Thirty years of genetic progress combined with seed ethylene priming improved drought resistance of wheat through long-term drought stress memory

Knowledge of short-term physiological adaption of wheat to soil water decit is well understood, but little is known about seed ethylene priming effect on long-term drought stress memory of dryland winter wheat. The 42 leading and new cultivars released between 1992 to 2017 were subjected to drought (45% Field water-holding capacity), and well-watered (75% Field water-holding capacity) conditions aim to screen cultivars with contrasting drought resistance and grain yield. Seeds primed with ethylene were subjected to both PEG-8000 and pure water to uncover ethylene-induced stress memory at both physiological and organ levels. Results showed that the soil water decit (45%FC) that occurred at the tillering stage resulted in 3.2% to 67.4% yield loss for 42 cultivars, which was determined mainly by the decrease in the fertile spike. Seed ethylene priming maintained leaf water by reducing root volume and dry weight, which played a crucial role in drought avoidance. Seed ethylene priming decreased malondialdehyde content by regulating auxin and abscisic acid signaling, reactive oxygen species scavenging capability, and osmotic regulation, which plays a crucial role in drought tolerance. Seed ethylene priming improved drought tolerance of the wheat through metabolic modication of carbon metabolism, glutathione metabolism, and phenylpropanoid biosynthesis . The seed ethylene priming induced long-term stress memory that improved tillering capacity and reduced wheat spikelets abortion, which provided extra 0.3 t ha -1 of grain yield. These results suggested that seed ethylene priming allowed the recall of long-lasting stress defensive memory, increasing grain yield by both drought avoidance and drought tolerance. to the IAA, CTK, and GA 3 decreased by an average of 18.1%, 26.6%, and 29.4%, respectively, as compared to the well-watered treatments. The primed plants grown at the soil water decit condition have greater ABA and ethylene contents than the unprimed treatments. These results suggested that plants primed with ethylene increased drought tolerance by increasing IAA, CTK, and GA 3 contents in early responses. information procession module, forty-six DEGs were enriched in the plant hormone signal transduction pathway, including 12 down-regulated genes related to auxin signal transduction, one upregulated gene related to zeatin auxin signal transduction, and two down-regulated genes connected to the ABA signal transduction pathway. The acid receptor PYR/PYL family were down-regulated; 13 genes encoding protein phosphatase 2C were upregulated; 1 gene encoding serine/threonine-protein kinase SRK2 was upregulated; 2 genes encoding ethylene response transcription factor were upregulated; 8 genes related to jasmonic acid signal transduction pathway were upregulated; and 8 genes encoding salicylic acid were upregulated. It was found that 100 differentially expressed genes were enriched in the glutathione metabolic pathway, in which three genes encoding glutamate-cysteine ligase were upregulated, six genes encoding glutathione reductase (NADPH) were upregulated, three genes encoding glutathione peroxidase were upregulated, and 85 genes encoding glutathione S-transferase were upregulated. Drought stress also induced 105 DEGs enriched in the phenylpropanoid biosynthesis pathways and 81 DEGs enriched in the phenylalanine metabolism pathways. A Total of 902 transcription factors were identied response to drought stress, and the top 5 leading transcription factors were RLK-Pelle_DLSV, MYB, RLK-Pelle_WAK, RLK-Pelle_LRR-I-1, and C 2 H 2 .


Introduction
The monitoring of global climate changes indicated that extreme climate events occurred more frequently and in more severity (IPCC 2016). Globally, a series of studies published from 1980 to 2015 have shown that the drought leads to 21-40% yield reduction in wheat, and the severity and duration of the drought determine the extent of yield loss (Daryanto et al. 2017;Fahad et al. 2017). Though all wheat stages could be affected by drought stress, the tillering and grain-lling are more susceptible than other wheat-growing stages, and the soil water de cit that occurred at tillering will cause an unrecoverable reduction in ear-bearing tillers and grain yield (Farooq et al. 2014). The direct grain yield loss by soil water de cit is up to 48.9 million tons as the soil water de cit will affect the soil water and root relation at all levels from molecular, cellular metabolic events and organs to the whole plant (Muscolo et al. 2015). Preventing such high yield loss is particularly crucial to address the challenges of climate change.
Drought avoidance and drought tolerance are the dominant approaches for drought resistance conferred by crops than drought escape, and the drought resistance is associated with physiology and biochemical responses at molecular, cellular, tissue, organ, and whole plant levels Li et al. 2021). Plants adapt to soil water de cit at the whole plant level by generating a more in-depth root system, smaller leaves, and rolling of leaves for reducing leaf water loss (Liu et al. 2006). To survive under extreme soil water de cit when the plant cannot avoid it, the plants have evolved sophisticated mechanisms to alleviate organelle damage when they perceive the signal of environmental deterioration. At the cellular level, percentage of leaf water in soil water deficit treatments compared with well-watered condition ) using the following equation: , in which FW refers to fresh weight for those plants primed with Ethylene, TW refers to dry weight for those plants without priming, and the DM refers to the dry weight of oven-dried leaves at 75 o C till constant weight.

Endogenous hormones of root hairs
The root hairs were collected and then frozen in liquid nitrogen immediately and stored at −80°C until analysis. The collected root hairs (1g) were homogenized in 5ml of cold 80% aqueous methanol overnight at 4 o C and then centrifuged at 10,000 × g for 20 min at 4 o C. After the crude extract liquid was purified by passing a C18 Sep-Pak Catridge, a 400ul of extract solution was condensed using an N 2 and then re-dissolved in 400ul phosphate buffer (0.01 mol/L, pH 7.4) prepared for phytohormones analysis. The indole-3-acetic acid (IAA) content was determined as following Tian et al. (2008). The gibberellin (GA 1 and GA 3 ) content was carried out using a previously described method (Castillo and Martinez 1997). Extraction and assay of ABA were conducted as described previously (Lang et al. 2018).

Membrane lipids peroxidation, antioxidation enzymes, and osmotic adjustment substance assays
The oxidative damage to the membrane lipids was estimated by malondialdehyde (MDA), a major thiobarbituric acid, as an indicator for membrane peroxidation (Wei et al. 2016). The fully expanded leaves (0.2 g) were ground in 5 mL of 95% (v/v) trichloroacetic acid (TCA), and the homogenate was centrifuged at 15,000 × g for 10 min at 4 o C. An equal volume of supernatant and 0.6% (w/v) thiobarbituric acid was added to the TCA (20%), followed by heating the solution at 96 • C for 30min. The MDA concentration was measured by recording 450nm, 532nm, and 600 nm using the formula: C = 6.45× (OD 532 −OD 600 ) −0.56×OD 450, and the MDA content was expressed as mol g −1 FW.
The enzyme activities of SOD, POD, and CAT were extracted and assayed as per the previous protocol ) with minor modifications. The frozen leaf tissue (0.3 g) was ground in 5 mL of 0.1 M Tris-HCl buffer (pH=7.8) containing 1 mM DTT, 1 mM EDTA and 1% PVP. The homogenate was centrifuged at 20,000 × g for 20 min at 4 • C. The supernatant was collected for measuring enzyme activities. Superoxide dismutase (SOD, EC 1.15.1.1) activity was assayed by monitoring the inhibition of photochemical reduction of nitro blue tetrazolium. Catalase (CAT, EC 1.11.3.6) activity was assayed by potassium permanganate titration and expressed as M H 2 O 2 g −1 FW min −1 . Peroxidase (POD, EC 1.11.1.7) activity was estimated as the guaiacol oxidation rate by measuring the rate of change in absorbance at 436 nm.
The amount of free proline was extracted from fresh flag leaves of the wheat, and it was measured by using the ninhydrin reagent as following the method of Wang et al (2019) with minor modification. After the leaves were ground in 5 mL of 3% (w/v) sulfosalicylic acid, the homogenate was centrifuged at 15,000 × g for 10 min at 4 • C. The collected supernatant was re-extracted with 2ml toluene-20, and the absorbance was recorded at 520nm. The free proline was expressed as mmol proline g −1 FW.
The total water-soluble carbohydrates (WSC) were determinate by the anthrone method (Yang et al. 2017). In brief, 0.1 g dry leaf powder was extracted three times, with 80% ethanol, in a water bath at 80°C for 30 min. The non-structural carbohydrates were diluted to a final volume of 25 mL, and the non-structural carbohydrates were measured spectrophotometrically at 620 nm using an anthrone reagent.
2.3.5 Transcriptome profiling of wheat root hair cell RNA extraction, cDNA library construction, and RNA sequencing: In each plot, the total RNA from root hairs was extracted using (Illumina, San Diego, USA) according to the manufacturer instructions (Illumina, San Diego, CA). In short, the Poly-A-containing mRNA was purified twice from the total RNA using poly-T oligo-attached magnetic beads. First-strand cDNA was generated using reverse transcriptase and random primers. Following the second strand cDNA synthesis and adaptor ligation, 200-bp cDNA fragments were isolated using gel electrophoresis and amplified by 18 cycles. The products were loaded onto an Illumina HiSeq2000 instrument and subjected to 100 cycles of paired-end (2 × 100 bp) sequencing. The processing of fluorescent images into sequences, base-calling, and quality value calculations were performed using the Illumina data processing pipeline (version 1.8).
Gene ontology and pathway enrichment of DEGs: The high-quality reads were obtained by removing low-quality reads (Q < 20), and then mapped onto the international wheat genome sequencing consortium (IWGSC) using Bowtie2 (v2.2.9). The differential expression level was estimated using R software with edgeR package (R version: 4.0, edgeR version: 2.3.52). A strict algorithm was developed to identify differentially expressed genes (DEGs) between samples, and false positive and false negative errors are performed using Benjamini and Yekutieli hypergeometric test (FDR) method (Wei et al. 2020). Expression levels that exhibited an absolute log2 (FC) ≥ 1 and an absolute -log10 (P) ≥ 1.41 as the threshold to be significantly differentially expressed genes (DEGs). Two-fold up-or down-regulation compared to the corresponding control condition were taken only into consideration for further analysis. Gene Ontology (GO) enrichment was based on AgriGO software with a hypergeometric statistical test and Hochberg (FDR). Pathway enrichment analysis of DEGs was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg/). This analysis identifies enriched signals transduction pathways and metabolic pathways in DEGs compared with the whole genome background.
Validation of differentiated expressed genes by qRT-PCR: Total RNA for each sample was pretreated with DNase I, and the firststrand cDNA was generated using a RevertAid First Strand cDNA Synthesis kit (Fermentas, Vilnius, Lithuania) with oligo (dT) 20 primer following manufacturer instructions. The SYBR-based qRT-PCR reactions (SYBR Green I, Osaka, Japan) were performed in a 10 μl reaction volume on a Light Cycler 480 system (Roche, Basel, Switzerland) using the following reaction conditions: 95°C for 1 min followed by 40 cycles of 95°C for 10 s and 60°C for 15 s. The detection threshold cycle for each reaction was normalized against the housekeeping gene expression level with primer sequences 5'-GACCGTATGAGCAAGGAGAT-3' and 5'-CAATCGCTGGACCTGACTC-3'). All qRT-PCR reactions were replicated three times, and the results were analyzed based on the ΔΔCT method (Roche), where ΔΔCT represents ΔCT condition of interest -ΔCT control. The results obtained were transformed to a log 2 scale.

Statistical analysis
Two-way ANOVA was applied to analyze the difference between treatments performed using SAS version 9.  The soil water de cit (75% FC) resulted in 42 cultivars' grain yields decreased by between 3.2% and 67.4%, compared with those under well-watered conditions. The higher grain yield observed under well-watered conditions indicated a more signi cant yield loss rate by soil water de cit ( Fig 1A). The rst two principles explained the 26.7% and 15.3% of the yield variations, where cultivars with PC1 scores >0 identi ed as high-yielding cultivars; PC2 scores >0 identi ed as higher drought resistance cultivars. The PCA1 mainly relates to root length, chlorophyll content, and leaf area, while PCA2 mainly relates to root length, chlorophyll content, leaf water loss, and yield loss. The loading scores combined with the relative contribution to the sum of squares revealed that the yield loss attributed mainly to fertile spikes and panicle weight (Fig1, Table 1). Cultivars derived the negative relationship between grain yield and yield loss rate to identify drought avoidance traits. The droughttolerant group includes ve cultivars that showed both high yield and drought resistance. The drought sensitivity group consists of seven cultivars with a low yield and higher yield loss rate when the wheat is grown in 45% FC compared with those grown under well-watered conditions (75% FC). The drought sensitivity (CM-16), intermediate , and tolerance (CM) cultivars with similar genetic backgrounds were screening from PCA biplot ( Fig 1B). There was a dramatic decrease in plant biomass by soil water de cit (45% FC) for the drought-sensitive cultivar (CM-104), while the drought-tolerant cultivars (CM-16) did not show a signi cant difference in the plant biomass compared with those well-watered plots (75% FC). The yield components con rmed that the yield loss by soil water de cit could be attributed mainly to the decrease in fertile spikes and less by grains number per spike (Fig 1 C).

Seed ethylene priming effect on yield associated wheat phenotypes
Seed ethylene priming increased grain yield mainly by fertile spikes (P<0.05, Table 2), and there were no signi cant differences in plant height, grain number per panicle, and 1000-kernel weight (P>0.05). The seed ethylene priming decreased the degraded spikelets by 24.2% compared with unprimed plants, and there was no signi cant difference for fertile spikelets.
Seed ethylene priming increased tillering capacity by 21.5%, and the tiller survival rate increased by 14.7%, compared with those without seed ethylene priming. These results suggest that seed ethylene priming increased grain yield mainly by the fertile spike that was resulted from the increased tillering capacity and tiller survival rate.

Seed ethylene priming effect on root morphology traits and shoot biomass under soil water de cit
The soil water de cit resulted in root length, surface area, and root volume decreased by averages of 39.3%, 30.0%, and 22.5%, respectively, compared with those under well-watered conditions, whereas the root diagram increased 16.4% (Table 3).
As a result of decreased root surface area, the soil drought (45% FC) leads to the plant height, shoot dry weight, and leaf area per plant decreased by 31.4%, 37.6%, and 62.7%, respectively. In contrast, the seeds ethylene priming decreased leaf area reductions, shoot dry weight, and root dry weight, but not the plant height. Signi cant drought and priming interaction was observed in seedling heights and root to shoot ratio, but not the root surface area and shoot dry weight. These results suggested that seed ethylene priming improved plant resistance to soil water de cit by reduced root dry weight and LAI losses.

Comparative phytohormones of root hairs and signal transduction pathways
The unprimed plants grow on soil water de cit condition (45%FC) lead to the root hair IAA, CTK, and GA 1+3 contents decreased by an average of 8.5%, 17.9%, and 15.4%, respectively, as compared to the plant grown under well-watered condition (Fig 2). In contrast, the primed plant is grown at water de cit condition (45%FC) lead to the IAA, CTK, and GA 3 contents decreased by an average of 18.1%, 26.6%, and 29.4%, respectively, as compared to the well-watered treatments. The primed plants grown at the soil water de cit condition have greater ABA and ethylene contents than the unprimed treatments.
These results suggested that plants primed with ethylene increased drought tolerance by increasing IAA, CTK, and GA 3 contents in early responses.
The differently expressed genes related to drought stress response and ethylene mediated tolerance response revealed that the ABA and CTK signals were upregulated, including PP2C, SnRK2, B-ARR, which further regulate stomatal closure and cell division. In contrast, the IAA signal including AUX1, AUX/IAA, GH3, and SAUR increased under soil drought, which further suppressed plant growth. Ethylene mediated drought tolerance associated with plant hormone signal was associated with Ethylene, IAA, and ABA signal pathways because the jasmonate ZIM domain (JAZ), ethylene response factor 1 (ERF-1) were down-regulated, and SAUR were upregulated response to enhanced drought tolerance. The above ndings demonstrate that the ethylene-mediated drought tolerance relies on diverse signal pathways, including upregulated IAA signal transduction and the down-regulated ABA, JA, and Ethylene mediated stress responses.
3.4 Root hair antioxidative enzymes, osmotic adjustment substance, and leaf water loss ROS-scavenging enzymes' activities were further examined to clarify the underlying ethylene-mediated short-term physiological adaption to soil water de cit ( Table 4). As expected, drought stress (45%FC) leads to overproduction of the ROS, which caused lipid oxidation as shown in the MDA content, and there were signi cant increases in the activities of SOD, POD, and CAT in both primed and unprimed treatments as compared to the well-watered treatment (75%FC). Compared with unprimed control, the primed plants showed lower CAT activity and soluble sugar content, whereas increasing the activity of SOD. There was a ve-fold increase in both SOD activity and soluble sugars content between well-watered and soil drought stress treatments, whereas primed plants decreased soluble sugar by 12.7% compared with unprimed plants. The proline in primed plants increased by 3.2% compared with unprimed plants. The plants grown under soil water de cit conditions (45% FC) led to 7.6%~6.9% of leaf water loss and 1.2~2.0% of root water loss than those grown under well-watered conditions (75% FC). The seed ethylene priming did not affect root moisture but reduced leaf water loss and leaf water content compared with those unprimed plants. These results suggested that seed ethylene priming reduced leaf water loss under soil water de cit by modifying ROS scavenging enzymes and osmotic adjustment substances.

Global gene expression of root hairs by transcriptome sequencing
Both primed and unprimed plants were subjected to 8% PEG-8000 to simulate drought stress for 1h and 12h to evaluate the similarities and differences in both drought-induced and Ethylene-induced transcriptomes responsive effects. The experiment performed deep RNA sequencing of 1-week old wheat seedling leaves using the Illumina sequencing platform. After removing low-quality reads, the eight materials yielded 79.3 million to 148.5 million clean reads (see Methods, supplemental le 2). The high-quality reads accounted for more than 99.3% of the raw reads, and the Q30 for each sample was above 90.86%. Clean reads of each sample were aligned with the reference genome, and the alignment e ciency ranged from 61.94% to 65.20%. Fold Change (> 2) and FDR < 0.001 were used as screening criteria for differentially expressed genes (DEGs). The controlled and drought-stressed groups were separated by the rst principle component representing 56.7% of the total variation and the second principle representing 19.4% of the total variance (Fig3 A). Drought stress-induced 3 197 DEGs at 1 hour after subjected to drought stress compared with controlled, of which 1953 genes were upregulated, and 1244 genes were downregulated ( Fig 3B). Seeds Ethylene priming induced 2586 DEGs, of which 1698 genes were upregulated and 888 genes were down-regulated. Both droughts induced DEGs and Ethylene induced DEGs at 12h after treatments were signi cantly higher than that of at 1h subjected to drought stress, of which 1744 and 2272 DEGs were upregulated, and 2743 and 2711 DEGs were down-regulated. The Venn diagram plot showed that there were 642 and 1779 genes associated with drought tolerance in both 1h and 24h after subjected to drought stress, in which 730 unique genes were expressed after subjected to drought stress ( Fig 3C). All the DEGs were analyzed for time-scale variation in their expression and were classi ed into 9 clusters ( Fig   3D). The clustering analysis also allowed us to sort genes responses according to their overall behavior. This analysis has demonstrated that rapid gene response could be classi ed into up to 9 clusters. We identi ed 1479 genes that were consistently upregulated (cluster 1), and 1352 genes (cluster 9) were always down-regulated repose to drought stress. Seeds Ethylene priming induced 22 genes that consistently upregulated in root hairs. In contrast, the 19 genes that consistently down-regulated in root hairs. The chromosome-wise gene expression pattern revealed nearly the same expression pattern in 1h and 24 h subjected to drought stress. DEGs associated with drought stress were mainly distributed on Chr3 BL (5%), Chr5 BL (5%), Chr5 DL (4%), and Chr2 BL (4%). In contrast, DEGs associated with Ethylene mediated drought tolerance were mainly distributed on Chr4 DS (6%), Chr5 BL (6%), and Chr3 B (5%).

Go and KEGG analysis of the differently expressed genes of root hairs
Go enrichment analysis and were placed into three major categories: "biological process", "cellular components", and "molecular function" (Fig 4). The transcripts of E0P0 vs. E0P1 related to drought stress response in the biological process category mainly involved metabolite process, cellular process, single organism process, response to the stimulus. The drought stress-induced DEGs in the cellular components category mainly expressed in the cell part, organelle, and membrane. The E0P1vs E1P1 related to Ethylene mediated drought tolerance that associated with increased gene expression in the cell wall (9.67% vs. 11.76%), chloroplast (13.44% vs. 18.62%), and cytosol (8.60% vs. 10.78%), and decreased gene expression in Golgi apparatus (16.12% vs. 7.84%). The most enriched GO term under molecular function was catalytic activity, which was 21.02% and 17.89% in the E0P0 vs. E0P1 and E0P1 vs. E1P1 comparisons. The other GO terms under molecular function were binding, electron carrier activity, transporter activity, and nucleic acid binding transcription factor activity.
The pathway-based analysis was performed using the Kyoto Encyclopedia of Genes and genomes pathway database (Fig 5).
The results showed that the drought stress-induced 3197 DEGs (1h) and 4487 DEGs (12h) involves 20 metabolites pathway and 17 functional pathways (Q-value < 1). Most of the DEGs associated with drought stress-related to plant hormone signals, glucosinolate biosynthesis, taurine, and hypotaurine metabolism, zeatin glutathione metabolism, avone, and avonol biosynthesis. In the environmental information procession module, forty-six DEGs were enriched in the plant hormone signal transduction pathway, including 12 down-regulated genes related to auxin signal transduction, one upregulated gene related to zeatin auxin signal transduction, and two down-regulated genes connected to the ABA signal transduction pathway. The acid receptor PYR/PYL family were down-regulated; 13 genes encoding protein phosphatase 2C were upregulated; 1 gene encoding serine/threonine-protein kinase SRK2 was upregulated; 2 genes encoding ethylene response transcription factor were upregulated; 8 genes related to jasmonic acid signal transduction pathway were upregulated; and 8 genes encoding salicylic acid were upregulated. It was found that 100 differentially expressed genes were enriched in the glutathione metabolic pathway, in which three genes encoding glutamate-cysteine ligase were upregulated, six genes encoding glutathione reductase (NADPH) were upregulated, three genes encoding glutathione peroxidase were upregulated, and 85 genes encoding glutathione S-transferase were upregulated. Drought stress also induced 105 DEGs enriched in the phenylpropanoid biosynthesis pathways and 81 DEGs enriched in the phenylalanine metabolism pathways. A Total of 902 transcription factors were identi ed response to drought stress, and the top 5 leading transcription factors were RLK-Pelle_DLSV, MYB, RLK-Pelle_WAK, RLK-Pelle_LRR-I-1, and C 2 H 2 .
Carbon metabolism, carbon xation, and phenylpropanoid biosynthesis were the top 3 crucial pathways associated with Ethylene mediated drought tolerance. The DEGs encode ribulose-bisphosphate carboxylase, glyceraldehyde-3-phosphate dehydrogenase (NADP+) glyceraldehyde-3-phosphate dehydrogenase were upregulated response to Ethylene mediated drought tolerance. The DEGs encoding peroxidase enriched in phenylpropanoid biosynthesis were a down-regulated response to Ethylene mediated drought tolerance. A total of 42 transcription factors were identi ed, which were associated with ethylene-mediated drought tolerance response, and the 3 top transcription factors were the RLK-Pelle_DLSV, NAC, and AP2/ERF-EPF.

Validation of DEGs using qRT-PCR
The qRT-PCR validation of the DEGs was performed using cDNA isolated from treatments. The selection of high up-and down-regulated genes was based primarily on regulating various pathways related to drought stress and ethylene mediated drought tolerance (Fig 6). Because the number of down-regulated genes was greater than that of upregulated genes in 12h after subjected to soil drought, the experiment chose more downregulated genes for validation through qRT-PCR. The expression level of each selected DEG for both priming and drought stress treatment compared to the expression level indicated by its transcriptome sequencing data. The differential gene expression detected by qRT-PCR was consistent with the up-and down-regulation trends of gene expression obtained from transcriptome sequencing (R 2 =0.89; n=24). The results were indicating that the analytical results of transcriptome sequencing were reliable.
4. Discussion 4.1 Genetic modi cation of drought avoidance traits combined with seed ethylene priming provided an extra 0.3t hm -2 yield The natural and arti cial selection of cultivars with high yield and drought resistance processes numerous physiological and molecular mechanisms adapt to drought stress while realizing the potential yield (Liu and Qin 2021). Identifying and utilizing traits associated with drought tolerance and drought avoidance is the most effective measure for reducing leaf water loss and stress injury (Kamphorst et al. 2019;Pouri et al. 2019;Katuwal et al. 2020;Liu and Qin 2021). Notably, the higher the grain yield under well-watered conditions (75 % FC), the more signi cant yield loss by soil water de cit (45% FC). It is possible that cultivars with increasing drought resistance indicating higher metabolic cost while the yield potential realization. The intrinsic properties of 42 leading and new cultivars acclimate to soil water de cit provided a prospect for drought tolerance improvement while yield realization. There is no doubt that wheat cultivars selected from well-watered conditions have little contribution to improving drought tolerance even through genetic gain in different wheat production regions. The breakthrough cultivars released in recent years have both yield and drought resistance provided insights for future improvements. The yield loss by soil water de cit was determined mainly by the decrease in fertile spike, and the drought persistence primarily relates to root length, root volume, leaf chlorophyll content, and leaf water loss rate, which plays a signi cant role in drought avoidance. Plants that avoid drought stress rely on maximizing root surface area, minimizing leaf area, and decreasing leaf water loss (Fang et al. 2017;Sadok and Schoppach 2019;Pace et al. 2015). High leaf water content has been associated with higher leaf photosynthesis levels, cell membrane stability, ROS-savaging enzymes, and grain yield (Wang et al. 2021;Larkunthod et al. 2018). Although deeper root exploration may enhance water absorption in droughtstressed conditions, the metabolic cost cannot compensate for the loss of aboveground plant growth. Similarly, small leaf and prolonged stomatal closure increased the capability of drought avoidance, but in the long run, it may lead to a signi cant reduction in grain yield (Li et al. 2021). Therefore, a further increase in root length is not a wise approach for both grain yield and drought resistance because it will cause a decrease in carbon economy at the whole-plant scale. A novel nding from the present study is that the seed Ethylene priming enhanced both root volume and shoot dry weight without further metabolic cost on the deeper root. The drought resistance conferred by crops with increased root volume that improved tillering capability and spikelet survival rate and provided extra 0.3 t hm -2 grain yield under soil water de cit conditions (45% FC).

Seed ethylene priming recalling long-lasting stress memory through an altering signaling pathway
Drought tolerance involves mechanisms operating at temporal scales, from short-term physiological feedback process, plants maintain leaf water, and long-term feedback strategies have participated in the trade-off between carbon accumulation and risks of deleterious soil water depletion. For most short-term physiological processes, the short-term (1h, 12h) response is involved in changes in stomatal control, hydraulic tissue conductance, osmotic adjustment, root-to-leaf growth response optimization of carbon metabolism (Tardieu et al. 2018). As expected, the ethylene-responsive factors 1 and 2 were downregulated at both 1h and 12h subject to drought stress, and the EIN 3 was upregulated compared with those grown under well-watered conditions. These results can be attributed to the ethylene-responsive factor (ERF) function involved in diverse plant reactions to biotic or abiotic stresses (Seo et al. 2010). ETR, EIN2, and EIN3 in soybean (Arraes et al. 2015), PYL in cotton (Liang et al. 2017), and JAR1 and MYC2 in soybean (Takagi et al. 2016) were reported to be positively correlated with drought tolerance. Seed ethylene priming allowed the recall of long-lasting stress defensive memory, which decreased the Ethylene perception signal's gene expression. The Ethylene perception signal and root to shoot ABA secondary signal involve a dialogue between ABA, IAA, and SA, and the ABA is the dominant hormone regarding its role in plant drought tolerance (Wang et al. 2017b;Zhu et al. 2019;Harb et al. 2020). The production of ABA in the root hair and then mobilized to the leaves will maintain plant water status through stromal regulation , and then postpone the development of water de cit in the aboveground (Liu et al. 2005). PP2C, PYL, and SnRK2 are the three critical components of the ABA signaling pathways; PP2C acts as a negative regulator, while SnRK2 acts as a positive regulator (Khan et al. 2019;Liang et al. 2018). In the present study, the decreased expression of ABA receptor genes PYR/PYL combined with increased expression of PPC2 and SnRK2 is crucial for maintaining leaf water under drought-stressed conditions due to their function in stomal regulation. Notably, it was reported that salicylic acid (SA) production also affects stomatal opening, photorespiration, and antioxidant defenses before any detectable change in the relative water content (Sanchez-Martin et al. 2015). The Ethylene mediated drought stress memory was associated with increased SA signal transduction gene TGA expression in 1h after subjected to drought stress but not 12h, which con rmed the crucial role of SA in the early stage of drought response. As a result of decreased ABA and Ethylene contents treated with seed Ethylene priming, the IAA responsive gene SAUR was signi cantly increased, whereas the Jasmonate ZIM-domain and ERF1/2 were signi cantly decreased. The identi ed phytohormone signals cross-talk opened promising avenues to maintain leaf water under soil water de cit.

Transcription factor and phytohormones co-mediated for ethylene mediated drought tolerance
The transcription factor played a central role in regulating gene responses to abiotic stress in plants (Golldack et al. 2014). The MYB, C2H2, NAC, AP2/ERF-ERF, and bHLH were identi ed and coupled with the soil water de cit in this study. These identi ed transcription factor families are well known in plants' stress tolerance (Zhang et al. 2015;Zhu et al. 2019 (Huang et al. 2015;Wu et al. 2016). NAC-coding genes' overexpression could improve drought tolerance and be more effective when under the control of stress-responsive or tissue-speci c promoters (Nakashima et al. 2014;Brasileiro et al. 2015). The transcription factor ERF has been reported to enhance osmotic tolerance by modulating the increase in stress-responsive gene expression, including MYB transcription factor, ABA synthesis gene, and photosynthesis-related genes (2010), thereby regulating shoot branching, lateral root, leaf senescence even the leaf wax accumulation (Nie et al. 2018;Ranjan et al. 2012). The bHLH transcription factor reported enhanced drought tolerance not only by regulating stomatal density, transpiration, photosynthesis, and avonoid biosynthesis (Zhai et al. 2016;Dong et al. 2014;Wang et al. 2016), but also associating with ABA biosynthesis, JA module component, and root development (Seo et al. 2011;Li et al. 2019). The loss of function of the SWI2/SNF2 chromatin remodeling ATPase BRAHMA (BRM) causes ABA hypersensitivity during post-germination growth arrest, demonstrating the crucial role of SNF2 on drought sensitivity of plants (Han et al. 2012). In rice plants, ALT1, an Snf2 family chromatin remodeling ATPase, reported alkaline tolerance negatively through enhanced defense against oxidative stress (Guo et al. 2014). The WRKY transcription factor's biological function in abiotic stress has been reported enhancing drought tolerance by ABA synthesis, AsA accumulation, and MAPK signal (Rushton et al. 2010;Ren et al. 2010). GmWRKY13, GmWRKY21, and GmWRKY54 have been reported to play crucial roles in regulating both drought and salt tolerance of soybean (Zhou et al. 2008). In barley, the overexpression of the C 2 H 2 type zinc nger member TaZFP1B resulted in an increase in drought tolerance (Cheuk et al. 2020), which can be mediated via a phytosulfokine receptor since overexpression AtPSKR1 improves growth in Arabidopsis.

Antioxidant enzyme and osmotic adjustment co-mediated ROS savaging involved in Ethylene mediated drought tolerance
Drought signals enable plants to slow down leaf water loss, stimulate the production of osmosis-regulating metabolites, activates the internal drought defense system to prevent the lipid oxidation of cell membrane (Singh et al. 2017;Gupta et al. 2020). The MDA content has been suggested as a marker of lipid peroxidation of the cell membrane in response to abiotic stress (Moller et al. 2007). The seed Ethylene priming reduced root hair MDA content, indicating the cell membrane's stressinduced injury was alleviated compared to the cultivars without seed ethylene priming. Lower MDA content can be mainly explained by increasing ROS savaging enzymes, osmotic adjustment, and metabolic alternation (Wang et al. 2017a;Hameed et al. 2011;Simova-Stoilova et al. 2009). The seed Ethylene priming increased SOD and CAT activities but not POD activities under water de cit. This is clear that seed Ethylene priming reduced oxidative stress by increasing ROS scavenging enzymes triggered to detoxify cells. The differently expressed genes associated with glutathione metabolism, phenylpropanoid biosynthesis, and avonoid biosynthesis con rmed the Ethylene mediated drought tolerance associated with the internal drought defense system. The osmolytes included the soluble sugar and proline contents, which have been reported to stabilize membranes and maintain protein structure at stressed conditions, which appears to be consistent with yield and yield stability (Singh et al. 2017;Blum 2017). A novel nding from the present study is that the seed Ethylene priming increased Proline content but signi cantly decreased soluble sugar content. This helps reduce the consumption of carbohydrates assimilates while maintaining osmotic balance, thereby partitioning more carbohydrates for leaf growth that would increase aboveground plant growth. Previous results reported that plant growth and photosynthesis remain affected by water de cit even though leaf cell turgor is maintained via osmotic adjustment (An-Ching and Boyer ;Bouchabk et al. 2006). Seed Ethylene priming increased root hair Proline level, but not soluble sugar content explains why the root length did not increase signi cantly in seed ethylene priming treatment. Seed ethylene priming increases root volume that enables more N to accumulate in the leaves, thereby affecting light capture and biomass accumulation. Seed Ethylene priming increased root hair proline level that enhances drought tolerance by maintaining the osmotic balance of cells and ultimately helps plant recovery from drought stress. These short-term physiological changes also have a profound long-term consequence for plants' carbon budget at plant scale. The long-lasting stress memory reduced plant biomass loss by maintaining the leaf area and chlorophyll content, which further increased fertile tillers and grain number, thereby reduced yield loss under drought stress.

DEGs involved in metabolites pathways and their expression pattern in response to drought stress
The root is the rst to encounter low soil moisture in the soil, and the transcriptome pro ling of root hairs provide a good target for evaluating genes and key process underling plant adaption and tolerance to soil water de cit. Some antioxidants and osmotic regulation substances such as ascorbic acid (AsA), glutathione, glutathione, and avonoid can be formed to effectively eliminate ROS and improve drought defense. Glutathione transferases are involved in protection under various stress conditions by detoxifying toxins, increasing oxidative stress (Singh et al. 2017). Glutathione S-transferases play a signi cant role under drought conditions by conjugating glutathione with electrophilic substrates to increase their solubility and facilitate further metabolic processing (Yilmaz and Iscan 2014). Expression levels of glutathione S-transferase and glutathione peroxidase activity under drought in ag leaves were found correlated with high yield stability (Galle et al. 2009). Genes encoding glutathione-disul de reductase and glutathione peroxidase were upregulated under drought stress. In contrast, genes encoding glutathione transferase were down-regulated for both 1h and 12h subjected to soil drought, indicating that upregulated gene expression involved in glutathione biosynthesis was responsible for drought tolerance in plants. The antioxidant enzymes constitute the " rst line of defense" against ROS-induced cell membrane peroxidation during drought stress, and the phenylpropanoids constitute a secondary antioxidant system. The phenylpropanoid biosynthesis was activated due to the depletion of primary antioxidant defenses and control cellular H 2 O 2 within a sublethal concentration range (Sun et al. 2011) explained genes expression involved in phenylalanine ammonia-lyase, cinnamoyl-CoA reductase, and trans-cinnamate 4-monooxygenase were increased under drought. The catalase-peroxidase and cinnamyl-alcohol dehydrogenase were signi cantly decreased, which explains the decreased plant growth under drought stress. Both these ndings indicate a potential role in drought stress responses of dryland winter wheat.
Seed ethylene priming induces physiological changes that affect the metabolism through the up-or down-regulation of genes involved in the phytohormone signal transduction, phenylpropanoid biosynthesis, carbon xation, carbon metabolism, galactose metabolism. Seed ethylene priming upregulated genes related with caffeate O-methyltransferase and peroxidase, which must have played a signi cant role in the seed Ethylene mediated drought stress memory. The catalase and nonspeci c peroxidase activities can directly inhibit the production of H 2 O 2 (Katuwal et al. 2020;Abid et al. 2018). It was speculated that seed Ethylene priming improved drought resistance related to its ability to scavenge ROS. The upregulated DEGs related to peroxidase prepared the plant response more quickly and actively to drought. It was reported that drought stress resulted in the reduction of leaf elongation by less lignin biosynthesis, which was highly corrected with the accumulation of caffeate O-methyltransferase (Vincent et al. 2005 . Therefore, seed ethylene priming increased root volume and leaf area resulted from the increased gene expression mainly involved in phenylpropanoids biosynthesis.

Prospect for future improvement
The short-term physiological changes do not necessarily indicate they affect the nal yield (Sadras and Richards). The seed Ethylene priming avoids yield loss of 363 kg hm -2 under 45% eld water capacity but does not necessarily indicate they can avoid yield loss under more severe drought stress. In this investigation, the extent of yield loss by seed Ethylene priming depended on soil water availability. A signi cant contribution from the present study is to deepen the current knowledge on ethylene-mediated drought tolerance memory. The root-level determinations of hormone signals were made based on root hairs, while the increased water retention capability was determined on leaf levels. It is speculated that the Ethylene mediated drought tolerance was mediated by root-to-leaf cross talk. Secondly, the drought stress memory is always accompanied by wheat growth, and as approved by the present study, the nal fertile spikes were signi cantly reduced by ethylene priming.
Previous results indicated that the drought stress memory was inherited to the next generation by epigenetic modi cation. However, how many generations can remember this precious memory remains unknown. Thirdly, seed Ethylene priming activates genes associated with Eth sensing signals and triggers multi-hormone signals, indicating a cross-tolerance to multiple stresses condition. It is crucial to reduce the adverse effects of global climate change on crop yield. Over the past decades, researchers have made great efforts to elucidate the complex genetic architectures and a regulatory mechanism involving drought tolerance in an increasingly changing variable environment. The identi ed agronomy traits and genes that exert their effects on drought tolerance capability remain mostly unknown. It is of great importance to integrate multi-omics into association mapping to bridge the knowledge gap. The integration of multi-omics data will provide integrative information that increases the current knowledge of drought stress memory and multi-stress adaption. Undoubtedly, the drought tolerance improvement by using genetic approaches or seed priming, the understanding of the complex regulatory network associated with drought tolerance will help ne-tune the adaptability of wheat to withstand the challenges in an increasing changeable environment. In any case, further studies are required to shed light on (1) the exact mechanisms controlling root-to-leaf signals transduction; (2) the exact generations of the existence of the drought stress memory; (3) multi hormone signals commination adapt to cross stresses tolerance.

Conclusions
The results obtained from a combination of controlled lab experiments with observations under eld drought-stressed conditions revealed that seed ethylene priming helped alleviate drought stress-induced growth inhibition by enhancing both short-term drought adaptation and long-term drought memory. Seed ethylene priming maintained leaf water by increasing root hair antioxidant enzymes and osmolytes regulation, resulting in lower membrane lipid damage than non-primed seedlings. The increased proline content and reduced soluble sugar content provided an economical approach for enhancing osmolytes regulation while realizing potential yield. The maintaining of root hair homeostasis endogenous IAA and ABA levels participates in recalling drought stress memory by affecting carbon metabolism, carbon xation, and phenylpropanoid biosynthesis. Due to these bene cial effects, the long-lasting stress memory increased tillering capacity and orets e ciency, which provided an additional 0.3t yield in the dryland farming system. Future studies on Ethylene priming-induced epigenetic modi cations, transgenerational stress memory, and cross stress memory are still required to fully elucidate the role of Ethylene priming on long-term stress memory in an increasingly changing environment.   Note: P, S, and FC represent the seed priming, soil drought, and field water holding capacity. Different letters indicate significance at the 5% probability level. Note: P, S, and FC represent the seed priming, soil drought, field water holding capacity. Different letters indicate significance at the 5% probability level.  The changes in root hair hormone contents (A) and hormone signal transduction pathways (B) of seedlings primed with ethylene grown under well-watered (75%FC) and drought stress (45%FC) conditions.

Figure 5
Top pathway entries of the DEGs related to drought response (A), and seed ethylene (Eth) priming mediated drought response (B).