Regulatory changes in TaSNAC8‐6A are associated with drought tolerance in wheat seedlings

Summary Wheat is a staple crop produced in arid and semi‐arid areas worldwide, and its production is frequently compromised by water scarcity. Thus, increased drought tolerance is a priority objective for wheat breeding programmes, and among their targets, the NAC transcription factors have been demonstrated to contribute to plant drought response. However, natural sequence variations in these genes have been largely unexplored for their potential roles in drought tolerance. Here, we conducted a candidate gene association analysis of the stress‐responsive NAC gene subfamily in a wheat panel consisting of 700 varieties collected worldwide. We identified a drought responsive gene, TaSNAC8‐6A, that is tightly associated with drought tolerance in wheat seedlings. Further analysis found that a favourable allele TaSNAC8‐6A In‐313, carrying an insertion in the ABRE promoter motif, is targeted by TaABFs and confers enhanced drought‐inducible expression of TaSNAC8‐6A in drought‐tolerant genotypes. Transgenic wheat and Arabidopsis TaSNAC8‐6A overexpression lines exhibited enhanced drought tolerance through induction of auxin‐ and drought‐response pathways, confirmed by transcriptomic analysis, that stimulated lateral root development, subsequently improving water‐use efficiency. Taken together, our findings reveal that natural variation in TaSNAC8‐6A and specifically the TaSNAC8‐6A In‐313 allele strongly contribute to wheat drought tolerance and thus represent a valuable genetic resource for improvement of drought‐tolerant germplasm for wheat production.


Introduction
Drought has caused a 13.7 % average loss in worldwide cereal production over recent decades (Lesk et al., 2016;Pennisi, 2008). Among these cereals, wheat (Triticum aestivum L.) is an important staple crop, mainly grown in arid and semi-arid regions worldwide. Its production is frequently compromised by water scarcity, which is exacerbated by the trends of climate warming and population growth on a global scale (Lesk et al., 2016). Therefore, improvement of drought-tolerant germplasm for higher yields in arid cultivation regions is therefore an important goal among wheat breeding programmes. However, drought tolerance is a complex trait controlled by many genes with minor effects. Several previously identified genes necessary for drought tolerance have been characterized for their involvement in signal perception, signal transduction and transcriptional regulation of other genes necessary for drought tolerance, including protein kinases and transcription factors (TFs), such as members of the AP2/ERF, MYB, bZIP and NAC TF families (Hu and Xiong, 2014;Liu et al., 2014;Nakashima et al., 2012). TFs are particularly attractive targets for genetic improvement of abiotic stress tolerance in crops because the engineering of a single TF can affect all of the genes in its respective regulon, consequently impacting the physiological process or trait for which they are responsible (Gahlaut et al., 2016;Hu and Xiong, 2014).
Notably, a number of these stress-responsive NAC TFs have been extensively validated through transgenic overexpression as a means to increase crop drought tolerance and to improve yield or biomass. For example, overexpression of SNAC1 in rice or wheat led to significantly enhanced drought tolerance and yield under severe drought conditions without any observable phenotypic changes or yield penalty under well-watered growth conditions (Hu et al., 2006;Saad et al., 2013); elevating the expression of OsNAC5, OsNAC6 and OsNAC10 in rice can also improve drought tolerance as well as grain yield under drought conditions (Jeong et al., 2010;Jeong et al., 2013;Lee et al., 2017). In addition, overexpression of OsNAC5, OsNAC6 and OsNAC10 in rice roots induces radial root growth, which is a key adaptation to drought stress through root growth and development in response to water-deficit conditions (Jeong et al., 2010;Jeong et al., 2013;Lee et al., 2017). Thus, previous research has established that stress-responsive NAC TFs can serve as useful candidates for improvement of stress-tolerant crop germplasm without introducing detrimental growth defects.
However, as of yet no quantitative trait loci (QTL) responsible for drought tolerance in wheat have been cloned, despite some reports of mapping information (Fleury et al., 2010;Gahlaut et al., 2017;Quarrie et al., 1994). In this study, we used linkage disequilibrium (LD)-based association mapping, a powerful tool for dissecting complex agronomic traits and identifying alleles that can contribute to crop improvement, in conjunction with candidate gene association analysis for detecting DNA polymorphisms (e.g. SNPs) or genetic markers within, or tightly linked to genes contributing to complex traits (Huang et al., 2011;Liu et al., 2017;Wang and Qin, 2017;Xiao et al., 2017). This approach has previously been successfully applied to detect allelic diversity of genes controlling crop drought tolerance, especially in rice and maize (Liu et al., 2013;Mao et al., 2015;Singh et al., 2015;Songyikhangsuthor et al., 2014;Wang et al., 2016;Xiang et al., 2017;Xiong et al., 2018).
Using the IWGSC, 2018 wheat genome (IWGSC, 2018), we identified and characterized 39 TaSNAC genes, which formed 13 homologous groups in wheat. The association between genetic variations within each candidate TaSNAC gene and drought tolerance, evaluated in terms of survival rate (SR) under severe drought stress at the seedling stage, was quantified using a candidate gene association strategy and a diversity panel consisting of 700 wheat varieties from global germplasm. A strong association was detected between TaSNAC8-6A and SR in the seedling stage. Further analysis showed that sequence variation in the promoter region of TaSNAC8-6A, specifically an indel in the ABRE motif, was associated with distinct patterns of gene expression in response to drought stress across several different wheat varieties. Additionally, overexpression of TaS-NAC8-6A up-regulated several putative NAC-targeted genes involved in auxin-and drought-response signalling pathways, subsequently stimulating the development of lateral roots and consequently enhancing tolerance to drought stress.

Identification and characterization of TaSNAC genes
In order to identify stress-responsive NAC subfamily genes in wheat, we conducted a BLAST search against the wheat genome database (IWGSC RefSeq v1.1) using Arabidopsis and rice NAC TFs as queries. Returned hits were pooled and redundancies were eliminated. Each sequence was manually examined to confirm NAC domain presence and location. In total, 260 NAC-domain proteins were identified in the wheat genome (Table S1). Based on Arabidopsis and rice NAC superfamily phylogenies (Nakashima et al., 2012), 39 wheat NAC TFs were classified in the stressresponsive NAC subfamily. Since wheat is hexaploid with three subgenomes (2n = 6x = 42; AABBDD), we further classified these 39 wheat NAC TFs into 13 homologous groups. The three homologous alleles were designated as TaSNACX_YA, TaSNAC-X_YB or TaSNACX_YD, where X denotes the homologous group number and Y the wheat chromosome where it is located. The previously identified proteins TaNAC2 (Mao et al., 2012) and TaNAC29 (Huang et al., 2015) were renamed as TaSNAC5-5A and TaSNAC13-2A, respectively ( Figure 1a).
To characterize the phylogenetic relationships between TaS-NAC proteins and their orthologs from rice, B. distachyon, maize and Arabidopsis, a neighbour-joining phylogenetic tree was reconstructed using full-length protein sequences ( Figure S1). These stress-responsive NAC proteins clustered into two distinct subgroups consistent with previous reports (Nakashima et al., 2012). Several NAC proteins from four monocot species displayed pairwise correspondences with high bootstrap support, suggesting high interspecific conservation. Furthermore, wheat NACs were phylogenetically closer to B. distachyon than to rice or maize, consistent with the closer evolutionary relationship between wheat and B. distachyon than with other species in this study. TaSNAC7-2A/B/D and TaSNAC8-6A/B/D, and TaSNAC12-2A/B/D and TaSNAC13-2A/B/D most likely represent duplicated genes in wheat since only one ortholog from each pair was found in the B. distachyon, rice and maize genomes ( Figure S1).
In order to isolate and functionally characterize individual TaSNAC genes, we cloned 39 of them from wheat cv. Chinese Spring and performed a transactivation activity screen of the encoded proteins using a yeast transactivation assay. Due to the high sequence similarity within groups of homologous proteins (>95%), we subcloned 13 genes, one from each homologous group, into a yeast GAL4-DNA binding domain expression vector, which was subsequently transformed into yeast reporter cells. The TaSNAC-GAL4 fusion proteins exhibited a relatively higher level of transactivation activity as demonstrated by their ability to grew well on selective medium (Figure 1b). These genes are distributed across 18 of the wheat chromosomes except chromosomes 7A, 7B and 7D. Gene structure analysis revealed that three TaSNAC genes have one exon, seven have two exons, 28 have three exons, and one has four exons ( Figure S2).

Expression profiles of TaSNAC genes
In order to better understand the function of each of the TaSNAC genes, we first examined TaSNAC expression in eight different wheat tissues under normal growth conditions by quantitative real-time PCR (qRT-PCR). The PCR primers were designed to amplify the homologous alleles at a conserved locus; for example, the relative expression level of TaSNAC1 represents the combined expression of all three homologous TaSNAC1 alleles (TaSNAC1-1A, TaSNAC1-1B and TaSNAC1-1D). The results revealed high variation among the expression patterns of different TaSNAC genes ( Figure 2a). TaSNAC1/2/6/7 had relatively low expression levels in the eight tissues, while TaSNAC3/11 were expressed at relatively high levels. TaSNAC4/8 were also expressed at relatively high levels in these tissues. However, TaSNAC4 had higher expression levels in seedling leaves, stems and roots, while TaSNAC8 had higher expression in seedling leaves and stems, flag leaves, spikes at 15-DAP (days after pollination) and 15-DAP grains. The remaining genes were highly expressed only in specific tissues. For example, TaSNAC5/12/13 were expressed in leaves; TaSNAC9 was expressed in 15-DAP spikes; and TaSNAC10 was expressed at high levels in seedling leaves, 15-DAP spikes and 15-DAP grains (Figure 2a).
To better understand the regulatory functions of individual TaSNAC gene in response to drought, the expression of 13 TaSNACs was examined by qRT-PCR in leaves and roots of threeleaf stage seedlings under drought stress. We observed a dramatic up-regulation of all TaSNACs in response to drought, especially in leaf tissue ( Figure 2b). The expression of TaSNAC1 and TaSNAC5 in leaves after drought stress was increased > 100fold, indicating extreme sensitivity to drought. In leaf tissue, some genes exhibited similar expression profiles, such as the two pairs of genes TaSNAC11/13 and TaSNAC8/12. Some TaSNACs were up-regulated after light/early drought stress, such as TaSNAC2/8/ 10/12 in leaves and TaSNAC2/4/5/6/7/9 in roots, suggesting that these genes are transcriptionally activated as part of an early response to drought stress. Other genes were up-regulated only after severe stress, for example TaSNAC1 in leaves and roots, suggesting that they participate in the response to severe drought. We also observed sets of genes that are differentially expressed in specific tissues during stress response, such as TaSNAC8/10/12/13, which were up-regulated in leaves but down-regulated in roots ( Figure 2b).
In order to determine which regulatory elements exerted control over TaSNACs, we next analysed the DNA sequence of the promoter regions of the 39 TaSNAC genes to identify cis-acting regulatory elements involved in the drought stress response, such as ABA-responsive elements (ABRE), MYB-binding sites (MYBR) and dehydration-responsive elements (DRE) (Yamaguchi-Shinozaki and Shinozaki, 2005). We found that all of the TaSNAC genes have one or more of these cis elements (Table S2), suggesting stressactivated regulatory control of these TaSNAC genes, and likely a pivotal role in the transcriptional response to drought. Collectively, these data indicate that different TaSNAC genes exhibit variation in their expression patterns, dependent on developmental stage, tissue type and drought conditions, thus suggesting diverse roles in developmental regulation during drought stress.

Association analysis of natural variation in TaSNACs with wheat drought tolerance
Previous studies have reported that stress-responsive genes TaSNAC5-5A and TaSNAC13-2A participate in regulation of stress response to drought in wheat (Huang et al., 2015;Mao et al., 2012). To further investigate whether genetic variation in any of the TaSNACs was associated with phenotypic differences in drought tolerance among varieties, we conducted an association analysis for each of the TaSNAC genes. Recently, 630 517 high-quality wheat SNP markers (660K SNP array) were identified from RNA sequencing or genotyping by sequencing (GBS) of a wheat natural diversity panel consisting of a 110 varieties collected internationally (Rasheed et al., 2017). Using this wheat SNP array, a total of 623 529 high-quality SNP markers with minor allele frequency equal to or more than 0.05 were identified in a larger, separate wheat panel consisting of a worldwide collection of 700 varieties ( Figure S3a,b). These SNP markers were used to identify DNA polymorphisms in each of these 39 TaSNAC genes.
Given the estimated LD decay rate of more than 200 kb in wheat varieties ( Figure S3c), we selected SNP markers within a 400-kb interval centred on each TaSNAC gene. All of these TaSNACs were found to be polymorphic, with an average of 19 SNP markers detected for each TaSNAC (Table 1). TaSNAC8-6A was found to be the most polymorphic, with 57 SNPs identified in this natural diversity panel. The drought stress tolerance phenotype of each wheat variety was also investigated, and the survival rate (SR) of seedlings under severe drought conditions was scored ( Figure S3d). Then, to assess whether there were genetic variations within or around TaSNAC genes associated with specific drought tolerance traits, an association analysis was conducted using the genotypic SNP data from the 39 TaSNAC genes and the drought tolerance phenotypic data from the 700 varieties. Three statistical models, including the general linear model (GLM), GLM + PCA and a mixed linear model (MLM), were applied to identify significant genotypic and phenotypic associations. Compared to GLM and GLM + PCA model, our results clearly showed that the MLM model significantly reduced the false-positive associations ( Figure S4).
The analysis detected 10 SNP markers around TaSNAC8-6A significantly associated with drought tolerance in this wheat panel under the MLM model (P < 0.01), including the most significantly associated SNP marker out of the total 753 SNPs detected around all TaSNAC genes (P = 3.29 9 10 -4 ) (Table 1; Figure S5a). We next analysed the 10 predicted genes within the 200-kb interval centred on TaSNAC8-6A ( Figure S5b). We  TaSNAC8-6A is associated with drought tolerance in wheat seedlings 1081 ( Figure S5c). Thus, TaSNAC8-6A was revealed to be a potentially important candidate gene significantly associated with drought tolerance in this wheat natural variation panel. TaSNAC8-6A contributes to wheat  drought tolerance To accurately identify genetic variations, TaSNAC8-6A was resequenced in 109 randomly selected variation panel accessions. The sequenced region encompassed a 3310-bp genomic DNA fragment corresponding to the full-length TaSNAC8-6A locus, including its promoter, exons, introns and both the 5'-and 3'-UTRs. In total, 43 SNPs and eight insertion/deletions (InDels) were identified (Table S3). An MLM-based association analysis was applied between all SNPs/InDels and the previously assigned drought tolerance phenotypes for each accession. Two newly identified genetic polymorphisms in the TaSNAC8-6A promoter were highly associated with drought tolerance (P < 1.0 9 10 À3 ), and specifically, InDel-313 showed the strongest association (P = 1.67 9 10 À5 ). Additionally, one nonsynonymous SNP (SNP731; P = 6.92 9 10 À3 ) was identified in the coding region ( Figure 3a; Table S3). The pairwise LD of these SNP markers indicated that InDel-313 was in high LD with InDel-409, but in low LD with other DNA polymorphisms ( Figure 3b).
To further compare the effects of the TaSNAC8-6A In-313 and TaSNAC8-6A Del-313 alleles on drought tolerance in wheat seedlings, we constructed two bi-parental F 2 populations by crossing drought-tolerant variety Pubing202 with two drought-sensitive varieties, Yangmai13 and GLUYAS EARLY. In these crosses, both Yangmai13 and GLUYAS EARLY have the TaSNAC8-6A Del-313 alleles, while Pubing202 carries the TaSNAC8-6A In-313 allele ( Figure S6d). qRT-PCR expression analysis showed that TaSNAC8-6A expression was higher in Pubing202, regardless of water availability, compared to Yangmai13 and GLUYAS EARLY ( Figure S6e). To more quantitatively determine the phenotypic contributions of the TaSNAC8-6A Del-313 and TaSNAC8-6A In-313 alleles to drought tolerance, the two F 2 segregation populations were treated with drought, and the plant survival rates after stress and re-watering were calculated. In the population derived from Yangmai13 9 Pubing202, the survival rate of TaSNAC8-6A Del-313 homozygous plants was~36%, lower than the survival rate of TaSNAC8-6A In-313 homozygous plants (58%). Similarly, in the population derived from GLUYAS EARLY 9 Pubing202, the survival rate was~33% for plants with a homozygous TaSNAC8-6A Del-313 allele, also lower than the survival rate of TaSNAC8-6A In-313 homozygous plants (67%) ( Figure S6f). Collectively, these data further support the premise that variation in TaSNAC8-6A contributes to the drought tolerant phenotype. In particular, the strong performance of plants carrying TaSNAC8-6A In-313 in drought conditions suggests that this allele represents a promising genetic resource for improvement of drought tolerance in more susceptible, high-yield wheat lines.
TaSNAC8-6A In-313 allele increases drought-activated transcriptional response to ABA signalling Sequence analysis of TaSNAC8-6A InDel-313 revealed that the 3bp insertion/deletion (TAC) was located in an ABRE motif (TACGTG) (Figure 4a). To verify whether the location of InDel-313 in the promoter region affected transcriptional levels of TaSNAC8-6A, we TaSNAC8-6A is associated with drought tolerance in wheat seedlings 1083 inserted promoter fragments from both TaSNAC8-6A In-313 and TaSNAC8-6A Del-313 upstream of a GUS (b-glucuronidase) to compare their transcriptional activation of the reporter in the presence and absence of ABA (Figure 4a,b). We found that regardless of whether plants were exposed or not to ABA, GUS expression driven by the TaSNAC8-6A In-313 promoter (C2; tolerant) was much higher than that driven by the TaSNAC8-6A Del-313 promoter (C1; sensitive). GUS expression was comparable to that of the C1 fragment when driven by the truncated promoter variants TaSNAC8-6A In-313 (C4) and TaSNAC8-6A Del-313 (C3), which lacked InDel-313 and the ABRE motif altogether. Furthermore, GUS expression driven by a truncated TaSNAC8-6A In-313 promoter variant carrying only the In-313/ABRE motif region (C6 construct) was significantly higher than that driven by the truncated TaSNAC8-6A Del-313 promoter variant carrying only the Del-313/ ABRE motif (C5 construct). In addition, exposure to 10 µM ABA significantly increased GUS expression driven either by the full length (C2) and truncated (C6) TaSNAC8-6A In-313 promoter fragments compared to their expression in the absence of ABA, though the C6 variant was considerably lower than the full length In-313 C2 construct (Figure 4b). These results showed that InDel-313 is a causal genetic element for TaSNAC8-6A up-regulation, and the insertion at this indel site increased TaSNAC8-6A expression.
In plants, AREB/ABF TFs are reported to directly target a core ABRE motif present in target gene promoters (Fujita et al., 2013). In order to identify the candidate AREB/ABF proteins that bind the ABRE motif in the TaSNAC8-6A promoter, we searched the wheat genome database using Arabidopsis AREB/ABFs as queries. We thus identified nine ABFs in wheat and designated them as TaABF1, TaABF2 and TaABF3 followed by chromosome locations. Phylogenetic analysis showed that TaABF1 was closely related to Arabidopsis ABF1/3/4, while TaABF2 and TaABF3 were closely related to ABF2 ( Figure S7a). qRT-PCR analysis revealed that all TaABFs were predominantly expressed in seedling leaves and were up-regulated in response to drought in leaf tissue (Figure S7b,c). We then conducted yeast one-hybrid assays, which indicated that TaABF1, TaABF2 and TaABF3 are candidate TaSNAC8-6A ABRE-binding proteins (Figure 4c). Furthermore, yeast cells carrying either the TaSNAC8-6A Del-313 (C1) or truncated TaSNAC8-6A Del-313 , consisting only of Del-313 (C5), promoter fragments grew slowly and with less binding activity than those transformed with the TaSNAC8-6A In-313 promoter fragment (C2) or truncated TaSNAC8-6A In-313 , carrying only In-313 (C6). These differences increased with ABA concentration (Figure 4c). Taken together, these data clearly indicate that TaABFs can directly bind with stronger affinity to the ABRE motif in the TaSNAC8-6A In-313 promoter than to that of TaSNAC8-6A Del-313 alleles to increase the TaSNAC8-6A expression and therefore enhance wheat drought tolerance.
Overexpression of TaSNAC8-6A confers seedling drought tolerance To better understand TaSNAC8-6A function in drought tolerance, we generated transgenic Arabidopsis expressing TaSNAC8-6A coding sequence from wheat cv. Chinese Spring. Two independent lines exhibiting high TaSNAC8-6A expression were selected for characterization of their stress response to drought (Figure 5a). We compared drought tolerance between transgenic and vectortransformed (WT) plants and found that transgenic Arabidopsis displayed significantly more robust drought tolerance, without inducing dramatic morphological changes under normal growth conditions. When WT survival was~37%,~85% of transgenic Arabidopsis plants survived in parallel water-withholding experiments (Figure 5b). The transgenic plants also developed more lateral roots and greater root dry mass with no difference in primary root length (Figure 5c, -Ramirez et al., 2013;Olsen et al., 2005). Analysis of promoter regions or 5'-UTRs of auxin-related genes YUCCA2, PIN3, PIN4, ARF6 and ARF8 confirmed the presence of one or both NAC-binding sites, suggesting these genes can potentially serve as TaSNAC8-6A regulatory targets. Expression of these auxin-related genes was elevated in transgenic Arabidopsis compared to WT, which likely contributed to their enhanced root system development (Figure 5e).

d). Previous reports revealed that NAC TFs selectively bind to either T[TAG][GA]CGT[GA][TCA] [TAG] or [ACG][CA]GT[AG]N{5,6} [CT]AC[AG] consensus sequences (Matallana
In order to explore the effects of constitutively TaSNAC8-6A expression in wheat, we overexpressed this gene in the wheat cv. Comparison of the effects on GUS reporter gene expression of two promoter sequence variants (with or without ABRE motif). Error bars, SD; statistical significance was determined by a two-sided t-test, *P < 0.05, **P < 0.01. (c) Yeast-one hybrid assay of TaABFs binding to the TaSNAC8-6A promoter fragments. The bait vectors containing full-length (C1 and C2) or truncated (C5 and C6) promoters and the prey vector containing TaABFs were introduced into yeast strain Y187. Yeast cells were grown on SD-Leu media with or without ABA. Fielder background and subsequently selected three independent overexpression lines for further testing. Protein blot analyses demonstrated that TaSNAC8-6A was properly expressed, translated and functional in transgenic wheat (Figure 6a). We then compared drought tolerance through side-by-side planting of transgenic and WT siblings and consistently observed two-to threefold increases in survival rate for transgenic lines over WT under water-deficit conditions (Figure 6b,c). Root system comparisons between overexpression lines and WT recapitulated findings in Arabidopsis that primary root lengths were comparable between transgenic and WT, while lateral root number and root dry mass were substantially increased in transgenic wheat compared to WT (Figure 6d-g).
To determine whether TaSNAC8-6A overexpression affected yield in addition to improving drought tolerance, we preliminary evaluated production under well-watered and water-limited conditions in the greenhouse. Under well-watered conditions, phenotype of transgenic plants was almost identical to WT ( Figure S8a), and similar to observations of seedling roots, transgenic lines produced more lateral roots than WT ( Figure S8b). Furthermore, water use efficiency (WUE) in the transgenic lines was significantly elevated over that of WT under drought stress, but not significantly different under well-watered conditions ( Figure S8c). Moreover, yield-related traits of transgenic plants were also similar to that of WT ( Figure S8d). However, under water-limited conditions, the transgenic lines exhibited slightly later flowering, wider leaves and higher grain yield than WT ( Figure S8e). In the present study, yield-related traits were investigated mainly in the greenhouse. Whether TaSNAC8-6A can significantly contribute to wheat yield in the field requires further detailed field-based investigation. Collectively, these data suggest that increased TaSNAC8-6A expression may improve auxin-related lateral root development and WUE, which further contribute to the enhanced drought tolerant phenotype. TaSNAC8-6A is associated with drought tolerance in wheat seedlings 1085 TaSNAC8-6A overexpression induces global changes to gene expression particularly in auxin-and droughtresponsive signalling pathways To better understand the regulatory network through which TaSNAC8-6A mediates a strongly tolerant phenotypic response to drought, we compared the transcriptomes of WT and transgenic OE3 and OE5 lines under both well-watered and water-deficit conditions. Under well-watered conditions, a total of 1451 and 554 genes (adjusted P < 0.01, fold change > 2 or < 0.5) were up-regulated and down-regulated, respectively, between transgenic lines and WT ( Figure S9a,b; Table S4). Gene ontology (GO) analysis revealed up-regulated genes in 'chitin metabolism', 'secondary metabolism', 'response to abiotic stress', 'response to drought' and 'response to auxin' biological pathways. In contrast, down-regulated genes were found in 'homeostatic process', 'energy quenching' and 'response to oxidative stress' biological pathways ( Figure S9c). These results support the qRT-PCR relative expression data indicating that TaSNAC8-6A up-regulates stress-and auxinrelated genes.
Under drought stress, 1229 and 1031 genes (adjusted P < 0.01, fold change > 2 or < 0.5) were at least twofold up-and downregulated, respectively, in the transgenic lines compared to WT (Figure 7a,b; Table S5). GO analysis showed that up-regulated genes found in stimulus-related pathways, including 'response to drought', 'response to abscisic acid (ABA)' and 'response to osmotic stress'. Interestingly, 'auxin-activated signalling pathway' and 'auxin polar transport' biological pathways had genes up-regulated under drought conditions, whereas 'flavonoid biosynthetic', 'secondary metabolite', 'photosynthesis' and 'oxidation-reduction process' pathways contained down-regulated genes ( Figure 7c). Furthermore, more genes categorized as 'drought response', 'ABA response' and 'auxin response' were up-regulated in drought-stressed samples than in well-watered samples (Figure 7c; Table S6). We hypothesized that these transcriptomic changes likely enhance root development in transgenic wheat to improve WUE during drought by increasing the roots' capacity to absorb soil moisture.

Discussion
Regulatory variation in TaSNAC8-6A correlates to wheat drought tolerance Although many genes necessary for drought adaptation have been cloned (Hu and Xiong, 2014;Zhu, 2016), the contribution by natural sequence variation to phenotypic variations remains largely unknown. Previous research has shown that genetic changes in cis regulatory regions can drive advances in plant evolution and breeding, often through subtle phenotypic changes induced by modification of the timing, pattern, or level of gene expression (Meyer and Purugganan, 2013;Wittkopp and Kalay, 2011). Our analysis revealed that the promoter InDel-313 polymorphism is highly associated with TaSNAC8-6A expression. Specifically, varieties with TaSNAC8-6A Del-313 show lower expression of TaSNAC8-6A under moderate drought stress and tend to be drought sensitive, while varieties carrying TaSNAC8-6A In-313 show higher expression and exhibit higher drought tolerance ( Figure S6b). The main reason is that the TaSNAC8-6A Del-313 allele lacks an ABRE motif and thus cannot bind with TaABFs, further lowering the expression of TaSNAC8-6A (Figure 4). In addition, in the F 2 segregation populations, TaSNAC8-6A In-313 homozygous plants showed higher tolerance to drought stress, while plants carrying TaSNAC8-6A Del-313 showed increased susceptibility to drought stress ( Figure S6d-f). We therefore propose that the TaSNAC8-6A In-313 allele is favourable for adoption by breeding programmes seeking to improve drought tolerance in wheat germplasm.
TaSNAC8-6A plays a positive role in wheat drought tolerance NAC TFs involved in regulating drought tolerance are promising targets in breeding for drought tolerant varieties (Hu and Xiong, 2014;Nakashima et al., 2012). TaSNAC8-6A expression was positively correlated with seedling drought tolerance under moderate drought stress, with higher expression observed among drought-tolerant accessions across a panel of 40 wheat varieties ( Figure S6b). When the TaSNAC8-6A was overexpressed in Arabidopsis and wheat, the drought tolerance of transgenic plants was significantly improved (Figure 5a,b; Figure 6a-c). These data suggest a positive regulatory role for TaSNAC8-6A in adaptation to drought stress in wheat.
We also investigated the yield of transgenic TaSNAC8-6A overexpression wheat lines and found that grain yields of overexpression plants were slightly higher under water-limited conditions, while under well-watered conditions, the yields were almost identical compared to WT ( Figure S8d,e). Importantly, under well-watered conditions, both transgenic lines exhibited more extensive root systems than WT ( Figure S8b), which could plausibly contribute to the observed yield gain. In agreement with this finding, similar effects on yield have also been observed in transgenic rice overexpressing OsNAC5, OsNAC6 and OsNAC9 (Jeong et al., 2013;Lee et al., 2017;Redillas et al., 2012). However, when OsNAC5, OsNAC6 and OsNAC9 were overexpressed in rice driven by a root-specific (RCc3) promoter, the grain yields were significantly increased under both well-watered and water-limited conditions (Jeong et al., 2013;Lee et al., 2017;Redillas et al., 2012). Future investigation would incorporate tissue-specific promoters, such as RCc3, for optimization of TaSNAC8-6A overexpression for wheat production under field conditions.

TaSNAC8-6A is involved in the regulation of wheat root development
Environmental disturbances have been tightly correlated with changes in root architecture, and to adapt to stress, plants can modulate the growth of primary, lateral or adventitious roots as well as root hair length and distribution (Malamy, 2005). The plant growth hormone auxin has been established as an influence on development of primary root length, number of lateral roots and gravitropism (Himanen et al., 2002;Zhao, 2010). However, little is yet known about which genes in the auxin signalling pathway may be directly targeted by NAC transcription factors. TaSNAC8-6A overexpression in both Arabidopsis and wheat enhanced their drought tolerance and improved root system development ( Figure 5; Figure 6), which was also observed in transgenic Arabidopsis overexpressing AtNAC2 (He et al., 2005) and transgenic rice overexpressing OsNAC6 (Lee et al., 2017).
This study also revealed the presence of NAC binding sites in the promoters of the auxin-related genes YUCCA2, PIN3, PIN4, ARF6 and ARF8. Relative expression analysis showed that these genes were transcriptionally elevated in transgenic Arabidopsis plants, suggesting that they may be targets of TaSNAC8-6A (Figure 5e). While in TaSNAC8-6A transgenic wheat lines, root transcriptome analysis revealed significant up-regulation of several auxin transporter genes and the auxin polar transport pathway were highly enriched (Figure 7c; Figure S9c; Table S6), in agreement with the known mechanisms of auxin translocation (Lavenus et al., 2013;Matthes et al., 2019). Our data also revealed similar up-regulation of several wheat homologs of canonical auxin signal transduction genes, such as ARFs, EINs and Aux/IAA (Figure 7c; Figure S9c; Table S6). The increased expression of these auxin signalling and transport genes caused by TaSNAC8-6A overexpression suggests a likely mechanism for enhanced root development that benefits plants during droughtinduced stress.
Based on our results, we proposed a model of how InDel-313 of TaSNAC8-6A functions in drought response (Figure 8). Accordingly, drought stress activates TaSNAC8-6A expression through ABA signalling or other regulatory pathways. In the ABA signalling pathway, drought could induce TaABFs expression, which activates TaSNAC8-6A depending on the ABRE element in its promoter. The TaSNAC8-6A In-313 allele has an ABRE element, which can bind with TaABFs and further enhance the expression of TaSNAC8-6A. The overproduced TaSNAC8-6A more strongly promotes drought response and further causes improved drought tolerance of wheat varieties. Thus, the natural variation of InDel-313 most likely functions as a molecular switch in ABA signalling to regulate TaSNAC8-6A expression and further plant drought tolerance.

Expression analysis of TaSNAC genes
Seeds of wheat cv. Chinese Spring were surface-sterilized in 1.0 % (v/v) Topsin-M (Rotam Crop Sciences Ltd., Taiwan, China) for 10 min, washed in deionized water and germinated on wet filter paper (25°C; 3 days). Germinated seeds were transplanted to hydroponic nutrient solution. For drought treatment, the threeleaf-stage seedlings were dehydrated at 20°C, 40-60 % relative humidity. Leaves and roots were collected at 0, 1, 3, 6, 12 and 24 h after dehydration. Total RNA was isolated using TRIZOL reagent (Invitrogen, Carlsbad, CA), and subsequently, all samples were treated with DNase I (Takara, Dalian, China) and RNA concentration was determined by Nanodrop 1000 (Thermo Scientific, Shanghai, China). cDNA was prepared using M-MLV Reverse Transcriptase, and qRT-PCR used SYBR Premix Ex Taq kit (Takara) on a Step One System (Applied Biosystems, Shanghai, China). The 2 ÀΔΔCt method was used to calculate relative expression using three biological replicates per sample. PCR conditions consisted of 95°C for 10 min, 40 cycles at 95°C for 15 s and 60°C for 30 s.

Yeast one-hybrid transactivation activity assay
Thirteen TaSNAC genes were individually cloned into the pGBKT7 vector to evaluate protein transactivation activity in the AH109 yeast strain. All transformed cells were screened on synthetic defined (SD) medium plates without tryptophan (SD/-T). The cell concentration of yeast transformants was adjusted to an OD 600 of 0.1, and the yeast cells were drop-inoculated onto SD/-T plates, or SD without tryptophan and histidine (SD/-T-H), or without tryptophan, histidine and adenine (SD/-T-H-A). The plates were incubated at 30°C for 3-6 days before photographing.
Drought tolerance evaluation of the wheat association panel 700 wheat varieties were phenotyped for drought tolerance in the cultivation pool (9.0 9 1.0 9 0.2 m, length 9 width 9 depth) in which 0.6-tons soil and 0.16-tons vermiculite were uniformly mixed with 0.35-tons water. The pool was divided into 700 plots and twelve replicate plants from each variety were grown together in a plot. Watering was withheld after seeds were planted and seedlings were grown in a greenhouse (16-h light/8-h dark light regimen, 16/14°C and relative humidity of 40 %). Soil water content (SWC) was recorded beginning at the first symptoms of leaf-wilting. Irrigation was applied to recover the surviving plants when clear differences in wilting were observed. Three days after rehydration, the survival rate (SR) of each genotype was recorded. Seedlings with viable green leaves were rated as survivors. Four independent, replicated experiments were conducted.

Association analysis of TaSNAC genes
Association analysis was conducted on 700 wheat varieties with corresponding phenotypic drought tolerance data obtained in this study. Among 623 529 high-quality wheat SNP markers with minor allele frequency (MAF) ≥ 0.05, 753 SNPs were found in the 400-kb interval around all 39 TaSNAC genes. The GLM model, GLM model with principle components (GLM + PCA) and the MLM model were all tested for detection of SNPs significantly associated with drought tolerance using the TASSEL 5.0 program (Bradbury et al., 2007). TaSNAC8-6A-based association mapping was performed on 109 representative wheat varieties. The TaSNAC8-6A promoter (~2.0 kb), coding regions (including introns) and 5'-and 3'-UTR sequences were amplified and sequenced. These sequences were assembled using DNAMAN and aligned using MEGA version 7. Nucleotide polymorphisms, including SNPs and InDels, were identified (MAF ≥ 0.05) among these genotypes and their association with the SR and pairwise LD were calculated with TASSEL 5.0.

Arabidopsis transformation and drought tolerance assay
The TaSNAC8-6A coding region in wheat cv. Chinese Spring was amplified and inserted at HindIII and SmaI sites into the pGreen vector (Qin et al., 2008), driven by the CaMV35S promoter. The plasmid construct was transformed into Agrobacterium tumefaciens strain GV3101 containing the pSoup helper plasmid. Arabidopsis thaliana ecotype Col-0 was then transformed using the floral dip method. Seedlings (T 3 ) with single TaSNAC8-6A insertions were identified for further analysis. For drought tolerance assay, seven-day-old plants were transferred into pots containing 130g soil/pot. At 30 days after germination, wellwatered plants were stressed by withholding water. Drought treatment lasted for 14 days, followed by a 6-day recovery with watering prior to counting the surviving plants. Minimum 20 plants of each line were compared with WT, and four independent experiments were conducted. For analysis of lateral-root development, seedlings were cultured on MS medium for 7 days and then transferred to fresh MS. Images were taken after 16 days and root dry mass was obtained from 30-day-old soilgrown plants. After cleaning, total roots were dried at 70°C for 24 h prior to measuring total root mass. Expression of auxinrelated genes was determined by qRT-PCR from root samples of 16-day-old seedlings grown on MS.

Wheat transformation and drought tolerance assay
The TaSNAC8-6A CDS from wheat cv. Chinese Spring was inserted into the pCAMBIA3301 overexpression vector driven by an optimized wheat ubiquitin promoter. Agrobacterium tumefaciens-mediated transformation introduced the plasmid into immature embryos of wild-type bread wheat (cv. Fielder) following the protocol reported by Ishida (Ishida et al., 2015). Transgenic-positive and sibling transgenic-negative (WT) plants in each generation were screened by PCR. The relative expression of TaSNAC8-6A in transgenic plants was measured by qRT-PCR, using protocols for RNA collection, cDNA synthesis and qRT-PCR described above. Three independent T 3 lines, OE3, OE5 and OE6, were planted side by side with WT in 2:1 soil: vermiculite. Three-leaf soil-grown seedlings underwent drought treatment for~30 days, followed by a 3-day recovery period with irrigation prior to counting surviving plants. At least 48 plants from each line were compared in each test, and three independent experiments were conducted. Lateral root development was measured using twenty-day-old hydroponic seedlings. Root material was then dried at 70°C for 72 h to obtain dry mass.

Evaluations of growth and yield components
Growth and yield trait evaluation of three independent T 3 overexpression lines (OE3, OE5 and OE6) and sibling transgenicnegative (WT control) plants, under well-watered and milddrought conditions, were performed in plastic boxes (0.50 9 0.35 9 0.20 m, length 9 width 9 depth) in a 2:1 soil: vermiculite mixture. Each WT/transgenic line was grown in groups of four plants/row in two identical plastic boxes, one wellwatered and one with slowly increasing drought. Well-watered plants were regularly irrigated until maturity. In drought-subjected boxes, plants were regularly watered until the head sprouting period; then, irrigation was stopped. Plants showed mild wilting at the beginning of flowering. Soil water content was recorded every other day after the initiation of water withholding. Growth and yield characteristics for all lines and treatments were recorded, and statistical data were based on 8 seedlings per line. The experiment was repeated twice.

RNA-seq analysis of transgenic wheat
Twenty-day-old hydroponically cultivated Ubi:TaSNAC8-6A transgenic and WT plants were used for root transcriptome analysis. Each total RNA sample was isolated using TRIzol reagent (Invitrogen) according to the manufacturer's instructions and was evaluated for quality and quantity with an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo alto, CA) and an Agilent RNA 6000 Nano Kit (Agilent Technologies). Library preparation was performed with a TruSeq RNA Sample Preparation version 2 guide (Illumina). Sequencing on an Illumina HiSeq 2500 system generated an average of 673 million paired-end reads (2 9 101 nucleotides) per library. RNA-seq data were analysed as described previously (Ram ırez-Gonz alez et al., 2018). Significant differential expression was defined as an absolute fold change > 2 and FDR < 0.01. GO categories matched with significantly up-and down-regulated genes were identified using the GOseq R package (Young et al., 2010).

ChIP-qPCR analysis
Twenty-day-old hydroponically cultivated Ubi:TaSNAC8-6A transgenic wheat and WT plants were used for ChIP assays, performed according to previously described protocols (Bowler et al., 2004). Seedlings were cross-linked with 1 % formaldehyde under vacuum for 10 min and ground in liquid nitrogen prior to isolation of nuclei. DNA was sheared by sonication into 200-to 500-bp fragments. Chromatin complexes were incubated with anti-TaSNAC8-6A polyclonal antibody, and immune complexes were pulled down using Protein A beads. The precipitated DNA was purified and dissolved in water for further analysis. A commercial polyclonal antibody to GAPDH (Santa Cruz Biotechnology, Santa Cruz, CA) was used as a negative control. All primers used in this research are listed in Table S7.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.

Figure S1
Phylogenetic tree of stress responsive NAC proteins in wheat, rice, B. distachyon, maize, and Arabidopsis. Figure S2 Chromosome location and gene structure analysis of the TaSNAC genes. Figure S3 Phylogenetic tree, LD decay, and survival rates of the 700 wheat varieties. Figure S4 Quantile-quantile (Q-Q) plot for association of SNPs and drought tolerance. Figure S5 Identification and molecular characterization of drought tolerance candidate gene TaSNAC8-6A in wheat. Figure S6 Association analysis between InDel-313 variant-driven expression of TaSNAC8-6A and the drought tolerant phenotype in wheat. Figure S7 Identification and expression analysis of TaABF genes in wheat genome. Figure S8 Agronomic traits of Ubi:TaSNAC8-6A transgenic wheat plants. Figure S9 Transcriptomic analysis of Ubi:TaSNAC8-6A transgenic wheat under well-watered conditions. Table S1 The information of two hundred and sixty NAC super family genes and their corresponding protein sequences in wheat. Table S2 Organization of cis-acting regulatory elements in the promoters of 39 TaSNAC genes. Table S3 Variations in the TaSNAC8-6A genomic region and their association with wheat drought tolerance. Table S4 Significantly up-or down-regulated genes in Ubi: TaSNAC8-6A transgenic wheat grown under well-watered conditions. Table S5 Significantly up-or down-regulated genes in Ubi: TaSNAC8-6A transgenic wheat grown under drought conditions. Table S6 NACRS motifs in the promoters of up-regulated genes involved in ABA-response, drought stress, and auxin-response.