insertion mutants), as well as aba2, were obtained from the Arabidopsis Biological Resource Center. Plant growth conditions were as described previously 17.

Gene expression analysis

Seedlings (3-week old) were used for the following treatments. Dehydration treatment was carried out at room temperature with 75% humility for seedlings harvested from MS plates and placed on filter paper. Other treatments were carried out for seedlings grown on MS with 250 mM NaCl, 50 μM ABA, and 50 μM MeJA. For gene expression analysis, total RNA was isolated from nitrogen-frozen seedlings using the hot-phenol method 21. Totally, 15 μg RNA was applied in each lane in RNA gel analysis. Hybridizations were performed with the α-32p-labeled ATAF1-specific probe derived from its C-terminal region.

Constructs for plant transformation

To create the transgenic construct for overexpressing ATAF1 in Arabidopsis, the cDNA was amplified by RT-PCR with the primers (5′-GGATCC TAA TGT CAG AAT TAT TAC AGT TG-3′ and 5′-ACTAGT CTA GTA AGG CTT CTG CAT G-3′), and cloned into the site of BamHI and SpeI in the vector pBA002Myc. In this construct, N-terminal ATAF1 was fused in frame to the C-terminal 6× Myc tag and the tagged protein was driven by the CaMV 35S promoter. The transformants were selected by screening seedlings with resistance against Basta (8 μg/ml) on an MS plate. Myc-ATAF1 expression level was checked by RT-PCR reaction and western blot with an anti-Myc monoclonal antibody.

Abiotic stress assays

Seeds were sterilized and suspended in 0.15% agar and plated on MS medium (including vitamins, M0222.0050, Duchefa Biochemie. B.V., The Netherlands) supplemented with 1.5% sucrose, then stratified in darkness at 4 °C for 3 days and transferred to growth chambers with the same environmental conditions described above. To conduct germination rate assay, seeds were plated on MS medium without or with ABA (0 μM, 0.25 μM, 0.5 μM, and 1 μM), NaCl (50 mM, 100 mM, and 150 mM), or 1 μM paraquat. To conduct root growth assay, seeds sown on MS plates were stratified for 3 days at 4 °C and were vertically grown for 4 days under normal conditions, and then seedlings were transferred to vertical square MS plates with or without 10 μM ABA. For drought stress, 10-day-old seedlings were transferred from MS plates to water-saturated soil for 7 days, and then water was withheld until plants showed evident drought-stressed phenotypes. For leaf water loss assay, rosette leaves excised from plants grown in soil were put on filter paper for air-dry treatment. The leaf weight was measured at a series of time points. For rose bengal treatment, 5-day-old seedlings germinated on 1/2 MS medium were transferred to 1/2 medium with 6 μM rose bengal.

Effects of ABA on stomatal aperture of plants. Epidermal peels from 5-week-old plants were kept for 12 h in the dark, incubated under light in buffer (10 mM KCl, 7.5 mM potassium iminodiacetate, and 10 mM MES-KOH, pH 6.15) at 20 °C for 3 h, and then treated with 0 and 50 μM ABA for 2 h before aperture measurements (measured with a ZEISS microscope).

Fungal culture and plant inoculation

B. cinerea was grown on potato dextrose agar for 7-10 days at 24 °C with a 12-h photoperiod before spore collection. Spores were collected by flooding the dish and filtering the suspension with gauze. The conidia were resuspended with potato dextrose broth, and the concentration was adjusted to 105 spores/ml. Whole plants or rosette leaves were placed in dishes containing 0.8% agar and inoculations were performed with a 3 μl droplet of B. cinerea spores. After inoculation, plants were kept under a transparent cover to maintain high humidity and transferred to a growth chamber with 21 °C day and 18 °C night temperatures.

( Supplementary information is linked to the online version of the paper on the Cell Research website.)