Maize is one of the most important crops in the world, with an annual global production of 1.21 billion tons in 2021 (https:// www.fao.org/). Mechanized harvesting of maize kernels is an effective way to solve the continuous increase of labor costs and improve production efficiency. However, mechanized harvesting has not yet been achieved in China and many other countries, and the mechanized harvesting area of maize kernel is less than 5% in China, partly due to a lack of suitable maize varieties. [1][2][3] Kernel dehydration rate (KDR) is a key determinant of moisture content at harvest that affects mechanized harvesting and kernel quality. [4][5][6] A high moisture content at harvest seriously limits harvesting efficiency and increases the cost of drying and storage. 7,8 This represents a considerable problem as most maize varieties in China have high grain water content at harvest, ranging from 30% to 40%, but the moisture content suitable for mechanized harvesting is from 15% to 25%, seriously limiting the widespread application of mechanical harvesting. 1,[9][10][11][12] Farmers usually delay harvest to reduce moisture content, which leads to a delay in the planting time of the next crop and greatly affects the annual yield of the crops, especially in the Huang-Huai-Hai region of China with a Wheat-Maize Rotation System. 1,4,7 High moisture content can also lead to kernel breakage during mechanical harvesting, 6,13,14 and can cause kernel mold, ear rot, and ear sprouting under warm and humid environmental conditions. 10,[15][16][17] Thus, increasing KDR and reducing kernel moisture content at harvest is essential and has become a major goal of modern maize breeding. 18,19 Many studies have identified quantitative trait loci (QTLs) controlling KDR in maize over the past two decades. 1,4,6,10,13,18,[20][21][22][23] However, very few QTLs have been cloned, 21 rendering the molecular mechanism of maize KDR unsolved. Furthermore, the genes controlling KDR are also rarely reported in plants, 24 and the underlying mechanisms are largely unknown.
Recent advances have shown that some so-called long noncoding RNAs (lncRNAs) have small open reading frames (sORFs) less than 100 amino acids in length. [25][26][27] These sORFs can be translated into small proteins, often referred to as micropeptides. [28][29][30] Many functional peptides translated from lncRNAs have been studied in humans, leading to molecular insights into their functions. [30][31][32][33][34] While a few functional micropeptides have been identified in plants, [35][36][37][38][39][40] relatively little is known about their roles.
Furthermore, the origin of new genes can generate and maintain remarkable phenotypic diversity. [41][42][43] These new genes can originate de novo from non-coding sequences and affect adaptation, morphological diversity, and speciation of organisms. [43][44][45][46][47] However, new genes in plants that are both novel and phenotypically impactful remain largely uncharacterized. 48 In this study, we fine-mapped and cloned a QTL for KDR in maize, Kernel Dehydration Rate1 (qKDR1), to a non-coding intergenic DNA sequence. qKDR1 regulates the expression of a 31 aa micropeptide encoding gene, qKDR1 Regulated Peptide Gene (RPG), located $10 kilobases upstream of qKDR1. Two MYBrelated transcription factors, ZmMYBST1 and ZmMYBR43, bind to the qKDR1 region to repress RPG expression. mi-croRPG1 (micropeptide of RPG ORF1) is derived from a sORF generated by a single nucleotide mutation at the base of the genus Zea that created a start codon. The microRPG1 micropeptide controls KDR by regulating the expression of genes in the ethylene signaling pathway ZmETHYLENE-INSENSITIVE3like 1 and 3 (ZmEIL1 and 3). Modulation of microRPG1 expression or exogenous application of the peptide in the distantly related dicot model Arabidopsis impacts KDR. Our findings shed light on the molecular mechanism underlying KDR and provide useful insights for future crop breeding, underscore the significant role of micropeptides, and present an excellent model for de novo gene evolution.
qKDR1 is the causal locus for KDR We mapped four QTLs for KDR in a maize recombinant inbred line (RIL) population developed from a cross between the maize inbred lines K22 and DAN340 49,50 (Figure 1A). The QTL with the largest effect, qKDR1 on the long arm of chromosome 1 and explaining 9.8% of the phenotypic effect, was selected for positional cloning. To examine the allelic effects, we developed a pair of qKDR1 near isogenic lines (NILs) (NIL DAN340 and NIL K22 ) (Figure 1B) from a heterogeneous inbred family library. 54 Compared with NIL DAN340 , NIL K22 had a more rapid KDR with up to 1.8% reduction in moisture content at harvest, but no obvious difference in other agronomic traits (Figure 1B; Table S1).
The QTL was further narrowed down to a 1,417 base pair (bp) intergenic non-coding region according to the maize B73 reference sequence version 4.0 (Figure 1C). NIL DAN340 had a 6,181bp insertion in this region compared with NIL K22 (Figure 1C; Table S2). To confirm the function of qKDR1, we deleted the target sequence using CRISPR-Cas9 in two different maize inbred lines, Zheng58 (similar to NIL K22 ) and B104 (similar to NIL DAN340 ) (Figure 1D). Both knockout lines had slower KDR compared with wild-type siblings (Figures 1E and 1F), with few effects on other agronomic traits in two different environments (Table S3). We next surveyed the natural variation of the 6,181bp QTL region insertion in a diverse maize inbred line population, 55 and identified two additional insertions or deletions (indels) in this region by polymerase chain reaction (PCR). Finally, five different haplotypes were identified (Figure 1G), and four of them, Hap2-Hap5 (with indel), had a lower KDR compared with Hap1 (Figure 1H). The collective data all show statistically significant differences based on different methods (Figure 1I). These results indicate that the 1,417-bp sequence of qKDR1 is a key determinant for KDR variation, as knocking out this sequence resulted in a reduced KDR in maize.
qKDR1 acts as a silencer An intriguing question is, as a non-coding sequence, how does the 1,417-bp sequence function. To answer this question, we performed transient transcriptional activity assays in maize protoplasts. The qKDR1 sequences from the two NILs, qKDR1 2D (a qKDR1 fragment of NIL DAN340 ) and qKDR1 2K (a qKDR1 fragment of NIL K22 ), were independently fused downstream of the b-glucuronidase (GUS) gene driven by a minimal Ubiquitin promoter (mpUbi) (named pUbi2D and pUbi2K) and using the firefly luciferase (LUC) gene as an internal control (Figures 2A and 2B). Compared with GUS activity of the mpUbi vector (pUbiVector), the activities of pUbi2D and pUbi2K were both significantly reduced. Moreover, the GUS activity of pUbi2K was lower than pUbi2D, demonstrating that GUS expression was reduced by the presence of both qKDR1 alleles, and the inhibitory effect of qKDR1 2K was greater than that of qKDR1 2D (Figures 2A-2C). This suggests that the different qKDR1 alleles act to repress target gene expression. We next fused a series of qKDR1 fragments with different lengths downstream of the GUS gene (named pUbi3D and pUbi3K-pUbi5K) (Figures 2A and 2B). Transient transcriptional activity assays indicated that all of them displayed a similar repressive activity to pUbi2K (Figures 2A-2C), but pUbi2D led to higher GUS activity. These results suggest that a 369-bp sequence of qKDR1 is the core silencer sequence and that the 6,181-bp insertion in NIL DAN340 reduces the silencing activity.
To identify the gene(s) regulated by qKDR1, we performed RNA sequencing (RNA-seq) on four NILs harboring different qKDR1 genotypes (Figure 2D). We rationalized that genes that were differentially expressed according to the genotypes of qKDR1 would be candidate target genes (Figures 2E and 2F). We identified 17 genes, of which three were upregulated and 14 were downregulated (Figures 2E and 2F; Table S4). Among these genes, only one, which we named RPG, was close to qKDR1, being located $10-kb upstream (Figure 1C). Consistent with the KDR phenotype, RPG was highly expressed in maize kernels after 30 days after pollination (DAP) (Figures S1D, S4A, and S4B). Real-time quantitative PCRs (qPCRs) revealed that RPG expression was lower in NIL K22 than in NIL DAN340 in the late stage of maize kernel maturation (Figure 2G). In the diverse inbred line population, we also found that the RPG expression in Hap2 and Hap5 (with indels in qKDR1) was significantly higher than (I) The five haplotypes with different qKDR1 genotypes in the diverse maize inbred line population showed significant phenotypic differences by using GLM (with population structure) and MLM (with population structure and kinship). Association studies were performed with general linear model (GLM) or mixed linear model (MLM) implemented by TASSEL (trait analysis by association, evolution, and linkage). [51][52][53] ANOVA, one-way ANOVA in (H). Data are represented as mean ± SEM. **p < 0.01, ***p < 0.001 (Student's t test). n is the sample size, a sample represents the phenotype from a plant (C, E, and F) or from a plot (average of multiple plants of a row) (H). See also Tables S1-S3.
that in Hap1 (without indel in qKDR1) (Figure 2H). As expected, RPG expression was also higher in both of the qKDR1 knockout lines (Figure 2I). Furthermore, chromatin interactions inferred from H3K4me3 and RNA polymerase II (RNAPII) occupancy 56 were also detected between RPG and the qKDR1 locus (Figure 2J). (J) Chromatin interaction analysis by paired-end-tag sequencing (ChIA-PET) between qKDR1 and RPG. Chromatin interactions inferred from H3K4me3 and RNA polymerase II (RNAPII) occupancy were detected between the RPG and qKDR1 loci. The bottom panel shows gene expression (RNA-seq). Light gray columns represent the locations of qKDR1 and RPG. Data from Peng et al. 56 Data are represented as mean ± SEM. *p < 0.05, **p < 0.01, NS, not significant (Student's t test). n is the biological replicate size, a biological replicate is an independent maize protoplast transformation experiment (C) or is from a plant sample (maize kernels at the late maturation stage) (G-I). See also Figure S1 and Tables S4 and S5.
To further confirm that RPG is regulated by qKDR1, we fused the 883-bp RPG promoter upstream of the firefly LUC gene (named pRPGVector), and the Renilla luciferase (REN) gene was used as an internal control. The qKDR1 fragments of different lengths from NIL DAN340 or NIL K22 were also fused downstream of the LUC gene (named pRPG3D and pRPG3K-pRPG5K) (Figures S1A and S1B). Compared with the RPG promoter-driven LUC activity, all of them displayed lower LUC activity (Figures S1A-S1C). This result is consistent with transient transcriptional activity assays of mpUbi-driven GUS activity (Figures 2A-2C). These results thus collectively suggest that qKDR1 functions as a repressor of RPG expression, and the 369-bp region of qKDR1 is sufficient for silencing activity.
Published chromatin immunoprecipitation sequencing (ChIPseq) data 57 showed that 43 transcription factors can bind to the qKDR1 region, two of which are annotated as transcriptional repressors (Table S5, Zm00001d000184, MYB-related-transcription factor 31, named ZmMYBST1 and Zm00001d029875, MYB-related-transcription factor 43, named ZmMYBR43), consistent with the role of qKDR1 acting as a silencer. ZmMYBST1 and ZmMYBR43 were expressed in a similar pattern as RPG with sequence similarity of 96% (Figure S1D), 58,59 suggesting that they may affect RPG expression. mRNA in situ hybridization was performed to confirm the expression patterns of ZmMYBST1, ZmMYBR43, and RPG, and the results showed that they were all expressed in the embryo, aleurone layer, and placento-chalazal region (Figure S1E). To evaluate the effects of ZmMYBST1 and ZmMYBR43 on the qKDR1-regulated expression of RPG, we performed transient transcriptional activity assays, in which the coding sequence of ZmMYBST1 and ZmMYBR43 driven by the CaMV 35S promoter were used as effectors, and pRPG3D and pRPG3K were used as the reporters, respectively (Figure S1F). Compared with the control, coexpression with ZmMYBST1 or ZmMYBR43 resulted in strongly repressed LUC activities of the two reporters (Figure S1G). Additionally, double mutants of ZmMYBST1 and ZmMYBR43 generated by CRISPR-Cas9 significantly decelerated KDR compared with wild-type siblings (Figure S1H). These results suggest that ZmMYBST1 and ZmMYBR43 target the qKDR1 region to repress RPG expression in vivo.
RPG encodes a functional micropeptide RPG was not annotated as a gene in maize (B73 version 4.0). However, high-depth RNA-seq reads mapped to and were presumably transcribed by the RPG region (Figure 3A). We thus performed rapid-amplification of cDNA ends (RACE), including both 5 0 -RACE and 3 0 -RACE, and found two full-length RPG transcripts, named T01 and T02, with lengths of 2,013 and 1,723 bp in NIL K22 (Figure 3B; Table S2). Next, to identify possible protein(s) encoded by RPG, we screened putative ORFs of the two transcripts using ORF finder (https://www. ncbi.nlm.nih.gov/orffinder) and identified twelve putative ORFs (Figure 3C), with the largest encoding a micropeptide of 58 amino acids. Small RNA-seq (sRNA-seq) revealed that RPG likely does not function by producing sRNAs (Figure 3D). However, ribosome profiling sequencing (Ribo-seq) revealed that RPG mRNA was ribosome bound in the ORF1, ORF2, and ORF3 re-gions, suggesting that it encodes one or more micropeptides (Figure 3E).
To assess whether RPG functions by encoding micropeptides rather than a lncRNA, we overexpressed the full-length RPG or a mutant RPG with all 12 ATG start codons mutated to CTG (Figure S2E). Overexpression of wild-type RPG resulted in a slower KDR (Figures S2A-S2D), whereas the non-translatable mutant RPG did not alter the KDR (Figures S2E-S2H). These results imply that RPG functions by encoding micropeptides rather than as a lncRNA.
We found two natural deletions in the RPG region in a diverse maize inbred line population. 55 These removed most of the ORF2 and ORF3 sequences, and the different haplotypes had no obvious KDR differences (Figures S2I-S2L). We also knocked out ORF2 and ORF3 by a 40-bp deletion induced by CRISPR-Cas9 (Figure S2K), and these lines had similar KDR compared with wild-type siblings (Figures S2M and S2N). By contrast, three frameshift mutations in ORF1 with 1-, 2-, or 4-bp deletions (Figures 3F and 3G) all led to a faster KDR (Figures 3H-3K). These findings indicate that ORF1 encodes the functional RPG micropeptide. We next evaluated the effect of ORF1 on the kernel moisture content across different environments. At harvest, the kernel moisture content of ORF1 knockout lines decreased between 1.8% and 17.0% (average of 7.1%) in different environments compared with wild type (Figures 3L-3O). Moreover, the microRPG1 knockout lines had no obvious effects on other agronomic traits in three different environments (Table S6). These results suggest that editing of RPG ORF1 has potential for breeding maize with higher KDRs.
To further investigate the influence of ORF1 on the KDR phenotype, we expressed ORF1 and ORF1m (in which the ORF1 start codon was mutated to CTG) fused with FLAG tags under the control of the Ubiquitin promoter, in maize inbred line B104 (Figure 4A). The wild-type ORF1 construct, which expresses the ORF1 micropeptide, significantly decelerated KDR (Figures 4B, 4C, S3A, and S3B). By contrast, expression of the ORF1m construct did not alter the KDR (Figures 4D, 4E, S3C, and S3D). These results provide additional evidence that the ORF1 micropeptide product of the RPG gene affects maize KDR.
To further confirm that the ORF1 was translated in maize, we fused a GFPm ORF, lacking a start codon to the C terminus of ORF1, and driven by the 883-bp RPG promoter in protoplasts (Figure 4F). Indeed, we observed fluorescence of the ORF1-GFPm fusion protein in transformed cells (Figure 4G). However, mutation of the ORF1 start codon from ATG to ATT abolished the expression of the ORF1-GFPm fusion protein (Figure 4G). We used anti-GFP antibodies in western blotting to confirm expression of the ORF1-GFPm fusion protein, and expression was absent when the ORF1 ATG was mutated (Figure 4H). Similar results were obtained when we used the CaMV 35S promoter in place of the RPG promoter (Figures S3E and S3F).
We next asked if we could detect the endogenously produced RPG ORF1 micropeptide. We probed western blots of total plant protein and different concentrations of the synthetic ORF1 micropeptide (ORF1p) using an anti-ORF1 antibody. Indeed, we detected the ORF1 micropeptide of the expected size in protein extracts from maize kernels (Figure 4I). The endogenous ORF1 micropeptide was also verified by immunoprecipitation (IP) (Figure S3G) and identification by mass spectrometry (MS) (Figures 4J and S3H-S3J). These results indicate that the ORF1 micropeptide is produced endogenously in maize kernels. Thus, we named this 31 amino acid residue micropeptide as mi-croRPG1 (micropeptide of RPG ORF1) (Figure 4J). microRPG1 appears not related to any known peptides in public databases. S2 and S6.
To ask how microRPG1 functions, we performed RNA-seq using microRPG1 knockout and overexpression lines compared with wild type as well as Arabidopsis treated with microRPG1. 384 genes were significantly downregulated in the microRPG1 knockout line and upregulated in microRPG1 overexpression line and exogenous micropeptide application, and 121 genes showed an opposite trend (Figures 5A and 5B). RPG is highly expressed in the late stage of maize kernel development (after 30 DAP) when the dry matter has been completely filled (Figures S1D, S4A, and S4B), implying that microRPG1 may function only during specific developmental periods. Since RPG is specifically expressed in the late stage of kernel maturation, we screened for differentially expressed genes with similar expression patterns to RPG by Z score (R2). As a (C) 11 genes specifically expressed in the late stages of seed (kernel) and endosperm displayed similar expression pattern with RPG in seed, endosperm, and embryo. Among them, Zm00001d047563 showed the highest expression level. The expression levels of 11 genes in various tissues from B73 by RNA-seq. 58,59 (D) Phylogenetic tree of EIN3/EIL in maize, rice, and Arabidopsis. (E-H) The expression levels of EIL genes are regulated by microRPG1. The expression level of ZmEIL1 (E) and ZmEIL3 (F) in microRPG1 knockout (ORF1-KO, 37 DAP kernels) and overexpression (ORF1-OE, 40 DAP kernels) lines. The expression level of AtEIN3 (G) and AtEIL1 (H) in silique (Arabidopsis ripe silique, 72 days after sowing) treated with water (control) or 2.0 mM of ORF1p. The expression levels of EIL genes were quantified using qPCR and normalized to maize or Arabidopsis ACTIN (n R 3).
(legend continued on next page) result, 11 genes (Figure S4C) had similar expression pattern to RPG in seed, endosperm, and embryo (Figure 5C), and these genes are more likely to be regulated by microRPG1 and play a similar role in the late stage of maize kernel development. Among them, Zm00001d047563 an ortholog of Ethylene-Insensitive3/Ethylene-Insensitive3-like 1 (EIN3/EIL1), a key component in ethylene signaling in Arabidopsis and rice (Figure 5D), [60][61][62][63][64][65] had the highest expression (Figure 5C). Zm00001d047563 has four paralogs in maize; however, only Zm00001d047563 (named ZmEIL1) and Zm00001d028974 (named ZmEIL3) were expressed in the kernel and had similar expression pattern to RPG (Figures S1E and S4D). To confirm the expression pattern of microRPG1 micropeptide, we performed immunohistochemistry. The results showed that it was expressed in the embryo, aleurone layer, and placentochalazal region (Figure S4E), consistent with the mRNA expression of RPG (Figure S1E). qPCR confirmed that ZmEIL1 and ZmEIL3 were upregulated in the microRPG1 knockout and downregulated in the microRPG1 overexpression lines (Figures 5E and 5F). This suggests that microRPG1 expression in the late stage of maize kernel development inhibits ethylene signaling, resulting in slower dehydration. Similarly, the Arabidopsis paralogs EIN3 and EIL1 were downregulated following exogenous application of microRPG1 (Figures 5G and 5H). These results indicate that microRPG1 represses the expression of EIN3-related genes in both maize and Arabidopsis.
To investigate the influence of ZmEIL1 and ZmEIL3 on the KDR phenotype, we next generated two knockout lines for ZmEIL1 and three for ZmEIL3 by CRISPR-Cas9 (Figures S4F and S4G). All of the knockout lines significantly decelerated KDR compared with wild-type sibling plants (Figures 5I and 5J). We also investigated the effects of ZmEIL1 or ZmEIL3 knockout on kernel quality traits and seed germination rate and found that knockout of EIL genes affected germination rate significantly and kernel quality slightly (Figures S5A-S5L). These results suggest that EIL genes play important roles in plant development and that the fine-tuning of their expression by microRPG1 can be beneficial for agricultural production.
Considering that EIN3/EIL1 are key components in ethylene signaling, we also applied exogenous ethylene to different maize lines and found that this could accelerate KDR (Figures 5K-5M). Furthermore, we tested five marker genes for ethylene signaling 62,[66][67][68][69][70][71][72] in microRPG1 knockout and overexpression lines and following exogenous micropeptide application. We found that these marker genes were downregulated when microRPG1 was overexpressed or exogenously applied and upregulated when microRPG1 was knocked out, supporting the idea that microRPG1 regulates ethylene signaling (Figures S5M-S5O). Our results indicate that ethylene signaling can affect KDR, suggesting a possible downstream mechanism for microRPG1 in controlling this trait. This finding is consistent with the well-known role of ethylene in stimulating fruit ripening. [71][72][73] Origin of microRPG1 micropeptide To investigate the origins of the microRPG1 micropeptide, we analyzed its syntenic regions in the Poaceae. We found homologous sequences in the genus Zea and Tripsacum but absent from other members of the Poaceae. Sequence alignments indicated that the 96-bp microRPG1 sequence was highly conserved in all maize and teosinte genomes (Table S2). Although similar sequences were present in Tripsacum dactyloides, they lacked a start codon, suggesting they are not translated. Phylogenetic analysis showed that a nucleotide (ACG to ATG) mutation led to the evolution of the ORF1 micropeptide (Figures 6, S6A, and S6B). These observations, together with the above finding that no known peptide associations exist in public databases, suggest that the microRPG1 micropeptide originated de novo from a non-coding sequence after the recent split between the genus Zea and Tripsacum $0.65 million years ago (mya) (Figure 6). 45,[74][75][76][77][78] microRPG1 functions similarly in Arabidopsis Given our demonstration that the microRPG1 micropeptide is able to control the ancient seed dehydration pathway in maize, we wondered whether it could exert similar effects in distantly related plant species. We therefore investigated whether exogenous application of synthetic ORF1p would function in Arabidopsis (Col-0). We applied different ORF1p concentrations ranging from 0.01 to 2 mM. Remarkably, silique ripening was significantly delayed at peptide concentrations of 1 and 2 mM (Figures S7A and S7B). A 2 mM concentration of ORF1p was selected for additional studies that again delayed silique ripening compared with plants treated with a control peptide (scORF1p), with the same amino acid composition but a scrambled sequence (Figures 7A-7E). Application of ORF1p significantly increased the moisture content of Arabidopsis seeds, but there was no significant difference in flowering time (Figures 7F and 7G). This result is similar to microRPG1 overexpression in maize, which leads to a slower KDR. To confirm this result, we generated three microRPG1 overexpression lines in Arabidopsis, and all of them significantly delayed silique ripening (Figures S7C and S7D). Our results indicate that microRPG1 micropeptide functions similarly in the distantly related Arabidopsis species as in maize.
The peptide treatment experiment suggests that ORF1p can enter cells, presumably by crossing the plasma membrane, echoing the behavior of some reported peptides. 79,80 To test this idea, we labeled ORF1p with fluorescein (FAM, carboxyfluorescein) (FAM-ORF1p) and incubated it with Arabidopsis seedlings. Interestingly, the fluorescein-labeled ORF1p rapidly (I and J) ZmEIL1 (I) and ZmEIL3 (J) knockouts had slower KDR in field trials performed in Jilin in 2022 (22JL). (K-M) Ethylene treatment resulted in a faster KDR in F1 of SK 3 KN5585 (K), in F1 of SK 3 Zheng58 (L), and in KN5585 (M). SK, KN5585, and Zheng58 are maize inbred lines. Pre-tre and post-tre represent pre-treatment and post-treatment, respectively. Data are represented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, NS, not significant (Student's t test). n is the biological replicate size, a biological replicate is from a plant sample (E-H). n is the sample size, a sample represents the phenotype from a plant (I-M). See also Figures S4 and S5. penetrated into the root ($0.5 h) but took longer to penetrate other parts of the seedling (Figures S7E and S7F). At 24 h after application, FAM-ORF1p was detected in most parts of the root and began to appear in the cotyledons (Figures 7H and S7E). FAM-ORF1p was detected in leaf veins (Figures 7H and S7G), suggesting that it is transported to the shoot tissues. To address whether ORF1p acts intracellularly in Arabidopsis rather than through cell surface receptors, we observed the subcellular localization of FAM-ORF1p. We found that FAM-ORF1p could enter root cells and localized in the nucleus, cytoplasm, and plasma membrane (Figures S7H and S7I), suggesting it may function through different mechanisms. These results were confirmed by incubation of FAM-ORF1p with maize protoplasts (Figure S7J). We also fused ORF1 with GFP and transformed it into maize protoplasts, and observed fluorescence in the nucleus, cytoplasm, and plasma membrane (Figure S7K). Taken together, these data suggest that the ORF1p micropeptide can enter cells and plays a similar role in Arabidopsis and maize, implying conserved function across plant species.
In this study, we identified and cloned qKDR1, a QTL for KDR in maize, and showed that it corresponds to a non-coding, intergenic DNA sequence. Two MYB-related-transcription factors bind to the qKDR1 region to regulate the expression of its putative target-regulated gene, RPG. RPG encodes a functional micropeptide of 31 amino acids, and this micropeptide alters KDR by regulating the expression of two key ethylene signaling genes (Figure 7I). We thus identified a Zea genus-specific micropeptide and constructed a possible regulatory pathway for maize kernel dehydration, providing useful insights and research directions for our in-depth understanding and applications of seed dehydration in plants.
Non-coding DNA sequences located in intergenic regions can function either as enhancers or silencers and may be bound by transcription factors to regulate the expression of genes located up to megabases away. 81,82 Several such regulatory elements have been functionally characterized in plants, including an enhancer of teosinte branched1 (tb1), 83 Vegetative S2.
to generative transition1 (Vgt1), 84,85 Kernel Row Number4 (KRN4), 86 and Upright Plant Architecture2 (UPA2) 87 in maize, as well as enhancer of Grain Width5 (GW5) 88 and silencer of Frizzy Panicle (FZP) 89 in rice. Here, we cloned a non-coding DNA sequence, qKDR1, that is located $10 kilobases downstream of RPG and can repress RPG expression. However, we have not yet revealed the molecular mechanism by which the qKDR1 region interacts with the RPG promoter, and this is a priority for future research. Here, we also identified a Zea genusspecific micropeptide encoded by the RPG locus. Although the core DNA sequences are also present in Tripsacum, a single nucleotide mutation (C to T) occurred exclusively in the genus Zea, creating a start codon that allowed translation of the functional microRPG1. There are various mechanisms for the origin of new genes, [41][42][43] but it is rare to report that a non-functional sequence becomes a functional gene through a single-base mutation, making this an excellent example to understand the origin of new genes.
The maize kernel moisture content suitable for mechanized harvesting ranges from 15% to 25%, but the moisture content of most maize varieties in China and many other global regions typically ranges from 30% to 40% at harvest. 1,[10][11][12] We found that microRPG1 knockouts decreased the moisture content at harvest by an average of 7.1% and up to 17.0%, depending on the environment and harvest time. Our findings suggest that manipulating microRPG1 to alter the KDR in maize may have great potential in future breeding for mechanized harvesting, since almost all maize lines containing the micropeptide. The microRPG1 knockout lines had no obvious effects on other agronomic traits in three different environments (Table S6). Interestingly, we found that microRPG1 alters KDR by regulating the expression of ZmEIL1 and ZmEIL3, and similar KDR phenotypes were observed in knockouts of ZmEIL1 or ZmEIL3, or following exogenous ethylene application to maize kernels. This hints that the ancient ethylene signaling pathway is involved in maize kernel dehydration. The knockout of EIL genes showed some adverse phenotypes, but microRPG1 knockouts did not. This suggests that fine-tuning of ZmEIL1 and ZmEIL3 expression by microRPG1 has obvious application value (Figures S5A-S5L; Table S6). By precisely regulating the spatiotemporal expression of additional genes in this pathway, we might find other ways to control maize kernel dehydration and eventually pave the way for greater understanding of the underlying mechanisms. The finding that exogenous application of microRPG1 invokes similar responses in Arabidopsis suggests that its utility may reach far beyond maize. Further exploring whether this micropeptide plays a role in other species and how it works will be our research focus in the coming years.
Limitations of the study qKDR1 is a non-coding regulatory sequence located in an intergenic region downstream of RPG; however, how the qKDR1 region interacts with the RPG promoter remains to be fully explored. The molecular mechanisms of microRPG1 have not been fully elucidated. Micropeptides may act intracellularly, or function through cell surface receptors. We found that mi-croRPG1 can be taken up by cells and localizes to the nucleus, cytoplasm, and plasma membrane, suggesting that it may function in multiple ways. In addition to experiments mentioned above, the discovery of intracellular targets or cell surface receptors of microRPG1 remains to be achieved.
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jianbing Yan (yjianbing@ mail.hzau.edu.cn).
Plasmids, transgenic plants, and unique/stable reagents generated in this study will be made available on request for scientific research, but we may require a payment and a completed materials transfer agreement for commercial purpose.
d All data are available in the main text and the supplemental information or at public databases. Sequence data were deposited in Genome Sequence Archive (GSA) 90 of the National Genomics Data Center 91 (https://ngdc.cncb.ac.cn/gsa), with accession numbers GSA: CRA013670, CRA013671, CRA013672, CRA013673. Experimental data of genetic mapping, fine mapping, or agronomic traits were depos-ited in Zenodo: https://doi.org/10.5281/zenodo.13924714. These accession numbers are also listed in the key resources table. STAR+METHODS KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Rabbit Polyclonal anti-ORF1 ABclonal This paper Rabbit Polyclonal anti-GFP Proteintech Cat # 50430-2-AP; RRID: AB_11042881 Mouse Monoclonal anti-actin ABclonal Cat # AC009; RRID: AB_2771701 Rabbit control IgG (IgG) ABclonal Cat # AC005; RRID: AB_2771930 Mouse Monoclonal anti-Rabbit IgG Antibody (M205) HRP GenScript Cat # A01827 HRP Goat Anti-Mouse IgG (H+L) ABclonal Cat # AS003; RRID: AB_2769851 HRP Goat Anti-Rabbit IgG Servicebio Cat # GB23303; RRID: AB_2811189 Anti-Digoxigenin-AP, Fab fragments Roche Cat # 11093274910; RRID: AB_514497 Bacterial and virus strains Trans1-T1 Phage Resistant Chemically Competent Cell TransGen Biotech Cat # CD501 Chemicals, peptides, and recombinant proteins Protease Inhibitor Cocktail MCE Cat # HY-K0010 RNase-free DNase I Huayueyang Biotech Cat # ZH0146 MUG Sangon Biotech Cat # A602251 Ethrel Coolaber Cat # CE5121 FM4-64 Thermo Fisher Scientific Cat # T13320 DAPI Thermo Fisher Scientific Cat # 62248 ORF1p This paper N/A scORF1p This paper N/A FAM-ORF1p This paper N/A Critical commercial assays Huayueyang plant RNA extraction kit Huayueyang Biotech Cat # ZH120 EasyScript one-step gDNA removal and cDNA synthesis supermix TransGen Biotech Cat # AE311 AceQ qPCR SYBR green master mix Vazyme Cat # Q111 SMARTerâ RACE 5'/3' kit Takara Bio USA, Inc Cat # 634858 BCA protein assay kit Sangon Biotech Cat # C503021 Tricine-SDS-PAGE gel preparation kit Sangon Biotech Cat # C641100 PierceÔ classic magnetic IP/Co-IP kit Thermo Fisher Scientific Cat # 88804 Dual-Luciferaseâ reporter assay system Promega Cat # E1980 SP6/T7 transcription kit Roche Cat # 10999644001 Deposited data sRNA-seq and Ribo-seq data This paper GSA: CRA013670, CRA013671 RNA-seq data This paper GSA: CRA013672, CRA013673 Experimental data of genetic mapping, fine mapping or agronomic traits This paper Zenodo: https://doi.org/10.5281/ zenodo.13924714 Experimental models: Organisms/strains Arabidopsis thaliana Col-0 Lab stock N/A Arabidopsis thaliana Col-0 p35S::ORF1-33HA/WT This paper N/A Maize inbred line B73 wild-type This paper N/A Maize inbred line B104 wild-type This paper N/A Maize inbred line B104 qkdr1 This paper N/A Maize inbred line Zheng58 wild-type This paper N/A (Continued on next page) EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS Maize and growth conditions Maize (Zea mays L.) seeds of qKDR1, microRPG1, ZmEIL1 and ZmEIL3 knockout lines and corresponding wild types were germinated on wet paper in 13 cm 3 13 cm Petri dishes for seed germination rate assays. Seeds were grown in plant growth chamber at 26 C and 70% relative humidity, with a photoperiod of 16-h-light (10000 lx)/8-h-dark. Etiolated seedlings of maize inbred line B73 were sown in soil nutrition bowls and grown in plant growth chamber at 26 C, 70% relative humidity, and in the dark. Field-grown plants were grown and cultivated in the experimental field in Hainan (Sanya; 109.19 E, 18.38 N), Hubei (Wuhan; 114.32 E, 30.58 N), Henan (Xinxiang; 113.81 E, 35.20 N), Liaoning (Shenyang; 123.47 E, 41.68 N), Jilin (Gongzhuling; 124.83 E, 43.51 N) or Beijing (39.9 N, 116.3 E), China.
Arabidopsis thaliana Col-0 seeds were surface sterilized with 75% ethanol (1 min) and 95% ethanol (5 min), then plated on halfstrength Murashige & Skoog (½MS) medium supplemented with sucrose and agar (Coolaber, Cat # PM10621-307), stratified for at least 3 days at 4 C, and then vertically grown in plant growth chamber at 22 C and 70% relative humidity, with a photoperiod of 16-h-light (10000 lx)/8-h-dark. After growing to 2$4 true leaves, plants were transplanted into soil nutrition bowls to grow.
Plant materials A recombinant inbred line (RIL) population of 201 lines derived from the cross between the maize inbred lines DAN340 and K22 49,50 were used to detect quantitative trait loci (QTLs) for kernel dehydration rate (KDR). A pair of near isogenic lines (NILs) were developed from heterogeneous inbred family (HIF) that was heterozygous at qKDR1 region, NILs homozygous for DAN340 and K22 across the target region were designated NIL DAN340 and NIL K22 , respectively. A total of 497 diverse maize inbred lines 55 were used for haplotype and genotype frequencies analyses. Out of these plant materials, a set of 339 maize inbred lines was used for ORF2 and ORF3 haplotype analyses; a set of 399 maize inbred lines was used for qKDR1 haplotype analyses.
In the field, we used a GE BLD5604 moisture meter to measure the moisture content. In order to eliminate the influence of developmental period on KDR, we controlled that every plot was pollinated on the same day and the pollination date of each line was recorded. The moisture content was measured on the same interval days after pollination for each line and its comparison group. For QTL mapping, 17 plants with the same genotype were planted in one row in the field, five randomly chosen plants were measured in each row, and the average value represented the value of this genotype. For fine mapping and transgenic functional validation, all lines and their corresponding comparison groups were planted in neighboring rows in the field. The transient moisture content was measured in two successive stages before harvest, and the kernels were harvested after the second moisture content measurement, so the second moisture content was the moisture content at harvest (Figure S2O). The kernel moisture content of each plant at each Continued REAGENT or RESOURCE SOURCE IDENTIFIER stage was measured twice, and the average value was taken as the phenotype of the kernel moisture content for this plant at that stage (n = sample size, a sample represents the phenotype from a plant). Kernel moisture content was obtained by measuring the kernels from the middle of the ear on the plant. Then, the area under the dry-down curve (AUDDC) was calculated to quantify the kernel dehydration rate. 21,92 A smaller AUDDC value represents a faster dehydration rate, while a larger AUDDC value represents a slower dehydration rate. To make a fair evaluation, we compared our method with the absolute moisture content obtained by the oven drying method which is typically performed in the lab. The absolute moisture content was calculated as (fresh weightdry weight) / fresh weight. To evaluate the reliability of our method, all mature kernels from the middle of the ear were measured by oven drying method and meter reading (our method, using a GE BLD5604 moisture meter) in two NILs and found that the correlation coefficient (R 2 ) was 0.84, indicating that our method is reliable.
The DAN340/K22 RIL population was planted in a randomized block design across five geographical locations in China: Hainan (Sanya; 109.19 E, 18.38 N) in 2013; Hubei (Wuhan; 114.32 E, 30.58 N), Henan (Xinxiang; 113.81 E, 35.20 N), Liaoning (Shenyang; 123.47 E, 41.68 N) and Jilin (Gongzhuling; 124.83 E, 43.51 N) in 2014. Moisture content of maize kernels was measured at four successive stages: 40, 46, 52 and 58 days after pollination (DAP), the area under the dry-down curve (AUDDC) is calculated to quantify the kernel dehydration rate. The best linear unbiased predictor (BLUP) value for each line was calculated using the mixed liner model to eliminate the influence of environmental effects. The BLUP values for each line were used as the phenotype to perform QTL mapping. QTL mapping was conducted using composite interval mapping implemented in Windows QTL Cartographer version 2.5 (https:// brcwebportal.cos.ncsu.edu/qtlcart/WQTLCart.htm). 93,94 Zmap (model 6) was used to detect QTL throughout the genome with a walking speed of 0.5-cM, and the default values for background controls of 5 for control markers, 10 for window size (cM) and forward regression method for regression method. A threshold for significant QTLs was determined by 1,000 permutations at a significance level of P < 0.05. The confidence interval for each QTL position was calculated using the 2-LOD drop of the peak. For qKDR1, the QTL confidence interval was 55.61 cM $63.11 cM on Chromosome 1, corresponding to 16.75 Mb $20.03 Mb.
To fine-map qKDR1, heterozygous individuals derived from HIFs (Heterogeneous Inbred Families) were planted to screen new recombinant events. 54 The heterozygous region of HIFs for qKDR1 was 16.81 Mb $23.62 Mb on Chromosome 1, and the HIFs did not detect QTLs in other heterozygous intervals except for the qKDR1 interval. We detected HIF parents in heterozygous regions except the qKDR1 interval, and then self-pollinated and selected the offspring that were homozygous in other regions for fine mapping. Identification of new recombinants using flanking markers in the qKDR1 region, and new molecular markers were developed to determine the breakpoints of identified recombinants. For each new recombinant, progeny tests were performed by comparing the KDR of NIL DAN340 and NIL K22 homozygous individuals from F3 families. 95 The significant difference of KDR between NIL DAN340 and NIL K22 homozygous individuals were compared using Student's t test. If NIL DAN340 and NIL K22 homozygous individuals show significant difference in KDR, the parental recombinant was assumed to be heterozygous for the target QTL; otherwise, the parental recombinant was homozygous. The QTL was narrowed down to a 1417-bp non-coding region (corresponding to B73 reference sequence version 4.0, Chr1:20007756-20009147) by integrating the QTL mapping information of all recombinants. The sequences of the primers used in fine mapping of qKDR1 in this study are listed in Table S7.
Total RNA was extracted from different maize tissues, including maize kernels (37DAP for microRPG1 knockout line, 40DAP for mi-croRPG1 overexpression line and four NILs with different qKDR1 genotypes, 45DAP for three haplotypes with different qKDR1 genotypes from a diverse maize inbred line population and qKDR1 knockout line, and 5DAP to 45DAP for RPG expression pattern of NIL DAN340 and NIL K22 ) and maize leaves (seedlings), or Arabidopsis ripe siliques (72 days after sowing) using the Huayueyang plant RNA extraction kit (Huayueyang Biotech, Cat # ZH120) and treated with RNase-free DNase I (Huayueyang Biotech, Cat # ZH0146) to remove genomic DNA. The first strand cDNA was synthesized by EasyScript one-step gDNA removal and cDNA synthesis supermix (TransGen Biotech, Cat # AE311). qPCR was performed on the real-time PCR (CFX96 Real-Time System, Bio-Rad) by using AceQ qPCR SYBR green master mix (Vazyme, Cat # Q111). The expression levels of RPG, ORF1, ZmEIL1, ZmEIL3, ZmPIF4, ZmEBF1, ZmLZF1, ZmERF1, ZmACO12, AtEIN3, AtEIL1, AtPIF3, AtEBF2, AtLZF1, AtERF1 and AtACO12 were quantified and normalized to that of maize or Arabidopsis ACTIN, respectively. Each group or genotype contained at least two independent biological replicates, where one biological replicate (n = biological replicate size) is derived from one plant sample (maize kernels, maize leaves or Arabidopsis ripe siliques), and each replicate was measured by at least two technical replicates with similar results. Primers for quantifying the expression levels of RPG, ORF1, ZmEIL1, ZmEIL3, ZmPIF4, ZmEBF1, ZmLZF1, ZmERF1, ZmACO12, AtEIN3, AtEIL1, AtPIF3, AtEBF2, AtLZF1, AtERF1 and AtACO12 are listed in Table S7.
Field trials of agronomic, flowering time and yield-related traits Field trials were used to measure agronomic and yield-related traits of two NILs and the gene-edited lines. All NILs and knockout lines and their corresponding wild-types were planted in neighboring rows in the field. For NILs, plants were grown in Hainan (Sanya; 109.19 E, 18.38 N), China, in 2018. All plants were self-pollinated, flowering-related traits were recorded in flowering period, two agronomic traits were investigated after pollination, and five yield-related traits were measured after harvest (Figure 1B; Table S1). For gene-edited plants of qKDR1 and its wild-type plants, plants were grown in Beijing (39.9 N, 116.3 E) and Jilin (Gongzhuling; 124.83 E, 43.51 N), China, in 2022. All plants were self-pollinated, flowering-related traits were recorded in flowering period, two agronomic traits were investigated after pollination, and five yield-related traits were measured after harvest (Table S3). For gene-edited plants of microRPG1 and its wild-type plants, plants were grown in Hainan (Sanya; 109.19 E, 18.38 N) in 2021, Beijing (39.9 N, 116.3 E) in 2022, and Jilin (Gongzhuling; 124.83 E, 43.51 N) in 2022, China, respectively. All plants were self-pollinated, flowering-related traits were recorded in flowering period, two agronomic traits were investigated after pollination, and five yieldrelated traits were measured after harvest (Table S6). N = sample size, a sample represents the phenotype derived from a plant. The large sample size differences between flowering time or agronomic traits and yield-related traits are due to self-pollination failure or poor pollination resulting in fewer plants being harvested in some cases.
Sequence and haplotype analysis of qKDR1 qKDR1 sequences of NIL DAN340 and NIL K22 were cloned through PCR using primers MCK57_3F and MCK57_R. Then, the sequences of NIL DAN340 and NIL K22 were determined using a series of primers (Table S7). Sequence comparisons and analyses were performed in MEGA X and BioEdit. Haplotype analysis of qKDR1 was performed by three pairs of primers (Table S7), of which MCK57_3F and MCK57_R were used to identify the distribution of a 6181-bp insertion (InDel6181), QDR_F and QDR_R were used to identify the distribution of a 50-bp deletion (InDel50), and MCK57_7F and MCK57_R were used to identify the distribution of a 234-bp insertion (InDel234) in the diverse maize inbred line population.
The full-length sequences of RPG in NIL K22 and NIL DAN340 in the late stage of maize kernels (45DAP) were identified by the rapid-amplification of cDNA ends (RACE). RACE experiment was performed using SMARTerâ RACE 5'/3' kit (Takara Bio USA, Inc, Cat # 634858) following the manufacturer's protocol. High quality total RNAs of NIL K22 and NIL DAN340 were extracted and used for the 5'-RACE and 3'-RACE with gene specific primers. The final products of RPG were verified by Sanger sequencing and compared with DNA reference sequence of maize (B73 version 4.0). The gene specific primers used in RACE were listed in Table S7.
RNA and small RNA sequencing Total RNA was extracted from different maize tissues (maize kernels at the late maturation stage at 37DAP for microRPG1 knockout line, 40DAP for microRPG1 overexpression line and NILs) or Arabidopsis tissue (ripe siliques, 72 days after sowing) using the Huayueyang plant RNA extraction kit (Huayueyang Biotech, Cat # ZH120) and treated with RNase-free DNase I (Huayueyang Biotech, Cat # ZH0146) to remove genomic DNA. The total RNAs were used to perform RNA sequencing (RNA-seq) or small RNA sequencing (sRNA-seq). For RNA-seq, the Illumina NovaSeq 6000 and DNBSEQ-T7 platforms were used for sequencing with paired-end reads generation. Cleandata of RNA-seq was obtained and compared to the maize reference genome (version 4.0, http://ftp.ebi.ac.uk/ensemblgenomes/pub/release-48/plants/fasta/zea_mays/dna) by RSEM v1.3.3 96 combined with Bowtie v2.4.1, 97 and the expression levels were calculated with maize annotation information (version 4.48, http://ftp.ebi.ac.uk/ ensemblgenomes/pub/release-48/plants/gtf/zea_mays). Then, the differentially expressed genes were calculated by EBSeq, 98,99 and the differentially expressed genes were selected with significances of P < 0.05 and R 1.2 fold change for NILs, P < 0.05 and R 1.5 fold change for microRPG1 knockout and overexpression, and P < 0.05 and R 3.0 fold change for microRPG1 exogenous application. For sRNA-seq, the Illumina NovaSeq 6000 platform was used for sequencing with single-end reads generation. Cleandata of sRNA-seq was obtained by using Cutadapt, 100 and mapped to the maize reference genome (version 4.0, http://ftp.ebi.ac.uk/ ensemblgenomes/pub/release-48/plants/fasta/zea_mays/dna) by ShortStack 101 combined with Bowtie. 102
In order to investigate whether RPG functions by encoding protein(s), ribosome profiling sequencing (Ribo-seq) was performed in a late stage of maize kernel maturation (40DAP). Library construction and sequencing were performed by Novogene (Novogene Co., Ltd.), the Illumina NovaSeq 6000 platform was used for sequencing with single-end reads generation. Cleandata of Ribo-seq was obtained and mapped to the maize reference genome (version 4.0, http://ftp.ebi.ac.uk/ensemblgenomes/pub/release-48/plants/ fasta/zea_mays/dna) by using TopHat v2.1.1. 103 GFP assay and subcellular localization GFP assay and subcellular localization were performed in maize protoplasts. For GFP assay, the protoplasts were cultured at 23 C in the dark for at least 16 h, and GFP fluorescence was imaged using a laser confocal microscope (Olympus, FV12000MPE). For subcellular localization of FAM-ORF1p in Arabidopsis and maize seedlings, images were collected using a laser confocal microscope (Olympus, FV12000MPE) in the presence of 10 mM FAM-ORF1p after incubation for 24 h. For subcellular localization of FAM-ORF1p in maize protoplasts, images were collected using a laser confocal microscope (Olympus, FV12000MPE) in the presence of 10 mM FAM-ORF1p after incubation for 24 h. For subcellular localization of ORF1-GFP in maize protoplasts, images were collected using a laser confocal microscope (Zeiss, LSM980). The primers used for GFP assay and subcellular localization were listed in Table S7.
The plant tissues were frozen in liquid N 2 and ground into powder. The powder or cells were suspended in IP lysis buffer (Thermo Fisher Scientific, Cat # 87788) or SDS lysis buffer (25 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 2% SDS and 5% glycerol) with 1 mM PMSF and 1% protease inhibitor cocktail (MCE, Cat # HY-K0010), and then sonication was performed for 30 cycles, with 5 s on and 10 s off for each cycle (Diagenode, Bioruptor Plus). The lysates were then centrifuged at 12000 rpm for 10 min at 4 C. The protein concentrations were determined by BCA protein assay kit (Sangon Biotech, Cat # C503021). The total protein for detection of endogenously produced ORF1 micropeptide was extracted by PTM BioLab Co. Inc. For detection of endogenously produced ORF1 micropeptide or ORF1-GFP fusion protein, the proteins were denatured at 96 C for 5 min and separated using 4-20% Tris-Glycine-SDS-PAGE (Sangon Biotech, Cat # C651105) with ColorMixed protein marker (Solarbio, Cat # PR1930) or SeeBlueâ Plus2 Pre-Stained standard (Thermo Fisher Scientific, Cat # LC5925). For detection of endogenously produced ORF1 micropeptide following immunoprecipitation (IP), the tissue lysates were denatured at 96 C for 5 min and separated using Tricine-SDS-PAGE gel preparation kit (Sangon Biotech, Cat # C641100) with Spectra Multicolor low range protein ladder (Thermo Fisher Scientific, Cat # 26628). The proteins were separated by SDS-PAGE, and then transferred onto a polyvinylidene fluoride (PVDF) membrane using a wet transfer apparatus (Bio-Rad Laboratories). Western blotting was performed using anti-GFP (1:3000) or anti-ORF1 (1:3000) antibodies. For the detection of actin in maize, western blotting was performed using anti-actin (1:5000) antibody. The antibodies used in western blotting were anti-GFP (Proteintech, Cat # 50430-2-AP), anti-actin antibody (ABclonal, Cat # AC009) and rabbit control IgG (IgG) (ABclonal, Cat # AC005). The polyclonal antibodies were generated in rabbit against a peptide (HSRAPTNKRAGLQNYC) from ORF1 micropeptide (ABclonal).
The ORF1 complexes in transgenic overexpression lines of full-length RPG (OEWT), transgenic overexpression line of ORF1 (OE) and wild-type plants (WT) were co-immunoprecipitated (co-IPed) using anti-ORF1 with IgG as a control at 4 C for 16 h. Co-IP experiments were performed using a PierceÔ classic magnetic IP/Co-IP kit (Thermo Fisher Scientific, Cat # 88804) following the manufacturer's protocol. Then, the complexes were separated using Tricine-SDS-PAGE for detection of endogenously produced ORF1 micropeptide via western blotting, and used for protein identification using mass spectrometry.
The transient transcriptional activity assays were performed in maize protoplasts. For the GUS-LUC transient transcriptional activity assays, a series of fragments of qKDR1 was generated and fused into the downstream of the b-glucuronidase (GUS) gene driven by a minimal Ubiquitin promoter (mpUbi), respectively. The firefly luciferase (LUC) gene was an internal control. For the dual-luciferase (LUC-REN) transient transcriptional activity assays, the 883-bp RPG promoter was fused into upstream of the firefly luciferase (LUC) gene in pGreenII 0800-LUC vector, and then named pVector. The Renilla luciferase (REN) gene was used as an internal control. The qKDR1 fragments from NIL DAN340 and NIL K22 were fused into the downstream of the LUC gene in pVector, respectively. To evaluate the effects of ZmMYBST1 and ZmMYBR43 on the expression of RPG regulated by qKDR1, the coding sequence (CDS) of ZmMYBST1 or ZmMYBR43 was fused into the downstream of the CaMV 35S promoter in pGreenII 62-SK vector, and used as effectors. The empty pGreenII 62-SK vector was used as a control. The maize protoplasts were collected from the leaves of 10$14-day-old etiolated seedlings of inbred line B73. The GUS signal was collected using 4-Methylumbelliferyl-b-D-glucuronic acid (MUG, Sangon Biotech) as a substrate, and was performed as described previously. 104,105 After transformation and incubation, the protein samples were divided into 3 3 20 mL replicates. GUS reaction buffer (0.03 M Na 2 HPO 4 , 0.02 M NaH 2 PO4, 0.01 M EDTA, 0.1% Triton, 0.07% b-mercaptoethanol, 0.1% MUG) was pre-warmed at 37 C, and 150 mL was added to all replicates at the same time. Then, each mixture was divided into 3 3 50 mL replicates. After incubation at 37 C for 1$5 h, 200 mL 0.2 M Na 2 CO 3 was added to terminate the reaction. The fluorescence signal of the final mixture was detected at 465 nm under excitation at 355 nm. The LUC and REN signals were collected using Dual-Luciferaseâ reporter assay system (Promega, Cat # E1980) following the manufacturer's protocol. Relative GUS activity was calculated by normalizing GUS activity to LUC activity, and relative LUC activity was calculated by normalizing LUC activity to REN activity. The primers used in the transient transcriptional activity assays were listed in Table S7.
The synthetic micropeptides (purity > 95%) were synthesized by Genscript (https://www.genscript.com.cn). The micropeptides were dissolved in water, and stored at À80 C until use. ORF1p: MILLPNHSRAPTNKRAGLQNYCPQRSCIRRG, scORF1p: NTRQCCGANL PQYRRIHLARLRGSMPNPKIS, 5-FAM fluorescein-labeled microRPG1 (FAM-ORF1p).
Arabidopsis seeds were surface sterilized with 75% ethanol and 95% ethanol, then plated on half-strength Murashige & Skoog (½MS) medium supplemented with sucrose and agar (Coolaber, Cat # PM10621-307), stratified for at least 3 days at 4 C, and then vertically grown in plant growth chamber at 22 C and 70% relative humidity, with a photoperiod of 16-h-light/8-h-dark.
Treatment groups adopted half-strength MS medium containing different concentrations (0.01 to 2.0 mM) of synthetic ORF1 micropeptide (ORF1p) or 2.0 mM scORF1p (with the same amino acid composition as the ORF1p but a scrambled sequence), and the control group replaced the micropeptides with water. After growing to 2$4 true leaves, plants were transplanted into soil nutrition bowls to grow.
Treatment groups of Arabidopsis were exogenously watered and sprayed with ORF1p or scORF1p until the siliques ripened, while the control group used water. The date of flowering, and the date when siliques began to ripen and the date when 50% siliques ripening were recorded. After harvesting the seeds, the fresh weight and dry weight after drying at 105 C were measured, and then the moisture content was calculated.
Micropeptide absorption in Arabidopsis 2.0 mM and 10.0 mM FAM-ORF1p was prepared and incubated with Arabidopsis seedlings for 0.5 h$24 h. After the treatment, seedlings were rinsed with ddH 2 O for 5$6 times. The fluorescence of FAM in plants was observed by fluorescence microscope (Nikon, SMZ25). FAM fluorescence was excited with a 492 nm laser and recorded in the 518 nm emission range. ImageJ was used to analyze the fluorescence images.
Synteny analysis of the region encompassing $20 flanking genes on each side of microRPG1 was performed in Poaceae plants (the genomes were from Ensembl Plants (http://plants.ensembl.org/index.html)) except the genus Zea and Tripsacum by using TBtools. 106 Synteny analysis of the region encompassing two flanking genes of microRPG1 was performed in the genus Zea and Tripsacum (the genomes were from maize-pangenome (https://maize-pangenome-ensembl.gramene.org) and MaizeGDB (https:// www.maizegdb.org)). The 96-bp sequence of microRPG1 with 20-bp of 5' and 3' flanking sequence or with 300-bp of 5' flanking sequence and 200-bp of 3' flanking sequence was compared with the syntenic sequences between the two flanking genes by using TBtools (P < 0.01). Bioedit was used to compare the microRPG1 and its homologous sequences in the genus Zea and Tripsacum. The divergence time of species refers to several published studies. [74][75][76][77][78]
The protein files for Poaceae plants in the tree were obtained from Ensembl Plants (http://plants.ensembl.org/index.html) and MaizeGDB (https://www.maizegdb.org). Subsequently, the identification of single-copy orthologous proteins was performed using OrthoFinder v2.2.6, 107 after selecting the longest transcript for each gene. The construction of the species phylogenetic tree was carried out using FastTree v2.1.10. 108 For EIN3/EIL in maize, rice and Arabidopsis, blastp was used to align ZmEIL1 homologous sequences in National Center for Biotechnology Information (NCBI, https://blast.ncbi.nlm.nih.gov) with reference proteins (refseq_protein). The homologous sequences of ZmEIL1 in maize, rice and Arabidopsis were aligned by MUSCLE in MEGA X. The phylogenetic tree was constructed using the maximum likelihood method in MEGA X.
The overlapping differentially expressed genes form microRPG1 knockout, overexpression and exogenous application were used to calculate Z-scores and the genes were considered as specific expression in the late stages of seed (kernel) and endosperm if they had a Z-score of 2 or higher. The Z-scores of the genes expression in 38DAP seed (kernel) and 38DAP endosperm were calculated, and the expression of genes in all tissues (except for the seed and endosperm of 26DAP and 32DAP) of seed, endosperm and embryo from B73 58,59 was used to calculate Z-scores.
The 30DAP kernels of maize inbred line B104 were fixed in a 4% polyformaldehyde solution (4g paraformaldehyde dissolved in 100 mL 1 3 PBS, pH 6.5$7.0) and used for mRNA in situ hybridization. The kernels were dehydrated in a series of ethanol concentrations and embedded, then sectioned to a thickness of 5 mm. To generate probes for ZmMYBST1, ZmMYBR43, RPG, ZmEIL1 and ZmEIL3, T7 promoter sequence (CATTAATACGACTCACTATAGGG) was added into reverse primers and probe fragments were amplified by PCR using gene specific primers listed in Table S7. Digoxigenin-labeled antisense probes were then transcribed using an in vitro transcription kit (Roche, Cat # 10999644001) following the manufacturer's protocol. RNA hybridization and immunologic detection were performed as described previously. 109 The signals were observed by microscope (Nikon, SMZ25).
The 30DAP kernels of maize inbred line B104 were dehydrated in a series of ethanol concentrations and embedded, then sectioned to a thickness of 5 mm. The sections were dewaxed and rehydrated, followed by antigen repair. Next, the tissues were incubated in 3% hydrogen peroxide solution for 25 min at room temperature to block endogenous peroxidase, then blocked in 3% BSA for 30 min at room temperature. The tissues were incubated with anti-ORF1 antibody (1:300) overnight at 4 C and subsequently incubated with the corresponding HRP conjugated secondary IgG antibody for 50 min at room temperature. Cellular nuclei were stained with hematoxylin. Immunohistochemistry was performed by Wuhan Servicebio Technology Co., Ltd.
Maize kernels were dried thoroughly and ground into powder. About 0.2g was weighed into a test tube and the weight recorded. Subsequently, 4.5 mL of sulfuric acid methanol solution was added to each test tube, and the mixture was shaken and mixed well, and then 100 mL internal standard was added. The solution was incubated in a water bath at 90 C for 2.5 h$3 h. The sample was subsequently cooled to room temperature, 2 mL ultrapure water was added and the mixture was shaken well, then 2 mL n-hexane was added and the mixture was shaken well. Following this, the mixture was centrifuged at 3000 rpm for 10 min at room temperature and the supernatant was taken. The oil contents were subsequently determined by liquid chromatography.
Maize kernels were dried thoroughly and ground into powder. About 0.02g was weighed into a test tube and the weight recorded. Subsequently, 1 mL anhydrous ether was added and the mixture was shaken well, then centrifuged at 8000 g for 10 min at room temperature and the supernatant was discarded. Subsequently, 1 mL 80% ethanol was added to the precipitate and shaken well. The mixture was incubated in a water bath at 80 C for 30 min, prior to centrifugation at 8000 g for 10 min at room temperature, after which the supernatant was discarded. Then 0.5 mL of ultrapure water was added to the precipitate and the mixture boiled in a water bath for 15 min to fully gelatinize the sample. After cooling, 0.5 mL of 6 mol/L hydrochloric acid was added and the mixture was heated at 95 C for 30 min, prior to being centrifuged at 8000 g for 10 min at room temperature, with the supernatant taken for measurement. After diluting with ultrapure water, 50 mL of supernatant was taken into a 1.5 mL test tube. 250 mL of anthrone reagent was added and mixed well prior to incubation in a water bath at 95 C for 10 min. After cooling to room temperature, the absorption at 620 nm was detected. The determination of kernel starch content was performed by Wuhan ProNets Biotechnology Co., Ltd.
Maize seeds of each genotype were immersed in water for about 8 h, and then transferred to wet paper in a 13 cm 3 13 cm Petri dish for germination. Seeds were grown in a plant growth chamber at 26 C and 70% relative humidity, with a photoperiod of 16-h-light/8-h-dark. Seed germination was considered complete when the root length was > 2 mm. The number of germinated seeds was counted 88 h after sowing. Each genotype contained 24$36 seeds per biological replicate, and three biological replicates were performed.
The moisture content of maize kernels was investigated before exogenous ethylene treatment (pre-treatment), and then 277 mM ethrel (Coolaber, Cat # CE5121) was prepared and used to treat maize kernels in the late stage of kernel maturation in F1 of SK 3 KN5585, F1 of SK 3 Zheng58, and KN5585 in the field. In detail, the ear husks were peeled back, then 10 mL was sprayed onto the kernels on each ear, followed by covering the ear with a paper bag, and this procedure was repeated after the solution evaporated ($12 hours later). Four days post-treatment, the moisture content of maize kernels was investigated again to evaluate the effects of ethylene treatment.
To generate overexpression lines, the related sequences driven by the Ubiquitin promoter were transformed into maize inbred line B104 at Wimi Biotechnology (Changzhou, China). The transgenic-positive and transgenic-negative plants were identified in each generation, and qPCR was performed to determine the expression levels of transgenes in the overexpression lines. The CRISPR-Cas9 technique was used to generate knockout lines. The target sites were designed and cloned into pCXB053 vector, and then transformed into maize inbred line B104 or KN5585 at Wimi Biotechnology. 110 The knockout lines were genotyped by PCR and sequencing to identify the mutant plants. The knockout lines with homozygous mutations were used for phenotypic analysis and field trials. The primers used in transgenic functional validation were listed in Table S7. All the overexpression and knockout lines and their corresponding wild-type were planted in neighboring rows in field and the moisture content of maize kernels was measured.
Data for quantification analyses are represented as mean ± SEM. Significance of difference was examined by Student's t test (*P < 0.05, **P < 0.01, ***P < 0.001) or one-way ANOVA followed by Fisher's LSD multiple-comparison test. Different letters at top of columns indicate significant differences at P < 0.05. The statistical analyses for all experiments were performed with the GraphPad Prism 9 software (https://www.graphpad.com/scientific-software/prism/). The exact value of n, what n represents and the statistical details of each experiment are described in the figure legends or tables. (legend continued on next page) Figure S7. microRPG1 delays silique ripening, FAM-ORF1p can be absorbed by root cells and detected in leaf veins, and the subcellular localization patterns of microRPG1, related to Figure 7 (A and B) Applying synthetic ORF1p with different concentrations in Arabidopsis (Col-0). Different concentrations of ORF1p (0.01-2 mM) were applied in Col-0. The concentration of 1 and 2 mM of ORF1p significantly delayed silique ripening compared with the control. Silique ripening days were recorded when the siliques began to ripen (A), and when 50% of the siliques were ripe (B). DAS, days after sowing. (C and D) microRPG1 overexpression delayed silique ripening in Arabidopsis. Silique ripening days were recorded when the siliques began to ripen (C), and when 50% of the siliques were ripe (D). DAS, days after sowing. (H) FAM-ORF1 is localized in the nucleus, cytoplasm, and plasma membrane in root cells of Arabidopsis seedlings in the presence of 10 mM FAM-ORF1p. (I) FAM-ORF1 is localized in the nucleus, cytoplasm, and plasma membrane in root cells of maize seedlings in the presence of 10 mM FAM-ORF1p. (J) FAM-ORF1 is localized in the nucleus, cytoplasm, and plasma membrane in maize protoplast cells in the presence of 10 mM FAM-ORF1p. (K) ORF1-GFP is localized in the nucleus, cytoplasm, and plasma membrane in maize protoplast cells. SCAMP1-RFP and GHD7-mCherry were used as markers for the plasma membrane and nucleus, respectively. Scale bars, 30 (H), 20 (I), and 10 mm (J and K). Data are represented as mean ± SEM. Different letters at top of columns indicate significant differences at p < 0.05 (one-way ANOVA followed by Fisher's LSD multiple-comparison test). *p < 0.05, ***p < 0.001 (Student's t test). n = sample size, a sample represents the phenotype from a plant.
Cell 188, 44-59, January 9, 2025 ª 2024 The Author(s). Published by Elsevier Inc.
Cell 188, 44-59, January 9, 2025
Cell 188, 44-59, January 9, 2025
d This study does not report original code. d Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request.
Cell 188, 44-59.e1-e8, January 9, 2025
Cell 188, 44-59.e1-e8, January 9, 2025
Cell 188, 44-59.e1-e8, January 9, 2025 e4
Cell 188, 44-59.e1-e8, January 9, 2025 e8
(legend on next page)