ZmEREB92 negatively regulates seed germination in maize by affecting embryo growth during imbibition

ZmEREB92 has been reported to regulate maize terpenoid phytoalexins previously [30]. Here, we observed that the Crispr/Cas9-mediated knockout mutants of ZmEREB92 we generated before exhibited stronger seed germination as well as seedling emergence than wild type KN5585 (Fig 1A–1G). Such superiority can be retained in seeds harvested from different cultivation areas (S1 Fig). Radicle emergence is known as the determinant event for seed germination, which is dependent on elongating of embryo and loosening of surrounding structures [31]. By observing the longitudinal section of imbibed seeds, we found that the percentage of embryo was markedly higher in ereb92 mutants at 6, 24 and 36 hours after imbibition (HAI) compared to those of KN5585, while no difference was found in quiescent embryos at 0 HAI (Fig 1E and 1F). Besides, similar pericarp thickness and water absorption rates were detected in KN5585 and ereb92 mutant seeds (S2 Fig), indicating that rather than hydrate-mediated embryo swelling, endogenous embryo growth potential might be the primary force to drive faster germination of ereb92 mutants.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Loss-of-function of ZmEREB92 enhanced seed germination in maize by promoting embryo growth during imbibition.

(A) The germination performance for KN5585 and ereb92 mutants (ereb92-2 and ereb92-6) at 5 days after imbibition (DAI) on the filter paper infiltrated with sterile water. (B) The time-course germination of KN5585 and ereb92 mutants for 1–7 DAI under normal condition. Error bars indicate mean ± SE (n = 3). (C) The phenotype of seedling emergence of KN5585 and ereb92 mutants after 10 days sowing in the soil. (D) The seedling emergence rates of KN5585 and ereb92 mutants were counted at 4, 5, 6, 8, 10 days after sowing in the soil. Error bars indicate mean ± SE (n = 3). (E) The longitudinal section of KN5585 and ereb92 mutant seeds at 0, 6, 24 and 36 HAI. The embryo region was marked with yellow dash line. (F) The percentage of embryo for KN5585 and ereb92 mutant seeds at 0, 6, 24 and 36 HAI. The relative embryo proportion is calculated with the formula: Embryo area/ whole seed area *100%. The areas were calculated by ImageJ software. Error bars indicate mean ± SE (n = 5). (B, D, F) Asterisks indicate significant difference from the control (KN5585) at each time point (Two-way ANOVA followed by Tukey test, **P <0.01). (G) The schematic longitudinal view of the embryo structure in maize. (H) The histological sectioning and cytological analysis of the radicle and plumule region of KN5585 and ereb92-6 seeds at 36 HAI. (I, K) The cell size in radicle (I) and plumule (K) of KN5585 and ereb92-6 indictaed in (G), respectively. A total of 50 cells from three different sections of each line were counted. (J, L) The cell number of radicle and plumule of KN5585 and ereb92-6 seeds indicated in (G). (G-J) The distribution of cell size is displayed by boxplot. Error bars indicates the value range and the box shows the medium and the upper and lower quartiles. The circles represent for individual datapoints of biological replicates in each line. Number marked for each data is the exact P value (Student’s t-test, *P<0.05, **P<0.01).

https://doi.org/10.1371/journal.pgen.1011052.g001

To dissect the mechanisms by which ZmEREB92 regulates embryo enlargement during imbibition, we investigated the number of superficial cells and size of the apical cells in both radicle and plumule of KN5585 and ereb92-6 seeds at 36 HAI (Fig 1G and 1H). In radicle, the average size of apical cells of ereb92 mutants was enlarged by 39.2% and the number of superficial cells also significantly increased by 27.2% compared with those of KN5585 (Fig 1I and 1J). Similar results were also observed in plumule cells with an increment of 18.1% in the size of the apical cells and 19.5% in number of superficial cells by mutation of ZmEREB92 (Fig 1K and 1L). These results demonstrated that the enhanced embryo growth was contributed by both cell expansion and cell division.

Transcriptomics reveal the role of ZmEREB92 in regulating hormone-related pathways in imbibed seeds

To understand the regulatory network of ZmEREB92, we employed RNA-seq using the seeds at 0 and 6 HAI of KN5585 and ereb92-6 mutant (S1 Data). Compared to 0 HAI, more than a thousand of differentially expressed genes (DEGs) were detected at 6 HAI in both KN5585 and ereb92 mutant seeds (Fig 2A). Notably, the DEGs were mostly enriched in the up-regulated genes in comparison group of ereb92_6h vs KN5585_0h (3411 genes) and ereb92_6h vs ereb92_0h (3767 genes) (Fig 2A and 2B), suggesting that the release of suppression on downstream genes by ZmEREB92 mutation might play a major role in facilitating seed germination. Further GO and KEGG analysis of these up-regulated DEGs were all revealed a significant enrichment in hormone-related pathways (Fig 2C and 2D, S3A and S3B Fig, S1 and S2 Tables). Besides, other pathways including cell wall biogenesis, carbohydrate metabolism, UDP-glycosyltransferase activity and MAPK signaling were also enriched (Fig 2C and 2D, S3A and S3B Fig, S1 and S2 Tables). Together, these results strongly imply the role of hormone signaling in ZmEREB92-mediated seed germination.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Transcriptomic profile reveals the role of ZmEREB92 in regulating hormone-related pathways.

(A, B) A and B, Number of DEGs (Log₂ (|FC|) ≥1, P<0.05) (A) and the Venn diagram shows the number of overlapped DEGs (B) in the comparison groups of ereb92_6h vs KN5585_6h, ereb92_6h vs ereb92_0h, KN5585_6h vs KN5585_0h and ereb92_6h vs KN5585_6h. Three independent experiments were performed for each sample at each time point. (C, D) GO (C) and KEGG (D) analysis of the DEGs in the comparison groups of ereb92_6h vs ereb92_0h and KN5585_6h vs KN5585_0h. (E) The heatmap shows the log₂FC (Fold change) of differentially expressed hormone-related genes in the comparison groups of ereb92_6h vs ereb92_0h or ereb92_6h vs KN5585_6h. (F) qPCR analysis for several selected hormone-related genes in seeds of ere92 mutant and KN5585 at 6 HAI. Error bars indicate mean ± SE (n = 4). Ef1a was used as the reference gene and relative expression level was normalized to one biological replicate of KN5585. Asterisks indicate significant difference (Student’s t-test, *P<0.05, **P<0.01).

https://doi.org/10.1371/journal.pgen.1011052.g002

We thus analyzed the hormone-related DEGs in ereb92_6h vs ereb92_0h or ereb92_6h vs KN5585_6h (Fig 2E and S3 Table). Among them, GA biosynthesis and inactivation genes were both up-regulated in ereb92_6h compared to ereb92_0h or KN5585_6h (Fig 2E). Such increased expression pattern was also detected for several GA signaling genes, including GA receptor genes ZmGID1 and ZmGID2 (Fig 2E), suggesting a more active GA metabolism and signaling in imbibed ereb92 mutant seeds. Besides, the expression of ZmABH1, the ABA catabolic gene, was largely increased in ereb92_6h compared to ereb92_0h or KN5585_6h (Fig 2E). Three PYL genes (ABA receptor genes) and a set of PP2C genes, the negative regulator of ABA signaling, were also showed increased abundance, while ABA signaling gene ZmSnRK2.3 was slightly down-regulated (Fig 2E and S3C Fig). Besides, multiple ethylene, auxin and brassinosteroid (BR) related genes were also up-regulated except for ZmEIN1 (Fig 2E and S3C Fig), suggesting these pathways might contribute to the enhanced germination in ereb92 mutant, too. Further q-PCR analysis confirmed our RNA-seq results, in which ZmGA20ox1, ZmGA2ox6, ZmEIL4, ZmEIL7 and ZmETR2 were up-regulated in ereb92 mutant at 6 HAI compared to those of KN5585, while ZmSnRK2.3 expression was reduced (Fig 2F).

The AP2/ERF family members are known to bind the GCC-box in target gene promoters [32]. We next performed cis-element enrichment analysis to investigate whether ZmEREB92-triggered transcriptomic changes are mainly dependent on the GCC-box. We divided the datasets from KN5585_6h vs 0h (S1) and ereb92_6h vs 0h (S2) into three groups, including all differentially expressed genes (ADEG), up-regulated genes (URG) and down-regulated genes (DRG) (S4A Fig). Non-differentially expressed genes (NDEG) were also analyzed as the negative control. We firstly scanned the promoters (1.5 kb upstream of ATG) of ADEG, URG and DRG in S2 and calculated the percentage of genes in each group that contains a certain Transcription factor binding site (TFBS) in their promoters. The TOP10 enriched TF families were shown and we found that in the ADEG group, only TALE was enriched higher than NDEG (S4B Fig). For the URG group, the ERF, TALE and LBD displayed higher enrichment than NDEG, among which ERF accounts for the highest proportion (S4C Fig). However, in the DRG group, no TF family showed higher percentage than NDEG (S4D Fig). We further analyzed the genes that are presented in S2 but not in S1 (S2-S1) and got similar results (S4E–S4G Fig). Together, we suggested that the GCC-box might have the major effect in transcriptional up-regulation rather than down-regulation in the ereb92 mutant during imbibition.

Ethylene signaling plays a major role to drive seed germination in ereb92 mutants

Our RNA-seq data revealed that differentially expressed hormone-related genes were mostly enriched in GA, ABA and ethylene pathways (Fig 2), we next asked which hormone plays the main role. By applying hormone or corresponding inhibitor treatments, we firstly found that ABA application inhibited the germination of both KN5585 and ereb92 mutants, but the disparity between them was still large (Fig 3A and 3B). As with ABA treatment, the paclobutrazol (PAC) treatment, which inhibits GA biosynthesis, also resulted in similar trends (Fig 3A and 3B). Notably, application of 1-Methylcyclopropene (1-MCP), the competitive inhibitor of ethylene receptor, almost eliminated the difference between KN5585 and ereb92 mutants (Fig 3A and 3B), indicating a more important role for ethylene-related pathway. These results were also supported by p-value analysis (S5 Fig). However, the content of ethylene precursor 1-aminocyclopropanecarboxylic acid (ACC) in ereb92-6 seemed slightly lower than that in KN5585 at 36 HAI. These results implicated that it might be ethylene signaling rather than ethylene biosynthesis that acts in ZmEREB92-regulated seed germination. Meanwhile, GA accumulation and ABA degradation were more intensive in ereb92-6 mutant than KN5585 during imbibition, resulting in a significant reduction of ABA/GA ratio (Fig 3C and S6 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Ethylene signaling plays a major to promote seed germination in ereb92 mutants.

(A) The germination performance at the for KN5585 and ereb92 mutants at 3 DAI under normal condition (CK), 50 μM ABA treatment, 100 mg/L Paclobutrazol (PAC, inhibitor of GA biosynthesis) treatment and 200 mg/L 1-Methylcyclopropene (1-MCP, inhibitor of ET receptor). (B) The germination rate of KN5585 and ereb92 mutants counted at 1, 1.5, 2 and 3 DAI under the treatments described in (A). Error bars indicate mean ± SE (n = 4). (C) The heatmap showed the Log₂FC of the hormone level relative to KN5585_0h. (D) The longitudinal section of the seeds of KN5585 and ereb92 mutants at 0, 24 and 36 HAI under 1-MCP treatment. Embryo region was sketched with yellow dash line. (E) The percentage of embryo for seeds of KN5585 and ereb92 mutant at 0, 24 and 36 HAI under 1-MCP treatment. The relative embryo proportion is calculated by ImageJ software. The circles are represented for individual datapoints of biological replicates in each line. Error bars indicate mean ± SE (n = 7). n. s. indicates no significant difference using one-way ANOVA followed by Tukey tests (P>0.05).

https://doi.org/10.1371/journal.pgen.1011052.g003

We then performed ethephon (ETH) and 1-MCP treatment on seeds of maize inbred line Mo17 to clarify the action of ethylene signaling in maize seed germination. Results showed that although ETH treatment only slightly elevated the germination rate at 2 HAI, 1-MCP treatment continuously inhibited seed germination (S7A and S7B Fig). The embryo proportion was also increased by ETH but decreased by 1-MCP treatment at 36 HAI (S7C and S7D Fig). These findings confirmed that ethylene signaling played a positive role in regulating maize seed germination. Similarly, 1-MCP treatment compromised the embryo enlargement in ereb92 mutants after imbibition (Fig 3D and 3E). We also investigated whether the expression of ZmEREB92 could be affected by exogenous ethephon (ETH) or 1-MCP treatment during imbibition. The results showed that ZmEREB92 was significantly down-regulated by ETH at 24HAI but strongly induced by the 1-MCP at 12 and 24 HAI (S7E and S7F Fig). Altogether, we demonstrated that ZmEREB92 regulated seed germination by affecting the embryo growth, which might be dominantly mediated by ethylene signaling.

ZmEREB92 directly represses the expression of ZmEIL7

Among those up-regulated ethylene-related genes in our RNA-seq data, the promoter of ZmEIL7 contains most abundant GCC-boxes (S8A Fig). To verify whether ZmEIL7 is a target gene of ZmEREB92, the 1500 bp fragment upstream the start codon was amplified and subjected to a dual-luciferase (LUC) reporter (DLR) assay (S8B Fig). In line with the increased expression of ZmEIL7 in ereb92 mutant, ZmEIL7 promoter activity was significantly decreased in the presence of ZmEREB92 (Fig 4A). Such suppression was released by deleting the GCC-box enriched region in the promoter (proEIL7-P1) or retain only 794 bp upstream fragment (proEIL7-P2) (Fig 4A), suggesting that the GCC-box enriched region might be the site regulated by ZmEREB92. Yeast-one-hybrid (Y1H) assays were conducted subsequently to reveal the direct binding of ZmEREB92 to ZmEIL7 promoter containing the GCC-box enriched region (Fig 4B). Further EMSA confirmed that the GCC-box enriched region (-1235 to -1194 bp) served as the binding site for ZmEREB92 (Fig 4C). ZmEREB92 was found to lack trans-activation activity in our previous work [30]. Sequence analysis showed that ZmEREB92 contains two EAR motifs at C terminus (Fig 4D), which confer the repression activity by interacting with co-repressors, such as TOPLESS protein [33]. Mutation of either EAR motif-1 or -2 was insufficient to affect proZmEIL7 activity, while double mutations of both EAR-motifs resulted in a complete loss of suppression on ZmEIL7 promoter (Fig 4D), indicating that the transcriptional repression of ZmEREB92 on ZmEIL7 depends on both EAR motifs. Besides, the mutation of EAR-motifs did not impair the binding of ZmEREB92 to ZmEIL7 promoter (Fig 4B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. ZmEREB92 inhibits the transcription of ZmEIL7 by directly binds to the G-boxes in ZmEIL7 promoter.

(A) Dual-Luciferase Reporter (DLR) assays in maize protoplast showing that ZmEREB92 negatively regulated ZmEIL7 transcription through the -1188 to -1500 bp region upstream to the start codon. The different promoter region of ZmEIL7 were co-transformed with empty vector (EV) or ZmEREB92. p35S-REN was used as the internal control. Error bars indicate mean ± SE (n = 3). Asterisks indicate significant difference (Student’s t-test, **P<0.01). (B) The repression of ZmEREB92 on ZmEIL7 transcription was dependent on two EAR-motif at C terminus. The EAR-motif was indicated by black line. Single or double mutations of two EAR-motifs (ΔEAR-1, ΔEAR-2 and ΔEAR-1/2) in ZmEREB92 were generated and co-transformed with ZmEIL7 promoter, respectively. Error bars indicate mean ± SE (n = 3). Different lowercase letters represent significant differences (one-way ANOVA followed by Tukey tests, P<0.05). (C) Yeast one hybrid assays suggest the direct binding of ZmEREB92 to ZmEIL7 promoter. The promoter of ZmEIL7 was co-transformed with AD-ERBE92 or AD-EREB92(ΔEAR1/2) and grown on selective medium (SD/-Ura/-Leu/AbA). AbA, Aureobasidin, 400 ng·mL^-1. Empty pGADT7 vector (AD) was also co-transformed as the negative control. (D) EMSA to show the direct binding of ZmEREB92 to the GCC-boxes in ZmEIL7 promoter. Two fragments containing GCC-boxes in ZmEIL7 promoter were labeled with biotin and incubated with ZmERBE92 purified recombinant proteins. 200-fold excess of unlabeled probes were used for competition. (E, F) The expression pattern of ZmEIL7 in KN5585 and ereb92-6 mutant at 2, 6, 12, 24 and 36 HAI under control (E) and 1-MCP treatment (F). Error bars indicate mean ± SE (n = 3). Ef1a was used as the reference gene and relative expression level was normalized to one biological replicate of 2 HAI of KN5585. Different lowercases or majuscules represent significant difference in KN5585 or ereb92-6 mutant, respectively (one-way ANOVA followed by Tukey tests, P<0.05). Asterisks indicate significant difference between KN5585 and ereb92-6 mutant at each time point (Student’s t-test, **P<0.01, n.s. no significant difference).

https://doi.org/10.1371/journal.pgen.1011052.g004

Expression analysis revealed that ZmEIL7 expressed steadily within 12 HAI and decreased from 24 to 36 HAI in KN5585. In contrast, it was increased from 6 to 24 HAI in ereb92-6, both of which were significantly higher than those in KN5585 (Fig 4E), but decreased at 36 HAI. However, under 1-MCP treatment, not only the induction of ZmEIL7 in ereb92 mutant was abolished, but also the difference between KN5585 and ereb92 mutant was eliminated (Fig 4F). Phylogenetic tree of all EIL members in maize, rice and Arabidopsis showed that ZmEIL7 has a closest relationship with rice OsEIL6 (S9A Fig), and several GCC-boxes were also found in the promoter of OsEIL6 (S10B Fig), but its function remains unclear. Together, we illustrated that ZmEREB92 directly binds to the GCC-box enriched region of ZmEIL7 promoter to suppress its expression, such inhibition is dependent on two EAR motifs at C-terminus of ZmEREB92.

α-amylase contributes to ethylene-mediated germination in ereb92 mutants by accelerating starch mobilization

With findings that the ereb92 mutants possess more active embryo after imbibition (Fig 1C and 1D), we are curious whether this growth potential is derived from embryo itself. To answer this question, we performed germination assay using separated embryos without endosperm. Both KN5585 and ereb92-6 mutant seeds germinated rapidly with an identical germination tendency (Fig 5A and 5B), suggesting that the growth potential might also contributed by endosperm. Starch degradation in starchy endosperm is known to provide carbohydrate and energy for seed germination and early growth [8]. Indeed, the α-amylase activities of ereb92 mutants were significantly higher than that of KN5585 (Fig 5C). Two amylase genes, ZmAMYa2 and ZmAMYb5 were identified as DEGs from our RNA-seq data (S3C Fig). Among them, the promoter of ZmAMYa2 contains an array of five GCC-boxes while ZmAMYb5 contains none (S10A Fig). Further q-PCR assays confirmed that the expression of ZmAMYa2 was higher in ereb92 mutant seeds at 6 HAI (Fig 5D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. ZmEREB92 directly inhibit ZmAMYa2 transcription and resulted in accelerated starch mobilization in imbibed seeds of ereb92 mutants.

(A) The germination performance at 24 HAI using separated embryos of KN5585 and ereb92-6 mutant. (B) The germination rate at 0, 24, 36 and 48 HAI of KN5585 and ereb92-6 mutant using separated embryos. Error bars indicate mean ± SE (n = 3). (C) α-amylase activity in KN5585 and ereb92 mutants at 24 HAI. Error bars indicate mean ± SE (n = 7). Asterisks indicate significant difference (Student’s t-test, *P<0.05). (D) The relative expression of ZmAMYa2 in ereb92-6 mutant and KN5585 at 6 HAI. Error bars indicate mean ± SE (n = 3). Ef1a was used as the reference gene and relative expression level was normalized to one biological replicate of KN5585. Asterisks indicate significant difference (Student’s t-test, **P<0.01). (E) DLR assays in maize protoplast showing that ZmEREB92 negatively regulated ZmAMYa2 transcription and such repression was partially dependent on two EAR-motifs. The promoter of ZmAMYa2 were co-transformed with EV, ZmEREB92 or ZmEREB92(ΔEAR1/2), respectively. p35S-REN was used as the internal control. Error bars indicate mean ± SE (n = 3). Different lowercases indicate significant difference (one-way ANOVA followed by Tukey tests, P<0.05). (F) EMSA to show the direct binding of ZmEREB92 to the GCC-boxes in A and C fragments of ZmAMYa2 promoter. Three fragments containing GCC-boxes in ZmAMYa2 promoter were labeled with biotin and incubated with ZmERBE92. 200-fold excess of unlabeled probes were used for competition. (G, H) The expression pattern of ZmAMYa2 in KN5585 and ereb92-6 at 2, 6, 12, 24 and 36 HAI under control (G) and 1-MCP treatment (H). Error bars indicate mean ± SE (n = 3). Ef1a was used as the reference gene and relative expression level was normalized to one biological replicate of 2 HAI of KN5585. Different lowercases or majuscules represent significant difference in KN5585 or ereb92-6 mutant, respectively (one-way ANOVA followed by Tukey tests, P<0.05). Asterisks indicate significant difference between KN5585 and ereb92-6 mutant at each time point (Student’s t-test, **P<0.01, n.s. no significant difference).

https://doi.org/10.1371/journal.pgen.1011052.g005

Subsequent DLR assays demonstrated that proZmAMYa2 activity was inhibited by the presentation of ZmEREB92 (Fig 5E). However, the double mutations of two EAR-motifs in ZmEREB92 only partially rescued the activity of proZmAMYa2 (Fig 5E), suggesting involvement of other repressors. The EMSA results further established that ZmEREB92 directly binds to the A and C regions of proZmAMYa2 (Fig 5F). By q-PCR analysis, we found that ZmAMYa2 shared a similar expression pattern with ZmEIL7. Phylogenetic analysis showed ZmAMYa2 is closely related to RAmy3D in rice (S9B Fig), which has a positive role in seed germination by facilitating starch degradation in rice [14]. A series of GCC-boxes were also enriched in the promoter of RAmy3D (S10B Fig), suggesting similar regulation of this gene by ERF transcription factors. Together, ZmEREB92 also regulates the starch mobilization during imbibition to affect seed germination, which is possibly mediated by the direct repression of ZmEREB92 on ZmAMYa2 expression.

Rice ortholog of ZmEREB92 is functionally conserved in regulating seed germination

Among all ERF VIII subgroup members in maize, rice and Arabidopsis, ZmEREB92 is phylogenetically closest to rice OsERF74 (S11 Fig). Syntenic analysis confirmed that OsERF74 is the ortholog of ZmEREB92 in rice (Fig 6A). Phylogenetic tree of ZmEREB92 and its other orthologs from major cereals and Arabidopsis also clustered ZmEREB92 and OsERF74 into one branch (Fig 6B). ZmEREB92 showed a maximum expression in embryo during kernel maturation, while OsERF74 expressed highest in reproduction organs, followed by endosperm during seed development (S12 Fig). To test whether OsERF74 exerts similar functions as ZmEREB92, we generated the Crispr/Cas9-mediated knockout mutants of OsERE74 (S13 Fig). Two independent mutant lines were obtained and showed enhanced seed germination performance (Fig 6C and 6D). A genome-wide scan for genes that underwent convergent selection in rice and maize domestication identified OsERF74 as a selected gene on chromosome 5 [34]. Consistent with this, the nucleotide diversity of OsERF74 promoter region was strikingly reduced in Oryza sativa ssp. japonica (hereafter, japonica) compared to wild rice Oryza rufipogon (hereafter, rufipogon), while genomic region of OsERF74 showed an increased diversity in Oryza sativa ssp. indica (hereafter, indica) (Fig 6E). ZmEREB92 was skipped in previous studies because it was not assembled in B73 RefGen_v4 (S14 Fig and S4 Table). We analyzed the nucleotide diversity across ZmEREB92 locus using a small maize and teosinte panel based on the public genome assembled by B73 RefGen_v5. As with OsERF74, similar loss of nucleotide diversity was detected in the promoter region of ZmEREB92 (Fig 6F).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Rice ortholog of ZmEREB92 negatively regulates seed germination and underwent selection during rice domestication.

(A) Syntenic analysis of ZmEREB92 and OsERF74 from maize chromosome 8 (B73 RefGen_v5, pink) and rice chromosome 5 (Nipponbare IRGSP-1.0, light blue), respectively. (B) Phylogenetic analysis of ZmEREB92 and its orthologs in major cereal crops and Arabidopsis using neighbor-joining method. The numbers shown next to the branches indicate bootstrap values from 1000 replicates. (C) The germination performance of Kitaake and Oserf74 mutants at 3 DAI under normal condition (25 °C). (D) Time-course seed germination rate from 0–7 DAI for Kitaake and Oserf74 mutants under normal condition. Error bars indicate mean ± SE (n = 3). Asterisks indicate significant difference from the control (Kitaake) at each time point (Two-way ANOVA followed by Tukey test, *P <0.05, **P <0.01). (E) Nucleotide diversity across the OsERF74 locus among O. sativa japonica (blue), O. sativa indica (green) and O. Rufipogon (pink). The genomic region of OsERF74 is shown at the X axis. (F) Nucleotide diversity across the ZmEREB92 locus among maize (blue) and teosinte (pink). The genomic region of ZmEREB92 is shown at the X axis.

https://doi.org/10.1371/journal.pgen.1011052.g006

We next profiled the expression levels of ZmEREB92 and OsERF74 in different maize and rice subspecies using the public high-throughput transcriptome datasets. The results revealed that ZmEREB92 is significantly down-regulated in maize compare to those in teosinte (S15A Fig) [35]. Additional analysis in 368 maize inbred lines showed that the temperate maize lines tend to have lower expression level of ZmEREB92 than tropical maize lines (S15B Fig) [36]. Besides, no difference was found between the stiff stalk (SS) and non-stiff stalk (NSS) heterotic groups of temperate maize (S15B Fig), suggesting that we should utilize other germplasm resources such as tropical/subtropical lines to explore the potential excellent alleles of ZmEREB92 for breeding programs. In rice, OsERF74 showed a lower expression in temperate japonica than those in rufipogon, indica and tropical japonica (S15C Fig) [37]. Using other transcriptomic datasets for indica and japonica, we found that OsERF74 displayed higher expression in indica compare with japonica (S15D Fig) [38,39], among which indica are usually distributed in tropical regions while japonica are mainly grown in temperate regions [40]. Such pattern is similar with the expression divergence of ZmEREB92 in temperate and tropical maize (S15B Fig). Our results suggest that the orthologous gene pair ZmEREB92/OsERF74 might have experienced convergently selection during maize and rice domestication with conserved function in seed germination control and show similar expression divergence.

ZmEREB92 and OsERF74 conservatively regulate the sensitivity to temperature during seed germination

Temperature is an essential environmental signal that sensed by seed to determine germination [41]. Previous studies revealed that both ZmEREB92 and OsERF74 could be induced by cold [42,43]. In our transcriptome data, a series of cold-related genes were up-regulated in ereb92 mutant, including CESAs, DREBs and CIPKs (S16 Fig). The q-PCR results confirmed that these genes expressed higher in ereb92 mutant seeds at 6 HAI (Fig 7A). Several GCC-boxes were found in the promoter region of these two genes and DLR assays proved that the transcription of both CESA genes can be strongly repressed by ZmEREB92 (Fig 7B and S10C Fig). These results imply that ZmEREB92 might also regulate seed germination under cold condition. To verify this, we firstly demonstrated that the expression of ZmEREB92 was indeed up-regulated by cold treatment but repressed in normal condition during imbibition (Fig 7C). We then observed a much stronger seed germination and seedling establishment for ereb92 mutants compared to KN5585 in cold condition (Fig 7D and 7E and S17 Fig), while such superiority was not detected under PEG treatment (S18A and S18B Fig). The expression of ZmEREB92 was only slightly induced at 12HAI but reduced to the control level at 24 and 36 HAI by PEG treatment (S18C Fig). Our results suggest a specific role for ZmEREB92 in cold tolerance. In rice, the mutation of OsERF74 also resulted in enhanced seed germination under cold stress (Fig 7F and 7G), suggesting conserved function for orthologous ZmEREB92/OsERF74 to regulate seed germination in a temperature-sensitive manner.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. The orthologous gene pair ZmEREB92/OsERF74 show conserved functions in seed germination under cold stress.

(A) Relative expression analysis of cold-related genes in seeds of ere92-6 mutant and KN5585 at 6 HAI. Ef1a was used as the reference gene and relative expression level was normalized to one biological replicate of KN5585. (B) DLR assays for the promoters of ZmCESA1 and ZmCESA6 co-transformed with EV or ZmEREB92. (C) The expression pattern of ZmEREB92 during imbibition under normal (28 °C) and cold (12 °C) condition, respectively. Ef1a was used as the reference gene and relative expression level was normalized to one biological replicate of control-2HAI. (A-C) Error bars indicate mean ± SE (n = 3). Asterisks indicate significant difference (Student’s t-test, *P<0.05, **P<0.01). (D) The seed germination performance of KN5585 and ereb92 mutants at 6 DAI under cold condition (12 °C). (E) Time-course seed germination rate from 1–7 DAI for KN5585 and ereb92 mutants. Error bars indicate mean ± SE (n = 3). (F) The germination performance of Kitaake and Oserf74 mutants at 6 DAI under cold condition (15 °C). (G) Time-course seed germination rate from 0–7 DAI for Kitaake and Oserf74 mutants under cold condition. Error bars indicate mean ± SE (n = 3). (E, G) Asterisks indicate significant difference from the control (Kitaake) at each time point (Two-way ANOVA followed by Tukey test, **P <0.01).

https://doi.org/10.1371/journal.pgen.1011052.g007

Plant materials and growth conditions

The CRISPR-Cas9 mutants of ZmEREB92 were generated from our previous study in KN5585 background [30]. The inbred line Mo17 was also used for germination assay in our study. All these lines were planted in the experimental fields of Sichuan Agricultural University in Chengdu, Sanya and Xishuangbanna at normal condition and seeds were harvested at 45–60 d after artificial self-pollination. The plant height (PH), ear position height (EPH), spike leaf length (SLL) and width (SLW) were measured at milk-ripening stage and the yield-related traits including kernel row number (KNR), kernel number per row (KNPR), ear length (EL), ear diameter (ED), corncob diameter (CD), ear weight (EW), corncob weight (CW) and hundred grain weight (HGW) were measured using the well-pollinated ears. A total of 10 plants were detected for each line as biological replicates. For seedling growth, the seeds were planted in a same pot and grown at the growth chamber at 28 °C/24 °C (day/night) and 16 h light/8 h dark photoperiod with constant irrigation. Three-leaf-stage seedlings were used for cold treatment at 4°C for 4 d with at least three biological replicates.

Germination assay

Seeds of different genotypes were sterilized with 10% NaClO solution for 20 min and washed 5 times with sterile water. After that, 30 seeds were evenly placed on the filter paper infiltrated with sterile water in the germination box and putted into growth chamber at 28°C with dark condition. Equal amount of sterile water was added every 24 h to keep the filter paper moist. To study the hormone effects on seed germination, water was replaced with 200 mg/L 1-Methylcyclopropene (1-MCP, Fresh Doctor, inhibitor of ET receptor), 100 mg/L Paclobutrazol (PAC, Solarbio, inhibitor of GA biosynthesis), 50 μM ABA (Sigma) or 50 μM ethephon (ETH, Solarbio). For cold and osmotic stress, the sterilized seeds were germinated under 12°C or with 13.5% PEG6000 treatment in dark condition, respectively. The radicle emergence was defined for germination and the number of germinated seeds was recorded every 24 h. Each assay was performed with at least 3 biological replicates. To investigate the internal seed structure during imbibition, the seeds at 0, 6, 24 and 36 h after imbibition (HAI) were cut longitudinally and observed under a stereomicroscope (Olympus-SZX10). The embryo proportion and pericarp thickness were calculated using the ImageJ software with 10 biological replicates.

Histological sectioning and cytological analysis

To observe the cell size and cell number, the embryos were isolated from the seeds of KN5585 and ereb92-6 at 36 HAI and then fixed in the FAA buffer (formaldehyde:acetic acid:ethanol:water = 5:5:63:27, V/V/V/V) for 36 h. The samples were sequentially dehydrated with a gradient ethanol, cleared in xylene series, embedded in paraffin and sectioned at 4 μm thickness. The sections were then rehydrated in the gradient xylene and ethanol series and washed with water. After stained with 0.1% toluidine blue for 5 min, the sections were observed and photographed under a light microscope (Leica, DM500). The epidermal cell numbers in radicle and plumule were counted with 5 biological replicates and cell size was calculated with the ImageJ software from 50 cells in 10 randomly selected views from three different sections of each line.

Water absorption rate

The seeds of KN5585 and ereb92 mutants were evenly placed on the same filter paper infiltrated with sterile water. The incipient weight of each seed (M_0h) was measured before imbibition. After 24 h imbibition, the weight of each seed (M_24h) was measured again. The relative water absorption rate of each line was calculated with 16 biological replicates using the formula: (M_24h-M_0h)/M_0h×100%.

Measurement of endogenous hormone contents

The seeds of KN5585 and ereb92-6 at 0 and 36 HAI were collected, grounded into powder with liquid N₂ and extracted with 1 mL methanol/water/formic acid (15:4:1, V/V/V). After centrifugation for 5 min at 4 °C, 1200 rpm, the extracts were concentrated and re-dissolved with 100 μL 80% methanol/water (V/V) solution for LC-MS/MS analysis. The endogenous GAs, ABA and ACC contents were detected by MetWare (http://www.metware.cn/) based on the AB SciexQTRAP 6500 LC-MS/MS platform. Three biological replicates from each line were analyzed.

RNA extraction and qRT-PCR

To determine the expression pattern during seed germination, the seeds of different genotypes were imbibed for 2, 6, 12, 24 and 36 h with or without a certain treatment and grounded into a fine powder in liquid N₂. The RNA was extracted with TRNzol reagents (Tiangen) according to the manufacturer’s instructions and subsequently 1 μg of RNA was used for cDNA synthesis with the reverse transcriptase (Vazyme). The quantitative real-time RT-PCR (qRT-PCR) was executed on a StepOne Plus Real-Time PCR System (Applied Biosystems, ABI) using the SYBR Green Master Mix (Takara). All expression levels were calculated by the ΔΔCT method using the elongation factor Ef1a as the reference gene [61]. All experiments were performed with at least three biological replicates. Primers were listed in S5 Table.

RNA-seq analysis

Total RNA was extracted from the of KN5585 and ereb92-6 at 0 and 6 HAI seeds with three biological replicates and RNA integrity was evaluated with Agilent 2100 bioanalyzer. The RNA libraries were prepared and sequenced by Novogene (Beijing) on an Illumina Novaseq platform. Raw data was subjected to FastQC software for quality control and then the clean reads were aligned to maize genome (B73 RefGen_V4, AGPv4) using Hisat2 v2.0.5. FPKM, which was generated by counting the reads numbers mapped to each gene by featureCounts v1.5.0-p3, and used to estimate gene expression levels. Differential expression analysis was performed using the DESeq2 R package (1.20.0) and defined with a corrected P-value<0.05 (Benjamini and Hochberg’s approach) and absolute foldchange (FC) ≥2. Differentially expressed genes (DEGs) were further implemented by clusterProfiler R package for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. The TBtools software was used to display the expression of DEGs by heatmaps.

Cis-element enrichment analysis

The upstream 1500 bp of the gene was extracted as the promoter sequence. The cis-elements were predicted using the Plant Transcriptional Regulatory Map (http://plantregmap.gao-lab.org/binding_site_prediction.php) with the P value < 10⁻⁵ as the threshold for filtering. The retrieved cis-elements were indicated as transcription factor binding sites (TFBS) and then the percentage of genes with each TFBS in promoter was calculated and compared. Genes that were not differentially expressed in any comparison were analyzed as the control. The bootstrap method was used to test the significant enrichment.

Phylogenetic analysis

A literature search or database search based on PhyloGenes website (http://www.phylogenes.org/) was performed to identify the group VIII subfamily ERF transcription factors, EIN3/EIL members and α-amylases in Zea mays (RefGen_V4), Oryza sativa (Japonica Group IRGSP-1.0) and Arabidopsis thaliana (TAIR10), and the sequences were downloaded from MaizeGDB (https://www.maizegdb.org/) and EnsemblPlants (http://plants.ensembl.org/index.html). Amino acid sequences were aligned in MEGA software with ClustalW method using the maximum likelihood method. The iTOL online website (https://itol.embl.de/) was used to visualize the tree. Accession numbers for sequence data used in phylogenetic analysis can be found in S6 Table.

Dual-luciferase reporter assay

The promoter sequences of downstream genes were downloaded from Phytozome v13.0 (https://phytozome-next.jgi.doe.gov/) based on Zea mays RefGen_V4. The promoters were cloned from the genomic DNA of B73 inbred line and then constructed into the reporter vector pGreenII 0800-LUC. The coding sequence of ZmEREB92 was cloned into the effecter vector pGreenII 62-SK driven by the 35S promoter. The promoter-LUC plasmids were co-transformed with ZmEREB92 into maize protoplasts via PEG-mediated transfection method as described previously and REN was used as the internal control [62]. After incubated for 16h, total proteins were extracted from the transfected cells with the dual-luciferase assay reagents (Promega) and analyzed on a luminometer (Thermo Scientific Varioskan LUX). The LUC/REN ratio was used to defined the promoter activity. At least four biological replicates were performed for each experiment.

Yeast one-hybrid (Y1H) assay

To test the direct binding of ZmEREB92 to ZmEIL7 promoters, the yeast one-hybrid assays were carried out based on the GAL4-AbA system (Clontech) [63]. The GAL4-activating domain fused ZmEREB92 (AD-EREB92) was obtained in our previous study and further mutation of two EAR motifs was performed to generate AD-EREB92(ΔEAR1/2). Promoter region of ZmEIL7 was cloned into the pAbAi vector and then linearized at BbsI or BstBI site before transformed into the Y1H gold strain. Subsequently, the AD-EREB92 or AD-EREB92(ΔEAR1/2) was transformed into yeast cells integrated with pZmEIL7-AbAi vector and the transformed cells were grown on the selective medium SD/-Ura/-Leu and SD/-Ura/-Leu/AbA for 2–4 d. p53 and empty pGADT-7 (AD) were also transformed as the negative control.

Electrophoretic mobility shift assay (EMSA)

The CDS of ZmEREB92 was connected into pET32a with a His-tag fused to the C terminus and then transformed into BL21 competent cells. The recombinant protein was induced by 0.5 mM IPTG and purified with Ni-NTA beads (Smart-Lifesciences). The promoter fragments of ZmEIL7 and ZmAMYa2 were labeled with biotin (Sango). Afterwards, 0.2 pmol labeled probes were incubated with equal amount of ZmEREB92-His protein for 30 min at 28 °C. The unlabeled probes were used as cold competitors and added with 200× amount for competition. The mixtures were further separated in 6.5% native page gel and transferred to the 0.45μm Nylon membrane (Sango). The probes were detected with a chemiluminescence EMSA kit (Beyotime). Probes were listed in S5 Table.

Analysis of α-amylase activity

The seeds of KN5585 and ereb92 mutants imbibed for 24 h were used to quantified the α-amylase activity with a micro α-amylase assay kit (Solarbio). Briefly, the imbibed seeds were ground to a fine powder in liquid N₂ and 0.8 mL ddH₂O was added into 0.1 g powder and mixed thoroughly. After 15 min incubation at room temperature, the supernatant was collected by centrifuging at 6000 g for 10 min and further used for the detection of α-amylase activity. Three biological replicates from each line were analyzed.

Bioinformatic analysis

Sequence alignments was performed with ClustalW method and the phylogenetic tree was constructed with the Maximum- Likelihood method using the MEGA software. All phylogenetic trees were visualized through iTOL online website (https://itol.embl.de/). The orthologous genes were identified using the public database including maizeGDB (https://www.maizegdb.org/), EnsemblPlants (http://plants.ensembl.org/index.html) and Gramene Maize (https://maize-pangenome.gramene.org/). The genomic sequences and annotations of maize chromosome 8 and rice chromosome 5 were downloaded from EnsemblePlants database based on Zm-B73-REFERENCE-NAM-5.0 and Oryza sativa Japonica Group IRGSP-1.0, respectively. Then the syntenic analysis between maize chromosome 8 and rice chromosome 5 was performed with TBtools [64]. The constitutive expression patterns of ZmEREB92 and OsERF74 from qTeller dataset (http://qteller.maizegdb.org) and Rice Expression Profile Database (RiceXPro, https://ricexpro.dna.affrc.go.jp/) were shown in heatmap drawn by TBtools [64–66].

The expression profiles of ZmEREB92 and OsERF74 in different species were derived from several public high-throughput RNA-seq data. The dataset used for the expression analysis of ZmEREB92 in wild and cultivated maize is deposited in GEO database under the accession number GSE30036, which sequenced the above-ground tissue of 38 maize genotypes and 24 teosinte genotypes [35]. We extracted the expression value of ZmEREB92 based on the probe ID CHR08RS119567227, CHR08RS119567281 and CHR08RS119568280. The transcriptomes for temperate maize and tropical maize cultivars are retrieved from the GenBank Sequence Read Archive (SRA) under the accession code SRP026161, which performed on the immature seeds of 15 days after pollination for these 368 lines [36]. The RPKM of ZmEREB92 in SRP026161 is normalized for analysis. The expression data for OsERF74 is collected from the rice library in Plant Public RNA-seq Database (http://ipf.sustech.edu.cn/pub/plantrna/) by searching the gene ID LOC_Os05g41780 and BioProject ID PRJNA428294, PRJNA597070 and PRJNA292458, in which PRJNA428294 performed on the panicles of wild and cultivated rice, PRJNA597070 performed on the young leaf of different cultivated rice accessions and PRJNA292458 performed on the seedlings of six accessions from indica and japonica [37–39,67].

Nucleotide diversity analysis

Since ZmEREB92 was not assembled in B73_RefGen V4 genome, a small panel of genomic sequences across ZmEREB92 locus from 26 maize accessions and 8 teosinte accessions were downloaded from maizeGDB based on Zm-B73-REFERENCE-NAM-5.0 reference genome. Genomic sequence across OsERF74 locus using the data from Rice Super Pan-genome Information Resource Database (RiceSuperPIRdb, http://www.ricesuperpir.com/), including 132 indica accessions, 57 japonica accessions and 23 rufipogon accessions [68]. Nucleotide diversity was calculated under a 50-bp sliding window and 25 bp step size for maize/teosinte comparison or a 100-bp sliding window and 25 bp step size for O. sativa/O. rufipogon comparison using DnaSP 5.0 software (http://www.ub.edu/dnasp/) [69].

Generation of rice Oserf74 knockout lines for seed germination assays

To generate the CRISPR/Cas9-mediated OsERF74 knockout mutation in rice, the sequence 5’-GGCACCGGCCAGCACTTCCGTGG-3’ from 207 to 226 bp of OsERF74 CDS was selected as the guideRNA and introduced into the pCBSG032 CRISPR/Cas9 vector. Then the construct was transformed into rice callus with the Agrobacterium tumefaciens strain EHA105. The positive transformants were identified by hygromycin soaking and sequencing confirmation. Two T₂ knockout lines were obtained for germination assays as described above for maize seeds. The difference is that the germination of rice seeds is at 25 °C for normal condition and 15 °C for cold condition.

Chilling treatment at seedling stage

The three-leaf-stage seedlings of KN5585 and two ereb92 mutant lines were used for chilling treatment. The seedlings were transferred to the growth chamber at 4°C with a 16 h light/8 h dark photoperiod for 4 days and then recovered at 25°C for 2 days. After recovery, the photographs were taken and the survival rates were calculated. As described previously, the relative injured area (%) was defined with the ratio of the wilting area to the whole area of the second leaves using the Image J software [70]. Each experiment was performed with at least three biological replicates.

Statistical analysis

The statistical analysis was carried out using the GraphPad Prism 8.0 software. Briefly, multiple comparison was performed using one-way or two-way ANOVA followed by Tukey tests or LSD tests and differences between two data sets were calculated by the two-side Student’s t test or Mann-Whitney U test.