Transcription factors have been shown to control the activity of multiple stress response genes in a coordinated manner and therefore represent attractive targets for application in molecular plant breeding. We investigated the possibility of modulating the transcriptional regulation of drought and cold responses in the agriculturally important species, wheat and barley, with a view to increase drought and frost tolerance. Transgenic wheat and barley plants were generated showing constitutive (double 35S) and drought-inducible (maize Rab17) expression of the TaDREB2 and TaDREB3 transcription factors isolated from wheat grain. Transgenic populations with constitutive over-expression showed slower growth, delayed flowering and lower grain yields relative to the nontransgenic controls. However, both the TaDREB2 and TaDREB3 transgenic plants showed improved survival under severe drought conditions relative to nontransgenic controls. There were two components to the drought tolerance: real (activation of drought-stress-inducible genes) and ‘seeming’ (consumption of less water as a result of smaller size and/or slower growth of transgenics compared to controls). The undesired changes in plant development associated with the ‘seeming’ component of tolerance could be alleviated by using a drought-inducible promoter. In addition to drought tolerance, both TaDREB2 and TaDREB3 transgenic plants with constitutive over-expression of the transgene showed a significant improvement in frost tolerance. The increased expression of TaDREB2 and TaDREB3 lead to elevated expression in the transgenics of 10 other CBF/DREB genes and a large number of stress responsive LEA/COR/DHN genes known to be responsible for the protection of cell from damage and desiccation under stress.
The Rab17 gene from maize belongs to a group of LEA/DHN genes. It is induced by abscisic acid (ABA) and water deficit (Vilardell et al., 1990). The maize Rab17 promoter has been tested in several heterologous systems. The promoter was tested in stably transformed tobacco and by transient expression in rice protoplasts and in both cases induction of the promoter by drought and ABA treatment was demonstrated (Vilardell et al., 1991). The activity of the 1.3-kb fragment of the Rab17 promoter fused to the GUS coding sequence was analysed in transgenic Arabidopsis plants and in ABA-deficient and ABA-insensitive mutants of Arabidopsis. Although the ZmRab17 promoter was active in the embryo and endosperm during late seed development, promoter activity decreased during seed germination and GUS activity was not enhanced by ABA and water deficit in transgenic plants. This suggests different molecular mechanisms for the Rab17 promoter activation in maize and Arabidopsis (Vilardell et al., 1994). Phylogenetic analysis of 5′-noncoding regions from the Rab16/17 gene family of sorghum, maize and rice revealed the absence of some important cis-elements in the promoters, which could explain some differences in the expression of Rab17-like genes in these plants (Buchanan et al., 2004). Although the activation and mechanisms of regulation of the ZmRab17 promoter were intensively studied, the application of this promoter for stress-inducible over-expression of genes in either maize or other crop plants has not been reported.
In this paper, we demonstrate that constitutive over-expression of two different wheat DREB factors leads to the substantial improvement of barley capacity to survive during severe drought and frost stresses. We show that this improvement is at least partially because of activation or suppression of the large number of other DREB/CBF genes and the consequent cascade of activation of downstream stress responsive genes, many of which are known to be directly involved in the protection of cells from damage caused by dehydration. We also show that the drought-stress-inducible ZmRab17 promoter, which has low or no basal expression in wheat in the absence of stress, is quickly and strongly activated in both wheat and barley by drought. Transgenic plants with stress-inducible over-expression showed little or no undesirable developmental traits such as stunted growth, dwarfism, delayed flowering and smaller spikes, traits which were observed in plants with constitutive over-expression of DREB factors. In contrast to wheat, in barley the Rab17 promoter is leaky and a pleiotropic phenotype can be observed in the well-watered transgenic plants although developmental setbacks are much less pronounced than in plants with strong constitutive overexpression of DREB genes.
Expression of TaDREB2 and TaDREB3 in well-watered plants and under different stresses
Full length cDNAs of TaDREB2 and TaDREB3 were isolated in a yeast one-hybrid screen from a library prepared from unstressed wheat grain using the DRE sequence from Arabidopsis (TACCGACT) as bait (Lopato et al., 2006). The phylogenetic relationship to other CBF/DREB factors from wheat and barley are shown in Figure 1.
TaDREB2 and TaDREB3 are both expressed in flower and grain tissues in the absence of stress. However, high levels of expression of TaDREB2 were also detected in roots, coleoptiles and embryo of germinating seed (Figure 2a). In the absence of abiotic stress, expression of TaDREB2 and TaDREB3 in leaf was very low while expression of both TaDREB genes was enhanced by drought. Drought stress induced TaDREB2 more strongly than TaDREB3 (Figure 2b). However, both genes were only weakly activated by cold (Figure 2c). Surprisingly, TaDREB2 was found to be strongly activated by wounding in grain with a lower level of activation by wounding in leaves (Figure 2d). No induction was seen in plants under salt stress and weak induction of TaDREB3 by ABA was only in leaf tissues (data not shown).
Full-length genomic clones, including promoter sequences, were isolated from the BAC library of Triticum durum cv. Langdon using full-length cDNAs of TaDREB2 and TaDREB3 as probes. Four BAC clones containing the same gene (#328 O11, #671 O3, # 741 C12, and #871 H2) were isolated for DREB2 and one BAC clone (#1111 G2) for DREB3; genes and promoter regions were identified and sequenced. The genomic sequences were designated TdDREB2 and TdDREB3L. The deduced protein sequence of TdDREB2 was identical to TaDREB2. However, the sequence of TdDREB3L was slightly different from TaDREB3. Both genes contain no introns. Analysis of 1972 and 2749 bp long promoter sequences of TdDREB2 and TdDREB3L, respectively, using PLACE software revealed potential stress-responsive cis-elements. Both promoters contain multiple abscisic acid responsive elements (ABREs), which may be responsible for the drought-inducible activation of both DREB genes. No DREs/CRTs or drought-related MYCR and MYBR elements were identified in the TdDREB2 promoter. However, it contains one site (CAATTATTG) specific for the class I HDZip transcription factors, some of which are known to be induced by ABA and drought (Agalou et al., 2008). The TdDREB2 promoter region also contains a binding site for the GATA-type zinc finger protein (AGATCCAA) associated with wounding-induced activation of some MYB transcription factors (Sugimoto et al., 2003). This element can be potentially responsible for the wounding-inducible activation of TdDREB2 as no GCC-boxes or other wounding-related elements were identified in this promoter. In contrast, the TdDREB3 promoter is rich in DRE/CRT, MYBR, and MYCR elements. Unfortunately, mapping of functional cis-element(s) was not successful because of the weak activity of both promoters (data not shown).
Slow development allows plants constitutively expressing TaDREB2 and TaDREB3 to recover after drought stress in a controlled environment
The coding regions of TaDREB2 and TaDREB3 were cloned into the pMDC32 vector under the 2 × 35S promoter (Curtis and Grossniklaus, 2003). This promoter drives strong expression in transgenic barley although it is two- to threefold weaker than the polyubiquitin promoter from maize (unpublished data). In contrast, in wheat, the 2 × 35S promoter is weak and inefficient for constitutive over-expression (data not shown). Constitutive expressions of TaDREB2 and TaDREB3 were, therefore, only analysed in transformed barley plants.
Eleven and 13 independent transgenic barley lines were obtained for TaDREB2 and TaDREB3, respectively, using the Agrobacterium-mediated transformation method (Tingay et al., 1997; Matthews et al., 2001). Southern blot hybridization indicated that most transgenic T0 lines had 2–6 copies of the transgene. In some plants, several copies of transgene were either inserted at a single site or situated very close together as no segregation was seen in four subsequent generations (data not shown). Expression levels of the transgenes were examined by RNA-blot analysis using total RNA from leaf tissue. Most T0 lines had high levels of transgene expression in leaves (Figure 3a). Analysis of transgenic plants was performed using progeny of T1–T3 generations. As experiments commenced using T1 plants which were not homozygous and often contained several copies of transgene, Northern blot hybridization was used to confirm transgene expression in each plant, and plant phenotypes were compared with levels of expression (Figure 3). Untransformed plants and plants with no transgene expression (null segregants) were used as controls. No significant difference was observed in development and stress tolerance between these two groups of control plants. In separate experiments, transgenic plants were produced with empty transformation vectors. These plants showed no differences to the other control plants.
Plant constitutively expressing TaDREB2 and TaDREB3 grew more slowly than control plants, produced less tillers and showed delayed flowering by 2–3 weeks under well-watered conditions (Table S1). The differences in plant size (plant height and number of leaves) 4 weeks after sowing, were associated with levels of transgene expression. Despite these differences of growth rate, plants with constitutive up-regulation of TaDREB2 displayed a phenotype similar to that of control plants at flowering (similar height, architecture and spike size, Figure 3a). In contrast, TaDREB3 plants only reached about 70% of the size of control plants at flowering stage (Figure 3b), with shorter spikes and lower yields in the T1 generation (Table S1). However, in later generations, the size of spikes returned to normal. No differences in fertility or grain size were observed between the transgenic lines and control plants. Both transgenic populations partially returned to normal phenotypes in the T2–T3 generations although transcription levels of the transgene, up-regulation of several downstream genes and stress tolerance remained unchanged.
Four-week-old control (C) plants and T1 or T4 generation transgenic plants were subjected to 18–21 days of drought stress (the drought survival procedure is described in Experimental procedures). Volumetric water content (VWC) in soil with the small transgenic plants changed more slowly than in pots with larger plants (VWC at 5%vs. 3% after 4 days without watering). As a result, control plants showed signs of stress including loss of turgor, leaf rolling and loss of chlorophyll much earlier than transgenic plants. Most transgenic lines showed no signs of stress for at least 2–3 days longer than control plants, some retained turgor and showed no wilting or other signs of stress for even longer (Figure 3c). Between 5% and 10% of control plants were able to recover after re-watering following the 18–21 days of drought. However, almost 100% plants of some transgenic lines survived and completely recovered within 1–2 weeks of re-watering; the smallest plants showed the quickest recovery. These results suggested that ‘improved’ drought tolerance of transgenic plants with constitutive over-expression of transgenes may be because of reduced water consumption resulting from slower growth and smaller size of transgenic plants. To confirm this hypothesis, we performed the drought survival test with two transgenic populations which constitutively over-express genes encoding HD-Zip class II and PHD-finger proteins that are not up-regulated by drought and have no relation to drought tolerance but strongly suppress plant growth in a similar manner to DREB factors (N. Kovalchuk and S. Lopato, unpublished data). The recovery of such ‘placebo’ plants was also higher than that of control plants although, survival of the ‘placebo’ transgenic plants under drought conditions was lower than for transgenic barley with constitutive up-regulation of TaDREB2 or TaDREB3.
Activation of stress responsive genes by constitutive expression of TaDREB2 and TaDREB3 in transgenic barley
One of the largest groups of genes up-regulated by drought, cold and salt stresses comprises the LEA proteins (Caramelo and Iusem, 2009). Figure 4a displays the levels of expression in transgenic and control plants of four different LEA genes from barley: HvDHN8, HvA22, HvCOR14B, and HvDHN5. A strong up-regulation of all these LEA genes was observed in three generations for three independent transgenic lines that constitutively over-expressed TaDREB3. The strongest up-regulation was shown for HvCOR14B, which was highly correlated with the transgene expression level (Table S2) and the weakest induction was for HvDHN8. Only mild (about 1.5–2-fold) up-regulation of HvDHN8 was observed in transgenic barley that constitutively over-express TaDREB2 (data not shown).
Seven barley CBF/DREB factors (HvCBF1, HvCBF3, HvCBF6, HvCBF10A, HvCBF11, HvCBF15 and HvCBF16) were found to be up-regulated by constitutive expression of TaDREB3 in all transgenic lines tested and over three consecutive generations (Figure 4b,c; Table S2). In contrast, three CBF/DREB factors (HvCBF2, HvCBF9, and HvCBF14) were down-regulated in the same lines. Expression of these 10 CBF/DREB factors were affected in exactly the same way in plants over-expressing TaDREB2 (Figure 4c). However, the magnitude of up- and down-regulation of particular HvCBFs/HvDREBs was transgene dependent (e.g. HvCBF6 is more strongly up-regulated in TaDREB2 than in TaDREB3 transgenic plants). In addition, up- or down- regulation levels of four barley CBFs/DREBs (HvCBF2, HvCBF14, HvCBF1, and HvCBF6) were generation dependent and were highest in T2 plants. This was also the generation where plant development appears to return to normal. Expression of other CBFs/DREBs (HvCBF9, HvCBF10A, HvCBF11, HvCBF3, HvCBF15) was generation dependent in the case of one transgenic line but was equally up- or down-regulated in all generations of other transgenic lines.
Transgenics with TaDREB2 and TaDREB3 over-expression showed activation of two cellulose synthases, HvCesA1 and HvCesA8, which are involved in the biosynthesis of primary and secondary cell walls, respectively, and can be potentially involved in the recovery after wounding (Figure S1). HvCesA1 is known to be involved in primary cell wall biosynthesis and was co-ordinately up-regulated with the transgene expression level in all tested transgenic lines (Table S2). However, HvCesA8 up-regulation was weaker and correlation with transgene up-regulation was poor (Table S2). Analysis of the expression of several wounding- and pathogenesis-inducible genes (HvHIR1, HvPR2_4, HvPR5, HvCAT1, and HvSOD2) gave inconclusive results (data not shown).
Constitutive up-regulation of TaDREB2 and TaDREB3 leads to improved frost tolerance
Several of the downstream genes up-regulated by over-expression of TaDREB2 and TaDREB3 are known as cold inducible or cold related. Consequently, frost tolerance in transgenic plants with constitutive over-expression of TaDREB2 and TaDREB3 was assessed.
A frost tolerance test, −6 °C for 12 h, was performed in a cold/frost cabinet on 3-week-old seedlings of transgenic and control barley plants. Under these conditions, all plants were severely damaged and only about 12% of control plants were able to recover after 2 weeks at normal temperatures. Under the same conditions, all the tested transgenic lines showed increased survival (Table 1; Figure 5), with survival of more than 50% in sublines of Line 5 (L5-7-4 and L5-4-2) for TaDREB2 and 45% in the progeny of Line 11 for TaDREB3 transgenic plants. No differences in development of the control plants or transgenic plants that survived exposure to frost treatment were detected after several weeks of recovery relative to the same lines grown under normal growth conditions. In an experiment where the minimum temperature was −4 °C, all control and transgenic plants survived. However, most of control plants showed extensive damage to leaves. In contrast, no or very little change was detected in transgenic plants (Figure 5).
Table 1. Results of the frost tolerance test
Transgenic line progeny
Number of tested plants
Number of survived plants
% of survived plants
Accuracy of data†
†Differences in recovery between transgenic lines and control plants were tested in a Pearson’s Chi-squared test (n.s., *, **, ***, mean nonsignificant differences, P-value <0.1, <0.01, <0.001, respectively).
Transgenic barley and wheat plants with drought-inducible expression of DREB factors
As noted above, constitutive over-expression of TaDREB2 or TaDREB3 led to reduced growth in transgenic barley plants and this appeared to at least partially account for the observed drought tolerance phenotype. To decrease or eliminate undesirable phenotypes, barley and wheat were transformed with constructs in which the 2 × 35S promoter was exchanged for a 634-bp long fragment of the drought- and salt-inducible Rab17 promoter from maize (Vilardell et al., 1990). Twenty independent barley lines were produced for each construct. For wheat transformed with pRab17-TaDREB2, 45 independent lines were generated and for pRab17-TaDREB3 construct, 18. The presence of the transgene was confirmed by PCR using specific primers from the Rab17 promoter and nos terminator.
In the case of barley, transgenic plants were generally slightly smaller than control plants. However, the difference in size was observed only in some plants with high levels of basal promoter activity (uninduced) and was less pronounced than in transgenic plants with constitutive over-expression of the same genes under 2 × 35S promoter.
In wheat, plants had a comparable phenotype to control plants (Figure 6a, before drought) for both transformed genes. Under well-watered conditions and under moderate water deficit (until 5% of VWC, −0.6 MPa), the stomatal conductance of the transformed plants were similar to that of control plants, decreasing from 238 ± 29 to 32 ± 3 mmol/m2/s1 with soil drying. This resulted in no difference in leaf water status, regardless to the soil water status, with leaf water potential decreasing only slightly with soil drying, from −0.87 MPa under well-watered conditions to −1.16 MPa under drought (Figure 6c).
Drought-inducible expression of DREB factors increases the plant survival after severe drought stress
Wheat plants transformed with pRab17-TaDREB2 and pRab17-TaDREB3 constructs showed no developmental setbacks during the first 3 weeks after germination. As the point water was withheld, transgenic wheat plants were the same or very similar to control plants. Drought recovery experiments were performed using the same conditions as above but the length of drought was 14 days (over the last 10 days the VWC in soil was lower than 3%). During the drought exposure, the behaviour of control and transgenic plants was nearly the same: all plants dried at a similar rate and were showing severe damage at the last day before re-watering although the transgenic plants remained marginally greener (Figure 6a). One week after re-watering, only 14% of control plants had recovered, whereas between 33% to 100% of the transformed plants recovered, depending on line (Figure 6a,b). The transgenic lines showing the strongest recovery from drought stress tended to be those showing the strongest induction of expression under drought stress. A larger collection of transgenic lines then used here would be needed to establish a clear correlation with expression levels.
Wheat plants transformed with pRab17-TaDREB3 recovered more quickly than plants transformed with pRab17-TaDREB2, and started to flower 3 weeks after re-watering; plants transformed with pRab17-TaDREB2 started to flower 3–4 days later than this (Figure 6a). Three weeks after re-watering, both transgenic populations looked healthy; they had similar size of spikes and number of fertile florets compared with well-watered control plants. Two of 20 control plants survived drought stress but recovered much more slowly than transgenic plants. They remained about one-third of normal size when they started to flower and subsequently produced fewer and smaller spikes compared to well-watered controls.
Activity of maize rab17 promoter in wheat and barley
Northern blot and Q-PCR analysis of the expression of DREB factors under Rab17 promoter were performed using leaf samples collected 1 day before watering was stopped, 3 days after plants showed clear signs of wilting (VWC in soil 2%) and, for some lines, 3 weeks after re-watering. Both Northern blots (data not shown) and Q-PCR analysis of barley plants revealed relatively high basal level of activity of the Rab17 promoter in leaves in the absence of stress (Figure S2a). Levels of basal activity differed between independent transgenic lines of barley. Developmental abnormalities observed in some plants correlated with levels of basal promoter activity. No or very little basal activity of the maize Rab17 promoter was detected in wheat by both Northern blot hybridization and Q-PCR. However, drought stress quickly and strongly activated the ZmRab17 promoter (Figures 7a and S2b). Exogenous DREB expression was limited to the duration of the stress and the first 1–2 days of recovery after re-watering. This led to minimal undesired changes in plant development but was sufficient to confer improved plant survival after drought stress. In both barley and wheat, re-watering led to the rapid down-regulation of the Rab17 promoter, but low levels of transgene transcripts were still detectable 2 weeks after re-watering (Figure S2b). We have not observed any negative influence of the low level transgene expression on flowering time, size, the number and shape of spike and size or shape of grain.
Barley plants with moderate levels of basal activity of Rab17 promoter were used in the frost tolerance test as a model for moderate constitutive expression. No induction of the Rab17 promoter by cold/frost temperatures was detected (Figure S2c). Moderate constitutive over-expression of both TaDREB2 and TaDREB3 led to considerable improvement of frost tolerance, which was, however, a bit lower then improvement in frost tolerance under the strong 2 × 35S promoter (Table. 1; Figure 5g). On the other hand, the moderate constitutive expression seen with Rab17-driven expression in barley significantly reduced the negative pleiotropic effects on development.
Activation of stress-inducible genes by inducible over-expression of DREB factors
Expression of nine wheat LEA/COR/DHN genes known to be induced by drought and cold were examined in the transgenic plants. The expression results were initially used to determine a ratio of expression levels under drought stress (time of sampling: 4 days after soil VWC reached 2%) relative to well-watered conditions. These data are then used to calculate the increase in induction of expression in transgenic plants relative to induction in control plants (Figure 7b). This reflects additional induction of these genes by DREB transgenes relative to induction solely by drought and potentially related to the effects of the endogenous DREB genes.
Although induction by the transgene reached 50-fold for some LEA/COR/DHN genes, most genes showed lower induction. Activation of some LEA/COR/DHN genes appeared to be specific for only one of the transgenes. For example, the induction of expression of TaRAB17 was much stronger in TaDREB3 transgenic lines, while induction of expression of TaWZY2 was stronger in TaDREB2 transgenic plants.
Both constitutive and inducible expression of TaDREB3 in transgenic barley plants lead to specific up-regulation of cold-inducible HvCOR14B gene. Levels of HvCOR14B correlated well with levels of TaDREB3 over-expression (Table S2). No up-regulation of HvCOR14B was detected in TaDREB2 transgenic plants (Figure S3).
Real and seeming drought tolerance
The grain of cereals has a generally high level of ABA relative to other plant organs and shows strong induction of expression of genes that are up-regulated in other plant tissues only in response to environmental stresses (Ali-Benali et al., 2005; Sreenivasulu et al., 2006). Some of these genes, including LEA/COR/DHN genes, are probably involved in the protection of cells during grain maturation and desiccation and help maintain cells and tissues viability until germination (Ali-Benali et al., 2005; Rorat, 2006). In the absence of stress, endogenous TaDREB2 and TaDREB3 were expressed at higher levels in flower and grain tissues, relative to leaves (Figure 2a). However, transcript levels of both genes were strongly up-regulated in leaves by drought and slightly by cold (Figure 2b,c). These patterns of expression suggest a role for TaDREB2 and TaDREB3 in protection of plant tissues from desiccation. Therefore, strong up-regulation of expression of these factors in wheat and barley may help to increase survival under water deficit.
One of the first reactions of plants to abiotic stress is to decrease growth rates (Boyer, 1970). This allows plant to decrease water consumption and save energy. Constitutive over-expression of most DREB/CBF genes tested so far in transgenic plants leads to stunted growth, mild or strong dwarfism, slower development and a delay in flowering time (Kasuga et al., 1999; Kim et al., 2004; Oh et al., 2007). One of the reasons for the smaller size could be because of down-regulation of gibberellin deactivating genes (Magome et al., 2004, 2008). This undesirable agronomic trait, although a natural physiological reaction of plants to drought, becomes severe in transgenic plants with strong constitutive over-expression of stress-related regulatory genes. These negative phenotypes were also observed in our transgenic plants with constitutive over-expression of TaDREB2 and TaDREB3 (Table S1). However, the slow growth of transgenic plants makes it difficult to compare transgenic with control plants and complicates analysis of changes in drought tolerance. Even small differences in plant development lead to errors in the assessment of drought tolerance and imprecise or incorrect conclusions. However, the analysis of the regulation of stress-inducible genes (Figure 4) suggested that the improvement in drought tolerance of plants seen with constitutive over-expression of TaDREB2 and TaDREB3 is not simply a reflection of stunted growth but also the result of increased protection of cells from desiccation on the molecular level.
Partial normalization of transgenic plant growth and flowering time became noticeable in the T2 generation and subsequent generations of transgenic plants. This normalization of development cannot be explained by diminishing of transcription levels of transgene, as transgene expression was assessed in each generation and was not seen to decline. It also seems unlikely that protein modification or turnover changed as drought and frost tolerance as well as levels of induction of some stress-inducible genes in T1–T4 generations of transgenic plants remained roughly the same and correlated with transgene expression. However, generation-dependent changes in gene expression were observed for several groups of downstream genes encoding CBF/DREB factors (Figure 4b,c) and some aquaporins (data not shown). For these genes, expression was down-regulated in T0 plants but returned to normal in later generations of transgenic plants. This suggests that plants use alternative regulatory pathways to normalize the expression of genes which are indirectly regulated by TaDREB2 and TaDREB3 and might otherwise disturb plant development in the absence of stress. It should also be noted that the plants selected for further study were those that showed the strongest expression levels rather than based on zygocity. Consequently, some of the variation in overall expression was likely because of selection for homozygous lines from the T2 generation onwards.
TaDREB2 and TaDREB3 are frost tolerance genes
Protein alignment of TaDREB2 and TaDREB3 to sequences of other DREB/CBF factors from barley and wheat (Figure 1) revealed that both proteins are close homologues of CBF factors, which were demonstrated to be involved in cold-stress response. The TaDREB2 is a close homologue of CBF7 from barley and from Triticum monoccocum. Southern blot hybridization to nullisomic-tetrasomic lines of hexaploid wheat (Sears, 1954) with full-length cDNA of TaDREB2 as probe revealed that the gene is located on Group 3 chromosomes of hexaploid wheat and most likely present as a single copy (data not shown). In contrast, TmCBF7 was mapped on chromosome 5A at or near the Fr-A2 locus of T. monoccocum and was previously described as related to frost tolerance (Vágújfalvi et al., 2005). TaDREB2 protein has highest sequence conservation with TINY from Arabidopsis thaliana, which belongs to the small subfamily of proteins with sequences distinct from both DREB1 and DREB2 subfamilies from Arabidopsis (Sun et al., 2008).
TaDREB3 is a close homologue of TmCBF5 (Figure 1). The TmCBF5 gene was mapped on chromosome 7A of T. monoccocum (Miller et al., 2006) and TaDREB3 also mapped on the Group 7 chromosome of Triticum aestivum and is also most likely a single copy gene (data not shown). However, the bread wheat orthologues of TmCBF5, TaCBF5, is differentially expressed in cold acclimated frost tolerant and frost sensitive lines relative to nonacclimated controls and hence remains an interesting candidate gene for the frost tolerance (Sutton et al., 2009). It was demonstrated that both TaDREB2 and TaDREB3 are weakly up-regulated by cold (Figure 2c).
Strong up-regulation of some of drought/cold stress-inducible LEA/DHN/COR genes was detected in transgenic barley plants with constitutive over-expression of TaDREB3 under normal growing conditions (Figure 4). Analysis of the expression levels of 10 different CBF/DREB genes, some of which were mapped in vegetative frost tolerance QTLs, suggests that activation of LEA/DHN/COR genes is a result of co-operative action of DREB3 and CBFs (Figure 4). As expected, levels of up-regulation of the stress-inducible genes in transgenic plants differ between plants having constitutive or inducible expression of the transgene: on the whole, up-regulation under the inducible promoter in transgenic barley is lower and some genes that were up-regulated under constitutive expression were not up-regulated when the inducible promoter was used (data not shown). This can be explained by different spatial patterns of transgene expression under 2 × 35S and Rab17 promoters and the relatively short period of activity of the Rab17 promoter.
Previous work with close homologues/orthologues of TaDREB2, TaDREB3 and several LEA/COR/DHN genes including Cor14 (Vágújfalvi et al., 2000; Miller et al., 2006; Ganeshan et al., 2008) suggested a possible involvement of these genes in cold and frost tolerance. Furthermore, constitutive expression of TaDREB2 and TaDREB3 leads to the constitutive expression of a large number of genes normally induced by cold stress. The elevated expression of these genes may reduce the requirement for cold acclimation of transgenic plants. Consequently, the survival rates improved for T2 and T3 transgenic barley seedlings expressing TaDREB2 and TaDREB3 under temperatures as low as −6 °C and with very short (several hours) acclimation (Figure 5). The survival rates of all the lines tested were higher than for control plants grown under the same conditions (Table 1).
Strong constitutive expression of transgenes led to negative developmental phenotypes, therefore weak constitutive expression was also investigated to see if frost tolerance was still observed. The ZmRab17 promoter was used as this promoter showed moderate basal level of expression in barley leaves in the absence of drought stress (Figure S2a,c). Although this promoter is strongly induced by drought, it is not induced by cold (Figure S2c) and, in the absence of drought stress, can be used as a moderate ‘constitutive’ promoter. Frost survival rates in barley plants expressing TaDREB2 or TaDREB3 under moderate ZmRab17 promoter were slightly lower than for plants with the strong 2 × 35S promoter but higher than for control plants (Table 1). Therefore, the Rab17 promoter from maize could be used together with TaDREB2 and TaDREB3 to improve both drought and frost survival rates in barley with minimal changes in plant development.
Reduction of negative effects on plant development
The promoter of the Rab17 gene from maize (Vilardell et al., 1990) was used to drive strong drought specific expression of TaDREB2 and TaDREB3. According to results of others (Close et al., 1989; Mundy et al., 1990) and results presented here, the Rab17 promoter has low basal activity in most plant tissues but is stronger in embryo and developing endosperm. It is rapidly and strongly induced by drought but not by cold. Genes similar to the maize Rab17 were isolated from other plant species, and in many cases promoter activity was similar (Michel et al., 1994). ABA-induced activation of the Rab17 promoter was studied in heterologous systems (in stably transformed tobacco and by transient expression in rice protoplasts) and its drought inducibility was initially explained by the presence of the ABRE in a 100 bp segment of the promoter (Vilardell et al., 1991). Later, several cis-elements (five putative ABREs and four other sequences) important for the strong activity were mapped in the promoter (Busk et al., 1997). Six of these elements were shown to be important for expression in embryos, whereas only three elements were responsible for the basal and stress-inducible expression in leaves. Among these elements was a new GC-rich rab Activator element, CACTGGCCGCCC, responsible for the low constitutive expression of Rab17 in maize leaves, and the drought responsive elements (DREs) (Busk et al., 1997). Finally, it was found that the ZmRab17 promoter is activated by ABA, drought and salt stress through the single DRE (DRE2). Two AP2 domain containing transcription factors from maize, designated DBF1 and DBF2, were isolated in a yeast one-hybrid screen using DRE2 as bait. Promoter activation by over-expression of DBF1 and repression by over-expression of DBF2 were demonstrated (Kizis and Pages, 2002). ABA appears to play a role in the regulation of DBF activity, and the ABA-dependent pathway was suggested as the regulatory mechanism that acted through the C-repeat/DRE element (Kizis and Pages, 2002). The results presented by Kizis and Pages (2002) suggested that the Rab17 promoter could be suitable for drought-inducible over-expression of DREB factors. The promoter is strong, has low basal activity in maize, induction of the promoter starts within several hours of the plant sensing water deficit, and its activity quickly returns to basal level after re-watering. Regulation of the promoter by DREB factors under stress conditions could potentially further increase activity of the promoter in transgenic plants by a feedback loop.
Analysis of the activity of this promoter in barley and wheat demonstrates that in these plants, the promoter behaves similarly to maize; under drought stress it is rapidly and strongly induced in leaves, whereas cold stress does not induce expression from the ZmRab17 promoter (Figure S2c). Different levels of basal activity were observed in transgenic barley and wheat if DREB factors were used as transgenes (Figure S2). However, these differences were not observed when transcription factors from other families were over-expressed by this promoter (data not shown). This difference may be because of differences in the ability of TaDREB2 and TaDREB3 to negatively influence expression of genes responsible for the basal activity of the promoter in barley as opposed to wheat.
The high basal activity of the Rab17 promoter in barley relative to wheat still leads to negative phenotypes. However, these are far milder than the negative phenotypes of transgenic barley with a strong constitutive promoter. In contrast, wheat transgenic plants before stress and after re-watering were difficult to distinguish from control plants.
Enhanced survival of transgenic wheat plants under drought stress can be explained by up-regulation of wheat LEA/COR/DHN genes to higher levels that under normal drought stress. In control plants, most of the tested LEA/COR/DHN genes have low to moderate basal levels of expression in the absence of stress and are strongly up-regulated by drought. In transgenic wheat plants under well-watered conditions, expression of these genes remained at the same level as in well-watered control plants. However, under drought, expression levels of the wheat LEA/COR/DHN genes examined were from 1.5- to 50-fold higher than in stressed control plants (Figure 7b).
All genes except TaRAB17 were more strongly up-regulated in TaDREB2 transgenic versus control plants. The strongest up-regulation was observed for Wcor18, Wcor80, TaRAB16.5, and TaRAB18. Wlt10 and Wcor410 were weakly up-regulated in both transgenics. These data and the results of the analysis of stress-inducible genes in barley indicated qualitative and quantitative differences in up-regulation of downstream genes by TaDREB2 and TaDREB3 which probably resulted in differences of the developmental and stress phenotypes of the transgenic lines.
Several important conclusions can be made from the results of this work. Constitutive up-regulation of TaDREB2 and TaDREB3 in transgenic barley plants improves survival rates under severe drought and frost stresses but this leads to negative developmental phenotypes. The undesired changes in plant development can be at least partially prevented by use of weak constitutive and stress-inducible promoters. The promoter used here had low activity in the absence of stress was induced by drought stress and quickly returned to basal levels after re-watering.
The enhanced stress tolerance appeared to result from the up-regulation of many stress-inducible genes involved in the protection of cell integrity under severe stress. Independent transgenic lines varied in their drought recovery rates (33%–100%) and this variation appeared to be related to the level of induction of expression under drought stress. However, a more extensive study would be needed to confirm this correlation.
Improved ‘survival’ under severe drought did not provide any advantage to transgenic compared to control plants during prolonged growth under water limited conditions. These conditions were sufficient for plant survival but negatively influence crop yields. Comparison of the transgenic and control plants that survived the drought stress showed no improvement in grain yield; in fact the transgenic showed a slightly reduced yield. However, the greatly increased plant survival should translate to improved overall yield under field condition. Clearly, field evaluation is now necessary to determine the efficacy of these transgenes.
For vegetative frost tolerance, good results were obtained by using moderate constitutive expression. Further improvement may be achievable by using weak cold-inducible promoters, with low basal activity and/or tissue specificity.
Both TaDREB2 and TaDREB3 strongly regulate many different CBF/DREB genes from barley. Although the TaDREB2 and TaDREB3 proteins are structurally very different, they appear to regulate similar DREB/CBF genes. It seems probable that the same situation applies to other studies where DREB/CBF factors were over-expressed. The stress tolerant phenotypes and regulation of downstream genes described in multiple ‘DREB’ papers may have resulted from the simultaneous activation or repression of other DREB/CBF factor(s). This may have lead to nonspecific binding and activation of nontarget promoters resulting in loss of specificity of transgene action and the strong negative phenotypes often reported. The use of weak, tissue-specific and stress-inducible promoters should partially alleviate this problem.
Plasmid construction and plant transformation
Full-length cDNAs of TaDREB2 (Acc. DQ353852) and TaDREB3 (Acc. DQ353853) isolated in the Y1H screen from a wheat grain cDNA library (Lopato et al., 2006) were used as templates. Coding regions of TaDREB2 and TaDREB3 cDNAs were cloned into: (i) the pMDC32 vector (Curtis and Grossniklaus, 2003) downstream of the vector’s duplicated 35S promoter and (ii) a pMDC32 vector in which the 2 × 35S promoter was excised using HindIII–KpnII restriction sites and replaced with a 634 bp fragment of the ZmRab17 promoter (Busk et al., 1997). All four constructs were transformed into barley (Hordeum vulgare L. cv. Golden Promise) using Agrobacterium-mediated transformation and the method developed by Tingay et al. (1997) and modified by Matthews et al. (2001). Wheat (T. aestivum L. cv. Bobwhite) was transformed using microprojectile bombardment as described by Kovalchuk et al. (2009). pRab17-TaDREB2-nos and pRab17-TaDREB3-nos fragments were excised from the respective constructs using PmeI and BsaXI, gel purified and co-transformed together with the pUbi-hpt-nos cassette (3676 bps fragment of the vector plasmid, cut with PmeI–SmaI) into wheat using microprojectile bombardment.
Isolation and analysis of genomic clones
Genomic sequences and 5′ upstream regulatory sequences of orthologues/homologues of TaDREB2 and TaDREB3 were isolated using the procedure described by Kovalchuk et al. (2009) from the BAC library prepared from T. durum L. cv. Langdon. The full-length cDNAs of TaDREB2 and TaDREB3 were used as probes. Four BAC clones with strong hybridization signals were isolated with TaDREB2 as a probe. All contained the same gene, which is identical to TaDREB2 and thus was designated TdDREB2. Only one clone with a strong hybridization signal was selected for TaDREB3. It encodes a close homologue of TaDREB3 and this gene was designated DREB3-like (TdDREB3L). The 1972- and 2749-bp long promoter sequences were isolated for TdDREB2 and TdDREB3L, respectively. The promoters were analysed for the presence of potential stress-related elements using PLACE software (http://www.dna.affrc.go.jp/PLACE/signalscan.html) and a database of plant cis-acting regulatory DNA elements (Higo et al., 1999).
Plant growth and stress conditions
For phenotypic analysis, plants were grown under glasshouse conditions with an average day and night temperature of 25 and 16 °C, respectively, with the day length extended to 15 h with supplemental lighting. T1 and T2 generation plants were monitored for changes in growth rate, plant height, heading time, number of tillers, spike phenotype, grain phenotype and yield. Null segregants from the transgenic lines and untransformed plants were used as controls.
Seedlings for the ‘survival’ drought tolerance test were grown under growth room conditions, with a 16 h day at 24 °C and night temperatures of 16 °C. Progeny of T1 and T2 generations of transgenic plants (10 plants per independent transgenic line) were grown either in 12-cm square pots or as pairs of control and transgenic plants in 20-cm pots for 3 weeks. The volumetric water content (VWC) of each pot was monitored at least every second day during the experiment and plants with the same VWC were used for comparison and documentation. Three weeks after germination, water was withheld. Seven to 10 days after the pots reached 2%–3% VWC and wilting was observed, the plants were re-watered. Plants were assessed for recovery after 1 and 3 weeks of re-watering, and stress-tolerant plants were transferred to the glasshouse for generation of seeds. Leaf samples were collected from well-watered plants 1 day before withholding water (for all tested transgenic and several control plants), 3–4 days after the VWC had reached 2% and transgenic lines with inducible promoter and several control plants showed clear wilting.
To check if differences in stomatal conductance and leaf water status could explain the differences of recovery after severe drought, an experiment was carried out at stable soil water content. Pots (22 and 22 cm in deep) were filled with 4 kg of soil and sampled for measurement of water content. One plant of each line (two DREB2 and two DREB3 transformed lines) and one control plant were sown in each pot. At the two-leaf stage, soil was dried and maintained at the target soil water content by watering every 2 days. Three different soil water contents were tested: well watered (VWC = 16%), VWC = 6% corresponding to a predawn soil water potential of −0.3 MPa and VWC = 4% corresponding to a predawn soil water potential of −0.6 MPa. Because a VWC of 2.5% corresponded to the wilting point, this level of water deficit was not tested. Stomatal conductance was measured at midday, the day after watering with a diffusion porometer (Decagon SC-1 Leaf Porometer, Pullman, WA, USA) on nonexpanding leaves. Leaf water potential was measured at the same time on nonexpanding leaves. Leaves were cut, placed into a Scholander-type pressure chamber (Soil Moisture Equipment Corp., Santa Barbara, CA, USA). The pressure at which extruded sap first began to wet the cut surface was registered as the opposite of leaf water potential.
The most appropriate temperature and length of treatment for assessing frost tolerance was determined as the treatment that killed most of the control barley plants. Seeds were sown in 20-cm pots and grown in the growth chamber with 12 h light at 23 and 18 °C at night temperatures. Seedlings at the four-leaf stage (approximately 3 weeks after germination) were transferred to a cold chamber (BINDER, Tuttlingen, Germany). To protect the plant roots from frost damage, pots were insulated. Temperature profiles differed in the severity and duration of the minimum temperature, which were either −5 °C for 1 h, or −6 °C for 1, 6, 12 and 14 h in preliminary tests performed on control barley plants. In all experiments, the ice nucleating agent SNOMAX® (Sno-Quip Pty Ltd, Mittagong, NSW, Australia) was sprayed onto plants as a 2 g/L solution 2 h before the chamber reached the minimum temperature. A HOBO data logger (Onset Computing, USA) was used to control and record temperature, light intensity and humidity in the chamber. After the frost treatment, plants were moved back to the growth chamber and maintained under the same growth conditions as before the frost treatment. The number of recovered plants was recorded after 24 h of recovery. Based on these preliminary tests, the minimum temperature of −6 °C for 12 h was selected to test frost tolerance in the transgenic barley plants. Only 10%–20% of control plants survived and recovered under this regime.
In each experiment, three sublines for each independent transgenic line and from each subline progeny of 12 plants were tested for frost tolerance. Three transgenic plants were grown along with one control plant in each pot (20 cm). The experiment was repeated at least three times for each transgenic line. Expression of the transgene was assessed by Northern blot hybridization using total RNA from tissues collected shortly before the frost treatment and when the temperature reached 4 °C (for ZmRab17 promoter only).
Analysis of gene expression
Transgene presence and expression in the T0–T4 generations of transgenic plants were analysed by either Southern or Northern blot hybridization as described by Sambrook and Russell (2001) or by RT-PCR and Q-PCR. Q-PCR analysis was performed using primers from the coding region and nos terminator for transgenes and primers from 3′ untranslated regions of endogenous stress-inducible genes. cDNAs prepared using SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA) from leaf tissues of control and transgenic plants collected before and during stress and during plant recovery were used as a template. Leaf tissues from several independent lines and several consecutive generations of transgenic plants were used for the analysis of downstream genes. Each PCR was repeated three times. The Q-PCR procedure was described by Burton et al. (2008). mRNA copy number for each tested gene was normalized against four control genes as described by Burton et al. (2008). Primer details appear in Table S2. Different tissues of T. aestivum cv. Chinese spring plants were used for tissue-based analysis of expression of endogenous TaDREB2 and TaDREB3. Inducibility by drought was analysed in several independent plants of T. aestivum cv. RAC875. Material was collected from plants grown under well-watered conditions and 3–4 days after the plants had started to show signs of drought stress. For cold inducibility, analysis of two independent, 6-week-old plants of T. aestivum cv. RAC875 were incubated at 4 °C, and leaf material was collected at 0, 1, 4, 24 and 48 h after plants were transferred to 4 °C. For the wounding experiment, leaf and 10–15 DAP old grain from two independent plants of T. aestivum cv. Chinese spring were wounded using a fine metal brush, and material was sampled at 0, 0.5, 1, 2, 3, 8 and 17 h after wounding.
We thank M. Pallotta and N. Bazanova for the technical assistance with isolation and characterization of BAC clones, Dr K. Oldach for providing us with Q-PCR primers for wounding-inducible genes, and Dr U. Langridge and R. Hosking for assistance with growing plants. This work was supported by the Australian Research Council, the Grains Research and Development Corporation and the Government of South Australia and the University of Adelaide.