The transcription factor bZIP68 negatively regulates cold tolerance in maize

Abstract

Maize (Zea mays) originated in tropical areas and is thus susceptible to low temperatures, which pose a major threat to maize production. Our understanding of the molecular basis of cold tolerance in maize is limited. Here, we identified bZIP68, a basic leucine zipper (bZIP) transcription factor, as a negative regulator of cold tolerance in maize. Transcriptome analysis revealed that bZIP68 represses the cold-induced expression of DREB1 transcription factor genes. The stability and transcriptional activity of bZIP68 are controlled by its phosphorylation at the conserved Ser250 residue under cold stress. Furthermore, we demonstrated that the bZIP68 locus was a target of selection during early domestication. A 358-bp insertion/deletion (Indel-972) polymorphism in the bZIP68 promoter has a significant effect on the differential expression of bZIP68 between maize and its wild ancestor teosinte. This study thus uncovers an evolutionary cis-regulatory variant that could be used to improve cold tolerance in maize.

The transcription factor bZIP68 negatively regulates cold tolerance in maize, and natural variation in the bZIP68 promoter underlies its difference in expression between maize and teosinte.

IN A NUTSHELL.

Background: Cold stress is an important environmental factor that limits plant growth, development, and geographical distribution. Maize originated in Mexico and is susceptible to low temperatures. Cold stress not only inhibits the germination and seedling growth of maize, but it also affects ear formation and grain filling, resulting in decreased crop production. However, the molecular and genetic basis of the cold response in maize remains unclear.

Question: We aimed to identify genes that regulate the cold response in maize and dissect the underlying mechanism.

Findings: We identified the transcription factor bZIP68, a negative regulator of the cold-stress response, from a population of transgenic maize plants overexpressing more than 700 maize genes. bZIP68 interacts with MPK8 (MITOGEN-ACTIVATED PROTEIN KINASE8), which is also a negative regulator of the cold-stress response. MPK8-mediated phosphorylation of the Ser250 of bZIP68 not only increases the protein stability of bZIP68 at 4�C but also improves its binding ability to the promoter of the transcription factor gene DREB1.7. Re-sequencing analysis of the promoter region of bZIP68 revealed that the nucleotide diversity (π) was strongly reduced in maize compared with its ancestor teosinte, supporting the notion that bZIP68 was selected during maize domestication. A 358-bp Indel was selected in the bZIP68 promoter region during maize domestication, which leads to increased bZIP68 expression and decreased cold tolerance in maize.

Next steps: We determined that the bZIP68 locus was a target of selection during early domestication; however, the cold-tolerance favourable allele of bZIP68 was not selected. Therefore, it will be interesting to explore why bZIP68 was selected during maize domestication. Meanwhile, we plan to generate cold-tolerant maize varieties by editing the bZIP68 promoter.

Introduction

Cold stress is a major environmental factor that severely constrains the growth, development, and geographical distribution of crops in nature (Sobkowiak et al., 2016). The adverse effects of cold stress on plants affect many aspects of physiology and biochemistry, including photosynthesis, respiration, enzymatic reactions, carbon/nitrogen (C/N) balance, secondary metabolism, and nutrient absorption. Exposure to colder temperatures may therefore lead to poor germination, stunted seedling growth, and even the death of tissues or entire plants, which lowers crop quality and delays harvest and may even result in crop failure (Allen and Ort, 2001; Marocco et al., 2005; Nguyen et al., 2009; Hussain et al., 2020). Maize (Zea mays) is a major food crop that has adapted to a wide range of environmental conditions worldwide. However, because it originated at tropical latitudes, maize is more susceptible to cold stress than other cereal crops, such as wheat (Triticum aestivum), barley (Hordeum vulgare), and foxtail millet (Setaria italica) (Juurakko et al., 2021; Meng et al., 2021). Therefore, it is important to elucidate the mechanism that regulates cold sensitivity in maize, which will enable us to introduce key cold-tolerance genes/modules from other species for breeding cold-tolerant maize varieties with vigorous seedling growth to enhance the distribution of maize in high latitude areas.

Cold stress signal transduction has been intensively studied in the dicot plant Arabidopsis (Arabidopsis thaliana) (Ding et al., 2020). Under cold stress, the expression of many genes is induced by DEHYDRATION-RESPONSIVE ELEMENT BINDING FACTOR1 (DREB1) transcriptional activators. DREB1 binds to the C-repeat element (CRT)/DRE cis-elements within the promoters of COLD-RESPONSIVE (COR) genes and modulates their expression, thereby leading to freezing tolerance (Thomashow, 1999; Jia et al., 2016; Song et al., 2021). DREB1s are regulated at multiple levels (Shi et al., 2018; Ding et al., 2020). Although several potential genes involved in cold tolerance in maize have been identified, the underlying molecular basis of cold tolerance is poorly understood. Zea mays DREB1A increased plant freezing tolerance when overexpressed in Arabidopsis (Qin et al., 2004). In addition, ZmDREB1A directly promotes the transcription of RAFFINOSE SYNTHASE (RAFS) by binding to its promoter and increasing cold tolerance in maize (Han et al., 2020). Protein kinases in maize are also involved in the response to cold stress. For instance, the calcium-dependent protein kinase ZmCPK1 reduced freezing tolerance when overexpressed in Arabidopsis (Weckwerth et al., 2015). Similarly, the overexpression of ZmMKK1, encoding MITOGEN-ACTIVATED PROTEIN KINASE KINASE1, improved the cold tolerance of transgenic tomato (Solanum lycopersicum) by blocking the accumulation of reactive oxygen species (ROS) and increasing the accumulation of osmoregulatory substances (Cai et al., 2014). We recently showed that MITOGEN-ACTIVATED PROTEIN KINASE8 (MPK8) negatively regulates cold tolerance in maize by destabilizing the type-A response regulator RR1. Furthermore, RR1 enhances maize cold tolerance by inducing the expression of DREB1s and CesAs (CELLULOSE SYNTHASE As), encoding two positive regulators of cold signaling in maize (Zeng et al., 2021).

The basic leucine zipper (bZIP) transcription factor family plays multiple roles in plant responses to abiotic stress. For example, abscisic acid (ABA) response element (ABRE) binding factors (ABFs), members of the bZIP family of transcription factors, function as core regulators in the ABA signaling pathway in Arabidopsis; these proteins are involved in plant responses to drought, osmotic, and salt stress (Yoshida et al., 2010, 2015; Fernando et al., 2018). Several bZIP members are involved in the cold response in rice (Oryza sativa). Notably, one functional polymorphism in bZIP73 determines the diversity in cold tolerance between japonica and indica cultivars that arose during artificial selection (Liu et al., 2018). bZIP73 interacts with bZIP71 and enhances cold tolerance by modulating ABA biosynthesis and ROS homeostasis (Liu et al., 2018). The maize genome harbors 125 bZIP genes encoding 170 distinct bZIP proteins, only a few of which have been reported to respond to abiotic stress (Wei et al., 2012). Zea mays bZIP4 and ABP9 (ABRE BINDING PROTEIN9) contribute to plant drought and salt stress responses (Zhang et al., 2011; Ma et al., 2018), while bZIP60 mediates the unfolded protein response during heat stress (Li et al., 2020). Nevertheless, the role of bZIP transcription factors in regulating cold tolerance in maize remains elusive.

In this study, we identified the bZIP68 transcription factor as a negative regulator of cold tolerance in maize. Transcriptome analysis revealed that bZIP68 acts mainly as a transcriptional repressor of a set of COR genes. In particular, the cold-inducible expression of DREB1 family genes is downregulated by bZIP68, which directly binds to the A-box/G-box in the DREB1.7 promoter. Furthermore, bZIP68 is phosphorylated by MPK8, a negative regulator of cold tolerance in maize, thereby promoting bZIP68 protein stability and DNA binding affinity under cold stress. Importantly, the bZIP68 locus showed a strong selective signal following maize domestication. Variants of a 358-bp insertion/deletion (Indel-972) polymorphism in the bZIP68 promoters of maize and teosinte significantly affect the transcription of bZIP68. This study thus provides a genetic resource for breeding cold-resistant maize varieties by editing the maize bZIP68 promoter.

Results

bZIP68 negatively regulates cold tolerance in maize

In a reverse genetic screen of a previously described population of transgenic maize plants overexpressing more than 700 maize genes (Zeng et al., 2021; Supplemental Data Set S1), we identified several genes that altered cold tolerance in maize. One gene encodes the bZIP transcription factor bZIP68. Transgenic lines overexpressing bZIP68 driven by the Ubiquitin (Ubi) promoter displayed a cold-sensitive phenotype compared with the wild-type (WT) LH244 inbred line after exposure to 4�C for 3–4 days (Figure�1, A and B). Ion leakage is an indicator of stress-induced plasma membrane damage. After cold treatment, the ion leakage of bZIP68-overexpressing lines was much higher than that of the WT (Figure�1C). A drop in ambient temperature is typically associated with increased osmolality of plant cells, making osmolality an indicator of the cell freezing point and plant cold tolerance (Kaplan et al., 2004). In agreement with their compromised freezing tolerance, osmolality was markedly lower in bZIP68-overexpressing plants relative to the WT (Figure�1D). We also generated transgenic maize plants overexpressing bZIP68 with a MYC tag and driven by the Ubi promoter (bZIP68-MYC #1 and bZIP68-MYC #2) (Figure�1E); these lines were hypersensitive to cold stress compared with the WT (Figure�1, F and G), indicating that the addition of the MYC tag does not interfere with protein function.

Figure 1.

bZIP68 negatively regulates cold tolerance in maize. A, Relative bZIP68 expression levels in the WT LH244 and 14-day-old bZIP68-overexpressing seedlings (OE1, OE2), as determined by RT-qPCR. Levels in the WT were set to 1.0. Data represent the mean of three independent experiments � sd. Different letters represent significant differences (P < 0.05, one-way ANOVA). B–D, Cold phenotypes (B), ion leakage (C), and osmolality (D) of WT and bZIP68-OE lines. Fourteen-day-old seedlings grown at 25�C were exposed to 4�C for 3 days and allowed to recover at 25�C for 24 h. E, bZIP68 protein abundance in bZIP68-MYC-overexpressing seedlings (#1, #2). bZIP68 was detected with anti-MYC antibody. HSP82 served as a control. F and G, Cold phenotypes (F) and ion leakage (G) of WT and bZIP68-MYC seedlings. H, Schematic diagram of bzip68-1 and bzip68-2 generated by CRISPR/Cas9-mediated genome editing. I–K, Cold phenotypes (I), ion leakage (J), and osmolality (K) of WT, bzip68-1, bzip68-2, and the F₁ progeny of the bzip68-1 � bzip68-2 cross. Fourteen-day-old seedlings grown at 25�C were treated at 4�C for 4 days and allowed to recover at 25�C for 24 h. L and M, bZIP68 protein abundance under cold stress. Fourteen-day-old bZIP68-MYC seedlings were kept at 25�C or treated at 4�C for the indicated time. bZIP68 protein was detected with anti-MYC antibody. M, Quantification of immunoblot results shown in (L) using ImageJ software. bZIP68 protein levels were normalized to HSP82. Protein levels in bZIP68-MYC seedlings grown at 25�C (0 h) were set to 1.0. Data represent the mean of three independent experiments � sd (**P < 0.01, Student’s t test). In (B, F, and I), representative pictures are shown. In A, C, D, G, J, and K, data represent the mean of three independent experiments � sd (n = 3 seedlings per replicate). Different letters represent significant differences (P < 0.05, one-way ANOVA).

We generated loss-of-function mutants of bZIP68, named bzip68-1 and bzip68-2, using the CRISPR/Cas9 genome editing system (Xing et al., 2014). The bzip68-1 mutant harbored a 169-bp deletion (from 521- to 689-bp downstream of the translation start site ATG) and bzip68-2 carried a 1-bp insertion (519-bp downstream of the ATG) as well as a 21-bp deletion (from 686 to 706 bp downstream of the ATG) in bZIP68 (Figure�1H andSupplemental Figure S1A), leading to frameshifts in the open-reading frame and premature termination of translation. We selected homozygous bzip68-1 and bzip68-2 mutant lines without the Cas9 transgene for further study. In contrast to bZIP68-overexpressing lines, bzip68-1 and bzip68-2 showed a cold-tolerant phenotype compared with the WT (Figure�1I), which was accompanied by reduced ion leakage and greater osmolality (Figure�1, J and K). We crossed bzip68-1 with bzip68-2 to generate F₁ progeny (Supplemental Figure S1A), which all exhibited a cold tolerance phenotype comparable to that of the single mutants (Figure�1, I–K), indicating that the loss of bZIP68 function results in increased cold tolerance in maize seedlings.

To further clarify whether bZIP68 affects the germination ability of maize at low temperature, we performed germination experiments at 25�C and 12�C. There was no significant difference in germination between WT and bzip68 at 25�C, while the germination rates of the bzip68 mutants were slightly higher than the WT when grown at 12�C. In contrast, the germination rates of bZIP68-overexpressing lines were significantly lower than the WT at 25�C and even lower at 12�C (Supplemental Figure S1B). These results indicate that bZIP68 plays a negative role in maize germination under normal and cold conditions. Taken together, these data suggest that bZIP68 negatively regulates the cold tolerance of maize at both the germination and seedling stages.

The C-terminus of bZIP68 contains the bZIP domain, which consists of a basic DNA-binding region and a leucine zipper domain (Jakoby et al., 2002). Using a yeast (Saccharomyces cerevisiae) one-hybrid assay, we determined that the bZIP68 transcriptional activation domain is located in the middle region (amino acids [aa] 138–275) (Supplemental Figure S1C), which is rich in proline and acidic residues (Droge-Laser et al., 2018). Consistent with its function as a transcription factor, a fusion protein between green fluorescent protein (GFP) and bZIP68 localized to the nucleus in the leaves of transgenic maize plants expressing bZIP68-GFP driven by the Ubi promoter (Supplemental Figure S1D). Furthermore, phylogenetic analysis indicated that bZIP68 is closely related to rice TRANSCRIPTION FACTOR RESPONSIBLE FOR ABA REGULATION1 (OsTRAB1) and Arabidopsis ABF2 (Droge-Laser et al., 2018; Supplemental Figure S1E and Supplemental File S1). The expression of bZIP68 and its homologous genes bZIP49, bZIP123, and bZIP4 was not induced by cold stress (Supplemental Figure S1, E and F). Interestingly, immunoblot analysis showed that bZIP68 protein abundance gradually increased after 12 h of cold treatment in Ubi:bZIP68-MYC transgenic plants (Figure�1, L and M). These data indicate that cold stress regulates bZIP68 at the protein level.

MPK8 interacts with and phosphorylates bZIP68

Previous studies showed that the phosphorylation of bZIPs affects their protein stability (Sirichandra et al., 2010). We performed amino acid sequence alignment of bZIP68 and its closest homologs in maize (bZIP49, bZIP123, and bZIP4), rice (TRAB1), and Arabidopsis (ABF2). The results revealed three conserved regions in the N-terminus of bZIP68. These regions contain phosphorylation sites (R-X-X-S/T, ST-X-X-ED) of SnRK2 and CDPK protein kinases (Choi et al., 2005; Supplemental Figure S2). To examine whether the cold-induced increase in bZIP68 abundance is due to phosphorylation, we immunoprecipitated bZIP68 from bZIP68-MYC overexpressing transgenic maize plants with an anti-MYC antibody, followed by immunoblotting with an anti-phosphotyrosine antibody. We detected a phosphorylation signal at the correct molecular weight for bZIP68-MYC (∼60 kDa), which was abolished by λ-PPase treatment (Supplemental Figure S3A), suggesting that bZIP68 is phosphorylated in planta. We predicted phosphorylation sites with the online tool NetPhos3.1a. The result showed that there were several phosphorylation sites of SnRKs and CDPKs in the conserved regions of the N-terminus of bZIP68 and two putative conserved MPK phosphorylation sites (SP/TP) that are close to the bZIP domain among different plant species (Supplemental Figure S3B).

Since we previously showed that MPKs are involved in cold tolerance in maize (Zeng et al., 2021), we screened MPKs for potential interaction with bZIP68 by yeast two-hybrid assay: Two clade C MPK subfamily members, MPK2 and MPK8, interacted with bZIP68 (Figure�2A andSupplemental Figure S3C). The transcriptional activation domain of bZIP68 (aa 138–275) was sufficient for these interactions (Figure�2B). Because MPK2 and MPK8 are close homologs in maize (Supplemental Figure S3D) and both negatively regulate cold tolerance (Zeng et al., 2021), we chose MPK8 for further study. We performed a glutathione S-transferase (GST) in vitro pull-down assay using the recombinant proteins His-bZIP68 (His-tagged ZIP68) and GST-MPK8 or GST alone. Immunoblotting showed that His-bZIP68 was pulled down by GST-MPK8 but not by GST (Figure�2C), suggesting that bZIP68 interacts with MPK8 in vitro.

Figure 2.

MPK8 interacts with and phosphorylates bZIP68. A, MPK2 and MPK8 interact with bZIP68 in yeast. Yeast cells were grown on synthetic defined (SD) medium lacking Leu and Trp (–Leu –Trp) or SD –Leu –Trp –His –Ade medium. B, MPK8 interacts with the middle region (aa 138–275) of bZIP68 in yeast. C, In vitro pull-down assay showing the interaction between bZIP68 and MPK8. Recombinant GST and GST-MPK8 proteins were immobilized on glutathione agarose beads and incubated with His-bZIP68 protein. bZIP68 was detected with anti-His antibody. D, BiFC assay showing the interaction between bZIP68 and MPK8 in N. benthamiana leaf cells. The YFP signal was visualized by confocal microscopy. Scale bars, 30 μm. E, Co-IP assay showing the interaction between MPK8 and bZIP68 in N. benthamiana leaves. Total proteins were immunoprecipitated with anti-GFP agarose beads. Co-immunoprecipitated MPK8-MYC was detected with anti-MYC antibody. F, MPK8^Y113C phosphorylates bZIP68 but not bZIP68S^250A in vitro. Recombinant MBP-bZIP68 and GST-MPK8^Y113C proteins were subjected to an in vitro phosphorylation assay. The autoradiogram (top) and the Coomassie Brilliant Blue-stained gel (bottom) are shown. G, In-gel kinase assay showing the kinase activity of MPK8 activated by cold treatment. Total proteins were extracted from 14-day-old seedlings of the WT, mpk8-1, and MPK8-OE2 treated at 4�C for 0, 6, and 12 h. Recombinant MBP-bZIP68 was used as the substrate. MPK8 kinase activity was detected by autoradiography. HSP82 served as a control.

To confirm the interaction between bZIP68 and MPK8 in planta, we performed a bimolecular fluorescence complementation (BiFC) assay in Nicotiana benthamiana. We observed the reconstitution of yellow fluorescence protein (YFP) in the nuclei of epidermal cells when co-infiltrated with bZIP68-YFP^N and MPK8-YFP^C constructs, but not when bZIP68-YFP^N was co-infiltrated with MPK11-YFP^C (Figure�2D), an MPK that did not interact with bZIP68 in yeast (Supplemental Figure S3C). We also performed a co-immunoprecipitation (co-IP) assay in N. benthamiana leaves co-infiltrated with MPK8-MYC and bZIP68-GFP constructs. We successfully co-immunoprecipitated MPK8-MYC with bZIP68-GFP, but not with control GFP alone (Figure�2E). Therefore, bZIP68 interacts with MPK8 in vitro and in vivo.

The interaction of bZIP68 with a kinase prompted us to determine if this transcription factor might be phosphorylated by MPK8, which we tested using an in vitro kinase assay. We previously showed that MPK8^Y113C (harboring a substitution of Tyr113 with Cys) is a constitutively active form of MPK8 (Berriri et al., 2012; Zeng et al., 2021). We therefore used GST-MPK8^Y113C for the assay. The online tool NetPhos 3.1a predicted putative MPK phosphorylation sites (SP/TP) at residues Ser229 and Ser250, located within the interaction interface between bZIP68 and MPK8 (Supplemental Figure S3B). Indeed, MPK8^Y113C phosphorylated both recombinant maltose binding protein (MBP) fused to ZIP68 (MBP-bZIP68) and MBP-bZIP68^S229A (with Ser229 mutated to Ala to prevent phosphorylation) (Figure�2F). In contrast, we detected no phosphorylation signal when Ser250 or Ser229 and Ser250 were mutated to Ala (Figure�2F). Liquid chromatography–tandem mass spectrometry (LC-MS/MS) of the recombinant proteins confirmed that Ser250 of MBP-bZIP68 is phosphorylated by MPK8 (Supplemental Figure S3E). These results indicate that MPK8 phosphorylates bZIP68 at Ser250 in vitro.

Furthermore, we performed in-gel kinase assays using protein extracts from WT, MPK8-OE, and mpk8-1 maize seedlings treated at 4�C, with recombinant MBP-bZIP68 protein as the substrate. The activity of protein kinases of ∼40 kDa (the molecular size of MPK8 is 42.2 kDa) was enhanced in extracts from WT seedlings exposed to cold treatment. Strikingly, this cold-induced increase in kinase activity markedly increased in MPK8-OE plants but was largely lost in the mpk8-1 mutant (Figure�2G). These data demonstrate that cold treatment facilitates the phosphorylation of bZIP68 by MPK8 in planta.

MPK8-mediated phosphorylation of bZIP68 decreases cold tolerance in maize

To further explore the biological relevance of the interaction between MPK8 and bZIP68, we performed a cell-free degradation assay to investigate whether MPK8 affects bZIP68 stability by phosphorylation. In the presence of ATP, recombinant MBP-bZIP68 was degraded more slowly in protein extracts from WT seedlings than from mpk8-1 cell extracts but more rapidly than in MPK8-OE cell extracts (Figure�3, A and B). Moreover, the recombinant mutated MBP-ZIP68^S250A variant protein was more unstable than MBP-ZIP68 in cell extracts from MPK8-OE seedlings. In addition, the 26S proteasome-specific inhibitor MG132 blocked the degradation of MBP-bZIP68 (Figure�3, A and B). These data suggest that MPK8-mediated phosphorylation of bZIP68 prevents its 26S proteasome-dependent degradation. To assess whether MPK8 regulates bZIP68 protein stability in planta, we co-transfected maize leaf protoplasts with bZIP68-GFP or bZIP68^S250A-GFP constructs and different amounts of MPK8-MYC plasmid DNA. bZIP68 protein abundance increased with increasing amounts of MPK8-MYC DNA when co-transfected with bZIP68-GFP, but not with bZIP68^S250A-GFP (Supplemental Figure S4, A–C).

Figure 3.

MPK8 enhances the stability of bZIP68. A and B, In vitro cell-free degradation assay showing that MPK8 stabilizes bZIP68. Total proteins extracted from 14-day-old WT, mpk8-1, and MPK8-OE2 seedlings were incubated with recombinant MBP-bZIP68 or MBP-bZIP68^S250A protein in the presence of 1-mM ATP. MBP-bZIP68 was detected with anti-MBP antibody. HSP82 served as a control. Protein levels without treatment were set to 1.0. The immunoblot results were quantified using ImageJ software. C and D, bZIP68 protein abundance in bZIP68-MYC and mpk8 bZIP68-MYC seedlings without (C) or with (D) 4�C treatment. bZIP68 was detected with anti-MYC antibody (C). The immunoblot results were quantified using ImageJ software. bZIP68 protein levels were normalized to HSP82. Protein levels without cold treatment were set to 1.0. E–G, Relative bZIP68 expression levels (E), cold phenotypes (F), and ion leakage (G) of WT, bZIP68-OE2, and bZIP68^S250A (#1, #2) seedlings. H and I, Cold phenotypes (H) and ion leakage (I) of WT, mpk8, bZIP68-MYC, and mpk8 bZIP68-MYC seedlings. In B and D, data represent the mean of three independent experiments � sd (*P < 0.05, **P < 0.01, Student’s t test). In E, G, and I, data represent the mean of three independent experiments � sd (n = 3 seedlings per replicate). Different letters represent significant differences (P < 0.05, one-way ANOVA).

We previously showed that MPK8 negatively regulates the cold tolerance of maize (Berriri et al., 2012; Zeng et al., 2021). We generated mpk8 loss-of-function mutants via CRISPR/Cas9 gene editing: mpk8-1 harbored a 1-bp insertion and mpk8-3 a 2-bp deletion, both in the first exon (Supplemental Figure S4, D and E). Both mpk8 mutants showed enhanced cold tolerance, as expected (Supplemental Figure S4, F–H). We crossed mpk8-3 with bZIP68-MYC#1-overexpressing plants to obtain mpk8 bZIP68-MYC lines (Supplemental Figure S4I). Under normal growth conditions (25�C), bZIP68 protein abundance was relatively stable in both genotypes (Figure�3, C and D). However, upon cold treatment, bZIP68 accumulated to higher levels in bZIP68-MYC transgenic seedlings, but not in mpk8 bZIP68-MYC lines (Figure�3, C and D). These results suggest that MPK8 promotes the stability of bZIP68 under cold stress.

To explore the role of bZIP68 phosphorylation in vivo, we generated transgenic lines accumulating the nonphosphorylatable form of this protein, bZIP68^S250A. We selected two independent transgenic lines that showed transcript levels comparable to those of the bZIP68-OE2 line for cold phenotypic assays (Figure�3E). In contrast to the hypersensitivity of bZIP68-OE2 seedlings to cold stress, bZIP68^S250A transgenic seedlings largely resembled the WT in terms of cold tolerance and ion leakage (Figure�3, F and G). Consistent with this result, the protein level of bZIP68^S250A did not increase but instead decreased under cold treatment (Supplemental Figure S4, J and K). Together, these results suggest that the phosphorylation of bZIP68 at Ser250 impairs cold tolerance in maize.

To dissect the genetic relevance of bZIP68 and MPK8 in regulating cold tolerance, we compared the cold tolerance of mpk8-3, bZIP68-MYC, and mpk8 bZIP68-MYC seedlings. After cold treatment, mpk8-3 showed enhanced cold tolerance, whereas bZIP68-MYC seedlings were hypersensitive to cold stress compared with the WT (Figure�3, H and I). mpk8 bZIP68-MYC seedlings largely phenocopied bZIP68-MYC seedlings in terms of cold sensitivity and ion leakage (Figure�3, H and I). These data suggest that bZIP68 functions downstream of MPK8 to negatively regulate cold tolerance in maize.

Identification of bZIP68-regulated COR genes by transcriptome analysis

To identify the potential target genes of bZIP68 that are responsible for the cold response, we conducted transcriptome deep sequencing (RNA-seq) analysis of 14-day-old WT and bzip68-1 seedings grown at 25�C or exposed to 4�C for 12 h in three independent experiments. We first performed a principal component analysis (PCA) and verified the repeatability among the replicates (Supplemental Figure S5). We then identified differentially expressed genes (DEGs) based on the criteria of a significant difference (P < 0.05) with an absolute fold-change ≥ 2. A total of 3,742 genes were identified in WT seedlings after exposure to cold (4�C–12 h) compared with seedlings before cold treatment (4�C–0 h), among which 2,024 (54%) were upregulated and 1,718 (46%) were downregulated (Figure�4A andSupplemental Data Set S2).

Figure 4.

Transcriptome analysis of bZIP68-dependent COR genes. A, Volcano plots representing the fold-change of DEGs in the comparison groups of WT–0 h versus WT–4�C–12 h, bzip68-0 h versus WT–0 h, bzip68–12 h versus WT–12 h, and WT–4�C–12 h versus WT–25�C–12 h (P < 0.05, absolute fold change ≥ 2.0). Gray dots represent genes without significant changes in expression. Three independent experiments were performed for each sample at each time point. B, Venn diagrams representing DEGs regulated by cold (top, left), bZIP68 (top, right), and both cold and bZIP68 (bottom). C, Clustering analysis of bZIP68-dependent COR genes. Heatmap shows the Log₂ (TPM) values of 1,339 DEGs. DEGs with similar expression patterns were clustered with the R package Pheatmap. D, GO analysis showing 10 of the top 20 enrichment terms of Clusters 2 and 3 shown in C, which was performed using Phyper (FDR < 0.05). Bubble charts show the GO terms. E, Heatmap showing the Log₂ (TPM) of enriched genes in the GO term Regulation of transcription, DNA-templated in Cluster 3 shown in D. F, Relative DREB1 expression in WT, bZIP68-OE, and bzip68 seedlings before and after 4�C treatment, as determined by RT-qPCR. Expression in the untreated WT was set to 1.00. Each bar represents the mean of three independent experiments � sd. Different letters represent significant differences (P < 0.05, one-way ANOVA).

To exclude possible effects from circadian rhythms, we compared the RNA-seq data of the WT treated at 4�C for 12 h versus plants kept at 25�C for 12 h and identified 4,079 DEGs that were more strongly affected by cold than photoperiod (Figure�4A andSupplemental Data Set S3). By comparing the 3,742 and 4,079 DEGs, we defined 2,383 overlapping DEGs as COR genes, including 1,412 upregulated genes and 981 downregulated genes (Figure�4B andSupplemental Data Set S4). To identify bZIP68-regulated DEGs, we compared the RNA-seq results from the WT and bzip68-1 with or without cold treatment (4�C) for 12 h, leading to the identification of 5,064 DEGs regulated by bZIP68; these genes included 3,561 DEGS from bzip68 versus WT without cold treatment and 2,265 from bzip68 versus WT after 12 h of cold treatment (Figure�4, A and B and Supplemental Data Sets S5 and S6). A comparison of 2,383 and 5,064 DEGs defined a total of 1,339 common genes that are regulated by both cold and bZIP68 (Figure�4B andSupplemental Data Set S7), which accounts for 56% of COR genes and 26% of bZIP68-regulated genes.

Cluster analysis grouped these 1,339 bZIP68-regulated COR genes into five clusters (Clusters 1–5) (Figure�4C and Supplemental Data Set S8). We performed gene ontology (GO) analysis of each cluster to explore the biological processes involving bZIP68. Only Clusters 2 and 3 have significantly enriched GO terms. We listed 10 of top 20 most significantly enriched terms. Cluster 2 is enriched in biological processes involved in the regulation of transcription, response to different stresses, and jasmonic acid signaling, and Cluster 3 is mainly involved in the regulation of transcription and various cellular processes (Figure�4D). Interestingly, high proportions of genes in Clusters 2 and 3 were enriched in the regulation of transcription (Figure�4D). Based on these results, we analyzed DEGs enriched in regulation of transcription in Cluster 3 by constructing heatmaps (using transcripts per million [TPM] values). Many different types of transcription factors are enriched in this term, including DREB1s, MYBs, bZIPs, bHLHs, and WRKYs (Figure�4E). DREB1 genes encode transcription factors that play crucial roles in cold stress responses in different plant species (Thomashow, 1999; Shi et al., 2018; Zeng et al., 2021). Notably, the heatmaps showed that the expression of DREB1 genes within Cluster 3 in the WT, including DREB1.1, DREB1.2, DREB1.3, DREB1.4, DREB1.5, DREB1.7, and DREB1.10, was highly induced by cold treatment, which even more pronounced in bzip68 under cold stress (Figure�4E). These data suggest that bZIP68 negatively regulates cold tolerance by modulating the expression of a set of COR genes, including DREB1 genes.

DREB1 genes are direct targets of bZIP68

RNA-seq data showed that a dozen DREB1 genes are regulated by bZIP68. We confirmed the upregulation of DREB1.7 and DREB1.10 expression in the bzip68 mutants and their downregulation in bZIP68 overexpression lines compared with the WT after cold treatment by reverse transcription quantitative polymerase chain reaction (RT-qPCR) (Figure�4F). Next, we examined the cold sensitivity of a weak dreb1.7 allele (named Δpro) harboring a 22-bp deletion in the DREB1.7 promoter (Xiao et al., 2020). The Δpro mutant showed lower DEREB1.7 expression, as expected, as well as greater cold sensitivity than the corresponding WT (Supplemental Figure S6). Combined with our recent findings on DREB1.10 (Zeng et al., 2021), these data confirm the key roles of DREB1s in promoting cold tolerance in maize.

The bZIP domain of bZIP-type transcription factors preferentially binds to motifs such as the A-box (TACGTA), C-box (GACATC), and G-box (CACGTG) elements of downstream target genes (Choi et al., 2000). Motif analysis identified a set of conserved core A-/G-/C-boxes in the promoter regions of these DREB1s, suggesting they are potential targets of bZIP68. To test this possibility, we performed chromatin immunoprecipitation (ChIP) assays with an anti-GFP antibody using 14-day-old bZIP68-GFP seedlings with or without cold treatment at 4�C for 12 h, followed by qPCR, using DREB1.7 as an example. Of the five putative binding sites (P1–P5) in the DREB1.7 promoter, ChIP-qPCR demonstrated that bZIP68-GFP bound to the P1 (A-box), P2 (G-box), and P5 (C-box) sites of the DREB1.7 promoter in bZIP68-GFP samples, but not in WT ChIP samples (Figure�5A). The fold enrichment at the P1, P2, and P5 sites substantially rose after cold treatment (Figure�5A), indicating that cold treatment promotes the binding of bZIP68 to the DREB1.7 promoter. We then performed electrophoretic mobility shift assays (EMSAs), which established that MBP-bZIP68 directly binds to the conserved P1 motif in the DREB1.7 promoter (Figure�5B). Increasing amounts of an unlabeled WT probe markedly abolished bZIP68 binding to the biotin-labeled probe, but the mutated unlabeled probe failed to compete for binding (Figure�5B). These results demonstrate that bZIP68 indeed directly binds to the DREB1.7 promoter.

Figure 5.

MPK8-mediated bZIP68 phosphorylation promotes its binding to the DREB1.7 promoter. A, ChIP assay showing the binding of bZIP68 to the DREB1.7 promoter in vivo. bZIP68-GFP seedlings were exposed to cold treatment at 4�C for 12 h or remained at 25�C. ChIP was performed using anti-GFP antibody. Immunoprecipitated DNA was quantified by qPCR using primers specific to regions within the DREB1.7 locus (P1–P5). Relative enrichment is represented as input (%). Data are shown as the mean of three independent experiments � sd. Different letters for each binding site represent significant differences (P < 0.05, one-way ANOVA). B, EMSA showing the binding of bZIP68 to the A-box in the DREB1.7 promoter. C, Dual-LUC assay showing that bZIP68 negatively regulates DREB1.7 transcription in maize protoplasts. Protoplasts co-transfected with Super:bZIP68-GFP and pDREB1.7:LUC were treated with or without 4�C for 1 h. 35S:REN was used as the internal control. Each bar represents the mean of three independent experiments � sd. Different letters represent significant differences (P < 0.05, one-way ANOVA). D, The effect of MPK8 on the transactivational activity of bZIP68 in N. benthamiana leaves. 35S:LUC was used as an internal control. Relative GUS/LUC value indicates the DREB1.7 expression level. pDREB1.7:GUS co-transfected with Super:HA-FLAG and Super:MYC was used as a control. HA-FLAG-bZIP68 (HF-bZIP68), HF-bZIP68^S250A, and MPK8-MYC were detected with anti-HA and anti-MYC antibodies. Each bar represents the mean of three independent experiments � sd. Different letters represent significant differences (P < 0.05, one-way ANOVA). E, EMSA showing that the binding affinity of bZIP68 to the DREB1.7 promoter is enhanced by MPK8. Recombinant GST-MPK8^Y113C was incubated with MBP-bZIP68 in kinase reaction buffer at 30�C for 30 min with or without 50-μM ATP. A biotin-labeled DREB1.7-P1 DNA fragment was incubated with MBP-bZIP68, MBP-bZIP68^S250A, or phosphorylated MBP-bZIP68 protein. MBP-bZIP68, MBP-bZIP68^S250A, and GST-MPK8^Y113C were detected by CBB staining. Quantitative data are shown as the mean of three independent experiments � sd. Different letters represent significant differences (P < 0.05, one-way ANOVA).

We also performed transient transactivation assays using a 1-kb promoter fragment of DREB1.7 in a dual luciferase (LUC) reporter system (Figure�5C). Accordingly, we co-transfected maize leaf protoplasts with the bZIP68-GFP effector construct and the pDREB1.7:LUC reporter construct. DREB1.7 transcription was induced by cold treatment, as expected. However, co-transfection of the LUC reporter with bZIP68-GFP repressed DREB1.7 expression at 25�C and this repression was even more pronounced after cold treatment (Figure�5C). Taken together, these results suggest that bZIP68 negatively regulates DREB1.7 transcription by directly binding to its promoter.

MPK8-mediated phosphorylation of bZIP68 enhances its binding to DNA

MPK8 was recently shown to be a negative regulator of DREB1 family gene expression (Zeng et al., 2021). Thus, we determined whether the phosphorylation of bZIP68 by MPK8 contributed to the transcriptional activity of bZIP68 on DREB1 genes. We performed a transient transactivation assay in N. benthamiana leaves using the DREB1.7 promoter driving the β-GLUCURONIDASE (GUS) reporter gene. The bZIP68 and MPK8 effector constructs were placed under the control of the Super promoter. Consistent with the above results (Figure�5C), the transient expression of bZIP68 significantly suppressed DREB1.7 expression (Figure�5D). Furthermore, the suppression of DREB1.7 expression mediated by bZIP68 was significantly enhanced in the presence of MPK8. In contrast, the bZIP68^S250A mutant largely alleviated the transcriptional repression of DREB1.7 compared with WT bZIP68 with or without MPK8 (Figure�5D). These results suggest that phosphorylation of bZIP68 by MPK8 is a prerequisite for the transcriptional repression of DREB1.7 by bZIP68.

We used competitive EMSA to examine whether MPK8 affected the association of bZIP68 to the DREB1.7 promoter. After incubation with MPK8^Y113C in kinase reaction buffer in the presence of ATP, we observed an increase in the binding affinity of MBP-bZIP68 toward the DREB1.7 promoter (Figure�5E). In contrast, the DNA binding ability of bZIP68^S250A drastically decreased upon incubation with MPK8^Y113C and ATP (Figure�5E). Combined with the above results, we conclude that the phosphorylation of bZIP68 at Ser250 by MPK8 modulates both its stability and DNA binding affinity.

bZIP68 was selected during maize domestication

Nucleotide diversity in maize reflects the history of human selection and migration (Buckler and Thornsberry, 2002; Yamasaki et al., 2007). A previous study revealed a set of candidate domestication and improvement genes that are directly related to maize evolution and breeding programs. bZIP68 is located in a ∼27.5-kb selected region of chromosome 4 (Hufford et al., 2012) and this region only contains this gene, suggesting that bZIP68 was likely selected during maize domestication. The nucleotide diversity of upstream regions of bZIP68 decreased during maize domestication (Hufford et al., 2012). We therefore re-sequenced the promoters and 5′-UTRs of bZIP68 in 21 maize accessions and 17 wild progenitor teosinte accessions (including 12 Z. mays ssp. parviglumis and 5 Z. mays ssp. mexicana) using HITAC-seq (highly parallel indexed tagmentation-reads assembled consensus sequencing). Comparative analyses revealed extensive nucleotide variation in the bZIP68 promoter and 5′-UTR, including 9 Indels and 89 single-nucleotide polymorphisms (Supplemental Figure S7). Maize domestication was accompanied by a decrease in nucleotide diversity due to the domestication bottleneck and serial founder effects (Wang et al., 2017; Allaby et al., 2019). Indeed, re-sequencing data analysis showed that the nucleotide diversity (π) of bZIP68 was 0.02792 in teosinte and 0.00136 in maize (Figure�6A). Thus, maize retained only 4.9% (π_maize/π_teosinte) of the nucleotide diversity of teosinte, indicating that nucleotide diversity in the bZIP68 promoter region decreased significantly during maize domestication from teosinte. Furthermore, we identified a 358-bp insertion at position –972 bp (Indel-972, B73 genome) upstream of the start codon (ATG) of bZIP68 in all 21 maize accessions, but not in any teosinte accession we tested (Supplemental Figure S7). We then genotyped the bZIP68 promoter in 207 maize inbred lines originating in the tropics, subtropics, and temperate zones by PCR. The insertion was present in all 207 maize accessions we examined (Supplemental Figure S8 and Supplemental Data Set S9), further supporting the notion that this 358-bp indel was selected during the early stage of maize domestication.

Figure 6.

The bZIP68 locus was a target of selection during maize domestication. A, Resequencing analysis showing that the nucleotide diversity (π) is significantly reduced in maize relative to teosinte. The 1.5-kb bZIP68 promoter region was re-sequenced in diverse maize (21) and teosinte (17) lines. Population nucleotide diversity was analyzed by DnaSP version 5.0 software. B, Relative bZIP68 expression levels in bZIP68-NIL^maize, bZIP68-NIL^teosinte, and bZIP68-NIL^het. Each bar represents the mean of three independent experiments � sd. Different letters represent significant differences (P < 0.05, one-way ANOVA). C, Relative leaf injured area in bZIP68-NIL^maize, bZIP68-NIL^teosinte, and bZIP68-NIL^het lines after 4�C treatment. Data represent mean � sd (n = 11 for NIL^maize, n = 23 for NIL^teosinte, and n = 13 for NIL^het) and different letters represent significant differences (P < 0.05, one-way ANOVA). D, Transactivation activity of the bZIP68 promoter in maize protoplasts. Schematic diagram of the reporter constructs used in the transactivation assay with the 35S: LUC/REN fused to the 1,188-bp promoter fragments from teosinte and the 1,492/972-bp promoter fragments from maize. Each bar represents the mean of three independent experiments � sd. Different letters represent significant differences (P < 0.05, one-way ANOVA).

We speculated that bZIP68 transcript levels might have changed during maize domestication as a result of the InDel detected above. To test this hypothesis, we examined the expression of bZIP68 in homozygous and heterozygous recombinant inbred lines from a maize-teosinte BC₂S₃ population, with maize cultivar W22 and teosinte 8759 as the parents (Supplemental Figure S9A; Liang et al., 2019; Tian et al., 2019). The self-progeny of the heterozygous hybrid family (HIF) MR1077 line, a heterogeneous inbred family that is heterozygous at bZIP68, comprised bZIP68-NIL^maize, bZIP68-NIL^teosinte, and bZIP68-NIL^het (Supplemental Figure S9, B and C). Notably, the expression of bZIP68 was significantly higher in bZIP68-NIL^maize than in bZIP68-NIL^teosinte before and after cold treatment, with bZIP68 expressed at intermediate levels in the bZIP68-NIL^het lines (Figure�6B). In agreement with a role for bZIP68 in negatively regulating cold tolerance, the bZIP68-NIL^teosinte lines exhibited a significantly lower relative leaf injured area than bZIP68-NIL^maize or bZIP68-NIL^het, the latter two being phenotypically comparable (Figure�6C). These results indicate that bZIP68-NIL^teosinte is more tolerant to cold stress than the other two genotypes due to its lower bZIP68 transcript levels.

Finally, to further investigate the contribution of natural variation at the bZIP68 promoter, we cloned the promoter fragments (1.5-kb upstream of the ATG) from two maize inbred lines (B73 and W22) and two teosinte accessions upstream of the firefly LUC gene for transactivation activity assays in maize leaf protoplasts. LUC activity driven by the maize promoters was significantly higher than that from the teosinte promoters (Figure�6D). Importantly, introducing the equivalent 358-bp deletion of teosinte into the maize constructs greatly weakened transcriptional activation (Figure�6D). Taken together, these data suggest that bZIP68 underwent strong selection during domestication and that the 358-bp insertion resulted in the transcriptional activation of bZIP68 in maize.

Discussion

Maize originated from tropical and subtropical regions and is sensitive to cold stress. Therefore, it is vital to elucidate the underlying genetic and molecular basis of cold sensitivity in maize for crop improvement through genetics and breeding. In this study, we determined that bZIP68 negatively regulates cold tolerance in maize by modulating the expression of a set of COR genes. We showed that bZIP68 directly binds to the promoters of DREB1 genes, encoding important transcriptional activators of cold tolerance, to repress their expression during the late stage of the cold response in maize (Supplemental Figure S10A), possibly to prevent over-response to cold stress caused by the excess accumulation of DREB1. We also demonstrated that MPK8 positively regulates the stability and transcriptional activity of bZIP68 via direct phosphorylation at Ser250. Intriguingly, the bZIP68 genomic region was the subject of strong selection during domestication from teosinte to maize. In agreement with this finding, we observed an increase in bZIP68 expression in maize cultivars compared with teosinte. Further natural variation analysis among maize inbred lines and teosintes identified a 358-bp insertion (Indel-972) in the bZIP68 promoter region that appears to have acted as an enhancer to promote bZIP68 transcription in maize (Figure�7).

Figure 7.

Model of the role of bZIP68 in regulating the cold response in maize. Domestication led to expression diversity of bZIP68 in maize and the ancestor teosinte. A 358-bp insertion in the bZIP68 promoter results in a continuously high bZIP68 expression level and cold susceptibility in maize. Post-translational modification of bZIP68 is crucial for its regulation of cold stress responses in maize. Under cold stress, MPK8 interacts with and phosphorylates bZIP68 at Ser250 to promote its stability and DNA binding affinity. The transcriptional repression of COR genes (e.g. DREB1s) by bZIP68 is enhanced, especially at the late stage of the cold stress response, thereby negatively regulating cold tolerance in maize.

Plants respond to changes in their immediate environment by activating signal transduction pathways and gene expression programs (Gong et al., 2020). Most signaling pathways relay extracellular and subcellular information to transcription factors to initiate stress responses (Chen et al., 2021). In Arabidopsis and rice, protein kinases, such as SNF1-RELATED PROTEIN KINASE 2.6 (SnRK2.6), MPK3/6, and BRASSINOSTEROID-INSENSITIVE2 (BIN2) have been shown to interact with and phosphorylate transcription factors to regulate their stability and plant cold tolerance (Ding et al., 2015; Li et al., 2017; Zhang et al., 2017; Keyi Ye, 2019). We previously demonstrated that MPK8 phosphorylates and promotes the degradation of type-A RR1 in the cytosol to inhibit cold-induced DREB1 expression and cold tolerance in maize (Zeng et al., 2021). Here, we showed that MPK8 phosphorylates bZIP68 in the nucleus to repress the expression of DREB1 genes. The identification of this kinase cascade-transcription factor regulatory module reveals the molecular basis of cold sensitivity in maize and thus provides helpful information for molecular breeding of crops.

Post-translational modifications are crucial for the physiological roles of bZIP transcription factors in plants. For instance, Arabidopsis SnRKs phosphorylate the bZIP family members ABFs to enhance their transcriptional activity or protein stability in response to abiotic stress (Kobayashi et al., 2005; Furihata et al., 2006; Sirichandra et al., 2010). Here, we demonstrated that the phosphorylation of bZIP68 at Ser250 by MPK8 not only improves its protein stability, but also promotes its DNA binding affinity in maize. Interestingly, the Ser250 residue is close to the DNA binding domain (basic region) and is located within a pocket structure that was predicted to be a likely metal ion binding domain using Alpha Fold (http://alphafold.ebi.ac.uk/). Magnesium (Mg²⁺) significantly enhances the DNA binding ability of the bZIP transcription factor CYCLIC-AMP RESPONSE ELEMENT BINDING PROTEIN (CREB) to promote its homodimerization in mammals (Schumacher et al., 2000). Hence, it is possible that phosphorylation at S250 affects the charge properties of bZIP68, which is conducive to attracting the binding of metal ions within bZIP68 to enhance its DNA binding activity.

Considering that bZIP68 is a negative regulator of cold tolerance in maize, it is important to identify the natural variations present in bZIP68 to distinguish cold-tolerant and cold-sensitive alleles. In this study, we showed that bZIP68^teosinte confers greater tolerance to cold stress than bZIP68^maize. Moreover, bzip68 mutants did not display significant differences in yield traits compared with the WT (Supplemental Figure S10, B–D). Therefore, the bZIP68^teosinte allele represents an excellent natural resource for improving cold tolerance in maize to enhance adaptation to high latitude regions.

During the process of maize domestication, the planting area gradually expanded from low latitudes to high latitudes, where maize gained cold tolerance to adapt to a cold environment. Interestingly, we found that the bZIP68 locus was a target of selection during early domestication; however, the cold-tolerant�elite allele of bZIP68 was not selected. Therefore, this selection signal may not be beneficial for cold adaptation in maize at high latitudes. Maize was domesticated through natural and human selection, which led to a stark decrease in genetic diversity due to the excessive pursuit of yield traits, while being accompanied by the accumulation of causal harmful alleles (Doebley, 2004; Huang and Han, 2012; Wang et al., 2017). We did not observe obvious differences in hundred-grain weight, plant height, or ear height coefficient between the bzip68 mutants and WT (Supplemental Figure S10, B–D), suggesting that the selection of bZIP68 was not associated with these yield-related traits. Since transcriptome analysis indicated that a large set of genes is regulated by bZIP68, it is possible that ZmbZIP68 was selected directly or indirectly with some other yet unknown agronomic traits during domestication. It is worth noting that only the bZIP68 gene is present in this selection region, so the retention of the cold-sensitive allele of bZIP68 might not be a hitch-hiking effect due to selection at linked genes during maize early domestication. Although the underlying mechanism of selection at the bZIP68 locus remains unclear, some cold-tolerant favourable alleles, including the bZIP68^teosinte allele, indeed exist in plants in their original low-latitude regions, which could be mined and used to improve cold tolerance to help crops adapt to high latitude regions. This strategy was successfully employed in the development of the cold-tolerant rice variety Norin-PL8 by introgressing chromosomal segments from the tropical variety Silewah responsible for cold tolerance into the template japonica variety Hokkai241 (Saito et al., 2004).

Our findings reveal that the 358-bp Indel-972 in the bZIP68^maize promoter appears to enhance bZIP68 expression. The improvement of maize cold tolerance may thus be achieved either by introgressing the bZIP68^teosinte allele into modern maize or by directly deleting the 358-bp region using CRISPR/Cas9 technology. Interestingly, we identified multiple conserved cis-elements in this 358-bp insertion by Plantpan3.0 (Chow et al., 2019), including CCAAT-box, W-box (TTGACT/C), GAATC (MYB), and AAATT (AT-Hook) elements, which may account for the transcriptional upregulation of bZIP68 (Supplemental Figure S10E). A previous trans-expression quantitative trait locus analysis with 503 maize accessions (Hirsch et al., 2014) indicated that a regulatory polymorphism for bZIP68 expression is located in the coding region of NF-YB10 (NUCLEAR FACTOR-Y subunit B10) (Supplemental Figure S10F). NF-YB interacts with NF-YA and NF-YC to form a complex, which acts as a transcription factor or as a repressor to bind to CCAAT-box elements (Laloum et al., 2013). Further investigation of the transcriptional regulatory mechanism of bZIP68 in the cold-stress response will contribute to the engineering of cold-tolerant crops via element-specific genome editing.

Materials and methods

Plants materials and growth conditions

All transgenic maize (Z. mays) plants and CRISPR-Cas9 mutants were generated by the Center for Functional Genomics and Molecular Breeding of Crops, China Agricultural University, Beijing. The dreb1.7 weak allele (Δpro) mutant was reported previously (Xiao et al., 2020). The heterogeneous inbred family (HIF:MR1077) was generated from maize-teosinte BC₂S₃ recombinant inbred lines derived from a cross between maize inbred line W22 and teosinte accession CIMMYT 8759 (Z. mays ssp. parviglumis) (Liang et al., 2019; Tian et al., 2019). Near isogenic lines (NILs) homozygous for W22 and 8759 across the target region were developed for bZIP68 designated NIL^maize and NIL^teosinte, respectively, which were planted in Hainan, China (18� 23′ 12″ N, 109� 10′ 56″ E). The teosinte DNAs and maize inbred lines used for promoter re-sequencing were described previously (Fang et al., 2020; Gao et al., 2021).

Maize seeds were planted in pots (30 cm � 20 cm � 15 cm, length � width � depth) containing plant ash, vermiculite, and Pindstrup soil mix (Denmark) (3:1:1) and grown at 25�C under a 16-h light/8-h dark photoperiod with 150 μmol m⁻² s⁻¹ white light and 60% relative humidity. For cold treatment, maize plants were grown in pots (30 cm � 20 cm � 15 cm, length � width � depth) for 14 days with irrigation and exposed to 4�C under a 16-h light/8-h dark photoperiod for 3–4 days using a cold chamber (Conviron CMP6010). After cold treatment, the seedling was recovered at 25�C for 1 day prior to photography.

Physiological analyses

The ion leakage assay was performed as described previously (Zeng et al., 2021). Leaves from 14-day-old maize seedlings (following 4�C treatment and recovery) were placed in a 15-mL centrifuges tube with 10 mL ddH₂O. The ion concentration was measured with a conductivity meter (METTLER TOLEDO FE30) and the value recorded as S1. The ion concentration of ddH₂O was measured as S0. The samples were placed in boiling water for 30 min and cooled to room temperature. The ion concentration was measured and recorded as S2. The results were calculated as follows: $\mathrm{ion}\mathrm{�}\mathrm{leakage}\mathrm{�}\left(\mathrm{\%}\right)\mathrm{�}=\mathrm{�}(\mathrm{S}2-\mathrm{S}0)/(\mathrm{S}1-\mathrm{S}0).$ Osmolality was measured with an OsmoTECH PRO Multi-Sample Micro-Osmometer. The leaves of 14-day-old maize plants (following treatment at 4�C for 24 h) were directly passed through a 1-mL medical syringe to obtain plant extract. Twenty-microliters of samples was used to measure osmolality.

Plasmid construction and plant transformation

The transgenic plants (bZIP68-OE1, bZIP68-OE2, bZIP68-MYC#1, bZIP68-MYC#2, bZIP68-GFP, bZIP68^S250A #1, and bZIP68^S250A #2) used in this study were derived from maize inbred line LH244. These constructs were generated by the cloning bZIP68 and MPK8 coding sequences into pCUN(m)-MYC/GFP or pBCXUN (Zeng et al., 2021). To obtain mutants of bZIP68 and MPK8 using CRISPR-Cas9 genome editing, fragments of the first intron were selected as guide RNA targets and cloned into pBUE411 (Xing et al., 2014). Transgenic plants were obtained by Agrobacterium-mediated transformation.

To construct the plasmids, the coding sequences of bZIP68 and MPK8 were amplified and cloned into the EcoR1 site of pGBKT7 to generate the BD-bZIP68 and BD-MPK8 plasmids. The coding sequences of bZIP68 and MPK8 were amplified and cloned into the EcoRI site of pGADT7 to generate the AD-bZIP68 and AD-MPK8 plasmids. To obtain the Super:bZIP68/MPK8-GFP(-MYC/GFP) plasmids, the coding sequences of bZIP68 and MPK8 were cloned into the KpnI site of pSuper:1300-GFP/MYC. The vector PGEX4T-1 was used to construct GST-MPK8 at the EcoRI site. The pMAL-c5X (BamHI) and pET32a (EcoRI) vectors were used to express MBP-bZIP68 and His-bZIP68 recombinant proteins. To perform the transient transformation assay in N. benthamiana, the coding sequence of bZIP68 was cloned into the SalI site of pCAM1307, which was used to construct 35S:HA-FLAG-bZIP68. The mutated sequences of bZIP68 were generated by site-directed mutagenesis and used as templates to obtain the MBP-bZIP68^S229A, MBP-bZIP68^S250A, MBP-bZIP68^S229AS250A, 35S:HA-FLAG-bZIP68^S250A, Super:bZIP68^S250A-GFP, and Super: bZIP68^S250A-MYC constructs. The constructs used to clone the MPK8 sequences with point mutations were described previously (Zeng et al., 2021). The 1.0-kb DREB1.7 and 1.5-kb bZIP68 promoter sequences were amplified and cloned into pGreenII 0800-LUC vectors to obtain the pDREB1.7:LUC and pbZIP68:LUC constructs, respectively. The pDREB1.7:LUC and pbZIP68:LUC plasmids were transformed separately into Agrobacterium (GV3101) together with the helper plasmid (pSoup-P19) for subsequent assays. The 1-kb promotor sequence of DREB1.7 was amplified and cloned into pCAMBIA1381 to obtain pDREB1.7:GUS. All constructs were generated using a Seamless Assembly Cloning kit (Clone Smarter). The primers and restriction enzymes used for plasmid construction are listed in Supplemental Data Set S10.

RNA extraction and RT-qPCR

Total RNA was extracted from the leaves of 14-day-old maize plants using TRIzol reagent (Invitrogen, USA) and reverse transcribed into cDNA with M-MLV reverse transcriptase (Promega). The expression levels target genes were detected using SYBR Green reagent (Takara) in a StepOnePlus Real-Time PCR System (Applied Biosystem) (Zeng et al., 2021). The maize polyubiquitin gene Ubi-2 was used as an internal control. Primer sequences for qPCR are listed in Supplemental Data Set S10. The experiments were independently repeated at least three times (three biological replicates).

RNA-seq analysis

For RNA-seq analysis, 14-day-old maize seedlings (bzip68-1 and WT LH244) treated with or without cold at 4�C for 12 h were used to extract total RNA from the second leaves with TRIzol reagent. To remove the influence of rhythm, total RNA was also prepared from WT (LH244) plants grown at 25�C for 12 h. RNA integrity was evaluated using a 2100 Bioanalyzer (Agilent Technologies) and the RNA libraries were constructed and sequenced by Novogene (Beijing). Paired-end sequencing of 150 base reads (Novaseq 6000 pe150) were generated on Illumina platform. The methods used for RNA-seq data analysis were described previously (Li et al., 2021). Raw data (raw reads) in fastq format were subjected to quality control by FastQC (v.0.11.9) and the clean reads were mapped to the reference maize genome (B73 RefGen_v4, AGPv4) using the HISAT2 program (v2.0.4) with default parameters. The read counts of each gene were obtained by FeatureCounts (v.2.0.1). PCA and DEGs were analyzed using the R package DESeq2 (v.1.30.0). P-values were adjusted using the Benjamini–Hochberg procedure. DEGs were selected based on the following criteria: P < 0.05 and log₂ (|fold change|) ≥1. Three independent replicates were performed for each sample at each time point. Cluster analysis was performed using the R package pheatmap (v.1.0.12; https://cran.r-project.org/web/packages/pheatmap/index.html). GO enrichment of DEG clusters was performed with the program Phyper (http://www.geneontology.org/). The significance of the GO terms was corrected using FDR < 0.05. A heatmap of the expression levels of the DEGs was constructed using TBtools software (Chen et al., 2020).

Phylogenetic analysis

Amino acid sequences of the homologs of bZIP68 were downloaded from Gramene protein database (https://www.gramene.org/) with the criteria of identify >50%, score >70. Amino acid sequences of 26 bZIPs and 18 MPKs identified in maize, rice, and Arabidopsis were aligned using MEGA 7.0 software with default pairwise and multiple alignment parameters. The phylogenetic trees were constructed based on the alignment results using the Neighbor-Joining method in MEGA 7.0 software with the following parameters: Poisson correction, complete deletion, uniform rates, and bootstrap (500 replicates) (Newman et al., 2016).

Protein subcellular localization

To determine the sub-cellular localization of bZIP68, the leaves of 14-day-old Ubi: bZIP68-GFP transgenic plants were used for protein localization assays. GFP fluorescence signals were collected under a Leica SP5 confocal microscope (excitation wavelength 488 nm). The leaves (5-mm strip) were stained with DAPI, a nuclear marker (100 ngmL^–1), under a vacuum at 25�C for 12 h. The DAPI fluorescence signals were collected under a Leica SP5 confocal microscope (excitation wavelength 405 nm).

Yeast two-hybrid assay

Yeast two-hybrid assays were performed as described previously (Li et al., 2017). Yeast cells (AH109) were co-transformed with BD-MPK constructs and AD-bZIP68 or the empty vector pGADT7. The yeast cells were grown on -Leu/-Trp or -His/-Ade/-Leu/-Trp selective medium for 2–3 days.

In vitro pull-down assay

Recombinant proteins GST-MPK8 and GST were expressed in Escherichia coli BL21 (DE3) at 16�C for 12 h and purified using glutathione beads (GE Healthcare). His-bZIP68 was expressed in E. coli BL21 Codon Plus (DE3)-RIPL at 16�C for 12 h and purified using Ni-NTA agarose (Qiagen). GST pull-down assays were performed as described previously (Ding et al., 2015). GST-MPK8 and GST were separately incubated with glutathione beads at 4�C for 2.5 h. His-bZIP68 was added to the reaction, followed by incubation for 1.5 h. The interaction signal was detected with anti-His antibody (Beijing Protein Innovation).

Co-IP assay

Co-IP assays were performed as described previously (Ding et al., 2015). The plasmids Super:MPK8-MYC and Super:bZIP68-GFP were transformed into Agrobacterium GV3101 and co-infiltrated into N. benthamiana leaves, followed by incubation at 25�C for 36 h. The empty-vector Super:GFP was used as a negative control. Total proteins were extracted from the samples with protein extraction buffer (50-mM Tris–HCl, pH 7.5, 150-mM NaCl, 20% glycerin, 0.1% NP-40, 1 mM DTT, and 1� protease inhibitor cocktail) and incubated with GFP agarose beads (ChromoTek) for 2.5 h. The immunoprecipitated samples were washed five times with washing buffer (50-mM Tris–HCl, pH 7.5, 150-mM NaCl, 20% glycerin, 0.1% NP-40, 0.1% Triton X-100) and subjected to immunoblot analysis. The protein interaction signal was detected with anti-MYC antibody (Sigma-Aldrich).

BiFC assay

BiFC assays were performed in N. benthamiana leaves as described previously (Ding et al., 2015). The full-length bZIP68 coding sequence was cloned into pSPYNE. The MPK8 or MPK11 coding sequence was cloned in pSPYCE. The constructs were co-expressed in N. benthamiana leaves for 36 h. The YFP fluorescence signal was detected under a Leica SP5 confocal microscope (excitation wavelength 488 mm).

In vitro phosphorylation assay

The in vitro phosphorylation assay was performed as described previously (Li et al., 2017). Recombinant GST-MPK8^Y113C was incubated with MBP-bZIP68, MBP-bZIP68^S229A, MBP-bZIP68^S250A, MBP-bZIP68^{S229A S250A}, or MBP protein in the reaction buffer (20-mM Tris-HCl, pH 7.5, 10-mM MgCl₂, 25-mM ATP 0.1-μCi [γ-³²P] ATP, and 1-mM DTT) at 30�C for 30 min. MBP protein was used as a negative control. The reactions were stopped by adding 5� sodium dodecyl sulfate (SDS) sample loading buffer and separated on a 10% (v/v) SDS–polyacrylamide gel electrophoresis (SDS–PAGE) gel. The phosphorylated signals were visualized by autoradiography (Typhoon 9410 imager).

In-gel kinase assays

In-gel kinase assays were performed as previously described (Ding et al., 2015). Total proteins were extracted from 14-day-old maize seedlings with in gel protein extraction buffer (5-mM EDTA, pH 8.0, 5-mM EGTA, pH 8.0, 25-mM NaF, 1-mM Na₃VO₄, 20% [v/v] glycerol, 10-mM DTT, 1� protease inhibitor cocktail, and 25-mM HEPES-KOH, pH 7.5). The proteins were separated on a 10% (v/v) SDS–PAGE gel containing MBP-bZIP68 as a substrate. The gel was washed three times with washing buffer (25-mM Tris–HCl, pH 7.5, 0.5-mM DTT, 5-mM NaF, 0.1-mM Na₃VO₄, 0.5-mgmL^–1 BSA, and 0.1% Triton X-100) at room temperature for 20 min each time. The gel was incubated in protein renaturation buffer (25-mM Tris–HCl, pH 7.5, 1-mM DTT, 5-mM NaF, and 0.1-mM Na₃VO₄) at 4�C for 1, 12, or 1 h. The gel was incubated in kinase reaction buffer (40-mM HEPES‐KOH, pH 7.5, 1-mM DTT, 12-mM MgCl₂, 0.1-mM Na₃VO₄, and 2-mM EGTA) at room temperature for 30 min, followed by incubation in fresh kinase reaction buffer supplemented with 70-�Ci [γ‐³²P] ATP and 9-μL 1-mM cold ATP at room temperature for 2 h. The gel was washed five times with 5% (w/v) TCA and 1% (w/v) sodium pyrophosphate for 20 min each time. The phosphorylated signals were visualized by autoradiography (Typhoon 9410 imager).

LC-MS/MS assay

An LC-MS/MS assay was performed to identify the phosphorylation sites of bZIP68 by MPK8 by Mass Spectrum Laboratory of China Agricultural University. MBP‐bZIP68 and GST‐MPK8^Y113C purified proteins were incubated in protein kinase reaction buffer containing 20 mM MgCl₂, 50 mM Tris–HCl, pH 7.5, 1 mM DTT, and 50 �M ATP at 30�C for 30 min. The reaction products were reduced by DTT and alkylated by iodoacetamide, followed by digestion with trypsin (pH 8.5) at 37�C for 12 h. The results were analyzed by LC‐MS/MS as described previously (Liu et al., 2017).

Cell-free protein degradation assay

The cell-free protein degradation assay of bZIP68 was performed as described previously (Ding et al., 2015). Total proteins were extracted from the leaves of 14-day-old maize plants in degradation buffer (50-mM Tris-MES, pH 8.0, 10-mM EDTA, pH 8.0, 0.5-M sucrose, 1-mM MgCl₂, 5-mM DTT). Equal amounts of total proteins from MPK8-OE, mpk8-1, and WT (LH244) plants were incubated with recombinant MBP-bZIP68 protein for the time indicated. The reactions were stopped by adding 5� SDS sample loading buffer and separated on a 10% (v/v) SDS–PAGE gel. bZIP68 protein was detected with anti-MBP antibody.

Transcriptional activity assay

The dual-LUC transcriptional activity assay was performed as described previously (Li et al., 2021). Maize protoplasts (from 14-day-old etiolated seedlings) were co-transfected with pDREB1.7:LUC reporter constructs and Super: bZIP68-GFP at 25�C for 14 h. 35S:REN was used as an internal control. Total proteins were extracted from the samples using dual-LUC assay reagents (Promega). The LUC/REN ratio was used to measure DREB1.7 promoter activity, as detected using a GloMax 20/20 luminometer.

The transcriptional activity assay was performed using a GUS reporter system as described previously (Ding et al., 2015). Nicotiana benthamiana leaves were co-transformed with pDREB1.7:GUS and 35:HA-FLAG-bZIP68, MPK8-MYC, or 35:HA-FLAG-bZIP68^S250A; 35S:LUC was used as an internal control. GUS and LUC activity were measured as described (Shi et al., 2012) and relative GUS activity (GUS/LUC) was calculated to determine the activities of pDREB1.7 using a microplate reader (TECAN F-4500).

EMSA

EMSA was performed as previously described (Shi et al., 2012). The recombinant MBP-bZIP68 protein was incubated with biotin-labeled probe, competition probe, or mutant probe using a LightShift Chemiluminescent EMSA kit (Thermo Fisher) at 25�C for 20 min. The binding reactions were stopped by adding 5� native sample loading buffer and separated on a native-PAGE gel. The binding signal was detected using a Nucleic Acid Detection Module kit (Thermo Fisher). The DNA sequences of the probes are listed in Supplemental Data Set S10.

ChIP assay

The ChIP assay was performed as previously described (Shi et al., 2012) with minor modifications. WT LH244 and bZIP68-GFP overexpressing seedlings were grown at 25�C in the greenhouse for 14 days under a 16-h light/8-h dark photoperiod and treated with or without cold at 4�C for 12 h. The leaves were crosslinked and chromatin was extracted from the cross-linked samples and fragments with an ultrasonicator. DNA fragments associated with bZIP68-GFP protein were incubated with GFP agarose beads (ChromoTek). The enriched DNA fragments were collected with a ChIP DNA Clean and Concentrator kit (ZYMO). Immunoprecipitated DNA was quantified by qPCR using the primers of target genes. Relative enrichment was represented by input (%). The primers are listed in Supplemental Data Set S10.

Nucleotide diversity analysis

The 1.5-kb promoter regions of bZIP68 in maize (21) and teosinte (17) were re-sequenced by HITAC-seq (Gao et al., 2021). Population nucleotide diversity (π) analysis was performed with DnaSP version 5.0 software (Librado and Rozas, 2009).

Statistical methods

Differences between two groups were assessed using a two-sided Student’s t test. For multiple comparisons, significance analysis was performed with one-way ANOVA followed by Tukey’s multiple comparison tests. The immunoblot results were quantified using ImageJ software. The statistical analysis was performed using GraphPad Prism version 8.0 (Supplemental Data Set S11).

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL libraries under the following accession numbers: bZIP68 (NP_001353894, Zm00001d050018); MPK8 (NP_001149495, Zm00001d014658); bZIP49 (XM_008666300, Zm00001d031790); bZIP123 (NM_001136667, Zm00001d005884); bZIP4 (XM_008658809, Zm00001d012294); DREB1.1 (XM_008672111, Zm00001d006169); DREB1.2 (NM_001146976, Zm00001d021205); DREB1.3 (XM_008654734, Zm00001d021208); DREB1.4 (NM_001159200, Zm00001d006170); DREB1.5 (XM_020541212, Zm00001d021207); DREB1.7 (NM_001177010, Zm00001d036003); DREB1.10 (NM_001154158, Zm00001d002618); ZmMPK11 (XP_008656705, Zm00001d011465); ZmNF-YB10 (NP_001130166, Zm00001d050242). Ubi-2 (NM_001329666, Zm00001d053838); AtABF2 (NP_849777, AT1G45249); OsTRAB1 (NP_001390561, Os08t0472000).

RNA‐seq data are available at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/) under series entry PRJNA778546, which can be downloaded at http://www.ncbi.nlm.nih.gov/sra/PRJNA778546.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. bZIP68 encodes a bZIP transcription factor.

Supplemental Figure S2 . Amino acid sequence alignment of ZmbZIP68 and its homologous proteins in Arabidopsis (AtABF2), maize (ZmbZIP123, ZmbZIP49), and rice (OsTRAB1).

Supplemental Figure S3. bZIP68 is phosphorylated by MPK8.

Supplemental Figure S4. MPK8 negatively regulates the cold response by stabilizing bZIP68 in maize.

Supplemental Figure S5. PCA of WT and bzip68-1 with and without cold treatment.

Supplemental Figure S6. DREB1.7 positively regulates cold tolerance in maize.

Supplemental Figure S7. Sequence comparison of bZIP68 in maize and teosinte.

Supplemental Figure S8. The 358-bp insertion is ubiquitous in maize inbred lines.

Supplemental Figure S9. Graphical genotype of the HIF.

Supplemental Figure S10. Expression quantitative trait locus analysis identifies regulatory polymorphisms for bZIP68 expression.

Supplemental Data Set S1. List of cold response candidate genes screened from a maize population overexpressing more than 700 maize genes.

Supplemental Data Set S2. Genes differentially expressed in WT plants with and without cold treatment (4�C–0 h versus 4�C–12 h).

Supplemental Data Set S3. Genes differentially expressed in WT plants under normal conditions (25�C–0 h versus 25�C–12 h).

Supplemental Data Set S4. COR genes that were specifically regulated by cold treatment without effects of circadian clock.

Supplemental Data Set S5. Genes differentially expressed in bzip68-1 compared with the WT plants at normal conditions (4�C–0 h-bzip68 versus 4�C–0 h–WT).

Supplemental Data Set S6. Genes differentially expressed in bzip68-1 compared with the WT plants under cold treatment (4�C–12 h–bzip68 versus 4�C–12 h–WT).

Supplemental Data Set S7. TPM value of 1,339 bZIP68-regulated COR genes.

Supplemental Data Set S8. Cluster analysis of bZIP68-regulated COR genes.

Supplemental Data Set S9. List of maize inbred lines for genotyping.

Supplemental Data Set S10. List of primers used in this article.

Supplemental Data Set S11. Statistical analyses.

Supplemental File S1. Sequence alignments used for phylogenetic analysis.

 

S.Y. conceived the project. Y.S. and S.Y. designed the experiments. Z.L. and Y.S. performed most of the experiments. Z.L. and D.F. performed the RNA-seq analysis. S.Z. performed the maize transformation. Z.L., X.Z., and J.T. peformed the maize domestication analysis. Z.L., X.W., R.Z., X.Y., F.T., J.L., Y.S., and S.Y. analyzed and discussed the data. Z.L., Y.S., and S.Y. wrote the manuscript with comments from all the authors.

The authors responsible for distribution for materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plcell) are: Yiting Shi (shiyiting@cau.edu.cn) and Shuhua Yang (yangshuhua@cau.edu.cn).