The MYB family is one of the largest TF families in plants, and it was named for the conserved MYB domain that is present in all eukaryotic MYB TFs (Katiyar et al., 2012). MYB TFs play important roles in the tolerance to various stresses in plants, including cold stress (Wu et al., 2019). The Arabidopsis CBF gene promoters contain a MYB recognition sequence and can be activated by MYB TFs (Li et al., 2019a). Unlike AtMYB15 in Arabidopsis, which is a negative regular of CBF gene transcription, two characterized maize MYBs are positive regulators of CBF genes (Figure 1B). Expression of the maize ZmMYB31 gene, which encodes an R2R3-MYB transcription factor, is induced at low temperature. Overexpression of ZmMYB31 in Arabidopsis was found to upregulate the expression of Arabidopsis CBF genes, thereby enhancing the resistance of transgenic Arabidopsis plants to low temperature and oxidative stress (Li et al., 2019a). Meng and Sui (2019) isolated another nucleus-located R2R3-MYB transcription factor, ZmMYB-IF35. Low temperature induced the expression of ZmMYB-IF35 in the cold-tolerant maize inbred line M54. Transgenic Arabidopsis plants overexpressing ZmMYB-IF35 showed improved cold tolerance, higher antioxidant enzyme activity, reduced levels of reactive oxygen species (ROS), and less ion leakage. ZmMYB-IF35 expression also positively regulated the expression of stress-related genes such as AtCBF2, AtCBF3, AtCOR1, and AtCOR2. Similar to the functions of maize ZmMYB genes, overexpression of a rice R2R3-MYB gene, OsMYB2, was found to improve cold tolerance, indicating that OsMYB2 is also a positive regulator of cold tolerance in rice (Yang et al., 2012). Further analysis of the expression patterns of 46 ZmMYB genes under different abiotic stress showed that 22 of these genes respond to different stress treatments. Among them, there were six ZmMYB genes that responded to cold treatment, but only ZmMYB53 expression was exclusively induced by cold stress (Chen et al., 2018).

MAPK or MPK (mitogen-activated protein kinases) proteins, a family of serine/threonine protein kinases, are involved in many important processes including stress signal transduction and development (Kong et al., 2013). The MAPK cascade pathway is a ubiquitous signal transduction module in eukaryotes that transmits biological signals from receptors to target molecules through a variety of intracellular and extracellular routes (Moustafa, 2014). The MAPK cascade component involves a three-kinase module, namely MAPK, MAPKK (MAPK kinase), and MAPKKK (MAPKK kinase; Xiang et al., 2021). The MAP kinase cascade can phosphorylate AtICE1 to promote its degradation, and thus is involved in the regulation of cold tolerance in Arabidopsis (Zhao et al., 2017; Figure 1A). In maize leaves, ZmMPK5 was found to be involved in the recovery of plants from cold stress (Berberich et al., 1999). Kong et al. (2011) isolated ZmMKK4, a group C MAPKK gene, from the root system of maize variety “Zhengdan 958,” and expression of the ZmMKK4 transcript was found to be up-regulated by low temperature exposure. Overexpression of ZmMKK4 in Arabidopsis increased the plants' cold tolerance, which showed that ZmMKK4 is a positive regulator of cold tolerance in maize (indicated by the green arrow in Figure 1B). Pan et al. (2012) identified ZmMPK17, a group D MAPK gene, that was induced by cold stress. The overexpression of ZmMPK17 enhanced cold tolerance in tobacco (Nicotiana tabacum L.) plants by affecting the antioxidant defense system. Cai et al. (2014) isolated and identified ZmMKK1, a group A MAPKK gene, from “Zhengdan 958”. Ectopic expression of ZmMKK1 in tobacco enhanced its cold tolerance, suggesting that ZmMKK1 is also involved in the response of plants to low temperature.

A very recent study identified a new cold regulation pathway, ZmMPK8-ZmRR1-ZmDREB1.10/ZmCesA2, in maize (Zeng et al., 2021; Figure 1B). The ZmRR1 (type-A Response Regulator 1) transcript level is slightly decreased, whereas the ZmRR1 protein level is increased by cold treatment of maize seedlings at 4°C. Overexpression of ZmRR1 leads to enhanced cold tolerance by accumulating and inducing the expression of the downstream genes, ZmDREB1.10 and ZmCesA2 (Cellulose synthase 2), suggesting that ZmRR1 acts as a positive regulator of maize cold tolerance. The ZmMPK8 protein, a negative regulator of cold tolerance, phosphorylates ZmRR1 at Ser15. A natural variation of ZmRR1 with a 45-bp deletion that encompasses Ser15 prevents its phosphorylation by ZmMPK8. At present, the ZmMPK8-ZmRR1-ZmDREB1.10/ZmCesA2 pathway is the best characterized cold regulation pathway in maize, and it provides an in-depth understanding of the molecular mechanism underlying cold tolerance in maize.

To date, functional investigations of many cold responsive genes have been performed in maize, but the functions of many of these genes were studied by overexpression in model plant species. As shown by the green and orange dashed arrows in Figure 1B, a total of eight maize genes were transformed into Arabidopsis, and two maize genes were transformed into tobacco. Only one study investigated the molecular function of ZmRR1, ZmMPK8, and their downstream genes ZmDREB1.10 and ZmCesA2, by either mutating or over-expressing the corresponding genes in maize. Transformation of maize genes into maize plants not only provides stronger evidence for gene functions, but also enables the straightforward elucidation of molecular pathways to provide in-depth biological insights. Furthermore, such native transformation would provide potential genetic resources for improving cold tolerance in maize.

Osmotic adjustment is one of the most important physiological mechanisms employed by plants to cope with many types of stresses. Under adverse conditions, osmotic adjustment substances help to maintain cell turgor and the capacity of tissues to retain water, and the contents of osmotic adjustment substances are positively correlated with plant stress resistance (Farhangi-Abriz and Torabian, 2017). The accumulation of osmotic adjustment substances can reduce the water potential in plant cells, prevent water outflow, and at the same time protect the structure of macromolecules in the cell (Li et al., 2013). The main osmotic adjustment substances in plants include soluble sugars, soluble proteins, and free proline (Zhang, 2014).

Soluble sugars are strongly hydrophilic, and they can reduce the stress damage to plant cells by reducing the water potential and maintaining the activity of some biological macromolecules (Guan et al., 2009). Soluble proteins can scavenge ROS (reactive oxygen species) and stabilize cell membrane structure (Nadarajah, 2020). Proline is a hydrophilic amino acid, and free proline functions to stabilize the metabolic process in the cytoplasm (Guan et al., 2009). Ma et al. (2015a) found that cold stress induces the expression of ZmP5CS1, a key gene for proline synthesis in immature maize embryos. Some studies have reported that low temperature treatments increase the contents of soluble sugars, proline, and soluble proteins in the seedlings of many maize varieties. The relative increases were larger in the varieties with strong cold tolerance compared to the varieties with relatively weaker cold tolerance. The lower the temperature, the more significant the difference between varieties (Li and Fang, 2018; Zhao et al., 2020). In conclusion, cold tolerant maize can maintain the water absorption capacity of cells and can reduce cold damage by accumulating higher levels of osmotic adjustment substances in response to cold stress.

Cold stress can destroy cellular homeostasis and cause the accumulation of high levels of ROS in plant cells. Excessive levels of ROS can be harmful to the cell membrane system because of lipid peroxidation (Nadarajah, 2020). Plant cells can induce the expression of a protective system to resist the threat posed by ROS. The protective enzymes within a plant cell, such as SOD (superoxide dismutase), CAT (catalase), POD (peroxidase), and APX (ascorbate peroxidase), cooperate with one another to scavenge the reactive oxygen free radicals so as to maintain normal physiological metabolic activities as much as possible to avoid damage to cell components (Nadarajah, 2020). Maize cells can stabilize the membrane structures and reduce cell damage by constantly oxidizing the reactive oxygen free radicals. Many studies have shown that cold stress enhances the activities of SOD, POD, and CAT in maize. The activities of the three main antioxidant enzymes and the relative expression of related genes are positively correlated with the cold tolerance of inbred lines (Wei et al., 2014; Yang et al., 2016; Li and Fang, 2018). For example, an early study showed that low temperature induces the expression of the APX gene, and the APX activity in cold tolerant maize was found to be higher than in cold sensitive maize (Pinhero et al., 1997). One study showed that the activities of protective enzymes decreased in the three cold-tolerant and three cold-sensitive maize inbred lines under cold stress, but the relative decrease in the cold-tolerant inbred lines was less than that in the cold-sensitive inbred lines (Peng et al., 2016). These differences could be due to many reasons, but we can conclude that the antioxidant enzyme system in cold-resistant maize lines is generally stronger than that of cold-sensitive lines.

MDA (malondialdehyde) is the main product of peroxidation of polyunsaturated cell membrane lipids and is a marker for oxidative stress. MDA is a highly reactive compound that can restrain the activity of cell protective enzymes and aggravate membrane peroxidation. When the cell membrane is damaged, a large number of electrolytes flow out of the cell, resulting in a surge of electrolytes (Wang et al., 2006). Therefore, relative conductivity and MDA levels are negatively correlated with cold tolerance in maize lines (Zhang, 2014). Several studies have reported that with increasing time of exposure to cold, the relative conductivity and MDA content increases. The cell membranes of cold tolerant varieties were less damaged, and the relative conductivity and MDA contents increased to a lesser extent in maize lines that show strong cold tolerance (Zhang, 2014; Peng et al., 2016).

Plants tightly regulate the levels of some hormones to cope with the changing environment (Lamers et al., 2020). ABA (abscisic acid) is an abiotic stress hormone that can improve the stress resistance of plants, and it participates in the regulatory response for a variety of abiotic stresses, including cold (Qin et al., 2021). The ABA contents in the roots of cold-tolerant lines were significantly higher than in the roots of cold-sensitive lines. Exogenous ABA treatment can promote the ability of maize seeds to germinate at low temperature (Zhang, 2011). GA (gibberellic acid) is also considered to play a role in plant cold resistance, but its effect is not as obvious as that of ABA (Eremina et al., 2016; Rihan et al., 2017). Studies have shown that the GA and IAA (indole-3-acetic acid; auxin) contents in the roots of maize seedlings decrease under cold stress, while the ABA content gradually increased (Janowiak et al., 2002; Zhang, 2011; Wei et al., 2014).

SA (salicylic acid) is another important plant hormone that is involved in cold tolerance in several plant species including maize, and its direct physiological effect is the change in the antioxidant enzyme activity of plants (Farooq et al., 2009). At low temperature, the increase in endogenous SA biosynthesis is closely related to the increase in antioxidant enzyme activity in maize seeds and during seedling growth (Wang et al., 2013). Cold stress in maize seedlings was not moderated by treating the seeds with SA (Gómez-Muñoz et al., 2018), but SA+H₂O₂ treatment at low temperature shortened seed germination time, improved seedling vigor, reduced cold damage to maize seeds, increased the activities of antioxidant enzymes and the expression levels of the corresponding genes, and thus improved cold tolerance (Li et al., 2017). Exogenous application of BL (brassinolide) also increased the germination rate, reduced cold damage to maize seedlings and caused an increase antioxidant enzyme activity, all of which resulted in an increase in plant biomass (Sun et al., 2020). Thus, manipulation of plant hormones appears to be an efficient way to alleviate the damage caused by cold temperature and to ensure the growth of seedlings at later stages of development.

Cold tolerance in plants is highly correlated with mineral nutrient levels (Waraich et al., 2012). The inhibition of growth in maize from cold stress is at least partly caused by indirect damage due to the decreased uptake of nutrients from the soil (Gómez-Muñoz et al., 2018). For example, the absorbance of K and Pi (orthophosphate) by maize roots is particularly affected by cold soil temperature (Bravo et al., 1981). Seed treatment with Mn/Zn can mitigate the negative effects of cold stress, and it resulted in increased biomass production in high-P soil but not in low-P soil, indicating that the absorbance of P is relieved by Mn/Zn treatment (Gómez-Muñoz et al., 2018). In addition, both Mn and Zn are key co-factors of several enzymes that are involved in the detoxification ROS caused by cold stress. Application of Si (silicon) is also a useful strategy to improve cold tolerance in maize seedlings during the early growth stage. The beneficial effects of Si seed treatments included the restoration of hormonal balances that were disrupted by cold stress and maintaining homeostasis of other micronutrients (Moradtalab et al., 2018). Therefore, appropriate plant nutrition is a useful strategy to alleviate the cold stress.

Cold stress significantly decreases photosynthesis in maize (Fracheboud et al., 2002). This decrease in photosynthesis may result from many factors, such as impaired chloroplast development, changes to the pigment composition, damage of the PSII reaction centers, and reduced activity of carbon cycle enzymes. The reduced photosynthetic ability at low temperatures also reflects the cold resistance of maize (Hund et al., 2008). Fracheboud et al. (2002) found that the chlorophyll a/b ratio in the cold tolerant lines was higher than in cold sensitive lines. Duran Garzon et al. (2020) showed that cold tolerance is related to higher chlorophyll content, higher G6PD (glucose-6-phosphate dehydrogenase) activity, and a higher sucrose-to-starch ratio. Many studies have used photosynthesis-related traits to quantify cold tolerance in maize. These studies are discussed below and are given in Tables 1, 2.

In general, cold tolerant maize may have stronger ability to accumulate more osmotic adjusting substances, stronger antioxidant enzyme system to stabilized ROS, stronger ability to maintain hormonal homeostasis, higher absorbance of mineral nutrient, higher photosynthetic ability, etc.

Quantitative genetic analyses of cold tolerance in maize have so far shown that heterosis, general and specific combining abilities, reciprocal maternal and non-maternal effects, additive effects, genotype, growth stage, and environmental factors are all involved in the expression of cold tolerance in maize. For example, Hodges et al. conducted two studies to explore the cold tolerance of maize at both germination and seedling stages using twelve hybrids derived from a complete diallel of four inbred lines (Hodges et al., 1997a,b). The results showed that the seed germination rate, the activities of CAT and APX in leaves at the 3-leaf stage, and the general and the special combining ability of the dry matter weight of leaves at the 4-leaf stage were significantly different. Revilla et al. (2000) proposed that cold tolerance has additive, dominant, and maternal effects. The experiment of Yan et al. (2017) showed that there was significant heterosis in maize cold tolerance at the seedling stage. Neta et al. (2020) used three cold-tolerant lines and three cold-sensitive lines to carry out partial diallel cross experiments to analyze the heterosis, general and specific combining abilities, reciprocal maternal and non-maternal effects, as well as the expression of CAT, APX, SOD and other genes. Results showed that there was heterosis and a reciprocal effect for germination under cold stress, and that the non-additive genes were more important. The genes that control cold tolerance depend on the particular materials used and the traits studied (Frascaroli and Revilla, 2018). Each growth stage of maize, including germination, emergence, and early seedling growth, is controlled by an independent genetic model; therefore, cold tolerance might also be regulated independently in the different growth stages (Frascaroli and Revilla, 2018). To make the situation more complicated, some studies have also reported that there are interactions between genes and the environment in the expression of cold tolerance in maize (Presterl et al., 2007).

The studies cited above show that the genetics of cold tolerance in maize is very complicated. However, with the broad application of genomic tools, considerable progress has been made in recent years in the identification of genetic loci and genes that are associated with maize cold tolerance. The information from these studies will facilitate a deeper understanding of the genetic mechanisms that control maize cold tolerance at the molecular level.

QTL (quantitative trait locus) mapping is a powerful tool for the identification and manipulation of loci underlying important and complex traits in agricultural crops. Recent studies have used different mapping populations, such as F_2:3 families, RILs (recombinant inbred lines), and BC (back cross) populations, to study the QTLs that control cold tolerance in maize (Table 1). Most of these segregating populations were constructed by crossing cold-sensitive and cold-tolerant inbred lines in order to increase the relative degree of variation in the cold-response phenotypes. Several cold responsive physiological traits, such as seed germination rate, root phenotypes, and seedling photosynthesis, were measured under both normal and cold conditions (Table 1). These studies identified many reliable molecular marker loci associated with QTLs or genes that regulate cold tolerance and could be further used for breeding of cold tolerant inbred lines or hybrids.

Fracheboud et al. (2002) performed a QTL analysis of five traits related to the function of the photosynthetic apparatus at low temperature using a set of 233 maize RILs. They identified 18 QTLs that were significantly correlated with the target traits, of which the main QTL for leaf development at low temperature was the main QTL for pigment composition. Fracheboud et al. (2004) used an F_2:3 population derived from the cross ETH-DH7×ETH-DL3 and detected 19 QTLs controlling chlorophyll fluorescence at low temperature. The major QTL for photosynthetic cold tolerance in maize seedlings is located on chromosome 6. Using the same population, Jompuk et al. (2005) detected a major QTL for cold tolerance located on chromosome 6 that corresponded to chlorophyll fluorescence and chlorophyll content. Hund et al. (2004) used an F_2:3 population constructed from the cross Lo964×Lo1016 to map QTLs for root and shoot development in maize seedlings under cold conditions, and found a dominant QTL located on chromosome 5. Presterl et al. (2007) performed QTL mapping on 720 DH (doubled haploid) lines and found a total of 18 QTLs associated with leaf chlorosis, leaf purpling (anthocyanin), and frost damage at low temperature. Rodríguez et al. (2008) performed QTL analysis by using an IBM (intermated B73×Mo17) population. None of the QTLs was identified under normal growth conditions, but two QTLs significantly correlated with leaf color at low temperature were located on chromosomes 3 and 6, and these two OTLs explained 14.2% of the phenotypic variation and 28.2% of the genetic variation. Guerra-Peraza et al. (2012) used the IBM302 population for QTL analysis and discovered a major QTL for photosynthesis-related traits on chromosome 5. The favorable allele of this QTL was contributed by Mo17 and appeared to be the major factor that explained the differential response of B73 and Mo17 to changes in the temperature at night. Rodríguez et al. (2013) used an F_2:3 population derived from the cross of a cold susceptible (A661) with a cold tolerant (EP42) inbred line to detected genomic regions related to the cold-induced albino phenotype. A major QTL on chromosome 2 was identified that explained 14% of the phenotypic variation. Using the same F_2:3 population, 10 QTLs related to photosystem traits were identified, with six of them at normal temperatures and four under cold conditions. A parallel meta-QTL analysis identified three genomic regions that regulate the development of maize seedlings at low temperatures (Rodríguez et al., 2014). Hu et al. (2016) performed a QTL analysis using 243 IBM Syn4 RILs and detected six QTLs associated with the low-temperature germination rate and six QTLs correlated with low-temperature primary root length. Of these, four pairs of QTLs were located in the same genomic regions. Yan et al. (2017) performed a QTL analysis with seedlings from an F_2:3 population (207 lines) obtained by crossing a cold sensitive line (K932) with a cold tolerant (Mei C) inbred line. Their analysis resulted in the detection of seven QTLs controlling four cold-related traits, and one of the QTLs explained 10.55–25.29% of the phenotypic variation.

While the above-cited QTL analyses used “regular” populations, several studies have used “advanced” populations to improve the accuracy of the study. For example, Shi et al. (2016) constructed two connected RIL populations that shared one parental line and were able to identify 26 QTLs associated with seed vigor traits under low temperature conditions. Fourteen of these QTLs were further integrated into five mQTL regions through a meta-analysis. Two of the mQTL regions located on chromosomes 2 and 9 had R² values >10% and were previously identified as QTLs for seed vigor traits. Using two cold-tolerant and two cold-susceptible inbred lines, Li et al. (2018) generated three connected F_2:3 populations to detect QTLs related to seed germination ability at low temperature. A total of 43 QTLs and three mQTL regions were detected. Yi et al. (2020) used 406 lines from a multi-parent advanced generation intercross (MAGIC) population and found 858 SNPs grouped in 148 QTLs that were significantly associated with cold tolerance-related traits, and most of the QTLs were located in specific regions, particularly bin 10.04.

Several studies combined QTL analyses with other genetic or genomic approaches such as GWAS (genome-wide association study) or transcriptome analysis to efficiently uncover further biological insights. From the same population used by Hu et al. (2016); Goering et al. (2021) selected a panel of 97 lines for QTL analysis of traits including chlorophyll concentration, leaf color, and tissue damage at low temperature. Two cold-related QTLs with high additive impact were detected. These authors further verified the two QTLs using transcriptome data and identified 13 candidate genes likely to be involved in controlling the cold responses. Recently, Jin et al. (2021) studied cold responses in maize using a joint analyses that combined QTL and GWAS. They first performed QTL mapping for POD activity using an F_2:3 population (210 lines) derived from cold-tolerant (W10) and cold-sensitive (W72) lines, and detected 12 QTLs significantly associated with POD activity and cold tolerance. They then conducted GWAS on a natural population consisting of 80 backbone inbred lines and found that four SNPs were significantly associated with POD activity at low temperature (Table 2). Using a joint analysis of the QTL and GWAS results, Zm00001d002729 was determined to be a potential cold tolerance gene. Overexpression of Zm00001d002729 increased the cold tolerance of maize seedlings by increasing POD activity and decreasing the MDA content and relative conductivity (ion leakage).

Table 1 shows that the genomic loci involved in the regulation of cold tolerance are distributed on almost all of the maize chromosomes, but most of these QTLs have not been further fine mapped or functionally characterized as cold tolerant genes, which prevents more detailed studies of cold tolerant mechanism in maize. One major reason for this may be the huge amount of work involved in gene mapping in large populations. In recent years, joint analysis using advanced populations or genomic tools has become increasingly popular to narrow down the search for candidate genes. Such an approach was recently used successfully to identify the major-effect cold regulating gene Zm00001d002729 (Jin et al., 2021), suggesting that it can be more widely applied in other studies.

Compared to biparental segregating populations, natural populations consisting of several hundred maize inbred lines have also been used for the genetic analyses of cold tolerance in GWAS (Frascaroli and Revilla, 2018). Although most of the cold-responsive physiological traits used in GWAS were similar to those in QTL analyses, including leaf color, root length, and seed germination-related traits (Tables 1, 2), GWAS offers increased mapping resolution and accuracy due to the higher level of genetic diversity in the mapping populations (Lipka et al., 2015).

Strigens et al. (2013) conducted the first GWAS in maize to dissect traits related to cold tolerance. Using a maize germplasm collection of 375 inbred genotypes with 56K SNPs (single nucleotide polymorphisms) GWAS identified 19 significant association signals that explained between 5.7 and 52.5% of the phenotypic variance for cold-related traits such as early growth and chlorophyll fluorescence. Huang et al. (2013) used 125 maize inbred lines to perform GWAS on 10 traits related to cold tolerance at the germination and seedling stages. A total of 43 SNPs were detected that were associated with cold tolerance and 40 candidate genes were predicted based on 31 of these SNPs. Yan et al. (2017) crossed the cytoplasmic male sterile parental line S-Mo17 with 338 different inbred lines to generate a test cross association mapping population. GWAS was performed on these 338 test crosses, and 19 significant SNPs associated with cold tolerance-related traits were detected. Hu et al. (2017) conducted GWAS on seed germination traits using 282 inbred lines of maize under normal and low temperature conditions. A total of 17 associated SNPs related to cold tolerance were identified, and seven of the SNPs were located in candidate genes. In a population of 222 maize inbred lines, Zhang et al. (2020) used GWAS to identify 30 SNPs related to cold tolerance during maize seed germination. Fourteen candidate genes directly related to the SNPs were found and further verified by gene expression analysis. Zhang et al. (2021) studied germination-related traits in 300 inbred lines under low temperature conditions. GWAS analysis revealed a total of 15 significant SNPs, and three genomic loci were repeatedly associated with multiple traits. Yi et al. (2021) evaluated a large panel of 836 maize inbreds, and GWAS analysis uncovered a total of 187 significant SNPs that could be integrated into 159 genomic regions that controlled seed emergence and traits related to early growth.

Despite its higher genetic mapping resolution, GWAS has not been without controversy. In particular, many of the cold-associated SNP markers identified in these studies were found to be located in the non-coding regions and generally thought to function in the regulation of gene expression. But which gene(s) do they regulate? There are large structural variations present in the genomes of different maize inbred lines, and the genetic loci predicted in the B73 genome may not represent all the genetic loci in other inbred lines. Hence, there may be some other unpredicted genes near or even located on cold-associated SNP markers, leading to inaccurate results.

Transcriptome analyses are also widely used to understand the molecular responses of maize to cold stress and to mine for cold tolerant genes. Li et al. (2016) used RNA-seq analyses to compare the transcriptomes of a freezing tolerant (KR701) and a freezing sensitive line (Hei8834) before and after cold treatment at the seedling stage and identified 948 DEGs (differentially expressed genes). GO (Gene Ontology) analysis revealed that the terms “binding functions,” “protein kinase,” and “peptidase activity” were over-represented in the DEGs. Li et al. (2019b) used RNA-seq to analyze the gene expression of cold tolerant (M54) and cold sensitive (753F) inbred lines under cold stress at the seedling stage. More DEGs were found in M54 than in 753F after both 4 and 24 h of cold treatment, indicating that the cold-responsive signaling networks were more active in the cold tolerant line. Li et al. (2020) analyzed the transcriptome of maize B73 seedlings under different low temperature conditions. In this study, 5,358, 5,485, and 5,312 DEGs were detected in response to cold stresses of 4, 10, and 16°C, respectively, and the expression of five genes including ZmDERB1 was significantly up regulated. Frey et al. (2020) selected 21 DH lines from a DH population (flint landrace “Petkuser Ferdinand rot”) based on their cold tolerance; 11 lines were cold resistant and 10 were cold sensitive. The transcriptomes of the 21 DH lines were analyzed after control and cold treatments. Here, 148, 3,254, and 563 DEGs were found to be related to cold treatment, cold tolerance, and growth rate at low temperature, respectively. Zhang et al. (2020) used RNA-seq to verify the correlation between the candidate genes and low-temperature tolerance and found 10 DEGs that were located in the linkage disequilibrium region of a GWAS analysis. Of these genes, two of them appear to regulate cold signal transduction and cell membrane fluidity. Li et al. (2021) analyzed transcriptomic changes in seeds of three sweet corn NILs and their parents under cold stress. A total of 20 DEGs were found to be highly related to low-temperature germination, and a gene encoding UDP-glucosyltransferase was hypothesized to be essential to cold germination in sweet corn. Waititu et al. (2021) conducted a comparative analysis of the transcriptomes of seedlings of 24 cold-tolerant and 22 cold-sensitive inbred lines under cold stress. A total of 2,237 DEGs were identified, which included 147 TFs belonging to 32 families such as MYB, ERF, NAC, WRKY, bHLH, MIKC MADS, and C2H2. Yu et al. (2021) studied the leaf transcriptomic response of two maize inbred lines with contrasting cold tolerance levels under a time series cold treatment. The results showed that cold tolerance in line B144 is due to active mediation of stomatal opening and protection of photosystem II from photooxidation by upregulating the expression of genes for D1 proteins, while the sensitive line Q319 was unable to close its stomata in response to cold.

In addition to RNA-seq analyses, other genomic tools have also been used to study the transcriptomic changes in maize to cold. Rodríguez et al. (2013) used microarray hybridization on the chlorophyll-less and chlorophyll-containing sections of leaves of maize inbred line A661 which shows a cold-induced albino phenotype. A total of 1,002 differentially expressed transcripts were identified between the two sections, and these DEGs were classified into 23 categories including genes in the tetrapyrrole biosynthesis pathway and photosynthesis. Di Fenza et al. (2017) conducted microarray analysis with four varieties, with two cold-tolerant and two less cold-tolerant lines, to identify genes that were differentially expressed under chilling conditions. A total number of 64 DEGs were identified in the two chilling-tolerant varieties, while no significant changes in expression were observed in less cold-tolerant lines. Another study used cDNA-AFLP to analyze the gene expression changes in response to cold stress and identified three maize genes, ZmMAPKKK, ZmCLC-D, and ZmRLK, that were possibly involved in the cold response (Yang et al., 2011).

To investigate the regulatory roles of miRNAs (microRNAs) in cold tolerance, Aydinoglu (2020) studied the miRNome (miRNA microarray) in seedlings of the maize hybrid ADA313 that were treated with cold temperature. In this study, 24, 6, and 20 miRNAs were specifically differentially regulated in the meristem, the elongation zone, and the mature zone by cold stress, respectively. This study highlighted the importance of miRNAs in the maize response to cold stress. Combined with large-scale bioinformatic analysis, Zhou et al. (2022) examined the transcriptome changes in seedlings of the inbred lines B73, Mo17, and W22 and the F₁ hybrids with all three combinations (B73xMo17, W22xB73, and W22xMo17) in response to cold or heat stress. They identified many stress-related DEGs among the different maize genotypes and assigned these expression changes to cis-or trans-regulatory mechanisms using the F₁ hybrid data. Their study answered the question of how sequence differences in cis-elements of different genotypes impact a gene's responsiveness to stress, and shed light on the prediction of a plant response to stress by implementing a more sophisticated model construction (convolutional neural networks).

Above transcriptomic analyses indentified tens to thousands of DEGs in individual studies. This astonishing variaiton in the numbers of DEGs indicates that a vast majority of the DEGs can not overlap to each other, leaving an almost impossible task to mine the conserved overlaping genes. In deed, when Sowiński et al. (2020) surveyed 13 independent studies addressing the transcriptomic changes in response to cold stress in maize seedlings at the V2-V5 stages, they found <500 DEGs were reported in more than one study and the rest are specific to individual studies. Furthermore, among the 13 independent studies survied, four of these 13 studies used moderate low temperature stress and the other nine used severe cold stress. By comparing the DEGs, it was found that the transcriptomic changes that occur in response to moderate low tempreture and severe cold stress are fundamentally different (Sowiński et al., 2020). Taken together, the variation in gene expression in different studies can be caused by the genotypes, tissue, developmental stage, experimental design and the strength of cold. Although the results of individual studies varied considerably, the “common genes” shared by several independent projects after carefully eliminate the differences can be attractive candidates for further functional studies.