The study was conducted between 2022 and 2023 at the experimental base of the Sichuan Academy of Agricultural Sciences (104°39′E, 31°00′N) located in Guanghan City, Sichuan Province, China. Two wheat cultivars, Chuanmai 104 (CM104, cold-tolerant) and Chuanmai 42 (CM42, cold-sensitive), which exhibit similar growth but differ in cold tolerance at the booting stage were sown in 24 x 21 cm pots on October 30, 2022 and harvested on May 5, 2023. The seeds of both cultivars used in the experiments were kindly provided by the Crop Research Institute of Sichuan Academy of Agricultural Sciences, China. CM42 was bred by using first-generation synthetic hexaploid wheat (Syn 769, Decoy 1/Aegilops tauschii 188) and CM104 was derived from CM42 (Liu et al., 2021). The soil used was composed of pH 7.46, organic matter of 31.76 g kg^-1, total nitrogen (N) of 1.98 g kg^-1, available N of 118.80 mg kg^-1, total phosphorus (P) of 1.64 g kg^-1, available P of 68.05 mg kg^-1, total potassium (K) of 16.68 g kg^-1 and available K of 157.75 mg kg^-1. Exactly 7 kg of soil was mixed with 4.1 g of NPK compound fertilizer in a ratio of 15:15:15 in each pot and used to sow 24 seeds per pot, with 16 seedlings retained post-emergence. The seedlings were then placed in the natural field until the plants reached the booting stage. The relative water content of the soil was maintained at 60-70% throughout the whole growth period. At the booting stage, the stems that grew uniformly were marked. All pots, except for the control treatments were then transferred to a smart climate chamber (Nanjing Hengyu Instruments and Equipments Co., Ltd, Nanjing, China) with 70% relative humidity and white light conditions (500 μmol m⁻² s⁻¹ during the day and 0 μmol m⁻² s⁻¹ at night). There were 100 pots for each cultivar, with one control (CK) and three cold stress treatments. The first treatment (T1) consisted of plants kept at 4 °C for 16h in light (L) and at -1 °C in darkness (D) once, while the second (T2) and third treatments (T3) consisted of plants maintained at the same conditions for 3 and 5 days, respectively. Spike tissue samples from 24 individual stems per treatment were pooled in four replicates after the cold treatment, immediately frozen in liquid nitrogen, and stored at −80°C for subsequent physiological and molecular analyses. Of which, molecular analyses (transcriptome, proteome and metabolome) were only carried out in the T3 treatment due to the more obvious phenotypic differences between the two varieties under this treatment. The remaining plants were transferred to the natural field environment to assess pollen viability at anthesis and seed setting rate at the maturity stage. Meteorological data (air temperatures and precipitation) during the field trial period were collected from a meteorological station adjacent to the field experiment site ( Supplementary Figure 1 ).

Pollen viability at anthesis was assessed using the 2,3,5-triphenyl tetrazolium chloride (TTC) staining method (Fábián et al., 2019). Fresh pollen was collected from the base and top of the spikes of different treatments and incubated in a 1.5 ml centrifuge tube with 0.1% TTC for 30 minutes at 37°C. The stained pollen was examined for viable and non-viable pollen grains under an Olympus BX51 light microscope (Olympus, Tokyo, Japan) at room temperature. The pollen abortion rate (%) was calculated by dividing the number of colorless pollen grains by the total pollen grains counted x 100%. A minimum of 1,000 pollen grains were studied for different varieties and treatments in triplicates.

At the maturity stage, 60 marked stems were harvested in triplicates for each genotype and treatment and used to record the fertility of superior florets per spikelet, the number of degenerate and fertile spikelets per spike, and the thousand-grain weight. The superior florets were defined as the two florets proximal to the rachis per spikelet, which usually carry seeds (Zheng et al., 2016). The seed setting rate of superior florets per spike (%) was calculated by dividing the seed number of superior florets per spike by two times the number of fertile spikelets per spike x 100%.

The contents of soluble sugar and protein, proline, ABA and gibberellin (GA₃), and the activities of catalase (CAT, EC 1.11.1.6), superoxide dismutase (SOD, EC 1.15.1.1), peroxidase (POD, EC 1.11.1.7) in young spikes under different cold treatments were measured using different assay kits from Suzhou Comin Biotechnology Co. Ltd. (Suzhou, China).

Total RNA was extracted from the young spike tissues of the control and T3 cold-treated CM104 and CM42 genotypes in triplicates using the RNeasy Plant Mini Kit (Qiagen Inc, Germany) and their quantities and qualities were evaluated using NanoDrop 2000 (Thermo Fisher Scientific, USA). Sequencing libraries were generated using the NEBNext Ultra™ RNA Library Prep Kit for Illumina (NEB, USA) and sequenced on an Illumina HiSeq 2500 platform (Beijing Biomarker Technologies Co., LTD., China). The in-house Perl scripts were used to remove the adapter, ploy-N and low-quality reads. These clean reads were then aligned to the wheat reference genome IWGSC RefSeq v2.1 (https://urgi.versailles.inra.fr/download/iwgsc/IWGSC_RefSeq_Assemblies/v2.1/) using HISAT2 v2.2.1 (http://daehwankimlab.github.io/hisat2/). Bowtie2 v2.5.0 (https://bowtie-bio.sourceforge.net/bowtie2/) and StringTie v2.2.0 (https://ccb.jhu.edu/software/stringtie/) were used to align the clean reads and calculate gene expression profiles, respectively. Gene function was annotated based on Nr (NCBI non-redundant protein sequences), Nt (NCBI non-redundant nucleotide sequences), KO (KEGG Ortholog database), GO (Gene Ontology), Pfam (Protein family), Swiss-Prot (A manually annotated and reviewed protein sequence database), and KOG/COG (Clusters of Orthologous Groups of proteins). Gene expression levels were quantified in FPKM (fragments per kilobase of transcript per million mapped fragments). Differentially expressed genes (DEGs) analysis was conducted using the DESeq2 v1.26.0 R package (https://github.com/thelovelab/DESeq2). The DEGs were identified using the threshold of |log2FoldChange| ≥ 1 and false discovery rate (FDR) < 0.01. Enrichment analysis of the DEGs for GO and KEGG pathways was conducted using GOseq v1.26.0 R package (https://github.com/lmika/goseq) and KOBAS v2.0 (http://kobas.cbi.pku.edu.cn/), respectively.

To validate the transcriptome data, 16 DEGs associated with cold stress were randomly selected for verification by quantitative real-time polymerase chain reaction (qRT-PCR) analysis in CM104 and CM42. The detailed primer sequences used are listed in Supplementary Table 1 . The wheat actin gene was used as an internal reference gene. The qRT-PCR was conducted on an ABI7500 Real-Time PCR System (Applied Biosystems, USA) with a total of 20 μL volume, containing 10 μL SYBR Green PCR-mix (DF Biotech., China), 2.5 μL of cDNA, 5.5 μL ddH₂O, and 1 μL of each primer. The PCR reactions were conducted at 95°C for 3 min, followed by 40 cycles of 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s. The relative gene expression levels of the DEGs were determined by the 2^^-△△Ct method (Livak and Schmittgen, 2001) in triplicates.

To further investigate the protein expression of cold-tolerant variety under cold stress, protein was extracted from young spikes of the CK and T3 cold-treated CM104 wheat in triplicates using the TCA-acetone precipitation method (Jorrin-Novo, 2014). Total protein concentration was measured by the Bradford Protein Assay Kit and the proteome was analysed using the tandem mass tags technology from Biomarker Tech. Co. (Beijing, China), as described by McLoughlin et al. (2018). The MS/MS raw data were analyzed using the MASCOT engine v2.2 (Matrix Science, UK) which is integrated into Proteome Discoverer 1.4 (Thermo Fisher Scientific, USA). Searches utilized a peptide mass tolerance of 15 ppm and a production tolerance of 0.02 Da, achieving a 5% FDR. Proteins were annotated using the KEGG and COG databases for functional annotation and enrichment analyses, respectively. Differentially expressed proteins (DEPs) were identified using the Student’s t-test (P < 0.05) and a fold-change threshold of >1.5 or <0.67.

To further evaluate the metabolite accumulation of cold-tolerant variety under cold stress, metabolites were extracted from spike samples of CK and T3 cold-treated CM104 in triplicates and processed for metabolomics analysis using the LC-MS system as described previously (Liu et al., 2023b). Raw data was collected using MassLynx V4.2. The data processing operations were processed by the Progenesis QI software based on the METLIN database and Biomark’s self-built library for identification. The classification and pathway information of identified compounds were searched in KEGG, HMDB and lipidmaps databases. The significant difference in each compound was determined by Student’s t-test. Orthogonal partial least squares discriminant analysis (OPLS-DA) modelled by the R language package “ropls” was used. The variable importance in the projection (VIP) value of the model was determined through multiple cross-validation. Differentially accumulated metabolites (DAMs) were identified by combining the different multiple, p-values, and the VIP value of the OPLS-DA model, with the following criteria: fold-change ≥ 1, p< 0.05, and VIP > 1. Finally, the significance of the DAMs in the KEGG pathway enrichment was assessed using a hypergeometric distribution test.

Agronomic and physiological data of the two cultivars under different cold treatments were analyzed by one-way analysis of variance with SAS, version 9.4 (SAS Institute, Cary, NC). Duncan’s multiple range test was performed to determine the statistically significant differences. Comparison of treatment/variety means was based on least significant difference (LSD) at a 0.05 probability level. Plotting was performed using Origin 2021 and Microsoft Office Excel 2020.

Different cold stresses had a significant effect on the pollen viability and seed-setting rate of the two wheat varieties. The pollen abortion rates at the anthesis significantly increased in all three treatments (T1, T2 and T3) in both wheat varieties compared to the controls, with more abortion rates in T3-treated samples ( Figure 1A ; Supplementary Figure 2 ) and higher pollen abortion rates in CM42-treated samples than in CM104. The cold stress significantly reduced the seed-setting rates in both varieties at the maturity stage, with lower rates in T1-, T2- and T3-treated CM42 samples than in CM104 cold-treated samples ( Figure 1B ; Supplementary Figure 3 ). However, there were no significant differences in the number of degenerate and fertile spikelets per spike, or thousand-grain weights across different cold treatments for each variety ( Supplementary Table 2 ).

The contents of the soluble and sugar protein, proline and ABA showed a significant upward trend from T3-, T2-, and T1-treated samples in both varieties to the controls, with those detected in CM104 also significantly higher than those found in CM42 ( Figures 2A–D ). On the contrary, GA₃ content significantly declined in both varieties, with CM104 showing a markedly lower hormone than CM42 under the same cold treatment ( Figure 2E ). The antioxidant enzyme activities of SOD, POD, and CAT initially increased and then decreased with prolonged cold treatment, with CM104 exhibiting higher enzyme activities than CM42 under the same cold treatment ( Figures 2F–H ).

A total of 9,656 proteins were found and analyzed in CM104CK_vs_CM104CT ( Supplementary Table 11 ). The principal component analysis of the proteome revealed a clear distinction between the two treatments ( Figure 5A ). The cold stress resulted in significant changes in 173 DEPs compared to CK, with 151 proteins upregulated and 22 proteins downregulated ( Figure 5B ). Of these, two dehydrin proteins, namely TraesCS7A03G1365300 and TraesCS7B03G1305200 were the most up-regulated DEPs (log₂FC: 4.70 and 3.04). The abundance of several cold-responsive proteins, including TraesCS2D03G0959100 (log₂FC: 2.80), TraesCS2B03G1138000 (log₂FC: 2.30), and two LEA proteins, namely TraesCS5A03G1181100 (log₂FC: 2.10) and TraesCS3A03G0416800 (log₂FC: 2.00) also substantially increased ( Supplementary Table 12 ). The results of hierarchical clustering algorithm analysis showed that the criteria for the identification of DEPs were reasonable ( Figure 5C ). The COG enrichment analysis revealed the participation of numerous DEPs in post-translational modification, protein turnover, chaperones, general function prediction, cell wall/membrane/envelope biogenesis, amino acid transport and metabolism, and lipid transport and metabolism ( Supplementary Figure 6 ). The KEGG enrichment analysis also indicated significant changes in key pathways, including photosynthesis - antenna proteins, metabolic pathways and glycerophospholipid metabolism that were significantly changed ( Figure 5D ).

A total of 938 metabolites were detected between CM104CK and CM104CT, with a decrease in the content of 586 and an increase in the content of 352 metabolites ( Supplementary Table 12 ). The principal component analysis revealed a notable distinction in metabolic patterns between the CK and CT groups ( Figure 6A ). The OPLS-DA analysis indicated significant changes in metabolite concentrations under cold stress, suggesting that certain variables could be used to differentiate between the CK and CT groups ( Figure 6B ). A comparison of the metabolite concentrations between CM104CK and CM104CT groups identified 180 DAMs, with 64 upregulated and 116 downregulated ( Figures 6C, D ; Supplementary Table 12 ). The four most up-regulated DAMs were LysoPC 22:6, 6’-O-Sinapoylsucrose, 4-O-(2’’-O-acetyl-6’’-P-coumaroyl-β-D-glucopyranosyl)-P- coumaric acid and 5-L-Glutamyl-L-amino acid (log₂FC>10, Supplementary Table 12 ). The DAMs were categorized into 21 groups, with carboxylic acids and their derivatives, fatty acyls, and organooxygen compounds as the three most abundant groups ( Figure 6E ). Additionally, KEGG enrichment analysis identified significant enrichment in 69 metabolic pathways, including amino acid biosynthesis, carbon metabolism, lysine degradation and 2-oxocarboxylic acid metabolism as the four most enriched pathways ( Figure 6F ).

To better understand the molecular mechanism associated with the tolerance of CM104 to low temperatures, we performed an integrated analysis of the multi-omics sequencing data. Comparative transcriptomic and proteomic profiling revealed DEGs with the same trend of mRNA and protein expression ( Supplementary Figure 7A , quadrants 3 and 7) and were enriched in 17 KEGG pathways after cold stress ( Supplementary Figure 7B ). Among them, those enriched to DEPs were photosynthesis-antenna proteins, glycerophospholipid metabolism, ether lipid metabolism, caffeine metabolism, and starch and sucrose metabolism ( Figures 7A, B ; Supplementary Figure 7C ). Integrated transcriptome and metabolome analysis revealed that DEGs and DAMs were distributed in 64 KEGG pathways, with starch and sucrose metabolism, carotenoid biosynthesis, arginine and proline metabolism, glycolysis/gluconeogenesis, and glycerolipid metabolism having the highest number of DEGs/DAPs ( Figures 7A, C ; Supplementary Figure 8 ). Proteomic and metabolomic analyses indicated the enrichment of DEPs and DAMs in 29 common KEGG pathways, with the top five pathways namely, glycerophospholipid, arginine and proline, ether lipid, pentose and glucuronate interconversions, and thiamine metabolism having the highest number of correlated DEPs/DAMs ( Figures 7A, D ).

Integrated analysis of the transcriptomics, proteomics, and metabolomics data sets identified significant enrichment in three common KEGG pathways ( Figure 7A ), including starch and sucrose, glycerophospholipid, and ether lipid metabolism. In the starch and sucrose metabolism pathway, most of the genes encoding sucrose-phosphate synthase (SPS, EC:2.4.1.14), sucrose synthase (SS, EC:2.4.1.13), alpha, alpha-trehalose-phosphate synthase (α,α-TPS, EC:2.4.1.12), glucose-1-phosphate adenylyltransferase (EC:2.7.7.27), α-amylase (EC:3.2.1.1) and granule-bound starch synthase (EC:2.4.1.242) were upregulated. Three metabolites, including D-Fructose 6-Phosphate, D-Glucose 1,6-bisphosphate and Trehalose 6-phosphate were downregulated, while three proteins namely, SS, glucose-1-phosphate adenylyltransferase and β-fructofuranosidase were upregulated in cold stressed wheat ( Figure 8A ). For the glycerophospholipid metabolism pathway, most of the genes encoding phospholipase D1/2 (EC:3.1.4.4), diacylglycerol kinase (EC:2.7.1.107), and phosphoethanolamine N-methyltransferase (EC:2.1.1.103) were upregulated, while most of the genes encoding glycerophosphodiester phosphodiesterase were downregulated. One metabolite, glycerol-3-phosphocholine was upregulated and another, dihydroxyacetone phosphate was downregulated, while all the two proteins, namely phospholipase D1/2 and phosphoethanolamine N-methyltransferase were upregulated in wheat exposed to cold ( Figure 8B ). In addition, the metabolite choline alfoscerate, and the gene encoding phospholipase D1/2 and its corresponding protein were implicated in the ether lipid metabolism pathway ( Figure 8B ).

The booting stage is the key stage in the differentiation and development of wheat spikes and is easily influenced by external temperature (Yu et al., 2022). Extreme low-temperature stress adversely impacts anther development, subsequently reducing seed setting rates, grain numbers per spike and yields (Liu et al., 2020). Low temperature also causes abnormal meiosis of pollen mother cells, delayed disintegration of the tapetal layer, and insufficient accumulation of anther starch grains, resulting in pollen abortion (Barton et al., 2014). In this research, pollen viability and seed set were significantly affected by cold stress in both varieties, with more severe effects in those with longer exposure to low temperatures. Chakrabarti et al. (2011) found a significant positive correlation (R ² = 0.65) between pollen germination percentage and seed setting rate in wheat crops, consistent with the findings of this experiment. The cold-tolerant variety CM104 exhibited a significantly lower pollen abortion rate than the cold-sensitive variety CM42 across all the treatments, thereby maintaining a high seed setting rate and mitigating the impact of cold stress on its yields.

Crop plants experience various physiological changes when exposed to low-temperature stress and the strength of cold tolerance in wheat is closely linked to changes in different osmoregulatory substances, hormone contents and protective enzyme systems (Thakura et al., 2010). The response to cold stress in different types of wheat is found to be divergent. Winter wheat requires a low temperature during the vernalization period and can survive cold conditions during the early vegetative stages (Ejaz et al., 2023). Compare to spring wheat, winter wheat showed higher cold tolerance, potentially due to their effective osmoregulation ability, high photosynthetic capacity, and lower lipid peroxidation under low-temperature conditions (Zhang et al., 2016; Hassan et al., 2021). As the main osmoregulatory substances, the increase in soluble sugars and proteins and proline in cold-stressed plants effectively increases the cytosol concentrations and lowers the freezing point of the cells, aiding in maintaining the integrity of the cell membrane (Chen et al., 2014). In our study, CM104 had higher levels of soluble sugars, soluble proteins and proline than CM42 under various cold treatments, suggesting that CM104 synthesizes more osmoregulatory substances to enhance its cold tolerance and mitigate cold damage. Cold stress also alters the hormonal balance in plants by affecting levels of ABA and GA₃ (Hassan et al., 2021; Soualiou et al., 2022). In this study, cold stress increased ABA but decreased GA₃ content in young spikes of both varieties. The ABA content was higher in CM104 than in CM42, whereas the GA₃ content was lower, which was conducive to promoting stomatal closure to maintain water balance and increase or maintain the content of chlorophyll, proteins and nucleic acids, thereby increasing the cold tolerance. The low-temperature stress induces excessive reactive oxygen species (ROS) in plants, disrupting cellular balance and causing severe cell membrane damage and lipid peroxidation. Therefore, plants have evolved a sophisticated enzyme system, including SOD, POD, and CAT, to mitigate ROS-induced plasma membrane damage (Zhang et al., 2016; Hail et al., 2017; Jiang et al., 2022). In our current study, CM104 exhibited higher antioxidant enzyme activities than CM42 under various cold treatments, suggesting its greater efficiency in ROS scavenging and reduced cold-induced damage in CM104 than CM42.

In crop plants, cold stress usually causes significant changes in gene expression patterns and cold tolerance is closely linked to the co-expression of these genes (Gusain et al., 2023). The TFs are crucial in plant development and environmental stress resistance, and various TF families, including AP2/ERF, bHLH, MYB, HSF, NAC, bZIP, and WRKY, play important roles in cold-responsive gene expression (Chen et al., 2014; Ahad et al., 2023). We identified more TFs in CM104 than in CM42, with the most differentially expressed TF genes in both cultivars belonging to bHLH, AP2/ERF, and HSF families. The AP2/ERFs are a significant TF family in plants, and are integral to plant growth, development, stress responses and biosynthetic pathways (Singh et al., 2002). The CBFs, which are members of the AP2/ERF family, regulate a series of cold-regulated (COR) genes as has been shown in the connection between CBF3 and the frost-tolerance locus Fr-A2, demonstrating that CBF3 positively regulates cold stress responses in wheat (Vágújfalvi et al., 2003). The TaCBF14 and TaCBF15 genes which enhance low-temperature tolerance in spring barley have also been identified in winter wheat (Solteísz et al., 2013), further supporting the roles of CBFs in regulating cold stress. The overexpression of ERF TFs, TERF2/LeERF2 significantly enhanced the cold tolerance in tomato and tobacco (Zhang and Huang, 2010), while TaPIE1, an ERF gene family activated the defense and ethylene signalling pathways in wheat, thereby improving its tolerance to cold stress (Zhu et al., 2014). In our study, all the ERF family DEGs were up-regulated in CM104, while some were down-regulated in CM42, indicating that these TFs may be crucial in CM104’s cold tolerance mechanism.

The bHLH family genes are the second largest TFs in plants and are crucial in plant stress response. Recent studies have identified numerous bHLH TFs that improve plant cold tolerance via the CBF-COR pathway or by enhancing ROS scavenging (Singh et al., 2002). Zheng et al. (2021) found that the apple bHLH homolog MdPIF3 negatively regulates cold tolerance. However, Yang et al. (2022) demonstrated that NtbHLH123 enhances cold stress in tobacco by regulating pathways and maintaining reactive oxygen homeostasis. In wheat, the overexpression of the bHLH gene TaMYC2 improved cold tolerance in Arabidopsis by interacting with TaICE41 to regulate the ICE-CBF-COR module (Wang et al., 2022). As important regulatory factors, HSFs play significant roles in stress responses to various biotic and abiotic stresses in wheat and other plants (Singh et al., 2002). For instance, HSFA1b enhances heat tolerance by activating OPR3 expression and promoting JA synthesis in wheat and Arabidopsis (Tian et al., 2020), while the overexpression of the wheat HSF genes, TaHsf3 and TaHsfA6f in Arabidopsis improves tolerance to heat, cold, and drought stress (Zhang et al., 2013; Bi et al., 2020). In this study, some bHLH and HSF TFs were up-regulated while others were down-regulated under cold stress in both varieties, indicating their complex response mechanism to low temperatures in wheat. In addition, other TFs such as the CSA and MTB2 TFs, which were uniquely identified in CM104, may be directly related to its cold tolerance but deserve further analysis in future studies.

Plant hormones are crucial in managing abiotic stress by activating cold response genes and coordinating various signalling pathways (Frascaroli, 2018; Gusain et al., 2023). In this study, we found genes related to phytohormone activities such as ABA in response to cold stress in wheat. A metabolomic analysis demonstrated significantly elevated ABA and Ja-L-Ile levels in wheat under cold stress and identified critical pathways linked to ABA/JA signalling as significant regulators of cold tolerance in wheat (Zhao et al., 2019). In addition, Abhinandan et al. (2018) showed the performance of essential functions in various abiotic stresses by the PP2C gene family, which is also a core component in the ABA-dependent signalling pathways. All DEGs related to PP2C were upregulated in the CM104 and CM42 under cold stress. However, the PP2C DEGs namely TraesCS1A03G1005000, TraesCS2A03G0000700 and TraesCS3A03G0542200) were unique to CM104, most likely due to the ABA-induced expression of PP2C, which regulates SnRK2 protein kinase activity to initiate ABA signalling (Zhao et al., 2019; Jiang et al., 2022). Therefore, the variation in the expression of PP2C between the two wheat cultivars indicates its potential significance in cold tolerance.

Auxin, a key regulator of plant growth and development, has an unclear role in response to abiotic stresses like cold (Abhinandan et al., 2018). It interacts with ABA and JA to regulate the expression of TFs, such as WRKYs, thereby enhancing cold tolerance in plants (Raza et al., 2023). The wheat auxin response factor (ARF) genes are also crucial in the interaction between abiotic stress responses and the ABA signalling pathway (Xu et al., 2020), while the auxin-responsive protein gene (IAA14) may play a role in wheat cold tolerance (Jiang et al., 2022). We annotated TraesCS6B03G0760600 and TraesCS6A03G0612400 as ARF8 and TraesCS5D03G0986600 as IAA14, and showed their upregulation in both cultivars, indicating their potential role in wheat cold tolerance. However, there is a need to study the role and function of auxin under cold stress in wheat.

Ethylene negatively regulates plant growth and influences various physio-biochemical mechanisms associated with biotic and abiotic stresses (Soualiou et al., 2022). For instance, ethylene-induced cold stress tolerance activates defense-related genes and regulates ROS-scavenging pathways via signalling cascades to mitigate the effects of cold stress (Raza et al., 2023). The upregulation of ethylene response factor (ERF) genes, TaERF1 and TaERF3 enhances tolerance to salt, drought, and cold in wheat (Abhinandan et al., 2018). In the present study, most DEGs related to ERF were upregulated in CM104 and downregulated in CM42, most likely due to the differences in cold resistance of the two varieties.

During cold stress in plants, a series of response gene expression products or cold-responsive proteins are produced via signal transduction to mitigate low-temperature damage and enhance cold tolerance (Frascaroli, 2018; Ahad et al., 2023). The LEA/dehydrin proteins are the key types of cold-responsive proteins, crucial in plant response to cold stress (Vítámvás et al., 2010; Kosová et al., 2021; Soualiou et al., 2022; Xu et al., 2022). For instance, the overexpression of maize dehydrin gene ZmDHN2b in tobacco enhances cold tolerance by decreasing malondialdehyde accumulation and electrolyte leakage (Xing et al., 2011), while the overexpression of Solanum habrochaites dehydrin gene ShDHN in tomato plants enhances its adaptation to cold stress by improving ROS scavenging and water-retention capacities (Liu et al., 2015). In wheat, the overexpression of the dehydrin gene Wdhn13 in Arabidopsis enhances freezing stress tolerance, while gene silencing increases its sensitivity to cold (Xu et al., 2022). Vítámvás et al. (2019) found a significant correlation between dehydrin transcript/protein accumulation and winter survival in both winter wheat and winter barley plants in the field conditions. In this study, many dehydrin genes and proteins such as TraesCS6A03G0900100 and TraesCS6B03G1082300 were highly up-regulated in CM104, indicating that they may improve cold tolerance in this variety. Future investigations are also needed to comprehensively explore the function of these important cold-responsive proteins in wheat.

In addition to genes and proteins, numerous metabolites such as p-coumaroyl regulate plant growth and development and confer cold stress resistance in plants (Lv et al., 2022). Other metabolites, such as sucrose, phenylalanine, glutamine, glutamate and proline enhance cold acclimation in tobacco (Song et al., 2024). In this study, cold stress significantly upregulated lysoPC 22:6 and 5-L-glutamyl-L-amino acid in wheat. LysoPC occurs in the cell membrane, with a functional role in its bilayer. The LysoPC level is often adjusted by lipid remodeling, which is crucial in cold stress tolerance in maize (Gao et al., 2024). On the other hand, 5-L-glutamyl-L-amino acid is linked to glutathione metabolism, a crucial pathway for oxidative stress defense and its increased levels aid in defense against high saline-alkali stress in rice leaves (Qian et al., 2023). Therefore, the significantly increased levels of LysoPC 22:6 and 5-L-Glutamyl-L-amino acid indicate their potential importance in cold tolerance in wheat.

Cold acclimation, deacclimation and reacclimation are complex, evolution-based physiological processes through which plants respond to their environment. These processes all involve multiple genes and metabolic pathways (Kosová et al., 2025). Integrated multi-omics analysis could provide comprehensive information on key metabolic pathways involved in cold stress (Yadav, 2010; McLoughlin et al., 2018). Our comprehensive analysis of the transcriptome, proteome and metabolome, showed that several associated DEGs, DEPs and DAMs were involved in starch and sucrose, glycerophospholipid and ether lipid metabolism. Starch and sucrose metabolism is a crucial plant metabolic pathway that significantly influences responses to abiotic stress (Chen et al., 2014) and may be closely related to cold resistance in wheat (Zhao et al., 2019; Jiang et al., 2022; Li et al., 2023). Soluble sugars, such as sucrose, glucose, trehalose and maltose involved in this pathway provide energy that enhances antioxidative capacity, maintains membrane stability, and mitigates cold stress damage (Zhang et al., 2016; Lv et al., 2022).

Our study demonstrated a significant increase in soluble sugars in both varieties under cold stress. Similarly, DEPs named SS and β-fructofuranosidase, and most DEGs encoding SPS, SS, α,α-TPS, and α-amylase associated with the regulation of sucrose, trehalose and maltose production, were highly upregulated. SS and β-fructofuranosidase hydrolyze sucrose to increase levels of glucose, which is crucial for osmotic protection and energy production during cold stress. In addition, the D-fructose 6-phosphate and trehalose 6-phosphate metabolites were downregulated, indicating that low temperatures might inhibit their synthesis or cause their decomposition into downstream substances of sucrose and trehalose. These results show that the metabolism of starch and sucrose improved, contributing to cold tolerance in CM104 at the booting stage.

The metabolic pathways associated with membrane lipids tend to change significantly under cold stress, since plant membranes are the first barrier to respond to external environmental stimuli (Chen et al., 2014; Hou et al., 2016). In the present study, the genes and proteins encoding phospholipase D1/2 and phosphoethanolamine N-methyltransferase, and the metabolite of glycerol-3-phosphocholine related to glycerophospholipid and ether lipid metabolism pathway were upregulated. These genes, proteins and metabolites are involved in plasma membrane synthesis and degradation. Phospholipase D is a crucial enzyme that plays a significant role in abiotic and biotic stress signalling in plants. It mediates plant defense responses mainly by participating in resistance signalling pathways associated with ROS (Wang, 2005). Phosphoethanolamine N-methyltransferase catalyzes the methylation of phosphoethanolamine to phosphocholine, which plays a significant role in abiotic stress response. Phosphoethanolamine N-methyltransferase transcript and its corresponding protein were strongly upregulated by low temperature in wheat (Charron et al., 2002), consistent with our findings. This indicates the important role of phospholipase D1/2 and phosphoethanolamine N-methyltransferase in response to cold in wheat. Besides, glycerol-3-phosphocholine is a potential biomarker for drought stress responses and also serves as an important nutrient source (Chen et al., 2022). In our study, glycerol-3-phosphocholine was upregulated, indicating its significant role in response to cold in wheat. Overall, our findings identified potential gene, protein, and metabolite candidates linked to metabolic pathways involved in cold responses during the booting stage in wheat. However, further research is required to confirm the specific functions of these genes, proteins, and metabolites, and to integrate them into future wheat breeding programs to aid in the selection and breeding of cold-tolerant wheat varieties.